PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON 109(4):749-754. 1996 The systematic status of Guatemalan populations of snakes allied with Ninia maculata (Reptilia: Colubridae) Brian E. Smith and Jonathan A. Campbell Department of Biology, University of Texas at Arlington, Arlington, Texas 76019-0498, U.S A. (Current address for BES) U.S. National Biological Service, Northern Prairie Science Center, 8711 37th St. SE, Jamestown, North Dakota 58401-7317, U.S.A. Abstract. — The Guatemalan colubrid snake, formerly recognized as Ninia maculata pavimentata, is elevated to a full species. This species is closely related to A^. maculata of southern Central America, but differs in segmental counts, relative tail length, various measures of head shape, and ventral col- oration. Three recently collected specimens of A^. pavimentata represent sig- nificant range extensions of this snake in Guatemala. Ninia pavimentata is re-described and ecological notes are provided. Resumen. — El coWbrido guatemalteco conocido como Ninia maculata pav- imentata es elevado al nivel de especie. Esta especie se encuentra cercanamente relacionada a Ninia maculata, del sur de Centro America, pero difiere de esta en numero de segmentos, largo relativo de la cola, medidas cefalicas y color- acion ventral. Tres especimenes de Ninia pavimentata recientemente colectados representan extenciones de rango significativas para esta serpiente en Guate- mala. Se redescribe Ninia pavimentata y se provee datos ecologicos. The apparent distribution of the colubrid snake known as Ninia maculata has been an enigma to biogeographers. Peters (1861) originally described A^. maculata from a single specimen taken in Costa Rica. Bo- court (1883) noted similarities between four specimens from Alta Verapaz, Guatemala, and snakes from Costa Rica, and described the Guatemalan specimens as a new sub- species, N. maculata pavimentata. In his re- vision of the genus, Dunn (1935) did not consider the taxonomic position of this sub- species, and apparently never examined the types. Stuart (1948) noted the large hiatus between the provenance of A^. maculata pavimentata in Alta Verapaz and the north- ernmost record of A^. m. maculata, given by Dunn (1935) as Jinotega, Nicaragua. Stuart (1948) predicted that intergrades would eventually be collected in the intervening area. Savage & Lahanas (1991) noted that no such specimens had been forthcoming, and they anticipated that A^. m. pavimentata would eventually be elevated to full specific status. The senior author, as part of an ongoing study of the genus Ninia, has examined 14 specimens of TV. m. pavimentata housed in collections, including several specimens from Guatemala collected far from the type- locality. It has become clear that these spec- imens share certain unique similarities with each other, but differ from the more south- em populations, indicating that the northern population deserves full species status. Materials and Methods Material from the Museum National d'Histoire Naturelle (MNHNP), Carnegie Museum (CM), Los Angeles County Mu- seum (LACM), Museum of Vertebrate Zo- ology (MVZ), University of Michigan Mu- seum of Zoology (UMMZ), Florida State Museum (UP), the University of Texas at Arlington (UTA), and the Universidad del Valle of Guatemala (UVG) was examined. 750 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASfflNGTON Standard scale counts and measurements of the head and body were taken on all specimens. Counts of paired characters were made on both sides of the body. The following counts were made: number of dorsal scale rows one head-length behind the angle of the jaw, at mid-body, and one head length anterior to the anal plate; num- ber of ventrals (method of Dowling 1951); number of subcaudals (counted on the left side, beginning with the first subcaudal in contact with an opposite subcaudal, the "best" method of Peters 1964); the number and position of supralabials, infralabials, supraoculars, loreals, preoculars, postocu- lars, anterior temporals, posterior tempor- als, chin tubercles (males only); and dark bands on the left side of the body. Ventral color pattern was scored using the catego- ries of Savage & Lahanas (1991:43, their figure 4). Several measurements of the head were made to the nearest 0.1 mm using a vernier caliper, and the following variables were normally distributed: head