446 FIRST REPORT UPON MAGNETIC WORK

the other hand may be assumed as practically the same at the same
instant of absolute time over the whole of such an area as that of
Maryland, for example. And so may the secular variation for a
period of time covering a few years be taken as practically the same
over the whole of Maryland. At present the westerly declination is
increasing at the rate of about 3' per year.

The magnetic survey of Maryland was fortunate in that the Wash-
ington Magnetic Observatory was so close by. Unfortunately, how-
ever, the sphere of usefulness of the latter in this respect was some-
what diminished partly by the fact of the electric car disturbance
which affects the work of this Observatory and partly by the inade-
quacy of the working force of the Observatory.

In consequence it has not been possible as yet to make an accurate
reduction of the observations of the magnetic survey for diurnal
variation and disturbance variation. The data called for that, was
necessary for this purpose the Washington Magnetic Observatory has
up to date failed to supply.

As it was highly essential that this first part be issued this year,
I had to content myself, for the present, with the following method
of procedure. I determined the mean diurnal variation for the
months of the survey from the four years of observation 1887-91
made by the Washington Magnetic Observatory before its unfortu-
nate removal to the new (and disturbed) site. The quantities are
given in a later chapter. This was applied and regarded as the
normal diurnal variation. I next obtained the permission of Com-
mander Phythian, in whose charge the Magnetic Observatory is
placed, to make a personal inspection of the magnetograph traces
during the period of the survey and by a mere superficial inspection
made a list of the more or less disturbed days. The list of magnetic-
ally disturbed days kindly furnished by the Director of the Toronto
Observatory, Professor Stupart, practically agreed with the list pre-
pared from the Washington traces. I next inspected the hourly read-
ings, such as were tabulated, and formed the hourly means where this
was possible. I assumed that these hourly readings for any par-
ticular month would give practically the same diurnal variation as