§ 299. A differentiation of another order occurring in the alimentary canal, is that by which a part of it is developed into a lateral chamber or chambers, through which carbonic acid exhales and oxygen is absorbed. Comparative anatomy and embryology unite in showing that a lung is formed, just as a liver or other appendage of the alimentary canal is formed, by the growth of a hollow bud into the peri-visceral cavity, or space between the alimentary canal and the wall of the body. The interior of this bud is simply a _cul-de-sac_ of the alimentary canal, with the mucous lining of which its own mucous lining is continuous. And the development of this _cul-de-sac_ into an air-chamber, simple or compound, is merely a great extension of area in the internal surface of the _cul-de-sac_, along with that specialization which fits it for excreting and absorbing substances different from those which other parts of the mucous surface excrete and absorb. These lateral air-chambers, universal among the higher _Vertebrata_ and very general among the lower, and everywhere attached to the alimentary canal between the mouth and the stomach, have not in all cases the respiratory function. In most fishes that have them they are what we know as swim-bladders. In some fishes the cavities of these swim-bladders are completely shut off from the alimentary canal: nevertheless showing, by the communications which they have with it during the embryonic stages, that they are originally _diverticula_ from it. In other fishes there is a permanent _ductus pneumaticus_, uniting the cavity of the swim-bladder with that of the gullet: the function, however, being still not respiratory in an appreciable degree, if at all. But in certain still extant representatives of the sauroid fishes, as the _Lepidosteus_, the air-bladder is “divided into two sacs that possess a cellular structure,” and “the trachea which proceeds from it opens high up in the throat, and is surrounded with a glottis.” In the _Amphibia_ the corresponding organs are chambers over the surfaces of which there are saccular depressions, indicating a transition towards the air-cells characterizing lungs; and accompanying this advance we see, as in the common _Triton_, the habit of coming up to the surface and taking down a fresh supply of air in place of that discharged. How are the internal air-chambers, respiratory or nonrespiratory, developed? Upwards from the amphibian stage, in which they are partially refilled at long intervals, there is no difficulty in understanding how, by infinitesimal steps, they pass into complex and ever-moving lungs. But how is the differentiation that produces them initiated? How comes a portion of the internal surface to be specialized for converse with a medium to which it is not naturally exposed? The problem appears a difficult one; but there is a not unsatisfactory solution of it.