200 BULLETIN 114, UNITED STATES NATIONAL MUSEUM. 



This indicates that the District of Columbia specimens have suf- 

 fered an actual reduction in the number of temporal scutes. In the 

 extent of this reduction they can be compared only with micwpholis, 

 elapsoides, and virginiana, reduced forms belonging to the same group, 

 but only distantly related to triangulum. 



Bearing in mind the peculiarities of the geographic variation in 

 ventrals, scale rows, and temporals we have now to consider the color 

 pattern. As will be evident from the descriptions, triangulum is 

 separated from syspila almost entirely on the basis of color pattern 

 differences. Typical specimens of the two forms are very different 

 appearing animals, and, although intergradation certainly takes place 

 along the common boundary of their ranges in Illinois, Indiana, and 

 Ohio, each form seems to retain its distinctive features very close to 

 the border of its range. Just how true this statement is for Kentucky 

 and Tennessee can not be told now for lack of specimens, but the few 

 we do possess indicate that syspila extends over about the western 

 third of these States. Throughout the northern tier of States the 

 typical pattern, as given in the description, is seldom departed from. 

 The dorsal blotches average about 47, are almost always more than 35, 

 and are usually accompanied by two more or less well-developed 

 series of lateral alternating spots on each side. But in the southern 

 AUeghenies the dorsal blotches average somewhat fewer, the higher 

 numbers of the north are rarely, if ever, attained, the blotches often 

 extend lower down the sides, the two rows of lateral alternating 

 blotches also often extend down lower, and are rather frequently 

 fused, either partially or completely, into a single series. This does 

 not mean that typical examples do not occur, but that they are less 

 frequent than in the north. This reduction in number of dorsal 

 blotches, accompanied by increase in width and consequent loss of 

 the upper series of alternating spots, reaches its extreme, as the dia- 

 gram shows (fig. 57), in New Jersey, Delaware, Maryland, and 

 northern Virginia. 



The complicated and characteristic pattern of the head follows, in 

 a general way, the pattern of the body. It is most perfectly developed 

 in the northern parts of the range. Even here the chevron on the 

 back of the head may appear only as a transversely oval spot, and the 

 white line between, over, and behind the eyes may be partly bridged 

 by extension of the brown pattern. But in practically all of these 

 northern specimens the half collar is completely replaced by the 

 chevron or its representative. Close to the range of syspila the half 

 collar is often partly or completely developed. Many specimens from 

 western Pennsylvania and northern West Virginia, while typically 

 triangulum in other respects, possess the half collar instead of the chev- 

 ron. South in the AUeghenies aU conditions of chevrons and half 

 collars may be found. Often when the half collar is present there 

 will be two notches in its anterior border or one in its posterior, rep- 



