Sponge Spicules and their Names. 213 



of end-spines in the parenchymal microhexactins; or possibly, 

 again, the parenchymal trabecular tissue was prevented from 

 developing a complicated system of meshes requiring to be 

 supported by centrifugal end-spines, as in the Hexasterophora. 

 The end-spines of amphidisks appear to function as balancers 

 for maintaining the orientation of vertical props, and the office 

 of the end-spines * of hexasters is probably to support strands 

 of a complicated network of secondary formation, the develop- 

 ment of such meshes apparently being unhampered in the 

 Hexasterophora. 



The correct appreciation of the relation — in regard to 

 absence or presence and direction — of the end-spine to the 

 actine is not, as it might at first seem, a trivial matter, but 

 important from the point of view not only of morphological 

 classification of spicules, but also of phylogenetic interpre- 

 tation. 



EXPLANATION OF PLATE VIII. 



Group A. Holactine Spicules (without end-spines). 



Fig. 1. Hexactin. 



Fiq. 2. Stauractin. 



Fig. 3. Triactin. 



Fig. 4. Diactin. 



Fig. 5. Microhexactin. 



Fig. 6. Actine of same (in glycerine), X 1400. 



Several of these figures are slightly diagrammatic, the outline of the 

 axial canals heing exaggerated. 



Group B. Astral Spicules (with end-spines). 



Hexadisk. 



Amphidisk. 



One end of amphidisk in optical section, showing axial canal 

 terminating at upper end of actine and not prolonged into the 

 centripetally directed teeth or end-spines (in glycerine), 

 X 1000. 



Part of hexaster (calycocome). 



Monoxyhexaster. 



Stauraster. 



Triaster. 



Diaster. 



* In the Ann. & Mag. Nat. Hist., Dec. 1909, p. 507, I spoke of end- 

 spines acting as spandrils. This was an error, because spandrils (or 

 spandrels) are spaces of a certain kind. I should have said " supporting 

 rods." 



Ann. & Mag. N. Hist. er. 8. Vol. v. 15 