length from the front face of the rostral to the angle of the jaw on the left side of the head, head width at the level of the angle of the jaw, snout length from the tip of the snout to the anterior margin of the orbit, head width at the widest point (immediately behind the orbit), length of the prefrontal suture, length of the parietal suture, and length and width of the frontal scale. Three juvenile Ninia pavimentata were used in the species de- scription, but were not used in statistical analyses. Other head scale measurements used in the description of head scale shape included rostral height-width and loreal length-height. Measurements of paired ce- phalic scales were taken on the left side of the head only. Measurements of snout-vent length and tail length were made with a standard metric ruler to the nearest 1 mm. To assess the differences between the Costa Rican and Guatemalan populations, a subset of morphometric measurements were entered into a multivariate analysis of co- variance (MANCOVA), using snout-vent length as a covariate. Variables that were found to be not normally distributed were dropped from this analysis, as noted above. Ventral and subcaudal scale counts are sex- ually dimorphic in species of Ninia (Savage & Lahanas 1991), with slightly higher ven- tral counts in females, and slightly higher subcaudal counts in males. Ventral and sub- caudal counts added together form the seg- mental count, a variable that was used in a study of Costa Rican N. maculata (Savage & Lahanas 1991) as a non-dimorphic mea- sure of vertebral number. In our study, we used segmental counts in a Mann- Whitney U test to compare populations of Ninia ma- culata (sensu lato). Our analysis indicates that Ninia pavimentata is distinctive from A^. maculata and we suggest that these taxa are specifically distinct. Table 1 provides the range of variation for standard scale counts in N. pavimentata. Statistical analy- ses were performed using Statistica version 3.1 (StatSoft, Inc. 1991). Results and Discussion Ninia pavimentata is a small colubrid snake from Guatemala (snout-vent length of largest male, 265 mm; largest female, 275 mm). It is characterized by a head slightly distinct from the neck; dorsal scales keeled and heavily striated, in 19 rows throughout; ventrals 142-147 in males, 136—147 in females (except as noted for UVG 382); and subcaudals 68-76 in males, 63-75 in females (except as noted for UVG 382). See Table 1 for standard scale counts. The subcaudals are paired and the anal plate is single. Insufficient data exist to document sexual dimorphism in ventral and subcaudal counts of A^. pavimentata. The head scala- tion follows a typical colubrid pattern of nine large, platelike scales. The rostral is 1 .35-2 times broader than high, and the in- temasals are paired, about half the size of the paired prefrontals. The frontal suture is 1.5-1.85 times as long as the prefrontal su- ture, except for one individual (MVZ 159860) in which the frontal scale is only slightly longer than the prefrontal suture. VOLUME 109, NUMBER 4 751 Table 1. — Selected features of Ninia pavimentata (* tail incomplete, # chinshields dessicated, chin tubercles observed but not counted). Museum no. Sex Snout-vent length (mm) Tail length (mm) Ventrals Subcaudals Segmental count Chin tubercles UTA R-7099 S 225 83 144 68 214 91 MNHNP 1192 S 226 82 147 70 217 # MNHNP 1994.226 3 209 83 142 74 216 63 MVZ 159860 6 222 95 143 76 219 79 MVZ 159861 S 265 103 142 75 217 75 UP 96321 S 106 40 141 74 215 — UVG382 $ 215 62 130 49 179 — MNHNP 1994.227 9 275 101* 145 66* 211* — MNHNP 1994.228 9 195 71 144 70 214 — UMMZ 89083 9 255 69* 144 49* 193* — CM 43954 9 203 75 147 75 222 — UTA R-37590 9 266 81* 136 53* 189* — LACM 40055 9 154 49 142 63 205 — There is a single frontal, about as long as wide, in all specimens. The suture between the paired parietals is about as long as the frontal scale in all specimens examined. There is a single supraocular, a pre- and postnasal, and the loreal scale is 1.5-2 times as long as high, except for three in- dividuals (LACM 40055, MNHNP 1192, and UMMZ 89083) in which the loreal scale is nearly square. Stuart's (1948) state- ment that the loreal is distinctively high and short appears to be relative, based on com- parisons with other species with horizontal- ly elongate loreals. There is no preocular (in MVZ 159860, one small