Dike — Recent Observations in Atmospheric Electricity. 201 



gave values about the same as on iand, and I see no reason for 

 supposing that the potential gradient should be less over the 

 sea than over the land. The conductivities obtained varied 

 considerably but not more so than at a hxed land station, 

 apparently exhibiting some tendency to decrease as the baro- 

 metric pressure increases. 



The mean conductivities for the voyage were : 



Ap=l'603 X 10 -4 electrostatic units 

 \ n — 1-433 X10- 4 " " 



This would make the vertical current somewhat greater 

 than what Gerdien found, assuming the same potential gradient. 

 The observations for a day usually extended over an hour or 

 more and were made up of alternate readings of the conduc- 

 tivity for positive and for negative electricity, A charge of 

 about 150 to 200 volts was put on the inner cylinder by means 

 of a dry pile and the divergence of the electroscope leaves 

 read. This reading was the most difficult part of the observa- 

 tion at sea since the rolling of the ship gave the leaves a pen- 

 clular motion, while the rise and fall of the ship as it rode over 

 the waves gave them a flapping motion, so that they alter- 

 nately spread farther apart as the ship sank into the trough of 

 the waves and fell together as it rose to the crest. The pen- 

 clular motion was avoided to some extent by mounting the 

 instrument on a gimbal stand and steadying it with the hands 

 to keep it horizontal, but the flapping was unavoidable and 

 limited the observations to moderately smooth seas. Only one 

 leaf at a time could be watched through the eye- piece, so that 

 the estimation of the mean position of the leaves was difficult. 

 However, the consistency of the results shows that some degree 

 of accuracy was obtained. 



Immediately after reading the divergence of the leaves, the 

 fan was rotated at a uniform rate for five minutes. At the 

 end of that time a second reading of the divergence was made 

 and the charge was removed by earthing the inner cylinder, 

 which was then charged negatively. After a wait of about one 

 minute to allow any of the former charge which might have 

 been absorbed by the insulation to be neutralized, a reading of 

 the divergence of the leaves for the negative charge was made, 

 in time to begin the rotation of the fan two minutes after the 

 end of the rotation for the positive charge. Closing both ends 

 of the cylinder to exclude air currents, a leakage test to deter- 

 mine the normal rate of leak over the insulation was made, 

 usually extending over fifteen to twenty minutes. 



The electroscope had to be calibrated of course, and this 

 was done at long intervals, as laboratories for testing electrical 

 instruments are not numerous in the Pacific. An apparatus 



