Aggregation theory 63 



Whether or no the curves drawn from these data fit the points as 

 determined may best be seen by actually drawing the curves, and 

 placing the points in relation to them. 



This is done in Figs. 31 and 32. 



To pass to a more complicated case that of a solution of haemo- 

 globin in a mixture of salts, namely Ringer's fluid: taking n = ^'l 11 

 and K ='00427, a curve is produced which satisfies the properties 

 of haemoglobin in Ringer's solution, as well as does the freehand 

 curve of Barcroft and Camis (9) . 



From Ringer's solution we may pass to the still more complicated 

 problem which is presented by blood. 



Here we have the advantage, in some cases at all events, of work- 

 ing with curves upon which much more time and labour have been 

 spent than have been claimed by the solutions of haemoglobin. 



Some forty or more determinations have been made on the normal 

 blood of Mr C. G. Douglas < 10) of St John's College, Oxford, some by 

 myself, some by Haldane and Douglas ; and as our methods differed 

 somewhat in detail, the fact that we arrived at the same result proves 

 that Douglas' normal curve has been determined with as much 

 accuracy as our present methods will admit of. Yet in spite of 

 the great number of determinations which have been made, the 

 curve as drawn freehand through the points is extremely close to 

 that calculated from the equation, with the values ^='000212, 

 n=2'5. This curve and the corresponding points are given in 

 Fig. 33. The curve differs very much from those of the solution 

 given just previously, inasmuch as the S-shape is much more evident. 

 This difference is due to the fact that the curve is determined in 

 the presence of 40 mm. C0 2 pressure, which was the pressure of that 

 gas in Douglas' alveolar air and presumably in his body generally. 



Each test which we apply, and which is met satisfactorily, leads us 

 on to another. We were drawn into the discussion which has formed 

 the subjects of this present chapter and the last by the discovery that 

 the bloods of different animals had not identical affinities for oxygen. 

 The theory we are testing must therefore not only hold true for the 

 blood of a particular person, but it must also hold good for the bloods 

 of different species. 



It has been tested in the case of man, sheep, dog and cat, and 

 seems to be true for them all. 



But the differences which have been found between different 

 species apply to a less extent to different individuals, at all events to 



