1897- MORPHOLOGY IN ZOOLOGICAL SCIENCE. 23 



enough to start with, since there are many specific distinctions far less 

 obvious than many individual variations. Many specific characters 

 are doubtless merely associated with other characters which really 

 determine the survival of their possessors, but this is true to a less 

 degree of generic characters, and as we pass over to family and 

 ordinal groups we get on safer and safer ground in assuming that we 

 are dealing with features directly due to natural selection. 



Why then has this feeUng of dissatisfaction with the method 

 of morphology grown up ? The main reason is, I think, a con- 

 viction that it proves too much. The most discordant views as 

 to the relationship of animals and their ancestry have been drawn 

 from the same facts, and there does not seem to be any court 

 of appeal before which rival views can be brought. Then again, 

 continued study has forced home the conviction, that the processes 

 of evolution are much more complex than was at first imagined, and 

 that so far from being a simple process from the less to the more 

 differentiated, the converse, viz., degeneyation or simplification of struc- 

 ture, is also going on ; further, that similar structures are sometimes 

 independently developed along different lines of descent, in virtue of 

 what was called " parallel " or " convergent evolution," but termed by 

 Professor Lankester homoplasy. Now the discovery of the great 

 principles of degeneracy and homoplasy, whilst it explained many 

 points, has caused considerable doubt as to the certainty of 

 morphological reasoning. For really, when armed with the principle 

 of progressive degeneracy as well as with that of progressive differ- 

 entiation, there is no limit to the powers of the evolutionary theorist ; 

 one can derive literally any one animal from any other by first deleting 

 all the obnoxious organs in the supposed ancestor, and then evolving 

 any number of new ones. Then again, if similar structures may have 

 been developed independently in two different stocks by the action of 

 similar external conditions, where is one to draw the line ? How far 

 is one justified in relying on similarity in structure as a criterion of 

 community of descent at all ? These, I think, are some of the 

 questions which have underlain the feeling of scepticism as to the 

 value of morphology which has crept over many zoologists, and which 

 have caused, I confess, much trouble and distrust in my own mind. 

 I now venture to suggest ways of looking at morphological facts, 

 which seem to me more fruitful than the ordinary methods, and which 

 have given me fresh hope in the pursuit of morphological study. 



First, it is a mistake to assume that in tracing a supposed line of 

 descent we are at hberty to assume that any conceivable variation 

 may have occurred, variations, for instance, the utility of which we are 

 not bound to explain. I think if we take specific and generic 

 distinctions as our units, we are bound to show that some parallel 

 change to the one we postulate has in all probabihty taken place. 

 Thus there has been an immense amount of discussion as to how the 

 pentadactyle limb was derived from the fin of a fish ; but no one, so far 



