Liophis trebbaui Roze

Myers, C. W. & Donnelly, M. A., 2008, The Summit Herpetofauna Of Auyantepui, Venezuela: Report From The Robert G. Goelet American Museum-Terramar Expedition, Bulletin of the American Museum of Natural History 2008 (308), pp. 1-147 : 123-130

Liophis trebbaui Roze Figures 74–77

Liophis trebbaui Roze, 1958a: 262–264 , fig. 11. Holotype AMNH R-61019 [lost, see Remarks] from [south slope] Auyantepui, [1100 m / 3609 ft fide AMNH catalog 5 Guayaraca area, about 1000–1100 m], Venezuela, collected by W. H. Phelps on ‘‘2 de abril, 1938’’ [5 February 4, 1938 fide AMNH catalog]. Donnelly and Myers, 1991: 46 (resurrection of L. ingeri and L. trebbaui ).

Liophis cobella trebbaui: Dixon, 1983a: 159 (includes L. ingeri ).

Liophis cobellus trebbaui: Dixon, 1989: 10 (incorrect subsequent spelling of cobella ).

MATERIAL: Camp 2, 1750 m: AMNH R-140231, from the 1994 AMNH – TERRA- MAR Expedition to Auyantepui. Cabanayén, 1230 m, (about 70 km SE Auyantepui): AMNH R-104803, collected by S. Gorzula in 1975. Marco de Fronteira BV8, Roraima, Brazil: MZUSP 9235–9236, collected by C. M. Carvalho in 1986.

The single specimen (figs. 74–76) from our expedition to Autantepui is a juvenile female, 221 mm total length, 35 mm tail length (15.8 % of total length); dorsal scales lacking apical pits, smooth, in 17-17-15 rows, with reduction occurring by fusion of rows 3 + 4 at ventrals 107 on left side/104 on right. Ventrals 174, subcaudals 54, paired; anal plate divided. Supralabials 8, 2nd–3rd touching loreal, 4th–5th in eye; infralabials 8, first four touching anterior genials. Maxillary teeth 16 + 2 on right side.

COLORATION: In life (fig. 74), AMNH R- 140231 was black with a thin, yellow nuchal bar followed by 40 ringlike yellow markings on body—the first 16 of these narrow crossbands being mostly unbroken above, the posterior ones broken dorsally, often with the two halves in alternating sequence. The ringlike crossbands are greenish yellow laterally and a deeper golden yellow dorsally. These markings are narrow, mostly about one-scale wide dorsally; they become wider on the lower sides, most conspicuously on the anterior body where they are about two-scales wide; markings are indistinct atop tail. Ventral and subcaudal surfaces transversely black-banded and checkered on pale greenish white. Some areas of the pale ventral color are transversely continuous with the much narrower yellow crossbands above, but the overall aspect is not of a ‘‘ringed’’ pattern (fig. 75).

Head black, with dull olive-yellow small spots on parietals and adjacent temporal region and with olive suffusions on snout. Supralabials heavily edged in black on light yellow (fig. 76); infralabials pale greenish yellow with black edging; genial plates and gular region white. Iris brown with a small, ill-defined patch of bronze above pupil. Tongue, including tips, black.

DISTRIBUTION AND NATURAL HISTORY

VENEZUELAN SPECIMENS: Some previous assignments of specimens to this taxon are in error (see under Comparisons below). As documented herein, we know of Liophis trebbaui from six specimens —four from Venezuela and two from northern Brazil. The Venezuelan material includes our one specimen taken on the summit of Auyantepui at 1750 m, and the two type specimens collected on its south slope at about 1000– 1100 m (see Comments on Type Specimens and Type Locality). A fourth Venezuelan specimen (AMNH R-104803) comes from Cabanayén, 1230 m elevation, about 70 km southeastward from Auyantepui (5 ° 369N, 61 ° 449W, fide Gorzula and Señaris, 1999: anterior genials. Maxillary teeth 17 + 2 on right side. There is a narrow nuchal ring (a shorter nuchal bar in the Auyán specimen) and 40 pale ringlike markings anterior to the tail (as in the Auyán specimen). Many of the pale markings are medially broken, with the two halves sometimes offset; they are noticeably wider on the lower sides (fig. 77). The labials and parts of the genials are edged with black; pale ventral surfaces are banded and checkered with black (but less heavily marked than in the Auyán specimen).

