778
ATMOSPHERIC
during a thunderstorm, the charge brought down by the rain, when greatest, was approximately at the rate of 76 x lO-14 coulombs per second per sq. cm. of earth’s surface. This was the maximum met with. In their paper in Terrestrial Magnetism, Elster and Geitel give graphical illustrations of their results. § 15. At most stations a negative potential gradient is exceptional, except during rain or fog. During rain it is a frequent but by no means invariable phenomenon, and it seldom persists long without a recurrence to positive. The alternations of sign during heavy thunderstorm rain are often both frequent and sudden, and a similar phenomenon is not unusually observed during thick fog. In some localities a negative potential gradient seems by no means uncommon, even on bright calm days. Thus Michie Smith (Phil. Mag. vol. xx. 1885, p. 456) observed numerous instances of negative potential at Madras during bright days in August and September. The phenomenon was quite common between 9.30 a.m. and noon, during westerly winds, which at Madras are very dry and usually dusty. The presence of dust seemed to Michie Smith a contributory cause, but dust in easterly winds was not accompanied by negative potential. At the Finnish polar station Sodankyla, in 1882-83, Lemstrom and Biese found that out of 255 observed occurrences of negative potential, 106 took place in the absence of rain or snow. The proportion of occurrences of negative potential under a clear sky was much above its average in autumn, and much below its average in spring and summer. In many cases Lemstrom and Biese observed no change of sign to follow on rain or snowfall. At Polhem, in Spitzbergen, Wijkander is reported to have observed negative potentials almost as often as positive; but the potential gradients at Polhem seem to have been abnormally low. At the polar station Godthaab, in 1882-83, negative potential seemed sometimes to be associated with aurora (Paulsen, Bull, de VAcad. . . . de Danemarke, 1894, p. 148). § 16. It has been found by Lenard, Elster, and Geitel and others, that the potential gradient is negative near a waterfall. The influence may extend to a considerable distance. Lenard ( Wied. Ann. xlvi. 1892, p. 584), who has specially investigated the matter, concludes that when pure water falls upon water the air in the neighbourhood becomes negatively charged ; the presence of dust in the air is, he believes, unessential. The effect can be produced inside a closed space (Maclean and Goto, Phil. Mag. vol. xxx. 1890, p. 148), but under such circumstances the presence of dust seems to increase the effect. A little impurity in the water markedly diminishes the charge, and may alter its sign. More recent experiments by Kelvin, Maclean, and Galt [Phil. Trans, vol. cxci. A, p. 187) have shown that when air is charged by dropping water, as in Maclean and Goto’s experiments, the charge is considerably greater in the air near the level of impact of the falling water than in that at a higher level. A sensible charge still remained, however, when the influence of the splashing was eliminated. In the opinion of Kelvin, Maclean, and Galt, this property of falling 'water renders an ordinary water-dropper unsuitable for use indoors, at least in a small room, though it constitutes no practical objection to its use out of doors. Exner (Wien. Sitz. (2) 1896, p. 467) observed an electrical action in the case of breaking sea-waves, noticing the spray to be negatively electrified. § 17. Electrification is generally admitted to exist in the atmosphere, but some authorities have held that it is carried by dust, or by ice and water particles forming clouds. Lord Kelvin’s view is that ordinary air, whether containing dust or not, can be electrified, either positively or negatively. This conclusion is supported by experiments by Kelvin and Maclean (Proc. Roy. Soc. vol. Ivi. 1894, p. 84 ; Phil. Mag. August 1894), and by more recent experiments by Kelvin, Maclean, and Galt {Phil. Trans, cxci. A, p. 187). In the earlier experiments the electrification was measured by means of a water - dropper. In the more recent ones, for the reasons specified in § 16, other methods were employed, and by means of needle-5points attached to an electric machine a charge of density 37 x 10 C. G.S. electrostatic units was given on one4 occasion to ordinary air, while a density as great as 22 x 10~ units was attained by means of an electrified hydrogen flame. After stopping the electric machine, the electrification gradually
