Salamandra quadridigitata Holbrook, 1842e:65

Pyron, R. Alexander & Beamer, David A., 2022, A nomenclatural and taxonomic review of the salamanders (Urodela) from Holbrook’s North American Herpetology, Zootaxa 5134 (2), pp. 151-196 : 174-179

Salamandra quadridigitata Holbrook, 1842e:65 , pl. 21

Holbrook (1842e) described the Dwarf Salamander, Eurycea quadridigitata (see Mittleman 1967), based on a series of specimens (primary syntypes), including some that he apparently collected himself in the “middle section” of South Carolina , others that were received from “ Mr. Couper ” (James Hamilton Couper, 1794–1866) and Dr. Harden from Georgia (presumably from their homesteads; see below), and still others from unnamed collectors from unknown localities in Florida. Schmidt (1953) later restricted the type locality to the “vicinity of Charleston, South Carolina ” without presenting further evidence. As noted elsewhere, Schmidt did this for many of Holbrook’s taxa based primarily on Holbrook’s residence in Charleston ( Adler 1976). Holbrook noted a modest resemblance of this new species to Salamandra guttolineata Holbrook, 1838a , but elaborated that “it differs in many and important characters.” He further noted that it is “the only one of our Salamanders with which I am acquainted that has four fingers and four toes, and from this circumstance is its specific name derived.” This is remarkable given that Salamandra scutata Temminck in Temminck and Schlegel, 1838 (= Hemidactylium scutatum , the Fourtoed Salamander) had been described four years earlier, and several specimens of H. scutatum in extant museum collections are attributed to Holbrook (ANSP 10590; MCZ A-1199, A-4400, and A-120827), though these likely post-date the publication of North American Herpetology .

Holbrook described only a single individual, presumably that illustrated in pl. 21 by J. Queen, which is nevertheless still a primary syntype rather than the holotype, since reference is made to multiple specimens in Holbrook’s possession. This specimen (ANSP 490) is extant and was incorrectly regarded as the “type” by Fowler and Dunn (1917) and Dunn (1926), and the “ holotype ” by Mittleman (1967), Malnate (1971), and Wray et al. (2017). Erroneous reference to ANSP 490 as the holotype does not constitute a valid lectotype designation under Article 74.5 or 74.7. Wray et al. (2017) then declared ANSP 490 “destroyed” and designated UF 178833 (field tag KW0004) as the neotype. We provide evidence that this designation did not fulfill the necessary criteria of the Code, and that ANSP 490 is extant, in good condition, can reliably be demonstrated to be a primary syntype. The neotype designation of UF178833 is consequently invalid under Article 75, and we hereby reverse it and designate ANSP 490 as the lectotype of Salamandra quadridigitata Holbrook, 1842e . Our evidence is as follows.

First, Article 75.1 states that “A neotype is the name-bearing type of a nominal species-group taxon designated under conditions specified in this Article when no name-bearing type specimen (i.e. holotype, lectotype, syntype or prior neotype) is believed to be extant and an author considers that a name-bearing type is necessary to define the nominal taxon objectively.” However, a putative name-bearing type was known to be extant at the time, and all four four-toed “dwarf” salamander species in Florida, Georgia, and Alabama were objectively well-defined by Wray et al. (2017) even without the existence of a type specimen for Eurycea quadridigitata . There was no taxonomic dispute as to which of the four species-level genetic lineages this name applied, and there was therefore no objective need for a neotype designation. The description by Holbrook (1842e) contained not only diagnostic color-pattern characteristics, but also specific attributes of the illustrated primary syntype.

