INTRODUCTION. 
IX 
an essential character in the diagnosis of subclasses of Fishes. In 
short, the terms “ Ganoid ” and “ Teleostean,” while convenient for 
use in alluding to well-defined bony-scaled types and modern bony 
fishes respectively, can no longer be employed as means of precise 
scientific expression. 
At the same time, however, that modern research has led to these 
difficulties, the combined results of comparative anatomy and palae¬ 
ontology have suggested an alternative classification, which seems to 
express ail the more important facts at present known. It is to be 
expected that any subdivision of a class into u orders ” or “subclasses” 
will gradually become less cogent as the earlier types are more 
fully revealed; but when all discoveries tend to prove that these 
subdivisions are divergent phyla, meeting only in remote antiquity, 
an approximately natural classification seems to have been attained. 
Among fishes, for example, it is now well known that, at least since 
Lower Devonian times, there have been two distinct plans of cranial 
structure, between which no definitely intermediate forms occur. 
As pointed out both by Stannius 1 and Huxley 2 , the upper segments 
of the mandibular and hyoid arches are directly fused with the 
chondrocranium in Chimcera , Protopterus , and their allies ; while they 
are loosely articulated, the upper segment of the hyoid arch forming 
a movable suspensorium, in all the Elasmobranchs and the so-called 
Ganoidei and Teleostei. These types of cranial structure are termed 
respectively the “ autostylic ” and “hyostylic ” 3 . It is now generally 
admitted that the first division passes through some of the early 
Dipnoan fishes into the Amphibia, and thus into terrestrial Verte¬ 
brates ; while it seems equally clear that the extreme specialization 
of the second division has resulted in the modern types of fishes—the 
vertebrates most completely adapted to an aquatic existence. 
Again, it will be observed that in the earliest known Palaeozoic fish- 
fauna there are representatives both of the autostylic and hyostylic 
types on the same primitive biological level, so far as the develop¬ 
ment of the appendicular skeleton and the axial skeleton of the trunk 
are concerned, but yet differing in the nature of the exoskeleton. 
Some families exhibit mere “ placoid ” dermal calcifications, traversed 
by delicate branching nutritive canals, these isolated plates not 
uniting even in the region of the branchial apparatus to form any 
covering of the clefts ; other families are well encased in dermal and 
* H. Stannius, ‘Hanclbuck der Zootoinie—Fisclie,’ (1846), pp. 18, 32. 
2 T. H. Huxley, ‘Elements of Comparative Anatomy’ (1864), pp. 195, 209. 
3 T. H. Huxley, Proc. Zool. Soc. 1876, pp. 40, 41. 
