113 
mind the limited amount of observations which we have for 
determining inequalities of long periods. 
{Note added on March 6th.) 
I take this opportunity of saying a few words on what I 
imagine to be the proper line of policy that should be pur- 
sued in this research. 
( 1 ) There are manifestly two stages in the investigation. 
In the first place we wish to ascertain whether there is any 
connexion between the state of the sun’s surface (as revealed 
by spots) and the meteorology of the earth, and in the 
second place we wish to find the nature and laws of this 
connexion should it be proved to exist. 
(2) If the various meteorological elements at the various 
stations of the earth are found to present the same periodic 
inequalities as those which characterise sun spots, this 
must be taken as decisive in favour of a connexion of some 
sort between the two, quite irrespective of the exact form 
of the inequalities. Nor will this evidence be invalidated 
if an inequality at one station should be different in form 
from that at another. 
(8) Assuming the probability (from the evidence already 
brought forward) of such a connexion, the most natural 
hypothesis is that which supposes that the sun has inequali- 
ties which affect his radiating power. Hence it is of great 
importance (as proposed by Professor Stokes and others) to 
ascertain by judicious actinometrical experiments whether 
the heating effect of the sun’s rays be in reality variable. 
(4) In absence of actinometrical results we have grounds 
for believing that the magnetic activity of the sun is great- 
est at epochs of maximum sun spots, and it seems most 
natural that the meteorological activity of our luminary 
should be greatest when his magnetical activity is greatest. 
From the reasoning of the paper to which this note is 
added we may conclude that there is no evidence which 
can be deduced from rainfall against this hypothesis. 
