102 COMPARISONS OF RESPIRATORY EXCHANGE. 



used were extremely large for the amounts of carbon dioxide to be 

 absorbed. The water-absorber following the carbon-dioxide absorbers 

 weighs 18 kilograms, and it is possible to weigh it to 0.1 gm. ; an error 

 of 0.1 gm. in weighing would be equivalent to about 0.5 to 1 per cent for 

 a period in which the carbon-dioxide elimination was 25 gm. When the 

 results of the whole experiment are considered, the percentage error be- 

 comes very much smaller, as the experiment is usually two hours in length. 



Errors in measuring the residual carbon dioxide. In the former 

 method of determining the residual carbon dioxide, it was possible 

 to have a larger error in the measurement of this factor than with the 

 later method, i. e., that by which the sample of air is taken from the 

 outgoing air-current and analysis made by means of the Sonden-Pet- 

 tersson gas-analysis apparatus. With the later method it is possible 

 to determine the percentage of carbon dioxide in the residual air to 

 within 0.002 per cent. With a volume of 950 liters, which is approxi- 

 mately the volume of the bed calorimeter, this would be equal to about 

 0.02 liter or 0.04 gm. This is an error far too small for consideration 

 in this connection. That the actual composition of the outgoing air 

 undergoes mathematically the same fluctuations in percentage com- 

 position as does the entire air of the chamber has been demonstrated 

 by analyzing samples taken at different points in the chamber, the 

 results showing no variations due to the difference in location. 



Error in determining the water-vapor in the chamber. The method 

 formerly used in determining the amount of moisture in the chamber 

 was that of diverting a small stream of air from the outgoing air-current 

 through a set of U -tubes, one of which contained sulphuric acid and 

 pumice stone. Twenty liters of air were passed through these U -tubes 

 and the increase in weight of the tubes was noted; from this increase in 

 weight and the volume of the sample, the amount of moisture inside of 

 the chamber was calculated. The errors in this method have to do 

 mainly with the errors in weighing. The U-tubes weighed about 60 

 gm., and the possible error was 1 to 5 mg. The amount of mois- 

 ture absorbed in the 20-liter sample, although varying, was usually 

 about 100 mg. It will be seen, therefore, that the error might be 5 

 per cent. This error chiefly affects the determination of the oxygen 

 consumption, as the calculation of the amount of oxygen present in the 

 apparatus at the end of an experimental period necessitates the deduc- 

 tion of the volume occupied by the water-vapor in the chamber air; 

 consequently any error in determining the latter volume results in a 

 similar error in the calculated volume of oxygen. 



The more recent method of determining the residual water- vapor, 

 that is, the use of the wet- and dry-bulb thermometer inside of the 

 chamber, gives determinations with a high degree of accuracy. The 

 determinations obtained by this method have been carefully checked 

 by duplicate determinations made with the Sonden hygrometer and 

 the results agreed very closely. 



