REPORT ON THE KERATOSA. 65 



The specimen proved to contain filaments, not being, however, particularly rich in 

 them. 



Colour. — Soft parts dirty white, skeletal-fibres brownish-yellow. 

 Habitat. — Baliia ; shallow water. 



Cacospongia collectrix, F. E. Schulze, sp. 



Oligoceras collectrix, F. E. Schulze, Zeitschr. f. wiss. Zool., vol. xxxiii. p. 34. 



The two Challenger specimens which I group in this species are so diff"erent in 

 external shape that, should we classify the Keratosa according to their form, we 

 should be obliged to refer them, at least, to difierent genera. Yet it is well known that 

 in most cases the external shape as well as the colour in the Porifera are of extremely 

 subordinate consequence, and I am the more inclined to regard both the specimens 

 in question only as individuals of the same species, as in the Dalmatian specimen of 

 Cacospongia collectrix described by F. E. Schulze we find a form of, so to speak, neutral 

 characters with respect to the differences above mentioned. These differences concern 

 the mode of growth and the quantity of foreign bodies enclosed in the parenchyma. 

 The Dalmatian specimen just alluded to is represented in Prof. Schulze's paper on 

 pi. ii. fig. 6, and one can see that is not crust-like but of massive shape. On 

 pi. iii. fig. 7, Prof. Schulze gives also a drawing of the anatomical organisation 

 of the specimen in question, and this drawing renders obvious that, contrary to the 

 specific designation of collectrix, its soft parts are comparatively free from foreign 

 enclosures. This latter character is common to the Challenger specimen of Cacospongia 

 collectrix from the Philippine Islands, while as to its external shape this specimen has 

 been found in the form of a crust. On the contrary, the second Challenger specimen (from 

 Japan) recalls vividly, so far as its exterior is concerned, F. E. Schulze's specimen from 

 Lesina, but its soft parts proved to be overloaded wdth foreign enclosures, almost 

 exclusively sand grains often 2 to 3 mm. in diameter. It is therefore plain that the 

 above differences may be important enough to justify the subdivision of the species into 

 independent varieties, but in no case of a greater consequence. All this, however, on 

 the supposition that hotli our specimens are really closely allied to Cacospongia collectrix, 

 F. E. Schulze. So far as the specimen from the Philippine Islands is concerned this is 

 beyond doubt. This specimen, together with the mussel-shells which it coats, is also 

 of a rather lunq>y, massive form ; a more attentive examination shows, however, that 

 this is due simply to the aggregation of the mussel -shells, and that the sponge itself is of 

 a well-marked crust-like appearance, sometimes as thin as a leaf of paper. Now, while in 

 some parts of it no skeleton is to be discovered at all, the others are propped by small 

 (2 to 7 mm. high and about O'o mm. thick), isolated horny trees very rich in enclosed 



(zool, CHiLL. EXP, — PAUT XXXI. — 1884.) Hh 9 



