THE president's ADDRESS. 41 



Cinachyra vaccinata there is a correlation between the modifi- 

 cations of the three kinds of triaene. 



Such variations as these strongly suggest the existence of 

 definite factors in the germ plasm, upon which the modification 

 depends. It appears also that in some cases a factor may influence 

 in the same kind of way the form of more than one type of spicule. 

 On the other hand, the facts with regard to the meristic varia- 

 tion of the triaenes suggest that each ray may be influenced by 

 a separate factor. 



The factorial hypothesis appears to be further supported by 

 certain very definite conclusions at which we are arriving as 

 to the general course of evolution followed by the Tetrax- 

 onida. On the whole, evolution seems to have been pro- 

 gressive, accompanied by increasing complexity of structure, 

 manifested especially in the skeleton. Along certain lines of 

 descent, however, it appears that the culminating point has been 

 passed and regressive evolution has taken place, resulting in 

 simplification of structure. This simplification seems to have 

 been effected by the complete dropping out of certain types of 

 spicule. This is strikingly exemplified by the case of the great 

 family Stellettidae and its derivatives. The progressive evolu- 

 tion of the stellettids was marked by a strong development of 

 triaenes of various kinds, accompanied by a beautiful series of 

 astrose microscleres, many of which have very remarkable and 

 easily recognisable forms. In his report on the Tetractinellida of 

 the Challenger Expedition Professor Sollas proposed the family 

 Epipolasidae for a number of species which differ from the 

 stellettids only in the total absence of triaenes. We know now 

 that this group is not a natural family, but consists of a hetero- 

 geneous assemblage of stellettids which have independently lost 

 their triaenes, and in some cases we can tell with certainty, by 

 the characteristic form of the asters, from what particular stel- 

 lettid genus an epipolasid species has been derived. We can even 

 do more than this, for in the genus Aurora there are three per- 

 fectly distinct triaene-bearing species each with a closely corre- 

 sponding epipolasid species. The genus Chondrilla, again, ap- 

 pears to have arisen from^stellettid ancestors by loss of all the 

 megascleres, and, finally, in Chondrosia, microscleres as well as 



