934 DOVE ON THE NON-PERIODIC VARIATIONS IN THE 
the “second winters” of 1807 and 1808 reached their maximum 
near the Alps. If we consider more closely the march of the 
differences in the tables, we see that the amount of deviation is 
greatest at one particular spot, from which it decreases in every 
direction. If, for instance, we find the maximum of heat or cold 
in the middle part of Germany, we may expect smaller differ- 
ences in the N. S. E. and W.; but if we pass certain limits we 
find great deviations in the opposite sense. We find however 
that the decrease in the amount of deviation does not take place 
with the same rapidity in all directions, it proceeds slower in the 
north and south than in the east and west direction. 
Fourthly : “ Similar relations of the weather are more frequent 
from south to north than from east to west.” In cases of very 
great deviations we have in the latter direction a double oppo- 
sition between Europe on the one hand and America and Asia 
on the other. The temperatures of the winter of 1821-22, and 
of the January of 1834, in Europe, were probably so high in 
consequence of the severe cold in Asia and America. In De- 
cember 1829 the maximum of cold was at Berlin; it was still 
very visible at Kasan; but in Irkutzk it was mild, while America 
enjoyed very warm weather. But generally Europe partakes of 
the same distribution with one of its neighbours. Sometimes 
Europe, Asia and America have all the same kind of weather, 
and in this case the opposites are found either in the north or 
the south. 
If these phenomena are due to conflicting and alternately 
prevailing currents, it is evident that we may determine the 
breadth of these currents ; if the chain of observations be extended 
over a larger space than the currents occupy, we shall find some- 
where or other the opposite extremes. Egede Saabye says of 
Greenland, “The Danes have remarked that when there is a 
severe winter in Denmark there is a mild one in Greenland, and 
vice versa.” The same applies to Iceland; and indeed so much 
so, that the transport of goods thither from Denmark partly de- 
pends upon it. If this distinction were one of north and south 
there could not be so considerable a disagreement between Pe- 
tersburg and Reikiavig, as is shown in the following tables :— 
