1992. The Journal of Arachnology 20 :18–2 4

PHRYNIDAE () FROM ANDROS ISLAND , BAHAMAS, WITH NOTES ON DISTRIBUTION PATTERNS, RECENT ORIGIN AND ALLOMETRY

D. Jonathan Browne: Dept. of Entomology, University of Pretoria, Pretoria 0002 , Republic of South Africa

ABSTRACT : Fieldwork on Andros Island produced two of phrynid amblypygi, margine- maculatus and viridiceps . New localities and biological data are presented . The species were found to be completely sympatric in two of the three localities where they were collected. The presence of amblypygi in the Bahamas is attributed to Recent dispersal from Cuba via Paleoprovidence, a land mass which emerged during the lower sea-levels accompanying Pleistocene glacials . Dispersal from Florida and Hispaniola is rejected . Significant isometric or allometric relationships between median prosomal length and pedipalp tibia length were not detected . The work generally points to the lack of much basic information on amblypygid distribution an d bionomics .

Two species of phrynid amblypygi are known variety of collecting techniques were employed . from the Bahamas. Collecting efforts from May through August 198 7 C. L. Koch is widespread in the northern Baham a (110 days) used flight intercept traps, malaise Islands (Quintero 1981) and Paraphrynus viri- traps, baited pitfall traps, shrub and tree beating , diceps (Pocock) is known from southern Andro s grass sweeping, blacklighting and hand collectin g and New Providence Islands (Banks 1906 ; Mul- [the most successful in terms of obtaining am- linex 1975). From published records, both ap- blypygi specimens] . The author was assisted b y pear to be uncommon to rare in the Bahamas. a technician throughout the collecting period . In- Furthermore, these species have never been re- tensive hand collecting included investigating all corded as being completely sympatric in the same caves and "banana holes" found, which often habitat (see Quintero 1983). required the use of rappelling equipment to gain The general lack of collecting of all s entrance. In addition stones, rocks, leaf litter and in the Bahama Islands, and especially Andro s logs were turned and the spaces beneath the bar k Island, was noted by the author several years ag o of dead trees examined to a height of 2 m. while undertaking literature searches . This wa s A large number of localities were sample d surprising because Andros Island is the largest (>100) which included all the recognised vege- island in the Bahamas (Fig. 1), and has the great- tation zones many times over. Most recent plant est diversity of vegetation types which would ecology workers have characterized ten distinc t presumably harbor the greatest number of ar- Bahamian vegetation types ; these are beach/ thropod species in the Bahamas . As a result of a strand, coastal rock, coastal coppice, interior familiarity based on many trips to Andros Island coppice, pineland, savanna, scrub, freshwater the author decided to undertake an extensive sur- marsh, saltwater marsh, and mangrove (Nick - vey of this island, employing modern mass col- rent et al. 1987). lecting techniques not used by earlier workers. All specimens recorded here were collected an d Recent collecting has revealed new localities and identified by the author and vouchers were de - biological data for two Bahamian amblypygi spe - posited by Dr. S. B. Peck (Department of Biol- cies on Andros Island. This report is offered as ogy, Carleton University, Ottawa) in the arach- a contribution to knowledge of the nology collection of the American Museum of of the West Indies . Natural History . The significance of allometric changes in am- METHODS blypygi has been previously studied and detaile d The project was intended as a general survey by Quintero (1983) for seven Cuban species. The of the arthropods of Andros Island, thus a wide significance of allometric changes in Bahamia n 18

BROWNE— FROM ANDROS ISLAND, BAHAMAS 19

Figure 1 .—Map showing major islands of the Bahamas in relation to the Antilles and Florida. Line around islands represents probable extent of land area during Pleistocene glacial low sea levels. Paleoprovidence is the name proposed for the largest and most central of these Pleistocene glacial islands . Water barriers to dispersal were less significant during these times. Species may have been lost as sea levels rose, and the fauna was force d into today's more restricted island areas.

