354 Mr. E. A. Minchin on the 



thereby their equiangular character. The triradiate systems 

 may become bilateral in form by alterations in the length of 

 a ray relatively to the other two, but never, except abnormally, 

 by changes in the angles at which they meet, at least in the 

 case of the principal spicules of which the general skeleton is 

 composed, and which are found in all parts of the sponge. 

 In other words, the triradiate systems may become " ungleich- 

 strahlig," but never " ungleichwinkelig." When triradiate 

 systems vv ith paired angles occur, sagittal triradiates as Hackel 

 has well termed them, they form a special dermal or other 

 layer restricted to some region of the sponge colony. Mon- 

 axcns may be present in addition to tri- and quadriradiates. 

 All the species of this group are very contractile and have a 

 well-developed contractile flat epithelium distinct from the 

 skeletogenous layer. In the forms without quadriradiate 

 spicules the contraction may reach a pitch at which the tubes 

 become solid. In those with quadriradiates the projecting 

 gastral rays exercise a hindrance to the contraction of the 

 tubes in proportion to the length of the rays, and contraction 

 is not possible to the same extent as in the " Ascetta " forms, 

 but it is still very marked. In the histology Vi^e may notice, 

 as a striking and very constant character, that in the collar- 

 cell the nucleus is situated at the base of the cell. In the 

 embryology the larva is a 2^<^'^^nchymeUa * — that is to say, 

 a ciliated blastula in which cells migrate inwards to form an 

 inner mass or future dermal layer, which is completely covered 

 by the ciliated or gastral layer. Finally, we may note that 

 the first spicules to appear after fixation are triradiate systems, 

 so that the young sponge passes through what may be called 

 a " Protascetta " stage +. 



In the hotryoides group, on the other hand, the principal 

 triradiate systems of the skeleton are sagittal, having become 

 bilaterally symmetrical through the basal rays meeting in 

 such a way as to form an unpaired and two paired angles. So 

 far as 1 know, equiangular triradiate systems never occur 

 among the spicules of this group. The sponges of this genus 

 are in contrast with those of the preceding through their 

 non-contractility, due apparently to the fact that the cells of 



* Using this term in a descriptive sense only, without wishing to imply 

 any homologies of the component layers. 



t I have found the parenchymella larva and the Protascetta stage in 

 coriacea, hlanca, cerehrum, reticulum, and contorta. In the last-named 

 Barrois has described and figured an amphiblastula larva ("Meuioires 

 sur I'embryologie de quelques £ponges de la Manche/' Ann. d. Sci. nat. 

 (6) iii. 1876, pp. 35-37, pi. xiv. tigs. 21, 22), but this statement must rest 

 on an errorof identification. Parenchymella larvae havealso been described 

 in clathrus and primordialis by Metschnikoff and Schmidt, 



