445 



in this group, and appears to have been seen only byHaeckel. A 

 skeleton made up entirely of stauractines might, however, be expected 

 to occur in the most ancient members of the group Hexactinellida, such 

 asProtospongia, and this expectation appears to be perfectly reahzed. 



While therefore neither the embryological nor the palaeontological 

 facts, if considered separately, furnish a complete proof of my theory, 

 yet if taken both together the accumulated weight of evidence becomes 

 much greater, especially when it is seen that on theoretical grounds the 

 facts are just what might have been expected and foretold. The palaeo- 

 zoic group of Stauractinophora is, in fact, one which it would be neces- 

 sary to invent, if it did not exist, on the theory of Hexactinellid phylogeny 

 here put forward. In theories of spicule-origin in sponges, there is 

 generally a tendency to assume that the forms with many rays are more 

 primitive than those with fewer. This assumption works very well for 

 Demospongiae, though even in this group Maas [4, 5j has described the 

 asters of Tetiiya as arising from fusion of tetraxon spicules. On the 

 other hand the reduction-theory, as it may be called, certainly does not 

 apply in the case of the Calcarea, where it is above dispute that the 

 triradiate spicules are more primitive than the quadriradiates, as was 

 recognized on general grounds by Haeckel and Schulze, and as I 

 may claim to have proved definitely by my observations upon the mode 

 of formation of these spicules. There is, therefore, no a priori reason 

 against the hexactine being derived from the stauractine by the addition 

 of rays to it, but, on the contrary, from the known facts in other sponges, 

 there is if anything rather a presumption in favour of such an origin. 



There is, however, one point in the evolution of the spicules which 

 requires further discussion, and which is, indeed, as great a difficulty on 

 Schulze's theory of the origin of the hexactines, as on my view that 

 the stauractines were the more primitive form. On either hypothesis 

 the question at once arises, why should the four rays that lie tangentially 

 in the body-wall between the chambers meet so definitely at right 

 angles? If they first appeared as stauractines during my stage Ä^ there 

 does not seem at first sight any reason why the rays should meet sym- 

 metrically at all. There would be nothing to prevent the rays enclosing 

 any sort of angle at their junction. Schulze's theory ascribes the 

 rectangular junctions of the rays to the arrangement of the chambers, 

 which would mean that the spicules first acquired their regular form in 

 stages B or C, and would also necessarily imply that the diverticula 

 of stage jB, or the chambers of stage C, had a regular arrangement in 

 rows crossing each other in a rectangular pattern, as shown in a, fig. 2. 

 But if we enquire further why the chambers should have been arranged 

 at any time in this manner, it seems impossible to suggest any reason 



