Generic Boundaries in Texas Cave Salamanders, and a Redescription of Typhlomolge Robusta (Amphibia: Plethodontidae) Author(S): Floyd E

Home , Eurycea neotenes

Generic Boundaries in Texas Cave Salamanders, and a Redescription of Typhlomolge robusta (Amphibia: Plethodontidae) Author(s): Floyd E. Potter, Jr. and Samuel S. Sweet Reviewed work(s): Source: Copeia, Vol. 1981, No. 1 (Feb. 10, 1981), pp. 64-75 Published by: American Society of Ichthyologists and Herpetologists (ASIH) Stable URL: http://www.jstor.org/stable/1444041 . Accessed: 10/01/2012 14:25

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

American Society of Ichthyologists and Herpetologists (ASIH) is collaborating with JSTOR to digitize, preserve and extend access to Copeia.

http://www.jstor.org 64 COPEIA, 1981, NO. 1

RIVAS,L. R. 1964. A reinterpretationof the concepts Wauchope and R. C. West (eds.). Univ. Texas "sympatric"and "allopatric"with proposal of the Press, Austin. additional terms "syntopic"and "allotopic."Syst. TURNER,B. J., AND R. K. LIU. 1977. Extensiveinter- Zool. 13:42-43. specific genetic compatabilityin the New World SOLTZ, D. L., AND R. J. NAIMAN. 1978. The natural killifishgenus Cyprinodon.Copeia 1977:259-269. history of native fishes in the Death Valley system. Nat. Hist. Mus. L.A. Co. Sci. Ser. 30. MUSEUM OF ZOOLOGYAND DIVISION OF BIOLOG- TAMAYO,J. L., ANDR. C. WEST. 1964. The hydrog- ICAL SCIENCES, THE UNIVERSITY OF MICHI- 84-121. In: Hand- raphy of Middle America, p. GAN, ANN ARBOR, MICHIGAN 48109. Accept- book of Middle American Indians, Vol. 1. R. ed 18 Dec. 1979.

Copeia, 1981(1), pp. 64-75

Generic Boundaries in Texas Cave Salamanders, and a Redescription of Typhlomolgerobusta (Amphibia: Plethodontidae)

FLOYD E. POTTER, JR., AND SAMUEL S. SWEET

Confusion presently surrounds the systematics of the salamanders of the genera Eurycea and Typhlomolgeof the Edwards Plateau of central Texas. Existing morphological evidence is reviewed and found to be inadequate to resolve the status of the genus Typhlomolge.New data on skull morphology supports the continued recognition of this genus. The advanced troglobite Typhlomolgerobusta Longley, 1978, is redescribed and shown to share the features which distinguish T. rathbunifrom the Texas Eurycea.Typhlomolge rathbuni and T. robustamay be parapatricin the San Marcos Pool of the Balcones Aquifer.

of parallelism and convergence characters are known or suspected to be sus- PROBLEMSare acute in the systematics of troglobitic ceptible of convergence. animals, resulting from a common tendency for Recent systematic treatments of the hemidac- cave colonization to occur repeatedly through- tyliine plethodontid salamanders of the Ed- out the range of ancestral populations, and wards Plateau of central Texas exemplify these from an apparent similarity in selective regimes difficulties of interpretation. The caves of the which characterizes the cave environment. The Edwards Plateau contain a diverse assemblage narrow circumscription of successful adaptive of troglobitic salamanders, comprising numer- modes in the evolution of troglobites is evi- ous independently-derived populations which denced by the existence of striking conver- display varying degrees of specialization in pig- gences in morphology, physiology and life-his- mentation, eye size and structure, relative limb tory features among diverse taxa of troglobitic length and head size, head shape and numbers animals (Poulson, 1964; Barr, 1968; Poulson of trunk vertebrae and teeth (Mitchell and Red- and White, 1969). For these reasons most meth- dell, 1965; Brandon, 1971; Sweet, 1978b). At ods of phylogenetic reconstruction are difficult present, each of three alternative taxonomic to apply in the resolution of systematic ques- arrangements of this group of six species has tions involving troglobitic taxa; even a large set a degree of acceptance in the literature. Conant of apparently synapomorphic character states (1975), Sweet (1977a, b, 1978a) Thomas (1976) may fail to demonstrate convincingly mono- and Collins et al. (1978) retain the generic des- phyly when a significant percentage of the ignations of the original authors (Eurycea nana

? 1981 by the American Society of Ichthyologists and Herpetologists POTTER AND SWEET-TYPHLOMOLGE ROBUSTA DESCRIPTION 65

