FOSSIL SALAMANDERS AND SALAMANDER ORIGINS

AMER. ZOOLOGIST, 5:319-334 (1965). FOSSIL SALAMANDERS AND SALAMANDER ORIGINS RICHARD ESTES Dept. f Bilgy, Bstn University, Bstn, Massachusetts SYNOPSIS. Recent (mstly unpublished) wrk indicates that the present classificatin E salamanders needs mdificatin. The fssil recrd f salamanders has been markedly imprved recently, and tends t supprt these cnclusins, but des nt ffer any clues t the rigin f salamanders r f the Lissamphibia as a whle. The structure f the sirenid salamanders, bth recent and fssil, des nt supprt the establishment f a separate Order Trachystmata fr this grup. Re-evaluatin f the Palezic amphibians suggests that the primitive temnspndyl labyrinthdnts (especially the Dissrphidae) shw the greatest number f resemblances t lissamphibians, and at this state f knwledge are the mst prbable grup f rigin. Nevertheless, at the present time, the ancestral grup cannt be indicated with any high degree f cnfidence. INTRODUCTION The recent review by Parsns and Williams (1963) has demnstrated that mdern amphibians pssess many characters in cmmn, and that the three mdern rders very prbably have descended frm a relatively restricted grup f fssil frms. Unfrtunately, the available evidence was sufficiently equivcal s that Parsns and Williams were unable t identify the grup f rigin with any f the knwn grups f Palezic amphibians. In brief, Parsns and Williams attempted t demnstrate the clse relatinship f the three mdern rders by shwing the essential identity f unique specializatins f the dentitin, ear, eye, viscera, and integument (ibid., p. 48). They cited many additinal characters that ccur as a grup nly in the three mdern rders, but that have been acquired r apprached by varius ther tetrapds. Still anther imprtant aspect f their wrk is the reinterpretatin f characters that have ften been used t indicate that the three mdern rders culd nt be clsely related. These are the nature f the vertebral clumn, the structure f the lfactry capsules, and the nature f the gymnphinan skull. The recent wrk f Szarski (1962) is als imprtant in indicating the clseness f relatinship f the mdern amphibians, This research was supprted in part by Natinal Science Fundatin Grant GB-1683. I am indebted t thse persns cited in the text fr pinins and availability f unpublished manuscripts. The phtgraphs were taken by Mr. Frederick Maynard, Bstn University. 319 and in cmbinatin with the wrk f Parsns and Williams (ibid.), Eatn (1959), and Reig (1964) frms an impressive bdy f evidence which indicates a trend away frm the plyphyletic views entrenched in the literature. Hwever, the latter tw wrkers have expressed dubts that the gymnphinans are related t the ther tw rders. Parsns and Williams have already cmmented n the dubts expressed by Eatn. Reig has nted especially that while the anterir part f the levatr bulbi muscle f the gymnphinans appears t be hmlgus with the similar structure f urdeles and anurans, the psterir part is reduced and seems t be equivalent t the reptilian levatr pterygidei. The reductin f the eyes in the gymnphinans seems a better explanatin fr the bserved differences, and if this is true, then the anatmy f the levatr bulbi des nt cntradict the relatinship f gymnphinans t the ther mdern amphibians in any fundamental way. My wn viewpint is t assume the reality f the Lissamphibia in the sense f Gadw, 1901. My purpse here is t discuss fur aspects f the general prblem: (1) the classificatin f the mdern salamanders; (2) the fssil recrd f salamanders and its relatively small cntributin t knwledge f the ancestry f the grup; (3) the relatinships f the Sirenidae t the ther salamanders; and (4) t make a renewed effrt t extract sme indicatin f the ancestry f the mdern frms frm the pre-lissamphibian fssil recrd. Dwnladed frm https://academic.up.cm/icb/article-abstract/5/2/319/183307 by guest n 03 Nvember 2018

320 RICHARD ESTES THE CLASSIFICATION OF MODERN SALAMANDERS Mdern salamanders are usually cnsidered t cnstitute five subrders within the Order Caudata: Cryptbranchidea, Ambystmatidea, Salamandridea, Prteida, and Meantes, fllwing Nble (1931). Nble's categries have remained essentially unaltered since 1931, since little infrmatin has been available t warrant a change. Hwever, recent wrk by Wake (1964), Larsen (1963), and Mnath (1965) substantiates the suggestins f Laurent (1947) based n the palatal dentitin, Dunn (1941) based n the sund-transmitting apparatus, and Teege (1957) based n vertebral develpment, that plethdntids are nt clsely related t salamandrids. Bichemical cmparisn f lactic dehydrgenases shws great similarities f Amphiuma and plethdntids in this system, rather than t salamandrids as Nble (ibid.) suggested (Salthe and Kaplan, 1964; Salthe and Kaplan, ms.). Thus the Salamandridea as cnstituted by Nble wuld be markedly altered. Hecht (1957) has pinted ut that the resemblances f the tw Recent genera f the Prteida are cmmn t larval frms, and that these animals may nt necessarily be related. I d nt believe that Hecht's evidence cnclusively demnstrates at this time that they shuld nt be in the same family, a cnclusin substantiated in part by their chrmsme number (Set, Pmerat, and Kezer, 1964). Whether r nt they are related t the Salamandridae, as suggested by Herre (1935), needs further study, but curtship prcedures (Salthe, in press) and serlgical relatinships supprt this view (Byden and Nble, 1933). The Meantes are a special prblem, and I will pstpne discussin f them until I have discussed the fssil recrd f salamanders in a general way. The reader is nt t assume that the unpublished wrk cited here is substantiated, r that I necessarily accept it in tt. It is cited as an indicatin that much imprtant wrk bearing n salamander classificatin is nw in prgress, and that in man) wa\;> it des nt supprt the classificatin f Nble. The study f salamanders is in t embrynic and t actively develping a state t prpse any new classificatin at this time. THE FOSSIL RECORD OF SALAMANDERS The hynbiids seem t have n fssil recrd at all. Auffenberg and Gin (1959) state that Wlterstrffiella may be a hynbiid, but I believe that Herre's (1950) riginal placement in the Ambystmatidae is crrect. The distinctly bicipital rib-bearers and prjectin f neural spine beynd the psterir zygapphyses indicate ambystmatid rather than hynbiid r cryptbranchid relatinships (Tihen, 1958, p. 19). The cryptbranchids are represented in the Oligcene, Micene, and Plicene f Eurpe, and the Micene f Nrth America (Westphal, 1958). Recent studies by Meszely (1965) indicate that the Nrth American fssil frms (Micene) are referable t Andrias, rather than t Cryptbranchus. The ambystmatids are knwn as late Cenzic fssils frm Nrth America (Tihen, 1958, p. 37), and Peabdy (1954) has made a gd case fr the presence f a Palecene Nrth American ambystmatid n the basis f trackways. A partial skeletn which I am nw describing frm the early Ecene f Nrth Dakta demnstrates the presence f an elngated, ribless, aquatic ambystmatid, with sme vertebral specializatins analgus t thse f the sirenids. There are a few fssils which indicate the presence f ambystmatids in the Palecene and Micene f Eurpe. These are Wlterstrffiella (nted abve) and Bargtnannia (Herre, 1955), respectively. The latter has strng anterir basapphyses as des the Nrth American Ecene frm nted abve, and is best placed as an ambystmatid, as Herre has suggested. Geyeriella (Herre, 1950), als frm the Palecene f Eurpe, has anterir basapphyses as in Bargmannia and ther ambystmatids, and cntrary t Herre is thus prbably nt a plethdntid, a grup in which the basapphyses (when present) are usually psterir, except in sme large specimens f Phtliudn glutinsus and P. juxlani. Gey- Dwnladed frm https://academic.up.cm/icb/article-abstract/5/2/319/183307 by guest n 03 Nvember 2018

