SPONGES J 2 1 



parenchymal strata, a point in which Hexactinellids are perhaps 

 more primitive than other sponges. 



The choanocytes, long unknown, have recently been discovered by 

 Scliulze, who describes them in Schaudinnia arctica as a uniform layer of 

 columnar epithelium, each cell bearing a collar and fiagellum. The body 

 of the cell is slightly constricted towards the middle, and expanded both 

 at its upper and lower ends. At the lower end the base of the cell forms 

 a foot-like plate, which contains the nucleus, and is in contact with the 

 similar basal plates of neighbouring cells to form a continuous protoplasmic 

 membrane, limiting the chamber towards the exterior and interrupted 

 only by the chamber pores or prosopyles. In surface view the basal 

 membrane shows a number of granular strands running from each nucleus 

 to its four neighbours, and so producing the appearance of a network or 

 lattice with ajjproximately rectangular or rhombic meshes ; this is the 

 membrana reticiduris formerly described by Schulze in the Challenger 

 material, and then but imperfectly understood. Finer strands, disposed 

 in an irregular manner, ramify in the meshes of the coarser network. 

 At their upper ends also the choanocytes are adherent to one another, 

 just below the origin of the collar, except where a prosopyle traverses the 

 chamber wall. In this way a continuous system of spaces is enclosed 

 between the narrowed middle portions of the cells. The collars are quite 

 separate from one another. The flagellum is connected with the basal 

 nucleus by an axial filament passing down through the body of the cell. 



4. Development. — Nothing is known of the embryology. Schulze 

 found only immature ova, of the usual type, in the Challenfjer material, 

 and no larvae or even segmentation stages. 



5. Classification. — The classification here adopted is that applied by 

 Schulze (1887) to recent forms, with a few subsequent additions or 

 emendations. In addition a certain number of fossil genera and families 

 have to be noticed, of which the exact position in Schulze's system is 

 not in all cases clear and cannot be determined without special in- 

 vestigation. ^ 



Sub-Class 1. Lyssacina, Z. 



The spicules of the skeleton either remain separate or are united at 

 a late period of growth in an irregular manner by siliceous masses or by 

 transverse synapticulae. 



Order 1. Hexasterophora, F.E.S. 



Hexasters always present in the parenchyma ; ciliated chambers 

 thimble-shaped, sharply separate from one another. 



Family 1. Edplectellidae, Gray. The dermal skeleton contains 

 sword-shaped oxyhexactines with long proximal ray. (a) SuB- Family 1. 



^ In his most recent work on American Hexactinellids [24] Schulze abandons the 

 subdivisions Lyssaciua and Dictyonina as a natural classification, and divides the 

 group into two orders; (1) Anii^hidiscophora, including tlie single family JFyalone- 

 matidae ; and (2) Hexasterophora, which is extended to include not only the 

 remaining families of Lyssacina, but also all the Dictyonina. 



