Eurycea subfluvicola, Steffen & Irwin & Blair & Bonett, 2014

Steffen, Michael A., Irwin, Kelly J., Blair, Andrea L. & Bonett, Ronald M., 2014, Larval masquerade: a new species of paedomorphic salamander (Caudata: Plethodontidae: Eurycea) from the Ouachita Mountains of North America, Zootaxa 3786 (4), pp. 423-442 : 431-439

Eurycea subfluvicola sp. nov.

Ouachita Streambed Salamander

Holotype. The holotype is an adult female collected by MAS and KJI on 25-March 2013 and deposited at the Museum of Vertebrate Zoology, University of California, Berkeley ( MVZ 269485 ; Fig. 3 ). Detailed measurements (mm) post preservation is listed in Table 4 . Though slightly faded in preservation, the holotype is similar in coloration and pattern as described in the species diagnosis. The holotype contains nine enlarged oviductal eggs that are visible through the venter.

Paratypes. Paratypes are one formalin fixed specimen ( AMNH A191438 ) and seven cleared and stained specimens deposited at the American Museum of Natural History in New York City , New York ( AMNH A191439 - A191441 ) and the Museum of Vertebrate Zoology ( MVZ 269485-269489 ) .

Type locality. A unnamed first order tributary of Slunger Creek, located in the Trap Mountains, a subdivision in the southeastern portion of the Ouachita Mountain physiographic province, Hot Spring County, Arkansas, USA. This locality is within the confines of Lake Catherine State Park.

Etymology. The specific name, s ubfluvicola, is an adjective derived from the Latin prefix, sub - meaning “below”, fluvius which is “a stream”, and colo meaning “to dwell.” The translation then is “dwells below the stream”, in reference to its existence below the streambed during xeric conditions.

Diagnosis. This species is assigned to the genus Eurycea based on strong molecular phylogenetic evidence, and overall similarity to other larval Eurycea . The genus Eurycea is highly heteromorphic, with no single character to diagnosis it from other spelerpine genera ( Wake 1966; Camp et al. 2009; Adams et al. 2009). However, members are consistently monophyletic in molecular phylogenetic studies ( Camp et al. 2009; Bonett et al. 2014).

The dorsum of Eurycea subfluvicola is primarily uniform amber/yellow background color, pigmented with numerous dark brown melanophores, which create irregularly shaped blotches throughout the dorsum and flanks ( Fig. 3 ). In most individuals irregularly spaced spots are formed by the absence of melanophores along the dorsolateral region of the trunk, possibly indicative of the lateral line. The semi-transparent venter is unpigmented, except for a few widely dispersed melanophores beneath the tail. Dorsal and ventral coloration is separated by a sharply defined ventral-lateral boundary along the trunk.

Morphological characteristics that distinguish Eurycea subfluvicola from syntopic E. multiplicata larvae are: 1) Eurycea subfluvicola is paedomorphic (i.e. presence of mature testes and ova, and deposition of fertilized eggs while in the larval form) and, based on our samples, is sexually mature by at least 32mm SVL. Eurycea multiplicata are never known to be paedomorphic and typically metamorphose between 24 to 41mm SVL. 2) Head and body shape differ significantly between the two species. PCAs show only minimal overlap in morphospace between body shape, and no overlap in head shape ( Fig. 4 ). Additionally, the Goodall’s F-test on geometric morphometric data was highly significant for heads, and bodies (F 82,270 6 =35.85, p<0.001 and F 18,630 =13.60, p<0.001). The same pattern was also seen when comparing only small E. subfluvicola (<40mm SVL) to E. multiplicata larvae (F 82,155 8 =25.53, p<0.001, and F 18,324 =4.94, p<0.001). 3) Relative to SVL, E. subfluvicola has a longer trunk, and a shorter and narrower head compared to E. multiplicata larvae ( Table 5 ). This may be driven by the fact that E. subfluvicola has a very long trunk relative to its head, but these differences are not significant. 4) In profile, Eurycea subfluvicola has a flat head and longer snout (ANCOVA, p <0.001; Table 5 ) with relatively smaller in diameter and depressed eyes, compared to E. multiplicata larvae (ANCOVA, p <0.05; Fig. 5 , Table 5 ). 5) Individuals of E.subfluvicola are more attenuate in body form than E. multiplicata larvae, having smaller diameter trunks and tail widths, with the latter being highly significant (ANCOVA, p <0.001; Table 5 ). 6) E. multiplicata have many more iridophores in their irises, whereas E. subfluvicola has reduced coloration in its iris making the eye primarily black in color. 7) A prominent black stripe of pigmentation is present along the lateral side of the snout and head in E. multiplicata , this stripe is absent in E. subfluvicola ( Fig. 5 ).

