332 N. E. BROWN ON THE STRUCTURE OF DIATOMS. 



of 3,000 diameters. The pores are best seen when the light is 

 not very brilliant. 



Triceratium favus. The structure of this diatom, as well 

 as that of several other species, has been described and illustrated 

 in a very interesting article by Floegel in the Journal of the 

 Royal Microscopical Society, 1884, vol. 4, p. 665, t. 9, figs. 21 and 

 22, and by Otto Miiller in the Deutschen Botanischen Gesellschaft, 

 1898, vol. 16, p. 387, t, 26, fig. 5, and 1899, vol. 17, p. 435, t. 29, 

 figs. 1 to 5. Both these authors figure and describe the valve 

 as consisting of honeycomb-like hexagonal chambers, which are 

 open at the outer surface and closed by a very thin perforated 

 plate at the inner surface of the shell. Floegel made sections of 

 the valve, and from his drawings of what he saw one would 

 expect his interpretation to be correct. Miiller's interpretation 

 is substantially the same. I have not made sections, but from 

 repeated observations of the external appearance of the valve 

 I am convinced that their interpretation is not correct. If 

 the outer surface of the shell of T. favus is examined under a 

 binocular, with a y^th oil-immersion objective, using either oblique 

 light or oblique light reflected from the under surface of the 

 cover-glass upon the object (the Leitz dark-ground illuminator, 

 when decentred, acts admirably for this purpose), a thin plate of 

 silex closing the external opening is very distinctly evident, for 

 light-reflections and shadows can be very clearly seen upon it, 

 and are seen to move over its surface when the mirror is slightly 

 moved. The appearance is represented in fig. 15, made from a 

 camera-lucida drawing, in which the outline was made by viewing 

 it under a monocular, with central light, at a magnification of 

 1,500 diameters, and the shading put in to show its appearance 

 as seen under a binocular at the same magnification with oblique 

 light, the chamber chosen being midway on the slope between the 

 apex of the convexity of the outer surface of the valve and the 

 margin. This closing membrane I believe to be very thin, and 

 probably any section of it that Floegel made would be nearly or 

 quite invisible, and therefore easily overlooked. I fail to detect 

 any pores in it, although I have examined it by several methods 

 of illumination ; but at the same time there is a faint indication 

 of some kind of fine-grained surface which may ultimately prove 

 to be pore-structure. 



Upon examining the inner surface of the valve at the same 



