670 Transactions of the Society. 



plate having a spherical opening in the centre of each hexagon. 

 Slack asserts he has seen real depressions and a real projecting 

 network, both composed of small spherules. Morehouse sees duplex 

 valves, the inner of which has spherical openings, the edges 

 representing the thickest parts of the valve ; the hexagonal net- 

 work of the outer valve lies in the depressions between those edges. 

 Across the mesh of the network extends a thin siliceous film with 

 most delicate anastomosing network, having its weakest point in 

 the centre. All sorts of things may be observed on this object if 

 one remains satisfied with a mere examination of the surface. The 

 nearest approach to the real state of affairs was made in 1880 by 

 Prinz (21). His figs. 7 and 8 represent tolerably well the vertical 

 chamber-walls, although by the method applied (the experiments 

 were made with thin rock sections) the real details would be difficult 

 to make out. This species viewed on the surface is to be dis- 

 tinguished by the large areola in the centre of the disk ; the section 

 shows that the entire cell has no longer the form of a coin, but, in 

 consequence of the slight curving of the disk, is like a bi-convex 

 lens. The areolse are slightly smaller than those of C. radiatus. 

 My somewhat numerous sections through one specimen are very 

 similar to those of C. radiatus, but I find nowhere a definite clue 

 for the existence of an outer opening. In fig. 29 a small portion 

 of a vertical section is given to the right, as observed with most of 

 the sections ; on the left are seen small portions in the edge of fine 

 slightly injured sections. Here I am in doubt whether the 

 T-shaped figures may not be produced by the splitting of the very 

 fine outer membranes which are in the girdle ; if everything in the 

 gum is uninjured one sees the membrane extending evenly across 

 the supports without indication of holes. It is curious that with 

 this species very often an air-bubble remains behind in the chambers 

 which hardly ever occurs with C. radiatus and Pinnularia, and 

 this might suggest that the fiuid gum does not enter through 

 openings but in the more difficult endosmotic process. It is 

 unnecessary to deal more in detail with the statements of Stephen- 

 son and Slack. The former has apparently arrived at his view 

 through examining a specimen in process of fission. 



§ 3. C. centralis Ehrenb. — This is a very large species, of 

 1/3 mm. in diameter, strongly convex, having in the centre a few 

 large areolae. I have before me a number of sections through a 

 valve which had unusually coarse markings, thus differing but 

 slightly from C. radiatus. Fig. 30 is a portion of one of the best 

 sections, and shows the chambers with the T-shaped sections of the 

 walls, suggesting arches; the surface view is like G. radiatus. 

 Here also no opening can be detected in the finest edge-portions, 

 therefore I believe I am accurate in stating that this species has 

 completely closed chambers. One must not be led away by the 



