THE RECAPITULATION THEORY 351 



ancestral for all Metazoa. The difficulty arising 

 from its varying mode of development in different 

 forms is however still unsolved, and embryologists 

 are not yet agreed whether the invaginate or delami- 

 nate form is the more primitive. In favour of the 

 former is its much wider occurrence ; in favour of 

 the latter the fact that it is easy to picture a series 

 of stages leading gradually from a unicellular proto- 

 zoon to a blastula, a diblastula, and ultimately a 

 gastrula, each stage being a distinct advance, both 

 morphological and physiological, on the preceding 

 stage ; while in the case of the invaginate gastrula 

 it is not easy to imagine any advantage resulting 

 from a flattening or slight pitting in of one part 

 of the surface, sufficient to lead to its preservation 

 and further development. 



Of larval forms later than the gastrula, the most 

 important by far is the Pilidium larva, from which 

 it is possible, as Balfour has shown, that the 

 slightly later Echinoderm larva, as well as the 

 widely spread Trochosphere larva, may both be 

 derived. Balfour concludes that the larval forms 

 of all Ccelomata, excluding the Crustacea and verte- 

 brates, may be derived from one common type, 

 which is most nearly represented now by the Pili- 

 dium larva and which " was an organism something 

 like a Medusa, with a radial symmetry." The 

 tendency of recent phylogenetic speculations is to 

 accept this in full, and to regard as the ancestor 

 of Turbellarians and of all higher forms, a jelly-fish 

 or Ctenophoran, which in place of swimming freely 

 has taken to crawling on the sea-bottom. 



