242 



of annvlata. (6) Still more important is the fact that annulata lies at the 

 point of convergence of three separate genetic series of the triangulum 

 group : The triangulum-syspila-a7naura-annulata series, the virginiana- 

 elapsoides-amaura-annulata series, and the gentilis-annulata series. 

 This situation fulfills therefore the most convincing of the criteria 

 formulated by Adams (1902, 122) for determination of centers of dis- 

 persal, namely, ^^ continuity and convergence of lines of dispersal." 



If we accept annulata as the stem form of the triangulum series we 

 may describe the derivation of the color patterns as follows: From 

 the plateau region of southern Mexico north to extreme southern 

 Texas we have a form, annulata (fig. 66), ringed with black and white 

 or yellow, the black more or less split with red dorsally but continuous 

 between the white rings on the belly, except at the extreme anterior 

 end of the body, where the red may cross the belly. The white rings 

 are nearly uniform in width on the dorsal scales and only slightly, if 

 any, widened on the ventrals; they are not encroached upon by the 

 black, but may be more or less mottled with black, particularly on 

 the sides. Figure 66 shows this pattern as exhibited by a typical 

 specimen from Puebla, Mexico. Passing to amaura (fig. 67), we see 

 that the white rings have become distinctly widened on the lower 

 rows of scales, that the black on the belly has become separated on 

 the midventral line, and that the red has increased in extent at the 

 expense of the black. The black has here ceased to encircle the body, 

 and has become restricted to a border for wide dorsal saddles of red. 

 Figure 68 is a typical example of sysjnla from Arkansas. This pat- 

 tern is derived from the last by a lateral contraction of the dorsal 

 red saddles, a narrowing of the black borders, the development of a 

 ventro-lateral series of black spots in alternation with the dorsal 

 blotches, a checking of the belly with small quadrate spots of black, 

 and an increase in the mottling of the white dorsal scales (no mottling 

 was shown on the white scales of figure 67 because the specimen was 

 much faded). The alternating spots and perhaps some of the small 

 quadrate spots may be the result of a pinching off, as suggested by 

 figures 67 and 68. Figure 69 shows the pattern of a typical triangulum 

 from northern Michigan. Whether the upper series of alternating 

 spots is the result of an upward migration of the first series (fig. 69) 

 is uncertain, but it appears much more likely that it is due to a pinch- 

 ing off of segments of the dorsal blotches in the course of their con- 

 traction. This is the end of the series in this direction. 



To derive gentilis we must return to annulata. The course of evo- 

 lution has here been different. Instead of a lateral shrinking of the 

 black and red, there has been a longitudinal shrinking. The begin- 

 ning was the same as for the other series. The white rings widened 

 on the lower rows of scales and on the ventrals, thus greatly increasing 

 the width of the white rings in proportion to the black on the belly. 



