le the funnel is still to be seen.

We thus learn how the sea is curved at every part, and therefore it is
natural to suppose that the earth is a sphere. When we make more careful
measurements we find that the globe is not perfectly round. It is
flattened to some extent at each of the poles. This may be easily
illustrated by an indiarubber ball, which can be compressed on two
opposite sides so as to bulge out at the centre. The earth is similarly
flattened at the poles, and bulged out at the equator. The divergence of
the earth from the truly globular form is, however, not very great, and
would not be noticed without very careful measurements.

The determination of the size of the earth involves operations of no
little delicacy. Very much skill and very much labour have been devoted
to the work, and the dimensions of the earth are known with a high
degree of accuracy, though perhaps not with all the precision that we
may ultimately hope to attain. The scientific importance of an accurate
measurement of the earth can hardly be over-estimated. The radius of the
earth is itself the unit in which many other astronomical magnitudes are
expressed. For example, when observations are made with the view of
finding the distance of the moon, the observations, when discussed and
reduced, tell us that the distance of the moon is equal to fifty-nine
times the equatorial radius of the earth. If we want to find the
distance of the moon in miles, we require to know the number of miles in
the earth's radius.

A level part of the earth's surface having been chosen, a line a few
miles long is measured. This is called the base, and as all the
subsequent measures depend ultimately on the base, it is necessary that
this measurement shall be made with scrupulous accuracy. To measure a
line four or five miles long with such precision as to exclude any
errors greater than a few inches demands the most minute precautions. We
do not now enter upon a description of the operations that are
necessary. It is a most laborious piece of work, and many ponderous
volumes have been devoted to the discussion of the results. But when a
few base lines have been obtained in different places on the earth's
surface, the measuring rods are to be laid aside, and the subsequent
task of the survey of the earth is to be conducted by the measurement of
angles from one station to another and trigonometrical calculations
based thereon. Starting from a base line a few m