REGULATION OF ENVIRONMENT _ 63 
their specific gravity equal to that of the water at 
whatever depth they may be, or even to counterbal- 
ance the effects of a float or weight attached to them. 
I have in my library Ludwig’s copies of Moreau’s 
papers. They are an interesting clue to what was 
passing through his mind in suggestions he made as 
to the possibility of oxygen secretion in the lungs. It 
was discovered by Bohr that the oxygen secretion in 
the swim bladder is, like salivary secretion, under 
nervous control; and Dreser found that oxygen secre- 
tion can be excited by pilocarpin, a drug which also 
excites secretion in other glands. 
The cells in the wall of the swim bladder which 
secrete the oxygen are columnar, and arranged like 
the cells of many other secreting glands, whereas the 
lung epithelium is extremely thin. Nevertheless the 
elementary structure of the lung is glandular, just as 
in the case of the swim bladder; and both lung and 
swim bladder are developed as outgrowths of about 
the same part of the alimentary canal. Before the 
lungs expand at birth the lung epithelial cells are cubi- 
cal, and similar to those of other secreting glands. 
That the secreting cells should be thicker in the swim 
bladder is natural considering the enormously greater 
pressure against which the cells have to secrete. 
The pressure difference against which oxygen can 
be secreted in the lungs is evidently quite limited. This 
is shown by measurements of the oxygen pressures in 
the blood in CO poisoning, when the stimulus to secre- 
tion is pushed up to what is presumably a maximum. 
If there were no limit the secretion of oxygen would 
