CISX NH OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES No. 151 August 13, 2001 MADAGASCAR Contributions to the Study of the Biodiversity of Madagascar Number 1 A MONOGRAPH OF THE MIGID TRAP DOOR SPIDERS OF MADAGASCAR AND REVIEW OF THE WORLD GENERA (ARANEAE, MYGALOMORPHAE, MIGIDAE) By Charles E. Griswold and Joel Ledford LIBRARIES California Academy of Sciences San Francisco, California =e MADAGASCAR Contributions to the Study of the Biodiversity of Madagascar Number 1 A MONOGRAPH OF THE MIGID TRAP DOOR SPIDERS OF MADAGASCAR AND REVIEW OF THE WORLD GENERA (ARANEAE, MYGALOMORPHAE, MIGIDAE) (Above) Female of Poecilomigas abrahami (O.P. Cambridge) from Dlinza Forest, Zululand, South Africa (Below) Nest of Moggridgea teresae Griswold on fallen log at Kirstenbosch, near Cape Town, South Africa Photographs by Teresa Meikle OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES No. 151 August 13, 2001 MADAGASCAR Contributions to the Study of the Biodiversity of Madagascar Number 1 A MONOGRAPH OF THE MIGID TRAP DOOR SPIDERS OF MADAGASCAR AND REVIEW OF THE WORLD GENERA (ARANEAE, MYGALOMORPHAE, MIGIDAE) By Charles E. Griswold and Joel Ledford California Academy of Sciences San Francisco, California SCIENTIFIC PUBLICATIONS Alan E. Leviton, Editor Katie Martin, Managing Editor © 2001 California Academy of Sciences, Golden Gate Park, San Francisco, California 94118 All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage or retrieval system, without permission in writing from the publisher. ISSN 0068-5461 Printed in the United States of America Norcal Printing, Inc., San Francisco, California Table of Contents Page PANDSUT ACE Weer tens theta hae teen tetas rch lar N caste cape satus up Se leeauigech esas oes geen null aides Reneyanl at saint teapee Go arcs 7 IMO GUCHIO MUR eprint yey tees ede one nee ceeu rake aihite ie ic csie its sbau eee emee rere Nereaeee re Scie A NOEs if MatenalspandiNlethod sie amnesia cite leer nes et custenric cup prereset tesla car necpausm suis ane butiaereat ace Stterrei ara nae ETE 8 INCKMOWIEd STIEMES) ey sra cents eycicls © aeeseme ae cee eecus Gey nero eaeie es lye ens eCPM Eel el cae tons) Sastayeetene Says ahSeuel srawevene 9 Materal¥exammined tier e weustar gest nari tcacuencoy ence sae icy eh urs ee ken cera Retna ature atgtars top en eit cepen ergy eae utes eue fede 11 BPX OM OMY fs yc tce istics sss sy soe teens hey Sse eo gomsceaarsb stiob salheks psiciessme tiie (eevee SHSM ani caireuctbeeniieel aimed esa vcs wiPel( ce elystlchtonenemayeaya owaNe 12 INVA HTS FeV ae eens cee cece ae UE a Po ee PPE reer PCE ete ein seamen PERN CoP ee ene 12 KeystoworldiGeneravand Malagasy,;species of Misidaceri ries vate ele escola eeha leiee re = ieee 12 IMbieaal exe OYE IBRO MOVIE ACEI oo fo cu dnd on Gor ooe HE dweheeouHadnmoopbodouorgugascae x 14 CGIOthGIGYSUSES IMO ete eee eT ote es Ser eer eee Con Ee ON me ayaa 14 ETELEKOMMISASMO S Cire mentee ere ete ay Smee RNY oy sl een aati ete e aA Meme een ecto crt leat eee custome 15 MallecomiganGolobottasablatnic kr ser acct ote eee ine a es ees een ne ee ee 15 IME SOS MUIR OC Ise ecye ayers a ete ra aeype easy a eeg 2h oer og ae ee cua CREPE Wea eee Ura sehen eu RDI CA 15 Moseriasca, OukaCambrid Ser eet ii tact cr pt hee tcc arcs ore kro 16 IPAM Loy AhegeKrs NWN) We eeetan he airore/e bas Cale eos MeO erro OTR Oo OER en ne oh to hamid one Ode 17, GolobojjiaiGuswoldiSqWedtordsmewacenus) mers sai eteeiy a eee ee aie ele eer weber: 17 MircrdaciotéMadasascantrsestie is» cos cranes ves eisusirecerauee papel cxtey airing sce seec oy es actesiPe pa) esa agetedes Cap aparece prog te 18 Raramiginaetketrumkevatc hen sy retire seer eis ets thence eae re te an cee oa e soar Rare ed eare rio REO ae 18 Geis MAO PAOM MVOC. ho curekedobodongosnomeerundoswodeoeEoconosnacadao4s 6 18 IMiGKOMESOMMATICOWANTLOCOCK@ ane eater o eee Cae 19 GenussParamigasiROcock tare srcavemicnsye nrececiey acme ute tert oi asec tens ee ees nie aey asec eres een ee O iRaramicas.alluavail (SIMON) ea pena eee ae ee ean Ce aay eee OTe eee) RaramigaNandasibe Raven.Me Was PeCleSrrqa seat haar. ene ae el oe ey ete ee Paramigas goodmani Griswold & Ledford, new species ............0 0000 c eee eee eee 22 Paramigas;macrops Griswold & Wedtord) new Speciesy\-n.-452 a9 aoe eee ee eae eee Paramigas manakambus Griswold & Ledford, new species ...........-.00000000 0-02 ee s 24 Paramigas milloti Griswold & Ledford, new species ............--00 eee e eee ee eee eee 25 Raramigasroracie Griswold é&edtordmewispecleSmiy se spas eee Lae ee iRaramigasipaulianin(IDTESCOrsc Canard) ews ie eee eee ee eee ee Paramigas pectinatus Griswold & Ledford, new species .......... 0.0 ee eee eee eee es 28 JACEE XAT (SWIMM) 34 wkh ao edekeeseaasedcowmonnadatoespaudunoadeagmen son oy) Paramicasiothorum Guswold\éWedtords new: SpecCleSs:2 2 se -wahey as ane ae eee 30 GenuseihpropOeussPOCOCKs marae er ary sears y eae ea toe naire Penner patie ieee ao ai a EE 3] MMO ROAS Tielageniyes (Gucite))\anacod@aaleducunoyocandsoos ve ateudnod@oddgeseouabaes 31 hy ro DOCUSIMIFANAUSEROCOCKa earn hata OC ee 33 BH VO PEE LICS tre erp een pera rete ht eee ge Cesena pe ne Skee URC oT RRS yO ARTE ceed ley ky Pahl Seca ee peo et aN a 34 DE AEEY cost k cicern te bees pri er estas sari a irks RG LOUaT erie) Perec OM io ml ees Pek sree MUS ee Ate, cee m sate ONCE iin & 34 @UbeTOUPS Hr He stent rr cee RS eS ain, Reni Ooo edt homies Uae Ry wii Sepmregan el ree aad en arses 35 ENTITIES azo ccted ats tees abe Shsten et nearer bbc eRe esc rel Mee Mec eeser RE ERP am steam es a te 35 reLemmedetneel amma aey se Cucenpe etme cet eterna Res ee ctu tes fo8 sco ath punta AIRE Uy ce shicseyeb nt cauenew ss eS Characterioptimnzationrandibranchisupporth eee cscs coy secs cee soles) cicero een 36 Characterndescrptionsrandhnterpretationsimem me oo tons oi iene ece oe erie ieee es O Chemin Moning Waleal in GhininNGiin: honéacen ces cooounmouno bana pe oo Moun oon gob. doodle 43 le ahilkoyezerayi Ope IM bieaTGRIS” Ws he Gio quater fore oa a) diols Gia a oroemtaeeo Clore rELE eit a ehetel one ora Ub iaiG la Wiad Celt ote oetoe elo 43 (CHERSIITCE LETS cise ores orton ienia Compre chcrcs area Raney cA ace me ecnere ees ee TREO IHR ee ia atariacatihiet ita Weeecaeietin 44 BIO SECO STAM Varo nner atone tae Centers ecm a int Noni eunaet ce) aren, Te py Rea Sonn ati ellcarertagtalegrs MeN ra UP abem cess) scape 44 GONCIMSTONS ar pee.feac yee uiew acca Morea eee RSM ete cease cai creme ne) caeh ean ied ealic ea ewteni am ung aR ere ea 47 Jbiteratuineycited Myre kwacetns ccan RRSP dee teterk on SOR Mey ae cence ceeeinis eaena arenas abet encima arta teehetan eed 47 Ei SULeS WIE O Steere tery cert en Crete arene aes Ne neta ear raerns eee dae ker ocean) heme ee Ol SPD EIN IXtaes tere easter e ete ae eyecare nee Cea Serie ere Me aay a NE ee aoe ne ee eta w ee ten eg 119 _ ' oaeens ote Sh i a ee. a ies ; a “snele i A MONOGRAPH OF THE MIGID TRAP DOOR SPIDERS OF MADAGASCAR AND REVIEW OF THE WORLD GENERA (ARANEAE, MYGALOMORPHAE, MIGIDAE) Charles E. Griswold (1)* and Joel Ledford (2) The Malagasy fauna of the spider family Migidae is monographed and a phylogeny for Malagasy species and exemplars of the world genera is proposed. This phylogenetic analysis corroborates the monophyly of Migoidea (Actinopodidae plus Migidae) and Migidae, and suggests that the Mala- gasy migids form a monophyletic group (Paramiginae, sensu Petrunkevitch 1939, but not Raven 1985), that Migas, Moggridgea, and Poecilomigas form a monophyletic group (Miginae Simon 1892), that the Calathotarsinae Simon 1903 is paraphyletic with respect to the Miginae, and that the South American migids form a monophyletic group. A key to the world genera and 14 species of Malagasy Migidae is presented. The following new species are described: Paramigas goodmani, P. macrops, P. manakambus, P. milloti, P. oracle, P. pectinatus, and P. rothorum, all authored by Gris- wold and Ledford, and Paramigas andasibe Raven. Paramigas subrufus Pocock is a new junior syn- onym of P. perroti (Simon). The male of Thyropoeus malagasus, the first male known for the genus, is described. Goloboffia new genus is proposed for Migas vellardi Zapfe 1961, giving Goloboffia vel- lardi new combination. The trap door spider family Migidae has long at- tracted attention because of its strikingly disjunct southern continent distribution. Migids are known from Australia, Africa, Madagascar, New Zealand, New Caledonia and the southern cone of South Amer- ica: virtually all parts of the former supercontinent Gondwanaland except the Indian subcontinent and Antarctica. Migids have figured prominently in discus- sions of spider biogeography and of southern disjunc- tions (Pocock 1903; Legendre 1979; Platnick 1981; Nelson & Platnick 1984; Griswold 1991a). Reasonably complete modern treatments exist for the African (Griswold 1987a, 1987b), Australian (Raven 1984; Main 1991), South American (Schiapelli & Gerschman 1975; Goloboff & Platnick 1987; Goloboff 1991), and New Zealand faunas (Wilton 1968). The faunas of Madagascar and New Caledonia have received less at- tention. Dresco and Canard (1975) briefly reviewed the Malagasy genera. Nothing has been published on the migids of New Caledonia since Berland (1924). With *To whom correspondence should be addressed: (1) Schlinger Cura- tor of Arachnida, Department of Entomology, California Academy of Sciences, Golden Gate Park, San Francisco California 94118 USA and Research Professor of Biology, San Francisco State Uni- versity; Fax: (415) 750-7228; Internet: cgriswold@calacademy.org; (2) California Academy of Sciences Summer Systematics Institute, Golden Gate Park, San Francisco California 94118 USA and De- partment of Biology, San Francisco State University; Fax: (415) 750-7228; Internet: jledford@calacademy.org. this paper, the 15th in a series treating the afromontane spider fauna (Griswold 1985, 1987a, 1987b, 1987c, 1990, 1991b, 1994, 1997a, 1997b, 1998a, 1998b, 2000, 2001; Griswold and Platnick 1987), we monograph the Malagasy migids utilizing all available material. Madagascar is widely recognized as being of great conservation importance (National Research Council 1980; Rasoanaivo 1990; Myers et al. 2000) because the island is known for high rates of endemism and unique occurrence of primitive members of otherwise wide- spread taxa (Myers 1988). Ongoing rapid habitat de- struction, particularly of forests, makes the collection, description, and study of the evolutionary and biogeo- graphic significance of the Malagasy biota particularly urgent. The spider fauna of Madagascar remains poorly known. Since the first Malagasy spider (Augusta glyphica Guérin 1839) was described more than 150 years ago, nearly 150 papers have been published on Malagasy spiders. The number of spiders recorded from the whole island only slightly exceeds 450 de- scribed species (Roth 1992a; Griswold, in press), sig- nificantly less than the 626 species recorded from the British Isles (Merrett et al. 1985, Merrett & Millidge 1992). Yet, nearly 400 species (new and described to- gether) have been collected from a single site (Parc Na- tional Ranomafana) in the southern part of the island (Roth 1992b), suggesting a rich fauna yet to be de- 8 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 scribed. The important afromontane spider families Cyatholipidae (Griswold 1997a) and Phyxelididae (Griswold 1990) have only recently been recorded from the island. Remarkably, a new family (Halidae) was recently described from Madagascar from hitherto undescribed species (Jocqué 1994). Alderweireldt & Jocqué (1994) suggest that the known component of the Malagasy spider fauna is around 10%. Our knowl- edge of the evolutionary relationships of Malagasy spi- ders is poorer still: of the more than 120 papers that deal explicitly with the taxonomy of Malagasy spiders, fewer than 10 present cladograms or other explicit phy- logenetic hypotheses. Clearly, much remains to be learned about this rich fauna. Our studies of the Malagasy migids benefit from two recent developments: (1) a well supported cladis- tic hypothesis of migid affinities and (2) new migid material collected in the course of general surveys of Madagascar biodiversity. Raven (1985) and Goloboff (1993a) proposed mygalomorph cladograms that con- firmed a Migoidea (sensu Platnick & Shadab 1976) comprising Migidae and Actinopodidae and suggested Ctenizidae and Idiopidae as close relatives. Our phylo- genetic analysis builds on this hypothesis. After initial collections of Malagasy migids in the late 19th century few of these spiders were collected for the next 100 years. Renewed interest in the Malagasy biota and con- cern for impending extinctions led to new efforts to survey the island. In particular, the efforts of Steve Goodman and Brian Fisher in the 1990s to collect ter- restrial arthropods and small vertebrates produced as many migid specimens as collected in the previous 100 years. Nevertheless, our attempts to monograph the Malagasy migids and understand their phylogeny are hindered by absent and ambiguous data. Of the ten world genera two (Mallecomigas and Micromesomma) are still known only from females. Malagasy migids re- main very rare in collections: we have examined only 113 specimens, adults and juveniles alike. Many species are known from single specimens, and collect- ing data may be absent or ambiguous. Of the 14 Mala- gasy species that we recognize, only one (7hyropoeus malagasus) is known with certainty from both sexes. Indeed, only once have a male and female of a Mala- gasy migid species been collected in association (in 1996 by Goodman in a pitfall trap at Vohimena)! Recognizing these shortcomings in the data, we at- tempt to understand the phylogenetic position of Mala- gasy migids, review the biology of the family, describe the Malagasy fauna, and provide a key to the world genera and Malagasy species. We are confident that continued exploration of the arthropod faunas of the southern continents will complete the picture of the Migidae and provide tests of the hypotheses presented herein. MATERIAL AND METHODS CONVENTIONS—Throughout the text, figures cited from previous papers are listed as “fig.” and those ap- pearing in this paper as “Fig.” All illustrations are by Jenny Speckels (JS), Joel Ledford (JL), Virginia Kirsch (VK) or Charles Griswold (CG) and are attributed in the figure captions. When referring to size, ‘small’ = <10mm, ‘medium’= 10-20 mm, and ‘large’ = >20 mm. In the discussion of biogeographic history “ma” refers to millions of years before the present. Records of dis- tribution in Madagascar presented in Figure 68 include localities for the juveniles listed in Table 1. DESCRIPTIONS—Each genus Is treated, a brief tax- onomic history is provided, major diagnostic features are listed, and published reports on the biology are summarized. For Malagasy and Afrotropical genera the description summarizes all included species: for other genera the description reflects the exemplars (see Table 2) supplemented by the literature. Species de- scriptions follow the format in Griswold (1987a). Each description is of a single individual of each sex, which is identified either as a type or by the locality at which it was collected. Illustrations cited in the description may not refer to the specimen described but serve to describe the features reported. For each sex there is also a section reporting the variation in the most con- spicuous and variable features. For variation in quanti- tative features three to five adult individuals represent- ing the full range in overall size were chosen to sample the species: the number (N) is reported at the beginning of each section. All measurements are in millimeters. Abbreviations used in the text and figures are listed in Table 3. Eyes are measured from above. The sternal sigilla are the concave regions near the sternal margin as viewed in oblique lighting, not the discolored area as- sociated with these structures. Spines are recorded as arising from the dorsal (d), prolateral (p), ventral (v) or retrolateral (r) surface of the segment and are recorded from proximal to distal. For example, tibia p 0-0-1, d 1-0-0, r 0-1-1 signifies that the tibia has a proximal spine on the dorsum, a distal spine on the prolateral side, and a median and distal spine on the retrolateral side. The convention r 0-0-la signifies that the retro- lateral spine is at the apex of the segment (a = apical). Especially on the lateral surfaces of the tibiae-tarsi of the female pedipalpus and legs I and II spines may be in diagonal series oriented in a proximo-dorsal to disto- GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 9 Table 1: Juvenile Paramigas from Madagascar (all specimens are in CASC) Antsiranana Province: Reserve Speciale Manongarivo, 10.8 km SW Antanambao (13°57.7'S, 48°26'E) elev. 400 m, rainforest, sifted litter, 8 No- vember 1998, B.L. Fisher Reserve Speciale d’ Anjanaharibe-Sud, 12.2 km WSW Befingotra (14°45'S, 49°26'E) elev. 1985m, montane rainforest, sift- ed litter, 25 November 1994, B.L. Fisher Toamasina Province: Presquil de Masoala, 5.3 km SSE Ambanizana, Andranobe (15°40'S, 49°58'E) elev. 425 m, rainforest, sifted litter, 21 No- vember 1993, B.L. Fisher SF Tampolo, 10 km NNE Fenoarivo Atn. (17°17'S, 49°26'E) elev. 10 m, littoral rainforest, sifting litter, 4 April 1997, B.L. Fisher Fianarantsoa Province: Reserve Speciale Ivohibe, 8 km E Ivohibe (22°29'S, 46°58.1'E) elev. 1200 m, forest, sifting leaf litter, 15-21 October 1997, B.L. Fisher Reserve Andringitra, 38 km S Ambalavao (22°12'S, 46°58'E) elev. 1680 m, montane rainforest, sifted litter, 23 October 1993, B. L. Fisher Reserve Andringitra, 38 km S Ambalavao (22°12'S, 46°58'E) elev. 1875 m, montane rainforest, sifted litter, 29 October 1993, B. L. Fisher Reserve Andringitra, 8.5 km SE Antanitotsy (22°10'S, 46°58'E) elev. 1990 m, rainforest, sifting litter, 6 March 1997, B. L. Fisher Toliara Province: Mahafaly near Eloetse, by Lac Tsimanampetsoa (24°10'S, 43°45'E) 15-16 September 1992, V. & B. Roth Southern Isoky-Vohimena Forest (22°41'S, 44°50'E) elev. 730 m, tropical dry forest, sifted litter, 21 January 1996, B.L. Fish- er Foret de Petriky, 12.5 km W Tolagnaro (25°03.73'S, 46°52.16'E) elev. 10 m, littoral rainforest, sifting litter, 22 November 1998, B.L. Fisher ventral direction, with the smallest spines most dorsal and the largest most ventral (Figs. 1, 24A): in this case the convention P 1-1-2-4-3 signifies that on the prolat- eral surface there are two proximal spines in a row fol- lowed by diagonal series of 2, 4 and 3 spines. We de- fine “thorns” as socketed setae that taper to the apex and that are less than 5x as long as the diameter at the base (Fig. 25D; Griswold 1987a, figs. 23-26). Thorns have a nearly smooth surface. Cuspules are socketed setae that are narrow at the base, widest medially, and blunt to pointed at the apex, and deeply grooved (Gris- wold 1987a, figs. 358, 359) or smooth (Figs. 43C—D). Localities in Madagascar may be hard to trace due to many identical or similar place names, abbreviation of place names on labels, illegible labels, and name changes over time. We have tried our best to decipher difficult labels. The recent compilation of field stations where insects have been collected in Madagascar (Vi- ette 1991) has been very helpful. SPECIMEN PREPARATION—Vulvae were examined in lactic acid. In some cases they were cleaned by ex- posure to trypsin or cleared with Clorox® bleach. Ex- amination was via Wild MSApo, Olympus SZH10, Leica MZApo and Leitz Ortholux II microscopes. Prior to examination with a Hitachi S-520, JEOL T- 200 or Cambridge Stereoscan Scanning Electron Mi- croscope (SEM), large structures were air-dried and delicate structures (spinnerets, whole mounts of juve- niles) were critical-point-dried. PHYLOGENETICS—On the preferred cladogram (Fig. 65) all nodes were assigned a letter (e.g., A, B). Throughout the discussions of the phylogenetic analy- sis and relationships among taxa these letters are used to refer both to nodes on the cladogram (e.g., node N) and clades distad to that node (e.g., clade N, compris- ing Migas, Moggridgea and Poecilomigas). For further discussion see ‘phylogenetic methodology’ under ‘Phylogenetics’ below. ACKNOWLEDGMENTS Support for this project came from National Sci- ence Foundation grants BSR-9020439 and DEB- 9020439 to Griswold and DEB-0072713 to Brian Fish- er and Griswold. Initial participation by Ledford was made possible by the California Academy of Sciences (CASC) Summer Systematics Institute, supported by NSF grant BIR-9531307. Additional space and equip- ment support was provided to Ledford by the Ento- 10 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 Table 2: Exemplars and other taxa examined for migid phylogenetic analysis (note: M=male; F=female). Outgroups: Actinopus rufipes (Lucas, 1834): EF from Brazil, Barra de Pirai, Rio de Janeiro, det. Bucherl (AMNH) Actinopus sp.: E from Tobago, Grafton Estate (AMNH) Actinopus sp.: FF from Colombia, Meta, Puerto Lleras, Lomalinda, elev. 300m., 73°22'W, 3°18'N, 15 April 1986, B. T. Car- roll (CASC), MM, same locality, March 1987 (CASC) Actinopus sp.: M from Argentina, Corrientes, 29 April 1964, E. S. Ross (CASC) Bothriocyrtum californicum (O.P. Cambridge, 1874): M from USA, California, Los Angeles Co., campus of University of California at Los Angeles, November 1955, M. Maher (CASC); F from Los Angeles Co., Eagle Rock, July 1991, C.E. Griswold (CASC) Idiops sp.: MF from South Africa, Natal, Pietermaritzburg, 1-15 October 1985, C. & T. Griswold (CASC) Non-Malagasy Migidae: Calathotarsus coronatus Simon, 1903: specimens from Chile, Valparaiso, Parque Nacional La Campana, Cerro La Campana, M from elev. 980m., 11 November 1986, F from 1080 m., 26 August 1981, R. Caderon (AMNH); also M from Santia- go Prov., Quilicura, September 1979, Pena (AMNH) and F from El Canelo, elev. 950 m., 8 September 1966, E. Schlinger & M. Irwin (CASC) Calathotarsus simoni Schiapelli & Gerschman, 1975: F from Argentina, Buenos Aires, Cerro Negro, Sierra de la Ventana, April 1974, Césari (AMNH) Heteromigas dovei Hogg, 1902: MF from Australia, Tasmania, Patersonia, 29 January 1926, V. V. Hickman (QMB) Migas vellardi Zapfe, 1961: holotype M from La Herradura, Coquimbo Province, Chile (UCS); F from Chile, Coquimbo Re- gion, Guanaqueros, 10 January 1984, E. Maury & P. Goloboff (MACN 8338) Mallecomigas schlingeri Goloboff & Platnick, 1987: holotype F from Chile, Araucania Region, elev. 610m in Cordillera de Nahuelbuta, 18 km W Angol, 10 February 1967, E.I. Schlinger (UCS at CASC). Note: This specimen has been trans- ferred to UCS. Migas exemplars: Migas gatenbyi Wilton, 1968: M (fragmentary) from New Zealand, Wellington, Town Belt, Oriental Bay, 41°17'S, 174°46'E, 11 January 1995, L.J. Boutin (CASC), F from same locality, 17 April 1995, L.J. Boutin (CASC) Migas taierii Todd, 1945: M from New Zealand, Patearoa (AMNH) Other Migas: Migas affinis Berland, 1924: holotype M from New Caledonia, Forét du Mt. Panié, 27 May 1911, Roux & Sarrasin (MNHN AR 4132) Migas distinctus O.P. Cambridge, 1879: EF from New Zealand, Portobello (AMNH) Migas giveni Wilton, 1968: F from New Zealand, N. Island, Waipoua Forest, Yakas Track, 2 February 1994, E.I. Schlinger (CASC) Migas nitens Hickman, 1927: F from Cornelian Bay, Tasmania, 28 August 1930, V. V. Hickman (MNHN AR4131) Migas paradoxus L. Koch, 1873: E from New Zealand, C. L. Wilton (AMNH) Migas variapalpus Raven, 1984: M from Australia, Queensland, O’Reilly’s, Lamington National Park, elev. 935m, 28°13'S, 153°08'E, 12 July 1992, C.E. Griswold (CASC) Migas sp.: M from Tauranga, New Zealand (OMD) Moggridgea crudeni Hewitt, 1913: F from Alicedale, South Africa, F. Cruden (AMSA) Moggridgea dyeri O. P. Cambridge, 1875: M from Uitenhage, South Africa, F. Cruden (AMSA) Moggridgea intermedia Hewitt, 1913: MEF from South Africa, Western Cape, Diepwalle Forest Station, 21 km. N. Knysna, 33°57'S, 23°09'E, elev. 540 m, 12 December 1996, C.E. Griswold (CASC) Moggridgea pseudocrudeni Hewitt, 1919: M from Alicedale, South Africa (AMSA) Moggridgea peringueyi Simon, 1903: M from South Africa, Cape, Oudtshoorn, 29 October 1949, B. Malkin (CASC); F (holotype) from South Africa, Cape, Matjiesfontein (MNHN 19274) Moggridgea rupicola Hewitt, 1913: MMF from South Affrica, Alicedale (AMSA) Moggridgea tingle Main, 1991: paratypes, from Walpole National Park in Western Australia, M from Big Tingle Tree, 16 July 1908 (WAM 89/330), FE from Valley of the Giants, 14 January 1990 (WAM 90/1112) and Deep River crossing at Manjinup, 14 January 1990, B.Y. Main (WAM 90/1113) Poecilomigas abrahami (O.P. Cambridge, 1889): F from South Africa, Natal, Sordwana Bay, elev. 0 m, 50 mi. E of Ubom- bo, 5 April 1958, E. S. Ross (CASC); F from South Africa, Uitenhage (AMSA); MM from South Africa, Natal, Pieter- maritzburg, Town Bush, elev. 3000 ft., 29°33°S, 30°21’E, 8 January 1984, T. Meikle (CASC), same data (NMSA). Poecilomigas basilleupi Benoit, 1962: MF from Tanzania, Tanga Region, West Usambara Mts., Mazumbai, 4°49’S, 38°30’E, elev. ca. 1400 m, 10-20 November 1995, C. Griswold, D. Ubick, & N. Scharff, (CASC) GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 11 Table 3: List of Anatomical and Other Abbreviations Used in the Text and Figures. a apical AER anterior eye row ALE anterior lateral eyes AME anterior median eyes d dorsal HS spermathecal head ITC inferior tarsal claws OAL ocular area length OAW ocular area width p prolateral PER posterior eye row PEE posterior lateral eyes PME posterior median eyes T retrolateral SS spermathecal stalk STC superior tarsal claws Vv ventral