iQii.] N. Annandai.e : Sponges and Vermctid Molluscs. 53 



one cladus is less well developed than the other three. They 

 can always be distinguished, however, from true phyllotriaenes 

 b}^ the fact that in the former this cladus, although apparentlj^ 

 vertical in direction and often only a little proliferated at the 

 tip, has never the regular pointed form of the vertical spike 

 of the phyllotriaene. I have no doubt that these ''inter- 

 mediate " desmas are simply 3'oung spicules that have not yet 

 become firmlj- united with their fellows. The true phyllotriaenes 

 are entireh* confined to the external surface, on which they form 

 a single reticulate layer, the branches of different spicules over- 

 lapping at the extremities and the spike pointing verticallj^ down- 

 wards. The subdermal channels lie directly under this layer and 

 the spikes project into their lumen. I have not found any of the 

 discoidal forms figured by Carter, op. cit., pi. xviii.figs. 2a, 2b; 

 but spicules agreeing with his figs. 2c and 2d are abundant in my 

 preparations (pi. ix, figs. 4 — 7). Possibly the discoidal spicules 

 are onl}- found in 3'oung or stunted sponges. The most note- 

 worth}^ difference between Carter's specimens and those of this 

 variet}^ is the apparent absence in the former of the slender 

 rhabdi that form a conspicuous feature in the latter. These 

 rhabdi vary greatly in length but are always very slender, the 

 longest measuring eg mm. X o'ooy mm. They are not inflated 

 in the middle or at the ends and their tips are bluntly pointed, 

 but one end is often stouter than the other. The longer specimens 

 are sinuous and hair-like. These rhabdi lie in a more or less vertical 

 position in the interstices of the skeleton, and more especially in 

 the walls of the efferent canals, sometimes forming loose strands in 

 the latter situation. They do not project on the surface of the 

 sponge. The microscleres (pi. ix, fig. 15) are somewhat irregular 

 amphiasters with a circle of spines round the shaft ; but the 

 spines are often asymmetrical in form and disposition. The spicules 

 measure about O'oi mm. in length and 0*006 mm. in greatest 

 l)readth. The amphiasters are found chiefly in the dermal layer, 

 in which they are densely scattered. 



Habitat. — Off the coast of the Ganjam and Vizagapatam dis- 

 tricts of the Madras Presidency in 15 to 30 fathoms : associated 

 with Spiroglyphus cnmniingi (Morch). 



Var. siliqunriac , nov. 

 (Plate ix, figs. 16 — 18.) 



Sponge denser and more massive than that of var. spiroglyphi, 

 of a dull grey colour diversified with large spots which are of a 

 deep purple colour and have irregular but well-defined outlines ; 

 oscula larger, main efferent canals wider and more vertical. 



Spicules and Skeleton. — The desmas, phyllotriaenes and amphi- 

 asters agree almost exactly with those of the var. spiroglyphi , ex- 

 cept that the desmas are a little stouter ; but the rhabdi are very 

 few or altogether absent. The sponge has the habit of collect- 

 ing spicules from other sponges that grow in its vicinity, and I 



