REPORT ON THE DEEP-SEA KERATOSA. 59 



which contain a soft medullar or pith-substance. But a closer examination of the 

 different Stannomidse, and a comparison with the Spongelidse and Psamminidae, have 

 convinced me that the network is the hydrorhiza of a symbiotic Hydroid ; this con- 

 jecture was finally proved by the discovery of the hydranths and gonangia (figured in 

 PI. II. figs. 6, 7, &c; see their description in the Appendix). I suppose that these strong 

 chitinous tubes of Stylactis, &c, replace the absent stout spongin-fibres of the skeleton in 

 the Stannoniidse, and that the want of these latter may be supplied by the development 

 of this curious symbiosis. The same remarkable condition is found among the 

 Spongelidse, in Psammophyllum (Pis. IV., V.), which connects this family with the 

 Stannornidse. 



Canal-System. — The Stannomidae seem to agree in the essential structure of the 

 canal-system with the closely-allied Spongelidse, with which they are immediately 

 connected by the transitional genus Psammophyllum (Pis. IV., V 7 .). Below the porous 

 dermal membrane, which is very distinct in Stannophyllum and Stannarium, there are 

 usually large subdermal cavities. These communicate with the internal canal-system, 

 which is expanded, together with the symbiotic hydrorhiza, between the two parallel 

 dermal plates of these foliaceous sponges. In Stannoma, where no distinct dermal 

 membrane was preserved, the canals in the cylindrical branches form a closer network, 

 with smaller meshes. The form, size, and disposition of the flagellated chambers seem 

 to be similar to those of the Spongelidse, but only traces of them could be found ; their 

 epithelia were destroyed in the same way as the exodermal epithelium of the outer 

 surface. 



Eggs and Gastrulse. — Having convinced myself that the Stannomidae are true 

 sponges (and not " gigantic Rhizopods," as was supposed by the first observers), it was, 

 of course, very important to confirm that opinion by the authentic demonstration of 

 eggs, and if possible gastrulse. For a long time I looked in vain for them, but at last I 

 was fortunate enough to find them in a single specimen of Stannophyllum globigerinum, 

 apparently better preserved than the others. After having stained it with carmine and 

 dissolved the calcareous pseudo-skeleton in hydrochloric acid, I found scattered here and 

 there in the maltha single amoeboid cells with a large vesicular transparent nucleus and a 

 dark nucleolus. The largest were so similar to the usual naked sponge eggs (especially 

 to those of Psammophyllum, PI. V. fig. 5, e, and of Spoyigelia), that I had no doubt as to 

 their egg nature, the more so as a few eggs were found in segmentation. Finally, some 

 larger dark ovate bodies, composed of granular cells which were found in the same speci- 

 men, may be its gastrula larvse ; they were, however, not sufficiently well preserved to 

 allow of a detailed description and drawings. 



