48 
Greville, on Asterolampra. 
The most variable of all the Barbadoes Aster olamprcs, not 
only in the number of rays_, which range from six to twenty, 
and in the cellulate or non- cellulate umbilicus, but also in 
the characters derived from the segments. Discs from the 
extreme ends of the series, might readily be taken for 
distinct species ; and yet it is exceedingly difficult to define 
the varieties. In drawing up the specific character, I have 
been unable to discover any part upon which I could place 
the least dependence, except the lateral cellules of the 
segments of the valve, which seem always to project more or 
less beyond the rest ; and it is mainly to modifications of the 
same part, that I have had recourse in separating the 
varieties. There is no rule with regard to the amount of 
space in the valve occupied by the segments themselves. It 
is very frequently half the radius, often considerably less ; and 
in some of the large varieties, where the centre is occupied 
with numerous cellules, the radius is generally divided into 
three nearly equal parts, viz., the segments, the hyaline area 
transversed by the umbilical lines, and the central group 
of cellules. (PL VII and VIII, figs. 23, 24.) With regard to 
the elongated cellules of the segments, they difi'er greatly in 
length in the more remarkable varieties, as do also the lateral 
cellules, compared with the intermediate ones. The lateral 
cellules are sometimes quite straight; at others, slightly 
curved inwards towards the apex. They are generally so far 
separated as to admit of the passage of the rays between 
them to the margin of the valve ; but the passage is some- 
times partially interrupted, in consequence of the cellules 
being occasionally somewhat enlarged toward their apices, 
and touching each other at that point. (Fig. 22.) Very 
rarely the lateral cellules are united throughout their whole 
length, and, consequently, the narrow portion of the ray is 
rendered obsolete. (Fig. 21.) Perhaps the most extreme 
variety is one which I at first considered distinct, and named 
A, splendida (fig. 25) ; and even now, I am not quite satisfied 
about its true position. The extraordinary length of the 
cellules, which constitute nearly the whole substance of the 
segments, distinguish the few imperfect valves I have seen. 
But the cellules of the segments in this species are so exceed- 
ingly variable, that a mere difi'erence in length, however 
great, will scarcely, I fear, afibrd a satisfactory diagnosis. 
Of this protean diatom I have examined a multitude of 
specimens, besides taking notes of about sixty valves, in order 
to ascertain as far as possible the range of variation. 
8. Asterolampra Bright welliana, Grev. — Valve very convex ; 
umbilicus somewhat eccentrical; segments unequal, having 
their inner margin composed of large oblong cellules, only 
