Horary Oscillations of the Barometer. 191 
are immediately replaced by the cold particles from above; and 
by this circulation the diffusion of heat is very rapid. But the 
exchange of particles between the upper and lower strata must 
occupy some time, however small the interval, and the consequence 
taust be that the barometer will measure by its fall the amount of 
the inequality. So on the other hand, in the process of cooling, in 
the absence of the sun, experiment has shewn that the lower strata 
of the air become more rapidly affected by radiation than the 
upper, and the total increase of weight from this cause, will be 
shewn by the rise of the mercurial column. 
If we trace this effect along any given meridian, we shall become 
sensible of the manner in which this influence operates. Beginning 
at the equator, the only circumstance which we have to appreciate 
is the irregularity of the lateral expansion or contraction. As the 
earth acquires warmth from the sun, the barometer falls; but the 
check which the incoming current from the poles sustains, must be 
felt along the whole line of its course; and its due velocity being 
opposed, without any adequate compensation in the upper current, 
the barometer would have a tendency to rise at all latitudes be- 
tween the equator and the pole. Assuming an intermediate 
station upon the same meridian, we should have the same effect 
produced by the unequal expansion of the lower current of the 
atmosphere, but opposed now by the impulse communicated from 
the equator. The fall of the barometer would only then represent 
the balance of the two effects, and must be less than at the equator. 
The further we proceed towards the pole, the more must this re- 
yulsive action accumulate, and the less must the balance of the 
two become, till at some neutral point they are exactly equal. 
Beyond this point, again, the former action would exceed the latter, 
and the barometer would rise in the higher latitudes, while it was 
falling in the lower. 
The results of the preceding table obviously coincide with such a 
gradual progress towards a neutral point: but up to the time when 
I published an essay upon this subject, there were no experiments 
to prove the corresponding opposite effect beyond this limit. By 
‘a careful examination, with this view, of the meteorological register 
P2 
