152 CARBON DIOXIDE IN EXPIRED ALVEOLAR AIR, 



aualysis. These capillary tubes were placed at intervals of 25 cm. 

 along the first two metres of the tube. Sampling was performed 

 by attaching burettes in which a vacuum of 50 to 100 c.c. was 

 produced by lowering the mercury reservoir, and clamping the 

 tube connecting the reservoir to the burette. The upper nozzles 

 of the burettes were connected by short pieces of pressure india- 

 rubber-tubing controlled with strong spring clamps. The 

 analyses were made in a Hempel's pipette with a sample of 50 c.c. 

 contained in a gas burette graduated in fifths of 1 c.c. The 

 readings were made to the nearest tenth. These readings were 

 easily possible as the divisions on the scale were 1 nun. apart. 

 As the change in volume during absorption varied from 1*5 c.c. 

 to 4 c.c, the analytical error might amount to 0*2 on the calcu- 

 lated percentage of carbon dioxide. A number of analyses were 

 done in duplicate, and the figures show the agreement between 

 the duplicates to be good. In no case did the difference in the 

 calculated percentage amount to more than 0*1. In the opinion 

 of Krogh(13), it is misleading to give the average alveolar 

 tension of carbon dioxide with an accuracy of more than 0*5 mm. 

 Hg., as the analytical figures should not be strained too far. 



Control of E.rper'mwntal J/efhod.— The air leaving the alveoli 

 of the lung traverses the tubular air-passages before reaching the 

 mouth. It is recognised that a certain amount of alveolar air 

 must be passed through these tubes before the air present in 

 them is dislodged. When the air issuing from the mouth travels 

 through a long tube, the same general conditions will prevail. A 

 considerable amount of expired air will be diluted with the air 

 in the tube before the whole of this is removed. It is now 

 known that, when a stream of gas is passed along a tube, an 

 axial stream traverses the centre of the tube, and that a ''spike" 

 of the entering gas is thrust into the air that is present. The 

 * 'spike" is more slender and elongated the more rapid the speed 

 of the entering gas. 



A series of experiments has been carried out to ascertain what 

 amount of gas was needed to wash out the tube, when passed 

 into the brass tube, in a small fraction of a second. Since a 