Specifically, Holbrook (1842e) gave the measurements of the syntype as “Length of head, 4 lines [0.33 inches = 8.5mm]; length of body, 1 inch [25.4mm]; length of tail, 2 inches [50.8mm]: total length, 3 inches 4 lines [84.7mm].” While the tail of ANSP 490 is broken off (apparently after preservation given the jagged or ragged edges around the break; Fig. 9 ), our measurements of the specimen match Holbrook’s almost exactly, with a head length of 8.5mm and a body length of 25.4mm (SVL=33.9) occurring essentially in the middle of any reasonable range of starting and ending points for taking those measurements on the specimen. Furthermore, some authors have described the specimen as “partially dissected” ( Mittleman 1967; repeated by Wray et al. 2017). Examination of the specimen does not reveal an abdominal dissection as might ordinarily be assumed from the wording, but rather incisions at the jaw commissures allowing the dentary to be retracted to examine the maxilla, palate, and lingual apparatus. Indeed, Holbrook (1842e) gave a detailed description of these structures, which could only have been allowed by a microscopic examination of this style of preparation.

Wray et al. (2017) disputed the type status of ANSP 490 because it was not explicitly designated as such by Holbrook (1842e), and was only cataloged later under a different genus ( Manculus Cope, 1869 ). However, numerically cataloging specimens was not common practice in Holbrook’s time, and very few of his types were deposited prior to publication (see Adler 1976; Burbrink et al. 2021). Later accessioning by a worker using prevailing nomenclature is not evidence against the type status of the specimen, particularly when it can be linked to the original description by historical tradition and the documentary evidence of physical characteristics. Thus, we reliably conclude that ANSP 490 is indeed a primary syntype of Salamandra quadridigitata Holbrook, 1842e , and presumably the specimen figured in pl. 21.

Wray et al. (2017) also suggested that a neotype is necessary to clarify potential taxonomic ambiguity in the allocation of the name Salamandra quadridigitata Holbrook, 1842e to the two parapatric species occurring in South Carolina , and the three other species distributed parapatrically to the west from the Florida Panhandle to eastern Louisiana. However, Holbrook’s description is thorough, and contained the diagnostic character of “abdomen is bluish silvery-white” which distinguishes it from E. chamberlaini Harrison and Guttman, 2003 with a yellowish venter. These are the “dark” and “light” morphs originally noted by Mittleman (1947), in part. Furthermore, E. chamberlaini typically has 17 vertebrae and costal grooves while E. quadridigitata has 19, and previous X-ray photographs of ANSP 490 revealed 19 apparent vertebrae (R. Highton, pers. comm. to Harrison and Guttman 2003). Our count of the specimen also reveals 19 apparent costal grooves.

The measurements of the syntype also distinguish it from E. chamberlaini and the smaller E. hillisi Wray et al., 2017 and E. sphagnicola Wray et al., 2017 . These are the other species of four-toed “dwarf” salamander that also occur in Florida, Georgia, and South Carolina with which E. quadridigitata could potentially be confused Wray et al. (2017). Using Holbrook’s measurements of the syntype compared to the morphometric analysis given by Wray et al. (2017), an SVL of 33.9mm is slightly larger than the range of 20.8–33.8 given for E. quadridigitata (n= 136), but significantly outside the ranges of 18.0–29.5 for E. hillisi (n= 20), 16.4–25.7 for E. sphagnicola (n= 36), and 22.6–32.3 for E. chamberlaini (n= 7).

Only Eurycea paludicola Mittleman, 1947 found west of the Pascagoula River in Mississippi also obtains this size; 22.3–37.2mm (n =20). Wray et al. (2017) noted several morphometric ratios that can discriminate between E. paludicola and E. quadridigitata , but these were not given by Holbrook (1842e) and we therefore do not attempt to calculate them here. However, while the two species may occur parapatrically in southeastern Louisiana ( Lamb and Beamer 2012), this is outside the original areas given by Holbrook (1842e) which contain only the four species listed above. Thus, E. paludicola is not relevant to any taxonomic ambiguity in the nomenclatural allocation of Salamandra quadridigitata Holbrook, 1842e . Rather, the data given by Holbrook (1842e) in the original description are sufficient to distinguish E. quadridigitata from all other similar or related species in its geographic range as originally defined.

