small bronchi.“ RoosEVELT 1) says: „When one takes a perfectly in- 
jected cast of the ultimate bronchus and air vesicles (which looks like 
a berry) and breaks it up (observing it with a low power lens), it 
becomes evident that the interior of the berry is made up of spheroidal 
bodies which are attached to one another at two or more points. 
These are the true air vesicles and the important point for us in 
DELAFIELD’s account is that these vesicles communicate freely with 
one another. This they do since it is possible to break off a number 
together, and then, by further breaking to show that there are many 
which present corresponding rough surfaces at the line of fracture.“ 
My own experiments lead to the other conclusion, i. e. the air- 
cells do not communicate. If we take a lobe of a lung and cut down 
to the point where the first branch of any size is given off from the 
main bronchus, divide the branch from the main trunk, tie a cannula 
in the main trunk, and inject fluid, none will pass to that part of the 
lung supplied by the divided branch, but this portion of the lung will 
remain collapsed and sharply marked off from the distended portion. 
I have injected fluid into a lung thus prepared until it reaches a point 
just short of rupture, but no fluid would flow from the divided branch. 
Again, if we inject a main bronchus of a lobe of a lung, prepared as 
above, with celloidin, and digest it in normal pepsin solution, it will 
be found that no celloidin has passed to that part of the lung sup- 
plied by the divided branch. If now we prepare a third lung as be- 
fore, and instead of leaving the divided branch uninjected, we inject 
it with a different colored celloidin from the main branch and digest 
it in pepsin it will be found that the two celloidins have not flowed 
together, but are distinct and can easily be separated. In my recon- 
struction of the lung I find no communication between the air-cells of 
the different air-sacs of a lobule or the cells of one lobule with those 
of another. 
We have already seen how irregular the cluster of air-sacs from 
a terminal bronchus is, and how they dove-tail into one another, thus 
making the breaking apart of a complete wax injection quite impossible. 
It is probably from this cause that ROOSEVELT got „corresponding 
rough surfaces“, or it may have been that he separated the air-sacs 
from an atrium, the ,,rough surfaces‘ being the broken sac passages, 
or he may have separated the atria, thus getting two or more broken 
surfaces. 
1) Medical Record, New York, Vol. 37, No. 8. — New York Medical 
Journal, March 7, 1891. 