SALAMANDERS 321 eriella may thus be referable t the Ambystmatidae alng -with Bargmannia. Scapherpetn and the related Lisserpetn (Estes, 1964, 1965) are abundant in Nrth American Cretaceus and early Cenzic depsits, and are essentially cryptbranchids in many aspects f their skull structure. In thers, ntably the cnstructin f the maxilla and the presence f nly tw bnes in the lwer jaw, they resemble the ambystmatids (specifically large larval Dicamptdn). The bdy f these animals may have been elngated, their skulls shw indicatins f reductin in ssificatin, and at least ne, prbably bth, had small thugh well-develped limbs. These features may indicate that they were permanently larval t sme degree, as in Amphiuma. These animals have been placed in a separate family Scapherpetntidae by Auffenberg and Gin (1959), but the validity f this family must await detailed stelgical studies f the large amunt f material f Scapherpetn nw available (Estes, 1964). The Salamandridae is represented by a number f undubted fssils frm Palecene and later sediments in Eurpe (Kuhn, 1962), pssibly in the early Cretaceus f Israel (Nev, 1964), and frm Oligcene and later sediments in Nrth America (Van Frank, 1955; Estes, 1963; Richmnd, 1964). The few meager Pleistcene recrds f Amphiuma in Nrth America (Brattstrm, 1953) are nw supplemented by a Palecene frm frm Wyming, prbably ancestral t the recent genus (Estes, ms.). The parasphenid referred t Amphiuma by Brunner (1956) frm the Pleistcene f Eurpe is unlike that f any salamander in the pssessin f a strng ventral keel, and is almst certainly frm a fish. The Prteidae is represented by Palaeprteus in the Ecene f Eurpe (Herre, 1935), and pssibly there als by the early Cretaceus Hylaebatrachus and the Micene Orthphyia, all f which are mre r less elngated animals and resemble the mdern permanent larval frms in their skull structure. There is sme evidence that the earliest knwn salamander, frm the late Jurassic f Wyming, belngs t this grup as well (Hccht and Estes, 1960). It is likely t have been a permanent larva, judging frm its lack f ssificatin and similarity t ecturus. The nn-desmgnathine plethdntids are nt knwn with certainty as pre-pleistcene fssils, unless the Plicene trackway referred by Peabdy (1959) was made by a member f this grup. The Eurpean Micene Dehmiella may belng here als (Herre and Lunau, 1950). The distinctive desmgnathine Prdesmdn first appears in the early Cretaceus f Texas (specimens cited by Hecht, 1963, p. 22, and nw being studied by me). It als ccurs in the late Cretaceus f Wyming (Estes, 1964), but then seems t disappear frm Nrth America. It reappears, surprisingly little changed frm the early Cretaceus representatives, in the Micene f France (specimens currently under my study). Its few divergences frm mdern desmgnathines can easily be ascribed t its prbably very elngated, aquatic habitus and its primitiveness. Several ther frms f permanent larval type als may be near the plethdntids, thugh the evidence fr this is equivcal. These are Balraclisaurid.es, frm Ecene (Estes, ms.) and Micene (Estes, 1963) depsits in Nrth Dakta and the suthern United States respectively, and the elngated Opisthlritn, frm the Cretaceus and Palecene f the Rcky Muntain regin (Estes, 1964; Auffenberg, 1961). The latter may be related t Batraclisanrides, but als shws sme skull similarities (perhaps nly analgus nes) t the mdern Gyrinphihis palleucus necturides, a netenic plethdntid (Lazell and Brandn, 1962). The sirenids have a lng fssil recrd, extending back t the late, and perhaps even early, Cretaceus. This grup will be cnsidered in greater detail belw. In summary, my studies f fssil salamanders indicate the fllwing very prbable situatins: (1) Mst f the fssil animals, especially the early nes, are structurally analgus t mdern permanent larval types. Many, perhaps mst, shw tendencies tward elngatin. This is perhaps because f the nature f the fldplain Dwnladed frm https://academic.up.cm/icb/article-abstract/5/2/319/183307 by guest n 03 Nvember 2018

322 RICHARD ESIES aquatic depsits in which the fssils ccur, especially in Nrth America, where the sutheastern regin frms the nly area in which salamanders f this type still exist (with the exceptin f the cave-dwelling Eurpean Prteus). (2) All f the available fssil salamanders are referable t, r clse t, mdern families. (3) As a result f their clse relatinships t mdern frms, their structure shws little r nthing which relates them t any particular grup f Palezic amphibians. At the mment, nly the sirenids seem t be ptentially useful in a discussin f the rigin f salamanders. Hwever, they are imprtant nly in regard t the questin f plyphyly; i.e., have the sirenids descended frm Palezic amphibians separately frm ther salamanders, r are they true Lissamphibia? THE RELATIONSHIPS OF THE SIRENIDAE Gin and Auffenberg (1955, 1957, 1958) have described a number f Cenzic and Cretaceus sirenids, and I have cntributed additinal evidence n the structure f the late Cretaceus frms, which are represented by skull and hyid elements as well as vertebrae (Estes, 1964). As a result f Gin and Auffenberg's studies, Gin and Gin (1962) have suggested that numerus peculiarities f skull and vertebrae f the sirenids justify placing them in a separate Order Trachystmata. They have suggested, in additin, that the sirens were derived frm the ai'stpd grup f fssil lepspndyls, in cntrast t a presumed rigin f salamanders frm the micrsaurs. A substantiatin f this allcatin has nt been published, but I prpse here t discuss varius aspects f the structure f sirenids in an attempt t test the validity f Gin and Gin's prpsitins. Gin and Auffenberg (1955) have described the well-develped frked alifrm prcesses n the psterir end f the neural arches f the vertebrae f sirenids. The assciated musculature has been described by Auffenberg (1959) and cmpared with that f Amphimna and ectmus. Bth Amphiiima and Siren are specialized in their epaxial and hypaxial musculature, and Auffenberg has demnstrated clearly that bth genera have mdified this musculature in smewhat different and parallel fashin, and that in sme ways Necturus and Amphiiima are mre similar t each ther than either is t Siren. Auffenberg's cnclusin is that "the differences [in epaxial and hypaxial musculature] between the tw genera seem t be basic and supprt the widely held cncept f their distant relatinship." Hwever, by his wn admissin, Siren and Necturus are mre similar t each ther than t Amphiiima in the mysepta and their cnnecting fiber tracts, and Amphiiima and Siren als resemble each ther in a number f features. In additin t this mylgical evidence, Siren and Amphiiima shw a number f characters in cmmn which may be parallelisms assciated with the elngated bdy frm, but which are nevertheless specifically similar enugh that rdinal separatin f the sirenids frm ther salamanders may be t extreme a view: Thus, bth have large Jacbsen's rgans, mdificatins f the chanal valve t permit vluntary cntrl, pendulent extensins f the auricles, and an unlbulated testis. Siren als shares with Cryptbranchus the presence f large thyrids and an eminentia septalis f the frebrain (ble, 1931). It thus seems premature t base any phylgenetic cnclusins n the musculature, until the epaxialhypaxial musculature f such genera as Prteus, Dicamptdn, and sme f the salamandrids is described as well. The alifrm prcesses themselves are highly develped in Siren and Pseudbran chits, but are much less well-develped drsally in the Cretaceus sirenid Habrsaurus (Estes, 1964, Fig. 37), and in additin similar structures are present in sme f the salamandrids. In Euprctus, fr example, there is a prminent ntch n the psterir surface f the neural arch int which muscles insert, and which is flanked by prminent alifrm prcesses very much like thse f sirenids (Fig. 1). This srt f structure als ccurs in the tail regin in Tinicha trsa (Teege, 1957). The develpment f the ntch seems t me t be an Dwnladed frm https://academic.up.cm/icb/article-abstract/5/2/319/183307 by guest n 03 Nvember 2018

