152 FUNGAL PATHOGENS OF SPIDERS He EVANS CAB International Institute of Biological Control, Silwood Park, Ascot, Berk s SL5 7PY, UK AND RA SAMSON Centraalbureau voor Schimmelcultures, P.O. Box 2 73 , 3740 AG Baarn, The Netherlands

This article deals exclusively with those spore penetration th an the remammg fung i which are proven or purported spider exoskeleton, or compared to that pathogens of spiders (Araneae, of insects. Invariabl y, th is is the part of Arachnida). They are usually included the spider which first becomes mummi- w ithin the umbrella terms of entomo- fied and which later bears most of th e genous or entomopathogenic fungi external fungal mycelium and fructifica- (Evans, 1982; Samson et 01. , 1987), since tions. It is suggested th at the majority of they share a similar morphology and fungal infections take place through the biology with some groups of insect spider abdomen, involving both pathogens. Purists may argue for more mechanical and enzymatic penetration, precise teminology, such as araneo- and that the once inside the genous or araneopathogenic fungi, to haemocoele enters a yeast phase which readily identify the pathogen-host subsequently produces a toxin lethal to associations involved. Indeed, many of the spider host. A review of fungal the mycologists mentioned in the follow- pathogenesis in insects can be found in in g pages have been guilty of confusing Samson et 01. (1987). There is no reason spiders with insects in their publications to believe that the process is significantly and all, including the present authors, different in spiders. have paid sparse attention to determin- The following is an account of the ing the spider hosts attacked by these history, ta xonomy and ecology of the fun gi. In part, this has undoubtedly been fungal pathogens of spiders: the first of due to a dearth of arachnologists whose its kind, although a few papers relating services can be called upon to give a specifically to descriptions of fungi on positive indentification. This poor inter- spiders have been published (Mains, disciplinary collaboration is evidenced 1950, 1954; Kobayasi & Shimizu, 1976). by the absence of any reference to fun gal It is not intended to be a comprehensive pathogens in published works on the review. The list of references has been biology of spiders, even when natural pruned as far as possible, without enemies are discussed (Foelix, 1982; hindering the more committed reader Nentwig, 1986). from gaining a ccess to the 60-70 There is abundant cir cumstantial specialist references con- evidence, both from field obs ervations sulted during the preparation of this and from analogy with closely-related in- article. sect fungi, to support the statement that most, if not all, of the fungi included HI STOR Y here, can be implicated directly in spider mortality. The cuticle of spiders is essen- That spiders are subject to lethal fungal tially similar in composition to that of in- infections has been known for some con- sects (Foelix, 1982). Nevertheless, the siderable time. Gray (1858) described hard exo cuticle, characteristi c of ins ect several entomopathogenic fungal species exoskeleton, is not present in the ab- and indicated that both the adult and egg dominal region (opisthosoma) of spiders. stages may be attacked . According to Physically, therefore, the opisthosoma Petch (1932), th e first illustration of a must be considerably easier for fung al fungus on a spider host was by Ditmar 153 in 1817 . This is reproduced in Gray's (Basidiomycotina) att acking or actively publication and the same species (Akan- parasitising spiders. There are records thomyces aranearum, see Fig . 7) was of such fungi, someti mes with araneo- redescribed by Cooke (1892), in his genous epithets, on spider hosts in the masterly treatise on entomogenous mycological lit erature but these appear fungi. For a more detailed insight into to be either secondary colonisers of the subject, however, we are indebted to spider cadavers or hyperparasitic on the industriousness of T Petch who nam- obligate fungal pathogens. The Ento- ed and described a large number of fungi mophthorales, for example, are com- from spider hosts. EB Mains made a monly reported on most groups of ter- shorter but no less effective contribution restrial arthropods but, seemingly, have to our knowledge of these specialised never been described from spiders. In- deed, the true pathogens are restricted to members of the Clavicipitales, and 00 : their hyphomycete anamorphs, which 00: 00001: :' ~ have de veloped mechanisms enabling 00 . ' them to invade, kill and colonise both in- dO : sect and spider hosts. Some genera and : ,- species occur only on spiders whilst others are less sel ective in their host ~a range. The following genera have been recorded on spiders and are either con- sistently or uniquely associated with , these hosts: Ascomycotina, Clavicipitales Cordyceps (Fr .) Link (Figs. 3-4) J, At least eight species have been confirm- ed on spiders (Kobayasi, 1941, 1982 ; Kobayasi & Shimizu, 1976, 1983 ; Mains, ,- 1954). Fi g. 1. Fungal genera on spiders. (AJ Hirsu- telIa;(BJHymenostilbe; (C) A kanthomyces :(DJ Torrubiella Baud. (Figs. 2D , 5) Cl athroconium; (EJ Gihellula; (F) Twenty-seven species have been describ- Granulomanus. (Bar = 10um). ed from spider hosts (Koba yasi, 1982; Kobayasi & Shimizu, 1983). fungi, whilst Y Kobayasi, has had a long association with fungal pathogens of spiders which began more than 40 years Deuteromycotina, ago with his classic monograph (Koba- yasi, 1941) and has continued virtually Akanthomyces Leb . (Figs. lC, 7) to the present day (Kobayasi & Shimizu, Three species reported so far from 1983) . Recent observations made by us spiders; both Cord ycep s and Torrubiella on material collected in tropical teleomorphs have been proposed for two ecosystems over the past 15 years, in- of these species (Samson & Evans, 1974). dicate that there are many species, and possibly genera, still to be described and Clathroconium Samson & Evans (Fig. ID) even more awaiting discovery. Monotypic genus, only on spiders; telemorph unknown (Samson & Evans, TAXONOMY 1982).

