METHODS OF MEASUREMENT. 27 



Tests for tightness. The care used in making all of the various joints 

 in the system air-tight was well worth the effort, for the apparatus was 

 surprisingly free from leaks. At the beginning of each day's experi- 

 menting a preliminary test for tightness was always made, and if the 

 pointer on the kymograph-drum showed any variations within 3 min- 

 utes, a leak was sought for. During the research further tests of 

 15 and 20 minutes were likewise made for possible leaks. Early in 

 the experiment these longer tests were made daily, but later they were 

 made but once a week or even once in two weeks. 



Tests of air in the system. Before beginning the experiment of the 

 day it was the practice to empty the bell of the spirometer and intro- 

 duce 4 or 5 liters of oxygen into the system. The ventilating appa- 

 ratus was then operated for a few minutes, the spirometer again 

 emptied, and refilled with oxygen. This prevented any danger of 

 deficiency in the oxygen-content of the air and an accumulation of 

 nitrogen. Duplicate analyses of the air after the completion of a series 

 of seven successive experimental periods with W. K. on one day gave 

 results for oxygen of 21.90 and 21.93 per cent. 



TREADMILL. 



A detailed description of the treadmill was given in the report of the 

 previous study in which this apparatus was used. 1 It consists of an 

 endless leather belt which travels over two broad wooden pulleys, sup- 

 ported on ball bearings, at the ends of a wooden frame. The mill is 

 actuated by a ]/?. h. p. electric motor. The belt between the pulleys 

 is supported by a considerable number of steel tubes, set close together 

 but without touching in a steel framework. Each tube is fitted at its 

 two ends with annular steel ball bearings. A rolling, frictionless sur- 

 face is thus provided for the man to walk upon, which is both sub- 

 stantial and smooth. The speed of the motor may be varied at will. 

 Some slight alterations were made in the treadmill for this research, 

 but it was essentially as used in the previous study with the subject 

 walking. The general arrangement of the apparatus is shown in the 

 frontispiece and in figure 1, page 19. 



Distance counters. The use of the two counters recording the num- 

 ber of revolutions of the front pulley of the treadmill was continued in 

 this research, except that the "continuous counter," for recording the 

 total number of revolutions, was moved to the rear of the treadmill 

 frame and operated by a wire from the front pulley. This change was 

 made so that the counter might be in view of the operator standing at 

 the absorber table. The other counter, which is shown in figure 5, 

 and records the number of revolutions of the pulley during the experi- 

 mental period, is known as the "period counter." It is fastened to 

 the front pulley of the treadmill and is operated by the bar J, which 



'Benedict and Murschhauser, Carnegie Inst. Wash. Pub. No. 231, 1915, p. 34. 