Bishop, 1941, E. neotenes Bishop and Wright, conclusions through an examination of a new 1937, E. latitans Smith and Potter, 1946, E. trog- troglobitic species which seems to be superfi- lodytes Baker, 1957, E. tridentifera Mitchell and cially intermediate between Eurycea tridentifera Reddell, 1965, and Typhlomolgerathbuni Stejne- and Typhlomolgerathbuni. The major method- ger, 1896, in order of increasing troglobitic ad- ological assumption in this investigation is that aptation). Following Mitchell and Reddell evidence of convergent adaptation to the cave (1965) and Mitchell and Smith (1972), Typhlo- environment among the Edwards Plateau molge has been widely considered to be a junior hemidactyliines would require the rejection of synonym of Eurycea: Reddell (1967, 1971), Bo- an hypothesis of monophyly in the group. gart (1967), Blair (1968), Raun (1971), Hen- dricks (1973), Thomas (1974), Tupa and Davis MATERIAL EXAMINED (1976), Russell (1976) and Longley (1978). Wake (1966) redefined and expanded the ge- In the course of this work 95 specimens of nus Typhlomolgeto include both E. tridentifera T. rathbuni ranging from 7.4 to 73.2 mm in and T. rathbuni; this arrangement has been ac- standard length (SL, tip of snout to posterior cepted by Brame (1967), Brandon (1968, 1971), margin of cloaca), and 144 specimens of E. tri- Leviton (1971), Raun and Gehlbach (1972), dentifera ranging from 14.0 to 46.0 mm SL have Dundee (1973), Gorham (1974), Besharse and been examined externally and by high resolu- Holsinger (1977) and Smith (1978). In addition radiographs. Five T. rathbuni (28.5-54.0 tion, Longley (1978) has employed the desig- mm SL) and 32 E. tridentifera (14.0-45.0 mm nations Eurycea (=Typhlomolge) rathbuni and SL) were available as cleared and stained prep- Typhlomolge(=Eurycea) rathbuni (sic), apparent- arations. In addition, 42 (7 cleared and stained) ly without intending to establish subgeneric sta- specimens of E. troglodytes(22.5-43.6 mm SL), tus for Typhlomolge. 21 (1 cleared and stained) specimens of E. la- At issue are the status and content of the ge- titans (9.5-47.8 mm SL), and 153 (17 cleared nus Typhlomolge,and the zoogeographic inter- and stained) specimens of E. neotenes(15.0-53.0 pretations of Wake (1966) and Mitchell and mm SL) representing 13 troglobitic populations Smith (1972). Wake suggested that Typhlomolge have been available, together with approxi- is a relict of a Miocene radiation of salamanders mately 3500 specimens representing about 115 ancestral to the genus Eurycea, a radiation also epigean populations of E. neotenesand two pop- represented by the troglobites Typhlotritonspe- ulations of E. nana. The repositories of this laeus and Haideotriton wallacei; in Wake's view, material are listed by Sweet (1978b). Cleared the Texas Eurycea derive from a subsequent and stained representative series of all genera (probably Plio-Pleistocene) invasion of the Ed- and the majority of species of the hemidacty- wards Plateau. Citing geologic evidence on the liine plethodontids have been available as ref- age of the plateau (discussed below), Mitchell erence material. and Smith (1972) interpreted the history of the The following abbreviations are employed in Texas hemidactyliines to date from Pleistocene the text: FEP, personal collection of the senior time. author; MVZ, Museum of Vertebrate Zoology, The question at the root of this disagreement University of California, Berkeley; TNHC, seems to be the following: is there a morpho- Texas Natural History Collection, University of logical dichotomy among the Edwards Plateau Texas, Austin; TTU, uncatalogued material hemidactyliines which is of sufficient magni- loaned by R. W. Mitchell and J. R. Reddell, tude to indicate their derivation from disparate Texas Tech University. ancestral stocks? In other words, does this as- consist of in semblage solely lineages evolving RESULTS AND DISCUSSION parallel (and thus properly congeneric under an operating assumption of monophyly), or is Status of Typhlomolge.-Had representatives of there some combination of parallel and con- the range in troglobitic specialization now vergent lineages? The present work has two known in the Texas Eurycea been available to objectives: to reevaluate existing evidence on Stejneger it is doubtful that the genus Typhlo- the status of the genus Typhlomolgein terms of molge would have been established. The histor- the likelihood of convergent origin for shared ical accident of the discovery of T. rathbuni long character states; and to test our phylogenetic before other paedogenetic troglobites were 66 COPEIA, 1981, NO. 1 known from the Edwards Plateau resulted in in E. tridentifera, but are nearly always present action, which is now viewed from a very differ- in large specimens, though they are uniformly ent perspective. The discovery of increasingly absent in T. rathbuni; this feature is presumably specialized troglobitic species of the Texas Eu- related to the reduced need for lateral bony rycea has delineated an adaptive trend in which support in the flattened skulls of these species. T. rathbuni is considered the end member, and The enlarged otic capsules noted by Wake in- has understandably led to a review of the re- dicate a general adaptive trend among troglo- lationship between the two genera. Baker bitic vertebrates (Poulson [1963] on amblyopsid (1957:336) anticipated this review in noting fishes). Thus, none of these features provides that Eurycea troglodytes"demonstrates a way that evidence for or against the recognition Typhlomolgecould have evolved from a Eurycea- ofTyphlomolge. type ancestor." The distinctiveness of T. rath- The literature contains indications of several buni was further eroded by the discovery of the other characters of potential value in resolving advanced troglobite E. tridentifera, leading the relationship between Eurycea and Typhlo- Mitchell and Reddell (1965:23) to reject the ge- molge. Their utility is a function of the likelihood nus for the reason that "the differences be- of convergent acquisition of similar or identical tween the neotenic Eurycea and T. rathbuni are states, a matter which is sometimes difficult to almost wholly ones of degree only." The evi- evaluate. These features are considered in turn dence for this conclusion is the existence of a below, and include: fusions of tarsal cartilages; concordant morphocline in external appear- the positional relationship of the jugularis ra- ance encompassing all of the Edwards Plateau mus of the facial nerve to the columellar stylus; hemidactyliines; however, this morphocline is and the overall pattern of modification of the composed of characters which can easily be anterior cranial elements. demonstrated to exhibit convergence on a The fourth and fifth tarsal cartilages of T. broad scale. Irrespective of the actual phylog- rathbuni were reported by Emerson (1905) to be eny, such evidence cannot constitute a suffi- fused into a single element. Wake (1966) noted cient argument for a monophyletic relationship a similar fusion in E. tridentifera, but found the among the members of this assemblage. tarsal elements to be discrete in his T. rathbuni. In supporting and redefining the genus Ty- These cartilages are also discrete in six tarsi ex- phlomolge, Wake (1966) noted that his speci- amined here; the unmodified condition thus mens of E. tridentifera (3) and T. rathbuni (1) occurs in eight of ten cases in T. rathbuni. shared several characters which distinguished Among 14 E. tridentifera representing four them from the remaining species of the Texas populations, tarsals four and five are fused in Eurycea. In addition to elongate limbs and a 21 of 28 cases (in three instances the fusion reduced number of trunk vertebrae, Wake cit- includes tarsal three as well). Tarsal fusion oc- ed enlarged otic capsules, absence of ossified curs in three of 14 instances in E. troglodytes,in orbitosphenoids, diapophyses which extend be- one of 20 tarsi of ten specimens of E. neotenes yond the lateral margins of the zygapophyses representing four troglobitic populations, and and presence of alar processes on the parapo- in six of 100 tarsi of E. neotenes representing physes as synapomorphies in E. tridentiferaand five epigean populations. Increased frequency T. rathbuni. Access to a larger sample permitted of tarsal fusion seems to correlate with increas- Mitchell and Smith (1972) to demonstrate that ing troglobitic specialization in species of Texas ossified orbitosphenoids are present in some E. Eurycea, but the derived condition is much less tridentifera, and that the vertebral characters frequent in T. rathbuni than in E. tridentifera. employed by Wake also occur in some individ- Kingsbury and Reed (1909) reported that in uals of all species of the Texas Eurycea; their T. rathbuni alone among plethodontid salaman- findings are supported by the material exam- ders the jugularis ramus of the facial nerve ined here (which includes the specimens avail- passes dorsal to the columellar stylus. The cra- able to Wake and to Mitchell and Smith). The nial nerves can be seen in some cleared and variation noted in these characters is primarily stained specimens, especially those which have ontogenetic in nature. In general, elongate dia- been preserved in isopropyl alcohol. The jugu- pophyses and well-developed alar processes are laris lies dorsal to the columella in each of three to be found in at least some larger individuals specimens of T. rathbuni examined (Fig. Ic); of all of the Edwards Plateau hemidactyliines. however, it is also dorsal in position in some The orbitosphenoids are delayed in appearance individuals of species of the Texas Eurycea, and POTTER AND SWEET-TYPHLOMOLGE ROBUSTA DESCRIPTION 67

