74 DR. FARADAY’S EXPERIMENTAL RESEARCHES IN ELECTRICITY. (SERIES XXVI.) 
not be restrained above, as in the case formerly supposed (2922.), but could gather 
in from space freely. From what has been said, however, it will be evident that 
such a conclusion can only be drawn with any degree of confidence from observa- 
tions made pretty equally over both hemispheres. 
2950. If we should ever attain a good knowledge of the annual variation for several 
stations in different parts of both hemispheres, it would help to give data by which 
the depth at which the magnetic power is virtually situated might be estimated ; for, 
as this power is expected to undergo undulations over very large portions of the 
earth’s surface by the annual changes of temperature (2884.), so they would differ in 
character and extent according as the origin of the lines should prove to be more or 
less deeply situated. 
2951. With regard to the many variations of magnetic force, not periodic or not so 
in relation to the sun, which yet produce the irregular and overruling changes already 
referred to (2945.), dependent, as I suppose, on local variations of the atmosphere, I 
may be allowed to notice briefly such points as have occurred to my mind. 
2952. The varying pressure of the atmosphere, over a given part of the earth’s sur- 
face, ought to cause a variation in the magnetic condition of that part of the earth. 
It is represented to us by a difference of three inches of mercury, or one-tenth of the 
weight of the atmosphere. Now the oxygen in a given space is paramagnetic in 
proportion to its quantity (2780.), and therefore it does not seem possible that that 
quantity over a given space of the earth’s surface, whether it be recognised by volume 
as above, or by weight as in a given volume at the earth’s surface, should be varied 
to the extent of one-tenth of the whole sum without producing a corresponding altera- 
tion in the distribution of the magnetic force ; the lines being drawn together and 
the force made more intense by an increase of the quantity or of the barometric 
pressure, and the reverse effects produced at the occurrence of diminished pressure. 
2953. At any spot which is towards the confines of that space where the air is 
increasing or diminishing in pressure, there will probably occur variations in the 
directions of the lines of force, and these will be more marked at such places as 
happen to be between two others, in one of which atmosphere is accumulating, whilst 
from the other it is retreating. Whether these changes (which I think must occur) 
produce by vicinity effects large enough to become sensible in our magnetic instru- 
ments, is a question to be resolved hereafter. To suggest the cause is useful, because 
to know of the existence, nature, and action of a cause, is important to the arrange- 
ment of the best means of observing and evolving its effects. 
2954. Winds and large currents of air above may often be accompanied by mag- 
netic changes if they endure for a time only. A constant stream like the trade-wind, 
may have a constant effect ; but if, when the arrangement of the lines of magnetic 
force through the atmosphere is in a given state consequent upon the condition of 
the atmosphere at that time, a wind arises which mixes regions of cold and warm 
