ON THE ANATOMY OF FISHES. 
245 
lost all trace of a posterior division, and the corresponding process of the fifth 
vertebra has atrophied altogether, or is rudimentary. 
The evolution of the mechanism presents no difficulty, inasmuch as the necessary 
factors are obvious in many Siluroids. The anterior division of the transverse 
})rocess of the fourth vertebra is often flexible {e.g., many Arioid genera), even while 
retaining its normal articular relations with the post-temporal, and in other Siluridac 
(e.^., Macrones) it not infrequently happens that the terminal portion of the process 
is in part formed by the ossification of the transverse membrane, while the protractor 
muscles can readily be regarded as derivatives from the anterior section of the dorso- 
lateral musculature which passes over, and is partially inserted into, the dorsal 
surfaces of the modified anterior transverse processes, on its way to the posterior 
face of the skull. 
We were at one time tempted to think that the post-temporal plates of Macrones 
and the allied genera might represent a form of “ elastic-spring " mechanism. Both 
in their relations to the anterior wall of the air-bladder and their growth as ossifica- 
tions of the transverse membrane, and the outer stratum of the tunica externa, these 
plates resemble the terminal expansions of an “ elastic-spring ” apparatus, but that 
they have the same functional significance is extremely doubtful. There are no 
special muscles for their protraction, nor any arrangement for elastic recoil. What- 
ever power of backward or forward movement they possess must be very slight, and 
must depend entirely on the corresponding swing of the loosely articulated post- 
temporal bones in conjunction with the whole of the pectoral girdle, the mobility of 
which is certainly well marked in these as in many other Siluridse. But it is 
extremely improbable that such pendulum-like antero-posterior movements of the 
pectoral girdle can have much effect on the post-temporal plates, seeing that their 
inner and ventral margins are closely attached by ligamentous fibres and by the 
dorsal edge of the transverse membrane to the crescentic distal portion of the 
modified transverse process. The connection is perhaps not so rigid but that it 
admits of a very slight forward or backward movement of the plates to an extent 
which in no case exceeds 1 mm., at all events in spirit preserved specimens. So 
slight a range of movement can scarcely be regarded as likely to produce any 
appreciable effect either on the air-bladder or its gaseous contents. For these reasons 
we prefer to consider the plates merely as bony buttresses for the purpose of 
effectually damping the anterior wall of the bladder ; they consequently aid in 
restricting any movements of vibration, or expansion and contraction, solely to the 
lateral walls of the anterior chamber. 
In addition to those genera and species that we have either been able to examine 
for the first time, or to revise, there are several others which, from the description ol 
various writers, must also be referred to the SiluridcB normales ; these are the genera 
llara, Pseudeutropius, Olyra, Chaca, Piratinga, Sorubim, Callophysus, Synodontis, 
Doras, and certain species of PlaXystonia and Arius. The details given of the 
