Me took Me cet r* Rpentch ah os MARIS Ala Nigra Ree acai aaa ae, ARORA RE ORANG eA A Be Aare PR Now ae es ee TD Oe SON Tore Sob SS an Lene er Tae 7 Sea sien : DMD Re Lear ee SSMU SALA ares t AS: WRENS ASR AE ae AFM ATE Re SM Ang te aty tate Utes Marat a9 i TACT NS ROSA TACO NRA A Ma eras ROMA BURMAN Acie mags Ne QED: Mai ae DMA MRED Eline MPA RE NCA AE LE EB RENN NE Rn NR APA ONS In hats WPAN te Ae al LAS WR tiny: SONS YER SONI Cee me RESTO RA AL Gala’ ty Cony, cena ere) Brn NR Aaa me ett ide ae Sor 9 AE TERA rO ra TRO ae AAA AL Ort WA SERN AIT Ae ec ye Whe Me RgNAL : MAM AEVALE RAG CNCN ATL me eA aaty SOUZA TAMEOR Ah ta tel hg PRE Marae mace ts eaten wate AMA TRA NAR WMP Ra ae 95 3 Me WL A AA RR NAR time Earn ba jp eytbat ae Aste Vidas WIA Mt fm ARM AIAG MANY tyre, PNA TA RS ATE ATRTAZL Ar § asco asa ae Whang tee LVM MMF Ate a ag, Mera tae nth iden Rote PLA MAS & GeEALA Gi Ath SALOME Asse raceid tases ard F7.% HARVARD UNIVERSITY e Library of the Museum of Comparative Zoology a) vu Ow ae a i i A } Hh fa mn. a ine f ni sn ae 4 i i te kid aie ' a i 0 all Or He iia ‘ i, i : Y at ; " i if a | ; 7 7 j Ecology and Evolution of the Gastrochaenacea (Mollusca, Bivalvia) with Notes on the Evolution of the Endolithic Habitat Joseph Gaylord Carter Bulletin 41 PEABODY MUSEUM OF NATURAL HISTORY YALE UNIVERSITY Ecology and Evolution of the Gastrochaenacea (Mollusca, Bivalvia) with Notes on the Evolution of the Endolithic Habitat JOSEPH GAYLORD CARTER Department of Geology, University of North Carolina BULLETIN 41 e DECEMBER 1978 PEABODY MUSEUM OF NATURAL HISTORY YALE UNIVERSITY NEW HAVEN, CONNECTICUT 06520 Bulletins published by the Peabody Museum of Natural History, Yale University are numbered consecutively as independent monographs and appear at irregular intervals. Shorter papers are published at frequent intervals in the Peabody Museum Postilla series. The Peabody Museum Bulletin incorporates the Bulletin of the Bingham Oceanographic Collection, which ceased independent publication after Vol. 19, Article 2 (1967). Communications concerning purchase or exchange of publications should be addressed to the Publications Office, Peabody Museum of Natural History, Yale University, New Haven, Connecticut 06520, U.S.A. © Copyright 1978 by the Peabody Museum of Natural History, Yale University. All rights reserved. No part of this publication, except brief quotations for scholarly purposes, may be reproduced without the written permission of the Director, Peabody Museum of Natural History. Printed in the United States of America MUS. COMP, ZOOL. LIRmARYy JUN 41 6 i980 rH aA R V ra Hig * a rw VERSITY CONTENTS ; LUIS TE COT 6 OY GN 22) OS eee a nce eee rear reo ene rae Vv ABS RAGH (Hngtish, German, Russian) 3: a... oe sce5- 152 oe << > ve ] iN OTS NEOs B90 BC 1 OU PE a eee a KO ee Ne gk 7 ji ENA EF XO) BOO BL © Th ie ae ear oe ee ar XS aoe ei Pers 8 2a VEACEERTALS- AINDEME THODS:. sic. fie v3 icc ce cer eee eres 9 3. DISTRIBUTION AND ECOLOGY OF THE TROPICAL WESTERN ATLANTIC GASTROCHAENIDS .......2....% 13 oe, UN DAUET ELQIS BY Selle abe eet waar ae IMLS Gay 2 RRC R LE Fh pe eran Bop AeA 16 IW ENTLY S neste Ruy ey Rae™ An SR OA nd CARS. Be eR OPI US Sea ethy 16 SSID NG AS Peto weg Oe or 2 Sa te rar Mee ishavs (Na ehate, WN NOLO aes 18 Gfenidiajand labial palps: 24 <5 vscice cia Sek as ee od ee 2] TONE SES Se eee OS 2 ie ae Reale ge Wetonies eevee Aa tencrs ee Rena i AMEE aE Weak ee ee AO RC ar SN AOD era eras Ter ay 22 Katestinal contents... co. 26 2. sakes asst ro 1h ase AOE oar a 24 SARS el gee ON OK on cla ston, ML aus ctteiar SEI cere nt eae em 26 OSE ME NIC ROS PRUC BURE iieskacercs de uca.ee oo aco ace eres 30 fe GA MEN DoAIND DEN TERION 25 sce ates) eacceieiererens 4 aie eie once eal Soars UREN CON WN Sie res ea tee ya7s Sc SER ch stats & Sia geteuc She ehaeye le mmo a cl aces ee SNUCAININ E ROrCDORING 3500 ences. cith icc tiene ree oe ake omnes 48 HOGAS PT ROCHAE NID ORIGING 33. a5 cc ooleess ois ste woe eles iiehen ar 52 11. EVOLUTION AND ADAPTIVE RADIATION WITHIN EDR GAS PROCEUAPE NACH Acs o. 2ae ae Ware oun eae oS see eles 59 Generalevolutionany, trends, o.... saree = s5/5 2 a es ees nee 59 Biv olen ns te Perio let, acti, aaah ee swcols 2) oo woke ooresn fe wi anoles ores? 61 Evolution of Gastrochaena (Rocellaria) and Gastrochaena UE OSIFOCRACHA) RE ca ee et ee eh acl ne a) steel nee 61 Evolution of Gastrochaena (Cucurbitula), Kummelia and [OTE TUTE Oe eS IR errs ere ee Ot ee ee te 63 12. CONVERGENCES BETWEEN THE CLAVAGELLACEA PON TONCOAS MiROMGEIAE NAGASE oe ne ceraccaiciniajsis, oe winie deelentes -F 71 ee ESOS SU IN ego oy da 6 ten’ op BSc gost MVS a Yeo ee ors Ne 73 Diversification within the endolithic habitat .................-- 73 Erosional instability and tropical endolithic communities ....... 75 Taxonomic distinction of Spengleria, Cucurbitula Gastrochaena and | ID REGS fon OUT MRED peo SRO OP Ri Ree oe UCU IRP EE PWT PEN aret mae EO APPENDIX A. List of the Soldier Key Bivalvia associated with Diploria skeletons, with references to their illustrations in other works... 80 PP PENTT XS Be (PaxOnOUMICiOULLIME 6250 «ce sno eke ke es tee ae es 81 PePEN DIX IG. Description of Hew SPECIES . ...........6 2.6.6 soe ene ne 83 eer oes see son Sb chahece macs o.8 we one wlelé ste ens 87 Pee n rishi REBERENGES © 2 s0... .el. Sass c eens oe Se eee nies oie oe ial ay i. Oe a) af 4) 7 bee nel ¥ bel 3 § f=. ri | ai ® at ; ' : 7 oi he Ss & eo | : = | V i .. , noe ei nd 7 . } . i Beng? be 38, i tore mie — oa eae: To ania) a paps an meatal hai aes tr: eer bs Be bien oben? 3 ewe 7a 7 noe af 3° ani an pee ae itl ts oe ee | ery 7 aA Ral oll (O00 NID UTR 69 NO FIGURES . Growth bands versus comarginal shell ridges ................. 10 EASE -TECQUEME): CISEEIMUMIOEIS. occ cf. nicl tage crt = ale eae 11 Ventral mantle and pedal aperture) .. 25. cece ncn i 17 SUB HOIS I sc igs eee ae eae ets an ona ee nr ooh 19 Ctenidium: of Gastrochaena (G:) Rians ©. 02 ans we ewe ee ce ees 90 Se Musculature un Speneleria MOStratG ian. teas 2 etd ee 93 . Musculature m Gastrochaena (R.) ovata... 62 owe eens oe 93 * Musculatureim Gastrochaena (G.)htans oe. 2-6 ve oie ooo 55 3 6 oic:5,5' 93 eisize frequency. of MIgEsted Panticles .....).- .. << ts. so tle 24 EE SHEN INOLP NOOO. 5 ooo o-oo. 5 Mele a oe so cis eee © mie ie eile neta ole ae 27 eeSiiell Helghtiengthiversus ave 1. n).) cei see ise oe ie 28 Bele GiGi VERSUS ARE = ch. fe ote ethets fae fete) tec e ours ie fee oie el 28 . Apparent abrasive efficiency versus burrowing rate ........... 99 . Shell microstructure of Spengleria rostrata ........+..+0 +22 eee 3] . Shell microstructure of Gastrochaena (R.) ovata .........2.0+045: a . Shell microstructure of Gastrochaena (G.) hians ..........600+45: 3] . Posterior periostracal spikes in Spengleria rostrata .........-+.++- 32 . Lateral anterior periostracal spikes in Spengleria rostrata ........ 39 . Anteroventral periostracal spikes in Spengleria rostrata .......... 39 . Acetate peel of periostracum of Spengleria rostrata ...........-- 39 . Posterior-anterior periostracal thinning in Spengleria rostrata ... . 33 . Periostracal spike distribution in Spengleria rostrata ...........-- 34 . Periostracal spike size and shape variation ............-----+--- 34 ') Retiostiacal spike length VErsusiape =. si. 02 mae ee is ee 35 . Periostracal spike width/length versus age .............------- 36 . Periostracal secretion in Spengleria rostrata ........+++++e sees 37 . Periostracal spikes (light microscopy) ..........-++--+++-+ses2: 3¢ . Periostracal spike (scanning electron microscopy) ........----- Ey = Penmostracal spike (ultrastructure) ~ 0.2 2.0.2. soe ees nce ens os: mt . Periostracal spikes in juvenile Gastrochaena .........++++++++++5 38 . Composite prisms in Spengleria rostrata (transverse section) ..... 40 . Composite prisms in Spengleria rostrata (etched shell exterior) .. . 40 RISUREROWiCrOSS-SCOROMIS. 0.60 2 yay oii is ass, oe aioe oie 2s alee ae 42 . Burrow shell chamber width/height versus age .............--- 43 . Shell chamber lining in Spengleria rostrata ........6+-+0++000555 44 . Siphonal burrow lining in Gastrochaena (G.) hians ........-+-+++: 44 . Probing tubules of Gastrochaena (G.) truncata .........++++00055 45 . Siphonal burrow aperture of Gastrochaena (G.) hians .........-- 45 . Burrow shell chamber of Spengleria rostrata .......-.+++++0055: 45 . Siphonal burrow apertures of Spengleria rostrata .........---+++: 45 . Burrow shell chamber of Gastrochaena (G.) hians ............++: 45 Pustirkow length WEISS AGE S22. roc oso cies oe 5 ae esis aivi s cies so ene 46 » Burrowing Tate Versus age, .25.04. 6.25. 00e- Ree Aa ee fork Ae 46 . Burrow latex casts of Gastrochaena (G.) hians ...............00- 47 - Anteroventral mantle of Gasirochaena(G.)hians ©2222-2220) 25 49 . Pedal probing by Gastrochacna(G.) tameaia: 520.2 2n toe 50 . Calcareous tube of fossil Gasirochazna (G)) sp. 2... eee 50 . Comparisons with possible gastrochaenid relatives ............. 53 J bstract.Of data m Rig 418i. 250 -acc sc ake ane, eee es 54 . Generic diversity (Gastrochaenidae, Hiatellidae, Permophordae) gs .56.05.5 aces wae tiem oe na see eee ee ae » Shell of Myeconcha spot Winters 1903. 3. 22602 = cee 57 : Shelliot Sangumolies? spot Ghromici9o2 ye ajc oh eee 57 . Shell of Spenglersa spengleri of Deshayes 1860 ........ 2.2% a: sone 58 : Shelltot Gastrochaena moreana: 2.0. oe os css ee 60 . Natural burrow casts of Gastrochaena moreana .:.............+: 60 Burrow form of Gasivrochacna: norcanae: - Os. on ee ance 2 ee 60 . Shell-‘and burrow of Gastrochaena,(h:) lunsley 3222 32 ss ee 62 . Calcareous igloo of Gastrochaena (Cucurbitula) cymbium ......... 64 . Burrow latex casts of Gastrochaena (Cucurbitula) cymbium ....... 64 . Shell of Gastrochaena (Cucurbitula) cymbium ...... 2, ie See 65 . Natural casts ot Kummeha americana tubes ..3.). 2 spe ee 67 > Caleareous tube of Eujfistulana mumia . <2. oe re ee 68 > Shellof Pufistulana miunia.s cs veces oq oe eee 69 . Calcareous tube of Clavagella muliangularts ........ 0... .0.004-- 79 . Shellofi€lavagella multangulants 0s an hoe shoe seeeto a aoe ane 79 . Comparisons among Spengleria, Gastrochaena s.s., ANG TROCEMATIA, | Lote. eA Sl tos Wd eo be er 78 a Abstracts of data in’ Fig? GO" 23 Kup Ku, ®nopuga, CLUA (Sol- dier Key, Florida), Spengleria rostrata, Gastrochaena (Gastrochaena) hians w Gastrochaena (Rocellaria) ovata OYeHb xOPOWO NpucnocobNneHb! K *KU3HV B TOHKMX UV 6bICTPO pa3mMbiBaeMbIx CyOCTpaTax. OHV O4eHb XOPOWO Npucnoco6unncb K 3TOU Cpegfe, Ppa3BuB, U3-3a HEOObIKHOBEHHON CBOeNM CNOCO6HOCTH CucpoHanbHOrO CTATMBaHUA UV YANNHeHNA, YHNKalbHbIN MEXaHu3M ANA HanpaBNeHHoro CBepneHuA YU CNOCO6HOCTb NpoTpaBNuBannaA XOAOB B NOpose U ux 6bICTpOro BOCCTAaHOBNeHMA B cnyyYae noBpex*GeHua. Gastrochaenacea BepoarTHee BCErO Ppa3sBUNaCb B TPNACOBbIN UNV PaHHun topcKun nepvog u3 *vBwUux B MeNnKUX xOMax permophorids unu grammysiids B NPOMEXYTOYHbIA NONYKAMeHMCTbIN Nepvog FTHeE3GOBaHUA. OBONHOUNA *KUBLUUX B KA@MeHUMCTOM rpyuTe lithophagids (Mytilacea) uv gastrochaenids morna npou3souTu B OTBET Ha 3KCNAHCUIO CpefbI B KAMEHMCTOM IpyHTe, CONYTCTBYIOLUYHO TONVACOBOMY uv topckomy nepvogam nponudepaunu cKnNepakTuHnaHcKkux (scleractinian) KOopannos. Gastrochaenids nogBeprnucb BTOPYYHOK afaNTMUBHOM pagnuayun B MENOBOKM VU TPeETMYHbIN Nepvovb!, B pe3syNbTaTe Yero O6pa3s0Banncb >*KUByLuMe B tTpy6kax Kummelia wu Eufistulana, a TakxxKe Gastrochaena (Cucurbitula), xoppbi KOTOPOU HanOMUHalOT NO CopmMe NpOAONbHO pa3spesaHHyto rpywy. *KuByujve B Tpy6Kax gastrochaenids u clavagellids npeactasnatot cob60onv Knaccuyeckun O6pa3ey CNHXPOHHOM 3BONIOUMOHHOM KOHBeEpreHuMUn BkKONOrMYeCKM UV @unoreHetTuyecku pasnunyHbix npegAkoB. Clavagellacea BepovATHee BCeroO Ppa3BUNNCb U3 *®KUBWMX B rnyOoKux xog“ax npeActasButTenen Pandoracea. AlMatomosBbie BOsOpOocnu, KOTOPbIe, BO3MO>KHO, ABNAHWTCA NpeocbnagsarouyumM SNEMeHTOM CbuTONNaHKTOHA, NO-BUAMMOMY HE ABNAHOTCA PNaBHbIM UCTOYHUKOM nutaHyva Ansa gastrochaenids B Cong>Kep Ku, nockonbky GnvaTOMOBbIe pakKOBUHbI PpeAKO BCTpeYaloTcA B COAeP>xKUMOM Ux oKeNyAKa. DanbHenuwMe uccnefoBaHuA MOryT NOATBEPAMTb BEPOATHOCTb TOPO, YTO OCHOBbIM MCTOYHMKOM MUTaHuA ANA gastrochaenids ABNAWTCA MUKpOOpraHu3MbI (Microbiota) OCagKOB unu NNAHKTOHHbIe OpraHv3Mbl, HE COAepxKaljVe MMHEPaNbHbIX PaKOBUH. OcdekTuBHaA COPTMpPOBKa NO BeNnnyuHe NpPOornoyeHHbIx YACTML, COOTHOCMTCA CO cKnagkamu >»a6p uc 6onee MHOFOUWCNeHHbIMM FNaBHbIMU CUG(OHANbHbIMU uynanbuamu y Spengleria rostrata. Spengleria HagoO CuvTaTb OTNMYalOWMMUCcA OT Gastrochaena Ha ypoBHe poga U3-3a YHWVKAaNbHOM HOXKHOU MYCKyNAaTypbl, COBEPWEHHO CAMOCTOATENbHbIX CUMOHOB, CKNagyatTbix *abp, yMeEPeHHOM peAyKuMM NepeAHeEM paKOBUHbI U 3agepxKuBaHNA NepvocTpakaNnbHONW KanbCuduKaunn ANA MexaHnvyecKoro cBpepneHuaA Ha B3pOCNOH cTaguu. Gastrochaena s.s. wu Rocellaria @yHAaMeHTanbHO OTNUYAaIOTCA Apyr OT Apyra Ha ypOoBHeE nNogApoAa TONbKO xapakTepom pevAykKuun nepeAHux KOHUOB PaKOBMH UW NpKCyTCTBMeM UNM OTCYTCTBYeM BbINyYKNOM MYCKyNbHOU (myophoral) nogAnopKu AnA nepeAHen 6 PEABODY MUSEUM BULLETIN 41 OTCYTCTBNeEM BbINYKNOUW MYCKyYNbHOU (myophoral) nognopku ANA nepeAHen COKPaTUTeNbHONM MbiWuUbI. MogpoA Cucurbitula cnepzyetT OTnNuYaTb OT Gas- trochaena s.s. wv Rocellaria u3-3a O6A3aTeNbHOrO CBOMCTBa NpvAaBaTb xo”y cbopmy, HanomMuUHalOLy!O NPOAONbHO paspesaHHytlo rpylWy, U3-3a 3aBeEPHyTON Ha nepeAHeM KOHUe MaHTUY VU U3-3a APyrux XapaKTEPHbIX OCOH6EHHOCTEN PaKOBUHbI xoma. Gastrochaena s.s., Rocellaria w Cucurbitula npeactasnatot co60nu 3 nogpoga pofa Gastrochaena. KanbcuduKkauna nepuoctpakyma y S. rostrata reHeTUYeCKW OTNMNYHAa OT Kanbcudpukayun HYKHUX CNOeB pakOBUHbi. MpucyTcTBUue NepnocTpakaNnbHON KanbCudpukalun y B3pOCNbIX OpraHv3MOB BCeX BUAOB COBpeMeHHbIX (Recent) Spengleria vu y HeKOTOpbIX UcKONaembIx Gastrochaena vu Eufistulana (Monogpbie OpraHu3Mbl) HABOAYT Ha MbICNb,3ITO WTO ABNAeTCA HaCneACTBeEHHON OCOOCHHOCTbIO CBEPXCeMeNCTBa. Cpega B KAMeHUYCTOM rpyHTe 6bINa KONOHM30BaHa ABYyCTBOpKamnh (Bivalvia) B Tpex NOCNeAOBAaTeNbHbIX BBONIOWUMOHHbIX cha3sax: 1) BO-NepBbIX, 3TO Obinu lithophagids u gastrochaenids (TpvacoBpin (?) uv OpCcKuH Nepuvosb!); 2) BO-BTOPbIX, 3T0 6bINnn pholads u hiatellids (topcKuu” uv MeNOBON Nepvo sb); 3) uv B-TpeTbUX, 3TO 6bINu NpegAcTaBuTenu HECKONbKUX CEMeNCTB HE *®XVUBWUX B KAMEHMCTOM IpyHTe, KOTOpbie 3acenunu 3Ty CpeAy B pasnuunble BPeMeHa KaNHO3ONCKOrO Nepvoga. B NE€PBOM BBONIOUMOHHONM chase npeo6naganu xumunyecku nNpoTpaBnuBatowne oco6u, rnaBHbiIM O6pa30m, Tponuyeckoro wu cy6Tponuyeckoro cy6cTpata Kap6oHaTa KanbunA, TOrfa Kak BTOpaA cha3a cocToANa u3 MexaHnyecKku CcBepnaAuwux oco6en pasnu4nbix cy6cTpaToB BCex MMaBHbIX TEMNepaTyPHbIxX 30H. Pholads uv hiatellids He Cmornu pa3BuTb 6oratyto u pa3sHoobpa3Hyt0, *UBWY!0 B KA@MHAX, (ayHy B TPONNYeECKOM Kap6OHATHOM cyOcTpate — BepoarTHee BCerO, BCNeACTBUe 3aCeneHnA 3TONM CpeAbI rpynnamu lithophagids u gastrochaenids co CBeEPNUNbHbIMY MeXaHv3Mamu Nyyuwero KayecTBa. ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA ACKNOWLEDGMENTS I am grateful to the members of my dissertation committee, Drs. A. Lee McAlester, Willard D. Hartman, Donald C. Rhoads and Kar] M. Waage for their valuable assistance and advice. Ellen Crovo and Muriel Hunter aided this study in granting me access to their gastrochaenid material. For her tireless assistance in retyping many drafts of this paper, and especially for her patience and understanding, I am grateful to my wife, Judy. This study was conducted while the author was a Woodrow Wilson and a National Science Foundation Graduate Fellow. Parts of this study were supported by Doctoral Research Grant GB 36048 of the National Science Foundation and by supplementary grants from the society of Sigma Xi and the Colgate University Research Council. University of North Carolina Joseph Gaylord Carter Chapel Hill, North Carolina 27514 1. INTRODUCTION The Gastrochaenacea are a compact group of eulamellibranch bivalves found in tropical, subtropical, and warm-temperate waters throughout the world. Like the Pholadacea, the gastrochaenids are primarily borers of hard substrata, but they are taxonomically much less diverse [Recent faunas comprise about 15 species, according to Boss (1971)] and they are restricted in their boring habit to calcium carbonate substrata. Along with the lithophagid borers, gastrochaenids form a major but commonly over- looked element of tropical endolithic bivalve faunas. The biology of several gastrochaenid species is known from the studies by Deshayes (1846), Pelseneer (1911), Lamy (1923, 1925), Atkins (1937) Purchon (1954, 1958), Duval (1963), and Dinamani (1967). Notes on gastrochaenid ecology and biology are provided by Sluiter (1890), Kuhnelt (1930, 1934), Otter (1937), Robertson (1963), Gohar and Soliman (1963c), and Soliman (1973). In spite of this abundant literature, little is known about the evolution of this group or the comparative ecology of its species, and the anatomy and ecology of the Western Atlantic gas- trochaenids are largely unknown. The only recent study of evolution within the Gastrochaenacea is that by Boss (1967) on the evolution of the genus Spengleria. At Soldier Key, Florida, populations of three gastrochaenids are found boring into skeletons of the coral Diploria. The occurrence of these closely related species in the same substrata provided a unique and rare opportunity for studying their comparative ecology under virtually identi- cal environmental conditions. The following study of Spengleria rostrata (Spengler) 1793, Gastrochaena (Gastrochaena) hians (Gmelin) 1791, and Gas- trochaena (Rocellaria) ovata Sowerby 1834 is based primarily on Soldier Key populations, supplemented by observations of living specimens from Dis- covery Bay, Jamaica, and Castle Harbor, Bermuda. 2. MATERIALS AND METHODS At Soldier Key, Florida, skeletons of the coral Diploria were collected and systematically dissected to record their macroscopic epilithic and en- dolithic fauna. Living gastrochaenids from Discovery Bay, Jamaica, and Castle Harbor, Bermuda, were observed in running water tanks both in partially opened natural burrows and in artificial Plexiglas burrows. Specimens were examined by dissection and histological thin-section. The thin-sections were stained in haematoxylin-eosin or Alcian Blue. Shells of all three species were mounted in Epon 815 epoxy resin (Miller/ Stephenson Chemical Co., Danbury, Connecticut) and sections through the shells were studied in thin-section and in acetate peels. Internal growth band counts were taken directly on photographs of acetate peels at a magnification of slightly over 400 diameters. As relative ages are important to this study, the procedure for determining relative age will be described in detail. In order to estimate age from internal growth banding it is common practice to count light and dark band pairs in a radial section from the umbo to the shell margin (see Pannella and MacClintock, 1968, among others). Unfortunately, direct internal growth band counts cannot be made for most gastrochaenids because their outer prismatic shell layer is irregularly developed in patches, or is largely abraded, and because the crossed lamellar microstructure of their inner shell layers obscures internal growth banding. Furthermore, estimates of relative age from “annual” depositional breaks were found to be inaccu- rate and unreliable because these are poorly expressed on gastrochaenid shell exteriors and because interruptions in shell deposition result from nonperiodic (e.g., traumatic) as well as seasonal causes. Although continuous, internal growth band counts were not obtaina- ble, the total number of internal growth bands for each specimen could nevertheless be estimated by integrating empirically determined relation- ships between the number of comarginal ridges counted on the shell exterior and numbers of internal growth bands comprising successive comarginal ridges. Except for periostracal features, comarginal ridges comprise the major concentric ornament in most or all gastrochaenid shells (Fig. 10). A single comarginal ridge consists of numerous succes- sively deposited internal growth bands, and it represents a major episode of shell secretion. For all three species the comarginal ridges along an umbonal-posterior radial section were counted by examination under light microscopy, and the number of internal growth bands per comargi- nal shell ridge was determined for a few widely spaced ridges where the outer prismatic shell layer is well enough preserved to permit continuous counts. These data, plotted in Figure 1, show that the number of internal growth bands per comarginal ridge increases with age. This age effect is especially pronounced in S. rostrata. Over the age intervals examined, the number of internal growth bands per ridge (y) is approximately a linear 9 10 PEABODY MUSEUM BULLETIN 41 100 @ o ae) @ 8 & o a. ” G50 (= is] fo} = A > 25 Q Le2) ° te) 100 200 300 400 comarginal ridges Fic. 1. Relationship between internal growth bands per comarginal shell ridge and number of comarginal shell ridges in a radial section. Key: Spengleria rostrata (dots). Gastrochaena (Rocellaria) ovata (squares). Gastrochaena (Gastrochaena) hians (triangles). Data are from single juvenile and single adult specimens for each species. function of the total number of comarginal shell ridges (x). These relation- ships vary between the species and are described by the equations: Spengleria rostrata: y = 0.50% +280 Gastrochaena (Rocellaria) ovata: y = 0.06% + 9.0 Gastrochaena (Gastrochaena) hians:; y = 0.07x + 11.0 On the basis of these equations, the total number of internal growth bands along the umbonal-posterior radial section can be estimated by integrating the appropriate function over the interval (x=o) to (x=total number of comarginal ridges). For example, for individuals of the three species with 100 comarginal ridges in the radial section, the total number of internal growth bands is computed as: Spengleria rostrata: 100 [ (0.56% + 21.0)dx Nes 0 [0.28x* + 21.0x + Cl] = 4900 internal growth bands Gastrochaena (Rocellaria) ovata: 100 iN, J (0.06x + 9.0)dx [0:03x7? + 9.0% + Clo 0 1200 internal growth bands Gastrochaena (Gastrochaena) hians: 100 f (0.07% + 11.0)dx 100 ( [(0.035x? + 11.0x + C] = 1450 internal growth bands II ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 1] Although this method of relative age determination may be less accu- rate than continuous internal growth band counts, it is a better indication of relative age than counts of comarginal shell ridges alone would provide. As is apparent in Figure 1, reliance upon comarginal shell ridges alone would have grossly underestimated the age of S. rostrata relative to the other two species. To minimize the possibility of aberrant age determina- tions resulting from traumatic (i.e., nonperiodic) interruptions in shell secretion, shells showing unusual interruptions in the pattern of forma- tion of comarginal ridges (about 5 percent of the collected populations) were excluded from age-related analyses. The similarity in average inter- nal growth band width for S. rostrata, G. (R.) ovata and G. (G.) hians (2:30; 2.47, and 2.59 um., respectively) suggests that the time period of internal growth band formation is identical for these species. G. Pannella (personal communication) found that in the pholad Penitella a single pair of light and dark internal growth bands generally represents two weeks of shell deposition. As calculated using this time-conversion factor, the modes of age frequency distribution for the live gastrochaenids collected at Soldier Key (Fig. 2) are separated by an average of 1.06 years, and the age ranges of the populations are 2-5 years, 3-13 years, and 5-18 years for S. rostrata, G. (R.) ovata and G. (G.) hans, respectively. frequency of individuals ° 2,500 7,500 10,000 5,000 age in growth bands Fic. 2. Age-frequency distribution for combined populations of Spengleria rostrata, Gastrochaena (Rocellaria) ovata and Gastrochaena (Gastrochaena) hians collected during March 1970, at Soldier Key, Florida. Apparent abrasive efficiency of the shell was computed from data of burrow length and amount of anterior shell abrasion. Burrow lengths used in these calculations exclude extensions of the burrow above the coral substratum (e.g., burrow elongations resulting from simple posterior ex- tension of the burrow lining) and posterior burrow elongation resulting from attempted coral or sponge overgrowth of the siphons. Anterior shell abrasion was measured by comparing the number of comarginal shell ridges truncated in the shell anterior relative to the total number of comarginal ridges present in the unabraded shell posterior. The original width of abraded anterior shell ridges was estimated for each species as the average width of newly secreted and yet unabraded ridges. Apparent abrasive efficiency of the shell is expressed in micrometers of coral ab- 12 PEABODY MUSEUM BULLETIN 41 raded per micrometer of anterior shell abraded. The use of burrow volume abraded rather than burrow length would have been more ap- propriate, but the volume of shell abraded is especially difficult to esti- mate, and the volume of coral abraded cannot be readily calculated be- cause gastrochaenids typically line their burrows with aragonitic laminae. However, the use of shell volume abraded rather than shell width abraded would not change the trend of the data because shells of all three species are comparable in thickness anteriorly. In S. rostrata the burrow is wider relative to the width and height of the valves than in the other two species. Therefore calculations of apparent abrasive efficiency based on burrow volume would only further accent the already comparatively high values plotted for this species in Figure 13. The opening moment of the ligament was determined by orienting shells with the commissure in a horizontal position, then placing weights on the uppermost (left) valve until complete valve closure was brought about. Since the weight of a single valve is almost insignificant compared to the opening momentat closure, these opening moment values are reason- able estimates of the opening moment that would have been obtained by correcting for the weight of the upper valve. Opening moments were measured on adult specimens that had been preserved in 50 percent ethanol for approximately one week. Insofar as the ethanol solution may have altered the mechanical properties of the ligaments, these data are significant only for suggesting relative ligament strengths among the specimens at hand. Intestinal contents from immediately anterior to the anus were spread on a glass slide in tapwater, smeared with very light pressure, and examined by normal and polarizing light microscopy. Measurements of maximum particle diameter were made directly from photomicrographs of the gut contents enlarged to a final magnification of 550 diameters. 