Effect of salts 47 



then we were able time after time to obtain solutions which were 

 identical with that figured by Bohr for haemoglobin. 



To test the effect of salts in a more systematic way we made 

 solutions of haemoglobin, divided them into various portions, added 

 various salts to the portions and compared their dissociation curves 

 with that of the original haemoglobin as prepared by Bohr's method. 



In Fig. 24 are given the dissociation curves of oxyhaemoglobin 

 in the solutions of electrolytes which are of greatest physiological 

 importance, namely sodium chloride and potassium chloride isotonic 

 with one another. We shall reserve the consideration of some 

 others till later. From this point constructive work on the dis- 

 sociation curve of blood commenced. In the red blood corpuscle 

 haemoglobin is dissolved in a solution of salts which differs very 

 widely in different members of the animal kingdom. But it is 

 evident that the first step towards building up the curve of blood 

 from that of haemoglobin is to investigate a haemoglobin solution 

 dissolved in the actual salts which are present in the red blood 

 corpuscles. But here a fresh complication arose, for the salts differ 

 in the corpuscles of each different species. We performed the follow- 

 ing experiment : a solution of haemoglobin was divided into two 

 portions, (a) and (6). To the portion (a) were added the salts of the 

 human red corpuscles as determined by the old, though still classical, 

 analysis of Schmidt"" 1 . To the portion (6) were added the salts of 

 the dog's red blood corpuscles as determined by the recent, and ex- 

 cellent, analysis of Abderhalden' 7 '. The dissociation curves of these 

 two solutions were determined in the presence of approximately 

 40 mm. pressure of carbonic acid, and gave the results depicted in 

 Figs. 25 and 26. 



It was at least evident that the two curves were quite different, 

 that the difference could by no possibility be explained by experi- 

 mental errors, and that one had arrived at at least one tangible 

 reason why the blood of one animal should differ from that of 

 another ; but even then it seemed that we were far from a complete 

 solution of the whole problem before us. It was not unlikely that 

 any number of other reasons might conspire to add or detract from 

 the differences we had observed, and therefore our next move w r as to 

 make as faithful a comparison as possible between the curves which 

 we had just observed, and the actual curves which could be obtained 

 from human and canine bloods respectively when exposed to the 

 same conditions as the haemoglobin solutions (a) and (b). We hoped 



