INTRODUCTION. 219 



actually but proportionately and are all comparatively rare. It must not be sup- 

 posed that these species are in any wise ancestral the one to the other; but are 

 simply examples chosen as typifying arrested stages pointing out the path by 

 which such a form as anthracinus, for instance, may have been derived from one 

 like cinerexis at first sight apparently wholly unrelated. The apparent dis- 

 similarity is bridged over by a series of hypothetical ancestral intergrades which 

 happen tobe illustrated by various scattered species. In anthracinus (syn. asper) , 

 copei fsyn. picturatus) and scaber there is a distinct sexual dichromatism,the males 

 being monochrome, the females fantastically banded and decorated with ocellate 

 spots. How^ever, a similar dichromatism exists in decoratus and torrei among 

 the species having granular scales. 



A condition similar to this has been independently arrived at in some 

 species of Gonatodes, far from closely allied. This apparently is not true of the 

 Jamaican richardsonii for I have traced eight specimens, all banded and of this 

 number both sexes were probably represented. Structurally this species is 

 closely related to the three others but is more primitive in colouration, probably, 

 and in lacking the middorsal zone of granules. This curious character has not 

 been mentioned previously. It is found that in several, indeed, in most of the 

 species with very large scales, there is a narrow zone of fine granular scales 

 running down the middorsal region. This probably has been at least three 

 times independently acquired in that the zone is rather wide and not very 

 sharply defined in lincenti (St. Vincent) and macrolepis (St. Lucia) ; while festus 

 (Martinique) shows no sign of a zone. The species from the Upper Lesser 

 Antilles, pidus (St. Kitts), sputator (St. Eustatius) and elegantulus (Antigua) 

 have a very narrow zone, sharply defined although the body-scales are not so 

 very large. In Porto Rico, Mona, and the Virgin Islands (macrolepis), the local 

 species although large scaled, has no middorsal zone. On the other hand, in 

 Cuba, Haiti, and Andros Island there are very large species with very large 

 scales, viz. scaber, copei, and anthracinus respectively which have the middorsal 

 granular zone well defined. Thus it may be observed that the character appears 

 in three separated groups of species. No explanation has been offered suggesting 

 why this character should be acquired. It seems reasonable, however, that it 

 may have to do with the flexibility of the body. The large overlapping scales 

 tend to form a sort of cumbrous armour and the expansion and contraction of the 

 body connected with breathing may be facilitated by this zone of small, non- 

 imbricating scales lying upon a flexible integument. Analogous, but more 

 obviously necessary, are the lateral grooves of soft skin or fine scales seen in 



