90 



Soraotiraos I found sco])iil(;8 with more than four, with six or even eight branches. 

 Tlu) branches of tliose HC()})ulea are, like tlioso of tlio tetrabrachial ones above 

 (ioscribod, dither of nearly uniform thickness througliout, or basally very thin 

 and tcrniinaily thickenetl in a chib-shaped manner; and they are also, like those, 

 d(nisely covered with spines, small at the base, larger towards the distal end and 

 directed backwards (pi. XV, f. 3). It is to be remarked that transitional scopule- 

 forms, connecting the extreme ones here described in different ways, are pretty 

 frcHjuent. Scopules with fewer than four branches are exceedingly rare. 



Uncinates, varying in length and thickness, are met with close to the dicty- 

 onal hoiu^coinb pret(-y fre(iuently. They ;iro arranged vertically to the surface 

 and usually penetrate the whole thickness of the body-wall. The outer half 

 of the spicule, nearer the dermal surface, is always thicker than the inner 

 liiiir, nearer the gastral surface, which latter is quite gradually attenuated to 

 a pointed end. Tlie spines which are generally very oblique, closely " anliegend " 

 to the shaft, point inwards. If the uncinates protruded beyond the surface they 

 could be compared to harpoons. 



Of the spicules irregularly scattered throughout the parenchyme I will first 

 describe those simple hexactines which measure 100-150/* in diameter and play 

 so important a part in the growth of the dictyonal network. In all the places 

 where this skeleton-net is growing in thickness or otherwise, great numbers of 

 such hexactinos are met with. Their centres are usually considerably thickened 

 and their rays are stout, straight, gradually attenuated, bluntly pointed and 

 irregularly covered with more or less numerous, small tubercles or vertical spines. 

 B(>sides these robust hexactines, most of which are to be considered as building 

 fnaterial for the dictyonal not, other, similar oxyhexactines are met with here and 

 tliere(pl. XV, f. l.'i). Theraysof these are longer and more slender and covered 

 111(1 II' or less densely with vertical, straight or outwardly curved spines. These 

 iiexactines are at first loosely scattered through the parenchyme, later they are ap- 

 parently also incorporated by the dictyoTial net, at least I have seen, in the mas- 

 sive, basal part of the skeleton, such slender, spiued hexactino slirmly attached to it. 



Another kind of parenchymalia are the hexasters- Some of them are very 

 peculiar, one axis being differentiated and usually elongated. These spicules are 

 ti(>ver connected with the dictyonal network. Curiously enough the distribution 

 of these hexasters is so irregular that one finds great numbers of them in some 

 places, whilst in others they are very scarce. They are scattered irregularly 

 tlu'ough tlu> chamber-layer, the subgastral and subdermal regions, but always 

 absent in the gastral and dermal membranes. 



The most frequent hexastor-form is that described by Wyville Thomson in 

 18()8 in his essay *" On the Vitreous Sponges " and there illustrated on p. 123. 

 It has a considerably elongated main axis. W. Thomson says of it : " One set 



• Aim. Mug. Nat. llidt. ser. 4, Vol. 1, p. lU. 