SALAMANDERS adaptatin fr increasing the length f the individual muscles withut adding t the FIG. 1. Vertebrae f sirenids and salamandrids t shw the develpment f the alifrm prcesses. A, Habrsaurus dilatiis, drsal vertebra (Late Cretaceus, Wyming), transverse prcesses restred frm ther specimens, abut 5. B, Euprclus platycephalus, drsal vertebra (Recent), 10. Anterir and psterir ends f centrum retuched t distinguish them frm the neural arch. C, Habrsaurus dilatus, Univ. Califrnia Mus. Palen. 54214, drsal vertebra (late Cretaceus, 'Wyming), psterir zygapphyses and trans\erse prcesses restred, 1.5. length f the vertebrae themselves. Many ther salamanders shw bifurcatin f the neural spines in sme parts f the vertebral clumn (e.g., ambystmatids, Amphiuma), and it even ccurs in the hypapphyses f sme salamandrids, e.g., Triturid.es chinensis. Thus while it is true that sirenids have emphasized this cnditin (especially in the mdern species), I am nt cnvinced that this need be interpreted as setting the sirenids far apart frm ther salamanders, especially when the presence f alifrm prcesses in sme salamandrids, their lesser develpment in fssil sirenids, and my interpretatin f ther evidence (see belw) is cnsidered. The salamandrids and the sirenids are als similar in the relatively large neural arch, in the extreme lateral placement f the lfactry lbes f the brain (Nble, 1931, p. 357), in the tendency t lay small numbers f eggs in many places and ver a lng perid f time, and in their serlgical relatinships (Byden and Nble, 1933). Reed (1920) shwed that there were similarities f sirenids t salamandrids in the mrphlgy f the ear regin. Hwever, sme f Reed's evidence was questined by Dunn (1941). The extreme larval nature f the sirenids makes this regin a difficult ne t interpret, and there are sufficiently equivcal aspects f the currently available evidence t warrant a thrugh re-study f this regin in the sirenids. There are a few characters in which sirenids resemble the lissamphibians in general, but which have been cited ften as evidence fr strng difference between them and ther mdern amphibians. The presence f hrny beaks n adult mdern sirenids is reminiscent f the cnditin in frg tadples, as well as shwing similarity t the relatively primitive urdeles Onychdactylus and Ambystma which, hwever, have nly the predentary sheath (Nble, 1931, p. 52). The separate cracid ssificatin in the sirenids is als a similarity t frgs, and is knwn therwise in amphibians nly in the nectridean Diplcaulus and in Seymuria. Any phylgenetic significance cited fr the presence f this element wuld be extremely difficult t maintain in animals which shw such strngly different Dwnladed frm https://academic.up.cm/icb/article-abstract/5/2/319/183307 by guest n 03 Nvember 2018

324 RICHARD ESTES FIG. 2. Siren lacerlina, smewhat diagrammatic palatal view f Mas. Cmpar. Zl. Harvard 136 (Recent), abut 2.5. ec, Ectpterygid: e, exccipital; f, frntal; f, fenestra valis; icf, internal cartid framina; m, maxilla; n, nasal; pi, palatine; pm, premaxilla, ps, parasphenid; pt, pterygid; q, quadrate; s, squamsal; se, sphenethmid; v, vmer. adaptatins and phyletic psitins as d the abve-mentined frms. Mst interesting in the present cntext are a number f characters which tend t set the mdern sirenids apart frm ther salamanders, but which d nt ccur in the less mdified fssil frms. Here mre discussin is necessary, since much f this infrmatin is still unpublished r has been published nly recently (Estes, 1964). First, the skull f Siren is mre elngate than that f mst ther mdern salamanders. Hwever, fssil material nt yet described shws that the skull f the Cretaceus sirenid Habrsaurus is relatively shrt, and has a blunt snut very much like that f ther salamanders (cf. Figs. 2, 3). As an analgue t this situatin, the skull f Ampliiitrna is als elngate in mdern frms, but is shrt and blunt in the Palecene fssil mentined abve. Sirenids are stated t lack a maxilla (ble, 1931, p. 181) and all but the palat- \mcriik 1 teeth; \et Hiibrsaimts had premaxillary, maxillary, and dentary teeth f a peculiar frm with expanded crwns (Estes, 1964, Figs. 34-36). This type f tth is seen in relatively smaller frm in sme living salamanders and frgs (e.g., Tihen, 1958, Fig. lh). In additin, the vmer is a distinct element, and its teeth are arranged in a pattern similar t that seen in larval salamanders (cf. Figs. 3, 7). The expanded palatines are similar t thse f sme plethdntids. Cntrary t Nble's statement, a small separate maxilla flating in cnnective tissue, and prearticular teeth are present in mdern Siren, and were first figured by Parker (1882). Unlike thse f mst lissamphibians, the teeth f fssil and Recent sirens are nt pedicellate, but the cnditin is reduced in the permanently larval Nectunts (Parsns and Williams, 1962) and absent in larval Rhyactritn (Larsen, 1963). Its absence in mdern sirenids is thus perhaps partly the result f prgressive paedmrphsis. It culd als be in part the result f descent frm an animal like Habrsaurus in which, as a result f the crushing dentitin, the pedicellate cnditin had been lst as inadaptive. Anther cnsideratin is the shape f the premaxillae, and the fact that their ascending prcesses are lateral t the nasals in the, r '/? y,' I ^ FIG. 3. Hnbrmauni'i dilattis, restratin f palate Hate Cretaceus, \V\ming), alx>ut 2. Dtted lin ale drawn b\ analg) with Recent Siren lacerliiin. Fi alilnt*\ iatins see Fie. 2. Dwnladed frm https://academic.up.cm/icb/article-abstract/5/2/319/183307 by guest n 03 Nvember 2018