species has not been previously examined . Due small (each roughly 0 .5 km2) and isolated local- to the small sample size, immatures, females an d ities of high interior coppice . No specimens of males were combined into a single data set (Tabl e either species were found in any of the other nin e 1) as did Quintero (1983) . The relationship be- recognized vegetation zones . Only a few speci- tween median prosomal length (x) and pedipal p mens of each species were collected so as not to tibia length (y) was calculated using the Stat- disrupt what appeared to be uncommon and iso- graphics simple regression package (Version 2 .6). lated populations . However, they were found to Both linear and multiplicative models were fit t o be very abundant throughout the period fro m the data. The best-fit model was chosen base d May to August 1987 . The two species bear the on R-squared values with the final relationshi p following information. between x and y expressed as a power curve [y = a x"]. Comparison of actual and predicte d Phrynus marginemaculatus slopes (Futuyma 1986), using the Student's t test, Distribution.—This species has the most up- followed techniques suggested by Sokal & Rohl f permost latitudinal distributional range of any (1981) and Hoel (1984) . Interpretation of results amblypygid species in the eastern part of Nort h adhered to methods proposed by Futuyma (1986 ) America. It has been previously recorded fro m and Packard & Boardman (1987). Statistical data Bermuda, southern Florida, Cuba, Haiti, Do- are presented in Tables 2 and 3 . minican Republic, Puerto Rico, Jamaica, and RESULTS AND DISCUSSION Antigua (Quintero 1983) and is widespread i n the northern Bahama Islands (Quintero 1981) . Both Phrynus marginemaculatus and Para- phrynus viridiceps were found to be resident o n New Records.—BAHAMAS: Andros Island (random Andros Island . Despite intensive and prolonged search of high interior coppices) : CDC Farm, Cricket searching, populations were located in only three Coppice, 19 July 1987, 1 female ; Shot-gun Coppice,

