268 Proceedings of the Royal Physical Society. 



is not enough in the case of vertebrates, e.g., it will be 

 necessary further to take into consideration the most fixed 

 structures — the skeleton and nervous system, the generative 

 organs, the heart and respiratory organs, and any rudi- 

 mentary or atrophied organs that may be present. When 

 breaking up the large groups into sub-groups, it will be 

 necessary to direct attention to the amount of elaboration 

 that has taken place, or to the specialisation any series of 

 organs has undergone, or to the appearance of new organs. 



Although the system of classification at present generally 

 in use is to all intents and purposes a natural system, and 

 although it does not fail to recognise in a general way the 

 genealogical affinities, it must be admitted that it is at the 

 best a most disjointed system, and that the definitions are 

 often drawn up in such a way that they seem to apply 

 equally well to several distinct forms or groups. How then 

 might the present arrangement be improved ? How can the 

 relationships be better expressed and the definitions made 

 more definite ? A system of classification ought to 

 enumerate all the information we possess about the various 

 groups and individuals composing them. This was the case 

 with the systems of Linnaeus and of Cuvier. Since their 

 time, however, we have not only collected an enormous 

 number of facts as ta the structure of animals, but we have 

 learned much of their distribution, development, and ancestral 

 history. Hence a classification now must not only express 

 all we know as to the structure, but also all we know as to 

 the ontogeny and phylogeny of the organisms considered. 

 The most suggestive and useful classification ought naturally 

 to assume the form of a genealogical tree, and the definitions, 

 instead of merely enunciating what are supposed to be the 

 ordinal generic and specific characters, ought to express as 

 far as possible the relationship indicated by this tree — to 

 interpret or explain these relationships, and at the same time 

 suggest the amount of elaboration and specialisation, or it 

 may be the amount of degeneration that has taken place. 

 Our knowledge of extinct forms is not yet sufficient to enable 

 us to follow out this ideal system ; but it may be possible, 

 without waiting until geologists fill up the many gaps, to 