.SAI.AMADF.RS 325 FIG. 4. A, Habrsaurus dilatus, restratin C drsal view f premaxillae and maxillae (late Cretaceus, Wyming), abut 2. B, Siren lacertina, drsal view f snut regin f Mus. Cmpar. Zl. Harvard 136 (Recent), abut 2.5. Fr abbreviatins see Fig. 2. mdern sirenids. Again this seems t have been different in Habrsaurus, which resembles such urdeles as Necturus in the premaxillnasal area in that the premaxillae are nt widely separated alng the midline and thus prbably excluded the nasals (Fig. 4). Byden and Nble (1933) have shwn hw the nasal arrangement in Siren can be derived frm that f sme hynbiids, r (in their view) less likely, frm that f sme salamandrids. The resemblance f the sirenids t the lepspndyls must nw be discussed. Gin and Gin (1962, p. 65) have stated that the frmer "shw marked resemblances t the ai'stpd-nectridean stck" "especially in the detailed structure f their vertebrae." Aistpd vertebrae have been mst recently studied by Baird (1964). A minr amunt f frking at the psterir end f the neural spine ccurs, but this als ccurs in a number f salamanders ther than sirenids, and t much greater degree than in aistpds. The pssible presence in the latter f basapphyses (present in urdeles) has been cited by Baird, but McGinnis (1964) disagrees, n the basis f better material. The presence f intravertebral spinal nerve exits resembles sirenids as well as a number f ther salamanders (Estes, 1964, pp. 81, 93). Thus if aistpds are related t any mdern salamanders, it is t all f them equally, with n special relatinship t the sirenids. Aistpds d nt at all resemble mdern salamanders in skull structure, in the pssessin f single-headed ribs, r in the cmplete absence f limbs and girdles. As a final nte n the relatinship f the sirenids and the aistpds, the status f the presumed aistpd Palaesiren beinerli frm the Permian f Eurpe must be discussed. It was riginally described by Geinitz (1864), and later figured by Geinitz and Deichmuller (1882, pi. 9, reprduced here as Fig. 5), wh stated that this animal was very similar t Siren. Baird (1964, fn. p. 14), in cmmenting n Fritsch's interpretatin f this animal as a gigantic aistpd, ntes that the vertebrae appear t be hurglass shaped nly externally; "they lack the funnel-like amphicely which characterizes the Aistpda." As Baird ntes, the specimen is unprepared, but the figure shws three vertebrae, each ten centimeters lng, giving a ttal length f 14 meters fr this amazing beast, accrding t vn Huene (1956). The latter has als published (ibid., Fig. 23) what appears t be a tracing f the riginal figure, and which if accurate resembles neither aistpds nr sirenids. The riginal figure indicates that the vertebrae are bscured drsally by sme srt f dermal armr, and it is difficult t see any similarity t sirens r t lepspndyls f any srt. The size f the vertebrae f this animal indicates that it may perhaps be a reptile since it is far larger than the largest knwn amphibian. The relatinships f the sirenids t the nectrideans is equally difficult t dcument. The vertebrae f the latter all have cmplex accessry articulatins, whereas in sirenids there are n accessry articulatins, even n the expanded caudal vertebrae. N indicatin f the develpment f alifrm prcesses is present in nectridean vertebrae, and all f the latter bear a characteristic striatin n neural and haemal spines. The tail f the nectridean Scincsaurus is als expanded adaptively as in Dwnladed frm https://academic.up.cm/icb/article-abstract/5/2/319/183307 by guest n 03 Nvember 2018

820 RICHARD ESTES A B KIC. 5. Palaesiren heinrrti. A, drsal view f vertebrae; ]!, ventral view f vertebrae; abut ne-third (frm Geinitz and Deichmuller, 1882, pi. 9). sirens and in sme ther salamanders, and must have been used in aquatic lcmtin as in the mdern grups. Hwever, Scincsaurus was very likely ne f the hrned frms, as nted by Parsns and Williams (1963, p. 45), and cnfirmed by Baird (in litt., 1965), and is thus a pr candidate fr salamander (r lissamphibian) ancestry. The vertebrae f the keraterpetntid nectrideans Diplcauhis and Diplceraspis (Beerbwer, 1963) have clsely-apprximated duble rib articulatins as in salamanders, but unlike thse f ther members f the family and f the superficially mre salamander-like urcrdylid nectrideans. The vertebrae f these grtesque hrned frms are f generally rather salamander-like frm, and shuld be a warning t thse wh wuld use grss vertebral similarities t dente ancestry f any mdern amphibian within a Palezic grup. As in the case f the aistpds, if the nectrideans are related t mdern salamanders, it is t all f them equally, with n special relatinship t the sirenids. Thus it seems clear (thugh much wrk remains t be dne) that there are n cmpelling reasns t separate the sirenids frm ther salamanders rdinally, fr there are t many shared characteristics with the mdern frms, and the similarities are heightened when fssil sirenids are cnsidered. It wuld be especially desirable t determine the presence in sirenids f such characters as the green rds f the retina, the papilla amphibirum f the inner ear, Dwnladed frm https://academic.up.cm/icb/article-abstract/5/2/319/183307 by guest n 03 Nvember 2018

SALAMANDERS 327 and the fat bdies, since these and ther characters were suggested by Parsns and Williams (1963) as indicatrs f the clse relatinship f mdern amphibians. THE ANCESTORS OF THE LISSAMPHIBIA Several pints must be nted befre the fssil recrd is reassessed t see whether r nt any suggestin can be made as t the grup frm which the Lissamphibia riginated. The salamanders and frgs exhibit many parallelisms within their wn grups, and it is thus difficult t assess the phylgenetic significance f many mrphlgical features. This is als the case in the Palezic frms, in which such features as skull shape, develpment f a duble ccipital cndyle, lss f the hypglssal framen, etc., have lng been knwn t ccur in unrelated lineages. Thus the prblem f relating tw grups separated by many millins f years is intensified. The adaptive radiatin f the Palezic amphibians prduced small, urdele-like types several times hw many times, we d nt knw. It is appealing t accept Rmer's suggestin (1947) that "branchisaurs" are the larvae f labyrinthdnts, and he has demnstrated this in a number f instances (see Lehman, 1955; Heyler, 1958, fr dissenting pinins). Hwever, can we be sure that all small, Palezic labyrinthdnts are larvae f sme larger frm? It is ften difficult t distinguish between the adult skulls f small species f living reptiles and the yung f a larger species, and the varying degrees f paedmrphsis and neteny achieved by mdern salamanders d nt make the assessment f maturity in the "branchisaurs" easier. I am nt suggesting reinstatement f the "Phyllspndyli," nr indicating that the members f this s-called grup are ancestral t salamanders r t the Lissamphibia as a whle. Reig (1964), hwever, has suggested recnsideratin f the "Phyllspndyli" as ptential ancestrs f the Lissamphibia, fllwing the suggestins f Lehman (1955) and Heyler (1958) that the "Phyllspndyli" are prbably a natural grup. Reig has stated that the latter grup resembles urdeles and anurans in: fenestratin f the palate, the presence f tw ccipital cndyles, the maxillquadratjugal cntact, the well-develped tic ntch, and reductin f the bnes f the cranial rf. As Parsns and Williams have nted, the first fur f these are present in a number f grups. The last character is nt true f the animals which have been referred t the "Phyllspndyli" and relatinship t them f the Lissamphibia is nt especially cnvincing (see Table 1). Much that has happened in the evlutin f mdern terrestrial salamanders is likely t have been essentially a late Meszic and Cenzic phenmenn. The mdern terrestrial frms are specialized in varius ways and prbably can tell us little abut lissamphibian ancestry. What seems imprtant t emphasize here is that salamanders, indeed mdern amphibians as a whle, very likely started ut as transfrmed, but primarily aquatic frms (as Leurgnathits and thers are tday), and that a number f grups have independently achieved sme degree f terrestriality. This phenmenn in the mdern frms was perhaps partly an adaptatin t remve the eggs frm the aquatic medium, in which they were subject t heavy predatin, either macr- r micrrganismal. Pssibly this predatin increased rapidly with the essentially Cretaceus and Cenzic evlutin f abundant fresh-water telests. Unfrtunately, the majrity f knwn fssil salamanders were prbably permanent larval types, and thus even less likely t shw indicatins f lissamphibian ancestry than are the mdern terrestrial frms. In any case, frm the nature f the fldplain depsits in which mst fssil salamanders have been fund, we are unlikely t find many specimens f mre ancient terrestrial types. Nw that sme f the prblems have been indicated, let us turn t the lissamphibian prttype cnstructed by Parsns and Williams (1963). When they applied their list f characters which such a prttype culd be expected t have t the Palezic amphibians, they fund that the Dwnladed frm https://academic.up.cm/icb/article-abstract/5/2/319/183307 by guest n 03 Nvember 2018

