444 Prof. W. J. Sollas on the 



It most commonly consists of a very short straight shaft with 

 two long spines radiating from each end, the plane containing 

 the spines at one end being frequently turned at right angles 

 to that containing those at the other, a disposition wliicli 

 suggests a spiral arrangement, not otherwise discoverable in 

 the shaft. Almost as frequently, however, all four spines lie 

 in one and the same plane {v. Kent, vii. fig. 18). The num- 

 ber of rays varies greatly: sometimes only two appear, 

 giving us a bent acerate form ; often only three, the trirudiate 

 so formed closely resembling the characteristic spicule of the 

 Calcispongiffi ; four is the commonest number ; but additional 

 rays are not uncommonly present, up to and possibly ex- 

 ceeding eight ; in instances where the number of rays exceeds 

 live the spiral tendency is more markedly displayed. The shaft 

 sometimes shortens and disappears ; and then the four-rayed 

 form resembles the quadriradiate of Dercitas {v. Kent, vii. 

 fig. 16). The fourth ray often appears as a sprout from one of 

 the rays of a triradiate. The sjnnes are usually sharply pointed, 

 but often become rounded at the ends (fig. 13) ; they are 

 smooth and only very rarely roughened. In size these 

 spicules vary enormously : the rays of the larger forms are 

 frequently 0-00o4 inch long, but they may reach 0"005 or 

 more ; in the smaller forms they are often no more than 

 0*00091 inch long. By multiplying the length of the rays 

 by 2 we get a close approximation to the length of the whole 

 spicule. 



The minute spicules are scattered without apparent arrange- 

 ment through the sponge ; the large spicules, on the contrary, 

 lie in fascicles or short fibres, which radiate from the centre 

 to the surface, the rays of the forks spreading out beneath the 

 skin, and the heads of the grapnels lying close beneath them, 

 in the angle between the rays and their shafts. The forks 

 appear never to extend outside the surface of the sponge ; but 

 the acerates and the grapnels project a considerable distance 

 beyond it. The proximal ends of these projecting spicules 

 appear about each fibre a little below the skin (PI. XVII. 

 fig. 6); and the spicules, diverging from each other, pass out 

 in a conical pencil, having its apex pointing inwards. To- 

 wards the base one finds in addition numerous spicules con- 

 vo'ging from the middle of the sponge towards conical 

 papillaj, from which they emerge as single fibres ; here the 

 base of the cone is inwards ; the fibres afterwards open out to 

 form the terminal tufts of diverging spicules, the greater part 

 of which appear to be grapnels. 



The Ectoderm.— niQ epidermis (PL XVII. fig. 32) is a 

 thin membrane everywhere investing the s^jonge, and bearing 



