102 A RESPIRATION CALORIMETER. 



the volume of alcohol delivered. This is due to the fact that this bu- 

 rette is one of the Geissler type, in which a ground glass rod serves as 

 a stopcock. Obviously, therefore, when this glass rod is removed the 

 actual amount of liquid delivered between the marks on the burette is 

 larger than when the rod is in place. This difference happens to com- 

 pensate almost exactly for the lower specific gravity of the alcohol 

 used, and consequently the number of cubic centimeters read on this 

 burette corresponds to the same number of grams of alcohol. 



Thus the calculation of the amount of alcohol admitted to the lamp 

 is based upon the loss in weight of the alcohol bottle and tubes and 

 the variations of alcohol level in the burette. The lamp here described 

 burns alcohol at the rate of about 20 grams per hour. 



The electric light placed just outside the window is so situated that 

 the rays of light fall upon a mirror which is inclined in such a position 

 as to illuminate brilliantly the gage tube, thus materially aiding in the 

 proper adjustment of the alcohol level at the end of the period. 



FREQUENCY AND DURATION OF EXPERIMENTS. 



In the earlier years of experimenting it was deemed advisable to con- 

 duct an alcohol check test immediately before each experiment with 

 man. With increasing skill in manipulation the necessity for these 

 frequent tests has in a large measure disappeared, and at present three 

 or four tests in a year are all that are required to control the apparatus. 



The experiments last from 8 to 36 or more hours. Recently experi- 

 ments of about 24 hours have been most common. The experiment 

 here reported was subdivided into three periods of 3 hours 54 minutes, 

 5 hours 44 ^ minutes, and 1 1 hours 52 minutes, respectively, the whole 

 experiment lasting 2 1 hours 30^ minutes. The total amount of alcohol 

 burned was 406.8 grams, apportioned among the periods as follows : 

 First period, 73.4 grams; second, 108.1 grams, and third, 225. 3 grams. 



CALCULATION OP THE ALCOHOL CHECK EXPERIMENTS. 



From the weight of the alcohol burned and the known factors cor- 

 responding to the theoretical amounts of carbon dioxide, water, and 

 oxygen per gram of alcohol, the theoretical quantities that should be 

 found by means of the respiration apparatus may be readily computed. 

 Inasmuch as the quantities of carbon dioxide, water, and oxygen, as 

 found by gain in weight of the absorbing system and loss in weight of 

 the oxygen cylinder must be corrected for the variations in the residual 

 amounts of these gases present in the system at the end of each period, 

 it is customary in tabulating the results of these determinations to in- 

 clude in the tables the data for the residual amounts. 