328 RICHARD ESTES Rmeriidae. Since Parsns and Williams based their discussin f the s-called "rep- tilian" micrsaurs n that genus, relatinship f this grup f micrsaurs t the liss- amphibians is exceedingly imprbable, Parsns and Williams have als shwn that mst prbable grups f rigin were the temnspndyls, the nectrideans, and the micrsaurs f s-called "reptilian" type, With respect t the latter, Baird (1958) presents evidence that Cephalerpetn is a captrhinmrph reptile f the family Bdy f mderate length Appendicular skeletn well-develped Pleurcentra present, dminant Ribs tw-headed Atlas f mdern type Skull relatively shrt, brad Palate f advanced type Otic ntch present Parietals frm psterir skull margin All lissamphibian skull bnes present Tempral regin fenestrated Maxilla cntacts quadratjugal Operculum psterir t stapes Occipital cndyles paired i Hypglssal framen present Anterir braincase bnes present Mentmeckelian lx>nes present Small stederms present Teeth pedicellate Number f lissamphibian characters present in all knwn members () Number f cases in which characters ccur in sme but nt all members (Oj Number f lissamphibian characters lacking in knwn members (O) Ttal pints indicating lissamphibian affinity TABLE 1. Tabulatin f lissamphibian characters by grup. O O x x x? x x 3 7 9 10 O O? x x?? 5 3 9 8 O O O x? x?? 5 5 8 10 O x x? x x? x x x 5 8 6 13??? 8 0 8 8????? 9 0 9 9? 12 0 7 12 Anthracsaurs Nectrideans Micrsaurs Temnspndyls "Phyllspndyli" Platystegs Amphibamus Dissrphidae x? x 11 2 5 13 Dwnladed frm https://academic.up.cm/icb/article-abstract/5/2/319/183307 by guest n 03 Nvember 2018

SALAMANDERS 329 the ther grups f micrsaurs lack any strng indicatin f lissamphibian relatinship. Parsns and Williams' characters are presented in graphic frm in Table 1. It shuld be stated here that I realize fully the dangers f wrking with an artificial r hypthetical mdel ancestr, but I believe the present cmparisn is useful in the cmparisn f trends tward the lissamphibian cnditin. Parsns and Williams nte that an atlas f mdern type wuld be expected in the ancestr f the Lissamphibia. Althugh they are prbably crrect in this, similar atlantes have been evlved in micrsaurs, nectrideans, and sterespndyls (Gregry, Peabdy, and Price, 1956, Fig. 16a; Beerbwer, 1963, Fig. 8a; Sawin, 1945, Fig. 6a), and the axis f mammals is a similar and analgus structure. Parsns and Williams expected that an dntid prcess wuld be present n the ancestral lissamphibian atlas, but the atlantes f many frgs, the salamanders Prdesmdn and Opisthtritn (Estes, 1964, Figs. 38g, 42a), and the sterespndyl Eupelr (Sawin, ibid.) are als quite similar and lack the dntid prcess; thus the ancestral atlas culd be f either type. Szarski (1962) has pstulated an early split between amphibian and amnite lines; a suggestin which is in accrd with present understanding f the nature f the seymuriamrphs and pssible rigin f reptiles frm primitive labyrinthdnts (Rmer, 1947, p. 93; Olsn, 1965; Gregry, 1965). This situatin was nted by Parsns and Williams (1963, p. 49) as an indicatin that lissamphibians might be descended frm the small, prbably cutaneuslyrespiring lepspndyls instead f the (usually) larger labyrinthdnts, which may have been using the ribs and cstal muscles fr respiratin as in amnites. Thugh Szarski has clearly delimited the differences between the tw methds f respiratin, it is difficult t believe that several methds were nt used by varius labyrinthdnts, and that cutaneus respiratin wuld nt be ne f them (especially in the smaller frms, fr I agree with Parsns and Williams that cutaneus methds wuld be insufficient fr large, active labyrinthdnts). When all f the characters which are either present in all r at least sme members f the grups are ttalled, it is apparent that amng the majr grups, the temnspndyls have the highest prprtin f lissamphibian characters expressed t ne degree r anther (Table 1). This, I think, is an indicatin that the temnspndyls are the grup f rigin f the lissamphibians r are clsest t that grup; at the least, this is the surce fr which the best case can be made. Eatn (1959) has als suggested a temnspndyl rigin fr the Salientia and Caudata, thugh he cnsidered the gymnphinans t be derived frm Lysrphus-like lepspndyls. The nly temnspndyls which shw any significant similarity t the Lissamphibia are the primitive frms with rhachitmus vertebrae. Here, then, is an immediate prblem fr thse wh wuld derive mdern amphibians frm this grup. As part f an attempt t answer this prblem, I wuld like t re-emphasize the reasn fr the ften-cited ancestry f salamanders frm the lepspndyls. It is nt, certainly, n the basis f similarity f the skulls f the tw grups a similarity which in fact is nt present, as has been shwn by Parsns and Williams (1963) and thers. This ancestry is, ultimately, based n the fact that bth grups have vertebrae which, in the adults, are frmed f nly ne piece, and that the lepspndyls are fr the mst part small animals, like the mdern amphibians. Minr details f the transverse prcesses have been cited as evidence fr nectridean ancestry f salamanders by Gregry, Peabdy, and Price (1956), but this evidence is t weak t d mre than exemplify the classical appeal f lepspndyls as ptential ancestrs. Lepspndyl vertebrae arc inferred by Williams (1959) t be pleurcentral in rigin by cmparisn with embrylgical cnditins in the Lissamphibia, especially the Gymnphina (see Panchen, 1963 fr a dissenting pinin n nn-embrylgical grunds). Parsns (cmments, this sympsium) has suggested that the lepspn- Dwnladed frm https://academic.up.cm/icb/article-abstract/5/2/319/183307 by guest n 03 Nvember 2018

