Sponge-fauna of Norway. 399 



so as to become three or four cells deep ; but usually it con- 

 sists of a single layer of cells only. Just below the epidermis 

 sphsero-stellates occur between the dermal cells, their rays pro- 

 jecting against the epidermal membrane (PL XVII. fig. 11, d). 



No vesicular nor gelatinous connective tissue is observable 

 in the dermis. 



The Globate Layer. — The structure of this does not differ 

 from that described as existing in Geodia Barretti, except by 

 the absence of vesicular connective-tissue cells from the trian- 

 gular spaces left between the fibrillar ligaments ; these cells are 

 replaced here, as elsewhere in Isops Pklegrmi, by gelatinous 

 connective tissue. The most exterior, and therefore oldest, of 

 the globates of the rind are very often hollow within, the 

 small central cavity which exists in the ordinary adult globate 

 having become enlarged to a great but variable extent. This 

 occurs as the result of an absorption which begins at the inner 

 ends of the trichites, and, extending radiately outwards, repro- 

 duces the early form of the young globate as a hollow cast 

 within the old one; the same result is brought about by exposing 

 the solid mature globates to the action of boiling caustic 

 potash, as described by me in a previous communication to 

 this Magazine (Ann. & Mag. Nat. Hist. ser. 4, vol. xx. 

 p. 292). 



Subcortical Layer (PI. XVII. fig. 4). — The purely fibrous 

 part of this (fig. 4,/), which lies immediately beneath the 

 globate layer, is very thin, and passes below into gelatinous 

 connective tissue (fig. 4, c), in which fibres like those of the 

 fibrous layer lie loose and more or less apart from each 

 other, and being consequently well defined are easily studied 

 in situ. They are hyaline and fusiform, with attenuated ends, 

 sometimes greatly prolonged ; a central axis is rarely visible ; 

 more usually the interior is occupied by an axial cavity ; gene- 

 rally it would appear empty, but sometimes contains a small 

 refringent spherule, which I take to be a nucleolus, and is some- 

 times filled with colourless granular material. The axial cavity 

 may be relatively very small, a mere slit in the centre of the 

 fibre ; or it may be large, perforating the whole length of the 

 fibre, and converting it into a genuine tube. The tube so 

 formed is liable to split open at one end ; when this happens 

 the slit wall uncurves and spreads out into a thin lamina. 

 The hyaline wall of the fibre frequently also becomes fibril- 

 lated and sometimes apparently laminated ; it then becomes 

 liable to exfoliation or defibrillation, as the case may be. 



The gelatinous connective tissue consists of a colourless, 

 structureless, soft matrix, containing numerous dispersed oval 

 nuclei surrounded by a small quantity of granular colourless 



27* 



