GREVILLE, ON DIATOMACES. 159 
possible to determine whether the rays were centrical or ex- 
centrical. The outline also of the valve is subject to varia- 
tion. When it is more or less ovate, the rays are excentrical ; 
but when the valve (in the same species) becomes more 
nearly orbicular, the rays then become more centrical. I 
am now speaking more particularly of those species which 
have a granulated structure and general diaphanous ap- 
pearance. In Spatangidum Arachne and 8S. heptactis of 
Brébisson, the rays are always more or less excentrical, and 
seem to constitute a really permanent character, which, 
taken in connexion with the more important difference of 
structure, viz., the distinct areolation, will justify, I think, 
the separation of these two species from Asteromphalus. 
With regard to the parts of the valve on which specific 
value may be placed in the two genera, we must be guided 
by the results of extended observation. I am inclined to 
believe that, closely as these genera are allied, the same parts 
are not of equal value in both. In Spatangidum, as I pro- 
pose to restrict that genus, the number of the rays in the 
several species seem to be constant. The rays themselves 
afford good characters, as well as the lines of the hyaline 
area (umbilical rays of authors). In Asteromphalus, on the 
other hand, the radiating lines of the hyaline area (including 
their insertion on the nucleal line) seem alone to furnish 
reliable distinctions. In studying the composition of these 
beautiful discs, of both Asteromphalus and Spatangidum, I 
have been led to take the view of Kiitzing, who considers these 
radiating lines as ‘‘ sepimenta imperfecta.” They terminate 
abruptly, precisely as in Asterolampra. The true rays may 
be said to commence from the outer edge of the hyaline area, 
for they are merely blank spaces in the areolation or granu- 
lation of the disc. It is true that as the intervals between 
the lines of the hyaline area are a sort of continuation of 
these blank spaces, they may, in this point of view, be con- 
sidered as the ‘“ enlarged bases” of the rays, and thus, ac- 
cording to De Brébisson’s idea, be directly imposed on the 
“base” of the median ray, or, as I prefer to call that part, 
the nucleus or nucleal line. In A. Brookei, however, such 
an arrangement can take place to a very partial extent, as 
Professor Bailey describes the “ umbilical rays” as “ flexuose, 
some simple, others branched, or two or more uniting before 
reaching the centre.” These lines are also occasionally 
forked in a species I have found in South African guano. 
It seems to me, therefore, advisable to take such characters 
as the lines of the hyaline area afford independently of the 
rays. 
