Wallich, on the Diatom-valve. 
143 
Inasmuch as the photographic representation so constantly 
appealed to is stated to be partly in perfect focus^ and partly 
out of focus^ whereas the structure is equally distinct on 
both parts, and the only difference observable consists in 
the reversal of the lights g,nd shadows^ it is much more pro- 
bable that one portion was as much without the true focus as 
the other portion was within it. 
From the woodcut of the photographic picture^ it would 
appear that the thickness of the walls of the hexagonal cel- 
lules is equal to half of their diameter. Now^ the striae being 
stated to be composed of dots, and the lines being estimated 
at 2oi^oo^^^ inch apart, the walls must measure in round 
numbers to oVo o^^ P^i'^ of an inch in thickness ; thus pre- 
senting a surface the outline of which ought to be readily 
resolved by the same powers that show the diagonal markings, 
— for instance, by a quarter-inch objective, — did the hexagonal 
structure really exist. 
It is highly improbable that hexagonal structure should 
present itself in one species or group of a well-marked genus, 
whilst a totally different structure is admitted to exist in the 
other species of the same genus. 
In proof of the rhomboidal structure, I beg, on the other 
hand, to offer the subjoined proofs. 
Under the application of any powers, ranging from \ to -xV 
of an inch focus, so long as definition remains unimpaired, the 
rhomboidal structure is invariably discernible \ the diagonal 
lines being predominant and visible, with perfect clearness, in 
the case of the rhomboidally marked group, whereas the rect- 
angular series is so in the other. 
The object retained in one position on the stage, when 
viewed under a given power, say a x'g-inch objective and 
a low eye-piece, exhibits oblique lineation and rhomboidal 
facetted spaces, with perfect definition; whereas, by re- 
placing the low eye-piece with a high one, and making any 
alteration of focus demanded by the change, the hexagon-like 
structure exhibits itself, but with imperfect definition. 
By causing the rotation of the slide, containing either the 
rhomboidally or rectangularly marked forms, at every forty- 
five degrees a fresh series of lines will predominate, according 
to the direction of the illuminating rays; each of the four 
series being, of course, twice repeated in one complete revo- 
lution, and the change of series therefore taking place eight 
times. 
* As the longitudinal and transverse series of lines in tiie rhomboidal 
group, and the diagonal series of the rectangular group of FlenosigmaSy 
require much more careful adjustment than the predominant series, for 
VOL. VIIl. p 
I 
