On the Classification of Animals. 267 



were adhered to. Lamarck, who stumbled upon the theory 

 of evolution, but was unable to marshal a sufficient array of 

 facts to support it, still seems to have believed in the pos- 

 sibility of arranging animals in a linear series, and, it may be 

 added, that at the present day this old notion has been 

 saddled to the theory of descent by those who have not been 

 carefully initiated into the mysteries of biology. The pos- 

 sibility of arranging animals in a linear series is, however, no 

 longer admitted by naturalists — they know that the relations 

 animals have to each other cannot be represented on a plane 

 surface, but only in space of three dimensions ; that in fact 

 the only possible method of classifying animals is to arrange 

 them in divergent groups and redivergent sub-groups, or, in 

 other words, to represent them in the form of a tree. Further, 

 various groups, e.g., orders, families, genera, and species, are 

 no longer looked upon as being always equivalent, but rather 

 as arbitrary divisions adopted for convenience, and not because 

 they have the same comparative value. If a natural system 

 of classification is founded on descent with modifications, it 

 must be strictly genealogical in order to be natural, hence 

 the various groups (orders, families, genera, etc.) must be 

 considered as expressing the amount of difference amongst 

 organisms which have descended from a common ancestor, 

 and which together form a branch or phylum ; and it may be 

 added, that the possibility of dividing groups into orders, 

 genera, etc., has resulted from members of these groups 

 undergoing elaboration and differentiation in different direc- 

 tions. This being so, the particular rank which any number 

 of individuals receives will depend partly on the amount of 

 elaboration and partly on the disappearance of the immediate 

 or remote ancestral forms, or of both. 



From what has been said, it will be evident that in drawing 

 up a system of classification it is necessary to distinguish 

 adaptive from essential characters, and to remember that no 

 arrangement can be natural that is not genealogical. The 

 safest characters therefore will be those that show an affinity 

 between organisms to each other or to an ancestral form. 

 When arranging animals into large groups, we should therefore 

 notice especially the developmental history. This, however, 



