388 University of California Publications in Zoology [Vol. 13 



relationships cannot be based on any single character ; it must be 

 a composite picture of the evidence furnished by all the characters. 

 In the following paragraphs there is given a summary of the evi- 

 dence furnished by one single small character, the morphology of the 

 minute structure of feathers. In itself it is obviously inadequate. 

 Taken in conjunction with the evidence furnished by other characters, 

 however, it is hoped that the evidences of relationships suggested by 

 epiphyology may help to bridge the gaps left by other comparative 

 studies, and so help to complete and perfect the chains of relationship 

 in avian phylogeny. 



Beginning with the Ratitae, the structure of the feathers of the 

 various orders included in the group gives very strong evidence in 

 favor of their being primitively rather than secondarily flightless 

 birds, the Struthionif ormes and Rheif ormes apparently being end 

 branches of one main evolutionary stem, and the Casuariiformes and 

 Apterygiformes similar end branches of another stem which is prob- 

 ably to be considered a little nearer to the line leading to the Cari- 

 natae (see summaries on pages 288, 290, 293, 295). 



Among carinate birds the most primitive feathers are to be 

 found among the penguins, and next above them the Colymbiformes 

 (for details, see pages 298 and 302). From this group two im- 

 portant lines of evolution are suggested, one leading up through 

 the Alcidae to the Laridae and Limicolae, and through the Rallidae 

 and Gruidae to the Galli, Columbae, and finally the Passeriformes, 

 while the other leads through the Procellariiformes and Ciconiiformes 

 to the Anseres and Falconiformes. While the writer does not be- 

 lieve that this is necessarily the true phylogenetic relationships of 

 the various groups, nor even contend that it is nearly the correct 

 interpretation, nevertheless with the facts now at hand, it satisfies 

 the conditions of the feather morphology better than any other 

 interpretation at present available. According to morpohology of 

 feathers there is strong evidence for the following hypotheses : 



(1) The Procellariiformes, while constituting a specialized off- 

 shoot, lie near the line of descent of the Ciconiiformes, the nearest 

 of the latter order to the primitive type being the Steganopodes, while 

 the Ardeae are the most specialized. 



(2) Flotus and Phalacocorax deserve to be separated into two 

 distinct families, the former showing remarkable likenesses to the 

 Cathartae. 



