55 



in closer contact, or possess bigger nuclei. This is e. g. the case at the periphery; but it likewise 

 occurs in other places. 



5. The skeleton. 



A. Spicules. PI. VIII, fig. i — 7; PI. IX, fig. 1—9; PI. X, fig. 1 — 7; PI. XI, fig. 1— 7; 

 PI. XII, fig. I — 12; PI. XIII, fig. 1—4; PI. XIV, fig. 2. 



The complete skeleton of Spiras t 'rei 'la purpurea is composed of tylostyli (resp. styli) and 

 spinispirae 1 ). In many cases, however, the latter are very scarce ; not unfrequently they were 

 even found to be absent. It is, of course, possible that the spinispirae are in such cases in 

 reality not absent, only exceedingly scarce. This may be concluded from the facts that 1" we 

 generally examine only a little fragment of the whole sponge and 2" I often found them to be 

 present in the very same specimen in which other authors did not find them. As a proof how 

 scarce they sometimes are, may be mentioned that e. g. in one preparation of S. K. 245 I found 

 one or two spinispirae, in another of the same sj^ecimen not a single one; this fact stands by 

 no means alone. These statements teach us that, if we examine specimens like 1971, 1975, 2049, 

 B. M. 11 and B. M. 23» which in their whole habitus seem to be identical, but some of which 

 possess spinispirae, others not, we have no right to separate them specifically on account of 

 the absence or presence of spinispirae. Now this is exactly what we frequently see. Dexdy does 

 not mention spinispirae in his Suberites inconstans var. globosa and var. maeandrina ; but I 

 actually found them in the very types, though the spicules in the former are very rare indeed. 

 The Siboga specimens g8, 931, 964 a-b, g66a-b are doubtlessly identical with Dendv's above 

 named sponge; they all possess spinispirae. The Siboga specimens 2030, 2031, 2032 are three 

 fragments, very likely from the same individual ; in 2031 and 2032 I did, in 2030 I did not find 

 spinispirae. 



I. Megas cle r es. 



The megascleres are typically distinct tylostyli; a few styli, however, generally are present 

 at the same time. In some cases other sorts of megasclera are met with viz. oxea and strongyla. 

 It is doubtful whether the former belong to the sponge. (Cf. infra). 



The size and the shape of the tylostyli are exceedingly variable. If we compare the 

 slender tylostyli of 1945 (PI. VIII, fig. 3 v-i) with the stout ones of 1418 (PI. XII, fig. 10a-;) 

 the difference appears to be so great that it might be taken as of specific value. The same 

 variation we find in others. Dendy's Spirastrella vagabunda B. M. 17 (PI. XI, fig. 1 a—-/j) has 

 slender tylostyli, whereas the same author's Spirastrella vagabunda B.M. 14 (PI. XI, fig. 6a— £) 

 has very stout ones. Dendy of course correctly dicl not specifically distinguish the two, for he 

 found severaJ transitions. In Spirastrella vagabunda B.M. 41 (Pb XI, fig. 5 a-Q and B. M. 24 

 (PI. XI, fig. Ja— £) both, slender and stout spicules occur; moreover they are not so stout as 

 in B.M. 14- Consequently Dexdy did not consider Spirastrella congenera Rdl. as a different 



1) The term is used in the sense proposed by me in 1902 p. 10S. 



