PRINCIPLES AND METHODS OF PALAEONTOLOGY. 369 



The laws of life cau only be obtained by observation of the facts of life 

 and generalization from those facts, and the philosophical naturalist 

 seeks that classilication which shall best enable him to remember facts 

 and generalizations already won, and shall most efficiently assist him to 

 obtain others. 



As Ciivier lias well expressed it, modern classification endeavors to 

 throw the facts of the structure of living beings into the fewest possible 

 general propositions. Each living being, therefore, has been compared 

 with all others, and tbose from which it is not separated by any constant 

 difference are grouped together as one '• species." The different species 

 have next been compared, and those which agree in some one or more 

 characters, wliile they differ from all others in these characters, are 

 urranged into a larger group, called a "genus." By a like procedure, 

 genera have been grouped into "families;" these into "orders," orders 

 into "classes," and classes into "subkingdoms," v/hich last are the pri- 

 mary subdivisions of the animal and vegetable "kingdoms" respectively. 



The resemblances and differences upon which the groups are founded, 

 being based on a comparison of the whole organization of living beings, 

 are thorough and fundamental, and, as it were, indicated by nature her- 

 self. Hence this mode of classification has been termed "natural," in 

 contradisfinction to those i)reviously leferred to, the divisions of which 

 are founded on insulated and superficial relations. 



But it is obvious that if animals and ]dants were not constructed upon 

 common plans, it would be impossible to throw them into groups 

 expressive of their greater or less degree of resemblance, such as those 

 of the natural classification. In fact, the doctrine of "common plan" 

 and of "natural classification" are but two ways of expressing the great 

 truth, that the more closely we examine into the inner nature of living 

 beings, the more clearly do we discern that there is a sort of family 

 resemblance among them all, closer between some, more distant between 

 others, but still pervading the whole series. 



There is yet another way in which this doctrine has been expressed. 

 In every group there is some average form, some form which occupies a 

 sort of central place, around which the rest seem to arrange themselves; 

 and this form may therefore be taken as the representative of the group, 

 as the nearest actual embodiment of the common plan. Such a form is 

 commonly called the fj/pe of the group ; and in this sense an antelope 

 might be termed the type of the Ruminanfia ; a dog of the Carnivora. 

 It is in this sense that tlie word "type" will be used in these pages; but 

 it is proper to remark that the term is not uncommonly applied to the 

 most characteristic and marked form of a group. In this sense a cat 

 ratlier than a dog would, perhaps, be selected as a typical carnivore. 



The phrase "family resemblance" has been used above, and it, perhaps, 

 expresses better than any other the sort of likeness which exists among 

 the members of a natural group; specific and generic alliance having the 

 same sort of relation as brotherhood and cousinhood. But it is import- 

 ant to remember that the classification of animals and plants stands on 

 its own basis, and is entirely independent of physiological considera- 

 tions. For the purposes of the classifier it is wholly immaterial whether, 

 as some maintain, "species" are immutable and have taken their origin 

 independently of one another, directly from the hand of the Creator; 

 or whether, as others think, they are indefinitely modifiable, and have 

 all resulted Ironi the changes induced by external influences upon some 

 common stock. If all forms of living beings were fossil, and we knew 

 nothing about life, the natural classilication of animals and plants would 

 beexactly what it is now; except as it might be affected by the resulting 

 24 s 