3. DISTRIBUTION AND ECOLOGY OF THE TROPICAL WESTERN ATLANTIC GASTROCHAENIDS The Gastrochaenidae comprise a relatively nondiverse but persistent component of endolithic faunas in corals and shells throughout the tropi- cal and subtropical Western Atlantic. Spengleria rostrata (Spengler) 1793 and Gastrochaena (G.) hians (Gmelin) 1791, the widest ranging species, are found throughout the Caribbean and in Bermuda. G. (G.) hians dominates the Florida Keys gastrochaenid fauna in terms of population density, whereas S. rostrata is generally rare here and elsewhere in the Western Atlantic. The remaining gastrochaenids show more limited geographic distributions. Gastrochaena (Rocellaria) ovata Sowerby 1834 is rare to mod- erately common throughout the tropical Western Atlantic, but is absent from Bermuda. At Bermuda this species is apparently replaced by the similarly short-siphoned Gastrochaena (G.) mowbrayi Davis 1903." At Castle Harbor, Bermuda, G. (G.) mowbrayi attains high population densities equaling or exceeding those of G. (G.) hians. Three other short-siphoned species are G. (G.) stimpsonii (Tryon 1862) from North and South Carolina and two similar species, one from the Gulf of Mexico [see “Rocellaria hans,” p. 218, in Andrews (1971)], and the other from Puerto Rico and the Bahamas. The following ecological notes refer to the three more common gastrochaenids of the Florida Keys, i.e., S. rostrata, G. (G.) hans and G. (R.) ovata. These notes are based primarily on populations collected near Soldier Key, Florida, where all three species are found in Diploria coral skeletons. Larger Diploria skeletons at Soldier Key, Florida, harbor at least 19 bivalve species, eight of which are borers [Gastrochaena (Gastrochaena) hians, G. (Rocellaria) ovata, Spengleria rostrata, Lithophaga nigra, L. antillarum, Pet- ricola typica, P. lapicida, and Botula fusca], and two of which are semien- dolithic nestlers (Arca imbricata and Paramya subovata). The endolithic and epilithic bivalve assemblages show higher diversities in coral substrata largely unprotected by living coral polyps. A list of the more common Soldier Key endolithic and epilithic bivalves is presented in Appendix A (below). Larger Diploria skeletons commonly show dense infestation by borers on both their upper and lower surfaces. Many of these disc-shaped corals have lost their initial attachment to the hard bottom, probably ‘Davis’ plate 4, figure 21 for G. Gastrochaena mowbrayi Davis and his cotype for G. mowbrayi at the United States National Museum (USNM 109562) are different species. Davis apparently mistakenly illus- trated Spengleria rostrata in this figure, but it is clear from his text description and from the burrows illustrated in his plate 4, figure 22, that he was referring to the species represented by his cotype. Davis’ mowbray? is clearly specifically distinct from G. (R.) ovata because of its much shorter siphons and subdued concentric ornamentation. Davis’ plate 4, figure 20, described by Davis (1903, p. 128) asa juvenile G. ovata, is likewise a specimen of S. rostrata. | Bay 14 PEABODY MUSEUM BULLETIN 41 through disintegration by borers. Nevertheless, the concave undersur- faces of the Diploria skeletons provide considerable settlement area be- cause they remain largely elevated above the sediment-water interface. The mytilacean Lithophaga nigra, the most abundant endolithic bivalve, bores upper and lower surfaces of Diploria but shows a marked preference for the center undersurfaces. Among the gastrochaenids, Gastrochaena (G.) hians and Gastrochaena (Rocellaria) ovata bore the coral margins, and Spengleria rostrata preferentially bores the underside of the coral margins. The synecology of bivalves inhabiting Diploria substrata from Soldier Key, Florida, and Castle Harbor, Bermuda, is discussed in greater detail by Carter (1976). Gastrochaenids are unique among Western Atlantic endolithic bivalves because of their unusual specializations for survival in thin and rapidly eroded substrata. Their relatively long siphons effectively isolate their shells from substratum erosion, and the probing behavior of their anterior pedal organ (see below) enables them to guide their burrows away from neighboring borers and unstable coral surfaces. Gastrochaenids also show an unusual capacity for repairing even extensive damage to their burrows by forming new calcareous burrow walls. Specimens of Spengleria rostrata and Gastrochaena (G.) hians from Discovery Bay, Jamaica, were apparently able to avoid predation by forming new calcareous burrow walls after nearly half of their burrow shell chamber had been naturally broken away. Natural burrow reconstruction has been described for Gas- trochaena (Rocellaria) laevigata from the Red Sea (Bertram 1936), and it is likely that most or all Recent representatives of the Gastrochaenacea are capable of burrow repair. Although a similar capacity for repair occurs in some more specialized species of Lithophaga, this does not occur in L. antillarum or L. nmgra. Not surprisingly, there is a general correlation between capacity for burrow repair and the secretion of calcareous burrow linings (as opposed to paste-type detrital linings) among species of Lithophaga. In addition to permitting burrow repair, the ability to secrete calcareous burrow linings may be adaptive for filling in the posterior of the burrow around the siphons. Because these linings are commonly thick in gastrochaenid burrows, they probably minimize the weakening effect of the borings themselves on the coral substratum. Schroeder (1972) has discussed calcareous burrow linings in this context, i.e., in terms of rein- forcing the coral substratum. This would clearly be adaptive in thinner and therefore more readily broken coral margins where the gas- trochaenids preferentially settle. In addition to their adaptations for life in rapidly eroded and broken coral margins, gastrochaenids show unique specializations for avoiding both coral overgrowth of the siphons and gastropod predation. Like Petricola typica but unlike the lithophagids, gastrochaenids can postpone or escape coral overgrowth by elongating their siphonal burrow toward the posterior. In G. (G.) hzans the siphons are especially extensible and retract- able, thereby enabling this species to survive extreme conditions of coral overgrowth or erosion (see Fig. 44). On the other hand G. (G.) hans differs from S. rostrata and G. (R.) ovata in lacking effective protection from predation by gastropods. The siphonal burrow aperture is sufficiently wide in G. (G.) hians to allow predation by certain naticid gastropods, judging from their characteristic borings in the posterior of some G. (G.) ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 15 hians shells. The small siphon diameter in G. (R.) ovata and the siphonal separation and projecting burrow linings (“baffles”) in S. rostrata are appa- rently effective in excluding gastropod predation, at least by naticids and muricids. A similar defense against gastropod predation is not observed in the Soldier Key lithophagids. But an alternative defense may have evolved in other species of Lithophaga (e.g., in L. bisulcata) where regular posterior encrustations on the shells may constitute effective barriers to gastropod boring. 4. ANATOMY Although the literature contains numerous references to general gas- trochaenid anatomy (e.g., Tryon 1882 and Lamy 1925) and a few refer- ences to their particular organ systems (e.g., Atkins 1937, Dinamani 1967, Duval 1963) there are few detailed studies of individual species. Notable exceptions include the work by Deshayes (1846) on Gastrochaena dubia; Fischer (1866) on “Fistulana” [=Eufistulana] grandis; Pelseneer (1911) on Gastrochaena machrochisma, G. dubia and Spengleria mytiloides; and more recently Purchon (1954) on “Rocellaria” [=Gastrochaena (G.)] cuneiformis, and Soliman (1973) on “Rocellaria” [=Spengleria] retzi. The present section provides comparative anatomical data for the Western Atlantic Spengleria rostrata, Gastrochaena (G.) hians, and Gastrochaena (Rocellaria) ovata. ManTLe. The mantle lobes of S. rostrata, G. (R.) ovata and G. (G.) hians are fused ventrally except at the pedal aperture (Fig. 3). The ventral mantle of all three species is muscular and highly contractile, but that of S. rostrata appears exceptionally thick and glandular. In these species the anteroven- tral mantle is normally expanded well beyond the shell margins, coming in contact with the adjacent anteroventral walls of the burrow shell chamber. In G. (R.) ovata and G. (G.) hans the pedal aperture is narrow, while in S. rostrata this is commonly greatly expanded relative to the diameter of the foot. In all three species the periostracum is initiated in a mantle groove near the periphery of the shell (Fig. 3C,D). This position of the periostracal groove requires that the mantle is fused by at most its inner and middle lobes. In G. (R.) ovata the periostracal groove is close to the shell margin, and the exposed ventral mantle appears otherwise smooth and featureless. In S. rostrata the periostracal groove is likewise situated close to the shell margin, but it is bordered by an inconspicuous mantle lobe (between the periostracal groove and the pedal aperture) that shows an apical longitud- inal groove. In G. (G.) hians the periostracal groove lies farther from the shell margin and it is bordered, as in S. rostrata, by a grooved mantle lobe. In G. (G.) hians this mantle lobe forms a prominent projecting ridge (Fig. 3C). G. (R.) ovata alone shows left and right elongate white glandular areas near the periostracal groove. The position of these glands (Fig. 3G) cor- responds to the zones of luminescence described for “Rocellaria” [=Gas- trochaena (Rocellaria) | grandis by Haneda (1939), but these glands in G. (R.) ovata were not observed to luminesce. Similar glands were described for “Rocellarta” [=Gastrochaena (Gastrochaena)] cuneiformis by Purchon (1954) who presumed that they secrete mucus for binding coral debris. In the Soldier Key gastrochaenids the lateral mantle within the mantle cavity is glandular and appears superficially irregularly folded. Deshayes (1846) and Clark (¢n Forbes and Hanley 1853, p. 135) described similar glands for Gastrochaena dubia and G. modiolina, respectively. Deshayes 16 ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 17 noted that these glands in G. dubia are divided into anterior and posterior masses, and he presumed that the anterior masses secrete an acid for boring. In the Soldier Key species these glands form continuous lateral masses lining the lateral mantle within the pallial line. In G. (R.) ovata and G. (G.) hians the glands appear coarsely rugose, whereas in S. rostrata they are more roundly inflated. Fic. 3. Ventral views of Spengleria rostrata (left), Gastrochaena (Rocellaria) ovata (middle) and Gas- trochaena (Gastrochaena) hians (right). The photographs of S. rostrata and G. (R.) ovata are of living specimens in artificial Plexiglas burrows; these two photographs are outlined in black to accent the outer edge of the shells. The view of G. (G.) hians is of a specimen preserved in alcohol; so its mantle and pedal aperture are greatly contracted. Legend: Sole of the foot in Spengleria rostrata. Expanded pedal aperture in S. rostrata and G. (R.) ovata; contracted pedal aperture in G. (G.) hians. Grooved mantle lobe in G. (G.) hians. Periostracum between the grooved mantle lobe and the shell margin in G. (G.) hians. Posterior of the inhalant siphon in S. rostrata, fully retracted. String of carmine particles visible through the semitransparent ventral mantle of G. (R.) ovata. The particles have been bound in mucus and are being transported to the base of the inhalant siphon for ejection. G. Elongate white glandular area in G. (R.) ovata. ol sige Ge ae G. (R.) ovata, G. (G.) hians and especially S. rostrata possess a muscular ventral mantle, and in specimens removed from their burrows, this mantle has been observed to pump water actively through the mantle cavity and out the inhalant siphon. In the relaxed condition [see S. rostrata and G. (R.) ovata in Fig. 3] the anteroventral mantle is held agape around the base of the foot and a steady inhalant stream of water passes through this opening from the burrow into the mantle cavity. When carmine particles are introduced into this anterior inhalant stream, they are quickly bound into mucous strings and passed posteriorly along two ventrolateral mantle grooves to the base of the inhalant siphon (Fig. 3F). The mucous strings are then expelled through the inhalant siphon by contraction of the ventral mantle while the pedal aperture is closed around the stock of the 18 PEABODY MUSEUM BULLETIN 41 foot. Mantle contraction is immediately followed by an apparent “gulping” of water through the dilating pedal aperture as the mantle cavity expands to its normal position, to be followed by another series of contraction and “gulping.” In G. (R.) ovata and G. (G.) hians this flushing cycle is repeated several times in rapid succession in the process of purging the inhalant siphon. In S. rostrata the mantle contractions are rarely repeated in quick succession, but the valves of this species close slightly during mantle contraction to assist the expulsion of pseudofeces. StPHONS. Unlike many Myidae, Hiatellidae and Pholadidae, gas- trochaenids do not protect their long siphons with a periostracal sheath. The inhalant and exhalant siphons of G. (R.) ovata and G. (G.) hians are externally fused into a single tube over most or all of their length (Fig. 4). Similarly fused siphons have been described for G. dubia and G. macro- chisma by Pelseneer (1911) and for “Rocellaria” [=Gastrochaena (Rocellaria) | ruppelli by Soliman (1973). In contrast, the inhalant and exhalant siphons in S. rostrata are completely separated. As this species matures, it spreads apart its siphons within the substratum through chemical erosion and accompanying deposition of aragonite. Divergence of the siphons in “Rocellaria” [=Spengleria] retzu has been attributed to coral overgrowth despite the observed occurrence of this species only in dead coral (Soliman 1973). All individuals of S. rostrata collected from Florida, Bermuda and Jamaica for the present study were found boring into coral skeletons or limestones 1n which living coral was not observed surrounding the inhalant and exhalant siphonal apertures. In all instances the siphons had diverged from one another within the hard substratum, forming an angle of 40 to 90 degrees. As the siphonal epithelium of S. rostrata erodes into the substratum on one side of the siphon, it secretes layers of prismatic arago- nite to fill in the previous burrow on the opposite side. Sections through the posterior of the burrow show that siphonal boring truncates the calcium carbonate linings of the burrow shell chamber as the animal penetrates deeper into the coral. The siphons of all three gastrochaenids are pale yellow to cream- colored externally but are brown to black internally. The extent of the internal pigmentation is proportional to the length of the siphons, being restricted to the posterior of the siphons inG. (R.) ovata, extending farther anteriorly in S. rostrata, and covering most of the interior of the siphons in G. (G.) hans. Some individuals of G. (R.) ovata are flecked with white on the interior of their siphons and on the siphon tentacles. In some S. rostrata and in many G. (R.) ovata siphonal pigmentation is lacking altogether, especially among the younger individuals. In S. rostrata and G. (R.) ovata the posterior apertures of the inhalant and exhalant siphons are guarded by contractile annular siphonal mem- branes. In G. (G.) hians only the exhalant siphon bears an annular mem- brane, and this is extremely thin, transparent and very mobile (Fig. 4). The siphons ofS. rostrata are fringed by about 250 minute tentacles arranged in four or five rows. In G. (R.) ovata the relatively large major siphonal tentacles are surrounded by an irregular row of minor ones, and the two annular membranes are also fringed by a delicate row of small tentacles. ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 19 el 1mm. S. rostrata G.(R.) ovata G.(G.) hians Fic. 4. Morphology of the siphon tips in Spengleria rostrata, Gastrochaena (Rocellaria) ovata and Gas- trochaena (Gastrochaena) hians. The directions of the inhalant and exhalant currents are indicated by the lower and upper arrows, respectively. The siphonal tentacles are drawn as they appear in the contracted state. The siphon tips of G. (G.) hians are much simpler, showing only a single row of intermediate size tentacles. As discussed below, the size and number of the major siphonal tentacles correlates with the degree of size sorting of ingested particles found in the intestines of these species. The interior of the long siphon tube of G. (G.) hians differs from the others in showing dorsal and ventral longitudinal grooves, expressed on the exterior of the siphons by a pair of dorsal and ventral ridges. In some specimens these features are also indicated by a pair of longitudinal grooves impressed upon the calcareous siphonal burrow lining. Similar longitudinal grooves were observed by Purchon (1954) for “Rocellaria” [=Gastrochaena (G.)] cuneiformis. In G. (R.) ovata the inhalant and exhalant siphons are separated internally for most of their length by a horizontal pallial septum. In this species and in S. rostrata (where the siphons form separated tubes) the posterior of the ctenidia are free and do not extend past the posterior shell margin. In contrast, the ctenidia of G. (G.) hians extend far past the posterior of the shell and are attached to the pallial siphonal septum, forming a long delicate partition between the inhalant and exhalant chan- nels (Fig. 5). 90 PEABODY MUSEUM BULLETIN 41 Fic. 5. Diagrammatic section through the ctenidium of Gastrochaena (Gastrochaena) hians: transverse section through the middle of the siphon. The directions of the inhalant and exhalant currents are indicated by the lower and upper arrows, respectively. Legend: O. Outer demibranch. I. Inner demibranch. S. Ctenidial septum. The inhalant siphon of S. rostrata and G. (R.) ovata differs from G. (G.) hians in having a transverse valve across its base. Similar valves have been described for S. mytiloides by Pelseneer (1911) and for “Rocellaria” [=Gastrochaena (G.)] cuneiformis by Purchon (1954). In S. rostrata the larger part of the valve is a crescent-shaped partition suspended ventrally from the pallial septum separating the inhalant and exhalant siphon tubes. This is accompanied by two smaller flaps projecting from ven- trolateral positions on the wall of the inhalant siphon. When contracted, these valves restrict access between the inhalant siphon and the mantle cavity. The valve in G. (R.) ovata appears more delicate and consists of a single membranous flap suspended ventrally from the siphonal septum separating the inhalant and exhalant siphon tubes. Contraction of this process blocks only one-third of the aperture at the base of the inhalant siphon. A siphonal valve does not occur in G. (G.) hians, undoubtedly because its inhalant and exhalant siphons are separated anteriorly by a ctenidial rather than a pallial septum (Fig. 5). According to Haas (1935) these siphonal processes function as check-valves to arrest the flow of water through the inhalant siphon. These valves might assist purging pseudofeces, because pseudofeces collect near the base of the inhalant ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 21 siphon. Partial blocking of the inhalant siphon would increase the veloc- ity of the purging water at this point. Siphonal valves are also known in certain Mytilacea and Mactracea (Yonge 1948, 1955). In all three species the siphons show great sensitivity to changes in light intensity. Shadows passing over the tips of the siphons induce partial adduction of the valves and retraction of the siphons away from the burrow posterior. The siphons are remarkably contractile, and even the long siphoned G. (G.) hians is capable of retracting these entirely within the shell. It is noteworthy that these species can apparently fully retract their siphons only after prolonged irritation. Sudden, intense irritation of the siphons in G. (G.) hians causes premature adduction of the shell and damage to the siphons by being pinched between the posterior shell valves. This behavior suggests that complete siphonal retraction may require a preparatory interval during which the siphonal hemocoels are partially deflated by draining their fluid to other parts of the mantle. Unlike S. rostrata and G. (R.) ovata, complete retraction of the siphons in G. (G.) hians is assisted by introversion of their bases and “tucking” the siphon into the anterior mantle cavity. CTENIDIA AND LaBIAL Patps. The eulamellibranch ctenidia of S. rostrata are distinctly plicate and thick, while those of G. (R.) ovata and G. (G.) hians are flat and considerably more delicate. As in many bivalves, the outer demibranch shortens anteriorly relative to the inner one, and only the inner demibranch is fused to a distal oral groove. These observations corroborate Pelseneer’s (1911) generalization that flat gills associate with fused siphons, whereas plicate gills associate with separated siphons in the Gastrochaenacea. Posteriorly in S. rostrata and G. (R.) ovata the outer demibranch is shorter than the inner one. In G. (G.) hians the outer and inner demibranchs become equal in length toward the posterior (Fig. 5). The ctenidia of G. (G.) hians are unique in extending far posteriorly into the siphonal tube, but the outer demibranch in S. rostrata is unique in forming a prominent supraaxial extension, resembling that described for “Rocellaria” [=Gastrochaena (G.) | cuneiformis by Purchon (1954). In all three species the ventral tips of the anterior filaments of the inner demibranch are inserted into and fused to a distal oral groove between the elongated labial palps [ctenidium-palp association type “two” of Stasek (1963) ]. The distal oral groove is especially long in S. rostrata and the labial palps are larger here than in the other two species. Foor. The foot of the Soldier Key gastrochaenids consists of a circular to slightly oval pedal disc with an anteriorly projecting pedal organ, a longitudinal byssal groove, and a posterior byssal gland and byssus cavity (Figs. 6-8). The anterior of the foot shows a small pedal gland located between the pedal organ and the byssal groove. These pedal apparatus- ses are similar to those described for G. dubia, G. macrochisma, and S. mytiloides (Pelseneer 1911), “Rocellaria” [=Gastrochaena (G.) | cuneiformis (Purchon 1954) and “Rocellaria” [=Spengleria | Gastrochaena retzi (Soliman 1973). Drawings of the foot in the tube-dwelling “Fistulana” [=Eufistulana | grandis do not show a well-formed pedal disc, but the presence of this in a 22 PEABODY MUSEUM BULLETIN 41 rudimentary state may be inferred from the pedal “scar” in the anterior of its tube (see plate 12, figure 3, in Fischer 1866). In S. rostrata the pedal disc is relatively wide, and in this species and G. (R.) ovata the contracted anterior pedal organ is cone-shaped (Figs. 6 and 7). Compared with S. rostrata the pedal discs of G. (R.) ovata and G. (G.) hians are smaller in diameter. The contracted pedal organ of G. (G.) hians appears relatively flattened and spatula-like (Fig. 8). S. rostrata is unique in that the anterior of its foot shows 5 or 6 chevron-shaped glandular corrugations im- mediately dorsal to the pedal organ, with the apex of each chevron point- ing ventrally toward the sole of the foot. The Gastrochaenacea have been characterized as losing their juvenile byssal attachment to adhesive or suctorial attachment in the adult stage (Pelseneer 1911, Otter 1937, Soliman 1973). On the other hand, Yonge (1963) noted that the byssus gland persists in adult gastrochaenids and that byssus threads are occasionally secreted. Among the Soldier Key species the foot is byssally attached in juveniles and adults alike, although to varying degrees. In all three species the byssal attachment is minute and is hidden from view by the pedal disc. Adults of G. (R.) ovata are firmly byssally attached and their byssus fibers occasionally pull away pieces of the burrow wall when these animals are removed from their burrows. The byssus in G. (G.) hians and especially in S. rostrata is more delicate, and in S. rostrata pedal attachment is apparently supplemented by adhesion by the surface of the pedal disc. In S. rostrata the burrow lining may be marked by a pedal “scar” showing the outline of the pedal disc, the byssal groove, and occasionally also the site of byssal attachment. Pedal scars are less common but are occasionally observed in burrows of G. (R.) ovata and G. (G.) hians. Other Recent species of Gastrochaena and Spengleria (Gohar and Soliman 1963c, Soliman 1973), “Fistulana” [=Euftstulana] (Fischer 1866) and cer- tain fossil species of Gastrochaena show pedal scars 1n the anterior of their burrows. These scars are potentially useful sources of information for reconstructing the pedal structure of fossil gastrochaenids. MuscuLaTvrgE. S. rostrata, G. (R.) ovata and G. (G.) hians are heteromyarian, with the anterior adductor muscles relatively small and attaching near the extreme anterior shell margin (Figs. 6-8 and 14-16). Reduction of the anterior adductor muscle is proportional to the degree of anterior shell reduction. Thus, the ratio of posterior to anterior adductor cross-sectional area 1s lowest in S. rostrata (averaging 5.0), slightly greater in G. (R.) ovata (averaging 6.7), and is greatest in the very inequilateral G. (G.) hians (averaging 12.1). The pedal musculature of all three species consists of anterior and posterior pedal retractors and anterior pedal protractors. S. rostrata also possesses pedal elevators (Figs. 6-8). In S. rostrata the pedal protractors each consist of a bifurcating muscle attaching ventral and slightly lateral to the anterior adductor. In contrast, the protractors in G. (G.) hians consist of single muscles that attach dorsal and posterior to the anterior adductor. In S. rostrata the pedal retractors are diffuse and pass peripheral to the visceral mass as they attach into the food. In G. (R.) ovata and G. (G.) hians the more concentrated retractor muscles pass directly through the visceral mass and they attach more exclusively to the byssus apparatus. ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 23 POSTERIOR PEDAL RETRACTOR ANTERIOR ADDUCTOR poUaL ETEVATOR | ANTERIOR PEDAL RETRACTOR POSTERIOR ADOUCTOR ANTERIOR PEDAL PROTRACTOR S. ROSTRATA PEDAL 6 ORGAN BYSSAL GROOVE BYSSUS CAVITY POSTERIOR PEDAL RETRACTOR ANTERIOR PEDAL RETRACTOR POSTERIOR ADDUCTOR ANTERIOR ADDUCTOR G. (R.) OVATA r/ PROTRACTOR BYSSAL GROOVE BySSUS CAVITY POSTERIOR PEDAL RETRACTOR ANTERIOR PEDAL PROTRACTOR ANTERIOR PEDAL RETRACTOR POSTERIOR ADDUCTOR TERIOR ADDUCTOR G. (G.) HIANS PEDAL : 8 ORGAN BYSSAL GROOVE BYSSUS CAVITY Fics. 6-8. The major pedal and shell adductor musculature in the tropical Western Atlantic gas- trochaenids. The epithelium is shown intact near the base of the foot to show the position of the anterior pedal organ and the longitudinal byssal groove. The anterior adductor muscle in Spengleria rostrata (Fig. 6) is shown lifted out of its normal position immediately dorsal to the anterior pedal protractors (see dashes). 24 PEABODY MUSEUM BULLETIN 41 G. (G.) hians differs from the other species in the attachment of its anterior pedal retractors to a pair of calcareous projections (myophores) extending laterally and posteriorly from the dorsal shell margins. In most G. (G.) hians the myophore consists of a simple triangular plate beneath the hinge line; but in a few individuals this plate may extend slightly above the dorsal shell margin, or it may bifurcate toward the posterior. The left myophoral plate is generally larger than the right by an average surface area ratio of 1.43/1. Distinct triangular myophores are not developed inS. rostrata or G. (R.) ovata, but in the latter species the anterior pedal retractors may attach to a pair of small, irregular knobs similar in position to the myophores of G. (G.) hians. In all three species the pallial musculature consists of well-defined muscular bundles attaching to the shell in irregular patches, thereby forming a discontinuous pallial line (Figs. 14-16). In G. (G.) hians the posteroventral pallial muscles form a distinct accessory adductor muscle. A similar accessory adductor is not found in S. rostrata or in G. (R.) ovata. INTESTINAL CONTENTS. Fecal pellets dissected from the intestines of S. rostrata, G. (R.) ovata and G. (G.) hians are shaped in plain rods (type “four” of Arakawa 1970). Intestinal contents dissected from adult specimens collected in March of 1970 consisted of calcium carbonate debris with traces of sponge spicules, diatoms and crustacean exoskeletons. Although intestinal contents for the three species are similar in average maximum particle diameter, size sorting appears to be slightly more efficient in S. rostrata than in the other species (Fig. 9). This better sorting correlates with See Gras Spengleria rostrata ___S~_ Gastrochaena (Gastrochaena) hians ____S Gastrochaena (Rocellaria) ovata frequency relative 10 5 pm. particle diameter Fic. 9. Size frequency of ingested particles removed from the intestine immediately anterior to the anus. Particle size was measured as the maximum length. The curves are based on size-frequency histograms for several hundred particle measurements for three individuals of each species. See materials and methods (Chapter 2) for sampling procedure. ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 25 ctenidial plication and the presence of smaller and more numerous major siphonal tentacles in S. rostrata (Fig. 4). The sampled gastrochaenids were collected from near the coral margins where their siphonal openings were elevated not more than 10 centimeters from the carbonate sand sub- stratum. Nevertheless, variation in the distance of their siphonal openings from the sand substratum may have biased the particle sorting data. The striking aspect of the intestinal contents is the dominance of calcareous debris and the paucity of organic debris and mineralized tests of planktonic organisms. This is surprising, considering that Smith et al. (1950) determined that diatoms were the most important component of the plankton in 1945 at Soldier Key, with dinoflagellates being less abun- dant. The zooplankton (mostly copepods, copepod nauplii, and tintinnids) rarely dominated the plankton at this locality. In the present samples, diatoms were rarely encountered in the intestinal contents, and in each sample the number of diatoms was exceeded by the number of sponge spicule fragments. If the March 1970 plankton at Soldier Key was not very different from what it was in 1945, then the present data suggest that most of the ingested particulate matter comes from resuspended carbonate - debris rather than from the plankton per se. Further study is needed to determine the source of primary nutrition of these bivalves, but it may likely come from the microbiota of resuspended sediment (e.g., bacteria) or from planktonic organisms lacking mineralized tests. D3. SHELES Unlike the common pholad borers, gastrochaenids lack conspicuous rasp- ing “teeth” on the shell anteriors, and their delicate shells do not appear obviously adapted for deep boring into hard substrata (Fig. 10). Their shells resemble the Pholadacea only in being thin, edentulous and broadly gaping anteriorly, and they differ from most pholads in lacking a perma- nent posterior shell gape and modification of the shells for rocking about a vertical axis. All three Soldier Key gastrochaenids show great variability in their proportion of shell height to length, but their shell growth is still best described as isometric (Fig. 11). As judged from shell length, the rate of growth inS. rostrata andG. (G.) hians is nearly constant over the age interval sampled, whereas shell growth rate in G. (R.) ovata slows markedly in older individuals (Fig. 12). G. (R.) ovata and G. (G.) hians are similar in having simple comarginal shell ridges that are greatly truncated anteriorly by abrasion. Every fourth or fifth comarginal ridge protrudes farther from the surface of the shell in G. (R). ovata, but the shell ridges show a more uniform height in G. (G.) hians. S. rostrata differs from the other species in having wider, more rounded comarginal ridges. More strikingly, these ridges are covered with minute aragonitic periostracal spikes (see shell microstructure, below). While all three gastrochaenids show considerable anteroventral abra- sion of the shell, the nature of this abrasion differs between S. rostrata and the other species. In S. rostrata abrasion reduces the anterior periostracal spikes to blunt stubs at the margin of the shell while only occasionally truncating a comarginal shell ridge. On the other hand, periostracal spikes are not present in G. (R.) ovata and G. (G.) hians, so abrasion often greatly reduces the height of their comarginal shell ridges. In these two species the latest-formed ridge initially projects prominently but is quickly partially or entirely abraded. Following this abrasion, a new comarginal ridge is se- creted, thereby starting a new cycle of abrasion and secretion. In all three species anterior shell abrasion is limited primarily to the surface of the latest formed comarginal ridge. Some G. (G.) hians and, less commonly, some G. (R.) ovata show evidence of additional minor abrasion on the umbones and on the lateral shell surfaces. Anterior truncation of the comarginal shell ridges is slightest in S. rostrata (50 to 70 percent of the ridges truncated) undoubtedly because these are protected by the calcified periostracal spikes. Anterior trunca- tion is much higher in the other two species, averaging 80 to 90 percent in G. (R.) ovata and over 90 percent in G. (G.) hians. The apparent abrasive efficiency of the shell (see Materials and Methods for definition) in S. rostrata is variable but generally much higher than in the other species, perhaps reflecting the varying importance of periostracal calcification or 26 ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 20 mm. Fic. 10. A. Spengleria rostrata (Spengler) 1793, YPM 9480, Discovery Bay, Jamaica. B. Gastrochaena (Rocellaria) ovata Sowerby 1834, YPM 9490, Discovery Bay, Jamaica. C. Gastrochaena (Gastrochaena) hians (Gmelin) 1791, YPM 9483, Discovery Bay, Jamaica. From top to bottom of figure: lateral exterior view, lateral interior view, dorsal view perpendicular to the hinge axis, ventral view perpendicular to the plane of the pedal gape, ventral view perpendicular to the hinge axis (showing hinge structure). 28 PEABODY MUSEUM BULLETIN 41 Shell height / length 5,000 10,000 age in growth bands Fic. 11. Variation of shell height/length ratio with age in Spengleria rostrata (dots), Gastrochaena (Rocellaria) ovata (squares), and Gastrochaena (Gastrochaena) hians (triangles). mm. shell length 7,500 10,000 5,000 age in growth bands Fic. 12. Variation of shell length with age in Spengleria rostrata (dots), Gastrochaena (Rocellaria) ovata (squares) and Gastrochaena (Gastrochaena) hians (triangles). chemical boring in protecting the ridges from abrasion (Fig. 13). The apparent abrasive efficiency of the shell in G. (R.) ovata and G. (G.) hians is more uniform throughout the sampled populations, and is only slightly higher in the former species. G. (R.) ovata may show a higher value than G. (G.) hians because its burrows are shallower; so its shell penetrates primar- ily into the outer, more friable layers of the coral skeleton. ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 29 ded 3 ky 8 um. coral bored /um. shell abr 3 °o 25 5 10 5 20 average burrowing rate (um. coral bored per growth band ) Fic. 13. Relationship between apparent abrasive efficiency (um. coral bored/um. shell abraded) and average burrowing rate in Spengleria rostrata (dots), Gastrochaena (Rocellaria) ovata (squares), and Gastrochaena (Gastrochaena) hians (triangles). 6. SHELL MICROSTRUCTURE Shells of all three species consist primarily of periostracum, composite prismatic, crossed lamellar (CL) and fine complex crossed lamellar (fine CCL) structures (from shell exterior to shell interior), with myostracal prismatic structure occurring at sites of muscle-shell attachment (Figs. 14-16). The periostracum is generally intact only in the posterior of S. rostrata, where this forms regular vertical corrugations. The periostracum is commonly 15.0, 10.0 and 7.5 ym. thick in the posterior of S. rostrata, G. (R.) ovata and G. (G.) hians, respectively, and this thins anteriorly in S. rostrata but thickens by about 20 percent in G. (R.) ovata and G. (G.) hans. The periostracum of S. rostrata is studded with aragonitic spikes arranged in concentric rows in the shell posterior (Fig. 17) and in concentric to oblique rows in the shell anterior (Figs. 18, 19). Calcified spikes also occur in parts of the periostracum extending the primary ligament anteriorly and posteriorly; here the spikes are irregularly arranged and more sparse than in other parts of the shell. The aragonitic spikes are clearly entirely contained within the periostracum posteriorly (Fig. 20), but toward the shell anterior the organic component of the periostracum thins and the spikes lie closer to the surface of the shell (Fig. 21). In the midlateral and anterior parts of the shell, the bases of the spikes are partially imbedded within the underlying composite prismatic shell layer, and they are strongly abraded along the anteroventral shell margins (Figs. 18, 19). The unabraded spikes vary considerably in average dimensions, those in the shell anterior being widest, and those in the dorsal periostracum being much narrower (Fig. 22). The spikes in any given fragment of perio- stracum may vary considerably in width and length (Fig. 23), but average spike width and length both increase regularly with age. In the shell posterior the average spike length increases geometrically with age (Fig. 24) whereas the average spike width/length ratio increases linearly with age at the time of secretion (Fig. 25). The secretion of individual aragonitic spikes is apparently initiated by the inner surface of the outer mantle fold near the periostracal groove, i.e., between the outer and middle mantle folds. A view of the outer surface of newly secreted periostracum between the outer and middle mantle folds (Fig. 26) shows a progressive increase in spike diameter away from the 30 ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 31 ~S. ROSTRATA 14 LIGAMENT SUPPORT — — 15 Fics. 14-16. Typical distributions of shell microstructure in the tropical Western Atlantic gas- trochaenids. Legend: stippling: myostracal prismatic structure. dashes (parallel to the shell margins): branching crossed lamellar structure. dashes (perpendicular to the shell margins): radiating crossed lamellar structure. triangles: triangular crossed lamellar structure. rectangles: blocky crossed lamellar structure. white: fine complex crossed lamellar structure. Although shown in white for the sake of simplicity, the shell exterior near the umbones may be composite prismatic, crossed lamellar, or complex crossed lamellar, depending on the depth of abrasion of the shell. The two exterior shell layers (periostracum and composite prismatic) do not appear in this view of the inner shell surface. The varieties of crossed lamellar and complex crossed lamellar structure are explained in detail by Carter (1976b). 32 PEABODY MUSEUM BULLETIN 41 Fic. 17. Surface of the posterior periostracum in Spengleria rostrata; scanning electron micrograph. Rodlike structure on the right is a sponge spicule. Fic. 18. Abraded tips of aragonitic periostracal spikes on the sides of the shell in the anterior of Spengleria rostrata. Note the alignment of the spikes in rows oblique to the larger comarginal shell ridges. Fic. 19. Aragonitic periostracal spikes on the anteroventral shell margin of Spengleria rostrata. Note the apical abrasion. Fic. 20. Acetate peel of a radial vertical section through the posterior periostracum of Spengleria rostrata. Note the thin outer layer of the organic periostracum pulled away from the tips of some of the aragonitic periostracal spikes. ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 33 S ey d < , a NS Ra anterior posterior Fic. 21. Diagrammatic representation of the anterior thinning of the organic periostracum along an anterior-posterior vertical section through the shell of Spengleria rostrata. Note the increased imbed- ding of the aragonitic periostracal spikes within the composite prismatic shell layer (¢) toward the anterior. Legend: a. periostracum with aragonitic spikes. b. acicular prismatic layer. c. composite prismatic layer. d. crossed lamellar and fine complex crossed lamellar layer. 54 PEABODY MUSEUM BULLETIN 41 Size and shape variation in spikes of S. rostrata s & 50 um. (spikes) 10 pm. (shell) : Fic. 22. Size and shape variation of periostracal spikes in Spengleria rostrata. The measurements indicated by the spike shapes represent averages of at least 60 measurements for each area of the _ periostracum circled in the diagram. Only unabraded spikes were measured. ym. spike width te) 10 ao 50 60 70 20 yy 30 pm. spike length Fic. 23. Variation in aragonitic periostracal spike shape and size within a small portion of perio- stracum removed from the posterior of the shell in Spengleria rostrata. (See area circled in the accompanying photograph.) periostracal groove and a simultaneous change in spike appearance from cloudy white to clear and refractive. The aragonitic spikes attain maximum size near the shell margin, just prior to reaching the sites of initiation of the underlying composite prisms. Individual aragonitic spikes appear crystallographically uniform under crossed-polarized light, with the crystallographic c-axis paralleling the spike length. Dissolution of isolated spikes in dilute HCI reveals an abundant water-insoluble organic matrix. The shape of aragonitic periostracal spikes may be constant or quite variable within a small fragment of periostracum. Several shape variations among thousands of spikes observed are illustrated in Figure 27. The most common shape (Fig. 27,a-7) isa simple cone with a hexagonal cross-section. Rarely, the spikes show distinct internal growth banding (Fig. 27, q-s) ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 35 190 1.80 > 1.70 1.60 um. spike length 109, 1.50 1.40 7500 ° : 2500 5000 E age (in growth bands) at time of secretion Fic. 24. Relationship between aragonitic periostracal spike length and age at time of secretion in Spengleria rostrata. Each data point represents an average of at least 60 length measurements for spikes isolated from within the posterior periostracum near the extreme posterior shell margin. The perio- stracum is sampled along a radial transect from the umbo to the posterior shell margin. Samples from the same specimen of S. rostrata are indicated by the same symbol (a dot, square, triangle, tilted square, upside-down triangle, or hexagon). indicative of periodic secretion on their bases. Apparent irregularities in secretion may also result in basal (Fig. 27, j-m, 0) or apical (/, m) bosses. Although the spikes are generally mutually isolated in the periostracum, they are rarely closely spaced in groups of two or three (Fig. 27, ) or they may appear as fused twins (0, p). Scanning electron microscopy of isolated spikes (Figs. 28, 29) shows an ultrastructure of elongate crystal laths usually 0.03 to 0.04 wm. thick and 0.45 to 1.06 wm. wide. Individual laths may be terminated by hexagonal crystal faces or they may show evidence of length-parallel acicular subunits about 0.03 wm. in diameter. This lathlike ultrastructure contrasts sharply with the polygonal prismatic ul- trastructure of aragonitic periostracal spikes described for the pandora- cean Laternula by Aller (1974). Spikelike periostracal calcification is rare or 36 PEABODY MUSEUM BULLETIN 41 0.45 0.40 x 3 °o = rea) @ Cc ® Se fe et 3 38 @ o & | a 2 A rt D Vv oO —_ © > oO 0.25 0.20 ‘ age (in growth bands) at time of secretion Fic. 25. Relationship between aragonitic periostracal spike width/length ratio and age at time of secretion in Spengleria rostrata. See explanation of Figure 24. absent in adults of Recent representatives of Gastrochaena. But the pres- ence of minute spikes on shells of the Upper Cretaceous Gastrochaena (Rocellaria) linsleyy (Appendix C) and Eufistulana ripleyana (Stephenson 1941) suggests that this is an ancestral feature of the superfamily. Aragoni- tic periostracal spikes have also been observed on the shell of a Pliocene Gastrochaena (Gastrochaena) from Florida (Fig. 30). In this fossil the spikes are restricted to the juvenile shell and were apparently not secreted by the adult. Spikelike periostracal calcification is retained in Recent gas- trochaenids only in the two surviving species of Spengleria and perhaps in the juvenile shell of Gastrochaena rugulosa Sowerby (USNM 184364; Pana- ma), Gastrochaena denticulata Deshayes, and certain Gastrochaena (Rocel- laria) (Carter 1976a). Most living bivalves lack distinct periostracal calcification of the kind present in S. rostrata; so this species is a striking exception to the rule. Aller (1974) described similar spike-shaped processes cemented to the shell ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 37 Fic. 26. Surface view of newly secreted periostracum in the posterior of Spengleria rostrata. The extreme posterior of the shell (out of focus) appears in the uppermost part of the photograph. The spike diameter increases from the bottom of the photograph (nearest the periostracal groove) toward the posterior shell margin. Fic. 27. Aragonitic periostracal spikes freed from within the posterior organic periostracum of Spengleria rostrata by dissolution of the organic matrix in NaOCl. Fic. 28. Aragonitic periostracal spike freed from within the posterior periostracum of Spengleria rostrata by dissolution in NaOCl. Scanning electron micrograph. Fic. 29. Aragonitic periostracal spike freed from within the posterior periostracum in Spengleria rostrata. Higher magnification of the area circumscribed by the parallelogram in Figure 28. Scanning electron micrograph. 38 PEABODY MUSEUM BULLETIN 41 Pe se eee Fic. 30a,b. Aragonitic spikes on the juvenile parts of the left valve of a Pliocene Gastrochaena (Gas- trochaena) from St. Petersburg, Florida (YPM 9597). Figure 30b is a higher magnification of the area circumscribed by the parallelogram in Fic. 30a. Note the absence of the spikes in the adult shell to the right (posterior) and bottom (ventral) in Figure 30a. exterior in the pandoracean Laternula. Although these calcified spikes are probably not entirely embedded within the periostracum in the shell posterior (Aller, personal communication, states that this possibility has yet to be explored), they resemble the spikes in S. rostrata in three respects. 1) In Laternula and S. rostrata at least some of the spikes are cemented to or ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 39 partially embedded within the outer prismatic shell layer and they are structurally distinct from this layer. 2) The aragonitic spikes are formed in a zone of mantle epithelium (the inner surface of the outer mantle fold) peripheral to the zone of initiation of the outer prismatic layer of the shell proper. In addition, the spikes are fully formed by the time the underlying prismatic shell layer is initiated. 