DISTRIBUTION OF TEMPERATURE ON THE EARTH. 225 
not arrive at any conclusion as to the distribution of temperature 
at this time ; we cannot tell in which places the temperature was 
too high, and in which too low. But when we know that the 
mean temperature of March during many years was at these 
places respectively — 0°50, 4°10, 4°87, 5°29, 5°60, 4°18, 3°38, 
—4°'54, we see instantly that at Reikiavig in 1836 the tempera- 
ture of March was 2°57 too low, at Boston normal, at London 
a little too high, and that this excess extended through Germany 
up to Petersburg, where it reached its maximum, the tempera- 
ture being there almost 6° too high. Such differences alone can 
be quantitatively compared, and they have the advantage that 
the faults arising trom imperfect instruments are eliminated. 
We must not assume that the true mean monthly tempera- 
tures are obtained from the above observations; but we shall 
make a nearer approach to truth if we compare the means of 
each place drawn from the same period of time. From 1807 to 
1824 we have observations for twenty places, extending from 
Madras to Torneo, and from Salem to Dantzic; for a few places 
some years are wanting. From 1820 to 1830 we have them for 
twenty-three places ; from 1828 to 1834 for thirty-five, &c. &c. 
(Tables here follow in which the observations are arranged in the 
above manner.) 
From these tables we may draw the conclusion “ that there at 
times exist certain causes which have considerable duration, and 
extend over a considerable space, which either raise or depress 
the temperature.” The following table of the yearly means will 
show this more plainly :— 
1807 to 1824, 1820 to 1830. Difference, 
Palerniosss scisssssser eee 13°37 
10:10 
761 
8°41 
Regensburg . 
DRMEEBAPT: sisccesccesssarenvoss 
Carlsruhe 
Berlin 
Torneo 
