THE ANCESTRY OF THE VERTEBRATES 537 



an unrelated form should, as an adaptation, so modify its 

 early stages as to resemble these echinoderm larvae as closely 

 as does the Tornaria, and the only alternative is to accept as a 

 very ancient common ancestor of both echinoderms and verte- 

 brates the form which all these larvce may be said to copy; a 

 form having the characteristics common to all, including bi- 

 laterally, minute size, transparency, locomotion by bands of 

 cilia, and pelagic life. The lineal descendants of this hypotheti- 

 cal ancestor chose two paths, the one leading to the Echino- 

 dermata, the other to Balanoglossus, the Tunicata, Amphi- 

 oxus, and eventually the Vertebrata. 



This theory, although incomplete and unsatisfactory in 



B 



FIG. 150. Comparison of Tornaria and Echinoderm larvae, lateral 

 views. [After BALFOUR.] 



(A) Tornaria. (B) Auricularia. (C) Bipinnaria. 



a, apical area; b, oral area; c, post-oral area; d, anal area. 



parts, is consistent with the most approved lines of biological 

 thought; it rests upon development as well as adult structure, 

 and bears the indorsement of the majority of investigators 

 at the present time. The weakest part of the argument is that of 

 the significance of the Tornaria larva ; and while the acceptance 

 of this gives us very little enlightenment, to abandon it would 

 be to sacrifice but little, and would render the gulf between 

 the adult Balanoglossus and other invertebrates only a little 

 more profound. To summarize in the words of two recent 