3) Spike formation occurs simultaneously with the secretion of organic periostracum. Since the spikes in both Later- nula and Spengleria are formed simultaneously with organic periostracum by the inner surface of the outer mantle fold, they may be properly regarded as periostracal structures. The spikes are therefore genetically distinct from the underlying prismatic shell layer. By inference from the present data, certain granular processes on the exterior of other pandora- ceans (e.g., Thracia pubescens; see Taylor, 1973) and the spherical granules described for a poromyacean by Runnegar (1974) (see his plate 5, fig. 8) may also be regarded as calcified periostracal structures. The wide dis- tribution of spike- and granule-like periostracal calcification among rep- resentatives of the Mytilacea, Permophoridae, Myoida (all four super- families) and Anomalodesmata (all six superfamilies) suggests that these structures appeared early in the evolution of the Bivalvia. According to Carter and Aller (1975) spicule-like periostracal calcification may have constituted a primordial molluscan shell. This theory finds support in the occurrence of radial rows of minute granules cemented to the shell ex- terior in the Middle Cambrian monoplacophoran Latouchella penecyrano (fig. 1OA-12 in Runnegar and Jell, 1976), an early representative of the most primitive molluscan class. In addition to showing aragonitic periostracal spikes, S. rostrata is unique in having a sparsely developed aragonitic acicular prismatic layer in the shell posterior between the periostracum and the underlying com- posite prismatic layer (Fig. 21 and Carter, 1976a). This acicular layer, which characterizes S. mytiloides as well as S. rostrata, thins and disappears anteriorly where the periostracum comes to lie in contact with the compo- site prismatic shell layer (Fig. 21). The acicular prisms occur in fan-shaped aggregates that only partially fill the cavity beneath the perio- stracum. The orientation of individual acicular prisms appears random, but there is a tendency for prisms pointing toward the shell margin to be longer, suggesting growth in a concentration gradient. Individual acicular prisms show no evidence of an organic matrix upon dissolution, and they are morphologically identical to a common crystal form of inorganically precipitated aragonite. The underlying composite prismatic shell layer is strongly developed in the posterior S. rostrata but is very weakly developed in the anterior of this species and over the entire shell of G. (R.) ovata and G. (G.) hians. The composite prismatic shell layer consists of radial first order prisms that bifurcate toward the shell margin (Figs. 31, 32). Radial, vertical sections show that these first order prisms are not strictly horizontal but rather bend toward the inner shell surface. The individual first order composite prisms consist of smaller second order prisms radiating at a high angle from a longitudinal central prism axis toward the surface of deposition. The inner shell layers beneath the composite prismatic layer consist of crossed lamellar (CL) and fine complex crossed lamellar (fine CCL) struc- 40 PEABODY MUSEUM BULLETIN 41 ie. 31k Acetate peel of a fyerical transverse section through the WIRE Hipeve layer in the posterior of Spengleria rostrata. The shell exterior is toward the upper part of the photograph. The boundary between the porous acicular prismatic (above) and composite prismatic (middle) shell layers appears near the upper part of the photograph. The boundary between the composite prismatic and crossed lamellar shell layers appears in the lower part of the photograph. Fic. 32. Exterior surface view of the shell posterior in Spengleria rostrata showing several first order composite prisms radiating from the umbo (upper part of photograph) toward the shell posterior (lower part of photograph). The periostracum and acicular prismatic shell layer have been removed to expose the composite prisms. tures with myostracal prismatic structures developed at sites of shell- muscle attachment. As seen on inner shell surfaces, the CL structure is generally found exterior to the pallial line, while the fine CCL structure is found mostly interior to the pallial line except in the shell posterior, where it also occurs slightly exterior to the pallial line. Figures 14 to 16 show the distribution in the shell of several varieties of CL structure defined by Carter (1976b). The first order crossed lamels in G. (R.) ovata are mostly of the branching concentric variety (BCL), while those in G. (G.) hians are of the BCL variety anteriorly and of the BCL, radiating (RCL), and blocky (BICL) varieties posteriorly. In S. rostrata RCL and BCL occur together in the shell anterior, while triangular (TCL) lamels flank BCL lamels in the shell posterior. The CL and CCL structures observed in the present species are uniformly present among the scores of individuals analyzed, but the distribution of each structure on the depositional surface varies with the age of the individual. Significant deviations from the distribution patterns shown in Figures 14 to 16 occur in gerontic individuals. In gerontic specimens the BCL structure may cover much of the depositional surface, even within the pallial line. 7. LIGAMENT AND DENTITION The ligament in the Soldier Key gastrochaenids is external and inserts on ligament nymphs posterior to the umbones. The primary ligament is extended anteriorly and posteriorly only by the periostracum, which: unites the valves along their entire dorsal margin. The ligament nymphs are shorter in S. rostrata than in G. (G.) hians and G. (R.) ovata, and in all three species the ligament and its periostracal extensions preclude pholad- like rocking of the valves about a vertical axis. The line of the ligament is oblique to the longitudinal shell axis, with the angle between the ligament and shell axis being greatest in G. (G.) hians and least in S. rostrata. Because of this high angle, G. (G.) hians can open its shell valves widely along the posterior margin while hardly increasing the gape between the valves anteriorly. The ligament nymph inG. (R.) ovata is smaller than in the other species, and this difference is expressed in the relatively small opening moment of its ligament. The ligament opening momentat shell closure for preserved specimens averages about 25 grams in adult S. rostrata and G. (G.) hians but only about 6.5 grams in G. (R.) ovata. At the point of maximum shell gape allowed by the width of the burrow, the ligament retains 28 to 40 percent of this opening moment, averaging about 7, 10, and 2 grams in§S. rostrata, G. (G.) hans and G. (R.) ovata, respectively. These latter values represent the relative forces that would be applied to the lateral burrow walls by the shell margins during boring, assuming that no additional pressure is supplied by muscular and hydrostatic forces. The Gastrochaenacea are commonly described as edentulous (Pur- chon 1954 and Olsson 1961, among others), and no species examined in this study shows a regular dentition. Rare specimens of G. (G.) hians from Discovery Bay, Jamaica, show a minute elongate ridge or boss in the right or left valve and a corresponding pit in the opposing valve, but this “cardinal dentition” is too irregularly developed to warrant homologies with the cardinal dentition of other bivalves. Apparently similar structures were noted for other gastrochaenids by Forbes and Hanley (1853, p. 129) and Lamy (1925). Lamy described the hinge as edentulous or showing a rudimentary cardinal tooth, forming a small callosity in older individuals. 41 8. BURROWS Unlike most endolithic bivalves, gastrochaenids secrete aragonitic burrow linings that form distinctive internal burrow shapes and commonly show considerable detail in sculpture and ornamentation. Typical internal bur- row shapes and burrow lining distributions are shown in Figure 33. The present observations largely confirm and extend those of Robertson (1963) for the three common Florida Keys gastrochaenids. Burrows of S. rostrata and G. (R.) ovata are less than twice as long as the shell and commonly curve gently in the dorsal direction. In contrast, burrows of G. (G.) hians are several times longer than the shell and may be straight or sinuously curved. All three burrows show two well-defined parts, a pos- terior siphonal burrow and an anterior shell chamber. Aragonitic linings in the form of an annular diaphragm or pointed “baffles” may constrict the siphonal burrow at its junction with the shell chamber. Asa result, shell mobility is restricted to the shell chamber. The shell chamber is only slightly larger than the shells in every dimension, and it varies from circular to slightly elliptical in cross-section (Fig. 34). This part of the burrow is wide enough to allow rotations of the shells about an anteropos- terior axis, and rotations by S. rostrata in partially opened burrows have been observed up to 90 degrees in either direction relative to the stationary foot. The lining of the siphonal burrow is smooth in G. (R.) ovata, but this Saas 1 cm. Fic. 33. Burrow cross-sections for adults of Spengleria rostrata (a), Gastrochaena (Rocellaria) ovata (b) and Gastrochaena (Gastrochaena) hians (c). The coral substratum is indicated by stipling; the aragonitic burrow linings are indicated by solid lines. The absence of a burrow lining is shown by a dashed line. The solid lines drawn between the siphon tubes of Spengleria rostrata suggest the configuration of accretion banding in the aragonitic deposits. 42 ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 43 = - ~ o g a fa e A SS a) <= — B10 Ff on "@24 m44@ «~ 8 ad A 4 - ® Qa E oO a= oO — 09 7) <= ” 0.8 ° 2,500 ; 5,000 7500 10,000 age in growth bands Fic. 34. Relationship between burrow shell chamber width/height ratio and age in Spengleria rostrata (dots), Gastrochaena (Rocellaria) ovata (squares) and Gastrochaena (Gastrochaena) hians (triangles). may show a slight annular constriction (i.e., diaphragm) in older speci- mens at the base of the siphons. In contrast, the siphonal burrow in G. (G.) hians shows irregular concentric ridges especially near the base of the siphons (Figs. 33c, 36). In S. rostrata the burrow lining projects promi- nently in the form of two pointed baffles at the base of each siphon. ‘These baffles appear identical to those described for another species of Spengleria by Soliman (1973). The siphon tube linings in S. rostrata also show minute knobs projecting about 0.3mm from the burrow walls, giving these linings a rough interior surface. The burrow linings in all three species show distinct accretion banding, and the linings in S. rostrata are commonly porous (Figs. 35, 36). The lining of the shell chamber is generally smooth, but this may appear locally rough and pitted where its prismatic micro- structure is unusually coarse. The thickness and distribution of the siphonal burrow linings depend largely upon the diameter of the siphons relative to the shell chamber and the direction of boring by the siphonal epithelium. The siphonal burrow linings increase in thickness with increasing difference in diameter be- tween the siphons and the shell chamber. Thus, the thick-siphoned G. (G.) hians secretes thin siphonal burrow linings, whereas the narrower siphoned S. rostrata and G. (R.) ovata secrete thicker linings. Because the siphons in S. rostrata spread apart by boring into the substratum, their burrow linings are thickest between the siphons and thin or absent on the opposite (i.e., “boring”) side of each siphon. The fused siphons of G. (G.) hians may also bore laterally into the substratum, similarly resulting in a thicker deposition of aragonite on one side (Fig. 33c). As in the siphonal burrow, the thickness and distribution of burrow linings in the shell chamber depend on the boring direction. This direction may be partly dorsal, ventral or lateral in addition to the prominent anterior direction. An unusual feature of many gastrochaenid burrows is the presence of minute tubules penetrating the burrow lining around the area of pedal 44 PEABODY MUSEUM BULLETIN 41 500 um. g of Spengleria rostrata. Note the lines of accretion paralleling the surface of deposition (horizontal in this photo- graph). Fic. 36. Acetate peel of a vertical section through the siphonal burrow lining of Gastrochaena (Gas- trochaena) hians. The coral substratum appears in the lower one-third of the photograph. The section is taken from near the base of the inhalant siphon, and is aligned parallel to the siphon length. attachment (Fig. 37). These are common features in burrows of G. (G.) hians, and are especially numerous in burrows approaching other borers or opposite surfaces of the substratum. For example, anterior burrow tubules were abundant in the two specimens boring near the coral margin in Figure 44. These tubules probably serve a probing function for guiding the boring direction (see below). All three gastrochaenids are capable of extending their siphonal burrow above the substratum, but this habit is typical only in G. (G.) hians. In this species the burrow lining invariably projects one to two centimeters above the substratum. When coral overgrowth threatens G. (G.) hians, it may further extend its siphons several centimeters beyond their normal length (Figs. 38, 44). These siphonal burrow extensions differ from those described by Otter (1937) for G. (G.) cuneiformis in showing only a partial calcareous partition between the inhalant and exhalant siphonal aper- tures. For S. rostrata, overgrowth by encrusting sponges is more of a problem than coral overgrowth because this species commonly settles the underside of coral margins. Because encrusting sponges seldom reach a considerable thickness (i.e., rarely over two or three centimeters), they induce only a slight lengthening of the siphonal burrow in S. rostrata (Fig. 40). The Soldier Key gastrochaenids show an amazing capacity for repair- ing even severe damage to their burrows. Burrow repair was observed for a Bermuda specimen of G. (G.) hians in a running water tank. On three successive days the mantle secreted aragonitic laminae along the anterior margins of a break in its burrow shell chamber (Fig. 41a,b). As shown in Figure 41b, these laminae overlap so that subsequent sheets reinforce and extend the previous ones. The formation of each lamina was accompanied by inflation of the anteroventral mantle and was initiated by the secretion of a mucous sheet. Initially supported by the mantle, the mucous sheet ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 45 Fic. 37. Probing tubules distributed over the anteroventral wall of the shell chamber of Gastrochaena (Gastrochaena) truncata, YPM 9281, from Pearl Islands, Panama. This view is from the burrow post- erior. The tubules penetrate the calcareous burrow lining and the coral substratum. Fic. 38. The siphonal burrow aperture of Gastrochaena (Gastrochaena) hians extended posteriorly apparently to avoid overgrowth by the coral Diploria clivosa. Fic. 39. Broken burrow shell chamber of Spengleria rostrata, showing the attached pedal apparatus and the contracted anteroventral mantle. Fic. 40. The siphonal burrow apertures of the Spengleria rostrata shown in Figure 39. An encrusting sponge surrounds the exhalant (left) and inhalant (right) apertures. Fic. 41a,b. Broken anterior shell chamber of Gastrochaena (Gastrochaena) hians showing the partially exposed shells and a newly secreted aragonitic burrow wall. a, dorsal view. b, anterior view. 46 PEABODY MUSEUM BULLETIN 41 became wrinkled when semirigid, and then hardened as the crystallization of aragonite was completed. As might be expected from their differing burrow lengths (see Figs. 33 and 42) the three gastrochaenids differ considerably in their life aver- aged rates of boring (Fig. 43). G. (G.) hians is the most rapid borer, and its boring rate decreases considerably with age. Although much slower bor- ers, S. rostrata and G. (R.) ovata show only a slight increase or decrease in boring rate with age, respectively. As illustrated by the latex burrow casts for G. (G.) hians in Figure 44, burrow lengths can also be greatly affected by erosional truncation (most common near the coral margins) or siphon elongation induced by threatened coral overgrowth. mm. burrow length 38 ° 2,500 5.000 - 7,500 10,000 age in growth bands i ‘ Fic. 42. Relationship between burrow length and age for Spengleria rostrata (dots), Gastrochaena (Rocellaria) ovata (squares), and Gastrochaena (Gastrochaena) hians (triangles). ny burrowing rate (um. coral bored per growth band) 2,500 7500 10,000 age in growth bands Fic. 43. Relationship between life-averaged burrowing rate and age at time of collection for Spengleria rostrata (dots), Gastrochaena (Rocellaria) ovata (squares), and Gastrochaena (Gastrochaena) hians (triangles). ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 47 new coral growth eroded coral of sa tok Ae ~ ke SOAS Rei NDEs us £2 BEE fat ore se a SEs SS ee LNT SRE aon eile Det Sess et Fic. 44. Latex casts of three adult specimens of Gastrochaena (Gastrochaena) hians in a diagratninatie representation of their actual positions in a Diploria substratum. Only one-half of the Diploria sub- stratum is shown in this figure. Note the attachment of the Diplora skeleton to the substratum on the far left. 9. MANNER OF BORING The role of mechanical abrasion in gastrochaenid boring is apparent from the strong anteroventral abrasion of their shells, but the precise manner of mechanical boring has never been observed. Gastrochaenid anatomy and shell morphology suggest that the anteroventral shell margins are abraded against the substratum by contraction of the pedal retractor muscles about the byssally or suctorially attached foot. Purchon (1954) and Yonge (1963) presumed that abrasion by the shell then occurs by closing or opening the valves, but Gohar and Soliman (1963c) suggested that abrasion is caused by anteroposterior shell movements with some rotation. Otter (1937) inferred from the oval burrow cross-sections in Gastrochaena (Gastrochaena) cuneiformis that this species bores by means of a rocking movement comparable to that observed in Pholas. Pholads typically rock their shells about a vertical axis (Nair and Ansell, 1968), but a comparable boring mechanism cannot occur in gastrochaenids because of their rela- tively straight, long hinges and prominent ligaments. Gohar and Soliman noted that abrasion by the shell margins is supplemented by dorsal and lateral shell abrasion, and they added that gastrochaenid burrow circular- ity requires rotation of the shells about an anteroposterior axis. The observed rotation of S. rostrata in its burrow (see above) suggests that this may be a significant aspect of the boring mechanism. Hancock (1848) suggested that gastrochaenids and other endolithic bivalves abrade sub- strata by means of siliceous particles imbedded in their pedal and mantle epithelium. Subsequent authors (e.g., Jeffreys 1865) were unable to substantiate Hancock’s theory, and siliceous particles were not found in the mantle of the Florida Keys gastrochaenids. One specimen of G. (G.) hians from Discovery Bay, Jamaica, showed crystals of aragonite superfi- cially impressed in its siphonal and anterior mantle epithelium (Fig. 45) but these structures are atypical and may have been formed only in re- sponse to irritation of the epithelium by crab commensals found within its mantle cavity. It is likely that the “siliceous” particles described by Hancock were in fact sand grains adhering to the epithelium or refractive lipid globules within the epithelium. The distribution of shell abrasion in G. (G.) hians and G. (R.) ovata suggest that their mechanical boring is accomplished primarily by the latest-formed, projecting, comarginal shell ridge. In S. rostrata abrasion also occurs on the lateral valve surfaces, but here the aragonitic perio- stracal spikes rather than the underlying ridges are the primary agents of abrasion. Abraded coral debris probably enters the anterior mantle cavity through the dilated pedal gape, where it is bound into mucous strings and passed posteriorly to the base of the inhalant siphon. The same mantle pumping activity that serves to expel these and other pseudofeces (see above) will also cause a strong water current to pass between the animal 48 ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 49 Fic. 45. Left: Ventral exterior view of Gastrochaena (Gastrochaena) hians showing the shell margin (far left), the folded periostracum (most of the middle portion of the photograph) and bundles of aragonite crystals on the exposed anterior mantle (far right). Right: three bundles of aragonite crystals removed from the anterior mantle by lightly scraping its surface. and the burrow wall, thereby sweeping abraded coral debris toward the pedal aperture and into the mantle cavity. Because of their restriction to calcareous substrata and the apparent ability of their delicate siphons to enlarge the posterior of the burrow, gastrochaenids have been considered to bore partially by chemical means (Otter 1937, Yonge 1963, Gohar and Soliman 1963c). The divergence of the siphons of Spengleria within the substratum (Fig. 33a) clearly requires a chemical boring mechanism, since the siphonal epithelium is relatively thin and is not protected by a periostracal sheath. Chemical boring is also required for the formation of the minute tubules that penetrate the anterior burrow linings in G. (G.) hians. The diameter of these tubules and their location near the pedal apparatus suggest that they are produced by the anterior pedal organ. Probing by this pedal organ would clearly be adaptive for directing burrowing, since gastrochaenids commonly bore relatively thin shells and unstable coral margins. This function is compati- ble with the commonness of probes in G. (G.) hans burrows nearing outer surfaces of the substratum or approaching burrows of neighboring borers. As shown by vertical sections through the anterior of the burrow, these probing tubules commonly follow a sinuous course, penetrating several millimeters of the coral substratum in addition to the calcareous burrow lining (Fig. 46). An unusual example of pedal probing is provided by a Pliocene Gastrochaena (Gastrochaena) sp. from North Carolina. The speci- men whose burrow is shown in Figure 47 initially bored into a pectinid shell, but formed its own aragonitic tube when it outgrew this substratum. A latex cast of the interior of this tube (Fig. 47, right) shows numerous anterior tubules that penetrate both the secreted tube walls and shell debris incorporated into the surface of the tube. Some of these tubules branch distally, suggesting multiple probings from the same penetration through the aragonitic tube. Additional evidence for chemical boring comes from certain Indo- Pacific species of Gastrochaena (Cucurbitula) boring into Spondylus shells. Gastrochaena (Cucurbitula) commonly penetrates the outer calcitic shell 50 PEABODY MUSEUM BULLETIN 41 Fic. 46. Penetration of the burrow lining and part of the coral substratum by a probing tubule of Gastrochaena (Gastrochaena) truncata, YPM 9281, from Pearl Islands, Panama. Note the truncation of accretion lines within the general burrow lining, indicative of a previous cycle of resorption and resecretion. The diagram is based on an acetate peel of a vertical section through the burrow anterior. Fic. 47. Lateral (left) and posterior siphonal (middle) view of the calcareous laminated and aggluti- nated tube of a Pliocene Gastrochaena (Gastrochaena) sp. from Duplin County, North Carolina (USNM 9979). layer of Spondylus and partly enters the underlying aragonitic layer. Chem- ical boring by Gastrochaena (Cucurbitula) is indicated by a distinct indenta- tion in its burrow wall that follows the contact between the calcitic and aragonitic layers of the host shell. This indentation is too sharply defined to result from differences in the mechanical abradability of the Spondylus shell layers, but it may result from more rapid chemical erosion of the aragonitic layer. This would be compatible with the higher thermo- dynamic instability of aragonite. _ Although chemical boring by the pedal, siphonal and anterior mantle epithelium in gastrochaenids seems certain, the nature of the boring agent is presently unknown. Deshayes (1846) proposed that the anterior half of ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 51 the interior pallial glands in Gastrochaena dubia produce acid for chemical boring, and similar suggestions were offered by Cailliaud (1856) and Carazzi (1903). Extensive lateral pallial glands are found in the tropical Western Atlantic species and in other gastrochaenids (see above and Pelseneer 1911), but the position of these glands within the mantle cavity suggests that they are not directly involved in chemical boring. These glands probably function like the similar glands in the burrowing and boring Hiatellacea, i.e., to secrete mucus for binding pseudofeces within the mantle cavity (see Hunter 1949 and Yonge 1971). According to Jacca- rini et al. (1968) Lithophaga lithophaga bores chemically by means of calcium-complexing secretions emanating from two dorsal pallial glands. Distinctive glands in a similar position have not been observed in the Soldier Key gastrochaenids. However, future investigations might explore the possibility that the chevron-shaped glands located immediately dorsal to the anterior pedal organ inS. rostrata, or other pedal glands, may secrete a chemical boring agent. Another possibility, suggested by Kuhnelt (1934), is that the glands participating in chemical boring are not concentrated, but rather occur throughout the entire exposed mantle epithelium. Sucha mechanism is plausible, considering that the mantle epithelium in many bivalves is capable of bringing about general decalcification of the inner shell surface (Dugal 1939, Crenshaw and Neff 1969) or eroding tubules through previously deposited shell layers (Oberling 1964, Taylor et al. 1969). 10. GASTROCHAENID ORIGINS Numerous fossils with probable gastrochaenid affinities are described in the literature of European Jurassic bivalve faunas, e.g., “Gastrochaena” infraliasina Terquem 1855 and several species of “Gastrochaena” described by Phillips (1829), Eudes-Deslongchamps (1838), Buvignier (1852), de ‘Loriol and Bourgeat (1888), de Loriol (1891), and Arkell (1929-1937). Many of these Jurassic forms have not been critically re-examined since their original description in the nineteenth century, and the mytilid versus gastrochaenid affinities of the more elongate of these species have yet to be satisfactorily demonstrated. But some of these fossils, like the Jurassic Gastrochaena moreana Buvignier 1852 (Figs. 54-56) and “Gastrochaena” [=Spengleria] recondita (Phillips 1829) show unmistakable gastrochaenid affinities in terms of their flasklike burrow, Rocellaria-like or Spengleria-like shell outline, and well developed siphons (judging from the shape of the siphonal burrow). These Jurassic forms generally differ from the modern species of Gastrochaena and Spengleria in their more restricted pedal gape and greater lateral compression in the shell anterior. Most current hypotheses regarding the origins of the Gas- trochaenacea are based on comparisons of anatomy and life habits be- tween Recent representatives of the Gastrochaenacea and Pholadacea. Purchon (1954) considered the Pholadacea as possible gastrochaenid rela- tives on the basis of their 1) common representation by borers with cal- careous burrow linings, 2) presumed homologies between the pholad apophysis and gastrochaenid myophore, and 3) possible similarities in siphon structure. But Purchon noted that differences in manner of bor- ing, stomach structure, and visceral ganglia do not support the view that gastrochaenids are an early offshoot from a pholadacean stock. Purchon (1954) concluded that “there is insufficient evidence to justify any view as to relationship between the Gastrochaenidae and the Adesmacea [Pholadacea].” Purchon’s negative conclusion is supported by the fact that the pholad apophysis and gastrochaenid myophore are not homologous, because they attach to different pedal muscles, i.e., to the anterior and posterior pedal retractors in Gastrochaena and Zirfaea respectively (see above and Nair and Ansell 1968). In addition, possible pholadomyoid ancestors of the Pholadacea (e.g., Myopholas and Girardotia; see Morris in Taylor et al. 1973, p. 291, and Runnegar 1974) are probably too specialized for a deep burrowing life habit to have given rise to the Jurassic gastrochaenids. Although the pholads are unlikely gastrochaenid ancestors, several other taxa remain as possible ancestors on the basis of their anatomical similarities (e.g., the Hiatellacea) or because of their representation by fossil forms morphologically similar to the early gastrochaenids [e.g., the Isofilibranchia (Modiomorphacea and Mytilacea), Permophoridae and Grammysiidae]. Certain representatives of these taxa resemble gas- 52 ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 53 trochaenids in their general shell form, reduced dentition or edentulous hinge, external opisthodetic ligament inserting on a ligamental nymph, and reduced anterior adductor muscle scars. In addition, certain living or fossil representatives of all these taxa show evidence of gastrochaenid-like periostracal calcification (Carter 1976a). Data of shell features and known or inferred anatomy and ecology for all these taxa are compiled in Figure 48. Features of soft anatomy for the extinct Modiomorphacea (Isofili- branchia), Grammysiidae, and Permophoridae are inferred from their closest living relatives, i.e., the Mytilacea (Isofilibranchia), Pholadomyidae or Pandoracea, and Carditidae, respectively. Data of labial palp and stomach structure were not available for the recent Pholadomyidae; so these are inferred from the Pandoracea, a later evolving pholadomyoid superfamily. Excluded from this tabulation are features (enumerated 10 9 palp grade Ey stomach grade" extensive pallial glands ligament fusion layer ventral mantle fusion elongate siphons ctenidium grade labial Ren Bee HIATELLIDAE” GA ES G ctr [PalPalalP |r |alPale| 3 [emammysiioae? —_[n7[a[Pala[po[ ae [po] al e Pa) « [easrrocuaenioaelc|P[r|e/alrlelalP lel? Fic. 48. Summary of characters varying among the Gastrochaenidae, Isofilibranchia (Mytilacea and Modiomorphacea), Permophoridae, Hiatellidae and Grammysiidae. Excluded here are characters found in all five groups (e.g., aragonitic periostracal spikes and anterior shell reduction) and specialized features found only in the Gastrochaenacea (e.g., the pedal probing organ). Symbols: present (P), absent (4), nacreous shell structure (N), crossed lamellar shell structure (CL), filibranch ctenidia (F) and eulamellibranch ctenidia (E). "Soft anatomy data inferred from the Recent Cardita (Carditidae; see Yonge 1969). *Soft anatomy data from Yonge (1971) unless otherwise noted. 