Of the calculable ratios used by Wray et al. (2017), values of SVL/total length (33.9/84.7=40%) and tail length/ total length (50.8/84.7=60%) for the syntype are inconclusive, occurring within the ranges given for all 5 species. Finally, the value for head length/SVL (8.5/33.9=25%) dramatically exceeds the range given for E. chamberlaini (18.0–21.7), while also modestly exceeding those for E. quadridigitata (15.4–24.4), E. paludicola (17.3–24.7), and E. hillisi (18.9–24.4), matching only that given for the much smaller E. sphagnicola (20.0–28.7). We do not attempt to re-measure the syntype here for these or other critical variables, but nevertheless conclude that it is clearly a rather large adult, female E. quadridigitata , and that Holbrook’s description unambiguously refers to this species.

Second, Article 75.3 lays out 7 qualifying conditions for a neotype, of which 75.3.4 requires “the author’s reasons for believing the name-bearing type specimen(s) (i.e. holotype, or lectotype, or all syntypes, or prior neotype) to be lost or destroyed, and the steps that had been taken to trace it or them.” Wray et al. (2017) stated that “this specimen [ANSP 490] is in very poor shape, partially dissected, and essentially destroyed,” but did not examine it personally and instead relied on published or secondhand accounts. On the contrary, it is clearly extant and in fairly good shape for a 180-year-old specimen. The documentary evidence provided above (including the dissection damage and morphometric data) allow it to be unambiguously assigned to both Eurycea quadridigitata as defined by Wray et al. (2017) and the original description of Salamandra quadridigitata Holbrook, 1842e .

Third, Article 75.3.6 requires “evidence that the neotype came as nearly as practicable from the original type locality.” Wray et al. (2017) restated the range of localities given by Holbrook (1842e) and Schmidt (1953) ’s restriction but did not examine potential sites. Whether ANSP 490 was collected by Holbrook himself in South Carolina , or was received from Florida or Georgia, remains unclear. However, the specimens from Couper and Harden likely originated from their plantations at Hopeton-On-The-Altamaha (GA: Glynn; 31.308, -81.460) and Woodmanston (GA: Liberty; 31.701, -81.475) from which they sent Holbrook specimens (see Stephens 1997). Similarly , while Wray et al. (2017) noted that “the middle section” of South Carolina contains both Eurycea chamberlaini and E. quadridigitata , the “vicinity of Charleston, South Carolina ” as restricted by Schmidt (1953) contains only E. quadridigitata , with numerous recent localities based on VertNet and iNaturalist records. In contrast, while Holbrook (1842e) reported receiving specimens from Florida, it is unclear that Wray et al. (2017) ’s neotype locality in Leon County, Florida was considered by Holbrook.

We provide strong supporting evidence that the specimen ANSP 490, regarded as a type for at least 100 years prior to Wray et al. (2017) is, in fact, an original primary syntype of Salamandra quadridigitata Holbrook, 1842e . Article 75.8 provides that “if, after the designation of a neotype, the name-bearing type ... of the nominal speciesgroup taxon that was ... presumed lost is ... found still to exist, on publication of that discovery the rediscovered material again becomes the name-bearing type and the neotype is set aside.” Thus, we conclude that UF 178833 is to be set aside as the neotype, and ANSP 490 again becomes a name-bearing type, of Salamandra quadridigitata Holbrook, 1842e , with the type locality “vicinity of Charleston , South Carolina’ ’ following Schmidt (1953). We hereby designate it the lectotype, rendering the other primary syntypes of unknown number, locality, and disposition, paralectotypes.

Figures

FIGURE 9. Lectotype (ANSP 490) of Salamandra quadridigitata Holbrook, 1842e, type locality in the vicinity of Charleston, S.C. The remaining primary syntypes, of unknown number and disposition, thereby become paralectotypes. This specimen was erroneously considered the holotype by Mittleman (1967), Malnate (1971), and Wray et al. (2017), who declared it destroyed and designated a neotype (UF 178833), which we reverse here. We were unable to match the handwriting on the label, but it appears modern and may be Malnate. Ruler is in mm.