COMPARATIVE OSTEOLOGY AND EVOLUTION OF THE LUNGLESS SALAMANDERS, FAMILY PLETHODONTIDAE DAVID B. WAKE1

ABSTRACT: Lungless salamanders of the family Plethodontidae comprise the largest and most diverse group of tailed amphibians. An evolutionary morphological approach has been employed to elucidate evolutionary rela­ tionships, patterns and trends within the family. Comparative osteology has been emphasized and skeletons of all twenty-three genera and three-fourths of the one hundred eighty-three species have been studied. A detailed osteological analysis includes consideration of the evolution of each element as well as the functional unit of which it is a part. Functional and developmental aspects are stressed. A new classification is suggested, based on osteological and other char­ acters. The subfamily Desmognathinae includes the genera Desmognathus, Leurognathus, and Phaeognathus. Members of the subfamily Plethodontinae are placed in three tribes. The tribe Hemidactyliini includes the genera Gyri­ nophilus, Pseudotriton, Stereochilus, Eurycea, Typhlomolge, and Hemidac­ tylium. The genera Plethodon, Aneides, and Ensatina comprise the tribe Pleth­ odontini. The highly diversified tribe Bolitoglossini includes three super­ genera. The supergenera Hydromantes and Batrachoseps include the nominal genera only. The supergenus Bolitoglossa includes Bolitoglossa, Oedipina, Pseudoeurycea, Chiropterotriton, Parvimolge, Lineatriton, and Thorius. Manculus is considered to be congeneric with Eurycea, and Magnadig­ ita is congeneric with Bolitoglossa. Two species are assigned to Typhlomolge, which is recognized as a genus distinct from Eurycea. No. new information is available concerning Haptoglossa. Recognition of a family Desmognathidae is rejected. All genera are defined and suprageneric groupings are defined and char­ acterized. Range maps are presented for all genera. Relationships of all genera are discussed. Plethodontid salamanders have undergone an adaptive radiation char­ acterized by structural, functional, and ecological parallelism. The relatively primitive desmognathines and hemidactyliines have remained close to the an­ cestral habitat, mountain brooks and streams in eastern North America. Evo­ lutionary trends in the direction of abandonment of the aquatic larval stages and attainment of direct development are evident in desmognathines. The tribes Bolitoglossini and Plethodontini have direct development in terrestrial habitats. Terrestriality has stimulated evolution in these two successful groups and both have extended far beyond the ancestral range. Paedomorphosis has played a significant role in influencing evolutionary patterns within the family. In hemidactyliines, paedomorphosis has led to paedogenesis in four genera. Paedogenesis is a highly specialized but regressive adaptation in these groups. Paedomorphosis has profoundly influenced ter­ restrial groups by allowing them to achieve reproductive maturity at early structural stages. The phenomenon has led to evolutionary progress and may have been the key to the success of the terrestrial groups. Ancestral plethodontids probably arose from a stock close to that which gave rise to ambystomatids. The subfamily Desmognathinae is a specialized early derivative of this stock. Probable desmognathine remains from Creta­ ceous formations suggest that subfamilial differentiation was a Mesozoic event. The tribe Hemidactyliini is the most generalized group, both in struc­ ture and ecology, and is probably close to the ancestral familial stock. The tribe Bolitoglossini may have been the earliest derivative of the plethodontine stock. It ranges to western North America, Europe, and South America, and is the only salamander group that penetrates the tropics. In addition it is the only group of plethodontids that no longer occurs in eastern North America. The final derivative, the tribe Plethodontini, probably differentiated in early Tertiary times. The group has extended its range to western North America but still has relatively primitive representatives in Appalachia. lDepartment of Anatomy and The College, The University of Chicago, 102S East S7th Street, Chicago, Illinois 60637 1 2 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 INTRODUCTION Primitive species of plethodontids occur in east­ Despite considerable progress in studies of popula­ ern North America, primarily in Appalachia. Ad­ tion genetics and the processes of speciation, biolo­ vanced species have spread to western North gists have made little headway in understanding the America, Europe, and Central and South America. mechanisms .of macro- and megaevolution. The Because of the variety of adaptive types among goal of this comparative osteological study of the plethodontid species, the family is veritably a "liv­ salamander family Plethodontidae is to elucidate ing laboratory." It is an ideal group in which to problems relating to evolution above the species study patterns of vertebrate evolution, particularly level. Salamanders are generalized, evolutionarily speciation, mechanisms and patterns of adaptive conservative tetrapods. They were chosen for study radiation, and the origins of higher taxa. because available evidence indicates that evolution­ This investigation has several objectives. Skeletal ary rates have been relatively slow in the group. variation is analyzed, particularly to elucidate The family Plethodontidae includes a large variety evolutionary relationships on generic and higher of adaptive types among its diverse species. Many levels. Implications for taxonomy and phylogeny: intermediate adaptive stages are extant, thus facili­ are emphasized and the biogeographic history of tating analysis of major evolutionary events. Hope­ the family is reanalyzed on the basis of these data. fully, this study will provide groundwork for A functional and evolutionary approach is em­ comprehensive analyses of evolutionary patterns in ployed to focus attention on factors involved in the the family Plethodontidae, as well as in other invasion of new adaptive zones and subsequent groups. adaptive radiations, that is, on major features of The amphibian order Caudata includes nearly evolution in the family Plethodontidae. 300 species of salamanders. Almost two-thirds of The following brief sketches of the plethodontid these comprise the family Plethodontidae. The genera are presented to provide general background family is distinguished from all others by absence information for the reader. of lungs combined with presence of a cutaneous Salamanders of the genus Desmognathus (eight depression (the nasolabial groove) extending frorri species), the dusky salamanders, are found in the each nostril to the upper lip of transformed indi­ eastern portion of North America from Quebec and viduals. The most comprehensive treatment of the Nova Scotia to the Gulf Coastal Plain. The stocky, family is Dunn's (1926) pioneer monograph, agile members of this genus occur in a variety of which is now, unfortunately, badly out of date. habitats, but most are aquatic to semiaquatic in While papers by Piatt (1935), Taylor (1944), habits. The most aquatic species are the largest, and Soler (1950), Tanner (1952) and others have pro­ the smallest species are terrestrial and even semi­ vided additional information, the introduction to arboreal. All are quick, alert forms, adept at leap­ "The Salamanders of the Family Plethodontidae" ing and climbing. (Dunn, 1926) has been the best source of informa­ An aquatic genus, Leurognathus (one species), tion concerning structure and phylogeny. The most resembles the larger species of Desmognathus in recent general reviews of theories of familial evolu­ habitus. It occurs in mountain brooks in the south­ tion are given by von Wahlert (1957) and Martof ern Appalachian Mountains. (1962). A recently discovered genus, Phaeognathus (one A major adaptive radiation has been the out­ species), is a primarily terrestrial form apparently standing feature of the phylogeny of the family. restricted to a small area on the Coastal Plain of Primitive and generalized plethodontids live in Alabama. The single species, an elongate form with semiaquatic habitats and pass through extended very short legs, lives in burrows. larval periods during their ontogeny. Some species Gyrinophilus (three species) is a group of large, are completely aquatic, but advanced species range stout salamanders with elongated trunks and short, from semiaquatic to terrestrial, with many of the stocky tails. The species occur in springs, seepages, intermediate stages still represented by living rivulets, and cave waters from Quebec to northern species. Many advanced species have abandoned Alabama and Georgia. One species is permanently the aquatic larval stage and undergo direct terres­ larval. trial development; they occupy a large variety of The strikingly colored (brownish to bright red, terrestrial habitats, from cold and barren highlands with dark spots) members of the genus Pseudo­ to heavily forested tropical lowlands. Some of the triton (two species) are very stout, short-legged terrestrial species are arboreal, and a few are semi­ species which occur in springs and seepages in fossorial, but most are surface dwellers that take eastern North American from New York to Flor­ daytime refuge in burrows or under surface cover. ida, and west to Ontario, Ohio, and Louisiana. 1966 EVOLUTION OF LUNGLESS SALAMANDERS 3 The nearly aquatic salamanders of the genus species have flattened bodies and relatively long Stereochilus (one species) have elongate trunks and limbs and digits. All species have enlarged jaw relatively short, stout tails. The poorly known genus musculature. inhabits swampy and slow-moving waters on the The relatively short-bodied species of Hydro­ Atlantic Coastal Plain from Virginia to Georgia. mantes (five species) occur in terrestrial habitats The slender, active members of the genus in limestone areas in the mountains of California Eurycea (twelve species) range over most of east­ and in Europe. The salamanders are most com­ ern United States, with isolated populations on the monly found under rocks and logs, and in caves. Edwards Plateau of Texas. The salamanders occur Batrachoseps (three species) is a genus of small, in a variety of habitats, from stream sides and attenuate salamanders with four toes and very long swamps to springs, caves, and subterranean waters. tails. All are terrestrial and utilize underground Some species fail to metamorphose and achieve burrows. The animals are found along the West sexual maturity in a larval state. The genus Coast of North America from northern Oregon to Manculus (one species) is discussed separately in northern Baja California. There may be isolated the data sections of this paper, but is considered by populations in southeastern Alaska and on the me to be congeneric with Eurycea. Nevado de Colima, Jalisco, Mexico. The unpigmented, blind, cave-dwelling members The seven remaining genera are collectively of the genus Typhlotriton (one species) inhabit termed the neotropical genera in this paper. Boli­ subterranean waters and water edge habitats in toglossa (fifty-two species) is a large, diverse, wide­ caves of the Interior Highlands. spread group of terrestrial to arboreal species. Typhlomolge (two species) and Haideotriton Included are species that are among the largest (one species) are permanent larvae that are blind (B. robusta) and smallest (B. rufescens, B. orestes) and unpigmented. The larger, spindly-legged Typh­ of terrestrial salamanders. Highland species usually lomolge occurs in localized underground water have unwebbed to partially webbed hands and feet, systems at the edge of the Balcones Escarpment in while webbing is more extensive in lowland forms. Texas. Haideotriton, smaller and with shorter legs, The species occur from northeastern Mexico is found in underground waters of southern Geor­ through Central America to central Bolivia, the gia and northern Florida. mouth of the Amazon River in Brazil, and possibly Hemidactylium (one species) is a diminutive, as far southeast as southern Minas Gerais, Brazil highly distinctive inhabitant of swamps, sphagnum (Wake and Brame, 1966). bogs, and neighboring woods. It has a basally con­ The genus Oedipina (sixteen species) is a group stricted tail, only four toes, and is strikingly colored of highly distinctive, elongate species with extreme­ (clear white ventrally, with large black spots). The ly long tails, very short limbs, and attenuated genus ranges from southern Canada to the southern bodies. All are terrestrial. The genus ranges from Appalachian Mountains, with several disjunct southern Mexico through Central America to west­ southern and western populations. ern Colombia and northern Ecuador. Plethodon (eighteen species) is a large genus of Pseudoeurycea (twenty species) includes a rather terrestrial salamanders found in forest floor and variable group of terrestrial to semiarboreal sala­ woodland habitats. The salamanders vary consider­ manders found from northeastern Mexico to south­ ably interspecifically, from relatively large, stout, ern Guatemala. Although most of the species are short-bodied and long-legged species to relatively of relatively moderate size, the genus includes small, slender, elongate and short-legged species. species that are probably the largest terrestrial The genus occurs in many areas of eastern and salamanders (P. bellii, P. gigantea). northwestern North America, and in the Rocky A variable group of relatively diminutive sala­ Mountains. manders is included in the genus Chiropterotriton Ensatina (one species) is a stout-bodied, long­ (sixteen species) . Members of the genus are largely legged terrestrial salamander that ranges from arboreal, but some occur in forest floor habitats and southern British Columbia to southern California. others live mainly in caves. The genus is found The genus Aneides (five species) is a group of from northeastern Mexico through northern Cen­ terrestrial to arboreal salamanders found in the tral America, and also in Costa Rica. Appalachians, the high mountains of New Mexico, Lineatriton (one species) is a poorly known form and along the West Coast of North America. Ter­ from central Veracruz, Mexico. This terrestrial restrial members of the genus resemble generalized animal is a small, elongate form with extremely spe-cies of Plethodon and have cylindrical trunks shortened limbs and a long tail. and relatively short limbs and digits. Arboreal Members of the genera Thorius (nine species) 4 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 and Parvimolge (three species) are the smallest ice General Research Support Grant FR 5367 to species of salamanders. These diminutive, attenu­ the University of Chicago and from the Dr. Wal­ ated salamanders are found in terrestrial and ar­ lace C. and Clara A. Abbott Memorial Fund of boreal habitats in wooded highland areas. Thorius the University of Chicago have aided terminal occurs in southern Mexico, north of the Isthmus of portions of the work. Tehuantepec, and Parvimolge is found in Veracruz, Mexico and in Costa Rica. MATERIALS AND METHODS This study is based on a morphological analysis of ACKNOWLEDGMENTS plethodontid salamanders, with an emphasis on This work is a revised version of a doctoral disser­ osteology. Materials have been prepared for oste­ tation completed at the University of Southern ological examination in a variety of ways. Pre­ California. My chief debt of gratitude is to Jay M. served specimens have been measured and sexed, Savage, an inspiring teacher, who suggested the then cleared in potassium hydroxide, stained with problem and offered constant encouragement and Alizarin Red-S, and stored in glycerine. Measure­ guidance. My wife, Marvalee H. Wake, has read ments of size included in the text are the distance the manuscript, offered constructive criticisms, and in millimeters from tip of snout to posterior angle given her continuing moral support. Arden H. of vent. Cleared and stained material has provided Brame, Jr., John L. Mohr and George B. Rabb most of the skeletons used in the study. Many have participated in stimulating discussions con­ skeletons have been prepared by maceration and cerning portions of the work, and have made many disarticulation, and many others by dehydration helpful comments. Charles F. Walker loaned me a and subsequent cleaning by dermestid beetle larvae very important collection of cleared and stained and sodium hypochlorite solution. Skeletons of salamanders housed at the Museum of Zoology, every plethodontid genus have been utilized in the University of Michigan. Others who have con­ study, and include the following (number skele­ tributed to the study include A. P. Blair, James P. tonized, number cleared and stained; all measure­ Bogart, Charles M. Bogert, Ronald A. Brandon, ments and locality information on file): Desmog­ William E. Duellman, Robert E. Gordon, Alice G. nathus aeneus 0, 2; D. auriculatus 1, 6; D. fuscus C. Grandison, W.R. Heyer, Robert F. Inger, James 1, 6; D. monticola 5, 1; D. ochrophaeus 1, .S:; D. Kezer, Jens W. Knudsen, Alan E. Leviton, James ocoee 0, 4; D. quadramaculatus 4, 12; D. wrighti A. MacMahon, Bernard S. Martof, Hymen Marx, 0, 1; Leurognathus marmoratus 0, 8; Phaeognathus Hermano Niceforo Maria, James A. Peters, Philip hubrichti 1, 1; Gyrinophilus danielsi 1, 6; G. pal­ J. Regal, William J. Riemer, Francis L. Rose, Rob­ leucus 0, 1 ; G. porphyriticus 2, 2; Pseudotriton ert C. Stebbins, Russell H. Tuttle, Thomas M. montanus 2, 3; P. ruber 12, 10; Stereochilus margi­ Uzzell, Jr., Barry D. Valentine, Ernest E. Williams, natus 0, 3; Eurycea aquatica 0, 3; E. bislineata 2, and Richard G. Zweifel. The many courtesies 12; E. longicauda 2, 13; E. lucifuga 0, 1; E. multi­ shown me by Thelma Howell, Director, Highlands plicata 0, 2; E. nana 0, 5; E. pterophila 1, 3; E. Biological Station, during a summer of research at tynerensis 0, 1; E. troglodytes 0, 1; Manculus that institution are gratefully acknowledged. Tech­ quadridigitatus 0, 9; Typhlotriton spelaeus 0, 8; nical assistance has been provided by Addye C. Typhlomolge rathbuni 0, 1; T. tridentifera 0, 3; Brown, Les Siemens and James Everett. Mrs. Haideotriton wallacei 0, 2; Hemidactylium scuta­ Eleanor Willis of the Department of Anatomy, tum 0, 6; Plethodon cinereus 2, 8; P. dorsalis 0, 3; University of Chicago, typed the manuscript. P. dunni 3, 4; P. elongatus 2, 9; P. glutinosus 1, 11; Laboratory and library facilities were provided P. jordani 5, 9; P. larselli 0, 1; P. longicrus 0, 1; P. by the Allan Hancock Foundation of the University neome:xicanus 0, 3; P. ouachitae 0, 1; P. richmondi of Southern California. The National Science 0, 7; P. vandykei 0, 7; P. vehiculum 2, 8; P. wehrlei Foundation generously supported my studies in the 0, 1; P. welleri 0, 6; P. yonahlossee 3, 10; Ensatina form of three Cooperative Graduate Fellowships. eschscholtzii 6, 7; Aneides aeneus 3, 10; A. hardii In addition certain aspects of the study have been 0, 11; A. ferreus 3, 25; A. flavipunctatus 2, 13; A. supported by grants of the Society of Sigma Xi lugubris 5, 12; Hydromantes brunus 0, 5; H. genei (to D. B. Wake), National Institutes of Health, 0, 1; H. italicus 0, 1; H. platycephalus 0, 5; H. Institute of Arthritis and Metabolic Diseases, Grant shastae 0, 1; Batrachoseps attenuatus 3, 26; B. A-3549 (to J. M. Savage), and National Science pacificus 2, 25; B. wrightiO, 10; B. sp. nov. (Monte­ Foundation Grant GB 3868 (to D. B. Wake). rey) 0, 22; B. sp. nov. (Bodfish) 0, 1; Bolitoglossa Funds from the United States Public Health Serv- adspersa 0, 7; B. borburata 0, 1; B. cerroensis 0, 2; 1966 EVOLUTION OF LUNGLESS SALAMANDERS 5 B. colonnea 0, 1; B. dunni 0, 1; B. engelhardti 0, 2; were exposed for from thirty seconds to ten minutes B. fiaviventris 0, 1; B. helmrichi 0, 1; B. lignicolor at instrument readings of twenty-five to forty kilo­ 0, 1; B. marmorea 0, 2; B. mexicana 0, 2; B. morio volts and five DC milliamperes. Much osteological 0, 1; B. nigroflavescens 0, 1; B. occidentalis 0, 1; B. information has been obtained in this manner. All orestes 0, 1; B. platydactyla 0, 3; B. robusta 0, l; radiographs are on file at the University of Southern B. rostrata 0, 2; B. rufescens 0, 3; B. salvinii 0, 2; B. California. savagei 0, 2; B. striatula 1, 1; B. subpalmata 6, 13; Extensive field experience in the Appalachian B. schizodactyla 0, 1; Oedipina bonitaensis 0, Highlands and eastern United States, on the Ed­ 4; 0. complex 0, 1; O. cyclocauda 0, 1; O. gracilis wards Plateau of Texas, in the mountains of New 0, 2; 0. inusitata 0, 5; 0. longissima 0, 2; O. par­ Mexico, in the Pacific Northwest, in California, in vipes 0, 3; 0. poelzi 0, 4; 0. pacificensis 0, 1; 0. Costa Rica, and in Peru has provided me with syndactyla 0, 2; 0. uniformis 0, 1; 0. sp. nov. valuable background information concerning sala­ (Costa Rica) 0, 5; Pseudoeurycea altamontana 0, manders and habitats. 1; P. bellii 0, 1; P. cephalica 0, 2; P. expectata 0, 1; P. gadovii 0, 1; P. goebeli 0, 1; P. leprosa 0, 3; P. NON-OSTEOLOGICAL FAMILY rex 0, 1; P. robertsi 0, 1; P. scandens O, 1; P. smithi CHARACTERS 0, 2; P. unguidentis 0, 1; P. werleri 0, 2; P. sp. nov. (Mexico) 0, 1; Chiropterotriton abscondens O, 3; Lunglessness was first reported among plethodontid C. arboreus 0, 1; C. bromeliacia 0, 2; C. chiropterus genera by Wilder ( 1894) . Subsequently it has been 0, 4; C. dimidiatus 0, 5; C. multidentatus 0, 5; C. discovered that members of other salamander nasalis 0, i; C. priscus 0, 1; C. xolocalcae 0, 2; families (Hynobiidae, Ambystomatidae, Salaman­ Parvimolge richardi 0, 1; P. townsendi 0, 2; Linea­ dridae) may have reduced lungs, and a few species triton lineola 0, 4; Thorius dubitus 0, 1; T. mac­ of these families lack lungs, but lunglessness is dougalli 0, 3; T. pennatulus 0, 2; T. pulmonaris apparently a universal character of plethodontid 0, 5. species. Skeletal material has also been available for the A small depression, the nasolabial groove, ex­ families Hynobiidae (Hynobius), Cryptobranchidae tends from each nostril to the lip in all fully meta­ (Cryptobranchus, Megalobatrachus), Sirenidae (Si­ morphosed plethodontids. Associated with the ren), Ambystomatidae (Ambystoma, Dicamptodon, groove is a system of glands, which, in many ple­ Rhyacotriton), Salamandridae (Chioglossa, Cynops, thodontid genera, forms a swollen protuberance. Notophthalmus, Paramesotriton, Salamandra, Tari­ The swelling is usually more obvious in males than cha, Triturus), Necturidae (Necturus), Proteidae in females, and may be drawn into a slender cimis (Proteus), and Amphiumidae (Amphiuma). that extends ventral to the lower jaw. Many specimens of most plethodontid genera The above characters, in combination, define have been dissected in order to study musculature the family Plethodontidae. The family is comprised and duct and nerve pathways. These specimens of over one hundred and eighty species distributed · have been stored in seventy per cent ethyl alcohol among the following genera: Desmognathus, Leu­ and dyed with borax carmine-picric acid (for rognathus, Phaeognathus, Gyrinophilus, Pseudo­ muscles and nerves) and toluidine blue (for carti­ triton, Stereochilus, Eurycea, Typhlotriton, Typh­ ·1age). All observations have been made with a lomolge, Haideotriton, Hen;zidactylium, Plethodon, stereoscopic dissecting microscope with magnifica- Aneides, Ensatina, Hydromantes, Batrachoseps, . tions of from ten to ninety diameters. Bolitoglossa, Oedipina, Pseudoeurycea, Chiroptero­ ·mstological preparations of the tongues of triton, Parvimolge, Lineatriton and Thorius. The Batrachoseps pacificus, Bolitoglossa subpalmata, last seven genera are collectively termed the neo­ Chiropterotriton chiropterus, Ensatina esch­ tropical genera in this report. The species Eurycea scholtzii, Eurycea bislineata, Plethodon cinereus, quadridigitata is retained as the sole representative Pseudotriton ruber, and Typhlotriton spelaeus have of the genus Manculus in the data sections only. been available. Materials were preserved in 8 % Plethodontid genera vary considerably in their formalin, Bouin's solution, or Zenker's solution, ecological requirements. They may be grouped as . embedded in paraffin, and stained with Azan using follows (in part after Dunn, 1926; von Wahlert, · standard techniques. 1957): r. X-ray radiographs, mostly stereoscopic, have 1. Primarily or strictly aquatic. Leurognathus, !; been prepared for many species, particularly of Gyrinophilus, Pseudotriton, Stereochilus, Eury­ F those not well represented in collections and not cea, Manculus, Typhlotriton, Typhlomolge, r available for direct skeletal examination. Plates H aideotriton. 6 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 2. Semiaquatic or semiterrestrial. Desmognathus, likely that Phaeognathus, a genus with terrestrial H emidactylium. adults (Valentine, 1963 c; Brandon, 1965), also 3. Terrestrial. Phaeognathus, Plethodon, Ensatina, has direct development. It has long been stated as Aneides, Hydromantes, Batrachoseps, Bolito­ fact that Hydromantes and Oedipus (= the neo­ glossa, Oedipina, Pseudoeurycea, Chiroptero­ tropical genera) are ovoviviparous or viviparous triton, Parvimolge, Lineatriton, Thorius. (Dunn, 1926; Noble, 1931). Recently Gorman (1956) has reported that H. shastae lays eggs, and The above classification is over-generalized and has shown that earlier reports of ovoviviparity in there are many exceptions. Desmognathus alone the genus are questionable. Peters (1863) stated has species that may be aquatic (quadramaculatus), that Bolitoglossa adspersa bears living young, and semiaquatic (monticola), semiterrestrial (ochro­ he was followed by subsequent authors. Niceforo phaeus), or almost terrestrial (wrighti), and several Maria (1958), however, found that B. adspersa species of Eurycea are semiaquatic. lays eggs. Brame and Wake (1963) reviewed liter­ Life histories of only a few of the many pletho­ ature reports and showed that there is no evidence dontid species are known, but the general mode of ovoviviparity or viviparity in neotropical sala­ of reproduction of most genera is understood (see manders. Dunn, 1926; Noble, 1927 b, 1931; von Wahlert, Young males of all plethodontids have unilobed 1957). Aquatic larvae are known to occur in testes, but in certain genera an emptied lobe is Desmognathus, Leurognathus, Gyrinophilus, Pseu­ moved forward and a second lobe appears caudally, dotriton, Stereochilus, Eurycea, Typhlotriton, separated from the first by a short, slender, non­ Manculus, Typhlomolge, Haideotriton, and Hemi­ functional region. A new lobe is added in the dactylium, and all have been available for study. second, fourth, and sixth years of sexual activity in Desmognathus exhibits considerable interspecific Desmognathus, with the new lobes becoming func­ variation in length of the aquatic larval period. tional in the third, fifth, and seventh years (Humph­ According to Organ (1961 a) D. quadramaculatus rey, 1922; Organ, 1961 a). The maximum number has the longest larval period, and the period is suc­ of lobes is five (Humphrey, 1922). cessively shorter in D. fuscus, D. monticola, and D. Multiple testes are of common and apparently ochrophaeus. Young of D. aeneus emerge from the constant occurrence in old adult males of Desmog­ egg in a semimetamorphosed condition, with gills, nathus, Leurognathus, Phaeognathus, Batrachoseps, gill slits, nasolabial grooves, eyelids, and a greatly Bolitoglossa, Oedipina, Pseudoeurycea, Chiroptero­ reduced dorsal tail fin (Valentine, 1963 a). Reduc­ triton, Parvimolge, Lineatriton, and Thorius. Other tion of gills begins immediately, and transforma­ than in the above genera multiple testes have been tion is completed in less than a week. Finally, encountered only in single specimens of Typhlo­ Organ (1961 a, 1961 b) has shown that D. wrighti triton spelaeus, Eurycea pterophila, and Aneides has no free living larval stage, and hatchlings lack ferreus. gills and tail fins. Multiple testes may be used with caution as an Most of the aquatic and semiaquatic genera have indicator of relationships, and genera in which the stream larvae with low dorsal fins, reduced gills, character consistently occurs fall into two groups: and well developed hind limbs. Pond larvae with ( 1) Desmognathus and its allies, and (2) Batracho., high dorsal fins that extend onto the trunk and seps and the neotropical genera. slowly developing, reduced hind limbs are found The remainder of this report deals primarily in Hemidactylium (Blanchard, 1923; Bishop, with osteological data. Detailed descriptions of the 1941), Manculus (Goin, 1951), and Stereochilus skeleton of Aneides, including illustrations of skulls (Schwartz and Etheridge, 1954). and other features, have been presented previously Paedogenesis, or retention of larval non-repro­ (Wake, 1963) . Original observations concerning ductive morphology throughout life, is character­ osteological characters in additional genera are to istic of Typhlomolge, Haideotriton, Gyrinophilus be found in many papers, including in particular the palleucus, and certain species of Eurycea (latitans, following: Baird (1951), Cope (1859, 1866, nana, neotenes, pterophila, troglodytes, tynerensis). 1869), Dunn (1926), Emerson (1905), Hansen Terrestrial direct development with no aquatic and Tanner (1958), Hilton (1945 a, 1945 b, 1945 larval stage is characteristic of Plethodon, Aneides, c, 1945 d, 1946, 1947 a, 1948, 1959), Joubert Ensatina, Hydromantes, Batrachoseps, Bolito­ (1961), Kingsbury and Reed (1909), Moore glossa, Pseudoeurycea, Chiropterotriton, and Par­ (1900), Monath (1965), Noble (1921), Parker vimolge, and is presumed to occur in the terrestrial (1877, 1882), Piatt (1935), Reed (1920), Smith genera Oedipina, Lineatriton, and Thorius. It is (1920), Soler (1950), Taylor (1944), Wieder- 1966 EVOLUTION OF LUNGLESS SALAMANDERS 7 sheim (1875, 1877), and Wilder (1925). Informa­ dentalis, two tiny palatal processes, and two ascend­ tion presented here is limited to those elements ing and posteriorly directed frontal processes (Fig. that demonstrate significant intergeneric variation. 1 ) . The latter arise separately and embrace the internasal fontanelle, and in all small larvae are separated for their entire length. This condition bas ANALYSIS OF VARIATION been found in all larvae of Desmognathus, Leurog­ Osteological Characters nathus, Gyrinophilus, Pseudotriton, Stereochilus, Premaxilla Manculus and H emidactylium. The condition is Cope (1866) first separated genera with two pre­ also common in Typhlotriton and Eurycea, but ex­ maxillae from those in which the bones have fused, ceptional individuals of E. nana and E. neotenes and these characters were adopted by most sub­ have frontal processes that fuse posteriorly, behind sequent workers. Dunn ( 1926) concluded that the fontanelle. A single Typhlotriton larva has a premaxillae are primitively separated in pletho­ single premaxilla with frontal processes that are dontids. His conclusion was probably based on the broadly fused behind the fontanelle. The typical fact that separated premaxillae are the rule in fishes larval condition is found in all Haideotriton and and early amphibians and are present in all mem­ Typhlomolge. The life history of Phaeognathus is bers of the primitive salamander families Crypto­ unknown, and all remaining genera undergo direct branchidae, Hynobiidae, and Ambystomatidae. development. In 1943 Grohman reported the appearance dur­ ing the ontogeny of Gyrinophilus of a suture be­ tween previously fused premaxillae. Larvae have a single premaxilla, but during metamorphosis a suture forms. In 1959 Grohman noted that a single premaxilla is present in the larvae of Gyrinophilus, Pseudotriton, and Eurycea, and stated "this may be considered primitive in the Salamandroidea al­ though it is undoubtedly a specialized situation in comparison with the conditions prevailing in the hynobiids and remote ancestral stocks." Grohman suggested that fused premaxillae be regarded as -1------;,..--•Parietal the ancestral condition for these genera, and that fusion be considered less specialized than separa­ Figure I. Anterior cranial elements of a small (23.9 tion. The condition in Gyrinophilus was considered mm.) larval Gyrinophilus danielsi. Drawn from a an advance over that of the other genera, in which cleared and stained individual. the premaxillae remain fused. Martof and Rose (1962) suggested that a single Despite numerous complexities and considerable larval premaxilla is the rule in plethodontids, as variation two premaxillary conditions are recog­ earlier suggested by Grohman. They examined very nizable in adult plethodontids. In most, a single small larvae of Gyrinophilus, Pseudotriton, Eu­ premaxilla is present and the frontal spines may be rycea, Desmognathus and Leurognathus and found separated, fused in front of the fontanelle, fused a single element with no evidence of two centers of behind the fontanelle, fused in front of and behind ossification. Rose and Bush (1963) accepted the the fontanelle, or fused for their entire length. This reasoning of Grohman (1959) that a single element general situation is found in Desmognathus, is primitive in plethodontids and rejected Dunn's Leurognathus, Phaeognathus, Pseudotriton, Stereo­ (1926) proposal that paired elements are primitive. chilus, Eurycea, Manculus, Typhlotriton, Typh­ They stated that fusion or separation of parts of lomolge, Haideotriton, Aneides, Bolitoglossa, the larval premaxilla, a single dentigerous rod with Oedipina, Pseudoeurycea, Chiropterotriton, Parvi­ two frontal processes, should be considered ad­ molge, Lineatriton, and Thorius. The other general vanced. Thus Eurycea, which usually retains a condition in which two premaxillae are present in larval-like structure, was considered more primitive adults and separated for their entire lengths is than Pseudotriton, in which the frontal processes found in Gyrinophilus, Hemidactylium, Plethodon, fuse, and Gyrinophilus, in which the premaxillae Ensatina, and Hydromantes. Grobman (1943) has separate. shown that small Gyrinophilus have a single pre­ Plethodontid larvae of all stages have a single maxilla, and Martof and Rose ( 1962) report a premaxilla which consists of a curved, toothed pars single premaxilla in small Plethodon and H emidac- 8 MEMOIR OF THE SOUTHERN CAUFORNIA ACADEMY OF SCIENCES VOL.4 tylium. One element is present in the smallest Ambystomatidae two premaxillae are present in Hydromantes (27.5 mm. standard length) exam­ larvae of all genera. Maxillae appear earlier during ined, but it is apparent that a suture is forming. Two ontogeny than in the plethodontids, and are moder­ premaxillae are present in all Ensatina examined ately well developed before metamorphosis proper (as small as 23.l mm.), with no sign of fusion. is under way. They are present in the permanently Exceptional conditions are found in some in­ larval arnbystomatids. In plethodontids maxillae do dividuals of Typhlotriton, Chiropterotriton, and not form until late in metamorphosis, and no maxil­ Thorius. A tendency toward premaxillary separa­ lae are present in the permanent larvae. In amby­ tion seems evident in large, old adults of Typhlo­ stomatids the premaxillae are attached to the skull triton. Average adults invariably have a single proper by means of articulations with the frontals, element, but with increasing size the toothed bar the vomers, and the maxillae. In plethodontids the narrows on the midline. In one large adult (55.3 premaxilla projects forward from the skull proper, mm.) a distinct suture is present and two premaxil­ to which it is attached only by the tips of the frontal lae occur. Curiously, there seems also to be a processes and by a tenuous articulation with the tendency toward fusion of frontal processes in vomers (Fig. 1). Two premaxillae projected an­ large Typhlotriton, although the individual with teriorly from the skull proper, lacking lateral articu­ two premaxillae had separated frontal processes as lations, are considerably less rigid than a single well. A single adult Chiropterotriton xolocalcae has structure resulting from fusion at the anterior ex­ two premaxillae which are joined by a single tooth tremity of the formerly paired structures. I suggest pedicel, a situation also rarely encountered in that the premaxillary condition observed in pletho­ Plethodon, which normally has separated elements. dontid larvae is an example of larval evolution, or One adult Thorius pennatulus has two distinctly caenogenesis (see DeBeer, 1958), which has re­ separated premaxillae. Ontogenetic series of these sulted from selection operative at early larval stages species have not been available for study. of the ancestral stock. The larvae of plethodontids Specialized species of Batrachoseps (B. attenua­ are carnivorous; they seize and manipulate prey tus, B. pacificus, two species nova) have a single by the closing of the premaxilla and dentaries on premaxilla. In B. attenuatus, B. pacificus, and one each other. The stronger the premaxilla, the more of the new species (Bodfish), the frontal processes efficient would be the feeding process, and indi­ arise separately, then fuse almost immediately and viduals having a single stronger structure would remain fused as far as the anterior margin of the probably be favored over those with paired, weak­ fontanelle, where the processes separate and di­ er elements. The attainment of the single larval verge. The processes fall short of the posterior premaxillary condition has doubtless been of im­ margin of the fontanelle. A small fontanelle, often portance in the success of plethodontids. triangular but sometimes reduced to a tiny circle, is Metamorphosis is a period of drastic molding present immediately above the pars dentalis and. and reorganization in plethodontids (Wilder, below the point of frontal process fusion. In the 1925). During metamorphosis several skeletal ele­ other new species (Monterey) eleven of fifteen in­ ments are lost, others are gained, and the remaining dividuals have the condition discussed above, but skull elements are extensively remodeled. In con­ the other four have frontal processes that remain sidering the evolution of plethodontids it is thus separated. The most primitive member of the essential that the total metamorphic process be genus, B. wrighti, has a single premaxilla in small considered apart from the changes that take place stages, but the frontal processes are separated. That in any given element. While this subject is con­ there is ontogenetic variation is evident. Two small sidered in detail elsewhere (see below), it bears individuals (21.8 mm., 29.1 mm.) have a single directly on the question of premaxillary evolution premaxilla. A median depression is present in the and is thus raised here. toothed bar of slightly larger individuals (32.4 It has long been customary to assume that one mm.), and at later stages (36.3 mm., 38.3 mm.) can find a more primitive condition in an earlier the single element of small individuals is separated rather than a later developmental stage of a given by a suture into two distinct halves, connected by species, and it is apparent that Grohman ( 1943) the pedicel of a single median tooth. As individuals and Rose and Bush (1963) have followed this become progressively larger (above 40 mm.), two reasoning in their work. Such practices have been premaxillae are present with no sign of fusion. decried by DeBeer (1958) who states numerous The larval and adult premaxillary conditions objections and offers many alternative suggestions must be considered separately if a cogent theory to explain commonly encountered phenomena. In is to be advanced to explain observed facts. In the the case of the plethodontid premaxilla it is clear 1966 EVOLUTION OF LUNGLESS SALAMANDERS 9 that the metamorphic process must be considered morphic pattern is retained (Fig. 2). The pattern is apart from the element itself. One must distinguish similar in Hemidactylium, Hydromantes, Pletho­ between a primitive metamorphic pattern, and the don, and presumably Ensatina. The metamorphic presumed primitive nature of a retained larval process is slowed in Typhlotriton, and is drawn out character. over a long period of an organism's life, but the Metamorphosis in primitive salamanders (hyno­ premaxillae of the largest individuals may finally bilds, ambystomatids) results in paired, separated separate. In regard to premaxillae the exceptional premaxillae. It is clear that such a condition is the Typhlotriton can be said to undergo delayed meta­ end result of a primitive metamorphic pattern and morphosis. Additional support is gained from the should be considered primitive regardless of the genus Batrachoseps. Four of the five species are condition present in larvae. Plethodontid larvae specialized and demonstrate numerous paedomor­ may be considered advanced over larvae of other phic tendencies, but the fifth, B. wrighti, is less families in the fused condition of their premaxillae. specialized and more primitive. Delayed metamor­ Premaxillary fusion in adult plethodontids is not phosis occurs in B. wrighti which results, in old primitive even though it is the condition of the adults, in two premaxillae while the other species larvae. There is considerable evidence that the never undergo premaxillary metamorphosis. primitive metamorphic pattern is retained in sev­ The evidence cited above concerning Typhlo­ eral primitive plethodontids, with resultant pre- · triton and Batrachoseps suggests that fused pre­ maxillary separation. My interpretation of the maxillae in adults of these genera (and possibly in events transpiring during metamorphosis in Gyrin­ Chiropterotriton and Thorius, see above) is the ophilus, for example, is that the primitive meta- result of paedomorphosis. Paedomorphosis also accounts for the fused premaxillae in Typhlomolge, H aideotriton, Gyrinophilus palleucus, Eurycea pterophila, E. nana, and E. tynerensis. It is prob­ able that the fused premaxillae in Pseudoeurycea, Chiropterotriton, Bolitoglossa, Oedipina, Parvi­ molge, Lineatriton, and Thorius are due in part to paedomorphosis. All are affected by paedomor­ phosis in other ways (see below), and the pre­ maxilla is relatively very small in the group. A simplification of parts has taken place in the neo­ tropical genera, however, and it is possible that an evolutionary trend toward simplification has been operative in unison with a trend toward meta­ morphic delay or failure. The skulls of Pseudotriton, Desmognathus, Leurognathus, Phaeognathus, Stereochilus, and Aneides appear to have been secondarily strength­ ened. All have solidly articulated skulls and all may have fused frontal processes. Fusion of premaxillae in adults of these genera is probably the result of selection in favor of increased strength of skull elements. A similar explanation can be advanced in regard to Eurycea and its derivative, Manculus. Eurycea may have arisen from an ancestor with a single premaxilla and a solid skull (Fig. 2). Skull weakening in advanced Eurycea is apparently a secondary event. In the Plethodontidae it can be assumed that Figure 2. Trends in premaxillary evolution. A. 'IYpical ~ of plethodontid larvae and embryos ( Gyrino­ larval and embryonic premaxillary fusion is the "6ilus danielsi), B. Adult pattern of primitive genera rule. The primitive metamorphic pattern results (Gyrinophilus danlelsi), C.-0. Adult patterns of ad­ in two premaxillae. I consider any modification of YaDced genera (C. Desmognathus monticola, D. Eury­ this primitive metamorphic pattern to represent an "'° bislineata, E. Pseudoeurycea leprosa, F. Stereo­ lAilua marglnatus, G. Oedipina parvipes). Not drawn evolutionary advance, whether the result of paedo­ to scale. morphic influence or of selective pressures favoring 10 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 the strengthened premaxilla which results from matidae), and may reflect the relative antiquity of fusion. In this discussion I have limited the topic the genus and its proximity to the familial ances­ to fusion or non-fusion of the toothed portions of tral stock. It is not firmly established, however, that the premaxillae. Fusions of the frontal processes short frontal processes are primitive. In the above occur only in forms in which fusion of the .toothed group palatal processes are generally moderately portions has occurred. developed, but the palatal shelf of Gyrinophilus is Several groups are recognizable on the basis of relatively broad. metamorphic effects on premaxillary structure, and The premaxillae of Desmognathus, Leurogna­ details of their premaxillary anatomy are discussed thus, and Phaeognathus have a characteristic struc­ below. The larval premaxilla is retained in rela­ ture, one not closely approximated by other pletho­ tively unmodified form in Typhlomolge and Haide­ dontids. Perhaps the most diagnostic features are otriton. The primitive metamorphic pattern is re­ the compactness, the close articulation of the pre­ tained in Gyrinophilus, Hemidactylium, Plethodon, maxilla with surrounding elements, and the close Ensatina, and Hydromantes. The final group in­ association of the frontal processes and toothed cludes genera that have some modification of the bar. The skulls of all are depressed, especially primitive metamorphic pattern, and may be sub­ Leurognathus, and it is impossible to distinguish divided as follows: (1) Desmognathus, Leurog­ between pars dentalis and pars frontalis where the nathus, Phaeognathus, (2) Pseudotriton, Eurycea, two areas join. The lateral margins of the pars Manculus, Typhlotriton, (3) Stereochilus, (4) dentalis extend posteriorly and a little dorsally and Aneides, (5) Batrachoseps, (6) Bolitoglossa, Oedi­ become the lateral margins of either pars frontalis. pina, Pseudoeurycea, Chiropterotriton, Parvimolge, The frontal processes of larvae resemble those of Lineatriton, Thorius. primitive plethodontid genera and have distinct The skulls of Haideotriton and Typhlomolge are shanks and blades. During ontogeny bone is added strongly depressed and the premaxillae are modified to the lateral margins of the shanks and the proc­ accordingly. The single pars dentalis is relatively esses become relatively very broad. The distance broader than in larvae, and is a straight structure, across the shanks is only slightly less than the width posteriorly recurved only at the lateral tips. Frontal of the toothed bar (Fig. 2). A small internasal processes arise separately and extend almost di­ fontanelle is embraced by the frontal processes and rectly posteriorly with little dorsal rise. The proc­ is increasingly reduced in size during ontogeny. In esses are very broad in Typhlomolge, narrower in old adult Leurognathus the fontanelle is reduced to Haideotriton, and in both embrace the internasal a tiny opening or is entirely closed (Moore, 1899). fontanelle. Frontal processes remain separate in Frontal processes fuse behind the fontanelle in the genera but may articulate behind the fontanelle. Leurognathus, Phaeognathus, and some Desmog­ Palatal portions are poorly developed. nathus (fuscus, auriculatus, ochrophaeus, monti­ In Gyrinophilus, Hemidactylium, Plethodon, cola), while in other Desmognathus (aeneus, ocoee, and Ensatina frontal processes embrace a moderate wrighti) the processes articulate but do not fuse. to large fontanelle but remain separated. The The relatively primitive D. quadramaculatus has processes are longest in Gyrinophilus where they unfused processes for a long time (fused in indi­ expand distally and articulate with each other be­ viduals more than 80 mm.). The posterior tips of hind the fontanelle. Post-fontanelle expansion is the frontal processes in this group of genera are less well developed in the other genera. The proc­ lateral in position, and following fusion the poste­ esses may articulate posterior to the fontanelle in rior border is concave. The palatal shelves are some species of Plethodon (see Wake, 1963), but relatively the broadest and best developed in the not in Hemidactylium or Ensatina. A unique condi­ family. tion is encountered in Hydromantes in which the Pseudotriton and Eurycea illustrate a tendency frontal processes are extremely short, especially in for fusion of the frontal processes posterior to the the American species. The slender processes fontanelle. In the smallest Pseudotriton larva avail­ abruptly terminate at about the anterior margin of able (17.2 mm.) the frontal processes have already the nasals, well anterior to the posterior end of the encircled the fontanelle and are in the process of fontanelle. In addition the anterior portion of the fusing. In more advanced larvae and in all adults snout is relatively deep in Hydromantes and the the frontal processes are solidly fused. Typical un­ frontal processes are almost vertically oriented. The fused frontal processes are found in Manculus and situation in Hydromantes, more than any other most Eurycea (tynerensis, longicauda, lucifuga, plethodontid, resembles that encountered in primi­ multiplicata). In the other Eurycea a trend toward tive salamander families (Hynobiidae, Ambysto- fusion of frontal processes is evident. A single 1966 EVOLUTION OF LUNGLESS SALAMANDERS 11 E. nana, a single E. pterophila, and all available odon is superficial, related to premaxillary separa­ E. aquatica have fused processes. Rose and Bush tion in organisms of roughly similar size. (1963) report that three of five transforming A single premaxilla occurs in all species of E. aquatica and eleven of twelve adults had fused Aneides, but in other regards the element resembles processes. Wilder ( 1924) reported that 4.5 per cent the paired premaxillae of Plethodon (Wake, 1963). of 109 larval E. bislineata from Massachusetts bad A distinct trend in the direction of increased fused frontal processes. All E. bislineata larvae ex­ strength of skull elements is evident in Aneides, amined by me have unfused processes. Rose and and is reflected by the premaxilla. A single element Bush ( 1963) report unfused processes in eleven of appears to represent a specialization within the twelve E. bislineata (size not specified) from Ala­ Plethodon-Ensatina-Aneides assemblage. An addi­ bama, Mississippi, and Louisiana. My data indicate tional modification of the premaxilla occurs in large an ontogenetic trend toward process fusion in the adult A. lugubris, possibly the most advanced and species. Processes are fused in four of six adults certainly the most specialized species. The frontal (37.3 - 39.5 mm.) from North Carolina, but are processes are very long, equaling or exceeding those separated in the two smallest individuals (36.9, in any other genus, and tend to fuse behind the 37 .2 mm.). Frontal processes are partially to fontanelle. Secondary bony growth in lugubris roofs totally fused in three adults (37.4 - 44.4 mm.) the fontanelle. from Indiana, but are not fused in one large The advanced species of Batrachoseps have a pe­ adult ( 42.5 mm.). It is obvious that geographic culiar fusion of the frontal processes above the pars and ontogenetic variation occurs. An unusual con­ dentalis but below the fontanelle; this condition is dition is evident in the frontal processes of Pseudo­ not encountered in any genus discussed so far. The triton in that, following fusion, secondary growth frontal processes are fused from their origin to the occurs and a single process, usually with a bifur­ fontanelle in Stereochilus, but in Batrachoseps the cated tip, grows posteriorly. Thus the postfontanelle processes arise separately, fuse for a very short portion of the frontal processes is large relative to distance, then separate and diverge posterolaterally other genera. Palatal shelves range from small around the fontanelle. Such a situation is found (Manculus) to moderately large (Pseudotriton) in elsewhere among plethodontids only in the neo­ the group. The premaxilla of Typhlotriton closely tropical genera. In most plethodontids the fonta­ resembles that of Eurycea. Typhlotriton may have nelle tends to be enclosed on all sides by the frontal paired premaxillae, however, and in a sense it processes, and a trend for fusion of the frontal joins this group of genera with Gyrinophilus, which processes posterior to the fontanelle is evident. In it resembles in many other features. Batrachoseps and the neotropical genera fusion of Perhaps Stereochilus should be grouped with the processes occurs in front of the fontanelle, if Pseudotriton, Eurycea, and Manculus. The entire at all, and the fontanelle tends to be pushed pos­ skull of Stereochilus is rather compressed and teriorly in reference to the premaxilla. The pro­ elongate, and the skull elements, including the cesses themselves are relatively shorter (relative to premaxilla, are modified accordingly. In adults, over-all skull dimensions) than in the more primi­ frontal processes are fused and arise from the pars tive genera, and there is a trend toward decrease dentalis as a single element. The processes separate in the over-all size of the premaxilla. at a moderate distance from the toothed bar to The neotropical genera resemble the advanced encircle the relatively small fontanelle. In small species of Batrachoseps in premaxillary structure .adults the processes completely encircle the fontan­ more closely than do any other plethodontid genera. elle and articulate with each other posteriorly. The The most striking difference between the two processes are in contact for a considerable distance groups is the great reduction seen in the neotropical · posteriorly but separate near their tips. In larger species, particularly in the pars dentalis. While a adults, the fontanelle is almost completely closed general reduction in size of the premaxilla has over. The frontal processes are completely fused taken place in Batrachoseps, the reduction is most anterior and posterior to the fontanelle and separate evident in the palatal and frontal processes, and the · only at the posterior exeremity (Fig. 2). Palatal pars dentalis remains a relatively large structure. shelves are moderately large. In the neotropical species the pars dentalis is ob­ · Hemidactylium has a larval premaxilla very simi­ viously very shortened relative to the total size of , lar to that of larval Eurycea. Adult premaxillae the premaxilla. This character separates the neo­ ; have relatively long toothed portions and broad, tropical genera from all other plethodontids. Palatal jttout palatal shelves. In these features they resemble processes tend to be poorly developed or absent ~Eurycea and related genera. Resemblance to Pleth- in the entire group. 12 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 The pars dentalis is always relatively small in the is formed by the frontal processes. It is instructive neotropical genera, but in specialized species of to follow the ontogenetic changes in 0. poelzi. In several genera (Chiropterotriton, Bolitoglossa, juveniles the frontal processes arise separately and Oedipina) and in all Thorius the portion is ex­ remain separated. The processes are very close to tremely reduced. The extreme is reached in 0. each other for most of their length and diverge parvipes in which the length is as great as the slightly only near the terminus. In small adult breadth and the structure is barely large enough to 0. poelzi the processes are fused from their origin support a single tooth pedicel (Fig. 2) . for about one-third of their length. The processes In all Parvimolge examined the frontal processes of old adults are fused from their origins for from arise separately and remain separated for their en­ one-half to three-fourths of their length. In general tire length. In all other neotropical genera at least the frontal processes in all Oedipina are fused for some fusion is present. one-third to three-fourths of their length, less in Pseudoeurycea normally has separated frontal young and more in old individuals. 0. parvipes processes, but a single P. leprosa has processes that illustrates the extreme condition in which a single, arise separately, then fuse, and finally separate to totally undivided frontal process arises from the diverge at the fontanelle. In P. smithi the processes midline of the pars dentalis and proceeds poster­ arise separately in small individuals, but are fused iorly to the anterior margin of the fontanelle, where from the pars dentalis to the fontanelle in large it is only slightly expanded (Fig. 2) . adults (Fig. 2). In summary, primitive premaxillae that separate Frontal processes arise and remain separated in relatively early in ontogeny are found in five rela­ all Chiropterotriton examined with the exception tively primitive genera of plethodontids, and in the of the specialized C. abscondens and C. nasalis. The most primitive species of the advanced Batracho­ processes of abscond ens and nasalis have a common seps. Premaxillary fusion occurs in forms with origin and remain fused until the fontanelle is strengthened skulls; fusion thus may either be the reached at which point they separate and diverge. result of a strengthening adaptation in the species, Frontal processes are solidly fused from their or of derivation from a stock in which such an origin on the pars dentalis to the somewhat pos­ adaptation had occurred previously. Secondary teriorly placed fontanelle in Lineatriton. At the fusions of the frontal processes also occur in some anterior margin of the fontanelle the processes sep­ of these forms. Fusions related to skull strengthen­ arate. The fontanelle is very narrow anteriorly, and ing have arisen separately in groups containing the the processes remain very close for about one-half following genera: the distance from the point of separation to the 1. Pseudotriton, Eurycea, Manculus, Typhlotriton, terminus. Most of the fontanelle is located pos­ Stereochilus. teriorly and is bounded by the anterior margins of the frontals. 2. Aneides. The pars dentalis of Thorius is very small but 3. Desmognathus, Leurognathus, Phaeognathus. gives rise to paired, extremely slender frontal pro­ Premaxillary fusion is characteristic of paedo­ cesses. In most individuals the processes arise sep­ morphic forms (retention of the larval premaxilla). arately and remain separated for about two-thirds It is difficult to trace the parallelism, but fusion of their length before diverging around the pos­ must have occurred at least as many as four times: teriorly placed fontanelle. The processes may be fused for short distances in front of the fontanelle 1. Gyrinophilus palleucus. in T. pulmonaris and T. macdougalli. 2. Typhlomolge, Haideotriton, and paedomorphic The frontal processes of most Bolitoglossa are species of Eurycea. separated for their entire lengths, and are relatively 3. Neotropical genera. long and well developed. In some species (robusta, 4. Advanced species of Batrachoseps. borburata) the processes embrace the fontanelle, and are rather broadly expanded posterior to the My ideas concerning morphological trends and fontanelle. In others the processes may arise sep­ relationships of premaxillary structure are illus­ arately, fuse at a point, then diverge around the trated in Figure 2. On the basis of total premaxillary fontanelle (adspersa, orestes, savagei, subpalmata). structure it is possible to place plethodontid genera The most specialized premaxillary conditions are in the following groups: achieved in the genus Oedipina. The fontanelle has 1. Gyrinophilus, Pseudotriton, Eurycea, Manculus, moved posteriad in reference to the premaxilla, and Typhlotriton, Typhlomolge, Haideotriton, Ster­ only the extreme anterior margin of the fontanelle eochilus, Hemidactylium. -1966 EVOLUTION OF LUNGLESS SALAMANDERS 13 2. Hydromantes. prefrontals and the space normally occupied by the 3. Plethodon, Ensatina, Aneides. prefrontals has been invaded by the maxillae. Simi­ 4. Desmognathus, Leurognathus, Phaeognathus. lar instances of bones enlarging into areas vacated by other elements have been discussed by Parring­ 5. Batrachoseps, Pseudoeurycea, Chiropterotriton, ton ( 19 5 6) . The nasals are very small in these Parvimolge, Lineatriton, Thoriw, Bolitoglossa, genera; the result of maxillary enlargement is that Oedipina. the facial lobes greatly exceed the nasals in size, a It is significant that each of the above groups situation not found elsewhere among the pletho­ except 4 has at least one genus that includes species dontids. with the proposed primitive condition-unfused In primitive salamander families (Hynobiidae, premaxillae. The fossil relative of Desmognathus Ambystomatidae) facial lobes arise from the an­ and allies, Prodesmodon, had unfused premaxillae terior portion of the pars dentalis and terminate (Estes, 1964); all modern genera of Group 4 have before the mid-point. The moderate-sized elements greatly strengthened skulls and it is likely that the in a pre-midpoint position encountered in Gyrino­ condition in Prodesmodon was primitive. philus and others are generalized, and probably resemble the ancestral condition. The conditions Maxilla in Batrachoseps (anterior movement), Desmog­ Facial lobes arise from the second quarter (from nathus and its allies (enlarged size with very broad anterior) of the maxillary pars dentalis in most bases), and Plethodon and its allies (posterior plethodontid genera, and the bulk of each lobe is movement) appear to be derived, specialized anterior to the midpoint of the maxilla. The lobes situations. of Batrachoseps arise from the anterior one-third Nasals articulate firmly and extensively with the of the pars dentalis, and the structures are well in pars facialis of the maxillae only in Desmognathus, advance of the midpoint. This forward movement Leurognathus, Phaeognathus, and Aneides lugu­ is greater than in any other plethodontid, and is bris. Moderate articulation is found in Gyrino­ probably related to the general shortening of the philus, Typhlotriton, and Stereochilus, and rather snout characteristic of the genus. No obvious an­ limited articulations are found in a number of terior movement of the lobes is seen in other species genera (Pseudotriton, Eurycea, Plethodon, Aneides, with greatly shortened snouts (e.g., Bolitoglossa Hydromantes, Batrachoseps, Bolitoglossa, Pseudo­ orestes). The only other genera in which the an­ eurycea, Chiropterotriton, Parvimolge, Lineatri­ terior margins of the facial lobes approach the ton). In many of these latter genera articulation is anterior tip of the pars dentalis are Desmognathus, limited, and may consist simply of a meeting at a Leurognathus, and Phaeognathus, in which the single point. Primitively the space between the lobes are very large and extend posteriorly beyond nasals and the maxillae is occupied by prefrontals, the midpoint of the element. lacrimals, or both (Ambystomatidae, Hynobiidae), , Facial lobes of the remaining genera arise well and extensive articulation is a specialized condition. posterior to the anterior tips of the maxillae. The In adult Gyrinophilus, nasals and maxillae are in lobes are primarily anterior to the midpoint of the contact, but prefrontals extend to the anterior mar­ pars dentalis and generally terminate anterior to gin of the nasals below the articulation. Nasals that point in Gyrinophilus, Pseudotriton, Stereo­ usually do not articulate with the maxillae in 'chilus, Eurycea, Manculus, Typhlotriton, and Pseudotriton, and prefrontals extend between the Hemidactylium. The lobes are of moderate size in elements to the anterior margins of the nasals. This most of these, but are relatively small in Eurycea situation is also encountered in some Eurycea, and and Manculus. In Plethodon, Ensatina and Aneides in Manculus, Typhlotriton, Hemidactylium, most · the lobes tend to arise from the central one-third Plethodon, and some Aneides and Ensatina; it is of the pars dentalis, and usually extend past the probably close to the ancestral plethodontid midpoint. The lobes are relatively small in Ensatina, condition. larger in Plethodon, and largest in Aneides. Hydro­ The maxillae are long, extending beyond the mantes has a moderate..gized facial lobe lying an- posterior margins of the eyes in Desmognathus, : . 1eriorly and terminating at about the midpoint of Leurognathus, Phaeognathus, Stereochilus, and the maxilla. The relatively small facial lobes of the Typhlotriton, and terminating at about the posterior e.eotropical genera terminate at about the midpoint margins of the eyes in some Bolitoglossa (e.g., mex­ ~: of the pars dentalis. icana). Maxillae are shorter iB other genera and f Desmognathus, Leurognathus, and Phaeognathus terminate between the centers and posterior mar­ f have greatly enlarged facial lobes. The genera lack gins of the eyes. In primitive salamanders (Hyno- 14 MEMOIR OF THE SOUTHERN CAUFORNIA ACADEMY OF SCIENCES VOL.4 biidae, Ambystomatidae) the standard condition apparent that the plethodontid genera fall into two is elongate maxillae which extend beyond the major groups; the first contains the genera Desmog­ eyes. The elongate condition in plethodontids may nathus, Leurognathus, and Phaeognathus, the sec­ be a reflection of this ancestral pattern. ond contains the remaining genera. A process arises on the inner margin near the tip of each long pars dentalis of Phaeognathus and Septomaxilla extends posteromedially toward the quadrate, ter­ Small paired septomaxillae are primitively pres­ minating a short distance before the quadrate is ent in plethodontids but are absent in some ad­ reached. This process is large, stout and expanded, vanced groups and in those species which fail to and has grown in the direction of the jugal liga­ complete metamorphosis (Typhlomolge, Haideotri­ ments, which are very short, dorsoventrally ex­ ton, Gyrinophilus palleucus, and paedogenetic panded and stout. Precursors of the process are Eurycea). evident as tiny projections extending out into the Septomaxillae are greatly reduced in Desmog­ jugal ligaments in Desmognathus and Leurogna­ nathus, Leurognathus, and Phaeognathus, and ap­ thus. It is well known that osseous processes often pear as tiny crescentic ossifications with relatively develop at tension points, and the development of great dorsoventral dimensions. Maxillae and nasals the processes in all three genera, and particularly are well separated from the septomaxillae. Other in Phaeognathus, is an indication of the relatively temperate genera have relatively large septo­ great tension in the jugal ligaments and of the maxillae, and the bones are particularly well de­ strength and rigidity of the skulls. The skull par­ veloped in Hemidactylium, Plethodon, Ensatina, tially raises to open the mouth in these three genera, and Batrachoseps. a unique situation in urodeles, and the development In Gyrinophilus, Pseudotriton, Stereochilus, Eu­ of processes is doubtless related to the new tensions rycea, Manculus, Typhlotriton, and Hemidactylium resulting from this drastic functional modification. the septomaxillae are particularly well separated Desmognathus, Leurognathus, and Phaeognathus from the maxillary facial lobes, with the least have greatly enlarged palatal shelves that are separation in Hemidactylium. Septomaxillae may broadly overlapped by the vomers. The shelves are be slightly covered by lateral portions of the nasals considerably larger than in any other genus, and in these genera. are approached in width only by Stereochilus. Maxillary facial lobes overlap the septomaxillae Gyrinophilus and Pseudotriton resemble Stereo­ slightly to considerably in Plethodon, and exten­ chilus, but have relatively narrower shelves. Palatal sively overlap the bones in Aneides. Septomaxillae portions of the maxillae of all other genera are are well separated from the facial lobes in Ensatina, narrower than in the above genera. The shelves but in Ensatina and in A. aeneus specialized lateral are subject to considerable ontogenetic and inter­ nasal processes may overlap the septomaxillae. specific variation, however, and in very large adults Maxillary facial lobes tend to overlap slightly or of several genera (e.g., Aneides, Plethodon) may be just contact the posterior margins of the septo­ relatively broad. Primitive salamanders seem to maxillae of Hydromantes, but extensively overlap have relatively narrow palatal shelves, but broad the septomaxillae of Batrachoseps. shelves are found only in relatively primitive Septomaxillae are of rather irregular occurrence plethodontids. Most specialized and advanced in the neotropical genera, but have been observed plethodontids, such as the neotropical genera, have in at least some individuals of all genera. There is narrow palatal shelves. a well marked evolutionary trend toward reduction Several unique maxillary specializations are en­ and loss of the bones in these genera. "7ben present countered in the genus Aneides. These will be dis­ the septomaxillae are small and well separated cussed more fully in a forthcoming paper (see also from the maxillary facial lobes and the nasals. "7ake, 1960; 1963). Septomaxillae are best developed in some species The American species of Hydromantes (brunus, of Chiropterotriton, and Rabb (1956) stated that platycephalus, shastae) have high, sharp-pointed presence of septomaxillae distinguished Chiroptero­ facial lobes with relatively narrow bases, while the triton from Pseudoeurycea. In a later paper Rabb European species (genei, italicus) have low, round­ (1960) wrote that septornaxillae are present only ed, broad-based lobes. in those species north of the Isthmus of Tehuante­ As maxillae appear during metamorphosis, they pec, but my observations conflict somewhat with are, of course, absent from the paedogenetic his. The bones are present in the following species groups. (N=north of Isthmus; S=south): C. arboreus On the basis of total maxillary structure it is (N), C. bromeliacia (S), C. chiropterus {N), C. 1966 EVOLUTION OF LUNGLESS SALAMANDERS 15 multidentatus (N), and C. priscus (N). Septo­ Nasals in Desmognathus, Leurognathus, and maxillae are absent in: C. abscondens (S), C. Phaeognathus are small and narrow, with a charac­ dimidiatus (N), C. nasalis (S), and C. xolocalcae teristic shield-like shape. The nasals are much (S). smaller than the maxillary facial lobes and slightly Septomaxillae, not previously reported in Pseudo­ overlap the premaxillary frontal processes. Lateral eurycea, are present in some individuals of P. portions of the nasals are in very close contact cephalicus and P. werleri, and Rabb (in. litt.) says with the maxillary facial lobes. the bones occur in western populations of P. Pseudotriton adults have nasals that broadly leprosa. Very small septomaxillae are present in overlap the premaxillary frontal processes and may Parvimolge townsendi, but the bones are absent in meet in the midline. Grohman (1959) stated that P. richardi. Septomaxillae are reduced to tiny splin­ nasals develop in contact during metamorphosis ters of bone in Lineatriton. Bolitoglossa normally and tend to separate during ontogeny; my observa­ lacks septomaxillae, but small remnants are present tions conflict with his and support Martof and Rose on either side of one B. rufescens, on one side of (1962) who said in regard to the nasals "they are another, and on one side each in one individual closer in adults than in earlier stages." Examina­ each of B. mexicana, B. platydactyla, and B. tion of eight cleared and stained P. ruber ranging in rufescens. Small septomaxillae are present in some size from 44.8 to 75.6 mm. indicates that nasals Oedipina inusitata, but are absent in other species. arise far apart and tend to grow slightly mediad Septomaxillae are absent in most Thorius, but are during ontogeny. Only one individual (73.0 mm.) present on one side in one T. pennatulus. Hilton bad the nasals in median contact. Grohman (1959) ( 1946 b) reported some indication of septomaxillae considered the Pseudotriton condition to be prim­ in the same species. itive for the family, but such a situation is not seen Generally septomaxillae are well developed in in more primitive families. Nasals are in contact primitive plethodontids. The bones are either re­ medially in hynobiids, but the bones are distinctly duced or lost in three separate evolutionary lines: different from those of Pseudotriton and are over­ (1) Desmognathus and its allies (always present, lapped by the premaxillary frontal processes. Re­ but reduced); (2) the paedogenetic species (absent cent work by Lebednika (1964) indicates that because species fail to metamorphose); ( 3) the medial portions of the nasals of hynobiids represent neotropical species (either true evolutionary reduc­ vestiges of the postfrontal bones of fishes. No evi­ tion and loss or the result of paedomorphosis) . dence of postfrontal contribution to the nasals of plethodontids was found, and the condition in Nasal Pseudotriton is probably specialized. Skulls of The nasals of plethodontids are subject to con­ Pseudotriton are relatively solid, rigid structures; siderable interspecific and intergeneric variation, secondary medial growth of the nasals is clearly but it is difficult to group genera on the basis of related to a strengthening trend. nasal structure. The primitive nasal condition is A parallel to the condition in Pseudotriton is seen probably close to that seen in Gyrinophilus, Pletho­ in large Ensatina, and to a lesser extent in Gyrino­ don, and other genera in which relatively large philus. The nasals undergo medial growth and over­ nasals only slightly overlap the frontal processes lap the frontal processes. Similar, but less extensive, and usually only slightly, if at all, articulate with medial growth is encountered in other genera as the maxillae. The bones are somewhat. variable in well. shape, but are usually quadrangular or broadly tri­ All anterior elements of Manculus are reduced, angular. Anterior borders of the bones slope pos­ and those of Thorius are very small and variably terolaterally. Posterior margins are irregular, but shaped. The strikingly reduced nasals of Thorius often terminate in poorly to well defined points. are in the form of slender crescents lying far pos­ Similar conditions are encountered in the primitive teriorly at the anterior margins of the orbits. Some Ambystomatidae. Since nasals appear during meta­ of the large-nostriled species of Chiropterotriton morphosis they are absent in the paedogenetic (e.g., nasalis) have tiny, crescentic nasals. species. Martof and Rose ( 1962) report the ap­ pearance of nasals earlier in the ontogeny of ; Pseudotriton than Gyrinophilus. The elements also Prefrontal ·appear relatively early in Eurycea, at a stage ap­ Prefrontals are primitively present in pletho­ proximating that of Pseudotriton. Nasals of Des­ dontids, but since the bones appear during meta­ mognathus and Leurognathus appear late during morphosis they are absent in Typhlomolge, Haideo­ metamorphosis. triton, and the paedogenetic species of Eurycea and 16 MEMOIR OF THE SOUTHERN CA.UFORNIA ACADEMY OF SCIENCES VOL.4

