ZOOLOGY AND BOTANY, MICROSCOPY, ETC. 
661 
alveoli of tlie shell. The interpolated ones (proceeding from above 
downwards) are at first very small, then larger but rectangular, and 
twice as long as wide, making the pattern one of alternate dots and 
rectangles ; as we pass to the right the rectangles run into each other 
obliquely, making a wavy white line, the dots of the alveoli proper 
being in the bends of the line, very much as in the longitudinal fibrils 
of print No. 11. This change, distortion, and multiplication of the dots 
is so entirely within our common experience in diatom-study, that I 
have no hesitation in explaining the longitudinal striated appearance in 
this patch as the result of the reduplicating of the dots by the intercalation 
of the rectangular ones, making in fact broken lines, which on so small 
a scale are sufficiently even to make continuous ones to the naked eye. 
On the other side of the midrib in the same print (No. 12) the rectangles 
and round dots are of nearly equal size, but they still make a faint 
longitudinal striation, diverging a little from the midrib as we pass 
from left to right. 
We thus have an ocular demonstration how a striated appearance 
may be made out of a tessellated one, when there is no question of 
continuous fibrils. Yet even this does not prove that the fibrils are not 
there. Of course all visual appearances under the Microscope have 
their cause in the structure of the object, considered in relation to the 
laws of transmitted and reflected light. The puzzle often is to tell 
what to attribute to each factor. I do not think it difficult to account 
for the tessellated appearance of dots and squares with alternate blue and 
red colour. To do so may require us to refer to some elementary 
matters in diatom-marking. 
Dr. Brebisson, at a very early day, divided the regular dotted 
markings of diatoms into three classes : (1) Quadrille rectangle droit — 
in squares parallel to midrib, e. g. Pleurosigma balticum ; (2) Quadrille 
rectangle oblique — in squares oblique to midrib, e. g. P formosum ; 
(3) Quinconce — quincunx or lozenge of 60° smaller angle, e. g. P. 
angulatum. This classification has been a good deal neglected, but has 
good claims to remembrance, and will assist me in explaining the 
phenomena before us. 
In Mr. Smith’s print No. 6 is well shown what I regard as the normal 
scheme of areolation of P. formosum. It will be seen to be a reticulation 
with meshes as nearly square as nature gives us in growing things. If the 
corners of these meshes be filled up, the included circles will still keep 
to each other the relative position of Brebisson’s oblique quadrille. 
The diminution of the round alveoli would not need to proceed far 
before the approximately rectangular mass of silex between the circles 
would be about as large in diameter as the circles themselves. Under 
the laws of optics, which we have already seen illustrated in print 
No. 12, the tendency of approximately rectangular details is to become 
more strictly so in the microscopical image. In Fig. 74 I have illus- 
trated this by a geometric diagram, of which one half shows the square 
reticulation, and the other the resulting tessellation of solid squares and 
round alveoli when the walls are thickened and the corners filled up. 
It will be noticed that when the corners are so filled as to make the 
alveoli circular, the interspaces are approximately square, and, being 
solid, will be red or pink by transmitted light when the alveoli are 
1891. ‘ 3 A 
