Memoirs of the Queensland Museum 33(2): 389-100

Home , Calapnita, Cambridgea, Carorita, Corythalia, Cotinusa, Crioa

proceedings of the

XII INTERNATIONAL CONGRESS OF ARACHNOLOGY

BRISBANE MEMOIRS OF THE VOLUME 33 1 1 NOVEMBER, 1993 QUEENSLAND MUSEUM PART 2

The preparation for this International Congress was superbly managed by two Organising Committees. All members worked very hard and I am very grateful for their cooperation. The Western Australian Museum was the centre of theScientific and Publications Committee of Dr Mark S. Harvey, Dr Bill R Humphreys, and Dr Barbara York Main inis group, lead by Mark Harvey, formed a fine team who very capably managed all scientific and publication matters from manuscript submission to acceptance through the Congress program and photographic awards.

The local Organising Committee included Dr Valerie Davies (Secretary), and Ms Tracey Churchill, Ms Jan Green 1

y Mr Phl1 LawIess and \7t? k!T myself That group managed diverse aspects of Congress organisation Mr David% Nebauer assisted committee members during the registration process. My preparation was initially assisted by Ms Julie Gallon. In July 1 99 Mr Philip Lawless 1 „ took up the reins from Julie and proved to be a very dedicated and astute lieutenant. Emeritus Professor George Davies assisted the Secretary throughout the entire Congress period.

Delegates from Russia, Kazakhstan, The People's Republic of China, Madagascar, Mexico, Hungary India Belgium, North America, Germany, and Namibia were provided with varying degrees of sponsorship by the Queensland Museum Board of Trustees and sponsors. Their travel arrangements, often highly complex and problem-rich, were very capably managed by Ms Joanne Lavelle, then of State of Corporate Travel. Many a Congress delegate will no doubt fondly remember the Committee reception at the International airport. Several appreciated the kind hospitality of 4 Mr Tony Luck, Microrim (Australia), of Camira, in an Aussie barbeque' before the Congress.

Organisation of the Post-Congress tour to Cape Tribulation, was greatly facilitated by the much appreciated contributions of Ms Linda Stanley, Ms Esther Cullen, and Cliff and Gail Truelove.

Editorial assistance was provided by Ms Gudrun Sanies, Philip Lawless and Mr Neale Hall.

The Congress organisers very gratefully acknowledge the financial donations of Australian Geographic Magazine Western Australian Museum, Amalgamated Pest Control Australian Museum, Museum of Victoria South Australian Museum, Commonwealth Bank, University of Queensland-Entomology & Zoology Departments Griffith University-School of Australian Environmental Sciences, QANTAS Airlines, Australasian Arachnologicai Society, CSIRO, Division of Entomology, and the British Council. Financial support was also received from International Conference Support (loan) and the Australian Tourist Commission. The Queensland Museum Board of Trustees provided generous support for the Congress and the publication of these proceedings.

Robert J. Raven Chairman, Xllth International Congress of Arachnology

October 11, 1993

3SJ Dr Joachim Atfis 80 Prof Malcolm Edmunds So Dr Hubert Hofer 45 DrPelckaUhnnen Ms Rachel Allan 17 DrMarfcElgw 11 DrS. Pidaniuhamy 15 Dr John Taienl 47 MrChnsHolden 52 Mr Peier Lmklaict Di Andre* Austin '2 Mr 1H Dr Ralph Platen 115 t*rKoithiT3n;tk.i Theodore Evans MO Dr Pdtet Honlfc 49 Dr Adam Locket 28 Prof Drf^onflaen 19 DrrVterFaipAealrer Norman Pfdinuk 111 DrOurigjiTarabacv M9 Dr Gustavo Honruga 2 1 Mr Alan J^ingbotinm 33 Dt Gary MrAlhertn Barrinn 44 MrRicriardFautifcr Toll", fib Mr Phil Taylor 33 Dr Peter Hudson 32 MrRt-ger towe 35 Dr Simon Pollard DrTiinoihy Beaton 81 !> Martin Fthncr 40 Ms Margaret Tin 1 1« Di Bill Humphrey; 42 DrVolkerMahncri 22 Dr MsEmilvBoIith-j 35 DtLynEontcr William Preston lift DrNohun Tsurusaki 117 Dr Glenn Hum 30 Prof Dr Brady A.R.tBcrt} Main 1)4 Dr Robert R,iven Allen 5A DrRd/Eoryet 58 MS Hch Hurmc 3ft Dr G»t>riele L'W Ms 43 Dr Barbara York Main J20 Dr George Roderick Cathy Car 48 Dr Rosemary Gi|1cipi« Drlgnaiio Vazquez 64 Dr Robert lackson 106 tames Dr Patrick Marc 27 Dr Christine Rolled Dr Canco 1 Dr Lenny Ms Maureen Glover 28 Dr Rud> Jocoue Virtcenl 105 Ms Sylvw Marc-Mallet 34 Dr Jerome R inner Mf Kefyn Catley % DrMfchaclGray Mr Cor Vinlt M MsSirithtCAriko 75 Mr Roland Mckay M*Tracey Churchill 38 16 Mrs Phyllis Rovner DrFrilfVulltttth Ms Jan Gitten Mr Joseph KToh 57 MrGrahmiMilledsc 77 DrJon CuddinEton 12? Ms Judy Crirmbaw MrFetertc Samu Ms Julianne Waldock 6t Dr AJviu Krwuki 121 Dr Gary Miller Dr Valerie Davits 68 Dr Peter Sihucndin^cr Mr Doug 82 Dr Charles Gnswoid 101 Dt Peter Ktotneo Wallace 107 MnPatncit Milter 90 Dt Paul 1 1 2 Prof George Davies Seldcn Dr Fnedrich 6 DrCrisunaGuerrera-Trejo 60 Dr Seppo Koponen Wallemiciri 91 WillixmMiller DtGjuJ V DrChrisultetleman 62 Mr Johanna* HenseheJ 111 Dr Oirwun Kmpf DrJorgWunderl.cn 67 Dr Hifckuga Ono U Dr Robert Suttr 11 Dr lanct Edmunds S3 Mr Dav.rt Hirst 104 Dr Frederic Ysnet 76 Mr Philip LavlCKh 61 Dr VtedjmJrOvtsrtarcnk'j 53 DrHirotiurni Suzuki 1

CONTENTS

PART 1 (Issued 30 June, 1993)

Bruce, N.L. Two new genera of marine isopod crustaceans (Cirolanidae) from Madang, Papua New Guinea J Cannon, L.R.G. New temnocephalans (Platyhclminthes): ectosymbionts of freshwater crabs and shrimps 17 Chimonjdes, PJ. & Cook, P.L. Notes on Pannularia MacGillivray (Bryozoa. Cheilostomida) from Australia ..*.., 41 Cook, A.G. Two bivalves from the Middle Devonian Burdelcin Formation, north Queensland 49 Short, J,W. Cherax cartalacoolah, a new species of freshwater crayfish (Decapoda: Parastacidae) from northeast Australia _ 55

Hooper, J.N.A. ? Kelly-Borges, M. & Riddle, M. Oceanapia sagitraria from the Gulf of Thailand T . . _ 61 Richards, S.J. & Johnston, G.R. A new species of Nyctimystes from the Star Mountains, Papua New Guinea 73 Van Dyck. S. The taxonomy and distribution ofPetaurus gracilis (Marsupialia: Petauridae), with notes on its ecology and conservation status k , 77 DWYER, P.D. The production and disposal of pigs by Kubo people of Papua New Guinea J23 David, B.&Dagg,L. Two caves , . . . . r . , 143 COUPER, PJ, A new species of Lygisaums de Vis (Reptiiia: Scincidae) from mideastem Queensland 163 Davies, M., Watson, G.R, Mcdonald, K.R., Trenerry, M.P. & Werren, G. A new species of Uperoleia (Anura: Leptodactylidae: Myobatrachinae) from northeastern Australia , , 167 Davies, V.T, The spinning field and stridulating apparatus of penultimate male Macrogradungula moonya (Araneae: Austrochiloidca: Gradungulidae) , 175 Hand, S. First skull of a species of Hipposideros(Brachipposideros) (Microchiroptrea: Hipposidendae), from Australian Miocene sediments ,*,.,., 1 79 Hand, S., Archer, m„ Godthelp, H. Rich, t T.H. & Pledge, N.S. Nimbadon, a new genus and three new species of Tertiary zygomaturines (Marsupialia: Diprotodontidac) from northern Australia, with a reassessment oiNeohelos 193

Healy, J,M. t Lamprell, K.L. & Stanisic, J. Description of a new species of Chama from the Gulf of Carpentaria with comments on Pscudochama Odhner (Mollusca: Bivalvia: Chamidae) , 21 Hunt, G.S. A new cavernicolous harvestman from lava lube caves in tropical Australia (Arachnida: Opiliones: Phalangodidac) 217 Ingram, G.J., Nattrass, A.E.O. & Czechura, G.V. Common names for Queensland frogs ..... jj\ Jamieson, B.G.M. Spermatological evidence for the laxonomic status of Trapezia (Crustacea: Brachyura:

Heterotremata) . , , r 225 Jamieson, B.G.M. A taxonomic revision of the eastern Australian earthworm genus Pcrissogaster Fletcher (Megascolccidae: Oligochaeta) 235 Kelly, S.R.A. On the alleged occurrence of the Early Cretaceous ammonite Simbirskites in Queensland 245 Lambkin, KJ. New information on the Australian small bittacids (Mecoptera) 253 Leach, GJ. & Hines. H.B. Frequency of observation of bird species in sub-coastal farmland in southeast Queensland 259 I

I.IMPUS, C.L., Couper, P.J. & Couper, K.L.D. Crab Island revisited; reassessment of the world's largest Flatback Turtle rookery after

twelve years , . , r 277 Paterson, R.A., Quayle, C.J. & Van Dyck, S.M. A Humpback Whale calf and two subadult Dense-beaked Whales recently stranded in southern Queensland 291 Quicke, D.LX & Ingram, S.N. Braconine wasps of Australia ,..,.. , 299 Richards, S.J. & Alford, R.A. The tadpoles of two Queensland frogs ( Anura: Hylidae, Myobatrachidae) ..*... 337 Ingram, GJ., Couper, P.J. & Donnellan, S. A new two-toed skink from eastern Australia < . „ 34] Webster, G.D. & Jell, P.A. Early Permian inadunate crinoids from Thailand _ . , 349 Williams, S., Pearson, R. & Burnett, S. Survey of the vertebrate fauna of the Dotswood area, north Queensland t . 361

Williams, S. ? Pearson, R. & Burnett, S. Vertebrate fauna of three mountain tops in the Townsvillc region, north Queensland: Mount Cleveland. Mount Elliot and Mount Halifax * , . . 379 NOTES Arnold, P. First record of ctcnophore Coeloplana (Benthoplana) meteoris (Thiel, 1968} from Australia 16 Catterall, C.P. A record of the Common Planigalc at Myora Springs: the first dasyurid from North Stradbroke Island 54 Porter, R. & Husband, G. A record of communal egg-laying in the skink Carlia tetradactyla , , . _ 60 Shea, G.M. New records of Lerista allanae (Squamata: Scincidae) ,220 Jbll, P.A. & Lambkin, KJ. Middle Triassic orthopteroid (Titanoplcra) insect from the Esk Formation at Lake Wivcnhoe T 258 SlMPENDORFER, C.A. Pandarid copepods parasitic on sharks from north Queensland waters _.,... 290 Nattrass, A.E.O. & Ingram, GJ. New records of the rare Grecn-thighed Frog , . . , 348 Jell, P.A. Late Triassic homopterous nymph from Dinmore, Ipswich Basin A 360

PART 2 (Issued 11 November, 1993)

PROCEEDINGS OF THE XII INTERNATIONAL CONGRESS OF ARACHNOLOGY INVITED PAPERS Selden, P.A. Fossil arachnids-recent advances and future prospects 389 POLIS.G.A. Scorpions as model vehicles to advance theories of population and community ecology: the role of scorpions in desert communities *,.,,,., 40 Elgar, M.A. Inter-specific associations involving spiders: klcptoparasitism, mimicry and mutualism 411 CONTRIBUTED PAPERS

Adis, J. & Mahnert, V. Vertical distribution and abundance of pseudoscorpions (Arachnida) in the soil of two different neotropical primary forests during the dry and rainy seasons 431 Austin, A.D. Nest associates of Ciubiona robusta L. Koch (Arancae: Clubionidae) in Australia 441 Baert. L. and Jocque, R. A tentative analysis of the spider fauna of some tropical oceanic islands 447 Benton, T.G.

The reproductive ecology of Euscorpiusjlavicaudis in England .. + 455 Canard. A. & Stockman, R.

Comparative postembryonic development of arachnids , 461 Catley, K.M. Courtship, mating and post-oviposilion behaviour of Hypochilus pococki Platnick (Araneae, Hypochilidae) ..„...,..„....*.* 469 Churchill, T.B. Effects of sampling method on composition of a Tasmanian coastal heathland spider assemblage .... .475 Davies, V. Todd.

A new spider genus (Araneae: Amaurobioidca) from rainforesls of Queensland, Australia , 483 Deeleman-Reinhold, C.L. An inventory of the spiders in two primary tropical forests in Sabah, North Borneo .491 Duffey, E. A review of factors influencing the distribution of spiders with special reference to Britain 497

Edmunds, J. The development of the asymmetrical web of Nephilengys cruentara (Fabricius) 503 Edmunds, M. Does mimicry of ants reduce predation by wasps on sal tic id spiders? 507 Fafrweather, P.G.

Abundance and structure of fossorial spider populations * . * ..... > P . .513 GILLESPIE, R.G. Biogcographic pattern of phylogeny in a clade of endemic Hawaiian spiders

(Araneae, Tetragnathidae) , , , , ,...519 Gromov, A.V. A new species of Karschiidae (Solifugae, Arachnida) from Kazakhstan ....,..,..,.,,., 527

HrRSTr D.B.

A new species of AmaurobioidesQ.P ^-Cambridge (Anyphaenidae: Araneae) from South Australia . . . ,529 H0RM1GA, G. Implications of the phylogeny of Pimoidae for the syslematics of Jinyphiid spiders

(Araneae, Araneoidea, Linyphiidae) * - 533 Humphreys, W.F,

. Criteria for identifying thermal behaviour in spiders: a low technology approach . S43 Hunt, G.S. & Maury, E.A. Hypertrophy of male genitalia in South American and Australian Triaenonychidae (Arachnida: Opiliones: Laniatores) 551 Jackson, R.R. & Wilcox, R.S.

Predator-prey co-evolution of Portia fimbriate and Euryattus sp., jumping spiders from Queensland . . .557 JOCQUfe, R. >v "We'll meet again , an expression remarkably applicable to the historical biogeography of

Australian Zodariidae (Araneae) + . . . 561 KONDO, A., CHAKL E. & FUKUDA, M. Basic architecture of the ovary in the golden silk spider, Nephila clavata 565 Koomen, P. & Peeters, T.MJ, New prey records for spider hunting wasps (Hymenoptcra: Pompilidae) from The Netherlands 571

Koponen, S.

Ground-living spiders (Araneae) one year after fire in three subarctic forest types, Quebec (Canada) . . .575 Kraus, O. & Kraus, M. Divergent transformation of chelicerae and original arrangement of eyes in spiders

(Arachnida. Araneae) ,,.,....*..,.,,.,.. v , . 579 LEHTINEN, P.T.

Polynesian Thomisidae - a meeting of old and new world groups . 585 Locket. n.A. Scorpion distribution in a dune and swale malice environment 593 Main, BY. From flood avoidance to foraging: adaptive shifts in trapdoor spider behaviour .599 Marc, p.

Intraspecific predation in Clubiona corticalis (Araneae: Clubionidae) h 607 Muller, M.C. & Westheide, W. Comparative morphology of Ihc sexually dimorphic orb-weaving spider Argiope bruennichi

(Araneae: Araneidae) , , . . , t 515 Platen, r. A method to develop an 'indicator value' system for spiders using Canonical Correspondence Analysis (CCA) 621 ROLLARD, C. The spiders of the high-altitude meadows of Mont Nimba (West Africa); a preliminary report 629 Rovner, J.S. Visually mediated responses in the lycosid spider Rabidosa rabida: the roles of different pairs of eyes , 535 Sunderland, K.D. & Topping, C.J. The spatial dynamics of linyphiid spiders in winter wheat 639 Suzuki, H. & Kondo, A. Morphology of the embryos at germ disk stage in Achaearanea japonica (Theridiidae) and Neoscotia nautica (Araneidae) 545 Tarabaev, C.K.

An experiment on colonization of karakurt {Latrodectustredecimguttaius, Black Widow spider) on island territories in Kazakhstan 55 \ Tarabaev, C.K., Zyuzin, A.A. & Fyodorov, A.A. Distribution of Latrodectus (Theridiidae), Eresus and Stegodyphus (Eresidae) in Kazakhstan

and Central Asia . . * 553 Tsurusaki, N. Geographic variation of the number of B-chromosomes in Metagagrella tenuipes (Opiliones, Phalangiidae, Gagrellinae) , 659 Uhl, G. Mating behaviour and female sperm storage in Pholcus phalangioides (Fuesslin) (Araneae) 667 Wishart, G.F.C. The biology of spiders and phenology of wandering males in a forest remnant (Araneae: Mygalomorphae) 575 Wunderlich, J. The Macaronesian cave-dwelling spider fauna 5gj Ysnel, F. Relationship between food intake and spider size in temperate zones: experimental model for an orb-weaving spider , gg7 Zyuzin, A.A. Studies on the spiders wolf (Araneae: Lycosidae). I. A new genus and new species from Kazakhstan, with comments on the Lycosinae 693 FOSSIL ARACHNIDS-RECENT ADVANCES AND FUTURE PROSPECTS

PAULA SELDI

Selden, P.A. 1993 1 1 11: Fossil arachnids-recent advances and future prospects. Memoirs of the Queensland Museum 33(2): 389-100. Brisbane. ISSN 0079-8835.

Until 5 years ago, the arachnid fossil record was sparse. It was dominated by a comparative wealth of forms in Carboniferous Coal Measure sediments, and near-modern forms from Palaeogene Baltic amber. Both these relatively well -documented sources and the few reported finds elsewhere in the record suffered from erroneous interpretations. In recent years, new interpretations of existing fossils and a few spectacular new finds have filled in the gaps in the record and changed our knowledge and views of the course of arachnid evolution. Particular examples arc: Devonian pseudoscorpions and spiders, book-lungs in Carboniferous scorpions Triassic mygalomorph spiders, and hlj^sstc and Cretaceous araneomorph spiders. Phylogenetic systematic analyses of extant arachnids have produced evolutionary scenarios which conflict wilh the observed fossil record in parts. 'The newly expanded knowledge of the fossa! record allows better tests forthecladograms. Future work on reinterprctation of known Carboniferous and Palaeogene fossils, on rare Mcso/.oic arachnids, and on arachnids in the earliest known terrestrial ecosystems in the Silurian will add to our knowledge of the fossil record of the arachnids and further enhance testing of phylogenetic hypofbcsc&JQAgfa&pliRda, A rachnida, Cheiiceralo. palaeontology, phytogeny, Pycn&gontda.

Paul A. Selden, Department of'Geology. University of Manchester. Manchester M 1 3 91*1+ United Kingdom; W November, 1 992.

For most of this century, one name dominated Coal Measures of Europe and North Amcric; Ihe literature on fossil arachnids, that of example: Mazon Creek, Illinois; Cosclcy. Alexander Petrunkevitch (1875-1964). Hngland; and Nyrany. Czechoslovakia. The Petrunkevitch (1955; in S termer, 1955) sum- Palaeogene occurrences arc mainly from Baltic marized the arachnid fossil record to mid-century amber. Although Trigonotarbida and a ques- (Fig 1) in the Treatise on Invertebrate tionable record of Araneae had been known from palaeontology' and although he published on the Devonian Rhynie Chert of Scotland since amber spiders after 1955. Lhc broad view Of the Hirst ( 1 923), they were omitted from the diagram. fossil record of chelicerates remained little (In addition, Petrunkevitch knew of undescribed changed until about a decade ago. Few workers Lebanese amber opilionids and some Cretaceous either published on fossil arachnids or disputed spiders from Manitoban amber). Petrunkevitch' s assignments during his lifetime. Petrunkevitch developed theories on the evolu- Only recently, during restudy of the fossils, have tion of arachnids, which resulted in his superor many errors and misinterpretations in his work dinal classifications of 1945 and 1949. He come to light. recognized a number of 'evolutionary tre Jn the fossil chelicerate record published in the such as the movement of the mouth rearwards 'Treatise* (Fig. I), the Merostomata (essentially from the Xiphosura to the arachnids, and the aquatic chelicerates) are separated from the reduction of the metasoma to a tail or pygidium Arachnida. Second, most of the arachnid side One of the most important characters used in his consists of dashed lines converging towards the classifications is the width of the connection be- base of ibe Cambrian, indicating lack of fossil tween presume and opisthosoma, i.e. reduction rj] record and uncertainty of affinities respectively the first abdominal somite to a pedicel Third, apart from one dubious palpigrade and Petrunkevitch (1945)divided the class Arachnida some scoqpions, there are no other records of into two subclasses, Latigastra and Caulog. Mesozoic (Triassic-Cretaceous) arachnids. on the basis of a broad or a narrow prosoma-opis- Fourth, there is a clear pattern in the temporal thosoma connection respectively. Later.,

distribution of the fossils, a concentration of Petrunkevitch ( 1949) added the subclass Solum records in the Upper Carboniferous, and many to the scheme to include solely his new order modem groups also occur in the Palaeogene Trigonotarbida which he considered exhibit (early Tertiary) The former records are from the wide and narrow junctions. Another subclass, the 390 MEMOIRS OF THE QUEENSLAND MUSEUM

versible evolutionary trends took groups down blind alleys-useful trends which proved lethal when taken to extreme or when environmental conditions changed. Characters could therefore J ill ft! Ifll be described as 'major* or 'minor', depending on the taxonomic rank they diagnose. Provided the

I 'rank' of a character is not decided a priori, there

Tertiary I l is no problem; however, difficulties arise when i mm+ l l TT character states do not clearly change at taxon 1 Merostomata , Arachnida boundaries. For example, in a diagnosis of the i i i i i i i Cretaceous | subclass Soluta Petrunkevitch, 1949 is: 'ab- i : " h" domen composed of 8 to 1 1 segments' *--H\ r7^ Jurassic i (Petrunkevitch, 1955, p. P107). Petrunkevitch

I I described this variability as the character being in

Triassic I I UfSH-H-rt* H" a 'labile' state. So, the subclass Soluta is diagnosed on the labile condition of the abdominal segmentation, the presence of either a broad or a

It I M j narrow junction between the opisthosoma and t prosoma (see above), and the overall resemblance of the coxosternal region to that in spiders [my italics]. Petrunkevitch (1955) argued that solutes are not spiders because of the combination of characters in the group, and addition- ally they showed a single series of marginal plates on the opisthosoma. Obviously, such a group could also be considered a collection of quite different animals placed together through their shared possession of a spider-like coxosternal region.

Restudy of fossil solutes reveals that the problem lies mainly in Petrunkevitch's inability to FIG. 1. Stratigraphic ranges of Chelicerata and correctly interpret fossil material. The number of Aglaspidida and presumed phylogenetic relation- segments in the Soluta is invariably 1 1 (Shear et ships (from St0rmer, 1955). al, 1987) but the number Petrunkevitch interpreted in each specimen differed according to its Stethostomata, was created at this time to accom- preservation. Thus, where a 2-segmented modate the orders Anthracomartida and Hap- pygidium was preserved, then 2 additional seg- topoda which supposedly have a broad ments were counted over specimens which did prosoma-opisthosoma junction and a unique not preserve this organ, and the short first ab- coxosternal region. Petrunkevitch's (1949) clas- dominal segment is not always visible in fossils. sification scheme, used in the Treatise', has not Similarly, the interpretation of the prosoma-opis- stood the test of time. Weygoldt and Paulus thosoma junction depended on how closely these (1979) noted its use in some textbooks but tagmata were conjoined in the fossil. pointed out severe deficiencies in the scheme Petrunkevitch described Trigonomartus pus- when other characters are taken into account. tulatus, and noted (1913, p. 104): 'The Petrunkevitch was a devout proponent of the cephalothorax being much harder, kept more or idea of the 'decoupling' of macroevolution and less its shape, and what appears on it as a median microevolution. He envisaged major features crest was in reality a median groove. The ir- (those which define higher taxa) originating by regular, polygonal depressions were evidently mutation or other accelerated evolution, whereas thickened areas of the chitin and formed in life minor morphological differences (those which low elevations.' But, two pages before he had separate species, for example) could provide only diagnosed the new genus thus: 'Carapace trian- long, slow evolution and rarely produced higher gular with a median crest in the posterior half, taxa (Petrunkevitch, 1952, 1953). Petrunkevitch covered with irregular polygonal depressions.' (1955) envisaged extinction occurring when irre- Thus he had recognized that the fossils were ;

FOSSIL ARACHNIDS V>l

external moulds but diagnosed the genus as ii or impression of ventral into dorsal surfaces and they were casts. The error perpetuated until 1955 vice versa, * Kjellesvig-Waering' s conclusions on when, in the Treatise" (p. PI 12), the diagnosis Junctional morphology and phytogeny, both in became 'Carapace triangular, high, with median this MS and his other work, are not without crest and a pustulose surface, without eyes. Ab- dispute, but his long experience with the domen with pustulose surface* Thus, pustules taxonomy of fossil chelicerates was generally were recognized but the median crest remained, reliable Kjellesvig-Waering wrote in his MS without explanation for the emendation. Further- The question of whether PhalangiotarbnJi more, eyes exist in Thgonomartus fPetrunk- Haasc, 1890, or Architarbida Petrunkevitch, cvitch, pi. 9, fig. 49 in the same place as in 1945 is the proper name for this order of arach- 1913, 1 Aphantomartus (Pocock 1911, PI. II, fig. 6). nids has not been settled, although it is difficult These two genera were synonymized by Selden to understand why any question should have and Romano (1983). As well ;»s misinterpreting arisen in the first place.' What Petrunkevitch did fossils, Petrunkevitch produced some illogical was to substitute an existing name w ith one based taxonomic arguments, in 1945. he erected the on better preserved specimens of the order: 'Wh;*i Aphantomartidae for eophrynids with 7 ab- is more reasonable than to regard the Family Arehitarbidae as the most characteristic one of dominal tergites I i.e. Aphantomartus areolatus Pocock, 1911).' In 1949, he erected the the Order and to emphasize this fact by using a Trigonomartidae, and, recognizing thai Aphan- proper derivative of the generic name for the tomartus had 8 abdominal tergites, not 7 he Order? (Petrunkevitch, 19*5, p. 11). t staled (p. 256): 'This means that the Family The above examples show that much work is Aphantomartidae becomes a synonym of needed on fossil arachnids already in collect Trigonomartidae, the number of abdominal seg- in addition to study of the many new fc ments having served as the only character of awaiting description distinction/ Why not place the contents of the new Trigonomartidae' in the existing Aphan THE FOSSIL RECORD tomartidje? Aphantomartidae has priority and (Fig. 2) was redefined by Selden and Romano (1983j Furthermore, illustrations purporting to differen- Arachnid Relatives liate Aphantomartus and Trigono/nartus The extinct aglaspidids are probably not (Petrunkevitch 1955, figs 80, I and 3) are un- chelicerates since they be^r neither cheliccr.it representative and merely emphasise different other features which would ally them wilh !he characters mr genus Fig. of same SO, 3 is not Chelicerata over any other arthropod group Aphantomartus areolatus, as stated in the text, (Briggs et al. 1979). The fossil record doc t but Pruvost fig. a copy from (1919, 42) of A. help to determine the systematic position of the 1 with drawn incorrectly . pococku eyes on enigmatic pyenogonids. Chdicerae are not a pa- Consider also the Phalangiotarbida.Kjellesvig- requisite for a chelicerate. SaneiacansBrigg* and Waering redescribed this group just before his Collins, 198S from the Middle Cambrian Bu | [ death in 1979, and the MS was being prepared for Shale of British Columbia lacks chelicerac but posthumous publication (see Kjellesvig-Waer- was included in the phylum because of a c ing, 1978). In the MS, Kjellesvig-Waering, a bination of characters unique to Chelicerata: six renowned taxonomic 'splitter 1 reduced , pairs of prosomal appendages, cardiac lobe, Petrunkevitch' s 10 genera and 13 species to four prosomaandopisthosoma, and anus at rear c genera and five species. He stated in the introduc- trunk segment. Sunctacaris was described as tion to his MS: "Seldom, if ever, has a fossil group sister to all other chelicerates, but may not be the with such uncomplicated, mostly easily deter- oldest cheliccratc because a dubious xiphosuran minable morphological characters, been sub- carapace of Lower Cambrian age, EalimuQu jected to such misunderstanding and careless and afaM.v(Moberg. 1892) was recorded from Oland. erroneous work as has the order Phalangiotarhida Sweden Xiphosura are the most primitive

Haasc, 1 890. The main reason for this state has chelicerates in existence and, though preViC been the complete failure of some of the workers allied with the Eurypterida in the Merostomata- in this group to understand fundamental paleon- most authors place Xiphosura with eithet the tological principles of preservation, for example, Scorpionida (Bcrgstrom, 1979, 19S3;Bergslrom molds and casts, external and internal, along with t-t a].. 1980, van der Hammen, 1985, 1986) or as results of compaction and consequent reflection M.frt to all other chelicerates (except 392 MEMOIRS OF THE QUEENSLAND MUSEUM

Sanctacaris) (Grasshoff, 1978;Boudreaux, 1979; the otherwise sparsely recorded Mesozoic, scor- Paulus, 1979; Weygoldt and Paulus, 1979; pions reported from the Triassic of France (Gall, Weygoldt, 1980), thereby rendering Meros- 1971), and the Cretaceous of Brazil (Campos, tomata an unnatural group. 1986) are currently under study.

Scorpions PSEUDOSCORPIONIDA Scorpions are the arachnid group with the ear- Many pseudoscorpions are known from the liest known ancestors; the most ancient known Tertiary (mainly in ambers, e.g. listed in scorpion is Dolichophonus loudonensis (Laurie, Schawaller (1982, table 1), and some are known 1 889) from the Llandovery of the Pentland Hills, from Cretaceous ambers of Lebanon (Whalley, near Edinburgh, Scotland. Kjellesvig-Waering 1980) and Manitoba (Schawaller, 1991). How- (1986) proposed a controversial classification ever, the most important fossil pseudoscorpions scheme. Stockwell (1989) produced a more ac- are well preserved specimens of Dracochela ceptable classification scheme of Scorpionida deprehendor (Shear et al, 1989; Schawaller et which included fossils, but it has yet to be pub- al 1991), in the Upper Devonian mudstones of lished formally. A linchpin of Kjellesvig- 7 Gilboa, New York. Only protonymph and Waering's classification was the supposed tritonymph are known which, though modern in Devonian gilled scorpion described as Tiphos- many aspects, cannot be assigned with con- corpio hueberi. Restudy of this material (Selden fidence to extant taxa because both diagnostic and Shear, 1992) revealed that it is not a scorpion characters in the fossils and cladistic assessment but an arthropleurid myriapod! of extant forms are lacking. The early Silurian record of scorpions could be interpreted as representing the earliest terrestrial animals since all modern scorpions are terrestrial. Solifugae However, all Silurian fossil scorpions occur in The Carboniferous solifuge, Protosolpuga car- marine or marginal marine sediments, and mor- bonaria Petrunkevitch, 1913, was described as phological features suggest an aquatic mode of being in a very poor state of preservation. It is life. Petrunkevitch (e.g. 1953) considered all fos- impossible to judge the validity of the identifica- sil scorpions were terrestrial, but other workers tion from the published photograph and drawing. (e.g. Wills, 1947; Stormer, 1970; Rolfe and Be- The only reliable fossil solifuge is Happlodontus ckett, 1984; Kjellesvig-Waering, 1986) argued proterus Poinar and Santiago-Blay, 1989, from for an aquatic habitat for Silurian scorpions at Oligocene Dominican amber. least. Evidence for aquatism among fossil scorpions are: gills and digitigrade tarsi, as well as the Opiliones absence of terrestrial modifications such as coxal apophyses, stigmata, book lungs, trichobothria, Until recently, Opiliones had a fairly typical highly developed pectines and plantigrade tarsi. arachnid fossil record, being known only from

There is overlap in the ranges of aquatic and Upper Carboniferous strata and Tertiary ambers. terrestrial scorpions but the first terrestrial forms In 1985 a specimen was discovered in Lower probably appeared the Devonian (Selden and Carboniferous rocks of East Kirkton, near Edin- Jeram, 1989). It is not easy to decide whether a burgh, Scotland (Wood et aL, 1985), and a year given fossil had an aquatic or terrestrial mode of later, one was described from the Lower life; the original environment of the enclosing Cretaceous of Koonwarra, Victoria, Australia sediment is commonly the best clue, but a recent (Jell and Duncan, 1986). Both of these unnamed find is worthy of especial note: well preserved specimens are long-legged opilionids but no fur- book lungs in a Carboniferous (Visean) scorpion ther identification is possible (pers. obs.). from East Kirkton, near Edinburgh, Scotland The order Kustarachnida Petrunkevitch, 1913 (Jeram, 1990). Few new records of fossil scor- is included with the Opiliones, following Beall pions have turned up in recent years although in (1986).

FIG. 2. Current knowledge of the fossil record of Aglaspidida, Pycnogonida and Chelicerata; data in Selden (1993). Solid lines denote actual occurrence in the stage(s) concerned; interrupted lines indicate presumed occurrence in intervening stages. ? denotes doubtful record. Note that taxon ranks are not equivalent; occurrences of important genera Sanctacaris (most plesiomorphic chelicerate) and Attercopus (oldest and most plesiomorphic spider) are shown separately. Stratigraphic resolution is to stage; abbreviations in second column refer to standard stage names (see e.g. endpapers of Briggs and Crowther, 1990). FOSSIL ARACHNIDS 393

HOL PLE PL I n=t MM

CHT RUP PRB g

LJT

HA DAN

J-- A ,\ CMP 5 A N

TUR CEN ALB

VLG

OXF

& I BTH -£ 8A J TOA £ X PLB SIN HET

NOR

LAD (A N S 1=1 SCY TAT

, I A

UF I JOJN

S A K

GZE KAS. MOS BSH B 3--s >. FAM a— E <1 GIV 1 4frefcapus EIF

SIG GEO PKD

LLY ASH I R D uo_ LLN ARG TRE

STD | Sanctacxns CRF EDI STU (

3JM MEMOIRS OF THE QUEENSLAND MUSEUM

I^tALA^lOrAKI-'ILV. Haptopoda The situation of ihe phatangiotarbids has been This monotypic order was established by described above. This group is only known from Pocock (191 J) on the basis of the subdivided ihe Upper Carboniferous but fossils are tarsus of the first leg. Petrunkevitch (1949) widespread in European and North American denned and reexamined the specimens, and coalfields. redefined the order based on a new interpretation of the abdominal segmentation. The group would repay restudy along with Anthracomartida and RlCINULEJ Trigonotarbida. Ridnulei axe known only from the Upper Car- boniferous ofEurope and North America, and the ANtHRACOMAfcTlDA New and Old World tropics ai the present day Together with Haptopoda, this order forms (their range extends outside the tropics mainly by Petrunkevitclvs 1949 subclass Stethostomata. In cavernicole species). A recent revision of the discussion of the rationale for separating fossils (Selden, 1992) revealed a greater diversity a Stethostomata from Soluta (Shear and Selden. in the Carboniferous than today, but based on an Shear et aL, 1987), it was concluded that essentially similar body plan. It appears that the 1986; the only feature separating anthracomartids from group has remained in warm, humid habitats trigonotarbids is two versus ooe rows of marginal (equatorial forest litter and caves) throughout its tergal plates the opisthosoma. Again, this com- geological history on mon Upper Carboniferous group needs careful restudy. Mites Actinedida (Pm&tigmaia) The oldest miles are Trigonotarbida from the Lower Devonian Rhynic Chert of Scot- Trigonolaibids are the best known extinct land (Hirst. 1923). Other Devonian Ac- arachnid group on account of their excellent tinotrichida are known from Gilboa, New York preservation in the Devonian Rhynic Chert of (Norton etaL 1988, 1989; Kethley <7tf/ , 1989), Scotland and Gilboa mudstones of New York and A few Jurassic and Cretaceous Actinotrichida are arc among the first known land animals (Jeram e known (e.g. Bulanova-Zakhavatkina, 1974; a/,» 1990). First described from Upper Car- Krivolutsky and Ryabinin. 1976: Sivhcd and boniferous rocks (Buckland, 1837;Fritsch, 1901; Wall work, but the majority of fossil mites 1978), Pocock, 1902. 1903, 1911), Hirst (1923) are oribauds from Bailie amber (e.g. Koch and described the first Devonian specimens (from Berendu 1854; Sellnick. 1918, 1931). Anac- Rhynic), and St0rmer (1970) described forms tinotrichida are very poorly represented in the from the Middle Devonian of Alken-an-der- fossil record; there arc no fossil Opilioacarida or Mosel, Germany. Trigonotarbida is one of the Holothyrida and only a tew, somewhat suspect, irjLhnids groups found relatively frequently records of Ixodida (e.g. Scudder. 1890) and in Palaeozoic terrestrial rocks of from Argentina Gamasida (e.g. ITirschniann, J971). Fossil mites (Pinto and Hiinicken, 1980), Spain (Selden wnd are probably found routinely in palynoiogical Romano. 1983), Czechoslovakia (Oplusril, preparations but are unreported. With the growth 1985), and Germany (Jux, 1982). of micropalaeontological techniques in >he study The exquisite preservation of the Rhynie Chert of fossil arthropods it is likely that many more meant that Hirst (1923) could describe minute fossil mites will be identified details of the trigonotarbids from that deposit.

Trigonotarbids from Gilboa (Shear et al. t 1987) Palpigradi not only confirmed Hirst's observations but also The preservation potential of palpigradcs is uncovered further morphological features of even lower than that of mites Their small size. these interesting animals. Later work lias shown thin cuticles and interstitial habitats makes them that some of the species described as trigonotar- difficult objects of study when Recent or fossil bids in 1987 were really spiders or other pul- Sternanhron ziueli Haase, 1890, from the Juras- monale arachnids (Selden et a/., 1991), but the sic lithographic limestone of Solnhoferi, Ger- systematic position of the Trigonotarbida, sister many . is doubtful; the only good fossil palpigrade to all other pulmonates, was strengthened A is Palaeokocncnia mordax Rowland and Sissom, trigonotarbid and centipedes, found together with 1980, from the 'Onyx Marble' quarries early land plants in Silurian (basal Prfdoli) sedi- (Pliocene) of Arizona. ments at Ludford Lane, Ludlow, England (Jcram FOSSIL ARACHNIDS 395

et al, t 1990), pushed back the earliest record of (1990b) argued (bra new group of Carboniferous land animals, by around 16 million years and 'labidognalhous liphistiomorphs 1 on the indicated thai trigonotarhids were among the ear- evidence thai |hc P ked chelicerae yet any liest terrestrial animals. spider with orthognath chelicerae would have them preserved if the carapace and palps were. Arawaf. This argument presupposes that orthognath chelicerae are always porrcct, which they may n04 Great strides have been made recently in spider be. To argue morphology from preservation (of minus iCoddington and Levi, 1991) and lack of it is a dangerous practice. concomitantly, new finds of fossil spiders have !) Until recently, no mygalomoxph spider was added to the ge olog ical record. The oldest spider known earlier than the lernnTy. Eskov and is Attercopus fimbhunguis Shear. Selden and Rolfe, 1987, from Gilboa; supposed spiders from Zonshtein (1990a) described some Rhynie (Hirst, 1923) and Alken-an-der-Mosel mygalomorphs from Siberia and Mongolia, plac- (Stormer, 1976) have been disproved (Selden el ing them in the modern Mecicobothriidae, Atypidae and Antrodiaetidae. They are excep- ai . 1 99 1 }. Attercopus is sister to all other spiders; the patella-tibia joint is a rocking joint but in a tionally well preserved, but poorly illustrated and more plesiomorphic state than other spider;Jack described; in contrast, the line drawings are of

ing the 'compression zone Y* of Manton ( 1977). high quality. In 1992. with the description of a Autapomorphies of the Attercopus clade arc: Triassic mygalomorph, Rosamygale^ our fimbriate paired claws, spinules on the palpal knowledge of the antiquity ofmygalomorphs was femur, and lack oftrichobothria: the latter feature more than doubled (Selden and Gall, 1992). This is puzzling. was placed in the extant family Hexathelidac>and In spite of descriptions of Devonian and Car- suggests a widespread distribution of the family boniferous araneomorph spiders (Archaco- across Pangaea before rifling of the supercoo- metidae Petrunkcvitch, 1949; Pyritaranetdae tinent. Hexatheb'ds show many plesiomorphic Petrunkevitch, 1953), none of those seen by the characters among mygalomorphs but neverthe- author could be proved to be a spider at all. less, mygalomorphs fflaj yet be found in Petrunkevitch seemed to concur with Frusch Palaeozoic rocks.

(1904) and Pocock ( 191 1 ) in their placement of Mesozoic spiders have only recently been dis- fossils in the Araneomorphae without question, covered. The oldest fossil araneomorph is even if he disagreed with their detailed descrip- Jurarcmeus rasriitsytit Eskov. 1984. placed in a tions. All audiors to place fossils of these seemed new family, Juraraneidae, in the Araneoidea. in the basis their Araneomorphae on of general Juraraneus, like the mygalomorphs descrihed by resemblance to particular groups of araneomorph Eskov and Zonshtein (1990a), is well preserved spiders rather than real characters. For example, but rather poorly documented for such an impor- Petrunkevitch (1953: 107) defined Pyritaraneidac tant find, so it is difficult to be sure whether the and redefined Arachaeomeudae as araneomorph placement is justified. Eskov (1987) has also spiders with segmented opisthosomae, differing described Archaeidae from the Jurassic of from each other by their laterigrade and prograde Kazakhstan from where Filistatidae are cunrenil) legs respectively. Nowhere isihe identification as being described (Eskov. 1990). araneomorphs questioned Esko\ and Zonshtein Recent finds of Cretaceous araneomorphs have < 1990a) considered segmentation of the opis- emphasized the diversity of a spider faiinfl nl thosoma in the Pyritaraneidae to be an artifact, modern aspect during this period. Unfortunately, but agreed that this family belongs in Araneomor- some show liule morphological detail (Jell and phae Selden et al. (1991 1 studied Archaeometa Duncan, 1986), but Selden (1990a) described nephilina Pocock, 1911 in the British Museum of (Natural History) and a plaster cast ofA. ctevomca specimens from the Lower Cretaceous north- Stormer, 1976 from the Senckenberg Museum, east Spain, beautifully preserved in lithographic concluding that neither species was a spider and limestone. The specimens included a deinopoid that A. devon'tca may not be an arachnid at all. 3nd a tetragnathid, so both eribellate and ecribel- Carboniferous ArthrolycosidaeFritsch, 1904 and late orb-web weavers were in existence at this Arthromygahdae Petrunkevitch. 1923 in the time. In broad terms, by the Tertiary, the spider British Museum (Natural History) can be placed fauna was almost identical to that of today, and with the mesothclcs because of the distinct tei- only 3 families are known to have become extinct gites on their opisthosomae. Eskov and Zonshtein Since Ihe Palaeogene (Eskov, 1990) 196 MEMOIRS OF THE QUEENSLAND MUSEUM

Uropygi Well preserved uropygids are found in Coal Measure rocks in Europe (e.g. Brauckmann and Koch, 1983) and North America. All are placed in the modern Thelyphonidae.

SCJIIZOMIDA Three species of schizomids are known from the Pliocene "Onyx Marble* quarries of Arizona and one from the Oligocene of China (Lin et al, 1988).

Amblypygi Fossil amblypygi arc known from the Coal Measures of Europe and North America and from Tertiary ambers (e.g. Schawaller, 1979). Amblypygi may be present in the Devonian of Gilboa; a possible pedipalp tarsus was figured by Shear etal (1984) and Ecchosispuhhribothrium Selden and Shear. 1991 may belong in this group (Selden etal. 1991).

ARACHNID PHYLOGENY t/an der Hsmmen 1980

Selden (1990b) discussed three recent phylogenetic hypotheses with the evidence of the fossil record (Fig. 3). A cladogram which accurately reflects evolutionary events predicts that successive dichotomies should occur in ascending chronological order, and a complete fossil record should show this. Weygoldt and Paulus's

( 1 979) analysis (Fig. 3c) predicts that palpigrades should occur in strata at least as old as Devonian because the more derived mites and pseudoscorpions occur in beds of (hat age. In their scheme, Opiliones occupy a derived position. Van der Weygoldt sna paillu*

Hammen ( 1 989; fig, 3b) suggested that Opiliones should occur the Cambrian since they are tenta- FIG. 3. Cladograms of relationships among the arach- tively shown as sister group to Xiphosura + Scor- nid groups as viewed by a, Shultz (1990); b, van der piones. Shultz (1989, 1990; Fig. 3a) also placed Hammen (1989); and c ? Weygoldt and Paulus ( 1 979). position Opiliones in a which predicts their Interrupted lines indicate uncertainty. presence in Silurian times. Since scorpions were aquatic then, so would opilionids have been. FUTURE PROSPECTS None of the phylogenetic analyses (Fig. 3) in-

corporated extinct groups. Whilst it is impossible to include ancestors in cladistic analyses, there is Work in progress includes: palaeophysiology no reason why well known extinct groups should of early terrestrial chelicerates-aquatic and ter- not be included, say at the Carboniferous level. restrial adaptations in eurypterids, scorpions, and Apart from the enigmatic palpigrades and the other Siluro-Devonian arachnids; paleobiology highly derived Schizomida, for which fossil of the Trigonotarbida; and Cretaceous spiders evidence is lacking, all arachnid orders were in from Canadian amber and the Santana Formation existence by that time. of Brazil. FOSSIL ARACHNIDS 397

Much of Petrunkcvitirh's work needs revision. Gejtielseorpion |Arachnida: Thelyphomda. A new phaiangiota/hid fauna has been collected Thelyphonidae] aus dem Namurium lunrcrc- West-Deutschlaii:! in recent years from a coal mine tip in Somerset Ober-Karbon) von England iBeall, 1991). Carboniferous lomologia Generalis 9: 63-73. BRIGGS, D.EG., BRUTON. D.L. & WHIT- Anthracomanida and Haptopoda need to be res- TINGTON, H.B. 1979. Appendages of the tudiecL particularly in relation to the now ex- arthropod Aglaspis spinifcr (Upper Cambrian, tremely well known and possibly related i)sin> and their significance. Palaeontology tngonotarbids. The identity of described 22: 167-180. Amblypygi is in little doubt, bui modem descrip- BRIGGS, D.E.G. & COLLINS, D 19S8 A Middle tions would be helpful. In need of critical ex- Cambrian ehchoerate from Mount Stephen amination are: the single fossil solifugc British Columbia. Palaeontology 31: 779-9':v. Protosolpuga from Mazon Creek, the supposed BRIGGS. D.E.G. & CROWTHER, P.R. (eds). 1990. palpigrade Sternarthron from the Jurassic of Ger- 'Paleobiology a synthesis'. (BlackwcII Scien- many, and the amber spiders described by tific Publications Oxford, London. Edinburgh, Boston, Melbourne i. I942 Petrunkcvirchi'eg. t 1950, 1938). The prob- BUCKJ.AND. W. 1837. 'The Bridgewater ireausev on lem with these spiders is that over the years some the power, wisdom and goodness of God us of the supposed 'amber' in collections has dis- manifested in the creation, Tiv.itise IV. Genii m*v coloured, which suggests II not truly may be and ice to

Palaeogene but . rather more recent copal or other natural the li 2nd Edit. (William ftcteertog: resins. I ondonX

Successful palaeoarachnoiogy requires BULANOVA-ZACHVATINA I; M 1974 New knowledge of both Recent arachnids and under- genera of oribatid miles Iron; the Lppci standing of styles of fossil preservation. Much Cretaceous of Tairoyr. Palcootologichcskij Zhur- naJ2: 141-144. previous work suffered from erroneous inter- CAMPOS. D.R.B. 1986 Pnmeiro Tegistro fossil dc pretations of one sort or another. Goals for future Scorpionoidca na Chapado do Araripe (Cretaceo work are: to understand the origin of the pTesent- Inferior), Brasil. Ana is da Academia Brasilcira de diversity of arachnids day and the relationships Cienoas58: 135-137. among the various groups, and the reconstruction CODD1NGTON, J. A. & LEVI, H.W. 1991. of ancient terrestrial ecosystems. tematics and evolution of spiders (Araneac). An- nual Review of Ecology and Systematic $ LITERATURE CITED 565-592. ESKOV, K. 1984. A new fossil spider family from the BEALL, B.S. 1986. Reinterpretation of the Knstarach- Jurassic of Transbaikalia (Araneac. Chcliccratai. roda. (Abstract). American Arachnology 34; 4. Neues- Jahrbuch fiir Geologie und Palaoniologie, 1991. The Writhlington phalangjotarbids: Iheu Monatshefte 1984: 643-653.

palaeobiLMogieaUigniflconec. Proceedings i>l 'uV 1^87. A new arehaeid spider (Chelieerata; AraOCac) Geologists Association 102: 161-168. fwffl the Jurassic of Kazakhstan, with tioIcn on

led ranges I BERGS 1R0M, J. 1979. Morphology of Fossil thcso-cal "Gondwanan" qI rco Dd arthropods as a guide lo phylogenelie relation- Neucs Jahrbuch fur Geologic und PalaotUolueie. ships. Pp. 3-56. In Gupta, A. P. (cd.) 'ArthropixJ Abhandlur.gen 175:81-106.

Phylogenv'. (Van Nostrand Reinhold: New 1990 Spider paleontology .present trends and I QI 190- Yoiky. ex pec tfil ions. Acta Zoologica Fennica 12V 198), Morphology and svstematics of early 127 arthropods, Ahhandluneen des Naturwt- ESKOV. K. &ZONSHTEIN, S. 1990a. First Mcsoz senschafllichcn Vcrcin in Hamburg 23: 7-42. rus e alomorph spiders from the LowerCretaccnus with notes the system BERGSTR6M, J., STURMER. W. & WINTER. C. nf Siberiaand Mongolia on and evolution ot the inlraorder Mygalomorphae 1 9 S . Poiaeoisopus. Palaeopantopus a n .1 Pahfolhea, pycno>:ontd arthropods from the iCbchcerata; Araneac). Neues Jahrbuch liii Lower Devonian Hunsriiek State. West Germany. Geologic und Palaontologie, Abhandlungen 178: Palaontolopjsche Zcilschrifl 54: 7-54, 325-368. BOUDREAUXrii.B. 1979. Signilkanee ol interseg- 1990b. A new classification for Ihe order Aranetda mental tendon system in arthropod phylogeny and (Araehnida; Chelieerata"!. Acta Zoolosica Fcn- a ninnophylclic classification of Arthropoda. Pp. nica \9Q. !29- 137. 551-86. In Gupta, A. P. (edj 'Arthropod FRITSCH, A. 1901. ^Faunadcr Gaskohle. IV. Penois- Phvlogeny\ (Van Nostrand Reinhold: New che Arachniden'. p. 56, pis 153-154, text-figs.

York ). 359*363. BRAUCKMANN r, & KQCH, L 3983 1904 'Palaco/.uische Aiachmdcu'. (A. FfiLsch Prothelyphunus riuttfra^u.K n tp., cm neuer Prague). 86pp., 15 pi. 398 MEMOIRS OF THE QUEENSLAND MUSEUM

GALL, J.-C. 197 1 . Faunes etpaysages du Gres a Voltzia 124pp. [Published posthumously with footnotes du nord des Vosgcs. Essai paleoecologique sur le by A. Menge]. Buntsandstein supdrieur. M6moires du Service de KRIVOLUTSKY, D.A. & RYABININ, N.A. 1976. la Carte Geologique d* Alsace et de Lorraine 34: Oribatid mites in Siberian and Far East amber. 1-318. Reports of the Academy of Sciences, USSR 230: GRASSHOFF, M. 1978. A model of the evolution of 945-948. [In Russian]. the main chelicerate groups. Symposia of the LAURIE, M. 1889. On a Silurian scorpion and some Zoological Society of London 42: 273-284. additional eurypterid remains from the Pentland HAASE, E. 1890. Beitrage zur Kenntnis der fossilen Hills. Transactions of the Royal Society of Edin- Arachniden. Zeitschrift der Deutschen Geologis- burgh 39: 515-590. chen Gesellschaft 42: 629-657. LIN QI-BIN, YAO YI-MIN, XIANG WEI-DA & X1A HAMMEN, L. VAN DER 1985. Functional morphol- YU-RONG. 1988. An Oligocene micropalaeocn- ogy and affinities of extant Cheliccrata in evolu- tomofauna from Gubei District of Shandong and tionary perspective. Transactions of the Royal its ecological environment. Acta Micropalaeon- Society of Edinburgh: Earth Sciences 76: 137-146. lologica Sinica 5: 331-345. 1986. Comparative studies in Chelicerata IV. MANTON, S. M. 1977. "The Arthropoda. Habits, func- Apatellata, Arachnida, Scorpionida, Xiphosura. tional morphology, and evolution'. (Clarendon

Zoologische Verhandelingen, Leiden 226: 1 -52. Press, Oxford), xxii + 527 pp.

1989. 'An introduction to comparative MOBERG, J.C. 1892. Om en nyupptackt fauna i block arachnology'. (SPB Academic Publishing: The af kambrisk sandstcn, insamlade af drN. O. Hoist. Hague), x + 576pp. Geologiska Foreningens i Stockholm HIRSCHMANN, W. 1971. A fossil mite of the genus F6rhandlingarl4: 103-120. Detuirolaelaps (A carina, Me so stigmata, NORTON, R.A., BONAMO, P.M., GRIERSON, J.D. Digamasellidae) found in amber from Chiapas, & SHEAR, W.A. 1988. Oribatid mite fossils from Mexico. 69-70. In Petrunkevitch, A. (ed.) Studies a terrestrial Devonian deposit near Gilboa, New of fossilifcrous amber arthropods of Chiapas, York. Journal of Paleontology 62: 259-269. Mexico. University of California Publications in 1989. Fossil mites from the Devonian of New York Entomology: 63, vi + 106pp., 3 pis. State. 271-277. In Channabasavanna, G.P. and HIRST, S. 1923. On some arachnid remains from the Viraktamath, C.A. (eds.) 'Progress in Acarology. Old Red Sandstone (Rhynie Chert Bed, Aber- Vol. 1. Systematics and Taxonomy of Acari*. deenshire). Annals and Magazine of Natural His- (Oxford and IBH Publishing Co. Pvt. Ltd: New tory (9) 12: 455-474. Delhi, Bombay, Calcutta). JELL, P. A. & DUNCAN, P.M. 1986. Invertebrates, OPLUSTIL, S. 1985. New findings of Arachnida from mainly insects, from the freshwater, Lower the Bohemian Upper Carboniferous. Vestn. Cretaceous, Koonwarra Fossil Bed (Korumburra Ustred ust. geol. 60: 35-42. Group), South Gippsland, Victoria. Memoirs of PAULUS, H.F. 1979. Eye structure and the monophyly the Association of Australasian Palaeontologists of Arthropoda. 299-383. In Gupta, A.P. (ed.)

3: 111-205. 'Arthropod Phylogeny' . (Van Nostrand Rcinhold: JERAM, A.J. 1990. Book-lungs in a Lower Car- New York). boniferous scorpion. Nature 342: 360-361. PETRUNKEVITCH, A. 1913. A monograph of the

JERAM, A.J., SELDEN, P.A. & EDWARDS, D. 1 990. terrestrial Palaeozoic Arachnida of North Land animals in the Silurian: arachnids and America. Transactions of the Connecticut myriapods from Shropshire, England. Science Academy of Arts and Sciences 18: 1-137. 250:658-661. 1923. On families of spiders. Annals of the New JUX, U. 1982. Somaspidion hammapheron n. gen. n. York Academy of Sciences 29: 145-180. sp.-ein Arachnide aus dem Oberkarbon der sub- 1942. A study of amber spiders. Transactions of the varistischen Saumsenke NW Deutschlands. Connecticut Academy of Arts and Sciences 34: Palaontologische Zeitschrift 56: 77-86. 119-464. KETHLEY, J.B., NORTON, R.A., BONAMO, P.M. & 1945. Palaeozoic Arachnida of Illinois. An inquiry SHEAR, W.A. 1989. A terrestrial alicorhagiid into their evolutionary trends. Illinois Slate mite (Acari: Acariformes) from the Devonian of Museum, Scientific Papers 3(2): 1-72. New York. Micropaleontology 35: 367-373. 1949. A study of Palaeozoic Arachnida. Transac- KJELLESVIG-WAERING, E.N. 1978. The Phalan- tions of the Connecticut Academy of Arts and giotarbida (Arachnida) of Mazon Creek. Geologi- Sciences 37: 69-315. cal Society of America, Abstracts with Programs 1950. Baltic amber spiders in the Museum of Com- 10: 258-259. parative Zoology. Bulletin of the Museum of 1986. A rcstudy of the fossil Scorpionida of the Comparative Zoology, Harvard University 103: world. Palaeontographica Americana 55: 1-287. 259-337. KOCH, C.L. & BERENDT, G.C. 1854. 'Die im 1952. Macroevolution and the fossil record of Bernstein befindlichen Crustaceen, Myriapoden, Arachnida. American Scientist 40: 99-122. Arachniden und Apteren der Vorwelf. (Berlin). 1953. Palaeozoic and Mesozoic Arachnida of ) -

FOSSIL ARACHNIDS 399

Europe. Gcoloajcal Society of America, Memoir SCUDDERSH 1890 lllvsirations of the Car- 53. xi+ 1-128 pp. boniferous Aractttiida of Nortli America, ol the

1955. Arachnida. p4M62 In Moore, R.C < cd orders Anthracomarti and Pcdipalpi. Mfcmdin of 'Treatise on invertebrate paleontology. Part P. the Boston Society of Natural History 4: 443-456. Arihropoda 2\ (Geological Society of America pis 39. 40. and University of Kansas Press: Bouldci. SELDEN, P.A. 1990a. Lower Cretaceous spiders tram Colorado and Lawrence Kansas), xvti + 18 1 pp. the Sierra deMontsech, north-east Spain. Palaerm 1958. Amber spiders in European collections lology 33: 257-285. TransacbonsoflbeCottruxticul Academy of Arts 1990b. The arachnid fossil record. Newsletter of the and Sciences 4 1 : 97-400. British Arachnological Society 58: 4-6. PINTO. ID. & HGNTCKBN* M,A. I9B0. 1992. The fossil ricinulcids. Transactions of the GondtyonAtechflt a new genus oi the order Royal Society of Edinburgh Earth Sciences 83: Trigonotarbida (Arachnida) Argentina, from 634, BoJetin de )a Academia National dc Cleric *.iv 1 993. Aglaspidida, Pycnogonidannd Cbcliceraia. In Cordoba, Argentina 53: 307-315. Benton, M.J. and Whyte, M.A. (eds) 'Tru (i POCOCK, R.l. 1902. Eophrynus ami allied Car- record 2'. (Chapman and Halt London, New boniferous Arachnida. Part I. Gcol* York, Tokyo, Melbourne. Madias), in iucsm. Magazine, decade 4 9; 439-448. SELDEN, P.A. & GALL, J.-C. J992. A Triage 1903. A new Carboniferous arachnid. Geological mygalumurrih spider from the northern Vosgcs Magazine, decade 4 10: 247-251. France. Palaeontology 35: 21 1-235. 191 1. A monograph of the terrestrial Carboniferous SELDEN, P.A. & JERAM, A.J. 1989. Palaeophv Arachnida of Great Britain. Monograph of the ogy of terrtstrialisation in the Chclicerata. Trans Palaeontographical Society 64: 1-84. actions of the Royal Soeierv of Edinburgh; Earth POINAR, G.O. SANTIAGO-BLAY, J.A. 19X9. A & Sciences 80: 303-310. fossil solpugtd, Happlodontus pnnerus, new SELDEN. PA. & ROMANO, M. 1983. Fust genus, new species (Arachnida: Solpugida) from Palaeozoic arachnid from Iberia: Aphantun; Dominican amber. Journal of the New York En- areolaius Pocock (Stephanian; Prov. Leon, N.W. tomological Society 97: 125-132. Spain), w'ich remarks on aphantomartid taxonomy PKUVOST, P. 1919. Introduction a I'elude du terrain Boletino del Instiluto Geologia e MfOeialogia houiller du Nord et du Pas-de-Calais. La faune d'Efcpana94: 106-112. continentale du terrain houiller du Nord de la SELDEN, PA. & SHEAR. W.A. 1992. A tnyriapod France. Memoircs pour scrvir a explication dc la identity foi the Devonian scorpion" TiphwcOrpio carte geologique dctaillec de la France, t Minister'.- hiwbfri, Bcriehi des nalurwissenschaftlicb des Travaux Publics, Paris) 5K4 pp Medi/iniscru n Vc- reins in Innsbruck. Supplement ROLFE, W.D.L & BECKETT. C 1984, AttfCOOkfflS M 10. 15*36 of Silurian Xiphosurida, Seorpionida. an.) Rhyl SELDEN. PA.. SHEAR. W.A. & BONAMO. PM. locarida. In Basset L M.G and 1 awson. I D \99l< A spidei und othei arachnids from Hie 'Auteeology o\' Silurian Organisms'. Sp man of New York, and reintcrpretati<>i Papers in Palaeontology 32: 27-37, Devoiiuui Araneae. Palaeontology 34: 241-2X1. ROWLAND J.M.&S1SSOM, WD 1980, Rcpojl SELLNICK. M. 1918. Die Oribatiden de. Bemsuni- fossil palpigradc from the Tertiary ol Arizona, ;;n. \ sammlung der Universitat Konigsberg. a review of the morphoJogyandsystcmadcBoflhc Seta del rhysikaliseli-Okonotnisehcn Gesellsehaft z.u order (Arachnida: Palpieradida). Journal of Konigsbcrg 59: 21-42, Arachnology 8: 69-86 1931. Milben Bernstein. lk'n»slcinforschuni»en IAWALLER, W. 1379. Erstnaehwcisdet Qftkmng im GeiBelspinncn in Dominican! schem Bernstein 2: 148-180, (Stuttgarter Bernsteinsammlung Arachnida, SHEAR, W.A., BONAMO, P.M., CJRIERSON, J.D.. Amhlypvgi). StuUgarterBeitraae , und ROLFE. W.D.I., SMITH. E.L. & NORTON. R.A. SerieB50: 1-12 1984. Early land animals in North Ann 1982. Spinnen der Pamilien Tctragnathidac. Science 224: 492-494, Uloboridae und Dipluridaein Dominicanischem SHEAR, W.A. & SELDEN. P.A. 19K6 Phylogenelic Bernstein und allgemeine Gcsichtspi lelation&hips of the Trigonotarbida, an extinct

(Arachnida. Araneac). Slutlgutlct Be i tragc /.ut order ol arachnids. Adas X Congrcsos Inter- Naturkunde. Sene B 89; 1-19. national dc Aracnologia, Jaca, Espana, 1986 I: 1991. The first Mesozoic pscudoscorpinn, from 393-397 Cretaceous Canadian amber. Palaeontology 34: SHEAR, W.A., SELDEN. PA., ROLFE, W.D.I., 971-976. BONAMO, P.M. & GR1ERSON. J.D. 1987. New SCHAWALLER, W„ SHEAR, W A & BONAMO, lem'.sinal arachnids from the Devonian of Gilbou, PM 1991 The lirsl Paleozoic DSClldascorpiioiUl New York (Arachnida. TrigonotarbjOal. (Arachnida. Pseudoscorpionida). A men can American Museum Nuvitatc-j 2901 l-M I. Museum Novitaies 3009: 1-17. SfllAR. W.A.. SCKAWALLER, W. ft BONAMO. 400 MEMOIRS OF THE QUEENSLAND MUSEUM

P.M. 1989. Record of Palaeozoic pseudoscor- 1976. Arthropods from the Lower Devonian (Lower

pions. Nature 341 : 527-529. Emsian) of Alken-an-der-Mosel, Germany. Part SHULTZ, J.W. 1989. Morphology of locomotor ap- 5. Myriapoda and additional forms, with general pendages in Arachnida: evolutionary trends and remarks on fauna and problems regarding in- phylogenetic implications. Zoological Journal of vasion of land by arthropods. Senckenbergiana the Linncan Society 97: 1-56. Lethaea 57: 87-183. 1990. Evolutionary morphology and phylogeny of WEYGOLDT, P. 1980. Towards a cladistic classifica- Arachnida. Cladislics 6: 1-38. tion of the Chelicerata. 8th International Congress SIVHED, U. & WALLWORK, J. A. 1978. An early of Arachnology, Wien, 1980: 331-334. Jurassic oribatid mite from southern Sweden. WEYGOLDT, P. & PAULUS, H.F. 1979. Unter- Geologiska Foreningens i Stockholm suchungen zur Morphologic, Taxonomie und Forhandlingar 100:65-70. Phylogenie der Chelicerata. Zeitschrifl fur STOCKWELL, S.A. 1989. 'Revision of the phylogeny Zoologischc Systematik und Evolutions- higher classification of scorpions and forschung 17: 85-200. (Chelicerata)'. Unpublished Ph.D. thesis. Univer- WHALLEY, P.E.S. 1980. Neuroptera (Insecta) in sity of California, Berkeley. amber from the Lower Cretaceous of Lebanon. ST0RMER, L. 1 955. Merostomata. P4-P4 1 . In Moore, Bulletin of the British Museum (Natural History), R.C. (ed.) 'Treatise on invertebrate paleontology. Geology 33: 157-164. Part P. Arlhropoda 2\ (Geological Society of America and University of Kansas Press: Boulder, WILLS, L.J. 1947. A monograph of British Triassic Colorado and Lawrence, Kansas), xvii + 181pp. scorpions. Palaeontographical Society Monographs 437, 441: 1-137. 1 970. Arthropods from the Lower Devonian (Lower Emsian) of Alkcn-an-der-Mosel, Germany. Part WOOD, S.P., PANCHEN, A.L. & SMJTHSON, T.R.

1 . Arachnida. Senckenbergiana Lethaea 5 1 : 335- 1985. A terrestrial fauna from the Scottish Lower 69. Carboniferous. Nature 314: 355-356. JRP10NS AS MODEL VEHICLES TO ADVANCE THEORIES OF POPULATION AND COMMUNITY ECOLOGY: THE ROLE OF SCORPIONS IN DESERT COMMUNITIES GARYA.POUS

Polis, G.A. 1993 Jill: Scorpions as model vehicles to advance theories of population and communitv ecologv: the role of scorpions in desert communities. Memoirs ofthe Queensland Museum 33(2): 401-410. Brisbane. ISSN 0079-8835.

The diversity (5-16 species) and abundance (0.2-1.0 individuals/m') of scorpions suggest that ihcy may be quite ecologically important in desert communities Ecological importance is considered in terms of population energetics (the quantity of energy and mass flo, through populations) and of regulation of community structure and dynamics (influence on the distribution and abundance ol other species). The energetic analysis provided three

conclusions 1 ) Scorpions monopolise a relatively large share of" animal biomass, particularly relative to vertebrates and other arthropods. 2) This relative success is due to a suite of autecological traits (very low metabolism and very Mgn assimilation efficiency leading EC low energy requirements; and great tolerances to water stress, heat and starvation) that allows scorpions to prosper under the unpredictable and low food availability conditions that characterise deserts. 3) However, these traits lessen their impact 00 energetics, prey and competitors. Thus the importance of scorpions relative to horneolherrnic vertebrates is less than an analysis of density and biomass suggest because scorpions require and process pre\ in quantities relatively low to their biomass. Nevertheless, as a group, scorpions arc probably important conduits of energy flow in deserts. Research on the interactions among scorpions and between scorpions and spiders strongly suggests that scorpions can play key roles in the structure and dynamics of their communities. Studies* in the deserts of California, Baja California and Namibia show that intragmld is predation by scorpions a major force determining the (temporal and spatial j distribution. abundance and age structure of populations of their cornpctilors/prey.Qrt/wifire, Scor- pionida, age structure, distribution inrray.uild predation, population end community erof-

GarvA. Polis, Department of fiioloay, Vondrrhilt Universes. NashvilU*, Tennessee 17235. U.S.A., 26 October, 1992.

Ecology is a rapidly developing area with a Polis, 1 986) and the patterns and processes affect- great deal of internal argument arid disagreement, ing community and food web structure (Bradl As in other fields, theory has advanced far more 1983; Polis, 1991a). rap.diythanci.ipmcalvv^rk.Conscqucntly.much j n this paper I evaluate the 'ecological theory is controversial in need of empirical and importance" of scorpions. Throughout, I indie evaluation. Scorpions possess a number of char- how research on scorpions has advanced our acteristics that make them ideal models to test and general understanding of ecology. Although advance ecological theory. Thus, it is relatively scorpions occur in almost all non-boreal habitats. easy to collect rapidly great amounts of data and I focus on desert scorpions because much to manipulate experimentally individuals and en- evidence suggests thai they may be particularly tire populations. Research on scorpions has con- important components of arid ecosystems and are tributcd to many areas of population, behavioral a relatively 'successful' group. They are diverse ! and community ecology: e.g.. the evolution of life some taxa are extraordinarily abundant, and, as a history theory (Polis and Farley, 1980), the evolu- group, they form a large proportion of the tion and ecology of age-structured populations biomass of all desert arthropods and easily ex- (Polts, 1984a. 198S; McCormiek and Polis, ceed the biomass of all desert vertebrates Then 1986a). the dynamics of cannibalism (Polis, success arises partially because they exhibit 1980a, 1981, 1984b), the dynamics of intraguild several physiological and ecological traits that predation (Polis and McCormiek, 1986b, 1987, pre-adapt them to the low and unpredictable food levels r deserts. Polis vt al., 1989), the evolution and ecology of ° foraging strategies (Polis. 1980b; Bradley, 1988), 'Ecological importance* can be expressed in patterns and processes in biogeography (Due and terms of either energetics and nutrient cycling J

402 MEMOIRS OF THE QUEENSLAND MUSEUM

(the quantity of matter and energy flowing How do desert scorpions measure against other through a population or functional group) or groups of consumers for these parameters? First, regulation of community structure and dynamics scorpions are relatively speciose in deserts. On (the impact on diversity, distribution arid abun- the average, 7.1 species co-occur in desert dance of other populations). throughout the world (range: 2-16 syrnpatric species, typically. 5-9 species; Polis, 1990) (see ENERGETICS AND NUTRIENT CYCLING Table 1 for comparison with other desert taxa). Furthermore, populations are often quite dense, The importance of desert scorpions on energy averaging>3200 individuals/ha with several and nutrient cycling is a function of the quantity species maintaining populations 5000-10,000/ha of prey bioma&s captured. The amount of cap- (e.s.Shorthouse, 1971;Lamoral. I978;Polisand tured biomass is a function ofthe density, popula- Farley, 1980; Polis and McCormick, 1986a; tion biomass, metabolism and efficiency of Bradley, 1986. Polis. 1990). On the average,

energy transfer. I now present a very rough ap- pions are reportedly more dense than all other proximation of the relative importance of scor- macroscopic animal taxa in these deserts except

pions to the flow of energy and nutrients in desert 'insects' , iermites\ and isopods (ants are undoub-

ecosystems. To approach this question, data were tedly also more dense) (Table 1 ) Since scorpions compiled from the literature reporting the diver- ire .imone; the largest of all terrestrial arthropods

sity, density and biomass of various groups of (adults ofmost desert species = 0.5- 1 Og; Polis and

desert taxa" (see Polis, 1990, 1991b; Polis and Farley, 1 980), these high densities produce rather Yamashita, 1991 for additional details). These large estimates of standing biomass (= density of data are not without bias and other problems. For individual species x mass of individual animal). example, studies tire not usually conducted in Each species population of desert scorpion

areas where the focal taxon is absent or rare; nor averaged 7. 1 5 kg/ha. Only termites (and probably are rare species studied as often as common ones. anls) support a greater population biomass per Consequently, the statistics presented here over- estimate density and biomass and should be taken

as a first approximation of actual parameters. Population per Biomass (Vg/ha) Taxon n However, the data are instructive and suggest that species per taxon scorpions could be a major link in the flow of Mammals 1.40 39.9 29 y through desert communities. Birds 0.02 0.9 25

| .]• inls n.57 6.8 46

'Insects* O.SH 521.2 31

t ] Density (no./ha) Diversity (no. species) n Termites 12.45 113.4 7

Mammals 43 + 1 30 28J ±21.3 14 7 Isopods 9.91 12.S

Birds U.fti i i 4o 7 ±27 14 Mfllipe 1.15 2.8 2 U&inls 3d8£ U U Spitlrr-i D i:>. 7 4 Insect! 35. 100 ±54.600 592 1 796 ID Scorpions 7.15 50.8 17

nsnnit'- - 4.025.000+ 2.793.000 9.l± 15 7 AH venrbmrcs 47.7

.i i . .i.nl'. : ,. n>± 336.850 3±tLfc 1 All afhropods 7U1-; i

Millipedes 1150*214 : -i±i i 7 AH maemfauna 755 a Spiders 322U±8«00 54.4±26.7 31 Solpugids 9.9±6.7 37 TABLE 2. Estimated population biomass for various Scorpions 32io± tsqu 7.1 + 2.5 5v: taxa taken from literature. Population biomass per species is average wet weight of one species per TABLE 1. Estimated average density and diversity of hectare. This statistic was sometimes reported; how-

major taxa of desert maerofauna reported from the ever, it was often calculated by multiplying the literature. Diversity is the mean number of species per average mass of an individual times the density. taxon from local sites in different deserts. Density Population biomass per taxa is calculated as the statistics were standardized to a per hectare basis. product of population biomass per species and Means axe reported with their standard deviations; n average diversity of that taxon. Note that data on ants = sample sue. Note that these statistics should be were insufficient lo include. Additional information viewed as very rough approximations rather than ts reported in Polis and Yamashita (1991). n refers to absolute values. Note that data on ants were insuffi- the number of species in a particular taxon for which cient to include here (see Polis and Yamashita, 1991 biomass data exist. Note thai these statistics arc only for further details) gross approximations of reality. SCORPIONS AS MODEL VEHICLES 403

species or per taxon (from the ordinal level down) ferred. Thus a gram of scorpion does not process than scorpions (Table 2). as much food as a gram of arthropodivorous When the biomass of all scorpions living sym- vertebrate. Overall then, desert scorpions are less patrically in desert areas is calculated (= mass of important in energy and nutrient cycling than individual species x average number of sympatric their diversity, abundance and biomass suggest. species), scorpions as a group exhibit a greater How much energy do desert scorpions process? biomass (= 50.8kg/ha) than all other taxa except We have two estimates. Polis (1988) calculated termites (=1 13.4Kg/ha) and the sum of all other that average populations of Paruroctonus insects (= 521 .2kg/ha; Table 2). Note also that the mesaensis used 9000 grams of prey/ha/year. The population biomass of scorpions is higher than Australian Urodacus yaschenkoi requires any one group of vertebrates (e.g., mammals and 7 900g/h a/year; this translates into 98,400 ants or lizards average 39.9 and 6.8Kg/ha) or all ver- 31,570 medium sized spiders eaten per hectare tebrates combined (47.7kg/ha). Overall, scor- per year (Shorthouse, 1971; Marples and Shor- pions form 6.7% of the biomass of all macrofauna thouse, 1982). It is uncertain exactly how such species combined, 7.1% of the biomass of all figures for individual scorpions or for the sum of macroarthropods and 1 06% of the biomass of all all sympatric species of scorpions compare to vertebrates. those for vertebrates. This is an interesting ques-

Thus it would appear that they are important tion to pursue. conduits for energy transfer in deserts. However, two characteristics lessen their importance. First, INFLUENCE ON COMMUNITY they exhibit the highest ecological (production) STRUCTURE AND DYNAMICS efficiencies (percent assimilated energy incorporated into new biomass) of all taxa that live in The second measure of ecological importance the desert (Table 3). Second, they exhibit meta- is to determine how a taxon influences the bolisms that are extremely low relative to other dynamics, distribution and abundance of other poikilotherms and endotherms (Table 4). Al- taxa. In theory, scorpions can influence desert though these features are powerful adaptations communities in many ways: as predators affect- that allow efficient use of food and partially ex- ing characteristics of their prey, as prey of other plain the success of scorpions in deserts, they also predators and as competitors of other function to decrease the amount of energy trans- arthropodivores. One of the basic questions in ecology is: 'What Group P/A% factors determine the distribution and abundance

Insectivorous mammals 0.86 of species?' To approach this question, researchers often focus on groups of similar species Birds 1.29 that use similar resources; especially those Small mammals 1.51 resources the supply of which may be limiting 'Other' mammals 3.14 and in demand (e.g., food). Such species groups 'Homeotherms' 3.1

Fish 9.77 Taxon Metabolic rate (mlV02/gm/hr) Social insects 10.3 Homeotherms (basal rales) invertebrates* 25.0 Terrestrial Mammals 0.07-7.4 Solitary herbivorous insects 38.8 shrew, 7.4 Solitary detritivorous insects 47.0 rodents 1.80 Solitary carnivorous insects 55.6 elephant 0.07 Spiders 45-60 Birds 2.3-4.7 Scorpions 68.2 Poikilotherms (at 25°C)

Insects 1.665 ±1.25

TABLE 3. Ecological (Production Efficiency) of Spiders 0.92 + 0.92 various animal taxa. This efficiency is equal tothe Scorpions 0.057 ±0.048 proportion of assimilated energy that is incorporated into new biomass (=Production/Assimilation). Note that the highest efficiencies are found in carnivorous TABLE 4. Metabolic rates of various animal taxa. The insects, spiders and scorpions. *Terrestrial inver- data are taken from many sources. The sample size tebrates do not include insects or arachnids, (primari- for insects is 82 species; for spiders, 8 species; and for ly from Humphreys, 1979; see Polis and Yamashita, scorpions, 7 species, (see Polis and Yamashita, 1991 1991 for further details). for further details). MEMOIRS OF THE QUEENSLAND Ml JSHL'M

are called guilds (= a group of species that use INTERACTIONS AMONG SCORPIONS resources in a similar way and are thus potential IN THE COACHELLA VALLEY competitors). For example, guilds of desert desert scorpion co-occur in granivores (birds, ants, rodents) all eat seeds, Four species of sandy habitats on the floor of the Coachella Val- regardless of specific differences in resource ac- ley (Riverside County* California). Three are in quisition. the family Vaejovidae [PantroctortUt me54* One approach to study guilds is to describe their Stahnkc, P. luteolus Gcrtsch and Vaejovis con- patterns or structure. By guild structure, we mean fusus Siahnke), one [Hadrurus aHzonen&U the diversity and abundance of species members; Ewing), the [uridac (Polis and McCormick, spatial and temporal patterns and resource use 1986a, I9R7). ParuroctoHUS mesaettsis form (i.e., niche characteristics). However, such >95% of all individuals wul occur si densities in descriptive studies, although common, are not the range of 0,2-0.5 individuals/ha; the other fully satisfying because they do not directly ad* three species are relatively rare. Each species

(fastis Ihe process that prodwe the observed requires 2-5 years to mature and is composed of patterns. tjI distinctly sized year classes. Size changes An alternative approach is to determine if guild greatly, eg., P. nwsaensis increase 60-80 limes members interact, and if so, can such interactions in weight from 0.03g (instar 2) to 2.0-2.5g (non- significantly shape guild structure. There are gravid adults). The year classes and species overlap to various in use insect and several possible ways that guild members can degrees of

arachnid prey ; average overlap among all species interact. These range from cooperation and is moderate to high (0.67 [prey size] and 0.43 mutualism to competition n predfttiort [prey laxa]). Thus, the scorpions potentially com- Such predation among guild members is called pete for food. intraguild predation (Polis and McCormtck, Ext eirSJA e intraguild predation occurs among 1987; Polis et rf.. 1989). Intraguild predation is the four species; Tabic 5 presents a matrix of who an ubiquitous interaction as- among many iA horn. Several factors characterize this in- semblages of potential competitor but has traguild predation 1) Each species was both an received little formal attention from cither intraguild predator and prey. Such mutual preda- theorists or empiricists (Po 50) tion occurs simply because the predator scorpion One major theme of my research with scor- was always larger regardless of the species com- pions has been to analyze lhe characteristics and bination involved (n= 170 scorpion-scorpion significance of intraguild predation- Many scor- predations). Thus scorpions of all species are pion prey items (other scorpions, spiders and vulnerable as they grow from small juveniles to full si/.e adults and predatory reversals (mutual solpugids) are also potential competitors with predations) arc common. For example, young P. scorpions (Polis, 1990; Polis. l991a ? b;Po!isand rnesavnsis scorpions are eaten by relatively larger Yamashila, 1 991 1 1 present information on three adult P. luteolus and Vaejovis confusus: adult P. systems; in each, scorpions; frequently eat species fiwsuensis prey on the (now) relatively smaller in the same guild of arthropod) vorous predators adults of these species. Thus age/size is a key and such intraguild predation significantly affects the distribution, abundance and population PREY I dynamics of these potential competitors. used .''..'llrW V, PREDATOR ti.un r.iut can Total | scorpions in these studies us models to delineate Huilrtirw: 11.0 6 1 123 \ 1 2J.9 the characteristics and dynamics rjf intraguild t' iuit\'l\t\ 0.0 33.3 6.7 6.7 46.7 predation. These studies illustrate how scorpions P tnesatfitsis 0-4 03 5.0 8 3 I4J have proved to be extraordinarily amendable for kWfovft U.IJ 00 a.o |4.0 12,0 ecological research, large amounts of data can be collected, interactions (e.g., feedings) observed TABLE 5. Scorpion-scorpion predation in the and quantified relatively easily, and individuals Coachella Valley . The entries represent the percent of the diet each specie? forms as prey for each of the or whole populations manipulated experimental- thai four scorpion species. The diagonal represents inly. For example, in the first study, field data were inispecific prcdaiion (cannibalism) (from Polis & Mc- collected on 130,000 individuals, 2000 feedings Cormiek, 1987). fI.ari=HatJrurus urizanensis: P.lut and 6000 individual scorpions were manipulated = Parurorfo/ius Iufeolus; P.mes- Paruroctonus in conlrolled field experime nwsoens/s: V.con -Vaejovis confusus. SCORPIONS AS MODEL VEHFCLKS 405

determinate of intraguild predalion: adults are the nificantly depressed the abundance of the rarer lators and immature individuals aie the prey species. Does intraguild predation Ififcew ise affect significantly more frequently than expected by their distribution'' Since predation is apparently a chance. 2) "The most common species (P. mesaen- key factor in the population dynamics of these sis) was the predator in 91% of all intraguild species, natural selection is expected to fevoi predations observed. Its average overlap in prey adaptations that reduce the probability that an use (= 0.44i with other scorpion species was individual will encounter its predator. Indeed •nd highest 3) Intraguild prcdation. at least by prey often avoid places and times that their P mesaensis, is significantly more frequent when predators frequent or where the probabilii availability was low: when <1% of the predation is high (see Polis and McCon

population was feeding, heterospecifies formed 1987). Typically, the large predatory entity (e.g. (

uf all diet items. In contrast when the percent scorpion species and - class) occurs in

v. feeding , onlj 3-296 of all prey were productive periods and microhabitats whereas other species of scorpion. 4) Mortality caused by smaller entities coexist by spatial segregation in intraguild predalion was generally an inverse a heterogeneous Ittbttat uimI by temporal ilis function of the density of both P. luteolus and V placement. confusus\ this resulted because much of the sur- The temporal and spatial distribution of smaller face activity ofthesotwospecit ddui ng age classes and species of Coaehella Valley (less productive) periods when P. mesaensis was pions reflect avoidance (in ecological and/or

absent from the surface. 5 ) intraguild p*cdaiion is evolutionary time) of larger age classes i»m|m>[|;ii)i in scorpion population dvii.rni,- . species. The overall distribution of P. luteolus Whcn analyzed as percent mortality of small and V, confusus tend to plaoe these species on the species (here = total number of individuals eaten surface during limes (in winter, late fall) and in by P. mesaensis divided by the total number ever place and) characterized by relatively low observed), intraguild predation by P. mesaensis surface popul I adult P. mesaensis These L.i I led gfl£ and &% of ail P luteolus and V. con- times and microhabitats support significantly less fu&US ever observed. prey than those used by adult P. mesaensis) ooo P. a feeding rale Can such high rates of mortality cause the rarity sequcnlly mesaensis has (2.9 significantly greater than all other species com- Of ihese species h is only possible to nppniach this question using field experiments. Removal of bined (1.70%). Further, the minority of P. > 2 lufeolus and V. confusus that forage when and >60007 . mesaensis (=3.2 Kg) from 300 ( 1 00m ) P. is active suffer quadrats over a 29 month period demonstrated where mesaensis a dispropor- - ltd chance of being eaten by P. thai the rarity of these species is caused substan- Inttuspealic (cannibalism) tially by intraguild predation from P. mesaensis. mesaensis, predalion has produced similar patterns of temporal dis- 6) lioth P lunolus ;jih! V foujtiwix (bill not // trihution, feeding mortality patterns mensis) increased significantly (600% and and among

age classes P. mesaensis* I • 'i-, 1 135%) in removal as Compared with 00 control of 980, 1984a), quadrats. It was speculated that the rarity of the Thus, intraguild predalion is an important fac- largest species iH arizonensis) is a result tor limiting the abundance and shaping the dis- bottleneck in adult recruitment; predalion by P. tribution of these scorpions Many other mesaensis killed >10% of all newborn H. assemblages of desert scorpions exhibit patterns arizonensis obsci-vcd during the study. 7) The age lhat suggest that scorpion-scorpion ptedatton is a structure of these smaller species was significant- major process shaping distribution and abun- iffereflt in removal areas: first year juveniles iu and McCormick, 1987; Polis, 1990; were 1.75 to 2.85 more abundant in removals but sec Bradley, 1988). JUS controls. This suggest that the numerical

ii mse by the rarer species would be even more INTERACTIONS AMONG SCORPIONS. dramatic if this experiment continued beyond its SPIDERS AND SOLPUGIDS 29 month period. Note that a plausible alternative IN THE COACHELLA VALLEY hypothesis exists: removal of P. mesaensis ed exploitation competition and thus al- Scorpions also frequently eat competitors other lowed the observed increases in density. A robust than other scorpions. For example, the diet of the test failed to detect competition in this system scorpion P, mesaensis consisted »>f 896 vpl lia and McCormick, 1986b, 1987). and 14% solpugids (Polls and McCormick, Thus, intraguild predation by P. mesaensis sig- 1986b). Docs such intraguild predation sig- 406 MEMOIRS OF THE QUEENSLAND MUSEUM

nificantly affect the distribution and abundance This occurs because scorpions are effective of these unrelated taxa? Spiders responded in the predators on these spiders. above experimental removal of >6000 P. However, this outcome is more variable on mesaensis by doubling in removal quadrats as isolated trees further from the river and along compared to controls. Surprisingly, neither sol- smaller washes entering the river. On some trees pugids nor all insects combined increased sig- close to the river, the abundance of scorpions and nificantly (all p >.05) in removal plots. These taxa spiders is similar to that found in the river. How- did not increase either because individuals dis- ever, some trees have no scorpions and great persed from removal areas, because the increase numbers of spiders (50-400/tree); some trees of spiders and smaller scorpions compensated for have neither scorpions nor spiders; and some, no the removal of P. mesaensis by eating surplus scorpions and few (<20) spiders. Overall, trees arthropods, or simply that scorpions exerted little without scorpions support significantly more impact on insect populations. The first explana- spiders (112.3±60.6, n = 21) than trees with tion is likely true for widely foraging solpugids scorpions (24.5 ±14.3, n =20) (p <0.001; only and is possible but unlikely for the more sessile experimental trees scored). This variation in insects. The second explanation is unlikely: the abundance and these distributions exist because biomass increase of spiders and scorpions repre- neither U. otjimbinguensis nor G. echinatus dis- sented <10% of the removed P. mesaensis perse far from the river, yet spiders disperse fur- biomass. The third explanation, difficult to ac- ther than scorpions. Dense spider populations cept, is nonetheless a real possibility: P. mesaen- occur only in more isolated trees where scorpions sis may take such a small proportion of all insects are absent. In trees quite distant from the river, that its removal does not affect insect density. neither species occur. Does intraguild predation by U. otjimbinguen- INTERACTIONS BETWEEN SCORPIONS sis scorpions on G. echinatus spiders produce AND SPIDERS IN THE NAMIB DESERT such patterns of distribution and abundance? Ad- ditions of Uroplectes over a one year period to scorpion-free trees (n = 11) highly significantly <.001) reduced G. echinatus populations to Predation by scorpions on spiders also appears (p 42% of that on control trees (n = 12). Removal of to be a key determinant of spider densities in the scorpions from trees (n= 8) also produced a high- Namib Desert (Polis and Seely, unpublished rely significant, 2.9 times increase in G. echinatus search, 1988, 1989). Both the scorpion Uroplec- as compared to control trees (n= 12) with their tes otjimbinguensis (Karsch) (Buthidae) and the full complement of scorpions. spider Gandanimeno echinatus (Purcell) (Eresidae) live under loose bark on larger Acacia These experiments showed that intraguild trees and are the major arboreal predators of predation was concentrated on young spiders and insects on such trees. Populations of G. echinatus could significantly alter age distributions (p are severely reduced when they co-occur locally <.001 for each of the following comparisons): on Acacia erioloba trees with the scorpion U. The smallest size class of spiders on experimental otjimbinguensis. This system is a model example trees represented 49% of the population (n = 728 how predator-prey interactions are complicated spiders) one year after scorpions were removed greatly by patterns of local distribution. Not all compared to only 31% (n = 221) on control trees suitable patches (Acacia trees) contain the full where scorpions remained; similarly, the smallest array of local species capable of existing within size class formed 48% of all spiders (n = 903) on the patch. Trees grow along the banks of dry trees where scorpions were not present compared rivers (e.g., the Kuiseb) and become less dense to 34% (n = 820) on those trees to which scor- and more sporadic with increasing distance from pions were added. the river bed. The local abundance of scorpions Thus differential dispersal and semi-deter- and spiders on each tree is a function of differen- ministic biotic interactions combined with differ- tial dispersal, extinction and a predator-prey ing isolation of patches are major determinants of relationship with U. otjimbinguensis scorpions the distribution, abundance and age structure of eating G. echinatus spiders. Although both these species. In general, historical and stochastic species are found on river trees, U. otjimbinguen- dispersal events in patchy environments are a sis is, with no exceptions, the numerically paramount factor explaining the distribution and dominant species (10-50 scorpions/trees) and G. abundance of predators and their prey and species echinatus is relatively uncommon (5-20/tree). of competitors (Polis, 1991b; Polis and SCORPIONS AS MODEL VEHICLES

Yamashita, 1991). Sucb conditions can produce nificantly greater secondary production of local extinctions or great variance in abundances arthropods than larger islands and mainland via deterministic bwtic interactions, but promote areas. However, high productivity is not from global coexistence. I suspect lhat such situations autochtonous primary production by terrestn nl are normal among many species living in the plants. Evidence strongly suggests that alloch- notoriously heterogeneous desert. Many such as- thonous production from the ocean is the prime semblages occur in patches; differential disper- source of productivity. Many islands are 'desert sal, local extinctions and 'hide-and-seek' islands' with limited primary productivity from •.lyuamics are undoubtedly extremely important terrestrial plants Several of the most productive in determining the exact structure of local as- small islands support 2-28 plants (individuals, not semblages. Unfortunately, little research has species) yet 200-1000 spiders can occur on a focused on these processes. The system with single cac scorpions and spiders on Acacia trees is ideally Small islands are more productive than laige suited to analyze such processes represent and islands and mainland areas for two reasons- F another example of the use of scorpions to ad- the Perimeter : Area ratio of an island decreases vance our comprehension of ecological proces- with size (perimeter is a linear function; an ses square function). Thus small islands have relatively more shoreline per unit area of Land and INTERACTIONS BETWEEN SCORPIONS consequently, receive relatively more nutrient AND SPIDERS ON ISLANDS and energetic input from marine drift (shore IN THE GULF OF CALIFORNIA wrack-algal and animal detritus) Second species-area relations are such lhat small islands This system shows several of the same general lack predators of nesting marine birds; thus small islands processes as the one on Acacia trees in the Namih support birge colonies of pelicans, gulls, petrels and illustrates the importance of predator-prey and terns interactions incurring between spatially struc- Marine detri tus and material from nesting birds tured populations Spider, scorpion and/or lizard are eaten by many (semi-) terrestrial arthropods. populations on small islands (approximately Some arthropods cat dead bird tissue (ej are 1-3 orders of magnitude more dense chicks, adults) and fish scraps However* dipterun than on larger islands (approximately 1- parasites are dense consumers of nesting birds 1000km*) and the mainland; significant negative and form an important conduit of energy from the relationships occur between island size and den- sea to spiders. Trapping shows that insect abun-

sity for each of these taxa. For cxjmple, the dance is highly significantly (p<0.00l f all com- scorpion Centruro\dcs exilicauda is 2-25 time> parisons) greater in bOill supralittoral zones and more abundant on small islands. Three major around nests than other areas on these islands and variables likely explain the great variance in on islands with colonies of marine birds cum spider abundance. 1) the presence ot scorpion pared to those without colonies (also see Due and predators (often absent from small islands); the Polis Dietary analysts shows thai 2) T 1985); dispersal and colonizing ability of spiders relati ve detritivorcus and parasitic arthropods act as con* to scorpions (better colonizers of small islands as duitz of energy from the sea when converted to compared to scorpions), and 3) differential cr.ci- large populations of terrestrial spiders, scorpions 2-5 gy flow from marine tu terrestrial systems (much and/or lizard , spiders aft times greater to small island more dense on islands with colonies of manne This research has been in progress tor four birds versus those without colon years (1989-1992) in the Midrift area of the Gulf Predation is also an extremely important

I island and 6 mainland site- hetween Bahia able determining the abundance of these taxa. On de Los Angeles (Baja California del Norte, Gulf Midrift islands, spider abundance is sig- Mexico) and Guaymas (Sonora. Mexico) (Pol is nificantly lower in the presence of scorpions ami unpublished). Scorpion and spider abundance lizards ('particularly Cetuturouies exiiicauda$x\d 3 were quantified at each site: spiders were counted Ufa stansburiana' 1990: 0,19/m cjlIus 1991: : on >4000 cacti (one sample unit) and >8000m of 2.5/nr I versus then absence (63.4/m'; 28.2/hv\ supralittoral shoreline (another sample unit). In- both p<.O05>, The importance of scorpions but sect abundance was estimated at these sites for not lizards on spider abundance is suggested by

>1000 trap days. the analysis u\~ 10 islands Deal : IS ill 2 Small islands (

408 MEMOIRS OF THE QUEENSLAND MUSEUM

(0.4 I'm* cactus) on all islamls with C ,i>f;-randa mity and predation is tequired to understand the (n=8> compared to those with lizards (15.3/nr distribution and abundance of these taxa. Each cactus: n=2). factor varies more or less regularly with island The final variables in this system are spatial size: generally, as size increases, secondary structure and colonization ability. Incidence productivity decreases, predation increases and functions (percent occurrence on islands clas- the importance of differential colonizing ability sified by area) describe llie ability of taxa to diminishes. Multivariate analysis allows statist! disperse to and successfully remain on islands of cal dissection to determine the relative contribu- different sizes. The incidence curves suggest that tion of each of these factors to the observed relative ability to colonize small islands roughly variance in spider density. This anaJysis (Table 6) can be ranked: Spiders> Via lizasds> Scorpions> shows that spider density is a significant positive predators >Songbirds. Coloniz-ation would function of prey availability and significantly [emiinistic if a particular taxon were alway s depressed in the presence of scorpions. present or absent for all inland-, of a particular size Arrhropodivorouv lizards arc a seemingly unim- class. In fact, incidence values for smaller b portant factor, explaining almost none of the size classes arc neither nor lOOvr suggesting variance in spider abundance on Midrift islands. that presence or absence of a particular taxon is In summary; Productivity sets potential maxi- somewhat stochastic. Thus some small islands mal population size. Small islands are much more exhibit high densities of spiders because scor- productive than lujgerislands because of the rela- pions are absent whereas other similar sized is- tive greater input of marine allochthonous lands exhibit low densities of spiders because productivity from drift and marine hints scorpions are present Colonizing ability establishes the insular species An integration of colonizing abilities, produc- combinations; species-area relations show thai larger islands are more diverse and support more types of predators. The realized abundan; 1990 Ant wtttalbanfc ' Jf . Ms F t»r* terrestrial taxa is limited by (intraguild) preda- Repression A ,.:. 0.8 DJXHffl tion. For example, if scorpion predators arc ab- Error 1 1.42 i [>9 sent, spiders are dense on small, high produetiv ity re 4.65 islands. When scorpions are present, spider den- Perimeter" Area 10.3 io»sh sity is lower (but still higher than on large islands SLwpion presence 4.29 0.0559 and the mainland) and the density f>F seoqiions IS Liaidfucspnce 9J2 d.5h:7 relatively high. As island size increases, produc- Cactus volume IJ0 2902 i tivity decreases hecause nest predators arc 1 Tofali*" this model =0-70 present (thus bird colonies disappear) and alloch- IWI dnUwnh lir*rds SS MS p p>H df thonous detrital input decreases as a function of Regression 3 3.13 1.04 11.01 u.nnru island Perimeter ; Area ratio. Eventually, as is- Error 16 1.51 0.09 I land size increases (with decreases in produc- Total 19 1.65 tivity and increases in predation), the abundance | PeriineiLf Arut 9.06 of spiders, lizards and scorpions decreases until Seorp«on presence 5.30 0jO285 abundance on very large islands approaches that

L * / ii rtl presence 1.1.76 0.9687 of the mainland. Strong predation from many

Total R' this model =0.671 sources occurs on the relatively low productivity mainland; consequently, populations of spiders, scorpions and lizards are quite low. TABLE 6. Multivariate regression of factors that may influence lain the Gulf of California in 1990 and 1991. The maximum R im- CONCLUSIONS provement technique is used; this produces the best model all the independent variables. Inde- given This paper presents various types of data to pendent variables include lizard presence, scorpion evaluate the 'ecological importance of scorpions presence perimeter: area ratio of island, mean cactus in deserts. Ecological importance was first con- volume/island and prey availability/island. The best sidered in of population energetics (the two variable model includes perimeter: area ratio and terms scorpion presence. Three variable mudcls arc quantity of energy and mass flowing through presented with lizards, The effect of lizards is always scorpion populations) and second, in terms of the noo-si^mficantly weak, regardless of what higher regulation of community structure and dynamics order model is used. (how intraguild predation by scorpions influen- SCORPIONS AS MODEL VEHKULS 409

ccs the distribution and abundance of their com- contributed their time and energy to this work. petitors/prey). The energetic analysis provided Key colleagues include Sharon McCoiinick. three conclusions: 1 ) Scorpions are quite diverse Mary Seely\ Steve Hunl. Rick Fleet. Mike Qin- and abundant in deserts. They monopolize a rela- lan. Tracey Wadsworth, Victor Fat, Chris \i tively large inure of animal biomass in desert Demse Due ; Tsunemi Yamashita and Sharon communities, particularly relative to verteh Lee-Polis. Bill Humphreys' careful reading has pud other arthropods 2) Their relative success is improved this paper greatly \ thank the fine Etsifl due to a suite of autccological traits that are at the Queensland Museum (especially Robert particularly suited |p the hursn nnd variable Raven) and the Western Australian Museum for climatic conditions of deserts. These traits (very orchestrating such a great congress and for low metabolism and very high assimilation ef- providrng support for my attendance. ficiency leading to low energy requirements: and great tolerances to water stress, heat and starva- LITERATURE CITED tion) preadapt them to prosper successfully under the unpredictable and low productivity food BRADLEY, R.A. 1983. Complex food webs and availability that characterize desert*. 3i However. manipulative experiments in ecology. Oik: these traits (low metabolism and high assimilation efficiencies) lessen their impact on ener- I9-8G. The relationship between population density getics, prey and competitors. Thus the of Punt rot fort us ufahensis (Scorpionida- Vaejovidae) and characteristics of its habitat. importance of scorpions relative to horncother- Journal of Arid Environments 11; 165-171. mic vertebrates is less than an analysis of deftsj ' v 1988. The influence of weather and biotic factors on arid biomass suggest because scorpions require the bchaviom of the scorpion [ParuMCtatOtt and process prey in quantities relatively low to u(ahcnsu). Journal of Animal Ecology S7 their biomass. Nevertheless, as a group, Scorpions 551 are probably important conduits nf energy flow DUE. A.D. & POLIS, C A. 1985. Diology of the inier- in deserts. lidal sewpion, Vacjom Uftomlh lournal of Zool- ogy 207 563-580. I he research on the interactions among scor- 1 986 Trends in scorpion diversity along Baja pions and that between scorpions and spiders the California peninsula. American Naturalist T2M: strongly suggests that scorpions can play key 460-468. roles in the structure and dynamics of the com- HUMPHREYS, W.F. 1979. Production and respifSttion munities in which they live. These studies m animal populations. Journal of Animal Ecology showed that intraguild predation by scorpions 48: 427-4S4. was a major force determining the (temporal and LAMORAL, B. 1978. Soil hardness, an important, and spatial) distribution, abundance and age structure limiting factor in burrowing scorpions of the of populations of their competitors/prey. How- genus Opisthaphtluilums C.l Koch, 1837 r- ever, as an important caveat, these interactions pionidae. Scorpionida). Symposium ol the Zoological Society of London 42: 171-181. must be viewed in I he context of the environment MARPLES, T.G. & SHORTHOUSE, D.J. 1982. An in which they occur. Dispersal ability, spatial energy and water budget for a population of Arid structure and productivity arc just some of the /.one scorpion Urodacus yctschenkai (Birula possible important factors that moderate the 1903). Australian Journal of Ecology 7: 119 predator-prey interaction between scorpions and MCCORMICK, S.J. & POLIS. G.A. 1990. their intraguild prey. prc<nr* and parasites. 294-320. In Polis, d A

The role of all ecologists is to integrate these led.. I 'Biology of Scorpions'. (Stanford University factors to produce a synthetic understanding of Pa*Ns. Stanford, California). The o\ the processes and dynamics that structure natural POLIS, O A 3980a. significance caimibalis the population dynamics and surface aciivifv communities. I suggest thai scorpions arc par- natural population of desert scorpions. Behavioral ticularly suited for this task and will continue to Ecology and Sociobiology 7: 25-35. be a productive vehicle to advance the thcoit-iK:i I 1980b, Seasonal and age specific variation in the and empirical body of ecology. surface activity of a population of desen pions in relation to environmental factors. Jour- ACKNOWLEDGEMENTS nal of Animal Ecology 49: 1-18 1981. The evolution and dynamics of intraspc.ifir prcdatioit Annual Review of Ecology and Sw This paper highlights the results of ) r>- 1 f> years Uanutfcs 12:225-251. mI scorpion researcn. It is impossible to thank 1984a. AgC structure component of niclic widlh and individually all the wonderful people that have intraspeciftc resource partitioning: can age 410 MEMOIRS OF THE QUEENSLAND MUSEUM

groups function as ecological species? American habitat, and the evolution of life history strategy. Naturalist 123: 541-564. Ecology 61: 620-629. 1984b. Intraspecific predation and "infant killing" POLIS, G.A. & MCCORMICK, S.J. 1986a. Patterns of among invertebrates. Pp. 87-104. In Hausfater, G. resource use and age structure among species of and Hardy, S. (eds). 'Infanticide: Comparative desert scorpions. Journal of Animal Ecology 55: and Evolutionary Perspectives'. (Aldine Publ. 59-73. Co.: New York). 1986b. Scorpions, spiders and solpugids: predation 1988. Exploitation competition and the evolution of and competition among distantly related taxa. interference, cannibalism and intraguild preda- Oecologia71:lll-116. tion in age/size structured populations. 185-202. 1987. Intraguild predation and competition among In Persson, L. and Ebenmann, B. (eds), 'Size desert scorpions. Ecology 68: 332-343. Structured Populations: Ecology and Evolution'. POLIS, G.A., MYERS, C.A. HOLT, R. 1989. The (Springer- Verlag: Heidelberg). & ecology and evolution of intraguild predation: 1990. Ecology. Pp. 247-293. In Polis, G.A. (ed.), potential competitors that eat each other. Annual 'Biology of Scorpions'. (Stanford University Review of Ecology and Systematics 20: 297-330. Press: Stanford). 1991a. Complex trophic interactions in deserts: An POLIS, G.A. & YAMASHITA, T. 1991. The ecology empirical critique of food web theory. American and importance of predaceous arthropods in desert Naturalist 138: 123-155. communities. Pp. 180-222. In Polis, G.A. (ed.), Communities'. 1 99 1 b. Desert communities: an overview of patterns "The Ecology of Desert (Univer- sity of Arizona Press: Tucson). and processes. Pp. 1 -26. In, Polis, G.A. (Ed), "The Ecology of Desert Communities'. (University of SHORTHOUSE, D. 1971. Studies on the biology and Arizona Press: Tucson). energetics of the scorpion, Urodacus yaschenkoi POLIS, G.A. & FARLEY, R.D. 1980. Population biol- (Birula, 1904). (Ph.D. dissertation, Australian Na- ogy of a desert scorpion: survivorship, micro- tional University: Canberra). INTER-SPECIFIC ASSOCIATIONS INVOLVING SPIDERS; KLEPTOPARASITISM, MIMICRY .AND MUTUALISM

MARK A ELGAR

Elgar, M.A. 1993 11 11: Inter-specific associations involving spiders: klcptoparasiusin, mimicry and mutualism. Memoirs of the Queensland Museum 35i2): 41 1-430. Brisbane. ISSN 0079-SS35

Many spiders have life-styles that involve a relatively close and prolonged association with another species; for example, between a specialist predator and its prey species, or a species may rely on another for either protection from predators or providing a suitable place to live. In asymmetric relationships, where individuals of one species benefit at the expense of the other, each species may act as a selection pressure on the other species. This can result in the evolution of specific adaptations and counter-adaptations thai are evident in at least three kinds of inter-specific associations between spiders. These associations, namely kleptoparasitism, mimicry and mutualism are reviewed here. Our understanding of the evolution of these fascinating systems remains limited, despite numerous anecdotal accounts, because only a few studies are experimental. The purpose of this review is two -fold: U3 Illustrate the use of comparative and experimental studies for understanding lhe evolutionary significance of these inter-specific relationships, and to highlight thusc gaps in our knowledge that might benefit from \h\h apptvacU.C\Inter-spedftc associations, spiders, kleptoparasithm, mimicry, mutualism.

MwkA. Lii>ar, Department qfZoology, University ofMelbourne, Parkvilie, Vtcxeri Australia; 18 January, 199X

Individuals of one species can affect in- (e.g. Endlcr, 1986; Daviess ai, 1989; Toft tiaU dividuals of another as a result of competition, 1991), but there is also some evidence for similar predation, parasitism or mutualism. The evolu- processes in predalor-prev systems (eg Brodie tionary implications of any association between and Brodie, 1991; Endlerf 1991 ). two or more species depends critically on the frequency and nature of the interaction* For ex- INTER SPECIFIC ASSOCIATIONS IN ample, a species may be the prey of many species SPIDERS of generalist predators. While these predators may represent an important selective pressure Research on the behaviour and ecology of favouring anti-predator responses in the prey spiders has, with a few notable exceptions, species, the adaptations of the prey may have focussed on issues involving single species (e.g little impact on the reproductive success of the Humphreys, 19S8). including foraging behaviour predators. In contrast, a predator that preys on (Reichert* and Luczak, 1982; VollrauY 1987a; only one species can be an important selective Uetz, 1992), habitat choice (e.g. Reichert and force favouring anti-predator adaptations in that Giltespie, 1986), intraspecific competition (e.g. species. In turn, the prey species anti -predator Reichert, 1982), courtship and mating (e.g. adaptations can exert selective pressure on the a Robinson and Robinson. 1982; Elgar, 1992) and predator, favouring improved predatory abilities social behaviour (Buskirk, 1981; Elgar and each species acts as a selection pressure on Thus, Godfrey, 1987; Uett, 1988). Nevertheless, some adaptations the other, favouring and counter- spiders have relatively specific and prolonged adaptations, perhaps leading to characteristics relationships with other species. These relation- that are increasingly specific to the relationship. ships often involve predation or avoiding preda- However, these improvements need not neces- tion, and perhaps a reason why interspecific sarily change the relative position of each interactions involving spiders have been protagonist (Dawkins and Krebs, 1979). The neglected is that spiders are frequently perceived evolution of these specific adaptations and as generalist predators; cursorial or wandering counter-adaptations depends on both the frequen- spiders attack any vulnerable prey that they can cy of the interactions and the effects of each find, while web-building spiders simply capture protagonist. Host- parasite systems provide a rich any prey that is caught in their web. However, (he seam of examples of such evolutionary processes view that spiders are generalist predators is mis- 113 MEMOTRS OF THE QUEENSLAND MUSEUM

leading. Many spiders prey on only a few specks Mysmenidae, Symphytognaihidae and using foraging techniques that include building Theridiidae (Table I). Of these spiders, the genus specialised webs; producing chemical com- Argyroses (Theridiidae) is the best documented pounds that attract prey; utilising the webs or (see Vollrath. 1984T 1987b). capturing capabilities of other spiders; mimicking prey behaviour; and cooperative foraging {see EVttJGNCe OP KLEPTOPARASmSM reviews in Stowe. 19S6: Nenrwig, 1987). Clearly, Argyrodes were originally described as com- the survival and reproductive success of both mensals: Argyrodes benefit by feeding on the predaior and prey will depend on their predatory prey items that are caught in the host's web. but and defensive behaviours, and the degree to the host is not disadvantaged hecause these prey li the predator depends on the prey as a items do not form part of its diet (e.g. Belt. 1 874). l of food. Not all associations between However, subsequent behavioural and ecological spiders and other invertebrates are predator-prey studies revealed that individual Argyrodes relationships, some species depend on other remove prey that might otherwise be consumed species for protection from p&daiOtti 01 provid- by the predator. These observations suggest that ing suitable places 10 live. This review will focus the relationship between Argyrodes and their on three mtei specific relationships involving hosts is more accurately described as klep spiders: klcptoparasitism. mimicry and toparasiiic rather than commensal (see Vollrath, mutualism. 1984,1987b). A detailed understanding of the nature of these In fact, klcptoparasitism may also be an inap- inier-specific associations will benefit from both propriate description. A kleptoparasiiic relation- experimental and companiiive studies The ship implies that one partner in the symbiosis former can provide insight into both 1hflies, damsel fies and wasps (see behaviour of the host (Vollrath, 1984. 1987b). reviews in Vollratfi, 1984. 1987b; Nentwig and There are several mechanisms by which klep- Heimer, 1987). Most descriptions of these guests toparasitic Argyrodes avoid detection or capture are anecdotal, and consequently the nature of the by the host: many species drop from the web relationship is poorly understood. The webs of when challenged by the host, A. antipodfanuA many spiders are also host to numerous other swings away from the web when the host is spiders thai obtain food from prey caught in the agitated (Whitehouse, 1986). and A. ultdans cnl& host's web. These spider guests commonly holes in the tangle web of its social spider bosfl referred to as kleptoparasites, arc represented in Anelosinms eximius, forming a tunnel that ap- at least four families, including the Dictynidae, parently facilitates escape (Cangialosi, 1991). .

INTER-SPECIFIC ASSOCIATIONS 413

Surprisingly, the evidence that the presence of number of kleptoparasites per web is greater in Argyrodes has a negative effect on the reproduc- larger colonies. For example, Nephila edulis tive success of the host has not been directly builds webs in aggregations, and webs in ag- sed. For example, there are no experimental gregations have higher kleptoparasite loads and t -knee that the growth rate or fecundity of the infestation rates than those found alone fElgar. host is reduced by the presence of Argyrodc 1989). Re-locating a tyefc away from an aggi any other genera of klep(oparasiles). Instead, the tion may reduce kleptoparasite load, but the host negative impact ofArgyroses on its host has been Subsequently does no( bench! from the foraging inferred primarily from either the behaviour of and predator defense advantages of living within the host i e.g. Larcher and Wise, 1985), or from an aggregation (e.g. see UfctK, 1 988) Moving web estimates of the energetic costs derived from the In reduce kleptoparasite load may not be loss of prey items obtained by Argyrades. For possible frit some stX ul •• example, the number of prey items consumed by Aneiositnus thai build substantial, permanent Neplrila clavipes is reduced with increasing num- Webs, Indeed, high klcpttt; re ap- bers of Argyrcxlcs on the web (Rypstra, 1981). parently responsible for tftc demise of some and A. uMatts removes around 26% of the prey Anelosinius colonics (Cangialosi, 1990b) but not items that are caught in the web of its hosi others (Vollrath, 1982). Anelosimuy exitnhts (Cangialosi, 199' Like their klcptoparasites ho^K individual | Vollrath (1981 > examined the potential costs of also react 10 variation in prey capture rates. The Argyroses by estimating the energetic require- feeding rates of klcptoparasites ^rc Likely w* be ments of a single kleptoparasite The daily en influenced by both Ihe prey capture rate or the requirements of the 3-4 mg A. elevaius is 0.82 J, host and the number of other kierxoparasites on about i>.5% of the daily requirements of its 975mg the web. Host web capture rales may vary at: (ftephih clavipes] post This proportion in- ing to both the location and the size of the web. creases with larger numbers of kleptoparasites The number of kleptoparasites increases w ith the per web; over 40 individuals have been counted web sue of several host species leg. Elgar, 1989; on a single Nephila web (although the average is Cangliosi. 1990a), possibly because larger 2.2 klcptoparasites per web), suggesting a poten- have higher web capture rates that can support tially high energetic cost of this relationship more kleptoparasites. Web-building spiders relo- (Vollrath, 1981). cate their webs Btf to prc> capture pates If KJeptoparasites exact a cost on host reproduc- (e.g. Gillespie and Caraco. 1987), dndArgyrodes tive success, then selection should favour any may behave similarly by moving to differ* r,i trait that enables the hosts to reduce ihat cost. webs (hut see lurcher and Wise, 1985). It would

There are several mechanisms by which I be interesting to establish experimentally might reduce the cost of kleptoparasilism: whether the emigration rate of individual Ar- recovering the prey from the kleptoparasite; ^TOi#s»ncreasesasaresultonowerwebcapture reducing the kleptoparasites access to the prey ; or rates or increased numbers ot conspecifics. If the Mmply abandoning the web and building another latter, it is possible that the distribution or Ar elsewhere. Interestingly, hosts appear td be inef- gyroJes within a population of hosts, particularly ficient at recovering prey (Vollrath, 1979a, b; those hosts that aggregate, could be predicted by Rypstra, 1981) although several host species the ideal tree distribution (see Milinski and reduce access to their prey by chasing the klep- Parker, 1991 j. toparasites (Cangialosi, 1990b) or concealing the A possible option for klcptoparasites thar e* prey in retreats [see Cangialosi. 1990b). Larcher perience a low feeding rate is b capture and .m.l Wise (1985) demonstrated experimentally consume the host before moving tn tbe web of that hosts are more likely to abandon webs when another host (e.g. Tanaka. 1984). Some species Argyroses are present ihun absent. Nephila of Argyti either obligate or facultative clavipes relocates its web when it is infested with predators of their hosts (see Table 1 1 Predatory large numbers of klcptoparasites (Rypstra, 1 98 ), 1 Argywdet can capture the host through mir although the behaviour of /V. clavipes may be a ing a prey item (e.g. Whitebouse. 1 986) o response to lower feeding rates, rather than to advancing toward the host and attacking it. Such numbers of kleptopMiasitt-s, a specialised form of prcdation i& not uncomni* m Social or communal spiders appear to have in spiders, and has been recorded in several fewer defensive options against high klep- families leg. Jackson, 1987; Jackson and b roparashe loads, and this cost may be higher if the 1982. Jackson and Bras-sington, |9g7 414 MEMOIRS OF THE QUEENSLAND MUSEUM

TABLE 1: Spiders thai are presumed to be kleptoparasites of web-building spiders. Families; Ag, Agelenidae; Am, Amaurobiidae; Ap, Aphantochilidae; Ar, Araneidae; CI, Clubionidac; Co, Corinnidac; De, Deinopidae; Pholcidae; Pro, Di, Dipluridae^ Er, Eresidae; Gn, Gnaphosidae; Lin, Linyphiidae; Lio, Liocranidae; Ph T Agg, Prodidomidac; Sa, Sallicidae; Td, Theridiidac; Tm T Thomisidae; Ul, Uloboridae; Zo, Zodariidae. Ag- gregates; Soc, Social; Sol, Solitary Web types: O, orb; T, tangle; S, sheet; F, funnel; Sp, spacc.Body sizes in mm.

Body size Host Guests Assocl Source Kleptopar*si(ic laxa per web i N Family Species Size Web Social Dicrynidae GnsNsuld&Metkle- Siegodyphus T Soc K Archaeodicryna ntova Er Gns wold (1987)

Heteropodida*

Olios d'tana Urn Badumna Candida Soc 20 K Jackson (1987)

O. sn. indet. Am Badumna Candida T ' Soc 10 K Jackson (1987)

' «

; { !l r T Soc 10 K Jackson (19S7) '. (arnan i i Stegodvphus sarasinorum

O. obesulus B. Ste eodvphus sarasinorum It Soc :: K Jackson (19871 M vsmenidoe

Isela okuncaiui Di AUothele terreth F Sol K Ciriswold(19S5i

KtUfui tntpiitiiui Di Thelerfwris kanclii F K Coylc etal. (1991) Mysmenopsis archeri Ph T K Baptisa(l988) M. copue Ar Cvrfophora O K Baert(1990)

M citlrelicotu Di K Coylee-la/. (1991)

M. cienga Ar Cyrtophora (.) K Baert(1990) M- dipluroamiga Di Diplura s K Vollnllh(1978)

M. furfiva L5 J_Di Jschnothelexera T Sol 14 K Coylc etal. (1991)

M. gamboa Di Dipiun ! S K Vollralh(1978) K Co\\cel at (1991,1 M. hauscar Di - M- isthnamigo Di Diplura S K VolIrath(I97H) CoyJe& Meigs Di IsclmoiheU sp. T Sol 4 K M. monticata (1989)

M- pachacutec Di K Coy]c etal. (1991>

i M palpatis Di K Coylc c( at l 1991

M tibialis Di K Coylecfa/. II99D

robins estralis |k Briitowe(I95S3 > 8 Ootwps pukher 2.0 | 1.5 | Am Aman fen |T SaUicidAe

Sinmeiha paetuta 7 7,0 Am Badtunna Candida Soc KP Jackson (1985) Symphvlognathidae

] VoHraihi 1978) Ofrirttapua baxano 3 1.3 1 Di Diplura sp. 1 40 | S Sol K Theridiidae

Ag Cambridgea sp. F Sal KP Whiiehuuse (1988a) antipodianus

A. antipodianus Ag Stipiudion !" Sol KP Wmiehouscl 1988a)

A. antipodianus Am Badumna tonginquus Sp Sol KP Wh)ieboiiscll9KKai

A. antipodianus I Ar Araneus crassa Sul KP Whnehou5e(19SSa)

A. antipodianus i Ar Lriophara nustulosi. 13 Sol KP Whiiehou^ei l9SKa)

A. unfipudiatms Ar Cvclosa trilobata 8 o Sol KP V. Iiiiehouse i >9S8aj

A. antipodianus Ar hcucaugc dromedoria o Sol K F Whitehousel 1988a)

A. (uUipodiantts Ph Phafcus phalangioides SP Sol KP Wbi(ehouse(T988a,l

J A. antipodianus Td Achaearanea T Sol KI Whitehousc (1988a)

A. antipodianus 3 : s ... Cyrtophora furta !: O Sol K mgzierat (1983)

-V antipodianus 3,0 2-5 \r Septula eduhs 21 O Agg K Elear(19B»j INTER-SPECIFIC ASSOCIATIONS 415

Body size Host Guests Assoc* Kleptoparasitic taxa Source per web 6 v Family Species Size Web Social

A. argyrodes Ar ' Cyrtophora citricola O K Vollrath(1984) A. atopus 2.4 3.4 Ar Nephila clavipes 25 O Agg 5 K Vo!!rath(1987)

A. attenuatus 17.0 9.3 none Eberhard (1979a)

Exline & Levi A. baboquinari 3.7 3.5 Td Latrodectus T Sol K (1962)

A. baboquivari 3.7 3.5 Ul Philoponella oweni 6 O Agg KP Smith Trail (1980) Exline & Levi A. cancellatus 3.2 3.8 Ag Agelenopsis F K (1962) Exline & Levi A. cancellatus 3.2 3.8 Ar Argiope aurantia 22 O K (1962)

Exline & Levi A, cancellatus 3.2 3.8 Ar Araneus strix O K (1962) Exline & Levi A. cancellatus 3.2 3.8 Ar Mecynogea lernniscata O K (1962) Exline & Levi A. cancellatus 3.2 3.8 Ar Metepeira labyrinthea O Sol K (1962) Exline & Levi A. cancellatus 3.2 3.8 Ar Nephila clavipes 25 O Agg K (1962) Exline & Levi A. cancellatus 3.2 3.8 Ar Verrucosa arenta 9 O K (1962) Exline & Levi A. cancellatus 3.2 3.8 Li Frontinella pyramitela 4 S K (1962) Exline & Levi A. cancellatus 3.2 3.8 Ph Pholcus space Sol K (1962) Exline & Levi A. cancellatus 3.2 3.8 Td Theridion tepidariorum 7 T K (1962)

A. caudalus 3.1 3.9 Ar Argiope argentata 16 O Sol K Smith Trail (1980) Exline & Levi A. cochleaforma 2.7 3.7 Ar Argiope O K (1962)

Exline & Levi A. cochleaforma 2.7 3.7 Ar Gasteracantha O K (1962)

A. colubrinus 25.0 none Eberhard (1986) Exline & Levi A. cordillera 3.6 3.1 Ar Gasteracantha O K (1962)

A. cxlindratus Ar Araneus ventricosus O Sol K Shinkai(1988)

A. dracus 2.3 2.6 Ar Nephila clavipes 25 O Agg 5 K Vollrath(1987)

A. elevatus 3.4 4.0 Ar Argiope argentata id 4 K Vollrath(1979) Exline & Levi A. elevatus 3.4 4.0 Ar Gasteracantha o K (1962)

A. elevatus 3.4 4.0 Ar Nephila clavipes 25 o Agg 5 K Vollrath(1979 Exline & Levi A.fictilium 10.0 5.0 Ar Araneus o KP (1962)

A.fictilium 10.0 5.0 Li Frontinella communis 4 s Agg KP Wise (1982)

A.fictilium 10.0 5.0 Ul Philoponella oweni 6 Agg KP Smith Trail (1980) Exline & Levi A.fictilium 10.0 5.0 none (1962)

A. fissifrons 7.0 Ag Agelena limbata 16 F Agg 2 KP Tanaka(1984)

A.fissifrons 7.0 Lin Linyphia s KP Tanaka(1984)

A. fissifrons 7.0 Td Theridion japonicum T KP Tanaka(1984)

A.fissifrons 7.0 Ul Philoponella sp. 3 O Agg 3 KP Elgar (pers obs)

A. fissifrons 7.0 Ul Uloborus varians KP Tanaka(1984)

A. flagellum none Eberhard (1986)

Exline &Levi A. globosus 2.3 2.3 Ar Nephila clavipes 25 o Agg K (1962)

A. incisifrons Ar Cyrtophora hirta 14 Sol K Elgar era/. (1983)

Table 1 . continued C

416 MEMOIRS OF THE QUEENSLAND MUSEUM

Table 1. continued

Body size Host Guests Assoc 4 Source Kleptoparasitic taxa per web 6 9 Family Species Size Web Social Gray & Anderson incursus 3.8 2.2 Td Achaearanea mundula 6 T Sol 5 KP A. (1989) Exline & Levi 24.0 19.0 none A. longissimus (1962) Robinson & miniaceus Ar Nephila maculata 43 O K A. Robinson (1973) Exline & Levi nephilae 1.7 2.2 Ar Argiope O K A. (1962)

A. nephilae 1.7 2.2 Ar Cyrtophora motuccensis 19 O Agg K Berry (1987) Exline & Levi A. nephilae 1.7 2.2 Ar Gasteracantha O K (1962) Exline & Levi A. nephilae 1.7 2.2 Ar Neoscona O K (1962) Exline & Levi nephilae 1.7 2.2 Ar Nephila O K A. (1962) Robinson & A. nephilae 1.7 2.2 Ar Nephila maculata 43 O K Robinson (1973) Exline & Levi A. pluto 3.9 3.7 Ar Argiope aurantia 22 O K (1962) Exline & Levi A. pluto 3.9 3.7 Ar Metepeira labyrinthea O Sol K (1962)

Exline & Levi A. pluto 3.9 3.7 Td Latrodectus T Sol K (1962) Exline & Levi A. proboscifo 2.6 2.9 Ar Gasteracantha O K (1962)

A. projiciens 4.0 3.2 Ar Metazygia sp. O Sol KP Eberhard(1986)

A. sp. A Ar Nephila clavipes 25 O Agg 2 K Rypstra(1981)

A. sp. B Ar Cyrtophora moluccensis 19 O Agg KP Lubin(1974)

A. sp. Ar Gasteracantha O 23 Vollrath(1981)

A. subdolus 2.6 2.8 Ul Philoponeila oweni 6 O Agg K Smith Trail (1980)

A. trigonum 4.2 2.5 Ag Agelena limbata 16 F Agg KP Suler etal. (1989) Exline & Levi A. trigonum 4.2 2.5 Ag Agelenopsis F (1962)

A. trigonum 4.2 2.5 Ar Mecynogea lemniscata O KP Wise (1982) Larcher & Wise A. trigonum 4.2 2.5 Ar Metepeira labyrinthea O Sol KP (1985) Exline & Levi A. trigonum 4.2 2.5 Lin Linyphia marginala 5 S KP (1962) Larcher & Wise A. trigonum 4.2 2.5 Lin Neriene radiata 5 S Sol KP (1985) Exline & Levi A, trigonum 4.2 2.5 Td Latrodectus T Sol (1962) Exline & Levi A. trigonum 4.2 2.5 Td Theridion zelotypum 4 T KP (1962)

A. trigonum 4.2 2.5 Lin Frontinella pyramitela 4 S KP Suter era*. (1989) Exline & Levi A. ululans 4.0 3.7 Ar Nephila clavipes 25 o Agg K (1962)

A. ululans 4.0 3.7 Td Anelosimus eximius T Soc K Cangialosi (1990a) Exline & Levi A. weyrauchi 4.3 3.8 none (1962)

Two categories of relationship ('Assoc') are defined: 'K' is Kleptoparasite only, and 'KP' means that the guest may also capture the host. Some species of Argyrodes may be incorrectly categorised as 'Kleptoparasite only because their predatory behaviour has not yet been observed. Taxonomy of Argyrodes follows Levi and Exline (1962) and thus Argyrodes, Ariamnes and Rhomphaea are not distinguished. .

INTERSPECIFIC ASSOCIATIONS *11

Body length! that also prey on their hosts I Table 2). How

t statistic kJcptoparasilc Uepiupjrasite males of these two groups of species are not only 3nd predaior significantly different in body size (Tabic 2). ¥ Argyrfhies 1,M±I»-2(14) 5.0 ±1 .1(6) 9.41** A rgyrodes that -prvy on their hosts also specialise A Argyroses 3.2 + 0.2 TM) 3.7+0.4(5) 1.21 on smaller hosts, compared with the size of the

Dimorphism i.d/v) J M+fiOSfH) |u.82±0.I4<5) 2.5H* hosts of those species of Argyrodes that are only ti<81 2 1.5 ±2.6 6.7 + 0.S(5i 3.47 *• kleptoparasites {Table 2). These compar. data show that, as predicted, ihe difference in s in- TABLE 2: Differences in kleptopaiasite and host body between A rgyrodes and its host is greater for length according to whether ihe kleptuparasite does those species that are primarily kleptoparv; or does not a No prey on their host- compared wiih those that are also predatory. lvalues axe mean lengths ± SE (sample sizes), Values These comparat i ve data suggest that selection has compared using l-test with pooled variance. Average ihai host body size measures were obtained for ,4 rgyrodes either favoured larger body size for species * ** are both kJeptoparasitic and prey on their hosts, that have multiple hosts. p< 0,05 ; p < (Mm of it has favoured a further reduction in body size and Jackson, 1986: Jackson and Hallas, 1990) A in those species that arc primarily klcptopara potential cost of this foraging strategy is that the The laHer argument is consistent with the Meptoparasitc may become prey to the host At that kVcploparasitrsm is a specialised foraging one host specie*:- has a relatively effective strategy that evolved from a more general klep defensive behaviour: when the predatory Ar- toparasitie and predatory lifestyle (VollratL >!i:\ is detected- the host simply cuts the web 1984). thereby collapsing it and ensuring that the The evolutionary sequence leading to the di predator cannot proceed further (eg, Eberhard. genee of these two foraging strategies within 1979b), Suteret ai (1989) report that female h Argyrodes is not known; oi>e may have evolved can discriminate between eonspecific miteia ' the other, or both may have diverged males, prey items and predatory A. trigonum,

a common web-building ancestor | apparently using chemical cues, and respond jc- Whitehouse, 1986). Thus, it is not possible, cordingjy. without an accurate phylogeny. to establish whether selection has favoured an increase in COMPARATIVE PATTERNS WiltllN AftGr/t.OOiS body size with the predatory- lifestyle, or a What evolutionary sequences arc responsible decrease in body size is associated wiih a klcp tor the diversity of predator)' specialisations hi toparasitie lifestyle. Indeed, the species placed Argyrodesl Smith Trail 11080) stresses the im- within the single genus Argyrodeshy Exline and

1 1 1 others portance of the kleptoparasites' ability to identify Le\ 962) have been placed by into three the vibratory signals generated by the host, thus genera; the Annnmes, the Rhomphaca. and the Argyrodes. In this classification, the allowi ng them to stal k and safely capture the host Ariamncs However, kleptoparasites that also attempt to and Rhampluu-a groups arc primarily fl capture their hosts nsk being captured the? ' predators and the Argyrodes group are klcp- ves. Consequently they may be more likely to topar (Whitehouse, 1987) Thus, the attempt to capture vulnerable hosts, such as those differences described above may be confounded (sec below) Resolv- that are smaller (e.g. Smith Trail, 1980; Larcher by taxonomic associations of these issues is most likely achieved and Wise, 1985). moulting (e.g. Vollrath, 1984). ing some ven the spiderHnp.s of the host by experimental manipulation of individuals Whiterrouse, 1986. within a species that shows both kleplopar and predatory behaviour. Ifrelative body size is important in determining the outcome of attacking the host, then predatory An additional pattern revealed by comparative species ofArgyttufesiiiay be largerthan prim analysis also deserves experimental ioveslign kleptoparasitic species, or the fi ay tend to Hon. The degree of sexual size dimorphism (male specialise on smaller hosts. These predictions are lenguVfemak length) eovaries significantly with supported by comparative data of body length ihe foraging strategics of Argyrodes, Males are measures for 20 species o\ Argyrodes (see Tabic smallerthan females in those species that prey on I). Species of Argyrodes were divided into two their host, consistent with patterns of size dmuu gioups. according io whether they preyed on their phism in almost all other spiders I e.ij. Elear et at. . hosts: females el Argyroses that are only klep- 199CV; Elgar, 1991, 3992; Vollrath and Parker. toparasitic are significantly smaller than those 1992). However, males of those species of Ar- '

418 MEMOIRS OF THE QUEENSLAND MUSEUM

Host taxa | Kleploparasite* Biology Host range

t statistic Ageienidae Kleptoparasite Kleptoparasite only and predator Agelena (2), Agelenopsts, Argyrodes (4) KP Catnbridgea, Stiphidion No. species* 18 6 Amaurobiidae Family 1.6±0.3 2.8 ±0.7 ] 89 Simaetha, Badumna KK Species 2.7 ±0.7 3.5 ±1.3 0.57 (2) Argyrodes Amaurobius Oonops K Argyrodes that are Araneidae TABLE 4: Mean host-ranges of either only primarily kleptoparasitic or they also prey Araneus (5), Argiope (4), Cyclosa, on their host.* refers to Argyrodes. Cyrtophora (2), Gasteracantha, Leucauge, Mecynogea, Metazygia, Argyrodes (18) KP Metepeira, Neoscona, Nephila (3), of Argyrodes. It is likely that both host-range and Verrucosa parasite-range will expand as more records be- Cyrtophora Mvsrnenopsis (2) KK come available. In contrast, many species appear Dipluridae to be host specific, with one species of klep- Allothele Isela K toparasite recorded from the web of only one Mvsrnenopsis certain Diplura, Ischnothele KK species of host. For example, in Peruvian (U) habitats, Argyrodes ululans is found only on the Diptura Cuhmagua K webs of the social spider Anelosimus eximius, Thelechons Kilifia K despite considerable effort searching for this Linyphiidae kleptoparasite on other potential hosts (Can- Frontinella (2), Linyphia, Neriene Argyrodes (4) KP gialosi, 1990a). Phulcidae Vollrath (1984) argued that Argyrodes can be Photcus Argyrodes (2) KK placed in two general categories; specialists that Pholcus Mysmenopsis K are host specific but behaviourally versatile, and Theridiidae generalists that invade the webs of many different Achaearanea, Theridion (3), Argyrodes (7) KP species but use relatively few techniques to ob- Anelosimus, Latrodectus tain food. Thus, Whitehouse (1988a) considers Uloboridae Argyrodes antipodianus a specialist, primarily Uloborus, Philoponella Argyrodes (4) KP because its behaviour is versatile, and adults are found primarily on the webs of Eriophora pus- TABLE 3: Summary of taxonomic distribution of tulosa. Dichotomies like these can be misleading kleptoparasites (see Table 1 spiders that arc host of because both host-specificity and behavioural for further details). KP= Kleptoparasites and versatility are most likely continuous rather than predators; K=kleptoparasites. * No. species in each discrete variables; A. antipodianus is found on the genus in parentheses. webs of several other hosts (Table 1). Further- gyrodes that are primarily kleptoparasitic are more, host-specificity may also vary between generally larger than their conspecific females populations, depending on the diversity and (see Table 2). What factors are responsible for abundance of potential hosts in different popula- this reversal of size dimorphism patterns within tions. For example, A. antipodianus in this group of spiders? One explanation is that Whitehouse's (1988a) study may be found competition between males for access to females primarily on Eriophora pustulosa because that is may be more intense for kleptoparasitic spiders, the most common host in her New Zealand and consequently sexual selection has favoured population. large male size in these species (see also The host range of Argyrodes may vary accord- Whitehouse, 1988b). ing to whether the species is both kleptoparasitic There is considerable variation in both the num- and predatory or whether it is only klep- ber of species that are host to each species of toparasitic. Purely kleptoparasitic Argyrodes kleptoparasite and the number of kleptoparasite may escape host-detection through specialised species found on each web-building host species behaviours, but these behaviours may be effec- (see also Vollrath, 1984, 1987b). For example, tive for relatively few host species. If so, the host Argyrodes cancellatus are found on the webs of ranges of primarily kleptoparasitic Argyrodes at least ten different host species from five may be less than for species ofArgyrodes that are families (see Table 3), while the orb-weaver also predatory. The comparative data provide Nephila clavipes is host to at least seven species little support for this prediction (Table 4); al- INTERSPECIFIC ASSOCIATIONS 419

though the host-range of primarily klep- described aJlopatric Ischnothele morphs that also tnpaiasitic species is less than the range of appear to be sister species. Third, diplurids may predatory species, the difference is not statistical- be more sensitive to web invaders than the hosts ly significant. of Argyrodes* and thus the kleptoparasitk be- The results of these inter-specific comparative haviours required to avoid detection by one host analyses within the genus Argyrodes should be species are not appropriate for another. In this interpreted cautiously. These patterns may be regard, it is noteworthy that Argyrodes are not confounded by an association between foraging known to invade diplurid webs, despite the tH strategy and taxonomic affinity, and thus the dif- taxonomic range of their hosts. ferences in body size or host range may be due to The relatively permanent nature of the host's Other, unknown features that differ between these web is a common characteristic of the hosts of all two groups. This possibility is especially relevant klepioparasites (Table I) Kleptoparasites that given the ambiguity of the taxonomic arrange- live on permanent webs may benefit by spending ment of this genus. Furthermore, some species of less lime searching for new webs compared with Argyrodes may be incorrectly assigned to those that are associated with hosts that frequent-

primarily kleptoparasite status simply through ly move their webs However, it may not be the lack of observations. Thus, the patterns may permanent structure of the web that is important, change when more data and/or a more accurate but rather the tenacity of the web site. For ex- phylogeny become available. Nevertheless, the ample, the large, nocturnal, Australian orb ms suggest several interesting questions that weaver Eriophora transmarina builds a new web

could be resolved by an experimental approach every evening and then destroys it the following dawn. Despite the temporary nature of its web, this spider is alsu hosi to many individual Ar- T SPBCJRCm OF K1EPTOPARASITH5 gyrodes, probably because it has a high web-site Both Argyrodes and Mystnenopsis belong to tenacity (M Hcrbcrstcin. unpublished data) web-building families and thus are relatively e phylogenetically (Coddington and Levi, 1991). However, the range and taxonomic af- MIMICRY BY SPIDERS finities of their hosts are substantially different

1 hie i ). Argyrodes have been recorded on the Many species of animals, including spiders, webs of 29 host genera from eight families resemble other, unrelated species. These (Ageienidae, Amaurobiidae, Araneidae, resemblances may be visual, chemical, be Linyphiidae,. Pholcidae. Psechridac Theridiidae. havioural or acoustic and are usually referred to UJoboridac), and some species have many hosts as mimicry. There are many different types H

• sre above), in contrast, !! of the 14 species of mimicry, whwrhhasprccipitaied some contra Mystnenopsis are found on diplurid hosts, with sy over its definition (e.g. Endler, 1981; Pasteur. the remaining species found on Cyrtopltora 1982). Two general forms of mimicry axe di (Araneidae) and Pholcus (Pholcidae). A com- guished in this review: defensive mimicry and parative analysis reveals a significant difference; aggressive mimicry. In the former, the mboetfc every species of Mysmerwpsis has only one host form is presumed to have evolved because the

species, while the host range for Argyrodes is 2.7 risk of predation (or parasitism) on the mimic bfi

<±0.7. n*l*) species, or 1.6 (±0.3) host reduced as a resuJt of IIS resemblance to the families. model. The lower mortality occurs because the

Why is Mystnenopsis more host-specific than receiver (the predator) does not usually prey on Argyrodes? There are several possible explana- the model, and fails to distinguish between it and tions. First, kleptoparasitism may have evolved the mimic. Aggressive mimics resemble some more recently in Mysmenopsis than in A rgyrodes, feature o{ (heir prey species thereby increasing and therefore the former klcptoparasitc has had the chance of capturing the prey model. Both Ie3$ time to expand its host tangc. Second, the forms of mimicry occur in several families of present associations between Mysmenopsis and spid: dipllifids may have evolved from a common an- The relationship between mimic, model and cestor and subsequently spcciatcd as host/klcp- receiver is asymmetric; only the mimic benefits toparasite pairs. Consistent will) this is Coyie and and any improvement in the mimic will be

Meigs ( 1989) description of two sister speci favoured rapidly by natural selection. Both the kleptcparasttes (Mysmcrt

420 MEMOIRS OF THE QUEENSLAND MUSEUM

i (axon fit Taxa Spider A V.UTll Fain Subfamily Species

Caxtianeira dubium CI Ponerinae Pachxcnndvh obscuricornis Reiskind(t977) Mazax pax Co Ponerinae Ectatomma ruidum Reiskind(l970) Mvrmecium bifaxciatum Co Formieinac Ctunponntus femoratus OHveOT([988)

Myrmecium bifusciatum. <„ Mvmucinae Megalomyrmex modesfa Ohvciraf WS81

Myrrnt'OU-i Cn Ponerinae Pachxcoodvla unideniata OftveirH

Myrtnecorypax cubanas Co Formieinac Camponotus planatus Myers and Sail ( 1 926)

M\rmei ot\pus fuligmosux Co Forrnicinae Camponotus planatus Jackson & Dnininumd 1. 1974) — Myrmecotypai rettenmeyert Cn Fonnieinae Camponotus sericeWeturis Reiskind(l977)

Sphecoivpus rtiger Co Ponerinae Pacfrvcotidvla villoxa Olivcira(!988)

Micaria pulicariu Gn Myrmicmac AcurUhomyrme.x niger Bristowfl(l958j

Micaria :.p. Un Myrmiemae Apha encgaste r beccarii HingMon(l927)

Micaria scintillans Gn Rormicinje Formica fusca Hristowe (1941 1

Phrurolithus festivus Lio Formieinac Lasias *N>' r Bnslowt i 1941 i

Phrurelithus minimus Lin Forrnicinae Formica Jttxi a Hri\Mwc(l94I)

Marietta furva Sa Forrnicinae Camponotus brevis Rcisiund(1977) ANT MIMICRY Mariettafurva Sa Formieinac Campanulas fastigams Reisttjnd(1977)

ONLY Mvrmarachne e}vn$ota Sa Pseudomvrmecinae Tftruponera anthracina Edmunds 1. 1978)

Mxrmaruchnv foetuses Sa htnnicinae Qecophvlla longwoda Fdmunds i 1978) Myrmarachne forrmraha Sa Formieinac Formica rufa Bris1owe(194l)

Mxrmaruchnt legpn Sa Forrnicinae Camponolus acvapimensh Edmund* (1978) Mvrmarachn? lupala v. Forrnicinae Polyrhachis Jackson (1986'i

M'vftnurachne paralUta Sa Ponerinae Pachxiondyla caritmlata Reiskind(1977j

Mvmui rachn%' pa raileio Sa r'vwrmac j'utvltt striatinodis Rciskmd(1977)

Mymmrachne plataleotdes Sa Fonnicmae Ocropbyila smaragdina Maihcw( I954l

Myrmarachne sp. ij Forrnicinae Foivrharhis simplei Hingslon(I927)

Myrmarachne sp. Sj Fonnieinne Prrnolepis longicorn is Minpsion0927)

Mvrmarachne s^ Sa Myrmlcinae ioteindica UfngstOn (192-7)

._ '., -.',.. Sa Forrnicinae Cft'npt'KOiUi pUivanv; .lav k son Dnimmond i 1974 Cullcr(l99M S\na j,' eh's: otx idenuilis Sa Fonivianae Myrmica emetiauui

Svnageles twiJentidts Sa Forrnicinae Lasius aUenus Cmlcr(199J)

l:iigelhaidn Syiuinelftvettdior Sa Forrnicinae LashlS niger i J970)

Svnemosvna >;p Sa PseudontyntiActnae PsetuU'tnvrmexme.xicaiui'j ReiskJnd»1977) S\aemosyna americana Sa Psendomynnev mac Pseudomvrmes boapis R.;i;l-in-m97;i

Synemosyna auratitiacti Sa Pseudomvrmet inae Fseudnntyrmes OIiveira(]988) Synemoxyna smithi Sa Pscudumymieeinae f'yruihjmyrrtuivlo'i^ma Myers and Sail (1926)

Sytiemoxvna smithi Sa Psciidoinvrmccinae PseMdomymuiJlaiiitu Myeis and Sail (1926)

Zuniea taeta Sa Formic inac Camponotusfemo ratus 01ivdra(l988) Zuniga magna Sa Ponerinae Pachvcotutyla viltosa 0]ivcira(J9S3,)

Altutea fbffmcaria n\ Mvnnicinac Cheianer croccenentre Rcisltind and Levi (1967) Dipoenu |Td MyrnVicinac Pheidole indica Hinpston (1927)

TABLE 5 (part). Taxonomic distribution of spiders that mimic ants, including those that also prey on the model, * spiders observed with dead ants. Family abbreviations given in Table 1.

food respectively (see Endler. 1991 ), and natural ves will depend upon the benefits to the mimic selection will favour models that have less and the costs of mimicry to the model. The costs resemblance to the mimic (although the strength to the model in aggressive mimicry are more of this selection will depend on the frequency complicated, and depend upon whether the model

with which the model is attacked) The degree of and the receiver (in this case, the potential prey) resemblance between model and mimic thatcvol- is the same individual. INTER-SPECIFIC ASSOCIATIONS Vi

Spidei La ion Ant l,i'.:; Somen Fara Suhfvnilv Species

Gjsiium CI Formicinae GflVijKimaiUSDgrfa Hingston (19271 4

Cosrttophtisii Sp. 1 Sa Fomiicmae CuniprtniHusdetritUi CunisiiySH)

Costnophusis sp. 2 Sa Formicinae Camponotusfulvofi Cun.;-r (1988)

• .,.-. Mymmrachne sji. sa Hormicinac Gifrtponow compr . Hm^tonn927) ANT MIMIC AND TulilmtishiiiUs Sd Formicinae Crif>ir«'/MWttJ Wins 0983)

i PREDATOR "Jo! iJciinfu.-u Tli Formictnac i .;>ph\Ua smun-.y-Jitut Cooper etal. (1990) Amwr/dfj aUntmatuiata TVn Formicinae Oetophyila smaraqdma Cooper et ai (1990)

"! : AmyeiatafoTUceps hi I uni-ii,. i...i.- Oecflphylla smaragdina Kingston |1 $27): Maihew (19541 - Bristovvoimi) ffucraiuuni Tm MyrmicnHie Cepholoies f Cryptocetus)

Apitattwchitus royersi Ap Itfynnicinati Tut nptocerus pustltus Oliveiru and Sazima (19X4 i

HpbronfjtttX braJlevi Zo Dolichodt nn;ie hidomyntte.K purpureas R. Allan (pers. coinm.)

