REVISION OF THE KING SNAKES. 241 



adult is quite distinct. Conjuncta is in all probability a derivative 

 of hoylii. A similar situation probably holds with respect to liol, 

 hrooM, and probably also with several other forms of this genus- 

 The probability is therefore strong that the fact that the young of 

 triangulum have red instead of brown blotches indicates that the 

 ancestor of triangulum was a red instead of brown spotted snake. 

 (6) If we have any regard for the evidence derived from the getulus 

 and calligaster groups that marked reduction in structural features 

 is evidence of specialization, we can not ignore the fact that triangu- 

 lum is the most reduced form in this whole series. This reduction is 

 expressed particularly in the great frequency of the scale formulae 

 21-19-17 and 19-21-19-17 and the large number of specimens that 

 possess a temporal formula lower than 2 + 3+4, i. e., 2 + 3+3, 2 + 2 + 3, 

 1+2 + 3. 



If we accept these reasons for not regarding triangulum as ancestral 

 we are led next to sijspila. But this form does not possess the pat- 

 tern that was alleged to be primitive in triangulum. In scale rows 

 and temporals it shows the same degenerative tendencies that are 

 more strongly expressed in triangulum. In fact, there appear to be 

 no good reasons for considering syspila as primitive. This leads us 

 directly to amaura. This form is restricted to the lower Mississippi 

 Valley — a region geologically new and therefore totally unsuited to 

 be a center of origin of a group of animals. Gentilis could scarcely 

 be regarded as an ancestral type as its color pattern is very evidently 

 a derived one, it is a reduced form in number of ventrals, and it 

 inhabits a region entirely unsuited for preservation of old types. 

 The only other possibilities are the Mexican forms nelsoni and 

 annuhta, and for present purposes it is immaterial which of these 

 may be the oldest. Nelsoni is less likely to be primitive because it 

 inhabits chiefly the coastal area, and it presents a derived appear- 

 ance in the extreme width of the red interspaces. There are for 

 annulata, however, several positive reasons for regarding it as the 

 ancestral type of the triangulum series. (1) In structural features 

 it presents the closest approach to the primitive types of the other 

 large groups: The loreal is oblong with no peculiarity in shape; 

 the scales rows are 23 or 21 (no specimens drop to 17 at the 

 end); the temporals are 2 + 3 + 4; the labials are 7 and 9; the chin 

 shields are normal; the proportions of head, body, and tail are 

 average ; the ventral plates average but slightly less than for splen- 

 dida, (2) It occupies a geographically favorable region, the Mexi- 

 can plateau. (3) It is in a region inhabited by the ancestors of the 

 getulus and calligaster groups. (4) It is in a region sho^\Ti to have 

 been the center of dispersal for another widespread and recent group 

 of North American snakes, the garter snakes. (5) The color pattern 

 of all the forms of triangulum may be most simply derived from that 



