28 PORIFERA. III. 



Finally I may direct attention to the fact, that I have in the present work placed the 

 genus Hymedesmia in the subfamily Ectyoninae, which perhaps might seem to go against the views 

 given above about the relationship of this genus to the group treated here; yet this is not so, since 

 firstly, the subfamilies Mycalinae and Ectyoninae, as I have already mentioned earlier, are scarcely 

 quite natural, and even if so it was not impossible, that genera, which on account of their characters 

 must now be placed in the Mycalinae, should be thought to have originally been derived from 

 Ectyouine forms. 



In the first part of this work I broke up Carter's Phloeodictyinac and placed its two genera 

 Phloeodictyon and Oceanapia in the Renierinae and Gelliinae respectively. In a work published in 1905 

 (Rep. of the Pearl Oyster Fisheries of the Gulf of Manaar, Part III, 165) Dendy keeps the subfamily 

 Phloeodictyinac under the family Desmacidonidae, and he refers to it the genera Phloeodictyon, Oceana- 

 pia, Histoderma, Sideroderma and Amphiastrella. His reasons for keeping these forms together 

 are mainly the same as those of Carter on founding the group, viz. the hard dermal layer 

 and the presence of fistulae, but on account of the chelae in Histoderma Dendy now removes the 

 group to the Desmacidonidae. In Part I of this work pag. 56 57 I gave the reasons which seemed to 

 me to necessitate the breaking up of the Phloeodictyinae, and the same reasons are still valid. While 

 Dendy thus lays stress on the outer shape and the presence of a solid dermal layer as the distinguishing 

 characters, but pays no attention to the spicules, I on the contrary follow the opposite way and take 

 first, as the most important character, the spicules and what may be deduced from them with regard 

 to the relationship of the forms. Dendy would also have difficulties in delimiting his subfamily; 

 Phloeodictyon and Petrosia are, as I have already declared (Part I, 1. c.), nearly related; Thiele says 

 (Zoologica XXIV, 2, 1899, J 9) tnat Phloeodictyon ( Rhizochalina) medium is an intermediate form; 

 Melonanchora, which has a solid dermal layer and papillae should, I think, also be referred to the 

 Phloeodictyinae and the same holds good with regard to several HymedcsHiia-spzcizs; also some species 

 of Gellins should be placed here according to the views of Dendy. On the other hand, I think that 

 the spicules give good hints towards a natural grouping; the oxea present in Phloeodictyon are spicules 

 typical for the Homorrhapidae ; Oceanapia and the species of Gellius, which are provided with a solid 

 dermal layer, have oxea of quite the same kind, and these together with their microscleres refer them 

 to the Heterorrhaphidae. The facts are quite otherwise with regard to the spicules in the four genera 

 in question; their spicules are diactinal (when fully developed), but of quite another type from those 

 occurring in the Homorrhaphidae and Heterorrhaphidae. That this is the case is shown, and beyond 

 doubt, by the development, since while the spicules in Homorrhaphidae and Heterorrhaphidae are really 

 diactinal, and also originate in this form, the spicules in the four genera mentioned are on the 

 contrary only secondary diactinal, but originate as monactinal; they have thus a development quite 

 as the dermal spicules in other Myxilleae, as I have described in Part II of this work, pag. 125. I 

 therefore take it to be quite certain, that these genera belong to the Myxilleae, and that their spicules 

 answer to the dermal spicules in the more typical Myxilleae, but here form the whole skeleton, of 

 which the interior skeleton however is generally weak. The facts present in Melonanchora emphysema 

 seem to me in the highest degree to confirm this view. The occurrence of chelae also shows that 

 these genera have nothing to do with Phloeodictyon or Ocratiapia ; that the chelae may sometimes 