Zodarion Zo Myrmicinac Messor tmrbunts Hiiigsioin 19271

Table 5. continued

Visual Mimicry: Spiders of Ants resemble moderately large ponerines that are within the spiders* colour range. Furthermore. Spiders that resemble ants are an especially different tnsr;ns of these spiders mimic intriguing form of mimicry that is poorly under- models of equivalent size; thus the small, black stood. Many of these spiders not .inly have an early instars mimic small myrrnicine ants, while extraordinary physical resemblence with their ant the older instars resemble medium sized attinc models, but also exhibit particular behaviours ants Such a close degree of resemblance at dif- that improves the illusion remarkably. Ant- ferent stages in the spiders' development sug- mimics, represented in at least six families of gests that the selection pressure favouring spiders and mimicking the four major subfamilies mimicry is very strong. of ants (Table 5), fall into two categories: those that appear to have little behavioural interaction Ant mimicry can provide at least three benefits, with the ants and generally avoid contort with depending upon whether the spiders prey on their them; and those spiders that specialise on captur- ant models. These benefits include protection ing and eating their ant models. There are no cle-ar from various predators, improved predatory suc- taxonomic affiliations between the species of cess on the ant prey, and both. Ant-mimics that spider mimics and the species of ant models; apparently do not prey on their models arc mostly ponerine, myrmecine and formicine ants are salticids, corinnids and a few gnaphosids (Table models for both clubionid, saltieid and other limics that prey on their models are mostly spiders. Nevertheless, certain species of ants ap- represented by thomisids and zodarids, although pear to be models for spiders more frequently there are also a few records of theridiids, corin- than others. For example, seven species of Catn nids and sidticids {Table 5). The record for the ptmotus are models to spider mimics and one saltieid species Myrmarachne (Hingston, 1927) species, C. femoratus, is a model for two corin- i 9 11 t-usual and unlikely to be typical because nthe r nids (Mynnecium) and the saltieid JAtnigu, five species of this large, ant-mimicking genus do noi species of the ponerine genus Pachycondyla are routinely prey on their model ants (e.g. Edmunds. models for five different spiders, and the weaver 1978). Only a few genera of salticids are ell U|t Uecophylla smamgdina is a model for three regular ant-predators (e.g. Jackson and van OI- species from Australia and India. It is not obvious phen. 1991, 1992). The theridiid Dipoena why spiders mimic these genera of ants more re&ernbles the decapitated head of a dead ant frequently than others. which are found in the refuse heap of the ant 111 Si Hingston (1927) suggests Uiat mimicry in this Some species of spiders mimic more than one species is aggressive because it allows the spidei species of ant. For example, the clubionid Cas- to live in the nest of the ants on whom it may prey. tianeim rica resembles species of both ponerine However, predation on these ant hosis by and myrmicine ants and the different mimetic Difxfena was not observed. forms depend upon developmental changes. COlotir variation in adult females, and sexual Perhaps the most vexing question concerning dimorphism (Reiskind, 1970). Male C. ncn defensive ant mimicry by spiders is establishing resemble,4/7n and Orfoniamochux, while females the identity of lhe receiver {i.e. the predaior or )

422 MEMOIRS OF THE QUEENSLAND MUSEUM

Spider Ant

Species Family Subfamily Species Source

Deinopts longipes De not specified Robinson & Robinson (1971)

Chrysilla lauta Sa not specified Jackson & van Olphen (1992)

Corythalta canosa Sa not specified Jackson & van Olphen ( 199 1

Habrocestum pulex Sa Formic inae Lasius sp., Prenolepis sp. Cutler (1980) Habrocestum pulex Sa Ponerinae Ponera pennsylvanica Cutler (1980)

Natta sp. Sa not specified Jackson & van Olphen (1992)

Natta rufopicta Sa not specified Jackson & van Olphen (1992)

Pystira orbiculata Sa not specified Jackson & van Olphen (1991)

Siler semiglaucus Sa not specified Jackson & van Olphen ( 1992)

Corythalia canosa Sa not specified Edwards etal. (1975)

Euryopis californica Td Formic inae Camponotus Porter & Eastmond (1982)

Euryopis coki Td Myrmicinae Pogonomyrmex Porter & Eastmond ( 1 982)

Euryopis funebris Td Formic inae Camponotus castaneus Carico(1978) Latrodectus hesperus Td Myrmicinae Pogonomyrmex rugosus MacKay(1982)

Latrodectus pallidas Td Myrmicinae Monomorium semirufus MacKay(1982)

Steatodafulva Td Myrmicinae Pogonomyrmex badius H611dobler(1971)

Achaearanea sp. Td Formic inae Oecophylla smaragdina Cullen(1991) Saccodomusformivorous Tm Dolichoderinae Iridomyrmex McKeown(1952) Strophius nigricans Tm Formic inae Camponotus crassus OHveira&Sazima(1985)

Zodahon frenatum Zo I'ormicinae Cataglyphis bicolor Harkness(1976)

TABLE 6: Spiders that specialise on ant prey but are not ant-mimics.Family abbreviations given with Table 1, parasitoid). Despite widespread reports and other spiders Tibellus (Philodromidae) and descriptions of ant-mimicry by spiders, few Phidippus (Salticidae). These spiders do not feed studies have addressed this question quantitative- on the ant M. americana, but more importantly ly. The visual nature of ant-mimicry suggests that they were less likely to attempt to capture the the spiders are gaining protection from visual mimic S. occidentalis than immature Phidippus enemies, including birds (e.g. Belt, 1874; Engel- (that are not ant mimics). hardt, 1970), wasps (e.g. Edmunds, 1993) and Spiders are also prey to a variety of other inver- other spiders (e.g. Cutler, 1991). It is unlikely that tebrates, especially pompilid and sphecid wasps the visual resemblance to ants provides the spider (e.g. Coville, 1987), and acrocerid dipterans (e.g. mimics with protection from either their ant Schlinger, 1987). These parasitoids are primarily models or other species of ants, because ants visual hunters and many myrmecophilous perceive the environment primarily by chemical, arthropods gain protection against these enemies rather than visual cues (see Holldobler and Wil- by associating with ants (e.g. Holldobler and Wil- son, 1990). Furthermore, many spiders that are son, 1990). Thus, ant-mimicry may reduce the risk of predation by sphecid and pompilid wasps. either specialist predators of ants (see Table 6) or Edmunds (1993) provides qualitative data sug- live in close proximity with ants (Table 7) are not gesting that ant-mimics Myrmarachne are less necessarily visual mimics. likely to be taken by the predatory wasp Pison Most of the diet of many spiders are other xanthopus than might be expected if this wasp spiders (e.g. Bristowe, 1941, 1958; Reichert and was indiscriminate in its choice of prey. Finally, Luczak, 1982; Nentwig, 1987). In contrast, ants it is interesting to note that no species of lycosid are not a common prey item for most spiders, have been reported as ant-mimics (see Table 5), although a few spiders are specialist predators of perhaps because these spiders are generally noc- ants (see Table 6). Thus, ant mimicry may pro- turnal foragers and are also seldom victim to vide some degree of protection from other sphecid wasps (see Coville, 1987). spiders. Experimental evidence of this possibility is provided by Cutler (1991), who examined whether ant-mimicry in the salticid Synageles Behavioural Mimicry: Courtship Vibrations occidentalism a mimic of the ant Mynnica Some spiders are renowned for preying ex- americana, reduces the risk of predation by two clusively on other spiders. Notable among these INTERSPECIFIC ASSOCIATIONS

Spider Am Species Family Subfamily Species Source

Trtrilus arientinus Ag Formicinae Formica Brisiowc

Eiltea puna 'Jn Forraicinac Camponolus incj Ncwnan (1982,1

Acanauchenius srurriti s Lin Formicinae Tetramorium caeipilum Brisxowc(1958)

; i ,nbolusformicarius Lin Formicinae Formicu obicuripes DomUlei Redder (1972)

Evanxia merens Lin Ftirmicinac } armu o fuu;.u Bristowe(!95S)

Masontur. Lin Formicinae f-'og onontyrmex Porter [1985)

Thyrtosihenius bic\alus l.in Formicinae Formica rufa Brisiowc(1958)

f'hruruitffut* I .:-, Formicinae CrenuHDfiaxter F'oncrtl9B5}

Mxandra Pro not specified Main (1976)

5a DoUchodennae Tapinonta melanncephalum SheparU & Gibson (1972)

TABLE 7: Spiders that have been found within or adjacent to the nests of anLs. Family abbreviations riven with

Table 1. are the inimelid or pirate spiders that invade the specialised form of predation by P. fimbriata may webs and attack the owners of other species of be responsible for the improved ability of Eu spiders (e.g. Bristowe, 1941). Many of these tu.% to recognise and defend itself from /' spiders are aggressive mimics. For example, the fimbriata, compared with other salticids. For ex- mimetid pirate spiders Mimetus and Era wait at ample, Euryattus recognises P. fimbriata the periphery of the web of the social spider potential predator, unlike another prey species Anelosimus studiosus (Brach. 1977). The Jacksonoides queenslandica. Interestingly, this mimetids then pluck on the web thereby attract- recognition ability is not present in another ing a host spider that is then captured and eaten population of Euryattus in which P. fimbriata arc The salbcid Portia is also well known for its absent. Experimental trials reveal that P. k ability to mimic the struggles of prey ensnared in fimbriata attacks and captures these n.< the web of other spiders. The investigating host spiders more frequently than spiders from die is then captured bv Portia { Jackson and Hallas, population that is exposed to P. fimbriata (Jack 1990). son and Wilcox, 1993) It is still not clear whethci Some species of Portia also mimic the male the two populations of Euryattus are conspecifics courtship behaviour of their prey species; a be- or represent two different species. The more dis- haviour that increases their chances of capturing tantly related the two populations, the less likely that the differences in behaviour are the result of the unsuspecting female t' Jackson and Hallas, 1986). If prey populations suffer high frequencies the presence or absence of P. fimbriata Never- of this form of mimicry, then Portia may act as a theless, u appears to be a fascinating example of crioa pressure favouring improved dis- how the foraging behaviour of a predator has criminatory abilities in the prey, thereby estab- apparently acted as a selection pressure influenc- lishing an an evolulionarily dynamic l arms race' ing the defensive behaviour of its prey. {Mensu Dawkins and Krebs, 1979), Evidence of this form of frequency dependent selection is Chemical Mimicry; Moths and Ants provided by Jackson and Wilcox (1990, 1993 k in Spiders produce a variety of chemicals thai ll>eir study of the predatory-prey relationship be- function to attract conspecifics. Female spiders tween two Australian salticids, Portia fimbriate from many different families product and Euryattus sp. pheromones that attract members of the opposite Euryattus females live in a nest comprising a sex (e.g. Lopez, 1987: Pollard etai, 1987). and rollcd-up leaf, suspended from rock ledges and Evans and Main (1993) show experimentally that iree trunks by silk guylines. Portia fimbriata is a pheromones may be important for maintaining versatile predator of many salticids and in a social cohesion in social spiders. Several taxa of

Queensland population, it preys on female spiders arc capable of inicr-specific chemical Eury&tlus sp. using vibratory displays that ap- communication, of which the most familiar is the parently mimic the courtship behaviour of remarkable form of chemicai mimicry by botes Eunanus males This behaviour luTes Euryattus spiders (see Stowe, 1986, 1988 for extensive females from their nest, and they are sub- reviews). Boras spiders, composing several sequently attacked by P fimbriata This genera within the Aranejdae, do no* construct 424 MEMOIRS OF mB QUEENSLAND Ml SEUM

orb- webs Bad swing at their prey a betas or simply take advantage of a safe teluge (a droplet of adhesive) attached to the end of a Whatever the reason, it is unlikely that they could silk thread. Betas spiders arc aggressive mimics remain in or nearant nests without some chemical and prey exclusively on male moths; the spiders protection, because anls rarely tolerate foreign produce a cbemicaUubstance that mimics the sex nest intruders Porter (1985) provided qualitative pheromone of its moth prey species (see Eber- evidence for the presence of ant recognition hard, 1977. 1980; Yeaigan, 19SS; Stowe, 1986, pberomones by introducing myrmccophltous 1988; Stowe era/.* 1987). The exact source of the spiders Masoncus into the nests of difl prey attractant compounds is not known, bui »s rogan&myrmtx antv MoXtitocasytttt, not attack- likely ED be emitted from the spider (Stowe ef til.. ed if they were re-introduced into their original 1987). The evolution of this specialised foraging nests, but the spiders were atUtcked and killed technique is particularly intriguing because it in- within minutes if they were placed in the nesc of volves two phages; the first comprises ihe produe- foreign Pugonv>n\ rnitx or other species of ants. tion of motli-atlracting chemicals (see also It iv not known whether these spiders actively Horton, 1979), and the second is the adoption of produce the appropriate pheromones. or whether a specialised use of silk together with the loss of they simply adopt it from the substrate ofthe tie St. the orb-web. Interestingly, anecdotal observa- The predatory behaviour of two Australian tions suggest that the spider swings the bolas in spiders may also involve chemical mim.uiy llu response to vibratory signals generated by the Australian basket-web spider SoccpdaflUAS for- flying moths (Main, 1976). mtvunms iThomisitlae) builds a basket-like web that appears to attract wandering The male location mechanism of at least seven tridomyrmci families of moths are exploited b> bolas spiders, ants that may venture into the basket web (Mc- Keown, taps the ant hut the range of moth prey species captured by 1952). Tne spider also with each species of bolas spider varies (Stowe, 1986, its legs, that may further mimic ant communication, and eventually captures the unsuspecting 1988; Stowe et al., 1987), Some spiders capture It if S. webs only one species of moth, while Mcistophortt cor- ant remains to be seen formivorous capture anls, and whether the nigera is capable of capturing at least nineteen do!) fridOmyerrtCM moth species (Stowe et ai, 1987). There are no ants are actively attracted to the basket-web. The relationship between obvious taxonomic affinities between the dif- extraordinary predatory an undescribed thcmliiti its weavei ant ferent groups of bol as spider and their moth prey and species (Stowe, 19S6, 1988). The variation in OecophyUa sniaragtlina prey (see Cooper et al. also represent an example of the use bolas spider prey-specificity is likely to be related 1990) may mimicry. This thcridiid constructs a to the bio-geographic distribution of potential of chemical moth prey, the chemical compounds produced by web made of several strands of silk suspended the spiders and the chemicals u&d as moth sex- between vegetation and additional strands that attractants. Furthermore, some compounds that are anchored to the substrate below. The anchor silk that is attract certain species of moth may inhibit attrac- part is a small white bead of very attractive to the ants. If the web is complete and tion df other moths (Slowe et oi r 19 an ant bites the silk it is catapulted into the web Since araneid spiders an capable of chemical above, where it is captured by the spider. The mimicry of moths, it is not unreasonable to expect bead of silk is often placed near ant lligtxu that ant-mimicking spiders may be capable of and can sometimes attract the attention of many producing chemical compounds that 'appease' individual O. smaragdtna that all attempt to bite ants. Many species of invertebrate myr- the silk. mecophiles produce chemicals that mimic ant communication chemicals (set- Hnlhlobler and M'JTUALlSTie ASSOCIATIONS Wilson, 1990) The production of these chemicals can reduce the risk of the am:-, attacking the myrmecophiles. One group of spiders that arc There arc few examples of mutualistic ass- likely to be capable of chemical mimicry arc tions between species of spiders or even bet those that live in ant nests {sec Tabic 7). Little is spiders and other organisms. This is surjxi known about these spiders, but some earlier given the widespread occurrence of mutualistic reports may have mistakenly recorded Ihwi ins m Other ta>a (e g Boucher et df\, living in ant nests, rather than adjacent to the nest - Smith and Douglas, 1987, Holldoh!er and

(see Bristowc, 1941 >. ll to not clear whether these Wilson, 1990). but may reflect the predatory na- spiders prey on the ant larvae within the ant nest. ture of spiders Tietjen el of. ( 19X7) describe an INTF.R -S PPCI PIC ASSOCIATIONS

interesting example of a mutualistic association was significantly reduced following experimen- involving the social spiders Mallos gregalis tal removal of both Afissifrons and Philoponella These spiders do not remove the remains of prey from the barrier-web. The lower growth rate from their nest, and this debris becomes a nutrient during the ten day experimental period may rep- base foT various yeasts. The odour of these yeasts resent a potential reproductive loss of around 30 is apparently attractive to various flies, that settle eggs (estimated from the weights of egg masses). on the prey carcasses and are then captured by the P. arttentatus probably benefits by increased cap- spiders. The association is likely to be mutualistic lure rales as a result of tlie increased area of tangle because the spiders provide food for the yeast and web generated by the webs of Philoponella, in & the yeast's presence attracts food for the spider. wav analogous to the webs of some social spiders

The relationship between spiders that live in (see Struhsaker, 1969; Uetz, 1988 j The addition itnts* nests and their am hosts may also be al wiebs may increase the probability of arresting, mutualistic for some species. For example, insecti. that then drop into the sheet web, without Shepard and Gibson (1972) found myr- being caught in the orb-web of Philoponella, It mecophilous >alticid spiders of the genus seems unlikely that P. argentatus benefits from Corimtta in 61 % of 50 nests of the dolichodcrine the presence of A.fissifrons. In fact, A.fissifrons ant Tapinorna melanoe^phalum. Interestingly, is more likely to have a negative effect on the nasi ant nests with Cotinusa had more brood per nest, because it feeds on prey items caught in the more workers per nest and more brood per worker barrier web and also may prey on Philoponella' than those nests without Cotinusa, Unfortunately, on two occasions, A fissifivns were seen feeding these difference wzre not examined statistically, on Philoponella. consistent with other repoi and the greater numbers of ants and brood in the the foraging behaviour of this species (sec Table nests with Cotinusa may be due to the larger size IX of the former nests. Nevertheless. Shepard and Gibson (1972) suggest that the spider uses the ant SOME CONCLUSIONS AND PROSPECTS nest as a foundation for die construction nf its vvch, and in return provides the ants with The relationship between klcpioparasitic Ar- protection from predators or parasites. ies and their hosts has been extensively examined* yet the effects of the association on the PSEPWUS ANO FtUWMKkUA fitness components of t'Whti Argyrodes or its host Many species of orb-weaving spiders in the are presently unqualified. Consequently, it may genus P^//op^/2c'//.iflI!orK>ridae)buiId their webs be inappropriate to call rhese species Hep- within the barrier webs ofoilu-rarancid, thcridiid loparasites be i) they may not take prey agelenid andpsechrid spiders (Struhsakcr, 1969; thatthehi .1 otherwise feed on and (b) their Lubin, 1986). These associations sere thought !o hosts may not suffer a fitness cost. Of course, be commensal; Philoponella has a place to build many other well documented host-parasite sys- a web, but it was assumed [hat their presence has tems similarly fail to quantify the fitness effects little effect on the host spider (e.g. Lubin, 1986). of the presumed parasite (sec Toft et a!., 1991). In Madang Province, Papua New Guinea a Nevertheless, circumstantial evidence that die species of Philoponella builds webs between the presence of Argyrodes has influenced the bin threads of the tangle web of a large psechrid of at least a few host species suggests that klcn Psechrusargentatus. Not all P lechnts webs have Iti&arasitiSftt is an evolutionary dynamic Philoponella, but as many as 15 males and relationship. Comparative analyses reveal h I. males can be found on a single host web. Like esting differences in the biology of klcp- many small uloborids, Philoponella is a com- loparasites that do, or do nnu also prey on their munal spider, with several orh-v. chs sharing sup- host, HoM e\ ei\ there are no obvious exphi rial port threads A ihcnduiiArgy rottenjlssifrons also for the evolution of this behav n patrols the barnet weh hii! is itevft ("mind on il«c There are interesting pamlLI between chemi- sheet web of the host spider. The number of both cal mimicry by the bolas spider and vibratory A ftsstfrons and Philoponella on a Single hoM mimicry by the sal tic id Portia; bath are examples web is positively correlated with the size of the of aggressive mimicry in which the mimic ex- host. ploits the mate-auraciiin* mechanism of the The relationship between P. argenuvas and model. They also illustrate the broad spectrum ol Philoponella appears to be mutual istic il sensoTy mechanisms that arc exploited and the unpublished). The growth rate of P a- range of phytogcnetic similarity between model 42n MKMOIRS OF THE QUEENSLAND MUSEUM

and mimic. The models are clearly disadvantaged irrespective of the true nature ot these relation- by the mimics, and selection is likely to favour ships, inter-specific associations involving mechanisms that allow the victims to distinguish spiders provide a rich seam of biologicaJ systems between their conspecifk mates and the spider that pose a variety of fascinating questions. predators. There is some evidence of this selection for Euryattus, the model of Portia, but there ACKNOWLEDGEMENTS are no data on the impact of bolas spiders on their model moth populations (but see Yeargan. 1 988), 1 am very grateful for the critical comments and nor is it known whether the ability of mate moths discussion provided by Rachel Allan, Jonathan to discriminate between conspecific female Coddington, William Eberhard, Malcolm Ed- pheromones and bolas spider mimics has munds, Theo Evans, Mary Whitehouse and espe- changed. One difference between these two cially Robert Jackson and Fritz Vollrath; not all mimicry systems is that the victims of Port'm arc agree with me. but their advice is appreciate d. female, but the victims of bolas spiders are male. Thanks also to Rachacl Bathgate for helping This difference may have implications for the compile the kleptoparasite data-base; Fritz relative strength of selection in these types Peru the fitness cost to the ants, as a result ofdefensive Bulletin de Institute Roya) Scicntditjuc dc Nation- mimicry by these spiders, has nut teen quantified. ale Belgique 60: 5-1S. Nevertheless, the degree of visual mimicry in BAPT1SA, R. L9SS. On teymetwpsis (Mysrnmid strung many spiders suggests that there has been fcteptopnrasite of Pnolcidae. American Arachnol- selection for this form of protection against ogy 38: 4. predators. The inter-specific variation in the de- BELT. T. 1^74. The Naturalist in Nicaragua'. (John gree of resemblance between spider mimics and Murray: Lc.^don). their ant models suggests an evolutionary process BERRY, J.W. 1987. Notes on the life history and be- reflecting differences in the discriminatory icrof the communal spider Cyrtophora mntur- censis (Doleschall) (Araneae. Araneidac) in Yap, abilities of the receivers. These differences may

Caroline Islands. Journal of Araehnoiogv 1 5; 309- also reflect the frequency with which the spiders 319 and ants co-occur, and the kind of substrate on BOUCHER, D.H, JAMES, S. and KEELER, KM, found. Finally, ant which both axe mimicry by 1982. The ecology of mutual ism. Annual Review spiders that also prey on their models begs the of Ecology and Systematic* 13: 315-347. question of whether special isation on ant prey BRACH, V. 1977. Anelosimus studiosus (Araneae: followed ant mimicry, or vice-versa. Theridiidae) and the evolution of quasisocialtty in iberidiid spiders. Evolution 31: 154-161 Mutualisms involving spiders have received BRISTOWE, W.S. 1941. 'The Comity of Spiders'. little attention, compared with other inter- Volume 2. (Ray Society: London). specific associations. There are several explana- I95S. 'The Wartdof Spiders*. (Collins: London.) tions: the Araneae may be characterised by an BRODIE, E D , in & BRODIF, ED, 1991, Evolution- absence of mutualisms; these mutualisms simply al) respotwe of predators to dangerous prey have not been detected; or non-mutualistic as- reduction of toxicity of newts atvi resistance of iliofts may even have been incorrectly in- garter snakes in island populations. Evolution 45: ferred. For example, the impetus of my study of 231 224 in the Arachnida. Pseckruh, Philoponella and Argyrode* was to BISK1RK.R.6. 19&I, Sociality pp. 281-367. In: H.R. Hermann (ed.). 'Social Insects'. reveal the fitness costs to the host of what ap- Volume II. (Academic Press: New York), peared to be a kleptoparasitic relationship. The CANGIALOSL K.R. 1990a. Lite cycle and behaviour correct nature of the relationship between the of the kleptoparasitic spider. ArgyrOdfiA uluians species was only revealed experimentally, and (Araneae, Theridiidae). Journal of Arachnology this is likely to be true of many other inter- 18: 347-358. specific associations described in this review. Bur 1990b. Social spider defense against klep- INTER-SPECIFIC ASSOCIATIONS 427

toparasites. Behavioral Ecology and Soctobiol- spider Massophara dizzvdeani sp. n. (Araneae). ogy 27: 49-54. PfeydieOT: 145-169. 1991. Attack strategics ol a spider xlepioparasiit: 1986. Web Kidding behavior of anapi&sympbylog- effects of prey availability and host colony size. nathid and mysmenid spiders (Araneae). Journal Animal Behaviour 4 J; 639-647. of Arachnology 14: 1*39-356. CARICO, J.E 1978. Predator behaviour in Ennvf»i EDMUNDS.M 1978.0ntheassociatxonbetweefiAfvr- funebris i.Hentz) (Araneae: Tneridiidae) and the moradme spp. (Salticidae) and ants. Bulletin of evolubotuiT, significance of web reduction Sym- lhe British Arachnoiogical Society 4: 149- 160. posium of the Zoological Society of London 42 1993. Does tmtnicry ol ants reduce predation by 51-58, wasps on salticid spiders? Memoirs of the CODDINGTON, J. A. & LEVI, H.W 1991, Sys- Queensland Museum 53: 507-512. tematica and the evolution of spiders (Araneae EDWARDS, G.B., CARROLL, J.F.. WHTTCOMB, Annua) Review of Ecology and Systematics 22: W.H. 1975. Strides auraia (Araneae: Salticidae). S92. a spsder predator of ants. Florida Entomologist 57

COOPER, D , WILLIAMSON, A. & WILLIAMSON. 337-346. C. 1990. Deadly deception. Geo 12(1): 87-95. ELGAR, MA. 1989. Kkptoparasitism: a cost ol ag- COVILLE R.E. 1987. Spider-hunting sphecid wasps. gregatine for an orb-weaving spider. Animal Be- Pp. 309-318. In: W. Nentwig fed*). 'Ecoph haviour 37: 1052-1054. ogy of Spiders*. (Springer- Verlag: Heidelberg*. 1991. Sexual cannibalism, sue dimorphism and COYLE, F.A. & MEIGS. T.E. 1989. Two new species courtship behavior in orb-weaving spiders of klcptoparasiie Mysrmmopsis (Araneae, Mys- (Araneae). Evolution 4S- 444-448.

menidae) from Jamaica. Journal of Arachnology 1 992. Sexual cannibalism in spiders and other inver- 17:59-70. tebrates. Pp. 129-156. in: MA. Elgar and BJ. COYLE, FA., O'SHIELDS, TC. & PERLM LITTER, Crespi teds). "Cannibalism: ecolugy and evolu- D.G. 1991. Observations on the behaviour of the tion among diverse taxa*. (Oxford University kleptoparasiuc spider Mysmenopsis furtira. Jour- Press: Oxford). nal of Arachnology 19: 62-66. ELGAR, MA. & GODFRAV. H.C.J. 1987. Sociality CULLEN, E. 1991. Observations on an ant -eating and sex ratios in spiders. Trends in Ecology and theridiid spider Australasian Arachnology 42; 3- Evolution 2: 6-7. 6. ELGAR. M.A. r GHAFFER. N. & READ, A. 1990. CURTIS, B.A. 1988. Do ant-mimicking Cosmophasis Sexual dimorphism in leg length among orb- prev on (heir Camponotut models? Cimbebasi3 weaving spiders: a possible role far sexual can-

i a 67-70. nibalism loutnal of Zoology, London 222: CUTLER. B. 1980. Ant prcdation by Habnnestim;

/w/rv(Hent/.) (Araneae: Salticidae). Zoologischer ELGAR. M.A.. POPE, B. & WILLIAMSON, I 1983. Anzcigcr. Jena 204; 97-101. Observations on the spatial distribution and 1991- Reduced predation on the antlike jumping natural history of Cyrfophora hirta (L. Koch) spider Sputgeles occidentalis (Araneae: Sal- (Araneae: Arancidac). Bulletin of the British

ticidae i. Journal of Insect Behavior 4: 401-407. Arachnological Society 6: 83-87.

DAVrES.N.B,BOURKE,A.F.G.& BROOKE. M.DE ENDLER. J. A. 1981 . An overview of die relationships L. 1989. Cuckoos and parasitic ants: interspecific between mimicry and crypsis. Biological Journal brood parasitism as an evolutionary arms race. oftheLinnean Society 16:25-31. Trends in Ecology and Evolution 4. 274-278. I9S6, 'Natural Selection in the Wild'. (Princeton DAWKINS, R. KJ*EBS JR. 1979. Arms races be- University Press: Princeton). & t

tween and within species. Proceedings of thr 1991 . Interactions between predators and prey. Pp. Roy at Society of London B 205: 489-511. 169-196 In: J R, Krvns and N.B. Davies (cifc). OONDALE. CD. & REDNER, J.H. 1972. A synonym 'Behavioural Ecology: an Evolutionary

proposed in Perimortes, a synonym rejected in Approach', f Blackwclls: Oxford). Walckenaera, and a new species described in ENGELHARDT. W. 1970 Gcstalt und Lcbcnswivc Coehlembvlus {Araneida; Erigonidae). Canadian der AmeiMenspunne SynagcfeS venator (Ideas'. Entomologist 104: 1643-1647. Zuglcich ein Beit rag am Ameiscnmimik- HBERHARD, W.G. 1977. Aggressive chemical ryforschung. Zoologischer Anzeiger 185: 317-

L bolas spider. I 'X: 1171- 334. mimicry* bv a Science 1)75. EVANS, T.A.& MAIN, B.Y. 1993. Attraction be twecn 1979a. Aryvrodw uitvrmatus (Theridiidae). a vvcb social crab spiders: silk pheromones in

that is nr.l ;\ snare. PsydtC B6: 407-413. sociaits. Behavioral Ecology 4: 99-105. 1979b. Predatory behaviour of an assassin spider, EXLINE, H. &LEVI, H.W 1962. American spiders of

Chorizopes sp. (Arancidac), and the defensive the genus Argyrodex (Araneae, Thcridiidae \. Bui bcbavioQi of its prey. Journal of the Boinba> lean of tiic Museum ofComparative ZooJog Natural History Society 79: 522*524. 75 204. 1980. The natural history and behavior of Ihi GILLESPIE, R.G &CARACO.T 1987 Risk set. 1

428 MEMOIRS OF THE QUEENSLAND MUSEUM

foraging Strategies of two spider populations. Lanka: predatory behaviour and cohabii Ecology 68. 887- 899. with social spiders. Bulletin of the British Arach GRAY. M.R. & ANDERSON, G.J. 1989. A new nological Society 7; 133-136 Australian Species of Argyrode:; Si in on JACKSON, R.R. &. BLEST, A.D. 1982. The biology of

(Araneoidea; Theridiidae j which preys on its host. Portiafimbriate!* a web-building jumping spider

Proceedings of the Ltnncan Society oi New South t Araneae SaJticidaet from Queensland: ariliui Wales 111: 25 30 lion of webs and predatory versatility. Journal of GRISWOLD, CO. 1985. tsela okuncana. a new germs Zoology. London 196: 255-293. and species of kleptoparasirie spider From JACKSON, R.R. & BRASSINGTON. RJ. 1987. The southern Africa (Araneae. Mysmcniaach Annals biology oiIPholcus pfmiartjiioidex (Araneae, Phol- of the Nauil Museum 27: 207 217 ctdae); predatory versatility, araneophagy and ag- GRISWOLD, C.E. 6c ME1KLE-GRIS WOLD/I". 1987. gressive mimicry. Journal of Zoology. Ixwdnn Archaeodh-tyna uUnu, new species (Araneae- 211:227-238 Dictynidae). a remarkable le of JACKSON. R.R. & HALLAS. SEA. 1986, Compara-

; gjOup-lb d spiders ipfaM spp . tive biology of Portia, afticanus, ft ciHnntamt. P. Ercsidae*. Novitales Araneae: American Museum ftmhriata P. labia ta, and P. si frilffZJ, 28^7; 1-il araneophagic, web-building jumping Spi HARKNKSS, R.D. 1476. Further observations on ih<- (Araneae- SaKieidfte): utilisation oi n relation between an ant Catuxlyphis btcolor |F.) picdatory Versatility, and mtraspecitV im-

(Hyro., Formicidae), and a spider Zodarium lions. New Zealand Journal ofZoologJ i frenatum (Simon) (Araneae Zodariidae), En 499

; lomologists Magazine : '21. Monthly 1990 Evolulionary Origins of displays used tn 1942. HARVEY, PH. & PAGEU MD 'TheCompara gresstvc mimicry to Ponfa a web-invadinp nvc Method in Evolutionary Bkriog) (Oxford araueopltagk jumping spidct (Ararn University Press; Oxford). ticidacl. New Zealand Journal OFZOofogy 17: HINGSTON. R.W.G. 1927. Field observations on spider mimics Proceeding? of the Zoological JACKSON, R.R & VAN OLPHEN. A, 1991, Prey Society. London 1927 841-838 capture techniques and prey p&fefl

B. 1 97 1 Sreatodafiitvct ("Jheridiidaej. HOLLDOBLER, urn orbiaiL'ta, «f)t- 77" I a spider that feeds wi handler ants Psyche eating jumping spiders (Araneae. Saltieidaet 202-208 Journal of Zoology. London 223: 577-591.

HOLLDOBLER, B ft WILSON, E.O. 199J. The .-.'.,." I9*vl i ton 1 1 and prey pre i r i Ants'. fHarva/d Unfrertft) Press: Cambridge). of Chrysilla. S'aHa and Siler, ant-eating jumping; 732pp spiders (Annieae, Satuadae i from Kcrv HORTON, C.C. 1979 Apparent attraction of moths by Lanka. Journal of Zoology, London 227; 163- webs of araneid spiders. Journal of Arachnoloxy 7:8 JACKSON, R.R. & WILCOX. R.S. 1990 Aggressive HLiMPHREYS.WF.1988.Thcccoiogyofspidersw.iJ) m behaviour I iii pj eific predatory ajtd &{0 Australia. Pp. 1-22. Ip predator recognition in the predator-prey interac- Austin and N.W. Heather Australian I ii ins of Portiafimbriate and Euryattus sp.. jump- Araehnology' .{ Australian Entomological Society ildets from Queensland Behavioral Ecology Miscellaneous Publication No. 5: Brisbane) and Sociobinlogy 26: 11! 119 JACKSON, J.F

spiders from Queensland. Mem i J the Of transfonimi-on.il mimicrv, American Midland Queensland Museum 33 557-560. Naturalist 91 24S 251. Th< JACKSON. R.R 1986 bi I JACKSON, R.R. 1985. The biology of Simuertu* MRMAN.E.A.R &

ogy of Tairria erebus \ A raneac. Gnaphosidae }, an paemla and S. thontcica. web-building jumping araneophagic spider from New Zealand: silk spiders (Araneae Sallicidae) from Queensland: utilisation and predatory versatility, New Zealand ..< '-habitation with social spiders. ulilizaii"n "I Journal of Zoology 13: 521-54 silk, predatory behaviour and intra-specifie inter- Experimental actions Journal of Zoology, London (B) I; 175- LARCHER. S.F. & WISE. D.H. 1985. 210, studies of the interaction between a web-invading spider and two host species. Journal of Arachnol- 1986. The biology of ant-like jumping spiders (Araneae, Salticidoe): prey and predatory be- ogy 13 43 59. haviour of Myrmoruthne with particular atten- LOPEZ, A. 1987. Glandular aspects of sexual biology tion to M lupara from Queensland. Zoological Pp Htl-132. 1". W Nem*ig(cd.>. 'E< Journal of the Linnean Society 88; 179-190. ogy of Spiders'. (Springcr-Vcrb*:* Heidelberg). 1987. The biology of Olios spp., huntsman spiders LUB1N.Y.D. 1974. Adaptive advantages and the evolu-

(Araneae, Sparassidae ) from Queensland and Sri tion of colony formation in Cyrtophora (Araneae: .

INTER-SPECIFIC ASSOCIATIONS 429

Araneidae). Zoological Journal of Ihc Ltawi predator of coliembola in harvester-ant colonics. Society 54: 321-339 Psyche 92: 145-151. 1986. Web building and prcv capture in the PORTER, S.D. & EASTMOND, D A |982.JBirj Uloboridae. Pp. I32-17J. In: W. Shear led.i cnii iTheridiidae). a spider that pre 'Spiders: Webs, Behavior and Evolution, (Stan- Pok'twonrsmu'' urils Journal of Arachnolovv 10: ford University Press: Stanford.) 275-277. MACKAY, W.P. 1982. The effect of predalion of REICHERT, S. E. 1932. Spider interaction strategic Western widow spiders fArancae: Thendiidae) on communication vs. coercion. Pp. 281-315. In

harvester ants (Hymenoptera: Formicidac >. P.N. Witt and J.S. Rovner (eds). Spider Com-

Oecolopia (Berlin) 53: 406-4J I munication: Mechanisms and Ecological MAIN. B.Y 1976. 'Australian Spiders'. (Collins: Syd- Significance' (Princeton University Press: Prtn- ney). tott> MATHEW. A.P. 1954. Observations on the habus of REICHERT, S.E. 8c GILLESPrE, R.G. 1986. Habitat choice .-.nil utilization two spider mimics of the red ant, Oeeopftylfa in wen-building spiders. Pp smaragdina (Fabr.j. Journal of the Bombay 23-4R. In: W. Shear led.). Spiders: Wtbg, Be Natural History Society 52: 249-263. haviorand Evolution (Stanford University I' Stanford). MCTVER, J. 19R7. On the myrmccomorph Coquitlvtui insignis Uhler (Hemiptera: Miridae): arthropod REICHERT. S.E. & LUCZAK. J. 1982. Spider fbr*g ing: behavioral responses to prey. Pp. 353-385- In: predators as operators in an anl-mimctic system. Zoological Journal of the Linnean Society 90: P.N. Witl and J.S. Rovner (eds), 'Spider ( 133-144, municatinn: Mechjr-, , ..mi BcolO| Significance'. (Princeton University Press: Prin- MCKEOVvN, KC. 1952. •Australian Spiders, their

cetoo I, Jives and habits' (Angus and Robertson: Sydney i. REISKIND. J. 1970. Multiple mimetic forms M1LINSKI. M. & PARKER. G.A. 1991. Competition ant-mimicking cluhtonid spider. Science 169: for resoutccs. Pp. 137-168. In: Krebs, J.R. and 587-588. J,B (eds). 'Behavioural Bcolog) 1977. Ant-mimicry tn Panamanian clubionid and evolutionary approach'. (Black wells: Oxfoi sallicid spiders fAraneae: Chibionidac. Sal- MYERS, J.G & SALT, G. 1926, The phenomenon of ticidae). Riotropiea 9* l-B, myrmccoidy. with new examples from Cuba. REISKIND, J. & LEVI. H.W. 1967. Anotm, an am- Transactions of the Entomological Socio mJmicking theridiid spider from New Calcdonu London 1926:427-435 .neae: Thendiid^eV Psvche 74: 20-23. NENTW1G, W. 1 987. The prey ofspiders. Pp. 249-263. ROBINSON. M.H & ROrMNSON, B 1971 The In- W NeRtWtgfed -ivxiology of Spiders'. predatory befiavior Of the ogre-faced spider (Springer-Verlag: Heidelberg). Dinapis tongipes F Cambridge (Aranear 19s NENTWIG, W. & HE1MER, S. Dinopidae) American Midland Natural aspects of spider webs. Pp. 211-225. In: W. 96. Nentwig (eo\). 'Eeophysiology of Spiders'. 1973 Ecology and behavior of the giant wood (Springer-Verlag: Heidelbei under Nephila maculata (Fabncius) in New NOONAN G.R. 1981 N61BS or. interactions between i Ilea. Smithsonian Contributions to Zoo low the spider ttlicu pu no (Gnaphosidac) and the unl 149: [.76, Cawp0t\0tti$ tnco in the Peruvian At- 1982 Courtship and mating behavior in spiders Bionopica 14: 145-148. Annual Review of Entomology 27: 1-20. OLIVEIRA, PS. 1988. Ant-mimicry in some Brazilian RYPSTRA, A.L. 1981. Trie effects of'kleptoparasWism salticid spiders S.4- and clubionid (Aranr>.- on prey consumption and web relocation in a ticidae, Chibionidac), Biological Journal of Ihe Peruvian population of the spider Nephila Linnean Society 33: 1-15. 0ikos37: 17^-182. OLIVEIRA. P.S & SAZIMA. 1. 1984- The adaptive SCHLfNGER, EJ. 1987; The biology of Acrocendat bases of ant-mimicry in a Neotropical aphan- (Dipteral: true endoparasitoids ol spinier*. Pp. toehilid spider (Araneac: Aphantochilidje r 3 19-327. Tn: W. Nentwig (cd). "Ecophysiologyof Biological Journal of the Linnean Society 22: Spiders', (Springcr-Vcrlag: Heidelberg)* 145-155. SHEPARD, M. & GIBSON. F. 1972. Spider-ant sym

PASTEUR. G 1982. A classifies review of blosts: Cntimf.i' spp, (Araneidae; Sail- rd mimicpy systems. Annual Review i>t Ecology £nd Tapinopta melanocephalmn (Hymenoptera: For- Systematic:-; 13: 169-199. imcidae) Canadian Entomologist 104: 1951-

POLLARD. SO., MACNAB. A.M. & JACKSON, J.J. 1954 1987. Communication with chemicals: SHfNKAI. A. 1988. A note, on the web silk theft by

pheramones and spiders. Pp. 1 53- 144. hi. W Argyrodet ivlindrauts (Thorell) (Araneae:

(ed.h 1 Nentwig "Eeophysiology of Spid Theridiidae). Acta Aracbnologica 36: 15 I (Springer-Verlag: Heidelherg) SMITH. D.D. & DOUGLAS, A.E. 1987. The Biology

TORTER. S.D. 1985. Maxoncus spider: a mirn. of Symbiosis { Edward Arnold: London ). { . .

430 MEMOIRS OF THE QUEENSLAND MUSEUM

SMITH TRAIL, D. 1980. Prcdation by Argyrodes 1 98 1 Energetic considerations of a spider parasite- (Thcridiidae) on solitary and communal spiders. spider host system, Revue Arachnologique 3: Psyche 87: 349-35?. 37-44. STOWE, M. K. 1986. Prey specialization in the 1982. Colony foundation in a social spider. Araneidae. Pp. 101-131. In, W. Shear (ed.). Zeitschrift fur Tierpsychologic 60: 313-324. 'Spiders; Evolution*. (Stan- Webs, Behavior and 1984. Kleptobiotic interactions in invertebrates. Pp. ford University Press: Stanford). 61-94 In: C.J. Barnard (ed.). 'Producers and 513-587. In: 1988. Chemical mimicry. Pp. K.C. Scroungers: Strategies of Exploitation and (ed.). "Chemical Mediation Spencer of Parasitism'. (Chapman and Hall: Beckenham) Coevolution*. (Academic Press. New York). 1987a. Growth, foraging and reproductive success. STOWE, M.K., TUMLINSON, J.H. & HEATH, R.R. Pp. 357-370, In: W. Nentwig (ed.). 'Eeophysiol- 1987. Chemical mimicry: bolas spiders which ogy of Spiders*. (Springer- Vcrlag: Heidelberg). attract male moth prey emit components of prey 1987b. Kleptobiosisin spiders. Pp. 274-286. In: W. sex pheromone blends. Science 236: 964-967. Nentwig (ed.). 'Ecophysiology of Spiders'. STRUHSAKER, T.T. 1969. Notes on the spiders (Springer-Vcrlag: Heidelberg), Uloboms mundwr (Chamberlin and Ivic) and Nephila davipes (Linnaeus) in Panama. American VOLLRATR F & PARKER, G.A. 1992. Sexual Midland Naturalist 82: 61-63. dimorphism and distorted se\ ratios in spiders. SUTER, R.B.. SHANE, CM. & HIRSCHEIMER, A.J. Nature 360: 156-159. 1989. Spider vs. spider: Froniirwlla pyramitela WHITEHOUSE, M.E.A. 1986. The foraging be- detects Argyroses trigonum via cuticular chemi- haviours f>M rgyrodes aniipoduma (Theridiidae), cals. Journal of Arachnology 17: 237-241, a klcptopaiasitc spider from New Zealand. New TANAKA, K. 1984. Rate of "prcdalion by a klcp- Zealand Journal of Zoology 13: 151-168. toparasitic spider. Argyrodes ftssijnms. upon a 1987. "Spider cat spider": the predatory behaviour large host spider. Ageterm limhaia. Journal of of Hhomphaea sp. from New Zealand. Journal of Arachnology 12: 363-367. Arachnology 15:355-362. W.J., TIETJEN, AYYAGAR1. L.R. & UETZ, G.W. 1988a. Factors influencing specificity and choice of 1 987. between social spiders and yeast: Symbiosis host in Argyrodes antipodidm (Araneae, their role in prey attraction. Psyche 94: 151-158. Thcridiidae). Journal o( Arachnology 16: 349- TOFT, C.A.. AESCHLIM ANN. A. & BOL1S, L. 1991 355. Parasite-Host Associations: Coexistence or Con- 1988b. To mate or fight? Male-male competition flict?* (Oxford University Press: Oxford). and alternative mating strategies in Argyrodes UETZ,, G.W 1988. Group foraging in colonial web- aniipodiana (Theridiidae. Araneae). Behavioural building spiders, evidence for risk sensitivity. Be- Process 23: 163-172. havioral Ecology and Sociobiology 22: 265-270. WING, K, 1983. Tutelinasitttiiis (Araneae: Salticidao; 1 992, Foraging strategies of spiders. Trends in Ecol- an ant mimic that feeds on ants. Journal of the ogy and Evolution 7: 155-159. Kansas Entomological Society 56: 55-58. VOLLRATH, F. 1978. A close relationship between two spiders. Curimagua bayano syneeious on a WISE, D.H. 1982. Predation by a commensal spider. Diplura species. Psyche 85: 347-353. Argyrodes tn^onum. upon its host: an experimen- 1979a. Behaviour of the kteptoparasitic spider Ar- tal study. Journal of Arachnology 10: 1 1 1-1 16. gxrodes elfx-atm Araneae, Thcridiidae). Animal YEARGAN. K.V. 1988. Ecology of a bolus spider, Behaviour 27: 515-521. Masiophora hittch'msoni: phenology, hunting tac- 1979b. Vibrations, their signal function for a spider tics, and evidence for aggressive chemical

klcptoparasile. Science 205: 1 149-1 151. mimicry. Oecologia 74: 524-530. VERTICAL DISTRIBUTION AND ABUNDANCE OF PSEUDOSCORPIONS (ARACHNIDA) IN THE SOIL OF TWO DIFFERENT NEOTROPICAL PRIMARY FORESTS DURING THE DRY AND RAINY SEASONS

JOACHIM ADIS and VOLKER MAHNERT

Adis,J.andMahnert, V. 1993 11 11: Vertical distribution and abundance of pseudoscorpions (Arachnida) in the soil of two different neotropical primary forests during the dry and rainy

seasons. Memoirs ofthe Queensland Museum 33(2): 43 1 -440. Brisbane. ISSN 0079-8835.

Pseudoscorpions were extracted from 0- 1 4cm soil depth in two dryland (upland) forests near Manaus, Brazil. In a primary forest on yellow latosol, about 1700 specimens per m were

obtained during the dry season and 1 1 35 during the rainy season. They accounted for 1 0- 1 5% of all arthropods extracted, excluding Acari and Collembola. In a primary forest on white sand soil (campinarana), about 530 specimens per m" were obtained during the dry season and 480 during the rainy season. They accounted for only 3-4% of all arthropods extracted, again excluding Acari and Collembola. Significant but different correlations were found in both forest types between the abundance of pseudoscorpions and changing moisture, temperature and pH conditions in relation to soil depth and season. Neither during the dry season nor during the rainy season was the abundance of pseudoscoq^ions in mineral subsoils higher in response to the changing soil moisture content in organic layers. This was reported for arthropods from forests in the seasonal tropics where periods without precipitation occur, Results are discussed at the species level. They are compared with published data on the vertical distribution and abundance of pseudoscorpion species from the yellow latosol of a secondary dryland forest (dry season and rainy season) from the same region. In zwei Festlandwaldern in der Umgebung von Manaus wurden Pseudoskorpionc aus 0-14cm Bodentiefe extrahiert. In einem Primarwald auf gelbem Latosolboden wurden

wahrend der Trockenzeit 1700 und wahrend der Regenzeit 1 135 Individuen pro m" nachgewiesen. Sic rcprasentierten 10-15% allerextrahierten Arthropoden, Acari und Collembola ausgenommen. In einem Primarwald auf WeiBsandbodcn (campinarana) wurden wahrend der Trockenzeit 530 und wahrend der Regenzeit 480 Individuen pro m nachgewiesen. Sie reprasentierlen nur 3-4% aller extrahierlen Arthropoden, Acari und Collembolen wiederum ausgenommen. Signifikante aber unlerschiedliche Korrelationen crgaben sich in beiden Waldtypen zwischen der Abundanz der Pseudoskorpionc und der sich andernden Feuchle. der Temperatur und dem pH in Bezug auf Bodentiefe und Jahreszeit. Weder wahrend der Trockenzeit, noch wahrend der Regenzeit, war die Abundanz der Pseudoskorpionc im mineralischen Unterboden als Folge auf die sich andernde Fcuchte im organischen Ober- boden hoher. Dies wurde fur Arthropoden in Waldern der saisonalen Tropen, die Trocken- perioden durchlaufen, nachgewiesen. Die Ergebnisse werden auf Artniveau diskutiert. Sie werden mit publizierten Daten iibcr die Vcrtikalvcrteilung und Abundanz von Pseudoskor- pionarten aus einem Sekundarwald auf gelbem Latosolboden im gleichen Gebiet verglichen. QPseudoscorpiones, abundance, seasonality, vertical distribution, Neotropics.

Joachim Adis, Tropical Ecology Working Group, Max-Planck-Institute for Limnology, Postfach 165, D-2320Ploen, Germany, in cooperation with National InstituteforAmazonian Research (INPA), Manaus, AM, Brazil; VolkerMahnert, Museum d'Histoire naturelle, Case Postale 434, CH-1211 Geneve 6, Switzerland; 23 November 1992.

In those wet but markedly seasonal tropics upland) forests experience a rainy season where periods without precipitation occur, ter- (December-May; average monthly rainfall 211- restrial arthropods are reported to migrate to 300mm) and a 'dry' (= drier) season (June mineral subsoils in the dry season as a response November; average monthly rainfall 42- to changing humidity in organic layers (Beck, 162mm). Annual precipitation is 2105mm (based 1964; Bullock, 1967; Goffmet, 1976; Lawrence, on 75 years of records from the meteorological 1953; Levings and Windsor, 1984; Liebermann station at Manaus, cf. Ribeiro and Adis, 1984). and Dock, 1982; Merino and Serafino, 1978; About 75% of the rainfall (1500mm) is recorded Petersen and Luxton, 1982; Rybalov, 1990; during the rainy season. This had no observable Strickland, 1947; Willis, 1976 and others), difference in vertical distribution of terrestrial Centrai Amazonian dryland (= non-flooded arthropods in primary forests on yellow latosol .

432 MEMOIRS OF THE QUEENSLAND MUSEUM

DRY SEASON RAINY SEASON

N/m 702.5 635.1 259.8 105.9

45 -|(%] r/oj 2 N/rr/ = 1,703 3 2: N/m -1,135.7

30 30 -

•- 15 - 15 **r r 0-3-5 3.5-7.0 70-10.5 10.5-UO 0-35 35-7.0 70-10.5 10.5-U.O Soil Depth (cm) Soil Depth (cm)

• |=nymphs I U odults — • = soil moisture content

FIG. 1 . Distribution of pseudoscorpions in soil and soil moisture content (both in %). Samples taken every 3.5cm to depth of 14cm in dry season and in rainy season in primary dryland forest on yellow latosol near Manaus,

Brazil. Total catch each season = 100%. Abundance for each soil layer (N/m ) and for total catch/ season 2 (SN/m ). and on white sand soil (Adis et al, 1989a, b, thaler (1956), Prance (1990), Pranced al. (1976) unpublished data; Morais, 1985). However, and Rodrigues (1967) provide detailed botanical results were based mainly on data for orders. descriptions of the forest. Microclimatic data are Evaluation of sampling data at the species level given by Decico et al. (1977), Marques et al. is now possible for pseudoscorpions, since iden- (1981) and Ribeiro and Nova (1979). The soils tification has been completed by the second near the Ducke Forest Reserve were described by author. Falesi and Silva (1969) as deep, yellow latosols, strongly weathered, excessively to very strongly STUDY AREAS acidic, of heavy texture in all profiles (A-C: 0-

1 10cm) and with clay content of the B horizon varying from 50-70%. In the area, where Primary Forest on Yellow Latosol study entomological long-term investigations have pre- Sampling was carried out during the rainy viously been undertaken (Adis and Schubart, (March) and dry seasons (October) of 1987 in a 1984; Morais, 1985; Penny and Arias, 1982), the dryland terra firme forest at the Ducke Forest soil carried a l-3cm thick humus layer (A in- Reserve (= Reserva Florestal A. Ducke, 2°55'S, ), 59°59'W) of the National Institute for terspersed with fine roots and a thin leaf litter, the surface. further details Amazonian Research (INPA, Manaus), situated covering most of For on the study area see Penny and Arias, 1982. on the Manaus-Itacoatiara highway (AM-0 1 0; cf Penny and Arias, 1982). The area sampled was During the 1987 rainy season, 322mm of rain classified as high terra firme forest (Takeuchi, were recorded in March (data from the 1961; cf. Brinkmann, 1971) with several species meteorological station at Ducke Forest Reserve, of large, broad-trunked trees (with or without provided by M. de N.G. Ribeiro at INPA, buttresses). The trees reached 44m in height Manaus). On the day of sampling (March 20, (average height: 22m) and formed a closed 1987) soil moisture content (= weight difference canopy. Approximately 235 species, representing between wet and dried soil samples in%) was 43 families of trees, were recorded in the study 20.0% at 0-3.5cm (= humus layer), 17.5% at area. Most frequent were Leguminosae, 3.5-7cm, 15.5% at 7-10.5cm and 17.7% at 10.5- Rosaceae, Lauraceae, Sapotaceae and 14cm soil depth, respectively (= mineral subsoil) Lecitidaceae. The forest had a patchy understorey (Fig. 1). Soil temperature at 10 a.m. decreased and a scanty herb layer. Guillaumet (1987), Lech- from 25.2°C in the top 3.5cm to 24.5°C at a soil 1

NEOTROPICAL PSEUDOSCORPIONS

DRY SEASON % N/rrT RAINY SEASON % N/rrV IOOjD tOUO 1,135.7

:-»;- —esk:: -:.- :- = M 1 nde ns,_-_-;;'--. ^6a.j:--i 1,159.7 3516. E

* - i *T1 1 7 minor 1 2 . T nra 1 1C

•••.•••J e.a 6 t. roivpl B1.B

,: 1 schusie.i A grocfl IB 33.7

4 gr S3 ccilis 3.7 62.4 i sctju^ton 19.2

P. homodentotus ] o.e |*.J I tenuis

i I 1 i I i 60 20 40 50 aot%]

2 : FIG. t 'ft. 2. Dominance land abundance (N/m ) of species of pseudosa ifpii ms extracted Irara I m of soil (0- depth) during dry season (October and rainy in 1987) Season (March 1987) primary dryland forest on y f tatosol nearManaus, Brazil Total catch of-each season == 100%. (See text for further explanation*. depth of The soil pH increased 3.3 in pling = 14cm from (March IS. 1988] soil moisture content | the top 3.5cm to 3.8 ill the lower layers (3.5.- weight difference between wet and dried sod 14cm). During the IM87 dry season. 139mm of samples in %) increased from 20.1% in the top rain were recorded in October (data from the 3.5cm to 26.2% at 10.5cm depth (= humus layer' meteorological station at Ducke Forest Reserve) and dropped to 9 9% at 10.5-I4cm (- white sand On the day of sampling (October 14, 1987) soil soil) (Fig. 4). Soil temperature at 1) a.m J moisture content increased from 1 3.6% in the top decreased from 24.7°C in the top 3.5cm to 24.5' C 3.5cm to 19.1% at a soil depth of 14cm (Fig )) at a soil depth of 14cm. The pH of the soil was and the soil pH increased from 3.3 lo 3.8. Soil 3.2 (0-3.5cm), 3.3 (3.5-7cm), 3.6 (7-!0.5cmiand temperature at J 1 varied a.m. slightly (24.6- 3.7 (10.5-14cmi, respectively. During the 1 D 24.7 C) from the surface to 14cm. dry season, 77mm of rainfall were recorded in August (data from the meteorological station of EMBRAPA/UEPAH, km 54). On the day of sam- Primary Forest on White Sand Soil pling (August 05. I9S8] soil moisture content Sampling was carried out during the rainy decreased from 21.9% in the top 3.5cm to $ season (March) and the dry season (August) of at 10.5cm depth (= humus layer] and to 9.3% al 1988 in a dryland campinarana forest of the 10.5-14em (= white sand soil, Fig 4). Soil Biological Reserve (approx. INPA/SUFRAMA temperature at 3 p.m. decreased from 27.1°C in , 2°30'S, 60°I0 W). at km 45 (formerly km 62) on the top 3.5cm to 2S.5°C at a soil depth of 14cm. the Manaus Boa Vista -174). highway (BR Cam- The pH of the soil was 3.6 (0-3.5cm), 3.5 (35- pinarana (= caatinga arborea) was classified as a 7cm). 3.5 (7-10.5cm) and 3,6 (10.5-I4cm>. elati \ ely light forest on white sand soil with respectively. The relative light intensity on the thin-stemmed trees 10-20m high, with occasional forest floor was about 2 \% (210 lx; comparative large, broad-trunked individuals, with or without value in the open air at 3 p.m., 10.000 lx (overcast buttresses, a patchy understorev and no herb layer sky);cf Brinkmann, 1970, 1971). (Guillaumet, I9S7, Anderson, 198 1; Lisb>>a. 1975). Data on geomorphologv and soil genesis METHODOLOGY are found in Chauvcl et al. ( 1987) Floral inven- tories have been given by Lisboa (1975). Braga

(1979), Anderson etal i 1975). Anderson ( i9sT) In both study areas, six soil samples were taken and Guillaumet (1987), while the microclimatic along a transect al random intervals with a split data were provided by Ribeiroand Santos ( 1975). corcr (a steel cylinder with lateral hittj In the study area, the white sand soil carried a diameter 2 I em, length 33cHl) which was driven humus layer that was 10-1 tern thick (Ao), into the soil with a mallet. Each sample was taken penetrated by a matting of roots and a sur- thin ? to a depth of 14cm and was then divided into four face-covering leaf litter. subsamples of 3.5cm each. Animals were ex- During the 1988 rainy season, 293mm of rain tracted from subsamples following a moditk- were recorded in February and 2K0mm in M- method of Kempson (Adis, 1987). All pseudn (data from the meteorological station of scorpions were separated by species [adults EMBRAPA/UKPAK, km 54). On the day of sam- (males and females), nymphai insiars| and iheil r r

434 MEMOIRS OF THE QUEENSLAND MUSEUM

DRY SEASON RAINY SEASON DRY SEASON RAINY SEASON

: 168.438.5 4JB 1107 28.9 N/m : 105.9 144 4.8 N/m

100 100 -i Id eobisium schusteri Tyronnochthoni us minor 2 2 •2. N/rrf = 192 ZN/m = 211.7 i— i ^N/m = 139.6

- 50 50

E \—\ i — i r

2 N/m': 1347 H4 77.0 4.8 N/m :231.0 5678 255j0 1055 2262 264.7 235.8129.9 100 -, 100 Brazilatemnus browni Microblothrus tridens

2 N/m =81.8 3: N/rrT = 1.159.7 ^ N/rrr= 856.6

50 -

fl BH r~n t r i r -3.5-7.0 -1Q5-KJ0 -3.5 -7.0-10.5-14.0

N/m : 48.1 14.4 4.8 24.1 4.8 100 -, Soil Depth (cm) Albiorix gracilis = nymphs = adults f | ] N/ m* = 625 2: N/m = 33.7 . —i X

50 -

i i 1 liin 1 i | r^ I r —— 3.5 -7.0-10.5 -14.0 3.5-7.0-10.5-14.0

Soi I Depth ( cm )

FIG. 3. Vertical distribution of five species of pseudoscorpions in soil (%). Samples taken every 3.5cm to depth of 14cm during dry season and rainy season in primary dryland forest on yellow latosol near Manaus, Brazil. 2 2 Total catch each season = 100%. Abundance for each layer (N/m ) and total catch per season (S N/m ).

2 abundance was calculated for 1 m . Vertical dis- disregarded) in the study area (Adis et al, un- tribution of adults and nymphs in relation to published data). During the dry season, about changing conditions of soil moisture content, 41% of all 1,700 (±71) pseudoscorpions were temperature and pH was statistically evaluated collected from the top 3.5.cm, 37% from below with the linear correlation test (Cavalli-Sforza, the humus layer (3.5-7cm) and 22% at 7-14cm 1972), using the original field data. depth. About 54% of all specimens collected represented juvenile stages (Fig. 1: nymphs). RESULTS Decreasing abundance of pseudoscorpions at greater soil depths was significantly correlated Primary Forest on Yellow Latosol with increasing soil moisture content (adults and Pseudoscorpions accounted for 10-15% of all nymphs: P<0.01, r = -0.998, nymphs only: arthropods extracted from 1 m soil of 14cm depth P<0.05, r = -0.956; n = 4). During the rainy 2 (10,000-12,000 ind./m , Acari and Collembola season, vertical distribution was similar (Fig. 1): . — )

NEOTROPICAL PSEUDOSCORPIONS 435

DRY SEASON RAINY SEASON

2 25 5D U43 62A 7Z1 N/m ; 351. 2 53.C ) 33.7 A 3.3 60-

2 2 N/rn = 533.8 2N/m =461.2 45 J

30-

-•^" - 15 • ~

0-35 35-7.0 7.0-10.5 10.5 -U JO 0-35 3.5-70 7.0-10.5 105 -K.0 Soil Depth (cm) Soil Depth (cm)

= nymphs I l = ad ul t s = soil moisture content

FIG. 4. Distribution of pseudoscorpions in soil and soil moisture content (both in %). Samples taken every 3.5cm

to depth of 1 4cm during dry season and rainy season in primary dryland forest on white sand soil near Manaus. Brazil. Total catch each season = 100%. Abundance for each soil layer (N/nr) and total catch / season (SN/nr

38% of the 1.100 ( ±46) pseudoscorpions were catch. Less abundant (> 1 0%) were Miratemnidac recovered from the top 3.5 cm, 29% from 3.5- (Brazilaiemnus browni Muchmore ; dry season: 2 2 7cm and 33% from 7-14cm soil depth. About 149±20 ind./m , rainy season: 82 ± 13 ind./m ) 61% of the total catch were nymphs. Decreasing and Ideoroncidae (Albiortx gracilis Mahnert). abundance of pseudoscorpions at greater soil Few differences in dominance were found for depths was significantly correlated with decreas- species captured during both seasons, however ing soil temperature (adults and nymphs: P<0.01 their abundance varied (Fig. 2). r = +0.993. nymphs only: P<0.05, r= +0.956; n = Except for A. gracilis, vertical distribution of 4). pseudoscorpions in the soil differed between Six species of pseudoscorpions were collected species but not within species when comparing in both seasons (Fig. 2). All were previously different seasons (Fig. 3). /. schusteri and B. reported from soils of Amazonian dryland forests browni were only found at 0-7cm soil depth, with (Mahnert and Adis, 1985). Syarinidae were most abundances being greatest in the top 3.5cm. This abundant, with Microblothrus tridens Mahnert is also true for 7. minor which lived, like A.

(dry season: 1 160 ±55 ind./m*, rainy season: gracilis, to a soil depth of 10.5cm. However, 2 860 ±24 ind./ra ), Idcobisium schusteri Mahnert occurrence was restricted to the upper 7cm in T. and Ideoblothrus tenuis Mahnert accounting for minor during the rainy season and in A gracilis

74-78% of the total catch (Fig. 2>. Chthoniidae during the dry season (Fig. 3 ). /. tenuis was found was the next most abundant family, with Tyran- only at a soil depth of 3.5-7cm (= below the nochthonius (T.) minor Mahnert (= Lagynock- humus layer) during both seasons, whereas M. thonius minor (Mahnert)): dry season: 212± 17.4 tridens occurred in all soil layers (0-l4cm). 2 2 significant correlation was ind./m , rainy season: 140± 14 ind./m ) and (dry During thedry season, season only) Pseudochthonius homodentatus observed between decreasing abundance of T. Chamberlin representing 12-13% of the total minor and greater soil depths (Fig. 3) and the

1 At this locality, B. browni is apparently smaller. It may have adapted to its environment, thus representing an c-co-species. 436 MEMOIRS OF THE QUEENSLAND MUSEUM

DRY SEASON % N/m RAINY SEASON % N/m

100.0 533.8 100.0 481.2

|^3.3 230.9 35.0 158,4

i4-. -.--.. 30_6 163.6 3 2.0 154.4

-:-:=:-;-z -;fi;ij . B. brown i 17. ] 91.3 M tridens 2 8.0 134.8

I 3.0 H.4 I , schusten 5.4 28.8 schusten

R homodentatus 1.8 9.5 A. arboncola 1.0 4.8

A. arboncola 0.9 4.8 T rolundimanus 1.0 4.8

P crassifemoratus 0.9 4.8 "1 20 60 (%) 20 60 (%

2 2 FIG. 5. Dominance (%) and abundance (N/m ) of species of pseudo scorpions extracted from lm soil (0-1 4cm depth) during dry (August 1988) and rainy (March 1988) seasons in primary dryland forest on white sand soil near Manaus, Brazil. Total catch of each season = 100%. (See text for further explanation). increased soil moisture content (P<0.05; adults all specimens collected represented juvenile and nymphs: r = -0.959, adults only: r = -0.968; stages (Fig. 4: nymphs). Decreasing abundance P< 0. 10; r = -0.936 for nymphs only; n= 4, respec- of pseudoscorpions at greater soil depths was tively). Similar results were recorded from A. significantly correlated with decreasing soil gracilis (P<0.05; adults and nymphs: r = -0.969, moisture content (P<0.05; adults and nymphs: r nymphs only: r = +0.974; n = 4). During the rainy = +0.965, adults only: r = +0.952; n=4) and season, decreasing abundance of 7". minor and B. decreasing soil temperature (P<0.05; adults and browni (Fig. 3) was significantly correlated with nymphs: r = +0.967; n=4). During the rainy increasing pH values at greater soil depths (T. season, vertical distribution was similar but more minor: P <0.05, r = -0.970 for adults and nymphs, pronounced (Fig. 4): 73% of the 480 ( ± 27) pseu- r = -0.951 for nymphs only and P<0.01,r = -0.994 doscorpions were taken from the top 3.5cm, 1 1 % for adults only; r = -0.998 B. browni: P<0.01, for from 3.5-7cm, 7% from 7-10.5cm and 9% from adults and nymphs and r = -0.995 for nymphs the white sand soil layer (10.5- 14cm). About 68% only; n = 4, respectively). In M. tridens, the of the total catch were nymphs. No significant increasing abundance of nymphs at greater soil correlation was found between the vertical dis- depths was significantly correlated with decreas- tribution of pseudoscorpions and changing con- ing soil moisture content (P<0.05, r = -0.977; n = ditions of soil moisture content, soil temperature 4). The decrease in abundance of adults, however, and pH. was significantly correlated with the decrease of Seven species of pseudoscorpions were ob- the soil temperature at greater soil depths tained during the dry and six species during the (P<0.05, r = +0.962; n = 4). This was also ob- rainy season (Fig. 5). All were previously served for nymphs of T. minor (P<0.05, r = reported from soils of Amazonian dryland forests +0.951; n = 4). Neoteny and potential par- (Mahnert and Adis, 1985). Chthoniidae were thenogenesis were confirmed for M. tridens, with most abundant, with Tyrannochthonius sexually mature tritonyrnphs (= males) being ab- (T.J minor Mahnert (dry season: 231 ±13 ind./nV\ sent (cf. Adis and Mahnert, 1990a; Mahnert, 2 rainy season: ind./m T. 1985). 154±15 ), (T.) rotun- dimanus Mahnert and Pseudochthonius homodentatus Chamberlin accounting for 33- Primary Forest on White Sand Soil 45% of the total catch (Fig. 5). Syarinidae was the

Pseudoscorpions accounted for 3-4% of all next most abundant family with Microblothrus tridens Mahnert (dry 2 arthropods extracted from 1 m" soi 1 of 1 4cm depth season: 164± 10 ind./m , 2 rainy season: 2 (14,000-15,000 ind./m , Acari and Collembola 135 ±8 ind./m ) and Ideobisium 2 disregarded) in the study area (Adis et a/., 1989a, schusteri Mahnert (dry season: 29 ±3 ind./m , 2 b). During the dry season, about 48% of the total rainy season 14 ±4 ind./m ) accounting for 31- 530 (± 22) pseudoscorpions were collected from 36% of the total catch. The Miratemnidae were the top 3.5cm, 27% from 3.5-7cm, 12% from represented by Brazilatemnus browni Muchmore 7- 2 10.5cm (= humus layer) and 13% from the (17-35%; dry season: 91 ±6 ind./m , rainy 2 white sand soil layer (10.5- 14cm). About 66% of season: 168 ±16 ind./m ). Less abundant (^ 1%) ) ) 9

NEOTROPICAL PSHUDOSCORPIONS AM

DRY SEASON RAINY SEASON DRY SEASON RAINY SEASON

N/m 14.4 L£ 9.6 14.4 N/m 154.0 52.9 192 4.8 12S9 24.1

100 -i 100 Ideobismm schusten Tyrgnnocntornus minor

2 4 3 N/m =2 8.8 jN/m -- 14.4 N/m" = 230.9 2: N/m" = 154.0

50 - 50

; a , f—r Q n

N/rr? 48.1 14 4 19.2 9.6 120.3 4.8 24.1 19.2 N/m 24.1 722 192 4ei 770 24,1 9.6 24.1 100 'Ui Brazilatemnus browni Microbtotfrrus tndens 0) 2 2 War = 91 3 iN/^ = 168 4 2 N/m = 163-6 SN/m = 134. a c 50 50 u

a> a Run El RaR T r 4-ATL A 1 r T I T 3.5 -70 -10.5 -44.0 -3.5 -70 -105 -14 G -3 5-7.0 -10.5-140 -35 -7.0-10.5-14.0 Soil Depth (cm Soil Depth (cm

nymphs ] = adults

FIG. 6. Vertical distribution of four species of pseudoscorpions in soil (%). Samples taken every 3.5em to depth

of 1 4cm during dry and rainy seasons in primary dryland forest on white sand soil ncarManaus, Brazil. Total 2 catch of each season = 100%. Abundance given for each soil layer (N/m") and for total catch per season (XN/m ).

were Idcoroncidae [Albiorix gracilis Mahncrt) tween abundance and soil conditions was ob- and Chernetidae (Pseudopikmusc rassifemoratus served in T. minor: its abundance decreased at Mahnert). The same four species (T. minor, B. greater soil depths as the soil moisture content browni, M. tridens and /. schusteri) were the most (P<0.05, r= +0.979 only for adults; P<0.10, r = abundant in both seasons, however, their +0.935 for adults and nymphs, r = +0.900 for dominance in the total catch per season varied nymphs only; n = 4, respectively) and the soil (Fig. 5). In two species, vertical distribution dif- temperature decreased (P<0,05, r = +0.967 for fered with the seasons. During the rainy season, adults and nymphs, r - +0.972 for adults; P<0. 10, r = +0.939 for nymphs only; n =4, respectively). T. minor and /. schusteri were found only from No significant correlation was found during the 0-7cm and in the top 3.5cm of soil, respectively. rainy season between the vertical distribution of Both species were found throughout the 0-14cm species and the abiotic factors investigated. In A/. soil sample during the dry season (Fig. 6). Their tridens sexually mature tritonymphs (= males) abundance was greatest in the top 3.5cm, inde- were absent, which confirms neoteny and poten- pendent of seasons. This is also true for B. browni tial parthenogenesis in this species (of. 4. 1 .). which, together with M. tridens, occurred throughout the 0- 1 4cm soi I sample during the dry season as well as the rainy season. Adults of M, DISCUSSION tridens were somewhat more abundant in the top 7cm. whereas no adults of/, schusteri were col- In Central Amazonia the abundance of pseudo- lected, probably due to low catch numbers. A. scorpions in primary and secondary dryland arboricola was restricted to the top 3.5cm. forests on yellow latosol accounted (inde- During the dry season, significant correlation be- pendently of seasons) for 3-5 Tr of all arthropods 438 MbMOIRS OF THE QUEENSLAND MUSRIM

extracted from 1m* soil of i4cm depth and for ture in the primary forest on white sand mm I (cf,

4. , 10-15% when Ac-ah and Collembola were ex- 1 4.2.). In addition, decreasing abundance with cluded (cf. Adis e? aL. 1987a, b, unpublished depth was also related to somewhat lower data). In the primary forest on white sand soil. temperatures recorded from the lower levels of their abundance was only 0.7-0.9% and 3-4%, soil in Ihe white sand forest. During the rainy respectively, due to large numbers of Pauropoda season, no correlation was found between species and Diplura (cf. Adis & or*-. 1989a. b). About abundance and soil moisture content in these two two-thirds of all pscudoscorpions recovered from forest types. However, in the forest on yellow the soil of the primary and secondary forest on latosol there was positive correlation between yellow latosol (61-67% of (he total cjuhi and decreased abundance of nymphs and lower soil about ihree-uUartcrs Of the primary white sand temperatures and negative correlation between forest (75-84*1 inhabited the lop 7cm. No decreased abundance of nymphs, nymphs and species were found to occur exclusively in the adults (P<0.05) and increased soil pH. Ai greater lower, mineral subsoil (e.g. in 10.5-14cm soil soil depths there was an even higher statistical (7". of negative correlation be- depth). Two species fniwrt H. bnwxn WW probability (P<0.01) more abundant in the upper, organic layer tn the tween decreased abundance of adults and in- and three forests which were in tod i Figs 3, 6: creased pH. Correlation between abundance fag. 3 in Adis and Mahncn. 1990a). M. tridens, T, il moisture content has not yet been reported for

- and B, hruymi were the most abundant T. minor from the secondary forest on yellow species, representing 89-95% of the total catch in latosol (Adis and Mahnert, 1990a). The more 0- the primary forests on yellow latosol and on white homogeneous occurrence of A/, tridens in 1 4cm sand soil. They represented 5> f the total soil depth (Figs 3, 6; fig. 3 in Adis and Mahnert catch in the secondary forest on yellow latosol, 1990a) was correlated with soil moisture content correlation between where /. tenuis and A \ were frequent as as well: there was positive well (22-33% of the total catch; cf. Adjs and decreasing numbers of nymphs and sod moisture Mahnert, 1990a). The similar spectuun of paeu- in the primary forest on yellow latosol during the doscorpion species in forests on yellow latosol rainy season and, correspondingly, pegai and on white sand soil and especially the lack ot relation between increasing abundance of endemic species in the latter forest type, confirm nymphs and decreasing soil moisture content at the geological results ufChauvel fll aL (1987). greater soil depths in the secondary forest on They reported that the white sand area inves- yellow latosol (cf Adis and Mahnert, 1990a) In tigated represents the final stage of podzolisation, adults, howeveT thcic wn% a posiiive correlation i.e. the transformation of clayey latOSOla to white between decreased abundance il moisture sandy podzols by (long-term) Weathering And content (secondary forest) and between leaching processes [f the white jtmdswc decreased abundance and lower temperatures at large dried-up old riverbed, species composition greater soil depths (primary would be different Irom the yellow latosol area Thett data all suggest if ips other than id Adi* aid Mahncn 1990b). soil moisture content (abiotic and/or biotic ones) Neither during the dry nor the rainy season: js welt cause the different distribution of Ihe abundance of pscudoscorpions in mineral pseudoscorpion species in the soil. A significant subsoils higher in response to the changing mois- correlation, for example, was found between Ihe ture content in organic layers. This phenomenon ahunCentral America (Walter andLietli. I960 67) and (dry season: P<0.01, r = +0.991, rainy sc, for Manaus (Worhcs, I986)J. Higher or lower P<6.05. r = +0.967; n - 4). This relationship hiiN abundance of (at least some) species at a distinct to be investigated in more detail, as sprii .^tail- soil depth seems lo he related to different abiotic represent a potential food source for pscudoscor- factors. For example, the decreasing abundance pions (cf. Jones, 1975; Rcssl and Bcier, 1958; of 71 minor with soil depth during the dry season Wcygoldt, 1969). Thus, complementary studies correlated with soil moisture content in two should be carried out on reproduction patterns, fbfl&l types: negatively with in E 'Moisture food availability and trophic structure over a content values in the primary forest on yellow one-year period (cf. Adis et aL, 1988). Data also latosol and positively with decreasing soil mois- show i hat the number of samples taken per soil 1

NEOTROPICAL PSEUDOSCORPIONS •1.i-j

layer (n= 6) was sufficient for a species Vertical distribution and abundance of arthropod* dominance of 5% of the total catch, as standard in the soil ofa Neotropical secondary forest during 374- S deviation did not exceed 203?. Results indicate die drv season. Tropical Ecology 28: 1 ADIS. J.. MORAIS. J.W.DE & MESQUITA. H.G.DE that terncole pseudoscorpions of primary dryland 1987b. Vertical distribution and abundance of forests on white sand soil and of secondary arthropods in the soil of a Neotropical secondary dryland forests on yellow latosol represent the forest during the rainy season. Studies on lerricole species spectrum of a primary dryland Neotropical Fauna and Environment 22; 1 89- 1 97 forest on yellow latosol, which has been altered ADIS, J., MAHNERT, V.. MORAIS, J.W.DE

ACKNOWLEDGEMENTS ADIS. J., RIBEIRO. L K. MORAIS, J.W.O CA VALCANTE, E.T.S. 1989b. Vertical distrihu The pseudoscorpion material was obtained lion and abundance nfarthropods from white sand during * soil of a Neotropical campinarana forest during tlte the post-graduate courses Eniomolo£ ic a I season. Studies Neotropical not! Field Ecology* of INPA/University Amazonas at dry on Fauna hri vironment 24; 201-21 1, Manaus (Brazil) in August 1988 and Tico-en- ANDERSON, A.B. 1981. White-sand vegetation ol tomological course on neotropical arthropods' of Brazilian Amazonia. Biotropiea 13: 199-210. the Chrisiian-Alhrechis-UniveTsity at Kiel (Ger- ANDERSON, A.B., PRANCE, GT. & ALBUQUKR many) in October 1990. All participants are QUE. B.w.P.DE i975.£srudoe sdtwe a vege thanked for their valuable collaboration. Inge das Campinas Amazotucas -III. A vegetacao Jen- Schmidt- Brockmann kindly typed the hosa da Campina da Reserva Biologica INPA manuscript, Susannc Hamann (both at MP1 in SUFRAMA (Manaus-Caraearaf, km 62). Acta made the drawings and Nancy Wetder cor- Amazonit -246. rected the English draft. BECfc L, [964, Tropisehe BodcitftuitB im Wecbsrl von Regen-und Trocker^ek Natur und Museum LITERATURE CITED 94:63*71 BRINKMANN. W.L.F. 1970. Relative light intensity mr.asiuvmer.ls in a secondary fnresl (capocira) ADIS, J. 1987. Extraction of unhropixis from ruMintpi- Hear Manaus, Amazonia. Bra/.iL Boletim do INPA cal soils with 8 modified Kempson apparatus.

i Manaus; Brazil), FesquisasFlorestais J7: 1-h Journal of Tropical Ecology 3: 131 138 1971 Light environment in tropical rain forest o| ADIS, J .& MAHNERT, V. 1985. On the natural history and ecology of Pseudoscorpiones (Arachnida) Central Amazonia. Acta Amazomca 1: 37-49. J.A. l967.The Arthrop-KJaol iropkal SO Is from an Amazonian hlackwater inundation forest. BULLOCK. Amazoniana9: 297-314. and leaf liUer. Tropical Ecology 8: 74-87. 1990a. Vertical distribution and abundance of Pseu- BRAGA. PIS. 1979. Subdivisao OtogcografVca, lipcm ir.vent.irio doscorpiones (Arachnida) in the soil of 3 dc vegeiacao, conservacao e florisUeo Neotropical secondary forest during the dry da floresta amazomca. Acta Amazonica 9 (Sup season and rainv season. Acta Zoologica Fennica plemeni-; 53-80.

Cfc 11-16. CAVALLi-SFORZA.L. 1972. Grundzugebiologisclv 1990b. On die species composition of Pseudoseor- medizinischer Statistik.* (G. Fischer. Sturt

piones from Amazonian dryland and inundation CHAUVEL,A..LUCAS,Y.&BOULET.R I forests in Brazil. Revue Suisse de Zoologie 97: dynamics of the Amazonian tcn-a-firme fon 49-53. the genesis of the soil mantle of the F$gl0l ADIS, J. & SCHUBART, H.O-R. 1984. Eooloj Manaus, Central Amazonia, Brazil. Expcrieruu research (>n arihrop

forest ecosystems with recommendations for DECICO, A.. SANTOS. H.M . RIBEIRO. M DE N G study procedures. 111-144. In J.H Cooley and & SALATI, E. 1977. Estudos climatologicos da

F.B. Golley (eds) Trends in ecological research Reserva FJorestal Ducke, Manaus, AM I for the 1980*/ (NATO Conference Series, Scries Gcolernperaturas. Acta Amazon icu 7. 485

I; Ecology. Plenum Press; New York, London j. FALESLI.& S1LVA.B.N.R.DA 1969. OaJtpfosds

ADIS. J., MORALS, J.W.DE & RIBEIRO, EF. 1987a. Manaus-llacoalzara. Sera ENtudados cEnsaios no. 440 MEMOIRS OF THE QUEENSLAND MUSEUM

soil populations and their role in 1 . Sccretariade Produgao do Amazonas. (IPEAN: analysis of fauna Belem, Brasil). decomposition processes. Oikos 39: 288-387. GOFFINET, G. 1976. Ecologie edaphique des milieux PRANCE, G.T. 1990. The floristic composition of the naturels du Haut-Shaba (Zaire). I. forests of Central Amazonian Brazil. 60-71. In Characteristiques, ecotopiques et syn6cologie Gentry, A.H. (ed.) 'Four neotropical rain forests.' comparee des zoocenoses intercaliques. Revue (Yale University Press: New Haven). d'Ecologie et de Biologic du Sol 12: 691-722. PRANCE, GX, RODRIGUES, W.A. & SILVA, GUILLAUMET, J.-L. 1987. The dynamics of the M.F.DA 1976. Inventario florestal de um hectare Amazonian terra-firme forest: Some structural de mata de terra firme km 30 da Estrada Manaus- and florislic aspects of the forest. Experientia 43: Itacoatiara. Acta Amazonica 6: 9-35. 241-251. RESSL, F. & BEIER, M. 1958. Zur Okologie und JONES, P.E. 1975. Notes on the predators and prey of Phanologie der heimischen Pseudoskorpione. British pseudoscorpions. Bulletin of the British Zoologische Jahrbiicher. Abteilung fur Sys- Arachnological Society 3: 104-105. tematic Oekologie und Geographie derTiere 86: LAWRENCE, R.F. 1953. 'The biology of the cryptic 1-26. fauna of forests with special references to the RIBEIRO, M.DEN.G. & ADIS. J. 1984. Local rainfall indigenous forests of South Africa.' (Cape Town). variability-a potential bias for bioecological LECHTHALER, R. 1956. 'Inventario das arvores de studies in the Central Amazon. Acta Amazonica um hectare de terra firme da zona "Reserva Flores- 14: 159-174. tal Ducke" muniefpio de Manaus.* (Amazonia RIBEIRO, M.DE N.G. & NOVA, N.A.V. 1979. Es- (Botanica) no. 3, INPA: Manaus, Brazil). tudos climatologicos da Reserva Florestal Ducke, LEVINGS, S.C. & WINDSOR, D.M. 1984. Litter mois- Manaus, AM. III. Evaporatranspiracao. Acta ture content as a determinant of litter arthropod Amazonica 9: 305-309. distribution and abundance during the dry season RIBEIRO, M.DE N.G. & SANTOS, A.DOS 1975. on Barro Colorado Island, Panama. Biotropica 16: Observacoes micro climaticas no ecossistema 125-131. Campina Amazonica. Acta Amazonica 5: 183- LIEBERMANN, S. & DOCK, C.F. 1982. Analysis of 189. the leaf litter arthropod fauna of a lowland tropical RODRIGUES, J.M.G. 1986. 'Abundancia e evergreen forest site (La Selva, Costa Rica). distribui^ao vertical de Arlhropoda do solo, em Revista de Biologia Tropical 30: 27-34. capoeira de terra firme.' (M.Sc. Thesis, INPA: LIS BoA, P.L. 1975. Estudos sobre a vegetacao das Manaus, Brazil). 80 p. Campinas Amazonicas. II. Observacoes gerais e RODRIGUES, W.A. 1967. Inventario florestal piloto revisao bibliografica sobre as campinas ao longo da estrada Manaus-Itacoatiara: Dados amazonicas de areia branca. Acta Amazonica 5: prcliminares. Atas do Simposio sobre Biota 211-223. Amazonica 7: 257-267. MAHNERT, V. 1985. Weitere Pseudoskorpione RYBALOV, L.B. 1990. Comparative characteristics of (Arachnida) aus dem zentralen Amazonasgebiet soil macro fauna of some tropical savannah com- (Brasilien). Amazoniana 9: 215-241. munities in Equatorial Africa: preliminary results. 1-11. MAHNERT, V. & ADIS, J. 1985. On the occurrence Tropical Zoology 3: and habitat of Pseudoscorpiones (Arachnida) STRICKLAND, A.H. 1947. The soil fauna of two from Amazonian forests of Brazil. Studies on contrasted plots of land in Trinidad, British West Neotropical Fauna and Environment 20: 21 1-215. Indies. Journal of Animal Ecology 16: 1-10.

MARQUES, A.OEO.R,RIBEIRO,M.DEN.G., SAN- TAKEUCHI, M. 1961 . The structure of the Amazonian TOS, H.M.DOS & SANTOS, J.M.DOS 1981. vegetation. II. Tropical rain forest. Journal of the Estudos climatologicos da Reserva Florestal Faculty of Science. University of Tokyo. Section Ducke, Manaus-AM. IV. Precipita^ao. Acta 3: Botany 8: 1-26.

Amazonica 1 1: 759-768. WALTER, H. & LIETH, H. 1960-67. MERINO, J.F. & SERAFINO, A. 1978. Variaciones 'Klimadiagramm-Weltatlas.' (G. Fischer: Jena). mensuales en la densidad de microartropodos WEYGOLDT, P. 1969. 'The biology of pseudoscor- eddficos en un cafetal de Costa Rica. Revista de pions.' (Harvard University Press: Cambridge, Biologia Tropical 26: 291-301. Massachusetts). MORAIS, J.W. DE 1985. 'Abundancia e distribuigao WILLIS, E.O. 1976. Seasonal changes in the inver- vertical de Arlhropoda do solo numa floresta tebrate litter fauna on Barro Colorado Island, primaria nao inundada.' (M.Sc. Thesis INPA: Panama. Revista Brasileira de Biologia 36: 643- Manaus, Brazil). 657. PENNY, N.D. & ARIAS, J.R. 1982. 'Insects of an WORBES, M. 1986. Lebensbedingungen und 1 Amazon forest. (Columbia University Press: Holzwachstum in zentral amazonischen New York). Uberschwemmungswaldern.ScriptaGeobotanica

PETERSEN, H. & LUXTON, M. 1 982.. A comparative 17: 1-112. NEST ASSOCIATES OF CLUBIONA ROBUSTA L KOCH

i ARANEAE: CLUBIONIDAE) IN AUSTRAL! ^

A.D. AUSTIN

Austin, . . A.D. 1993 II II' Nest \ i|es of Clubiotta robusta L Koch (Araneue. Clubinnidac) in Australia. Memoirs of the Queensland Museum 33(2): 441-446. Brisbane. ISSN 0079-8835.

The nests of Cluhkma robusta, whether or not they contain a resident female spider, arc usually inhabited by other invertebruio:. here termed nest associates. The abundance and diversity of nest associates for C robusta arc described from sampled ncsiv kssocj recorded were; other spider*. Acarina. Mollusea and insects of the orders Collembola. Coleoplcra, Hcteroptera, Hymenoplera, Lepidoptera Neuroptera. F'socopiera and Thysanop- lera. The presence of a resident female was a feclox which affects the abundance and v> of associates. Biological notes on each group of nest associates are given. Results for C. robusiu arc compared Willi data on other spiders in Ihe same habitat, and other categories or" nest associates used in other studies Q4 raneae, CM poros-itn ne\i

A.D. Austin, Deportment of Crop Protection, Waire Campus Lhvie>.u>\ af Adcli;wu\ f>,0. Glen Osmond. South Australia 5064, Australia; 18 August, 1992.

Most vagabond spiders build a silken retreat in During a study of the biology of C. robusta in which they moult, mate, lay eggs and protect South Australia (Austin, 1984a* 1 988), incidental themselves and their eggs from predators and observations showed that 1 } most nests contained extreme environmental conditions (Jackson. at least some associates. 2) associates inhabited 1979; Jackson and Griswold. 1979; Pollard and both vacant and occupied nests, and 3) the size Jackson, 1982; Austin, 1984a, 1988; Wolf, 1990). range of associates appeared to differ between Different species construct a great variety of vacant and occupied nests. This study then sets retreats which vary in size, structural complexity out to detail thediversru and abundance of nest and the location in which they are placed. For associates found in C. whnsia nests, and the Chilnona robusta L Koch, a moderate-sized factors that might determine their number and dubionid (body length about 15 mm) which in- size. habits the corticating (shedding) hark of eucalypt trees in south-eastern Australia, two types of MATERIALS AND METHODS retreats have been recognized: flimsy, thin silk- walled moulting chambers constructed by Nests of C. robusta were collected from under juveniles, and thicker walled nesis constructed by the bark of eucalypt trees with large plates Of mature females for egg laying (Austin, 1984a) corticating bark at their bases (i.e. Eucnlx Whatever the structure of spider retreats, they camaldulensis, E. cladvealyx, E. leucoxylon and often attract an array of other invertebrates E. viminolis) in the Mylor area of the Adelaide (termed 'nest associates') which utilize them in a Hills, South Australia (sec Austin. 1984a). Nests

Variety Of ways f Auteru 1925; Jackson and Gris- were collected at random in the summer ( Deceit) wold, 1979; Griswold. 1986). This aspect of ber-February) over three years (1979-19X1) spider retreats lias been examined only rarely. A Nests constructed only in that summer were in- study of the nest associates of the salticid Phidip- cluded in samples. The whiter, fresher colour of pus johnsoni Peckham and Peekham in the the silk was used to distinguish them from nests U.S.A. by Jackson and Griswold (1979) »g the built in previous seasons. Samples included only known detailed work. Jackson and Griswold vacant nests, nests containing a resident spider, (1979) divided nest associates into eight and nests containing a resident spider and c^t*- categories, following the three described by mass or hatched juveniles, in the field, after bark

Au ten ( 1925), based on their biological relation- plates were removed from trees, nests were gently ship with the resident spider, its eggs or the nest teased away and placed in screw-top plastic vials itself (sec Table 1). Most of these biological (40 mm din.). A large while dish prevented au categories have also been recognized far two eidenial 1- ies or resident spi eresid species in South Africa (Griswold, 1986). In the laboratory, when ncM associates were en- 442 MEMOIRS OF THE QUEENSLAND MUSEUM

Category Description was observed directly (e.g., parasitoids or

Endoparasitoids of spiders that seek out and/or predators of eggs or adult C. robusta-1 Ogcodes la emerge from their hosts inside nests sp., Ceratobaeus masneri Austin), or for those Ectoparasites and ectoparasitoids of spiders that seek lb species whose biology was well enough known out and/or emerge from their hosts inside nests to rule out all categories except one (e.g., 2 Endoparasitoids of in eggs nests Lepidoptera pupating in nests or Hymenoptera 3a Predators of eggs in nests (including 'ectoparasitoids') which specifically parasitize other spiders). For 3b Predators of spiders in nests other associates, multiple categories are listed to Scavengers feeding on dead spiders or other material 4 indicate the range of biological associations that in nests might exist. For many species it was impossible Predatory spiders that adopt nests as a predatory 5 device to determine whether nests were being used as a specific refuge or the species was there by acci- Spiders that adopt the nests of other spiders as a 6 refuge or substitute for constructing their own nests dent (categories 7/8). For others, their association

Organisms (other than spiders) that adopt nests as a with C. robusta appeared to vary. Gnaphosid and 7 refuge or pupation site segestriid spiders were observed in the field prey- 8 Accidental inhabitants ing on subadult C. robusta (category 3b, see 9 Conspecific males below), as well as constructing their own nests inside those of the host spider (category 6). TABLE 1 . Categories of nest associates, modified after Jackson and Griswold (1979) (see text for further Nest associates seemed to occur both in vacant information). nests and those occupied by a female spider, there appearing to be fewer and smaller associates in countered, details of the nest were recorded along the latter. To test these hypotheses, i.e. that oc- with any relevant biological information for the cupied nests had fewer and smaller associates, the associate, and they were then preserved in 70% fauna from 30 occupied nests and 30 vacant nests alcohol for identification. Parasitized spider eggs were compared (Table 3). Of the 40 recognized and the larvae and pupae of mantispids, pom- taxa that were collected during the 1979-81 sur- pilids, Diptera and Lepidoptera were left to vey, 28 were found in the 60 sampled nests, with emerge prior to preservation. As in Jackson and Diptera and Neuroptera being the ordinal taxa not Griswold (1979), nest associates only within the recorded. The 30 occupied nests had substantially nest, within the silk matrix of the nest wall, or fewer nest associates (total 147) compared with under the nest were collected, not those complete- vacant nests (227) (see Table 3). Using a Wil- ly outside or near the nest. To compare differen- coxin test (Zar, 1984) occupied nests contained ces in the diversity, size and abundance of significantly fewer associates per nest than associates between vacant nests and those oc- vacant nests (Z = 2.36, P <0.02), and significantly cupied by a resident female, 30 nests of each type fewer associates larger than 2.5 mm in length (Z were collected at random in the field during = -4.38, P <0.0001). The two major groups of January 1980 and examined. associates that seemed to differ most notably Voucher specimens of each nest associate were between vacant and occupied nests were other lodged in the Department of Crop Protection species of spiders and Coleoptera, which were insect collection, Waite Campus, University of both mostly represented by individuals much Adelaide, South Australia. greater than 2.5 mm long. This specific size was differences size of nest as- RESULTS chosen to test for in sociates because previous work on C. robusta has shown that this is about the minimum size for 400 nests sampled during the summers of Over potential prey that females spiders can physically to yielded species nest 1979 1981, over 40 of handle (Austin, 1988). Hence, one suggested associates including Acarina, Araneae, Collem- cause for fewer large associates in occupied nests bola, Coleoptera, Diptera, Heteroptera, is that they would be killed or driven away by the Hymenoptera, Lepidoptera, Neuroptera, Psocop- resident spider. tera, Thysanoptera and Mollusca (Table 2). The biological association of each associate with C.

robusta (categories listed in Table 1), whether Biology of Nest Associates determined by direct observation or only as- Assigning nest associates to one or more sumed, is also given for each species in Table 2. biological categories first required that those used The exact category of individual associates could by Jackson and Griswold (1979) be modified. In be assigned only for those species whose biology particular, endo- and ectoparasitoids have been NEST ASSOCIATES OF CLUBTONA 443

Associate C. robusta Associate C. robusta Taxonomy of nesl associate Taxonomy of nest associate Category Category 1 11 III I Tl 111

Acarina: Mesostigmata Chloropidae Pseudogaurax sp. 3a +

Phytoseiidae Seiulus sp. 1 4/7/8 |+ \5 \2 Heteroptera Acarina: Prosligmata Reduviidae Empicoris rubromaculatus 3a/7/8 + 4 2 (Blackburn) BdelHdae Bdellodes harpax (Atyeo) 4/7/8 + 2 2 Pentatomidae Nezjara viridula L. 7/8 + Cyta latirostris (Hermann) 4/7/8 + 1 1 Poecilometis sp. 7/8 + Anystidae genus & sp. indet. 4/7/8 + 6 7 Hymenoptera ?Eupalopsellidae genus & sp. indel 4/7/8 + 1 2 Eulophidac Tetrasiichus sp. v> + Erythraeidae genus & sp. indet 4A7/8 + 3 2 Formicidae Iridomyrmex sp. 3a + 3 Araneae Scelionidae Ceratobaeus clubionus Clubionidae Clubiona cvciadata Simon 3a/6/8 + 3 2 7/8 + 2 2 Austin C.robusta L.Koch 6 9 + 8 2 C. lamponae (Hickman) 7/8 -i G Clubiona spp. juveniles* 6/8 + 16 11 C. masneri Austin 2 + 4 7 Gnaphosidae Intruda sp. 3a/3b/6/8 + 2 - Gnon sp. 7/8 1 Scotospilus Hahniidae sp. 3a/6/8 + - Trissolcus sp. 7/8 1

Lamponidae Lampona sp. 3a/3b/6 /8 + 1 Pompilidae Epipompilus sp. 3a + Sallicidae Myrmarachne sp. 3a/6/8 + 3 Lepkloptera

Ser\aea sp. 3a/6/8 1 + - Limacodidac Doratifera sp. (pupa) 7 1 Segestriidae Ariadna sp. 3a/3b/6 '8 + 2 genus & spp. indet. (larvae)* 7/8 + Collembola Neurnptera

Entomobryidae Drepanura sp. 1 & 2* 4/7/8 + 11 7 Mantispidae Austromantispa sp. 1 b/3a + Neanuridae Setanodosa sp. 4/7/8 + 5 3 genus & sp. indel. (larva) lb/3a + o lo Coleoptera Psocoptera Cocci nellidae 4/7/8 + 5 2 Lepidopsocidae genus & spp. indet.* 4/7/S + 2 5 Scymus parallelus Blackburn Liposcelidae Liposcelis sp. 4/7/8 + 6 8 Dermestidae Anthrenus sp. 4/7/8 + 2 1 ThysanopteraThripidae genus & spp. Trogoderrna antipodum Blackburn 4/7/8 + 1 4/7/8 + 6 7 indet.* Tenebrionidae genus & spp. indet.* 4/7/8 + 11 6 Mollusca: Gastropoda Helicidae 4/7/8 + 1 Diptera Cochlicella ventrosa (Ferussac)

Acroceridae '.' Ogcodcs sp. la *

TABLE 2. Nest associates in nests of Clubiona robusta and category of associate to which they belong (see Table

I and text). I = associates found at least once (+) in occupied or vacant nests over three-year period (1979-81 ); II = no. of nests in which each associate occurred in sample of 30 vacant nests (Jan 1980); III = no. of nests in which each associate occurred in sample of 30 nests occupied by 9 spiders (Jan 1980); * = nest associate represented by more than one species. distinguished (categories la and lb), and pred- Scavengers, refuge seekers and accidental in- ators of spiders and their eggs divided into sep- habitants. Most nest associates belong to one of arate categories (3a and 3b). The term 'parasitoid' these categories (25/40 taxa), but without detailed

(Askew, 1971) is used for species that kill their biological studies on individual species, it was host (e.g., scelionid egg parasitoids), and 'par- impossible to determine in which specific cat- asite' for species that feed from but normally do egory they belonged. The major groups are Acar- not kill their host (e.g., mantispid larvae feeding ina, Collembola, Coleoptera, Pentatomidae on adult spiders, see below). 'Ectoparasitoids' of (Heteroptera), Lepidoptera, Psocoptera (New, eggs, a term used by some authors, are not distin- 1974), Thysanoptera and Mollusca. In addition, guished from egg predators (Austin, 1985; La- three parasitic wasps can be categorized as either Salle, 1990). Also, a new category has been refuge seekers or accidental inhabitants; Cera- added; that of conspecific males (category 9, see parasitoid of below). tobaeus clubionus Austin, an egg Notes on the biology of nest associates were other clubionid spiders (not C. robusta -Austin, made both in the field and on nests and associates 1984b), Ceratobaeus lamponae (Hickman), an kept for short periods in the laboratory, during the egg parasitoid of Lampona spp. (Hickman, three-year survey of nests. 1967a), and Trissolcus sp. and Gryon sp., egg 7

444 MEMOIRS OFTHE QUEENSLAND MUSBUM

Nest residents C. robustu nests as a predatory device (category ik-s f|Q 9 resident v 5, see Jackson, 1976; Jackson and Griswold,

No of nesis sampled - 30 1979). '; eggs. ToCi! no. of associates 227 I Parasitoids and predators of C. robusta ilic Mean no. of assocuies/nest 7.67 45W An endoparasitic wasp, Ce, masneri, was

Total no. of associates >2.5imn 72 13 most commonly noted species (category 2) and vacant nests, occupied nests % of associates^.? rmn 31.7 8.9 was associated with without eggmasses, and occupied nests with egg- Mean no. of associate sfaeftt> 2.5iwn D 1 masses This wasp caused the highest level of egg TABLE 3. No. til nest associates for vacant nests and mortality (22-25%, Austin, 1984b, 1988) of all nests occupied by resident 9 spider. species found feeding on eggs. Pour species thought to be specific predators (/ectoparasites' parasiioids of pentatomid bug eggs (Masner. of other authors) were seen feeding on C. robusta 1976). eggmasses only rarely [category 3a'r These in- Other spider* that use C. robusta nests. More cluded a chlompid fly Pseudogaurax sp., on than 50% of vacated nests were occupied by other three occasions (Austin, 1985). a eulophid wasp, Spider*, including conspecitlcs, many of which Teltwtichus sp. (LaSalle, 1990), a pompilid had constructed moulting retreads or their own wasp. EpipotnpfltiS sp (Pollard, 1982), each on nests inside that of C robusta (category 6). Six- one occasion, and a manlispid. Austromonthpa teen of 30 vacam Qfests contained juvenile gp., on two occasions. The only gencraUst ChbJona which, in most cases, could not be predator of eggs recorded was the ant Indumyr identified to species. Some of these may have nwx sp which was observed feeding on cggnia;- come from pre. iousty hatched eggmasses in the scs and hatched eggs several times, but only in same nest. However. 1 1 of 30 occupied nc>ts also Meant nests (Austin. 1988). Other possible contained Ctubiona juveniles, 8 of which con- predators of eggs include a reduviid. EfftplvortS tained gravid female C. robusta and no pre rubromaculatus Blackburn, and other spiders. viousiy hatched eggmasses, indicating that these

juveniles had migrated from othet i n out of 60 nests sampled in January 1980 contained Assoc7jAi£s of Other Spiders on Eucai.ypi conspeciflc adult nudes, twd cohabitating with Bark mature females and eight found in oMef, • aeated During the 1979-81 survey, while collecting nests Oilier spiders were encountered more rare- nesis of C. robusta. nests of three relatively com- ly in the survey and only in vacant rteftlS All mon spiders were also collected and their nest species, except Km the hahnud Scowspiius sp.and associates examined. Observations on these three the segestriid Ariadna sp , at least once con- specie* are presented here. Other species were structed retreats or ncsN inside nests of C. robus- eithertoo rarely encountered to report on bete Of ta were not found associated with their nests. Para-iiioids and parasites of'adult C. robus.'a. Ser.ata sp (Salticidac). This same specie Only on two occasions were such organisms one very similar" was figured as Plerippui recorded. One aerocerid fly (lOgcades sp.) was vaiidus Urquhart by Hickman (1967b), who reared from a female C robust** (category la) described some aspects of its biology in Tas- (Schlinger, 19 A a single mantispid larva mania^ The spider constructs a flattened, cifi sibly Aitstrvmantispa sp.) was found at- nest (30-40 mm dia.) of dense, while silk, similar tached to the pedicel of a gravid female inside a in appearance and density to cotton wool Twen- nest (category lb). Some manbspid* are now- ty-five nests were collected: 20 occupied by known to act as ectoparasites nf female spiders as female spiders, and 10 contained eggiuasa small larvae and then as predators of the eggs as hatched juveniles. Only Srt associates were iden- larger larvae I'Redborg, 19H2; Redhwg and tified from t hese nests: A ustmmantispa sp. 2 (two

Macleod, 1983) ue sis i, Ogcodes sp. (one nest), PsewJogaura

Predators ofadult C. robusta The only poten- 2 » seven nests), Indomyrmex Sp. (one nest), tial predators of adult C. robusta collected during Amhrenus sp. (one nest) and an unidem the study were other large spiders, but no such carabid beetle (one nest). Indomyrmes sp. was predatiun was noted (category 3b). Four cases of feeding on eggs inside a vacant nest predatiun on subadult C robusta outside of Austromantispa sp., Ogcodes sp. and Ps. retreats were recorded, two each by haruda sp. gaurax sp. in these nests were different from and Ariadna sp. Also, no spider were noted Li&ing those in C robusta nests. Mites, Collembola and NEST ASSOCIATES OF CLUB IONA 445

Psocoptera were seen in some nests but were not examined in this study harbour a large and diverse extracted and identified. No scelionid wasps were invertebrate fauna that acts as a reservoir for C. reared from eggmasses. robusta nests. Also, the apparent high and consis- Bredajovialis (L. Koch) (Salticidae). This easi- tent abundance of this fauna means that there is ly recognized salticid builds a disc-shaped nest probably a much greater chance of nests being (about 30 mm dia.) with two prominent tubular inhabited, compared with the places P. johnsoni entrances opposite each other (Hickman, 1967b). constructs nests, i.e., under rocks, pieces of wood, The surface of the nest is hardened and parch- and in hollow reeds (Jackson, 1979). ment-like, but the inside is lined with dense, Occupied nests of C. robusta have smaller nest flocculent, white silk. Seventeen nests were col- associates compared with vacant nests presumab- lected: six had resident, gravid females, and five ly because resident spiders are known to kill and had females with eggmasses. Only two associates feed on intruders larger than 2.5mm. Also, were identified, and both were feeding on eggs. females are known to guard their eggmasses and These were Idris sp. (Scelionidae) and Pseudo- kill even distasteful intruders in nests (Austin, gaurax sp. Another scelionid is known to 1988). However, the aggressiveness of resident parasitize B. jovialis eggs elsewhere in Australia spiders does not explain why there are fewer (Hickman, 1967a; Austin, 1981), but this species small associates (<2.5 mm) in occupied nests (see was not found in South Australia. As for Servaea Table 2). One possible reason for this is that the sp., mites, Collembola and Psocoptera were seen activity of female spiders in building, repairing in a few nests but were not extracted and iden- and tidying nests, disturbs many small inver- tified. tebrates and they move away. Ariadna sp. (Segestriidae). This species builds The notes on nest associates of other spiders a flattened oval or tubular nest with walls of given here pose some questions. Generally the similar construction and silk density to that of B. nests of the three species examined have struc- jovialis. Only nine nests were collected, all con- turally more complex nests than does C. robusta, tained resident females, and four contained egg- and this is associated with them having far fewer masses. The only nest associates recorded were nest associates. In the light of this association, the larvae of a hymenopteran egg predator in one further research on nest associates could profitab- nest. These failed to develop and thus could not ly examine the structure of nests as a factor be identified. No mites, Collembola and Psocop- regulating the number and types of nest as- tera were seen in the nests of this species. sociates, as could differences in the behaviour of resident spiders inside their nests. DISCUSSION Finally, the biological categories proposed by Jackson and Griswold (1979) for the nest associates of P. johnsoni were directly applicable The number of species of nest associates col- to the associates recorded here for C. robusta and lected from C. rohusta nests and the major taxa the other three spiders examined and, thus, represented were similar to that found in nests of probably can be applied to the associates of any a salticid, P. johnsoni (Jackson and Griswold, nest-building vagabond spider. We have 1979). However, C. robusta differed in that many modified their scheme in two points: 1) to distin- more contained associates: of the 60 sampled guish between endo- and ectoparasitoids, be- nests 55 (92%) contained at least one associate, cause these two types of parasitoids have whereas only 1 .2 % of P. johnsoni nests were basically different biologies (Austin, 1985; La- occupied by associates. No nests of P. johnsoni Salle, 1990), and 2) to include conspecific males occupied by a resident female spider contained as a new category, which were found in both associates, whereas most occupied C. rohusta occupied and vacant nests. nests did (87%). P. johnsoni nests also differed in that a much greater proportion of recorded associates were spiders (66%), whereas for C ACKNOWLEDGEMENTS robusta* even excluding Acarina and Collem- bola, only 26% of associates were spiders. The I wish to thank the following people for iden- reason for these differences is unclear, but is tification of specimens: Val Davies, Mike Gray, probably due to differences between the habitats Penny Greenslade, Gordon Gross, Mark Harvey,

of the two spiders and the size and nature of David Lee, Eric Matthews and Tim New. I also potential invertebrate faunas that exist at the two thank Mike Keller for assistance with statistical locations. Certainly, the bark of eucalypt trees analyses, Sally Collins for assistance with field 446 MEMOIRS OF THE QUEENSLAND MUSEUM

work and comments on the manuscript, and Paul PhkUppus johnsoni (Salticidae. Araneae). Psyche Dangerfield for comments on the manuscript and 83: 243-255. proof-reading. 1979. Nests of Phidippus johnsani (Araneae, Sal- ticidae): characteristics, pattern of occupation, and function. Journal of Arachnology 7: 47-58. LITERATURE CITED JACKSON, R.R. & GRISWOLD, C.E. 19*79. Nest associates Of Phidippus johnsoni (Araneae, Sal- ASKEW, R.R. 1971. 'Parasitic Insects*. (Heinemann: ticidae). Journal of Arachnology 7: 59-67. London). 316pp. LASALLE, J. 1990. Tcirastichinac (Hymenoptera: AUSTIN, A.D. 1981. Hkkmanelta, a new genus of Eulophidae) associated with spider egg sacs. Jour- Scelionidae from Australia (Hymenoptera: Proc- nal of Natural History 24: 1377-1389. totrupoidea). Journal of the Australian En- MASNER, L. 1976. Revisionary notes and keys to tomological Society 2ft 303-308. world genera of Scelionidae (Hymenoptera: Proc- 1984a. Life history ofciubiona rubustu L. Koch and totrupoidea), Memoirs of the Entomological related species (Araneae, Clubionidae) in South Society of Canada 97: 1-87. Australia. Journal of Arachnology 12; 87-104. NEW, T.R. 1974. Psocoptcra from nests of the colonial 1984b. The fecundity, development and host spider bceuiicus Candidas (Koch) (Dicfynidac) in relationships of Ceratobaeus spp. (Hymenop- Western Victoria. Australian Entomological tera; Scclionidae), parasites of spider eggs. Magazine 2: 2-6. Ecologieal Entomology 9: 125-138. POLLARD, S.D. 1982. Epipompilns insularis 1985. The function of spider egg sacs in relation to (Hymenoptera: Pompilidae), a parasitoid of hunt- parasitotds and predators, with special reference ing spiders. New Zealand Journal of Zoology 9. to the Australian fauna. Journal of Natural His- 37-40. tory 19:359-376. POLLARD, S.D. & JACKSON. R.R. 1982. The biol- 1988. Guarding behaviour, eggmass shape and the ogy of Club'tona camBridgei (Aranea.

eggsac iti Chibtona robliSta L. Koch < Araneae. Clubionidae); intraspecillc interactions. New Clubionidae). Pp.X7-95.ln A.D. Austin and N.W. Zealand Journal of Ecology 5: 44-50. Heather (eds). 'Australian Arachnology'. (The REDBORG, K.E. 1982. Interference by the manlispid Australian Entomological Society. Brisbane). Montispa uhlfri with the development of the AUTEN, M. 1925. Insects associated with spider nesis. spider Lxcosu rabida. Ecological Entomology 7: Annals of the Entomological Society of America 187-196. 18*240-250, REDBORG, K.E. & MACLEOD. E.G. 1983. GRISWOLD,T.M. 1986. Nest associates of iwo species CHmacieUu hrunnea (Ncuroptera: Mantispidac):

of group-living Eresidae in South Africa. Adas X ;i manlispid thai obligately boards spiders. Journal Congress; Internationale Arachnologie of Natural History 17: 63-73.

Jaca/Espuna 1 : 275. SCHLINGER, E.I. 1987. The biology oi Acrocendae HICKMAN, V.V. 1967a. New Scelionidae (Hymenop- (Diplera): true endoparasitoids of spiders. 319- tera) which lay their eggs in those of spiders. 327. In W. Nenlwig (ed.K 'Ecophysiology of Journal of the Entomological Society of Australia Spiders'. (Springer-Verlag: Berlin). (N.S.W.)4: 15-39. WOLF, A. 1990. The silken nests of the ciubionid spider 1967b. 'SomcCommonSpidcrsofTasmania'. (Tas- Cheiracanihiutn penny! and Cheiracanihiutn manian Museum and Art Gallery: Hobart). pwwiprium (Araneae, Clubionidae). Acta 112pp. Zoologica Fennica 190: 397-404. JACKSON, R.R. 1976. Predalion as a selection factor ZAR, J.H. 1984. 'Biostaiisucal Analysis'. (Prentice- in the mating strategy Of the jumping spider Hall: New Jersey). 718pp. A TENTATIVE ANALYSIS OF THE SPIDER FAUNA OF SOME TROPICAL OCEANIC ISLANDS

L. BAERT and R. JOCQUE

Baert, L. and Jocque. R. 1993 II II; A tentative analysis of the spider fauna of sume tropical oceanic islands. Memoirs of the Queensland Museum 33(2): 447-454. Brisbane. ISSN 0079-8835.

The spider Fauna of oceanic islands in the Pacific (Galapagos. Juan Fernandez,, Easter island, Marquesas, Hawaii), the Atlantic (Saint Helena) and the Indian Ocean (Comoros, Seychelles) is analysed. Family composition, size frequency distribution, species/area relationship, the origin of the faunas and their qualitative compositions are discussed. The colonization of the islands by spiders appears to be greatly influenced by the accessibility of the island, its habitat diversity and the speciation capacity of particular groups. Most of the fauna is composed of species transported by man. Thorough analysis of island faunas is still hampered by the lack of complete published data sets. Les faunes araneologiques des iles oceaniqucs siiuees dans 1c Pacifique (Galapagos, Juan Fernandez, Marquises, Hawaii) 1'Atlantique (Sainte-Helene) et l'Ocean Indien (Cornores, Seychelles) sont analysees. Leur composition au niveau dc la famille, la distribution des frequences de taille, leur origine et leur composition qualitative sont disculees. La colonisation des Pes par les araignccs semblc ctrc fbnement influencce par racccssibiliie dc cesTlcs, leur diversite en habitats et la capacite de speciation de certains groupes. Une grandc partic des faunes est composee d'especes apponees paT Thomme. Une analyse complete des iles ocdaniqucs esi encore impossible a cause du manque dc donnces completes publiees.

\jAraneae l island zongeograph \

Leon Baert, Koninklijk Belgisch Instituut voor Natuurwetcnschappen, Vaufierxtrtwt 2$ />'- 1040 Brussels, Belgium; Rudy Jneque, Ktminklijk Museum VOO) Midden- Afrika, B-3080 Tervuren, Belgium; 26 October, 1992.

Most studies of island spiders have been con- ber of species in each family was expressed as a fined to the description and enumeration of percentage of the total number of species. The species- A general explanation of the composition Renkoncn-index was used to calculate the per- of the araneologica! fauna of islands has not been centage of similarity. tried. The availability of many almost complete species lists of remote oceanic and granitic is- RESULTS AND DISCUSSION lands allows us to analyse these spider faunas and try to understand their compositions. Only a few Family Similarity data-sets (those of Galapagos and Saint Helena) Because the islands are in different are nearly complete and most others are likely to zoogeographical provinces, a comparison of ihcir grow considerably. For several islands the faunas on the species level is not informative. species lists included many unidentified species Similarity on family level was surprisingly high submitted for study to us or present in other between the Galapagos, the Hawaiian Ar- collections (e.g. Cocos: Natural History Museum chipelago (Pacific) and Saint Helena (Atlantic, of Los Angeles County). Nevertheless we have reaching ca 70% (Fig. 1). A second group (cfl tried to Find patterns in the composition of the 60% similarity) consists of the Seychelles ami tropical island spider faunas. Cocos Islands. These distinct vegetation belts, as on the main islands of the First group, seem* MATERIAL AND METHODS important and induces a similar family composition. These islands were initially covered entirely Apart from the spiders of the Galapagos and the by forest. The southern pacific Raster Island and Comoros, all data used are from the literature Juan Fernandez- now covered by a gramineous, (Table 1). Hence, in the analysis of the size dis- so called pampa- vegetation, have little simit tribution, we had to use the total length, which is in family constitution. a less reliable measure than carapace width or the size of another sclcroiised body part. Body Length Frequency Distribution To compare the family composition, ihe num- In 1fpp

448 MEMOIRS OF THE QUEENSLAND MUSEUM

HSghflfl Area, Distance to nearest AgR h. land/ archipelago elevation Literature Consulted TABLE 1 (km' continent (km) (my > (m) Summary data S-J'J CbeostfOO] 46.6 500 (Cosla Rica 1 Hn?ue and Miller (1981) (age is ap-

SttbdcnaU&HJ i:u 705 1800 (Africa) 14 1 Benoil(I977) proximate) of islands and Easier I |PA) I65.fi 530 3760 (Chile) 2.5 ! Skoiubcrg ( 1920). Berland (1924) archipelagos Juan Fernandez. 1. (JF) 178 1500 660 (Chile) 10 | Sternberg 1954) considered. b), Saaristo 1978) Seychelles Arch. (SE) m 914 932(MaJaga>car) 65 Benoil (1978a. ( Berland Marquesas Arch. {MA I 1274 1 2(H) j 6.100 (Ten) | (1935, 1939l

Galapagos. Arch. (GA) 7KM2 I7i>7 960 (Ecuador) 9 Jackson ( tyfi5>. Bacn tt at []9&ya, M 1

i 1 Arch. (HA > 16504 yjxrs laooo (Mexico) TO Suman (1964 1965 i Gtv.u, \\ 197 (1nH5i | I C'l-mnrcvArch- | 2237 2300 300 (Madagascar) tocqifc

the Seychelles seem on average smaller than those from the collections made with pitfall traps in West Africa (Table 2). As only the difference

5 GA-RCJ is significant (x~ = 23.S4; i < 0.02) it seems that the discrepancies are further meaning- less. However, complete data from the islands and fully comparable data from the continent might reveal more differences in average si/e of island and continental spiders. We expect selection to favour small spiders during the colonisation process. Thus, many more small spider species are likely to be present on islands than on the continent. Smaller spiders live in higher densities than bigger ones and ihc possibility that they reach an island is statistically higher than for larger spiders. Moreover, smaller spiders arc usually better aeronauts although young stages of FIG. L Dendrogram of the similarities between the larger spiders have often been observed to bal-

: LndxAvrchipelagos on family IcveL loon as well. The chance that a raft-rafting being an important mean of colonisation-contains a and secondary forest in Ivory coast (Jocque, 'propagule' is also likely to be higher for smaller 1980), middle-size spiders in the class of 4-6mm spiders than for larger ones. dominate (Fig. 2). However, data for Ivory Coast However, the possibility for survival of spiders arc based entirely on pitfall trap samples. That is probably belter larger may bias the size frequency distribution towards that reach an island for than for smaller ones but the enormous differen- higher or lower averages. On the other hand, the ces in specific ecophysiological characters make- collections from islands are clearly biased it difficult to generalise. towards larger spiders. Earlier collectors are more likely to have taken conspicuous large The Seychelles are granitic continental frag- webspiders than small soil dwelling spiders A> ments and must have had a well-balanced fauna for the other analyses it is probably too early to when they separated from the mainland. Hence, detect relevant differences in size distributions on they should have elements of a continental spider different islands as compared to the situation on fauna. Apparently this is not so. Some unusually the continent. large soil-dwelling spiders do exist on the island Only spiders from Galapagos. Coeos Island and (see below), but the smallest size class seem more important on that archipelago, just as on the TABLE 2. Mean body length (mm), number of species others. The Seychelles are a group of small is-

and standard deviation ( RCJ = Ivory Coast ). lands with the largest not exceeding 15km". When ocean levels were higher these islands SH SE PA CO GA |MA IJA RCl JjF must have been tiny and the chance for extinction '. 6.75 5.57 6.45 4.79 5.52 7.63 6.40 682 very high. Larger spiders occur in smaller popula-

, n n 125 29 r36 31 Ul 127 l U tions. Thus, they arc more liable to become ex-

1 1 S ! 4.47 3.47 4.20 4.23 .1 4M 1 [5.511 3.35 9fl 3 91 tinct hence the otherwise unexpected dominance SPIDERS OFOCF.ANIC ISLANDS S44

id smaller spiders on the Seychelles. The spider species present (Fig. 3.2: r = 0.7l6, p>000l i !|1 fauna of Seychelles is probably too little known both cases there is a significant correlation. because the two main expeditions (Benoit 1978, S;taristo were in 1978} the dry season. Origin ok Faunas Frequency distribution histograms for the As the distance from an island from the main- smaller Pacific islands (Easter, Juan Fernandez land increases, its component of continental and Cocos) and for the Marquesas archipelago, species decreases. show very irregular shapes. The distribution Very striking is the much lower percentage of -/cms to have several peaks. continental species on the African islands as com- There may be several reasons for this: pared to those of the Pacific The same is true for 1 the fragmentary knowledge of their spider the southern Pacific islands Juan Fernandez and fauna. Few spider species are listed for these Easier Island This could be explained by their islands in the literature. Occasional samples, isolated position. Juan Fernandez lies in the Olid made by other xoologists during brief visit die of the northerly directed Humboldt Current, :*r Island, yielded important new records of whereas Easter Island js situated a long distance spider species That is apparently true of our from the continent in the centre of the South

knowledge of the spider fauna of most tropical Pacific Gyre, making ihc arrival by rafting ralln t islands. unlikely.

2 Cocos and Easter LLmd arc still 100 young or The presence on the Galapagos islands of are so isolated (Easter Island in the centre of the northern, central fwith Antillean elements), as South Pacific Gyre; Juan Fernandez amidst the well as southern American elements, could he northerly directed Humboldt Current) that they explained by their special situation astride the arc difficult to colonise and an equilibrium in equator. The archipelago lies in such a posi

their spider communities has not yet been that it rs reached by the warm NiJlo Current in ihc reached rainy season and by the cold Humboldt Current 3 due to their limited size and low habitat in the dry season. The Californian Cunent of the diversity these islands cannot support more Northern J lemisphere runs southwards along the species and do indeed have several size distribu- North American coast and reaches the Panama n peaks. Basil) where it is warmed up and turns tow the Galapagos as the Nino Current. The southern Humboldt Current runs northwards alon; Rf.latton between Island Area and 5h south American coast and turns westward near Density the equator towards the Galapagos. Rafts Tropical African oceanic islands are far richer easily be transported from the north as well as in species and their species/area relationship is from the south. A floating raft takes about two to much higher (0.48-0.81) than for tropical Pacific four weeks to reach the islands from the South oceanic islands (Tabic 3). The very small Cocos American mainland (Schatz, 1991). Island lying the closest to a continental landmass Furthermore, there was a broad connection be- js exceptional in this respect (0.60). Tropical tween the Caribbean region and the eastern Pacific archipelagos (0.01 -0.04) and the southern Pacific urea from 48 my until 3.5 -3 my ago pacific islands (0.14-0.18) have relatively low (Woodring, 1959; Jones and Hasson, 1965) with species/area relationships- It is not clear why the a sea current running from the Atlantic to the African offshore islands are so much richer than Pacific (Petuch.1982). The Panama isthmus was the others The source areas may be richer but that plugged some 3 my ago by the Caribbean Plate must be proven by thorough studies of the con- which was shoved in between the north and south tinental spider faunas. Many more species may American plates. At that time the Galapagos is- have been introduced on the African islands. This lands had already emerged from the sea. It is is probably true for Saint Helena that ha acceptable that many fauna elements of the

unusual large proportion (44%) of introduced Caribbean reached the Galapagos at that I spiders, mainly from temperate regions. It is Nearly 82% oflhe known spider species of Cocos doubtful whether that is b sufficient explanation. Island arc continental species There is a relationship between altitude The Marquesas have more species of Pacific

i I highest elevation of the island and the number origin. This can be expected because of the posi-

i:l species present i Rg, 3.1; r = 0.677, p> 0.01) tion of this archipelago at the matgin of the and between area of the island and the number of Polynesian province (NE end of the south ,

450 MEMOIRS OF THE QUEENSLAND MUSEUM

SIZE FREQUENCY DISTRIBUTION SIZE CLASSES COCOS SIZE CLASSES PASQUA IVORY COAST

ll.ll .

SIZE CLASS (mm)

SIZE CLASSES GALAPAGOS SIZE CLASSES SAINT HELENA SIZE CLASSES MARQUESAS

lllli.. .Willi. Jllli.

SIZE CLASSES SECHELLES SIZE CLASSES HAWAII SIZE CLASSES JUAN FERNANDEZ

IIIl . . iillii., . III. i.

FIG. 2. Frequency distribution histograms of total body length (over 2mm size classes) for each island or archipelago.

equatorial archipelagos chain). In contrast, been sampled only superficially or not at all. Few Hawaii and Galapagos have only a low percent- families have been thoroughly revised and new age of Polynesian species as the distances be- studies will be needed once the 'black holes' have tween them and the Polynesian archipelagos are been filled. For instance, the figures in table 2 for too extensive. Galapagos are based upon 70% of the total of the The term 'endemic' must be used now with recognised species. Hence, 46 species have yet to great care because the continental faunas are far be identified. At this stage in our knowledge one from being well known. This is especially so for is never sure that the species one describes from the neotropical spider fauna. Many regions have an oceanic island really has a dispersion restricted

3. spider fauna Continental Pacific Cosmopolitan TABLE The Island/Archipelago S S/area Endemic distribution origin species of islands and archipelagos. Percentage of species with Cocos I. (CO) 30 0.63 82% (A) 18% distribution 1° continental Galdpagos Arch. (GA) 146 0.02 32% (A) 4% 9% 60% (A=American; E=African; Hawaii Arch. (HA) 168 0.01 11% (A) 5% 19% 64% Au=Australian), 2° cos- Marquesas Arch. (MA) 48 0.04 3 ! % 20% 50% mopolitan and 3° species

Juan Fernandez I. (JF) 32 0.18 17% (A) 17% 70% known only from island or

Easter I. (PA) 23 0.14 22% (AU) 66% 12% 'Endemic'. Values based only described species. Seychelles Arch. (SE) 131 0.48 9% (E) 18% 65% on

St Helena I. (SH) 98 0.81 7% (E) 44% 45% SPIDERS OF OCEANIC ISLANDS 45

altitude/species relationship island attitude vs. Linyphiidae species log species number of species

jr D

i i ^- n 2.2

B 2 -

1,1 n n^ a i i^f °a 1.6 u^y n ^^ D njs M O Q. pa a y'w i D ma an

i i i ' 1 1 9 1 — i 2.2 2,4 2,6 2.1 3.2 3,4 3,6 log altitude attitude (m)

Area/species relationship relationship area/species Linyphiidae log species number of species 16

log area log area

FIG. 3. 1,2. Relationship between no. of spider species FIG. 4.1, 2. Relationship between no. of linyphiid (N) and: 1, altitude (highest elevation) of island or species (L): 1, and altitude (L = 0.0035 alt(m) + archipelago (logN = -0.6224 log alt.(m)); 2., island/ar- 1.0544; 2., and island/archipelago area (L = 3.7886 chipelago area (logN = 0.2686 log area + 1.0269). log area - 4.0344) (separate islands in archipelagos). (S = no. species). to that island for it may not yet have been found on the continent. We may therefore seriously Juan Fernandez is very striking. Though close to question whether we can validly use the propor- the mainland, it is a rather isolated island (Hum- tion of endemics for the analysis of an island boldt Current) with an environment rendered fauna. However, we consider that the percentage harsh by the very special climatological condi- of endemics we now recognise reflects the rough tions. proportion of real endemics that will eventually The proportion of cosmopolitan species on the be shown to exist. islands varies between and 20%, with two strik- Cocos Island seems to have few (18%) ing exceptions: Saint Helena with ca 44% and 'endemic' species. It lies relatively close to the Easter Island with 60% (this figure is based on continent. Once the distance to the continent ex- only 10 species). Most of these species are intro- ceeds 900km, we find an 'endemic' proportion in duced and restricted to human settlements. This between 50 and 65%. The high percentage of shows the great impact man has had on the spider 'endemics' on Hawaii (64%) is probably due to fauna of Saint Helena. its old age and thus the long period of isolation of All the islands considered here do have human the archipelago. The high percentage for the settlements, native on Hawaii, Easter Island and Galapagos islands (60%) may perhaps be due to Marquesas, but immigrated in historical times on the fragmentary knowledge of the South the others. Human settlements invariably result in American spider fauna (the main reason why one deforestation for arable land and subsequent in- third of the Galapagos spider fauna is not yet troduction of ubiquitous species. Most often this identified). is detrimental to the original fauna. Even more The high percentage (70%) of 'endemics* on damage can be done with the introduction of MEMOIRS OF THE QUEENSLAND MUSEUM

domesticated animals (cattle, dogs, cats, rodents, climatic conditions, the impression may exi rt that etc.) and cultivated plants Accidental introduc- lhey arc the determining factor) The presence of tions can also have serious consequences as for ;n, important number of Linyphiidae on high is- instance the Litile Fire Ant (WflOTWWW lands was already illustrated for the Con Q:*ropuitctaia)on Galapagos. This species haunts iJi>.t|ikl. 1985) This is particularly true for is- large areas, devastating nearly all living animal lands. There is a significant correlation between organisms. the altitude of island and the number of species

of Unyphiidae present (Fig. 4. 1 : r = 0.780, p< area or the island and the itative Analysis 0.001) and between number of linyphiid species present (Fig ^.2: r = Qualitative Analysis ol (he distribution of par- 0.6SS, rx 0.001). There is a third aspect relevant ticular groups on islands is perhaps moTe reveal- to the colonisation of islands: parapatric specla- than quantitative analyses which are tion on the spot The Lycosidae arc a good ex- hampered by the incompleteness df tuanv dati ample. Wolf spiders might be expec ted to be good gets. colonisers. Their juvenile stages arc active bal- Spiders that do not balloon do not seem well looners, and transport by rafting is also a likely adapted to colonise oceanic islands. Good ex- means of dispersal if only because lycosids fcrte amples arc seen in the Zodariidac. Only four ven- common on banks of rivers and in marshes species, if we include Crypt&thete, are present on from where rafts are supposed to be derived. the islands considered. On the Seychelles wc find However, the number of insular species is quite the endemic Cryptothelc allmudi Simon; the low compared to the high number of species in Comoros have Avecua radiasu Jocque on Grande continental tropical areas. (Galapagos: 6 species; Comore and Dforts telugaws locqui on Moheli Hawaii: 1 1; Cocos: 1; Juan Fernandez: 2; Saint (Jocqud, 1986). Zodarim trispinosum Suman is Helena: 7; Seychelles: 2: Comoros: 3 ). Moreover, known from Hawaii. The presence of a Zmi almost all species from islands in thus study have on Hawaii, completely outside its main distribu- no continental distribution and must be con- tion is pu/2ling. but is probably explained by an sidered endemics of each island. Only Bristowict- introduction as is the presence of Z. fuivonigrum lu is both (Simon) in North America. The Mygalomorphae seycheRensis (Bristowe) known from are found only on two of these archipelagos Six Seychelles and Comoros Their apparently high speciation related to this. species in three families (Benoit, 1978) occur on rate may be Of gpc the Seychelles Five of these are endemic, the note, some species at high altitudes on siwne sixth species. indrtidalis (Legendre and islands are apparently derived from species at low case thai Benoit). is also found on Madagascarand Grande altitudes, A well documented is ol 0ns- towiella the closely re- Comore, where it occurs together with the en- on Comoros where two demic MoggridgeQ ncsiota Griswold. The lated species have clearly differentiated habitats: presence of these spiders, all but one apparently RristnwieUa seyrheflensis living in shori gps.s.sy true endemics, on the Seychelles is explained by vegetation from sea level to about 3500m and its granitic nature, which implies that lhey have BristowieUa kartalensis Alderweireldt living in had a fauna from the moment they were separated recent lava flows with sparse vegetation from from the continent. The Comoros are probably about 600m upwards ( Alderweireldt, 1988). more easily colonised than other islands, possibly Another interesting case is that of the Hawaiian by rafting, as that archipelago is close to its source cave dwelling lycosids which have strongly area, Madagascar, for which the spider fauna is reduced eyes or none and which clearly speciated unfortunately poorly known. on the islands themselves (Gertsch, 1973). Linyphiidac can be considered excellent On Galapagos, a group of six species apparent- coloniser* mainlv because many frequently bal- ly derived from the most common one (*Lyc 3*). loon. However, Unyphiids appear to be able to It occurs over the whole archipelago and occurs occupy few habitats. Jocque 0984, 1985) ex- mainly in coastal salt marshes. Another species 5* plained that interference competition in tropical I 'Lye i occurs only on the low island Espanola lowland with ants is apparently too important to The four other species are confined to the high allow the presence of many iinyphiid species. As pampa zones of the volcanoes Sierra Negro and

( ants are less common at higher altitude many Cerro Azul ' Lye 1 \) on Isabella, on San Cristobal

2') ( more linyphiid species tend to be present in high- l 'Lye t on Santa Cruz *Lyc 4\i and the Alccdo land than at low elevations (e.g. Scharft, 1992). volcano on Isabella CLyc 6'). The revision of this (Since ant diversity and density are linked to remarkable species-group i< in preparation. SPIDERS OF OCEANIC ISLANDS 453

A fourth example of such segregated parapatric Introduction. Revue de Zooloeie Africaine 92: populations is found on Juan Fernandez where 2 390-404. Lycosa species are found, one living along the 1978b. Contributions a Fetude de la faune terrestre des ties granitiques de Farchipel des Sechelles. coast, the other living in the higher pampa zone. Araneae Orthognatha. Revue de Zoologie These statements reveal three important factors Africaine 92: 405-420. influencing the composition of island faunas. In BERLAND, L. 1924. Araignees de File de Piques et the first place, there is the accessibility of the des Ties Juan Fernandez. In 'The Natural History island, mainly its distance to a source area. In the of Juan Fernandez and Easter Island'. Vol. III. second place, the diversity of the island's habitats Zoology: 419-437. is important. Particular spiders such as 1935. Araignees des ties Marquises. In 'Marquesan mygalomorphs and zodariids are only present on Insccls-IF. Bemice P. Bishop Museum Bulletin those islands that are easily reached or already 114:39-70. 1939. Nouvelles araignees marquisiennes. had a fauna when they became isolated. Other In 'Mar- quesan Insects-IIF. Bernice P. Bishop Museum families, although good colonisers, appear to be Bulletin 142:35-63. restricted by the ecological conditions of the is- GERTSCH, W. 1973. The cavernicolous fauna of land they can reach. Speciation appears to be the Hawaiian lava lubes. 3. Araneae (Spiders). Pacific third factor which may be important in the Insects 15: 163-180. colonisation of habitats that can hardly be HOGUE, C. & MILLER, S. 1981. Entomofauna of reached by the normal ways of dispersal. The Cocos Island, Costa Rica. Atoll Research Bulletin effect of niche pressure (Jocque, 1982) is likely 250: 29pp. to be an important mechanism in this respect. JACKSON, M. 1985. 'Galdpagos, a natural history guide'. (The University of Galgary Press). CONCLUSION JOCQUE, R. 1980. 'Verspreidings-, aktivitcits- en groeipatronen bij spinnen (Arancida), met spe- ciale aandacht voor de arachnofauna van de The faunal composition we now find on many Kalmthoutse heide*. (Doctoraatsverhandeling islands is far from being natural. At the same time Rijksuniversiteit Gent). 181 pp. we know little about the arthropod fauna of most 1982. Niche pressure and the optimum exploitation oceanic islands. This makes the analysis and hypothesis. Biologisch Jaarboek Dodonaea 50: comparison of the faunas very hard. The in- 168-18L 1984. Considerations concernant Fabondance rela- fluence of speciation processes is probably large- tive des araignees errantes et des araignees a toile ly overlooked in connection with the vivant au niveau du sol. Revue Arachnologique compensation of extinction. 5: 193-204. 1985. Linyphiidae (Araneae) from the Comoro Is- LITERATURE CITED lands. Revue de Zoologie Africaine 99: 197-230. 1986. Ant-eating spiders from the Comoros ALDERWEIRELDT, M. 1988. On the genus Bris- (Araneae, Zodariidae). Revue de Zoologie towiella, with the description of B. kartaiensis n. Africaine 100: 307-312. sp. from the Comoro Islands (Araneae, JONES, D.S. & HASSON, P.F. 1965. History and Lycosidae). Bulletin of the British Arachnological development of the marine invertebrate faunas Society 7: 269-272. separated by the central american isthmus. Pp. BAERT, L, MAELFA1T, J.-P. & DESENDER, K. 325-356. In Stehli F.G. and Webb S.D. (eds). 'The 1989a. Results of the Belgian 1986 -expedition: Great American Biotic Interchange'. Araneae, and provisional checklist of the spiders MACARTHUR R.H. & WILSON, E.O. 1967. The of the Galapagos archipelago. Bulletin van het theory of island biogeography'. (Princeton Koninklijk Belgisch Instituut voor Natuur- University Press: new Jersey). 198pp. wetenschappen 58: 29-54. PETUCH, E.J. 1982. Paraprovincialism: remnants of 1989b. Results of the Belgian 1988-expedition to palcoprovincal bounderies in recent marine mol- the Galapagos islands: Araneae. Bulletin van het luscan provinces. Proceedings of the Biological Koninklijk Belgisch Instituut voor Natuur- Society of Washington 95: 774-780. wetenschappen 59: 5-22. SAARISTO, M. 1978. Spiders (Arachnida, Araneae) BENOIT, P. 1977. Araneae. Introduction. In 'La Faune from the Seychelle Islands, with notes on tenrestre de File de Sainte-Helene. Quatrieme taxonomy. Annales Zoologici Fennici 15:99-126. partie'. Annalen van het Koninklijk Museum voor SCHARFF, N. 1992. The Linyphiid fauna of Eastern Midden-Afrika, Zoologische Wetenschappen Africa (Araneae, Linyphiidae). Distribution, pat- 220: 12-188. terns, diversity and endemism. Biological Journal 1978a. Contributions a Fetude de la faune terrcstre of the Linnean Society 45: 117-154. des lies granitiques de Farchipel des Sechelles. SCHATZ, H. 1991. Arrival and establishment of Acari 454 MEMOIRS OF THE QUEENSLAND MUSEUM

on oceanic islands. Pp. 613-618. In Dusbabek F. SUMAN, T. 1964. Spiders of the Hawaiian Islands: andBukva, V. (eds). 'Modern Acarology'. Vol.2. Catalog and Bibliography. Pacific Insects 6: 665- (Academia, Prague and SPB Academic Publish- 687. ing: The Hague). 1965. Spiders of the Family Oonopidae in Hawaii. SKOTTSBERG, C. 1920. Notes on a visit to Easter Pacific Insects 7: 225-242. Island. Pp. 2-20. In 'The Natural History of Juan WOODRING, W.P. 1959. Geology and paleontology Fernandez and Easter Island'. Vol. 1. of canal Zone and adjoining parts of Panama. 1954. A geographical sketch of the Juan Fernandez Description of Tertiary mollusks (Gastropoda: Islands. Pp. 89-192. In 'The Natural History of Vermetidae to Thaididae). United States Geologi- Juan Fernandez and Easter Island'. Vol. 1. cal Survey Professional Paper 306: 147-240. THE REPRODUCTIVE ECOLOGY OF EUSCORPIUS FLAVICAUDIS M ENGLAND

T.G. BENTON

Benton, T.G. 1993 1111: The reproductive ecology of Euscorpiusflavicaudis in England. Memoirs of the Queensland Museum tt(2): 455-460. Brisbane. ISSN 13079-8835.

