ATMOSPHERIC The Edinburgh and London data are from the papers by Mossman already mentioned, those for Tilsit are due to Kassner (see Met. Zeit. for 1894, p. 237). Table XIY. Mean Annual Number of Thunderstorms. Decade ending 1800 1810 1820 1830 1840 1850 1860 j 1870 1880 1890 Edinburgh 4-0 4-9 57 7-7 6-7 5*7 6-5 5-4 10-6 9-4 London 7-9 9-5 8-3 11- 11-8 10-5 11-9 | 9-6 15-7 13-0 12- 12-1 16-1 15-3 11-9 17-6 21-8 Tilsit . Mossman expresses the opinion that the apparent recent increase in the frequency of thunderstorms at Edinburgh and London is, to some extent at least, a true phenomenon. Unquestionably the two sets of figures show corroborative features—notably the fall in the decade 1861 to 1870—but it is difficult to feel certain that the notable extension in the limits of both cities may not have exercised an influence. The diminished intensity in the decade 1861-70 is, it may be added, in accord with Lawson’s vital statistics. In Germany various authorities believe in a great recent increase in thunderstorms. Von Bezold {Berlin. Sitz. 1899, No. 16) has given a table showing the number per million of insured houses struck in Bavaria for each year from 1833 to 1897. Prior to 1868, the number struck in one year never exceeded 78 per million; between 1868 and 1883 it exceeded 100 per million in five, but only five, years ; since 1884 it has only once been less than 100 per million, and it exceeded 200 per million in 1895, 1896, and 1897. Kassner (reviewed Met. Zeit. 1890, p. 385), dealing with similar insurance statistics for Central Germany, concluded that during the twenty-six years 1864 to 1889, destructive lightning strokes had increased 126 per cent, as against an increase of only 11 per cent, in houses. Von Bezold seems to accept a large increase in thunderstorm intensity in Germany as an undoubted fact, and he is disposed to associate it with the great increase in smoky factory chimneys, and in the number of wires and rails all over the country. It is possible, of course, that changes in the ordinary height or construction of buildings, or influences which only an insurance expert could assess, may have had an effect. § 33. Table XV. gives data based on the number of trees of different species struck by lightning during a series of recent years in the forests of Lippe, in Germany, as quoted by Henry (l.c.). Table XV. Beech. Larch. Ash. Birch. Species of Tree. I Oak. Fir. 21 Total struck 159 59 I 20 5 1 Mean liability . j 57 39 I 5 In calculating the liability, allowance is made for the relative numbers of the different species. The different years are treated separately, the liability of beech to be struck being taken as unity throughout. Prohaska (see Met. Zeit. 1899, p. 128) gives somewhat analogous data for Styria. He concludes that oaks, poplars, and pear trees are especially liable to be struck, while beech is exceptionally safe. Jonesco (see Henry, l.c.) has concluded from experiments that fresh wood, rich in starch, but poor in fatty material, possesses the greatest electrical conductivity, and that the conductivity in some species of living trees varies a good deal with the season of the year. He believes that the liability to be struck by lightning is greater the higher the conductivity. Allowance must, however, be made for the fact that different species of trees attain different heights, and that one species may be common in marshy, another on rocky ground. According to Heilman (quoted by Henry), the liability to lightning stroke on different soils in Germany may be put at: chalk 1, clay 7, sand 9, loam 22. Differences of this kind might influence the apparent liability of different species of trees to be struck. § 34. Luminous discharges from pointed objects are not infrequent in mountainous districts, especially during thunderstorms. On the Sonnblick, where the