64 
MR. A. W. RtCKER AXD DR. T. B. THORPE ON A MAGNETIC 
in practice in the field this method is troublesome. It is difficult to see the image 
unless the mirror is very bright, and there is fairly strong sunshine. On days when 
the weather and sunshine are uncertain it is very annoying to have to spend time 
when the sun is visible in making preliminary adjustments. From and after our visit 
to Stornoway in 1884, fie., after observations had been made at about a dozen stations, 
we adopted a different plan. The mirror was frequently adjusted either indoors in 
accordance with the directions of the Admiralty ‘ Manual,’ or by means of some 
elevated object in the field. In the latter case the error of parallelism was first got 
rid of by observing the object when the observer’s back was turned towards it, and 
the mirror was nearly vertical. The mii-orr was adjusted until the image appeared in 
the same position as before when it was reversed in its bearings. The object was then 
viewed by reflection when the instrument was turned through 180°, fie., when the 
mirror was nearly horizontal and the axis of the mirror was made perpendicular to the 
axis of the telescope. By repeating these processes twice or thrice a perfect adjust¬ 
ment was obtained, and it was found that under ordinary circumstances the instrument 
could be carried about from place to place for some time without this adjustment being 
seriously affected. To make certain, however, that all was right, we always (when 
possible) took observations of the sun both in the “front” and “back” positions, i.e., 
with the mirror approximately horizontal and vertical and reversed the mirror in its 
bearings in both positions. By the latter precaution the error of parallelism (if any 
existed) was eliminated, and the observation thus afforded data for the calculation of 
the error of collimation, and for the corrections which in consequence of this error 
must be applied to the “ front ” observations. 
By this method the actual adjustment of the mirror could be made at any time or 
place. If the time at our disposal was short, or the weather uncei'tain, the instrument 
was regarded as ready for use as soon as it was set up in the field, and thus no time 
was lost, but the observations were so taken that any error of adjustment could be 
calculated and allowed for. Lastly the agreement of the measurements made in the 
front and back positions gives a valuable test both of the adjustments and observations. 
In the case of the observations made in 1888 with Magnetometer No. 60, the 
method of correcting the mirror by the image of the cross wires was again resorted to. 
Some modifications of the instrument which we have introduced [loc. cit.) now make 
this method practicable, and it was found to work well under the new conditions. 
Diurnal Variation and Disturbance. 
In the survey of 1857 corrections for diurnal variation and disturbance were applied to 
the declination observations only. These two quantities are intimately connected. The 
total deflections of the continuously recording instruments at any instant are the sums 
of the normal diurnal variations and the disturbances. If the latter are assumed to 
occur simultaneously and in equal intensity over an area such as that of the United 
