OBSERVATIONS ON ATMOSPHERIC ELECTRICITY. 587 



than the cloud, but a careful study of the observations will show that this is 

 not quite an accurate statement of the case, as some of the observations show 

 a high potential outside, but near a thick mist. Of course most clouds are 

 dissipating at the outside, so that a low potential will, if I am right, be usually 

 found there. 



It seems worth pointing out that, if we neglect the readings made in mist, 

 the highest were got on the 5th, when it was almost perfectly calm and there 

 was a thin but very decided haze. The temperatures during the day were also 

 above the average. 



The electrometer used throughout my observations was Thomson's small 

 portable electrometer, No. 43, which gives a reading of about 24 divisions 

 for 100 Daniell cells, when tested at a temperature of about 90° F. I find, 

 however, that there is a large temperature effect on the zero readings, and I 

 have not been able to test whether or not this affects the value of the scale 

 readings. In Madras the leakage is so great and the temperature range so 

 small that I could never be quite sure of the effect of temperature on the zero 

 reading, but it was different in the comparatively dry atmosphere of Dodabetta, 

 where the instrument retained its charge well and where it was used in 

 temperatures varying from about 38° to 69° in the shade, but really with a 

 much wider range since the day observations were made in the sun, in which a 

 black bulb thermometer rose to from 125° to 138°. The accompanying curve 

 (PI. CXLVI.) shows the changes in the zero readings during four days, and an 

 approximate curve for the shade temperatures during the same period. The 

 temperatures are only a very rough approximation, as during the time that the 

 sun was above the horizon they were taken between the outer and inner walls 

 of the tent near the door, on the side away from the sun, no better shade being 

 available ; at night the thermometer was placed outside 4 feet from the 

 ground.* It will be seen from these curves that on the 8th the zero rose from 

 1202 at 14 h to 1261 at 23 h a rise of nearly 5 per cent. At first this large 

 temperature change gave me considerable trouble, as I found that the earth 

 readings before and after the air reading differed considerably from each other, 

 but this difficulty was reduced to a minimum by leaving the electrometer on 

 the observing post for a few minutes before making the observation. It will be 



The minimum thermometer gave readings varying from 32°"5 on the 7th to 40° on the 5th and 

 6th, with an average of 36°. It is interesting to compare this with the reading made in the Dodabetta 

 observatory, and I have accordingly taken the observations for the five years from 1848 to 1852 for the 

 corresponding days of January. I find the lowest recorded temperature is 40° - 4, while the average is 

 45° - l, or 9° higher than the average which I got. Only a small part of this difference can be ascribed 

 to the unknown error of the thermometer used then, and it is equally improbable that the present year 

 was so much colder than any of the years during which the observations were made. The only remain- 

 ing explanation is that, as popularly believed, the observer took his thermometers inside the hut that 

 so he might escape the necessity of going out into the cold mist-laden air after nightfall. 



VOL. XXXII. PART III. 5 E 