20 THE JOURNAL OF ARACHNOLOG Y

Table 1. —Measurements of Median Prosomal Length They are protected from annual burning of th e (x) and Pedipalp Tibia Length (y), in mm, from spec - surrounding pine forests and savanna (the tw o imens collected from Andros Island during July 1987. most common vegetation zones) . The canopy of A = Phrynus marginemaculatus, B = Paraphrynus vir- a coppice is dense with a cool and wet, but sparse , idiceps. understory layer . The vegetation grows on hon- B ey-combed limestone which affords the ambly- Specimen A pygids many small holes in which to retreat . Am- number x y x y blypygi appear to favor cool, wet habitats 1 3 .984 3 .992 5 .140 4.99 8 (Quintero 1983) which explains their preferenc e 2 5.280 5.450 7.085 8.100 for the high interior coppice . The coppices ar e 3 4 .290 4 .423 6.140 6.140 separated by wide stretches of arid savanna an d 4 5 .040 4 .783 4.235 4.000 pine forest. Two of the three coppices in whic h 5 2 .950 2 .790 3.589 3.400 these species were found (London Ridge and Shot - 6 4 .558 4 .940 3.790 3.280 gun Coppice) lie approximately 60 km from eac h 7 3.060 2.890 other. The third coppice (Cricket Coppice) i s 8 5 .685 5 .465 within four km of Shot-gun Coppice. Pocock (1893) recorded Paraphrynus viridiceps from 19 July 1987, 2 females (1 with 12 eggs), 4 males; southern Andros Island, which is over 100 k m London Ridge, 24 July 1987, 1 male. from the populations recorded here, and per- Variation.—Body length 6.32—13 .98 mm; median manently separated by a wide salt-water gap fro m prosomal length 2 .95—5 .28 mm; left pedipalp tibia the central and northern sections of the island . length 2.79—5 .45 mm. Color varies from wheat yellow Although amblypygids are known to run rapidly, to black in both males and females. their dispersal capabilities are unknown. No specimens were found in the dry savanna or pine Paraphrynus viridiceps forest. If non-habitable areas lie between cop- Distribution .—This species is limited to the pices, then four km may be as much a barrier t o Bahamas and Cuba (Quintero 1983). Paraphry- gene flow as 60 km . At present it is not known nus viridiceps has been previously recorded fro m if amblypygids disperse between coppices in the Cuba (Quintero 1983). The holotype male was Bahamas. However, widespread habitat destruc- collected in New Providence and described by tion has occurred several times on Andros Island Pocock (1893). Additional Bahamas records in- as a result of logging ; therefore, the present ap- clude Andros Island (South Bight) and Ne w parent isolation of these populations may be a Providence (Mullinex 1975) . recent condition . Sympatry: Partial sympatry of species range New Records.—BAHAMAS: Andros Island (ran- s dom search of high interior coppices) : CDC Farm, between Phrynus and Paraphrynus has been re- Cricket Coppice, 19 July 1987, 1 male; London Ridge, ported previously (Quintero 1983). However, 24 July 1987, 4 females, 2 males, 1 immature. Quintero (1983) "doubts" that "species range s Variation.—Body length 6.29—15 .04 mm; median will overlap to a major extent" due to "compet- prosomal length 3 .06—7 .08 mm; left pedipalp tibia itive exclusion" . On Andros Island Paraphrynus length 2.89—8 .10 mm. Color varies from wheat yellow viridiceps and Phrynus marginemaculatus were to black in both males and females. found to be completely sympatric in the sam e habitat. They were also observed to intermingle Local Distribution and Habitat Preferences : freely. However, these observations do not dis- High interior coppices occur on elevated part s prove the occurrence of competitive exclusion. of Andros Island . This community is the most The absence of specific information relating t o diverse of the vegetation zones on the island . specific niche requirements for either species Dominant woody plant species in this vegetation makes it difficult to support or refute competitiv e zone include Bursera simaruba, Metopium tox- exclusion. iferum, Ficus aurea, Exothea paniculata, Calyp- Recent origin of Bahamian amblypy- tranthes pallens, Drypetes diversifolia, Clusea ro- gids: Three possible sources of Bahamian am- sea, Psychotria angustifolia and Nectandra blypygi must be considered. These are Florida , coriacea (Nickrent et al. 1987). The surface of Cuba and Hispaniola. Dispersal of flora and fau- the high coppice is very much eroded and slightly na from Florida into the Caribbean is considere d depressed which tends to accumulate moisture . to be a very rare event . Many insect groups (Eck-

BROWNE—PHRYNIDAE FROM ANDROS ISLAND, BAHAMAS 2 1

Table 2. —Statistics for regressions of Median Prosomal Length (x) versus Pedipalp Tibia Length (y), comparin g linear (L) and multiplicative (M) models. Data from specimens collected from Andros Island during July 1987 . A = Phrynus marginemaculatus, B = Paraphrynus viridiceps, a = intercept, b = slope.

Sp. Model a ± SE b ± SE R2 SE estimate A L -0.295 ± 0.59 1.08 ± 0 .13 94 .18 0.250 A M -0.153 ± 0.16 1.11 ± 0 .11 95 .92 0.05 3 B L -1 .208 ± 0.45 1.23 ± 0 .01 96.89 0.335 B M -0.338 ± 0.11 1.20 ± 0.07 97 .81 0.057

