RESPIRATION 415 



tions of barometric pressure do not sensibly affect the correction, but at 

 high altitudes the correction must evidently be diminished in the pro- 

 portion of about 0.1 cc. for every 100 mm. of diminution in atmospheric 

 pressure. 



If the blood is taken, not from the living body, but from a saturating 

 vessel, the gases dissolved physically must be calculated on the same 

 principle, allowing for their pressures in the vessel. ^^ 



The method just described has been tested for accuracy in several 

 ways. In the first place it has been found that when blood fully saturated 

 with air at room temperature is placed under the sodium carbonate 

 solution in the ordinary way and then agitated after the gauges have 

 become steady, there is no sensible variation in the reading of the burette 

 afterwards. The constancy of the reading can be relied on to .002 cc. 

 with careful work. Hence the percentage saturation can be relied on to 

 0.5 per cent, or the oxygen capacity per 100 cc. of blood to o.i cc, if the 

 measuring pipettes are properly calibrated. This is as good a result as 

 could be obtained with 20 cc. of blood by means of the original ferri- 

 cyanide apparatus. The present method is therefore as exact as the 

 original one for determining the oxygen capacity of blood, but is quicker 

 and more convenient. By using sodium carbonate instead of ammonia 

 solution the errors due to diminution of the vapor pressures of ammonia 

 and water on mixing the blood with the solution are eliminated, while 

 the use of saponin, first introduced by C. G. Douglas, produces the laking 

 of the blood which is necessary in order to allow the ferricyanide to 

 act on the oxyhaemoglobin. The fact that, as has been found by Meakins 

 and Davies, haemoglobinometer estimations coincide within i per cent 

 with the results by this method furnishes further confirmatory evidence. 



The new apparatus gives sharper results than the constant volume 

 method which Barcroft and I described in 1902. This is, I think, partly 

 due to the larger volume (2 cc.) of blood employed; partly to the fact 

 that the disturbance due to the use of ammonia solution is avoided and a 

 sharper index of temperature equilibrium is given by the two gauges 

 of the present apparatus ; and partly because the gauge levels are always 

 at the same place, whereas in the constant-volume apparatus the gauge 

 levels shift to places wide apart, so that small errors due to varying 

 capillarity of the gauge tubes are apt to tell. It is thus difficult, with the 

 constant-volume apparatus, to avoid errors within 2 per cent on either 

 side of the actual percentage saturation. 



^^ In the paper by Barcroft and myself where we first described the constant 

 volume blood-gas apparatus, the correction for gas in simple solution was un- 

 fortunately given incorrectly; and this doubtless accounts for the somewhat 

 distorted forms of the dissociation curves of oxyhaemoglobin in Barcroft's earlier 

 experiments on this subject. 



