443 



the radially directed ray would then have i)ierced the gastral layer, 

 which no spicules ever do in the Hexactinellida; but rather the form of 

 spicule known as the stauractine, a tetractine in which the four rays 

 meet each other at right angles and lie in the same plane; in other 

 words a spicule belonging to the triaxon series, but in which the two 

 radially directed rays are lacking. Spicules of this type are of common 

 occurrence in the Hexactinellids, but are generally regarded as derived 

 by reduction of the rays of a regular hexactine. While I am by no 

 means prepared to deny that the stauractines ordinarily met with in 

 adult Hexactinellids at the present day, may in many, perhaps in all 

 cases, have originated in this way from hexactines, I am nevertheless of 

 ojiinion that there is strong evidence, apart from such speculations as I 

 have set forth above, for regarding the stauractine as an older type, phy- 

 logenetically, than the hexactine, and for deriving the hexactine primi- 

 tively from the stauractine by the addition to the latter of the radially 

 directed rays; this addition having taken place at the time when the 

 gastral layer became folded to form the flagellated chambers. 



In support of the hypothesis that the stauractine represents the 

 most archaic type of spicule in the Triaxonia, two classes of facts may 

 be adduced. 



In the first place, Ijima has succeeded in finding larvae in various 

 stages of development in the tissues of some of his specimens. In these 

 larvae spicules are formed before they are set free from the maternal 

 sponge, and the first formed spicules are stauractines inali 

 cases. In the second place, in all palaeozoic Hexactinellids of wliich 

 the spicular structure can be made out, it is found in all cases to con- 

 sist entirely or at least principally of stauractines, for which 

 reason Schrammen [8] has founded an order Stauractinophora to in- 

 clude the palaeozoic families Protospongidae, Dictyospongidae, 

 and Plecto spengi dae, all characterized by the possession of a ske- 

 leton composed of stauractines. Schrammen has, it is true withdrawn 

 his name Stauractinophora in his later publication [9], and merely groups 

 the three above-mentioned families together under the heading Incertae 

 Sedis, but in my opinion this is a retrograde step. The stauractinal 

 skeleton is especially well seen in the oldest known fossil sponge Proto- 

 spongia, from the Cambrian, in which it was described by Sollas [12". 

 The description given by Sollas also brings out another point of great 

 importance in the present connection, namely the the extreme thinness 

 of the body wall in Protos pongia, which inclined Sollas to the belief 

 that the layer of stauractines seen by him represented only the dermal 

 spicular layer of the sponge. On the hypothesis, however, that Proto- 

 spongia was in a stage of evolution in which the gastral or collar-cell 