P* -7 0. C. sq. Os. C. S. ppt. r.j. eb. 1

m" ch. 1. b

5 mm f

a C

Fig. 1. Left lateralviews of the posteriorskull, hyobranchialelements and trigeminaland facial nerves of: a) Euryceaneotenes (MVZ 119938, 34.8 mm SL); b) E. tridentifera(TTU uncat., 37.0 mm SL); and c) Typhlomolge rathbuni(TTU uncat., 28.5 mm SL). The jugularis ramus of the facial nerve is characteristicallyventral to the columellarstylus in E. neotenes,but follows a dorsal route in most E. tridentiferaand all T. rathbuniexamined. Unique features of T. rathbuniillustrated here include a parietal spur, a cartilaginous bridge above the trigeminal nerve, and the origin of the ceratohyalligament from the squamosal. Key to abbreviations:ch., ceratohyal;ch. 1., ceratohyalligament; c.s., columellar stylus; eb. 1, epibranchial 1; m., mandible; o.c., otic capsule; os., orbitosphenoid;p., parietal;ppt., palatopterygoid;q., quadrate;r.j., ramusjugularis; sq., squamosal.

appears to be a rough correlate of the degree Kingsbury and Reed (1909), and in Typhlotriton of troglobitic adaptation. Examination of 17 and Haideotriton as well. specimens representing four populations of E. The positional relationships of nerves and tridentiferashows the jugularis ramus in a dorsal skeletal elements are generally considered to be position in 27 of 34 cases (Fig. ib). It is unilat- evolutionarily conservative, and are commonly erally ventral in three specimens and bilaterally accorded substantial weight in phylogenetic ventral in two. Among other species, the jugu- studies. For example, the same character dis- laris is bilaterally dorsal to the columella in one cussed above is known to occur elsewhere E. troglodytes and bilaterally ventral in four among salamanders only in Proteus anguinus specimens. In 11 individuals from three trog- and Necturus maculosus (Kingsbury and Reed, lobitic populations of E. neotenes the ramus is 1909), and has figured strongly in recent dis- dorsal in 11 cases, ventral in eight, and perfo- cussions of the relationship of these genera rates the columellar stylus in three. The ramus (Larson and Guthrie, 1974; Hecht and Ed- is unilaterally dorsal in one of four epigean wards, 1976). It is clear that this condition of specimens of E. neotenes, but otherwise follows dorsal passage of the jugularis ramus has orig- a ventral route (Fig. la), as is also the case in a inated independently at least twice among the single specimen of E. nana. The jugularis ra- four genera, which are all characterized by mus is uniformly ventral in position in our ma- elongate, anteriorly-rotated suspensoria. The terial of the hemidactyliine genera examined by columellar stylus in salamanders is attached to 68 COPEIA, 1981, NO. 1 the suspensorium either directly or by means rathbuni is a subadult (28.5 mm SL). Fig. 2 il- of a short ligament (Kingsbury and Reed, 1909; lustrates dorsal and lateral aspects of the skulls Francis, 1934; Fox, 1954; Monath, 1965; Wake, of E. neotenes, E. tridentiferaand T. rathbuni. It 1966); with elongation of the suspensorium this is evident that E. tridentiferadiffers from E. neo- point of attachment is shifted anteroventrally, tenes principally in the enlargement and antero- placing the columellar stylus in a more ventral lateral extension of the suspensorium; no pro- position with respect to the braincase. The col- portional differences are evident in the anterior umellar anlage derives from neural crest tissue, cranial region or in the braincase. In contrast, and is present as a dense mass prior to or con- the skull of T. rathbuni shows extensive broad- temporaneous with the formation of the com- ening and appreciable elongation of the pre- ponents of the facial nerve (Landacre, 1921; maxilla, frontals, vomers and palatopterygoids, Horstadius and Sellman, 1946). Hence the de- while the suspensorial region more closely re- veloping jugularis ramus may simply follow the sembles that of E. neotenes than it does that of path of least resistance around the columellar E. tridentifera. stylus or its ligamentous connection to the sus- The proportional differences among these pensorium; rather than representing a feature skulls are most easily understood via the graph- of value in phylogenetic reconstruction the dor- ic method of deformed coordinates (D'Arcy sal path of the nerve seems to be a correlate of Thompson, 1917). This technique has the ad- a type of modification in jaw mechanics com- vantage of demonstrating complex and often mon among paedogenetic salamanders. This interrelated alterations in form in a single sim- feature can no more be considered a synapo- ple diagram. The method is suitable as a first morphy in Eurycea and Typhlomolgethan it can step in selecting variables for quantitative anal- be so interpreted between these genera and yses, or alone in general comparisons of form; Necturus and Proteus, or between the latter pair. a quantitative method for constructing and At this point a brief summarization may be analyzing deformed grids has been recently useful as a point of departure. None of the fea- proposed by Bookstein (1978). In practice, a tures which has been used to argue either for reference specimen (the skull of E. neotenes in the inclusion of Typhlomolgein Eurycea or for this case) is drawn in a grid, and the coordinates the inclusion of E. tridentifera in Typhlomolge of the intersecting lines are transferred to comprises valid evidence for these views. Most drawings of the specimens to be compared to or all of these characters are directly related to the reference object. Connecting lines produce a common pattern of adaptation to subterra- the deformation required of the reference ob- nean life, and seem to be easily acquired in a ject to cause it to resemble each of the others, convergent fashion by a diverse array of trog- providing a visualization of the nature and de- lobitic salamanders and fishes. While there is gree of the proportional alterations involved. no direct supporting evidence, the weight of Numerous examples of the application of this inference from these data suggests that the in- technique are presented by D'Arcy Thompson clusion of T. rathbuni in Eurycea represents a (1917), Huxley (1932) and Bookstein (1978). more reasonable solution than does the recog- The deformation grid for E. tridentifera(Fig. nition of a genus composed of T. rathbuni and 2c) illustrates both an hypertrophy of the sus- E. tridentifera. A final point of evidence has pensorial region and a lack of modification in bearing on the validity of the remaining taxo- the anterior skull elements and braincase in nomic option, that of the continued recognition comparison to E. neotenes. The skulls of trog- of a monotypic genus Typhlomolge(as modified lobitic E. neotenes, and E. latitans and E. troglo- below). dytes provide a series of intermediate stages in Wake (1966) commented that the anterior this proportional trend (Mitchell and Smith, portion of the skull of T. rathbuni is extremely 1972:Figs. 21, 23). The deformation grid for broad and depressed, whereas the anterior cra- T. rathbuni (Fig. 2e) illustrates a radically dif- nial elements of E. tridentifera, while slightly ferent set of proportional modifications involv- depressed, are not broadened and do not differ ing the anterior cranial elements, and a lack of substantially from the condition characterizing enlargement in the suspensorial region. The hemidactyliine larvae in general. This point of skull of T. rathbuni cannot be derived by ex- distinction is easily seen in the skull photo- trapolation of the trend which characterizes graphs presented by Mitchell and Smith (1972), troglobitic adaptation in the Texas Eurycea; it despite the fact that the figured specimen of T. represents an entirely different solution to the POTTER AND SWEET-TYPHLOMOLGE ROBUSTA DESCRIPTION 69