The reproductive ecology of the scorpion Euscorpius flavicaudis was studied both in the field in England (at a colony dating back 120yrs), and under semi-natural conditions in the laboratory. Before the mating season males become vagrant to search for females. On

, encountering a pregnant female or one with young, in her 'burrow ' ihe male may mate-guard liei until her period of maternal care ends and she becomes receptive to him. Large males have a higher mating success than small males. Two instars of adult males exist in this population of scorpions: large 7ih instars and small 6th instars. and the reasons for this appnrently paradoxical situation are explored. QEu.\<-y/rpius, matc-guurding, mating, life- ktSWl v

T.G. Benton, Department ifZntdogy, University of Cambridge, Dawning St.. Cuwhr'tdgt., CB2 3EJ, United Kingdom; present address: School of Environmental Sciences. University of East Anglia, Norwich, NR4 7TJ, United Kingdom; 7 December, J 992.

Evolutionary and behavioural ecology are rela- associated with buildings (Wanless, 1977). tively young disciplines; and have to date con- Colonies have been reported from several loca- centrated on using the more 'familiar' animals as tions in England, but only one, at Sheemess , subjects: birds, mammals and insects. These dis- Docks, Kent (51°26'N, 0°45 E) has lasted munv ciplines have largely passed arachnids by. In- years and is of considerable size. The first record creasingly, however, we are beginning to realise Of scorpions at Sheemess is a label on a specimen the utility of arachnids as model animals in help- of Euscorpius flavicaudis in the Natural History ing to elucidate general evolutionary principles. Museum, London: Taken in Sheemess Dock-

Although there has been considerable increase yard, 1870, J.J. Walker' (P.D. Hillyard, in lit.). in behavioural ecological interest in spiders in recent years, scorpions have been largely ig- METHODS nored. This is perhaps because they are difficult animals with which to work - long-lived, inac- Field Observations tive, elusive and nocturnal. This paper outlines At Sheemess Docks, most present buildings some of the main findings of a study of scorpion and the perimeter wall date from 1823 (Benton, reproductive ecology. It is unfortunately incom- 1992a). Scorpions arc common in areas little plete: posing more questions than it solves; bill, 1 frequented by people; and arc especially common hope that it illustrates the fact that, as in so many living in the perimeter wall. This old wall is about areas of their biology, scorpions arc 4m high and 66cm thick. It shows the effects of idiosyncratic, fascinating and more complex than time, with the mortar between the bricks crum- they may first appear. This study combines the bling away, creating cracks' whichthescorp power of controlled laboratory experiments and readily inhabit. My study area was a 104m length the ease of laboratory observations with ihe of perimeter wall (bounding ornamental and reality of what is observed in the field, and, by so vegetable gardens). The field study was OOfl doing, for the first time the events surrounding the ducted from January I9S8 to July 1989, during courtship of a scorpion are explored and placed which 92 visits were made to Sheemess Docks in an ecological context. (totaling 500 hours of observations). Scorp were observed using a portable ultra-violet light. THE STUDY ANIMAL Benton (1992a) details field methods

Euscorpius flavicaudis (de Geer), common Laboratory Observations throughout southern Europe (Birul;i. 19)7), also To allow more complete observations and ev occurs in England, most probably ;t> a icsuit of perimentation a laboratory colony of Euscorpius introductions This species is a successful was set up, modelled on the habitat at Sheemess. colonising species (Fage, 1928; Wan less,. 1977; This consisted of a plywood wall (2.44m x Lourcnco & Vachon, 1981) and is often found 5.79m). covered in sandpaper, with 144 slots (2.5 4*i MEMOIRS OF THE QUEENSLAND MUSEUM

,i 5, km) cut into the face. Into these slots were seen. The numbers build up during the spring and filled artificial "cracks* 12 cm long. These were peak in late summer (Benton, 1992a) before open to the wall and closed at the far end. A layer decreasing again to winter levels. As these scor- of soil was placed at Ihe base of the wal I . Th is soil pions live for several years (Benton. 1991b) this was periodically damped to mimic rain, and each pattern reflects varying patterns of behaviour, week woodliec (the scorpion's natural food) were rather than gross mortality and natality. Ft* rims' reJcased onto it. This colony was kepi under a of the year, most scorpions seen are in their crack natural photopcriod. and as near natural tempera- (for November to June only 8 ± 1 1% of all scor- ture and humidity as possible. Observations and pions observed are out of their cracks). Hov. experiments were conducted at night, when the in summer, on average, 24 ±9% of scorpions wall was illuminated by lamps fitted with deep- observed were on the wall surface. Adult males red filters (invisible to scorpions: Machan, 1968). make up most surface-active animals 3-4x more Unless being used inan experiment, the scorpions males are seen on the wall surface than adult (collected at Sbecrncss) were allowed to move females over the summer (Figs la, b) freely around the wall Benton (1991*0 details The number of females giving birth or with colony setup. young peaks when female activity on (he waJl surface (Fig. la) is least. Females with young

i Hi' 'lis EXPI I Mi remain deep within their cracks, and were: never In this study of male contest behaviour for observed to venture onto the wall surface Indeed, females during the mating season, 1 1 males were females in the last singes of pregnancy also ap- used On the artificial wall 40 x 40cm enclosures peared to show a marked reduction in activity. were made around a single crack A female car- Most matings at Shccrness (Fig. lb) are just rying young was placed in this crack, and later an after the number of females with young peaks experimental male was added. Thirty minutes (Fig. 1 a). Known females with young were found alter this male had encountered and entered the to be unreceptive to males until 2 ±1.2 days crack After encounter- a second male was added. (n=l2) after the juveniles had moulted and ing the crack the second male would attempt to climbed off her dorsum (Benton, 1992b* enter. The males would then meet, assess each About 3 weeks otter male surface-activity other and fight. This procedure was then imme- begins, the number of matings observed peaks diately repeated. The two replicates constituted a (Fig. lb). Only at this time can males andfen contest' (and ui all cases the results were unam- be found together in the same crack. These biguous, and identical for each replicate). In total periods of cohabitation averaged 10±9.9 nights 1 10 contests were conducted (with each male as

(range I -30; n=19: data from field). Cohabiting 'owner' against every other male as intruder). occurred only prior to mating, and whilst the Contest order was randomly selected. Other female was either carry ing young, or in late preg- methods are detailed in Benton (1992b). and scor- nancy (Benton, 1992b). Males were aSSOCi pion mensuration in Benton (1991bi sometimes with females living in shallow cracks RESULTS 1 it=3)i in which case the males spent the day in a deep crack nearby and 'commuted* to the female's crack to spend the night sitting at the Background entrance. The identity of the male mating with a Like mo^i scorpions, &, Jluviawdis is noctur- specific female was known for 15 cases of nal. During the day and over the winter, they are cohabitation: in 14 the male had been recently not visible as they have retreated to the back of ci "habiting (P<0.00:o with that female. Females their cracks in the wall. At night, depending on of this species do not appear to assess males, and the season, most scorpions evident are at the will mate with any male present when they be- entrance to their cracks where they remain im- come receptive (Benton, 1992b). If a male's en- mobile, with their claws outstretched and open, counter rale with females is low, and a female's The most common prey items are isopods {Por- tec Cptivity is predictable, then a male encounter- i ( vUiosaiher){hA% ) and then COnspOCffics I _ ing a soon-to-be receptive female may do better (Benton. l9Wa) waiting for her to become receptive, rather than leaving her and risking trying to find a female Seasons nearer receptivity (Grafen and Ridley. 1983); and Scorpion activity at Shccrncss is highly this seems to be the case with these scoipions seasonal. Over the winter very few (if any) can be (Benton, 1992b). This behaviour can also benefit 1

REPRODUCTIVE ECOLOGY OF ELSOORPtL'S FLAWCAUDIS 457

To investigate male mate-guarding contests 15 1 Fig 1a 1 10 contests were staged between 1 1 males (each male as 'owner' against each of the others as 'intruder"). The proportion of contests won correlated very strongly with measures of scorpions size, and most strongly with pedipalpal claw

length frs=0.97 t P<0.001). This is to be expected as £ flaxicaiidis uses its claws as its main offen- sive weapon. In 80% of contests, the larger- clawed male won: when the srnallcr-clawed male- 1 won (2(r%), it was the "owner in most (91%) contests. Both relative claw -size and ownership -,.,-, r •; i"i 4ft n 1 status had highly significant effects on contest Days outcome (two-way size: F=68.5; di 15- ANOVA, FTfl 1b J ,36; P

P<0.000l ): iarge-clawed males usually won contests, but if contestants were closely matched in 10-J size then ownership status decided the outcome (giving an advantage equivalent to about 11% longer claws).

Size and Sexual Advantage: In the laboratory, males differed in reproductive success: some males mated twice, someone*: 4C0 and some not at all. Larger males mated more often than small males (Fig. 2). This size advantage is for two reasons. Firstly, as described above, larger males are better competitors for PIG. 1 Reproductive season of Kuscorpiusjlavicaiidis in England, 1988. (a) Seasonality of adult female female-occupied cracks. Secondly, about 40 surface-activity and no. of births. Female surface-ac- matings are not preceded by mate-guarding. tivity is smoothed over two weeks: nightly count These fall into three broad categories: non-preg- shown as average ofpreceedingand following week's nant females at the start of summer, females n-ot counts. births (for each half-month) estimated No. of found by males before their maternal care has from no. of females giving maternal care, and date at ended and (most rarely) females who have mated which this ended, (b) Seasonality of adult male sur- already (this is possible as a spermatocleutrum is face-activity (smoothed over two-weeks), and half- not secreted in this species). Large males obtain monthly count of no. of matings (observed no. of matings plus no. of spermatophorcs found). Data more matings from all three categories than small originally, in part, in Benton 1992b. males. This is because, for each category of mating, females are initially unwilling to o the females for two reasons. Firstly, itensures that When a female is mate-guarded prior to her they are mated. Secondly, a major advantage of receptivity, she encounters the male frequently, and upon becoming receptive begins courting maternal care in scorpions is to prevent predation without aggression. Conversely, when a male of the juveniles (Benton, 1991a), so having a male encounters a non-pregnant female, or one without at the front of the crack prevents entrance by other young he immediately attempts to court. The scorpions which may aid this role. male grabs (or attempts to grab) the female's In all cases of cohabitation, the male stations claws and stings her (Benton, 1990). Indeed, the himself near the entrance, with the female behind start of a courtship {without mate-guarding) is Males attempting to enter the crack en- him- indistinguishable from a cannibalistic attack. counter the currently cohabiting males, and each Large males are better at mating with these un- tries to grasp the other's claws. Within a very few willing females as, unusually for scorpions, large seconds one male retreats and tlees from the males can be larger than females. In this specie. crack. Males therefore tight for 'possession' of a and others (Polis, 1980), size (especially clav. crack occupied by a female in the period before size) is a good predictor of the outcome of can- she becomes receptive. nibalistic contests. At the start of courtship, males 458 MEMOIRS OF THE QUEENSLAND MUSEUM

trie hobo thria lmm -0.5 0.5 1.0 2.C

Mating Success

FIG. 2. Relationship between size (claw-length) and mating success in laboratory males free on wall surface (mean ± SD). Difference between groups is significant (Kruskal-Wallis, H=l 1.4, P=0.003). sting females, and larger males can eat adult females (and often do: 10 cases observed in the laboratory), therefore, large males essentially give females the choice of being cannibalised or lmm accepting courtship. Small males obtain fewer matings as they have a smaller (or nonexistant) size-advantage over females (Fig. 3) and therefore are more likely to flee from those females who are not immediately willing to court.

Size Dimorphism Sexually mature adult males are recognised by FIG. 3. Claws of two adult 8 6 (a, b) and one adult 9 the secondary sexual characteristic of a notch in (c) same scale. Note secondary sexual characteristic: notch in fingers. the pedipalpal fingers (Fig. 3). Adult males can differ markedly in size, and this is especially been reported in scorpions, but is known to occur noticeable when one considers claw size, which in some spiders (e.g. Vollrath, 1980; Watson, increases allometrically with body size, such that 1990) and occurs widely across the animal larger males have disproportionately larger claws kingdom (Ridley, 1983). For precopulatory mate- (Fig. 3). The frequency distribution of adult male guarding to evolve, males must gain an advantage claw sizes at Sheerness was dimorphic (Fig. 4). This dimorphism also occurs with other measures by staying with a female prior to her becoming of size, such as prosoma length (Benton, 1991b). receptive rather than searching for a receptive This dimorphism probaly arises because there are female. Two criteria may determine this situa- two instars of adult male in the population at tion: firstly, if a male can predict when a female Sheerness: 6th and 7th (see also Benton, 1991b). is going to become receptive, and, secondly if receptive females are difficult to find. The former DISCUSSION may occur if females become receptive following a moult (e.g. Watson, 1990), or as in the case of Two points are noteworthy. Firstly, the these scorpions, after maternal care. The latter criterion may result from female receptivity reproductive ecology of Euscorpiusflavicaudis is much more complex than previously imagined. being very limited in time (e.g., in Gammarus Naively placing pairs of scorpions together in a female receptivity is limited to a brief period laboratory situation to watch the courtship would between moulting and the hardening of the present a very misleading picture because the amphipod's exoskeleton: Grafen and Ridley, most significant behaviours occur before actual 1983), low population densities (and so low en- courtship. Precopulatory mate-guarding has not counter rates) or because of high male mortality REPRODUCTlVE ECOLOGY OF EUSCORPIUS FLAVICALTMS 459

is equal. This ESS could be maintained by frequency dependent selection (Benton, 1992b): large males do relatively better w ben rare (as they

suffer little competition) so their frequency is increased in the population by natural selection. However, when large males are common their gains are reduced, because of increased competition between them, to a point where they are

offset by the cost of delaying maturity. Hence, %[ i evolutionary equilibrium is reached between the 3 18 8 IS 8 ft a t> t> ^ C3 r— genes controlling maturation at the 6th and 7th instars. en £ i * * 6 Secondly, an alternative explanation for this C law length f/nm) paradox is that there may be phenoty pic plasticity in the maturation stage gene that can produce FIG 4. Distribution ofelaw-si7.es in adult 6 3 is sig- A nificantly non-normal. Filled bars arc observed dis- a variety of phenotypes depending on the tribution; stippled bars are expected normal vironment is said to exhibit phenotypic plasticity distribution given mean and standard deviation of (Lessels, 1991). The 'optimal* time at which to 2 data.(x =2U.P<0.005>. mature may depend on factors such as food availability, and thus growth rate (Stearns ai»d during mate -searching (Vollrath, 1980). The Koella, 1986). For example, if food is plentiful,

period of female receptivity is reduced where the and an individual is large for its cohort it may be pdteitiit]/ Of a male is ensured by being ihe first best to mature early. Conversely, if an individual mule to mate with a female (so called first-male is small for its cohort then it may be better to delay sperm priority), which may be common in arach- maturity until more food is encountered. This nids (e.g. Vollrath, 1980; Watson, 1990). A virgin seems to be the case in manv spiders (e.g., female is far more valuable to a male than one Deevey, 1947: Vollrath. 1980). this subject will which is already mated. Thus, Euscorpius be discussed further elsewhere (Benton, in fhsvicaudis, first-male sperm-prinrity and low prepjration), population densities, coupled with the female's receptivity being predictable, may make it more LITERATURE CITED profitable for a male encountering a female engaged in maternal care lo mate-guard her rather BENTON. T.G. 1990. 'The behaviour and ecology of than risk searching for another female nearer scorpions.'' (Unpublished Ph.D. thesis. University receptivity of Cambridge) 1991a. Reproduction and parental care in the soot- Secondly, there seems to be a paradox. Adult pion. Euscorj)iu.\rflax icuudis. Behaviour 1 17: 20- males exist as two instars in the population. This 28.

itself is not a novel finding (sec Francke and 1991b. The lilt* history Ol littscorpiux Jlav'icaudix es, 1982)' M IS surprising that any males ma- (Scorpioncs.Chaetidac). Journal of Araehnology ture at the smaller instar since the smaller males 19: 105-110. seem to be at such a disadvantage in obtaining 1992a. Determinants of male mating success in a mates. Natural selection would quickly eliminate scorpion. Animal Behaviour 43: 125-135 !992b. The ecology Euscorpius flavicat-u any tendency to mature at a disadvantageous size. of England. Journal of Zoology, London 226: 331- The paradox suggests two explanations. First, 368. large males may have an advantage only in a BIRULA, A.A.B. 1917. 'Scorpions: Fauna of Russia mating season. As large males have more instars, and adjacent countries; Arachnoidea, Volume 1: tiiL> probably take longer to mature. Although Scorpioncs.' (Petrograd: translated from the Rus- sian by the Israeli Scientific males mate more in the short-term, the younger Programme of Trans- lation, Jerusalem, 1965). they mature presumably the more seasons in DEEVEY, G.B. 1949. The developmental history of which to find mates. Perhaps there is a mixed Latrodecius nutcTans (Fabr.) at different rates of evolutionarily stable strategy such that the CESS) feeding. American Midland Naturalist 42: 189- short-term gains enjoyed by 'large males* are 219. offset by the losses due to having one less season GRAFEN, A. & RIDLEY, M. 19K3. A model ol man in which to mate, so that, on average, the lifetime guarding. Journal of Theoretical Biology 102 reproductive success of Marge* and 'small* males 460 MEMOIRS OF THE QUEENSLAND MUSEUM

FAGE, L. 1 928. Remarques sur la dispersion en France POLIS, G.A. 1980. The significance of cannibalism on

et 1' acclimation en France de V "Euscorpius the demography and activity in a natural popula- flavicaudis" (De Geer). Association Francais pour tion of desert scorpions. Behavioural Ecology and 1'avancement des sciences, La Rochelle 1928: Sociobiology 7: 25-35. 650-652. RIDLEY, M. 1983. 'The explanation of organic diver- FRANCKE, O.F. & JONES, S.K. 1982. The life history sity.* (Clarendon Press: Oxford). of Centruroides gracilis (Latreille). Journal of Arachnology 9: 223-239. STEARNS, S.C. & KOELLA, J.C. 1986. The evolution LOURENCO, W.R. & VACHON, M. 1981. Comple- of phenotypic plasticity in life-history traits: ments a la description d'Acanthrothraustes predictions of reaction norms for age and size at brasiliensis (Mello-Leitao, 1931) (=Teuthraustes maturity. Evolution 40: 893-913.

brasiliensis Mello-Leitao, 1931), synonyme d'- VOLLRATH, F. 1 980. Male body size and fitness in the Euscorpiusflavicaudis (Geer, 1778) (Scorpiones, web-building spider Nephila clavipes. Zeitschrift Chactidae). Journal of Arachnology 9: 223-228. fur Tierspsychologie 53: 61-78. LESSELS, C.M. 1991. The evolution of life histories. WANLESS, F.R. 1977. On the occurrence of the scor- Pp. 32-68. In Krebs, J.R. and Davies, N.B. (eds.) pion Euscorpius (DeGecr) at Sheer- 'Behavioural Ecology: an evolutionary flavicaudis ness Port, Isle of Shcepey, Kent. Bulletin of the approach'. 3rd Edition. (Blackwell Scientific Pub- British Arachnological Society 4: 74-76. lications: Oxford). MACHAN, L. 1968. Spectral sensitivity of scorpion WATSON, P. J. 1990. Female-enhanced malecompeti- eyes and the possible role of shielding pigment tion determines the first male and principle sire in effect. Journal of Experimental Biology 49: 95- the spider Unyphia litigiosa. Behavioural Ecol- 105. ogy and Sociobiology 26: 77-90. COMPARATIVE POSTEMBRYONIC DEVELOPMENT OF ARACHNIDS

ALAIN CANARD and ROLAND STOCKMANN

Canard, A. and Stockman, R. 1993 11 11: Comparative postembryonic development of arachnids. Memoirs ofthe Queensland Museum 33(2): 461-468. Brisbane. ISSN 0079-8835.

A common model is used to describe the growth of various arachnid groups. In these predators there is retardation of development of the first instars, along with greater maternal care for the clutch and even, as far as some groups are concerned, in a viviparous development. Mites, which have very diversified biologies, have developments which have evolved in many different ways. Les developpements des differents groupes d'arachnides sont decrits en suivant une trame commune. II apparait ainsi une Evolution des groupes de predateurs qui se traduit par une augmentation du retard de developpement des premiers stades, en liaison avec des soins a la ponte croissants, avec meme pour certains groupes des developpements vivipares. Les Acariens, de biologies tres diverses, ont en consequence des developpements qui ont evolue* dans des voies tres differentes. [^Development, arachnids, growth.

Alain Canard, Laboratoire deZoologie et d Ecophysiologie, Universite de Rennes 1, France; Roland Stockmann, Laboratoire de Phxsiologie des Insectes, Universite de Paris VI, France; 19 March, J 993.

Analysis of developmental types used in arach- embryonic, starts with the first divisions of the nids is very great diverse in the terms used to egg and continues with the formation of tissues. describe these phenomena. This diversity tends to After hatching or birth, the development is partially hide some similar points in the develop- qualified as postembryonic. The postembryonic mental process. Authors often use terminology organism is then covered with an external integu- and analyse specific to a single group or species, ment and develops outside of the egg membrane instead of referring to general concepts relating or the female's genital tract. Various organs ap- to all arthropods. pear and develop, some will not be functional This study compares the developmental until late developmental stages (e.g. genital or- processes using the same terms, and we will only gans). Externally, the development is shown by use specialised terms if it is absolutely necessary. changes in the cuticle. Reiteration of this concept All developments cannot be discussed in detail may seem pointless, but the study of arachnids here; definitions and more precise descriptions requires some explanations about hatching and can be found in Canard and Stockmann (1992). birth.

Hatching, i.e. the opening and release from the METHODS egg's membrane, can be a long process. It may take a few days for some spiders, and sometimes Our study is based mainly on literature, en- some moults can be observed between the begin- riched with our observations on arachnid growth, ning of the opening of the egg's membranes and particularly on spiders and scorpions. their complete liberation. Hence, the post- First, we will evoke the main concepts and embryonic period does start with the opening of definitions used and will then define the different the egg's membranes. This does not make it scales of development for each taxon. Taxa which necessary to look for a phenomenon before hatch- are exclusively predatory are here represented by ing as, for example, Legendre ( 1 958) and Vachon to the increasing level of care devoted to the (1 958b) did as they chose the 'inversion' to define clutch. Mites, which have diverse biologies, will the beginning of the postembryonic period. be studied separately later. Moreover, this proposition has one drawback: it makes the postembryonic development begin at RESULTS a time when the organism is not covered by a tegument. Concepts and Definitions Birth appears as a well and easily defined phenomenon, without ambiguity. However, in postembryonic development, hatching and birth pseudoscorpions, the organism leaves the The development, defined as initially female's genital tract and moves into a ventral ,

462 MEMOIRS OF THE QUEENSLAND MUSEUM

, free tile, re pn>y lite on yolk twdmg rtdh ifw mottor

fmchmg

Jffl. " T .' v raJ h* "* 4 _ i^^tg^i^^fig^^y,i egg \JU >p7[>fj3ltp4li Jfi *s 37 ^108 *B -r FIG. 2. Postembryonic development scheme of a soJ- pugid, Eremobates ditrangoruts (after Murna, 1966). PIG- I. Stages of postembryonic development of an opilionid, Ltobunum rotundum (Phaiangidae, Pal- DIFFERENT DEVELOPMENTAL ROUTES

patores) (after Gucutal, 1 943 and Naisse, 1 959). (m = moult, Ji = incompleted juvenile, J = juvenile), Opilionids The eggs are laid isolated from each other and brood pouch ? which is like an external extension of the female's genital tract. Once in the pouch, then abandoned. The emergent animal depends development continues and the organism Mil] on its yolk reserves it looks like a juvenile har- receives nutritive fluid from the mother. There- vestman but morphologically differs from fol- fore, the start of the postembryonic development lowing instars by the unpigmented integument pf pseudoscorpions should be when they leave and lack of some characteristics, such as un- the pouch, rather than when (hey leave the formed chcliccrac, unsegmented tarsus, absence female's genital tract, as others have stated. of median eyes. etc. Therefore it is an incomplete luvemle: Jii (term taken from Holm, 1940 con- (= 'lar idmgtoJiihcithiL. 1NSTARSANDS! '..SKS cerning spiders) 1965). It also has temporary organs (one or two The i nstar is the organism between two moults. Generally, among arthropods, Ihc external form egg teeth located on dorsum of cephaknhorax). does not significantly change between tbt>e (wo After the first moult, the animal differs from the stages, except for the short periods of pre- and imago only by its size and by sexual eharac- post-exuviation. However, among some mites, teristics. It is a juvenile, the second of the juvenile the external morphology and biology are phase J: (~ 'nymphc' according to Juberthie. modified while the cuticle remains. Hence, Herik- 1965). Although active, it still lives on its yolk ing (1882) distinguished two forms, called insUirs reserves, After one moult, the juvenile opilionid?- ('stade"), one active and the other motionless. (J3) scatter and then feed on prey they catch. Although this use of the term 'instar* was fol- Usually, 6-7 iuvemle instars occur before lowed only by a few authors to describe the comes an imago, more rarely there are 5 to 8. The development of few mite?., another acarologist number of moulNnuy vary from with individuals (Grandjean, 1938), considered that the term was of a species, and according to the developmental too indefinite and proposed the term 'stase'. The conditions; this number is generally the same fur definition of stase changed later, but this term is both sexes. the basis for an evolutionary concept (Grandje;m On becoming a breeding instar (irnagos) the 1954; Andre and Jocque, 1986; Andre, 1989). opilionid cease mi.ulltng (adults), although some 5-fi (Juberthie, We use 'instar' here in its usual definition, may live for be > bars 1965). which means an organism between two moults (for endocrinal considerations, see Canard and SOUR.IGIUS Stockmann, 1992). To keep the general defini- The female isolates herself in burrow hut does tion, we will discuss the concept of stase later. give care to her eggs. When the animal presents separate biologies The animal which hatches is incomplete, mo- linked with two different aspects during the same tionless and lives on its yolk reserves. It is unpig- instar, we call them 'forms' and give them two mented, has no eyes and no racquet organs. Two different names. types may be distinguished. The first is in the Galcodidac (Vachon, 1958a, Junqua, 1966). It

Succf-ssion i >i- Jnstars docs not look like the imago and keeps the aspect We limit our s.tudy to a mnrph<>-bio logical it had while under the constraint of the description of the successive instars. These in- membranes of the egg; it is a foetal instar F [- stars can follow each other b phase (Vachon, Marve* of Vachon, 1958a or Junqua, 1966, = 1953) in which all instars are of same kind. These i-embryo" of Murna, 1966). The second type different types are defined in Table 1 esent in some Solpugidae (Lawrence, lw7) s

COMPARATIVE POSTEMBRYONIC DEVELOPMENT OF ARACHNIDS 463

free life, feeding on prey life on yolk free We, feeding on prey rnS m9 mtO with the mother

' 9 wt mi ns W V -' Nk ® FIG. 3. Scheme of the different stages of the postembryonic development of an amblypygid. Taran- tula marginemaculata (after Weygoldt, 1970). FIG. 4. Scheme of the different stages of the pos- and some Karschiidae (Thaler, 1982). It looks tembryonic development of Typopeltis stimpsonii like the imagos (juvenile), but lacks some struc- (Thelyphonidae) (after Yoshikura, 1965). tures and has temporary organs like thorny con- adapted to this life, so with reference to Scor- tinuations of the legs. It is an incomplete juvenile: pions, we call it the pullus: JP (= 'embryon' of Jii (= 'primarlarve' of Thaler, 1982). Pereyaslawzewa 1901, and 'praenymphe' of further moult results in an animal which has A Weygoldt 1970). all the adult organs, except those linked with The pullus moult while on the mother and then reproduction. This juvenile is, according to the leaves her. Through this moult it acquires all the species, the first or the second of the phase (= characteristics of the adult (except the reproduc- 'nymphe' of Vachon, 1958a; Junqua, 1966; tive organs). It is a juvenile instar: J2. After a short Muma, 1 966). It remains with the female until its gregarious period, the juveniles scatter and live integument is hard enough, then it disperses and on their own, feeding themselves. The number of lives on its own, feeding itself on prey it catches. juvenile instars may vary between individuals The number of instars before reaching the status (Weygoldt, 1970) and, after one year, they be- of imago varies according to the individuals. The come an imago. imagos have a short life expectancy comprising The imagos of both sexes go on moulting and, only one instar (adult). under good conditions, the females keep growing after this moult. Therefore, there are no 'adults'. Amblypygids The female carries the eggs under its abdomen Uropygjds in a brood pouch generated by the genital tract The female isolates herself in a burrow, and during egg-laying. The incubation period of the lays her eggs in a newly secreted transparent clutch may last 3 months (Weygoldt, 1970). ventral sac. Hatching takes place in the brood pouch. Hatching occurs in the brood pouch. It cor- The organism released from the egg's responds to a quasi-simultaneous release of the membranes has a foetal aspect with the prosoma egg's membranes and of the integument of an bent towards the abdomen and its appendages instar similar to the foetal instar of the tight along the body. It is very incomplete (ap- Amblypygids: F (= 'primarlarve' of Kastner, pendages incompletely segmented, absence of 1949 and 'prelarva' of Yoshikura, 1965). This setae and of sensory organs, etc.), motionless and instar, which was already formed in the egg, has lives on its yolk. It is a foetal instar: F (= a very short postcmbryonic life. 'deutembryo' of Weygoldt 1970). The animal released after hatching and after the After one moult, the animals leave the brood first moult, can move and climb upon the pouch and attach themselves under the females' mother's back. In general morphology, it looks abdomen. They still live on their yolk, can move, like the adult (juvenile phase), but lacks some and look like imagos (juvenile), but some struc- organs (median and lateral eyes are not yet tures are lacking and the internal organisation is visible, flagellum unsegmented, etc.). It has par- incomplete (digestive tract, circulatory system, ticular organs linked with its life on the mother, etc.). It is an incomplete juvenile instar which has including pad-like organs at the tip of the legs, particular organs on the legs, a dorsal continua- instead of claws. It is a pullus: JP (= tion at the distal end of the tibiae (Weygoldt, \sekondarIarve' of Kastner, 1949; 'larva' of 1970) and, in some families, an adhesive organ at Yoshikura, 1965). After a 'diapause', there is a the tip of the tarsus. It lives on the mother and is moult which releases a juvenile instar: J2 (= ,

464 MEMOIRS OF THE QUEENSLAND MUSEUM

Forms Characteristics Symbols

embryonal aspect unsegmented foetal instar different motionless instars which do not feed appendages morphology from themselves imagos segmented appendages nymph

active instar larva Non-breeding incomplete no temporary organs juvenile same morphology as several organs which do not function temporary organs linked to life imagos pullus on mother

only non-functional genital organs juvenile

Breeding (= imagos) with moulting imago

TABLE 1. Characteristics of different kinds of instars.

life on yolk free life, feeding on prey

FIG. 5. Scheme of postembryonic development of Prokoenenia wheeleri (after Rucker, 1903).

'pullus' of Kastner, 1949; = 'protonymph' of Yoshikura, 1965). They are still gregarious but move around the burrow and begin to feed themselves. After one winter spent together, they then moult and scatter. Subsequent moults usually occur annually, but may be less frequent (Yoshikura, 1965). FIG. 6. Postembryonic development in two spider that (a) abandons its clutch (Larinioides cor- Imagos do not seem to moult (adults). Their species nutus) (after Ysnel, 1992), and (b) cares for young size of each species does not vary much. (Philaeus chrysops) (after Bonnet, 1933; Canard, 1984). Palpigradids Others keep their clutch with them in their silk The sexual biology of palpigradids is almost chamber (e.g. salticids). unknown. Moreover, nobody has ever succeeded Hatching occurs in the cocoon and it some- in breeding palpigradids. Therefore, knowledge times takes a few hours before the first instar is about their postembryonic development is based released. This instar is generally foetal: F (= on observations of natural populations. Under 'prelarve' of Vachon, 1958b, = 'pullus' of these conditions three immature instars have been Canard, 1984). Among some orthognathids it determined for many species (Conde, 1984). remains intrachorional and is therefore not pos- The three instars of Prokoenenia wheeleri cor- tembryonic. Its very thin cuticle bursts during respond to juvenile instars (Rucker, 1903). Their hatching when the egg membranes open. Among morphology evolves in a quantitative manner several species which give care to the clutch, (number of bristles, articulations of the flagellum, there is a series of 2 or 3 instars of this kind: Fi evolution of genital parts, etc.). F2, F3 (Canard, 1 987). As they cannot move, they stay in the cocoon. Spiders The following instars are mobile and look like The degree of maternal care given to the clutch a spider (juvenile) but the first one or two still lack varies between species. Some spiders abandon some adult characteristics: they are incomplete their cocoon (e.g. araneids). Others carry it in juvenile (Ji) (= iarves' and 'prenymphes' of their chelicerae or attached to their spinnerets. Vachon, 1958b). In some cases, the first instar of COMPARATIVE POSTEMBRYONIC DEVELOPMENT OF ARACHNIDS 465

life on yolk free life, feeding on prey Irfe with the mother on birth free life, feeding on prey yolk or nutritive fluids hatching m3 ^ ml m2 rr\3 J3 Nw 94

FIG. 7. Scheme of postembryonic development stages of ricinuleid Ciytocellus palaezi (after Pittard and Mitchell, 1972)." FIG. 8. Postembryonic development of a pseudoscor- this phase is very incomplete, but in some others, pion, Chelifer cancroides (after Vachon, 1938). some characters can only be shown absent using retains some non-evolved characteristics (sen- an SEM. These first incomplete juvenile instars sorial system, silk glands, digestive tract, etc.). It live on their yolk, but some also feed on un- is an incomplete juvenile: Jii (= 'larve IF and developed eggs, which they can pierce with a 'protonymphe' of Vachon, 1938; 'protonymph' cheliceral blade. Dispersal takes place after the of Weygoldt, 1969). Some species apparently moult which releases a juvenile equipped with all retain a foetal aspect (Judson, 1990). The first free its organs (J). Juveniles then live on prey they instar sometimes remains and moults in the cham- catch (= 'nymphes'; Vachon, 1958b). The total ber constructed by the female (Weygoldt, 1969). number of juvenile instars may vary within a The animal lives alone after this moult. The num- species. Males often become an imago in fewer ber of following juvenile instars is fixed to two: instars than females. J2 (= 'deutonymphe'), J3 (= 'tritonymphe'). Each Female orthognathids and filistatids can still instar can be identified through its moult. In labidognathids, imagos do not moult trichobothriotaxy (Vachon, 1938). Imagos do not any more (adults). In nature, all male spiders die moult any more. without moulting. Scorpions Ricinuleids Eggs hatch in the female's genital tract Hatching releases an active individual, which (viviparous species) or soonafter laying catches prey but which has only three pairs of legs (ovoviviparous species). and therefore does not present the general arach- Newly born scorpions climb onto the mother's nid characteristics. It differs from the imago, and back and remain there, living on yolk reserves. is a larva: L. The three instars that follow They resemble an adult (juvenile) but is incom- resemble the imago, and possess 4 pairs of legs; plete (without trichobothria, unpigmented in- however, they lack genitalic structures. They are tegument, without specific bristles, etc.). It has juveniles instars: Ji, J2, J3 (= 'nymphes' of Pit- temporary organs linked to life on the mother's tard and Mitchell, 1972). There is only one im- back, such as legs without claws but bearing aginal instar (adult). adhesive organs at their tip. This incomplete juvenile is a pullus: JP (= 'larve' of Vachon, P.SEUDOSCORPIONS 1940). The eggs are laid in a brood pouch where they After one moult on the mother's back, the feed upon maternal nutrients with the aid of the juveniles periodically move to the ground, where embryonic membrane. Growth of the embryos they begin hunting and eating for the first time. bursts the chorion and the external side of the They disperse afterwards and live alone. The brood pouch to which they remain attached by the second instar is a complete one (J2). The number buccal region. Their form is not differentiated. of juvenile instars may vary according to sex. Within a few seconds, the mother injects a nutri- Mostly there are 6 to 7 instars, but it may vary tive fluid which trebles their volume (= 'larves from 5 (Orthochirus) to 10 (Diplocentrus) (Polis, gonflees' of Vachon, 1938; 'deutembryons* of 1990). Imagos do not seem to moult, but it may Weygoldt, 1969). Organogenesis continues and a be possible (Stockmann, 1968). moult occurs which releases an animal which emerges by an anterior cephalothoracic tooth. Mites The released instar looks like the imago, but The Acari have many more developmental 166 MEMOIRS OF THE QUEENSLAND MUSEUM

lifconyofc free life feeding on noslnost I I

H M\ZL

free life ** stching (04] ml w3 ' FIG. 9. Postcmbryonic development Stages 01 a scor- m2 yr- pion, Euscorpius italicus (after Angermann. 1957). L jTptgt J3 ' types than other arachnids. The eggs are often ffl abandoned. There is a first instar which has a foetal aspect, but it remains intrachorionic, ex- 94 cept in some cases (Coineau, 1977). This instar can be compared with foetal instars of other FIG. 10. Succession of stages in postcmbri development of mites, (a) Argasidae. OfhtthodoT06 arachnids (F), but it is not postembryonic. maritime (after Guiguen, 1990); (b) Orih.iiul.w The first free instar is obviously different from (Carahodes wllmarmi) (after Bellido. 1983). the imago because it usually has only 3 pairs of legs. It moves and can feed itself. It is a larva: L. tcmatic group of the imago, therefore they are not The following instars are eight-legged and only larvae, but genital organs and furthermore some differ from the imagos by some quantitative or other structures are lacking. This 'incomplete' sexual characteristics: they are juveniles (J) (= situation is not always easy to observe mor- ^nymphes'). The number of instars is often fixed phologically. However, biological information is for a species, at 1 or 2, more often a maximum of useful, because these instars are nearly always 3, but up to 4-5 in the Argasidae. unable to live on their own. After these immature instars imagos appear, Trie pullus (Pavlovsky, 1924) is an incomplete which do not moult anymore (adults) (= juvenile instar with special adaptations to life on 'prosopon' of Renter, 1909). the mother, such as pad-like organs on the tip of In the thrombidiids, there are periods of inac the legs. tivity between larval and juvenile stages and be- The presence of two morpho-biological forms tween juvenile and adult stages. At such times, during the same instar among mites is rare the animal is covered by the original cuticle, but amongst the arthropods {except Diptera), but it secretes a new tegument under it, which becomes does not raise any problems of description and the tegument of the next instar. In many other does not require any fundamental vocabulary cases, motionless instars can he distinguished, changes: it simply requires more accurate defini- sometimes comparable to real nymphs or to tions. specific survival-forms, which allow for dispersal. Chronological and Morpijo-biological DISCUSSION AND CONCLUSIONS Descriptions In the development of a systematic group of Terminclogv and Concepts arachnids, except in ricinuleids, pseudoscotpions Most terms used in other arthropods can also be and perhaps palpigradids, there is no fixed num- used in arachnids. We have used only the original ber of instars Therefore it would be unwise to terms of pullus. foetal instars and incomplete base a study on few species and to fix the juvenile. chronology, because some still unknown

The foetal instar, although it has been defined developments may modify the established sys- for arachnids (Canard, 1987), is not specific to tem. Theoretical systems of this kind were this group. Ft is evident in some insects and proposed by Reuter (1909) for mites and by myriapods (= 'prolarves\ 'prelarves\ 'pseudo- Vachon (1958b) for spiders. Thus, the constant

foetus' , etc.). presence of three post-larval juvenile instars The incomplete juvenile instars belong to the (=nymph) in mites stated by Reuter and often juvenile phase of which Ihey form a part (Jii followed (protonymphe, deutonymphe, followed by J2). One can recognize the sys- tritonymphe.i does not conform to most species. COMPARATIVE POSTEMBRYOMC DEVELOPMENT OF ARACHNIDS 467

in which there are less than 3 juvenile instars and and, unlike other arachnids, arc not all predators. even less to those with 4 to 6 instars (e.g. Or- Therefore, they have followed different evolu- mihodoros marinrmts). tionary pathways and sometimes metamorphosis Developmental analyses based on a fixed num- takes place, with larvae ar>d nymphs (similar in ber of instars (chronological) establishes com- these cases to those of holometabolk: insect mon points between instars of different species, With 'survival' instars which enable them to dis- in order to envisage evolutionary pathways. But perse. these common points will always remain This evolution of clutch orjuvenile care by the hypothetical, and moreover, these pathways can mother docs not indicate phylogenetic relations

be elucidated without this system. Therefore we in the different orders, because it is a general do not wish Jo use a method with no decisive phenomenon within the animal world and can advantages and, because of iis fixed character, appear independently in different groups. limited development descriptions, because Species which do not suit the established system LITERATURE CITED xcluded. In mites, for example, the number of instars is considered fixed to one larval, three ANDRE, H.M. 1 9»9. The concept o! slasc. Pp. 3- 1 4 In juvenile instars and the adult. However, in many H.M. Andre and J.C Lioitt (eds). 'L'ontoe species one or more juvenile instars are missing ct le Concept tic Stase chaz lea Aniiropodc/. AGAR- Wnvrv and sometimes there are more than three juvenile

ANDRE, H.M. A jOCQUfi, ft. I9H6. The detin.iu ,i instars. For mites there is, depending on the BtascsJn spiders and other arachnids. Memoir species, a variable number of instars. laSocieie' Koyalc Beige d'Emomologie 33: 1-14. ANCERMANN, H. 1957. Ubcr verhalten, Spcr- Evolutionary Pathways mophorenbildung und Sinncsphysiologic von EuscorptHs Hbst. und Vcrwandtcn Immature instars of arthropods are adapted to kalians Alton (Scorpiones, Cbactidae). Zeitschrift fttr special lifestyles or environments and sometimes Tierpsychologie 14: 276-302. differ from those of imagos. Often these adapta- BELUDO, A. 1983. Etude morphoJogique ei tions influence the morphology so deeply that it ^cologique de Carabodcs wilhmvwi Bernini, is difficult to distinguish the imago from the ]°75 (Acan, Oribatci) dans une formation immature forms (larva). Such different es art piunnii^ de la landc armoricainc. Thdsc 3cmc marked among the arachnids llarvae absent ex- cycle Rcnncs 1. 79 pp. cept in miles and rictnuleids). BONNET, P. 1933. Le cycle vital de Phitoeus chtysops (Anin&dc. A good correlate probably exists between the Pi»da Salticidc). Archives de Zoologie ExperimcntaJe75: 129-144 increased level ofcart by the moiher to its clutch CANARD, A. 1984. Contribution 1 la connai&sancc dii and the increased number of incomplete instars at developpement, de Ideologic el tie the start of the development. T^cophysiologie des Arancides dc landes ar-

i Moreover, the biology and morphc 1< * >f ear ! >gy y Fnoric^ines. These Doci. Etat, Rennes: 389 pp,» instars can be observed" and explained as adapta- annexe. t52pp. tions to life in the cocoon or with the mother, e.g., I9S7. Analyse nouvelle du developpement pes* Uvc temporary organs such as distal, pedal pad- ^ivunnaire des araignees. Revue Arflch- nologique 7:91-128. like organs of pullus (attaching to the mother) oj? CANARD, A. & STOCKJvlANN, R. 1992. Don tl>£ ehclicera) Made of;some incomplete juvenile comparers sur le developpement poslembrynn spiders (feeding on undeveloped eggs). The i ialie des .-trachnides. Bulletin de la Socicte Scien- viviparous cases do depend on the same kind of uTi*juedeBrclagne61 n°h.s. 1, 1990: 19-50. evolutionary processes. 1 COINEAl , Y. 1977. La premiere prelarvc cllatostau- The evolution of many arachnids has probably que eonnuc chcz Ics Acaricns. Acarologia 19: been characterized by growing care of the clutch 4tv54 Correlated with the increasingly and later CONDE. B. 1984. Les palpigrades: quclques aspects development of the first instars. Thus, the instars morpno- bjologiques Revue ArachnoJoeique 5: are both incomplete and regressed, because these 133-143. GRANDJEAN. P I93X, SitrVontogenie des Acarkn-t adjectives depend on the point of view con- Comptes-rendus de I* Academic des Sciences. sidered: ontogenetic or evolutionary. This cor- Pahs 206: 146 150

responds to the deux temps' of Grand jean i 19541 1954 Lbs deux scries de temps et revolution. Bul- 'stale and, ton certain extent, loihe approach' and letin biolo^ique dc la France el la Belgiqut* 88: the 'staso approach' of Andre* (\9> ] 413-433 Mites present great diversity in their biology JTAL, J. 1943. Du developpement (Xtftembqwir MEMOIRS OFTHE QUEENSLAND MUSEUM

nairc dc Phulangium opitio L. Bulletin de la 7ER, E. 1909. Zur Morphologic und OftfiQgcnle Soctete Zoologjqve dc France 57: 98- 100. der Acariden. Acta Societatis Scientarum Fen

GUIGUHN. C. 1 990. Cycle el ecologie & Ornuhodorus nicae36(4): 1-288. (AlecrnmhiuKl nuiritbnus Vermeil el Marguct RUCKER, A. 1903. Further observations on Koettenia. lAearina: Argasidae) parasite d**0|seati Zoologische Jahrbucher. Systematik (Okolpgie),

Bulletin de la Soeictc francaisc dc Parasitologic, Geographic und Biologie 1 8: 40 1 -434 Supplement 1: 316. STOCKMANN. R. 1979. Developpcment DO* HEMKINO, M. 1882. Beitnigc /ui Anatomic. Entwick- terabryonnaire ct cycle d'intermus chez un Scor- lungsgechichte und Biologic von Tromhidium pion Buthidae: Buthotas minax occidental'^

fuligmostufi Herm. Zeitschrift fur wis- (Vachon et Stoekmann ). Bulletin du Museum Na- Sttischnftlicbe Zootogie 37; 553-663. tional d'Histone Natnrellede Paris. 4emeserie I: HOLM. A- 1940. Studieu iJr*er die Entwicktnng und die 405- 420. EntwickhmgsbhMogie der Spinnen. Zoologjske THALER. K.. 1982. Die Priiiiartarvc der Waizcospili- Biddttg Uppsafg r9: 1-244, nen Gyllipus ci cyprlatica Lawrence (Arachnida. JUBERTHIE, C. 1965. Donnees sur reeologie. le Solifugae, Karschiidae). Mitteiluneen der devefoppement el la reproduction des Opijions, Schweizerischen Entomologischcn Gcscllschaft Rcvuc d'F.cologie et de Biologic du Sol 2:377- 55; 93-05. 397. VACHON, M. 1938. Recherches anatonuqtu JUDSON, M.L.I. 1990. 'nieremarkaMepn.tsmyinphof biologiques sur la reproduction et le Psetu1ochlhonius{Ch\:\om\h\, ChlhoniidacV Bul- u'ev cloppement des Pseudoscorpions. These Dol 1

I lettade|a.Soci6(^Europ i taig&n Etat, Pans: 1-207; Annates des Sciences Narurcl- h.s. I: I59-161 les. Zoologie, serie t f: 1-207. JUNQUA, C. I966. Recherches biologiques et lu>- W0. Sur la systcmatique des scorpions. Memoires tophvsiologiques sur un Solifuge sabarien. These du Museum National d'Histoire Naiurelle de

Doci Etat Paris. 1 24 p. Paris 2: 241 260. KASTNLR, A. 1949. Zur Enlwicklungsgrchietoe yafl 195Sa. La larvc dc Galeodes arabs C.L. K. (Arach- V'msoma Tkelyptumus rututotus L. fPedipalpn. 2. nide, Solifuge). Comptes Rendus de 1'Academie Tcil. Die Enlwkklung der Mundwcrk/c des Sciences, Paris 246: 477- 480. Beinhuften und Slema. Zoologische Jahrbucher, 1958b. Contribution a l'etude du developpeincni Anatomic und Ontogenie 70: 169-197. post-emhryonnaire des Araignees. It:' notr: LAWRENCE, R.F. t!M7. Some- observations on ihe generaiiies et nomenclature des siades. Bulletin eggs and newly hatched embryos vfSolpitgn ho.\- de la Soci6te Zootogique de France 82, 1957 tijis White (Arachnida). Proceedings of the (1958): 337-354, Zoological Society o i London 1 17: 429-434. P. 1969. The Biology of LEGENDRE, R. 1958. Contribution a l'etude du WEYGOLDT r Pseudoscorpions*. (Harvard University Press: devcloppcmcnl embryonnairc des Araignees. Cambridge, Mass.). Bulletin dc la Soeiere Zooloeiqne dc France S3: 60-75. 1 970. Lcbenzv klus und postembryonisehc Entwick- lung der Gcissclspinne Tarantula mar MILLOT. J. 1949. Ordrc des Amblvpygcs. 563-588. In: gmemaeulato C.L. Koch (Chcliccrata. P.P. Grasse (cd). Traite de Zoologie'. Vol 6. Masson: Paris. Amblypygi) im Lahoratonum. Zeii.schritl liir Morphologic der Tiere 67: 58-85. MUMA. M.H. I96G, The life cycle of Eremnhuus 1975. UnterSUChUngen Zur EnnbryoJogje und Mor- durangenus ( Arachnida: Solpugida). Florida En- lqgj»1 49; 23^-24: phologic der Geisselspinne Tarantula mar- NAISSE.L 19.59, Neurosecretion el glandes endocnoes wacvlatQ C.L. Koch (Arachnida. chci les Opilions. Archives de Bioloejc 70: 217- Amblvpygi, Tarantulidac). Zoomorpholoeie 82: 264 L37-199. PAVLOVSKY, E.N. 1924. Studies on ihe organization YOSHIKURA, M. 1965. Pi>stembryomc development of a whip scorpion, Typopehix sttmpsonii (Wood). and development ofsa )i ptoSi Quarterly Journal of Microscopical Science, London 68: 615 640. Kumamoto Journal of Science. Series B Bioli PEREYASLAWZEWA, S. |90I, OevHoppement 7:21-50. embryonnaire des Phryncs. Annates des Sciences 1975. Comparative embryology and phylogeny of Naturclles, Zoologie. 8ernc scne, 13: 117 304. Arachnida Kumamoto Journal of Science, Series P1TTARD, K. & MITCHELL, R W 1972. Compara- B Biology 12:71-142. tive morphology of the lite stages of Cryptoceltas YSNEL. F.1992. Impact trophiquc er valcur bioin-

pnSacz* I Arachnida, Ricinulci). Graduate Studies dicatrice d'une population d'Araigne'es: Exempte

Texas Tech University I. 1-7/. dune cspecc a toile geometrique Lariniuides cur- 1

POLIS. G. 1990. Biology of Scorpions . (Sfettfbri nutusi Araneidae). These LJniversitc' Rcnnes 1217 M-tsiiy Press: Stanford). pp. ;

COURTSHIP, MATING AND POST-OVIPOSITION BEHAVIOUR OF HYPOCHILUS POCOCk'l PLATNICK (ARANEAE, HYPOCH1LID

K.M..GA1LB)

Catley. KM. 1993 11 II: Courtship, mating and post-oviposition behavmu, » >t HypOi pococh Ptatrricfe (Araneac, Hvpochilidaej. Memoirs of the Queensland Museum 33 (2): 469-474. Brisbane. ISSN 0O79-BS35

Couitstltp and mating in Hypochilus poeocki Platnick is described for Ihe first lime for any member of the supcrfamily Hypochiloidea. Five phases of male behaviour are recognised; pre courtship, non-contact courtship, contact courtship, copulation and posl-copularory behaviour. Chemotacbc stimulation seems to be the prime releaser of male courtship behaviour which involves web-tugging, mutual leg-stroking and female guarding. Post- QVipositioa behavioCD IS described and the role of a previously undescrihed sheet web,

constructed by females after oviposit ion as well as early instar spiderlings. is discussed in terms of its phylogeneticin-.plie;jiK»ns,["j//v/'.-<^V7/7^v, CGnrlshty* phytogenetics, sexUal st '• unction,

K.M. Catley, Department of Biology* Western Carolina £/«n Cullowfur,

The significance of undertaking etiological and exhibit an interesting pattern of disjunct cn-

l studies of spiders in the family Hypoehilidac demism (Catley, 1991 ; Huff and Coyle, l stems from the phylogenetic position of the fami- Males moult to maturity later than females, typi- a relic taxon at the base of ihe Infraorder cally appearing in early August. They move

Araneomorphae (Platnick, 1977; Foistcr et al, t tensively (presumably in search of females) and 1967). Diagnosis of behavioural units that are do not associate with penultimate females, sug- used in such processes as web construction (Cod- gesting that there is no first male sperm dington, 1986) and courtship (Platnick, 1971; precedence (Eberhard el ul. in press). Hulversen, 1976. Coyle and O'Shields. 1990) may prove useful in future phylogenetic anal} sis. MATERIALS AND METHODS This paper describes the courtship and mating Hypochilus behaviour of pococki in the Despite extensive day and night observation laboratory and attempts to recognise potentially during 1990 and 1991, no courtship or mating informative behavioural sequences. Egg sac con- behaviour was observed in the field. Hypochilus struction is also described and attention to drawn can be very difficult to maintain in the laboratory a previously undescrihed silk construct, a Veil they are affected adversely by changes in web* built by post egg-laying females as well as humidity and only occasionally can they be per- instar spiderlings. Apart from one incom- early suaded to construct a web, making feeding plete observation by Fergusson ( 1972) of mating problematical. Following several unsuccessful in Hypochilus thorelli Hoffman {—Hypochilus attempts to establish a mature female in the kf) this is the first detailed description of pOCOi a laboratory, a single specimen collected in July reproductive behaviour for any member of the 1990 from Wolf Creek watershed. Cullowhee. supcrfamily Hypochiloidea. Jackson County NC. established a web in a The genus Hypochilus comprises live species 50x30cm glass tank. The tank was furnished with confined to the Southern Appalachian highlands a sloping wooden framework (45°angle) covered Ot eastern North America, two species from with sandpaper to simulate an overhanging rock central and northern California and o\\q from ledge The floor was covered with vermiculite to central Colorado. Two additional species, one a depth of 5cm and was kept very moist. from New Mexico and the other from the San After the female attacked and killed the first Bernardino Mountains of southern California male introduced into the arena, she was allowed will soon be described (Catley in prep.). These to feed for a period of one week. The second male traplogyne cribellale spiders build characteristic was introduced on August 28 1 990 and a total of lampshade' webs on rock surfaces often close to 7 hours 47 mins of maJe/fcmalc encounters were i mining water. All species appear to be allopatnc Filmed using a Panasonic WV-D5000 vidci 470 MEMOIRS OF THE QUEENSLAND MUSEUM

FEMJ-\L£ IrJg^WJ-ai'Uri

in center of introduced lampshade

contacts female web

manipulates web with pedipalps and legs no response / X grooms legs and 'bobbing' pedipalps \ / palpal gesticulation

no response

1 web-tugging

1 \

'rests' |

attacks retreats N / / \

leaves web but tugging on lampshade no attack, breaks down wall approachs male

mutual leg stroking

palpal insertions

takes up 'guarding' position

3 Q-o.

FIG. 1. Sequence of courtship and mating behaviour in Hypochilus pococki formulated from encounters in the laboratory involving two males and a single female (see text for explanation). )

COURTSHIP TNHYPOCH1LUS POCOCKI 47 <

cordcr fitted with a Micro-Nikkor 55mm close-up web as a result of male tugging and approached lens. Sessions were annotated by verbal com- the male, waving her front pair of legs. The male ments recorded through the audio channel of the held his ground (unlike previous female ap video camera. Data analysis was achieved using proaches which appeared aggressive and from slow motion iuvd freeze frame functions of a which the male quickly withdrew). A long ses- VCR. sion of mutual leg-stroking followed, involving mainly legs one and two. The male maintained a RESULTS constant, very rapid stroking of the female* s legs (mainly metatarsi and tibiae) and body as she The description of courtship and mating appeared to become progressively more catalep- presented here is a composite derived from tic, and eventually assumed the acceptance pos- laboratory observations of only two males and a ture. This stroking session lasted for 3 unn single female and therefore may not be repre- secs-more than twice the period of copulation sentative. itself. Male courtship behaviour can be divided into five distinct phases; pre-courtship, non-contact COPULATION courtship, contact courtship, copulation and post- Palpal insertions. Following mutual leg-strok- copulation (Fig. 1). The following behavioural ing the female oriented her abdomen at 45° to the units were diagnosed. substrate and adopted a semi-cataleptic acceptance posture. The male faced the female and Maie Behavioural Units advanced, with pedipalps fully extended, to ;j position where his dorsal cephalothorax was ad- ('Rh COURTSHIP jacent to the female's sternum (Hg, 2). It was nm Ley and pedipalpal grooming. Within a few entirely clear whether or not the male tapped on minutes of a male being introduced into the arena. the female's genital area with his palps prior to the legs and pedipalps were used to manipulate insertion as described for some araneids (Robin- female silk. web constructed by the same A son and Robinson, 1980). Such apparent tapping female but which had been abandoned for several may simply be attempts to locate the opening to weeks elicited the same response. Rubbing the the bursa copulatnx tarsi and metatarsi of legs 1, 2 and 3 together, as The palps were inserted alternately, the right well as drawing them and the pedipalps between followed by the left, each insertion lasted 3-10 the open chelicerae (and presumably the endites seconds with the whole insertion sequence-lasting occurred immediately after the first and sub- I min. 22 sees. To achieve insertion from this sequent encounters with the female web. The position requires thai the palpal organ be twisted pedipalps were rubbed vigorously across the silk through 90° at the same time as the pedipaip is for extensive periods, this was followed either by straightened. The copulatory phase ended abrupt- drawing them through the mouthparts, as ly when the female broke away from the male, described above, or by rapid pedipalpal gesticula- who was immediately pursued some distance tions. from the web. At no time during courtship or Bobbing. Flexing of the legs resulted in the mating was the male observed to lay down silk whole body moving up and down relative to the substrate. Bobbing is often interspersed between sessions of leg and pedipaip grooming. POST-COPULATION 'Guarding "posture. After copulation, the male, NON-CONTACT COURTSHIP after a brief period of palpal grooming, took up a Web-tugging. The male pulled on the cnbellate characteristic position close to the female, often lampshade web of the female with his first pail of touching her. His legs were extended parallel lu legs. actions, repeated in bouts lasting 3-7 Such the substrate, with the first three pair diro I seconds interspersed with periods of inactivity anteriorly, and the fourth pair directed posterior- result in lasting from 5 sees, to several mins., can ly. The first pair of legs were held in such . the side wall of the lampshade being partially position such that the femora were at 30° to the destroyed. longitudinal axis of the prosoma, while the remaining podiies tended 10 converge dist CONTACT COURTSHIP over the female. This seems to be a characteristic Leg-stroking. The female eventually left the position seen often in the Held. It was maintained r: MEMOIRS OF THE QUEENSLAND MUSEUM

FIl - 2. Mating position of Hypochilus pococki (for the sake of clarity not all male appendages arc illustrated) for 2.5 hours after which the female attacked and 17, and the final egg-sac was constructed on badly injured the male. November 6. Each was suspended from the framelines of the web while particles were incor- Female Behavioural Units porated in the outer layer of the egg-sac. This involved the spider descending to the substrate non-contact and carrying pieces of vcrmiculitc back to the Attack, The female failed for long periods to web in her chelicerae (in the field egg cases are show any response to the male's web-tugging but covered with particles of moss or lichen (Fergus- did eventually respond by rushing out of the son.. 1972; pcrs. obs ). lampshade and pursuing him. Three attacks ap- Two egg-sacs were 'screened in' by a particular peared in earnest, with the male withdrawing type of sheet web not previously described for rapidly. The fourth response was instigated more Hypoehilus, Observations from the field show slowly (with reduced speed and vigour); this that a similar 'veil web' is also constructed by change in intent' appeared to be sensed by the early instarspiderlings (Catley, 1991 ) The silk is male, who did not retreat. This encounter led of very different appearance from either directly to contact courtship and copulation (Fig. framework or cribellate silk and appears as a dense, finely woven sheet. Its purpose, either as a vertical 'veil' in front of egg-sacs (typically CONTACT COURTSHIP those suspended in a fissure in the rock), or as a Leg stroking. Sec male behaviour. barrier underneath which a number of very early instar spiderlings build their regular prey catch- copulation ing lampshade webs, may be protective. Acceptance posture- A semi-cataleptic position with the abdomen held at 45° to the substrate (Fig. DISCUSSION

2 ) occurring after contact courtship (extensive leg stroking) with the male. The function of courtship in spiders may be most simply expressed as: alerting the female to POST-COPULATION the presence of the male, the possible suppression Egg laying ethology. Twenty days after copula- of female predatory behaviour and stimulation of tion, the female laid eggs and constructed an the female to accept copulation. Variations on egg-sac. First a saucer-shaped disc of pink silk this basic theme have been voiced by several was laid down onto which the eggs were authors including Bristowe (1958), Crane (1949) deposited. This was then closed up to form a and Platnick (1971); most expand the concept to flattened sphere The pink colour of the silk has include elements of species specific recognition. been corroborated by several observations in the advertisement of sexual availability and the field, however, when the egg-sac is ready to be functioning of a releaser system. Other com- covered in cryptic material its colour is off-white. ponents which may also be important are male A second egg-sac was produced on October 4, behavioural elements designed to ensure his post- followed by a much smaller egg-sac on October copulation survival. :-:TSMIP IN HYPOCHILUS; POCOCKJ 473

Accounts of courtship and mating in other more a role in placating the female, resulting in her

i live spider taxa are scant, but include infor- adopting the acceptance posture. It may well be mation on the Mesotbelae: Liphistiidae, Hep- homologous with the "leg fencing' behaviour tuthrla (Haupt, 1977); Mvgaiomorphae: seen in some diplurid spiders (Coylc and O'- Nemesiidae (Buchli. I%2); Atvpidac (Clark, Shields, 1990). 1969); Dipluridae (Coyle and O'ShieJd -. [990), Copulation was achieved in the mating position Whereas behavioural observations for spiders in (Fig. 2; position 1 of Kaston, 1981 1. Alternate the Araneoclada arc available from Bristowe palpal insertions, as observed in H\porhilt4S, (1941, 1958) various families. Crane (1949) lor should be considered plesiomorphic when en- saltici'ds, and Robinson and Robinson (1978, countered in other iran^omorph spiders (using 1980) fox arancidftt this account of courtship and Hypochilus as an outgroop) and as apomorphic mating in HypodiHus is the first to be published when palps are insen Siuuil fot Any lower (non- Araneoclada) arancomorph taneous palpal insertion has been documented in spider the following families, Dysdcridac. Scgcstriidac. The courtship and mating repertoire of Oonopidac, Scytodidae andPhoIcidac (Bfistfl

Hypochilus is relatively undcrived and given the 1958) hut appears no*, ftS Bristowe conclude i^i n; 1 ics phylogenelic position the behavioural be plcsiomorphic for the Araneoclada. Simul- units which comprise it may be hypothesised to tari&ous Insertion rmj well prove to t represent the plcsiomorphic condition of synapomorphy for the higher haplogyncs, Dys- arancomorph reproductive behaviour in general. dcioidca plus 'Scvtodids* (Coddington and Levi. Given a larger data base within the Araneomor- 1991) phae. comparison of behavioural characters will The pust-copulatoxy position taken up by the allow correct polarity decisions to he made and male is believed to represent a guarding posture the resulting data set, combining both mor- Hypochilus pococki males have been shown con- phological and ethological characters, should clusively not to associate with penultimate puv, idc stringent test loge-ny. a more of phy females 1 Eberhard el al. in press). Hence, that tin- Leg and pedipalpal grooming behaviours per- numerous oc: .hen males were found in formed by the male upon contact with the female this position in the field, suggest that they rt web may be indicative of the occurrence ol sent in fad, post-mating situations, and that the female pheromone on the web. Such prc-contact male was most likely guarding the female and encounters were necessary lo release web-iug- thus his chance of paternity. ging behaviour in these observations (o=4) and is The implications of the 'veil' web produced by

i stent with Platnick's (1971 ) hypothesis thai the female following egg-sac construction ic chcmotactic stimuli are prime releases of male quire further comment. Hypochilus egg-sacs arc courtship behaviour in haplogyne spiders. Il has superbly cryptic anil mXtfl ate not concealed b) been suggested that the longer anterior legs of such a 'veil' web (Shear, 1969; Fcrgusson, 1972; mobility male Hypochilus confer superior when Catley, 1 99 1 ). The behaviour of some females in locating females (Fergusson, 1972; liber hard el concealing egg-sacs might have primitively rep- al in press). However, such pronounced sexual resented a selective advantage from vertebrate or dimorphism may have had its origin in courtship hymenoptcran predation behaviour. The extreme length of the first legs, Accepting ihccladistic hypothesis of Forsterrf maximising the distance between male and by al, (1987) on the rel at mnships of the female during web-tugging, should increase the Hypochiloidca and Ausrrochkloidea, data on male's surviving attacks. chance of female Such sheet web construction in these taxa suggests that interactions also for may provide an opportunity the lampshade web of Hypochilus is autapomor sexual selection by female choice to occur (Elxi phic. Ectatostictti davidi (Simon), the sister taxon hard, 19X*)), ihc female testing male 'fitness' by of Hypochilus, constructs a sheet web (Li and repeated attacks. rln Zhu> 1 984 1 as all kAOWtl membcts of

Contact courtship was initiated when the AustrocbiliducfForster^rtf/.. 1 987). Ontogenetic female eventually left the lampshade and ap- cv idence, based on the observation that very early also sheer web proached the male. Such behaviour may not how- inslarspiderlings construct a | ever he typical. Fcrgusson ( 1972) reported that in ley, 1991), also supports this hypothesis and lends the one encounter he observed the male some weight to the suggestion that the plesiomot scrambled into the lampshade With the female. phic web construct of araneomorph spiders was The long period of mutual leg-stroking may play the sheet web. 474 MEMOIRS OF THE QUEENSLAND MUSEUM

ACKNOWLEDGEMENTS LEY, K. M. (in press). Correlation between sper- mathecal morphology and mating systems in spiders. Biological Journal of the Linnean Society. Grateful thanks are due to Dr. Frederick Coyle FERGUSSON, I. C. 1972. Natural history of the spider for the loan of video recording equipment, Rudolf Hypochilus thorelliM&rx (Hypochilidae). Psyche Meier for help with translations and Dr. William 79: 179-199. Shear for his useful comments on an early draft FORSTER, R.R., PLATNICK, N.I. & GRAY, M.R. of the manuscript. I should like to acknowledge 1987. A review of the spider superfamilies financial support from the Organising Committee Hypochiloidea and Austochiloidea (Araneae, of the XII International Congress of Arachnol- Araneomorphae). Bulletin of the American ogy, the Graduate School of Cornell University Museum of Natural History 185: 1-116.

and the Grace Griswold Fund which allowed me HAUPT, J. 1977 Preliminary report on the mating be- to attend this Congress. haviour of the primitive spider Heptathe la kimurai (Kishida) (Araneae, Liphistiomorphae). LITERATURE CITED Zeitschrift fur Naturforschung 32: 312-314. HELVERSEN, O. VON 1976. Gedanken zur Evolution BRISTOWE, W.S. 1941. The Comity of Spiders, vol. derPaarungsstellung bei den Spinnen (Arachnida, Araneae). 3: 13-28. II. (Ray Society: London). Entomologica Germanica 1958. The World of Spiders. (Collins: London). HUFF, R. P. & COYLE, F. A. 1992. Systematics of BUCHLI, H. 1962. Note preliminaire sur Hypochilus sheari and Hypochilus coylei, two l'accouplement des araignees mygalomorphes southern Appalachian lampshade spiders Nemesia caementaria, Nemesia dubia et (Araneae, Hypochilidae). Journal of Arachnology Pachylomerus piceus. Vie et Milieu 13: 167-178. 20: 40-46. CATLEY, K.M. 1991. The phylogenetic relationships KASTON, B.J. 1981. Spiders of Connecticut. State of the species of the lampshade spider genus Geological and Natural History Survey of Con- Hypochilus (Araneae, Hypochilidae). (Un- necticut. Department of Environmental Protec- published Masters thesis, Western Carolina tion. Bulletin 70. University). LI, ZHONSHAN & ZHU, CHAUNDIAN, 1984. Ec- CLARK, D. J. 1969. Notes on the biology of Atypus tatosticta davidi (Simon, 1888) of China affinis Eichwald. Bulletin of the British Arach- (Araneae: Hypochilidae). Journal of the Bethune 36-39. nological Society 1: Medical University 10(5): 510 (in Chinese). CODDINGTON, J. A. 1986. The monophyletic origin PLATNICK, N. I. 1971. The evolution of courtship of the orb-web. Pp. 319-363. In Shear, W. A. (ed). behaviour in spiders. Bulletin of the British Arach- 'Spiders. Webs, behavior, and evolution'. (Stan- nological Society 2(3): 40-47. ford University Press: Stanford, California). 1977. The hypochiloid spiders: a cladistic analysis CODDINGTON, J. A. & LEVI, H.W. 1991. Sys- with notes on the Atypoidea (Arachnida, tematics and evolution of spiders (Araneae). An- Araneae). American Museum Novitates 2627: nual Review of Ecology and Systematics 22: 1-23. 565-92. ROBINSON, M.H. B. 1978. The COYLE, F. A. & O'SHIELDS, T. C. 1990. Courtship & ROBINSON, evolution and mating behavior of Thelechoris karschi of courtship systems in tropical araneid (Araneae, Dipluridae), an African funnel web spiders. Symposia of the Zoological Society of 42: 17-29. spider. Journal of Arachnology 18: 281-296. London CRANE, J. 1949. Comparative biology of salticid 1980. Comparative studies of the courtship and spiders at Rancho Grande, Venezuela. Part IV. An mating behavior of tropical araneid spiders. analysis of display. Zoologica 34: 159-214. Pacific Insects Monograph 36: 1-218 EBERHARD, W. G. 1985. Sexual selection and animal SHEAR, W. A. 1969. Observations on the predatory genitalia. (Harvard University Press:Cambridge). behavior of the spider Hypochilus gertschi Hof- EBERHARD, W. G, GUZMAN-GOMEZ, S. & CAT- fman (Hypochilidae). Psyche 76: 407-417. H

EFFECTS OF SAMPLING METHOD ON COMPOSITION OF A TASMANIAN COASTAL HEATHLAND SPIDER ASSEMBLAGE

TRACEY B. CHURCHILL

Churchill, T.B. 1993 II II: Effects of sampling method on composition of a Tasmanian coastal heathland spider assemblage. Memoirs of the Queensland Museum 33(7: V 475 Brisbane. ISSN 0079-8835.

The composition of a spider community in coastal hcathlands of north-east Tasmania was derived from a 16 month survey using pitfall traps, sweep net and visual search sampling methods Incorporated into a replicated, standardised sampling program. This interpretation of composition is shown to rely on the relative efficiency of the three collecting methods to sample the taxa present. Since mature spiders arc required to confirm species identity, the differential selection of age and sex classes by the methods is illustrated. Whilst pitfall traps catch a greater number of taxa (at all taxormmic levels) and adult spiders, certain ia\a are not or barely represented by this widely used technique. The subjective nature of the visual search method allows for the potential to target mature spiders. Limits of the sampling methods are emphasised in response to a growing dependence on survey data for the assessment of biodiversity .[jAraneae, methodology, biodiversity, heathland. invertebrates, community.

Tracey B. Churchill Division of Environmental Science. Griffith University. Nathan,

Queensland 4 1 , Australia; 6 November, 1992.

Surveys of spider communities in Australia catch adults will effect the accuracy of a species have primarily been motivated by specific list. taxonomic interests in this relatively unknown By adopting a suite of collecting techniques to faunal group. Whilst this has lead 10 invaluable target spiders both on the ground and in vegeta- improvements in taxonomy, such collections are tion, the chances of sampling all taxa present are increasingly being utilised to extract data for the increased and thus data more useful for com- making of critical conservation management munity studies are collected (Uetz and Unzicker. 1976). Accordingly, a combination of pitfall trap, decisions. It is therefore necessary that further sweep net visual was selected consideration be given to the factors that can and search methods

for a 16 month survey of spiders in the nqrth-ea s I affect the interpretation of survey results. coastal comer of Tasmania. The area is largely The primary factor that limits the comparability developed as sheep and cattle grazing propertns of data from different locations or times is the although to the seaward side of the remaining method used to sample spiders. Different coastal Eucalyptus and Casuarina forests is often methods can preferentially sample certain a margin of heathland dominated by members of mierohabitats and/or particular taxa tMerrett and the Proteaceae, Casuarinaceae, Ep&cridsu Snazell, 1983). For example, the commonly used Papilionoideae and Xanthorrhocaceae. An id pitfall trap selects ground active species (Duffey, creasing impact of recreational and residential 1974; Merreit and Snazell, 1983; Lowrie, 1985) development threatens the remaining heathland and the use of this method alone can produce (Kirkpatrick, 1977). In this paper, the composi- species lists that under-represcnt more sedentary tion of spiders in the heathland community is or foliage inhabiting members of the community. inferred by the list of spider taxa collected over The effectiveness of different sampling techni- the survey period. The relative efficiency of the ques can be influenced by behavioural differen- three sampling methods in capturing dominant ces between not only taxa, but also age or sex taxa is then compared to illustrate how the choice classes nf a given species. For example, males of of method can influence the final interpretation many species are more readily captured by pitfall of community composition or species ridioea traps than females due to their active search for a mate (Merrett, 1967, 1968), which may represent MATERIALS AND METHODS ground activity in an otherwise foliage dwelling taxa. Since mature specimens are usually re- Spiders were collected using pitfall Irai quired to identify species or genera, the ability to sweep net and by visual searching, each of which 476 MEMOIRS OF THE QUEENSLAND MUSEUM

was standardised for effort and replicated. At microhabitat. but a bias was shown towards those monthly intervals from October 1986 to January secured within web retreats, as such groups may 1988, sampling was carried out during a one week not be amenable to capture by the previous two field trip. Two replicate 90m** sites were selected methods. at each of two study areas, Waterhouse Point and Hddystone Point. Within each site there were nine RESULTS AND DISCUSSION 18m" plots placed 18m apart in three rows of three. This allowed for three replicate plots per Composition sampling method, allocated initially at random. A total of 8,625 spiders were collected using all following sampling For the relevant plot the three methods over 16 months, and these spiders routine was employed, a) Nine pitfall traps were comprised 130 species of the Araneomorphae in set apart in a 3 x 3 matrix using 9cm 4.5m 97 genera and 34 families (see Table 1). Names diameter traps; b) one sweep sample of 50 sweeps could not be allocated to 26% of genera and 92% was taken using a 28cm diameter net in a 12x3 of species, indicating further that many m area and c) visual searching for 30 minutes was Australian groups need taxonomic revision made over a 3 x 3m area. Spiders were preserved (Davies, 1985; Raven, 1988). The most diverse in 70% alcohol, identified to species where pos- families in terms of the number of species were sible, and lodged with the Queen Victoria the Salticidae (14 spp.) and Gnaphosidae (11). Museum, Launceston, Tasmania. followed by the Theridiidae (9), Zodariidae (9), The three sampling methods were considered Thomisidae (8) and Araneidae (8). The four most to be complementary in their selection of taxa abundant species were, in decreasing order occupying different strata. Pitfall traps sample Diaea sp. (5.8%), Badumna vandiemani (5.3%). spiders mobile on the ground, in contrast to sweep Odo sp. (4.3%), Hestimndema sp. (4.1%). netting which targets spiders in the foliage. The number of spiders falling into pitfall traps Visual searching can reveal spiders in any depends on their activity (Mitchell, 1963;

Lycosidae

Zoridae Amaurobiidae Gnaphosidae

Zodariidae Thomisidae

Prodidomidae

Salticidae

Theridiidae

Linyphiidae

Micropholcommatidae

Clubionidae

600 1200 1800 2400 Number of individuals

FIG. 1. Number of individuals (shaded bars) and adults (black bars) for ihe twelve dominant spider families. "

SAMPLING METHODS ANDTASMANIAN SPIDERS 477

Amaurobiidae Heterupodidae Genus S sp.2 * Genus S sp.3 Genus Asp. 1 Neosparassus sp. * Genus S sp.4 Genus A sp.2 Linyphiidae Genus T sp. Genus B sp. * Laetesia sp. 1 Genus C sp. Genus U sp. Laetesta sp.2 Genus D sp. Genus V sp Laetesia sp.3 Genus W sp. Amphinectidae Laetesia sp.4 Genus X sp. Amphinecta mitvinus (Simon, 1903) Genus N sp. Mamoea sp. Genus O sp. Stiphidiidae Araneidae Lycosidae Biaimi sp. * Corasoides australis Butler, 1929 Cyclosa sp. Artoria sp.l * Stiphidionfaceium Simon, 1902 Eriophora biapicata (Koch, J 871) Artoria sp.2 Gasteracantha mimu'Thorell, 1859 Artoria sp.3 Tetragnathidae Artoria * Genus E sp. sp.4 Deliochus sp. Artoria spp. Genus F sp. Phonognatha sp. Genus sp. Lycosafunesta (Koch, 1849) G Tetragnatha sp. Genus H sp. Lycosa speciosa (Koch, 1879) Theridiidae Genus I sp. Lycosa sp, Clubionidae Micropholcommatidae Ac/jrt?tffYm«/sp. Dipoena sp. Cheiracamhiwn sp. Micropholcomma sp. I Episinus sp. Clubiona sp. 1 Micropholcomma sp.2 Euryopis sp. Clubiona sp. 2 Texiricelta Sp.l * Phoroncidia trituberculata (Hickman, Clubiona sp. 3 Te.xtricella sp.2 1951) Genus J sp. Te.xtricella sp.3 Sffd/otfasp.l Mimctidae Corinnidae Steatoda sp,2 A sudtpus sp. Australomimetus sp. Steatoda livens (Simon, 1895) * Castianeira sp. Milurgidae Tberidion sp. Supunna sp. Miturga sp. 1 Thorn is idae Cyatholipidae Miturga sp.2 Cymbacha sp. *. Uliodon velox (Hickman. 1930) Hanea sp Diaea sp. * Matilda sp. Vliodon sp. Sidymella sp. 1 Desidac Mysmenidae Sidymella sp.2

AuStmusia sp. Genus Psp. Sidymella sp.3 Badumna vandiemani Gray, 1983 * Nicodamidae Sidymella sp.4 Forsterina Sp, Sidymella longipes (Koch. 1874) Nicodortm tm'lana2ffnihusi\)Tt$toaSt, 1893) Tuakana sp. Stephanopis sp. Oecohiidae Dictynidae Tnxopidae Oecobius annulipes Lucas, 1846 Cattevophthalmus sp. 1 Laestrygtmes setosa Hickman, 1969 Oonopidae Catlevophthahnus sp.2 Trochanteriidae Gnaphosidae Orchestina sp. Corimaethes sp. Genus Q sp. Anzaciasp.] * Oxyopidae Zodariidae Anzacia sp.2 * sp. * Eitica sp. * Genus R sp. Afffe/Wl Asteron "reticulation" Megamynnaekkm sp Pararchaeidae Micarta sp. 'Australatica" Sp. Pararchaea sp. Trachycosmus sp. * Habronestes sp.l Pisauridae Zelotes sp. 1 Habronestes sp.2 Zehles sp.2 Dolomedes sp. Habronestes "bradlcyt Zf/o/t'j.sp.3 Prodidomidae Neostorena sp .1 Neostorena sp.2 Genus K sp. * Molycria sp. Genus L sp. Nostera sp. Salticidae Hadrotarsidae Zoridae Lycidas sp. //aWro/armvsp.l * Argoctenus sp. Maratus sp. * I/admtarsus sp.2 Hcstimodema sp. Opisthoncus sp. Genus M sp. 0

TABLE 1 . List of spiders collected from Tasmanian coastal heathlands. Aslcrisk indicates the 20 most abundant species. 478 MEMOIRS OF THE QUEENSLAND MUSEUM

Greenslade, 1964; Uetz and Unzicker, 1976; B PHMALES Merrett, necessarily actual 1983) and not on MALES abundance in the community (Merrett, 1967; B SUBADULT FEMALES Merrett and Snazell, 1983). Individuals active on in PENULTIMATE MALES JUVENILES foliage presumably experience a higher prob- S ability of being knocked into a sweep net and for the visual search method the chance of noticing spiders would be increased by their activity (Cur- tis, 1980). Therefore it is stressed that references r 50 to abundance in this paper relate to numbers caught and not population size.

With respect to the number of individuals, the 10 collections were dominated by the Lycosidae (26% of the total), Zoridae (10%), Thomisidae (9%) and Zodariidae (9%) (Fig. 1). The families Salticidae, Amaurobiidae, Desidae and Gnaphosidae then account for the next 22%. With

the exception of the zodariids and zorids, these 20 families are amongst the largest in Australia (Raven, 1988). The Araneidae, which are other-

wise the most abundant Australian spider family L0 (Raven, 1988), comprised only a minor com-

ponent of this collection ( 1 .4%).

The dominance hierarchy (Fig. 1 ) is determined by the number of adults in each family. However,

as it is not standardised as to whether the whole PITFALL TRAP SWEEP NET VISUAL SEARCH data set (including immatures) or only the adult 2. individuals in each sampling data are used to describe patterns of family FIG. Percentage of method for the different age and sex classes. dominance in a given community, a comparison is made to both (Fig. 1). The interpretation of net (75%) to visual search (84%) methods. From relative abundance of families is affected by these results pitfall traps seem to be the mast which category is used. Given that only adult data efficient at selecting mature spiders from the are useful for comparisons at the generic or coastal heathland community. species level, adult data seem the better choice for total of families, genera, and assessments of biodiversity. The number species can be compared to that collected by each sampling method (Table 2). At each taxonomic Effects of Sampling Method on Composition level, pitfall traps sample the most taxa (between Pitfall traps collected the most individuals 87-94% of the total), followed by visual search- (6212), followed by visual searching (1900) and ing (41-66%) and sweep netting (25-41%). It is sweep netting (513). However, as the sampling relevant to point out that the results presented effort of pitfall traps far exceeds that of sweep here were derived over a 16 month survey period. netting and visual searching, comparisons of taxa The likelihood of recording certain taxa using a between methods are made relative to the total of given sampling method for typically shorter sur- each method. vey periods depends on the relative ease with

Due to a reliance on acquiring mature spiders which they are collected by that method (it also from surveys to confirm species and generic level depends on the temporal abundance of taxa, to be identifications, Fig. 2 presents the differential discussed elsewhere). The percentage of adult distribution of age and sex classes for each sam- Total Pitfall Sweep Visual pling method. Pitfall traps clearly caught the greatest percentage of males (35%) and visual Family 34 32 14 23 searching, the least (3%). Females are also col- Genus 97 84 30 45 lected more by pitfall traps, although the dif- Species 33 130 113 | 33 ference between methods is not as distinct. Accordingly, the percentage of immature spiders TABLE 2. The number of taxa in total, and for each increases from pitfall traps (46%), through sweep sampling method at three taxonomic levels. SAMPLING METHODS AND TASMANIAN SPIDERS 479

Lycosidae

Zoridae ttmmm Amaurobiidae Gnaphosidae SBS Zodariidae

Thomisidae l illllllllll I llllinMlllMllMIIIIIIIIIIIIKIIIIillllllllMIIIMIHIIIITTl Prodidomidae

Salticidae Visual search (n=303) Theridiidae m Sweep net (n=130) Linyphiidae B Pitfall trap (n=3315)

Micropholcommatidae

Clubionidae

10 20 30 40 50 % Adult total for each sampling method FIG. 3. Percentage of the adult spider total for each sampling method for the 12 dominant spider families.

i Pitfall trap a 80 Sweep net X4J 0J Visual search B 70

(H OJ Ph 60 >r

a bO

: .:.

Clubiona sp.1 Clubiona sp.2 Lycidas sp. Maratus sp. Servaea sp, Clubionidae Salticidae

FIG. 4. Percentage of the adult spider total for each sampling method of Clubionidae and Salticidae. 480 MEMOIRS OF THE QUEENSLAND MUSEUM

spiders sampled by each method varies over the species of clubionids and salticids illustrates that 12 dominant families (Fig. 3). Ground dwelling the differential selection of taxa by sampling spiders such as gnaphosids or zodariids are ex- method also operates at the species level (Fig. 4). clusively caught by pitfall traps. Despite lycosids Within the Clubionidae, despite the two species and zorids being sampled by all methods, there is being collected by all three methods, Clubiona a reduced chance that they would be represented sp. 1 was almost exclusively sampled by pitfall by sweep net and visual searching over a shorter traps, whereas Clubiona sp.2 was more often survey period. If only pitfall traps were used, the taken by a sweep net or visual searching. Similar- probability of representing the families ly, the three dominant salticid species were Thomisidae or Salticidae is markedly reduced. preferentially sampled by the three different Linyphiids, due to their habit of building low methods. Hence, if one sampling method was webs under the foliage are primarily amenable to favoured over any other, especially for a shorter capture by visual searching. Consequently, the survey period, many species would be omitted results show that there is a greater probability of from the final species list. representing some families over others according to the method used. IMPLICATIONS FOR THE FUTURE The number of species in each family are dis- SURVEY OF SPIDER COMMUNITIES tributed differently across sampling methods (Table 3). Whilst the ground dwelling spiders There is currently no spider sampling technique such as the gnaphosids and zorids have all their that is unbiased. The success of any method is species falling into pitfall traps, other families usually related to certain aspects of spider behave a pattern of species distribution across haviour and therefore generally represents an in- methods quite different to the distribution of incomplete range of taxa. Whilst this may be dividuals. For example, thomisids may be best readily acknowledged by arachnologists, the sampled using sweep netting and visual searching limitations of a given method is not always (Fig. 3). Yet, these methods inadequately sample clarified in the interpretation of community com- all the thomisid species collected (Table 3). The position. This is particularly important when non- contrast is explained by the two most abundant arachnologists utilise the information as thomisids, Diaea sp. and Cymbacha sp. being representative of the whole community. collected mostly by sweep net and visual search Despite the use of three sampling methods in methods. Further examples include all clubionid this survey, the species list is not unbiased. In this species being sampled by all three methods (com- study, sweep net and visual search methods were pared to a very unequal distribution of in- carried out during the day and may therefore not dividuals) and a greater number of salticid select nocturnally active taxa. Visual searching species being sampled by pitfall traps (compared included looking at the ground, but effective sam- to this method catching the least number of sal- pling of leaf litter was not undertaken, and this ticids). micro-habitat can harbour distinctive families The number of individuals of the dominant (Raven, 1988). The visual search method is also is Pitfall Sweep subjective in terms of where the search focus Family Total Visual trap net directed. Attention was paid in this survey to

Amaurobiidae 5 5 1 sample spiders in positions (particularly in nests)

Clubionidae 5 3 3 3 that were not as vulnerable to the other two collecting methods. Where the objective of the sur- Gnaphosidae 11 I I vey is to estimate taxonomic composition of the Linyphiidae 6 6 1 6 spider fauna, the efficiency of both the visual Lycosidae 7 7 search and sweep net methods could be improved Micropholcommatidae 5 5 1 by avoiding the collection of distinctly immature

Prodidomidae 1 I (J spiders. Salticidae 14 9 7 6 Also, the equipment design can effect the num- Theridiidae 9 6 2 3 ber of individuals and taxa caught (eg., for pitfall

Thomisidae 8 8 2 3 traps see Luff, 1975 and Curtis, 1980). Temporal factors can further influence which taxa are sus- Zodariidae 9 9 ceptible to capture by a given method and as TABLE 3. Number of species in total and for the three discussed by Abraham (1983), this can be related sampling methods for 12 dominant families. to seasonal migration of spiders between vegeta- SAMPLING METHODS AND TASMANIAN SPIDERS 481

tive strata. To enhance the comparability of sur- Catalogue of Australia. 3. Arachnida'. (Australian vey data, there is therefore a need to standardise Government Printing Service: Canberra). methodology, equipment design, sampling effort DUFFEY, E. 1974. Comparative sampling methods for grassland spiders. Bulletin of the British Arach- and timing. Yen and Butcher (1992) also make nological Society 3: 34-37. this plea in respect of terrestrial invertebrate sur- GREENSLADE, PJ.M. 1964. Pitfall trapping as a veys for the ultimate goal of conservation. method for studying populations of Carabidae Methodological limitations need to be taken into (Coleoptera). Journal of Animal Ecology 33: 301 - account during the final interpretation of tax- 310. onomic lists for a more useful assessment of the KIRKPATRICK, J.B. 1977. 'The Disappearing Heath'. fauna. These aspects are stressed in the light of a (Tasmanian Conservation Trust Incorporated). rapidly growing reliance on such data sets for LOWRIE, D.C. 1985. Preliminary survey of wandering conservation and management in Australia, and spiders of a mixed coniferous forest. Journal of the need to critically assess invertebrate survey Arachnology 13:97-110. methods for estimating the loss of biodiversity LUFF, M.L. 1975. Some features influencing the ef- worldwide (Coddington et a/., 1991). ficiency of pitfall traps. Oecologia 19: 345-357 MERRETT, P. 1967. The phenology of spiders on ACKNOWLEDGEMENTS heathland in Dorset: I. Families Atypidae, Dys- deridae, Gnaphosidae, Clubionidae, Thomisidae Salticidae. Animal Ecology A survey of spiders of the north-east coastal and Journal of 36: 363-374. heathland was supported by the Plomley Founda- 1968. The phenology of spiders on heathland in tion at the Queen Victoria Museum, Launceston, Dorset: I. Families Lycosidae, Pisauridae, Tasmania, who also funded curation of half of the Agelenidae, Mimetidae, Theridiidae, Tetrag- spider collection. Specimens were identified and nathidae, Argiopidae. Journal of Zoology (Lon- the curation completed with resources kindly pro- don) 156: 239-256 vided by the Arachnology section, Queensland 1983. Spiders collected by pitfall trapping and Museum (Brisbane) and with taxonomic advice vacuum sampling in four stands of Dorset heath- from Mark Harvey (Nicodamidae), Rudy Jocque land representing different growth phases of (Zodariidae), Roily Mackay (Lycosidae), Robert heather. Bulletin of the British Arachnological Raven (all other taxa) and Marek Zabka (Sal- Society 6: 14-22 ticidae). Completion of the project has been poss- MERRETT, P. & SNAZELL, R. 1983. A comparison ible through an Australian Postgraduate Research of pitfall trapping and vacuum sampling for as- Award at Griffith University, Brisbane. sessing spider faunas on heathland at Ashdown Forest, south-east England. Bulletin of the British LITERATURE CITED Arachnological Society 6: 1-13. MITCHELL, B. 1963. Ecology of two carabid beetles, Bembidion lampros (Herbst) and Trechus quad- ABRAHAM.J. 1 983. Spatial and temporal patterns in a ristriatus (Schrank) Il.Studies on populations of sagebrush steppe spider community (Arachnida, adults in the field with special reference to the Araneac). Journal of Arachnology 11:31 -50. technique of pitfall trapping. Journal of Animal CODDINGTON, J.A., GRISWOLD, C.E., DAVILA, Ecology 32: 377-392. D.S., PENARANDA, E. & LARCHER, S.F. RAVEN, R.J. 1988. The current status of Australian 1 991 . Designing and testing sampling protocols to spider systematics. Pp. 37-48. In, Austin, A.D. and estimate biodiversity in tropical ecosystems. Pp. Heather, N.W. (eds) 'Australian arachnology*. 44-60. In Dudley, E.C. (ed.) 'The unity of evolu- The Australian Entomological Society Miscel- tionary biology: proceedings of the fourth interna- laneous Publication No. 5. tional congress of systematic and evolutionary biology'. (Dioscorides Press: Portland, Oregon). UETZ, G. E. & UNZICKER, J.D. 1976. Pitfall trapping CURTIS, D. 1980. Pitfalls in spider community studies in ecological studies of wandering spiders. Journal (Arachnida, Araneae). Journal of ArachnologY 8: of Arachnology 3: 101-111. 271-280. YEN, A.L. & BUTCHER, R.J. 1992. Practical conser- DAVIES, V. TODD 1985. Araneomorphae (in part). vation of non-marine invertebrates. Search 23: Pp. 49-125. In Walton, D.W. (ed.) 'Zoological 103-105.

A NEW SPIDER GENUS (ARANEAE: AMAUROBIOIDEA) FROM RAINFORESTS OF QUEENSLAND, AUSTRALIA

VALERIE TODD DAVIES

Davies, V. Todd. 1993 11 1 1: A new spider genus (Araneae: Amaurobioidca) from rainforests of Queensland, Australia. Memoirs of the Queensland Museum 33(2): 483-489. Brisbane. ISSN 0079-8835.

Malala gen.nov. is described and its relationships discussed; descriptions of two new species, Malala lubinae sp.nov. and Malala gallonae sp.nov. arc given, including their spigot morphology. Malala is considered incertae sedis within the Amaurobioidca (sensu Leh- tinen).[y\raneae, Amaurobioidea, new taxa, spigot morphology.

Valerie Todd Davies, Queensland Museum, PO Box 3300, South Brisbane, Queensland 4101, Australia; 28 October, 1992.

Many amaurobioid spiders in Australia no longer spin a web; they are usually vagrants either in litter or foliage and typically the forelegs have strong ventral spination. Two species of a tree- frequenting genus are described here. Abbreviations. Measurements: cephalothorax length (CL) and width (CW), abdomen length (AL) and width (AW). Eyes: anterior median (AME), anterior lateral (ALE), posterior median (PME), posterior lateral (PLE), anterior row (AR), posterior row (PR). Spinnerets: anterior (ALS), median (PMS), posterior (PLS). Collectors: R.J. Raven (RJR), N. Hall (NH), V.E. Davies (VED). On figures: ac aciniform spigot; als, anterior spinneret; at, anal tubercle; c, con- 1-5. Malala lubinae. cephalothorax ductor; cp, protuberance on 6 chelicera; cy, FIGS 1, 3, 9. 1, (lateral), 3, eyes and chelicerae (frontal), 2, 4, 5, 6\ cylindrical spigot; e, embolus; fb, frontal bristle 2, eyes and chelicerae (frontal), 4, chelicerae on 6 chelicera; fs, frontal spur on 6 chelicera; (ventral), 5, palp (ventral). Map, major ampuliate spigot; map, minor ampullate spigot; ms, median sclerite; n, nubbin; pi, Description piriform spigot; pis, posterior spinneret; tf, Straw-coloured ecribellate spiders with little or tegular flange; to, tarsal organ. All specimens are no pattern on abdomen. Carapace rounded dor- lodged in the Queensland Museum. sally, gradually declining in height behind fovea

(Fig. 1). Anterior row of eyes straight, posterior Malala gen. nov. row slightly recurved from above, procurved from front. Canoe-shaped tapetum in indirect eyes. Clypeus x 2 AME; narrow median sclerite Type Species above chelicerae (Figs 2,3). Sternum longer than Malala lubinae, sp. nov. wide, broadly truncate anteriorly, pointed posteriorly. Labium about as wide as long, a little Diagnosis more than half the length of the endite. Four-5 Three-clawed, pale clubionine-like with long retromarginal cheliceral teeth; proximal ridge 3-5 spinose legs. Eyes in almost straight rows; AME leading to promarginal teeth, 2 small medial much smaller than other eyes. Narrow median teeth between first and second marginals. sclerite above chelicerae. Chelicera with one or Without long retrolateral filamentous setae at two long frontal bristles proximally; pointed base of fang. Legs long (1243), trochanters un- frontal spur distally in 6. 8 palp without median notched. Coxae I longer than IV. Without apophysis; elaborate tibial apophysis with dorso- plumose (ciliate) hairs on body or legs. Numerous retrolateral element. strong ventral spines on tibiae and metatarsi I and 484 MEMOIRS OF THE QUEENSLAND MUSEUM

m? FIGS 6-12. Malala spp. epigyna. 6-9, M. tubinae, ventral, ventral (cleared), posterior, dorsal. 10-12, M. gallonae, ventral, lateral, dorsal.

II; metatarsi long, without preening combs; tarsi short. Trichobothria in a single row on metatarsi FIGS 13-15. Malala lubinae. 13, tarsal organ and and tarsi. Bothria grooved, collariform (Fig. 14); bothrium, 14, trichobothrial base, 15, 6, chclicera tarsal organ with pear-shaped opening (Fig. 13). (prolateroventral). Superior claws 7-8 teeth, inferior claw 2 teeth; 2 pairs of fringed accessory claw setae present; 9 wide 1:0.8; endite 1:0.6. Chelicera with 5 palpal claw 2-3 teeth. Colulus. Spinnerets: ALS retromarginal teeth and 5 promarginal; second much larger than PLS; one major ampullate proximal tooth largest. Chelicera with 2 stout (gland) spigot and nubbin on ALS. Epigynum frontal bristles; low retrolateral swelling near small with scape-like median structure; anterior base of fang; fang with slight ventral projection. gonopores. Male palp with spiniform embolus, Length of coxae 1 :IV is 1:0.6. Leg measurements conductor. Tibial apophysis with 3 membraneous are given in Table I. processes; the anterior process branched or un- Notation of spines. Femora: palpi, DO- 1-1; I, branched, the posterior ones 'toothed' and facing Dl-1-1, PO-0-2, RO-0-2; II, Dl-1-1, PO-1-2, each other. Respiratory system with 4 tracheal R0-M;III,D0-M,PO-2-l,RO-l-l;IV,Dl-l-l, tubes in abdomen; broad inner tubes branching PO-0-1, R0-0-1. Patellae: palpi, D0-0-1; Tibiae: and rebranching; slender outer tubes simple.

Etymology The generic name is derived from 'malal' an aboriginal word for spider.

Malala lubinae sp. nov. (Figs 1-9, 13-17,20-27)

Material Examined TYPES. Lamington National Park, southeastern Queensland. Holotype: 9, Nagarigoon, 8.iv.l976, VED, NH, S20349. Paratypes: Nagarigoon, 1 3, 8.iv.l976, VED, NH, S20350; 4 9, 2j, S20351; Binna Burra, 1

1 l-12.ii.1981, Y. Lubin, RJR, VED, S20353; 1 9 with spiderlings in sealed leaf, S20354; Mt Hob- wee, 1 9, 3-8.iv.1976, RJR, VED, S20355; O'- Reillys, 1

Description Female: CL 3.6, CW 2.6, AL 3.8, AW 2.5. FIGS 16-19. 6 palps. 16, 17, M. lubinae. 16, left palp, Ratio of AME:ALE:PME:PLE is 4:10:9:10; 17, tibial apophysis. 18, ]9,M.gallonae. 18, left palp, 19, tibial apophysis. width of AR is 2 1 , PR is 25. Sternum longer than NEW SPIDER GENUS 485

Q.imniS.ekU 1.06E2 8044/00 A44

®*j

1 ' /*Mfk ilVlap^w ' Jmm *72 .n

22 JP Mm ,|t.,- A44 1 ejiiiiis.iku 3.85E2 8002/00 0**15.0 JcU 177E3 0014/00 A13

FIGS 20-25. 9 Af. /hW/ki* spinnerets. 20, spinneret group. 21, 22, 25, right ALS. 23, PMS. 24, right PLS 486 MEMOIRS OF THE QUEENSLAND MUSEUM

FIGS 26-3 1 . Malala spp. spinnerets. 26, 27, penultimate 9 M. lubinae. 26, right ALS, 27, major ampullate spigot and nubbin on ALS. 28-31 6 M. gallonae. 28, spinneret group, 29, right ALS, 30, PMS, 31, left PLS. NEW SPIDER GENUS w,

palpuDl-O-LPl-l-O; I. P2-M.V7-6-2, Rl-I-l; prickly leaf blade of Calamus muelleri, the 11, P2-1-L V5-5-1, Rl-2-0 HI, P2-1-0, V2-1-2, lawyer vine.

R2-3-0; IV pl-1-1, VI- 1-2. Rl-I-l. Metatarsi: 1, -l,V4-3-LRM-l;HP2-1-LV4-2-3.RM- Malala gallonae sp. nov. 2-1. IV. 2;Ul,D2-2-2,Pl -0-LV2 RLQ-1; D2-2- !F.gs 10-12.18. 19.28-31) 2, Pl-0-1. V2-2-1. RO-1-1 Tarsi: palpi, P2-1-0, VO-0-2. Palpal tarsi swollen distally. Dorsal abdomen with slight foliate pattern. Epigynum very Material Examined small (Figs 6-9). TYPES. Holoiype: 9. Beilenden Ker Ranee. 1054 m. l7~K-24.Xt.1981, Harthwatch/ Spinnerets: ALS much larger than PLS I 20). ALS with one large major ampullale spigot Queensland Museum Expedition, S20357. Paiatvpes. Same data as holotvpc, I 5, S20358, i Map), a nubbin (n) and about 30 piriform (pi) spigots arranged in 2 groups (Figs 21,22, 25). The "S20359: 1 ?, S20360. 1 d\ same locality, major ampullatc spigot and nubbin are shown 25 31 X.I98I.S2036L 1 ^.Malaan State Fo. more clearly in a penultimate female (Figs 26. 20-4.iv.197S. RJR. VED. S 20362 Upper Boulder Creek, 11 km Tully, 850- 1000m, 1 PMS (Fig. 23 1 with large minor ampullatc NW spigot (map). 3 aciniform gland spigots (ac) and 2, 17-18.xi.l984.L Gallon, VED,S20363; I S, one cylindrical gland spigot Icy). PLS (Fig. 24) S20364; 1 d, 6.xii.89, C.B. Monteith, S20365. \\ ith one large cylindrical spigot and lOacinifonm All in northeastern Queensland. spigots. Female: CL 2.7. CW 2.0. AL 3.2. AW 1.9. Other females varied in size: CL 2 9-3.9, Cty Ratio of AME ALE:PME:PLE is 4:8;9:8. 2.3-2. 9, AL 3.S-4.4. AW 2.3-2.8. Chelicerae with single stout elongate frontal Male: CL 3.1, CW 2.3. AL 3.3. AW 2.0. Ratio bristles crossing each other. Four retromarginal. of AME:ALE:PME:PLE is 5:9:8:9. Chelicerae 3-4 promarginal teeth on chelicerae. Fang with divergent; 2 stout setae ((b) crossing over frontal- slight ventral projection. Length of coxae 1:P- i ly with lesser one above; pointed spur (Is) distal 1 OS- Leg measurements arc given in Table I, to these. Fbur teeth on retromargin of chehcera, Notation of spines. Femora: palpi, D0-0-1. I. 4 on promargin; conical protuberance (cp) near DIM, PO-2-1, R2-l-2;H,Dl -M, PO-2-1. Rl base of fang: fang long with marked ventral 2-1; III, DCV-l-1. PO-2-1. R0-2-1, tVf DI-M, projection (Figs 2.4, 1 5 1. Length of coxae I:IV is PO-O-L R0-O-1. Patellae: palpi, D0-0-L Tibiae:

1:0.7. Leg measurements arc given W Table I palpi. D1-0-l.P2-l4):LP2-l-1,V6-5-l,R2-l-l: P2-l-0 Notation of spines: Femora: palpi, D0-1-I: I. 1L P2-1-L V5-5-1. R2-1-I: III. t V0-2-0 y r DI-l-L P0-0-2. RO-l-1. It. Dl-M, PO-2-1, R0- Rl-2-0: I\ Pl-I-0, VO-1-0. Rl-I-0. Metatarsi: I. II. ] 2 111, DO- 1-1, PO-1-1, RO-1-1; IV. Dl-1-1, P2-1-1, V4-2-3, RM-2; P2-1-1, V4-2-2, R2- O-l;HI.D2-2-2.Pl-0-LV2-2-l.R1-0-l;lVD2- P(M)-1 , RO-0- 1 . Patellae: palpi. D0-O-1 . Tibiae: L

P2-1-1, V7-4-1, RM-1; II. P2-1-L V6-4-1, Rl- 2-2. P1-1-I, V2-2-I. R0-0-L Epigvnum i l-l;ni,P2-]-fi.V2-l-2.RI-l-l;IV.P2-0-l,Vl- 10-12) 1-2, Rl-I-l. Metatarsi: 1. P2-I-I. V4-2-2, Variation in size: CL 2 6-2.8. CW 2.0-2.1. AL Rl-l-I; II. P2-M. V3-3-2, Rl-l-I; III, DfM-L 3.0-3.4, AW 1.8-2 .5

P2^| f V3-2-2»Kl-l*l;1V;p0-M l Pl-M t V^ Male: CL 2.4, CW t.8. AL 2.9, AW 1.4. Ratio 2-2, Rl-I-l. Of AME:ALE:PME:PLE is 4:8:9:8. Chcliccrac 3 palp (Figs 5,15); tegulum extended to form slightly divergent; single enlarged setae crossing anterior prolateral flange; embolus, arising over frontally. pointed spur distally; without .rally on tfigUlUtt), tapers to apohlt; conductor protuberance near base of fang. Three retromar- membraneous with pointed sclerosed tip: cynv ginaL 4 promarginal teeth. Leg measurements arc bial tip shorter than length of tegulum. Tibia with given in Table 1 . Notation of spines is similar to 3 retrolatcra' processes, the anterior one £ M. lubinae, branched. 6 palp (Figs 1 St 1 9): tegulum without prolateral Variation in size CL 3.1-3.3, CW 2 4-2.5. AL flange, embolus arising prolaterally. Cymbinm 3.6-3.8. AW 2.1-2.2. tip a little shorter than length of tegulum. Tibia with 3 reirolaicral processes, the anterior one wiih Biology 2 branches. The spiders were collected from branches or Spinnerets: ALS (Fig. 29) have a large major foliage of trees in the rainforest. The egg sac was ampullatc spigot, a nubbin and about 20 piriform placed with the spider in a sealed portion of the spigots. PMS (Fig. 30) have a large minor ampul- :

MEMOIRS OF THE QUEENSLAND MDSEUM

TAULE I. Leg measurements (mm) of Malata spp. Desidac: Dictynoidea) has simple median brinks * Bowed. thus invalidating the branching as a synapomorphy for the dictynoids. Gray's evidence suggests Ug r- "' i Nlc Ta TpuI thai ihe branching may be important only at the M. lubina? generic level in the same way as the extension of - 1 ? 4.2 9 palp 1.5 05 D.4 the trunks into the cephaJothorax is regarded. In

1 .1.5 i : 3JJ 31 0.9 12.6 their recent paper on the higher systematic* of li 2.7 1.2 3.0 2.6 !..;; 10 1 Araneae, Coddington and Levi 0991) state that III 2.0 10 1.6 2.0 0.7 '-3 'these superfamilies are among the largest cladis- IV 2.6 1,0 23 2.9 0.9 97 tic problems at the family level* and that the 6 p'ilp 1.5 0.5 07 - 1.2 3.9 difficulty of definition 'partly stems from

1 3.2 1.2 3.7 3.1 1 12.: heterogeneity within families'. I would add that

i M II Z.5 2 5 \ 5 0.9 9.7 the few Australian genera on which some ofthese families are based compared with the large num- 111 1.9 u u IW 6.4 ber yet to be described, also contributes to this rv £.3 P*J ii M H.6 difficulty. Af. xaitonae

- Vpaip ! ! 05 0.7 i u 3.3 Most Australian rainforest cribellatcs which

O.tj I 3.1 1.0 3.6 -.1 il.7 range in size from small to very large have a

11 2.6 1.0 2.9 2.6 O.S 9.9 nondescript but recognisable abdominal pattern (see Lehtincn, 1967: divided cribellar in 1.8 U.S 1.5 1.9 0.6 6.6 435), a field, a single row of tarsal trichobothria increas- IV 23 UK 2.0 2.5 0.8 M ing length distally a of metatarsal ' in row - and <5 p^ip II IJ -1 0.5 1.2 l trichobothria. The palp has a complex tibial : $ i 2,9 1.0 3.9 3.1 0.9 :! apophysis and a median apophysis is usually ii 2.5 0.9 2.7 2.5 0.8 M present. I am persuaded io return to Lehtinen s in IS 0.7 1.3 1.8 0.6 t". classification and regard these Australian spiders IV 2.0 M7 J.B 2.3 0.7 75 as belonging in the Amaurobioidea.

Only one dictynid, Callevophthalmus is late spigot and 3 acini form spigots. PLS {Fig. 31) described from Australia It is found in grass have 10 acini form spigots. and open forest and not, so far as I know, m

Variation in size: CL 2.0-2.8, CW 2.0. rainforest It is a small spider with a distinct abdominal pattern and an entire cribellar field. It DISCUSSION has no tarsal trichobothria and a single metatarsal trichobothrium. The 6 palp has a small tibial Malnla. a 3-clawed ecribellatc, has a single row apophysis, a large T-shaped conductor extending beyond the edge of the retrolaterally of tarsal trichobothria of increasing length distai- cymbium has no median apophysis. The chelicerae ]y and a single row of metatarsal trichobothria. and 6 prolatcrally The 6 palp has a complex tibial apophysis and are indented and have a proximal frontal processMhese modifications do iHttai u:;n no median apophysis. The £ chelicerac diverge 'i distally and have a pointed frontal spur distally. io be homologous with those of Mala/a Col The median tracheal trunks are branched- The levophtluilmus is similar to Arangina from New Zealand and to Dictyna. anterior lateral spinnerets have one major ampullate gland spigot and a nubbin. From these and This is the first description of spigot morphol- other characters Malala clearly belongs in the ogy in an ecribcllatc amaurobioid. In general, amaurobioid/dictynoid complex of families. In cursorial spiders use silk only as a drag-line, and his world revision of cribcllates, Lchti nen (1967) forthe construction of the cocoon in the 9. Thus, recognised one major superfamily Amaur- it is expected they will have fewer kinds of ohioidca with 6 families. Describing New spigots than the web-spinning (cribellate) Zealand spiders- Forster (1970) and Forster and amaurobioids Maiala has 3 acini form spigots on Wilton (1973) defined two groups, Dictynoidea the PMS and 10 on the PLS. The typical small and Amaurobioidea, based on the branching or acini form spigots which providesilk for swaihiug non-branching respectively of the median prey (Kovoor, 1987) are not present in MuUila. ncal ruitiks. However, Gray (1983) found that The lack of these in contrast to their abundance Forsterirw, an acknowledged close relative o\' in cribellate amaurobioids and dictynoids (Cod Badtunna (classified by Forster and Wilton in dington. 1990b) is probably associated with NEW SPIDER GENUS 489

Malala? s nomadic existence. The presence of one LITERATURE CITED major ampuJlate gland spigot (Map) and a nubbin on the ALS is found in diciynoids (notably in CODDINGTON, J.A. 1990a. Ontogeny and homology Dictyna) and in the Orbiculariae (Coddington, in the male palpus of orb-weaving spiders and 1990a) as well as in some other taxa (Platnick et their relatives, with comments on phylogeny al„ 1991). The nubbin (presumably a remnant of (Araneoclada: Araneoidca, Dcinopoidea). Smith- a second major amputate spigot) may be sonian Contributions to Zoology 496: 1-52. homoplasious in Malala, 1990b. Cladistics and spider classification: The fringed accessory claw setae present in arancomorph phylogeny and the monophyly of Malala are probably an adaptation to tree- or orbweavers (Araneae: Araneomorphae: Or- biculariae). ActaZoologica Fennica 190: 75-87. foliage-dwelling and are similar to those found in some New Zealand spiders (Forster, 1970; figs CODDINGTON, J.A. & LEVI, H.W. 1991. Sys- tematica and evolution of spiders (Araneae). An- 28-3 1 ) from a similar habitat. They are unlike the nual Review of Ecology and Systematic* 22: serrate setae which are involved in handling silk 565-592. in araneoids. The presence of a narrow median FORSTER, R.R. 1970. The spiders of New Zealand, sclerite above the chelicerac in Malala may be Part HI. Otago Museum Bulletin 3: 1-184. apomorphic. Marples (1962) mentions that a tri- FORSTER, R.R. & WILTON, C.L. 1973. The spiders angular sclerite is present in Mawchia but not in of New Zealand. Part IV. Otago Museum Bulletin Paramafachia. 4: 1-309. At present 1 am unable to place Malala in a GRAY, MR. 1983. The taxonomy of the semi-corn- family and consider it incenae sedis within the muna! spiders commonly referred to the species, Amaurobioidea (sensu Lehtmen). Ixeuticus Candidas (L. Koch) with notes on the genera Phryganoporus, Ixeuticus and Badumna ACKNOWLEDGEMENTS (Araneae, Amaurobioidea). Proceedings of the Linnean Society of New South Wales 106: 247- 261.

I am grateful for the support of the Council of KOVOOR, J. 19S7. Comparative structure and his- the Australian Biological Resources Study for tochemistry- of silk-producing organs in arachnids. funding the survey of rainforests during which Pp 160-186. In, Nentwig, W. cd. 'Ecophysiology of Arachnids'. (Springer- Verlag:BerJin). some of this material was collected and forfinan- enbdlate cially supporting Chris Lambkin who did the LEHTINEN, P.T. 1967. Classification of the spiders and some allied families, with notes on the lay-out for the figures. I acknowledge the help evolution of the suborder Araneomorpha. Annales given by Earthwatch and the Center for Field Zoologici Fennici 4: 199-468. Research, Boston, Mass., U.S.A. for supporting MARPLES, B.J. 1962. The Matachiinae. a group of the Queensland Museum's expedition to Mt Bel- cnbcllatc spiders. Journal of the Linnean Society lenden Ker. I of thank Don Gowanlock the of tendon 44: 701-720. Electron Microscopy Centre, University of PLATNICK, N.I., CODDINGTON, J.A., FORSTER, Queensland for the scanning electron R.R. & GRISWOLD, CE. 1991 . Spinneret mor- micrographs and staff of the Queensland phology and the phylogeny of haplogyne spiders Museum for their willing help in the preparation (Araneae, Araneomorphae). American Museum of this paper Novitatcs3G)6: J -73.

AN INVENTORY OF THE SPIDERS IN fWO PRIMARY TROPICAL FORESTS IN SAB AH. NORTH BORNEO

CHR1STA L. DEELEMAN-REINHOLD

Deeleman-Reinhold, C.L. 1993 11 1 1 ; An inventory of the spiders in two primary tropical

forests in Sabah, North Borneo. Memoirs of die Queensland Museum 33(2 1: 491-495. Brisbane. ISSN 0079-8835.

Collecting trips were made to a primary rainforest area at 1500- 1900m altitude (Mt Kinabalu National Park) and a primary lowland rainforest (Danum Valley Field Centre) in Sabah, North Borneo. For comparison, a strongly degraded secondary forest in the town Kola Kinabalu was also sampled. All the material, with the exception of the mygalumorphs and salticids, has been identified and compared with collections from Sarawak, Kalimantan and

Sumatra. 254 species were distinguished in approximately 1 20 genera, 35 could be identified as known species, seven of which were clearly synanthropic, the rest are undescribed. 207 species were found in one locality only: 85% of the species from Kinabalu, 70% of the species of Danum and 50% of the species from the town park. Widespread species were found mainly in the Araneidae, Pholcidae, Oonopidae, Clubionidac and Salticidae. A list of the genera and species is given. {jBio diversity, rainforest. Asia, Aranear.

Chnstu L. Det'leman-Reinhold, Sparrenhmti $, 4641 GA Qswruinrht, The Netherlands; 8 April, 1993.

Tropical rainforests, covering only 6% of the linen, 1979, 1981. I9S2; Levi, 1982, 1983; Earth' s surface, are believed to harbour more than Okuma, 1988; Platnick and Murphy, 1984; Wan- half of all terrestrial animal species, of which less less, 1987). than 10% arc described at present (Stork and In a privately undertaken program of inventory- Ciyston, 1990). Inventorying the spider fauna of ing spiders of primary and secondary forests tii rainforests in south-east Asia has given spider south-east Asia, during the Use 14 years 1 have taxonomy a new turn, especially so after the been engaged, with the help of other, partly introduction of a new sampling method that tar- autochthonous collectors, in surveying the spider gets canopy arthropods. fauna of south-east Asia, mainly Indonesia, Most rainforests in Asia have now been Malaysia, Thailand, Sri Lanka and the Philip- destroyed or degraded, but the number of un- pines. As part of this initiative, I made three in the part described species is still overwhelming. The collecting trips to Sabah north-eastern rapid destruction of our rainforests is an incentive of Borneo. k to securing as much data as possible as ... an extensive program of inventorying aimed at es~ METHODS timating diversity of species ... is essential for a fuller understanding of the role of biodiversity in InJune 1979, July 1980 and April-May 1991, ecosystem function' (Coddington a a/., 1992). collected spiders in the primary rainforest of Records of spiders from Borneo arc extremely Mount Kinabalu National Park at altitudes of poor. With more than 30,000 species of spiders 1500- 1900m, In May 1991. 2 days were spent described so far, only about 160 named spider collecting at an altitude of 50Om (Poring Hot species have been described or recorded for the Springs). In May 1 99 1, spiders were collected in whole of Borneo; 65 of these are saiticids. Only lowland primary forest around the Danum Valley 96 spider species were reported from Borneo Field Centre in eastern Sabah. For comparison, before World Warl.Wanless and Hillyard(19K4) some, time was also spent collecting in the town present a list of species collected during the park in Kots Kinabalu. araehnological survey of Gunuug Mulu National The spiders were collected by hand pirfrj Park, Sarawak, 360 species were collected in all sweeping, litter sieving and pitfall trapping on the families, 38 of which were identified with known ground. All araneomorph spiders, with the excep- species, another 14 with reserve, 20 identified tion of the Salticidae have been identified (Tables species are salticids. From Sabah. a mere 20 1-3). The collected spiders were compared with spider species arc known, all published since most specimens of the above mentioned south- 1979 (Deeleman-Reinhold, 1980. I9S7; Lch- cast Asia collection. Identification was done as )1 ) ) 1 ) 1 ) ) ) 1 ) )1

492 MEMOIRS OF THE QUEENSLAND MUSEUM

Oormpidat* Otucilia sp. also at ;>* (1J, 500m Meotipa $p (] Dysderinasp. (I) fofleatAs sp. (] ), Sabah F'lujn>ncidia sp. (2), Sabah Gumasomorpha sp, (2», Sabah sp.fh sp. in*

sp. (2) Teuiamus sp. ( I Theridicn sp. (4 lschnathymi\ sp. (4) OrthabttiasfrOl Sabah 5pp. (3)*, 1 on canopy walk Opopaea ? Corinninoe (1) Undescribed genus (2) Orchestifia sp. (1) New genus ll) sp. 1 1 >, also at 500m sp- (1), Sabah, only below 600m Gnaphimdae Undescribed genus f 1 Ptectopilux sp. (I Jacaena sp- (11, on the lawn Mimetidae Xyphinus Icmmscaius Deeieman Palpimanidae Mimetia sp ( 5p.(l).» Bougrius 3) sp. (1) Theridiosomat idae Undcscribed genus ( ] ), also lower in RAcAfl Atdiirndwe Piito sp. ( 1 dary forest Att£tt0 Sp. ( 1 i Thvridiosomti sp, (I I Te*r»blemmidi»c Mysmeniriae AbieMttta MuUnellu sp. [3) I Jndcscribcd genu 3 Bometftjjma Undescribed genus'' U) U ) Sabphya Thnmistdae Tetragnalhidue Leucauge cflebesiuna Walckcnaer. circunuKptctanxDcclcinan. also lower in Borbnropaaus sp. ( I i widespread secondary forest Lycoptts sp. ( 1 ). also ai 5(X)m roherti spp.(i),Sa^ah Deeieman Iwfwntnopf Sp, I 1 >, canopy walk*

GJenognatha sp. ( kinabaluann Deeieman sp. (1) 1 bispinosa Deeieman Phrynumchne sp. (1) Af*yu/asp.(4)

Ovyopklae Undescribed genus 1 1 2)

Undescpbed IT j. also Ochyrnc t ra I id . if '.(J sp. (1). canopy walk* genus {2 at 500m

PsUpdenxs sp. ( 1 Pi sun rid ae Undescribed gequs 1" i U, abp al 500m

Spcit-era sp- i ) Polybara rone idae | sp. (I) A Undcscribed genus (I Lycosidae Ara/vitvsp, (1 StyU'didiH* Pardmu sp ( | | Argiope remwunJiii Dolcschull, widespread Scvtcdei 11 pallida Dolescliall, Widespread, Undescribed gfcnbi I aemuhi (Walckcnaer), widespread

Pholcidae Hippasinae 1 1) Cyclosa bifida Doleschall, widespread Vlhina sp. ft), Sabah llahnidue Cyrtophora sp. \ 1

Aliiira sp. ( I > ? Eriophora Sp, (1)

Spentufphoru sp. (1 >, Sabah Huhnta i2) Costeraointhii sp. U »* miser Bristowc, widespread* Hcrsilhdae MUottia brevipes There IK widespiead

Belisana Sp. (1) Henilia sp 1 1 Ntr.K-'-'-'na nuututi L. Koch, WOfU Iropici Undescribed genus ( 1 1 heridiidjit' Undescribed genus (1 HeUrnpodidae Actoforu/ieu iL. Koch), munduh wide- l.inyphiidae Hcterapcxiasp. (I) spread Sertdiic heccam l"horell, widesprrad sp. { I jcanopy walk* tepida riorum (C.L. Koch), worldwide Kuala sp (J) Pnimcoplt sp, (1) i-p(l) 1* Punimtionetusp.(\) GHuwsp.(J)* upp. (3 Vonievrtrsp. (3) Undcscribed genus ( 1 ), in grass Atithximux sp. (1 ), canopy walk*

Undescribed genus [1 ). canopy walk* Argyrafa \iphia> Tboreu, widespread

i bribed genus II (|) Clenidac Kkomphoca sp. ( 1

Undescribed genus 111 (J I GteniiSBp, (I) Chnsr.o Sp. ( I I Undescribed genus IV (|) Clubionidae s.I. Coleosoma sp. (2), Sabah

Undescribed genus ( Clubioninne Ciwcifutfa sp. (J) V 2) Uloboridae Cheiracanlhiuni sp { ( sP u>- Clttbuma sp. (4) Dipcena sp. (5) PhiUtponelUi sp. (U sp. (1) canopy wrfl ' Sp '!")* b'hborus [ugubrh Thorell, widesprr.Kl* '-> sp.(t>widesp i Efti&itOtssp' Fstvhridue PhrurolitliiimL Junula sp, ( 0. Sabali Pxechnts fcitwbalt* Levi

Table 1, Spiders from Mount Kinabalu, 1500- 1900m (Headquarters and Power Station) and 500m (Poring Hot Springs), primary rainforest, 18 collecting days in April-May, June and July. Family order is phyiogeneuc*. List gives no, of undescribed species in parentheses and notes on species. *= only at 500m much as possible with the aid ofmodem revisions but, guished. Of these, 35 species could be identified where these do not exist. I had to rely on die keys in as described species, seven of which are clearly

Simon ( 1892-1 903) and the Latin descriptions (with- synanthropic. out illustrations) of Thorell (1877-1899) and Simon. On Mt Kmabulu (1500- 1900m). 135 species Many nineteenth century types deposited in Genova, were collected in 18 days (41 species represented Paris and London were studied. Only species of by one specimen only); 25 species were collected which adults were collected are considered here. in two collecting days at Poring Hot Springs, lower down on the mountain slope at 50O-6O0m; RESULTS four of these were shared with the 1500-1 900m site (see Table 1). For six of the 19 described and From the three main prospected localities in named species this is the tvpe locality Sabah, a total of 254 species from most spiders (Dceleman-Reinhoid. 1980, 1987'; Levi, 1982). families (for practical reasons the mygalomorphs 132 species (85%) were collected only in and the salticids were excluded) could be distin- Kinabalu; 24 species were also found elsewhere. ) 1 ) ) ) ) 1 )1 ) )1 )

FOREST SPIDERS OF SABAH, NORTH BORNEA 493

Oonopidae Sfeflfeiittr sp. 1 1 ), Sabah 0.:\(.iriuhi fcp. ( 1) IhKdervm >p. (1) Corinninac Uipttena ny (2) sp. Gamasomorplta (2), Sabuh Undescribed genus ( Episintis 1 sp. {ft Ischnottrvrrus peltifer (Simon) world tropics P&lpimonida* Swiiife vp. (I). S*hab sp (5) Bcagriiis sp. (I) Thcrvlion sp. l4) Upopaea ? sp. (11 Sabah Zodariidae

Undescribed genus ( I Orrhevina sp. yl) Malinetla sp. (2) Flcrtopilus sp. {1}, Sabaii Thumisiidae Mimelidat

Xyphinus sp. (1) Borbotwparats sp. (1 AfjfaM&usp. (1) Tit rah It/in ii id ai' Loxobates sp. (1), in logged area Mysinenid&e 1

Al'temma sp. (I) ( in taenia sp. '), togged area Undescribed genus ( 1 Othyroceraridaie Pentraensspil) AnapidiM? Synemasp. (1) 1 Pxeiidanapis parorulu< Simon, widespread ,V/'e

Glenognatha sp. ( SmertngopUs paitidtii (Blackwall), world Potybaea sp. (1) !

tropic-i Oxyitpidac Aruncidne J f fiolcus sp. (23 i Inn logged area O.xyupes- Imeatipea C.L. Koch, widespread Cae.rostrLs sp. (1) rmaptwra (1). "sp. $p£ sp. Sabuh ) Cyclasa bifida DoJescltal), widespread Bttlisana lapponia Thorell, sp. (1) mperba widespread mulmeinfiisis (Thorell), widespread Heleropodidae Lycosidse Gusiewivnihu 5p. (I J Hetempoda sp. (1 Gaa subarmatu Thorell. widespread sp. ( 1 ), in logged area. Wiuhcosa himtcinvo iTboreH), widespread, Larinia phih'vtU-a L. Koch, widespread CWpj sp. ( 1 j in logged area; J ClenioW i ur,ioKapULU'ola (Thare)l), wjilespread Milotna irifaaetata Thorell, widespread Ctenmsp, 0) Hahniidac Neoscona nausica L. Koch, world tropic. Gnaphosiri&e Attsiru sp, (1 sp- (I) Mkythus sp. (l)wiUi- 1 HerKtliidue /V/^sp.fU Clubioiudae si. HersUiasp, {)) Undescribed genus 1 (1) Clubkminae Thtrridiidat Undescribed genus n f I ChtiracatUhium sp. ( Acluitunmeu sp, ( I 1 Unyphiidav Chthio'Ui iij sp, Argynxlex sp. ( I Undescribed genus (I) t'astianeirinae Cephulobares sp. 1 1 AmHus sp. (1 Chrysst) sp. (1) LllfvhrH-idae

ftnuolflhinac ColcasotyQ Sp [] i, Sabab PhikponvlLi Sp, [23

Tabic 2. Spiders from Danum Valley Field Centre, primary lowland forest, 8 collecting days in May; some species,

mostly Thomisidae ? in freshly logged area. Family order is 'phylogenetic' . List gives no. of undescribed species in parentheses and notes on species.

Compared to the lowland catches, a the lack of data only, or is a high percentage of predominance of Linyphiidae was found. endemic species real? This phenomenon oc. iin Tn primary lowland forests around Danum Val- much more frequently in some families than in ley Field Centre in East Sabah, 90 species were others. Quite often, in adjacent localities a sister collected in 9 days (Table 2); 14 species have species is found. In a long-term inventory of a 1-2 2 been previously described. Of these, 67 species lorn area on the northern side of the Sibolangit (70%) were only found at Danum, and 23 were range, on Gunung Leuser in Sumatra (Deeleman- also found elsewhere. Jn a freshly logged area, Reinhold, unpublished data), spiders were col- thomisids were particularly diversified. lected once a week for two years. A similar study In the secondary forest of Signal Hill in the was conducted on the other side of the ridge. Les> township of Kota Kinabalu, 16 species were col- than half of the species were found on both sides lected, 7 of which could be identified to species. of the range! Therefore, endemism in spiileis Eight species were found also elsewhere, and 8 seems characteristic of primary rainforests, even species (50%) were collected only on that site. though the real extent of distribution ranges will only be revealed after long and extensive sam- DISCUSSION pling. For example, recent studies on south-cast Asian Linyphiidae (Millidge and Russell-Sntiili, The main conclusion is that in tropical forests, 1992) report 27 species, 26 of which new, spider species known from only one locality are described in 1 1 new and four known genera; .'Ii enormously preponderant even though all dis- new species were recorded from only one locality tribution types from cosmotropical to very (see also Scharff, 1992).

restricted ranges were encountered. Also ? widely distributed species were often In a total of 254 species from the three localities found in human-made habitats. In such habitats (Tables 1-3), 207 were collected atone locality most species described in the last century were only, 92 of which were 'singletons'. Is this due to found. In the course of identifying large south- ) ) 1

494 MEMOIRS OF THE QUEENSLAND MUSEUM

Oonopidae Psilochorus sp. (l)widespread sp.CO (Simon), world iropics Ctenidae Ischnothyreus peltifer Tetragnathidae sp.(D Ctenus sp. (1) Ptectoptilus sp. (1) Chibionidae s.l. Leucauge sp. (1) Oedignatha scrobiculata Simon, widespread Ochyroceratidae Araneidae Psiloderces sp. (1) Palpimanidae punctigera Doleschall, widespread Theotima minutissima (Petrunkevitch), Boagrius sp. ( 1 Neoscona Theridiidae world tropics Uloboridac Pholcidae Janula sp. ( 1 Uthina luzonica Simon, widespread Theridion tenuissima Thorell, widespread Uloborus humeralis Hasselt, widespread

Table 3. Spiders from town-park Signal Hill, Kota Kinabalu (2 collecting days). Family order is 'phylogenetic List gives no. of undescribcd species in parentheses and notes on species. east Asian collections it appeared that the Some genera have been particularly speciose in majority of the species described prior to the early primary forest. In Ischnothyreus I found 1 20th century occur in habitats created by humans species in Sabah (10 undescribed); in Theridion rather than in the rainforests. Thus, the spider 11; in Dipoena 8; and in Clubiona 8 (all un- fauna of the latter is still almost unknown. described). A high degree of endemism seems to occur in One final remark on diversity. Among the certain families; other families which include a strongly represented families, diversity in the fol- relatively high number of widely distributed lowing families appears to be higher than species are Araneidae, Gnaphosidae, Oonopidae, average: Pholcidae, Clubionidae s. lat., Tetrag- Pholcidae and Salticidae. Occasionally, one or nathidae, Araneidae, Linyphiidae. two species in a family are able to disperse con- This study indicates that, when replacing siderably, whereas their relatives have remained primary forest by secondary plantations, the loss limited to a restricted area. Among the best dis- of species diversity of spiders is enormous. perses are some of the smallest known litter- dwelling spiders, with a body length of less than ACKNOWLEDGEMENTS 1mm, which independently seem to have developed methods to overcome the vicissitudes Mr. Mh. Yusof of SERU, Kuala Lumpur kindly of ballooning, e.g. the tiny armoured anapid provided me with a research permit for 1991. Pseudanapisparoculus Simon is distributed over AKZO Resins, Bergen op Zoom donated the al- much of tropical south-east Asia both in primary cohol for preservation. and secondary forests. The small ochyroceratid Theotima minutissima (Petrunkevitch) and the LITERATURE CITED oonopid spider Ischnothyreus peltifer (Simon) are distributed over the palaeo- and neotropics, CODDINGTON, J. T HAMMOND, P., OLIVIERI, S., congeners. where they live side by side with local ROBERTSON, J., SOKOLOV, V., STORK, N. Also larger spiders have been found to be widely & TAYLOR, E. 1991. Monitoring and inventory- distributed in humid forest, such as some ing biodiversity from genes to ecosystems. Pp. Cyclosa, Argiope, Acusilas, Neoscona and 83-117. In, Solbrig, O. (ed), 'From genes to Gasteracantha species, but also the delicate, al- ecosystems: a research agenda for biodiversity.' (IUBS: Paris). most transparent pholcid Calapnita venniformis DEELEMAN-REINHOLD, C.L. 1980. Contribution to Simon. the knowledge of the southeast Asian spiders of The number of small-range species in both the families Pacullidae and Tetrablemmidae. primary and secondary evergreen forests seems Zoologische Mededelingen 56: 65-82. to be enormously higher than we are used to in 1987. Revision of the genus Xyphinus Simon temperate climates. Very few wide-spread (Araneae: Oonopidae). Acta Arachnologica 35: species seem to occur naturally on Mount 41-56. Kinabalu; more were found in lowland forest. LEHTINEN, P.T. 1979. Spiders of the Oriental- Australian region I. Lycosidae: Venoniinae and It is premature to estimate the total number of Zoicinae. Annales Zoologici Fennici 16: 1-22. species present. Richest in species probably is the 1980. Spiders of the Oriental-Australian region III. family Salticidae. Also numerous in species are Tetrablemmidae, with a world revision. Acta the Theridiidae, Oonopidae, Araneidae, Zoologica Fennica 162: 1-151. Clubionidae s.l. and Tetragnathidae in that order 1982. Spiders of the Oriental-Australian region IV. (see also Wanless and Hillyard, 1 984 for Gunung Stenochilidae. Annales Zoologici Fennici 19: Mulu). 115-128. FOREST SPIDERS OF SABAH, NORTH BORNEA 495

LEVI, H.W. 1982. The spider genera Psechrus and one by one. New Scientist ,1 1 August 1990: 43- Fecenia (Araneae: Psechridae). Pacific Insects 24: 47. 114-138. THORELL, T. 1877. Studi sui Ragni Malcsi e Papuani

1983. The orb-weaver genera Argiope, Gea and I. Annali del Museo Civico di Storia Naturale Neogea from the western Pacific region Giacomo Doria, Genova 10: 341-634. (Araneae: Araneidae, Argiopinae). Bulletin of 1878. Studi sui Ragni Malcsi e Papuani II. Annali the Museum of Comparative Zoology, Harvard del Museo Civico di Storia Naturale Giacomo 150:247-338. Doria, Genova 13: 1-317. MILLIDGE, A.F. & RUSSELL-SMITH, A. 1992. 1881. Studi sui Ragni Malesi e Papuani 111. Annali Linyphiidae from rainforests of Southeast Asia. del Museo Civico di Storia Naturale Giacomo (Araneae). Journal of Natural History 26: 1367- Doria, Genova 17: vii-xxvii, 1-720. 1404. 1890a. Studi sui Ragni Malesi c Papuani IV, 1. Annali del Museo Civico di Storia Naturale OKUMA, C. 1988. Five new species of Tetragtujttot Giacomo Doria, Genova 8: 1-419. from Asia (Araneae: Tetragnathidae. Esakia 26; (2) 1890b. aliquot in 71-77. Diagnoses Arancarum novarum lndo-Malesia inventarum. Annali del Museo PLATNICK, N.I. & MURPHY, J.A. 1984. A revision Civico di Storia Naturale Giacomo Doria, of the spider genera Trachyzelotes and Urozelotes Genova 30: 132-172. (Araneae, Gnaphosidae). American Museum 1 892. Studi sui Ragni Malcsi c Papuani IV, 2. Annali Novitates 2792: 1-30. del Museo Civico di Storia Naturale Giacomo SCHARFF, N. 1992. The linyphiid fauna of eastern Doria, Genova 31: 1-490. Africa (Araneae: Linyphiidae) - distribution pat- WANLESS, F.R. 1987. Notes on spiders of the family terns, diversity and endemism. Biological Journal Salticidae 1 . The genera Spartaeus^ Mintonia and of the Linnean Society 45: 1 17-154. TaraxelltL Bulletin of the British Museum of SIMON, E. 1892-1895. 'Histoire Naturelle des Natural History (Zoology) 52: 107-137. Araignees' vol. 1. 2nd edition. (Encyclopedic WANLESS, F.R. & HILLYARD, P.D. 1984. Arach- Roret: Paris). nological notes from Gunung Mulu National Park

1897-1903. 'Histoire Naturelle des Araignees' vol. with a list of the spiders recorded from Borneo and 2. 2nd edition. (Encyclopedic Roret: Paris). a preliminary list of the harvestmen of the Park. STORK, N. &GASTON, K. 1990. Counting species Sarawak Museum Journal 51 (ns) 30: 53-64.

A REVIEW OF FACTORS INFLUENCING THE DISTRIBUTION OF SPIDERS WITH SPECIAL REFERENCE TO BRITAIN

ERIC DUFFEY

Duffey,E. 1993 ] I 1 1: A review of factors influencing the distribution of spiders with special reference to Britain. Memoirs of the Queensland Museum 33(2): 497-502. Brisbane. ISSN 0079-8835.

Available data on factors influencing spider distribution are synthesized using mainly British and other European information, The importance of landscape history is stressed, in frag- menting ancient natural habitats and creating new ones. It is postulated that this has reduced the number of specialists and allowed a range expansion of pioneer and euryoecious species. A tentative classification of life strategies is proposed. Habitat diversity and microspatial distribution are discussed with examples. OSpiders. hubiiat pfejerencei aphic cita- tribution, life strategies.

Eric Dujfev, Cergnc House, Wadenhoe. Peterborough. PEH 5ST. United Kingdom; 3 November 1992.

