METHODS OF MEASUREMENT. 21 



in the volume of air in the closed circuit. 1 In the present research, 

 a large spirometer (G in figure 1) was employed, 8 liters in capacity, 

 which was built on the principles already described in reports from this 

 Laboratory. 2 This spirometer was placed on a small table at the left 

 of the treadmill and close to the subject. A light, non- viscous oil was 

 used in the spirometer rather than water. None of the subjects com- 

 plained of odor from the oil; in fact, none of them knew that oil was 

 used. 



During the experiments, when severe muscular work was being 

 performed, the respirations of the subject became so deep that the 

 movements of the spirometer-bell were too large to be recorded on the 

 usual kymograph-drum. To reduce the movement of the pointer, 

 the thread supporting the bell was passed through a pulley to which 

 the counterpoise and the pointer were attached. (See g, fig. 1.) The 

 movement of the pointer was thus reduced one-half, but this had its 

 disadvantage in that it doubled any error in the reading, since all 

 readings must be multiplied by 2. As in the walking experiments, 

 the oxygen consumption was frequently 8 to 10 times the resting value, 

 this doubling of the error played no role save in the "rest" experiments. 



In measuring the rate of oxygen consumption in certain tests in this 

 research, no oxygen was admitted for a portion of the experimental 

 period and the lower capacity limit of the spirometer was thus reached 

 in a few minutes. For these few tests, use was made of the double 

 spirometer shown in figure 2. A duplicate spirometer A is attached to 

 the principal one B by a large tube E and a 3- way valve D. The 

 oxygen is introduced by means of the connection C. Both spirometers 

 were filled with pure oxygen before an experiment, and the usual 

 readings taken on the main spirometer B. As soon as the subject was 

 connected with the ventilating system, the bell of the spirometer B 

 fell rapidly with each respiration as the oxygen was consumed. When 

 the oxygen was at as low a level as seemed wise, the three-way valve D 

 was opened and oxygen from the duplicate spirometer A was forced 

 into B by pushing down the bell of A. The valve was then closed 

 and A was again filled from the oxygen cylinder through the connec- 

 tion C. This was repeated as often as necessary. The kymograph 

 curve thus shows a succession of hills and valleys (see fig. 15, p. 183), 

 due to the fact that the pointer rose on the scale of B as the oxygen 

 was consumed and then sank when the supply from the reservoir A 

 was forced into the main spirometer B. The time of filling spirometer 

 B was scarcely 2 seconds, and not over one or two respiration tracings 

 were lost in the process. 



Benedict, Am. Journ. Physiol., 1909, 24, p. 345. See, also, Carpenter, Carnegie Inst. Wash. 

 Pub. No. 216, 1915, p. 24. 



"Benedict, Deutsch. Archiv f. klin. Med., 1912, 107, p. 172; Carpenter, Carnegie Inst. Wash. 

 Pub. No. 216, 1915, p. 37. 



