OF THE SPONGIADiB. 49 



rosette. Fig. 297, Plate XVIII, represents a rosette- 

 shaped group containing about the usual number of 

 spicula. 



Besides the rosette-shaped groups in Hymeniacidon lingua, 

 there are a considerable number of these spicula dispersed 

 over the surfaces of the membranes, but the attachment of 

 these spicula is more frequently at the middle of the shaft 

 than at the smaller end of the spiculum, their normal point 

 of attachment. In the single and separate mode of dis- 

 position they are performing the office of equi-anchorate 

 spicula, and the mode of their attachment is varied accord- 

 ingly ; but under these conditions they are rarely ever so 

 fully developed, nor do they attain the same size as those 

 which form the radiating groups. Notwithstanding the 

 numerous groups and dispersed spicula of the inequi- 

 anchorate form, this sponge is also abundantly furnished 

 with bihamate spicula of various forms, but they are never 

 congregated hke the anchorate ones. 



The same radiating mode of arrangement occurs in 

 a parasitical Australian sponge from . Preemantle, but the 

 form of the terminations of the spicula is very different 

 from those of Hymeniacidon lingua. The distal termination of 

 each of the inequi-anchorate spicula is shortened in length, 

 but expanded laterally to a considerable extent, and 

 its terminal edge is furnished with three thin pointed 

 teeth. The distal end has two small expanded and raised 

 wings, projected in the direction of the inner cm-ve of the 

 spiculum, and so disposed as to cause it to resemble very 

 closely an engineer's spanner for bringing up to their 

 bearings projecting square-headed screws. Thus, although 

 the forms of the termination of the two varieties of spicula 

 vary to a considerable extent, the principles of their 

 structure and purposes are in perfect unison. Pig. 135, 

 Plate VI, represents a single spiculum highly magnified 

 to display the peculiarity of their structure. 



As may be imagined, from their office and situation in a 

 thin stratum of a gelatinoid sarcode, they are at all times 

 small, and in many cases so minute as to require a micro- 

 scopic power of at least 600 linear to render their structure 