Figure 3. Dorsal and ventral views of skull of adult female Phaeognathus hubrichti. Cartilage stippled. Jugal ligaments lined. Posterior patch of vomerine teeth outlined. Line equals 5 mm.

Figure 4. Dorsal and ventral views of skull of adult male Stereochilus marginatum. Cartilage stippled. Line equals 5 mm. 1966 EVOLUTION OF LVNGLESS SALAMANDERS 17

Figure 5. Dorsal and ventral views of skull of adult male Plethodon ;ordani. Cartilage stippled. Posterior patch of vomerine teeth outlined. Line equals S mm.

Figure 6. Dorsal and ventral views of skull of adult female Bolitoglossa subpalmata. Cartilage stippled. Posterior patch of vomerine teeth outlined. Line eauals 5 mm. 18 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 Gyrinophilus. The bones also have been lost in Route of the Nasolacrimal Duct other genera including Desmognathus, Leurogna­ Eye glands are associated with the lower eyelids thus, and Phaeognathus, in which the prefrontal of metamorphosed salamanders, and the secretions area is invaded by the maxillary facial lobes and of the glands are conducted to the external nares the expanded anterior portions of the frontals. by the nasolacrimal ducts (Noble, 1931). These­ All Hydromantes lack prefrontals; the prefrontal cretion is apparently an eye lubricant, and the region is invaded by relatively large nasals and, in excess drains through the ducts (Noble, 1931). the American species, by enlarged maxillary facial Since eyelids, eye glands, and nasolacrimal ducts lobes. develop during or just before metamorphosis, all Prefrontal loss in Batrachoseps and various neo­ are absent in Ty phlomolge, H aideotriton, and the tropical genera is associated with trends toward paedogenetic species of Gyrinophilus and Eurycea. reduction and simplification of general skull struc­ N asolacrimal ducts of hynobiid and primitive ture as well as paedomorphic trends. The bones ambystomatid (Dicamptodon, Rhyacotriton) sala­ are absent in all species of Batrachoseps except B. manders pass through horizontally oriented fora­ wrighti. Prefrontals are absent in small B. wrighti, mina in the lacrimal bones, elongate, longitudinally but appear during ontogeny and are present as very oriented elements located between the nasals and small dots of ossified tissue in adults over 45 mm. the maxillary facial lobes. Near their anterior ends (large for the species). Enlarged nasals occupy the ducts pass through evacuations or grooves in some of the prefrontal region in Batrachoseps. the septomaxillae, and open in the nasal cavities. Among the neotropical genera prefrontals are Relations of the ducts to bony elements have present in all Pseudoeurycea (moderate-sized) , changed in advanced ambystomatids, in salaman­ Parvimolge (small), and Lineatriton (small). Pre­ drids, and in plethodontids. Larsen (1963) has sug­ frontals are inter- and intra-specifically variable in gested that intercalation of a lacrimal between the Thorius, and are represented by tiny slivers of bone prefrontal and the maxilla reduces stability of the in some, but appear to have fused with the nasals maxilla on the neurocranium and that the lacrimals in others. have been lost in higher groups due to selective Rabb (1960) stated that prefrontals are present pressures acting to fix firmly the maxillae to sur­ in species of Chiropterotriton which occur north of rounding elements. Lacrimal loss has led to rela­ the Isthmus of Tehuantepec, but absent in species tively extensive prefrontal-maxillary articulations south of the Isthmus. Prefrontals are present in all in primitive plethodontids. In · addition the pre­ northern species that I have examined, but are also frontals of genera such as Ambystoma have taken present in the southern C. nasalis, and present or over the function of the lacrimals, and the ducts absent in the southern C. abscondens. The bones pass through the prefrontals. are absent in the southern C. bromeliacia and C. The primitive nasolacrimal duct route of pletho­ xolocalcae. dontids is best seen in Gyrinophilus, in which the Prefrontals are present in primitive Bolitoglossa, ducts extend from the eyes through the tissue of but in a number of advanced species they may be the canthus rQlltralis to the anterior ends of the lost (colonnea, orestes, robusta), or fused with the nasal capsules. Anteriorly the ducts proceed ven­ nasals (adspersa, cerroensis). Hansen and Tanner trad and open in the nasal cavities. The ducts leave no or only slight impressions on the underlying. (1958) report greatly reduced prefrontals in B. bones. Similar routes are followed in Stereochilus occidentalis and B. rufescens, but the bones are and Typhlotriton. absent in my material. Intraspecific variation occurs Ducts of the remaining plethodontid genera have in some species (savagei, subpalmata), and the wholly or in part abandoned the primitive dorsal bones may either fuse with the nasals or remain integumentary routes for routes closer to the cranial discrete. elements. Each prefrontal of Pseudotriton has a Prefrontals are absent in all species of Oedipina, narrow anterolateral projection, similar to that of and the nasals and maxillary facial lobes have Gyrinophilus, which extends between or slightly grown into the vacated area. ventral to the maxillary facial lobe and the nasal. The remaining plethodontid genera have rela­ The prefrontals differ, however, in that those of tively well developed prefrontals. The highest de­ Pseudotriton have tips which are slightly to con­ velopment is achieved in Aneides, in which the siderably evacuated. In addition each prefrontal specialized members of the lugubris group have has a distinct, lengthwise dorsal groove or trough. very large prefrontals which interlock with the Nasolacrimal ducts pass from the eyelids over the maxillary facial lobes (see Wake, 1963). prefrontals and along the depressions or grooves 1966 EVOLUTION OF LUNGLESS SALAMANDERS 19 to the anterior margins of the prefrontals. The of Batrachoseps, but small posterolateral nasal con­ ducts proceed ventrad through foramina, the walls cavities are usually associated with foramina which of which are formed by the evacuated prefrontal are otherwise integumentary. tips on three sides, and by integument on the fourth. Prefrontals are absent in Hydromantes, and the Paths of the ducts in Plethodon, Ensatina, and ducts extend across the maxillary facial lobes, Aneides (except A. lugubris) are similar to those then through foramina formed by concavities in in Pseudotriton. Relatively large prefrontals, evacu­ the posterolateral boundaries of the nasals and ated anteriorly, extend from the orbits nearly to facial lobes. the nares; the ducts extend dorsally and anteriad In Pseudoeurycea, Parvimolge, Lineatriton, and across the bones, then ventrad through the anterior primitive species of Chiropterotriton and Bolito­ evacuations. The ducts pass posterior and then glossa the ducts pass dorsally over the prefrontals ventral to characteristic lateral nasal lobes in and enter foramina formed by anterior prefrontal, Ensatina, and the foramina are formed of nasal and dorsal maxillary, and posterolateral nasal con­ prefrontal material. An extensive maxilla-nasal cavities. Nasals are reduced in size in C. nasalis, articulation effectively separates the prefrontals and the prefrontals bear distinct, fully enclosed from the narial region in Aneides lugubris. The an­ foramina on their anterior extremities. In those terior cranial elements of the species are covered Chiropterotriton and Bolitoglossa which have lost with dense osseous accretions, and the ducts enter prefrontals, and in all Oedipina, the nasals and foramina running lengthwise in the accretions and maxillae usually articulate extensively, and the proceed to the anterior ends of the prefrontals. ducts pass dorsad initially, then move ventrad There the ducts enter grooves in the thickened through concavities in the posterolateral nasal mar­ lateral margins of the nasals, and enter the nasal gins. Slightly to moderately concave dorsal margins cavities at the anterior ends of the bones. of the maxillary facial lobes also contribute to the The prefrontals in Eurycea and Manculus are foramina. moderate to small and lie farther back than in the An extreme of the trend described above is genera discussed above. The bones are primarily reached in Thorius, in which the prefrontals are associated with the anterior orbital margins, and, greatly reduced and play no role in the nasolacrimal although they extend between the maxillae and dqct routes. The nasals have been forced posteriorly nasals, are walled off from the osseous nasal aper­ by the shortening of the snout and enlargement of tures by lateral processes of the nasals. In both the nostrils, and now form the anterior margins genera the prefrontals have pronounced anterior of the orbits. Nasal-maxillary articulation is slight. concavities, and, in some Eurycea, distinct laterally The ducts initially extend anteriad from the eyelids placed foramina. The concavities and/ or foramina in a dorsal position, but very soon enter partially may be associated with small lateral concavities of closed foramina in the posterior orbital margins of the nasals, mid-dorsal concavities of the maxillary the nasals and proceed ventrad to the nasal facial lobes, or both (see also Wilder, 1925). The capsules. nasolacrimal ducts extend across the prefrontals, Presumably the most specialized plethodontid and, in some instances, dorsal to the posterior por­ condition is that encountered in Desmognathus and tions of the facial lobes, before entering the for­ allied genera. Wilder ( 1913) was unable to find amina. The point at which the ducts run ventrad nllSolacrimal ducts in Desmognathus, and the ducts is relatively posterior to that found in Pseudotriton. appear to be absent in Phaeognathus and Leurog­ Processes of the nasals articulate with the max­ nathus as well. illary facial lobes in Hemidactylium, but the pre­ On the basis of the arrangement of the antero­ frontals are similar to those of Eurycea and bear lateral cranial elements, and the route of the naso­ pronounced anterolateral concavities. The ducts lacrimal ducts, the following generic groupings are pass over the prefrontals, through the prefrontal evident: concavities, then under lateral nasal processes. Prefrontals play no role in the route of the 1. Gyrinophilus, Stereochilus, Typhlotriton. nasolacrimal ducts in the remaining genera of 2. Pseudotriton. middle lattitudes. Prefrontals of Batrachoseps 3. Plethodon, Ensatina, Aneides. wrighti are extremely reduced, and do not extend 4. Eurycea, Manculus, Hemi(iactylium. · between the nasals and maxillae. The ducts pass 5. Hydromantes, Batrachoseps, Bolitoglossa, Oedi­ across the maxillary facial lobes and through pina, Pseudoeurycea, Chiropterotriton, Par­ · foramina in the interspace between the maxillae vimolge, Lineatriton, Thorius. and nasals. Duct routes are similar in other species 6. Desmognathus, Leurognathus, Phaeognathus. 20 MEMOIR 0# THE SOUTHERN CAUFORNIA ACADEMY OF SCIENCES VOL.4 The above groupings reflect the general trends ticulation between bilateral counterparts limited from primitive dorsal to advanced ventral naso­ to extreme anterior tips; elements very widely lacrimal duct routes. This is a morphological trend separated posteriorly; small anterior processes that may reflect phylogeny. It is likely that the ten­ attach vomers to premaxilla. Gyrinophilus, dency for posterior movement of the nasolacrimal Pseudotriton, Stereochilus, Eurycea, Manculus, duct .foramina has proceeded in parallel along Typhlotriton, Haideotriton, Typhlomolge, Hemi­ several phylogenetic lines, with parallelism most dactylium. evident between groups 3, 4, and 5. 2. Vomers in shape of upper quarter circles, with Vomer small posteromedial projections; virtually com­ Vomers of two distinct types are encountered in plete articulation of bilateral counterparts along larval and paedogenetic plethodontids (Fig. 7): midline; no anteriorly directed processes, and 1. Vomers in shape of inverted L's with rounded no attachment of vomers to premaxill~. corners and varying amounts of expansion; ar- Desmognathus, Leurognathus.

A B

alatopterygoid

Figure 7. Vomers and palatopterygoids in larval plethodontids. A. Desmognathine pattern (Desmognathus quadramaculatus), B. Plethodontine pattern (Gyrinophilus danielsi). Not drawn to scale.

Larval ambystomatid vomers resemble condition vomerine growth posteriolateral to the tooth series 1, and it apparently is the more generalized and (Fig. 8). Fontanelles are open and of moderate primitive of the two plethodontid conditions. The size in all genera, and vaulting is moderately to well larval vomers of Desmognathus and Leurognathus developed in all but Typhlotriton. Preorbital pro­ articulate extensively with each other. The most cesses are present primitively. The vomerine tooth plausible explanation for this articulation is that it is series originate on the processes and proceed an­ a strengthening compensation for the loss of suspen­ teriomediad almost to the midline where the series sorial articulation of the palatopterygoids (which sharply turn and proceed initially posteromediad, see), with which the larval vomers articulate finally posterolaterad, to terminate below the pos­ posteriorly. terior portions of the parasphenoid. Such a pattern Plethodontid genera may be placed in three is characteristic of Gyrinophilus, Pseudotriton, groups on the basis of adult vomerine structure. Stereochilus, and Typhlotriton. All four have well Gyrinophilus, Pseudotriton, Stereochilus, Eurycea, developed and stocky preorbital processes that Manculus, Typhlotriton, and Hemidactylium form extend beyond the lateral margins of the internal one group that is characterized by: ( 1) presence nares, but fall far short of the vomerine body of sharply arched vomerine tooth series, (2) bony margins. Anterior and posterior portions of the 1966 EVOLUTION OF LUNGLESS SALAMANDERS ~~ 21 tooth series are continuous in the four genera. The Hemidactylium has slender, toothless preorbital bilateral series are joined for a short distance at processes that extend to the lateral margins of the about the level of the anterior margins of the internal nares. The anterior tooth series are arched orbitosphenoids in Stereochilus. Lateral margins of anteromedially, but to a lesser extent than in other the vomerine bodies are projected a little posteriad, members of the group. In very small individuals below the preorbital processes, in Stereochilus and ( 16 mm.) the anterior and posterior tooth series are drawn into spinous, posterolaterally projecting are continuous, but in adults there is a toothless processes that are diagnostic of the genus; pre­ gap. orbital processes are also directed rather strongly The posterior portions of the tooth series, or the posterolaterad, and thus are not overlapped by the posterior tooth patches, are widely separated past body processes. the parasphenoid midpoints in all members of the group. The posterior teeth are in relatively narrow series or patches which contain from two to six diagonal rows at the point of maximum width. A B Plethodon, Ensatina, Aneides, Hydromantes, Batrachoseps, Bolitoglossa, Oedipina, Pseudoeury­ cea, Chiropterotriton, Parvimolge, Lineatriton, and Thorius form a second group. The farthest anterior extent of the vomerine tooth series is at the extreme lateral extremity of each series, or slightly medial to that point, rather than near the midline. The an­ terior tooth series extend along the lateral and pos­ terior margins of the vomers, and there is no Figure 8. Vomers of adult plethodontines. A. Pletho­ posterolateral vomerine growth beyond the series donine and bolitoglossine pattern (Plethodon jordani), (Fig. 8). Anterior and posterior toothed areas are B. Hemidactyliine pattern (Pseudotriton ruber). Arrow normally widely separated, particularly in adults. indicates bony growth behind tooth series. Not drawn to scale. Fontanelles are normally moderate to very large, and vaulting is moderately to well developed. Rela­ In contrast to Stereochilus with its compressed tively slender preorbital processes that extend to skull, Typhlotriton has a broad, flat skull. In old at least the lateral margins of the internal nares are adults the vomerine series of teeth are very sharply present in all genera of this group except Batracho­ arched and the posterior portion of each series seps, Thorius, some species of Aneides (ferreus, curves back on itself as it leaves the vomer proper. ffavipunctatus, lugubris), and some Chiropterotri­ Thus the series are very widely separated in the ton (bromeliacia, dimidiatus, nasalis). Small pro­ interorbital region, and in the largest individuals cesses are evident in some Batrachoseps, but are come to lie, for a short distance, along the orbito­ virtually absent in most. The longest preorbital sphenoids rather than the parasphenoid. The series processes occur in Ensatina, where they extend are usually sinuous, however, and curve back below beyond the lateral margins of the vomerine bodies. the parasphenoid, where they extend to the pos­ The posterior tooth patches may be continuous terior end of the parasphenoid. At the latter point with the anterior tooth series in very young indi­ the series proceed medially and approach each viduals (e.g., Ensatina at 23.l mm.), but toothless other on the midline. The pattern in Typhlotriton gaps are normally present in adults. The posterior . is unique both in the degree of separation of the patches vary from very well separated in Hydro­ aeries and their posterior extent. The body of each mantes and Batrachoseps, to moderately or slightly vomer is also unusual in the genus and bears a separated in most genera, to closely appressed but posteriorly directed process that forms the lateral unfused in Thorius and some species of Bolito­ margins of the nares. glossa. In genera such as Pseudoeurycea and Eurycea and Manculus have short preorbital Chiropterotriton the tooth patches may be in con­ processes that fall short of the lateral edges of the tact anteriorly but well separated posteriorly. In · internal nares. Metamorphosing individuals have general, patches in genera of this group contain >tooth patterns similar to those described above, more teeth than those of the former group, and but adults normally have the posterior patches vary from three to four rows at maximum in some eeparated by toothless gaps from the arched an­ species of Pseudoeurycea, to as many as ten to terior series (usually continuous in E. aquatica; twelve rows in some species of Bolitoglossa and Rose and Bush, 1963). Plethodon. 22 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 Parasphenoids in the first two groups of genera that the highly aquatic Leurognathus has lost the extend dorsally above the vomers proper to points internasal glands, which aid in food capture in equal to, or somewhat in advance of, the anterior terrestrial habitats. Fontanelles in Desmognathus, orbitosphenoid margins. The area of overlap is and to a lesser extent in Phaeognathus, may be relatively great, and at no point do the parasphen­ closed over secondarily by medial growth of the oids and vomers articulate squarely. This is also fontanelle borders. These genera differ further from the condition encountered in primitive salamander Leurognathus in that dorsal growth of the vomerine families. In the third group (Desmognathus, bodies forms bony walls around the posterior and Leurognathus, Phaeognathus) parasphenoids are posterolateral portions of the fontanelles, while the greatly shortened in old adults, and articulate processes are absent in Leurognathus because skull squarely with vomerine posterior margins at points depression places the premaxillae and vomers near the orbitosphenoid midpoints. This articula­ essentially in contact. tion is considerably more rigid than in other groups, The internal nares have shifted laterally in and is a specialized condition associated with the Leurognathus, and the preorbital processes are very trends toward skull strengthening in the third closely applied to the vomer bodies, with only group. The specialization involves not only para­ narrow lateral narial slits. Nares are in the normal sphenoid shortening, but also relative vomerine position in Desmognathus and Phaeognathus. Pre­ lengthening. Thin portions of the parasphenoids orbital processes of Phaeognathus and, especially, in Desmognathus may extend fairly far anteriorly, of Leurognathus are long and extend laterally dorsal to the vomers. beyond the vomerine bodies. The processes are Vomerine tooth pattern is highly distinctive in shorter in Desmognathus, and lateral margins of the third group. In young Desmognathus and many the bodies grow posterolaterad to extend below adults the anterior teeth are in short, arched series the tips of the processes. Preorbital processes tend that do not extend onto the preorbital processes. to articulate with the antorbital portions of the Anterior vomerine teeth are lost in the larger in­ frontals in all three genera, a situation not en­ dividuals, especially males, of some species (e.g., countered in the other generic groups. D. monticola, D. quadramaculatus). Anterior teeth Posterior vomerine patches are well separated are in relatively long, straight series in Phaeog­ from the anterior series and slightly separated nathus, with slight anterolateral curves; the series from each other in adults of the third generic form relatively very small angles with the skull group. The patches are well developed, and have axes, and are located considerably more posteriorly five or six diagonal rows of teeth. The anterior than in other plethodontid genera. The slender pre­ series and the posterior patches are continuous in orbital processes of Phaeognathus are toothless. late larvae and metamorphosing individuals in D. Teeth are usually absent in Leurognathus, but may quadramaculatus. At these stages the anterior be present in some populations (Martof, 1962). series are highly arched, and the apices are antero­ In toothed individuals of Desmognathus and medial in position. From each apex the series slope Leurognathus, and in Phaeognathus, considerable laterally and terminate before the medial margins amounts of bony posterolateral growth occur of the internal nares are reached; in the other direc­ beyond the tooth series in a manner similar to that tion the series extend posteromediad, then postero­ encountered in genera of the first group. laterad, and finally terminate on the posterior, Vomerine bodies are relatively small and very ventral surfaces of the parasphenoids. The general flattened in Phaeognathus and Leurognathus, but pattern closely resembles that seen in adult the somewhat larger elements of Desmognathus are Stereochilus. slightly vaulted. The vomerine bodies are broadly overlapped dorsally by the palatal shelves of the Frontal premaxillae and maxillae in Desmognathus and Frontals primitively are large and occupy most Leurognathus, but the vomers overlap the pre­ of the interorbital area. Ambystomatid and hyno­ maxillary palatal shelves in Phaeognathus. Fon­ biid frontals are long, and extend beyond the tanelle closure was reported in Leurognathus by posterior margins of the orbitosphenoids, and such Moore (1899), and was employed as a generic a condition is probably primitive in the pletho­ character by Dunn (1926) and Noble ( 19 31 ) . dontids. The evolutionary trend in plethodontids Distinct fontanelles are present in young individuals appears to be toward a decrease in frontal size, and in small adults, however, and are progressively associated with an increase in the relative size of closed over with increasing size. Dunn ( 1926) has the parietals. Plethodontid genera may be grouped stated that fontanelle closure is related to the fact as follows on the basis of this character: 1966 EVOLUTION OF LUNGLESS SALAMANDERS 23 1. Frontals long, extend beyond posterior margins widely separated posteriorly. The elements are es­ of orbitosphenoids. Desmognathus, Leurogna­ sentially in contact between the anterior margins thus, Phaeognathus, Gyrinophilus, Pseudotriton, of the orbits, but rapidly diverge posteriorly, and Stereochilus. leave much of the dorsal portion of the brain un­ 2. Frontals short, fall short of posterior margins covered by bone. The frontals of Thorius are also of orbitosphenoids. Eurycea, Manculus, Typh­ reduced, and leave middorsal gaps that are con­ lotriton, Typhlomolge, Haideotriton, Hemidac­ siderably smaller than those of Batrachoseps. tylium, Plethodon, Ensatina, Aneides, Hydro­ The most distinct group of plethodontid genera mantes, Batrachoseps, Bolitoglossa, Oedipina, based on frontal structure contains Desmognathus, Pseudoeurycea, Chiropterotriton, Parvimolge, Leurognathus, and Phaeognathus, in which the Lineatriton, Thorius. bones are dense, stout, relatively thick, and en­ larged. Anterior expansion is great. Ventrolateral The genera may also be grouped on the basis of extensions of the expanded regions have invaded presence or absence of distinct lateral processes on the antorbital regions between the maxillary facial ·the posterior margins of the frontals: lobes, and the maxillary palatal shelves and vomer­ ine preorbital processes. This feature is one on 1. Frontals with medially sloping posterolateral which Cope's (1866) family Desmognathidae was margins; no distinct lateral processes. Des­ based. The antorbital processes are specializations mognathus, Leurognathus, Phaeognathus, Gy­ associated with skull solidification and strengthen­ rinophilus, Pseudotriton, Stereochilus, Typhlo­ ing, and are found in no other plethodontid genera. molge, Haideotriton. The expanded anterior portions of the frontals have 2. Frontals with relatively straight or only slightly grown into the area vacated by the prefrontals, and medially sloping posterolateral margins; poorly are extensively overlapped by the maxillary lobes to very well developed lateral processes which in a firm, extensive articulation. Posteriorly, along overlap parietals on posterior margins. Eurycea, the orbital margins, distinct lateral ridges are Manculus, Typhlotriton, Hemidactylium, Pleth­ present. odon, Ensatina, Aneides, Hydromantes, Bolito­ glossa, Oedipina, Pseudoeurycea, Chiroptero­ Parietal triton, Parvimolge, Lineatriton, Thorius. Generic groupings are recognizable on the basis of parietal structure. The first group includes Many exceptions are found among species of Plethodon, Ensatina, Aneides, and Hydromantes. · the genera of the last group, but every genus has at In these genera the parietals are relatively flattened i least some species that fit the category. The best with broad, concave, posterior depressions (some : developed processes are found in Eurycea, Man­ IargeAneides), slight depressions (Aneides, Pletho­ '"·culus, Hemidactylium, and the neotropical genera, don, Ensatina), or essentially no depressions (Hy­ htqd the processes are apparently all that remains dromantes ). In all four genera the orbital edges of . of the greatly reduced frontals of Batrachoseps. In the parietals are laterally rather than ventrally di­ LTyphlotriton the posterolateral borders slope medi- rected, and the bones have no lateral aspects. The 1 ally, and the small posterior processes are located orbitosphenoids are moderately to slightly over­ · more medially than in other genera. There is no lapped by the depressed posterior portions except ··indication of lateral processes in ambystomatids in Hydromantes. The general condition in these '. or hynobiids, and the posterolateral borders slope genera is similar to that found in primitive sala­ :. medially. Lack of lateral processes appears to be manders, and is perhaps close to the ancestral ~primitive in plethodontids. The lateral processes condition of the family. i.may compensate for general skull weakening, since The second group includes Gyrinophilus, Pseu­ they tend to be better developed in forms with dotriton, Stereochilus, Eurycea, Manculus, Typhlo­ weakly developed skulls (e.g., Manculus). triton, Typhlomolge, Haideotriton, and Hemidac­ Frontals of Batrachoseps and Oedipina have tylium. Parietals of this group have posterior well developed anteromedial spinous projections depressions that vary from marked to shallow, and that extend beside and below the premaxillary the bones have distinct lateral aspects with ventrally . frontal processes. Anterior expansion of the directed edges that significantly overlap the orbit­ frontals is marked in Hydromantes, and is also evi­ osphenoids in metamorphosed individuals. Pos­ dent in various species of many other genera. terior depression is marked in H emidactylium, Notable frontal reduction occurs in two genera. Stereochilus, Eurycea, and Manculus, but is poorly In Batrachoseps frontals are narrow and are very developed in Gyrinophilus, Pseudotriton, and 24 MEMOIR OF THE SOUTHERN CAUFORNIA ACADEMY OF SCIENCES VOL.4 Ty phlotriton. The relatively smooth parietals of the other groups. Parietals are divisible into two Typhlomolge and Haideotriton resemble those of sections. The anterior sections are high and very larvae of the other genera, but are somewhat better flat, and are on the same level as the frontals. developed. Lateral processes and distinct, obliquely Neither frontals nor anterior portions of parietals oriented crests are present posteriorly in Typh­ are covered by musculature. The anterior portions lomolge. The posterolateral margins of the parietals articulate very firmly with the frontals. Pronounced are raised to form portions of the parietal-otic lateral ridges are present, and each bone proceeds crests in Pseudotriton, Gyrinophilus, Stereochilus, ventrad and slightly mediad from the ridge; thus and some species of Eurycea (aquatica, bislineata). there is a distinct lateral aspect to each bone. The A third group includes Batrachoseps, Bolito­ posterior portions of the parietals have been im­ glossa, Oedipina, Pseudoeurycea, Chiropterotriton, pressively molded by the strong ligaments that ex­ Parvimolge, Lineatriton, and Thorius. The parietals tend from the atlas vertebrae to the mandibles (Fig. of these genera are either smooth or have slight 9), and are in the form of deep, concave, almost posterior depressions. Well developed, lateral parie­ U-shaped troughs that are directed anterolaterad, tal spurs are present in all genera, and are diagnostic into the orbits. The parietals and anterior portions of the group (this structure is well illustrated by of the occipito-otics are drawn into anterolateral Wiedersheim, 1877, Fig. 9b, in Batrachoseps). orbital extensions of the troughs. The distinction The spurs are small lateral processes that are di­ between anterior and posterior portions of the rected sharply ventrally over the area of the ascend­ parietals is strong in Desmognathus and Leurog­ ing processes of the suspensoria. It is possible that nathus, but is less well marked in Phaeognathus, the spurs may have arisen as compensations for the in which the anterior portions are relatively lesser generally weakened condition of the dorsal roofing proportions of the bones and the parietals form bones in members of this group. The spurs appear significantly greater amounts of the troughs. to be neomorphic developments in these genera, In Pseudotriton and some Aneides (lugubris and are probably not morphologically related to the group) the parietals tend to fuse medially in large lateral portions of the parietals of Eurycea and individuals, and a median crest forms. Parietals Manculus which may, in very small species, re­ are separated by medial spaces in larvae and paedo­ semble the spurs. genetic adults. Separation in adults of non­ Genera of the final group (Desmognathus, paedogenetic species (Hydromantes, Batrachoseps, Leurognathus, Phaeognathus) have highly special­ Thorius, Eurycea multiplicata, Bolitoglossa platy­ ized parietals that differ markedly from those of dactyla) may be due to paedomorphosis. A huge