mology Department, University of California, Davis. Additional support came from the Exline-Frizzell and In-house Research Funds of the CASC and the Schlinger Foundation. The 1998 CASC expeditions to Ranomafana, Madagascar were supported by grants from the Oracle and McBean Foundations. Robert Raven provided the specimens and de- scriptive notes on Paramigas andasibe, which new species is attributed to his authorship. Specimens and observations essential to the completion of this study were obtained through fieldwork in Australia, Mada- gascar, and New Zealand. This fieldwork was made possible through permissions granted by a variety of government bodies. Permits to do research in Australia were granted by the Queensland National Parks and Wildlife Service, facilitated by Dr. R. Raven of the Queensland Museum. Permits to do research in and ex- port specimens from Madagascar were granted by the Association Nationale pour le Gestion des Aires Pro- tegées (ANGAP) and Direction des Eaux et Foréts of the Ministre d’Etat a L’ Agriculture et au Development Rural, under Accordes de Collaboration of the Xerces Society (facilitated by Dr. C. Kremen, Mr. C. Ramili- son, and Ms. B. Davies of that organization) and Insti- tute for Conservation of Tropical Environments (ICTE) (facilitated by Drs. R. van Berkum and Ben- jamin Andriamihaja). A permit to collect specimens in New Zealand protected areas was granted by Te Papa Atawhai (Department of Conservation). Research in Tanzania was made possible through a Research Per- mit from the Tanzania Commission for Science and Technology (COSTECH) and Residence Permit Class C from the Tanzanian Department of Immigration, and export of specimens made possible by a CITES Ex- emption Certificate from the Wildlife Division of the United Republic of Tanzania, facilitated by Professor Kim M. Howell of the University of Dar-es-Salaam. Research at Mazumbai was made possible by Dr. S. A. O. Chamshama, Dean of Forestry, Sokoine University, Morogoro, and Mr. Modest S. Mrecha, Officer in Charge, Mazumbai Forest Reserve. Research in South Africa was made possible through a permit from the Department van Waterwese en Bosbou. The following are thanked for their hospitality: (Australia) Drs. Robert Raven and Val Davies; (Bel- gium) Dr. Rudy Jocque; (France) Drs. Christine Rol- lard and the late Jacqueline Heurtault; (New Zealand) Drs. Lyn and the late Ray Forster; (UK) Mr. John and the late Mrs. Frances Murphy, Mr. Paul Hillyard; (USA) Drs. Norman Platnick and Petra Sierwald. For generous hospitality and able assistance in Madagas- car, especially in fieldwork, I thank the following: Mr. Riya Andriamasamanana, Ms. Marie Jeanne Raheri- lalao, Mr. Samuelson Randrianarisoa, Mr. Christophe Roland and Ranomafana Park Ranger Mr. Emile Raje- riarison. Dr. David Kavanaugh deserves special thanks for help in the field as do the members of the California Academy of Sciences/Parc Botanique et Zoologique Tzimbazaza (CAS/PBZT) 2000 Ant-Spider course: Daniella Andriamalala, Balsama Rajemison, Jean- Jacques Rokotoarison, Jean Claude Rakotonirina and Helian Ratsirarson. Assistance with Scanning Electron Microscope examination was provided by Mr. Darrell Ubick (CASC), and Mr. Vijay Bandu, Mrs. Connie Bowland, and Mr. Antony Bruton (University of Natal). Darrell Ubick and Mr. Keith Dabney assisted in ways too numerous to mention. A draft of the manu- script was critically read by Fred Coyle, Pablo Goloboff, Norm Platnick, Robert Raven and David Ka- vanaugh. Agreement with all of our conclusions is not implied This is contribution #14 from the CASC Center for Biodiversity Research and Information (CBRI). MATERIAL EXAMINED The following institutions and individuals lent or otherwise provided specimens: Ms. Lisa Joy Boutin, Tasmania, Australia; Dr. Ariel Camousseight, Museo Nacional de Historia Natural, Universidad de Chile, Santiago (USC); Mr. Peter Croeser, Natal Museum, Pietermaritzburg, South Africa (NMSA); The late Dr. Ray Forster, Otago Museum, Dunedin, New Zealand (OMD); The late Dr. Maria Elena Galiano, Museo de Ciencias Naturales “Bernardino Rivadivia”, Buenos Aires, Argentina. (MACN); Dr. Fred Gess, Albany 12 Museum, Grahamstown, South Africa (AMSA); Dr. Mark Harvey, Western Australian Museum, Perth, Australia (WAM); Mr. Paul Hillyard and Mrs. Janet Beccaloni, The Natural History Museum, London, UK (BMNH); Dr. Rudy Jocqué, Musée Royal de L’A frique Centrale, Tervuren, Belgium (MRAC); Dr. Torbjérm Kronestedt, Naturhistorika Riks-museet, Stockholm, Sweden (NRS); Dr. Norman I. Platnick, American Mu- seum of Natural History, New York, USA (AMNH); Dr. Robert J. Raven, Queensland Museum, South Bris- bane, Australia (QMB); Dr. Christine Rollard and the late Dr. Jacqueline Heurtault, Muséum National d’His- toire Naturelle, Paris, France (MNHN); Mrs. Barbara and the late Mr. Vincent D. Roth, Portal, Arizona; Dr. Petra Sierwald and Mr. Phil Parillo, Field Museum of Natural History, Chicago, USA (FMNH); The other specimen depository listed is ZFMK (Staatliches Mu- seum fiir Naturkunde, Karlsruhe, Germany). TAXONOMY Migidae Diagnosis —Distinguished from all other mygalo- morphs by having the combination of characters rastel- lum absent (Figs. 2B, 5B, 9B), fangs quadrate and keeled (Figs. 2C, 43A), thoracic fovea straight to re- curved (Figs. 1, 2A), and PLS with spigots on median and distal articles only (Figs. 3A—B). Description——Small to large sized mygalo- morphs, 5-45 mm in length. Sparsely setose. Carapace smooth in female (Fig. 1), rugose in males (Figs. 6, 52), length 1—1.25 x width; caput low to arched. Ocu- lar area 0.40—0.65 x width carapace; AER straight to slightly recurved (Figs. 2A, D). Thoracic fovea re- curved, in some specimens straight (Figs. 29, 52), 0.12—0.40 x width carapace. Chelicerae lacking rastel- lum, fangs quadrate, with keels (Fig. 43A), in most species obliquely oriented (Fig. 2C), pro and retromar- gins of fang furrow with teeth (Figs. 24E, 43B). Ster- num length 1.0—1.50 x width, with | pair sigilla oppo- site coxae II-III (Fig. 5B). Cuspules (socketed, blunt macrosetae) present on labium and pedipalpal coxae at least of females (Figs. 2C, 43C, D); serrula absent (Fig. 43C). Leg formula 4123 or 1423; legs I and II of fe- males with strong series of spines (Figs. 1, 57) on pro- and retrolateral margins of tibiae, metatarsi, and, in some species, tarsi, these spines in most species form- ing diagonal rows; legs III and IV with dense patterns of spinules and stout setae, especially dorsally and pro- laterally on patellae and on metatarsus III (Figs. 25A—C, 44D); scopulae absent from females, present beneath at least some tarsi in males (Figs. 30B, 31F, OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 60A). STC with tooth or single row of teeth, ITC with or without teeth. Trichobothrial pattern: pedipalpal tar- sus with a dorsal row of 4-6, row may divide proxi- mally, tibiae with dorsobasal converging rows of 3-6, metatarsi with group of several near dorsoapex, poste- riad of this point trichobothria form a single dorsal row that may extend to metatarsal base, tarsi with numerous trichobothria (10-25) on their dorsal surface (Fig. 27D); trichobothrial base distally embedded, with proximal hood, smooth or weakly corrugiform (Figs. 24C, 42D, 55C; Griswold 1987a, figs. 41-42). Tarsal organ in apical half of segment (Fig. 27D), oval to rouna, weakly domed, with concentric ridges (Figs. 24D, 42D, 55D; Griswold 1987a, figs. 43-44). Posteri- or lateral spinnerets 3— segmented, basal segment longest, median and distal approximately equal in length, distal segment domed, length less than 2/3 di- ameter, median segment with 12-40 spigots, apical with 30—SO; anterior lateral spinnerets single segment- ed, separated at base by diameter, 0.5 x length PLS, with ventroapical group of 4-20 spigots (Figs. 3A—B); male with rectangular field of epiandrous spigots ante- riad of epigastric furrow, spigots evenly spaced, sepa- rated by approximately % their length. Spermathecae (Figs. 17A—C, 35A-E) paired, simple, unbranched, straight to distally curved, with pores throughout, un- sclerotized or sclerotized. Male with or without clasp- ing processes on legs, with (Fig. 30B) or without (Fig. 11D) megaspine at retroapex of metatarsus I, pedipal- pus with bulb simple, piriform (Figs. 7A-C, 11A-C, 16A, 20A, B, 32A—D, 54A-C). KEY TO WORLD GENERA AND MALAGASY SPECIES OF MIGIDAE 3(2) Patellae I, II, and IV with ventral setae normal, ta- pering from base to apex (Fig. 13C); fang with basal tooth (Figs59D) ioe. aera oe eee 4 —Patellae I, II, and IV with ventral patches of lamellate setae (Figs. 15C, 16B, 61D); fang without basal tooth bands ja atiov exrapSoniieoy cist oe ER aE RECS Moggridgea 4(3) Dark dorsal and lateral maculations forming trans- verse bands or annuli on tibiae and metatarsi (Fig. 18); metatarsus IV preening comb absent. . . Poecilomigas — Tibiae and metatarsi lacking such dorsal and lateral GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS maculations (Fig. 13C); metatarsus IV preening comb present in some species (Fig. 59C) Migas 5(2) Tibia III cylindrical, convex dorsally at base (Figs. 4, 8, 14D, 59A); caput setose, with rows or at least a few setae posteriad of the ocular area (Figs. 8, 12A) . — Tibia III concave dorsally at base (Figs. 1, 25A—C); caput lacking setae except some species with pair of prefoveal setae (Figs. 2A, 23A, 50) 6(5) Patellae lacking ventral lamellate setae (Figs. 12C, D); fang unarmed (Figs. 5B, 12B) — Patellae with ventral lamellate setae (Fig. 14C); fang with basal tooth (Fig. 14B) Goloboffia vellardi (Zapfe) 7(6) Pedipalpal tibia cylindrical, tapering to apex, un- modified, without cluster of spines (Fig. 8); OAW less than 0.5 x width of caput (Figs. 8, 12A) — Pedipalpal tibia swollen ventrally, with dorsoapical cluster of spines (Figs. 4, 5C); OAW greater than 0.5 x width of caput (except C. pihuychen) (Fig. 4) Calathotarsus 8(7) Cuspules on pedipalpal coxae restricted to base near labium (Fig. 12B) Brey shots Mallecomigas schlingeri Goloboff & Platnick —Cuspules on pedipalpal coxae extending for length of Coxae (hig% 9B) ere eateries Genie Heteromigas 9(5) Sternal sigilla deeply excised, lunate (Figs. 51B, 58B) 10 10(9) Fang unarmed (Fig. 58B); eyes reduced, PLE di- ameter less than 0.4 x ALE (Fig. 57); total length greater than 30mm... . Thyropoeus mirandus Pocock — Fang with basal tooth (Fig. 51B); eyes normal, PLE diameter greater than 0.5 x ALE (Fig. 50); total length less than 20mm..... Thyropoeus malagasus (Strand) 11(9) Leg coxae with only ordinary, slender setae (Fig. 34B); carapace lateral margin evenly rounded posteri- OnlyA(E12533) Pres es err case are le — Leg coxae II and III with thorns (Figs. 23B, 25D); carapace lateral margin narrowed posteriorly (Fig. 22) Micromesomma cowani Pocock 12(11) Dorsally legs I and II with slender procumbant setae at apices of tibiae and bases of metatarsi (Figs. DA CHAUB ian eyses sree eorste wos. ero siieke Sears 13 — Dorsally legs I and II with thorns at apices of tibiae and bases of metatarsi (Fig 33) Paramigas goodmani, new species 13 13(12) Dense vestiture of long, curved, silky hairs be- neath at least tibiae and metatarsi I and II (Figs. 33, 41B, 47C) — Long, curved, silky hairs sparse or absent from be- neath tibiae and metatarsi I and II (Figs. 1, 28B) . . 15 14(13) Spermathecae elongate, length greater than 1.67 x head diameter, head diameter less than 2.30 x diam- eter stalk (Figs. 35A—C, 47 G-H) Schaar ileqey veel ateroey ce Paramigas perroti (Simon) — Spermathecae short, length less than 1.67 = head di- ameter, head diameter greater than 2.14 x diameter stalk (Figs. 40A, B, 41C, D) Paramigas oracle, new species 15(13) Tibia I with fewer than 30 ventrolateral spines (Fig. 46B); prefoveal setae small (Fig. 46A) to absent (Fig. 48); spermathecae short, length less than 1.20 x head diameter, head diameter greater than 3.50 x di- ameter stalk (Figs. 45A—C) 16 — Legs I and II densely spinose, tibia I with more than 40 ventrolateral spines (Fig. 1); prefoveal setae large (Fig. 39A); spermathecae long, length greater than 1.63 x head diameter, head diameter less than 1.81 x diameter stalk (Figs. 35E, 39C)................. Paramigas milloti, new species 16(15) Carapace with conspicuous prefoveal setae (Figs. 44A, 46A) 17 — Prefoveal setae minute to absent (Figs. 28A, 48). 18 17(16) Legs I and II dorsally with patches of procum- bant setae at apices of tibiae and bases of metatarsi sparse (<10) (Fig. 44C) asthe Paramigas pauliani (Dresco & Canard) — Legs I and II dorsally with patches of conspicuously serrate, procumbant setae at apices of tibiae and bases of metatarsi dense (>20) (Fig. 46B).............. Paramigas pectinatus, new species 18(16) Labium with fewer than 10 cuspules; dense net- work of reticulate striae on lateral margins of caput (Fig. 28A) and longitudinal banding patterns on legs (Eigs28B) ip oeearotioor Paramigas alluaudi (Simon) — Labium with more than 15 cuspules (Fig. 49B); lat- eral margins of caput without striae and legs not longi- tudinally banded (Fig. 48) Pe nierein aerate ec Paramigas rothorum, new species 19(1) Fang with basal tooth (Fig. 53B) ......... 20 — Fang lacking basal tooth (Figs. 30A, 37B) ..... 24 20(19) Tibia I unmodified, cylindrical; sternal sigilla flat (Fig. 30A); pedipalpal tarsus with apical spinules (igs) 20A=B S37 E—G) aaa veneer ee ae 21 14 — Tibia I swollen prolaterally (Figs. 52, 55A); sternal sigilla deeply incised (Fig. 53B); pedipalpal tarsus lacking apical spinules (Figs. 53E—G, 54B-C)...... Thyropoeus malagasus (Strand) 21(20) OAW <0.41 x carapace width; embolus longer thangbulby(ii ga (AN rece cae a ctetalecicncbere ra 22 —OAW >0.44 x carapace width; embolus length equal to bulb (Legendre & Calderon Gonzalez 1984, fig. X, Goloboffia vellardi (Zapfe) 22(21) Scopulae present beneath tarsi III and IV (Fig. 60A); femora with dorsal spines short or absent (Gris- wold 1987b, fig. 70) 23 — Scopulae absent from tarsi III and IV (Fig. 60B); femora with dorsal spines long (Griswold 1987b, fig. GD) ee ereg ee cee cpehaosrod stacey tuey Res an eucvortaavels Migas 23(22) Dark dorsal and lateral maculations or annuli on tibiae and metatarsi; embolus straight (Figs. 20A—B). ate pied Oa eee ero binitatn a caiman Poecilomigas — Tibiae and metatarsi lacking such dorsal and lateral maculations or annuli; embolus reflexed (Raven 1984 ISNAIKS)) A rae Bare act a. cr Migas variapalpus Raven 24(19) Femur I venter rounded Figs. 7D, 30B); apical lobes of pedipalpal tibia subequal (Figs. 7B, 11C, 32C) 5 hI OEE ROE RAS ee re een Ro 25 — Femur I venter carinate (Fig. 60C); ectal apical lobe of pedipalpal tibia longer than mesal (Figs. 16A, 32D) Serer dta ets COR Cea ec NRGH Rese una Moggridgea (Africa) 25(24) Caput low, not highly arched, height less than 2.0 x height at thoracic fovea; retroapex of tibia I with megaspinei(Figs! 7D) 30B) vase are certs ets ees 26 —Caput highly arched, height greater than 2.3 x height at thoracic fovea; retroapex of tibia I lacking mepgaspine) (Epa) epee ees ss ae eee Heteromigas 26(25) Metatarsus I straight, cylindrical or slightly swollen retroapically (Figs. 29,36) ............ 27 — Metatarsus I swollen for apical half (Figs. 6, 7D). . Calathotarsus 27(26) Metatarsus I cylindrical, unmodified (Figs. SL GN3 SC) Spree e Aachs ps. sien airs tan cae saitodal ites op SEN 28 —Metatarsus I pale and swollen retrolaterally (Figs. 29, 30 Be SiIDIE) wae aciaks Gham arches s BCS De care Paramigas andasibe Raven, new species 28(27) Pedipalpal tarsus with apical spinules (Figs. 37E-G); embolus length less than or equal to bulb lenothi(Gies33SE—G) maces aaeise See ee 29 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 — Apex of pedipalpal tarsus lacking spinules; embolus length much longer than bulb (Fig. 16A).......... Meee tronic ote ees cle Moggridgea (Australia) 29(28) Tarsi III and IV cylindrical, with weak ventral scopulae (Fig. 36); tibia I with 3-4 retroventral spines (ig. 37 @)i saa Paramigas macrops, new species — Tarsi III and IV swollen, curved, sausage shaped, with dense ventral scopulae (Fig. 38D); tibia I with only retroapical megaspine (Fig. 38C)............ Boor ete Paramigas manakambus, new species MIGID GENERA NOT FROM MADAGASCAR Calathotarsus Simon, 1903 (Figs. 4, SA-E, 6, 7A-F, 62A-B, 63A, 65, 66) Calathotarsus (type species C. coronatus Simon, 1903) comprises three species from Argentina and Chile. The genus was described by Simon (1903), the Argentinean fauna was reviewed by Schiapelli and Gerschman (1973), the Chilean fauna was reviewed by Legendre and Calderon Gonzalez (1984), Goloboff and Platnick (1987) and Goloboff (1994), and new species were described by Schiapelli and Gerschman (1975) and Goloboff (1991). Possible synapomorphies for the genus are the modified female pedipalpal tarsus bear- ing a ventral expansion and dorsal group of cuspules (Figs. 5C) and bent male metatarsus I (Figs. 6, 7D) that is swollen for the apical half. The very wide ocular area (Fig. 4), mentioned as a synapomorphy by Goloboff (1991) may relate ‘Migas’ vellardi to Calathotarsus. Calathotarsus are medium sized (12—20 mm) with an arched caput (Fig. 5A) in the female, wide ocular area (except in C. pihuychen Goloboff) and with rows of setae on the caput. The thoracic fovea is simple or may have a weak posterior extension (Fig. 4). The chelicer- al fang furrow has denticles between the tooth rows (Figs. 5D, 7E) and intercheliceral tumescence in the male, the cuspules of the pedipalpal coxae are clustered near the base (Fig. 5B), the sternal sigilla are flat, and there are no thorns on the leg coxae. Female tarsi I and II have spines, patellae II] and IV and tibiae and metatarsi III are densely spinulose, tibiae III are cylin- drical, metatarsi III and IV have broad retrolateral preening combs of long and short setae (Figs. 62A—B), and the ITC have one to a few teeth (Fig. SE). The spermathecae are short, cylindrical and without easily distinguished head and stalk (Fig. 63A). Males have a megaspine on metatarsus I, the pedipalpal tibia lacks spinules, and there are scopulae beneath at least tibiae GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 15 III and IV. Biology has been reported by Schiapelli and Gerschman (1973, 1975) and Goloboff (1991). All Ca- lathotarsus appear to be burrowers rather than nest builders. Goloboff reports that Calathotarsus corona- tus and C. pihuychen make burrows closed with a thick and rigid trapdoor (Goloboff 1991 figs. 9-11) that has several small pits on the inner surface that presumably mark where the spider inserts its fangs to hold the door shut. Heteromigas Hogg, 1902 (Figs. 8, 9A-D, 10, 11A-F, 63B, 65, 66) Heteromigas (type species Heteromigas dovei Hogg, 1902) comprises two species from eastern Aus- tralia. Hogg described H. dovei from Tasmania (Hogg 1902), Raven (1984) described H. terraereginae from Queensland and Raven and Churchill (1989) added notes on the morphology and biology of H. dovei. Pos- sible synapomorphies for the genus are long pedipalpal tarsi of males (Fig. 10) and loss of the male retrolater- al tibia | megaspine (Fig. 11D). Heteromigas are small to medium sized (8—18 mm) with an arched caput (Fig. 9A) in the female, moderately broad ocular area (ocu- lar area width 0.38—0.44 x carapace width) (Fig. 8) and with rows of setae on the caput. The thoracic fovea has a posterior extension in most specimens. The chelicer- al fang furrow has denticles between the tooth rows and the male lacks intercheliceral tumescence (Figs. 9C, 11E). The cuspules of the pedipalpal coxae extend to the apex, the sternal sigilla are flat, and thorns are absent from the leg coxae (Fig. 9B). Female tarsi I and II have spines, patellae III and IV and tibiae and metatarsi III are densely spinulose, tibiae III are cylin- drical and preening combs are absent. Tarsal claws of the female have simple or divided teeth (Fig. 9D) and those of the male have multiple teeth (Fig. 11F). The spermathecae of H. dovei are long and have the head narrower than the stalk (Fig. 63B) whereas those of H. terraereginae are short and cylindrical. Male metatar- sus I lacks a megaspine and the pedipalpal tibia lacks spinules, and there are scopulae beneath at least metatarsi III and IV. Biology has been reported by Raven (1984) and Raven and Churchill (1989). Het- eromigas terraereginae lives in the ground in burrows that are fitted with circular trapdoors impregnated with soil. Heteromigas dovei also makes burrows in the soil that are up to 10 cm deep with a soil-impregnated lid. Mallecomigas Goloboff & Platnick, 1987 (Figs. 12A-F, 63C-D, 65, 66) This monotypic genus (type species Mallecomigas schlingeri Goloboff & Platnick, 1987), known only from Chile, was proposed by Goloboff and Platnick (1987) for a unique female specimen. A possible synapomorphy for Mallecomigas is spermathecae that are long and have the head narrower than the stalk (Figs. 63 C—D). Mallecomigas are large (22 mm) with an arched caput, narrow ocular area (width 0.43 x carapace width) (Fig. 12A) and with rows of setae on the caput. The thoracic fovea is simple with a faint, shallow posterior extension. The cheliceral fang fur- row lacks denticles between the tooth rows (Fig. 12E), the cuspules of the pedipalpal coxae are clustered near the base, the sternal sigilla are flat, and thorns are ab- sent from the leg coxae (Fig. 12B). Female tarsi I and II have spines (Fig. 12C), patellae III and IV and tibi- ae and metatarsi III are densely spinulose, tibiae III are cylindrical (Fig. 12D) and metatarsi III and IV have preening combs of setae that are uniformly long and separated at the base by distances greater than their di- ameter (Goloboff & Platnick 1987:9). The comb on metatarsus IV is inconspicuous, being formed of very thin setae. The spermathecae are characteristic: they are long and have the head narrower than the stalk (Figs. 63C—D). Tarsal claws have few teeth (Fig. 12F). The male is unknown and nothing has been published on the biology. The collector indicated that the speci- men had been dug from a deep hole in the ground with- out a door (E. Schlinger, pers. commun.). Migas L. Koch, 1873 (Figs. 13A-E, 14A-F, 19C, 59A-C, 60B, 63E-F, 65, 66) With 34 described species from Australia, New Caledonia, and New Zealand, Migas (type species Migas paradoxus L. Koch, 1873 from New Zealand) 1s the largest migid genus. It comprises a disparate as- semblage of species united by the presence of a tooth near the base of the fang, preening combs in most species, and lacking the synapomorphy (tibial banding) of Poecilomigas. We do not know the genus in detail but our species examined do include the type species and comprise species from the full geographic range of the genus: Migas affinis Berland from New Caledonia, M. distinctus O. P. Cambridge, M. gatenbyi Wilton, M. giveni Wilton, M. paradoxus L. Koch, and M. taierii Todd from New Zealand, and M. nitens Hickman and M. variapalpus Raven from Australia. Most of our other data come from the literature. Our analysis sug- gests that the Chilean species Migas vellardi is mis- placed in Migas and is more closely related to the other South American genera Calathotarsus and Mallecomi- gas (see Goloboffia). The composition and affinities of Migas constitute the greatest remaining problem in migid taxonomy. The species from Australia, New 16 Caledonia and New Zealand are similar in being small to medium sized (6—-17mm) with a nearly flat caput no higher than the thoracic region (Fig. 13A), ocular area moderately wide (OAW 0.37—0.49 x carapace width) (Fig. 13B) and with well developed rows of setae on the caput leading to a pair of enlarged prefoveal setae. The thoracic fovea is simple. The cheliceral fang fur- row lacks denticles between the tooth rows (Fig. 13D) and intercheliceral tumescence in the male may be present (some CASC and AMNH specimens deter- mined as M. variapalpis, although these may be a dis- tinct species [Goloboff 1991:71]) or absent (holotype of M. variapalpis [Raven in Goloboff 1991:71]), the cuspules of the pedipalpal coxae extend to the apex, the sternal sigilla are flat, and thorns are absent from the leg coxae. Female tarsi I and II lack spines (Fig. 13C), patellae III and IV and tibiae and metatarsi III are sparsely spinulose and tibia III is cylindrical (Fig. 59A). A dense vestiture of long, curled hairs beneath legs I and II may be present (e.g., M@. gatenbyi) or ab- sent (e.g., M. giveni), leg I and II dorsal procumbant setae are absent (M. affinis has retrodorsal patches of thorns at the bases of metatarsi I and II that resemble those of the Malagasy Paramigas goodmani but it lacks procumbant setae on the tibiae) and preening combs of widely spaced or mixed long and short setae are present in most species (Fig. 59C), though absent at least from M. affinis and M. distinctus. Tarsal claws of the female are simple or have a few teeth (Fig. 13E). Males typically lack scopulae from tarsi III and IV (Fig. 60B) but have a metatarsus I megaspine and spin- ules on the pedipalpal tarsus. The spermathecae of at least some New Zealand species are folded so that they appear mushroom-like (Fig. 19C): this morphology may prove to be a synapomorphy uniting at least some Migas species. Males have spinules at the apex of the pedipalpal tarsus. The biology of several species of Migas has been reported. Species from New Zealand and Australia ap- pear to make nests or burrows, and some species (e.g., M. distinctus) are reputed to make both (Wilton 1968: 77). The biology of New Zealand species was summa- rized by Wilton (1968). He reports that Migas