50 



COMPARISONS OF RESPIRATORY EXCHANGE. 



care is of course taken to insure that all of the observations are made 

 as uniformly as possible. 



It is of interest to note the average accuracy of meters in actual 

 experimenting. An opportunity was given for observing this in con- 

 nection with a study on the effect of a carbohydrate-free diet upon four 

 young men during the winter season of 1912-13. Both meters and 

 oxygen cylinders were used in these experiments. The type of oxygen 

 cylinder and valve employed will be subsequently described. The 

 cylinders were weighed to approximately 0.01 gm. on the balance 

 regularly used in connection with the respiration apparatus; the meters 

 were read as usual, and the barometer and temperature observed during 

 each period of admission. Each of the four meters was in charge of 

 each of four observers at various times, so that the series of results prob- 

 ably represents as nearly as can be the actual range of accuracy with 



TABLE 8. Results of Cathcart's experiments on the effect of the rate at which 

 oxygen is passed through the meter. 



these meters in use. Table 9 shows the correction factors obtained, 

 assuming that the loss in weight of the oxygen cylinder was accurately 

 measured and that there was no leak of oxygen during the experiments. 

 From an examination of the results it would appear that the range in 

 percentage accuracy is 2 per cent, that the average deviation for the 

 four series was from 0.37 to 0.75 per cent, and that the majority 

 of the figures are within this variation. Three of the observations with 

 meter No. 2 do not appear to have been made with sufficient care, 

 i. e., the first one on December 27 (100.4 per cent) and the first two on 

 December 28. On the latter date there was evidently a compensation 

 error which brought the first value well above the average and the other 

 considerably below. In general, however, the figures for meter No. 2 

 are reliable. Similarly, it is believed that the percentages for the 

 other three meters are representative of the accuracy with which one 

 can use the meter. 



Mention has been made of the various types of valves and connec- 

 tions which have been used with the oxygen cylinders. As has been 

 stated, the reduction valves supplied with the oxygen cylinders or 

 which were purchased in Europe were at times so inefficient that the 

 substitution of the meter proved of much advantage. Subsequently 

 it was found that a needle-valve, sold by the Charles E. Beseler Co., of 



