780 



SCIENCE 



[N. S. Vol. XXXII. No. 831 



itself. Such were the successive inferences 

 of the theory of electrons by ionization as 

 later developed, though not known in 1891, 

 when my work in that subject began. As 

 an attack upon the problem, the observed 

 magnetic field of the earth was subdivided 

 into a normal field and the several deflect- 

 ing magnetic fields mentioned above, the 

 vectors being expressed in appropriate 

 rectangular and polar coordinates. These 

 several fields, (1) diurnal, (2) meridional, 

 (3) perturbation, were built up by compu- 

 tation over the entire earth, and they form 

 the systems of magnetic component fields 

 to be accounted for by superposed mag- 

 netic systems, or by systems of induced 

 currents of electric ions in motion. The 

 formulas pertaining to this subject, espe- 

 cially from 0. Heaviside's papers, were 

 assembled in 1900 for use by meteorolo- 

 gists. The beautiful lectures on the the- 

 ory of electrons, by H. A. Lorentz, 1909, 

 will guide the reader much more fully in 

 all these interrelated subjects. My 36-inch 

 model of the diurnal magnetic deflecting 

 field exhibited a very complex series of 

 vectors of three types, which have the spe- 

 cial characteristic that they are intimately 

 associated with the diurnal convection of 

 the atmosphere, and the temperatures, va- 

 por pressures, barometric pressures and 

 wind velocities. Now, this diurnal convec- 

 tion is confined to a stratum only one or 

 two miles thick, and therefore the magnetic 

 vectors must be caused by the movement of 

 the electric ions in the lower and not in 

 the higher strata of the atmosphere, so that 

 this magnetic field can be caused only indi- 

 rectly by the incoming radiation, that is, 

 by the motion of the ions in low-level con- 

 vection currents. It is quite probable that 

 this lower field of electricity comes from 

 the radioactive processes in the ground, but 

 the stream lines of + ions in the lower 

 strata can readily be drawn from an in- 



spection of the magnetic vectors, whatever 

 their physical origin. Eliminating the 

 normal and the diurnal magnetic fields, 

 there remains the second or meridional 

 field which consists of vectors nearly per- 

 pendicular to the normal lines of magnetic 

 force around the earth and lying closely in 

 the planes of the magnetic meridians. 

 These vectors form the picture of a per- 

 meable shell lying in an external field of 

 magnetic force, and they slowly surge back 

 and forth as a reversing system every few 

 days, southward and northward, as if the 

 electric currents made circuits in the air 

 along the lines of magnetic force, high over 

 the equator and returning as ground cur- 

 rents through the earth 's outer layers. On 

 collecting such vectors at stations in all 

 parts of the earth in long tables, it was seen 

 that the reversals had a periodic action 

 easy to detect and mark off by dates. A 

 least square solution gave the period 26.68 

 days, and thus associated it with the ob- 

 served synodic rotation at the equator of 

 the sun. At that time the commonly ac- 

 cepted synodic period was assumed to be 

 about 26.00 days, derived from least square 

 solutions on auroras, thunderstorms and 

 magnetic fields, and therefore not in har- 

 mony with current opinions. The recent 

 spectroscopic work of the Mt. Wilson ob- 

 servatory, however, fixes the equatorial 

 period of rotation at the level of the photo- 

 sphere as 26.70 days with a smaller period 

 in higher levels of the sun's atmosphere. 

 We infer that the mixed radiation from the 

 sun should be sorted out by elevations, and 

 by latitudes, as well as by longitudes, in 

 order to complete the solution of this prob- 

 lem. This meridional magnetic field at the 

 earth carries one inference with it. If the 

 normal terrestrial magnetic field is equiva- 

 lent to a system of east-west electric cur- 

 rents, then this disturbing field can not be 

 accounted for by strengthening and weak- 



