Identity of Navicula crassinervis, &e. By W.H. Dallinger. 177 
may have relatively coarse striation. But the arrangement of the 
hemispheres is, in frustules of all sizes, unalterably the same. 
Now as we do not select the variable but the invariable 
characteristics of the dog to define it specifically, on what principle 
should we ignore an énvariable characteristic of N. rhomboides 
(and all other Naviculz), and attempt to determine it specifically 
upon characteristics which are found, when large numbers are 
examined, to be constant only in their variability ? 
That this variation is ever recurring, I may further illustrate. 
I have lately been examining a fine gathering of N. rhomboides 
from Llyn-cum-Bychan, North Wales. The frustules are very 
large, and the beading or striation comparatively coarse. They 
were taken in the living state, and only cleaned for examination, 
and not for mounting. Fig. 2, A, B, C, gives the outlines of forms 
of frustules found repeatedly. I need not point out the varieties 
of form or “shape” which they present; they are sufficiently 
obvious. But I may remark that in all the “striation” was 
identical. Now I would ask, do these varieties of “shape” con- 
stitute specific differences? or rather, are they held to do so? 
Because, if they are, we should have a distinct and scientific method 
for their detection. 
Again. In an examination of a good-sized pocket-collecting 
bottle full of “ material,” in a well-cleaned state, of smaller forms, 
such as have been called F. Saxonica and N. crassinervis, I have 
detected and drawn the forms seen in Fig. 3, A, B, C, no less than 
five times, quite independently of each other ; but the striation was 
alike in all. 
Finally, the “nodules” were commented on as exhibiting 
variety, In my drawings; and this was taken as, apparently, 
important, and the form of the nodule combined with the “ shape ” 
of the whole frustule was suggested as the basis of specdfic de- 
markation. I can only say that it would have been by no means 
impossible to have selected “nodules,” in every case, alike; and 
however unlike any two frustules may be, it would be possible to 
find the intermediate links. In Fig. 4 are the midribs of three 
forms, occurring in the Llyn-cum-Bychan and “ Cherryfield ” 
gatherings not infrequently; and the same are also to be found 
in “Lancashire” and “Bennis Lake” cleaned material. These 
special varieties were drawn from the Llyn-cum-Bychan speci- 
mens, and it will be seen that B is simply a link between the 
extremes A and ©. And whilst it is not invariably so, I find a 
tendency towards the association, in the frustules, of such a nodule 
as C, Fig. 4, and such a termination or apex as A, Fig.1; whilst a 
midrib terminating as in A, Fig. 4, is associated with such an apex 
as C, Fig. 1. This tendency is (quite accidentally) depicted in the 
drawings sent with my last paper, and still preserved in the Plates. 
So far am I from seeing the need of multiplying species on 