scale is present on the left side and two small scales on the right side; on both sides the loreal still con- tacts the eye), two postoculars (rarely three or, more rarely, one), one anterior temporal, and two posterior temporals. Supralabials and infralabials usually 7 (rarely 6), with supralabials 3 and 4 in contact with the eye except for UF 96321, in which only supra- labial 3 contacts the eye on the left side, and only supralabial 4 contacts the eye on the right side. The eye is moderately large, usually just slightly smaller in diameter than the loreal length. The mental is about twice as wide as long, and the first pair of infralabials are in contact, separating the mental and anterior chinshields. The first four infralabials are in contact with anterior chinshields (except in LACM 40055, in which the third and fourth infralabials are fused on the right side). The fourth infra- labial is in contact with the posterior chin- shields (except as noted above for LACM 40055). The anterior chin shields are 1.4- 1.8 times longer than the posterior chin- shields. Males have prominent raised chin tubercles on both chinshields and the first four pairs of infralabials. There are 63-91 chin tubercles on the anterior chinshields and first pair of infralabials. In alcohol the dorsal ground color is usu- ally reddish brown, with the exception of UTA R-37590, which is badly formalin- blackened, obfuscating any dark dorsal crossbands. The other specimens we ex- amined have 34-50 well-defined black crossbands. Approximately two-thirds of these crossbands are continuous across the dorsum, others end at the mid-dorsal scale row. The anterior band is always continu- ous across the dorsum. Immediately in front of this band is a pale nuchal collar, indis- tinct and narrow, which is usually no wider than one or two scale rows at mid-dorsum, and slightly wider laterally. In seven of 12 snakes this collar is not continuous across the dorsum. The supralabial and infralabial scales are cream colored, with black edges. The distinctive checkerboard ventral pattern is diagnostic for the species. Costa Rican 752 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASfflNGTON N. maculata can be similar (see Savage & Lahanas 1991: their fig. 4G and 4H, for il- lustrations of venters similar to A'', pavimen- tata), but Guatemalan specimens are much more boldly marked than any Costa Rican specimens that we have examined. Ninia pavimentata can be distinguished from its close relative N. maculata by hav- ing a broader, blunter, more spatulate head. The upper lips of TV. pavimentata are some- what flared compared to A^. maculata, and are easily seen in dorsal view. Additionally, the dorsal pre-orbital surface of the head of N. pavimentata is distinctively convex, while this surface in A^. maculata is nearly flat. The snout of A^. pavimentata is fairly blunt, whereas that of N. maculata is more angular. The statistical analyses, while not robust with such small sample sizes, assist in in- dicating the nature of the differences be- tween the two species. The MANCOVA re- sults separate the species using the follow- ing variables: Ninia pavimentata has a rel- atively longer tail than A^. maculata if = 29.17, P = 0.000039, df = 1,18); A^. pavi- mentata has a narrower head at the level of the angle of the jaw (f = 5.55, P = 0.030, df = 1,18), probably reflective of the spat- ulate shape of the head of N. pavimentata; and A^. pavimentata has relatively larger pa- rietal scales (f = 4.59, P = 0.046, df = 1,18). In addition, A^. pavimentata has a sig- nificantly higher segmental count than N. maculata (Z-adj = -2.97, P = 0.0030). There is little overlap between segmental counts of A^. pavimentata and N. maculata (Table 1). Three snakes warrant special mention. Two of these are significant extensions of the range of Ninia pavimentata, which had previously been known only from the high- lands of Alta Verapaz and adjacent Baja Verapaz. One specimen (UTA R-37590) was recently collected in the province of San Marcos, the first specimen of A^. pavi- mentata to be collected on the Pacific ver- sant of Guatemala. Examination of this specimen reveals that it is dark (probably formalin-blackened) but otherwise it agrees with A^. pavimentata. Another specimen (LACM 40055) from near Barillas in the Sierra Cuchumatanes, Huehuetenango, is fairly typical of the species. Finally, a