BRITISH SPONGE OF THE GENUS RAPHIODESMA. 103 



them, we trace the connection of genera, and thus extend the 

 genealogy. Raphiodesma is a progression, perhaps, we may say, to 

 a higher development, from the simpler structm-e seen in Hymenia- 

 cidon. In this latter, there is no sort of regularity in the arrange- 

 ment of spicules ; but, in the former, there is a beginning at least 

 of order in the fasciculi and their disposition, which naturally leads 

 on to another higher development in the genus Desmacidon. Now, 

 in following out this natural progression, it is interesting to observe 

 that we carry with us the same character of spicules. 



We have seen how, in Raphiodesma, there is an attempt to form a 

 more regular skeleton than that of Hymeniacidon, but that the net- 

 work is never connected. In the next genus, Desmacidon, we find 

 this done and perfected. Let us now see which of the species allies 

 itself closest with the genus in question. 



Desmacidon consists of a fibrous network, well-marked and 

 defined, " composed entirely of spicules arranged in accordance 

 with the axis of the fibre, cemented together, and thinly covered with 

 keratode." Out of twelve species which Dr. Bowerbank has 

 tabulated, six present us with forms of spicules associated together 

 precisely like those in Raphiodesma. This is remarkably shown in 

 D. copiosus, where are two forms of sub-clavate spicules in the 

 skeleton, as in R. sordida, small tricurvate and also two kinds of 

 anchorate spicules very closely resembling those of the latter 

 species, as well as the bihamate form. So that the two species of 

 two dififerent genera are in this particular nearly identical. In D. 

 cegagropila there are the same general forms, but without the tri- 

 curvate. D. constrictus presents us with the same forms as the latter, 

 though with a remarkable development in the anchorate spicules as 

 well as in those of the skeleton. D. similaris has acuate skeleton 

 spicules, anchorate tricurvate and bihamate spicules ; D. Peachii 

 the sub-clavate spicule again with a divergent form of the bihamate, 

 &c. Lastly, that most interesting species, D. rotalis, agrees in 

 possessing similar bihamate and anchorate forms with an acuate 

 skeleton spicule.* 



Thus, I think, we clearly perceive in these three genera a pro- 

 gressive development from a lower to a higher type, if we can so 

 call that, which has a more complete network in a more definite 

 order ; and this progress, associated with similar forms of spicules, 



* For these details I must refer to Dr. Bowerbank's " Spongiadae," Vol, 

 iii., Plates Ixxii., pp. 265, 357; Ixiii., Ixxxiii., Ixxi., p. J 83; Ixxxix., p. 319; 

 Ixiii., xc, p. 327. 