yj11,1

L4-/'\l' \ bI \ I

Fig. 2. Dorsal and lateral views of the skulls of: a), b) Euryceaneotenes (MVZ 120347, 40.5 mm SL, skull length 7.0 mm); c), d) E. tridentifera(MVZ 120564, 44.8 mm SL, skull length 8.1 mm); and e), f) Typhlomolge rathbuni(USNM 51321, 51.2 mm SL, skull length 13.1 mm). The skulls are not drawn to scale, and only the right mandible is illustratedin dorsal view. The grid lines in Fig. 2c, d, e, f demonstrate the deformations necessaryto derive the skulls of E. tridentiferaand T. rathbunifrom the skull of E. neotenes,as discussedin the text.

design problems encountered by troglobitic sal- the limitations imposed by paedogenesis and amanders, and may be taken to indicate that T. troglobitic specialization make a comparable rathbuni is not a member of the same lineage as evaluation difficult for Typhlomolgeand Haideo- the Texas Eurycea. We believe that the existence triton. Wake's first assumption was that the de- of different design solutions to life in the same gree of troglobitic adaptation displayed by a environment represents strong evidence against species is a direct correlate of the duration of an hypothesized monophyletic relationship its occupation of the cave environment; while among the Edwards Plateau hemidactyliines, recognizing that such factors as population size and recommend the continued recognition of and the intensity of selective regimes can act to the genus Typhlomolge. negate this assumption, it is nonetheless one tacitly made by most workers. In the absence of Origin of Typhlomolge.-Wake's (1966) inter- biochemical or other evidence to the contrary, pretation of the historical zoogeography of the we accept this assumption in its general form, central Texas hemidactyliines is not necessarily and infer that T. rathbuni has been a troglobite supported by the acceptance of Typhlomolgeas for a greater length of time than has any of the a lineage distinct from Eurycea. This model troglobitic species of Eurycea. seems to be based on one observation and two In proposing that salamanders ancestral to assumptions, one of which has been challenged Typhlomolge reached the central Texas region by Mitchell and Smith (1972). Three specialized in early Tertiary time, Wake (1966) assumed troglobitic genera (Typhlomolge,Typhlotriton and the existence of conditions permitting the col- Haideotriton) sharing affinities with Eurycea oc- onization of caves during the middle Tertiary. cur in areas now peripheral to the Appalachian Mitchell and Smith (1972) and others working center of diversity of that genus. Osteological with the troglobitic fauna of the Edwards Pla- and myological evidence is good that Typhlotri- teau (Barr, 1960; Holsinger, 1967) have main- ton is a less-derived taxon than is Eurycea tained that few if any caves were present in this (Wake, 1966; Lombard and Wake, 1977), but region prior to late Pliocene or early Pleistocene 70 COPEIA, 1981, NO. 1 time; Mitchell and Smith cited this chronology fissure through which the spring issued was ex- as evidence against a dual ancestry for the Ed- cavated to a depth of about 6 m; no increase in wards Plateau hemidactyliines. Against this water flow resulted, but several white salaman- view, the regional geologic literature provides ders emerged from the fissure and four were ample evidence that the Edwards Plateau was collected by Mr. Wilson. Left unattended, two defined as a structural feature by the late specimens were consumed by a heron. The re- Eocene (Sweet, 1978b, reviewed this evidence); maining pair was presented to C. S. Smith of recent work (Abbott, 1974, 1975; Mench, 1978) Southwest Texas State College, but by 1961 suggests that the Balcones Aquifer system now only the larger specimen remained, and was inhabited by T. rathbuni was developed in Mio- donated to the senior author by W. K. Davis. cene time. Briefly noted, deltaic deposits along This specimen was designated as the holotype the Texas coast demonstrate a shift from dis- of Typhlomolgerobusta in an unpublished thesis tant-source to local-source fluvial systems in the (Potter, 1963). Subsequent to 1951 the type lo- late Eocene (Fisher, 1969; Fisher et al., 1970), cality was covered by the Blanco River and indicating the elevation of the Edwards Plateau. filled with gravel and silt; the present river The Oakville Sandstone (of early to middle channel lies to the east of this site, which has Miocene age) contains reworked fossils and been located (Russell, 1976) and partially ex- chert deposits (Weeks, 1945) which can only cavated. have come from erosion of the lower Creta- Detailed examination of the sole remaining ceous strata of an elevated plateau (Sweet, specimen convinced us that it does not repre- 1978b) from which a considerable thickness of sent an aberrant individual of Typhlomolgerath- overlying deposits had already been stripped buni; work on a formal description of this sal- away. Cave habitats were certainly present at amander was in progress when the name least along the eastern edge of the Edwards Typhlomolgerobusta was inadvertently occupied Plateau by