2936 DOVE ON THE NON-PERIODIC VARIATIONS IN THE 
great difference in the thermic effect of the same mean direction 
of the wind at places which are far apart. 
If we consider the isothermals as fixed curves, constructed on 
the surface of the earth, we do so erroneously. The lines in 
reality wander between two extremes, which we call the isotheral 
and isochimenal lines; they have moreover various curvatures, 
which in different years are not always the same at the same 
season. These non-periodic changes of form are caused by 
winds which extend over large surfaces, and therefore at certain 
places the change of form in the curves which follow each other 
takes place in one direction, while at other places the direction — 
of change is exactly opposite. Out of the several directions of 
the wind, which are continually varying at any particular spot, 
we may deduce one mean direction which is distributed over a 
large surface, which, if not exactly everywhere identical, at least 
changes uniformly. Within its range the variations from the 
mean take place generally in the same sense, and this range will 
form a “ weather system.” The determination of these systems 
would be the same for Meteorology as that of the isothermals is 
for Climatology. 
The continual movement of the atmosphere prevents the ex- 
istence of a distinct line of separation between two neighbouring 
systems, and the more distinctly separated they have existed at 
one time, with so much the more violence will they afterwards 
modify each other. If in winter severe cold and mild weather 
border on each other in a direction from west to east, the spring 
will appear early in that system in which the winter has been 
mild, at the same time that in the other system there is a tem- 
perature perhaps but little above zero, because all the heat that 
is produced is employed in melting the large masses of ice which 
have been formed. The warm air of the one system will not 
always be able to resist the pressure of the cold air of the other. 
The more rapidly the heat has risen the more sudden will be 
the entrance of the cold air. On this account the spring will 
be unpleasant from frequent changes between warm and rough 
weather. These changes are usually repeated several times with 
diminished force, and at length cease altogether suddenly, for 
then the summer weather has obtained the ascendency. 
As an instance of this may be mentioned the rough spring of 
1835, for remarks on which vide Dove, Pogg. Ann. xxxvi. p. 318.) 
In the tables, which embrace a period of fifty years, we fin 
