326 
TERRESTRIAL AND SOLAR RELATIONS 
This magnetic system has constituted a difficult problem for 
solution, as it is necessary to have a simple, world-wide cause 
capable of producing these diverse effects. 
The most prominent fact is. the inversion of vectors as be 
tween the two hemispheres, and it is easy to show that the 
diurnal convection is oppositely directed in reference to the 
normal magnetic field, positive in the southern hemisphere and 
negative in the northern hemisphere. 
In the Tropic zone the air rises nearly vertically by day and 
falls by night; in the Temperate zones it flows toward the poles 
by day, and toward the equator by night, being oppositely 
directed in each hemisphere relative to the positive direction 
of the magnetic field; in the Arctic and Antarctic zones the 
movement is upward by day toward the sun and downward 
at night. These five zones of circulation are marked off from 
each other by the high-pressure belts in latitudes + 30° and — 30° 
and by the low-pressure belts in latitudes + 66° and — 66°. 
The zones of circulation agree with the zones of magnetic vectors 
as defined in 1892. 
Fig. 64 contains a scheme of the circulation vectors (black), 
and the magnetic vectors (dotted), as derived from the two 
sources indicated. There is remarkable agreement so far as 
the observational data extend, and the corresponding portions 
of the circulation adopted by natural inference agree with the 
parts that are known. It is generally true, (l) that the circula 
tion vectors and the magnetic vectors are at right angles to each 
other, and (2) that the turning points in both systems coincide 
in all parts of the five zones. The conclusion is almost imperative 
that the circulating vectors, through the generated ions in 
streams, induce the observed magnetic deflecting vectors. 
While there is much to be done by observations fully to verify 
this theory, it is clear that the main features of both the systems 
are in remarkable conformity to the known facts of the observa 
tions. The horizontal and vertical components of the two sets 
of vectors in Fig. 64 should be united in one set of spacial vectors, 
in order that this system may be properly comprehended. The 
evidence is very strong that the magnetic variations depend