"It is not difficult to draw a hypothetical plan of the development of the air-bladder as a breathing organ. In the two families of fishes just mentioned, whose air-bladders indicate that they once possessed the air-breathing function and have lost it, we perceive the process of formation of an air-breathing organ beginning over again under stress of similar circumstances. The larval development of the air-bladder points significantly in the same direction. In fact we have strong reason to believe that air-breathing in fishes was originally performed, as it probably often is now, by the unchanged walls of the oesophagus. Then these walls expanded inwardly, forming a simple cavity, partly closed by a fold of membrane, like that of the Ophiocephalidæ. A step further reduced this membranous fold to a narrow opening, leading to an inner pouch. As the air-breathing function developed, the opening became a tube, and the pouch a simple lung, with compressing muscles and capillary vessels. By a continuation of the process the smooth-walled pouch became sacculated, its surface being increased by folding into breathing cells. Finally, a longitudinal constriction divided it into two lateral pouches, such as we find in the lung of the Dipnoans. This brings us to the verge of the lung of the amphibians, which is but a step in advance, and from that the line of progress is unbroken to the more intricate lung of the higher land animals. "The dorsal position of the bladder and its duct would be a difficulty in this inquiry, but for the fact that the duct is occasionally ventral. This dorsal position may have arisen from the upward pressure of air in the swimming fish, which would tend to lift the original pouch. But in the case of fishes which made frequent visits to the shore new influences must have come into play. The effect of gravity tended to draw the organ and its duct downward, as we find in the Crossopterygians and in all the Dipnoans, and its increased use in breathing required a more extended surface. Through this requirement came the pouched and cellular lung of the Dipnoans. Of every stage of the process here outlined examples exist, and there is great reason to believe that the development of the lung followed the path above pointed out.