CRITICAL DISCUSSION OF RESPIRATION APPARATUS. 265 



It is recommended that so far as possible all respiration apparatus 

 be controlled by means of some method in which a known quantity 

 of the gases is measured. For instance, candles, alcohol, ether, or other 

 combustible materials may be burned, and, since their composition is 

 definitely known, the oxidation products and oxygen requirement may 

 be definitely measured and compared with the actual determinations 

 made with the apparatus. It must be pointed out that such control 

 tests only prove that the apparatus is theoretically accurate, but does 

 not necessarily prove that all experiments made upon men with this 

 apparatus will give accurate results. Too frequently an apparatus 

 which has been proved to be theoretically correct has been used by 

 investigators in a way in which it was not intended to be used or the 

 experiments were not carried out under proper conditions or were 

 not sufficiently controlled. Far-reaching conclusions and theoretical 

 deductions have then been drawn from a very few experiments. The 

 determination of the respiratory exchange of man in short periods and 

 particularly of the respiratory quotient is a very difficult problem. 

 Conservatism in the acceptance and interpretation of results is therefore 

 strongly recommended because of the great number of variable factors 

 involved in any respiration experiment and because of the great necessity 

 of repeated observations before one can be absolutely certain of the 

 results obtained. 



I desire to express my thanks to Miss A. N. Darling for much assis- 

 tance in the preparation and editing of the manuscript and to Professor 

 Francis G. Benedict for advice and helpful criticism throughout this 

 investigation. 



Nutrition Laboratory of the 



Carnegie Institution of Washington, 

 Boston, Massachusetts, March 17, 1915. 



€> 



