158 CEYLON PEARL OYSTER REPORT. 



yellowish colour, aggregated in rounded masses of very varying size. These may he 

 symbiotic al^ae. 



R.N. 213, 217 (both from deep water outside pearl hanks, Gulf of Manaar) ; 247. 



Family : DESMAOtnONTP/F, 



Sigmatomonaxonellida in which some of the microscleres are cheki? (except when 

 these have been lost by degeneration). 



The presence of microscleres in the form of chelae constitutes a natural character by 

 which the Desmacidoniche are, as a rule, easily distinguished from all other sponges. 

 Unfortunately, however, the chelae are verv apt to disappear by degeneration, 

 especially in the sub-families Ectyonina? and Phlo?odictvinae, and we have then to 

 depend upon other characters such as the presence of spined echinating styli for 

 guidance in classification. That the chela, one of the most remarkable forms of 

 microsclere known to us, has originated by modification of the sigma, there can, I 

 think, be little douht. In the Report on the " Challenger" Monaxonida* we showed 

 that in ontogeny the chela arises from a sigmoid form (Esperetta mammiformis) and 

 that sigmata and chelae must therefore be grouped in the same category. Nor are 

 intermediate forms of adult spicules unknown to us, such as the curious " bipocilli " of 

 the genus Iophon, especially those of Iophon chelifer, Ridley and Dendy (1), and, 

 most notable of all, the bidentate sigmata of Topsent's Gellivs bidens (64). 



We are, therefore, justified in regarding the Desmacidonidae as derived from the 

 Haploscleridae by modification of the sigmoid microscleres into chelae, though why this 

 modification should have taken place is verv hard to understand. It is extremely 

 difficult to see how the verv peculiar and highly specialized chelate form of microsclere 

 can he of any special advantage to its possessor, and we have here one of those 

 numerous cases in which, so far as we can see at present, the theory of natural 

 selection signally fails to account for the facts. 



I have already pointed out that the sub-family Phloeodictyinae must be transferred 

 to the Desmacidonidae on account of the presence of chelate microscleres in the 

 genus Histoderma, certain species of which are obviously very closely related to 

 Phlceodictyon and Oceanapia. For the purposes of this Report the three sub-families 

 Esperellinae, Phlceodictyinae and Ectyoninae will be sufficient. TorsENT's sub-family 

 Dendoricinae appears to me to be an unnatural group which cannot he maintained, the 

 differentiation of the ectosomal megascleres being far too general and widespread a 

 character to be utilised as distinguishing the sub-family, whose members fall very 

 naturally in one or other of the remaining sub-families. The sub-family Buharinse, 

 proposed by the same author, has, I am glad to see, heen again abandoned by him in 

 his latest work (62). 



* Page xx. 