Many attempts have been made to categorise within a restricted area some species are found the habitats of spiders according to their preferen- only, or mainly, in 2 very different habitats, foi ces for dampness, dryness, shade, light, warm or example sand-dunes and marshes; (c) some cool temperatures. In addition, increasing species are so widespread that they are difficult knowledge derived from extensive collections to characterise in terms of habitat preferences: 1 mack during the last 50 years in Britain and the species confined to only one specialised habitat in rest of Europe have made it possible 10 dmw are generally few number and often rare. Ex- distribution maps (Locket eta!., 1974; Ransy and amples are presented in this paper- Baert, 1%5. 1987a, 1987b, lansscn and Baert, Nomenclature follows Roberts (1985-87) to- spiders and Anon. (1964-80) for plants. 19S7; Ransy et al. T 1990; Alderweireldt and Macifait. 1990; Canard, 1990), even though their [jlCOmpletene&S is acknowledged. The distribu- LANDSCAPE HISTORY AND PATTERN tion of some species shows a clear influence of OF SPIDER DISTRIBUTION latitude and longitude, with north-south or east- west trends, or an association with major habitat Bri'.ain was largely forested before human set formations. This paper discusses some factors tlemcnt but forests are now much modified scat- which influence spider distribution, such as tered remnants. This extensive environmental history and geographic range, habitat landscape change will have favoured some species and dis- preferences, adaptation to man-made biotopes advantaged others. For evample, when the sur- and small-scale environmental differences, with viving rare species associated with major the object of categorising the occurrence of formations in Britain arc listed (Table 1 ) most are species according to life strategy. recorded for coastal systems which-apart from Islands of modest size such as Bntain are the more limited mountain tops-arc the li poorer in species than adjacent continental areas. modified components of the present-day For example, the spider fauna of Belgium, a near landscape. Species which arc regarded as 'char neighbour, which is only 1/Sth the size of Britain acteristic' of major habitat formations (Ram and has 75km of coastline compared with 1977) (Table 2) may also be identified. The high Britain's 19000km, and a lower landscape diver- numbers recorded for grassland, dry heaths and sity, has almost as many recorded species (600] coasts probably reflect a great expansion of I species forest iKeckenbosch et al a 1977) as Britain's 622 by open-ground after clearance. kebens, 1985-87). The extent of habitat modification and distur- Habitat labels based only on local information bance is difficult to quantify because the type and may be unsatisfactory, since I in Britain and the intensity of change varies from place to place In rest of Europe ) we know that (a] some widespread the case of wetlands, however, there is a common species are found m different habitats according factor in that alteration to the water table has a to where they occur in their geographic range; (b) greater impact on the fauna than other uses made 498 MEMOIRS OF THE QUEENSLAND MUSEUM

TABLE 1. Numbers of British spider species in the TABLE 2. Numbers of British spider species assessed IUCN categories Endangered, Vulnerable and Rare, as characteristic of various major habitats (compiled assigned to major habitats (compiled by P. Merrett in by P. Merrett in Ratcliffe, 1977) Bratton, 1991). Grasslands 74

Dry lowland heaths 63 Coast- dune, shingle, saltmarsh, cliff 24 Coastal formations 53 Dp; lowland heaths 14 Deciduous woodlands 34 Fens-mesotrophic to eutrophic 12 Fens 31 Deciduous woodlands in lowlands 10 Wet heath/bog 21 Caledonian (ancient) pine forest, Scotland 5 Open moorland 17 Grasslands-acidipholous to calcicolus 7 Caledonian pine forest 14 Wet heath/bog—oligotrophic 7

Open moorland, uplands and mountains 4 This provides evidence that they disperse over long distances but survive in only a few places. of these areas by man. In 1969 11 fens in East Agricultural land has been colonised by species Anglia were surveyed by a group of arach- whose natural habitats are sand-dunes and stony nologists hand-collecting for a total of 10 hours open ground, for example Troxochrus per site. Each hourly collection was kept separate. scabriculus (Westring, 1851) and Milleriana in- The similarity between the faunas of the 11 fens errans (O. P. -Cambridge, 1884). Porrhomma in terms of abundance of each species present was convexum (Westring, 1861) is also frequent on assessed using Mountford's Index of Similarity cultivated land but elsewhere occurs in caves, (Mountford, 1962). Three groupings were mines, culverts (Locket and Millidge, 1953), and derived, of which 2 were more similar than either under stones by stream and lake shores (K. was to the third group. The 8 fens in groups 1 and Thaler, pers. comm. and Duffey, unpublished). 2 had survived with little change to their water The stone-filled filter beds of sewage works regimes at that time, while the third group had attract another cave species, Lessertia dentichelis suffered falling water tables and hence vegetation (Simon, 1884), together with Leptorhoptrum changes (Duffey, 1974). The mean numbers of robustum (Westring, 1851), whose natural species for groups 1 and 2 were 53.2 and 50.0, habitat is freshwater marshes and wet meadows, respectively, and 34.6 species for group 3. Most and also E. longipalpis and P. convexum. The of the rarer and more specialised fen species environment of filter-beds is completely artificial (Marpissa radiata (Grube, 1859); Hygrolycosa and uniform with stable temperature and high rubrofasciata (Ohlert, 1865); Neon valentulus humidity, forming a 'super habitat' in which few Falconer, 1912; Carorita paludosus Duffey, species occur but in much higher numbers than 1971; Centromerus incultus Falconer, 1915; found in nature. Maso gallicus Simon, 1894) were recorded in The ability of some species to live successfully groups 1 and 2. in two contrasting habitats was first noted by Bristowe (1939) and described as HABITAT VERSATILITY IN SPIDERS 'diplostenoecism* by Duffey (1974) and 'doppel- tes okologisches Vorkommen' by Schaefer and Tischler (1983). The best-known examples are The ability of many spiders to live successfully those species found on coastal dunes and also in in a range of different environments has been mnrshcs-Synagelesvenator (Lucas, 1836), Tibel- little studied. The best known examples are lus maritimus (Menge, 1875), Clubiona juvenis pioneer species that are good aeronauts and the Simon, 1878, Hypomma bituberculatum (Wider, first to colonise newly created habitats. Meijer 1834) and Thanatus striatus C.L. Koch, 1845. (1977) discusses this for Dutch polders, and Duf- The last is also found in dry grasslands. Sitticus fey (1978) for croplands and grasslands. Not all rupicola (C.L. Koch, 1837) is widespread on aeronauts behave in this way. Erigone arctica stony mountainsides in central Europe (White, 1852) and E. longipalpis (Sundevall, (Proszynski, 1978) but in England occurs only on

1 830) are often abundant on coastal driftlines and coastal shingle banks. saltmarshes but are rare inland. The former is Competitive relationships may also influence found on mountainsides in Sweden (the late A. choice of habitat and hence distribution. Zelotes Holm, pers. comm.) and both have been recorded electus (C.L. Koch, 1834) is the characteristic on inland saline areas as well as in sewage works. species of this genus on coastal sand-dunes in FACTORS INFLUENCING SPIDER DISTRIBUTION 499

Britain whereas Z pusillus (C.L. Koch, 1833) is associated with inland heaths, dry grasslands and open, stony ground. However, Z. electus was found (Anon., 1979) to extend only as far north as south-east Scotland and at higher latitudes was replaced on the coastal dunes by Z. pusillus. All these species can adapt to different environmental conditions and emphasise the need to study habitat selection throughout the whole geographic range of a species before its distribution status can be understood.

EFFECTS OF HABITAT FRAGMENTATION

Several distribution maps in Locket et al. (1974) show widely scattered records for certain species which may indicate a decline from a former continuous distribution. Lepthyphantesmidas Simon, 1884 is an ancient forest specialist and is associated with loose bark and dead timber or birds' nests made of twigs. It FIGS 1A-C. Distribution of spiders on 2.5x2. 5m plot is only from 4 scattered sites in Britain divided into 25x25cm quadrats on a Danish heath- known land. where ancient forest survives. Similarly, wet- 1A. Vegetation map: Blank space, heather; lined lands have suffered serious losses, including the shading, grass tussocks; cobbled, mosses, lichens and Fen Basin, East Anglia, which was progressively small stones. drained from Roman times and lost the remaining large areas of marsh in the mid- 19th century. The local lycosid Hygrolycosa rubrofasciata (Ohlert,

1 865) survives in the two remaining fenland areas

1 "\ and also in numerous small fen relicts around the 2 2 2 2 -/ \ s .\ 1 d d 3 4 2 6 2 * 3:

s s 5/ 4 3 6 2 3 '•-.••-"' l 4 t 1 s s s\ 3 2 6 3 4 r 3 d © d ft/ 1 3 5 7 3 3 4 2 / \ \ 5 1 2 3 5 1 6 1 1 2 14 7 4 2 1 > 2 1 2 2 5 3 3 3 7 1 3 3 / d 6 9 5 5 1 3 6 2 1 ~ t ) v_ 7 2 2 4 7 2 5 7 , . 2 j 3 I J L

d s

FIG. IB. Trichopterna cito. A small web-spinning linyphiid almost confined to open ground with short, FIG. 1 C. Other species preferred cover of heather plant sparse vegetation; not a known active aeronaut nor a or grass tussocks, e.g. Scotina gracilipes (Blackwall, pioneer (Table 3) but a Narrow Specialist. Spread of 1 859) (s) and Dipoenaprona (Menge, 1 868) (d). Each heather eliminates it. record refers to a catch of 1-5 individuals. 5(X) MEMOIRS OF THE QUEENSLAND MUSEUM

TABLE 3. Classification of life strategies and adap- Pardosa palustris (L., 1758) is the most active tability to environmental diversity based on British aeronaut of the common lycosids found in Britain spiders. and northern Europe (Richter, 1970). Few have been recorded in a wide variety of habitats, in- PIONEER SPECIES cluding marshes, grasslands, heathlands, and Active aeronauts which disperse freely, exploiting agricultural crops, but occasionally is found to be newly created open ground where competition is dominant, as in moist hay meadows in river val- low. Widely distributed in temporary or changing leys and on some sand-dunes. In a survey of dune habitats such as agricultural cropland, gardens, systems in Scotland (Anon., 1 979) it was the most urban areas, leys and other types of disturbed abundant lycosid, although it was not recorded on ground or vegetation. all the dune areas which were studied. It seems GENERALISTS that the active aeronautic behaviour of P. palustris enables it to move about freely and its

Common species with a capacity to adapt to a wide ability to exploit many different habitats gives it range of semi-natural habitats and permanent ar- a high ranking as an opportunist. Kessler (1973) tefacts in the man-made environment. May be also showed that P. palustris is able to produce difficult to assign to a particular habitat. more eggs per unit of body size under field conditions than 7 other Pardosa species. This may be BROAD SPECIALISTS an advantage when colonising new areas.

A. Widespread, euryoecious, or 'characteristic' Adaptability to environmental diversity varies species (Table 2) associated with major habitats from species to species, so that a gradient exists such as deciduous woodland, marshes, heaths or from those which are very widespread (usually ancient grassland, but may be found in many dif- many species) to those which appear to be con- ferent variants of the chosen major formations. fined to a specialised habitat (relatively few species). The different components of this B. Diplostenoccious species, mostly widespread gradient are outlined in Table 3, and modified and associated with 2 different habitats but usually from Duffey (1975a). more common in one than the other. Occasionally much more abundant in man- made habitats than in the natural environment. This grouping grades HABITAT DIVERSITY AND into species successful in 3 or more different en- MICROSPATIAL DISTRIBUTION vironments.

NARROW SPECIALISTS The structure of the vegetation, the litter layer and physical features of the environment have a Stenoecious species which seem confined to clearly strong influence on spider distribution and defined habitat. Includes rare species in low num- species composition (Duffey, 1962, 1968, 1974; bers and others which may be locally abundant, Edwards et a/., 1975; Uetz, 1991). There is although confined to a restricted area because the evidence that all species, however widespread, habitat is scarce. are influenced in some way by habitat structure. This was shown by comparing the fauna in dif- margins of the Fen Basin. Are these relicts of a ferent quadrats of a simple vegetation type on a former more extensive distribution? Danish heathland (Duffey, 1974). Fig. 1 shows heathland severe Lowland has also suffered the differences in species and numbers of spiders losses, having been reduced to many small frag- between 100 25x25cm quadrats in a block ments by agricultural reclamation with a conse- measuring 250x250cm. The vegetation consisted loss biodiversity quent of (Webb, 1990). Eresus of a heather plant {Calluna vulgaris (L.) Hull, niger (Petagna, 1787) was formerly more 1808) in one corner of the block, and a few widespread in heaths in southern England but is scattered grass tussocks of Deschampsiaflexuosa now known from only one locality where the ((L.) Trin., 1836), while most of the area was population is small. covered with mosses, lichens and small stones. In contrast, some species are able to live in a The species in Fig. 1 occurred in clearly defined range of different habitats, and thus overcome the microhabitats. Trichopterna cito (O.P.- problem of fragmentation. If this characteristic is Cambridge, 1872), which was widespread in the also combined with active aeronautic dispersal, open ground of moss and lichen, avoided the the chances of finding isolated suitable environ- heather, and was also absent from 4 quadrats on ments will be further enhanced. For example. the right-hand side of the sampling area which FACTORS INFLUENCING SPIDER DISTRIBUTION 501

had been trampled. Similar results are reported by ACKNOWLEDGEMENTS Jocque (1973), who studied the fauna of different types of woodland litter layers in Belgium. On a My grateful thanks to Malcolm Mountford and heathland ranging from dry to wet areas, Snazell David Spalding for statistical help with the Index (1982) sampled the spider fauna by pitfall trap- of Similarity. ping at 154 random points over 12 months. By Principal Components Analysis on the 45 most LITERATURE CITED numerous species he was able to show a wide ANON. 1964-80. 'Flora Europaea'. 5 vols. (Cambridge gradation from specialised to broad habitat University Press, Cambridge). preferences, which conform well with the 1979. 'The invertebrate fauna of dune and Machair categories in Table 2. sites in Scotland, Vol.I\ (Project 469 Report to In an experiment in tall grassland in England, Nature Conservancy Council: Peterborough). faunal changes were recorded in grass litter ALDERWEIRELDT, M. & MAELFAIT, J.-P. 1990. modified by trampling. Samples of grass litter 'Catalogus van de Spinnen van Belgie, Deel VII, Lycosidae'. (Studicdocumenten No. 61, enclosed in nylon-mesh bags (20x20x8cm) Koninklijk Belgisch Instituut voor Natuur- received 3 different treading treatments (5 wetenschappen: Brussel). treads/month, 10 treads/month and an undis- BRATTON, J.H. (ed.) 1991. Arachnida. 121-219. In turbed control series) (Duffey, 1975b). There 'British red data books. 3. Invertebrates other than were 25 replicates in each case. After 12 months insects'. (Nature Conservancy Council: Peter- borough). the volume of the litter in the controls had fallen BRISTOWE, W.S. 1939. 'The comity of spiders, Vol. by 50 % due to natural decay, but those having 1'. (Ray Society: London). 10 treads/month fallen 81 the had had by %. Of CANARD, A. (ed.) 1990. Araignees et scorpions de 10 most frequent species, 5 were eliminated by l'oucst de la France: catalogue et cartographic the treading after 12 months, and the total num- provisoire des especes. Bulletin de la Societe Scientifique de Bretagne61(l): 1-302. bers of all species fell by 84 %. Of the 5 remaining DUFFEY, E. 1962. A population study of spiders in species, 3 showed little response to the treading, limestone grassland: description of study area, their numbers falling only marginally. The higher sampling methods and population characteristics. level of treading in this experiment was very light Journal of Animal Ecology 31: 571-599. compared with that on a public footpath in a 1968. An ecological analysis of the spider fauna of 641- popular area, but the effect on the litter fauna was sand dunes. Journal of Animal Ecology 37: quite dramatic. 674. 1974. 'Nature reserves and wildlife' (Heinemann CONCLUSIONS Educational Books: London). 1975a. Habitat selection by spiders in man-made environments. Proceedings of the 6th Interna- The modification of the European landscape tional Arachnological Congress. Amsterdam, through many centuries has been the greatest 1974:53-67. influence determining the distribution of spider 1975b. The effects of human trampling on the fauna of grassland litter. Biological Conservation 7: species. Natural and near-natural habitats are now 255-274. rare, as are the specialist species associated with 1978. Ecological strategies in spiders, including them. Today many species have adapted to man- some characteristics of species in pioneer and made environments, whether permanent or tem- mature habitats. Symposia of the Zoological porary, and those preferring open-ground Society of London 42: 109-123. E. GREEN. M.B. 1975. linyphiid spider conditions have greatly expanded their range. A DUFFEY, & A biting workers on a sewage-treatment plant. Bul- gradation of life strategies (or adaptability) from letin of British Arachnological Society 3: 130- pioneer species to narrow specialists is proposed. 131. Superimposed on these factors are the influences EDWARDS, C.A., BUTLER, C.G. & LOFTY, J.R. of major climatic zones and the microspatial 1975. The invertebrate fauna of the Parks Grass plots, II. Surface fauna. Report Rolhamsted Ex- variations in the abiotic environment. All these perimental Station, 1975(2): 63-89. features should be considered when defining the JANSSEN, M. & BAERT, L. 1987. 'Catalogus van de distribution and habitat characteristics of a Spinnen van Belgie, Deel IV, Salticidae'. species, drawing on evidence from the whole (Studiedocumcnten No. 43, Koninklijk Belgisch geographic range. Instituut voor Natuurwctcnschappen: Brussel). 502 MEMOfRS OF THE QUEENSLAND MUSEUM

JOCQUE, R. 1973. Spider fauna of adjacent woodland Oxyopidae, Pholcidae, Pisauridae, Scytodidac, areas with different humus types. Biologisches Segestriidae, Eusparassidae, Zodariidae, Jahrbueh Dodonaea41: 153-178. Zoridae. Documents de Travail No. 46 (Institut KEKENBOSCa J.* BOSMANS, R. & BAERT, L. Royal des Sciences Naturelles de Belgique: 1977. 'Liste des araignees de la faune de Bruxelles). Belgique'. (Tnslitut Royal des Sciences Naturelles RANSY, M., KEKENBOSCH, J. & BAERT, L. 1990. de Belgique: Bruxelles), Catalogue des araignees de Belgique, Pt 6 KESSLER, A. 1973. A comparative study of the Clubionidae et Liocranidae. Documents de production of eggs in eight Pardosa species in the Travail No. 57. (Institut Royal des Sciences field (Araneidae, Lycosidae).Tijdschrifl voorEn- Naturelles de Belgiques: Bruxelles).

tomologie 1 16: 23-41. RATCLIFFE. D. A. (ed.). 1977. Spiders. Pp. 95-101. In LOCKET, G.H. & MILLIDGE, A.F. 1953. British *A nature conservation review. Vol. 1". spiders'. 2 vols. (Ray Society: London). (Cambridge University Press: Cambridge). LOCKET, G.H., MILLIDGE. A.F. & MERRETT, P. RICHTER, C.J.J. 1970. Aerial dispersal in relation to 1974. 'British spiders. Vol. III'. (Ray Society: habitat in eight wolf spider species (Pardosa, London). Araneae, Lycosidae). Oecologia, Berlin 5: 200- ME1JER, J. 1977. The immigration of spiders 214. (Araneidae) into a new polder. Ecological En- ROBERTS,MJ. 1985-87. 'The spiders of Great Britain tomologv2: 81-90. and Ireland'. 3 vols. (Harley Books: Colchester). MOUNTFORD, M.D. 1962. An index of similarity and SCHAEFER. M. & TISCHLER, W. 1983. its application to classificatory problems. Pp. 43- 'WorterbiicherderBiologie; Oekologie'. (Gustav 50. In Murphy. P.W. (ed-). 'Progress in soil Fischer: Jena). zoology*. (Butterworths: London). SNAZELL, R. 1982. Habitat preferences of some PROSZYNSKI, J. 1978. Distributional patterns of the spiders on hcalhland in southern England. Bulletin Palaearctic Salticidae (Araneae). 335-43. In Mur- of the British Arachnological Society 5: 352-360. phy, P.W. (ed.) 'Symposia of the Zoological Society of London No. 42. Arachnologv' UETZ, G.W. 1991. Habitat structure and "spider forag- RANSY, M. & BAERT, L. 1985. 'Calalogus van de ing. Pp. 325-348. In Bell, S.S., McCoy, E. D. and Spinnen van Belgie, Deel D. De Cribcllatae*. Mushinsky, H.R. (eds). 'Habitat structure, the (Studiedocumenten No. 25, Koninklijk Belgisch physical arrangement of objects in space'. (Chap- Instituut voorNatuurwetenschappen: Brussel). man and Hall: London). 1987a. Catalogue des araignees de Belgique, Pt 3. VAN W1NGERDEN, W. 1977. 'Population dynamics Les Araneidae. Documents de Travail No. 36. of Erigone arctica (White) (Araneae. (Instilut Royal des Sciences Naturelles de Belgi- Linyphiidae)'. Doctoral thesis. Free University of que: Bruxelles). Amsterdam. 1987b. Catalogue des araignees de Belgique, Pt 5. WEBB, N. R. 1990. Changes: on the heathlands of Anyphaenidae. Argyronetidae, Atypidae. Dys- Dorset, England between 1978 and 1987. Biologi-

deridae, Mimciidae, Ncsiicidac, Oonopidiie, cal Conservation 51 : 273-286. THE DEVELOPMENT OF THE ASYMMETRICAL WEB OF NEPH1LENGYS CRUENTATA (FABRICIUS)

JANET EDMUNDS

Edmunds, J. 1993 1 I 1 1: The development of the asymmetrical web ofNephilengvs cruentata (Fabricius). Memoirs of the Queensland Museum 33(2): 503-506. Brisbane. ISSN 0079- 8835.

In the field, young Nephilengys cruentata (Fabricius) remain on the barrier threads of their mother's web for about ten days, before moulting. Some do not initially disperse far, and may even remain on the barrier threads to build their first web. These first webs are complete orbs, As the spiderlings grow the hub of the web is spun approximately three quarters of the way up the web; later there are no spirals above the hub- At this stage some spiders join the hub to the barrier, forming a *teni* under which the spider rests. The true reireai, of a tube that is closed at the end away from the hub, develops gradually as the spiders grow \\j[\\ivr,C]Nephilengys, web development* evolution of webs.

Janet Edmunds, Mill House. Mill Imu>\ Goo\uurgh. Pre\ton PR32JX, United Kingdom; 12 November, 1992.

Nephilengys cruentata (Fabricius) is common RESULTS throughout tropical Africa. In Ghana it lives in savanna and forest edge areas. Under natural The cocoon is laid close to the web, often nn a wall or ceiling When the spiderlings emerge, conditions it attaches its web to trees, but it has they remain close to where they have emerged in frequently adapted to attaching it to dwellings. It a tight bunch on the barrier web for about ten was very common at Legon in such situations. days, spinning extra irregular threads. After the The adult female is large but the mate is much first post emergence moult, they disperse, but smaller. The webs of larger spiders are usually a many remain close to the mother's web, some roughly triangular partial orb The hub is at the even building then 'irsi catching web in the upper apex, where there is a cylindrical retreat, threads of the barner web closed at the end away from the hub. There is an When the spiders are grouped into different size extensive barrier above and behind the viscid classes based on leg 1 length, the extent of the the increases with each succes- web; the main attachment between the two is at web below hub the hub. The hub and retreat arc often in the angle between a wall and the ceiling or overhang. This paper presents data on the shape of the web of N. cruentata from the earliest instars to adult . METHODS

Spiders of all ages were observed living on the Position ot hub verandahs of a private dwelling and of the Zool- ogy Department at Legon, but it was not possible to follow individual spiders in the field or captivity. Web size and length of the first leg of each spider were measured, although the relation between spider si ?e and instars was not determined. The vertical asymmetry, lateral asymmetry and Leg Length shape (circularity) ot the web were determined FIG. 1. Changes tn vertical asymmetry in webs of N. from the ratios of vertical radii, horizontal radii cruentaiit with different leg lengths. At lop; ratio of and the two diameters respectively, all measured upper to lower radius = 0; Between centre and top: to the nearest 5mm; N for all figs is given in Table ratio of upper to lower radius >0

upper to lower radius = ] . I. .

504 MEMOIRS OF THE QUEENSLAND MUSEUM

Position of hub

Central

Between centre & side

At side

Leg length Leg length FIG, 2. Changes in lateral asymmetry in webs of N. cruentata with different leg lengths. At side: ratio of two horizontal radii = 0; Between centre and side: ratio of two horizontal radii >0 <1; Central: ratio of Web shape two horizontal radii = 1 FIG. 3. Changes in shape in webs of N. cruentata with different leg lengths. Longer than wide: vertical sive class, while that above the hub increases diameter > horizontal diameter; Central: vertical slightly before decreasing to zero (Table 1). diameter = horizontal diameter; Wider than long: The vertical asymmetry increases with size vertical diameter < horizontal diameter. (Fig. 1): spiders with leg 1 <6mm spin complete orbs. As they grow, less web is spun above the viscid web, the barrier increases (Fig. 4). The hub, so that the hub is then about three quarters very first webs have no barrier or just a few of the way up the web. Larger spiders, with leg 1 threads, which are attached to the orb at only a >19mm have no web above the hub. Because of few points. But over some days more threads are the difficulty in following individual spiders, the added, until the barrier is a fairly dense cone time span over which these web changes occurred shaped tangle behind the orb, and more firmly is not known. fixed to the web. Small pellets of detritus, similar Early webs are symmetrical laterally (Fig. 2), in size and shape to the spiderling and of the same but webs of larger spiders show greater variation pale grey colour, usually occur in the barrier and may be markedly asymmetrical laterally; at (Edmunds and Edmunds, 1986). These may the extreme the hub is completely to one side. deflect attacks of potential predators. Some early webs are also circular or nearly so The development of the retreat increases with (Fig. 3), however very few webs of spiders with the size of the spider (Fig. 5). Spiders with the hub leg 1 > 6mm have the vertical and horizontal diameters of the same size. During these changes in the proportions of the

TABLE 1. Body size and web dimensions (in mm) in

different size classes (based on leg 1 length) of Nephilengys cruentata.

Barrier Leg ] length <6 7-12 12-18 >19

Body size <3.5 3.5-7.0 7.0-9.0 >9.0

N 8 14 14 9

Hub-web toprmean 38.1 47.1 19.2

range 25-50 0-80 0-100

Hub-web botlom: 38.1 109.3 180.0 478.9 mean Leg length range 25-50 45-170 140-260 200-700

Web Width: mean 73.8 117.9 177.4 431.1 FIG. 4. Development of barrier in webs ofN. cruentata range 70-90 80-180 80-290 240-850 with different leg lengths. WEB OF NEPHILENGYS 54)5

hub. Like Nephilengys, there are barrier vvhv though on both sides, and often above the orb. These are supportive and probably also defensive (Robinson and Robinson, 1973). Herennia or natissima (Dolcschall) constructs a very long web, with almost parallel sides iRobinson and Lubin, 1979). The spider sits in a cup-shaped depression of dense silk that forms Lhehub. It is towards the top of the web. though there arc

several sticky spirals above it. The spider h;v

barrier web. but as it builds very close to tree

trunks, it would be difficult to find a place to build one.

Larger juvenile Nephilinae build webs like the adults, and few observations have been publish- R6UMI a ed on the very early webs. The first wel Nephila cfavipes (Linnaeus) are circular (Corn- stock, Levi and Levi, 1968). FIG. 5. Development of reireat in webs gfM crueniaia 1948; Brown

v. ith different leg lengths. Christenson (1983) give measurements of the webs of N. clavipes spider! ings between 2 at the centre or part of the way up the web rest at 9mm body length. The webs show an increase in grow. the web hub and there is no retreat. When the hub vertical asymmetry as they However, UK) like Nephilengys cruentaia, even the instars truil 1 1 1 «uilt at or very close to the apex, the first stages of a retreat are formed as a denser group of spin the first catching webs had the hub ap- threads within the barrier close to the hub. As proximately two thirds of the way up the more threads axe added, this becomes a roof-like Both species seem to spin the hub at the top ol reach similar size. with tcilt between the hub and the barrier, or some- web when they a Webs spirals above the hub are spun by ju\c times an open ended tube. Spiders with leg 1 >] 9mtn normally had a closed tubular retreat, like Nephila maadaia (Fabricius) (Robinson and an adult. Robinson, 1973, figv 5 6), and possibly bv Nephila senegalensis (Walckenaer) (Clan Early webs are very fine, and it is not easy to ornutis see the details of the indiv idual threads. However, 1987). The webs of juvenile Herennia Lubin, 1979) are also a complete web appears to be spun in one opera- sima (Robinson and more like a complete orb. tion, unlike larger webs which are spun in two or more sections on different days. The temporary The fact that the earliest webs of at least sonic spiral does not appear to be left up, though more species of Nephilinae are complete orbs indicates detailed observations would be needed to confirm that the orb is a primitive characteristic in them, this. Larger spiders leave the temporary spiral in and that the incomplete orbs of the larger spiders the finished wpb. are derived. There are other spiders which have u highly modified orb-web but build a more c Wi DISCUSSION plete orb as a juvenile. Spiders of the genus Scoloderus build extremely elongated inv-. The webs ol most adult female Nephilinae are ladder webs, with the hub towards the base. The not typical orbs. Most Nephiki species build in- juveniles of $coioVe the hub, including in some instances a few (Stowe, \97%).ln Araneusatrihastulus from New sticky spirals. In Nephila plumipes large, probab- Zealand, which builds a web that is elongated ly mature, females were seen in Brisbane, both above and below the hub, some of the Australia, that had several sticky radii above the proportions of juvenile webs are less extreme hub. However, in at least some individuals, these (Forster and Forster, 1985). It would be interest- appeared to have been laid as pendulum turns, ing to find the st iv early webs of other spiders rather than complete spirals (pers. obs.). Nephila with atypical webs, such as that of the spider thai does not huild a retreat and the spider rests at the built the ladder web observed by Robinson and $06 MEMOIRS OF THE QUEENSLAND MUSEUM

Robinson (1972) in New Guinea, oow identified COMSTOCK, J.H. 1948. The spider book'. Revised as close to Tylotlda (El>cihard, 1990b). Other and edited by WJ. Gertsch (Cornstock Publish- aspects of web construction and use may be more ing: Ithaca). 1975. inverted ladder* orb derived in adults than in juveniles, such as the EBERHARD, W.G. The web of Scoloderus sp. and the intermediate orb of angle of the spring line and resting position of the Eusmla (?) sp. (Araneae: Araneidae). Journal of spider in Epeirotypus sp, from Costa Rica (Eber- Natural History 9: 93- 106. hard. J 986). However, in the west African 1982. Behavioral characters for the higher Has Paramtxeus cyrtoscapus (Pocock), the conical sifkation of orb-weaving spiders Evolution 36; horizontal webs of juveniles arc more derived 1067-1095, than the planar vertical of adult females »Ed- 1986 Ontogenetic chances in the web of rrnmds, IV tfr&typiis sp. (Araneae, Tberidiosomatidac). If it is confirmed that young Mepfiilengys Journal of Arachnology 14: 125-128. crueniata build webs in one picef and destroy 1990a. Earlv stages of orb construction by temporary spirals^ then the adult behaviour of Philoponelia v'tcina, Leucauge mariana, and leaving the temporary spirals in place is Nephila clavipes (Araneae, Uloboridae and presumably aJso a secondary modification. There Lgnathidae), and their phylogcnetic implica- s. Journal Arachnology 18: 205-234. are other characteristics of the webs of of Function and phylogeny of spiders webs. Nephilinae that are derived. Despite their size, the 1990b. Annual Review of Ecology and Systematic* 21. webs of adults have a finer mesh compared to 14 1-372. some other orb weavers (personal observati EDMUNDS. J. 1978. The web o\ Paraneus [itej <:yr- hard (1981 1990m) concludes that some toscapus (Pocock 1899) (Araneae: Araneidae) in characteristics of the web building behaviour of Ghana. Bulletin of the British ArachnologiccM Nephila clavipes, such as the unique method of Society 4: 191-196. laying the sticky spiral, are an adaptation to spin- EDMUNDS, J,& EDMUNDS, M. 1986. The defensive ning a tightly meshed WMfib, &Ve*l though other mechanisms oforb weavers (Araneae: Araneidae) aspects (e.g. frame construction} are primitive, in Ghana. West Africa. Pp. 73-89. In Eberhard, Levi (1986), Eberhard (1990a) and Coddington W.G.. Lubm. V.D. and Robinson. MR (ed.s). (1990) classify the Nephilinae with the Metinae Proceedings of the Ninth International Con^.rrss of Arachnology, Panama (Smithsonian Institu- and tetragnattiids, rather than with Aranein and tion: Washington D.C.). its relatives. The derived nature of ihe partial orb FORSTF.R, L.M & FORSTF.R, R R 1985 A deriva- of the larger Nephilinae would be consistent with tive of the orb web and its evolutionary sig- this classification. nificance. New Zealand Journal ol Zoology 12: 4SS-465. ACKNOWLEDGEMENTS LEVI, H.W. 3986. The neouopieal orb- weaver genera ysometa and Homalt.'wcr/} (Araneae: Tetrag-

I would like to thank Prof D.W. Ewer for the nathidae). Bulletin ofthc Museum of Comparative DSe of facilities while working at the University Zoology 151:91-215

l of Ghana, and Malcolm Edmunds for construc- LEVI, H.W. & LEVI, L.R. 1968. A guide to sp.uers tive comments. and their kin'. (Golden Press: New York). 160pp. ROBINSON. M.H. & LUBIN, YD. 1979. Specialists L1TERATURL and generalists: the ecology and behavior ofSQtttf wcb-building spiders from Papua New Guinea. Pari fie insects 21: 97-132. BROWN, S.G. & CHR1STENSON T.E. [983. The relationships between web parameters and ROBINSON, M.H. & ROBINSON, B. 1972. The struc- the spiderling predatory behavior in the orb-weaver ture, possible function and origin of remark- Nephila clavipes. Zeitschrift fur Tierpsyehologie able laddcr-wcb built by a New Guinea orb-web 63:241-250. spider (Araneae: Araneidae) Journal of Natural CLAUSEN, J.H.S. 1987. On the biology and behaviour History 6: 687-694. of Nephila wnegatensis senegalensis (Walck- 1973. Ecology and behavior of the giant wood

enaer, 1 837). Bulletin of the British Arachnologi- spider Nephila macula ta (Fabricius) in New calSocieiy7: 147-150. Guinea. Smithsonian Contributions to Zoology CODD1NGTON, J.A. 1990. Ontogeny and homology 149: 1-76. in the male palpus of orb- weaving spiders .mil STOWE, M.K. 1978 Observations of two nocturnal their relatives, with comments on phylogcny oibweavc/s that build spceiali/.cd webs: (Araneodada: Arancoidca, Deinopoidea). Smith- Scotpdtws cordatus and Wixin sctypCL (Araneae: sonian Contributions to Zoology 496: 1-52. Araneidae) Journal of Arachnology 0: 141-146. DOES MIMICRY OF ANTS REDUCE PREDATION BY WASPS ON SALTICID SPIDERS ' MALCOLM EDMUNDS

Edmunds, M. 1993 11 Hi Does mimicry of ants reduce predaiion by wasps on snfticid spiders? Memoirs of the Queensland Museum 33(2). 507-512. Brisbane. ISSN 0079-8835.

A common predator of spiders at Legon, Ghana, is the wasp Pison xamhopus, 96% oi whose prey were salticids. The data from cells built by 31 wasps containing over 800 spiders arc used to examine whether mimicry of ants gives protection against Pison. Comparison of salticids in wasp cells with those found on nearby vegetation shows that fewer ant mimics {Myrmurachne spp.) are taken than one would expect if wasps were capturing salticids in proportion to their occurrence, but also that some individual wasps specialize in capturing Myrmurachne, Implications lor the searching image hypothesis are discussed. \Z\Baieswn mimicry, ant-mimicry, Salticidae, predatiorl, search image. Pison. Myrmarachne.

Malcolm Edmunds, Department of Applied Bioloyv, University of Central Lancashire, Preston I PR 2HES United Kingdom; 12 November,"!992.

Ant mimicry has evolved in several spider fam- associated with them would also escape preda- ilies* e.g. Salticidae. Clubionidae, Thomisidae, tion. He further attempted 10 show lhat ant- Aphantochilidae and Theridiidae (Kingston, mimicking spiders arc less often taken by the 1927; Mathew, 1954; Reiskind and Levi, 1967; wasp Pison xanthopus than are other salticids, but Rciskind, 1977; Edmunds, 1974, 1978; Wanlcss, there were rather little data available at that time. i978;01iveiraandSazima, 1984, 1985;01iveira, More recently, the American ant mimic 1988). Some of the most precise morphological Synageles occidentalis was found being preyed and behavioural resemblances to ants occur in the on less than are salticids salncid genus Myrmarachne which is widespread much non-mimicking by in both the Old and New Worlds, especially in the the philodromid Tibeltus oblongus, and the large tropics (Wattless, 1978). AtLegon, Ghana, three salticid Phidippus vlarus ignored or avoided species of Myrmurachne are common, each one Synageles and ants in exactly the same way, while closely resembling one species of ant when it is continuing to attack non-mimicking salticidv full grown but a different ant species when it is (Cutler, 1991) immature (Edmunds, 1978). This mimicry could More data on the prey of Pison xanthopus arc have two advantages for the spider: given here to test the theory that ant mimicry 1. It could deceive the ants and so enable the gives protection against spider-hunting wasps. -.pivJcr to creep up and prey on them. This is aggressive mimicry as may possibly occur in the MATERIAL thomisid Amyciaea forticeps (Mathew, 1954), and is well documented for the aphantochilid Mud cells of Pison xanthopus were collected Aphantochilus roqersi (OHveira and Sazima, from window frames at the University of Ghana, 1984), Lcgon, Ghana between 1969 and 1973. In- 2. It could deceive a predator into mistaking it dividual wasps build from I -6 cells in a row. Each for an ant which that particular predator docs not cell is stocked with 5- 1 paralysed salticid spiders cat, This is b;ite.sian mimicry. and an egg is laid on one of these. The spiders Myrmarachne do not normally attack ants, so were preserved in alcohol and identified, usually there is no support for aggressive mimicry (Ed- to genus or species. They were also classed as munds, 1978; Wanlcss, 1978) Indeed whenever cither good ant mimics {Myrmarachne), poor ant an ant comes near they use their acute eyesight mimics (Cosmophasis sp.) or non-mimics (other and quick reactions to avoid contact This genera). Most of Che spiders are now in the col- gests a danger of being killed if caught by the 3nts (Edmunds, 1978; Wanlcss, 1978). Direct lection of the Natural History Museum, London. evidence supporting batcsian mimicry is sparse. Identification of the spiders was confirmed |» Edmunds (1974. 1978) argued lhat because few F.R Wanless, and the wasp was determined by insectivorous birds prey on ants, the ant mimics the tatc Professor O.W. Richards. SQfi MEMOIRS OF THE QUEENSLAND MUSEUM

Habitat taken by wasps correspond with those found in one particular habitat then the wasps are probably hunting in that habitat. Spider — The habitats are: leaf litter; short, regularly cut in w g a grass; long grass and herbage; l-3m high shrubs; i & s B bO and tree trunks. The spectrum of spiders in wasp •z> C JZ g similar to those in shrubs 1 3 'Si cells is most found and Cood ant mimics trees (Table I). The canopy was not sampled but fauna. However, it is un- Mvrniarttchne foenisex Simon + + + probably has a similar likely that the wasps were hunting in short grass, Mvnnarachne elongate S/Aimbalhy + •- + leaf litter or long grass. Mvrmaruchne legon Wattles* + +

Myrniaruchne uvira Vfcttess + RESULTS Poor ant mimics

Castnophasit «n. + + + + + N on- mimic? Spiders in Wasp Cells and on Shrubs

Thx?»e sp. ^r ; + + + Some variation in the numbers of spiders f/W/MJ Sp. + caught by each wasp (Table 2) is due to the

(iAWJfimsy. + + different numbers of cells completed by the Cells full lar- *//.:, .v .p. + + wasps when collected. with grown the in (\Ui-bt'tus rufopicrus Simon + vae or pupae were ignored since spiders them were reduced to carapace cuticles, but cells -»• "s aditnsont ( Audouin) with eggs or young larvae contained spiders that 't'-hr-nki-lto modesta \ essert + were intact and so are included. The spiders Menemems sp. + + caught by each wasp came from 2-9 cells, e.g. un P$ettdici»$ 5p 4- 4 3 Feb 1973, the first cell contained a pupa, the Other non-niiuiicMrig tallic Ids + + second a full grown larva, and the third had nine spiders but no egg. The wasp presumably was live TABLE 1. Salttcid spider-. 111 bflbitafe at Legon. killed before fully provisioning this cell. May-July 1973. Of 872 spiders found, 837 were salticids (Table k RATIONALE 2); 160 were good' ant mimics (i.e. Myr* marachne spp.), judged by the human eye, and a further 15 can be classed as poor (behavioural) If ant mimicry deceives wasps so that they do ant mimics (i.e Cosmophasis spp,). In 1973, not capture ant-mimicking spiders as often as every shrub in Zoology (twice) and in Botany non-mimics, then the proportion of ant mimics to (once) was searched (see Edmunds, 1978). All non-mimics should be lower in wasp cells than in salticids found were scored as cither a good the natural environment. If wasps are not mimic, a poor mimic or a non-mimic. The dif- deceived then the proportion of ant mimics ferences between types are highly significant should be the same. The tesi of tins hypothesis is : (X c» = 49.04,p<0.001; Fig. 1 ). Pison clearly take to compare the incidence of ant mimics in wasp significantly fewer good ant mimics than they do cells with those found in the field poor mimics or non-mimics compared with their incidences in the environment. WHERE DOES PISON HUNT? However, wasps searching by running quickly over vegetation are unlikely to find spiders rest- First, the wasp's hunting range must be estab- ing in their retreats beneath or between leaves. So lished so that a random sample of salticids can be perhaps the comparison should be made between collected from the same place. Pison is small, the numbers of spiders in wasp cells and the difficult to follow in flight, and was observed numbers foraging on leaves (excluding those in hunting on only a few occasions. Each time it was retreats). These figures arc also given in the upper running and making short flights over leaves of part of Fig. I (the black bars only); 61.9' shrubs. It was never observed on the ground or in spiders on shrubs were good mimics compared grass, but as I spent more time examining shrubs with 19.1% in wasp cells. This too is highly than any other habitat, this is not conclusive. I significant, again indicating that wasps take many therefore collected salticids from several dif- fewer good mimics than poor or non-mimics : ferent habitats at Legon. If the species of spider (X i2i = 64.15, p<0.001). The proportion of poor WASP PREDATION ON ANT-MIMICKING SALTICIDS 509

collected on shrubs in February and in May 1973 (Edmunds, 1978); (2) that the three wasp cells

(taken 22 Jan- 3 Feb 1973 ) from close to where the shrubs were surveyed had between them eight Myrmarachne and 54 other salticids compared with 24 Mynnarachne and 26 other salticids on the shrubs on Feb 2; this difference is highly 2 that significant (\ i = 15.0, p< 0.001 ); and (3) the data (Table 2 ) show no clear evidence of spiders occurring at particular sites or in certain seasons. So, while there may be some variation in spider species at different times and places at Legon, such variation is unlikely to account for the very different numbers of salticids taken by wasps assuming that they take different species in proportion to their frequency in the population.

662 No Wasps Near Dangerous Ants Most Myrmarachne taken by Pison were black and identified as M. legon and M. elongaia {'Wan- less, 1978) but some immature specimens could not be identified. Only one specimen of the very common M. foenisex, in a total of 160 Myr- marachne, was taken. It was a juvenile whose body was red-brown and black (Edmunds, 1978, Fig. I), quite unlike Oecophylla, but very similar to the smaller ant Crematogasier castanea which only lives close to Oecophylla (Edmunds, 1978). MJhrnisex c\ose\y associates with the aggressive red weaver ant Oecophylla longinoda, of similar Good Poor Non- colour. Adult M.foemsex are probably too large for Pison To attack, but since it captures many mimics mimics mimics young black Myrmarachne one might expect it to take young M.foemsex as well. Young M. legon

FIG. 1 . Comparison ofgood, poor and non-ant mimick- are quite similar to young M. foenisex, but they ing salticids on shrubs and in cells of Pison xan- do not associate with Oecophylla nor with C. castanea. Many young M, legon but very few M /' mirnics taken by wasps is not significantly dif- foenisex were taken by Pison. Hence, ferent from that of the non-mimics, so in the probably avoids hunting on plants overrun by analysis that follows the poor (behavioural) Oecophylla. mimic Cosmophasis is treated as a non-mimic. H Pison does not hunt on plants with Oecophyl- la, then salticids found with these ants need to be Spatial or Seasonal Variation omitted from the comparison of ant mimics and non-mimics on shrubs and in wasp cells (Table These comparisons are of spiders on shrubs in 3). Of salticids on shrubs, 49% arc good ant February and May 1973 with spiders in wasp cells mimics compared with ]9% in wasp cells. This collected over 3.5 years. The relative numbers of difference is highly significant, and remains so if Mynnarachne and of other salticids vary spiders in retreats are excluded <0.001). throughout the year and over different parts of the (p University campus, and this may account for the Wasps clearly take fewer Myrmarachne than differences in proportions of spider prey taken by one would predict if they caught them tn propor-

wasps. Evidence against this possibility is ( 1 ) that tion to their occurrence in the environment. This there was no significant difference in the relative is therefore good evidence for the defensive value numbers of Mynnarachne and of other salticids of ant mimicry against predalion by Pison. :

510 MEMOIRS OF THE QUEENSLAND MUSEUM

Mimics Non-anf rnimick.inn spiders wasps who between them

Date (place Mv Co Ps Rh |Te Mc VI M. Th Sa Mo Fi Os Of .!! took only eight Myr- out of spiders 28.9.69 Zool. :* 26 marachne 73 (11%), while it failed to 29.9,69 Zool r. i 11 protect them from the 13.1.71 8.i| 6 29 1 36 fourth wasp which took 34 If, 1 1.2.71 Bol- u 1 6 2 2 ! -- Mvrmarachne out of 46 12.3.71 jHill 4 J44 I I 1 2 prey (74%). Overall 25 in- 1 :i 1 1.4.7! ?xjol. 1 29 dividual wasps took only ' ; 34 19.4.71 Maths 32 38 Mvrmarachne out of 5272 ISSER 2 28 15 11 693 spiders (5.5%), in-

1 3 26 15.2.72 Zool 22 dicating that they had been 18 15.2-72 /.„: id 2 4 2 deceived by the mimicry.

33' , 18.3.72 7i M.I. The other seven wasps "'--: i 26 27 ..72 24 1 took 122 Mvrmarachne

1 37 14.4.72 IS SEP. 36 out of 1 79 spiders (68.2%) "" 21 1 6 Zool, 20 1 indicating that they had

.'' Zool. 2S 1 1 J overcome the defence of

22.5.72 IS.SI R 2^1 1 25 mimicry to the extent that

5.7,72 jISMIK 23 2 30 they preyed almost ex- - 6 72 ISSER 25 clusively on ant mimics (a

"} 1,11.72 Hill 5 v. I |l 45 wasp on 6 July 1972 took

! 23 Mvrmarachne and no 8.1 1.72 Zool. 15 16 1 other spiders, while in the • !3.n 7: Zoot. 19 i 45 other six wasps Myr- 15.11.72 Zool 18 1 W 1" maradme taken were 18.1,73 JHill 26 27 3 never less than 45%). 22.173 Bot b Im J 8 2 I 43 For the poor ant mimic.

24.1.73 |fioL J 9 10 Cosmophaxix, the 15

3.2.73 Bot. i 1 2 9 spiders taken by wasps 1.3.2.73 Agric 11 11 (Table 2) represent 1 .7% of i - 21 ; H Hill 6 I i all salticids taken, or 2. 1%

2'..' n 1 36 Hilt 1 33 1 excluding Myrmarachne. } 28 27.5.73 ISSER 25 1 This is less than their rela- Agiu. 19 19 tive frequency on shrubs

-J roidi 160 15 538 51 19 15 11 6 "-i 3 2 3 |33 872 (2. 4%^ of all" spiders or 5.1% excluding Myr- TABLE2. Spiders in cells of Pison xtmthopus &i Logon, Ghana, 1969-1973. Key marachne), but the dif- to Places ai University ofGhana, Legon: Faculty af Agriculture, Agile.; Depart. fe rences are no t s jp of Botany, Bot.; Institute ot"Statistical, Social and Economic Research, ISSER; „;f;„„„nificant.f However,u^.,», ,*,,«.- ^«.one 36 Legon Hill, Hill; Depart, of Mathematics, Maths; Depart, of Zoology, Zool 1 as ,on 3 ' ^augm i - P ~> j I — 17 Thomisidae. 6 Thendiidae. 5 Clubionidae. 2 Oxyopidac. 1 Anncidae. )^ 2 3 - - Cosmophasis while all Sclenopidae, t Philodromidae. Oxyopidae. Gnaphosidae. Taxon head is abbreviated in order. My, Myrmarachne sp.; Co, Cosmophasis sp.; Ps. other wasps very rarely Pseudicius sp.; Rh, Rhene sp.; Te, Telamania sp.; Me, Menemerus sp.; Vi, took them. Therefore, even Viciria sp., So, Sonoita Ughrfaoti; Th> Htyeite sp.; Sa, Saitis sp.; Mo, Mogrus poor mimicry of ants ap- sp.; Fi, Fissident sp.; Os, other Sallicidac; Of, other spider families. pears to give some protection against most wasps, Do Individual Wasps Hunt Specific Prey? but occasionally a wasp will specialize on this All wasps do not take a similar spectrum of species, just as other wasps do with Mvr- spiders, but each individual preys on one or two marachne. species of spider (Tabic 2% Thus the first wasp in the table preyed on Rhene sp„ the second and Hunting by the Searching Image Method third on Pseudkius Sp. and the fourth on black The term searching image was used to describe species of Myrmarachne, Ant mimicry was ob- the way in which tits (Paridac) collect caterpillars viously of defensive value against the first three for their young: each bird tends to bring insects 1

WASP PREDATION ON ANT-MIMICKING SALT1CID5 511

'TABLE 3. Percentages of Myrmarachne (Myrm) and hypothesis makes no prediction about concern | $ other salticids on shrubs at Lcgon, and in cells of tion on one prey interfering with finding other Pison xanthopus. Figures in brackets are numbers prey. These observations do noi explicitly test and percentages found in open (i.e. not in retreats). either prediction, but I consider that Pison is noi

Source of spider* | N ) My nil Others hunting by adjusting ils search rate because dif-

Shrubs-all daia |82(63) 52.4 (61.W 47;#38.i) ferent wasps concentrate on different prey imply- 46.3 ing that capture of several prey of one species Wa&p cells 837 19 1 80,9 (59.K) interferes with capture of other species. The two Shrubs -omilling. M. wasps at 1SSER on 5lh and 6th of July 1972 took jbffii5tx&aL\ spideis with ants Qecophylhi 53(41) 49.1 (S&5) Sl9f»I.S) very different prey If they had been hunting & Crcmatogayitr simply with different search rates then the first 25.5 wasp should have taken sonic Mymu.n;: /»•,-< (34.6) | Wasp cells S3G 1?0 810 wlnle the second should have taken some Pseudicius This lota] concentration ofpredominantly a single species for several days on one species of prey is not easy to explain the search and then to abruptly switch to another species on rate hypothesis. ffinbergen, I960) Tinbergen hypothesized that the birds recognised caterpillars by particular If Pison is not capturing prey by adjusting its character which they "assimilated in a kind of search rate, (here are at leasi thice ollu-t hUfl learning process* and that in their search for prey methods that might result in the concentration on in they looked foi these particular characters. Later particular species of prey shown Tabic 2> •is have used the words 'searching image* t. Wasps might have some sort of preference mi 'search image' interchangeably, but its i/ picv rather than another, This is unlikely definition has been refined following Croze if the preference is based on taste because the tlo (1970>.Dawki us (197 1 ),Krebsil 973 r, Lawrence wasps fling bui not eafl the prey, and Allen 1.1983) and Guilford and Dawkins 2. Wasps could he searching in different

1 1 >H7). it is now generally understood to mean a place perceptual change in a predator that temporarily a. Individual wasps might search in different increases its ability to detect particular cryptic mierohabitais. If each species of spider lives in a prey which it has recently encountered. Search- slightly different microhabitat on the shnjbs then ing image needs to be distinguised from various individual wasps could catch different species of other types of preference that predators may show spider. Evidence in favour of this is the wasp on for particular prey (Krebs, 1973; Lawrence and 18 Mar 1972 which concentrated on ihomisidund

Allen, 1983), so it is probable that the behaviour other non-salticid spiders. Because it initially ohscrved by Tinbergen which he called hunting caught some of these spiders it is reasonab a by means of 'specific searching image' would suppose that it learned to hunt in particular areas noi be today considered a proven example of or in a particular way such thai it continued m searching image behaviour. catch these spiders instead of salticids. Individual Pison xanthopus clearly concentrate b. Wasps might search in one area before going

00 one ui a Few species of prey; is this because it on to another. If spiders arc clumped instead of hunts using a searching image or by some other being randomly spaced, then individual wasps Method? In a recent critique of the search image hunting in the same general area could capture concept Guilford and Dawkins (1987, 1988a.b) different species of spider. I often found two of argue that all studies purporting to demonstrate three spiders of one species on a shrub so the searching images can actually be explained better distribution is clumped rather than random. How- in terms of variation in search rate. They make ever, there is no exclusion of one species by two predictions which can distinguish between another, and rhe clumping may simply be of the two hypotheses. First, ihe search rate recently mated pairs or of parents with young that hypothesis predicts that mimetic prey should take have failed to disperse. This impression has not longer to find than non-mimics, so wasps that ken quantified, htn I oonsidei it unlikely thai it have learned to find them will slow down their could explain the extreme specialization on one search rate. The searching image hypothesis species (Tabic 2). makes no such prediction. Second, the searching 3. The wasps may have a perceptual searching image hypothesis predicts that because of percep- image as implied by Croze (1970). Wasps of tual specialization, wasps concentrating on one other genera can learn the configuration of

-.i-i' of prey will ignore others. The search rate landmarks near ibeh (see Tinbergen 512 MEMOIRS OF THE QUEENSLAND MUSEUM

J 958), and this presumably involves some per- spiders (Ataneldae) in Ghana. Acta Zoologica ceptual memory not unlike that in ourselves when Fennical90: 117-122. M. 1974. 'Defence in animals: a survey of we search for something with a particular image EDMUNDS, anti-predator defences'. (Longman: London). 357 in our mind. pp. It is unclear which of the above methods of prey 1978. On the association between Myrmarachne capture or if another method (e.g., Krebs, 1973) spp. (Salticidac) and ants. Bulletin of the British is involved. But certainly some such behaviour Arachnological Society 4; 149-160. occurs both in Pison and in other wasps, e.g. the GUILFORD. T. & DAWKINS, M.S. 1987. Search sphecid Chalybion fuscipenne (J. Edmunds, images not proven; a reappraisal of recent 1990). evidence. Animal Behaviour 35: 1838-1845. 1988a. Search image versus search rate: a reply to CONCLUSIONS Lawrence. Animal Behaviour 37: 160-162, 1988b. Search image versus search rate: two different ways to enhance prey capture. Animal Clearly, ant-mimicking spiders are common, Behaviour 37: 163-165. more so than non-mimics in some habitats. Ant HINGSTON, R.W.G 1927. Field observations on mimicry may give protection against wasps such spider mimics. Proceedings of the Zoological Society of London 56: 841-858. as Sceliphron which prey on spiders of various KREBS J.R. 1973. Behavioural aspects of probation families (Edmunds, 1974; J. Edmunds. 1990). Pp. 73-tH. In Balcson, P.P.G. and Klopfer. P.M. The results confirm the suggestion from a (eds) 'Perspectives in ethology. Vol. l\ (Plenum preliminary, less rigorous study (Edmunds, Press: London). 1974), that ant mimicry is of protective value LAWRENCE. E.S & ALLEN, i.A. 1983. On the term against the specialist predator Pison xanthopus. 'search image'. Oikos 40: 313-314. The evidence that behavioural ant mimicry MATHEW, A.P. 1954. Observations on the habits of protects Cosmophasis against Pison is not con- two spider mimics of the red ant, Oecophylta smarngdina (Fabr.). Journal of the Bombay clusive; it may be protective against other wasps Natural History Society 52: 249-263. or against birds. Wanless (1978) and Curtis OL1VEIRA, PS 1988. Ant-mimicry in some Brazilian (198.8) have evidence thai spiders of this genus sallicid and elubionid spiders (Araneae: Sal- prey on ants, so this mimicry could also be ag- in njae, Clubionidae). Biological Journal of the gressive rather than defensive in function. Linnean Society 33: 1-15. OUVRIRA. P-S. & SAZ1MA I., 1984. The adaptive ACKNOWLEDGEMENTS bases of anl-mimicry in a neotropical aphantochilid spider (Araneae: Aphamochilidaei. Biological Journal of the Linnean Societv 22: I grateful to Fred Wanless for help wiih am 145-155. identification of the salticids, Paul Hillyard for 1985. Ant-hunting behaviour in spiders with em- identifying the other spiders, and Janet Edmunds phasis on Strophiits nigricans (Thomisidac). Bul- for critically reading the paper. letin of the British Arachnological Societv 6: 309-312. LITERATURE CITED KHISKIND, J. 1977. Anl-mimicry in Panamanian elubionid and salticid spiders (Araneae: Clubionidae, Salucidae). Biolropica 9: 1-8. CROZE, R 1970. Searching image in carrion crows. REJSKIND. J. & LEVI, H.W, 1967. Anatea, an ant- Zeitschrift fiirTicrpsyehologie. Supplement 5; I- mimicking theridiid spider from New Caledonia 85. (Araneae: Thendiidae). Psyche 74: 20-23. CURTIS, B.A. 1988. Do ant-mimicking Cosmophasis TIN BERGEN, L. I960. The natural control of insects prey on their Campontfius models? Cimbcbasia in pine woods. I. Factors influencing the intensity (Windhoek) 10: 67-70. of predation by songbirds. Archives CUTLER, B. 1991. Reduced prcdaiion on the antlike Necrlandaises de Zoologie 13: 265-343. jumping spider Synageles occidentalis (Araneae: TINBORGEN, N. 1958. 'Curious naturalists'. (Country SaJticidae). Journal 401-407. ofInscctBehavior4: Life: London). 280 pp. DAWK1NS, M. 1971. Perceptual changes in chicks: WANLESS. F.R. 197S. A revision of the spider genera another look at the 'search image' concept. Bt'lippo and Myrmctrachne (Aruneae: Sallieidae) Animal Behaviour 19. 566-574. in the Ethiopian region. Bulletin of the British EDMUNDS, J 1990 Wasp predation on orb web Museum (Natural History) (Zoology) 33: 1-1 39. ABUNDANCE AND STRUCTURE OF FOSSORIAL SPIDER POPULATIONS

PETER G. FAIRWEATHER

Fairwcuiher, P.G. 1993 11 11: Abundance and structure of fossorial spider populations. Memoirs of the Queensland Museum 33(2): 513*518. Brisbane. ISSN 0079-8835.

Preliminary Findings from an ecological study of large fossorial spiders, especially rnygalomorphs, made in two locations in eastern New South Wales are given. The abundance ol burrowing spiders was assessed in eight habitats: dry sclerophyll forests on sandstone and shale substrata, wet sclerophyll forests on sandstone, pastures, suburban gardens, pine windbreaks along roadside verges, coastal cliffs, and coastal swamps. The spiders included several species of Lycosidae, two Idiopidae and one Hexathelidac. Each species was restricted in its range of habitats and dominated the burrowing spider assemblage in only one or two habitats- The population structure of burrow sizes is described and compared lor dense populauons.QVflrt^tt*. Idiopidae, Hexathelidae, Lycosidae, ecohgy, burrowing, sampling, distributions, habitats, size frequencies.

Peter G. Fairweather. Graduate School of the Environsnent, Macquarie University, New South Wales, 2109 Australia; 5 November, 1992,

The field ecology of burrowing spiders in Aust- hy Brisbane Waters National Park (Benson & ralia has rarely used quantitative methods (but sec Fallding. 1981). Study habitats were located ut Humphreys, 1988). Nor were field experimental m altitude At each location, all sites were manipulations used to test specific hypotheses, as within 2km of each other. I sampled six habitats in other branches of ecology (Underwood, 1990). in each location, although only three were present

I hese first quantitative observations of the abun- in both locations (Table I). dance, size structure and habitat associations of large burrowing spiders from two locations near Sampling and Analysis of Data Sydney are part of environmental studies of fun- The sampling. October 1991 -March 1992. was nchveb and trapdoor spiders. non-destructive. Quadrats (0.25 m") were ran- The abundance, distribution and ecological in- domly placed and then searched for individual teractions of fossorial spiders have been studied spiders, burrows and trapdoors. All rocks, logs overseas in detail Isee eg Buchli, 1969; Luins, and litter in each quadrat were overtumet' 1978, McQueen, 1983; Conlcy. 1985; Fernan- searched. I counted spiders, burrows or wehs of dez-Montraveta et .•;»'„ 1991; Miller & Miller, each species in each quadrat. 1 sampled 10 quad- 1991). The general biology, systematica and rats in each of two sites in each habitat at each evolution of Australian mygalomorph spiders location (n=240 quadrats in all). The paired sites have been studied (see Main, 1 976, Raven, 1 9R8 ). were replicate patches of habitat at the same The few quantitative ecological studies from elevation, as similar as possible, at least 100m Australia focussed on arid habitats (e.g. Main, apart and 100- 1000m" in area. Site boundaries 1987: Kotzman, 1990), upland sites (e.g. were located randomly on habitat maps before Humphreys, 1976), or used pitfall trapping (e.g. sampling. Spatial variability in each habitat was Curry et ol. 1985). In contrast, this study was on assessed by comparing spider densities in each lite warm temperate coast of eastern Australia. pair of sites. This design allowed a three-factor hierarchical METHODS ANOVA to examine the partitioning of spider abundance (as density-no. per quadrat) among Study Sites and Habitat Types the fully nested factors of Location, Habitat with- Two locations near Sydney, New South Wales, in Locations and Site within Habitat (Sokal and

and situated on Hawkesburv sandstone sub- Rohlf. 198 1 ) ANOVAs were done after assump- Ntratum, were sampled. Galston (33°4rS, tions were checked using Cochran's test for het- 150°21'E) is a semi-rural village that has become erogeneity of variances Means and standard suburban since 1972 and retains pockets of bush. errors of spider densities were calculated for cad:

Study habitats were located at 1 70-2 1 0m altitude. habitat and site All statistical and graphical ai.ai , Patonga (33°30 S, ISriS'B) is a coastal village yses were performed with SYSTAT Version 5 on the Hawkesbury River estu3ry and surrounded software (Wilkinson, 1990). 514 MEMOIRS OF THE QUEENSLAND MUSEUM

Location Habitat type Trees Shrubs Ground layer Litter Common spider Less common spiders

Galston Wet sclerophyl forest (WSF) dense dense dense, rocks dense, moist Atrax wolf

Dry sclerophyll forest (DSF) sparse dense little, rocks dense, dry none M. rapax, wolf

Garden lawns (GL) none few grass none none M. rapax, wolf

Shale turpentine forest (STF) dense few grass, logs sparse, moist Atrax wolf, M. rapax

Pine road verge (PRV) pines few some grass deep, dry M. gracilis wolf M. rapax

Pasture none none grass little none none

Patonga Wet sclerophyll forest (WSF) dense dense dense, rocks dense, moist none wolf, M rapax, Atrax

Dry sclerophyll forest (DSF) sparse sparse little, rocks dense, dry none wolf

Sea cliffs (SC) few few some roots, rocks .sparse, moist M. gracilis wolf, M. rapax

She-oak swamp (SS) sparse none some dense, moist M. rapax none

Garden lawns (GL) none few grass little none M. rapax, wolf

Sports ground none none grass none none none

KEY TO HABITATS STF: open forest in Fagan Park on ridgetop shale Howell DSF: characteristic open woodland (Benson &FaHd- Wianamatta (Benson & 1990), dominated turpentine, Syncarpta ing 1981; Benson tfHowcll 1990) in exposed posi- by glomuhfem tions on Hawkesbury sandstone plateaux, otner lT?™ included E. punctata, E pamculata E dominated by Angophora costata. Eucalyptus acmenwides E. resmifera, E globoidea and A. flonbunda', shrub layer mainly of young trees; haemastoma, E. gummifera and E sparsiflora ; un- ground cover of grasses, logs and sparse litter. derstorey of sclerophyllous shrubs of Fabaceae, ProteaceaeandMyrtaceae;!itterIayerdensebutdry; SS: estuarine swamp dominated by she-oaks, sandy soil; many ant nests and sandstone outcrops. Casuarina glauca; ground layer oiJuncus kraussi, she-oak needles saltmarsh succulents; lowest WSF: in moist gullies and other sheltered locations; and area inundated by highest tides, but areas around many of the same tree species in the canopy, also E. * s d st°ne ontxmps higher and drier. Bushfire burnt piperita and E. eximia; compared with D§F; shrubs *P k and creepers more mesophyllic and denser, with th,s hab,lat and dry sclerophyll forest early 1991. pockets of smaller trces-Pittosporum undulatum, SC: sea cliffs immediately above rock platforms in Tristaniopsis laurinamd Ceratopetalumapetalum; the estuary; vertical walls of loamy soil covered litterlayerdense, moist, compacted; ant nests fewer; with sparse leaf litter, mosses; some creepers and soil with more humus but about the same amount of grasses common around roots of trees and snrubs on rocky outcrops. cliff edge. The least extensive habitat in area; 10 m

or ,ess h a,on seasnore - GL: cultivated exotic grasses regularly mown to <4 ™8 S cm high, and probably also treated with fertilisers. At Galston, the last habitat was pasture grazed by herbicides and pesticides; little bare ground and horse and cattle. Grass height, 6-40 cm; grasses almost no litter. probably fertilised infrequently; patches of bare soil

rare - PRV: pine trees (Pinus radiatd) along road verges, rarely mown and graded; deep litter of pine neecfles; At Patonga, a lawn and bare ground area around a soil with humus-rich, dry layer above hard clay- sporting ground was the last habitat, very similar to loam; understorey some Acacia and Pittosporum the pasture at Galston except grass mown regularly shrubs, & grasses. to <4 cm high.

TABLE 1. Habitat characteristics, with spiders and occurrence of rocks, logs or tree roots. Common spiders =

density > 1 per m . Abbreviations used in text are also given for each habitat.

Species were identified by burrow charac- March) in some habitats to assess temporal varia- teristics and observations of spiders seen at the tion between, before, and after the breeding burrow entrances either at night or during late season (roughly mid-sampling). In particular, afternoons on overcast days. Initially, I ex- abandoned burrows (with signs of decay and cavated at least 30 burrows of each type to obtain unoccupied) were noted. Burrows under rocks specimens for more thorough identification, and and logs were counted in some habitats because to investigate burrow structures and food funnelwebs seemed restricted to such locations (e.g. short burrows in unconsolidated soils). remains. I measured maximum diameter of all burrows in quadrats (and outside them at some RESULTS sites to increase sample sizes to >30). I tested for differences among size frequency distributions of burrow diameters using Kolmogorov-Smir- Spider Species nov (KS) 2-sample tests. Six species in three families are here grouped Sampling was repeated in late summer (Feb.- into four ecological types. Misgolas gracilis BURROWING SPIDER ECOLOGY 515

(Rainbow &Pulleine, 1918) is a large idiopid that builds deep, oblique burrows with a trapdoor among leaf litter in friable soils. The lids varied from flimsy and merely silk-covered with a thin layer of dirt to quite robust plugs for older spiders; this may be related also to the amount of litter present. In moist areas, the lid often grew moss and liverworts, M. gracilis was found only in SC and PRV habitats (Table 1).

Misgolas rapax Karsch, 1 878 is also large with a burrow like M, gracilis but without a trapdoor. Often litter and vegetation around the burrow entrance were incorporated into the flared opening. Their burrows were more vertical than those of M. gracilis. The idiopids were identified using Main (1985), Mascord (1970. 1980). M. rapax was found in eight habitats but abundantly only toSS Clfcbte I), Atrax robustus Cambridge, 1877 is the Sydney funnel web spidcr(Hcxalhelidae). Several similar species are known from areas near the study

i ions, but all spiders collected were identified U$ A. rohustus using Gray (1988), Scott (1980), Main (1985), Mascord (1970, 1980) (some smaller spiders were minimally confamilial). Most spiders were found under rocks and logs where characteristic silk tubes led to shallow FIG. 1. Abundance (no. per 0.25nT) of each species burrows made in mostly unconsolidated soils and group versus habitat. A at Galston, habitats are: humus. Thus, all spiders were examined in the I=\vsf. 2=dsf. 3=STF. 4=prv, 5=GL, 6=pasture. B ai Patonga- habitats are: 1=wsr 2=dsf, 3=SC. field but collections were limited to avoid deple- 4=SS. 5=GL, 6=sport ground. MR=Af. rapax, MG=Af. tion of the populations and for safety. Atrax was xracMs, fw= Atrax and wouMycosids. Means and luund in onlv two habitats but was a dominant in standard errors calculated from n = 20 quadrats in both (Table 1>. each habitat (i.e. sites were pooled). Wolf spiders (Lycosidae) built narrow, vertical burrows without lids and with much flimsier silk sparse populations (of only lycosids and M. linings than did M. rapax. They were identified rapax) were found in open habitats-DSF and GL using McKay (1985) and references therein. The The abundance in the five occupied habitats at species excavated were Lycosa godeffroyi L each location showed differences between Koch, 1865, L. leuekartii (ThoreW, 1870) and habitats and species groups. An ANOVA of total Pardosa serrata (L, Koch, 1877), the latter with spider density among locations/habitats/ sites characteristic palisades around the burrow. Other (Tabic 2) showed no significant difference be- species but not excavated include L.furcillata L tween the locations but large differences among Koch, 1 867, L. pictiventris L. Koch, 1 877, and L. the habitats within locations. At Galston, highest palabunda L. Koch, 1 877 Due to this uncertainty spider densities were found in PRV and WSF over the exact identity of the occupants of some habitats, fewer in STF and very few in GL and burrows, 1 lumped data on all lyeosid burrows DSF habitats. At Patonga, SC had the greatest 1 into 'wolf spiders I found these in nine habitats densities, followed by WSF and SS habitats, and but most commonly in WSF i at Patonga only ) and very few in the GL and DSF habitats. PKV (Table I). Habitats with dense spider populations in either location were dominated by a particular species. Abundances At Galston, Atrax was the most common spider No spiders or burrows were found in quadrats in the WSF and STF habitats but was not found sampled in the pastures or sports ground (Table in other habitats. At Patonga, Atrax was found 1. Fig. 1), although lycosids had been seen there only in Ihe WSF habitat. M. rapax occurred in Neither habitat will be discussed further Very four habitats at Patonga and Galston, but 516 MEMOIRS OF THE QUEENSLAND MUSEUM

5? LUiA I 4 5 • 7 a 12 3 4 9 6 7 8 10

DENSITY I NUMBER i i s 4 a a 7 a AJ 7 a £ • i a s * 9 12 3 4 C_3 SITE

FIG. 3. Abundance of Atrax under rocks and logs in some habitats, expressed as no. per rock or log, and density (no. per m of rock or log microhabitat). Area of each rock and log estimated from product of two, LllIlL perpendicular linear dimensions in contact with the i c a 4 s a 7 a 2 3 4 9 8 7a ground. Sites (sample sizes) were: 1=WSF at Galston (n = 43 rocks); 2=DSF at Galston (h=27 rocks); 3=STF at Galston (n =20 logs); and 4=WSF at Paton- ga («= 18 rocks).

and Rohlf, 1981). For total spider density and M. __» rapax alone, the two sites sampled within each 1 E 3 « 6 S 7 8 12349979 habitat differed significantly (Table 2) . Because the sites were chosen randomly from the total SIZE CLASS habitat, this result indicates medium-scale variability in abundance of total spiders and M. FIG. 2. Representative size frequency distributions of burrow diameter for dense populations of each rapax (i.e. at scales of about 100m). species. Diameters grouped into eight equal size classes (to show large range with reasonable numbers in Size Frequencies each):l=0-4.9mm;2=5-9.9;3=10-14.9;4=15-19.9; Sample sizes of burrow diameter of sufficient 5 = 20-24.9; 6 = 25-29.9; 7 = 30-34.9; and 8 ^ 35mm across the burrow entrance. A) Atrax from WSF, number could be obtained only in habitats with Galston, n = 95; B) Atrax from STF, Galston, n = 59; dense populations. The burrow size structure C) Atrax from WSF, Patonga, n = 50; D) lycosids (Fig. 2) of these populations showed differences from PRV, Galston, n = 38; E) M. gracilis from SC, (by KS tests) among species that were consistent Patonga, n - 189; F) M. gracilis from PRV, Galston, across habitats. Very small burrows (< 5mm) = = n 83; G) M. rapax from SS, Patonga, n 102; H) were not found for M. rapax and only commonly M. rapax from GL, Galston, n = 79. for lycosids. Atrax showed bimodality in two habitats with the modes occurring at sizes cor- dominated only in the SS habitat. M. gracilis, in responding to pre-reproductive juveniles and ma- contrast, was found in only one habitat in each ture females (Fairweather, unpublished data). location, but dominated the spider assemblage in The other species had more unimodal burrow both. In the SC habitat at Patonga, this species had diameter frequency distributions. Lycosid bur- the greatest mean density found in this sampling 2 rows were much narrower on average than the (>12 m ). Lycosids were found in five habitats mygalomorphs (Fig. 2). M. gracilis had the at Galston and four at Patonga but they were largest burrows overall. never dominant. No species group was either positively or negatively correlated with any other in these samples (for all r, P> 0.05, n = 200). Temporal Variation The patchiness of mean spider abundance in Sampling before and after the breeding season any habitat was examined by the Sites within showed few changes. The proportion of burrows Habitat factor in the nested ANOVA (see Sokal of M. gracilis in the PRV habitat (Galston) that BURROWING SPIDER ECOLOGY 517

were abandoned and decaying increased from Factor df f-ralios

4.6% (n= 64) in December to 9.6% (n = 60) in All M. Atrax M. gmtitis Wolt March. These abandoned burrows did not, how- spiders rapax ever, differ in size from the occupied ones (P> Location 1 (.. i*j 1,2! |0,3K 1 91 0.1X13 0.05 by KS tests), suggesting no size selectivity Habiiat y 73.43* B.93 0.88 I3.fi" in either mortality or abandonment by breeding 10 33 25 LI 2> Si: y.54« 31 males. The proportion of Atrax that were juvenile TABLE 2. Three-factor, hierarchical ANOVA of den- (i.e. <15mrn and with no enlarged pedipalps on sity (number per quadrat) of total spiders and each males) increased from November to February species group analysed separately. ^Significance at from [n = 67) to (n = across all three 12% 44% 84) P <0.05 level; df, degrees ol iTeedojpjdffor residual, occupied habitats. 180. and for total, 199. Toial N=2Q0 0.25 m2 quad rats; only Jive habitats used here MlCROHABlTATS ferem species. As well as the above habitat Within each habitat, burrows were found more segregation, microhabitat preferences were also frequently in particular situations. For example, shown by several groups, most strikingly tor M. rapax in the SS habitat iPatonga) were only Atrax, which was found in moist areas under found in areas around rock outcrops and none in shelter. Predictive relationships of abundance the lower, inundated part of the swamp. M. with environmental variables te.g. soil nutrients, gracilis was most abundant in moist, mossy organic matter, compaction and moisture; patches in SC habitat (Paconga) and in areas amount, moisture and temperature; size and depth covered with litter rather than bare ground in the of shelter) may be established. Experiments on PRV habitat (Galston). All Atrax webs en- the effects of shelter, litter and moisture condi- countered were seen under or against cither rocks tions on the abundance of these spiders arc •r toga, although thorough searches were also needed. litter in grass made amongst and clumps. This 3, Fossorial spiders respond adversely to many prompted sampling centred on rocks or logs in human impacts on their environments. This has three habitats at Galston and one at Patonga (Fig. implications foi the interaction of these Spiders 3), The abundance of Atrax under rocks and logs with people and their activities. Few or no fos- habitats. in differed with None were found DSF, sorial spiders were resident in habitats that lacked despite abundant rocks. Counts of webs per rock a litter layer or were regularly mown, watered, or lo^ w ere similar ul the three occupied habitats. r treated w ith chemicals or graded . The spide- When expressed on a per area basis (i.e. rrr of burrow in such open habitats (Fairweather pcrs. rock or log), the densities were much greater and obs. f, so perhaps the conditions may not be a] differed among these habitats (Fig 3), tive to prey. There is some longer term evui. of declines in two of these populations associated DISCUSSION with increasing Urbanisation, direct distu bushfirc and vegetative change (Fairweather This study suggests several hypotheses. unpuh il.u;ii 4. established /. These species rarely encounter each other in Dense populations have been for nature, suggesting little VOmth ftlicn aittirs several years, at least Size frequency distributions of burrows in each dense among them. I located each species group in more revealedjuveniles lhan one habitat, but they tended to dominate population, therefore recruitment had occurred different ones. /t/nu were favoured by apparently and no population was relict. Several very large more moist conditions under shelter: although burrows were present in the populations of the exfoliated rocks were abundant in the drier DSF three mygaloraorphs, probably indicating habitat, no Atrax were found beneath them. A/. matriarchs [sensu Main, 1987). The abundance gntcilisVNB found in relatively exposed positions and size structure did not alter from October lo f&ea cliff and road verge) with the most compact March, which implies short-term stability for soil, whereas M. rapax dominated more open these long-lived spiders. Behaviour consistent habitats (in terms of the litter and ground layers'). with breeding behaviour over summer was seen Lycosidswere the most widespread group, which for the mygalomorphs. Lycosids with egg sacs may reflect that data of several species were were seen only in spring and autumn. lumped Cliaracien.sahon ol habitats regarding 5, Prediction by some populations may $tf\ soil, litter and vegetation conditions is needed. influence the assemblages of their prey* \1 2. Specific habitat characteristics favour dif- gracilis and Atrax were quite dense in particular S1S MEMOIRS OF THH QUEENSLAND MUSEUM

microhabitats, wiih some very large spiders; this. jroyi (L. Koch 1865) (Araneae: Lycosidae! . 59-80. and their predatory habits suggest that their role Dal of Animal Ecology 45:

1 . The ecology of spiders with special reference as predators in the ground-layer ecosystem would 98fl to Australia. Pp. 1-22. In: Austin, A.D. aixl be worth further study.

Hea(her? N.W. (eds). 'Australian Arachnology'. In conclusion, large fossorial spiders are not (Australian Entomological Society Miscel- evenly distributed across a variety of habitats, and laneous Publication No. 5: Brisbane). each habitat is dominated in numerical terms by KOTZMAN, M. 1990. Annual activity patterns of the one or few species. Although the study was done Australian tarantula Seletwcosmia stirfingi in two contrasting locations, the generality of (Araneae, Theraphosidae) in an arid area. Journal these results awaits scrutiny with further data as of Arachnology 18: 123-130.

. J. 1978. Studies on populations of the tunnel does the cause of any of the patterns described for LAFNG D qteb spider Porruthele antipodtana. Tuat.'»r;> 23: the first time here. 67-RI. MAIN, B.Y. 1976. 'Spiders*. (Collins: Sydney) ACKNOWLEDGEMENTS 1985. Arachmda: MygaJomorphae. Pp. 1-48. In Walton, D.W. (ed.) Zoological Catalogue of Australia Vol. 3\ (Australian Government This study Utras supported by a Macquarie Publishing Service; Canberra). University Research Grant and benefited from 1987. Persistence of invertebrates in small areas; the assistance of N. Babicka. G. Napier com- case studies of trapdooT spiders in Western mented upon a draft of the manuscript and two Australia. Pp. 29-39. In Saunders;, D.A., Arnold, anonymous referees improved its clarity. G.W., Burbridge A.A.and Hopkins, AJ.M (eds) 'Nature conservation: the role of remnants of LITERATURE CITED native vegetation'. (Surrey Beatty & Sons- Syd- ney). MASCORD, R. 1970, 'Australian spiders in colour*. BENSON, D. A: HOWELL, J. 1990. Taken for gran ted: (Reed: Sydney). the bushland of Sydney and its suburbs'. (Kan- 1980. 'Spiders of Australia: a field guide' (Reed; garoo Press: Sydney). Sydney). BENSON, J.S. & FALLDING, H. 1981 Vegetation MCKAY, R..I. 1985. Arachmda: Araneomorphae: survey of Brisbane Water National Park and en- Lycosidae. Pp. 73-88. In Walton, DAV. fed..) virons. Cunninghamia 1: 79-1 13. Zoological Catalogue of Australia Vol V 1969. behavior in (lie BUCHL1, H.H.R. Hunting (Australian Government Publishing Service: Can- American Zoologist 9: 175-193. Ctenizidae. berra) CONLEY, MR. 1985. Predation versus resource MCQUEEN. D.l. 1 983. Mortality patterns for a popula- limitation in survival of adult burrowing wolf tion ol burrowing wolt spiders, Ceolycosti Oecologia 71- spiders (Araneae: Lycosidae). 67: domifvx (Hancock), living in southern Ontario.

Canadian Journal of Zoology 61 ; 2758-2767. CURRY, S.J., HUMPHREYS, W.F., KOCH, L.E. & MILLER, P.R. & MILLER. G.L 1991, Dispersal and MAIN, B.Y. 1985. Changes in arachnid com- survivorship in a population of GeOfyCOXd fur* munities resulting from forestry practices in karri r/co/a (Araneae, Lycosidae i. Journal of Arachnol- forest, south-west Western Australia. Australian ogy 19:49-54. Forest Research 15: 469-480. RAVEN. RJ, 1988, The current status ol Australian FERNANDEZ-MONTRAVETA, C. LAHOZ ( spider systematica. Pp. 37-47 In Austin, A.D. and BELTRA, R. & ORTEGA, J. 1991. Spatial dis- Heather, N.W. (eds) 'Australian Arachnology'. tribution tarermilu it vent rix of Lycosa fuse < (Australian Entomological Society Misccllnnt n i (Araneae, Lycosidae) in a population from central Publication No. 5: Brisbane). Spain. Journal of Arachnology 19: 73-79. SCOTT, G. 1980. 'The funnelwcb' (Darling Downs GRAY, M.R. 1988. Aspects of the systematica of the Institute: Touwoomba). Australian funnel web spiders (Araneae: SOKAL, R.R. & ROHLF, F.J. 1981. 'Biometry'. Hexaihelidae: Alracinae) based upon morphologi- (Freeman: New York). cal and electrophoretic data. Pp. 1 13-125. In .Aus- UNDERWOOD, A.J. 1990. Experiments in ecology tin, A.D. and Heather, N.W. (eds) 'Australian and management: their logics, functions and inter- Arachnology'- (Australian Entomological Society pretations. Australian Journal of Ecology 15:365- Miscellaneous Publication No. 5: Brisbane). 389, HUMPHREYS. W.K 1970. The population dynamics WILKINSON, L 1990. SYSTAT. the system for of .m Australian wolf spider. Geotycosa (odef* statistics', (SYSTAT: Evanstan). ;

BIOGEOGRAPHIC PATTERN OFPHYLOGENY IN A CLADE

OF ENDEMIC HAWAIIAN SPIDERS ( ARANEAE, TETRAGNATHIDAE

ROSEMARY G. GILLESPIE

Gillespie. R.G. 1993 II 11: Biogcographic pattern of phytogeny in a clade of endemic Hawaiian spiders (Araneae- Tetragnaihidae'f. Memoirs of (he Queensland Museum 33(2): 519-526. Brisbane. ISSN 0079-8835.

The birOia of the Hawaiian archipelago offers an ideal system with which in study the dynamics behind the evolutionary process, both because the islands harbour many specinse lineages, and because they are arranged within a chronological lime frame. Over the past 5 years I have begun to uncover an unexplored radiation of one of Hawaii's most abundant and conspicuous invertebrate groups: the spider genus Tetragnatha. The current study focuses on a small clade within the lineage, in which all the component species have

abandoned web-building, instead foraging as cursorial predators. I examine 2 primary questions: 1. What has been the relative importance of strict geographic isolation (populations on different volcanoes) versus divergence between contiguous habitats (populations on the same volcano) in the evolution of this clade? 2. Does thephylogeny indicate a pattern of ecological and distributional change which could suggest that ecological rather than sexual

shifts may underlie species formation? I generated a phylogeny based on morphological characters, and compared this phylogeny to the biogcographic pattern oi Lhe Hawaiian Islands. The resulLs suggest that, for this clade of cursorial species, speciation requires strict

geographic isolation, and ecological (more than sexual) shifts appear to play SI role in

initialing divergence. Considering the islands as a series of evolutionary snapshots, 1 would also speculate that speciation is commencing on the youngest island (Hawaii), and develop" [fig on the adjacent older island of Maui- [jTetragnatha, phyloveny, Hawaii speriation, adopatry,

Rosemary G. Gillespie, Hawaiian Evolutionary ffology Program, University of Humid ot Manoa, Honolulu, Hawaii 9682Z U.S.A.: 28 October, 1992.

Species represent one of the basic units of to minimize the resources jointly used by both evolution, yet the processes by which they arc species, leading to further ecological divergence formed remain poorly understood (Mayr, 1963) (Mayr, 1963; Grant, 1986). Studies of the Hawaiian biota have lent considei The Hawaiian archipelago (Fig, 1) provides a able insight into the mechanisms underlying natural laboratory for studies of speciation speciation. These studies are highlighted by the (Simon, 1987). First, the extreme isolation of the Hawaiian Drosophifa. in which sexual selection islands has allowedrepeatcdandcxplosivediver through female choice appears to play an integral siflcation of species in a large number oflincag.es role in inducing species formation among small including honeycreepers (Bergen 1981; Fre- populations colonizing geographically isolated ah. 1 987) Jand snails (Cooke ef a/., 1960), crick- islands (Carson, 1968; Carson and Kancshiro, ets (Otte, 1989) and drosopbilid flies (Kaneshiro 1976; Kaneshiro. 1988). One may ask whether it and Boake, 1987). Further, the islands are a Berks is possible to generalize from these studies that of volcanoes arranged within an identifiable adjustments in the sexual environmental largely chronological time frame; the currently high is- responsible for driving species radiations. Other lands range from Kauai, the oldest and most studies outside the Hawaiian Islands have found eroded, to Hawaii, \hc youngest, highest and that ecological changes in isolation arc more im- largest, with 5 separate volcanoes. portant in driving species proliferation (Mayr, This study uses a lineage of spiders to examine 1963; Grant, 1986). When a species is released speciation patterns within the context of tin- from interaction with related species, by Hawaiian archipelago. The spiders belong to lhe whatever means, it may broaden its habital use long-jawed orb-weaving genus Tetrugtiathu, iihd exhibit much more variation among in- which comprises a large number of endemic dividuals (Lack, 1971; McCune, 1990). The ar- species in the Hawaiian Islands (Gillespie, 1991, gument is that if such a re productively isolated 1992), Outside the archipelago, Tetragnatha arc incipient species were reunited with its parent, among the most widespread and conspicuous selection could act on ihe ecological variability spiders worldwide, yet collectively they are also 520 MEMOIRS OF THE QUEENSLAND MUSEUM

KAUAI

Waialeale 5.1 myr O'AHU Koolaus WaianaesW^*-6 myr 3 -» 3.7 myr WW MOLOKA'1 1.8 myr MAUI West Maui 1.3 myr

Direction of Haleakala LANA'I w movement of 0.8 myr 1 .3 myr tectonic plate & relationship HAWAII to island Kohalas age (myr) 0.43 myr Mauna Kea Hualalai 0.38 myr 0.40 myr N

Ma una Loa Kilauea t 0.20 myr 0.10 myr

FIG. 1 . Major land masses of the Hawaiian archipelago, indicating approximate age, and direction of movement of tectonic plate.

one of the most homogeneous, both in morphol- haviour. Because explanations for species forma- ogy (elongate bodies and long legs) and ecology tion in the Hawaiian Drosophila rely heavily on (orb web generally built over or near water) the elaborate courtship behaviour of the group, (Wiehle, 1963; Levi, 1981; Gillespie, 1986, the absence of such behaviour in the Hawaiian 1987). Hawaiian species of Tetragnatha repre- Tetragnatha suggests that alternative explana- sent a paradox, exhibiting considerable morph- tions might be required to account for species ological and ecological diversity. Until 1992, the proliferation. The questions I address in this study sole reference to endemic Hawaiian repre- are: 1. What has been the relative importance of sentatives of the genus was based on descriptions strict geographic isolation (taxa diverge on dif- of a single species by Karsch (1880) and 8 species ferent volcanoes) versus divergence between by Simon (1900, redescribed by Okuma, 1988). I contiguous habitats (taxa diverge on the same have now described an additional 16 species (Gil- volcano) in the evolution of the spiny-leg clade

lespie, 1 99 1,1 992), and have collected more than of Hawaiian Tetragnatha? I generated a 60 new taxa that span a broad spectrum of phylogeny for the clade based on morphological colours, shapes, sizes, ecological affinities, and characters, and then compared the phylogeny to behaviours. In terms of courtship behaviour, the biogeographic pattern and history of the is- however, Hawaiian representatives of the genus lands. 2. Does the phylogeny indicate a pattern of display the simple cheliceral locking mechanism ecological and distributional change which could

characteristic of the genus (Levi, 198 1 ; Gillespie, suggest that ecological rather than sexual shifts pers. obs.). may underlie species formation?

Here I examine a small clade (the 'spiny-leg' clade) within the radiation of Hawaiian Tetrag- METHODS natha. Representatives of this clade are characterized by a cursorial habit, and do not build webs Collection and Ecological Measurements (Gillespie, 1991). Further, in common with other Spiders were collected by visual night search-

representatives of the genus (Levi, 1981), but in ing at various times of the year between 1 987 and striking contrast to the Drosophila radiation, 1991 in wet, mesic and dry native forest in all of these spiders display minimal courtship be- the currently high Hawaiian Islands (Kauai, PATTERN OF PHYLOGEN V OF HAWAIIAN SPIDERS S21

Qahu. Molokai, Lanai, Maui and Hawaii; Fig. 1). RESULTS Habitats from which spiders were taken were scored as wet (>450cm average annual rainfall), Collection and Ecological MeaSURHMH mesic (250-450cm) or dr>' (<250 cm). Elevation Representatives of the spiny-leg clade of was categorized as low medium-low (2000m). Mierohabitat associations were T. brevignatha Gillespie, T. restricta Simon and determined by categorizing the specific site from T. quaslmoda Gillespie, which occur in wet, which an individual was collected (roots, fern mesic. and sometimes dry. forest. The ranges fronds, against bark, etc.) (Gillespie, 1987). over which the different species were found is listed in table 2. Mierohabitat associations were loose, although the bright green species (T. tan- Phylogenetic Analysis udas Gillespie. T. polychromatQ Gillespie, T, brevignatha^ T. macracantha Gillespie, T. 1 used a cladistic approach (Hefting, 1966) waikamoi Gillespie and T. kauaiensis) were col- based on morphological characters to determine almost entirely from leaves, whereas die relationships among the spiny-leg clade of lected darker coloured T. kamakou Gillespie, T. per- Hawaiian Tetragnatha. I scored a total of 30 reirai Gillespie, T. pilosa Gillespie. T. quasimodo characters relating to cheliceral armature f upper and T. restrtcta were collected from brown or and lower tooth rows), leg spination, and colour red-brown substrates. of the cephalothorax and abdomen (Table 1). In addition, 1 scored characters from the detailed siruciure of the male palp using a Hitachi S-800 PHVL.onF.vFiic Analysis scanning electron microscope. I used a Hawaiian When characters were scaled for equal weight- web-building species of Tetragnatfra, 7, ing regardless of number cf states and unordered, stelarahusta Gillespie as an outgmup in the a total of 7 most parsimonious trees were analysis because molecular data indicate that this generated (consistency index 0.517. retention species belongs to a closely related sister clade of index 0.509). Subsequent weighting by successive approximations had little effect on the tree the spiny-leg species (H.B. Croom r pers. comm.l. topology, and gave a single tree of unweighted Characters were analyzed as unordered states length 76 (consistency index 0.725, retention (i.e., any character state permitted to transform index 0,765). Fig. 2 shows the tree with explana- directly into any other state) using Fitch (Fitch, tions of the characters defining each node. The 1971) and Wagner (Farris, 1970) parsimony in characters defining species are marked as bars. PAUP (Swofford, 1990) under the accelerated transformation method of optimization. Charac-

Rf 1 .\T10NSHIP BETWEEN SPECIES PhYLOGENY AND ter states were polarized as primitive or derived Island Biogeography by outgroup comparison (Maddison et al. t 1984). and characters were scaled for equal character As can be seen from this phylogeny based on morphological characters (Fig. the most close- .

this I changes have accompanied species fonnah I whereas representatives of species on compared the resulting phylogcny to the Maui occur only in mesic forest at middle eleva-

. -ographic locations of the component laxa tions Distributions ofrepresentatives of the clade wiibin the Hawaiian archipelago. on East Maui show some anomalies. In particular, — '

522 MEMOIRS OF THE QUEENSLAND MUSEUM

Kiel KllU pit mac pol tant wak brev perr kam rest J !!.,'-:

1 ! I 1 1 1 ! 1 1 9 first tooth: tiny/moderate size/a* large at others 1 1 2

' ' ' i

2 <5 first tooth absent rtnunp/fingcr 2 2 2 I i 2 2 2 2 2

3 d* *«r f first looth down margin): benl up/Slraigot/beiH down 1 1 1 1 1 1 I 1 1 2

I] 1 1 1 4 1 6 'sl'closc to T' (second looth down m3rpn>?

(lower present? 1 1 1 1 1 D 5 | d apical tooth cheltcera)

6 No. laigc teeth on lower margin 6 chelieera; 1/ 2 1 1 1 I 1 1 1 1

7 '5 lower tooth row: short/ ton-: 1 1 I 1 1 1 1 1

1 1 1 8 A teeth 3 and 4 on lower margin much smaller man rest? 1 1 1 I 1 1 1

1 J 9 d teeth 5 onwards larger Lhan 3 and 4? 1 1 i I 1 1 1 1 1 !

10 r?' fusi two teeth well separated? 1 1 1 1 1

Curl on terminal projection, conductor none/ slight/ ? 11 1 1 1 2 1 I 1 2 complete

Terminal projection of conductor points: straight/ 9 12 2 2 2 u 2 2 2 2 2 I backward/ forw aid

- l I I I Cap of conductor tip; shallow/ deep 1 1 L

14 Cap ridge of conductor tip: lateral/ medial 1 1 l,< 1

Backward projection of conductor tip: above/ at same level/ IS 1 1 1 1 I 1 below cap

1 i 16 Spur of conductor tip: indistinct/ prominant ll 1 1 I I 1 I 1

: ;r,<> .,] ._< ,-, j 'ii.'M.,«- 1 I 17 ii|-v .lngled up.' su ,i\.::v< out/ hooked down 1 2 2 1 1 2 jo

Floor and spur base of conductor lip: at same level/ 18 o u 1 1 1 1 I I separated

19 Separation of conductor cap and pleats: large/ small I 'J o

1". 20 Cap of conductor: * \dd medium/ high 2 1 2 2

21 Cap of conductor: rounded/ pointed riehV flai (j 2 1 1 2 2

1 22 Tip of conductor Iwists to show underside? ] I 1 1 I I 1 1 I 1

23 Venter color; translucent/ dark 1 1 1 1 1 1

1 II t! l 24 Venter pattern; plain/ rne Oiled Spots. 1 1 2 1 1 1 n ]

1) l 25 SierTium color, i_r>m*luecn'V opaque 1 1 1 1

1 i i i i :t) Orb webs buili? 1 1 ] i i 1

27 Tip of ct conductor proieciion: Hunt/ pointed I I 1 ! i 1 J 1

receptacles: swelling/swelling Seminal no angled down/ 9 28 2 2 2 2 i 1 2 2 angled up

29 Dorsum color: brown/ variable/ green I 2 2 2 2 2 a 30 Tibial spines tlatenil.medial. dorsal): 332/ 442/ 552 1 2 2 1 2 ! 1

TABLE 1 . Characters used for generating phytogeny . stel = T. stelarobusta; kau = T. kauaiensis; pil = T. pilosa: mac = T. mucracantha; pol = T. (.wlychromata: tant = T. tantalus; wak = T. waikatnai: brev = T. brevtgnatha, pen* = T. perreirai; kam = T. latrnakou; rest = T. restricta; quas = 7. quasimocfo. there are three bright green species, one endemic phylogeny based on morphological characters bo this volcano (T. macracantha), one shared with therefore strongly suggests that strict geographic West Maui (T waikamoi) and T, brevignatha isolation (between islands only) is necessary for shared with Hawaii Island. The East Maui species the initiation of species formation. Such isolation exhibit parapatric ranges, with only very narrow appears also to underlie speciation events in the zones overlap, and are more closely related to of Hawaiian Drosophila (Carson and Templeton ? species on other islands rather than to each other, 1984). The phylogeny of the Hawaiian spiny-leg Tetragnatha also indicates that species colonize DISCUSSION in a generally southerly direction, with the most ancestral taxa occupying the oldest island, Kauai. Differentiation between species of the spiny- In addition, colonization of the most recent island leg clade of Hawaiian Tetragnatha appears never (Hawaii) may be associated with ecologicaJ to have occurred on the same mountain mass: in release: populations of each of the three species no situation are two sister species found on the that have colonized Hawaii Island, T, quasimodo, same volcano, or even on the same island. This T restricta and T brevignatha, occupy a broad PATTERN OFPHYLOGENY OF HAWAIIAN SPIDERS 523

T. macracattthn

a 9

II T. kamakou 10,11

12,13,14,15,16

HfHW-T-pemr,™

. restricta H I T

FIG. 2. Phylogeny of the Hawaiian spiny-leg Tetragnaiha based on morphological characters. Explanations are given for characters defining each node; characters defining species are marked as bars. Sketches of the tip of the male conductor (left) and the upper surface of the apical portion of the male chelicera (right) are included for ease of comparison. Arrows point to the location of a species in the archipelago, with the size of an arrow tip being approximately proportional to the size of the distribution of a given species.Character changes defining each node are as follows. 1, Conductor terminal projection: short->long. 2, Conductor cap ridge: lateral- >medial. 3, Conductor cap: roundcd->pointed. 4, Conductor cap lip: blunt->pointed. 5, Colour: brown/variable->green. 6, First upper cheliceral tooth lost. 7, First lower cheliceral tooth lost. 8, First 2 lower margin cheliceral teeth-> well separated. 9, Conductor backward projection at leveI-> below cap. 10, Conductor cap: low->high. 11, Backward projection conductor spur->anglcd down. 12, Venter: pale->dark. 13, Abdomen colour: green->brown. 14, First dorsal cheliceral tooth~>finger. 15, Cheliceral 'ri tooth-> closer to T\ 16, First 2 lower cheliceral tccth-> closer. 17, Conductor backward projection hooked up-> angled down. 18. Lower tooth row Iong-> short.

range of habitat types. In particular, 7. brevig- the Hawaiian Drosophila. The Hawaiian natha is found in almost every habitat type on Drosophila generally demonstrate single volcano Hawaii Island, whereas representatives of the endemism, one species having its closest relatives species on East Maui are confined to a narrow on an adjacent volcano. In contrast, the

band of mesic forest at middle elevation. phylogeny I have generated for the spiny-leg There are some distinct differences between the species of Hawaiian Tetragnaiha suggests a non-

pattern of phylogeny I have generated here for the uniform and disjunct pattern. Possible explana- Hawaiian Tetragnaiha and patterns suggested for tions for the Tetragnaiha pattern may best be 524 MEMOIRS OF THE QUEENSLAND MUSEUM

ISLAND HAWAII MAUI MO LA OAHU KA M. Kca Kh Hu W Haleakala Ka La Wainaes Ko Wai VOLCANO Mauna Loa S w w E E Saddle E E N N E E w Elevation 1-2 0-1 1-2 0-1 1-2 0-1 1-2 0-1 1-2 1-2 1-2 1-2 0-1 1-2 0-1 1-2 1-2 1-2 1-2 0-1 1-2 0-1 1-2 tantalus X polychromata X X brevignatha X X X X X X X X X macracantha X X X waikamoi X X X X kauiensis X kamakou X x X perreirai X pilosa X quasimodo X X X X X X X X X X X X X X X X X X

TABLE 2. Tetragnatha species collected at different sites (islands, volcanoes and elevations, x 100m) through the Hawaiian Islands. Islands: MO= Molokai; LA= Lanai; KA= Kaui. Volcanoes: M. Kea= Mauna Kea;Kh= Kohala; Hu = Hualalai; W= W. Maui; Ka= Kamakou; La= Lanaihale; Ko= Koolaus; Wai = Waialeale. considered by viewing the Hawaiian archipelago tion (Kaneshiro, 1976, 1983; Carson etaL, 1989). as a series of evolutionary snapshots, with specia- Indeed, it is possible that both T. macracantha tion starting on Hawaii Island and developing on and T. brevignatha arose through hybridization, East Maui. The three species on the recently which may play an important element in the likely to relatively formed Hawaii Island are be formation of species in general (Endler, 1989). recent colonists that have expanded their range As sexual discrimination and ecological adap- and habitat use. Such ecological release sub- tation develop, invaders would presumably lose sequent to colonization is considered an impor- their ability to colonize an occupied land mass. tant step in initiating species divergence in The pattern of distribution of representatives of Galapagos finches (Grant, 1986). However, the spiny-leg clade on older islands suggests that widespread species on Hawaii Island are the remarkably homogeneous, and none are endemic closely related taxa cannot maintain coexistence to the island. It may be that Hawaii Island is too on the same land mass unless they have under- young for speciation to have occurred in the gone sufficient ecological divergence. The situa- spiny-leg Hawaiian Tetragnatha. The situation tion on East Maui may therefore represent an suggests that considerable movement of in- unstable state: ultimately, a single species will dividuals occurs within the island, and gene flow take over the land mass, as a result of introgres- between islands has been sufficient to prevent sion or competitive displacement. speciation during the period of existence of The mechanism I have proposed for speciation Hawaii Island. among representatives of the spiny-leg Hawaiian The adjacent older volcano of East Maui was Tetragnatha remains speculative. However, the once part of the island complex, 'Maui Nui' repeated ecological release of newly forming taxa (comprising Molokai, Lanai, East and West strongly suggests that ecological changes have Maui). This island was likely first invaded by T played some role in initiating species divergence, tantalus. Males may be better colonists than as does the finding that two populations (Maui females (Bishop, 1990), but spiderlings would versus Hawaii) of an apparently diverging also arrive, and eventually give rise to a popula- species (T. brevignatha) differ only in terms of tion that would expand its range on that island. their habitat occupation. I suggest that, unlike the However, colonists would continue to arrive on Hawaiian Drosophila in which sexual selection Maui Nui, and, at least initially, the original has been heavily implicated in the speciation colonists would not be reproductively isolated from the secondary colonists of the same species. process (Kaneshiro, 1983; Kaneshiro and Gid- It is also possible that, if the secondary colonists dings, 1987), ecological factors (range expan- included closely related heterospecifics, sion, reinvasion, competition) may be more hybridization might occur, as newly forming taxa important among the spiny-leg species of tend to have poorly developed sexual discrimina- Hawaiian Tetragnatha. .

PATTERN OF PHYLOGENY OF HAWAIIAN SPIDERS 525

ACKNOWLEDGEMENTS proach to character weighting. Systematic Zool- ogy 18: 374-385. 1970. Methods for computing Wagner trees. Sys- The study was supported by grants from the tematic Zoology 19: 21-34. Hawaii Bishop Research Institute and the Hawaii 1989. The retention index and the rescaled consis- Natural Area Reserves System with additional tency index. Cladistics 5: 417-419. support from the Smithsonian Institution, the FITCH, W.M. 1971. Toward defining the course of Bishop Museum, Haleakala and Hawaii Vol- evolution: minimal change for a specific tree canoes National Parks, Maui Land and Pineapple topology. Systematic Zoology 20: 406-416. Company, the Nature Conservancy of Hawaii, FREED, LA., CONANT, S. & FLEISCHER, R.C. Pacific Tropical Botanical Gardens, the U.S. Fish 1987. Evolutionary ecology and radiation of and Wildlife Service, the University of Maryland Hawaiian passerine birds. Trends in Evolutionary (Dept. of Entomology), and the University of Biology 2: 196-203. GILLESPIE, R.G. 1986. 'Between population com- Hawaii (Dept. of Zoology). I am indebted to the parison of resource acquisition in die long jawed following for their assistance in collecting orb weaving spider Tetragnatha elongata\ Ph.D specimens: R. Bartlett, J. Burgett, H. L. Carson, Dissertation. (University of Tennessee: Knox- I. Felger, J. I. Gillespie, J. Halloran, J. Jacobi, K. ville: Tennessee). Y. Kaneshiro, J. Kiyabu, L. Loope, D. Lorence, 1 987. The mechanism of habitat selection in the long A. C. Medeiros, S. Montgomery, C. Parrish, S. jawed orb weaving spider Tetragnatha elongata Perlman, W. Perreira, D. Preston, V. and B. Roth, (Araneae, Tetragnathidae). Journal of Arachnol- R. Rydell, W. Stormont, J. Strazanac, M. White ogy 15: 81-90. and K. Wood. I am grateful to J. A. Coddington, 1 991 Hawaiian spiders of the genus Tetragnatha: I. Spiny Leg Clade. Journal of Arachnology 19: H. B. Croom, H. W. Levi, S. R. Palumbi, N. I. 174-209. Platnick and to G. K. Roderick for providing 1 992. Hawaiian spiders of the genus Tetragnatha II. advice, discussion and comments. Species from natural areas of windward East Maui. Journal of Arachnology 20: 1-17. LITERATURE CITED GRANT, P.R. 1986. 'Ecology and Evolution of Darwin's Finches'. (Princeton University Press: BERGER, A.J. 1981. 'Hawaiian Birdlife'. (University Princeton, New Jersey). of Hawaii Press: Honolulu). HENNIG, W. 1966. 'Phylogcnetic Systcmatics'. BISHOP, L. 1990. Meteorological aspects of spider (University of Illinois Press: Urbana, Illinois). ballooning. Environmental Entomology 19: KANESHIRO, K.Y. 1976. Etiological isolation and 1381-1387. phylogeny in the planitibia subgroup of Hawaiian

CARSON, H.L. 1968. The population flush and its Drosophila. Evolution 30: 740-745. genetic consequences. Pp. 123-137. In Lewontin, KANESHIRO, K.Y. 1983. Sexual selection and direc- R.C. (ed). 'Population Biology and Evolution'. tion of evolution in the biosystematics of the (Syracuse University Press: Syracuse, New Hawaiian Drosophilidae. Annual Review of En- York). tomology 28: 161-178.

CARSON, H.L. & KANESHIRO, K.Y. 1976. 1 988. Speciation in the Hawaiian Drosophila, Bios- Drosophila of Hawaii: Systcmatics and ecological cience 38: 258-263. genetics. Annual Review of Ecology and Sys- KANESHIRO, K.Y. & BOAKE, C.R.B. 1987. Sexual tematics. 7:311-346. selection and speciation: issues raised by CARSON, H.L. & TEMPLETON, A.R. 1984. Genetic Hawaiian drosophilids. Trends in Ecology and revolutions in relation to speciation phenomena: Evolution 2: 207-211. The founding of new populations. Annual Review KANESHIRO, K.Y. & GIDDINGS, L.V. 1987. The of Ecology and Systcmatics 15: 97-131. significance of asymmetrical sexual isolation and CARSON, H.L, KANESHIRO, K.Y. & VAL, F.C. the formation of new species. Evolutionary Biol- 1989. Natural hybridization between the sym- ogy 21: 29-43. patric Hawaiian species Drosophila silvestris and KARSCH, F. 1880. Sitzungs-Berichte derGesellschaft Drosophila heteroneura. Evolution 43: 190-203. Naturforschender freunde zu Berlin. Jahrgang. COOKE, C, MONTAGUE, J. & KONDO, Y. 1960. Sitzungvom 18:76-84. Revision of Tornatellinidae and Achatinellidae LACK, D. 1971. 'Ecological Isolation in Birds'. (Har- (Gastropoda, Pulmonata). B.P. Bishop Museum vard University Press: Cambridge, Mas- Bulletin 221: 1-303. sachusetts). ENDLER, J.A. 1989. Conceptual and other problems in LEVI, H.W. 1981. The American orb-weaver genus speciation. Pp. 625-648. In Otte, D. and Endler, Dolichognatha and Tetragnatha north of Mexico J.A. (eds). 'Speciation and its Consequences'. (Araneae: Araneidae, Tetragnathinae). Bulletin of (Sinauer Associates: Sunderland, Massachusetts). the Museum of Comparative Zoology Harvard FARRIS, J.S. 1969. A successive approximations ap- 149:271-318. 526 MEMOIRS OF THE QUEENSLAND MUSEUM

MADDISON, W.P., DONAGHUE, M.J. & MAD- OTTE, D. 1989. Speciation in Hawaiian crickets. Pp. DISON, D.R. 1984. Outgroup analysis and par- 482-586. In Otte, D. and Endler, J.A. (eds). simony. Systematic Zoology 33: 83-103. 'Speciation and its Consequences*. (Sinauer As- MAYR, E. 1963. 'Animal Species and Evolution'. sociates: Sunderland, Massachusetts). (Harvard University Press: Cambridge, Mas- SIMON, C. 1987. Hawaiian evolutionary biology: an sachusetts). introduction. Trends in Ecology and Evolution 2: MCCUNE, A.R. 1990. Evolutionary novelty and 175-178. atavism in the Semionotus complex: relaxed selec- SIMON, E. 1900. Arachnida. Fauna Hawaiiensis 2: tion during colonization of an expanding lake. 443-519, pis. 15-19. Evolution 44: 71-85. SWOFFORD, D.L. 1990. PAUP 3.0s. Phylogenetic OKUMA, C. 1988. Redescriptions of the Hawaiian analysis using parsimony, version 3.0. (Illinois spiders of Tetragnatha described by Simon Natural History Survey: Champaign, Illinois). (Araneae, Tetragnathidae). Journal of the Faculty WIEHLE, H. 1963. Tetragnathidae. Tierwelt of Agriculture Kyushu University 33: 77-86. Deutschlands 49: 1-76. fEW SPECIES OF KARSCHIIDAE (SOLIFUGAE. ARACHNIDAi FROM KAZAKHSTAN

ALEXANDER V. GROMOV

Gromov, A.V. 1993 ll 11: A new species of Karschiidae (Solifugae. Aratlwudu) hum Kazakhstan. Memoirs of the Queensland Museum 33(2): 527-528. Brisbane, ISSN 0079- 8835.

The new solpugid species Karschia mangistauensis of the family Karschiidae is described from material collected in south-western Kazakhstan. The relationship of the new species is given, as well as lhe list of solpugids currently known from Kazakhstan- En se basant surle materiel recohe au Sud-Ouest du Kazakhstan on decrit une nou vel le espece de solifuge Karschia mangistauensis de la famille Karschiidae. On donne les afftnitcs de la nou velle espece aussi que la listc des solifuges connucs a present au Kazakhstan. {jSoipugids, Karschiidae, Karschia, Kazakhstan.

Alexander V. Gromov, Institute of Zoology Kazakhstan Academy of Sciences, AkiXdem- gorodok, Alma-Ata 32. 480032 Kazakhstan Republic, CIS; 8 March 199 i

The solpugid fauna of Kazakhstan is poorly Kugusem Well, 68.5 km E of Akkuduk Village. known. Only 13 species were recorded from the Yeraliev District, (43°10'N, 54°S3*E)1 2-5 May region by Biruia (1938): these are Karschia 1990, S.I Ibraev, 1 S\ Sulykkyzylsai Well, 69.5 :arudnyi Biruia, Eusintonia turkestana Kraeplin, km NE of Akkuduk Village, Mangistau District. D , D Anoplogylippus dswigaricus Roewer, A. rick- (43 28 N,54 43-E). 12 May 1991, E.E. Kopdvk- mersi (Kraeplin), Hemigylippus latnelliger baev. Holotype and 1 9 paratype preserved in Biruia. Galeodes araneoides Pallas, G. lurkes- Zoological Institute, St Petersburg [Leningrad], tanas Kraepclin, G. caspius Biruia. G. Justus the remaining material in the author's collection. Biruia, G. palia Si Birula.Paragaleodes heliophilus Heymons, F_ pallidas Biruia and Diagnosis Daesia rossica Biruia. The new species is closely related to Karschia The present paper concentrates on material col- cotnifera Walter, 1 889 from Turkmenistan, from lected from Mangyshlak and Ustyurt Plateaus which it differs by the colouration, by the shape (south-western Kazakhstan). Solpugids were of upper modified mcsolateral seta near the h.ise preserved and studied in 70% alcohol using a of chcliccral fingers, by the number of teeth on binocular microscope MBS-1. Their determina- the fixed finger and by the spinulation of tion was done according to Walter ( 1 889). Biruia pedipalps. (1938) and Roewer (J 932- 1934, 1941 ).

MORPHOLOGY AND BIOLOGY Description Family KARSCHIIDAE Male (holotype). Total length 21 mm. Body colouration light yellow-brown, with greyish head and thorax and grey-yellow abdomen with Karschia mangistauensis sp. nov. darker tergites. Chelicerae yellow, distally in- (Figs 1-10; table I) cluding brownish-black teeth. Pedipalps (Fig 4) proximal part of femur yellow, distal part of Material Examinhd femur and the remaining segments greyish. LdgS

TYPES. Holotype 6 , Zhylandy Cape. Yeraliev yellow. Ocular tubercle (Fig. 1) dark, with District. Mangyshlak Plateau, Mangistau Area. numerous hairs, sparse short setae and 2 long , South- Western Kazakhstan, (43WN, 51°39 E), setae. Near the base of the eheliceral fingers there 2 May 1991, A.V. Gromov. Paratypes: South- is a mesal row of long setae: the upper two are western Kazakhstan: Mangistau Area: Man- strongly modified and thickened (Figs 2, 3). Ar- gyshlak Plateau: 2 6 , 2 9, same data, except 2-4 mature of pedipalps (Fig- 5): protarsus (basitar- May 1991, K.U. Balmukanov, A.V. Gromov, sus, metatarsus auct.) with 9 promesolateraJ

K.B. Dzhankurazov; 1 <$, Aktau City [Shev- spines, tarsus with 1 mesobasal spine. Third ab- chenko], (43°H 7 N, 51°39'E), 28 April 1991, dominal segment with 46 broad ctenidia, fourth

KU. Balmukanov; Ustyurt Plateau: 1 6, one with 19 ctenidia (Fig. 6). 528 MEMOIRS OF THE QUEENSLAND MUSEUM

Length

Chelicerae 4.4. 1.6 wide

Propellidium 2.4, 3.5 wide

Pedipalp: total (with coxa) • s._

tibia 4.9

basilarsus 3,8

Lep I twtlh coxa) >A 1

Lef! II (With coxa) :-!

Leg 111 (with coxa) I6.<

Lep IV (with loxu) 25 -'

TABLE 1. Measurements (in mm) of Karschia mangistauensis, sp. nov.

Biology

Night solpugtd in clayey desert under stones during day.

ACKNOWLEDGEMENTS

1 am grateful to Messrs K.U. Balmukanov and KB, Dzhankurazov (Kazakh State University, Alma-Ata) for their assistance in capturing of the new species, Mr S.I. Ibraev and Mr E.E. Kopdyk- FIGS 1-10. Karschia mangistauensis, sp. nov. 1-6, bacv (Institute of Zoology, Kazakhstan Academy male; 7-10, female. 1, propeltidium, dorsal view. 2, of Sciences, Alma-Ata) for their material kindly right chelicera, ectal view. 3, modified setae near presented to me, arid to Dr A.A. Zyuzin (Institute base of fingers, and flagcllum on leftchcliceral fixed of Zoology, Kazakhstan Academy of Sciences. finger, mesal view. 4, 5 right pedipalp, mesoventral Alma-Ata) for translation of manuscript. I am view; coloration (4); spinulation (5). 6. ctenidia on also indebted to Dr Ch. K. Tarabaev (Institute of fourth sternite of abdomen, ventral view. 7, right Zoology of Kazakhstan Academy of Sciences, chelicera, ectal view, 8, right cheliceral fixed finger, Alma-Ata) who presented my paper at the XII ventral view. 9, genital opcrcula, ventral view. JO, ctenidia on fourth sternite of abdomen, ventral view. International Congress of Arachnology (Bris-

Scale line = I mm. bane, Australia).

Body length of paratypes 17-21 mm, number of LITERATURE CITED promesolatcra! spines 6-10, mesobasal spine of tarsus sometimes absent, fourth abdominal segment with 17 or 19 ctenidia. BIRULA, A.A. [BIALYMTSKU-BIRULA], 1938. Solifugae. In; Fauna of the USSR, Arachnida, vol, Female paratypc. Body colouration lighter than 1(3). (Academy Of Sciences of the USSR; Mos- in male, with darker pedipalps. Distal part of cow-Leningrad, (in Russian) femur IV light brown. The head behind ocular ROEWER, CE 1932-1934 Solifugae, Palpigradi. In: tubercle with slight longitudinal light brown line. Dr H. G. Bronn's Klassen und Ordnungcn des Ocular tubercle as in male, occupying less than Tierreichs, Teil 5, Abteilung 4, Buch 4, 723 pp. 1/3 of clypeus width. Fixed fingers arc straight Bremen. from above, their length no more than the width 1941. Solifugen 1934-1940. Veroffentlichungcn bus dem Deutschen (Colonial- und Ubcrscc- of chelicerae- Armature of chelicerae as in Figs Museum in Bremen 3(2): 97-192. 7, 8. Genitalia (Fig. 9) with pale rosy ectolateral WALTER, A, 1889. Transkaspische Galeodiden. In: setae. Fourth abdominal segment with 19 pale Wisscnschaftlichc Ergcbnisse der im Jahre 1886 rosy ctenidia which are thicker than in male (Fig. in Transkaspien 1 (Zoologische Abthcilung). 10)' Lieferung6: 103-117. -

A NEW SPECIES OF AMAUROBIOIDES O.P. -CAMBRIDGE (ANYPHAENIDAE: ARANEAE) FROM SOUTH AUSTRALIA

D.B. HIRST

Hirst, D.B. 1993 11 11: A new species of Amaurobioides O.V. -Cambridge (Anyphaenidae: Araneae) from South Australia. Memoirs of the Queensland Museum 33(2): 529-532. Brisbane. ISSN 0079-8835.

The littoral spider Amaurobioides isolatus sp. nov. is described from South Australia and is the first new species of the genus from the Australian mainland. The male palp ofA. litoralis Hickman is re-illustrated. Biogeography is discussed. \Z\Araneae, Anyphaenidae, Amaurobioides, new species.

David B. Hirst, South Australian Museum, North Terrace, Adelaide, South Australia 5000, Australia; 6 November, 1992.

The genus Amaurobioides O . P i c ka rd series of 1 6 and 3 9 9 from Tasmania) deposited Cambridge, 1883 occurs in New Zealand in AM, were examined. Hickman made no special (Cambridge, 1883), Campbell Island (Hogg, type designations but the male seen (AM KS6410) 1909), South Africa (Hewitt, 1917) and was is labelled 'Holotype\ Hence, the label designation reported from Tasmania (Hickman, 1949) and is invalid (Art. 73iii). The series also includes two Chile (Forster, 1970). Forster (1975) included the females of an undescribed species (Forster, 1970). south-east coast of Australia in the distribution of Forster, when noting that this type material com- the genus but that was not cited by Main (1981) or Davies (1986). The specimen(s) to which prised two sympatic species, stated that the larger Forster (1975) referred were deposited in the form was A. litoralis while the smaller form was Australian Museum, Sydney (AM) but were not undescribed. labelled as Amaurobioides (unpublished data) and have not been found (pers. comm., M. Gray). CO No further material has been reported from main- .2 1.8 L X land Australia. Once Amaurobioides was found f0 • X on the rocky shoreline of the eastern side of Gulf

St Vincent in South Australia , deliberate searching in similar habitats on Eyre Peninsula, Fleurieu 1.7 Peninsula and Kangaroo Island showed the species was widespread. X X Types of A. litoralis Hickman, 1949 (syntype C

1.6 r X C E \ -m c -

0) *-> '_ X CO 1-5 r

- 3 It

1.4 \

Species

FIG. 1. Southern portion of South Australia showing FIG. 2. Comparison of sternum length: width ratios: a, the distribution of Amaurobioides isolatus. Shaded b, A. litoralis: small form (a); large form (b); c, A. coastline may include suitable habitats. isolatus. m, mean. MEMOIRS OF THE QUEENSLAND MUSEUM

Amaurobioides isolatus sp. nov.

(Figs I, 2, 4-9, Table!)

Material Examined , TYPES. Holotypc 6*, Blanche Point, 35°I5 S, 1&°2S*E, l.iii.1986, DH, N1992206. Paratypcs; allotype 9, same data as above, N1992207; V (wilh , spideriings), Elliston, 33*39' S. 134°53 B, 31.iii,1987. DH, D. Lee, N1992218; 9, juv., same data, N1992216-7, 9, 5 penult 6\ 2 penult. ?. same data

as holotypc but 25x1986. NI992208-15: penult. ? ? Petrel Cove, 35*^5, I38°36E, l.ii.1991. DH, D N1992219, 9, 4juv., Point Avoid, 34 4 PS, 135'19'E, 31.iii.1987, DH,"D. Lee, N 1992221; 9 (with spiderlings). juv.. same data, N1992220; 2 9, juv.. Point Hllen/K.l., 36WS. 137°11'E, Nkxi.1987, DH, N 1992222-3; 9 (with eggsac). Point Tinlinc, K.I.,

35°59'S, 137°37'E, 1 Lxi.1987, DH, N1992224; 6"?, Port Willunga, 35°16S, 138°28'E. in retreats ,5m above base of large rock on beach. 14, vi. 1992, DH, N1992234-5; 4 9, juv., West Bay, K.I., 35°54'S. I36M2E, 6,xi.l987, D, N1992225-9. All in South Australia and deposited in South Australian Museum.

Diagnosis The S of A. isolatus is recognised by the shape of the primary conductor. A. isolatus further differs from the two Tasmanian species in having FIG. 3. A. litoralis. Tibial apophysis and cymbium of syntype & AM KS6410\ ventral the sternum relatively broader while the 2 epigynum is smaller and less sclerotised. The 6

The <$ is similar in size to the sole smaller 9 of has spination of tibia 1 identical to the 9 , a feature the type series, and not to the two larger 2 9 as shared with A. maririmus O.P.-Cambridge, from suggested by Forster (also captions for figures of which it is separated by the genitalia. the two forms appear to be transposed as the larger specimen figured is called The small form). Description This is supported by comparisons of the sternum MALE. CL 3.79, CW 2.56. AL 4.48, AW 2.41. length:width ratio (Fig. 2). Ratios: <$ , 1.65 and Colour in alcohol: Carapace yellow-brown, smaller type 9 1.67; larger 9 9, 1.77, 1.81, 9 yellow posterior to fovea and laterally to above A. isolates 1.48-1.60 (n=14. mean =1.56) and 2 anterior legs, caput dark orange-brown, clypeus 1.51, 1,60, Although based only and lateral edge of face brown- Chclieerae red- Australian material seen, theconspccific 6* of the brown, darker distally. Maxillae and labium larger and undescribed form would probably at- orange-brown, anterior margins cream. Sternum tain a ratio in excess c\ 1 .70. cream, margins orange Legs cream-yellow, The 6 palp OfA. litoralis is re-illustrated here; anterior metatarsi and tarsi darker, coxae cream. the ?epigvnum was adequately illustrated in Palp cream, cymbium brown. Abdomen cream Forster (1970, tigs 487, 488). with dark red-brown pattern dorsally and laterally, typical for genus. Venter darker around METHODS AND MATERIALS spiracle and posteriorly to spinnerets, dark area extending anteriorly from spiracle in two narrow Hairs are omitted from illustrations. 9 genitalia lines almost to epigastric furrow. were dissected then cleared in lactic acid. All Carapace elongate, longer than broad in ratio

in > longitudinal fovea measurements arc millimetre.?. \bbreviu(tons ! 19:1 3 short imperfect CL, carapace length; CW, carapace width; AL; (presumably straight in normal specimens]. abdomen length; AW, abdomen width, MOQ. Striae barely evident adpressed setae black, median ocular quadrangle; aw, anrnmr Width; upright setae on caput between fovea and ocular

p\v. posterior width; L or I , length, W, width; K.I.. region. Eye group occupies less than half width Kangaroo Island; DH. D. Hirst. ,iut.AME0.10,ALE0.17.PME = PLE0.18. .

SOUTH AUSTRALIAN AMAUROBIOIDES 531

Legl Leg 2 Leg 3 Leg 4 Palp

Fe 3.32(2.67) 3.30(2.61) 2.71(2.46) 2.98(2.51) 1.50(1.40)

Pa 1.53(1.59) 1.56(1.48) 1.33(1.39) 1.47(1.50) 0.59(0.65)

Ti 3.40(2.50) 3.23(2.43) 2.24(1.82) 2.74(2.32) 0.65(0.66) Me 3.19(2.34) 2.99(2.25) 2.40(1.96) 2.59(2.15) -(-) ^J\ Ta 1.67(1.42) 1.42(1.29) 1.02(0.95) 0.97(0.95) 1.26(1.24) Total 13.11(10.52) 12.50(10.06) 9.70(8.58) 10.75(9.43) 4.00(3.95) FIGS 8-9. A, isolatus. 8-9, cleared vulva of paratype 9 N 1992208, 8, ventral, 9, dorsal. TABLE 1. Leg lengths of Amaurobioides isolatus. Values are for holotype S (allotype 9). dorsal median secondary prong (Fig. 5), right apophysis lacking accessory prong (Fig. pos- Interspaces: AME-AME 0.05, AME-ALE 0.04, 6), sibly normal state. PME-PME 0.11, PME-PLE 0.12, AME-PME FEMALE (as 6* except as follows). CL 4.21, 0.14, ALE-PLE 0.08. MOQ; aw: pw: 1 = 0.26: 0.47: 0.39. Width of clypeus to AME 0.10. CW 2.89. AL 6.97, AW 4.16. Chelicerae with three teeth on both margins. Ster- Eyes. AME 0.11, ALE 0.20, PME 0.19, PLE 0.20. Interspaces; 0.08, num: L 2.12, W 1.40. Legs. (Table 1) Scopula AME -AME AME-ALE sparse, particularly on posterior pairs. Leg spina- 0.06, PME-PME 0.12, PME-PLE 0.14, AME- = tion; no spines on patellae or tarsi. Ventral tibiae PME 0.17, ALE-PLE 0.10. MOQ; aw: pw: 1 and metatarsi spines usually paired, occasionally 0.29: 0.50: 0.46. Width of clypeus to AME 0.06. Sternum: L 2.30, W 1.48. Legs. (Table 1) single. Leg I, fe d3 pi , ti v6, me v3. Leg II, fe d3 pi, ti pi v6, me v4. Leg III, fe d3 pi rl, ti pi rl Epigynum. Fossa broad, anterior margins v5 (v6 on right), mep3r3 v6. Leg IV, fed3pl rl, vaguely defined by what appears to be sub- ti rl v3, me p2 r2 v5. Palp, fe d3 pi, pa with one cuticular sclerotisation of vulva (Fig. 7). Vulva of stout bristle dorsally at distal end, ti many long paratype N1992208 shown in Figs 8, 9 [allotype stout bristles, cymbium with three short weak not dissected]. spines prolaterally. Palp. Tibia with left apophysis having short Variation

CL of paratype 6 , 3.48; the tibial apophysis of both palps lack accessory prongs and the fovea is

straight. CL of paratype $ ; 3.22-4.74, mean 4.35 (n=13).

Etymology The specific epithet, isolatus, reflects the isolated nature of the species distribution.

NATURAL HISTORY

Habits are similar in other Amaurobioides. Silk retreats are assumed to be permanent although many spiders collected or observed at Blanche Point in February were wandering over the rocks at night. Several penultimate 6 6 9 $ were collected in this way and kept alive for some time, one 6 eventually maturing 3 months later. Most insects placed with the spiders, including moths and terrestrial isopods, were not fed upon. Small flies (mostly Drosophila) were more readily accepted (Forster, 1970: 167). Littoral isopods were not tried as a food source (Hickman, 1949). Since then a specimen of A. isolatus at Blanche Point lunged at and grasped a small littoral isopod half FIGS 4-7. A. isolatus. 4-5, Left tibial apophysis and its size with its anterior legs, hesitated and then of ventral, retrolateral; cymbium holotype 6 , 4, 5, 6, released the isopod, retreating to its position at the right tibia and apophysis; 7, epigynum of allotype 9 entrance to the nest. Small flies, often seen resting 532 MEMOIRS OF THE. QUEENSLAND MUSEUM

in sheltered areas ai night along the coast, are Australia presumably because continual changes another likely food source. Females with spiderl- to the coast have occurred gradually. Littoral ings (26 and 42) in the retreat were found in spiders have moved with it through dispersal and mid-autumn at both Point Avoid and Ellision colonising adjacent new areas of rocky terrain or while spiderlings and an eggsac were found in re-establishing itself into old areas within a separate retreats in November on Kangaroo Is- suitable distance from the receding or insurgent land. sea. The sympatry of species of Amaurobioides in New Zeahmd and Tasmania probably occurred BIOGEOGRAPHY by migration over land following uplift or weathering of unsuitable coastline between two Most areas sampled were within the splash separated species and exposing a continuous zone on rock faces sheltered from the full velocity rocky habitat, rather than by dispersal on silk of the sea, but retreats were seen in an exposed lines. and treacherous area at Cape du Couedic. Kan- garoo Island. Known populations are separated ACKNOWLEDGEMENTS by unsuitable or sandy coastline of varying lengths (Fig. 3). Amaurobioides may have dis- I am grateful to Dr M Gray (AM) lor ihc ban persed across the intervening ocean by parachut- of types of the Ta&manion Amaurobioides and to ing on silk lines or drifting on flotsam to settle Mr L N Wcolscn to commenting on an early and inhabit their present littoral environments. draft of the manuscript. Forster (1970) considered that claim to be over- stated as a number of distinct forms some sym- LITERATURE CITED patic, were present in New Zealand. Platnick (1976) reinforced that in the Laroniinae. - On some new genera and Since Australia's separation and subsequent CAMBRIDGE. OP 1883, species of spider*. Proceedings of ihe Zoological drift from the other continents of Gondwanaland. Society of London 1883:352-365. die South Australian coastline has been altered by DAILY, B., MILNES, A.R, TWIDALE, C.R., changes in sea level during the last ice-age The BOURNE, J.A. 1979. Geology and geomorphol- cliffs rocky of Blanche Point and southern coast ogy. In "Natural History of Kangaroo Island*. of Kangaroo Island which now provide a habitat (Royal Society of South Australia: Adelaide). for Amaurohi(fides were during the last formed DAVIES, V T 1 9XfV 'Australian spider,, Arancae. Col- ice-age over Tertiary deposits and have since lection, preservation and identification . heen uplifted and weathered (Daily etui, 1979). (Queensland Museum Booklet No. 14. Brisbane). One population in D'Estrees Bay, Kangaroo Is- FORSTER. R.R. 1970, The spiders of New Zealand land, consisted of only a few individuals on a Part HI. Otago Museum Bulletin No. 3: 1-1 84. small isolated rock outcrop backed by a low 1975. The spiders and harvestmen. In Kuschel, G. (cd.) 'Monographic biologicac: biogeography sandstone ledge. Similarly, near Port Willunga and ecology of New Zealand*. (W. Junk: Nether and 2km south of Blanche Point, individuals exisl lands). on a few older limestone rocks remaining on the HEWITT, J. 1917. Descriptions of new South African wide sandy beach which are reached by the nor- Aruchnida. Annals ol Ihc Natal Museum 3: 687- mal high tide. Although backed by cliffs, no 711. spiders were found on these. Cliffs also abut the HICKMAN, V.V. 1949. Ta-.i.niniMMiitin.ai spiders with sea in areas north of Blanche Point to Marino llOtts on their respiratory systems, habits and Rocks but as yd AmAur&biohfes has not been taxonomy. Papers and proceedings of the Royal found. An 'ides is unlikely to be fi Si--ieiy of Tasmania 194& $1-43. from Victor Harbouron the llcuneu Peninsula to HOGG, H.R 1909. Spider, and Opilionf.s from the at least Robe in the south-cast of South Australia. Subantareric Islands of New Zealand. The Su luetic Islands of New Zealand I, Wellington Extensive sand dUlfeS winch now form ihc 1909, [55-181. Coorong and much of the coast to the south are MAIN. BY. 1981. Australian spiders: diversity, dis- pan of that atea built up during inundation by the tribution and ecology, in KeaSl. A. (cd.) 'Ecologi- sea in the Miocene and the Pleistocene, and the cal Biogeography of Australia* Vol. 2. (W. Junk, few rock) outcrops noty present have remained Netherlands). isolated since. PLATNICK, N.I. 1976. Drifting spiders or continents?: In summary, Amaurobioides is one of the few wcariancc biogeography of die spider subfamily true Gondwanan spider genera left in South Lamniinar, Systematic Zoology 25; 101-109. IMPLICATIONS OF THE PH YLOGENY OF PIMOIDAL FOR THE SYSTEMATICS OF LINYPHIID SPIDERS

( ARANEAE, ARANEOIDEA, LINYPHIIDAE i GUSTAVO HORMIGA

Hormiga, G. 1 993 1 J 1 1 : Implications of the phylogcny of Pirooidae for the systcmatics of linyphiid spiders ( Araneae, Araneoidea, Linyphiidae). Memoirs ofthe Queensland Museum 33(2): 533-542 Brisbane. ISSN 0079-8835.

The araneoid family Pimoidae (new rank) is hypothesized 10 be the States group of Linyphiidae. Lonisfagea Brignoli is a junior synonym of Pimoa Chambcrlin and [vie (new characters synonymy). The thai support the monophyly of Pimoidae and of Li nypluidac pi I \\ Pimoidae are discussed. Explicit outgroup comparison to the closest relative; ol linyphiids

li r e. pimoids) allows studies of character evolution and character polarisation within linyphiids and the assessment of previous phylogenetic hypotheses for the family- I'lelimi nary data on the implications ol pimoid phytogeny for linyphiid systematic* are evaluated, based mainly on morphological characters. Linyphiid monophyly is discussed. La lamilia arancoidc Pimoidae (nucvo rango) es, hipoteHicamcntc, el grupo hermano de Linyphiidae. El gencro Louisfagea Brignoli se considera sinonimo dc Pimoa Chamberlm and Jvie (nueva sinonimia). Sc discuten los caracteres que apoyan la monofilia de Pimoidae

y de i inyphiidae ma's Pimoidae. La utilization explfcita del criteria de comparacion con el grupo externo de los linifidos (cs deeir, los pimoidos) pcrmite estudiar \s evolution y polari/aeirtn de caracteres en linifidos. asi corno la evaluacion de antcriorcs hipoiesis filogencticas sobre esta familia. Se evaluan los datos preliminares, basados en caracteres

morfologicos prineipalmente, sobre las implications de la filogenia de los pirnoidos para la sistematicade los linifidos. Tambien se diseute la monofilia de Linyphiidae. [jPimoidae, lunphtidae, Pimoa, cladislks. phylogcny. monophyly, homology.

Gustavo Hormiga, Department of Entomology, NHB 105, National Museum of Natural History. Smithsonian Institution, Washington DC 20560> V.SA and Maryland Centerfor Systematic Entomology. Department ofEntomology University Mankind, College Park. { of MD 20742. U.S.A.; 27 October. 1992.

