REVISION OF THE KING SNAKES. 113 



the dark areas. This is well shown by a specimen from New Orleans 

 (fig. 32). In this the symmetrical oval spots occupy only the areas 

 that were without white spots in the specimens from central and 

 southern Texas, and the spots which were present in these specimens 

 are present here too, and with much the same unsymmetrical orienta- 

 tion as was shown initiated in figure 31. That is, Jiolbrooki has 

 retained the pattern of splendida, as developed in Texas, but has 

 added something to it. 



The form niger is developed from holbrooki by a fading of the newly 

 derived white spots between the bands and a contraction of the white 

 in the crossbands and sides (fig. 33). The exact derivation of getulus 

 (fig. 35) from nige?' (fig. 33) is uncertain, due to scarcity of specimens 

 from the critical region. A specimen from Anniston, Alabama, sug- 

 gests that the transition to getulus was accomplished by a disappear- 

 ance of alternate cross bands on the back and possibly by fusions of 

 the light areas on the ends of the ventrals. In this specimen the 

 crossbands are scarcelv distinguishable but seem to be reduced in 

 number. A specimen from Marietta, Georgia (fig. 34), has the pat- 

 tern of getulus, but the crossbands are very narrow. The change 

 from this to the typical pattern was accomplished by a widening of 

 the white parts, thus making the crossbands conspicuous and bringing 

 out the chain pattern that was initiated in the splendida of central 

 and southern Texas and remained inconspicuous in holbrooki. The 

 pattern of such a typical getulus is shown in figure 35. 



Floridana was developed from getulus by a basal lightening of each 

 dark scale, and an increase in the number of crossbands (fig. 36). 

 Here, as in Tiolbroolci, we have a spotted pattern that is at the end of 

 an evolutionary series, but it is a decidedly different sort of spotted 

 pattern from that of splendida (fig. 30) in that it bears the vestiges 

 of an earher pattern while that of the latter apparently does not. 

 In extreme southern Florida this series of patterns is carried one step 

 farther and to its logical conclusion in hroolcsi (fig. 37), in which even 

 traces of the crossbands appear to be lost. Numerous structural 

 features, however, mark it as a specialized type. 



Yumensis is formed from splendida even more simply. It occurs 

 as follows: Westward from southern New Mexico splendida becomes 

 somewhat altered. The essentials of the process are fundamentally 

 the same as the changes eastward. The white centers lose their sym- 

 metrical form and arrangement and assume various shapes. Small 

 dark areas develop on the sides opposite the crossbands (fig. 29). 

 The doi-sal dark areas increase in length and breadth. At this point 

 we have the fundamentals of another getulus pattern, but, apparently 

 by chance, rings are produced instead of a chain pattern. The inter- 

 mediate forms upon which this outline is based show that the white 

 spots separating the lateral dark areas from the dorsal ones are 



