show the order of magnitudes correctly. If so the effect on N/a
is 0.6 X io"6. For the A2 to A7 I have taken the root of the sum
of the squares of the 2n i normalised coefficients given in the
Report.
It is desirable that these harmonics be better determined: but
I have no plan for this except with the natural growth of data.
Meantime I think that the determination of N is reliable to the
amount stated.
I am equally satisfied that a determination of the integral with
alone, over a limited portion of the whole globe may be very faulty.
Although and A/ are made equal at only 8 specified values of
I have found that the two functions are sensibly equal at the inter
mediate values, for which I have found agreement by computation.
The functions show sensible difference for values between 30° and
50° but that can reasonably be neglected. Examination of the
curves;in figure x, in which I have plotted /.sin 4". M) s™
in blue and red, and of their difference in green, shows that the
latter has negligible values beyond say 250: and one might even be
content with neglect of values beyond 20°.
In this way, by use of Afor one may correctly ignore the
portions of the Earth beyond 20° from the computation point.
200
EARTH-SHAPE AND POTENTIAL
Y
20° 40°' 60° 60° 100° 120° 140° 160°
f sin y
A f sin Y
(f-if)s\r,y