Significantly, perhaps, there are no Engyodontium de Hoog (Fig. 2C) known representatives of the lower fungi Three species on spiders; Torrubiella (Mastigomycotina) or the higher fungi teleomorph reported (Gams et al. , 1984). 154

Fig. 3. Cordyceps caloceroides on Mugale spider, Brazil. Fig. 4. C. cylindrica on trapdoor spider hidden in tunnel nest. Brazil. Fig. 5. Perithecia of Torrubie11a sp. on small salticid spider, note orange ostioles emerging from yellow mycelial weft around perithecia, Ecuador. Fig. 6. Close-up of ostiolar region with emergent capitate asci. 155

~7 .. 9

8 ...... I 0

Fig. 7. Akanthomyces aranearum, Ecuador. Fig. 8. Hymenostilbe sp. on salticid spider, Brazil. Fig. 9. Granulomanus synanamorph of Gibellula sp., Sulawesi. Fig. 10. VerticiIIium anamorph from ascospore culture of Cordyceps caloceroides, Brazil. 156 Cibellula Cav. (Figs. lE, 11-17) Verticillium Nees per Link (Fig. 10) Four species validated, genus confined Two species recorded on spider hosts to spider hosts (petch, 1932; Mains, 1950; (Cams, 1971), but several of the Samson & Evans, 1973; Kobayasi, 1982; araneogenous Cordyceps species have Kobayasi & Shimizu, 1983). Synana- been shown to produce Verticillium morph: Cranulomanus de Hoog & Sam- anamorphs in culture (pers. obs.). son (Fig. IF). It is surprising that some of the plurivorous, cosmopolitan insect Hirsutella Pat. (Fig. lA) pathogens, such as bassiana One authenticated species (Evans & Sam- (Bals.) Vuill. and Metarhizium son, 1982) but Kobayasi & Shimizu (1976) anisopliae(Metsch.) Sorok., are not more mention a Hirsutella anamorph when common on spider hosts. The complete describing a new Cordyceps taxon from absence of any spider pathogen in the spiders. Entomophthorales is intriguing. In the genus TorrubieIIa, the promi- Hymenostilbe Petch (Fig. IB) nent perithecia develop directly on the Two species reported to attack spiders; spider cadaver, typically on the abdomen teleomorph connection unproven (Sam- and buried to some extent in a dense, fre- son & Evans, 1975). quently brightly-coloured mycelium. As in most of the fungi included here, the mycelium also extends onto the substrate helping to fix the spider in position. The perithecia in the genus Cordyceps develop on aerial structures (stromata). The associated anamorph can occur directly on the mycelium or on syn- nemata of varying complexity. ECOLOGY Most of the fungal-infested spider specimens examined by the authors have been from subtropical and tropical forest ecosystems and this reflects the richness of these habitats as sources of entomo- pathogenic Clavicipitales (Evans, 1982). The diseased spider hosts are usually found on the undersides oflower storey shrubs and herbs, but occasionally on tree boles. Most are free-living adult spiders (possibly Salticidae), although Fig. 2. Fungal genera on spiders. (A) clutches of eggs and smaller spiders, en- Nomuraea; (8) PaeciJornyces; (C) Engyodon- cased in tough silken threads or bags, are tium; (D) Torrubiella (ex M E 8oudier, 1885). also subject to attack. The larger forms (Mygalidae) occur in the leaf litter or Nomuraea Maubl. (Fig. 2A) buried in the soil in their silk-lined tun- One species only but common on spiders nels. There is evidence that tropical orb- (Kobayasi, 1941; Samson, 1974); Cor- weaving spiders of the genera Argiope dyceps teleomorph now confirmed and Nephila (Araneidae) are also infected (Evans, 1982). by pathogenic fungi, especially by Nomuraea atypicola (Yasuda) Samson, Paecilomyces Bain. (Fig. 2B) dying in their webs or nearby on abnor- Two species associated with spiders mal reduced webs (W Nentwig, pers. (Samson, 1974); teleomorph probably in cornm.), The latter fungus has an in- the genus Cordyceps. teresting host range since it is also 157 prevalent on large trapdoor spiders in a centre of diversity of these specialised Japan (Kobayasi, 1941), producing spider pathogens (H C Evans, Unpubl.). elaborate synnemata on the subterra- Although it has not been possible to nean hosts. Synnematal