phenomenon is of frequent occurrence, St Elmo’s fire has been found to answer to a discharge sometimes of positive, sometimes of negative electricity (Elster and Geitel, Met. Zeit. 1891, p. 321, and 1894, p. [68]). The colour and appearance of the light differ in the two cases, red predominating in a positive, blue in a negative discharge. Stade {Met. Zeit. 1898, p. 238) also describes the appearance of St Elmo’s fire of different signs, as observed on the Brocken. § 35. A description of the more prominent features of aurora will be found in the Ency. Brit. vol. xvi. pp. 177, 178, 183. Table Auroras. XVI. somefrequency statistics more recentlyinpublished as to contains the relative of auroras different months of the year. The data for Scotland and London are from papers by Mossman (see Met. Zeit. 1898, p. 307), the rest are quoted by Ekholm and Arrhenius {Kongl. SvensJca Vetenkaps-
ELECTRICITY
781 The figures are Akademiens Handlingar, Bd. 31, No. 2, 1898). percentages of the mean number for the year. Table XVI. Relative frequency of Auroras throughout the Year. United From 16° S. N.E. Scot- London Norway States to 39° S. land (122 (189 years). (1861-95). (1871-93). (1783-1894). years). 411-5 68-6 10-9 January5 . 15-2 9-2 510-5 127 February . 1 159-4 12-0 10-2 8-8 March 4-1 10-9 12-7 107 7-1 April 3-8 2-2 40-3 8-0 May 1-1 o-o o-o 3-8 67 June 0-4 1-9 o-o 2-4 77 July 0-8 4-4 57-2 7August 12'9 14-5 9-6 10-9 .U-9 September 1616-9 15-8 10-3 21-1 October . 12-0 9-6 15-0 7-8 November 10-4 9-6 611-9 5-8 December. 6Table XVI. is in general agreement with the statement in the Ency. Brit, that auroras are most frequent near the equinoxes. In comparing winter and summer data, especially in high latitudes, it must be remembered that faint auroras are invisible in twilight. § 36. Table XVII. may be regarded as a continuation of one in the Ency. Brit. vol. xvi. p. 178. It gives Wolf’s relative sun-spot numbers alongside of numbers proportional to the frequency of auroras. In the case of Edinburgh the figures are the actual number of auroras as given by Mossman {l.c.). The other data are from a paper by Ekholm and Arrhenius {K. Svenska Vet.-Akad. Hand. Bd. 31, No. 3, 1898). Table XVII. Sion-spot and Auroral Frequency. Year.
Wolfs Number.
1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894
73-9 139-1 111-2 101-7 66-3 44-6 17-1 11123-4 6-0 32-3 54-3 59-6 63-7 63-5 52-2 25-4 136-8 6735-6 73-8 84-9 78-0
Auroral Frequency. Edinburgh, j N. America. 16° to 39° S. 5 3 19 123 21 408 60 565 8 105 12 781 8 10 1443 11 4 741 6 0 676 2 2 877 L 1 258 0 405 0 1 12 646 5 2 804 17 5 860 138 2 686 28 3 533 0 5 728 974 6 0 610 3 1 759 9 0 336 1 0 358 1 7 732 1 9 614 48 2 1067 9 12
§ 37. During the polar year, September 1882 to September 1883, numerous aurora observations were made at many of the stations. Those at the Finnish station, Sodankyla, have been discussed by Lemstrom and Biese {Exped. Polaire Finlandaise, t. hi., Helsingfors, 1898). In addition to the more usual phenomena the Finnish observers noticed a variety of others. On a good many occasions, in the absence of ordinary aurora, they saw a yellowish-white illumination, showing in the 7spectroscope the characteristic auroral line (wave-length 5569 x HT mm., according to Lemstrom). On some occasions, in the absence of any phenomenon visible to the unaided eye, Lemstrom saw the auroral line wherever he turned the spectroscope. The most outstanding phenomena, however, described by Lemstrom and Biese are artificially produced luminosities, in the shape usually of flames, but occasionally resembling auroral rays. The auroral line was usually detected, but was feeble. The apparatus consisted of a number of sharp points connected by wires carried on insulators, the whole enclosing an area of several hundred square metres on the top of one or other of a
