The Sponges of Lake Biwa. 23 



breed or which survive 1 »reeding continue their growths on similar 

 lines to their or their neighbours' autumnal offspring. 



It is not possible to lay down any definite rules as to the dates 

 on which these seasonal changes occur. It is probable, indeed, 

 that they are influenced not only by actual temperature, which of 

 course varies from year to year, but also by other factors in the 

 environment and possibly by individual idiosyncrasy on the part of 

 the sponge. Generally speaking, however, it is clear that, just as 

 the seasons are somewhat later in Japan than they are in northern 

 and central Europe, so the fresh- water sponges are somewhat later 

 in attaining their full development, which rarely occurs in L. Biwa 

 before September or the latter part of August. 



This lateness of growth is illustrated in particular by the pro- 

 duction of gemmules, if a general statement is permissible — for 

 here again other questions but those of temperature and season in- 

 trude themselves. All the species, moreover, do not produce 

 gemmules simultaneously. Speaking quite generally, it is 

 safe to say that Ü. lacuslris is as a rule the latest to do so, and S. 

 semispongitta perhaps the earliest. We no' iced in the autumn of 

 1915 that gemmules were already being produced commonly in the 

 latter species as early as the beginning of October, whereas as 

 late as the beginning of November there was as yet no 

 rudiment of them visible with the aid of a hand-lens in 

 S. lacustrls. Another specific difference of the same kind was 

 demonstrated in a remarkably clear manner in the small 

 lake at Komatsu on the west side of L. Biwa at the latter date. In 

 this lake, which is connected with the 'main body of water by a 

 narrow channel, we obtained numerous specimens in which either 

 S. lacustrls or S. fragills, or the two species together, coated the 

 stems of reeds (PI. I, Fig. 10). While in S. lacustris there were 

 no gemmules and vegetative growth was evidently still active, S. 

 fragllis invariably contained a large number of these reproductive 

 bodies and in some cases was already completely degenerate, the 

 sponge-substance in extreme examples having almost completely 

 disappeared, leaving a pavement-layer of gemmules with remain? 

 of the bare skeleton clinging to the reed-stem. 



