1906.] IN MEXICAN LIZARDS. 319 



the rank of local races. It so happens that, for instance, the 

 Balsas and the Southern Guerrero groups clifier rather much 

 from each other, but they do so only in the aggregate ; whilst not 

 a single individual combines all the extreme combinations of those 

 very extremes of variations which singly are quite fi'equent in the 

 same group. In other words, each individual still remains within 

 the pale, and it almost appears as if the excursion of one or more 

 characters well beyond the average of another species were care- 

 fully counterbalanced by the most typical behaviour of the rest 

 of the characters. Still, we have in these kaleidoscopic C. depjjei 

 a very fair example of the way in which they might settle into 

 different, definable races, varieties, or subspecies. 



There remains the question, lohether 0. cleppei, as a species, can 

 always he distinguished from other Gnemidophori, especially from 

 small-sized individuals of the striped G. immuiahilis *. Now, it 

 so happens that, although these two kinds overlap absolutely in 

 every one of their characters, I have not found one amongst more 

 than 200 individuals examined about which there could be any 

 doubt. At least, it so happened that the occurrence of exceptional 

 extremes was always counterbalanced by such a typical develop- 

 ment of the remaining features, that uncertainty was set aside. 

 This is of great importance. The two species, with much the 

 same distribution, and structurally so closely allied that they 

 overlap by all their characters taken separately, do not " run into 

 each other." We have to conclude that they ai-e no longer 

 nascent, but well established forms. It seems probable that they 

 have sprung from the same . not remote ancestor, and they now 

 are " specifically " distinct, so much indeed that they now have 

 not only practically the same distribution in Mexico, but that they 

 can live side by side. Then* difference in size is sufficient to 

 exclude interbreeding. It would be a joure assumption that an 

 exceptionally large male G. deppei might pair witli a small 

 G. immutahilis. But supposing that hybrids were possible, such 

 intermediate sj)ecimens have not yet been found. 



I have caught one immature G. immutabilis, length 81 mm., 

 near Rincon (text-fig. 71 E), at the southern foot of the Cajones, 

 which in most structural respects bears a striking resemblance to 

 a young male G. deppei, length 61 mm., from Rio Balsas, where 

 G. deppei alone occurs with G. mexicanus. 



The Rincon specimen reveals itself as G. immutahilis by the 

 following characters : — (1) Collar lead-coloured, rest of under parts 

 pale ; in a male G. dejjjjeioi this size the whole chest and abdomen 

 would be deep blue-black. (2) The centre pair of the eight 

 stripes is broken up into series of short streaks. (3) Humerus 

 with at least 7 rows of large scales. On the other hand, the 

 Rincon specimen has only 7 femoral rows, like the Balsas specimen, 

 and three of these rows extend right down to the knee, an 

 essentially dejjjyei-like character. Both specimens have 3 complete 



* For comparison with C. sexlineatus, see p. 306. 



