39 



shaped processes", another "consists of hollow, finger-shaped processes or "fistulae" rising from 

 a sandy base to a total height of about 50 millims". In these terms typical specimens of trop. 

 digitata are very well characterised (PI. I, fig. 6 — 9). We have about the same features in 

 5. congenera, S. punctulata and S. solida. The surface is generally more or less shining, not 

 hispid, sometimes more even, another time somewhat corrugated vertically. In several respects 

 such specimens which I bring together as trop. digitata resemble typical specimens of trop. 

 pyramidalis. In the former the processes are on the whole more prominent, being longer; in 

 the latter we can distinguish one or two chief processes as a pyramidal cone with a varying 

 number of much smaller, secundary ones (PI. III, fig. 1). The shape of the processes in trop. 

 digitata is generally distinctly digitate-, in pyramidalis both primary and secundary processes 

 tend to lateral compression. The specimens 426 a and 426 b show this very plainly. In 426 c 

 there are less processes — only a few just at the base. In 1045 there are no secundary pro- 

 cesses at all. The characteristic pyramidal shape we find also in some specimens of Stibcritcs 

 wilsoni e. g. B. M. 22 and in Spirastrella areolata (B. M. 2) ; these are, like 1045, destitute or 

 almost destitute of secundary processes. 



Ridley & Dendy described a sponge, which they called Spirastrella papillosa. They 

 say (1887 p. 232): "Sponge.... massive, erect, sessile; shape conical, broad at the base and 

 tapering gradually to an obtuse apex, where are situated several large oscula .... The entire 

 surface, except in the immediate neighbourhood of the oscula, is covered with numerous, closely 

 placed papillae of considerable size . . . ." At the end of the description the authors state: "the 

 peculiar warty appearance of the surface seems to be the most characteristic feature of this 

 species...." We have seen that the specimens 92 a-1 and 426 c! are identical with it. Whereas the 

 Challenger dredged only a single specimen the Siboga brought home more than a dozen and so we 

 have a better conception of the real character. They are all massive, but they are by no means 

 always broadest at the base. Specimen g2 c (PI. III, fig. 5) is more or less oval; g2 e and g2 1 

 have a more conical shape. Not unfrequently they have more than one top (92 a, 92 b, g2 d, 

 g2i and g2j)- The greater part of the surface is covered with papillae or tubercles, giving to 

 the sponge the warty appearance. However, the various specimens possess this feature in different 

 degrees-, thus in 92b and g2d the turbercles are comparatively low and few; in 92c and 92e 

 they are high and numerous. Spirastrella cunctatrix B. M. 13 externally resembles the Siboga 

 specimens; likewise some parts of B. M. 10 (PI. III, fig. 3). Intermediate between these distinctly 

 warty specimens and typical specimens of trop. digitata are 614 a and 614 b. But also in another 

 respect we find transitions between them. The surface in trop. digitata often exhibits rugae, 

 which we do not find typically in trop. tuberosa. But it must be remembered that Carter says 

 (1886 p. 115) of S. cunctatrix var. porcata (B.M. 1) that the surface "presents a number of 

 thick rugae running from the base towards the apex . . . ." Indeed, the external appearance of 

 S. cunctatrix B. M-. 1 is quite different from that of S. cunctatrix B. M. 13- But in S. cunctatrix, 

 B. M. 10 (Pk III, fig- 3) we find the characters of both united. 



However different such specimens like g2c (PI. III, fig. 5), 426 b (PI. III, fig. 1), g48 c 

 (PI. I, fig. 7), and ig45 (PI. III, fig. 2) may be, we learn by the above statements that they 

 gradually pass into each other. The original crust, by becoming thicker and enveloping its 



