252 RESPIRATION 



secretion during work. His experiments did not appear to show 

 any active secretion, and it would have been extraordinary if 

 they had. 



I now come to the main point of Barcroft's criticisms. Hartridge 

 had at first calibrated his instrument by ascertaining its readings 

 with what he believed to be known mixtures of oxyhaemoglobin 

 and CO-haemoglobin. He subsequently found that his calibrations 

 had been quite incorrect; and in order to secure correct calibration 

 he finally had recourse to the very tedious method of pumping 

 out the CO and oxygen from the blood mixture after adding 

 ferricyanide, and determining the CO and oxygen by analysis, 

 using the general method which I followed in originally testing 

 the accuracy of the ferricyanide method for blood gases. In his 

 paper 38 Hartridge says of his first method that "experiments 

 made since to discover the cause of the error have shown that 

 with the method of mixture employed complex interactions take 

 place between the two portions of solvent." Let us expand this 

 somewhat mystic statement. He was working with blood diluted 

 with water to about a twentieth. One portion of this he saturated 

 with CO, and another portion with air. These were then mixed. It 

 was apparently expected that the result would be a mixture con- 

 taining half the haemoglobin saturated with CO and the other 

 half with oxygen. Now if one dilutes blood to a twentieth and 

 saturates with CO, the solution will contain about one volume of 

 CO in combination with haemoglobin to two and one-half in 

 simple solution ; and when this is mixed with an equal proportion 

 of the solution saturated with air the CO in simple solution in the 

 first part will straightway combine with the haemoglobin in the 

 second part, and turn out the oxygen, the result being that prac- 

 tically the whole of the haemoglobin combines with CO. With 

 the method first adopted by Hartridge it was clearly impossible 

 for him to calibrate his instrument. 



Our colorimetric method of determining the saturation of 

 haemoglobin with CO had repeatedly been tested against mix- 

 tures previously prepared, the most scrupulous precautions (de- 

 scribed in three different papers) being, however, taken to avoid 

 errors arising from the solubility of CO. Barcroft, however, infers 

 that because Hartridge's calibration failed with the method of 

 mixtures, ours was presumably also inaccurate : whereas Hart- 

 ridge's final calibrations were made with the blood pump, which 



38 Hartridge, Journ. of Physiol., XLIV, p. 9, 1912. 