Figure 9. Skull, cervical vertebra and first trunk vertebra of adult female Desmognathus ochro­ phaeus (46.7 mm.). Note atlas-mandibular ligament, shape of cervical vertebra, and pterygapo­ physes extending posterolaterally above postzygapophyses of trunk vertebra; all are characteristic features of the subfamily Desmognathinae. \1966 EVOLUTION OF LUNGLESS SALAMANDERS 25 ·space separates the parietals of Batrachoseps ticularly broad in Typhlotriton, and, while less ~smaller in wrighti than in other species), and broad, are noticeably expanded in Gyrinophilus, "moderate spaces are present in Thorius. The state­ Pseudotriton, Stereochilus, Eurycea, Manculus, and ment by Cope (1889) that the parietal bones of Hemidactylium. Broad anterior parasphenoid tips '()edipina are "two small oval lateral scales" and are typical of plethodontid larvae, and are found in are unossified is incorrect. The bones are well ossi­ adult Typhlomolge and Haideotriton. A tendency 'fted and completely roof the posterior portions of toward tip narrowing is evident in the remainder [:the brain in Oedipina. of the genera, but the tips may be slightly expanded ~:::.: in some species of Pseudoeurycea and Chiroptero­ Orbitosphenoid triton, and relatively broad in large species of other genera (e.g., Bolitoglossa robusta, Plethodon yon­ ..•. Orbitosphenoids are absent in Typhlomolge and ahlossee ). An extreme of reduction is achieved by i{itaideotriton, and in small larvae of other genera. Chiropterotriton abscondens, in which the anterior fTbey are present in large larvae, and in the paedo- . tips are narrowed to sharp, shortened points in •~f8netic species of Eurycea. Well developed orbito- front of which the orbitosphenoids articulate. iaphenoids are present in adults of other genera, Broadened posterolateral parasphenoid pro­ , except in Thorius, in which the elements are small cesses are found in primitive salamanders (hy­ . and do not articulate dorsally with the frontals and nobiids, ambystomatids) in which they do not •·parietals. ordinarily articulate with the quadrates. The con­ The optic fenestrae are always enclosed in bone dition encountered in Stereochilus appears to be 'and are located from near the midpoints of the the most primitive in the family, and two major bones (in some Aneides), to very near the pos­ types of modifications have occurred. Hypertrophy terior margins. The fenestrae in Typhlotriton have of the processes has taken place in the first group been reduced to minute foramina which reflect the of genera, and reduction and eventual loss of the reduced nature of the visual system, including the processes occurs in most members of the second drastic reduction of the optic nerves, in the genus. group. The truncated anterior parasphenoid tips encountered in the first group are unlike those Parasphenoid found in other salamander families, and the mod­ Plethodontids may be placed in two groups on erately long, overlapping tips of the second group of the basis of parasphenoid structure. The first group genera are apparently close to the primitive condi­ contains Desmognathus, Leurognathus, and Phae­ tion. Furthermore, the broad or expanded para­ ognathus. The parasphenoids are shortened and sphenoid tips of some genera (e.g., Typhlotriton, fall short of the anterior margins of the orbito- Gyrinophilus) resemble those characteristic of · lj>henoids. In addition· the anterior portions are hynobiids, ambystomatids, and salamandrids, and tn:incate and strongly concave to U-shaped. The probably represent the primitive plethodontid bones articulate along their anterior margins di­ condition. rectly with the vomers in moderately firm sutures. 1'lie posterior portions of the parasphenoids are Occipito-Otic . very broad with large posterolateral processes that Otic capsules of plethodontids are considerably articulate directly with the enlarged quadrates. variable in size, but genera cannot be grouped on · · The parasphenoids are not notably shortened in that basis alone. In some the otic capsules are the second group (all other plethodontid genera), noticeably large (Ensatina); in others they are · ind extend to the anterior margins of the orbito­ greatly reduced (Oedipina). Most genera have cap­ sphenoids or beyond. The anterior portions are sules of moderate size. moderately concave (but never U-shaped) to al­ Distinctive otic crests allow some groupings to most fiat. Anteriorly the bones overlap the vomers, be made. Desmognathus, Leurognathus, and Phae­ and posteriorly parasphenoids fail to articulate ognathus have posteromedial-anterolaterally orient­ with the quadrates. Broad, posterolateral wings, ed crests that extend along the posterior margins which fall short of the quadrates, are present in of the otic-parietal ·depressions. The crests are Stereochilus, and somewhat smaller processes are prominent and relatively high, and are formed ex­ present in Pseudotriton. lil Gyrinophilus rounded clusively of otic capsular material. No other genera processes are present, but they are less well de­ have crests even remotely similar. veloped than in Pseudotriton. The processes are Double ctests, usually separated, are found in faintly indicated, or absent, in all other genera. Gyrinophilus, Pseudotriton, Stereochilus, and some Anterior portions of the parasphenoids are par- species of Eurycea (aquatica, bislineata). The an- 26 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 terior crests are formed of parietal and otic projec­ condyles, attached directly to the lateral margins tions, and more posterior, lateral crests are formed of the foramen magnum, are characteristic of of otic and squamosal projections. The crests are primitive salamanders (hynobiids, ambystomatids), about equal in over-all height and extent, and are occur in most plethodontid genera, and are doubt­ relatively isolated with rather limited bases. In all less primitive in plethodontids. Occipital condyles four genera the crests tend to be spine-like, and are of Desmognathus, Leurognathus, and Phaeogna­ directed laterally or slightly posterolaterally. Otic thus are borne on the distal tips of moderately long, and parietal contribution is about equal in the an­ cylindrical pedicels or stalks. Stalked occipital terior crests of Gyrinophilus and Stereochilus, but condyles were first noted by Cope ( 1866) and were . parietal participation is sharply decreased in Pseu­ one of the characters on which his family Desmog­ dotriton and Eurycea. Otic and squamosal partici­ nathidae was based. Mandibles of these three gen­ pation is about equal in the lateral crests of era are held largely immobile, and skulls move on Gyrinophilus and Stereochilus, while squamosal condylar-atlas joints. The stalked condyles increase participation is sharply increased in Pseudotriton the distance between the skull proper and the and Eurycea. The crests reach a maximum develop­ vertebral column, and provide space for vertical ment in Pseudotriton and Gyrinophilus. In Gyri­ skull movement. nophilus one or two small projections may be present in the region between the two crests, and the crests may be joined to form high, curved Suspensorium lamina in large Pseudotriton (circa 90 mm.). Plethodontids differ from other, more primitive Distinctive high, wing-like otic crests are present salamanders in lacking bony pterygoids as adults. in most species of Aneides (Wake, 1963). The Pterygoids have been replaced by cartilaginous crests are extensive and are oriented in a postero­ pterygoid processes of the suspensoria, which ex­ lateral-anteromedial direction. Posteriorly the crests tend anterolaterally toward the maxillary tips. The are continuous with long, low, posteriorly directed processes are shortest in those forms in which the squamosal crests. maxillae are longest, and are particularly short in Moderate crests similar to those in Aneides are Desmognathus, Leurognathus, and Phaeognathus. present in Pseudoeurycea smithi. The crests have Cartilaginous portions of the suspensoria are parietal rather than squamosal participation, and relatively large in most plethodontid genera, and the small, accessory, anterior crests proceed over the ossified portions ( quadrates) are limited to distal anterior semicircular canals perpendicular to the regions and do not articulate with the occipito-otics main crests. Slight otic crests are present in some or the parasphenoids. The quadrates communicate other species of Pseudoeurycea (bellii, goebeli). with the skull proper by means of large, cartilagin­ Small anterolaterally oriented crests, or antero­ ous, basal suspensorial portions. This situation is laterally directed small projections, are found in similar to that commonly encountered in primitive some individuals of Plethodon, Ensatina, Batracho­ salamanders, and is evidently primitive in pletho­ seps, Chiropterotriton, and Bolitoglossa. Unique, dontids. In large adults of some genera (e.g., Pseu­ laterally oriented and dorsoposteriorly directed dotriton, Aneides) there is a tendency for thin, crests are present in Plethodon elongatus (Wake, bony, dorsomedial quadrate processes to develop, 1963). but the processes do not reach to the otic capsules. All crests are apparently specializations related Ossified portions of the suspensoria are extensive to provision for additional space for the origin and in Desmognathus, Leurognathus, and Phaeogna­ insertion of enlarged amounts of mandibular ad­ thus, but the cartilaginous portions are small. The ductor musculature, and are most prominent in quadrates extend below the squamosals to articu­ forms in which the jaw musculature is greatly late broadly and firmly with the otic capsules and enlarged. No otic crests have been observed in with posterolateral parasphenoid processes. Crania Manculus, Typhlomolge, Typhlotriton, Haideo­ raise in opening the mouths of these three genera, triton, H emidactylium, Hydromantes, Oedipina, and forces between the crania proper and the Parvimolge, Lineatriton, or Thorius, and crests are quadrates are much greater than in more general­ absent in most species of Batrachoseps, Eurycea, ized forms in which the quadrates serve simply as Chiropterotriton, and Bolitoglossa. points of suspension of the mandibles. In most plethodontid genera the quadrates serve as the Occipital Condyles fulcra for mandibular motion; in Desmognathus Pletbodontid genera may be placed in two groups and its allies the articulars serve as fulcra for on the basis of occipital condyle structure. Sessile cranial movement. 1966 EVOLUTION OF LUNGLESS SALAMANDERS 27 Palatopterygoid Opercular Plate Palatopterygoids are present in larval and paedo­ The opercular apparatus of salamanders has re­ genetic plethodontids, but disappear during meta­ cently been reviewed by Monath (1965). On the ·morphosis. Larval plethodontids may be placed in basis of an incomplete study of plethodontids he 't\vo groups on the basis of palatopterygoid struc­ stated that primitive, aquatic, semi-terrestrial, and ture. Palatopterygoids are large and elongate in terrestrial plethodontid genera retain a well-de­ ,Gyrinophilus, Pseudotriton, Stereochilus, Eurycea, veloped columella, but that columellar reduction >Manculus, Typhlotriton, Typhlomolge, Haideotri- has occurred independently in two lines (essentially ton, and Hemidactylium. The bones are expanded the attached and free-tongued groups of von anteriorly, but are usually drawn into relatively Wahlert, 1957). The major trend has been clear narrow, elongate, posterior processes which articu- since Reed's (1920) study. Although I disagree • late directly, or almost articulate, with the quad­ with the phylogenetic speculations of Monath con­ rates. The condition in these genera is similar to cerning plethodontids, the overall trend is clearly that encountered in primitive salamanders, and as Reed, Dunn (1941), and others in addition to probably represents the primitive plethodontid Monath have suggested. · condition. Desmognathus, Leurognathus, and Phaeognathus Palatopterygoids are small and teardrop-shaped, have large opercular plates which bear distinctive and have posterior points in Desmognathus and columellae on their anterolateral borders. The · Leurognathus. The bones are separated from the columellae are very short and stocky, and arise as • quadrates by distances that approximate the total broad-based, somewhat flattened processes. lengths of the palatopterygoids. This condition ap­ Columellae of the remaining plethodontid gen­ pears to be advanced within the family. era, when present, are tubular rods of small diameter. Salamander columellae are primitively Squamosal rod-like (Reed, 1920), and the situation in these The squamosals are most highly developed and genera is probably primitive in the family. Columel­ '1rongest in those species of various genera in which lae are present and well developed in Gyrinophilus, head and jaw musculature is well developed (e.g., Pseudotriton, Stereochilus, Eurycea, Manculus, · Desmognathus and allies, Aneides lugubris, Gyri­ Typhlotriton, Typhlomolge, Haideotriton, Hemi­ nophilus, Pseudotriton, Pseudoeurycea smithi). dactylium, Plethodon, Ensatina, Aneides, and Genera with reduced skeletal elements and head Hydromantes. Columellae of Batrachoseps, re­ muscles (e.g., Manculus, Batrachoseps, Chiroptero­ ported to be absent by Reed (1920) , Dunn ( 1926), triton, Thorius, Oedipina) have thin, narrow, splint­ and Piatt (1935), and to be present by Hilton like squamosals. Squamosal processes contribute (1945 a) and Monath (1965), are present in all to the otic-squamosal crests in Gyrinophilus, Pseu­ individuals examined. Columellae are well de­ dotriton, Stereochilus, and Eurycea, and the squa­ veloped in B. wrighti and are only slightly reduced mosals may have relatively large, dorsal raised re­ in the other species. gions in Desmognathus, Leurognathus, and Phaeog­ Neotropical genera are characterized by a trend nathus. Posterior squamosal crests are present in toward columellar reduction and loss. Primitive large Aneides. In general variation is closely related Chiropterotriton have the best developed columel­ lae, but in advanced species (e.g., C. bromeliacia, to function, and is more significant evolutionarily C. xolocalcae, C. nasalis, C. abscondens) columellae on the interspecific than on the intergeneric levels. are lost. Very small columellae are present in The squamosals bear small, well-developed, Lineatriton. Pseudoeurycea has columellae that are slender, cylindrical processes in all Thorius (Tan­ usually vestigial and small. Vestigial, very short ner, 1952). The processes arise above the mid­ columellae are present in Bolitoglossa dunni and in points of the anteroventrally sloping, posterior one individual of Parvimolge townsendi, but are squamosal edges and proceed almost directly pos­ absent in all other species of both genera, and in teriorly in horizontal planes. The quadratopector­ all Thorius and Oedipina. When columellae are ab­ alis musculature, which in other genera attaches sent the columellar processes of the suspensoria to the quadrates, has shifted to the squamosals in extend to the opercula. Thorius, and the processes have apparently de­ The opercula are essentially absent in some adult veloped in response to pressures resulting from the Chiropterotriton multidentatus, except for small shift. These processes have been encountered else­ amounts of tissue at the bases of the relatively well where only in a single large individual of Oedipina developed columellae. complex. Reed (1920) stated that Typhlomolge should 28 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 not be included in the family Plethodontidae be­ tively small prearticulars of Hydromantes, Batrach­ cause the columellae contribute considerably more oseps, Bolitoglossa, Oedipina, Pseudoeurycea, to the plate portions than in other members of the Chiropterotriton, Parvimolge, Lineatriton, and family. Similar, but less extreme conditions are en­ Thorius are dorsomedially inflected. countered in H aideotriton and Eurycea pterophila, Desmognathus, Leurognathus, and Phaeognathus however, and appear to be related to the paedo­ have well developed, anterior, dorsal prearticular morphic nature of these groups. The situation cer­ projections located just anterior to the prearticular tainly does not warrant exclusion of the involved shelves. The atlas-mandibular ligaments insert on species from the Plethodontidae. the projections. Similar but considerably smaller projections are present in Stereochilus, some large Mandible Eurycea bislineata, and adult Typhlotriton. Slight The primitive salamander mandibular pattern is indications of the processes are occasionally en­ retained by the majority of plethodontid genera. countered in other genera. Dentaries and prearticulars are of moderate to Dentary bulk is relatively greatest in Desmogna­ small size. Each mandible is curved primitively, thus, Leurognathus, and Phaeognathus, and this so that the mandibular rami form broadly rounded bulk is approached, but not matched, by large, arches. Meckelian grooves are open for much of specialized species such as Aneides lugubris. The the lengths of the dentaries, and at the anterior dentaries are relatively slender in most genera. ends of the prearticulars. Raised, pointed, dentary processes are located Specialized mandibles are found in Desmogna­ in the coronoid region in Typhlotriton, and about thus, Leurognathus, and Phaeognathus. In these match the prearticnlars in height. Laterally in­ genera the dentaries are large, massive structures flected, flange-like processes are present in large that are particularly broad ventrally. The dentaries Plethodon and Aneides. Other genera either have are straighter than in other genera, and the resultant rather poorly developed or no dentary processes. mandibular arches are considerably more pointed Very small coronoids that bear teeth occur on than in more primitive genera. The meckelian the inner margins of the dentaries in larval grooves are completely, or almost completely Desmognathus, Leurognathus, Gyrinophilus, Pseu­ closed, and the anterior extensions of the prearticu­ dotriton, Stereochilus, Eurycea, Manculus, Typhlo­ lars are entirely enclosed within the 0-shaped triton, and H emidactylium, and in adults of the dentaries. These specializations are related to func­ paedogenetic species. No indication of the element tional changes in the three genera. The organisms has been seen in embryos of species which undergo force their way beneath rocks and into crevices, direct development. and the rigid, relatively straight, strengthened mandibles are presumably of significance in such Hyobranchial Skeleton behavior. Larval structure The mandibular condition in Desmognathus, Desmognathus and Leurognathus have four Leurognathus, and Phaeognathus is most closely larval epibranchials, but three are present in larvae approached by Stereochilus, in which the mandibu­ of all other genera, and in the paedogenetic adults. lar arches are relatively elongate and pointed, and Bolitoglossa subpalmata embryos, removed from the anterior portions of the meckelian grooves are egg capsules, have adult configurations with a single closed. pair of epibranchials, and the same is true of hatch­ The prearticulars in plethodontids are strongly ling Plethodon and Batrachoseps. Dent (1942) has inflected dorsomedially and shelves are small and reported three epibranchials in an embryo ( un­ directed medially in Gyrinophilus and Pseudotri­ specified size) of Plethodon cinereus. ton. In Stereochilus, Eurycea, Manculus, Hemi­ Larval hyoids are found in adults of the paedo­ dactylium, and Typhlotriton the shelves initially genetic species, and all have branchial plates. The are inflected medially, then dorsomedially. Similar branchial plates of Typhlomolge are notably elon­ stiuations are encountered in Desmognathus, gated and proportionately longer than in other Leurognathus, and Phaeognathus, but the portions genera. The ceratobranchial arms are swept pos­ on which muscles insert are relatively larger than teriorly and make considerably smaller angles with in other genera. The shelves are also directed dorso­ each other than in other genera. medially in Plethodon, Ensatina, and Aneides, and the inflected portions are relatively very large in Adult structure adult A. ferreus, A. flavipunctatus and A. lugubris. Several important papers (Piatt, 1935; Tanner, Prearticulars are very small in Ensatina. The rela- 1952) have dealt with tongue morphology, includ- 1966 EVOLUTION OF LUNGLESS SALAMANDERS 29 ing some aspects of the musculature. It is especially level of the midpoint of the first basibranchial, and important that the attachment of the tongue to the at that point small processes proceed antero­ anterior margin of the mouth be understood. medially; the processes serve as the point of Genioglossal muscles primitively extend as well attachment of suprapeduncularis muscles. The developed strap-like structures from the area of the ceratohyals extend well anterior to the processes mandibular symphysis dorsally into the tongue. as attenuated rods. They bind the anterior border of the tongue and Tanner ( 1952) has indicated that the ceratohyals aid in returning the extended tongue into the of Bolitoglossa are noticeably shorter than those of mouth. The primitive condition is encountered other neotropical genera, and that the elements do ·in Desmognathus, Leurognathus, Phaeognathus, not proceed much beyond the origin of the rela­ Plethodon and Aneides (the muscles considered to tively broad suprapeduncularis muscles. Shortened be diagnostic of desmognathines by Tanner, 1952, ceratohyals are also present in Hydromantes, and are the genioglossals) . The fibers proceed for a both Bolitoglossa and Hydromantes lack sublingual abort distance medially from their origin on the folds, structures related to the anterior elongation dentaries, then swing abruptly dorsally and insert of the ceratohyals. Sublingual folds occur in all in the tongue. The genioglossals are modified and neotropical genera except Bolitoglossa, and in a little elongated in Ensatina and Hemidactylium. several midlatitude genera (e.g., Pseudotriton), but Each muscle originates somewhat posteriorly, along the folds appear to be artifacts of the structure of the medial margin of the dentary, and proceeds the ceratohyals and of the adetoglossal condition. almost directly medially, then abruptly moves dor- There is no evidence to support Tanner's assurances . sally into the tongue. A bizarre arrangement is that all folds are homologous structures. seen in Batrachoseps, in which the fibers originate near the articular end of the mandible and proceed Comua · anteriorly following the curve of the mandible. Near the anterior end of the mandible the fibers Paired comua are associated with the anterior proceed medially, resembling the medially directed portions of the first basibranchials in all genera · muscle in Ensatina and H emidactylium, and finally except Hydromantes, in which the tongue muscula­ move dorsally at the midline. ture normally associated with the comua is firmly · Anterior portions of the tongue are held tightly attached to the anterior basibranchial tip. The re­ • in the mouth by the stout genioglossals in Desmog- mainder of the genera may be grouped as follows: • nathus, Leurognathus, Phaeognathus, Plethodon, 1. Cornua distinctly separated from the first basi­ · and Aneides. In Ensatina and Hemidactylium branchials, joined by connective tissue. Desmog­ .· movement is possible, and tongues may be pro- nathus, Leurognathus, Phaeognathus, Gyrin­ pelled for some distance out of the mouth. The ophilus, Pseudotriton, Stereochilus, Eurycea, · tongue of Batrachoseps may be protruded almost Manculus, Typhlotriton, Hemidactylium, Pleth­ •. as far as in those genera which lack genioglossals: odon, Aneides, Ensatina. . Gyrinophilus, Pseudotriton, Stereochilus, Eurycea, 2. Comua closely applied and broadly attached to Manculus, Typhlotriton, Hydromantes, Bolito­ the first basibranchials, often seemingly or . glossa, Oedipina, Pseudoeurycea, Chiropterotriton, actually continuous with the basibranchials. Parvimolge, Lineatriton, and Thorius. In Stereo­ Batrachoseps, Bolitoglossa, Oedipina, Pseudo­ . chilus and Typhlotriton, the tongues are attached eurycea, Chiropterotriton, Parvimolge, Linea­ anteriorly by fleshy, non-muscular tissue. triton, Thorius. Ceratohyal Cornual attachment is very clear for most of the Anteriorly expanded ceratohyals are character­ genera. However, the comua of Typhlotriton are istic of most plethodontids, but the structures are relatively broad-based and superficially appear to attenuated anteriorly in Hydromantes, Oedipina, be in direct continuity with the first basibranchial. southern species of Chiropterotriton (bromeliacia, Histological preparations reveal that the joint abscondens, see Rabb, 1958), Lineatriton, Parvi­ region contains densely packed cells similar to molge richardi, and Thorius (see Tanner, 1952). cartilage cells, but interstitial substance is absent in Attenuation occurs only in specialized groups, and the joint region. The cartilages are not in direct is a specialized condition. continuity. Histological preparations of Batracho­ The attenuated ceratohyals of Oedipina are diag­ seps and Chiropterotriton clearly show that the nostic of the genus (Tanner, 1952). The posterior junction of the comua and first basibranchial is an cylindrical portions flatten slightly at about the area containing cartilage cells and interstitial sub- 30 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 stance; there is no joint. The joint is sharp and clear similar positions. They differ in that the lingual in Ensatina, with the region between the cartilages cartilages are squat (usually) rather than elongate, consisting of densely packed fibroblasts and con­ and separated from (usually) rather than con­ nective tissue fibers. tinuous with the first basibranchials. The genera of group 1 may be placed in sub­ Plethodontids have long been grouped on the groups. The narrow, elongate cornua of Desmogna­ basis of tongue attachment. Uzzell (1961) intro­ thus, Leurognathus, and Phaeognathus arise from duced the useful terms deto- and adetoglossal to narrow bases, and are about one-half or more of the refer to presence and absence respectively of an­ total length of the first basibranchial. Those of terior tongue attachment, or to the attached and Gyrinophilus, Pseudotriton, Stereochilus, and H emi­ free-tongued conditions. Detoglossal genera are dactylium are moderately broad-based structures Desmognathus, Leurognathus, Phaeognathus, Ster­ one-fourth to one-half the length of the first basi­ eochilus, Typhlotriton, Hemidactylium, Plethodon, branchial; in contrast to those of the other three Aneides, Ensatina, and Batrachoseps, but of these genera the comua of Stereochilus are expanded the genioglossal muscles have been lost in Typhlo­ distally rather than pointed or cylindrical. The triton and Stereochilus, and elongated from greater small, tapering, pointed comua of Eurycea, to lesser extents in Batrachoseps, Ensatina, and Manculus, and Typhlotriton are less than one-half H emidactylium. It is significant that anterior first the length of the first basibranchials. Old adults of basibranchial extensions are present as consistently Typhlotriton have comua with slight distal expan­ well developed, distally knobbed structures only in sion. The large, distally expanded, wing-like cornua genera in which the genioglossal muscles are short of Plethodon, Ensatina, and Aneides are one-third and strap-like (see Piatt, 1935), and in which es­ (Ensatina) or from two-fifths to three-fourths the sentially no movement of the anterior portions of length of the first basibranchials. Expanded distal the tongue out of the mouth is possible (Desmog­ portions are extremely thin. Proximal portions are nathus, Leurognathus, Phaeognathus, Plethodon, cylindrical. Aneides). Ensatina, anatomically detoglossal but Batrachoseps has the longest cornua of the group with a fairly protrusible tongue, has relatively 2 genera, and its long, thread-like cornua curve smaller processes than the above genera, and the gracefully in 180° arcs with total lengths of from processes are pointed rather than distally expanded. one-third to one-half the length of the first basi­ Histological preparations show that the region just branchial. The cornua of Bolitoglossa strikingly re­ anterior to the comual attachment in Ensatina is semble those of Batrachoseps, but are straighter and strongly depressed and contains densely arranged shorter (one-fifth to one-third times first basi­ cartilage cells but notably small amounts of inter­ branchial length) . Cornua of the remaining genera stitial substance. This is in marked contrast with are successively shorter and inconspicuous; each is regions craniad and caudad where the cells are equal to less than one-fifth of the first basibranchial widely separated by large amounts of interstitial length in Pseudoeurycea, Chiropterotriton, and substance. This region is apparently flexible, and Oedipina, and less than one-tenth in some Chirop­ the anterior extension bends ventrally during terotriton (e.g., bromeliacia, abscondens), Par­ tongue flipping. A joint apparently forms in the vimolge, Lineatriton, and Thorius. depressed region in some large individuals. Dissec­ tions reveal that portions anterior to the attachment Lingual cartilage of the comua are occasionally separated from the Piatt (1935) reported lingual cartilages to be first basibranchials in large Ensatina, and form present in Gyrinophilus, Pseudotriton, Eurycea, small, triangular, cartilaginous structures on which and Manculus, found "forerunners" in Stereochilus insert hyoglossal muscle fibers. These structures and Typhlotriton, but found no trace in other must be considered lingual cartilages, and their plethodontids. He considered lingual cartilages to appearance in Ensatina demonstrates the relation­ be "paleotelic" features, homologous with the ship of the anterior processes or first basibranchial "Sebnenplatte" of Salamandra, but his work was extensions and the lingual cartilages. Both anterior biased by his assumption that plethodontids had extensions and lingual cartilages are absent in been derived from salamandrids. Batrachoseps and Hemidactylium, the other genera It is apparent that lingual cartilages are homo­ that have genioglossal muscles. I am unable to cor­ logous with the anterior extensions of the first roborate Hilton's (1959) observation that Hemi­ basibranchials of primitive salamanders, as sug­ dactylium has a lingual cartilage. gested by Hilton (1947 a). Both serve as the sites Stereochilus and Typhlotriton have true lingual of hyoglossal muscle insertion, and both occupy cartilages despite Piatt's ( 1935) report that a 1966 EVOLUTION OF LUNGLESS SALAMANDERS 31 "Sehnenplatte" or membranous forerunner is pres­ condition. Lingual cartilages may be related to flip­ ent. In Typhlotriton lingual cartilage formation ping the tongue pads, but they are lost in many occurs rather late during ontogeny. In small adults accomplished tongue-flippers (e.g., Bolitoglossa) the cartilages have a membranous appearance, as and are not functionally essential. Presence of reported by Piatt. Microscopic sections of a tongue lingual cartilages may represent a relatively ad­ of an adult Typhlotriton (53.1 mm.) reveal that vanced, functionally significant stage in the loss of a well developed lingual cartilage is present and the anterior first basibranchial extensions. Evidence that it is attached to the anterior end of the first is strong for parallel gain and subsequent parallel basibranchial at either lateral margin by dense loss of lingual cartilages as part of an over-all mor­ aggregations of cartilage cells. On the midline only phological trend which is as follows: dense connective tissue fibers hold the vertically 1. Well developed anterior extensions of the first oriented cartilage to the horizontal basibranchial. basibranchials present; no indications of lin­ Well developed lingual cartilages are present in gual cartilages. Desmognathus, Leurognathus, Gyrinophilus, Pseudotriton, Eurycea, and Mancu­ Phaeognathus, Plethodon, Aneides. lus. In these genera the structures are irregular, 2. Anterior basibranchial extensions small, point­ squat, moderately broad elements that resemble ed, flexible or partially detached; lingual carti­ those of Stereochilus and Typhlotriton rather than lage-like structures thus formed. Ensatina, some those of Ensatina. Pseudoeurycea. The only remaini.rig midlatitude genus, the truly adetoglossal Hydromantes, lacks both lingual car­ 3. Completely detached lingual cartilages; first tilages and anterior basibranchial extensions. basibranchial extensions absent. Gyrinophilus, Tanner (1952) corroborated Piatt's observation Pseudotriton, Eurycea, Manculus, Stereochilus, that lingual cartilages are absent in neotropical some Chiropterotriton, Parvimolge townsendi, genera. Rabb (1955), however, reported lingual Typhlotriton,? some Hemidactylium. cartilages to be present in some individuals of a 4. Poorly defined tissue present at first basibran­ newly described species, Chiropterotriton priscus, chial tips which may represent lingual cartilage and I have found the cartilages in some Pseudo­ vestige or remnant; no anterior basibranchial eurycea (bellii, cephalica), some Chiropterotriton extensions. Some Typhlotriton, some Pseudo­ (chiropterus, multidentatus), and in Parvimolge eurycea, some Chiropterotriton. townsendi. Small triangular structures similar to 5. Both anterior basibranchial extensions and lin­ those of Ensatina are present in Pseudoeurycea gual cartilages absent. Hemidactylium, Hydro­ werleri, and similarly shaped structures, but more mantes, Batrachoseps, some Pseudoeurycea, definitely separated from the first basibranchials, some Chiropterotriton, Lineatriton, Oedipina, are present in Chiropterotriton multidentatus. The Parvimolge richardi, Bolitoglossa, Thorius. cartilages of other species are rounded distally rather than pointed, and are short, squat structures. The lingual cartilage of plethodontids is clearly The cartilages are never as distinctly separated from not derived from the lingual process of the basi­ the first basibranchials, nor as broad and fiat, as branchial of porolepiform :fishes, as suggested by those of Gyrinophilus, etc. The poorly defined, in­ Jarvik (1963), but is a new element that appears completely detached tissue found at the anterior for the first time in plethodontids and is clearly ends of the first basibranchials in Pseudoeurycea related to the evolution of tongue flipping. Piatt · goebeli, P. leprosa, Chiropterotriton dimidiatus, (1935) has discussed the homologies of the lingual . and C. bromeliacia may represent vestigial lingual cartilage. He considered it to be the homologue of cartilages. the "Sehnenplatte" of Salamandra. The "Sehnen­ It is apparent that lingual cartilages are homo­ platte" is a ligamentous plate which serves as the logous with the anterior first basibranchial exten­ site of insertion of the hyoglossal and abdomino­ sions of anatomically and functionally detoglossal hyoideal (anterior extension of rectus abdominis, genera. Separation of anterior extensions and sub­ also called rectus cervicis profundus) musculature. sequent formation of lingual cartilages appears to This ligamentous plate is present as a well devel­ be associated with acquisition of a degree of tongue oped structure in microscopic sections of the freedom. It may be assumed that retention of an­ tongues of plethodontids dorsal to the first basi­ terior extensions is functionally undesirable in at­ branchial in the posteroventral portion of the tainment of the adetoglossal condition, since well tongue pad. Both lingual cartilage and ligamentous developed extensions are absent in all genera that plate are present in such forms as Typhlotriton. The have even a slight tendency toward the free-tongued hyoglossal musculature originates on the lingual 32 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 cartilage and at least some of its posterior fibers of other genera, despite their relatively slight insert on the plate. Clearly the cartilage and plate expansion. are functionally distinct, non-homologous struc­ The genera of group 3 have essentially con­ tures. The presence of lingual cartilages in pletho­ tinuous comua which arise from the anterior ends dontids in no way supports Piatt's theory of of the basibranchials, and give the impression, more plethodontid origin from salamandrids. Since all apparent than real, that the anterior ends are ex­ evidence indicates that lingual cartilages occur only panded. When present, expansion is gradual, not in plethodontids that have attained some degree of sharply defined. tongue freedom, Piatt was correct in stating that the lingual cartilage is not homologous with the Proportions of articulated elements otoglossal cartilages of ambystomatids, which serve Plethodontid genera may be grouped according as tongue pad supports. to the relative proportions of the articulated ele­ ments as follows: First basibranchial 1. Sequence of length from longest to shortest; Plethodontid genera may be grouped on the basis ceratobranchial I-( ceratobranchial II, basibran­ of first basibranchial shape: chial I, or epibranchial). Desmognathus, Leu­ 1. Major basibranchial expansion near midpoint. rognathus, Phaeognathus, Plethodon, Aneides. Desmognathus, Leurognathus, Phaeognathus, 2. (Ceratobranchial I or epibranchial)-basibran­ Hydromantes. chial I-ceratobranchial II. Ensatina. 2. Major expansion slightly posterior to attachment 3. Epibranchial-( ceratobranchial I or basibran­ of cornua. Gyrinophilus, Pseudotriton, Stereo­ chial 1)-ceratobranchial II. Manculus. chilus, Eurycea, Manculus, Typhlotriton, Hemi­ 4. Epibranchial-ceratobranchial 1-(ceratobranchial dactylium, Plethodon, Ensatina, Aneides. II or basibranchial I). Eurycea, Batrachoseps 3. No major expansion, but basibranchial broad­ wrighti. est at anterior end. Batrachoseps, Bolitoglossa, 5. Epibranchial-ceratobranchial 1-ceratobranchial Oedipina, Pseudoeurycea, Chiropterotriton, Par­ II-basibranchial I. Gyrinophilus, Pseudotriton, vimolge, Lineatriton, Thorius. Stereochilus, Typhlotriton, H emidactylium. 6. Epibranchial-basibranchial 1-ceratobranchial 1- Hydromantes has cylindrical basibranchials that ceratobranchial II. Hydromantes, most Batra­ have thin, flattened, flange-like expansions on either choseps, Bolitoglossa, Oedipina, Pseudoeurycea side. The greatest width is equal to about 3.5 to 5 (one species, P. bellii, has ceratobranchial I and times the least diameter (width ratio) of each basi­ basibranchial l reversed), Chiropterotriton, branchial. If the flanges are disregarded the first Parvimolge, Lineatriton, Thorius. basibranchials resemble those of group 3. In the other genera of group 1 each first basibranchial is The first ceratobranchials and the epibranchials slightly flattened in the area of greatest expansion, are of nearly equal length in Ensatina, and the epi­ but no flanges are present. Width ratios are from branchials are the longest elements in all of the 1.5 to 3. remaining genera, except those of group 1. Epi­ Thin lateral extensions arise from the anterior branchial length is clearly associated with acquisi­ one-third of each basibranchial in Eurycea, Man­ tion of tongue freedom; the longer the epibranchials culus, and Typhlotriton. Width ratios are from 5 (relatively), the farther the tongue can be extended to 6, the greatest in the family, and basibranchials from the mouth. Significantly, short epibranchials are paddle-shaped. The basibranchials of Pseudo­ (at least shorter than the first ceratobranchials) are triton and Gyrinophilus are rather flattened, and found only in the five genera (group 1, above) have width ratios of 2.5 to 3. Width ratios are less which have no osteological or myological modifica­ than 2.5 in Stereochilus, Hemidactylium, Pletho­ tions for even partially freeing the anterior tongue don, Ensatina, and Aneides, but the expanded areas margins. The condition in which the first cerato­ are much nearer the anterior ends in the first two. branchials are the longest elements and in which Plethodon, Ensatina, and Aneides differ from there is no tongue freedom is the condition in primi­ H emidactylium in lacking sharply defined areas of tive families; it is clearly the primitive plethodontid expansion; Stereochilus is intermediate between the condition. two conditions. The first basibranchials of Stereo­ The first are longer than the second ceratobran­ chilus and Hemidactylium resemble those of Gyri­ chials in all genera, and are longer than the first nophilus and Pseudotriton more closely than those basibranchials in all but group 6. The first cerato- 1966 EVOLUTION OF LUNGLESS SALAMANDERS 33 branchials are relatively shortest in Hydromantes from the mouth. For efficiency in transmission of (about one-half the length of the first basibran­ force it is functionally advantageous for the lines chials), and longest in Stereochilus (about one and of force between articulated elements to be rela­ two-thirds times first basibranchials). tively straight. In all plethodontids the second cera­ The genera may be grouped according to second tobranchials articulate directly with the posterior ceratobranchial length: end of each first basibranchial, while the first cera­ tobranchials are attached by ligaments to the side 1. Second ceratobranchials less than 0.8 times first of the basibranchial. Primitively the first cerato­ basibranchials. Hydromantes, most Batracho­ branchials are the more dominant force trans­ seps, Bolitoglossa, Oedipina, Pseudoeurycea, mission elements, and the dominance is maintained Chiropterotriton, Parvimolge, Lineatriton, in most midlatitude plethodontids. In the eastern Thorius. North American adetoglossals epibranchials are not 2. Second ceratobranchials more than 0.8 but less greatly elongated, but the first ceratobranchials than 1.3 times first basibranchials. Batrachoseps have acquired specialized braces of two types: wrighti, Desmognathus, Leurognathus, Phaeog­ ( 1 ) attachment behind the specialized first basi­ nathus, Gyrinophilus, Pseudotriton, Eurycea, branchial anterior expansion (Manculus, Eurycea), Manculus, Typhlotriton, Plethodon, Aneides, (2) cartilaginous struts attached to the expanded Ensatina. ends of the ceratobranchials (Gyrinophilus, Pseudo­ 3. Second ceratobranchials more than 1.3 times triton). In Hydromantes, Batrachoseps, and the first basibranchials. Hemidactylium, Stereo­ neotropical genera the primitively weak, relatively chilus. long, slightly sigmoid-shaped second ceratobran­ chials have become strongly shortened and straight, The relative dimensions of the first and second apparently as a result of selection pressures associ­ ceratobranchials may also be used to group the ated with attainment of the free-tongued condition genera: and correlated lengthening of the epibranchials. In these genera the first ceratobranchials have been 1. First ceratobranchials of significantly greater weakened and anteriorly bowed, and in some gen­ bulk than second ceratobranchials; minimal era (e.g., Thorius) have lost their direct articulation diameter of first equal to or greater than with the basibranchial. As a result of these changes maximal diameter of second. Desmognathus, there is a direct, rather straight line of force trans­ Leurognathus, Phaeognathus, Gyrinophilus, mission from each epibranchial through the short, Pseudotriton, Eurycea, Manculus, Stereochilus, stout ceratobranchial to the first basibranchial. The Typhlotriton, Hemidactylium, Plethodon, En­ mechanics of tongue action have recently been satina, Aneides. discussed by Uzzell ( 1961), who cites the enlarged 2. Second ceratobranchials of significantly greater second ceratobranchials and their tendency to cal­ bulk than first ceratobranchials; minimal di­ cify in certain neotropical genera (see below) as ameter of second equal to or greater than evidence that pressure applied at the posterior ends maximal diameter of first. Hydromantes, Batra­ of the epibranchials is probably transmitted to the choseps, Bolitoglossa, Oedipina, Pseudoeurycea, first basibranchial through the second rather than Chiropterotriton, Parvimolge, Lineatriton, first ceratobranchials. Thorius. As suggested above the length of the epibran­ chials is highly significant, and the genera may be First ceratobranchials that are stouter than sec­ grouped as follows: ond ceratobranchials are typical of primitive sala­ 1. Epibranchials less than 1.25 times first basi­ manders, and such a condition is primitive in branchials. Desmognathus, Leurognathus, plethodontids. All of the genera of the latter group Phaeognathus, Plethodon, Ensatina, Aneides. (2) are functionally free-tongued (although Batra­ choseps is anatomically detoglossal), and all have 2. Epibranchials 1.5 to 2 times first basibranchials. greatly elongated epibranchials. The subarcual Gyrinophilus, Pseudotriton, Stereochilus, Eury­ rectus muscles envelop each epibranchial, and at­ cea, Manculus, Typhlotriton, Hemidactylium, tach anteriorly to each ceratohyal. When these Thorius. muscles are contracted, the epibranchials are 3. Epibranchials 2 to 3.5 times first basibranchials. moved forward. As a result the entire articulated Hydromantes, Batrachoseps, Bolitoglossa, Oedi­ hyobranchial skeleton, surrounding muscle, and pina, Pseudoeurycea, Chiropterotriton, Parvi­ connective tissue (i.e., the tongue) are extended molge, Lineatriton. 34 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 Since the basibranchials in those genera which Second basibranchials are absent in seventy-seven have anterior basibranchial extensions are propor­ Batrachoseps of five species examined by me. A tionately longer than in those that lack the ex­ very well developed but nonossified and noncalci­ tensions, the epibranchials have also been compared fied inscriptio tendinis is found in the longitudinal with the first ceratobranchials: throat musculature of Batrachoseps in the same 1. Epibranchials less than or almost equal to length position as a second basibranchial, and less well of first ceratobranchials. Desmognathus, Leu­ developed structures are found in other genera of rognathus, Phaeognathus, Plethodon, Ensatina, group 2. The element is ligamentous, not cartilagi­ Aneides. nous, and is considered a second basibranchial vestige by Tanner (1952). 2. Epibranchials more than 1 but less than 1. 7 The genera of group 1 may be further grouped times first ceratobranchials. Gyrinophilus, Pseu­ on the basis of second basibranchial structure: dotriton, Stereochilus, Eurycea, Manculus, Typhlotriton, H emidactylium. 1. Width of second basibranchials in adults greater 3. Epibranchials 1.8 to 6 times first ceratobran­ than length of first basibranchials. Gyrinophilus, chials. Hydromantes, Batrachoseps, Bolito­ Pseudotriton, Eurycea, Manculus, Hemidactyl­ glossa, Oedipina, Pseudoeurycea, Chiropterotri­ ium. ton, Parvimolge, Lineatriton, Thorius. 2. Width of second basibranchials two-thirds to three-fourths times first basibranchials. Stereo­ The epibranchials of group 3 genera are remark­ chilus, Typhlotriton. ably long. The longest occur in Hydromantes, in 3. Width of second basibranchials less than two­ which epibranchials may be 5.9 times the first thirds times first basibranchials. Desmognathus, ceratobranchials. Very long epibranchials are also Leurognathus, Phaeognathus, Plethodon, En­ found in Chiropterotriton (maximum, 4.6 times satina, Aneides. first ceratobranchials), Oedipina ( 4.0), Parvi­ molge (3.7), Pseudoeurycea (3.2), Bolitoglossa Second basibranchials are very strongly triradiate (2.9), Batrachoseps (2.6), and Thorius (2.7). The with very long arms and long, medial, posteriorly shortest adult epibranchials in group 3 are found in extending processes in Gyrinophilus, Pseudotriton, Batrachoseps wrighti ( 1.8). Of the group 1 and 2 Eurycea, Manculus, and Typhlotriton. The struc­ genera only Eurycea has epibranchials that are tures are distinctly triradiate, but have shorter more than 1.4 times the first ceratobranchials lateral and posterior processes than the above gen­ (l.33-1.67). era in Stereochilus, Typhlomolge, and Haideotri­ ton. In the latter genera, and in Typhlotriton, Second basibranchial second basibranchials are somewhat cruciform Presence or absence of second basibranchials in with arms strongly swept posteriorly. Second basi­ adults serves as a character to group plethodontid branchials of remaining genera are usually bowed genera: with arm-tips extended posterolaterally, and with no posteromedial projections. Small posteromedial 1. Second basibranchials present in adults. Des­ enlargements may be present occasionally. Second mognathus, Leurognathus, Phaeognathus, Gy­ basibranchials are notably reduced in size in rinophilus, Pseudotriton, Stereochilus, Eurycea, Phaeognathus and some species of Plethodon (e.g., Manculus, Typhlotriton, Hemidactylium, Hai­ dunni). deotriton, Typhlomolge, Plethodon, Ensatina, Aneides. Calcification 2. Second basibranchials absent in adults. Hydro­ mantes, Batrachoseps, Bolitoglossa, Oedipina, Calcifications of elements other than the second Pseudoeurycea, Chiropterotriton, Parvimolge, basibranchials occur in some genera. They are ap­ Lineatriton, Thorius. parently not of universal occurrence in any single species or genus, and are more common in the An extremely tiny second basibranchial, a small larger members of a given population (Uzzell, dot of ossified tissue, is present in the largest 1961; Wake, 1963). The following list contains all Pseudoeurycea smithi examined (73.2 mm.). Cope observations of calcified elements known to me, (1889) illustrated a second basibranchial in Batra­ and is a composite of my personal observations as choseps, and Piatt (1935) reported a reduced sec­ well as those of Wiedersheim (1877), Piatt (1935), ond basibranchial, but Hilton (1947) and Uzzell Hilton (1947a), and Uzzell (1961). Personal (1961) say the structure is absent in their material. observations ( *), unique observations ( * *) : 1966 EVOLUTION OF LUNGLESS SALAMANDERS 35 1. Ceratohyals (calcification limited to region of 1. Desmognathus, Leurognathus, Phaeognathus. posterior hook unless otherwise noted). Larval 2. Plethodon, Ensatina, Aneides. Eurycea (extensive, into expanded portions) **, 3. Gyrinophilus, Pseudotriton, Stereochilus, Eury­ Eurycea**, Typhlotriton**, Hemidactylium**, cea, Manculus, Typhlotriton, Haideotriton, Batrachoseps* *, Oedipina (extensive, at hook Typhlomolge, Hemidactylium. and into expanded portions) * *. 4. Hydromantes, Batrachoseps, Bolitoglossa, Oedi­ 2. First basibranchials (calcification usually lim­ pina, Pseudoeurycea, Chiropterotriton, Parvi­ ited to portions of central cores) . Desmognathus molge, Lineatriton, Thorius. (entire element)*, Leurognathus**, Manculus, Typhlotriton**, Hemidactylium**, Plethodon*, These groups correspond exactly with my pro­ Aneides**, Lineatriton*, Thorius**. posed tribal grouping (see under systematic sum­ 3. Comua of first basibranchials. Desmognathus* *, mary). At least one genus in each of the four Leurognathus (extreme proximal portions groups has genioglossal muscles and is thus tech­ only)**. nically detoglossal (see Uzzell, 1961). Assuming 4. First ceratobranchials (central portions) . Leu­ the above groups to be natural, adetoglossy has rognathus (proximal portions only)**, Man­ arisen in parallel in groups 3 and 4 and may be culus, H emidactylium * *, Parvimolge town­ evolving in group 2 (through Ensatina). Piatt sendi*, Lineatriton*. (1935) also concluded that free tongues had arisen 5. Second ceratobranchials (central portions or in parallel in groups roughly corresponding to my complete). Leurognathus**, Manculus, Hemi­ groups 3 and 4. He based his conclusions on the dactylium * *, Parvimolge townsendi*, Linea­ premise that the lingual cartilage, a supposed triton*. paleotelic character, was not present in Hydro­ mantes and the neotropical genera. Tanner (1952) 6. Epibranchials. Larval Gyrinophilus*, larval has discussed the illogic of this argument, and it Eurycea*, larval Stereochilus* *, Batrachoseps has been shown (see above) that Piatt's data were (extreme distal tips)**, Parvimolge townsendi incorrect. Nevertheless there is some congruence (proximal portions), Lineatriton (proximal between Piatt's generic groupings, based on gen­ portions). eralities concerning hyobranchial skeleton and four Uzzell (1961) has recently reported sequence of muscles, and mine. calcification of hyobranchial elements in Parvi­ molge and Lineatriton, and has postulated close Dentition relationship on the basis of sequence similarities, as Each plethodontid tooth is composed of a crown Rabb (1955) had earlier suggested. Calcification and a pedicel (Parsons and Williams, 1962). In sequences reported by Uzzell are: (1) Lineatriton, those genera with generally strengthened jaw clos­ second ceratobranchials-epibranchials-first basi­ ing adaptations (Desmognathus, Leurognathus, branchials, (2) Parvimolge townsendi, second Phaeognathus, Aneides) the pedicels are particu­ ceratobranchials-epibranchials, (3) Manculus, first larly well developed. A remarkable situation is basibranchials-second ceratobranchials-first cerato­ found in Phaeognathus in which the crowns are branchials. In simple distribution Leurognathus extremely shortened, very stocky, and blunted. The and Hemidactylium resemble the condition report­ pedicels are very large and greatly strengthened, ed in Manculus (all may have calcified first cera­ and are almost double the length of the crowns. tobranchials, none have calcified epibranchials) Species with moderately to weakly developed skulls more closely than that reported in Parvimolge and have relatively small, fragile pedicels. Lineatriton (both may have calcified epibranchials, All larval teeth are unicuspid, and during meta­ none have calcified first ceratobranchials). Uzzell morphosis there is a gradual replacement with ( 1961 ) reports that the second ceratobranchials bicuspid crowns. Adult teeth are considerably vari­ calcify before the first in Manculus; in at least some able as to shape and size, and both sexual and Leurognathus the opposite occurs. ontogenetic variation occur. Maxillary, dentary, and vomerine teeth are nor­ Synthesis of hyobranchial characters mally bicuspid with large lingual cusps and small, A synthesis of all of the above information con­ crescentic labial cusps, but some exceptions occur cerning hyobranchial skeletal structure results in (some Aneides, adult male Hydromantes, Eurycea). the following generic groupings (see Figs. 10 and Small individuals of all plethodontid species ex­ 11): cept Aneides lugubris have bicuspid premaxillary 36 MEMOIR OF THE SOUTHERN CAUFORNIA ACADEMY OF SCIENCES VOL.4

'--..:::=:::.-.co ----Bl

~B2

Figure 10. Hyobranchial apparatus. Ceratohyals removed. A. Plethodon yonahlossee, B. Des­ mognathus quadramaculatus, C. Typhlotriton spelaeus. Abbreviations: Bl, first basibranchial; B2, second basibranchial; CO, cornua of first basibranchial, Cl, first ceratobranchial; C2, second ceratobranchial; E, epibranchial; LC, lingual cartilage. Line equals one mm.

Figure 11. Hyobranchial apparatus of bolitoglossine supergenera. A. Hydromantes italicus, B. Batrachoseps pacificus, C. Bolitoglossa robusta. Line equals one mm. 1966 EVOLUTION OF LUNGLESS SALAMANDERS 37 teeth, and bicuspid teeth occur in adults of most Thorius lacks maxillary teeth, but they occur in species. Males of the more primitive genera some numbers in several populations (see Gehl­ (Gyrinophilus, Pseudotriton, Stereochilus, Typhlo­ bach, 1959). Reduction in number of maxillary triton) have premaxillary teeth that are only slightly teeth also occurs in Aneides, especially in members modified, but the teeth are elongated and directed of the lugubris species group (Wake, 1963). In anteroventrally in adult males of at least some general, it may be stated that reduction in maxillary species of Desrnognathus, Leurognathus, Eurycea, dentition is a specialization encountered in highly Manculus, Hemidactylium, Plethodon, Aneides, specialized and advanced groups. Reduction in Hydromantes, Batrachoseps, Bolitoglossa, Oedi­ number coupled with hypertrophy in individual pina, Pseudoeurycea, Chiropterotriton, Lineatriton, tooth size is also a specialized condition. Parvimolge, and Thorius. The elongated teeth may be unicuspid and spine-like, or the labial cusps may be greatly elongated and either projected antero­ Vertebral Column ventrally, or ventrally hooked (see also Noble, Regional di[ferentation 1927 a). Hooked premaxillary teeth are common A single cervical vertebra, the atlas, occurs in in male Plethodon, Pseudoeurycea, Chiropterotri­ all salamanders. Trunk vertebrae in salamanders ton, and Bolitoglossa. Sexual dimorphism is more vary from 12 to over 60, and from 13 to 24 in pronounced in advanced than in primitive genera, plethodontids. A single sacral vertebra is typical and elongated premaxillary teeth have unquestion­ of all salamanders, except sirenids. From 2 to 3 ably appeared more than once. There is no evidence caudosacral vertebrae are present in plethodontids, that the character has disappeared and reappeared but caudal vertebrae are extremely variable, and in the family, the views of Dunn (1927) and are subject to interspecific, individual, ontogenetic, Noble (1927 a) to the contrary notwithstanding. and sexual variation in number. Maxillary teeth are greatly reduced in number Significant variation is encountered in the num­ or absent in some Bolitoglossa ( chica, colonnea, bers of trunk vertebrae of plethodontids. Range occidentalis, rufescens), some Oedipina (elongatus, and modal numbers of trunk vertebrae are pre­ alfaroi), and most Thorius. It has been stated sented in Table 1. The table is based primarily on (Taylor, 1944; Smith and Taylor, 1948) that my observations. Ranges of variation of trunk

TABLE 1. Variation in number of trunk and caudosacral vertebrae.

TRUNK VERTEBRAE CAUDOSACRAL GENUS RANGE MODE VERTEBRAE Desmognathus 15 15 2 Leurognathus 15 15 2 Phaeognathus 21-23 22 2 Gyrinophilus 18-20 19 3 Pseudotriton 17-18 18 3 Stereochilus 18-20 19 3 Eurycea 14-21 15 3 Manculus 17-18 18 3 Typhlotriton 17-20 18 3 Typhlomolge 13-14 14 3 Haideotriton 14 14 3 H emidactylium 15 15 3 Plethodon 14-24 17 3 Ensatina 14-15 14 3 Aneides 15-18 16 3 Hydromantes 13-14 14 3 Batrachoseps 16-21 20 2-3 Bolitoglossa 14 14 2 Oedipina 18-22 20 2 Pseudoeurycea 14 14 2 Chiropterotriton 14 14 2 Parvimolge 14 14 2 Lineatriton 14 14 2 Thorius 14 14 2 38 MEMOIR OF THE SOUTHERN CAUFORNIA ACADEMY OF SCIENCES VOL.4 vertebrae doubtless will be extended for some (19-20). In two genera, Eurycea and Plethodon, genera, and I have relied on the literature for cer­ great intrageneric variation is encountered; gen­ tain genera (Phaeognathus data from Brandon, eralized species have low numbers of trunk verte­ 1965; Typhlotriton and Gyrinophilus from Bran­ brae (e.g., E. bislineata, E. longicauda, E. lucifuga, don, 1966; Plethodon from Highton, 1962). A 15-16; P. vandykei 15-16), but advanced, elongated modal number of 14 is found in ten genera, four species are also found (E. multiplicata, E. tyner­ have modes of 15, one has 16, one has 17, three ensis 20-21; P. cinereus, P. richmondi 18-24). have 18, two have 19, two have 20, and one has Elongation in some groups is correlated with in­ 22. The genera thus cluster about the lower and crease in the numbers of trunk and caudal verte­ higher extremes. A brief survey of the numbers brae (Oedipina, Batrachoseps, Phaeognathus), but of trunk vertebrae in relatively primitive families only numbers of trunk vertebrae are noticeably in­ is instructive: Salamandra '14-15, Salamandrina creased in Gyrinophilus and Stereochilus, and tails 13, Pachytriton 12, Chioglossa 13, Euproctus 13, are relatively short. Pleurodeles 15, Notophthalmus 13, Ambystoma Most plethodontid genera have three caudosacral 14-16, Rhyacotriton 16-17, Dicamptodon 14-15, vertebrae, but two distinct generic groups have two: Hynobius 15-16, Batrachuperus 17, Onychodac­ ( 1) Desmognathus, Leurognathus, Phaeognathus, tylus 11. A tendency for low or relatively low num­ (2) Pseudoeurycea, Chiropterotriton, Lineatriton, bers of trunk vertebrae in primitive groups is Parvimolge, Thorius, Bolitoglossa, Oedipina (i.e., obvious. Numbers on the order of 13 to 17 are the neotropical genera). Batrachoseps is the only probably closer to the primitive plethodontid genus that shows variation in numbers of caudo­ number than those from 18 to 24 range. Thus the sacral vertebrae. B. wrighti and an undescribed relative constancy of the low vertebral number of species may have either two or three, but all other most neotropical genera appears to be a primitive species have only two. Despite this fact, basal tail character in the otherwise specialized and advanced breakage is more common between caudal verte­ group. brae one and two than in front of caudal one, and Elongation, whether associated with increases in one is easily misled as to the number of caudosacral numbers of vertebrae or not, and attenuation have vertebrae. Most primitive salamanders have three arisen in parallel a number of times in plethodontid or four caudosacrals, but Cryptobranchus, some evolution. Trunk elongation has appeared inde­ species of Ambystoma, Chioglossa, and probably pendently at least seven different times: (1) other genera, have only two. Three seems to be Phaeognathus, (2) Gyrinophilus-Stereochilus, (3) the standard number in most hynobiids, salaman­ Eurycea-Manculus, (4) Plethodon, (5) Batracho­ drids, and ambystomatids, and is probably the seps, (6) Oedipina, (1) Lineatriton. Attenuation primitive plethodontid number. Reduction in cau­ has appeared at least as many times, and is found dosacral numbers must have evolved in parallel in to some extent in Phaeognathus, Stereochilus, the family. Eurycea, Manculus, Typhlotriton, Plethodon, Accurate information concerning numbers of Aneides, Batrachoseps, Bolitoglossa, Lineatriton, caudal vertebrae is difficult to obtain because of and Oedipina. Attenuation correlated with over-all the common occurrence of autotomy or breakage size diminution is apparent in Batrachoseps, Man­ of the tail and the regeneration of true vertebrae. culus, Chiropterotriton, Parvimolge, and Thorius. In general most adult plethodontids have between The elongation achieved in each major group of 25 and 50 caudal vertebrae, but more than 50 are plethodontids usually involves both trunk and tail, found in at least five genera: Eurycea (lucifuga, and in most instances the numbers of trunk verte­ longicauda), Manculus, Batrachoseps, Lineatriton, brae increase. Several exceptions occur. In Eurycea and Oedipina. Batrachoseps and Oedipina com­ longicauda and E. lucifuga generalized low num­ monly have more than 60, and Oedipina, with the bers of body vertebrae are maintained, and the longest tails in the family, often has more than 75 trunks are relatively non-elongated, but the num­ and may have 100 (0. longissima). Very short bers of caudal vertebrae are increased. In Linea­ tails are present in Hydromantes and Parvimolge; triton attenuation and elongation results in part maximum numbers of caudals in both are less than from vertebral elongation. Elongation of Oedipina 25. Tail breakage is commonly encountered in all is accompanied by increases in the vertebral num­ genera except Hydromantes and in some species bers of all species. The primitive Batrachoseps of Aneides (aeneus, ferreus, lugubris), in which wrighti has relatively low numbers of trunk verte­ tails are used in locomotion. Tails are very stout brae ( 16-17) , while the highly specialized species proximally in Desmognathus, Leurognathus, (B. attenuatus, B. pacificus) have high numbers Phaeognathus, Gyrinophilus, Pseudotriton, Stereo- 1966 EVOLUTION OF LUNGLESS SALAMANDERS 39 chilus, and in large species of other genera nathus, Leurognathus, and Phaeognathus. Tension (Eurycea, Aneides, Plethodon), and break.age is exerted by highly developed atlantal-mandibular usually limited to distal portions. Tails separate ligaments (Fig. 9), holds the mandibles relatively from the trunks at basal constrictions located just rigid, and forces within the ligaments have been posterior to the last caudosacral vertebrae in H emi­ sufficiently great to result in the development of dactylium, Ensatina, Bolitoglossa, Oedipina, Pseu­ prearticular projections and marked posterior ele­ doeurycea, Chiropterotriton, Parvimolge, Linea­ vation of the atlas. The pedicels are fused in these triton, and Thorius, but breakage can occur distally genera, and very large, broad, dorsal bosses are as well. In other genera breakage can occur at present. The bosses are relatively low anteriorly, almost any point along the tail. but rise steadily posteriorly. Precipitous drops occur posteriorly, and high, broad, bony cliffs are Cervical vertebra formed. The ligaments attach to posteriorly located, Atlases of two very different types occur in the transverse ridges on top of the bosses. Because the plethodontid genera. Vertebrae similar to those of mandibles are relatively fixed, the crania must be primitive salamander families occur in the majority raised to open the mouths of members of these of the genera, but highly specialized elements genera; such a situation is not encountered in other are present in Desmognathus, Leurognathus, and salamanders. Occipital condyles of Desmognathus, Phaeognathus. Leurognathus, and Phaeognathus are very convex, Primitive atlases have well developed, anteriorly and articulate with enlarged, very concave, atlantal directed, odontoid processes that bear large, condylar facets. In addition the odontoid processes laterally oriented facets that articulate with the are very short. Ventral portions of the processes inner walls of the foramen magnum. Anterior mar­ are greatly reduced, and the anterior margins are gins of the process are slightly concave, and the markedly concave and eroded. The lateral articu­ processes are deeply concave dorsoventrally. Paired lar facets of the processes, which primitively articu­ enlarged condylar articulating facets which meet late with the walls of the foramen magnum, are the occipital condyles are borne on stout antero­ very small and fail to enter the foramen; rather they lateral processes. The facets are almost flattened articulate with the inner, ventral portions of the to slightly concave. Conical centra arise from the condylar stalks. Because of these modifications, midpoints of the vertebrae and proceed posteriorly the cranio-atlantal joint is considerably more flex­ and just slightly ventrally. Neural arch pedicels ible in these genera than in other plet~odontids. arise on either side of the centra, meet dorso­ The atlases of Desmognathus, Leurognathus, medially, and form characteristic bony bosses and Phaeognathus are raised above the level of the where the mandibular adductor muscles attach. trunk vertebrae, and the centra are thus directed Hydromantes, Bolitoglossa, Oedipina, Pseudo­ posteroventrally. The angle is so severe in Desmog­ eurycea, Chiropterotriton, Parvimolge, and Linea­ nathus that the dorsal, posterior surfaces of the triton have weakly fused pedicels, with little or no centra are on about the same level as the odontoid dorsal boss formation. The pedicels fuse without processes. In preserved individuals elevation of the forming bosses in Manculus, Batrachoseps (attenu­ atlases over the trunk vertebrae is increased, ap­ atus, pacifi,cus), and Thorius, but some low ridges parently as a result of shortening of the mandibular­ which serve as muscle attachments may be present. atlantal ligaments, and the result is the character­ Median, anteriorly placed, raised nodes or nodule­ istic "pose" of the genera noted by many authors. like processes form in Aneides, Plethodon, Eurycea, The atlases of other plethodontids are not obviously Typhlotriton, Typhlomolge, Haideotriton, Batra­ raised. choseps (wrighti), Pseudoeurycea (unguidentis, smithi), Chiropterotriton (nasalis), and Bolitoglossa Trunk vertebrae (adspersa, cerroensis, dunni). The nodules are Problems relating to vertebral structure and evo­ crater-like in most, and especially in Haideotriton lution are being studied currently by the author. and Typhlomolge. Well developed, raised bosses Much of the information presented here must be are present in Gyrinophilus, Pseudotriton, Hemi­ considered preliminary, and no new information dactylium, Ensatina, and some Aneides. Well de­ will be presented at this time concerning several veloped, nodule-like bosses are present on the aspects of vertebral morphology. In particular posterior portions of the neural arches in Stereo­ questions relating to the fate of the notochordal chilus, and are associated with well developed, canal, the ontogeny of the intervertebral articula­ vertical crests. tions, evolution of the vertebral centrum, and Highly specialized atlases are found in Desmog- mechanism of tail autotomy are deferred to a later 40 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 date. Reference is made t-0 my earlier statements medial hypapophyses which are better developed concerning these points (Wake, 1963). than in any other genus. Hypapophysial keels are well developed on the anterior vertebrae of Des­ Centrum structure mognathus, but poorly developed posteriorly. Hy­ Central dimensions vary considerably in pletho­ papophyses are best developed in large species dontids, and are correlated to a large extent with (D. quadramaculatus, D. monticola), poorly devel­ elongation and relative thickness of trunk muscula­ oped in the small species (D. aeneus, D. wrighti). ture. Posterior central diameter of adult, mid­ The best developed hypapophyses in the above trunk vertebrae is from about 2 to 2.8 times the genera are anteriorly placed, knob-like proc115ses central length ( centrum ratio) in Desmognathus, with small, posteriorly oriented keels, and small, Leurognathus and Phaeognathus. The smallest ratio posteriorly placed low keels. The keels may be con­ is found in the elongated Phaeognathus in which tinuous on anterior vertebrae in Leurognathus. The the centra are relatively short and stout. hypapophyses of Phaeognathus differ sharply from Typhlomolge and Haideotriton are short bodied, those of Desmognathus and Leurognathus. Well relatively weakly muscularized, aquatic genera for developed, uninterrupted keels extend virtually the which the water's buoyancy provides the major full length of the centra on about the first seven body support, and in both the vertebral centra are vertebrae, and progressively lower keels, or low small. Central ratios in both are about 3, and reflect ridges, are present on the more posterior vertebrae. the small size of the posterior portions of the centra Anterior knobs, so conspicuous in the other genera, rather than increased central length. Related genera are poorly developed and obvious only anteriorly. (Gyrinophilus, Pseudotriton, Stereochilus, Eurycea, Remaining plethodontids lack well developed Manculus, Typhlotriton and Hemidactylium) have hypapophyses, but remnants occur in Pseudotriton, central ratios of from 2.2 to 2.5, and have neither Stereochilus, Eurycea and very large Typhlotriton. very stout nor very elongated centra. Slight ridges are present on the posterior portions Relatively stout vertebral centra are present in of the sixth vertebral centrum of Pseudotriton, and Ensatina ( centrum ratio 2.5) and Aneides ( cen­ somewhat larger ridges are present on the posterior trum ratio 1.8 to 2.4). Plethodon may be divided portions of centra three through six in Stereochilus. into two groups: (1) an eastern group with rela­ Small, keel-like projections are present on the tively stout centra and ratios of 2.5 to 2.8 (cinereus, posterior margins of centra four or five, or both, in dorsalis, glutinosus, jordani, neomexicanus, ouachi­ Eurycea bislineata, and on centrum five in E. tae, richmondi, wehrlei, welleri), (2) a western aquatica, but none occur in other species. Old group with slenderer centra and ratios of 2.9 to 3.2 Typhlotriton have a tiny vestige on centrum five. (dunni, elongatus, larselli, vandykei, vehiculum). I Long, well developed, very stout, paired basa­ am unable to recognize the three major types of pophyses are present on the posterior central mar­ vertebrae in Plethodon proposed by Highton gins in Desmognathus and Leurognathus, and the (1962). processes are only slightly less well developed in Hydromantes vertebrae have ratios of 3 to 3.9, Phaeognathus. Basapophyses are well developed in and the centra are somewhat elongated. Central many Eurycea, moderately developed in Typhlo­ diameter is relatively small. The slender, elongated triton, present on a few anterior vertebrae, but species of Batrachoseps have small, semi-elongated variously developed posteriorly in Pseudotriton, centra with ratios of 3 to 3.5. The neotropical gen­ Manculus, and Typhlomolge, small and variously era also have rather small centra. Pseudoeurycea developed in Hemidactylium and Stereochilus, and and Chiropterotriton have relatively the largest virtually absent in Haideotriton and Gyrinophilus. centra and the lowest central ratios (2.5 to about Ensatina lacks basapophyses, and the processes are 3). Centra are somewhat slenderer and more absent or only slightly indicated in Plethodon and elongated in Bolitoglossa (ratios 2.8 to 3.5). Very Aneides; freely projecting processes are never pres­ slender, elongated centra with high ratios are found ent. The processes are absent in Batrachoseps and in the remaining genera (Parvimolge about 3, Hydromantes, and are irregularly distributed in the Thorius 3.4 to 3.7, Oedipina 3.0 to 3.5, Lineatriton neotropical genera. Basapophyses are absent in 4.5). Lineatriton is an exceptionally elongated form Oedipina, Pseudoeurycea, Parvimolge, Lineatriton, that has maintained a primitive low number of and Thorius. Some Chiropterotriton (some indi­ trunk vertebrae. Elongation has involved pro­ viduals of chiropterus, priscus, dimidiatus) have nounced elongation of the vertebral centra, and variously developed and irregularly distributed, thus of the vertebrae. small processes. A number of presumably advanced All vertebrae of Leurognathus have ventro- members of the genus Bolitoglossa (lignicolor, 1966 EVOLUTION OF LUNGLESS SALAMANDERS 41 platydactyla, ftaviventris, striatula, mexicana, sal­ are well developed on the first three to five verte­ vinii) have large, well developed, projecting basa­ brae, and are apparent to vertebrae eight to eleven. pophyses. Other Bolitoglossa (borburata, cerroen­ The processes are well developed only on the first sis, occidentalis, rufescem, subpalmata) have one or two vertebrae of D. quadramaculatus, and occasional small processes, and the remaining are not apparent past the sixth vertebrae. No ptery­ species have none. gapophyses are present in D. wright/. Well devel­ oped processes that extend as spinous projections Neural arch structure beyond the zygapophyses are present on the first Neural crests, primitively present in plethodon­ four vertebrae of Phaeognathus, but the elements tids, are subject to intergeneric, serial, and onto­ are reduced posteriorly and are not apparent past genetic variation. A detailed discussion will be the ninth vertebra. Moderately developed ptery­ presented at a later date. gapophyses are present on the first vertebrae in Hyperapophyses are moderately to well devel­ Leurognathus, but they are very reduced by the oped on the trunk: vertebrae of all plethodontids, fourth and only slightly indicated on more posterior and two conditions are encountered: (1) hypera­ vertebrae. The functional significance of pteryga­ pophyses arise united and remain united on all, or pophyses appears to lie in providing additional almost all, trunk vertebrae, (2) hyperapophyses attachment sites for dorsal spinal muscles, which arise united or separated; if united, only on the raise the skulls of the three genera to open their anterior vertebrae, and separated posteriorly. Many mouths. kinds of variation are encountered, but Leurog­ nathus, Phaeognathus, Gyrinophilus, Pseudotriton Transverse processes ruber, Eurycea multiplicata, Ematina, Hydroman­ Diapophyses and parapophyses are long and well tes italicus, H. brunus, and H. genei have condition developed in Desmognathus, Leurognathus, and 1. The remainder fall in group 2, except for peculiar Phaeognathus, and extend far beyond the lateral conditions encountered in some species of Desmog­ zygapophysial margins. The posterolaterally di­ nathus (monticola, quadramaculatus) and in Pseu­ rected processes are relatively straight. The centra dotriton montanus in which the hyperapophyses are of mid-trunk: vertebrae are 0.7 to 0.9 times the divided on a few anterior vertebrae, but united on distance across the parapophysial tips ( centrum­ most vertebrae. In addition the hyperapophyses of parapophysial ratio). The parapophyses and dia­ Typhlomolge and Haideotriton are united for most pophyses are entirely free, or are joined proximally of their lengths, but are separated at their tips. only, in Desmognathus and Leurognathus, but are Hyperapophyses tend to be united in hynobiid joined by thin bony webbing for their entire salamanders, and united on most vertebrae in am­ lengths in Phaeognathus. Parapophyses are directly bystornatids. In ambystomatids a tendency for divi­ below or just slightly anterior to the diapophyses. sion and elongation is apparent on the posterior Alar expansions are usually present on the anterior vertebrae. It is apparent that united processes are margins of the parapophyses, especially on anterior more primitive than separated ones in salamanders and mid-trunk vertebrae. Alar processes are par­ in general, and probably in plethodontids as well. ticularly well developed with distinct anterior pro­ The genera of plethodontids in group 1 are rela­ jections in D. monticola, D. quadramaculatus, tively primitive, based on other characters. Leurognathus, and Phaeognathus. Well developed Desmognathus, Leurognathus, and Phaeogna­ dorsal processes are present on the diapophyses and thus differ from all other plethodontids in the pos­ parapophyses of D. monticola, and slightly smaller session of pterygapophyses on all or some trunk: processes are evident in D. quadramaculatus and vertebrae (Fig. 9). The posterior margins of the Leurognathus. The processes correspond with neural arches extend posterolaterally above and raised processes on both rib heads. beyond the postzygapophyses, and form wing-like Centrum-parapophysial ratios are about 0.8 to or spine-like processes. The processes are always 0.85 in Pseudotriton, Gyrinophilus, and Stereochi­ largest and longest on the first vertebrae, and lus, about 1 in Typhlomolge and Hemidactylium, gradually become reduced in size posteriorly. All and 1 to 1.4 in Eurycea, Manculus, Typhlo­ trunk: vertebrae of large D. auriculatus bear ptery­ triton, and Haideotriton. Parapophyses and dia­ gapophyses, but the processes become progressively pophyses are moderately long and extend beyond reduced in size posteriorly. Pterygapophyses are the zygapophyses in Gyrinophilus, Pseudotriton, present on about the first seven vertebrae in D. Stereochilus, Typhlomolge, and Hemidactylium, monticola, with at least rudiments on all trunk: but are very shortened and do not extend ( dia­ vertebrae. In most Desmognathus pterygapophyses pophyses) beyond the zygapophyses in Eurycea, 42 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4. Manculus, Typhlotriton, and Haideotriton. The joined to their tips on the third vertebra, and com­ diapophyses and parapophyses are fused proxi­ pletely fused to form cylindrical processes with mally but free distally in Pseudotriton, Gyrinophi­ single distal articular facets on succeeding verte­ lus, and Stereochilus, and are free for their entire brae. Apparently the parapophyses contribute most lengths in Eurycea, Manculus, Typhlotriton, Typh­ to the single processes, and the diapophyses are. lomolge, Haideotriton, and Hemidactylium. Para­ lost. pophyses are slightly anterior to the diapophyses in Centrum-parapophysial ratios are from 0.85 to Gyrinophilus, Pseudotriton, Stereochilus, and 1.2 in Batrachoseps, and the transverse processes Hemidactylium, but are greatly in advance of the are long, extending beyond the zygapophyses. The diapophyses in Eurycea, Manculus, Typhlotriton, processes are initially straight, but bend posteriorly Typhlomolge, and Haideotriton. In Gyrinophilus just past their midpoints in a manner similar to and allies the posterior margins of the parapophy­ Hydromantes. Transverse processes are separated ses are located posterior to the anterior margins of for most of their lengths, but are joined for short the diapophyses. In Eurycea and allies the dia­ distances proximally. As in Hydromantes the para­ pophyses are borne near the midpoints of centra, as pophyses are almost directly below the diapophyses. in other plethodontids, but the parapophyses arise Alar expansion in Batrachoseps and Hydromantes from the anterior one-third of the centra and are is virtually nonexistent. entirely in advance of the diapophyses. The ad­ Centrum-parapophysial ratios are from 0.85 to vanced position of the parapophyses and the very 1.25 in Bolitoglossa, Oedipina, Pseudoeurycea, and short transverse processes of Eurycea and allied Chiropterotriton, slightly more than 1 in Parvi­ genera are unique characters among the pletho­ molge and Thorius, and about 1.3 in Lineatriton. dontids. The Lineatriton ratio reflects elongated centra Centrum-parapophysial ratios of Plethodon and rather than shortened transverse processes. Trans­ Aneides are from 0.85 to 1.2, and those of Ensatina verse processes of all neotropical genera extend are about 0.8. Diapophyses and parapophyses are slightly to well beyond the zygapophyses. Trans­ moderately long and extend well beyond the zyga­ verse processes are completely separated or joined pophyses. Transverse processes are separated proxima11y only in Chiropterotriton, Parvimolge, completely in Ensatina, most Plethodon, and some Thorius, most Pseudoeurycea, and many Bolito­ individuals of Aneides; joined proximally but sep­ glossa. The processes are joined for most of their arated distally in some Plethodon (wehrlei, ciner­ lengths in some Pseudoeurycea ( altamontana, eus) and Aneides (aeneus, ferreus); and joined smithi) and many Bolitoglossa (colonnea, dunni, almost to their tips in some Plethodon ( elongatus, engelhardti, marmorea, morio, orestes, platydac­ neomexicanus, richmondi) and Aneides (flavipunc­ tyla, rostrata, savagei, subpalmata). The processes tatus, hardii, lugubris). Parapophyses are either are initially separated in Lineatriton, but are com­ directly below or just slightly anterior to the dia­ pletely joined past the third vertebra, and the pophyses at their origins. Transverse processes are diapophyses are lost by the fourth or fifth vertebrae; generally rather straight, and extend laterally with single processes are present on succeeding verte­ slight posterior angles (less than 30°) from the brae. A similar situation is encountered in Oedipina perpendicular with the vertebral axis. Alar expan­ in which single processes are present by the third or sions are generally absent, but relatively broad, fourth vertebrae. Lineatriton and Oedipina re­ shelf-like structures extend from the centra to semble the European species of Hydromantes in points near the tips of the parapophyses both an­ this regard, and the character is not found else­ teriorly and posteriorly in A. lugubris, and, to a where in the family. The parapophyses are borne lesser extent, in A. flavipunctatus. almost directly below, or slightly in advance of, Centrum-parapophysial ratios are 0.8 to 0.9 in the diapophyses in most neotropical species. The Hydromantes, and the transverse processes are parapophyses of some groups (some Chiroptero­ long, extending well beyond the zygapophyses. The triton and especially Parvimolge) may be distinctly transverse processes are initially straight, but may anterior to the origin of the diapophyses. Trans­ bend posteriorly past the midpoints of the centra. verse processes are usually initially straight, but The processes are almost entirely separated in H. are slightly to distinctly bent posteriorly past their platycephalus and H. shastae, but may be joined midpoints. The processes extend primarily laterally, for about one-half their lengths in H. brunus. The but are slightly to moderately swept posteriorly in transverse processes of the European Hydro­ most genera, and are very strongly swept posteriorly mantes (italicus, genei) are joined proximally or in Lineatriton and Oedipina. Exceptions are 0. are entirely separated on the first two vertebrae, complex and 0. parvipes in which the processes are 1966 EVOLUTION OF LUNGLESS SALAMANDERS 43 almost perpendicular to the vertebral axis. Slight Small Manculus and Eurycea, and all Haideotriton, anterior alar expansions are present on the para­ lack caudal transverse processes. pophyses of several Chiropterotriton (abscondens, H emidactylium has transverse processes on about multidentatus, nasalis). Very large alar expansions the first nineteen vertebrae, and all but the first that form moderately broad shelves are present in are midcentral. front of the parapophyses in Lineatriton, and much Ensatina has rather poorly developed processes smaller expansions are present posteriorly. Alar past the first vertebrae, and none occur past the expansions are evident on the anterior margins of tenth or eleventh vertebrae; all but the first few the parapophyses of several Oedipina (inusitata, are located midcentrally. Transverse processes are complex, bonitaensis, pacificensis, longissima), and somewhat better developed in Plethodon, and vir­ may be drawn into anteriorly directed, pointed tually all are midcentral. Midcentral processes are processes. Alar expansions are usually present but moderately developed and present on most caudal poorly developed in other Oedipina, and are poorly vertebrae of Aneides. developed or absent in other neotropical genera Cylindrical, midcentral transverse processes are than those discussed (see Tanner, 1950, for present on most caudal vertebrae of Hydromantes. illustrations) . Batrachoseps has rather complex, broad-based, strutted processes on most caudal vertebrae in Caudal vertebrae intermediate positions; some processes are at the anterior ends of the centra, while others are slightly Tails of plethodontid salamanders are of three posterior but ahead of midcentrum. Bolitoglossa, general types: Oedipina, Pseudoeurycea, Chiropterotriton, Par­ 1. Very stout proximally, tapering distally. Des­ vimolge, Lineatriton, and Thorius all have trans­ mognathus, Leurognathus, Phaeognathus, Gy­ verse processes on virtually all caudal vertebrae; rinophilus, Stereochilus, Typhlomolge, Haideo­ the processes are located at the anterior ends of triton. the centra and are swept moderately to strongly 2. Relatively slender proximally, tapering some­ anteriorly. what distally. Eurycea, Manculus, Typhlotri­ Unusual neural crests are found in Oedipina; ton, Plethodon, Aneides, Hydromantes. usually only the first one or two caudal vertebrae (but as many as ten) have single medial crests. 3. Constricted proximally, expanding beyond con­ Commencing on about the second vertebra the striction, but tapering toward distal tips. H emi­ crests divide and form double, parallel crests on dactylium, Ensatina, Bolitoglossa, Oedipina, either side of the midline. The crests are broadest Pseudoeurycea, Chiropterotriton, Parvimolge, basally and angle toward one another dorsally. Lineatriton, Thorius. Double caudal crests are constant, unique charac­ ters of Oedipina. Batrachoseps has slight indications of tail con­ striction, and rather bridges the gap between types Ribs 2 and 3. Some Eurycea (e.g., E. longicauda) bridge Ribs vary considerably in length in the family, types 1 and 2. and also vary serially within individuals. Mid-trunk First caudal vertebrae of Ensatina and H emi­ ribs are from 0.6 to 0.75 times the transparapo­ dactylium, and to a lesser extent, the neotropical physial distance (rib-parapophysial ratio) in Des­ genera, are highly specialized structures of impor­ mognathus and Leurognathus, but are relatively tance in tail autotomy (see also Wake, 1963.) short in Phaeognathus (0.4). Ratios are from 0.65 Detailed discussion is deferred to a later date. to 0.8 in Gyrinophilus, Pseudotriton, Stereochilus, Midcentrally located transverse processes are and Hemidactylium, about 0.9 in Typhlomolge, borne on virtually all caudal vertebrae of Desmog­ about 1 to 1.2 in Manculus, Typhlotriton, and nathus, Leurognathus, and Phaeognathus, and the Haideotriton, and from a little over 1 to about 1.5 processes are relatively stout, particularly in in Eurycea. Ensatina has relatively short ribs (0.6), Leurognathus. but ribs are longer in Plethodon (0.67 to about 1) Transverse processes at midcentrum are mod­ and longer still in Aneides (0.85 to 1.25). Hydro­ erately developed and present to about the tenth mantes and Batrachoseps have moderately short to eleventh caudal vertebrae of Gyrinophilus, ribs with ratios of about 0.6 to 0.75. A distinct Pseudotriton, Manculus, and Typhlomolge, to trend toward rib shortening is encountered in the about the sixteenth of Stereochilus, and almost to neotropical genera. The trend is best exemplified the end of the tails of Eurycea and Typhlotriton. by the genus Chiropterotriton in which relatively 44 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 primitive species have moderately long ribs (priscus, reversed. Limb length apparently increases as the ratio 1 ) , but more specialized species have short species of Eurycea become increasingly troglobitic ribs (abscondens, ratio 0.35-0.4). Rib-parapo­ (see Mitchell and Reddell, 1965). The relatively physial ratios are about 0.75 to 0.8 in Thorius, 0.4 longest limbs in the family are encountered in the to 0.75 in Pseudoeurycea (lowest figure in alta­ e~tremely specialized troglobite, Typhlomolge. The montana, highest in werleri), 0.4 to 0.67 in Boli­ elongated limbs of troglobitic species may be adap­ toglossa, (lowest figure in adspersa, highest in tations which enable the salamanders to walk about engelhardti), 0.9 in Parvimolge townsendi, and quiet cave pools with their bodies elevated above 0.5 in P. richardi. Ribs are especially shortened in the bottom debris. Oedipina (0.2-0.3) and Lineatriton (0.2-0.25). Distinct, well developed tibial spurs are present Members of the latter two genera are elongate, in most genera. The spurs are truncate and reduced attenuate forms, and the shortened and reduced in Typhlotriton, and are small and very slender in ribs are probably related to the change in habitus. Typhlomolge and Haideotriton. Ensatina lacks Rib heads are of moderate length and are well tibial spurs, but well defined dorsal ridges are separated in most genera. The heads are greatly present. Very small tibial spurs, borne on the mid­ separated and rather long in Eurycea, Manculus, sections of the bones and falling far short of the Typhlotriton, Typhlomolge, and Haideotriton. Rib proximal ends, are found in Hydromantes. The heads are rather short in Batrachoseps. Fusion of spurs are present in most species of Batrachoseps, the rib heads occurs in Aneides lugubris, Hydro­ but are absent in most populations of B. attenuatus. mantes brunus, Phaeognathus, some Batrachoseps A trend toward spur reduction is apparent in neo­ attenuatus, Pseudoeurycea altamontana, Bolito­ tropical genera, reduced spurs being present in glossa orestes, and B. subpalmata. Usually the last Lineatriton, Parvimolge townsendi, most Pseudo­ one or two ribs of all species are unicipital. In eurycea, Chiropterotriton, and Thorius, and in Hydromantes genei, H. italicus, and in all species of some Bolitoglossa (dunni, engelhardti, helmrichi, Oedipina only the first three or four pairs of ribs rufescens). Spurs are absent in all Oedipina, most are bicipital, and the remainder are unicipital. All Bolitoglossa, some Chiropterotriton (bromeliacia, but the first four or five pairs of ribs are unicipital nasalis, xolocalcae ), and some individuals of various in Lineatriton. species of Thorius and Pseudoeurycea.

