112 BULLETIN 114, UNITED STATES NATIONAL MUSEUM. 



with its description), the various forms of the getulus group all have 

 a pattern that is fundamentally one of narrow transverse white bands 

 on the back, separated by a ground color of black (figs. 27 to 37). 

 The numbers of these crossbands or rings, as the case may be, have 

 been averaged and plotted, for each form, in the accompanying 

 diagram (fig. 26). Two contrasted conditions are shown — the high 

 numbers of splendida and lioTbrooki, and the low numbers of getulus 

 and the west coast forms (the situation with foridana has already 

 been explained under the discussion of that form) . It is noteworthy 

 that the extremes of the group — the coast forms — ^have approxi- 

 mately the same variation and average in number of bands. The 

 fundamental difference in pattern between these forms is that the 

 bands in the west coast forms (figs. 27 and 28) are continued around 

 the body as rings, while in getulus {-^g. 35) they fork on the sides and 

 the connections across the belly are imperfect; from above both body 

 patterns look almost identical. It seems to the writer quite out of 

 reason to suppose that any of the west or east coast forms could have 

 changed into the spotted condition of splendida and TiolbrooM and 

 then emerged v/ith the pattern exhibited on the other coast. The 

 fact that the coast forms have patterns so strikingly similar suggests 

 that, like the diverse modifications in the pattern of snout and labials, 

 they have been produced by independent evolution from a similar 

 beginning. Then, of necessity, either Jiolhroold or splendida must be 

 examined for ancestral possibilities. The former is, however, suffi- 

 ciently rejected on structural grounds. 



The pattern of splendida as exhibited by examples from southern 

 New Mexico and northern Chihuahua briefly is this: Ground color 

 black; back crossed by about 50 or 60 narrow white bands formed of 

 black scales with oval white centers; sides with a white oval center on 

 each scale, the white occupying more area toward the belly; latter 

 black, except that the ends of about two gastrosteges, opposite the 

 dorsal white bars, are white. The important thing to notice is that 

 the white centers are all very approximately oval and oriented the 

 long way of the scales. Figure 30 shows a typical example from 

 southern New Mexico in which the black spaces on the back are very 

 small and the white centers are all very nearly uniform in orientation. 



Going eastward in the range of splendida we find that in western 

 Texas the white centers are often angular or long and narrow and 

 tend to lose their symmetrical orientation. This is sho^vn by a 

 specimen from Keeves County, Texas (fig. 31) . In central and south- 

 ern Texas this change is more marked, and there is a decided tendency 

 for some of the white centers on the sides opposite the crossbands to 

 fade, causing a series of small dark areas to appear on the sides, in 

 alternation with the dorsal dark areas. This is the beginniag of the 

 chain pattern of niger and getulus. It is obscured, however, in 

 liolbrooki by the appearance of an oval white center in each scale of 