wort 1988—Halictidae ; Liebherr 1988 — Platyn- of these species throughout Paleoprovidence. us; Nichols 1988—Scaratinae; Ramos 1988 — Rising sea-levels and inundation of most of Pa- Homoptera ; Slater 1988—Lygaeidae ; Wilson leoprovidence have formed the present Bahama 1988—Formicidae; Peck 1989—south Florida Islands in the last few thousand years . This re- insects) and trees (Tomlinson 1980) are believe d duction in area would also have reduced specie s to have dispersed from the Caribbean north, via numbers to present levels and restricted thei r Cuba or the Bahamas, to south Florida (Patter- movement between the newly isolated islands. son & Stevenson 1977) . The strong northern Allometry: Both the linear and multiplicative movement, from the southern Caribbean, o f models had significant fits to the data for bot h storms, prevailing winds and the Gulf Stream species (Table 2) . However, the R-squared values supports this argument . The author agrees with were higher for the multiplicative model, indi- this assessment and with the implication that cating that a power curve is the best-fit line . southern movement of flora and fauna fro m Allometries between closely related species ar e Florida to the Bahamas can be considered a very not congruent; but significantly different inter - rare event. specific differences between wild and laboratory Immigration and certain residency could hav e reared specimens have been reported, althoug h only been possible since the re-emergence of th e Quintero (1983) ascribes these differences to Bahamas after Pliocene flooding. The northern methodology. Quintero (1983) reported isomet- Bahamas (including Andros Island) are part of ric growth for seven Cuban species of amblypygi . the Bahama Bank, considered to be exposed con- It is more probable that he proved allometric tinental shelf (Lee 1951 ; Burke et al . 1984; Don- growth for reasons which are detailed below . nelly 1988). During the lower (some 100 m) sea- Demonstration of an allometric or an isomet- levels accompanying Pleistocene glacials and as ric relationship depends on the value of both th e recently as 18,000 years BP, the northern Ba- intercept and slope of the best-fit line . Packard hamas were broadly and continuously adjacen t & Boardman (1987), in their comprehensive re- to Cuba via a land mass known as Paleoprov- view of allometric analysis, state : "When a plot idence, with the latter separated from the former of some variable of interest yields a straight lin e by only a few kilometres of open water (Fig . 1). passing through the origin of a graph with linea r Immigration from Cuba during this time is th e co-ordinates, the variable varies isometrically most likely route, rather than the "stepping- with body size . When the line is curvilinear or stone" route from Hispaniola via the southern when it does not pass through the origin, how - Bahamas. This hypothesis is reflected in the cur- ever, the variable varies allometricallywith body rent distribution of Bahamian trees. Sixty-three size". Since neither line passes through the ori- species are common to the Bahamas and Cuba, gin, the Bahamian amblypygi species in this study while only twenty-nine are common to the Ba- do not exhibit isometric growth (Table 2; Fig. 2). hamas, Cuba, Hispaniola and the Lesser Antilles It is also evident that Quintero (1983) demon- (Patterson & Stevenson 1977) . Therefore th e strated allometric growth, rather than isometric widespread presence of Phrynus marginemacu- growth, as none of the best-fit lines which he latus throughout the northern Bahama Island s presented passed through the origin. This is not and the more restricted distribution of Para- an unexpected finding since size-related varia- phrynus viridiceps to only a few of these island s tion in most physiological variables is allometric is a reflection of a once widespread distribution (Packard & Boardman 1987) .

22 THE JOURNAL OF ARACHNOLOGY

P E D 6 marginemaculatu s P 5 A - y=0.86x1 ." L P 4

T I 3 B

A 2

L 1 E A G o T 2 .950 3.984 4.290 4 .558 5 .040 5 .28 0 H MEDIAN PROSOMAL LENGT H

P E D 1 0 viridiceps P A 8 y = 0.71 x 1 .2 ° L P

T I B 4 A

L 2 _ E B N G o T 3 .060 3.589 3.790 4.235 5 .140 5 .685 6 .140 7 .08 5 H MEDIAN PROSOMAL LENGT H

Figure 2.- Allometric growth curves with linear co-ordinates : Median Prosomal Length (abscissa) versu s Pedipaip Tibia Length (ordinate), measurements in mm . Slopes of both graphs are not significantly different from one and therefore an allometric relationship is not demonstrated . A = Phrynus marginemaculatus, B = Paraphrynus viridiceps .