Limb Elements Mesopodial Elements Primitively plethodontid limb bones have car­ Primitively eight carpals and nine tarsals are tilaginous heads, but calcifications occur in all present (Desmognathus, Leurognathus, Phaeogna­ Thorius, Lineatriton, Parvimolge, and some indi­ thus, Gyrinophilus, Pseudotriton, Stereochilus, viduals of Oedipina. Apparently these calcifications Eurycea, Typhlotriton, Haideotriton, Typhlomolge, are strengthening compensations for paedomorphic Plethodon, Ensatina, Hydromantes, Pseudoeury­ weakening. cea). All have primitive carpi with ulnars and Relative limb length is highly variable, and centrals separated except Hydromantes and Pseu­ description is made more complex by allometric doeurycea in which the elements are in contact. growth and ontogenetic, sexual, geographic, inter­ All of the above genera have primitive tarsi with specific, and intergeneric variation. Certain of the fourth distal tarsals much larger than the fifth, attenuate genera (Phaeognathus, Stereochilus, which are the only tarsals that do not touch the Manculus, Batrachoseps, Oedipina, Lineatriton, centrals. Parvimolge, Thorius) have obviously shortened Aneides differs from the above genera in that limb elements. Long limbs tend to be the end result the ulnars articnlate eXtensively with the centrals, of trends toward increased scansorial activity in and the fifth distals are the largest tarsals and articu­ several forms (Eurycea lucifuga, Aneides lugubris, late broadly with the centrals. The intermedia and A. aeneus, A. ferreus). Moderately long limbs are ulnars are fused in all carpi of A. hardii, and seven present in Ensatina, Hydromantes, and primitive carpal elements are present. species of Plethodon and Bolitoglossa. Hind limbs The remaining genera have reduced numbers of are much longer and stouter than the fore limbs in carpals, tarsals, or both, in some or all species. Desmognathus, Leurognathus, and Phaeognathus. Batrachoseps, Manculus, and Hemidactylium have Fore limbs are a little shorter than the hind limbs lost their fifth toes, and have also lost the fifth in other genera, but the limbs are of about equal distal tarsals. Manculus, Hemidactylium, Batra­ length in Ensatina. Femurs are always longer than choseps wrighti, and some populations of B. pacifi­ humeri except in Ensatina where the situation is cus have eight carpals, but other Batrachoseps have 1966 EVOLUTION OF LUNGLESS SALAMANDERS 45 seven, the result of fusion of ulnars and intermedia. central fusions result in six carpals in Thorius. Tarsal patterns similar to those in Aneides are Distal tarsals 4 and 5, the intermedia and fibulars, encountered in several Chiropterotriton (abscon­ and distal tarsals 4 and 5 and the centrals may fuse, dens, arboreus, chiropterus, dimidiatus, multiden­ and six to eight tarsals result. Intraspecific varia­ tatus, priscus), but in two advanced species (xolo­ tion is present in Thorius, and individuals may vary calcae, bromeliacia) the primitive plethodontid by one element from one side to the other. pattern is found. Another advanced species, C. Eight tarsal elements are present in most Bolito­ nasalis, has seven carpals and eight tarsals as a glossa as a result of fusion of distal tarsals 4 and 5. result of radial-intermedial and fourth and fifth A tendency for elimination of the third distal car­ tarsal fusions. C. abscondens may have seven or pals from their articulation with the centrals, and eight mesopodial elements as the result of ulnar­ subsequent reduction of carpal size, is apparent intermedial or third and fourth carpal, and fourth within the genus, especially in the group of species and fifth tarsal fusions. All ankles of three indi­ associated with B. dunni and B. engelhardti. Distal viduals have seven tarsals; wrists usually have carpals 1 and 2 meet distal carpals 4 and distal seven carpals, but one wrist each of two individuals carpals 3 are effectively walled-off from articulation has eight. with the centrals. The trend culminates in B. occi­ The tarsal patterns unique to Aneides and the dentalis and B. rufescens in which distal carpals 3 primitive species of Chiropterotriton apparently and 4 fuse to form single, strengthened, more have functional and evolutionary significance. In efficient elements, and seven carpals result. A simi­ the primitive plethodontid tarsus the lines of force lar situation is encountered in the tarsi of B. occi­ from the first, second, third, and fourth digits are dentalis and B. rufescens, in which distal tarsals 3, directed through the first three distal tarsal elements 4, and 5 fuse, and seven tarsal elements are present. to the large centrals. Lines of force from the fifth This condition is also encountered in one of three digits are directed through the relatively small B. platydactyla examined. fourth distal elements (distal tarsal 5) to the fibu­ The most extreme mesopodial specializations lars. The relative weakness of this arrangement is occur in Oedipina. The limbs, hands and feet are illustrated by parallel loss of the fifth distal tarsals greatly reduced in size, and are considerably less and digits in three genera. The arrangement in functional than in primitive members of the family. Aneides and Chiropterotriton is one in which all From five to seven carpal and from five to eight lines of force are relatively straight, and all are tarsal elements are present as the result of the directed to a single pivotal point in each tarsus, the following fusions (arranged in probable fusion centrale (see illustrations in Wake, 1963). Since sequence) : ulnar-intermedial, distal carpal 4- the lines of force emanate from single points, they central, distal carpals 1, 2, and 3; distal tarsals 4 may be swung or rotated about these points. Rela­ and 5 (fused in all species) , distal tarsals 4, 5, and tively greater spread across the outer digital tips central, fibular-intermedial. Species of Oedipina than in the primitive arrangement is thus possible; vary as follows in the number of mesopodial the increased spread has obvious advantages for fusions: bonitaensis 6 carpals, 7 tarsals; complex organisms with climbing propensities. Both Aneides 6, 6; cyclocauda, uniformis 1, 7-8; gracilis 5, 5-7; and Chiropterotriton contain species that are adept inusitata 6, 6; longissima 6, 7; pacificensis 6, 7; climbers. Apparently their mesopodial arrangement poelzi 1-8, 7-8; syndactyla 6, 6; sp. nov. 7, 7-8. is a prospective adaptation that has significant In Oedipina in which the third distal tarsals are adaptive value in arboreal and scansorial but not not fused with some other elements, they are very terrestrial organisms. small and resemble the condition in Thorius. Fur­ Parvimolge townsendi has six or seven carpal ther similarities of the two genera are the reduction elements as a result of ulnar-intermedial and pos­ in size of distal carpals and tarsals 1 and 3, and sible fourth distal carpal-central fusions. Distal shifts of the articulation of the first metacarpals tarsals 4 and 5 have fused, and eight tarsal elements and metatarsals from distal elements to centrals 1. are present. The latter condition is also found in some special­ A peculiar and unique situation is encountered ized Chiropterotriton (e.g., abscondens). in Lineatriton. Eight tarsal elements are p.resent Mesopodial calcifications of probably functional and distal tarsal 5 has fused with the foUrth or and phylogenetic significance occur in certain has been lost; metatarsal 5 extends into the area plethodontids. Cope (1869) stated that Thorius seemingly vacated by the tatsal. Eight carpals are has ossified mesopodial elements, and differs in present. this character from other plethodontids. Rabb Ulnar-intermedial and fourth distal carpal- (1955) reported calcified mesopodials in Parvi- 46 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 molge townsendi. Uzzell (1961) found mesopodial seven tarsals (fibulare, intermedium, centrale, cen­ calcifications in only twelve of thirty-four adult trale 1, distal tarsals 1-2, 3, 4-5). I have ~ot seen Thorius, thirteen of seventeen Parvimolge town­ calcified mesopodials in Lineatriton. sendi, and three of thirteen Lineatriton. He found Calcifications occur in several additional genera. eight tarsals calcified in twelve ankles of seven in­ Mesopodial calcifications observed include: fibu­ dividuals of Thorius, with lesser numbers in other lars of one Desmognathus quadramaculatus; fibu­ species. Uzzell suggested the following calcification lars and tarsal intermedia of two Desmognathus sequences: (1) carpus. intermedium-ulnare; radiate, ocoee (three ankles) ; carpal and tarsal intermedia centrale 1, or carpals 1 and 2; carpal 3; (2) tar­ and fibulars of a single Leurognathus; tarsal inter­ sus. outer-most tarsal (either tarsals 4-5, or tarsals media, fibulars, distal tarsals 5, one radiate, and 4-5-centrale); intermedium; fibulare, or centrale; one distal tarsal 1-2 of one Eurycea bislineata; tibiale, or tarsals 1-2; tarsal 3, or centrale 1. Uzzell tarsal intermedia and fibulars of one Typhlotriton; used slightly different terminology and calls cen­ tarsal and carpal intermedia, ulnars, fibulars, tar­ trals 1 of the carpi and tarsi the first carpals and sal centrals 1, distal carpals 1-2, 3, 4, and distal tarsals, respectively. tarsals 1-2, 3, 4, 5 of a single Ensatina; distal Sequence of calcification in Parvimolge and carpals 1-2 of a single Plethodon neomexicanus; Lineatriton was not determinable, but Uzzell distal carpals 1-2 and one distal tarsal 1-2 of a reported six elements with as many as eight centers single Pseudoeurycea unguidentis; one ulnare, fibu­ of calcification in Parvimolge carpi, and eight tar­ lars, one tibiale, distal carpals 4, one distal tarsal sal elements with as many as six calcifications. 1-2, and distal tarsals 4-5 of one Chiropterotriton Lineatriton was reported to have from one to five abscondens; and ulnar-intermedia, fibular-inter­ carpal and from one to three tarsal calcifications. media, tibials, carpal centrals, centrals 1, radials, Uzzell (1961) suggested that fusion of the inter­ distal carpals, 1-2, distal tarsals 4-5, and one distal media and ulnars, and close approximation of the tarsal 1-2 of a single Oedipina sp. nov. intermedia and fibulars reflect concentration of It is apparent that mesopodial calcifications have locomotor forces. In both wrists and ankles the appeared independently in several evolutionary outer or postaxial mesopodials fuse before the inner lines, but calcification is certainly more common or preaxial elements, and it is likely that these are in Parvimolge townsendi and Thorius, and prob­ the elements of greatest importance in force trans­ ably in Lineatriton, than in the other plethodontid mission. Mesopodial fusions are far more common genera. Postaxial calcifications predominate, but in the preaxial than in the postaxial portions of the preaxial calcifications appear first in some indi­ carpi and tarsi throughout the family. However, viduals. The most serious objections to Uzzell's fusions (e.g., distal elements 1 and 2) are present (1961) suggestions concerning the functional sig­ in the preaxial portions of all plethodontids, and nificance of mesopodial fusions and calcifications using Uzzell's reasoning for postaxial regions, the are the preaxial distal elements 1 and 2 fusions preaxial area might be the most significant primi­ universal in the family, and the tendency for loss tively with the postaxial portion gaining importance of the fifth toes and fifth distal tarsals in the post­ secondarily. axial regions in several evolutionary lines. Rabb (1955) interpreted several characters of Parvimolge townsendi to be the result of paedo­ Digits morphosis, and Uzzell (1961) suggested that calci­ Primitive plethodontids have four fingers and fied mesopodials may compensate for the paedo­ five toes, and the primitive phalangeal fomulae are morphic and presumably weak condition of the 1-2-3-2 for each hand and 1-2-3-3-2 for each foot. adult feet of Lineatriton, Parvimolge, and Thorius. Such formulae are characteristic of all Desmogna­ Uzzell also suggested that calcified mesopodials of thus, Leurognathus, Phaeognathus, Gyrinophilus, Parvimolge and Lineatriton reflect possible close Pseudotriton, Stereochilus, Eurycea, Typhlotriton, relationship, but thought the conditions in Thorius Haideotriton, Typhlomolge, Ensatina, Aneides, might be due to parallelism. Hydromantes, Pseudoeurycea, Chiropterotriton, The only material available to me of Parvimolge and Parvimolge. Hemidactylium, Manculus, and townsendi has been that used by Uzzell. I am unable Batrachoseps have lost the fifth toes and one pha­ to add to his data. Some, all, or none of the carpals lanx from each fourth toe; the foot formula is and tarsals of Thorius may be calcified. The maxi­ 1-2-3-2 (rarely 0-2-3-2 in B. attenuatus). Single mum numbers of calcifications observed by me in phalanges have been lost from the fifth toes of most Parvimolge are five carpals (ulnare-intermedium, individuals of two species of Plethodon (larselli, centrale, centrale l, distal carpal 4, radiale) and neomexicanus) and the formula is 1-2-3-3-1; all 1966 EVOLUTION OF LUNGLESS SALAMANDERS 47 other Plethodon have the primitive formula. Some mouth opening mechanism by means of which the Lineatriton have the primitive formula, but reduc­ mandibles are held relatively rigid and the skulls tions may occur to 1-2-3-1 in the hand and proper are raised, extensive skeletal and muscular 1-2-3-2-1 in the foot. Thorius may have the primi­ modifications related to functional changes in tive formulae, but intra- and interspecific variation mouth opening mechanics, and pterygapophyses occurs and formulae variations are: hand (0-1)-2- and well developed hypapophysial keels on ante­ 3-(1-2), foot (0-1)-2-3-(2-3)-(1-2). Most species rior trunk vertebrae. of Bolitoglossa retain the primitive formulae, but Characterization: ( 1 ) skulls rigid, depressed, B. occidentalis has lost one phalanx from each streamlined, composed of dense, heavy bones, fourth toe (foot 1-2-3-2-2), and formulae varia­ (2) larvae lack protruding gill rami, ( 3) larval tions in B. rufescens are: hand (0-1)-(1-2)-(2-3)­ skulls composed of ossified premaxillae, frontals, (1-2), foot 1-(1-2)-(2-3)-(2-3)-(1-2). Extensive parietals, occipito-otics, opercula, vomers, pala­ phalangeal reduction occurs in Oedipina, in which topterygoids, parasphenoids, quadrates, squamo­ only certain 0. poelzi and 0. sp. nov. retain the sals, dentaries, prearticulars, and coronoids, ( 4) primitive formulae; formulae variations are: hand four larval epibranchials, ( 5) larval vomers in (0-1)-(1-2)-(2-3)-(1-2), foot (0-1)-2-(2-3)-(2- broad median contact, ( 6) larval palatopterygoids 3)-(1-2). The commonest reductions occur in well separated from quadrates, (7) unicuspid larval digits one, four, and five of the feet, and digits one teeth, bicuspid adult teeth, ( 8) single larval and and three of the hands. adult premaxilla, (9) premaxillae very broad and A tendency for reduction of phalangeal elements depressed, internasal fontanelles small or closed and loss of digits in the postaxial portions of the over, ( 10) frontal processes of premaxillae stout hands and feet is evident. Reduction is most com­ and broad, arise broadly from toothed portions, monly encountered in attenuate, short-limbed ( 11) maxillae long, extend beyond eyeballs, small species. Uzzell ( 1961) has suggested that the post­ to large posteromedially oriented jugal processes axial portions of the carpi and tarsi are the areas present, ( 12) maiillary facial lobes very large, of greatest transmission of forces from the lower articulate extensively with frontals, ( 13) palatal limbs to the digits, but postaxial phalangeal and shelves of premaxillae and maxillae very broad, digital reductions suggest that other areas may be ( 14} septomaxillae present, but reduced, ( 15) as Important. nasals small, articulate firmly with maxillae and premaxillae, ( 16) prefrontals absent, ( 17) vom­ erine preorbital processes toothless, ( 18) anterior SYSTEMATIC SUMMARY vomerine teeth present or absent, (19) anterior Family Plethodontidae vomerine teeth discontinuous, (20) posterior vom­ Definition: adults lungless with nasolabial grooves, erine teeth borne in dense, separated patches, (21) no bony pterygoids, no lacrimals, otic and occipital anterior portions of frontals very large, expanded, elements fused, columellae fused with opercula or occupy prefrontal and antorbital regions, ( 22) absent, vomers with anterior laterally oriented frontals flat dorsally with sharp angular ridges tooth series and posterior multiple series borne along orbital borders, bones uncovered by muscu­ below parasphenoids, no angulars; larvae present lature, (23) deep parietal-otic troughs extend as or absent, if present with three or four epibran­ bony processes into orbits, (24) well developed chials, fused premaxillae, teeth on premaxillae, crests border posterolateral margins of parietal­ vomers, palatopterygoids, dentaries, and coronoids. otic troughs, (25) large posterolateral parasphe­ Range: eastern and western North America with noid processes, (26) parasphenoids relatively short, some species in central regions; Middle America abut against posterior margins of vomers in simple from Tamaulipas and Nuevo Leon in northeastern articulations, (27) quadrates very large, articulate Mexico and Nayarit in western Mexico to eastern with parasphenoids and occipito-otics, cartilagin­ Brazil and central Bolivia; the Maritime Alpes of ous portions of suspensoria relatively small, ( 28) southeastern France, the Piedmont of Italy, and orbitosphenoids large, well developed, quadrangu­ the island of Sardinia. lar, (29) squamosals large, solidly anchored to Content: two subfamilies, twenty-three genera, skulls, ( 30) occipital condyles stalked, ( 31 ) colu­ about one hundred eighty-four species. mellae short, stocky, rather flattened, (32) denti­ tion well developed, ( 3 3) meckelian grooves Subfamily Desmognathinae essentially closed, dentaries 0-shaped anteriorly, Definition: plethodontid salamanders having four ( 34) detoglossal; genioglossal muscles present, rather than three larval epibranchials, a unique (35) cornua elongate, attaching to first basibran- 48 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 chials behind anterior extensions of latter, (36) maxillae joined by septum-like processes which lingual cartilages absent, (37) well developed first form fontanelle borders; hypapophysial keels well basibranchial anterior extensions, (38) no marked developed on anterior trunk vertebrae, not present anterior first basibranchial expansion, (39) first or only slightly indicated on posterior vertebrae; ceratobranchials longest of articulated hyobran­ neural crests present only on first two or three chial elements, ( 40) first ceratobranchials consid­ trunk vertebrae; parapophyses and diapophyses erably larger than second, ( 41 ) epibranchials less separated for most of lengths; ribs more than one­ than 1.25 times first basibranchials, and shorter half the distance across parapophyses; stout, stocky than first ceratobranchials, ( 42) second cerato­ habitus; limbs moderately long; aquatic larvae, or branchials short, less than two-thirds length of first direct development. basibranchials, ( 43) atlases large and stout, bear Range: eastern North America, particularly very large, markedly concave, condylar articular in Appalachian Mountain region. facets, ( 44) odontoid processes of atlases very Content: aeneus Brown and Bishop, 1947; auri­ small, ( 45) posterior margins of atlases very large, culatus (Holbrook, 1838); fuscus (Rafinesque, broad, markedly raised, (46) fifteen or twenty-one 1820); monticola Dunn, 1916; ochrophaeus Cope, to twenty-three trunk vertebrae, (47) two caudo­ 1859; ocoee Nicholls, 1949; quadramaculatus sacral vertebrae, ( 48) larval vertebrae structurally (Holbrook, 1840); wrighti King, 1936. amphicoelous, intervertebral cartilages become calcified in adults and resemble classical opistho­ Leurognathus Moore, 1899 coelous condition, (49) hypapophysial keels well Definition: desmognathine salamanders with developed on at least the first few trunk vertebrae, fifteen trunk vertebrae; small maxillary jugal proc­ or on all trunk vertebrae, (50) hyperapophyses esses; internal nares open laterally near lateral well developed, tend to be united, but may be margins of vomers; vomerine preorbital processes divided on anterior vertebrae, ( 51 ) well developed, slightly separated from vomers proper, extend stout basapophyses on posterior central margins, laterally beyond vomer bodies; only in young an (52) pterygapophyses present on some anterior internasal fontanelle which becomes progressively trunk vertebrae, ( 5 3) parapophyses and diapo­ closed during ontogeny; skulls strongly depressed, physes long, separated or fused, (54) centra shorter premaxillae and vomers in close proximity with than distance across tips of parapophyses, (55) no septum-like connecting processes; hypapophysial ribs usually absent on last trunk vertebrae, (56) keels strongly developed on all trunk vertebrae, ribs less than 0.75 times distance across parapo­ particularly anteriorly; neural crests present only physes, (57) tails stout proximally, no basal con­ on first two or three trunk vertebrae; parapophyses striction, (58) well developed tibial spurs, (59) and diapophyses separated for most of lengths; hind limbs much larger than fore limbs, (60) ribs more than one-half the distance across the fingers and toes long and slender with distinctive parapophyses; stout, stocky habitus; limbs moder­ tendon arrangements and fleshy palms, ( 61) eight ately long; aquatic larvae and adults. carpals and nine tarsals, ( 62) distal tarsals 5 do Range: southern Appalachian Mountains of not articulate with central tarsals, ( 63) four fingers eastern United States. and five toes, ( 64) phalangeal formulae 1-2-3-2, Content: marmoratus Moore, 1899. 1-2-3-3-2, (65) adults aquatic to terrestrial, (66) aquatic larvae in most, but trends toward direct Phaeognathus Highton, 1961 terrestrial development, ( 67) pigmented, lobed Definition: desmognathine salamanders with testes in old males, ( 68) small gularis muscles twenty-one to twenty-three trunk vertebrae; very present posterior and external to very large quad­ large, expanded maxillary jugal processes that rato-pectoralis muscles. extend almost to quadrates; internal nares open Content: three genera-Desmognathus, Leurog­ medially; vomerine preorbital processes well sep­ nathus, Phaeognathus. arated from vomers proper, extend laterally beyond vomer bodies; an internasal fontanelle; vomers and Desmognathus Baird, 1850 premaxillae joined by septum-like processes which Definition: desmognathine salamanders with fif­ form fontanelle borders; anterior cranial elements teen trunk vertebrae; small maxillary jugal proc­ sculptured; hypapophysial keels present on all esses; .internal nares open medially; vomerine trunk vertebrae, best developed anteriorly; neural preorbital processes well separated from vomers crests present on all trunk vertebrae; parapophyses proper, and not extending laterally beyond vomer and diapophyses joined for virtually entire lengths; bodies; an internasal fontanelle; vomers and pre- ribs less than one-half the distance across the 48 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 chials behind anterior extensions of latter, (36) maxillae joined by septum-like processes which lingual cartilages absent, (37) well developed first form fontanelle borders; hypapophysial keels well basibranchial anterior extensions, (38) no marked developed on anterior trunk vertebrae, not present anterior first basibranchial expansion, (39) first or only slightly indicated on posterior vertebrae; ceratobranchials longest of articulated hyobran­ neural crests present only on first two or three chial elements, ( 40) first ceratobranchials consid­ trunk vertebrae; parapophyses and diapophyses erably larger than second, ( 41 ) epibranchials less separated for most of lengths; ribs more than one­ than 1.25 times first basibranchials, and shorter half the distance across parapophyses; stout, stocky than first ceratobranchials, ( 42) second cerato­ habitus; limbs moderately long; aquatic larvae, or brancbials short, less than two-thirds length of first direct development. basibranchials, ( 43) atlases large and stout, bear Range: eastern North America, particularly very large, markedly concave, condylar articular in Appalachian Mountain region. facets, ( 44) odontoid processes of atlases very Content: aeneus Brown and Bishop, 1947; auri­ small, (45) posterior margins of atlases very large, culatus (Holbrook, 1838); fuscus (Rafinesque, broad, markedly raised, ( 46) fifteen or twenty-one 1820); monticola Dunn, 1916; ochrophaeus Cope, to twenty-three trunk vertebrae, (47) two caudo­ 1859; ocoee Nicholls, 1949; quadramaculatus sacral vertebrae, ( 48) larval vertebrae structurally (Holbrook, 1840); wrighti King, 1936. amphicoelous, intervertebral cartilages become calcified in adults and resemble classical opistho­ Leurognathus Moore, 1899 coelous condition, (49) hypapophysial keels well Definition: desmognathine salamanders with developed on at least the first few trunk vertebrae, fifteen trunk vertebrae; small maxillary jugal proc­ or on all trunk vertebrae, (50) hyperapophyses esses; internal nares open laterally near lateral well developed, tend to be united, but may be margins of vomers; vomerine preorbital processes divided on anterior vertebrae, {51 ) well developed, slightly separated from vomers proper, extend stout basapophyses on posterior central margins, laterally beyond vomer bodies; only in young an (52) pterygapophyses present on some anterior internasal fontanelle which becomes progressively trunk vertebrae, ( 53) parapophyses and diapo­ closed during ontogeny; skulls strongly depressed, physes long, separated or fused, (54) centra shorter premaxillae and vomers in close proximity with than distance across tips of parapophyses, (55) no septum-like connecting processes; hypapophysial ribs usually absent on last trunk vertebrae, (56) keels strongly developed on all trunk vertebrae, ribs less than 0.75 times distance across parapo­ particularly anteriorly; neural crests present only physes, (57) tails stout proximally, no basal con­ on first two or three trunk vertebrae; parapophyses striction, (58) well developed tibial spurs, (59) and diapophyses separated for most of lengths; hind limbs much larger than fore limbs, (60) ribs more than one-half the distance across the fingers and toes long and slender with distinctive parapophyses; stout, stocky habitus; limbs moder­ tendon arrangements and fleshy palms, ( 61) eight ately long; aquatic larvae and adults. carpals and nine tarsals, ( 62) distal tarsals 5 do Range: southern Appalachian Mountains of not articulate with central tarsals, ( 63) four fingers eastern United States. and five toes, (64) phalangeal formulae 1-2-3-2, Content: marmoratus Moore, 1899. 1-2-3-3-2, (65) adults aquatic to terrestrial, (66) aquatic larvae in most, but trends toward direct Phaeognathus Highton, 1961 terrestrial development, (67) pigmented, lobed Definition: desmognathine salamanders with testes in old males, ( 68) small gularis muscles twenty-one to twenty-three trunk vertebrae; very present posterior and external to very large quad­ large, expanded maxillary jugal processes that rato-pectoralis muscles. extend almost to quadrates; internal nares open Content: three genera-Desmognathus, Leurog­ medially; vomerine preorbital processes well sep­ nathus, Phaeognathus. arated from vomers proper, extend laterally beyond vomer bodies; an intemasal fontanelle; vomers and Desmognathus Baird, 1850 premaxillae joined by septum-like processes which Definition: desmognathine salamanders with fif­ form fontanelle borders; anterior cranial elements teen trunk vertebrae; small maxillary jugal proc­ sculptured; hypapophysial keels present on all esses; -internal nares open medially; vomerine trunk vertebrae, best developed anteriorly; neural preorbital processes well separated from vomers crests present on all trunk vertebrae; parapophyses proper, and not extending laterally beyond vomer and diapophyses joined for virtually entire lengths; bodies; an intemasal fontanelle; vomers and pre- ribs less than one-half the distance across the 1966 EVOLUTION OF LUNGLESS SALAMANDERS 49 parapophyses; stout but elongated habitus with group, discontinuous in adults of all others, (20) extremely long trunks and long, attenuated tails; posterior vomerine teeth borne sparsely or densely limbs and digits markedly shortened; life history on narrow to broad, continuous or discontinuous unknown, but adults are terrestrial burrowers. posterior vomerine projections, (21) anterior por­ Range: known only from the pine woods of the tions of frontals moderately large and expanded, Coastal Plain in south-central Alabama. do not occupy antorbital regions and rarely occupy Content: hubrichti Highton, 1961. prefrontal regions, (22) frontals moderately rounded with no angular orbital ridges, frontals Subfamily Plethodontinae slightly to extensively covered by adductor mandi­ Definition: plethodontid salamanders having bulae anterior musculature, (23) no parietal-otic less than four larval or embryonic epibranchials, troughs, but many have posterior parietal depres­ normal salamander mouth opening mechanisms in sions, (24) well developed squamosal-otic and which the skulls proper remain relatively rigid and parietal-otic crests in some genera, others have the mandibles are lowered, and trunk vertebrae various types of otic crests, or no crests, (25) which lack pterygapophyses and well developed posterolateral parasphenoid processes moderate to hypapophyses. small, or absent, (26) parasphenoids relatively Characterization: ( 1) skulls variable in shape long, broadly overlap vomers, (27) quadrates and rigidness, usually relatively broad, moderately moderate to small, do not articulate with either high, and not strongly constructed; distinct tenden­ occipito-otics or parasphenoids, cartilaginous por­ cies in many groups for skull reduction and weak­ tions of suspensoria moderately large, (28) orbito­ ening, (2) larvae with protruding gill rami, (3) sphenoids large, moderate, or small, posteriorly larval skulls comprised of ossified premaxillae, constricted in some genera, (29) squamosals large frontals, parietals, occipito-otics, opercula, vomers, to small, anchorage to skulls variable, (30) occipi­ palatopterygoids, parasphenoids, quadrates, squa­ tal condyles sessile, ( 31 ) columellae present except mosals, dentaries, prearticulars, and coronoids, in some neotropical genera, usually moderately ( 4) three larval epibranchials, ( 5) larval vomers long and cylindrical, (32) dentition poorly to well in narrow anteromedial contact, well separated developed, (33) meckelian grooves open, dentaries posteriorly, (6) larval palatopterygoids with long, C-shaped anteriorly, ( 34) detoglossal or adeto­ posteriorly oriented processes that approach or glossal; genioglossal muscles present or absent, articulate with the quadrates, (7) unicuspid larval (35) comua broad and long to short and attenuate, teeth, bicuspid or specialized unicuspid adult attaching to or continuous with first basibranchials, teeth, (8) a single larval premaxilla, adults with ( 36) lingual cartilages present or absent, ( 37) ante­ single or paired elements, (9) premaxillae highly rior first basibranchial extensions present only in variable in shape, broad to narrow with a tendency some groups which lack lingual cartilages, (38) for depression; intemasal fontanelles usually large, first basibranchials anteriorly expanded or not, tend to close during ontogeny of a few species, (39) first ceratobranchials or epibranchials longest ( 10) premaxillary frontal processes highly vari­ of articulated hyobranchial elements, ( 40) first able, usually arise narrowly from toothed portions, ceratobranchials larger than second in most genera, ( 11 ) maxillae of variable lengths but usually fall smaller in one generic group, ( 41 ) epibranchials short of posterior eyeball margins, no jugal proc­ more than 1.5 times first basibranchials; epibran­ esses, (12) maxillary facial lobes moderate to chials shorter than first ceratobranchials in two small, frontal-maxilla articulation normally absent, genera, but longer in others, (42) second basi­ ( 13) palatal shelves moderately broad in some branchials present or absent, width more or less species, narrow in most, very small in some, (14) than first basibranchial length, ( 43) moderate­ septomaxillae primitively large, tend to be reduced sized atlases bear large condylar articular facets or lost in one generic group, ( 15) nasals small to that are relatively flat to slightly concave, ( 44) large, articulations highly variable from genus to large atlantal odontoid processes, ( 45) posterior genus, but articulation with maxillae usually neither atlantal margins relatively unmodified, ( 46) thir­ firm nor extensive, (16) prefrontals primitively teen to twenty-four trunk vertebrae, (47) two or present, lost in a few genera, (17) vomerine pre­ three caudosacral vertebrae, ( 48) vertebrae mostly orbital processes present in most, absent or reduced amphicoelous, but parallel tendencies toward a in a few; teeth may or may not extend onto proc­ secondary opisthocoely, (49) hypapophysial keels esses, but do so primitively, (18) anterior vomer­ usually absent, rudiments present on one or two ine teeth present, ( 19) anterior and posterior vom­ vertebrae in a few species, ( 50) hyperapophyses erine teeth continuous in primitive members of one usually well developed and divided, but united in 50 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 some species, ( 51) basapophyses present or absent, (7) anterior portions of parasphenoids strongly when present poorly to moderately developed, expanded, (8) parietals depressed posteriorly, with (52) no pterygapophyses, (53) parapophyses and distinct lateral aspects; no posterolateral parietal diapophyses short to long, separated or united; spurs, (9) otic crests, when present, double, diapophyses absent on posterior trunk vertebrae of involving parietals and squamosals; crests process­ three genera, (54) centra shorter or longer than like, (10) columellae present and well developed, distance across tips of parapophyses, (55) ribs ( 11) detoglossal or adetoglossal; genioglossal present on last trunk vertebrae in some, absent in muscles present or absent, (12) comua distinctly other genera, (56) ribs from 0.2 to 1.5 times dis­ separated from first basibranchials, (13) no tance across parapophyses, ( 57) tails stout to anterior first basibranchial extensions, ( 14) first slender proximally, basal tail constriction present basibranchials moderately to greatly expanded an­ or absent, (58) tibial spurs present or absent, (59) teriorly, ( 15) first ceratobranchials of significantly hind limbs only slightly larger than fore limbs, greater diameter and bulk than second cerato­ (60) fingers and toes highly variable in shape, no branchials, ( 16) epibranchials longest of articu­ fleshy palm pads; digits long and slender with lated hyobranchial elements; 1.5 to 2 times the first pointed to expanded tips in most, or short and basibranchials, longer than but less than 1. 7 times stout, or joined, or united by webbing, ( 61) five the first ceratobranchials, ( 17) second basibran­ to eight carpals, five to nine tarsals, ( 62) distal chials with widths greater than two-thirds first tarsals five do not articulate with central tarsals basibranchial lengths, ( 18) thirteen to twenty-one except in two genera, ( 63) four fingers and five trunk vertebrae, (19) three caudosacral vertebrae, toes, except for three genera which have four toes, (20) transverse processes may or may not extend (64) primitive phalangeal formulae 1-2-3-2; 1-2- beyond zygapophyses, (21) diapophyses and para­ 3-3-2; various reductions occur with extreme reduc­ pophyses present on all but last one or two verte­ tion possible 0-1-2-2, 0-1-2-2-1 in forms with five brae, ribs all bicipital, (22) ribs 0.6 to 1.5 times toes, ( 65) adults aquatic, terrestrial, subterranean, distance across parapophyses, (23) tibial spurs or arboreal, ( 66) aquatic larvae in one group, well developed, (24) aquatic larvae with three direct terrestrial development in most species, (67) epibranchials, some paedogenetic species remain testes pigmented or unpigmented, lobed testes aquatic, adults of other species aquatic to semi­ occur in some groups, ( 68) quadrato-pectoralis terrestrial. and gularis, or only gularis muscles. Range: eastern North America with some species Content: three tribes, twenty genera-Gyri­ in central North America north of Mexico. nophilus, Pseudotriton, Stereochilus, Eurycea, Content: eight genera-Gyrinophilus, Pseudo­ Typhlotriton, Typhlomolge, Haideotriton, Hemi­ triton, Stereochilus, Eurycea, Typhlotriton, Typhlo­ dactylium, Plethodon, Aneides, Ensatina, Hydro­ molge, Haideotriton, Hemidactylium. mantes, Batrachoseps, Bolitoglossa, Oedipina, Pseu­ doeurycea, Chiropterotriton, Parvimolge, Lineatri- Gyrinophilus Cope, 1869 ton, Thorius. · Definition: adults paedogenetic or fully trans­ formed, eyes well developed and functional; pre­ Tribe 1. Hemidactyliini maxillae separate, fused in paedogenetic species; Definition: plethodontine salamanders differing parietal-otic and squamosal-otic crests; well devel­ from all others by having aquatic larvae; differing oped orbitosphenoids; anterior and posterior vom­ from the Plethodontini in lacking anterior first erine teeth continuous; adetoglossal with lingual basibranchial extensions, and from the Bolitoglos­ cartilages and no genioglossal muscles; eighteen sini by having large, ossified second basibranchials, to twenty trunk vertebrae; tranverse processes of and larger first than second ceratobranchials. trunk vertebrae extend beyond zygapophyses; no Characterization: ( 1 ) premaxillae fused or sep­ basal tail constriction; four fingers and five toes. arate, ( 2) internasal fontanelles· primarily enclosed Range: eastern North America from southern by premaxillary frontal processes, or fontanelles Maine and Quebec, and Ohio, to northern Alabama secondarily closed, ( 3) maxillary facial lobes and Georgia. almost entirely anterior to midpoints of toothed Content: danielsi (Blatchley, 1901), palleucus portions, ( 4) septomaxillae and prefrontals present McCrady, 1954; porphyriticus (Green, 1827). and well developed, ( 5) vomerine tooth series sharply arched anteriorly, continuous or not with Pseudotriton Tschudi, 1838 posterior vomerine teeth, ( 6) vomers proper extend Definition: adults fully transformed; eyes well posterolaterally beyond anterior tooth series, developed and functional; premaxillae fused; parie- 1966 EVOLUTION OF LUNGLESS SALAMANDERS 51 tal-otic and squamosal-otic crests; well developed muscles; seventeen to twenty trunk vertebrae; orbitosphenoids; anterior and posterior vomerine transverse processes of trunk vertebrae do not teeth continuous; adetoglossal with lingual cartil­ extend beyond the zygapophyses; no basal tail ages and no genioglossal muscles; seventeen or constriction; four fingers and five toes. eighteen trunk vertebrae; transverse processes of Range: Interior Highlands of Missouri, Arkan­ trunk vertebrae extend beyond zygapophyses; no sas, Kansas, and Oklahoma. basal tail constriction; four fingers and five toes. Content: spelaeus Stejneger, 1892. Range: eastern North America from New York and Ohio to Florida and Louisiana. Typhlomolge Stejneger, 1896 Content: montanus Baird, 1849; ruber (Sonnini, Definition: adults blind, gilled, paedogenetic; 1802). premaxillae fused; no otic crests; no orbitosphe­ Stereochilus Cope, 1869 noids; no posterior vomerine teeth; larval attached tongues; thirteen or fourteen trunk vertebrae; Definition: adults fully transformed; eyes well transverse processes of trunk vertebrae extend developed and functional; premaxillae fused; pari­ beyond zygapophyses; no basal tail constriction; etal-otic and squamosal-otic crests; well developed four fingers and five toes. orbitosphenoids; anterior and posterior vomerine Range: Balcones Escarpment region of south­ teeth continuous; detoglossal with lingual cartilages central Texas. and no genioglossal muscles; eighteen to twenty Content: rathbuni Stejneger, 1896; tridentifera trunk vertebrae; transverse processes of trunk (Mitchell and Reddell, 1965). vertebrae extend beyond zygapophyses; no basal tail constriction; four fingers and five toes. Haideotriton Carr, 1939 Range: Atlantic Coastal Plain from Virginia to Definition: adults blind, gilled, paedogenetic; Georgia. premaxillae fused; no otic crests; no orbitosphe­ Content: marginatus (Hallowell, 1857). noids; no posterior vomerine teeth; larval attached tongues; fourteen trunk vertebrae; transverse proc­ Eurycea Rafinesque, 1822 esses of trunk vertebrae do not extend beyond Definition: adults paedogenetic or fully trans­ zygapophyses; no basal tail constriction; four formed; eyes well developed and functional; pre­ fingers and five toes. maxillae fused; parietal-otic and squamosal-otic Range: southern Georgia and north-central crests, or no crests; well developed orbitosphenoids; Florida. anterior and posterior vomerine teeth discontin­ Content: wallacei Carr, 1939. uous; adetoglossal with lingual cartilages and no genioglossal muscles; fifteen to twenty-one trunk Hemidactylium Tschudi, 1838 vertebrae; transverse processes of trunk vertebrae Definition: adults fully transformed, eyes well do not extend beyond zygapophyses; no basal tail developed and functional; premaxillae separate; constriction; four fingers and four or five toes. no otic crests; orbitosphenoids well developed; Range: eastern North America from Quebec and anterior and posterior vomerine teeth discontin­ New Brunswick to Florida, Oklahoma, and Texas. uous; detoglossal with genioglossal muscles and Content: aquatica Rose and Bush, 1963; bis­ no lingual cartilages; fifteen trunk vertebrae; trans­ lineata (Green, 1818); latitans Smith and Potter, verse processes of trunk vertebrae extend beyond 1946; longicauda (Green, 1818); lucifuga Rafin­ zygapophyses; basal tail constriction; four fingers esque, 1822; multiplicata (Cope, 1869); nana and four toes. Bishop, 1941; neotenes Bishop and Wright, 1937; Range: eastern North America from Nova Scotia pterophila Burger, Smith and Potter, 1950; quad­ and Ontario to Arkansas, Louisiana, and Florida. ridigitata (Holbrook, 1842); troglodytes Baker, Content: scutatum (Schlegel, 1838). 1957; tynerensis Moore and Hughes, 1939. Tribe 2. Plethodontini Typhlotriton Stejneger,1892 Definition: plethodontine salamanders which Definition: adults essentially fully transformed, lack aquatic larvae; differing in addition from the but eyes poorly developed and nonfunctional; pre­ Hemidactyliini by having anterior first basibran­ maxillae fused or separate; no otic crests; orbito­ cbial extensions, and from the Bolitoglossini by sphenoids with very small optic fenestrae; anterior having large, ossified second basibranchials and and posterior vomerine teeth continuous; deto­ larger first than second ceratobrancbials. glossal with lingual cartilages and no genioglossal Characterization: (1) premaxillae fused or sep- 52 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 arate, (2) internasal fontanelles primarily enclosed ereus (Green, 1818); dorsalis Cope, 1889; dunni by premaxillary frontal processes, secondarily Bishop, 1934; elongatus Van Denburgh, 1916; closed by osseous ornamentation in one species, glutinosus (Green, 1818); jordani Blatchley, 1901; ( 3) maxillary facial lobes arise from central one­ larselli Burns, 1953; longicrus Adler and Dennis, third of toothed portions, ( 4) septomaxillae and 1962; neomexicanus Stebbins and Riemer, 1950; prefrontals well developed, (5) vomerine tooth ouachitae Dunn and Heintze, 1933; richmondi series gently curved anteriorly, not continuous with Netting and Mittleman, 1938; stormi, Highton and posterior vomerine teeth, ( 6) vomerine tooth series Brame, 1965; vandykei Van Denburgh, 1906; marks posterolateral extent of vomers proper, (7) vehiculum (Cooper, 1860); wehrlei Fowler and anterior portions of parasphenoids not expanded, Dunn, 1917; welleri Walker, 1931; yonahlossee ( 8) parietals slightly or not depressed posteriorly, Dunn, 1917. no lateral aspects or parietal spurs, (9) otic crests, Ensatina Gray, 1850 when present, single involving otics only or otics and squamosals; crests long and continuous, (10) Definition: adults differ from Aneides and !'le­ columellae well developed, ( 11 ) detoglossal, short thodon in having basally constricted tails with genioglossal muscles present, (12) cornua dis­ highly specialized first caudal vertebrae, and in tinctly separated from first basibranchials, ( 13) lacking tibial spurs. anterior first basibranchial extensions well devel­ Range: far western United States from southern oped, ( 14) first basibranchials moderately ex­ British Columbia to extreme southern California, panded just anterior to midpoints, (15) first west of the Cascade-Sierra Nevada divide. ceratobranchials of significantly greater diameter Content: eschscholtzii Gray, 1850. and bulk than second ceratobranchials, (16) first Aneides Baird, 1849 ceratobranchials longest of articulated hyobran­ chial elements in most, but epibranchials equal Definition: adults differ from Plethodon and first ceratobranchials in one species; epibranchials Ensatina in having single premaxillae and special­ less than 1.25 times first basibranchials, and ized tarsi in which all elements articulate with the shorter than or same length as first ceratobran­ centrals. chials, ( 17) second basibranchials with widths Range: Applachian region of eastern North less than two-thirds first basibranchial lengths, America, mountainous regions of south-central ( 18) fourteen to twenty-four trunk vertebrae, (19) New Mexico, and far western North America west three caudosacral vertebrae, (20) transverse proc­ of the Cascade-Sierra Nevada divide from Van­ esses extend well beyond zygapophyses, (21) dia­ couver Island and Oregon to California and north­ pophyses and parapophyses present on all but last ern Baja California. one or two vertebrae, ribs all bicipital, (22) ribs Content: aeneus (Cope and Packard, 1881 ) ; 0.6 to 1.25 times distance across parapophyses, ferreus Cope, 1869; flavipunctatus (Strauch, 1870); (23) tibial spurs well developed, or absent, (24) hardii (Taylor, 1941); lugubris (Hallowell, 1849). direct terrestrial development with no aquatic lar­ val stage, adults terrestrial to arboreal. Tribe 3. Bolitoglossini Range: eastern and western North America with Definition: Plethodontine salamanders which species in the Rocky Mountains, the Interior High­ differ from the Hemidactyliini and Plethodontini lands, and Texas. by lacking second basibranchials and having larger Content: three genera-Plethodon, Ensatina, second than first ceratobranchials; differing in Aneides. addition from the Hemidactyliini by lacking aquatic larvae, and from the Plethodontini by having epi­ Plethodon Tschudi, 1838 branchials that are much longer than the first cera­ Definition: adults differ from Aneides in having tobranchials. two premaxillae, and from Ensatina in having tibial Characterization: ( 1) premaxillae fused or sep­ spurs and lacking basal tail constriction. arate, (2) anterior margins only of internasal fon­ Range: eastern North America to the Interior tanelles formed by premaxillary frontal processes, Highlands and Texas, but particularly in the Appa­ ( 3) maxillary facial lobes almost completely ante­ lachian region; western North America, primarily rior to midpoints of toothed portions, ( 4) septo­ west of the Cascade-Sierra Nevada divide, but with maxillae and prefrontals present or absent, trends populations in the Rocky Mountains of Idaho­ for reduction and loss of both, (5) vomerine tooth Montana, New Mexico, and? Utah. series gently curved anteriorly, not continuous Content: caddoensis Pope and Pope, 1951; cin- with posterior vomerine teeth, ( 6) anterior vom- 1966 EVOLUTION OF LUNGLESS SALAMANDERS 53 erine tooth series marks posterolateral extent of to high elevations in the Cascade and Sierra Nevada vomers proper, (7) anterior portions of parasphe­ regions; two species in Europe from north-central noids slightly or not expanded, (8) parietals slightly and western Italy to extreme southwestern France, or not depressed posteriorly; no lateral aspects, but and on the island of Sardinia. usually with well developed posterolateral parietal Content: brunus Gorman, 1954; genei (Schlegel, spurs, (9) no otic crests, (10) columellae present 1838); italicus Dunn, 1923; platycephalus (Camp, in some, but definite trend for reduction and loss, 1916); shastae Gorman and Camp, 1953. ( 11 ) adetoglossal, no genioglossal muscles; modi­ fied detoglossal with extremely specialized, elon­ Supergenus Batrachoseps gate genioglossal muscles in one genus, (12) Definition: plethodontine salamanders with fused cornua present or absent, when present apparently or separated premaxillae; extremely reduced or no continuous with first basibranchials, (13) no prefrontals; posterolateral parietal spurs; some­ anterior first basibranchial extensions, (14) first what reduced columellae; highly modified deto­ basibranchials slightly or not at all expanded ante­ glossal tongues with extremely elongated and riorly, (15) second ceratobranchials shorter, but slender genioglossal muscles; relatively long slen­ of significantly greater diameter and bulk than der cornua; omohyoideus muscles; sixteen to first ceratobranchials, (16) epibranchials longest twenty-one trunk vertebrae; two or three caudo­ of articulated hyobranchial elements; 2 to 3.4 sacral vertebrae; slight basal tail constriction; four times the first basibranchials; 1.8 to 6 times the toes; and multilobed testes. first ceratobranchials, ( 17) no second basibran­ Content: one genus--Batrachoseps. chials, (18) fourteen to twenty-two trunk verte­ brae, (19) two or three caudosacral vertebrae, Batrachoseps Bonaparte, 1839 (20) transverse processes extend beyond zygapo­ Definition: see above. physes, (21) diapophyses and parapophyses pres­ Range: western Oregon and California, and ent on most trunk vertebrae of most genera, but extreme northwestern Baja California; ? south­ diapophyses lost on most trunk vertebrae of sev­ eastern Alaska and 1 Jalisco, Mexico. eral genera; ribs bicipital or unicipital, (22) ribs Content: attenuatus (Eschscholtz, 1833); pacif­ 0.2 to 1 times distance across parapophyses, (23) icus (Cope, 1865); wrighti (Bishop, 1937). tibial spurs relatively small or absent, (24) direct terrestrial development with no aquatic larval Supergenus Bolitoglossa stage, adults terrestrial to arboreal. Definition: plethodontine salamanders with fused Range: far western North America west of the premaxillae; well developed, reduced, or no pre­ Cascade-Sierra Nevada divide from ?Alaska and frontals; posterolateral parietal spurs; columellae Oregon to northern Baja California, Mexico; the reduced or absent; adetoglossal, with no genio­ total neotropical (Mexico to Brazil and Bolivia) glossal muscles; slender, sinuous, relatively short and European (Italy, France, Sardinia) range of comua; no omohyoideus muscles; fourteen or the family. eighteen to twenty-two trunk vertebrae; two caudo­ Content: three supergenera, nine genera­ sacral vertebrae; slight to marked basal tail con­ Hydromantes, Batrachoseps, Bolitoglossa, Oedi­ striction; five toes; and multilobed testes. pina, Pseudoeurycea, Chiropterotriton, Parvimolge, Content: seven neotropical genera-Bolitoglossa, Lineatriton, Thorius. Oedipina, Pesudoeurycea, Chiropterotriton, Parvi­ molge, Lineatriton, Thorius. Supergenus Hydromantes Definition: plethodontine salamanders with sep­ Bolitoglossa Dumeril, Bibron, and Dumeril, 1854 arated premaxillae; no prefrontals; no posterolat­ Definition: salamanders with prefrontals present eral parietal spurs; well developed columellae; or absent; no posteriorly directed, rod-like squa­ adetoglossal, with no genioglossal muscles; no mosal processes; columellae extremely reduced or cornua; no omohyoideus muscles; fourteen trunk absent; no sublingual folds; fourteen trunk verte­ vertebrae; three caudosacral vertebrae; no basal brae; most ribs bicipital; distal tarsals 4 and 5 tail constriction; five toes; and unilobed testes. fused. Content: one genus--Hydromantes. Range: northeastern Mexico to eastern Brazil and central Bolivia. Hydromantes Gistel, 1848 Content: adspersa (Peters, 1863); altamazonica Definition: see above. (Cope, 1874); alvaradoi Taylor, 1954; arbores­ Range: three species in California at from low candens Taylor, 1954; biseriata Tanner, 1962; bor- 54 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 burata Trapido, 1942; brevipes Bumzahem and .Range: Nuevo Leon and Tamaulipas in north­ Smith, 1955; capitana Brame and Wake, 1963; eastern and Nayarit in western Mexico to Guate­ chica Brame and Wake, 1963; cerroensis Taylor, mala. 1952; colonnea (Dunn, 1924); cuchamatana Content: altamontana (Taylor, 1938); bellii (Stuart, 1943); dofl,eini (Werner, 1903); dunni (Gray, 1849); brunnata Bumzahem and Smith, (Schmidt, 1933); engelhardti (Schmidt, 1936); 1955; cephalica (Cope, 1865); expectata Stuart, epimela Wake and Brame, 1963; jlavimembris 1954; firscheini Shannon and Werler, 1955; gado­ (Schmidt, 1936); f/aviventris (Schmidt, 1936); vii (Dunn, 1926); galeanae (Taylor, 1941); gigan­ franklini (Schmidt, 1936); helmrichi (Schmidt, tea (Taylor, 1938); goebeli (Schmidt, 1936); 1936); hypacra (Brame and Wake, 1962); ligni­ leprosa (Cope, 1869); melanomolge (Taylor, color (Peters, 1873); lincolni (Stuart, 1943); 1941); nigromaculata (Taylor, 1941); rex (Dunn, macrinii (Lafrentz, 1930); marmorea (Tanner 1921); robertsi (Taylor, 1938); scandens Walker, and Brame, 1961) ; mexicana Dumeril, Bibron, and 1955; smithi (Taylor, 1938); sulcata (Brocchi, Dumeril, 1854; morio (Cope, 1869); nicefori 1883); unguidentis (Taylor, 1941); werleri Darling Brame and Wake, 1963; nigrescens (Taylor, 1949); and Smith, 1954. nigroflavescens Taylor, 1941; occidentalis Taylor, 1941; omniumsanctorum (Stuart, 1952); orestes Chiropterotriton Taylor, 1944 Brame and Wake, 1962; palmata (Werner, 1897); Definition: salamanders with prefrontals pres­ pandi Brame and Wake, 1963; peruviana (Boulen­ ent or absent; no posteriorly directed, rod-like squa­ ger, 1883); phalarosoma Wake and Brame, 1962; mosal processes; columellae present, reduced, or platydactyla (Gray, 1831); riletti Holman, 1964; absent; sublingual folds; fourteen trunk vertebrae; robusta (Cope, 1894); rostrata (Brocchi, 1883); most ribs bicipital; distal tarsals 4 and 5 separated rufescens (Cope, 1869); salvinii (Gray, 1869); or fused, when separated 5 usually larger than 4 savagei Brame and Wake, 1963; schizadactyla and articulates with centrale, or, if 4 larger than Wake and Brame, 1966; schmidti (Dunn, 1924); 5, prefrontals are absent. sima (Vaillant, 1911); sooyorum Vial, 1963; Range: southern Nuevo Leon and southwestern striatula (Noble, 1918); subpalmata (Boulenger, Tamaulipas through eastern and east-central Mex­ 1896); vallecula Brame and Wake, 1963; veracru­ ico to Guatemala, El Salvador, Honduras, and cis Taylor, 1951; yucatana (Peters, 1882). Costa Rica. Content: abscondens Taylor, 1948; arboreus Oedipina Keferstein, 1868 (Taylor, 1941); barbouri (Schmidt, 1936); brom­ Definition: salamanders with no prefrontals; no eliacia (Schmidt, 1936); chiropterus (Cope, 1863); posteriorly directed, rod-like squamosal processes; chondrostega (Taylor, 1941); dimidiatus (Taylor, no columellae; sublingual folds; eighteen to 1939~.; lavae (Taylor, 1942); magnipes Rabb, twenty-two trunk vertebrae; most ribs unicipital; 1965; megarhinus Rabb, 1960; mosauri (Woodall, distal tarsals 4 and 5 fused. 1941); multidentatus (Taylor, 1938); nasalis Range: southern Mexico through Central Amer­ (Dunn, 1924); picadoi (Stejneger, 1911); priscus ica and western Colombia to north-central Ecuador. Rabb, 1956; xolocalcae (Taylor, 1941). Content: alfaroi Dunn, 1921; bonitaensis Tay­ lor, 1952; collaris (Stejneger, 1907); complex Lineatriton Tanner, 1950 (Dunn, 1924); cyclocauda Taylor, 1952; elongata Definition: salamanders with prefrontals; no (Schmidt, 1936); gracilis Taylor, 1952; ignea posteriorly directed, rod-like squamosal processes; Stuart, 1952; inusitata Taylor, 1952; longissima columellae reduced; sublingual folds; fourteen Taylor, 1952; pacificensis Taylor, 1952; parvipes trunk vertebrae; most ribs unicipital; no distal (Peters, 1879); poelzi Brame, 1963; syndactyla tarsal 5. Taylor, 1948; taylori Stuart, 1952; uniformis Kef­ Range: central Veracruz, Mexico. erstein, 1868. Content: lineola (Cope, 1865). Pseudoeurycea Taylor, 1944 Thorius Cope, 1869 Definition: salamanders with prefrontals; no Definition: salamanders with prefrontals ex­ posteriorly directed, rod-like squamosal processes; tremely reduced or absent; posteriorly directed, columellae greatly reduced; sublingual folds; four­ rod-lik:e squamosal processes; no columellae; sub­ teen trunk vertebrae; most ribs bicipital; distal tar­ lingual folds; fourteen trunk vertebrae; most ribs sals 4 and 5 separated, 4 larger than 5, and 5 does bicipital; distal tarsals 4 and 5 fused. not articulate with the centrale. Range: southern Veracruz and Puebla to Guer- 1966 EVOLUTION OF LUNGLESS SALAMANDERS 55 rero and Oaxaca, Mexico, north of the Isthmus of genus, Desmognathus, was placed in the Desmog­ Tehuantepec. nathidae, and the other genera in the Plethodonti~ Content: dubitus Taylor, 1941; macdougalli dae. Cope ( 1869) erected the family Thoriidae for Taylor, 1949; maxillabrochus Gehlbach, 1959; the genus Thorius. Cope (1859, 1866, 1869) was minutissimus Taylor, 1949; narisovalis Taylor, the first to present osteological information of any 1939; pennatulus Cope, 1869; pulmonaris Taylor, significance and to characterize and discuss rela­ 1939; schmidti Gehlbach, 1959; troglodytes Tay­ tionships of salamander families. Desmognathids lor, 1941. were said to differ from plethodontids in having stalked rather than sessile occipital condyles, opis­ Parvimolge Taylor, 1941 thocoelous rather than amphicoelous vertebrae, This genus, recognized by recent authors, is a and a frontal process extending into the superpala­ poorly defined and apparently unnatural assem­ tal vacuity. Thoriids were said to have opisthocoe­ blage of three distinct species, and diagnosis is not lous vertebrae and ossified carpals and tarsals. possible. The species may prove to be members of Cope thought thoriids combined the characters of the genera Chiropterotriton and Pseudoeurycea, the Desmognathidae, Plethodontidae, and Amby­ but at least one, P. townsendi, appears sufficiently stomatidae. distinct to warrant a separate genus. The two Boulenger (1882) reduced the three families of species which have been examined (P. richardi, Cope to two subfamilies (thoriids included in Des­ P. townsendi) have small prefrontals; no posteriorly mognathinae) of the Salamandridae, hut Cope directed, rod-like squamosal processes; extremely (1889) rejected this work. Cope later modified his reduced or no columellae; sublingual folds present; views (1894 a) and placed Thorius in the Desmog­ fourteen trunk vertebrae; and bicipital ribs. Distal nathidae. As a result of his discovery of Hapto­ tarsals 4 and 5 are fused in P. townsendi. glossa, an elongate Costa Rican genus with an Range: central and southern Veracruz, Mexico, attached tongue and opisthocoelous vertebrae, and north-central Costa Rica. Cope (1894 b) recognized a subfamily Thoriinae Content: praecellens Rabb, 1955; richardi Tay­ within the family Desmognathidae. The subfamily lor, 1949; townsendi (Dunn, 1922). included Haptoglossa, Thorius, and Typhlotriton (described as a desmognathid by Stejneger, 1892). EVOLUTIONARY RELATIONSHIPS The unique type of Haptoglossa is lost and the AND TRENDS status of the genus remains enigmatic (see also Taylor, 1944) . Relationships The family Typhlomolgidae was erected for the Major Familial Divisions single genus Typhlomolge by Stejneger and Bar­ The family Plethodontidae, as proposed by Gray bour (1917), despite Emerson's (1905) study (1850), included species now included in the fam­ which showed it to be closely related to Eurycea. ilies Hynobiidae and Ambystomatidae as well as the The family received no support and :Rowler and following four groups: Plethodon; Desmognathus Dunn ( 1917) simply stated "Typhlomolge is a and Hemidactylium; Batrachoseps, Spelerpes permanent larva of some plethodont." (which included species now placed in Eurycea, On the basis of his studies of vertebral articula­ Manculus, Pseudotriton, Gyrinophilus, and Pseu­ tions, Moore (1900) suggested that desmognathids doeurycea), Geotriton (=Hydromantes), and Oedi­ be included in the family Plethodontidae. He was pus (=Bolitoglossa); Ensatina. Hallowell (1856) followed by most subsequent workers (Stejneger used family endings for three groups included in and Barbour, 1917; Fowler and Dunn, 1917; Gray's single family: (1) Plethodontidae-Ple­ Dunn, 1917 and 1926; Noble, 1931; Herre, 1935; thodon, Aneides, Desmognathus, (2) Bolitoglos­ Piatt, 1935; Schmidt, 1953; Conant, 1958). Dunn's sidae-Pseudotriton, Spelerpes, Batrachoseps, Geo­ (1926) classical work, which has long been the triton, (3) Hemidactylidae-Hemidactylium. The standard authoritative reference on plethodontids, groups were not precisely defined and never be­ recognized no formal groupings below the family came established in the literature. level. Cope (1859) recognized two groups on the basis Recently there has been an attempt to revive of tongue attachment; those with free tongues the family Desmognathidae. Smith and Taylor (including Batrachoseps) were called Spelerpeae, ( 1948) recognized two families of lungless sala­ those with attached tongues, Plethodontae. In manders, and two subfamilies (Thoriinae and Ple­ 1866 Cope revised his views and recognized two thodontinae) of plethodontids. The Thoriinae families of present day plethodontids. A single included only Thorius. Soler (1950) presented evi- 56 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 dence of the distinctiveness of Desmognathus and than any desmognathine genus (see below). The Leurognathus, and argued for recognition of a plethodontine genera are probably morphologi­ family Desmognathidae. In a rather uncritical cally more similar to the ancestral common stock review based on the literature, von Wahlert (1957) than are the desmognathines, and the Plethodon­ recognized a superorder Plethodontoidea (sug­ tinae must be considered the most primitive sub­ gested by Laurent, 1947) with two families, Des­ family (see Fig. 12). mognathidae and Plethodontidae. Two subfamilies of plethodontids were recognized: Plethodontinae (Aneides, Batrachoseps, Ensatina, Hemidactylium, Plethodon, Stereochilus) and Thoriinae (remaining nondesmognathine plethodontids) . Recent authors (Teege, 1956; Conant, 1958; Highton, 1961; Mar­ tof, 1962; Wake, 1963; Estes, 1964 and 1965; Monath, 1965) have not recognized a family Desmognathidae. Plethodontid salamanders fall into two major assemblages that I consider to be of subfamilial rank. The subfamily Plethodontinae contains twenty of the twenty-three genera, and is the more generalized and progressive group. The more spe­ cialized subfamily Desmognathinae is comprised of three genera. It is difficult to discuss the sub­ families in terms of one being more primitive or advanced than the other. Primitive members of both groups have aquatic larvae, but desmogna­ thine larvae have four epibranchials and pletho­ dontines have three. In some respects (e.g., palatal elements, gill rami) plethodontine larvae are the more generalized, but several plethodontine lar­ vae have secondary specializations (elongate snout and reduced eyes of Gyrinophilus, pond specializa­ Figure 12. Diagrammatic representation of the relation­ tion of Hemidactylium, Manculus and Stereochi­ ships of the major groups of plethodontid salamanders. lus) . Some plethodontines retain the primitive pre­ maxillary metamorphic pattern, but others do not Functional specializations of four types in the nor do the modem desmognathines. Primitive desmognathines have led to drastic skeletal mod­ characters found in primitive plethodontines but ifications. These modifications sharply distinguish not in desmognathines include presence of naso­ desmognathine from plethodontine genera, and lacrimal ducts, large septomaxillae, well developed include: ( 1) streamlining, a response to selective prefrontals, toothed vomerine preorbital processes, forces operative in the primitive, swift mountain continuous anterior and posterior vomerine tooth brook environments, (2) skull strengthening, a series, amphicoelous vertebrae, cartilaginous antor­ response to selective forces correlated with use of bitals, convex frontals, and open meckelian heads as wedges, ( 3) new mouth opening mechan­ grooves. Other primitive characters (tibial spurs, isms, the result of selective forces in the primitive mesopodial numbers and configuration) are shared environment and related to peculiarities in use of by desmognathines and primitive plethodontines. the heads, ( 4) vertebral and limb modifications, Desmognathines have many unique morphological the results of locomotor specializations. specializations, but subfamily-wide plethodontine Streamlining is a common occurrence in verte­ specializations are not apparent. The desmogna­ brates which inhabit aquatic environments, and thines probably diverged from a pre-pletbodontine especially those that occur in swift streams. Cer­ stock at an early date (Mesozoic). Desmognathines tainly the remarkable desmognathine skull flatten­ have become highly specialized, but have retained ing is the most important of the streamlining mod­ a number of primitive characters. The Plethodon­ ifications. Flattening imparts an elongate wedge­ tinae is a considerably more generalized and diverse shape to the skulls, a factor of particular advantage group than the Desmognathinae, yet some pletbo­ to these organisms which not only inhabit rapidly dontine genera can be considered more advanced flowing brooks, but also hide under rocks and in 1966 EVOLUTION OF LUNGLESS SALAMANDERS 57 rock crevices by forcing their way head first (Dunn, relatively minor importance in desmognathines; 1926; Martof, 1962). The anterior cranial elements apparently the anterior dorsal spinal muscles have of desmognathines are relatively dense, thickened, assumed the primary desmognathine mouth open­ closely articulated bones, and the dorsal surfaces ing function. The dorsal spinal muscles are excep­ of the skulls are solid and relatively smooth. The tionally well developed in desmognathines and frontals and anterior margins of the parietals ·are extend onto the skull to insert on raised otic crests, also fiat and smooth. Head muscles that primitively behind the parietal-otic troughs, and on the parie­ originate on the dorsal surfaces of the frontals and tals in front of the troughs. The muscles originate parietals have shifted their origins lateroventrally, on the dorsal surfaces of the anterior trunk verte­ and arise from the lateral margins of the bones, brae. Unique, moderately to well developed verte­ below .the ridge-like dorsolateral margins, and bral pterygapopbyses provide additional area for from the orbitosphenoids. Reduction or absence of dorsal spinal muscle origin. vomerine vaulting is also directly associated with Despite the fact that the adductor mandibulae the streamlining and skull depression. musculature contains ligaments, it plays a role in Other features associated with streamlining lowering the desmognathine skull onto the mandi­ include well developed eyes which are neither bles. Additional musculature also is brought into prominent nor strongly protuberant, fore limbs play. The quadrato-pectoralis musculature is considerably smaller than hind limbs, and very remarkably well developed in desmognathines, smooth transitions from relatively small heads to extending as stout bands from the quadrate region cylindrical trunks and stout tails. to the gular fold (see Piatt, 1935); contraction of Skulls of plethodontines remain relatively rigid the quadrato-pectoralis muscles pulls the skull in reference to the vertebral column, and the man­ down while contraction of the adductor mandibu­ dibles are lowered to open the mouths. In contrast lae muscles raises the mandibles. The two sets of the mandibles of desmognathines are held rela­ muscles effectively close the mouth. tively rigid and the skulls proper are raised some­ When the skull swings dorsally on the atlas, the what during mouth opening. The mandibles are mandibles are also raised. Mandibular elevation far from immobile in Desmognathines (see Noble, ceases, however, as the depressor mandibulae 1931), but are much more so than in plethodon­ muscles contract and facilitate mouth opening by tines. It has been suggested (Dunn, 1926; Conant, restricting dorsal mandibular motion. The skull 1958) that the relative immobility of the lower swings dorsally and movement of the quadrates on jaws stiffens the anterior portions of the bodies so the articulars occurs. Since the atlas is a relatively the salamanders can more readily force their way solid fulcrum and vertical motion of the mandibles under objects. Functional modifications of the jaw is restricted, skull movement propels the mandibles musculature immobilize the mandibles. Slips of forward. Movements of the quadrates on the articu­ the adductor mandibulae musculature (adductor lars perhaps has served as a selective force that has mandibulae extemus) extend from the prearticu­ led to strengthened quadrates. Desmognathine lars to the atlas in all plethodontids, but desmog­ quadrates are enlarged and firmly articulated with nathines differ from plethodontines in that the slips the parasphenoids and occipito-otics, and anchored contain stout, relatively inflexible ligaments that by strong jugal ligaments to the maxillae. The great limit the ventral movement of the mandibles rela­ tensions in the ligament are evidenced by the for­ tive to the trunk. mation of moderate to very large maxillary jugal Desmognathines have stalked occipital condyles processes in the desmognathines. The squamosals that effectively move the skull away from the atlas are stout and articulate firmly with the quadrates and provide room for swinging the skull up and and occipito-otics. The results of the above modifi­ down to open and close the mouth. Further mod­ cations are that the desmognathine quadrates are ifications facilitating skull movement on the verte­ solidly anchored to the brain case, and the pos­ bral column are the drastic reduction in size of terior portions of the skull are as strong and rigid the odontoid process and movement out of the as the anterior portions. foramen magnum. In addition the occipital con­ Primitive desmognathines are aquatic organisms dylar articular facets are strongly convex and are living in swift streams. A number of morphological received by enlarged, strongly concave atlantal specializations are related to stream locomotion. cotyles. Among these are the heavily muscularized trunks Muscular modifications have evolved to enable and stout, compressed tails. Development of opis­ skull elevation. The depressor mandibulae muscles thocoely facilitates increased intervertebral move­ lower the mandibles of plethodontines, but are of ment, the result of increased trunk and tail flexure. 58 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 Desmognathine trunk vertebrae are also charac­ favored dorsal and ventral fontanelle closure as terized by the presence of hypapophysial keels and parts of trends toward skull strengthening. Leurog­ basapophyses. Auffenberg (1961) and Estes (1964) nathus is unique in having very long vomerine pre­ suggest that such processes relate to the develop­ orbital processes that extend beyond the vomerine ment of muscles aiding lateral undulation and to bodies to articulate firmly with the maxillary palatal swimming efficiency. shelves. Leurognathus differs further in having A final feature associated with aquatic habits better developed hypapophysial keels on more and the particular behavior of desmognathines is vertebrae, and fused hyperapophyses on all trunk the hypertrophy of the bind limbs, which are much vertebrae. larger than the fore limbs. The large hind limbs One specialization found in Desmognathus but apparently play important roles in the locomotion not in Leurognathus is the presence of large, calci­ of desmognathines, particularly in aiding the sala­ fied plaques within the mouths under the eyes (the manders to force their way under and between "crushing plates" of Noble, 1931). The large, objects. irregularly-shaped plaques are found in at least There is no question as to the distinctiveness of three species (D. quadramaculatus, D. monticola, the Desmognathinae, but if the differences which D. ochrophaeus), and are apparently unique fea­ separate desmognathines and plethodontines are tures of the genus. carefully examined, it is apparent that most result Since Moore's (1899) description of Leurogna­ from adaptation to a particular habitat. Thus thus the status of the genus has not been seriously streamlining, skull strengthening, mouth opening questioned. Osteological data indicate that Leurog­ modifications, vertebral modifications, and limb nathus probably arose as a specialized branch of a modifications are all directly related to way of life Desmognathus ancestral stock that may have been of the salamanders, and are specialized adaptations very similar to modem D. quadramaculatus (see to the environment in which they occur. also Dunn, 1926). It is more difficult to separate Desmognathines and plethodontines are closely D. quadramaculatus from. L. marmoratus than related, and it is likely that they have had some from other Desmognathus. One might question the common ancestor. The differences between the advisability of generic recognition of a single groups are sharp, but when analyzed carefully, are species so obviously closely related to a larger, shown to be the results of relatively minor speciali­ rather diverse genus. I agree with Martof (1962) zations in the Desmognathines. So many character­ that recognition is justifiable because Leurognathus istics are held in common by the plethodontines has entered upon a new way of life, one not fol­ and desmognathines that recognition of a family lowed by any species of Desmognathus, and is Desmognathidae as proposed by Cope (1869) and morphologically distinct from all Desmognathus. recently advocated by Smith and Taylor ( 1948) Leurognathus is the most aquatic of any of the and Soler (1950) appears unwarranted. metamorphosed plethodontids and is restricted to aquatic situations (Martof, 1962). The distinctive Relationships of the Desmognathine Genera morphology, ecology, and life history of the species, Desmognathus and Leurognathus are the best L. marmoratus, warrant separation from Desmog­ known of the three desmognathine genera, and nathus and placement in a distinct genus. similarities and differences of the two have been Highton (1961) announced the discovery of a discussed to some extent by Moore (1899), Dunn remarkable new genus and species, Phaeognathus (1926), and Martof (1962). The genera are basic­ hubrichti, a very large, elongate, terrestrial bur­ ally similar, and differences are primarily the result rower. Valentine (1963 b, 1963 c) and Brandon of the peculiar selective pressures within the strictly ( 1965) have recently provided some information aquatic environment to which Leurognathus has that supports Highton's suggestion that Phaeogna­ adapted. Leurognathus has a more flattened skull thus is a desmognathine, and I have corroborated and a sharper, more pointed snout than Desmogna­ many of their observations; additional support has thus. Internal nares are reduced to small, laterally been presented above. Phaeognathus differs sharply placed slits, and the internasal fontanelles of large and strikingly from Desmognathus and Leurogna­ Leurognathus are completely closed. The fontan­ thus in many characters. Desmognathus and Leu­ elles primitively contain glands that furnish lubri­ rognathus are rather stout, short-bodied forms with cant utilized while swallowing by salamanders in fifteen trunk vertebrae, but the slender, elongate terrestrial or semiterrestrial situations; Leurogna­ Phaeognathus has twenty-one to twenty-three; limbs thus has become so completely aquatic that such and digits of Phaeognathus are greatly shortened in lubrication is unnecessary, and selection bas comparison with the other genera. Phaeognathus 1966 EVOLUTION OF LUNGLESS SALAMANDERS 59 is a terrestrial burrower, a habit not encountered be related to new types of locomotion and feeding. elsewhere in the Desmognathinae, and several There is no question concerning the distinctive­ characters are apparently related to burrowing. ness of Phaeognathus. Apparently it is a relatively The skull is rather bullet-shaped and only slightly advanced, specialized genus derived from a :flattened. The head is small relative to body size. Desmognathus-Leurognathus ancestral stock at a Skin covering the snout is co-ossified to the anterior relatively early date. cranial elements, which are sculptured and eroded. Of the three desmognathines Desmognathus is Particularly striking are the large, spatulate, max­ the most generalized and primitive. Leurognathus illary jugal processes which extend almost to the is closely related to Desmognathus and has entered quadrates. Size of the processes testifies to the a new adaptive subzone--the aquatic environment. extreme tension in the jugal ligaments, and to Phaeognathus is more closely related to Desmog­ the over-all strength of the skulls. thus than to Leurognathus, and was probably de­ The mandibular adductor series of muscles of rived from the common ancestral stock at an earlier desmognathines consist of two major bundles, the date than was Leurognathus (see Fig. 13). Phaeog­ adductor mandibulae anterior and the adductor nathus bas entered a new adaptive zone--the mandibulae externus (terminology of Edgeworth, terrestrial environment-ili an opposite evolution­ 1935)_. The former extend from the medial margins ary direction from that of Leurognathus. Desmog­ of the prearticulars to the lateral walls of the brain­ nathus bas primitive species which occupy the case, particularly the dorsal portions of the orbito­ primitive semiaquatic niche (see below), but the sphenoids and the ventrally directed lateral lips of genus is undergoing an adaptive radiation and in- the frontals. The adductor mandibulae extemus are the muscles called temporalis by various authors (e.g., Soler, 1950) and extend from the atlases to the prearticulars. Phaeognathus has larger adductor mandibulae anterior and smaller adductor man­ dibulae extemus muscles than Desmognathus and Leurognathus. Unique, lateral, frontoparietal ridges of Phaeognathus serve as points of attach­ ment for the enlarged adductor mandibulae ante­ rior musculature. In addition the dorsal spinal musculature is better developed in Phaeognathus than in other desmognathines, and the points of insertion of most of the musculature are the sites of the raised terminal projections of the prominent fronto-parietal ridges. As a result of the relatively large amounts of musculature which insert on the parietals anterior to the parietal-otic troughs, the parietals are less :flattened and the two parietal re­ gions are less distinguishable in Phaeognathus than in Desmognathus or Leurognathus. The quadrato-pectoralis and especially the gularis muscles are considerably less developed in Phaeog­ nathus than in the other desmognathine genera. The gularis muscles are barely perceptible and are Desmognath inae of little functional importance, and the quadrato­ pectoralis muscles are relatively slender. It is obvious that skull raising specializations are less well developed ili Phaeognathus than ili Desmog­ nathus or Leurognathus. The strong development Figure 13. Diagrammatic representation of relation­ of the dorsal spinal musculature could as well be ships within the subfamily Desmognathina.e. a burrowing as a mouth opening adaptation. Phaeognathus differs further from the other eludes species which have successfully entered desmognathilies ili characters of the vertebrae and terrestrial niches. Desmognathus is the most suc­ vomers (see above). The characters have no obvi­ cessful and evolutionarily the most plastic of the ous connection with ancestral conditions and may three genera. 60 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 Relationships of the Plethodontine Genera two subfamilies (Plethodontinae: Aneides, Batra­ Tribal relationships choseps, Ensatina, Hemidactylium, Plethodon, Stereochilus; Thoriinae: Bolitoglossa, Chiroptero­ By far the largest number of species and genera triton, Eurycea, Gyrinophilus, Hydromantes, of lungless salamanders are included within the Lineatriton, Magnadigita, Manculus, Oedipina, subfamily Plethodontinae. These species may be Parvimolge, Pseudotriton, Pseudoeurycea, Thorius, placed conveniently in three natural subgroups or Typhlotriton, ?Eladinea, ?Haideotriton, ?Typhlo­ tribes on the basis of structural similarities. The molge). His treatment of Stereochilus was novel generic assemblages are sufficiently distinct to in­ and without explanation. The only recent authors dicate continuity of descent within each tribe, but to recognize a subfamily Thoriinae have been not to warrant subfamilial recognition. Certain Smith and Taylor (1948) and they clearly included genera are somewhat intermediate between two only Thorius. The family group name Bolitoglos­ tribes, and no tribe is as distinct a group as the sidae (Hallowell, 1856) bas clear priority over subfamily Desmognathinae. Thoriidae (Cope, 1869) as used by von Wahlert The groupings proposed here differ from those (his Thoriinae). This was recognized by Monath of other authors, and some historical information (1965) who follows von Wahlert but refers to is necessary as an introduction to the following members of the latter's subfamily Thoriinae as discussion. Dunn's (1926) comprehensive work "bolitoglossines." Brame (1960) has also referred recognized two major groups of non-desmogna­ to this group as the bolitoglossine genera. thlne genera, the attached tongued Plethodon group The tribe Hemidactyliini is restricted in distri­ (Plethodon, Ensatina, Aneides, Hemidactylium, bution to eastern North America and contains eight Batrachoseps) and the free tongued Eurycea group genera: Gyrinophilus, Pseudotriton, Stereochilus_, (Eurycea, Manculus, Gyrinophilus, Pseudotriton, Eurycea, Typhlotriton, Typhlomolge, Haideotn­ Hydromantes, Oedipus) with Typhlotriton, Typhlo­ ton, and H emidactylium. The tribe Plethodontini molge, and Stereochilus intermediate. Most subse­ has a discontinuous distribution in eastern, central, quent authors have followed Dunn. Noble (1925, and western North America, and contains three 1927 a, 1927 b, 1931) suggested that there might genera: Plethodon, Aneides, and Ensatina. Mem­ be a relationship between Hydromantes and Pletho­ bers of the tribe Plethodontini are called pletho­ don. Herre ( 1935) saw an Oedipus-Hydromantes donines in this paper. The tribeBolitoglossini occurs group arising from a Plethodon stock and derived in western North America, Mexico, Central and the desmognathines from a Pseudotriton-like ances­ South America, and Europe, and contains nine tor, but otherwise followed Dunn. Piatt (1935) genera: Hydromantes, Batrachoseps, Bolitoglossa, recognized Stereochilus (Stereochilus, Typhlotriton, Oedipina, Pseudoeurycea, Chiropterotriton, Parvi­ Typhlomolge) and Gyrinophilus (Gyrinophilus, molge, Lineatriton, and Thorius. The bolitog~os­ Pseudotriton, Eurycea, Manculus) groups in one sine group bas the most genera, the most species, assemblage, and Plethodon (Plethodon, Aneides, the most diversity, and the most specialization of Ensatina, Hemidactylium) and Oedipus (Oedipus, the assemblages of the Plethodontinae. Hydromantes, Batrachoseps) groups in another. All members of the Hemidactyliini have aquatic Piatt's arrangement was unique in that he pro­ larvae, but all members of the other two tribes posed parallel evolution of free tongues. He stated have direct terrestrial development. This is a char­ "Hydromantes and Oedipus have no close affinities acter of major importance and one of the main to the Gyrinophilus group but arose from early reasons for considering the Hemidactyliini the Plethodon stock by means of a form like Batra­ most primitive of the three tribes. The hemidac­ choseps." Piatt's arguments were not particularly tyliine genera are aquatic to semiterrestrial as convincing and he has not been followed by sub­ adults, while the other tribes are strictly terrestrial sequent workers. Tanner (1952) recognized two or arboreal. groups of free tongued genera which he thought Hemidactyliine genera have larger and ~tronger were derived from a common free tongued ances­ second basibranchials than other plethodontids, and tor. His groups were: (1) Gyrinophilus, Pseudo­ the character is apparently related to aquatic life. triton, Eurycea, Manculus and (2) Hydromantes, The bony second basibranchials are associated Bolitoglossa, Magnadigita, Pseudoeurycea, Chiro­ with the hyobranchial musculature, forming trans­ pterotriton, Oedipina, Parvimolge, Lineatriton, verse septa in the rectus cervicis of adults. n;e Thorius. Tanner suggested that the relationships of second basibrancbials presumably play a role m Batrachoseps were with Plethodon and rejected strengthening this musculature, which is of great Piatt's proposal. Von Wahlert (1957) recognized importance in gill movements of larvae, and fiota- 1966 EVOLUTION OF LUNGLESS SALAMANDERS 61 tion and submergence mechanisms of adults. genera. Generalized and presumably primitive Second basibranchials are present in plethodonines, features of the bolitoglossines included paired pre­ but are relatively smaller than in hemidactyliines; maxillae in primitive species, and low and invari­ the elements are absent in all bolitoglossines. able numbers of trunk vertebrae in most genera. Several additional characters associated with The Hemidactyliini and Plethodontini include hyobranchial structure are among the most diag­ relatively primitive genera, and are rather general­ nostic of the tribes (see above under hyobranchial ized groups in comparison with the Bolitoglossini. skeleton). Primitive features found in some or all hemidac­ Several skull characters separate the tribes. tyliines include aquatic larvae, primitive nasolac­ Significant anterior parasphenoid expansion occurs rimal duct pathways, retention of genioglossal only in the hemidactyliine genera. The anterior muscles, retention of all primitive plethodontid vomerine tooth rows mark the posterior margins cranial elements, relatively long maxillae, con­ of the vomers proper in plethodonines and bolito­ tinuous anterior and posterior vomerine tooth glossines, but bony vomerine growth extends pos­ series, and separated premaxillae. terior to the tooth series, below the parasphenoids, Certain hemidactyliines have specialized condi­ in hemidactyliines. Anterior and posterior vomer­ tions which include loss of the fifth toes, presence ine tooth series are continuous only in four hemi­ of basal tail constriction, great first basibranchial dactyliine genera. Maxillary facial lobes arise from expansion, genioglossal muscle loss (most genera), the central one-third of the maxillae in plethodon­ attainment of true adetoglossy, separation of the ines, but the lobes are entirely anterior to the anterior and posterior vomerine tooth series, mid-points in the other tribes. The parietals of shortened transverse processes, presence of hypa­ plethodonines and bolitoglossines are relatively flat, pophysial rudiments, presence of well developed but those of hemidactyliines have distinct lateral basapophyses, attainment of paedogenesis, and aspects. premaxillary fusion. Ribs tend to be shorter, and vertebral centra Plethodonines are primitive in having detoglossy longer and slenderer in bolitoglossines than in the and genioglossal muscles in all species. In addition other two tribes. they have all the primitive cranial elements, and A number of specialized characters occur in at two of the three genera have paired premaxillae. least some species in two or more tribes, and most Specializations found in all species include separa­ or all are the result of parallel evolution. These tion of anterior and posterior vomerine tooth include proximal tail constriction (Hemidactyliini, series, advanced nasolacrimal duct routes, and Plethodontini, Bolitoglossini), loss of fifth toes, direct terrestrial development. Certain species (H,B), raised atlantal bosses (H,P), elongation, have other specializations including loss of tibial and increased numbers of trunk and tail vertebrae spurs and of vomerine preorbital processes, (H,P,B), otic crests (H,P), fused premaxillae strengthened skull elements, fused premaxillae, (H,P ,B) , loss of vomerine preorbital processes specialized maxillae, specialized arrangements of (H,P,B), attainment of significant tongue freedom mesopodial elements, and hypertrophy of quadrato­ and ability to protrude tongues for significant dis­ pectoralis musculature. tances from the mouths (H,P,B), loss of genio­ All available evidence indicates that the hemi­ glossal muscles (H,B), reduction and loss of tibial dactyliine line is the most primitive and generalized spurs (P,B), and tendencies toward opisth0coely of the three. A major consideration is the retention (H,P,B). of an aquatic larval stage in all members of the The Bolitoglossini is obviously the most special­ tribe. Plethodonines have a more primitive tongue ized and advanced of the three tribes. Specialized structure with retention of anterior attachments, and advanced conditions found only in some to all strong genioglossal muscles, and relatively short bolitoglossine genera include unicipital ribs and epibranchials; however, one hemidactyliine genus, loss of the diapophyses on most trunk vertebrae, Hemidactylium, also retains both anterior attach­ loss of prefrontals, loss of septomaxillae, loss of ment and genioglossal muscles. Features in which columellae, loss of maxillary teeth, loss of omo­ plethodonines are apparently more primitive than hyoideus muscles, fusion of premaxillary frontal hemidactyliines are primarily those related to processes, extensive mesopodial fusions, second tongue structure, and in most other characters basibranchial loss, and presence of parietal spurs. hemidactyliines are definitely more primitive. Some bolitoglossines have reduced numbers of The plethodonine line is closer to the hemidac­ chromosomes (Kezer, in litt.). The above charac­ tyliine line than is the bolitoglossine line (see ters are found in several to many bolitoglossine (Fig. 12). The primary feature relating hemidac- 62 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 tyliines and bolitoglossines is the attainment of The genera share a number of characters, includ­ true adetoglossy, but evidence bas been presented ing: ( 1) similar anterior cranial elements, in a which indicates that parallelism likely bas occurred. broad sense, ( 2) broad palatal shelves, ( 3) parie­ The above evidence indicates that hemidactyl­ tal-otic and squamosal-otic crests (shared with iines are probably closer to the ancestral pletho­ some Eurycea), ( 4) continuous anterior and pos­ dontid stock than any desmognathine, plethodo­ terior vomerine tooth series (shared with Typhlo­ nine, or bolitoglossine. Desmognathines were triton and some Eurycea), (5) broadly similar probably the earliest derivatives of the stock, fol­ hyobranchial structure, ( 6) lingual cartilages lowed by bolitoglossines and finally by plethodo­ (shared with Eurycea), (7) no genioglossal muscles nines. Within the Plethodontinae the tribe Hemi­ (shared with Eurycea and Typhlotriton), (8) dactyliini is apparently the central, primitive moderate first basibranchial expansion (shared group, the Plethodontini is somewhat more with Hemidactylium), (9) moderate-sized cornua, advanced, and the Bolitoglossini is the most spe­ (10) high modal numbers of vertebrae (shared cialized and advanced of the tribes (see Fig. 12) . with Typhlotriton), (11) proximally fused diapo­ physes and parapophyses, ( 12) long transverse processes (shared with Hemidactylium), (13) Relationships of the hemidactyliines diapophysial origins just slightly in advance of Generic groupings parapophysial origins (shared with Hemidactyl­ Three distinct generic assemblages of hemidac­ ium), (14) short ribs (shared with Hemidactyl­ tyliine salamanders are apparent: ( 1) Gyrino­ ium), ( 15) stout, relatively short vertebral centra, philus, Pseudotriton, Stereochilus, (2) Eurycea, ( 16) long frontals. Typhlotriton, Haideotriton, Typhlomolge, (3) Pseudotriton and Gyrinophilus resemble each Hemidactylium (see Fig. 14). other more closely than either resembles Stereo­ Gyrinophilus, Pseudotriton, and Stereochilus chilus. In addition to the above characters they are relatively primitive, rather closely related share enlarged adductor mandibulae anterior groups, but all represent unquestionably distinct muscles, which insert on top of the frontals and evolutionary lines worthy of generic recognition. virtually cover the interorbital regions. This mus- $' ,,, ,,, <:- . '> ...... _'> ...... _'> cf cf ..._o 0 ...... _o, .....~ ~ ,.... 0 "' Cl ~ ...... ~ () ..!" Cl 0 oq 0 s bo 0 bo -.SI <>"' ,o ..... ·$ '> ~ bib 'b ~ .... ~ $' ...... "' ' ~ J...~ "" ~ q_"" 45' :t J...~ ::?

