PARASITIC WEAVERBIRDS 9 



pli3^1ogenticall3'' conservative character as the pelvic musculature, the 

 carduelines agree with the ploccids and not with the fringillids. This 

 statement agi'ees with Tordoff's suggestion that the cardueline finches 

 are to be looked upon as ploceids, and it seems unnecessary to estab- 

 lish a new family Carduelidae for them and the estrildincs. Certainly'', 

 passerme families are slightly marked at best, and to increase their 

 number does not help to solve the elucidation of their relationships. 



The nearest relatives of the viduines are the estrildines, and then 

 the ploceines. As I have noted above, the viduines are apparently 

 an early offshoot from the estrildine stock, as they also possess a 

 number of characters of the ploceine stem. Wliile this arrangement 

 seems fairly clear, it is difficult to point to any particular existing 

 speries or genera in these groups as the most probable ancestor of 

 the viduines. 



In his 1917 diagram, Chapin showed the viduine line branching off 

 from near Clytospiza. This guess is as good as any that can be made; 

 I know of no other genus that fits the assumed requirements of a pos- 

 sible ancestor any better than does this one. Clytospiza is a fairly 

 generalized, smallish weaver with no specialized feathers or pecuJiar 

 habits. In the pattern of the mouth markings of the nestlings, in the 

 pure v/hite color of its eggs, and in that it makes no nest of its own 

 but uses abandoned nests of other birds, such as coucals and weaver 

 finches, Clytospiza fits into the picture that one might imagine as an 

 ancestral stock for the present parasitic widow birds. 



The present distribution of Clytospiza is perhaps more strictl}'' 

 tropical, and does not extend as far to the south as might be ex- 

 pected of the ancestral stock, which in the southern savannas is 

 thought to have given rise to the ^^duines, but we have no way of 

 knowing if Clyiospiza vras not more extensive meridionaUy in earfier 

 times than now. Furthermore, I use Clytospiza not in the literal 

 sense of the present genus, but as a portion of the phyletic stem of 

 the family from which the viduines diverged. 



We are now in a position to attempt a more detailed arrangement 

 of the little branches on the viduine stem. Prior to Chapin's revision 

 of the classification of the family (1917) the long-tailed viduas were 

 grouped together with such unrelated but parallel developments as 

 the euplectine genera Coliuspasser, Diatropura, and Drepanoplectes 

 merely on the basis of the long rectrices in the nuptial plumage of 

 the adult males. As Chapin showed, however, the elongation of the 

 rectrices is a character that has developed independently in different 

 sections of the family, and in ways that are reallj^ quite dissimilar. 

 Thus, the patterns of tail growth "in, for example, Coliuspasser and 

 Vidua have nothiug in common but their length and their seasonal 

 molt. In Vidua and Steganura it is only the two median pairs of 



