208 REPORT—1846, 
In urging a reconsideration of the value and significancy of these charac- 
ters, I may repeat that in mammals the mastoid constantly presents them, 
whilst the squamosal very rarely has the first, and not often the second cha- 
racter. It must also be remembered that the squamosal loses its connection 
with the frontal and progressively decreases in the mammalian class to less than 
the dimensions of the mastoid itself, as e. g. in echidna (fig. 12), whilst in this 
monotreme the mastoid, s, besides its connections with the parietal and exocci- 
pital, extends forwards to articulate with the alisphenoid, 6. If ossification 
were restricted in mammals to no. s, fig. 11, in reference to 16, which re- 
mained cartilaginous, then no. s would have the same relation to the otocrane, 
or in other words, would contribute the same protection to the acoustic laby- 
rinth, which no. s, fig. 5, performs in fishes; the external semicireular 
canal at least would be protected in the mastoid by both: only in mammalia 
the mastoid would also extend over the posterior canal. The petrosal loses 
no part of its essential character as the capsule or outer tunic of the laby- 
rinth by becoming ossified, nor is it less recognisable in fishes within the 
mastoid, by remaining membranous or cartilaginous, than is the sclerotic 
capsule of the eye in its chamber or orbit; which capsule, in like manner, 
presents all the corresponding histological modifications in one or other part 
of the vertebrate series. The*mask which has concealed the true features of 
resemblance in tke human mastoid to that of fishes, is simply the petrosal 
ossified and cemented to it. But the squamosal presents no such relations to 
the bony capsule of the semicircular canals in any mammal. Even the 
connection of the squamosal with the tympanic bone is, as we have seen, far 
less constant and intimate in mammals than the connection of the mastoid 
with the tympanic*. 
In the anatomical description of the existing ganoid fishes which M. 
Agassiz has unfortunately called ‘ Sauroid +,’ the bone no. s is described as 
* From the remark in p. 53, t. ii. pt. ii. ‘Recherches sur les Poiss. Foss.,’ it would seem 
that the circumstance of the extension of the tympanic air-cells into the mastoid, in certain 
mammalia, had weighed with M. Agassiz in determining its homological characters. 
+ All the characters by which these highly organized fishes approximate the Reptilia are 
found, not in the highest, but in the lowest order of that class, viz. in the batrachia, and herein 
more especially in the salamanders. The air-bladder of Lepidosteus resembles the lung of 
the serpent in its singleness, and those of the salamander in the degree of its cellularity ; 
some parts of the structure being peculiarly piscine. The bifid air-bladder of Polypterus 
resembles the lungs of the salamandroid menopome and proteus, in the want of cellular 
walls. The characteristic large bulbus arteriosus and its numerous rows of valves, which 
distinguish the ganoids from most other osseous fishes, are retained in the menopome, but 
are not present in any saurian. The anterior ball and posterior cup of the vertebra of Le- 
pidosteus are repeated in the salamander and pipa, but in no existing saurian. The laby- 
rinthodont character of the teeth of Lepidosteus was developed to its maximum in the great 
extinct reptiles (Salamandroides, Jager), which, by their double occipital condyle, denti- 
gerous double vomer, and biconcave vertebrz, were essentially Batrachia, not Sauria; and 
which combined characters now found in the lower salamandroid Batrachia, with the dental 
ones borrowed from fishes, and but feebly manifested by the most fish-like of saurians 
(Ichthyosaurus). All the so-called sauroid fishes retain the characteristic piscine articular 
concavity on the basioccipital for the atlas: it is, however, very shallow in the polypterus ; 
and is also extended transversely, with the lateral borders or angles so prominent, that, as 
M. Agassiz well remarks, “ it needs very little to change this transverse articulation with its 
two lateral ridges into two distinct articular condyles,” /. c. p. 71. But this would convert, 
pro tanto, the polypterus into a batrachian, not into a saurian. So far as the character of a 
single convex occipital condyle is valuable as a mark of affinity to the Sauria, it is present 
in a fish of a different order from the ganoids, and with much fewer approximations in other 
respects to the reptilian class, viz. in the Fistularia tabaccaria. There remains, therefore, 
only the character of the enamelled scales which the polypterus and lepidosteus present in 
common with all the lower organized ganoids, and which to a certain extent resemble the 
bony scutes of the crocodilia. If the deposition of calcareous matter in and upon the skin 
were not essentially a retention of a very low type of skeleton; if it were not presented by 
