THE AMERICAN NATURALIST. [Vol. XXXVII. 



having tails 25.5 per cent of the body length, one has 78 uro- 

 steges and the other only 70. The accompanying diagram (I) is 

 designed to show the variability in length of tail and in number of 

 urosteges, and also the correlation between those two characters. 



5. Greater Variability of Females. — One of the most inter- 

 esting facts brought out during this investigation is that female 

 water snakes are far more likely to vary from the normal than 

 are males. This is not a matter of size, for many of the aber- 

 rations are among the small snakes, and they do not seem to be 

 more frequent among large specimens. Thus, of 1 1 females 

 over 900 mm. in length, 45 per cent were normal as regards 

 the labial plates and number of scale rows (the points in which 

 the variability is most marked), while of 12 snakes less than 

 800 mm., less than 42 per cent were normal. In both erythro- 

 gaster and sipedon there are, normally, 8 labial plates on each 

 side of the upper jaw and 10 on each side of the lower; the 



normal formula therefore is — — — • Now, of the iq male 

 10-10 



erythrogasters examined, 16 possessed the normal number and 

 arrangement of the labials, and of 33 male sipedons, 25 were 

 normal ; of 52 males, therefore, 41, or 79 per cent, were normal 

 as regards the labials. Of the 8 female erythrogasters, only 4, 

 and of 30 female sipedons, only 14, were normal ; of 38 females, 

 therefore, only 18, or 47 per cent, were normal as regards the 

 labials. The 20 abnormal females show 26 variations from the 

 normal, and of these 21, or 80 per cent, are added plates, while 

 the 11 abnormal males show 16 variations, of which only 9, or 

 56 per cent, are added plates. Granting that these cases are 

 too few to determine any general law of variability, they are 

 nevertheless suggestive. Turning now to the number of scale 

 rows on the back, we find additional evidence of the greater 

 variability of females. The number of such rows is, normally, 

 23, counting where they are most numerous, which is usually 

 about one-third of the total length, back of the head. Of the 

 52 males examined, 45 had 23 rows, 3 had 24, and 4 had 25 ; 

 thus 86.5 per cent were normal. Of the 38 females, 29 had 

 23 rows, 6 had 24, and 3 had 25 ; thus, 76 per cent were 

 normal. It ought perhaps to be emphasized that this increase 