Linyphiids are one of the dominant spider systematics and character evolution, by explicitly groupsiniheHoIarcticregion.Despitetheirover- stating phylogenetic relationships in terms of whelming diversity and involved taxonomic his- synapomorphies rather than by the more specula- tor}* the phylogenetic structure of the family and tive approaches that have commonly been used

their relationship to other araneoids are very irt traditional linyphiid higher systematics. This

poorly understood. In this paper I present some approach enables us to evaluate comparative preliminary data on the systematics of pimoids, morphological data (or any other kind of biologi- confirming their sister-group relationship to cal data) in a eiadistic context. Hypotheses op Imyphuds, and on the cladislic structure of a phylogeny and character homology hypotheses small sample of linyphiid genera. A revision and are indistinguishable because every hypothesis numerical cladislic analysis of the pimoids and of homology is a hypothesis of monophyletic 1 the sample of linyphiid taxa (Horrniga, in press), grouping (Patterson, 1982). Finally, the present together with detailed character information, will study allows for a preliminary test of the be'published elsewhere shortly. The study of Lhe phylogeny of the linyphiid subfamilies proposed phylogeny of the pimoids requires the inclusion DY Wundcrlich (1986). of at least a sample of linyphiids (their putative sister-group) in order to assess character state MATERIALS AND METHODS polarities by means of outgroup comparison. It is in such a context that the present study should be Taxa considered, since the small sample of genera used Nine linyphiid, five pimoid, and two norv here can by no means account for the whole range linyphiid araneoid genera that are possible out- of linyphiid diversity. However, quantitative groups of the pimoid-linyphiid complex arc Cladislic analysis of the data presented here in this study. The linyphiid taxa selected rcpre- enables for some testable hypotheses on linyphiid sent the subfamilies and tribes used by Wundcr- $34 MEMOIRS OF THE QUEENSLAND MUSEUM

lieli ii986> in his phytogeneifc scheme for press). The removal of crispa would leave the Linyphiidae (given here in parentheses): remaining species of Louisfagea as a paraphyletic Linyphta irta?\gularis (Clerck) and Micro- genusx Louisfagea is regarded here as a junior linyphia dona (Chamberlin and Ivie)(Liny- synonym of Pbnoa Chamberlin and Ivie (NEW phiinae, Linyphiini); Bolyphantes luteolus SYNONYMY). Throughout this paper the taxon I'Blackwall) and Lepthypkantes tenuis (Black- name Linyphiidae (linyphiids) does not include wall)Walckenaeria directa (O. P. -Cambridge) (Erigoninae ); Haphms dilons Characters T (Urquhart) and S ovafroneto vulgaris Blest The data set contains 47 characlers (Table 1 ): (Mynogleninae); and Siemonyphantrs blauvettae 33 male and female genitalic characters, 5 spin- Gerlsch (Stemonyphantinae), The punoids neret spigot characters, 7 other morphological (which contain 21 species, including 11 new somatic characters, and 2 behavioral characters species (Hormiga, in press) are represented here The data consist mostly of original observations, by five species: Pimoa (=Louisfagea) rupicota but a few characters have been extracted from the (Simon), P. (=Lottisfeigea) crispa (Page), P at~ literature. Although this data set integrates infor-

Maculosa (Keysetlmg), P. breviata Chamberlin mation from several character systems, it espe- and ivie, and P. curvaia Chamberlin and Ivic. cially focuses on male palp and spinneret spigot Teiragtwtha versicolor Walckenaer and Zygiella morphology. The methods of study and of homol- \-uotatu (Cfcrckl 1 as outgroups of the ogy assessment of spinneret spigot morphology pimoid-linyphiid cladc. The affinities of2yg/£//a follow those of Coddington (1989). The work on arc problematic: the genus is currently placed in linyphiid morphology (including the descriptive Tetragnathidae. although not long ago il was studies on male palp, spinneret spigot, and thought to belong in Arancidac. Recent analyses tracheal system morphology) will also be pub- of Araneoidca relationships by Coddington and lished elsewhere. Scharff suggest that Zygiella is either sister to Araneidae or Araneinae, re, it is the most basal Analysis taxon within arancids or basal within the araneine The data set was analyzed using the computer clade (Scharff Coddington, and perv comm,). program for phylogcnctic analysis Hcnnig86 ver. 1.5 (Farris, Multislatc characlers Taxonomic nott: I have used taxonomic I98&) were decisions that will be soon discussed in greater treated as non-additive (unordered). detail elsewhere- The Pimoinae Wundertieh are raised to familial status tPimoidac, NEW RANK) RESULTS and are therefore removed from the Linyphiidae. Treating pimoids as a linyphJid subfamily Characters produces a great change in the diagnosis or Character distributions arc summarized in Linyphiidae- since it is largely based in male Tabic L The desmitraehcatc hj sign) genitalic chaiaiMrr> '-vlveh auc absent in the (sensu Millidgc, 1984; character 35) is a Pimoidae [eg,, rniersegmenial parucymbium. synapomorphy of the erigonine cJade. I have not loss of the arancoid conductor, loss of the able tii confirm some of the tracheal mor- araneoid median apophysis, presence of a radix phologies described by Millidge (1986). I have and a column, etc.). Once it is established thai examined the tracheal system of several

ids and linyphiids are sister- groups, the as- erigonine genera i Erigone aletris Croshy and signment of ranks i^ arbitrary. The exclusion of Bishop. E. psychrophila, Gonatium tubens pimoids renders Linyphiidae more homogeneous (Blackwall), Grammonotuangusta Dondaie, and and easier to diagnose. Louisfagea Brignoli, as Hypselistesftorens (O.P.-Cambridgej) and have presently defined, is polyphyletic (Horrniga. in not found evidence of the median tracheae opeo-

TABLE I. Rows represent characters columns lox j. is -.; is — and Hie first State 'state 0\ the and 'state !\ctc. *T represents missing data, and ' non-apphcahje states. The last two columns give the consistency fade* (CI) and the weight (W) assigned to the character in the successive character weighting analysis (sec text). Taxon

numbers: 0= Tctragnatha versicolor. 1 = Zygiella x-ttofaia, 2 = Unyphia triangularis* 3 = Microlinyphia dona, 4 = Botyphantes luteolus, 5 & Iwpfhyphantcs tenuis, 6 = Erigone psychrophila, 1 = Walckenaeria directa, 8 = Haplinis diloris, 9 = Navafroneta vulgaris, 10 = Stemonyphantes blauvelme. The remaining taxa are species ol

Pimoa: 1 I -rupicola, 12 -crispa, \3=al(tocu/aJa. \4 = hreviatu, 15 ^curvata. Characters: I -30, male genitalia; 31-33. female genitalia; 34-40, somatic morphology; 41-45, spinneret spigot morphology; 46. 47 T behaviour. PIMOIDAE AND SYSTEMATICS OF LINYPHIIDAE 535

2 K 1 3 4 5 6 7 9 L0 11 12 13 14 15 CI W Cymbium morphology: without dorsoectal denticulated 1 process (DDP); with DDP Q 1 1 1 1 1 1.00 10

2 DDP denticles: numerous few (20); (<20) 1 ] 1 LOO 10

3 Pimoid cymbial sclerite (PCS): absent; present 'J 1 1 1 1 Q I LOO 10

PCS -cymbium connection: sclerotized, 4 rigid; membranous, flexible 1 ) 1 1 LOO 10

5 PCS membranous ridge: absent; present 1 LOO 10 6 PCS shape: elongated anteroposteriorly; reversed U; J I 2 I 1 LOO 10

7 Paracymbium attachment: I integral; intersegmental 1 1 I ! ! 1 . ! 1 (» Q 0.50 4

Paracymbium morphology: straight; large-pointed apex; 8 U or J; lingui form- fused to PCS; triangular; short- 1 3 3 3 3 3 3 3 3 2 4 6 5 5 5 LOO 10 procurved; St type

9 Paracymbium apophyses: present; absent 1 I 1 1 I 1 1 1 1 1 1 1 0.50 3

JO PetioIe:otherwise; 1 fused to subtegulum 1 1 1 1 1 I 1 1 I 1 1 i 1 i LOO 10

11 suture: Tegular conspicuous; subtle or absent 1 i 1 LOO 10 12 Mynoglenine tegular apophysis: absent; present [ I LOO 10 13 Suprategulum: absent; fused; articulated i I ; 1 ! 1 2 D o LOO 10

14 apophysis: i i Median present; absent 1 1 1 l 1 1 1 1 1 1 0.33 3

15 Conductor: present; absent 1 I I 1 l 1 1 1 1 LOO 10 16 form: small undivided; Conductor and large and bilobate 1 i LOO 10

17 ' Embolus length: long and filiform; short 1 i 0.50 2

IS Embolic absent; present ! 1 1 I I } membrane: — 1 D 0.50 2 19 Pimoid embolic process (PEP): absent; present D 1 ] i 1 ! LOO 10

20 PEP conformation: undivided; divided I LOO 10

21 PEP base: narrow; wide and lamelliform i 1 i 1 .00 10

1 22 Radix: absent; present 1 i I 1 1 ! 1 1 0.50 4

; (distal 1 i ' 23 Column Itaematodocha): absent; present 1 1 , 1 1 1 1 0.50 A

24 Fickert's gland: absent; present ! 1 Q LOO 10

25 Terminal apophysis: absent; present — I 3 ( 1 1 0.50 3

26 Lamella characleristica: absent; present ! 1 (3 — 1 1 1 .00 10

dpedipalpal tibial apophysis: absent; dorsal, rounded; 27 a 2 2 3 1 . 1 I ! LOO 10 retrolaleral; ventral

[1 28 6 pedipalp tibial spines: not clustered; distal row o I 1 1 00 10

- 29 Prolateral trichobothria in male palpal tibia: two; one i i 1 1 1 1 i> 4

30 Retrolaleral trichobothria in 6 palpal tibia: 2: 4; 3;>4 1 Q 2 2 2 2 2 2 2 0.50 3

less ; 31 Epigynum form protrudes: than its width; more — 1 1 i LOO 10

32 Dorsal plate of epigynum, projections: absent; present 1 1 LOO 10

1 Atrium: absent; — : 1 33 present 1 0.50 2

; 34 Mynoglenine cephalic sulci: absent; present i) G Q 7 1 ! !) LOO 10

-: 35 Tracheal system: haplotracheate; desmitracheate 1 i o a 1.00 10

i 36 Ectal chelicerae smooth; with ] 1 1 1 1 1 1 I of d: stridulatory striae i 0.33 1

- ? 37 Retrolaleral teeth 9 chelicera: 3; >3; 2 1 ; i i 7 2 2 2 0.50 5

38 9 pedipalpal tarsal claw: present; absent : ] a 0.50 2

1 39 Leg autospasy: otherwise; at patella-tibia 1 ! ; 1 ; 1 i i 1 1 1 LOO 10

40 Trichobothrium metatarsus IV: present; absent 1 1 1 G D D 0.50 3

41 PMS: with anterior aciniform brush; without J 1 I 1 ! ! i i 1 ! 1 I i LOO 10

<-• 42 Aciniform spigots in 9 PMS: > 1 ; 1 ; absent Q D 2 2 1 : 2 2 0.66 10

i 43 PLS mcsal cylindrical spigot base: same size; enlarged 1 1 l s 1 1 1 1 i 1 1 LOO 10

44 PLS aciniform field: random spigots; elongated field 1 (1 i 1 1 1 1 i i LOO 10

? 45 Aciniform spigots in 9 PLS:>1; I; absent D a a o o ] 1 .L 2 2 0.66 10

-' '' •) 9 7 7 •) 46 3 spins sperm web while: above sperm web; below it ? 1 ? 1 7 i LOO 10

•? '.' r t > 7 7 1 9 47 6 position during ejaculation: above sperm web; below 7 i 7 1.00 10 MEMOIRS OP THE QUEENSLAND MUSEUM

ing directly to separate spiracles, as Millidge (1965, I%9. 19S3). Blest and Pomcroy (197S) reported. The mentioned erigomnes posses a studied the sexual behavior of Haplinis diloris. 1 Intchcfil atrium [contra Millidge, 1986.57). have used data from van Helsdingen's observa- which, using an aqueous solution of chiorazjol tions as valid for the different species of Le;>- black, stains similarly to the rest of the tracheal thyphantts and Microlinyphia in my data, under »s visible at m, The spiracle most both ends, the assumption that there is no variation for the there is where it is wider and rounded, although characters under study at the intragencric level. a slit connecting both ends. Such ends are not a For Erigone psychrophila, 1 have also used data closed circle (i.e., lbey;iioin;i separate spiracles), from other species in the same genus, namely E. as Millidge's illustrations seem to suggest (e g, deniifKtlpis and E. hngipalpis (Gcrhardt (1927, his Figure 5>, since they arc open at its inner part 1 923) cited in van Hclsdingen (1983)). The male to the interconnecting slit. Siemonyphantes position during the construction of the sperm web bluuvdtae and AUomengea piwiatu (EniextoiO (fork and web) and during ejaculation is below have tracheal atria opening through u single die sperm web in crigonines and mynoglenines, spiracle, contrary to Millidge* s assertion thai in above (only the web, the fork is constructed hi ah genera the atrium opens via two spirac and from below) in Linyphiini and Micronetini. Atria opening via 9 xinglr spiral I also found in Dtapetisca alteranda Chamberlin, •omerux sylvattcux (Black wall). Lep- An.m thypluuuesflavipes(B\ackwM)*L- rewttjst Black- The data (Table 1) were analyzed using the wall), and L. intricatus (Emerton). These latter implicit enumeration option ol'HennigS6, which genera were also reported by Millidge [the first found four equally parsimonious cladograms one only implicitly) to have the atrium opening with a length of 81 steps and consistency and via two spiracles. In the two latter species the slit retention indices of 0.74 and 0.81 respectively.

is very similar to the one reported here for the These four topologies differ in the interrelation- erigor.ines, with matkedly wider round ends (this ships of pirnoids and in the position of fact might have caused them to be taken as hi Sfemonyphanies, which in one of the four two spiracles). cladograms is sister to the pirnoids, This uu:cr The spinneret spigol morphology ehara. topology is the result of the parallel loss of the (41-45) support the monophyly of the pirnoids aeiniform fields in pirnoids and linyphiids. Deac- and of the pimnid-linyphiid blade (Hormiga. m tivating the characters that account for the num-

1 Linyphiid and pirroid spigot morphology ber of aeiniform spigots in the PMS and PLS (42 s consistent with the araiveoid groundplan (Cod- and 45, respectively) and using the implicit ilingiori, 1990b; Peters and Kovoor, 1991; Hor- enumeration option three cladograms arc ob- in press). pimoid-linyphiid clade miga. The tained. These three cladograms arc the same as

1 he aeiniform brush found in primitive or- PMS those obtained with the 'active' characters, with hicidarians (character 41 but so do many tctrag- )i the exclusion of the topology that clusters nalhidsandlhc ihei idiids. Pirnoids and linyphi ids Stenwnypltantes with pirnoids. Successive char- share the position of the mesa! cylindrical spigOt acter weighting (Farris, 1969; Carpenter, 1988) on the periphery of the PLS, hu: this is not ex- was used, as implemented by Hennig86, to clusive to the pimoid-linyphiid chute: it is also choose a cladogram from the set of four equally found iftZygtelfa K-twtata (pers. ohserv.) and in some other i^iragnathids (Coddinglon, pers. parsimonious cladograms. A single iteration cumm.; Platnick etaL, 1990). An enlargement of produced one cladogram (Fig. 1), which cor- the base of the peripheral cylindrical spigot of the responds to one o( the original set of four. This PLS (chaiaeter 43} is characteristic of pirnoids result is stable in a second iteration. Because this and linyphiids. Pirnoids have drashcalh reduced cladogram is based on the most consistent char- the PMS and M S aeiniform fields (Characters 42 ter:: rs it is prefeired as a hypothesis for explaining and 45): they either have one or none aeiniform the relationships of this sample of tax*. The spigots on each spinneret. Stemonyphunft's has cladistic analysis of this selection of pi moid la.xa also lost, presumably in parallel^ the acini fo-ni produces results (i.e. tree topologies) fully con- spigots in both the PMS and the PLS, gruent with those obtained in Hormiga I in press), The use of the mating sequence and the transfer in winch a total of 20 pimoid species wlic of sperm (characters 46-47) as taxonomic eh: analysed togethei with the same sample nf in Linyphiids was studied by van Helsdi linyphiids and the two outgroup genera. PIMOIDAE AND S YSTEMATICS OF LINYPHIIDAE 537

PIMOIDAE LINYPHIIDAE *" if NS $ &* <*$> & ^ s c$rfJjrjrjJSfj^> ^ ^- <£

- -29 30 4-30 18 38 '33 -f 36

- -33 13

2, 6, 8, 11, - -12 16, 21, 45 - -20 27

- -34

10 42 26, 40, 46, 47

14 31 45 -f-30

-4-8 22 4-19

23 '27

4-41 30

'37 - -44 4-42 7,8, 14, 15, 22, 23, 29, 37 - -45

9, 36, 39, 43

FIG. 1 . Preferred minimum length cladogram for the taxa and characters in Table 1 (three equally parsimonious alternative topologies exist; see text). The cladogram length is 81 steps; the consistency and retention indices are 74 and 81, respectively.

DISCUSSION Paracymbial apophyses are secondarily absent (i.e. lost) in the pimoid-linyphiid clade, although MONOPHYLY OF THE PlMOID-LlNYPHIID CLADE they are regained in the Micronetini. All these characters (except the patellar autospasy) exhibit Pimoids and linyphiids emerge as a some degree of homoplasy. monophyletic assemblage, unambiguously supported by the presence of cheliceral stridulatory Millidge (1988) rejected the inclusion of striae, patellar autospasy, and the enlargement of Linyphiidae in Araneoidea, and instead related the peripheral cylindrical spigot base in the PLS. them to Agelenidae (s. to.), Amphinectidae, and MEMOIRS OF THE QUEENSLAND MUSEUM

other taxa currently placed in Amaurobioideaand With the increasing number of studies that use Dictynoidea. His hypothesis on the exclusion of quantitative cladistic methods it is becoming linyphiids from Araneoidea has been elegantly clear that homoplasy is quite common (Cod- rebutted by Coddington (199Ub), who stated thai dington and Levi, 1991). Coddington (1990b) linyphiids exhibit 9 out of the 10 synapomorphies noted that many of the most useful characters for chat support the monophyly of Araneoidca. rhe inference of araneomorph phylogeny wv:rc Pimoidx share live same 9 araneoid synapomox homoplasious. Griswold (in press), in his study of phics. Peters and Kovoor (1991) studied the the Lycosoidea, arrived at a similar conclusion structure, histochemistry., and function of the for female genitalic characters. Linyphiids are not spinning apparatus of Linyphia triangularis and an exception, Lirul this fact will not surprise most concluded that the data did not provide an lin.phiid taxonomists- For example, interseg- dictation of el ttionship between mental paracytnr.i.i icharacter 7). similar to the

Linyphiidae and Agelenidac. Certainly I in phiid are also found in Teiragnadxa ', typo, and Milltdgc's hypothesis lacks character support. HygfWtha (Levi, 1981:274. 286). Millidge clearly argue in favor The available data of the < 1988:258) considers the two latter cases as in- inclusion of the pimoid-linyphiid eladc in tegral paracymbia. diflerent from the linyphiid 1 % Aianemdea. Furthermoie Millidge type, In which case the Lomcplasy would be idiosyncratic method of phylogenetic inference is removed from this character. But regardless of flawed because, among olher things, it seen possible instances of homoplasy, the interseg- suggest the use of symplcsiomorphies (by mental paracymbium is a putative synapomorphy 'reversing" the outgroup comparison method) to for linyphiids. Coding the paracymbia! morphol- establish family relationships (p. 254). !t is well ogy (eliaracter 8) is not an easy task. 1 have taken known that grouping by plesiomorphic ehanicic a conservative approach by coding it with a high slates produces paraphyletie groups (Hennig, number of character states (seven), in pan due to 1966) and therefore should be avoided. the high morphological diversity of this si i ULtui j. A major problem in araneoid phytogeny is the The coding used produced no extra length (the placement of the tbendiid and the hnyphiid- character s consistency index is 1 ), bul by itself pimoid lineages, which the orb web architec- m it provides little grouping information (only two ture has been lost (Coddington and Levi, 1991). states occur in more than one taxon). The si a suggested another Cy tholipidae have been as are thus 'ordered' by the tree topology genet possible sister group of linyphiids (Coddington, by all characters (the final optimization of the 1990a). While the sheet web might support this character on the cladogram was done by hand, hypothecs, the evidence provided by the mor- because several equally parsimonious optimiza- phological data is, at ll>e moment, inconclusive tions exist). Blest (1979) and Wunderlich (1986) consider that mynoglenines and erigonines share NlONCJPHYLY OF PlMOUXS the same type of paracymbium ('simple Pi moid monophyly is supported by nine paracymbium*). Van Helsdingen (1986:122) ar- synapomorphies. six of them from male palpal gued against this view by pointing out that many morphology and two from spigot morphology. It linyphiine genera alsohave the so-called 'simple' is interesting to note lhat nunc of 'he pimoid- paracymbium. Although in some cases linyphiid synapomorphies refer to palpal mor- erigonines and mynoglenines seem to have morphology, which is quite different in these two phologically 'simpler* paracymbia (short lineages and ought reiloci b very old time of proximal and distal branches, sometimes J- divergence an

M _ Nuftr. L^ CHARACTER ANAl VSI*. AND >haped). The paracymbium type found in

Cladistic Strucribe of Linypi I. Stetrumyphantes is considered by Millidge (1988) Linyphiid monophyly is supported by eight as an intermediate form between the integral and synapomorphies. Seven of these eight characters intersegmental types. This latter type ofparacym- (i.c. all except character B) are homoplaVious. bium is inferred to be the primitive State for PIMOIDAE AND SYSTEMATICS OF LIN YPIH1DAE 539

linyphiids. This state is subsequently iransfom it d does not bear Ihe column (wiuch in some cases is into lite paracymbium morphology found in the far from it, e.g. Pseudafroneta, in Blest' s figure rest of linyphiids, Coding mynoglenines and 597) and has no sperm duct going through it. I crigonincs as sharing (he same unique character have interpreted the mynoglcnincs as lacking a state (i.e.. the 'simple* paracymbium), as Blest suprategulum (sensu Saaristo) and coded its and Wunderlich have suggested, produces no tegular apophysis as a structure synapomorphic changes in the cladogram topology for mynogJenines and not homologous io the suprategulum ('mynogleninc tegular The presence of a median apophysis and a apophy character 1 2). However, the tegular apophysis of conductor on ihe tegulum 15 regarded as Haplinis seems be functionally plcsiomorpbic for arancoids (Coddington, Io analogous to the suprategulum in some linyphiids (van 1990a) Pimoids have a conductor and a median Helsdingen (1965, 1969); Blest "and Pomeroy apophysis (Hormiga, in press). The absence in (1978)) in engaging the socket of the epigynal imyphiids of a true (i.e. tegular) conductor and a scape, but data on the functioning of the genitalia median apophysis (Coddington, 1990a) are SCfOSS taxa arc Sttll very scarce. The regarded as synapomorphies for Imyphiids (char- suprategulum of Stemonyphantes is articulated to acters 15 and 14, respectively). Two potential the tegulum by means of a membranous connec- synapomorphies of linyphiids. the radix and the tion (van Helsdingen, 1968:134; pens, observ column, are waiting forresolution of theouigroup ) and is different from the rest of linyphiid o! the pimoid-linyphiidctadeinorderto be tested. suprategula which are fused Xo the tegulum (char- The linyphiid radix (character 22) might be acter 13). The cladogram in Figure 1 suggests tike homologous to the araneid radix, and therefore possibility of independent origins for these two plcsiomorphic for linyphiids, if araneids are the types of suprategula, and therefore questions its sister group of pinxud-linyphnds 1 Coddington. homology (secondary absence of the 1990a). The same happens with the linyphiid supnilegolum in the mynoglenines -versus inde- column or stalk (sensu Saaristo, 1971; character pendent gains-requires one additional step). 23) that connects the radix to the tegulum/suprategulum. The column could be The linyphiid embolic membrane homologous to the distal haematodoeha if Helsdingen, 1969 1 is &(M bOfltolo^twis to the araneids arc sister to the pimoid-linyphiid cladc arancoid conductor because oi their different (Coddington, 1990a). If that is not the- case, the position (but see Coddington. 1990a: 16). The homology of the linyphiid radix and column with embolic membrane (character 18) is a putative Us presumed equivalents in Araneidac might be synapomorphy for all linyphiids, with the ex- refuted, and these characters would function as clusion of the basal genus Stemonyphantes. The synapomorphies of linyphiids (this latter alterna- "embolic membrane' Ma rah nypbia is not an out- tive is the one mapped on Figure 1). Linyphiids growth of the column, as in most linyphiids (van seem to have reduced the number of prolateral Helsdingen, 1986:123), but a structure 'arising trichobothria in the male palpal tibia (character from (the) membranous connection of radix, base 29) from two (pimoids and outgroups in the data of embolus, and dorsal side of lamella' set 1 to one. However this putative synapomorphy Helsdingen. 1970:6). I have interpreted it as not

;- of linyphiids might loose generality (i.e. might be homologous to the column-positioned emboli refuted) in a data set with a larger sample of taxa. membranes, but the nature of this membrane

The. same might happen to the number of remains dubious. The alternative, i & coding it ;in retrolateral teeth on the female chelicera (charac- an embol ic membrane shifted to a radical position ter 37), which is four or more in all but one of the in Microlinyphia, produces no change in llie linyphiid taxa tfl the data set, and acts as putative cladogram topologv. synapomorphy for linyphiids. The terminal apophysis (character 25) is a

The linyphiid Kupraiegulum (ctafftctei 1 J) rs a synapomorphy for erigonines plus linyphiines, projection of the tcgulum that bears the column but its interpretation offers several problems. The and through which the sperm duct pusses (SflJiri * first is its homology with its homonvm inaraneidN to 11971, 1975); Millidge (1977); Coddington (Saaristo 1971, 1975: Coddington, 1990a). Such <>a)) However, the suprategulum might nut homology is dependent, among other things, (Ml a be homologous across all linyphiids. The tegular sister-group relationship between araneids and

!• projection thai Blest (1979, Tigs 596-602) and liityp ii pimoids), but even snihe homol-

Blest and Pomeroy (1978, figs 2, 4) call the ogy is not obvious. Zygietla x-notata (which un- 'suprategulum' in New Zealand mynoglcnincs considered here as an arancid) lacks anything 540 MEMOIRS OF THE QUEENSLAND MUSEUM

similar to a terminal apophysis, pimoids lack the (similar epigynal atria are also present in other radix (therefore, we do not know if they ever had linyphiids, not included here, that are not closely such apophysis), and basal linyphiids (i.e. related to the Linyphiini; Millidge, 1984; van Stemonyphantes and the mynoglenines) have Helsdingen, in lift,). Three synapomorphies sup- simple radices and no terminal apophysis. The port the monophyly of Micronetini: the paracym- cladogram in Fig. 1 suggests independent origins bial apophyses, a short embolus (it also occurs in (i.e. non homology) for the terminal apophysis in Erigone), and the presence of Fickert's gland in araneids and linyphiids. If the embolic division of the radix. Stemonyphantes is interpreted as simple (and not The nature of the clypeal glands is another simplified) it also suggests that complex embolic interesting problem in linyphiid evolution. divisions in araneids and in linyphiids arose inde- Whether the mynoglenine sub-ocular sulci are or pendently. This latter interpretation would ques- are not homologous to the male erigonine post- tion the monophyly of araneids plus linyphiids ocular sulci is a matter of debate. Mynoglenine (e.g. Coddington, 1990a: 14). Second, and at a sub-ocular sulci are found both in males and less inclusive level, not all erigonines and females (they are very similar in both sexes; linyphiines have a terminal apophysis. Evalua- juveniles also have functional sulci, at least in the tion of the homology of these apophyses requires species of Haplinis studied by Blest and Taylor, a more detailed cladistic structure for the family 1977), they do not play any active role during the (i.e. more taxa and more characters). Another courtship (at least in the species studied by Blest radical sclerite, the lamella characteristica (char- and Pomeroy, 1978), and they probably elaborate acter 26), is a putative synapomorphy for the defensive secretions (Blest and Taylor, 1977; but linyphiines. Further support for the monophyly of this latter hypothesis has not been empirically Micronetini plus Linyphiini is given by the loss tested, although the unique ultrastructure of the of metatarsus IV trichobothrium (character 40) clypeal secretory cells is consistent with the syn- and the position of the male during the construc- thesis of a toxic product). On the other hand, tion of the sperm web and during ejaculation erigonine post-ocular sulci (as well as the (characters 46 and 47). The phylogenetic infor- cephalic elevations) are found (mostly) in adult mation provided by the latter two characters males. These erigonine sulci usually have pores should be regarded as provisional, because of the associated with glands that are cytologically dif- high number of missing entries for these charac- ferent from those of the mynoglenine sulci (Blest ters in the matrix. According to Blest and and Taylor, 1977; Schaibleer a/., 1986; Schaible Pomeroy (1978) Haplinis is unique among and Gack, 1987), and they play an active linyphiids in having an expansion of the palp mechanical role during the courtship (they are prior to its locking to the female genitalia, while gripped by the female cheliceral fangs). Never- in the remaining linyphiids for which this trait is theless, these erigonine glands are not always known the male first locks its palp to the associated with cephalic specializations. epigynum and then expands the haematodocha. Mynoglenine and erigonine ocular sulci can be However, in a recent study on African linyphiids interpreted as homologous structures within the Scharff (1990:62) described a similar expansion same transformation series or as two independent prior to locking for Neriene kibonotensis developments. The available evidence is not easi- (Tullgren). More data on the distribution of this ly interpreted in either way. The mynoglenine and character are needed in order to establish it as a erigonine sulci differ in their position, in the mynoglenine synapomorphy. cytological structure of their associated glands, Erigonine monophyly is supported by the and in their behavioral role. It seems that the retrolateral tibial apophysis of the male palp, the available data argue against the homology loss of the female palpal claw, and the des- hypothesis, since they fail to meet the classical mitracheate tracheal system (sensu Millidge, homology criteria of position and detailed 1984). In the present dataset the epigynal atrium similarity. Congruence with other character sys- (character 33) is the only synapomorphy support- tems offers a powerful test of the homology ing the monophyly of Linyphiini. An epigynal hypothesis of the sulci. Blest (1979:165) argued atrium is also present in the mynoglenine genus that the most economical hypothesis (i.e. par- Haplinis (Blest, 1979:100) but absent in simonious) 'would suggest that the sulci of the Novafroneta (Millidge 1984:241). The mynoglenine type gave rise directly to the kind cladogram suggests independent origins for these found in Erigoninae'. Mapping his hypothesis on two atria; its homology is therefore questionable his (op. cit. cladogram y p. 172, which in parenthi- 1 ,

PIMOIDAF. AND SYSTEMATICS OF LINYPH1IDAL 541

eft] notation can be summarized as: monophylelic group, even in a schematic manner Mynogleninae (Linyphiinae, Erigoninae)) re- My study should be considered only a prelimi- quires the gain of the mynoglenine type of sulci nary sketch of linyphiid relationships. Clearly, a in the common ancestor of all linyphitds, much larger sample of laxa is needed before the profound modifications (morphological, main monophylelic groups can be established. cyiological, and behavioral) of the sulci to The addition of new tax3 and new characters achieve the erigonine type of sulci (either in the might affect the cladogram topology presented afl£e£tral engonines or at the level of the here. As we have seen, non-homoplasious char- linyphiine-engonine ancestor) and finally the acters for wide ranges of taxa are more the excrp loss of the sulci (and its accompanying glands and tion than the rule, and different character systems behavior) in the linyphiines. The alternative often delimit conflicting monophylelic groups, hypothesis (i.e. non homology of mynoglenine When large numbers* of taxa and characters arc and erigonine sulci) maps on the mentioned studied quantitative studies are imperative cladogram as two independent gams of the two Cladistic studies provide explicit and tesi types of sulci. The evolution in parallel of the hypotheses of relationship and are recognized as erigonine and mynoglenine sulci would then ac the most reliable method for retrieving ll»e count for their differences. Although the latter phylogenelic pattern thai underlies organic diver- hypothesis is more parsimonious (in both Bk\t s sity. Not until this approach is adopted will ad and my cladogram) this question contiol be truly varices in linyphiid higher classification become tested until more data (taxa, particularly those a reality. with any type of sulci and/or glands, arid information on the glands) are included in the data set. ACKNOWLHDGEMENTS This is due to the effect that mynoglenine and erigonine cladogram topologies might have CO I am indebted to Drs. J.A_ Coddington, C. Mit- the optimization of the characters) on the ten C.E. Griswold, NT. PLatnick, and N SchanfT [inyphiid cladogram. Only then we will be able for helpful discussion and comments on an earlier ui usses alternative hypotheses on the evolution version of this manuscript. Financial support for of these cephalothoracic specializations. this study was provided by the University of The linyphiid tracheal system needs to be Maryland and the Smithsonian Institution. studied in detail and re-evaluated New morphological descriptions are needed, since at least LITERATURE CH some of the available comparative data are inaccurate (see above). Millidge's (1986, figure 12 BLEST. A.D. 1979. Lmyplmdac-Myrjoglenmac, Pu 95-173, scheme for the evolution of the tracheal system In RR Furster and \D West. TV spiders of New Zealand, part V. Otago Museum in linyphiids is therefore not valid, because it in Bulletins: 1-175. partially based on inaccurate data. BLEST A.D. & POMEROY. G. 1978. The sexual I* The most parsimonious hypothesis to explain haviovr and genital mechanics of three species uJ the data presented in this study is the cladogram Afvnogiene* & TAYLOR. H.H. 1977. Hie ciypcat the three equally parsimonious alternatives that glands of Af\nn$l*vu's Simon and some otli-.r e*»st) relationships di ftcrent from those proposed linyphiid spiders. Journal of Zoology (London) by Wunderlich (1986:106). The mynoglenines 183: 475 | are considered here to be relatively basal CARPENTER. J.M. 1988 Choosing nrnoni multiple linyphiids, while Wunderlich suggested them as equally paisuuoniou.-. cladoirrams. Cladisu sister to the erigonines. Both hypotheses agree on 291-2%. considering the pimoids and Stemonypfaintes as CODDINGTON. J.A. 19W. Spmuciei mIK Splgo! mor- the most basal clades, and on the monophyly of phology: evidence for the monophyly ol orbwearv* the Micronetini plus the Linyphiini. To use either ing spiders, Cyrtophorinae (Aiaueidac), and the group Thcridiidac plus Ne.siieidac. Journal :>f of these two phytogenies as a classi ficatiofi would Arachnology 17:71-95. be premature. Wunderlich did noL explicitly list 1990a. Ontogeny and homology in the male palpus the synapomorphics that define the monophylctic of orb-weaving spiders and their relatives. wirh groups in his cladogram, synapomorphie comments on phytogeny ( Araneoelada: mixed up with diagnostic characters (some of Araneoidt.-a. DctROpoidca] Smithsonian Con- which are not synapomorphic), and there is no tributions tO /.oology 496: 1-52. mention of the genera included in each I990h ChuiistieK and spider classification: MEMOIRS OF THE QUEENSLAND MUSEUM

Arancomorph phylogcny and the monophyly of LArancac, Aranoomotphac). American Museum orbweavers {Araneae: Aranemorphae; Or- Novitatcs 3016: 1-73. bkulariae). Acta Zoologica Fennica 190: 75-87. SAARLSTO, M.I. 1971. Revision of the genus Maw CODDINGTON, J. A. & LEVI, H.W. 1991. Sys- Oj?.-Cambridge (Araneae, Linyphiidac). Annales tematic* and evolution of spiders (Araneae). An- Zoologici Fcnnici 8. 463- 482. nual Review ot Ecology and Systematica 12 1975. On the evolution of the secondary genital 565-592. organs of Lepihyphantinae (Araneae.

HARRIS. J.S. 1969. A successive approximations Bp Linyphiidac ), Proceedings of the 6th Internation- pranch lo character weighting. Systematic Zool- al Arachnological Congress (Amsterdam, 1974): 21-25. ogy I B: 174-385. Hcnnig86, version 1.5. (Computer program SCHAIRLE, LI & C.ACK, C. 1987. Zur Morphologic Histologic biologischen Bedeutung der ihuted by its author, 41 Admiral Street, Port Ulld Kopfstrukturen einiger Arten der Gattung JellvtNi>n Station, NY 1 1776i. Oiplrn'ephalus (Araneida, Linyphiidac, GRISWOl.D, C.E. (in press), investigations into tin Erigoninae). Verhandlungen des Naturwis- phylogcny Of lycosoid spiders and their kin 171- (Arachnida. Araneae, Lycosoidea). Smithsonian scnsschaftlichcn Vercins in Hamburg 29: Contributions to Zoology. 180,

, C. PAULUS. H.F. 1986. HENNIG, W. 1966. 'Phylogcnetic Systematica SCHAIBLE. U GACK, & Zur Morphologic, Histologic und biologischen (University of Illinois Pre*,s: Urbana). 263pp. Bcdeulung der Kopfstrukturen mnnnlichcr HORMIGA, G, In press, A revision and cladistic Zwergspinnen (Linyphiidae: Erigoninae). analysis of the spider family Pimoidae (New Zoologischc Jahrbuchcr (Systemalik) 113(3). Rank- Araneoidea, Araneae). Smithsonian Con- 389408. tnbtKions to Zoolof SCHARFF, N. 1990. Spiders of the family Linyphiidac nrb-'.veuxei . H.W. 198!, rhc D genera from the L^ungwa mountains, Tanzania Daliehogfiurfm and fi'tragnatha north of Mexico [Araneae). Entomoluuica Scandinavica, Supple- iilvic: Araneida, Trlragnathinae), Bulletin of ment No. 36: 1-95. the Museum oi Comparative Zoology L49: 271- VAN HELSDINGEN, P.J. 1965. Sexual behaviour of

/ -cpthyphantes ieprosus (Ohlcrt) (Araneidac: M1LLIDGE. A.F. 1977. The conformation of the malt- Linyphiidae), with notes on Ihc function of the palpal organs of Linyphiid spidcis and its applica- genital organs. Zoologische Mededclingen

tion to the ta\onomicandphyiogenctic uual) ts I (Leiden) 41: 15-42. die family (Araneae: Linyphiidac). Bulletin ol the 196$, Comparative notes on the species of the British Arachnotogical Society 4; 1-60. h ol a re t i C genus Stemonyphames M e n g e 1WI. The taxonomy of the Linyphiidac. based (Araneida, Linyphiidac). Zoologische i fiK'fly on the cpigynal and tracheal characters Mededclingen i Leiden >43< 10): 1 17-139 . (Araneae: Linyphiidac). Bulletin of the British 1969. A reclassilication of the species of Ltnyphut Arachnological Society 6: 229- 267. Latrci lie. based on the functioning of the genitalia 19K6 A revision of the tracheal structures of the (Araneida, Linyphiidac) I. Zoologische Vcihan- Litiyphitdae (Araneae). Bulletin of the British delingcn (Leiden) 105; 1-303 Society 7: 57-61. Arachnological 1970. A reclassification ot the SpCCtCS of Linyphia rhc relatives of the Linyphiidae; phylogenetic Latrei lie, based on the functioning of the genitalia

problems at the family level t Araneae). Bulletin | Araneida, Linyphiidae) II. Zoologische Verhan- ol the British Arachnological Society 7:253-268. delingen (Leiden ) 111: 1-86. PATTERSON, C. 1982 Morphological characters and 1983. Mating sequence and transfer of sperm as a homology Pp. 21-74. In Jovsey. KA and Friday, taxonomic character in Linyphiidac (Arachnida: A.E. (eds) 'Problems of Phylogcnetic Araneae). Verhandlungen des Naturwis- Reconstruction'. (Systematic* Association Spe- sensschaflliehen Vcreins in Hamburg 26. 227- cial Volume No. 21. Academic Press: London and 240. N<*w York), jyS6. World distribution of Linyphiidae, Pp. 121- PETERS, H M.&KOVOOR.J 1991-Thesilk-produc- 126. In Eberhard W.G., Lubin YD., Robinson big system of Linyphkt triangularis {Araneae, B.C. (eds) 'Proceedings of the Ninth Internation- Linyphiidac) and some comparisons wilh al Congress of Arachnology. Panama. 1983'. Aianeidae. Structure, histochemistry and func- (Smithsonian Institution Press: Washington, tion, /oomorphology III: 1-17. D.C.). k PLATNJCK, N.L, CODDINGTON. LA.. FORSTER. WUNDERLICH. J. 1986. Spinncnfauna gestern und R.R. & GRISWOLD, C.E. 1991. Spinneret mor- heute\ (Erich Bauer Verlag bei Quelle and Meyer phology and the phylogcny of haplogync spiders Wiesbaden). CRITERIA FOR IDENTIFYING THERMAL BEHAVIOUR IN SHDliKS A LOW TECHNOLOGY APPROACH

W.F HUMPHREYS

Humphreys, W.F 1993 1 ! I J: Criteria for identifying thermal behaviour in spiders: a low technology approach. Memoirs ofthe Queensland Museum 33(2): 543-550. Brisbane- ISSN 0079-8835.

The widespread occurrence of thermal behaviour in diumally active web spiders is either

largely ignored or not recognised. Thus it obfuscates some explanations of the function of the stabilunentum on spiders webs. Thermal behaviours of four spiders (Nephila edulis, N. macutata, Gasteraeantha minax and Neogea sp.) are examined using technology which is often inappropriate for field studies. Many thermal behaviours are recognised as well as behaviours which facilitate thermal behaviour. The thermal correlates of these behaviours arc established. Some observational criteria are derived, which require only simple equipment, by which thermal behaviour* may be recognised in the field and distinguished from other behavioural patttms.[jAmncae, orb-weavers, thermal behaviour, thermoregulation, stahilimentum.

W.F. Humphreys, Western Australian Museum, Francis Street, Perth, Western Australia 60(H), Australia. 26 October, 1992,

Many orb-web spiders remain active at the web gressively the. heat loading, behaviours thai are hmb during the day. where they can continue to themselves mostly graded (Humphreys, 1992). feed, mate, and defend their web site from the These include stilting, drooping, orientation, same and other species, as well as produce or front leg raising, abdomen pointing, posturing, respond to attractant signals (acoustic, tactile, front leg rotation and web abandonment. As- visual or chemical). The seasonal and diurnal sociated behaviours include silk laying and agita- duration of this activity can be extended by adopt- tion (Humphreys. 1992 and unpublished), In ing behaviours that warm the spider (Robinson addition, the use of a disx: stabilimentum as a sun and Robinson, 1974, 197K; BiereandUelz, 1981) shade, suggested by Robinson and Robinson when it would otherwise be too cold or else that (1973: 283), is an effective thermal behaviour in prevent it from overhauling (l.tibin and Henschel, Neogea sp. (Humphreys, 1992), Such behaviours 1990; Humphreys. 1978, 1987a, 1991, 1992). not only maintain the animal within its heat

I consider below mainly behaviours that tolerance range, but also serve to maintain I he prevent overheating. In essence, in hot weather body temperature (T*>) within a narrow range the spider postures so as to align the long axis of (presumably some optimum temperature) for ex-

its body with the sun's rays and in this position it tended periods of time (Humphreys, 1974, 1978, tracks the sun's apparent movement during the 1991). As body temperature has wide implications in physiological, behavioural, ecological day - Such behaviour is not exclusive to spiders of open country nor in tropical climates but is found and evolutionary contexts (Willmer, 1991), it is also in both temperate and tropical forest spiders important to recognise thermal behaviour in (Biereand Uetz, 1981; Humphreys, 1991, 1992 spiders in order to allow different types of ex- ami unpublished)- The standard interpretation of planations for their behaviours.

this thermal behaviour relies on a simple physical That the thermal behaviour of spiders is not model; posturing minimises the projected surface beiftg rCCOgni$e

Unwin, 1 98 1 : but see Humphreys, 1 986) or, in an provide mechanical support (Robinson and anti-predator hypothesis, minimising the sil- Robinson, 1973), to function as anti-predator houette against the hrightest background, else the devices (Eberhard, 1973, Edmunds and Ed- body area most brightly illuminated. munds, 1986;Lubin, 1986), to attract prey (Ewer, Recent observations have shed some light on 1972; Craig and Bernard, 1990) or to collect this seemingly simple process and revealed a water (Olive, 1980; sec also Ewer, 1972; Robin- sequence of behaviours serving to reduce pro- son and Robinson, 1973: 283). Neogea sp. in MEMOIRS OF THE QUEENSLAND MUSEUM

Papua New Guinea uses a disc stabilimcntum as sociobiological problems There needs to bo a pArteo), which, together with a sequence of some observational criteria using only low tech- other behaviours, each themselves graded, nology by which thermal behaviour may be reduces its heal loading (Humphreys, 1992). The recognised and distinguished from other be- demonstration that stabilimenta may he used in a havioural patterns. thermoregulatory rule raises questions concern- The examination of thermoregulation in orb- ing many observations which have been inter- weaving spiders is problematical as most preted as supporting the anti-predator role of methods used on vagrant spiders (thermal stabihmenta. preferendum apparatus, thermocouple implanta- Some observations and deductions have tion, lemperature transmitters) arc not ap- resulted in an anii-prcdator function being propriate. The most promising apparatus for such ascribed to the stabilimenuim. However, in the studies is the use of remote infrared telemetry absence of other, information, these observations (Suier, 1981; Humphreys, 1991, 1992), although are equally open to interpretation in terms of model spiders may be effectively used to deter- thermoregulatory hypotheses (see Humphreys, mine Te (Riceherl and Tracy, 1975; Hensehel tt 1992). For example -1. Only spiders that remain oi. 1992). at the hub of the web during the day produce Can criteria be established which would enable stiibilimenla (Eberhard, 1973). 2. Spiders with workers . using cheap and readily available equip- shuttle one side of the stabilimenta may from web ment, to establish that spiders ure behaving if] a to the other (ibid.). 3. The legs assume an 'aligned manner consistent with thermoregulation and to posture' by day but not by night (ibid.). 4. The identify the lhermal conditions under which they amount of silk used is directly related to openness initiate such behaviour? of the habitat (Marson. 1947), Eberhard (1973) found that Ulohorus divtrsus METHODS Marx used more silk in its stabilimentumon light than on dark nights (see also4. above). The trend for larger stabihrnentum at brighter Mies Observations on Nephiki edulis (Labillardiere) camouflaged those spiders most susceptible to were made both on Rottncst Island and in Perth. attack. However, without pertinent behavioural on Gasteracuntlni minax Thorell in the south- data, it is not possible to refute the hypothesis that west, of Western Australia, and on Nephita the open sites arc more exposed to direct sunlight nweuliitij (Fabric tus) in mangrove at Port Benoa, and thus that the stabilimenta are used to protect Bali, Indonesia the spider from ultra violet light (but sec Craig Spider temperatures were recorded from 1 100 and Bernard, 1990) or to reduce its heat load. -1500 hours using an infra-red thermometer Posturing by spiders that remain by day at the described elsewhere (Humphreys, 1991, 1992). web hub is rarely mentioned in the literature on The temperature was recorded of undisturbed stabilimenta and defense mecham r ex- spiders resting above and below the hub of the ample, although Edmunds and Edmunds (1986: web. bo ! h in the shade and in the sun. It was 83) found that species of Argute, Septula, recorded at intervals and as soon as possible (<3s) U'ucau^e, Cyrtophora and the Gastcracanthinae after a change in behaviour. Spider behaviour remain at the hub of the web during the day, they changed according to the incident light; this made no mention of posturing. In the varied because the site was sometimes shaded by Australasian region species in all these genera or by clouds. To induce more behavioural readily posture and track the sun (Humphreys, sequences the spider was sometimes shaded ar- 1991, 1992; W.R Humphreys, unpublished). tificially and the direction and strength of the Clearly, thermoregulation hypotheses need incident sunlight adjusted usi ng a mirror A plane more consideration in field studies uf diurnally mirror was used to alter the apparent position of active spiders. More recently this has occurred the incident radiation at about the same intensity (Henscbel el al., 1992; Humphreys. 1991. 1992; as the natural sunlight and a concave mirror was Lubin and Hensehel, 1990; Ward and Hcnschcl, used to alter the apparent position of the incident 1992) but such studies often require expensive radiation and at an intensity continuously vari- equipment to examine the thermal behaviour. able from greater to less than the intensity of the Such equipment may he inappropriate to a field direct sunlight. In the field, control of intensity biologist primarily interested in observation and AftS cl tide in the wind owing to movement of the

manipulation to examine behavioural or web , and hence the spider; control of the intensity THERMAL BEHAVIOUR IN SPIDERS 545

Behaviour Mean s.d. n 1 Range f wnt tegs raised: legs 1 and II are raised off the web aligned parallel to the incident radiaiwn; Refuse in shade 32.6 a 2.10 J 2 1 29 .0-35 J and ihis OCCURS as a graded sequence with the first pair Repose in partial sun 32.4 ah ! J0 19 [295-343

';! m - being rai*ed before the second pair. Repose in sun ; 1.66 47 |3l6-3R;i Orientation: the angle of the saggital plane of ihc

D spider is rotated to lie parallel to the solar azimuth TABLE 1 . Mean temperature {Tb C) of N. erfalis in while Repose position on Rottnest Island, Western the Jong axis of the body stays in the p the Australia. Tb of spiders in Repose differs according of web. lo energy intensities oi their location (shaded, partly Parr orientation: the spider is not in the Repose shaded, sunlight ANOVA - Fs %j$ = 19.895, position and has partly orientated its saggital plane I'<0 .001). Common letters include means not differ- between the Repose position and the orientated ing significantly (Fisher's PLSD at «= 0.05). position Posturing: change in the angle between the web of radiant energy is therefore relative and greater plane and the anterior-posterior axis of the spider. or less than the natural incident radiation. Repose'.posit ion: spiders occupy the lower or upper following temperatures arc The mentioned: T ; , T surface of the hub with the prosoma pointing = of ambient shaded air; Tb = of the spider s body downwards; the anterior-posterior axis of the which by default is the abdomen (Taj), other spider is parallel to the plane of the web. the thorax (Tifa). The environmental temperature Rotate fronr legy. following front legs raising die (TV) which is used as a shorthand for the effective 1 are rotated ftirwimlv such that Ihey ar:- heat load on the spider taking account of all Stretched out in from of the spider f \nd lie in rhr energy gains and losses. Means are followed in shade of its body when it is fully postured, this may parentheses by the standard deviation of the occur a> a graded sequence with the first pair being meat), and sample si7.e rotated before the second pan. Definitions required for this discussion arc p. Silk laying: adjusts the web stuctrurc near the hub en below (sec also Humphreys, 1991 1 1992). apparently to aid leg placement the better lo pos Abandon web: the spider leaves the web, often after ture and orientate to achieve the full Fabian posi- a sequence of very agitated movements, and moves tion. This facilitates subsequent thermoregulatory to the shade provided by the objects to which the behaviour but is not itself thcrrnorcgulatoi main anchor lines ol the web arc attached. y Stan to posture: when the spider changes the angle Abdomen pointing: ihe abdomen alone is orien- between its anterior poslcnot axis and the plane oJ tated to the sun as a prelude to full posturing. This the web such thai its long axi-- is parallel to (be behaviour is strongly tcprcscnted in some spei incident radiation .'Humphreys 1992). SMiing:, describes the 'standing on tiptoe* be- Agttailon and body Uft: the spider appears agitated haviour of scorpions used to prevent overheating and circles its body around the web's hub and in (Alexander and Ewer, 1958); here it descr the process the body is raised away from the web. similar behaviour m spiders (Humphreys, 1992). The latter is sometimes seen on its own and they are included here under the same behaviour. This

hody lift is not comparable to stiltine (Humphreys, ;dbnce 1992). Drooping: the spider hangs limply from the back Statements otherwise unsupported are based on legs with apparent loss of hydrostatic pressure; the my unpublished observations.

appearance is like that adopted by B spider imme- Repose Position and Thermoregulation diately after moulting while the new cuticle is Spiders adopt the repose position ifTe is below hardening. some critical level and they do so whether they Fabian position (Humphreys. 1991): the spider arc in shade or sun, hence, direct sunlight alone aligns its long axis parallel to the direction of does not cause spiders to posture, However, IV in incident sunlight with ihe prosoma facing away sun is higher than in shade (Table 1). For ex- from the sun. This position may be achieved by ample, during cool weather (low T«) in the orientation and/or posturing. When the incident and during hot weather in the shade all in- sunlight is parallel to the web plane then the Fabian dividuals are in the Repose pavilion on their v position may be the same as the Repose position if not otherwise engaged in activities such as (Humphreys, 1991). Continued adoption of the maring, web building etc

Fabian position results in the long axis of the spicier When Te is not sufficient to cause posturing in G. kfng the sun during the day. rmnuA the proportion of spiders in the Repose MEMOIRS OF THE QUEENSLAND MUSEUM

position during daylight docs not differ between Such activity pattens have been reported for many 2 = heliothcrmal spiders (Humphreys, 1978, 1987a, sunlit (36/38) and* shaded 132/3?) sites (\ i 0.016. P^ 0.90). 1987b) and other taxa (e.g. reptiles: Heatwole,

The mean Tt> temperature of ft. edutis in the 1970). Repose position was directly related to the inten- S- edutis abandons the web at 44.8°C (±0-50, 3) sity of the incident radiation such that spiders in and moves into shade. In Perth, Western Aust- the sun were hotter than partly or fully shaded ralia, when the shaded air temperature was extreme spiders (Tabic I). (46.2°C) many N. edidis failed to seek cooler Spiders on non-horizontal webs almost invariably places and fell dead from their shaded webs rest on the underside of the web with the prosnma through heat stress (G.A. Harold, pers. coonm, pointing down. However, spiders resting in posi- 1991). tions other than the Repose position should not be Large spiders assume the Fabian position earlier in taken as proof of thermoregulatory behaviour be- the day Lhan do small spiders (e.g., A?. vxacuiataY- cause some species, such as Verrucosa and this is consistent with the thermoregulation

Cyclase reputedly adopt a head up stance i Foelix, hypothesis because larger bodies have a higher 1982: 139), as does G. minax at night and Ar- tempeiature excess (Willmer and Unwin, 1981). gyrodes amipodianus OP. Cambridge, generally. However the threshold temperatures for given behaviours could be size related and lower in small Orientation and Posturing: Evidence for lhan in large spiders. This is consistent with the Thermoregulatory FUNCTION seeming generality that the tolerance zones of temperatures ex- In hot weather, spiders orient or posture on the animals are related to the example, very small N~ edulis start web to attain the Fabian position and then track perienced. For posture at 36.0°C (±2.39. 4>. significantly the sun's apparent movement. They do this ir- to cooler, average of 4.4°C =13.279. respective of web orientation. Heat and sunlight by an (Fsijrj the same are needed to obtain Lhese behaviours. On very p=0.002), than adult spiders undergoing hot days, spiders may leave the web altogether behaviour (40.4°C ±2.16.27). and seek shade. Large spiders assume Fabian posture earlier than small ones possibly because Orientation and Posturing in Response to under given environmental conditions large Manipulation spiders reach higher body temperatures. How- redirected and intensified ever, small spiders may have lower threshold Experiments with sunlight can be conducted to influence the be- Tb's. haviour of spiders to assist in determining In hot weather an individual spider in the sun will whether the behaviour is thermoregulatory use reorientation aiwVor posturing to align the without having to measure body temperature. The anterior-posterior axis of its body parallel to the results arc consistent for several species includ- direction of incident sunlight with the prasuma N. edulis, N. tnacuiata. facing away from the sun and thus achieve the ing Arachnum higghsi, G. minax and Neogea sp- Fabian position (Robinson anil Robinson, 1974, 1978; Humphreys, 1991, 1992). The spider will in cool sunny weather when the spiders are in the adjust this position during the day and track the Repose position, additional healing (by con- redirected sunlight using a concave mir- apparcntmovementofthesun(Humphceys, 1991 ). centrating ror), results in spider orientating-and/or In hot weathei all individuals in the sun orientate the posturing if necessary-to assume the Fabian posi- in l he same- direction irrespective uf the orientation tion. The redirected sunlight does not alone alter of their webs, namely they all assume the Fabian position by posturing andVor Teonentatton the behaviour of the spider (sunlight redirected at natural intensity using a plane mirror). Hence, l Humphreys, 1991). posturing is dependent on the intensity of the heat Heat atone does not cause the thermal behaviour applied. because in hot weather tinder heavily overcast conditions spiders do not assume the Fabian posi- In hot weather a spider which has assumed the tion. However, if intermittent direct sunlight Fabian position will resume the Repose position if

it artificially strikes the spider it assumes the Fabian position clouds obscure the sun or is shaded intermittently even if lower than the natural insolation is reflected it of a concave ror. On very hot days spiders may assume the Fabian onto hy means m» position in the morning and afternoon but leave the Spiders in the Fabian position in the sun will, if web to seek shade during the middle of the day. artificially shaded, assume a new Fabian position THF.RMAL BEHAVIOUR IN SPlDbkS 547

HG. 1. Thermoregulatory postures adopted by golden web spider, M clavipes. Lines projecting from force successive positions of sun (S1,S2, S3) indicate corresponding orientations of long axis of spider's body. Lateral views (a,c) and plan view (b) of web show posture assumed in response to ventral (through the web) insolation; h lateral insolation; c dorsal insolation (redrawn from Robinson and Robinson. 1974).

if the direction of the sun is artificially changed by increasing temperature of the abdomen (Humph- means of a mirror. reys, 1992). When a spider in the Repose position

In hot weather when the spiders arc in the Repose on the disc stabilimentum is heated by the sun it

position in the shade, redirection of unconcen- initially 'stilts*, as has been described for scor-

trated sunlight, by means of a plane mirror or a pions. The spider gradually raises its body away

concave mirror, results in the spider orientating from the disc surface until it has full downward and/or posturing if necessary to assume the Fabian extension of all legs and seems to be standing on position. 'tiptoe', thus removing the body of the spider as In hot weather, when the spiders are in a Fabian far as possible from the disc's surface. position, adding reflected sunlight from a plane On further heating, Neogea sp. progressively mirror results in the spider orientating with respect orientates its body and then gradually postures,

to the mean angular direction of the two sources of starting with the abdomen. It rotates the tip of the incident radiation (when angle <90°). A spider abdomen towards the incident radiation and this posturing between two heat sources will turn to- minimises the projected surface area of the ab- wards the one increased in intensity and vice versa. domen exposed to the sun. As the posturing Hence, the Fabian position moves towards the develops the prosoma also is aligned with the incident radiation of the mirror if the heat reflected abdomen so that the entire spider is orientated from the mirror is increased and towards incident prosoma from the sun with minimal silhouette area radiation of the sun if the heat reflected from the exposed. Eventually the legs themselves are rot- mirror is decreased. ated forwards until they are parallel to the long axis If redirected light comes from above the horizontal of the spider in which position they arc in the plane of the web spiders rapidly change their shadow of the abdomen, as is the prosoma; this is Fabian position; if redirected light comes from the full Fabian Position from which the spider below the honzontal plane of the web, a naturally tracks the sun (Humphreys, 1991, 1992). By these impossible position, spiders appear confused and means the spiders potentially can obtain very fine change position frequenUy and some species never control of their silhouette area and hence on their achieve the Fabian position (e.g. N. maculoia). temperature. In N. edulis many behaviours were recognisable as Cascading Behaviours similar to those observed in other spiders (e.g. Many thermally related behaviours that have orientation, posturing, agitation: Humphreys, been categorised are themselves graded so that 1991, 1992), whereas others have not previously they each develop progressively rather than been reported or recognised (e.g. drooping). Some switch from one state to another behaviours recognised may be components of the

For example as Neogea sp. warms in the sun, il same behavioural sequence For example, Agita- exhibits a progression of distinct behaviours, each tion, in which the spider circles around the bub. of which is graded and which are associated with involves the body being raised slightly from the 548 MEMOIRS OF THE QUEENSLAND MUSEUM

web, a behaviour sometimes seen on its own. Both Behaviour Mean s.d. n Range

- — behaviours are included here under the same Repose see Table 1 — —

category. Thermally there appears little difference Part orientation 33.9 0.57 6 33.1-34.7

between three categories recognised here as dif- Drooping 36.5 1.57 8 35.4-38.8

ferent behaviours (agitation and body lift, front Orientation 38.7a 2.30 9 34.2-40.3 start to posture). conducted legs raised, and Work Agitation and body lift 38.7a 2.09 21 35.2-43.5

under more controlled conditions in the laboratory : l ront leps raised 38.9a 2.84 M 35.0-43.3 may separate thermally these behaviours or allow Start to posture 4D 4b 2.16 27 34.8-43.6 their pooling using more rigorous criteria. Silk laying 41 Ibc 1.34 6 39.3-42.6 Eleven behavioural categories are recognised in N. Rotate front legs 42.4cd 0.99 10 41.2-44.3 edtdlis, ranging from Repose to web abandonment Abandon web 44.8d 0.50 3 44.3-45.3 which occur between Tb of 33.9 and 44.8°C (Table

2). The spider temperature associated with many TABLE 2. Mean temperature (Tb °C) of A/, edulis on of these behaviours is significantly different from Rottnest Island, Western Australia, associated with others. Some of these behaviours reduce the different behaviours (ANOVA - Fs 8,93 = 15.233, projected body surface area exposed to the sun and P<0.001). Common letters include means not differ- thus, under the predictions of the physical model, ing significantly (Fisher's PLSD at a= 0.05). result in lower equilibrium body tempera- should orientation (Humphreys, 1992) makes no sense ture, all else being equal (e.g. orientation, postur- under alternative hypotheses but is entirely con- ing, leg raising, leg rotation and web sistent with, and predicted from, the ther- abandonment). Other behaviours may not be thermoregulation hypothesis. moregulatory but are associated with the onset of The body temperature of a spider is a complex the next behaviour in the graded series (e.g. agita- function of many intrinsic factors (size, morphol- tion and body lift) or facilitate a subsequent stage ogy, attitude, physiology, reflectance, etc. ) as (e.g. silk laying to enable correct leg placement for well as factors extrinsic to the individual (e.g. full posturing). wind speed, turbulence, air temperature, incident As in Neogea sp. (Humphreys, 1992), some be- radiation and its spectral characteristics; Mon- haviours themselves form a graded series which teith and Unsworth, 1990). It is because Tb is a should proffer gradually increased ther- complex function of intrinsic and extrinsic fac- moregulatory effects. Both front leg raising and tors that a poor predictor of thermal front leg rotation occur initially in the front legs makes Ta the observation that spider followed by the second pair of legs. In addition behaviour. Hence, a contralateral legs are not necessarily lifted or may not always assume the Fabian position (or rotated at the same time. other presumptive thermoregulatory behaviour) at the same Ta does not imply that the behaviour DISCUSSION has no thermoregulatory significance. For example, Lycosa godeffroyi Koch in Canberra

began basking at much lower Ta in winter (4°C) Two classes of observation refute the than in summer (17°C) and reached 35°C on clear hypothesis that posturing serves an anti-predator winter days at Ta of 11°C (Humphreys, 1974, role as stated in the introduction. 1978); the latter shows the dominant role of Firstly, if the Fabian position reduce the sil- boundary layer effects for such surface dwelling houette area against the brightest part of the sky spiders. Although orb-weaving spiders are often as an anti-predator defense (i.e. to make them less high above the ground, such boundary layers may visible) then they should posture to the sun under clear conditions irrespective of the intensity of assume more importance in those orb weaving the sunlight; they do not. Furthermore, under spiders that incorporate a surface in their web conditions of patchy heavy clouds (cumulus and (e.g. Neogea sp. and leaf curling species; eumulo-status) against a clear sky, the spiders Humphreys, 1992). should assume a Fabian position with respect to These many classes of observation support the the brightest sector of the sky; they do not. hypothesis that the posturing and/or reorientation Secondly, many thermally related behaviours that spiders undergo in intense sunlight is of are themselves graded so that they are exhibited thermoregulatory significance. Many are not progressively as the spider warms. This provides alone adequate to support unequivocally the ther- the strongest evidence for the thermoregulation moregulation hypothesis (e.g. Table 3: 4, 6 , 9), hypothesis because partial stilting, posturing or some, in combination with others, support the THERMAL BEHAVIOUR IN SPIDERS 549

Behavioural response and thermal TABLE 3. Summary of Condition at spider Manipulation consequences characteristics of ther-

1 Cool weather in shade Nil Repose; Th=Ta moregulatory behaviour in orb- spiders. 2 Cool weather in sun Nil Repose; Tb>Ta weaving Definitions: Spiders as- 3 Cool weather in sun =S Repose; Ti,>Ta sume Fabian position in 4 Cool weather in sun >S Orientate and/or posture; TS >TC hot but not in cold weather. '" Hoi weather in shade Nil Repose; TY^Tj S denotes sunlight 6 Hot weather in shade = sor >s Fabian redirected onto the spider at Orientate and/or postureX tracks sun; about intensity of natural 7 Hot weather in sun Nil minimise 1Y, sunlight using a plane mir- 8 Hot weather in sun «5 Repose ror (=S) or at greater or less

9 Hot weather in sun =Sor>S Fabian than natural intensity using a concave mirror or Repose; in extremis may suffer heat death (>S < 1 Very hot day in shade Nil without posturing S). The spider is simul- taneously artificially 1 1 Very hot day in sun Nil Seeks shade shaded (s) or not (S). Num- All spiders orientate in same direction; 12 Population in hot sun Varied web orientation bers with * are considered minimise 1 1, alone, and numbers fol- 13 Population in hot sun Large & small spiders Large spiders posture earlier in day than small lowed by common letters Spider postures mid-way between two 14 Hot weather in sun =5 perpendicular to sun considered together, to incident heal sources are support strongly the ther- Grade n behaviour; behavioural cascade 15* Grade n-1 behaviour >S culminaling in Fabinii moregulation hypothesis.

16 Hot weather in sun =S below horizontal Apparent confusion in some species;> Tb

17 Cool weather in shade =s Posture to > exposure to heal source; Tb>Ta hypothesis (e.g. 1-7), while others support no tangling thermoregulatory from other be- other hypothesis (e.g. 15). haviours. While the emphasis here has been on behav- A thermoregulator with the battery of finely iours that reduce the heat load, spiders should use graded behaviours seen here should, under ideal behaviours to warm them in order to enhance the conditions, be able to maintain a near constant time they are at optimal temperatures. This is the body temperature under a wide range of environ- case in burrow dwelling lycosids (Humphreys, mental conditions. In practice their temperatures fluctuate with every air movement, at 1974, 1978, 1987a, 1987b) as well as in orb web markedly least partly owing to their small thermal capacity. spiders which may seasonally orientate their If the spiders are innately incapable, owing to webs to maximise the projected surface area to their small mass, of precise thermoregulation, warm more or faster (Carrell, 1978; Tolbert, why have they developed such a wide range of 1979). sophisticated behaviours which should permit While there is an indication of size related precise thermoregulation? effects in thermoregulatory behaviour in Nephila spp., as may be expected theoretically using a ACKNOWLEDGEMENTS simple physical model, no such size effect was observed in Stegodyphus lineatus Latreille It is a pleasure to acknowledge the occasional (HenscheU/ a/., 1992). assistance of Rae Young in the field and the While the thermal behaviour of spiders is much Young family for allowing me to work on their more sophisticated than has been accepted, the property 'Mandalay'. I thank Mark Elgar and presumed advantages of such fine tuning are not Yael Lubin for their many constructive com- understood. None the less, the recognition of such ments. Some of this work was undertaken during behaviour is an important aspect of field studies the tenure of a Christensen Research Institute and the means to do so are required, especially for Fellowship and with the approval of the Madang smaller spiders which are intractable subjects for Provincial Council; the advice and assistance of direct recording of temperature in the field. How- Matthew Jebb and his staff is greatly appreciated. ever, the sensible use of this schema should allow The work was funded by the Australian Research easy appraisal of the overt body positions of Grants Scheme (#D181/15274) and the Austral- spiders in the field as to their likely thermoreg- ian Research Council (#A18831977 and ulatory significance and it should assist in disen- #A 18932024). » .

sv. MEMOIRS OF THE QUEENSLAND MUSEUM

LITERATURE CITED Western Australia. Behavioural Ecology and Sociobiology 28: 47-54. Stamlimenta as parasols: shade construction ALEXANDER, A.J. & EWER, D.W 1958. Tempera- 1992. lure adaptive behaviour in the scorpion. Opis- by Neo$ea sp. (Araneae: Araneidae, Argiopinae) ihopkthaimus tatimanus Koch. Journal of and its thermal behaviour. Bulletin of the British Experimental Biology 35; 349-359 Arachnological Society 9: 47-52 Y.D. 1986. building and prey capture in BIERE. J.M. & UETZ, G.W. 1 981 Web orientation in LUBIN, Web the spicier Micrathena graciti.s (Araneae: Uloboridae. Pp. 132-171. In Shear, *W. A. (Ed). Arane>dae). Ecology 62: 336-344. Spiders: webs, behavior, and evolution. (Stanford CARREL. IE 1978. Behavioural thermoregulation University Press: Stanford, California), during winter in an orb-weaving spider LUBIN. Y.D. & HENSCHEL. J.R. 1990 Foraging at possum of the Zoological Society of London 42: the thermal limit: burrowing spider {Seoihyru, 41-50. Eresidae) in the Namib Desert dunes. Oecologia CRAIG. C.Lv & BERNARD, G, D. 1990. Insect attrac- 84:461-467, tion to ultraviolet-reflecung spider webs and web MARSON, J. 1947 Some obscivations on the ecologi-

decorations. Ecology 71 : 616-623. cal variation and development of Ihe cruciate zig- EBERHARD, W.G. 1973. Slabilimcnta on the webs of zag camouflage device of Ar$lop6 pufchem Uloborus diversus (Araneae: Uloboridae) and (Thor.), Proceedings of the Zoological Society of other spiders. Journal of Zoology, London 171: London 117:219-227 367-384. MONTEJTH. J.L. & UNSWORTH, M.H. 1990. Prin- EDMUNDS, J. & EDMUNDS, M. 1986. The defensive C ipit-s of environmental physics. (Edward Arnold. mechanisms of orb weavers (Araneae; Araneidae London). in Ghana. West Africa. Pp. 73-89. In Eberharu. OLIVE. C.W. 1980. Foraging specializations in orb- Robinson, W.G., Lubin, Y.D. and B.C. [Eds) weaving spiders. Ecology 61: 1 133-1144. Proceedings of the Ninth International Congress RIECHERT. S.E. & TRACY, CR. 1975 TV of Arachnology, Panama, 1983. (Smithsonian In- balance and prey availability : Bases fox a model stitution Press: Washington, D.C.K relating web-site characteristics to spider EWER, R.F 1 972. The devices in the web of the- West reproductive success. Ecology 56: 265-284. African spider Argiope flavipulpis. Journal of ROBINSON, M.H. & ROBINSON, B. 1973. The Natural History 6: 159-167. stabilimenta of Nephila ciaVlpes and the origin of R-F. 1982. Biology of Spiders. (Harvard POEUX, stabiliinentum-building in Araneids. Psyche 80: University Press: Cambridge. Massachusetts). 277-288. HEATWOLE, H. 1970. Thermal ecology of Ihe Desert 1974. Adaptive complexity: the thermoregulatory Dragon Amphibolurus inermis. Ecological postures of the golden-web spider Nephila Monographs 40: 425-457. ctavipt's at low altitudes. American Midland HENSCHEL, J.R.. WARD, D. & LUB1N. Y. 1992. The Naturalist 92: 386-396, importance of thermal factors for nest-site selec- 1978. Thermoregulation in orb-weaving spiders: tion, web construclion and behaviour of St4tgo* new descriptions of thermoregulatory postures dyphus lineatus (Araneae: Ercsidac) in the Ncgev and experiments on ihe effects of coloration. Desert. Journal of thermal Biologv 17: 97-106. Zoological Journal of the Linncan Society, Lon- HUMPHREYS, W.F 1974. Behavioural ihcrrnoirguiu don 64: 87-102. lion in a wolf spider. Nature. London 25 1 Behavioural 503. SUTER. R.B. 1981. thermoregulation: solar orientation in Frontinellu communis 1 978. The thermal biology of Geolyrosa godfffnq p iLinyphiidae), a spider. and other burrow inhabiting Lycosidae (Araneae) 6-mg Behavioural Ecol- ogy and Sociobiology 8: 77-K1 in Australia. Oecologia, Berlin 31: 319-347. 1986. Heal shunting in spiders. Pp. 41-46. In Con- TOLBERT, W.W. 1979. Thermal stress of the orb- weaving spider Argiope irifasciuta (Araneae). oxesso Intemacionale Arachnologia. vol. I. (Ed. J.A. Bairiemos). Jaca. Spain. Oikos 32: 386-392. 1987a, The thermal biology of the wolf spider WARD. D &. HENSCHEL, J.R. 1992. Experimental Lycout torentula (Araneae: Lycosidae) in north- evidence thai a desert parasitoid keeps its host ern Greece. Bulletin of ihe British Arachnologi- cool. Ethology 92:135-142. cal Society .73 117-122. WILLMER, P. 1991. Thermal biology and mate ac- 1987b. Behavioural temperature regulation. Pp. 56- quisition in eclotherms. Trends in Ecology and 65, In. Nentwig, W fed). 'Ecopbysioloey of Evolution 6: 396-399. Spiders*. (Springer-\''erjyg: H-.ilm.) WILLMER. PC, ft UNWIN, DM. 19X1 Field 1991. Thermal behaviour of n small spider analyses of insect heat budgets: reflectance, size (Araneae: Araneidae: Arancinuc) on spinifcx in and "healing rates. Oecologia 50: 250-255. HYPERTROPHY OF MALE GENITALIA IN SOUTH AMERICAN AND AUSTRALIAN

TRIAENONYCHIDAE < ARACHNIDA: OPILIONES: LAN1ATORES)

GLENN S. HUNT AND EMILIO A. MAURY

Hunt. G.S. and Maury, E A. 1993 U 1 1: Hypertrophy of male genitalia in South American and Australian Triaenonychidae (Arachnida: OptJiones. Lamatores). Memoirs of the Queensland Museum 33(2): 551-556. Brisbane. ISSN 0079-8835.

Hypertrophy of male genitahc elements, particularly the stylus, is described and discussed.

A stylus is regarded as hypertrophied if stylus length is sub-equ3l to or longer than truncus length. Greatest hypertrophy occurs in the Australian species Chmiella distincta Forster

(stylus x4.5 truncus) and a new genus, new species from South America ( \2.5 truncus). Other species discussed are Anutcanobunus jubenhiei Mufioz-Cuevas from South America, and C. mmwa Forster. RHynchobunus arrogans Hickman. Tasmanobunus parvus Hickman,

loxnumonuncia ip. T Allobunus distinCtUS Hickman and 71\elbunus tmrabiUs Hickman from Australia. To accommodate an elongate stylus, the truncus is often shortened, and the genital operculum and sternum modified so that the genital orifice is located more anteriad. Hypertrophy of the stylus may be associated with the hypertrophy or reduction of other terminal elements. In Cluniella spp. and the two South American species penetration of the Stylus occurs along a very long vagina; the spermathecae are situated at the base of the ovipositor. Hypertrophy may have evolved as aconsequence of sexual selection. (jOpiliones, Triaenonychidae, male genitalia, hypertrophy, morphology, sexual selection.

Glenn S. Hunt, Division ofInvertebrate Zoology, A us/ralian Museum, P. O. Box A2S5, Sydney SoUth> N*W South Wate&t 2000. Australia, EtmBo A. Maun, Museo Ar^entino de Ciencias Niitumies 'Bernardino Rivaduvw \ Casilia dt Corrco 220 StiCurS&l 5, 1405 liuenos Aires. Argentina; 3 November, 1992.

The triacnonychid penis comprises a basal trun- in the ovipositor (Fig. 5c), unlike the usual con- cus supporting an apical complex which includes dition where the spermathecae occur sub-apical- the stylus and associated plates, processes and ly. Therefore, the long stylus probably penetrates setae. In the primitive condition, three sets of almost the whole length of the ovipositor to teach plates are present (Fig. 1 A): the dorsolateral plates the spermathecae. The dorsolateral plates of the which are embryologically derived from the trun- penis are elongate, gradually tapering to x2 cus and the dorsal plate and ventral plate ventral plate length (Fig. 2a). The dorsal plate is embryologically related to the stylus (Martens. either lacking or intimately fused with the stylus; 1986). Both the dorsal and ventral plates were the latter is suggested by the subterminal lateral apparently primitively paired but are now fused, processes on the stylus which may be at least basally. The ventral plate carries setae. homologous to terminations of the dorsal plate Certain taxa have undergone loss or reduction of (see Thetbunus mirabilis below. Fig. 5A). The plates (Martens. 1 986; Hunt and Hickman. 1993) ventral plate is reduced in size, and the number of A few taxa have undergone extreme hypertrophy inferior setae is reduced from three to two pairs. in the length of the stylus with one or more of the The extreme stylus is accommodated within the associated plates frequently showing correlated body by shortening of the truncus, and by elon- hypertrophy, or reduction, depending on the gation of the genital operculum and posterior taxon. invagination of the sternum which together shift the genital opening anteriad. The sternum mar- HYPERTROPHIED STRUCTURES gins tend to follow the genital operculum (Fig

5E; cf. female genitosternal region, Fig. 5f) but LUASW. when the operculum is lifted the shape resembles that in Fig. 5H. The most extreme hypertrophy of the stylus known for the family occurs in Cluniella distinau Ciunttilaminuta Forster, 1955, which overlaps Forster, 1955 of SE Queensland and NE New in distribution with C distincta, has undergone South Wales (Fig. 2a). The stylus is x4.5 truncus less radical elongation of the stylus (Figs 2C-D) length. There is correlated morphological change Nevertheless, the stylus is xl.4 truncus length in the female where spermathecae occur basally The spermathecae are also basal despite the 552 MEMOIRS OP THE QUEENSLAND MUSEUM

tion in the truncus where penis sheath membranes attach is basal. The differences between the two species are not due to infraspecific variation, such as male dimorphism reported in some Australian genera (Hunt, 1985). Characters distinguishing C mtnuta include a very long calcaneus compared

with astragalus in leg 1.

atowcanobunusjubertwei munoz-cuevas, and New Genus, New Species (Maury, in prep.) A. juberthiei from southern South America shows moderate hypertrophy of the stylus with

stylus length about 1 .3 truncus length (Figs 3a-B). An undesenbed species referrable to a new genus (Maury, in prep.), but apparently with close genitalic affinities to A. juberthiei, has undergone greater hypertrophy with stylus length x2.5 truncus length (Fig. 3c). The dorsolateral plates avu present in both species, but the dorsal plate is apparently lost (or intimately united with the stylus). The truncus in the undescribed species is greatly shortened, in A, juberthiei much less so. Both species have an elongate genital operculum and a short, subtriangular sternum (for example, see Figs 5(H). When closed, the genital operculum virtually obliterates the sternum. Unlike the condition in Cluntella, the constriction is situated mid-way along the truncus in both species (Figs 3a-D). As in Cluniella, the spermathecae are situated at the base of the ovipositor at the end of a long vagina (Fig. 5l>: A. juberthiei). There appear to be no significant infraspecific variations in size or shape of the genitalia in these species.

ftffYNCHOBUNVSARROGAKS HlCKMAN This species, from NE Tasmania, also has a prominent constriction at about mid-way along the truncus (Fig. 4a). R. arrogans belongs to a FIG- 1. Variation in 6 genitalia of some Australian lineage distinguished by loss of the dorsolateral Triaenonychidae. A: Hickmanoxyomma eavaiieum\ plates. The dorsal plate envelopes the stylus B: Glypiobunus $tgnatUS\ C: Tasmutumyx montamis to it. R. D: Dt'pristes serripus (muscle shown); E: Lomanella basally and terminates ventral arrogans raniceps\ F: Clttniella dislincfa (muscle shown); c = is very close to, if not congeneric with, Glyp- constriction for membrane attachment; dip =dor- tobunus signatus Roewer (Hunt, in prep.). The solateral plate, dp = dorsal plate; is =• inferior seta; s stylus is xl.3 truncus length. Trie spermathecae

=stylus; ss =supcrior seta; t =truncus; vp =ventral are subterminal. The ventral plate is reduced but plate. still carries the normal complement of setae for this lineage (Figs 4A-B). R. arrogans has an elon- shorter stylus. The dorsolateral plates arc elongate genital operculum but lacks a triangular ster- gated to a similar extent to C. disdncta while the num. ventral plate, though reduced, bears 3 pairs of inferior setae. The truncus is less shortened, the TASMANO&l was parvus Hickman genital operculum is not elongate and the sternum This species is closely allied to, or possibly is long and narrow. In both species, the constric- congeneric with, R. arrogans and G. signatus as HYPERTROPHY OF MALE GENITALIA 553

it lacks dorsolateral plates, its stylus carries a subdistal dorsal barb, and the dorsal plate terminates ventral to the stylus (Figs 4C-D). The stylus is subequal in length to the truncus and the dorsal plate has undergone complementary elongation. The constriction around the truncus is more subdued and situated somewhat more basal- iy-

Tasmanonuncia n.sp. This species belongs to the same general

lineage as R. arrogans in that it lacks dorsolateral plates (Hunt, in prep.). However, the stylus carries a subdistal dorsal tuft of hairs instead of a single barb (Figs 4E-F). Stylus length is subequal to truncus length and, as with Tasmanobunus parvus, the dorsal plate is greatly elongate.

AlLOBUNUS DISTINCTLY HlCKMAN A. distincms, from NW Tasmania, has a prominent constriction mid-way along the truncus. The stylus is modestly elongate, being subequal in length to the truncus. The ventral plate, however, is hypertrophied, uniting with other sclerotisations to form an elongate cylinder around the proximal half of the stylus (Figs 4G-H). The male ofA. distinctus has a triangular sternum and a moderately elongate genital operculum.

Thelbunusmirarius Hickman Unlike the Rhynchobunus lineage, this NE Tas- manian species belongs to a lineage with well developed dorsolateral plates. The stylus is hypertrophied but the problem of space is partly solved by the stylus being twice folded back on itself (Fig. 5a). The dorsal plate is also elongate and closely integrated with the stylus, bifurcating into two lobes just below the stylus tip. The ventral plate setae are greatly hypertrophied and differentiated into different forms (Fig. 5b). This species also has a short triangular sternum (though not unlike the female) and an elongate genital operculum (Figs 5J-K). Thelbunus n.sp. has a less elaborate stylus and

its sternum less modified and genital operculum less elongate (sternum illustrated in Fig. 6).

Other Opiuones Hypertrophy of the stylus has also been recorded in the neopilionid Ballarra spp (suborder Palpatores) from Australia (Hunt and Coken-

FIG. 2. Hypertrophied 6 genitalia of Cluniella spp. A,

B: C. disfincta , A = lateral, B = ventral view of ventral

and dorsolateral plates. C t D: C. minuta , lateral and ventral, 554 MEMOIRS OF THE QUEENSLAND MUSEUM

FIG. 3. Hypertrophic!! 6 genitalia of Soulh American Triaenonychidae. A, B: Araucanobunus juberthiei, lateral and ventral. C, D: New genus, new species (Maury, in prep.)* lateral, ventral (pan). dolpher, 1991). Unlike the situation in Cluniella spp. and the South American species, the spermathecae are greatly elongate, opening into a short vagina but ending basally in the ovipositor. The long stylus is presumably inserted down the elongate spermatheca, not an elongate vagina as occurs in Cluniella. Other taxa exhibit elongation of the truncus, e.g., species of the triaenonychid genus many FIG. 4. Hypertrophied 6 genitalia in Tasmaman Tri- Lomanella (Hunt and Hickman, 1993). aenonychidae. A, B: Rhynchobunus arrogans , lateral

and ventral. C, D' Tasmanobunus parvus , lateral and DISCUSSION ventral. E, F: Tasmatwnunda n.sp. (Hunt, in prep.). lateral and ventral. G. H: Allobunus distinctus, lateral and ventral An elongate stylus is frequently, but not invariably, associated with an elongation of the seem initially accommodated by shortening of genital operculum and a shortening and broaden- the truncus (C. minuta condition). Changes to the ing of the sternum which together serve to elon- genital operculum and sternum evolved later. gate the cavity in which the penis lies (also, Cluniella spp. and the two South American shortening of the truncus may compensate for the species described above show close correspon- elongate stylus). dence in many features associated with stylus These changes in genitosternal architecture hypertrophy: great elongation of stylus, seem to have arisen independently in several modification of sternum and genital operculum, distantly related gerera. Within the one genus, shortening of truncus, and spermathecae situated both normal and hypertrophy-related genitoster- basally in the ovipositor at the end of a very long nal architecture can occur. For example, Cluniel- vagina- la minuta has evolved a moderately elongate The question is whether these features indicate stylus while retaining a 'normal' long, narrow a close phylogenetic relationship or whether they sternum and rounded genital operculum. C. dis- are examples of convergence. The latter is sup- tincia has undergone much greater elongation of ported because: the stylus with consequent changes in the sternum 1. Modifications to genitosternal architecture and genital operculum. appear to have arisen independently within the It seems more probable that the situation in C Cluniella lineage, evolving from the 'normal' minuta is more 'primitive', while the syndrome condition as occurs in C. minuta. of extreme characters in C. distincta represents a 2. Attachment of penis sheaths is basal in more derived condition. Thus, in the Cluniella Cluniella and mesial in South American species. lineage, space limitations of an elongate stylus 3. The penes of the South American species HYPERTROPHY OF MALE GENITALIA 555

tion by inflating the ovipositor so that it partly

' stylus. assuming that ' ' engulfs the Thus, the male C. minuta has not undergone a reversal in genitosternal architecture, the derived ovipositor seems to have evolved before shortening of the male sternum and elongation of the genital operculum. Thus, the genomes of Cluniella and the two South American taxa have the capacity lo allow quite remarkable convergence in a syndrome of characters. The overall effect appears the same. but the details differ. Why have such vastly elongate styluses evolved, particularly to the extreme shown by C. distincta? Sexual selection by female choice is

favoured by Eberhard ( 1 985) as the most generally applicable explanation for 'extravagant* genitalia. Eberhard proposes that 'male genitalia function as 'internal courtship', devices to increase the likelihood that females will actually use a given male's sperm to fertilize her eggs

rather than those of another male' . In the case of Cluniella and the South American species, it is postulated that males with the largest styluses have greater success than males with smaller styluses and hence contribute more of their genes to the next generation. The genitalia may stimu-

FIG. 5. Hypertrophy of genitalia in Thelbunus late the female prior to or during copulation and mirabilis and modified structures associated with so activate the appropriate responses, or it might genitalia hypertrophy in various Triaenonychidac. A, provide the right mechanical and sensory fit" B: T, mirabilis, lateral and detail of lateral showing during copulation. The basal spermathecae in the hypertrophy and modification of ventral plate setae. ovipositor of Cluniella spp. and the two South C,D: Ovipositor in Cluniella distincta and American species (and the long stylus matching Araucanobunus juberthiei respectively showing the long spermathecae in Ballarra spp.) suggest basal seminal receptacles. E, F: Genitosternal region that the correct mechanical fit may at least be part of C distincta, 6 and 9 respectively. G, I: of the answer. Genitosternal region of South American new genus. ncwspccics(Maury,inprcp.). & and 9 respectively; A search for congeners and an analysis of inter- H = shape of sternum in & after genital operculum and intraspecific variation, as well as behavioural removed. J, K; genitosternal region of 71 mirabiHs, 6 studies, may yield further data to resolve these and 5 respectively: go = genital operculum; sp = questions spermatheea; st = sternum; v = vagina (not delineated in C distincta). ACKNOWLEDGEMENTS show closer affinity with the South American genus Thaenonyehoides (see Maury, 1987)) The first author was assisted by an Australian rather than with Australian genera. Biological Resources Study grant. Drs Michael 4. Apart from the basal spermathecae and long Gray (Australian Museum) and A Roig-Alsina vagina, the ovipositor of the South American (Museo Argentino de Ciencias Naturales), and species appears to be of the typical triacnonychid Mr James Cokendolpher (Lubbock. Texas), com- form. The. ovipositor of both Cluniella spp. is mented on the manuscript. Mr Roger illustrations Sue highly derived in having a very membranous tip, Springthorpe did the and Ms in lacking well developed sensory lobes, and in Lindsay the SEMs. carrying vestigia! setae. In some specimens examined the membranous tip was inflated and had LITERATURE CITED ballooned* out the genital orifice. This morphology suggests that the female may assist penetra- EBERHARD. W G I98S. Sexual Selection and 556 MEMOIRS OF THE QUEENSLAND MUSEUM

*&4

FIG. 6. Genitosternal region in Thelbunus spp. and variation in sternum; A = sp. nov.; B T. mirabilis. Genital operculum lifted; note recess in sternum where folded stylus fits. Scale bars= 50Gu.m.

Animal Genitalia'. (Harvard University Press: (Arachnida: Opiliones: Triaenonychidae). Cambridge, Mass). 244pp. Records of the Australian Museum 45:81-1 19. HUNT, G.S. 1985. Taxonomy and distribution ofEqui- MARTENS, J. 1986. Die Grossgltederung der tius in eastern Australia (Opiliones: Laniatores: Opiliones und die Evolution der Ordnung (Arach- Triaenonychidae). Records of the Australian nida). Actas X Congreso Internacional de Arac- nologia, Jaca, Espana 1: 289-310. Museum 36: 107-125. MAURY, E.A. 1987. Triaenonychidae Sudamericanos. HUNT, G.S. & COKENDOLPHER, J.C. 1991. Ballar- IV. El genero Triaetionychoides H. Soarcs 1968 rinae, a new subfamily of harvestmen from the (Opiliones, Laniatores). Bolctin de la Sociedad dc Southern Hemisphere (Arachnida, Opiliones, Biologfa de Conception, Chile 58: 95-106. Ncopilionidae). Records of the Australian MUNOZ-CUEVAS, A. 1973. Descripcion dc Museum 43: 131-169. Araucanobumtsjuberthiei gen. et sp. nov. de Tri- HUNT, G.S. & HICKMAN, J.L. 1993. A revision of the aenobunini de Chile (Arachnida, Opiliones, Tri- genus Lonumelia Pocock and its implications for aenonychidae). Physis, Buenos Aires, Secc. C family level classification in the Travunioidea 32(84): 173-179. s

PREDATOR-PREY CO-EVOLUTION OF PORTIA FMBR1ATA AND EURYATWSSP., JUMPING SPIDERS FROM QUEENSLAND

ROBERT R. JACKSON and R. STiMSON WILCOX

Jackson, R.R. and Wilcox, R.S. 1993 1 1 11: Predator-prey co-evohKiofi of Portia fimhriata and Euryattus sp., jumping spiders from Queensland. Memoirs of the Queensland Museum 3312 V 5*57-560. Brisbane. ISSN 0079-8835.

Portia is a salticid ihat preys on other spiders and Euryattus sp is a saliicid thai nesls inside suspended rollcd-up leaves. Portia and Euryatius arc sympatric at a she near Cairns but not known to be sympatric at other sites studied. Portia from the Cairns site practices a unique prey-specific predatory behaviour against Eutyaltus, and Euryattus from this site is efficient at detecting and defending itself against Portia. Euryattus, but not Portia, is present at a site near Davies Creek which, although only ca 15km from the Cairns site, is more xeric and at a higher elevation. Three types of tests were carried out to compare Portia's efficiency at

eaiching adult allopatrie versus sympatric Euryattus (Test I ). allopatrie Euryatius juveniles versus juveniles of another salticid species on which Portia is known to prey (Test 2) and allopatrie versus sympatric Euryattus juveniles (Test 3). In these tests, Portia behaved

similarly toward allopatrie (Davies Creek) and sympatric (Cairns) Euryattus, except that it ailaeked and killed allopatrie more often than sympatric Euryattus. Allopatrie Euryattus in rast to Cairns Euryartuv. appeared not to recognize an approaching Portia as a pred&tOt.OPortia fimhriata, Euryattus, Jacksonoides, co-evolution, allopatry, symputrv

Robert R Jackson, Depot imrnt uj Zoology, University of Canterbury \ Chrtstchurch J, New Zealand; R Stimson Wilcox, Department of Biological Sciences, State University of New York at Binghamton, Binghamton, New York J 3902-6000, U.S.A.; 30 November, 1992.

Portia is a genus of'specialized jumping spiders a time for Euryattus to come out of its nest (J.k If (Salticidae) that prey on other spiders (Jackson son and Wilcox, 1990). Often, Euryattus actively and Hallas, 1986). Portia is a detritus mimic and defends itself by leaping at Portia and driving it has a unique, slow, choppy style of locomotion away (Jackson and Wilcox, 1990). This is un- that seems to preserve its crypsis. There are seven usual behaviour for a sallirid. From thousand* of described species of Portia, distributed in the observations of interactions between P. fimbriate tropics of Africa, Asia, and Australasia (Wan and many different species of sakkids (Jackscm 1978). A population of Portia fimbriate (Doles- and Hallas, 1986). it is evident ifiac Euryaiius is chal!) in Queensland uses specialized behaviour more efficient than other salticids at recognizing to catch other species of salticids (Jackson and and defending itself against an approaching Par- Blest. 1982). This population of P. fimbriate also uses a prey-specific predatory behaviour against tie. Also, in laboratory tests (Jackson and Wilcox. females of a particular sympatric salticid, £//rya?- \990), Euryottus; readily recognized an approach- lus sp. (Jackson and Wilcox, 1990). ing Portia as a potential predator, whereas. Jack- sonoides queenslandica. another salticid on Euryatius females suspend a dead, rolled-up which P. fimbriate feeds (Jackson and Blest, leaf by strong guylines from rock ledges and tree 1982), did not recognize P. fimbriate. This sug- trunks, then use the leaf as a nest (Jackson. 1985). gests an evolutionary arms race' ( sensu Dawkins Portia has never been observed to attempt to catch Euryaftusby going inside the rolled-up leaf. and Krcbs, 1979) between Euryattus and P. However, in Queensland, P. fimbriate uses fimhriata. Frequent prcdation by P. fimhriata on vibratory displays to lure Euryatius females from Euryattus may have favoured special abilities in their nests (Jackson and Wilcox, 1990). These Euryattus to recognize and defend itself against displays apparently mimic courtship displays of P fimbriate. This, in turn, may have resulted in Euryattus males (Wilcox and Jackson, unpubl. the evolution of refinements of P. fimbriate* data). Other species of Portia and populations of predatory behaviour To test this hypothesis, we P. fimbriate in areas from which Euryattus is must compare the behaviour of Euryattus in absent do not perform these displays (Jackson and populations with and without Portia. Recently, Wilcox- 1990). such an opportunity arose when Euryattus were Queensland P, fimbriate will wait for hours at found in an area in which Portia was not known. MEMOIRS OF THE QUEENSLAND MUSEUM

MATERIALS AND METHODS type of tests using sympatric Euryattus adults in an earlier study (Jackson and Wilcox, 1990) to see whether Portia's capture efficiency against Cages, maintenance, terminology, basic testing allopatric Euryattus adults was greater than procedures and analysis arc given in Jackson and against sympatric Euryattus adults. Test 2 using Wilcox (1990). Laboratojy cultures of sympatnc allopatric Euryattus juveniles was compared to tu/yattus* J. queenslandica and P. fitnbriata type 2 tests in Jackson and Wilcox (1990) using were established, using spiders collected from sympatric Euryattus juveniles and J. queenslan- rainforest near Cairns at aboul sea level (see dica juveniles. We already know that Portia cap- Jackson, 1985; Jackson and Hallas, 1986). A tures J. queenslandica juveniles more efficiently laboratory culture of allopatric Euryattus was than it captures sympatric Euryattus juveniles established from spiders collected in an Acacia- (Jackson and Wilcox, 1990). Here we examine Eucalyptus woodland beside Davies Creek, near whether Portia's capture efficiencies against Davies Creek National Park in the Alherton these two salticids vary when Euryattus- a %l- Tableland (about 15km from the study site near lopatnc. Test 3 enabled us to compare Portia's Cairns and at c. 500m elevation). Portia has never efficiency at capturing allopatric and sympatric In en recorded from this and other Atherton Euryattus u ven tie s Tableland habitats (Wanlcss, 1978; Jackson, un- j publ. data). Unless noted otherwise, all spiders Adult body length is c.8mm for both J [£&ied were reared in the laboratory from eggs of queenslandica and P. fimbriate and for both populations of Euryattus. Jackson and Wilcox i collected spiders. No individual spiders

1 , ratio were used in more than one test. In this paper, we 1 990) used three sue classes defined by the /' refer to Euryattus frorr} Cairns and Davies Creek of prey to predator body volume, when testing as "syrnpairic Euryattus* and "allopatric fimbriate with juvenile salticids; small (0.1- Eitryattvft respectively. There were no evident 0.25). medium (0.5-i), and large (1.5-2). Only differences related to general behaviour between two of these (medium and large) were used here. these two populations ofEuryattus. In particular, MeNeinai tests fi«r significance of changes similar leaves were suspended by were used for statistical analyses of the results nests and males courted with similar vibratory' from Tests 2 and 3, these tests being designed as displays. paired comparisons (Sokal and Rohlf, 1981):

The systematica of the genus Eu i j utUiS remains each individual Portia was used in one test with uncertain Whether the two populations of one salticid and another test with the other saltieid Euryattus we studied are one or two different 48 h earlier or later (decided randomly) Yates* species is not now known. Voucher specimens corrections were applied to the McNemar tests, were deposited at the Honda Collection of and the Bonferroni adjustment (sec Rice, 1989) Atthropods (Gainesville) and the Queensland was made to significance levels whenever single Museum. data sets were used in multiple comparisons.

We conducted three tests. In Test 1 , Portia was given access to an adult allopatric Euryattus RESULTS female in her nest. In Test 2, on alternate days, Portia was given access to a juvenile (2-3mm in body length) allopatric Euryattus and a juvenile

I in (2-3mm>/. queenstundica in a bare cage fi e.no Trsr BVWATTVS Aon it Nfst nest or other objects present). In Test 3, on alter- P. Jhnbriata behaved similarly toward ul nate days, Portia had access to a juvenile (2- lopatric (herein) and sympatric (Jackson and Wil-

3mm) of an allopatric and a juvenile of a cox, 1990) Euryattus, except that it attacked and sy mpatric Euryattus in a bare cage. To begin each killed allopatric Euryattus more frequently than test, Portiu type of was placed into a cage con- sympatric Euryattus {Fig. 1 , test of independence, taining the other spider shortly after lights came P<0.01). Allopatric Euryattus appeared less on in the laboratory (0800 hours). Spiders were prone lhan sympatric Euryaf/us to recognize P. observed continuously until predation occurred fimbriate as a predator: 85% of the P. fitnbriata or until 4h had elapsed. Each test was either got onto the leaf with allopatric Euryattus. but identical or at least similar to tests carried out only 43% got onto the leaf with sympatric Ev • previously (Jackson and Wilcox, 1990). tus\ 23% of sympatric Euryattus, but only 4% of Data from Test 1 using allopatric Euryattus allopatric Euryattus, drove P. fhnhriata away adults were compared to data from the identical (Fig. 1). —

PREDATOR-PREY COEVOLUTION 559

U i'j

it h -.. 'j - an — 35 100 ^ 301 25 60 25

20 no 15

•-,:,. 10 8 5 ?n o

I I I 1 I P twfsu»'i P pursued P cap'urwj P captured P captures

:n.;,. i I I 1, finrt J nn'v k nnlv and f ia\U P nnlul.,,1: gnd EF

' ! .- .i;.,vi. '(! ,- .,,...,! ., n „,-.,, fir' F .ia.i, . r .-.l-v. .,,!.„ „,.,

C frevei gof close guyline or onto lea' h IIh FIG. 3. Responses of P. fimbria ta (P) tested with large (see text) allopatnc Euryaltus (E) and sympatric J. FIG. 1. P. fimbriate* (?) lested with adult allopatnc queenslandica (J), 25 paired tests: each Portia tested (Davies Creek) and sympatric (Crystal Cascades) with one Euryattus and, on an alternate day, with one Euryattus females (E) in suspended, rollcd-up leaves. Jacksonoides (see text). Data lor *P pursued neither J T Data tor synipatric Euryattus from Jackson and Wil- norE and *P captured neither J norE' not displayed. cox (1990). Close: on leaf or guyline connected to leaf, or dropping on dragline toward leaf. For each outcome of lest, number given above bar and percent- Euryattus more often that it caught sympatric age is read from axis. Euryattus (Figs 4, 5).

Test 2: }\jveu\le Euryattus and Jacksonoimss DISCUSSION queenslandica There was no evidence that Portia captured or Only one population of Portia fimbriata from stalked J. queenslandica more frequently than Cairns studied (Jackson and Wilcox, 1990) is allopatric Euryattus (Figs 2, 3 T McNemar tests, sympatric with Euryattus. Euryattus suspends a NS). Allopatnc Euryattus did not appear to rollcd-up leaf for a nest, and this is the only recognize Portia as a predator any more readily salticid sympatric with the Cairns Portia, or with than did J. queenslandica. any other Portia studied, known to do this. Only the Cairns Portia is known to use a prey-specific Test 3: Cairns and Davies Creek Euryattus predatory behaviour against Euryattus. The sym- Juveniles patric (Jackson and Wilcox, 1990), but not the

There was no evidence that Portia stalked sym- allopatric , Euryatius appears readily to recognize patric Euryattus any more frequently lhan al- and defend itself against Portia. In fact, the allopatric Euryattus, but Portia caught allopatric io o

jo 12 80

)'i

201

_r_;_ n !."! i.l Q ... ,,,! ,-i'i £ putaued l r.ifii i ' tWlUfWl P rjwturft] -i A only end A s only a w* s h i " ... i ued p pui^ued P Mtwen P capture! P captuivi! p nil red

i und E J only E only h -I !" FIG. 4. Responses of P. fimbriata (?) to medium size FIG. 2. Responses of P. fimbriate (?) to medium size (see text) sympatric (S) and allopatric (A) Euryattus

(see text) allopatric Euryattus (E) and sympatric J. sp. juveniles. 38 paired tests: each Portia tested with queenslandica (J). 40 paired tests: each Portia lested one sympatric and, on an alternate day, with one with one Euryattus and, on an alternate day, with one allopatric Euryattus (see text). Data for 'P pursued Jacksonoides (sec text). Data for *P pursued neither J neither S nor A' and *P captured neither S nor A* not

nor E' and 'P captured neither J norE* not displayed. displayed. 560 MEMOIRS OF THE QUEENSLAND MUSEUM

35 cal reading of the manuscript. Financial support 30 was provided by grants from the National Geographic Society (3226-85) and the U.S.-Ncw Z5 Zealand Cooperative Program of the National 20 Science Foundation (NSF Grant BNS 8617078). 151

ID LITERATURE CITED

5 DAWKLNS, R. & KREBS, J.R. 1979. Arms races be- d tween and within species. Proceedings of the

irsue P pursued ::: i - i[iiliif'.j capUnfl Royal Society, London B 205: 489-5 11. .:.!', A cn!v nil A S uuly A unly 5 iirfl A JACKSON, R.R. 1985. The biology of Euryattus sp. indet., jumping spider (Araneae. FIG. 5. Responses of P. fimbriate (?) to large (see a web-building Salticidae) Queensland: utilization of silk, text) sympatric (S) and allopatric (A) Euryattus sp. from tntraspecific interac- juveniles. 28 paired tests: each Portia tested with one predatory behaviour and 145-173. sympatric and, on an alternate day, with one allopatric tions. Journal of Zoology. London (B) 1: Euryattus (see text)* Data for *P pursued neither S nor JACKSON, R.R, & BLEST, A.D. 1982. The biology of A' and *P captured neither S nor A' not displayed. Portia fimbriate, a web-building jumping spider (Araneae, Salticidae) from Queensland: utilization of webs and predatory versatility. Journal of lopatric Euryattus appears to be no better than J. Zoology, London 196: 255-293. JACKSON, R.R. & HALLAS, S.E.A. 1986. Compara- queenslandica at escaping predation by Portia , tive biology of Portia africerui, P. albimana, P. whereas Portia captured J. queenslandica more fimbriate, P. labiate, and P. schultzu aranco- efficiently than it captured the sympatric Euryat- phagic, web-building jumping spiders (Araneae, tus (Jackson and Wilcox, 1990). The ability of the Salticidae): utilisation of webs, predatory ver- Cairns Euryattus appears to be a predator- satility, and intraspecific interactions. New specific antipredator behaviour. Zealand Journal of Zoology 13: 423-489. Population differences were evident despite JACKSON, R.R. & WILCOX, R.S. 1990. Aggressive there being no known prior experience of the mimicry, prey-specific predatory behaviour and predator by the prey or the prey by the predator predator-recognition in the predator-prey interac- under laboratory rearing conditions in this and the tions of Portia fimbria ta and Euryattus sp., jump- earlier (Jackson and Wilcox, 1990) study. These ing spiders from Queensland. Behavioral Ecology and Sociobiology 26: 1 1 1-1 19. findings suggest that, in the Cairns area, Portia RICE, W.R. 1989. Analyzing tables of statistical tests. and Euryattus appear to have acted as selective Evolution 43: 223-225. agents on theevolution of each other's behaviour. SOKAL, R.R. & ROHLF, FJ. 1981. 'Biometry'. (Freeman: San Francisco). ACKNOWLEDGEMENTS WANLESS, R.F. 1978. A revision of the spider genus Portia (Araneae: Salticidae). Bulletin of the We thank Curt Lively. Mark Elgar and British Museum (Natural History) Zoology 34: Marianne Willey for useful discussion and criti- 83-124. "WE'LL MEET AGAIN", AN EXPRESSION REMARKABLY APPLICABLE TO THE HISTORICAL BIOGEOGRAPHY OF AUSTRALIAN ZODARIIDAE (ARANEAE)

R. JOCQUE

Jocque, R. 1993 1111: "We'll meet again", an expression remarkably applicable to the historical biogeographv of Australian Zodariidae (Araneae). Memoirs of the Queensland Museum 33: 561-564- Brisbane ISSN 0079-8835.

The largest subfamily of the Zodariidae, (he Zodariinae, contains 3 I genera and has endemic representatives on all tropical continents The sequence in this elade first indicates that the Australian zodariid fauna is the result of a combination of vicariance and dispersal events Only three zodariid genera are not endemic to Australia. It is argued that two of these.

Mallinella and Asceua t have reached Australia by dispersal over forest-covered areas. It is remarkable that the endemic Australian genera in the Zodariinae are more closely related to the African ones than are the Neotropical genera which is in contrast with the current ideas on chronology in plate tectonics. A possible explanation might be found in the past and present distribution o\' forests. The appearance of this type oi vegetation on the Crelaeenus- Tertiary boundary, might also be invoked to explain bipolar distributions in Africa. La plus importantedes sous-families des Zodariidae, les Zodariinae. eminent 31 genres, doni des endemiqucs sur chaque continent tropical. La sequence dans cc grand cum lodiquC d'abord que la composition de la faune australienne serait Je resultat iTunc emnbin.n do vicariance el de dispersion. Seuls trois genres trouves en Auslralie ny sonl pas endemiqucs. On avance P argument selon Icquel deuv d'entre cuk ont rcussi a alteindrc 1'Ausiraln. par dispersion a travers des aires couvertes dc forets. 11 csi rcmarquable que les genres endemiques australiens soient plus proches des genres afrieains que de ccua d' Amerique du

Sudcequi eontred il 'antes enncernant la clironologie tie la derive de>. plaques. Une explication possible pourrait se trouverdans la distribution passeeet actuclle des forets £quatoriales. La genese de ce type de vegetation, a la fin du Crctaed, pourrait cgalement expliquer la distribution bipolaue de certaines vieillcs lignecs cTinvcrtcbrcs. vy^frica, Gondwanaland, distribution, vicariance.

Rudy Jocque, Koninktijk Museum root Midden-Afrika, B-3080 Teryuren, Belgium; 13 October. 1992.

The Zodariidae arc a medium-si7.e pantropicai Genus Distribution Mabitai family of mainly nocturnal, ground-living •.!„, ..,,,,.• '. i.ustrali i?ircuniforcst spiders. Except for the members of the subfamily Asteron Australia steppe -woudlflnd Storenomorphinae they are virtually all bur- Cyriocttti southern Gondwana dunes rowers to some degree. Some simply dive into Hiibroncslcs Australia steppe -woodland sand (Psammodtwn) or hide in litter (Asceua. ' hiirim South and Central America ciicumfofl i Mallinella) whereas others make a complex Matlinetla Piilucoiropitti pic;il fureM retreat consisting of a burrow with a trapdoor Neoslarena Australia steppe -woodland \Fsammorygma* Neasiorena) or an igloo-like Storeno Australia steppe -wwulland construction of pebbles or grains of sand (Diores, Starosa Australia steppe -woodland Zodariori). This sedentary lifestyle and the fact -circurr.foresl that zodariids do not balloon explains why most Tentdos South and Central America steppe | species have small distribution ranges. It also makes them an ideal subject for zoogeographical TABLE 1. Distribution and habital of some zodariid genera. studies, moreover since a cladistic analysis of the family is available (Jocque, 1991}. representatives of these genera on any other con-

tinent (Table 1). THE AUSTRALIAN ZODARIIDAE Three genera have representatives elsewhere and in fact have the centre of their distribution outside Australia. The first one is Cyrioetea* the Most Australian zodariids, estimated at several only genus in the Cyriocteninae, found exclusive hundred species, belong to genera endemic to the ly in sandy habitats such as arid dunes and sand Australian continent. There is no trace of any deserts. Il has a typical Gondwanan distribution —

MEMOIRS OFTHE QUEENSLAND MUSEUM

i— • — tropical forest organisms (e.g., Mallinella)

originating ill Africa, have been able to reach India via a northern itinerary. No forest connection has ever existed via the mediterranean and > ._•' die Arabian peninsula. This paradox is discussed below ->v Virtually all Zodariidac arc restricted to ' -t habitats lying in a climatic zone with a marked dry season. In the Neotropics only a few species - (in Ishania and Tenedos) seem to have adapted to moist forest and no ?xxiariids have so far been I Ar>proximate distribution of MalUnelhi. found in Amazonia. In Australia no true forest- inhabiting species belonging to endemic genera with representatives in South Africa, Chile aivd appear to be present, although some species in the Australia iCyrioctea raveni Platnick and Griffin) genera Neostorena, Asteron, Storosa and Both the other genera, Mallinella and Asceua, Habronestes, occur in dense woodland. But the have an enormous distribution which appears to most common and widely distributed forest in- be linked with old world forests. One species of habiting genus is doubtlessly Mallinella. This each has reached the northern tip of Australia. genus has an enormous distribution (Fig. 1): it The Australian zodariid fauna is thus apparent- occurs in virtually all African forests from ly composed of three different stocks: a strictly Senegal in the west to the Chimanimani Mts. endemic one, a southern Gondwanan one and a (Zimbabwe) (Jocque. unpublished data) in the third element thai could be Quoted ftfl old world south and Kenya and Ethiopia in the east, but is forest fauna. To understand this composition we remarkably absent from the South African forests have to go back to an era before the breakup of which emphasizes its inability to cam unforested Gondwanaland. zones. It is present in montane rainforest in India The Zodartidae indeed have 3 basic and Nepal and occurs in virtually all dense forests Gondwanan distribution with the more in the Far East, as far as New Guinea. One species plesiomorphic iaxa represented in Africa, has reached the northeastern tip of Australia. This Australia and South America. These are the is apparently a very recent event as that species, Cvriocteinac, the Lachesaninae and the more Mallinella zebra Tborelk occurs on both New plesiomorphic members of the Zodariinae. The Guinea and Australia. Considering the fractionat- latter taxon is now known to include what has ing of this genus in the rest of its distribution been described as the subfamily Storeninae range and the fact that no speciation has occurred (Jocque, 1991, 1992). Thirty-one out of 47 genera in this case, there is little doubt that the arrival now belong in the Zodariinae. However, the more occurred during the Pleistocene ice-ages, during apomorphic members of the subfamily (Femoral which there was a land-bridge linking New- Gland Clade or FGC), having several sy napomor- Guinea and Australia (Keast. 1983). A well docu- phies (femoral gland. Jack of cfulum, flattened mented case of a less recent Asian invader is that incised hairs, fused chelicerae) are present only of Tamapsis (Hersiliidae) which is supposed to in Africa (including part of the Palaearctic) and have arrived in northern Australia during the tropical Asia. The same applies to the Miocene (Baehr, 1988). This relatively early ar- Storenomorphinae and the Cydrelinae which are rival has resulted in clockwise colonization of restricted to tropical .Africa and tropical Asia. almost the entire circumference of the Australian continent and the appearance of many strongly When Africa got finally separated from other related species. major landmasses, slightly more than 100 niybp, these three groups were apparently not yet in A second recent zodariid arrival is that of an existence. The bulk of the zodariid fauna outside undescribed species of Asce.ua (Jocque, in that continent is therefore supposed to be derived preparation) which can be characterized as a cir- of the plesiomorphic iaxa present at the time of cum-forest genus. Its ecological status is fairly

the :--plit-i»ft. However, from the above il is clear similar to those genera mentioned above, which

that Indta < and part of South East Asia 7) carried occasionally occur in dense woodland or even

;» much more modem assemblage* of /V«iariidae rainforest in the case of the South American when it moved towards its present position. In- Ishania and Tenedos. The distribution of Asceua deed, there is no evidence that at least purely (Fig. 2) is even larger than that of Mallinella and ZODARIID BIOGEOGRAPHY 563

Nelson and Platnick, 1980). However, not only is there much controversy about the original position of many plates, the timing and the mechanics of the movement are also subject to a debate. The expansion theory of Carey (1975) and Shields (1979, 1983) positions several plates in different places than does the classical theory (Norton and Sclater, 1979; Owen, 1983). According to the

latest data (Veevers et al. y 1991; Scotese et al. y 1988; Powell et aL, 1988) it appears that India split off from Gondwanaland together with the FIG. 2. Approximate distribution of Asceua. southern continents Australia and Antarctica. How then can one explain the fact that India and the patterns overlap only partially. Although Sri Lanka share many apomorphic monophyletic several to in rain- Asian species are known occur taxa with Africa (e.g., Hermippus, Storenomor- forest (Bosmans and Hillyard, 1990; P. phinae, Cydrelinae, members of the FGC)? Schweninger pers. comm. 1989; C. Deeleman- Briggs (1987) thinks that India, on its course to Reinhold, pers. comm. 1988; J. Murphy, pers. Asia, became again attached to north-east Africa com. 1988) no representatives are found in by the end of the Cretaceous, a remarkable African lowland rainforest. On the other hand it hypothesis that might explain the apparent faunal is present in rather dry areas in southern Africa. anomaly. It would certainly explain how forest its very Considering restricted distribution area in dwelling spiders like Mallinella and Asceua have Australia, it is supposed to have arrived there at reached Asia. the same time as Mallinella. The main question that is raised by the zodariid Two closely related lines of zodariines which cladogram is why the Australian taxa are more had been separated for about a hundred million strongly related to the African than to the South years thus met again. American ones. Assuming that Mallinella was initially better A possible explanation is that the few typical adapted to rainforest conditions, it would be in- forest inhabiting zodariid genera did only evolve teresting to study if the arrival q{ Mallinella zebra after the main continents had broken away from has induced a shift in the rest of the zodariid fauna Africa. Broad-leaved forests arose at least during in the area it now occupies, and whether it is still the Cretaceous but became dominant in expanding. equatorial conditions on the Cretaceous-Tertiary boundary (Wolfe, 1987; Wolfe and Upchurch, GENERAL DISTRIBUTION OF THE 1986, 1987). It is strange that in the context of ZODARIINAE African biogeography no mention has been made of the dramatic shift that has been brought about The monophyletic assemblage of the by the appearance of broad-leaved evergreen forest. at level in this Zodariinae is largely composed of genera which The conditions ground kind are endemic to particular continents (Jocque of forest and the former conifer forests were 1991, fig. 41). Remarkably, the African genera probably as different as they are for instance in are apparently more closely related to the present-day rainforest and miombo in Africa. The vegetation very Australian Zodariinae than to those in South faunas of these types are equally America. The five Neotropical genera are indeed different. Dense miombo woodland and submon- near the root of the zodariine branch whereas the tane evergreen forest within each other's sight on Australian ones are situated in between the the Nyika Plateau in Malawi, contain completely different litter-layer Afrotropical and the Palaeotropical ones (the lat- spider faunas with hardly ter group has an apparently recent outflow into any genus overlap in the spider families that have the Palaearctic). The reduced area-cladogram been studied (Zodariidae, Linyphiidae, Lycosidae, Tetragnathidae; Jocque, unpublished resulting from these relationships is given in Fig. data). 3. This statement is at least puzzling: it is in contrast with the timing of major plate tectonic The present day bipolar distribution of many events that caused the breakup of Gondwanaland, ancient African taxa of ground-dwelling inver- as deduced from geological data and cor- tebrates might therefore be explained in the light roborated by many biogeographical data (e.g. of the evolution of broad-leaved rain forest 564 MEMOIRS OF THE QUEENSLAND MUSEUM

JOCQUE. R. 1991, A generic revision of the spider family Zodariidae (Araneae). Bulletin of the

American Museum of Natural History 201 : 1-160.

3 992. A new species and the first males of Suffasia with a redelimitation of the subfamilies of the Zodariidae (Araneae). Revue Suisse de Zoologie 99: 3-9.

KEAST, J. 1983. In the steps of Alfred Russel Wallace: Biogeography of the Asian-Australian interchange zone. Pp. 367-407. In Sims, R., Price, J. & Whalley, P. (eds). 'The emergence of the biosphere.* (Academic Press; London and New York). NELSON. G. & PLATNICK, N.L 1980. A vicariance approach to historical biogeography. Bioseience 30; 339-343 NORTON, 1.0. & SCLATER, J.G. 1979. A model for FIG. 3. Area cladogram of tropical continents as based the evolution shield of the Indian Ocean and the on the clade of the Zodaninae. breakup of Gondwanaland. Journal of Geophysi- cal Research 84: 6803-6830. This novel of vegeta- during the Cretaceous, type OWEN,H.G. 1983. 'Atlas of continental displacement, 1 tion with a very new structure and hence ecologi- 200 million years to the present. (Cambridge cal conditions, split most existing taxococnoscs University Press: Cambridee) that had to adapt to the new forest conditions. In POWELL, C.McA., ROOTS, S.R. & VEEVERS J.J. the Zodariidae only very few genera succeeded 1988. Pre-breakup continental extension in East in doing so. In Africa only Mallinella really Gondwanaland the early opening of the eastern 261-283. developed into a specialised forest spider genus. Indian Ocean. Teetonophysics 155: C.R., GAHAGAN, L.M. LARSON, R. The remote relationship of South American and SCOTESE. & 1 988. Plale tectonic reconstructions ofthe Cretaceous African Zodariidae may well be the result of and Cenozoic oceanic basins. Teetonophysics 155: at the the breakup of similar events time of 27^18. favouring the dispersion of par- Gondwanaland SHIELDS, O. 1979. Evidence for initial opening of the ticular genera, hampering that of others. Pacific Ocean in the Jurassic. Palaeogeography, Palaeoclimatology, Palaeoecology 26: 181-220. LITERATURE CITED 1983. Trans-Pacific biotic links that suggest earth expansion. Pp. 199-205. In S.W. Carey (cd). The BAEHR, B. & BAEHR, M. 1988. Zur Systematic expanding earth, a symposium.' (University of PhyJogeme und Zoogeographie der Tamopsis Tasmania: Uobart).

arnhemensis-tircumvidetis-tropictJ Gruppe VEEVERS, JJ , POWELL, C.Mca.. & ROOTS, S.R. (Araneae. Hersiliidac) aus Australien. XL 1991. Review of seafloor spreading around Europaisehes Arachnologischcs Colloquium, Australia. I. Synthesis of the patterns of spreading. TUB-Pokumentation Koneresse und Tagunecn Australian Journal of Earth Sciences 38: 373-389. 38:238-249. WOLFE, J. 1985. Late Crctaceous-Cenozoic History of BOSMANS, R. & HILLYARD, P. 1990. Spiders of the deciduousness and the terminal Cretaceous event. family Zodariidae from Sulawesi* Indonesia Paleobiology 13:215-226. {Arachnida; Arancac; Zodariidae). Bulletin of the WOLFE, J. & UPCHURCH, G. 1986. Vegetation, British Arachnologic.il Society 8: 147-160. climatic and tloral changes across the Cretaceous-

BRIGGS, J. 1987. 'Biogeography and plate tectonics'. Tertiary boundary. Nature 324: 148-152.

(Elsevier: Amsierdani-New York). 1 987. North American nonmarinc climates and vegeta- CAREY, W. 1975. The expanding earth: an essay tion during the late Cretaceous. Palaeogeography, review. Earth Sciences Review 11: 105-143. Palacoclimatology, Palaeoecology 61: 33-77. BASIC ARCHITECTURE OF THE OVARY IN THE GOLDEN SILK SPIDER, NEPHILA ClAVATA

AKIO KONDO. EIKO CHAK1 and M1TSUHIRO FUKUDA

Kondo. A. ? Chaki, E. and Fukuda T M. 1993 Mil: Basic architecture of the ovary in the golden silk spider, Nephila clavora. Memoirs of the Queensland Museum 33 (2): 565-570. Brisbane. ISSN 0079-8835

In gravid females of the golden silk spider, Nephila davata L Koch, the ovaries are composed of flat cistemae. To clarify whether the cisternal structure is primarily or secondarily formed by pressure from mature eggs, ovary development was examined from early

subadults to mature spiders. Serial paraffin sections erf the early subadults definitely show that a pair of longitudinal furrows, presumptive lumina of the ovaries, lined with loose cell

layers, penetrate perpendicularly into the ovarian tissue. So the bctSJC architecture ofthe OVflfl in N. clavafu is primarily Hal cislerna. Im Falle von reifen Weibchcn von Nephila clavata sctzen stch die Ovarien aus flachen Bctiteln zusarnmen. Um heruusz.ufinden, ob diese strukturelle Flitchc angeboren ist oder durch den Druck der reifen Eicrn gebildel wird, wurdc die Entwicklung dcr Ovarien vom fruhen Subadult bis zum reifen Adult untersucht. In den seriellen Paraffinstuckchen van fruhen Subadult zcigte sich, dap ein Paar J"iigitudinale Furehen, die mil /erstrcutcn £ei- Icnschichlen verstarkt wurden, d.h. die presumptive Lumen, perpcndiklar in die Ovarium- gewebe gehen. Daraus kann man sehliepcn, da0 die Grundarchitektur von diesen Ovarien angeboren aus flachen Beuleln besteht. \jSpider, Nephila, man, histology,

Akio Kondo. Eiko Chakt and' Mitsuhira Fukuda, Department ofBiology. Faculty ofScience, TohoW2 University, 2-1, Mixama 2 chotne, Funabashi-slu, Chiha, 274 Japan: 21 December

The golden silk spider, Nephila clavani L. clavaia were collected in Funabashi, Central K .>. 1, is able to deposit egg masses twice (Kondo. Japan, from late August to early September.

I988;ShimojaJia ) Pnorto the first oviposi- ) 197] After removal of the cuticle, the entire body or tion, gravid females possess complicated ovaries the opisthosoina was fixed in FAA (formalin 5. whose lumina are expanded among mature eggs. ethanol 15, glacial acetic acid 1) for 5-21 hours, After the second oviposirion, the ovarian lumen dehydrated in graded strength of ethanol, em- of exhausted females is not closed and contains bedded in paraffin and sectioned at 6-lOu.m. immature oocytes and yolk granule debris. In Before sectioning, a cut surface of the sample both cases, the basic architecture of the ovary is containing mature eggs was treated with distilled not clearly identified. In vigorous females imme- water (Kondo. [969), Serial paraffin sections diately after first oviposilion, the lumen of (he were stained with Mayer's acid-haemalaum and ovary is completely closed and the ovarian eosin. epithelium protrudes many young oocytes into the body cavity. In this case the o\arian tubes RESULTS elongate dorsally and ventrally to form flat cisternal as if the lumina have been squeezed between StlH mature eggs, already ovulated at the first oviposi- ADULTS tion, side by side {Kondo and Chaki, 1991). A pair of ovaries was situated behind the book extending sac surrounded Generally, however, the spiders have a pair of lung to the cloaca! and tubular ovaries or ring-like ovaries. The present by the midgut glands dorsally and by silk glands laterally and ventrally (Figs Although the study tries to clarify whether the cisternal ar- 1,2), ovarian tissue was divided in two at the anterior chitecture of the ovary of A', clavafa is primarily or secondarily formed by pressure from mature and posterior ends, the left and right ovaries were not separated from each other and formed single rod-shaped tissue for the most pari. MATERIALS AND METHODS Maximum width and thickness of the ov. tissue were 250u.m and 150u.ni in a small sub-

adult with ecphalolhorax 1.6 mm wide, 2.6 im l Four female subadult and two adult Aft long, and 500u,m and 220jim in a large subadult 566 MEMOIRS OF THE QUEENSLAND MUSEUM

FIGS 1,2. 1. Longitudinal section of a large subadult.cs: cloacal sac, m: midgut, mg: midgut gland, ov: ovary, sg: silk gland. 2. Transversal section of a female immediately after final moulting, m; midgut, mg: midgut gland, ov;ovary. sg: silk gland. Scales: 0.5 mm. whose cephalothorax was 2.8 mm and 4.2 mm (Fig. 3). The loose cell layer may be the ovarian respectively. wall or the ovarian epithelium. Then the furrows In transverse sections the ovarian tissue was seemed to be surrounded by a single layer of the trapezoidal. In a small subadult, two furrows oocytes. Most oocytes, whose cytoplasm was lined with a loose cell layer penetrated from the strongly basophilic, were cuboidal or polygonal upper side of the trapezoid into the ovarian tissue and 40-50ujti in diameter. OVARY ARCHITECTURE IN NEPHILA CLAVATA 567

* -J* «

FIGS 3, 4. Transverse section of ovarian tissue in small subadult (3) and in large subadult (4). Scales: 50>m.

The spherical germinal vesicle was 25-35u-m in The cells constituting the ovarian wall were more diameter and contained an extremely basophilic or less columnar and had ellipsoidal nuclei in the and slightly eosinophilic nucleolus. Yolk basal part and vacuolated protoplasm in the distal granules had not formed. The vitelline body or part facing the presumptive lumen of the ovary. Balbiani's yolk nucleus was not observed. In the large subadult, profiles of the oocytes Each oocyte was connected with the ovarian closely resembled the oocytes in the small sub- wall through a short egg stalk or the funiculus. adult, though they increased slightly in number 568 MEMOIRS OF THE QUEENSLAND MUSEUM

FIGS 5, 6. 5. Transverse section of ovary in female immediately after final moulting. A. Median dorsal part of ovary. B. Distal part of F-shaped lumen of ovary. 6. Lumen of ovary filled with eosinophilic fine granules in a

gravid female, f: eosinophilic fine granules, oc: oocyte, oe: ovarian epithelium. Scales: 50p.m. and volume (Fig. 4). A prominent feature was the Adults In an adult presumably immediately after its appearance of narrow lumina in the ovaries, while final moult, the width of the entire ovaries was vacuolate cells in the ovarian wall remained in approximately 1 .0 mm and the lumen of the ovary depth at both the anterior and posterior ends of was F- or T-shaped, branching laterally (Fig. 2). the ovarian tissue. The left and right ovarian epithelia were not OVARY ARCHITECT!TRF IN NEPHILA CLAVaTA

fused, but connected with the oviducts. Total basophilic cells by the final moult. Since the number of oocytes counted in one female was OOcytec arc connected with this epithelium 1.390 including the small oocytes. No oocytes through egg stalks, it seems to be the ovarian were observed above the dorsal side of the epithelium in the strict sense, though flat cells are hoiizontally developing lumina aligned along the outside of this epithelium. Oocytes reached a diameter of 170ji.ni. Ger- Subsequently the Oocytes begin vitellogeitesis minal vesicles were 40-60|xm in diameter. The and accumulate eosinophilic yolk granules in nucleolus was collapsed and dispersed into their cytoplasm. The lumina of the ovaues con- nucleoplasm. Cytoplasm of the oocyte was less tinues to branch amongst the oocytes and on the. basophilic (Fie. 2). The eosinophilic yolk surface of the mass of the oocytes, and they arc granules were neither formed nor accumulated. filled with eosinophilic fine granules which The ovarian wall seemed to be a simple columnar respond to fine granules attaching to die sui or pseudostratified epithelium composed of of the chorion of the lycosid egg (Kondo, 1969). slender* compactly arranged and extremely the spheres on die chorion (Humphreys, 198"^ philie cells. Flat cells were lined up between 1987) or cement substance of the egg nus> the ovarian epithelium and the oocytes (Fig. 5). (Kondo and C'lmki, 1991). In a gravid female whose ovaries reached 4.0 In A* clavata, the ovarian epithelia of the left mm in width, the ovary contained many develop- and n>;hl nvaiics arc not directly connected with ing oocytes 3S0-460u.m in diameter, and tew each other, though in archacid spiders they arc immature oocytes. The developing oocytes were connected through a bridge in the posterior por- tilled with eosinophilic yolk granules and had a tion of the ovaries to form a H-shaped organ terminal vesicle 50-60fim m diameter. The (Traeiuc and Legcndrc. 1970). Interstitial tissue. lumen of the ovary branches among the develop- however, does not exist between the left and right ing oocytes. The lumen of the ovary contained ovaries, so the ovaries of A! clavata lorm a single nophihe fine granules which were presumab- rod-shaped organ. This rod-shaped structure w as ly a kind of cement substance used to hold the egg confirmed not only from both transverse and ukiss after oviposition horizontal sections but also by dissection of the (Fig ( 6) opisthosoma. DISCUSSION In conclusion, in N clavata the basic architecture of the ovary is primarily composed of flat cisterns. The formation of the ovary in Ncphila clavata W&3 examined from the early subadult stage The LITERATURE CITED small subadult employed was possibly a nymph of the 6th or 7th instar prior to the penultimate HUMPHREYS, W.F. 1983. The uirface of spiders' moult. In any case, at an early stage, the lon- tv-'t'.s. Journal of Zoology, London 200: 303-316. gitudinal ovarian walls arise as loose cell layers 1987, TllC accoutrements of Spiders' eggs penetrating from the dorsal side ol the ovarian (Araocae), with an exploration oftheir functional tissue perpendicularly into the mass of the importance. Zoological Journal of the Linnean oocytes. The epithelial cells are vacuolated in Society, London 89: 171-201. KONDO, A. 1969. The One structures ofthe early spider their distal part and form a lumen which is I- embl v.i. The Science Report of the Tokyo Kyoiku shaped in ttansversc .srvtioii. Then, tlu* ovaiian Daig&U Section B No. 207: 47-67, pis 1-8. epithelium constructs flat rather than cylindrical I'/XK, Effects of incubation temperature on the .'siom.-ie embryonic development and the oviposition in By the final moult the lumina have been formed the si Jk spider. Nepluln clavata L. Koch. Proceed- jn the vacuolated poilion of the ovan.ni ings Of Arthropodan Embryological Society <>t epithelium. Immediately ulterthe final moulting, Japan No. 23:21-22. while the oocytes increase in number and volume, KONDO. A. & CHAKI. E. 1991 Histological studies a few branches of the lumen arise, mostly m the on Hie ovaries ot the golden silk spidet, Ncphila L. Koch, before and after oviposition. posterior half of the ovary, forming F- or T- ChlVara Proceedings of Arthropodan Lnibryolugica! •itupcd profiles in transverse sections Simul- Society of Japan No. 26: 9-10. taneously the oocytes grow and become less SJUMOJANA. M. 1971. Studies on the genus Ncphila basophilic, perhaps due to a cytoplasmic in ( life in Okiflftwa Island. 1 ) Studies on the cycle of e. The li »ose cell layer in the small subadult u. Biological Magazine Okmav ii-msforms gradually into columnat epithelium I IB (In JapOJteftd I, Composed Of slender, compact and sirorigly TRACtUC r & LBGENDRE mr 1970 Doimfe 570 MEMOIRS OF THE QUEENSLAND MUSEUM

anatomiques sur l'appareil genital femelle des Ar- domadaires des Seances de 1* Academie des Scien- chaeidae (Araneides). Comptes Rendus Heb- ces,SerieD, Sciences Naturelles 270: 1918-1921 NEW PREY RECORDS FOR SPIDER HUNTING WASPS HYMENOPTERA: POMPILIDAE) FROM THE NETHERLANDS

P. KOOMEN andT.MJ PEETERS

Koomcn. P. and Peelers, T.MJ. 1993 II 11. New prey records for spider hunting wasps (Hvmcnoptera: Pompilidae) from The Netherlands. Memoirs of the Queensland Museum 33(2): 57 1-574. Brisbane. ISSN 0079-8833.

In the southern Netherlands, 56 spider prey were collected, and identified, together with their pompiud predators. Those data are compared with data from the literature for northwestern Europe. Anoplius spp. seem to prey mainly upon ground dwelling spiders. A strong preference for the wolf spider Trochosa terricola was found for Anoplius v'talicus\ in agreement with data from Bristowe (1948). Episyron rufipes and Caliadurgusfastvun ellUS attack only orbweb building spiders. Data for Auptopus carbonarius and Dipogon subinter- medium suggest that these species prefer to hunt on vertical planes like tree-trunks. Amch- nospiiu rufa. Pompilus anereus, and Priocnemis spp. appear to be non-selective. Dans le sud des Pays-Bas, 56 araignccs, proies de Pompilides, on 6t6 rccoltees avec lcurs prcdaieurs ct identificcs. Lcs donnces acquiscs ont etc comparers avec les cellcs dc la litteraturc concernant les observations rial! sees aux Pays-Bas ct dans' les pays voisins. Anoplius spp. montre unc nelte predilection pour les araignees terricolcs. Conforme'ment aux

donnlcs de Bristowe ( 1 948). Anoplius xiaticus preiere de loin Trochosa (erncolu, Episyron rufipes el Caliadurgusfascitilellus s'attaqucnt seulcmcni aux araignccs qui construiscnt des toilcs orbiculaires. Les observations sur Auphpus carbonarius et Dipogon submtermedius suggcrenl que ccs especcs chassent de preference sur des plans vcrticaux comme des vieux mors ou des troncs d'arbres. Arachvsp'da tufa, Pompdus cinereus et lcs Prujcnemis spp. scmblent etredes especcs non select'ives. iy\raneue. prcdation, Pompilidae, selectivity, The Netherlands'.

Peter Koomen. Havikshorst 34, NL-23J7 AL Letdett. The Netherlands; Theo M.J. Peeters, P.O. Box 9258, NL-5000HG Tilhurg, The Netherlands. 6 November, I99Z

Female Pompilidae attack and paralyse spiders catch what they can, or prefer, for some reason, with their sting. Each captured spider is the food some species of prey over others. This paper of a larva that hatches from an egg laid on the reports on the first results of such a study which spider. Before oviposition, most European pom- began in 19&7. pilids transport their prey to burrows excavated in the soil, or to pre-existing cavities, e.g., behind METHODS AND MATERIALS lurk or between stones (Gauld & Bolton, 1988). Knowledge of pompilid prey is limited for ob- Pompilids were studied at several locations. \ious reasons: most hymenopterists are interested mainly in the southern Netherlands, near of Til- unly in pompilids and not so much in their prey, burg- Unless stated otherwise, the second author Arachnologists tend to accept prey ofPompilidae located pompilids by eye and intercepted prey- gratefully, but without identifying their transporting females using glass tubes and ait predators. As a consequence, data on prey hardly insect-net. Prey were collected and preserved ifl exist, are out of date and/or are unreliable because alcohol. Predators were collected only if iden- of changed taxonomic views, or do not go beyond tification in the field was not possible. Wasp* generic level (e.g., Day, 1988; Oehlke and Wolf, were identified with the aid of Wolf (1972), Dav 1987). However, collectively, all published data (1988) and Oehlke and Wolf (1987). Identifies- suggest some prey specialization, but it is not tion of spiders was based mainly on Roberts clear if this is due simply to opportunistic be- (I985ab, 1987). Not all spiders could be iden- h;*viour of pompilids, catching only (or mainly) tified to species level, because some were the most abundant prey. A way to gather more juvenile, thus lacking the genital organs neces- knowledge concerning possible preferences of sary for a proper identification. Autecological pompilids, is to catch pompilid predators with data on spiders are derived from Jones (1983) and their prey, and to accurately identify them both. MaurerandHanggi (1990). All spiders and wasps Autecological infonnation on the spiders may collected will be lodged in the National Museum give a clue as to whether pompilid species only of Natural History. Leiden, The Netherlands 1 1;

572 MEMOIRS OF THE QUEENSLAND MUSEUM

RESULTS Priocnemis minuta (Vander Linden): Gnaphosidae (1 juv., Moergestel, 17.viii.1988). Priocnemis susterai Haupt: Drassodes cupreus In total, 56 pompilid prey were collected. Data (Blackwall) (1 9, Boxtel, 21.V.1988). are summarized below. For each specimen the specific date and collecting site are given because DISCUSSION both may have influenced hunting behaviour of the wasps. When more than one prey record is The new observations can be compared to data available for a single pompilid species, common from the literature on pompilids in The Nether- characteristics of the prey are given that may lands (Benno, 1969; Bouwman, 1915ab, 1916; point to a preference of the actual wasp. Chrysanthus, 1947; Lefeber and Van Ooijen, Walrecht, and nearby POMPILINAE: 1988; Thijsse, 1907; 1936) countries (Bristowe, 1948; Day, 1981, 1988; Anoplius concinnus (Dahlbom;: Pardosa sp. (1 juv., Oehlke and Wolf, 1987; Wahis, 1948, 1955, Thorn, 13.x. 1990); Trochosa ruricola (Degeer) (16, Thorn, 13.x. 1990). Both are wolf spiders, Lycosidae. 1962; Wolf, 1971). A summary with references They were excavated from a, supposedly, single bur- included is given in Table 1. Comments for each row between pebbles in a gravelpit, after a female A. pompilid species follow hereafter. concinnus had been observed showing nest building Anoplius concinnus was found with wolf activity. spiders (Lycosidae) as prey. This is in agreement Anoplius infuscatus (Vander Linden): Lycosidae (1 with the literature, where several genera of juv., Moergeslel, 30.iv.1988); Alopecosa accentuata Lycosidae are mentioned. (Latreille) (1 juv. Nicuw-Bergen, 6.ix.I991). Both are Both in The Netherlands and in other countries, wolf spiders. recorded Anoplius viaticus (Linnaeus): Agroeca brunnea a preference for wolf spiders has been (Blackwall) (1 9, Beegden, 12.V.1991); Trochosa ter- for Anoplius infuscatus. Beside these, ricola Thorell (1$ excavated, Alphen en Riel, Thomisidae and Agelenidae are mentioned. Most 5.V.1989; 19+1

1 , Drunen, 9.v. 1 , Loon op Zand, 12.iv. 1 9 1989; 9 991 species. So, the available data may point to a 19 excavated, Moergestel, 24.iv.1989; 29, do., preference for ground dwelling spiders. 27.iv.1989; 29, do., Moergestel; 29, do., 5.V.1989; Anoplius viaticus appeared to prey almost ex- 1 9,do.,29.v.l989; 1 9,da,23.vi.l989; 109+16% do., clusively the spider terricola 29.iii.1990); Pardosa monticola (Clerck) (19, Beeg- on wolf Trochosa den, 18.V.1989). All three species are ground dwelling (Lycosidae). These observations are neither spiders. restricted to one locality, nor to a short period. Arachnospila rufa (Haupt): Alopecosa fabrilis Only in two of 41 cases (5%), A. viaticus caught

(Clerck) ( 6 , Nieuw-Bergen, 1 .ix. 1 99 1 ). other species: Pardosa monticola (Lycosidae) Auplopus carbonarius (Scopoli): Haplodrassus sp. ( and Agroeca brunnea (Clubionidae). This sug- juv. found in a beehive in one of a group of barrel- gests that A viaticus did not simply catch the most shaped cells typical of A. carbonarius (Grandi, 1961: abundant wandering spider, but rather searched 75; Wahis, 1949: 99), Baarlc-Nassau, 1991); Clubiona actively for T. terricola. In literature on The brevipes Blackwall (19, Udenhout, 1 l.viii.1990). Netherlands, the sp. Thijsse Both are nocturnal hunters often hiding in silk retreats Lycosa mentioned by during the day. (1907) might refer to almost any modern lycosid Episvron rufipes (Linnaeus): Larinioides cornutus genus. From data on prey elsewhere, the list given (Clerck) (19, Helvoirt, 16.viii.1991); Nuctenea by Bristowe (1948) matches our data very well: umbratica (Clerck) (1 juv., Moergestel, 26.vi.1990). he found in only seven of 47 cases (15%) prey Both are orbweb building spiders. that were not Trochosa terricola. However, other Pompilus cinereus (Fabricius): Lycosidae (1 juv., papers indicate several families and genera. Some Moergestel, 4. viii. 1987). are hunters by day like Lycosidae, while others are nocturnal hunters (Drassodes) or inhabitants PEPSINAE: of sheet webs (Agelena). Common attributes of Caliadurgus fasciatellus (Spinola): Agelenatea redii the recorded spider genera are their medium to (Scopoli) (1 juv,, Moergestel, 4.x. 1989). large size (5-13 mm) and the strata they occupy: Dipogon subintermedius (Magretti): Segestria on or near ground level. Most also prefer rather senoculata (Linnaeus) (1 9 found in a cavity in an old pollard-willow (Salix alba), Tilburg, 23.vii.1991). dry and/or sandy places as does A. viaticus Priocnemisfennica Haupt: Clubiona terrestris Westr- (Thanatus, Aelurillus, Pardosa monticola, ing(l9,Hilvarenbeek, 29.ix.1990). Trochosa terricola), but this is not true for PRE V RECORDS FOR DUTCH SPIDER HUNTING WASPS S7J

PfPdMOf F'rcv

AttOptius concinnus Lvcusidat (D2): Arctom (O&W), Pardnxa s ., ( K&P, L&VO. O&V ), Trochasa fftri oU K&P) f \

Thomisidiw (D2) Thanatus (O&W) Lvctwida? (D2, K&P); Ahpecoso uccentuaia (K&P>, An-fiora (O&W), Pardoxa Aiuypltux infusratui (O&W), /Vritw (O&W), Thau (O&W), Irvtrhoxa (L&VO. O&W), Xrrolycoxa (O&W). Ag*knid»e(D2): 2V«ru (O&W)

Gnaphnsidae(D2); Drasswlvs lO&W), Clubionidac: A&vw

Ancpliifi viuthus Sub Turenlula a.). A. aculeata (*W3), A pufvenilenta (B, W2: sub TarentuLi carinatu), Arrtnxn (O&W), .•}, pflitS (Bl.

/Waw i o&W j. r\ her&nsis (W2l, P womirefo (B. K&P). Tw/wju (O&W). r. ntrric&te (B, K&P, W

Ksaundae Pisaura mirabilh (O&W. W2). AgelcnidB* t D2 k Agitata (O&W).

Gnapiwsidw (D2>: Dras;odts CO&W j. Gwsphosa (O&W). Cluhnnidae (D2>: Cheirutatutdum (U&Wj. SaMieid**- 4r MhMwtMr p . r fh ! "'" fD2): V-/urrt/w (O&W). Lyeosidae fD2>: Alopecosa (O&W: suh 7dwirafi& A /b/.n/i.v(PAK|. Irothmu (O&W),

ZorogiKHliu" A>r,.ynK (Wr>) SltfUlrHdME (02): SegfVrut(\\'c>L {innphotidae (T>2): Aphu>tlaisk HaphdraSSUSSp, [K&P) l'lubki»fdi«MD2): Cheiracaiuhium (Wu), Chtbiana f»m?pf.r(K&P). Aupfapus CRfo), C Anvphi»crtidac(D2). Th«mUklt*ir(D2j PhUodromus (Wo). SaHicidae(D2>: DendrypbajitexiVJo), fciu/r/iufWol,

r l SiUKwfWo), l#cwfeto«lt>2); /..vanwIWol /rf>dlOit7(M o). Agctenid»e(D2»:4 i-Wt-/ur(W .:0. Tclragnalhidae: Mfto(Wa).

LyrfMcida? » L)2: only some CMM) TclrnRnothidae A*cm (D2). Aranvidae Agelinnteu ndii i B. sub rlrtttaclu t ). 1''{n:vr»nrt4ftpe% A/ttTtttti (I>2. O&W), A. Jiai/t'mtirji\lB), Aw'fJ/v(0&W), Gibbaranea gibbcxu (L&VO; sub Artmeux gibbasus),

/_(i. Ihloidti i mttutits (B: :*ut> A. foliuia, K&P), Nuctenea andrrnltcn (K&Pi

Atypidae. AfttMU sp. (C). Zuropvidjw; Zflropsis (Q&!W), CnaphosidaelDI ) i'tt-mtttLha (O&W). Clubionidtte

fDl) W&'raniWhiumiOfcW) Zoridae (Ot): Zb«i<0&W>. Thomi

i'utiadutgus Telragnathidac: AitAii (D2. O&W). Af segmentota (L&VO). Araneidae: /Igf/r/wf^j r^»/(V ( k&P) />r*;«€wf (D2. faxiUtttUui 0&W),A dW«f (B2)./t dirjdntintu* (Be, B2. W1),A. quadrants (Wl).