This specimen was one of a small collection donated to the American Museum by Gorzula in 1977, for collaborative study with then Research Associate Janis Roze. The 246). According to the AMNH catalog, this specimen was collected at Cabanayén by Stefan Gorzula in 1975.

The specimen AMNH R-104803 (fig. 77) is a juvenile female that compares well with the one from the Auyán summit (compare figs. 74 and 76 with fig. 77). It is 200 mm total length, 35 mm tail length (17.5 % of total length); dorsal scales lacking apical pits, smooth, in 17-17-15 rows. Ventrals 170, subcaudals 52, paired; anal plate divided. Supralabials 8, 2nd–3rd touching loreal, 4th– 5th in eye; infralabials 10, first five touching unidentified specimen was assigned a catalog number and tagged, but it was held aside until Roze’s research materials were dispersed into the main collection in the late 1990s. The specimen is listed as Liophis sp. by Gorzula and Señaris (1999: 172, 250), who, however, gave the locality as ‘‘Parupa, CVG Agricultural Station’’, 1220 m at 5 ° 419N, 61 ° 319W. The locality ‘‘Parupa’’ is shown as ‘‘Campo Parupa’’ on Huber and Berry’s topographic map of the Venezuelan Guayana (issued in Steyermark et al., 1995), about 20 km NE Kavanayén. We assume Cabanayén (5Kavanayén) to be correct, since this name was provided for AMNH cataloging, although it makes little difference for present purposes. Cabanayén and Campo Parupa are in reasonable proximity on the Gran Sabana, roughly 70–90 km SE–ESE of Auyantepui (some AMNH specimens collected by Gorzula are catalogued as ‘‘Parupa, nr. Cabanayen’’). Gorzula and Señaris (1999: 246, 250) described Cabanayén as being in a ‘‘mixture of open savannas, scrub, and partially burned forest’’ and Parupa as ‘‘open savanna [and] riverine forest’’.

BRAZILIAN SPECIMENS: Two specimens were collected by Dr. Celso Morato de Carvalho in February 1986, in Roraima, near Vila Pacaraima, border marker BV (Brazil- Venezuela) 8, about 4 ° 299N, 61 ° 099W. Myers visited the locality in June 1993, in company with Drs. Carvalho and W. R. Heyer. The site is a rocky stream east of town, at about 830 m elevation in a very dense and humid evergreen forest. The locality is some 200 km SE Auyantepui, suggesting that Liophis trebbaui may be distributed fairly extensively in the uplands east of the Río Caroní.

Both specimens are juvenile males. Following data are for MZUSP 9235 and 9236, respectively: Total length (tail length) 203 (35) and 222 (42) mm; tail length 17.2 % and 18.9 % of total length. Smooth dorsal scales lacking pits in 17-17-15 rows. Ventrals (subcaudals) 169 (53) and 173 (57); anal plate divided. Supralabials 8, 2nd–3rd touching loreal, 4th– 5th in eye; infralabials 10/10 and 11/10, 1st– 5th touching anterior genials. Maxillary teeth on right side 17 + 2 and 18 + 2. A narrow nuchal ring (as in fig. 77, not a dorsal bar as in fig. 75); 39 and 38 pale ringlike markings anterior to tail, the rings relatively complete in MZUSP 9236, more extensively broken in MZUSP 9235; some rings medially broken with the two halves slightly offset; all rings except first (nuchal ring) noticeably wider on lower sides. Pale markings in temporal region and on frontal plate similar to specimen in figure 77. The supralabials, infralabials, and genials are dark edged, relatively lightly so in MZUSP 9235, blacker in MZUSP 9236. Venter with transverse black bars (narrower than in fig. 75), many broken and offset to give a checkered appearance.

