CALCAREA. 45 



in the middle of a solid jelly. There is no epithelial Hack, l>ut the epithelium 

 appears as the lining of the iiicurrent canals, which are more or less circular pipes 

 through the jelly. The pore cells have not been made out with certainty in this 

 species, hut the cells lettered p.c. in Figs. 51 and 53 arc prol>ably the pore cells, 

 which of course are closed. 



Apopyles. The apopylc usually has the appearance of an oval opening in a 

 thin elastic membrane stretching across the end of the flagellated chamber. It is 

 probable that the apopylcs can be closed in all species. 



Excurrent Canals. In Leucandra <jeltitint>m the excurrent canals close by the 

 contraction of the epithelial lining. When the canal contracts, the spiculcs arc left 

 in their normal position, and the lining cells stretch the gelatinous mesoderm behind 

 them, so that it is drawn into the canal. The apical rays do not appear to interfere 

 with this action. In cross-section the contracted canal, therefore, appears as an 

 irregular ring of spicules filled up with jelly, in the centre of which lies the contracted 

 group of lining cells (see Figs. 45 and 46). It is probable that similar contractility 

 may be possessed by other species which have a strongly developed mesodcrm ; 

 as a rule, however, the excurrent canals do not close. 



Gastral Cavity. In some Homocwla the gastral cavity <n lie closed solid.* 

 The author does not know of any instance in which this occurs among the Ifi'trrocorla. 



Osculum. The osculum may be closed in a variety of ways : 



(1) By simple contraction of the surrounding mass (v/., Lnu'andra primiijcnm). 



(2) By the folding inwards of the oscular collar (>'.<)., h-iuxindra f/flatiiwsa), etc. 



(3) By means of a diaphragm across the mouth (>:</., Megapogon raripilm). 



Diaphragms have been found in several of the new species, but not in good 

 enough preservation to enable their structure or method of closing to be investigated. 



SPICULES. 



The dimensions given for the spicules in the detailed descriptions of the several 

 species in Part II. have generally l>een taken from the drawings, which were traced by 

 camera lucida accurately to scale, and show as far as possible a representative selection ; 

 they must not !> understood to !> cither limiting (maximum and minimum) dimen- 

 sions or accurate average dimensions. To ascertain the true limiting and average 

 dimensions would require a detailed examination of a large numl>er of specimens and 

 great care in the preparation of the spiculc slides, in order to insure that all the 

 spicules from each specimen were mounted. There appeared to l>e no advantage to 

 be gained in the present instance by attempting such accuracy, which, indeed, would 

 only have been possible in the few cases where a sufficient number of specimens 

 existed. 



When the facial rays of a triradiatc or quadriradiate spicule arc " folded," /.., do 



Minchin (1). 