Larval morphology of plethodontids is highly conserved, and E. subfluvicola is very similar to the larvae of other species of Eurycea ( Wake 1966) . Trunk vertebral counts (between atlas and sacrum) are 21-22 (n=7, mode=21), whereas syntopic and nearby populations of E. multiplicata (n=17) have 21 vertebrae. The trunk vertebrae of E. subfluvicola also appears to be longer and narrower than E. multiplicata ( Fig. 6 ). The phalanges are well ossified with cartilaginous caps, exhibiting the standard plethodontid phalangeal formulae: 1-2-3-2 for the forelimbs, and 1-2-3-3-2 for the hindlimbs ( Shubin & Wake 2003). Carpal counts are eight, and tarsal counts are nine.

The quadrate and squamosal, which are part of the mandibular suspensorium, are collectively more robust in E. subfluvicola than in E. multiplicata ( Fig. 6 ). As in most spelerpine larval forms, E. subfluvicola contain 3 cartilaginous epibranchials, and in some individuals these structures are partially ossified (two of five specimens), whereas no ossification is apparent in E. multiplicata larvae (n=10; Fig. 6 ).

Phylogenetic Reconstruction and Divergence Time Estimation. As in previous phylogenetic analyses our Cytb phylogeny strongly supports the monophyly of the Eurycea multiplicata complex, which includes E. multiplicata , E. tynerensis , E. spelaea ( Bonett & Chippindale 2004; Bonett et al. 2014), and now E. subfluvicola ( Fig. 7a ). Eurycea subfluvicola is sister to the clade that includes all representative E. multiplicata populations from throughout the Ouachita Mountains (BAPP=0.99). Uncorrected pairwise sequence divergence between E. subfluvicola and E. multiplicata ranges from 13.5-16.6% (mean=15%) for Cytb. This is more than two to five times the divergence of many recognized sister species of plethodontid salamanders ( Chippindale et al. 2000; Hillis et al. 2001; Jockusch et al. 2002; Garcia-Paris et al. 2000; Vences & Wake 2007).

The phylogeny of Rag1 ( Fig. 7b ) also shows E. subfluvicola as sister to all E. multiplicata (BAPP=0.77), with relatively high sequence divergence between these taxa for Rag1, uncorrected p=0.9-2.3% (mean=1.4%), which is considerable given that it is a much more conserved nuclear gene ( Hoegg et al. 2004). What is most informative about the DNA sequence data is that it shows E. subfluvicola is highly divergent from syntopic E. multiplicata for both a mitochondrial (n= 32 specimens) and a nuclear (n= 15 specimens) gene ( Fig. 7 ).

Bayesian species-tree and divergence time estimation based on mitochondrial (Cytb and Nd4) and nuclear (Rag1) gene trees shows strong support for the sister relationship of E. subfluvicola and E. multiplicata ( Fig. 8 ). If we constrain the basal node of all spelerpine plethodontids to be 49 million years old ( Bonett et al. 2014), then the divergence between E. subfluvicola and E. multiplicata dates to approximately 18 million years ago.

