642 COMPARATIVE ANATOMY 



neurenteric canal of the embryo. The facts of ontogenesis, therefore, 

 according to Delsman, support his theory that the neural tube is the 

 homolog of the stomodeum of annelids. (Fig. 55) 



In one of his preliminary papers, Delsman (1913) makes a suggestion 

 of vertebrate ancestry which will appeal to some who remain unconvinced 

 by the theory just outlined. This suggestion is that the characteristic 

 differences which separate phyla, such as the chordates, annelids, and 

 molluscs, may have arisen by changes in the region of growth by which a 

 larval form like the trochophore is converted into an adult. Morphology 

 has tended to lean heavily upon evidence from comparative anatomy 

 for phylogenetic conclusions, especially in the absence of decisive paleonto- 

 logical evidence. But the growing conviction that evolutionary progress 

 has been dependent primarily upon changes in the chromatin of germ 

 cells has made speculations which assume the evolution of one phylum 

 from another by the transformation of highly differentiated adults seem 

 less probable than they did to earlier morphologists. The suggestion 

 that three great phyla may have arisen from a simple larval form like a 

 trochophore, therefore, sounds attractive. Delsman suggests that, just 

 as the elongated annelid develops from the trochophore by the elongation 

 of the region posterior to the blastopore separating mouth and anus as 

 far as possible from one another at opposite ends of the body, precisely 

 so molluscs and chordates may have originated from a trochophore-like 

 ancestor by change of the growth center. Most molluscs grow, not 

 antero-posteriorly, but dorsally. In them, the mouth and anus remain 

 relatively near one another, while an elongated dorsal sac is formed, in 

 which the intestine and liver are bent into a U-shaped tract. Again, 

 by the elongation of the region anterior to the blastopore so as to separate 

 this from the anterior end of the body, a chordate form like the larval 

 amphioxus is produced. Such assumptions as these avoid the pitfalls 

 which are met in attempts to convert a complex adult of one phylum into 

 that of a different phylum. Furthermore, they seem to accord with 

 recent discoveries in genetics and embryology. If, then, this suggestion 

 of Delsman is actually in accordance with the facts of ontogenesis, it 

 follows that the divergence of the chordate phylum from the other inverte- 

 brate phyla, occurred before the annelid phylum made its appearance. 

 But Delsman himself appears to have abandoned his attempt to derive 

 chordates and other phyla from a trochophore larva and to favor the 

 opinion that the adult annehd is in the direct line of chordate ancestry. 

 (Figs. 521, 523) 



THE ARTHROPOD THEORY 



The old theory of Geoffrey St. Hilaire (1822) that an arthropod might 

 be converted into a vertebrate by reversing dorsal and ventral sides and 



