FROM ORDINARY DAYS OF THE ELEVEN YEARS 1890 TO 1900. 
263 
conductivity may be due to contributions from some, perhaps many, previous days. The 
solar influences may be of different kinds, taking different times to travel from the sun 
and decaying at different rates. Sunspots again maybe evidence of some effect on the sun 
which is shared by the solar system, but which takes some time to travel the distance 
separating the earth from the sun. The diurnal inequality can be derived only from 
a combination of days. Thus it does not enable us to compare magnetic conditions 
and sunspot frequency on individual days. There is even a difficulty in comparing 
the run of magnetic conditions and sunspots during successive months, owing to the 
annual variation in the daily range. This difficulty can, however, be fairly sur¬ 
mounted if we express the range for each month as a percentage of the mean range 
from all months of the same name in the 11-year period. The percentage values thus 
obtained for R and R ; in H appear in Table XLIII. along with the corresponding 
Wolfeu’s sunspot frequency, the latter in heavy type. 
The general tendency for the percentage figures in Table XLIII. to be large in 
years of many, and small in years of few sunspots, is of course obvious. But when 
we compare successive months, we see that with rise of S we may have rise or fall of 
R, and that R and R' not infrequently change in opposite directions. There being 
132 months, there are 131 passages from one month to the next. If we allow l/2 in 
cases where the value is the same in two consecutive months, we find that R and R' 
changed in the same direction in 84 cases, S and R changed in the same direction in 
65|- cases, while S and R/ changed in the same direction in 7'2\ cases. A large 
monthly value for R' may be due to only two or three highly disturbed days, excluded 
from the ordinary days, still the number of cases in which R and R/ changed in 
opposite direction is larger than would have been expected. The figures quoted above, 
by themselves, afford no evidence of a connection between S and R in individual 
months, and only slight evidence of a connection between S and R'. This differs from 
what was observed in the case of D. There R and S increased or diminished together 
in 75 cases out of 131, while R r and S increased or diminished together only in 
68 cases. 
In many instances the values of S for consecutive months differ so little that 
accident might play a considerable part. If we confine ourselves to the 52 cases in 
which S changed by at least 10 units, there was agreement in the direction of change 
♦ 
in 30 cases as between R and S, 
,, 28 ,, ,, R r ,, S, 
,, 34 ,, ,, R „ R. 
This is decidedly more favourable to a connection between R and S in individual 
months. 
In a second investigation the months of each year were arranged in two groups of 
6 , consisting respectively of the months of largest and least sunspot frequency. The 
