Wallich, on the Diatom-valve. 
14t 
ing as tliey are situated on plane or curved portions of the 
valve ; thus proving most clearly_, that the figure of the cel- 
lules is modified^ as I have asserted^ by the configuration of 
the entire frustule. 
Advancing from the discoidal group^ presenting haxagonal 
cellulation^ we are gradually led_, first through Campylo discus, 
and then Surirella, to the Naviculoid group^ exhibiting what 
is called " striation/^ 
In Campylodiscus cribrosus we meet with circular or sub- 
circular markings. In C. Hodgsonii, the circular and the 
canaliculate are combined_, or rather occur at definite portions 
of the valvular surface; the canaliculi being arrayed con- 
formably to the curvature of the valve. This last feature is 
to be seen still more markedly in C. spiralis, whilst in 
Surirella fastuosa we have the connecting bond between 
Campylodiscus and the Naviculoid forms; specimens of 
C. fastuosa, from the Channel Islands, presenting a distinct 
flexure at right angles to the true axis of the valve, as in the 
last-named genus. 
Although our knowledge of the precise share taken in the 
secretion of the siliceous valve, by the primordial utricle, is 
lamentably deficient, certain facts crop out, here and there, 
which it may be well to record under one head, with a view 
to facilitate further inquiry. We know, for instance, that the 
frustules of the Diatomacese^like the fronds of the Desmidiacese 
are encased in a gelatinous layer or envelope. In some 
genera, this envelope is highly developed ; in others it is not 
so. But from the invarible obscurity of the markings upon 
all, until the siliceous surface is cleaned by the application of 
acids or heat, it is certain that such envelope exists indis- 
criminately in the whole tribe. Judging, therefore, from the 
impermeable character of the siliceous wall, it is highly 
probable that the gelatinous stratum is secreted by the pri- 
mordial utricle, through the marginal aperture of the valve, 
much in the same way as the epiderm of the molluscous 
shell is secreted at the margin of the mantle. Like the 
latter, moreover, it is probably intended as a protection to 
the subjacent structure. Of its highly elastic nature we have 
ample proof, as was shown in a paper communicated by me 
to this Society a short time ago. We can hardly doubt its 
vitality, therefore; and we are thus furnished with presumptive 
evidence that the invisibility of the motile filaments, whose 
existence I endeavoured to demonstrate inferentially, is due 
to the same cause that enables this gelatinous stratum to defy 
our optical appliances. 
Another element of difficulty in the resolution of valvular 
