168 MISS JOAN B. PROCTER ON THE 



keel is pi-op.er to L. atrov and the low long keel to L. lanceolatus. 

 Dr. Anrlersson has shown, however, that the latter form is really 

 the Coluber atrox of Linvijeus. so that if the former is maintained 

 as a distinct species on account of its scale-structure, it will be 

 necessaiy to alter the name which has usually been given to it. 

 I suggest that of L. ajjinis Gray, as the specimens of Bothrops 

 affinis (Catalogue oi Snakes (1849), p. 7, specimens / & o), which I 

 have examined, answer the definition, and this appears to be the 

 earliest name which can be applied to it. 



The dorsal scale of the typical L. atrox L. is of a long narrow 

 diamond shape, usually about twice as long as broad, and bearing 

 a simple keel extending to the apical pits in the tip of the scale. 

 That of L. affinis is broader in proportion, more rounded, and 

 sometimes truncate behind ; it bears a short keel, supported upon 

 an extremely convex area, leaving only a naiTOw margin of flat 

 scale. This convexity of the scales is so pronounced in some speci- 

 mens that they can be distinguished from the jJreceding form by 

 touch alone. The scales of other specimens, however, ai-e inter- 

 mediate in type ; they may be of the narrow form with the long 

 keel distinctly swollen at the base, or, while maintaining the long 

 keel of atrox, they may be both broad and convex as in affinis. 

 Other individuals present scales of both the extreme types. On 

 examining the middle third of the body in one of these, I find that 

 the median dorsal scales are of the high short type, and the lateral 

 dorsals of the long low type. Further examination shows that, 

 in every case, the anterior part of the snake is of the atrox 

 type and the posterior of the affinis type. The evolution of the 

 scale-structure is thus clearly shown. The long low-keeled scale 

 gradually broadens, whilst its keel swells along its base and 

 shortens, until it becomes completely transformed into the affinis 

 type. This transformation is, as described above, completed upon 

 the median dorsal scales sooner than upon the laterals. 



As regards the snout of L. jararaca, its shape is very slightly 

 diflferent from that of L. atrox, in that the canthus lostralis is 

 somewhat more obtuse, and the scales of the upper surface are 

 slightly larger than those on the vertex. 



These characters are not very distinct, as several specimens of 

 the typical L. atrox pi'esent snouts of a similar form ; in fact, 

 one specimen (Cat. Sn. iii. p. 536, spec, v) is indistinguishable 

 from L. jararaca (spec, g, Cat. Col. Sn. [1858] p. 226) in this 

 respect. Dumeril & Bibron, and A. E. Brown, both mention 

 this point as one of the distinguishing features between the 

 two forms ; and Schlegel (Phys. Serjj.), in his figure (pi. xix. 

 fig. 1) of the dorsal view of the head of the Jararaca, represents 

 the difference between the scales on the snout and those on the 

 rest of tlie head as very considei'able. This is certainly a point 

 of variation which I cannot consider proper to the Jararaca 

 alone, having found no fewer than six young specimens of 

 L. atrox with this chai^acteristic. 



The different types of form and lepidosis described above are : 



