380 Clark. The Classificaiion of Birds. [^^^^ 



These ten types are so constant and in general so easily recog- 

 nized, it seems to me they might well be made the central charac- 

 ters of ten orders ; and we should find that such orders are not 

 unnatural groups, but are characterized by many other important 

 features. For example, the birds with the charadriiform pterylosis 

 are nidifugous, schizognathous, with two carotids and aquincubi- 

 tal wings. There are, perhaps, other characters, but I have not 

 attempted to determine them all. Such a group might well be 

 called the Charadriiform es. Or the birds with the falconiform 

 pterylosis are all nidicolous, desmognathous, with aquincubital 

 wings, two carotids and epignathous, cered bill, and may well be 

 designated as Falconiformes. That it is not unnatural to associate 

 the Parrots with raptorial birds, will appear to anyone who will 

 examine Gadow's comparison of the two groups, which shows that 

 out of forty characters they have twenty-nine in common, includ- 

 ing those which seem to me must be granted to be of the most 

 significance. 



It is of interest to see how basing classification primarily on the 

 pterylosis will affect the position of certain doubtful forms. As 

 is well known, the Tinamous will belong with the Galliformes and 

 the Sand Grouse with the Columbiformes. The Flamingo is dis- 

 tinctly pelargiform. Opisthocomns is not at all galliform, but, 

 curiously enough, is quite distinctly falconiform. The Bustards 

 are clearly pelargiform and thus quite separate from the Charad- 

 riiformes. Psophia, on the other hand, is apparently more like 

 Plover than Stork. The Auks are not nearly related to the Gulls 

 and Terns, but are not far from the Petrels, and very possibly the 

 Penguins are merely a further specialized shoot of the same 

 branch. 



It would give a very wrong impression, were I to close this 

 paper without referring to any of the difficulties which lie in the 

 way of making such use of the pterylosis for a primary character, 

 as I have suggested. The number of orders would not be re- 

 duced thereby as much as was to be hoped, but ten is certainly 

 better than twenty. More important than this, however, is the 

 fact that the orders are by no means equally well-defined. The 

 Anseriformes, for example, are difficult to characterize except by 

 the very distinctive pterylosis, while the Colymbiformes and Galli- 



