400 SMITHSONIAN MISCELLANEOUS COLLECTIONS [vol. 5° 



scheme shows that the enlargement is greatest at the posterior end 

 of the thorax, and decreases anteriorly. The abdominal sacs must, 

 therefore, take in the most air, the sacci intermedii posteriores less, 

 the sacci intermedii anteriores still less, and the saccus interclavicu- 

 lar least of all. As the external wall of the thorax, the respiratory 

 muscles, and the lungs are all symmetrical, we would, if the air-sacs 

 were merely accessory bellows for the lungs, also expect that each 

 pair of sacs would be symmetrical. And this symmetry should be 

 most clearly pronounced in the largest of all the sacs — that is, the 

 abdominal ones — because, as above mentioned, these sacs are the ones 

 which change their volume most during respiration. In fact, how- 

 ever, this pair of sacs is altogether asymmetrical and dissimilar in 

 size. It is true that the sacci intermedii posteriores are, like the 

 sacci abdominales, asymmetrical, and dissimilar in size in a reverse 

 sense, but their capacity is so small when compared with that of the 

 abdominal sacs that they do not by any means compensate the dis- 

 turbance occasioned by the asymmetry of the latter. The power 

 necessary to work a small air-sac efficiently as a bellows is very dif- 

 ferent from that required to work a large one. The asymmetry of 

 the abdominal sacs would, therefore, if they really were, as Baer 

 believes, bellows, make it necessary that the muscles compressing 

 them should also be asymmetrical, which is not the case. The sacci 

 intermedii anteriores and posteriores often consist of two parts, one 

 of which is frequently connected with the other by means of a small 

 opening, but often not directly connected with the lung at all. The 

 complete separation of this air-space from the lung can only have 

 been produced by a secondary closing of the original opening. Such 

 a secondary closing would certainly not have occurred if this space 

 were a bellows employed for breathing. Sometimes, through this 

 bipartite division, the air-sac proper is greatly reduced in dimensions, 

 often by half its size: another asymmetry. Besides, it may be 

 remarked, that the two ostia forming a pair, especially those of the 

 interclavicular sac, are often asymmetrically placed. 



All this appears in a very different light if we consider the thoracic 

 air-sacs as structures selectively developed for the purpose of increas- 

 ing the size of the thorax without increasing its weight and facil- 

 itating the movements of the organs in the thorax, chiefly the heart. 

 Then it becomes evident that the shape of the air-sacs is of no im- 

 portance, and that their asymmetry is simply caused by the asym- 

 metry of the spaces they have to occupy between the asymmetrical 

 viscera. 



