Dr. 0. Wucherer on the Craspedocephali of Bahia. 243 



of separating the species of this genus. His diagnoses do not agree 

 exactly with those of Schlegel in his * Essai,' nor with those of 

 Dumeril and Bibron in their ' Erpetologie Generale/ I may therefore 

 be excused if I offer the following remarks on my specimens. In 

 my former paper I stated that I had neither seen Ci'aspedocephalus 

 laneeolatus nor C. brasiliensis. At the present time I have examined 

 very nearly forty specimens of " Jararaca," all of which, except three, 

 agree sufficiently in every character, and are, according to the de- 

 scriptions of herpetologists, referable to C. atrox. These three spe- 

 cimens show certain slight differences which justify a doubt of their 

 specific identity with the others. 



Dr. Gray mentions C. atrox as having seven upper labial shields. 

 Schlegel, in his ' Essai,' i. p. 189, and again ii. p. 535, describes this 

 species as having eight labial shields ; still this may perhaps be con- 

 sidered a mistake, for in his plate 19 of the above work G. atrox is 

 represented as having only seven upper labial shields. Dumeril and 

 Bibron make no allusion to this character in G. atrox. Now all the 

 specimens of G. atrox which I have had occasion to examine have 

 seven upper labial shields. Only one has on one side eight, which 

 must be considered an irregularity. 



Dr. Gray describes C. brasiliensis as having nine or ten upper 

 labial shields, the hinder ones of which are smaller ; Schlegel de- 

 cribes it as having nine ; and Dumeril and Bibron do not mention 

 the number of labial shields at all. 



The three specimens differing from those of G. atrox mentioned 

 above have all eight upper labial shields on each side, the last one 

 narrower than the last one in G. atrox. 



A statement I made in my former paper, that ray specimens of G. 

 atrox differed from those described by herpetologists in having fewer 

 longitudinal rows of scales, I now take the opportunity to rectify. 

 The number of longitudinal rows of scales in the species of this genus 

 is not always mentioned as a specific character, and indeed it does not 

 appear very serviceable as such. Schlegel's G. jararaca, the C. bra- 

 siliensis of Dr. Gray's catalogue, has twenty-seven rows of scales ; 

 of G. atrox he says (Essai, ii. p. 536), " On compte quelquefois 29 

 rangees d'ecailles," leaving it perhaps hence to be inferred that it has 

 generally a lesser number, or twenty-seven, like the one just de- 

 scribed, which is G. brasiliensis. Dumeril and Bibron (vii. p. 1509 

 and p. 1511) give to G. atrox from twenty-nine to thirty-two, to 

 G. brasiliensis twenty-seven rows. All my specimens of G. atrox, 

 with few exceptions, have twenty-seven rows of scales, a few having 

 twenty-five. Of the three specimens differing from them, two have 

 twenty-five and one twenty-three rows of scales. 



Schlegel and Dumeril and Bibron draw some specific differences 

 from the shape of the head, the former saying (ii. p. 535) that the 

 snout of G. atrox is more conical, by which I suppose is meant more 

 rounded, Dumeril and Bibron stating that the sharp edge on the 

 anterior part of the head is almost effaced, and does not reach back 

 to the orbits, furthermore that the scales on the anterior part of the 

 head are comparativeh- much larger than on the posterior part xsx C. 



