27 
that tlie class Pisces slowly changed from its palaeozoic or 
embryonic to its modem and adult structure. At a time 
when it was imagined that the older fishes and reptiles and 
Crustacea were in all respects larval or embryonic, the sup- 
position was at least plausible. But the doctrine of Pro- 
gressive Development, consistent as it doubtless appears with 
those views of modification of species which the writer holds 
in common with so many naturalists of our day, derives no 
support from palaeontology. Silurian Crustacea, Devonian 
Pishes, Carboniferous Amphibia, Jurassic Mammals are in 
no sense embryonic. What animals of highly general struc- 
ture may have inhabited the seas and plains of the pre- 
Cambrian world we know not. So far as the accessible 
stratified rocks and their revelations are concerned, the forms 
both of animal and of vegetable life are differentiated and 
complex from the first. 
Later research has in no way confirmed the hypothesis of 
Agassiz. Yogt has since (1845) shewn that the adult salmon 
is not only heterocercal like the young, but heterocercal to a 
much greater degree. Huxley has demonstrated a different 
kind of asymmetry in the Stickleback and EeL Here the 
chorda dorsalis (the aggregation of cells corresponding to the 
vertebral centra) is prolonged beyond the extremity of the ver- 
tebral column, and enclosed within a hollow styloid bone 
(the urostyle). Kolliker has detected so many fresh cases of 
extremely heterocercal tails under the guise of homocercity, 
that he even doubts whether a single instance is to be found 
of a truly symmetrical termination of the vertebral column in 
fishes. 
The process of growth which ends in this singular trans- 
formation is as follows : — A young salmon just emerged from 
the egg is already slightly heterocercal. The chorda dorsalis 
has an upward turn which becomes from day to day more 
pronounced. The hoemal arches or lower processes of the 
