REVISION OF THE KING SNAKES. 109 



other more western forms, is intensified in getulus into distinct forks, 

 a character that is even more strongly developed in Jloridana. Since 

 this feature is unique for the genus, we must consider it a strong reason 

 for excluding these forms from the list of ancestral possibilities. 



In connection with the positive evidence of specialization in the 

 forms east of splendida derived from the characteristics of the loreal, 

 chin shields, and copulatory organ, we must now consider that the 

 reduced scutellation of these forms is also a result of specialization. 

 We are forced to conclude, therefore, on structural grounds, that 

 Jloiidana, getulus, niger, holbrooTci, and probably the forms west of 

 the Rockies, must be excluded from consideration as possible ancestors 

 of the group. 



Turning now to the pattern, we find that its variations are truly 

 remarkable, ranging as they do from cross bands, to small spots, 

 to rings, and stripes, and to various combinations of these. In fact, 

 the pattern is the most variable of all the characters upon which 

 specific value is placed. It may therefore be expected that pattern, 

 if anything, will give a clue to relationships. 



Since the head pattern is the least variable, we may consider that 

 first. One of its strikingly constant characters is the pair of white 

 bars on the prefrontal and internasal plates. In the Pacific coast 

 forms these are broad, occupying most of the area of the scutes, and 

 usually so broadened as to largely lose their transverse character; 

 in getulus and Jloridana they are narrow and usually close to the 

 anterior and antero-lateral borders of the scutes; often they are 

 curved, or broken in the middle, or reduced to spots; in Tiolhrooki 

 they are pretty constantly narrow transverse bars, the chief varia- 

 tion being to reduction, breaking into spots, and curving; in yumensis 

 they are strong, symmetrical transveree bands on the anterior por- 

 tions of the scutes; in splendida they vary from the yumensis form 

 at the extreme west to the Tiolhrooki condition in the east; in the 

 central portion of its range these bars are narrow, rather irregular, 

 sometimes broken, or somewhat bent or curved. These bars, there- 

 fore, reach their extremes of diversity in the east and west coast 

 forms, and are developed most perfectly and symmetrically in Arizona 

 and in the Mississippi Valley. If we feel it necessary to assume as 

 the starting point the perfect and symmetrical condition, we must 

 explain why, starting with either liolhroolci or yumensis, evolution has 

 proceeded in one direction to an extreme condition (the nearer coast 

 forms), and in the other through a reduced condition {splendida), 

 to a perfect condition again, and then to another variable extreme. 

 To the writer the only logical explanation is to consider the condi- 

 tion presented by splendida in southern New Mexico as the starting 

 point. Here we have the beginnings of the white bars, imperfectly 

 formed as symmetrically placed lightened areas on a black head; evo- 