Hemidactylii n i

Figure 14. Diagrammatic representation of the relationships of the hemidactyliine genera. 1966 EVOLUTION OF LUNGLESS SALAMANDERS 63 culature covers only the posterolateral frontal ily. The larvae of Gyrinophilus are somewhat spe­ extremities in other hemidactyliines. Second basi­ cialized (elongate snout, reduced eyes) and adults branchial width is greater than first basibranchial are adetoglossal with elongated anterior cranial length in the two genera, but lesser in Stereochilus. elements, but in most other features the genus is Both Gyrinophilus and Pseudotriton have large, generalized; it may well be the most primitive adetoglossal tongues, but Stereochilus has a rela­ plethodontid, as suggested by Dunn (1926). tively small tongue with a membranous anterior The Eurycea-Typhlotriton-Haideotriton-Ty phlo­ attachment. molge line appears to have been derived from a Orobman (1959) placed Gyrinophilus in the Gyrinophilus-Pseudotriton ancestral stock. Typhlo­ synonymy of Pseudotriton on morphological triton is more primitive than Eurycea, and paired grounds, but Martof and Rose (1962) disagreed premaxillae, relatively long snouts, nasal-maxilla and reported that the two genera differ in skull articulation, and absence of nasolacrimal duct bone density ( Gyrinophilus light, Pseudotriton impressions ally it with Gyrinophilus. Typhlotriton heavy) snout length (G. long, P. short), frontal differs from other members of its line, and re­ and toothed portions of premaxillae fused (P.) sembles Gyrinophilus, Pseudotriton, and Stereo­ or not ( G.), and nasals elongate and contacting chilus in having continuous vomerine tooth series. maxillae ( G.), or broad and separated from maxil­ Eurycea resembles Pseudotriton more closely lae (P.). My observations corroborate their state­ than Gyrinophilus. Eurycea and Pseudotriton share ments in large part, but the nasals are in contact several important characters, including fused pre­ with the maxillae in some Pseudotriton. Other sig­ maxillae, relatively short snouts, marked nasolacri­ nificant differences include presence of ventro­ mal duct impressions, and hypapophysial vertebral medial quadrate projections, nasolacrimal duct remnants. Lingual cartilages and parietal-otic and impressions on the anterodorsal prefrontal sur­ squamosal-otic crests, all present in Gyrinophilus, faces, and broad and extensive overlap of the pre­ Pseudotriton, and Stereochilus, are present in maxillary frontal processes by the nasals in Pseu­ Eurycea but absent in Typhlotriton. dotriton but not in Gyrinophilus. Typhlotriton and Eurycea are closely related, as Stereochilus resembles Pseudotriton in having is evident from the number of common characters, fused premaxillary toothed and frontal processes, including: (1) similar hyobranchial structure, de­ and hypapophysial remnants on some trunk verte­ spite presence of membranous anterior attachments brae. Moderate maxilla-nasal articulation and no in Typhlotriton, (2) greatly expanded first basi­ nasolacrimal duct impressions relate Stereochilus branchials, (3) short cornua, (4) well developed to Gyrinophilus. Numerous characters separate sublingual glands of unknown function which are Stereochilus from both genera. The snouts and poorly developed or absent in other genera, (5) heads are greatly compressed laterally, and this extremely short transverse processes, ( 6) complete compression is reflected in the extensive fusion of parapophysial and diapophysial separation, (7) the premaxillary frontal processes anterior as well parapophysial origin greatly in advance of dia­ as posterior to the fontanelles, ontogenetic closure pophysial origin, (8) ribs longer than distance of the intemasal fontanelles, and medial juncture across parapophyses, (9) elongate centra, (10) of the vomerine tooth series below the anterior well developed basapophyses. portions of the parasphenoids. Apparently Typhlotriton is related to the Gyri­ Despite its specializations Stereochilus is a rela­ nophilus-Pseudotriton-Stereochilus assemblage, but tively primitive form, not far removed in most it is closer to Eurycea. Many similarities between features from an idealized familial ancestral stock. Typhlotriton and Eurycea suggest probable descent One of the most primitive characters of Sterep­ from an ancestral stock related to that from which chilus is the presence in adults of a distinct, well Gyrinophilus and Pseudotriton were derived. Sep­ developed lateral line system on the heads. The aration of Typhlotriton and Eurycea from the more system is better developed than in any other primitive genera is based primarily on the striking plethodontid genus (see Hilton, 1947 b; 1950). hyobranchial and vertebral differences. Stereochilus diverged from the ancestral pletho­ Hemidactylium is a member of the Hemidacty­ dontid stock at a somewhat later date than the liini but its relationships to other hemidactyliine stock which gave rise to the desmognathines, but genera are not clear. The outstanding primitive it may have been one of the earliest hemidactyliine characters are retention of separated premaxillae, derivatives. detoglossy, and genioglossal muscles. Hemidacty­ Gyrinophilus, Stereochilus, and Pseudotriton lium has low numbers of trunk vertebrae, and is form the most primitive generic group in the fam· the only non-paedogenetic hemidactyliine that 64 MEMOIR OF THE SOUTHERN CAUFORNIA ACADEMY OF SCIENCES VOL.4 does not have a tendency for opisthocoely. It re­ distal tarsals and toes, in being smaller, in having sembles Gyrinophilus and allies in vertebral and smaller anterior cranial elements, in having pond hyobranchial characters, but resembles Eurycea in rather than stream larvae, and in being restricted having separated anterior and posterior vomerine to lowland habitats. The many striking similarities tooth series. The nasolacrimal ducts follow routes far outweigh the differences, however, and include similar to those in Eurycea, and their paths are similar hyobranchial structure, similar vertebrae, impressed on the prefrontals. Specialized features and similarly shaped and arranged cranial elements. of Hemidactylium include loss of the fifth toes and Cranial similarities include presence of relatively fifth distal tarsals, and presence of basally con­ short frontals and very distinct posterolateral stricted tails with concomitant vertebral modifica­ frontal processes. Both Eurycea and Manculus tions. lack well developed vomerine preorbital processes, Hemidactylium has a puzzling mosaic of primi­ but other hemidactyliines have the structures well tive and advanced characters indicating specializa­ developed. The posterior maxillary tips are rela­ tion of a relatively primitive stock. It is closer to tively short and sinuous in both, and as a result, the the plethodonine line than is any other hemidactyl­ subocular grooves intersect the lips below the eyes iine. H emidactylium may be a specialized derivative at the posterior maxillary tips. A connecting groove of a primitive stock ancestral to all hemidactyliines. is present in many genera, but actual continuation Despite the statements of Dunn (1926), Noble of the subocular groove to the lip occurs elsewhere (1931), Piatt (1935), and others to the effect only in Thorius. that Hemidactylium is a derivative or close rela­ Unquestionably Manculus is distinct from all tive of Plethodon, there can be little question con­ other Eurycea, but when differences are weighed cerning its closer relationship to the hemidactyl­ against similarities, it is apparent that Manculus is iine line. Bishop (1941) intimated that Hemi­ simply a specialized offshoot of the Eurycea line. dactylium was more distantly removed from Pletho­ No purpose is served by isolating the single species don than had formerly been suggested, but he did in a separate genus. Other genera which have lost not mention possible relationship to other genera. fifth toes (Hemidactylium, Batrachoseps) warrant The aquatic larval stage of Hemidactylium is generic recognition on the basis of marked differ­ clearly a primary, primitive character of great ences from other groups in numerous characters. importance, not a derived feature as suggested by It is recommended that the species quadridigitatus Noble (1927 a). be placed in the genus Eurycea, and that Manculus The tribe Hemidactyliini presents a mosaic pat­ be considered a subjective junior synonym of tern of primitive and advanced characters. It is ob­ Eurycea. vious that many ancestral and intermediate groups are no longer extant, and that parallel loss has Status of the genera Haideotriton and Typhlomolge occurred in regard to several characters. Following completion of this paper a new paedo­ genetic and troglobitic species, Eurycea tridenti­ Status of the genus Manculus fera, was described (Mitchell and Reddell, 1965). The genus Manculus, with a single species, M. The authors considered the new species to be in­ quadridigitatus, has been recognized as forming a termediate between Typhlomolge and Eurycea and distinct group by most authors since 1869, when suggested that the former be placed in the syn­ the genus was proposed by Cope. The sole excep­ onymy of the latter. Through the kindness of James tion was Dunn (1923), who suggested that the P. Bogart I have recently had an opportunity to single species be placed in the genus Eurycea. examine three specimens of E. tridentifera and one Dunn (1926) reiterated his position, stating, "I of E. troglodytes. Information concerning these cannot think the genus Manculus worthy of recog­ species is limited to this portion of the paper. My nition. The sole difference from Eurycea is the investigations, discussed below, indicate that the absence of the fifth toe." Dunn's suggestion was not genus Typhlomolge is distinct from Eurycea and followed and Noble (1927a) was soon recog­ contains two species (tridentifera and rathbuni). nizing Manculus; the genus has been recognized by Paedogenetic plethodontids occur in four general all subsequent authors. regions: 1) northern Alabama and southeastern No one bas questioned the close affinity of Tennessee (Gyrinophilus palleucus), 2) southwest­ Manculus and Eurycea. Noble (1931) frankly ern Georgia and north-central Florida (Haideo­ stated, "Manculus is a dwarf form of Eurycea • •. triton wallacei), 3) eastern Oklahoma (Eurycea It differs chiefly in the loss of the fifth (outer) toe." tynerensis), and 4) central and west-central Texas, Manculus differs from Eurycea in lacking fifth along the Balcones Escarpment and on the Ed- 1966 EVOLUTION OF LUNGLESS SALAMANDERS 65 wards Plateau (Typhlomolge rathbuni, T. tridenti­ margins, and well developed orbitosphenoid bones. fera; Eurycea latitans, E. nana, E. neotenes, E. In addition all differ from T. rathbuni and resemble pterophila, E. troglodytes). Evolution of paedo­ larval Eurycea in general skull dimensions and in genesis has apparently occurred independently in the relative development of skull elements. All each of these areas. G. palleucus is clearly assign­ have very well developed columellae, thus resem­ able to Gyrinophilus, and is not a close relative bling Typhlomolge rather than Eurycea. Mitchell of other paedogenes. Haideotriton shows as many and Reddell (1965), citing Burger, Smith and similarities to larvae of such forms as E. bislineata Potter (1956) and Baker (1957), state that the and E. longicauda as to the Texan paedogenes. E. second basibranchial (posterior basibranchiuln) is tynerensis seems to be more closely related to E. continuous with the hyobranchial apparatus proper multiplicata of the Interior Highlands than to the in T. tridentifera and T. rathbuni and separated Texas Eurycea. The Texas Eurycea form a quite from it in the other species; the elements are also distinct group and there is no clear evidence of continuous in my E. nana, E. pterophila and E. close relationship to any single species of trans­ troglodytes. formed Eurycea. T. tridentifera resembles the paedogenetic Texan The three epigeal species, E. nana, E. neotenes, Eurycea in general proportions of the anterior cran­ and E. pterophila, are the least specialized of the ial elements, but in few other characters. Most of Texan species and resemble generalized plethodon­ the skeletal similarities are in features common to tid larvae. Mitchell and Reddell (1965) have cor­ all plethodontine larvae. The otic capsules are rectly stated that specializations related to sub­ large and resemble those of T. rathbuni. Like T. terranean life become increasingly pronounced in rathbuni, the orbitosphenoids are absent, the diapo­ the order E. latitans, E. troglodytes, T. tridentifera, physes are long and extend beyond the lateral and T. rathbuni. Specializations include such fea­ margins of the zygapophyses, the numbers of tures as eye degeneration, skin depigmentation, trunk vertebrae are reduced (14 in two specimens, limb attenuation and elongation, snout depression, 13 in one), and the adpressed limbs overlap. Small truncation and elongation, and reduction in num­ alar processes are present on the anterior margins ber of trunk vertebrae. Mitchell and Reddell agreed of the parapophyses as in T. rathbuni. Such proc­ with Baker (1957) that each species has evolved esses are not found in other paedogenetic Eurycea. independently of the others since the original Basapophyses are absent; the processes are very colonization of the subterranean habitat by the well developed in E. pterophila, and are present ancestral stock, and that present similarities of the and usually well developed in the other Eurycea. species are likely due to convergent evolution. Basapophyses are very poorly developed in T. Unquestionably Typhlomolge rathbuni is the rathbuni. Several unique features occur in tridenti- most highly specialized and distinctive species of 1era. The columellae are longer than in any other the Texan group. The species is much larger, has plethodontid and have a distinct bow (the columel­ longer limbs, more extremely specialized head and lae are also very long in T. rathbuni) . In addition snout, more reduced eyes, and is more depigmented the columella may have a dorsolateral spur near the than any of the other species. In addition its dis­ opercular-columellar junction. The columellar tinctive skull is greatly flattened and broadened. process of the suspensorium is ossified (or heavily The toothed portion of the premaxilla is much calcified), a unique condition among pletbodontids. broader and the frontal processes are broader and There are only eight tarsals in the single tarsus more depressed than in any of the other species. available, a result of the fusion of distal tarsals The palatopterygoids are very widely separated four and five. Bye reduction reaches an extreme and are longer and more spatulate than in the other in T. rathbuni and Haideotriton wallacei, in which species. There are no orbitosphenoids. The pari­ lenses are absent. Lenses are absent in about half etals bear distinct crests, and the otic capsules are the tridentifera known (Mitchell and Reddell, disproportionately large. Transverse processes are 1965), but are present in other Texan paedogenes. relatively long and extend beyond the lateral mar­ T. tridentifera is much smaller than T. rathbuni, gins of the zygapophyses. There are only 13 or 14 has more pigmentation, lacks parietal crests, and trunk vertebrae. has a l~ extremely specialized snout (although The paedogenetic species of Eurycea (tridenti· snout flattening is more extreme than in any /era excluded) all differ from T. rathbuni and re­ Eurycea). semble nonpaedogenetic species of Eurycea in hav­ Larvae of non-paedogenetic plethodontids are ing more than 14 trunk vertebrae, diapophyses all basically very similar with only minor tub­ which extend no farther than the zygapophysial familial differences. Interpretation of the charac- 66 MEMOIR OF THE SOUTHERN CA.UFORNIA ACADEMY OF SCIENCES VOL.4 ters of paedogenetic species has been partially dis­ than T. tridentifera and approach those of T. rath­ cussed by Dundee (1962) who emphasized prob­ buni in size. The vomers are very broad and "L" lems of definitions of higher categories. Interpreta­ shaped, and the palatopterygoids are large and tion is especially difficult in the case of the Texan spatulate; both elements resemble those of T. rath­ paedogenes, between which greater differences are buni. The skull presents an almost rectangular ap­ found than occur between the larvae of distantly pearance, unique in the family, because of the related genera. The fact that morphoclines exist in greatly expanded snout and relatively small otic regard to several characters with one of the less capsules. specialized Eurycea paedogenes on one end and Dundee (1957) reported some success in arti­ T. rathbuni invariably on the other is not neces­ ficially inducing transformation in T. rathbuni, but sarily evidence that all have had a common ancestor the species metamorphosed to a lesser extent than or that these are chronoclines. No information paedogenetic Eurycea. Atrophy of labial folds, concerning selection pressures in the various en­ gills, and fins was noted. The gill slits remained vironments is available, but it is known that the open. Maxillae, maxillary teeth, and a small un­ surrounding terrain is inhospitable to transformed paired anterior ossification (septomaxilla or nasal?) hemidactyliines. In order to survive in the area any appeared. One of the specimens of T. tridentifera hemidactyliine would have been forced to remain in examined was kept alive by James P. ~ogart for the larval habitat or to enter the permanent waters five months and was treated with thyroxin as well of springs and underground caverns. The Texan as with pituitary implants. The gills are greatly re­ paedogenes are a diverse lot and need not have a duced and only the anterior slit remains open. Some common ancestor. It seems inadvisable to refer T. changes are apparent in the labial region. Skull rathbuni to Eurycea until more information con­ shape is virtually identical to that of smaller, un­ cerning many aspects of the biology of all the treated individuals despite the appearance of an species involved becomes available. elongate pars dentalis of the maxilla and a series of Mitchell and Reddell (1965) suggest that T. maxillary teeth. Small, paired nasals (?) are also tridentifera is intermediate between T. rathbuni present. Coronoids are present despite the fact and the known paedogenetic Texan Eurycea. It is that they are normally among the first of the larval obvious from the above discussion that tridentifera elements to disappear during metamorphosis. The is, in regard to most characters, much more similar margins of the palatopterygoids are eroded. Appar­ to T. rathbuni than to any other species. Because ently T. tridentifera transforms under the influence the diversity within the Texan paedogene group is of metamorphosing agents to about the same degree so great and because the vertebral and orbitosphe­ as T. rathbuni, and to a much lesser degree than noid characters enable the group to be conveniently paedogenetic Eurycea (E. nana, E. neotenes, E. divided, I choose to recognize the genus Typhlo­ tynerensis; Dundee, 1962). molge and refer to it the species rathbuni and In contrast to the paedogenetic species of Gyri­ tridentifera. The species E. troglodytes is the near­ nophilus and Eurycea, Haideotriton wallacei under­ est to an intermediate but it is closely related to goes only a few minor integumentary changes, other Eurycea and is intermediate only in being loses only a single bone (coronoid) and gains no largely depigmented, in having reduced eyes, and new bony elements when treated with metamor­ in having relatively elongate, attenuated limbs. phosing agents (Dundee, 1962). Dundee concludes In some ways the genus Haideotriton is more that Haideotriton, like Typhlomolge and Necturus, similar to the larvae of non-paedogenetic Eurycea appears to be genetically resistant to thyroid secre­ than are either T. rathbuni or T. tridentifera. The tions. He states further that H. wallacei is the most vertebrae of Haideotriton, like Eurycea, have short resistant to thyroxin of all plethodontids investi­ diapophyses and lack parapophysial alar processes. gated. Haideotriton is clearly a close relative of Limbs are not nearly as long as in Typhlomolge. Eurycea. Because the genus is biologically distinct H. wallacei resembles Typhlomolge in having poor­ from all others and because there is no evidence ly developed basapophyses, 14 trunk vertebrae, that it evolved from a direct ancestor either of lenseless eyes, in lacking orbitosphenoid bones, Typhlomolge or of the paedogenetic Eurycea, the and in being greatly depigmented. The snout of genus Haideotriton is maintained. H. wallacei is more truncate and the otic capsules Mitchell and Reddell (1965) have correctly are smaller than in Typhlomolge. The skull is rela­ noted the lack of information concerning evolu­ tively much larger, both longer and broader, than tionary rates and selection pressures in subterra­ any Eurycea. The toothed portions of the premaxil­ nean habitats. Until such information becomes lae are much larger than those of any Eurycea or available it is not advisable to assume that all the 1966 EVOLUTION OF LUNGLESS SALAMANDERS 67 subterranean plethodontids of the Edwards Plateau unexpanded anterior first basibranchial extensions) entered the underground environments approxi­ and in vertebral characters (fused hyperapophy­ mately at the same time. It is more likely that those ses, lack of caudal transverse processes, shorter with the more extreme adaptations entered the ribs). environments first. Perhaps the ancestral Typhlo­ Plethodon ·and Ensatina resemble each other molge stock entered the subterranean habitats at a and differ from Aneides in sharing such generalized relatively early date, and the Eurycea somewhat characters as paired premaxillae, unmodified later. Anatomical evidence indicates that both gen­ maxillae, relatively weak prefrontal-maxilla articu­ era evolved from the same general evolving ances­ lations, relatively unmodified dentition, and primi­ tral stock but that Typhlomolge was derived from tive tarsal and carpal configurations. Plethodon has a pre-Eurycea ancestor and the other Texan paedo­ lower numbers of maxillary, dentary, and vomer­ genes from a Eurycea or very Eurycea-like ances­ ine teeth than Ensatina, and has a great amount tor. Haideotriton seems not to be closely related of vertebral number variation. Plethodon is a gen­ to Typhlomolge. It probably originated from a eralized group, and only a few specializations oc­ Eurycea-like stock in the Southeast. cur, including elongation of certain species Comments concerning Typhlomolge found else­ ( cinereus, elongatus, neomexicanus, richmondi), where in this paper are based on study of T. rath­ unusual otic crests (elongatus ), and reduction in buni and do not necessarily refer to T. tridentifera. fifth toe phalangeal number (larselli, neomexi­ canus). Relationships of the plethodonines Aneides is so similar to Plethodon that it is Ensatina, Plethodon, and Aneides are a compact tempting to postulate direct origin from Plethodon. group of closely related genera which form the Primitive species of Aneides (especially female tribe Plethodontini. I have discussed their relation­ hardii) differ from Plethodon only in having fused ships recently (Wake, 1960; 1963), and only a re­ premaxillae and modified tarsal configurations. Al­ view is presented here. most all of the many specializations which occur in The plethodonine genera appear to have been the advanced species of Aneides are either feeding derived from a relatively primitive, common ances­ or climbing adaptations. Advanced species of tral stock. Plethodon is the central and most gen­ Aneides differ sharply from Plethodon because of eralized genus, and the other two are specialized these specializations, which include reduction in offshoots. Ensatina is a primitive genus, and was number but increase in size of all teeth, sabre-like, probably derived from a Plethodon-Aneides an­ enlarged, unicuspid maxillary and dentary teeth, cestral stock. It is more closely related to Plethodon enlarged otic crests that serve as points of origin than to Aneides. Aneides resembles Plethodon of enlarged cephalo-hyo-mandibulare musculature, much more closely than it does Ensatina. Char­ specialized interlocking maxilla-prefrontal articula­ acters which relate Aneides to Ensatina are those tions, loss of the vomerine preorbital processes, characteristic of the tribe and are shared by Pletho­ greatly elongated limbs, bifurcated and decurved don as well. Aneides is the most specialized and terminal phalanges, greatly enlarged, edentulous advanced of the genera, and is closely related to posterior maxillae with cleaver-like cutting edges, Plethodon. My ideas concerning plethodonine re­ hypertrophied quadrato-pectoralis musculature and lationships are illustrated in Figure 15. prehensile tails. Primitive features of Ensatina include paired Despite the multitude of differences separating premaxillae (shared with Plethodon), very high advanced species of Aneides from Plethodon, the numbers of maxillary, dentary, and vomerine teeth, presence of such primitive species as A. hardii very long vomerine preorbital processes, low num­ clearly reveals close relationship of the two genera bers of trunk vertebrae, poorly developed otic and suggests origin from either a common ancestral crests, and relatively large otic capsules. Special­ izations found in Ensatina include basally con­ stock or of origin of Aneides from Plethodon. stricted tails, shortened first caudal vertebrae, /tezationships of the bolitoglossines elongated limb elements, especially fore limbs, ac­ quisition of considerable tongue freedom, rather Bolitoglossine salamanders form three groups elongated genioglossal muscles, and possibly prim­ sufficiently distinct to warrant recognition of super­ ordial lingual cartilages. Ensatina differs further genera. These groups are the supergenus Hydro­ from Aneides and Plethodon in details of hyo­ mantes with a single genus, the supergenus Bat­ branchial construction (shorter cornua, Ja~set._e,pi­ rachoseps with a single genus, and the supergenus branc'bials and second basibranchials, ' slftall and Bolitoglossa with seven genera (Bolitoglossa, Oedi.. 68 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL. 4

