XXII.] CLASSIFICATION. 289 



others, again, by tlie discovery of larval forms. And, with Many 

 the clearer understanding of affinities, we have come to see i^^^eu d\s-^ 

 that some apparent affinities are not real affinities at all, covered, 

 but merely adaptive modifications. Tims the whale tribe apparent 

 are not, as they appear on a superficial view to be, a group |i"ks have 

 connecting the Mammalia with the fishes, but a group of not to be 

 Mammalia modified for an aquatic life, and having no re- ^'^'^ °"'^'^* 

 semblance to fishes except a merely external one. Such a tribe. 

 conclusion as this, though merely negative, is as important 

 as any positive conclusion : perhaps more so, for the pur- 

 pose of clearing away difficulties and anomalies in classi- 

 fication. But it is likely that oiu' organic classifications 

 will ever remain incomplete and in some degree provi- 

 sional, because the materials of our knowledge are incom- 

 plete. Much has no doubt been lost that cannot be re- 

 covered. It is probable that whole classes once connected I^ost links, 

 the lower animal forms with the Yertebrata, which have 

 died out during geological time, and, being without shells 

 or skeletons, have left no fossil remains. And, what is 

 quite as important in accounting for apparent gaps in the 

 series, it is likely that forms that once underwent a meta- 

 morphosis have ceased to undergo it, and have come to be 

 developed by a direct process. Had the Cirrhipedes lost 

 their metamorphoses and come to be developed directly from 

 the egg, we should probably never have guessed how 

 closely they were connected with the Crustacea ; and had 

 they been known only by fossil shells, their resemblance to 

 the MoUusca, which is only external and adaptive, would 

 certainly have been misunderstood, and would have caused 

 them to be mistaken for molluscans. But, though I do 

 not believe that all the gaps in our classifications will ever 

 be filled up, I believe it will before long be generally 

 recognised that affinities are divergent and re-divergent 

 without ever reuniting ; as they ought to be on the 

 hj'pothesis that true classification is genealogical, and that 

 the true bond of affinity is community of descent. 



It is to be observed, that on the genealogical theory of Affinity 

 classification, affinity by descent does not necessarily in- from re-*^ 

 volve resemblance ; and the groups which are most nearly semblance. 



u 



