INTRODUCTION. 
XXV 
skull; and Hasse four great suborders, based upon the condition of 
the axial skeleton. 
With regard to the earliest of these classifications—the subdi¬ 
vision into Sharks and Rays—it must be regarded as approximately 
■natural. For although some Sharks (e. <7., Scylliidse) live upon the 
sea-bottom almost as constantly as the typical Rays, and although 
a certain amount of depression of the trunk and elongation of the 
slender tail naturally result from this circumstance, the pectoral 
fins never tend to enlarge, and the anal fin in no case disappears. 
On the other hand, even in such little-modified members of the Ray- 
series as the freely-swimming Pristidae, the pectorals have so far 
enlarged as to grow forwards and turn the gill-clefts to the ventral 
aspect, while the anal fin is completely wanting; and every grada¬ 
tion can be traced from this type to the most modified Trygonidae 
and Myliobatidse. 
At the same time, it must be remembered that, if the two sub¬ 
divisions just mentioned are solely defined in the ordinary manner 
(i. e ., Sharks with lateral gill-clefts, and Rays with ventral gill- 
clefts), all survivors of the primitive families of the Ray-series will 
become included unnaturally among the Sharks. The Squatinidae 
and Pristiophoridae, for example, possess lateral gill-clefts, like 
Sharks ; but the structure of the vertebrae, the partial growth 
forwards of the pectoral propterygium in Squatinn, and several 
striking resemblances existing between Pristiophorus and Pristis and 
Rhinobatus, all point to the Squatinidae and Pristiophoridae as pro¬ 
bably survivors of ancestral Rays. Moreover, the lowly family of 
Spinacidae may be as justly placed in one group as in the other, 
so far as the situation of the gill-clefts is concerned; but from the 
circumstance that in some of the typical genera (e. g., Acanthias ) 
considerable depression of the trunk is accompanied by a semi- 
ventral disposition of the clefts, while the anal fin is totally absent, 
it seems most philosophical to place the family provisionally with 
the less differentiated Rays. 
External features, indeed, though suggesting a broad natural 
classification, do not suffice for precise subordinal diagnoses; and 
it is therefore necessary to take into account the distinctive features 
presented by the endoskeleton. 
It might be supposed, at first sight, that the various modifications 
of the cranium and mandibular and hyoid arches would afford 
some satisfactory basis for the definition of subordinal groups ; but 
Palaeontology combines with modern zoological results to demon- 
