PALAEONTOLOGY AS A DISCIPLINE. 47 



that because a number of allied groups display a certain struc- 

 ture, their common ancestor must also have possessed it. 

 This may have been the case, but it is almost as likely not to 

 have been, because the structure in question may have been 

 many times independently acquired. While the comparative 

 method frequently enables us to discriminate between the two 

 classes of phenomena, it generally does not do so, and it never 

 can give entire certainty upon this point. 



On comparing the humerus of the horses with that of the 

 camels, we find in each a characteristic difference from other 

 artiodactyls and perissodactyls and agreement with each other, — 

 a feature which may be described in brief as the duplicity of 

 the bicipital groove and presence of a bicipital tubercle. It is 

 a priori probable that such an isolated resemblance between 

 two widely separated groups is due to convergence, and yet 

 the comparative method can give us no assurance that this is 

 not a primitive ungulate character retained in these two series 

 and lost in the others. Having recovered the various extinct 

 genera of both these phyla, we may trace out the gradual trans- 

 formation of the humerus and definitely show that the resem- 

 blance has been independently acquired at a comparatively late 

 period and is not a case of a persistent primitive feature. 



In short, the difficulty of reaching firmly fixed conclusions 

 upon questions of homology and relationship by the exclusive 

 use of comparative anatomy lies in the fact, that this method 

 deals only with the modern assemblage of animals, a mere 

 fragment of that which has existed in former times. It is like 

 attempting to work out the etymology of a language which has 

 no literature to register its changes. 



The second method of morphological inquiry, embryology, 

 has had a somewhat chequered career. Not many years ago it 

 was universally regarded as the infallible test of morphological 

 theory and the principle that the ontogeny repeated the 

 phylogenetic history in abbreviated form was accepted, almost 

 without question, as a fundamental law. But this view has 

 fallen somewhat into discredit. The admission which very 

 early had to be made, that " cenogenetic " features of develop- 

 ment were imposed upon or substituted for those due to 



