THE EVOLUTION OF ANIMAL FUNCTION 331 



which appears at a comparatively early stage. Yet we cannot 

 fail to see that there is some truth in the Biogenetic Law. On 

 the whole the successive stages of ontogeny do in a rough way 

 follow the lines of phylogenetic development, though there are 

 many short cuts and modifications. It would seem then that 

 the underlying fact, on which the Biogenetic Law rests, is 

 simply that modifications introduced into the course of ontogeny 

 do not always affect all stages of the process, but tend to appear 

 rather in the later stages than in the earlier. This is probably the 

 form in which we must accept the law for our present purpose. 



If we now return to the difficulty under consideration we 

 shall find that this statement may, with reasonable probability, 

 be of real assistance in the inquiry. If every modification 

 introduced into ontogenetic development necessarily affected 

 all stages of the process, we should be in no better position^ 

 after examining all stages of ontogeny, than we are after 

 examining only its final result. We should only find in different 

 species a number of wholly different ontogenies between which 

 the connection of succession would not be traceable, just as we 

 have a number of different results of ontogeny. If, however^ 

 some modifications appear at a comparatively late stage of onto- 

 geny, we have groups of ontogenies whose earlier stages are 

 alike though their later stages are different. As soon as we begin 

 to recognise such groups we are getting free from the bewilder- 

 ing state of things in which the functional capabilities differ 

 with every species. We may even, to put the matter somewhat 

 hopefully, find such groups co-extensive with different phyla 

 of the animal kingdom, though the adult functional capabilities 

 found within any one phylum are differentiated one from 

 another to such an extent that their inclusion in one group 

 on their own merits would not suggest itself. If we can only 

 recognise such groups as these, we shall be on the road towards 

 such a classification of functional ontogenies as will bring us 

 into touch with those broad facts of evolutionary succession 

 which are tolerably ascertained on morphological grounds. In 

 this way and for these reasons, I believe that we are likely to 

 learn far more of the past evolution of any functional capability 

 by examining its ontogenetic development in comparatively few 

 species, than we could hope to know if we were acquainted in a 

 vastly greater number of species with the final condition which 

 functional capability reaches in the adult organism. 