330 RICHARD ESTES dylus and lissamphibian vertebrae may be hmlgus t bth pleurcentriim and intercentruin. Because f embrylgy, this is nt prbable in the Lissamphibia; the embrylgy f the lepspndyls is utside discussin. He has als made the interesting pint that in the sterespndyls, in which nly the intercentrum was develped in the adult, the intercentrum culd be expected t be intersegmental in psitin fr functinal reasns. Thus the criterin f pleurcentral hmlgy by intersegmental psitin in the embry, as suggested by Williams, might nt apply in the adults, whatever the situatin in the embry. It is the pinin f Williams (1965) that tw basic questins are being argued; ne f embrylgical develpment and segmentatin, and anther f adult structure and mytmal relatinships. Whatever the basic cmpnents may be embrylgically, the vertebra grws and may shift its psitin during grwth. It is thus pssible t have an adult vertebra f quite different segmental psitin than its embrylgical predecessr. The studies f Eatn (1959) and (especially) Williams (1959) cnfirm that further careful study f the embrylgical develpment f amphibian vertebrae is needed t shed light n this prblem, but their wrk has made it acceptable, I think, t state that it is as pssible t derive the vertebrae f mdern amphibians frm a primitive rhachitmus type as it is frm a lepspndylus type. In any case, all amphibians (including lepspndyls and lissamphibians) must ultimately have been derived frm rhipidistians with a prt-rhachitmus vertebral clumn, since there is little evidence at present t supprt the view that lepspndyls r lissamphibians were derived frm rhipidistians separately frm the ther amphibians (Reig, 1964; Thmsn, 1964; and ther papers cited by the latter). Thus it is academic t attempt t justify a transitin which, in this view, has in fact ccurred. Whether r nt this has happened nce, t a cmmn ancestr f lepspndyls and lissamphibians, r separately in the tw grups is t greater interest. It seems clear that in this regard it will be mre critical t btain new Palezic fssils, than t cntinue t search the present data fr new interpretatins. When individual tenmspndyls are examined fr evidence t lissamphibian affinity, n ne frm emerges as being significantly like the prttype lissamphibian. Hwever, the restratin f the primitive edpsid rhachitme Platystegs lricatum (cnsidered a synnym f Dendrerpetn by Carrll, 1964a) frm the middle Pennsylvanian f Nva Sctia (Watsn, 1956) shws mre individual similarity t the Lissamphibia than d mst labyrinthdnts. Still, with the many resemblances indicated by tabulatin f the characters (Table 1) Platystegs des nt appear n first view t be a ptential lissamphibian ancestr r relative (Fig. 6b). The dissrphid Amphibamus (Fig. 6a), with the mdificatins t the wrk l Gregry (1950) and Watsn (1940) necessitated by the recent studies f Carrll (1964b), shws mre f the lissamphibian characters than des Platystegs. Amphibamus is an erypsid rather than an edpsid as is Platystegs, and is als frm the middle Pennsylvanian. Bth f these animals will be discussed belw as ptential lissamphibian relatives. Eatn (1959) has stated that a primitive characteristic f at least frgs and salamanders was a mvable basipterygid articulatin. This cnditin is als characteristic f sme adult Palezic types (prbably f all primitive nes), and was prbably always the case in larval frms. Watsn (1956) has described the mvable basal articulatin f a number f primitive rhachitmus amphibians, and has shwn the presence f a strng internal prcess f the pterygid in these animals, including Platystegs. The cmparable area in Amphibamus is nt well preserved, but seems t have been similar (Carrll, 1964b). A similar (but nt identical) situatin is als seen in large specimens f the ambystmatid salamander Dicamptdn (Fig. 7). The basicranium f Platystegs is unfrtunately nt knwn, but related frms indicate the present.e l a single cndle (which culd becme duble by lss f the basicripital) Dwnladed frm https://academic.up.cm/icb/article-abstract/5/2/319/183307 by guest n 03 Nvember 2018

SALAMANDERS S81 but n indicatin f the presence f an perculum. Otherwise, strng reductin f the ectpterygid and palatine culd result in essentially the same kind f palatal structure as in Dicamptdn. The same cnditins hld fr Amphibamus, and in additin, the ccipital cndyle is distinctly duble in dissrphids, as in lissamphibians (ibid., p. 234, Figs. 5, 13). Parsns and Williams (1963) nted that B JIG. G. A, Ampltibamus lyelli, restratin f palate, 1 (frm Carrll, 1964, Bull. Nfus. Cm p. Zl. Har\:ircl). B, Platystegs (=: Deiulrerpeln) lrirahim, restratin f palate, abut 1.25 (after Watsn, 1956;. Fr abbreviatins see Fig. 2. FIG. 7. Dicamptdn ensatus, Mus. Cmp. Zl. Harvard 4365, netenic male, palatal view f skull (Recent), 2. the fenestratin f the salamander and frg skull was psterir and ventral t the rbit. Since the maxilla cntacts the quadratjugal in Platystegs^ lss f the jugal wuld prduce fenestratin and a lwer tempral arch much like that f the Lissamphibia. These changes wuld be pssible nly if the main axis f stress came t lie alng the lngitudinal axis f the skull as it des in the lissamphibians, and the apparent absence f stress-bearing anterir braincase elements in Platystegs and its relatives is a majr difference frm the mdern grups. On the cntrary, the presence in dissrphids (including Amphibamus) f a well-defined sphenethmid (Carrll, 1964b, Figs. 7, 22) is a resemblance t lissamphibians. Drsally, the skull f Platystegs is like that f many ther labyrinthdnts, and thus des nt resemble that f lissamphibians. Hwever, the psterir skull table bnes are small, and frm Watsn's restratin are in gd part ccipital. Enlargement f the prminent supratempral prcess f the squamsal might give sufficient supprt t the psterir part f the skull, s that further reductin in the psterir bnes wuld be pssible. The atlas f sme dissrphids resembles Dwnladed frm https://academic.up.cm/icb/article-abstract/5/2/319/183307 by guest n 03 Nvember 2018

332 RICHARD E.STES that f lissamphibians, being frmed f a neural arch fused t a cylinder in Dissrphus and Cacps (Carrll, 1964b, p. 205); thugh this is nt knwn with certainty fr Amphibamus, it seems t have been similar (ibid., p. 234). At least ne dissrphid (Tersmius) has an pening fr the hypglssal nerve in the exccipital. Again this is nt knwn with certainty in Amphibamus, but since Carrll (1964b, p. 244) cnsiders the latter ancestral t Tersmius, it can be presumed t have ccurred in Amphibamus as well. It is apparent frm the abve cnsideratins that bth Plntystegs and Amphibamus, especially the latter, resemble the lissamphibians in many ways. Neither has enugh f the imprtant lissamphibian characteristics (r evidence f trends in the develpment f these characteristics) t be ancestral, but they intensify the case fr primitive temnspndyl relatinships f the Lissamphibia, as has already been suggested in part by Eatn (1959). Anther dissrphid nw being described by Eatn may be even mre significant fr lissamphibian phylgeny, since it has what appears t be an pisthtic strngly reduced t a flat, val piece f bne psterir t the stapes, and thus in appearance and psitin similar t the perculum f mdern amphibians. This situatin culd imply that the s-called pisthtic f prteid salamanders is nt really hmlgus with the pisthtic f Palezic frms. The crllary f this statement, that the prteids were derived frm Palezic frms befre the frmatin f an perculum, wuld be impssible t maintain in view f their clse resemblance t the ther salamanders. Hwever, there is n reasn t assume a ne-tne relatinship between perculum and pisthtic; bth are ssificatins f the psterir wall f the ear capsule and bth culd ccur tgether. The statement f Nble (1931, p. 215) that transitry pisthtic ssificatins ccur in Ambystma, ecturus, Siren, and a few ther urdeles shuld be fllwed up in this regard. In any rase, the cmparisns given abve indicate that amng knwn Pale/ic amphibians, the dissrphids are the mst plausible lissamphibian relatives (Table 1), thugh they certainly cannt be cited as the grup f rigin at this precarius state f knwledge. In summary, the clsest resemblances between mdern and Palezic amphibians are with the temnspndyls (specifically the Dissrphidae) rather than with the lepspndyls. The latter grup has held cnsiderable appeal fr sme wrkers searching fr ancestrs f the mdern frms. This appeal cannt be dcumented in any significant way s far as nectrideans and aistpds are cnsidered, and the same can be said f the micrsaurs, as stated by Parsns and Williams (1963) and by Gregry (1965). It is certainly true, hwever, that n individual temnspndyls shw tendencies tward skull bne reductin r the develpment f pedicellate teeth, bth f which shuld be demnstrated in any immediately ancestral frm r primitive lissamphibian. Thus, until mre infrmatin is available, the search fr lissamphibian ancestrs remains little mre than grasping at straws. REFERENCES Auffenberg, W. 1959. The epaxial musculature f Siren, Amphiutna, and Neclurus. (Amphibia.) Bull. Flrida State Mus. Bil. Sci. 4:253-265. 1961. A new genus f fssil salamander frm Nrth America. Amer. Midi. Nat. 66:456-465. Auffenberg, W., and C. Cin. 1959. The status f the salamander genera Srapherpetn and Hemitrypus f Cpe. Am. Mus. Nvit. 1979:1-12. Baird, D. 1958. The ldest knwn reptile fauna. Anat. Rec. 132:407-408.. 1964. The aistpd amphibians surveyed. Mus. Cmpar. Zl. Harvard Brevira, N. 206, 17. Beerbwer, J. 1963. Mrphlgy, paleeclgy, and ph\lgen) f the Perm-Penns\lvanian amphibian Diplceraspis. Bull. Mus. Cmpar. Zl. Harvard 130:34108. Byden, A., and G. Nble. 1933. Relatinships f sme cmmn Amphibia as determined by serlgical study. Am. Mus. Nvit. 606:1-24. Brattstrm, B. 1953. Recrds f Pleistcene reptiles and amphibians frm Flrida. Quart. J. Flrida Acad. Sci. 16:244-248. Btunner, G. ]9. r >6. achtrag /ur Kleinen Teufelshdhle bei Pttenstein (OberfrankenV F.in Cl>eigang vn der let/ten intergla/ialen Riss- Wiirm-W'arm-fauna /ur Wiirm-1-Kaltfauna. Dwnladed frm https://academic.up.cm/icb/article-abstract/5/2/319/183307 by guest n 03 Nvember 2018

