310 DR. H. GADOW OX EVOLUi'ION [Mav. 20, 



thumb procedure, applied to 86 specimens, gives the following- 

 result : — 



This tabulation is naturally very imperfect, not to say ci-ude, 

 but the results will not be altered Avhen we throw the intermediate 

 or doubtful cases of 6 to 7 rows either into the column with 6 or 

 into that with 7 complete rows. There were, for instance, amongst 

 the Oaxaca specimens four with 5 or 6 rows, but amongst those 

 of Guerrero only two such doubtful cases. The fact remains, that, 

 if we take Tierra Colorada as norma], those further south tend to 

 possess fewer femoral scales, whilst those further north, at Rio 

 Balsas, have decidedly more scales. This does not seem to be the 

 result of mere chance, because, as we shall see, the Balsas speci- 

 mens are remarkable for othei- features, although they are not 

 peculiar to them. Amongst the specimens from the State of 

 Oaxaca no difference whatever caii be discerned with reference to 

 their habitat nearer to or further away from the coast. 



At any rate the amplitude is considerable, the extremes being 

 two specimens with an imperfect 5th row, and one with 8 rows ; 

 the latter specimen, from Tequesixtlan, is in every other respect 

 a typical C. de-ppei. 



Inner side of tibia protected by 2, 2|, or 3 complete rows of 

 enlarged scutes. 



Preanal regioti. — As a i-ule there are 3 laige scales which form 

 a triangle, with two at the base and a third of ecjual size on the 

 top, but sometimes a fourth scale forms the apex ; or the two 

 basal scales are small, whilst the third is correspondingly enlarged. 

 In the majority the whole cluster is separated from the ventrals 

 by a very short isthmus, sometimes so short that tlie cluster and 

 the vetiti-als almost touch each other. The isthmus itself is mostly 

 covered with very small scales, and when these are larger the}' 

 are fewer in numbers. The whole character is worthless. 



Femoral pores. — The number of pores frequently differs on the 

 right and left side by one, rarely by two, in the same individuaL 

 There is no preference for one side. The smaller number in these 

 cases of asymmetry is always due to one or more pores remaining 

 undeveloped at the proximal, or more frequently at the distal end 

 of the series. For comparison, the higher number should there- 

 fore always be taken. The amplitude of individual variation is 

 considerable, wdien we recollect that these pores correspond in 

 numbers with transverse rows of scales. But unfortunately the 



