Vol. XVlin Clark, The Classification of Birds. 37 1 



do not correspond with the orders of other classes, being based on 

 far less important structural conditions. That there is good 

 ground for such belief is shown by the fact that the class Crustacea, 

 with as many species as Aves, is very generally grouped in a dozen 

 orders or less ; the Gastropod Mollusks, with nearly fifty per cent 

 more species than Aves, are almost universally included in three 

 orders ; .while the schemes for avian classification contain twenty^ 

 orders or more, or, worse still, are divided into gejis, super-orders, 

 super-families, and other indefinite groups which, to an elementary 

 student only makes " confusion worse confounded." That this is 

 a real difficulty in giving ornithology its proper place in a course 

 of zoology, other teachers besides myself can testify. 



The cause of this trouble, it seems to me, is to be found in the 

 importance that has been placed on characters which are by no 

 means fundamental in the structure of birds. Originally the orders 

 of birds were based on characters of the bill and feet ; but it was 

 long ago recognized that those characters are very unreliable, be- 

 cause so readily modified according to habits and food. In seek- 

 ing more stable characters, ornithologists turned to the skull and 

 other parts of the skeleton, the muscles, the wings, and even the 

 viscera. But as our knowledge of avian anatomy has increased, we 

 have been forced to admit that in all these points, changes of habit 

 are soon followed by changes of structure, and it becomes a mat- 

 ter of great difficulty to trace real relationship. Owing to the 

 large number of possible combinations of characters, which orni- 

 thologists regard as of more or less importance, the comparatively 

 homogeneous group of birds has been split up into numberless 

 orders. The remedy is to be found in a rearrangement of avian 

 characters, with a careful estimate of their relative value, so that 

 those that are least liable to change shall be accorded the most 

 weight. In Gadow's well-known scheme for the classification of 

 birds, published in 1892, he made use of more than forty characters, 

 to determine the mutual relationship of the groups. A careful ex- 

 amination of this list shows a very large number which are of 

 slight value because of their marked tendency to be easily modified, 

 while others are omitted which ought to be of great value because 

 of their slight tendency to vary. For example, there are no char- 

 acters of which use is made, connected with the reproductive, ex- 



