RESPIRA TION 629 



epithelium is not at all what would be expected in a secreting organ. The cells 

 are quite thin, and very unlike those of the gas gland of the fish, where, as we have 

 seen (page 361), there is an obvious reason why oxygen should be actively secreted. 

 A gas must be produced and absorbed in adaptation to the pressure at different 

 depths. It has been said indeed that, in the bird, where the need of extra supply 

 of oxygen would be supposed to be greater than in the mammal, the pulmonary 

 alveoli are devoid of epithelial lining altogether. If the exchange were by diffusion 

 alone, the direct contact of the walls of the blood capillaries with the alveolar air 

 would be of advantage. Again, unlike a secreting gland, oxygen passes with equal 

 facility in either direction. Breathing fire damp, for example, causes instantaneous 

 unconsciousness through loss of oxygen from the blood to the gas in the alveoli. 



Hartridge (1912, I) introduced an improvement in the carbon monoxide 

 method of Douglas and Haldane, by substituting observation of the change of 

 position of the absorption bands, which is produced by carbon monoxide, instead 

 of the mere visual comparison of the colour of two solutions. Using 'his new 

 method, Hartridge (1912, 2) investigated the effects of producing oxygen want in 

 the tissues in three ways, by breathing mixtures containing carbon monoxide, by 

 lowering the oxygen tension of the air breathed, and by doing work. He was 

 unable to find any evidence of crxygen secretion by the lungs in any case, but, as 

 was stated above, Douglas and Haldane now hold that it is not to be detected 

 until acclimatisation has been developed. 



Bohr introduced the consideration of the rate at which oxygen could pass through the 

 pulmonary epithelium and capillary wall, and calculated, entirely from theoretical data, what 

 he called "invasion" and "evasion" coefficients. The conclusion to which he came was 

 that the diS'erence between the tension of oxygen in the arterial blood and that in the 

 alveolar air could only be accounted for by secretion on the part of the cells. Krogh, 

 however (1910, 1), made direct experiments on the rate at which oxygen passed from water 

 into a gas bubble, and found that the "invasion coefficient" is really nearly seven times 

 that calculated by Bohr. It seems possible that the solubility of oxygen in water, which 

 enters into the formula, is altered at the contact surface between the epithelium and the 

 alveolar air, owing to the action of surface forces, a fact neglected by Bohr. We saw above 

 that the solubility of gases depends on the surface tension of the liquid solvent (page 54), and 

 that a low surface tension increases the solubility (Christov). It is quite possible that the 

 surface tension of the liquid covering the membrane of the lung alveoli may have a very 

 low surface tension, owing to presence of lipoid. If this were so, the solubility of oxygen 

 in it might be much greater than that reckoned by Bohr. From the invasion coefficient 

 it can be calculated how much oxygen can pass into the blood in a given time, and, although 

 it appears that it is sufficient to satisfy the conditions of rest on the diffusion theory, it is 

 held by Barcroft (1914, p. 216) that diffusion will not account for the large amount of oxygen 

 used in exercise, or under the conditions of low oxygen tension as in rarefied air. It is to 

 be remembered that the calculation requires knowledge of the quantity of blood passing 

 through the lungs. Krogh and Lindhard (1912) determined this experimentally in man, 

 and found that, in muscular work, it might rise to as much as 21 '6 litres per minute, instead 

 of the much smaller number taken by Bohr (1909) as the basis of his calculation. On p. 228 

 we find the following calculation. In a particular experiment it was found that 162 c.c. 

 of oxygen per litre of blood passing through the lungs was taken up and utilised ; that is, 

 85 per cent, of the difference between arterial and venous blood. In muscular work, 2,700 c.c. 

 of oxygen were consumed per minute. Hence, if we take 21 litres per minute as the cardiac 

 output, according to the measurements of Krogh, we find that 162 x 21 =3,400 c.c. of oxygen 

 per minute can be taken up by the lungs more than enough to satisfy requirements. 

 Similarly, the work of Patterson and Starling (1914) shows that the amount of blood sent 

 out by the heart, when working under optimal conditions, is very much larger than previously 

 assumed. Taking the data available, it can be shown that the amount of oxygen which the 

 blood can carry from the .alveolar air by diffusion is considerably in excess of that found to 

 be consumed under any muscular work hitherto determined. Marie Krogh (1915) has made 

 further experiments and finds that diffusion is quite capable of explaining the maximum 

 amount of oxygen consumed in muscular work. 



At the same time, it must be admitted that we have no explanation for the 

 results of Douglas and Haldane on Pike's Peak. It seems very desirable that 

 the experiments should be repeated on lower animals by the aerotonometer 

 method of Krogh. The difficulty is that the animals must be kept for some 

 time under reduced oxygen pressure and the experiments made under anaesthesia. 

 As regards the latter factor, the adherents of the secretion theory may make 

 the objection that the narcosis paralyses the secretory power of the cells ; but 

 it has no such effect on other glands. There is one fact in the data given by 



