and their Relationsliljj to the Corals. 115 



both layers of the body-wall (ectoderm and entoderm) and are 

 prodneed by the mutual separation of the cells' at changeable 

 points. There is no special canal-wall. The situation and 

 numher of the cutaneous j)ores are not constant^ hut changeable^ 

 in 1 i/ nth us and the most nearly allied Calcispongite {Leuco- 

 solenia, Clistoli/nthus). New ones form themselves, whilst the 

 previously formed pores are again obliterated by the union of 

 the cells which have moved asunder. The pores behave in 

 this manner also in Leucosolenia (a stock-forming Olynthus) 

 and in Clistol>jnthus (an Olyntlius with the mouth closed up). 



In the larger and more highly developed Calcispongia3 the 

 simple and inconstant cutaneous pores gradually become con- 

 verted into })ermanent and constant canals, which acquire a 

 l)roper wall by the extension of the flagellate epithelium of the 

 stomachal cavity upon their inner surface throughout the 

 whole of the ectoderm (as in the family Sycaridaj). Among 

 these the genera Si/cum and Dunstervillia have hitherto been 

 most accurately examined ; and in these the cutaneous pores 

 have become developed into very considerable canals, which 

 are quite regularly arranged, and traverse the wall of the body 

 in a radiating direction. All previous observers, however, 

 have overlooked the fact that these radiating canals not only 

 open inwardly into the stomach and outwardly at the sur- 

 face of the body, but also all stand in direct communica- 

 tion with each other. The walls between the individual 

 closely contiguous radiating canals are, in fact, perforated in 

 all parts like a sieve, and interrupted by numerous apertures 

 of communication, or conjunctive jJores^ through Avhich each 

 canal communicates with all its neighbours. In some genera 

 the regular radiating canals ramify outwards in the same way 

 as the irregular parietal canals in the Avails of the Dyssycidas. 



The most remarkable development of the canal-system is 

 attained, however, in Cgathiscus, which is nearly allied to 

 Sycarium and Sycum, and in which the horizontal partitions 

 between the superimposed radiating canals become absorbed, 

 whilst the vertical partitions between the canals lying side hy 

 side persist. By this means is produced a system of radial 

 perigastric chambers^ which is exactly analogous to the corre- 

 sponding system of perigastric cavities radially surrounding 

 the stomach in the corals. The only distinction is, that the 

 direct communication between the stomachal cavity and the 

 chambers surrounding it takes place in the corals by the open- 

 ing of the stomach and perigastric chambers below into the 

 common basal space of the body-cavity situated beneath them, 

 in Cyafhiscus, on the contrary, by longitudinal rows of aper- 

 tures (stomachal pores) which perforate the partition between 



