ARRANGEMENT AND CLASSIFICATION 57 



ledge of afiSnities. Since these affinities are the actual work of nature, 

 no naturahst will ever be able or indeed desire to alter the consequences 

 of a recognised affinity. The general arrangement will thus become 

 ever more perfect and less arbitrary, according as our knowledge 

 of affinities becomes greater. 



The case is different with classifications : that is to say, with the 

 various hues of demarcation that we have to draw at intervals in the 

 general list both of animals and plants. In truth, so long as there are 

 gaps remaining to be filled in our list owing to many animals and plants 

 not having yet been observed, we shall always find these Unes of 

 demarcation, which appear to be drawn by natvire herself ; but this 

 illusion will vanish as our observations accimiulate. Have we not 

 already witnessed the effacement of a great number, at least in the 

 smaller divisions, by reason of the numerous discoveries of naturalists 

 during the last half century ? Except for the hues of demarcation 

 resulting from gaps to be filled, those which we shall always have to 

 draw wiU be arbitrary and therefore changeable, so long as naturaUsts 

 do not adopt some conventional principle for their guidance. 



In the animal kingdom such a principle is that every class should 

 comprise animals distinguished by a special system of organisation. 

 The strict execution of this principle is quite easy, and attended only 

 with minor inconveniences. 



In short, although nature does not pass abruptly from one system 

 of organisation to another, it is possible to draw boundaries between 

 each system, in such a way that there is only a small number of animals 

 near those boundaries and admitting of doubt as to their true class. 



The other fines of demarcation which sub-divide classes are usually 

 more difficult to estabfish, since they depend on less important 

 characters ; and for this reason are more arbitrary. 



Before examining the true classification of animals, let me endeavour 

 to show that the list of fiving bodies should form a series, at least as 

 regards the main groups ; and not a branching net-work. 



Classes, should FpRM a Series in the Arrangement 

 or Animals. 



Man is condemned to exhaust all possible errors when he examines 

 any set of facts before he recognises the truth. Thus it has been denied 

 that the productions of nature in each kingdom of living bodies can 

 really be arranged in a true series according to their affinities ; and 

 that there exists any scale in the general arrangement either of animals 

 or plants. 



Naturafists, for instance, have noticed that many species, certain 

 genera and even some families appear to a certain extent isolated 



