48 GASEOUS METABOLISM OF INFANTS. 



of this identical apparatus and the proof of its accuracy for measuring 

 the results of half-hour periods after an amount of carbon dioxide 

 had been developed approximately equal to that produced by an infant. 

 From the experience with the large respiration chambers in the Nutri- 

 tion Laboratory, it became increasingly evident that the discrepancies 

 shown in the alcohol check tests for short periods were to be ascribed 

 not to errors in the absorption of carbon dioxide or to the calculation 

 of the amount of oxygen produced, but to discrepancies in the meas- 

 urement of the small quantities of alcohol necessary for the control 

 test. The alcohol required for producing (when burned inside the 

 respiration chamber) an amount of carbon dioxide equivalent to that 

 given off by a small infant corresponds to about 1.00 or 1.50 grams per 

 half hour. To measure this with an accuracy of 1 per cent and to 

 insure that the measurement represents not only the amount of alcohol 

 introduced but the actual amount consumed, involves much experi- 

 mental work. Mr. T. M. Carpenter, of the Laboratory staff, has 

 recently conducted experiments in which this second difficulty has been 

 overcome. 



In these experiments a small piece of capillary copper tubing was 

 carried through the walls of the respiration chamber by means of the 

 tube commonly used for the stethoscope, and then bent upward to form 

 a minute lamp. The exterior end of the copper tube was connected 

 by capillary rubber tubing to a glass burette of very fine caliber which 

 could easily be read to 0.01 c.c. The burette and the capillary rubber 

 tubing were suspended by a cord running over a pulley, so that both 

 the burette and the rubber tubing hung free in the air. When this 

 burette was very slowly and gradually raised, alcohol flowed with great 

 regularity through the copper tube into the chamber, where it burned 

 quietly. By reading the level of the alcohol in the burette at the 

 beginning and the end of any given experimental period, the absolute 

 amount of alcohol introduced could be accurately determined. A 

 small wooden pulley attached to the vertical upright of a Porter kymo- 

 graph was used for raising the burette regularly. An extremely even 

 elevation of the burette could be secured by adjusting the speed of 

 the rotating fan so that the amount of alcohol introduced per half hour 

 ranged not far from 1.00 to 1.25 grams. The whole apparatus is 

 shown in figure 4, in which may be seen the copper tube extending 

 through the two walls of the respiration chamber, the flexible rubber 

 tubing, one end of which is attached to the copper tube and the other 

 to the burette, and the wooden pulley and kymograph in position to 

 raise the burette as desired. 



When beginning the experiment, the lamp was lighted, the kymo- 

 graph set in motion, and after the lamp had burned a few moments 

 and regularity of combustion was assured, the cover was put in place. 

 At the end of a preliminary period of not far from 15 to 20 minutes, the 