SALAMANDERS 333 Jahrb. Gel. Pal. Mnakh. 1956:75-100. Carrll, R. 1964a. Persnal cmmunicatin.. 1964b. Early evlutin f the dissrphid amphibians. Bull. Mus. Cmp. Zl. Harvard 131:163-250. Dunn, E. 1941. The "percularis" muscle f salamanders. J. Mrph. 69:207-215. Eatn, T. 1959. The ancestry f mdern Amphibia: a review f the evidence. Univ. Kansas Publ. Mus. at. Hist. 12:155-180. Estes, R. 1963. Early Micene salamanders and lizards frm Flrida. Quart. J. Flrida Acad. Sci. 26:235-256.. 1964. Fssil vertebrates frm the late Cretaceus Lance Frmatin, eastern Wyming. Univ. Califrnia Publ. Gel. Sci. 49:1-180.. 1965. A new fssil salamander frm lintana and Wyming. Cpeia, 195:90-95. Gadw, H. 1901. Amphibia and reptiles. In S. F. Harmer and A. E. Shipley (eds.), The Cambridge natural histry, 668 p. MacMillan, Lndn. Geinitz, H. 1864. Palaesiren beinerti Gein., ein neues Reptil aus der unteren Dyas vn Oelberg bei Braunau. Neues Jahrb. Min. Gel. Pal., 1864, 513-516. Geinitz, H., and J. Deichmuller. 1882. Die Saurier der unteren Dyas vn Sachsen. Palaentgraphica 29:1-46. Cin, C, and W. AulFenberg. 1955. The fssil salamanders f the family Sirenidae. Bull. Mus. Cm par. Zl. Harvard 113:498-514.. 1957. A new fssil salamander f the genus Siren frm the Ecene f Wyming. Cpeia 2: 83-85.. 1958. New salamanders f the family Sirenidae frm the Cretaceus f Nrth America. Chicag at. Hist. Mus., Fieldiana, Gelgy, 10:449-459. Cin, C, and O. Gin. 1963. Intrductin t herpetlgy. W. H. Freeman and C., San Francisc, 341 p. Gregry, J. 1950. Tetrapds f the Pennsylvania!! ndules frm Mazn Creek, Illinis. Am. J. Sci. 248:833-873. Gregry, J. 1965. Micrsaurs and the rigin f Captrhinmrph Reptiles. Am. Zlgist 5 (2)., F. Pea bdy, and L. Price. 1956. Revisin f the Gymnarthridae, American Permian micrsaurs. Pea bdy Mus. Nat. Hist. Yale Bull. 10. 77 p. Hecht, M. 1957. A case f parallel evlutin in salamanders. Prc. Zl. Sc. Calcutta, Mkerjee Memr. Vl., 283-292.. 1963. A reevaluatin f the early histry f the frgs. Part II. Systematic Zl. 12:20-35., and R. Estes. 1960. Fssil amphibians frm Quarry Nine. Peabdy Mus. Nat. Hist. Yale Univ. Pstilla n. 46, 13 p. Herre, W. 1935. Die Schwanzlurche der mittelecanen (berlutetischen) Braunkhle des Geiseltales und die Phylgenie der Urdelen unter Einschluss der fssilen Frmen. Zlgica, Stuttgart, 33:1-85.. 1950. Schwanzlurche aus dem Palecan vn Walbeck. Zl. Anz. (Klatt Festschrift), 145: 286-301.. 1955. Die Fauna der mizanen Spaltenfiillung vn Neudrf a. d. March (CSR), Amphibia (Urdela). Sitz.-Ber Akad. Wiss. Wicn 164:783-803. -, and H. Lunau. 1950. eue fssile Schwanzlurche aus dem Burdigalium. Neues Jahrb. Gel. Pal., Mnatsh. 1950:247-252. Heyler, D. 1958. Revisin des Branchisaurus ilc la regin d'autun. Ann. Pal. 43:47-111. Kuhn, O. 1962. Die vrzeitlichen Frsche und Salamander, ihre Cattungen und Familien. ]h. Ver. vaterl. Naturde. Wurttenberg, 117:327-372. Larsen, J., Jr. 1963. The cranial stelgy f netenic and transfrmed salamanders and its bearing n interfamilial relatinships. Univ. Washingtn, Ph.D. thesis. Laurent, R. 1947. La dispsitin des dents vnieriennes chez la urdeles supcyieurs et sn imprtance phylgenetique. Bull. Mus. Ry. at. Belg. 23:1-4. Lazell, J., and R. Brandn. 1962. A new st)gian salamander frm the suthern Cumberland Plateau. Cpeia 2:300-306. Lehman, J.-P. 1955. Phyllspndyli. In Piveteau, J. (ed.), Traite de palent. Paris, Massn et Cie. 5:227-249. McGinnis, H. 1964. The skeletal mrphlgy f Pldegelhntia, a Permcarbniferus ai'stpd. Unpublished M.A. thesis, Univ. Califrnia. 154 p. Meszely, C. 1965. Nrth American fssil cryptbranchid salamanders. Am. Midi. Nat. (in press). Mnath, T. 1965. The jjercular apparatus f salamanders. J. Mrph. 116:149-170. Nev, E. 1964. Fssil urdeles in early lwer Cietacius depsits f Makhtesh Ramn, Israel. Natuie 201:415-416. Nble, G. 1927. The value f life-histry data in the study f the evlutin f the Amphibia. Ann. New Yrk Acad. Sci. 30:31-128.. 1931. The bilgy f the Amphibia. McGraw-Hill, ew Yrk, 574 p. (Dver Reprint, 1954). Olsn, K. 1965. Relatinships f Seymuria, Dindectcs, and Chelnia. Am. Zlgist 5 (2). Panchen, A. 1963. The hmlgies f the labyrinthdnt centrum. Prc. VI Internat. Cngr. Zl. Washingtn, D. C. 1:161. Parker, W. 1882. On the structure and develpment f the skull in the Urdeles. Trans. Zl. Sc. Lndn 11:171-214. Parsns, T., and E. Williams. 1962. The teeth f Amphibia and their relatin t amphibian ph)- lgeny. J. Mrph. 110:375-389.. 1963. The relatinships f the mdern Amphibia: a re-examinatin. Ouart. Rev. Bil. 38:26-53. Dwnladed frm https://academic.up.cm/icb/article-abstract/5/2/319/183307 by guest n 03 Nvember 2018

