244 BULLETIN 114, UNITED STATES NATIONAL MUSEUM. 



than for any of the subspecies of triangulum, except nelsoni, of the 

 west coast of Mexico. In proportions, dentition, and scalation it is 

 so close to the average for the genus that it is difficult to conceive of 

 it as much specialized or as a derivative of pyrrJiomelaena. In fact, 

 for regarding it as a subspecies of the latter there is no evidence. 

 We do not know whether their ranges meet, but the specimens on 

 hand show no approach of one form to the other in pattern or scala- 

 tion. It is even possible that multicincta may be only distantly 

 related to pyrrhomelaena. Its color pattern is similar, to be sure, and 

 their ranges are practically adjacent, but if pyrrJiomelaena did not 

 exist we would probably not be unwilling to regard multicincta as the 

 west coast representative of the subspecies of triangulum. The 

 discovery of the Patzcuaro specimen, indicating, as it does, the occur- 

 rence in southern Mexico of a close relative of multicincta, puts a new 

 face on the situation. Nothing definite can be said, however, imtil 

 more specimens of the new form, rutJiveni, shall have been found. 



PyrrJiomelaena gives much evidence of specialization in (1) its 

 high number of labials, (2) its atypical pattern, (3) long tail, (4) high 

 dentition, and (5) v/ide head. If we examine its color pattern 

 {^g. 71) we notice that while there is the same marked tendency for 

 the black to overspread the red dors ally, the former narrows decidedly 

 on the sides and often fails to reach the ventrals^a decided difference 

 from gentilis and multicincta; and the black of the belly, contrary to the 

 situation in the latter two forms, is concentrated chiefly opposite the 

 dorsal white rings, and it is the red that most regularly crosses the 

 belly. It is difficult to say how important this distinction may be. 

 The prominent specializations in structural features indicate old age 

 in the form and show that there may have been time for the pattern 

 to develop these peculiarities starting from the annulata condition. 

 The suspicion can not be avoided, however, that the present simi- 

 larity in the patterns of multicincta and pyrrJiomelaena are due to 

 convergent evolution. 



It is noticeable that multicincta is less removed structurally from 

 the average of the genus than pyrrJiomelaena, as it is less removed in 

 style of color pattern from gentilis and annulata. 



On the whole it would seem best to place these two forms with the 

 triangulum group on the basis of similarity in general style of color 

 pattern — that is, rings of red, yellow, and black — and to regard pyr- 

 rJiomelaena tentatively as an isolated form, at least as old as elapsoides, 

 and multicincta as the isolated representative on the Pacific coast of 

 the rest of the triangulum group. 



The preceding discussion may be summarized as follows : 



1. MicropJiolis must be regarded as a specialized derivative of 

 polyzona; the relationship of the latter to nelsoni and annulata is 

 close but not clear. 