BROWNE—PHRYNIDAE FROM ANDROS ISLAND, BAHAMAS 23

Table 3.-Statistics for Student's t test to determine whether the slope of the multiplicative regression differ s significantly from one for each of the Andros Island amblypygi species . If b = 1, then an allometric relationshi p is not demonstrated (Futuyma 1986). This is the case for both species.

Species t N — 2 P Phrynus marginemaculatus 0.246 4 0.3 > P > 0 .4 Paraphrynus viridiceps 0.529 6 0.3 > P > 0 .4

While the data in this study did not exhibit an abaeoid wings. My tolerant parents, Ron and isometric relationship, it cannot be concluded Mary Browne, have kindly given me much en- that an allometric one is demonstrated by de- couragement and financial support while com- fault. An allometric relationship also depends on plOing this project and over many years. Finally the value of the slope (b) (Futuyma 1986). Al- I Would like to thank Prebendary Philip Hus- lometry is demonstrated only if 1 1 . If b = baInds and Albert Ernest Browne who provide d 1, then y is a constant proportion of x and al- most of the funding for the field work . lometry is not demonstrated . This is the case with both species of Bahamian amblypygi fro m Andros Island (Table 3); therefore, allometry i s LITERATURE CITED not demonstrated. This is in contradiction to Banks, N. 1906. Arachnida from the Bahamas. Bull. Quintero (1983) who reported an allometric re- American Mus. Nat. Hist., 22:185-189 . s Burke, K., C. Cooper, J . F. Dewey, P. Mann, & J. L. lationship for the same variables in seven specie e of Cuban amblypygi. Quintero (1983) did no t Pindell. 1984. Caribbean tectonic and relative plat motions. Pp. 31-63. In The Caribbean—South detail his methodology so no reason for this dis- American Plate Boundary and Regional Tectonic s crepancy can be determined at this time, save (W.E. Bonini, R. B. Hargraves, & R . Shagam, eds .). for disproportionate sample sizes. Geol. Soc. America Mem ., 162. DOnnelly, T. W. 1988. Geological constraints on Ca - ACKNOWLEDGMENTS ribbean biogeography . Pp. 15-37. In Zoogeography of Caribbean Insects . (J. K. Liebheer, ed.). Cornell The author wishes to thank the people of An- University Press, Ithaca, New York. dros Island, without whose co-operation this Eiekwort, G . C. 1988. Distribution patterns and bi - project would never have been realised. Thanks ology of West Indian sweat bees (Hymenoptera: are especially due to Mr. D. Myles who ably Halictidae). Pp. 231-254. In Zoogeography of Ca- assisted under difficult field conditions. Dr. G. ribbean Insects. (J. K. Liebheer, ed.). Cornell Uni- Dennill (Dept. of Entomology, University of Pre- versity Press, Ithaca, New York . toria) and Dr . S. B. Peck (Dept . of Biology, Carle- Fiituyma, D . J. 1986. Evolutionary Biology, 2nd ed . ton University) provided invaluable editorial as- Sinauer Associates, Inc ., Sunderland, Massachu - setts. 600 pp. sistance. Dr. S. Chown and Mr. M. Vogt (Dept. - y Hoel, P. G. 1984. Introduction to Mathematical Sta of Entomology, University of Pretoria) kindl tistics. 5th ed. John Wiley and Sons, Toronto. 43 5 assisted with the statistical analysis . To Mr. J. PP. C. Cokendolpher, Dr. G. L. Miller, Dr . D. Quin- Lee, C. S. 1951. Geophysical surveys on the Bahama tero and an anonymous reviewer I wish to exten d Banks. J. Inst. Petrol., 37:633-657 . my sincere thanks for suggesting many improve- Liebherr, J. K. 1988. Biogeographic patterns of West ments to this manuscript. I also wish to thank Indian Platynus carabid beetles (Coleoptera) . Pp. Panagiotis Karandis, the mad Greek, for stim- 121-153. In Zoogeography of Caribbean Insects, (J . ulating conversation and procuring a seemingl y K. Liebherr, ed.). Cornell University Press, Ithaca , endless supply of mental lubricants . To my Zhitt New York. NIullinex, C . L. 1975. Revision of Paraphrynus Mo- & Jive II companions, Lou "Captain Ahab" D e a Vogt and Jon "Hard- reno (Amblypygida: Phrynidae) for North Americ Vries, Mark "Robo-scrop" and the Antilles. Occ. Papers Calififornia Acad. Sci. , copy" Pio, my sincere thanks for their useful , 116:1-80. albeit few, twisted and rather bizarre, ideas . Prof. Nichols, S. W. 1988. Kaleidoscopic biogeography of C. H. Scholtz (Dept. of Entomology, University West Indian Scaratinae (Coleoptera : Carabidae). Pp. of Pretoria), my PhD supervisor, kindly allowed 71-121 . In Zoogeography of Caribbean Insects. (J. me to delve into when he probabl y K. Liebherr, ed.). Cornell University Press, Ithaca , would have preferred me to keep looking at scar- New York.

