26 
less to describe the arrangement of Agassiz. In geological 
text-books it still holds a respected place, tjiough banished 
from zoology. We have four groups based upon tegumentary 
organs, ganoid, with (usually) rhomboidal or polygonal 
scales ; ^^lacoid, with shagreen or prickly tubercles ; ctenoid, 
with membranous scales jagged on the hinder edge ; cycloid, 
with rounded scales destitute of enamel. 
Agassiz further called attention to the distinction between 
heterocercal fishes (such as have the superior lobe of the tail 
larger, and containing a prolongation of the vertebral 
column), and homocercal (such as have a symmetrical tail). 
He believed that all fishes anterior to the chalk were hetero- 
cercal. It is not unusual to see the two terms used as if they 
designated two groups or orders of fishes, but no such syste- 
matic value is assigned by Agassiz to the distinction, although 
he exaggerated its importance, and misconstrued its anato* 
mical significance. 
A few words will show how superficial and destitute of 
classificatory importance is the distinction of heterocercal and 
homocercal tails. Many years ago Yon Baer and Yogt shewed 
that the tail is unsymmetrical in the embryonic state of the 
Carp, and also of Coregonus, a genus of Salmonidce. This 
curious fact was regarded by Agassiz as reconcileable with the 
homocercal character of the Salmonidce. He seems for a time 
to have believed that all fishes are heterocercal in an early stage 
of development, but that the true homocerci become symmetrical 
in the adult state, whereas the plagiostomous fishes and others 
never pass beyond the embryonic stage. Looking upon the 
geological distribution of the two types of structure, the 
heterocerci being supposed to include all known, palaeozoic 
fishes, Agassiz maintained that there was a parallel between 
the development of the individual and the development of 
the class. Just as the Salmon or Carp is first heterocercal and 
grows gradually homocercal, so it seemed possible to affirm 