Scgeslriidac: iV,e«/rit; (O&W). y Wiocut&fa (1>2. KftPf. SaHicki»e: fiflfrfcttl (CM ..'•'hi:', "itvimj

,"' :. n'nemis fenmca Cluhionidae: Clubiwui rerreiTn fiK&P). Lycosidae: Panh'M puffoh. ID2.I Prigenemu mitmtu Gn^phf^idae (K&P).

faiocfltmUt susterai Gnaphosidue: />ruj;.Tat^.t (O&W), /? cuprcut (K&P)

TABLE 1 . Summary of the prey records in the IHcr.Hurc tor the pompilid specie? for which this study gives new - data. References arc abbreviated; B = Bnsiowc, 194S; Be = Benno t 1969, Bl = Bouwman, 1415b; tt2 Bouwman, 1916; C = Chrysanlhus, 1947; Dl = Day, 1981; D2 = Day. 19R8; K&P = Koomen & PeetCTS, Ihis ^ludy; L&VO = Lefeber and Van Ooijen, ]988; O&W = Oehike and Wolf. 1987; T = TJiljise, 1907; Wa = Walrccht, 1936; Wo -Wolf, 1971, Wl = Wahis, 1948: W2 = Wahis, 1955; W3 = Wahis, 1962

AgTv&ctl brunnea, an inhabitant of wet places. So, bonarius is mainly a vertical hunter cannot be although our observations and Bristowe ( 1948) supported or falsified. Several authors note that suggest a strong preferervee for nnlv OWt species A. carbonarius has a habit of amputating the legs of Lycosidae, data from other countries show a of its preys (e.g.. Grandi, 1961: fig. 48; Oehlkc much broader spectrum of prey. This may point and Wolf. 1987). This did not occur in the two to a preference for Trochosa lereivola, that is cases we nnted. Both the Haplodrassus and the abandoned when the latter species is not avail- Clubitma prey still had all of their legs. able. Episvrwt rttf\ as found to prey upon hVO Arachnospila rufa is not selective: spiders with arancid spiders. Using data from other countries £*, various sizes, and from several habitats and SUalfl (see Table 1 ) il can be deduced (hat rufipes have been recorded. Our observations support specializes on orbweb building spick these data. For Pompilus cinercus many different prey are Vox Auplopus carixmarius, the observations known, including, burrowing spiders (Arypus), that Segestriidac, Clubuma hrv-\ tjn a, nrbweh huihlin^ > ( (Arttncus), 9Dd Anyphaenidae, and Dendryphantes were prey xcrophylic iArctosa pcri;a, Aclurillus) and items may point to a preference for spiders that bygrophylic species iOohmedes). This suggests walk along, or hide in crevices in, vertical planes, mi so at all. Tn dale, only one juvenile e,g. walls or Lrec-trunks. Species with (at least wolf spider can be added. occasionally) similar habits are found in most of Caliadurgusfasciatellus is a second pompilid i lie other families and genera recorded (Zoropsis, species that is obviously specialized in capturing Aphantaula.x, Haplodrassus^ Philodromus, Evar- orbweb building spiders The /Irani prey r-'V.*;, I ). Situcu\,Avi-}ena > MeUi\, hut withum SfX fas well within these data (Table lificauons, the hypothesis that A. car- The recorded prey of Dtpogan subintermedius SM MEMOIRS OF THE QUEENSLAND MUSEUM

arc in Conformity with its observed hunting GAULD, I. & BOLTON. B. (eds.) 1 988. The Hymenop- places: tree-trunks and piles of fences. Segestria tera. (British Museum (Natural History): Oxford University Press). senoculata is known to build its retreat in these 332pp. GRANDI, G. 1961. Studi di un entomologo sugli Im- places, and jumping spiders from the genus Sal- enottcri superiori. Bolletino dell'lstiluto di En- ticus may also be found there. tomologia dcirUmversita di Bologna 25: 1-659. For the three species of Priocnemis, viz., P. JONES, D. 1983. "A guide to spiders of Britain and fennicxi. P. tnmuta. and P. susterai, reliable data Northern Europe.' (Hamlyn: London). from the literature are few, because formerly LEFEBER. V. & VAN OOIJEN, P. 1988. these species were not distinguished as separate Vcrspreidingsadas van de Nederlandse Spinnen- taxa. Prey recorded so far do not suggest any high doders (Hymenoptera: Pornpilidae). Nederlandse 1-56. degree of selectivity. Faunistische Mededelingcn, Leiden 4: MAURER, R. & HANGGL K. 1990. (Catalog der ACKNOWLEDGEMENTS Schweizerischen Spinnen. Documcnta Faunistica Hclvetiac 12: 1-33+ !89pp. OEHLKE, J. & WOLF,H. 1987. Bcitragezurlnsekten- Messrs. Br. V Lefeber (Maastricht ), J. de Rond Fauna der DDR: Hymenoptera-Pompilidae. (Almere). and R. Wahis (Gembloux, Belgium) Bcitrage /.ur Entomologie. Berlin 37(2). 279-39(1 are thanked for their help in identifying the Porn- ROBERTS. ML 1985a. The spiders of Great Britain pilidae, and Mrs J.D. Prinsen (Wagcningen) for and Ireland 1: Atypidae to Thcridiosomatidae/ the identification of Alopecosa fabrilis. Messrs. (Harley Books: Colchester). 1985b. spiders oF Great Britain Ireland 3: J.E. de Oude (Den Haag.) and R. Wahis are ac- The and Colour Plates-Atypidac to Linyphiidae. (Harley knowledged for translating the summary into Books: Colchester). French. We thank Mr. J.C Felton (Brighton, 1987. 'The spiders of Great Britain and Ireland 2: U.K.) for improving the English text. Linyphiidae and checklist.' (Harley Books: Col- chester). LITERATURE CITED TH1JSSE, J.P. 1907. De wegwcsp {Pompihts viaticum. De Levende Naluur 12(4): 66-68, BENNO, P. 1969. Aantekeningen bij eruge Neder- WAHIS, R. 1948. Nidification de Calicurgus landse wegwespen (Hymcnoptcra, Pompilidac). hyalimtius F. (HymtSnoptcre Pompilidac). L'- Entomologische Bericbtcn, Amsterdam 29: 72- Entomologiste4: 210-213. 76. 1955. Contribution a P elude des Hymenoptera

BOUWMAN, BE 1915a. De wegwespen en hun Pompilidac. 1 . Lcs Priornemis beiges voisins du parasict. (Porttpi!&t$ en Cerop&tes), De Levende Prh'cnetrns periurbaior Harris (fiisctis auet,). Naiuur20m:226-229. Bulletin et annates de la Societc Royalc d'~ 1915b. De wegwespen en hun parasict. [Pompiius Entomologie de Bclgique 91(3/4): 92-108.

- en Ceropales). De Levcnde Ntfuvr 20< 1 3); 24 1 1962. Les Pompilides beiges de la collection Adol- 246. phc Crevecoeur (Hymenoptera: Pompilidac). BR1STOWE, W.S. 1948. Notes on the habits am) pre) Bulletin et annales de la Socicte Royale d'- of twenty species of British hunting wasps. Entomologie dc Bclgique 98(19): 324-33S. Proceedings of the Linnean Society of London 1986. Catalogue systematique et codage des 160: 12-37. Hymcnoptcres Pompilides de la Region Oucst- CHRYSANTHUS, P. 1947. Spinnen en Hymenoptera Europdenne. Notes fauniques de Gembloux 12: Aculeata. Tijdsehrift voor Entomologie 88: 482- 3-91. 488. WALRECHT, B.J.J.R. 1936. Dc levcnsloop van DAY. M.C. 1981. A revision of Pompihts Fabricius Pseudagemu carbonuria Scop, De Levende (Hymenoptera: Pompilidac), with further Natuur4H.5): 148-153.

nomene latum 1 and biological considerations. Bul- WOLF, H. 1971. Prodromus insectorum Bohcmos- letin of die British Museum (Natural History). lavakiae 10: Hymenoptera Pompiloidea. Acta

Entomology 42t 1 ): 1-42. faunistica entomologica Musei Nationalis Praeae 1988. Spider wasps. Hymenoptera Pornpilidae. 14(3): 1-76. Handbook For the Identification of British Insects 1972. Hymcnoptcra -Pompilidac. Insecta Helvetica,

6(4); 1-flft. Fauna 5: 1-176. GROUND-LIVING SPIDERS (ARANEAE) ONE YEAR AFTER FIRE TN THREE SUBARCTIC FOREST TYPES, QUEBEC (CANADA)

SEPPO KOPONEN

Koponen, S. 1993 11 11: Ground-living spiders (Araneae) one year after fire in three subarctic forest types.. Quebec (Canada). Memoirs of the Queensland Museum 33(2): Brisbane. ISSN 0079-8835.

The ground-living spider fauna was studied one year after fire ustng pitfall traps in three forest types of subarctic Quebec. July-August 1990. About 30 species, of the total 47 found at burned sites, were regarded as pioneer or colonizer species. Spiders captured commonly at burned sites included e.g. Gnaphosa microps Holm, G. muscorum (L. Koch), Pardosa hyperborea (Thorell), P. uiniana Gertsch, Trochosa terricakj Thorell, Alopecosa acuteara (Clerck), Diplocentria hidentaia (Emerton). and Sisis rvtundus (Emerton). Some species were found only or predominantly at unburncd sites; e.g. Pardosa rnoesta Banks and Lcpthyphantes complicate (Emerton). En juillet-aout 1990, unc annee apres des meendies, la faune des araignees habitant sur le SOl des trois types de foret du Quebec subarciique a 6\c ctudicc en utilisant des pieges-fosscs

Environ 30especes t d'un total de 47 trouvees dans les sites brules. ont etc* considerees comme desespeces piorvniercs ou colonisatrices. Les araignees capturccs gcn<5ralemcnt dans les sites bruits comprennent parexemple Gnaphosa microps Holm, G. muscorum (L. Koch), Pardosa hyperborea (Thorell), ft uiniana Gertsch, Trochosa rerrtcoiu Thorell. Alopecosa aculeata

(Clerck), Diplocentria bidentata (Emerton), et Sisis rotundatus { Emerton) . Quelques especes ont et6 trouvees seulement ouprincipalement dans les sites non brules, parexemple, Pardosa rnoesta Banks et Lcpthyphantes complicalus (Emerton). {jAraneae, forest fire, subarctic, Canada.

Seppo Koponen, Centre d'efudes nordtques. Universite Laval. Ste-Foy. Quebec G!K 7P4, Canada; present address: Zoological Museum, University of Turku, SF-20^00 Turku, Finland; 27 October, 1992

Fire history, effect of fire on forest vegetation, zone of Quebec, Canada. Forest fires occurred in and postfire succession of plant cover have been midsummer 1989; samples were collected in studied in detail in northern Quebec (e.g. Payette 1990 during the first postfire summer. // a/., 1989). By contrast, very little is known about the effects of fire on spiders in northern MATERIALS AND METHODS forests of North America. In central Alaska. Be- ekwith and Werner (1979) suggested that the decrease of many arthropod populations after fire The study areas were situated 1) at Lac can be attributed to increasing spider populations Ekomiak (53°2.TN, 7T"30*W), south of La jnd prcdation by spiders. Data from temperate Grande/Radisson and 2) at Kuujjuarapik (Paste* , forests of North America are also markedly scant de-1a-Baleine;55°17'N. 77°48 W) on the eastern (e.g. Pearse. 1943; Buffmgton, 1967). Pearse coast of Hudson Bay (Fig. 1). Both study areas (1943) listed about 35 species of spiders from were near the northern limits of the boreal forest. burned pine forests in North Carolina; however, 1 . At Lac Ekomiak, large areas of forest (thous low. individual numbers were ands of sq. km) were burned in midsummer 1989 In forests of northern Europe, postfire spider The study sites situated near the border of this faunas and their succession have been studied extensive fire area were in dry and mesic forest. both after natural fires and prescribed burning The main tree species were Pinus batiksuina ..• ,g« Huhta, 1971; Sehaefer, 1980; Haugc and Lamb, and Picva mariana (Mill.) B.S.B. at dry Kvamrne, 1983; Koponen, 1988, 1989). Data sites; Picea mariana and Lurlx lahcina (Du Rot) fu>m subarctic forests of Europe arc available K. Koch at mesic sites. The ground layer of the only from northern Finland (Huhta, 1971; dry sites was characterized by Cladonia ai>d Koponen, 1988. 1989). Pleurozium\ that of the mesic sites by Sphagnum This paper deals with the effects of fire on and PleunKium. The fire had been very intensive ground-living spiders at the beginning of fauna! and had killed all trees and destroyed field and succession in three forest types in the subarctic ground layer vegetation totally. The trapping 576 MEMOIRS OF THE QUEENSLAND MUSEUM

FIG. 1 (left). Postfire study sites at Lac Ekomiak and Kuujjuarapik, Quebec. Broken line indicates the northern limit of the continuous (subarctic) forest.

LDB LDC LMB LMC KB KC TABLE 1 (right). Total number of individuals in major Lycosidae, inds. 59 16 4.S 32 104 28 spider families, total number of species, diversity (H) Linyphiidae, inds. 44 35 34 22 110 172 and evenness (E) for pitfall trap material at Lac Ekomiak and Kuujjuarapik, Quebec. LDB = Lac Gnaphosidae, inds. 6 8 6 3 11 3 Ekomiak dry burned, LDC = Lac Ekomiak dry un- Others, inds. 6 5 6 14 3 = burned, LMB = Lac Ekomiak mesic burned, LMC Total, iu

Evenness E 0.87 0.88 0.90 0.84 0.73 0.77 period was 24 July-20 August 1990; traps were not changed during the period. with covers x 12cm) 2. At Kuujjuarapik, a small area (about 50 x They were provided (12 15m) of Picea glauca (Moench) Voss forest was against rainfall and litter, and there was 2-3cm burned in July 1989. The site is in an isolated, space between the cover and the ground. Traps small-sized woodland near forestline. Ground were placed in a line at each site, average distance layer was dominated by Pleurozium and Em- between the traps being 2m. vegetation petrum. The field and ground layer The indices used are Shannon-Wiener index of was destroyed but the intensity of fire had been diversity: less than at Lac Ekomiak. The trapping period was 14 July- 19 August 1990. The traps were H = changed once (1 August). -$>0(fo*2/>D

Ten traps were placed at each site (dry burned, and evenness: E = H/logS (S = number of species, dry unburned, mesic burned and mesic unburned) pi = proportion of total sample belonging to the at Lac Ekomiak and five traps at each site (burned and unburned) at Kuujjuarapik. Pitfall traps were /th species). The spider material is deposited in plastic cups (diameter 6.5cm, height 7cm) with the Zoological Museum, University of Turku, ethylene glycol (2.5cm) as a preservation liquid. Finland. 1

GROUND-LIVING SPIDERS AFTER FIRE i

Species n %BS silt 'preference' Species n 9 BS 'preference'

Alopccosi? aculealu 6 100.0 only at bumed Pordosa hvptrbareu 16 |flD.G only at burned site

Pardoso hyperborea 15 93.3 strongly io bunted PtKadicnemis americana 6 100.0 only at burned site

Oiplfcemha bidenfata 10 90.0 ilron^Jy lo hunted An torn alpi$er\a 9 T, r, lo burned site

g Trochoid Terricola 16 75.0 lo bumed Pardpsa uintana 100 74.0 to burned site > PardowsfPr* 31 58.1 equally cecum' ng Gnupb'jya muicttn/m 111 70.0 slightly to burned site 5 Gnaphosa spp. 8 50.0 equally occurring undus 10 60.0 equally occurring

Ajtyneta oln'acea 7 42.9 equally ocx m tog Srucottu c montunus 22 59.1 equally occurring

'. 13 15.4 to control A^vueia allosubtifii 19 15.? n;> L'.-.nui <] site

Pardam hypcrbortu 10 *UI strongly to burned latithorax obtusus 13 \5A rot site

-j ,, Diploceturrj t sdemata 6 y ? 3 lo burned fspthyphamfx alpimtK 14,3 i , nnm.1 sitf

r,r, i i.-0|i'i»'l " . E Gtuiphoxit micropx 9 b6.7 slightly to burned Hilairfl herniosa 1 n n .

Pardota&pp. 4: 6o.7 slightly to burned 1 LeptirfptuMTescvmpticants 10 5.3 sir.-iiglv to control site U | Qzvpttta 13 46.2 equally occurring i gaftochi Hilntro herniosa 7 34.3 to control TABLE 3. Common spiders ( - 6) trapped at bumed !':.' - »5 0.0 ' only ai control l and unburned (control) sites at Kuujjuarapik, Quebec,

n - total no. Of individuals trapped, both sites • TABLE 2. Common spiders (^ 6) Irapped at burned bitted; % BS =percentagc individuals caught at and unburned (control) sites, at Lac Ekootiafc, burned silc; * = adults of H herniosa and juv Quebec, n = total number of individuals trapped, both Hilatra specimen?. sites combined; %BS= percentage individuals caught #

io>> borea (Thorcll) in both study areas and A lopecosa juvenile H'daira specimens. acuieata (Clerck) a; Lac Ekomiak (Tables 2. 3). The species group of Pardosa uintana Gertseh, RESULTS P. rnackenziana (Keyseriing ) and P. xerampelhta (Keyserlingi, including many juveniles, as well A total of 772 spiders was collected from 6 as Arciosa alpigena (Doleschal) at Kuujjuarapik. study sites (3 burned, 3 unburned) in northern and Trochosa terricola Thorell at Lac Ekomiak. Quebec during July-August. 1990 Individuals ol also were more abundant at burned than unburned two families, Linyphiidae and Lycosidac, clearly sites. dominated all collections; individuals of The gnaphosids caught, Gnaphosa micropx Gnaphosidae ranked third among trap captures Holm and G. muscorum (L.Koch), were slightly (Table 1). more abundant at burned than unburned sites. Of Both at Lac Ekomiak and at Kuujjuarapik, trap the linyphiids (Erigoninae). Diplocentna huL-n captures were higher at burned sites than at un- lata (Emeiton) and Pocadlcnemxs americana burned sites. This was mainly due to the greai Millidge apparently 'preferred' burnt areas numbers of Lycosidac caught at open bumed (Tables 2. 3 sites. The figures for the two most abundant families were (Lac Ekomiak and Kuujjuarapik Many species were represented by less than 6 individuals captured, consequently noi in- combined): Lycosidae 208 at burned and 76 at and

in . Sev eral unburned sites, Linyphiidae 188 and 229 respec- cluded the Tables 2, 3 were found only at bunietl Mies. This included: tively. In general, at the burned sites, the diversity group Gnaphosa ftarvula Banks. Zeiotesfratris Chamberhn, Par- I'll) was higher or equal compared to the un-

(Thorel 1 1. nelti Pec-khan burned controls (Table I). dosafunijera Neon Peckham. Sislcus apertus (Holm), Ceratkelus Altogether 56 species were caught, 37 spei atriccps (O.P -Cambridge), Horcotesquadricns- were trapped at Lac Ekomiak and 34 at Kuuj- (Emcrton), Sciasies (Emerton), juarapik. Number of species from burned sites rams truncates Tunugyna debifil (Banks). Waickenaeria WHS 47; from unburned sites 37; 28 species were aJrotibialisO.P -Cambridge VV. eastanea (Emei- common to buth burned and unburned sites. Linyphiidae (Erigoninae and Linyphiinac), ton). W, directa (O.P.-Cambridge) and IV. trkor- Lycosidac. and Gnaphosidae were numerically nijr lEtnerton). dominant in species number, with 31 (24 and 7), Several species seemed to lack habitat 9 and 6 respectively. specificity and were equally found in marked Species that actively colonized the burned sites numbers at hoth burned and unburned sites. Such [Deluded, among the lycosids, Pardusa hyper- species included the linyphiids. Sisis rotundas S7B MEMOIRS OF THE QUEENSLAND MUSEUM

(Emerton), Sisicotnts montanus (Emerton) and ACKNOWLEDGEMENTS Agyneta olivacea (Emerton); and the thomisid

Ozyptiia gertschi Kuratu. These generalist I thank Louise Filion and the staff of Centre species must be regarded as colonizer species d'etudes nordiques for working facilities and because of their common occurrence at burned generous help at Kuuijuarapik and Quebec City. sites. Although only in ii few cases {Pardosa Luc Sirois provided help at Lac Ekomiak, and hyperboreo and P. uintana) 'preferences' lo Heli Hurme participated in the field work. burned areas were statistically significant, about 30 of the 47 species caught at burned sites can he LITERATURE CITED regarded as potential colonizers in the subarctic postfire forests investigated. ft&CKW1TH,R.C& WERNER, R. A. 1979. Effects of (ire on arthropod distribution. Pp. 53-55. In Species that clearly avoided burned sites were Viereck, L.A. and Dyrness, C.T.

103- < (Koponen, 1988), 1 DIVERGENT TRANSFORMATION OF CHELICERAE AND ORIGINAL ARRANGEMENT OF EYES IN SPIDERS (ARACHNIDA, ARANEAE)

OTTO KRAUS AND MARGARETE KRAUS

Kraus, O. and Kraus, M. 1993 11 II: Divergent transformation of chelicerae and original arrangement of eyes in spiders (Arachnida, Araneae). Memoirs of the Queensland Museum 33(2): 579-584. Brisbane. ISSN 0079-8835.

In various higher taxa of the Araneae (e.g., Mesothelae, Migidae, Hypochilidae), the chelicerae and their fangs show an intermediate position between those commonly called orthognathy and labidognathy. This stage is considered to form part of the ground pattern of spiders; accordingly, it is called plagiognathy (new term). It is concluded that plagiognathy gave rise to orthognathy and labidognathy as divergent adaptational developments. In most instances, plagiognathy is correlated with the maintenance of the original (plesiomorphous) arrangement of the lateral eyes (= ALE + PLE + PME) in triads or semi-triads. The previous assumption that orthognathy and the arrangement of eight eyes in two subparallel rows are characters that were already present in ancestral spiders is refuted. Bei verschiedenen hoheren Taxa der Araneae (z.B. Mesothelae, Migidae, Hypochilidae) weisen die Cheliceren sowie deren Klauen eine intermediare Position zwischen Orthognathic und Labidognathie im ublichen Sinne auf. Diese Anordnung wird als Teil des Grundmusters der Echten Spinnen angesehen und hierfiir die neue Bezeichnung Plagiognathie eingefiihrt. Von diesem primaren plagiognathen Zustand werden sowohl die Orthognathic als auch die Labidognathie als divergente Entwicklungen mit unterschiedlichem Anpassungswert ab- geleitet. In den meisten Fallen ist Plagiognathie korreliert mit dem Erhalt der urspriinglichen (plesiomorphen) Anordnung der Seitenaugen (VSA + HSA + HMA) in Form von Triaden oder Semi-Triaden. Die bisherige Annahme ist nicht langer aufrecht zu erhalten, wonach Orthognathic und die Anordnung von 8 Augen in zwei Querreihen als Komponenten des Grundmusters der Spinnen angesehen worden waren. \Z\Araneae, plagiognathy, orthognathy, labidognathy, lateral eyes, triads.

Otto Kraus, Margarete Kraus, Abteilung fiir Phylogenetische Systematik, Zoologisches Institut und Zoologisches Museum, Universitat Hamburg, Martin-Luther-King-Platz 3, D-2000 Hamburg 13, Germany; 12 November, 1992.

It is generally believed that the chelicerae in thognathy had been maintained in a considerable spiders can be arranged in either of two different number of higher taxa. ways, described by the terms orthognathy and Simon (1892: 64, 82) pointed out that the labidognathy. Orthognathy is commonly thought Liphistiidae and Migidae had arrangements of the to represent the primitive (plesiomorphic) char- chelicerae that did not fit very well into the acter stage (Foelix, 1982:3;PlatnickandGertsch, generally accepted orthognathy/labidognathy 1976: 13). At first glance, this view seems to be scheme. Later authors ignored such deviations', supported by the fact that a strictly orthognathous however, and continued to base the distinction of arrangement of these mouthparts is also present two major subtaxa of spiders-Mygalomorphae in the outgroup of the Araneae, i.e., in the (=Orthognatha) and Araneomorphae (=Labido- Amblypygi. Hence, the idea that orthognathy is a gnatha) on different positions of the chelicerae. plesiomorphic feature seems to be the most par- Kaestner alone remarked on the intermediate ar- simonious explanation. Accordingly, labidog- rangement of these mouthparts in Actinopodidae nathy is regarded as a derived (apomorphic) and in Hypochilus, but apparently he too con- feature. Kaestner (e.g., 1952, 1953a, b) presented tinued to adhere to the typological ortho- arguments supporting the assumption that gnathy/labidognathy concept. One main aspect of labidognathous, i.e., cooperating chelicerae had his study was therefore to classify the chelicerae various functional advantages. He produced a in Hypochilus as orthognathous or labido- model (Fig. 1) illustrating the transformation of a gnathous. 'primitive' orthognathous arrangement into the In this paper, we will present relevant facts, labidognathous position. However, it is difficult most of them already known for decades, and to imagine how this could have happened discuss conclusions allowed by alternative con- gradually, and Kaestner did not explain why or- cepts. MO MhMOlXS OF THE QUEENSLAND MI IS

FIG. 1. Transformation of chelicerae as supposed by Kaesrner. A, orthognathy, left ehelicera omitted, Front of prosoma nearly vertical. B, labidognathy. dotted lines and arrows indicate how front of prosoma (with chelicerae) shifted from original vertical to a horizontal position (rotation of basal segments of chelicerae not indicated). C, suggested economy of relatively small cooperating labidognalhous chelicerae compared with a single orthognathous chclicera (hatched); both seize objects of same size. -i Prom Kaestner, 1953b).

This approach leads directly into a critical discussion of another generally accepted dogma in arachnology—that the eight eyes present in the ground pattern in spiders were originally arranged in two more or less parallel rows. In the nearest outgroups (Amblypygi, Uropygi). however, the arrangement of these eight eyes is quite different: the lateral eyes form triads on both sides of the prosoma- Such triads also occur in certain FIG. 2. Position of chelicerae and fangs, lateral and spiders. We therefore also plan to adopt a some- ventral views. A-B, Mesothelae: Liphisttus sp. C-D, what unconventional approach, discussing the Atypidae: Atypus affiriis. E, Migidae: Migas quintus. F, Aclinopodidae: Missulena accuiona. G, question as to whether the presence of such triads Hypochilidac: Hvporhiius gerlscht H, H. ihorellj- in various subtaxa of the Araneae could be a (A-B from Millot, 1949; CD, F, H from Kaestner, persisting plestomorphic character expression 1952; E from Wilton, 1968).

RESULTS AND INTERPRETATIONS ments in the Migidae, referring to 'chelicerestxes courtes, convexes a la base, mais ensuite brusque- Chelicerae ment inclinees, presque verticalement ...* (see FACTS. Fig. 2e). Kaestner (1952: 118) studied Sason robustum (O. P.-Cambridge 1 883) as a repre- Orthognathy is commonly regarded as t sentative of the Barychelidae and characterized plesiomorphic. However, precisely those spiders the chelicerae as short and subvertically inclined that have the greatest number of plcsiomorphies In the same paper, Kaestner demonstrated that in common (Platnick and Gcrtsch, 1976) do not obliquely arranged chelicerae were also present show an orthognathous position of their in the Actinopodidae (Fig. 2f); he described the chelicerae: in the Liphistiidae (Figs 2a-b) the situation in Missulena occatoria (Walckenaer, basal segment (paturon) of these mouthparts is 1 ) and concluded: 'I cannot see any biological relatively short, inflated and obliquely posi- 805 reason for such conditions. But as torsions of this tioned. Further, the longer axis of this basal seg- kind play an important role in the origin of ment is orientated obliquely downwards, and not labidognathy, it is interesting to sec that they [the horizontally and paraxially as in 'true* Orthog- torsions] also occur in the Orthognathia' natha (Figs 2c-d). The corresponding position of may (transl. from German). the fangs is also oblique, and anything but paraxial, in contrast to the position of the fangs in It is worth mentioning that the chelicerae even Atypus, for example (see Simon, 1892: 64). of the oldest known spider, Aiterocopus fimbriim- The same situation as in Liphistius'is also found guts (Shear. Selden and Rolfe. 1987) (Middle- jn various subtaxa of the Mygalomorphae. In Devonian), had short basal segments and also

1892, Simon (. 82) described similar arrange- fangs (Selden et a/., 1991 , e.g., plate 1. figs 6-8). TRANSFORMATION OF SPIDER CHELICERAE 581

of the chelicerae. Accordingly, the plagio-

gnathous position present in the ground pattern - t the Araneae has been secondarily transformed in two different directions, both apomorphic character states: orthognathy and labidognathy (Fig. 3), Wc see various arguments in support of hypothesis:

a) It explains why orthognathy is not encountered in the Mesothelae [Liphisiiu^ Hep- mtht>Li\. bt The absence of orthognathy in representatives of several mygalomorph families is explained. Ib-j fact that the Hypochilidac arc not FIG. 3. Orthognathy (left) and labidognathy (right) ets is BpOsnorphk ctiarft s derived from plagiog- labidognathous explained by the simple as- nadiy (bottom). sumption that the original plagiognathy has been maintained in this group of the Araneomorphae. Unfortunately, their original position un- is Nonetheless, in all other Araneomorphae (this known. means in the Neocribellatae, the sister taxon to Chelicerae with an oblique position are also the Hypochilidae) labidognathy has been achicv in taxon to found a that unquestionably belongs ed and is regarded as an apomorphy of this taxon. the Araneomorphae (= Labidognatha auct.y. the This conclusion is not invalidated by the fact that Hypochilidac (Figs 2g-h>. Again, it was Kaestner superficially orthognathous arrangements orig- (1952: 132) who studied details. He concluded inated secondarily in a few sexually dimorphic dial the mouthparts in Hypochilus were of the araneomorph ta.xa (e.g., in males of the salticid orthognathous type in construction and expressed genus Myrmarachne). the view (Kaestner, 1952: 114)that 'the majority d) Kaestner s typological and entirely theoreti- of important characters present in Hypochilus is cal model suggesting how a supposed transition accordance with the Orthognathia, whereas the 111 from orthognathy to labidognathy could come number of features present in Labidognatha only about (Fig. 1 ) is replaced by a new concept (Fig. is very low. For this reason, I must remove the 3). This postulates divergent and gradual evolu- genus from the suborder Labidognatha and either tionary change of the ground pattern, that is to it in the set it in place Orthognathy or up a say, of plagiognathy. 1 suborder of its own (transl. from German). e) Kaestner' s complicated assumption that INTERPRETATION obliquely arranged chelicerae originated in pQfg Kaestner maintained that labidognathy was an lie! both in the Mygalomorphac and the Araneo- .ukanced character state, which had developed morphae is replaced by a simple, comprehensive from an orthognathous ground pattern by gradual hypothesis transformation (Kaestner, 1953a: 60; Fig. 1). He The only remaining conflict seems to be that felt that Hypochilus (and the Hypochilidae) reflected in the strictly orthognathous position of should be regarded as transitory stages and ex- the chelicerae in the most closely related out- plained the oblique position also present in the groups of the Araneae (Amblypygi, Uropygi). If

Baryehelidae and Actinopodidae as a parallel our 'plagiognathy hypothesis" is correct, it must development. Furthermore, he regarded the be assumed that orthognathy in the Araneae is a *semi-orthognathous' chelicerae in Dysdera different and thus independent secondary (Dysdendae) as intermediate. Kaestner thought, development within the mygalomorph spiders then, that various transitory stages still existed There is no question but that this contradiction forming a "phylogenctic link' between the two needs some further examination. extreme character states. Preliminary investigations have already sug- We reject this judgement based on typology. gested that orthognathy in Amblypygi may be and postulate that an oblique position of the different from orthognathy in spiders: the basal chelicerae, including the fangs, really represents segment in amblypygid chelicerae has a long. the plcsiomorphic situation (Fig. 3). As a new stout apoderne at its proximal dorsolateral border, term is needed, we would like to suggest which reaches deeply into the broad. Hal 'plagiognathy' to designate this original position prosoma. This peculiarity is lacking in plagio- 5S2 MEMOIRS OF THE QUEENSLAND MUSEUM

gnaihous and also in orthognathous chelicerae of - spiders. We expect that more detailed studies on the functional morphology, including the mus- " culature, will demonstrate that orthognathy in uropygids and amblypygids differs from ortho- i ' gnathy in spiders. This would support our view n and could perhaps constitute point f) in the list of 1 positive arguments above. • i

O I Eves C Surprisingly, plagiognathous spiders (for example Mesotheiae, Migidae, Hypochilidac) share a special arrangement of the eyes (Fig> 4g, e, c): anterior lateral, posterior lateral and posterior ' median eyes are grouped closely together. This E prompts the following remarks on the question as to how the eyes were grouped in the ground FIG. 4. Position of median (black) and lateral eyes in pattern of the Araneae. Trigonotarbida. Amblypygi and Araneae. A, As designations widely used in taxonomie - descriptions (AME, ALE, PME, PLE) disregard nf prosoma (from Shear et at., 1987). B, Amblypygi: Damon sp. C, HypochiJidae:Hyptw:/ii/H.y the origin of these 'ocelli', some notes on the gerttchi. D, Atypidac: Atypus affinis. E. Migidae. homolgy of the eyes ot spiders may be ap- Poecilomgassp. F, Pholcidae; Pholcus circutaris. CJ, propriate to ensure that we understand each othei McMiIhelac: Liphisfius sp. H, Dysderidae; Dysifani the anterior median eyes (AME) will be called Sp I. Aeelenidae: Agelena sp. (Not to scale). 'median eyes' by us, as they are homologous with the median eyes of other arthropods (for example Araneae * Amblypygi + Uropygi + Schiz-omida). 4 those in Xiphosura, ocelU' in insects, and the Devonian trigonotarbids had the usual two three components of crustacean naupHus eyes). median eyes, and the lateral eyes were repre- other eyes, three on each side, will be called AH sented by up to 9 ( 12 ?) lenses (Fig. 4a). Three of 1 'lateral eyes (ALE + PLE + PME), as they are these were major lenses, while the others were homologous with the paired original compound minor lenses arranged in the interspace between eyes in arthropods, for example, in xiphosurar.v the major ones (Shear et fll, 1987). This kind of PACTS transformation of the original compound eyes triad lateral eyes It is commonly believed that an arrangement clearly indicates that a of major pul- in two transverse rows of eight eyes is plesmmor- is a feature of the ground pattern of the monaies as a whole; accordingly, the loss of the phir Only two weak aspects support this s however: (a) there is no reason at all to doubt that minor lateral eyes could be regarded as an aut- the presence of eight eyes forms part of the apomorphy of all other pulmonates, including araneid ground pattern, and (b) their arrangement spiders. This secondary reduction of the minor in two rows is widely observed both in the lateral eyes may explain why most triads are not Mygalomorphae (for example the Actinopod- perfectly closed, not even in amblypygids (Fig. idae; see Simon. 1892: 79, figs 81-83) and in the 4b). Araneomorpha (for example Araneidae. Euspar- The peculiarity 'lateral eyes in triads' is com- assidac, Thomisidac ). monly used as a character in spider identification On the other hand, lateral eyes more or less keys, but as far as we can tell, its potential bearing distinctly grouped in triads occur in various on phylogeny has never been discussed. Could it groups of spiders. Mesotheiae, Vligidae and be that triads of lateral eyes are part of the ground Hypochilidac have already been mentioned. Al- pattern in the Araneae'7 most perfect triads occur in Pholcidae (Fig 4T). A survey of how the lateral eyes are positioned The same is true of Amblypygi (Fig. 4b) and in representatives of higher taxa of spiders shows Uropygi, the direct outgroups to spiders! that almost perfectly 'closed' triads (as in phol- The arrangement of the eyes in the extinct cids) are rare. In most instances, the three lateral Trigonotarbida deserves special attention. Ac- eyes on each side are somewhat dissociated. In cording to Selden etal. ( 199 1 : 254), they form the addition to the Mesotheiae and Migidae already sister group of all other pulmonale taxa (= mentioned, we should also like to draw attention TRANSFORMATION OF SPIDER CHELICERAE 583

to the Atypidae (Fig. 4d) and to the illustrations (e.g., 1894: 517), that is to say, the occurrence of in Raven's comprehensive study of the Mygalo- diads can be explained as part of the original triad. morphae (1985). In many cases the posterior lat- To some extent, the question remains open, as eral and the posterior median eyes are closely to how it is possible to distinguish between eye connected, with some distance between them and positions that can be regarded as more or less the anterior laterals. Hypochilus shows slightly modified triads and other positions, with secon- dissociated triads (Fig. 4c). The Dysderidae (Fig. darily approximated ALE and PLE. 4e) and Oonopidae are six-eyed spiders, having the median eyes completely reduced. In dys- PERSPECTIVES derids, the lateral eyes are closely grouped together, the the resembling arrangement of Apparently, plagiognathy is part of the ground laterals in the Mesothelae. In many groups within pattern of the Araneae. Developments in the the Araneomorphae, diads are present instead of directions of orthognathy and labidognathy can triads. They are formed by the ALE + PLE, with easily be explained as divergent evolutionary the separated. This arrangement can be PME changes (Fig. 3). The question therefore arises of found in Austrochilidae and especially in most how these might be correlated with functional Theridiidae and Linyphiidae, for example. Diads aspects. As an impetus for further discussion, we also occur in groups characterized by a secondary propose the following working hypotheses: loss of the PME, such as Scytodidae. a) In the Mygalomorphae, orthognathy may be INTERPRETATION correlated with the capture of prey on the ground. The assumption that eight eyes arranged in two Under such conditions, the two parallel fangs of transverse rows were already present in the the chelicerae can easily penetrate the victim on ground pattern of the Araneae is not supported by a substrate like two stabs of a dagger. It seems any concrete fact; nor would this at all correspond remarkable that a semi-orthognathous position of with the situation in the nearest outgroups. It the chelicerae has originated secondarily in the 1 would mean that triads and triad-like arrange- Dysderidae: they kill woodlice on the substrate. ments of the lateral eyes in spiders were classifi- b) In the Araneomorphae, the origin of labido- able as parallel developments (homoplasies). gnathy may be correlated with the evolution of This is unlikely. In accordance with the position capture webs (sheet, frame, orb webs etc.). These of the eyes in the Amblypygi and Uropygi, we could make it more efficient to bite the prey with expect that the laterals were primarily grouped as two opposing chelicerae or fangs, whereas triads (ALE + PLE + PME). This hypothesis is plagiognathous and, even more, orthognathous supported by five arguments; chelicerae might not penetrate but rather push a) Triads of major lateral eyes (lenses) already away the victim: there is no longer any substrate existed in Devonian Trigonotarbida; hence, triads 'supporting' prey animals. apparently form part of the ground pattern of all c) Plagiognathy and the maintenance of lateral pulmonates among arachnids. triads or semi-triads of eyes apparently form part b) The postulated configuration is in good of the ground pattern of spiders; these features are agreement with the arrangement of the eyes in the confined to more 'primitive' groups. The direct outgroups. presence and the various types of transformation c) Triads and semi-triads present in various of these two characters should be integrated into groups of the Mygalomorphae and also of the current concepts on the phylogeny of the Araneae Araneomorphae must no longer be explained by (see, for example, Raven, 1985; Coddington, assumed parallel origin. 1990). At present, our view of features of an d) Various types of somewhat dissociated lat- araneid ground pattern and succeeding evolution- eral eyes can be explained by a secondary separa- ary changes seems to be somewhat at odds with tion of the ALE or of the PME from the others, various published cladograms; they hence could which frequently remain in contact with each be partially wrong. We feel that this conflict may other. be due to the possibility that characters assumed e) Simon's 'oculi laterales utrinque contigui* to be synapomorphies in various cladograms (see,

1 But see Kaestner (1953a: 62). He believed that the position of the chelicerae in Dysdera was a 'phylogenetic link' between orthognathy and labidognathy. Unfortunately, he was not aware that the first postembryonic stages were nearly labidognathous, with relatively shorter basal segments and only slightly oblique fangs (pers. observ.). In Dysdera, the final semi-orthognathous position was gradually acquired in later instars. S84 MEMOIRS OF THE QUEENSLAND MUSEUM

e.g., Platnick and Shadab, 1976, fig. 1; Raven, der Lahidognathie. Zoologische Jahrbiicher 47-68. 1985, fig. 1) may well turn out to be symples- (Anatomie)730): 1953b, Die Mundwerkzeuge der Spinnen, ihr Bau, iomorphies; e.g., Raven's characters 35 (eyes ihre Funktion und ihre Bedeutung ftir das System. spread widely across the prosoma; same as Plat- 3. Teil: Die Cheliceren der Labidognatha (Di- nick and Shadab* s character 1) and 36 (male pneumones). Mitteilungen aus dem Zoologis- pedipalps: conductor of bulb present; see Kraus, chen Museum in Berlin 29(1): 3-54. figs 12, 14-16). 1978, KRAUS, 0. 1 978. Liphistius and the evolution ofspider genitalia. Symposia of the Zoological Society of ACKNOWLEDGEMENTS London 42: 235-254. MILLOT, J. 1949 Ordre des Araneides (Araneae). Pp. 589-743. In Grasse, P.P. (ed.), Traite de Zoologie. We are indebted to Prof. Dr. H.M. Peters, Anatomic Systdmatique, Biologic (Masson: Tubingen, and Prof. Dr. H.W. Levi, Cambridge, Paris). Massachussets, for critical advice. We express PLATNICK, N.I. & GERTSCH, W.J. 1976. The subor- cladislic American our thanks to Dr, W.A. Shear, Hampden-Sydney, ders of spiders: A analysis. Museum Novitates 2607: 1-15. who draw our attention to the situation in the PLATNICK, N.I. & SHADAB, M.U. 1976. A revision Trigonotarbida and in other fossil arachnids. Dr. of the mygalomorph spider genus Neocteniza R. Graz, kindly provided early posi- Horak, (Araneae, Actinopodidae). American Museum cmbryonic stages of Dysdera. Novitates 2603: 1-19. RAVEN. R.J. 1985. The spider infraorderMygaloniDf- LITERATURE CITED phae (Araneae): ciadistics and systematic*. Bul- letin of the American Museum of Natural History 182(1): 1-180. J.A. 1990. Ontogeny and homology CODDINGTON, SELDEN, P.A., SHEAR, W.A. & BONAMO, P.M. in male palpus of orb-weaving spiders and the 1991. A spider and other arachnids from the their relatives, with comments on phytogeny Devonian of New York, and reinterpretations of (Araneoclada: Arancoidea, Deinopoidea). Smith- Devonian Araneae. Palaeontology 34(2): 241 - sonian Contributions to Zoology 496; 1-52. 281. FOEL1X, R.F. 1982. 'Biology of spiders'. (Harvard SHEAR, W.A., SELDEN, P.A., ROLFE, W.D.I., University Press: London). Cambridge and P.M. & GRiERSON. J.D. 1 987. New BONAMO ? 306pp. terrestrial arachnids from the Devonian of Gilboa, KAESTNER, A. 1952. Die Mundwerkzeuge der Spin- New York (Arachnida. Trigonotarbida). Amer-

nen, ihr Bau, ihre Funktion und ihre Bcdeulung ican Museum Novitates 2901 : 1-74.

fur das System. 1 . Teil: Orthognatha; Palacocrib- SIMON, E. 1892-1895. 'Hisloire naturelle des ellata. ZoologischeJahrbiicher(Anatomie)72(l): Araignees. Vol. 1\ (Librairie eneyclopedique de 101-146. Roret: Paris). 1084pp. 1953a. Die Mundwerkzeuge der Spinnen. ihr Bau, WILTON, C.L. 1968. Migidae. In Forster, R.R. *Thc ihre Funktion und ihre Bedeutung ftir das System. spiders of New Zealand, vol. U*. Otago Museum 2. Teil: Herleitung und biologische Bedeutung Bulletin 2: 73-126. .

POLYNESIAN THOMISIDAE - A MEETING OF OLD AND NEW WORLD GROUPS

PEKKA T LEHTINEN

Lehlinen. P.T. 1993 1J II. Polynesian Thomisidac- a meeting of oLd and new worJd groups.

Memoirs of the Queensland Museum 33(2): 585-591 . Brisbane. ISSN 0079-8835.

The Polynesian thomisid fauna is postulated as consisting of an Hawaiian-cast Polynesian New World group, living mainly in isolated populations in the mountains and of representatives of two western lowland groups originating from Australia and Southeast Asia. The former group has apparently spcciaied into numerous endemic species, while the latter groups arc represented by a singLe. widespread species and a rare Tongan species, respectively. The ranges of the eastern and western groups do not overlap. The species of New World origin have been described or previously attributed to Mi$umena> Misumenops, or

Ss'tiaema. All such species are included here in Mecaphesa Simon, 1 900 with the type species

from Hawaii. MisumenopsF.O. Pickard-Cambridge, 1 900 has the type species from Eastern Brazil and has no close relatives in Polynesia or in the Old World. A widespread Old World group is also recognized here and tentatively included in Massuha which appear to be related to the New Gnrnean Lo.xonoretes. Dictea as currently recognized is polyphylelic and the

species occurring in west Polynesia {Diaeapraetexta (L. Koch, 1 865)) is postulated to belong to a group requiring a new generic delimitation and name. Hedana subtilis L. Koch, 1874, also of Asian origin is here regarded as having affinity with Tharrhaka. The poorly described thomisid species of the isolated, southernmost island group of Polynesia, Rapa Island has not been studied. [2/iranent, Thomisidae, Polynesia, biogeography

Pekka £ Ishtinen, Zoological Museum, Utliversff) nf Turku, 20500 Turk,.-. Ftntawll 11 March, 1993.

Many zoogeographical discussions, including Theridion, Huhma, or Liucauge. Some recent spiders, are flawed because of poor taxonomy. papers on Polynesian spiders have been publish- The zoogeography of the Polynesian spider fauna ed (eg. Maiples, 1955a, b. 1957, 1959, I960, has been discussed by Berian"d( 1927, 1928, 1929, 1964; Berry and Bcatty, 1987; Realty and Berry, 1930. 1933, 1934a. b. c. 1935a. b. 1937. 1938a, 1988: Realty et al, 1991). I have previously b, 1942, 1947), but his discussions were based on discussed some aspects of the spider zoogeog- unreviscd taxa. His conclusions were often af- raphy of the Pacific region (Lehtinen, 1980). fected by misidentifications and unevenoess of Polynesian thomisids have also been described data available. Most spiders for these papers con- by Strand (1913). sisted of assorted samples made by non- The evolution of Polynesian Thomisidae has specialists and many were synanthropic species found near villages. resulted in the most striking example of local speciation of spiders in Polynesia. The I have been working towards a zoogeographi- Thomisidae discussed comprise only the sub- cal synthesis of the Polynesian spider fauna for family Thomisinae in the sense of Suman (1970) ten years. Extensive personal field work in moun- and various other authors. The Philodromidaeare tain tops, but also in the disturbed zone has been nota sister groupof Thomisidae, but rather of the the most important method in the elimination of anthropochorous dispersal and distinctly Heteropodidae. geological history oi the Polynesian ar- synanthropic species from all speculations on the The origin of the fauna of natural habitat-. chipelagoes is now well known (Wilson, 1963; Duncan and McDougall. 1974; Dalrymple etai, The taxonormc revision of all families present 1975). An ancient continent of Pacifica has been see ms to be necessary for any valid proposed (with differing placement and size) zoogeographical conclusions. As the first step 1 marginally affecting to the historical zoogeog- have carried out a "generic" revision of Polynesian families which allows the placing of raphy of Polynesia (e.g. Nur and Ben-Avraham, most Polynesian species groups of spiders into 1 977 ; Craw, 1 983). Most geophy sicists agree that named or still unnamed groups of supraspecifie the Polynesian islands are not parts of ancient taxa instead of zoogeographical ly useless con land masses bfuken by processes of the plate ccptfl such as the "worldwide' Misumenops. tectonics, but rather chains of current islands and u

5ft MEMOIRS OF THE QUEENSLAND MUSEUM

seamounts representing former islands (Dal- apparendy retained many plcstomorphie charac- rympleefaL 1975). ters Some groups with striking individual adap- The origin of the Polynesian fauna therefore tive modifications (e.g. Heriaeus and Runcinia) can be explained only by long distance dispersal may be closely related to this group. from different directions (Gressitt and There are most probably many other endemic Yoshimoto, 1963) and partly by intrapolynesian species of Thomisinae in the mountains of French speciation processes within the island chains ( cf Polynesia, but a revision is excluded here. also Carson, 1984;Fosberg, 1991) The use of the The definition and delimitation of thomisine ba$ic principles of the vtcariance hiogeography genera has been based traditionally on a few

< Nelson and Platnick, 1981) will be essential for adaptive characters, including number, length, explaining the origin of groups of biota with and tvpe of setae on the carapace (Simon, 1895; complex patterns in their recent ranges. Craw's Mello-Leitao, 1929, Schick, 1965; Tikader. (1978) variant of panbiogeography, latex named 1980; Dondale and Redner 1978; Levy 1985: spanning-treebiogeography by Platoick and Nel- Ono, 1988), but little attention has been paid to son ( 1 988) is a useful method for comparisons of general patterns of the genital organs and type of area including also permanently oceanic island sexual dimorphism. In contrast to conventions in groups and it can be recommended for the the taxonomy of other spider groups, the naming analysis of many other groups of spiders. When of individual setae of the thomisid carapace has patterns of distribution are very simple and been used by some recent specialists (Schick. anthropochorous dispersal in historical time has 1965; Dippenaar-Schoeman. 1983). This ter- not thoroughly obscured the original patterns (cf. minology is widespread in aearinc taxonomy conclusion* with Stoddart. 1968), can be made Some genera have been very obscurely defined Craw's method such as have actually long been and therefore all catalogues list them as being used zoogeographers (e.g. Gressitt, 1961) by very widespread and species rich, e.g. Misumeixa, concept parsimonious area before the of 'the most Misumenops, Diaea, and Synaema. All these relationship' was defined and named. generic names appear in thomisids described or In spite of the current taxonomic confusion of listed from Polynesia. Even a superficial com- the Thomisidae on a global scale some parison of the descriptions or type material from generalisations on the zoogeography of the fami- many species of these genera (L. Koch, 1874; ly are possible in the Pacific area. This paper Kulczynskj. 191 1; Chrysanthus, 1964; Tikader. presenis the suggested relationships and 1980) reveals tl^at '.hey are typical "w aste-baskef zoogeography of the Thomisidae of Polynesia groups, v. here most species are not closely related according to rich new material and results of my to the respective type species unpublished revisional work. The phylogenetic classification of the west Polynesian Diaea and the east Polynesian TAXONOMIC REMARKS M\um*nops has been time-consuming, as all basic taxonomic work on Indo-Pacific and The nominate subfamily of Thomisidae should Neotropical Thomisidae was done before modern ailed Thomisinae, although the name taxonomic principles and methods became estab- Misumeninac has been widely used, also, e.g lished. Most structural characters used as generic recently by Dippenaar-Schoeman (1983). Two criteria in Thomisidae seem to be minor conver- groups of greenish or yellowish species without gently evolved adaptations. The type species of abdominal modifications are easily recognizably all three large widespread genera in question, one with conspicuous modifications in the ocular Misiimenops, Diaea, and hlisumeiui, are atypical area [Thomizus-group), the other without such or 'peripheral' species, not closely related to the modifications yMisumefuj -group). The limitation Pacific species. Actually the placing of many of thomisinc groups has been vague, Simon tropical species in these three genera have been

(1895) originally listed Misumena > Heriaeus, and repeatedly changed, depending mainly on em- Diaea in different tribes, while at the other ex- phasis laid on single adaptive characters, e.g. treme, the Mis umena-group of Dtppcnj type and pattern of setae on carapace, pattern of Scboeman (1983) includes not only Thomism, leg spines, eye pattern, etc. hut also Runcinia, The section of the carapace is variable in Ihe No phylogeny of tbomisine groups is known Polynesian groups of Thomisidae. Nevertheless, and detailed discussion fa beyond the scope of in Mecaphesa, sympatric species may be best i Ins study. However, the Misumena-gToup has identified by differences in length and density of -

POLYNESIAN THOMISIDAE 587

S3Z

of Indo-Pacific "Diaca" —* total range "=rj It ill! total range or Mecaphesa range or "Diaca" practexta o range of Mecaphesa in Pacific Islands

y: range and origin of Polynesian *Tharrta lea r

FIG. 1 . Geographic ranges and source of Polynesian thomisid groups.

the carapace setae. The shape of the carapace is but by Suman (1970) in Misumenops (14 spp),

variable in Mecaphesa, while the shape of the Synaema (1) and Mecaphesa (3 spp.). I have abdomen is variable in both 'Diaea and checked the type material of all Hawaiian Mecaphesa, even within one population. thomisids preserved in ihe Bishop Museum and, The colour pattern is variable also, although the in my opinion, both male and female genitalia of 'usual' colouration for most species provides a Synaema naevigerum Simon, 1900 are much reasonable guide to identification, if large closer to the genitalia of all Hawaiian populations are available. * Misumenops' than those of the type ofSynaernOy S. globosum (Fabricius, 1775) from Europe. The POLYNESIAN THOMIS1DS OF NEW relative width of the ocular area is certainly a WORLD ORIGIN parallelism in true Synaema and the Hawaiian 1 * Synaema . The blunt setae of Mecaphesa s. str. Most east Polynesian ihomisids have been long have been independently modified, and the three included in Misumenops F.O. Packard species constitute a sister group of the Hawaiian 9 Cambridge, 1900 (Berland, 1933. 1934b, 1942; 'Misumenops and "Synaema" together. The geni- Suman, 1970), although Roewer (1954) trans- tal organs of both sexes are again more or less ferred all Hawaiian species to Misumenoides F.O. similar and not at all related to Oxyptila or Pickard-Cambridge, 1900. Heriaeus. as claimed by Simon (1900). Suman Many Hawaiian and North American species of (1970) was not familiar with the Palaearetic

Misumenops sensu Schick ( 1 965) were originally thomisids and had no opinion on this matter, but described xnMisumena orDiaea, and Neotropical be published useful drawings of the genitalia of species also in Metadiaea Mello-Leitao, 1929. all Hawaiian thomisids. This group of The adaptive radiation of Hawaiian 'Misumenops' is also present in Japan, as Thomisidae indicates that essential changes in the Misumenops kumadai Ono, 1985 and in western shape of the carapace are possible withoul other North America at least 13 species (re/er-group). than minor changes in the male palpal structure. listed by Schick (1965) in Misumenops The Hawaiian thomisid species were listed by (Misumenops). Misumenops inclusus Banks,

Simon (1900) in Misumena (6 spp.), Dlaea (2 1 902 from Galapagos Islands and M. sjoestedti spp.), Synaema (4 spp.), and Mecaphesa (2 spp. ), Bcrland, 1924 from Juan Fernandez Islands arc 538 MEMOIRS OF THE QUEENSLAND MUSEUM

SE Asian/Ansiralian origins (western lowlands) Neotropical origins (eastern highlands) Uncertain origin (Rapa)

' lUtua fttnetiXlQ belongs in probably new group Species previously attributed lo Mtsumena. Mi:,urncnops rupacmii Misumenops and Synaema HedunasubtUis belongs y&Tharrhalta

TABLE J. Hypothesis of generic placement and zoogeographic origins of species groups of Thomisidac represented in Polynesia.

additional members of this genus in the Pacific and with genital organs of both sexes close to regroo, Runcinia. Their male tibial apophysis is similar to including the characteristic ribbed The generic name Misumenops is here reserved Mecaphesa, for the group of Neotropical species that are tip. The New Guinean Loxoporetes known only the female is probably related to, or even unambiguously related (0 hi. maculissparsa by (Keyserling. 1891), a species with a well congeneric with this group, of which most species described as developed tutacular process in cymbium and a have been Misumena. This group be Massuria Thorell, 1 887 complex of tibial apophyses that is widely dif- will probably named ferent from any Pacific species. M. pollens and it might be a plesiomorphic sister group of (Keyserling, 1880) and M. pallida (Keyserling, the widespread and widely sympatric Rum. 1880) were recently revised by Rinaldi (1983) without comparison to the type >peeies. The>e POLYNESIAN THOMISIDS OF OLD widespread Neotropical species arc not close to WORLD OR AUSTRALIAN ORIGIN Si. maculissparsa, but they may remain in the same genus. the other luino, the concept of On The greenish thomisids from Samoa and Tonga Mecaphesa is here widened to also include raosl islands have been described as different species Polynesian, some other Pacific and many north (L. Koch, 1874; Rainbow, 1902: New Hebrides; American 'Misumenops, the more pie si fr- Strand. 1913) all referred to Diaea, A critical morphic branch of this genus. 'Misumenops' survey of several large populations reveals that rapaensh Berland (Borland. 193*) from the iso- there ?s only one widespread species, *D." lated Island with a terrestrial fauna of Rapa praetextaiL. Koch, 1865)with large infrapnpttla peculiar affinities probably belongs elsewhere. liOfttl variation in the colour pattern, but quite The widespread Holarctic M. tricuspidatus small variation in the structure of the geniial (Fabricius, 1775) is removed from this genus, but organs. In contrast to east Polynesian and its final generic placement must wait for a more Hawaiian thomisids this species lives in the complete revision of Thomisinae: Misumcnini. \ c ..'ctatiori of lowlands and is also common in Fiji The structure of the mate tibial apophysis, includ- and Vanuatu. ing '.he rnicrostructurc of its tip, is different from At least D. sttcta and D. limbata (Kulczynski, all other thomisids known to me. hi. japonicus 1911) within the widespread and common 1 is ve | Rosenberg and Strand, 906) a retail or Melanesian Diaea spp. as well as the cast a member of Diaea, while the asperaJus group of Australian D. multopunctata L. Koch, 1 874 and Misumenops (Schick, 1965) may belong to D. prasittti L. Koch, 1 876 are congeneric with D. the coloradensis group represents Metadiaea and ."xta. This group of Australian-Polynesian a distinct genus^ not close to Mecaphesa or * Diaea" deserves generic status, but until some Misumenops. ihornisine genera from the Indo-Pacific area The synonymic history of Metadiaea isconfus- have been revised the erection of a new genus oo, as the authors discussine this problem would be hasty. The type species of Diaea. (Toledo Piza, 1937; Caponacco. 195^ Rinaldi Araneus dorsums Fabricius, 1775, probably

1 983) have based their opinions on the data from together with some other Palaearctic species has gpeiees other than the type species, M. ft male and female genital organs resembling Mello- Lejtao, 1929 from Minas Gerajs, Brazil. I Heriaeus (Loerbroks. 1983). The deviating non- agree with Rinaldi (1983) in transferring the other genitalic characters of Heriaeus are adaptations species to Misumenopspallida- group, but not the to life in the desert. lype of the genus, and Metadiaea remains a valid Hedana subtilis L. Koch, 1874 was described American genus probably including also North from one male and a juvenile from Tonga,

American species. western Polynesia. Most probably it is not con- There is a widespread Southeast Asian-New generic with the Australian type species H. Guinean group with short scutate male abdomen gracilis L Koch, 1874 and several other species POLYNESIAN THOMfSIDAE 589

from Southeast Asia to New Zealand. It has nol tional collecting both in the lowland as well as in been compared with the type of Tharrhalea from the mountains of Rarotonga by myself* no N. Australia, but it seems, at least, to belong to Mecaphesa spp. have been observed. The ab- the same tribe as 71 macututa from New Guinea. sence of Diaea praetexta in disturbed lowland H. pallida Koch, 1876, described from juvenile habitats of Rarotonga seems to show that specimens from Tonga most probably is a anthmpochorotts dispersal of this species is not synonym of//, subtilis or even '0/ praetexta. In very effective, although its frequency in the more addition to these two very old records there is also western archipelagoes can be partly explained by a recent record of a subaduk female from Tboga. this type of short distance dispersal. Hedana and Tharrhalea have been catalogued in The Australian Thomisidae appear to have ar- Stephanopinae (Simon, 1895; Roewer, 1954; rived through Melanesia, where Fiji and Vanuatu Bonnet, 1957), but 7/.* subtilis belongs to share the same common lowland species, 'Diaea' Thomisinae. praetexta and several closely related species are

I have seen relatives of H. subtilis (? Thar- pTesent in New Guinea and Eastern Australia rhalea) in New Guinea and southeast Asia and The other west Polynesian thomisid species, the range of this unnamed group is more or less *TharrhaJea' subtilis belongs to a group that has similar, but possibly extending farther a wider range including southeast Asia. In spite northwards, when compared to thatof the 'Diaea of intensive field work in Samoa and Tonga praetexta-gmup. during 1991-92, (here are still no mon thomisids known in the western archipelagoes of ZOOGEOGRAPHICAL CONCLUSIONS Polynesia There are some other Polynesian spiders of

Neotropical origin (Anyphaenidac, B< The Polynesian thomisid fauna has apparently Theridiidae. etc.), but most spider families repre- arrived from two opposite directions, South sented in Polynesia are of Melanesian, southeast America and Melanesia. These two elements are Asian or New Zealand origin. not known to be mixed in any pari of Polynesia map in Fig. 1 i ACKNOWLEDGEMENTS New World element (Mecaphesa) probably first arrived in Hawaii, where an explosive This study is part of a long range programme speciaiion has taken place, resulting in 17 known on Indii-Pacific spiders, including field and species (Suman, 1970: in three genera), mc$l museum work dunng the last 20 years. This work having a small range up in the mountains. has only been possible through die generous Galapagos Islands (1 or several spp.) and Juan cooperation of numerous persons, who could nol Fernandez Islands constitute another possible be listed separately here. Useful comments were source of immigration for the Bast Polynesian also presented by two referees and an editor. The Mecaphesa spp. They have further evolved to at map was kindly drawn by Ms. Maija Mustoncn, least six, but probably more local endemic?. and the original English text was checked by Mrs There are also sympatrie montane species, al least Alice Moore. in the Marquesas Islands, but mosi probably also inTahiti. The majority of East Polynesian species LITERATURE CITED live on mountain tops, but a few species have

; iccassionally been recorded also in lowland. BEATTY. I A & BERRY. J,W. 1988. Four new The Oriental-New Guinean genus rjf the tea "t Paratheurrta (Arancac, Dcsidac) from Afiswnena-group* here tentatively called Mas* Ihc Pacific. Journal of Aiaclmnloyy 16: 339-341 sunn has not been found in the intervening BEATTY. J.A.. BERRY. J.W.. & M1LLTDGE. AF Melanesian archipelagoes and must be excluded J99I. The Linyphhd spiders of Micronesia and from possible sources of origin of the east Polynesia with notes on distribution and habitats. Polynesian Mecaphesa, although both groups Bulletin of the British Arachnological Societv 8: 265-274. belong to Misumemm. a group probably older BERLAND, L_ 1924. Araignees tie 1'ile de Palques et than any mid-Pacific archipelagoes. des Tics Juan Fernandez. The Natural History of There are no known ihomisids in the Central Juan Fernandez and Easter Island 3: 419-437. Polynesian islands (Cook Islands, Tokclau Is- 1927. Note sur les Araignces recueillies aux fles lands. Niue. etc.) (Marples, 1955b, 1957. I960, Marquises par le R.P. Simeon Delmas. Bulletin 1964) or in the low coral islands north of Samoa du Museum National d'Hisioirc Naiurelle. Paris (Rainbow, 1 897; Rocwcr, 1944). In spite o( addi- 38: 366-368. :

590 MEMOIRS OF THE QUEENSLAND MUSEUM

1928. Remarques sur la repartition et lcs affinites Radovsky, F.J., Raven, P.H. and Sohmer, S.H. des Araignees du Pacifique. Proceedings of the (eds), Biogeography of the Tropical Pacific. Pan-Pacific Science Congress, 3, Tokyo (1): CHRYSANTHUS, FR. 1964. Spiders from South New 1044-1054. Guinea VI. Nova Guinea, Zoology 28: 87-104. 1929. Araignees (Araneida). Insects of Samoa and CRAW, R. 1978. Two biogeographical frameworks: other Samoan terrestrial Arthropoda, vol. 8: 35- Implications for the biogeography of New 78. Zealand. A review. Tuatara 23: 81-114. 1930. Les Araignees des Ties avoisinant la Nouvelle 1983. Panbiogeography and vicariance cladistics: Z£lande et les relations entre l'Australie et 1'- Are they truly different? Syst. Zool. 32: 43 1-437. Amerique du Sud. Compte Rendu de Stances, DALRYMPLE, G.P., SILVER, E.A. & JACKSON, 1 90-94. Societe" de Biogeographie, Paris 930(60): E.D. 1975. Origin of the Hawaiian Islands. In: 1933. Araignees des Tics Marquises. Bulletin of the Skinner, B.J. (ed.), Earth's history, structure, and Bernice P. Bishop Museum 114 (Pacific En- materials. Readings from American Scientist: tomological Survey Publication 7): 39-70. 112-126. 1934a. Les Araignees du Pacifique. Contribution a DIPPENAAR-SCHOEMAN, A.S. 1983. The spider ]'6lude du peuplement zoologique et botanique genera Misumena, Misumenops, Runcinia and des Ties du Pacifique. Publ. Soc. Biogeographie Thomisus (Araneae, Thomisidae) of southern 155-180. 4: Africa. Entomology Memoirs of the Department 1934b. Les Araignees de Tahiti. Bulletin of the of Agriculture, Republic of South Africa 55: 1-66. Bernice P. Bishop Museum 113 (Pacific En- DONDALE, CD. & REDNER, J.H. 1978. The crab Survey 97-107. tomological Publication 6(21)): spiders of Canada and Alaska. Araneae: 1934c. Araignees de Polynesie. Annales de la Philodromidae and Thomisidae. The Insects and Societe Entomologique de France 103: 321-336. Arachnids of Canada 5: 255 pp. (Supply and 1935a. Nouvelles Araignees marquisiennes. Bul- Services Canada: Hull). letin of the Bernice P. Bishop Museum 142 DUNCAN, R.A. & MCDOUGALL, 1. 1974. Migration (Pacific Entomological Survey Publication 8(4)): of volcanism with time in the Marquesas Islands, 31-63. French Polynesia. Earth and Planetary Science 1935b. Les Araign6cs des Ties Marquises et la Letters 21: 414- 420. Biogeographie. Compte Rendu de Seances, FOSBERG, F.R. 1991. Pacific Oceanic Island Societe de Biogeographie, Paris 100: 27-28. Biodiversity in a geohistoric perspective. 1937. Comment les Araignees ont peuple le Pacifi- Proceedings of the XVII Pacific Science Con- que. Bulletin de la SocitSte Oceanographique 1 gress, Honolulu: 71-73. 77-80. GRESSITT, J.L. 1961. Problems in the zoogeography 1938a. Araignees des Nouvelles Hebrides. Annales of Pacific and Antarctic Insects. Pacific Insects de la Societe Entomologique de France 107: 121- Monograph 2: 1-94. 190. GRESSITT, J.L. & YOSH1MOTO, CM. 1963. Disper- 1938b. Liste des Araignees recueillis a Maupiti (Ties sal of animals in the Pacific. Pp. 283-293. In, de la Societe) par M. Ropiteau. Bulletin de la Gressitt, J.L. (ed.). Pacific Basin Biogeography. Societe etude Oceanographique 6: 47-48. L. E. 1871-1886. 1942. Polynesian spiders. Occasional Papers of the KOCH, & KEYSERLING, Die Arachniden Australiens, nach der Natur be- Bernice P. Bishop Museum 17: 1-24. schrieben und abgebildet. 1489 (Bauer and 1947. Araignees. XVI. in Croisieredu Bougainville pp. Raspe: Niirnberg). (Thomisidae 1874: 492-610 aux Ties australes franchises. Memoires du and 1876: 742-824). Museum National d'Histoire Naturelle, Paris 20: 53-54. KULCZYNSKI, W. 1911. Spinnen aus Sud-Neu- Guinca. Nova Guinea 9 (Zoology 2): 109-148. BERRY, J. & BEATTY, J. 1987. Biogeography of Pacific Island spiders: a progress report. American LEHTJNEN, P.T. 1980. Arachnological zoogeography Arachnology 36: 3. of the Indo-Pacific Region. Verhandlungen des 8. BONNET, P. 1957. Bibliographia Araneorum 2(3): Intemationalen Arachnologen-KongreB, Wien, 1927-3026. (Lcs Artisans de L'Imprimerie 1980:499-504. Douladoure: Toulouse). LEVY, G. 1985. Araneae: Thomisidae. Fauna Palaes- BOSENBERG, H. & STRAND, E. 1906. Japanische tina, Arachnida, vol. II: 1-1 15. (Israel Academy of Spinnen. Abhandlungen der Senckenbergischen Sciences and Humanities: Jerusalem). Naturforschenden Gesellschaft in Frankfurt a. LOERBROKS, A. 1983. Revision der Krabbenspin- Main 30: 93-422. nengattung Heriaeus Simon (Arachnida: CAPORIACCO, L. DI, 1954. Araignees de la Guyane Araneae: Thomisidae). Verhandlungen des Natur- Francaise du Museum d'Histoire Naturelle de historischen Museum in Hamburg (N.F.) 26: 85- Paris. Commentationes Pontificiae Academia 139. Scientiarum 16: 45-193. MARPLES, B.J. 1955a. Spiders from Western Samoa. CARSON, H.L. 1984. Speciation and the Founder Ef- Journal of the Linnean Society of London, Zool- fect on a New Oceanic Island. PP. 45-54. In, ogy 42: 453-504 POLYNESIAN THOMISIDAE 591

1955b, Spiders from some Pacific Islands. Pacific Misumeninae do Brasil (Araneae, Thomisidae). Science 9: 69-76. Revista Brasileirade Entomologia 27: 147-153.

1957, Spiders from some Pacific Islands II. Pacific ROEWER, C.F. 1944. Einige Araneen von Prof. Dr. Science 11:386-395. Sixten Bocks Pazifik-Expedition 1917-1918,

1959. Spiders from South Pacific Islands III. The Meddelanden fran Goteborgs Museum Zoologis- Kingdom of Tonga. Pacific Science 13: 362-367. kaAvdelningl04: 1-10. 1960. Spiders from some Pacific islands, part IV. ROEWER, C.F. 1954. Katalog der Araneae von 1758 The Cook Islands and Niue. Pacific Science 14: bis 1954, vol. IIA. (Institut Royal des Sciences 382-388. Naturellcs de Belgique: Bruxcllcs). SCHICK, R.X. 1965. The crab spiders of California 1964. Spiders from some Pacific Islands V. Pacific Science 18: 399-410. (Araneae, Thomisidae). Bulletin of the American Museum of Natural History 129: 1-180. MELLO-LEITAO, C.F. DE 1929. Aphantochilidas e SIMON, E. 1 895. Histoire Naturelle des Araignees, vol. Thomisidas do Brasil. Archivos do Museu 1(4): 761-1084. (Librairie Encyc1op6dique de Nacional, Rio de Janeiro 31: 9-358. Roret: Paris). NELSON, G. & PLATNICK, N.I. 1981. Systematics 1900. Arachnida. Pp. 443-519. In, Sharp, D. (ed.), and biogeography: cladistics and vicariance. Fauna Hawaiiensis, or the Zoology of the (Columbia University Press: New York). Sandwich Isles, vol. 2. (Cambridge University

NUR f A. & BEN-AVRAHAM. Z. 1977. Lost Pacifica Press: Cambridge). Continent. Nature 270: 41-43. STODDART, D.R. 1968. Isolated island communities. ONO, H. 1988. revisional study of the spider family A Science Journal 4: 32-38. Thomisidae (Arachnida, Araneac) of Japan. (Na- STRAND, E. 1913. Neue indoaustralische und tional Science Museum: Tokvo). polynesische Spinnen des Senckenbergischen *1988. PLATNICK. N.I. & NELSON, G. Spanning-trce Museums. Archiv fur Nalurgeschichte 1913 A 6: biogeography: shortcut, detour, or dead-end? Sys- 113-123. tematic Zoology 37: 410-419. SUMAN.T.W. 1970. Spiders of the family Thomisidae RAINBOW, W.J. 1897. The Arachnidan fauna of in Hawaii. Pacific Insects 12: 773- 864. Funafuti. Memoirs of the Australian Museum 3: TIKADER, B.K. 1980. Thomisidae (crab-spiders). The

105-125. Fauna of India, Araneae, vol. I, part 1: 1-247. 1902. Arachnida from the South Seas. Proceedings TOLEDO PIZA, S. 1937. NovosThomisidos do Brazil. of the Linnean Societv of New South Wales 26: Folia de Clinic. Biologia, Sao Paulo 9: 179-183. 521-532. WILSON, J. TUZO, 1963. Continental drift. Scientific R1NALDI, I.M.P. 1983. Conlribuicao ao estudo das American 208: 86-100.

SCORPION DISTRIBUTION IN A DUNE AND SWALE MALLEE ENVIRONMENT

N.A. LOCKET

Locket, N.A. 1993 II II: Scorpion distribution in a dune and swale malice environment. Memoirs of the Queensland Museum 33(2): 593-598, Brisbane. ISSN 0079-8835.

A system of stable and vegetated dunes, separated by occasionally flooded swales, contains populations of six scorpion taxa. Urodacus yaschenkoi (h\ru\a) digs deep burrows in the soft soil of the dunes, but the shallower burrows of U. armatus Pocock are concentrated at their base, extending onto the swale for ca. 50m. Lychas jonesae Glauert, L. variatus (Thorell) and Isometro'tdes angusticaudis Keyserling occur on the swale but not the dunes cophonius kersluovi Glauert has occasionally been found in Htier beneath malice trees on the dunes. Soil hardness may account for U. yaschenkoi occurring only on the dunes. Isometroides, a spider predator^ occurs where spider burrows are found Predalion by the larger Vwdacus may aecounJ for the buthids not extending onto the dunes,Q Stor^otts, Australia, ecology, habitat.

N.A. Locket, Department ofAnatom;, and Histology, University ofAdelaide, Box 498, GPO,

Adelaide, South Australia 5001 1 Australia; 3 November. 1992.

An area of scrub near Bern in the South specimens referred to /. atigusticaudh fa Australian Riverland consists of u scries of stable ticipation of that work. dunes bearing malice trees and shrubs inter- Many factors influence the distribution of scor- spersed with flat grassy swales, which show signs pions, including latitude and climate, type or of occasional flooding. This locality contains a terrain and soil hardness. Biotic factors such as population of Urodacus yaschenkoi. Blaeklight- intra-guild competition and predation are also mg nn and between the dunes showed that five important (Polis, 1990). The latitude of the study other scorpion taxa are present in the area but not site, ca 74°, is within the range of maximum uniformly distributed. Observations arc now scorpion diversity, given by Polls (1990) as presented on these populations and their distribu- 38°, Within (his belt up tn ca lOspeciesmay ocnn tion across the dune- swale syslejn sympatrically in the northern hemisphere, though Two of the taxa found do not accord with the most locales have only 3-7. Diversity is gre descriptions given by Koch (1977) One ,i in dcverl regions, with 24 of the 28 communities Lychas, corresponds to Gljuer.'s (1925) unil- uf six or more being in subtropical deserts, where iustrated description of L jomtiae, apparently terrain may range from loose sand to hard packed frc>ma single specimen collected near KalgoorHe. stony ground. Soil hardness was closely related Western Australia. Koch (1977), who examined to species distribution in Opisthophthalmus bv the holotype, included L. jonesae in /.. tttar- Larnorai (1978). Bradley (1986) showed ibu: moreus but did not give reasons fordoing so populations of ihc burrowing Paruroctonus present specimens resemble L. alexandrinus, utahensis were denser on soft soils than hard, and widely distributed in arid Australia (Koch. 1977, that its burrowing Was impaired by the harden! i \y map 4), but differ significantly from it and from of the soil due cc rain. Vegetation patterns corre- L. marmoreus. They agree closely, except in lated both with soil type and scorpion popuhn colouration, with the (markedly faded) type of L but areas with different scorpion densities con- jonesae- An illustrated redescription ofL.jonesae tained similar biomass of potential prey. Bradley is in preparation, and the present specimens also demonstrated the destruction caused by referred to that species meanwhile. flooding to a population of P. utahen The other belongs to the genus Isometroides, Polis (1990) show* that biotic factors may well which Main (1956) suggested was monospecific, affect the composition and balance of a popula- all specimens being referable to/. vesc\is % a view tion. He gives strong evidence on the important supported by Koch (1977). The Bern specimens influence of scorpions on the community in agree better with the descriptions of Keyserling which they live, by competition through preda- (1885), Kraepelin (1916) and Glauert* (1925. tion on other invertebrates or cannibalism of

1963) of I. aru'itsticauiiis, and are certainly not smaller scorpion species, and of smaller con- typical of / vcsri/s. An account of the genus specific instars. Individual scorpions of different Isometroides is in preparation, and the present species but similar size may well compete direct- 594 MEMOIRS OF THE QUEENSLAND MUSEUM

Dale V.vascit U annat. t, august- I variatus LJonesae Berri. rainfall 1"! 1") 1 fi Sep 89 4 n

24 Mar 90 4 7. 4 u 1

"""

1 1 l^.V-i" 1, I : 3

20 0cl90 10 7 10 Nov 90 6 27

15 Dec 90 11 2 21 D- mm rain, mean 15 Jan 91 11 16 -^ mm rein, median

Days of rain !6Mar9l !! :;•;

1-1 w>: -m 14 2 2 I

2 Nov 91 5 l

7-8 Dec 91 9 6 5 1 71)

1 1 4\ I » I * I I ' I 2 Feb 92 1

JFMAMJJASOND ! 1 29 Feb 92 b 11 1 59

2 May 92 j 6 1 TOTAL 34 52 78 G 265 Berri, max. and min. temperatures

TABLE 1. Catches of scorpions, blacklighting only-

parts of the transect containing the characteristic burrows of Urodacus yaschenkoi were noted. On one occasion in summer all the burrows within Max.. "C the transect were marked with numbered tags. Min..*C Burrow identification was confirmed and specimens obtained by trapping burrows away from the transect. The traps were plastic vending machine cups, ca 200cc capacity, dug into place at the burrow mouth, the burrow opening directly fMAMJJASOND at the lip of the cup. The traps were visited at least Month In after sunset, when scorpions had emerged from the burrows and fallen into the cup.

FIG. 1. Temperature and rainfall, Berri (nine year Smaller burrows, without the curved entrance average). of U. yaschenkoi, were identified as those of U. armatus by digging out the scorpion. On three ly for food, since most are generalist predators, occasions U. armatus was identified by black- perhaps leading to one species becoming light at a burrow mouth which was then marked dominant. Different sized scorpions may take and examined next day. different groups of prey, based on size, though the Blacklighting was carried out, away from the young of a large species may compete directly transect but within the dune-swale system, on with adults of a smaller when the individuals are each visit by two persons, working parallel to of comparable size. each other and approximately 100m apart. Black- METHODS lighting, commencing at various times after sunset, on some occasions as soon as it became dark enough and on others up to 1 .5h after sunset, was The site is about 8 km NNE. of Berri, 140°38'E, continued for ca 2h. , 34°I3 S. in the South Australian Riveriand. The Urodacus yaschenkoi and U. armatus could be climate at Berri is warm and dry (Fig. 1). A distinguished in the field by blacklight. transect 10m wide was laid out from dune to dune Isometroides* except very small individuals, was across a swale, some 300m. The nature of the soil distinguishable from the Lychas species, but the was noted, and trees and shrubs more than a few two Lychas could not be told apart until the catch cm high plotted to the nearest lm. was examined indoors. This took place on return Single-, or on one occasion two-day, visits from the field, when notes on weather, capture were made to the study site on fourteen occasions, sites and behaviour were written. Detailed plots on moonless nights when possible, between of specimen location were not made, but it was March 1989 and May 1992. On each visit the possible to assign the taxa to dune, dune base or M ALLEE SCORPION DISTRIBUTION

FIG. 2. View of transect across swale, looking north from southern dune J. sandy soil al base of dune. 2, low bank beside old track. 3, track. 4, swale. 5, far dune. See- also Fig<; 3 and 4. swale. Identifications were confirmed on return to trace the fate of individual burrows. Individual to the laboratory. U yaschenkoi usually adult males, were found by blacklighting in December-February, mainly RESULTS on the dunes but including one ca. 50m onto the swale U. yaschenkoi were also observed at their burrow mouths by blacklight in summer. The transect, running N-S, extends 300m from the top of one dune, ca 6m high, across the flat Urodacus armuius burrows were occasionally the dunes, but were at the swale, to the top of the next dune (Fig. 2). The found on most dune dunes bear scattered rnallee trees, Eucalyptus base and for up to 50m onto the swale. U. ar- mostly immature, found blacklight oleosa and E. bruchycalyx, numerous bushes of matus, by had native hop, Dodonaeu sp., and clumps of a similar distribution , with occasional examples Spinifex. The swales are lightly covered with up to ca 100m onto the swale. Immature in- grass and other low plants, with sparse bushes of dividuals have sometimes been found clinging to native hop and of Cassia nemophila var. low vegetation within a few cm of the ground U platypodia (Fig. 3l armatus, recognisable by blacklight from U. yaschenkoi by their squat pedipalps, have been seen Catch data are summarised in Table 1 and at the burrow mouth between October and distributions in Fig. 4. Urodacus yaschenkoi bur- February rows occur on the dunes, mainly in the open but a few under light cover. In January 1991, 80 Three species of bulhid. Lychas jonesae, L burrows were located and marked within the tran- variatus and Isometroides angusticaudis have sect, but by October, 56 tags either were not been found, only on the swale, by blacklighting related to a visible burrow or had been disturbed L jonesae, commonest overall, though on oc- by animals. Of the three V. armatus burrows casions outnumbered by /. angusticaudis, occurs marked while blacklighting, two had been closed all across the swale, but often near low vegetation by next day, and would have been missed apart near the dune base. Some have been found cling- from the marker. No further attempts were made ing to stems within a few cm of the ground. 596 MEMOIRS OF THE QUEENSLAND MUSEUM

200 300m U. yaschenkoi _ ,

o^tt Dune DUNE 270 l«U

armatus 270 X u. < Z Track % 60 o O^D 210

15H

Isometroides 180

SWALE I . ychas ionesae

T . ychas variatus 30 -6 O o ° 8 burrows Spider 150 o '-. o o i . 120

120 210 Dune

• : t. " "'. • ' . .-.Vpp g> 100 b Track 90 Spinifex S V Eucalyptus brachycalyx 1 Eucalyptus oleosa 2 Old track Dodonaea sp. • U. armatus Cassia nemophila var platypoda o X A\ 60

FIG. 3. Plot, to scale, of distribution of major plants across transect. Distances in metres.

have 30 mostly head down, but none been found DUNE high in bushes. Lychas variatus has been taken occasionally, U. yaschenkoi never more than two and frequently none, in an 0m evening. Insufficient have been found to comment on their distribution on the swale.

Isometroides angusticaudis has mostly been FIG. 4. Distribution of scorpions across dune and swale found on the surface by blacklight, among low system. Urodacus yaschenkoi occurs on the dunes, U. grass rather than taller vegetation and extending armatus at their bases and the three buthids on the swale. up to but not above the dune base. A few have been dug by day from the burrows of lycosid DISCUSSION spiders, abundant on the swale but not the dunes. One Cercophonius kershawi (identified from Many labels in collections state a locality, Acosta (1990)) was found in May 1992 by kick- without details of microhabitat. The population ing over leaf litter beneath mallee trees while disparity now described from localities a few blacklighting. (Two juvenile Lychas variatus meters apart suggests that it may be helpful to were caught in the same way on dunes within 1 define localities more precisely. km of the transect in June 1987). Lamoral (1978), studying two burrowing MALLEE SCORPION DISTRIBUTION 597

species of Opisthophthalmus, found clear separa- The habit of clinging to vegetation close to the tion of two otherwise sympatric species, strongly ground, also observed in immature U. armatusby correlated with soil hardness, a factor probably G.T. Smith (personal communication), may also important in Urodacus. Koch (1978) found enable scorpions to avoid wandering predators. the depth and tortuosity of Urodacus burrows Some scorpions, e.g. Centruroides exilicauda in within a species greater with aridity. There is also America, are frequently found in bushes well off a species difference: U. yaschenkoi and U. the ground, but such climbing has not been seen hoplurus dig deeper and more spiral burrows than in the present case. U, armatus. Koch (1981) noted that Urodacus The total catches of buthids suggest that Lychas scorpions show little correlation with soil type, jonesae is dominant on the swale, though on but more with softness and the chance of reaching occasions more Isometroides have been caught. water by burrowing. Shorthouse (1971), Shor- L. variatus is much less common than either. thouse and Marples (1980) and Koch (1978) have Isometroides is known to be a specialist burrow-

described the spiral burrows up to 1 m deep of U. ing spider predator. Probably the Lychas species yaschenkoi in loose sandy soil. The dune soil at are less specialised, though little is known of their Berri is of this type. Most burrows of U. armatus, diet; one instance of L. jonesae eating an imma- seldom more than 30cm deep, are in the firmer ture U. armatus was the only act of predation soil at the dune base and swale, though some observed. Insufficient is known of the habits of occur up the dunes, in what appears otherwise to the two Lychas to indicate why the smaller should be U. yaschenkoi territory. be commoner. Koch (1981) considered three Australian zones, a moist temperate southern, semi-arid to ACKNOWLEDGEMENTS desert central and humid tropical northern, associating various scorpions with these zones: I thank Deirdre Locket and G.R. Johnston for assistance in the field with plant identifica- Urodacus yaschenkoi and U. armatus he and regarded as mainly central forms. He found scor- tion and Rae Sexton for assistance with plant pion distribution not correlated with vegetation identification; Dr D.C. Lee (South Australian Museum), Dr A. Yen (Museum of Victoria) and type, single species occurring in a wide range of M.S. Harvey (Western Australian habitats. He suggested that range-determining Dr Museum) factors include temperature, precipitation and for access to specimens in their care; and the Bureau of Meteorology for climatic data. biotic factors, e.g. competitive exclusion. He also examined morphological characters, suggesting that large size, longer metasomal segments and LITERATURE CITED spines, more granulation, and higher pectine ACOSTA, L.E. 1990. El genero Cercophonius Peters, tooth counts are aridity-linked in Urodacus, 1861 (Scorpiones, Bothriuridae). Boletin de la while large size, light colouration, more granula- Sociedad de Biologia de Concepcion (Chile) 61: tion, higher pectine tooth a less counts and 7-27. prominent subaculear tooth are aridity- linked BRADLEY, R.A. 1986. The relationship between traits associated in Lychas. He regarded population density of Paruroctotuis utahensis Isometroides •&$ showing the culmination of these (Scorpionida: Vaejovidae) and characteristics of

buthid traits. Of the sympatric species at Berri, its habitat. Journal of Arid Environments 1 1: 165- both Urodacus are pale and smooth, their pectine 172 1925. Australian Scorpionidea. Part 1. counts largely overlap, but U. yaschenkoi is large GLAUERT, L. Proceedings of the Royal Society of Western and U. armatus small. L. variatus and /. angus- Australia 11: 89-1 18. ticaudis are both pale but mottled though L. 1963. Check list of Western Australian scorpions. jonesae is small and dark, with subaculear tooth - Western Australian Naturalist 8: 1 8 1 1 85. intermediate between the other two buthids. KEYSERLING, E. 1885. Part 32. Pp. 1-48. In Koch, L. The swale, where Isometroides has been col- and Keyserling, E. 'Die Arachniden Australiens

nach der Natur bescrieben and abgebildet* . (Bauer lected by blacklight, contains numerous lycosid & Raspe: Niirnberg). spider burrows, up to 30cm deep in hard soil. KOCH, L.E. 1977. The taxonomy, geographical dis- Isometroides was recognised by Main (1956) as tribution and evolutionary radiation of Australo- a spider predator collected her their and by from Papuan scorpions. Records of the Western burrows. Four have been so collected in the Australian Museum 5: 81-367. present study, but no concerted digging has been 1978. A comparative study of the structure, function done. and adaptation to different habitats of burrows in 598 MEMOIRS OF THE QUEENSLAND MUSEUM

the scorpion genus Vrodacus (Scorpionida, Scor- MAIN, B.Y. 1956. Taxonomy and biology of the genus pionidae). Records of the Western Australian Isometroides Keyserling (Scorpionida). Museum 6: 119-146. Australian Journal of Zoology 4: 158-164. 1981. The scorpions of Australia: aspects of their POLIS, G.A. 1 990. Ecology. Pp. 247-293. In Polis, G.A. ecology and zoogeography. Pp. 875-884. In (Stanford (ed.) ? 'The biology of scorpions.' 'Ecological Keast, A. (ed.), biogeography of University Press: Stanford). Australia'. (Junk: The Hague). SHORTHOUSE,D.J. 1971. 'Studies on the biology and KRAEPELIN, K. 1916. Results of Dr E. Mjoberg's energetics of the scorpion Urodacus yaschenkoi.' Swedish scientific expeditions to Australia 1910- (Unpublished Ph.D. Thesis, Australian National 1913.4. Scolopendridenund Skorpione. Arkiv fur University: Canberra). Zoologi 10: 1-43. LAMORAL, B.H. 1978. Soil hardness, an important SHORTHOUSE, D.J. & MARPLES, T.G. 1980. Obser- and limiting factor in burrowing scorpions of the vations on the burrow and associated behaviour of genus Opisthophthalmus C.L. Koch 1837 (Scor- the arid zone scorpion Vrodacus yaschenkoi pionidae, Scorpionida). Symposia of the Zoologi- Birula. Australian Journal of Zoology 28: 581- cal Society of London 42: 171-181. 590. FROM FLOOD AVOIDANCE TO FORAGING: ADAPTIVE SHIFTS IN TRAPDOOR SPIDER BEHAVIOUR BARBARA YORK MAIN

Main»B Y 1993 11 1 ): From flood avoidance to foraging: adaptive shifts in trapdoor spider behaviour. Memoirs ofthe Queensland Museum 33(2): 599-606. Brisbane. ISSN O079-K81S

F< -ssonal habits protect many mygalomorph spiders from boih biotic and environmental factors. However burrows in some habitats are vulnerable to sheet flooding. Spidcrg in diverse habitats counter the hazard of flooding in various ways. A comparative account of adaptive specialisations, particularly of burrows, in rlood avoidance is presented. Such primary modifications sometimes lead to new foraging opportunities. Examples are given of modified foraging as a consequence of burrow adaptations. The origins ol such constructions are hypothesised in relation to changing climatic regimes and modified habitats. The idiopid genera Homagotu7 Rainbow and Nenhomogona Main are restored from synonymy. \Z\Mygaiomorphae, Trapdoor spiders, burrows, flood avoidance, foraging, Australia.

Barbara York Minn, Zoology Department, University of Western Auiuaho. Nedlonds, Western Australia 6009, Australia; 5 November, 1992,

Most mygalomorph spiders arc terrestrial and of flooding and shows that these primary adapta- cither make burrows, silk tubes or webs or com- tions have sometimes secondarily created new- bine a web with a burrow. A few are arboreal and foraging opp' h:i spiders In pann cither make tubes in the bark of trees or have webs the previously unreported nest of an undescribed in crevices or under bark; their nests are generally species of Aname which exhibits both 'primary' not associated with foliage. Of the 15 currently and 'secondary' responses in its flood-avoidunce recognised families of Mygalomorphae (Raven, tactics, is described also.

J 985), 10 occur in Australia and of these, seven have at least some representatives with trapdoors. FLOOD AVOIDANCE BEHAVIOUR A burrow provides protection from the physical environment, from weather conditions and from In certain terrestrial habitats, burrows and silk predators and parasites; it provides a brood cham- lubes and webs face the physical hazard of flood- ber for eggs and spiderlings A burrow is also a ing. Generally the complexity of habitats in rain-

lair from which a spider perceives and ambushes i and mcsophyl] forest provides some sort of or intercepts prey: thus it provides a foraging buffer against flooding, there is greater capacity hase. All these functions have been commented for absorption of rainfall in the vegetation and on many times in the literature over the last titter of a forest floor than there is for example in century, the earliest comprehensive study pc desert or semi-arid country where sheet floo ly being that of Moggridge (1873, 1874). on bare ground is a common phenomenon. liven In its simplest form a burrow has an 'open' so some rainforest situations are subject to water entrance from which a spider makes short sorties logging and flooding associated with torrential in pursuit of prey. In extreme habitats, the en- downpours. Likewise where rain is markedly sea- trance may be sealed with silk or soil for added sonal as in monsoon rainforests the alternation of protection during certain times of the year, for wet and dry conditions means that some habitat- example during summer drought or winter snow experience sudden saturation or inundation. Ma in falls. Folding collars and finally hinged doors (1976, 1982b) and Cloudsley-Thompson (1 give maximum security and protection. The pro- 1983) discussed some of the burrow modi. tective advantage of a door has been shown to be lions of mygalomorphs which prevent flooding. offset by the hinge-line inhibiting the foraging Avoidance of flooding is different to behaviours area of a spider but nevertheless many door build- whereby spiders withstand Immersion by cm-Ins er- have overcome this disadvantage by various ing the body in a bubble of air modifications to the burrow (Coyle, 1981; Main, In summary, in avoiding flooding of nest s 1986). mygalomorph spiders have adapted behavioural- This paper discusses the adaptive behavioural ly in several ways:

responses associated with structure and site, ( ) By the nest sire-behaviour prior to, nest 1 moving of Australian mygalomorph spiders, to the hazard during or after the event. 600 MEMOIRS OF THE QUEENSLAND MUSEUM

SUSFACE raACING (5) By adopting an arboreal instead of terrestrial habitat. Examples of these adaptations will now be given.

(1) Mygalomorphs elsewhere have been recorded as moving nest sites prior to inundation. The Amazonian web-weaving diplurid Ischno- thele guianensis makes webs on tree trunks in the rainforest which is seasonally inundated. Spiders move their web sites higher up the trunks as the water level rises (Hofer, 1991). The Australian diplurid genus Cethegus (curtain web spiders), which makes copious webs sometimes associated with a burrow retreat, amongst stones, or against logs or shrub stems, is widespread across tropical Australia, the interior and southwestern Western Australia (Main, 1960, 1991a and unpublished;

Raven, 1 984). A study on a species at Durokoppin in the Western Australian wheatbelt showed that the spiders move their web sites after rain damage (Main, unpublished). Another species which occurs amongst the rocks of water courses in the Kimberley rainforest patches (Main, 1991a) presumably moves web sites during the 'wet' when the habitat is inundated. Although dependent on a web for prey capture these spiders are remarkably agile on the ground, and thus could readily decamp if flooded out of their nest sites, follow- FIG. 1. Flood avoidance specialisations of mygalo- ing which a could readily reconstructed morph burrows: (a) soil and pebble levee ofKwonkan web be wonganensis; (b) bath plug door (Idiopidae); (c) (unlike some architecturally complex burrows of dome or 'cap' of an Aganippe species (Idiopidae); (d) other mygalomorph families). 'sock' of Anidiops (Idiopidae); (e) 'pebble' and collar (2) Even 'open' burrows, if lined with silk closure of burrow of Stanwellia nebulosa (Nemes- frequently have collapsible or folding collars iidae); (f) side shaft and escape 'hatch' of burrow of which can prevent flooding, for example as in Aname diversicolor (Nemesiidae); (g) side shaft with many nemesiids and some Misgolas species, e.g. door in nest of Idiommata sp. (Barychelidae); (h) profile of nest of a Teyl sp. (Nemesiidae) with en- M. pulchellus (Rainbow & Pulleine) (pers. obs.)- trance atrium and subterranean doors (and sideshaft); Collars are sometimes strengthened with litter or (i) aerial tube of Misgolas robertsi (Idiopidae) sup- by a surrounding pile of soil or silk-bound peb-

ported on tree trunk; (j) extended tube with trapdoor bles (Fig. la) e.g. the levees of some nemesiids of Aganippe castellum (Idiopidae) supported against including Kwonkan wonganensis (Main) (orig- stem of shrub; silk and soil (k) tube of 'turret-build- inally described as Dekana (Main, 1977, plate ing' Aname sp. (Nemesiidae) in foliage of shrub. 15)) (see also Main (1981, plate 1, fig. d) with reference to the levee of a 4 diplurine' burrow). (2) By permanently reinforcing the nest e.g. The ctenizid Conothele, the actinopodid burrow (and/or opening) against inundation-pre- Mis- vention. sulena and some barychelids have secure doors and parchment or canvas-like, silk linings which By modifying the nest structure that (3) so part are more-or-less impermeable. Most idiopid of it is safeguarded against inundation-preven- species construct well defined burrows and all tion. except some species of Homogona\ Neo- (4) By modifying the nest so that it is (in part) homogona and Misgolas have a trapdoor. Many sited beyond the reach of intermittent flooding- idiopid species which have adapted to flood- prevention. prone habitats, ranging from rainforest to desert

'The idiopid genera Homogona Rainbow and Neohomogona Main were synonymised with Cataxia Rainbow by Raven (1985). They are here formally restored from synonymy and diagnosed in Main (1985). v

FLOOD AVOIDANCE IN TRAPDOOR SPIDERS m

and particularly in bare ground or sloping creek function as brood chambers and also sometimes bftAKft, frequently have permanently reinforced us protection from predators it is probable that Hie hurrows including thick plaster walk coated with original fuiKtion was that they offset flooding of dense silk, and thick, close fitting bath-plug like the main shaft of the burrow (see Fig. 3g). Con- doors or lightly fitting caps (Figs lb c). Such sidering the habitat of Teyi species this is une- r secure nests also provide protection agains; quivocal. parasites adverse physical condi- predators, and (4) Burrows of some species are continued as tions, including dedication tubes above the ground or litter surface and may O) Some species of Idsopidae, Nemesiidae, be strengthened into free standing paiisatk. Barychelidae and Hexathdidae safeguard par- the attachment of leaves, twigs and debris) which ticular areas of their burrows against inundation, deflect sheet flooding in bare ground or prevent Such safeguards include (a) special blocks in the immersion in water-soaked litter. L'xamples in- 4 lumen or (b) sideshafts with 'escape batches Off clude the tubes of Hofnogona cunkularius Main iDe or more sidechambers with 'internal* trap- (Main, 1983, fig. 15). Misgolas fursutus (Rain- doors. Examples of the above are; bow & Pulleine) (perx. obs.i and Neohomogo-na [a] Some species which have cither an open wirhnyj ami jV hoIgaJiupensisMain (Main, 1985, entrance (e.g., Misgolas spp.) or a flimsy door fig. 219) Tubes that extend a considerable Jis tancc above the substrate are attached to rocks, Anidiops) have a collapsible poliaJ | 'sock', tree buttresses, exposed roots e.g. Misgolas sec Main, 1957, 1976. plate 9 and 1983. Eigs209 ( 210)consisting of a detachable section of the silk robe rt si (Main and Mascord) and related spe lining which can be pulled downwards by the (Main and Mascord, 1974, pi. la, b: Mascord. spider thus blocking the burrow (Fig. Id). Water 1970 pi. 2 fig. 4)i logs e.g. Cataxia macuiai,* seeping down the walls or flowing into the bur- Rainbow (Main, 1969, figs 30, 31} or stems of

is deflected by the infolded 'neck* of the sock shrubs e.g Aganippe casielhtm Main (Main and soaks into the surrounding soil leaving the 1986. figs 2, 4i; 1987. fig. 6). The first examples lower section of the silk lined burrow unflooded. have open tubes, sometimes with flanged, colla]

A similar structure in Sianwellia nebtdosa (Rain- sible collars but,A. costellum ha* a trapdoor i bow & Pulleine) (Fig, Lei is further reinforced by li, j). All are effective in flood avoidance by an attached, artificial pebble which seals the having the entrance above the 'flood level' ful lower silk lined part of the burrow (Rainbow & lowing a delude. Pulleine, 1918, pi. 20; Main, 1964, 1972 figs 22a. The previously undescribed but remarkable b, 1976). Although these devices are generally burrow/tube of an undescribed species otAmant' idered a projection against predators (see extends as a turret-like tube amongst supporting Main, 1956a regarding Anidiops (=Gait4s) their foliage of shrubs. The genus Aname which is original function was probably prevention widespread but endemic in Australia and Tas- against flooding (see Main, 1976 regarding Stan- mania, occurs in varied habitats, is taxonomically wellia). diverse and includes many undescribed species. lb) NestS of some species have silk sideshafts The turret-building species occurs in semi-arid with 'escape' hatches at the surface (see Fig. 10 country in southeastern Western Australia which probably enable the spiders to escape- Eyre Peninsula in South Australia. Nests have flooding as well as intrusive predators. Examples been observed in nuilleeApinifex associa*

' are the burrows of "wishbone' * gpidefS t dypJit.\/Trt(hliu'\ (Figs 2a* b, c) but also in yenus Aname (Main, 19S2a. fig. 1; 1976, pi. 10) mulga and amongst chenopod shrubs in seasonal- and Misgolas ornata (Rainbow j (Main, 1985: ly swampy habitats. The spiders make shallow 33). burrows lined with silk wbidi extend ;•* tuhes Into (c) Sideshafts with 'internal* trapdoors aic con- the foliage of supporting tussocks or shrubs and structed by several undescribed species i»f . open either w ithin the foliage or above the canopy

nesiidae) (earlier attributed to Ixtvmuus, vec I Fig Ik)- The outside of the tube is heavily bur Main (1976, pp. 86-88, fig. 21)), by Hadrm irregularly coated with soil. During and alter rain, lllcxathelidac) (see Main, 1964:40. fff. j; 1976, sheet flooding occurs in such sites, and particular- fie. 18c; Gray, 1984, fig. 3 1 (described 9&Atrox)), ly in the sandy loam of mallee/spinifex associa Idiommata (Barychclidac) (Main, 1976, fig. J9d; lions, water flows in a slurry around the spinifcx Raven & Churchill, 1991: 35) and Missulena hummocks. Burrows appear to be deep

i . jtlnopodidael (Main, 1956b). While the land/of remade) after rain and the sodden spoil sideshafts of all species studied are known to dumped on the lip of the tube and outside the 602 MEMOIRS OF THE QUEENSLAND MUSEUM

PRIMARY ADAPTATION SECONDARE' ADAPTATION

opening Flood avoidance through modified burrow Specialised foraging in response to Unmodified foraging structures 'new' opportunities

—open* burrows: collars^ folding collars & levecs surface Ground —cloved* burrows; reinforced walls & doors surface — snb-surcae closures oftunvn (open or closest surface s Subterranean burrow 1 o SO —bupcrmimcrary chambers with dour* subsurface (pitfall captures) oto u. extension of open rubes agamtf vertical support arboreal and semi-aerial

extension of closed burrow (wifh door) against Elevated — arboreal plant stem

extension of open tube into foliage foliage canopy

CORTICOLOUS — open mp i losed tubes ui biiV a/boreal

TABLE 1 . SiLing ofmygalomorph burrows primarily modified in flood avoidance and those secondarily resulting in specialised foraging methods. * open = without door; closed = with trapdoor. entrance. Thus the turret is heightened progres- NEW FORAGING OPPORTUNITIES sively while much of the pasie-h'ke spoil pours down the outside wall, solidifying into a kind of Foraging 'techniques' of mygalomorphs are stucco as it does so. Tubes up to 25cm have been closely associated with burrow structure. Bur- observed within the supporting scaffolding of rows of some species are constructed with special spinifex hummocks. The sometimes bulky struc- adjuncts that enhance foraging, such as silk trip tures belie the relatively small size of the ar- threads or attached twiglines. Main (1957, 1982b, tificers (up to 4.5mm carapace length). 1987) described some combinations of burrow structure, morphology and foraging behaviour in Occasionally free-standing tubes (but small and relation to the habitats occupied and noted the only up to litem high), away from shrubs, are relative efficiency of the various foraging (see found Main, 1982b, Eg. 2). Most unsup- strategies. ported tubes may collapse (at ground level) and In modifying behaviour and burrow structure in destroyed during flooding while at least the be flood avoidance, some species have as a conse- aerial section of those amongst the tussocks quence been exposed to 'new' foraging oppor- remain relatively intact. tunities. Table I summarises the categories of 1 (5) Mygalomorphs which have their nests sited 'primary burrow modifications and sitings as- completely on the trunks of trees are found only sociated with flood avoidance and any in very wet forests. Their nests in such situations 'secondary' changes of foraging behaviour appear well protected against inundation. Reinforcement to burrows and doors which do not involve alteration of siting of the entrance, do Australian examples include theraphosids which not generally result in modification of foraging make silk retreats under bark in north Queensland method. However, any modifications which alter (pers. observation), the nemesrid, Chenisionia the siting of burrow entrances away from the i-i!losa Rainbow and Pulleine which makes surface of the ground do result in specialised defined silk tubes in crevices of Casuarina bark (predatory) foraging behaviour. in the karri forest of southwest Western Australia A striking example of a subterranean entrance (pers. observation), species of the ctenizid. Con- is that constructed by a group of Teyl species. othele* in tropical Queensland (Main 1976) il- Although most species of Teyl have open bur- lustrated in McKeown (1963), the barychelid rows, a few build doors in flood avoidance Sason in north Queensland (Raven, 1986) and Several species have surface doors in bare open ground but the foraging method is similar to that rnigids e.g. Moggridgea tingle Mam (Main. of species with open holes (Main. 1986). Several 199 lb) C. viUosa and Af. tingle both make nests species which occur in deep litter in seasonally inthe ground also, in sites not vulnerable to water flood-prone, braided creeks and the flats around logging. This is probablv secondary habitat a salt lakes in southern Western Australia have adopted in response to gradual drying of the evolved a further specialisation. An open 'false climate and environment where the bark of trees entrance* leads into a cup-like pouch or atrium would be subject to dessication. which has on us inner wall a trapdoor (entrance FLOOD AVOIDANCE IN TRAPDOOR SP1DKKS 603

FIG. 2. Aname sp. Habitat and turrets (a) mailee/spinifcx shrubland cm sandv loam soil west of Balladonia. Western Australia; (b) spinifex shrubs with 'turrets'; (el turret in Spioiiex, note open entrance. proper) which leads into the main shaft of the provide a double safety valve m the eve burrow; a second trapdoor leads from the main flooding (Main, 1 976). The false entrance/atnum fthftfl into a blind side&haft (Fig. Ih) l These doors functions like a pit-fall trap which caceln tf 604 MEMOIRS OF THE QUEENSLAND MUSEUM

crawling insects that arc then seized by the spider aridification. Paradoxically while probably as it straddles the open door inside (Fig. lh) making simple tubes and burrows in a "primi- (Main, 1982b fig. 7d>. The pit-trap also provides tive", uniformly wet habitat the spiders have been a clear foraging area for the subterranean exposed to the the hazard of sheet flooding as a entrance-proper, unencumbered by the litter result of periodic deluges associated with a dryer, which would otherwise impede the plug-like but seasonal climate. Nevertheless, clearly a few door. Notable examples of elevated nests (sup- such species do belong to taxonomie groups ported against tree trunks), the primary adapta- which already had burrows and doors well tion of which has been flood avoidance but which adapted to sudden deluges in rainforest and nrv has secondarily resulted in arboreal foraging, mesophylic habitats, e.g., species of Arlmnitis. those of Misgolas robertsi (see Mascord, 1970 Much of the semi-arid woodland, shrubland p. ! 2; Main 1976. plate B5) and Aganippe caste!- and sakbiish steppe of southern Australia enjoyed I 34 and fig. 6). urn (Main. 1986, 1987 p. a consistently wet climate and supported Spiders, the nests of which are sited on the mesophytic forest and rainforest interspersed trunks of trees but which have no contact with the with sw ampland duri ng the early Tertiarv ( Cook- and are thus cortienhiu*., similarly have ground son, 1973; Cookson & Pike, (953; 1954; Trus- an arboreal foraging behaviour. Con- adopted weh\ 1990). With progressive drying of the othelc (possibly Cvtwiheie arboricota Pocock), climate and development of a markedly seasonal Sason and Chenisionia viliosa antl ttf&ggridgea weather pattern the vegetation was altered, leav- (irah? are thus all arboreal foragers although the ing some relictual, moist microhabitats in jux- latter two axe recognised as being facultative in taposition with other seasonally alternating their behaviour. Even Idiociis species which ex- wet/dry microhabitats some of which now extend their tubes above the tide level (or are scaled perience periodic, sudden inundation against tidal ri rnctimes against mangrove Certain species of trapdoor spiders whose stems (Abraham, 1924,), are probably in effect taxonomie relatives in rainforest and mesophytic arboreal foragers habitats lack specialised burrow structures have However, until recently there was no known responded to seasonal sheet flooding in arid example of an Australian mygalomorph which habitats by modifying burrow structures in flood fed in the foliage or canopy of vegetation. There- avoidance. Notable types ofburrow modification fore, the turret-building species of Aname is of in the western part of the continent, which are interest. The lubes open either in gaps amongst evolved in direct the foliage or above (he canopy of tussocks and hypothesised as having shrubs. The spiders presumably do not venture response to (geologically) historical changes in the burrows Aganippe far, if at all. away from the entrance (and are the environment are of unlikely to crawl down the outside of the tube to castelluvt, Teyl species with both surface and the ground) and thus are dependent cm insects subterranean doors and the aerial tubes of the creeping amongst and on the foliage and flying Aname species Main 1 1986) discussed the prob- insects as they alight- This would mean that they able origin of the semi-arboreal ijestofA caSJel- are a truly foliage or canopy forager as distinct tmn. iis anti-flooding function and adoption of a from arboreal spiders which forage on the trunks new' foraging behaviour This dual adaptation of trees and shrubs. probubly evolved during the late Miocene/Pliocene in response to fluctuating BURROW MODIFICATIONS as A wer/dry season* and fragmentation of the RESPONSE TO HISTORICAL landscape into a mosaic of wet and dry ENVIRONMENTAL CHANW-s uhahitats. It also has been postulated earlier

I Main, 1 976) that the single surface door of some The arboreal tubes and most corticolous nests Teyl species evolved in response to in mesophylic and rainforest habitats have climaiic/habiiat changes towards a seasonal turret t-J i« snu in continuously humid habitats weather regime and more open habitats The n? si species lias piobabiy, similarly (trtU where some mils may experience extreme odhc Aname flooding). In contrast, of those flood-adapted developed contemporaneously, in response to the nests which occur now in semi-arid and and drying of the climate, in different geographic habitats most have evolved (in situ) from habitats but which also occur in areas subjected mesophylic denizens bill as a result of a general to some seasonal, torrenlial or at least very heavy drying of the habitat associated with climatic 1 which causes flash flooding. .

FLOOD AVOIDANCE IN TRAPDOOR SPIDERS MS

1981. Australian | ITERATURE CITED spiders: diversity, distribution and ecology. Pp. S07-852. In Keast, A. (cd.) 'Ecologi-

' ABRAHAM, H.C 1924. Some mygalomorph spiders cal biogeography of Australia (Junk; The from the- Malay Peninsula- Proceedings of the Hague). Zoological Society, London 1924; 1091-1 I1M. 1982a. Notes on the revised fnxonomic position of CLOUDSLEY-THOMPSON, J.L. 1982. Desert adap- the Black Wishbone spader Dekana diversictdor tations in spiders. Scientific Reviews on Arid Hogg (Mygulomorphae: Dipluridac). Journal of

Ihe Royal Society of Western Australia bS: . Zone Research 1 ; 1-14. 1983. Desert adaptations in spiders Journal of And 1 982b. Adaptations td arid habitats by Environments 6: 307-31 7. mygalomorph spiders Pp 273*283 In Barker, COOKSON, I.C. 1953. The identification of the W.R. and Grccnsladc. P.J.M. (eds) 'Evolution oi sporomoiph PkySocktdidkeS with tkicrydum the llora and fauna of arid Australia*- (Peacock Implications; Frcwvillc). and its distribution 111 southern Tertiary deposits. 1983. Systematic* of the trapdoor spider genus Australian Journal of Botany 1 : 64-70. COOKSON, l.C. & PIKE, KM, 1953. The Tertiary Homogona Rainbow (Mygalomorphac: occurrence and distribution of t'oducurpus (Sec- eteni?idne- Homogoninac) Journal of the lion Oaorycarpus) in Australia and Tasmania. Australian Entomological Society 22: 81-92 1985, Further studies the systcruatics of ctcnuid Australian Journal of Botany I; 7 1 "82 on trapdoor spiders: a review of the Australian 1 954 . The lossi I occurrence of PhyUocladux and I wo

gencv, Araneae: Mygalom«.(|>h..;- I other Podocarpaceous types in Australia ( 'tenmdaei. Australian Journal ot Botany 2. 60-88. Australian Journal of Zoology, Supplemeniary Scries 108: 1-84. COYLF., FA 1 98 1 . Hie role of silk in prey capture by non-arancomorph spiders. American Arachnof- 56- Trapdoors of Australian mygalomorph 7 ogy 24: 6. spiders- protection or predation' Pp. 95-102. In s CRAY, M.R. 1984. The laxonomy of Ihe Atmx BarfteMOS, J. A. fed.) Actas X Congrcso Inter- adetaidenxis species-group (Miicroihelinac; national de Aracnologia U (Jaca). Mygalomorphac) with notes Oil burrowing be- 19X7. Persistence of invertebrates in small ftcais haviour. Records of the South Australian Museum case studies of trapdoor spiders in Wl IS: 441-452. Australia. Pp. 29-39. fn Saunders, D.A., Arnold. HOFER, H. 1990. The spider community (Arancae) ol C.W., Burbidge A. A. and Hopkins. AJ.M. (pds) 'Nature conservation: role remnants » Central Amazonian blackwiter inundation the of of forest (igupo). Acta Zoologicu Fenmca 190: 173- C vegetation/ (Surrey Bcatty: Chipping 179 Narfl

MAIN, B.Y. 1956a. Taxonomy and biology dI the 1991a, Kimberlcy spiders: rainforest strongholds. genus Ixonwttotdes Kcyscrling (Scorpionida). Pp. 271-293. In McKen/ie. N.I., Johnston, R.B. Australian Journal ofZoology 4. 158-164. and Kendrick, P.G (eds) "Kimberlcy rainforests* (Surrey Bcatty. Chipping North). J 956b. Observations on the burrow and natural his- 1991b. Occurrence of the spider tory of l he l rap-door spider MisxuU'iw trapdoor gCMIS (Ctenizidne). Western Australian Naturalisi 5j Moggridgm in Australia with descriptions of two 73-80. new species (Arancae; Mygalomorphac: 1957. Biology of aganippinc trapdoor spiders Migidac). Journal of Natural History 25: (MygMomorphae: Ctcni?,idnc>. Australian Jour- 397 nal of Zoology 5: 402-473. MAIN. BY A MASCORD, R. 1974. Description and i960* The genus CetheeUs Thotcll (Mygalomor- natural history of a "lube-building" specie 0l phac. Maeroihcliuue). Journal oi the Royal Dyareyapy fro m New South Wales and Society of Western Australia 43: 30-34. Queensland (Mygaloinoipliue. Cteni/.idacJ. Jour- 1964 'Spiders of Australia.' (Jnearanda: Brisbane). nal of the Entomological Society of Australia H- 15-21. 1 969. The trap-door spider genus Catania Rainbow (Mygalomorphac: Ctcni/.idae) taxonomy und MASCORD. K. 1970 Austiahun spideis in colour." natural history. Journal of the Australian fin (A.H. and A.W. Reed: Sydney). toinologica) Society 9: 15-21. MCKBOWN. K. 1963. 'Australian spiders.' lA^us 1972. Tbe mygalomorph spider genus Stanwcllia and Robertson; Sydney). Rainbow & Pulleine(Dipluridne) and its relation- MOGGR1DGE, J.T. 1873. "'Harvesting ants and trap ship to Aname Koch and certain other diplurinc door spiders.' (Reeve. London). genera. Journal of the Roval Society of Western I874w 'Supplement to harvesting aiUS and trap-duoj Australia 5: 100-114 spiders.' (Reeve: London).

1976. "Spiders/ (Collins; Sydney). R AINBOW, W.J. & PliLLEINE, R 1 1 1 91 S Austn.liiin 1977. Spiders. Pp. 100-107. In Kcnneally, K. (ed.) trap-door spiders. Records of the Australian

'The natural history nf the Wnngon Hills' 1 Museum 12: 81-169.

nook. 1 1, t Western Australian Naturalist' % Club: RAVEN. R.J. I9S4. Systematic* of the AustrabanCur- Perth). tain-weh spiders (Ischnothelinae: Dipiuridae: 606 MEMOIRS OF THE QUEENSLAND MUSEUM

Chelicerata). Australian Journal of Zoology, Sup- RAVEN, R.J. & CHURCHILL, T.B. 1991. A revision plementary Series 93: 1-102. of the mygalomorph spider genus Encyocrypta 1985. The spider infraorder Mygalomorphae Simon in New Caledonia (AraneaeBarychelidae). (Araneae): cladistics and systematics). Bulletin In Chazeau, J. & Tillier, S. (eds). 'Zoologia of the American Museum of Natural History 1 82: Neocaledonica 2.* Memoires du Museum Nation- 1-180. al d'Histoire Naturelle, Paris (A) 149: 31-86. 1986. A revision of the spider genus Sason Simon TRUSWELL, E. 1990. A curious and diverse flora. Pp. (Sasoninae, Barychelidae, Mygalomorphae) and 25-31. In Neville, S. (ed.) 'The Australian en- its historical biogeography. Journal of Arachnol- vironment, taking stock looking ahead.' ogy 14: 47-70. (Australian Conservation Foundation: Canberra). INTRASPECIFIC PREDATION IN CWBIONA CORTICALIS (ARANEAE: CLUBIONIDAE)

P. MARC

Marc, P. 1993 II 1 1: Intraspecific predaiion in Clubiona corticalis (toaneac: Clubionidac) Afemoirs of the Queensland Museum 33(2); 607-614. Brisbane. ISSN 007WO

Clukktna corticalis is supposed to be a useful biological control agent of Jepidopteran larvae Tortricidac) in orchards. Thus, the levels of its intraspecific prcdation especially when bred in captivity must be determined. Predator) is a risk at mating and dispersal. Cannibalism during mating, studied at 13°C, 20°Cand 27°Cand without prey in cages, remains low (1.5%) during the first 24 hours at all temperatures. After 24 hours cannibalism increased and was significantly higher at 27°C than al 13°C after one week. Males were 723fr of the victims. Males can mate twice with different without increasing the risk of cannibalism. The spiders were studied for 12 days after dispersal. Until Day 3, no intraspecific prcdation occured, whatever the availability of prey for juveniles. Intraspecific prcdation is very much reduced by the availability of sufficient prey, and a female given prey did not feed on its own progeny. Clubiona conkolis pourrait etre utilised en verger pour lutter contre des larves de

lepidopteres (Toruicidae). 11 ctait done important de connattre Ics niveaux dc prcdation intraspccifiquc chez eelte cspece, lorsqu'clle est ef evce en captivity. II csl montre- qu'il y a ) des risques dc predaiion a deux moments du developpemenu ' accouplcmcnl et la dispersion. Le cannibalisme durant I'accouplemenl, etudie a 13°C, 20°C el 27°C sans proie dans les

enceintes, resle faible a toutcs les temperatures ( 1 .5%) pendant les premieres 24hcures. Apres 24 heures, le cannibalisme augments. I) est significativement plus fort a 27°C qu'a I3°C apres une scmaine. Les males sonl les victimes dans 72% des cas ct pcuvent s'accoupler deux fois avee des femellcs dilTcrcntes sans augmenter les nsques de cannibalisme. Les jcunes araignces sont etudices pendant douze jours apres la dispersion. Quelle que soit la dispouibilite en proies.aucun cas de predation intraspecifiquen'imervient jusqu^aulroisieme jour Ensuite, la disponibilite en pmies dimunue de futon trcs significative 1c niveau de canabaJisme. A noter qu'une femelle disposant de proies, n'exerce aueuoe predaiion sur sa progefiiiure.Qfiitf/cjgv, Eurcpc, laboratory, Araneue, France, breeding, auxiliary, in-

traxprcijlc predation, Clubiona t orticatis, dispersal, mating.

J I . Marc, Laboraioire dc Zontoyie el d'lirophysiologie, L.A. 1NRA, Universite etc Hennes I,

dc Heaulieu, 35042 R< fin ( WA'i. }9 January. 1993.

Clubionids arc wandering, nocturnal spiders In Anyphaena accentuate (Walckcnacr), (Marc, 1990a). and certain species arc efficient in Philodromus cespiturn (Walckcnacr) and Diaea caterpillar control in orchards (Mansottr et al., donala (Fabricius) (the most abundant non- web 1980; Marc and Canard, 1989). The study of the spinning spiders in apple orchards), intraspecihe araebnofauna in apple orchards led us to lake an and interspecific recognition minimizes can- active interest in whether Clubiona corticalis nibalism (Marc, 1992). Whether this is true oi'C (Walckenaer)canbe used in biological control of corticalis is not known. C. corticalis occurs in pest caterpillars In the laboratory, all instars of very high densities on pines (Marc, 1990b) where C. corticalis can consume large quantities of they build silk nests under bark for refuge during lepidopteran larvae, especially tortricids, harmful ihc day (Mare, 1990a). Partially flaking bark arc to orchards (Marc, unpublished data). However, favoured shelters on which concentrations of 6- 2 the efficacy of using C corticalis in biological 12 nests can be found in 10-15cm (often the control remains uncertain. Spiders are generally clustered nests are stuck together). Furthermore, envisaged as having high rates of intraspecific nests of other clubionids, Clubiona brevipes and interspecific predation (Bristowe, 1958; Blackwall, C. leueaspis Simon and Ceto laiL Greenstone, 1978). If thisis true for C. corticalis, (Canestrini), often adjoin those of C. corticalis. then its effectiveness in biological control would These aggregations suggest low intraspecific and be reduced, i.e., an increase in spider densities in interspecific predation in nature. However. the field would increase cannibalism. Thus, data laboratory bred spiders for use as auxiliaries may on rates of cannibalism in C conuaiis are present special problems. In captivity, imr;i needed specific predation at critical moments of develop-

608 MEMOIRS OF THE QUEENSLAND MUSEUM

ment may be artificially high because contact RESULTS AND DISCUSSION between individuals is more frequent. Mating and dispersal are likely to be critical periods (Fig. 1). Cannibalism during Mating This work determines levels of cannibalism in C. In the final moult the reproductive organs are corticalis when bred in captivity. completely formed. Trophic, and locomotory activities different behaviour in two sexes MATERIALS AND METHODS and very of the C. corticalis species also appear (Marc,

1 992). In the field, males become adults about 1 Subadult spiders (one instar before maturity) days before females (Marc, unpublished data). were collected in winter, in a forest under the bark Males then seek females, inside whose nests of Pinus sylvestris. Removed from the field in mating takes place. Because of this asynchrony Ependorff micro-tubes, they were separated into of maturity, males frequently mate with females males (recognisable at this instar by a bulging soon after the latter moult and at which time palpal tarsus) and females. Each individual was females do not feed; they remain in their nest then placed in Petri dishes (9 cm in diameter and without feeding for 1-2 days after moulting 1 .5 cm high) for breeding. Damp cotton wool kept (Marc, 1990a). These two days appear, therefore, humidity high in cages. Individuals were fed with to be an especially favourable time for the male two flies (Lucilia, Diptera) three times per week. to mate because, during this time, the female is Unless otherwise specified all procedures and only slightly aggressive. Males may also build a observations where carried out in the laboratory. nest adjoining those of the subadult females and All enclosures are here referred to as 'cages'. thus fertilise them soon after the final moult. This was also observed by Austin (1984) in C. robusta Cannibalism during Mating L. Koch and by Wolf (1990) in Cheiracanthium pennyi Cambridge and Cheiracanthium In each of 1 35 Petri dishes, one male was placed punctorium (Villers). The mating of many other with a virgin female, and no food was added for species takes place just after the final moult of the week of the experiment. Breeding cages were females. Jackson (1980a) reported that 161 kept at 13°C, 20°C and 27°C under long photo- species use tactics of cohabitation (the male stays period (L-D=16-8), and were observed several with the immature female to mate with her after times a day. Males and females used for mating she has undergone her adult moult). had all moulted to adults at least two weeks before. Here, successful mating could be obtained several weeks after the final moult of both sexes. The male approaches the female by tapping on Cannibalism After Dispersal her nest. Copulation, which takes place inside the Dispersal, in this species, is when the juveniles female's nest, lasts for about 3 hours. The male, at instar 2 leave the female's breeding-nest. Can- mounted on the female's back and facing the nibalism, from dispersal (when the first juvenile same direction, forms an angle of about 45° to her. leaves the breeding-nest) for 12 days was ob- From this position, he applies each palp consecu- served on eight egg masses from mating carried tively into her epigyne. One male can fertilize the out. The egg masses were inside the female's eggs of at least 3-4 females. Mating can also take cages (25ml Petri dishes). These observations place outside the female's nest without can- were carried out at 20°C and in a L:D photoperiod nibalism. On the other hand, in other species, of 16:8. Two groups of 4 egg masses each were intraspecific interactions outside the nest were set up: more 'lethal' (Jackson, 1988a; Jackson and Mac- - no prey for the juveniles. In two cages the nab, 1989). female had been separated from her progeny. Intraspecific predation was observed in 13% of - Drosophila provided for juveniles. The the cases for all matings (Fig. 2). The most female was taken out of one dish. favourable conditions to reduce the risks of can- Lucilia were given to the females left in the nibalism during mating would have been to place cages every day from the second day after the the males in the female's cages just after the dispersal of the first of the juveniles. Trie caged female's moulted, or, at least, in the presence of spiders were observed 2-3 times a day for 1 2 days prey. Yet, here, males were brought to females and the status-alive, death by cannibalism, or which had moulted at least 15 days before. Fur- death by other causes, or still inside the nest- thermore, no food was provided. Because, to be recorded. practicable for production of auxiliary popula- i

INTRASPECIRC PREDA I ION IN CLUBIONA CORTICAUS 609

., .1,., i ...,..., ,.., :

''-:< E3-!IEEl -j \- 1

PrVt|3I> r- . J ...... iv ... i -.j Aoul r-L •

Bajiod .

rr'i.i 33 rwrnperotur* ("C) - FfW K -

HG. 2, lnuaspeeilic prcdation rate during Ban >i t ii.i mating ex- Me cycte tftjration of Oubiona corffcofc periment on C. cortu aJfs without prey at 13, 20 and 2TC. Symbols indicate the period of predation, n = no. of pairs tested; values on histogram bars =no. 1. Life cycle corticatis. FIG. of Clubiona Terminology cases recorded. i n Canard ( i 987). In.l, in.2 and in3 indicate the foetal msiars, and in.l to in. 10 the juvenile ones (last this has no effect on the number of egg masses or juvenile instars, 9 and 10 not present in every in- t on the descendants produced by the female. At dividual). Numbers in parentheses indicate duration 13°C no intraspeeific predation occurred before iu days. * (2-3) refer to Time spent by juveniles close 48 hour*, and the females were fertilized normal- to breeding-nest just after dispersal. ly and laid fertile eggs in the same wav as at2CfC and (tons for biological control, the time needed for 27°C. cure during laboratory breeding must be minimal. In all pairs where intraspeeific predation oc- Synchronized mating was more practicable. curred oivce in the week after the start of the Likewise, the absence of prey in the mating cages experiment, males are more often (72%) the vic- Simplifies the work. Cannibalism observed after tim. Wc did not observe a relation between one week (13%) was very much higher than that weight and (he individuals eaten, and sometimes hi the field. smaller individuals ate larger ones. There is no significant sexual dimorphism in C. corticalis. Predation before 24 hours, whatever the In contrast, even though size is not important, temperature, were rare ( 1 .5%) (Fig. 2), but there- males do move more than females (Marc, 1990a). after increased with time and after one week the Conceivably, this weakens males rclativr [0 rate of cannibalism was significantly lower (p) at similar sized females. Furthermore, males who a lower temperature (7,5% at I 3°C and 18.5% at have smaller trophic needs and mate several 2'?°C). times with different females display less aggres- The spider's activity is probably related lo the siveness towards their partners. This may partly temperature. Indeed, locomotory activity in- explain why males were more often the victims creases with temperature and greatly increases Of cannibalism. the probability interindividual encounters and, of A male can mate once or twice without sig- therefore, the risks of aggressiveness, especially nificantly increasing the rate of cannibalism aftrr in the absence of prey. one week (9% for one mating and 14% for two Cannibalism in 24 hours was rare, whatever the matings; p>0.05). However, after two fu temperature, even though the trophic needs of matings (3 or 4), intraspeeific predation increases females, especifically after adult moulting, must significantly to 33% (p<0.05) if compared he high considering ovule production. Tins can male which mated only once. However, the male be explained simply by the smaller need for food is not always the victim. after only one day of fasting, when this lime is In C reirhiim Schenkel, males and females can sufficient to guarantee mating. Indeed, the pair- mate several times without cannibalism ing time was quite short. Within 5 hours of the (Hc-ngmei andHongquan. 1987). Equally, in in- individuals being brought together, 75% of the teractions of C. cambridgei L. Koch, no ean- matings had begun, and mating itself lasted 3 nibalism occurred between virgin males and hours. Wc observed, however, that one male females during the 38 interactions studied (Pol- could mate several times with the same female lard and Jackson, 1982). Generally, species of which lengthens the interaction, but it seems that Clubiona do not seem to be very aggressive 611) MEMOIRS OF THE QUEENSLAND MUSEUM

6 A»re —•— Out bre«dlng-neil Oui ot breeaing-neai

—Q-- D«a1h by cannibalism Death by canntbol'Bm

—o— Other mortality Other mortality

Days

FIG. 3. Survival of juveniles (2nd instar) of C. corticalis during 12 days after dispersal, without prey. FIG. 4. Survival of juveniles (2nd instar) of C cor- ifcofein the presence of prey. toward each other during mating- Apparently, to eliminate cannibalism in C. corticatis, males and jtl over 4 days (Fig. 3). There was no sig- females should be brought together at about 13°C nificant difference between the cages with a and individuals should be kept paired for about female and those with one removed (p>0.05). 24 hours. Furthermore, the presence of prey at Thus, females did not attack their own progeny G 13 C should make the cohabitation of males and after dispersal for at least the first 12 days and on females perfectly feasible for several days the condition that prey was available Can- without intraspecific predation occurring. Such nibalism between the juveniles first began 3-4 conditions during mating should minimize the days* after the first dispersal and coincided with manpower needed and lower production costs. the expression of the first agonistic behaviour observed. Next, the number of juveniles decreased until there are about 10 individuals per Cannibalism After Dispersal cage towards the eleventh day. Juveniles leave the breeding-nest, built by the female, after 1-4 days. Then, they remain grouped When the juveniles were fed Drosophita (Fig. around the nest for about 2-3 days before disper- 4), the female's influence on this period of her progeny's similar. the other sal. This gregarious phase lasts 17-20 days at development was On 20°C (15-17 days in the breeding -nest and 2-3 hand, intraspecific predation was almost non-cx- days around the nest) (Fig. 1). It is only from the istant in the first 6 days and continued to be juvenile instar 2 (as defined by Canard, 1987), minimal afterwards, being about 5% after 10 that juveniles begin to hunt Until leaving the days. There is a highly significant difference be- breeding-nesc juveniles use on their vitelline tween the two groups after the 12 days of study reserves and, some attack the undeveloped eggs (p<0.001) Thus, with prey in the breeding cage, in the nest. Indeed, weight differences between cannibalism decreased and was almost juveniles leaving the breeding-nest indicate the eliminated. Rypscra (1983), likewise, found in existence of trophic activity in some individuals* several spider species that intra-individual as there are no significant weight differences tolerance increased and cannibalism decreased

known to occur between eggs in the same batch when maintained at extremely high prey levels - Also, in the spider (Lecaillon, 1905). Furthermore, this KiaftXetaL (1986) were able to prolong the trophic activity does not seem linked to the juvenile social period by giving juveniles abun- female feeding her juveniles by regurgitation or dant food. Austin (1984) recorded a high mor- the consumption by the juveniles of a trophic tality in the breeding of C robusta. cannibalism egg-mass as, for example, in Atnaurohius being one of the two major causes of mortality (Amaurobiidae) (Tahiri et at.. 1989). At this On the other hand, in nature, spiders, which arc potential prey of the in point, mortality is the highest in many species highest density the en- (Austin, 1984). It may be even more delicate in a vironment, only represented only 3% of prey captive breeding situation as the juveniles, in actually consumed (Austin, 1984). Austin sug- large numbers in the cages, would devour each gested that ihe highest mortality occurs at the other. dispersal instar. The dispersal of the juveniles given prey can The rest of the development presents fewer INTRASPECIFIC PREDATION IN CLUB IONA CORTICALIS 611

Fam Genus or species Authority Familv Genus or species Authority

Club Supuntia picta Jackson & Poulsen (1990) Anyphaenidae Anyphaena accentuata Marc (1992)

Gnap Taieria erebus Jarman & Jackson (1985) Olios diana .O. Heteropodidae lamarch ,0. obesulus Jackson (1987) Liny Oedothorax insecticeps Kiritani etal (1972) O. sp.

Lvco Lycosa pseudoannulata Kiritani f r a/. (1972) Porrhothele Hexathelidae Jackson Pollard (1990) Pardosa lugubris Edgar (1969) antipodiana &

P. purbeckensis Schaefer(1974) Philodromidae Philodromus cespitum Marc (1992)

P. ramulosa Yearganfl975> Salticidac Holoplatys sp. Jackson & Harding( 1 982) Pirata piraticus Schaefer(1974) Myrmarachne lupata Jackson (1982b)

Gerhardt (1924), Bristowc Thomisidae Diaea dorsata Marc (1992) Mime Ero aphana*. E.fttrcata* (1941), Czajka (1963), Canard (1984)

Mimetus maculosus* , M. TABLE 2: Literature review of araneophagic spiders: Jackson & Whitehouse (1986) sp.* k species which are little or not araneophagic. , species Oxyo Pcucetia viridans Turner (1979) kleptoparasitic which are not araneophagic. Phol Holocnemus pluchei Blanke(I972) days (3-5) in the cages did not result in in- Jackson & Brassinglon (1987), Pholcus phalangioides* Jackson & Rowe (1987) traspecific predation, but behaviour of escape and avoidance was observed. Similar observations Brettus adonis * , B. Sail Jackson & Hallas (1986a) cingtdatus"* had been made on sub-adults and adult females Cocalus gibbosus Jackson (1990b) bred at 25-30 individuals per cage (30x20x20

Cobanus mandibularis Jackson (1989) cm) over 2 months. Cyrba algerina * Jackson & Hallas (1986a) Therefore, in breeding C. corticalis, the

C. ocellatd°* Jackson (1990c) provision of Drosophila to juveniles at the disper-

Euryatlus sp. Jackson (1985a) sal stage should be sufficient to eliminate in-

1 '* Geiotia sp. Jackson (1990d) traspecific predation. Then, 5 days after dispersal Jacksonoides when all juveniles are out of the breeding-nest Jackson (1988a) queenslandicd3 and no cannibalism has occured, the division of

Phaeacius malayensis, P. Jackson & Hallas (1986a), these juveniles at instar 2 with about 5 per Petri sp. Jackson (1990a) dish, should prevent cannibalism later. Plexippus pavkulli Jackson &Macnab( 1989) Ph idippus Johnson i Jackson (1977) CONCLUSIONS Portia * Jackson & Hallas (1990> Jackson (1982a, 1986b). Intraspecific Jackson & Blest (1982), predation in C. corticalis during Portia fimb riata°* Jackson & Hallas (1986b), two especially susceptible periods of develop- Jackson & Wilcox (1990) ment (mating and dispersal) involves limited * P. africatxa , P. Jackson & Hallas (1986b) risks of cannibalism which can be eliminated. albimana"* Mating must be at 13°C. Males in the enclosures P. labiata *, P. shultzC* Jackson & Hallas (1986b) with females must be limited to 24 hours, and Simaetha paetula Jackson (1985b) dispersing juveniles must have sufficient prey. Tauala lepidus Jackson (1988b) Cannibalism should not, therefore, be an obstacle Scyl Scylodes longipes Nentwig(1985) to the mass breeding of this species for biological The! Achaearanea camura Jackson (1988b) control. Achaearanea tepidiorum Rypstra(1986) Rates of intraspecific and interspecific preda- Rbomphaea Enders(1974) tion have often been considered very high in spiders. The main enemies of spiders are often TABLE 1: Literature review of araneophagic spiders: other spiders (Bristowe, 1941; Foelix, 1982). species principally or strongly araneophagic. *, 'ag- Certain species partially practice araneophagy gressive mimicry' = to perform a variety of vibratory (e.g. Pardosa lugubris (Walck-enaer), Lycosa behaviour in which the prey-spider responded as it annulata Thorell), and a few make it their normally would to its own prey. °, oophagy. speciality (e.g. Mimetus, Ero, Portia) (Table 1). problems of cannibalism. Six groups ofjuveniles In fact, the species most studied for cannibalism are araneophagic in nature. In contrast, in other were bred together with 5 per Petri dish from species, araneophagy appears to be almost non- instar 2-6 cannibalism was noted. Further- and no existant (Table 2) even without prey (e.g. more, the periodic absence of prey during a few Anyphaena accentuata, C. corticalis, Diaea dor- .

fi12 MEMOIRS OFTHr: QUEENSLAND MUSEUM

Lit balance nature. sata, Philodromus cespitum) and it is absent in to the role of spiders the of the Zoological Society of London social spiders such as M alios gregalis (Simon) Symposium of VI ! 83-193. (Jackson. 1979, 1980), Behaviour of a few spider HENGMEI, Y. &. HONGQUAN. W.B. 1987 A study species, eannert be applied to all. The degree of OD the biology of the spider Clubiona rett araneophagy of a species must be based on only ActaZoologica Sinica 33: 255-261 thdtone. Therefore, ihe levels of intraspecitlc and JACKSON. R.R. 1977. Prey of the jumping spider interspecific ptedation amongst the more abun- Phidippusjohnsont (Araneae: Salticidac). Journal dant species in agrosystems in which a spider is of Arachnology 5: 145-149. a possible biological control agent must be 1979. Comparative studies of Dictyna and Malios studied. A simple^method has been finalised for (Araneae, Dictymdae). II. The relationship be- spiders which do no4 spin a web (Marc, 1992). tween courtship, mating, aggression and cannibalism in species with differing lypcs of social ACKNOWLEDGEMENTS organisation. Revue Arachnologique2: 103-132. 1980. Comparative studies of Dictyna and Malios: V Tolerance and resistance to starvation. Psyche Thanks to Didier Capdevillc for his help in 87:211-220 breeding the species studied . 1982a. The biology o\'Portia> a web-building jumping spider {Araneae, Salticidac) from LITERATURE CITED (^jeensland: intraspecific interactions. Journal of Zoology (London) 196: 295-305 AUSTIN, AJ>. 1984. Life history of Chduunu robiuM 19#2h. The biology of ant-likc jumping spiders: inlrnspecific interactions oj Myrmaruchne lupaia L. Koch and related species (Araneae, Journal of the l Zooloeical Clubkmidae) in South Australia. Journal Of Araneae, Salticidac). Arachnology 12:87-101. Linnean Society 76: 293-219. BLANKE, R 1972. Llntersuchungcn zur 1985a. The biology ot Euryaftus sp. uidet., a web- Qkophysiologie und Okelhologic von Cyr- buikiing jumping spider (Araneae, Salticidac) in from Queensland; Utilization of silk, predatory tophora citrkvUi Forskal (Araneac v Araneidae I Andahisicn. Forma Function 5: 125-206. behaviour and intraspecific interactions. Journal of Zoology (London) 1: 145-173. BR1STOWE. W.S. 1941 . The Comity of Spiders' Vol 2, (Ray Society: London). 1985b. The biology of Simactha pactula and S. thoracica, web-building jumping spiders 1 V5». The Clubionoidca - Gnaphosidae,

' Amneac, Salticidac) from Queensland: co- Clubionidacand Anyphacmdac. Pp. 1 16-132. In, The World of Spider*' Ed, Collins (revised habitation with social spiders, utilization of silk. 1971: predatory behaviourand intraspecific interaction. 175-210. CANARD, A. 1984. Contribution a la connaissancc du Journal of Zoology (London) 1: developpement, dc t'cculogie, ct de 1986a. Cohabitation of males and juvenile females: leeopbysiologic des Aranc'tdcs de landea «r a prevalent mating tactic of spiders. Journal of

moricaines. (These de Doctoral Etat, Renncs, Natural History 20: 1 193-1210 versatility, and the 389pp* annexe i 1986b. Web building, predatory 1987. Analyse nouvelle du developpemeiu pos evolution of ihe Salticidac. Pp. 232-267. In,

tcmbryonnaire des araigne"es. Revue Arach- Shear, W. A (ed ). Spiders: Webs, Behavior, and nologique7: 91-128. Evolution'. (Stanford University Press; Stan- CZAJKA, M. 1963. Unknown facts of the biology of ford)

the spider Erofunata (ViJlers) (Mirnetidae, Aran- 1987. The biology of Olios spp , huntsman spiders

eae). Polskie Pismo Entomologie/.ne 33: 229-23 1 (Araneae, Sparassidac) from Queensland and Sri ENDERS, F. 1974. Vertical stratification in orb-web Lanka: predatory behaviour and cohabitation social spiders. Bulletin the British spiders t AraneidaC: Araneae) and a consideration with of Arach- of other methods of coexistence. Ecology 55: nological Society 7; 133-136.

317-328. 1 988a, The biology oUacksnnoides qucenslandka, EDGAR. W.D. 1969. Preys and predators of the wolf a jumping spider (Araneae. Salticidae) from

spider Lycosa lugubris. Journal of Zoology f Lon- Queensland: intraspecific interaeuons, web-in- don) 159:405-411. vasiQH) predators, and prey. New Zealand Journal 15: 1-37, FOELIX, R-F. 1982. 'The Biology of Spiders' i Har- of Zoology vard University Press: Cambridge, Mass). 1988b. The biology oi Tuuuiu Itpulus, a jumping GERHARDT. U, *1924 Wcttere studicn fiber die spider (Araneae: Salticidac) from Queensland: bioloeie der spinrven. Archiv lur Narurgeschichtc display and predatory behaviour. New Zealand 90: 85-192. Journal of Zoology 15: 347- 364. GREENSTONE. M H 1978 The numerical response 1939. The biology of Cobanus mandibularis, a in prey availability of Pardosa rumulasa (Mc- jumping spider (Araneae: Salticidae) fromCosla

10 i Araneae: Lyco'tdac) and its relationship Ricy; intraspecific interactions, predatory he- " 3

INTR ASPECIFIC PREDATJON IN CLUBIONA CORTICALiS 61

haviottr, and siik utilisation. New Zealand Jour- PorHwtheie antipodiana (Araneae: nal of Zoology 16: 383-392. He.xathelidae't. New Zealand Journal of Zooloav 1990a. Ambush predatory behaviour of Phaeucius 17:499-526. matayensis and Pfmeacius sp. indct., sparlneine JACKSON, R.R. & POULSEN, B.A. 1990. Predatory

- jumping spiders (Arancae'. Salt iciiioc.) froin t rop versatility and sncraspeeifie interactions ofSupun-

cal Asia. New Zealand Journal of Zoology I na picta (Araneae: Clubionidae). New Zealand C 4 .» 1-498. Journal of Zoology 17: 169-184,

1 990b. Predatory and nesting behaviour of Cocalus JACKSON, R.R &ROWE. R.J. 1987. Web-invawvt gibb&SUS* a spartacine jumping spider (Arancae and araneophagy by New Zealand and Australian Saltiridae) from Queensland. New Zealand lour- pholcxl spiders. New Zealand Journal of Zoology

na! of Zoology 1 7 483-400 14; 139-140. 1990c Pled wt;iiiliTy and tnttaspecific inter- JACKSON. R.R.&WHITEHOUSE,M.E.A. 19B*>.The actions of Cyrhti tdjjermtl and Cyrha cicW/fJ/0, hiology of New Zealand and Queensland pirate web-invnding spartacine jumping spiders vpidets (Araneae, Mirnctidae): aggrciiive (Araneae: Sailieidae). New Zealand Journal n. mimicrv, araneophaev and prey specialization.