Variation. Variation is based on 24 specimens of E. subfluvicola (8 males, 16 females). All 24 individuals were adults, based on presence of mature or maturing egg follicles, or enlarged testes ( Figs. 3 and 9 ). Male SVL ranged in size from 34.9-45.7mm (mean=40.8) and between 31.4-48.0mm (mean=41.1) for females. There are no somatic differences between the sexes, and only minor differences in coloration among individuals. We found 0.1- 0.3% nucleotide divergence (uncorrected p) in Cytb and 18 sites with heterozygous positions across all individuals of E. subfluvicola for Rag1.

Distribution and Habitat. The known distribution of Eurycea subfluvicola is currently limited to the Trap Mountains, a range of generally east-west trending ridges in the southeastern portion of the Ouachita Mountain physiographic province. These uplands are composed of Paleozoic marine sedimentary rocks; principally the Devonian/Mississippian Arkansas Novaculite and younger Mississippian Stanley Shale formations. Stream valley alluvium deposits overlie the Stanley Shale and predominantly consist of Arkansas Novaculite gravels and cobbles. Eurycea subfluvicola is only known from one locality where it can be found in two nearby sites. The two sites are a 15m section of Slunger Creek and a 50m section of an unnamed tributary within the Slunger Creek alluvial valley; approximately 135m apart from one another. During our 2013 surveys, we found more individuals of E. subfluvicola (approximately 50 observations) in the longer stream section, while fewer individuals were found in the smaller stream section (<10 observations). Surface stream flows are seasonally ephemeral in these two stream reaches. However, surface flow persists, both upstream and downstream, when the streambed is completely dry at these two sites. This suggests a subsurface flow in the stream channels or hyporheic zone connectivity between these points, which may provide habitat for E. subfluvicola during periods of low rainfall. Our ability to observe E. subfluvicola is dependent upon the rise of the water table during winter and spring, which is similar to the seasonal observability of paedomorphic populations of E. tynerensis in the western Ozark Plateau ( Bonett & Chippindale 2006). The Slunger Creek valley supports two metamorphic, stream-dwelling plethodontids ( E. multiplicata and Desmognathus brimleyorum ). Dominant vegetation at the type locality is second growth mixed pine-hardwood forest composed of Pinus echinata and several hardwoods ( Acer , Celtis , Cornus , Liquidambar, Ulmus , and Quercus ).

Life history and ecology. Most aspects of E. subfluvicola life history are derived from observations in captivity, and reproductive status has been determined by examining the gonads through the transparent venter. Enlargement of testes was first observed in the field in February, and reached maximum size in the lab by late June and July. Like related E. multiplicata and E. tynerensis , E. subfluvicola typically has single lobed testes ( Sever 1974; MAS pers. obs.), but one individual possessed multiple lobes ( Fig. 9b ). Mature or maturing egg follicles have only once been observed in the field (late November), however, follicles started maturing in the lab (at 20 o C) in June and were robust by August. We have not yet found definitive juveniles of E. subfluvicola in the field, which suggests that early development (<32mm SVL) may be primarily subterranean.

Coinciding with peak testicular size, the five largest captive males began to show signs of metamorphosis (i.e. gill reduction, changes in head morphology and coloration). For four individuals, the process took over 2.5 months, and metamorphosis was never completed before the animals died. Eurycea multiplicata can complete metamorphosis in less than one month after initiation under similar conditions (MAS pers. obs.). Two of the partially metamorphosed E. subfluvicola showed remodeling of the hyobranchial apparatus towards the adult form ( Alberch et al. 1985). This includes the loss of larval epibranchials, and formation of the adult epibranchial, and the reshaping of the ceratohyal, basibranchials, and ceratobranchials. Adult maxillary morphology ( Wilder 1925; Wake 1966) started to form in three specimens. The small external gills were also never completely absorbed. These individuals demonstrate that facultative metamorphosis might be possible in this species, however, given that four of the five large males died during this process (with the 5 th individual still undergoing metamorphosis 5-February 2014), indicates that metamorphosis appears to be a grave event, as seen in some other paedomorphic salamanders that have naturally or experimentally metamorphosed ( Dundee 1957, 1962; Brandon 1976). It is notable that partial metamorphosis only occurred in the largest, most reproductive males (based on testicular examination), while females showed no signs of metamorphosis (n=13). We hypothesize that high levels of endogenous androgens may induce metamorphosis of male E. subfluvicola in captivity. Studies have shown that testosterone, but not estradiol, can increase the rate of metamorphosis in other salamander species ( Norris et al. 1973). Surface activity of Eurycea subfluvicola is primarily nocturnal, with individuals leaving the cover of rocks or leaf litter during the day. During periods without surface flow, individuals of E. subfluvicola presumably follow the water table using interstitial spaces in the streambed gravel. The subterranean ecology of E. subfluvicola in the hyporheic zone is completely unknown.

