Greville^ on New Diatoms. 
53 
error (if it be one) lies in the curved costse being made nearly 
twice as long as they ought to be. This character is very 
constant. Another error consists in the processes being made 
to rise but little above the level of the median space^ whereas 
in all my specimens the line commences at the contracted 
neck;, leaving the heads of the processes quite clear. I have 
thought it right, in the mean time, to include Bailey^s diatom 
(or rather the figure from his drawing) in my few words of 
description ^ but it is quite possible that the two may be dis- 
tinct, and indeed highly probable, if the Gulf-stream diatom 
should prove to be a living species. 
Porpeia quadmta, n. sp., Grev. — Valve, as seen in front 
view, forming nearly a parallelogram ; the angles rounded, 
very slightly produced ; median space nearly straight, the 
costae descending vertically at the inner angle of the pro- 
cesses, and then curving inwards and approximating, but not 
quite parallel with the suture. (Fig. 20.) 
Hah. Barbadoes deposit, Cambridge estate ; in slides com- 
municated by C. Johnson, Esq. ; very rare. 
It is very interesting to find the leading character of the 
genus so conspicuously carried out in this and the following 
species. The Substance is hyaline, and, with the exception of 
a row of very minute marginal puncta in the nearly straight 
median space and at the apex of the processes, and a very 
few larger ones scattered near the curved plates, the surface 
is smooth. Length of valve '0035''; but this is a variable 
character. 
Porpeia ornata, n. sp., Grev. — Frustule forming a paral- 
lelogram ; valve in front view punctate, produced at the angles 
into very slightly convex processes, rounded at the corners ; 
median space quite straight, the costse taking an outward 
curve, then becoming exactly parallel with the suture, and 
approximated at their apices. (Fig. 21.) 
Hah. Barbadoes deposit, Cambridge estate; in slides com- 
municated by C. Johnson, Esq. ; very rare. 
A beautiful species, which, when first observed, a couple of 
years ago, puzzled me exceedingly. Several examples have 
occurred since, and there can be no doubt now regarding its 
true position. Length of valve in the finest specimen disco- 
vered -0040^^ 
Hemiaulus. 
Hemiaulus symmetricus, n. sp., Grev. — Valve in front view 
produced at the angles into linear horns, sharply truncate at 
the top, and tipped with a single spine ; structure cellulate^ 
