DEPARTMENT OF TERRESTRIAL MAGNETISM. 327 



The summaries in both tables 3 and 4 support Dr. A. Nippoldt's view, as 

 based upon the Galilee observations in the Pacific Ocean in 1907-1908, that 

 the effect of the land is to increase the value of the conductivity as measured 

 at sea. 



From a summary of the various results thus far obtained at sea, the following 

 deductions in regard to the mean values of the elements may be drawn : The 

 potential gradient is of the same order of magnitude over the sea as over the 

 land; the radioactivity of the air over ocean areas far removed from land is 

 small, compared to that found over land; and that the ionization over the 

 ocean is at least as large as that found over land. 



Investigation of certain causes responsible for uncertainty in the measurement of atmos- 

 pheric conductivity by the Gerdien conductivity apparatus. C. W. Hewlett. Terr. 

 Mag., vol. 19 (Dec. 1914). 



The object of the writer was to test experimentally the behavior of the 

 apparatus under various conditions. It was shown by Dr. Swann that the 

 theory of the apparatus does not necessitate a constancy of the velocity over a 

 cross-section of the air-current, and that, if the potential is below a certain 

 minimum value depending on the total air-flow, the correct value of the con- 

 ductivity is obtained. By covering up half of the space opposite the fan, a 

 considerable irregularity was produced in the air-flow. For low charging 

 voltages, however, the conductivity was practically the same as without the 

 obstruction. As the voltage is increased beyond a certain point, the meas- 

 ured conductivity, without the obstruction, becomes less than that with it, 

 but this is simply due to the fact that the critical voltage is less with the 

 obstruction than without, owing to the smaller air-flow in the former case. 



As the measured conductivity was found to depend upon the presence or 

 absence of the funnel, it was thought advisable to investigate the exact effect 

 of the latter. Certain theoretical considerations point to the conclusion that, 

 in the absence of the funnel, the maximum allowable voltage is reduced owing 

 to the fact that the stream lines of the air which enter the funnel near the 

 edges suffer sharp bends in that vicinity. Experiments were consequently 

 made to determine whether the effect of the funnel was less important for low 

 charging voltages. It was found that for charging voltages of 10, 30, 50, 70, 

 100, the relation of the conductivity without the funnel to that with the 

 funnel were 0.98, 0.96, 0.98, 0.94, 0.89; hence, the theoretical conclusion that 

 the ratio is practically unity for low voltages, and less than unity for higher 

 voltages, is borne out. 



Experiments have also been made to determine the effect of the charge 

 induced on the earthed portions of the apparatus owing to the existence of the 

 potential gradient. It is found that under certain conditions this effect may 

 be considerable. 



* The usual theory of the Gerdien apparatus neglects the effect of the charge 

 collected by the rod which supports the central cylinder. Experiments were 

 made to measure directly this charge, and it was found to be far from negligible 

 in comparison with that collected by the central cylinder. 



In the use of the Gerdien apparatus it is important that the charging 

 potential used shall always be sufficiently low. The most sensitive region of 

 the electroscope supplied with the instruments is usually in the neighborhood 

 of 150 to 200 volts. It has been found, however, that the maximum allowable 

 voltage varies considerably from day to day, presumably owing to variations 

 in the specific ionic velocities. Usually a charging potential of 150 volts is 

 far too high. 



