190 MARINE BIOLOGY OF THE SUDANESE RED SEA. 
the Staurorrhaphidz, form an extremely compact and definite group, exactly 
comparable to that formed by the two families Sycettidee and Grantide. The 
Staurorrhaphidee are immediately derivable from the Chiphoride by the 
development of a dermal cortex. 
These two families would thus form a branch diverging from the main 
Grantillid stem almost immediately after the appearance of the primitive 
prochiact, the point of divergence being marked by the development of the 
apical ray, transforming the prochiact into the chiactine. 
It is possible to derive the genus Grantilla in two entirely different ways. 
According to one view there is no connection whatever between the 
Chiphoridze and Staurorrhaphide: on the one hand and this genus on the 
other. Grantilla would then be a direct development from the typical 
Heteropid form, by the bending of the paired rays of some of the sagittal 
triradiates. Against this view there are however several, to my mind, very 
serious objections. It implies that two spicules of peculiar and extremely 
similar shape and occupying similar positions in the sponge-wall, namely, the 
subgastral prochiact and the chiactine, should have been developed inde- 
pendently at two different times in the phyletic history of the Heterocela. It 
necessitates, further, our belief that these two spicules should have developed 
by entirely different methods from spicules of entirely different type; for 
Jenkin (9) is of the opinion that the chiactines are developed from the gastral 
quadriradiates. Another objection to this view is the presence of subdermal 
quadriradiates in Grantilla quadriradiata; so that if the Grantillide are to be 
derived from the Heteropids, it must be assumed that the spiculation of all 
these sponges is extremely liable to variation. This interpretation of these 
sponges would be in direct opposition to the principles of classification 
suggested by Dendy (2), who lays very great stress upon the stability of the 
composition and arrangement of the skeletal elements in the different families 
of the Heteroceela. It is, however, possible, as Professor Dendy has suggested 
to me, to give to the Grantillidee an interpretation in which none of these 
difficulties appear, namely, that they are a primitive group, derived from the 
same ancestors as the Chiphoride and the Staurorrhaphide, and representatives 
of the original line of descent from which the Heteropidee and the Ampho- 
riscide have been derived by specialization of their respective characters. 
The Grantillidee would thus be immediately derived from the primitive 
prochiact-bearing stock by the appearance of a dermal cortex. That a dermal 
cortex has originated more than once in the history of the Heteroccela is 
made evident by a consideration of the two families Grantide and Stauror- 
rhaphide. To derive the Staurorrhaphide from the Grantidee would neces- 
sitate the diphyletic origin of chiactines ; to derive them from the Chiphoride 
necessitates a similar diphyletic origin for the dermal cortex. Of these two 
alternatives the latter is by far the most simple; the direct continuity of the 
line Chiphoridee-Staurorrhaphide, the peculiarly restricted distribution of 