NATURAL HISTORY: Our specimen of Liophis trebbaui seems to be the first one from the Auyán summit. It is a juvenile found during the afternoon, in a shallow, rock-bottom pool containing Tepuihyla tadpoles (fig. 7 bottom). The diet of the similar Liophis torrenicola includes fish at 1000 m on neighboring Cerro Guaiquinima, but fish are absent from the higher summit of Auyantepui (although doubtless present on the lower slopes), where anurans and their tadpoles are the likely food source. No other specimens were seen despite intensive search.

COMMENTS ON TYPE SPECIMENS AND TYPE LOCALITY

The holotype (AMNH R-61019) of Liophis trebbaui is lost. Roze (1958a: 262; 1966: 182) published a photograph of what is inferred to be the adult female holotype. The catalog tag unfortunately appears to have been removed before the specimen was photographed. The specimen was not in the collection of primary type specimens when Myers joined the American Museum staff a decade after its publication, and it has not been found in the main collection. A search of the herpetology archives shows it as an unidentified snake (AMNH R- 61019) on a 1957 invoice, along with specimens later designated as types of Anilius scytale phelpsorum Roze and Neusticurus racenisi Roze , but the Liophis holotype alone was not marked as having been returned, an unfortunate omission that went unnoticed.

Roze (1958a) only gave Auyantepui as type locality, but the AMNH catalog records the holotype as having been collected by W. H. Phelps on February 4, 1938, at an elevation of 1100 m. The 1100 m camp (5 Guayaraca) of the 1937–1938 expedition is the same locality given for the single MBUCV paratype of L. trebbaui , which was collected in 1956 by Pedro Trebbau at 1020 m; Tate (1938a: 474) corrected the elevation to 1038 m and Dunsterville (1965: 168) gave ‘‘nearly 1,000 metres’’ for Guayaraca. The type locality of Liophis trebbaui is essentially the same as that of Neusticurus racenisi (footnote 27) and may reasonably be corrected to: ‘‘south slope Auyantepui, Guayaraca area, about 1000–1100 m’’.

The single paratype (MBUCV 3049) of Liophis trebbaui was said to be a juvenile (size not stated), with a juvenile coloration of transverse bands of black and white (preserved color). The adult female holotype was uniformly olive gray above; total length was given as 570 mm, but the originally stated ‘‘13 mm’’ of tail length (Roze, 1958a) clearly is an error, being much too short (2.3 % of total length) for the number of subcaudals (59) given. Howev- er, Roze’s (1966: 182) subsequent correction to ‘‘130 mm’’ (5 22.8 % of total length) possibly errs in the other direction. Other data provid- ed by Roze (1958a) for the type specimens of L. trebbaui —including ventral and supralabial color patterns, scale counts, method of scale-row reduction, and dentition—seem consistent with the new specimens shown in figures 74–77.

COMPARISONS

The black-edged or dark-suffused supralabials with pale centers separates Liophis trebbaui at a glance from its geographic neighbors L. ingeri , in which the lips are not well differentiated from the greenish head color ( Gorzula, 1992: photo 161; also Gorzula and Señaris, 1999: photo 114), and L. torrenicola , which has a conspicuously pale upper lip (Donnelly and Myers, 1991: fig. 28A; Mägdefrau et al., 1991: 23). 35 Other differences seem to pertain, but specimens are too few to

35 Kornacker (1999: 109, photo 40) included both Liophis torrenicola and L. trebbaui in a checklist, but the color photograph labelled ‘‘ Liophis trebbaui ’’ is misidentified. The photograph (credited to H. Mägdefrau on p. 270) shows a very small specimen of the pale-lipped L. torrenicola from Cerro Guaiquinima, apparently the same juvenile earlier illustrated in Mägdefrau et al. (1991: 23) under the name ‘‘ Liophis cf. longiventris ’’ (an illustration subsequently identified as L. torrenicola in Donnelly and Myers, 1991: 51). satisfactorily delimit ontogenetic and intrapopulational variation of any one character.