middle Miocene time in concert with by Longley (1978). Following citation of the the development of the Balcones Aquifer, and type locality, Longley stated that the specimen were thus available for colonization within the "... exhibited considerable morphological vari- limits of the chronology proposed by Wake. ation from other forms. It was different since The cave systems of the plateau interior are it had a broad, stocky body and moderately more recent features, and their colonization by short, thick limbs. In the thesis it was described populations of Eurycea neotenes continues at as a new species, Typhlomolgerobusta (Potter, present (Sweet, 1977a, 1978b). Cave systems in 1963)." (Although Longley's information de- the region inhabited by E. tridentifera (Sweet, rives entirely from Potter [1963], the author- 1977b) are in an advanced stage of develop- ship of the name cannot be Potter in Longley, ment, and are probably considerably older than 1978, by reason of the exclusion of unpublished the present systems developing elsewhere on theses from eligibility as valid references by the the Edwards Plateau. International Commission on Zoological No- menclature.) We therefore present the follow- Redescriptionof Typhlomolge robusta.-The ex- ing redescription. istence of a single specimen of a distinctively salamander from the San specialized troglobitic Typhlomolgerobusta Longley Marcos Texas, has been region, Hays County, Figs. 3, 4 noted in several accounts (Potter, 1963; Red- dell, 1967; Russell, 1976; Longley, 1978; Sweet Holotype.--Texas Natural History Collection and Potter, 1978). Based on information given 20255, a mature female, collected beneath the by Potter (1963), Russell (1976) and W. K. Davis Blanco River, 178 m elevation, 5 airline km NE (pers. comm., 1970), the history of this speci- of the Hays County Courthouse, San Marcos, men is as follows. During the summer of 1951 Hays County, Texas, on 23 July 1951 by the effects of a protracted drought reduced or MacBride B. Wilson. eliminated surface water supplies throughout central Texas. Attempting to maintain local Diagnosis.-A stout-bodied, depigmented pae- gravel-washing operations, a crew led by dogenetic salamander with very reduced eyes, MacBride B. Wilson was detailed to investigate 13 trunk vertebrae, robust limbs which overlap a small spring in the dry bed of the Blanco Riv- by one costal fold when adpressed, and a thick er northeast of San Marcos. A narrow vertical tail with moderately high fins. Typhlomolgero- POTTER AND SWEET-TYPHLOMOLGE ROBUSTA DESCRIPTION 71

Fig. 3. Dorsal views of the holotype of Typhlomolgerobusta (TNHC 20255; upper specimen)and a specimen of Typhlomolgerathbuni (FEP 10). Scale bar = 50 mm. busta differs from troglobitic species of Eurycea from eye to base of third gill ramus, and 15.1 and resembles T. rathbuni in vertebral number, mm (10.7 mm) in width at the level of the jaw the disproportionate lateral expansion of the articulation. The eyes are 0.3 mm (0.46 mm) in anterior cranial elements (premaxillae, frontals, diameter, deeply buried beneath the skin, and vomers and palatopterygoids), the absence of 5.2 mm (4.9 mm) apart. The trunk is stout and ossified orbitosphenoids, and an arcuate (as op- somewhat depressed, 13.6 mm (9.1 mm) in posed to angular) lateral outline of the man- maximum width; axilla-groin length 29.7 mm dible. Typhlomolgerobusta is distinguished from (26.4 mm); there are 12 (12) costal grooves. T. rathbuni by its robust body form (Fig. 3), The limbs are long and comparatively thick: longer trunk and slightly shorter limbs (ad- forelimb length 17.3 mm (19.1 mm); upper pressed limbs overlap by 5-9 costal folds in T. forelimb width 3.3 mm (1.6 mm); hindlimb rathbuni), broad, rounded tail, and rounded (as length 19.6 mm (21.6 mm); upper hindlimb opposed to oval) skull outline in dorsal view width 3.8 mm (1.8 mm). The feet are small but (Fig. 4). stout; digits 4:5, 3-2-4-1 and 3-4-2-5-1 in order of decreasing length; phalangeal formulae 1-2- Description.-The holotype measures 57.1 mm 3-2 and 1-2-3-3-2. The tail of the holotype mea- in standard length and 100.8 mm in total sures 43.7 mm (44.8 mm); it is compressed oval length. In the following series of measurements in basal cross-section and bears a comparatively the mean values for ten T. rathbuni (54.0-60.6 high dorsal fin beginning as a fold on the pos- mm SL, f 56.1 mm, and 94.0-103.2 mm in total terior trunk and extending about 2 mm beyond length, X 100.9 mm) are noted in parentheses. the last caudal vertebra to form a rounded tip The head of the type is large and very broad, (the tail fin of T. rathbuni is slender, arising measuring 14.3 mm (14.2 mm) from snout to above the cloaca and diminishing to a point at gular fold, 20.7 mm (19.4 mm) from snout to the last caudal vertebra). base of third gill ramus, 13.7 mm (14.3 mm) The skeletal morphology of the holotype has 72 COPEIA, 1981, NO. 1

melanin is present on the dorsal peritoneal membrane. A small incision reveals the holotype to be a gravid female with an undeter- mined number of 0.8-0.9 mm yolked ova.