334 RICHARD ESTES Peabdy, F. 1954. Track\va)s f an ambystmid salamander frm the Palecene Mntana. J. Palen. 30:731-740.. 1959. Trackways f living and fssil salamanders. Univ. Calif. Publ. Zl. 63:1-72. Reed, H. 1920. The mrphlgy f the sundtransmitting apparatus in Caudate Amphibia and its phylgenetic significance. J. Mrph. 33: 325-387. Reig, O. 1964. El prblema del rigin mnfiletic plifiletic de ls anfibis, cn cnsideracines sbre las relacines entre anurs, urdels y apds. Amegheniana 3:191-211. Richmnd, N. 1964. Fssil amphibians and reptiles f Frankstwn Cave, Pennsylvania. Ann. Carnegie Mus. 36:225-228. Rmer, A. 1947. Review f the Labyrinthdntia. Bull. Mus. Cmp. Zl. Harvard 99:1-368. Salthe, S. 1965. Curtship prcedures and the phylgeny f the salamanders. Cpeia (in press)., and N. Kaplan. 1964. Micrcmplement fixatin and enz)me evlutin in the Amphibia. Am. Zl. 4:395 (Abstr.). Sauin, H. 1945. Amphibians frm the Dckum Triassic f Hward Cunty, Texas. Univ. Texas Publ. 4401:361-399. Set, T., C. Pmerat, and J. Kezer. 1964. The chrmsmes f Necturus as revealed in cultures f leuccytes. Am. Nat. 97:71-78. S/arski, H. 1962. The rigin f the Amphibia. Quart. Rev. Bil. 37:189-241. Teege, M.-J. 1957. Studien zur Entwicklung und Gestalt der Urdelenwirbel. Zeitschr. Wiss. Zl. 160:95-163. Thmsn, K. 1964. The ancestry f the tetrapds. Science Prgress 57:451-459. Tihen, J. 1958. Cmments n the stelgy and phjlgeny f amb)stmatid salamanders. Bull. Flrida State Mus. 3:1-50. Van Frank, R. 1955. Palaclaricha ligcenica, new genus and species, an Oligcene salamander frm Oregn. Mus. Cmpar. Zl. Harvard. Brevira 45:1-12. vn Huene, F. 1956. Paliintlgie und Phylgenie der Niederen Tetrapden. Jena, Gustav Fischer Verlag 716 p. Wake, D. 1964. Cmparative stelgy and evlutin f the Iungless salamanders, Family Plethdntidae. Univ. Suthern Califrnia, Ph.D. thesis. Watsn, D. 1940. The rigin f frgs. Ry. Sc. Edinburgh, Trans. 60:195-231.. 1956. The brachypid labyrinthdnts. Bull. British Mus. Nat. Hist. Gel. 2:315-392. Westphal, F. 1958. Die tertiaren und rezenten eurasiatischen Riesensalamander (Genus Andrias, Urdela, Amphibia). Palaentgr., Ab. A. 110: 20-92. Williams, E. 1959. Gadw's arcualia and the develpment f tetrapd vertebrae. Quart. Rev. Bil. 34:1-32.. 1965. Persnal cmmunicatin. Dwnladed frm https://academic.up.cm/icb/article-abstract/5/2/319/183307 by guest n 03 Nvember 2018

Des the cmparative anatmy text yu are nw using ffer yur students all these advantages? an up-t-the-minute, authritative treatment f the endcrine system discussins f the significance f the pineal bdy, thymus, urhypphysis, median eminence, hypthalamic-hypphyseal prtal system and the neurmast system chapter summaries cvering the entire scpe f each chapter citatins within the text t recent literature in areas f active current investigatin Yu'll find all these and many mre utstanding features in this fascinating new textbk Kent COMPARATIVE ANATOMY OF THE VERTEBRATES Fig. 60. A chick embry f apprximately 33 hurs incubatin. By GEORGE C. KENT. Jr.. Ph.D. Prfessr i Zlgy, Luisiana State University. Publicatin date: May. 1965. 471 pages. 7" x 10", with 344 illustratins. Price. $9.25. Here is ne f the mst up-t-date new textbks in this subject area, clearly written and attractively illustrated with mre than 340 drawings and diagrams. It ffers yur students a realistic understanding f the changing patterns f anatmy, the influences causing these changes, and the factrs making it pssible fr these changes t be perpetuated. The authr makes it abundantly apparent, withut belabring the pint, that: (1) species are mdificatins f species; (2) the adult is a mdificatin f the embry; (3) individual differences as well as species differences exist; and (4) structure and develpment are bradly determined by inheritance and adaptively mdified by natural selectin. The first 4 chapters f this new text cntain helpful backgrund infrmatin. The remaining 12 chapters discuss the basic architectural pattern f each f the 12 vertebrate systems and specific mdificatins as they are seen in representative vertebrates. In keeping with the recent dramatic advances in the investigatin f neurendcrine interrelatinships, this new bk emphasizes the physilgical rle f the endcrines, their cntrl by the envirnment, their relatinship t the nervus system and their influence n structure and functin. Yu will find interesting discussins f the nature f neursecretins, the neurhemal rgans, the hypphyseal prtal system, and the rle f the brain, especially f the diencephaln, in neurendcrine functins. Dwnladed frm https://academic.up.cm/icb/article-abstract/5/2/319/183307 by guest n 03 Nvember 2018 THE C. V. I0SBY COMPANY 3207 Washingtn Bulevard Publishers St. Luis, M. 63103

RECENT ADVANCES IN CETACEAN RESEARCH WHALES, DOLPHINS AND PORPOISES EDITED BY KENNETH S. NORRIS The first bk t prvide an interdisciplinary and internatinal apprach t current cetacean research. It represents the results f a unique meeting f physical scientists and bilgists whse research int a wide variety f scientific disciplines, including systematics, fisheries management, zgegraphy, natural histry, anatmy, physilgy, hydrdynamics, acustics, linguistics, and behavir, has brught them t share a cmmn interest in the mst recent advances in cetlgy. The chapters in this bk were selected frm amng the papers presented at the First Internatinal Sympsium n Cetacean Research, which was held in Washingtn, D. C. in 1963. The sympsium, cnducted by the American Institute f Bilgical Sciences, was spnsred by the Bilgy Branch f U. S. Naval Research. Kenneth S. Nrris is Assciate Prfessr f Zlgy at the University f Califrnia, Ls Angeles. 771 pages, 133 plates, 124 figures, 35 tables. $15.00 UNIVERSITY OF CALIFORNIA PRESS Berkeley 94720 NEW! Cllegiate Dictinary f ZOOLOGY ROBERT W. PENNAK University f Clrad The first ne-vlume dictinary t fully encmpass the wrking language f zlgy. Extensively crss-referenced bk supplies 19,000 entries defined as they are used in all zlgical disciplines. 1964. $8.50 Order direct frm: THE RONALD PRESS COMPANY 15 East 26th Street, New Yrk 10 Dwnladed frm https://academic.up.cm/icb/article-abstract/5/2/319/183307 by guest n 03 Nvember 2018