The ventrolateral widening of the pale crossbars in juvenile Liophis trebbaui (figs. 74, 77) resembles the triangular widening of the lateral bars in juvenile L. torrenicola (Donnelly and Myers, 1991: fig. 28A)—a resemblance suggestive of possible affinities with Brazilian L. longiventris and L. frenatus from well south of the Amazon ( Dixon, 1983a, 1989). However, there seems to be no question of possible conspecificity. Two juvenile specimens of L. longiventris (AMNH R- 114263–114264) differ from the Venezuelan material in having numerous, very thin, pale dorsal crosslines, with only every other line confluent with the triangular or rounded top of a pale ventrolateral area (similar to fig. 1B in Dixon, 1983a). We have not seen adult longiventris , but an adult male L. frenatus (MZUSP 10160) is slender and remarkably attenuated—body width only 7–8 mm by 560 mm total length—quite unlike the habitus of torrenicola , trebbaui , and most other Liophis .

As previously discussed (Donnelly and Myers, 1991), there is no real evidence that any tepui Liophis is conspecific with Liophis cobella . 36 Dixon (1983a) had submerged L. trebbaui and L. ingeri together as the subspecies Liophis cobella trebbaui , which was said to have the following distribution: ‘‘Known only from the Chimantá and Auyán Tepuís, and km marker 144 of the El Dorado– Santa Elena [de Uairén] Highway, Bolivar, Venezuela’’ ( Dixon, 1983a: 159; 1989: 10). The references to Chimantá and Auyán tepuis refer to Liophis ingeri and L. trebbaui , respectively, whereas the last locality is based on a third species, Liophis miliaris (sensu lato).

NOTES ON LIOPHIS MILIARIS IN VENEZUELA

GRAN SABANA SPECIMENS ( L. MILIARIS

SENSU LATO): The ‘‘km marker 144’’ local-

36 In a useful checklist of Liophis, Dixon (1989) incorrectly changed the name cobella to ‘‘ cobellus ’’, evidently under the assumption that the specific name is a normally declined adjective. However, it is a noun (possibly an Amerindian vernacular name). Linnaeus (1758: 218) named Coluber Cobella in his customary way of capitalizing initial letters of specific names used as nouns in apposition.

ity cited above for ‘‘ Liophis cobella trebbaui ’’ was referenced by Dixon (1983a: 164) to KU 167584, a snake collected by W. E. Duellman in 1974, at an elevation of 1210 m. Although the specimen (fig. 78) is here assigned to Liophis miliaris sensu lato, it may represent an unnamed species. It is a female 695 mm in total length, including 122 mm tail length (17.6 % of total); dorsal scales lacking scale pits, smooth, in 17-17-15 rows; ventrals 176; paired subcaudals 55; anal plate divided; supralabials 8, 2nd–3rd touch loreal, 4th–5th in eye; infralabials 10/9, first five touching anterior genials; maxillary teeth 16 + 2 on right side. In life (Duellman’s fieldnotes and transparency), the dorsum was black with a greenish yellow spot on each scale; venter cream with black markings; iris dark reddish brown; tongue black. The top of the head is black, with dense pale spotting; the supralabials, infralabials, and genials have heavy black margins (fig. 78).

A second, larger miliaris -like snake with a color pattern and scutellation virtually identical to the above is AMNH R-114801, collected by Gorzula in 1975 at Cabanayén (elevation 1230 m fide Gorzula and Señaris, 1999: 246). It is a female 860 mm in total length, 143 mm tail length (16.6 % of total); ventrals 178, subcaudals 52. Head scales as above, except 10 infralabials on both sides; 15 + 2 maxillary teeth. A juvenile specimen of Liophis trebbaui (fig. 77) was collected at the same locality.

With total lengths of 695 and 860 mm, the aforesaid specimens (KU 167584, AMNH R- 114801) here assigned to Liophis miliaris sensu lato are appreciably larger and more muscular than the adult holotypes of Liophis torrenicola (457 mm), L. trebbaui (570 mm), and L. ingeri (630 mm).

