SUNSPOTS AND BRITISH WEATHER. 
245 
readings taken in the winter. This, however, is not apparently 
the case, the differences which do occur seemingly following no 
rule. The summer temperature of 1867 was decidedly low ; but 
that of 1868 was decidedly high; from 1868 until 1873 the 
summers, with slight oscillations, became colder and colder ; 
from 1873 to 1875 they became slightly warmer again; but, 
with 1876 a fall set in, and the summer of 1877 was the coldest 
we have experienced for many years past. 
The winter means display equally inexplicable irregulari- 
ties. The years 1868, 1869, 1872, 1874, and 1877, enjoying 
more or less mild winters, and the years 1867, 1870, 1873, and 
1876, experiencing, more or less, severe weather. In the years 
1867 and 1877, which are both years of Sunspot minimum, it 
will be noticed great differences exist in the winters ; that of 
1867 is the coldest of the whole period, whilst that of 1877 is 
among the warmer ones. The following curves (fig. 3) give the 
actual mean temperatures of the summers and winters of this 
period. The first curve with the scale on the left gives the 
winter, and the second curve with the scale on the right the 
summer means. 
In conclusion, it has been attempted to connect the electrical 
phenomena of our atmosphere with solar physics ; but with the 
same unfortunate result. The average number of days on which 
thunderstorms occurred during each summer of the period was 
forty-nine ; and the following table shows, for each year, the 
departure from that mean : — 
In 1867 
. 
— 5 days. 
1868 
+ 4 „ 
1869 
- 5 „ 
1870 
- 16 „ 
1871 
+ 12 „ 
1872 
+ 20 „ 
1873 
- 6 „ 
1874 
o „ 
1875 
+ 5 „ 
1876 
- 6 „ 
1877 
- 3 „ 
These figures, like the former, give no evidence of any vari- 
ation, either with or against the number of Sunspots. Both 
1867 and 1877, the years of Sunspot minimum, show a slight 
deficit in the number of days with thunder ; but on the other 
hand, 1870, which is the year of maximum Sunspots, shows a 
much more considerable falling off, while 1871 and 1872 have 
a very large excess. 
With the whole mass of evidence before us then, it becomes 
almost imperative to confess that, as practical guides in fore- 