specimen from near La Fran- cia, Izabal, near the Honduran border (UVG 382) is unusual not only because it is the only Ninia pavimentata known from the lowlands of Guatemala, but also because it shows a combination of features typical of either A^. maculata or A^. pavimentata. This snake has a smooth and level surface on the dorsal pre-orbital portion of the head, sim- ilar to A^. maculata, but it has large parietal scales typical of A^. pavimentata. The snout is fairly blunt like A^. pavimentata, but ven- tral and subcaudal counts are more similar to N. maculata. This specimen apparently represents an isolated population restricted to the lowlands of eastern Guatemala and perhaps adjacent Honduras. Despite its morphology, we do not consider it an "in- tergrade" in the sense of Stuart (1948). This population is separated from the most geo- graphically proximate populations of A^. maculata in Costa Rica by about 600 km and may have become isolated relatively soon after the vicariant event that separated A^. maculata and A^. pavimentata. The area in which this specimen was collected is not well known, and further specimens from this area are needed to fully resolve the sta- tus of this snake. We tentatively allocate this specimen to A^. pavimentata, but rec- ognize that it may ultimately prove to be distinct. With respect to the ecology and habits of Ninia pavimentata, very little is known. Most specimens have come from pine-oak or cloud forest at elevations of 1120 to 1825 meters (Fig. 1). A single lowland specimen was collected below 300 m. Ninia pavimentata is rarely encountered, either because it is rare or because of secretive behavior. Specimens have been found most frequently in the highlands of Alta Verapaz (with one specimen from nearby Baja Vera- paz), but recent specimens collected in the VOLUME 109, NUMBER 4 753 HONDURAS meters 500-999 KOO-1999 EL SALVADOR kms Fig. 1. Localities for Ninia pavimentata in Guatemala. departments of San Marcos and Huehueten- ango indicate that this species ranges more widely throughout the Guatemalan high- lands than previously thought. Material examined of Ninia pavimenta- ta. — Guatemala: Alta Verapaz: Municipio de San Cristobal Verapaz, Baleu, 1350 m (CM 43954); 11 km W San Cristobal Ver- apaz, 1120 m (UF 96321); Finca Chichen, 1410 m (UMMZ 89083); Finca Volcan, 25 km by road NW Senahu, 1300 m (MVZ 159860, 159861); no specific locality (MNHNP 1192, 1994.226, 1994.227, and 1994.228 — syntypes). Baja Verapaz: Cerro Verde, ca. 1500 m (UTA R-7099). Huehue- tenango: Sierra de los Cuchumatanes, Ba- rillas, 1700 m (LACM 40055). Izabal: Fin- ca Santa Isabel, near La Francia (UVG 382). San Marcos: Aldea La Fraternidad, Finca La Esperanza, 1825 m (UTA R-37590). Acknowledgments The following individuals graciously loaned specimens in their care: Ellen Cen- sky, Carnegie Museum of Natural History; David Auth, Florida State Museum; Robert Bezy, Natural History Museum of Los An- geles County; Ivan Ineich, Museum Nation- al d'Histoire Naturelle; Harry Greene, Mu- seum of Vertebrate Zoology; and Arnold Kluge, University of Michigan Museum of Zoology; and Michael Dix, Universidad del Valle of Guatemala. This paper is based in part upon work supported by the Texas Ad- vanced Research Program under Grant No. 003656-001. 754 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASfflNGTON Literature Cited demie der Wissenschaften zu Berlin 186L922- 925. Bocourt, M. E 1883. Mission Scientifique au Mexique P^^^^s, J. A. 1964. Dictionary of herpetology. Hafner et dans TAmerique Centrale.— Livraison 9: Publishing Co., New York, 392 pp. 529-592 pis 31-35 Savage, J. M., & P. N. Lahanas. 1991. On the species Dowling, H. G. 1951. A proposed standard system of of the colubrid snake genus Ninia in Costa Rica counting ventrals in snakes.— British Journal of and western Panama.— Herpetologica 47:37- Herpetology 1:97-99. 53. Dunn, E. R. 1935. The snakes of the genus Ninia.— Statsoft. 1991. Statistica. Version 3.1. Statsoft, Inc., Proceedings of the National Academy of Sci- Tulsa Oklahoma. ences 21:9-12. Stuart, L. C. 1948. The amphibians and reptiles of Peters, W. 1862 [dated 1861]. Mittheilung uber neue Alta Verapaz, Guatemala. — Miscellaneous Pub- schlangen des koniglichen zoologischen muse- lications Museum of Zoology, University of ums. — Monatsberichte der Preussischen Aka- Michigan, No. 69:1-109.