INTRODUCTION. 
XI 
osteology, so far as known, this fish seems to agree very closely 
with the existing Atherines ; but it differs in having one continuous 
dorsal fin, with the right and left halves of each spine not com- 
pletely fused together. Another genus, Syllwmus. from the English 
and North American Chalk, is nearly similar. The Crossognathidoe 
are therefore probably the Cretaceous forerunners of the Percesoces, 
which were already completely differentiated into their surviving 
families before the close of the Eoceno period. Atherina, Mugil, 
and Sphyrana all date from the Upper Eocene or Oligocene. The 
only noteworthy extinct genus, the European Oligocene Cobitopsis, 
belongs to the Ammodytidte and retains the pelvic fins which have 
disappeared in the existing genera. 
Hemibranchii. 
Among extinct Tertiary fishes there are several so closely similar 
to existing Hemibranchs and Lophobranchs, that there need be no 
hesitation in associating them with the latter. As yot, however, 
they afford no clue to the origin of this group. The comparatively 
generalised Gasterosteidoe are not represented; but all the other 
known families seem to date back to the Eocene period. Even the 
Centriscidae and Syngnathidoe have scarcely changed since that 
time. The highly-specialised Ampin sile seems to have persisted 
without any essential modification in its armour ; while Hippo- 
campus appears to have only lost its caudal fin in the interval. 
Acanthopteryqii. 
The study of the existing fauna has already suggested that the 
oldest true aeanthopterygian fishes are comprised among the 
Berycoids and Scombroids. It is tborefore interesting to find that 
most, if not all, of tho typical Cretaceous Acanthopterygii are 
referable to these two groups. The Upper Cretaceous genera are, 
indeed, very little specialised; and it seems probable that they 
actually originated at about the period of deposition of the Chalk in 
the northern hemisphere. All aro characterised by a single dorsal 
fin, no distinct cleft having yet appeared between the spinous and 
persistently soft portions. Some undoubtedly Berycoid genera 
(e. g. Hoplopteryx, p. 396) exhibit an excessive development of 
mucus-cavities on the head. 
One genus ( Prolates ), from the uppermost Cretaceous horizon in 
France, is probably a true I’ercoid. It marks the dawu of the 
