1891.] NEW-YORK MICROSCOPICAL SOCIETY. 75 



and an intermediate one consisting of a hexagonal grating or a 

 plate perforated with holes. The three in apposition would con- 

 stitute the shell with its alveoli. As a matter of fact, I do not 

 know that more than two such plates have yet been discriminated 

 as components of the diatom-shell, and my present belief is that 

 there are but two, unless Mr. Smith's investigations shall establish 

 a third. As I understand his observations, they are consistent 

 with the foregoing general theory of the shell-structure, and are 

 the basis for a theory of a fibrous film exterior to those above 

 described and superposed upon them. To this I shall presently 

 return. 



That the structure of the diatom-shell is the same in very thin 

 and finely marked species as in the more robust, has always 

 seemed probable to investigators, and the argument from analogy 

 has been used with confidence. This, however, has not been 

 allowed to discourage the study of the finer forms, and the im- 

 provement of our glasses has been utilized by bringing finer and 

 finer details under direct observation. The separability of the 

 shell into two plates has been noted not only in small and finely 

 marked Coscinodisci, like C. subtilis, but in the Actinocycli. In 

 PUurosigma {F. fermosiim, P. angulaiuni, and P. balticuni) the 

 broken margin showing what Mr. Smith calls the "postage stamp '" 

 fracture has been seen and photographed, and in a number of 

 photomicrographs of my "broken-shell series" (June, 1884) I 

 showed this and what 1 regarded as separated upper and lower 

 films of the valve of P. angu latum in the same specimen.* 



Such being the state of our knowledge and theory on the 



* la a summary at the end of a series of articles on " Structure of the Diatom-Shell " 

 in the Am. M. M. J. (March to June, 1884) I said, p. 109: 



"1. The diatom-shell is usually formed of two laminae, one or both of which may be 

 areolated, and may be strengthened by ribs, which have been described both as costse 

 and as canaliculi. 



"2. The normal form of the areolae is a circle, and these when crowded together take 

 a hexagonal and sub-hexagonal form. 



"3. The areolae are properly pits or depressions in the inner siu-face of one of the 

 laminae, so that when two laminae are applied together the exterior surfaces of the 

 shell thus formed are approximately smooth and the cavities are within. 



"4. The apparent thickening on the exterior of the lines bounding the areolae in 

 some species, as Eupodiscus argus, etc., is not in contravention of, but in addition to, 

 the formation above described. 



" 5. However fine the marking of diatom -valves may be, the evidence from the color 

 of the spaces between the dots and of the dots themselves supports the conclusion that 

 they follow the analogy of the coarser forms, in which both fracture and color are 

 found to prove that the dots are areolae and the weaker places in the shell." 



