Jan., 1902,] Varlatwns in the Water-Snake. 185 



from the neck. The pre-anal scale was found to be bifid in sev- 

 eral of the snakes. Since the gastrosteges represent modified 

 scales — scales such as occur on the dorsal and lateral surfaces of 

 the body — these bifid scales ma}^ represent a primitive condition, 

 where the vxntral scales were similar to the dorsal scales. This 

 bifidity of gastrosteges is not at all common in the adults of this 

 species. Hence it seems probable that during succeeding moults 

 the normal gastrostege is finally obtained. A fact that points to 

 this conclusion, indirectly however, is that on examining the 

 labials, the fourth lower labial in specimen number 4 was found 

 normal, but the epidermis, which had been loosened by the form- 

 alin, was seen to have a bifid labial corresponding to, and lying 

 immediatel}^ above, the fourth labial. This shows that in this 

 case at least the labial was changed from a bifid to a single plate. 



The point of exit of the umbilical cord (z. e., the yolk-sac and 

 the allantois), the so-called dermal umbilicus, being an old 

 structure, would be definite in position to a certain degree. The 

 number of gastrosteges anterior to this was found to vary to the 

 amount of ten scales in dift'erent individuals. 



In respect to the scutae of the head, it ma}- be said that little 

 variation was found. The shape of the rostral, vertical, nasals, 

 etc., presented little perceptible differences in the several indi- 

 viduals. However, the number of post-oculars was found to be 

 different in difierent snakes and on opposite sides of the same 

 snake. In column 6 of the table these variations are shown. 

 The first number represents the scales on the left side, the second 

 number those on the right side of the snake. Where the figures 

 are the same, as e. ^., 3-3, there is no variation in the number of 

 postoculars on the two sides of the head. 



The number of longitudinal rows of .scales was found to vary 

 in each specimen and in different parts of the same specimen. 

 Thus, in tracing a row of scales from head to tail, the row would 

 often end in a V-shaped plan, formed by the approximation of the 

 rows lying on either side of the row in question. The same was 

 true of the mother. Therefore no attempt was made to tabulate 

 them. 



The plate on page 184 shows the 33 3-oung from a dorsal view. 

 They should be counted from right to left to correspond wath the 

 table. It will be seen that the color pattern varies in the 

 different snakes. All gradations, from regularly arranged saddle- 

 shaped markings to chain-like figures are found. The greatest 

 irregularity in the markings occurs in the region over the heart. 

 The post-occipital collar is entire in .some, as in specimen No. i ; 

 in others, siich as No. 12, this collar is cut into lateral moieties, 

 all gradations existing between the two. 



If Cope's conclusions are correct that, in reptiles at least, color 

 variations arise posteriorly and advance anteriorly, the regular 



