206 ANNUAL REPORT SMITHSONIAN INSTITUTION, 19 3 5 



Compton's results are typical, and are particularly significant 

 because of the wide range of latitudes, elevations, and longitudes 

 covered. Higher intensities occur at high latitudes and the increase 

 with respect to the value at the Equator is 14 percent at sea level, 22 

 percent at an elevation of 2,200 meters, and 33 percent at 4,300 

 meters. 



9. THE SURVEY OF ANGULAR DISTRIBUTIONS 



Following the discovery by the writer and J. C. Street of an east- 

 west asymmetry of the coincidences of alined counters on Mount 

 Washington, N. H., in 1932, a survey for the study of this effect was 

 planned with the cooperation of the Carnegie Institution of Wash- 

 ington and was begun early in 1933. Up to the present time the 

 survey includes measurements at the station^ indicated on the map 

 of figure 4. Confirmatory results of significance have also been 

 reported by numerous other observers. 



The magnetic directional effect is manifest as an asymmetry, or a 

 difference of intensities from eastern and western azimuths at the 

 same zenith angle. The magnitude of the asymmetry is conveniently 

 expressed as the ratio of the intensity difference to the average in- 

 tensity for the two directions. In this form it is equal to the in- 

 tensity of the unbalanced charged component in the range of rigidi- 

 ties determined by the cut-off values for the two angles, this intensity 

 being expressed as the fraction due to this component of the total 

 number of coincidence counts. Only the relative intensities are in- 

 volved, and the measurements do not have to rely upon the calibra- 

 tion of the sensitivity of the instrument. Changes of sensitivity dur- 

 ing a measurement, however, must be avoided. Because of the rapid 

 change of intensity with zenith angle, caused by atmospheric absorp- 

 tion, this angle must be kept accurately the same in both azimuths. 

 With a system of frequent rotations automatically controlled by 

 clock works, about an accurately placed vertical axis, and with read- 

 ings taken automatically, the problem of keeping a constant sensi- 

 tivity under field conditions has been largely overcome and intensity 

 ratios have been determined with an accuracy approaching the theo- 

 retical limit of the statistical fluctuations in the total number of rays 

 counted. In many instances intensity differences of the order of 10 

 percent have been measured with an error of less than 1 percent. 

 The apparatus which has been developed for this purpose is repre- 

 sented in plate 3. 



The results of the asymmetry survey are combined in figure 5. 

 For each station the asymmetry, defined as above, is plotted against 

 zenith angle. The ordinates are thus equal to the unbalanced charged 

 component in the range of rigidities determined by the correspond- 



