Bull. nat. Hist. Mus. (Zool.) 59(1): 1-9 



Issued 24 June 1993 



A new snake from St Lucia, West Indies 



THE N ;L 



HISTC UM 



GARTH UNDERWOOD 



Department of Zoology, The Natural History Museum, Cromwell Road, London SW7 5 Bl 



"9 JUL 1993 



Ph^b'ENTED 



ZOOLOGY LiBRA RV 



Synopsis. The Clelia clelia reported from Dominica and St Lucia are reinvestigated. The specimens concerned are 

 recognised as a new species, Clelia errabunda. It is derived in relation to the mainland species C. bicolor and C. 

 rustica and is the primitive sister species of the mainland C. clelia, C equatoriana and C. scytalina. As such it is 

 interpreted as an island relict. Records from Dominica and Guyana are rejected, the St Lucia records are confirmed. 



INTRODUCTION 



Nearly a hundred years ago Boulenger (1896) reported the 

 South American snake Oxyrhopus clelia, now known as 

 Clelia clelia, from the Windward Islands of Dominica and St 

 Lucia, as well as from mainland localities. Ever since then 

 this species has been listed for these two islands, either in the 

 genus Clelia (Barbour, 1930; Schwartz & Henderson, 1988) 

 or Pseudoboa (Barbour, 1935, 1937). There are other records 

 of this snake for St Lucia but this remains the only record for 

 Dominica. 



Boulenger (1896) distinguished two varieties of "Oxyrho- 

 pus clelia': A, with 19 scale-rows at mid-body, and B, with 17 

 scale-rows. Under variety 'B' Boulenger lists: two specimens 

 from 'The city of Mexico', three specimens from 'W. Ecua- 

 dor' and one specimen each from 'Demerara', 'Dominica' 

 and 'St Lucia'. 



Bailey (1970, in Peters & Orejas-Miranda) gives a key to 

 the mainland species of Clelia. The city of Mexico specimens 

 key out as Clelia scytalina, the Ecuador specimens as C. 

 equatoriana. The remaining three specimens reach the 

 scytalina-equatoriana couplet of the key but are not clearly 

 assignable to either mainland species. Boulenger's variety 'A' 

 specimens key out as C.c. clelia and C.c.plumbea, the two 

 subspecies recognised by Bailey. Clelia was later reported 

 from the island of Grenada. It was described by Greer (1965) 

 as Clelia clelia groomei; Bailey places this in the synonymy of 

 C.c. clelia. 



Further examination of these remaining three specimens 

 shows that they represent an unrecognised species. In the 

 hope of determining the affinities of the new form the 

 specimens of Clelia in the collection of The Natural History 

 Museum, London were examined. A Paris Museum specimen 

 from St Lucia was also examined. The seven species recogn- 

 ised by Bailey are represented. It should be noted that 

 Scrocchi and Vihas (1990) have now sunk occipitolutea in the 

 synonymy of clelia. 



MATERIALS AND METHODS 



Each specimen was examined in respect of external features, 

 anterior viscera, maxillary teeth and, in the case of males, 

 hemipenis. For a representative of each species dissections 

 were made to show the superficial jaw muscles, ligaments and 



labial glands and the upper and lower jaws. 



The ventral scale count is according to Dowling (1951). 

 The notation for the scale-rows on the trunk allows a detailed 

 record in a single line of type using characters on a word 

 processor keyboard. An oblique transverse scale-row is iden- 

 tified by the ventral scale from which it passes backwards, as 

 shown by M.A. Smith (1943, fig. 10). Where there is a 

 scale-row fusion the upper of the two merged rows is indi- 

 cated. For example male specimen BMNH 89.4.8.2 is 

 recorded as: 



V 210 >>5> 17 (36;5</33;5<) 19 (145;4>/142;4>) 17 



Ventral scale count 210. Three scale-row fusions from the 

 back of the head onto the neck, the third recognisable as due 

 to the fusion of row 5 with the row below. Seventeen rows on 

 the neck through oblique row 36 left/33 right; row 5 splits 

 giving rise to 19 rows through oblique row 145 left/142 right; 

 row 4 merges with row below giving 17 rows. Most specimens 

 were recorded down the left-hand side only (Table 1, see 

 Appendix). 



The subcaudal count starts with the first scale on the left 

 side which makes full contact with a scale obliquely forward 

 and opposite. The dorsal scale-rows on the tail are somewhat 

 irregular around the base but settle down to even row 

 stretches: 8, 6, 4, 2. The fusions are rarely symmetrical so that 

 there are short odd number transitions between the even 

 stretches. Each transition is included in the preceding, 

 higher, even row stretch. The transverse rows are identified 

 by the subcaudal scale-row pairs from which they arise 

 obliquely. The tail of the above specimen was recorded as: 



79 prs, >8 (10) 8 (25) 6 (45) 4 (73) 2 



More than eight longitudinal rows through oblique row ten, 

 eight (and seven) through row 25, six (and five) through row 

 45, four (and three) through 73 and two rows to the terminal 

 scute. From this record the lengths of the five stretches are 

 calculated as the basis of comparison of the individual speci- 

 mens. The above becomes: 



C79 >8:10 8:15 6:20 4:28 2:6 



This feature shows marked sexual dimorphism. 



For selected specimens scales were mounted on slides to be 

 examined for the presence of pits and tubercles (Underwood, 

 1963). The wet scales are laid on a dried film of polyvinyl 

 alcohol lactophenol mounting medium. When dry the slide is 

 then covered using Canada balsam. For each specimen the 



