60 



Christopher W. Brown and Carl H. Ernst 



are listed in Table 1; the first 13 are those of Gloyd (1940) and were also 

 used by Pisani et al. (1973). However, the method of counting cross- 

 bands was probably different in this study; the band was not counted if 

 it was interrupted by at least one scale of ground color (Fig. 3). 



Specimens not from localities of probable intergradation were clas- 

 sified a priori into one of the two forms (subspecies) based on that race's 

 distribution as defined by Gloyd (1940). The 21 probable intergrades 

 (Fig. 2) not used in the analyses were classified into groups based on 

 their localities: those from within the range of C. h. horridus, those 

 from within the range of C. h. atricaudatus, and those from localities 

 lying between the two ranges. 



Fig. 3. Method of counting dorsal scale rows (numbered at top) and crossbands. 

 Of the three apparent bands, only one complete crossband would be counted 

 here, for only one is uninterrupted by any scales of ground color. 



