Redescription and validation of Bothriechis supraciliaris (Serpentes: Viperidae)

The populations of pitvipers from south western Costa Rica, have traditionally been identified as Bothriechis schlegelii (Berthold). However, in 1954 E. H. Taylor described one specimen from the area as a new subspecies, B. schlegelii supraciliaris. Werman returned supraciliaris to synonymy with schlegelii four decades later. However, morphometry and color pattern in a SW Costa Rica population (25 specimens) differ from those of specimens (N=57) from other parts of Costa Rica and from descriptions of South American specimens. Here the epithet Bothriechis schlegelii supraciliaris Taylor 1954. is reestablished as a valid taxon and elevated to specific rank as B. supraciliaris stat.nov. It is closely related to B. schlegelii from which it differs by its color patterns based on a uniform ground color with polymorphic dorsal designs, and its lower counts of ventral and caudal scales.

almost always uniform and lacks secondary pigments, while the dorsal ground coloration of specimens from elsewhere in the country is highlighted by a variety of dots, spots or bars of secondary pigmentation ( Figs. 1 and 3). Furthermore, the "oropel" morph of typical B. schlegelii, yellowish to golden orange, so well known from other Costa Rican sites is extremely rare in the General-Coto Brus area, where only one specimen (immature male) was seen in ten years of observations by one of the authors (L.D.G.).
Ventral coloration outside Coto Brus is equally enriched by scattered dots and spots, while in specimens from Coto Brus the ventral aspect is devoid of pigments at least in the anterior two thirds of the body length (Fig. 3), the caudal portion rarely has diffuse, dark, purplish brown hues mostly visible in preserved material.
Lepidosis: Specimens from Coto Brus and those from elsewhere in Costa Rica differ in number of ventral and caudal scales (Fig. 4). There are no significant differences in length and number of dorsal scales (Fig. 4).  Based on these results, we conclude that Bothriechis supraciliaris warrants specific recognition and it is here redescribed: Bothriechis supraciliaris (Taylor 1954), Fig. 1.

Diagnosis:
Closely related to B. schlegelii from which it differs by its color patterns based on a uniform ground color with polymorphic, dorsal designs and its lower counts of ventral and caudal scales.
Description: Pitviper of small to moderate site (maximum 600 mm), prehensile tail and strongly triangular outline of head and, from Taylor's original description: "Front of snout rounding, covered by about 70 strongly keeled scales anterior to level of the supraoculars; rostral approximately as wide as high, not visible above: on canthus a tiny median scale above rostral; each undivided nasal bordered above by three scales with free outer edges; loreal separated from the upper preocular by one small scale; these three scales bordered above by five scales, three of which have the keels strongly elevated into soft, flattened "spines" and forming a line with the two elevated supraorbital spines: the supraoculars divided, each bearing one or two small, spine-like outer projections: supraoculars separated from the eyes by a row of ten small granular scales: the next row consists of about five scales, two of which are the elevated supraorbital spines, this row separated from orbit by eight or nine small granules, two or three small postoculars: a long narrow subocular runs under orbit to or very close to the lower preocular: three preoculars, the median and lower border pit: second labial separated from lower preocular by small scale, broadly entering pit and with a scale bordering it above forms the anterior border of pit: the second labial has no partial suture entering from its anterior edge; three to five small lower loreals (prefoveals); 9-10 supralabials, the first continuously bordering the nasal: the third, fourth, and sixth separated from the subocular by three scale rows; infralabials 12-12, only two of which touch the first chinshields; first pair of chinshields much larger than the three following pairs, but distinctly smaller but perhaps a little wider than the first pair of labials; temporal scales keeled strongly, the keels compressed and elevated; about 56 scales across head at angle of jaw; no scales between nasal and rostral. Scale formula: 25, 23, 23, 23, 19, 19; ventrals (counting from first widened scale), 146, caudals 46, anal single; the terminal scute of tail rather large; the tail a tip with six rows of scales; 15 in the postanal region; all scales except outer row, and scales on underside of head, keeled." Our scale counts of the holotype agree with Taylor's, except that the caudals are 47. There is a marked sexual dimorphism: the females are longer and thicker (pers. observ. of 14 females and 4 males) as in B. schlegelii. Specimens collected in resting positions were found curled in a loose circle and mostly on horizontal, wide surfaces, while typical schlegelii often droop and loop on forks of stems and branches or form a tight S-shaped pattern on almost vertical, rough-barked tree trunks.

