3 i2 THE TANGANYIKA PROBLEM. 



in Fig. I — d., is very exceptional. The axial thread, in contrast to spicules of class 

 (a.), present a dilation corresponding to each swelling on the spicule. 



(7) The spicules of this class are of variable and irregular form, since the 

 individual amphioxea or amphitornota which form them may fuse at any point and 

 at any angle (Fig. 1 — b.). As a rule these compound systems are formed from 

 spicules from class (a.), though occasionally a spicule of class (/3) is found to take 

 part in their formation. 



With regard to their origin, two suppositions are possible ; first, that they are the 

 result of irregular growth, and branching of a single spicule derived entirely from a 

 single scleroblast ; secondly, that they arise by fusion of spicules primitively distinct, 

 and formed each by its own scleroblast. Fig. 1 — b. might be taken as evidence of the 

 former view, but such forms as that represented in Fig. 1 — b. render such a supposition 

 highly improbable, to say the least. The view that these spicular systems are of 

 compound origin receives strong support from the way in which their axial threads 

 cross one another instead of branching. If these irregularities arose as outgrowths 

 from one spicule formed in one mother cell, it might well be expected that their 

 axial threads should be also formed as outgrowths from that of the main spicule ; 

 but this is certainly not the case in many spicules of our Spongilla, as can be seen 

 from the figures. In another sponge, which is probably a monaxonid of the family 

 Axinellidce, viz., Tricentrium muricatutn (Pallas 1756), Ehlers, 1870 {=Plectronella 

 papulosa, Sollas, 1879), there are branched spicules in which the axial threads are 

 continuous throughout, a fact which may indicate that the spicules themselves owe 

 their form to branching. It seems clear, therefore, that the irregular spicules of 

 Spongilla moorei have, in many cases, been produced by fusion. Judgment must 

 be suspended for the present with regard to those systems in which no discontinuity 

 can be detected in the axial threads of the component spicule rays ; such spicules 

 may be simply branched. The question cannot be decided until the actual origin of 

 the spicules has been studied ; and the same may be said for Tricentrium. Since 

 now it has been shown that the triradiates and quadradiates of the Ascons are formed 

 by fusion, there is no inherent improbability in a similar process occurring in other 

 cases. 



Spicules of a similar character to the compound systems here described have been 

 figured by many authors in various Spongillidae {Spongilla aspinosa, Potts) Lubomir- 

 skia intermedia, Dybowski). All these authors regard them as abnormalities, but 

 in moorei they are so frequent that they must be considered as a normal feature of 

 the species. It is possible that in other Spongillidre these systems have not received 

 the attention they deserve. 



In addition to the spicules described above there are small masses of silica in 

 Spongilla moorei, comparable with those found in Spongilla aspinosa (Fig. 1 — e.). 



(B) The Arrangement of the Spicules to form Fibres, etc. — The spicules 

 which form the polyspiculous fibres belong mainly to the first and third classes 

 above described. Spicules of the first class form the greater part of the fibres, while 

 others lie about in the sponge tissue, presenting for the most part an iiregular 

 method of arrangement, though many such spicules are placed so as to bridge over 

 the spaces between the fibres in a perfectly definite way. Spicules of the second 

 class, which are far less numerous than those of the first, seldom participate in the 

 formation of the fibres, but, as a rule, lie scattered irregularly between the fibres. 



