The structure of the body-wall in Hexactinellida is so constant as to make it possible to give a general description applicable to all members of the group. It is of considerable thickness, but a large part is occupied by empty spaces, for the actual tissue is present in minimum quantity. In the wall the chamber-layer is suspended by trabeculae of soft tissue, between a dermal membrane on the outside and a similar gastral membrane on the inner side (Fig. 89). Thus the water entering the chambers through their numerous pores has first passed through the ostia in the dermal membrane and traversed the subdermal trabecular space; on leaving the chambers it flows through the subgastral trabecular space and the ostia in the gastral membrane, to enter the paragaster and leave the body at the osculum. The trabeculae and the dermal and gastral membranes together constitute the dermal layer. This conclusion is based on comparison with adults of the other groups, for in the absence of embryological knowledge no direct evidence is available. According to {199}the Japanese investigator, Isao Ijima,[234] the dermal and gastral membranes are but expansions of the trabeculae, and the trabeculae themselves are entirely cellular, containing none of the gelatinous basis met with in the dermal layer of all other sponges. There is no surface layer of pinacocytes, the cells forming the trabeculae being all of one type, namely, irregularly branching cells, connected with one another by their branches to form a syncytium. In the trabeculae are found scleroblasts and archaeocytes. The chambers have a characteristic shape: they are variously described as "thimble-shaped," "tubular," or "Syconate," and they open by wide mouths into the subgastral trabecular space. Their walls have been named the _membrana reticularis_ from the fact that, when preserved with only ordinary precautions, they are seen as a regular network of protoplasmic strands, with square meshes and nuclei at the nodes. This appearance recently found an explanation when Schulze, for the first time, succeeded in preserving the collared cells of Hexactinellids.[235] Schulze was then able to show that the choanocytes are not in contact with one another at their bases, where the nuclei are situated, but communicate with one another by stout protoplasmic strands. The form of the choanocyte can be seen in Fig. 91. [Illustration: FIG. 90.—Portion of the body-wall of _Walteria_ sp., showing the thimble-shaped flagellated chambers, above which is seen the dermal membrane. (After F. E. Schulze.)]