THE HERPETOLOGY OF HISPANIOLA 249 



Relationships. — In order to understand the Hispaniolan form, it is 

 needful to know something of the variation of other Celesti on nearby 

 islands. The Cuban and Puerto Rican forms are unlike costatus in 

 build, resembling Sauresia sepsoides in their weak legs and very elon- 

 gate body. 



There remain the Jamaican forms and the one from Navassa. The 

 two Jamaican Celesti, occiduus and impressus, can scarcely be distin- 

 guished biometrically from each other, as they are practically identi- 

 cal in scalation and in proportion. There is one difference between 

 them, however, that can be reduced to figures — the proportion of 

 width to length of head. In occiduus the head is relatively long for 

 its width, being scarcely swollen in the region of the temples. In 

 impressus the snout seems much shorter in comparison, while the tem- 

 poral region is much heavier, even in the very young. In taking 

 actual measurements I found that the greatest width of the head in 

 occiduus when applied from the posterior border of the ear-opening, 

 reached to the anterior corner of the eye or very slightly beyond. In 

 impressus the same measurement reached over halfway from the 

 anterior corner of the eye to the nostril. 



In coloration there seems to be no difficulty whatsoever in distin- 

 guishing at once the two Jamaican species. As Dr. Barbour once 

 described it, 11 impressus may be told by its "narrow brown bars, 

 from eighteen to twenty in number, which are broken and alternate 

 at the median line." There is little trace of the dark lateral stripe 

 behind the shoulder even in the very young. There is an increased 

 area of dark pigmentation, however, on the side of the neck in front 

 of the shoulder. The other Jamaican species, occiduus, of which Dr. 

 Barbour and I both consider crusculus to be a synonym, has a very 

 distinct dark lateral stripe in all the young and half-grown specimens, 

 while even the large adults have dark patches on the sides, which are 

 prolongations of the uneven and much broken dorsal chevrons. The 

 exact pattern of the back in the adult of this species is subject to 

 great variation as to the width and length of the dorsal markings. 

 One constant feature in the adult is the occurrence of a light longi- 

 tudinal stripe with a dark brown stripe on either border; this begins 

 at the occiput in full, strong colors, continuing backward and gradu- 

 ally fading out somewhere above the arm insertion. The very young 

 are almost without dark pigment on the back; the series of 16 half- 

 grown galliwasps from Arntully shows well the appearance first of 

 the small dark chevron-shaped bars and the eventual concentration 

 of pigment anteriorly at the borders of a central light area, which 

 assumes the form of a fairly unbroken median light stripe with the 

 dark borders not always very sharply divided from the adjacent halves 



11 Bull. Mus. Comp. Zool., vol. 52, No. 15, p. 298, 1910. 

 226849 — 41 17 