3Soft anatomy data inferred from Recent Pholadomya candida (Pholadomyidae; see Runnegar 1972) unless otherwise noted. ‘Data from Nakazawa and Newell (1968). °Data from Yonge (1948). ‘Data inferred from the Recent Pandoracea (see Stasek 1963 and Purchon 1958). 7Data from Taylor et al. (1969, 1973) unless otherwise noted. ’Data from various authors in the Treatise on Invertebrate Paleontology (R. C. Moore, ed.) Part N, Mollusca 6, (1969). ®Data from Pelseneer (1911) unless otherwise noted. 1TData from Stasek (1963). "Data from Purchon (1958). 54 PEABODY MUSEUM BULLETIN 41 above) found in all the compared taxa and specialized features unique to the Gastrochaenacea. The data of Figure 48 are abstracted diagrammati- cally in Figure 49 to show relative unweighted similarities with the Gas- trochaenacea. GASTROCHAENIDAE SISOFILIBRANCHIA PERMOPHORIDAE GRAMMYSIIDAE HIATELLIDAE Fic. 49. Abstract of the data compiled in Figure 48, showing unweighted similarities between the Gastrochaenidae and four possible ancestors. Each line connecting the Gastrochaenidae and another taxon represents one similarity from Figure 48. The two dotted lines represent possible but unverified similarities. Although the Hiatellacea and the Isofilibranchia resemble the gas- trochaenids in their representation by rock borers, these borers appear too early or too late in the fossil record or are too dissimilar in shell form to be likely gastrochaenid ancestors. The semiendolithic nestling Ordovician modiomorphid Corallidomus (Whitfield 1893) appears much too early to have given rise to the gastrochaenids. The Mytilacea are represented by Carboniferous through Permian forms morphologically similar to Lithophaga, but an endolithic habit has yet to be demonstrated for any Upper Paleozoic lithophagid (see Pojeta and Palmer 1976). Among other early lithophagids, the Permian Lithodomina is too specialized in terms of its internal ligament, and the Jurassic Inoperna is too dissimilar in its large size and shell form to be likely gastrochaenid ancestors (see generic diagnoses in Soot-Ryen_1969 and Pojeta and Palmer 1976). Furthermore, data of soft anatomy and shell mineralogy suggest that the Isofilibranchia are the least likely gastrochaenid ancestors among the possibilities considered in Figure 49. Many Mytilacea have an outer calcitic prismatic shell layer that differs mineralogically and microstructurally from the aragonitic outer prismatic layer observed in gastrochaenids (Oberling 1964 and Taylor et al. 1973). The Hiatellacea show greater anatomical similarity to the Gas- trochaenacea than any other taxon represented in Figure 49, but they are doubtful direct gastrochaenid ancestors for a variety of reasons. Prior to the Jurassic, the hiatellids may be represented only by a Triassic Panopea. Keen (1969b) considered the Permian Roxoa to be a hiatellid, but Run- negar and Newell (1971) subsequently allied this genus with the Pholadomyidae. Panopea and Roxoa are unlikely gastrochaenid ancestors because of their permanent posterior siphonal gape, a feature unknown in any fossil or Recent gastrochaenid. Hiatella, the only hard substrate borer in the Hiatellacea, does not appear until the Upper Jurassic (Keen 1969b), too late to have given rise to the Gastrochaenacea. In addition, the present data of functional morphology support Purchon’s (1954) suggestion that the endolithic habit evolved independently in Hiatella and the Gas- ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 55 trochaenacea. Unlike Gastrochaena, Hiatella moves its shell freely 1n its burrow, apparently using its siphons as a fulcrum for abrasion, and rock- ing its valves about a dorsoventral axis (Hunter 1949). Some anatomical similarities are also expressed differently in Hiatella and the Gas- trochaenacea. Whereas accessory ventral adductors in Hiatella form a muscular floor to the mantle cavity (Hunter 1949), the ventral adductor in Gastrochaena (G.) hians is a single, relatively small muscle near the base of the siphons. Extensive pallial glands are found in Hiatella along either side of the midventral region of the mantle cavity (Pelseneer 1911), whereas analogous pallial glands in Gastrochaena do not occur in a midventral position. In addition, the Hiatellacea differ microstructurally from the gastrochaenids. Whereas gastrochaenids show reclined composite prisms in their outer shell layer, hiatellids show irregular prisms with no distinct substructure (in Hiatella and Cyrtodaria) or a unique vertical composite prismatic structure (in Panopea). The first order lamellae of the crossed lamellar layer are relatively large and sharply defined in gastrochaenids, but these are smaller and poorly defined in hiatellids (Carter 1976b). Finally, and perhaps most significantly, the Hiatellacea had just evolved and were not diverse at the generic level when the earliest gastrochaenids appeared in the Triassic or Jurassic (Fig. 50). On the other hand, certain Permian permophorids resembled the modern Gastrochaena (G.) hians (Fig. 10c) in their mytiliform lateral profile, and may have also been similar in their adaptations for adult byssal attachment. The Permian permophorid Myoconcha sp. of Winters (1963) shows an anteroventral byssal sinus and anterior umbones reminiscent of modern gastrochaenids (Fig. 51). But these similarities are probably con- vergent, because many Middle and Upper Jurassic Gastrochaena and Spengleria had not yet evolved a comparable degree of anterior reduction and lateral shell inflation. Other Permian forms, including some “Per- mophorous” (Permophoridae) and certain forms questionably allied with the Grammysiidae were morphologically similar to the Jurassic through Recent species of Spengleria. Sanguinolites? sp. of Chronic (1952) (Gram- mysiidae?) resembles Spengleria in its rounded anterior margins, flattened posterior triangular area set off by radial ridges, concentric ornament, moderately anterior and low umbones, and possibly in its edentulous hinge (see Chronic 1952 and the present Figs. 52 and 53). Forms like Sanguinolites? might be considered unlikely gastrochaenid ancestors be- cause their shell shape is suggestive of a shallow burrowing rather than an epibyssate life habit. But curiously, Spengleria has retained many shell and anatomical features commonly associated with shallow burrowing i in the modern Bivalvia. Separated siphons have apparently characterized Spengleria since its earliest (Jurassic) appearance in the fossil record (see below), and this feature is clearly more characteristic of modern shallow burrowers than hard substrate borers. Completely separated siphons occur in many shallow burrowing Tellinacea, but in no other endolithic bivalve besides Spengleria. This genus also resembles modern shallow burrowers in its diffuse pedal retractor musculature (see Yonge’s 1969 discussion of evolutionary trends in pedal musculature in the Carditacea). Considering the shell and anatomical similarities between Spenglerra 56 PEABODY MUSEUM BULLETIN 41 millions of years a Gastrochaenidae Hiatellidae Carboniferous| 399 i parhanhoriiae Fic. 50. Generic diversity through time for the Gastrochaenidae, Permophoridae and Hiatellidae. The Gastrochaenidae are regarded as comprising three extant genera (Spengleria, Gastrochaena, and Eufistulana) and one extinct genus (Kummelia). The genus Gastrochaena includes the subgenera Gas- trochaena s.s., Rocellaria and Cucurbitula. Following Runnegar and Newell (1971) the “hiatellid” Roxoa, a possible ancestor of the Hiatellidae, is classified with the Pholadomyidae. The data are otherwise abstracted from Keen (1969a, 1969b) and Chavan (1969). and many modern shallow-burrowing bivalves, it is not unreasonable to assume that this genus evolved more or less directly from shallow- burrowing ancestors, i.e., from Permian or Triassic forms morphologi- cally similar to Sanguinolites? sp. of Chronic (1952). By this hypothesis, the immediate gastrochaenid ancestors had not become morphologically specialized for epibyssate nestling prior to their assumption of the en- dolithic habit. Instead, forms like Sanguinolites? sp. may have evolved through a semiendolithic nestling stage in protected microhabitats. This would bring their mantle epithelium in contact with coral or shell substrates, thereby permitting evolution of chemical boring. Unlike mechanical bor- ing, chemical boring would require little change in the ancestral shell and anatomical features except for elaboration of glandular tissues involved in the process of calcium carbonate erosion. Gastrochaenids retaining a primarily chemical boring mechanism (e.g., Spengleria) would keep the lateral shell profile, separated siphons, and diffuse pedal musculature of ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA ey Fic. 51. Myoconcha sp. (Myoconchinae, Permophoridae) Permian Supai Formation, Arizona. Photo- graphic reproduction of plate 8, figure 15 (middle; valve interior; AMNH 28035: 1), figure 16a (above; valve exterior; AMNH 28035/1:1) and figure 16b (below; dorsal view of both valves; AMNH 28035/ 1:1) from Winters (1963). The lower scale bar refers to the upper and lower figures; the upper scale bar refers only to the middle figure. FIG. 52, Sanguinolites? sp., Permian Kaibab Formation, Arizona, rubber squeeze from an external mold. Photographic reproduction of plate 10, figure 16 from Chronic (1952). their shallow burrowing ancestors. They would become morphologically specialized for boring primarily in their anteroventral pedal gape. This permits permanent pedal attachment and the application of a wide area of mantle epithelium to the burrow walls for chemical boring. A corollary of this hypothesis is that Spengleria represents a primitive stock from which the more mechanically boring gastrochaenids evolved. Natural selection for efficient mechanical boring would result in the anterior reduction, streamlined lateral profile, united siphons, and concentrated pedal mus- culature characteristic of modern Gastrochaena. Unfortunately, the taxonomic affinities of Sanguinolites? sp. and similar Upper Paleozoic forms possibly ancestral to Spengleria have not been satisfactorily determined. It is therefore uncertain whether the Gas- trochaenacea can be regarded as likely derivatives of the Permophoridae (subclass Heteroconchia) or the Grammysiidae (subclass Anomalodes- mata). Chronic (1952) indicated that his Sanguinolites may be related to 58 PEABODY MUSEUM BULLETIN 41 Fic. 53. Spengleria spengleri (Deshayes) 1857, Eocene, France. Photographic reproduction of plate XVII, figure 1 (middle; valve interior), figure 2 (below; ventral view of both valves) and figure 3 (above; valve exterior) from Deshayes (1860). Permophorous, and Elias (1957) subsequently erected Eopleuorphorous to include Sanguinolites tricostatus (Portlock) and similar edentulous Permo- Carboniferous forms resembling Permophorous in general shell outline, subdued anterior umbones, anterior myophoric buttress, and a shell or- nament of minute papillae, concentric ridges, and posterior radial keellike ridges. Newell and La Rocque (1969) considered Eopleurophorous a possible synonym of Sanguinolites, presumably largely on the basis of its edentulous hinge. But this criterion may not be definitive, because the hinge dentition is also subdued in certain forms presently classified with the Per- mophoridae. Also, some taxa possibly related to the Grammysiidae show heterodont-like differentiated cardinal and lateral teeth (e.g., Alula squamulifera; see Fig. 27 in Runnegar and Newell 1971). Therefore, if the Gastrochaenacea evolved from shallow infaunal permophorids or gram- mysiids similar to Sanguinolites? sp., the subclass affinities of these ances- tors cannot presently be resolved. 11.. EVOLUTION AND ADAPTIVE RADIATION WITHIN THE GASTROCHAENACEA GENERAL EVOLUTIONARY TRENDS. As presently hypothesized, Spengleria evolved as a chemical borer from shallow infaunal burrowing ancestors through an intermediate semiendolithic nestling stage. Except for its stocky foot and anteroventral pedal gape, Spengleria retained the major shell and anatomical features of shallow burrowers like the Permian San- guinolites? sp. of Chronic (1952). The Jurassic gastrochaenid faunas of Europe are presently well enough known to compare the early representa- tives of Spengleria and Gastrochaena with their modern counterparts. These earlier gastrochaenids were generally characterized by narrower and short- er pedal gapes than the modern species (compare Figs. 10 and 54). Within Spenglerna, the Jurassic through Eocene species were also more laterally compressed than the modern Spengleria rostrata (compare Figs. [OA and 53). In addition, certain Jurassic Gastrochaena show burrow casts with slight separation of the posterior siphon tubes (Fig. 56), perhaps representing an early transition from the fully separated condition in Spengleria to the fused siphons characteristic of modern Gastrochaena. By Miocene time, and probably much earlier, Spengleria and Gas- trochaena had expanded their pedal gapes laterally and lengthwise, and Spengleria had also laterally inflated its shell anterior (e.g., see Spengleria emilyana from the Miocene of Florida; Vokes 1976). This parallel evolu- tionary expansion of the pedal gape in Spengleria and Gastrochaena permit- ted application of the anteroventral mantle and shell margins over a wider area of the anterior burrow chamber, thereby facilitating both chemical and mechanical boring. A wider pedal gape allows for boring in the entire burrow anterior with minimal rotation of the valves about the stationary foot. Inasmuch as some permophorids and grammysiids secreted calcified periostracal spikes (Carter 1976a), Spengleria probably inherited these structures from its burrowing ancestors. Among modern gastrochaenids, Spengleria is unique in its retention of prominent periostracal spikes over the exterior of its adult shell. The modern Gastrochaena seldom, if ever, secretes aragonitic spikes in its adult stage, but these are secreted by juveniles in a few modern species (Carter 1976a). Periostracal spikes are more common in fossil representatives of Gastrochaena, and these cover the entire shell in the Cretaceous Gastrochaena (Rocellaria) linsleyi (see below). It is likely that calcified periostracal spikes initially functioned to increase friction between the shells and the substratum in shallow burrowing and semiendolithic nestling ancestors of the gastrochaenids. Aragonitic spikes may likewise have assisted chemical boring by scraping debris from the burrow walls in forms like Spengleria and, presumably, in early representa- tives of Gastrochaena. But the spikes became less important as Gastrochaena became increasingly specialized for mechanical boring, because they were 59 60 PEABODY MUSEUM BULLETIN 41 Fics. 54-56. Gastrochaena moreana Buvignier 1852, YPM 10002, Jurassic (Malm), Stramberg, Czecho- slovakia. Fic. 54. Shell valves. Upper left: lateral exterior view of right valve. Irregularity on the posterior of the shell is sediment filling the burrow. Center left: dorsal view of right valve and internal cast of left valve. Lower left: ventral view perpendicular to the plane of the pedal gape. The shell interior is partially filled with crystals of calcite. Fic. 55. Natural internal limestone casts of three burrows. The burrow on the right shows the internal cast of the left valve visible in Figure 54 (left, center). The substratum is a scleractinian coral. Fic. 56. Diagrammatic reconstruction of the center burrow in Figure 55, drawn after removal of the surrounding coral substratum. The figure represents an internal cast of the burrow. The burrow was lined only near the extreme posterior of the siphonal portion (i.e., toward the upper right, near the burrow aperture). The three scale bars are graduated in millimeters. functionally replaced by stronger comarginal shell ridges. By Pliocene time, calcified spikes were restricted to the juvenile shell (Fig. 30) or entirely lost in most species of Gastrochaena. Calcified periostracal spikes have likewise been secondarily reduced during the course of evolution of Eufistulana. Stephenson (1941) described minute spikes covering parts of the Cretaceous Eufistulana ripleyana, but these are limited to the juvenile shell or are entirely lost in the Recent Eufistulana (see below). Whereas Spengleria and Gastrochaena specialized for chemical and combined chemical and mechanical boring, respectively, certain other gastrochaenids partially or entirely abandoned the endolithic habitat for a ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 61 tube-dwelling existence. The fossil Kummelia retained a shell similar to Gastrochaena (Fig. 61), but evolved a combined tube-dwelling and free- burrowing existence. The modern Eufistulana is more strictly a permanent tube dweller, and its shell has evolved a graceful sculpture unique in this superfamily (Figs. 62, 63). Evolutionary transitions leading from Spengleria or Gastrochaena to Kummelia and Eufistulana are presently un- known. But the modern semiendolithic Gastrochaena (Cucurbitula) may provide some indication of what the transitional life habit may have been like. G (Cucurbitula) has evolved an obligatory “igloo”-forming habit in which the shell chamber is partially enclosed by the substratum and par- . tially contained by calcareous laminae (Figs. 58-60). In some species of G. (Cucurbitula), the calcareous “igloo” comprises the largest portion of the burrow, so the animal is almost entirely a tube-dweller. As a final note, the available fossil data are sufficient to indicate that shell rotation was incorporated into the mechanism of boring early in the evolutionary history of Spengleria and Gastrochaena. With the exception of the Upper Cretaceous Gastrochaena (Rocellaria) linsleyi (Fig. 57) and the modern Gastrochaena (Cucurbitula) (Fig. 59), post-Jurassic gastrochaenid burrows typically show rounded anterior cross-sections. Nonrotation of the shell in the modern Gastrochaena (Cucurbitula) is clearly a secondary specialization made possible by its strongly reflected mantle (see below). EVOLUTION OF Spengleria. Spengleria is particularly interesting because it has retained many of the shell and anatomical features of its presumed permophorid or grammysiid ancestors, Spenglerea is represented in the Middle Jurassic by “Gastrochaena” sp. (Palmer 1974) and in the Upper Jurassic by S. recondita (Phillips) 1829 and S. corallensis (Buvignier) 1843. The latter two species resemble the Recent Spengleria rostrata in their lateral profile, moderately anterior umbones, posterior triangular area, posterior truncation, and winglike projection of the posterodorsal shell margins. Undoubted fossil Spengleria are also known from the Paleocene (S. cumitariopsis, Georgia, Harris 1896), Eocene (S. spengleri, France, Des- hayes 1857), Miocene (S. emilyana, Florida, Vokes 1976; see also Cossmann and Peyrot 1909 and Boss 1967), and Pleistocene (S. rostrata, Key Largo Limestone, Florida; Carter, personal observation). The two Recent species of Spengleria [S. rostrata (Spengler) 1793 and S. mytiloides (Lamarck) 1818] occur in the tropical and subtropical Western Atlantic and Indo-Pacific regions, respectively. As noted previously, the Jurassic through Eocene Spengleria differs from the modern S. rostrata in its more restricted pedal gape and more pronounced anterior lateral compression (see Arkell 1929-1937, his pl. 43, figs. 1-4; and Buvignier 1852, his pl. VI, figs. 1-6). Judging from the burrow casts illustrated by Buvignier (1852), the Jurassic Spengleria posses- sed completely separated siphons, but the siphon bases were not con- stricted by baffles projecting from the burrow wall, as in the modern Spengleria (see Buvignier 1852, his pl. VI, figs. 19-20). Except for these minor differences in shell and burrow form, the Jurassic and modern Spengleria are surprisingly similar, and this genus has been evolutionarily conservative. EVOLUTION OF Gastrochaena (Rocellaria) and Gastrochaena (Gastrochae- na). If Spengleria is in fact a primitive genus morphologically similar to the 62 PEABODY MUSEUM BULLETIN 41 Fic. 57. Shell and burrow of Gastrochaena (Rocellaria) linsleyi n. sp., holotype, YPM 10216a, Upper Cretaceous Ripley Formation, Coon Creek, Tennessee. Specimen removed from a shell of Cucullaea vulgaris Morton. s\. B-F. From top to bottom of figure: lateral exterior view, lateral interior view, dorsal view perpen- dicular to the hinge axis, ventral view perpendicular to the plane of the pedal gape, ventral view perpendicular to the hinge axis (showing hinge structure). Diagram of a latex cast of the burrow interior of the holotype, same magnification as 4. The dotted lines in B, D, and E indicate the outer (exterior) surface of the Cucullaea vulgaris substratum. Legend: B_ View from the ventral shell margins C View from the outer (exterior) surface of the Cucullaea substratum D_ View from the anterior shell margins E View from the posterior shell margins F View from the inner (interior) surface of the Cucullaea substratum a_ Anterior of burrow shell chamber d Dorsal aspect of burrow shell chamber e Exhalant siphon ? Inhalant siphon / Position of left shell valve p Posterior of burrow shell chamber r Position of right shell valve v_ Ventral aspect of burrow shell chamber ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 63 shallow burrowing gastrochaenid ancestors, then Gastrochaena probably evolved from Spengleria or from Spengleria’s immediate ancestors through specialization for efficiency in mechanical boring. Increased reliance upon mechanical boring required streamlining of the shell and siphons into a cylindrical tube, reduction of the shell anterior, and eventually replace- ment of calcified periostracal spikes with stronger comarginal shell ridges. Anterior shell reduction in Gastrochaena s.s. and Rocellaria increased the mechanical leverage of the anterior pedal retractor muscles by bringing their attachment sites closer to the rasping shell margins. Gastrochaena s.s. is more specialized for mechanical boring than Rocellaria in terms of its generally stronger anterior reduction and more prominent myophores. Gastrochaena has been reported from the Triassic and Lower Jurassic (e.g., G. infraliasina Terquem 1855; see also references to Triassic species compiled by Diener, 1923). But the oldest presently confirmed member of this genus is G. moreana Buvignier 1852 from the European Middle and Upper Jurassic (Figs. 54-56). G. moreana and the Upper Cretaceous G. (Rocellaria) linsleyi (Fig. 57) differ from most post-Mesozoic Gastrochaena 1n their smaller pedal gapes. Like Spengleria, the genus Gastrochaena shows a general evolutionary increase in the width and length of its pedal gape, and the introduction of siphonal baffles in certain later Cenozoic species. The subgenera Gastrochaena s.s. Rocellaria are difficult to distinguish in the older fossil record because their hinge structures are often not preserved, and also because transitional forms were more common than at present. But the subgenus Gastrochaena s.s. was well differentiated at least by Oligocene or Miocene time, when it was represented by the Western Atlantic G. (G.) ligula Lea 1846, G. (G.) rotunda Dall 1898, and numerous other species. Gastrochaena s.s. comprises the majority of gastrochaenid species In modern endolithic faunas. EVOLUTION OF Gastrochaena (Cucurbitula), Kummelia, AND Eufistulana. The subgenus Cucurbitula and the genera Kummelia and Eufistulana are characterized by their replacement of the ancestral endolithic habit with an obligatory “igloo” or tube-dwelling habit. It is not uncommon for Gastrochaena s.s. and Rocellaria to construct nearly complete calcareous tubes after boring through substrata too thin to contain their shells (Fig. 47). Even Spengleria occasionally secretes a partial calcareous tube in re- pairing severe damage to its burrow. But tube formation is clearly faculta- tive in these three latter taxa, and does not characterize individuals in- habiting adequate shell and coral substrata. In contrast, “igloo” and tube dwelling are the preferred life habits among representatives of Cucurbitu- la, Kummelia and Eufistulana. Cucurbitula typically bores shallowly into the exterior of other shells and then completes the dorsal half of its burrow by secreting a calcareous “igloo” (Figs. 58-60). Superficial “igloo” formation is obligatory in Cucur- bitula, because this habit characterizes individuals boring even thicker substrata. The Indo-Pacific Gastrochaena (Cucurbitula) cymbium precedes “igloo” formation by boring shallowly into a shell substratum, and then emerges on the bored surface to secrete the calcareous “igloo” walls. Thus, G. (C.) cymbium is fully endolithic as a juvenile, and its “igloo” is constructed only as the later juvenile and adult burrow increases in size (Fig. 58). The “igloo”-forming stage is assumed after a more abbreviated juvenile en- 64 PEABODY MUSEUM BULLETIN 41 Fic. 58. Dorsal (a) and lateral (6) views of the calcareous “igloo” secreted by Gastrochaena (Cucurbitula) cymbium Spengler 1783, YPM 10218a, Calapan, Mindoro, Philippines, boring into Plicatula muricata Sowerby. Same specimen as in Figures 59 and 60. The cupules comprising the “igloo” represent successive stages of anterior burrow enlargement. The scale bar represents 10 millimeters. Fic. 59. Latex cast of the burrow interior of Gastrochaena (Cucurbitula) cymbium Spengler 1783, YPM 10218a, Calapan, Mindoro, Philippines. From top to bottom of figure: lateral, ventral, and dorsal views. The lateral view is slightly oblique to show the impressions of the umbones in the burrow shell chamber. The ventral view shows the impression of the pedal attachment scar in the middle of the burrow shell chamber. Photographs prepared by William Pirowski, 1974, at Colgate University (Hamilton, N.Y.). ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 65 Fic. 60. Gastrochaena (Cucurbitula) cymbium Spengler 1783, YPM 10218a, Calapan, Mindoro, Philip- pines. 66 PEABODY MUSEUM BULLETIN 41 dolithic stage 1n certain other representatives of this subgenus [e.g., in G. (C.) tasmanica; see Laseron and Laseron 1952. ] The flattened shape of the shell chamber in Cucurbitula requires that its shell does not rotate during the process of boring and “igloo” formation (Fig. 59). Boring and “igloo” formation are both accomplished by the ventral mantle, which is strongly reflected over the anterior of the shell valves (Gould 1861). Cucurbitula “igloos” are typically constructed of cup- shaped calcareous walls, with the size of the cups increasing toward the anterior of the burrow (Fig. 58). Burrow enlargement in Cucurbitula 1s apparently accomplished by periodic resorption of the burrow anterior, followed by the secretion of new anterior calcareous cups. In addition to its obligatory semiendolithic habit, Cucurbitula is readily distinguished by its unusually long and wide pedal gape and dorsoventral compression of the shells and burrow (Fig. 60). Cucurbitula resembles certain Gastrochaena s.s.1n its possession of pointed baffles projecting from the burrow lining at the base of the siphons. Tryon (1862) indicated that Cucurbitula is widely distributed in the fossil record. But Tryon possibly included under this name many facultative tube-dwelling representatives of Gastrochaena s.s. and Rocellaria. Fossil Cucurbitula as originally defined by Gould (1861) and as described here is presently known only from a single Tertiary species from Italy (see Brocchi 1814, his pl. XI, fig. 14a,b). The Recent species of Cucurbitula are restricted to the Indo-Pacific region and Australasia (Sturany 1899, Lynge 1909, Lamy 1923, Laseron and Laseron 1952). Although apparently closely related to the other subgenera in Gastrochaena, the phylogenetic origins and time of appearance of Cucur- bitula are presently unknown. Among the Mesozoic Gastrochaena, the Upper Cretaceous G. (R.) linsleyi resembles modern Cucurbitula in its flattened burrow shell chamber and shallow burrowing habit (Fig. 57). However, the shells of Cucurbitula and G. (R.) linsleyi are strikingly diffe- rent, and there are presently no known morphological intermediates. Unlike Cucurbitula, both Kummelia and Eufistulana have entirely aban- doned the endolithic habit for a specialized free-burrowing, tube-dwelling existence. Both taxa appeared in the fossil record during Cretaceous time, but only Eufistulana is represented in Upper Cenozoic and Recent faunas. Kummelia is known only from the Cretaceous through Eocene Kummelia americana (Gabb) 1860 from Europe and eastern North America (Holzap- fel 1889, Wade 1926, Stephenson 1937, Richards et al. 1958, Palmer and Brann 1965-1966). Shells of Kummelia are similar to modern Gastrochaena except for their greater elongation (Fig. 61, middle and right). Natural internal casts of K. americana tubes show widely spaced annular constric- tions (Fig. 61, left). Stephenson (1937) interpreted these to represent successive anterior tube walls partially removed during the process of periodic anterior tube resorption and secretion. Unlike many gas- trochaenids (including Eufistulana), Kummelia probably lacked a single, major constriction of the burrow lining at the junction of its siphonal and shell chambers. Complete Kummelia tubes are presently unknown, but partially dissolved tubes from the Paleocene of Maryland (Fig. 61, middle and right) indicate that the animal secreted an exteriorly smooth, conical tube at least in the burrow posterior. In a few Kummelia the shells are preserved some distance from the end of an irregularly shaped tube ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 67 Fic. 61. Natural internal casts of tubes of Kummelia americana (Gabb) 1860. Left: regularly shaped tube showing several widely spaced annular constrictions of the tube lining. Each constriction may repre- sent a previous position of the anterior tube wall. Successive anterior tube walls were probably resorbed as the siphons increased in length. USNM 496382, Vincentown Formation, Lower Eocene, New Jersey. Middle and right: two views of an irregularly shaped tube showing the right shell valve. The straight edge on the upper right portion of the right figure probably represents the impression of a posterior tube wall. USNM 496381, Aquia Formation, Paleocene, Maryland. The scale on the left represents 10 millimeters. anterior. This suggests that Kummelia employed considerable anterior- posterior shell movement to accomplish its periodic burrow expansion. This also reinforces the inference t ee Kummelia tubes did not elias dis- tinctly separated siphonal and shell chambers. The shells of Eufistulana differ from Kummelia and all other gas- trochaenids in their extreme elongation, sharply truncated anterior, and unusually long and wide pedal gape. The Indo-Pacific E ufistulana mumia (Spengler) 1783 is strongly compressed dorsoventrally, and it lacks pro- jecting myophores (Fig. 63). The elongate, conical tube of Eufistulana shows well-defined siphonal and shell chambers separated by a double- walled, elliptical diaphragm (Fig. 62). This partition greatly restricts shell movement in the anterior-posterior direction, but does not interfere with rotational activity within the shell chamber. Burrow elongation i in Eufistulana is apparently accomplished by episodic resorption and resecretion of the medial diaphragm and the anterior tube wall. Kthnelt (1934) suggested that Eufistulana increases its 68 PEABODY MUSEUM BULLETIN 41 Fic. 62. Right: lateral view of calcareous tube of Eufistulana mumia (Spengler), 1783, YPM 9589, Singapore. Left: lateral view of latex cast of tube interior. Middle: diagrammatic longitudinal cross- section through tube, showing the elliptical diaphragm between the posterior siphonal burrow and the anterior shell chamber. Scale represents 10 millimeters. burrow length in increments comparable to the shell length. By this hypothesis, the medial diaphragm represents the position of the previous anterior tube wall, and an entirely new shell chamber is formed during each phase of burrow elongation. Alternatively, Eufistulana may increase its tube length in smaller increments, i.e., by more frequent resorption and resecretion of its diaphragm and burrow anterior. This latter hypothesis seems likely, because the shell and shell chamber are nearly the same length in all representatives of Eufistulana examined by the author. In addition, Eufistulana’s tube shows numerous closely spaced concentric ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 69 Fic. 63. Eufistulana mumia (Spengler 1783), YPM 9589, Singapore. From top to bottom of figure: lateral exterior view, lateral interior view, dorsal view perpendicular to the hinge axis, ventral view perpendicular to the plane of the pedal gape, ventral view perpendicular to the hinge axis (showing hinge structure). 70 PEABODY MUSEUM BULLETIN 41 accretion bands on its anterior, and its medial diaphragm occasionally shows evidence of closely spaced successive diaphragm positions. Eufistulana is represented in the Upper Cretaceous by E. linguiformis (Weller) 1907 and E. whitfieldi (Weller) 1907 (both from New Jersey) and by E. ripleyana (Stephenson) 1941 from Mississippi and Texas. Of these three species, at least E. rpleyana retained calcified periostracal spikes in radiating and irregular rows. Among Recent species of this genus, spikelike calcified periostracal structures are generally absent or are re- stricted to the posterior of the juvenile shell (e.g., in an Australian Eufis- tulana clava; USNM 714790). Definite fossil Eufistulana are also known from the Eocene of Europe, northern and eastern Africa, and Pakistan (Glibert 1936, Eames 1951) and from the Oligocene of Germany (Koenen 1894). Like Cucurbitula, the modern species of Eufistulana are probably restricted to the Indo-Pacific region and Australasia. There can be little doubt that Eufistulana evolved from endolithic or semiendolithic ancestors. Even the modern Eufistulana is capable of pene- trating calcium carbonate substrata that interfere with its burrowing direc- tion (see Sowerby and Fulton 1903 and Smith 1907). In these instances, Eufistulana’s boring mechanism is almost certainly chemical, because its shells are delicate and show little evidence of mechanical abrasion on their margins. Possibly Eufistulana evolved from Kummelia or from a less specialized representative of Gastrochaena retaining the ancestral chemical boring mechanism and adult periostracal calcification. 12. CONVERGENCES BETWEEN THE CLAVAGELLACEA AND GASTROCHAENACEA Some representatives of the Clavagellacea are so strikingly similar to the Gastrochaenacea that they deserve special consideration. Like Eufistulana, most Clavagellacea construct flasklike calcareous tubes that surround the soft tissues, and which may be perforated anteriorly (compare Figs. 47 and 64). Like the Gastrochaenacea, some Clavagellacea pump water through their pedal aperture, and in both the Clavagellacea (Purchon 1960) and some tube-dwelling gastrochaenids this habit may serve as an accessory burrowing mechanism. Additionally, both superfamilies contain primarily tropical, long-siphoned, eulamellibranch bivalves that are represented by mechanical and chemical borers (Soliman 1971). Some clavagellids resem- ble gastrochaenids in their periostracal calcification (Carter 1976a) and some even resemble Gastrochaena (Gastrochaena) in having a myophore-like structure projecting from beneath the hinge (Fig. 65). Based onan analysis of their Recent and fossil shells and anatomical data, further similarities between these superfamilies can be summarized as follows: 1. Anterior umbones (see Deshayes 1857). 2. Mostly well-developed pallial sinus (see Deshayes 1857). 3. Extension of the ctenidia past the posterior shell margin (Purchon 1956b). . Annular siphonal membranes (Purchon 1956b). . Stomach construction, grade “four” (Purchon 1958). . Accessory ventral adductor muscles (Soliman 1971). But the Clavagellacea and Gastrochaenacea also show certain dis- similarities suggesting that they are not closly related. For example, the free shell valves of many Cretaceous and early Tertiary Clavagellacea (1.e., the valves not fused to the calcareous tube; see Deshayes 1857) closely resemble Mya and Panopea. This shell form indicates that the immediate ancestors of the clavagellids were probably specialized for deep rather than shallow burrowing. Like Panopea (Hiatellacea), the early Clavagel- lacea commonly show a permanent posterior siphonal gape and subequal adductor muscle attachment scars, and they lack gastrochaenid-like con- centric ornamentation. Although the Gastrochaenacea have evolved tube-dwelling species, their shell valves are never incorporated into the tube. Partial or complete fusion of one or both shell valves to the tube occurs in most Clavagellacea except certain early forms (e.g., Clavagella cornigera Schafheutl, see annotation, p. 170, in Smith 1962). Also unlike Gastrochaena and Spengleria, endolithic clavagellids abrade only the free shell valve and perhaps also the periostracum-covered mantle against the substratum, using the attached valve to provide leverage (Soliman 1971). Even among endolithic clavagellids, the foot is totally unlike that in the Gastrochaenacea. The clavagellid foot lacks both a circular pedal disc and an anterior pedal probing organ (Soliman 1971). The Clavagellacea re- HD Ov 71 72 PEABODY MUSEUM BULLETIN 41 Fic. 64. Calcareous tube constructed by Clavagella multangularis Tate (USNM 159380). Note the fusion of the left valve to the aragonitic tube. 3mm 3mm Fic. 65. The free valve of Clavagella mutangularis Tate removed from the aragonitic tube in Figure 64. Note the radiating lines of aragonitic periostracal spikes on the valve exterior (left) and the deep pallial sinus (right). The center photographs show the hinge from lateral (upper) and ventral (lower) views. semble some Pholadomyidae or Pandoracea more than the Gas- trochaenacea in their type “E” ctenidial structure (Atkins 1937), “fourth” mantle aperture (Allen 1958, Lacaze-Duthiers 1883), nacreous shell microstructure (Taylor et al. 1973) and periostracum-covered siphons (Purchon 1956b; Runnegar 1972). If, as is apparent from a description of a young Humphreyia by Smith (1911), some clavagellids also have a lithodesma, then this superfamily probably evolved from the Pandoracea. The tube-dwelling Clavagellacea and Gastrochaenacea, both of which appeared in the Upper Cretaceous, would then have descended from ecologically and phylogenetically dissimilar ancestors. Whereas the Gas- trochaenacea evolved from shallow-burrowing permophorids or gram- mysiids, the Clavagellacea evolved from deep-burrowing pandoraceans. 13. DISCUSSION DIVERSIFICATION WITHIN THE ENDOLITHIC HasiTaT. The endolithic habit is one of the most specialized and, appropriately, one of the last major life styles to evolve in the Bivalvia. Facultative epifaunal and boring bivalves have been described from the Ordovician (e.g., Corallidomus scobina; see Pojeta and Palmer 1976) but these are not truly endolithic. By the present definition, endolithic bivalves both excavate and are largely enveloped by their hard substratum. The Ordovician Corallidomus merely abraded hard substrata with its ventral shell margins, i.e., much like the modern epilithic Arca imbricata Bruguiere. The Carboniferous and Permian fossil record shows several species morphologically similar to the modern chemical borer Lithophage (see Merla 1931, Frebold 1933, and Wanner 1940 for the Permian forms). But according to Pojeta and Palmer (1976), the presumed endolithic habits of these Paleozoic Lithophaga are yet unverified. The Pholadacea are questionably represented by a Carboniferous Martesia (Turner 1969), but it is possible that this superfamily evolved only during the Mesozoic, i.e., from pholadomyoids similar to Myopholas and Girardotia. The latter two genera show striking similarities in shell form to certain modern pholads like Pholas latissima Sowerby (YPM 9590, Philippines). In any event, the fossil record of abundant and diverse endolithic bivalves clearly does not appear until the Triassic or Jurassic. Middle and Upper Jurassic tropical corals and hardgrounds are commonly infested by truly endolithic species of Gastrochaena, Spengleria, and Lithophaga (Arkell 1929-1937, Palmer 1974, Palmer and Fursich 1974, Fuirsich and Palmer 1975). Aside from the Mytilacea and Gastrochaenacea, only two other bivalve superfamilies are presently known to have evolved endolithic species during the Mesozoic. These are the Hiatellacea, represented by the Jurassic through Recent Hiatella (Keen 1969b; Hunter 1949; Yonge 1971) and numerous Jurassic and Cretaceous Pholadacea (Turner 1969). En- | dolithic bivalves evolved from various other primarily epifaunal or in- faunal bivalve stocks during the Cenozoic. These include the mytilid subfamilies Modiolinae and Crenellinae [including Botula, Gregariella, Fungiacava, and certain Modiolus (see Otter 1937, Yonge 1955, Gohar and Soliman 1963a, Keen 1971, and Goreau et al. 1972], certain Arcidae (Frizzell 1946, Olsson 1961), Tridacnidae (Purchon 1955b), Petricolidae (Otter 1937, Yonge 1958, Narchi 1975), Myidae (Yonge 1951), and Clavagellacea (Soliman 1971). In summary, the fossil record indicates that bivalves have invaded the endolithic habitat in a succession of adaptive radiations, i.e., in the Triassic (?) or Jurassic (by lithophagids and gas- trochaenids), in the Jurassic and Cretaceous (by the pholads and Hiatella), and in the Cenozoic (by representatives of several primarily nonendolithic stock groups). Insofar as chemical boring is generally considered more “specialized” than mechanical boring among modern bivalves, it is surprising that the 73 74 PEABODY MUSEUM BULLETIN 41 earliest adaptive radiation into the truly endolithic habitat apparently consisted of chemical borers. The modern Gastrochaenacea comprise chemically and combined chemically and mechanically boring forms, but their early representatives were probably largely chemical borers (see above). Based on their shell form, verified endolithic lithophagids have probably always been largely chemical borers. But the early evolution of chemical borers might be expected because chemical boring requires less specialization of shell form than mechanical boring. Except for their chemical boring apparatus and pedal structure, the early endolithic Spenglenia and Lithophaga may have retained much the same anatomy and shell form as their presumed shallow burrowing or epilithic nestling ances- tors. As suggested above, Spengleria may well have evolved its endolithic habit without greatly modifying its ancestral lateral profile, siphon struc- ture, and pedal musculature. Considering that lithophagids may have evolved their endolithic habit at least as early as the gastrochaenids, it is puzzling that they never evolved deep-burrowing, long-siphoned forms with comparable direc- tional mobility. In comparison with gastrochaenids, the Lithophaginae have remained unspecialized in terms of their simple grades of mantle fusion and siphon formation and their pedal structure (Yonge 1955; 1963). One may only speculate that the ancestors of the endolithic lithophagids differed from early gastrochaenids in possessing fili- branch rather than eulamellibranch gills. According to Stanley’s (1968) hypothesis, only the more efficient (i.e., eulamellibranch) gills would have been preadaptive for the evolution of long siphons, and hence deep burrowing. Gill pumping efficiency may have also permitted gas- trochaenids, but not lithophagids, to evolve narrow constrictions (i.e., baffles) in their burrow linings at the base of their siphons. Many early gastrochaenid burrows resemble fossil and Recent lithophagid burrows in their lack of siphonal baffles. But by Upper Cretaceous or early Tertiary time, siphonal baffles and medial diaphragms became common among representatives of both Eufistulana and Gastrochaena. This evolutionary innovation apparently occurred during or immediately fol- lowing the Upper Cretaceous radiation of naticid and muricid gas- tropods (see Soh] 1969). It is therefore interesting to speculate that siphonal baffles were adaptive for excluding the probosci of these in- creasingly important predators. Faced with this same increase in predation pressure, certain lithophagids may have evolved an alternative “baffle” against predatory gastropods in the form of thick posterior encrustations, as in the modern Lithophaga plumula (see Turner and Boss 1962, Soot-Ryen 1969). Unlike the lithophagid-gastrochaenid phase in the evolution of the endolithic habitat, the following Jurassic-Cretaceous phase saw the appearance of primarily mechanical rather than chemical or chemical- mechanical borers. The Mesozoic pholads became highly specialized for mechanical boring through their evolution of prominent rasping spines and a unique mechanism for rocking the valves about a dorsoventral axis (Purchon 1955a, 1956a; Evans 1968a,b; Ansell and Nair 1969). Hiatella (superfamily Hiatellacea) is a facultative borer-nestler found generally in temperate and colder waters (Hunter 1949; Keen 1969b). Insofar as the ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 75 Triassic (?) and Jurassic lithophagids and gastrochaenids were limited to carbonate substrata, they were largely restricted to warmer (i.e., carbo- nate rich) marine environments. In contrast, the mechanical boring mechanism of the pholads and Hiatella permitted their successful coloni- zation of a wider variety of substrata in cool and warm water marine environments. Interestingly, despite the Cretaceous and Cenozoic proliferation of pholads, the lithophagids and gastrochaenids maintained their domi- nance in terms of species-level diversity and population density in tropi- cal marine carbonate substrata. Conversely, despite the great ecologic and taxonomic diversity of modern pholads, they have evolved relatively few species in tropical corals and shells. Among modern pholads, the Western Atlantic Diplothyra smith Tryon and the Eastern Pacific Penatella conradi Valenciennes are endolithic in calcium carbonate substrata, and the latter has apparently evolved a combined chemical-mechanical boring mechanism (Smith 1968). But both taxa are more common in subtropical and cooler marine than in tropical marine environments (Turner 1954, 1955; Andrews 1971; Abbott 1974). Certain other modern pholads are occasionally found in tropical marine corals and shells (see Olsson 1961 and Abbott 1974), but these species seldom occur in densities comparable to gastrochaenids and lithophagids. This preliminary analysis suggests that the modern marine endolithic habitat is ecologically and taxonomi- cally partitioned largely between the lithophagids and gastrochaenids (ancestrally chemical borers; now inhabiting largely tropical marine coral and shell substrata) and the pholads and hiatellids (ancestrally mechani- cal borers; now inhabiting a number of substrata in a variety of marine temperature realms, but not as successful as gastrochaenids and lithophagids in tropical marine carbonates). One may only surmise that lithophagids and gastrochaenids maintained their dominance in tropical corals and shells because of their prior occupation of this habitat, and because of their competitive chemical or combined chemical and mechan- ical boring mechanisms. EROSIONAL INSTABILITY AND TROPICAL ENDOLITHIC COMMUNITIES. The same advantage of mantle fusion and siphon formation that enabled many Mesozoic bivalves to successfully inhabit unstable soft sediments (see Stan- ley 1968, 1972) probably contributed to the success of gastrochaenids in colonizing erosionally unstable hard substrata. The Jurassic Spengleria and Gastrochaena were undoubtedly limited in their directional mobility within the substratum. This is suggested by their generally straight or only slightly curving burrows, and by their generally short siphons. The siphons in some Jurassic Spengleria were relatively long, but their complete separation precluded abrupt departures from the initial boring direction. In many later Gastrochaena the siphons became much longer in relation to burrow length, and the inhalant and exhalant channels became fused into a single, narrow tube. In conjunction with this siphon streamlining, siphonal elongation permitted deeper burrowing and increased direc- tional mobility. Increased siphon length also permitted freedom of bur- row shortening in the case of erosional truncation, or burrow elongation in the case of threatened siphon overgrowth. 