AFFINITIES OF THE GENUS SIPHONIA. 807 



Prom these specimens, which I believe to be chiefly forms of >Si- 

 phonia Webster i, Sow., slices were taken in the following direc- 

 tions: — 1. Along the longitudinal axis of the sponge, longitudinal 

 sections (PL XXY. fig. 7 a) ; 2. Transversely through its centre, trans- 

 verse sections (PL XXY. fig. 7 b); 3. Parallel to the vertical axis, but 

 at some distance from it nearer the exterior, tangential sections. In 

 examining these sections we find the internal canals, owing to their 

 having been partially filled in with chalk marl and other earthy 

 material, are distinguished from the other parts by being more 

 opaque and lighter in colour, so that with reflected light they are 

 dull grey, and by transmitted light almost black in appearance ; 

 they are not, however, limited by definite walls, but shade into the 

 surrounding substance, i. e. the phosphatic material which now 

 occupies the place of the original sarcode of the. sponge. This is 

 more or less transparent, dark brown by reflected, and light amber 

 or yellowish brown by transmitted light ; it is everywhere traversed 

 by the skeletal network, which is transparent and colourless, and 

 always sharply defined from its matrix. 



Skeletal Network. — The structure of this may be best examined 

 with powers of from 60 to 140 diameters and by transmitted 

 light. 



Selecting a favourable portion (PL XXVI. fig. 1) of any one of the 

 sections, we observe, well marked off from the surrounding brownish 

 phosphate, a small circular colourless area (0*002 of an inch in dia- 

 meter) (fig. 1 a), which is the cut end of a smooth cylindrical rod that 

 has been traversed at right angles by the plane of the section ; from 

 this circle, then, i. e. from the end of the cylindrical rod, radiate 

 three smooth cylindrical arms, 0*005 to 0*016 of an inch long and 

 0*002 of an inch broad, which terminate, either without or with 

 previous bifurcation, by dividing into a number of rounded or hemi- 

 spherical apophyses, between w^hich are left corresponding rounded 

 concavities, the w T hole termination roughly resembling a small bunch 

 of grapes. The arms make various angles with each other, main-. 

 taining no constanc} 7- in this respect. 



The structure thus described is that of a Lithistid spicule(PL XXYI. 

 fig. 2), which, as it exists in recent sponges, likewise presents us 

 with a principal cylindrical shaft, also dividing into three chief ra- 

 diating arms, which terminate after one or, it may be, two bifurca- 

 tions, by breaking up iuto a cluster of botryoidal processes. 



The bunch of irregular knobs and sockets which ends one spicu- 

 lar ray fits into and interlocks with the similar cluster at the end of 

 the ray of au adjoining spicule, the knobs of the one fitting into the 

 sockets left between the knobs of the other. 



This articulation affects all the spicules alike ; the ends of none of 

 them are free : and thus a skeletal network results which is as re- 

 sistant and rigid as that of the Yitreohexactinellids, though it is 

 arrived at in a totally different way ; for in the latter sexradiate 

 spicules are cemented into a glassy fibro-reticulate structure by a 

 coating of silica which completely envelops them, while in the Li- 

 thistids no such cement is present, the \ unction of their trifid spi- 



3g2 



