34 THE ORIGIN OF THE NERVOUS SYSTEM 



gradient. But as the time of attachment of the planula 

 approaches, this gradient gradually becomes less and 

 less distinct, and the planula attaches itself by the end 

 which was originally apical, a fact which has long 

 been known. After attachment the original gradi- 

 ent disappears completely, a new gradient in the oppo- 

 site direction arises at the opposite end, and the first 

 hydranth arises from the high end of this new gradient. 

 In other words, the phenomenon which Loeb and others 

 have called heteromorphosis is a normal feature of 

 hydroid development. Again, in the polyclad Turbel- 

 laria the early stages of development show a very distinct 

 anteroposterior susceptibility gradient, the head arising 

 from the most susceptible region, but in the later stages, 

 so far as they have been examined, a reversal in the 

 susceptibility relations occurs, at least in the superficial 

 regions, and in the adult worms the ectoderm of the 

 head is less susceptible than more posterior levels. 

 These data on hydroids and polyclads have not yet 

 been published in full. In the annelids a second gradient 

 in the opposite direction from the primary antero- 

 posterior gradient appears, at least superficially, in the 

 posterior regions as the result of the development of 

 the posterior growing region (Hyman, 1916; Child, 

 1 91 7<f). But, whatever the changes, they are definite 

 and orderly and associated with the course of develop- 

 ment in each species. 



As development and differentiation progress the 

 indications of more or less specific relations between 

 particular regions or organs and particular agents 

 become more frequent. In Planaria dorotoccphala, for 

 example, we find that the lateral margins of the body 



