ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 367 



cesses may be regarded as a forerunner of metameric segnaen- 

 tation (Fig. 15). 



Fib. 15. — Diagram of an Annelid (J), Mollusc (B), primitive Annelid (C), 

 and Coelenterate (Z>) larva. 



To showj further, that a reasonable explanation of some of 

 the salient points in the embryology of segmented animals 

 may be given on the above view, without resorting to impos- 

 sible gastrulas, I suggest the following. Since in Coelente- 

 rates mesoderm cells are constantly forming from the ectoderm, 

 it is probable that their growing tentacles contain at least 

 some mesoderm and endoderm cells not derived from the 

 primitive layers. If such a tentacle were transformed into the 

 body of a segmented animal, its elongation and the formation 

 of the layers in the embryos would naturally take place by the 

 rapid division of large terminal cells like those found in all 

 rapidly elongating organs of both plants and animals, and, 

 indeed, just as it probably takes place in the Coelenterate ten- 



