ATMOSPHERIC ELECTRICITY 779 they have attributed to A values differing from one series of ex- rays do really show this difference, negative ions serving as conpel iments to another. Exner himself (Wieji. Sitz. xcix. Abth. ii. densation nuclei under considerably less supersaturation than 1891, p. 601) has applied his formula with A = 1400 and k~l-l5 positive ions. He found, however, that the non-dust nuclei, to observations at Vienna, Wolfenbiittel, St Gilgen, and India. always present in small numbers in the atmosphere, and the Arranging his observations in groups, each of which answers to a nuclei produced in moist air by ultra-violet light do not behave as narrow range of vapour density, he compares the mean observed free ions ; and whilst there is considerable reason to believe that and calculated values of P for the several groups. The tendency ionization develops in these nuclei in the process of supersaturafor the mean P to be small when the mean q is large is decided. tion this remains to be proved. Supposing ions to act as conElster and Geitel (Wien. Sitz. ci. 1892, p. 703) applied Exner’s densation nuclei in the atmosphere, Wilson’s experiments seem formula to groups of observations made by them at Wolfenbiittel to that this action would be chiefly limited to negative on quiet, comparatively clear days. They found a fair general ions,indicate and that the rain so formed would mainly carry down negative agreement between the observed and calculated mean values of P electricity. As he himself remarks, however, Elster and Geitel’s for the groups, but within the same vapour density group the (§ 14) do not show any conspicuous preponderance individually observed values of P often differed conspicuously. experiments of negative electricity in rainfall. Again, the supersaturation This had been noticed previously by Exner himself. Independ- required before even negative ions act as condensation nuclei is ent critical examinations of Exner’s theory have been made by considerable, and Aitken {Nature, March 1900, p. 514) believes Chreo (Proc. Roy. Soc. vol. lx. 1896, p. 96), Braun (review Met. that the number of dust particles usually present in the atmoZeit. 1896, p. [78]), Trabert {Met. Zeit. 1897, p. 106), and Gockel sphere is so large as to afford little opportunity for ions to act. {Met. Zeit. 1897, p. 281), dealing respectively with observations and Geitel on the other hand {Terrestrial Magnetism, vol. iv. at Kew, Bamberg, Meiningen, and Ladenburg. In each case the Elster 1899, p. 213) regard the action of ions, positive as well as negative, conclusion reached was that the relation between potential gradient in all probability a factor of importance. and vapour density is not a unique and fundamental one, as as §23. Everett (PAW Trans. 1868, p. 347), in discussing the Exner supposes. When observations taken at all seasons of the Kew atmospheric electricity results, specified certain points of year are combined together, the great majority of instances of resemblance between the diurnal variation and that of barometric large vapour density (or of high temperature) will be found to pressure.. The two elements agree in having two maxima and have occurred in summer, when the potential gradient is normally minima in their diurnal inequalities, and the times of their low ; thus there can hardly fail to be a certain amount of agree- occurrence, at least at Kew, are fairly similar. Hann {Met. Zeit. ment between observation and any theory which associates - low 1889, p. 106j; 1890, p. 29 ; 1891, p. 113) has dwelt on this repotential with high vapour density (or high temperature). semblance in reviewing atmospheric electricity observations at § 21. In 1887 Hertz (Wied. Ann. xxxi. p. 983) found that the various and has further pointed out that the relative imdischarge between two poles was facilitated by the presence of portancestations, of the semi-diurnal as compared with the diurnal term in a Ultraultra-violet light. Subsequent experimenters (see J. J. Fourier analysis, in the case of electric potential as in that of baroviolet ^ homson’s Recent Researches in Electricity, arts. 39 to metric pressure, appears less in arctic than in temperate latitudes, light. thatthat thevarious seat of metallic this action was the cathode, and is greatest in tropical stations. and found further, substances—notably § 24. In the papers mentioned in § 20, Chree, Braun, Trabert, freshly-polished zinc—when insulated became positively electrified, and Gockel all describe a closer association of high potential or rapidly lost a negative charge when exposed to ultra-violet gradient with low temperature than with low vapour density. light from any source, including the sun. This species of dissipa- Trabert seems almost disposed to regard this association as fundation is confined to negative charges, and depends principally on mental, absence presumably of, disturbing causes such as the material of the charged surface and the freshness of its polish; rainfall. in Ifthe this were the case, we ought to encounter exceptionally but it also depends on the nature of the surrounding gas and its high potential gradients at arctic stations, especially at night in pressure. . The property, though prominent in potassium, sodium, winter. At Cape Thorsden, however, so far as the recorded posiand zinc, is inconspicuous in most common materials at the earth’s tion of the water-dropper enables one to judge, the potential surface; Elster