678 A. W. WATERS ON FOSSIL CYCLOSTOMATOUS 
surface at any rate, and sometimes over all, larger openings giving a 
peculiar ornamentation to the surface; these large surface-openings 
do not, however, enter directly as such into the zocecial tube, but at 
the base there are one or more minute openings leading to the zocecial 
cavity (see Pl. XXX. fig. 8). In Idmonea radians these large 
exterior openings all occur at the junction of two zocecia; and in 
some Hornere, as H. reticulata, the posterior surface is divided into 
reticulations, in each of which there is one or more of these large 
openings. 
Still another character which may be of some use in testing the 
correctness of classification is the presence of ‘‘numerous minute 
rays” (Crag. Polyz. p. 122) in the zocecia and canaliculi, or as [ 
called them in Lichenopora radiata (Bry. Naples. p. 276) “ delicate 
hair-like teeth,” which occur in several Lichenopore and then have 
globular terminations. 
In Heteropora pelliculata, Prof. Nicholson has figured them (Ann. 
& Mag. Nat. Hist. 1880, p. 8), but he does not seem to have noticed 
that they had globular terminations, though in sections made from a 
specimen which he kindly gave me I find this to be the case (see 
Pl. XXXI. fig. 28). 
In what I call Heteropora cervicornis (Recent Heterop. Journ. R. 
Micr. Soe. vol. ii. p. 392), which MacGillivray has since described as 
Densipora corrugata, there are short obtuse teeth inside the zocecia, 
and in some Cyclostomata these are still further reduced, so that the 
inner surface appears irregular without definite teeth. This is, 
however, a character that will probably but seldom be available in 
studying fossils. 
The examination of the Chilostomata has shown that the mode of 
growth is in most cases of secondary importance, and that the form 
of the zocecia must be considered as of far greater value. This not 
unnaturally leads us to see that with the Cyclostomata also slight 
differences of growth must not be used as generic characters ; but 
wemust not allow this to carry us too far, though what has been learnt 
in the Chilostomata will convince us that so long as the classification 
of the Cyclostomata is based upon so few characters we may yet be 
far from a natural one. 
Such characters as whether there is one layer of cells, as in the 
Lepralic of the older writers, or two, back to back, as in Eschara, or 
whether the zoarium is reticulate, have been unhesitatingly abandoned 
as of no generic value ; and in the same way there are reticulated 
Idmonee, Hornere, Filisparse, and many other genera, and the 
differences between Diastopora, and Berenicea or Mesenteripera 
cannot be considered as very great; but in the Cyclostomata there 
are some such characters as the occurrence of definite bundles of 
zocecia in Frondipora, Fasciculipora, &c., which must be looked 
upon as of far different importance, and in the same way the presence 
of interspaces *, forming cancelli, between the zocecia does so far 
seem a character of great value, and these seem to indicate a different 
origin of the zocecial tube. 
* By using the word “interspaces”’ it will be seen that I do not accept Prof. 
Nicholson’s suggestion that the cancelli ‘‘ were occupied by a set of zodids.” 