Remarks.-It is evident that T. robusta is refer- able to the genus Typhlomolgeand supports the distinction 'previously drawn between that genus and Eurycea, if T. robusta is in fact specifi- cally distinct from T. rathbuni. The external appearance of the holotype might result from edema or storage in hypotonic preservative, though an edematous gravid specimen of E. tridentifera (M.V.Z. 120563) shows no involve- ment of the head, limbs or tail, and no comparable condition has been noted among nu- 0 3 merous poorly-preserved specimens of T. rathbuni. Internal examination to re- mm I appears I solve this in favor of the 4. Dorsal view of the skull of the of question recognition Fig. holotype of T. robusta; the trunk musculature of the ho- robusta 20255), the lat- Typhlomolge (TNHC showing is and the distinctive erally-expandedanterior cranial elements, broad pal- lotype strongly developed, skull bears no evidence of deform- atopterygoid and elongate suspensorium character- pathological istic of the genus Typhlomolge. Drawn from ity; neither feature is represented or ap- radiographs;left mandible and right coronoid omit- proached in the series of T. rathbuni examined. ted. The salamanders of the genus Typhlomolge are distributed along the scarp of the Balcones Fault Zone at the eastern margin of the Ed- been examined radiographically. The skull wards Plateau. In an obscurely published re- (Fig. 4) is rounded in outline and depressed, port Russell (1976) provided substantial new with laterally-expanded anterior cranial ele- information on the occurrence of these sala- ments and long, bowed, anteriorly-directed sus- manders in relation to local geology and hy- pensoria; the frontals are laterally produced drology, and expanded the known range of dorsal to the antorbital cartilages; the palatop- Typhlomolgerathbuni (Fig. 5). In addition to the terygoids reach the parasphenoid medially and localities reported by Uhlenhuth (1921) (John- articulate with the quadrates posteriorly. The son's Well, Ezell's Cave, Wonder [=Beaver] orbitosphenoid (which can be clearly seen in Cave and the U.S. Fisheries Commission arte- radiographs of E. tridentifera) is not present as sian well), Russell documents a new locality an ossified element. Premaxillary, vomerine, (Rattlesnake Cave) northeast of San Marcos, palatopterygoid, coronoid and dentary teeth and points out that since about 1940 a nearby number 41, 30, 20, 30 and 62, respectively; natural fissure (Primer's Well) has been con- these values are within the ranges for teeth of fused with the Johnson's Well site of Uhlen- T. rathbuni of comparable size (means for six huth. Longley (1978) added San Marcos specimens 54.0-57.5 mm SL: 40, 31, 21, 26, Springs to the list of localities for T. rathbuni. and 71 teeth, respectively). As in T. rathbuni the The type locality for T. robusta and each of vertebral column of T. robusta consists of the the seven reported localities for T. rathbuni all atlas, 13 trunk, one sacral, three caudosacral, lie within a few meters of the trace of the San and 33 caudal vertebrae (rathbuni, 22-23, ic 30 Marcos Springs Fault, in a zone less than 6 km caudal vertebrae). The maximum diameter of in length (Fig. 5). As Russell indicated, this pat- the limb bones cannot be determined from ra- tern could reflect merely the local distribution diographs, but the elements do not seem mark- of springs, caves and wells rather than the ac- edly thicker in T. robusta than in T. rathbuni. tual ranges of T. rathbuni and T. robusta. Hy- The coloration of the preserved holotype is drologic data reviewed by Russell indicated the a uniform dark ivory (stated to have been existence of a relatively open system of large "white" in life); no indication of external guan- anastomosing water passages throughout an ophores or melanophores can be seen, though elongate oval region (termed the San Marcos POTTER AND SWEET-TYPHLOMOLGE ROBUSTA DESCRIPTION 73

7' SanMarcos/I so ;i$ Fal 41?rgatOvl SanMarcos River BlancoRiver

Fig. 5. Map of the San Marcosarea, Hays County, Texas, showing the known distributionof Typhlomolge rathbuni(sites 1-7) and T. robustain relation to major faults of the Balcones Fault Zone. Key to numbered sites: 1, Johnson's Well; 2, Primer's Well; 3, Ezell's Cave; 4, Wonder (=Beaver) Cave; 5, U.S. Fisheries Commissionartesian well; 6, San MarcosSprings; 7, RattlesnakeCave; 8, type localityof Typhlomolgerobusta.

