d48 CHIKANOSUKE OGAWA 
to 1 per cent silver nitrate was unsuccessful. Such reagents as 
ammonia silver, protein silver, protargol, gelatin silver, and 
osmium silver were used. Only the osmium silver showed some 
intercellular cement lines along the blood-capillaries of respira- 
tory canaliculi. I found that solutions of silver nitrate above 
3 per cent also gave results of this kind. If the respiratory 
epithelium were impregnated by this, the cement lines, which 
are seen on the capillaries, must be continuous with those of the 
epithelial cells of the flute-holes; but, on the contrary, I found 
that the cement lines do not pass over to the epithelium of the 
flute-holes, but to the endothelium of the capillaries beneath 
the epithelium of the flute-holes. Thus it was brought out that 
even by injection of silver solution into the trachea itself the 
-endothelium of the capillaries are impregnated. In order to 
make my results more certain, I injected the silver solution in 
the blood-vessel system and found that images obtained were 
just like those resulting from injection into the trachea. 
Only the embryological evidence remains to be discussed in 
relation to this problem. Juillet claims to have succeeded in 
impregnating the respiratory epithelium of advanced chick 
embryos with silver solutions. Hence I also studied the embryos 
of the chicken, duck, and goose at several stages of incubation 
by means of silver impregnation as well as ordinary staining, 
and was convinced that, though the flute-holes are lined by one 
layer of cuboidal or flat cells, no epithelium is visible in the respira- 
tory canaliculi themselves; at least, impregnation of the respira- 
tory epithelium never takes place. I am inclined to think that 
Juillet mistook the epithelium of flute-holes for respiratory 
epithelium. 
With this, the presence of respiratory epithelium in birds 
became exceedingly doubtful. The fact that no analogous case 
has been found in which the epithelial coating of a free surface 
is lacking has led Oppel and other authorities to conclude that 
the lung of the bird must have respiratory epithelium in spite 
of the absence of actual confirmation. If it were only the usual 
silver method which failed here, such an inference might be 
warranted, but it is difficult to comprehend why no respiratory 
