GEODINELLA ROBUSTA. 209 



In the spicule-preparations I have seen a few fragments which appeared to 

 be parts of spicules much thinner and hjnger than the rhabds and the rhabdomes 

 of the ordinary teloclades. It has been stated above that on the surface indi- 

 cations of hirsute spicules, broken off short, are observed. It is possible that 

 these slender spicule-fragments are parts of those hj^jothetical hirsute spicules. 

 If hirsute spicules are really present they are probably teloclades with long 

 and slender cladomes. Since, however, I have not been able to find a single 

 cladome of such a spicule in all the numerous preparations made, I doubt 

 whether the sponge possesses any teloclades of this kind and rather incline to 

 the view that these fragments are parts of foreign spicules. 



The microscleres are larger oxyasters passing into smaller oxysphaerasters, 

 small strongylosphaerasters, and sterrasters. The strongylosphaerasters are very 

 numerous on the outer surface and also occur in small numbers in the walls of 

 the cortical and choanosomal canals. The oxyasters (and oxysphaerasters) 

 are quite numerous in the choanosome, chiefly in the canal-walls; a few 

 also occur in the cortex. The sterrasters are densely packed in the sterraster- 

 armour and are, in radial sections, frequently also found scattered in the choano- 

 some. Whether they naturally occur there, or whether those observed there 

 have been brought to this position in cutting, I cannot say. Among the ordi- 

 nary sterrasters a few sterrasters with abnormal (hypertrophic) rays, sterroids 

 occur. These appear to be irregularly scattered among the others. 



The isoadine rhabds (Plate 1, figs. 22, 23; Plate 4, figs. 8, 9, 11, 12, 16, 24) 

 are usually straight or slightly curved and spindle shaped, tapering towards 

 the ends; more rarely shortened and cylindrical. The ends of the spindle- 

 shaped ones are usually blunt, less frequently sharp pointed; the ends of the 

 cylindrical ones simply rounded off or slightly thickened. All these forms, the 

 spindle-shaped, sharp-pointed, and blunt amphioxes, and the shortened cylindri- 

 cal, simple amphistrongyles, and terminally thickened amphityles are connected 

 by transitional forms to such an extent that it appears advisable to describe 

 them together. These spicules attain a maximum length of 2.8 mm. Within 

 this limit their longitudinal dimension is variable, particularly so in the specimen 

 of var. carolae from Charlotte Sound, where a good many of the isoactine rhabds 

 are very much reduced in length. In this specimen the isoactine rhabds are 

 0.37 2.5 mm. , usually 1 .6 2.2 mm. long and 40-65 n thick. As will be seen from 

 the measurements tabulated below the thickness is not in proportion to the 

 length, the shortest spicules being nearly, if not quite, as thick as the longest. 



