THE AMERICAN NATURALIST. [Vol. XXXVIV 



exact numerical series, yet it is interesting to see how the sepa- 

 ration of erythrogaster from sipedon, so evidently shown by the 

 diagram of gastrosteges, is emphasized by this diagram of color. 



6. Variability. — It has already been shown that female water 

 snakes are more variable than males in the number of scale 

 rows and labials. It is interesting to see that in both these 



respects, as well as in color, sipedon is far more variable than 

 erythrogaster. This is clearly shown by the table on the 

 opposite page. 



It may further be stated that there are two specimens of ery- 

 throgaster, one male and one female, which are apparently normal 

 in every respect, as they have the average number of gastro- 

 steges (151) and of urosteges (77 and 67 respectively), the 



proper number of labials on each jaw ( J, and the proper 



number of scale rows (23). Such a normal individual of sipe- 

 don is not to be found among my 63 specimens. In Table II 

 may be found another illustration of this same point, for it 