con- struct unbranched trap-door nests that comprise a “flat- tened silk tube with a hinged lid with exposed surfaces camouflaged and strengthened with material gathered from the surrounding terrain. Nests may be on the trunks of trees, in moss on banks, under stones, in crevices . . . or in burrows in sand or clay banks” (Wilton 1968:77, figs. 221, 412). Nests have one or in rare cases two doors. Todd (1945) studied the biology of Migas. Raven and Churchill (1989) report that OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 Migas nitens makes burrows directly in the soil with a thin door whereas M. plomleyi makes a camouflaged silken, sac-like nest. Moggridgea O. P. Cambridge, 1875 (Figs. 1SA-F, 16A-B, 17A-C, 32D, 60C, 61A_D, 62C_D, 65, 66) Moggridgea (type species Moggridgea dyeri O. P. Cambridge, 1875) comprises 33 species from Africa and some surrounding islands and from Australia. This genus was described in 1875 by Cambridge for Mog- gridgea dyeri from South Africa, placed in the Migeae by Simon (1892) and transferred from the Miginae to Paramiginae by Raven (1985). Moggridgea are record- ed from mainland Africa by Benoit (1962), Cambridge (1875), Griswold (1987a), Hewitt (1913a, 1913b, 1913c, 1914, 1915a, 1916, 1919), Lawrence (1928), Purcell (1903), and Simon (1903), from the surround- ing islands of Bioko (Griswold & Ubick 1999), the Co- moros (Griswold 1987a), Principe (Simon 1907), and Socotra (Griswold 1987a). Griswold (1987a) mono- graphed the African species and Main (1991) described two new species from Australia. Our analysis suggests that Moggridgea should be returned to the Miginae (sensu Simon 1903) (placement supported by a low caput and loss of spines from female leg tarsi) and that the dorsobasal depression on female tibia III, consid- ered by Raven (1985) to be a synapomorphy of the Paramiginae, has arisen in parallel in Moggridgea. Possible synapomorphies for the genus are the lamel- late setae beneath at least patellae I, Il and IV (with ho- moplasy in Goloboffia vellardi), a dorsobasal depres- sion on female tibia III (with homoplasy in the Mala- gasy migids), and a median sclerotized band across the spermathecae. Thorns on the leg coxae were postulat- ed to be a synapomorphy uniting Moggridgea and Mi- cromesomma (Griswold 1987a:109); our new analysis suggests that these have evolved in parallel. Mog- gridgea are small to large (S—25mm) with a low caput (Fig. 15A), moderately broad ocular area (OAW 0.42—0.86 x caput width) (Fig. 15B) and with at least a few setae on the caput. Prefoveal setae may be absent, small, or enlarged. The thoracic fovea is simple or may have a weak posterior extension. The cheliceral fang furrow lacks denticles between the tooth rows (Fig. 15E) and intercheliceral tumescence in the male may be present (M. tingle) or absent (most African species). The cuspules of the pedipalpal coxae extend to the apex, the sternal sigilla are flat, and thorns on the leg coxae may be present or absent. Female tarsi I and II lack spines (Fig. 15C), patellae III and IV and tibiae GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS and metatarsi III are densely spinulose, tibiae III have a weak (most species) to moderate (M. breyeri Hewitt) dorsobasal depression (Figs. 15 D, 61 A~B) and metatarsus IV has a preening comb (Figs. 62C—D) of a few, closely spaced setae, in most species on a com- mon base (some lack this comb, e.g., M. anactenidia Griswold). Peculiar, lame!late setae are present beneath at least patellae I, II and IV, occurring in both sexes of most species (Figs. 16B, 61D). Tarsal claws may have simple (Fig. 15F) or multidentate (Griswold 1987a fig. 37) teeth. The spermathecae are short to long, cylindri- cal to sinuate, with the head slightly wider than the stalk (M. ampullata has the head narrower than the stalk: Griswold 1987a fig. 80) and have a median trans- verse sclerotized band (Figs. 17A—C). Males lack tibia I megaspines (Fig. 16B) and spinules on the pedipalpal tarsus (Fig. 16A), most species have scopulae beneath metatarsi III and IV and at least most of the African species have male femur I carinate ventrally (Fig. 60C) and the apical lobes of the pedipalpal tarsus unequal (Figs. 16A, 32D). The biology of African species was summarized by Griswold (1987a) and that of Aus- tralian species by Main (1991). The majority of African species for which biology is known make a bag or sac-like nest, oval to pear shaped, constructed of silk mixed with bits of the surrounding substrate, pro- vided with a single wafer or cork trap door, and placed in a niche or crevice in the substrate (Griswold 1987a, figs. 2-3). Most nest building species have been taken from the trunks of trees though some build their nests on rock faces or stone walls. The Australian M. tingle constructs similar nests (Main 1991, fig. 1). Some species make a true silk-lined tubular burrow, excavat- ed in the substrate and usually terrestrial. At least the African M. mordax, M. peringueyi, M. terrestris and M. terricola and the Australian M. australis are bur- rowing species. Poecilomigas Simon, 1903 (Figs. 18, 19A-B, 20A-C, 59D, 60A, 65, 66) This genus (types species Poecilomigas pul- chripes Simon 1903) comprises three species from southern and eastern Africa. The genus was proposed by Simon (1903), revised by Griswold (1987b), and Griswold (1998b) provided additional data on behavior and taxonomy. Possible synapomorphies for the genus are annulate tibiae (Fig. 18) and, implied by our analy- sis, the loss of preening combs from metatarsus IV. Poecilomigas are small to large sized (females 6.0-26.0 mm) with a low, nearly flat caput, moderate- ly wide ocular area (OAW 0.5—0.65 width caput) and 17 with rows of setae on the caput and enlarged prefoveal setae (Fig. 18). The thoracic fovea is simple. The che- liceral fang furrow lacks denticles between the tooth rows (Fig. 20C) and intercheliceral tumescence occurs in the male, the cuspules of the pedipalpal coxae ex- tend to or near to the apex, the sternal sigilla are flat, and thorns are absent from the leg coxae. Females are with or without dense, long silky hairs beneath legs I and II, tarsi I and II lack spines (Fig. 18), patellae III and IV and tibiae and metatarsi II] are weakly spinu- lose, and preening combs are absent. The spermathecae are short to long, cylindrical, uniformly sclerotized and have a well-defined head and stalk (Figs. 19A—B). Males are with or without a retroapical megaspine on tibia I, have spinules at the apex of the pedipalpal tar- sus (Figs. 20A, B), and have scopulae beneath tarsi IT and IV (Fig. 60A). Biology has been reported by Cam- bridge (1889), Pocock (1895), Hewitt (1915b), Gris- wold (1987b) and Griswold (1998b). Both P. abrahami and P. basilleupi appear to build nests, usually on the trunks of trees. Although Cambridge, Pocock, and He- witt report P. abrahami making nests with one or two doors, Griswold found only 2-door nests for this species and it is likely that the 1-door nests were made by other sympatric migids. The nests of P. abrahami are vertical, situated in a crevice and woven of silk and bits of the substrate, and are provided with a wafer type door at the upper and lower end (Griswold 1987b, figs. 16, 17). Poecilomigas basilleupi makes a similar nest with a single door at the upper end (Griswold 1998b, figs. 2-3). Goloboffia, new genus (Figs. 14A-F, 63E-F) Type species—Migas vellardi Zapfe 1961, male holotype from La Herradura, Coquimbo Province, Chile, deposited in Museo Nacional de Historia Natur- al, Universidad de Chile, Santiago, examined. Etymology—Named in honor of Pablo Goloboff, student of phylogenetic systematics and collector of many new and interesting South American mygalo- morphs. Gender feminine. Diagnosis ——Distinguished from migids other than Calathotarsus by having a very wide ocular area (OAW >0.6 x carapace width) and from migids other then Moggridgea by having ventral lamellate setae on the patellae. Goloboffia is distinguished from Calatho- tarsus by lacking a ventral expansion and dorsal group of cuspules on the female pedipalpal tarsus and bent male metatarsus I and from Moggridgea by lacking a dorsal excavation on female tibia III and having tarsal spines and spinules on the tibia of the male pedipalpus. 18 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 Note——We transfer Migas vellardi because it does not belong in Migas. It shares with the other South American genera Calathotarsus and Mallecomigas the synapomorphies of pedipalpal coxa cuspule distribu- tion that is concentrated at the proximal edge near the labium, slender spinules densely situated on patella- tibia III of females, and preening combs that consist of many setae that extend part way around the ventral apex of the segment and that occur on both metatarsi III and IV. We place Migas vellardi in the new genus Goloboffia because it differs from both Calathotarsus and Mallecomigas in several features both diagnostic and synapomorphic for these genera. Synapomorphies for Goloboffia are the low caput, basal tooth on the fang, ventral lamellate setae on the pate!lae, apical spinules on the male pedipalpal tibia and loss of the male metatarsus I megaspine. Description—The male was described by Zapfe (1961: 153) and redescribed by Calderon and Calderon Gonzalez (1983) and Legendre and Calderon Gonzalez (1984) and the female was described by Goloboff and Platnick (1987). Small (6—8mm); caput raised above thoracic region (though not highly arched), the ocular area very wide (OAW >0.6 x carapace width); with few setae on the caput and small prefoveal setae (Fig. 14A); thoracic fovea simple, nearly straight; chelicerae with anteromedian brush of stout setae (like Calatho- tarsus), fang with basal tooth and fang furrow with few (female) or many (male) denticles between tooth rows (Fig. 14E); male with intercheliceral tumescence; cus- pules of pedipalpal coxae restricted to base (Fig. 14B); sternal sigilla flat; female tarsi I and II with spines (Fig. 14C), patellae I, II and IV with ventral lamellate setae (Fig. 14C), patellae III and IV and tibiae and metatarsi III densely spinulose (Fig. 14D); metatarsi II] and IV with broad preening combs; female tarsal claws slen- der with short teeth (Fig. 14F); spermathecae short and cylindrical (Figs. 63E—F); male lacks tibia I megaspine; pedipalpal tarsus with apical spinules; scopulae beneath male metatarsi II and IV. Natural History—Goloboff (1991) reports that Migas vellardi is terrestrial and makes burrows closed with a thin and rigid trapdoor lacking beveled edges (Goloboff 1991, figs. 12-13). The door fits over the burrow entrance and has two mesal holding marks on the inner surface that presumably mark where the spi- der inserts it fangs to hold the door shut. Composition—One species. Distribution—Northern Chile (Fig. 66). MIGIDAE OF MADAGASCAR Paramiginae Petrunkevitch 1939 Myrtaleae Simon 1892: 84 (unavailable through homonymy of type genus) Paramiginae Petrunkevitch 1939: 154 (nomen novum). Roewer 1942: 192. Bonnet 1958: 3329. Raven 1985: 144. Diagnosis——Migids with reduced caput setation, with prefoveal setae only or lacking setae altogether posteriad of ocular area (Figs. 1, 2A, 33, 50), femur III with a ventral membrane that extends at least 4/5 the length of the femur (Fig. 42C), and tibia III with a deep dorsobasal excavation (Figs. 25A, C, 42A—B) and an- terior diagonal ridge (Fig. 25A, 42A), and most species with an angular and strongly tripartite thoracic fovea (Figs. 2A, 22, 46A). Genus Micromesomma Pocock 1895 Micromesomma Pocock 1895:190 (type species, by mono- typy, Micromesomma cowani Pocock 1895). Roewer 1942: 192. Bonnet 1957:2886. Dresco & Canard 1975:783-788. Raven 1985:145. Griswold 1987a:109-117. Platnick 1989: 71. Dippenaar-Schoeman & Jocqué 1997:77. Platnick 2001. Diagnosis——Distinguished from all other migid genera, except some Moggridgea, by having thorns on the ventral surfaces of coxae II and III (Figs. 23B, 25D), and from Moggridgea by having tarsal spines in the female (Fig. 22), a group of dorsal thorns at the apices of tibiae I and II (Figs. 24A—B), and lacking lamellate setae beneath the patellae. Note.—A possible synapomorphy for the genus is the thorns on leg coxae. Description—Small to large, 6.0-22.0 in length. Sparsely setose. Carapace smooth, length 1.14—1.21 x width, posteriorly narrowed (Fig. 22); caput with me- dian ocular seta, prefoveal setae present. Ocular area 0.43—0.52 = width caput; AER straight to slightly re- curved, ALE equal to or smaller than AME; PER slightly recurved, PLE equal to or smaller than PME; ocular quadrangle narrowed anteriorly. Thoracic fovea tripartite, recurved, width 0.19—0.30 x width carapace. Sternum length 1.30—1.53 x width, margin sinuate; se- tose along margin, sparsely setose on surface; sigilla oval to irregular, shallowly excavated, length 2.25—3.0 x sternum width (Fig. 23B). Thorns present at least on coxae II and III. Leg formula 4123; tibiae, metatarsi, and tarsi I and II ventrally with sparse distribution of short setae; femur II] with ventral membrane extending to base, tibia II] with deep dorsobasal excavation and anterior ridge (Figs. 25A—B). Spermathecae with broad head and narrow stalk (Figs. 23D—-E, 26A—C). Male unknown. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 19 Natural History—Unknown. Composition——One species. Distribution—Central Madagascar. Micromesomma cowani Pocock 1895 (Figs. 22, 23A—E, 24A-E, 25A-B, D, 26A-C, 65, 68) Micromesomma cowani Pocock, 1895:190—191 (two speci- mens from Central Madagascar, purchased of Rev. Deans Cowans, lectotype female, here designated in order to en- sure the name’s proper and consistent application, and paralectotype female, BMNH 82.26, in BMNH, exam- ined). Roewer 1942:192. Bonnet 1957:2886. Dresco & Ca- nard 1975: 783-788. Raven 1985:145. Griswold 1987a:109-117. Platnick 1989:71. Platnick 2001. Note——The types are from “Central Madagascar” (Pocock 1895: 191). There are no precise locality data with any specimen, so the distribution within Mada- gascar remains a mystery. Diagnosis——Same as for genus. Description—Female (lectotype): Total length 19.1. Carapace dark yellow-brown; caput and ocular area light yellow-brown with dark striae extending along lateral margins and from PME to thoracic fovea, black surrounding each AME and extending between each ALE and posterior eyes; chelicerae dark yellow- brown, coxae and trochanters yellow-brown, sternum light yellow-brown except darker at margin, pedipalpal coxae yellow-brown shading to pale yellow along an- terior margin; legs and pedipalpi dorsally dark yellow- brown, light yellow-brown ventrally on patellae and tibiae; abdomen faded to yellow-white. Carapace 6.0 long, 5.1 wide, height at thoracic fovea 0.29 x carapace width; smooth. Caput inclined (Fig. 23A); height 1.33 x that at thoracic fovea, 0.74 x carapace width; medi- an ocular setae situated behind and between AME, two pairs of setae positioned anteriad to this; clypeus 0.55 x length OAL, margin procurved. Thoracic fovea re- curved and tripartite, width 0.19 x that of carapace, 1.33 x wider than long, with pair of prefoveal setae. Ocular area width 0.52 x caput, 2.22 x wider than long; AER 2.0 wide, 1.14 x width PER. Ratio of eyes: AME: ALE: PME: PLE: 1.0: 2.13: 1.4: 1.2, diameter AME 0.27; AME separated by 0.45 x their diameter, PME by 5 x their diameter. Ocular quadrangle 1.43 x wider than long, posterior width 1.64 x anterior. Sternum 4.1 long, 3.1 wide, widest behind coxa II and narrowed an- teriorly, setose laterally and sparsely setose on surface; single thorn situated posteriad to labial suture; coxae I with 3—5, II with 17—20, and III with 35—37 thorns; sig- illa 2.2 long, 1.0 wide, shallow, irregularly shaped, ad- jacent to coxa II, width 0.31 x width sternum, distance between 0.5 x distance from margin. Labial sigilla dis- tinct, swollen; labium with 27 and pedipalpal coxae with 34-36 cuspules; labium 0.85 long, 1.15 wide, pedipalpal coxae 1.95 long, 1.1 wide, apex produced to a sharp point. Chelicerae 2.1 long, promargin of fang furrow with 3 teeth, retromargin with | large basal tooth, pro-and retromargin interspersed with 7 denti- cles, with pale swelling at base of tooth rows. Femur I 0.84, tibia I 0.49, femur IV 0.83, and tibia IV 0.52 x carapace width. Spination: pedipalpus: tibia v1-0-0, tarsus p 1-1-1-1, r 0-1-1-0; leg I: tibia p 2-4-2-4, r 3-4- 3-3-3, d 14 thorns at apex of tibia, metatarsus p 5-4-la, r 4-4-3-la, tarsus p 1-1-1, r 1-1-1; leg II: tibia p 2-1-3- 1-3, r 1-2-2-1-2, d 27 thorns at apex of tibia, metatar- sus p 3-3-2-la, r 0-2-2-2, tarsus p 1-1-1-1, r 1-0-0; leg III: patella with approximately 60, tibia with approxi- mately 110, metatarsus with approximately 130, and tarsus with approximately 70 spinules; leg IV: patella with approximately 70 spinules and tibia with approx- imately 80 stout setae, metatarsus p 0-0-1-la. Femur II with proximal ventral rows of 10-15 stout setae. STC teeth (pro-retro): I, II, HI (2-2), IV (1-1); ITC simple, pedipalp claw with | tooth (Fig. 23F). Leg measure- ments (Femur + Patella + Tibia + Metatarsus + Tarsus = [Total]): I: 4.3 + 2.3 + 2.5 + 2.1 + 1.1 = [12.25]; II: 3.75 + 2.25 +2.2 + 1.8 + 1.5 = [11.25]; III: 3.05 + 1.9 +2.15+1.9+ 1.5 =[10.5]; IV: 4.25 + 2.45 +2.7+2.15 + 2.15 = [13.7]; pedipalpus: 2.9 + 1.25 + 1.4 + (absent) + 1.25 = [6.25]. Abdomen 11.0 long, 7.0 wide, sparse- ly covered with short setae. Spermathecae not removed from type specimens, but glandular covering visible through cuticle suggests they are like those described from other specimens below (Figs. 23D—E, 26A—C). Variation (N = 4)—Total length 16.5—19.4; height at fovea 0.2-0.29 x carapace width. Caput width 0.73—-0.78 x carapace width, height 1.33—2.08 x height at thoracic fovea; width ocular area 0.43—-0.52 x caput width, PLE 0.4—0.75 x PME; clypeus length 0.33—0.55 x OAL; thoracic fovea width 1.33—2.12 x length. Che- liceral fang furrow interspersed with 4-8 denticles (Figs. 23C, 24E). Sternum with or without cuspules, sigilla oval to irregularly shaped; labium with 19-39, pedipalpal coxae with 40-45 cuspules; coxa I with 0-6, coxa II with 24-37, and coxa III] with 35-45 thorns. Tibia I (Fig. 24A) with 11-19, metatarsus I with 7-14 retroventral, tibia II with 6-15 proventral spines. STC with 1—2 teeth. Spermathecae (N=6 specimens) length 0.55—0.84 x distance between them and 1.12—1.35 x head diameter, diameter head 3.12-4.0 x diameter stalk, head length 2.144.0 x length stalk (Figs. 23D-E, 26A-C). Material Examined—MabDaGascar: (no other lo- cality) 1954 A. Verdier (3 females, MRAC 147.158); Le Barbier No. 29-1920, “Fage det. Revu par Dresco 20 en 1974” (19 females, MNHN AR4136, 1 female, CASC); Central Madagascar, Rev. Deans Cowans (lec- totype and paralectotype females, BMNH). Genus Paramigas Pocock 1895 Paramigas Pocock 1895:188 (type species, by monotypy, Paramigas subrufus Pocock 1895 [=Myrtale perroti Simon 1895]). Simon 1892:32. Roewer 1942:193. Dresco & Ca- nard 1975:783—788. Raven 1985:144. Platnick 1989:73. Dippenaar-Schoeman & Jocqué 1997:77. Platnick 2001. Legendrella Dresco & Canard 1975:786 (type species, by monotypy, Legendrella pauliani Dresco & Canard, 1975). Brignoli 1983:119. First synonymized by Raven 1985:144. Diagnosis.—Distinguished from all other migid genera by the presence of retrodorsal procumbant setae or thorns at the apices of tibiae and bases of metatarsi I and II (Figs. 27A-C, 46B), from Moggridgea also by having spines on female tarsi I and II (Figs. 1, 28B), and from all other genera except some Moggridgea and the Malagasy Micromesomma and Thyropoeus by hav- ing tibia III with a deep dorsal basal concavity (Figs. 36, 42A—B). Males have spinules at the apex of the pedipalpal tarsus (Figs. 30E, 32C, 37F), a retroapical megaspine on tibia I (Figs. 30B, 37C), and a tripartite thoracic fovea (Figs. 29, 36, 38A). Note——Possible synapomorphies for the genus are procumbant and stout dorsal setae at the base of metatarsi I and II and apical spinules on the male pedi- palp tarsus. Description—Small to large sized mygalo- morphs, 6—22 mm in length. Carapace smooth in fe- male, rugose in male, length I—1.3 x width in female, 0.98—1.02 x width in male, posteriorly narrowed in fe- male, rounded in male; caput with median ocular seta, prefoveal setae present (Fig. 1) or absent (Fig. 48). Oc- ular area 0.41—0.55 x width caput; AER straight to slightly recurved, ALE equal to or smaller than AME; PER straight to slightly recurved, PLE equal to or smaller than PME; ocular quadrangle narrowed anteri- orly. Thoracic fovea recurved, tripartite, 0.11-0.25 x width carapace. Fang furrow of females with pale swelling at base of tooth rows (Fig. 43B). Sternum length 1.10—1.53 x width, margin sinuate; setose along margin, sparsely setose on surface; sigilla oval to ir- regularly shaped, shallow, length 0.21—0.42 x sternum width. Labial suture distinct; cuspules present on labi- um and pedipalpal coxae of female, absent from male (Fig. 30A); thorns absent from coxae. Leg formula 4123 (female) or 1423 (male); coxae, legs and pedi- palpi setose; legs I and II with strong series of spines on pro-and retrolateral margins (Figs. 1, 27A); retrodorsal procumbant setae or thorns at the apices of OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 tibiae and bases of metatarsi I and II (Figs. 27A-C); tibiae, metatarsi, and tarsi I and II ventrally with sparse (Figs. 44C, 48) or dense (Figs. 27A, 33, 47C) distribu- tion of long filiform setae; legs III and IV with dense patterns of spinules and stout setae pro-and retrodor- sally gradually changing to slender setae distally in fe- male, spination weak in male; femur III with ventral membrane extending to base (Fig. 42C), tibia III with deep dorsobasal excavation and anterior ridge; scopu- lae absent from female, present in male. Male with retrolateral tibia I megaspine, spinules at apex of pedi- palpal tarsus, bulb uniform, embolus simple and slen- der, embolus short, 0.25—0.55 x bulb length. Sper- mathecae (Figs. 35A—E, 40A—D, 45A—C) paired, sim- ple, unbranched, straight to distally curved, length 0.67—-1.2 x distance between bases, with pores throughout, unsclerotized, with distinct head and stalk, head 1.8—7.5 x stalk width. Natural History—The biology of only a few species is known. All Paramigas adults for which col- lection data are available have been arboreal but some adult females have been taken from pitfall traps and some juveniles have also been sifted from leaf litter. Pocock (1895:190, figs. la, 1b) described two trap door nests received with the specimens of Paramigas subrufus. These are oval with a single wafer type door, woven of silk, and covered on three sides and the door with moss, lichen and pieces of bark (Figs. 21A—B). Several nests of P. oracle were collected from depres- sions on tree trunks and vines at Ranomafana. These nests are oval with length 2 x the spider’s total length, with a single wafer type door at the upper end. The ex- posed surfaces are covered with moss, lichen and pieces of bark, effecting a superb camouflage. All nests are made of fragments of bark and lichen woven to- gether with silk (Figs. 21C—D; Pocock 1895, figs. la, 1b). Composition—Eleven species. Distribution——Probably all of Madagascar (Fig. 68). Paramigas alluaudi (Simon) (Figs. 28A—E, 45C, 65, 68) Myrtale alluaudi Simon 1903: 133 (type female, Madagas- car, Fort Dauphin, MNHN, examined). Paramigas alluaudi, Roewer 1942: 193. Bonnet 1958: 3329. Platnick 2001. Diagnosis.