AMPHISPONGIA. 155 



filiform spicules regularly arranged in the direction of the length of the sponge. 

 Neither the spicules of the upper nor those of the lower part of the sponge appear 

 to have been organically attached together. No canals can be seen, nor in the 

 majority of the specimens is there any indication of a cloacal cavity ; but in one or 

 two examples the conical spicules of the basal portion are arranged so that a small 

 tubular cavity exists in the central axis, but this does not appear to extend into the 

 upper part of the sponge. 



All the examples of this sponge are shown on fractured surfaces of a soft brown 

 micaceous shaly rock. The spicules have been completely dissolved away, and only 

 moulds in the matrix remain. These moulds are very clearly defined ; they are usually 

 empty, but occasionally they are filled with a powdery ochraceous material, probably 

 peroxide of iron. This unfixvourable condition of preservation and the close arrange- 

 ment of the spicules make it very difficult to discover their true forms. Thus Salter 

 regarded the radiating conical tubes of the lower part of the sponge as the moulds of 

 bundles of spicules, whereas I believe them to be the cast of simple conical spicules. 

 If Salter's views were correct, the separate spicules of the bundles should have shown 

 some impression in the moulds, whereas the interior of these moulds is perfectly 

 smooth and even. The occurrence of these hollow conical moulds, detached from 

 the sponge and scattered singly throughout the rock, also shows that they were filled 

 by single spicules, and not by closely grouped bundles. The smaller spicules of the 

 upper portion of the sponge usually present the appearance of so many minute linear 

 or fusiform hollows closely arranged side by side, with a minute circular hole in the 

 centre of each individual. In this position only casts of three rays are exposed, but 

 when detached from the mass, which frequently occurs, the spicules are seen to possess 

 four rays in a plane at right angles to each other, and in the meeting-point of the 

 rays a central hole is present in some instances, indicating the existence of a fifth 

 ray, and possibly a sixth ray may have projected outwards. 



Ampldspongia was regarded by »Salter and also by Bowerbank as a Calcareous 

 sponge. The supposed triradiate form of the spicules and their complete dissolution 

 supported this opinion. But the true cruciform figure of the majority of the spicules 

 and the hexactinellid characters of others, at once show the affinity of the sponge to 

 siliceous Hexactinellids. The proportions of the spicules also correspond closer with 

 those of Hexactinellid sponges than with Calcareous forms. The complete dissolu- 

 tion of the spicules is no longer an argument in favour of their original calcareous 

 nature, for we now know that fossil siliceous spicules most readily undergo disso- 

 lution and replacement. The presence of peroxide of iron filling many of the casts 

 is also in favour of the original siliceous constitution of the spicules, as this material 

 is usually present in the moulds of undoubted siliceous sponges. 



Whilst the form and proportions of the spicules in the upper portion of this 

 species clearly show a relationship to those of Lyssakine Hexactinellids, their 



x2 



