SALPA IN RELATION TO EVOLUTION OF LIFE. 205 



Among these are some, the narcomedusse and trachomedusae, 

 in which the hydra larva completes its individual development and 

 grows up into a sexual medusa, and others, the anthoraedusae and 

 leptomedus?e, in which there is polymorphic differentiation into 

 hydranths, and sexual medusse or gonophores. The presumption 

 in morphology, whenever there is nothing which demands the 

 contrary, is that the simple is older than the complex, and we should 

 therefore be justified in holding that the craspedota without poly- 

 morphism, where each hydra grows up into a medusa, are more 

 primitive than those with polymorphism, even if there were no 

 positive evidence, and I shall now show that this view is supported 

 by positive proof. Omitting hydra, which is too aberrant and too 

 much isolated to be available for comparison, the hydroids are, 

 almost without exception, larval or sexually immature, and the 

 reproductive persons are all medusse, or else gonophores which 

 show by their structure that they are degenerated medusae ; even 

 hydra itself is probably a hydroid with gonophores in an extreme 

 stage of degradation. The affinities of hydra are doubtful, however, 

 but in most cases we may state definitely that the hydroid is larval 

 in its sexual nature ; and all the hydroids which we know must be 

 traced back to an ancestor with a larval hydra stage and an adult 

 sexual medusa stage ; for we must attribute to inheritance from a 

 common ancestor all they have in common except what can be shown 

 to be due to secondary modification. 



The question of the origin of the craspedota then narrows itself 

 down to this ; was the sessil mode of life, and the habit of forming 

 cormi which has resulted from it, acquired before or after the 

 evolution of the medusa ; was the hydra-like ancestor of the hydro- 

 medusae, a solitary pelagic animal as a sessil cormus. 



In discussing this question we must keep in mind the following 

 considerations ; first, that the history of the polyps and acraspeda 

 has no bearing on the subject, since these animals belong to another 

 line of descent ; secondly, we have the very remarkable fact that 

 the sex of all the progeny of a hydroid egg is usually fixed from 

 the start, and is either exclusively male or exclusively female. The 

 egg of a hydractinia may give rise to hundreds of hydranths and 

 machopolyps before it produces blastostyles, and these grow up 

 before they produce medusa buds ; the number of medusa buds 



