298 DISCOVERY REPORTS 



B.M. 87. 5. 2. 116, takes the form of a number of dark brown fragments of a low- 

 growing sponge with the surface thrown into numerous folds, resembling the adult form 

 of /. pattersotii. The main acanthostyli have the same form as those of the 

 juvenile stages of the same species (Fig. 23 c), while the echinating acan- 

 thostyli have much the same form as those of the adult forms (Fig. 21). 

 The tornota are truncate at the ends where they are beset with a few 

 spines (Fig. 22 c). Chelae are rare, but the characteristic trifoliate, oc- 

 casionally bifoliate, bipocilla are present in abundance (Fig. 24 c). The 

 second specimen, B.M. 87. 5. 2. 158, from Prince Edward Island, of 

 which nothing remains except a microscopic preparation, has the same y^ 



skeleton as the foregoing except that the anisochelae are more numerous 

 and the acanthostyli slightly longer. The next specimen, B.M. 87. 5. 2. 

 157, the "young specimen" from the Cape of Good Hope (fide Ridley jX 



and Dendy 1887, p. 120), is an extremely thin crust growing on a piece 

 of coral. In this the main acanthostyli are slightly longer than in the 

 lectotype and are strongly spined throughout their length, more particu- 

 larly at the base and at a point just below the apex (Fig. 23 b). Echinating pjg 2 1 . lophon 

 acanthostyli are present and the tornota have an oval head at each end proximum (Rid- 

 beset with numerous spines (Fig. 22 6). The latter, in addition to being ley): atypical 

 scattered among the meshes of the main skeleton, also form a tangential ^f ma ingacan- 



° -11 thostyle, x 400. 



skeleton in the dermis, where they are thickly packed but not arranged 

 in a definite reticulation. Anisochelae, measuring o-o 12-0-021 mm. long, are present, the 

 larger being arranged in rosettes. The bipocilla bifoliate, but occasionally trifoliate with 

 the smaller end multi-dentate, are sparingly present (Fig. 24 b). 



Lambe's (1893) specimens from Vancouver appear to be almost identical with the 

 lectotype, while Thiele's (loc. cit.) diff'er apparently only in the smaller size of the 

 acanthostyli. In both cases, however, no echinating acanthostyli were recorded. 

 /. chelifer ostia-magna, Wilson, from the Galapagos Islands, difi^ers in the larger size of 

 the acanthostyli and tornota. Again no echinating acanthostyli are recorded. The ex- 

 ternal form, moreover, is lamellar. The same external form is found in a group of 

 specimens presented to the British Museum by the late Dr Gilchrist, B.M. 04. 12. 1. 109, 

 from the Natal coast, some of which are massive with low lamellate outgrowths and some 

 purely lamellate. 



Among the Discovery sponges are a number which approximate to the type of 

 /. chelifer and must be considered in relation to it. In fact, despite obvious diff"erences, 

 it can be shown that they are actually conspecific with it and that these differences are 

 merely fluctuating variations, perhaps ecological in origin. 



These specimens vary from encrusting to irregularly massive, massive and spreading, 

 sub-clathrate or flabellate (and erect). The surface in the massive specimens may be 

 smooth, ridged, meandrine, plicate or bearing small flabellate processes. The colour 

 is always dark brown and the texture friable. The skeleton is a sub-isodictyal 

 reticulation of acanthostyli, echinated at the nodes by smaller acanthostyli, and 



