to
a firm bearing and snow is packed around the box. Then something,
usually a skin, is thrown over the snow, partly to prevent any possible
warmth from the sun melting the snow and shifting the bearing of the
box; partly to protect the eyes of the observer from the intense
reflected glare of light from the snow.

The mercury trough of the artificial horizon is placed on top of the
level box, and the mercury, which has been thoroughly warmed in the
igloo, is poured into the trough until it is full. In the case of the
special wooden trough devised and used on the last expedition, it was
possible to bring the surface of the mercury level with the edges of the
trough, thus enabling us to read angles very close to the horizon.

The mercury trough is covered with what is called the roof--a metal
framework carrying two pieces of very accurately ground glass, set
inclined, like the opposite sides of the roof of a house. The object of
this roof is to prevent any slightest breath of wind disturbing the
surface of the mercury and so distorting the sun's image in it, and also
to keep out any fine snow or frost crystals that may be in the
atmosphere. In placing the trough and the roof on the top of the
instrument box, the trough is placed so that its longer diameter will be
directed toward the sun.

A skin is then thrown down on the snow close to the box and north of it,
and the observer lies down flat on his stomach on this, with his head to
the south, and head and sextant close to the artificial horizon. He
rests both elbows on the snow, holding the sextant firmly in both hands,
and moving his head and the instrument until the image or part of the
image of the sun is seen reflected on the surface of the mercury.

The principle on which the latitude of the observer is obtained from the
altitude of the sun at noon is very simple. It is this: that the
latitude of the observer is equal to the distance of the center of the
sun from the zenith, plus the declination of the sun for that day and
hour.

The declination of the sun for any place at any hour may be obtained
from tables prepared for that purpose, which give the declination for
noon of every day on the Greenwich meridian, and the hourly change in
the declination.

Such tables for the months of February, March, April, May, June, and
July, together with the ordinary tables for refraction to minus 10 deg.
Fahrenheit, I had with me on pages torn from the "Nautical Almanac a