function in N. quantify the data further, spider popula- atypicola is clear cut. In Gibelluia it may tions are at a low level in most undisturb- be two-fold: improving dispersal efficien- ed tropical forests. It is concluded, cy by lifting the conidiophores into a therefore, that fungal pathogens con- more exposed position, thereby taking stitute a significant spider mortality fac- advantage of any air movement; aug- tor in such ecosystems. This is not to say menting sporulation capacity, particular- that such fungi are rare or absent in lyon small spiders, by increasing the other, less humid, temperate regions. spore-bearing area through complex Leatherdale (1958), for example, lists 13 branching. The apparent success of species of fungal pathogens on spiders Gibellula as a spider pathogen may also in the UK. The predominantly temperate be due to its ability to diversify its Asian collections of Kobayasi (1982) and resources. The Granulomanus Kobayasi & Shimizu (1983), furthermore synanamorph could be directed at poten- testify to their ubiquitousness. tial spider hosts coming into physical It is now becoming increasingly ap- contact with the infected cadaver, rather parent that spiders are an important than for aerial dispersal. However, component in the biological control of movement of inoculum within the pest insects. Are fungal pathogens of ecosystem is a speculative subject since spiders, therefore, a negative component we have no information on how and in terms of pest control? Several when the various types of spiders arguments for and against can be put for- become infected. ward. Spiders do prey on pests and, con- Evans (1974) regularly recorded en- sequently, any mortality factor of tomopathogenic fungi on spiders in arachnids must be looked upon as forest habitats in Ghana. Six collecting detrimental to the crop and hence to sites (20m') were sampled monthly over man. Nevertheless, spiders are also a one-year period and, after ants and coc- natural enemies of beneficial insects and cids, spiders were found to be the group can destroy large numbers of insect of arthropods most affected by fungal predators. Moreover, they also compete pathogens. Over 150 specimens, repre- with the latter for prey. Whichever argu- senting 15 species in 10 genera, were col- ment is valid, and it could be either given lected. Samson & Evans (1973) also the complexity of tropical crop recorded high levels of spider mortality ecosystems, the status of fungal in Ghanaian cocoa farms, collecting 82 pathogens should be carefully evaluated specimens of Gibelluia pulchra(Sacc.) in any modelling of spider population Cav. from 10 sample trees. Petch (1932) dynamics. reported the same fungus as common at There seems to be no better way of en- lower (tropical) altitudes in Sri Lanka. ding this article than by drawing upon Another species of Gibellula, as yet the still relevant remarks from one of the undescribed, was collected in epizootic earliest reviews on en to mapathogenic proportions over several years in fungi: "It is hoped, in conclusion, that Ecuadorian cocoa farms. A more recent these pages may awaken some curiosi- survey was carried out in a monsoon ty, if not interest, in favour of these forest habitat in the Far East (Sulawesi, singular productions, and thus lead to Indonesia) and, out of a total of 690 further inquiry as to the species of in- fungal-infected arthropods, gathered sects that unfortunately become the during a three-week visit, 137 were bases of the fungoid parasites" (Gray, spider hosts, or approximately 20% of 1858). the collection. Members of the genus We would like to thank Dr C Prior for Gibellula were especially well permission to use Fig. 11 and Ms G represented and this island seems to be Godwin for photographic assistance. 158

Figs. 11-17. Variations on a theme - the genus Gibellula. Fig. 11. Gibellula leiopus, sugarcane spider, Trinidad. Figs. 12-17. Undescribed Gibellula spp. on small, free-living salticid spiders from tropical forests. 159 REFERENCES

