Sponge-fauna of Madras. 155 



Suberifes mconstans, var. globosa. (PI. IX. figs. 1, la.) 



Sponge (PI. IX. fig. 1) massive, sessile, attached, irregularly 

 spherical, averaging about 160 millim. in diameter. Surface 

 uneven, but without digitate processes, very minutely hispid. 

 Irregularly scattered over the surface are numerous large, 

 more or less circular openings, which must be regarded as 

 the oscula ; these average in diameter about 4 millim. in one 

 specimen, while in the other they are smaller. Colour light 

 brownish orange. Texture hard and woody, incompressible ; 

 internally cavernous. 



Skeleton very irregular, composed of dense masses of loosely 

 aggregated tylostyli, without any defined fibres. At the 

 sm-face the tylostyli are mostly arranged in brushes with their 

 apices projecting outwards. 



Spicules (PI. IX. fig. la) large, stout, usually slightly curved 

 tylostyli, with well-marked, somewhat elongated heads of the 

 " enormi-spinulate " type and with gradually sharp-pointed 

 apices. Size about 0*57 by 0*022 millim. 



There are two specimens of this variety, agreeing fairly 

 closely with one another in external form. 



Suberites mconstans ^ var. mceandrina. (PI. X.) 



The single specimen (PI. X. fig. 1) consists of a great, hemi- 

 spherical, cake-like mass, attached by a very broad base to a 

 mass of calcareous debris. Average diameter about 300 millim. 

 The upper surface (PL X. fig. 1 a) is uneven, and is also fur- 

 rowed by numerous, closely-placed, very deep, meandering 

 grooves or elongated pits, each of which is about 2-3 millim. 

 broad, and has slightly prominent margins. There are also 

 usually numerous very small pits between the grooves. The 

 general surface is again very minutely hispid. Colour light 

 brownish orange. Texture hard, woody, and incompressible. 



Skeleton and spicules as before, except that the spicules 

 appear to be a trifle shorter. 



The meandering pits on the surface, from which I have 

 taken the name of this variety, are doubtless homologous with 

 the circular pits on the surface of the last variety, from which 

 we may imagine them to be derived by a process of lateral 

 elongation. In cases like the present it is of course an open 

 question as to what are to be considered the true oscula. 



Suberites mconstans^ var. digitata. (PL IX. fig. 2.) 



There are three specimens which I refer to this variety. 

 They differ considerably from one another in external appear- 



11* 