—Females are distinguished from other Paramigas that lack a dense vestiture of long silky hairs beneath legs I and II by the dense lateral network of striae on the caput (Fig. 28A), longitudinal banding pattern on legs (Fig. 28B), and the thinly distributed GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 21 cuspules on the pedipalpal coxae (<20) and labium (<8); male unknown. Description—Female (holotype): Total length 12.7. Carapace, chelicerae, and legs light brown, ven- ter, coxae, and trochanters yellow-brown, lateral mar- gins of caput yellow-brown with light brown band ex- tending from PME to thoracic fovea; dense lateral net- work of light brown striae on caput and extending from behind ocular area to thoracic fovea, also radiating from thoracic fovea to carapace margin (Fig. 28A); carapace with dark rebordered margin; ocular area light brown, black surrounding AME, behind ALE and between PLE and PME; clypeus dark; femora, patellae and tibiae with narrow dorsal and broad pro-and retro- lateral maculate bands, these united at base of patellae and bases and apices of femora and tibia, area between bands may have reticulate markings, tarsi yellow- white, unmarked; abdomen dark purple-brown includ- ing basal segment of spinnerets, distally segments yel- low-brown. Carapace 2.9 long, 2.55 wide, height at thoracic fovea 0.39 x carapace width; smooth. Caput inclined, height 1.4 x that at thoracic fovea, width 0.76 = cara- pace width; median ocular seta present with a pair of setae anterior to this, prefoveal setae minute, visible at greater than 80x magnification; clypeus height 0.4 times length OAL, margin straight. Thoracic fovea re- curved, tripartite with weak posterior depression, width 0.23 x that of carapace, 6.0 x wider than long. Ocular area width 0.41 x caput, 2.13 x wider than long; AER 0.8 wide, 1.0 x width PER. Ratio of eyes: AME: ALE: PME: PLE: 0.53: 1.0: 0.60: 0.80, diame- ter AME 0.1; AME separated by 1.25 x their diameter, PME by 3.25 x their diameter. Ocular quadrangle 0.40 x wider than long, posterior width 1.53 x anterior. Sternum 2.35 long, 1.67 wide, widest behind coxa II and narrowed anteriorly, sparsely setose; sigilla ir- regularly oval, adjacent to coxae II and III, width 0.25 x width sternum, distance between 0.71 x distance from margin. Labium with 6 and pedipalpal coxae with 16-18 cuspules; labium 0.5 long, 0.6 wide, pedipalpal coxae 0.72 long, 0.55 wide, apex produced to a blunt point. Chelicerae 0.45 long, fangs broad, flaring pro- laterally before apex, promargin of fang furrow with 3 teeth, retromargin with 4 teeth (Fig. 28D). Femur I (Fig. 28B) 0.80, tibia I 0.44, femur IV 0.76, and tibia IV 0.41 x width carapace. Spination: pedipalpus: tibia p1-0-0, tarsus p 1-1-1, r 1-0-0; leg I: tibia p 1-1-1-1-1, r 1-1-1-0, metatarsus p 1-3-2, r 1-3- 2; tarsus p 1-1, r 1-1; leg Il: tibia p 1-0-2, r 0-1-1-1, metatarsus p 3-1-1, r 1-0-2, tarsus p 0-2, r 1-0; leg I: patella with approximately 16, tibia with approximate- ly 20 spinules, metatarsus: vO0-0-la, tarsus with 4—5 proapical spinules; leg IV: patella with approximately 50 slender spinules, metatarsus v 0-0-la. Femur II with proximal ventral row of 5—6 stout setae, retrodorsum of tibiae I and II with 7-8 apical, metatarsi I and II with 3—6 basal stout procumbant setae. STC teeth (pro- retro): I, II (2-2), IM, IV (1-1); ITC simple, pedipalp claw with | tooth (Fig. 28E). Leg measurements (Femur + Patella + Tibia + Metatarsus + Tarsus = [Total]): I: 2.05 + 1.17 + 1.12 + 0.87 + 0.70 = [5.91]; II: 1.82 + 1.05 + 0.87 + 0.75 + 0.72 = [5.21]; Ill: 1.5 + 0.9 + 0.8 + 0.77 + 0.75 = [4.72]; IV: 1.95 + 1.12 + 1.05 + 0.87 + 0.87 = [5.86]; pedipalpus: 1.25 + 0.67 + 0.72 + (absent) + 0.47 = [3.11]. Abdomen 3.0 long, 2.25 wide, sparsely covered with short setae. Spermathecae with broad head and short, narrow stalk, length spermathecae equal to dis- tance between and to their head diameter, diameter head 3.7 x diameter stalk, head length 3.6 x length stalk (Figs. 28C, 45C). Material Examined—MabaGascar: Toliara: Fort Dauphin, 8 November 1901, Ch. Alluaud (Myrtale al- luaudi female, type, MNHN 26248, AR4123). Distribution—Known only from the type locality at the southern tip of Madagascar (Fig. 68). Paramigas andasibe Raven, new species (Figs. 29, 30A—G, 31A-H, 32A-C, 65, 68) Types.—Holotype male from An’ Ala, 9 km E An- dasibe, 840 m, 15 February 1995, primary rainforest, F. Glaw, deposited in ZFMK. Note——Robert Raven provided us with the speci- men and descriptive notes. We attribute authorship of the new species to him. Etymology—tThe specific name is a noun in ap- position from the type locality. Diagnosis——Distinguished from other Paramigas males by having pallid, swollen, weakly sclerotized areas on the retrolateral surface of metatarsus I (Figs. 30B, 31D—E) and AME smaller than ALE (Fig. 29). Description—NMale (holotype): Total length 17.1. Carapace dark red-brown with a dark brown band mar- ginal surrounding; caput with light red-brown areas along lateral margin; ocular area dark red-brown, black behind ALE and between PME and PLE; chelicerae dark red-brown; sternum, coxae, and trochanters or- ange-brown except sterno-labial junction red-brown; dorsal surface of legs I-IV red-brown gradually fading to yellow-brown on tips of tarsi and on ventral sur- faces, femora with dorsal and retrodorsal longitudinal dusky bands; prolateral surface of metatarsus I fading from red-brown to yellow-brown, retrolateral surface 22 yellow-white, bulging, tarsi I and II yellow-brown; pedipalpi yellow-brown; abdomen dark purple-brown; spinnerets and book lung covers yellow-brown. Carapace 8.0 long, 7.9 wide, height at thoracic fovea 0.24 x carapace width; strongly rugose. Caput low, height 1.15 x that at thoracic fovea, width 0.51 x carapace width; with minute pair of prefoveal setae; median ocular seta present; pair of setae situated ante- riad of this; clypeus height 0.34 x length OAL, trans- versely striate, margin straight. Thoracic fovea tripar- tite (Fig. 29), weakly recurved, width 0.15 x that of carapace, 1.71 x wider than long. Ocular area width 0.59 x caput, 2.13 x wider than long; AER 2.45 wide, 1.08 x width PER. Ratio of eyes: AME: ALE: PME: PLE: 1.0: 1.3: 0.6: 0.5, diameter AME 0.5; AME separated by 0.2 x their diameter, PME by 3.33 x their diameter. Ocular quadrangle 1.39 x wider than long, posterior width 1.6 x anterior. Sternum 6.4 long, 4.2 wide, widest behind coxa II and narrowed anteriorly, sparsely setose laterally and on surface; sigilla irregularly shaped, adjacent to coxae I, 0.33 x width sternum, distance between 0.1 x dis- tance from margin. Labium and pedipalpal coxae lack- ing cuspules; labium 1.5 long, 1.4 wide, pedipalpal coxae 3.3 long, 1.6 wide, apex produced to a blunt point. Chelicerae 1.6 long, promargin of fang furrow with 4—5 teeth and retromargin with | basal tooth and 4-5 distal denticles (Fig. 31G) Femur I 1.15, tibia 1 0.83, femur IV 1.08, and tibia IV 0.89 x width carapace. Scopulae beneath distal % of tarsus I and beneath entire tarsi I-IV (Figs. 31E, F). Spination: leg I: patella p 0-2-2, v 0-0-4, tibia p 1-0-0- 0, r 2-3-1-la, metatarsus p 2-2-0-1, r 1-0-1-1, tarsus p 1-1-1, r 1-1-1; leg Il: patella r 0-0-1, tibia p 0-0-1-1, r 2-3-0-2, tibia p 2-2-2-1, r 2-0-0-2, tarsus p 1-1-1-1, r 1- 1-0-0; leg III: patellae with approximately 41 prodorsal and 18 retrodorsal spinules, tibia with approximately 50 apical spinules, metatarsus v 0-0-0-4; leg IV: patel- la with approximately 150 prodorsal spinules, metatar- sus v 0-0-2-3. STC teeth (pro, retro): I, II (1-2), II, IV (1-1) (Fig. 31H). Leg measurements (Femur + Patella + Tibia + Metatarsus + Tarsus = [Total]): I: 9.1 + 4.5 + 6.6 + 6.1 + 2.0 = [28.3]; Il: 8.0+4.04+5.6+5.3+2.2 = [25.1]; Ill: 6.1 + 3.2 + 4.2 + 4.1 + 2.7 = [20.3]; IV: 8.6 + 3.9 + 7.1 + 5.8 + 3.5 = [28.9]; pedipalpus: 4.9 + 2.2 + 3.7 + (absent) + 1.5 = [12.3]. Pedipalpus (Figs. 31A—C) with femur 0.62, tibia 0.46 x carapace width; femur 3.26, tibia 2.46 x length tarsus; tibia widest basally, height 0.48 x length; tarsus with approximately 10 slender apical spinules; bulb width 1.06 x tarsus length; embolus length 0.34 x bulb width. Abdomen 7.5 long, 5.3 wide, sparsely covered OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 with short setae. Variation (N=4)—Total length 8.15—17.1. The size difference between the largest and smallest speci- mens is dramatic but specimens of intermediate size exhibit intermediate development of most distinctive characters. Prosoma yellow-brown to dark red-brown, legs orange-brown to dark red-brown. Caput 0.52-0.6 x carapace width, height 1.28—1.6 x that at fovea; tho- racic fovea width 0.28-0.29 x caput width, straight to slightly recurved; fang furrow with or without denti- cles (Figs. 30C, 31G). Spines of leg I small and incon- spicuous in smallest individual (Fig. 30B) to stout in largest (Figs. 31D0150E). STC II with 2-3 teeth (Fig. 30G). Femur I 0.48—1.15 x carapace width; pedipalpal femur 0.55—0.62 x, tibia 0.35—0.47 x carapace width; embolus length 0.25—0.37 x bulb (Figs. 30D—F, 31A-C, 32A-C). Natural History—The Ambohitantely specimen was collected in pitfall traps at 1450m in disturbed transitional montane mossy forest; other specimens are recorded from forest. Additional Material Examined—MADAGASCAR: Toamasina: Manakambahiny near Vavatene, forét, February 1995, A. Pauly (2 males, MRAC). Antana- narivo: 1450m in R.S. d’Ambohitantely, 24 km NE Ankazobe (18°10.1'S, 47°16.6'E), in pitfall traps, 7-12 December 1997, S. Goodman (1 female, FMNH). Distribution—East central Madagascar in An- tananarivo and Toamasina provinces (Fig. 68). Paramigas goodmani, new species (Figs. 33, 34A—E, 35D, 65, 68) Type.—Holotype female collected at 80m eleva- tion in Forét de Vohibasia, 59 km NE Sakaraha (22°27.5'N, 44°50.5'E), Toliara Province, Madagascar, 10-16 January 1996 by S. Goodman, deposited in FMNH. Etymology—tThe specific name honors Steve Goodman, whose extraordinary collecting efforts in Madagascar discovered the type and numerous other new species. Diagnosis.—Distinguished from all other Parami- gas that have a dense vestiture of long, silky hairs be- neath patellae-metatarsi I and II by having dorsal thorns at the apices of metatarsi I and II (Fig. 33). Description—Female (holotype): Total length 20.5. Carapace (Fig. 33) dark red-brown with a dark brown band surrounding its margin; caput light red- brown along lateral margin; ocular area dark red- brown, dusky between AME and black mesad of ALE and PLE; chelicerae dark red-brown; sternum, coxae, and trochanters orange-brown (Fig. 34B); legs and GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS pedipalpi dark red-brown, except orange-brown tarsi, with orange-brown setal bases giving a mottled pat- tern; abdomen dark purple-brown including book lungs and epigastric furrow; spinnerets dark red-brown (Fig. 34B). Carapace 8.9 long, 8.2 wide, height at thoracic fovea 0.31 x carapace width; smooth. Caput inclined (Fig. 34A), height 1.30 x that at thoracic fovea, width 0.73 x carapace width; median ocular seta situated an- teriad of AME and twelve setae positioned anteriad to this; clypeus 0.30 x length OAL, margin straight. Tho- racic fovea recurved and tripartite, width 0.21 x that of carapace, 1.8 x wider than long; with pair of prefoveal setae (Fig. 33). Ocular area width 0.50 x caput, 2.30 x wider than long; AER 3.0 wide, 1.09 x width PER. Ratio of eyes: AME: ALE: PME: PLE: 1.0: 0.66: 0.66: 0.5, diameter AME 0.6; AME separated by 0.33 x their diameter, PME by 2.25 x their diameter. Ocular quadrangle 1.48 x wider than long, posterior width 1.68 x anterior. Sternum 8.0 long, 5.3 wide, widest behind coxa II and narrowed anteriorly, setose along lateral margin and sparsely setose on surface; sigilla indistinct, shal- low, irregularly oval, adjacent to coxa II, 0.13 x width sternum, distance between 0.57 x distance from margin (Fig. 34B). Labium with 41, pedipalpal coxae with 46-48 cuspules; labium 1.3 long, 1.65 wide, pedipalpal coxae 3.0 long, 1.65 wide, apex produced to a blunt point. Chelicerae 2.0 long, promargin of fang furrow with 4 teeth, retromargin with 6 teeth, pro- and retro- margin interspersed with 1—3 denticles (Fig. 34D). Femur I 0.84, tibia 10.53, femur IV 0.85, and tibia IV 0.54 x carapace width. Ventral surfaces of patellae, tibiae, metatarsi, and tarsi of legs I and II densely cov- ered by long filiform setae (Fig. 33). Spination: pedi- palpus: tibia p 1-0-0, tarsus p 1-1-1, r 1-0-1; leg I: tibia p 1-2-3-3, r 2-3-3-6, d 12 thorns, metatarsus p 2-6-2, r 3-3-3-3, d 26 thorns, tarsus p 1-1-1, r 1-1-0; leg II: tibia p 2-3-3, r 1-3-4-2-3, d 15 thorns, metatarsus p 3-3-3-2- la, r 3-3-la, d 28 thorns, tarsus p 2-1-1-1, r 1-1-0; leg III: patella with approximately 69, tibia with approxi- mately 75, metatarsus with approximately 40, and tar- sus with approximately 12 spinules; leg IV: patella with approximately 80 slender spinules. STC teeth (pro-retro): I (3-2), II (3-3), IH, IV (1-1); ITC simple, pedipalp claw with | tooth (Fig. 34E). Leg measure- ments (Femur + Patella + Tibia + Metatarsus + Tarsus = [Total]): I: 6.9+ 4.0 + 4.4 + 3.0 + 2.0 = [20.3]; II: 6.2 +3.84+404+2.9 + 1.9 = [18.8]; Hl: 5.0+3.1+3.2 + 2.6 + 2.1 = [16.0]; IV: 7.0+4.0+45+3.9+2.5= [21.9]; pedipalpus: 4.0 + 2.1 + 2.6 + (absent) + 2.2 = [10.9]. 23 Abdomen 11.6 long, 8.7 wide, sparsely covered with short setae. Spermathecae with narrow head and long stalk, length spermathecae 0.78 x distance be- tween them and 1.85 x head diameter, diameter head 1.81 x diameter stalk, head length 0.90 x length stalk (Figs. 34C, 35D). Material Examined—Only the type. Distribution—Known only from the type locality, an isolated montane forest in south-western Madagas- car (Fig. 68). Paramigas macrops, new species (Figs. 36, 37A—H, 65, 68) Type—Holotype male collected at elev. 440 m in Reserve Naturelle Integrale d’ Andohahela, parcel 1, 12.5 km NW Eminiminy (24°37.6'S, 46°45.9'E), To- liara Province, Madagascar, 19-28 October 1995 by S. Goodman, deposited in FMNH. Etymology—the specific name refers to the large anterior median eyes. Diagnosis——Distinguished from other Paramigas males except P. manakambus by having AME larger than ALE (Fig. 36), the thoracic fovea recurved and metatarsus I cylindrical, and from P. manakambus by having a retroventral row of 4 spines on tibia I (Fig. 37C) and tarsi III and IV cylindrical (Fig. 36). Description—Male (holotype): Total length 6.15. Prosoma (Figs. 36, 37A) light yellow-brown with dark yellow-brown areas extending along lateral margins of caput and ocular area darkening to black at edge of clypeus; chelicerae dark yellow-brown; ocular area dark except between PME; sternum, coxae, and trochanters light yellow-brown (Fig. 37B); dorsal sur- face of legs I-IV dark yellow-brown fading to light yel- low-brown on tips of tarsi and on ventral surfaces; pedipalpi light yellow-brown; abdomen pale purple- brown; spinnerets and book lung covers light yellow- brown. Carapace 2.6 long and wide, height at thoracic fovea 0.19 x carapace width; weakly rugose. Caput in- clined (Fig. 37A), height 1.5 x that at thoracic fovea, width 0.55 x carapace width; ocular area with single setae situated anteriad of AME and a pair of setae situ- ated anteriad of this; clypeus height 0.35 x length OAL, margin recurved. Thoracic fovea tripartite and recurved, width 0.17 x that of carapace, 1.8 30 mm body length), the absence of a basal tooth on the fang (Fig. 58B), and the broad and short ocular area that is more than 3 x wider than long (Fig. 57). Description—Female (holotype): Total length 44.1. Carapace dark yellow-brown with faint dark lon- gitudinal reticulate markings extending along margins of caput and medially from PME to thoracic fovea (Figs. 57, 58A). Chelicerae, pedipalpi, and legs dark 34 yellow-brown, femora with dorsal longitudinal bands, integument darker beneath dorsal and lateral setal bands on patellae-tibiae. Sternum, labium, and coxae dark orange-brown except yellow-brown transverse band at sterno-labial junction (Fig. 58B). Abdomen dark brown including spinnerets. Abdomen split open and stuffed with cotton, connected to prosoma by a pin. Carapace 18.5 long, 17.0 wide, height at thoracic fovea 0.26 x carapace width; smooth. Caput highly arched (Fig. 58A), height 2.1 x that at thoracic fovea, length 0.52 x carapace width; median ocular seta pres- ent with 13 setae positioned anteriorly; clypeus length 0.03 x length OAL, margin straight. Thoracic fovea deep, T-shaped, width 0.25 x that of carapace, 8.6 x wider than long (Fig. 57). Prefoveal setae present. Ocular area width 0.65 x caput, 3.5 x wider than long; AER 8.8 wide, 1.1 x width PER. Ratio of eyes: AME: ALE: PME: PLE: 1.0: 0.3: 0.6: 0.5, diameter AME 0.6; AME separated by their diameter, PME by 6.27 x their diameter. Ocular quadrangle 2.25 x wider than long, posterior width 2.51 x anterior. Sternum 16.1 long, 9.5 wide, widest behind coxae II and narrowed anteriorly; sigilla adjacent to coxa II, 0.06 x width sternum, distance between 1.16 x dis- tance from margin (Fig. 58B). Labium with 28 and pedipalpal coxae with 39-43 cuspules; labium 3.7 long, 4.7 wide, pedipalpal coxae 7.2 long, 4.3 wide, apex produced to a blunt point. Chelicerae 5.0 long, promargin of fang furrow with 4 teeth, retromargin with 3 teeth, pro and retromargin interspersed with 16 denticles (Fig. 58E). Legs I and II with dense ventral distribution of long, slender setae. Femur I 0.64, tibia I 0.44, femur IV 0.61, and tibia IV 0.41 x width carapace. Spination: pedipalpus: tibia p 1-1, r 0-2, tarsus p 0-3, r 2-2; leg I: tibia p 2-3-3-4, r with 29 spines in two longitudinal rows, metatarsus p 2-4-3-3-2-2-4-3, r 3-4-3-2-3-3-4-8, tarsus p 1-1, r 3-0; leg II: tibia p 0-1-1-2-2-2-2-1, r 2- 3-5-5-3-3-2, metatarsus p 3-3-5-8-2, r 4-2-2-2-5-4, tar- sus p 1-1, r 2-0; leg II: patella with approximately 90, tibia with approximately 140, and metatarsus with ap- proximately 90 spinules; metatarsus with approximate- ly 180 and tarsus with approximately 80 slender spin- ules; leg IV: patella with 60 spinules, metatarsus v 0-0- 0-3a. STC (pro, retro) I (2-1), I (2-2), III, IV (1-2), ITC simple, pedipalp claw with a single large tooth. Leg measurements (Femur + Patella + Tibia + Metatar- sus + Tarsus = [Total]): I: 10.9 + 6.0 + 7.6 + 5.6 + 1.6 = (31.7); Il: 9.6 + 6.0 + 6.5 +5.5 + 1.9 = [29.5]; III: 8.7 +5.0 + 5.1 +5.2 + 3.2 = [27.7]; IV: 10.5 + 8.0 + 7.0 + 7.1 + 3.6 = [36.2]; pedipalpus: 7.7 + 3.2 + 5.1 + (ab- sent) + 4.1 = [20.1]. OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 Abdomen 23.0 long, 21.1 wide, sparsely covered with short setae. One spermathecae dissected out, with very narrow base and broad apex, length 1.79 x maxi- mum diameter, maximum diameter 4.75 x minimum (Figs. 56B, 58C). Variation (N=3)—Total length 32.0-44.1; height at fovea 0.21—0.26 x carapace width. Caput 0.76—0.84 x carapace width, 2.0-3.0 x height at thoracic fovea; width ocular area 0.56-0.69 x caput width, OAW 3.54.0 x OAL, diameter ALE 1.14~1.28 x AME, PLE = PME; clypeus length 1.12-1.54 x OAL, margin weakly procurved to straight; thoracic fovea width 1.92-3.53 = length. Fang furrow interspersed with 14-16 denticles (Figs. 58E—-F). Sternal sigilla width 0.03—0.07 x sternum width, distance between 0.6—2.0 x distance from sternal margin; labium with 25-40, pedipalpal coxae with 35—SO cuspules. Tibia I with 12-29, metatarsus I with 30-32 retroventral spines, tibia II with 11-18, metatarsus II with 20-22 proven- tral spines. Legs I and II may have sparse to dense vestiture of fine ventral setae. STC I and II have 1-3 teeth (Fig. 58G). Spermathecal length 1.61—1.79 x maximum diameter, maximum diameter 4.75—5.14 x minimum (Figs. 56A—C, 58C—D). Distribution—Probably southern Madagascar: the type locality is vague but the Paris specimens are recorded from Ft. Dauphin at the southern tip of the is- land (Fig. 68). Material Examined——MapaGascar: “S. Central Madagascar,” 1894, J. Last, BM1894.2.27.2 (holotype female of Thyropoeus mirandus Pocock, BMNH). To- liara: Fort Dauphin (3 females, MNHN AR4135). PHYLOGENETICS Data A data matrix was assembled (Appendix) com- prising 45 characters scored for all 14 Malagasy species, three exemplars of the outgroup families Id- iopidae, Ctenizidae and Actinopodidae, and 11 exem- plars of the 6 non-Malagasy migid genera Calathotar- sus, Heteromigas, Mallecomigas, Migas, Moggridgea and Poecilomigas. This matrix was analyzed under a variety of parameters to obtain trees of minimum length and of maximum fit (sensu Goloboff 1993c). Exemplars for the non-Malagasy migids and out- groups are listed in Table 2. When possible the type species of each genus was chosen, i.e., Calathotarsus coronatus, Heteromigas dovei, Mallecomigas schlin- geri, and Poecilomigas abrahami. Poecilomigas basilleupi was also included because it differed from P. abrahami in several characters of potential phyloge- GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS netic importance (e.g., male tibia I megaspine, dense setal vestiture beneath legs I and II of female). The type species of Moggridgea (M. dyeri) was inappropri- ate because it is known only from females. Instead, Moggridgea species were chosen to reflect the phylo- genetic diversity of the genus as proposed in Griswold (1987a): M. peringueyi from the peringueyi group, M. intermedia from the quercina group, and M. rupicola from the crudeni group. The Australian Moggridgea tingle was also included to test the monophyly of the genus. Exemplar selection for Migas was perhaps the most difficult problem. The species that we examined include the type species and comprise species from the full geographic range of the genus: Migas affinis from New Caledonia, M. distinctus, M. gatenbyi, M. giveni, M. paradoxus, M. taierii and a male of an undescribed species from New Zealand, and M. nitens and M. vari- apalpus from Australia. In our dataset the exemplar Migas comprises data from female M. gatenbyi and male M. taierii; M. vellardi from Chile is included as a separate taxon to test its placement within the Migidae. All Malagasy taxa are coded from the specimens listed in the monograph. Outgroups The Migoidea, comprising Migidae and Actinopo- didae, has been well supported in the studies of Plat- nick and Shadab (1976), Raven (1985) and Goloboff (1993a). In order to polarize characters within the Migoidea additional outgroups are necessary. Raven proposed Ctenizidae and Idiopidae as successive out- groups to the Migoidea and Goloboff depicted the same families as forming a trichotomy with the Migoidea. Accordingly, we selected representatives of the Actinopodidae (Actinopus), Idiopidae (/diops) and Ctenizidae (Bothriocyrtum) as outgroups to the Migi- dae. Exemplars were chosen on the basis of available material of both sexes at CASC; in addition, the exem- plars of the Actinopodidae and Idiopidae are the type genera of their families. Analysis We analyzed our data with Hennig86 (Farris 1988) and NONA 1.8 (Goloboff 1993b) to find minimum length trees and Pee-Wee 2.6 (Goloboff 1997) to obtain trees that maximize implied weights across all charac- ters (“fittest sensu Goloboff 1993c), with all charac- ters equally weighted (wt = 1) and unordered. We dis- cuss the results with reference to our preferred clado- gram (Fig. 65). Hennig86 analysis using the options t; bb*; gave 42 trees of 96 steps with c.i (consistency index) = 0.52 and ri. (retention index) = 0.76. The 35 strict consensus of these trees is like Figure 65 except for lacking clades G through K. The results of these analyses were subjected to several rounds of succes- sive weighting using the routine xs; w; and the trees obtained were compared with the weights reset to 1. Successive weighting gave 36 trees of 97 steps (1 step longer than the minimum). The consensus of these trees is like our preferred tree (Fig. 65) except for no resolution among the Paramigas species (i.e., nodes G—K were absent) and in lacking clade C, i.e. the placement of Heteromigas is ambiguous. We imple- mented Nona shuffling the input order of taxa 50 times (command mult*50). Trees were calculated accepting (amb =) and rejecting (amb -) ambiguous branch sup- port. Accepting ambiguous branch support gave 45 trees, rejecting ambiguous branch support gave only four trees: all were of 96 steps. In each case the con- sensus gave no resolution among the Paramigas species (i.e., nodes G—K were absent). We imple- mented Pee-Wee shuffling the input order of taxa 50 times (command mult*50), rejecting ambiguous sup- port (default) and using the concavity values | to 6. Concavity | gave 16 trees of fit 309.1. The strict con- sensus of these trees groups Heteromigas with clade S (Calathotarsus, Goloboffia vellardi, and Mallecomi- gas) and provides no resolution within Paramigas. The other concavity values each gave the same four trees: concavity 2 fit = 340.2, concavity 3 fit = 359.9, con- cavity 4 fit = 372.4, concavity 5 fit = 380.7 and con- cavity 6 fit = 387.2. In each case the strict consensus was like our preferred tree except for no resolution within Paramigas. The least restrictive assumption (accepting am- biguous branch support) allows numerous trees that suggest no resolution with Paramigas. More restrictive assumptions (i. e., rejecting ambiguously supported branches or preferring fittest trees derived with most concavity values) gave a smaller set of possibilities: 4 trees. Although the consensus of these also suggests no resolution within Paramigas, examination the 4 trees reveals a simple problem: resolution is lost in the con- sensus because P- manakambus and P. macrops may trade positions. Each in turn may be the sister group of P. oracle or belong to clade H. In one tree P. macrops is the sister group of P. oracle and P. manakambus be- longs to clade H, and in three others P- manakambus is the sister group of P. oracle and P. macrops belongs to clade H. The first tree is also most fully resolved. Preferred tree We believe that the most efficient character opti- mization and most realistic discussion of evolution 36 should be done on the shortest, best resolved tree pos- sible. A tree is a hypothesis, and maximum resolution and minimum length make the maximum number of predictions that can be tested. We prefer the tree that is of minimum length and most completely resolved given only unambiguous node support (Fig. 65). This tree requires 96 steps and has a consistency index of 0.52 and a retention index of 0.76. Readers should keep in mind that equally short but less resolved solutions for Paramigas occur, and that their consensus offers no resolution within Paramigas. Character optimization and branch support We used MacClade 3.0 (Maddison and Maddison 1992) and Clados 1.2 (Nixon 1992) to optimize char- acters on the tree. If optimizations were ambiguous, we usually resolved them using the ACCTRAN option (Farris optimization), which favors secondary loss over convergence to explain homoplasy and therefore max- imizes homology. Character optimizations and evolu- tion are discussed below. Branch support indices (Bre- mer 1994) were calculated with Nona for the clado- gram depicted in Figure 65 using the options hold25000 bsupportS. The “Bremer Support” (“Decay Index”) for a given node in the shortest un- constrained tree is the number of additional steps re- quired in the shortest trees for which that node collaps- es. Due to lengthy calculation times the search for branch support was truncated at values of 5, therefore the Bremer support values reported range between 0 and 5 or greater. Bremer support (decay indices) for the nodes in Figure 65 are A (2), B (3), C (1), D (4), E (5), F (1), G—K (0), L (4), M (2), N 5), O (4), P (4), Q (3), R (1), S (2) and T (1). CHARACTER DESCRIPTIONS AND INTERPRETATIONS We describe the characters and their states and also discuss their evolution as implied by our analyses. Character evolution is referred to our preferred clado- gram (Fig. 65). 