440 FIRST REPORT UPON MAGNETIC WORK

could be occupied in the same period of time. This is made clear in
the next paragraphs.

Aside from the actual error of observation at the time of observa-
tion, in the determination of a magnetic element, declination for
example, we have in addition the error introduced in reducing the
observation to the epoch selected. A part of the observation error
of the declination is entirely astronomical, i. e., due to an error in the
determination of the true meridian. Throughout the survey the
meridian was determined by means of alt-azimuth observations on the
sun, using the alt-azimuth attachment belonging to Magnetometer
No. 18. The sun was observed, on the average, at about two hours
before or after noon. Occasionally it happened, owing possibly to
the sun having been obscured previously, or on account of lack of
time, that the solar observations had to be made, if made at all, within
an hour of noon. When possible, such times were invariably avoided.
If the instrument was in good adjustment and carefully leveled, and
the solar observations were made so as to eliminate, as far as possible,
any outstanding defects in the adjustment of the instrument, then the
error in the determination of the true meridian consisted of the error
of pointing on the sun and the error due to an imperfect value of the
latitude assumed in the formula of computation. ¹ Now the pointing
error, by suitably arranging the observations, can be reduced so that
the error therefrom will fall within the reading error of the vertical
and horizontal circles attached to the instrument used. If these
circles read by opposite verniers to the nearest minute of are, so that
half-minutes can be estimated, the error of pointing can be reduced
so as not to exceed one minute, which degree of accuracy must at
present suffice for magnetic work. The latitude error is a function
of the hour angle, and is of contrary sign for observations made before
and after noon. In order to eliminate the latitude error it is neces-
sary either to observe at such an hour angle that the error will fall
within the limit of accuracy, or to make observations in the morning
and in the afternoon at about the same hour angle.

¹ The present maps of Maryland, with the exception of certain portions,
cannot be assumed as furnishing sufficiently accurate values of the geo-
graphical co-ordinates.