COOKE, M C (1892). Vegetable Wasps and of Vegetable Wasps and Plant Worms. Osaka, Plant Worms. London: Society for Pro- Japan: Hoikusha Publ. Co., 280 pp. moting Christian Knowledge, 364 pp. LEATHERDALE, D (1958). A host catalogue of EVANS, H C (1974). Natural control of arthro- British entomogenous fungi. The Entomo- pods, with special reference to ants [Fermi- logist's Monthly Magazine 64, 108-110. cidae], by fungi in the tropical high forest MAINS, E B (1950). The genus GibeIIula on of Ghana. Journal of Applied Ecology 11, spiders in North America. Mycologia 42, 37-49. 306-321. EVANS, H C (1982). Entomogenous fungi in MAINS, E B (1954). Species of Cordyceps on tropical forest ecosystems: an appraisal. spiders. BuIIetin of the Torrey Botanical Club Ecological Entomology 7, 47-60. 81, 492-500. EVANS. HC & SAMSON, R A (1982). Entorno- NENTWIG, W (Ed.) (1986). Ecophysiology of genous fungi from the Galapagos Islands. Spiders. Berlin: Springer-Verlag, 430 pp. Canadian Journal of Botany 60, 2325-2333. PETCH, T (1932). Gibellula. Annales FOELlx, R F (1982). Biology of Spiders. Cam- Mycologici 30, 386-393. bridge, Massachusetts: Harvard University SAMSON, R A (1974). PaeciIomyces and some Press, 306 pp. allied Hyphomycetes. Studies in Mycology GAMS. W (1971). Cephalosporium-urtige, 6,119 pp. SchimmelpiIze (Hyphomycetes). Stuttgart: G SAMSON R A & EVANS, H C (1973). Notes on Fischer, 262 pp. entomogenous fungi from Ghana. I. The GAMS, W, HOOG, C S DE, SAMSON, R A & genera GibeIIula and PseudogibeIIula. Acta EVANS, H C (1984). The hyphomycete genus Botanica Neerlandica 22, 522-528. Engyodontium a link between VerticiIlium SAMSON, R A & EVANS, H C (1974). Notes on and Aphanocladium. Persoonia 12, 135-147. entomogenous fungi from Ghana. II. The GRAY, R C (1858). Notices of Insects that are genus Akanthomyces. Acta Botanica Known to Form the Bases of Fungoid Neerlandica 23, 28-35. Parasites. London: Privately Printed, 22 pp. SAMSON, R A & EVANS, H C (1975). Notes on KOBAYASI, Y (1941). The genus Cordyceps and entomogenous fungi from Ghana. III. The its allies. Science Reports of the Tokyo genus HymenostiIbe. Proceedings of the Bunrika Daiguku 5, 53-260. KoninkIijke Nederlandse Akademie van KOBAYASI, Y (1982). Keys to the taxa of the Wetenschappen, Amsterdam 78, 73-80. genera Cordyceps and TorrubieIIa. Transac- SAMSON, R A & EVANS, HC (1982). tions of the Mycological Society ofJapan 23, Clathroconium, a new helicosporous 329-364. hyphomycete genus from spiders. Canadian KOBAYASI, Y & SHIMIZU, D (1976). Some Journal of Botany 60, 1577-1580. species of Cordyceps and its allies on SAMSON, R A, EVANS, HC & LATGE, J P (1987). spiders. Kew BuIIetin 31, 557-566. An Atlas of Entomopathogenic Fungi. Berlin: KOBAYASI, Y & SHIMIZU, D (1983). Iconography Springer-Verlag (In Press).

BMS GREETINGS CARDS AND WALLCHARTS Greetings Cards: Four, full colour cards are available, depicting Agaricus macrosporus, Flammulina velutipes, Lactarius torminosus, and Tylopilus felleus. Each is a reproduction of an original painting by Ray Cowell. Price: 25p per card (inel envelope), or £10.00 for forty-eight, plus postage. Wallcharts: Three posters, Edible Fungi, Poisonous Fungi, and Lichens &-Air Pollu- tion, are available. The first two were designed by Ray Cowell, and produced by the British Museum (Natural History) in collaboration with British Petroleum. The Lichen poster was designed by Claire Dalby. SPECIAL PRICE for Members and Associates: £1.00 plus postage (please note that wallcharts sent by post will be folded). Both Greetings Cards and Wallcharts are available from the Foray Secretary (Dr D W Minter), and will also be available at foray meetings. Please print your name and address with orders.