REVIEWS. 
277 
formations, until at present its sole representative is tlie African Foly- 
pterus. The genus which approaches it most closely is the Gyroptychius of 
M'Coy. Glyptopcmus is another closely-allied genus, as is evidenced by 
the structure of its skull. The angles of the scales of Gyroptychius are 
apt to become rounded off so as to present a transition from the rhomboid 
to the cycloid contour. It is then the less surprising to find fishes with 
cycloid scales so similar in their organization to Grlyptolsemns, Gyropty- 
chius, and Glyptopomus as imperatively to demand a place near them in 
any natural arrangement. The description of Holoptychius, compared 
with that of Glyptola:mus, Avill show their essential alliance, but the scales 
are inform and sculpture widely different. Platygnathusis closely allied 
to Holoptychius. Ghjpiolepis, with its remarkably ornamented scales, is 
closely allied to Holoptychius. These six genera, then, possess characters 
in common, and constitute a family of Ganoids, the Glyptodipteeini, 
and which may be subdivided into a rhombiferous group, containing 
Glytolcemus, Glyptopomus, and Gyroptychius, with diphy cereal tails ; and 
a cycliferous group, containing Holoptychius, Platygnathus, and Glypto- 
lepis. The family of vSaueodipteeini, distinct from, although allied to the 
Glyptodipterini, comprises not only the genera Osteolepis, Diplopterus, 
and Triplopterus (?), but also, Professor Huxley believes, the Mega- 
lichthys of the Coal. The Saurodipteri]ii and Glyptodipterini being sepa- 
rated from other pala30zoic fishes as well-defined but closely-allied families, 
the author goes on to consider what others can be ranged with them, or, 
in other words, what are the limits and what the importance of the larger 
group formed by the association of these families. The Ctenododi- 
PTEEiNi, a family just established by Pander for the reception of Dipterus 
and its immediate allies, must, he considers, take its place in close juxta- 
position to the Saurodipterini and Glyx^todipterini, seeing that it possesses 
all those structural peculiarities which are common to those two fiimilies ; 
but the former differ in the smoothness of their scales and other points, 
but chiefly in the peculiar form of the lower jaw, which much resembles 
that of a Coelacanth, and in their dentition. In the next place, the true 
CcELACANTHiNi havc a no less well- defined right to occupy a similar position, 
but the Professor restricts the group to Coelacanthus, Undina, and Ma- 
cropoma. The type species of Cadacanthus, that on which the genus was 
founded by Agassiz, is the C. granulatus of the Magnesian Limestone, 
which in all the great features of its organization is similar to Undina ; so 
that, contrariwise, any fish which difl'ers in essentials very widely fr jm 
Undina can be no Coelacanthus. As the case stands, then, there is no 
evidence of the supposed distinction between Coelacanthus and Undina; 
while, on the other hand, a recent comparison of well-preserved speci- 
mens of Undina and Macropoma has led to the conviction that these two 
genera are not much less closely allied. All the structural characters 
which are among the peculiarities of Undina are equally well marked in 
Macropoma, except that the teeth are more distinct and cylindrical. But 
further than this, as Dr. Mantell originally suspected and as Professor 
Williamson has since demonstrated, Macropoyna exhibits the peculiarity, 
seemingly without a parallel among fishes of other families, of having the 
walls of its air-bladder ossified. Now Professor Huxley finds good evi- 
dence of the existence of a similarly ossified air-bladder, not only in 
Undina, but in a ell-preserved specimen of a new genus of CcDelacanth* 
from the Lias, in the Museum of Practical Geology. Thus it appears to 
be certain that fishes closely allied to Ccelacanthus granulatus, forming a 
well-defined family, have ranged in time, with remarkably little change, 
* Holophagus Guln, described in this Decade by Sir P. Egerton. 
