OF THE SPONGIAD^E. 105 



surface, as the fragments of the dermal membrane which 

 remain all seem to cover that side of the fibres. Generally 

 speaking, there is some difficulty in detecting the double 

 series of spicular organs at the angles of the network, but a 

 reversal of the object beneath the microscope immediately 

 removes all doubt on that subject. 



In Grantia coinpressa and ciliata the intermarginal cavities 

 appear to attain their highest degree of development, and 

 are multiplied and expanded to such a degree as to almost 

 supersede every other organ. The whole sponge in these 

 species is formed of a great accumulation of elongated cells 

 or cavities, closely adjoining each other and angular by com- 

 pression. Their conical distal terminations, abounding in 

 pores, represent the external surface of the sponge, while 

 their valvular proximal ends form the inner surface, in con- 

 junction with the shallow cavities, into the distal ends of 

 which each cell discharges its contents. These shallow 

 depressions, intervening between the intermarginal cavities 

 and the cloaca, are all that remains to represent the incur- 

 rent portion of the interstitial systems so largely developed 

 in the Halichondroid sponges, the great cloacal cavity 

 entirely superseding the excurrent spaces and canals (Figs. 

 312 and 313, Plate XXI). 



In these species of Grantia there is no doubt regarding 

 the existence of cilia, the whole of these great cavities being 

 completely lined with them. 



It is a question whether the intermarginal cavities share, 

 in common with the interstitial canals, in the function of 

 the assimilation of nutriment, or whether they are devoted 

 solely to the aeration of the fluids of the animal ; and this, 

 if we consider the structure and extent of the interstitial 

 canals in the Halichondroid sponges, is probably the case. 

 In Granlia the abundant provision of cilia in those cavities 

 at once stamp them as breathing organs ; and although 

 cilia have never yet been satisfactorily proved to exist in 

 the intermarginal cavities of the Halichondroid sponges, 

 there can be no reasonable doubt of their being the hoino- 

 logues of the large ciliated cavities in Grantia cowpresm 

 and other similarly constructed sponges. Now, in these 