24 THE JOURNAL OF ARACHNOLOGY

Nickrent, D. L., W. H. Eshbaugh & T . K. Wilson . ribbean and continental American amblypygid fau- 1987. The vascular flora of Andros Island. Uni- na. Studies on the Fauna of Curacao and other Ca- versity of Illinois, Urbana, Illinois . 182 pp. ribbean Islands, 196 :1-54. Packard, G. C. & T. J. Boardman. 1987. The misuse Ramos, J . A. 1988. Zoogeography of the Auchen- of ratios to scale physiological data allometricall y orrhynchous Homoptera of the Greater Antilles with body size . Pp. 216-240. In New Directions in (Hemiptera). Pp. 61-71 . In Zoogeography of Carib - Ecological Physiology . (M. E. Feder, A . F. Bennett, bean Insects. (J. K. Liebherr, ed .). Cornell Univer- W. W. Burggren & R . B. Huey, eds .). Cambridge sity Press, Ithaca, New York. University Press, Cambridge. Slater, J. A. 1988. Zoogeography of West Indian Ly- Patterson, J. & G. Stevenson. 1977. Native trees o f gaeidae (Hemiptera) . Pp. 38-61 . In Zoogeography the Bahamas . Bahamas National Trust, Nassau. 128 of Caribbean Insects . (J . K. Liebherr, ed.). Cornell PP. University Press, Ithaca, New York. Peck, S . B. 1989. A survey of insects in the Florida Sokal, R. R. & F. J. Rohlf 1981 . Biometry : The Keys: Post-Pleistocene land-bridge islands : Intro- Principles and Practice of Statistics in Biological duction . Florida Ent ., 72:603-612. Research. 2nd ed. W. H. Freeman and Co ., San Pocock, R . I. 1893. Contributions to our knowledg e Francisco . 859 pp. of the arthropod fauna of the West Indies . III. Di- Tomlinson, P. B. 1980. The Biology of Trees Native plopoda and Malacopoda, with a supplement on th e to Tropical Florida . Harvard University Printing Arachnida of the class Pedipalpi . J. Linn . Soc. Zool., Office, Allston, Massachusetts . 480 pp. 24:473-544 . Wilson, E. O. 1988. The biogeography of the West Quintero, D ., Jr. 1981 . The amblypygid genus Phry- Indian ants (Hymenoptera : Formicidae) . Pp. 214- nus in the Americas (Amblypygi, Phrynidac) . J. Ar- 231 . In Zoogeography of Caribbean Insects . Cornel l achnol., 9:117-166. University Press, Ithaca, New York. Quintero, D ., Jr. 1983 . Revision of the amblypygid spiders of Cuba and their relationships with the Ca- Manuscript received April 1991, revised August 1991 .