ii50 PERCY SLADEN TRUST EXPEDITION 



centrum is then developed between the basal portions of the rays, and the whole spicule 

 at the same time increases in size (figs. 4 e, 4 e'). As the centrum grows larger the rays 

 grow longer and increase in number. They may still be either sharp-pointed (fig. 4/) or 

 truncated (fig. 4/'), but remain slender, so that their bases stand far apart from one 

 another on the surface of the spherical centrum. As growth proceeds the rays begin to 

 thicken and become bluntly conical. The radius of the centrum is now growing more 

 rapidly than the length of the rays, so that the intervals between the latter become to 

 a large extent filled up, the ends of the rays forming blunt conical projections on the 

 surface of a sphere (fig. ig). Short spines are now developed at the ends of the rays 

 (fig. 4 h). The spheraster has now reached what appears to be the adult condition in 

 Aurora (Stelletta) aurora Hentschel [1909]*. The adult condition in our species (fig. 4 c) 

 appears to be arrived at by further development of the terminal spines and by their 

 fusion with one another to form the enlarged extremities of the rays, and with those of 

 adjacent rays to form bridge-like connections between these extremities. There seems to 

 be little doubt that we have here a very interesting case of the recapitulation of phylo- 

 genetic history, the penultimate stage in the development of the sterrospheraster, with 

 terminally spined rays, representing an ancestral condition which is still retained as the 

 adult form in A. aurora. The oxyaster with which the development commences, and 

 the various intermediate stages, no doubt also represents ancestral adult forms. 



The sterrospheraster and its developmental stages are abundantly scattered through 

 the choanosome. The adult form is especially abundant in the cortex. 



(4) Small, irregular spherasters without any distinct rays but merely a nubbly 

 surface (fig. 4^). Varying in diameter up to about 0-0082 mm., but usually a good deal 

 smaller. These spicules look as if they might owe their origin to the separation of the 

 enlarged extremities of the rays of the sterrospherasters. They resemble these extremities 

 closely in shape, and usually also in size, though some of them are larger. I have, 

 however, found no evidence of such an origin, and they are probably independent spicules, 

 homologous with the strongylasters of Aurora aurora. They are scattered abundantly 

 throughout the sponge, but are especially numerous in the cortex. 



(5) Oxyasters (fig. 4 h) ; minute, slender-rayed, with no distinct centrum and about 

 5 or 6 rays; total diameter about 0'0165 mm. It is very doubtful whether these can 

 be sharply distinguished from the early stages of the sterrospheraster. 



The condition of the specimen does not permit of my giving a detailed account of the 

 canal-system and histology, but the following particulars may be noted. There is a thin, 

 fibrous cortex, about 0'2 mm. thick, densely packed, especially in its outer part, with 

 sterrospherasters and spherasters. This cortex is pierced at intervals by chones. Each 

 chone is divided into ectochone and endochone and the two are connected by a very 

 narrow and fairly long canal. The ectochone terminates at the surface of the sponge in a 

 single inhalant pore. The endochone opens into a subcortical crypt, from which, indeed, 

 it is not sharply distinguishable, and the subcortical crypts unite together to form large 

 irregular spaces from which the incurrent canals take their origin. The entire arrangement 

 is very similar to that figui^ed by Keller for Stelletta siemensi [1891, fig. 56]. 



* Compare especially Hentschel's figure of A. aurora var. arenosa (loc. cii., p. 363, fig. 7). 



