Dr. R. H. Traquair—Carboniferous Selachii. 83 
archipterygium, intermediate between the truly biserial one of 
Xenacanthus and the pectoral fins of ordinary sharks. If this inter- 
pretation is correct, then, along with Xenacanthus, this specimen is 
a witness against the lateral fold theory of the paired fins, at present 
so popular with anatomists and embryologists. Into that question 
I shall enter on another occasion, meanwhile so much is clear, that 
if we have before us the pectoral fin of Cladodus, and I do not doubt 
that we have, the affinity between that genus and Chlamydoselachus 
is not quite so close as Mr. Garman maintains, seeing that in his fish 
the pectoral fin shows the ordinary arrangement of basal pieces, 
though the metapterygium has two segments. 
What then of the Ctenacanthus theory ? No spine is seen in con- 
nection with the Hast Kilbride Cladodus, but as the body is absent, 
Spines may have been borne by the fish when complete. Again, in 
the Eskdale Ctenacanthus the form and structure of the pectoral fin 
are not shown, and though I interpreted its one imperfect tooth 
as “‘Cladodont,” I am willing to leave that an open question. It 
may be hybodont, and the hybodont form with its vertically com- 
pressed base must not be confounded with the cladodont type with 
its base horizontally flattened and irregularly elliptical or reniform. 
And in one of the instances which have been advanced to prove the 
connection of Cladodus with Ctenacanthus, a mistake has certainly 
been made. So far as I have seen them, the teeth which are found 
associated with the Coal-Measure Cienacanthus hybodoides, Eg., do 
not belong to Cladodus mirabilis, Ag., as has been asserted, but are 
allied to Hybodus in their narrow, compressed non-expanded bases. 
Mr. J. W. Barkas long ago expressed his opinion that ‘‘ most of these 
so-called Cladodi are in reality Hybodi” (M. Rev. Dent. Surgery, 
February, 1874), though he seems to think that the great difference 
between Cladodus and Hybodus lies in the former having the outer- 
most denticles larger than the intermediate ones, and consequently 
admits some of these Coal-measure specimens to the genus Cladodus. 
Ctenacanthus hybodoides has therefore nothing to do with Cladodus, 
and as regards the other species, I rather think that, if we knew the 
creatures to which they belonged, they would turn out to repre- 
sent several types, possibly very different frum each other. But of 
this I have now no doubt, namely, that the Cladodontide, whether 
they had spines or not, or whatever the shape of their spines if they 
had any, constitute a very different family from the Hybodontidae— 
while the latter on the other hand were closely allied to the Cestra- 
ciontide. For if Tristychius be a Hybodont, we have now some clue 
to the structure of an ancient representative of the family. 
Tristychius, Agassiz. 
A specimen of Tristychius, from Eskdale, allied to, if not identical 
with Agassiz’s T. arcuatus, shows the greater part of the body with 
the head, one pectoral fin and two dorsal fins. Hach of the dorsal 
fins has a spine in front. The pectoral shows two large basal pieces 
which J interpret as mesopterygium and metapterygium, the proptery- 
gium being either small or fused with the mesopterygium as in 
