276 OBSERVATIONS ON THE DEVELOPMENT OF 
Gobtide, Pediculot?, Pleuronectide, Lophobranchit, Plectognathi, and 
others).” 
I have omitted Heckel’s account of the vertebral column of 
the Pycnodonts which precedes the long and important extract here 
given, as less immediately germane to the present subject. Suffice 
it to say, that he admits altogether three modes of termination of 
the chorda dorsalis: 1. The end is naked or unprotected by any 
ossification, as in Paleozoic Fishes and existing Ganotdez. 2. Its 
unossified end is protected to a greater or less extent by lateral 
roof-like plates, as in the Sea/montd@; these Heckel calls Szegurv. 
3. The end of the chorda is enclosed within the anterior cavity of the 
body of the terminal vertebra, as in the Percid@, &c. 
I shall bring forward grounds for believing that Heckel is mis- 
taken as to this third mode of termination, and that in these fishes 
the end of the chorda really extends far beyond the anterior cavity 
of the last vertebra. 
In 1854, Stannius published (as a part of the new edition of the 
‘Handbuch’) his ‘Zootomie der Fische, beyond all comparison the 
best and most exhaustive work on the subject which has yet ap- 
peared. The structure of the fish’s tail is discussed at p. 29, but 
very unaccountably all mention of Heckel’s researches is omitted. 
In the Blenntde, Ophiodint, Tenioide, Muranoide, Fistulariea, the 
last caudal vertebra is said to end in a slight point. In Cyclopterus, 
Calltonymus, the Pleuronectida, and Plectognathi, “ the end of the 
last vertebra becomes flattened and slender, and is prolonged into 
a vertical broad plate, consisting of two quite symmetrical halves, an 
upper and an under.” 
A more detailed (but otherwise essentially similar) account to 
that of Heckel, of the tail of the Sa/monide, is next given, and the 
like structure is said to obtain throughout life in the Ganozdez, in 
Esox, Hyodon, &c., while it is transitory in Cyprinide, Characine, and 
others. 
In conclusion, Stannius points out that “many fish which pass for 
homocercal, show unmistakable traces of original heterocercality.” 
Having verified Stannius’s account of the structure of the caudal 
extremity in the salmon, but seeing no reason to doubt—what was 
generally admitted—that other Teleostean fish were truly homo- 
cercal, I pointed out, in 1855,’ that the foundation of the doctrine 
of Vogt and Agassiz was thereby destroyed. For Vogt’s observa- 
tions were made on a salmonoid fish, and a right comprehension 
of the structure of the tail in such fishes showed, that so far from the 
1 Friday evening meetings of the Royal Institution. 
