232 DOVE ON THE NON-PERIODIC VARIATIONS IN THE 
considerably and continually (and always on the same side) from 
those of adjoining places, we may assume that there have been 
errors in the observations; such errors may be caused by a 
change of instruments, a change of their situation, or of the time 
of observing. From these tables then the following conclusions 
may be drawn :— 
First: “The tropical atmosphere of the Indian basin does 
not appear to exert any visible influence on the weather of Ku- 
rope.” It cannot be denied that the earlier or later setting-in 
of the monsoons in particular years stands in connexion with 
the preceding and following movements of the atmosphere; but 
from the observations at Madras we are not able to trace any 
direct connexion with the thermal relations. 
Secondly: “The temperature of the zone of the trade winds 
in the Atlantic Ocean is in evident connexion with the weather 
of the neighbouring temperate zone.” As proof of this may be 
adduced the warm winter of 1827-28, and the low temperature 
of 1829. The uncommonly high temperature of January and 
February, 1828, in Havanna, is still seen very plainly at Marietta, 
and it is also very evident at Salem, Boston, and Concord, up 
to Montreal, and is also visible in the observations at New Haven, 
N. Bedford, Cambridge, Clinton, Lowville, Washington, Hud- 
son, Albany, Middleburg, Onondaga, Auburn, Cherry Valley, 
and Canandaigua; and it is remarkable that in the more south- 
ern localities, January was relatively the warmest ;—in the more 
northern on the contrary, February. The extreme cold in Eu- 
rope from November 1829 to February 1830, first makes its 
appearance at Havanna in January 1830, and remains there until 
April, at which time a milder weather has commenced in the 
temperate zone. 
Thirdly: ‘The cold of winter generally extends itself from 
north to south, but an unusual warmth moves in an opposite 
direction. By north and south, we mean these directions as 
modified by the rotation of the earth.” Regarding the more 
accurate direction in which an unusual extreme extends itself; 
this will depend in the first place upon the spot where it first 
appears in its greatest power, and upon the relative position of 
the places of observation to this spot. In January 1814 this 
locality appears to have been situated to the north of Sweden, 
for the January was coldest in Stockholm, Torneo, Carlisle, 
Dunfermline, and London; these places being named in the 
