348 MOSTLY MAMMALS 
probable that after the young fish are hatched they are 
retained for some time by attachment to the walls of the 
chamber. In the true pipe-fishes (Syngnathus), on the 
other hand, the task of looking after the nursery falls to 
the males, which are provided with a long pouch on the 
under-surface of the tail, formed by a fold of skin arising 
on each side, and the two meeting in the middle line. 
How the eggs are conveyed into this pouch I am totally 
unaware, but when once there, they are completely enclosed 
by the junction of the edges of the two folds of skin, and 
thus remain till they are hatched into minute eel-like pipe- 
fish, which soon make their way into the world by thrusting 
open the folds of the pouch. In the sea-horses the 
development is carried one stage farther, the nursing- 
pouch being completely closed along the middle line, and 
only communicating with the exterior by means of a small 
aperture at the anterior end, through which the eggs are 
by some means or other introduced, and by which in due 
course the young make their escape. Certain pipe-fishes 
(Doryichthys) differ from the ordinary forms in that the 
males have the pouch situated beneath the abdomen instead 
of under the tail; and it is not a little remarkable that in 
certain allied genera (JVerophis, etc.) the eggs are simply 
attached to the lower surface of the abdomen of the male 
without the development of a pouch. We have thus an 
excellent instance of the evolution of a special organ, so 
far as the abdominal pouch is concerned; but it would seem 
highly probable that the caudal pouch of the allied forms 
must have been independently evolved, in which event 
we should have a remarkable example of parallelism in 
development. 
Although many fishes retain their eggs within their 
bodies until the young are hatched and attain a consider- 
