672 R. KIRKPATRICK. 



are raostl}' scattered at the surface, iu which they lie hori- 

 zontally. 



The oval rings found deeper down in the sponge usually 

 have thinner rims. In one instance six rings followed at 

 equal intervals on one side and five on the other side of the 

 mass of sponge filling an open crypt. Hence I called the 

 upper part of the sponge Noronha scalarif ormis. 



The very fine primitive simple sigmata are commonly found 

 in the neighbourhood of the flagellated chambers. 



There seems to be no transition between the sigmata and 

 the clavidiscs. At the same time the clavidiscs probably 

 developed from some such form. In one or two of the 

 Tnyocytes acting as sphincters round the apopyles there seemed 

 to be an appearance of a slender curved axial rod of silex. 

 Possibly the slender sigmata may originally have come into 

 existence owing to the presence of sphincters, which surround 

 not only the pores and oscules, but also the apopyles of the 

 flagellated chambers. 



To sum up, normally there are five kinds of spicules in 

 Merlia, viz. tylostyles, long rhaphides, trichodragmata, clavi- 

 discs, and slender sigmata. Rarely thicker sigmata and toxa 

 occur. 



PI. 35, figs. 11-15 show abnormal forms of spicules, all fountl 

 in one specimen. 



(2b) The Calcareous Skeleton. 



When a living sponge is taken from the water presently the 

 semi-transparent, red, fleshy surface sinks a little, and 

 the porcelain-white calcareous skeleton becomes visible. 

 Under a lens it is possible to see the very minute circular or 

 polygonal meshes of a fine network, and the still more minute 

 tubercles rising from the nodes. In dried specimens the 

 flesh forms merely a thin yellowish pellicle, covering the 

 surface, which has a uniformly granular aspect due to the 

 tubercles below, with here and there a polygonal pattern 

 where the soft tissues have sunk more deeply into the spaces 

 of the skeleton beneath. 



