676 A. W. WATERS ON FOSSIL CYCLOSTOMATOUS 
In a large number of Cyclostomata the ovicell is a nearly globular 
or pyriform sac, as in Crista; in others it is an irregular inflation, 
partly enclosing some of the zocecia; and in nearly all these cases 
the wall of the ovicell has small pores, of the same size as those in the 
zocecial wall, but much closer together than in any other part 
of the zoarium. In Jdmonea radians, Lamk., however, there is a 
much more elaborate structure; for here the front has large ridges 
and large openings, but at the side there are plates surrounded by 
raised borders, and these plates are perforated by very numerous pores 
placed close together *. In Hornera the ovicell is usually, or perhaps 
always, on the dorsal surface; and as this is the case in Jdmonea 
wregularis, perhaps it should be removed to Hornera, for which 
there seemed before some reasons; in other genera the ovicell occurs 
on the front. 
In the Lichenoporide the ovicell is, as a rule, an irregular infla- 
tion; but in a specimen of Lichenopora which no douht is LZ. nove- 
zelandew, Busk, the ovicell is considerably raised, and oceupies the 
whole of the centre of the zoarlum. ‘The central part of this raised 
ovicell is flat, and is surrounded by a raised meandering line, and 
in this flat portion there is a semicircular opening. 
There is considerable variety in the form and size of these ovi- 
cellular openings, and it seems that this character should always be 
given, if possible, in describing any species. ‘The opening is very 
frequently infundibuliform ; sometimes directed forwards; in other 
cases, as Lubulipora, sp., directed backwards. In Hornera lichen- 
oides, Smitt, and in my specimens of H. frondiculata, the opening is 
lateral ; but Mr. Busk gives lt as superior. 
The next character of the Chilostomata mentioned was that fur- 
nished by the rosette-plates, and I have shown that there is a great 
detail variety in the way in which the connexion between the various 
zocecia takes place through these rosette-plates. In the distal wall the 
number of these plates varies from 1 to 4, and in the lateral wall from 
2 to 8 or more; and further, the rosette-plate itself varies in shape and 
size, and also in the number of openings or connecting-points. Having 
wade preparations of the rosette plates of a large number of species, 
I find that almost all show some characteristic points, and consider 
that they furnish most useful specific characters ; but in the Cyclosto- 
mata there are no rosette-plates, and so we are led to ask What is the 
equivalent of soimportant a structure? and can it be used in classifica- 
tion? In order to answer this question, I have prepared a large num- 
ber of sections of the calcareous framework of recent and fossil Cyclo- 
stomata, and find that in the walls of the tubular zocecia there are 
usually two kinds of pores or openings: first, the very small ones 
which open to the exterior of the zoarium, and these we will call 
surface-pores; it is these which cause the dotted surface familiar 
in Crisia, most Idmonce, Tubulipore, Stomatopore, Entalophore, 
* From Naples I have an Jdmonea, perhaps J. ¢riforis, Hell., with very 
curious ovicells. In the place of a series of zccecia there is a raised chamber, 
in shape like the bag of a bagpipe, with a few large pores on the surface, giving 
it much the appearance of the ovicell of Hornera Frondiculata, 
