144 Transactions of the Royal Society of South Africa. 
noon and midnight, and seldom at VI or XYIII. The observations made 
use of are those for the day upon which perigee occurred, together with 
the day before and the day after, 864 days in all, necessitating con- 
siderably over 20,000 simple subtraction sums. The labour is almost 
prohibitive. 
In the table on p. 146, column 1 indicates the hours of the lunar day^ 
U.M.P. being lunar noon, the mean of the first and twenty-fifth being 
accounted lunar midnight, or L.M.P. 
Column 2 gives the hourly means of the deviations F — v -f 10 for all 
the perigee periods as defined above. 
Column 3 gives the means of the deviations when the moon near perigee 
culminates between X and XIV. These we call noon " culminations. 
Column 4 the same for the four hours XXII to II. These we call 
" midnight " culminations. 
Column 5 the same for the remaining hours. These we call "horizon" 
culminations, since the horizon is nearly the sun's mean position for these 
hours. 
All the tabular numbers have been smoothed in threes by Bloxam's 
process in order to straighten out minor asperities. Such smoothing is 
perhaps open to criticism as unnecessary, and, at any rate, it diminishes 
somewhat the range of the velocity deviations attributable to the moon. 
According to column 2 there is one principal maximum and one principal 
minimum of velocity deviation, the former about XXIII, and the latter near 
moonrise, with a range of 0-235 mile an hour. Secondary maxima occur at 
IX and XVI ; secondary minima at XII and XX. 
It is interesting to compare this perigee range and the range given in the 
previous paper (which we may regard as the range for the moon's mean dis- 
tance) with the moon's least and mean distance from the earth. Calling 
the mean distance unity, the least distance is about 0*933, so that the ratio of 
the cubes of the distances is as unity is to 0 -81. Now the ratio of the ranges 
of wind velocity for mean and for perigee distances is as unity is to 1*175. 
or, say, 0*85 to 1 : that is, the ratio of the ranges of velocity is very nearly 
the same as the inverse ratio of the cubes of the moon's distances. And it 
would be still more nearly the same if we were to take the moon's mean 
distance for the three days about perigee. 
Column 3, derived from 226 noon perigees, and column 4, derived from 269 
midnight ones, give two remarkable curves. With minor fluctuations the 
former has a minimum at IV and a maximum at XVI, with a range of 0*7 mile 
an hour ; whereas the latter has definite double maxima and minima (like a 
diurnal atmospheric pressure curve) at X and L.M.P. and at VI and XVI, with 
a whole range of 0*5 mile an hour. Column 5 gives a relatively featureless 
curve, excepting that it vibrates practically in opposite phases to that of 
column 3, with a range of 0*25 mile an hour. These are large ranges, but 
