VAEIATIONS OF GARTER-SNAKES. 137 



material, will doubtless be corrected with more specimens, but they 

 indicate the general conditions. 



Affinities. — The possession by this form of two preoculars in the 

 southern part of its range is very interesting, since it is in this region 

 that it approaches nearest to the region inhabited by angustirostris , 

 and the division of the preoculars (a rare trait in the genus) furnishes 

 a strong inducement to consider its presence in both forms an evi- 

 tlence of direct relationship. The derivation of either of these forms 

 from the other is not as improbable as it may seem at first sight, for 

 they are closely similar, especially in the adjacent parts of their 

 ranges. They have apparently the same number of dorsal, labial, 

 ventral, and subcaudal scales and tail length, besides possessing in 

 common the characters of more than one preocular and absence of the 

 tlorsal stripe. Hammondi differs from angustirostris in generally hav- 

 ing two instead of three preoculars, the eye in contact with two 

 labials, not well-defined lateral spots, and in the presence of stripes. 

 Since angustirostris often has but two preoculars, and indeed appar- 

 ently to the south grades into a form (:melanogaster) which has but 

 two, it is not impossible that the number also decreases in the northern 

 part of the range. We have already seen that the other scale charac- 

 ter (the separation of the eye from one labial) suffers a similar fate. 

 The only apparent differences between angustirostris and hammondi 

 that remain, then, are the presence of the stripes and the obscurity of 

 the lateral spots. These differences are comparatively slight, but 

 the fact that the dorsal stripe is nearly always w^anting in hammondi 

 renders the differences still less. In fact, in the general type of color- 

 ation as well as in the arrangement of the ocular scales, hammondi is 

 similar to melanogaster, and I believe that there is strong evidence for 

 concluding that angustirostris is genetically midway between these 

 two forms. 



Its resemblance to angustirostris decreases in the northern part of 

 its range, and hammondi comes to resemble another form (elegans), 

 which replaces it to the northward. Its resemblance to this form is 

 so close that specimens from near the range of the latter (Inyo, Kern, 

 Tulare, and Fresno counties, California) are distinguished with the 

 utmost difficulty. This is brought about by the fact that hammondi 

 usually has a single preocular in the northern part of its range and 

 occasionally a dorsal stripe, and it can with certainty be distinguished 

 from elegans only when it has either two preoculars or no dorsal stripe. 

 However, sufficient proof that these forms actually intergrade is 

 wanting. Their relations will be considered more fully when the 

 affinities of elegans are discussed. 



