ORGANIC EVOLUTION 265 



different appearance, and the differences between them 

 may sometimes be correlated with their geographic dis- 

 tribution, and sometimes not. 



3. The close adherence to structural type in the mem- 

 bers of a single group that is modified for great diversity of 

 habit and environment; and, conversely 



4. The superficial similarity wrought in different struc- 

 tural types, when they are modified to a common mode of 

 existence. 



5. Correlations of structure; when one part of any type 

 is modified for a different sort of life, other parts are modified 

 in harmony therewith. The foot a of figure 148 is never 

 associated with the beak b, or with any other beak in the 

 series, except with beak of the type a. This is the sort of 

 concordance that makes the interpretations of fossil frag- 

 ments possible. 



6. Vestigial structures; why should these exist at all, 

 except they be ancestral ? 



7. The tendency of all embryos to recapitulate group 

 characters. Why should such a tendency exist, but for 

 age-long heredity? 



The palaeontologic record is exceedingly fragmentary, and 

 especially lacking in the more simple forms, that would be 

 most significant to us. The phylogenetic record is broken 

 by the absence of connecting forms between the groups, 

 existing organisms being only the twigs of branches that are 

 often widely separated. The ontogenetic record is perver- 

 ted by marked departures from the original course of 

 development. But, notwithstanding these difficulties, 

 which are so great as to make it easy to err in the interpreta- 

 tion of nature's genealogies, the evidence of descent is 

 thoroughly convincing. It is the more so because of the 

 way in which each of the partial records supplements and 

 corroborates the others, and it is certainly significant that 