ELECTRICITY disappeared, but could be detected for many minutes. In the earlier experiments little difference was observed between very dusty air and air as dustless as possible. Negative electrification seemed to disappear, when the machine was stopped, more rapidly than positive. In the later experiments it was found possible to give a small electrification to air in other ways, e.g., by shaking it up with water or solutions of various kinds, or blowing it in bubbles through water or aqueous solutions. § 18. Linss {Met. Zeit. 1887, p. 352) found that an insulated paste-board cylinder, covered with tinfoil and charged either positively or negatively, lost its charge gradually when in the free atmosphere. The potential Y existing after dissipation during time t was connected with the original potential V0 by a formula of the type Y = where q though constant on a particular occasion varied appreciably from day to day. The formula implies that the rate of fall of the potential varies as its existing value (counting the earth’s potential as zero). Linss’s experiments have been repeated by Elster and Geitel {Terrestrial Magnetism, vol. iv. 1899, p. 213). Their dissipator consisted of a hollow cylinder of blackened copper, insensitive to any direct photo-electric effect. When used out of doors, it was protected from rain, wind, and direct sunlight by an insulated hood. Regular daily observations were made for a long time, the loss during a given interval being measured for both positive and negative charges. In addition to numerous experiments at Wolfenbiittel, a considerable number were made on the Brocken and on the Alps, near Zermatt. Elster and Geitel denote the fraction of the charge lost per minute by E, distinguishing the sign of the charge by the addition of + or -. In the lowlands they found on the average E + = 1 ’3 per cent. In other words, positive and negative charges were lost, on the whole, equally fast, and on the average a charge of 100 units lost 1-3 units in one minute. This result agrees pretty well with that obtained by Linss. Neither wind nor mist was found to possess much influence, and, somewhat contrary to expectation, the most rapid loss of charge was observed on days when the air was specially clear and free from dust. In mountainous districts they found a change in the phenomena, as may be seen from the following typical results, obtained during the ascent of a peak near Zermatt:— Position. Zermatt valley On level ground . Near end of ridge . Peak, Gorner Gratz
Height above Sea. Metres. 1620 2600 3000 3140
E(+).
E(-).
Per cent. 4-5 4-4 27 07
Per cent. 4'4 6'8 7-0 6-6
From these and similar observations Elster and Geitel conclude: (1) that normally in valleys high above the sea the rate of dissipation is greater than near sea-level, whilst remaining independent of the sign of the charge ; (2) that on mountains, especially near peaks, a negative charge is lost much more rapidly than a positive charge. A diametrically opposite phenomenon was observed by them near waterfalls. For instance, near a fall at Zermatt they obtained 16 '2 per cent, for E( +) as against 1 ‘9 per cent, for E( -). They attribute the loss of charge to the existence of small positively and negatively charged masses in the atmosphere. In the lowlands or in mountain valleys the two species are equally numerous on the average, but positive masses largely predominate near mountain peaks and negative masses near waterfalls (cf. § 16). § 19. It has long been a disputed question whether vapour from a charged liquid carries off part of the charge ; it is rather a vital point in some theories respecting atmospheric electricity, such as that of Exner. Yarious experimenters {e.g., Blake, Mascart, Sohncke, Pellat) have advanced what they regarded as crucial evidence in favour of one view or the other. The question is discussed by J. J. Thomson, in his Recent Researches in Electricity and Magnetism, p. 54, and Trabert has treated it more recently in the Met. Zeit. for 1899, p. 377. Several supporters of the view that the vapour carries off a charge, e.g., Lord Kelvin {Proc. Roy. Soc. 31st May 1894, p. 84), have pointed out that if it be conceded that a drop of pure water can be charged and evaporated, the charge must be carried off by something which is either aqueous vapour or is unknown to science. § 20. In Exner’s theory of atmospheric electricity aqueous vapour is the essential element. If, as before, P be the potential gradient in volts per metre, and q be the density of aqueous vapour present in the atmosphere, then, according to Exner, P = A/{l+kq), where A and k are constants. In its extreme form the theory would imply that the above formula applied with A and k absolute constants in the case of every recorded value of P on clear days. Elster and Geitel, however, in their applications of the formula, seem generally to regard P and q as answering to the mean values during a day’s observations. On other occasions
