HUXLEY, ON THE SKELETON OF FISHES. 43 
structure is similar in principle. The tail is truly heterocercal. 
What answers to the urostyle is divided into two portions— 
the anterior of which supports the anterior hypural apophysis, 
the posterior the posterior; and the last is not only superior 
to the anterior hypural apophysis, as is the case in the 
Gasterosteus, but projects beyond it posteriorly. Seeing the 
close resemblance in the structure of the tail which exists 
among all Acanthopteri—inasmuch as the hypural apophyses 
resemble more or less closely those of the stickleback, and 
always bear the principal caudal fin-rays, I make no doubt that 
what is true for Gasterosteus is true for all, and by a parity 
of reasoning, that what is true for Anguilla and Salmo is good 
for all Malacopteri; and I therefore do not hesitate to draw 
the conclusion that the Ctenoidet and Cycloidei of M. Agassiz, 
so far from being homocercal, are in truth excessively hetero- 
cercal; that is to say, more completely heterocercal than the 
great majority of Hlasmobranchit. 
In the heterocercal tails of the Teleostei there are, how- 
ever, at least two well-marked varieties or grades of structure 
—the one, which might be called gymnochord tails, having 
the end of the notochord unprotected by ossification in its 
wall, asin the Steguri of Heckel; the other, which might 
be termed steganochord, having the end of the notochord 
enveloped in a styliform osseous coat, which there seems 
reason to believe represents the centra of two vertebre. 
As a common, if not universal, character of the Teleostean 
heterocercal tail, by which it is distinguished from that of 
Elasmobranchii, we have the peculiar development of the 
epiural and hypural apophyses. 
But if it be true that all Ctenoids and Cycloids are hetero- 
cercal, it is clear that the ground of the argument of MM. 
Agassiz and Vogt is completely cut away, so far as mere 
heterocercality goes. The ancient and the modern fishes 
are precisely on the same footing, and if the palzozoic 
Ganoidei and Elasmobranchii really represented embryonic 
conditions of existing Teleostei, they ought to be all strictly 
homocercal, for it is only m the embryonic state that a 
Teleostean is really homocercal. 
On the other hand, however, if homocercality and hetero- 
cercality are left out of the question, there can be no doubt 
that such facts as those brought forward by Heckel respecting 
the Pycnodonts show that in certain families of fish, at any 
rate, there has been a gradual change from a quasi-embryonic 
condition of the vertebral column to one more resembling 
that of an adult Teleostean. So perhaps it may be admitted 
that the structure of the tail in some modern Ganoids is more 