76 PEABODY MUSEUM BULLETIN 41 Studies of modern Western Atlantic tropical endolithic communities suggest that the evolution of directional mobility and siphonal elongation provided gastrochaenids with an ecological advantage over lithophagids for life in erosionally unstable substrata (Carter 1976). Certain other gastrochaenid features have likewise contributed to their success in thin and rapidly eroded coral and shell substrata. Probing of the hard sub- stratum by the pedal probing organ enables gastrochaenids to guide their burrows away from other borers and opposite coral or shell surfaces. Pedal probing holes are presently known in fossil gastrochaenids only in the Upper Cenozoic le.g., in the Gastrochaena (Gastrochaena) tube in Fig. 47]. But this adaptation may have evolved earlier in this superfamily, judging from the occurrence of pedal probing organs in both the modern Spengleria and Gastrochaena. Gastrochaenids also show an exceptional capacity for burrow repair and for anterior and posterior extension of their burrows in thin or overgrown substrata. Certain lithophagids are capable of repairing their broken burrows, and some even extend their siphonal burrow above the substratum by secreting calcareous laminae (e.g., Diplothyra smith Tryon 1862, North Carolina). But most lithophagids do not extend their siphonal burrows far beyond the sub- stratum, and none show a capacity for posterior burrow extension com- parable to Gastrochaena. The successful adaptation of gastrochaenids to erosionally unstable substrata is reflected in their spatial zonation in Diploria skeletons from Florida, Jamaica, and Bermuda. As described by Carter (1976) gas- trochaenids generally settle later in the cycle of coral disintegration when these substrata are more rapidly eroded. Gastrochaenids generally reach their highest population densities near the thinner, exposed coral mar- gins. In contrast, lithophagids commonly enter early in the cycle of coral disintegration when the upper coral surfaces are still protected by a com- plete cover of living polyps. In addition, many lithophagids reach their maximum population densities near the protected centers of the coral undersurfaces. As in soft sediment environments, erosional instability therefore appears to be an important factor influencing the spatial dis- tribution of bivalves in tropical endolithic communities. Asa final note, it is interesting to speculate that the adaptive radiation of the Triassic or Jurassic lithophagids and gastrochaenids may have been stimulated by the contemporaneous major expansion in tropical marine carbonate substrates. The Triassic and Jurassic periods saw the appear- ance and diversification of reef-building scleractinian corals. This evolutionary radiation culminated in the development of extensive her- matypic reefs in the Middle and Upper Jurassic (Wells 1969a,b). Accord- ing to Jackson et al. (1971) the cryptic habitats created by Jurassic reefs presented vast new areas comparatively free from competition with many common epifaunal taxa. These cryptic habitats and the coral substrata may have set the stage for the initial major radiation by the Bivalvia into the endolithic habitat. Taxonomic DisTINCTION OF Spengleria, Cucurbitula, Gastrochaena, AND Rocellaria. There is disagreement in the literature concerning the sub- ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA Cr generic (Tryon 1862, Dall 1898, Lamy 1925, Prashad 1932, Keen 1969a) versus generic (Pelseneer 1911, Olsson 1961, Boss 1967, Abbott 1974) taxonomic distinction accorded Spengleria Tryon 1862. Tryon (1862) proposed Spengleria as a new subgenus “to separate from Rocellaria s.s. those species which are elongate-cuneiform, truncated at the posterior end of the shell, and having a triangular space, radiating from the beaks posteriorly to the margin, elevated slightly above the general surface of the shell, and ornamented with transverse lamellae” (1862, p. 472). But Tryon’s diagnosis was not sufficient to exclude Gastrochaena truncata Sow- erby 1834, which has a truncated posterior and a triangular area of elevated periostracum. This species is properly regarded as Gastrochaena (Gastrochaena) because of its strong anterior reduction and prominent myophoral plates. Examination of G. (G.) truncata from Spondylus shells collected near the species type locality (the Bay of Panama) shows that its posterior raised triangular area consists of rugulose periostracum. Further- more, the siphons of G. (G.) truncata are fused for most of their length, as in the Western Atlantic G. (G.) hians. The other three species included by Tryon (1862) in Spengleria (i.e., G. mytiloides Lamarck 1818, G. plicatilis Deshayes 1855, and G. rostrata Spengler 1793) are properly regarded as representatives of Spengleria as presently emended. After thoroughly studying Spengleria mytiloides, Gastrochaena dubia, and G. macrochisma, Pelseneer (1911) suggested that the following charac- ters of Spengleria warrant its generic distinction from Gastrochaena: 1. Less anterior umbones. 2. Anterior adductor muscle more equal in size to the posterior adduc- tor than in Gastrochaena. 3. Completely separated siphons. 4. Plicated ctenidia. 5. Anterior point of the foot (the anterior pedal organ) reduced or absent in Spengleria. 6. Pedal protractor muscle present in Spengleria. The present study shows that the fifth and sixth characters are not unique to Spengleria. The anterior pedal organ is likewise small in Gastrochaena (Rocellaria) ovata, and both G. (R.) ovata and G. (G.) hians possess pedal protractor muscles. In order to better resolve the taxonomic distinction of Spengleria, data of shell morphology, anatomy and ecology for Spengleria rostrata, Gas- trochaena (G.) hians, and G. (Rocellaria) ovata are tabulated in Figure 66 and summarized in Figure 67. Excluding the features in common among all three species, S. rostrata resembles G. (G. ) hans only in its ligament, whereas S. rostrata and G. (R.) ovata are alike only in their common lack of certain specializations for mechanical boring unique to G. (G.) hians. Comparing Spengleria with the genus Gastrochaena (including Gas- trochaena s.s. and Rocellaria), it is apparent that many of their anatomical differences are directly or indirectly related to specializations in the latter for mechanical boring. Gastrochaena is more specialized for mechanical boring in terms of its increased anterior shell reduction, centralized pedal musculature, laterally expanded pedal gape, and (in most Recent species) 78 PEABODY MUSEUM BULLETIN 41 periostracal truncated| posterior |myophore | anterior |ligamental| ligament spikes posterior reduction nymph strength ie eae adductor pallial inequality sinus coarse |separated|} annular velar siphon |ctenidial | ctenidia tentacles | siphons |membrane flap length plication length (inhalant) ee ae a ee ees ee ee ae burrowing |size sorting rate in hindgut Fic. 66. Summary of characters varying among Spengleria rostrata, Gastrochaena (Rocellaria) ovata and G. (Gastrochaena) hians (indicated on the left by S., R. and G., respectively). Symbols: present (P), absent (A), strong or great (S), moderate or intermediate (M), weak or less (W), long (L) and short (sh). Ss G = sof \A = WAVY AV) siphons ecological and ctenidia total similarity characters Fic. 67. An abstract of data compiled in Figure 66, showing the unweighted similarities between species pairs of Spengleria rostrata (S.), Gastrochaena (Rocellaria) ovata (R) and Gastrochaena (Gastrochaena) hians (G.). Each line connecting the species indicates a similarity. ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 79 functional replacement of calcified periostracal spikes with comarginal shell ridges. ‘To accommodate anterior shell reduction, Gastrochaena has decreased the width of its anterior adductor muscle, and has shifted dorsoposteriorly the attachment site of its anterior pedal protractor mus- cles. Gastrochaena also differs from Spengleria in its fused siphons, flat ctenidia, and less numerous major siphonal tentacles. Considering also the distinct fossil record of Spengleria, these anatomical differences also cer- tainly warrant the generic level distinction of Spengleria from Gastrochaena. There is also disagreement in the literature concerning the taxonomic distinction accorded Cucurbitula Gould 1861. Gould (1861, p. 22) origi- nally based this taxon on a complex of distinctive characters, including: 1. A mantle that entirely envelopes the anterior shell valves. 2. The exterior ornament and curved, elongate shape of the shells. 3. The “artichoke-like” or bulbous structure of the “igloo,” which is made up of successive calcareous cups. Among subsequent authors, only Tryon (1862) and Kuroda et al. (1971) have followed Gould in recognizing the type species of Cucurbitula [Fis- tulana lagenula Lamarck] and similar forms as generically distinct from Gastrochaena. Fischer (1866) maintained that Cucurbitula should not be separate genus because its identity depends largely upon its “igloo”- forming habit. According to Fischer this habit occurs in many other gastrochaenids. Fischer’s (1866) recommendation was apparently fol- lowed by numerous subsequent workers, including Lynge (1909), Lamy (19225-1923; 1925); Olsson (1961), and Keen (1969a): Not all Cucurbitula secrete a capsule with an “artichoke-like” exterior, but capsule formation does appear to be obligatory here. Unlike Gas- trochaena s.s. and Rocellaria, Cucurbitula invariably secretes a calcareous capsule regardless of the substratum thickness. Cucurbitula can penetrate entirely through thinner substrata to construct a capsule on the surface opposite that initially settled. But even when settling thicker substrata, Cucurbitula nevertheless forms a capsule by changing its burrowing direc- tion and exiting on the surface initially penetrated. Because of its reflected mantle, ena elongate shell, and this obligatory capsule formation, Cucurbitula can properly be regarded as subgenerically distinct from Gas- trochaena s.s. and Rocellaria. Considering Gastrochaena s.s. and Rocellaria, the available data show their representatives to be similar in their ctenidia and siphons but diffe- rent in their degree of specialization for mechanical boring. Both taxa have undergone evolutionary specialization for mechanical boring, 1.e., by evolving anterior reduction, lateral and posterior expansion of the pedal gape, a centralized pedal musculature, and replacement of perio- stracal calcification with comarginal shell ridges. But Gastrochaena 1s further specialized for mechanical boring in terms of its greater anterior reduction and prominent myophores. Because these taxa differ only in this degree of specialization and because their species are morphologically quite similar in the earlier fossil record, they should be regarded as only subgenerically distinct. Inasmuch as Gastrochaena s.s. represents a grade of specialization for mechanical boring, further study of the fossil record may show this taxon to be polyphyletic. APPENDIX A LIST OF SOLDIER KEY BIVALVIA ASSOCIATED WITH DIPLORIA SKELETONS, WITH REFERENCES TO THEIR ILLUSTRATIONS IN OTHER WORKS. Endolithic species: Botula fusca Gmelin. Warmke and Abbott (1961) pl. 31, fig. d. Gastrochaena (Gastrochaena) hians (Gmelin). Morris (1973), pl. 32, fig. 1, listed as Rocellaria hians. Gastrochaena (Rocellaria) ovata (Sowerby). Warmke and Abbott (1961), pl. 44, fig. k, listed as Gastrochaena hians. Lithophaga antillarum Orbigny. Morris (1973), pl. 13, fig. 16. Lithophaga nigra Orbigny. Warmke and Abbott (1961), pl. 31, fig. m. Petricola lapicida Gmelin. Warmke and Abbott (1961), pl. 44, fig. e. Petricola typica (Jonas). Warmke and Abbott (1961), pl. 44, fig. b, listed as Rupellaria typica. Spengleria rostrata (Spengler). Warmke and Abbott (1961), pl. 44, fig. g. Semiendolithic species: Arca imbricata Bruguiere. Warmke and Abbott (1961), pl. 30, fig. e. Paramya subovata Conrad. Morris (1973), pl. 30, fig. 11. Epilithic species: Arca zebra (Swainson). Warmke and Abbott (1961), pl. 30, fig. 1. Barbatia cancellaria Lamarck. Warmke and Abbott (1961), pl. 30, fig. j. Barbatia candida Helbling. Warmke and Abbott (1961), pl. 30, fig. 1. Barbatia domingensis (Lamarck). Warmke and Abbott (1961), pl. 30, fig. d. Chama macerophylla Gmelin. Warmke and Abbott (1961), pl. 4, fig. c; pl. 3:7, fig. b: Came antillarum Orbigny. Warmke and Abbott (1961), text-fig. 31, fig. c, d. oprvnion radiatus (Anton). Warmke and Abbott (1961), pl. 32, fig. a. Lima lima (Linnaeus). Warmke and Abbott (1961), pl. 34, fig. f. Plicatula gibbosa Lamarck. Warmke and Abbott (1961), pl. 34, fig. g. 80 APPENDIX B TAXONOMIC OUTLINE The following taxonomic outline summarizes the generic versus sub- generic rank presently suggested for the gastrochaenid taxa discussed in this paper. Designations of genotypes are taken directly from the taxonomic review of this superfamily by Keen (1969a). This summary is not intended to represent a taxonomic revision of this superfamily, nor is this a particular endorsement of the genotypes outlined by Keen (1969a). In some instances, the genera and subgenera are accompanied by refer- ences to the more important taxonomic literature. i Lh IIl. FV: Genus Gastrochaena Spengler 1783. Type species: Gastrochaena cuneiformis Spengler 1783 by subsequent designation, Children 1822. See Kennard, Salisbury and Wood- ward (1931) for discussion of Children’s type designations, and see Keen (1969a). A. Subgenus Gastrochaena s.s. Spengler, 1783 (as presently re- stricted). Type species: Gastrochaena cuneiformis Spengler 1783. B. Subgenus Rocellaria de Blainville 1828 (issued in 1829). Type species: Gastrochaena modiolina Lamarck, 1818, by monotypy. C. Subgenus Cucurbitula Gould 1861. Type species: Fistulana lagenula Lamarck, 1801, by monotypy. See Olsson (1961) and Keen (1969a). Genus Spengleria Tryon 1862 (as presently restricted). Type Species: Gastrochaena mytiloides Lamarck, 1818, by subsequent designation, Stoliczka (1871; see his “Synoptical list” of “type- species”). Genus Eufistulana Eames 1951 [new name for Fistulana Bruguiere 1789 (1792)]. Type species: Gastrochaena mumia Spengler, 1783, by original des- ignation, Eames, 1951. Genus Kummelia Stephenson, 1937. Type species: Gastrochaena americana Gabb, 1860, by original des- ignation. Discussion: The names Kummelia and Polorthus are presently a source of taxonomic confusion. Turner (1966; 1969, p. N741) indicated that Gastrochaena americana Gabb 1860 is the genotype of Polorthus Gabb 1861 by Gabb’s original designation. Earlier, how- ever, Whitfield (1885, p. 203) expressed the opinion that Gabb (1861) founded the genus Polorthus on Teredo tibialis Morton. Teredo tibialis forms clusters of tubes, and on this basis is probably not at all closely related to Gastrochaena americana, i.e., to Kummelia as dis- 81 82 PEABODY MUSEUM BULLETIN 41 cussed in the present paper. Although Turner (1969) erred in indicating that Gabb (1861) designated G. americana as the type of Polorthus, her type designation must stand unless Whitfield’s (1885) discussion can be constructed as a previous type designa- tion. Clearly, if G. americana is the type of Polorthus, then this species is unavailable as the type of Kummelza. APPENDIXE DESCRIPTION OF NEW SPECIES Family Gastrochaenidae Gray, 1840 Shell fairly small, thin, more or less elongate, equivalve, inequilateral, prosogyrous, and widely gaping anteroventrally or along the entire vent- ral margin. Hinge edentulous or nearly so; ligament external and opis- thodetic and inserting on ligamental nymphs. Anterior adductor scars reduced relative to posterior ones. Siphons well developed; siphon forma- tion reflected in Recent and fossil representatives by a deep pallial sinus and an elongated posterior siphonal burrow. The siphonal and shell chambers of the burrow are generally clearly differentiable in casts of the burrow interiors. In many gastrochaenids these two chambers are sepa- rated by pointed “baffles” or by an annular constriction (diaphragm) produced by the calcareous burrow lining. Chemical and mechanical borers in calcium carbonate substrata, facultative tube-dwellers, and ob- ligatory tube-dwellers in tropical, subtropical and less commonly warm temperate waters. Genus Gastrochaena Spengler, 1783 Endolithic boring, facultative tube-dwelling, and obligatory “igloo” form- ing gastrochaenids with low to moderate elongation of the shells and anterior or nearly anterior umbones. Spikelike calcification of the perio- stracum, moderate lateral inflation of the shell anterior, a restricted an- teroventral pedal gape, and a flaring posterior siphonal burrow are primi- tive features, most of which are modified by Cenozoic time. Most Cenozoic species show restriction of periostracal calcification to the juvenile shell or a complete loss of this feature, greater lateral inflation of the shell anterior, and posterior and lateral expansion of the pedal gape. Subgenus Rocellaria de Blainville, 1828 (1829) Endolithic and facultative tube-dwelling gastrochaenids with relatively short siphons and umbones lying near but not at the anterior shell margin. The hinge line is straighter and somewhat thicker than in the other subgenera of Gastrochaena (i.e., Gastrochaena s.s. and Cucurbitula), and is 83 84 PEABODY MUSEUM BULLETIN 41 commonly horizontally flattened toward the anterior shell margin. The hinge line forms a sharp angle with the anteroventral shell margin (as viewed from the side), except in specimens showing extreme abrasion on this part of the shell. The exterior ornament commonly consists of distinct and rather regular comarginal ridges, and the posterior periostracum is generally thin and inconspicuous. The burrow generally takes the form of a simple flask-shaped tube without a prominent diaphragm or “baffles” at the base of the siphons. Some species show an elongated, horizontally flattened area on the anterior of the hinge extending in an anterior- posterior direction, and this may be developed posteriorly into a small, triangular myophore. Other species show more irregular thickenings at the attachment sites of the anterior pedal retractor muscles. But most species lack distinct, spatulate myophores projecting prominently into the umbonal cavity. Gastrochaena (Rocellaria) linsleyi, n. sp. Fig. 57, A-F DEscRIPTION. Shell small (mean length of 4 specimens 4.9 mm, ranging from 4.3 to 6.1 mm), elliptical posteriorly and widely gaping anteroven- trally (Fig. 57A). Umbones prosogyrous and lying near but not at the anterior shell margin. Shell valves ornamented with concentric ridges except on the prodissoconchs, which appear as distinct, smooth, shiny caps on the umbones. Anteriorly and laterally the concentric ridges are studded with numerous conical spikes aligned in concentric to slightly oblique rows and imbedded within or cemented to the shell exterior. These spikes (presumably periostracal in origin) are absent from the shell posterior, but may originally have been present in the organic periostracum in this part of the shell. The external opisthodetic ligament inserts on two well- developed ligament nymphs that extend nearly 1/4 the shell length from the umbo toward the shell posterior. These nymphs and the anterior edentulous hinge form an angle of about 16 degrees with the anteropos- terior shell axis. The shell anterior is only slightly laterally inflated, and the anteroventral pedal gape is restricted to the shell anterior, extending less than one-half the shell length toward the posterior. The juvenile valves show a broad, radial furrow in the midlateral position extending ventrally from the edge of the prodissoconch. This furrow does not persist into the adult growth stage and is consequently visible in larger shells only near the umbones (Fig. 57A). The shell structure is branching crossed lamellar near the shell margins and irregularly crossed lamellar toward the shell interior (see Carter 1976b for definitions). An outer prismatic shell layer may be present, but this has not yet been confirmed by sectioning. The shell interior appears glossy white and shows no pedal, adductor or pallial muscle attachment scars. The burrow of G. (R.) linsleyi shows distinct siphonal and shell cham- bers, but these are not separated by a calcareous annular ring or by pointed baffles projecting from the burrow lining (Fig. 57B—F). Latex casts of the ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 85 burrow interior indicate short, presumably fused inhalant and exhalant siphon tubes. The orientations of the siphons indicate that the shell com- missure plane was essentially parallel to the exterior surface of the Cucul- laea substratum. The burrow shell chamber is elliptical and relatively spacious in comparison with the shell dimensions, so it would have permit- ted both rotational and slight anterior-posterior shell movement. Types. Holotype: YPM 10216a. Type locality: Coon Creek, 250 yards east of Dave Weeks’ house, 3% miles south of Enville, Tennessee. This is the “Dave Weeks place” locality described by Wade (1926, p. 9). Stratigraphic position: Upper Cretaceous. Additional specimens: YPM 10216b-—d. MatTeERIAL. The species is based on four specimens, one of which (the holotype, Fig. 57A) is nearly perfect and is accompanied by a latex cast of the burrow interior. All four specimens are permanently deposited in the Yale Peabody Museum. G. (R.) linsleyi and several associated endolithic species (Lithophaga conchafondentis Gardner, Martesia procurva Wade, an endolithic sponge, and possibly Lithophaga ripleyana Gabb) were found boring into the ventrolateral exterior surfaces of an articulated shell of Cucullaea vulgaris Morton. OccurRRENCE. Shells of this species are presently known only from the type locality. The distinctive burrows of G. (R.) linsleyi observed in the upper valves of Exogyra costata Say from the Prairie Bluff Chalk, 2.8 miles south of the intersection of Highways 21 and 263, northwest of Green- ville, Alabama. Comparisons. The burrow of G. (R.) linsleyi resembles the Cretaceous G. dilatata Leymerie 1842 in its short siphonal burrow and more or less horizontal orientation of the burrow shell chamber relative to the surface of the substratum. But judging from Reuss’ (1845-1846) drawing of G. dilatata (see “Fistulana dilatata d’Orbigny,” Reuss’s plate 37, fig. 9) its siphonal burrow differs from G. (R.) linsleyi in showing no evidence of distinct inhalant and exhalant tubes. The Upper Cretaceous Gastrochaena ostreae (Geinitz) shows distinct inhalant and exhalant tubes in the siphonal burrow (see plate I, figs. 1 and 3 in Zazvorka 1943) but its siphons are separated at their base by a short extension of the posterior burrow shell chamber. In addition, Zazvorka’s (1943) description of G. ostreae indicates that this species is considerably larger than G. (R.) linsleyi. G. ostreae and G. (R.) linsley: similarly show a radial furrow in the shell valves, but this is more pronounced in the former and persists into the adult stage. G. (R.) linsleyi differs from most Cenozoic gastrochaenids in its relatively restricted an- teroventral pedal gape and relatively subdued lateral inflation in the shell anterior. Discussion. The lateral compression, restricted pedal gape, and calcified periostracal spikes in G. (R.) linsleyi are primitive features that strengthen the hypothesis of evolution of Gastrochaena (Rocellaria) from Spengleria. G. (R.) linsleyi represents an early stage in the evolution of Gastrochaena in which calcified periostracal spikes were functional for mechanical boring throughout life. Most Cenozoic Gastrochaena utilize projecting comarginal 86 PEABODY MUSEUM BULLETIN 41 ridges on the valve margins to abrade the substratum often without the aid of projecting periostracal spikes. It is interesting that G. (R.) linsleyy shows an ontogenetic loss of the radial furrow in the shell, since a similar furrow is found in adults of the Upper Cretaceous G. ostreae (Geinitz). G. (R.) linsleyi is assigned to the subgenus Rocellaria on the basis of its simple hinge line, nearly anterior but not terminal umbones, relatively short siphons, and simple burrow lining at the base of the siphons. It may be noted, however, that the delicate hinge and relatively subdued concen- tric ornament of G. (R.) linsley: are more typical of the modern species of Gastrochaena s.s. than Rocellaria. The naming of this species is dedicated to Robert M. Linsley of Colgate University. LITERATURE CITED Abbott, R. T. 1974. American Seashells, 2nd ed., Van Nostrand, New York. 663 pp., 6405 figs. Aldrich, T. H. 1903. Two new species of Eocene fossils from the lignitic of Alabama. Nautilus 17: 19-20. Allen, J. A. 1958. Observations on Cochlodesma praetenue (Pulteney) [Eulamellibranchia]. J. Mar. Biol. Assoc. U. K. 37: 97-112. Aller, R. C. 1974. Prefabrication of shell ornamentation in the bivalve Laternula. Lethaia 7: 43-56. Andrews, J. 1971. Sea Shells of the Texas Coast. Univ. Texas Press, Austin, Tex. 298 pp. Ansell, A. D. and N. B. Nair 1969. A comparative study of bivalves which bore mainly by mechanical means. Amer. Zool. 9: 857-868. Arakawa, K. Y. 1970. Scatological studies of the Bivalvia (Mollusca). Adv. Mar. Biol. 8: 307-436. Arkell, W. J. 1929-1937. A monograph of British corallian Lamellibranchia. London, Palaeontol. Soc., London, 1936. i-xxxviii + 392 pp., 56 pls. Atkins, D. 1937. On the ciliary mechanisms and interrelationships of lamellibranchs. Part III. Types of lamellibranch gills and their food currents. Q. J. Microsc. Sci. 79: 375-421. Bertram, G. C. L. 1936. Some aspects of the breakdown of coral at Ghardaqa, Red Sea. Proc. Zool. Soc. London [1936]: 1011-1026, 4 figs., 1 pl. Blainville, H. M. D. de 1828 (1829). See vol. 57, p. 244 in F. G. Cuvier [ed.] Dictionnaire des sciences naturelles . . . Suivi d’un biographie des plus célébres naturalistes . . . par plusieurs professeurs du Jardin du Roi, et des principales écoles de Paris, 61 vols.; Strasbourg, Paris. Boss, K. J. 1967. On the evolution of Spengleria (Gastrochaenidae; Bivalvia). Amer. Malacol. Union Ann. Rep. for 1967: 15-16. 