200 WALKER-ARNOTT, ON ARACHNOIDISCUS, ETC. 
in his fig. 124 a, and a’; when this occurs the median line, 
formerly straight, becomes also spirally twisted.* 
The genus Amphiprora, as appears to me, may be readily 
distinguished from Pleurosigma by attending to these con- 
siderations. The extreme thinness of the connecting zone in 
the latter renders it almost impossible to obtain a F. V., 
unless in fluid agitated by a drop of spirit of wine, while in 
the former the F. V. is readily detected, the connecting zone 
being of considerable breadth; in Pleurosigma the F. VY. is 
less in breadth than the breadth of the valve, while in Amphi- 
prora it is generally the reverse. 
I have not seen any Amphiprora with the principal or 
coarser striz oblique, although, from the facility of torsion 
in its valves such instances may occur. But as all diatoms, 
with striz composed of dots, have four rows of striz, two 
diagonal, one horizontal, and one longitudinal; and as the 
visibility of each depends, when delicate, on the position the 
valve presents to the illuminating oblique pencil of light, 
the closer or more difficult striz are sometimes seen when 
the others are not, and thus may be occasionally mistaken 
for the predominating or coarser ones, which alone are made 
use of in specific characters. I may here remark that when 
the dots are placed so as to form rectangles, the transverse 
and longitudinal lines are always the most remote, and there- 
fore predominate; and it is generally supposed that, when 
the dots are quincuncial, the diagonal lines are always most 
apparent ; but this conclusion is not correct, for when the 
diagonal lines make, with the transverse, an angle greater 
than 60°, the transverse rows are more remote than the 
diagonal, and when the angle is less than 30°, the longitudi- 
nal rows are the more remote and easily detected. In the 
quincuncial structure, therefore, the diagonal lines predomi- 
nate only when the angle of inclination is more than 30° and 
less than 60°; but the transverse and longitudinal cannot 
both preponderate in the same species. 
To distinguish Navicula from those species of Amphiprora 
which have the frustule straight and no ale, is considerably 
more difficult, and I doubt if there be any characters more 
readily available than those mentioned by Smith, viz., the con- 
* The facility of twisting and the amount seem to depend on the smaller 
or greater quantity of silex assimilated by the frustule; and this again, in 
all diatoms, varies in the same species according to the vigour of its growth, 
arising from locality, season of the year, and other incidental causes. Our 
whole knowledge of these beings is as yet in an embryo state, and will be 
best promoted by extensive morphological observations on species about 
which all are agreed ; till then the limiting characters of species, and the 
species themselves, must be very unsatisfactory. 