Pool) about 25 km in length and 10 km wide, ders in at least one cave in this area (Russell, which trends NE-SW beneath the edge of the 1976). Edwards Plateau in the vicinity of San Marcos. This zone is limited on the northwest by steep ACKNOWLEDGMENTS hydrologic gradients indicative of small or poorly-connected passages, and on the south- We wish to thank Glenn Longley of South- east by the trend of the Comal Springs Fault, west Texas State University for his ready co- which in the San Marcos area demarkates an operation in the later stages of our work, and eastern band of the Balcones Aquifer character- express our appreciation to Hobart M. Smith, ized by high solute concentrations. Populations Clark Hubbs, Richard V. Melville and Arden of Typhlomolgemay range throughout the San H. Brame, Jr., II, for their interest and advice Marcos Pool, and could have access to a much in matters of nomenclature. We are grateful to larger area along the eastern edge of the Ed- numerous individuals who facilitated exami- wards Plateau (Russell, 1976; Longley, 1978). nation of specimens under their care; special The probable range of T. robusta is difficult thanks are due Robert W. Mitchell and James to judge except to note that it does not include R. Reddell (Texas Tech University) for the loan the known localities for T. rathbuni. The type of an important collection of troglobitic mate- locality lies in a downfaulted block of Austin rial. Harry W. Greene, David B. Wake and Chalk overlying the Edwards Limestone else- John E. Cadle kindly provided constructive where inhabited by T. rathbuni. The Austin comments on the manuscript. Chalk is of a lithology unsuited to the development of caves, leading Russell to propose that the salamanders found at the type locality LITERATURE CITED had from in the dispersed passages underlying P. L. 1974. Calcitizationof Edwards Edwards Limestone via the trace of the San ABBOTT, Group dolomites in the Balconesfault zone south- Marcos located some 30 m to the aquifer, Springs Fault, central Texas. Geology 2:359-362. east. Russell that T. in- suggested robusta may . 1975. On the hydrology of the Edwards habit deeper levels of the Balcones Aquifer Limestone, south-central Texas. J. Hydrology than does T. rathbuni; however, this type of dis- 24:251-269. tribution should result in specimens of both BAKER, J. K. 1957. Euryceatroglodytes: a new blind species being discharged from San Marcos cave salamanderfrom Texas. Texas J. Sci. 9:328- 336. Springs, where only T. rathbuni has been re- T. 1960. The cavernicolousbeetles of to date 1978). The absence of BARR, C., JR. ported (Longley, the T. robusta from localities within the central subgenus Rhadine, genus Agonum (Coleop- por- Amer. Midl. Natur. 64:45-65. tion of the San Marcos Pool that tera:Carabidae). may suggest --. 1968. Cave and the evolution of tro- its distribution the Balcones to ecology lies in Aquifer globites, p. 35-102. In: Evolutionarybiology. Vol. the north and east of the Blanco River. There 2. Th. Dobzhansky,M. K. Hecht and W. C. Steere are unverified reports of large white salaman- (eds.). Appleton-Century-Crofts,New York. 74 COPEIA, 1981, NO. 1

BESHARSE, J. C., AND J. R. HOLSINGER. 1977. Gyri- ships: the proteid salamanders-a test case. Amer. nophilus subterraneus, a new troglobitic salamander Natur. 110:653-677. from southern West Virginia. Copeia 1977:624- HENDRICKS, F. S. 1973. Systematic re-evaluation of 634. the central Texas paedogenetic salamanders by BISHOP,S. C. 1941. Notes on salamanders with de- utilizing multivariate analysis. H.I.S.S. News-J. scriptions of several new forms. Occ. Pap. Mus. 1:111. Zool. Univ. Mich. 451. HOLSINGER, J. R. 1967. Systematics, speciation, and , AND M. R. WRIGHT. 1937. A new neotenic distribution of the subterranean amphipod genus salamander from Texas. Proc. Biol. Soc. Wash. Stygonectes(Gammaridae). U.S. Nat. Mus. Bull. 259. 50:141-143. HORSTADIUS, S., AND S. SELLMAN. 1946. Experimen- BLAIR, A. P. 1968. Amphibians, p. 211-271. In: Ver- tal Untersuchungen fiber die Determination des tebrates of the United States. W. F. Blair, A. P. knorpeligen Kopfskelettes bei Urodelen. Nova Blair, P. Brodkorb, F. Cagle and G. A. Moore. Acta Reg. Soc. Sci. Upsal. Series 4 13:1-170. McGraw-Hill Book Co., New York. HUXLEY, J. 1932. Problems of relative growth. Meth- BOGART, J. P. 1967. Life history and chromosomes uen and Co., London. of some of the neotenic salamanders of the Ed- KINGSBURY, B. F., AND H. D. REED. 1909. The col- wards Plateau. Unpubl. MA Thesis, Univ. Texas, umella auris in Amphibia. J. Morphol. 20:549-628. Austin. LANDACRE, F. L. 1921. The fate of the neural crest BOOKSTEIN,F. L. 1978. The measurement of biolog- in the head of urodeles. J. Comp. Neurol. 33:1-43. ical shape and shape change. Lecture notes in LARSON, J. H., JR., AND D. J. GUTHRIE. 1974. Paral- biomathematics 24. Springer-Verlag, New York. lelism in the Proteidae reconsidered. Copeia BRAME, A. H., JR. 1967. A list of the world's Recent 1974:635-643. and fossil salamanders. Herpeton 2:1-26. LEVITON, A. E. 1971. Reptiles and amphibians of BRANDON, R. A. 1968. Structure of the eye of Hai- North America. Doubleday and Co., New York. deotritonwallacei, a North American troglobitic sal- LOMBARD,R. E., AND D. B. WAKE. 1977. Tongue amander. J. Morphol. 124:345-351. evolution in the lungless salamanders, family Pleth- . 1971. North American troglobitic salaman- odontidae. II. Functional and evolutionary diver- ders: some aspects of modification in cave habitats, sity. J. Morphol. 153:39-79. with special reference to Gyrinophilus palleucus. LONGLEY, G. 1978. Status of Typhlomolge(=Eurycea) Bull. Nat. Speleol. Soc. 33:1-21. rathbuni, the Texas blind salamander. Endangered COLLINS,J. T., J. E. HUHEEY, J. L. KNIGHT AND H. M. Species Report 2. U.S. Fish and Wildlife Service, SMITH. 1978. Standard common and current sci- Albuquerque, New Mexico. entific names for North American amphibians and MENCH, P. 1978. Diagenesis related to Miocene reptiles. Soc. Study Amphib. Rept., Herp. Circ. 7. faulting in the Edwards Aquifer, San Antonio area, CONANT, R. 1975. A field guide to reptiles and am- Texas. Geol. Soc. Amer. Program (South-Central phibians of eastern and central North America. Section) 10:22. Houghton Mifflin Co., Boston. MITCHELL, R. W., AND J. R. REDDELL. 1965. Eurycea DUNDEE, H. A. 1973. A synoptic classification of the tridentifera, a new species of troglobitic salamander modern Lissamphibia. Tulane Univ., New Orleans. from Texas and a reclassification of Typhlomolge EMERSON,E. T., 2ND. 1905. General anatomy of Ty- rathbuni. Texas J. Sci. 17:12-17. phlomolge rathbuni. Proc. Boston Soc. Natur. Hist. - , ANDR. E. SMITH. 1972. Some aspects of the 32:43-76. osteology and evolution of the neotenic spring and FISHER,W. L. 1969. Facies characterization of Gulf cave salamanders (Eurycea, Plethodontidae) of cen- Coast basin delta systems, with some Holocene an- tral Texas. Ibid. 23:343-362. alogues. Gulf Coast Assoc. Geol. Socs. Trans. MONATH, T. 1965. The opercular apparatus of sal- 19:239-261. amanders. J. Morphol. 116:149-170. -- , C. V. PROCTOR,JR., W. E. GALLOWAYAND J. POTTER, F. E., JR. 1963. Gross morphological vari- S. NAGLE. 1970. Depositional systems in the Jack- ation in the genus Typhlomolgewith a description son Group of Texas. Ibid. 20:234-261. of a new species. Unpubl. MA Thesis, Univ. Texas, Fox, H. 1954. Development of the skull and asso- Austin. ciated structures in the Amphibia with special ref- POULSON, T. L. 1963. Cave adaptation in amblyopsid erence to the urodeles. Trans. Zool. Soc. Lond. fishes. Amer. Midl. Natur. 70:257-290. 28:241-304. . 1964. Animals in aquatic environments: an- FRANCIS,E. T. B. 1934. The anatomy of the sala- imals in caves, p. 749-771. In: Handbook of phys- mander. Oxford Univ. Press, London. iology: adaptation to the environment. D. B. Dill GORHAM,S. W. 1974. Checklist of world amphibians (ed.). Amer. Physiol. Soc., Washington, D.C. up to January 1, 1970. New Brunswick Mus., Saint , ANDW. B. WHITE. 1969. The cave environ- John. ment. Science 165:971-981. HECHT, M. K., AND J. L. EDWARDS. 1976. The de- RAUN, G. G. 1971. Herpetology of Texas caves, p. termination of parallel or monophyletic relation- 100-107. In: Natural history of Texas caves. E. L. POTTER AND SWEET-TYPHLOMOLGE ROBUSTA DESCRIPTION 75

