NEW FORMS OF MARINE DIATOMACEZ. 527 
giving to the detached segment a very peculiar character. The segments are 
marked by strong, coarse strive, about 20 in 0:001’. In one focus (fig. 97) the 
segments are well seen in the frustule; in another, fig. 97 6, the frustule is seen 
to consist of 9 or 10 convergent bars, which are coarsely granulate. In fig. 97 
the lateral segments are seen to be nearly in apposition, with a narrow space 
between them, of varying width, from the undulations of the ventral margin. 
The inner line of each lateral segment is very slightly curved; the nodule lies 
nearest to the ventral margin. The detached segments, figs. 97 c, and 97 d, are 
precisely like Cymbellze, and for a long time I considered them as such, with those 
of A. granulata and A. cymbifera. But the view in fig. 97 showed the real 
nature of these forms. Want of space alone prevents me from giving figures of the 
entire frustule corresponding to the segment in fig.97 d. The reader is requested 
to compare fig. 97 with the corresponding view of A. G'revilliana, fig. 89. 
This very striking species occurs only in the stony Loch Fine dredging, so often 
alluded to, where it is rather frequent, both entire and in detached segments. 
99. Amphora cosiata, Sm. Pl. XIV., fig. 99. This species is described by 
Professor SmirH in his Synopsis, vol. i., where the entire frustule, and also a half 
frustule, are figured. But little is said of its peculiar structure, and the detached 
segment is not figured. For this reason, and to show the analogy between this 
species and the three preceding ones, I have figured a detached segment (fig. 99). 
Form of segment highly arcuate, very broad in the middle. Ventral margin 
straight, or slightly concave, but often, as in the example figured, which is a 
rather small one, with a rounded prominence in the middle close to the nodule. 
The ends are capitate. Striz coarse, conspicuous, about 14 or 16 in 0-001’, 
thicker and stronger near the dorsal margin. Length from 0-002” to 0-0033’ ; 
breadth in the middle 0:0012’ to 0-0016". 
It will be seen that in this species also the segment resembles a Cymbella, 
although it is a very broad and highly arcuate one. When the segments are 
united, as in the entire frustule, it is not easy to see their real characters. The 
backs of these segments, or longitudinal bars, are, as in Professor Smiru’s accu- 
rate figures, marked by very coarse distant granules, which give no indication of 
the peculiar striation of the detached segments. Hence it was very long before 
I was able to detect the component parts of the frustule when detached, or to re- 
fer the form shown in fig. 99 to A. costata. But some specimens, as in the pre- 
ceding species, when carefully focussed, clearly show their true nature. 
100. Amphora bacillaris, n. sp. Pl. XIV., figs. 100 and 100 6. Form of frustule 
linear, narrow, with somewhat rounded ends, which are subacute. Length about 
00017’; breadth about 0-0003’. In one focus it exhibits two lateral portions se- 
parated by a middle space, the sides of which are perfectly straight, the ends 
beautifully rounded. In another, the whole frustule is seen to be composed of 
very narrow bars, separated by very sharp lines, converging on the ends, and 