Pl et hod on tin i

Figure 15. Diagrammatic representation of the relationships of the plethodonine genera. 1966 EVOLUTION OF LUNGLESS SALAMANDERS 69 pina, Pseudoeurycea, Chiropterotriton, Parvimolge, supergenera, and advanced characters found in at Lineatriton, Thorius). least several genera include septomaxillary, maxil­ The supergenera Bolitoglossa and Batrachoseps lary tooth, and columellar loss, tendencies toward share several important characters that are absent opisthocoely, acquisition of basapophyses, distinct in Hydromantes, including fused premaxillae (with tail constrictions, phalangeal modifications, and fused frontal processes in some species of each mesopodial fusions and reductions. Advanced group), similar basibranchial comua, posterolat­ characters shared with one or the other of the eral parietal spurs, columellar reduction, and bolitoglossine supergenera include bilobed testes, multilobed testes. James Kezer has told me (in reduced numbers of chromosomes, tendencies for litt.) that Batrachoseps and the supergenus Bolito­ unicipital ribs, fused premaxillae, and adetoglossy. glossa (he has not examined Parvimolge) have Hydromantes is apparently closer to the bolitog­ haploid chromosome numbers of thirteen, while lossine ancestral stock than are either of the other Hydromantes and all other plethodontids examined two supergenera. Both Hydromantes and Batra­ have fourteen. choseps are more closely related to the advanced One species of Batrachoseps (B. wrighti) has Bolitoglossa than to each other, but Batrachoseps paired premaxillae in adults and is similar in that may be closer to Bolitoglossa than is Hydromantes respect to Hydromantes, but the evolutionary trend (see Fig. 16). in Batrachoseps is in the direction of premaxillary fusion, as in the supergenus Bolitoglossa. Hydro­ mantes and Batrachoseps resemble one another, and differ from supergenus Bolitoglossa, in having three rather than two caudosacral vertebrae and in the universal presence of septomaxillae. How­ ever, two caudosacral vertebrae occur in at least some individuals of all species of Batrachoseps, and in all members of three species. Hydromantes is very similar to certain members of the supergenus Bolitoglossa (e.g., Bolitoglossa) in habitus and external morphology. In addition Hydromantes and the neotropical genera resemble each other, and differ from Batrachoseps, in having truly free tongues with no genioglossal muscles, low numbers of trunk vertebrae, tendencies for uni­ cipital ribs and loss of diapophyses, loss of omo­ hyoideus muscles, and five rather than four toes. Batrachoseps and Hydromantes are relatively primitive, and retain such primitive features as Bolitoglossini three caudosacral vertebrae (some individuals) and paired premaxillae (B. wrighti). Batrachoseps is highly specialized, however, and is elongated with increased numbers of trunk vertebrae, multi­ Figure 16. Diagrammatic representation of the relation­ lobed testes, tendencies for fusion of the premaxil­ ships of the bolitoglossine supergenera. lary toothed and frontal processes, loss of pre­ frontals, reduced frontals and parietals, well Tanner (1952) presented a diagrammatic family developed parietal spurs, and no fifth toes and fifth tree of the free-tongued plethodontids, and dis­ distal tarsals. It retains comua, genioglossal mus­ cussed relationships of these genera. His study was cles, and modified detoglossy, all relatively primi­ based on throat musculature and hyobranchial tive features, but Hydromantes has lost both comua cartilage structure, and he assumed that the free­ and genioglossal muscles and has adetoglossy. tongued genera formed a natural group. Tanner Hydromantes is the most generalized of the three suggested than an ancestral .stock had given rise to supergenera, retaining such primitive characters as two diverging generic groups, one containing Gyri­ unilobed testes, relatively large and separated nophilus, Pseudotriton, Eurycea, and Manculus, premaxillae with relatively attenuated frontal and the other containing Hydromantes and the neo­ processes, large columellae, and spurless parietals. tropical genera. Hydromantes was considered to be Bolitoglossa is the most advanced of the three close to Bolitoglossa and Magnadigita, but the 70 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 other neotropical genera were placed in a separate Oedipina and Bolitoglossa are offshoots in dif­ line. Within the latter group three divergent evolu­ ferent directions from the ancestral stock that gave tionary lines were recognized: (1) Thorius, (2) rise to the relatively generalized members of the Oedipina, (3) Parvimolge, Pseudoeurycea, Chiro­ Thorius group. Bolitoglossa is the more primitive pterotriton, Lineatriton. Tanner stated that he con­ of the two, and both contain highly specialized sidered association of Batrachoseps with Hydo­ species. Specialized characters of Bolitoglossa in­ mantes and the supergenus Bolitoglossa, as had clude loss of the septomaxillae and columellae by been proposed by Piatt (1935), to be entirely un­ most species, and the prefrontals and tibial spurs warranted. by many, presence of basapophyses in many spe­ Although I strongly disagree with Tanner's con­ cies, vomerine reduction in many species, tenden­ clusions concerning relationships of the neotropical cies for increased amounts of hand and foot web­ genera to other plethodontids, our views of rela­ bing, and presence of mesopodial fusions. Oedipina tionships within the neotropical group are similar, has a number of specialized and advanced char­ with one notable exception in the case of Thorius. acters, including high vertebral numbers, many Salamanders of the supergenus Bolitoglossa pre­ unicipital ribs, no septomaxillae (most species), sent a bewildering mosaic of characters and an prefrontals, columellae, or tibial spurs, the most extremely complex evolutionary pattern. Three specialized and advanced premaxillae in the family, major generic groups may be recognized: (1) the and extensive mesopodial fusions and phalangeal Bolitoglossa group with one genus and about fifty losses. species, (2) the Oedipina group with one genus The Thorius group of genera differs from Boli­ and about sixteen species, ( 3) the Thorius group toglossa and Oedipina in that distinct tendencies with five genera (Thorius, Pseudoeurycea, Chiro­ for intervertebral cartilage calcification are evident pterotriton, Parvimolge, Lineatriton) and about and transitional stages between amphicoely and forty-six species. opisthocoely occur. Included in the Thorius group Bolitoglossa differs from the other genera of the are the most generalized and primitive species of supergenus Bolitoglossa in lacking sublingual folds. the supergenus Bolitoglossa, but other species are The folds are advanced characters and are related among the most highly specialized and advanced in to the elongation of the ceratohyals into the an­ the family. Primitive Chiropterotriton retain such terior portions of the mouths. Within the Bolito­ generalized and primitive features as septomaxillae, glossini, Bolitoglossa, Hydromantes, and Batracho­ prefrontals, lingual cartilages, well developed colu­ seps are the only genera that lack the folds. mellae, vomerine preorbital processes, unfused pre­ Wake and Brame (1963a, 1963b) and Brame maxillary frontal processes, tibial spurs, nine tar­ and Wake (1963) have presented evidence that sals, and eight carpals; but primitive species have the genera Magnadigita and Bolitoglossa in the advanced tarsal arrangements. The most advanced sense of Taylor (1944) are not recognizable and Chiropterotriton may lack septomaxillae, pre­ have assigned all species to Bolitoglossa. It has frontals, columellae, lingual cartilages, vomerine been possible to examine skeletons of only about preorbital processes, and tibial spurs, and may have half of the species but all sections of the genus are fused premaxillary frontal processes, seven car­ represented. A number of osteological characters pals, and eight tarsals in a primitive configuration. that vary interspecifically have been discovered, but All members of the genus retain bicipital ribs. Ad­ none coincide with the old generic division based vanced species have strongly opisthocoelous verte­ on amount of webbing (complete in Bolitoglossa, brae. less extensive in Magnadigita). Detailed discussion Primitive Pseudoeurycea have prefrontals, lin­ of osteology and evolution of Bolitoglossa is de­ gual cartilages, reduced columellae, vomerine pre­ ferred to a later date, but it is clear that a mosaic orbital processes, unfused premaxillary frontal distribution of characters occurs. Species groups processes, tibial spurs, eight carpals, nine tarsals, which bridge the old genera are recognizable, but primitive tarsal arrangements, and bicipital ribs. only on internal characters; none warrants generic These features are common to most species. Septo­ rank. maxillae are absent in most species. No Pseudo­ Oedipina differs from all other genera of the eurycea have truly opisthocoelous vertebrae, but supergenus Bolitoglossa in having eighteen to some intervertebral cartilage calcification occurs. twenty-two rather than fourteen trunk vertebrae. Pseudoeurycea and Chiropterotriton are very The marked elongation of the trunks and tails are closely related, and can be separated consistently specializations related to the evolution of special­ only on the basis of tarsal structure. Primitive and ized behavior and ecology of Oedipina. most advanced species of Chiropterotriton have 1966 EVOLUTION OF LUNGLESS SALAMANDERS 71 specialized tarsi in which the fifth distal tarsals are able, and final assignment of the three species re­ larger than the fourth and articulate with the cen­ mains uncertain. I have not examined skeletal ma­ trale. All Pseudoeurycea retain primitive tarsi in terial of P. praecellens (known only from the which the fourth distal tarsals are larger than the unique type). P. townsendi has been considered to fifth, and articulation of the fourth tarsals with be very closely related to Pseudoeurycea by Tanner the fibulars eliminates the fifth tarsals from articu­ ( 1952) on the basis of similarities in throat muscu­ lation with the centrals. Some advanced species of lature, and Uzzell (1961) has suggested that the Chiropterotriton (xolocalcae, bromeliacia) have the species is closely related to Lineatriton because of primitive configuration encountered in Pseudoeu­ similarities in hyoid and mesopodial calcification rycea but they can be distinguished from all patterns. P. townsendi resembles Chiropterotriton Pseudoeurycea because they lack prefrontals. Prim­ in having septomaxillae and primitive species of itive Chiropterotriton can usually be distinguished both Chiropterotriton and Pseudoeurycea in having from Pseudoeurycea by having septomaxillae, but such generalized features as columellae (reduced) , advanced species lack septomaxillae and some spe­ bicipital ribs, prefrontals, vomerine preorbital proc­ cies of Pseudoeurycea may have the structures. esses, and tibial spurs. The species differs from Tanner (1952) found differences in the details of both genera in being dwarfed, in having distinc­ hyobranchial structure between the genera, but I tively shaped hands and feet (see Taylor, 1944), cannot corroborate his observations. Despite the in having distinctive, large, dorsal glands, and in fact that primitive members of Chiropterotriton having the fourth and fifth distal tarsals fused. The and Pseudoeurycea are very similar, it is apparent species differs further from most Pseudoeurycea in that two groups are represented and the two genera having septomaxillae, and from all members of that are retained. genus in having eight rather than nine tarsals. Tanner (1952) considered Pseudoeurycea to be The species richardi was placed in Parvimolge the most primitive of the neotropical plethodon­ by Taylor (1949), presumably on the basis of hand tids. Pseudoeurycea and Chiropterotriton share and foot shape. Rabb (1955) described P. praecel­ several primitive and generalized characters, but lens, including it in Parvimolge because it shared each retains primitive characters reduced or lost in several features with P. townsendi, including simi­ the other. As a genus Chiropterotriton has radiated lar feet, coloration, conspicuous dorsal glands, and and specialized to a greater extent than has Pseudo­ tail shape. Dorsal glands are absent in richardi, and eurycea, but primitive members of both genera are Rabb suggested that absence of the glands, differ­ almost equally primitive. H the total species com­ ences in foot structure, and enormous range dis­ plements are considered, the largely ground-dwell­ junction {richardi in Costa Rica, praecellens and ing, morphologically relatively uniform Pseudoeu­ townsendi in Veracruz, Mexico) imply that rich­ rycea must be considered more primitive than the ardi is not congeneric with praecellens and town­ largely arboreal, morphologically diverse Chirop­ sendi. terotriton. Primitive members of the two genera are Septomaxillae, reduced columellae, lingual carti­ very closely related and have diverged only slightly lages, and mesopodial and hyobranchial calcifica­ from common ancestral stocks. tions all occur in P. townsendi, but not in P. The three remaining genera of the Thorius group richardi. Nothing is known concerning the meso­ apparently diverged from a pre-Pseudoeurycea­ podial pattern in richardi, but there is no reason Chiropterotriton stock and have become highly why richardi cannot be placed in Chiropterotriton specialized. Parvimolge and Lineatriton are prob­ despite some differences in external foot morpho­ ably closer to each other than either is to any other logy. It is true that the characters lacking in richardi genus, but the two are also relatively closely related but present in townsendi are those often lost in to both Pseudoeurycea and Chiropterotriton. southern species of Chiropterotriton. This fact The single species of Lineatriton is an elongate might support the suggestion that richardi is a mem­ form that has unicipital ribs, elongated vertebrae ber of Chiropterotriton, but it is also possible that with only parapophyses on most vertebrae, and the same losses have occurred in parallel in Parvi­ noticeably shortened limb elements. Septomaxillae molge. Until more information concerning the spe­ have been lost. The genus is somewhat closer to cies presently assigned to Parvimolge becomes Pseudoeurycea than to Chiropterotriton, but it is available the problem is insoluble, and richardi is distinct from both. Resemblance to Oedipina is tentatively retained in the genus. probably the result of parallel evolution. Thorius probably represents the earliest diver­ The genus Parvimolge is probably an unnatural gent stock of the line that leads to Pseudoeurycea assemblage. Adequate material has been unavail- and Chiropterotriton (Fig. 17). Dwarfing is severe 72 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 in the genus, and the cranial elements are greatly preted as compensations for paedomorphic influ­ reduced in size and modified in appearance. The ence that has led to general skeletal dwarfing and genus is further characterized by such specializa­ reduction (see below, Paedomorphosis). Calcified tions as loss of columellae and vomerlne preorbital mesopodials have been considered diagnostic of the processes, variable loss of septomaxillae, high fre­ genus since the time of Cope (1869), but the quency of mesopodial calcifications, fusion of distal character lacks reliability. The largest Thorius tarsals 4 and 5, intervertebral cartilage calcification, that I have examined (28.2 mm.) has uncalcified and presence of posteriorly directed squamosal mesopodials (see also Uzzell, 1961). There is no processes. Primitive characters retained by the justification in removing Thorius from other neo­ genus include bicipital ribs, fourteen trunk verte­ tropical genera and placing it in its own subfamily, brae, and tibial spurs. Highly modified prefrontals the Thoriinae, as Cope (1893), and more recently are usually present. This mosaic pattern of general­ Smith and Taylor (1948), have suggested. I also ized and specialized characters indicates that mod­ consider the genus less distinct from other neo­ ern Thorius is the specialized end point of a line tropical genera than did Tanner (1952), and think derived from a stock ancestral to other genera of that Thorius is closer to Pseudoeurycea, Chiropter­ the Thorius group. Calcification and ossification of otriton, Parvimolge, and Lineatriton than are such elements as the ends of long limb bones, inter­ either Oedipina or Bolitoglossa. vertebral cartilages, and mesopodials may be inter- To briefly summarize, the tribe Bolitoglossini

() <:- ...... Cb Cb ~ 0 ,o ..... ,o. ~ C> ,o"' ...... 0 Cb .... 0 .... ·~ () ·~ " .~"' . o°' ~ ., Cb ~ ~ ... $" ... Cb ·~ 0 b ·~ Q..q, Cb t:l'Jo " Q..C> iJ ~ c

Figure 17. Diagrammatic representation of relationships within the supergenus Bolitoglossa. contains three distinct lines. The ancestral stock group Pseudoeurycea and Chiropterotriton are the has given rise to the relatively unspecialized most primitive genera, and Thorius is the most Hydromantes. Two main specialized branches specialized and advanced. Thorius probably rep­ have resulted in the related supergenera Batracho­ resents the first divergence from the group ancestral seps and Bolitoglossa (see Fig. 16). Within the stock, and Lineatriton and Parvimolge have di­ supergenus Bolitoglossa there has been a further verged somewhat more recently. Bolitoglossa and divergence, with the Thorius group of genera Oedipina represent offshoots in different evolu­ closest to the ancestral stock. Within the Thorius tionary directions from the ancestral stock of the 1966 EVOLUTION OF LUNGLESS SALAMANDERS 73 Thorius generic group. Oedipina is the most spe­ dontid parasphenoid teeth were ontogenetically de­ cialized and advanced plethodontid genus. rived from the anterior vomerine teeth. The vom­ erine arrangement of plethodontids, Dunn argued, Trends could most easily have been derived from the pos­ Ecology and Life History teriorly directed vomerine tooth series typical of salamandrids. In 1926 Dunn included six families The ancestral adaptive zone (Proteidae, Ambystomatidae, Pleurodelidae, Sala­ Simpson {1953: 199) states that at any instant mandridae, Plethodontidae, Amphiumidae) in the in time environments and organisms of a given suborder Salamandroidea. He declared that "the taxonomic level define a field or type of adaptation Plethodontidae can be most easily derived from characteristic of the group. These adaptive types the Salamandridae." Dunn's ideas have been wholly can be visualized if represented on paper as bands accepted by most recent workers (e.g., Noble, or zones that form sorts of grids (Fig. 18). Zone 1931; Piatt, 1935; Herre, 1935; Tanner, 1952; breadth is a reflection of general adaptation; the Auffenberg, 1961). Noble (1931) added one mod­ broader the zone, the more. generally adapted are ification - the ambystomatids were placed in a the organisms. Such zones involve abstraction and separate suborder, the Ambystomatoidea. Since oversimplification but are useful in explaining evo­ that time salamandrids and plethodontids have lutionary patterns and pathways. Simpson empha­ continued to be associated in the same suborder, sizes that such a zone is not geographic, physical, and ambystomatids have been placed in a separate or even environmental in character, but "an adap­ suborder (Goin and Goin, 1962). tive zone, representing a characteristic reaction and Laurent (1947) thought the vomerine tooth pat­ mutual relationship between environment and or­ terns observed in salamandrids and plethodontids ganism, a way of life and not a place where life is represented separate evolutionary derivations, and lived." I conceive of an adaptive zone as the sum he suggested that both conditions could have total ecological niche relationships of a group of evolved from an ambystomatid type. He also cited organisms with the same base level of general similarities in the vertebrae of plethodontids and adaptation. Thus modem reptiles can be said to ambystomatids (primitively amphicoelous versus occupy a terrestrial adaptive zone in a broad sense, opisthocoelous in salamandrids), presence of septo­ because of the general adaptational level associated maxillae and costal folds in both families (absent with amniotic egg evolution. Birds, close reptilian in salamandrids), and absence of paraoccipital relatives, have entered a new aerial adaptive zone processes and fronto-squamosal arches in both as a result of a base level of general adaption (in (both present in primitive salamandrids). Laurent the broadest sense) associated with the evolution concluded that the suborder Salamandroidea of of feathered flight. In order to establish the ances­ recent workers was an artificial assemblage, and tral adaptive zone of plethodontid salamanders it suggested that the Plethodontidae be housed in a is essential to briefly consider the relationships of third suborder, the Plethodontoidea. Laurent's Am­ the family to other salamander families, and the bystomatoidea includes only the Ambystomatidae, nature of the ancestral plethodontid stock. the Plethodontoidea includes only the Plethodonti­ dae, and the Salamandroidea includes the Sala­ Plethodontid origins mandridae, Amphiumidae, Proteidae, and Sireni­ Cope (1866, 1867, 1869, 1889) considered dae. As an alternative Laurent suggests inclusion of plethodontids to be closely related to and possibly all of the above families in a single suborder, the derived from ambystomatid salamanders. Stejneger Salamandroidea. (1907) placed the families Ambystomatidae (in­ Teege (1956) found that plethodontid vertebrae cluding hynobiids), Salamandridae, and Pletho­ were more similar to those of ambystomatids than dontidae (including desmognathines) in a super­ to salamandrids. She was unable to find supporting family Salamandroidea, a group first proposed by data for Dunn's theory. Larsen (1963) found little Sarasin and Sarasin (1887-1890) for all salaman­ to support salamandrid origin of plethodontids ders except amphiumids. Dunn ( 1922) completed based on his study of cranial morphology, and the transition from Cope's ambystomatid theory suggested that plethodontids and ambystomatids by stating "it is probable that some primitive sala­ were rather closely related. Both groups were mandl\d . . . gave rise to the much degenerate thought to have arisen from a prohynobiid an­ Plethodontidae." His statement was based on a cestor. Recently Monath (1965) has shown that study of the otic apparatus, and on the work of the opercular apparatus in plethodontids is basic­ Wilder (1920) which demonstrated that pletho- ally quite distinct from that of salamandrids. He 74 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 has suggested that plethodontids were not derived Dunn (1917), Wilder and Dunn (1920), and from salamandrids, but from a primitive sala­ Dunn (1926). All plethodontids are lungless and mander stock. lack ypsiloid processes on the pelvis. These losses During the course of this study I have had oc­ are presumed to be rheotropic adaptations. The casion to investigate salamandrid, hynobiid, and reasoning is that lungs are hydrostatic organs and ambystomatid anatomy. In addition to data pre­ disadvantageous for stream bottom dwellers. Ypsi­ sented by Laurent, Teege, Larsen and Monath, I loid processes aid in emptying the lungs, and are have noted similarities in the throat musculature of little value in lungless forms. Wilder and Dunn of salamandrids and hynobiids, and in plethodon­ (1920) reviewed literature reports and showed tids, ambystomatids, and hynobiids. Ambystoma­ that lunglessness or greatly reduced lungs are asso­ tids and plethodontids also resemble one another, ciated with mountain brook environments in and differ from salamandrids, in hyoid and meso­ European salamandrid genera (Euproctus, Sala­ podial structure. Larval and adult palatal and vom­ mandrina, Chioglossa). The list of nonplethodontid erine structure of plethodontids is much more lungless forms should also include the ambystoma­ similar to that observed in hynobiids and amby­ tid genus Rhyacotriton of northwestern North stomatids than that in salamandrids. America, and the Asian hynobiid genus Onycho­ I recognize four suborders in the order Caudata. dactylus; both are mountain brook forms. Because The suborder Cryptobranchoidea includes the fam­ the entire family Plethodontidae is lungless, Wilder ilies Cryptobranchidae and Hynobiidae. Members and Dunn reasoned that the species have had a differ from the other suborders in having external common lungless ancestor, and that this ancestor fertilization and in retaining distinct angulars in likely inhabited mountain brook environments. the lower jaws. The suborder Sirenoidea includes The southern Appalachian region (Appalachia) only the family Sirenidae, a group recently raised contains the greatest diversity of pletbodontids, to ordinal level as the Order Trachystomata by including all of the mountain brook genera and the Goin and Goin (1962). Sirenids differ from other most primitive members of the family. Appalachia salamanders in lacking hind limbs and in the pecu­ is a very ancient upland, and presumably this rela­ liar structure of their vertebrae (see Goin and tively stable area has been the region of origin, Auffenberg, 1958), but resemble salamanders in the primary distribution center, and the dominant cranial morphology, particularly of the palatal ele­ area of survival of the family Plethodontidae. ments, and in hyoid morphology. I agree with The ancestral adaptive zone can be characterized Estes ( 1965) in considering ordinal status un­ as a way of life in semiaquatic, montane, warm warranted. The suborder Ambystomatoidea in­ temperate niches such as probably existed in late cludes the families Ambystomatidae and Pletho­ Mesozoic times in what is now eastern North dontidae, and differs from the remaining suborder, America. the Salamandroidea (Salamandridae, Proteidae, Necturidae, and questionably Amphiumidae), in Evolution within the adaptive zone that primitive ambystomatoid genera have vomer­ The Desmognathinae and the Hemidactyliini ine preorbital processes which bear laterally ori­ have remained primarily within the ancestral adap­ ented series of teeth, and have septomaxillae. tive zone, but the Bolitoglossini and Plethodontini My hypothesis of the present state of salamander both have occupied new terrestrial adaptive zones. evolution is that the modern and progressive sub­ In addition there are tendencies in desmognathine orders Ambystomatoidea and Salamandroidea rep­ and hemidactyliine genera toward splitting the old resent separate evolutionary lines that have di­ zone into subzones and toward moving out of the verged independently from an ancestral stock zone (see Fig. 18). similar to modem hynobiids. Ambystomatoids Within the genus Desmognathus the most im­ dominate the New World, and salamandroids domi­ portant trend obviously is the one leading in the nate the Old World, but representatives of both direction of increased terrestriality (see below) , but occur in both regions. primitive species have remained in the ancestral The foregoing account has been necessary be­ adaptive zone. Phaeognathus is probably a terres­ cause most previous theories concerning pletho­ trial species that has left the old adaptive zone. dontid ancestry have assumed origin from sala­ Leurognathus bas remained in the ancestral zone, mandrid stocks. Despite their apparent error in but selection has resulted in special adaptation and establishing plethodontid ancestry, the most ac­ in restriction to a specialized segment of the an­ ceptable theory concerning the ancestral adaptive cestral environment - the mountain stream bot­ zone (their ancestral home) was advanced by toms. Martof (1962) has recently considered in 1966 EVOLUTION OF LUNGLESS SALAMANDERS 75 some detail the ecology of Leurognathus. He states gnathus should not be fitted into the predominant that the genus is restricted to cool mountain streams adaptive sequence toward terrestrialism; the evo­ and that only during times of droughts do Leuro­ lutionary trend in the genus has been from semi­ gnathus and D. quadramaculatus (the most aquatic aquatic to strictly aquatic existence. Desmognathus) utilize a common habitat. Leuro- The tribe Hemidactyliini has remained closer to

Figure 18. Adaptive zones and subzones in the family Plethodontidae. la, semiaquatic zone; lb, paedogenetic subzone; le, aquatic subzone; ld, swamp-bog subzone; 2a, terrestrial zone; 2b, arboreal subzone; 2c, semifossorial subzone. the ancestral adaptive zone than any other group icauda inhabits streams and springs, but also the of plethodontids. Evolution within the group has twilight zone of caves, and E. lucifuga is essentially been characterized by movements out of the ances­ limited to the twilight zones of limestone caves tral mountain brook environment into closely ad­ (Hutchison, 1958). Several species of Eurycea are jacent environments and by restriction to special­ paedogenetic and have adapted to specialized ized aquatic environments. Gyrinophilus and aquatic situations. E. tynerensis, E. nana, E. neo­ Pseudotriton have remained very close to the an­ tenes, and E. pterophila inhabit springheads and cestral environment and occur in spring and stream cold swift streams. E. latitans occurs in cave waters, habitats. A single species, Gyrinophilus palleucus, and E. troglodytes occurs in the waters of a large is a paedogenetic form in subterranean waters in solution cave on the Edwards plateau of Texas Tennessee and Alabama. (Baker, 1957; 1961). These species occur only at Eurycea has become somewhat diversified. Prim­ the periphery of the generic range. Specialization itive species such as E. aquatica, E. bislineata, and into restricted portions of the ancestral habitat or E. multiplicata inhabit brooksides and springs, sit­ into specialized adjacent habitats has allowed survi­ . uations that resemble the ancestral habitat. E. long- val in areas beyond the general generic range. 76 MEMOIR OF THE SOUTHERN CAUFORNIA ACADEMY OF SCIENCES VOL.4 Typhlomolge, Haideotriton, and Typhlotriton mognathinae are particularly instructive in eluci­ are all close relatives of Eurycea. All are found in dating how occupation of new adaptive zones may areas peripheral to Appalachia in specialized have occurred. Although primitive desmognathines aquatic habitats. All are unpigmented, blind, and remain in the primitive adaptive zone, a significant subterranean and have apparently survived in diversification bas occurred within the subfamily. rather unfavorable climatic regions following cli­ Occupation of a new aquatic adaptive subzone by matic change by adapting to specialized, relatively Leurognathus has been discussed above. A ter­ very stable, underground habitats. restrial trend is evident in Desmognathus. Organ E. quadridigitata is more terrestrial than other (1961 a) recently completed a significant analysis Eurycea, and is found in low swampy areas, trickles of the ecology of five species of Desmognathus from springs, and sphagnum bogs (Bishop, 1943; which occur in a variety of habitats on Whitetop Carr, 1940). Hemidactylium, Stereochilus, and E. Mountain, Virginia. The most aquatic and general­ quadridigitata have entered pond and swamp habi­ ized is D. quadramaculatus, a species that inhabits tats, and this movement is associated with occu­ streams and stream banks. D. monticola represents pancy of new adaptive subzones (Fig. 18). Appar­ the next evolutionary stage and is found on stream ently three separate and parallel evolutionary lines banks and in seepage areas, habitats more terrestrial are represented. All breed in ponds and swamps in than those of quadramaculatus. D. fuscus is slightly relatively quiet waters. The larvae resemble each more terrestrial than monticola, and D. ochro­ other and differ from all other plethodontids, in­ phaeus is unquestionably more terrestrial than cluding sympatric species, in possessing relatively fuscus. D. wrighti is the terminal and most ter­ high dorsal fins that extend onto the trunks of the restrial species within the series; it is primarily a animals (Bishop, 1941; Goin, 1951; Schwartz and forest floor form with only breeding females and Etheridge, 1954). Goin compared the rudimentary hibernating individuals ever found in aquatic sites. hind limbs of hatchlings of H emidactylium with The species arranged in order of increasing ter­ those of E. quadridigitata and noted that both have restriality are quadramaculatus-monticola-fuscus­ toeless lobes at early stages. Both have four toes ochrophaeus-wrighti. Arranged in order of decreas­ as adults, and that fact may be related to the slow ing larval period, the species are quadramaculatus­ rate of development of the hind limbs in larval f uscus-monticola-ochrophaeus-wrighti; wrighti has stages. Hind limbs of Stereochilus larvae appar­ direct terrestrial development with no free-living ently are better developed than the others. Move­ larvae. Organ presented survivorship curves which ment out of the stream and into pond environments show that male survivorship increases in the series has been accompanied by acquisition of pond-type quadramaculatus - monticola - fuscus-ochrophaeus­ larvae. Significantly all have ranges that are periph­ wrighti, a trend identical to the aquatic-terrestrial eral to those of the primitive hemidactyliine genera series. Female survivorship in quadramaculatus and which have stream-type larvae. E. quadridigitata monticola is about equal to that of males of the is semiaquatic, Stereochilus is aquatic, and H emi­ same species, but mortality of mature females of dactylium is terrestrial in the adult stage. fuscus, ochrophaeus, and wrighti is considerably higher than that of males of the species at the Occupation of new adaptive zones and subzones same ages. Organ has shown that brooding females Simpson (1953) has discussed the biological of ochrophaeus and wrighti return to aquatic sites, basis of evolutionary or radiative movements into and, as a group, have higher mortality rates than new adaptive zones, and the acquisition of new those parts of the populations that remain ter­ ways of life. Ip order for populations to enter new restrial during the brooding season. In addition the adaptive zones there are, according to Simpson, more terrestrial the species, the higher is the pro­ three essentials. The populations must have physi­ portion of mature males to mature females in the cal access to the new zones. They must also have populations; this is another indication of the higher the necessary preadaptations which will be of sig­ mortality of mature brooding females in aquatic nificant adaptive value in the new way of life, that nesting sites. Plethodontid evolution has been is they must have evolutionary access. Finally, com­ characterized by parallel invasion and occupation petitive encounters in the new zones must be at a of terrestrial adaptive zones (see below) . As a minimum so that the populations have ecological result of Organ's study it is now known that sur­ access to the zones. Only when populations have all vivorship is increased in terrestrial habitats within types of access can occupation of new adaptive a series of related species and within populations zones begin. of a single species. Selection favors terrestrial Evolutionary trends within the subfamily Des- over less terrestrial individuals within the popu- 1966 EVOLUTION OF LUNGLESS SALAMANDERS 77 lations. It is likely that the populations involved in larval types ranging from permanent larvae crossing the adaptive threshold between the semi­ (Typhlomolge) and species with long larval stages aquatic and terrestrial zones (probably in early (Desmognathus quadramaculatus, Gyrinophilus) to Tertiary) were under strong environmental selec­ species with brief larval stages (Desmognathus tion pressures, and that the populations evolved aeneus, Hemidactylium). The larvae may be stream sequentially at very rapid, progressive rates. or pond types. The variety of larvae in modern Although the trend in Desmognathus is toward forms and the presence of distinct terrestrial trends increased terrestriality, the genus has not entered a in a single, primitively aquatic genus (Desmog­ new adaptive zone. One species, D. wrighti, has nathus) are indications of the evolutionary plasticity abandoned the aquatic larval stage and is essentially of plethodontid developmental processes. This plas­ terrestrial, but it is a highly specialized form that ticity might be considered the prospective adapta­ is unlikely to give rise to an adaptive radiation. tion which provided evolutionary access to the Even if the species has the necessary evolutionary new terrestrial adaptive zone. plasticity, it is doubtful that successful occupation Diversification usually follows occupation of new of the new adaptive zone will be possible, for the adaptive zones with the extent of subsequent adap­ terrestrial niches in Appalachia are already largely tive diversity roughly proportional to the distinc­ occupied by species of Plethodon and Aneides and tiveness of the new adaptive type and extent and ecological access is not available. diversity of the new territory (Simpson, 1953). Available information indicates that Phaeog­ The diversification is related to the number of new nathus is a terrestrial species (Valentine, 1963 b; adaptive subzones to which the evolving popula­ 1963 c; Brandon, 1965). Characters that suggest tions have access. In the case of the Plethodontini terrestriality include terrestrial adult habitat, co­ and Bolitoglossini diversification has been, relative ossified skin on anterior cranial elements and to other salamander groups, very great. elongation as a result of the addition of trunk Within the Bolitoglossini terrestriality has been vertebrae (both adaptions related to terrestrial achieved and movement into several adaptive sub­ burrowing), and large, yolky oviducal eggs re­ zones has been initiated. The tribe as a whole has sembling those of terrestrial plethodontines. a high base level of general adaptation, but most Phaeognathus apparently represents an invasion of genera are very specially adapted to specific the terrestrial adaptive zone by a specialized, environments. dead-end group. Hydromantes is associated with a rather re­ All genera of the tribes Plethodontini and stricted habitat in limestone areas. The species Bolitoglossini have achieved true terrestriality. All occur on the surface under rocks and other debris, lack free living larvae, and have directly develop­ but individuals also inhabit caverns. Survival of ing embryos that are retained within the egg mem­ this relatively ancient group has been related to branes; eggs are laid on land. The two assemblages association with relatively stable microenviron­ are distinct morphologically, and terrestriality in ments. the groups may have evolved in parallel. Movement Distinct evolutionary trends in the direction of to land is truly movement into a new adaptive zone trunk and tail elongation and general skeletal re­ and adoption of a new way of life for salamanders. duction are found in Batrachoseps. The most Terrestriality has resulted in a significant elevation primitive species (B. wrighti) has relatively low of the base level of general adaptation, which has numbers of trunk vertebrae, moderately long tails, been followed in parallel lines by significant evo­ and relatively long limbs, but the advanced species lutionary diversifications and adaptive radiations. have many trunk vertebrae, very long tails, and very The terrestrial environment borders the ancestral small limbs. All have lost fifth toes and fifth distal mountain brook habitat, the presumed home of tarsals. These specializations indicate that Batra­ plethodonine and bolitoglossine progenitors, and choseps has undergone an evolutionary modifica­ physical access to the new adaptive zone was no tion of both locomotor morphology and behavior, great problem. At the time the adaptive shift was and these changes are apparently associated with proceeding (Cretaceous or early Tertiary periods) occupancy of a new environment. Storer (1925) the ancestral environment was probably crowded stated that Batrachoseps uses the excavations made with salamanders and competition was keen. By by earthworms and insects, and interpreted elonga­ contrast the terrestrial environment adjoining was tion of the body, increase in the number of trunk relatively unpopulated by potential competitors so vertebrae, reduction in size of limbs and feet, and that the pioneers had ecological access to the new tail elongation as specializations for subterranean zone. Today plethodontids have a wide variety of life. From personal acquaintance with Batracho- 78 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 seps in the field it is apparent to me that the south­ erate to high elevations from southern Mexico to ern California species spend considerably more northwestern Colombia, and have little to mod­ time in subterranean refuges than they do on the erate hand and foot webbing. Most primitive species surface and that morphological specializations are are ground-dwellers living under rocks and surface directly correlated with behavioral and ecological debris in primarily unforested highlands. Some specializations. The genus is in the process of in­ extend to moderate elevations and may occur in vading a new semi-fossorial or fossorial adaptive bromeliads as well as on the ground (e.g., B. subzone. Many plethodontids have genetically fixed subpalmata). Throughout its range Bolitoglossa has numbers of vertebrae. Apparently variability in invaded the tropical lowlands and expanded its the numbers of trunk vertebrae and attenuate peripheral range. Invasions of lowlands have in­ habitus have given Batrachoseps evolutionary ac­ variably been accompanied by increased webbing, cess to the new adaptive subzone. flattening of the hands and feet, and increased The supergenus Bolitoglossa is unique in being arboreality. Fully webbed hands and feet provide the only tropical group of salamanders. Within the more surface area and are of great significance in neotropics the group has undergone a considerable forms such as Bolitoglossa, which depend on sur­ and significant radiation. The supergenus comprises face tension to support their weight while walking seven tropical genera and over 100 species-almost and climbing on smooth, moist leaf surfaces. two-thirds of the plethodontid species. Pletho­ Genetic potential for variation in hand and foot dontids probably arose in what is now the Nearctic morphology and subsequent molding by selection region, and the primary dispersal center has been are the most important factors providing evolu­ northern; but a second center of dispersal is today tionary access to the lowland arboreal subzone. the southern and eastern borders of the Mexican The problem of ecological access is particularly Plateau, and it is inappropriate to speak of the acute in tropical lowlands. The "salamander family Plethodontidae as a Holarctic or Nearctic niches" of higher latitudes and elevations are filled group. -primarily by frogs and possibly by snakes and Most tropical plethodontids occur at high eleva­ lizards. Ecological access has been available, how­ tions in relatively cool habitats, and only two genera ever, in arboreal habitats, the route taken by (Bolitoglossa, Oedipina) have successfully invaded Bolitoglossa, and in semi-fossorial habitats, the the tropical lowlands (Lineatriton and Parvimolge route followed by Oedipina. may also occur in tropical lowlands but both have The final group, tribe Plethodontini, probably very limited distributions) . left the semiaquatic adaptive zone at a later date Neotropical species occupy a variety of terrestrial than the Bolitoglossini, and as a group is more niches. Lineatriton and Oedipina are greatly elon­ generally and less specially adapted. Species of gated with reduced limbs and feet and extremely Aneides and Plethodon still occur in Appalachia, long tails. Not much information is available con­ but the bolitoglossines all occupy ranges far from cerning the habits of the two, but apparently both the ancestral center. Two of the three plethodonine are subterranean to some extent and utilize burrows genera, Plethodon and Ensatina, are primarily of earthworms, mammals, etc. (Dunn, 1928). ground-dwellers and are most frequently en­ Other genera (Pseudoeurycea, Thorius) are more or countered under rocks, logs, and other surface lit­ less surface litter forms found under rocks and logs ter, or out wandering about on the forest floor. and under the loose bark of fallen trees. Some Most or all species of the two genera spend con­ Chiropterotriton are terrestrial and occur in mossy siderable amounts of time below the surface. areas on banks and cliffs. Ensatina is a rather stout-bodied form which uses Species of the genera Bolitoglossa, Chiroptero­ burrows of larger organisms such as rodents as a triton, Pseudoeurycea, and Parvimolge are moving means of getting below the surface (Stebbins, into a new adaptive subzone-the arboreal sub­ 1954). Species of Plethodon are more elongate zone. Chiropterotriton is a genus of primarily high­ than Ensatina. Slender attenuate forms such as land, climbing species. The group has an adaptation P. cinereus, P. richmondi, and P. elongatus have which has apparently given it evolutionary access reduced limbs and feet and are able to enter bur­ to the new subzone; this feature is the articulation rows of relatively small size. Some of the larger, of all distal tarsals with the centrals (see above, longer-legged species are capable of some scansorial mesopodial elements) . The mesopodial pattern is locomotion; I have encountered P. jordani at dis­ apparently of significance in facilitating climbing tances one to three feet above ground on tree by allowing greater digital spread. trunks. Other large species such as P. yonahlossee Primitive species of Bolitoglossa occur at mod- inhabit burrows of relatively large diameter. One 1966 EVOLUTION OF LUNGLESS SALAMANDERS 79 species, P. dunni, has secondarily returned to a first is paedogenesis in which sexual maturity is semiaquatic habitat in the Pacific Northwest and is accelerated and all members of the species retain commonly encountered in rocky, moss-covered larval morphology (except reproductive system) seepage areas. The primary habitat of both Pletho­ throughout life. Paedogenesis is genetically fixed, don and Ensatina, however, is at the surface of the leads to specialization and degeneration, and is of ground away from running water. little significance as far as future phylogenetic The genus Aneides is an excellent example of a progress is concerned. De Beer (1958) used the group entering a new adaptive subzone, the arboreal term neoteny to describe the permanent larval way of life. The most primitive species, A. hardii, condition of salamanders, but he fails to distinguish is a Plethodon-like form that lives in, under, and between environmentally induced and genetically on fallen logs. Three other species (A. aeneus, A. fixed conditions. The later (called permanent ferreus, A. lugubris) are arboreal to a large degree neoteny) perfectly fits his description of paedo­ and have numbers of specializations related to genesis. The other major condition which results arboreal life (depressed body, long limbs and from paedomorphosis is a phenomenon called ar­ digits, bifurcated terminal phalanges, prehensile rested development or differential metamorphosis. tail). A detailed analysis of evolutionary patterns Differential metamorphosis is a term that describes and trends in Aneides has been prepared and will the paedomorphic pattern in which metamorphic be published elsewhere. processes are extended over a considerable period A final species, A. flavipunctatus, is a short­ of time, with some elements (bones in the present legged salamander with a cylindrical body. It is case) completing metamorphosis early, others very often encountered in very moist situations such late, and some not at all. This condition is more as brooksides (Myers and Maslin, 1948). It is characteristic of terrestrial than aquatic groups of similar to Plethodon dunni in this regard and has plethodontids and has been of great importance a much greater affinity for wet habitats than other in providing an escape from specialization in the Aneides. Return to semiaquatic habitats (but with ancestral adaptive zone and access to new adaptive retention of the terrestrial "way of life") has oc­ zones. curred only in the Pacific Coast species of Pletho­ Paedogenesis occurs only in the following don and Aneides, since such ecological niches are species, all members of the tribe Hemidactyliini: occupied by Desmognathus, Eurycea, etc., in the Gyrinophilus palleucus (three distinct, allopatric East. races), Eurycea latitans, E. nana, E. neotenes, E. pterophila, E. troglodytes, E. tynerensis, Typhlo­ Paedomorphosis molge rathbuni, T. tridentifera, and Haideotriton Paedomorphic influence and wallacei. Based on the morphological evidence pre­ plethodontid evolution sented above, paedogenesis has probably evolved Paedomorphosis is a term used to denominate in parallel at least four, and possibly more times. several morphological modes or patterns of evolu­ All paedogenetic species are highly specialized, and tion. Paedomorphosis involves characters that are their distribution is somewhat to greatly peripheral present or make their appearance in the young to the main tribal range. Paedogenesis may have stage of an ancestral group and which in the arisen initially in response to selection pressures ontogeny of a descendent appear: ( 1) in the young operative in marginal habitats during arid periods. and adult stage, producing a substitution of a new Most paedogenetic species occupy subterranean adult condition for the old, resulting in progressive waters exclusively or to a considerable degree. Cer­ deviation in the ontogeny of the descendent from tainly aquatic larvae of ancestral populations had that of the ancestor; or (2) in the adult, by a rela­ physical access to such environments, and the very tive retardation of the development of the bodily fact that ancestral populations had aquatic larvae structures as compared with the reproductive or­ provided evolutionary access to the new aquatic gans (DeBeer, 1958). Paedomorphosis is and has or subterranean adaptive subzone. Subterranean been of extreme importance in the evolution of waters are notably free of competitors, and the plethodontid salamanders, and the general principle invaders also had ecological access. involved may be restated as the mode by which Differential metamorphosis has not been a major larval and youthful characters in the ancestor can feature of the evolution of the desmognathines. influence adult characters in the descendent (Gar­ Perhaps shortening of the larval periods and speed­ stang, 1922; De Beer, 1958). ing of the metamorphic processes in advanced Conditions of two distinct types have resulted Desmognathus are related to paedomorphosis. from paedomorphic influence in plethodontids. The Small adult size and resemblance in osteological 80 MEMOIR OF THE SOUTHERN CAUFORNIA ACADEMY OF SCIENCES YOL.4 structure to young stages of primitive species Neotropical salamanders (supergenus Bolitog­ characterize D. aeneus and D. wrighti. Examples lossa) have been affected by paedomorphosis more are retention of anterior vomerine teeth in adults, than any other plethodontid group with the excep­ relatively slight sexual dimorphism; and relatively tion of the paedogenetic species. Examples of thin, light cranial elements. These features are a paedomorphic influence include fusion of premaxil­ result of paedomorphic influence, and sexual ma­ lae and absence of second basibranchials in all turity is achieved at a morphological stage roughly genera. Examples found in individual genera follow. equivalent to a juvenile stage of such species as D. Bolitoglossa: failure of prefrontals to develop, or quadramaculatus. Both species have distinct ter­ irregular prefrontal development in a number of restrial tendencies. advanced species belonging to different species Apart from the paedogenetic effect, paedo­ groups (rufescens, occidentalis; savagei, adspersa, morphosis has had some influence in the Hemidac­ orestes; cerroensis, robusta; colonnea); failure of tyliini. Hemidactylium and Eurycea quadridigitata septomaxillae to develop in most species; partially have only four toes. The larvae of both differ from to fully webbed feet (a condition encountered in those of other plethodontids in that the hind limbs young or embryonic stages of other genera). develop slowly and are represented in hatchlings by Oedipina: failure of prefrontals and septomaxil­ undifferentiated lobes. Failure of the :fifth toes to lae to develop; weak development of limbs and feet, develop may be the result of differential initial with reduced phalanges and variable numbers of growth rates and delayed limb development. Paedo­ mesopodial fusions; high percentage of unicipital mcrphosis may also be responsible for the relatively ribs; extremely small premaxillae; weak skull weuk cranial elements and tenuous cranial articula­ articulations. tious of such species as E. quadridigitata and E. Pseudoeurycea: failure of septomaxillae to ap­ multiplicata, diminutive size of E. quadridigitata, pear in all but a few species, which have irregular retention of a lateral line system on the head of degrees of development. adult Stereochilus, relatively small eyes of Sterea­ Lineatriton: failure of septomaxillae to become chilus, and relatively poorly developed eyes, eye­ well developed; diminutive size; reduced limbs and lids, and nasolacrimal glands of Typhlotritan. feet. Bolitoglossines have been influenced by paedo­ Tharius: variously developed septomaxillae; all morphosis to a greater degree than any other ter­ cranial elements greatly reduced in size; diminutive restrial plethodontid group. Hydromantes has size; reduced limbs and feet; enlarged nostrils of retained the primitive metamorphic pattern in re­ some species. gards to premaxillary structure, but paedomorpho­ Two genera, Chiropteratriton and Parvimolge, sis may be responsible for the failure of prefrontals deserve special comment. Primitive Chiropterotri­ to develop, the relatively weak skull articulations, ton are but slightly affected by paedomorphosis, and for .the high numbers of unicipital ribs in the but an evolutionary gradient related to paedomor­ European species. phosis exists in the genus on a roughly north to Batrachoseps has been profoundly influenced by south axis (see also Rabb, 1960). The northern paedomorphosis, and differential metamorphosis is species usually have prefrontals, septomaxillae, apparent in the modern species. Premaxillae are tibial spurs, and vomerine preorbital processes, but fused in the young of all species and in the adults of the elements are absent in some of the southern all except B. wrighti. Premaxillary metamorphosis species. In addition southern species demonstrate is thus delayed relatively long in wrighti, and is features considered by Rabb (1960) and me to be never achieved in the other species. A further ex­ strengthening compensations in certain elements to ample of differential metamorphosis is the appear­ paedomorphic weakening in others .. These compen­ ance in very large wrighti of tiny prefrontal bones, sations include calcification of normally cartilagi­ which are absent in other species. Other paedomor­ nous structures such as mesopodial elements, heads phic effects evident in members of the genus include of limb bones, and intervertebral cartilages, fusion weak development of frontals and parietals, in­ of premaxillary frontal processes, and mesopodial complete cranial roof formation, incomplete devel­ fusions. These compensations are also encountered opment of vomers with no preorbital processes, in varying degrees in Batrachaseps, Lineatriton, absence of second basibranchials, weak limb and Parvimolge, Oedipina, and Thorius. Table 2 is a foot development, and absence of fifth toes. Pre­ representation of the distribution of paedomorphic frontals and vomerine preorbital processes are characters and compensations to paedomorphosis among the last elements to develop during primitive in some species of Chiropterotriton. It is apparent plethodontid metamorphosis. that the southern, more advanced species have more 1966 EVOLUTION OF LUNGLESS SALAMANDERS 81 paedomorphic characters than the more primitive ton. It is also possible that evolution is following northern species. The first four species (priscus, similar patterns in Chiropterotriton and Parvi­ chiropterus, multidentatus, dimidiatus) occur north molge, and richardi may be an advanced member of the Isthmus of Tehuantepec; the remainder occur of the latter. Further speculation must await in­ south of the Isthmus. C. dimidiatus and C. multi­ formation concerning P. praecellens and the meso­ dentatus occur sympatrically at some localities, podial structure of richardi. otherwise the chart is arranged on a geographical As a group the tribe Plethodontini has been little north-south axis. Some of the data in the table are affected by paedomorphosis. at variance with information presented by Rabb ( 1960), but the important point is that evolutionary specialization and advance in Chiropterotriton is The evolutionary significance of paedomorphosis associated with a geographic trend and with trends The major role played by paedomorphosis in in the direction of increased paedomorphosis. plethodontid evolution is demonstrated by different Parvimolge townsendi of Veracruz, Mexico, is adaptive trends that relate to the phylogeny of two strongly affected by paedomorphosis (diminutive distinct groups. Paedogenesis permits survival of size, reduced limbs and feet, mesopodial fusions, several species of hemidactyliines in specialized calcified mesopodials, calcified intervertebral carti­ and restricted habitats. These habitats are found lages, reduced septomaxillae and columellae) . It primarily in regions that are not favorable for the resembles P. richardi of Costa Rica in some char­ survival of related, non-paedogenetic species. Ex- acters (see above) but differs in others. The two . treme specialization within such unusual habitats species probably represent different evolutionary and consequential loss of evolutionary plasticity lines, one of which (townsendi) may have arisen (potentiality for future evolutionary progress) from a Pseudoeurycea-Chiropterotriton ancestor, characterize the paedogenetic forms. Differential the other from a Chiropterotriton stock. If placed metamorphosis, a pattern most apparent in the in Table 2, richardi would have three paedomor­ bolitoglossine genera, is associated with occupancy phically related characters (no septomaxillae, no of and survival in terrestrial habitats. The bolito­ columellae, calcified intervertebral cartilages), glossine genera are all highly specialized, but they somewhat fewer than other southern Chiropterotri- have avoided overspecialization through paedomor~

