CONTEMPORARY ADVANCES IN PHYSICS 



161 



and his school. F'ig. 4 embodies the data of three, in each of vvhicli 

 an air-chamber was lowered into a high-lying snow-fed '^ lake (Millikan's 

 reasons for choosing such I have already stated). Crosses stand for 



Q 7 



6 7 8 9 10 II 12 13 14 15 16 17 18 19 



DISTANCE BELOW SURFACE OF ATMOSPHERE IN METERS OF WATER 



Fig. 4 — Ionization ascribed to cosmic rays as function of depth beneath top of 

 atmosphere (early curve of Millikan and Cameron). 



data obtained in Arrowhead Lake, California (altitude 1550 metres) ; 

 circles correspond to Muir Lake, California (3580 metres); the dots to 

 which arrows point are the reward of a long hard journey, for they 

 show data from a lake — Miguilla — no less than 4570 metres above the 

 sea, in the Bolivian Andes. All of the measurements here plotted were 

 taken at depths of a metre or more below the water surface, but none 

 so great as Regener's — five metres was the maximum depth at Miguilla, 

 fifteen at Arrowhead and twenty at Muir.'' 



The points of the three sets lie close to a single smooth curve. Each 

 of the three, however, has been shifted horizontally, to make allowance 

 for the different depths of air overlying the three lakes: it has been 

 assumed that the air over Miguilla is the equivalent of 5.95 metres of 

 water, over Muir to 6.75 and over Arrowhead to 8.6 metres. To 

 figure out the equivalents it is necessary to know the distribution-in- 

 height of the air (Millikan got it from the Smithsonian Institution 



■• Regener of course admits that Lake Constance is not of this character, but by 

 later experiments (briefly mentioned in his note in the Physikalische Zeitschrift) he 

 found that radioactivity of its water was not distorting his results. 



' Data were obtained at lesser depths than a metre, but the values of ionization 

 are higher than would be inferred by prolonging the curve which is valid for greater 

 depths. This recalls a feature of Regener's data, but here the excess is ascribed by 

 Millikan not to soft cosmic rays, but to the radio-activity of the overlying air and the 

 nearby mountains. Of the points plotted in Fig. 4, the "two which fall farthest 

 from the curve correspond to single readings, and hence should be given little weight 

 in comparison with points which represent the means of three or four readings" 

 (Millikan & Cameron), 