Figures

FIGURE 3. Dorsal and ventral images of the female holotype (MVZ 269485) of Eurycea subfluvicola. Preserved October 2013.

FIGURE 4. Principle Component Analysis (PCA) of adult E. subfluvicola (dark grey) and larval E. multiplicata (light grey), demonstrating minimal overlap in head (top) and body (bottom) morphospace. Landmarks and semilandmarks are shown in Figure 2.

FIGURE 5. Side by side lateral and dorsal images of adult E. subfluvicola (top) and larval E. multiplicata (bottom) illustrating head shape differences between syntopic species.

FIGURE 6. Cleared and stained images of E. subfluvicola (top) and E. multiplicata (bottom) skulls. Yellow arrows indicate quadrate and squamosal bones which are collectively more robust in E. subfluvicola than E. multiplicata. Red arrows point to trunk vertebrae which appear more elongate in E. subfluvicola than E. multiplicata.

FIGURE 7. The 50% majority rule phylogram from a Baysian inference analysis of mitochondrial and nuclear genes. Partitioned Bayesian analyses of Cytb (a) and Rag1 (b), based on 15 million generations, showing E. subfluvicola (bold) and its sister taxon E. multiplicata. Posterior probabilities indicating node support are shown for major nodes of ingroup clades. For ingroup taxa, the label is the field number followed by the locality number (1 through 30; Figure 1; Table 1). Note that locality number 5 includes E. subfluvicola and sympatric E. multiplicata. E=Eurycea and P=Pseudotriton.

FIGURE 8. Bayesian species tree chronogram of spelerpines plethodontid salamanders including E. subfluvicola (bold). The species tree and divergence times estimates from *BEAST are based on mitochondrial (Cytb and Nd4) and nuclear gene (Rag1) trees (see Methods). Numbers subtending each node are posterior probabilities of node support for major lineages. Node bars indicate 95% highest posterior density on divergence dates. Some redundant taxa were pruned. E=Eurycea, G=Gyrinophilus, P=Pseudotriton, S=Stereochilus, and U=Urspelerpes. AR=Arkansas, NC=North Carolina, OK=Oklahoma, and TX=Texas.

FIGURE 9. Images of reproductive status in E. subfluvicola. (a) Female with developing ova in ovaries. (b) Male with enlarged testicular lobes, and smaller lobules in development. (c) Developing embryo of E. subfluvicola (16 days old).

Tables

TABLE 4. Detailed measurements (mm) of Eurycea subfluvicola holotype.

Total Length	67.9
Standard Length (SVL)	42.0
Head length (snout to gular fold)	5.1
Head depth at posterior angle of jaw	1.8
Head width	5.0
Eye to nostril (left)	1.5
Eye diameter (horizontal)	0.8
Interorbital distance	1.4
Snout to forelimb insertion	9.5
Trunk Length	27.5
Width at mid-body	5.5
Depth at mid-body	4.2
Tail length	26.0
Tail width at base	3.4
Tail depth at base	3.0
Forelimb length to tip of longest digit (left)	4.6
Hindlimb length to tip of longest digit (left)	5.4