SNAKES OF THE GENUS PITUOPHIS 17 



scale reduction in Elaphe and Lampropeltis on the one hand, and in 

 Thamnophis and the allied genus Matrix on the other. The discovery 

 of a third genus corresponding in method with the two latter genera 

 is exceptionally interesting in its relation to the question of generic 

 relationships. However, only the study of many more genera can 

 determine the value of the method of scale reduction in indicating 

 intergeneric affinities. 



In Pituophis, as in all the other forms studied, there is an evident 

 correlation between scale row formulae and sex. In every form of the 

 genus the formula is on the average higher in females than in males, 

 and in almost every case the abnormally high formulae that occa- 

 sionally occur are found in female specimens. This sexual variation 

 in the diameter of the body might be expected in correlation with the 

 egg-laying function of the females, and the undoubted retention of 

 large eggs for a considerable period. 



The high formulae are generally found, furthermore, near the 

 probable center of origin for the form, and there is always a noticeable 

 tendency toward a general decrease in the number of scale rows away 

 from the center of distribution. Although there is no appreciable 

 difference in the diameter of individuals of approximately the same 

 size, but with different scale formulae, nevertheless the geographic 

 correlation between the decrease in scale rows and the decrease in 

 other scale characters occurring in some forms seems to indicate that 

 scale reduction, in these forms at least, is evidence of a general dwarfing 

 toward the periphery of the form. Thus, for example, in sayi and 

 other forms there is a fairly constant decrease in numbers of scale 

 rows and ventrals from south to north, and the average size of the 

 individuals seems to be greater from the southern than the northern 

 part of the range, while all the largest specimens examined are southern. 

 On the other hand, the forms possessing the smaller average formulae 

 for the genus are not always the smaller in general size and may, on 

 the contrary, be distinctly larger than forms with much higher average 

 formulae. This is true of the subspecies of melanoleucus, which, 

 although possessing comparatively low scale formulae, are the largest 

 forms of the genus. Furthermore, in the subspecies of melanoleucus 

 the individual scales are proportionately larger than in any other 

 forms. Thus the theory of dwarfing can be applied only to intra- 

 specific variation, and not to variation between different forms. 



Variation in numbers of ventrals and caudals and in proportionate 

 tail length. — As stated above, the ventral surface is covered with a 

 series of large transverse scutes, the ventrals anterior to the vent in a 

 single series, and the caudals on the under surface of the tail in a 

 paired series. The number of scutes in these series varies within 

 definite limits in each form and has been generally recognized as of 

 primary importance as a systematic criterion. In this genus, how- 