Liophis miliaris had not been explicitly recorded from Venezuela when Dixon (1983b) revised the species, although KU 167584 was the source for Hoogmoed’s (1979b: 276) indication of the maximum elevation (1210 m) attained by L. miliaris and its placement in ‘‘Western Guiana’’ (M. S. Hoogmoed, personal commun.).37 Dixon’s (1983a) mistaken assignment of that specimen (KU 167584) to ‘‘ L. cobella trebbaui ’’

37 An explicit Venezuelan record of Liophis miliaris was provided by Kornacker et al. (2002), who reported a specimen (MHNLS 1634) from km 67 on the El Dorado– Santa Elena de Uairén road in southeastern Bolívar state. These authors also corrected an earlier photographic record ( Kornacker, 1995) of a living specimen of L. miliaris that had been misidentified as L. reginae zweifeli .

presumably was based on meristic data, similarity in head color pattern, and knowledge that a pale-spotted color pattern occurs in several nominal species of Liophis . However, L. trebbaui does not have a spotted pattern— Dixon (1983a: 157) seems to have extrapolated this from Roze’s (1958b: 304) description of L. ingeri , which Dixon regard- ed as a synonym of L. trebbaui . Roze mentioned una apariencia reticulada produced by pale areas on the head plates and dorsal scales on the anterior body of L. ingeri . This weakly spotted anterior pattern may prove diagnostic of L. ingeri , but it is not the vivid ‘‘salt and pepper’’ pattern that characterizes the larger Venezuelan specimens here assigned to L. miliaris (e.g., compare fig. 78 with photos of L. ingeri in Donnelly and Myers [1991: fig. 30] and especially in Gorzula and Señaris [1999: photo 114]).

The two specimens of Liophis miliaris s.l. mentioned above come from the northern part of Venezuela’s Gran Sabana (southeastward of Auyantepui). They do not match descriptions of any of the seven subspecies recognized by Dixon (1983b, 1989). Liophis miliaris chrysostomus , of lowland Amazonia, is similar in having a heavily checkered black venter but, unlike the present specimens, the underside of the tail tends to be unmarked (see Dixon, 1983b: fig. 3C). Furthermore the black ventral checkering in chrysostomus is mostly straight-edged with sharp corners (as is commonplace among snakes with such patterns), whereas in the present specimens the heavy markings have mostly nonlinear edges giving a different aspect from normal checkering (compare fig. 78 with fig. 3C in Dixon, 1983b).

OTHER SPECIMENS (L. M. MILIARIS): We are aware of two additional Venezuelan specimens of Liophis miliaris . (1) AMNH R-114755 comes from Santa Elena de Uairén (Estado Bolívar), near the Brazilian border (some 130 km SE of the Cabanayén locality mentioned above). It is a female (preserved in a hardened state and not accurately measurable) that resembles the two Gran Sabana specimens in having dark-edged labials and genials, and a spotted (salt-and-pepper) color pattern, but which differs from them as follows: a lower number of ventrals (ventrals 159, subcaudals 51); head dark brown without pale spots; pale venter inconspicuously marked with thin, pale gray streaks across the ventral plates; subcaudal surfaces unmarked; pale spots on scales of the first dorsal row enlarged to cover most of the scales. The last character causes the pale color of the first-row scales to nearly fuse with the pale venter. In these aspects of color pattern, AMNH R-114755 resembles photographs of the Linnaean holotype of Liophis miliaris ( Gans, 1964, fig. 23 [a composite of 5 photos]). Thus, the AMNH specimen from Santa Elena (misidentified as L. cobella in Gorzula and Señaris, 1999: 170) fits Dixon’s concept of Liophis miliaris miliaris , which occurs nearby in northern Guyana ( Dixon, 1983b: fig. 1).

(2) The second specimen is MHNLS 1634 from km 67 on the El Dorado– Santa Elena de Uairén road (Estado Bolívar); Kornacker et al. (2002: 32) give the coordinates as 06 ° 189N, 61 ° 189W. The locality is some 180 km (airline) N of Santa Elena de Uairén, roughly 130 km ENE of Auyantepui, and nearly 100 km NNE of Cabanayén. It was identified as Liophis miliaris miliaris by Kornacker et al. (2002), who provided a photograph in dorsal view and scale data including 156 ventrals and 52 subcaudals (sex not stated). The numbers of ventrals and subcaudals of MHNLS 1634 from km 67 are close to the 159 and 51 counts given above for AMNH R-114755 from Santa Elena de Uairén. These two specimens resemble Gran Sabana specimens (e.g., fig. 78) in the spotted salt-and-pepper dorsal pattern, but head and ventral patterns are quite different.

