vin CLASS I PICA TION 1 1 5 



more than a means by which the distinction between certain 

 objects and certain other objects could be grasped by the mind 

 of the founder and communicated to others. Classifications 

 based on such characters became so general, and were often so 

 carelessly and ignorantly put together, that they threatened to 

 bring the whole system into disrepute ; whereas I have little 

 hesitation in saying that, in reality, classification is one of the, 

 if not the, most important aims and ends of the study of 

 morphology. It is the best contribution which we can make 

 towards the solution of the great biological problem, for a true 

 classification, viewed by the light of the derivative hypothesis, 

 is nothing more or less than an expression of the actual 

 amount of affinity between different objects. An order, a 

 family, or a genus is no longer a group of animals linked 

 together by some arbitrarily selected characters, but a group 

 supposed to have been descended from a common ancestor, and 

 to have become, by whatever process, gradually differentiated 

 from other groups of animals. 1 



As in such groups, when once established, there can be 

 no crossing with other groups (as in human families, to which 

 they are sometimes compared), all the resemblances which are 

 found between members of different groups must either be 

 characters inherited from the common ancestor, perhaps lost 

 through many generations and reappearing at a subsequent 

 period by the process of " reversion," or they must be characters 

 having merely analogical resemblances i.e. characters devel- 

 oped by variation, but which, owing to similarity of conditions 

 of existence or other causes, have become similar to each other. 



To discriminate between these two classes of characters, 

 namely, those that are essential and fundamental, or, in the 

 language of one school, are dependent on conformity to type, 

 or, according to the derivative hypothesis, are inherited from 

 remote ancestors, and those that are modifications to suit the 



1 In examining into the validity of the derivative hypothesis, much is to be 

 expected from the study of the geographical distribution of animals and plants, 

 both in present and in former times ; but such study will be quite in vain unless 

 morphologists have first determined correctly the affinities of the animals and 

 plants treated of. It is obvious that all inferences from geographical or 

 palseontological research are useless, unless the classification on which they are 

 based is at least approximately true to nature. 



