350 THE RECAPITULATION THEORY 



mutual relations of the large groups of animals, or 

 of the points at which these diverged from one 

 another. 



In attempting to interpret these early ontoge- 

 n&icjstages as actual ancestral forms, beyond which 

 development at one time did not proceed, we must 

 keejTcIearly in view the various disturbing causes 

 whiclTtend to falsify the ancestral record ; such as 

 the influence of food yolk, or of habitat, and the 

 tendency of diminution in size to give rise to 

 simplification of structure, a point of importance if 

 it be granted that these free larvae are of smaller 

 size than the ancestral forms to which they corre- 

 spond; If on the other hand, in spite of these 

 powerful modifying causes, we do find a particular 

 larval form occurring widely and in groups not 

 very closely akin, then we certainly are justified in 

 attaching great importance to it, and in regarding it 

 as^ having strong claims to be accepted as ancestral 

 for these groups. 



Concerning these larval forms, and their possible 

 ancestral significance, our knowledge has made no 

 great advance since the publication of Balfour's 

 memorable chapter on this subject ; and I propose 

 merely to allude briefly to a few of the more strik- 

 ing instances. The earliest, the most widely 

 spread, and the most famous of larval forms is the 

 gastrula, which occurs in a simple or in a modified 

 form in some members of each of the large animal 

 groups. It is generally admitted that its signifi- 

 cance is the same in all cases, and the evidence 

 is very strong in favour of regarding it as a stage 