Lundelius, Jr., and B. H. Slaughter (eds.). Gulf - 1978b. The evolutionary development of the Natur. Hist., Dallas, Texas. Texas Eurycea (Amphibia: Plethodontidae). Un- - , ANDF. R. GEHLBACH. 1972. Amphibians and publ. PhD Dissertation, Univ. Calif., Berkeley. reptiles in Texas. Dallas Mus. Natur. Hist. Bull. 2. AND F. E. POTTER, JR. 1978. The validity of REDDELL, J. R. 1967. A checklist of the cave fauna -~,the genus Typhlomolge:a test with a new species. of Texas. III. Vertebrata. Texas J. Sci. 19:184-226. Program Joint Annual Meeting A.S.I.H., H.L., . 1971. A checklist of the cave fauna of Texas. S.S.A.R., Tempe, Arizona. VI. Additional records of Vertebrata. Ibid. 22:139- THOMAS,R. A. 1974. A checklist of Texas amphib- 158. ians and reptiles. Tech. Ser. 17, Texas Parks and RUSSELL,W. H. 1976. Distribution of troglobitic sal- Wildlife Dept., Austin. amanders in the San Marcos area, Hays Co., Texas. . 1976. Ibid. Report 7601. Texas Assoc. Biol. Invest. Trogl. Eu- THOMPSON,D'ARCY W. 1917. On growth and form. rycea, Austin. Cambridge Univ. Press. SMITH,H. M. 1978. Amphibians of North America: TUPA, D. D., AND W. K. DAVIS. 1976. Population a guide to field identification. Golden Press, New dynamics of the San Marcos salamander, Eurycea York. nana Bishop. Texas J. Sci. 32:179-195. - , AND F. E. POTTER, JR. 1946. A third neotenic UHLENHUTH, E. 1921. Observations on the distri- salamander of the genus Eurycea from Texas. Her- bution and habits of the blind cave salamander, petologica 3:105-109. Typhlomolgerathbuni. Biol. Bull. 40:73-104. STEJNEGER, L. 1896. Description of a new genus and WAKE,D. B. 1966. Comparative osteology and evo- species of blind tailed batrachians from the subter- lution of the lungless salamanders, family Pletho- ranean waters of Texas. Proc. U.S. Nat. Mus. dontidae. Mem. So. Calif. Acad. Sci. 4. 18:619-621. WEEKS, A. W. 1945. Oakville, Cuero, and Goliad for- SWEET, S. S. 1977a. Natural metamorphosis in Eu- mations of Texas coastal plain between Brazos Riv- ryceaneotenes, and the generic allocation of the Tex- er and Rio Grande. Amer. Assoc. Petrol. Geol. as Eurycea (Amphibia: Plethodontidae). Herpeto- 29:1721-1732. logica 33:364-375. TEXAS PARKS AND WILDLIFE DEPARTMENT, AUS- 1977b. Cat. Amer. Am- Eurycea tridentifera. TIN, TEXAS 78744 AND DEPARTMENT OF BIO- 199.1-199.2. phib. Rept.: LOGICAL SCIENCES, UNIVERSITY OF CALIFORNIA, . 1978a. On the status of Eurycea pterophila SANTA BARBARA, CALIFORNIA 93106. Plethodontidae). 34:101- Accepted (Amphibia: Herpetologica 17 1979. 108. Nov.