114 BULLETIN 114, UlSTITED STATES NATIONAL MUSEUM. 



weaker on one side of the crossband than on the other, so that the 

 dorsal dark areas become fused with the lateral on one side of the 

 crossband, and on the other the white is strengthened. Similarly the 

 lateral dark areas fuse with a ventro-lateral series, and thus become 

 complete on the belly. It now only remains for these newly formed 

 rings to be perfected. This occurs by a whitening of the white rings 

 and a fading of the white that may still form ragged edges projecting 

 into the dark areas. All these transitional stages may be observed 

 in specimens from the vicinity of Tucson. With the perfection of 

 the pattern of yumensis (fig. 28), the only change necessary to pro- 

 duce the pattern of hoylii (fig. 27) is an extension of the white in the 

 white rings to the whole of the scales in these rings. 



Thus the patterns of the coast forms of the getulus group result 

 from the tendency of the pattern of tj^pical splendida to become 

 broken into an alternating series of black and white areas. 



That splendida may well be the parent form of the getulus group 

 is indicated by its present distribution. It appears to be confined 

 to the region between central Texas and eastern Arizona. Its north- 

 ern limit is unknown, but is doubtless the mountains of central or 

 northern New Mexico. That it extends south into Mexico is as cer- 

 tain as anything can be, but how far we can only surmise. For 

 southern Mexico, where considerable collecting has been done, there 

 is no record of splendida, nor of any form allied to it. We ma}^ there- 

 fore say that the arid plateau region of the southwest is its home. 

 This has often been commented upon as being a favorable center for 

 preservation and for dispersal of reptilian life. The present diversity 

 of forms may be considered the result of an east and west dispersal, 

 with consequent modification in each major environment. 



The foregoing discussion may be sunoimarized as follows : ' 



1. Since each member of the getulus group is nearest and evidently 

 directly related to the form inhabiting the adj acent region, and since 

 the forms are in linear arrangement geographically, one of them 

 must therefore present a condition from which all the others may 

 be derived (californiae and nitida are ruled out of consideration here 

 for reasons discussed under the former) . 



2. Of the eight forms entering the discussion, two — conjuncta and 

 jioridana — are eliminated because their geographic position and the 

 evidence for their derivation from the adjacent continental forms 

 is too strong to allow them to be considered as ancestral possibilities. 



3. Comparative study of loreal and chin shields indicate that the 

 east and west forms present specializations in these scutes. 



4. The proportions of the head and its scutes indicate that splendida 

 is the most closely allied to the other forms of the genus, and that 

 both the east and the west forms of the getulus group are specialized. 



5. Getulus, jioridana, niger, and liolbrooki are excluded from con- 

 sideration as possible ancestral forms on account of the unique 



