386 University of California Publications in Zoologij f Vol. 13 



scales of reptiles and on the hair of mammals shows promise of 

 bringing out facts concerning them also, which will he of taxonomic 

 value. But in all of these the range of possible or probable mocli- 

 fiability is very slight as compared with that of feathers, on account 

 of the much greater complexity and minute structural units of the 

 latter, and therefore the morphology of feathers is of greater value 

 than lepidology or any other branch of epiphyology from a taxo- 

 nomic point of view. 



2. Principal Modifications of Structure Useful in Taxonomy 



There are a great many different parts of feathers and plumage 

 which show phylogenetic modifications, among which some are of 

 great value and can almost certainly be depended upon to be of 

 importance in showing relationships, while others are as plainly of 

 very doubtful value. From the foregoing systematic study of the 

 different groups of birds, the relative value of the different structures 

 and arrangements of different parts of feathers has made itself 

 apparent, and the following general conclusions may be drawn. 



The distribution of plumules is a character which, in itself, is of 

 little value, on account of its great variability, in a few cases, within 

 a single group, and on account of the adaptive value of the plumules, 

 as shown by their presence in all the lower orders of water-birds and 

 their recurrence in kingfishers amongst a group in which there is 

 elsewhere a tendency for plumules to be reduced or entirely lost. 

 However, in conjunction with the condition of other structures, the 

 distribution of plumules is of some phylogenetic importance. 



The aftershaft is of more importance, and its presence or absence, 

 and form if present, may be depended upon to a considerable extent 

 as showing phylogenetic tendencies. 



The quill, even of the remiges, is too variable within groups to be 

 of any great value except in a few cases, as, for instance, in the 

 i Anseres, where the disproportionate length of the calamus is of diag- 

 nostic value. The condition of the ventral groove and of the stria- 

 tions on the sides are of little value except in a few cases. 



The rami have a number of characters which are significant, 

 namely, the thickness of the pithy portion, whether of one or more 

 layers of cells (see Mascha, 1904), the width relative to the shaft, 

 the relative size and form of the ventral ridge, and the presence or 

 absence of villi on its ventral edge in the outer vane. 



The distal barbules of the remis'es have a number of characters 