The above specimens of L. m. miliaris resemble each other in lacking conspicuous pale spotting atop the head and in ventral pattern. Gilson Rivas Fuenmayor kindly provided a photograph of MHNLS 1634 in preservative that shows an orange-tinged venter with markings virtually the same as those of AMNH R-114755—the venters are marked inconspicuously with thin gray streaks across some ventral plates, the subcaudal surfaces are immaculate, and pale spots on scales of the first dorsal row are enlarged and nearly fused with the ventral color.

TWO KINDS OF LIOPHIS ‘‘ MILIARIS ’’ IN VENEZUELA: Gans (1964) surveyed the variation and synonymy of Liophis miliaris (Linnaeus) in southeastern South America; he suggested that it is a polytypic species but did not formally recognize subspecies. Dixon (1983b) enlarged the study to include northern populations of other miliaris -like snakes. Dixon associated pre-existing names to seven subspecies, of which the two northernmost ones ( miliaris and chrysostomus ) appear to be widely allopatric (map in Dixon, 1983b: fig. 1); see Giraudo et al. (2006) for comments on southern populations. Dixon’s (1983b: 796, 799) conclusion that the type specimen of Linnaeus’ miliaris came from Suriname is particularly noteworthy, since it defined the nominotypical subspecies— L. m. miliaris —as the allopatric population in northern Guyana, Suriname, and French Guiana. The nominotypical form is now known in southeastern Venezuela from Santa Elena de Uairén (AMNH R-114755) and from km 67, about 180 km to the north (MHNLS 1634)— at elevations apparently well below 1000 m.

Liophis miliaris miliaris seems to be replaced at higher elevations (. 1000 m) in the Gran Sabana by an unnamed miliaris -like snake of very different aspect (KU 167584, AMNH R-114801). This form seems to differ from lower-elevation L. m. miliaris in having a higher ventral count, in having a conspicuously spotted head and, especially, in having a distinctively patterned ventral surface (fig. 78). The photograph of a living specimen from ‘‘Kavac, Gran Sabana’’ probably belongs to this population, judged from its vividly patterned head ( Kornacker, 1995 [as L. r. zweifeli ]; Kornacker et al., 2002: fig. 2).

ON LIOPHIS ZWEIFELI : As demonstrated by Kornacker (1995) and Kornacker et al. (2002), specimens of Venezuelan Liophis ‘‘ miliaris ’’ can easily be confused with Liophis zweifeli (Roze) (5 Liophis reginae zweifeli sensu Dixon, 1983c ), a snake with a similar salt-and-pepper dorsal pattern (Roze, 1966: fig. 39); the similarity is real and collections should be checked with that in mind. The taxon zweifeli can be distinguished from Venezuelan ‘‘ miliaris ’’ in having apical pits on the dorsal scales (although these may be few and hard to find), in a lower number of ventrals (140–148 fide Roze [1966: 165], 132– 148 fide Dixon [1983c: 115), and in a longer tail with higher number of subcaudals (78–86 fide Roze [1966: 165], 69–88 fide Dixon [1983c: 127]).

Also, Liophis reginae zweifeli sensu Dixon possibly differs from Venezuelan L. ‘‘ miliaris ’’ in lacking black markings on the underside of head and lower part of the supralabials. The ventral color pattern seems to be variable in zweifeli ; the holotype was described as having ventral scales heavily mottled with black on the posterior two thirds of the body (Roze, 1959: 5), but in an available paratype (AMNH R-59430) the venter is mostly pale, with minimal black edging laterally over most of the venter (subcaudals immaculate). A concise description of the taxon Liophis reginae zweifeli and re-evaluation of its relationships are needed. The extraordinarily variable and wide-ranging Liophis reginae (see map in Dixon, 1983c: 118) is likely a composite species; one species ( Liophis oligolepis Boulenger ) already has been resurrected from its synonymy, as summarized by Cunha and Nascimento (1993: 73–74).