282 University of California Publications in Zoology (Vor. 13 
able, in some cases being scarcely noticeable, while in others it is 
very apparent. In the Steganopodes, for instance, the difference in 
feather morphology in some of the different families is very great, 
while in the families of Passeriformes, which, as a matter of fact, 
are hardly more than supergenera, it is extremely difficult to dis- 
tinguish between even widely separated ones. This difference in 
degree of differentiation also holds true for groups of higher rank. 
As intimated above, to be comparable the feathers whose parts are 
to be compared must be approximately similar, since there is fre- 
quently more variation between different kinds of feathers on a 
single body than between corresponding feathers of birds of different 
orders. For example, the barbules of a remex of Larus differ in 
their minute structure from those of a breast feather of the same 
genus far more than they differ from those of a remex of a loon, 
for instance. 
2. Classification Adopted 
The problem of what recognized system of classification to follow 
in the study of comparative feather morphology presents itself at 
this point. To the mind of the writer the system which represents 
most clearly the true relationships of birds according to the present 
status of our knowledge concerning them, and one that is coming 
into very general favor with ornithologists in this country as 
well as in Europe, is that presented by Knowlton and Ridgway 
in the Birds of the World (1909). This classification, as stated by 
Knowlton, is essentially the same as that used by Gadow (1891), 
modified in some details by the later researches of ornithological 
workers. Although this classification was adopted in the present 
study as a mere working basis, it was found that as far as feather 
morphology was concerned it is apparently a more natural grouping 
than any other; yet, as will be shown in the following pages, there 
are some possible changes in it suggested by feather structure, and 
a hypothetical revision of it, based primarily on the latter, will be 
suggested at the close of this paper. 
In the systematie discussion of the various groups, the grouping 
and suecession used by ‘Knowlton has been used with only two ex- 
ceptions. The Struthioniformes, Rheiformes, Casuariiformes and 
Apterygiformes have been included under a common heading Ratitae, 
as has usually been done, while the Crypturiformes have been dis- 
sociated from these and placed immediately after the Galliformes, 