1. OAW/ Caput width: (0) less than 0.41, (1) greater than 0.45. A traditional character used to support the mono- phyly of the Migoidea has been the broad distribution of eyes on the caput (Platnick & Shadab 1976; Raven 1985; Goloboff 1991). This feature optimizes at node A as a synapomorphy for the Migoidea. OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 2. Ocular area width: (0) normal for Migoidea, (1) extra wide. The ocular areas of Calathotarsus coronatus (Fig. 4), C. simoni (but not C. pihuychen) and Goloboffia vellardi are especially wide, greater than 0.60 x cara- pace width (Fig. 14A). Most migids and their out- groups have the ocular area with less than 0.40 x cara- pace width, and even the wide ocular areas of Actino- pus and Thyropoeus mirandus (Fig. 57) are less than 0.55 x carapace width. The extra wide ocular area op- timizes as a synapomorphy at node T for Calathotarsus plus Goloboffia vellardi with presumed reversal in C. pihuychen. 3. Prefoveal setae: (0) absent, (1) present, small, length <1 interdistances, (2) present, enlarged, length >| interdistances. A pair of setae is usually present on the caput near the anterolateral margins of the thoracic fovea (Figs. 1, 18). These setae may be small in or absent from males. Bothriocyrtum and Idiops have at least three bands of setae extending from the PME to the thoracic fovea. In Idiops an enlarged pair of setae, which we code as pre- foveals, is closer to the ocular area than to the thoracic fovea. In Actinopus females there are setae near the an- terior margin of the thoracic fovea, which the highly- arched caput renders nearly invisible when viewed from above. In some Migidae the pre-foveal setae are absent, in others conspicuously enlarged. Our clado- gram implies a complex evolutionary history for these setae. Small prefoveals (state 1) is the most likely ple- siomorphic state (Fig. 46A). Also coded as state | are the infinitesimal prefoveals of Paramigas alluaudi (Fig. 28A) and some Paramigas males. Enlarged pre- foveals (state 2) have evolved in parallel five times: at node Q as a synapomorphy for Poecilomigas plus Migas (Figs. 13A—B, 18), and in Idiops, Moggridgea tingle, Paramigas pauliani (Fig. 44A), and P. milloti (Fig. 1). Prefoveals have been lost in parallel (state 0) four times: in Moggridgea intermedia (Figs. 15A—B), Thyropoeus malagasus (Figs. 50, 51A), and in Parami- gas oracle (Fig. 41A) and P. rothorum (Fig.48). 4. Caput setation: (0). Extensive, with four to many setae posteriad of ocular area. (1) reduced, with prefoveal setae only or lacking setae altogether poste- riad of ocular area. Idiopids, ctenizids and actinopodids have the caput with two to several rows of setae extending back from the ocular area. Most Migidae also have setae on the caput. Migas (Figs. 13A—B), Heteromigas (Fig. 8), Mallecomigas (Fig. 12A) and Poecilomigas (Fig. 18) may have three strong longitudinal rows of setae; most GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS Moggridgea (Figs. 15A, B), Calathotarsus, and Goloboffia vellardi have two to three weak rows; Mog- gridgea tingle, other Moggridgea and Migas affinis have at least a row of 2-3 setae as well as prefoveal setae. The Malagasy Migidae lack setae on the caput posteriad of the ocular area with the exception of pre- foveal setae (Figs. 2A, 23A, 50). This loss is a synapo- morphy at node D. 5. Female caput height/ caput length: (0) highly arched (height > 0.72 x length), (1) low (height < 0.7 x length). When viewed from a lateral perspective, the fe- male carapace of Actinopodidae, Idiopidae, Ct- enizidae, and some Migidae (i.e., Calathotarsus [Fig. SA], Thyropoeus [Figs. 51A, 58A], Heteromigas [Fig. 9A]) is strongly arched. This character is most dramat- ic in the Actinopodidae. The arch may be absent or re- duced in males. Most migids have the caput only slightly higher than or equal to the height at the tho- racic fovea (Figs. 13A, 15A, 23A, 34A). A ratio com- paring the height of the caput to its length (from tho- racic fovea to clypeus) quantifies this character. The arched caput is plesiomorphic with a low caput derived in parallel three times: as synapomorphies at nodes E and N, and in Goloboffia vellardi. 6. Fovea: (0) procurved, (1) recurved. The thoracic foveae of the outgroups Actinopodi- dae, Idiopidae, Ctenizidae are strongly procurved. All Migidae have the fovea recurved: even Thyropoeus malagasus, which has the median part of the fovea slightly procurved, has the corners recurved (Fig. 50). The recurved fovea (Fig. 2A) is a synapomorphy at node B for the Migidae. 7. Fovea shape: (0) simple, (1) tripartite. Some migids have a longitudinal, posterior groove connecting with the transverse foveal groove (Figs. 2A, 22, 41A). Pocock (1895) was the first to recognize this tripartite fovea in Paramigas subrufus and used it to diagnose the genus Paramigas. Both males and fe- males of Heteromigas and all Malagasy migids except Thyropoeus malagasus have a conspicuous posterior groove, though the groove in Heteromigas is shallow (Figs. 8, 10) whereas that of the Malagasy migids is deep. As noted by Griswold (1987a, figs. 106-116, 149-150) the character may be variable at least within some Moggridgea. Some Calathotarsus individuals may have this weakly developed and it is faintly visi- ble in Mallecomigas (Fig. 12A). We consider the groove reduced in these taxa, and code their foveae as simple. Interestingly, all Malagasy Paramigas share this fovea shape including very small immature indi- 37 viduals (Fig. 2A). The tripartite condition optimizes as plesiomorphic for those taxa with recurved foveae with the simple fovea arising twice: as a synapomorphy at node M and in Thyropoeus malagasus. 8. Cheliceral tooth row number: (0) 2 rows only, (1) with denticles scattered between the rows of large teeth. Some Migidae and all outgroup taxa except /diops show complex dentition patterns on the chelicerae in- cluding large teeth and much smaller denticles, which are scattered on the fang furrow between the tooth rows and beneath the fang when it is closed. /diops has a promarginal row of large teeth and a retrobasal row of small teeth, which we have not coded as denticles. Denticles occur in Calathotarsus (Fig.5D), Heteromi- gas (Fig. 9C), Thyropoeus (Figs. 51D, 58E-F) and Mi- cromesomma (Fig. 23C). Goloboffia vellardi has 2-3 denticles (Fig. 14E), as do some Paramigas (Fig. 39D). Other migids have only 2 rows of teeth without denti- cles (Figs. 13D, 15E, 41E-F). The plesiomorphic state is equivocal at the base of our cladogram, but denticles optimize unambiguously as plesiomorphic for the Migoidea. Loss of denticles occurs in /diops, at nodes H and K within Paramigas, and possibly at node M with their reappearance at node T as a synapomorphy for Calathotarsus plus Goloboffia vellardi. Alterna- tively their presence in Calathotarsus and Goloboffia vellardi may be a plesiomorphic retention: this requires parallel loss at node N and in Mallecomigas. 9. Intercheliceral basal swellings: (0) absent; (1) present. This is a small unsclerotized swelling located basally on the cheliceral promargin proximad of the teeth (Figs. 24E, 43B) in Actinopus, Migas, Poecilomi- gas, Paramigas, Micromesomma and some Mog- gridgea. This feature appears to have arisen three times: in Actinopus and as synapomorphies at nodes E and N (with loss of this feature in Moggridgea peringueyi). 10. Male intercheliceral tumescence: (0) absent; (1) present. Intercheliceral tumescence refers to pale, unscle- rotized areas presumed to be of glandular origin on the interior surfaces of the chelicerae in some male myga- lomorphs. This feature has been recorded by Raven (1985) and scored and used cladistically by Goloboff (1993a). On our cladogram (Fig. 65) intercheliceral tumescence arises as a synapomorphy at node M with loss within Moggridgea at node P. 11. Rastellum: (0) absent, (1) present. A traditional character used to unite the Migidae 38 has been the absence of a rastellum at the apex of the chelicerae (Raven 1985, Goloboff 1991). This charac- ter is corroborated by our analysis as a synapomorphy at node B for the Migidae. 12. Fang shape: (0) round, (1) quadrate and keeled. A traditional character used to support the mono- phyly of the Migidae has been the presence of strong keels on the fang (Goloboff 1991, Raven 1985). The keeled fangs of Migidae are also quadrate in shape with minute serrations basally on the pro-and retrolat- eral margins (Figs. 2C, 43A). This character is corrob- orated by our analysis as a synapomorphy at node B for the Migidae. 13. Fang basal tooth: (0) absent, (1) present. The monophyly of the Miginae (Migas and Poe- cilomigas) has been suggested by the presence of a small basal tooth on the fang (Fig. 59D) (Griswold 1987a, 1987b; Raven 1985). Goloboff and Platnick (1987) were first to recognize potential problems with this character as a similar tooth is also present in Actinopus. Thyropoeus malagasus has a basal tooth (Fig. 51B) although its sister species, 7: mirandus, does not (Fig. 58B). Our analysis suggests that the basal tooth has arisen in parallel 4 times: as a synapomorphy at node Q for Migas plus Poecilomigas, and as autapo- morphies in Actinopus, Goloboffia vellardi, and Thy- ropoeus malagasus. 14. Fang orientation: (0) vertical, (1) diagonal. The diagonal fang orientation found in the Actinopodidae and Migidae (Figs. 2C, 12B) was used by Raven (1985) in support of the Migoidea. The char- acter is further corroborated as support at node A for the Migoidea in our analysis. 15. Pedipalpal coxa cuspule distribution: (0) across coxa to apex, (1) proximally near labium. Females of all Chilean Migidae (including Goloboffia vellardi) share a cuspule distribution that is concentrated at the proximal edge of the pedipalpal coxa near the labium (Figs. 5B, 12, 14B), which is a synapomorphy at node S. Bothriocyrtum also has this state, but this is most parsimoniously interpreted as a parallelism. 16. Pedipalpal coxa cuspule anterior distribution: (0) broadly across coxa to apex, (1) only proximally near labium. Goloboff and Platnick (1987) noted that Heteromi- gas and Mallecomigas have the cuspules on the pedi- palpal coxae extending to the anterior face of the seg- ment (rather than confined to the ventral surface) and OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 suggested that this feature is primitive and might indi- cate that either of these genera, or both together, might represent the sister group of all remaining migids. In fact cuspules extend onto the anterior surface of the pedipalpal coxae in all taxa studied. In Actinopus and Heteromigas cuspules extend broadly onto the anterior surface, whereas in all others the cuspules extend only onto the anteromedian corner. We have optimized broad anterior extension as a synapomorphy for Migoidea at node A with reversal to anterior extension only at the corner as a synapomorphy at node C. Alter- natively broad extension may have arisen in parallel in Actinopus and Heteromigas. 17. Thorns on coxae IE-III: (0) absent, (1) present. Thorns are stout, shortened setae that have a length 5 x or less than their basal diameter (Figs. 23B, 25D). They differ from cuspules in gradually tapering to a point rather than first expanding from the base (Fig. 25D). Pocock (1895) used the distribution of thorns on the coxae as support for his new genus Mi- cromesomma. Thorns occur in both females and males of some Moggridgea, though reduced in size and num- ber in the latter (the male of Micromesomma is un- known). Griswold (1987a) also used the distribution of thorns on the coxae in his analysis of Moggridgea and suggested that these features indicate a sister-group re- lationship between Micromesomma and Moggridgea. Our cladogram refutes this, suggesting that thorns arose in parallel in Micromesomma and Moggridgea. 18. Anterior sternal sigilla: (0) present, (1) absent. Migidae lack the anterior sternal sigilla, which are present in their outgroups. Goloboff (1993) used this character (his character 69: “posterior” is a typo [Goloboff pers. commun.]). Loss of anterior sternal sigilla is a synapomorphy at node B for the Migidae. 19. Deeply excavate, lunate shaped sternal sigilla: (0) absent, (1) present. Pocock (1895) recognized the deeply excavated, lunate shaped sternal sigilla as diagnostic for Thy- ropoeus (Figs. 51B, 58B). Raven (1985) recognized the significance of similar sigilla in Heteromigella malagasa Strand and synonymized Heteromigella with Thyropoeus. Our analysis corroborates this morpholo- gy as a synapomorphy at node L for Thyropoeus. 20. Female tarsi I and II spines: (0) present; (1) ab- sent. Most migids and their outgroups have spines on female tarsi I and II (Figs. 1, 14C). Migas (Fig. 13C), Poecilomigas (Fig. 18), and Moggridgea (Figs. 15C, 16B) lack these spines, a synapomorphy at node N. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 21. Female leg I and II shape: (0) rounded; (1) dorsoventrally flattened. When viewed from a lateral perspective, the tibi- ae, metatarsi, and tarsi of legs I and II appear dorsoven- trally flattened in all Migidae although they may be less flattened in males. We quantified this character by measuring the cross section of the tibia I apex at the tibia-metatarsus joint. /diops and Actinopus have near- ly round tibiae I and II, with the ratio of breadth to height less than 1.1, whereas Bothriocyrtum and all Migidae have the ratio of breadth to height greater than 1.2. Goloboff and Platnick (1987) suggested that Mal- lecomigas had these segments more rounded than other migids, but we find the cross section of Mallecomigas tibia I to be little different from that of Migas. This character cannot be polarized with our data set, but comparison to other mygalomorphs suggests that flat- tened legs (shared by Bothriocyrtum and the Migoidea) are derived. We optimize flattened legs as a synapo- morphy at node B with parallelism in Bothriocyrtum. 22. Patellar ventral lamellate setae: (0) absent, (1) present. Recognized by Cambridge (1875) and Hewitt (1913c), these erect, distally expanded setae (Fig. 61D) beneath at least patellae I, II and IV were postulated as a synapomorphy for Moggridgea by Griswold (1987a). Main (1991) discovered these setae in two new species from Australia, which she placed in Moggridgea. The setae occur in both males and females of the African species but only in females of the Australian M. tingle. We have found these setae beneath patellae I, II and IV in a female of the Chilean migid Goloboffia vellardi Zapfe (Fig. 14C). Although erect setae may occur in a similar position beneath the patellae of Migas gatenbyi these do not have the characteristic lamellate shape. Our analysis suggests that lamellate setae are a synapo- morphy at node O for Moggridgea (Figs. 15C, 16B) with parallel evolution in Goloboffia vellardi. 23. Leg I and II long, fine, curved ventral setae: (0) absent to weak, (1) dense. The densely distributed, long, fine, curved ventral setae that extend beyond the spine tips of legs I and II of females (Figs. 27A, 41B) were first recognized by Pocock (1895) in Paramigas subrufus and mentioned as a diagnostic character for Paramigas by Raven (1985). Griswold (1987b) recorded such setae in Poe- cilomigas abrahami. These setae are very densely dis- tributed. In Migas gatenbyi more than 15 setae can be counted on a transverse line between the spine rows on tibia I (Fig. 13C); in Poecilomigas abrahami, and Paramigas oracle (Figs. 27A, 41B), P. goodmani (Fig. 39 33) and P. perroti (Fig. 47C) there are more than 20. Some specimens of Thyropoeus mirandus have long, curved setae beneath legs I and II: although these are in longitudinal rows rather than being evenly distributed we code this state as present. Coding 7: mirandus either way makes no difference to the resulting trees. These setae do not occur in males. Our analysis suggests that the dense setal vestiture has arisen in parallel within Paramigas at node J, in Thyropoeus mirandus, in Poe- cilomigas abrahami and in Migas. It could be a synapomorphy at node Q for Migas plus Poecilomigas, but, given the variability in Poecilomigas and Migas, we prefer to minimize its importance until Migas is re- vised. 24. Dorsal setae at apex of tibia I and II: (0) erect, (1) procumbant and stout. A conspicuous retrodorsal patch of thickened procumbant setae is present on the apices of tibiae | and II near the tibia-metatarsus joint of Paramigas (Figs. 27A—C). These setae may be dense (e.g., Paramigas perroti, P. pectinatus) or sparse (e.g., P. pauliani, P. alluaudi). Micromesomma has thorns in the same position (Figs. 22, 24A—B). In most other specimens examined the setae in this region are evenly distributed and erect (Figs. 13C, 14C). Thyropoeus mi- randus has extensive longitudinal dorsal bands of procumbant setae on legs I and II (Fig. 57): whereas there are retrodorsal procumbant setae at the tibia- metatarsus joints such setae also occur dorsally and ex- tend to mid-segment. We have not coded the condition in 7: mirandus as homologous to that in Micromesom- ma and Paramigas. These setae do not occur in males. Our analysis suggests that these setae are a synapo- morphy at node E for Micromesomma plus Paramigas. 25. Dorsal setae at base of metatarsus I and II: (0) erect, (1) procumbant and stout. A conspicuous retrobasal patch of thickened procumbant setae is present on the bases of metatarsi I and II near the tibia-metatarsus joint of Paramigas (Figs. 27A—B). Like the previous character these setae may be dense or sparse. Micromesomma lack procum- bant setae in this position (Figs. 24A, B). These setae do not occur in males. Our analysis suggests that these setae are synapomorphic at node F for Paramigas. 26. Femur III ventral membrane: (0) short, (1) elongate. The length of the ventral membrane of the femur- patella joint on femur III may be restricted to 1/4 (Fig. 59B) to 1/3 (Figs. 15D, 61C) the length of the femur (most migids and their outgroups), or extend at least 4/5 the length of the femur (Malagasy migids including 40 Micromesomma, Paramigas and Thyropoeus) (Figs. 42C, 44D). This character appears to be independent of the tibia III dorsobasal excavation (see below) found in Malagasy migids: in other taxa with a tibia III dor- sobasal excavation (e.g., Moggridgea; Ctenizidae Um- midia) the membrane is short (Figs. 15D, 61C). This feature does not occur in males. Our analysis suggests that this morphology is a synapomorphy at node D for the Malagasy Migidae. 27. Tibia III dorsobasal excavation: (0) absent, (1) present. Most Migidae and their outgroups have tibiae III that are cylindrical (Figs. 4, 8, 14D, 59A). The pres- ence of a dorsobasal excavation on tibia III was used by Raven (1985) to support inclusion of Moygridgea in the Paramiginae. The dorsobasal excavation is a de- pressed, glabrous, usually shiny dark area at the base of tibia III (Figs. 1, 25A—C). In most Moggridgea (in- cluding M. tingle from Australia) it is small and shal- low (Fig. 15D, 61A—B), although in some species it may be absent (e.g., M. crudeni, Griswold 1987a fig. 18). In Moggridgea breyeri the dorsobasal excavation is deep and equals nearly one quarter the length of the segment (Griswold 1987a, fig. 86). In all Malagasy Migidae the dorsobasal excavation is deep and equals nearly one third of the segment length. All dorsobasal excavations, whether shallow or deep, are coded as ho- mologues. A vestige of the dorsobasal excavation may or may not occur in males: it is visible, at least as a darkened area, in the males of Malagasy Paramigas (Figs. 29, 36). Our analysis suggests that the dor- sobasal excavation has arisen in parallel in Mog- gridgea (node O) and the Malagasy Migidae (node D). When this character was weighted until forcing the monophyly of Moggridgea plus the Malagasy Migidae the resulting trees were 4 steps longer than minimum length trees. 28. Tibia III anterior surface: (0) convex, (1) with diagonal ridge. The anterior surface of tibia III of most migids and their outgroups is convex (Figs. 14D, 59A, 61A). The Malagasy migids (except Thyropoeus malagasus) have a weak to strong diagonal ridge extending from the midpoint of the dorsobasal excavation to the middle or nearly to the apex of the segment (Figs. 25A, 42A, 47D). No vestige of this ridge is found in males. We optimize this as a synapomorphy at node D for the Malagasy Migidae with loss in Thyropoeus malagasus. Alternatives are parallel evolution in 7Jhyropoeus mi- randus and clade E. OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 29. Spines on female patella and tibia III: (0) stout, (1) slender. The spines on the anterior surfaces of patella and tibia III of most taxa studied are short and stout, with a length that is less than 5 times base width (Figs. 1, 25). Heteromigas (Fig. 8), Calathotarsus (Fig. 4), Malle- comigas (Fig. 12D), and Goloboffia vellardi (Fig. 14D) resemble one another in having slender spinules with length greater than 10 times base width. This state op- timizes as a synapomorphy at node S with parallel evo- lution in Heteromigas. 30. Dark dorsal and lateral maculations forming bands on leg tibia: (0) absent, (1) present. Many migids have dorsal and lateral maculations on the leg tibiae and metatarsi (e.g., Fig. 28B). In Poe- cilomigas these marks are overlain by diffuse dark bands that completely surround the segment forming conspicuous basomedian leg bands most prominent on the tibiae (Fig. 18). This banding was commemorated by Simon’s naming the genus Poecilomigas (Simon 1903) and was proposed as a synapomorphy for this genus by Griswold (1987b). Our analysis corroborates this character as a synapomorphy for Poecilomigas at node R. 31. Preening comb at apex of metatarsus IV: (0) absent, (1) mixed and/or separate at base (calathotar- sine type), (2) identical and on common base (Mog- gridgea type). We define preening combs as rows or clusters of apical setae that differ conspicuously from their sur- roundings and that may be regularly arranged (see also Raven 1984:381; Griswold 1987a; Goloboff & Plat- nick 1987). In most migids and their outgroups, setae at the metatarsal apex do not form combs. Calathotar- sus, Mallecomigas, Goloboffia vellardi, most Migas (but not M. affinis) and most Moggridgea have unam- biguous preening combs at the apex of metatarsus IV. The nature of these setae may differ among taxa. In most Moggridgea the comb setae are similar in length and thickness and arise juxtaposed from a common base (Figs. 62C—D). In Calathotarsus (Figs. 62A—B) and many Migas long setae are widely spaced and may alternate with short setae. In Goloboffia vellardi and Mallecomigas (Goloboff & Platnick 1987:3, 9) and other Migas (Fig. 59C) the setae are uniformly long and separated at the base by distances greater than their diameter. Vestiges of these combs occur in males, e.g., Calathotarsus males have well developed combs on metatarsi III and IV, and in Moggridgea males there is a comb on metatarsus IV. Most migids may be scored unambiguously as having or not having combs. Her- GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS eromigas dovei has 3-5 slender setae regularly arranged at the retroapex on metatarsus IV whereas other Heteromigas species have nothing resembling a comb (Raven 1984, pers. commun.). We score the comb as absent in Heteromigas. A comb of widely- spaced setae optimizes as a synapomorphy at node M with loss of the comb occurring at node R for Poe- cilomigas. The Moggridgea type of comb optimizes as a synapomorphy at node P for the African Moggridgea. 32. Preening comb at apex of metatarsus III: (0) absent, (1) present. Calathotarsus (Fig. 4), Mallecomigas, and Goloboffia vellardi (Fig. 14D) have combs at the apex of metatarsus III that may be similar to those at the apex of metatarsus IV. The metatarsus HI comb of Goloboffia vellardi alternates long and short setae, dif- fering from the comb of uniform setae on metatarsus IV. This comb is a synapomorphy at node S uniting Ca- lathotarsus, Mallecomigas and Goloboffia vellardi. 33. Preening comb extent: (0) narrow, (1) extend- ing around more than one fifth of segment circumfer- ence. Whereas the preening combs of Moggridgea and Migas consist of a few closely-spaced setae (Figs. 59C, 62C—D), the combs of Calathotarsus, Mallecomigas, and Goloboffia vellardi consist of many setae that ex- tend part way around the ventral apex of the segment (Figs. 4, 62A—B). These broad combs are most con- spicuous when the segment is viewed end-on. Opti- mization of this character is ambiguous. We prefer the broad comb as a synapomorphy at node S uniting Ca- lathotarsus, Mallecomigas and Goloboffia vellardi. Al- ternatively the narrow comb could be a synapomorphy at node N uniting Moggridgea, Migas, and Poecilomi- gas (where it is lost). 34. STC III and IV: (0) claw larger than tooth, (1) tooth larger than claw. The superior tarsal claws of legs II] and IV may have one (e.g. Paramigas pauliani) to three teeth (e.g., Paramigas macrops) beneath the apex (Figs. 37H, 441). In most taxa the teeth are shorter than the apex of the claw but in both Thyropoeus malagasus (Fig. 51E) and 7. mirandus (Fig. 58G) the tooth is as long as or longer than the apex of the claw. This is particularly pronounced on STC IIL. In /diops the tooth of STC III is also large, though not as large as in Thyropoeus. This enlarged tooth is a synapomorphy at node L uniting the two species of Thyropoeus. 