Zoology 17- IS7 i , Joumafof Zoology (London) (A), 210: 279-303 Predatory and siltc utilisation befuvu JACKSON. R.R. & WILCOX. R.S 1990. Aggressive Gclotia sp. tndei. (Aianeac. Sailieidae! Spar- toy, prey-specific predatory behaviour and laeinae), a web-invading aggressive* mimic from predalOT-recognition an predator-prey inter.;

1 Journal of Zoology 17: Sri anka New Zealand of Portia ft wbriata and Eusyattus sp., jumping 475-4S2. spKiers froon Queensland. Behaviour Ecology and J hi KS< >N. K.K. & BLEST, A.D. 1981 The Mo]Oj Sociobiolagy 26: 111-119. Pnrriti fl/nhrium, a welvbuilding jumping spider JARMAN. EA.R.

bmlogy of Phofrus phaksttgioid&S I Arancae. Phol- KIRITANI. K.. KAWAHARA. S.. SASABA, V. & cidac): predatory versatility, arancophugy and ag- KASUJli F, 1972. Quantitative evaluation of gressive mimicry. Journal of Zoology (London:- .:iion by spiders on the green rice leafhopper, 211:227-238. Neptovtettix vinticeps Uhlcr, by a sight-count JACKSON.R.R, & WALLAS, S.E.A. 1986a. Predator, method. Researches on Population Ecology. versatility intraspecific interactions spar- and of Kyxito University 13: 157-200.

. in« jumping spiders (Arancae: Salticidae i: KRAFFT, B„ HOREL. A & JULTTA, J.M. 1986. In- Brants adoniSt tingulatus, Cyrba and R olgerifldi fluence of food supply on the duration of the Phanuius sp. indet. New Zealand Journal of gregarious phase of a maternal-social spider, 15:491-520. Coek-tc? temsfrM (Araneae, Agelenidae) Jour- 198M>. Comparative biology of Portia afmana, P. nal of Arachnology 14: 219-226. nlbimana, P.Jimhriatu, P. iabiatu. and P. shuttzL LECA1LLON, A. 1905. Nouvelles recherche* sur Ja arancopnagic. web-building jumping spiders biok>gie et la psychologic des Chiracarthion. Bu I (Araneae: Salticidae): utilisation ol wehs. Jelin dc V Association Philomatique d' Alsace el de predatory versatility Bfld intraspecific interac- Lorraine 7: 224-252. tions. New Zealand Journal of Zoology 13: 42?- MANSOUR, F, ROSEN, D.. SHULOV. A. & PI .Al IT, H.N 1980 Evaluation of spiders as biological .i i i 1900 . :m displays used in ag- contml agents oTSpodoptem littonalis larvae on gressive mimicry by Portia, a web invading apple in Israel. Acta Oecologiea, Oecologia Ap- ataneophagic jumping spider t anosr Sal- plieata 1: 225-232. ticidae). New Zealand Journal ol Eoologj I / 7-23. MARC. V. 1990a. Nycthennal activity rhythm of adult Clithiona corticalis (Walckcnaer, 1802) (Araeh JACKSON. R.R. & HARDING, DP I9R2. In nida. C'uhionidae) Acta Zc>ologica Fennica 190: traspecific uircTBi tioiu of Hofoplaty& sp. indet.. a

27v ; New Zealand jumping spidci lAraucuc liddae). New Zealand Journal of Zoology 9: I990tx Oonn(5cs surle peuplemeni d'Araneides; des 510. tomes de pins. C.R. XII Coll. Europccfl d'- 255- JACKSON, R.R. & MACNAB. AM 1989. Display Araehnologie, Paris (1); 260. mating, and prni.itory behaviour of the jumping 1992. Intraspecific and interspecific interactions be- spider Phximms paykuiiii (Aianeac: Swticioac). tween spiders from apple orchards. Comptes New Zealand loiim.il of Zoology 16; l*>l 166, Rendu XIII Coll. European d'ArarachnoIogie, Neochatel(Sui JACKSON. R.R. k POLLARD. SU 1990, IH- traspe<:ific interactions and the function ol MARC. P. ^ CANARD, A. 1989. Us essais courtship in rnygolonrorph spiders: a study ol' d'utrlxahon ties Araignees en lultc biotogique. 614 MEMOIRS OF THE QUEENSLAND MUSEUM

Bulletin de la Society Scientifique de Bretagne 14: gen zur Bedeutung der interspezifischen Konkur- 149-172. renz bei 3 Wolfspinnen-Arten (Araneae: NENTWIG, W. 1985. Feeding ecology of the tropical Lycosidae) einer Salzweise. Zoologische spitting spider Scytodes longipes (Araneae, Jahrbucher, Systematic (Okologie), Geographie Scytodidae). Oecologia (Berlin) 65: 284-288. undBiologie 101: 213-235. 1986. The prey of spiders. Pp. 249-264. In, W. TAHIRI, A., HOREL, A. & KRAFFT, B. 1989. Etude Nentwig (ed.), 'Ecophysiology of Spiders'. preliminaire sur les interactions mere-jeunes et (Springer- Verlag: Berlin). jeunes-jeunes chez deux especes ftAmaurobius POLLARD, S.D. & JACKSON, R.R. 1982. The biol- (Araneae, Amaurobiidae). Revue Arachnologi- ogy of Clubiona cambridgei (Araneae: que8: 115-128. Clubionidae): intraspecific interactions. New TURNER, M. 1979. Diet and feeding phenology of the Zealand Journal of Ecology 5: 44-50. green lynx spider, Peucetia viridans (Araneae: RYPSTRA, A.L. 1983. The importance of food and Oxyopidae). Journal of Arachnology 7: 149-154.

space in limiting web-spider densities; a test using WOLF, A. 1 990. The silken nest of the clubiorud spiders field enclosures. Oecologia (Berlin) 59: 312-316. Cheiracanthium pennyi and Cheiracanthium 1986. High prey abundance and a reduction in can- puncturium (Araneae, Clubionidae). Acta nibalism: the first step to sociality in spiders Zoologica Fennica 190: 397-404. (Arachnida). Journal of Arachnology 14: 193- YEARGAN, K.V. 1975. Prey and periodicity of Par- 200. dosa ramulosa (McCook) in alfafa. Environmen- SCHAEFER, M. 1974. Experimented Untersuchun- tal Entomology 4: 137-141. COMPARATIVE MORPHOLOGY OF THE SEXUALLY DIMORPHIC ORB-WEAVING SPIDER ARGIOPE BRUENNICHI (ARANEAE: ARANEIDAE)

MONIKA C MULLER AND WILFRIED WESTHEIDE

Muller, M.C. and Westheide, W. 1993 11 11: Comparative morphology of the sexually dimorphic orb-weaving spider Argiope bruennichi (Araneae: Araneidae). Memoirs of the Queensland Museum 33(2); 615-620. Brisbane. ISSN 0079-8835.

Although the web building spigots of the glandulae aggregataeand the glandulae fiagellifor- mes are not functional in mature males, adult Argiope bruennichi and A. lobaia males are able to build webs. The structure of these webs is described. The aciniform spigots on the intermediate spinnerets of A. bruennichi males have degenerated to a greal extent. Culture experiments with A. bruennichi enabled us to follow differences in the development of female and male morphology. The presence of the tubuliform spigots in the sixth instar suggests one possible evolutionary hypothesis concerning sexual dimorphism in spiders. Nach der Reifehautung sind die aggregaten und flagelli (brmen Spulen-welche als Triade die Fangfaden sezemieren-auf den hinteren Spinnwarzen der Mannchcn nur rudimentar ausgebildet.Trot/.dieserReduktionspannenaduUc^r^/op^/jrH^nrtfc/i^undA. lobataMSnnchcn Radnetze, deren Strukturbeschricben wird. Auf den mittleren Spinnwarzen subadulterund adulter A. bruennichi Mannchcn ist ein hoher Prozentsatz der aciniformen Spulen unvollstandig ausgebildet. Durch Aufzucht der Wcspenspinnc war cs moglich, morphologic sche geschlechlsspezifischc Untcrschicdc in der Postembryonalcntwicklung fcstzustellen. Das friihe Auftreten der tubuli formen Spuien irn weiblichen Spinnapparat kannals Argument ftir cine Hypothcsc zur Evolution des Sexualdimorphismus interpretiert wcrden. Q% Development, sexual size dimorphism, spinning apparatus, male orb-webs, tubuliform fpt$0t.

Monika C. Muller and Wilfried Westheide, Universiiat Qsnabruck, Fachber. lhalogie/Chemie, Spezielle Zoologie, D-49069, OsnabriicL Germany; 28 October, 1992.

Sexual dimorphism varies across taxa, but the (Scopoli, 1 772) were collected from Dabas (Hun- question of whether highly dimorphic species gary ), and subadult males and cocoons from Wit occur as a result of selection for large female or tenberg-Lutherstadt (Germaj>y). small male size remains controversial. Specimens of A. bruennkhi were raised in- Gerhardt (1924) observed that both car- dividually from the third instar. Some individuals nivorous feeding habits and cannibalism en- from each instar were fixed in Camay's fluid danger male spiders before, during and after immediately after moulting. The lengths and copulation (Elgar and Nash, 1988). Darwin widths of the prosoma were measured. For SEM (1890) and Brislowe (1929) suggested that their studies the spinning apparatus was removed Atid Small size protects the males from the females the spinnerets were dehydrated in ethanol and which do not recognize males of reduced size as critical point dried with carbon dioxide. After prey. In contrast, Gerhardt (1924) and Vollrath sputtering with gold, they were analysed with a

1 1 980) argue that 'the females have evolved Jo be Cambridge Stcreoscan 250. For light microscopy larger, allowing greater egg production* the spinnerets were separated into smaller pieces i Vollrath, 1980: 165). and embedded in Swarm-fluid. Morphological investigations of sexual size Subadult males of A. bruennichi were kept in dimorphism are rare; for example Sefcjguchi plexiglass frames from the penultimate stadium i l°55a, b) compared the spinning apparatus in until they died male and female spiders. The present investiga- Early instars were fed with greenflies tion examines whether or not there are mor- (Aphidina) and Drosophita melunogaster while phological data that could be used to assess these older stages were provided with various insects different ideas, by analysing the spinning ap- captured in the field, paratus in different instars of Argiope bruennichi. RESULTS MATERIALS AND METHODS Adult Akgiope Males can Spin Weds Adult males and females of Argiope brarnnuht The posterior spinnerets of adult males 616 MEMOIRS OF THE QUEENSLAND MUSEUM

SEXUAL DIMORPHISM IN ARGIOPE BRUENN1CHI 617

equipped only with rudimentary projections of II III IV V VI VII VIH 6 9

the triad spigots (Fig. 1). These aggregate and 1.3 1.4 1.4 1.6 1.9 1.8 2.0 1.9 1.9 flagelliform spigots-which usually furnish the S.D. 0.09 0.08 0.04 0.09 - - - 0.14 0.56 1 capture threads-obviously do not function. Nevertheless, adult males ofA. bruennichi and A. TABLE 1. Argiope bruennichi. Length/width- lobata spun rudimentary orb-webs (Fig. 2). These quotients and standard deviation of the opisthosoma webs were destroyed every day during our study for different instars (II-VIII) and adults (n = 62). and the spiders renewed them almost daily. Sub- vals between the moults varied (instar III to IV: adult males spun normal orb-webs with the typi- 17-73 days; instar IV to V: 1 1-52 days). Follow- cal stabilimentum until the last moult. Webs spun ing the sixth instar (when subadult males could after the terminal moult were smaller than sub- be first determined) the sexes developed dif- adult webs. They showed an irregular structure; ferently (Fig. 15). This pattern of development of but radia and a spiral were present. This spiral sexual dimorphism is less apparent in the consisted of a few turns with remarkably increas- prosoma. ing space outwards thus showing a closer resemblance to the auxiliary spiral than to the postembryonic development of the capture spiral. No droplets adhered to the spiral Spinning Apparatus threads when the spiders were sprayed with water The development of the spinnerets was docu- using an atomiser. They seemed to be thinner than mented by counting the piriform and aciniform capture threads and were not sticky. Prey thrown spigots. None of these fusules was found in the into the webs did not adhere to the spiral threads. second instar (Fig. 10). Six piriform (anterior A microscopical investigation revealed that their spinnerets), two aciniform (intermediate spin- structure also strongly resembled auxiliary nerets) and three aciniform spigots (posterior threads (Figs 5-7). The auxiliary nature of the spinnerets) were counted for third instar in- web was confirmed by comparing the connection dividuals (Figs 11-13). The development of the points of different types of threads with the radia spigot number was uniform until the sixth instar. (Figs 8-9). From that stadium on, the sexes developed differently as shown for the aciniform spigots on the Intermediate Spinnerets of Adult Males posterior spinnerets (Fig. 16). Although the The number of piriform and aciniform spigots anterior and intermediate spinnerets develop in for each sex were counted under the light micro- the same way (Fig. 16), the differentiation of the scope. There were fewer bases than apical parts sexes is less obvious. of the aciniform spigots in the intermediate spin- Tubuliform spigots were also first observed in nerets of males. SEM examinations showed that the sixth instar of female spiderlings (Fig. 14). a surprisingly high number of apical parts were degenerated (Figs 3-4). Only 19.2% (6 in- DISCUSSION dividuals) of the aciniform spigots on the intermediate spinnerets were fully developed. This The degeneration of the triad spigots on male degeneration was less in subadult males. posterior spinnerets during the terminal moult was first described by Sekiguchi (1955b) and POSTEMBRYONIC DEVELOPMENT OF BODY SlZE subsequently documented for other species. The shape of the opisthosoma changed during These morphological reductions in males were the development of A. bruennichi. The spherical explained by changes in their behaviour: adult form changed into an elongated one, reaching the males cease spinning webs and instead search for proportion of adult spiders at the sixth instar. This females. Emerton (1878) and McCook (1890) pattern is clearly shown by the length/width- mentioned that small webs were spun by Argiope quotient of the opisthosoma in each instar (see aurantia males, but no information about the Table 1). The development of both prosoma and structure of these webs was provided. This study opisthosoma was nearly uniform for all in- shows that A bruennichi and A. lobata males are dividuals until the sixth instar, although the inter- capable of spinning webs. The spiral threads of

FIGS. 1-9. 1. Triad spigots at posterior spinneret of adult male. Apical parts of aggregate (ags) and flagelliform (fs) spigots missing, bases vestigal. 2. Orb-web of adult Argiope bruennichi male. 3. Intermediate spinneret of adult male. Many apical parts of aciniform spigots missing. 4. Detail of 3. 5-9. Different structures of male webs. 5: Capture thread (subadult male); 6: Auxiliary thread (subadult male); 7: Spiral thread (adult male); 8: Junction of a capture thread with a radia; 9: Junction of a spiral thread with a radia. 618 MEMOIRS OF THE QUEENSLAND MUSEUM

a ::

FIGS 10-14. Developmental stages of spinning apparatus. 10: Second instar; anterior (A), intermediate (I) and posterior spinnerets (P) are undeveloped; 11-13: Third instar; 1 1 : Anterior spinneret with six piriform (ps) and two ampullatc (ams) spigots; 12: Intermediate spinneret with two aciniform (acs) and two ampullate (ams) spigots; 13: Posterior spinneret with three aciniform spigots (acs) and the triad (consisting of two aggregate tags) and one flagelliform (fs) spigots; 14: Detail (if posterior spinneret of 9spiderling at sixth instar. Tubuliform spigot (ts) is clearly to differentiate from aciniform spigots (acs). Note arrangement of triad: aggregate (ags) and flagelliform (fs) spigots stay apart from each other, typical for subadults. these webs resemble auxiliary threads, and there- recognized because the threads were not covered fore are not furnished by the triad glands which with glue droplets. However, it is possible that still exist immediately after the last moult (see these threads are not auxiliaries: Vollrath and Sekiguchi, 1955b"). The auxiliary thread type was Edmonds (1989) found that the glue is soluble in SEXUAL DIMORPHISM FN ARGIOPE BRUENN1CHI 61*

spigots of the posterior spinnerets furnish the balloon lines, then their degeneration on the intermediate spinnerets would not disadvantage the

- e — i males. Since adult males do not depend on the additional function of the aciniform threads (in tea \** 8 terms of prey wrapping) r their degeneration may ET be interpreted as a morphological adaptation to an altered style of life. These results suggest that the development of sexual takes in instar. r 1 dimorphism place the sixth Townley et at. (1991) reported tubuliform spigots in Araneus cavaiicus in the fourth instar,

I III II V V v'l krfl / r-. which is equivalent to the sixth instar of Argiope. 1b Townley et al. (199)) suggested that the muiib»r CI inatara 1 tubuliform spigots are present that early in order to 'stake out sites for the functioning spigots of mature females, because the tubuliform glands t'.M 1 are poorly developed and do not serve any func-

i 1 --1 tion at that time . This explanation seems unlikely a because the spinnerets are reorganized and the O W number of fusules increases with each moult. 40 In contrast, we consider the existence of the tubuliform spigots already at the sixth instar in

Id i (3* female spiderlings to be an indication for the

1 hypothesis that phylogenetically earlier females

1 reached maturity at this developmental stage.

9 I T ( 1^ r III IV V V VII J" In females the tubuliform spigots may indicate 16 the penultimate stadium as do the swollen palpi nianbi.tr of Instai* in males. While the males become mature, FIGS 15-16. Poslcmbryonic development of Argiope females undergo another series of moults (3-4) lo bruennichi: 15: Opisthosoma length; 16: Number of reach maturity. During their phylogcnetic his- acini form spigots at posterior spinneret. tory, the females in the subadult stage-more exactly in the penultimate stadium-undergo a water and Peters and Kovoor (1991) argue that prolongation of their development resulting in a the glue does not necessarily fall into droplets. larger body size, directly correlated with higher egg production. Therefore, sexual size dimor- It is not clear why adult males produce these phism have evolved to produce larger rudimentary webs, but their poor design suggests may females, a hypothesis as especially presented hv that it is unlikely that they function to catch prey. Gerhardt f 1924) and Vollrath (1980). That adult males rejected food offered with tweezers can support this assumption. ACKNOWLEDGEMENTS The females in spider genera that exhibit extreme size dimorphism are usually hemisessile. We wish to thank Dr. Peter Sacher (Wiuen- Thus courtship and mating is achieved by male berg-Lutberstadt. Germany) for valuable discus- mobility. Male mobility is achieved by using sions and for the Argiope material. bridging lines and ballooning (Peters, 1990). Mc- Cook (1890} described these balloon lines as CITED consisting of quite a number of threads that LITERATURE remain separated from one another, which sug- BRJSTOWE, W.S. 1929. The mating habits of spiders, gests that they are furnished by the aciniform with special reference lo the problems surround- spigots. The degeneration of these aciniform ing sex dimorphism. Proceedings of the Zoologi- spigots on the intermediate spinnerets of adult cal Society of London 21: 309-358, Argiope bruennichi males occur because the may DARWIN, C 1890. 'Die Abstammung des Menschen remaining spigots are sufficient to produce the und die geschlcchtliche ZuchtwahP. (Bd. 1. E. balloon lines. Alternatively, these threads may Schweitzerbart'sche Verlagsbuchhandlung, Stut- originate from other spigots. If the aciniform tgart). 620 MEMOIRS OF THE QUEENSLAND MUSEUM

EMERTON, J.H. 1878. The structure and habits of organs between male and female spiders. Science spiders'. (Salem). Reports of the Tokyo Kyoiku Daigaku University, ELGAR, M.A. & NASH, D.R. 1988. Sexual can- Section Zoology 8: 23-32. nibalism in the garden spider Araneus 1955b. The spinning organs in sub-adult geometric diadematus. Animal Behaviour 36: 151 1-7. spiders and their changes accompanying the last GERHARDT, U. 1924. Neue Studien zur Sexual- moulting. Science Reports of the Tokyo Kyoiku biologie und zur Bedeutung des sexuellen Daigaku University, Section Zoology 8: 33-40. Gropendimorptrismus der Spinnen. Zeitschrift fur TOWNLEY, M.A., HORNER, N.V., CHERIM, N.A., Morphologic und Okologie derTiere 1: 507-538. TUGMON, C.R., TILLINGHAST, E.K. 1991. MCCOOK, H.C. 1890. 'American spiders and their Selected aspects of spinning apparatus develop- spinning work', vol. 2. (Philadelphia). ment in Araneus cavaticus (Araneae, Araneidae). PETERS, H.M. 1989. the structure and On glandular Journal of Morphology 208: 175-192. origin ofbridging lines used by spiders for moving F. 1980. to distant places. Acta Zoologica Fennica 190: VOLLRATH, Why are some spider males 309-314. small? A discussion including observations on Nephila clavipes. 8. Intemationalcr Arachnologen PETERS, H.M. & KOVOOR, J. 1991 . The silk-producing system of Linyphia triangularis (Araneae, KongreB, Wien, 165-169. Linyphiidac) and some comparisons with VOLLRATH, F. & EDMONDS, D.T. 1989. Modula- Araneidae. Zoomorphology 111: 1-17. tion of the mechanical properties of spider silk by SEKIGUCHI, K. 1955a. Differences in the spinning coating with water. Nature 340: 305-7. A METHOD TO DEVELOP AN 'INDICATOR VALUE' SYSTEM FOR SPIDERS USING CANONICAL CORRESPONDENCE ANALYSIS (CCA) RALPH PLATEN

Platen, R. 1993 11 1 i: A method to develop an 'indicator value' system for spiders using Canonical Correspondence Analysis (CCA) Memoirs of the Queensland Museum 33(2): 621-627 Brisbane, ISSN 0079-8835.

Multivariate Canonical Correspondence Analysis (CCA) is used to denvc 'indicator values' for spiders, similar to those used for plants. The data set consists of activity abundance values of spiders, sampled by pitfall trapping in various habitat types (mires, woods, dry grassland) in die Berlin area. Light exposure, soil moisture and temperature were also measured at the sites. Species scores are plotted as a function of environmental factors within an ordination diagram. The method used to determine the indicator value from this ordination diagram is presented- The system of indicator values is regarded as a suitable method to evaluate sites and areas easily. Advantages and limitations are discussed. Mil Hilfc der multivariaten statisiischen Methodc Kanonische Korrespondcnz Analyse (CCA) wcrden Zcigcrwerte fur Webspinnen, ahnlich denen fur Pflanzen ermiitell. Die Entw'icklungdiesesZeigcrwertsysternsunddessenAnvvenduiigwirdirnPrinzipt»cschrieben. Doi verwendete Daicnsatz besteht aus Aktivitatsabundanzwcrtcn von Spinncn, die mil Bodcnfallen in unterschied lichen Biotoptypen (Mooren. Waldern und Trockenrasen) im Qcblel von Berlin gelangenwurden. Die an den StandoncngcmcsscncnabioiischcnFaklorcn Licht, Tcmperatur und Bodenfeuchte werden mil in die CCA einbezogen. An Hand von ,«iclcn wird der Weg erlaulert, Zeigerwene aus Ordinationsdiagrammen iu ermhieln. Mit Hille ciniger Arten werden Anwendungsbereich und Beschrankungcn des Zeiger- wcrtsyslems aufgezeigt und diskutiert. Das Zeigerwertsystem wird als eine brauchbare Methodc betraehtet, urn Standorte und Untersuchungsgebiete relativ leicht mit Hilfe der Spinncn zu bewcrtcn. [jAraneae, indicator value, multivariate analysis.

Ralph Platen, Institut fiir Bodenzoologie und Okologie, Freie Universitat Berlin, Tietzenweg 85-87, W-lOQOBerlin-45, Germany; 12 January 1993

Ecosystems change under anthropogenic in- been developed, e.g. lor soil organisms by flucnecs faster than their structures and functions Wodarz et a!. (1992), for epigaeic predatory

. .in be analysed. It is therefore difficult to make arthropods (spiders and ground beetles) by well-founded comments about their ability to Haenggi (1987) and Platen (1989. 1992), for withstand external pressure, or about possibilities spiders by Martin (1991) and for ground beetles for their protection or renaturalisation. The com- by Mossakowski and Paje (1985). Some of these plex ecological questions this deficit poses will evaluation systems describe the ecological be require field work involving as many environ- haviour of the species in the field very preo mental factors and groups of organisms as pos- (Martin, 1991), or allow a differentiated evalua- sible. tion of sitse or areas of studv (Mossakowski and Paje, 1985; Haenggi, l987;*P!aten, 1989). Some GOALS evaluation systems, however, have the disadvantage that lengthy calculations are necessary

foT synoptic evaluation for different sites or a A first step is the description of the ecological (Wodarz et al.. 199*; Haenggi, In other behaviour of species in the field. A further step is 1987). cases parameters are used in the caleiil.itu.os to derive evaluations for the sues, biotopes or which arc noistabictoriimeand/orlocality. such areas of study from the ecological behaviour of the species. For example when establishing as a Iow locaI abundance of a species, or the of individuals of a species caught in whether an area should be protected, tor purposes numbers a car (Mossakowski and Paje, 1985). The evalua- 1 .1 planning and biotope-management as well as y t.on systems mentioned can also only be appla:.! when studying the changes at the sites under ,0Cal1 whcre as a result °f intensive field work, anthropogenic influence, an efficient evaluation Y * the ecological behaviour of specie* atoogahiotie sysiem wVich beyond that is easy to handle will gradients is known. l.'c necessary. A number of evaluation systems have recently A much simpler method would be the applica- 1 1

622 MEMOIRS OF THE QUEENSLAND MUSEUM

liOfl of an indicator value system similar to thai RESULTS for plants of Ellenberg et al. 99 1 ). It would then no longer be necessary to redetermine and re- Ordination Diagrams evaluate the ecological behaviour of a species for The CCA results with abiotic factors light ex- each local investigation, since this would already posure and temperature, ns wrll n£ Soil water be contained in the key values Nor would the content, are shown graphically (Figs 1. 2). The evaluation involve complicated calculations. horizontal axis corresponds to the first CCA axis My aim has been to develop just audi an in- and the vertical to the second CCA axis. The I I I dicator value system for spiders species of spider are represented by an *X\ togcthcr with an abbreviation of the name as Far MATERIAL AND METHODS as possible. Using CANOPLOT it was also possible In determine the coordinates and the name of a species which could not be presented unam- Data biguously in the diagram. Initially the axes of the site factors light ex- The data consisted of the activity abundances posure and soil water content are extended of spider species. These were determined using beyond the origin (Fig. I). The factor along the ground traps in the Berlin area for open and 'environmental axes' increases in the direction of wooded sites in moors, in various types of forest the arrow. The origin marks the mean value for and for heathland and semi-dry and dry meadows. the entire data set Species whose position lies The investigation period was a full year in each the arrowhead an environmenta I a X between of 1 case. Activity abundance is defined by and the origin have a larger weighted mean. Heydemann ( 1 953) as the number of individuals. Where the origin is between the arrowhead and which has been trespassed a borderline (which is the position of the species, its weighted mean iv represented by the diameter of the pitfall trap) smaller than the overall mean. For interpretation within a certain period of time. Parallel to the trap a perpendicular is projected for each species in catches the following abiotic factors were also turn onto the environmental axis according to measured: their sequence (cf. Jongman ei at, 1987). X\s- The soil water content (measured as the per- ticus nitwit and Thanatm arena rius occupy the centage by volume of water in the upper soil brightest sites, and Pardasa agrestis and layer), the light exposure using the method Baryphyma prarense the warmest sites [Rg, I i described by Fnend and the effective (1961), Species at extremes of the axes represent limits temperature after Pallmannn et td (1940). The of the area for a two-factor system, and thus form

sites are described in detail in Platen * 1 9S9». the start and end points of the indicator value scale. The distance between these points is measured and divided into five equal parts. The General species are then noted for each class with the The activity abundance of the spider species appropriate indicator value. and the measurements of the abiotic factors are The determination of indicator values for three analysed using Canonical Correspondence factors requires at least two ordination diagrams. Analysis (CCA; Jongman et at.. 19S7), using the Initially an indicatorvalue is assigned for all three program CANOCO Version 3.10 (Braak, 1988. individual factors, then for all combinations of 1991). The results of this analysis arc displayed two factors. The result al ways remained the same. as ordination diagrams using CANODRAW For the representation Of all three factors in an (Smilauer, 1 990 ordination diagram the class of a species cmi. Before running CCA the spider data had been however, change, in some cases considerably, masked according to dominance in a formal way: since the relative spatial distances of the species species which did not have an activity abundance in Ihrce-factor constellations cannot be repre- of at least ] % a: a site were removed from the data sented in a two-dimensional coordinate system set. This meant that of the original 281 S| wiihoui distortion. For the determination ol in- species only 111 remained for the further dicator values from the data of individual en- analysis. vironmental variables the solution (Figs 1 . 2 ) is Furthermore a transformation of the WW data an optimum. was carried out. Instead of the abundance values Moisture is strongly negative correlated with their square roots were used. the 1st CCA-Axis (Inter-set correlation: -0.919) INDICATOR VALUES FROM CANONICAL CORRESPONDENCE ANALYSIS 623

Temperature

1st CCA-Axis

FIG. 1. CCA ordination diagram wilh 1 1 1 spider species represented by an V and environmental variables moisture and temperature represented by arrows. The part of the arrows to derive indicator values are devided inlo five parts ("M1-M5 andTI-T5 respectively). For further explanation sec text.

(Fig. I) which means that the horizontal species cannot claim to be comprehensive or generally distribution is best explained by light and less by valid. Some examples will show the similarities temperature (Inter-set correlation with 1st CCA- and differences between the indicator values and Axis: 0.26, with 2nd CCA-Axis: 0.538). Hence, other methods of determining ecological be- the vertical species distribution is best explained haviour. along the 2nd CCA-Axis. The distribution of Xysticus ninnii is centered The Inter-set correlation between light and 1st exclusively on dry meadows. Fi, L5» and T4 CCA-Axis is 0.955 which means that the data set reflect this ecological behaviour well. again is best explained by this abiotic factor. Diplocephalus permixius: Occurring mostly in Temperature is strongly correlated with 2nd wei alder forest-habitats characterised by high CCA-Axis (Inter-set correlation with 1st CCA- soil water content, low light exposure and low Axis: 0.0367, with 2nd CCA-Axis; 0.6177). temperatures. This is expressed with adequate precision by the indicator values F5, LI and T2. Indicator Values (Table I) Diplocephalus picinus: F2, LI and Tl charac- The last two columns contain details of the terise its habitat preferences, namely shadowy ecological type and habitat type in which the sites with low pH in dry mixed forests. species predominantly occurs in the Berlin area Pardosa agrestis: In this case the indicator

(after Platen et fli, 1991 ). The data are intended value does not reflect the ecological behaviour, only to demonstrate the principle of this method. as a result of the inadequacy of the data set. The In view of the limited data set the indicator values species occurs mostly on arable farmland and 624 MEMOIRS OF THE QUEENSLAND MUSEUM

2nd CCA-Axis Temperature

Xy « -. i nn

Light IstCCA- Axis

FIG. 2. CCA ordination diagram with 1 J 1 spider species represented by an V and environmental variables light and temperature represented by arrows. The part of the arrows to derive indicator values arc divided into Five parts (L1-L5 and TUTS respectively). For further explanation see text. other open, rather dry habitat types (Platen et al., evaluation systems (Mossakowski and Paje, 1991). While T5 describes the behaviour in terms 1985;Haenggi, 1987; Platen, 1989). of temperature, the indicated soil water content Some reservations concerning the applications (F5) is too high and the light exposure too low of indicator values are necessary. As already em- (LI). This is because in the data set it only oc- phasised by Ellenberg et al (1991), indicator curred on one site (a former moor which was still values describe the ecological behaviour of wet in comparison to the sites on mineral soils), species in a multiple system of biotic and abiotic so that as an outsider it had an extreme position factors, from which those chosen in this paper are in the ordination diagram. The data set did not regarded as the key abiotic factors for the spiders. include arable farmland sites. They do not, however, describe their physiological optima. Indicator values for animals underlie more Comments restrictions as those for plants because animals This method is a relatively easy one to deter- are mobile and often change their habitat for mine indicator values for spiders and other soil overwintering (Schaefer, 1976). Therefore in the arthropods (cf. Platen, 1992). strict sense they are valid only for Adults which Indicator values have the advantage that with does not change their habitat within their period only a few values the ecological behaviour of a of maturity. As juvenile animals were not in- species can be characterised. Mean indicator cluded in the data set this holds true for this values can be calculated, and in the absence of the investigation. measurements or vegetation surveys they allow a Indicator values should never be used in the rough quantitative assessment of environmental same way as measurements. They are ordinal, variables for a investigation site, habitat type or and are not suited for use in statistical (including an area of study. The use of indicator values multivariate) methods requiring higher scale obviates the need for complicated calculations for data. the evaluation of sites, as required by some The indicator values determined above arc al- INE>ICATOR VALUES FROM CANONICAL CORRESPONDENCE ANALYSIS 625

ways valid only for the data sel used. Since they program for canonical community ordination b) [partial |dcucndcdj [canonical correspondence were limited to only a few lypes of biotope the | pnncip.il results above can only be regarded as being a first analysis, components analysis and redundancy (Version 2.1)1. Technical approximation. The results of the analysis are anafyslg report LWA SK-02. Grocp Landbouwwtskunds, ly dependent on the type and number of Wageningcn habitats types and of the frequency with which 1990, Update notes: CANOCO Version 3.10. various species occur there. The combination nf Manuscript. Agricultural Mathematics Croup, a wet. light site (F5, L5) is not represented by an Wageningcn.

value, it is indicator although relevant for a num- hLLhNBERG. H-. WEBER. RE.. DULL, R., WIRTH, ber of species (Drepanotylus uncatus* AnttsUa V., WERNER. W. & PA LESSEN, D. clegans) (Table 1 ). However, since almost 2/3 of Zcigerweru wan PI'lan/cn inMiUcleutopa. Scrii*a the 30 sites investigated were we*, and most GeoboUmcit XVIFl: 1-248. species occurred with almost the same frequency prifnd, D.T.C 1961, A simple method oi measuri integrated light values in the held. Ecology 12 in wet habitats, these species grouped elo.se U i the origin Species which occur frequently, but only 577-580. at one or two dry sites with very high Hgbl HANGGI, A. !987. Die Spinnenfauna d*r FcuclllgehietC lies Grolkn Mooscs, Kt. Bern. 2 posure are far ftoin ihc origin, so thai there is a Die Bcurteilung des Nalurschut/.wcrtes natur higher differentiation of the axis over the bright nahcr Standoric anhand der Spinnenl'ann.v Mii- range. tcilungcn der Nauirforsehcnden Gcscllschaft in value s> A generally valid indi'catQi stem would BeroN.F 44: 157-185. to analyse all known spider of Ger- need specks HEYDEMANN. B. 1953. Agrarokologischc many or Central Europe for all existing types of Frobleroatik, dnrgetan an Untcrsuchungcn tlbcr biotope {abiotic factor combinations) in one data die Tierweh der Bodenoberfluche der Kultnr- set. from which the indicator values could the felder. (Thesis Kiel). 4^ P p derived. The scale could then be expanded, or JONGMAN, R.H.G., BRAAK, C.J.F TER & TONCr other environmental factors, such as the biotope EREN.O.RRfcds,), 1987 'DaMmilysiMn com- structure, could be included. A problem would be munity and landscape ecology.' (Pudoc Books: the large number of measurements required Wageningcn). MARTIN, D. 1991. Zur Autokologie der Spinn. r A further problem is that the CCA only depicts (Ar3chnida: Araneae) 1 Cbarakteristik der Habit species correctly in the ordination digram jj atausstattung und Prafcrcnzverhalten epigaischcr have an unimodal response curve along a factor Spinnenarten. Arachnologia Mitteilungen 1 ;5-26. gradient (Jongman et ai, t 1987), By plotting the MOSSAKOWSKI, D. & PAJE. F. 1985. Ein Bewei- frequencies of species at all habitat types sorted tungsverfahren von Raumeinhctlcn an Hand der the ii is according to lc. els of a factor, possible 10 Carabidenbestande. Verhandluneen, Gesellschaft determine bi-modul, multi-modal or continuous Okologic in Bremen 1983 13: 747-750. all habitats responses. The coordinates of (he PALLMANN, H., EICHENBERGER, E. & HASEL- where the Species occurs with high frequency can ER. A. 1940. Einc Met node der Tempcraturmev- be entered in the ordination diagram, making it sung bei okologischen oder bodenkundlichcn possible to recognise a corresponding range of Untersuchungen. Berichte. Schweizerischc occurrence on the environmental axis. For this Botanischc Gcscllschaft 50: 337-362. species, as is the case with some plant species, an PLATEN, R. 1989. Struktur der Spinnen- und indifferent response to this factor for the species LaufkaTcrfannn (Araeh .; Araneida, Col.. Carabidae) bccmtlutitcr Moorstan- or to give a range of the indicator value (cf. amhropoden dorle in Berlin (West); laxonomischc, raumtichc Ellcnbergef a/.. 199]) may be possible. und zcillicbc Aspckte. (Thesis. Techmsciie ACKNOWLEDGEMENTS Univcrsitat Berlin) 470 pp. PLATEN, R lf MOR1TZ, M. & BROEN, B.v 1991

I .isle der Webspinnen- und WcberkncchiarU-n

For ideas, criticism and discussion 1 would like (Arach ; Aroneid.i, Opilionida) des Berliner to thank Florian Bemmerlein-Lwx, Nuremberg, Raumc* und i hi c Auswcrtung fiir Ktturech«lzzwecke(RcrteLi8ie) 243-275 Eh Prof. Lasco Mucins, Vienna, and Prof. Gcrd Pp A. Auliagen, R. Plato and H. Sukopp (eds). 'Rote Weigrnann, Berlin Listen ||er gefahrdeten Pfian/en und Tic BcHin\(LiindsLlutisentvv.; Umwcltforsch. S6). I 1TFRATURECTTED PLATEN, R. !992. Die Lutwieklung eines Zeigw- wcrtsystcms fur Lnufkiifer (Col.; Carabidae) ml

- BRAAK, C.J.F. TGR 1988. CANOCO a FORTRAN Hilfccmer "Canonical Correspondence: Analysts" 626 MEMOIRS OF THE QUEENSLAND MUSEUM

(CCA). Verhandlungen der Gesellschaft fiir SMILAUER, P. 1990. CANODRAW. A companion Okologie 21: 321-326. program to CANOCO for publication-quality SCHAEFER, M. 1976. Experimentelle Untersuchun- graphical output. (Ithaca). gen zum Jahreszyklus und zur Uberwinterung von WODAR2, E., AESCHT, E. & FOISSNER, W. 1992. Spinnen (Araneida). Zoologische Jahrbiicher, Der Bodenbiologische Index (BI) - ein quantita- Systematik (Okologie), Geographic und Biologie tives MaB fiir die Bqdenqualitat. Verhandlungen 103: 127-289. der Gesellschaft fiir Okologie 21: (in press).

TABLE 1. Indicator values for soil water content (F), light (L), temperature (T) of 1 1 1 common spider species of varying types of habitat. ET=Ecological type: h = hygrophilic, (h) = weakly hygrophilic, x=xerophilic, (x)=weakly xerophilic, eu= euryoecious open-space dwellers, w=forest type, (w) = also in open spaces, hw=sparse forest species, (h)w=inhabits mesophilic deciduous forests, (x)w=inhabits forest on acid soil, h(w)=depending on type of preferred habitat: inhabits unwooded wet habitats or sparse forest. - = no preferred habitats. Family (C): Ag, Agelenidae; Dy, Dysderidae; Gn, Gnaphosidae: Ha, Hahniidae; Li, Linyphiidae; Lc, Liocranidae; Ly, Lycosidae; Ph, Philodromidae; Pi, Pisauridae; Sa, Salticidae; Te, Tetragnathidae; Tr, Theridiidae; Tm, Thomisidae; Zo, Zoridae.

SPECIES | F | L | T | ET \c SPECIES F L T ET C

Wet Forests W. cucullata (C.L. Koch) 2 2 (x)w Li

Pachygnatha listen Sundevatl 5 3 hw Te W. dysderoides (Wider) 4 3 (x)w Li

Bathyphantes approximates (O.P.C.) 4 3 h(w) Li IV. monoceros (Wider) 2 2 2 (x)w Li

B. nigrinus (Westring) 5 3 hw Li Euryopisflavomaculata (C.L. Koch) 4 4 (x)(w) Tr

Diplocephalus permixtus (O.P.C.) 5 2 h(w) Li Trochosa terricola Thorell 3 2 2 (x)(w) Ly

Diptostyla concoior (Wider) 4 .'; (h)(w) Li Xerolycosa nemoraiis (Westring) 2 2 (x)(w) Ly

Gonatium rubellum (Blackwall) 3 2 hw Li Cicurina cicur (Fabricius) 4 3 (x)(w) Ag

Porrhomma pygmaeum (Blackwall) 4 3 h(w) Li Agroeca brunnea (Blackwall) 4 3 (w) Lc

Walckenaeria atrotibialis (O.P.C.) 4 2 (w) Li Haplodrassus soerenseni (Strand) 1 1 (x)w Gn

Pirata hygrophilus (Thorell) 5 3 h(w) Ly Zelotes subterraneus (C.L. Koch) 3 3 (x)(w) Gn

Deciduous forests Ozyptila praticola (C.L. Koch) 2 2 (x-)w Tm

Centromerus sylvaticus (Blackwall) 4 2 (h)w Li Waterside sites

Ceratinella brevis (Wider) 2 Li (h)w 1 1 | 3 Gnathonarium dentation (Wider* 5 j i 4 | h Li Diplocephalus latifrons (O.P.C.) 3 2 (h)w Li Moors

Gongylidium rufipes (Sundevall) 3 2 (h)(w) Li Agyneta cauta (O.P.C.) 5 3 h(w) Li

Lepthyphantes pallidus (O.P.C.) 2 2 (h)(w) Li Diplocephalus dentatus Tullgren 5 3 h(w) Li

L. tenebricola (Wider) 4 2 (h)w Li Drepanotylus uncatus (O.P.C.) 5 3 h Li

/... zimmermanni Bertkau 2 1 (h)w Li Erigonella ignobilis (O.P.C.) 5 3 h Li

Microneta viaria (Blackwall) 3 2 (h)w Li Lepthyphantes mengei Kulczynski 5 3 h(w) Li

Neriene clathrata (Sundevall) 4 3 (h)w Li Lophomma punctatum (Blackwall) 5 3 h Li

Pardosa lugubris (Walckenaer) 4 3 (h)w Ly Notioscopus sarcinatus (O.P.C.) 5 3 h Li

Dry mixed forests Oedothorax gibbosus (Blackwall) 5 4 h Li

Harpactea rubicunda (C.L. Koch) 1 1 (X)W Dy Silometopus elegans (O.P.C.) 5 5 h Li

Abacoproeces saftuum (L. Koch) 2 1 (x)w Li Tallusia experta (O.P.C.) _> 4 (h) Li

Centromerita concinna (Thorell) 2 2 (x)(w) Li Walckenaeria alticeps Blackwall 5 4 h(w) Li

Centromerus pabuiator (O.P.C.) 3 2 (x)(w) Li W. kochi (O.PC) 5 4 h Li

Diplocephalus picinus (Blackwall) 2 1 (x)w Li W. nudipalpis (Westring) 5 4 h Li

Gonatium rubens (Blackwall) 2 1 (x)w Li W.vigilax (Blackwall) 5 3 h Li

Lepthyphantes angulipalpis Arctosa leopardus (Sundevall) 5 3 h Ly 2 1 (x)w Li (Westring Hygrolycosa rubrofasciata (Ohlert) 5 3 h Ly

LepthyphantesJlavipes (Blackwall) 2 1 (x)w Li Pardosa pullata (Clcrck) 5 3 h Ly Macrargus rtifus (Wider) 4 2 (x)w Li !' latitans (Blackwall) 5 3 h Ly Minyriolus pusillus (Wider) 2 1 (x)w Li P. piraticus (Clerck) 5 3 h Ly

Panamomops mengei Simon 1 1 (x)w Li P. piscatorius (Clerck) 5 3 h Ly Tapinocyba insecta (L. Koch) 2 2 (x)w Li P. tenuitarsis Simon 5 4 h Ly Walckenaeria acuminata Blackwall 4 3 (x)w Li Trochosa spinipalpis (F.O.P.C 5 3 h(w) L ) > INDICATOR VALUES FROM CANONICAL CORRESPONDENCE ANALYSIS 627

TABLE 1. continued

SPECIES F L T ET C SPECIES F L T ET C

Dolomedesfimbriatus (Clerck) 5 I 3 h Pi Tegenaria agrestis (Walckenaer) 4 1 4 (x) Ag

Antistea elegans (Blackwall) 5 1 4 h Ha Heathland

Reeds Tricca lutetiana (Simon) 1 3 2 (x) Ly

Baryphyma pratense (Blackwall) |5 |l 5 [h JLi Zelotes latreillei (Simon) 2 1 1 (x) Gn J Wet Meadows Dry grassland

Allomengea scopigera (Grube) 3 1 2 h Li Pachygnatha degeeri Sundevall 2 2 3 eu Te

A. vidua (L. Koch) 4 1 2 h Li Meioneta beata (O.P.C.) 1 4 3 X Li

Ceratinella brevipes (Westring) 4 ! 3 h Li Trichopterna cito (O.P.C.) 1 5 3 X Li

Erigonella hiemalis (Blackwall) 4 1 3 (h)(w) Li Troxochrus scabriculus (Westring) 2 1 1 X Li

Oedotboraxfuscus (Blackwall) 4 1 2 eu Li Typhochrestus digitatus (O.P.C.) 1 5 3 X Li

0. retusus (Westring) 5 1 4 eu Li Steatoda phalerata (Panzer) 1 5 3 X Tr

Pelecopsis mengei (Simon? 3 1 1 h Li Alopecosa cuneata (Clerck) 2 3 3 X Ly

Pardosa palustris (Linn6) 1 4 3 eu Ly A. pulverulenta (Clerck) 4 1 3 eu Ly

P. prativaga (L. Koch) 4 1 3 eu Ly Xerolycosa miniata (C.L. Koch) 1 4 4 X Ly

Cough-grass sites Hahnia nava (Blackwail) 1 4 3 X Ha

Centromerita bicolor (Blackwall) 2 |3 |3 |(x)(w) Li Agroeca proximo (O.P.C.) 4 1 3 (x) Lc

Ruderal sites Zelotes electus (C.L. Koch) 1 5 3 X Gn

Baihyphantes parvulus (Westring) 5 I 3 eu Li Z longipes (L. Koch) 2 4 5 X Gn

Pocadicnemis pumila (Blackwall) 4 1 3 eu Li Thanatus arenarius L. Koch 1 5 4 X Ph

Stemonyphantes lineatus (Linn6) 4 2 3 (x) Li Xysticus kochi Thorell 1 4 3 X Tm

Trochosa ruricola (Dc Gcer) 3 2 3 eu Ly X. ninnii Thorell 1 5 4 X Tm

Arable fields Aelurillus v-insignitus (Clerck) 1 5 4 X Sa

Bathypbantes gracilis (Blackwall) 4 1 3 eu Li Phlegrafasciata (Hahn) 1 3 2 X Sa

Erigone atra (Blackwall) 3 2 3 eu Li No obvious habitat preferences

E. dentipalpis (Wider) 4 2 3 eu Li Cnepbalocotes obscurus (Blackwall) 3 1 3 eu Li

Pardosa agrestis (Westring) 5 1 5 (x) Ly Zora spinimana (Sundevall) 4 1 3 eu Zo

THE SPIDERS OF THE HIGH-ALTITUDE MEADOWS OF MONT NIMBA (WEST AFRICA): A PRELIMINARY REPORT

C. ROLLARD

Rollard, C. 1993 1111: The spiders of the high-altitude meadows of Mont Nimba (West Africa): a preliminary report. Memoirs ofthe Queensland Museum 33(2): 629-634. Brisbane. ISSN 0079-8835.

Spiders are abundant in the high-altitude meadows of the Nimba mountains, in Guinea. Collections have been carried out in this ecosystem where grass ground cover is dominant:

this preliminary study at the family level concerns specimens collected in March 1991. It already gives some data on the localisation of the spiders. More than 20 families are represented along the mountain tops. A provisional list of these spiders has been drawn up. Most specimens were Araneidae, Gnaphosidae, Hersiliidae, or Salticidac. Their distribution in the herbaceous stratum as well as along an altitude gradient between 800-1700m is being analysed. Les araignees sont abondantes dans les prairies de haute altitude des Monts Nimba, en Guinee. Des recoltes ont 6te effecluees dans cet ecosysteme ou la couverture herbacee est dominante. Cette etude preliminaire au niveau des families concerne les specimens collected en mars 1991. Elle apporte deja quelques elements sur la localisation des araignees. Plus de 20 families sont representees sur ces sommets et unc liste provisoire en a ete etablie. La plupart de ces araignees sont des Araneidae, des Gnaphosidae, des Hersiliidae ou des Salticidae. Leur distribution est analysee dans la strate herbacee ansi que selon un gradient altitudinal entre 800 et 1700m. C\Spiders, biogeography, Africa, montane.

Christine Rollard, Museum National d'Histoire Naturelle, Laboratoire de Zoologie (Arthropodes), 61 rue de Buffon, 75005 Paris, France; J 4 January, J 993.

The Mont Nimba biosphere reserve, located in ravines (Lamotte, 1958). This environment is West Africa, is the subject of a multidisciplinary characterized by the strong contrast between the study as part of a UNESCO pilot project. These dry and rainy seasons. It presents a characteristic mountains have been classified as an Integral fauna with several endemic species. Natural Reserve since 1944. Over the past 50 In this paper, preliminary data dealing with the years, many more animals have been collected localisation of the different families of spiders during several scientific expeditions directed by collected in March 1991 are presented. The over- Professor Maxime Lamotte (Lamotte, 1943). all study will lead to a more detailed inventory of Many papers have been published on the Nimba the spiders along the mountain ridges, as well as mountain range. However works on the spider a better knowledge of their distribution with al- fauna are non- existent. Hence the organization titude and the relative abundance of the different of the spider populations are still little known. species. Especially in this tropical region, no work deals with the ecology of spiders except for the research ENVIRONMENT AND CLIMATE initiatives in the Ivory Coast, in the savanna of Lamto (Lamotte, 1943, 1967; Blandin, 1974; Blandin and Celerier, 1981). Mont Nimba is situated in High Guinea, near the borders of Liberia and the Ivory Coast (Fig. With this subject of research in mind, another 1 ). It extends from SW to NE for about 30 km into field trip was made to Guinea in March 1991. The guinean territory. All the crests stand over new collection of spiders made there compli- 1000m. The mountainside is steep and notched ments those of Mr Lamotte and his associates. by valleys with sheer slopes. No trees or shrubs This abundant material is in the process of being are present on the crests. Only some small trees classified. An attempt is being made to describe of the inferior savanna grow at lower altitude on the structure and function of the spider com- the slopes (Schnell, 1966). Above 900- 1000m, munity in this tropical ecosystem. the forest is confined particularly to the ravines. The programme focuses on the spiders of the The mountain range is covered by herbaceous high-altitude meadow, relatively less-frequently plants with a grassland structure. The term of collected than those of the savanna or the head of montane or sub-montane has been given to this .

630 MEMOIRS OF THE QUEENSLAND MUSEUM

RELATIVE FREQUENCY

t 20 40 60

Agelenldae Areneidoe Clublomdae Cohnmdee Clem doe Heterapodidee Gnaphosidee Hersi Indue Linyphndoe Liocronidee Lycosidee Oxyopldee Plseunaee Salticidae Scytodidee Selenopidee Sicemdae Tetragnelhidae FIG. 1. Map of Mont Nimba, West Africa. Situation of the various sample zones along the MontNimba crest. Theiidudoe Thomisidae crest vegetation (Schnell, 1987). In this type of Zcdarndaa n - 394) meadow, located in the guinean pan of Nimba, the iow (about 20-30crn or less) graminaceous FIG. 2. Total no. spiders taken during March 199! species Loudetia kagerensis is abundant and constant. This grass forms a predominant group as- METHODS sociated with other tall or short species, varying from site to site. until now, the high-altitude meadows of Mont Nimba receives abundant rainfall and a Up Mont Nimba have rarely been sampled. The dry season not exceed! ng three and a half months. present study lasted 20 days (8-27 March) with Generally, high-altitude meadows are often only 14 days of sampling. The first phase con- covered in fog during the rainy season, from May sisted in obtaining an overall idea of the spider to November. Precipitation is fine and stable. communities over the whole crest. The collecting During the other periods of the year, clouds scale program was prepared as follows: 4 days in the slopes and progressively cover the crest. Thus Sempere, Grands Rochers; 4 days on the Nion the humidity, which is closely bound to the de- crest, 2 days in Grands Rochers. R. Mallard; 3 gree of precipitation and nebulosity, varies with days on Mont Leclere and 1 day in Ziela, P. altitude and also along the crest. For example, the Richaud. crest of Nion, spreading upwards to Mont R. No strictly quantitative sampling methods were

Molard in the NE, is wetter than either Mont P. used. Several gathering squares (lm x lm) were results insufficient. In Richaud or the region of Sempere (Leclere et al., made but the are addition, these quadrats have not well materialized 1955). Nevertheless, seasonal variations do exist. been because the transport of materials was not easy- In the meadow, this factor does not seem to be Only the sizes were marked by various elements very important to the ecological cycle of the found on the ground. Furthermore, this type of fauna. analysis is rare in these environments. Some in- The spiders listed here were collected in March formation on the vertical distribution of the 1991, during the transition period just before the spiders was obtained by the use of beating and rainy season. Sudden storms or regular strong ground sweeping methods. A few specimens precipitations occur in the laic afternoon from found on the high sections of the grass were April onwards. captured, but this is certainly not a representative SPIDERS OF MONT N1MBA 631

Araneidae Salticidae

Hersilidae Setenopidae

Hetergpodtflae Teiragnarhidae . upper part of Oxyopidae Thomisidae

' ~7~- ~ . stems and ?w 5S leaves Agelemdae Oxyopidae

Connmdae Pisaundae

Clubiomdae Saiiicjdae

Gtefi flae Sicawdas Liocranidae stem base- Lmyphiidae Thomisidae Lycostdae

Scylodr!,-f I Saltiadae

Theridiidae

Zodanidae

superficial soil

FIG. 3. Distribution of families in cpigean environ- this crest, of the roughly 100 families known

ment. worldwide (Table 1 ). This relatively high number provides a good idea of the diversity of this en- sample of the spicier fauna of the herbaceous vironment. The high-altitude meadow-like stratum. plateau savanna characterized by the grass So, in the low herbaceous stratum, samples species Loudetia, is usually considered to be one were made by visual searching in vegetation and of the poorest habitats. However, this type of under stones, using forceps and poolers. Collect- environment clearly possesses an important ing was only carried out during the daytime and variety of spiders. was therefore not exhaustive. The time spent on Furthermore, the relative frequencies of these Mont Nimba was rather short: the researchers spiders gives another indication of their diversity stopped collecting at the end of an hour at each (Fig. 2). The physionomy of the araneological site. community is characterized by the predominance A mean of 40 individuals was collected per day, of Araneidac which represent around 21% of by one to three collectors depending on the days determined spider families. Four other families and the availability of the expedition members. were common: Gnaphosidae (57/394), Her- RESULTS siliidae (59/394), Salticidae (46/394) and Thomisidae (35/394). Wc note that these spiders are generally large and therefore easier to find. Families Identified The same observation can be applied to other For each sample, the spiders were sorted and families which are more easily collected, espe- enumerated by family. Young and adult spiders cially for a certain size. However, the fewer were counted together. Immature spiders repre- Clubionidae, Lycosidae, Oxyopidae, Pisauridae, sent about 54% of all specimens found. Selenopidae and Tetragnathidae in the collec- Mygalomorphs were collected, but not in great tions could indicate that they are less abundant in numbers: 0.5% (2/435). For the moment, the in- the meadow during this period of the year. The ventory of this suborder has not been established specimens belonging to most other families arc at the family level. Similarly, about 9% (39/435) generally small and consequently poorly col- of the collection is still at the level of indeter- lected. In addition, Liocranidae and Zodariidae minate araneomorph. Therefore, we are only able are active and move quickly on the soil, so they to present results of the determinate araneomorph frequently evaded capture. families, which correspond to 394 specimens. The names used follow Brignoli (1983) and Plat- nick (1989), Vertical Structure of the Spider Community At least 21 families have been recorded along The quality of a sampling must take into con- 632 MEMOIRS OP THE QUEENSLAND MUSEUM

II Agelenidae Hersiliidae Scytodidae

9 DO tO 00 01 IB CO 1 BL Aiancidae Liny phi idae Selenopidae Clubionidae Liocranidae Sicariiloae Oxyepldae i 9 Corinnidae Lyooffldac Tetragnathidae i Hdse 1 Ctenidac Oxyopidae Thcridiidae .i lypfl] a |£ 1 Heteropodidae Pisauridae Thomisidae irudqg 2 Gnaphosidae Salticidae Zodariidae Sclenopiaae i

Cl-i-Mi-Jg* 1

1 TABLE 1. Lisl of spider families on Mont Nimba : rudae Aaele^". 4 sideration the biology and the size of the spiders. fiidse 5

; "i-. 26 All families are represented, in spite of the small e Gnaphosidae 5 1 7 3 5 number of specimens. Around two-thirds of these G itvoniclse a 5 t D families frequent the herbaceous stratum and idse most are diurnal (Fig. 3 ).

i tfae 4 i 5 c stems leaves of On the upper part of the and iiotae i -1 9

; plants, the spider is composed of B : : 2 t 1 o community Hwsill " i

eight families of which the Araneidae, Tetrag- rhdmlsu :j 3 t 15 6

ia tiidae, lion-iliiij.ic and rhomisulag are the Lfcflsft : 2 4 K : 3

' t most common. The former two build their webs . pdjaaa

,.., 2 ' about 20 cm above the soil. Of all Araneidae, I Pisa " Oxyopidae and Salticidae, 25, 2 and 1 specimens respectively were collected by beating and : ..:- i sweeping of the ground. Ten families are present at the base of the stems or in the center of the

. clumps, Clubionidae being the most abundant. The elubionids are nocturnal hunters and easily TABLE 2. Altitudinal distribution of spider families in found in nests among the vegetation. Five meadow of Mont Nimba. families occur in the superficial soil layer. The provide many specimens, with the exception of Lycosidac constitute most of the collections. Tetragnathidac. These spiders are perhaps rare or Gnaphosidae are mostly nocturnal hunters found difficult to observe, but it is all the more interest- in nests among stones. Scytodidae are aJso found ing to note that their distribution is limited to a uocturnally active around stones. certain altitude. In the same way, Eusparassidae Thus, each level of this epigean environment and Pisauridac are found only up to 1200m, in seems to possess its own spider community, char- low vegetation. The lack of data, between 1000 acterized by its family composition. Neverthe- and i200m altitude, for the Araneidae and Her- less, some of them such as the Salticidae, siliidae is probably due to sampling problems. Thomisidae and Oxyopidae are present at all Spiders of these families as well as the Salticidae, levels of the vegetation and the soil surface. Only Thomisidae and Lycosidac. arc certainly present analysis at the species level will permit the at the different altitudes, and it is likely that the clarification of die distribution of the spiders in same is the case for the Zodariidae. The family each stratum. diversity seems to increase slightly with altitude. The collections made on slopes from 800m Ai.titupinai. Distribution upwards concentrated particularly on Mont The spiders were collected along the crest, Lcclcrc (Table 3). Here too, we note the diversity mainly situated above an altitude of 1200m. The of the spider fauna, with 14 families present of

results are presented by altitudinal classes of the 2 1 listed for Mont Nimba. The same families 200m. principally because of the small numbers are found at the highest altitudes. Only of spiders, and the various sampling zones are Zodariidae and Clubionidae are not encountered indicated in Fig. 1. below 1400m. Salticidae, Hersiliidae, An overall view indicates that some families Thomisidae and Lycosidac can be found from appear to be better represented at the highest 800 to 1600m, altitudes, from 1200m to around 1700m (Table We also compared the spider families found at 2). The numbers ofGnaphosidae and Clubionidae three points along the crest: P. Richaud, Grands regularly increase. Most other families did not Rochcrs and Nion crest, between 1200 to 1600- SPIDERS OF MONT NJMBA 633

MT TUDE ir. ,, families present only on P. Richaud and Grands Rochers. However, more precise information, at 1000 I 600 the species level will, hopefully, be available in Sicamdae 1 the future. Ctenidae 1

Liocramdae 1 COMMENTS Olonldae 4

Zodanidae 7 Spiders occupy an important place among the

Gnaphos'dae 2 i a invertebrate fauna of Mont Nimba. They are rep-

Araneidae A 3 1 resented by about twenty families which is a relatively large number for this type of highland Sahicdae A 1 2 3 meadow. This study is only a first approach; the He'S'h.dae 6 1 2 diversity of the spider community according to Thomisrdae 3 3 1 1 stratum and altitudinal level will certainly prove Lycosidae 2 4 1 3 to be rewarding, both quantitatively and qualita- Heieropodidae 2 1 tively, with the determination of species. Publi- 3 1 Pisauriaae cation of the final results will probably be delayed Scylodidae 1 because of taxonomic difficulties. Nevertheless, this study already gives some TOTW. NUMBER OF SPECIMENS results at the family level. Prudence in the inter-

results is necessary . because the NUMBER OF pretation of these FAMILIES families do not always form homogenous ecological units. The sampling methods used TABLE 3. Altitudinal distribution of spider families on must also be taken into consideration. In the MontLecIerc. tables and figures, we sec that the distributions 1700m altitude (Table 4). Seven families are found are dependent on the collecting effort. In present on the Nion crest, clearly a lower diver- this study we used mostly visual-hunting, with no sity than on P. Richaud with eleven families. Six collecting at night. So we only have a partial families were observed at all three places. Only sample of the spider families, mainly repre- Tetragnathidae were found south of the crest. senting those with diurnal activities. As yet, pit- This place seems to be more humid than the fall traps have not been used to intercept others at different periods of the year. The spiders nocturnal spiders. Nevertheless, an estimation of found there are generally hygrophilous species. the spiders present along the Mont Nimba crest For the moment, we cannot provide definitive and their spatial distribution has been made. results on the ecological requirements of the We observe several components of this spider

i i a. in - i

1400 1600 ALtffUDE -— *M Oxyopidae 7 thA * r, ci n scytodidae I ' 1 Linyphitdae ,i... '.'.'|^;.l. L "I"..' 3 : GnftptiDsidae 9

. Aranfiiriae 5 3 2 HsTsiliidaa J ^rir.hfiMii-.L 3 3 "damisidae 3

HersllHCac 22 1 3 lycos

I l.'-V 3 Rati Clubromdae TtiragnaUildae 16 " -.: t, : 31 I f.nrr sidaa 6 . I u l.ycandae 7 I OTA. IJLiWE-.

' rji '.r

:..r -:.-r.; .<,

NE sw

TABLE4. Altitudinal distribution of spider families along Nimba crest, from SW to NE: (a) Pierre Richaud; (b) Grands Rochers; (c) Nion crest before Richard-Molard, n_W MEMOIRS OF THE QUEENSLAND MUSEUM

community, not forgetting the movements of ACKNOWLEDGEMENTS wandering spiders. There are groups on the soil surface principally characterized by Lycosidae I thank Professor M. Lamotte for allowing mc and Salticidae; three other families, Liocranidae, to participate in this project, for providing me Zodariidae and Theridiidae exist in smaller num- with the opportunity of undertaking the mission bers. Groups in the lower part of the vegetation in Guinea, and for all his advice; Denise includes ten families, the most with few repre- PaJczjewski and Mark Judson for reviewing the sentatives; Clubionidae. Thomisidae and Sal- manuscript . ticidae exist in great numbers. A last group exist of the upper part of the herbaceous stratum where LITERATURE CITED Araneidae, Hersiliidae and Tetragnathidae are found in a great numbers; the five other families BLANDIN, P. 1974. Les pcuplcmcnts d'Araignees de are less well represented - la savane de Lamlo. Bulletin de liaison des Cher- The comparison between the northern and cheurs de Lamto 3: 107- 135* P. 1981. Les southern parts of the crest indicates a possible BLANDIN, & CELERIER ML Araignces des savanes de Lamto (Cote dT voire). tendency of one family (Tetragnathidae) to prefer Publication Laboratoire de Zoologie, Ecole Nor- greater humidity. The altitudinal distribution male Supeneure, 21,2 fasc., 586 pp. shows that some families, such as Araneidae, BR1GNOLI, P. 1983. 'A catalogue of the Araneae Salticidae, Hersiliidae, Thomisidae and described between 1940 and 198T. (Manchester Lycosidae, are present from 800 to 1752m. On University Press. Manchester). the contrary, other families are preferentially GILLON, D. & GILLON. Y. 1974. Comparison du localised at the highest or lowest altitudes. The pcuplcment d'invertebres de deux milieux herbaeds oucst-africains: Sahel el savane spider families found between the altitudes 8013 preforestiere. La Terre et la Vie, 28: 429- 474. and 1000m, can be considered as being roughly LAMOTTE, M 1943. Premier apereu sur la faunc du comparable with those present in typical savanna Nimba. Faculte des Sciences Univcrsitaircs, Paris, with Loudetici- In this environment, there are a n.845. great number of spider families (Gillon and Gil- 1958. Le cycle ecologique de la savane d'altitude du Ion, 1974). More data will be required to confirm Mont Nimba (Guinee). Annalcs dc la Socictd these tendencies, as well as their presence on the Roy ale de Zoologie de Belgique 89: 1 19-150. slopes according to a greater altitudinal gradient 1967. Recherches ecologiques dans la savane de

Lamto (C5le dTvoire) : presentation du milieu et and along the whole of the Mont Nimba crest. In du programme de travail. La Terre et la Vie 21: addition, some comparisons among sites, includ- 197-215. ing absence of families, might be artefacts, due to LECLERC. J.C.. RICHARD-MOLARD, J., relative rareness of representatives LAMOTTE, M.. ROUGERIE, G. & PORTERES, The study of all spiders collected during the R. 1955. La chafne du Nirnba. Essai previous expeditions directed by Prof. Lamotte geographique, Memoire de 1'I.F.A.N.. 43. 270 pp. will certainly provide supplementary elements to PLATN1CK, N.I. 1989. 'Advances in Spider the various points mentioned in this paper It will Taxonomy 19S1-1987\ (Manchester University Press: Manchester). be necessary to characterize, with more precision, SCHNELL, R. 1966. Guinee. Acta Phytogeographica the araneological fauna of Mont Nimba. Blandin Suecica 54: 69- 72. and Celerier( 1981 (already noted the misreading 1987. Les formations herbeuses monlagnardes des of this fauna in West Africa- In addition the monts Nimba (Guest africain). Bulletin du well-collected envitonments are essentially Museum National d'histoirc naturelle, Paris, 4. savanna rather than high-altitude meadow. ser.9(2): 137-151 VISUALLY MEDIATED RESPONSES IN THE LYCOSID SPIDER RABIDOSA RABIDA: THE ROLES OF DIFFERENT PAIRS OF EYES

LS. ROVNER

Rovner, J.S. 1993 11 11: Visually mediated responses in the \ycos\d spiter Rabidosa rahida: the roles of different pairs of eyes. Memoirs of the Queensland Museum 33(2):635-638. Brisbane. ISSN 0079-8835.

Video images of conspeciflcs were presented to Rabidosa rabida (Walckenacr) (Arantne: Lycosidae) to study the roles of different pairs of eyes in a wolf spider. Four groups of spiders had one pair ofeyes occluded, and four had all but one pair occluded. Various control groups were also tested. The PLE were essential for sizeable orientation turns of up to about 160° The PME served for rapid, long distance approaches toward the stimulus; they also initiated orientation turns of up to about 50°. If close to the stimulus, the ALE initiated small turns of up to about 20° and mediated small approaches. The AME did not mediate any responses. Courtship could be triggered in males via the PLE or the PME. In females, only the PME

mediated receptive display responses to temporally patterned leg 1 movement* seen in anterior or lateral views of a courting male. [JAraneae, Lycoshlut. eyes, visii'n. communication

Jerome S. Rovner, Department ofBiological Sciences, Irvine Halt. Ohio Univt Itotft Ohio 45701 U.S.A.; 3Juh\ 1992

Very little behavioural research has been car- 1991) in Athens County, Ohio, USA I he ded out on vision in lycosid spiders, compared to methods of maintenance and the laboratory con- the extensive studies on salticid spiders ditions during testing have been described pre- (reviewed in Forster, 1985). In lycosid spiders, viously (Rovnen 1989). Spiders were not used in occlusion of the eyes has been used to test for testing until 1 week after the final moult. Tests differences in the roles of the main vs. secondary were conducted between 1000 and 2200 hours. eyes (Homann, 1931) as well as the anterior vs. For each test the spider's home cage, with its posterior eyes (Acosta *?/«/., 1982); however, the resident within, was placed with its narrow front results of both of these studies were confounded side facing a small television set (black and white by the lack of controls for vibratory stimuli. The Magnavox BH-3907; screen = 9.2cm wide, only successful behavioural investigation of the 6.6cm high), which received a playback signal role of particular eyes in lycosids (Magni et al. y from a Sony recorder (SL-HFR70). The clear 1964) demonstrated the importance of the AME front side of the plasUc cage (the other sides were in astronomical orientation and was ac- opaque) was 7cm wide and was located 3cm from unuplished by covering all but a single pair of the screen. 1 removed the cage cover and, if eyes in each group tested. Occlusion of only one necessary, gently positioned the spider with an pair of eyes at a time has proven useful in several artist brush to insure that the screen would be behavioural studies of salticid vision (e.g., within the visual field of the spiders useable Fofster, 1979) eyes (1 used separate brushes for males and In the present study of the lycosid spider females). Then, a glass cover Tone for each sex) Rabidosa rabida (Walckenacr), I used both was placed on top of the cage. A front-silvered methods of occlusion to examine the roles of mirror fixed at 45° to the floor was 0.5m above different pairs of eyes in detecting visual stimuli the cage. A video camera (JVC GX-8NU'i and mediating appropriate responses. By using aimed at both the mirror and an adjacent, second, video images as stimuli (Clark and Ueiz, 1990), identical television set receiving the same signal 1 eliminated the possibility that vibrations or from the playback recorder as the first set. Ttus chemicals from conspecifics could confound the yielded (on a second, identical recorder) a kind of results. split-screen recording. On the left was a dorsal view of the spider, which facilitated the mea METHODS AND MATERIALS ment of turning angles (accurate to the nearest 5°

) and speeds of locomotion. On the right was a Penultimate Rabidosa rabida (formerly Lycosa view of the video image concurrently being rabida) were collected in late June (1990 and presented to the spider Thui, the relationship of 636 MEMOIRS OF THE QUEENSLAND MUSEUM

the video stimulus to the spider's response could assumed would only occur in this study as visual later be analyzed. responses. Subjects used for video presentations were There were eight experimental groups, each recorded against a plain, pale background and having 10 spiders of each sex. In four experimen- illuminated evenly by a 32-W, circular fluores- tal groups a single pair of eyes was occluded; in cent bulb 0.5m above the arena. The camera was four others, all but one pair was occluded. Since located at a distance yielding a screen image the one of the untreated control spiders had failed to size of the actual subject (average body lengths: respond to video playback within two trials and female=18mm; male=12mm). The 15-min video since the preliminary tests had indicated that playback for females was a pheromone-stimu- painting the eyes can reduce responsiveness, I lated, courting male; his occasional position gave unresponsive experimental spiders addi- changes provided nearly equal proportions of tional opportunities to respond, up to a limit of anterior and lateral views of his display (total five trials. One or more days separated consecu- number of courtship bouts = 51). The 10-min tive trials undergone by any individual spider. video playback for males was a lateral view of a female walking to and fro in an elongate glass RESULTS arena (passes across the screen = 1 5 leftward and 15 rightward). Some preliminary tests made use During testing of control groups, all of the of prey images provided by a 10-min video of untreated females and all but one of the untreated three crickets (Acheta domesticus, 10mm body males showed 'orientation' and/or 'long-range length) walking to and fro. approach.' Most (16/20) did so in the first trial. To cover the eyes of spiders, I painted them 'Orientation' involved a single, rapid pivot that with two coats of water-based enamel (Top Color resulted in the spider facing the image. This pivot

Hobbylack, Pelikan AG). That this insured com- had a mean speed ( SD) of 121 ± 51 .77s (N= 12), plete occlusion had been established previously about eight times faster than turns that occurred (Rovner, 1989). Spiders were tested one or more during wandering, 16 ± 12.47s (N= 11). A 'long- days later. range approach' covered a distance of up to

I ran two preliminary tests to see if painting or 10cm, the maximum being limited by cage depth its related procedures lowered responsiveness. (12.5cm). The speed of an approach, 5.9 ±3.32 All the eyes of five females were covered with cm/s (N = 22), was about ten times faster than two separately applied coats of clear paint. Five walking during wandering, 0.6±0.31cm/s (N = other females were briefly anaesthetised (carbon 36). (The above speeds are not the maxima at- dioxide) and then restrained for 6 hr, thereby tained, since I included the acceleration and duplicating the procedure used when painting the deceleration phases in each bout of locomotion.) eyes. However, at the two times that paint would No orientation or approach responses occurred in have been applied (0 hr and +3 hr), I used water the empty arena or fully blind controls. instead. The next day, when exposed to video Results of the eye occlusion experiments are images of crickets, the clear-painted spiders were summarized in Table 1. Spiders with any one of less responsive than were the water-brushed the four pairs of eyes occluded were still capable spiders. The latter seemed as responsive as un- of orientation; those with occluded PME did not treated spiders. show approach. When all but one pair of eyes was There were three control groups, each with 10 occluded, the only group failing to show orienta- females and 10 males; these 60 individuals were tion was the one limited to use of the AME; and given two trials apiece. One group consisted of the only group that still showed long-range ap- untreated spiders exposed to the conspecific proach was the one with useable PME. When video playback, to determine how readily fully close to the cage front, a few spiders having only sighted spiders respond. A second group, also useable ALE did perform a near-field approach untreated, was exposed to a 10-min video of an response, edging forward less than one body empty arena, to determine if that alone had a length. For all the responding experimental stimulating effect. (Light from the screen = about groups, the mean number of trials ( ± SD) needed 800 lux; incident light from above = about 350 to obtain either an orientation or a long-range lux.). The third group consisted of fully blinded approach to the image was 1 .2 ± 0.53 (N=63) for spiders exposed to the conspecific playback, to females and 2.4 ± 1 .24 (N = 55) for males (Mann- see if such spiders would perform behaviours Whitney CMest, Z= -6.545, P<0.0001). (orientation, rapid approach, or display) that I Spiders with useable PLE showed orientation VISUAL RESPONSES IN RABIDOSA RAB1DA 637

Long range Oncnlation Display may have prevented some visually stimulated ('.indtti.-n approach males that were facing the stimulus at the outset * r 9 9 s 5 J from being included in the courtship tcnal for Untreated 6 6 8 7 3 5 some experimental groups.) Only males with use- Rmpiy rircna able PLE Of PME show ed courtship accompanied Fully blind D D by orientation, and only males with non-occluded N0PLE5 6 7 ID 4 5 : PME showed courtship accompanied by ap- in No PMEs 6 U 4 proach (Tabic 1 1

No ALE* 10 7 9 5 4 1 Nn AMEs 10 6 9 6 4 3 DISCUSSION

11! Dills PL! i 7 D 5

Only PMEs 9 6 19 5 3 2 Data presented here indicate that R. rabida's

Only ALE* 3 1 posterior eyes play major roles in mediating Only AMEs u n I) D responses to important V isual stimuli, as was earlier predicted for lycosids on anatomical grounds TABLE 1. Number of ? and 6 spiders responding to (Homann, 1931; Land, 1983). The serve the video playback. N = 10 spidcrs/sex/eondition. Each PLE same function that they in salueid spider was allowed more man one [rial to respond: do spiders, thai controls, up to two trials: experimental*, up to five of detecting stimuli in an extensive visual field (rials. Only 63 that also performed orientation or and initiating the largest orientation turns. They approach are included under 'Display*. cannot mediate approach behaviour. On the other hand, the PME are essential for mediating rapid, turns of up to about 160°. Those with only the long-range approaches to stimuli. They can also PME useable performed orientation turns ofup to initiate oncnlation turns of up to5(F. about 50°. If near the stimulus, those with only As to the anterior eyes, m close range the ALE the ALE useable showed orientation turns of up can play small roles in orientation and approach. to about 20°. This contradicts Land's (1981) suggestion that During playback of male courtship, only those only the large posterior eyes are involved in prey females with non-occluded PME performed leg- capture. As predicted for lycosids by Land [ibid, ), waving receptive displays (Table I). Trus brief the AMEof /?. mhuh play no role by themselves display occurred 3.0±1.9s (.V = 30) after the in mediating turns or approaches toward a target. abrupt termination of the male's courtship bout. Whether they serve for other than polarized light The receptive display was sometimes performed detection (Fvlagni ct ai. 1964) remains to be ex- unimpeded, this being the case in females which plored. had not yet reached the front of the cage. How- The present findings on the anterior eyes of R. ever, it was usually constrained, since most rabida may not apply to all lycosids. For ex- females quickly approached the cage front and ample, unlike ft rabida > inArctosa variarm the edtheij anteriorlegs agaiosi the wail. Still, the AME are larger than the ALE {ibid.). Also, local timing of their response remained precise. tiering is found tn Ihe retinae of the AME ol Females showed this display while seeing Geolycosa godefftoyt hut not in the type genus anterior or lateral views of the- courting male. Lycvsa (Blest and O* Carroll, 1 989), Furthermore, Up to 6/1 of the males tested in each condition the finding of high nocturnal sensitivity in the performed courtship display, which occurred in anterior eyes of Lycosa tarcntula (Carncaburu el every group. It even occurred in two males during qL, 1990) suggests that an understanding of the the empty arena playback and in three of the fully capabilities of the anterior eyes of lycosids will blinded males, such daUi reflecting a previously require examining different species under various noted readiness of R. rablda to sometimes court illumination conditions. Finally, I must point out in response to subnormal stimuli (Rovncr, 1968). that I did not test for any possible collaboration Perhaps these non-visually stimulated courtships between different pairs of eyes. Collaborative were triggered by mechanical cues resulting from visual mechanisms were revealed in salticid my moving the cage to the testing site or from my spiders by ipsilalcral blinding (Fofster, 1979). positioning the spider with a brush (to duplicate As in the salticuls (Forster, I985)> orientation ihe procedure used on experimental spiders). For and approach arc the initial behavioural respon- this reason, only courtships occurring during tri- ses of lycosid spiders to moving images, whether als in which orientation or approach also occurred prey or mating partners. Thus, the perform. were scored as visually initiated displays. (This of such behaviours docs not specify the spider's 63* MEMOIRS OF THE QUEENSLAND MUSEUM

motivational state. However, the fact that twice LITERATI/RECITED as many trials were needed to obtain these behaviours in males as in females suggests that such ACOSTA. J., ORTEGA, J. & RODRIGUEZ- initial responses relate primarily topredation, i.e., SANABRA, F. 1982. Vision and predatory' be- that adult females, on average being hungrier than haviour in Lycosa JasciivetUris. 6th European 21-22. males, are more responsive to visual stimulation. Neuroscience Congress. Malaga. Spain: A subsequent occurrence of display behaviour BLEST, A.D.&G'CARROLL,D. 1989. The evolution of the tiered principal retinae of jumping spidery indicates a switch to sexual motivation. (Araneae: Salticidae). Pp, 155-170, In R N.Singh For display to occur in females the PME have and N.J. Strausfeld (eds). 'Neurobiology of sen- to be useable, and they are sufficient by themsel- sory systems', t Plenum Press: New York). ves for mediating this response. The fact that CARRICABURU. P.. MUNOZ-CUEVAS, A. & OR- lateral as well as anterior views of the courting TEGA-ESCOBAR, J- 1990. Elcctrorctinography circadian rhythm in Lycosa taremula male elicited responses suggests that the cue is and

I' Araneae. Lvcosidae). Acta Zoologiea Fennica movement of the male's black leg I: it extends " 190:63-67 forward in a pumping-like motion of increasing CLARK, D.L. & UETZ, G.W. 1990. Video image frequency and then abruptly flexes back. Unlike recognition by the jumping spider Maevia inclem- sahicid face-to-face courtship, in which the per* ' Araneae: SaJticidae). Animal Behaviour 40: ceived form of the male may be as important for : -S90. the female as his behaviour {ibid.), female R, FORSTER, L.M. 1979. Visual mechanisms of hunting rabida need not view the male's anterior. Ap- behaviour in Trite ptaniceps, a jumping spider parently, detection of a temporal pattern of move- (Araneae: Salticidae). New Zealand Journal of ment suffices for recognition. Zoology 6: 79-93. 1985. Target discrimination in jumping spiders As tn visual !y-!riggered courtship display, male lAraneae:Saiticidae). Pp. 249-274. In RG. Barth differ salticids, which re- R. rabida from male (ed). 'Neurobiology of arachnids'. (Springer- quire one particular pair of eyes, the AMEU/^/.i. Verlag: New York).

Stimulation of either pair of a male R. rabuhi S HOMANN.H. l93I.BeitragezurPhysiologiederSpin- posterior eyes suffices for triggering courtship. If nenaugen. HI. Das Sehvermogen der Lyeosiden. the PLE are involved, orientation precedes dis- Zeuschritt lur Vergleiehende Physiologie 14: 40- play. If the PME are involved, approach can 67.

precede display, although it need not do so. LAND, M.F. 196 1 . Optics and vision in invertebrates. Pp. 471-592. In H. Autrum (ed). 'Comparative Given that the movement-detecting PLH arc physiology and evolution of vision in inver- sufficient for courtship onset rn R, rabida, an tebrates. Handbook of sensory physiology adequate stimulus in this species may be any VI1/6B\ (Sprinuer-Vexlag: New York). object of an appropriate size, speed, and perhaps MAGNL F.. PAPL R, SAVELY, H.E. &TONGIORGI,

movement pattern entering the extensive visual P. 1 964. Research on the structure and physiology field of the PLE. Since the other three pairs of of the eyes of a lycosid spider. I!. "The role of lycosid eyes are also assumed to be movement different pairs of eyes in astronomical orientation. Archives llaliennes de Biologie 102: 123136. detectors (Land, 1981), there may be no need to ROVNF.R, J.S 1968. An analysis of display in the require- their involvement in additional visual lycosid spider Lycosa rabida Walckenaer. Animal analysis before initiating courtship. However, Behaviour 16: 358-369. this may not be true for all lycosids. In particular. 1989. Wolf spiders lack mirror-image responsive- female Pardosa laura have been hypothesized to ness seen in jumping spiders. Animal Behaviour use form vision forspecies discrimination (Suwa, 3S: 526-533. (984), [folate R hutra likewise do m>, eyes other SUWA, M. 1984. Courtship behavior of three new than the PLE would be have to be used lor such forms in the wolf spider Pardosa laum complex. form analysis in order to inititatc courtship. Journal of Ethology 2: 99-107. THE SPATIAL DYNAMICS OF LlNYPHHD SPIDERS IN WINTER WHEAT

K.D. SUNDERLAND andCJ. TOPPING1

Sunderland, K.D. and Topping, CJ. 1993 1111: The spatial dynamics of linyphiid spiders in winter wheal. Memoirs of the Queensland Museum 33 (2): 639-644. Brisbane. ISSN 0079-3835

The densiu of linyphiid spiders was monitored accurately throughout the growing season

in a field of winter wheat in southeast England in 1990 and 1 991, Numbers increased untii harvest in 19*^1, but declined before harvest in 1990, possibly due to drought conditions. The pattern of natality in 1991 closely mirrored the pattern of change in density, suggesting that reproduction, rather than immigration, was the predominant factor underlying (he increase in density. Aerial activity, as measured by deposition traps and a rotary trap in the field, and a suction trap at the edge of the field, increased progressively during the growing season. Results from a short-term field caging technique, used to measure net migration rates, indicated that there was little immigration before July {thereafter high sampling variance, caused by aggregation in weedy patches, precluded meaningful analysis).QA/t7/iea

Keith Sunderland and Chm Topping, Horticulture Research International, Utttehamplon, 7 West Sussex BNJ7 6LP, England; current address; Scottish Agricultural College, Craihstone, Bucksbum, Aberdeen AB2 9TR. Scot/and, UK; S September, I99L

There are few studies of ihc population Density Sampling dynamics oi predators. For example, Stiling Twenty-five 144m 2 squares were marked out (1988), in an examination of the incidence of inside a 60 x 60m area (adjacent to one edge of density-dependence in invertebrate populations, the field in 1990 and 30m from the nearest edge quotes 62 population dynamics studies; of these in 1991). Fifteen density samples were taken at each sampling interval (approximately weekly) 60 relate to phytophages, 2 to parasitoids and following a Latin Square design. The sample Will none- to predators. There are also few population consisted of a randomly-selected 0.5m 2 area of dynamics studies of migratory species, because ground delimited by a metal ring and sampled of lhe methodological problems involved in using a vacuum insect net (D-vac). Vegetation quantifying migration. The present study, of the and the top lem of ground, within the 0.5m- population dynamics of linyphiid spiders (the sampled by D-vac. were then immediately hand species concerned are all migratory predators), searched for spiders. The suction catch was kept was undertaken to collect some basic data in this at 10°C and returned to the laboratory for live- neglected area, and also because these species are sorting. Therefore spiders were collected from 2 known to be valuable predators of crop pests 7.5m of habitat (crop, ground surface and imme- diate sub-surface) on each occasion (Topping and (Sunderland el ni, 1 986). This paper summarises Sunderland, 1992). The D-vac collected only the main trends for total linyphiids; consideration about 50% of the total number of spiders present will be given to individual linyphiid species in in a sample unit, the other 50% were uncovered later publications. by hand-searching (sec also Sunderland et aL, METHODS AND MATERIALS 1987).

Natality Total spider density was measured throughout Adult female linyphiids were collected fron most of the growing season of 1 990 in a 17ha field the field, adjacent to the 60 x 60m area, at weekly ol winter wheat (c.v. Pastiche) in southeast intervals and incarcerated individually in 9cm England. In 1991, density, natality and migration diameter plastic Petri dishes lined with moist were measured in a 3ha field of winter wheat lev. filter paper. The dishes were returned immedi- Riband), 24km from the 1990 study field. The ately to a ventilated box m the study field and fields were treated with agrochemica! applica- examined at weekly intervals. Mean daily tions, following normal farm practice: insec- temperatures in the Petri dishes did not diverge ticides were no! required in either year. from field temperatures (measured on the ground ) )

64fJ MfcMOIRS OP THEQUEENSLAND MUSEUM

surface. undeT weed cover) by mote than I°C which therefore received only aerial immigrants. grew, the deposition traps were I warmer in spring, cooler in summer). The dishes As the crop were inspected at weekly intervals and the fol- progressively raised on wooden supports to main- lowing statistics recorded fj) proportion of tain the level of the fluid surface constantly at ca. spiders producing eggsacs in the first week of 5cm above the top of the crop canopy, The traps incarceration, (ii) lime (days) to emergence of were emptied at approximately weekly intervals. $pjderbngs,(iii) number of spiderlings emerging, Nomenclature follows Roberts (1987). and (iv) number of undeveloped eggs (by dissection of the eggsac). These results were used, in RESULTS conjunction with information on the density of adult females, to calculate daily natality rates for The study was based on data from more than each species and then combined to give a com- 39,000 individuals belonging to 53 specie posite spider natality curve, species were dominant (Table 1). All species belong to the family Linyphiidae, with the excep- tiriiagiiathid Pachygnatha degeeri. Net Migration Rate tion of the Lepthyphantes tenuis was the most abundant Migration to and from small areas of the crop species in 1991 and the second most abundant in was suppressed by the use of ten itainJess steel 1 990. Species composition was similar in the two spider-proof cages. The cages were circular, 2 years, the only notable differences being Uiat 0.5m by lm tall t aj*d made of mesh with 3 x 2.5 Meioneta rurestris was relatively more abundant perforations mm' 2 (too small to allow passage of in 1990, the reverse being true for Oedothorax firs? insist linyphiid spiders). The bases of the spp. Here, all spiders are treated as a group. cages were sealed with sufficient compacted soil In 1990, density of total spiders increased in Total to prevent entry or exit of any spiders. : spring to reach a peak of 7Snv on 18 June and spider density inside the ca^es was assessed as thereafter declined, apart from a short-lasting above. Becuus.: d were moved to a new peak (made up entirely of immature spiders) just location within the Latin Square each week, there before harvest ( Fig. 1 ) The pattern was different was assumed to be insufficient lime for the in 1991 (Pig. 2); spider density built up in two processes of natality and mortality to be sig- sups (the first in May/June, the second in August) nificantly affected by the changed microclimate 2 to reach a peak of 123m- just before harvest. inside the cages; therefore differences in the Mean air temperatures in June and July were change in density from one week to the next a slightly higher in 1990 ( 15. i C) than 1991 between caged and uncaged parts of the crop ill^C), but rainfall was considerably lower were considered to be a measure of net NfigftU (68mm in 1990 cf 19lmm in 1991). The semi- drought conditions in the summer of 1990 may Aerial Densii v OF SWDEBS have had a deleterious effect on spider survival.

A 46cm Propeller Suction Trap (Taylor, 195? ), with an air throughput of TOm"" uiin J and a sam- -' M 1 Species 1 1991 pling height of 142cm above the ground surface. Mettmeui rurestris (C.L. Koch) -:_ > 5.4 located at the edge of the study field, was emptied

:•. l^pthypbatUes tenuis (Blackwall 27.H 1 J daily. In addition, a rotary insect net was used to

1 Erigane atra (Blackwall) 9.S *.. collect spiders within the field, 25cm above the

MilUriana inerrans (O.P. -Cambridge I 5 - 0.4 lop of the crop canopy. The 10m long rotor arm

J Eri^ane promiscua (OTVCambndpe j 5.0 travelled at 6.3m sec~ and the 56 x 25cm net at the end of the rotor arm (which was designed to ErigmtK dentipalpiB (Wider) : 2 5.0 sample air isokineticaUy, (Taylor, 1962)) Pmchysnailui degeeri Sundcvall :.o n :

3 1 processed 53.8m min' and was also emptied Bathyptwmes xa:ctiix (BlnekwaM) 2,7 H ')

'.. • daily. To measure rates of input of spiders into Oc- dot he rat fttscus 1 RUckwalt : i

the field, a set of seven deposition traps were Oettuthorai retuxus (Weslring) 2.1 3.9 deployed at 15m intervals between the rotor and Fitnamtitnops suit:i'from (Wider) 1.0 0.1

the 60 x 60m area. Each trap consisted of a 10cm Otilothrtrax apicatus (B\uc\w^\\} 0.3 6.9 2 deep, I m fibre glass tray , filled with water and 1 Olhers* 5,0 6.6 ethylene glycol (20:1) plus 1% detergent, fitting 2 inside a 1.6m metal tray containing the same TABLE I. Species composition of adult spiders in samples in 1990 1991 fluid. The outer tray acted as a barrier to prevent density (n=1562) and (n= 1457). spiders walking from the crop into the inner tray. SPATIAL DYNAMICS OF LINYPHIIDS 641

Dale M 95% CL SPIDER DENSITY IN A FIELD 9 April 2.1 (-1.2-5.4) OF WINTER WHEAT 23 April 2.2 (-2.3-6.7) IN 1990

1 Harvest 1 May -2.0 (-8.7-4.7)

14 May 2.7 (-11.8-17.2)

2! Maj -12.8 (-29.5-3.9)

1 1 June -16.5 (-33.9-0.9)

18 June -8.0 (-23.1-7.1)

3 July 0.7 (-13.2-14.6) M*W 9 July 2.9 (-10.4-16.2) j I I 17 July -5.1 (-26.7-16.5) Jan Feb Mar Apr May Jul Aug Sep 23 July 1.8 (-20.5-24.1)

SPIDER DENSITY IN A FIELD 29 July -9.5 (-43.8-24.8) OF WINTER WHEAT 5 August 24.2 (-19.9-6S.3)

IN 1991 (solid line) 19 August 12 7 (-41.4-66.8)

9 September -10.0 (-52.7-32.7) 95% CLOF MIGRANTS 24 September 7.1 (-19.6-33.8) I 60 (vertical dashed line)

TABLE 2. Indices of net aerial migration, M (95% CL), in 1991.

circumstantial evidence that reproduction (as op-

Feb Mar Apr May Jun Jut Aug Sep posed to immigration) is the predominant process driving increase in density of spiders in the field. SPIDER DENSITY IN CAGED Data on migration can be examined in the light of AREAS OF WINTER WHEAT this hypothesis. Aerial activity of spiders, as IN 1991 measured by catches in the 1.4m suction trap at the edge of the field, tended to increase steadily E 60 from March to August, and this was followed by a much larger increase in September (Fig. 5). A similar pattern of aerial activity was evident inside the field, as indicated by the catch in deposition traps and in the rotary trap (Fig. 6). To assess whether there had been a a net gain or loss of Jan Feb Mar Apr May Jun Jul Aug Sep spiders from the field over a particular period, the density of spiders inside caged areas was com- FIGS. 1-3. Total spider density in a field of winter pared with the density outside,to give an index of wheat: 1990 (Fig. 1); 1991 (solid line) and 95% CL net 2 migration (M); of number of migrants m~ (vertical dashed lines) = (F2-F1)-(C2-F1) = F2-C2 (Fig. 2); in caged areas, 1991 (Fig. 3). M where Fl and F2 are densities in the uncaged

Density in the caged areas of the crop in 1991 part of the field in weeks 1 and 2 respectively, and followed a similar pattern to that in uncaged areas C2 is the density in the caged part of the field in (Fig. 3). 95% confidence limits increased during week 2. (F2-F1 represents change in numbers due July and August (Figs 2, 3) due to aggregation of to natality, mortality and migration, whereas C2- spiders in weedy patches of the crop (significant- Fl represents change in numbers due to natality ly more spiders in weedy than bare areas; paired and mortality alone, because migration was sup- t-test, n = 5 dates, p = 0.05). pressed by caging). The standard error of M is The pattern of change in density with respect calculated as the square root of [SE F22 + SE 2 to time is examined, below, in relation to natality C2 ]; the 95% CL's on M are therefore large and migration for 1991. because they are compounded of two standard

2 1 Natality rates were c. 4m day in the spring, errors. Values of M are shown in Table 2. Positive 2 1 2 briefly 10- 15m" day in late July, then 8m" day"' values indicate immigration and negative values in August. The pattern of daily natality was indicate emigration. However, all 95% CL's span similar to the patterns for immature spider densi ty the range from negative to positive and therefore (Fig. 4) and total spider density (Fig. 2); this is no significant net migration can be demonstrated. 642 MEMOIRS OF THE QUEENSLAND MUSEUM

DAILY CATCH OF SPIDERS IN A 1.4 M SUCTION TRAP AT THE EDGE OF A FIELD OF WINTER WHEAT IN 1991

Jan Feb Mar Apr May Jun Jul Aug Sep

4. Total spider density and total FIG. immature ( •) Feb Mar Apr May Jun spider natality (continuous line) in a field of winter wheat in 1991. DAILY CATCH OF SPIDERS IN A ROTARY TRAP • These 95% CL's are plotted on the curve of total and IN 7 SQ M OF q DEPOSITION TRAPS """ density (Fig. the extent the dotted spider 2); of IN A FIELD OF WINTER WHEAT line above and below the density curve shows the IN 1991 amount of immigration and emigration, respectively, that could have occurred between any two dates. The following conclusions can be drawn; (i) any migration that may have occurred in April was small (i.e. on a similar scale to density sampling variance), (ii) if immigration occurred in A / May and June it must also have been on a small 0--0-. J*&#

scale, but there could have been a large emigra- * * Mti I t 1 1 tt t ' ' I' I 'I " " tion, and (iii) in July, August and September 95% Feb Mat Apr May Jul Aug Sep CL's were very large (due to spider aggregation, see above), with no obvious bias in favour of FIGS. 5, 6. Daily catch of total spiders in winter wheat either immigration or emigration. field in 1991. 5. In 1.4m suction trap at edge of field;

6. In rotary trap (•) and in 7m of deposition traps . DISCUSSION densities of lycosid and linyphiid spiders vary greatly according to location (Table Densities There appear to be no previous quantitative 3). arachnological studies in which the seasonal pat- are often lower in graminaceous crops (Table 3). 2 terns of natality and migration are compared with The peak spider density of 123m in the present the seasonal pattern of density using consistent study is comparable with, if somewhat greater units. Examples of other arachnological studies than, densities recorded in cereals by other involving density or natality estimation are given authors (Table 3). In common with the present below. The majority of investigations where den- investigation, nearly all the studies in Table 3 sity has been measured are for grassland; peak reported large confidence limits due to aggrega-

Species Family Habitat Density Authors)

Geolycosa godeffroyi (L.Koch) Lycosidae pasture 1 Humphreys, 1976

Trochosa terricola Thorell Lycosidae grass heath 70 Workman, 1978

Pardosa palustris (Linnaeus) Lycosidae alpine meadow 9 Steigen, 1975 TABLE 3. Oedothoraxfuscus (Blackwall) Linyphiidae pasture 155 De Keer and Maelfait, 1987 Maximum density Erigone atra (Blackwall) Linyphiidae pasture 318 De Keer and Maelfait, 1988 estimates (number

Erigone arctica (White) Linyphiidae dune grass 330 van Wingerden, 1977 m ) in a range of

Total spiders Festuca grass 840 Duffey, 1962 arachnological studies. Total spiders rye grass 43 Alderwiereldl, 1987

Total spiders maize 49 Alderwiereldt, 1987

Oedothorax and Erigone Linyphiidae winter wheat 53 Nyffeler and Benz, 1988

Total linyphiids winter wheat 60 Fraser, 1982

Total spiders winter wheat 75 Sunderland 1987 SPATIAL DYNAMICS OF LiNYPHULn M.-v

lion of spiders There seem to be no previous phenology of ballooning spiders at two locations publications describing the seasonal pattern of in eastern Texas. Journal of Arachnology 13: HI* spider natality, but a few authors

f>o\i([, Koch 1 865) ( Araneae: Lycosidae). Jour- short-term field cages in the present study

nal of Animal Ecology 45 : 59-80. piovided useful estimates of the upper limits 10 MEIJER. J. 1976. The immigration of spiders migration (except when sampling variance be- (Araneida) info a new polder. Ecological En came very large) and it is expected that this tomolugy 2: 8} -90, technique will yield better results when data are NYFFELER,M.&BENZ,G. 1988. Prey andpredaior. analysed Tot individual species. In addition, when importance of micryphanlid spiders in winter the rotary trap has been calibrated, it should be wheat fields and hav meadows Journal of Applird possible to estimate rates of aerial immigration Entomology 105; 190-197. and emigration from a comparison of deposition ROBERTS, MJ. I9S7. The spiders of Great Britain and rotor catches. and Ireland, Vol. 2.*(Hartcy Books: Colchester England) ACKNOWLEDGEMENTS SCHAEFER, M. 1978. Some experiments on the regulation of population density in the spider Fioronia buccutenla (Arancida: Linyphiidac). We thank Mt Bewsey and Mr Reeves (HRI) for Symposia of the Zoological Society of London 42.

'ructing the rotary trap ? Mr Fenlon (HRI) fur 203-210. statistical advice, and Dr Jepson and Mr Thomas STE1GEN. A.L. 1975, Energetics in a population of "i »' Southampton University) lor useful discussions. Pardoia palustris (L.)( Araneae, Lvcosidaej The work was funded hy tbo Natural Environ- Hardangervidda. Pp 129-144 tn F Wielgo ment Research Council (Jomi Agriculture and (ed). 'Fennoscandian tundra ecosystems Part 2V (Springer: Berlin). Environment Programme) and the Ministry of STILrNG. P. 1988. Density-dependent processes And Agriculture, Fisheries and Food. The first author fee) factors in insect populations. Journal of is grateful to the British Council for the oppor- Animal Ecology 57: 581-593. tunity to deliver this paper in Australia. SUNDERLAND, K.D. 1987. Spidcisand cciealaptuds in Europe. Bulletin SROPAVPRS 1987/X/l; 82- LITERATURE CITED 102. 1991. The ecology ol spideisinccrcala Proceedings ALDERWEIRELDT. M. 1987. Density fluctuations ol of the 6th International Symposium on Pests and spiders on maize and Italian ryegrass fields Diseases ot Small Grain Cereals and Maize. Mededclingen van de Fakultcit LamtbaiHv- Board of Plant Protection. Halle/Saale, Germany wcLctischappen Kiiksuiuvcrsitcit Gfittl 52. 273- 1:269-280. 282. SUNDERLAND, K D. FRASER, A.M. & DIXON 1986. Field laboratory studies on DEAN f DA & STERLING. W.L 1085 Size and AFG and 644 MEMOIRS OF THE QUEENSLAND MUSEUM

money spiders (Linyphiidae) as predators of TOPPING, C.J. & SUNDERLAND, K.D. 1992. cereal aphids. Journal of Applied Ecology 23: Limitations to the use of pitfall traps in ecological 433-447. studies exemplified by a study of spiders in afield SUNDERLAND, K.D., HAWKES, C, STEVENSON of winter wheat. Journal of Applied Ecology 29: T., MCBRIDE, L.E., SMART, L.E., SOPP, P.I., 485-491. POWELL, W., CHAMBERS, R.J. & CARTER, WINGERDEN, W.K.R.E. VAN 1977. 'Population O.C.R. 1987. Accurate estimation of invertebrate dynamics of Erigone arctica (White) (Araneae, density in cereals. Bulletin SROP/WPRS Linyphiidae)'. (Thesis, Free University: Amster- 1987/X/l: 71-81. dam). TAYLOR, L.R. 1955. The standardization of air-flow 147pp. Life cycle and population in insect suction traps. Annals of Applied Biology WORKMAN, C. 1978. 43: 390-408. dynamics of Trochosa terricola Thorell (Araneae: heath. Ecological 1 962. The absolute efficiency of insect suction traps. Lycosidae) in a Norfolk grass Annals of Applied Biology 50: 405-421. Entomology 3: 329-340. MORPHOLOGY OF THE EMBRYOS AT GERM DISK STAGE IN ACHAEARANEA JAPON1CA (THERIDHDAE) AND NEOSCONA NAUTICA (ARANEIDAE)

HIROHUMI SU2UKI AND AKIO KONDO

Su7.uki 7 H. and Kondo, A. 1993 11 II: Morphology of ihe embryos, at germ disk siagc in Achaearanea japonica (Theridiidae) and Neoscona naniica (Araneidae). Memoirs of the

Queensland Museum 33(2 »: 645-649 Brishane. iSSN 0079-8835.

Embryos at the germ disk stage were investigated by electron microscopy in Achaearanea japonica (Theridiidae) and Neoscona naniica (Araneidae). In both spiders, the germ disk was composed of spherical cells, which had almost no large yolk granules. In the inner part of the embryo, several large yolk granules were packed by cell membrane with various organelles and glycogen crannies similarly to lycosid spiders. In Achaearaneajaponica there were very flal cells which possessed several large yolk granules in the surface region where the germ disk was not formed. In Ncosrrma nuuiica cells were not observed in that region at all, so the packages of large yolk granules were exposed directly to penvitelline space The araneid type can be distinguished from Ihe agelenid and theridiid type-

Die Embryonen von,4f hiuaruneajapifn!i-iiiThcT\d\\diic)undN^(n' OftQ naniica [ Araneidae) im Stadium dcr Kennscheibc wurden millels des Elcktroncnmikroskops untcrsucht. Die Kcimscheiben der beiden Spinnen bestanden aus spharischen Zellen, die groBe Dollcrkomchen nur selten haltcn. In dem inncrcn Ted von dem Embryo wurden manche groBen Dotterkdmchen mil verschiedenen Organellen und Glykogenkornchen von dcr Zellmembran gepackt, wie im F;ille von den lycosiden Spinnen. Im Falle von Achaearanea japonica gab es feehf llache Zellen, die manche grotton Dollcrkomchen batten, in dent oberriachlichen BcZtrfc, wo die Kcimscheibe nicht gehildel wurde- Im Falle von Neoscona naniica wurden die Zellen in dem Bc^irk schlicBlich nichl beobachl. also waren die Paeke ten OouerW-mchcn direki in der Periviiellinhohlc entbloBt. Der anmeide Typm BOB sich von dem agelemdeu Typus und dem iheridndcn Typus unterscheiden. \J$pideK Achaearanea, Neoscona, embryo, germ disk, morphology*

ftirohUmi SHZtiki and Ak'tO Kondo, Department of Biology, Faculty of Science. John University. 2-1. Miyawa 2 chome. hiwahoshi-sht, Chiha 274, Japan; 29 October. 1992.

Three types of germ disk formation are known on the germ disk region and few cells remain on in the embryos of spiders In the most common tbe other hemisphere ( Montgomery, 1 909). In the type known in Agelenidae, the germ disk is third type found in many Araneidae, a rip appears formed on a hemisphere of the egg as the result in the blastoderm, so the yolk mass is exposed of transformation of squamous blastoderm cells (Sekiguchi, 1957). Then all blastoderm cells take in this region into spherical cells. Egg surface of thfi other hemisphere is covered with squamous part in germ disk formation, and any cells are not cells (Holm, 1952). In the second type known in observed in the region where the germ disk is not Theridiidae, the most blastoderm cells converge formed. A comparative study of the spider embryos at germ disk stage was carried out under

light microscope ( Kondo and Yamamoto, 1975). Tbe study of germ disk formation under electron microscope was executed in lycosid spiders,

whose germ disk formation is the agelenid type (Kondo, 1969, 1970). We had to examine whether remaining cells connect each othcrornot

in theridiid type and whether extreme thin cells exist or not at the superficial region where the PIG. I. The embryo at germ disk stage in Achaean germ disk is not formed in araneid type. In present japonica (a) and Neoscona nautica (b). In A.

iaponica a u-^cvlK;»re found on region where germ study, electron microscopic investigation of the disk is not formed, in N. mmiica, exposed yolk n embryos at germ disk stage was carried out in is found on thai region. Scalc=0.2mm. Achaearanea japonica and Neoscona naulica. 646 MEMOIRS OF THE QUEENSLAND MUSEUM

& ps

r* = •

FIGS. 2-8. 2. A. japonica. Peripheral region of two germ disk cells. (0.5jxm.) Arrowhead: Desmosome-like structure, 3. A mid-body. Many microtubules, (ljxm.) 4. A germ disk cell. Main components of cytoplasm are fatty granules (fg) with medium electron dense matrix, and no large yolk granules. (lOjun.) n: nucleus, ne: nuclear envelope. 5. Mitochondria have high electron dense matrix and the cristae are found faintly. (0.5jxm.) 6. Cup-shaped mitochondrion. (0.5jim). cm: cell membrane. 7. Ring-shaped mitochondrion. (0.5 jim.) 8. Fatty granule (fg) lacking complete limiting membrane, but partly enclosed by smooth-surfaced endoplasmic reticulum (er). (ljxm.) gg: glycogen granules, m: mitochondria, ps: perivitelline space. Scales in parentheses. ,

EMBRYO MORPHOLOGY IN TWO SPIDERS Ml

MATERIALS AND METHODS Mitochondria had a high electron dense matrix, and the cristae were found faintly (Fig, 5). Many Achaearaneujaponica (Bbsenberg and Strand) figures of mitochondria showed oval or a (Theridiidae) and Neoscona naatica (L. Koch) bars, and several showed cups (Fig. 6) or rings

(Araneidae) were used here. In A. juponica, the (Fig. 7 i eggs collected in August were used. In N. nautica Smooth-surfaced endoplasmic reticula were the eggs laid in glass tubes at laboratory were often found enclosing fatty granules (Fig. 8) used. The observation of the live eggs was carried Rough-surfaced endoplasmic reticula were no* ou! in liquid paraffin, where the opaque chorion observed. became transparent. The eggs were fixed for 3 Typical Golgi bodies were rare. Vesicles were D hours at 4 C in 2% paraformaldehyde and 2.5% generally observed in the cytoplasm. The glutaraidehyde solution in 0.1 M phosphate buff- glycogen granules, 0. 1 p.m in diameter, were very er, pH 7.4, containing 0.2M sucrose Through high electron dense, and scattered. fixation, the eggs were cut in half with a tungsten "The superficial region where the germ disk was needle. After rinsing more than one hour with the not formed was occupied by remaining flat cells. same buffer containing 0.2M sucrose, the These cells were about lOOftm in length, abou! samples were postfixed for one hour at 4°C in 2% 25u.m in thickness at the central part, but often i'smic acid in 0.1M phosphate buffer, pH 7.4, less than Iprn near the peripheral one (Fig. 9). without sucrose. Afterrinsing with the same buff- The diameter of nuclei was about 13p.m. Des- er without sucrose, samples were dehydrated in rnosome-Iike structures were observed between rthanol series, transferred to propylene oxide. remaining cells (Fig. 9). and embedded in Quetol-812. Ultrathin sections Several large yolk granules occurred in these were cut on a ultra-microtome, LKB-4800, remaining cells. The largest yolk granule was sunned with uranyl acetate and lead citrate, and 2G>m in diameter. Vesicles were sometimes ar- examined under Hitachi HU-12A electron micro- ranged along the large yolk granules (Fig. 10). scope. Thick sections were prepared simul- The interior of the embryo was filled with yolk taneously, and stained with methylene blue for packages composed of several large yolk light microscopy granules, various organelles and glyc

granules, and enclosed bv cell membrane - RESULTS id.

(feast \ntcA A CH* $A R Jl HBA M t'OMCA The eggs were spherical and 0.5mm in Ellipsoidal eggs of A. nausica had longer axis diameter Typical theridiid type germ disk forma- measuring 1 ,2mm and shorter axis measuring Lion was observed (Fig, la). At 25°C, the eggs imm. At 23°C\ 45 hours were needed from took 24 hours to the germ disk stage after oviposi- oviposition to establish the germ disk (Fig. lb). lion. The germ disk was a single layer of spherical The germ disk was formed as a single layer cells, but the cells piled up in its central region. cell diameters composed of spherical cells, but the cells piled up The were about 45pm, and those nuclei Between these in its central region. The diameter of germ disk of were about 20pm. cells, there structures but Cells was about 30pm,. and that of nuclei was were desmosome-hke no in- about 15p.m. Desmosome-like structures were terdigitations. observed between germ disk cells at the superfi- Various types of lysosome-like bodies were cial region (Fig. 2). hut interdigitations were not observed (Figs 12-14). Mitochondria often observed. Mid-bodies were observed rarely (pig, crowded around the nucleus (Fig. 15). Several Golgi (Fig. 3). Narrow cytoplasmic bridge connected cells bodies were observed 16) Othercom- adjacent to each other, and contained many ponents of cytoplasm were similar to those ofA microtubules. japanica The main components of cytoplasm were tatty No blastoderm cells were observed in the surface region the disk was not granules, 1- 3pm in diameter, with a matrix of a where germ foi (Fig. medium electron density (Figs 4, 8). The limiting 17). membrane was often obscure. The germ disk cell Ad -ilmoM do large yotk granules. Fine yolk DISCUSSION granules, less than 5yun in diameter, were observed. In this study, some cup- or ring-shaped 648 MEMOIRS OF THE QUEENSLAND MUSEUM

n 16 vm

mitochondria were observed. These types of were figured in embryo of A. tepidariorum

mitochondria were not reported in ihe embryos of (Suzuki and Kondo, I99l). In N. nautica, many they were the lycosid spiders (Kondo T 1969, 1970). but mitochondria found surrounding EMBRYO MORPHOLOGY IN TWO Sl'IDHKS M9

yolk cells. In this investigation, detailed observa- vm- tion of yolk cells was not carried out. The embryo at gemi disk stage in A. japonica had very flat cells with several large yolk granules (Fig. 18). Distinct differences of cytoplasm were noi observed between spherical germ disk cells and flat remaining cells. As in A. tepidariorum

(Suzuki and Kondo, 199 1), except for the large yolk granules and extreme flat shape in the remaining cells, the tine structure of embryo at ... :,,, ./ ... .-...., germ disk stage in A. japonica was similar to that in lycosid spiders belonging to agelenid type. FIG. IS. Schematic figures of embryos at germ disk- In ,V, nautica, the yolk packages were exposed stage in A, japonica (left) and N. nautica (right). In directly to the perivitelline space (Fig. 18). both spiders, the germ disk is composed of spherical cells, which have almost no large yolk granules. The interior of the embryos is filled with yolk packages LITERATURE CITED (yp). In A. japonica, there are very flat cells which possess seveial large yolk granules in the surface HOLM, A. 1952. Experimentelle Unlersuchungcn fiber region where the germ disk is not formed. In fV Bntwicklung uod Entwickluogsphysiologic nautica. any cells are not found in that region, so the dL-sS;!;nnciicmbryus. ZoologiskaBidrag Iran Op ypUt packages are exposed directly to perivitelline psalo 29: 293- 424.

space (ps). vm; vitelline membrane. KONLXX A. 1 %V. The fine structure of" the early spider cmbrvo. The Science Reports of the Tokyo Kyoiku Daigaku. Section B 207: 47-67. nucleus. This phenomenon was reported in L i .i7fi Morphological study on the spider's lycosid spiders (Kondo, 1969), In N. nautica, embryonic cells at germ disk stage. Japanese many lysosome-like bodies, described also in Journal ofDcvelopmentel Biology 24: 20-21. (in lycosid spiders (Kondo, 1969), were observed, Japanese) however histocbemica] studies arc needed for KONDO. A. & YAMAMOTO, N. 1975. Comparative final identification. morphology on the early spider embryos. Atypus 63: 13. (in Japanese) In both spiders, A. japonica and N. nautica, MONTGOMERY, T. H. 1909. The development of interdigilarions were not observed, bill they were Thcridium, an Aranead, up to the stage of rever- reported in disk region of lycosid spiders germ lurri. Journal of Morphology 20: 297-352. I Kondo, 1970). SEKJGUCHL, K. 1957. Reduplication in spider eggs produced by centnfugation. The Science Reports Jn the inner part of the embryo in both spiders, o\ the Tokyo Kvoiku Daigaku Section B 8: 227- several large yolk granules were packed by cell 230 membrane with various organelles and glycogen ,' SUZUKI H. & KONDO, A. 1 99 1 . Fine structure of the granules. These structures were described as yolk germ disk in the theridiid spider, Achaearanea spheres in lycosid spiders (Kondo, 1969) Since tepiditrionem (C. Koch). Proceedings of nucleus was not observed in them, these packages AruVopodan Embryologica! Society of Japan 26: of large yolk granules were distinguished from 11-12.

MG. 9-17. 9-1 1. A. japonica 9. Extreme Hat shape in peripheral part of remaining cells. A dcsmosoinc-like structure (Arrowhead) is found between cells. (lu.m). 10. A large yolk granule included in remaining cell.

Arranging vesicles along yolk granule. (0.5 u.m). 1 1 . Peripheral part of yolk package in inner part of embryo. Large yolk granules, a ring-shaped mitochondrion (m), vesicles (v). and glycogen granules (arrowhead) are

packed by cell membrane. (1 u.m).

12 17. rV. nautica 12. A lysosome-like body including amorphous matrix. A ring-shaped mitochondrion (m)is found. (IjAnt) 13. A lysosome-like body including many small vesicles (lu-m). 14. A lysosome-like body including several double membranes or myelin-like structure. (lp,m). 15. Mitochondria around nucleus In)

showing nuclear pore (arrowhead) and nuclear envelope i ne)( 1 jim). 1 6. A Golg) body. (0.5^m). 1 7. Superficial region where germ disk is not formed. A yolk package composed of large yolk granules (yg), a cup-shaped mitochondrion (m), glycogen granules (arrowhead), and vesicles are exposed directly to perivitelline space (ps). (l(im). Abbreviations: fg: fatly granule, fy: fine yolk granule, gg; glycogen granules; ps: perivitelline Space, yg; yolk granule vm: vitelline membrane. Scale line in parentheses.

AN EXPERIMENT ON COLONIZATION OF KARAKURT (LATRODECTUSTREDECIMGUTTATUS, BLACK WIDOW SPIDER) ON ISLAND TERRITORIES IN KAZAKHSTAN

CHINGIS K. TARABAEV

Tarabaev, C.K. 1993 11 11: An experiment on colonization of karakurt (Latrodectus tredecimguttatus, Black Widow spider) on island territories in Kazakhstan. Memoirs of the Queensland Museum 33(2): 651-652. Brisbane. ISSN 0079-8835.

To create an artificial, controllable population of Latrodectus tredecimguttatus (karakurt) with the aim of collecting venom, an experiment on mass colonisation of southern population spiders on an island territory was carried out. Retardation of the overwintering stage under laboratory conditions ensured the availability of large numbers of karakurt for colonisation and eliminated its uncontrollable reproduction in neighbouring territories. Pour creer une population artificiele bien controlee des Latrodectus tredecimguttatus (karakurt) dans Ie but d'obtenir du venin on a mis une experience de la dissemination des Araignees de la population m£ridionale sur le terriloire insulairs. Le delai du stade hivernant dans les conditions de laboratoire assure la stabilisation du nombre de masse de karakurt pour leur colonisation et elimine leur reproduction non-controlee sur les lerritoires adjacents. \Z\Colonisation, karakurt, egg sacs, Latrodectus,

Chingis K. Tarabaev, Institute of Zoology of Kazakhstan National Academy of Sciences, Akademgorodok, 480032 Alma-Ata, Kazakhstan Republic; 19 March, 1993.

In Kazakhstan, mass annual collections of a numerous, yet controllable karakurt population. venom-producing arachnids are carried out to To do so we retarded the development of spiderl- obtain their venom for medicinal and other pur- ings and then released the 2nd instars over the poses. Black widow spiders ("karakurt*, island. As a result, the spiderlings avoid the dis- Latrodectus tredecimguttatus (Rossi)) are caught astrous late frosts but the postembryos of the new in the greatest numbers. At the same time, their generation in their egg sacs would not have time abundance varies from year to year: quite often to develop into overwintering 1st instar spiderl- periods occur during which one can hardly find a ings. This phenomenon is therefore the necessary single specimen (see Marikovskij, 1956; Levi, condition for the possibility of creating a control- 1983; Tarabajev, 1990). A bite from a karakurt is lable karakurt population, as well as for elimina- an appreciable danger. Hence, during periods of tion of uncontrollable mass reproduction of great abundance, it must be controlled. In connec- spiders on the neighbouring territories. For the tion with this problem, we carried out an experi- intensification of degree-days deficit effect, the ment on karakurt colonisation on an island with spiderlings from 500 egg sacs of Latrodectus the aim of creating an abundant and controllable tredecimguttatus collected in Uzbekistan artificial population. (southern population from Dzhizak Steppe) were used in our experiment. Before the colonisation MATERIALS AND METHODS we made a census of the native population of karakurt on the island.

The experiment was performed on the small During winter, egg sacs of southern population island 'Malyj' (1.4km) in Alakol' Lake spiders with overwintering first instar spiderlings (46°08*N, 81°52'E) near the northern border of were kept in the laboratory (temperature 0-5°C). the known karakurt distribution: the 2nd instar In the second half of May these were placed into spiderlings emerging from egg sacs are often a gauze-covered 20 litre vessel at room tempera- affected here by the late frosts which occur in ture (18-22°C), for their reactivation from winter April-May (Tarabajev, 1990). In contrast, the diapause. After moulting in their egg sacs, many development of spiders is critically restricted by spiderlings emerged; then the vessel which con- placed in a refrigerator (4-5°C) the shortened warm season: if the postembryos tained them was within the egg sacs have no time to develop into until June. the overwintering 1st instar spiderlings, they die Before mass colonisation, a census of the during winter (Marikovskij, 1956). This natural karakurt population on 'Malyj' Island was phenomenon stipulated the possibility of creating carried out by the visual investigation of the 652 MEMOIRS OF THE QUEENSLAND MUSEUM

whole island territory fii for the settling by the In August-September, mass collection of southern population spiders of karakurt (ca. 8500 females must be carried out. These females are 2 2-3 m ). kept in collection boxes for weeks until they lay their egg sacs in these boxes. (This RESULTS phenomenon was first noticed by us when studying the technique of mass collecting from the While making a census of the native karakurt field). Females are subsequently used for obtain- population on 'MaJyj' Island before mass ing venom while egg sacs are kept at room colonisation on 12 June 1988, we found one nesi temperature until the 1st instar spiderlings from the previous year with two empty egg sacs, emerge (overwintering stage). After that the egg and two more old nests. Six living karakurt sacs must be kept in a refrigerator at 0-5°C until specimens of 4-5th instar were also found- By late the following season. August there were 3726 nests from the southern After tbe reactivation of spiderlings in spring, population, or one specimen per 2-3in * Three they are released over the island, which they females of the native karakurt population were recolonised effectively. Every August-Septem- also found: they differed by their larger si/e (no ber, mass collection of females is carried out, and females were measured). In 28 nests of southern the cycle is renewed. population spiders there was only one egg sac per Thus, the indubitable advantage of our method nest: no egg sacs were found in the rest, while in is the elimination of uncontrollable mass three nests of native population spiders there reproduction of karakurt and the absence of any were two egg sacs in each. Dissection of egg Sacs necessity of special egg sacs collecting for confirmed our views. In four egg sacs of the colonisation. native karakurt population there were postembryos, in two there were first instar spiderl- LITERATURE CITED ings, while only eggs were found in 20 egg sacs of the southern population. LEVI, H.W. 1983. On the value of genitalia structures in 1989, 74 nests Of 100 nests examined May and coloration in separating species of widow were without egg sacs-some ncsls were ruined; spiders {Lalrodectus sp.) (Arachnida: Araneae: 26 nests each had one egg sac, all eggs were dead. Theridiidae). Verhandlungen des Naturwis- These results confirmed that due to the artificially senschaflliehen Vereins zu Hamburg (NF) 26: retarded development of southern population 195-200. spiders in the northern conditions of AlakoF Lake MARIKOVSKU, P.I. 1956. [Tarantula and karakurt: the eggs had died within the egg sacs as they had morphology, biology, toxicity]. (Kirghiz SSR of Sciences Publishers: Frunze.), [in insufficient time to develop to the overwintering Academy Russian] first instar spiderlings (the average Alakol' area TARABAJEV, CH. 1990. Winter Hosts and late frosts temperature in September is no more than 10- as the reason of karakurt (Black Widow Spider, I5°Q. Latwdevtus trtdecimguaatus) depression in therefore propose the following scheme for We Kazakhstan. Bulletin de la Societe Europeenne the creation of many controllable artificial black d Araehnologie (Comptcs Rendusdu XII Col- widow populations for the purpose of obtaining loque Europeen d'ArachnoIogie. Paris (France). venom. 2-4 juillel 1990), No. hors serie 1: 346- 348. DISTRIBUTION OF LATRODECTUS (THERIDIIDAE), ERESUS AND STEGODYPHUS (ERESIDAE) IN KAZAKHSTAN AND CENTRAL ASIA

CHINGIS K. TARABAEV, ALEXEY A. ZYUZIN and ANDREY A. FYODOROV

Tarabaev, C.K., Zyuzin, A.A. and Fyodorov, A.A. 1993 11 11: Distribution of Latrodectus (Therid'iidae), Eresus and Stegodyphus (Eresidae) in Kazakhstan and Central Asia. Memoirs of the Queensland Museum 33(2): 653-657. Brisbane. ISSN 0079-8835.

The distributions of three Latrodectus species (L tredecimguttatus, L dahli and L pallidas), and of three eresid species (Eresus niger, E. tristis and Stegodyphus lineatus) within the Kazakhstan-Central Asian region are analysed based on the literature and on original data. Latrodectus tredecimguttatus presumably occupies almost the entire Kazakhstan territory, while two other widow species are local and more southern: both of them are First recorded here within Kazakhstan. Preliminary morphological and mating analyses show that both 'Latrodectus tredecimguttatus' and 'Eresus niger within the territory of Kazakhstan and Central Asia are composite species: the first consists of at least one species different from the European L tredecimguttatus, and the second consists of at least three separate species. On a analyse* la distribution de trois especes du genre Latrodectus (L. tredecimguttatus, L. dahli et L pallidus) et de trois especes de la famille Eresidae (Eresus niger, E. tristis et Stegodyphus lineatus) dans la region du Kazakhstan et de V Asie centrale d'apr&s le donnees litteraires et originales. L'aire d'habitation de Latrodectus tredecimguttatus comprend, hypotheriquement, presque tout le territorie du Kazakhstan, tandis que deux autres especes sont locales et plus meridionales: Tune et 1* autre sont mentionnees ici pour la premiere fois pour Kazakhstan. L*analyse morphologique pr^alable et les experiences d'accouplement montrent que Latrodectus tredecimguttatus aussi bien que Eresus niger au Kazakhstan et a

1' Asie Centrale sont des especes collectives: dont la premiere se compose, au moins, d'une

espece differente de L tredecimguttatus d'Europe et I 'autre se compose, au moins. des trois especes s6paTies.[jLatrodectus, Eresus, Stegodyphus, distribution.

Chingis K. Tarabaev, Alexey A. Zyuzin, Andrey A. Fyodorov, Institute of Zoology, Kazakhstan National Academy ofSciences, Akademgorodok, 480032 Alma-Ata, Kazakhstan Republic; 19 March, 1993.

The genus Latrodectus (Theridiidae) and the tredecimguttatus and E. niger, must be family Eresidae have not been studied very much revised, in Central Asia. Before 1950 only one Latrodec- tus species was known from the former territory MATERIALS AND METHODS of the U.S.S.R -L. tredecimguttatus (Rossi), or 'karakurt'. Two years later, Spassky (1952) first reported L pallidus O. Pickard-Cambridge from This work is based on material collected mainly the western, desert regions of the Turanian by us in Kazakhstan and Central Asia. Our spec- zoogeographic province ^Turkmenistan ter- imens were compared with those from Europe ritory]. Later Charitonov (1954) described the and North Africa. Spiders were examined in 70% new subspecies L. pallidus pavlovskii from Turk- alcohol using binocular microscopes MBS-1 and menistan (the so-called 'white karakurt'). Twen- MBS-10. Preliminary experiments on mating be- ty years later Tystshenko and Ergashev (1974) tween European L. tredecimguttatus and Widows found in Uzbekistan another black widow from Kazakhstan were also carried out. species, dahli Levi. the Eresidae, L Amongst Abbreviations: BIN, Biological Institute, Nov- three species have previously been recorded: osibirsk , Russia; IZAj institute of Zoology, Eresus niger (Petagna), E. tristis Kroneberg and Alma_ Ata; MCZ, Museum of Comparative Zool- (LatmUe). Stegodyphus lineatus ogV) Cambridge, Massachusetts, USA; MNHN, This paper deals with new collections and Museum National d'Histoire Naturelle, Paris, data relating to the two groups and species France; ZISP, Zoological Institute, St. Peters- distributions. Also, we suggest that previous burg; ZMMU, Zoological Museum of the Mos- broad species concepts, especially within L. cow University. )

654 MEMOIRS OF THE QUEENSLAND MUSEUM

.-.,

FIG. I: Distribution of Unrodectux spp. in Kazakhstan and other Central Asian Republics. Legend; 0= L tredecimguttatus\ D= L dahli; - L. pallidas pavlovskii; oblique hatching s distribution oTL tredcarnguttatus after Marikovskij (1956); cross hatching = suggested distribution of L pallidum within the distribution of/.. tredt'cimxuiiaius. Black figures = published dala; white figures = original data.

RESULTS AND DISCUSSION ubris (Dufour), etc. (Chantonov. 1932). Rossi- kov (1904) devoted a monograph to this species Family Theridiidae and Marikovskij (1956) analysed the biology and distribution of L. tredecimguttatus within the Latrodectus Walckenaer former USSR- According to this author karakurts are found over almost all of Kazakhstan (Fig. 1). Remarks However, existing difficulties in the systematica 1 The northern border of the widow spiders dis- of Latrodectus species (see Levi, 1983), as a rule, tribution within the former USSR seems to pass result in too wide an interpretation of their dis- near 52°N (Fig. I). tribution. When investigating Latrodectus species, we found that adult females of L Latrodectus tredecimguttatus (Rossi tredecimguttatus from Italy had a light spotted 'Karakurt' abdomen while Kazakhstan specimens were completely black. DrG. Schmidt (pers. comm.) Rkmakkn considers the karakurt from Kazakhstan to be the Within the former USSR this species has been species Latrodectus lugubhs (Dufour) described known under the names L conglobaius C.L. from Egypt. Preliminary experiments on mating Koch. L crebus Savigny and Audouin, L. lug- carried out in 1991 together with Mr D. Weick- ubris Motchoulsky, L tredecimguttatus vur. lug- mann-Zwoerner (Germany) showed that ) . 1 ,

LATRODECTUS. ERESUS ANDSTEGODYPHUS IN KAZAKHSTAN 653

Kazakhstan specimens could not cross with our spiders with the type material or topolypcs ol European L. tredecimgauatas [two males ftora L pallidas is necessary. Kazakhstan were used). At present the distribution of L iredecimgu.1- Material Examined tutus as delimited by Marikov&kij 1.1956) must be Turkmenistan: 1*2, Tashauz Area, Shakh- ed. as we now suppose- the traditional L senem, under Artemisia, 9 Oct 1983, O.S. tredecimguttatus" in Kazakhstan *o be a separate Soyunov. Uzbekistan: 4 x, Dzhizak A l species. Kyzylkum' state farm, 16 Jul 1982, N.E. Br- gashev. Kazakhstan 19, South Kazakhstan Material Examined Area, Kyzylkum Desert near Tabakbulak VilL '1991 Italy La?.io, nearPriverno. 16 May 1962, 24 Aug , A.A. Zyuzin, B.M, Gubin 1

fields and stones, H. Levi (MCZ). Kazakhstan 9 of Tabakbulak population in laboratory) (IZA I and Central Asia, many males and females from different localities (IZA). Family Erksilvm

Latrodeetus dahli Levi Eresus Walckenaer

REMARKS Eresus niger (Pctagna) This very localised species among the Cential A^jan republics occurs in Uzbekistan only Remarks 1990,): ^henko&Brgashcv, 1974;Frgashcv T Published and original data on the distribution in Kazakhstan, it was found by us first in the of this species within the Sibero-Kazakhstan- Kyzylkum Desert (two females, det. Dr Y.M. Central Asian region, in the European part of Marusik)(Fig. 1). We have compared our Russii haritonov, 1932) and in Europe (sec specimens with the female paratype ofL, dahti Bonnet, 1956) show ihat the northern limit of its from Iran and found their resemblance in the area seems to pass near 56°N: thus, E. niger is hairiness of the abdomen's dorsum (long slender theoretically distributed over the whole spines and rather long setae between them: sec Kazakhstan and the Central Asian region also Tystshenko and Ergashev, 1974, fig. 3) Preliminary analysis of material we have at our Nevertheless, the epigynal opening in all our disposal show that at least three separate species specimens is 4 limes as wide as long, while in the of the 'Eresus niger* complex occur in female paratype it is only 3.4 times as wide as Kazakhstan (Fig 2). The acute deficit uf long; slight differences are also in the vulvae (ci\ specimens of both se\es taken from the same Levi, 1959, figs 11, 12; Tystshenko and Er- place is the main obstacle to a detailed taxonomic gashev, 1974, figs 4, 5). As the male of L. dahli study. tmm the type locality is up to now undescribed. According to Merrett and Millidge (1992), the wc cannot be sure that our specimens belong to correct name for the species Eresus niger (Ptlng- real L. dahli. na) is Eresus cinnaberinus (Olivier) (sec also Lchiinen, 1967, p. 233). Mavlrial Examined Iran: 1 paratype V, Bushire, Persian Gulf Material Examjni

Uzbekistan: 1 I (MCZ). d y ?,KashkadarjaArca, France: 2<5.2?2 juv.. Col du Ccris, 15 Sep Kazakhstan: Karshi Steppe. N.E. Rrgashev 2 9 1908; BanvuK 3>) May 1909 (MNHN, No. AK SoulJl Kazakhstan Area, Kvzylkum Desert, 77.5k S3S). Spain: 3 6. 4 9. La Granja, Jun I90X of Chardara Vill., 5-6 Jun 1989, Tarabaev. NW (MNHN, No. AR 837). Mongolia, i rj, 'Potanin, FyodbKJY, Zyuzin (IZA Schenkeldet. 1946* (MNHN, No. AR 852). Him

gary: 1 6 .Csakbereny, Vertcs.20Sep 1991. V.V. Latrodeetus pallidus pavlovskii Chantonov Dubaiolov, V.G. Mordkovitch. Russia: 3 c£, J) (Fig. Bashkir Reserve. Bashart i.V Stebaev; 1 ej,

Novosibirsk Area, 1 3k W of Karasuk Vill., 7 Sep Remarks 1989; 5 d\ ibidem, 27k SE of Zdvinsk Vill

Within Kazakhstan, it was found by us first in Malyje Chany Lake, 10 Sep 1989, V.P. Pekin.

the Kyzylkum desert, which is probably its* north- Turkmenistan: I c?, Kopetdag Reserve, 15 Maj ernmost limit. To clarify the taxonomic position 1988, Karpenko;2 tf, 25 Aug 1988. ibidem, !5k

oft pallidas pavlovskii, thorough comparison of W.if Firyu/a Vill. Mount Diishak, 2100 m; I 656 MEMOIRS OF THE QUEENSLAND MUSEUM

u 40 .

l FIG. 2: Distribution of eresid spiders iii Kazakhstan and other Central Asian Republics. Legend: 0= Eresus niger complex: I, 2, 3* = E.

Dubatolov. V.K. Zinchenko; 1

Males of E. tristis con be readily separated from (ZMMU No. Ta 1 104); 1 rf, Alma-Ata Area, 5k those of E. niger by the black colour of their NE of Kapchagaj City, A.A. Fyodorov (IZA). :

LATRODECTUS. ERESUS, AND STEGODYPHUS IN KAZAKHSTAN 657

Turkemistan: 3 cf, Western Kopetdag Ridge, spiders in Uzbekistan']. ('Fan* Publishers: Tash- near Kara-Kala Vifl., 7 Feb 1979, L Morozova kent), [in Russian] KRONEBERG, A. 1875. Arancae. Jn: Fcdtschcnko, (ZISP). Tadzhikistan: 1 <3, 'turn to Chashma, f A. P., 'Reise in Turkestan, Zoologischer Their, 9.V.1986,Itka (BIN). vol.2; 1-58. Moscow. LEHTINEN, P.T. 1967. Classification of the cribcllate Stegodyphus Simon spiders and some allied families, with notes on the evolution of the suborder Araneomorpha. Annates (Latreille) Stegodyphus lineatus Zoologici Fennici 4: 199-430. Kroneberg) i=Eresus arenahus LEV], H.W. 1959. The spider genus Lairodecms (Araneae, Theridiidac). Transactions of the Remarks American Microscopical Society 28: 7-43, This species was previously known from 1983. On the value of gcnitalic structures and Turkmenistan, Uzbekistan and Southern coloration in separating species of widow spiders Kazakhstan. We have found it considerably {Larroilectus sp.) (Arachnida: A ran eae northwards and eastwards (Fig. 2). Theridiidac). Verhandlungcn des Naturwis- senschafltichcs Vcreins zu Hamburg, (NF) 26: Material Examined 195-200, and karakurt: Many males and females from: Spain; Italy MARIKOVSKIJ. P.l. 1956. ['Tarantula morphology, biology, toxicity']. (Kirghiz SSR (Sicily); Algeria; Tunisia (Kairouan) (MNHN Academy of Sciences Publishers; Frunze), [in No. AR 785, 786); different parts of Kazakhstan Russian] and Central Asian republics (sec Fig. 2) (IZA). MERRETT. P.&MILL1DGE, A.F. 1992. Amendments lo the cheek list of British spiders. Bulletin of (he ACKNOWLEDGEMENTS British Arachnological Society 9: 4-9. NENILIN, A.B. & PESTOVA, M.V. 1986. [The spiders at our dis- For placing the necessary material of the family Eresidae in the USSR fauna], posal and for assistance in experiments we are Zoologichcskij Zhurnal 65: 1734-1736. [in Rus- grateful to: Prof. H.W. Levi (Cambridge, Mas- sian] sachusetts, USA). Drs D V. Logunov, V.V. ROSS1KOV, K.N, 1904. [The venomous spider Dubatolov (Novosibirsk), K.G. Mikhailov (Mos- karakurt], Agricultural monograph. Trudy Byuro cow), V.T. Ovtsharenko (St. Petersburg) (all Rus- po Entomologii 5 (2) 1-232. St. Petersburg. |in sia), Ch. Rollard (Paris, France), and Mr D. Russian] Weiekmann-Zwoerner (Weissenburg. Ger- SIMON, E. 1895. Arachmdcs recueillis par Mr G. many). Potanine en Chine et en Mongolie (1876-1879), Bulletin de PAcadcmte ImpenaledesScieri- LITERATURE CITED St. Petersburg, [5)2 (4): 331-345. SPASSKY, S.A. 1952. |The spiders of the Turanian zoogeographic province]. Eniomologicheskoje BONNET, P. 1956. 'Bibliographia Araneoninv, vol OhKi/j-enije 32: 192-205. [in Russian] 2(2): 919-1026. (Douladoure: Toulouse). V.N. 1937. CHARITONOV, D.E. 1932. Katalog der Russisehen SPASSKY, S.A. & SHN1TNIKOV, Spinnen. Annuaire du Musce Zoologique, [Materials on the spider fauna of Kazakhstan]. 265- Leningrad 32 (Beilage): 1-206. Trudy Kazakhstanskogo Filiala AN SSSR 2: 1954. |Thc new representative of the genus 300, [in Russian] Latrodectua from Turkmenia (Lairodccfus pal- TYSTSHENKO. V.P. & ERGASHEV. N. 1974. (Aranei Theridiidae), a lidas O.P. Cambr, subsp. pavlovskii n )). \Luirodeaus dahtt Levi T Zoologichcskij Zhurnal 33: 480-485. [in Rus- species of venomous spiders new in the USSR sian] fauna]. Entomologicheskoje Obozrenijc 53: 933- ERGASHEV, N.E. 1990. Ideology of the venomous 937. [in Russian].

GEOGRAPHIC VARIATION OF THE NUMBER OF B-CHROMOSOMES IN METAGAGRELLA TENUfPES (OPILIONES, PHALANGIJDAE, GAGRELUNAE)

N. TSURUSAKI

Tsuiusaki, N. 1993 II 11: Geographic variation of the number of B-chromosomes in Metagagrella teitutpes (Opiliones, Phalangiidae, Gagreihnae). Memoirs of the Queensland Museum 33(2); 659-665. Brisbane. ISSN 0079-8835.

Chromosomes of Metagagrella letwipes (L. Koch) (Arachnida, Opiliones. Phalanyndac. Gagrellmae) were surveyed in 8 populations in japan. Almost every individual examined had one or more (highest number per eel! was 19) supernumerary or B-chromosomcs, in addition to the basic set of chromosomes (2n=18). These B-chromosomes are heterochromalic. and during meiosis they appear to behave as univalents. The number oj B-chromosomes varied both among and within populations. The number aiso fluctuated to some extent among cells from the same individual, suggesting nondisjunction at mitotic anaphases. No correlation could be elucidated between the number of B -chromosomes and externa] morphologies or habitat type. The number of B-chromosomes may affect growth rale, and in turn reduce the synchrony of breeding within a population. Die Chromosomen von Metagagrella tenu'tpes (L. Koch) (Arachnida, Opiliones, Phol.in- giidae, Gagrellinae) wurden in 8 japanischen Populationen untersueht. Fast jedes ubcrpruftc

Individuum wies eines oder mehrere (die hochste Anzahl pro Telle war 1 9) iiherztfhJigc oder B-Chromosomen zusatzlich zum Standardchromosomensat? (2n = IS) auf Dicsc B- ChromosomensindheierochromatischundschcmensichwahrcndderMcioscalsl'nivalenle zu vcrhallen. Die Zahl derB-Chromosomcn variiert sowohl zwischen als auch innerhalb von Populationen. Sie schwankt in gewissem AusmaB auch zwischen Zellen aus ein und demselben Individuum, was auf Non-Disjunction bei milotischen Anaphases schlieBenlaBt. Zwischen dcr Anzahl der B-Chroniosomcn und dcr autk?ren Morphologie oder Habitattypen konntcn kcinc Korrctationen gefunden werden. Es bestchl die Moglichkeil, daB die Zahl der B-Chromosomen die Wachsturnsrate bcfinHusseTi und in der Folge die Synchronic der Fortpflanzung innerhalb cincr Population reduzieren l&nnJQOpiliones, Metagagrella ienuipes, Japan, B-chromosomes, geographic variation.

Nolnu> Isurusaki, Department nf Hiol<>g\, F&ttifoy of Education, Tottori Universitv, Tvtlon 680, Japan; 29 October. 1992.

