312 
On the Measure of Temperature , [Oct. 
fore, our globe was surrounded with hydrogen, its altitude would be about 14|th. 
times higher than our atmosphere is. 
Hence a means of determining the altitude of the atmospheres of any of the 
celestial bodies; and reci procall y, having the altitudes and the nature of the airs, 
their attractive forces. And hence, too, a proof of the small attractive forces of 
comets, which have been found by other methods, with a means ot computing 
them, at least approxiiuatively. 
5th By (1) reduced to (l) it appears that the velocity of sound, at the surface, 
is independent of the pressure of the atmosphere ; and by (2), that the pressure 
in the higher regions is dependent on this very velocity, and varies with it, being 
greater or less as this is greater or less; this apparent paradox is eaily explained : 
for at the surface, the pressure results fro n the total quantity of air, but at a 
given altitude, from the total quantity, minus that below, which depends on the 
temperature at the surface, and thence on the velocity of sound. 
6th. Our barometric formula (2) requires no aid from temperatures of the ex- 
ternal air. It includes all that is needful within itself, and merely requires that 
the barometers be of one, or reduced to one, temperature. Even this it might 
do without. But as I have elsewhere remarked, Laplace’s formula in this respect 
is singularly helpless ; it not merely affords no means of finding the difference 
of temperatures, but cannot do without it. 
7th. By the help of the formula here given, the time sound takes to travel 
over any given space, oblique as well as horizontal, a problem, I believe, that has 
never been attempted, may easily be determined. For instance, it a be the altitude 
of the generation of sound, b that of the auditor, and q the distance between the 
two, the time in seconds is 
5 */ s . — LL - , 
g (b— a) C 3 \/‘2 3*/23 
( 7 ) 
and the time of travelling vertically from the top to bottom of the atmosphere, or the 
contrary.— *S in which S is the horizontal velocity at the surface. If, therefore, 
S 1089, 4 as we have computed it at 3e° Fahr., this time is 4 ni 4/ s 4. 
I might here observe, by way of conclusion, that should any one teel disposed 
to compute a table from our barometric theorem for the more easy measuring ot 
heights it would be advisable to do it for 52* Fahr. The altitude being taken tor 
this temperature, and multiplied by twice the number of degrees which the tem- 
perature of the lower barometer may be above or beneath 53°, 1 gVo 1 * 1 °* t ie F r0 " 
duct will be the only correction required, and is to be added or substracted to the 
preceding altitude, just as the lower barometer’s temperature happens to exceed or 
^Fo r °the° c e°iit igrade the rmomc te r , the table had better be computed for o temp, or 
tlie freezing point. 
ly On the Measure of Temperature, and the Communication of Heat. 
Bij M. M. Dulong and Petit. 
part ii. 
ON THE LAWS OF COOLING. 
The first received views on the communication of heat are to be found in the 
works of Newton. 1 This great philosopher assumes, « priori , .that a heated uo > 
exposed to a constant cooling process, such as that of a uniform current o r, 
would lose, in each moment of time, a quantity of heat proportional to the c* a. • 
its temperature, over that of the cooling medium ; and that consequently these 
of heat, in equal and successive portions of time, would form a decreasing g 
cal progression. Kraft, and after him Richmann 2 have endeavoured to vert y 
1 Newton Opuscula , t. ii. p. 423- 
3 Nov. Com. Ac. Petrop.t. i. p. 195. 
