130 I. IJIMA : HEXACTINELLIDA. I. 



interspace between and around tliera except in places along the 

 convex sides of any two adjoining chambers, where, as it not 

 infrequently happens, the outer surfaces of these may lie to a 

 greater or less extent in direct contact with each other. The 

 point in question may best be studied on either real or optical 

 sections across the chambers. The said interspace, whieli scarcely 

 needs be pointed out as a part of the incarrent lacunar system 

 of the external trabecular layer, is broadest where three or some- 

 times more chambers at a time give boundary to it (PL V, tig. 43). 

 The cornel's of such a sjiace extend into the exceedingly narrow 

 cleft between every two chamber?, which cleft, as above men- 

 tioned, may at times be obliterated by the coming in contact of 

 the opposite surfaces. Even in the latter case, I think, the contact 

 surfaces, so far as the parts of the wall concerned show the 

 characteristic reticular structure, are not in actual fusion, though 

 sometimes a fusion may occur at points where the chambers come 

 in contact at the rim (marginal membrane), in which part the 

 chamber-wall is, as will soon be seen, structurally the same as 

 the trabecular. 



The intercameral space above referred to is traversed by fine 

 branching trabeculce (fig. 43, fr.), which extend between chamber- 

 walls or connect these with the supporting spicules and serve to 

 keep the chambers expanded and in position. Close to the 

 chamber- ritn the trabeculse are replaced by an exceedingly thin, 

 fairly continuous membrane — the comiecting membrane or mem- 

 brana reuniens of F. E. Schulze (PI. IV, fig. 22, cm.) — which 

 thus spans the interspace left between the circular apopyles and 

 joins these together. The connecting membrane may be said to 

 shut off the intercameral incurrent space from' the excurrent 

 lacunaj of the internal trabecular layer. Occasionally there exist 