35. ITC III size: (0) large, greater than 1/3 length of STC III, (1) small, less than 1/5 length of STC II. 41 Goloboff and Platnick (1987) noted that Malle- comigas resembles the Actinopodidae in having a nor- mal (rather than reduced) ITC III, implying that the large inferior tarsal claw is a shared primitive feature. We quantified ITC III size by comparing its length to that of STC III. /diops, Actinopus, Heteromigas dovei, Mallecomigas schlingeri and Thyropoeus mirandus have ITC III that are at least 1/3 the length of the STC and all other taxa studied have reduced ITC. Large ITC optimize as primitive with reduced ITC appearing in Bothriocyrtum and at node C within the Migidae. The large ITC of Mallecomigas and Thyropoeus mirandus are reversals. 36. Male femur I venter: (0) convex, (1) carinate. Male femora are typically cylindrical or laterally flattened (Figs. 7D, 11D, 30B). The ventral surface of femur I (and femur II of many species) of African Moggridgea is so strongly laterally flattened that it forms a ventral carina or ridge (Fig. 60C; Griswold 1987a, fig. 13). The carinate femur I optimizes as a synapomorphy at node P for the African Moggridgea. 37. Male pedipalpal tibia apex form: (0) Ectal lobe blunt, not or only slightly longer than mesal, (1) ectal lobe pointed, much longer than mesal. The apex of the male pedipalpal tarsus has two lobes that are of near equal length in most migids and their outgroups (Figs. 7A—C, 11A—C, 32C, 54B). The ectal lobe is much broader than the mesal, and in Mog- gridgea intermedia, M. peringueyi and M. rupicola (as well as most other African Moggridgea) this lobe is pointed and extends far beyond the apex of the mesal lobe (Figs. 16A, 32D). The morphology of unequal apical lobes of the pedipalpal tarsus optimizes as a synapomorphy at node P for the African Moggridgea. 38. Male pedipalp tarsus apical spinules: (0) ab- sent, (1) present. One or more elongate spines, conspicuously stouter than the surrounding setae, may be present at the apex of the male pedipalpal tarsus in some Migidae (Poecilomigas [Figs. 20A—B], Paramigas [Figs. 32C, 37E-G], and Goloboffia vellardi) and Idiopidae. Our analysis suggests that apical spinules evolved four times: as synapomorphies at node F for Paramigas (the male of Micromesomma is unknown) and node Q for Migas plus Poecilomigas, and in /diops and Goloboffia vellardi. 39. Male retrolateral tibia | megaspine: (0) absent, (1) present. We define a megaspine as a spine that is conspic- uously larger that others near it (Figs. 7D, 55B). 42 African Moggridgea lack enlarged spines at the retroapex of femur I (Fig. 16B), as do Heteromigas (Fig. 11D) and Goloboffia vellardi. Poecilomigas abrahami lacks such a spine but Poecilomigas basille- upi has an enlarged retroapical spine (Griswold 1998a fig. 7). Similar enlarged spines occur in Moggridgea tingle, Migas variapalpus, and Calathotarsus (Fig. 7D). A single large apical spine is present on the retro- lateral surface of tibia I on all Malagasy migids. This is at the apex in Paramigas macrops (Fig. 37C) and P. manakambus (Fig. 38C) and arises from a subapical spur in Paramigas andasibe (Figs. 30B, 31E). Idiops has two and Bothriocyrtum several enlarged retroapical spines: we have coded these as homologous to the migid megaspine. Coding these as non-homologous makes no difference to the resulting tree. If we consid- er the spines in J/diops and Bothriocyrtum as megaspines this feature optimizes as plesiomorphic with megaspines lost in parallel five times: in Actino- pus, Heteromigas dovei, Goloboffia vellardi, Poe- cilomigas abrahami, and as a synapomorphy at node P for the African Moggridgea. Alternatively loss of the megaspine could be a synapomorphy at node A for the Migoidea with regain at node C within the Migidae. 40. Male metatarsus I shape: (0) cylindrical, (1) curved, with retrolateral pale swelling. The males of Paramigas andasibe have metatarsi I that are curved, with a retrolateral pale swelling (Figs. 29, 30B, 31D—-E). Calathotarsus (Figs. 6, 7D) and Bothriocyrtum have male metatarsi that are strongly bent retrolaterad at mid-segment; we do not consider this as homologous to the state discussed here. This feature is an autapomorphy for Paramigas andasibe. 41. Male tarsus III shape: (0) cylindrical, (1) curved, sausage shaped, strongly convex ventrally. Most rastelloid males have tarsi III and IV cylin- drical (Figs. 6, 10, 36). Males of the Malagasy migids Paramigas andasibe (Fig. 31F) and P. manakambus (Fig. 38D) have tarsi III and IV weakly curved, slight- ly swollen, and strongly convex ventrally. We optimize this as a synapomorphy at node G, predicting its pres- ence in the undiscovered males of P. alluaudi, P. pecti- natus, P. pauliani, and P. rothorum. 42. PLS apical segment: (0) triangular, (1) domed. As noted by Raven (1985), the apical segment of some mygalomorph spiders is triangular in shape, most easily observed in the Ctenizidae. All Migidae and Actinopodidae (the Migoidea, clade A) share a domed apical segment of the PLS (Figs. 3A—B). 43. PLS spigot distribution: (0) all articles, (1) me- dian and apical articles only. OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 Goloboff and Platnick (1987) first recognized the unique distribution of spigots on the PLS of Chilean Migidae and suggested it as a potential synapomorphy for the family. All migid taxa surveyed have their spig- ots limited to the median and apical segments of the PLS (Figs. 3A, B), corroborating this condition as a migid synapomorphy at node B. 44. Distribution of spermathecae sclerotization: (0) uniform, (1) a median band. Most migids and their outgroups have the sper- mathecae evenly sclerotized (Figs. 35A—E, 64A-C). The degree of sclerotization may vary from light to heavy. Uniquely, Moggridgea (including M. tingle from Australia) have a narrow band of sclerotization across the middle of the spermatheca (Figs. 17A-C). This sclerotized band may even be visible through the cuticle. This is a synapomorphy at node O for Mog- gridgea. 45. Spermatheca shape: (0) short, cylindrical, HS equal to or slightly wider than stalk, (1) head narrower than long stalk, (2) long stalk with wider head, (3) head much wider than short, narrow stalk, (4) clavate. The spermathecae of Actinopus (Fig. 64B) and Bothriocyrtum (Fig. 64C) are short, stout and have the head only slightly wider than the stalk. Similar sper- mathecae occur in Goloboffia vellardi (Figs. 63E—F) and Calathotarsus (Fig. 63A). Heteromigas dovei (Fig. 63B) and Mallecomigas schlingeri (Figs. 63C, D) have spermathecae that are long and have the head narrow- er than the stalk. A slender stalk and broad head is characteristic of most migids, 1. e., Migas (Fig. 19C), Moggridgea (Figs. 17A—C), Poecilomigas (Figs. 19A—B) and several Paramigas (Figs 35A—E, 40A—B). Idiops (Fig. 64A) and the Malagasy migids Paramigas rothorum (Fig. 45B), P. alluaudi (Fig. 45C), P. pecti- natus (Fig. 45A), P. pauliani (Figs. 40C—D), and Mi- cromesomma (Figs. 26A—C) have the head much wider than short, narrow stalk. The stalk is especially narrow in Paramigas rothorum, P. alluaudi, P. pectinatus, and P. pauliani. Thyropoeus mirandus (Figs. 56A—C) and T. malagasus (Fig. 56D) have clavate spermathecae that taper gradually from the base to the head: that of 7: mirandus arises from a very narrow base. State 0 (short, cylindrical) optimizes as plesiomorphic. State | (head narrower than long stalk) arises in parallel in Heteromigas and Mallecomigas. State 3 (head much wider than short, narrow stalk) arises in parallel in /d- iops and as a synapomorphy at node E uniting Parami- gas and Micromesomma, and state 2 (long stalk with wider head) is a synapomorphy for clade N (Mog- gridgea plus the Miginae) and at node I within GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS Paramigas. State 2 is predicted to occur in the undis- covered female of Paramigas macrops. Characters not included in data matrix Goloboff and Platnick (1987:7) state that unlike most migid genera, Heteromigas and Mallecomigas have short spines on tibiae and metatarsi I and II and imply that retention of this primitive feature might in- dicate that either of these genera, or both together, might represent the sister group of all remaining migids. In fact Heteromigas and Mallecomigas spines are not particularly short: the longest spines on tibiae and metatarsi I and II of these genera have a length/base width ratio equal to that of many species of Migas, Moggridgea and Paramigas. PHYLOGENY OF MIGIDAE Our preferred tree (Fig. 65) contains the maximum resolution supported by the data. We discuss clades on this tree with their synapomorphies listed in parenthe- sis (Character: State). Clade A (Migoidea) — Synapomorphies are the wide ocular area, with OAW/ Caput width greater than 0.45 (1:1), diagonal fang orientation (14:1), and domed PLS apical segment (42:1). A potential synapomorphy, ambiguously optimized, is the cuspules on the pedipal- pal coxae that extend broadly onto the anterior surface (16:1). Clade B (Migidae) — Synapomorphies are the re- curved thoracic fovea (6:1), loss of the rastellum (11:1), quadrate and keeled fang (12:1), loss of anteri- or sternal sigilla (18:1), flattened female anterior legs (21:1, ambiguous) and PLS spigots limited to median and apical articles only (43:1). Clade C (all migids other than Heteromigas) — The reduced ITC III is a synapomorphy (35:1). A po- tential synapomorphy, ambiguously optimized, is the cuspules on the pedipalpal coxae that are restricted to the anterior median corner (16:0). Clade D (Malagasy migids, Paramiginae sensu Simon) — Synapomorphies are the reduced caput seta- tion, with prefoveal setae only or lacking setae alto- gether posteriad of ocular area (4:1), femur III ventral membrane that extends at least 4/5 the length of the femur (26:1), and tibia III with a deep dorsobasal ex- cavation (27:1). The diagonal ridge on tibia II (28:1) is an ambiguous synapomorphy. Clade E (Micromesomma plus Paramigas) — Synapomorphies are the low female caput, with height <0.7 = length (5:1), cheliceral basal swellings (9:1), procumbant and stout dorsal setae at the apices of tibi- 43 ae I and II (24:1), and spermathecae with the head much wider than the short, narrow stalk (45:3). Clade F (Paramigas) — Synapomorphies are the procumbant and stout dorsal setae at bases of metatar- si I and II (25:1) and male pedipalp tarsus apical spin- ules (38:1). The latter feature could instead be a synapomorphy uniting Micromesomma and Parami- gas, as the male of Micromesomma is unknown. Clade G (six species of Paramigas) — The synapomorphy is male tarsi III and IV that are weakly curved, slightly swollen, and strongly convex ventrally (41:1). This has been observed only in Paramigas an- dasibe (Fig. 31F) and P. manakambus (Fig. 38D) and is predicted to occur in the species of clade G that are known only from females. Clade H (five species of Paramigas) — Loss of denticles from between the cheliceral teeth is the synapomorphy (8:0). Clade I (five species of Paramigas) — Spermath- ecae with a long stalk with wider head (45:2) is a synapomorphy uniting these species. Clade J (four species of Paramigas) — The dense- ly distributed, long, fine, curved ventral setae that ex- tend beyond the spine tips of legs I and II of females are the synapomorphy (23:1). Clade K (Paramigas macrops plus P. oracle) — Loss of denticles from between the cheliceral teeth is the synapomorphy (8:0). Clade L (Thyropoeus) — Synapomorphies are the deeply excavate, lunate shaped sternal sigilla (19:1), ITC III and IV with the tooth longer than the claw (34:1) and clavate spermathecae that taper gradually from the base to the head (45:4). Clade M (Calathotarsus, Goloboffia, Mallecomi- gas, Migas, Moggridgea and Poecilomigas) —Synapo- morphies are the simple, recurved fovea lacking poste- rior extension (7:0), loss of denticles from between the cheliceral tooth rows (8:0, ambiguous), presence of male intercheliceral tumescence (10:1) and metatarsus IV preening comb (31:1). Clade N (Miginae sensu Simon) — Migas, Mog- gridgea and Poecilomigas are united by the synapo- morphies low female caput, with height < 0.7 x length (5:1), basal cheliceral swellings (9:1), loss of spines from female tarsi I and II (20:1) and by having the spermathecae with a long stalk with wider head (45:2). Clade O (Moggridgea) — Synapomorphies are the patellar ventral lamellate setae (22:1), tibia II dor- sobasal excavation (27:1), and narrow band of scleroti- zation across the middle of the spermatheca (44:1). Clade P (African Moggridgea) — Synapomor- phies are loss of male intercheliceral tumescence 44 (10:0), metatarsus IV preening comb setae that are sim- ilar in length and thickness and arise juxtaposed from a common base (31:2), carinate male femur I venter (36:1), male pedipalpal tibia apex with the ectal lobe pointed, much longer than mesal (37:1) and loss of the male retrolateral tibia | megaspine (39:0). Clade Q (Migas plus Poecilomigas) — Synapo- morphies are the enlarged prefoveal setae (3:2, am- biguous), basal tooth on the fang (13:1), and male pedi- palp tarsus apical spinules (38:1). Clade R (Poecilomigas) — Synapomorphies are the dark dorsal and lateral maculations forming annuli on the leg tibiae (30:1) and loss of the metatarsus IV preening comb (31:0). Clade S (South American Migidae) — This newly proposed clade comprising Calathotarsus, Mallecomi- gas and Goloboffia vellardi is united by the synapo- morphies of pedipalpal coxa cuspule distribution that is concentrated at the proximal edge near the labium (15:1), slender spinules densely situated on patella- tibia III of females (29:1), a preening comb at the apex of metatarsus III (32:1) and preening combs that con- sist of many setae that extend part way around the ven- tral apex of the segment (33:1, ambiguous). Clade T (Calathotarsus plus Goloboffia vellardi) — Synapomorphies are the extra wide ocular area, greater than 0.60 = carapace width (2:1) and den- ticles scattered between the cheliceral tooth rows (8:1, ambiguous). CLASSIFICATION This phylogenetic analysis corroborates the mono- phyly of Migoidea (Actinopodidae plus Migidae) (Platnick & Shadab 1976; Raven 1985; Goloboff 1993a) and Migidae. We reject the Calathotarsinae (sensu Simon 1903) and note that it is paraphyletic with respect to the Miginae and Paramiginae. On the other hand, our results corroborate both the Miginae (sensu Simon 1903) and Paramiginae (Petrunkevitch 1939, sensu Simon 1892, but not sensu Raven 1985). Simon (1892, 1903) believed that the migids fell into three groups. His Migeae (Miginae), comprising Migas, Moggridgea and Poecilomigas, included taxa that have simple thoracic foveae and third tibiae and low caputs. His Calathotarsae (Calathotarsinae), com- prising Calathotarsus and Heteromigas, included taxa that have caputs raised as in Actinopus. His Myrtaleae (Paramiginae Petrunkevitch 1939), comprising Mi- cromesomma, Paramigas, and Thyropoeus, included taxa that have tripartite foveae and dorsobasal excava- tions on tibiae III]. Raven (1985) reexamined migid classification and modified the limits of Miginae and OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 Paramiginae by transferring Moggridgea from the Miginae to Paramiginae. He noted that many Mog- gridgea species have a dorsal excavation on tibia III. He retained the Calathotarsinae although noting (Raven 1985:57) that they appear to lack a synapomor- phy. Griswold accepted Raven’s relimited Paramiginae (Griswold 1987a) and Miginae (Griswold 1987b). Goloboff and Platnick (1987) described Mallecomigas and the female of Migas vellardi. Migas vellardi was tentatively placed in the Miginae based on the basal tooth on the fang and low caput. Goloboff and Platnick (1987) noted that Mallecomigas retained presumably plesiomorphic features, e.g., rounded tibiae and metatarsi I and II (shared with Heteromigas) and a nor- mal (rather than reduced) ITC III, and suggested that Mallecomigas, Heteromigas, or both together might represent the sister group of the remaining migids. Our conclusions differ from some of those of Raven (1985) and of Goloboff and Platnick (1987). Although the dor- sally excavate tibia III is included in the dataset (char- acter 27) for both Moggridgea and the Paramiginae it is most parsimonious to place Moggridgea in the Migi- nae with Migas and Poecilomigas. When character 27 was weighted to force the monophyly of the Paramigi- nae sensu Raven 1985 (including Moggridgea), result- ing trees were 100 steps, 4 steps longer than the mini- mum for unweighted data. Our results suggest that the South American Migas vellardi, Mallecomigas and Calathotarsus form a clade. Migas vellardi does not belong in Migas, and we propose the new generic name Goloboffia for it. Goloboffia appears to be the sister genus of Calathotarsus, but to place Migas vellardi in Calathotarsus would render this genus heterogeneous in that Migas vellardi has ventral lamellate setae on the patellae but lacks the characteristic Calathotarsus synapomorphies of modified female pedipalpal tarsus bearing a ventral expansion and dorsal group of cus- pules and bent and apically swollen male metatarsus I. Mallecomigas is not a relatively primitive migid. We include the shape of tibiae and metatarsi I and II in the dataset (character 21) but we find the cross section of Mallecomigas tibia I to be flattened and little different from that of Migas. The large ITC of Mallecomigas (character 35) appears to be an independent reversal to the plesiomorphic form. Our results do agree with those of Goloboff and Platnick (1987) in suggesting that Heteromigas is a primitive migid: it is probably the sister group to the rest of the family. BIOGEOGRAPHY Due to their strikingly disjunct southern distribu- tion (Fig. 66) the migids have long drawn the attention GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS of biogeographers. Pocock (1903: 352) first offered an explanation for their distribution suggesting that they had arisen in the Afro-Mascarene area (i.e., Africa plus Madagascar) and subsequently migrated to Australia- New Zealand and later to South America. Legendre (1979) next commented on their distribution, writing after the widespread acceptance of continental drift as an influence on biogeography. Legendre accepted migid distribution as strong evidence for Gondwana- land, considering them “couvrent incontestablement l’ancienne Gondwanie et ont été souvent invoquées comme preuve de |’existence du continent de Gond- wana” (Legendre 1979:45—46). Platnick was the first to examine the migids in the context of cladistics and vicariance biogeography (Platnick 1981; Nelson & Platnick 1984). Lacking a cladistic treatment of the family, Platnick accepted the then current classification as the best available phylogenetic hypothesis (Platnick 1981:90, fig. 7) and used the then conventional hy- pothesis of Gondwanaland breakup (i.e., [[Africa, Madagascar] [New Zealand [Australia, South Ameri- ca]]]). Recognizing that changes in migid classification and geological interpretation might alter the basic data, he nonetheless found that two components in migid classification were consistent with vicariance theory and explained by Gondwanaland breakup (Platnick 1981:89, fig. 6, components 4 [Australia plus South America] and 5 [New Zealand plus Australia plus South America]). Since Platnick’s work our under- standing of Gondwanaland breakup has changed some- what (Smith et al. 1994) and with our new cladogram several components of migid classification have changed. Main (1991) described new Moggridgea from southern and western Australia, suggested that the genus may have originated in the Jurassic (140 ma), and compared the distribution of Moggridgea to pseu- doscorpions, scarab beetles, and midge flies (see below). Griswold (1991a) examined migids as part of the Afromontane biota (Griswold 1991a; White 1978, 1983). Using a parsimony method to combine cladistic biogeographic information from the spider families Microstigmatidae (Microstigmata), Migidae (the Mog- gridgea quercinca group) and Phyxelididae (the Vi- doleini and Phyxelidini, then considered Amaurobi- idae), he proposed a cladogram for areas of temperate and montane forest in Africa and Madagascar (Gris- wold 1991a, fig. 5) that contained a sister-area rela- tionship between Madagascar and the mountains of East Africa and predicted that members of the Mog- gridgea quercinca group should be discovered in Madagascar. Dispersal cannot be ruled out a-priori in the case 45 of migid disjunctions although mygalomorphs like migids have long been considered poor in dispersal abilities and therefore excellent subjects for studies of historical biogeography. Most mygalomorphs are not thought to disperse by ballooning (becoming airborne and floating on silken threads) though spiderlings of some mygalomorphs exhibit stereotyped ballooning behavior and float for short distances on silken threads (Coyle 1983, 1985). Rafting on floating vegetation also cannot be ruled out for migids. Many migids are tree dwellers with tightly-sealed trap door nests. As Raven (1980) pointed out, a gravid female in a sealed trap door nest on a floating tree trunk might succeed in dis- persing over water. At least the ancestors of Mog- gridgea nesiota may have arrived on the oceanic is- lands that make up the Comoros in this way. To favor a vicariance explanation for the disjunctions in migid distribution these patterns must correlate with some potential vicariance event in earth history and/or be general, that is, not characteristic of the migids alone. As shown below, migid disjunctions are both correlat- ed with events in earth history and similar to distribu- tion patterns found in other organisms. Our new cladogram bears upon previous hypothe- ses and suggests several new insights. The Migidae can reasonably be considered a Gondwanan family, occur- ring in several parts of that former southern supercon- tinent (Africa, Australia, Madagascar, New Caledonia, New Zealand, and South America) and nowhere else (Fig. 66). An area cladogram (Fig. 67) derived from our taxon cladogram (Fig. 65) suggests several inter- continental relationships. Australia appears three times on the cladogram: related to Africa (Moggridgea), re- lated to New Zealand and New Caledonia together and to eastern and southern Africa (clade Q: Migas plus Poecilomigas), and related to the areas Madagascar, South America, Africa and Australia, New Zealand and New Caledonia together (this relationship speci- fied by Heteromigas). The complex area relationships shown by Australia suggest that this continent might be subdivided for biogeographic studies. For example, Moggridgea suggests a relationship of western and southern Australia to Africa, the Miginae suggest a re- lationship of eastern Australia to Africa, New Zealand and New Caledonia, and Heteromigas suggests a rela- tionship of eastern Australia to all the other parts of Gondwanaland occupied by migids. Africa appears twice, once related to Australia (Moggridgea) and once to Australia, New Zealand and New Caledonia (clade Q: Poecilomigas plus Migas). South America (repre- sented by clade S, comprising Calathotarsus, Malle- comigas and Goloboffia vellardi) is related to Africa, 46 Australia, New Zealand and New Caledonia together (represented by clade N, comprising Moggridgea, Migas and Poecilomigas). Madagascar arises near the base of the cladogram (clade D: Paramiginae) and is the sister area of all other Gondwanan components in- habited by Migidae, i.e. there is a fundamental split be- tween Madagascar and the remainder of Gondwana- land. This is compatible with a recent, widely accepted hypothesis of Gondwanaland breakup. Smith, Smith and Funnell (1994) map a water gap between the east coast of Africa and west coast of Madagascar that may date back to the Sinemurian (200 ma); these areas re- mained in contact via Antarctica to the south. Mada- gascar (plus India) had separated from the remainder of Gondwanaland by the Aptian (120 ma). Other Gond- wanan fragments remained in contact at this time. The fundamental division between the Malagasy Paramigi- nae and migids of the rest of Gondwanaland suggests that the family existed at least 120 million years ago and that the ancestors of Paramiginae may have been stranded on the Madagascar-India plate during the Apt- ian. The fact that Heteromigas (from eastern Australia) is plesiomorphic to even this potentially ancient split suggests that the Migidae may have geographically dif- ferentiated before the breakup of Gondwanaland. The distribution of the remaining migids cannot be simply explained by continental drift and Gondwana- land breakup. The simplified hypothesis of Gondwana- land fragmentation suggested by modern works (e.g., Smith et al. 1994) is ((Madagascar, India) (Africa (New Zealand (South America, Australia)))). This is of course an oversimplification. Gondwanaland was not homogeneous, nor was or is any of its fragments. Nu- merous tropical west Africa-tropical South America disjunctions on the one hand and Australia-New Zealand-Chile disjunctions on the other suggest the fragmentation of tropical and temperate biotas that were already distinct before the breakup of Gondwana- land. Migidae exhibit at least two cases of trans-Indian Ocean disjunction: within Moggridgea (Africa-Aus- tralia) and clade Q (Poecilomigas [Africa] and Migas [Australia, New Caledonia and New Zealand]). Trans- Indian Ocean disjunctions occur in several other arthropod groups. The millipede families Harpago- phoridae and Sphaerotheriidae occur in Africa, Mada- gascar, south Asia and Australia (Kraus 1978). The millipede distributions include taxa in Madagascar and south Asia, a situation not known in the trans-Indian Ocean Migidae. Trans-Indian Ocean disjunctions (for taxa that do not also occur in South America) have been called “East gondwanwan” by Kraus (1978). This conception of “East Gondwana” differs from others OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 (e.g., Truswell 1977; Pielou 1979) in which East Gond- wana included India, Antarctica, Australia, New Zealand and New Caledonia, but excluded Madagas- car, which was considered part of “West Gondwana- land” along with Africa and South America. The spider family Cyatholipidae describes another trans-Indian Ocean disjunction, occurring in Africa, Madagascar, Australia and New Zealand but probably absent from South America (Griswold 2001). Harvey (1996a, 1996b) records several trans-Indian Ocean disjunctions in pseudoscorpions that he attributes to fragmentation of former ranges that encompassed parts of Gond- wanaland. Cranston, Edward and Colless (1987) record an Africa-Australia disjunction in the chironomid midge fly genus Archaeochlus, and Mathews (1976) and Howden (1981) report Africa-Australia disjunc- tions in various groups of scarab beetles. These dis- junct distributions may result from vicariance of a biota that was already differentiated in Gondwanaland. Given the Aptian age (120 ma) for Migidae implied by a basal split in the family it is likely that continental drift has influenced the distribution of the non- Paramiginae, probably acting on formerly restricted distributions and shaped by extinctions. The solution to migid biogeography requires a better understanding of the environmental history of Gondwanaland as well as a more complete picture of migid phylogeny and dis- tribution. Our study refutes the prediction of Griswold (199 1a) that Madagascar should be home to close rela- tives of migids from the East African mountains. No members of Moggridgea have been found in Madagas- car, although the species Moggridgea nesiota lives on the nearby Comoro islands (Griswold 1987a). The East Africa-Madagascar sister area relationship proposed by Griswold (1991a) does appear to be a general pattern for spiders, though. It has been replicated by new data for two groups of cyatholipid spiders (Griswold 2000, 2001) and a pair of zorocratid spider genera (Griswold 1993, 2000), and parallels vicariance patterns suggest- ed for lemurs (Yoder et al. 1996) and tenrecs (Asher 1997). What can be said of the distribution of migids within Madagascar? Unfortunately the sampling is un- even and the cladogram is poorly resolved. Indeed, the consensus of equally short cladograms (but not neces- sarily equally preferable: see discussion above under “Preferred tree”) offers no resolution among Madagas- car migids. On the preferred cladogram (Fig. 65) some distributions of interest do emerge. With the exception of Paramigas goodmani from the southwestern part of the country, all clade J species (Paramigas perroti, P. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS macrops and P. oracle) occur in the central-eastern part of the country. The sister group of clade G, Paramigas milloti, occurs in the north. Clade H comprises one species from the eastern escarpment (Paramigas man- akambus), three species from the south or southwest of the country (Paramigas alluaudi, P. pauliani, and P. pectinatus) and one (Paramigas rothorum) from the far north. It is unfortunate that relationships cannot be fur- ther resolved among the members of this eastern, northern and western clade. A special relationship be- tween northern and southwestern species has been sug- gested by Raxworthy and Nussbaum (1997, fig. 5.5), who performed a parsimony analysis of endemicity (PAE) on the distributions of 200 Malagasy reptile species. Unfortunately, understanding the biogeogra- phy of migids within Madagascar will require far denser sampling than has been done to date. CONCLUSIONS Migid trap door spiders are an ancient group whose complex distribution has been influenced by continental drift but defies simple explanation. They may have been geographically differentiated even be- fore the breakup of Gondwanaland. It appears that the Paramiginae evolved from ancestors stranded by conti- nental drift on Madagascar, probably while this was still connected to India. This suggests that we need to look for Migidae in south Asia, particularly Sri Lanka and southern India. No migids have been found in these areas, but their highly cryptic habits may have led to them being overlooked. Much remains to be learned, especially about the Malagasy fauna. We need to asso- ciate more males and females, observe the biology of Thyropoeus and Micromesomma, and document the distribution of Micromesomma. The monograph and phylogeny presented herein is only a preliminary look at this ancient and fascinating group of Malagasy spi- ders. LITERATURE CITED ALDERWEIRELDT, M., and R. JoCQUE. 1994. Biodiversity in Africa and Europe: The Case of Spiders (Araneae). Bi- ologische Jahrbuch Dodonaea 61:57—-67. ASHER, R. J., 1997. African and Malagasy tenrecs: a biogeo- graphic parallel with lemuriform primates? American Journal of Physical Anthropology, supplement, 24: 69-70. BENOIT, P. L. G. 1962. Migidae nouveaux du Musée Royal de |’ Afrique Central. Revue de Zoologique et de Botanique Africaines 66:276-282. BERLAND, L. 1924. Araignées de la Nouvelle Calédonie et 47 des iles Loyalty. Pages 159-255 in F. Sarazin and J. Roux (eds.), Nova Caledonia, Zoologie, 3 (2). Berlin. BOonnerT, P. 1957. Bibliographia Araneorum. Toulouse, 2 (3): 1927-3026. BONNET, P. 1958. Bibliographia Araneorum. Toulouse, 2 (4):3027-4230. BonneT, P. 1959. Bibliographia Araneorum. Toulouse, 2 (5):4231—S058. BREMER, K. 1994. Branch support and tree stability. Cladis- tics 10:295-304. BRIGNOLI, P. M. 1983. A catalogue of the Araneae described between 1940-1981. Manchester University Press, 755 Pp. CALDERON GONZALEZ, R. 1983. Revision systematique des Araignées mygalomorphs du Chili, suivie d’une etude sur le development post-embryonnaire de Tryssothele pissi Simon, 1899 (Mygalomorpha, Dipluridae). These, Academie de Montpellier, Université des Sciences et Techniques du Languedoc. CAMBRIDGE, O. P. 1875. On a new genus and species of trap- door spider from South Africa. Annals and Magazine of Natural History 4(16):317-322. CAMBRIDGE, O. P. 1889. On some new species and a new genus of Arachnida. Proceedings of the Zoological So- ciety of London 1889:3446. CARPENTER, J. M. 1988. Choosing among multiple equally parsimonious cladograms. Cladistics 4:291—296. CoyLe, F. A. 1983. Aerial dispersal by mygalomorph spider- lings (Araneae, Mygalomorphae). Journal of Arachnol- ogy 11:283-286. CoyLe, F. A. 1985. Ballooning behavior of Ummidia spider- lings (Araneae, Ctenizidae). Journal of Arachnology 13:137. CRANSTON, P. S., D. H. D. Ep>warp and D. H. CoLLess. 1987. Archaeochlus Brundin: a midge out of time (Diptera: Chironomidae). Systematic Entomology 12:313-334. DIPPENAAR-SCHOEMAN, A. S., and R. JocQUE. 1997. African Spiders: an identification manual. Plant Protection Re- search Institute Handbook 9. Pretoria, South Africa. 392 pp. Dresco, E., and A. CANARD. 1975. Sur une espéce nouvelle de Mygale de Madagascar: Legendrella pauliani, type d’un genre nouveau (Araneae, Migidae). Bulletin du Muséum National d’Histoire Naturelle, Paris, (ser. 3), 306 (Zool. 216):783—788. FARRIS, J. S. 1969. A successive approximations approach to character weighting. Systematic Zoology 18:374-385. Farris, J. S. 1988. Hennig86 1.5. Microcomputer program available from Dr. Arnold Kluge, Museum of Zoology, University of Michigan, Ann Arbor, Michigan, 49109- 079 USA. Go.oporr, P. A. 1991. A new species of Calathotarsus (Araneae: Migidae) from Chile. Journal of the New York Entomological Society 99 (2):267-273. GocosorF, P. A. 1993a. A reanalysis of mygalomorph spider families. American Museum Novitates (3056): 1-32. GoLosorF, P. A. 1993b. Nona 1.6. Computer program avail- 48 able from J. M. Carpenter, Dept. Entomology, American Mus. Natural History, Central Park West at 79th, New York, New York 10024 USA. Go.osorF, P. A. 1993c. Estimating character weights during tree search. Cladistics 9:83—92. GoLosorF, P. A. 1994. Migoides de Chile, nuevas 0 poco conocidas (Araneae, Mygalomorphae). Revista de la Sociedad Entomologica de Argentina 53: 65—74. Go.osorF, P. A. 1997. Pee-Wee 2.6. Computer program available from J. M. Carpenter, Dept. Entomology, American Mus. Natural History, Central Park West at 79th, New York, New York 10024 USA. Go .osorfF, P. A., and N. I. PLATNICK. 1987. A review of the Chilean spiders of the superfamily Migoidea (Araneae: Mygalomorphae). American Museum Novitates (2888):1-15. GRISWOLD, C. E. 1985. A revision of the African spiders of the family Microstigmatidae (Araneae: Mygalomor- phae). Annals of the Natal Museum 27:1—37. GRISWOLD, C. E. 1987a. The African members of the trap- door spider family Migidae (Araneae: Mygalomor- phae). 1: The genus Moggridgea O. P.- Cambridge, 1875. Annals of the Natal Museum 28(1):1—118. GrIswoLp, C. E. 1987b. The African members of the trap- door spider family Migidae (Araneae: Mygalomor- phae). 2: the genus Poecilomigas Simon, 1903. Annals of the Natal Museum 28:475—497. GRISWOLD, C. E. 1987c. A review of the southern African spiders of the family Cyatholipidae Simon, 1894 (Araneae: Araneomorphae). Annals of the Natal Muse- um 28:499-542. GRISWOLD, C. E. 1990. A revision and phylogenetic analysis of the spider subfamily Phyxelidinae (Araneae, Amau- robiidae). Bulletin of the American Museum of Natural History 196:1—206. GRISWOLD, C. E. 1991a. Cladistic Biogeography of A fro- montane Spiders. Australian Systematic Botany 4 (1):73-89. GRISWOLD, C. E. 1991b. A revision and phylogenetic analy- sis of the African spider genus Machadonia Lehtinen (Araneae: Lycosoidea). Entomologica Scandinavica 22:305-351. GRISWOLD, C. E. 1994. A revision and phylogenetic analysis of the African spider genus Phanotea Simon (Araneae: Lycosoidea). Annales, Musée Royal de |’Afrique cen- trale, Sciences Zoologiques 273:1—83 GRISWOLD, C. E. 1997a. The spider family Cyatholipidae in Madagascar (Araneae, Araneoidea). Journal of Arach- nology 25:53-83. GRISWOLD, C. E. 1997b. Scharffia, a remarkable new genus of spiders from East African Mountains (Araneae, Cy- atholipidae). Journal of Arachnology 25: 269-287. GRISWOLD, C. E. 1998a. The nest and male of the trap-door spider Poecilomigas basilleupi Benoit, 1962 (Araneae, Migidae). Journal of Arachnology 26: 142-148. GRISWOLD, C. E. 1998b. Wanzia fako, a new genus and species of spider from Cameroon (Araneae, Cyatholipi- OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 dae). Entomologica Scandinavica 29:121—130. GRISWOLD, C. E. 2000. Afromontane spiders in Madagascar (Araneae, Araneomorphae: Cyatholipidae, Phyxelidi- dae, Zorocratidae). Pages 345-354 in W. R. Lourengo and S. M. Goodman (eds.), Diversity and Endemism in Madagascar. Mémoires de la Société de Biogéographie, Paris. GRISWOLD, C. E. 2001. The living world genera and Afrotropical species of cyatholipid spiders (Araneae, Orbiculariae, Cyatholipidae). Memoirs of the California Academy of Sciences, No. 26. 251 pp. GRISWOLD, C. E. (in press) Spiders. Jn S. Goodman and J. Benstead (eds), Natural History of Madagascar. Univer- sity of Chicago Press. GRISWOLD, C. E., and N. I. PLATNICK. 1987. On the first African spiders of the family Orsolobidae (Araneae, Dysderoidea). American Museum Novitates 2892:1—14. GRISWOLD, C. E., and D. UBick. 1999. Spiders—Araneae. Pages 7-67 in C. E. Griswold (ed.), Arthropod surveys on Bioko, Equatorial Guinea. Report to Central African Regional Programs for the Environment (C.A.R.P.E.). 71 pp. GUERIN, F. E. 1839. Gastéracanthes sculptée et de Feisthamel, nouvelles especes d’Aranéides. Revue zo- ologique par la Société cuviérienne, Paris: 109-111. Harvey, M. S. 1996a. The biogeography of Gondwanan pseudoscorpions (Arachnida). Revue Suisse de Zoolo- gie, hors ser.: 255-264. Harvey, M. S. 1996b. Small arachnids and their value in Gondwana biogeographic studies. Pages 155-162 in S.D. Hopper, J. Chappill, M. S. Harvey, and A. George (eds.), Gondwanan Heritage: Past, Present and Future of the Western Australian Biota. Surrey, Beatty and Sons, Chipping Norton. HewiItt, J. 1913a. Description of a new trapdoor spider from Cape Colony. Annals of the Transvaal Museum 4:47. Hewitt, J. 1913b. Descriptions of new and little known species of trapdoor spiders (Ctenizidae and Migidae) from South Africa. Records of the Albany Museum 2: 404-434. Hewitt, J. 1913c. Descriptions of new species of Arachnida from Cape Colony. Records of the Albany Museum 2:462-481. Hewitt, J. 1914. Descriptions of new Arachnida from South Africa. Records of the Albany Museum 3:1—37. HewIrt, J. 1915a. Descriptions of several new or rare species of Araneae from the Transvaal and neighborhood. An- nals of the Transvaal Museum 5:89—100. HEwIrt, J. 1915b. Notes on several four-lunged spiders in the collection of the Durban Museum with descriptions of two new forms. Annals of the Durban Museum 1:125—-133. HEwIrt, J. 1916. Descriptions of new South African spiders. Annals of the Transvaal Museum 5:180-213. HEwIrt, J. 1919. Descriptions of new South African spiders and a Solifuge of the genus Chelypus. Records of the Albany Museum 3:196—215. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS Hoa, H. R. 1902. On some additions to the Australian spi- ders of the suborder Mygalomorphae. Proceedings of the Zoological Society of London 1902 (2):121—142. Howpen, H. F. 1981. Zoogeography of some Australian Coleoptera as exemplified by the Scarabaeoidea. Pages 1007-1035 in A. Keast (ed.), Ecological Biogeography of Australia. Junk, The Hague. INTERNATIONAL COMMISSION ON ZOOLOGICAL NOMENCLA- TURE. 1999. International Code of Zoological Nomen- clature, 4th edition. The Natural History Museum, Lon- don. 306 pp. JocquE, R. 1994. Halidae, a new spider family from Mada- gascar (Araneae). Bulletin of the British Arachnological Society 9:281-289. Kocn, L. 1873. Die Arachniden Australiens. Nurnberg: 369-472. Kraus, O. 1978. Zoogeography and Plate tectonics. Intro- duction to a General Discussion. Verhandlungen der Naturwissenschaftlichen Verein in Hamburg (NF) 21/22: 3341. LAWRENCE, R. F. 1928. Contributions to a knowledge of the fauna of South West Africa. v. Arachnida. Annals of the South African Museum 25 (2):217-312. LEGENDRE, R. 1979. Les araignées et la dérive des continents austraux. Annales de Biologie 18:37-68. LEGENDRE, R., and R. CALDERON GONZALEZ. 1984. Liste sys- tématique des araignées mygalomorphes du Chili. Bul- letin du Museum National d’Histoire Naturelle, Paris, Zoology 6(4):1021—1065. MappiIson, W. P., and D. R. MAppISsoN. 1992. MacClade, ver. 3.0. Sinauer Associates, Sunderland, Mass. Mam, B. Y. 1991. Occurrence of the trapdoor spider genus Moggridgea in Australia with descriptions of two new species (Araneae: Mygalomorphae: Migidae). Journal of Natural History 25:383-397. MatTHews, E. G. 1976. A revision of the Scarabaeine dung beetles of Australia, III. Tribe Coprini. Australia Journal of Zoology, Suppl. Ser. 39:1—52. MERRETT, P., G. H. LOCKET, and A. F. MILLIDGE. 1985. A check list of British spiders. Bulletin of the British Arachnological Society 6:381-403. MERRETT, P., and A. F. MILLIDGE. 1992. Amendments to the check list of British spiders. Bulletin of the British Arachnological Society 9:4-9. Myers, N. 1988. Threatened biotas: “Hot Spots” in Tropical Forests. The Environmentalist 8(3):187—208. Myers, N., R. A. MITTERMEIER, C. G. MITTERMEIER, G. A. B. DA Fonseca, and J. KENT. 2000. Biodiversity hotspots for conservation priorities. Nature (London) 403: 853-858. NATIONAL RESEARCH COUNCIL. 1980. Research Priorities in Tropical Biology. National Academy of Sciences, Washington D. C. 116 pp. NELSEN, G. and N. PLATNICK. 1984. Biogeography. Carolina Biology Readers, Scientific Publications Department, Barrington, North Carolina 119:1—16. Nixon, K. C. 1992. Clados, version 1.2. Program and docu- 49 mentation, available from author at L.H. Bailey Horto- rium, Cornell University, Ithaca, New York 14853. PETRUNKEVITCH, A. 1928. Systema Aranearum. Transactions of the Connecticut Acadademy of Arts, Sciences and Letters 29:1—270. PETRUNKEVITCH, A. 1939. Classification of the Araneae with key to the suborders and families. Pages 139-190 in A. Petrunkevitch (ed.), Catalogue of American Spiders. Transactions of the Connecticut Acadademy of Arts, Sciences and Letters, vol. 33. PIELOU, E. C. 1979. Biogeography. John Wiley & Sons, New York. 351 pp. PLATNICK, N. I. 1981. Spider biogeography: past, present, and future. Revue Arachnologique 3:85—96. PLATNICK, N. I. 1989. Advances in spider taxonomy: a sup- plement to Brignoli’s A Catalogue of the Araneae de- scribed between 1940 and 1981. Manchester. 673 pp. PLATNICK, N. I. 2001. The World Spider Catalog, Version 2.0 (Online Catalog). P. Merrett and H. D. Cameron (eds.), The American Museum of Natural History, New York (http://research.amnh.org/entomology/). PLATNICK, N. I., and M. U. SHADAB. 1976. A revision of the mygalomorph spider genus Neocteniza (Araneae, Actinopodidae). American Museum Novitates 2603: 1-19. Pocock, R. I. 1895. On new genera and species of trapdoor spiders. Annals and Magazine of Natural History 6 (16):187-197. Pocock, R. I. 1903. On the geographical distribution of spi- ders of the Order Mygalomorphae. Proceedings of the Zoological Society of London 1903:340—-368. PURCELL,W. F. 1903. New South African spiders of the Fam- ilies Migidae, Ctenizidae, Barychelidae, Dipluridae and Lycosidae. Annals of the South African Museum 3 (4):69-142. RASONAIVO, P. 1990. Rain Forests of Madagascar: Sources of Industrial and Medicinal Plants. Ambio 19: 421-424. RavEN, R. J. 1980. The evolution and biogeography of the mygalomorph spider family Hexathelidae (Araneae, Chelicerata). Journal of Arachnology 8:251—266. RAVEN, R. J., 1984. Systematics and biogeography of the my- galomorph spider family Migidae (Araneae) in Aus- tralia. Australian Journal of Zoology 32 (3):379-390. Raven, R. J. 1985. The spider infraorder Mygalomorphae (Araneae): cladistics and systematics. Bulletin of the American Museum of Natural History 182:1—180. Raven, R. J., and T. B. CHURCHILL. 1989. A new species of Migas (Araneae, Migidae), with notes on Heteromigas in Tasmania. Bulletin of the British Arachnological So- ciety 8 (1):5-8. RAxwortTHy, C. J., and R. A. NUSSBAUM. 1997. Biogeo- graphic patterns of reptiles in eastern Madagascar. Pages 124-141 in S. M. Goodman and B. D. Patterson (eds.), Natural Change and Human Impact in Madagas- car. Smithsonian Institution Press, Washington, D.C. RoEweR, C. F. 1942. Katalog der Araneae von 1758 bis 1940. Bremen: Natura 1:1—1040. 50 RoTH, V. 1992a. Preliminary list of spiders recorded from Madagascar with synonymies, present placement of species, and a partial bibliography. Unpublished techni- cal report, Institute for Conservation of Tropical Envi- ronments, 18 pp. Rotu, V. 1992b. Spiders of Ranomafana National Park and immediate area. Unpublished technical report, Institute for Conservation of Tropical Environments, 13 pp. SCHIAPELLI, R. D., and B. S. GERSCHMAN DE PIKELIN. 1973. La familia Migidae (Simon, 1892) en la Argentina (Araneae, Theraphosomorphae). Physis (Buenos Aires) Sec. C, 32 (85):289-294. SCHIAPELLI, R. D., and B. S. GERSCHMAN DE PIKELIN. 1975. Calathotarsus simoni sp. nov. (Araneae, Migidae). Ph- ysis (Buenos Aires) Sec. C, 34 (88):17-21. Simon, E. 1891. Descriptions d’espéces et de genres nou- veaux de la famille des Aviculariidae. Annales de la So- ciété entomologique de France 61:300—312. Stvon, E. 1892. Histoire Naturelle des Araignées, 1(1), Roret, Paris, pp. 1-256. Stmon, E. 1895. Histoire Naturelle des Araignées, 1(4), Roret, Paris, pp. 761—1096. Stmon, E. 1903. Description d’Arachnides nouveaux. An- nales de la Societe entomologique Belgique 47:21-39. Simon, E. 1907. Arachnides recueillis par L. Fea sur la Cote occidentale d’Afrique. Ire partie. Annali del Museo Civico di Storia Naturale di Genova (3) 3:218-323. Situ, A. G, D. G. SMITH and B. M. FUNNELL. 1994. Atlas of Mesozoic and Cenozoic Coastlines. Cambridge Univer- sity Press, Cambridge. 99 pp. OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 STRAND, E. 1908. Arachniden aus Madagaskar, gesammelt von Herrn Walter Kaudern. Zoologische Jahrbiicher, Abteilung fiir Systematik 6:453-488. Strand, E. 1934. Miscellanea nomenclatorica zoologica et palaeontologica, VI. Folia Zoologica et Hydrobiologica 6:271-277. Topp, V. 1945. Systematic and biological account of the New Zealand Mygalomorphae. Transactions of the Royal So- ciety of New Zealand 74:375-407. TRUSWELL, J. F. 1977. The Geological Evolution of South Africa. Purnell, Cape Town. 218 pp. VIETTE, P. 1991. Principales localités ou des Insectes ont été recueillis a Madagascar. Published by the author. 89 pp. WHITE, F., 1978. The afromontane region. Pages 464-513 in M. J. A. Werger (ed.), Biogeography and Ecology of Southern Africa. Junk, The Hague. White, F., 1983. The vegetation of Africa: a descriptive memoir to accompany the Unesco/AETFAT/UNSO vegetation map of Africa. United Nations Educational, Scientific and Cultural Organization, Paris. 356 p. WILTON, C. L. 1968. Migidae. Pages 74-126 in R. R. Forster and C. L Wilton (eds.), The spiders of New Zealand, Part II. Otago Museum Bulletin. YODER, A., M. CARTMILL, M. RUVOLO, K. SMITH and R. VIL- GALYs. 1996. Ancient single origin for Malagasy pri- mates. Proceedings of the National Academy of Sci- ences, USA, 93:5122—5126. ZaAPFE, H. 1961. La familia Migidae en Chile. Investiga- ciones Zoologicas Chilenas 7:151—157. FIGURES 1-68 GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS Ficure |. Paramigas milloti, new species, holotype female, dorsal . Illustration by JS. 5.0mm 53 54 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 FIGURE 2. Paramigas spp., juveniles. A, B. Reserve Speciale Ivohibe, Madagascar. C, D. Ambanizana, Masoala, Madagas- car. A. Cephalothorax, dorsal. B. Cephalothorax, anterior. C. Apex of sternum, labium, and mouthparts, ventral. D. Ocular area, dorsal. Scale bars: A = 1000 um, B = 430 um, C = 500 um, D = 250 um. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 55 Ficure 3. Morphology of Paramigas. A, B. Juvenile from Ambanizana, Masoala, Madagascar. C, D. Paramigas oracle, new species, female from Talatakely, Madagascar. A. Spinnerets, ventral. B. Left spinnerets, ventral. C. Pedipalpal tarsus claw. D. Tarsus I claws. Scale bars: A= 250 um, B = 120 um, C = 100 um, D = 200 um. 56 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. FIGURE 4. Calathotarsus coronatus, female from Cerro La Campana, Chile, dorsal. Illustration by JS. 151 GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 57 0.3mm FiGuRE 5. Calathotarsus female morphology. A, B. Calathotarsus simoni from Cerro Negro, Argentina. C-E. Calathotarsus coronatus from Cerro La Campana, Chile. A. Cephalothorax, lateral. B. Cephalothorax, ventral. C. Right pedipalp, prolater- al. D. Dentition of right chelicera. E. Right tarsal claws: pedipalp, leg I retrolateral, leg II retrolateral, leg III prolateral, leg IV prolateral. Illustrations A-C by JS, D by JL, E by CG. 58 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 Ficure 6. Calathotarsus coronatus, male from Cerro La Campana, Chile, dorsal. Illustration by JS. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 59 0.3mm Ficure 7. Calathotarsus coronatus, male from Cerro La Campana, Chile. A-C. Left pedipalp tibia and tarsus. A. Prolateral. B. Ventral. C. Retrolateral. D. Leg I, retrolateral. E. Dentition of right chelicera. F. Left tarsal claws: leg I retrolateral, leg IT retrolateral, leg II prolateral, leg IV prolateral. Hlustrations A-D by JS, E by JL, F by CG. 60 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 Ficure 8. Heteromigas dovei, female from from Patersonia, Tasmania, dorsal. Illustration by JS. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 61 0.5mm FiGuRE 9. Heteromigas dovei, female from from Patersonia, Tasmania. A. Cephalothorax, lateral. B. Cephalothorax, ventral. C. Dentition of right chelicera. D. Left tarsal claws: pedipalp, leg I retrolateral, leg IT retrolateral, leg III prolateral, leg IV prolateral. Illustrations A-B by JS, C by JL, D by CG. 62 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 Ficure 10. Heteromigas dovei, male from from Patersonia, Tasmania, dorsal. Illustration by JS. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 63 0.5mm Ficure 11. Heteromigas dovei, male from from Patersonia, Tasmania. A—C. Left pedipalp tarsus. A. Retrolateral. B. Ventral. C. Prolateral. D. Left leg I, retrolateral. E. Dentition of right chelicera. F. Left tarsal claws: leg I retrolateral, leg II retrolat- eral, leg III prolateral, leg IV prolateral. Hlustrations A-D by JS, E by JL, F by CG. 64 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 3.0mm Cue 3.0mm FiGuRE 12. Mallecomigas schlingeri, holotype female. A. Cephalothorax, dorsal. B. Cephalothorax, ventral. C. Left leg I, retrolateral. D. Right leg III, prodorsal. E. Dentition of right chelicera. F. Tarsal claws: right pedipalp, left leg I retrolateral, left leg II retrolateral, right leg III prolateral, right leg IV prolateral. Illustrations A-D by JS, E by JL, F by CG. