XXIV 
IXTEODUCTIOX. 
and as I am not avrare of any definite separation between it and 
Discoeavea, it seems safest now to place the whole of the series 
from Discocavea to Radiopora in the Trepostomata. Discocavea 
may be a decadent Trepostome, but it is quite possible that it 
may be cancellous, in which case the genus should be removed 
from the Eadioporidae and restored, with any living cancellous, 
non-mesoporous species now included with the Discoporellidae 
(‘Lichenoporidae’), to the Cancellata. 
This question may, however, well await the result of a revision 
of the living species assigned to that family, and more definite 
knowledge of the structure of living Tholopora. 
The Classification of the Cyclostomata. 
The Cyclostomata are the characteristic Cretaceous Bryozoa, and 
the Cretaceous fauna supplies the best materials for the natural 
classification of that Order. 
The classification of Cyclostomata is beset with two great 
difficulties. One is comparatively superficial, as it is only 
quantitative ; the other is fundamental and qualitative. 
The quantitative difficulty is that the variability of Cyclostomata 
is so great that there is an irreconcilable difPerence of opinion as 
to the value of the characters used as generic distinctions. Some 
heroic authors are prepared to repudiate all questions of con- 
venience, and try to follow rigid and logical rules. They decline, 
for example, to recognize the difPerence between adnate and erect 
growth as of generic value, and the retention of Prohoscina and 
Stomatopora has been declared retrograde. 
It is no doubt true that under some circumstances a zoarium 
that, under normal conditions, would be adnate, may be forced to 
become partially free and erect. Thus if a Berenicea grow attached 
to a thin cylindrical stem, the growing edges of the zoarium will 
meet from opposite sides of the stem ; and their further adnate 
growth in this direction being thus prevented, the two edges may 
project from the stem growing back to back, as a free bilaminar 
sheet. Such cases are exceptional, and even their free portions are 
really adnate, as they consist of two sheets growing adnate to one 
another. 
Analogy with other classes of animals supports the probability 
of so great a difPerence in mode of life as that between an erect or 
