412 
DR. T. B. ROBINSON ON THE REDUCTION OE ANEMOGRAMS. 
maxima in each octant do not belong to the same years ; while the amount of discord- 
ance is so great as to almost exclude the idea of any law. I looked for one in the 
direction already noticed. In 1860 the sun-spots were at a maximum, in 1856 at a 
minimum ; and if they exert any influence it must have been considerably less in 1857 
than in 1860. The products of velocity and time were accordingly examined in these 
years, and that for 1860 is 4167 greater than in the other. But this result is reversed 
by 1863, which exceeds 1860 by a still greater amount, 5223 ; and evidently many 
decennial periods must be examined before any reliable conclusion can be attained as to 
this influence. 
The same lawless irregularity may be observed in the maximum velocities of separate 
years. The highest in the period before us is 71 miles in November 1861, the lowest 
19, in July 1860. Far higher velocities than these are sometimes attained, but only 
for a few minutes. It holds also as to the number of hours when the velocity exceeds 
twenty-five miles. As instances : in January 1863 this number is 146, in 1857 it is 23 ; 
in April 1859 it is 40, in 1861 it is 0 ; in November 1861 it is 90, in 1860 it is 5. It 
occurs also, though not intensely, in the hours of calm. It may have some interest to 
give the mean velocity for each month irrespective of the direction. 
Table II. 
Month. 
Velocity. 
Total miles. 
January 
13-51 
70336 
February 
12-82 
60422 
March 
13-00 
67691 
April 
11-62 
51587 
May 
7-78 
39664 
June 
4-24 
35353 
July 
6-59 
34343 
August 
7-29 
35986 
September 
8-02 
39513 
October 
9-12 
45568 
November 
9*97 
47671 
December 
12-98 
166498 
Here also there seems little indication of equinoctial gales. March is a trifle more 
windy than February, but September less so than October. The yearly sums also 
do not show any special relation to the solar spots; the total in 1857 = 79865; 
in 1860 = 73067 ; but in 1863=95583. The total miles in the seven years=590672, 
and the mean velocity during that time is 9*729. 
II. The most obvious way of dealing with the west and south components of V is to 
derive from them interpolation formulae for each year involving periodic functions of 
the time, and deduce from the coefficients of these formulae in successive years some 
general law. This, however, seems impracticable, for the components differ so widely 
in successive years as to preclude any hope of reconciling them. As a specimen of this 
discordance I give the values for the first hour of the series for January 1 : — 
