50 THE ROYAL SOCIETY OF CANADA 



blood pressure was recorded by the usual methods from one of the 

 femoral arteries and a cannula was inserted in the opposite artery for 

 the purpose of collecting samples of blood on which to measure the 

 degree of oxygen unsaturation. This was done by using Barcroft's 

 differential blood gas manometer as follows: a narrow straight pipette, 

 graduated in 0. 1 ces., was connected with the artery and 2 quantities 

 of 0.2 c.c. each of the blood were delivered immediately, and without 

 coming in contact with air, under about 0.3 c.c. of a weak solution 

 of sodium carbonate, and saponin, contained in the two small " bottles" 

 of the apparatus. One of the bottles was rotated so as to lake the 

 blood and render it completely saturated with oxygen and both were 

 then attached to the manometer tubes and placed in a water bath at 

 room temperature until there was no further shrinkage of the fluid 

 in the manometer. On now shaking the manometer the blood in the 

 bottle that had not previously been shaken became laked and ab- 

 sorbed oxygen from the air of the bottle so that a slight negative 

 pressure resulted, the extent of which was indicated by the manometer. 

 Finally a solution of potassium ferricyanide was mixed with the laked 

 saturated blood on one side and the positive pressure created by the 

 oxygen thereby evolved also read on the manometer. From these 

 two values the percentage of unsaturation was calculated. 



Samples of alveolar air were also collected by the method de- 

 scribed elsewhere, and analysed in the Haldane apparatus. In an 

 actual experiment the procedure was to measure the volume of air 

 breathed, to analyse samples of alveolar air during several short 

 periods of time while the animal was still breathing outside air, and 

 when the results were found to be tolerably constant to take 

 samples of blood. The inspiration valve was then connected with 

 the rubber bag and the volume of breathing measured and the alveolar 

 air analysed at regular intervals of 10 minutes, duplicate samples of 

 blood being also taken usually at the same periods, although some- 

 times it was necessary to omit one or more of these. 



Results 



Although observations have been made over a wide range of 

 oxygen percentages the most important for our present purpose are 

 those in which alveolar air contained between 8.3 and 10.8 per cent, 

 of oxygen. From these experiments the following results are of 

 interest : 



1. During the first ten minutes after causing the animal to 

 breathe oxygen-poor air, the respiratory volume becomes increased 



