316 INVERTEBRATE MORFHOLOGT. 



sexually mature male and female individuals which, since they differ from 

 the parent form in possessing more highly developed eyes as well as a 

 more perfect adaptation of the parapodia for swimming, probably leave the 

 sponge and swim about freely in the ocean distributing their sexual elements. 

 If this be the case, this species presents both colony formation and alterna- 

 tion of generations, the latter phenomenon being also manifested by other 

 species of 8yllis and by Autolytus, in which buds are produced linearly, 

 differing from the parent in the structure of the parapodia and separating 

 to lead a free existence as male and female individuals. A modification of 

 this process is found in certain species of Nereis, in which at the time of 

 sexual maturity the posterior segments of the body develop setse more 

 perfectly adapted for active locomotion by swimming than were those of 

 the immature form ; these sexually mature forms were at one time 

 referred to a separate genus, Eeteronereis. 



The Phylogeny of the Polychata. — The origin of the Polychseta has been 

 within recent years the subject of considerable discussion. The discovery 

 of the wide distribution of the Trochophore larva led to the supposition 

 that it was an ancestral form, from which the Polychaeta had been de- 

 veloped by a process of linear budding, each metamere of the Polycheet 

 body being equivalent to the original Trochophore and the adult organism 

 being a co-ordinated succession of Trochophore individuals. Other authors, 

 however, who do not see iu metamerism the result of a budding of the 

 individual, but rather the multiplication of its subordinate parts, are in- 

 clined to refer the Annelids to a Nemertean-like ancestor and to consider 

 the Trochophore larva a purely secondary adaptation. Between these two 

 views it is difficult to decide, and it is possible that in their plain statement 

 neither is quite correct, though each may contain certain elements of 

 truth. 



It seems exceedingly probable that a larval form which is met with in 

 the life-history of the Annelids and MoUusca, as well as in a modified form 

 in other groups, has some ancestral significance. It is diflicult on any 

 other hypothesis to explain its occurrence in widely different groups, since 

 it seems hardly probable that it could have arisen independently in several 

 instances. Convergent evolution could hardly be carried to such an ex- 

 tent as to produce in the MoUusca, quite independently of any genetic 

 relationships, a larva resembling in all its structure that of an Annelid. 

 If the Trochophore occurred only in the Annelids, it might be quite possible 

 that it had made its appearance in the life-history of some primitive 

 Annelid as a secondary modification of a more primitive larva, and had 

 reappeared subsequently in the life-history of all forms descended from this 

 Annelid ancestor, but this would not explain its occurrence also in the 

 MoUusca, unless it be supposed that the members of this group have been 

 derived from the primitive segmented Annelid, a view that has little to 

 recommend it. The working of the biogenetic law (see p. 143) is interfered 

 with in innumerable instances, and the distinguishing between examples of 

 its action and secondary modifications is the most difficult task of the 