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 65 FiGurE 13. Migas gatenbyi, female from Wellington, New Zealand. A. Cephalothorax, lateral. B. Cephalothorax, dorsal. C. Right leg I, retrolateral. D. Dentition of right chelicera. E. Left tarsal-claws: pedipalp, leg I retrolateral, leg II retrolateral, leg III prolateral, leg IV prolateral. Hlustrations A-C by VK, D by JL, E by CG. 66 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 0.2mm FIGURE 14. Goloboffia vellardi, female from Guanaqueros, Chile. A. Cephalothorax, dorsal. B. Cephalothorax, ventral. C. Right leg I, retrolateral. D. Left leg II, prolateral. E. Dentition of right chelicera. F. Left tarsal claws: pedipalp, leg I retro- lateral, leg II retrolateral, leg III retrolateral, leg IV prolateral. Illustrations A-D by JS, E by JL, F by CG. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 67 0.3mm Ficure 15. Moggridgea intermedia, female from Diepwalle, South Africa. A. Cephalothorax, lateral. B. Cephalothorax, dor- sal. C. Right leg I, retrolateral. D. Right leg III, prolateral. E. Dentition of right chelicera. F. Left tarsal claws: pedipalp, leg I retrolateral, leg II retrolateral, leg III prolateral, leg IV prolateral. Illustrations A-D by VK, E by JL, F by CG. 68 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 FIGURE 16. Moggridgea peringueyi, male from Oudtshoorn, South Africa. A. Left pedipalp tibia and tarsus, retrolateral. B. Right leg I, retrolateral. Ilustrations by VK. e Bi FiGuRE 17. Spermathecae of Moggridgea spp., dorsal. A, B. M. tingle paratype (WAM 90/1112), Valley of the Giants, Aus- tralia. C. M. intermedia, Diepwalle, South Africa. Arrows to bands of sclerotization. Scale bars for A, C = 0.2 mm, for B = 0.1 mm. _——_ gee a GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS Ficure 18. Poecilomigas basilleupi, female from Mazumbai, Tanzania, dorsal. Illustration by JS. 69 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 70 FiGurE 19. Spermathecae of Migidae, dorsal. A. Poecilomigas basilleupi, Mazumbai, Tanzania. B. Poecilomigas abrahami, Sordwana Bay, South Africa. C. Migas gatenbyi, Wellington, New Zealand. HS - spermathecal head, SS - spermathecal stalk. Scale bars: = 0.2 mm. _—_ 0.5mm FIGURE 20. Poecilomigas basilleupi, male from Mazumbai, Tanzania. A, B. Left pedipalp patella-tarsus. A. Ventral. B. Retro- lateral. C. Dentition of right chelicera. Illustrations A, B by JS, C by CG. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 71 FIGURE 21. Nests of Paramigas. A, B. P. perroti (lectotype of P. subrufis). C, D. P. oracle, new species, holotype. A, C. Outer. B, D. Lateral. Scale bars: = 1 cm. 72 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 5.0mm FIGURE 22. Micromesomma cowani, female from MRAC, dorsal. Illustration by JS. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 783 5.0mm ao ee Pe : e@ @ ofe @ ee ® ooo e@ ee G e@ : ® 5.0mm 0.3mm 0.3mm FIGURE 23. Micromesomma cowani. A-C, E. Female from MRAC. D. Female from MNHN. F. Lectotype female from BMNH. A. Lateral. B. Ventral. C. Dentition of right chelicera of two individuals. D, E. Spermathecae, dorsal. F. Tarsal claws: pedipalp, left leg I retrolateral, left leg II retrolateral, left leg II prolateral, right leg IV prolateral. Illustrations A, B by JS, C-E by JL, F by CG. 74 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 FiGuRE 24. Micromesomma cowani, female from MNHN. A. Leg I, retrolateral. B. Tibia-metatarsus I junction, dorsal, show- ing thorns. C. Tarsus IV trichobothrial base. D. Tarsal organ IV. E. Right chelicera. Arrow to intercheliceral basal swelling. Scale bars: A= 1000 pm, B = 231 um, C, D = 38 um, E = 600 um. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS TS Ficure 25. A, B, D. Micromesomma cowani, female from MNHN. C. Paramigas oracle, female from Talatakely, Madagas- car. A-C. Tibia II. A. Prolateral. B. Retrolateral. C. Dorsal. D, Coxa III showing thorns. Arrow to prolateral ridge of tibia II. Scale bars: A, B = 750 um, C = 600 um, D = 300 tm. 76 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 FIGURE 26. Spermathecae of Micromesomma cowani, dorsal. A, C. MNHN females. B. MRAC female. Scale bars: = 0.2 mm. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS He FIGURE 27. Paramigas oracle, new species, female from Talatakely, Madagascar. A. Tibia-metatarsus I, retrolateral. B. Retro- lateral view of tibia-metatarsus I junction showing serrate, procumbant setae. C. Dorsal view of serrate, procumbant setae at apex of tibia I. D. Tarsus IH, dorsal, showing trichobothrial distribution and tarsal organ (arrow). Scale bars: A= 1000 um, B = 430 um, C = 75 wim, D = 270 um. 78 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 1.0mm 0.1mm 0.2mm FIGURE 28. Paramigas alluaudi, female holotype of Myrtle alluaudi. A. Cephalothorax, dorsal. B. Right leg I, retrolateral. C. Spermathecae, dorsal. D. Dentition of right chelicera. E. Left tarsal claws: pedipalp, leg I retrolateral, leg ‘II retrolateral, leg III prolateral, leg IV prolateral. Illustrations A, B by JS, C, D by JL, F by CG. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 2.0mm FiGuRE 29. Paramigas andasibe Raven, new species, male from Ambohitantely, Madagascar, dorsal. Illustration by JS. 79 80 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 G 0.3mm FiGuRE 30. Paramigas andasibe Raven, new species, male from Ambohitantely, Madagascar. A. Cephalothorax, ventral. B. Right leg I, retrolateral. C. Dentition of right chelicera. D—F. Left pedipalp tibia-tarsus. D. Prolateral. E. Ventral. F. Retrolat- eral. G. Left tarsal claws: leg I retrolateral, leg II retrolateral, leg III prolateral, leg IV prolateral. Illustrations A, B, D-F by JS, C by JL, G by CG. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 81 1.0mm 0.5mm FiGure 31 Paramigas andasibe Raven, new species, holotype male. A—-C. Left pedipalp patella-tarsus. A. Prolateral. B. Ven- tral. C. Retrolateral. D. Right patella-tarsus I, dorsal. E. Right patella-tarsus I, retrolateral. F. Left tibia-tarsus III, retrolater- al. G. Dentition of right chelicera. H. Tarsal claws: right leg I, retrolateral, left leg II retrolateral, left leg II] and right IV, pro- lateral. Illustrations A-F by JS, G by JL, H by CG. 82 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 FIGURE 32. Pedipalpi of male Migidae. A-C. Paramigas andasibe Raven, new species, from Manakambahiny, Madagascar, right pedipalp. A. Patella-tarsus, retrolateral. B. Patella-tarsus, prolateral. C. Tarsus, dorsal. D. Moggridgea pseudocrudeni from Alicedale, South Africa, apex of pedipalpal tarsus. Arrows to ectal lobe of tarsus. Scale bars: A, B = 1000 pm, C, D= 380 um. 83 GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS .Omm 5 Illustration by JS. dorsal. I > holotype female new species, Sj FiGuRE 33. Paramigas goodmani 84 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 0.3mm FiGure 34. Paramigas goodmani, new species, holotype female. A. Lateral. B. Ventral. C. Spermathecae, dorsal. D. Denti- tion of right chelicera. E. Left tarsal claws: pedipalp, leg I retrolateral, leg II retrolateral, leg III prolateral, leg IV prolateral. Illustrations A, B by JS, C, D by JL, E by CG. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 85 FiGuRE 35. Spermathecae of Paramigas spp., dorsal. A. P. perroti (large female, MRAC). B. P. perroti (small female, MRAC). C. P. perroti (syntype of Paramigas subrufus). D. P. goodmani, holotype. E. P. milloti, holotype. HS — spermathe- cal head, SS — spermathecal stalk. Scale bars: = 0.4 mm. 86 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 FIGURE 36. Paramigas macrops, new species, holotype male, dorsal. Illustration by JS. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 87 i = : 0.3mm H FIGURE 37. Paramigas macrops, new species, holotype male. A. Lateral. B. Cephalothorax, ventral. C. Right leg I, retrolat- eral. D. Dentition of right chelicera. E-G. Left pedipalp patella-tarsus. E. Prolateral. F. Ventral. G. Retrolateral. H. Left tarsal claws: leg I retrolateral, leg II retrolateral, leg III prolateral, leg IV prolateral. Illustrations A-C, E-G by JS, D by JL, H by CG. 88 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 0.3mm FIGURE 38. Paramigas manakambus, new species, holotype male. A. Cephalothorax, dorsal. B. Dentition of right chelicera. C. Right leg I, retrolateral. D. Right leg IV, prolateral. E-G. Left pedipalp patella-tarsus. E. Prolateral. F. Ventral. G. Retro- lateral. H. Left tarsal claws: leg I retrolateral, leg II retrolateral, leg III prolateral, leg IV prolateral. Illustrations A, C-G by JS, B by JL, H by CG. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 89 FIGURE 39. Paramigas milloti, new species, holotype female. A. Lateral. B. Ventral. C. Spermathecae, dorsal. D. Dentition of right chelicera. E. Left tarsal claws: pedipalp, leg I retrolateral, leg II retrolateral, leg III prolateral, leg IV prolateral. Il- lustrations A, B by JS, C, D by JL, E by CG. 90 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 FiGure 40. Spermathecae of Paramigas spp., dorsal. A. P. oracle, Talatakely, Madagascar. B. P. oracle, holotype (SS fore- shortened in this view). C. P. pauliani (small paratype of Legendrella pauliani). D. P. pauliani (large paratype of Legendrel- la pauliani). HS — spermathecal head, SS — spermathecal stalk. Scale bars: = 0.3 mm. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 91 1.0mm 0.2mm 0.3mm 0.3mm Ficure 41. Paramigas oracle, new species, female. A-C, F, G holotype. D, E. Talatakeley, Madagascar. A. Cephalothorax, dorsal. B. Right leg I, retrolateral. C, D. Spermathecae, dorsal. E, F. Dentition of right chelicera. G. Left tarsal claws: pedi- palp, leg I retrolateral, leg II retrolateral, leg III prolateral, leg IV prolateral. Illustrations A, B by JS, C-F by JL, G by CG. 92 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 FIGURE 42. Paramigas oracle, new species, female from Talatakely, Madagascar. A. Tibia III, prolateral. B. Junction of patel- la-tibia II. C. Venter of femur II] showing membrane. D. Tarsal organ and trichobothrium on tarsus III. Arrow to prolateral ridge of tibia III. Scale bars: A = 600 um, B = 250 um, C = 750 um, D = 38 um. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 93 FIGURE 43. Paramigas oracle, new species, female from Talatakely, Madagascar. A. Fang. B. Dentition of fang furrow (arrow to basal swelling on chelicera). C. Right pedipalpal coxa. D. Cuspules on right pedipalpal coxa. Scale bars: A = 500 um, B = 430 um, C = 600 um, D = 120 um. 94 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 FIGURE 44. Paramigas pauliani, female. A-E. Holotype of Legendrella pauliani. F—1. Paratypes of Legendrella pauliani. A. Cephalothorax, dorsal. B. Cephalothorax, ventral. C. Right leg I, retrolateral. D. Left leg III, prolateral. E, F. Dentition of right chelicera. G, H. Spermathecae, dorsal. I. Tarsal claws: pedipalp, left leg I retrolateral, right leg II retrolateral, right leg Ill prolateral, left leg IV prolateral. Illustrations A-D by JS, E-H by JL, I by CG. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 95 FiGure 45. Spermathecae of Paramigas spp., dorsal. A. P. pectinatus, holotype (narrow stalks of spermathecae partially hid- den). B. P. rothorum, holotype. C. P. alluaudi (holotype of Myrtale alluaudi). Scale bars: = 0.2 mm. 96 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 6 (ep 6 0.2mm FIGURE 46. Paramigas pectinatus, new species, holotype female. A. Cephalothorax, dorsal. B. Right leg I, retrolateral. C. Spermathecae, dorsal. D. Dentition of right chelicera. E. Left tarsal claws: pedipalp, leg I retrolateral, leg I retrolateral, leg III prolateral, leg IV prolateral. Illustrations A, B by JS, C, D by JL, E by CG. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 97 0.5mm FIGURE 47. Paramigas perroti, female. A-E, I. Lectotype of Myrtale perroti. H. Paralectotype of Paramigas subrufus. F, G. MRAC specimens. A. Cephalothorax, dorsal. B. Cephalothorax, ventral. C. Right leg I, retrolateral. D. Left leg II, prolater- al. E, F. Dentition of right chelicera. G, H. Spermathecae, dorsal. I. Left tarsal claws: pedipalp, leg I retrolateral, leg II retro- lateral, leg III prolateral, leg IV prolateral. Illustrations A-D by JS, E-H by JL, I by CG. 98 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 2.0mm FIGURE 48. Paramigas rothorum, new species, holotype female, dorsal. Illustration by JS. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 99 B ee (( 0.2mm FiGure 49. Paramigas rothorum, new species, holotype female. A. Lateral. B. Ventral. C. Spermathecae, dorsal. D. Denti- tion of right chelicera. E. Left tarsal claws: pedipalp, leg I retrolateral, leg II retrolateral, leg HI prolateral, leg IV prolateral. Illustrations A, B by JS, C, D by JL, E by CG. 100 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 2.0mm FiGuRE 50. Thyropoeus malagasus, female from Vohimena, Madagascar, dorsal. Illustration by JS. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 101 SS AS. Meaney 4h) nay ‘i ava: and ahh Ey i i He tds 0.2mm E 0.3mm FiGuRE 51. Thyropoeus malagasus. A-D. Female from Vohimena, Madagascar. E. Holotype female of Heteromigella mala- gasa. A. Lateral. B. ventral. C. Spermathecae, dorsal. D. Dentition of right chelicera. E. Left tarsal claws: pedipalp, leg I retrolateral, leg II retrolateral, leg III prolateral, leg IV prolateral. Illustrations A, B by JS, C, D by JL, E by CG. 102 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 Ye 2.0mm FIGURE 52. Thyropoeus malagasus, male from Vohimena, Madagascar, dorsal. Illustration by JS. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 103 0.3mm FiGuRE 53. Thyropoeus malagasus, male from Vohimena, Madagascar. A. Lateral. B. Cephalothorax, ventral. C. Left leg I, retrolateral. D. Dentition of right chelicera. E-G. Left pedipalp patella-tarsus. E. Retrolateral. F. Ventral. G. Prolateral. H. Left tarsal claws: leg I retrolateral, leg II retrolateral, leg III prolateral, leg IV prolateral. Illustrations A-C, E—G by JS, D by JL, H by CG. 104 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 FiGuRE 54. Thyropoeus malagasus, male from Vohimena, Madagascar, right pedipalpus. A. Tibia—tarsus, prolateral. B. Tar- sus, dorsal. C. Tarsus, retrolateral. Scale bars: A= 1000 um, B = 380 um, C = 600 um. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 105 FiGuRE 55. Thyropoeus malagasus, male from Vohimena, Madagascar. A. Right tibia I, dorsal. B. Apex of right tibia I, retro- lateral. C. Trichobothrial base, metatarsus I. D. Tarsal organ I. Scale bars: A= 1200 um, B = 500 um, C, D = 30 um. 106 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 FIGURE 56. Spermathecae of Thyropoeus spp., dorsal. A. T. mirandus, small female, Ft. Dauphin, Madagascar. B. T. miran- dus, holotype, left. C. T. mirandus, large female, Ft. Dauphin, Madagascar. D. T. malagasus, Vohemena, Madagascar. Scale bars: A-C = 0.4 mm, D = 0.2 mm. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 107 FIGURE 57. Thyropoeus mirandus, female holotype, dorsal. Illustration by JS. 108 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 0.5mm FiGuRE 58. Thyropoeus mirandus, female. A-C, E. Holotype female. D, F, G. Ft. Dauphin, Madagascar. A. Lateral. B. Cephalothorax, ventral. C. Left spermatheca, dorsal. D. Spermathecae, dorsal. E, F. Dentition of right chelicera. G. Left tarsal claws: pedipalp, leg I retrolateral, leg II retrolateral, leg III prolateral, leg IV prolateral. Illustrations A, B by JS, C-F by JL, G by CG. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 109 FiGuRE 59. Morphology of Miginae. A-C. Migas gatenbyi, female from Wellington, New Zealand. D. Poecilomigas abra- hami, female from Grahamstown, South Africa. A. Patella-tibia III, prolateral. B. Leg III ventral, showing membrane, arrow to proximal extent. C. Metatarsus-tarsus TV junction, showing widely spaced comb setae (arrows). D. Fang base showing tooth. Scale bars: A= 600 um, B = 750 um, C = 200 um, D = 190 um. 110 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 FiGuRE 60 Morphology of male Migidae. A. Poecilomigas abrahami, from Town Bush, Pietermaritzburg, South Africa, metatarsus-tarsus IV showing ventral scopula. B. Migas sp., from Tauranga, New Zealand, metatarsus-tarsus IV showing lack of ventral scopula. C. Moggridgea pseudocrudeni, from Alicedale, South Africa, femur I ventral showing carina. Scale bars: = 500 um. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 11] FIGURE 61. Moggridgea intermedia, female from Diepwalle, South Africa. A. Patella-tibia III, prolateral. B. Patella-tibia IIT junction. C. Leg III ventral showing membrane (arrow to proximal extent of membrane). D. Lamellate setae beneath patella IV. Scale bars: A = 600 um, B = 250 um, C = 1000 um, D = 150 um. 112 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 FIGURE 62. Metatarsus IV combs of Migidae. A, B. Calathotarsus coronatus, female from El Canelo, Chile, metatarsus IV apical comb, retroventral. C. Moggridgea crudeni, female from Alicedale, South Africa, metatarsus IV comb. D. Moggridgea dyeri, female from Uitenhage, South Africa, metatarsus IV preening comb base. Scale bars: A = 200 um, B = 100 pm, C = 200 um, D = 25 um. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS 113 FIGURE 63. Spermathecae of Migidae, dorsal. A. Calathotarsus simoni, Sierra de la Ventana, Argentina. B. Heteromigas dovei, Patersonia, Tasmania. C, D. Mallecomigas schlingeri, holotype. D. Close up of left spermatheca. E, F. Goloboffia vel- lardi, Guanaqueros, Chile. F. Close up of left spermatheca. HS - spermathecal head, SS - spermathecal stalk. Scale bars A, B, D, E = 0.2 mm, C = 0.4 mm, F = 0.1 mm. 114 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 FiGuRE 64. Spermathecae of migid outgroups, dorsal. A. /diops sp., Pietermaritzburg, South Africa. B. Actinopus sp., Lo- malinda, Colombia. C. Bothriocyrtum californicum, Eagle Rock, California, USA. Scale bars A, B = 0.2 mm, C = 0.4 mm. 115 S$ ‘(a (E) 0 “(h) d “(h) 0 (GS). N “(@) W “(b) 7 (0) A—9 “(Dd (5) a “(b) “( se : : 1) a(S) a ( ‘ 4 1) L pure (Z) Le Saeco otek mem ma pare an reese Gay isi din ue les eva metig ap een eetel es é JIJoig ‘$9 ANDI (Ley (1)94 (Lor Wt (ev VW (tz (Leb (er (Leb (9 (ise g (o)g4 (1)8z 6) (Le (zz (Dor (o)8 Aa He (o)2 = L 2) (2)sp UA 2 (e)sp al ae (ive : ee (Hee (e)sb Ss (6 (w)sp (W)s (ee (L)sp ()se Nn (Ws (Le (ez (o)6e (se S (Wer (ep N (o)6e (itz g (Wee Zz (zz (st ie (et (os (ez 6z (ist a (i)sz b (Wee Sie (6 (Ze } ‘ wo 2 Ss & & g % 8 o S ¢ O° & & & o << & & & ® ~ & A ~ ad) ~ O S¥ & S ~ ~~ Se ~~ v SS 2 z y S: v FIGURE 67. Area cladogram implied by preferred migid cladogram. E = eastern, NC = New Caledonia, NZ = New Zealand, S = southern, W = western. GRISWOLD AND LEDFORD: MADAGASCAR MIGIDS A Thyropoeus malagasus B Thyropoeus mirandus C Paramigas alluaudi D Paramigas andasibe E Paramigas goodmani F Paramigas macrops G Paramigas manakambus H Paramigas milloti | Paramigas oracle J Paramigas pauliani K Paramigas pectinatus L Paramigas perroti M Paramigas rothorum @ juvenile Paramigas ? Micromesomma cowani B CJ 117 FIGURE 68. Map of Madagascar showing records of Migidae. Locality data for juvenile Paramigas are listed in Table 1. APPENDIX Character by taxon matrix. Rows represent characters. The first state listed is coded as "0", the second as "1", etc., "2" = unknown, "-" = non-applicable. Columns represent taxa. Character statistics are for our preferred tree (Fig 65, which has length = 96 steps, consistency index = 0.52, retention index = 0.76, total fit [from PeeWee, concavity =3] = 359.9). The final five columns give the number of steps (St), the consistency index (CI), the retention index (RI), the weight from successive weighting from Hennig86 (Wt) and the fit from Pee-Wee. Taxon abbreviations are (vertically): Id = Idiops, Ac = Actinopus, Bo = Bothriocyrtum, Co= Calathotarsus spp., Ma = Mallecomigas schlingeri, Ht = Heteromigas dovei, Mg = Migas gatenbyi female + Migas taierii male, Mv = Goloboffia vellardi, Mp = Moggridgea peringueyi, Mr = Moggridgea rupicola, Mi = Moggridgea intermedia, Mt = Moggridgea tingle, Po = Poecilomigas abrahami, Pb = Poecilomigas basilleupi, Ta = Thyropoeus malagasus, Tm = Thyropoeus mirandus, Ms = Micromesomma cowani, Av = Paramigas milloti, ls = Paramigas pectinatus, Pu = Paramigas alluaudi, Rm = Paramigas oracle, Le = Paramigas pauliani, Vb = Paramigas goodmani, Md = Paramigas rothorum, Mb = Paramigas manakambus, Ah = Paramigas macrops, Pp = Paramigas perroti, An = Paramigas andasibe. Character abbreviations are (horizontally): abs = absent, ant. = anterior, apic = apical, aut = autapomorphy, Calath = calathotarsine type, cl = clavate, cusp = cuspules, dist. = distribution, dorsoap. = dorsoapical, F = femur, H = height, H/L = height divided by length, ITC = inferior tarsal claw, L = length, | = long stalk with head, Max. = pedipalpal coxa, med = median, Moggr = Moggridgea type, mt = metatarsus, n = head narrower than stalk, ns = narrow stalk, OAW = ocular area width, Pat = patella, palp = pedipalpus, PLS = posterior lateral spinnerets, procumb. = procumbant, prox. = proximal, prs = present, s = short, cylindrical, scler. = sclerotization, segs = segments, SS = sternal sigilla, T = tibia, t = tarsus, vent = ventral, W = width, < = less than, > = greater than, M = male, F = female. IABCMHMMMMMMP PTTMAIPRLVMMAPA Carapace dcooatgvpritobamsvsumebdbhpn St CI RI Wt Fit 1. OAW/ Caput W: < 0.41; > 0.45; OLOLTIVVIDD UTE 00) 50.0) 101 00 2. Wide OAW: normal; extra wide; -0-1000100000000000000000000 1 1.00 1.00 10 10.0 3. Prefoveal setae: abs; small; enlarged; 2111112111022201121102101111 9 0.22 0.22 0 3.0 4. Caput setation: extensive; reduced; 000000000000001111111111--1- 1 1.00 1.00 10 10.0 5. Caput H/ caput L: arched; low; 0000001111111100111111111111 3 0.33 0.71 2 6.0 6. Fovea: procurved to straight; recurved; OOORPT TA eee O01 200! 02 0'.0 7. Recurved Fovea: simple; tripartite; ---0010000000001111111111111 2 0.50 0.90 4 7.5 IABCMHMMMMMMP PTTMAIPRLVMMAPA Mouthparts dcooatgvpritobamsvsumebdbhpn St CI RI Wt Fit 8. Cheliceral tooth row #: 2; plus denticles; 0111010100000011110000100011 5 0.20 0.63 1 4.2 9. Intercheliceral basal swellings: abs; prs; 010000100?111100111111111111 4 0.25 0.62 1 5.0 10. M intercheliceral tumescence: abs; prs; 0001-0110?01110--------- OO= Om ZERO SO MOS OF a aaee7i5 11. Rastellum: prs; abs; 0001111111111111111111111111 1 1.00 1.00 10 10.0 12. Fang shape: round; quadrate & keeled; OO OPP ele L100; 15710.082'0) S05) 13. Fang basal tooth: abs; prs; 0100001100001110000000000000 4 0.25 0.40 1 5.0 14. Fang orientation: vertical; diagonal; 0101111111111111111111111111 1 1.00 1.00 10 10.0 15. F max cusp dist: across max; prox; 001110010000000000000000--0- 2 0.50 0.66 3 7.5 16. F max cusp ant dist: broad; corner only; 010001000000000000000000--0- 2 0.50 0 0) Ws IABCMHMMMMMMP PTTMAI PRLVMMAPA Legs and sternum dcooatgvpritobamsvsumebdbhpn St CI RI Wt Fit 17. Thorns on coxae II-III: abs; prs; 0000000001000000100000000000 2 0.50 (0) Oe Wate 18. Anterior sternal sigilla: prs; abs; 0001111111111111111111111111 1 1.00 1.00 10 10.0 19. Deeply excavate SS: abs; prs; 0000000000000011000000000000 1 1.00 1.00 10 10.0 20 Srethl,. Tl spines!:-prey; abs: 000000101111110000000000--0- 1 1.00 1.00 10 10.0 21. F ant. leg shape: round; flattened; OO Dae ae ee TEL ELT 2 OF 50) 0 0) oS 22. Pat vent lamellate setae: abs; prs; 0000000111110000000000000000 2 0.50 0.75 3 7.5 23. F I-II fine, vent setae: abs; dense, long; 000000100000100100001010--1- 4 0.25 0.40 1 5.0 24. F T I,II dorsoap setae: erect; procumb; 000000000000000011111111--1- 1 1.00 1.00 10 10.0 25. F Mt I,II dorsoap setae: erect; procumb; 000000000000000001111111--1- 1 1.00 1.00 10 10.0 26. F F III vent membrane: short; elongate; 000000000000001111111111--1- 1 1.00 1.00 10 10.0 27. F T III basal concavity: abs; prs; 0000000011110011111111111111 2 0.50 0.80 4 7.5 28. F T III anterior ridge: abs; prs; 000000000000000111111111--1- 2 0.50 0.88 4 7.5 29. F P-T 3 spination: stout; slender; 000111010000000000000000--0- 2 0.50 0.66 3 7.5 30. T dark maculations: abs; prs; 0000000000001100000000000000 1 1.00 1.00 10 10.0 31. Mt IV preening comb: abs; Calath; Moggr; 0001101122210000000000000000 3 0.66 0.83 5 7.5 32. Mt III preening comb: abs; prs; 0001100100000000000000000000 1 1.00 1.00 10 10.0 33. Preening comb form: narrow; encircling; == 1 1-010 000 <—'— n\n (ej ; <; 0000000000000011000000000000 1 1.00 1.00 10 10.0 35. ITC 3 size: large; small; 0011001111111110111111111111 4 0.25 0.25 0 5.0 119 120 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 151 Male characters 36. M F I venter: convex; carinate; 37. M palp t form: ectal=mesal; ectal>>mesal; 38. M palp t apical spinules: abs; prs; 39. M Retrolateral T I megaspine: abs; prs; 40. M mt I shape: cylindrical; pale swelling; 41. Mt III shape: cylindrical; swollen; Spinnerets 42. PLS apical segment: triangular; domed; 43. PLS spigot dist.: all segs.; med & apic; Female genitalia 44. Spermathecal scler.: 45. Spermathecal shape: uniform; Sense ds: median band; risi77 es; IABCMHMMMMMMP PTTMAI PRLVMMAPA dcooatgvpritobamsvsumebdbhpn St CI 0000-0001110000--------- 0.0) 0000-0001110000--------- 1 1.00 1000-0110000110--------- 4 0.25 1011-0100001011----- 5 0.20 0000-0000000000--------- 367 5 OW) 0000-0000000000--------- ale AES (OO) IABCMHMMMMMMP PTTMAIPRLVMMAPA dcooatgvpritobamsvsumebdbkhpn St CI 0101111111111111111111111111 1 1.00 0001111111111111111111111111 1 1.00 IABCMHMMMMMMP PTTMAIPRLVMMAPA dcooatgvpritobamsvsumebdbhpn St CI 000000001111000000000000??0? 1 1.00 3000112022222244323323237?2? 7 0.57 yo SMITHSONIAN INSTITUTION LIBRARIES TT 3 9088 01302 6430