ATMOSPHERIC MAGNETISM — ANNUAL VARIATION. 
55 
with the whole system of magnetic lines. But as the magnetic poles do not coincide 
with those of the earth, or with what may be called the poles of the changing tem- 
peratures, so a cause of difference in direction will here arise. 
2886. Again, it may be that as oxygen is cooled, its paramagnetic power may in- 
crease in a more rapid proportion than that of the change of temperature, so that the 
chief alteration of the disposition of the earth’s force may be in the extreme northern 
and southern parts ; and in combination with the holding power of the earth (2907.) 
may even cause a change the reverse of that expected above in lower latitudes. If 
in our winter the lines of force were to close together in the polar parts and to open 
out in lower latitudes, the balance of magnetic force would just as well be sustained 
as if all the lines in our hemisphere were to be compressed and strengthened, and be 
compensated for by a corresponding change in the south. In the former case, each 
hemisphere would balance its own forces, in the latter they would be balanced 
against each other. There can, I think, be no doubt, that as far as the mass of the 
earth and the space above our atmosphere are unchangeable in relation to annual 
and diurnal variation, so far they would tend to restrain any variation which might 
depend only on the varying temperature and state of the air ; holding as it w^ere the 
two sides of the variations, the increase and diminution of intensity, or the right and 
left hand in change of direction, nearer together than they otherwise would be. 
2887. Further, if it be supposed that the whole of a hemisphere is affected at once 
in the same direction by change of temperature, it wall not he affected alike, but differ- 
ently in dfferent latitudes, because of the difference in amount of that change. 
2888. The difference of land and water (2880.) will still further break up any ex- 
pected uniformity of the general result, and cause that certain parts of the cooling 
hemisphere shall increase in power more in proportion than other parts ; and when 
these parts lie on opposite sides of the magnetic meridian of any given place, they 
would probably have power to cause an alteration in the declination of the needle at 
that place. 
2889. As the annual changes of temperature are less at the equator than in parts 
more north or south, so there, probably, little or no annual variation would occur; 
none indeed as regards the varying temperature or expansion of the air, but only that 
portion which is consequent upon the alternate changes of the parts on its opposite 
sides (2884.). 
2890. Another effect, which may be considered as an annual variation, but which 
is connected with the diurnal change, may be expected. As the daily changes in 
temperature of the atmosphere, influential upon a given place in north or south 
medium latitudes, are greater in extent in summer than in winter, so the correspond- 
ing magnetic variations may be expeeted to vary also, being larger in the northern 
hemisphere, when the sun is on the north side of the equator, and less when he 
is present in the southern hemisphere, and producing like correspondent change 
there. 
