Specific Diagnosis in the Diatomacese. By Thomas Comber. 431 
spaces. In Eaphidieae, they may lie on either side of the raphe ; or 
transversely to the valve, forming a pseudo-stauros ; or midway 
between the raphe and the margin. In discoid forms they may be 
either central or radial. All are more or less inconstant ; yet nume- 
rous species have been proposed for extremely slight variations ; and 
in Navicula two of the main sections of the genus, adopted by Prof. 
Grrunow, the “ Lyratoe ” and the “ Hennedyeae,” have been based upon 
them, although even the two typical species, N. lyra and N. Henne- 
dyii, are themselves connected by a series of intermediate forms. 
The arrangement of the striae in the megafrustule of a species, in 
some cases differs from that in the ordinary form. A slide, tor which 
I am indebted to the late Mr. Haughton Gill, shows the megafrustule 
of a species of Amphora , with a large central blank. It therein 
differs so materially from the parent frustules, that it would certainly 
be regarded as a distinct species, had not its formation by their union 
been observed. The cultivation of diatoms, as originally suggested 
by Dr. Miquel of Paris, has probably much to teach us in this 
respect. 
The distinction formerly relied on, as separating the two genera 
Eu'podiscus and Aulacodiscus , was the presence or absence of radial 
blanks between the processes and the centre of the valve ; but more 
complete observation has shown that this feature is quite unreliable, 
and consequently Mr. Eattray, in his monograph of Aulacodiscus, 
includes in that genus even the original typical species of Eupodiscus, 
such as the old E. Argus and E. Bogersii. 
A character used for specific diagnosis, chiefly in the Crypto- 
raphidieae, which is also very variable, is the relief of the valve. The 
valve may be flat, or more or less convex, or with an elevation or 
depression in the centre ; and on such characters numerous species 
have been formed ; yet the two valves of a frustule may differ com- 
pletely in this respect ; or the terminal valves of a filament differ 
considerably from the others. A notable instance occurs in Tricera- 
tium Montereyii Brightwell, described as differing from T. arcticum 
in the centre of the valve being elevated in a very peculiar manner ; 
but the front view of a frustule sometimes shows that one valve is 
quite flat, the other quite conical, in fact, in the same gathering, 
specimens of these forms and of Biddulphia baldena, which is only 
another form of the same species, may be found, some with both 
valves flat, some with one valve flat and the other convex, and some 
with both valves convex. 
There only remains to notice one more character too variable to 
be at all relied on, namely, the presence or absence of spines or apiculi, 
whether marginal or variously placed on the surface of the valve. 
The species of Stephanopyxis and Systephania were mostly distin- 
guished from each other by differences in these appendages ; but there 
can be little doubt that several of them should be united. By such a 
character, too, Ehrenberg distinguished his genus Odontodiscus, now 
