MAGNETISM, TERRESTRIAL (see 17.353). During 1909-21 further developments occurred in various directions in the study of terrestrial magnetism; and these are dealt with below.
Instruments. The intercomparison of the magnetic instruments of different countries has long had a recognized importance, and a good many comparisons have been effected since the early part of the 20th century, especially by the Department of Terrestrial Magnetism of the Carnegie Institution of Washington. A difficulty increasingly realized is the uncertainty whether any magnetometer can be considered invariable to the degree of accuracy ordinarily aimed at, viz. lV(o-ooooi C. G. S.) in force and o'-i in direction. Until this difficulty is surmounted the establishment of any inter- national standard would probably be premature. The use of a coil and an electrical current promises greater accuracy for force measure- ments than the ordinary magnetometer, with less risk of undetected change. Coil instruments for the measurement of H (horizontal force) have accordingly been designed in various countries, including Russia, Japan, Britain and America, but the accuracy attained remains for investigation. An additional reason for the development of an electrical method of measuring H is that the Q term in the deflection formula 2wr a (i +Pr 2 +Qr < ) where m is the magnetic moment, r the distance is not really negligible under ordinary conditions. This complicates the observation by requiring three deflection distances, and the apparent changes in P and 6 in some instruments raise doubts as to the measure of success attained. An instrument for the direct measurement of the vertical force would possess advantages, especially in high magnetic latitudes.
Survey. Survey work on land and sea has gone on continuously under the auspices of the Carnegie Institution, and many of the results have been published. This work includes a survey of Aus- tralia by E. Kidson, and surveys of numerous remote regions in Asia, Africa and South America. Other recent surveys are those of New Zealand by C. Coleridge Farr, Belgium by A. Hermant, and Spain (preliminary) by U. de Azpiazu and R. Gil. There have been re-surveys of the British Isles by G. W. Walker, and of Japan under the general guidance of A. Tanakadate. A survey of India has long been in progress.
Diurnal Inequality Potential. The derivation of the diurnal variation from a potential suggested by A. Schuster has been fur- ther considered. Assuming the derivation possible, W. van Bem- melen and S. Chapman have found the part of the potential repre- senting internal forces relatively larger than Schuster did.
A. van Vleuten, examining the hypothesis critically, especially as applicable to quiet-day phenomena, has concluded that the balance of evidence is against it. A difficulty in the way of any general theory of the diurnal variation is the varied influence of disturbance. Not merely is the influence different at different places. At the same place it differs for the several elements, varies according to the season of the year, and is not the same at the same season of different years.
Eclipse Phenomena^. Several attempts have been made to ascertain how magnetic phenomena are affected by an eclipse of the sun. No final conclusion has yet been reached, but it seems fairly certain that if an eclipse effect exists, it is small, and probably represents a slight retardation in the changes natural to the hour. _ Sudden Commencements. Another inquiry depending on coopera- tion has been into the simultaneity or non-simultaneity of the so-called " sudden commencements " (S.C.s) of magnetic storms. Several independent investigators have failed to find any certain difference between the times of commencement at remote stations. It has been found, however, that S.C.s differ markedly in type in high and low latitudes, and their apparent duration may differ considerably at different stations. The investigation is complicated by the great difference in sensitiveness and m type exhibited by magneto-graphs.
De Bill " Characters." The study of magnetic phenomena has benefited from the international scheme which has had its head- quarters since 1906 at De Bilt, Netherlands. Each cooperating station assigns to every day a " character " figure, o, I or 2, accord- ing as the day is quiet, moderately disturbed or highly disturbed. The arithmetic means of the figures so assigned are given in the annual De Bilt lists, so that all days have international " character " figures^ ranging from o-o to 2>o. The intercomparison of the " char- acters " assigned at different stations has confirmed the view that disturbance is seldom, if ever, confined to one area.
Twenty-seven Day Period. The existence of an authoritative international measure of disturbance for individual days has enabled the reality of a " 27-day period " in magnetic disturbance to be investigated without any possible prejudice. The phenomenon has been found to exist for quiet as well as for disturbed days. A day which follows a conspicuously disturbed (or quiet) day after an interval of from 26 to 28 days has decidedly more than the average chance of being itself a disturbed (or quiet) day.
Diurnal Variation. Again, comparisons have become possible of the diurnal inequalities in days, undoubtedly representative of quiet and of disturbed conditions. It has been found that in high southern latitudes and presumably equally in high northern latitudes the diurnal inequality is exceedingly sensitive to dis- turbance, and more especially so in winter. For instance, at the station of the Scott Antarctic Expedition in 1911-2 the diurnal
inequality derived from the five days of largest international " char- acter " figure in each of the four midwinter months had in each of the elements a range fully four times that of the inequality derived from the international quiet days (five a month). The type of the inequality was in this case but slightly affected by disturbance.
