36 
PROFESSOR P. M. DUNCAN ON THE PERSISTENCE OF 
lower portions of the interseptal loculi ; but it is curved and arched, and is dissepimental 
rather than horizontal and tabulate. Their fasciculate columellse and faint pellicular 
epithecas are remarkable structures ; and their costal arrangement, by which the septum 
corresponds with the intercostal space, is eminently characteristic of some Rugosa. 
They differ from the Oyathaoconidce in having an endotlieca; but their completely 
lamellar septa and their distinct costae associate them with the next, or rather the first 
family of the Rugosa — the Stauridce. 
The Stauridce were formed into a family by MM. Milne-Edwards and Jules Haime in 
1850*, and it was differentiated as follows: — 
The septa are well developed, and are formed of perfect laminae, which extend unin- 
terruptedly through the length of the visceral chamber ; they are united laterally by 
lamellary cross dissepiments, and they are arranged in four systems, usually character- 
ized by the presence of four large septa arranged in the shape of a cross. The wall is 
well developed and imperforate. 
The family contained in 1850 two genera of compound and two of simple corals. 
The first are, of course, out of the line of the present communication, except that 
one of them, the Holocystis of the Lower Greensand, offers a remarkable proof of the 
persistence of the rugose type. 
The second or simple coral genera are Polyccelia and Metri ophy l lum . 
Polyccelia has no columella, and the dissepimental tissue is in the form of horizontal 
tabu he, and in Metriophyllum the septa are grouped in four fasciculi. Had a species 
of Polyccelia a fasciculate columella and a few arched dissepiments, it would represent 
one of the Conosmilice with the tetrameral type — the Conosmilia lituolus for instance. 
The manner in which curved or arched dissepiments are associated with and follow 
tabulae in the same rugose corals may be seen in many specimens of Carboniferous 
species, so that the distinction between the two conditions is not so great as was thought 
formerly. The absence of a columella is a generic distinction. 
Conosmilia , according to the theory of its being a persistent type, should be admitted 
into the Stauridce, in the neighbourhood of the genus Polyccelia. 
MM. Milne-Edwards and Jules Haime classify the Stauridce as the first family of 
the Rugosa, and the Cyathaxonidce as the second ; and the distinction is the want of 
endotliecal structures in the last-named natural division. 
YI. If the occurrence of a tetrameral septal arrangement in a Miocene genus in which 
there is a species with the normal Neozoic hexameral type has any significance with 
reference to older forms, corresponding phenomena should be more common in more 
ancient faunas, — that is to say, the secondary strata should contain a greater number of 
tetrameral and octomeral types combined with the hexameral than the tertiary deposits ; 
and the fossil corals of the oldest secondary rocks should retain greater evidences of 
their descent from Palaeozoic Rugosa than those of a later date. 
The following data maybe advanced in proof of the occurrence of these requirements. 
* Op. cit. vol. iii. page 324. 
