270 DE. P. CHALMERS MITCHELL OX THE 



Some parallels may make the complexity of the problem of the caeca apparent. The 

 Passeres display all diets and one type. 



The Owls and the Falconiformes have almost identical diet, the larger forms of both 

 being carnivorous, the smaller forms chiefly insectivorous ; the Owls all exhibit apocen- 

 tricity A, the Falconiformes apocentricity B. 



Fowls and Pigeons (except the fruit-eating Pigeons) have a similar diet ; Fowls all 

 exhibit apocentricity A ; Pigeons apocentricity B or rarely C. 



The Gulls have a similar diet consisting of fish and garbage ; most exhibit apocentricity 

 B, but Stercorarius is archecentric. 



On the other hand, Phcenicopterus is distinguished from other Ciconiiformes by a 

 nearly purely vegetable diet, and it has the longest caeca in the group. The Mergansers 

 are distinguished from the other Anatidse by their typically piscivorous diet, and they 

 alone among the Anatidte have apocentricity B as in most other fish-eaters. 



CHARACTERS AND CLASSIFICATION. 



In the systematic descriptive part, my task was to treat the characters of the patterns 

 displayed by different birds as nearly as possible as if the gut were the whole animal, and 

 the various phylogenetic figures and the three Plates display what I take to be the 

 relations of the intestinal tracts, and not necessarily the relations of the possessors of 

 these tracts. I have been taking, in fact, the anatomical structure as the unit, and not 

 the individual or the species. In a sense, this is a return to the old Hunterian method ; 

 but its purpose reflects on the new problem of the nature and evolution of varieties and 

 species of genera and families. Granting that the Plates attached to this paper represent 

 with approximate accuracy the phylogeny of the intestinal tract in birds, we have yet to 

 learn the relation of the phylogenetic tree of this structure to the phylogenetic trees of 

 other structures, and the relation of all these to the phylogenetic trees of these im- 

 permanent combinations of characters that we call species. Although the coincidence of 

 such trees is frequently assumed, there is no a priori reason to support such a proposition ; 

 and there is much recent work on the nature of characters and of their inheritance to 

 throw doubt on the proposition. The nature of the anatomical structure in any organism 

 depends in the first place on the nature of certain material transmitted from the parents 

 in the fertilized ovum (naturally it does not matter to the argument whether the trans- 

 mitted substance be what we call "matter" or "state of matter" such as mode of 

 rhythm). Among other writers de Vries (6) has recently brought strong experimental 

 evidence to show that at least in plants the hereditary material is composed of inde- 

 pendent units which may be sorted out and recombined in each sexual generation. Of 

 such independent variables underlying the fully developed anatomical structures of 

 animals, we know practically nothing as to their number, nature, or modes of sorting 

 out and recombination as they pass from generation to generation. Next, the anatomical 

 structures of animals depend on the environment in which the combination of transmitted 

 units come to maturity as actual adult organs ; that is to say, they depend on the 