1971. Critical estimate of the number of Recent Mollusca. Occas. Pap. Mollusks Mus. Comp. Zool. Harvard Univ., 3 (40): 81-135. Brocchi, G. 1814. Conchiologia Fossile Subapennina con Osservazione Geologiche sugli Apenninie sul Suolo Adiacente. Milano, Stamperia Reale. Tomo secondo. 1012 pp., 16 pls. Bruguiére, J. 1789. Tableau Encyclopédique et Méthodique. [Uncompleted text, pp. 85-132, and pls. 96-286 were published in 1972.] Vol. I, fasc. 2: p xii [1792], pl. 167, fig. 17-22. Buvignier, A. 1843. Sur quelques fossiles nouveaux des departements de la Meuse et des Ardennes, Mém Soc. Philomat. Verdun. 2: 226-255, pls. II-VI. 1852. Statisque géologique, minéralogique, minérallurgique et paléontologique, du départe- ment de la Meuse. J. B. Bailliére, Paris. pp. ili and 1-694 (text): Atlas, pp. 1-532 and 32 pls. Cailliaud, F. 1856. Mémoire sur les Mollusques perforants. Nat. Verh. Bataaf. (Koning. Hollandsche) Maatsch. Wet. Haarlem, Verz. 2, 11: 1-58, 3 pls. Carazzi, D. 1903. Contributo all’istologia e alla fisiologia dei Lamellibranchi. Int. Monatschr. Anat. Physiol. 20: 57-90, 2 pls. Carter, J. G. 1976. The ecology of a tropical Western Atlantic endolithic bivalve community. Geol. Soc. Amer. Abstr., Vol. 8, no. 2: 148-149. Carter, J. G. and R. C. Aller 1975. Calcification in the bivalve periostracum. Lethaia 8: 315-320, 2 figs. Children, J. G. 1822 (1822-1823). Lamarck’s Genera of Shells. Quart. J. Sci. Lit. Arts, Vol. XIV, No. 27, for Oct. 1822, pp. 64-86; No. 28, for Jan. 1823, pp. 298-322; pls. II-VI. Chronic, H. 1952. Molluscan fauna from the Permian Kaibab Formation, Walnut Canyon, Arizona. Geol. Soc. Amer. Bull. 63: 95-166, 11 pls. SS aer M. 1908. Pélécypodes du Montien de Belgique. Extraits Mém. Mus. Roy. d’Hist. Nat. Belg. 5: 1-76, 8 pls. Cossmann, M. and A. Peyrot 1909. Conchologie néogénique de |’Aquitaine. Bordeaux Actes Soc. Linn. 63: 73-293, 7 text-figs., 7 pls. Crenshaw, M. A. and J. M. Neff 1969. Decalcification at the mantle-shell interface in molluscs. Amer. Zool. 9: 881-885, 6 figs. Dall, W. H. 1898. Contributions to the Tertiary fauna of Florida; with especial reference to the Miocene silex-beds of Tampa and the Pliocene beds of the Caloosahatchie River. Part IV. Trans. Wagner Free Inst. Sci. Philadelphia 3: i-vii and 571-947, pls. 23-35. Davis, C. A. 1903. Notes on the Mollusca of the Bermuda Islands. Nautilus 17: 125-130. Deshayes, G. P. 1846. Examen anatomique du gastrochene de la Mediterraneé. Contrib. R. Acad. Sci. Paris 22: 37-38. 1855. Descriptions of new shells from the collection of Hugh Cuming, Esq. Proc. Zool. Soc. London 1854: 317-371. 1857 (1856-1866). Description des animaux sans vertebres decouverts dans le bassin de Paris 87 88 PEABODY MUSEUM BULLETIN 41 pour servir de supplement a la description des coquilles fossiles des environs de Paris, comprenant une revue générale de toutes les espéces actuellement connues. J.-B. Balliere, Paris, 3 vol. text., 2 vol. atlas, 912 pp. 187 pls. See pp. 81-392 (published in 1857). Diener, C. 1923. Lamellibranchiata triadica. Part 19, pp. 1-259. Jn Fossilium Catalogus, I: Animalia. W. Junk, Berlin. Dinamani, P. 1967. Variation in the stomach structure of the Bivalvia. Malacologia 5: 225-268. Dugal, L. P. 1939. The use of calcareous shell to buffer the product of anaerobic glycolysis in Venus mercenaria. J. Cell. Comp. Physiol. 13: 235-251. Duval, D. M. 1963. The comparative anatomy of some lamellibranch siphons. Proc. Malacol. Soc. London 35: 289-295, 12 figs. Eames, F. E. 1951. A contribution to the study of the Eocene in Western Pakistan and Western India. B. The description of the Lamellibranchia from standard sections in the Rakhi Nala and Zinda Pir areas of the Western Punjab and in the Kohat district. Phil. Trans. R. Soc. London Vol. 235, Ser. B: 311-482, 9 pls. Elias, M. K. 1957. Late Mississippian fauna from the Redoak Hollow Formation of southern Oklaho- ma, Part 3, Pelecypoda. J. Paleontol. 31: 737-784, 1 text-fig., pls. 89-97. Eudes-Deslongschamps, M. 1838. Mémoire sur les coquilles fossiles lithophages des terrains secon- daires du Calvados. Mem. Soc. Linn. Normandie, Ser. 2, Vol. 6: 220-229, pl. IX. Evans, J. W. 1968a. The effect of rock hardness and other factors on the shape of the burrow of the rock-boring clam Penitella penita. Palaeogeog. Palaeoclim. Palaeoecol. 4: 271-278. 1968b. Factors modifying the morphology of the rock-boring clam Penztella penita (Conrad, 1837). Proc. Malacol. Soc. London 38: 111-119, 4 pls. 1968c. Observations on the borers in the shell of the bivalve Placopecten magellanicus. Amer. Zool. 8: 795. Fischer, P. H. 1866. Anatomie des Fistulanes. J. Conchyl. 14: 321-335, 2 pls. Forbes, E. and S. Hanley 1853 (1848-1853). A History of the British Mollusca and Their Shells. London J. Van Voorst. Vol. 1, Ixxx+486 pp.; Vol. 2, 557 pp.; Vol. 3, 616 pp.; Vol. 4, 302 pp., A-SSS+ 133 pls. Frebold, H. 1933. Weitere Beitrage zur Kenntnis des Oberen Palaozoikums Ostgronlands. I. Die Fauna und stratigraphische Stellung der oberpalaozoischen “Weissen Blécke” (Kap Stosch Forma- tion) Ostgrénlands. Medd. Grgnland 84: 3-61, pls. i-iv. Frizzell, D. L. 1946. A study of two arcid pelecypod species from western South America. J. Paleontol. 20: 38-51, pl. 10. ' Fursich, F. T. and T. J. Palmer 1975. Open crustacean burrows associated with hardgrounds in the Jurassic of the Cotswolds, England. Proc. Geol. Assoc. 86: 171-181, 5 figs. Gabb, W. M. 1860. Descriptions of new species of American Tertiary and Cretaceous fossils. J. Acad. Nat. Sci. Philadelphia, Ser. 2, Vol. 4: 375-406, pls. 67-69. 1861. Notes on Cretaceous fossils with descriptions of a few additional new species. Proc. Acad. Nat. Sci. Philadelphia, Nov., 1861. pp. 363-367. Gardner, J. A. 1928. The molluscan fauna of the Alum Bluff Group of Florida. Part V. Tellinacea, Solenacea, Mactracea, Myacea, Molluscoidea. U.S. Geol. Surv. Prof. Pap. 142E: 185-249, pls. 29-36. Glibert, M. 1936. Faune malacologique des sables de Wemmel, I. Peleécypodes. Mém. Mus. Roy. d’Hist. Nat. Belg, 78: 1-241, 75 figs., 7 pls. Gmelin, J. F. 1791. Caroli a Linné Systema Naturae per regna tria naturae. Editio decima tertia. Leipzig, Germany, Vol. 1, Part 6, Cl. 6, Vermes, pp. 3021-3910. Gohar, H. A. F. and G. N. Soliman 1963a. On two mytilids boring in dead coral. Pub. Mar. Biol. Sta. Ghardaga, Red Sea, 12: 205-218. 1963b. On three mytilid species boring in living corals. Pub. Mar. Biol. Sta. Ghardaqa, Red Sea, 12: 65-98, 18 figs. 1963c. On the rock-boring lamellibranch Rocellaria riippelli (Deshayes). Pub. Mar. Biol. Sta. Ghardagqa, Red Sea, 12: 145-157. Goreau, T. F., N. I. Goreau, and C. M. Yonge 1972. On the mode of boring in Fungzacava eilatensvs (Bivalvia: Mytilidae). J. Zool. London 166: 55-60, 2 pls. Gould, A. A. 1861. Descriptions of shells collected by the North Pacific exploring expedition (con- tinued). Proc. Boston Soc. Nat. Hist. 8: 14-40. Gray, J. E. 1840. Synopsis of the Contents of the British Museum. 42nd ed., British Museum, London. 370 pp. 1847. A list of the genera of Recent Mollusca, their synonyms and types. Proc. Zool. Soc. London 15: 129-219. Haas, F. 1935. Bivalvia. Band 3, Abt. 3, Teil I, xii: 1-9847nm H. G. Bronn [ed.], Klassen und Ordnungen des Tierreichs, 562 figs. (See pp. 865-984). Hancock, A. 1848. On the boring of the Mollusca into rocks, and on the removal of portions of their shells. Annals and Mag. Natur. Hist. 2 (2): 225-248, 1 pl. Haneda, Y. 1939. Luminosity of Rocellaria grandis (Deshayes) (Lamellibranchia). Kagaku-Nanyo (South Sea Science) 2: 36-39, 6 figs. Harris, G. D. 1896. The Midway Stage. Bull. Amer. Paleontol. 1(4): 156 pp., 15 pls. Holzapfel, E. 1889. Die Mollusken der Aachener Kreide. 3. Classe Lamellibranchiata. Palaeontogr. 35: 139-263, pls. 8-29. ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA 89 Hunter, W. R. 1949. The structure and behavior of Hiatella gallicana (Lamarck) and H. arctica (L.) with special reference to the boring habit. Proc. R. Soc. Edinburgh, Ser. B, 63: 271-289. Jaccarini, V., W. H. Bannister, and H. Micallef 1968. The pallial glands and rock boring in Lithophaga lithophaga (Lamellibranchia, Mytilidae). J. Zool. London 154: 397-401. Jackson, J. B. C., T. F. Goreau, and W. D. Hartman 1971. Recent brachiopod-coralline sponge communities and their paleoecological significance. Science 171: 623-625, 2 figs. Jeffreys, J. G. 1865. British Conchology, III. Marine Shells. J. van Voorst, London. 394 pp. Keen, A. M. 1969a. Superfamily “Gastrochaenacea,” pp. N700-N702in R. C. Moore [ed.] Treatise on Invertebrate Paleontology, Part N, Mollusca 6. Geol. Soc. Amer. and Univ. Kansas Press. 1969b. Superfamily “Hiatellacea,” pp. N699-N700 in R. C. Moore |ed.] Treatise on Inverteb- rate Paleontology, Part N, Mollusca 6. Geol. Soc. Amer. and Univ. Kansas Press. 1971. Sea Shells of Tropical West America. Stanford Univ. Press, 2nd ed., 1064 pp. Kennard, A. S., A. E. Salisbury, and B. B. Woodward 1931. The types of Lamarck’s genera as selected by J. G. Children in 1823. Smithson. Misc. Coll. 82, No. 17, 40 pp. Koenen, A. von 1894. Das Norddeutsche Unter-Oligocan und seine Mollusken-Fauna. Leiferung VI. Revision der Mollusken-Fauna des Samlandischen Tertiars. Abh. Geol. Spec. Preussen und den Thuring. Staat. 10 (6): 1249-1392, pls. 87-99. Kuhnelt, W. 1930. Bohrmuschelstudien I. Paleobiologica 3: 51-91. 1934. Bohrmuschelstudien II. Paleobiologica 5: 371-405. Kuroda T., T. Habe, and K. Oyama 1971. The Sea Shells of Sagami Bay. Edited by Biol. Lab., Imperial Household, Maruzen Co., Ltd., Tokyo. Lacaze-Duthiers, H. de 1883. Morphologie des Acéphales. Premier Mémoire. Anatomie de I’ Arrosoir. (Aspergillum dichotomum, L. Reeve). Arch. Zool. Exp. Gen. (2) 1: 1-68. Lamarck, J. B. P. A. de M. de 1801. Systeme des animaux sans vertebres ou table général des classes, des ordres, et des genres de ces animaux. Paris. viii + 432 pp. 1818 (1815-1822). Histoire naturelle des animaux sans vertebres . . . précedée d'une introduc- tion offrant la détermination des caractéres essentiels de l’animal, sa distinction du vegetal et des autres corps naturels, enfin, l’exposition des principes fondamentaux de la zoologie. Paris, Vols. 1-7. See Vol. 5 (publ. 1818), 612 pp. Lamy, E. 1922. Notes sur les espécies Lamarckiennes appartenant a la familie des Gastrochaenidae. Bull. Mus. d’Hist. Nat., Paris 28: 307-311. 1923. Les gastrochénes de la Mer Rouge. Bull. Mus. Nat. d’Hist. Natur. 29: 391-395. 1925. Révision des Gastrochaenidae vivants du Muséum National d'Histoire Naturelle de Paris. J. de Conchyl. 68: 284-319. Laseron, C. F., and J. Laseron 1952. A shell that builds a house. Proc. R. Zool. Soc. New South Wales, 1951-1952: p. 39. Lea, H. C. 1846. Descriptions of some new fossil shells from the Tertiary of Petersburg, Virginia. Trans. Amer. Phil. Soc., n.s., 9: 229-274, pls. 34-37. Leymerie, A. 1842. Suite du mémoire sur le terrain crétacé du département de l’Aube. Mem. Soc. Geol. France. 5 (1): 1-34, pls. 1-13. Linnaeus, C. [Linné] 1758. Systema Naturae, per Regna Tria Naturae, Secundum Classes, Ordines, Genera, Species, cum Characteribus, Differentiis, Synonymis, Locis. 10th ed., Holmiae, 2 vols., 824 Loriol, C. L. P. de 1891. Etudes sur les mollusques des couches corraligeénes inférieurs de Jura Bernois. III Partie. Mém. Soc. Paléontol. Suisse 18: 175-258, 20 pls. Loriol, C. L. P. de and E. Bourgeat 1888. Etudes sur les mollusques des couches coralligenes de Valfin (Jura). III Partie. Mém. Soc. Paléontol. Suisse 15: 225-364, 14 pls. Lynge, H. 1909. The Danish expedition to Siam, 1899-1900, Part IV, Marine Lamellibranchiata. Det Kongel. Dansk. Vid. Selsk. Skr. 7, Rack. Nat.-Urv. og Mathem., Afd. 3, Vol. 4: pp. 1-299, pls. 1-5. Merla, G. 1931. La fauna del Calcare a Bellerophon della regione dolomitica. Padua, Universita di Padova, Istituto Geologico, Memorie 9: 1-221, pls. i-X1. Morris, P. A. 1973. A Field Guide to Shells of the Atlantic and Gulf Coasts and the West Indies. 3rd. ed., W. J. Clench [ed.], Houghton Mifflin Co., Boston. i-xxvili+330 pp., 76 pls. Nair, N. B. and A. D. Ansell 1968. The mechanism of boring in Zirphaea crispata (L.) (Bivalvia: Pholadidae). Proc. Roy. Soc. London Ser. B, 170: 155-173, 10 figs. Nakazawa, K. and N. D. Newell 1968. Permian bivalves of Japan. Mem. Fac. Sci. Kyoto Univ. Ser. Geol. Mineral. 35: 1-108, 11 pls. Narchi, W. 1975. Functional morphology of a new Petricola (Mollusca, Bivalvia) from the littoral of Sao Paulo, Brazil. Proc. Malacol. Soc. London 41: 451-465. Newell, N.D. and A. La Rocque 1969. Family “Grammysiidae,” pp. N819-N823 in R. C. Moore led. ] Treatise on Invertebrate Paleontology, Part N, Mollusca 6. Geol. Soc. Amer. and Univ. Kansas Press. Oberling, J. J. 1964. Observations on some structural features of the pelecypod shell. Mitt. Natur- forsch. Ges. Bern, 20: 1-63, 6 pls., 3 figs. Olsson, A. A. 1961. Mollusks of the tropical Eastern Pacific, particularly from the southern half of the Panamic-Pacific Faunal Province (Panama to Peru). Panamic-Pacific Pelecypoda. Paleontol. Res. Inst., New York, 574 pp. 86 pls. Orbigny, A. d’ 1844 (1843). Paléontologie Francaise. Terrains Crétacés. G. Masson, Paris, Vol. 3, Mollusques, 807 pp., pls. 237-489. 90 PEABODY MUSEUM BULLETIN 41 Otter, G. W. 1937. Rock-destroying organisms in relation to coral reefs. Sci. Rep. Great Barrier Reef Expedition 1: 323-352, 3 figs., 2 pls. Palmer, K. V. W. and D. C. Brann 1965-1966. Catalogue of the Paleocene and Eocene Mollusca of the southern and eastern United States. Bull. Amer. Paleontol. 218: 1-1058, 5 pls. Palmer, T. J. and F. T. Fursich 1974. The ecology of a Middle Jurassic hardground and crevice fauna. Palaeontology 17: 507-524, pls. 75-77. Pannella, G. and C. MacClintock 1968. Biological and environmental rhythms reflected in molluscan shell growth. J. Paleontol. 42 (5) Supplement: 64-80. Pelseneer, P. 1911. Les lamellibranches de l’expédition du Siboga. Partie anatomique. Siboga-Expéd. Monogr., Part 61, Vol. 53a, 125 pp., 26 pls. (Leiden). Phillips, J. 1829. Illustrations of the geology of Yorkshire; or a description of the strata and organic remains of the Yorkshire coast: accompanied by a geological map, sections, and plates of the fossil plants and animals. York, T. Wilson and Sons. xvi + 192 pp., 23 pls., map. Pojeta, J. Jr.and T. J. Palmer 1976. The origin of rock boring in mytilacean pelecypods. Alcheringa 1: 167-179, 5 figs. Prashad, B. 1932. The Lamellibranchia of the Siboga Expedition (exclusive of the Pectinidae). Systematic part II. Pelecypoda. Siboga-Exped. Mon. 53c: 353 pp., 9 pls. E. J. Brill, Ltd., Leiden. Purchon, R. D. 1954. A note on the biology of the lamellibranch Rocellaria (Gastrochaena) cuneiformis Spengler. Proc. Zool. Soc. London 124: 17-33, 11 figs. 1955a. The structure and function of the British Pholadidae (rock-boring Lamellibranchia). Proc. Zool. Soc. London 124: 859-909, 19 figs. 1955b. A note on the biology of Tridacna crocea Lam. Proc. Malacol. Soc. London 31: 95-110. 1956a. A note on the biology of Martesia striata L. (Lamellibranchia). Proc. Zool. Soc. London 126: 245-258, 7 figs. 1956b. A note on the biology of Brechites penis (L.) (Lamellibranchia). J. Linn. Soc. London Zool. 43: 43-54, 8 figs. 1958. The stomach in the Eulamellibranchia; stomach type IV. Proc. Zool. Soc. London 131: 487-525, 14 figs. 1960. A further note on the biology of Brechites penis (L.) (Lamellibranchia). Proc. Malacol. Soc. London 34: 19-23, 1 fig. Reuss, A. E. 1845-1846. Die Versteinerungen der bohmischen Kreideformation, mit Abbildungen der neuen oder weniger bekannten Arten, gezeichnet von Joseph Rubesch. Stuttgart, E. Schweizer- bart’sche Verlagsbuchhandlung, 1-148, 51 pls. Richards, H. B. et al. 1958. The Cretaceous fossils of New Jersey. Part 1. New Jersey Bur. Geol. Topogr., Paleontol. Ser., 1-266, 46 pls. Runnegar, B. 1972. Anatomy of Pholadomya candida (Bivalvia) and the origin of the Pholadomyidae. Proc. Malacol. Soc. London 40: 45-58, 3 figs., 2 pls. 1974. Evolutionary history of the bivalve subclass Anomalodesmata. J. Paleontol. 48: 904-939, 10 figs., 5 pls. Runnegar, B. and P. A. Jell 1976. Australian Middle Cambrian molluscs and their bearing on early molluscan evolution. Alcheringa 1: 109-138, 11 figs. Runnegar, B. and N. D. Newell 1971. Caspian-like relict molluscan fauna in the South American Permian. Bull. Amer. Mus. Natur. Hist. 146: 1-66, 27 figs. Schroeder, J. H. 1972. Fabrics and sequences of submarine carbonate cements in Holocene Bermuda cup reefs. Geol. Rund. 61: 708-730. Sluiter, C. P. 1890. Uber die Bildung der Kalkréhren von Gastrochaena. Natuurk. Tijdschr. Ned. Ind. Batavia Deel 50: 45-60, 1 pl. Smith, E. A. 1907. Note on Fistulana mumia perforating a valve of a Dosinia. Proc. Malacol. Soc. London 7: 203, 2 figs. 1911. Nore on the very young stage of the genus Humphreyia. Proc. Malacol. Soc. London 9: 23-25, 2 figs. Smith, E. H. 1968. Functional morphology of Penitella conradi relative to shell penetration. Amer. Zool. 85.799. Smith, F. G., R. H. Williams, and C. C. Davis 1950. An ecological survey of the subtropical inshore waters adjacent to Miami. Ecology 31: 119-146. Smith, L. A. 1962. Revision of the Clavagellacea. Veliger 4: 167-174. Sohl, N. F. 1969. The fossil record of shell boring by snails. Amer. Zool. 9: 725-734, 15 figs. Soliman, G. N. 1969. Ecological aspects of some coral-boring gastropods and bivalves of the north- western Red Sea. Amer. Zool. 9: 887-894. 1971. Ona new clavagellid bivalve from the Red Sea. Proc. Malacol. Soc. London 39: 389-397, 2 figs., 1 pl. 1973. On the structure and behavior of the rock-boring bivalve Rocellaria retzii (Deshayes) from the Red Sea. Proc. Malacol. Soc. London 40: 313-318, 2 figs., 1 pl. Soot-Ryen, T. 1969. Superfamily “Mytilacea,” pp. N271-N280 in R. C. Moore [ed.] Treatise on Invertebrate Paleontology, Part N, Mollusca 6. Geol. Soc. Amer. and Univ. Kansas Press. Sowerby, G. B. (The first) 1834. Characters of new genera and species of Mollusca and Conchifera, collected by Mr. Cuming. Proceedings for March 25, 1834. Proc. Zool. Soc. London, Part II. 1834, pp. 21-22. ECOLOGY AND EVOLUTION OF THE GASTROCHAENACEA oH Sowerby, G. B. (The third) and H. C. Fulton 1903. Note on a specimen of Fistulana clava, Lamk., perforating a shell of Mitra interlirata, Reeve. Proc. Malacol. Soc. London 5: 345 and pl. 16, figs. 8-9. Spengler, L. 1783. Beskrivelse over en nye slaegt af toskallede muskeler, som kan kaldes Gastrochaena, 1 tre foranderlige Arter, hvoraf hver boer i et forskielligt drmehuus. Nye Samling K. Dan. Vidensk. Selsk. Skr. 2: 174-183, 17 figs. 1793. Beskrivelse over el nyt slaegt af de toskallede konkylier forhen af mig kaldet Chaena saa og over det Linneiske slaegt Mya, hvilket ngiere bestemmes, og inddeles i tvende slaegter. Skrivter af Naturhistorie-Selskabet 3: 16-26. Stanley, S. M. 1968. Post-Paleozoic adaptive radiation of infaunal bivalve molluscs—a consequence of mantle fusion and siphon formation. J. Paleontol. 42: 214-228, 13 figs. 1972. Functional morphology and evolution of byssally attached bivalve mollusks. J. Paleontol. 46: 165-188. Stasek, C. R. 1963. Synopsis and discussion of the association of ctenidia and labial palps in the bivalved Mollusca. Veliger 6: 91-97, 5 figs. Stephenson, L. W. 1937. The stratigraphic significance of Kummelia, anew Eocene bivalve genus from New Jersey. J. Washington Acad. Sci. 27: 58-64. 1941. The larger invertebrate fossils of the Navarro Group of Texas. Univ. Texas Publ. 4101: 1-641, 95 pls. Stoliczka, F. 1871 (1870-1871). Cretaceous fauna of Southern India. Vol. III. The Pelecypoda, with a review of all known genera of this class, fossil and recent. Mem. Geol. Surv. India, Paleontol. Indica, ser. 6, Vol. 3: xxii+535 pp., 50 pls., Calcutta. [pp. 1-222, pls. I-XII, published 1870; pp. 223-535, pls. XIII-L, published 1871]. Sturany, R. 1899. Lamellibranchiaten des Rothen Meeres. Expedition S. M. Schiff “Pola” in das Rothe Meer. Denkschr. Kaiser. Akad. Wissen. Math.-Natur. 69: 255-295, 7 pls. Taylor, J. D. 1973. The structural evolution of the bivalve shell. Palaeontology 16: 519-534, 5 text-figs., 1 pl. Taylor, J. D., W. J. Kennedy, and A. Hall 1969. The shell structure and mineralogy of the Bivalvia. Introduction, Nuculacea-Trigonacea. Bull. Br. Mus. Nat. Hist., Zool., Suppl. 3: 1-125, 77 figs. 29 pls., 16 tbls. 1973. The shell structure and mineralogy of the Bivalvia II. Lucinacea-Clavagellacea. Conclu- sions. Bull. Br. Mus. Nat. Hist., Zool. 22: 253-294, 33 figs. 15 pls., 22 tbls. Terquem, M. O. 1855. Paléontologie de I’étage inférieur de la formation liasique de la province de Luxembourg, Grand-duché (Hollande), et de Hettange, du département de la Moselle. Mém. Soc. Géol. France, Ser. 2, Vol. 5, Mem. No. 3: 219-343, pls. XHI-XXVI. Tryon, G. W., Jr. 1862. Synopsis of the Recent species of Gastrochaenidae, a family of acephalous Mollusca. Proc. Acad. Nat. Sci. Philadelphia, for Dec. 1861, 13: 465-494, 1 fig. 1882. Structural and Systematic Conchology; An Introduction to the Study of the Mollusca. Vol. I., 453 pp., 140 pls., published by the author, Academy of Natural Sciences, Philadelphia. Turner, R. D. 1954. The family Pholadidae in the Western Atlantic and the Eastern Pacific, Part I-Pholadinae. Johnsonia 3: 1-64, 34 pls. 1955. The family Pholadidae in the Western Atlantic and the Eastern Pacific, Part IlI— Martesiinae, Jouannetiinae, Xylophaginae. Johnsonia 3: 65-160, pls. 35-93. 1966. A survey and illustrated catalogue of the Teredinidae. Mus. Comp. Zool. (Harvard Univ.), 265 pp., 64 pls. 1969. Superfamily “Pholadacea,” pp. N702-N741 in R. C. Moore [ed.] Treatise on Inverte- brate Paleontology, Part N. Mollusca 6. Geol. Soc. Amer. and Univ. Kansas Press. Turner, R. D. and K. J. Boss 1962. The genus Lithophaga in the Western Atlantic. Johnsonia 4: 81-1 16, ls. 57-75. Vokes, H. E. 1976. Notes on the fauna of the Chipola Formation-XIX. On the presence of Gastrochaena (Spengleria) (Mollusca: Bivalvia). Tulane Stud. Geol. Paleont. 12: 161-162, 2 text-figs. Wade, B. 1926. The fauna of the Ripley Formation on Coon Creek, Tennessee. U.S. Geol. Surv. Prof. Pap. 137: 1-272, 72 pls., 2 figs. Wanner, C. 1940. Neue permische Lamellibranchiaten von Timor. University of Amsterdam Expedi- tion to the Lesser Sunda Islands 2: 372-395. Warmke, G. L. and R. T. Abbott 1961. Caribbean Seashells. Livingston Publ. Co., Narberth, Pennsyl- vania. 348 pp., 44 pls., 34 text-figs. Weller, S. 1907. A report on the Cretaceous paleontology of New Jersey. New Jersey Geol. Surv. Paleontol. Ser., Vol. 4: 1107 pp., 111 pls. Wells, J. W. 1969a. Aspects of Pacific coral reefs. Micronesica 5: 317-322. 1969b. Order “Scleractinia,” pp. F328-F369 in R. C. Moore [ed.] Treatise on Invertebrate Paleontology, Part F, Coelenterata. Geol. Soc. Amer. and Univ. Kansas Press. Whitfield, R. P. 1885. Brachiopoda and Lamellibranchiata of the Raritan clays and greensand marls of New Jersey. U.S. Geol. Surv. Monograph 9, xx +338 pp., 34 pls., map. 1893. Contributions to the paleontology of Ohio. Ohio Div. Geol. Surv., Rept. Vol. 7: 407-700, pls. 13-56. Winters, S. S. 1963. Supai Formation (Permian) of eastern Arizona. Mem. Geol. Soc. Amer. 89: 99 pp., 7 figs., 9 pls. 92 PEABODY MUSEUM BULLETIN 41 Yonge, C. M. 1948. Formation of siphons in lamellibranchs. Nature 161: 198. 1951. Structure and adaptations for rock-boring in Platyodon cancellatus (Conrad). Univ. California Publs. in Geol. 55: 401-408, 3 text-figs. 1955. Adaptation to rock boring in Botula and Lithophaga (Lamellibranchia, Mytilidae) with a discussion of the evolution of this habit. Q. J. Microsc. Sci. 96: 383-410, 19 figs. 1958. Observations on Petricola carditoides (Conrad). Proc. Malacol. Soc. London 33: 25-31, 3 figs. 1963. Rock boring organisms pp. 1-24 in Mechanisms of hard tissue destruction, R. F. Sognnaes [ed.], Amer. Assoc. Adv. Sci. Publ. 75. 1969. Functional morphology and evolution within the Carditacea (Bivalvia). Proc. Malacol. Soc. London 38: 493-527, 23 figs. — 1971. On functional morphology and adaptive radiation in the bivalve superfamily Saxicavacea [Hiatella (=Saxicava), Saxicavella, Panomya, Panope, Cyrtodaria]. Malacologia 11: 1-44, 35 figs. Zazvorka, V. 1943. Gastrochaena ostreae (Gein) a Lithodomus a (Reuss) z ceského utvaru kridového. Vést. Kral. Ceské Spoleé. Nauk. Praha. 10: 1-13, pls., i UNPUBLISHED REFERENCES Carter, J. G. 1976a. The ecologic and phylogenetic significance of the periostracum in the Bivalvia. Ph.D. dissertation, Yale Univ., New Haven, Connecticut. 1976b. The structural evolution of the bivalve shell, with notes on the ecologic and phylogenetic significance of crossed lamellar structures. Ph.D. dissertation, Yale Univ., New Haven, Connecticut. Palmer, T. J. 1974. Paleoecological studies in the Middle and Upper Bathonian of central England and northern France. Ph.D. dissertation, Oxford Univ., England. Robertson, P. B. 1963. A survey of the marine rock-boring fauna of southeast Florida. Master’s Thesis, Univ. Miami, Coral Gables, Florida. arvard MCZ Libra ANAT 4 066 305 wesrse ey, Ae eres Mato E EMEA MEAT e ce ODE Ans che hen oe ts \ Wing #22 Seas SSA 0 AY ere 6 8,4 tapers PET sve SARE we ne Gath shee ura Aaa my be widveceaae ASA GAMPAN OT LAE, PLT Wee wo yy nensge wath Pty NTNRRS ES Patience. eg, SOAS Fat aie ato rar rae PE vO he nh ate pe oF s PMs tenes : a ¢ SURPASS E eRe tens? base Eola : gag Rae PEM REE! ois 27, PARA Bid 22.5005 bb mo EMRE S yoy SDRAM Ores caer *} TEP EISS SU ore ana sy PTW cy cere PA np Fs uae ‘ TES Esyeb oer ete cet eteee. LEAR Ha dco, rewh irk ene of ob : SDL ores WOES wer tas pe leee OSpeeiem trae . MEAD e hero re” Pre nae Odie nine ree AON ceaneen, none serene . oes Leto adel bebe eee wies sg : + PRP eee ns a, vr » fs a eat CANN NORTE. Ee Fatale at SRS Boy IDR EN ey ‘ 3 7 . 3 nee irre : : PAN eo t i? xt 7 7 FUP SW: 9 3424 Fig ce EA ted de Dae er Cee Pune pet erie WARS. PN Sa a, che. 7 4 : ; Ae oe : ~ 2 ose? E feather | = i edn or ee Bk TY Mama Ce e AY B's SAA th oe 2 , ’ 1 - AoC) rea OPH 4 ’ +} Pema dart 957% ie drys . AY aoe eS 1