8 BULLETIN 2 08, UNITED STATES NATIONAL MUSEUM 



2 in Melignothes, and 1 in Pseudofringilla) . It seems that the process 

 of differentiation was probably completed (at the present level) 

 long enough ago to allow for the subsequent disappearance of inter- 

 mediate stages, and also that the main lines of differentiation came 

 not so much in chronological sequence from earlier derivatives from 

 the original stock but directly from the original stock itself. As far 

 as may be judged from the structure and external character of the 

 present species, there were three radiating branches from the original 

 ancestral stock. Bearing in mind the probable similarity of the 

 primitive honey-guides to the nonspecialized barbets, it seems that 

 this original stock was generally more similar to the large-billed 

 species of Indicator (subgenus Indicator) than to any of the other 

 present forms. Adding to this the likelihood of the family being 

 originally a forest-dwelling group, it would seem that Indicator 

 maculatus of the West African forests is as near as any to this ances- 

 tral group. The development of /. maculatus is thus the base of the 

 first line of descent, and from it, or from something akin to it, devel- 

 oped two branches, one leading through /. variegatus to /. archi- 

 pelagicus and /. indicator and the other leading from 7. variegatus 

 to Melichneuies rohustus. It may be noted that it is only in that 

 branch of the phylogenetic tree of the gi-oup that leads from 7. varie- 

 gatus to 7. indicator that we find evidence of the guiding habit. 

 Whether this remarkable habit is also exhibited in 7. archipelagicus 

 is, unfortunately, unknown, and further information on that species 

 is apt to be very slow in coming in. That 7. variegatus is less advanced 

 than 7. indicator is in agreement with the slight development of 

 guiding in the former as compared with its great development in the 

 latter. The intermediate position of 7. variegatus between 7. macula- 

 tus on the one hand and 7. archipelagicus and 7. indicator on the other 

 is further suggested both by its coloration (losing the heavily macu- 

 lated pattern of its "ancestor" and approaching the plainer pattern 

 of its "descendants") and by its voice (its purring note being appar- 

 ently reflected in the catlike mia-ow krruuuu of 7. archipelagicus). 



A second line of differentiation — expressed chiefly in the reduction 

 in the size of the whole organism, particularly the length of the bill — 

 may be traced from the hypothetical ancestral stock to that which 

 bifurcated into the subgenera Melignothes and Pseudofringilla. 

 The former of these produced 7. minor and its races, and then, by 

 further reduction in over-all size, 7. exilis and its subspecies; the 

 latter is evident only from its one known end product, 7. xantho- 

 notus. 



The third line of descent, involving originally chiefly the narrowing 

 of the bill, produced Melignomon zenkeri, and then, through the loss of 

 one pair of rectrices and the development of a softer, fluffier plumage, 