The harvestman Metagagrella wnuipes (L. MATERIALS AND METHODS Koch) is widespread throughout Japan but with a peculiar habitat and distribution pattern, in specimens used for southern Japan, this opilionid is typically coastal. The chromosome examina-

tion Hsted (Table 1 Fig 1 .. Cytologic^ data Howcver, in northern Japan, it also occurs inland &* were obta,ned preparations r, and prefers open habitats, e.g. parks and gardens. ^»ft tes or ovaries of field-collected adults and penul thatareatfectedbymoderatehumandisturbance.>u / er * ju ^ . u — r 4 u fi Jhe lechni used is dtsCTl ^ 6 in this Dunni5 a chromosomal survey of spee.es, Tsurusakl ( , 9g5) and>surusak ; and Cokcndol- the number of chromosomes varied both within pher n99o) rn somc ce]ls pattcms simiiar |0 and among populations, with a range of 2n=18 to C-bands were observed despite the fact that these 36. Moreover, the number of chromosomes cells had received no special treatment. In many

varied somewhat from one cell to another in individualsthe number of B-chromosomes van. .1 almost all individuals. Close examination of from cell to cell. In such cases, the chromosome mitotic and mciotic chromosomes revealed that number of each is represented by a modal mini the karyotype of this species is usually composed bcr. of2n=I 8 standard chromosomes and one or more For two populations (Nakajima and supernumerary or B-chromosomes, and that the Maruyama), lengths of femur I of the specimens latter cause the overall chromosome number to used in the chromosome preparation wcte vary- Here. I will describe the karyotype, nature measured with an eyepiece graticule. of the B-chromosomes, and pattern of geographic Detailed collection data on the specimens c.v variation in the B-chromosomc nurr.bci amined are listed in the appendix. 660 MEMOIRS OF THE QUEENSLAND MUSEUM

Campus of Hokkaido Univ. Wakasakanai

Sunagawa

Botanical Garden

6<7M* : 3-10(6.2)

Nakajima Shirahama

: 0-3 26*4* (17) 14*6? : 4-11 (6.6) %> "*« Awa-Amatsu '• 13*: 1-6(3.5)

FIG. I . Map of populations of M. temtipes, Japan, used here (double open circles) with records compiled from

literature (e.g. Suzuki, 1 973; Suzuki and Tsunisaki, 1983) and data in the appendix (solid circles). No. of samples and range of B-chromosome number in each study population are shown, with means in parentheses.

RESULTS osomes alone, was determined to be 2n=18 for both males and females (Figs 2-3). Karyotypes and Nature of B-Chromosomi-s /. Standard karyotype (Fig. 3) The chromosome number of this species varied Autosomes consist of 8 pairs of medium-sized enormously among and within populations- From metacentrics (Nos. 4 and 5), submetacentrics comparisons of the karyotypes of individuals (Nos. 1, 3, 7, 8), and subtelocentrics (Nos. 2 and with the lowest chromosome number, 2n=18, to 6). The X chromosome, the largest, is sub- those from others with 2n=I9 and greater, and telocentric, while Y is submetacentric and similar from analyses of both meiotic chromosomes and in size to chromosome No. 1. In C-banded mitotic unintentionally obtained C-banded chromosomal metaphases, centromeric regions of the A- spreads, the chromosome number of the standard chromosomes were positively stained (Figs 4-5). karyotype, which consists of so-called A-chrom- No prominent differences were found among the

' TABLE 1 . B-chromosome numbers in 8 populations of Metagagrella temupes. notes: S = seashore habitats, 2 1 = inland habitats e.g. parks, fields; Range, mean and mode not for all cells counted but for values represented 3 4 by mode of each individual; Calculated only for samples with>5 individuals; One or more were juveniles.

7 Locality and habitat type in parentheses Date Specimens examined Number of B-chromosomes 3 min. max. mean mode SD 4 Wakasakanai (S) 9-1X-1985 Id 6 6 6 6

Sunagawa (I) 8&15-X-I986 3 6 2 6 4.7 6 -

- Campus of Hokkaido Univ. (I) 18-1X1981 1 6 IS IK 18 18

Botanical Garden. Sapporo 1982 (1) 5-IX-198 5 6,1 9 4 10 6.2 4 2.40

as above, 1986(1) R-V1I1&9-IX-1986 8o\29 3 9 6.2 5, 7. 8 1.93

Moray ama, Sapporo (I) 5-1X1986 204.3 9 3 15 7.0 6 2.75

Awa-Amalsu(S) 27-VIIM984 13 d 1 6 3.5 4 1.39 9"* Shirahama (S) 29-VH-1983 14 4.6 4 11 6.6 6 l 60

Nakajima (S) 26-VII1 I9S6 26rf,4 9 3 1.7 1,2 I 02 B-CHROMOSOMES IN METAGAGRELLA TENU1PES 6ftl

H II II If !••• * .. I

A __*** • lift111 l mbl • i^lOS'S

14 •* Ik If Xt ft* *« M fW It***********.*** A «*. 2rM8*l9b<

FIG. 3. Karyotypes of male Mtfagagrtltti fewipef FIG. 2. Representative chromosome complements at based on ihe photographs in Fig. 2. A. Nakajima; B. spermatogonia! mitotic ractflptoase of male Maruyama; C. campus of Hokkaido Univ. B*

Mciagagrelhi \enuipes. A, Nakajima. 2n=l 8 ( 1 SA*s}; chromosomes are arranged on the second row in B B, Maruvama, 2n=28 (ISA's + 10BV); C campus of andC. Scale = 5u,m. Hokkaido Univ., 2n=37 (ISA's + I9BY). Scale =

5 pirn. latitude (Spearmann's coefficient ofrankcorre I j lion, rs=0Ll9, n=8, P>0.05) and habitat types standard karyotypes of individuals from various (seashore or inland: Mann-Whitney U-tcst, populations. P>O.0.U 2. B-chramosumes In addition to a set of standard chromosomes DISCUSSION described above, almost all chromosomal spreads contained at least one B -chromosome. These were meta- or submetacentric, equal to or smaller Characteristics of the B-Chromosomes than the shortest pairs (No. 8) of the A -comple- According to Jones and Rees (19S2), B- ment (Fig. 3). In C-banded chromosome spreads, chroinosomcs have been reported in over 1000 the B's were heteropyenottc. darkly staining species of plants and more than 260 species of along their total lengths {Figs 4-5). During animals. In Arachnida, however, only \l meiosis, the B's remained univalent (Fig. 4-5) species of Acari have been shown to have B\; even when the cell earned 2 or more. The number Apanomma ftmhnartwi and two specie- of B's varied considerably from one cell to the Hai'mtiphvsalts (Oliver and Bremner, 19 other within a single individual, possibly due to Oliver era/., 1974). at during mitosis. nondisjunction anaphase The B-chromosomes in Metagagrella tcnuipes Numeric variation in B's among individuals have the following characteristics thai arc typical within populations was also evident (Fig. 6). No for B's recorded in various other organisms significant differences were found in the number (White. 1973: Jones and Rees. 1982; Wcrrcn et of B's between the sexes from any population uU 1988, Shaw and Hewitt, 1990; Jones. 1991): (Mann-Whitney U -tests for each of four popula- (i) they are smaller than most members of the sexes were sampled, tions where both P=0.48- A-complemcni; (2) they appear to be comprised 0.96). of a large amount of constitutive heterochromatin, (3) they remain univalent Geographic Variation in Number op B\s during meiosis; (4) the number of B's vanes from The number of B-chtomosomes varied sig- one cell to another even within an individual, nificantly among populations (single classifica- indicating they display nondisjunction at anaphases of spermatogonia! mitoses. tion ANOVA. F=34.6: d.f =7: P<0.00J) (Figs 1 and 6, Table I). The lowest and the highest However, the frequency and the number of B's population means of the number of B's were 1.7 were rather unusual. In this species, every popula- for Nakajima (n=30) and 18 for the campus of tion sampled contained B's and the frequency

Hokkaido University (n=l ), respectively, and the within a population was up to 87% in Nakajima means of the other populations lay between the and 100% in the other 7 populations. The number two extremes, range = 3.5-7. However, no sig- of B's retained per individual of this species was nificant correlations could be detected between also extraordinarily high; the population means the number of B's and characteristics such as had a range from 3.5 to 7 in 6 out of 8 populations "

662 MEMOIRS OF THE QUEENSLAND MUSEUM X + I v'

* I

* V # f **A

^P ^ — : *5* jt K

#• " * £ O €

B B FIG. 4. C-banded mitotic metaphase (A) and meiotic metaphase I (B) in 6 M. tenuipes from Sunagawa, FIG. 5. C-banded mitotic metaphase (A) and diakinesis with 2 B-chromosomes (arrowed). Scale = 5|xm. - metaphase I (B) in 6 M. tenuipes from Maruyama,

(Table 1 ). The higher end was found in the single with 6 B-chromosomes (arrowed). Note: 9 bivalents male sampled from the campus of Hokkaido formed by 9 pairs of A-complement of chromosomes. = University, whose modal number of 18 B's was Scale 5jxm. exhibited by 30 cells and the maximum number exophenotypic effect due to B's, some show a of 19 in 3 cells (Figs 2-3). These are among the relationship between the number of B- highest numbers in far of B's animal species so chromosomes and the rate of development recorded, close to the number 'about 20' in Xylota (Hewitt and East, 1978; Harvey and Hewitt, nemorum (Diptera: Syrphidae) (Boyes and Van 1979). Thus, there is a possibility that the Brink, 1967). presence of B-chromosomes retards the cell cycle due to the additional DNA or its organization Effect of B's on Phenotype which is possibly different from A- Metagagrella tenuipes shows marked variation chromosomes. In turn, these effects may in- both among and within populations in external fluence the growth and development of the whole morphology, such as body size, leg lengths, de- organism (Jones and Rees, 1982). Such influen- gree of development of a spine on the dorsal ces might be related to an unusual feature of the scutum, number of noduli on the legs, and coloration of the body (Suzuki, 1973; Suzuki and life history of this species, namely, high Tsurusaki, 1983). However, no correlation was variability among individuals in the time to reach found between these characters and the number maturity. The duration of coexistence ofjuveniles of B-chromosomes (Fig. 7). These facts are con- and adults of this species at Maruyama is es- sistent with the observation that the B's are C- timated to be about three weeks, whereas it is less band positive. If these B's are indeed genetically than 1-2 weeks in other species of opilionids inactive, numerical variation of B's would lead to having no B-chromosomes, such as Oligolophus no effect on the phenotype of their owner. aspersus, Leiobunumjaponicum, and two species Although few studies have demonstrated any of Nelima from the same locality (Tsurusaki, :

B-CHROMOSOMES JN METAGAGRELLA TENUIPES (><^

9n 40 H # Nakajima Botanical Garden 1982 8- • Maruyama X=6.2

Boianical Garden 1 986 * * X=6.2

Maruyama20o, 3^ X=7

t. Awa-Amatsu 1 3 c? K-3.5

1 i i 1 i i 2 4 S 8 10 12 Number of B-chromosomes

FIG. 7. Relation of length of femur 1 to number of B-chromosumcs in 2 populations (Nakajima and Maruyama) of M. U'tiuipes. No significant correlation between 2 variables in each population [r = 0.305

(P>0.l) for Nakajima: r = 0.065 (P>0.1) for Maruyama), although both characters show prominent differences between two populations. Two

J 2 3 4 5 8 7 8 9 10 11 12 13 U 15 variables may be negatively correlated I. although Number nf B-chromosomes spuriously) between populations. Leg lengths decrease with increase in latitude (Tsurusaki, Uft- FIG. 6. Frequency distribution of B-chromosomc num- publ.) but the number of B's does not. bers in 6 samples from 5 populations. The other 3 populations (Wakasakanai, Sunagawa. and campus frequencies in the Botanical Garden at Sapporo of Hokkaido Univ.) omitted due to paucity of which were sampled in 1982 and 1986 (Pig. 6). If specimens surveyed. B-chromosomes were inherited in a non-Men- unpublished data based on weekly field-collec- dclian manner, and were neutral in phenotypic expression, not tions muxJc in 1979). The same phenomenon is the frequency of B\s would be also inferred from field-data from various stable. For example, it might be expected that the localities all over Japan. This versatility in die crossing of a male with 3 B's and a female with timing of final molting of this species might be 3 B's would produce some offspring with 4 or ascribed to the numerical variation of B- more B\s. However, no individuals with more chromosom.es. Further study is needed. than 3 B's arc found in Nakajima population. This indicates that some selection pressure limits the of that individual retain in Geographic Variation in Number of B*s number B's one can a It is still uncertain what The fact that the B's are found in every popula- particular population. factors the population and the tion over a wide geographic range of this species determine mean the B's. this species indicates that they arc of rather ancient origin. range of number of However, useful in various aspects of B's Marked morphological variation among the B- may be studying controversial chromosomes also attests to their long evolution- and 'selfish' DNA, including the issues on the level ot selection discussed by Wer- ary history. However, it is also possible that they aJ.. and Hewitt might be produced de novo from the A-comple- ren el (1988) and Shaw (1990). ment by recurring mutation. Although B's appear widespread across the ACKNOWLEDGEMENTS species range, the number of B\s considerably varied among populations, with a fairly wide 1 would like to thank Dr. R.G. Holmberg o\ range from 1.7 at Nakajima to 18 on the campus Athabasca University, Alberta, and two of Hokkaido University. In each population, the anonymous referees for useful comments and number of B's may be stable for at least several careful reviews that improved the manuscript. An years; there was no significant difference in their abstract in German was kindly prepared by Dr .1 664 MEMOIRS OF THE QUEENSLAND MUSEUM

Gruber of Naturhistorisches Museum Wien. Late of ticks. III. Chromosomes and sex determination Dr. T. Ito and the staff of the Seto Marine Biologi- in some Australian ticks (Ixodoidea). Annals of the Entomological Society America 837- cal Laboratory, Kyoto Univ. and Dr. S. Otsuka, of 61: 844. Hiroshima Univ., provided facilities for OLIVER, J.H., TANAKA, K. & SAWADA, M. 1974. chromosome preparation of the Shirahama Cytogenetics of ticks (Acari: Ixodidae) 14. population. The following persons kindly Chromosomes of nine species of Asian provided material used to map the range of Haemaphysalines. Chromosoma 45: 445-456. Metagagrella tenuipes: Drs. Sk. Yamane SHAW, M.W. & HEWITT, G.M. 1990. B- (Kagoshima), A. Otaka (Hirosaki), N. Yoshida chromosomes, selfish DNA and theoretical (Sapporo), Ms. T. Sato (Tokyo), and Messrs. T. models: where next? 197-223. In Futuyma, D.J. Kuwahara (Wakkanai), T. Tanabe (Tokushima), and Antonovics, J. (eds). 'Oxford Surveys in K. Ishii (Gumma), H. Okada (Himeji), Y. Evolutionary Biology, Vol. T. (Oxford Univer- sity Press: New York). Nishikawa (Osaka), N. Nunomura (Toyama), M. SUZUKI, S. 1973. Opiliones from the South-west Is- Yamashita (Iki). This work was partly supported lands, Japan. Journal of Science of the Hiroshima by Grants-in-Aid nos. 63740434, 02854100, and University, Ser. B, Div. 1 (Zoology) 24: 205-279. 03740393 from the Ministry of Education, SUZUKI, S. & TSURUSAKI, N. 1 983. Opilionid fauna Culture, Japan. Science and of Hokkaido and its adjacent areas. Journal of the Faculty of Science, Hokkaido University, Ser. VI, LITERATURE CITED Zoology 23: 195-243. TSURUSAKI, N. 1985. Taxonomic revision of the BOYES, J.W. & VAN BRINK, J.M. 1967. Leiobunum curvipalpe-group (Arachnida, Chromosomes of Syrphidae III. Karyotypes of Opiliones, Phalangiidae). I. hikocola-, hiasai-, some species in the tribes Milesiini and Myolep- kohyai-, and platypenis-subgroups. Journal of the tini. Chromosoma 22: 417-455. Faculty of Science, Hokkaido University, Ser. VI, HARVEY, A.W. & HEWITT, G.M. 1979. B- Zoology 24: 1-42. chromosomes slow development in a grasshop- TSURUSAKI, N. & COKENDOLPHER, J.C. 1990. per. Heredity 42: 397-401. Chromosomes of sixteen species of harvestmen HEWITT, G.M. & EAST, T.M. 1978. Effects of B (Arachnida, Opiliones, Caddidae and Phalan- chromosomes on development in grasshopper giidae). Journal of Arachnology 18: 151-166. embryos. Heredity 41: 347-356. WERREN, J.H., NUR, U., & WU, C.-J. 1988. Selfish JONES, R.N. 1991. B-chromosome drive. American genetic elements. Trends in Ecology and Evolu- Naturalist 137:430-442. tion 3: 297-302. JONES, R.N. & REES, H. 1982. 'B Chromosomes'. WHITE, M.J.D. 1973. 'Animal Cytology and Evolu- (Academic Press: London). tion, 3rd ed.\ (Cambridge University Press: OLIVER, J.H. & BREMNER, K.C. 1968. Cytogenetics Cambridge). B-CHROMOSOMES IN METAGAGRELLA TENUIPES 6^5

APPENDIX. New material of M. tenuipes used in Fig. 1 . Data in order: locality, altitude if available and needed, no. individuals, no. specimens dissected in parentheses if needed (Value may not bi in Table 1. since no countable chromosomal spreads were obtained for some specimens), date collected, collector (NT = N. Tsurusaki; YN=Y. Nishikawa).

hokkawo: l.Rebun.Kabukai, Kabukai Elementary School, 10m, 1 juv., 1 1-viii-I990,yn. l.Rishiri;

I Oshidomari. ca. 20m, I 5, 8-vm-1985 t Y. Kuwahara; Oshidomari, 5m, 9, 2 juv., 7-9-vm-1990 ;

S. Ueno. M. Sato, yn; Fujino, 10m, I juv., 7-vin-1985, nt. Wakkanai: Soya Point, lm, 2 3,39,1 juv., 2-1X-1988, nt; Wakkanai Port, 2m, 4 9, 5 juv., 7-vin-1985, S. Ishimaru, nt. Esashi-gun,

Hamatonbeisu-ch6,Bcniya-Genseikaen, 2 6*, 1 9„8juv., 14-15-vih-199G\yn, S. Ucno. Tcshio-gun;

Toyotonu-cho. Wakasakanai, on coastal sand dune: 1 juv. (I juv), ^-viii-1985, nt; 1

1 2 (2 SapporoV campus of Hokkaido Univ 1 Utashmai, 6 (1 o*). 8-X-19S6; 6 6\ 15-x-1986.nt\ ,

<5, I 9, 18-IX-1981, nt; Botanical Garden of Hokkaido Univ.: 9o.9? (5 tJ,] 9). 5-IX-I982, nt; 13 o\ 16 9,5-ix-1984; 7 juv. (7 juv.), 8-vn>1986; 3 3 (3 d ), 9-ix-1986, nt. Sapporo, Maruyama,

23 tJ. i9 9 (20 6 , 3 9 ), 5-ix-1986, nt. Tomakomai, Tomakomai Experiment Forest of Hokkaido

I Univ., 1 6, 3 9. 19-IX-1980. NT- vamagata pref : Obanazawa-shi, Obanazawa, & t 1 9,4juv,

2y-vm-1983, A. Otaka; 6 d, I 9, 4-X-1984, A. Otaka; Obanazawa-shi, along Route 347, 2 9, y-x-1988, T. Tanabe. chibaprei- : Awa-eun, Amatsu-Kominato-cho, Awa-Amatsu, Saneiri coast: 1 ij, 1 9,2juv.,21-v-J9S3; 1 6, lS-ix-1983, K lshii; 2 o\ 1 9, 1 8-xii-l 983, K . ishii. Awa-Amatsu,

Saneiri, Matsucabana, under growth of maritime fern, Cxrwmium falcatum, 1 -2m, 23 6 . 2 1 S (16 6). 27-vm-1984, nt. tokyo pref.: I. Hachijo, Mine, 2 4. 2 9, 17-18-X1-1983, H Okada. TQYAMi pref. Toyama-shi: Yokogoshi. I 9. 20-X-1978; 1 juv., I8-V-I990, N. Nunomura; Hamakurosaki,

Plnus thtmksrgi foresi. i juv.. 2-V-1979; 3- S, 3 9, 6-vin-1980» N. Nunomura. wakayamaprek;

Nishimuru-min,Shirahama-ch6, Shirahama. Seto Marine Biological Lab., i d,3 9. 10-vii-1983, S.

Otsuka; 29 J.4I 9,43juv.(16o\ 1 9 ,9 juv.),29-vn-1983,NT tottoripree: hvami-gun, IwariUH Uradome coast, Kamogaiso, 4d, 39, 7-IX-1988. nt. R. G. Holmberg; Ketaka-gun, Ketaka-eho, Yatsukarni, Anedomari coast, lo% 39. 5-ix-l9S7. nt. OKAVAMA pref: Oku-eun, Ushimado-cho.

Ushimado Marine Biological Station of Okayama Univ., Benten-iwa. 1 A . 4 9 . IO-vui-1983, T. Sato.

Hiroshima pref.: L Nomi-jima, Irukanohana: 12 juv., 13-vi-l 976; 1 o*. 4-vn-1977, nt . EHiMEPRtt :

Onsen-gun, Nakajima-cho, 1. Nakajirna. Okushi: 3 V, 6-7-vm-197l; 1 9 , 13-vm-I974, nt; Okushi, Seno-hana,51 cJ/69 9 (30 d. 6 9), 26-vni- 1986, nt.T. Nakajirna: Himcgahama, 1 V. 13-vm-1974, nt, I. Takashima off Himegahama, 3 d, 3 9, 6-vm-1971, nt Maisuyama, Mt Fukumi, Fukumiji, 860m, 2 juv., 6-ix-l970, nt Nagasaki pref: Iki-gun, Lshida-cho, Kukishoku, 4 d, 5 9 ,2 juv.. 20-vi-1990, M. Yamashita. klimamoto pref. Ushibuka-shi, 1 Tojima, 2 juv M. Yoshikura;

Amakusa-gun, Itsuwa-cho, Oniike, 3 d, 1 9 , 2 juv., 8-VU-1958, M. Yoshikura; Matsushima-cho,

1. Aitsu.4 ct , 1 9, 10-vm-1957 M. Yoshikura; Matsushima-cho, Macjima, 2 6 , 4 9, 20-vu-1963,

M. Yoshikura. kagoshimapref.; I. Yakushima. Shiratani-Unsui-kyo to Ml. Miyanoura-dake, 1 9.

2-5-VIII-1983, 11. Okada. I. Tanegashima: Hamada, ljuv., ll-vn-1983, Sk. Yamanc; Makigav,

I juv., 3-5-V-1984, Sk. Yamane. Kumage-gun, Kamiyaku-cho, 1. Kuchinoerabu-jima. S , 27-lV- 1984. Collector unknown-

MATING BEHAVIOUR AND FEMALE SPERM STORAGE IN PHOLCUS PHALANGIQIDES (FUESSLIN) (ARANEAE)

G. UHL

Uhl, G. 1 993 11 11: Mating behaviour and female sperm storage in PhotCUS phalangioides (Fuesslin) (Araneae). Memoirs oftlie Queensland Museum 33(2); 667-674. Brisbane. ISSN 0079-8835.

Females of P'hole us phalangioides do noi possess receptacuia scminis but store transferred spermatozoa in their genital cavity. The spermatozoa are embedded in glandular secretion that is discharged from two accessor)' glands situated in the posterior wall of the genital cavity. The gland cells belong toacomplex type ofclass 3 cells according to the classification ofNoirot and Qucnncdy( 1974, 1991). With the sperm mass of a single copulation the females are able to fertilize several batches of eggs, although the sperm might be easily washed out with the passage of the first batch of eggs (Forster, 1980). Nevertheless, females allow repeated copulations. The first copulation usually took over an hour but subsequent copulations lasted only a few minutes, no matter whether the female mated with the same or with a different male. Copulations after egg-laying tended to be long. Male spiders might achieve reproductive advantage in copulating with any female they meet and females might have an interest in filling up their storage capacities. Phoicus phalangioides Wcibchen besitzen keinc Receptacula seminis im iibliehen Sinne, sondern speichern die wahrend einer Kopulation iibertragenen Spermien im Hohlraum des Uterus externus. Die Spermien werden dort in ein Sekret eingelagerl, welches von zwei akzessorischen Driiscn produziert wird. Diese Drusen bcfindcn sich in derposterioren Wand des Uterus externus. Die DruscnzcHcn sind nach einem Klassifikaiionssyslcm, das fur epidcrmale Driiscnzcllcn von Inscktcn erstclll wurdc (Noirot and Quennedy, 1974, 1991), einem komplizicrtcn Typ der Klasse 3 zuzuordncn. Mit den Spermien, die wahrend cincr etnzigeu Kopulatiou tibertragen wurden, konncn die Wcibchen mehrere Eigelegc bclruclttcii (Uhl. in press a), entgegender Annahme. die Spermien kOnnten wahrend der erstcn Eiablage leicht ausgewaschen werden (Forster, 1980). Die Wcibchen lassen dennoch mchrcrc Kopulationen zu, wobei die erste Kopulation gcwohnlich liber cine Stunde daue-rt, jede wcitcre Kopulation schon nach wenigen Minutcn abgebrochen wird, unabhangig da von ob Cs sich urn das selbc odcr um ein neucs Marmchen handclt. Nach ciner Eiablage lassen die

Wcibchen wieder lange Kopulationen zu r Fur die Mannchen mag es von Vortei! sein sich mil jedem Wcibchen zu paarcn dem sic begegen, und Wcibchen konntcn ein Intcrcsse daran haben ihre Speicherkapazitat voll ausxusehopfen. QAraneae, Pholcidae, Pholcus photon- gioidcx, sperm storage, glands, secretion, infrastructure, repeated matings, copulation duration,

Cahriele Uhl, Instilut fur Biologic I (Zoologie), UntversiUit Freiburg LBr. t Albertxtrafle 2 Ja t 7800 Freiburg t.Br.. Germany; 28 October, 1992.

Most female spiders store sperm within storage amount and fertility of the spermatozoa trans- Structures lhat are spatially separated from the fcrrcd during a single copulation. This would be genital cavity. Some 'primitive* spider families risky as the spennatoznu might be defective 08 suchasDiguetidae.Liphistiidae, Arehaeulaeand insufficient 1 export the females to fill tip their Pholctdae retain the sperm mass within the gem- storage car>acitv bv means of repeated mating* at

UJ cavity itself (Forster. 1980)- Forster assumed . . |easl |fl< t Maying when few clumps of sper- that the bursal storage mode has little survival - majE flfl rcmai(; jn thc |ta , ^ flftcr - - ..- , * . value as the sperm mass gets flushed out with the 4 „ "V , , c .F 3 ,. :. ovipcsition.h However, it the female docs noi passage of the eggs and therefore, storing the , . . ,he opportunity populate repeatedly the stored spermatozoa in spatiallv separated storage strut- 10 spermatozoa can be sufficient tor fertilizing fol- mres would eliminate the need for repeated in- lowing egg batches successfully. This study will semination. Despite the bursal storage mode, female Phoicus phalangioides (FucssYm) are ahle also give a brief morphological account on the to fcTtizilc numerous batches of eggs with the bursal storage mode m P. phtttangiQifa and will sperm of a single insemination [Uhl, in pfeft a), present histological aiod nliriLStructura] findings However, the females probably do not rely on the on thcglandulartissuc that exudes its product into 668 MEMOIRS OF THE QUEENSLAND MUSEUM

FIG. 1. Dorsal wall of genila! cavity of P. phakmgioides. A, Pore plates viewed from genital cavity; B, Pores that exude secretion, C, One pore plate from its dorsal side, glandular tissue removed; D, cuticular ductules of accessory glands. the genital cavity for sperm storage. For more behaviour the spiders were kept in couples in detailed information see Uhl (in press b, c). plastic boxes ( 1 6.5x9x6. 5cm) and their behaviour was recorded day and night on video tape. MATERIAL AND METHODS Five different experimental set-ups were used in order to answer the indicated questions:

Juvenile Pholcus phalangioides were reared 1. Duration of copulation: Virgin females were individually in the lab. To investigate mating offered unexperienced males. BEHAVIOUR AND SPERM STORAGE IN PHOLCUS 669

2. Copulation duration with sperm-depleted males: Virgin females were offered recently experienced males (J/2 hour after termination of copulation), 3. Repeated matings with same partner: Pre- viously virgin females were kept with previously unexperienced males for up to 20 days. 4. Repeated matings with changing partner: J/2 hour after their first copulation females were offered unexperienced males. In order to check the influence of box-size On mating behaviour 4 experiments were carried out using 2.5 times bigger boxes (18x12x1 1 5cm). 5. Duration of copulation after egg-laying: Post-oviposition females that had mated once were allowed to copulate again with unexperienced males. For SEM studies adult females were anaesthetized, dissected and fixed in 70% ethanol or Bourn. In order to investigate the sclerotized parts of the female genital tract, the female genitalia were put in 5% NaOH solution until the soft parts were dissolved. Some genitalia were opened or cut with a sharp razorblade to locate the sperm mass in the genital cavity. They were dehydrated in ethanol, CP-dried, sputter coated with gold and examined in a Zeiss Semco Nanolab 7. For light- and electron microscropy the spiders were anaesthetized and dissected in glutaral- dchyde. After fixation in 2% osmium tetroxide/- glutaraldehyde, they were post-fixed in osmic acid (modified after Franke et a/., 1969), dehydrated in graded series of alcohol followed by propylene oxide and embedded in Epon. The FIG. 2. A, Genital pfafc '"lipped back, dorsal wall o! semithin sections (0.7- i urn) were cut with glass genital cavity revealed. Two pore plates, one con- knives on a Reichert OmU3 and stained with cealed by secretory 'plug'; B, secretory 'plug' in tnluidin. Ultrathin sections were cut with glass cutsapittally;C Sperm mass in female genital cavity ? knives and diamond knife. They were stained secretion in female genital bract. with uranyl acetate, counterstained with lead citrate and examined in a Zeiss EM9 electron copulations were long (x: 56.8 minutes). The microscope. males probably refill Their copulatory organs prior to the second copulation. The filling was RESULTS observed in only one case by chance and was not detectable on the video recording.

Mating Behaviour 3. Unexperienced males and females were kepi in pairs up to 20 days. Six females out of 12 1 Virgin females copulated with unex- allowed copulation from time to time: one female perienced males over an hour (x=64.5 minutes; copulated twice, four females copulated 3 times, sd=26.6; shortest duration: 16min, longest dura- one female copulated five times. The second and tion: 122 min; n=42). This supports findings of the following copulations were always very short, Reagan and Reagan ( 1 989) who investigated 1 04 they took only 2 to 5 minutes. Copulation dura- pairs (mean copulation duration: 72.3 min: tion tended to decrease in successive matings. sd=43.3; shortest: 10 min, longest: 304 min). 2. Five males that had mated with virgin 4. Unexperienced males were brought to females (long copulation I) were brought half an females that had already mated once (long houraftercopulation to another vugin female. All copulation 1). In 9 of 1 4 cases the females allowed M

tM MEMOIRS OF THE QUEENSLAND MUSI- .1 1

Copulation duration oenialsct-lip n (muU

1. Virgin 9 with unexperienced <5 42 64.5(16-122:26.6)

recently 2. Virgin t Villi 5 56.8(37-75; 119) experienced 6

J. Second copulation with same ct 6* 3.6(2-5; UCi

4 Second copulation wilh 9* 2.6(1.5-5; 10} UBMpcncncci.' '

Posi-oviposilion 9 wiih 5. 7 59*(2i-IW;3t^ unexperienced rf

TABLE I. Mating behaviour in P. phatangioUks ay a function ofieniale reproductive history. Only cases nf copulation given- Range and standard deviation in parentheses. further copulation. Again, second copulations basted only a few minutes (1-5-5 min). Control cNpenmcnls using bigger boxes showed that 3 females out of 4 allowed repeated matings (4.5;

2; 1 + 1 min). 5. Seven females that had mated once (long copulation I) and were kepi separately afterwards, had access to males after oviposition. Copulation duration was long. Copulation duration seems to depend on the reproductive history of the females (Table 1 ).

The Genital Cavity The dorsal (posterior) wall of the genital cavity is characterized by iwo oval pore plates (Fig. 1a). The pore plates converge in direction of the ridges FIG. 3. Semithin section of accessory glands. A, Lon- and grooves that make up the heavily sclerotized gitudinal section. Arrows marknutiei of different cell valve which separates the genital cavity from the types. Ba; basal lamina, E2: outer envelope celL G: oviducts. Both plates arc perforated by pores of gland cell, m: microvilli. P: pore piale. s: secretion of 3ou.rn in diameter (Fig. IB). The pores are in the gland cells, S: secretion in genital cavity; B. contact with gland cells that exhude their glan- Transverse section. Arrows and stars mark nuclei of dular secretions into the cavity (Fig. Itf). different cell types. Scale lines: 20firn. The tissue-free pore plates reveal the canal zones of the glandular tissue when looked at from The Acchssoky Glands their dorsal side (Fig. Ic). Situated in cavities, The glandular tissue is composed of highly cone-shaped hollow structures are apparent (Fig. elongated ceils (Fig. 3). Different cell types form Id). These are the distal regions of the canals that the gland with various nuclei at different levels open into the uterus as the the externus pores of (Fig. 3a). The gland consists of lightly stained pore plate. Proximally the canals change into thin cells whose nuclei lie close to the basal lamina, ductules thai exhibit a rough surface after 6- and densely stained cells with more distal nuclei. !0u.m(Fig. ]D». Nuclei of other celt types lie mainly in the centre.

Sperm Storage Secretory vesicles are apparent in the densely stained cells (the gland cells G). Accumulations The accessory glands discharge their products of such vesicles arc found in the apical second through the pores of the pore plates into the third of the glandular tissue. Each accumulation genital cavity and form two portions of secretory forms two portions of tightly packed secretory 'plugs'. (Fig. 2,\). During copulation, the male globules that discharge their contents into a com- transfers sperm mass into the female secretion mon reservoir which is homogeneously coloured (Fig. 2H). The spermatozoa are surrounded by (Fig. 3a. H). individual secretory envelopes, they are coiled and inactive (Fig. 20. Bordering on the gland's orifices is a /one of .

BEHAVIOUR AND SPERM STORAGE IN PHOLC'US

greyish coloration that is formed by the light coloured cells (outer envelope cells E2). Histology and ukrastmcture show that the glandular tissue includes many similar units, each prnvided with a cuticular ductule that leads to the pore plate (Fig. 3a). Each unit comprises two gland cells and two envelope cells. The two gland cells (Gl. 2) join each other to form a common reservoir (Fig. 4, 5B). They are rich in granular endoplasmic reticulum (Fig. 5a), mitochondria and dense secretory vesicles and exhibit numerous golgi complexes in the supranuclear region. The vesicles vary in size (up to 1.5u,m in diameter) and get more numerous in the distal cell region. They are enclosed in a close-fitting membrane which is obscured by the matching density of the mature granula. The inner envelope

cell (El ) surrounds and partially separates the two gland cells (Fig. 5b, C, d) and forms the proximal pari of the ductule (Fig. 4c). The outer envelope cell (E2) surrounds all of the previously mentioned cells. Its cytoplasm is poor in organelles (Fig. 5a, b, d). It produces the distal part of the ductule and forms numerous microvilli that gather round the ductule and the orifice (Fig.SE) and represent the greyish zone visible in the semi thin section of Fig 3A. The gland cells and the outer envelope cell form a so-called basal labyrinth adjacent to the basal lamina (Figs 4, 5A). The glandular units are separated from each other by elongated epithelial cells.

DISCUSSION

The glandular units of the accessory glands in P. phalangioides belong to class 3 cells (Noirot and Quennedy. 1974, 1991). According to that classification, a gland cell is associated with a cuticular ductule that has been secreted by a 'canal' cell. In P. phalangioides however, there are two gland cells that are always connected by a common microvilli region, one inner and one outer envelope cell that form a double ensheath- ing of the gland cells. Moreover, both envelope cells take part in producing the canal that leads to the pore plate. Therefore, the glands studied here belong to a complicated type of class 3 cells. There is some information on gland structures in female spider genitalia. The glands of the re- FIG. 4. Schematic reconstruction of one glandular unit cepiaculumseminisofTelemidae belong to class of accessory glands Ba: basal lamina, BL: basal I type of gland cells (Lopez and Juberthie-Jup- labyrinth, U. ductule. El. inner envelope cell, H2: eau, 1983). and in some Theraphosidae De Carlo \' outer envelope cell G 1 2: gland cells, N: nuclei;'-. ( 1 973) stated a class 1 composition. However, the pore plate, light microscope study of Kovoor (1 98 1) indi- 672 MEMOIRS OF THE QUEENSLAND MUSEUM :

BEHAVIOUR AND SPERM STORAGE IN PHOLCL'S

eates that a more complex gland composition as matozoa arc liable to be washed out du in P phalangioides might not he exceptional. Mtion (Forster. 19B0) Nevertheless, The glandular units of P. phalangioides could males of P. phalangioides succeed in produ after single I unction as a two-component-sy^tcm 89 the outer several fertile batches of eggs a envelope cell shows conspicuous microvilli that mating (Uhl, m press a surround the orifices of the ductules. This indi- :h females do not depend on repeated cates secretory activity although the outer en- insemination, they allow further copulati velope cell contains no stainable secretory These always last only a few minutes in con droplets orgranules. Their product could have got to the first copulation that ia^ts over an hour and Lost dunng fixation or. the cells produce there is no apparent difference in copulation dura- release their product only on demand. Not all the tion between second copulations with the same or secretory activities of cells are accompanied by with a different male. If new males are able lo microscopically detectable accumulation of the replace sperm ofa previous male, longer copula- product in the cytoplasm. Apart from that, the tions would be expected. Such short copulation* outer envelope cell exhibits a basal labyrinth. It may suggest that mates are able ss female contributes to enhance the exchange of RKi ecules virginity or reproductive history during inset between the haemolymph and the cells (Berridge of their palpal structures and then decide on fur- and Oschmanm 1972) and characterizes active thermvCilmen: of time and energy. to cells that take up or transfer material from or It has yet to be in bother successive ihe haemolymph. copulations result in a transfer of spermaioy The glandular secretions might serve various all, which will aba give information on sj functions such as nutrition of the sperm (Co precedence. Provided they transfer sperm and a/., 1983; De Carlo, 1973; Engelhardt. 1910; fertilise at least some eggs, it would be advant- Forster. 1980), pheromone production (Kovoor, ageous for males to male with any female the) 1981) or sperm displacement from the spcr- meet The females, on the other hand, can be mathecae into the genital cavity during oviposi- expected to fill their storage structure with as |iOft{ Forster etal, 1987; Lopez, 1987; Lopez and many spermatozoa as possible to achieve ihe. Juberthie-Jupeau, 1983). Bngnoli (1976) and eft possible reproductive success. Depend- Lopez and Juberthie-Jupcau (1983) considered ing on the amount of sperm already aceumi activation of sperm prior to fertilization. The in the genital cavity females might allow further glandular tissue might be responsible for trigger- copulations and hence, decide on copulation during activation via a secretory product that gets ation. Indeed, there is wmc evidence that the released exclusively before oviposition. Further, female terminate s copuiaLion the females might achieve advantages from resorbing the sperm mass out of the genital cavity. ACKNOWLEDGEMENTS

1 consider the secretion in the female genital

tract of P. phalangioides serve:- primarily as a F wish to thank my supervisor Prof. Dr. P. depot for the sperm that guarantees successful Wcygoldt (Freiburg) for making this work pos- Storage as the onset of the female receptivity sible. Dr. M. Schmitt (Bonn) and two anonymous corresponds with the time needed to fill the geni- referecN gave helpful comments on the manu-

tul cavity with glandular secretion (Uhl. in press script. Without a grant from the Studienstifiung

a l. Concerning any either possible functions des Deuischen VoJkes and additional support specific investigations are still lacking. from the Organizing Committee participation at The bursal storage mode is considered a 'prim- the Xllth Congress of Arachnology, Bns Hive" mode with little survival value as the sper- would have been impossi:

FIG, 5. Ullrathin secfions of accessory glands. A) Longitudinal section. Nuclear region of gland cells, b.

1 i Longitudinal section. Two ^land cells join to form common microvilli region, surrounded by twoenvelopc 1

forming first part of canal (arrow). C) Longitudinal section. Mierovi I! i region. Arrow shows cleaving bactctium. D) Transversal section. Beginning olr ductule. Gland cells joined, enveloped twice. E) Microvilli region ol outci cell, onirr envelope cell close to pore plate. Small ductule (arrow); BL: basal labyrinth. El : Inner envelope E2. envelope cell, cr: endopbsmaiie reticulum, GJ: gland cell 1. G2: eland cell 2. n: nucleus, m: microvilli, s secretory droplets. All scale lines. 2\i.m. 674 MEMOIRS OF THE QUEENSLAND MUSEUM

LITERATURE CITED KOVOOR, J. 1 98 1 . Une source probable de pheromone sexuelles: les glandes tegumentaires de la region BERRFDGE. M.J. & OSCHMAN, J.L. 1972. genitale des femelies d*araign£es. Alti dclla Transporting epithelia'. (Academic Press, New Societa Toscana di Scienze Naturali, Memoire York, London). Serie B, Supplement 88: 1-15. BRIGNOLI, P.M. 1976. Ragni d'ltalia XXIV. Note LOPEZ, A. 1987. Glandular aspects of sexual biology. sulla morfologia dei genitali intcrni dei Pp. 121-131. In Nentwig, W. (ed.). 'Ecophysiol- Scgestriidae e cenni sulle specie italianc ogy of spiders*. (Springer: Berlin, Heidelberg, (Araneac). Fragmcnta Entomologica 12; 19-62. New York). COYLE, F.A., HARRISON, F.W., MCGIMSEY, W.C. LOPEZ, A. & JUBERTWE-JUPEAU, L. 1983. Struc- & PALMER, J.M. 1983. Observation on the structure ct ultrastrucfure de la spermatheque chez ture and function of spermathecae in haplogyne Teletna tenella Simon (Araneae, Telemidae). spiders. Transactions of the American Micros- Me"moires Biospdologiques 10: 413-419. copical Society 102: 272-280. NOIROT, C. & QUENNEDY, A. 1974. Fine structure DE CARLO, J.M. 1973. Anatomia microscopica de las of insect epidermal glands. Annual Review of Entomology 19:61-80. espermatecas de los gencros Grammostola y Acanthoscurria. Physis Seccion C, Buenos Aires 1991. Glands, gland cells, glandular units: some 32(85): 329-342. comments on terminology and classification. An- ENGELHARDT, V. 1910. Beitrage zur Kenntnis der nates de la Societd Entomologique de France vveiblichen Copulationsorgane einiger Spinnen. (N.S.)27: 123-128. Zeitschrift fur wissenschaftliche Zoologie 96: 32- REAGAN, N.L. & REAGAN, A.P. 1989. Mating and 117. egg production in Pholcus phalangioides.

FORSTER , R .R. 1 980. Evolution of the tarsal organ, the American Arachnological Society, Newsletter 40. respiratory system and the female genitalia in UHL, G. (in press a). Sperm storage and repealed egg spiders. Verhandlungen des S.lntcrnationalcn production in female Pholcus phalangioides Arachnologen KongreB, Wien 1980: 269-283. (Fuesslin) (Araneae), Bulletin de la Soriete FORSTER, R.R., PLATNICK, N.I. & GRAY. MR Neuchateloise des Sciences Naturelles. Actes de 1987- A review of the spider superfamilics la XHIeme Colloque Europeen d'Arachnologie, Hypochilidae and Austrochilidae (Araneac, Neuchatel 2.-6. Sept. 1991. Araneomorphae). Bulletin of the American (in press b). Genital morphology and sperm storage Museum of Natural History 185: 1-J16. in Pholcus pluilangioides (Fuesslin) (Araneae). FRANKE, W.W., KRIEN, S. & BROWN, R.M. 1969. Acta Zoologica, Stockholm 74. Simultaneous gluiaraldehyde-osmium tetroxide (in press c). Ultrastructure of the accessory glands fixation with postosraication, an improved fixa- in female genitalia of Pholcus phalangioides tion procedure for electron micoscropy of plant (Fuesslin) (Araneac Arachnida). Acta Zoologica, and animal cells. Histochemistry 19: 162-164. Stockholm 74. THE BIOLOGY OF SPIDERS AND PHENOLOGY OF WANDERING MALES IN A FOREST REMNANT (ARANEAE: MYGALOMORPHAE)

GRAHAM F.C. WISHART

Wishart, G.F.C. 1 993 1111: The biology of spiders and phenology of wandering males in a forest remnant (Araneae: Mygalomorphae). Memoirs of the Queensland Museum 33(2): 675-680. Brisbane. ISSN 0079-8835.

Eight syntopic mygalomorph species, four of one genus, are recognised as inhabitants of a small remnant forest. Over six years 1,207 mature males were trapped and collected from a nearby home swimming pool and the wandering times of mature males of the species are compared. The population density within the forest is estimated. It is suggested the high number is because ofan 'edge effect* supported by a reduction in predator numbers, an 'island effect' and that mygalomorphs have low dispersion powers, long life cycle and sedentary life sty\c.\3Mygalomorphae, Hadronyche, Misgolas, Stanwellia, community, forest, remnant, phenology, population, syntopic, wandering,

Graham F.C. Wishart, 'Scalloway*, Willowvale, Gerringong, New South Wales 2534, Australia; 23 November, 1992.

Phenological studies of spiders commonly con- taxonomic revisions of particular groups (Coyle, sider the distribution and abundance of species in 1971; Raven, 1984). Main (1982) synthesised relation to habitat variation or disturbance (Peck male wandering data for her studies of arid zone and Whitcomb, 1978; Koch and Majer, 1980). spiders. Studies of male spider wandering patterns are less This paper presents male wandering patterns common and are usually associated with over a six year period for six species of syntopic

FIG. 1. Aerial view, easterly aspect, from study site towards Gerringong township. 676 MEMOIRS OF THE QUEENSLAND MUSEUM

FIG, 2. Aerial view, south-westerly aspect, of study site. Pool length is 10m.

forest mygalomorphs and also discusses possible forest but is dense and extensive around the forest ecological effects which could account for the edge (not allowed for in the forest dimensions) viability of the spider population. The long term preventing the entry of cattle. data gathering was enabled by the serendipitous The site is on the edge of a ridge and slopes location of an in-ground swimming pool acting rapidly downwards to the north east. The soil is as a large pit-fall trap close to remnant forest. of volcanic origin. The surface soil is thin and spread over a basalt bedrock. Large but moveable STUDY AREA rocks are prolific. Leaf litter is ca 5cm deep.

The study site is on the property 'Scalloway' METHODS 150°47'23 H (34°44'irS t E) (Figs. 1,2), near Ger- ringong, N.S.W., and is a remnant piece of mar- Wandering mygalomorph spiders trapped in a ginal Complex Notophyll Vine Forest (Bywater, 10m long swimming pool 15m from the forest 1978) at an altitude of 1 10m. It is a fragment (ca edge were collected daily from 1 July 1985 to 30 1 June 1991. Variables possibly affecting the num- 95m x 55m) of the original 'lllawarra Scrub , varying forest types which occupied much of the bers of spiders trapped, such as domestic lighting, coastal strip east of the escarpment in South East- maintenance of grounds and ability of different ern Australia between the towns of Stanwell Park species to wander further or more quickly than in the North and Bomaderry on the Shoalhaven others, were not taken into account. For three River in the south, a distance of 75km. Land years from 28 August 1986 to 25 October 1989 2 clearing removed over 80% of this original the burrows within a plot 1.5m were examined vegetation (Fuller and Mills, 1985) and took weekly. place mainly between 1850 and 1910, the remnants being restricted to land unsuitable for graz- RESULTS ing because of steep gradients and rock outcrops. The forest overstory includes the trees Syncar- Community Composition pia glomulifera and Alphitonia excelsa and Mature males of eight species of mygalomorph several vine species forming a discontinuous spiders were collected from the pool, viz.: Mis- canopy at a height of about 12m. The understory golas hubbardi Wishart, 1992 (Idiopidae), M. consists mainly of the small tree Commersonia dereki Wishart, 1992, M. kirstiae Wishart, 1992, fraseri and the exotic woody scrambling shrub M, robertsi (Main and Mascord, 1974), Lantana camara. L. camara intrudes within the Hadronyche sp. (Illawarra group) and Atrax sp. . .

SPIDER BIOLOGY IN A FOREST REMNANT 677

£ 5 - 3

-- & 3

° 2 -h 3

I. i Il l l fH m H nMMH HMMM tl J i WUI Jrll AWM m imni H t H i rmm t

f»- K n op CT1 pi - cy «y c? O) ty (0 di r^- 6 * O 4 *- 04 r-.j CVJ »- ftl w «- «- ft DAYS OF THE YEARS

FIG. 3. Total number ofMisgolas hubbardi captured on same days of years for six years from 1 .7.85 to 30.6.91

(Hexathelidae), Stanwellia hoggi (Rainbow, (based on the assumption that the ratio of the 1914) (Nemesiidae), and Kiama lachrymoides number of burrows occupied by all mature female Main and Mascord, 1971 (Cyrtaucheniidae). The spider species to the number of burrows occupied Hadronyche sp. is that referred to as Hadronychc by M. hubbardi is equivalent to the ratio of the sp. 20 by Gray (1987). The Atrax sp. is smaller number of captured mature male spiders of all but similar morphologically to A. robustus species to the number of captured mature male Cambridge, 1877. It differs also in that males spiders of M. hubbardi) indicates there should be wander from October to December whilst those 41 mature female mygalomorphs per rrr in the of A. robustus wander mainly in January (Gray, 1 .5m" plot. 1986). They are considered here to be different In the 1.5m* plot the maximum number of species. burrows counted at one time was 55. These included open burrows, those known to exist but Population Density temporarily sealed and very small open burrows Because much of the surface area of the forest of which most failed to persist. There were 27 floor is covered by large rocks and occupied by burrows larger than 8 mm in diameter of which trees the 1.5 m" plot cannot be representative of 10 were each occupied by a mature female M. the whole forest area. However, an extrapolation hubbardi. Due to physical limitations in accurate-

6 +

£ 3 P

- l T I I I + tlTTT H M l fH m iT MM l H H IIIIIII 11 M TT M H I TT M I I H M I M TTTIT HM T MH TI M I' HM H I N- t- r~ o> o» o o co cj to o — OJr-C\J~«l w »- r- n DAYS OF THE YEARS

FIG. 4. Total number of Misgolas dereki captured on same days of years for six years from 1 .7.85 to 30.6.91 678 MEMOIRS OF THE QUEENSLAND MUSEUM

7 -

i, e

§ 5 +

£ 4

U zCD ' r 1 4

r TTTfTTTTTTTTTTTTTTTTTITP TT flW>TTrrMl>T-T>TTriTrTf'TTTTTT nTTTmTTTTTTTrTTTTriTI I HH T THU r ni I H l ll ll H F l'T I N r~ i-. UJ a c* f- c. o a wcvieg — — c« i*i n -w Li", Lf) :j3 tp 'J Xf "** - * ^ f*. Q o — ui b> e*j is a s a s $ a fi « 4 ru a gs - cu s DAYS OF THE YEARS

FIG. 5. Total number of Misgolas kirstiae captured on same days of years for six years from 1 .7.85 to 30.6.91 ly counting small burrows and recognizing the K. lachrymoides were too small for deductions to presence of the burrows of M. kirstiae and be conclusive.

Hadronyche sp. the total count must fall short of During the six year period, 1 1 female 5. hoggi the actual mygalomorph population. were also trapped in the pool during the months from March to October. Phenology Of Wandering Males (Table 1) Wandering mature males of M. hubbardi were DISCUSSION found almost throughout each year (Fig. 3). The extensive wandering period of M. robertsi is in- No previous reports have been found of either terrupted during January and February when the eight species of mygalomorph spiders or four frequency of capture is reduced (Fig. 6). The species of one genus (Misgolas) of spiders exist- wandering period for S. hoggi is also long, ex- ing elsewhere syntopically or sympatrically. tending over six months (Fig. S). As is customary The high proportion of mature female S. hoggi to expect of male mygalomorphs, the wanderings collected is surprising, 11 females to 26 males, of M. dereki (Fig. 4), M. kirstiae (Fig. 5) and H. and 10 of the 1 1 were taken at times coincidental sp. (Fig. 7) are restricted to short and more precise with the male wandering period. The number annual periods. The collections ofAtrax sp. and reflects the long-legged, male-like morphology

of the female adapting it to roam. Further, that only one male K. lachrymoides was collected No. Total Peak active may indicate the inability of the of this Species peak trapped male (nipped periods periods per day species to wander a long distance a reflection in

April-July 58 \)-i this case of the male's short-legged female-like Misgolas hubbunti 197 Nov. -J an. 20 0-2 morphology. The usual sexual dimorphism of Misgolas dereki 153 Apr-June 98 n-7 mature burrowing mygalomorph spiders (female stout Misgolas kirstiae M Sept. -Nov. 98 0-7 bulky, legged; male less bulky, long legged) is contradicted in these species. :-, two Misgolas robertsi l Nov. -May 96 0-4 Hadronyche 606 March- May y6 0-29 Some possible reasons for the dense mygalomorph spider population are offered. StonweUia kostt* 26 May-Nnv too n-i First, predators of mygalomorph spiders may Atmx sp 5 Oct.-Nov 100 0-2 have become extinct in the area following human Kiama lachrxmoides 1 Nov settlement. For example, the bandicoot Total 1207 (Perameles nasuta Geoffroy, 1 804), reputed to be a mygalomorph spider predator (Main, 1976; TABLE 1. Summary of numbers and timing of male Preston-Mafham, 1984) and once very common spiders trapped in pool from 1 July 1985 to 30 June in Willowvale, has been rarely seen by the author 1991. and not at all for 15 years. Here then, possibly, is SPIDER BIOLOGY IN A FOREST REMNANT 679

CM — CM DAYS OF THE YEARS

FIG. 6. Total number of Misgotas robertsi captured on same days of years for six years from 1 .7.85 to 30.6.91 a paradox where habitat destruction and the intro- 'Scalloway' site by ca. 400m of pasture land there duction of feral animal pests, may increase, not is a paucity of mygalomorph spider burrows and decrease, mygalomorph spider numbers by few insects suitable for prey. Burrows are more decreasing spider predators. common near the forest edge, and so too is insect Second, the concentration may result from an life with grasshoppers, moths and crickets 'Island Effect' where it is found that the popula- prolific. Because the 'Scalloway' forest remnant tion of an individual species can be far greater is small it is in effect a forest edge throughout than expected when the species is insulated with an abundance of suitable prey. Laurence within a small ecological system (separate pers. (1991) states, in the tropics, forests near edges comms, M.Gray and R.Raven). Also Main ( 1 987) exhibit striking changes in microclimate, vegeta- proposed that mygalomorphs are admirably fitted tion structure and composition, disturbance to persist in small isolated areas because of their regimes, and invasions of species from adjacent low dispersion powers, long life cycle and seden- habitats. Thus, in fragmented systems, species tary life style. that tolerate edge conditions are often favoured'.

Finally, in a larger (ca. 300 ha) forest with I suggest that this is the explanation for the complete canopy and separated from the presence of this dense mygalomorph population.

id aa M

34 -j-

32 --

c/i 30 - IE 3 26 — I 26 2 24 - M £ 22 M to *

fe 18

£ 1Z % ID

o mi 1IIT H T iT ti T m irmm Trmmmn i " imiiiuiiiiiiiiiiiTmmiTUtW

•— wi o» c\i d> (*) V •- w t- S «m i** *• DAYS OF THE YEARS

FIG. 7. Total number of Hadronyche sp. captured on same days of years for six years from 1,7.85 to 30.6.91 6S0 MEMOIRS OF THE QUEENSLAND MUSEUM

2 *S 1

03U TT>irni Hn iriTTTT mH TnTTITTTTT H r' H ri|lTfT I TM' I H'TITT TT TT l T T T TT i n T r TTT TTHTTH T IH TTnmim f T T T T T T r T T*T T T T f T*T l*T T

IT J - « >- m

FIG, 8. Total number of Stanwellia hoggi captured on same days of years tor six years from 1,7.85 to 30.6.91

ACKNOWLEDGEMENTS Australia. Journal of the Royal Society of Western Australia 63: 2 1-28. LAURENCE, W.F. 1991, Ecological correlates of ex- My thanks to the Australian Museum and Dr tinction pronencss in Australian tropical rain Michael Gray for the provision of facilities and forest mammals. Conservation Biology (5) 1: 79- advice; Mr Paul Askew for field assistance; Mr 89. Harry Mitchell for aerial photographs; Mr MAIN. B.Y. 1976. Spiders'. (Collins: Sydney). George Browning for helpful criticism; and Dr 1982. Adaptations to arid habitats bv mvgalomorph spiders. Pp. 273-283. at. Robert Raven of the Queensland Museum for In Barker, VV.R* et (eds). 'Evolutions of the and fauna of arid computing suggestions and encouragement. flora Australia", (Peacock Publications: Frcwville). Especially do I thank my neighbour, Dr Peter 1987. Persistence of invertebrates in small areas; Linklater, for the execution of statistical work and case studies of Trapdoor spiders in Western the preparation of the graphs and Dr Bill Australia. Pp. 29-39. In Saunders. D. et ai (eds). Humphreys of the Western Australian Museum 'Nature conservation: the role of remnanLs of

* for comments on the data and guidance. I am native vegetation, (Surrey Bcatty and Sons: Syd- grateful to the CSIRO Science and Industry En- ney). dowment Fund for its support. MAIN. B.Y. & MASCORD. R.M. 1971. A new genus of diplurid spider (Araneae: Mygalomorphae) from New South Wales. Journal ol" the Australian LITERATURE CITED Entomological Society 6: 24-30. 1974. Description and natural history of a ''tube- building" species Dyarryops BYWATER, J. 1978. Distribution and CCDlcgy of rain- of from New South forest vegetation and fauna in the lllawarra. Wales and Queensland (Mygalomorphae: (Thesis, University of Wollongong). Ctenizidae). Journal of the Australian En- 15-21. - tomological Society 8: CAMBRIDGE. O.P 1 877. On some new genera and species of Araneidea. Annals and Magazine of PECK, W.B & WHITCOMB. W.H. 1978. The phenology and populations ol ground surface, cursorial Natural History (4) 19: 26-39. spiders in a forest and a pasture in the south central COYLE, FA* 1971 . Systematica and natural history of United States. In Mcrrett, P. (ed.). 'Arachnology'. the mygalomorph spider genus Amrodiaeius and Symposia of the Zoological Society of London 42: related genera ( Araneae. Antrodiaetidae). Bulletin 131-137. of the Museum of Comparative Zoology 141: 1984. 'Spiders of 269-402. PRESTON-MAFHAM, R. & K. the world." (Blandford Press: Dorset), FULLER, L. & MILLS, K. 1985. Native trees of central RAINBOW, WJ. 1914. Studies in Australian lllawarra. (Weston and Co.: Kiama). Araneidac. Records of the Australian Museum 10: GRAY, M.R. 1986. A systematic study of the funnel 187-270. web spiders (Mygalomorphae: Hexalhelidac RAVEN, R.J. 1984. A new diplurid genus from eastern Atracinae). (Unpublished PhD. Thesis. Mac- Australia and a related Anam? species quarie University). (Diplurinae: Diplundac; Araneae). Australian I9S7. Distribulion of the funnel web spiders. Pp. Journal of Zoology, Supplementary Series 96: 2-32 3 1 1 . In Cuvueevich. J.. Pcarn, J, & Davie, P. 1-51. Icds). 'Toxic plants and animals'. (Queensland WISHART, G.F.C. 1992. New species of the trapdoor Museum: South Brisbane). spider genus Misgoias Karsch (Mygalomorphae: KOCH, L.E. & MAJER, J.D. 1980. A phenological Idiopidae) with a review of the tube-building investigation of various invertebrates in forest and species. Records of the Australian Museum woodland areas in the south-west of Western 44.263-278. b c1 .

THE MACARONESIAN CAVE-DWELLING SPIDER FAUNA (ARACHNIDA: ARANEAF JORG WUNDERLICH

Wunderlich, J. i 993 II II: The Macaronesian cave-dwelling spider fauna. Memoirs ofthe Queensland Museum 33(2): 681-686. Brisbane. ISSN 0079-8835.

The composition of the cave-dwelling spider fauna of the Macaronesian Islands-Madeira, the Azores and the Canary Islands-is compared with the endemic epigean spider fauna of liiese archipelagos. The giades of adaptations in the cave-dwelling spiders are compiled, and the following questions are discussed: from which geographic regions did the stem species come? What can be said about the evolution of the species? How old are the cave-dwelling spider species'/ Die Fauna der Hohlcnspinnen dcr Makaroncsischcn inseln - Madeira, A/.oren und Kanari- 'nseln- wird mit der endemischen epigiiischen Fauna dieser Archipclc vcrglichen. Der Grad der Anpassung an das Hdhlenlebcn wird untcrsucht und vcrglichen; die folgendcn Fragcn wcrden diskuticri; Wo licgt dcr Ursprung dcr Stammarten? Was kann iibcr die Evolution und das Alter der hohlen-bewohnenden Artcn gesagt warden l\y\raneae t uoglohiies. Conation and Macaronesian Islands, Island biology, biogeography, evolution,

J&tg Wundcrltch. II\ndenbury,s!r. 94, 0-75334 Straubenhardl J, Germany; 7 December, J 992.

Some island groups in the northern Ailantic- Three species of those listed (Table 1 ) arc no! thc Canary Islands, the Azores, the Archipelago restricted to a single island: of Madeira, the small Ilhas Selvagens and by 1 Meta bourneti Simon. 1 929 (Tetragnathidac) most authors

Islands-are called Macaronesian Islands (Fig, 1 1. on Tenerife (Canary Islands); The Macaronesian Islands are mainly of Volcanic 2. Agraecina cafmriensis Wunderlich, 1991 origin, only the Eastern Islands (Fuerteventura, (Liocranidae) is known from caves on Gran

I .anzarole) are partly of continental origin and Canariaand Tenerife (Canary Islands); have been probably connected with Africa some 3 Rugathodes pica (Merrett and Ashmole, million years ago. 1989) (Theridiidae) is known from caves on Pico Fajal (Azores). The first troglophilic and t r o g 1 o 1 1 i and Macaronesian spiders were described in 1985 Here I deal with five questions I. What is the from Tenerife. Now cave-dwellers are known composition of the Macaronesian fauna of from Madeira (1 species), from the A/ores (1 troglophilic and troglobitic spiders and what are species) and from the Canary Islands (at least 17 the differences to the epigean fauna? 2 Which endemic species, see Wunderlich, 1991). by far species are extremely well adapted as cave most species are known from Tenenfc: at least 1 dwellers? 3. From which geographic regions did 4. spec ies = 1 0% of the endemics (and perhaps there the stem species come? What can be said about are hundreds of mostly undeseribed insect the evolution of the species? 5. How old are the species of different orders ). cave-dwelling spider species? Especially on the Canary Islands there are many caves. The best studied system of caves on CaVF-DyvEI UNO AND EpJGHAN ShDBRS

Tenerife - cueva del Viento, cueva Reventon - is The most diverse spider families arc shown in more than 16km long; the length of all 2-3. In the Canarian troglophilic and Macaronesian caves is perhaps more than 100km, troglobitic cave-dwellers (Fig. 2): Dysderidae (at and only a few have been examined intensively. least 35%), Linyphitdae (25%) and Pholcidae (25%), the sum of these 3 families is 85%. (No MACARONESIAN CAVE-DWELLINC Oecobiidae). SPIDERS In the epigean endemic species (Fig. 3) the composition is quite different: Dysderidae 16%, Linyphiidac 15%, Pholcidae 13%, the sum of Nearly all cave-dwelling spiders are endemics these 3 families is 44%, only half compared with of one island or even only one cave (Table 1 ): the cave-dwellers. In the families Dysderidae 682 MEMOIRS OF THE QUEENSLAND MUSEUM

rzi

h?40iju I

500km

I Bilwtfltim, AFRICA T- "JtBWl

4t TENEKtrpO rtrrrrF.VEimi»» 4 I'Cn •[.n.. r |,1„ . f'ier> e=>_-n*iirrnn t « ^_ CANARY ISLANDS /

FIG. I. The Macaronesian Archipelagos. Some Macaronesian cave-dwelling spiders (underlined) and their European and Macaronesian relatives.

(only Dysdera) and Pholcidae (Pholcus and Sper- mented, and the legs are nearly of normal length. mophorides) all epigean genera have evolved This species has been introduced from Europe or cave-dwelling species; in the Linyphiidae only North Africa. members of 5 from 21 genera with endemic 2. Not strongly adapted is Agraccina canarien- species have evolved cave-dwellers (= 25%). sisy but very variable in the depigmentation and Dysdera is the genus richest in species in caves in eye reduction (grades 1-3, Figs 4- 5). The and out of caves on the Canarian and variation is intrapopular. This seemingly Macaronesian Islands. Oecobiidac probably do troglophilic subterranean species is not restricted not find their prey -ants- in the caves. to caves. 3. In four Canarian species of Spermophorides, Highly Adapted Cave-dwellers? the eyes are more or less reduced (Figs 8-1 1); the Different grades of adaptation to cave life in the eyes of an epigean Spertnophorides sp. from Macaronesian spiders is evident in three struc- Tenerife are normal (Fig. 12). Cave-dwelling tures (Table 1): the size of the eye lenses, the body Spennophorides spp. are not strongly related. pigmentation and length and slendemess of legs. The known species occur on four different is- Some true cave spiders of Dysdera have lands. So the eye reduction must have been evolved independently four times. reduced eyes f but neither depigmentation nor long and slender legs. I do not know the explana- 4. Rugaihodes pica is restricted to caves. The tion. Thus, perhaps the eye reduction is the best spiders are strongly depigmented, they have indicator regarding the grade of adaptation to strongly reduced eyes and the legs are long and cave life in spiders. For discussion below I choose slender (grades 3-4, Fig. 6a; cf Fig. 6b the related the following five spider species. epigean R. acoreensis).

1. Meta bourneii (Tetragnathidae, Tenerife) is 5. In Troglohyphantes oromii (Ribera and Bias- restricted to deeper parts of caves, but the eyes co, 1986) (Linyphiidae, Tenerife), the eyes are are not reduced, the body is only slightly depig- tiny or completely absent (Fig. 13), body and legs MACARONESIAN CAVE-DWELLING SPIDERS 683

Di Leg Liocranidae Species RE. Dep T Nestlcidae 5% Meta bourneti Wpal.T =£1 ^1 <2 5% Agraecina canadensis CI (GC,T) 0-3 0-3 ] 1-7 Tetragnathidae Linyphiidae 5% 25% Dysdera labradaensis am 2 2 D. ratonensis CI (LP) 2 2

Lepthyphantes palmeroensis CI (LP) 1 i 1 3

D. chioensis CI (T) 3 1 4

D, ambulotenta CI(T) 3-4 3-4

Dysderldae Pholcus baldiosensis CI(T) 2 3 5

35% Spermophoridesflava CI(GC) 2 3 1 6

S. reventoni CI(T) 1-2 1 2 ^5

Pholcidae D. esquiveli CI(T) ?3-4 1 1? :-o 25% S. fuertevenlurensis CI(F) : 3 *=1 *s6

Walckenaeria cavernicola CI(T) 1-2 1-2 3 6 FIG. 2. Composition of Canarian families of S. jusfoi CI (EH) 2-3 3 3 =£9 troglophilic and troglobitic spiders based on 18 D. unguimmanis CI(T) 4 2-3 =£4 -10 species. Centromerus sexoculatus M • 4 2-3 =£10 are completely depigmented, the legs are very Rugathodes pico AZ 3 3. 3-4 =£10 long and slender (all grades between 3 and 4). The ?Melopobactrus cavernicola CI(T) 3-4 3 3-4 -10 adaptations in Canarionesticus quadridentatus Troglohyphantes oromii CI(T) 3-4 4 3-4 -11 (Nesticidae, Tenerife) and in IMetopobactrus Canarionesticus cavernicola (Linyphiidae, Tenerife) are similar. CI(T) 4 3 4 11 quadridentatus These species show the strongest adaptations to cave life. TABLE 1 . The Macaronesian troglophilic (at least the first two and perhaps the first five species) and Origin of the Stem Species troglobitic spider species and the grades of their adap- Macaronesian cave-dwelling neoendemic tations from (normal structures as in epigean taxa) spiders and their European relatives. The to 4 (= eyeless or almost so/ completely depigmented/ hypothetical origin of all species is the West- very long legs, the species listed below); AZ = Mediterranean area, most came from Europe Azores, CI = Canary Islands, EH = El Hierro, F = Fuerteventura, = Gran Canaria, LP = La Palma, (Fig. 1), e.g. species of Rugathodes to the Azores GC M = Madeira, T = Tenerife, Wpal = West Palearctic). - Madeira seems to be a 'stepping stone*, Di, distribution; R.E., reduced eyes, Dep, depigmen- Centromerus to Madeira. Troglohyphantes came tation; Leg, long & slender legs; T, Total. perhaps from Spain to Tenerife, but the sister species is unknown. Walckenaeria came from North Africa to Tenerife, Agraecina came from Europe or North Africa to Tenerife (and Gran Canaria). The occurrence of cave-dwelling species of the Dysderldae different genera on Tenerife - e.g. Walckenaeria, heridiidae Troglohyphantes and Agraecina - are remarkable (Fig. 1). The highest Macaronesian mountain, the Linyphiidae 3718m high Teide on Tenerife, seems to be a 15% Pholcidae 'catcher' of aeronautic (ballooning) spiders 13% which came from the Western Mediterranean Lycos idae area. This finding is supported by the relation- 5% ships of the endemic spider fauna of the Teide and Dictynidae the Canadas, an area surrounding this high moun- 3% Gnaphosidae tain (cf. Wunderlich, 1991: 104-107). 5% Evolution of Macaronesian Cave-dwelling Spiders In some spiders, e.g. Canarionesticus quadridentatus and Troglohyphantes oromii, no re- epigean FIG. 3. Composition of Canarian families of lated epigean species is known, and almost spiders on than 400 species. based more nothing can be concluded about their evolution. . 1

684 MEMOIRS OF THE QUEENSLAND MUSEUM

cFoV^

i .

, ; - , 10 6b /'^ / 9 «

.,:....-, juAiuiotli

11 ;

t *?n j«? or. 1

FIGS 4-7. 4-5. Variable eye reduction in subterranean Canarian Agraecina canadensis Wunderlich, 1991

(Liocranidac). 6, 7. Rugathodes. 6 , 8 body with eyes FIGS 8-1 1. Eye reduction in Canarian cave-dwellers and right tibia I from the Azores: 6a, of the cave- spiders: 8, Spermophorides rfven/owWunderlich, dwelling R. pica: 6b, of the epigean R. acoreensis. 1991; Fig. 9, S. fuertecaverisis\ 10, S.jlava Wunder- 7. Right male pedipalps ventral: 7a, R. bellicosum lich, 1991; U.S.justoi Wunderlich, 1991. 12. Eyes (Simon, 1873), Europe; 7b, R. madeirensts Wunder- of the epigean S. tenoensis Wunderlich, 1991, lich, 1987, Madeira; 7c, R.ptco (cavcrnicolous) and Tenerife. 13. Sometimes completely eyeless prosoma R. acoreensis (epigean), Azores. These species show of cave-dwelling Tragiohyplxantes oromii, Tenerife. no differences, (E = embolus).

possibility that this species is a paleoendemic (Several species are not well studied, e.g. Dys- relict (Peck, 1990: 372-373). dera species, or only one sex is known). In this Based on the very similar genital organ, in connection Troglohyphantes oro/nii is of special Rugatkodes pica and acoreensis they arc identi- interest because nearly all species of this genus cal, I found some strongly related spider species are troglophilic or troglobitic cave-dwellers as (Table 3). In the cave-dwelling R. pico (Table 1 well as the species from Tencrifc. I can imagine and discussed above), the adaptations to cave life that there has never existed an epigean stem and the differences in some non-genital structures species on Tencrifc and that the ancestors, per- compared with epigean species are very distinct: haps as a cocoon, has been transported by a bat the eye reduction, the depigmentation and the directly from a European cave to a cave on prolongation of the legs, Fig. 6a; compare Fig. 6b Tenerife, Several species or subspecies of bats fly of the epigean species, which is also known from from Europe via Madeira to the Canary Islands the Azores. Otherwise the genital structures in the (Dr. Biscoito, Mus. Munic. Funchal, Madeira, two Azorcan species show no differences in both pers. comm.). This hypothesis would explain why sexes (Fig. 7c), but there are distinct differences there is no epigean Troglohyphantes spp. on to madeirensts from Madeira (Fig. 7b) and bel-

Tenerife or another Canary or Macaronesian Is- licosum from Europe (Fig. 7a). So I do believe land. Otherwise 1 do not want to exclude the that Ft. pico and R. acoreensis are true sibling ;

MACARONFS1 AN CAVE-DWELLING SPIDFRS r-HS

;i.ir..Ni>_-_,, , ._,. ,.-: Paleaoendemic species Ntocndemics

The epigean sibling or sister species/stem species b known, usually from tame island

Ctinariunesticus quiitiridentatm iNtslicidac, 1 > this group belong most Macaronesian Mem boumeti Tent-rife). No known relative. cave spiders, >cr Mow (genesis/evolution), Dysderidae, Pholcidae, Linyphiidae, TheridiidAc and UoiTtuiidac.

TABLE 2. Historical groups of Macaronesian cave-dwelling spiders. Position of Troglohyphantus unmtit tfi list

is uncertain; it ha* no known epigean relative. See below; possible transport by a bat. species (or even subspecies?) and that R. Spermophondes, Lepthyphantes, Walckenacria, acoreensis also is the stem species of R pUo, Altdia, Zimirlna and others. These spiders have In SpermopkorideSt differences in genital ami moderately reduced eyes and are more on less non -genital structures arc distinct in both sexes, depigmented, they seem lo be troglophilous and and perhaps they are not sister species, The noimacrocaveTmcolous, but microcavernicolous remaining species in this list arc known only from or myrmceophilous. Agraeci/ta t unariensis is a females, species of this stratum but it aJso penetrates caves. i'A has dis- Marti n et al. ( ] 989} assume thai some Canarian closely related species been newly troglobites can be considered as relict species covered in a Romanian cave). which evolved after changes in the climate 'since 2. Another preadaptation is offered in the con- there have been alternating wet and dry periods ditions in Ihe laurisilva (the relict laurel forest);

..causing important changes in the fragile insulai 1 humidity and low changes in temperature. ecosystem on the surface/ (See below For example 1 found the lamisilva species Rugathodes). Furthermore the yearly seasonal Troglonaia madvirensis Wundcrlich, 1987 changes-hot and dry summers, cooler and more (Aftapfdae s.l.) in the wet and light part near the humid winters especially on the Eastern Canary entrance of a cave on Madeira (Sao Vincentc). I Inlands—did perhaps initiate vertical movement of collected i V of Lepthyphanfes mauli Wunder- some species into the ground in the summer and lieh, 1991 < Linyphiidae) in the same part of this also autecological changes for instance in cave; this species probably also came from the Centromerus fuerteventurensis Wundcrlich, wet forest near the cave.

1991 (Linyphiidae). In this species, which is nol 3. The third kind of niotopes that lead U>

.1 cave-dweller, the eyes are reduced as well as in troglobitic conditions are under stones and under species of Scotargus (Linyphiidae) and Altelta leaves on trees at places with a high humidity nwr (Dictynidae) of other Canarian islands (Wunder- stretches of water. At such places, under stones lich, 1991). as well as under leaves twi trees. I found on the iicoree/tsLi, Borges and Oromi ( 1 99 ) do prefer the ^adap- Azores R the epigean sibling/stem 1 tive shifting theory' of Howarth (l°73) 'This Species of the cave-dwelling R. pica. Under a theory does not invoke isolation during climatic stone over flowing water on Tcncrifc I found lor volcanic?) changes but instead proposes that partly depigmented Spiders 6f WakMmtrkl the partly adapted anceslors shifted inio newly .>i!ha developed niches.' In my opinion this theory can well explain the genesis of some Macaronesian Agog 01 M ^CARONBSIaN Cavk-dwitlling Spider rave-dwelling spider species including the ones Specks listed above (sec Peck, 1990: 366-368). Especial- The age ol the Macaronesian caves remains ly there exists large caves and lava rubes on the unknown. The youngest Macaronesian Islands, Canary Islands, and there arc many ecological the Western A/orean and the Western Canarian niches. Islands, are only very few (1-2?) million years From the Macaronesian Islands and us spiders old Two spider species. PholcUS nufitat.iis and

I know three kinds of preadaptations which sup- Walekcmieria cavernirola, are cave-dweDeis al port the 'adaptive shifting theory' the Canadas on Tenerife. The greatest ag|

1. A lot of Canarian species arc known as these species could be the same as the age of this hypogean spiders and were caught using special part ofTenenfe, at least 200 000 (up U>2 million) traps in the ground in the so-called 'mesocavcr- years shallow stratum' (Wunderhch, 1991: 11). RpgQihadcs pico is known only from caves of FrOm This stratum and from accidental captures the Azorean islands Pico (a young island) and

litl. 1 these . undertones I know Canarian species of Dysthn,. FajjI. After Ashmoleiin 991) Azi 686 MEMOIRS OF THE QUEENSLAND MUSEUM

TABLE 3. Macaronesian Cave-dwelling species Epigean relatives Island Family cave-dwelling troglo- Spermophorides fuertecavensis S. 'fuerteventurenxis Fuerteveniura (Canary Is.) Pholcidae philic or troglobitic

Centromerus sexocutatus, and sp. nov. C. variexatus Madeira Linyphiidai* spiders and their nearest relatives, Lepthyphantes palmeroensis Lfurcabill\ La Palma (Canary Is) Liny phi idae epigean possible stem/ sister/sibling ',.. ('; Walckenaeria cavern'scola W leidfen.six '._='-.t. ;-::;-> is. 1 Linyphiidae species from same is- Rugathodes pica R.acoreensisi Pico and Pajal (Azores) Theridiidac land.

Islands had perhaps a land bridge during the last for his kind comments to the first version of the glaciation. So this species probably evolved from manuscript. its closely related epigean stem species, near or identical with R. acoreensis, perhaps not later LITERATURE CITED than 10 000 years ago (= 10 000 generations). From genital structures, in both sexes there are no BORGES, P. & OROMI, P. 1991. Cave-dwelling differences, speciation should have happened, ground beetles of the Azores (Col., Carabidae). geologically, not long ago, that means at the end Memoires de Biospeoiogie 18: 185-191. HOWARTH, F.G. 1973. The cavemicolous fauna of of the last glaciation. But this idea is very vague;

Hawaiian lava tubes. 1 . Introduction. Pacific In- perhaps the speciation happened much later, and sects 15: 139-151. R. pied was transported by bats from one island MARTIN, J.L. et at. 1989. Sur les relations entre lc to the other only very few thousand years ago troglobieset les especes epigees des lies Canaries. (Wunderlich, 1991: 200-201). Wunderlich Memoires dc Biospeoiogie 16: 25-34. ( 1991) gives further details of these cave-dwell- PECK, S.B. 1990. Eyeless arthropods of the Galapagos ing spiders and their taxonomy- Islands, Ecuador: composition and origin of the cryptozoic fauna of a young, tropical, oceanic archipelago. Biotropica 22: 366-381. ACKNOWLEDGEMENTS WUNDERLICH, J. 1991. Die Spinnen-Fauna der Makaronesischen Inseln. Taxonomie, Okologie, Biogeographie und Evolution. The spider fauna of I like to thank the Deutsche Forschungs- the Macaronesian Islands. Taxonomy, Ecology, gemeinschaft, Bonn, for the loan of optical instru- Biogeography and Evolution. Bcitraege ments, and I also like to thank S.B. Peck, Ottawa, Araneologie, vol. 1, 619 pp. RELATIONSHIP BETWEEN FOOD INTAKE AND SPIDER SIZE IN TEMPERATE ZONES: EXPERIMENTAL MODEL FOR AN ORB-WEAVING SPIDER

FREDERIC YSNEL

Ysnel, F. 1993 1111: Relationship between food intake and spider size in temperate zones: experimental model for an orb-weavinc spider. Memoirs ofthe Queensland Museum 33(2): 687-692. Brisbane. ISSN 0079-8835-

A method is presented that allows calculation of the energy intake by an orb-wea vi ng spider,

Larnnoui^s \ omutus { Araneae: Araneidae) under natural conditions throughout the spider's life. A laboratory study provides several relationships between individual energetic con-

sumption and size of spiders depending on the thermal conditions of their environment. I observe a preferred temperature (21°C) at which spiders have ihe biggest consumption. A model (pseudo-cubic spline) is constructed for the calculation ot energy intake by each juvenile instar. Energy requirements of the adult population are estimated from the repor- duction rate observed in the Held. The energy requirements under natural conditions and the total weight of prey consumed by the population in the course of hs biological cycle can be infcJTed. In the mesophUous heathland investigated, the total fresh weight of prey consumed

by the population during the life eycie is 18.2kg.ha' . Ccttc etude vise a relier les laux dc survie, de croissance et de reproduction d une espece Orbitele Lartruoides comutus (Araneae: Araneidae), au nombre de proics consommecs en milieu naturel au cours du cycle biologique. Cette analyse est ddduite d une approche bioenergetique. Au iaboraU'ire. !es consommations energcliqucs el la croissance des

araignees sont testees en foneiion de I environnement thcrmiquc. Ccs experiences metteni en Evidence la presence d un optimum ihermique d ingestion qui modi fie la croissance et la consommation des individus. Un module d ajustement (spline pseudo-eubiqui.-) liant la

temperature, la laille des araignees et la temperature ambiante est propose pour cslimcr I energie ingenue en phase juvenile. L energie ingcrcc en phase adultc est eslimec en eomparant les parametres taux de reproduction -consommation calcules en elevage, au taux de reproduction observes en milieu naturel. Les besoms cncrgciiques sont ensuitc convenis en quantitc

de proies capturces, Sur la lande mesophile etudiee, la population capture en poids frais, 1 8.2 kg de proies par hectare au cours du cycle biologiquc .\Z\Population energetics, Araneae, Araneidae, Lari/iioides comutus.

Frederic Ysnel, Laboraioire d'EvolutUm des Systemes Naturels ct Modifies, URA 696 du

CNRS, Campus de Beaufieu, University de Rennes /, 35042 Rennes Cedes, Frame; & March, 1993.

Among the many studies already carried out However, no study has been done in temperate about the trophic spectrum of the araneids climates to link the characteristics of population fNenlwig, 1987), some suggest that the popula- dynamics of spiders with the quantity or the lion of spiders-wandering spiders or non-migrant quality of trophic resources. My work on spiders-ean utilize a significant proportion of the Larinioidex cornutus (Clerk, 1758) (Araneae: secondary productivity of natural areas (Kajak, Araneidae) (Ysnel 1989) describes demographic 1967; Robinson and Robinson 1970; van Hook, evolution and reproduction rate of a population a mcsophilous heathland in western France 1971; Blandin and Celerier, 1981; Nvffeler, "V (Ysnel in press). Besides this, laboratory studies 19821. Moreover, some studies emphasize that showed ' heTm nd W thegrowthincrement(VollrathJ9S8Uherateof « ° A f .^ ?? *# . . si/c Tf^ ,,, and food intake (Vsne, 1990). This study , r , n .fv,, rk-iA\ .l. ; reproduction^Riechcr...1974; W«c. 1979) or the ^ combine resu|t s tQ m ' ^^ ^ ^ density of individuals (Kajak, 1977) observed in _ the n mbcr of cv ^ d bv a namra , , a the populations can Fluctuate with the number of - tjon of idcrs du jls dcve f opmen , prey caught. Thus, the characteristics 01 population dynamics of spiders partly depend on the MATERIALS... „.„„, ., A AND.»*— .METHODS.^.™„^,-, quantity of prey captured; they can R ivc data about the secondary productivity and consequently about the biological resources of natural This study is based on the calculation of the biotopcs. energy intake during the postembryonic develop- 3

688 MEMOIRS OF THE QUEENSLAND MUSEUM

13°C 16.5X 21°C 26°C Inslar n C±SD n C ±SD n C ±SD n C±SD J2 _ . 7 19.03(5.03) 8 18.75^6.38 7 12.70(3.3) TABLE I: Average

7 S 50.54(24.39) 2K 74 i 13.971 J3 ! 44.60(11.25) 7 values of consump- u • 9 116.70(26.9) 14 83.88(37.86) 12 74.900753 tion (C, joules) during each juvenile JS 2 165 9 241.62(71.75) 13 134 00(32.5) 9 181.66(111.9! inslar. (n = number Jfi . . 7 471.95(190.12) 12 333.90(86,27) 9 595(196.56) of individuals n J 7 6 143M(3?3) 7 1999.14(317.15) i 647.70(166.65) 3 668.39(22.55) tested, SD = stand- L i 2581.5 18 4 1 120.69 (456) ard deviation).

- - 1 1029.6 19 !

H l |i.-X.i un aulas) C [) V U.-S :

l B : 4S, ( = DM

t 31 \ '..-.. . il , . J, r.

:': 16"*'

1 n 50. r (KV-ll

FIG. 1. Theoretical relationship between spider size

and consumption in intermoult periods (n = no. of FIG. 2. Pseudo-cubic spline showing relationship be- measurements, r= correlation coefficient). tween spider size (T PI V), consumption in intermoult period (C) and air temperature ("PC). ment of L. cornutus. To estimate energy requirements during juvenile development, I refer to vomitaria and Lucilia seriedta). Three times a results on the individual energy consumption of week, all spiders received a fixed number of prey spiders at different temperatures (Ysnel, 1990); according to their age. Juvenile spiders were kept thus, only the main references of the experimental under a light regime of LD 12. 1 2. Furthermore, 8 conditions are described. Young spiders from spiders collected in the field at stages 4 and 6 and cocoons reared in the laboratory were divided were reared in the laboratory at 13°C. For these into three groups by temperatures- 1 6. 5°C. 2 PC spiders, the energy consumption was studied after and 26°C. (These values are a little lower than the first moult in captivity. those in my previous work; they agree with more As adults, only females feed (adult males no accurate values using an electronic ther- longer build webs). In nature, females can mate mograph). Spiders required a varied diet for sur- in autumn but cgg-Iaying occurs only under long vival therefore they were reared to maturity using days from May to August. During this period, three prey species: the first instars were fed with females lay, on average, only one egg-sac and adult Drosophila melanogaster and the last in- disappear soon after laying ("Ysnel, in press). stars were fed with callipnorid flies (Calliphora Therefore, energy consumption is estimated in

13°C 16.5°C 16°C Inslur TABLE 2. Average ]t±sd n y ± iu a T±SD in T±SD K.W. lengths of tibia IV (T, j 12 - 7 0.27*0.033 B 0.28 ±0.024 CL27±Q.026 P<.0J mm) and comparison of values for I6.5°C. 7 0.37 ±0.0 IS 8 0.37 ±0,032 P<0.2 21°C J3 040±(UM3 9 a and 26 C tn = no. of - t 7 14 6 0.56 ±0,07 047 ±0. D5 9 [) 3H±O09 P<0.05 measurements, SD = ') 15 2 0.87 o,9i ±o.u: 13 0,'>i-0 0S 10 0.86 ±0.1 iVtHiK standard deviation K.W. . : J6 6 1.24 ±0.186 10 0.92 ±0 105 9 1 42 ±n P<0.08 = Kruskall- Wallistest).

-'.: 1 '- 17 6 1.68 ±0.1 7 .. ? 11 1.21 ±0.143 3 l.34±h.l>4 fVO.IU

_ - IS 1.49 ±0.11 1 |l.89

- K' |- I 66 RELATIONSHIP BETWEEN FOOD INTAKE AND SPIDER SIZE 689

13°C 16.5°C 2\°C TABLE 3. Consump- tion of 9 and no, of r ( Yegj: T Nl C | C/egg Nl C J Nl C C/. g? eggs /cocoons at 3 1083.15 2.25 115 9.41 2.15 82 772.85 9.425 2.13 78 797.1 10.2 temperatures. (T = length of tibia 4; Nl 1215.20 2.32 131 9.27 2.40 95 679.20 7.14 2.10 87 719.3 8.26 = no. eggs/first egg- sac; C=energy intake 133300 2.50 144 9.25 2.28 96 700.40 7.30 2.40 140 1338.9 9.50 by / in joules; C/egg =energy intake for 2.02 108 948.70 8.77 2.13 50 569.13 11.38 1.87 40 600.71 2.01 pro-duction of 1 .. - Z03 105 2.10 75 \ 80 50 644.31 2.90 egg). 2.47 12K - 2.24 65 - 2.02 57 723 81 0.05

2 4? m 1.98 58 871.8 15.03

2 43 126 - -

. .'.,-: . i ..i .'] | , < i 11 i N |

1S00

s labwaiery dfla (3j

1 Ltiata(t)i

FIG. 3. Relationship between size or females (T) »nd energy (C) required to elaborate first egg-sac. adults based on the energy intake needed to FIG. 4. Relationship between no.of eggs (N) produced produce one cocoon. With this aim, 22 over- in first cocoon and size (T) of females. wintering fecund females were collected and cohorts (Ysnel, 1992). For each juvenile instar, reared in the laboratory under long days (LD =16 the energy consumption was inferred from the : 8) and different temperature conditions (13°C, average size of spiders (length of tibia TV) and the 16.5°C, and 21°C). After the females oviposited, thermal environment in the field. Daily we determined the relationship between the ener- meteorological data were classified in four gy intake, the female size (length of tibia IV) and temperature classes (T< 16°C; 16.5°C 22°C) and the average instar the adult, the For each juvenile and ener- value of each class calculated. The energy in- difference be- gy intake was determined by the gested is determined from the relative proportion the tween the energy in whole captured prey and of each phase of temperature during one inter- that in the food remains. Calorific determinations moult period The average value of the caloric for all prey and food remains were made using a a gfl '" i i ! ' u Parr bomb calorimeter (Ysnel, J 990). r< 16 T From the phenology of different instars in nature, the developments of the species have been worked out (Ysnel, in press). Spiderlings emerging from egg sacs laid from late spring to early summer become adult before winter and form a first cohort of individuals (cohort CI). Spiderl- ings which appear later (end of summer) form a second cohort in the population (cohort C2); they arc still immature in winter and become adults early next summer. In analysing the demographic evolution of the population (from 20 m sample FIG. 5. Relationship between energy required to areas) during the life cycle, the number of surviv- produce one egg (C/egg/JouIcs) and size of 9 (T). ing spiders per instar was counted for the two 1

h90 MEMOIRS OFTH£ QUEENSLAND MUSEUM

table 4. of Method 1 Method 2 Comparison consumption (C. joules) ol Size JTNEL cm (M-. TNEF C(Jj (field) Size TNEF[c/cggr.?j- 173.8 1421 4 173,8 111.05 Theoretical no. of egg*

1 2.3 107.8 9S i : 217.0 1981,6 217,0 9.02 1957.8 (laboratory, TNEL>, in

> 2.6 147.5 I230.fi 260.9 2176.4 |2.6 260 8.30 2165 1 (field, TNEF).

individual Lonsumption equivalent of insects captured in the field is 22.6

./eight (Y). 1 used the conversion factor calculated for calliphorid flies (Y = yJ0262: Ysnel, 1990).

RESULTS AND DISCUSSION

Individual Consumption, Sptder Size and juvenile Thermal Conditions of Juveniles FIG. 6 Estimated individual consumption per instar under field conditions, The number of instars needed to reach maturity varied with the growth conditions. In Araneidae. highest consumption but the growth increment is the first free-living instar is called J2 (Canard, paradoxically the smallest. From instars J4 on- D 1 987). Prom S to 8 immature instars may various- wards, individuals kept at 21 C are significantly ly precede adulthood. Average consumptions per smaller in size than individuals kept at 1 6 5V instar are calculated for four temperatures (Table 26°C (Kruskall-Wallis test, Tabic 2). Using Ihe 1). Average consumption also varies with instar energy budget equation of Pctnisewicz and Mac- and growth conditions. In each group, an overall Fadyen (1970), the energy dissipated as heat connection between the instar of development of (respiration) could be increasing (while the ener- a spider and the energy intake of that instar cannot gy retained a^ growth is decreasing) and the be deduced because individual variation is con- spiders must be more active at 21°C. In contrast, siderable. However, for all individuals, whatever when the temperature deviates from the thermal the instar, there is a good correlation between preferendum, the energy allowed for the activity individual consumption and spider size the inter- of spiders must be less important (probably be- moult (Table 2). For each temperature with all cause spiders progressively fall into quiescence instars merged, we can fit all the observed values or aestivation) and the energy retained as growth to four exponential graphs corresponding to four increases. different equations (Fig. 1 ). The four curves show One can estimate the energetic requirements of l hat the individual consumption varies with temperature. For the tested temperature range, spiders C (p)(x, y), by interpolating the pooled and spiders of similar size, I observed a preferred data to a single regression model linking spider temperature of consumption' (2 1°C) at which the size and air temperature during the intermoult, to energy intake at the intermoult is maximal the individual energy consumption. The method, (Ysnel, 1990). based on the 'homogeneous spline functions with Near the thermal preferendum, spiders have the several variables' (Duchon, 1980) allows us to

Cohort '."'! Cohort C: Population TABLE 5. Estimated mean energy

' Instar Size I md. n CI \sizt Cind. n C2 N/nrf ( ).!/!:. consumption per juvenile instar in

0,260 247 7528 0.26 il 159 6996 20.3 n:-h 1 2: n |30.5 J field conditions of individuals

J ' >' ^:7 13 127 o.4n li 1 U.JM |j h 38 92 5S3I 1 648 0.98 (Cind.). all Cohorts 1 (CI) & J

11 (C2), & of total population 11 M S5S 60.9 102 6212 1 1 > 7 101 8 29 23SI 6,6 9164 0,77 tC).

i.,-.'i •>>- =abuudancc of different insiars; 13 IH7.7 16240 0.7S 15ft 28 442-1 5.7 20664 1 74 N/nr= spider density; kg/ha

1 J6 1.140 1 264.8 ;7I 1K80I J. 14 400.54 , in 1 ; 2.50 I 27 4,9 29216 j =mean total weight (kg/ha ) of food 17 1 442.3 1 18 7961 | 1.36 641.K |7 4493 12454 D.s 1.23 consumed/instarU Ysnel. 1992). . )

RELATIONSHIP BETWEEN FOOD INTAKE AND SPIDER SIZE 691

1 sizci SD N C/eg£ E/Fn \c - method 2: In nature, we can estimate the

'. ill." I 1:945*0. IE 169 1931.7 24 4h7f ! number of eggs produced by females using the Cohort C? 1.91 ±0.13 \M 11.66 1912.2 |37 7(1751 regression equation I (Fig. 4). The energy consumption is then calculated by referring to the TABLE 6. Estimated mean energy consumption of 1 values of ^C/egg given for the females of the adults (N = estimated no. of eggs/cocoon; E/egg = same size (Fig 5, regression equation IT). energy intake to produce one egg, E/F= energy intake Theoretical values of the consumption of by 5 ; n: = no. ? 5 per cohort; C = total consumption females derived by the two methods are similar i / cohort). (Table 4). Hence, the energy required to produce adjust significantly Ihe whole experimental one egg is not influenced by the feeding regime values to a pseudo-cubic spline (Fig. 2) by means of \piders- The number of eggs per cocoons is of the following formula : dependant on female size and on the quantity of food intake. We use the ratio C/egg to estimate Ip)((l -u):+(v - V()1 C(p)(x,y)=E to ^+alAH

spider size at intermoult, y = air temperature, ai, EKFfcflY RFOUIRFMr-NT.SIM NATIIkl otl, a2 and (3 are calculated by the solution of a For individuals of the same juvenile instai linear system. consumption vanes with time of year of the This model estimates the energy requirements spiders appear (Tabled). Individuals of cohort C2 during an intermoult period of any juvenile instar. have, in most instars, an higher consumption than under these experimental conditions. those of cohort CJ although the average size in the two cohorts is similar (Fig. 6). Hence, the Energy Consumption during Adult Phase spiders of cohort C2 live closer to the thermal The individual consumption of all females prcferendum than those of cohort CI is plotted against to the caloric equivalent prey (Tabic 3) spider gjze | Fig. 3); According of the the regression line can be used as a calibration caught in natural areas (22.6 J/mg dry weight), we curve to estimate the consumption of females can estimate the food consumption of the- orb- independent of thermal environment. The num- weaving population. In the mesophilous heath- ber of eggs per cocoon also showed a significant land investigated, the total fresh weight of prey linear relationship (Fig. 4) with female size- consumed by the population in its biological 2 However, die number of eggs in cocoons in the cycle is 0.036 kg/20 m which amounts lo 18.2 laboratory (line 2) and those in nature (line 1) kg/ha. The amount of food ingested varies with clearly indicate that breeding females produce the hsstiur. The high mortality rate (Table 5) fewer eggs in the laboratory. Two hypotheses can reduces total consumption in the first three be advanced to explain this result: juvenile instars. In later instars. although the

- the food consumption of breeding females in spider density decreases, food intake by the the laboratory is lower than that of females in population increased owing to a concommilant nature; consequently fewer eggs arc produced m increase in individual consumption. The quantity the laboratory. of prey caught by adult females is higher than for

the whole juvenile population ' - for females of the same size, the number -f (Table 6. Fig eggs produced in nature is favoured by a varied diet. CONCLUSIONS The ratio 'energy consumption/number of eggs produced' gives an index (C/egg) of the energy The species responds to variation of the trophic required to produce one egg (Table 3). The values conditions in both body size and the number of 1 of 'C/egg are dependant on female size I Fig. 5). eggs produced. These two biological parameters Two methods arc now used to study the energy constitute indices of" reference to estimate the consumption of females. quantity of prey caught by Ihe population during - method i: For females of a given size, the the biological cycle, Compared with the results regression equation I (Fig 3) allows the calcula- nf others (Table 7), the food intake by the population of the energy consumption to produce a given tion of Larinioides conwtus represents from XT number of eggs. When females produce more 36.4<& of the total food consumption of the spider eggs (as in nature). Wt estimate the energy con- communities of different ecosystems. In other sumption using a simple rule of 3. respects, the food ingested in the biotope Studied 692 MEMOIRS OF THE QUEENSLAND MUSEUM

— Bioiopc Prev killing rate (kp/ha/vearl Countrv DUCHON, J. 1980. 'Fonction splines homogenes a

' > : ' plusieurs variables (These dc Doctorat d'Etat, 1 Reed beJi of lake i usiria Universile Superieure de Mathematiques Grassland 52 I. S \ Generalcs, Institut National dc Pans Grignon.) Grassland >I50 Swfl/jcrland KAJAK, A. 1 967. Productivity of some populations of Grassland i:m Poiatd web spiders. Pp. 807-820. In K. Pelrusewicz;

. [Forcn tea Il ; maiq 'Secondary Productivity of Terrestrial Ecosys-

J Spartino swamp 1 2. 5 OS-A. tems.' (Warsawa). I 1977. Analysis of consumption by spiders under TABLE 7. Prey killing rate by spider communities of laboratory and field conditions. Ekol. Pol. 26: ecosystems (after Nyffeler, vegetation of different 409-427 1982). NENTWtG.W. 1987. The prey of spiders. Pp.249-263, I'Oort intake In W. Nentwig (ed). 'Ecophysiology of Spiders.*

ins i tw; (Springer: Heidelberg). in NYFFELER, M. 1982. Field studies on the ecological role of spiders as insect predators in agrusvstcms. (D.Sc. Thesis. Swiss Fed. Inst Techno!., Zurich, Aku-Photodruekl PETRUSEWICZ. K. & MACFADYEN. A. 1970. Productivity of terrestrial animals. IBP Handbook 13 (BlackwcH Scientific Publications: Oxford). RIECHERT. S.E. 1974. Thoughts on the ecological signifance of the spiders. Bioscience 24: 352-356 ' "Pt^TT^TTrg ROBINSON, M.H. & ROBINSON, B. 1970. Prey J2 13 j4 ,]5 Muds town caught by a sample population of the spider Ar- giopeargtmtafti (Araneae: Araneidae)in Panama: FIG. 7. Food intake by population in biological cycle. a years census data. Zoological Journal of the Linnean Society 49: 345-358. i% five limes as important as the food ingested by VAN HOOK, R.l. 1971. Energy nutrient dynamics of the population studied by Kajak ( J 967 ) in a Polish spider and orthopteran poulalions in a grassland grassland. This is the only analysis of the charac- ecosystem. Ecological Monographs 4 i : 27 1 28$. teristics of population dynamics of an orb- weav- VOLLRATH, F. 1 988. Spider growth as an indicator of ing spider in a temperate Tegion and emphasizes habitat quality. Bulletin of the British Arach-

1 7-2 9, the importance of die secondary productivity. nological Society 7: 2 1 WTSE, D.H. 1979. The effect of an experimental in- ACKNOWLEDGEMENTS crease in prey abundance upon the reproductive rates of iwo orb weaving spider species (Araneae: AroneidaehOecologia, Berlin 41: 289-300, 1 wish to express my most sincere thanks to Dr YSNEL, P. 1989. Caracterisaiiondes stades dc devclop- Y. Lafranche (Institut de Recherche en Math- pement dc Lanniuides cumulus (Araneae. Ar- cmatiquesdeRennes). who greatly contributed in * uiopidae) par la trichobothriolaxie. C.R. XI the mathematical treatment of data and helped in EuropaVsehes Arachnologisehcs Colloq- the production of the pseudo-cubic spline. I Tub-Dokumcntation Kongrcssc und Tagungen should also like to acknowledge Pascal Dahyot- Berlin 38: 72-78. CornSe who improved the translation. 1990. Energy consumption of the orb-weaving

spider barmoidci comutus ( Araneae. Aran- LITERATURE CITED eidae) during the postembryonie development under different temperature conditions. XI Inter- national Congress of Arachnology, Turku, Fin- BLAND1N, P. ft CBLERIER, ML. 1981 Us land. Acta Zoologica Fennica 190: 409-414,

i\ I Araignees tfes savanes de Lflmta (Cote d i Organisation des peuplemcnls, bilans 1992. Impact trophique et valeur bioindicairice energeuques, place dans reeosvsteme. Pubis Lab. d'une population d' Araignees: exemple d'une ZiOOl. E.N.S. Pans 21 (2 fuse.):' 1-586. esp&ce a toile georo^trSquc LariHlottfes cottitttus LANARD, A. 1987. Analyse nouveilt du (Araneidao) (These dc Doctoral d'Universitc". developpemcnt rX'Sternhryonnairc des Araignees. University de Rcnnes L France). 217 pp. Revue ArachnoIogiLjue 7: 91-128. (in press). Data points for a study of population

CUMMINS, K.W & WUYSCHECKJ C 197 1 . Calor- dynamics of an orb- weaving spider {Lannioides cmc ic equivalents for investigations in ecological cornutotx, Araneae. Araneidae). Actes du XlII energetics. Mitt. Internal. Verein. Limnol, (16): ii>que Europeen d'Arachnologic, Neuchatel. 158 pr Suisse, 1991 STUDIES ON THE WOLF SPIDERS (ARANEAE: LYCOSIDAE). I. A NEW GENUS AND SPECIES FROM KAZAKHSTAN. WITH COMMENTS ON THE LYCOSINAE

ALEXEY A, ZYUZIN

7yi)/in i 7 A A. 993 11 11; Studies on the wolf spiders (Araneae. Lycosidae). L A new genus and new species from Kazakhstan, with comments on the Lycosinae. Memoirs of the Queensland Museum 33(2): 693-700. Bnsbanc. ISSN 0079-8835,

A new genus Ocidicosa (Araneae: Lycosidae: Lycosinae) is established and a new species (hulhosa supermirabilis, sp. nov. is described from South-western Kazakhstan. Relation- ships of the new species are analysed. The role of some morphological features of burrowing lycosids is specified. Tegular (median) apophysis in all members of the subfamily Lycosinae ts declared to serve as the functional conductor of the embolus: the mechanism of its action

« ilyscd. The role of some structures and the use of corresponding names is specified, the comparison of the main subfamilies is given. The structure of the tribe Trochosini Zyuzin. 1990 is revised: this tribe is divided into two subtribes including non-burrowing and burrowing forms. On etablit un genre nouveau Ocidicosa (Araneae; Lycosidae: Lycosinae) et decril une espeee nouvelle Ocidicosa supermirabilis sp. n. du Sud-Ouest du Kazakhstan. Les affmites de lespece nouvelle sont analysees. On precise !e role des quelques indices rnorphotogiques dea I.ycosides creusanls. On a determine que 1" apophyse tegulaire (median) chez tous les membres de la sous-lamillc Lycosinae sen de conducleur fonctionne! d'cmbolus: le

mecanisme de sont action est analyse. On precise 1c role de quelques structures el I' usage des noms correspondanls. la comparaison des prineipaks sous-families est donnee. On fait une revision de la structure de latribu Trochosini Zyuzin. 1990: cette tribu se diviseendeux sous-tribus inserant des formes non-creusantes et crcusantes. \jLycosidae, Ocidicosa, conductor.

AlexeyA. Zvuzin, Institute ofZoology, Kazakhstan National Academv of Sciences, Akitdem-

gorodokt 480Q32 ALMA-ATA 32, Kazakhstan Republic; I H March, 1993.

Burrowing spiders of the family Lycosidae tion, dissected tegular apophyses of males were within Kazakhstan are poorly known. Records attached lo female epigynes (Fig. 5). Eye meas- exist only for Allohogno singoriensis (Luxmann, urcments are given in eyepiece micrometer units \11Q\ Lycoui nonimanni (Tiwvcll, 1875) (= L. (x32). Genitalia and their parts for scanning narbonensis auct,, non Latrcille. 1806), and L electron microscopy were preserved in ethanol. alticeps (Kroneberg. 1875) (see Charitonov,. air-dried, mounted on stubs, gold-coated and c\- 1932; Dubinin, 1946). However, of those, only annned in a JEOL JSM-T200at i5kv. the distribution in Kazakhstan of A singoriensis Abbreviations: ALE, anterior lateral eyes; has been well studied (see Marikovskij, 1956). AME, anterior median eyes; ARE, anterior row The only large non-burrowing lycosid reported of eyes; AZ, private collection of the author, L, from Kazakhstan is Hogna radiata (Latretlle) leg; LC, carapace length; P, palp; PLE, posterior (sec Schmidt, 1895: Aral Sea and Mangyshlak lateral eyes; PME. posterior median eyes; TA. Plateau). That report is very doubtful, as the tegular apophysis; VVC, carapace width; ZMMU. author studied juvenile specimens only. Zoological Museum of the Moscow University

MATERIAL AND METHODS MORPHOLOGY AND DISCUSSION

Spider material I collected in 1989 tn Oculieosa gen, nm Karynzharyk sands. South-western Kazakhstan, are used here. Specimens were captured at night Type SPECSES with the use of a miner's head torch: spider's eyes Oculicow suprrmirabilix sp. nov. reflect torch light at the distance of 20-25m and even more. Spiders were examined in 70% al- Diagnosc- cohol using hmoeular microscopes. To under- Medium size (body length 12-20mm). stand the relative position of different parts of the Cephalothorax: head strongly elevated, thorax in male and female genital apparatus during copula- both sexes behind ALE is evidently descending 694 MEMOIRS OF THE QUEENSLAND MUSEUM

p| f_: \ e Mr AME L E

. 'iVH- j 0.93-0.98 0.33-0.36 0.23-0.24

PLE 1.02-1 07 0. 34-0.38 0.23-0.24

ami: 2.80-3.00 2,67-2 93 - 0.63-0.71

- A) i 4.204,42 1 4. 10-4.30 1.40-1.58

TABLE 1. Relative size of eyes in Octdicosa super-

mirabiiis, sp. nov. (26, 2 ? )

towards abdomen (Fig. l) t lateral sides of head are almost vertical. Carapace is relatively narrow, ratio LC/WC is 1 .44- 1 .50. PME and PLE are very large, height of ocular field is 0.39-0.42 of LC (Fig. 2). ARE recurved, AME larger than ALE

(Fig. 3). Row I is 1.64-1.70 times shorter than row 2; row 3 is 1.19-1 .23 times wider than row 2.

Clypeus narrow, its height less than 1 diameter of AME. Retromargin of cheliceral fang furrow with 2 large equal teeth. Ti + Mt 1 and II with 2 ventral pairs of spines except apical ones. Base of embolus in lateroapical position, TA is transverse lamella with narrow stout process by its base, directed ventrad, and with sclerotized edge situated distally (Figs 6-8). TA on its inner (dorsal) side with deep narrow transverse sinuous channel opened at the distal end of sclerotized edge. Epigyne has narrow anterior part and widened posterior (genital) parts where transverse genital part of septum is situated (Fig. 4).

FIGS 1-4. Oculicosasupermirabilis,sp. nov ,: 1, female Distribution carapace, lateral view. 2, ditto, dorsal view.. 3, ditto, South-Western Kazakhstan. frontal view. 4, epigyne. 5, position of tegular apophysis on epigyne in Atopecosa atneata (Clerek) Oculicosa supermirabiiis sp. n. during copulation. Abbreviations: ap. anterior pocket (Figs 1-4,6-8) of epigyne; ch. channel of tegular apophysis; ch. op., channel opening; eg, cpigynal groove; gen.op, genital Material Examined opening of epigyne; gps, genital part of septum; sp, TYPES. Holotype o\ South-Western septal pedicle; vs. ventral spur('hook') of tegular apophysis evident through cuticle. Kazakhstan, Mangistau Area, Yeraliev District, 37 km S of Akkuduk VilL, Karyn^haryk sands ? Description (42°38 N. 54°03*E), elav soil, S.I. Ibraev and Female. Carapace brown or light brown. A. A. Zyuzin (ZMMU), 14-15 May 1989. Lateral bands wide, continuous, forming wide Paratypes: 7 ct, 3 2, same data; 1 6 . 3 9, same data, except 15-16 May 1989 (I 9 paratype in light 'cheeks' at the sides of head. Median band ZMMU, remainder in AZ). represented by yellowish rhomboidal spot lying

Female (carapace length = 7.0mm) Male (carapace length = 7.05 mm)

Leg Femur Patella Tibia Mel. 1 :H:-'J:- T.Mai r/CL Pfniui Patella Tibia Met. 1 tiri-u:; Jmi.iI T/CL

-. S I V 2. 40 3.70 3.40 ; ho 16 50 2.36 6.20 2.70 15 5 50 2.50 22.05 3 13

2 i.ffl : IS : v; 3.35 1.80 15.65 2.24 6.10 2.60 4.80 5.60 2.60 21.70 3.08

3 4.60 2.00 3.20 4.00 1.K0 15.60 2 23 6.00 2.35 4.60 6.70 : 70 22.35 3 .7

4 5 ,;i i 2.30 4.30 5.70 2 25 20.35 2.91 7.30 2.50 5.80 8.50 3.25 :t 35 3 8S

- i !'" J so - P 2 70 1.60 2 do 1 80 12 3 20 130 L60 2.10 8.40 !

TABLE 2. Oculicosa supermirabiiis, sp. nov.; length of leg and palp segments in millimetres. T/CL. total leg length/ carapace length; met., metatarsus. A NOW LYCOS1D GENUS FROM KAZAKHSTAN AND THE TROCHOSINI ««

above median furrow (Fig 2); sometimes this spot is triangular with its base near PLE. Ocular field yellowish-brown, with dark spots around eyes. Carapace evidently descending behind PLE and is supplied with dense fur-like whitish hairs at its edge. Legs, coxae covered dorsally with dense whitish long hairs. Remaining segments uniform yellowish; their length given in Table 2,

Leg spination: femora 1 and II with 2 promesolateral spines; tibiae 1 and II with 2-2 ventral spines and 2 small apical ones, 1-1 mesolateral, no ectolateral spines; tibiae 111 and

IV with 1-1 dorsal; metatarsi I and 11 with 2-2 ventral + 4 apical, 1-1 mesolateral, no ectolateral spines; tarsi 1 IV dorsally with 2 submedian long setae. Palp: all segments uniform yellowish, but tarsus distally a little darker, tarsal claw long, slightly curved, wilh 4 very small teeth. Ab- domen: sides with dispersed dark points, dorsal pattern consists of brownish lanceolate stripe with dark margins, series of light spots lying below, and whitish spots of different size around. This pattern is covered wilh dense whitish pubescence. Ventral side above epigastrum dark, with dense dark hairs and bristles, sparse dark hairs arc also on spinnerets and at stigmal area, the remainder uniform whitish, but sometimes there arc 2 closely situated, almost parallel longitudinal dark lines. Coxae ventrally yellowish-brown, sternum yellowish with some dark hairs on sides, labium and maxillae yellow-brown. Chelicerae reddish-brown, basal segment is covered with long dense whitish-grey hairs and bristles in front. Retromargin of cheliceral fang furrow with FIGS 6-8, Oculxcasu sitpermirabMs, sp. nov.: t>, malt* palpus, ventral view. 7, ditto, lateral view. 8. male 2 teeth, promargin with 3. Body length 15.0-20.2 tegular apophysis with synembolus. apical view. Ab- mm, LC 6.5-8.0 mm. Epigyne: Fig. 4. breviations: cb. op., channel opening o\' tegular Male. Body covered with white hairs. Pattern apophysis; dep, tcgulaT depression; pi, palea; pr, stout of" both carapace and abdomen as in female, but process of tegular apophysis; scd, sclcroh/xd edge oi carapace covered with many whitish-grey tegular apophysis, sem, synembolus. adpresscd hairs. Ocular field wilh long and dense whitish hairs Carapace edged with a narrow short and adpresscd hairs, tibia and cymbium with hairs. stripe of long dense fur-like whitish (silvery) (except its distal part) long and dense Dentation of cheliceral furrow as in Perflate. Bod\ hairs directed anteriad; many short, adpresscd length 12.0-13.8 mm, 6.3-7.3 mm. Palpttfi; silvery hairs also at sides of head and form LC Figs 6, 7. 'cheeks*. Legs: colouration as in female, coxae dorsally with long and dense silvery hairs. Length of legs as in Table 2. Leg spination: femora \ and Ecology id llasin female; tibiae 1-1 V with 1-1 dorsal spines, Ail aduits were collected in clayey parts sometimes only prodorsal remains, laterally on Karyn/haryk sands each side 1-1, ventrally 2-2 -f 2 apical: metatarsi

I and li with 1-1 lateral on each side, ventrally 2-2 Discussion

+ 5 apical; tarsi I and II dorsally with 2 submedian Oculicosa suf?ermirabilis is closely related to long setae. Palp: segments yellow, cymbium dis- some species of Lyrosa s. lat., namely Lycoso t.illv brownish: all segments covered with while alticeps (Kroneberg, 1875) and L. medicu hairs: femur with sparse hairs, patella wilh many (Pocock. 1889), but differs from them by the 696 MEMOIRS OF THE QUEENSLAND MUSEUM

Lycosirwe Evippmae Pardos>iniic Wadicoranatf Vc w-Mv:' Pir.itin.ie

1 a2-a3 a2 a2 a2 u-;u ill\ ;i I a1-a2 A 2 bl; seldom M 3 hi hi hi bl-b2 bl, seldom b2

cl; J usually c 1 ; seldom c2 & cJ cJ cJ cl c! seldom c2

4 dl; seldom M dl d2 d2 a: dl-d3 e2 el el c2 el e3

6 n n n-a fl fl n 7 gi gi gi g2 gl gl

J) hi hi hi h2 h2-hJ hi ? 9 i2. seldom i3 J il i4 i5

' 11 kl;.Jc5,\lsuaJJyk2; seldom k7 k2 kl. usually k2 k3 k4 k5-fcft; seldom U

CHARACTERS 9 Male palp, bed of lip of resting embolus; il, small tegular i3„ 1 . Size: a I , small; a2, medium; a3, large-very large depression; i2, enlarged tegular depression; deep dor- i4. .Carapace, cephalic area - h 1 , high ; hi, narrow, protrud- sal channel of TA; tegular depression on upper tegtdar 2 h3 , ing basally with antenoi row of eyes process; i5, deep ascending tegular groove 3, Carapace, posterior ocuUr trapezium; cl, trapezoidal; c2, 10. Male palp, character of conductor; jl, deep dorsal wide trapezoidal; c3, +/-ijuadra(it;ulai transverse channel of TA; j2, deep dorsal longitudinal 4 Carapace, si/e of anlerior eyes: dl . variable; 62, small; channel of TA; ]3. thick well-sclerolized basal part of d3 f comparatively large palea concealed by tegulum; j4, opened (free) transverse 5. Male palp, origin of embolus: cl, mcsolaterul; cJ, sclerotized lateral process of the basal part of palea; j5, laleroapical, e3, distal (apical) opened large apical; j6, combined with embolus (single 6. Male palp, shape of embolus: fl, long curved spine; r2, complex) enlarged at tip; 0, short, combined wiih COtUtu

7. Male palp, tegulum: g 1 . no sclerotized processes; g2. with 1 1. Epigyne: kl, variable; k2, median inverted T-shaped 1-2 stout well sclerotized procc plate; k3, entire plate with 2 parallel oblong grooves

8- Male palp, tegular (median! apophysis: hi, thick, well above; fc4, simple entire plate; fc5. simple hairy plate; k6, sclerotized; h2, weakly r,cleroDZed (membranous); hJ, hairy plate with lateral scienter; k7, posteriorly protrud- absent ing hairy plate

2 TABLE 3. Comparison of lyeosid subfamilies Notes: ' Wadicosa; after Lehtincn & Hippa. 1979; Hipposa: Aulonkr, Hygrolycnsa; Xerolycosa; Evippa.

shape of epigyne and tegular apophysis, by the press against the carapace when moving in nar- profile of carapace (cf. Kroneberg, 1875, p!. IV, row burrows: the length of femur IV is slightly fig. 28; Pocock, 1889, pi. XIII, fig. I), as well as longer than the carapace slope. Long-legged by relatively larger PME and PLE, presence of males of burrowing iycosids very often have a

only 2 teeth at the retromargin of chelicerae t in low flattened carapace (as well as females of the L. aiticeps and L, medka 3), colouration of the genus Lycosa s. str., e.g. Lycosa tarantula and L Ventral side of abdomen and smaller body length- rwrdtnanni see Zyuzin, 1990): this compensates The descending carapace tn both sexes of for the lack of descent, facilitates the folding of

Oculicosa supermi nihil is, fur-like hairs at the very long femora and improves the mobility of

edges of carapace, and the long and dense whitish these spiders. I suggest that the strongly descend- hairs on the dorsal side of coxae indicate a bur- ing carapace not only in females but also in males rowing way of lite isee also Zyuzin and Zarko. (as in Oct4(icosa\ together with comparatively

1989; Zyuzin, J 990), although 1 could not find the narrow carapace, indicates the burrowing way of burrows of this species. Our investigations life from the early juvenile stages up to their

showed that carapace pubescence together with imaginal moult: therefore, the c!i striburion ofsuch dense hairs on the coxae considerably diminish Species is probably very restricted. On the con- the friction between the coxae and the edges of trary, in Lycosa nordtnanni and AitohogtKX s'tn-

i . in the much more active males these goriensis with their flattened carapace in males, features arc more pronounced and supplemented mature females seem to be more or less burrow- with many dorsal adprcsscd whitish-grey hairs on ing, while the juveniles, especially early stages, the carapace thus facilitating their movements in are active. This feature undoubtedly facilitates burrows. The role of carapace descent is as fol- aerial dispersion of juveniles: as a result, both of lows The comparatively long femora III and IV these species are widely distributed. A NEW LYCOS1D GENUS FROM KAZAKHSTAN AND THE TROCHOS1N1 697

I AL MORPHOLOGY Lycosinae is *synembolus\ as n al com-

panies the embolus and has Ihe Same direction I Further discussion concerns the terms suppose that the embolus during copulation conductor' and 'terminal apophysis', as their enters the TA channel together with the laminat interpretation by different authors is sometimes syncmbolus which directs the embolus to the rather contradictory. channel and probably locks the last as a stopper, fully excluding tin- deviation of the ewtbl Conductor Thus, the syncmbolus plays the role of auxiliary conductor, In species of the Alo put* The lip of the resting embolus in both Par verulenta group the syncmbolus is fused with the dosinae (at least in Pardosa and A> osa) base of the palea, so that only an ectal tooth and Lycosinae lies in the oblong depression of the remains at its external side: in this case all the • legnlum: in Pardosinac This depression i r.tiln i base of the palea goes to the wide antechamber* small and spoonlike (Figs 9, 12), while in before the channel. 1 \\ osinae it is enlarged, sometimes MfOfi glj ;uui

( 1 correctly regards the usually forms the tegular lobe (Fig. ID; Dood Dondale 986) subfamily 1986, figs 12, 13). In an unexpanded palp of name Hippasinae to be a junior synonym of the Pardosinae the depression of the legulum fully Lycosinae. In representatives of the genus Hip- separates the embolus from the trueTfunctionaO pasa TA also serves as the functional COttduCtOI conductor which is the transverse Well- despite ihc lack of an allium m sonic Hippazn sclerotized groove situated near the base of the species (e.g. Hippasa descrlicoia Simon, and //, terminal part (shield, palea) of the genital bulb cinerea Simon) and die agelenid habit nf the and almost fully concealed by the tegulum (Figs spiders, I regard them, as well as allied genera, to 9, 12). In members of Lycosinae, the enlarged represent the tribe Hippasmi Simon, slat nov. in depression of the tegulum is regarded as the con- the subfamily Lycosinae (see Table 3). ductor of the embolus (see Dondale and Redner, Besides the Lycosinae. the conduclOfl \i R 1979; Dondale, 1986), though this bed for the sented by the tegular apophysis in two other testing embolus does not fit to assist the exact subfamilies Evippinae (type: EvipfM Simon) and movement of the embolus tip to the female Allocosinae (type: Allocosa Banks) copulatory opening. Lehtinen and Ihppa '1979) 1. Evippinae. Members of tins subfamily have write: 'We are aware that "conductor" is not a a hooked longitudinal TA which somewhat very suitable name for the Differ pan of the resembles the transverse troehosoid 'hook' of ^sid division, because is not al- embolic it Lycosinae As in the species of Lycosinae, the functional conductor'. Our n ays a inves members of Evippinae have a distinct channel on have shown that the deep transverse channel on the inner (dorsal) side of TA and a pcdicled ihc inner (dorsal) surface of tegular apophysis septum widened posteriad (Zyuzin. 1 985, fl| opened at its narrow distal end and diagnostic for 16, 20-22), Bui. despite these similarities, the all members of the subfamily Lycosinae (Don- genera Evippo and Xeralyro.sQ have a number of dale. 1986) is intended for the embolus and un- features which allow them to be regarded as doubtedly directs its tip to the copulatory members of the separate subfamily. Thus, the opening: thus, the TA of Lycosinae Serves as the channel on the dorsal side of TA in Evippinae is functional conductor. The mechanism of opera- longitudinal (in ail Lycosinae it is transverse); the of such ;» conductor during copulation is whole embolus IS Kiiuarc.d IP a deep deprcs '. n in Alap'jcosu cuntota (Clock) (Pig, 5). and forms a very characteristic recurved flat loop; while the hooked ventral spur of the TA comes has apical position; into contact (forms a hook-up) with the anterior the base of embolus always an is strongly reduced, and pockets of the epigyne, the ventral rib of the the palea the epig 'hook' enters the longitudinal epigynal groov grooves are very shallow and lie at the level Of that the channel opening lying at the distal end of the scpiiil pedicle. X~ n Evippo spp. the TA comes into proximity with th- .•.>pn';iiory syncmbolus is transformed into a narrow functional conductor which is constantly situated in K • -(ciiing of the epigyne. channel of TA and projected beyond the TA Above the embolus and behind the TA in I ies of Lycosinae is situated a narrow, shar- limns- in this cast- TA y. an auxiliary pened laminar process (see Figs 6 X) usually ; In Xemiycosa spp the functional con- itightly grooved on its ventral side: the prop ductor is represented by TA: the embolus con- luirnc for this laminar process of palea in stantly lies in a channel (sec Zvuaui, 1985); the 698 MEMOIRS OF THE QUEENSLAND MUSEUM

FIGS9-12. Scanningelectron micrographs: 9, 10, 12, genital bulbs dissected from cymbia; 11, palea with embolus

and conductor dissected from genital bulb). 9, Pardosa sodalis Holm. 1 0. Hogna radiata (Latreille). 1 \, Pardosa chionophila L. Koch. 12, Pardosa turkestanka (Rocwcr). Abbreviations: con, conductor dcp, tegular depression; emb, embolus; pi, palea; prp, process of palea; sem, synembolus; TA, tegular apophysis; term, terminal

apophysis; ucon, upper branch of conductor. Scale bar=0. 1 mm. synembolus is strongly reduced and fused with Dondale, 1986). Besides, the basal part of the the semi-transparent extension of the embolus. pardosoid palea probably serves as an auxiliary 2. Allocosinae. In Ailocosa spp. TA is double- conductor directing the embolus into the channel branched, the channel is situated on the dorsal of TA in the expanded palpus. side of the narrow basal branch and holds the tip In the genus Pirata and allied genera (Piratula, of the resting embolus; and the atrium of the Aulonia, Hygrolycosa) the functional conductor epigyne is lost (see Dondale and Redner, 1983; is combined with a short thin embolus in a com- A NEW LYCOSID GENUS FROM KAZAKHSTAN AND THE TROCHOSINI 699

mon sickle-shaped complex resting in a deep and channel of TA (see above), the role of the true narrow ascending tegular groove. The distal posi- terminal apophysis is fulfilled by the ventral tion of the well sclerotized conductor in repre- process(es) of TA fixing the last on the female sentatives of the subfamily Venoniinae (Venonia epigyne. The designation of the ectal tooth of and allied genera: see Lehtinen and Hippa, 1 979) palea in the Alopecosa pulverulenta group as the does not allow us to include Pirata in the subterminal apophysis (see Kronestedt, 1990) is also family Venoniinae, as Dondale ( 1 986) did. Pirata incorrect: actually this tooth is the synembolus and allied genera probably deserve to be included (see above). in the separate subfamily Piratinae (type: Pirata Formerly (see Zyuzin, 1990), I restricted the

Sundevall, 1 832) (see Table 3). Lycosini to burrowing lycosids only, and erected the new tribe Trochosini for non-burrowing Terminal Apophysis genera of Lycosinae. Herein the structure of both these tribes is revised: thus, I include in Very often the palea in Pardosinae (at least in Trochosini only those genera that are charac- Pardosa and Acantholycosa) above the embolic terized by the very peculiar TA which has a division is supplied with a stout, very sclerotized transverse lamella with a ventrally directed process: many authors (e.g. Holm, 1947; Ton- trochosoid 'hook', or spur, and the epigynal sep- giorgi, 1966; Kronestedt, 1975) designate this tum with a distinct narrow pedicle, widened process as the terminal apophysis. Dondale posteriad, very often in the shape of an inverted (1986) writes '... the terminal apophysis ... is T'; epigynal grooves on either side of septal believed to assist the finding and penetration of pedicle are rather deep and distinct. Both non- the copulatory opening by the embolus tip'. It is burrowing and burrowing lycosids are included therefore obvious that, to play this very important in this very large tribe, undoubtedly having a role, the terminal apophysis must be situated im- common origin; in accordance with this view the mediately above the end of the conductor (Figs 9, tribe Trochosini is divided into two subtribes: 1 1 ; Kronestedt, 1975, fig. 3); sometimes the den- Trochosina Zyuzin, stat. nov. (including the non- tiform terminal apophysis is situated directly at burrowing genera Trochosa, Alopecosa s. str., the outer part of the conductor (see Dondale and Hogna s. str., Schizocosa), and Geolycosina, sub- Redner, 1984, figs 21, 25, 26). At the same time, trib. nov. (including the burrowing genera there are many cases when the much larger paleal Arctosa s. str., Geolycosa, Allohogna with a very process is situated far above the conductor, i.e. so characteristic carapace profiles: see Zyuzin, thai it cannot assist the exact penetration of the 1990, fig. 1). There are many species throughout embolus tip into the female copulatory opening the world, including African ones, which also (Fig. 1 2); however, such a process is also wrongly belong to the Trochosini: the generic and sub- designated as the terminal apophysis (see Lowric tribal position of these remain obscure due to the and Dondale, 1981, fig. 10; Dondale and Redner, artificial system of Roewer (1959-1960). 1984, fig. 5; Dondale, 1986, fig. 7). Tongiorgi As shown in Fig. 5, the length of the epigynal (1966, fig. 1) correctly designates the true ter- groove and septal pedicle in species of Trochosini minal apophysis and the laminar process: the is correlated with the length of the ventral spur of destination of such a process is otherwise, e.g. to TA. protect the resting embolus, or to make an engagement during copulation. As to the tribe Lycosini, I place here only the members of Lycosa s. str. with their very peculiar An incorrect designation of terminal apophysis genitalia, and some allied species referred to is also used by Buchar(1976, figs 7, 8): in his fig. 'Allocosa', 'Hog/ia' and probably Metatroch- 8 it is a mere tubercle of the palea, while in fig. 7 osina (Roewer, 1959-1960, figs 124, 126, 129, {Pardosa thaleri) this author confuses it with the 219,304-305,517). narrow laminar conductor sharpened at the tip and characteristic for the Pardosa bifasciata group. The. similar conductor shape, also desig- ACKNOWLEDGEMENTS nated as the terminal apophysis, is in the species 'Pardosa' oncka Lawrence (see Kronestedt. For the loan of material and necessary technical

1987, fig. 4C). assistance I am very grateful to the following In some works the synembolus of Lycosinae is persons: Dr Ch. Deeleman-Reinhold (Os- also called the terminal apophysis (see Dondale sendrecht, The Netherlands); Dr CD. Dondale and Redner, 1979, 1990; Dondale, 1986). But, as (Ottawa, Canada); Mr S.I. Ibraev (Alma-Ata, the synembolus only directs the embolus to the Kazakhstan); Drs D.V. Logunov (Novosibirsk), :

7IX> MEMOIRS OF THE QUEENSLAND MUSEUM

Y.M. Marusik (Magadan) and K.G. Mikhailov oncka Lawrence from Africa, with notes on its (Moscow) (all Russia). For the kindly given pos- generic position. Journal of Natural History 21: 967-976. sibility of use the scanning electron microscope 1990. Separation of" two species standing a* in my work 1 am very thankful to Dr V.K. Lebsky Ahipecosa uculeaia (Clerck) by morphological, and Mr V.M. Semyonov (St. Petersburg, Russia). behavioural and ecological characters, with I am also indebted to Mr K.N. Plakhov (Alma- remarks on related species in the puherutentu Ata), who organized the scientific trip to Man- group (Araneac. Lycosidae). Zoologica Scripia gyshlak and Ustyurt Plateaus. 19203-225. LEHTnNEN. P.T. & H1PPA, H. 1979. Spiders of the LITERATURE CITED Oriental-Australian region. I. Lycosidae: Vcnoniinae and Zoicinae. Annales Zoologici Fen- nici 16: 1-22. BUCHAR, J. 1976, Or* a - ycosiden (Araneac) aus Nepal. KJiumbu Hinial 5: 201-227. LOWRJE, DC. & DONDALE, CD. 1981. A revision CHARITONOV, D.E 1932. (Catalog der Russischen of the nigra group of the genus Pardosa in North Spinncn. Beila.sc zum Annuaire du Musce America (Araneac. Lycosidae). Bulletin of the Zoologique, Leningrad, 32. 206pp. American Museum of Natural History 170: ]25- DONDALE. CD. 1986 The subfamilies of wolf 139. spiders (Araneac: Lycosidae). Adas del X Con- MARIKOVSKIJ. P.I. 1956. |Tarantula and karakurt.

grcso Aracnoloeieo, Jaca, Espana, vol. I: 327 Morphology, biology, toxicity). (Academy of 332. Sciences of Kirghiz S.S.R.: Frunze). 281 pp. [in DONDALE. CD & REDNER, J.H. 1979. Revision of Russian) the wolf spider genus Alapea>$Q Simon in North POCOCK. R.l. 1889. Aracrmida, Chilopoda and Crus- America (Araneae: Lycosidae). Canadian En- tacea. In: Aitchison, J.E.T. The Zoology of die- tomologist 111: 1033- 1055. Afghan Delimitation Commission. Transactions 1983. The wolf spider genus AttOQQM in North and of the Linnean Society of London (2) 5, Zoologv Central America [Araneac Lycosidae). L10-12I,

Canadian Entomologist 1 15; 933-964, ROEWER. C F. 1959-196(1 Araneac Lycosacformia II 1984. Revision of the mitvma group ol the wolf (Lycosidae*. Exploration du Pare National de spider genus Pardosa (Araneac: Lycosidae). r Upemba. Mission G.F. De Witle, fascicule 55. Psyche 91: 67-1 17. lnstitut des Pares Nationaux du Congo Beige. 1990. The wolf spiders, nurseryweb spiders, and 1040 pp. (Bruxclles). lynx spiders of Canada and Alaska (Araneac: SCHMIDT. P. 1895. Beitrag vat Kenntniss der Lycosidae, Pisauridac, and Qxyopidae). The In- Laufspinncn (Araneac Ciligradae Thor.) sects and Arachnids of Canada 17. 383 pp Rnsslands. 7.oologisehcJahrbuchcr, Abteilungfur

(Agriculture Canada: Ottawa t. Svstemalik 8: 439-484. DUB[NFN,V.B. 1946. [Inhabitants of mammal holes in TONGIORGLP. 1966. Italian wolf spiders of the genus South-Kha/akhstan Area and their significance Pardosa (Araneac: Lycosidae), Bulletin of the for man}, Izveslija Akademii Nauk Kazakhskoj Museum of Comparative Zoology 134. 275-334. S.S.R., Parasitologies! Series 4: 93-102. (in Rus- ZYUZIN A.A. 1985. [Generic and suhfamiJial criteria sian! in the systematic* of the spiders family Lycosidae HOLM, A. 1947. Egenlliga Spindlar. Araneac. Fam. (Aranei), wilh the description of anew genus and 8- 10. Qxyopidae, Lycosidae och Pisauridac. two new subfamilies). Proceedings of the SvensK SpindcUauju vol; 3. 48 pp. Zoological Institute. Leningrad 139: 40-51. [in KRONEBERG, A. 1875. Araneac. In Fedtschenko. Russian) A.P. Reise in Turkestan, Zwlocischer Theil 2: 19911. Studies on burrowing spiders of the family 1-58. Lycosidae (Araneae). 1. Preliminary data on KRONESTEDT. T. 1975. Studies on species ol structural and functional features. Acta Holarctie Pardcso groups (Araneac, Lycosidae!. Zoologica Fcnmca 190: 419-422. I. Redescription ofPetftfosa tflbomacuhto Emcr- 2YU2IN, A.A. & ZARKO. M.V. 1989. The Structural ton and description of two new species from North and functional features of digging spiders of the America, with comments on some taxonomic family Lycosidae (Araneae). In: XI International characters. Zoologiea Scripta 4. 217-22*. Congress of Arachnology, Turku, Finland, 7-12 1987. On some African and Oriental wolf spiders August, 19X9. Abstracts. Reports from the Depart-

(Araneac, Lycosidae): redescription of Pardosa ment of Biology. University ofTurku 19: 1 16. . 11

CONTENTS (continued) JOCQUE, R. "We'll meet again'", an expression remarkably applicable to the historical biogeography of Australian Zodariidae (Araneae) 561 Kondo, A., Chaki, E. & Fukuda, M.

Basic architecture of the ovary in the golden silk spider, Nephito clavata , ...... 565

KOOMEN, P. & Peeters, T.MJ. New prey records for spider hunting wasps (Hymenoptera: Pompilidae) from The Netherlands 571 KoponeN, S.

Ground-living spiders (Araneae) one year after fire in three subarctic forest types, Qu6bec (Canada) . . 575

Kraus, O. & Kraus, M. Divergent transformation of chelicerae and original arrangement of eyes in spiders

(Arachnida, Araneae) ....,., • v ....».,.. , 579 Lehtinen, PX Polynesian Thomisidae - a meeting of old and new world groups 585 Locket, N.A. Scorpion distribution in a dune and swale malice environment 593 Main, B.Y.

From flood avoidance to foraging: adaptive shifts in trapdoor spider behaviour : 599

Marc, P. Imraspecific predation in Ctubiona corlicalis (Araneae: Clubionidae) 607 MtJLLER, M.C. & WESTHEIDE, W. Comparative morphology of the sexually dimorphic orb-weaving spider Argiope bruennichi

(Araneae: Araneidae) , » , 615 Platen, R. J A method to develop an 'indicator value system for spiders using Canonical Correspondence Analysis (CCA) ,....> 621 ROLLARD, C The spiders of the high-allitude meadows of Monl Nimba (West Africa): a preliminary report 629 ROVNER, J.S. Visually mediated responses in the lycosid spider Rabidosa rabida: the roles of

different pairs of eyes » 635 Sunderland, K.d. & Topping, C.J. The spatial dynamics of linyphiid spiders in winter wheat 639 Suzuki, H. & Kondo, A. Morphology of the embryos at germ disk stage in Achaearanea japonica (Theridiidae) and Neoscona nautica (Araneidae) 645 Tarabaev, C.K. An experiment on colonization of karakurt [Latrodectus tredec'tmguttatus, Black Widow spider) on island territories in Kazakhstan ..,,,.,,.... 65 Tarabaev, C.K., Zyuzin, A.A. & Fyodorov, a.A. Distribution of Latrodectus (Theridiidae), Eresus and Stegodyphus (Eresidae) in Kazakhstan

and Central Asia ... . > , , . — 653 TSURUSAKJ, N. Geographic variation of the number of B-chromosomes in Melagagrella tenuipes

(Opiliones, Phalangudae, Gagrellinae) , ,.....,.,.., ...,.,....».. * ... . 659 Uhl, G. Mating behaviour and female sperm storage in Photcus phalungioides (Fuesslin) (Araneae) ...... 667 Wishart, G.F.C. The biology of spiders and phenology of wandering males in a forest remnant (Araneae: Mygalomorphae) 675

WUNDERLLCH, J. The Macaronesian cave-dwelling spider fauna 68 YSNEL F. T Relationship between food intake and spider size in temperate zones: experimental model for an orb- weaving spider 687 Zyuzin, A.A. Studies on the wolf spiders (Araneae: Lycosidae). 1. A new genus and new species from

Kazakhstan, with comments on the Lycosinac . » . . . 693 1

CONTENTS INVITED PAPERS Selden, P.A. Fossil arachnids-recent advances and future prospects 389 Pous, G.A. Scorpions as model vehicles to advance theories of population and community ecology: the role of scorpions in desert communities 4 401 Elgar, M.A. Inter-specific associations involving spiders: kleptoparasitism, mimicry and mutualism 411 CONTRIBUTED PAPERS Adis, J. & Mahnert, V. Vertical distribution and abundance ofpseudoscorpions (Arachnida) in the soil of two different neotropical primary forests during the dry and rainy seasons 431 Austin, A.D. Nest associates of Clubiona robusta L. Koch (Araneae: Clubionidae) in Australia 441 Baert, L. & JOCQUfc, R. A tentative analysis of the spider fauna of some tropical oceanic islands 447 Benton, T.G. The reproductive ecology of Euscorpiusflavicaudis in England 455 Canard, A. & Stockman, R. Comparative postembryonic development of arachnids 461 Catley, K.M. Courtship, mating and post-oviposition behaviour of Hypochtt us pococki Platnick (Araneae, Hypochilidae) 459 Churchill, T.B. Effects of sampling method on composition of a Tasmanian coastal heathland spider assemblage 475 Davtes, V. Todd. A new spider genus (Araneae: Amaurobioidea) from rainforests of Queensland, Australia 483 Deeleman-Reinhold, C.L. An inventory of the spiders in two primary tropical forests in Sabah, North Borneo 491 Duffey, E. A review of factors influencing the distribution of spiders with special reference to Britain 497 Edmunds, J. The development of the asymmetrical web of Nephilengys cruentata (Fabricius) 503 'Edmunds, M. Does mimicry of ants reduce predation by wasps on salticid spiders? 507 Fairweather, P.G. Abundance and structure of fossorial spider populations 513 Gillespie, R.G. Biogeographic pattern of phytogeny in a clade of endemic Hawaiian spiders (Araneae, Tetragnathidae) 5 19 Gromov, A,V. A new species of Karschiidae (SoHfugae, Arachnida) from Kazakhstan 527 Hirst, D.B. A new species of Amaurobioides O.P.-Cambridge (Anyphaenidae: Araneae) from South Australia 529 Hormiga, G. Implications of the phytogeny of Pimoidae for the systematics of linyphiid spiders (Araneae, Araneoidea, Linyphiidae) , 533 Humphreys, W.F. Criteria for identifying thermal behaviour in spiders: a low technology approach 543 Hunt, G.S. & Maury, E.A. Hypertrophy of male genitalia in South American and Australian Triaenonychidae (Arachnida: Opiliones: Laniatores) 55 Jackson, R.R. & Wilcox, R.S. Predator-prey co-evolution of Portiafimbriate and Euryattus sp., jumping spiders from Queensland . . .557

(continued inside cover)