3§ 



with a thickness of about 5 cm. The sponge crust is 1 — 10 mm. thick and leaves less surface 

 of the substratum free than 1455- Specimen 1421 consists of three or four pieces evidently broken 

 from one piece. It is again growing on coral or coralline algae, but the substratum in many 

 places disappeared, probably being destroyed in the way we savv in 1455. The result is that 

 in some parts a kind of trabeculae is formed, which however clearly shows to have originally 

 been incrusting, the diameter of the trabeculae like the crust measuring 2 — 5 mm. (PI. V, 

 fig. 14). Only in this way can I explain the aberrant shape of this specimen, which corresponds 

 remarkably well with the "variety" (B.M. 38) Ridley & Dendy describe (1887 p. 229). They 

 say, that the Challenger dredged a fine specimen, "consisting of a number of stout, irregular, 

 anastomosing trabeculae, forming together a sessile, cavernous mass . . . ." It is clear that this 

 specimen 1421 is no more a mere crust; in 1402 the process is gone still farther. Both in 1402 

 and 1421 the first traces are visible of free elevations or papillae. They gradually lead to 

 specimens like 300 (PI. I, fig- 1) 410, 1092, 1391 by evolution of two things: becoming thicker 

 and producing processes. Very remarkable in this respect are 1075 and 1403 (PI. I, figs. 8 and 

 5 respectively). Only small portions of the crust are visible, but the processes are comparatively 

 large; in 1075 one of these (broken off) measures 3 to 4 cm. (PI. I, fig. 9). We see at the 

 same time that the shape of these processes is rather variable. Lindgren's S. aurivillii has 

 much the same appearance. Still a step farther from the original crust are 614 a and 614 b. 

 They represent irregular lumps; but in the whole appearance of the processes 614 a-b resemble 

 300, 410 etc. The specimens 300, 410 and 1945 all come from the same locality ("Saleyer", 

 station 213); on the other hand 1455, which externally and anatomically is so alike 1945, 

 is found "off North Ubian". Judging from the external appearance the above mentioned 

 specimens are obviously most closely allied. Anatomically we see, however, considerable diffe- 

 rences : large robust spinispirae in 1945, none in 300 ; a thick, shining pseudo-cortex in the 

 latter, nothing of it in the former etc. etc. We have demonstrated, however, that all this 

 does not give us the right to distinguish species. Schmidt says (1864 p. 36) that " Suberites 

 bistellahis" forms incrustations, resembling leaves, u deren Nervatur oft tauschend durch das 

 oberflachlich verlaufende Wassergefasssystem dargestellt wird". Lendenfeld (1897 P- 55) has 

 likewise observed these superficial, tangential canals and states: "Die Haut, welche diese Kanale 

 bedeckt, ist auch im lebenden Schwamme haufig etwas eingesunken ; in Weingeistexemplaren 

 ist dieselbe so stark versenkt, dass man an ihrer Stelle tiefe Furchen sieht". I found the same 

 in a specimen from Naples (described in MS.) and also in 1031- In other specimens nothing of 

 this arrangement is visible. 



Specimens like e. g. 1092 and 1391, forming cakes or cushions with a tendency to produce 

 conical or digitate processes most gradually lead to typical specimens of trop. digitata, such 

 as 948 a-f and 5. vagabunda. In 948 c e. g. the processes have about the same size and appearance 

 as the larger one in 1075 (Compare PI. I, figs. 7 and 9). Ridley (1884 p. 468) says of 5". vaga- 

 bunda: "massive, attached by broad base, tending to grow up into large nodular elevations.... 

 General surface slightly verrucose (in spirit), more so in large dry specimens". Dendy has shown 

 afterwards (1905) that the sponge is exceedingly variable in every respect. As to the external 

 one of his four "varieties" "consists of a massive base rising up into a few short, stout, finger- 



