Gradation of Heat in the Aimosphere. 27 
mond, we shall obtain results quite erroneous and unsatisfactory. 
The decrease of heat in the atmosphere, as determined by the 
ascent of balloons, seems to follow a slower rate than in the case of 
altitudes on the earth’s surface. There can be no doubt that this 
manner of experimenting is free from many causes of irregularity 
to which terrestrial observations are subject. We might therefore 
hope that by this means much light would be thrown on the gra- 
dation of heat in the atmosphere; but a sufficient number of 
accurate experiments are wanting to establish a conclusion in 
which confidence can be placed. In the case of the ascent of 
Gay-Lussae, we obtain a rate of nearly 95 fathoms to a centesi- 
mal degree, which is not extremely different from the mean found 
by terrestrial altitudes. 
If we could abstract from the many and powerful causes by 
which the natural and regular propagation of heat in the atmo- 
sphere is continually disturbed, there is no doubt that the tem- 
perature would decrease nearly in the same proportion that the 
height increases. But this must not be understood in a sense 
strictly literal and mathematical. If we conceive the height of 
a column of air to be divided into portions corresponding to the 
same given difference of temperature, it is much. more probable 
that these portions will form a progression inereasing or de- 
creasing slowly, than that they will constitute a series of perfectly 
equal increments. Such however are the anomalies attending 
-observations of the temperature of the atmosphere, that it is ex- 
tremely difficult to determine by experiment, whether the heat 
decreases in a less or greater ratio than the height increases. 
Accordingly the opinions of philosophers on this point are divided, 
The late Professor Playfair, in his Outlines of Natural Philosophy, 
supposes that the increment of altitude necessary for depressing 
the thermometer one degree, is a quantity continually increasing 
as we ascend higher. But the contrary opinion, that the heat 
decreases more rapidly than the height increases, is more ge- 
nerally prevalent; and it seems to deserve the preference, be- 
cause it is adopted by those philosophers, such as Humboldt and 
Ramond, who have more particularly directed their attention to 
this research. 
We may draw from the theory of the astronomical refractions 
an argument in favour of the conclusion, that the heat of the at- 
mosphere decreases in a greater ratio than the height increases. 
It has hitherto been found to be impossible to reconcile the hori- 
zontal refraction of the stars with the actual rate of the decrease 
of heat in the atmosphere. If we adopt the law of a uniform 
decrease of temperature, and take the rate at 90 fathoms to @ 
centesimal degree as found by experiment, the horizontal refrac- 
tion thence determined will exceed the true quantity by about a 
D2 minute 
