183 



outer opening. When the sponge contracts the pores close and then 

 in sections look like amoeboid mesoderm cells. In fact the «größere, 

 körnige Zellen« described by von Lendenfeld (3) in his forms B 

 and C of this sponge (p. 212, 213 and 216) are in all probability only 

 closed pores. 



3) Mesoderm. The spicules always have on each arm one, some- 

 times two cells. Each of the cells is closely applied to the spicule 

 sheath and has a rounded, or more usually oval nucleus slightly 

 smaller than the ectoderm nuclei, with a chromatin network thicke- 

 ned in several points. The protoplasm is very clear and free from 

 granules and it is often extremely hard to see its limits. Sometimes 

 the cell contains two nuclei close together; often there are two cells 

 on one arm of a spicule. In many cases the cell has a simple fusiform 

 shape, lying along a spicule; in others it is placed so as to connect 

 the arms of two spicules ; or finally it is often connected with other 

 spicule cells by processes. The impression one gains by a careful 

 study of these cells is that the spicules lie in a continuous cell network. 

 Besides these cells the mesoderm contains potato-shaped wandering 

 cells , of a greenish yellow colour and so full of refractive granules 

 that the nucleus is only visible by careful focussing as a clear (or 

 stained) space, and ova with colourles protoplasm containing granules 

 varying in size and appearance and a distinct large spherical nucleus 

 and nucleolus. Both these kinds of cells can easily be distinguished 

 from closed or not yet open pores by the colour and granules of the 

 protoplasm and the structure of the nucleus. The stellate mesoderm 

 cells so often described, e. g. by von Lendenfeld (3) (p. 211 «zahl- 

 reiche, sternförmige Hindegewebszellen«) are exceedingly rare, if not 

 entirely absent. What have been mistaken for such are the spicule 

 cells, owing to the spicule itself being nearly always more or less 

 displaced in sections. 



4) Endoderm. The collar cells vary in shape, according as the 

 sponge is expanded or contracted. In the first case (Fig. 2) they are 

 short and broad, and have very long collars, but when the sponge con- 

 tracts they become laterally compressed and are then columnar and 

 narrow, with a very low collar. The base of each cell is rounded 

 and the cells are in close contact without any intervening substance, 

 appearing in surface views as polygonal from mutual pressure. In 

 the normal condition they are without any projections , but when ob- 

 served living they can often be seen to throw out numerous fine pro- 

 cesses, which is always, however, a sign of cessation of activity and 

 death. The figures of these cells given by von Lendenfeld (3) 

 (PI. IX, Figs. 33, 34, 35) appear to me, therefore, to represent purely 



