Oil the <( Glass-Rope " Hyaloncriia. 89 



pointed projections, usually turned towards the point of the 

 needle, rarely bent backwards like barbs from either end 

 towards the centre. In all these awl-shaped spicules a delicate 

 central canal is very apparent, and, in all of them, at or near 

 the middle of their length, one or two fine cross canals cut the 

 central canal at right angles, exactly as in the long needles of 

 the coil. When the cross canals have an appreciable length, 

 two or four slight bulgings on the outer surface of the needle 

 indicate their position (Plate, Fig. 5). 



From this form we pass by an easy transition to a second 

 class of very generally distributed spicules. In cases where a 

 single cross canal only is developed, this canal has become 

 produced into two arms at right angles to the original spicule, 

 and the primary and secondary branches together form a 

 cross (Plate, Fig. 8). When two transverse canals are 

 produced, a star of four secondary branches cross the main 

 shaft, giving origin to the remarkable six-spoked forms (Plate, 

 Fig. Gj. The larger of this group are usually nearly smooth, 

 but very minute spicules of the same type, with all the branches 

 feathered and their ends curved (Plate, Fig. 10), are very 

 common, clustering in groups round the larger styles. Fre- 

 quently one of the halves of the needle is undeveloped, and a 

 form is produced in which the single long shaft, represent- 

 ing one half of the original spicule, stands out from a whorl of 

 four transverse branches ; all the rays are feathered. Such 

 spicules line in large numbers the internal cavities and passages 

 of the sponge. The cross is placed against the bounding felt 

 of spicules, and the long barbed ray projects into the space, 

 apparently to prevent the entrance of foreign matters (Plate, 

 F.g. 2). 



The most remarkable spicules are the two forms repre- 

 sented in Plate, Figs. 3 and 11. The larger of these (Plate, 

 Fig. 3) consist of a strong shaft roughly tuberculated, 

 with a well-marked central canal showing the characteristic 

 cross processes. From either end of the shaft from seven to 

 nine long teeth curve gracefully backwards, ending nearly 

 opposite the centre of the shaft in fine points. These singular 

 spicules are most abundant in the cortical layer, though they 

 are found likewise scattered irregularly through the substance 

 of the sponge. The other set (Plate, Fig. 11) are exceed- 

 ingly minute, only to be detected by a power of about 300 

 diameters ; they resemble those just described in general 

 form, but the recurved hooks are united by a kind of silicious 

 web, beyond which the point of the hooks project slightly, 

 so that the expanded ends of these spicules are singularly like 

 umbrellas. They resemble most remarkably in form and size 

 the " amphidisci " of the gemmules of Spongilla ; they are not, 