TABLE 2. Paedomorphic characters and compensations in Chiropterotriton. Characters I"' £ ~ g ;§ := =c:tl li i;i.. := -e 1 u c:tl 8< Jg i;i.. ~ i;i.. ~ ~f ~ g e~ =~ ii ~] ~i;i.. 3 .s::.o._ ~.sa 7J"' ~Q ~Q0) ~1ii i:i 8< .~o =~t;1,,, -g "' 0) t.l El~ 0) g Cll:::l '8i:i..~ "' lH 0 0) .go -as §S .... .l:l O.S! §"!§ t !9 p,.Q 8:= 41) t.l O!"' o> ·--.s::.o·- - ::l !a a ~"' Species ~g ~Q ~a! E= &'! 8a! £~ ~z ~~ ~a ....=";l u ~ --- priscus 0 5chiropterus 0 ~ . multidentatus 0 dimidiatus x x 2 . xolocalcae x x x x 4 . bromeliacia x x x x x s nasalis x x x x x x x 7 s 7 ~ abscon.dens x x x x x x x 82 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 phosis. This appears to be associated with and, in and has led to the great differentiation of the group. fact, may be responsible for a great increase in Differential metamorphosis is most obvious in evolutionary plasticity. peripheral and marginal habitats where either selec­ Paedogenetic plethodontids are degenerate spe­ tion is very strong or entrance into a foreign en­ cies living in habitats that are secluded, difficult, or vironment was difficult for the ancestral stock. The unfavorable from the standpoint of the major genus Batrachoseps provides an example of the group, viz., the primitive hemidactyliine genera. significance of these paedomorphic effects. B. - These habitats are invariably on the margins of the wrighti, an inhabitant of environments similar to present generic or tribal ranges. Such a distribution those occupied by primitive plethodontids, is in­ suggests that the species are descended from rem­ fluenced but slightly by paedomorphosis. The other nants of species which inhabited the regions when species, all with peripheral or marginal distribu­ conditions were more favorable for their survival tions, are strongly affected by paedomorphosis. than at present. They probably entered their pres­ Selection in the latter species probably favored ent habitats in response to strong eliminative pres­ populations or individuals in which reproductive sures in the environments of the adults of ancestral maturity occurs at an early developmental stage. A populations. Such elimination may have been ran­ similar example has been discussed earlier in re­ dom and nonselective with the organisms virtually gard to the southward spread of the genus Chirop­ defenseless against such severe climatic factors ·as terotriton in the Neotropics. drought or extremes of heat or cold. Permanent Metamorphosis and ontogenetic anatomical occupancy of the larval habitat or of physically changes require energy expenditure beyond main­ adjacent and accessible habitats, such as subter­ tenance levels. Thus, an additional factor in the ranean water systems, was necessary for survival. evolution of paedomorphosis is the selection in Under conditions of nonselective elimination there marginal habitats of individuals that could repro­ is selection for survival of progeny through in­ duce at relatively early overall developmental creased fertility, and, particularly in salamanders, stages. Paedomorphic individuals expend relatively acceleration of the life cycle with earlier repro­ less energy in order to contribute to the gene pools ductive maturity. Accordingly, individuals do not than non-paedomorphs. Therefore the paedo­ complete ontogeny earlier than individuals of re­ morphs could reproduce at higher rates than non­ lated species, but certain systems (e.g. reproduc­ paedomorphs, and the populations would evolve in tive) mature while others do not. Most systems are the direction of the former. Biotic selective pres­ characterized by larval morphology. sures are strong in the bolitoglossines, particularly Selection probably favored neotenic individuals in such densely populated habitats as the tropical over those that transformed, with eventual genetic lowlands. These pressures have contributed to a fixation of neoteny, i.e., paedogenesis. The pre­ substantial increase in the significance of paedo­ mature maturation of reproductive organs linked morphosis in bolitoglossines, relative to more with degeneration or underdevelopment, or both, primitive pletbodontids. of other organs is typical of this type of selection Evolutionary stagnation has resulted from pae­ and characterizes the paedogenetic plethodontids. domorphosis in the paedogenetic hemidactyliines. Results of such selection in other organisms have In contrast, paedomorphosis in the bolitoglossines been summarized by Schmalhausen (1949). has increased evolutionary plasticity and has ex­ The second kind of paedomorphosis (differential erted a profound directing influence in their phylo­ metamorphosis) is associated with the evolutionary geny. Bolitoglossines may have been the first activity of the tribe Bolitoglossini. Terrestrial ple­ plethodontid group to abandon the aquatic larval thodontids lack aquatic larvae. Their metamorphic stage and to become truly terrestrial. Ancestors of changes tend to be spread over a greater period of this group were probably the first terrestrial pletho­ time and tend to be less extreme than in those dontids to extend their ranges out of Appalachia. species with larval stages that undergo complete Bolitoglossines must have begun specializing very transformation in a relatively short period of time. early, so that they were no longer capable of ef­ In terrestrial plethodonines most metamprphic fectively competing with the relatively generally changes occur before hatching. It is by differential adapted and vigorous plethodonines when the metamorphosis in the bolitoglossines that some of latter invaded terrestrial habitats. Elimination of these changes (premaxillary separation, develop­ bolitoglossines in Appalachia resulted. Bolitoglos­ ment of prefrontals, etc.) occur relatively late in sines in the West were forced to continue trends life, or not at all. Differential metamorphosis has toward increased specialization as a means of avoid­ activated the variational potential of bolitoglossines ing competition. Only a fragment of a presumed 1966 EVOLUTION OF LUNGLESS SALAMANDERS 83 early radiation survived, and these were the an­ Paedomorphosis may result from a relatively cestors of the modem supergenera. Only one group, simple genetic mechanism. There is some sugges­ the tropical salamanders of the supergenus Bolito­ tion that control or regulatory genes, in some ways glossa, has been successful, and all of its species are analogous to the operator genes of microorganisms, moderately to profoundly affected by paedomor­ exist in vertebrates, and these may influence the phosis. The tropical environment is foreign to sala­ activity of numerous structural genes or gene com­ manders, and the association of paedomorphosis plexes (Ingram, 1961; Manwell, 1963). Cooper with the tropical salamander radiation suggests that (1965) has demonstrated that cartilage inducing it provided a means of evolutionary escape from activity in embryonic stages of vertebrates is a overspecialization for ancestral stocks. The highly stage-specific phenomenon, with no single "in­ adaptive tropical radiation has resulted in the evo­ ducer" having a dominant role. Presumably there lution of several genera and over 100 species of is an intricate pattern of gene activation and inac­ salamanders. Paedomorphosis seems to have been tivation. Regulator genes, or their equivalent, may a means of releasing the evolutionary potential of play very important integrating roles in these de­ the group. velopmental patterns. It is possible, however, that At the present time plethodontid evolutionary polygenic systems, operative at early stages of activity seems to be most intense in the wet tropical development, rather than single genes, are involved lowlands, a densely populated environment. In­ (see Waddington, 1956). Modification of these vasion and radiation in such an environment by mechanisms could have very widespread effects, salamanders, which, except for this single group, leading to differential metamorphosis and other are otherwise restricted to northern temperate en­ evolutionary phenomena. Selection for favorable vironments is highly significant. The vigor with combinations of such genes or gene complexes which the radiation has proceeded and the high could result in the fixation of new morphological degree of species differentiation attest to the evolu­ arrangements. Favorable new arrangements, once tionary plasticity of the group. The fact that all of introduced in populations, might play key roles in the more advanced species of the various genera initiating major new evolutionary trends, especially are increasingly affected by paedomorphosis dem­ if they provide access to a new habitat or a new onstrates its dynamic, progressive influence in the adaptive zone. group. The active role of paedomorphosis has been Genetic fixation of paedomorphic characters is the dominant feature in the evolution and adaptive closely associated with the evolution of compensa­ radiation of the tropical salamanders. tory adaptations. Selection for developmental ar­ Although little factual supportive evidence has rythmia or partial developmental failure would not been provided, paedomorphosis has been consid­ be successful without concomitant compensatory ered by Garstang, DeBeer, Rensch, Schmalhausen, changes in the genetic background of the popula­ and other biologists to be an important evolutionary tion as a result of counter selection. Examples are process. Few attempts have been made to reinter­ evident in some neotropical salamanders. In mem­ pret the early work on paedomorphosis in the light bers of the paedomorphic genus, Thorius, diminu­ of modem genetic and evolutionary theory. As a tive size and weak bony and muscular development result paedomorphosis is often assigned only a are compensated for and strengthened by calcifica­ minor role in theories of the evolution of higher tions of normally uncalcified cartilaginous struc­ levels of organization, and receives only passing mention in major treatises on evolution (e.g. Simp­ tures such as the ends of long bones, the mesopodial son, 1953; Mayr, 1962). Recently additional facts elements, and the intervertebral cartilages. have been made available. Manwell (1963) cites Accumulating evidence indicates that paedomor­ the almost identical nature of biochemical and phosis is of extreme importance in providing new physiological properties of the hemoglobins of adult opportunities to populations of organisms. Paedo­ brook lampreys (Entosphenus lamottei) and am­ morphosis might initiate and even direct major mocoete larvae of other species as evidence of evolutionary and phylogenetic trends, and provide paedomorphosis in the former. Several kinds of a mechanism of rejuvenation for overspecialized morphological evidence of paedomorphosis are or slowly evolving groups. Evolutionary break described in the present paper. These and other throughs which result may be involved in the adop­ data indicate that a reinvestigation of the general tion of new ways of life and may form the basis biological significance of paedomorphosis and other for subsequent adaptive radiations. Paedomorpho­ ontogenetic phenomena (DeBeer, 1958; Rensch, sis clearly is a major feature of pletbodontid evolu­ 1959) might have fruitful results. tion and phylogeny, and investigations of the 84 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 significance of the phenomenon in other groups of had opisthocoelous vertebrae. Dehmiella had stout organisms are suggested. neural crests that were considerably higher than the posterior neural arch margins, a condition not BIOGEOGRAPHY AND PHYLOGENY encountered in plethodontids. Transverse processes were much straighter than in most plethodontids The Fossil Record and were almost perpendicular t6 the central axis. The only undoubted plethodontid fossils are The vertebrae of Dehmiella were apparently some­ Pleistocene remains of Plethodon glutinosus of what broader than those of any modem pletho­ essentially modem structure from the Wisconsin dontid, but were amphicoelous. Certainly the ver­ and Illinoian ages of Florida (Holman, 1958; tebrae of the two genera bear little resemblance to 1959 a; 1959 b). Peabody (1959) reported track­ those of Hydromantes, the only modem European ways of several salamanders, including Batra­ plethodontid, or to other plethodontids. Experience choseps, from the Lower Pliocene of Tuolomne bas shown that parallelism is common in salaman­ County in the Sierra Nevada of California. The der evolution, and basapophyses, found only in a Batrachoseps trackway was referred to B. pacificus few plethodontids, are not sufficient reason for as­ on the basis of large size and body and limb dimen­ signment of fossil forms to the family Plethodonti­ sions, but it :fits equally well an undescribed species dae. which occurs near the fossil locality. These are the Certain Mesozoic species may be plethodontids only New World Cenozoic plethodontid records. or close relatives. Auffenberg ( 1961) described a Herre and Lunau (1950) described a new genus new genus and species, Opisthotriton kayi, from and species of fossil salamander, Dehmiella schind­ vertebrae collected from the Lance Formation, wolfi, from the Lower Miocene (Burdigalium) of Upper Cretaceous, of eastern Wyoming. Opistho­ southern Germany, and Herre (1950) described a triton was tentatively assigned to the Salamandridae new genus and species, Geyeriella mertensi, from on the basis of very long maxillae and opisthocoel­ the Paleocene of Germany. Both are known only ous vertebrae. Hypapophysial keels and basapophy­ from vertebrae and are said to resemble pletho­ ses led Auffenberg to consider the genus to be close dontids because of the presence of well developed to plethodontids. basapophyses. Such processes are well developed Estes (1964) examined much more material of in desmognathines and some plethodontines, but Opisthotriton than was available to Auffenberg. He not in salamandrids. Ambystomatids and hynobiids showed that material (a maxilla) on which pro­ have either differently shaped processes at the front posed salamandrid relationship was based is refer­ ends of the centra, or none at all. Dehmiella was able to Scapherpeton, a fossil cryptobranchid. Estes considered to be a plethodontid, but Geyeriella considered the vertebrae of Opisthotriton to be was thought to be a forerunner of the plethodontids. definitely opisthocoelous. On the basis of vertebral Auffenberg ( 1961 ) suggested that Geyeriella characters, including presence of pterygapophyses, may represent a side line of the main evolutionary Opisthotriton is questionably referred to the sub­ sequence leading to the Plethodontidae and that family Desmognathinae. Dehmiella represents the first "true plethodontid." Opisthotriton differed from all modern pletho­ He assumed a close salamandrid-plethodontid re­ dontids in lacking internasal fontanelles, and had lationship, and this assumption obviously influenced hour glass-shaped palatopterygoids unlike those of his judgment. Estes (1964) stated that Dehmiella any plethodontid. It further differed from desmog­ is probably a plethodontid, but that the status of nathines in having two premaxillae and open meck­ Geyeriella is uncertain. elian grooves, but both are primitive characters that I have not examined the material upon which might be expected in an ancestral group. Transverse the genera Geyeriella and Dehmiella are based, but processes were considerably longer than those of it is doubtful that either is a plethodontid. The any modem plethodontid, but otherwise the :figures character upon which the genera are based (basa­ of the vertebrae are similar to those of modem pophyses) is an adaptive feature that has arisen a desmognathines. The maxillae are strikingly differ­ number of times within the Plethodontidae. It is ent from those of all plethodontids. Estes thinks apparent from the descriptions and illustrations Opisthotriton was a larval or semilarval form. that both fossil genera differed in a number of Estes (1964) described a new salamander genus structural features from modem plethodontids. and species, Prodesmodon copei, from the Upper Geyeriella differed in the extreme posterior place­ Cretaceous of Eastern Wyoming. The genus is also ment and great length of the transverse processes, known from the early Cretaceous of Texas (Hecht, and in the relative neural arch length. It apparently 1963) . The species was considered by Estes to be a 1966 EVOLUTION OF LUNGLESS SALAMANDERS 85 desmognathine on the basis of disarticulated bony tion have since modified the landscape, but Ap­ elements. Important characters of the genus in­ palachia has provided a relatively stable environ­ clude: opisthocoelous vertebrae with strongly pro­ ment for the development of the biota since the end truding condyles, strongly developed hypapophy­ of Permian (Graham, 1964). ses, pterygapophyses, and basapophyses; large, It was stated in rather unequivocal terms by concave, atlantal condylar articulating facets; small Dunn (1926) that plethodontids have been derived odontoid processes; alar processes; paired premaxil­ from a salamandrid stock in the general present­ lae; no intemasal fontanelles; closed meckelian day Appalachian region during late Eocene to Mio­ grooves; sculptured anterior cranial elements; sinu­ cene. Evidence has been presented above to show ous tooth bearing dentary margins; small maxillary that the plethodontid-salamandrid relationship is facial lobes; interlocking mandibular symphyses; not close, and that it is highly unlikely that the two non-pedicellated teeth; no ribs. This is a mosaic of groups have had a common ancestor. The report primitive and advanced characters unlike any pat­ of highly specialized desmognathine salamanders tern found in modem plethodontids. Lack of ribs in early Cretaceous (Estes, 1964, 1965; Hecht, and intemasal fontanelles, and presence of non­ 1963) suggests that family characteristics of the pedicellated teeth and interlocking mandibular Plethodontidae were established during Mesozoic. symphyses distinguish Prodesmodon from all ple­ The probability that the specialized Prodesmodon thodontid genera. Most of the characters suggest a occupied swampy environments beyond the west­ relationship to desmognathines, but modem des­ ern periphery of Appalachia in Cretaceous indi­ mognathines have a single premaxilla and large cates that plethodontids had then been in existence maxillary facial lobes. Paired premaxillae are to be for some time, and suggests that if plethodontids expected in primitive desmognathines, however, did arise in what is present-day Appalachia, they and on the basis of the information presented by may have originated in mid-Mesozoic or earlier. Estes, Prodesmodon apparently is a specialized (no Based on morphological similarities, plethodon­ ribs, vertebral specializations) offshoot of a primi­ tids may have arisen from an ambystomatid or pre­ tive desmognathine stock. ambystomatid ancestral stock, which in tum was Research on fossil salamanders is actively being probably the New World derivative of the ancestors pursued by Estes, and further comments are de­ of the Old World Hynobiidae. The family Ambysto­ ferred until results are made available. matidae is an endemic North American family which is more primitive than the family Pletho­ Historical Backgrounds and the dontidae. Modem ambystomatids apparently have Primary Dichotomy less evolutionary plasticity than plethodontids, but the family has undergone a moderate diversifica­ Origin of plethodontid salamanders in what is tion. Significantly, one genus, Rhyacotriton, occu­ now called the Appalachian Highland physiogra­ pies mountain brook habitats in northwestern phic province (Atwood, 1940), or Appalachia, was North America, and it is virtually lungless. It is first discussed by Wilder and Dunn (1920), and postulated that a similarly adapted population, but the proposal was expanded and detailed by Dunn one with great genetic and evolutionary plasticity (1926). Since that time the theory has gained wide and a higher base level of general adaptation, suc­ acceptance. Three of the four major groups of cessfully occupied the mountain brooks of Appala­ plethodontids occur in Appalachia, including all of the primitive genera. Wilder and Dunn (1920) and chia during Mesozoic times and gave rise to the Dunn (1926) developed the thesis that mountain plethodontid radiation. brooks were the ancestral plethodontid habitats; Because the fossil record is scanty and incom­ mountain brook plethodontids occur only in Ap­ plete it is impossible to date the primary dichotomy palachia. Evolution of the family requires an area of the Plethodontidae. Divergence of the desmog­ that has been relatively stable geologically and cli­ nathine line from the more generalized stock that matically for a very long time, at least since mid­ gave rise to the present plethodontines must have Mesozoic; Appalachia is such a region. Thornbury taken place in Mesozoic. Certain characters (e.g., (1954) states that mountain building affected separated premaxillae) of Prodesmodon indicate Appalachia as early as the Ordovician and contin­ that it was relatively close to the ancestral stock. ued intermittently throughout the Paleozoic. The The closed meckelian grooves, opisthocoelous ver­ Appalachian revolution at the close of the Paleozoic tebrae, and shape of cranial elements indicate that was apparently the culmination of mountain build­ the primary plethodontid dichotomy had already ing trends. Erosion, sedimentation, and peneplana- occurred, and that the Cretaceous Prodesmodon 86 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 was a specialized desmognathine offshoot which rivatives. Salamanders and Arcto-Tertiary deriva­ had left the ancestral habitat. tives are both relatively ancient groups. Their as­ Tertiary events are responsible for the sub­ sociation is probably ancient and the communities familial evolutionary patterns seen in the pletho­ may have evolved rather slowly throughout Tm­ dontids. Because the fossil record is scanty, paleo­ ary. Savage (1960) has suggested that correlation botanical and paleoclimatic data must be used to of geologic and geofloral history forms a firm basis infer the past history of the Plethodontidae. for development of concepts concerning historical The evolution of flowering plants has been re­ bigeography of animal groups; this approach is viewed by Axelrod (1960) who states that close followed in the present work. relationship between modern plants and fossils of Historical elements of the North American her­ the later Oligocene makes it possible to reconstruct petofaunas have been discussed by Savage (1960). the habitats and climates under which Tertiary Of particular interest is the concept of an Old floras lived, to outline the belts of Tertiary vegeta­ Northern Herpetofaunal Element first proposed by tion, and to explain the evolution of modern pat­ Dunn (1931) and further developed by Savage. terns of plant distributions. Three broad, world­ The Old Northern Element as defined by Savage wide belts of distinctive vegetation were present in consists of descendents of temperate circumpolar early Tertiary: the Arcto-Tertiary Geoflora in the groups found in the far northern continental areas North, the Antarcto-Tertiary Geoflora in the South, in early Tertiary but forced southward as the and the very broad Tropical-Tertiary Geofl.ora in tropical limits were restricted by long-term Tertiary between. Many modifications occurred during Ter­ climatic trends. Savage emphasized the historical tiary, but the modern floristic patterns can be de­ association of the Old Northern Element with the rived from these three belts. Plants are the domi­ Arcto-Tertiary Geoflora, and his analysis aptly fits nant members of biotic communities, which in tum the family Plethodontidae. Savage further subdi­ are evolutionarily conservative units. Information vides the Old Northern Element into four historical about modern plant-animal associations allows one zoogeographical complexes. These complexes and to make hypotheses concerning past events. the plethodontid genera I consider to be members Tertiary events resulted in numerous reorganiza­ of each are: tions of the early Tertiary pattern. The tropical L Eastern American Complex. Desmognathus, vegetational belt gave rise in the New World to the Leurognathus, Gyrinophilus, Pseudotriton, Neotropical-Tertiary Geoflora which is ancestral Stereochilus, Eurycea, Typhlotriton, Typhlo­ to modern neotropical forest formations. A major molge, Hemidactylium. geofl.ora, the Madro-Tertiary, developed in situ in southwestern North America in early Tertiary in 2. Western American Complex. Hydromantes, response to increasing cooling and aridity which Batrachoseps, Ensatina. forced the old communities out. This geoflora was 3. Central American Complex. Bolitoglossa, derived primarily from Neotropical-Tertiary fl.oral Oedipina, Pseudoeurycea, Chiropterotriton, elements. The Madro-Tertiary Geoflora expanded Lineatriton, Parvimolge, Thorius. steadily in the increasingly arid Southwest from 4. Southeastern American Complex. Phaeogna­ late Eocene·throughout Tertiary, and has given rise thus, Haideotriton. to the modern semiarid woodland, chaparral, thorn forest, arid scrub, desert grassland, and desert vege­ Species groups of Aneides and Plethodon are mem­ tation of southwestern North America. Of con­ bers of both 1 and 2. In the following sections the siderable interest to the present study are the probable history of these genera is reviewed. modem forest derivatives of the Arcto-Tertiary Geoflora. The more important of these in North America are the Eastern Deciduous, Hemlock­ The Desmognathines Hardwood, Boreal, Rocky Mountain, Pacific, and Desmognathines have remained close to the Southeast Evergreen Forests. In addition disjunct plethodontid source region. Some species of Des­ elements of the Geoflora survive today interspersed mognathus today occupy what might well be the with Neotropical-Tertiary derivatives in the high­ primitive habitat. A moderate diversification has lands of Mexico, Guatemala, and other Central occurred in the subfamily, however, and .some American countries (Braun, 1950; Martin and groups have left Appalachia (Fig. 19). Harrell, 1957; Axelrod, 1960). Plethodontid sala­ Prodesmodon represents the earliest known di­ manders are today primarily members of commu­ vergence from the generalized desmognathine nities dominated by Arcto-Tertiary Geofioral de- stock. It left the ancestral habitat and familial 1966 EVOLUTION OF LUNGLESS SALAMANDERS 87 source region, became morphologically specialized, The Plethodontines and occupied a peripheral swampy habitat by early Tribal Origins Cretaceous; divergence must have occurred at a The plethodontines are more generalized mor­ relatively very early date. phologically than are the desmognathines and have Phaeognathus may represent a somewhat later, remained closer to the ancestral plethodontid but still very early divergence (Fig. 13). Ancestors stock. The plethodontine stock gave rise relatively of Phaeognathus left Appalachia, became terres­ early to three major evolutionary lines (Fig. 12). trial, and occupied the Coastal Plain at the peri­ The first probably diverged from the generalized phery of Appalachia. The Coastal Plain slowly stock at the beginning of Tertiary. The immediate emerged from the sea starting in early Tertiary and factors in the divergence of this group, the fore­ has been in existence since Eocene (see also Gra­ runners of the Bolitoglossini, were the loss of aqua­ ham, 1964). Today Phaeognathus appears to be a tic larvae, acquisition of terrestrial habits, and oc­ Coastal Plain relict, a specialized survivor of a cupation of terrestrial niches by generally adapted group that was probably more widespread during populations with genetic and evolutionary plas­ Tertiary than at present (see Valentine, 1963 a, ticity. Occupation of the new adaptive zone pro­ for further speculations concerning the genus). vided impetus for the subsequent diversification Another divergent desmognathine line is repre­ and adaptive radiation of the Bolitoglossini. sented by Leurognathus, a southern Appalachian Several types of evidence support the theory of endemic. It has adapted to a specialized, strictly early origin of the bolitoglossines. The group today aquatic environment, and it is impossible to date contains three distinct, specialized subgroups. It has its divergence. Adaptation to the new adaptive sub­ differentiated to a greater degree than either of the zone has taken place very slowly and over a long other two tribes. The high degree of diversification period of time. The modem genus Leurognathus is suggests a long history, and morphological differ­ probably the culmination of long Tertiary trends, ences suggest an ancient separation from the an­ and divergence from the ancestral stock has been cestral stock. Bolitoglossines are distantly peri­ more recent and less extreme than that of Phaeo­ pheral to the source and dispersal center of the gnathus. family and are the only plethodontid group that no Desmognathus has remained relatively close to longer occurs either in Appalachia or its border­ the ancestral desmognathine stock, both morpho­ lands. It is the only group that has invaded Europe logically and ecologically. The genus is centered in and the Tropics. Finally, a Batrachoseps trackway Appalachia, but has entered the Coastal Plain, the from the Mio-Pliocene border of California (Pea­ Interior Highlands, and adjacent areas. Movement body, 1959) indicates that the genus was of essen­ into peripheral lowland areas by species such as D. tially modem construction at that early date. The auriculatus and D. fuse us has been accompanied above facts suggest a very early origin of the Boli­ by adaptation to slower moving waters than in the toglossini and an early spread with subsequent dis­ ancestral habitat, and invasion of the Interior High­ placement by other plethodontids. lands has probably been via lowland routes. Several The second evolutionary line, which has given montane species (D. ochrophaeus, D. aeneus, D. rise to the Plethodontini, diverged from the gen­ wrighti) have become increasingly terrestrial. Major eralized plethodontid stock at a later date than the trends in Desmognathus are clearly in the direction bolitoglossine ancestors. This divergence also in­ of dwarfing and increased terrestriality. Desmo­ volved entrance into a new adaptive zone with con­ gnathus is ecologically the most diverse plethodon­ comitant loss of aquatic larvae and occupation of tid genus, but it is remarkable uniform osteologi­ terrestrial niches by generally adapted populations. cally. Most of the osteological differences are re­ There are several reasons for thinking the pletho­ lated to size and development stage. donine divergence followed rather than preceded Evolution within the Desmognathinae and par­ the bolitoglossine evolution. Members of the group ticularly within Desmognathus has proceeded in left Appalachia and reached the Rocky Mountains situ in southern Appalachia. The course of evolu­ and the Pacific Coast, but the group as a whole is tion within the subfamily has involved a successive not as peripherally distributed as the bolitoglos­ occupation of distinct ecological niches in the same sines, and two of the three genera are still repre­ general environmental and geographic region sented by Appalachian species. Modem pletho­ (Dunn, 1917; 1926; Hairston, 1949; Organ, 1961 donines are more primitive morphologically than a) . Radiation within the subfamily is a miniature of the bolitoglossines, but this fact may be due to the that in the family as a whole. more generalized nature of the plethodonines and 40

30

Desmognathus \ Leurognathus \ .. \ Phaeognathus \ Ensatina

Hydromantes

Figure 19. Distribution of Desmognathus, Leurognathus, Phaeognathus, and Ensatina. Figure 20. Distribution of Hydromantes in California. Figure 21. Distribution of Hydromantes in Europe. 1966 EVOLUTION OF LUNGLESS SALAMANDERS 89 their more recent association with the generalized tic Coastal Plain. Numerous primitive features plus plethodontine ancestral stock. several specializations (e.g., skull compression, The final group, the Hemidactyliini, has remain­ habitat) suggest a relatively early divergence from ed primarily in Appalachia and its borderlands. It the tribal stock. Stereochilus divergence and dif­ is closer morphologically to the ancestral stock ferentiation probably dates from the establishment than is any other plethodontid group. Evolution of the Coastal Plain in early Tertiary. and differentiation has been centered in eastern Atwood (1940) divides the Appalachian High­ United States. lands Physiographic Province into two parts: the northeastern division (New England-Acadian) and the southwestern division (Hudson Valley to cen­ History of the Hemidactyliines tral Alabama). Gyrinophilus is found throughout The history of the hemidactyliine genera is both divisions, but Pseudotriton is associated only closely tied to the Cenozoic history of the eastern with the latter. Pseudotriton has a more extensive, United States. Only one genus, Gyrinophilus, is but more southerly range than Gyrinophilus (Fig. now limited in its distribution to the general Appa­ 22 and 24). Although Pseudotriton occurs in lachian region, but two other genera, Eurycea and springs and seeps at high elevations and may be Pseudotriton, occur widely in Appalachia (Fig. sympatric with Gyrinophilus, it can survive in the 22, 24, 25). Appalachia is the center of dispersal warm waters of low elevations: Pseudotriton has of the tribe. Dispersal has been centrifugal in all successfully invaded the Coastal Plain from Louisi· directions from Appalachia. ana to the mouth of the Hudson River, and P. mon· Gyrinophilus is morphologically the most primi­ tanus, in particular, has adapted to very muddy, tive hemidactyliine and it has ·remained close to mucky areas. Thus in its ecology and distribution, the ancestral home. Today the generic range cor­ as well as in its morphology, Pseudotriton demon· responds to the approximate limits of ancient Ap­ strates evolutionary advance over Gyrinophilus. palachia. Gyrinophilus is found in mountain brook Eurycea ranges more widely than any of the habitats, but may occur in springs and woodland above genera and has successfully invaded the seepage areas. Coastal Plain, the Interior Highlands, and the Paedogenetic populations that occur in caves in Balcones Escarpment-Edwards Plateau areas of southeastern Tennessee and northern Alabama Texas in addition to occupying the entire Appa­ have been assigned to Gyrinophilus (G. palleucus) lachian Highland region (Fig. 25). Presumably the on the basis of their resemblances to larvae of genus arose relatively early in Tertiary. Eurycea Gyrinophilus porphyriticus and G. danielsi. The has undergone a moderate evolutionary diversifica­ salamanders have been induced to metamorphose tion. The three widest ranging species, E. longi­ by Dent, Kirby-Smith, and Craig ( 1955), Dent and cauda, E. lucifuga, and E. bislineata, are all found Kirby-Smith (1963), and Blair (1961). These in the Appalachian Highlands, and E. aquatica is authors state that G. palleucus will undergo com­ found on the southern borders of Appalachia in plete metamorphosis in the laboratory, and that northern Alabama. Both longicauda and bislineata metamorphosed individuals closely resemble other also occupy favorable microhabitats (seepages and species of Gyrinophilus; but Blair (1961) reports swamps near springs and rivers) on the Coastal that the premaxillae remain fused, a larval feature. Plain but retain primitive stream larvae. E. quad­ Further, the vomers retain larval configuration. It ridigitata has adapted (pond larvae) to the swamps is apparent from X-ray photographs presented by of the Coastal Plain. Dent and Kirby-Smith (1963) that fused pre­ The ranges of E. lucifuga and E. longicauda ex­ maxillae and larval vomers are also retained in their tend from Appalachia through a narrow corridor supposedly completely metamorphosed specimens. across northwestern Kentucky and southern Illinois In most characters G. palleucus does resemble oth­ and Indiana to the Interior Highlands Physiograph­ er Gyrinophilus, but generic assignment remains ic Province (Atwood, 1940) in Missouri, Arkansas, tentative. It could also be a remnant of a primitive eastern Oklahoma, and extreme southeastern Kan­ Gyrinophilus-like stock. The distribution of G. sas. Two distinct species of Eurycea (tynerensis palleucus is that expected of a paedogenetic species and multiplicata) and the genus Typhlotriton, a - on the periphery of the range of its non-paedo­ Euryceo. relative, are confined to the Interior High­ genetic congeners. lands. Other plethodontines found in the Interior Stereochilus is the derivative of an early hemi­ Highlands include the endemic species, Plethodon dactyliine stock that invaded and successfully . ouachitae and P. caddoensis, and P. glutinosus adapted to the swampy environments of the Atlan- ranges southwest from Appalachia in a pattern sim- 30

Gyrinophi lus Stereochilus Haideotriton

Figure 22. Distribution of Pseudotriton. Figure 23. Distribution of Hemidactylium, Typhlomolge, and Typhlotriton. Figure 24. Distnoution of Gyrinophilus, Stereochilus, and Haideotriton. Figure 25. Distribution of Eurycea. 1966 EVOLUTION OF LUNGLESS SALAMANDERS 91 ilar to that of E. longicauda and E. lucifuga. Dis­ western and southwestern plants (derivatives of the junct western and southwestern populations of P. Madro-Tertiary Geoflora; see below) invaded the cinereus, P. dorsalis, and Hemidactylium scutatum Edwards Plateau. Mesophytic species of the mixed also occur in the Interior Highlands. forest survive as relicts in only a few spots west of Five species of paedogenetic Eurycea (latitans, the Mississippi River. Humid corridors were re­ nana, neotenes, pterophila, troglodytes) and the established during Pleistocene pluvials below the paedogenetic genus Typhlomolge, derived from a glacial borders. The primary corridor between the pre-Eurycea stock, are endemic to the Balcones Appalachians and the Interior Highlands was Escarpment-Edwards Plateau region of central through the Hill Section of the Western Mesophytic Texas. Disjunct populations of P. glutinosus also Forest Region of Braun (1950), which extends occur in the region. For biogeographic purposes from the Knob region of western Kentucky through the entire plethodontine fauna of the Interior the Norman and Crawford uplands of southern In­ Highlands and of Texas is discussed as a unit, al­ diana, and the Ozark Hills of southern Illinois. though both hemidactyliine and plethodontine spe­ This corridor has been especially important for cies are involved. plethodontine distribution. The problem of disjunct populations of Appa­ Plethodontines in the Interior Highlands and lachian organisms in the Interior Highlands and the Edwards Plateau represent three levels of on the Edwards Plateau has been discussed by num­ divergence from the ancient Appalachian popula­ erous authors. The works of Braun (1950), Smith tions. The earliest level includes the two endemic (1957), Dowling (1958), Blair (1958), and Con­ genera, Typhlotriton of the Interior Highlands ant ( 1960) are pertinent to this discussion. Dowling and Typhlomolge of the Edwards Plateau borders. discussed the geology and paleoclimatology of the Ancestors of these genera may have entered the Interior Highlands, and stated that the region has regions in early Tertiary when favorable climatic been above sea level since Pennsylvanian times. conditions prevailed and relatively uniform forest Dating the time of isolation of the Interior High­ communities extended from Appalachia into Mex­ lands is controversial and probably impossible ico and to the West Coast. They have survived with the data at hand, but most authors agree that drier surroundings than are generally considered it was a Cenozoic event; Atwood (1940) suggests suitable for survival of plethodontids by occupying mid-Tertiary as the time of isolation from eastern subterranean aquatic habitats. Paedomorphosis highlands. has been the key to the success and survival of The modem Interior Highlands Physiographic both genera. Typhlomolge is paedogenetic and is Province, as described by Atwood (1940), includes restricted to underground water systems. Typhlo­ the St. Francois Mountains and Salem Plateau, the triton, which is probably not closely related to Ozark Plateau, and the Boston Mountains, all north Typhlomolge, occurs in surface creeks in the larval of the Arkansas River Valley, and the Ouachita state, but the adults are restricted to caverns. Mountains to the south of the valley. The province The second level of differentiation is represented is bounded on the north and west by the Prairie by the endemic species of Appalachian genera Province, on the east by the Mississippi Lowland, which occur in the two regions. These include and on the south by the Coastal Plain. Eurycea latitans, E. nana, E. neotenes, E. ptero­ Braun (1950) has discussed the significance of phila and E. troglodytes of the Edwards Plateau, disjuncts of Appalachian forest plants which occur and E. multiplicata, E. tynerensis, Plethodon ouach­ in the Interior Highlands and canyons in the Ed­ itae, and P. caddoensis of the Interior Highlands. wards Plateau. She interprets these disjuncts as Six of the seven species of Eurycea are paedoge­ relicts of a former more widespread mesophytic netic, and the seventh (multiplicata) is sometimes forest (the East American Element of the Arcto­ neotenic (Conant, 1958). P. ouachitae and P. cad­ Tertiary Geotlora; Axelrod, 1960). Today the In­ doensis are much better adapted to dry conditions terior Highlands are drier than Appalachia and than P. glutinosus (Pope and Pope, 1951). Dowling Oak-Hickory communities dominate. Braun sug­ (1958) has suggested that isolation and differentia­ gests that a mixed mesophytic forest was continu­ tion of the Interior Highland endemics dates from ous across eastern North America and into the Miocene to Pliocene on the generic level and the Mexican Highlands in Early Tertiary. Interior late Pliocene to early Pleistocene on the specific regions were drying in middle Tertiary and the level. Blair (1958) attempted to explain the distri­ mixed forest was forced to retreat northward and bution of the relict species of Plethodon and Eury­ eastward. At this time the Oak-Hickory Associa­ cea on the basis of Pleistocene events. These times tion evolved in the Interior Highlands, and drier seem much too recent. Typhlomolge and Typhlotri- 92 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 ton probably were differentiated when Tertiary Interior Highland populations of H emidactylium drying was beginning to drive plethodontids east­ are relicts of a former more southerly distribution ward. They survived in specialized microhabitats to (see below). ' which they became increasingly more highly The corridor by means of which Pleistocene in­ adapted. The species of Eurycea are probably de­ vasion of the Interior Highlands occurred extends scendents of populations "trapped" in isolated from the Appalachians through the hill sections favorable microhabitats as increasing Tertiary of western Kentucky and southern Indiana, the aridity forced the genus as a whole eastward. Ozark Hills of southern Illinois, and the St. Francois Paedomorphosis bas been a universal means of Mountains and Salem Plateau of eastern and south­ survival for these endemics. They have survived by eastern Missouri. The entire corridor region is oc­ remaining for all (the paedogenetic species) or cupied today by Eurycea longicauda, E. lucifuga, much (multiplicata) of their lives in waters that are and Plethodon glutinosus. Plethodon cinereus ex­ surrounded by terrain largely or wholly unsuited tends south to the Ohio River and west to the for urodele terrestrial or semiterrestrial existence. Illinois border in Indiana and has populations on In these ni'ches they survived the late Pliocene and the Salem Plateau. P. dorsalis extends as· far west Pleistocene droughts. as the Mississippi River, afid Hemidactylium has a Plethodon caddoensis and P. ouachitae have relict Salem Plateau population. To the north of been considered to be members of the yonahlossee the corridor region the distribution of plethodontids species group, the most primitive of the nine groups is limited by prairies, and to the south by the Coastal of Plethodon recognized by Highton (1962). The Plain. The corridor is most constricted in southern other members of the group, P. longicrus and P. Illinois. There is a striking correlation between the yonahlossee, have limited ranges in the southern northern limits of the ranges of several of the Appalachians. Highton suggests that the prototype species, particularly E. longicauda and E. lucifuga, of the group was once widely distributed in the and the maximum southern extent of continental eastern United States. The primitive nature of P. glaciation during the Kansan glacial period (the caddoensis and P. ouachitae, and the fact that they farthest southern glacial penetration) as mapped are adapted to live in relatively dry environments by Smith (1961) for the salamanders, and Flint (Pope and Pope, 1951; Dowling, 1958) suggest a and co-workers (1958) for the glaciers. Glaciated rather early divergence and an in situ adaptation to terrain does not provide suitable habitat for most the increasing dryness of the Interior Highlands plethodontids. from mid-Tertiary to Recent. The highly specialized, paedogenetic Haideo­ The final level of differentiation is represented by triton occurs in limestone caverns and underground populations of primarily Appalachian species that water systems of the Coastal Plain in southwestern are subspecifically distinct from the main popula­ Georgia and northern Florida and was probably tions to the east. These include the Balcones Es­ derived from a Eurycea-Iike ancestral stock rela­ carpment populations of Plethodon glutinosus, the tively early. Invasion of the coastal plain by Ouachita Mountain populations of P. cinereus, the Haideotriton ancestors may have occurred as early Ozark Plateau populations of P. dorsalis, and the as Eocene. Interior Highland populations of Eurycea longi­ The relationship of H emidactylium to other cauda. Undifferentiated Interior Highland popula­ hemidactyliine genera is not clear, but the com­ tions of Eurycea lucifuga and Hemidactylium bination of primitive and specialized characters in­ scutatum, and Balcones Escarpment and Interior dicates a relatively early origin from a hemi­ Highland populations of P. glutinosus also may be dactyliine stock. H emidactylium has departed from included with this group. Some of these popula­ the primitive way of life and is rather terrestrial tions are truly disjunct (Balcones Escarpment P. in habits; it breeds in sphagnum bogs and has pond glutinosus; Interior Highland P. cinereus, P. dor­ larvae. Hemidactylium has followed favorable en­ salis, and H. scutatum); ranges of the other species vironments to the north and northwest and extends are continuous with those of more easterly popula­ considerably farther to the northwest (western tions. I agree with Dowling (1958), Smith (1957), Wisconsin) than any other hemidactyliine (Fig. and Blair (1958) that the distribution of this final 23) . Disjunct relict populations of the genus occur group is the result of Pleistocene and Recent in sphagnum areas in the southern Appalachians, events. It is likely that invasion of the regions oc­ northern Florida, eastern Louisiana, the Ouachita curred from the east during pluvial periods, possibly Mountains of Arkansas, and the Salem Plateau of with the differentiated populations entering rela­ southeastern Missouri. It is apparent that H emi­ tively early, the undifferentiated entering later. The dactylium occupied more southerly regions during 1966 EVOLUTION OF LUNGLESS SALAMANDERS 93 periods of Pleistocene glacial maxima than it does climatic, physiographic, and biotic changes in. its now, and that following the last glacial retreat it former habitat, and survived only in a few rather has spread rapidly into formerly glaciated regions isolated, specialized habitats in widespread seg­ (e.g:, Michigan). The relicts of lower latitudes are ments of its former range. Two species, H. italicus the remnants of populations that once occupied of Italy and France and H. genei of Sardinia, are the Pleistocene refugia (see Smith, 1957). Conant sole living European plethodontids. The route by (1960) thinks the disjunctness of Hemidactylium which Hydromantes entered Europe, probably in in Missouri and Arkansas, of Plethodon glutinosus early Tertiary, was the Bering Land Bridge between in Texas, and of other amphibians and reptiles may northeastern Asia and northwestern North Amer­ be attributed to climatic changes during the Xero­ ica. Simpson (1947) and others think the bridge thermic Interval following Climatic Optimum in was above sea level for most of Tertiary, and was post-Wisconsin times. submerged only for considerable lengths of time in middle Eocene and middle to late Oligocene. It is likely that Hydromantes reached Europe as early The First Terrestrial Encounter as early Eocene through an Arcto-Tertiary Forest -the History of the Bolitoglossines corridor on the Bering Land Bridge. Absence of Divergence of the bolitoglossine ancestral stock Hydromantes from northwestern North America proceeded very rapidly following movement of the and northern Eurasia may be the result of extinc­ primitive bolitoglossine stock across the adaptive tion of most populations during unfavorable cli­ threshold between the aquatic and terrestrial adap­ matic periods of late Pliocene and Pleistocene. tive zones. Establishment of the group probably Modern species of Hydromantes are all limited to occurred very early, possibly in late Cretaceous specialized and rather isolated limestone areas, and or early Tertiary. The group spread out from Appa­ all species, both European and American, have lachia into terrestrial communities dominated by relatively small ranges. elements of the Arcto-Tertiary Geoflora, which ex­ Several authors have proposed theories in con­ tended in a broad belt across Holarctica as a mix­ flict with that suggested above for Hydromantes ture of conifers and deciduous hardwoods. The dispersal. Dunn (1926) suggested Hydromantes Arcto-Tertiary Geofiora linked western Europe, the reached Europe in late Miocene or Pliocene. Noble alpine belt, northeastern Asia, western North (1931) thought plethodontids reached Europe by America, and eastern North America from early way of Greenland "at a time when this northern through middle Tertiary, and provided the con­ region enjoyed a warmer climate." Schmidt (1946) tinuous forest corridor that has presumably been suggested that the relation between west European essential for plethodontid dispersal. and west North American faunas is based on a Bolitoglossines probably reached the West Coast later Tertiary, Pleistocene, and Recent dispersal of North America some time in early Tertiary via across the Bering Land Bridge. terrestrial forest corridor routes. Subsequent di­ The genus Batrachoseps is an ancient group of vergence of the plethodonine terrestrial species highly specialized salamanders. The most primitive may have brought plethodonines and bolitoglos­ and generalized species is B. wrighti, which occu­ sines into competition, with resultant extinction of pies forests derived from the Arcto-Tertiary Geo­ the less well adapted bolitoglossines in Appal~chia flora in the Cascade Mountains in northern and and surrounding eastern North America. Absence central Oregon. The species is the only one re­ of bolitoglossine species from the Great Plains and stricted to the presumed ancestral environment. Rocky Mountains is almost certainly related to B. attenuatus ranges from southern Oregon to Tertiary drying and cooling, and inhospitable southern California, and occurs in a variety of habi­ Quaternary climates. tats-in forests derived from the Arcto-Tertiary Three distinct groups of bolitoglossines survive Geofl.ora to areas dominated by derivatives of the today, the supergenera Hydromantes, Bolitoglossa, Madro-Tertiary Geofiora. B. pacificus of the low­ and Batrachoseps. Of these Hydromantes is the lands of Southern California and northern Baja most primitive and may be relatively little changed California is restricted to coastal plant communi­ from the ancestral bolitoglossine stock (Fig. 16). ties derived from Madro-Tertiary elements. Two The present range of Hydromantes is a classic and possibly more undescribed species occupy dis­ example of disjunct distribution and relict popula­ junct, relict ranges in areas of Arcto-Tertiary de­ tions (Figs. 20 and 21). It is apparent that a once rived plant communities in the Santa Lucia Moun­ widespread, primitive, but rather specially adapted tains and southern Sierra Nevada of California, and ancestral group was unable to adapt to the Tertiary the high Sierra San Pedro Martir of northern Baja 94 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 California. This distributional pattern (Fig. 26) plant and animal. The entire assemblage shifted to suggests a once far-ranging group which bas been the south, especially at high elevations, as the restricted to relatively small areas in what are for tropical border continued to move southward in plethodontid salamanders specialized environ­ Oligocene and Miocene in response to long range ments. Evidence that the genus once had a more Tertiary cooling trends. extensive range is based on two marginal records. The Madro-Tertiary Geofiora was developing in The first is the collection of the types of Batracho­ the increasingly arid southwestern United States seps caudatus from southeastern Alaska (Cope, and northern Mexico from Eocene throughout 1889). No further individuals of the species have Tertiary. Increased lowland aridity favored expan­ been found, and subsequent authors (Dunn, 1926; sion of the Geotlora between the main portions of Stebbins, 1951) have questioned the authenticity the Arcto-Tertiary belt to the north and the Arcto­ of the locality data. There has been little collecting Tertiary-Neotropical-Tertiary "ecotone" isolated in in Alaska and there is little reason to question the the highlands to the south. Presumably the long­ record. The second is the collection of a single im­ range Tertiary climatic trends resulted in effective mature individual of Batrachoseps from the north­ separation of the ancestral populations of the super­ ern slope of the Nevado de Colima at about 7,000 genus Bolitoglossa from other bolitoglossine popu­ feet elevation in the state of J alisco, Mexico. The lations to the north. collector was Hans Gadow, a noted biologist, who It is apparent that Appalachia is the primary discussed conditions of the collection in some detail ·dispersal center of plethodontid salamanders, but in two separate publications (1905, 1908). He was it is equally apparent that the moist borders of the aware of the significance of his discovery from the Mexican Plateau have acted as a secondary center first. The specimen, in the British Museum, has of evolution and dispersal. Seven genera and over been examined by Stebbins and Lowe ( 1949), Hen­ one hundred species of bolitoglossine salamanders drickson (1954), Arden H. Brame, Jr. and me. All have been derived from the ancestral Tertiary popu­ concur that it is a Batrachoseps. The Mexican rec­ lations. A number of species remain in the ancient ord makes good biogeographic sense for a bolito­ "ecotonal" situation, but several lines have suc­ glossine genus, and I consider it to be authentic. It cessfully adapted to other habitats where they are is possible that other populations of Batrachoseps associated with derivatives of the Neotropical occur as relicts in the Sierra Madre Occidental of Tertiary Geofiora. Significantly the more struc­ Mexico. Failure of collectors to obtain additional turally primitive and generalized species have material may be due to the relative lack of intensive remained in the highland "ecotone." collecting during the right seasons in the critical Three distinct lines of tropical bolitoglossine evo­ areas. lution are apparent (Fig. 17). Perhaps the earliest The entrance of the supergenus Bolitoglossa into divergence from the ancestral stock was by Bolito­ Mexico has recently been discussed in some detail glossa. The genus is found in the Mexican High­ (Brame and Wake, 1963), and only a review will lands but has also extended into the Central Ameri­ be presented here. It is proposed that a bolito­ can and South American highlands as far south as . glossine stock ancestral to the supergenus Bolito- Ecuador and Venezuela. Throughout its range glossa was associated in early Tertiary with elements various species have invaded and adapted to the of the Arcto-Tertiary Geofiora at relatively high lowland habitats where many species have become latitudes. Axelrod (1960) has stated that a mixing arboreal. Lowland species of Bolitoglossa range of elements of the Arcto-Tertiary and Madro­ from Mexico through Central America to Bolivia Tertiary Geofioras started in middle Cretaceous and Brazil (Figs. 27 and 29). Invasion of the low­ and formed what he calls a broad "ecotone." The lands and acquisition of arboreal habits has ap­ "ecotone" extended southward along low mountains parently occurred in parallel several times in the into Mexico during Cretaceous and early Eocene. genus. Highland species are usually both more Species of plants, or paired-species, which now oc­ primitive and more generalized than lowland cur disjunctly in Appalachia and the Mexican and species. With only a few exceptions (e.g., rufescens, Guatemalan highlands, probably entered the area occidentalis) the more northerly species of any at this time (see also Braun, 1950; Martin and habitat type are more primitive and generalized Harrell, 1957); these plants are relicts of the old than their more southerly ecological equivalents. world-wide Arcto-Tertiary Geofiora, and similar Stuart (1954) has suggested that Magnadigita (= species occur today in eastern Asia. Plethodontid primitive highland Bolitoglossa) is autochthonous salamanders may also have entered the "ecotone" in Central America, and has spread north and south along with their Arcto-Tertiary associates, both from the Guatemalan Highlands. What may be the 1966 EVOLUTION OF LUNGLESS SALAMANDERS 95

120 100

50 I

40

30

Batrachoseps

SCALE

10 200 400 600 800 1000 1200 1400 KILOMETERS

Figure 26. Distribution of Batrachoseps. 96 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4

