176 MILLER. [Vol. VIII. 



sac-passage and air-sac. Usually it is quite difficult to deter- 

 mine the atria and air-sacs unless we have a complete series of 

 sections to examine ; I have already mentioned several points of 

 distinction, but for general purposes attention to the following 

 points will aid one to decide. The bronchi may be recognized 

 by their containing smooth muscle fibres and by their diameter. 

 Lying in more or less of a circle about the bronchus may be 

 seen three, four or five openings nearly circular in outline and 

 separated from the bronchus by a number of smaller openings; 

 the first are the atria and the second are air-cells. On the 

 outside of the circle of atria may be seen large irregular open- 

 ings, the air-sacs, and grouped around each air-sac a number of 

 smaller openings, their air-cells. 



The above description shows that the lung of mammals is 

 constructed after the plan of that of the birds, reptiles and 

 amphibia. In general it may be viewed as a conglomeration 

 of crocodile's lungs. In crocodiles, as in birds, these atria 

 (they are not bronchi) give rise to many air-sacs. In the 

 crocodile there is but one terminal bronchus; in birds several; 

 and in mammals there are many. In both crocodiles and birds 

 each atrium gives rise to many air-sacs ; in mammals each 

 atrium gives rise to but few air-sacs (see the various cuts). 

 In reptiles the lung often shows a double structure ; one or 

 more air-bags and many air-sacs and cells. In birds this same 

 combination is retained, only there are a great many air-bags 

 extending far beyond the thorax, and in addition there is a 

 communication of the atria with one another. In mammals 

 the specialization is only single, i.e. atria with their air-sacs, 

 but no air-bags. 



From time to time it has been asserted that there is com- 

 munication between the air-cells of adjoining lobules. This 

 question I have already discussed in other papers. I think it 

 sufficient to state that in careful reconstructions it is impossible 

 to find any communication, and my corrosion preparations sug- 

 gest the same. Moreover, it is impossible to inject a fluid from 

 one bronchiole to a neighboring bronchus, which certainly could 

 be done easily were there any communication between the air- 