" Character " v. " Magnetic Activity." For these and similar investigations the really important thing is to discriminate between the magnetic conditions on days of the same month or season. For this purpose the existing De Bilt scheme seems eminently satis- factory. But it is otherwise when we wish to compare different years. There is a natural tendency to vary the standard of " character " according as the year is quiet or disturbed. " Character " i-o is apt to mean decidedly less disturbance near sunspot minimum than near sunspot maximum. To remove this difficulty, F. Bidlingmaier sug- gested the use of " Magnetic Activity," defined as the mean value throughout the day of ir (a'+^+y-) where a, /3, 7 represent the departures of the 3 magnetic components from their normal values. Apart from the difficulty of assigning normal values, the amount of labour entailed appears prohibitive. An alternative, suggested by C. Chree, to employ the square of the absolute daily range, or some quantity based thereon, has been tried at Eskdalemuir and De Bilt. It seems fairly workable, but entails sensibly more labour than the " character " scheme.
Variation of Disturbance with Latitude. A comparison of the disturbances recorded in 1911-2 at the Scott Antarctic station with those at various other stations extending from Mauritius to Sitka (Alaska) has shown a marked tendency for disturbance to be simultaneously great in high northern and southern latitudes, and to be less at intermediate stations, whether north or south of the equator. Again, comparison of auroras with magnetic " character " figures, got out for individual hours as, well as for individual days, has confirmed the view that an intimate relation exists between aurora and magnetic disturbance. This has long been accepted as true of large magnetic storms and aurora in countries such as England, where both phenomena are rare ; but it seems also true of less extreme cases of the phenomena, which abound in high latitudes. The two phenomena at a station in high latitudes do not perhaps as a rule wax and wane together, but if we select a number of days (or hours) when there is bright aurora, the mean magnetic " character " of these days (or hours) is well above the average. It is doubtful whether visible aurora is not the rule rather than the exception in high latitudes. Faint aurora is undoubtedly sometimes seen on days which magnetically are quieter than the average.
Theory. The various phenomena we have referred to the synchronous occurrence of magnetic disturbance in different partg of the earth, its special development in high latitudes, the en- hancement of the regular diurnal variation during disturbance, the association of disturbance with aurora, and the " 27-day period " all accord in a general way with the theory pretty generally held that magnetic disturbance and aurora are both due to the emission from the sun of some species of electrical radiation. On the nature of the radiation there is, however, a lack of agreement. Kr. Birke- land, one of the earliest and most eminent supporters of the theory, believed in /3 (negative) rays, possibly of greater velocity than any of artificial origin. C. Stormer, another eminent Norwegian authority, favours a (positive) rays. Difficulties in the way of either theory undiluted have been discussed by A. Schuster, S. Chapman and F. A. Lindemann. One of the chief difficulties is the scattering to be expected from the mutual repulsion of ions all of one sign.
When approaching the earth, ions, whether positive or negative, would naturally spiral round the lines of magnetic force, and so be concentrated in high latitudes. This fits in with the special develop- ment of aurora and disturbance in high latitudes, but leaves unex- plained the occasional appearance of aurora in latitudes like those of central and southern Europe. The presence of ions should increase the electrical conductivity of the atmosphere, thus enhancing the amplitude of the regular diurnal variation, supposing that due, as is generally believed, to electrical currents in the upper atmosphere. The fact that disturbance enhances the diurnal inequality relatively more in winter than in summer would follow naturally from the very probable hypothesis that ordinary solar radiation itself in- creases the conductivity of the atmosphere. In high latitudes we may suppose that in winter the electrical currents associated with aurora play the part which in summer is largely played by direct solar radiation.
In the simplified mathematical theory of aurora and magnetic dis- turbance developed by C. Stormer the earth is treated as magnetized as it would be if only the first order Gaussian harmonic existed. Everything is then symmetrical round the axis of the first harmonic, whose north end according to J. C. Adams was in 1880 at 78 24' N. and 68 4' W., but moving west through about o -I per annum.
In a theory of magnetic storms S. Chapman also assumes sym- metry round an axis regarded as magnetic latitude 90. He has derived mean results from 40 storms recorded at 12 stations arranged in 5 groups according to magnetic latitude. Confining himself to storms having an S.C., he regards the " general storm " phenomena, i.e. the phenomena not dependent on local time, as determined by the " storm time," or time elapsed since the S.C. The phenomena varying with local time he regards as composed of the ordinary diurnal variation + a " local storm variation," determined like the