DISCUSSION
Among neotropical pitvipers Bothriechis schlegelii is known particularly for its impressive color variation throughout its geographical range (Werman 1984, Campbell & Lamar 1989. Its usual habitat is lowland rainforests although it is also known from moderate to high elevations (Campbell & Lamar 1989, Crother et a!. 1992. In Costa Rica B. schlegelii is amply distributed in the Caribbean and southern Pacific lowlands, but the Pacific range is geographicaly isolated from the rest of the country (Crother et al.1992).
Recently, the systematics of the genera Bothrops and Bothriechis has been the subject of diverse morphological and biochemical studies (Campbell & Lamar 1989, Shätti et al. 1990, & Kramer 1991, Cadle 1992, Crother et al. 1992, Werman 1992, Kuch and Freire 1995, Wüster and McCarthy 1996. Taylor's (1954) keen observation of a single specimen (Type, KU. 31997) was notable and gross differences shown by the material at hand were correct, although the main parameters used for his description were not, as he based the distinction on the aspect and morphology of some scales that often are ontogenetically variable within the species, such as the divided supraoculars and the color of the tail tip.
This study shows that, despite normal variation, specimens from Coto Brus display a notable and constant coloration divergence from the rest of the country.
Coloration of the Costa Rican taxon is equally and strongly related to some morphotypes from Colombia and Ecuador as described and illustrated in Campbell & Lamar (1989). The close resemblance and affinity of the SW Costa Rican Pacific versant supraciliaris to the NW South American schlegelii seems to support the hypothesis of an ancestral continuum of a schlegelii stock, a hypothesis further enhanced by the existence of other cases (Solórzano 1995), such as Porthidium volcanicum Solórzano and its vicar P kinsbergii, or by Lachesis muta melanocephala Solórzano & Cerdas (1986) whose color pattern and scaling is closer to the South American L muta muta than to L muta stenophrys from the Atlantic versant of Costa Rica. In contrast, Zamudio and Greene (1997) defend a greater affinity between this and stenophiys.
Unlike Werman (1984), we used a larger sample that has evidenced significant differences in the number of ventral and caudal scales. Allozyme comparisons of five individuals from Coto Brus, one from Upala in NE Costa Rica, and one from Ciudad Neily, 35 km SW of Coto Brus, support the differentiation (Crother, unpublished).
The geographical range of supraciliaris seems to be confined to mid and higher elevations (800-1700 m) (Appendix 1). For schlegelii, the range is ampler but mostly confined to lowland rainforests outside and within SW Costa Rica. To date, we have found no overlap of the ranges of the two taxa. There is a closer similarity in squamation between Costa Rican supraciliaris and Colombian schlegelii (see Renjifo 1979 for summary statistics of Colombian populations).
The species redescribed here is thus far only known from the Valle del General and the Coto Brus altiplano in SW Costa Rica, and its existence suggests that an exhaustive revision of the systematics of Bothriechis schlegelii throughout its range, from Northern Penú to Mexico, would likely reveal similarly isolated and discrete genotypes.

AKNOWLEDGMENTS
We are grateful to William W. Lamar, Mahmoud Sasa, José M. Gutiérrez and two anonymous reviewers for suggestions to greatly improve this manuscript, and to Alberto Leon S. Peters iA. & B. Orejas-Miranda. 1970. Catalogue of the for his drawings. Michael and Patricia Fogden supplied some photographs. The organization for Tropical Studies, provided logistical support for the two senior authors.