422 FIRST REPORT UPON MAGNETIC WORK

say, of including in the survey the observation of the action of the
earth's magnetism in its entirety, not simply to observe one angular
component, e. g., magnetic declination. Regarded from every possible
point of view, from the practical as well as from the purely scientific,
this course recommends itself.

What would we think of the astronomer who endeavored to adduce
celestial laws by observing simply one celestial co-ordinate, e. g., right
ascension? By massing such right ascensions together in a statistical
fashion he could compile certain statistical laws which would undoubt-
edly possess some value, but that he could never get at the real physical
laws governing the phenomena is too apparent to require further
argument. Nay, his statistical deductions might even lead him to
adopt totally erroneous physical theories. Or, suppose the meteor-
ologist should attempt to frame laws for weather prediction by
observing and massing together simply one meteorological element,
such as barometric pressure or temperature? True, this is the very
thing that he must largely do in the present stage of meteorology, for
the prime reason that he has not yet learned how to reach all the
factors that control and shape the phenomena of the weather. He
fully recognizes this and is making an earnest endeavor to get at the
true physical laws by enlarging the scope of his methods of observa-
tion. And so the magnetician must be careful to start in the right
way at the very beginning.

It is true that the practical needs of the land surveyor are largely
satisfied, if not almost entirely, by a magnetic declination survey,
such as for example the one so thoroughly carried out by the State
Geological Survey of New Jersey, but erroneously designated a " mag-
netic survey. " Declination surveys will supply the necessary data,
such as declination, and when repeated, likewise the secular change of
declination, but this data will always remain on the empirical plane,
so long as simply declination data are at hand. We can never hope
to reach the true laws of nature in this way.

The advantage to be derived in the deduction and verification of
secular change data by a combined treatment whenever possible of all
the magnetic elements involved, instead of the hitherto independent