160 ESSAY ON CLASSIFICATION. Part I. 



parts ; before we can combine families into natural groups, we have to make a 

 thorough investigation of their Avhole structux'e, and compare it ■\vith that of other 

 families. So form is characteristic of famihes ; and I can add, from a careful investi- 

 gation of the subject for sevei'al years past, during which I have reviewed the whole 

 animal kingdom with reference to this and other topics connected with classifica- 

 tion, that form is the essential characteristic of famihes.^ I do not mean the mere 

 outline, but fonn as determined by structure ; that is to say, that famihes cannot 

 be well .defined, nor circumscribed within their natural limits, -nnthout a thorough 

 investigation of all those features of the internal structure wliich combine to deter- 

 mine the form. 



The characteristic of the North American Chelonians which follows, may serve 

 as an example how this subject is to be treated. I will only add here, that how- 

 ever easy it is at first, from the general impression made upon us by the form 

 of animals, to obtain a glimpse of what may fairly be called famihes, few inves- 

 tigations require more patient comparisons than tliose by which we ascertain 

 the natural range of modifications of any typical form, and the structural features 

 upon which it is based. Comparative anatomy has so completely discarded every 

 thing that relates to Morphology ; the investigations of anatomists lean so uniformly 

 towards a general appreciation of the connections and homologies of the organic 

 systems which go to build up the body of animals, that for the purpose of imder- 

 standing the value of forms and their true foundation, they hardly ever afford any 

 information, unless it be here and there a consideration respecting teleological rela- 

 tions. 



Taking for granted, that orders are natiu-al groups characterized by the com- 

 plication of their structure, and that the different orders of a class express the 

 different degrees of that comphcation ; taking now further for granted, that families 

 are natural groups characterized by their form as determined by structural pecu- 

 liarities, it follows that orders are the superior kind of division, as we have seen 

 that the several natural divisions which are generally considered as orders, contain 

 each several natural groups, characterized by different forms, that is to say, con- 

 stituting as many distinct families. 



After this discussion it is hardly necassary to add, that families cannot by any 

 means be considered as modifications of the orders to which they belong, if orders 

 are to be characterized by the degrees of comphcation of their structure, and families 



' These investigations, which liave led to most Dr. A. A. GoulJ, and which I would not allow to 

 interesting results, have delayed thus far the publi- appear before I could revise the whole animal king- 

 cation of the systematic part of the Principles of dom in this new light, in order to introduce as much 

 Zoology, undertaken in common with my friend, precision as possible in its classification. 