and Geitel, however(IFied. Hran. xliv. 1891, p. 722), gradient was exceptionally low. At Sodankyla, Lemstrom and found sensible traces of it in fluorspar and a few other minerals! Biese found the mean potential gradient for the colder half of The earliest recognition of ultra-violet light as a possibly important the day was that considerably less than that for the warmer half during factor in atmospheric electricity, seems due to Arrhenius (Met. winter and spring, and that for the year as a whole, the mean Zeit. 1888, pp. 297 and 348). Whilst mentioning the possible during the warmer hours stood to that during the colder effect of ultra-violet light incident on the earth or on particles in gradient hours in the ratio 823 : 735. Again, while the chief maximum in the atmosphere, he relies mainly on its making the air it traverses the diurnal inequality of the potential gradient usually occurs after conduct electrolytically. Elster and Geitel, on the other hand, sunset, the principal minimum at many stations occurs before sunattach main if not exclusive importance to its action at the earth’s rise, when the temperature is at its lowest. Finally, whilst the surface.ra From experiments with electric light they concluded that diurnal inequality of the potential gradient usually shows two the te of loss of a negative charge by a small sphere of freshly maxima and minima, that of temperature has normally but one. amalgamated zinc, was connected with J, the intensity of ultraObservations by Elster and Geitel on 19th August 1887, violet light,^by a linear relation z = a + M {Wien. Sitz. ci. Abth. and§ 25. on 22nd January 1898 {Met. Zeit. 1888, p. 27, and ii. 1892, p. 703). Having found a and b experimentally, they 1899,byp. Ludwig 281), agreed in showing a diminution of potential gradient determined J from day to day from the observed rates of loss of during the totality of a solar eclipse, followed by a smart rise after charge in sunshine. They tried several formulse connecting J totality. These observations possess a certain significance in conwith the potential gradient P, the simplest being of the type nexion with the ultra-violet light theory of Arrhenius, to which P —A/(a + /3J), where A, a, (3, are constants. In comparing ob- they seem opposed. Sudden changes of potential gradient are, served and calculated values of P, they followed Exner’s example, so frequent, that much weight cannot be assigned to arranging the observations in* groups according to the value of j! however, observations so limited in number. Observations under somewhat Whilst some of the series of observed and calculated mean group analogous conditions, but pointing to an opposite conclusion, are values show a surprisingly good agreement, individual observations described by {Met. Zeit. 1899, p. 481). During several days’ within the same group often differ conspicuously. More recently observations Gockel in a small oasis, El Mora near Biscra, he found a Brillouin and Buisson (reviewed Met. Zeit. 1898, p. 38) have marked tendency in the potential gradient to rise rapidly just after found that dry ice is sensitive to ultra-violet light, the influence sunset. being of the order of one-tenth of that observed with zinc. As § 26. Lemstrom {Brit. Assoc. Report for 1898, p. 808) has been the ice surface melts, the effect rapidly diminishes, becoming experimenting 1885 on the influence of artificial electric fields vanishingly small when a complete layer of water has formed. or currents on since the growth of plants. The influence is, he savs, This discovery has been questioned by Benndorf. If confirmed, m general highly beneficial; for instance, it shortens the time it would increase the probability that ultra-violet light plays an required by strawberries and raspberries to ripen, and increases the important part in atmospheric electricity. It may, however, be injurious, especially during bright § 22. Recently, owing mainly to the researches of C. T. R. yield. unless abundance of water is provided. Lemstrom believes Wilson, evidence has accumulated as to the existence in atmo- days, atmospheric electricity to play an important part in the growth of spheric air of nuclei—other than dust particles—which act as vegetation in high latitudes, and he assigns a special rdle to the ■centres of condensation for water vapour, when supersaturation is needles of fir pine trees. E. H. Cook {Brit. Assoc. Report produced artificially. Wilson found that ions acting as nuclei for 1898, ^ p. and 809) describes experiments somewhat similar to can be called into existence by Rbntgen rays, Uranium rays, and Lemstrom s. He observed a beneficial effect on plants grown ultra-violet light acting on negatively charged zinc or even on under a pole—especially the negative pole—of a Wimshurst moist air. J. J. Thomson {Phil. Mag. Dec. 1898, p. 533) pointed machine or an induction coil. Cook refers to experiments aiming ■out that if the positive and negative ions should differ in their at the direct utilization of atmospheric electricity. power of condensing water vapour, then a cloud might form round the one set of ions, and that as the cloud particles fell under § 27. The beginning and ending of thunderstorms at a gravity, electrical separation would take place. Wilson {Phil. station are seldom very clearly defined, and the duration Trans, cxciii. p. 289) has found that ions produced by Rontgen is very variable, so that the exact significance to be at-
