2 58 COMPARATIVE ANATOMY 



stand in much closer relationship to one another than classes 

 in another phylum. If there were a complete genealogical 

 record it would be possible to make the various grades of 

 classification correspond to degrees of relationship; but, as 

 the record cannot be complete, such a correspondence is 

 impossible, and we can only deal according to the best of 

 our judgment with the evidence at our disposal. A classifica- 

 tion, then, can only be regarded as an approximate representa- 

 tion of animal affinities ; but, such as it is, it should represent 

 the fullest extent of our knowledge. Classifications must 

 change as knowledge advances. 



Digitized by Microsoft® 



