Ixiv THE VOYAGE OF H.M.S. CHALLENGER. 



The Heterorrhapliidae constitute, perhaps, the least natural of our four families. In 

 many respects they occupy an intermediate position between the Homorrhaphidae and the 

 Desmacidonidge. As their name implies, they are possessed of different kinds of spicules, 

 as opposed to the Homorrhaphidas, in which we never find more than one kind. Micros- 

 clera are present in the group, but these are never of the chelate type. The five sub- 

 families, viz., Phloeodictyinse, Gelliinse, Tedaniinse, Desmacellinse and Hamacanthinse, into 

 which the Heterorrhaphidse are divided, are not so intimately related inter se as are 

 either the two subfamilies of the Homorrhaphidae or of the Desmacidouidae. 



The Phloeodictyinse form a compact and very remarkable group. Both in external 

 form and in anatomical characters they appear to be strongly contrasted with other 

 subfamilies of the Heterorrhaphidse. The first known specimen of the group was dredged 

 by the Rev. A. M. Norman off the Shetland Islands in 1861. It was evidently only a 

 fragment, and was named by Bowerbank Isodictya rohusta} In 1864 Mr. J. G. 

 Jeffreys dredged another specimen, also off the Shetland Islands, which was much 

 mutilated by the dredge, and was named by Bowerbank Desmacidon Jeffrey sii.^ Later 

 discoveries showed that these two fragments belonged to one and the same species, of 

 which the Rev. A. M. Norman was the first to describe an entire specimen. He 

 established the genus Oceanapia for this interesting form.^ 



In 1870 Schmidt founded his genus Rhizochalina* the type species being Rhizo- 

 clialina oleracea. Although in the present Report we keep these two genera distinct, yet 

 we have very strong reasons for believing that they ought to be united, and then the 

 genus Oceanapia would be the sole representative of the subfamily. Finally, in 1882, 

 Carter established ^ his " Group " " Phlceodictyina," and included therein nine species, all 

 partaking more or less of the characters of the original Oceanapia rohusta. Unfor- 

 tunately the material at our disposal is not in very good condition for anatomical 

 investigation, but an examination of thin, stained sections has led us to believe that 

 future researches, under more favourable circumstances, will lead to very interesting 

 results. 



Concerning the Gelliinse we have very little to say. The subfamily is characterised 

 by the extreme simplicity of the spiculation, which in the case of the genus Gellius is 

 identical with that of the genus Oceanajna. This identity might be thought to indicate a 

 close relationship between the two subfamilies. Possibly it does indicate some connection, 

 but we do not think that this can be a very intimate one, for, although the Gelliinse 

 present us with great diversity in external form, yet we know no instances of anything 

 like an approach to the highly characteristic form of the Phlceodictyinge ; in other words 

 the sponge is never divisible into body and fistulas, and this is a very important 



1 Mon. Brit. Spong., vol. ii. p. 304. ^ Mon. Brit. Spong., vol. ii. p. 347. 



3 Brit. Assoc. Report for 1868, p. 334. * Spong. Atlant. Gebiet., p. 35. 



' Ann. and. Mag. Nat. Hist., ser. 5, vol. x. p. 117. 



