2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 95 



EXPERIMENTAL 

 Since the method of measuring carbon dioxide used in the present 

 experiments is new, it will first be very briefly described. The method 

 is based upon the experimental fact that carbon dioxide gas is a very 

 powerful absorber of a certain band of infrared radiation. Two 

 millimeters thickness of pure carbon dioxide at N. T. P. absorb 78 per- 

 cent of the radiation in a band from 4.2 to 4.3 fx. Thus in a long opti- 

 cal path, small concentrations of carbon dioxide will cause marked 

 absorption of this radiation. No other gas or vapor, including water 



2000 4000 6000 (000 10000 



Fig. I. — Carbon dioxide assimilation curves. (After Hoover, Johnston, and 

 Brackett.) Ordinates, carbon dioxide assimilated. Multiply by 0.25 to obtain 

 cubic millimeters per minute. Abscissae, carbon dioxide concentration. Multiply 

 by 41 X 10"" to obtain volume percent. Parameters, light intensities. Multiply 

 by 3.56 x ID"* to obtain watts/cnr, or by 4.96 to obtain foot-candles. 



vapor, common to air absorbs this band of radiation. Hence the 

 method is specific for carbon dioxide alone, and is not affected by 

 humidity changes. Because of its nature, the method has almost 

 negligible temperature and pressure corrections. It is as sensitive as 

 the best chemical methods since at small concentrations it will detect 

 one part in a million (by volume) of carbon dioxide. It is practically 

 instantaneous, the response of the galvanometer-thermocouple system 

 (about 5 seconds) determining this. Because of this latter fact, for 

 most purposes it is many times more sensitive than any chemical 

 method, since it will detect this one-millionth part of carbon dioxide 

 in 5 seconds. 