100

0 MilH

Oedipina

____j .... Mi lea

Figure 27. Distribution of Bolitoglossa and Thorius in Middle America. Figure 28. Distribution of Oedipina in Middle America. The genus also has been reported from Chiapas, Mexico. Figure 29. Distribution of Bolitoglossa and Oedipina in South America. 1966 EVOLUTION OF LUNGLESS SALAMANDERS 97 most primitive species is a Guatemalan highland main populations to the south by dry gaps (e.g., form, B. dunni, which retains columellae, pre­ San Louis Potosian Gap). Martin (1958 a) refers frontals, primitive vertebrae, long limbs, slightly these populations to a humid montane group of the webbed hands and feet, primitive numbers of Northeast Madrean Herpetofaunal Component. carpals and tarsals, tibial spurs, and primitive The salamanders of this component must be con­ vomers. The most specialized species appear to sidered one of several in the Central American be B. rufescens of the northern lowlands and mem­ Complex of the Old Northern Herpetofaunal Ele­ bers of the South American palmata and altama­ ment of Savage (1960). Martin (1958 a) suggests zanica groups (Brame and Wake, 1963). that the northeastern isolates reflect post-pluvial A second phylogenetic line contains the highly confinement of a habitat formerly widespread. Sev­ specialized, greatly elongated species of Oedipina. eral species (C. multidentatus, P. bellii, P. cepha­ These bizarre organisms are attenuate forms with lica) occur on both sides of the ecological barrier elongated trunks, reduced limbs, and extremely and the suggestion seems reasonable. long tails (often over seventy-five per cent of total Both Pseudoeurycea and Chiropterotriton have length). The high degree of specialization, adapta­ crossed the Isthmus of Tehuantepec and entered tion to lowland environments, and penetration of the mountains of southern Mexico and Guatemala. South America all suggest early divergence from Chiropterotriton extends even farther to the south the common tropical bolitoglossine ancestral stock. into northeastern Costa Rica. The increasing spe­ Two major sections are evident. One contains rela­ cialization of the southern Chiropterotriton has tively short-bodied forms with few or no maxillary been previously discussed. Stuart (1957) has sug­ teeth and relatively large, essentially fully webbed gested that Pseudoeurycea moved southward over feet (elongatus, parvipes, complex); they are found the high plateaus of Chiapas and Guatemala after over the entire range of the genus from Mexico, crossing the low, hot Isthmus of Tehuantepec in the British Honduras, and Guatemala to north-central Pleistocene. Stuart ( 1954) thinks that Pseudoeury­ Ecuador. The other section includes relatively long­ cea is unquestionably a recent entrant into Nuclear bodied species with many maxillary teeth and rather Central America from Mexico, and the slight small, syndactylous feet (all other species) which penetration of Pseudoeurycea into Central America range from Guatemala to western Panama. Oedi­ supports this suggestion. Stuart (1954) also sug­ pina is the only plethodontid genus found primarily gests that Chiropterotriton reached Guatemala by in tropical lowlands and is the only tropical genus relatively low Pacific mountain-slope routes. In that does not occur in the highlands along or near view of the great differentiation and greater south­ the southern and eastern margins of the Mexican ward penetration of Chiropterotriton, it is possible Plateau. Costa Rica is the present dispersal center that the genus may have crossed the Isthmus of and apparently has been the center of evolution Tehuantepec at an earlier date than Pseudoeurycea. of this, the most specialized and advanced pletho­ Numerous biogeographic problems of consider­ dontid genus (Figs. 28 and 29). able interest and importance to the local and The third and final line includes those species regional distribution of the genera Pseudoeurycea, now assigned to the genera Pseudoeurycea, Chi­ Chiropterotriton, Bolitoglossa, and Oedipina in ropterotriton, Thorius, Parvimolge, and Linea­ Guatemala are discussed by Stuart ( 1935, 1943, triton. All are found in mountains in and around 1948, 1950, 1951, 1954, 1957, 1958) and Schmidt the southern margins of the Mexican Plateau (Figs. (1936). 27, 30, 31). Thorius, Lineatriton, and probably To summarize, the history of the neotropical ·Parvimolge (depending on the assignment of the plethodontids has involved mainly southward Costa Rican species, richardi) are limited to the movements along the Central American axis. Major area north of the Isthmus of Tehuantepec. The movements have probably been via highland routes. primitive species of both Pseudoeurycea and Chi­ Two genera, Bolitoglossa and Oedipina, have been ropterotriton are also found north of the Isthmus. in Nuclear Central America for a long time and The mountainous southern and eastern margins of may have arisen there. Others (Pseudoeurycea, the Mexican Plateau have been the center of Chiropterotriton, ?Parvimolge) are probably more evolution of the group. recent invaders. It seems likely that northward Pseudoeurycea and Chiropterotriton extend far­ movements of Bolitoglossa and southward move­ ther to the northeast than any other bolitoglossines, ments of Pseudoeurycea across the Isthmus of occurring in the mountains of southwestern Tehuantepec by highland routes may have occurred Tamaulipas and southern Nuevo Leon. These during Pleistocene. Stuart (1954) has stated that northeastern populations are separated from the geographic continuity existed during Pleistocene, ...... ·~ '9 ,•

C hiropterotriton . Lineatriton • I

1 Parvimolge * 'L.·Ps,!r-•-u_d_o• __ u_r_y_c_•_a~f!!i~·~,~~~~.l ·-~~~~~'--~­ 1• IGO Figure SO. Distribution of Chiropterotriton and Lineatriton. Figure 31. Distribution of Pseudoeurycea and Parvimolge. 1966 EVOLUTION OF LUNGLESS SALAMANDERS 99 when a temperature drop of about 5° C would have learned about the history of Aneides by studying permitted the majority of highland species to move the history of the geofl.ora. All plethodonine genera across Nuclear Central America. While the degree are now members of communities in which fl.oral of southward distribution is dependent in part upon elements derived from the Arcto-Tertiary Geo­ the time of entrance into Nuclear Central America, flora dominate or play important roles, so that the vagility of the group must also be considered Lowe's suggestions concerning Aneides are at least (Stuart, 1954). Bolitoglossa and Oedipina are in part applicable to the tribe as a whole. clearly early derivatives of the ancestral· stock, but The history of the Arcto-Tertiary Geofl.ora in the greater southward penetration of Bolitoglossa North America has been outlined in broad form may be more a reflection of greater vagility than of by Chaney, Condit, and Axelrod (1944), Axelrod greater antiquity. (1948, 1950, 1956, 1957, 1960), Braun (1950, 1955), and MacGinitie (1958). In early Tertiary the Tropical-Tertiary Geofiora formed a broad belt The Second Terrestrial Encounter which in North America extended as far north -the History of the Plethodonines as southeastern Alaska on the West Coast and pos­ The second terrestrial group, the Plethodontini, sibly to Nova Scotia on the East Coast (Axelrod, is found today in eastern and western North 1960). Above the tropical border was the circum­ America, with several Rocky Mountain isolates. polar Arcto-Tertiary Geoflora. Tertiary cooling and The ancestral stock probably diverged from a drying trends resulted in a southward shift of the primitive plethodontine group in what is now Appa­ tropical border and a concomitant southward shift lachia later than the divergence of the bolitoglossine of the Arcto-Tertiary Geoflora. A broad "ecotone" stock. Neither terrestrial group is ancestral to the formed where the two major fl.oral groups met. As other. Although bolitoglossines may have arisen a result of climatic changes during middle and late earlier, both groups were probably well established Tertiary a number of segregated elements evolved, by Oligocene. two of which, the East American Element and the There are three genera of plethodonines. Aneides West American Element (Chaney, Condit, and has one species (aeneus) in southern Appalachia, Axelrod, 1944; Axelrod, 1960), are of primary one (hardii) in the high mountains of south-central importance to this discussion. New Mexico, and three (ferreus, ffavipunctatus, Ancestral plethodonines invaded and adapted to lugubris) on the Pacific Coast (Fig. 32). Plethodon the terrestrial environment in the Arcto-Tertiary has eleven species in eastern North America of forests of Appalachia and began to move westward which six (welleri, richmondi, wehrlei, jordani, in the favorable habitats provided by the forests, longicrus, yonahlossee) are primarily Appalachian probably sometime in early Tertiary. Secular cool­ species, two (ouachitae, caddoensis) are endemic to ing and drying trends reinforced by the rainshadow the Interior Highlands, two (cinereus, dorsalis) oc­ following the uplift of the Rocky Mountains dis­ cur in both areas, and one (glutinosus) occurs in rupted the Arcto~Tertiary belt in central North Appalachia, the Coastal Plain, the Interior High­ America. Forests were replaced by savannah and lands, and the Edwards Plateau of Texas. Five possibly some grassland. It is difficult to date the species of Plethodon ( elongatus, dunni, larselli, disjunction, but Miocene is my best estimate. In an stormi, vehiculum) are restricted to the Pacific important discussion of late Tertiary biogeography Coast of northwestern United States and southern of the Great Basin, Shotwell ( 1961) reviewed the British Columbia, one ( vandykei) occurs in western distribution of fossil horses from late Miocene to Washington, northern Idaho, and northwestern the end of Tertiary. He records Merychippus Montana, and one (neomexicanus) occurs in the (Protohippus) from as far north as northwestern highlands of north-central New Mexico (Fig. 33). Montana as early as the Barstovian faunas (late The single species of Ensatina (eschscholtzii) ranges Miocene). This is an indication of savannah and from British Columbia to extreme southern Califor­ possibly grassland habitats (see Shotwell, 1961; nia along the Pacific Coast and the slopes of the 1963, for correlations of fossil horse genera and Cascade-Sierra Nevada system (Fig. 19). habitats). He also indicates that central North Lowe (1950) suggested that the history of America was covered by dry facies of Arcto­ Aneides is related to the Cenozoic history of the Tertiary vegetation during late Miocene; such a Arcto-Tertiary Geoflora in North America. It is situation was probably unsuitable for continuous assumed by Lowe and by me that the present day plethodontid occupation. Pliohippus is present in association of Aneides with Arcto-Tertiary floral Clarendonian faunas (early Pliocene) in many elements is an ancient one and that much can be areas of central United States, coinciding with the I ~- /

30

I

30

Plethodon

Figure 32. Distribution of Aneides. Figure 33. Distnoution of Plethodon. 1966 EVOLUTION OF LUNGLESS SALAMANDERS 101 expansion of open grassland (Shotwell, 1961). and Ting (1960, 1961), and Shotwell (1961, 1963) However, MacGinitie (1962) doubts that large are particularly pertinent to this discussion. De­ grassland areas existed until Pleistocene or Recent velopment of the xeric Madro-Tertiary Geoflora times. MacGinitie has shown that what is now in the interior southwestern United States began in northern Nebraska was a savannah region in late early Tertiary. During Tertiary there was a general Miocene with N eotropical-Tertiary relicts, pine-oak drying of western North America favoring expan­ forest elements (Madro-Tertiary), and a swamp sion of the Madro-Tertiary elements at the expense cypress component of the East American Element of the Arcto-Tertiary elements. On the basis of of the Arcto-Tertiary Geoflora. All such habitats geofloral distribution and the distribution of fossil are now unsuitable for plethodontid dispersal. Thus horses, Shotwell (1961, 1963) has made the fol­ at some time from middle Miocene to early Pliocene lowing suggestions concerning the history of the the environment of central North America became Great Basin in late Tertiary. In Barstovian times inhospitable for plethodontids, and the species in the mesic Arcto-Tertiary elements extended the east and in the west were separated. throughout the Great Basin but were mixed with The theory proposed here for dating of the dis­ drier Madro-Teniary elements in the south. Shot­ junction of the ranges of plethodonine genera is in well states that by Clarendonian times the Madro­ conflict with that of Dunn (1926), who did not Tertiary Geoflora extended northward, and Arcto­ have the benefit of the important paleobotanical Tertiary elements had largely been displaced from literature of the past thirty years. My theory is in the central Great Basin except for moister local fairly close agreement with that of Lowe (1950), areas (high elevations, stream borders). Pliohippus which placed disjunction of Aneides in late Mio­ has been found in Clarendonian deposits in much cene with complete disjunction during early of the Great Basin. This fact indicates that savan­ Pliocene. nah or grassland was present. Both habitats are Plethodonine genera were well established by the unfavorable for plethodontids. By Hemphillian time of the disjunction. Two of the three genera times (middle to late Pliocene) increasing cold (Aneides, Plethodon) occur on both sides of the stopped the northward extension of the Madro­ grassland barrier, and forest corridors are not Tertiary elements, but the present semiarid charac­ known to have existed across the barrier. ter of the Great Basin was already established. No Dating of the disjunction of the ranges of Pletho­ woodland elements were present in Hemphillian don and Aneides provides some indirect evidence times in the northern Great Basin (Shotwell, 1963). on the relative age of the genera. Divergence of Later Tertiary and Quaternary changes in the Aneides from a Plethodon-like ancestral stock must flora of the Great Basin were the result of secular have occurred at an early date. By the time of the drying and cooling reinforced by Pleistocene warp­ range disjunctions both genera were widespread, ing and faulting which raised the Sierran crest from perhaps with transcontinental distributions. Generic five thousand to nine thousand feet (Axelrod, 1950; differentiation may date from Oligocene or early 1957; Axelrod and Ting, 1960; 1961). This uplift, Miocene. according to Axelrod (1957), would have ac­ Aneides and Ensatina may have arisen inde­ counted for a ten to fifteen inch reduction in rainfall pendently from Plethodon in western North Amer­ to the east. Increasing continentality of climate, ica. While the most primitive species of Plethodon rain shadow effect of the Sierran uplift, and severe are found in eastern North America (Highton, cooling of Pleistocene glacial periods all con­ 1962), the most primitive Aneides (hardii, see tributed to the disappearance of plethodontids over Wake, 1963) is found in the west (New Mexico) most of the Rocky Mountain and interior range of and Ensatina is limited to the Pacific Coastal re­ the group. The distribution of the West Coast gion. Four of the five Aneides are western species species was further severely limited by the major and the fifth (aeneus of Appalachia) is a morpho­ Sierra Nevada uplift of Pleistocene times (see Axel­ logically specialized species that occupies a highly rod and Ting, 1961). specialized arboreal-rock crevice habitat (Gordon, Three species of plethodonines occur today in 1952; Wake, 1963). the Rocky Mountains. Plethodon vandykei of Plethodonines were probably widespread in coastal Washington has disjunct populations in western United States in Miocene. Elimination of northern Idaho and Montana. Several other am­ the group from most of the area east of the Cascade. phibians are also found in both areas (Ascaphus Sierra Nevada divide and west of the Great Plains truei, Dicamptodon ensatus, Taricha granulosa) occurred subsequent to niidcontinental disjunction. and at least one lizard (Gerrhonotus coeruleus) is The researches of Axelrod (1957, 1958), Axelrod distributed continuously across intervening moun- 102 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 tain ranges in southern British Columbia. Disjunct CONCLUSIONS AND SUMMARY distribution of P. vandykei is probably the result The large, diverse family Plethodontidae con­ of Pleistocene or Recent events. tains more than two-thirds of the living species P. neomexicanus and A. hardii of New Mexican of salamanders. The morphology, ecology, and be­ highlands are apparently relicts of the mid-Tertiary havior of the species have shown them to be the Rocky Mountain fauna as suggested by Lowe most progressive and advanced group within the (1950). Murray (1957) thinks final isolation may order Caudata. The general adaptive level of the have occurred as late as late Pleistocene. Martin family is high, relative to other salamanders, and a (1958 b) states that 4,000 to 4,500 foot altitudinal successful radiation of specially adapted species displacement of biotic zones occurred during full­ has occurred. Because so many of the adaptive glacial periods in the Southwest. It is possible that types have survived, the family has been ideal for final southern displacement of the species may have the study of major features of evolution, especially occurred at that time. Survival of the two species the morphological bases for adoption of new ways is related to the fact that they have been able to of life and the ensuing adaptive radiations that adapt to favorable microhabitats in the relatively typically characterize such events. inhospitably cold Rocky Mountains. A functional and evolutionary morphological Blair (1958) has suggested that P. neomexicanus analysis has been employed to outline the evolu­ may have arisen from populations that dispersed tionary patterns of individual bony elements, func­ directly across Oklahoma in Wisconsin times, and tional anatomical units, and groups of organisms. that A. hardii may have been connected with the This analysis has led to a new classification of the eastern population as recently as late Pleistocene. family and has formed the basis of theories con­ The New Mexican highland flora as a whole is cerning major features of evolution within the clearly derived from the West American Element family. of the Arcto-Tertiary Geoftora. This indicates a The twenty-three genera and one hundred eighty­ western, not eastern origin for the salamanders as four species of plethodontids may be grouped in well (see Chaney in Chaney, Condit, and Axelrod, two subfamilies based on morphological evidence. 1944). Both species are isolated at altitudes above The more specialized subfamily Desmognathinae 8,000 feet. They are well defined species, not closely is the smaller and has the more restricted geographic related to any others. It is likely that they are dis­ range. Its three genera (ten species) live in eastern placed southern relicts of the ancient Rocky Moun­ North America, centered in the southern Appa­ tain fauna, and their evolutionary history is not lachian mountains, an area presumed to be the an­ related strictly to Pleistocene events. cestral home of the family. The larger and more Pacific Coast species have undergone a moderate generalized subfamily Plethodontinae includes amount of diversification. Ensatina eschscholtzii is twenty genera grouped in three tribes. The most the most primitive in morphology, the most gen­ primitive tribe, the Hemidactyliini (eight genera, erally adapted in ecology, and the most widespread twenty-four species) is concentrated in eastern of the western plethodonines. Other species have North America. The tribe Plethodontini (three more restricted ranges, and are generally more genera, twenty-four species) occurs in both eastern specialized (see Peabody and Savage, 1958, for and western North America. The remaining nine detailed biogeographic analyses of western species). genera are included in the most advanced tribe, To summarize, the plethodonines occurred all the Bolitoglossini (one hundred twenty-six species), across the continent in early Tertiary. Middle and a group that ranges from western North America late Tertiary events are responsible for the present into the Neotropics, and has two species in Europe. generic disjunctions. Plethodonines are more suc­ It is the only plethodontid group that lacks repre­ cessful than other high latitude terrestrial genera sentatives in eastern North America. and were apparently the victors in competitive en­ The two subfamilies of lungless salamanders counters with the bolitoglossines in the east. West­ differ markedly in the functional morphology of the ern plethodonines have also had notable success, head. In the Desmognathinae the lower jaw can and the bolitoglossine genera have survived the be lowered only until a ligament extending from the plethodonine invasion only by specializing. The mandible to the atlas vertebra becomes taut. The failure of plethodonines to reach the tropics and skull is then raised to open the mouth further. the Old World is probably the result of relatively Morphological modifications directly and indirectly recent derivation from the plethodontine ancestral associated with this functional shift are great and stock. the desmognathine salamanders differ considerabiy 1966 EVOLUTION OF LUNGLESS SALAMANDERS 103 from the more generalized plethodontines. The in tropical environments and has contributed to a three tribes of plethodontines differ mainly in relatively large scale adaptive radiation and evolu­ anatomical details, such as hyobranchial configura­ tionary diversification. tion, form of cranial elements, and vertebral The evolution of a great variety of special adap­ differences. tive features has been associated with the adaptive Each of the four major groups of plethodontids radiation of the lungless salamanders. Many ex­ has undergone a secondary adaptive radiation. As amples have been cited, among them new feeding a result structural and ecological parallelism are mechanisms in desmognathines, skull strengthening common phenomena in the family, and include the adaptations in forms that force their way head first following examples. Tongues are primitively at­ under rocks or make burrows ( desmognathines) tached to the front of the mouth in salamanders, and in forms that have adopted new feeding tech­ including the Desmognathinae and Plethodontini. niques (Aneides), limb shortening in burrowers of In one genus each of the Hemidactyliini several genera, toe loss, basal tail constriction and (Hemidactylium) and Bolitoglossini (Batracho­ adaptations for tail autotomy, specialized climbing seps) tongues are attached, although in a modi­ adaptations in the limbs, hands and feet, eye re­ fied form, to the anterior margins of the mouth. duction and loss in cave dwellers, size diminution, Parallel evolution of free tongues has occurred in gigantism, and others. The radiation has been ex­ these two tribes and free tongues are found in tensive and complex, with the adaptations appear­ nearly all species of the groups. In addition there ing in parallel and multiple parallel lines found is a trend in the direction of tongue freedom in within the family and familial subgroups. The re­ some Plethodontini (Ensatina). Primitive members sult is an evolutionary pattern that is characterized of the family have aquatic larvae and are aquatic by the mosaic distribution of primitive and ad­ to semiaquatic as adults. There are distinct parallel vanced, generalized and specialized characters on trends in the direction of terrestrialism with loss of all taxonomic levels. aquatic larvae. All species of the Bolitog1ossini and Knowledge of the late Mesozoic and Cenozoic Plethodontini are terrestrial and there are definite history of the Northern Hemisphere, fossil history tendencies in the direction of terrestrialism in cer­ of the family, ecology and distribution of the living tain species of the other groups as well (Desmogna­ species, and evolutionary relationships within the thus aeneus and wrighti; Hemidactylium scutatum). family has formed the basis for a theoretical con­ Elongated trunks have evolved in parallel in each sideration of the phylogeny and biogeography of of the four groups and within genera of a single the family. Ancestors of the family were probably group. Basal tail constriction and associated special­ derived from a stock close to that which gave rise izations have arisen independently in the three to ambystomatid salamanders. Appalachia is the tribes. The fifth toe has been lost in three separate present center of dispersal and earlier suggestions genera. Many additional instances of parallelism that this region is the source area are accepted. have been cited. Parallelism is the rule rather than Divergence of the desmognathine line from the the exception in the evolution of the family. main line of plethodontid evolution occurred very One genus dominates each group in terms of early, probably by the beginning of Cretaceous numbers of species and size of geographic range times. Plethodontine tribal divergence probably (Desmognathus, Eurycea, Plethodon, Bolitoglossa). dates from earliest Tertiary. The first plethodontine Within each of these genera a tertiary level of adap­ group to differentiate may have included the ter­ tive radiation bas occurred. Diversity of size, habi­ restrial ancestors of the modern bolitog1ossines. The tat, behavior, and coloration characterizes each of plethodonines are thought to have arisen after the the genera. bolitoglossines because they have a less extensive Paedomorphosis has been a very important range, have undergone a less extensive adaptive morphological mode of evolution in the family. radiation and diversification, and they have living Species of four genera of hemidactyliines are representatives in eastern North America. Possibly paedogenetic. This developmental phenomenon has competition with the more vigorous plethodonines permitted survival of specialized forms in perma­ led to elimination of the older and more specialized nently aquatic and underground habitats in regions bolitoglossines from eastern North America. The unsuitable for their more generalized relatives. Hemidactyliini represent the diversification of the Paedogenetic species are so highly specialized that ancestral stock. Some of the genera (e.g., H emidac­ they are .probably evolutionary dead-ends. By con­ tylium) diverged from the stock very early, perhaps trast the phenomenon of differential metamorphosis as early as the ancestors of the other tribes. The has given bolitoglossines access to new ways of life group is thus less compact than the other two tribes. 104 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 There is evidence that the family remains evolu­ mander from Texas. Texas Jour. Sci., 9(3): tionarily active. Entrance into the tropics has given 328-336. the salamanders access to new ways of life, and 1961. Distribution of and key to the neotenic such genera as Thorius, Oedipina, and Bolitoglossa Eurycea of Texas. Southwestern Nat., 6( 1): may be initiating new phylogenetic trends. Activity 27-32. of these groups is attested by the facts that forty Bishop, S. C. per cent of all living species of salamanders are 1941. The salamanders of New York. New York found within a single supergenus that ranges from State Mus. Bull., 324: 1-365. Mexico into South America, and more than twenty­ 1943. Handbook of salamanders. Ithaca, New five per cent of the species of plethodontid sala­ York: Comstock Puhl. Co., Inc., 555 pp. manders are members of the genus Bolitoglossa. Blair, A. P. 1961. Metamorphosis of Pseudotriton palleucus with iodine. Copeia, 1961 (4): 499. LITERATURE CITED Atwood, W.W. Blair, W. F. 1940. The physiographic provinces of North 1958. Distributional patterns of vertebrates in the America. Boston: Ginn and Co., 536 pp. southern United States in relation to past and present environments. In Hubbs, C. L. Auffenberg, W. (ed.), Zoogeography. Washington: Amer. 1961. A new genus of fossil salamander from Assn. Adv. Sci., publ. 51(17): 433-468. North America. Amer. Midi. Nat., 66(2): 456-465. Blanchard, F. N. 1923. The life history of the four-toed salamander. Axelrod, D. I. Amer. Nat., 62: 156-164. 1948. Climate and evolution in western North America during middle Pliocene time. Evo­ Boulenger, G. A. lution, 2(2): 127-144. 1882. Catalogue of the Batrachia Gradientia s. Caudata in the collection of the British 1950. Evolution of desert vegetation in western Museum, ed. 2. London: British Museum, North America. Carnegie Inst. Wash. Publ., 127pp. 590: 215-306. 1956. Mid-Pliocene floras from west-central Ne­ Brame, A. H., Jr. vada. Univ. California Publ. Geol. Sci., 33: 1960. The salamander genus Oedipina in North­ 1-322. ern Central America. Bull. So. Calif. Acad. 1957. Late Tertiary floras and the Sierra Nevadan Sci., 59(3): 153-162. uplift. Bull. Geol. Soc. Amer., 68: 19-46. 1958. Evolution of the Madro-Tertiary Geoflora. Brame, A. H., Jr. and D. B. Wake Bot. Rev., 24: 433-509. 1963. The salamanders of South America. Los 1960. The evolution of flowering plants. In Tax, S. Angeles Co. Mus. Contrib. Sci., 69: 1-72. (ed.), Evolution after Darwin. Chicago: the University of Chicago Press, Vol. 1, The Brandon, R. A evolution of life, pp. 227-305. 1965. Morphological variation and ecology of the salamander Phaeognathus hubrichti. Copeia, Axelrod, D. I. and W. S. Ting 1965(1): 67-71. 1960. Late Pliocene floras east of the Sierra Ne­ 1966. A re-evaluation of the status of the sala­ vada. Univ. California Puhl. Geol. Sci., mander Typhlotriton nereus Bishop. Copeia, 39(1): 1-118. 1966(3): 555-561. 1961. Early Pleistocene floras from the Chagoopa surface, southern Sierra Nevada. Univ. Cali­ Braun, E.L. fornia Publ. Geol. Sci., 39(2): 119-194. 1950. Deciduous forests of eastern North Ameri­ ca. Philadelphia and Toronto: Blakiston, Baird, I. L. S96pp. 1951. An anatomical study of certain salamanders 1955. The phytogeography of the unglaciated east­ of the genus Pseudoeurycea. Univ. Kansas ern United States and its interpretation. Sci. Bull., 34, pt. 1(6): 221-265. Bot. Rev., 21: 297-375.

Baker,J.K. Burger, W. L., H. M. Smith, and F. E. Potter 1957. Eurycea troglodytes: a new blind cave sala- 1956. Another neotenic Eurycea from the Ed- 1966 EVOLUTION OF LUNGLESS SALAMANDERS 105 wards Plateau. Proc. Biol. Soc. Wash., 63: Dent,J.N. 51-58. 1942. The embyronic development of Plethodon cinereus as correlated with the differentia­ Carr, A. F,, Jr. tion and functioning of the thyroid gland. 1940. A contn'bution to the herpetology of Flori­ J. Morph., 71: 577-601. da. Univ. Florida Publ. Biol. Sci. Ser., 3(1): 1-118. Dent, J. N. and J. S. Kirby-Smith 1963. Metamorphic physiology and morphology Chaney, R. w.. C. Condit, and D. I. Axelrod of the cave salamander, Gyrinophilus pal­ 1944. Pliocene ftoras of California and Oregon. leucus. Copeia, 1963(1): 119-130. Carnegie Inst. Washington Publ., 553: 1- 407. Dent, J. N., J. S. Kirby-Smith, and D. L. Craig 1955. Induction of metamorphosis in Gyrinophi­ lus palleucus. Anat. Rec., 121-129. Conant,R. 1958., A field guide to reptiles and amphibians of the United States and Canada east of the Dowling, H. G. 100th meridian. Boston: Houghton Miftlin, 1958. Geographic relations of Ozarkian amphib­ 366pp. ians and reptiles. Southwestern Nat., 1 ( 4) : 174-189. 1960. The queen snake, Natrix septemvittata, in the interior highlands of Arkansas and Mis­ souri, with comments upon similar disjunct Dundee, H. A. distn'butions. Proc. Acad. Nat. Sci., Phila­ 1957. Partial metamorphosis induced in Typhlo­ delphia, 112(2): 25-40. molge rathbuni. Copeia, 1957(1): 52-53. 1962. Response of the neotenic salamander H aide­ Cooper. o.w. otriton wallacei to a metamorphic agent. 1965. fDd.uction of somite chondrogenesis by car­ Science, 135: 1060-1061. dlage and notochord. A correlation between inductive aetivities and specific stages of Dunn,E.R. cytodifferentiation. Devel. Biol., 12: 185- 1917. The salamanders of the genera Desmogna­ 212. thus and Leurognathus. Proc. U.S. Natl. Mus., 53: 393-433. Cope,E.D. 1922. The sound-transmitting apparatus of sala­ 1859. On the primary divisions of the Salaman­ manders and the phylogeny of the Caudata. dridae with descriptions of two new species. Amer. Nat., 56: 418-427. Proc. Acad. Nat. Sci., Philadelphia, 1859: 1923. Mutanda herpetologica. Proc. New England 122-128. Zool. Club, 8: 39-40. 1866. On the structure and distribution of the 1926. The salamanders of the Family Plethodon­ genera of the arciferous Anura. J. Acad. tidae. Northampton, Mass.: Smith College, Nat. Sci., Philadelphia, ser. 2, vol. 6, pt. vii+ 441 pp. 1(2): 67-112. 1927. On the relationships of certain plethodont 1867. A review of the species of the Ambysto­ salamanders. Copeia, 165: 102-106. midae. Proc. Acad. Nat. Sci., Philadelphia, 1928. The habitats of plethodontidae. Amer. Nat. 19: 166-211. 680: 236-248. 1 1869. A review of the species of the Plethodon­ 1931. The herpetological fauna of the Americas. tidae and Desmognathidae. Proc. Acad. Copeia, 1931(3): 106-119. Nat. Sci., Philadelphia, 1869: 93-118. 1941. The "opercularis" muscle of salamanders. 1889. The Batrachia of North America. Bull. U.S. J. Morph., 69(2): 207-215. Natl. Mus., 34: 1-525. 1894a. On a collection of Batrachia and Reptilia Edgeworth, F. H. from southwest Missouri. Proc. Acad. Nat. 1935. The cranial muscles of vertebrates. London: Sci., Philadelphia, 1893: 383-385. Cambridge Univ. Press, 493 pp. 1894b. Second addition to the knowledge of the Batrachia and Reptilia of Costa Rica. Proc. Emerson, E. T., II Amer. Philos. Soc., 31(142): 333-347. 1905. General anatomy of Typhlomolge rathbuni. Proc. Boston Soc. Nat. Hist. 32(3): 42-76. DeBeer,O. 1958. Embryos and ancestors, ed. 3. London, Ox­ Estes, R. ford Univ. Press, 197 pp. 1964. Fossil vertebrates from the late Cretaceous 106 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 Lance formation, Eastern Wyoming. Univ. Gray, J.E. California Publ. Geol. Sci., 49: 1-187. 1850. Catalogue of the specimens of Amphibia in 1965. Fossil salamanders and salamander origins. the collection of the British Museum, ed. 1. Amer. Zoologist, 5: 319-334. Part II. Batrachia Gradientia, etc. 72 pp.

Flint, R. F., R. B. Colton, R. P. Goldthwait, and Grohman, A. B. H. B. Willman 1943. The appearance during ontogeny of a suture 1959. Glacial map of the United States east of the between previously fused dermal bones. Rocky Mountains, ed. 1. New York: Geol. Anat. Rec., 87: 211-214. Soc. America. 1959. The anterior cranial elements of the sala­ manders Pseudotritlm and Gyrinophilus. Fowler, H. W. and E. R. Dunn Copeia, 1959(1): 60-63. 1917. Notes on salamanders. Proc. Acad. Nat. Sci., Philadelphia, 1917: 7-28. Hairston, N. G. 1949. The local distribution and ecology of the Gadow,H. plethodontid salamanders of the southern 1905. The distribution of Mexican amphibians and Appalachians. Ecol. Monog., 19: 47-73. reptiles. Proc. Zool. Soc. London, 2: 191- 224. Hallowell, E. 1908. Through Southern Mexico. London: 1856. Description of several species of Urodela, Witherby and Co., xvi 527 pp. + with remarks on the geographical distribu­ Garstang, W. tion of the Caducibranchiate division of 1922. The theory of recapitulation: a critical re­ these animals and their classification. Proc. statement of the biogenetic law. J. Linn. Acad. Nat. Sci., Philadelphia, 8: 6-11. Soc. London, Zoology, 35: 81-101. Hansen, A. M. and W.W. Tanner Gehlbach, F. R. 1958. A comparative osteological study of certain 1959. New plethodontid salamanders of the genus species belonging to the genus Bolitoglossa Thorius from Puebla, Mexico. Copeia, (Amphibia). Great Basin Nat., 18(3-4): 1959(3): 203-206. 85-100.

Goin,C.J. Hecht, M. 1951. Notes on the eggs and early larvae of three 1963. A reevaluation of the early history of the more Florida salamanders. Ann. Carnegie frogs. Part II. Systematic Zool., 12: 20-35. Mus., 32(2): 253-263. Hendrickson, J. R. Goin, C. J. and W. Auffenberg 1954. Ecology and systematics of the salamanders 1958. New salamanders of the family Sirenidae of the genus Batrachoseps. Univ. California from the Cretaceous of North America. Publ. Zool., 54(1): 1-46. Fieldiana-Geology, Chicago, 10(33): 449- 459. Herre, W. 1935. Die Schwanzlurche der mitteleociinen (ober­ Goin, C. J. and 0. B. Goin lutetischen) Braunkohle des Geiseltales 1962. Introduction to herpetology. San Francisco: und die Phylogenie der Urodelen unter Ein­ Freeman, 341 pp. schluss der fossilen Formen. Zoologica, 33 (87): 1-85. Gordon, R. E. 1952. A contribution to the life history and ecol­ 1950. Schwanzlurche aus dem Paleociin von Wal­ ogy of the plethodontid salamander Aneides beck. In Neue Ergebnisse und Probleme der aeneus (Cope and Packard). Amer. Midl. Zoologie (Klatt-Festschrift). Leipzig: Geest Nat., 47(3): 666-701. und Portig, 286-301.

Gorman,J. Herre, W. and H. Lunau 1956. Reproduction in plethodont salamanders of 1950. Neue fossile Schwanzlurche aus dem Burdi­ the genus Hydromantes. Herpetologica, galium. Neuen Jahrb. Geol. Paliion., 8: 247- 12(4): 249-259. 252.

Graham, A. Highton, R. 1964. Origin and evolution of the biota of south­ 1961. A new genus of lungless salamanders from eastern North America: evidence from the the Coastal Plain of Alabama. Copeia, fossil plant record. Evolution, 18: 571-585. 1961(1): 65-68. 1966 EVOLUTION OF LUNGLESS SALAMANDERS 107 1962. Revision of North American salamanders the diphyletic origin of the tetrapod tongue. of the genus Plethodon. Bull. Florida State Kungl. Svenska Vetenskapakad. Hand., 9: Mus., Biol. Sci., 6(3): i35-367. 5-74.

Hilton, W. A. Joubert, P. J. 1945a. The skeleton of Batrachoseps. J. Ent. Zool., 1961. Contributions to the cranial morphology of 37(1): 8-17. Pseudotriton ruber ruber (Sonnini). Ann. 1945b. The skeletons of Aneides, Plethodon, Univ. Stellenbosch, 36(8): 389-418. Eurycea, Manculus quadridigitatus, Hemi­ dactylium, Ensatina, Gyrinophilus, and Kingsbury, B. F. and H. D. Reed Pseudotriton. J. Ent. Zool., 37(2): 43-68. 1909. The columella auris in Amphibia. J. Morph., 20(4) : 549-628. 1945c. The skeletons of Desmognathus, Leurogna­ thus, Typhlotriton, Stereochilus, and Ple­ Larsen, J. H., Jr. thopsis. J. Ent. Zool., 37(3): 75-84. 1963. The cranial osteology of neotenic and trans­ 1945d. The skeletons of Hydromantes, Typhlo­ formed salamanders and its bearing on in­ molge, and Haideotriton. J. Ent. Zool., 37 terfamilial relationships. Ph.D. Thesis, Uni­ (4): 98-102. versity of Washington, Seattle, 205 pp. 1946. Skeletons of Mexican and Central American salamanders of the family Plethodontidae. Laurent,R. J. Ent. Zool., 38(1): 1-8. 1947. La disposition des dents vomeriennes chez 1947a. The hyobranchial skeleton of Plethodon­ les urodeles superieurs et son importance tidae. Herpetologica, 3(6): 191-194. phylogenetique. Bull. Mus. Roy. Hist. Nat. 1947b. Lateral line sense organs in salamanders. Belgique, 23(11): 1-4. Bull. So. Calif. Acad. Sci., 46(3): 97-110. 1948. The vertebrae of salamanders. J. Ent. Zool., Lebednika, N. S. 40(3): 47-65. 1964. Development of the nasal bones in the caudate amphibia. Doklady Akademii Nauk 1950. Dermal sense organs of Stereochilus. Her­ SSSR, 159: 219-222. petologica, 6(5): 135-136. 1959. The tongue in Urodela. Herpetologica, 15 Lowe, C. H., Jr. (3): 137-140. 1950. The systematic status of the salamander Plethodon hardii, with a discussion of bio­ Holman,J.A. geographical problems in Aneides. Copeia, 1958. The Pleistocene herpetofauna of Sabertooth 1950(2): 92-99. Cave, Citrus County, Florida. Copeia, 1958 ( 4) : 276-280. Macginitie, H. D. 1959a. Amphibians and reptiles from the Pleis­ 1958. Climate since the late Cretaceous. In Hubbs, tocene (Illinoian) of Williston, Florida. C. L. (ed.), Zoogeography. Washington: Copeia, 1959(2): 96-102. Amer. Assn. Adv. Sci., publ. 51(2): 61-79. 1959b. A Pleistocene herpetofauna near Orange 1962. The Kilgore flora: a late Miocene flora Lake, Florida. Herpetologica, 15(3): 121- from northern Nebraska. Univ. California 125. Publ. Geol. Sci., 35(2): 67-158.

Humphrey; R.R. Manwell,C. 1922. The multiple testis in urodeles. Biol. Bull., 1963. The blood proteins of cyclostomes. A study 43(1): 45-67. in phylogenetic and ontogenetic biochemis­ try. In Broda!, A. and Fiinge, R. (eds), The Hutchison, V. H. Biology of Myxine. Oslo: Scandinavian 1958. The distn'bution and ecology of the cave University Books, 372-455. salamander, Eurycea lucifuga. Ecol. Monog., 28: 1-20. Martin, P. S. 1958a. A biogeography of reptiles and amphibians Ingram, V. M. in the Gomez Farias region, Tamaulipas, 1961. Gene evolution and the haemoglobins. Na­ Mexico. Misc. Publ. Mus. Zool. Univ. Mich­ ture, 189: 704-708. igan, 101: 1-102. 1958b. Pleistocene ecology and biogeography of Jarvik,E. North America. In Hubbs, C. L. (ed.), Zoo­ 1963. The composition of the intermandi'bular di­ geography. Washington: Amer. Assn. Adv. vision of the head in fish and tetrapods and Sci., publ. 51 (15): 375-420. 108 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 Martin, P. S. and B. E. Harrell 1927b. The value of life history data in the study 1957. The Pleistocene history of temperate biotas of the evolution of the Amphibia. Ann. in Mexico and eastern United States. Ecol­ New, York Acad. Sci;, 30: 31-128. ogy, 38(3): 468-480. 1931. The biology of the Amphibia. New York: McGraw-Hill, 577 pp. Martof, B. S. 1962. Some aspects of the life history and ecology Organ, I.A. of the salamander Leurognathus. Amer. 196la. Studies of the local distribution, life his­ Midl. Nat., 67(1): 1-35. tory, and population dynamics of the sala­ mander genus Desmognathus in Virginia. Martof, B. S. and F. L. Rose Ecol. Monog., 31: 189-220. 1962. The comparative osteology of the anterior 1961b. Life history of the pigmy salamander, Des­ cranial elements of the salamanders Gyrino­ mognathus wrighti, in Virginia. Amer. Midl. philus and Pseudotriton. Copeia, 1962( 4): Nat., 66(2): 384-390. 727-732. Parker, W. K. Mayr,E. 1877. On the structure and development of the 1962. Animal species and evolution. Cambridge, skull in the urodelous Amphibia. Part I. Mass.: Harvard University Press, 797 pp. Philos. Trans. Roy. Soc. London, 167(2): 529-597. Mitchell, R. W. and J. R. Reddell 1965. Eurycea tridentifera, a new species of trog­ 1882. On the morphology of the skull in the Am­ lobitic salamander from Texas and a re­ phibia Urodela. Trans. Linn. Soc. London, classification of Typhlomolge rathbuni. Zoology, ser. 2, 2(3): 165-212. Texas J. Sci., 17: 12-27. Parrington, F. R. Monath, T. 1956. The patterns of dermal bones in primitive 1965. The opercular apparatus of salamanders. J. vertebrates. Proc. Zool. Soc. London, 127 Morph. 116: 149-170. (3): 389-411. Parsons, T. S. and E. E. Williams Moore, J.P. 1962. The teeth of Amphibia and their relation to 1899. Leurognathus marmorata, a new genus and amphibian phylogeny. J. Morph., 110: 375- species of salamander of the family Des­ 390. mognathidae. Proc. Acad. Nat. Sci., Phila­ delphia, 1899: 316-323. Peabody, F. S. . 1900. Post-larval changes in the vertebral articula­ 1959. Trackways of living and fossil salamanders. tions of Spelerpes and other salamanders. Univ. California Publ. Zool., 63 (1) : 1-72. Proc. Acad. Nat. Sci., Philadelphia, 1900: 613-622. Peabody, F. S. and J. M. Savage 1958. Evolution of a Coast Range corridor in Murray, K. F. California and its effect on the origin and 1957. Pleistocene climate and the fauna of Burnet dispersal of living amphibians and reptiles. Cave, New Mexico. Ecology, 38(1): 129- In Hubbs, C. L. (ed.), Zoo geography. Wash­ 131. ington: Amer. Assn. Adv. Sci., publ. 51 (8): 159-186. Niceforo Marla 1958. Secci6n Herpetol6gia. Bol. Inst. La Salle, Peters, W. Bogota, 45(198): 1-16. 1863. Femere Mittheilungen iiber neue Batrachier. Monats. Konig. Preuss Akad. Wiss., Berlin, 1863: 445-470. Noble,G.K. 1921. The anterior cranial elements of Oedipus Piatt, J. and certain other salamanders. Bull. Amer. 1935. A comparative study of the hyobranchial Mus. Nat. Hist., 44(1): 1-6. apparatus and throat musculature in the 1925. An outline of the relation of ontogeny to Plethodontidae. J. Morph., 57(1): 213-251. phylogeny within the Amphibia II. Amer. Mus. Novit., 166: 1-10. Pope, C. H. and S. H. Pope 1927a. The plethodontid salamanders; some as­ 1951. A study of the salamander Plethodon oua­ pects of their evolution. Amer. Mus. Novit., chitae and the description of an allied form. 249: 1-26. Bull. Chicago Acad. Sci., 9: 129-152. 1966 EVOLUTION OF LUNGLESS SALAMANDERS 109 Rabb,G.B. Schwartz, A. and R. E. Etheridge 1955. A new salamander of the genus Parvimolge 1954. New and additional records from the North from Mexico. Breviora, Mus. Comp. Zool., Carolina Coastal Plain. Herpetologica, 10 Harvard Coll., 42: 1-9. (3): 167-171. 1956. A new plethodontid salamander from Nuevo Uon, Mexico. Fieldiana-Zoology, Shotwell, J. A. Chicago, 39(3): 11-20. 1961. Late Tertiary biogeography of horses in the northern Great Basin. J. Paleon. 35(1): 1958. On certain salamanders of the plethodontid 203-217. genus Chiropterotriton. Occ. Pap. Mus. Zool. Univ. Michigan, 587: 1-37. 1963. Pliocene mammalian communities of the Juntura Basin. In Shotwell, J. A., R. G. 1960. A new salamander of the genus Chiroptero­ Bowen, W. L. Gray, D. C. Gregory, D. E. triton from Chiapas, Mexico, with notes on Russell, D. Taylor, The Juntura Basin: related species. Copeia, 1960( 4): 304-311. studies in earth history and paleoecology. Trans. Amer. Philos. Soc., new ser., 53 ( 1) : Reed,H.D. 7-21. 1920. The morphology of the sound-transmitting apparatus in caudate Amphibia and its phy­ Simpson, G. G. logenetic significance. J. Morph., 33(2): 1947. Evolution, interchange, and resemblance of 325-387. the North American and Eurasian Cenozoic mammalian faunas. Evolution, 1 (3): 218- Rensch, B. 220. 1954. Die phylogenetische Abwandlung der On­ 1953. The major features of evolution. New York: togenese. In Heberer, G. (ed.), Die Evolu­ Columbia Univ. Press, xx + 434 pp. tion der Organismen. Stuttgart: Fischer, 103-130. Smith, H. M. and E. H. Taylor 1948. An annotated checklist and key to the am­ phibia of Mexico. Bull. U.S. Natl. Mus., Rose, F. L. and F. M. Bush 194: 1-118. 1963. A new species of Eurycea (Amphibia: Cau­ data) from the southeastern United States. Smith, L. Tulane Stud. Zoo!., 10(2): 121-128. 1920. The hyobranchial apparatus of Spelerpes bislineatus. J. Morph., 33: 527-583. Sarasin, P. and F. Sarasin 1887- Smith, P. W. 1890. Zur Entwicklungsgeschichte und Anatomie 1957. An analysis of post-Wisconsin biogeography der ceylonesischen Blindwiihle lchthyophis of the prairie peninsula region based on dis­ glutinosus. In Ergebnisse Naturwissenschaft­ tributional phenomena among terrestrial licher Forschungen auf Ceylon. Wiesbaden: vertebrate populations. Ecology, 38(2): Kreidels, 2: 153-263. 205-218. 1961. The amphibians and reptiles of Illinois. Bull. Savage, J.M. Illinois Nat. Hist. Surv., 28(1): 1-298. 1960. Evolution of a peninsular herpetofauna. Systematic Zoo!., 9(3): 184-212. Soler, E. I. 1950. On the status of the family Desmognathidae. Univ. Kansas Sci. Bull., vol. 33, pt. 2(12): Schmalhauaen, I. I. 459-480. 1949. Factors of Evolution (American transla­ tion). Philadelphia and Toronto: Blakiston, Stebbins, R. C. xiv+ 327 pp. 1951. Amphibians of western North America. Berkeley and Los Angeles: Univ. California Sclunidt, K. P. · ------Press. 539 pp. 1936. Guatemalan salamanders of the genus 1954. Natural history of the salamanders of the Oedipus. Zool. Ser. Field Mus. Nat. Hist., plethodontid genus Ensatina. Univ. Califor­ · c;:laicago, 20(17): 135-166. nia Publ. Zool., 54(2): 47-124. 1946.oOll the zo0geography of the Holarctic re­ gion. Copeia, 1946(3): 144-152. Stebbins, R. C. and C. H. Lowe, Jr. :'*'JA claed: list of North American amphibians 1949. The systematic status of Plethopsis with a :.. nMid niptilea, ed. 6. Chicago: Amer. Soc. discussion of speciation in the genus Batra­ lcbthy. Herpt., 280 pp. choseps. Copeia, 1949(2): 116-129. 110 MEMOIR OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES VOL.4 Stejneger, L. H. 1944. The genera of plethodont salamanders in 1892. Preliminary description of a new genus and Mexico, pt. I. Univ. Kansas Sci. Bull., 30, species of blind cave salamander from pt. 1(12): 189-232. North America. Proc. U.S. Natl. Mus., 15: 1949. New salamanders from Costa Rica. Univ. 115-117. Kansas Sci. Bull., 33: 279-288. 1907. Herpetology of Japan and adjacent terri­ tory. Bull. U.S. Natl. Mus., 58: 1-577. Teege, M-J. Laurinat 1956. Studien zur Entwicklung und Gestalt der Stejneger, L. H. and T. Barbour Urodelenwirbel. Zeit. wissen. Zoologie, 160: 1917. A check list of North American amphibians 95-163. and reptiles, ed. 1. Cambridge: Harvard Univ. Press, 125 pp. Thornbury, W. D. 1954. Principles of Geomorphology. New York Storer, T. I. and London: Wiley, 618 pp. 1925. A synopsis of the Amphibia of California. Univ. California Publ. Zool., 27: 1-342. Uzzell, T. M., Jr. 1961. Calcified hyoid and mesopodial elements of Stuart, L. C. plethodontid salamanders. Copeia, 1961(1): 1935. A contribution to a knowledge of the her­ 78-86. petology of a portion of the savanna region of central Peten, Guatemala. Misc. Publ. Valentine, B. D. Mus. Zool. Univ. Michigan, 29: 1-56. 1963a. Notes on the early life history of the Ala­ 1943. Taxonomic and geographic comments on bama salamander, Desmognathus aeneus Guatemalan salamanders of the genus Oedi­ chermocki Bishop and Valentine. Amer. pus. Misc. Publ. Mus. Zool. Univ. Michigan, Midi. Nat., 69(1): 182-188. 56; 1-33. 1963b. The plethodontid salamander Phaeogna­ 1948. The amphibians and reptiles of Alta Vera­ thus; external morphology and zoogeogra­ paz, Guatemala. Misc. Publ. Mus. Zool. phy. Proc. Biol. Soc. Washington, 76: 153- Univ. Michigan, 69: 1-109. 158. 1950. A geographic study of the herpetofauna of 1963c. The plethodontid salamander Phaeogna­ Alta Verapaz, Guatemala. Contrib. Lab. thus: collecting techniques and habits. J. Vert. Biol. Univ. Michigan, 45: 1-77. Ohio Herpt. Soc., 4(1-2): 49-54. 1951. The herpetofauna of the Guatemalan Pla­ teau, with special reference to its distribu­ Waddington, C. H. tion on the southwestern highlands. Contrib. 1956. Genetic assimilation of the bithorax pheno­ Lab. Vert. Biol. Univ. Michigan, 49: 1-71. type. Evolution, 10: 1-13. 1954. Herpetofauna of the southeastern highlands of Guatemala. Contrib. Lab. Vert. Biol. Wahlert, G. von Univ. Michigan, 68: 1-65. 1957. Biogeographische und Oekologische Tatsa­ 1957. Herpetofaunal dispersal routes through chen zur phylogenie amerikanischer northern Central America. Copeia, 1957 Schwanzlurche. Zoo!. Jb. Syst., 85 (3): 253- (2): 89-94. 282. 1958. A study of the herpetofauna of the Uaxac­ tun-Tikal area of northern El Peten, Guate­ Wake,D.B. mala. Contrib. Lab. Vert. Biol. Univ. Michi­ 1960. The osteology and relationships of the sala­ gan, 75: 1-30. mander genus Aneides Baird. M. Sc. Thesis, University of Southern California, Los An­ Tanner, W.W. geles, 181 pp. 1950. A new genus of plethodontid salamander 1963. Comparative osteology of the plethodontid from Mexico. Great Basin Nat., 10: 37-44. salamander genus Aneides. J. Morph., 113 1952. A comparative study of the throat muscula­ (1): 77-118. ture of the Plethodontidae of Mexico and Central America. Univ. Kansas Sci. Bull., Wake, D. B. and A. H. Brame, Jr. 34, pt. 2(10): 583-677. 1963a. The status of the plethodontid salamander genera Bolitoglossa and Magnadigita. Taylor, E. H. Copeia, 1963(2): 382-387. 1940. New salamanders from Mexico, with a dis­ 1963b. A new species of Costa Rican salamander, cussion of certain known forms. Univ. genus Bolitoglossa. Rev. B~ol. Trop., 11 ( 1) : Kansas Sci. Bull., 26(12): 407-439. 63-73. 1966 EVOLUTION OF LUNGLESS SALAMANDERS 111 1966. Notes on South American salamanders, Wilder, I. L. Whipple genus Bolitoglossa. Copeia, 1966(2): 1913. The life history of Desmognathus fusca. 360-363. Biol. Bull., 24(5): 293-343. 1920. The urodele vomer. Anat. Rec., 18: 349. Wiedersheim, R. 1924. Variations in the premaxillary of Eurycea 1875. Salamandrina perspicillata und Geotriton bislineata. Amer. Nat., 58: 538-543. fuscus. Versuch einer vergleichenden Anat­ omie du Salamandriden. Genua: Instituts 1925. The morphology of amphibian metamor­ der Sordo-Muti, 207 pp. phosis. Northampton, Mass.: Smith College, xviii + 161 pp. 1877. Das Kopfskelet der Urodelen. Gegen. Morph. Jb., 3: 352-548. Wilder, I. L. Whipple and E. R. Dunn Wilder, H. H. 1920. The correlation of Iunglessness in salaman­ 1894. Lungenlosen Salamandriden. Anat. Anz., ders with a mountain brook habitat. Copeia, 9: 216-220. 84: 63-68. CONTENTS .ABSTRACT 1

INTRODUCTION 2 Acknowledgments 4 Materials and Methods 4 Non-osteological Family Characters 5

ANALYSIS OF VARIATION 7 Osteological Characters 7 Premaxilla 7 Maxilla 13 Septomaxilla 14 Nasal 15 Prefrontal 15 Route of the Nasolacrimal Duct 18 Vomer 20 Frontal 22 Parietal 23 Orbitosphenoid 25 Parasphenoid 25 Occipito-Otic 25 Occipital Condyles 26 Suspensorium 26 Palatopterygoid 27 Squamosal 27 Opercular Plate 27 Mandible 28 Hyobranchial Skeleton 28 Larval structure 28 Adult structure 28 Ceratohyal 29 Comua 29 Lingual cartilage 30 First basibranchial 32 Proportions of articulated elements 32 Second basibranchial 34 Calcification 34 Synthesis of hyobranchial characters 35 Dentition 35 Vertebral Column 37 Regional differentiation 37 Cervical vertebra 39 Trunk vertebrae 39 Centrum structure 40 Neural arch structure 41 Transverse processes· 41 Caudal vertebrae 43 Ribs 43 Limb Elements 44 Mesopodial Elements 44 Digits 46

SYSTEMATIC SUMMARY 47 Family Plethodontidae 47 Subfamily Desmognathinae 47 Desmognathus 48 Leurognathus 48 Phaeognathus 48 Subfamily Plethodontinae 49 Tribe Hemidactyliini 50 Gyrinophilus 50 Pseudotriton 50 Stereochilus 51 Eurycea 51 Typhlotriton 51 Typhlomolge 51 Haideotriton 51 H emidactylium 51 Tribe Plethodontini 51 Plethodon 52 Ensatina 52 Aneides 52 Tribe Bolitoglossini 52 Supergenus Hydromantes 53 Hydromantes 53 "'"· Supergenus Batrachoseps 53 Batrachoseps 53 Supergenus Bolitoglossa 53 Bolitoglossa 53 Oedipina 54 Pseudoeurycea 54 Chiropterotriton 54 Lineatriton 54 Thorius 54 Parvimolge 55

EVOLUTIONARY RELATIONSHIPS AND TRENDS 55 Relationships 55 Major Familial Divisions 55 Relationships of the Desmognathine Genera 58 Relationships of the Plethodontine Genera 60 Tribal relationships 60 Relationships of the hemidactyliines 62 Generic groupings 62 Status of the genus Manculus 64 Status of the genera Haideotriton and Typhlomolge 64 Relationships of the plethodonines 67 Relationships of the bolitoglossines 67 Trends 73 Ecology and Life History 73 The ancestral adaptive zone 73 Plethodontid origins 73 Evolution within the adaptive zone 74 Occupation of new adaptive zones and subzones 76 Paedomorphosis 79 Paedormorphic influence and plethodontid evolution 79 The evolutionary significance of paedomorphosis 81

BIOGEOGRAPHY AND PHYLOGENY 84 The Fossil Record 84 Historical Backgrounds and the Primary Dichotomy 85 The Desmognathines 86 The Plethodontines 87 Tribal Origins 87 History of the Hemidactyliines 89 The First Terrestrial Encounter: The History of the Bolitoglossines 93 The Second Terrestrial Encounter: The History of the Plethodonines 99

CONCLUSIONS AND SUMMARY 102

LITERATURE CITED 104 Eurycea longicauda, an adult female from Deputy Cave, Jennings County, Indiana. MEMOIRS OF THE SOUTHERN CALIFORNIA ACADEMY OF SCIENCES

The MEMOIRS of the Southern California Academy of Sciences is a series begun in 1938 and published on an irregular basis thereafter. It is intended that each article will continue to be of a monographic nature, and each will constitute a full volume in itself. Contributions to the MEMOIRS may be in any of the fields of science, by any member of the Academy. Acceptance of papers will be determined by the amount and character of new information and the form in which it is presented. Articles must not duplicate, in any substantial way, material that is published elsewhere. Manuscripts must conform to MEMOIRS style of volume 4 or later, and may be examined for scientific content by mem­ bers of the Publications Committee other than the Editor or be sent to other competent critics for review. Because of the irregular nature of the series, authors are urged to contact the Editor be­ fore submitting their manuscript.

DAVID K. CALDWELL Editor

Communications concerning the purchase of the Memoirs should be directed to the Treasurer, and communications concerning manuscripts should be directed to the Editor, both in care of the Los Angeles County Museum of Natural History, Exposition Park, Los Angeles, California 90007. Printed by Anderson, Ritchie & Simon, Los Angeles, California