1898] IMPERFECTION OF THE GEOLOGICAL RECORD 329 
lake in the Devonian or Old Red Sandstone period, and have been 
worked for commercial purposes from time immemorial. — Fossil 
fishes have been known and collected from them for more than 
seventy years. Every exposure has been searched by expert collec- 
tors, whether in the cliffs or in quarries. Yet, Palaecospondylus has 
only been found in one very thin stratum in a single quarry, where 
it occurs, not as a rarity, but in countless numbers. It seems as if 
a shoal of the species had been accidentally destroyed and suddenly 
covered up; and it is a fortunate accident that a small quarry has 
been opened at the precise spot. 
Our knowledge of the earliest shark-like fishes exhibiting the 
most primitive type of paired fins is almost equally scanty. 
Cladoselache, in a state fit for accurate scientific study, has hitherto 
been met with only in a flagstone at. the base of the Carboniferous 
formations in Ohio, U.S.A. Teeth of the same kind have been 
known for many years from several parts of the northern hemi- 
sphere; but the complete fish has only been discovered in Ohio 
within the last decade, and even now the skeleton of the head and 
vertebral axis remains practically unknown. 
In some instances the old Palaeozoic forms of fish-life which 
withdrew to the comparatively peaceful realms of rivers and fresh- 
water lakes after the vigorous period of their race was past, and 
survived until the present day, were thus entirely lost to geological 
records. For example, three detached teeth from the English 
oolites and scarcely more from corresponding rocks in Colorado, are 
the sole known traces of the Dipnoan fishes between their world- 
wide distribution at the dawn of the Mesozoic era and the scattered 
remnants which still survive in the fresh-waters of South America, 
Africa, and Queensland. Similarly, there is no doubt that Polypterus 
and Calamoichthys existing in the fresh-waters of tropical Africa, 
are the direct and little-altered descendants of some of the Palaeozoic 
fringe-finned ganoids; but we have still not found even a trace of 
them in the Mesozoic or Tertiary strata in any part of the world. 
They must have lived somewhere, but the geological record, so far 
as explored, is too imperfect to afford a clue to their whereabouts 
and history. 
The case of the Amphibia or Batrachia is still more remarkable ; 
though perhaps they, too, have been fresh-water animals since the 
Palaeozoic era. It is definitely proved that some of the early lung- 
breathers belonged to this class ; for traces of gill-arches are occasion- 
ally observed in young individuals, showing that they breathed by 
gills in their immature state (Branchiosaurus, Archegosaurus). It is 
also certain that these primitive Amphibia were the dominant type 
of vertebrate life from their appearance until the middle part of the 
Permian period. In early Mesozoic times, however, they suddenly 
