06 
Geeville, on New Diatoms. 
In the figure above referred to^ the side view of the valve is 
only -0028" in lengthy and I have seen it still smaller. In 
the trilocular form, however, under consideration, the length 
of the valve extends from -00-15" to -0088", the latter 
measurement being greatly beyond any recorded specimens 
of even the septemlocular valve. A very noticeable character 
invariably accompanies this condition of the species, in the 
different position of the spines and arrangement of the 
cellules in the central elevation of the valve. In the valve, as 
hitherto known, the cellulation of the middle elevation radi- 
ates from the centre, where two or three small spines are 
placed, as is accurately represented in SmitVs ^ Synopsis.' In 
the valves before us, on the contrary, the cellules do not 
radiate from the centre; for a transverse line crosses the 
elevation on which the spines, sometimes in a row, sometimes 
at each extremity, are placed, and the cellules are given off 
at right angles to this line, except at each end of it, where 
they radiate, as from a centre, towards the margin. In one 
of my figures it will be seen that there are six spines in a 
row, while in the larger valve the spines only occupy each 
extremity of the transverse line. It is true that in fully 
developed states of the species, as in large examples of 
quinquelocular and septemlocular individuals, there is some- 
times seen a somewhat oval centre of radiation, but this is 
evidently only a slight deviation from the normal disposition, 
and quite different from the transverse line invariably present 
in the valves now figured (figs. 3^ 4). The slight deviation 
above mentioned takes place, as I have remarked, (and that 
only rarely) , in large, fully developed frustules ; but in the 
case now before us the more important deviation occurs in 
trilocular frustules. Some difference also appears to exist in 
the outline of the side view of the valve, the ends being some- 
what sharper, and consequently rendering the form more 
decidedly elliptical. The divisions in the front view, again, 
are more turgid, and the middle elevation more pro- 
minent. 
When frustules are observed considerably larger than the 
average size of the species to which they belong, they are 
generally pronounced to be sporangial. No conjugation has 
been observed in the BiddulpMce , but it has been assumed 
that something analogous to it must take place; and, according 
to one of our most acute and original observers, Mr. Thwaites, 
^^'^the two kinds of endochrome may be developed at the 
opposite ends of one frustule as easily as in two contiguous 
frustules, and give rise to the same phenomena as ordinary 
conjugation. However this may be, we do not as yet know 
