11 
Distribution Heat over the Globe. 
not a theory : It is an essay essentially different from those we 
have quoted, and, as its learned author calls it, a determination 
of the mean heat found empirically by the application of co- 
efficients furnished by observation. The method of Mayer is 
analogous to that which Astronomers pursue with so much suc- 
cess, w hen they correct by small steps the mean place of a planet, 
by means of the inequalities of its motion : It does not present 
the result of the solar action disengaged from the influence of fo- 
reign circumstances ; but, on the contrary, it estimates the tem- 
peratures such as they are distributed over the globe, w'hatever 
be the cause of that distribution. The mean heat of two places 
situated under different latitudes being given, we find by a simple 
equation the temperature of every other parallel The calcula- 
tions of Mayer, according to which the temperatures decrease from 
the equator to the poles, as the squares of the sines of the latitudes, 
give results sufficiently precise, when the place does not differ 
much in, longitude from that of the regions where the empiri- 
cal co-efficients have been obtained. But, even in the northern 
hemisphere, wffien we apply the formula to places situated 70° 
or 80° to the east or west of the meridian of Paris, the calcula- 
ted results no longer agree with observation. The curve which 
passes through those points whose temperature is 32®, does not 
coincide with any terrestrial parallel. If, in the Scandinavian 
Peninsula, we meet with this curve under the 65th or 68th de- 
* The formula given by Mayer was T rr: 24 cos Lat. ; orT = 12-f-12 cos 
2 Lat. for Reaumur’s scale; and T rz; 84 — 32 sin 2 Lat., or T rr 58 + 26 cos 2 Lat. 
for Fahrenheit’s scale. Since the publication of Humboldt’s memoir, M. Daubuisson 
has resumed the subject of the earth’s temperature in his Traite de Geognosie, tom. i. 
p. 424. Paris, 1819. He gives the following formula, which is almost the same as that 
of Mayer, for finding the mean temperature, according to the Centigrade scale, viz. 
T = 27® cos 2 Latitude. This formula, which is superior in accuracy to Mayer’s, gives 
all the temperatures in defect for latitudes below 42°, and in excess for all the higher 
latitudes, as appears from Daubuisson’s table. It is therefore obviously defective. 
M. Daubuisson, however, considers it as applicable principally between the paral- 
lels of 30° and 60° of N. lat. It ought to be remarked, that in the above formula, 
27° has been assumed as the mean temperature of the equator, in order to make the 
results agree with observations made in the temperate regions, whereas the mean 
temperature of the equator, as ascertained by' Humboldt, is 27°.5 ; and if this were 
used in Daubuisson’s formula, it would make the differences still more in excess— 
Ed. 
