110 ZOOLOGY sect. 



'i, teased out in sea-.ater, and the broken fragments examined 

 under a tolerably high power of the microscope, groups of these 

 collared cells will be detected here and there, and in many places 

 the movement of the flagella will be readily observed. The 

 flagellum is flexible but with a certain degree of stiffness, 

 especially towards the base, and its movements resemble those 

 which a very supple fishing-rod is made to undergo in the act of 

 casting a long line — the movement being much swifter and 

 stronger in the one direction than in the other. The direction 

 of the stronger movement is seen, when some of the cells are 

 observed in their natural relations, to be from without inwards. 

 It is to these movements that the formation of the currents of 

 water passing along the canals is due. The collars of the cells in 

 specimens teased in this way become for the most part drawn back 

 into the protoplasm. 



The short passage or excurrent canal, which leads inwards from 

 the flagellate canal to the paragastric cavity, differs from the 

 former in being lined by flattened cells similar to those of the 

 paragastric cavity ; it is partly separated from the flagellate canal 

 by a thin diaphragm (Fig. 81, di, and Fig. 83), perforated by a 

 large circular central aperture — the apopyle (ap) — which is capable 

 of being contracted or dilated : its opposite aperture of com- 

 munication with the paragastric cavity, which is very wide, is 

 termed the gastric ostium of the excurrent canal. 



The effect of the movement of the flagella of the cells in the 

 flagellate canals is to produce currents of water running from 

 without inwards along the canals to the paragastric cavity. This 

 causes water to be drawn inwards through the prosopyles from 

 the incurrent canals, and, indirectly, from the exterior through the 

 perforated membranes at the outer ends of the latter. 



Between the ectoderm of the outer surface and of the incurrent 

 canals, and the endoderm of the inner surface and of the flagellate 

 canals, are a number of spaces filled by an intermediate layer — 

 the mcsoglcea — in which the spicules of the skeleton are 

 embedded. Each spicule is developed from cells termed sclero- 

 blasts, which migrate inwards from the ectoderm. Each ray is 

 formed by the agency of a separate scleroblast, so that there are 

 three at least of the latter for each triradiate, and four for each 

 tetraradiate spicule. The spicules (Figs. 80 and 81, sp) are 

 regularly arranged, and connected together in such a way as to 

 protect and support the soft parts of the sponge. Most are, as 

 already noticed, of triradiate form. Large numbers, however, are 

 of simple spear-like or club-like shape (sp) ; these, which are 

 termed the oxeote spicules, project on the outer surface beyond the 

 ectoderm, and are arranged in dense masses, one opposite the 

 outer end of each of the ciliated canals, this arrangement pro- 

 ducing the pattern already referred to as distinguishable on the 



J 



