22 Art. 1.— N. Annandale and T. Kawamura : 



In summer the phase of S. dementis that occurs on stones in 

 the lake is a somewhat peculiar one, more massive than is usual in 

 the species and having large round oscula, the main exhalant canals 

 connected with which are usually of small calibre and some- 

 what deeply buried. The surface is often somewhat nodular. 

 but there are no branches. Towards the end of summer 

 or in early autumn the sponge begins to increase its superficial 

 area by spreading out in a thin layer. At the same time slender, 

 elongate branches begin to grow out from the surface, while the 

 oscula formed in the peripheral, younger parts are smaller, and the 

 canals opening into them more superficial than in the more mas- 

 sive central region. 



Changes of a similar nature, but much less pronounced, occur 

 in all the commoner, shallow- water species so far as the external 

 appearance of the sponge is concerned, and it is possible that the 

 dwarfed phase of S. semispongilla is no more than an arrested sum- 

 mer form. 



It is in 8. dementis, however, that the most remarkable sea- 

 sonal change has been observed. In this species it is the structure 

 of the skeleton rather than the external form that is affected, the 

 winter growth having a much more diffuse and open network 

 than that produced in summer and autumn. In many sponges of 

 this species the change seems to be gradual and specimens are to 

 be found, especially in the latter part of autumn, that have the 

 laxer skeletal structure throughout their substance. Specimens also 

 occur, however, in which the change is quite abrupt and the sponge 

 can be seen even by the naked eye to consist in vertical section of 

 two quite distinct layers, the uppermost of which has the skeleton - 

 fibre much more widely separated than the lower layer (PI. I, Fig. 

 8). Such sponges have only been found in winter. As sexual 

 reproduction takes place freely in this species from about the middle 

 of September to the middle of October, it seems probable that the 

 majority of the parent sponges die and disintegrate after breeding, 

 that the young sponges which grow up from larvae in autumn do 

 so with considerable rapidity and have a delicate skeleton consisting 

 of comparatively few spines, and that the sponges which do not 



