670 



R. KIRKPATRICK. 



slender tylostyles^ with pointed ends outwards, along with a 

 few slender rhaphides, which form more or less vertical pillars 

 of support round the large ectosomal spaces and incurrent 

 canals. In contracted sponges one to four of these vertical 

 wisps are drawn down into the upper crypts, but in expanded 

 specimens the wisps are outside the crypts and form supporting 

 pillars to the ectosome and canals (PI. 32, fig. 10, and PL 

 33, fig. 3). Sometimes a bundle of spicules lies transversely 

 on the floor of an upper open crypt. Only very rarely are 

 spicules of any kind found in the lower ciypts, but neverthe- 



Text-fig. 2. 



^iJl^'hr^' 



Sections of masses of ci-yjjt-tissue showing siliceous spicules. 

 a. An elongated calcocyte in neck of crypt. Soft crypt-tissue 

 mostly disintegrated owing to insufficient fixation. X 22.5. 



less they do occur tliere. (The probable reason for this rarity 

 is explained in Section 5). 



Numerous microscleres in the form of oval rinars — for 

 which I propose the name " clavidiscs "- — are scattered about 

 on the surface, and also, but much less abundantly, deeper 

 down. 



' In the report on the " Discovery "' Tetraxonida I have used the 

 term "tyle" in place of "tylostyle," because it was short, and by way 

 of antithesis to " amphityle, " but I now return to the commonly used 

 designation " tylostyle." 



* Clavis, key, referring to the key-hole notches ; discus, quoit. 



