17 
Distribution of Heat over the Globe. 
tions of different months, do not exceed 1°.8, and they are very 
regularly positive or negative, according to the order of the sea- 
sons. M. Arago ^ has examined for seven years the observations 
of noon. They give for Paris 5° A more than the mean tempera- 
ture of the whole year. Upon high mountains in the tempe- 
rate zone, the difference is scarcely *|*. By the application 
of coefficients, variable according to the latitude and the eleva- 
tion, we may deduce the true mean temperatures from observa- 
tions made at any particular period of the day, nearly in the 
same manner as we can ascertain the latitude of a place from 
altitudes of the sun, taken out of the meridian. 
If we do not stop at two observations of the maximum and 
minimum, but add a third observation, we commit an error more 
or less serious, if we divide simply by three the sum of the ob- 
servations, without attending to the duration of the partial tem- 
peratures, and to the place which the third observation occupies 
between the last terms of the series J.' Experience proves, that 
the mean temperatures of the year, obtained by two or three ob- 
servations, do not differ sensibly, if the intermediate observation 
is sufficiently distant (four or five hours) from the observation 
of the maximum and minimum. Whenever, therefore, we do 
not take into account the duration of the intermediate tem- 
peratures, we should prefer the two observations of the ex- 
treme temperature, which is the method most generally adopted. 
We shall content ourselves with pointing out the errors to which 
* The mean of the observations at noon at Paris 'was 56°.84' ; at Clermont 
in Auvergne (elevation 1348 feet), 56°,30; at Strasburg (elevation 453 feet), 
55°.22. Bullttin de la Soc. Pkilom. 1814, Oct. p. 95. — H. 
*|- At the Hospice of St Gothard. Ephem. Soc. PaU 1785, p. 47, 
$ Example . — Gn the 13th June, at 4^ in the morning, 46°.4 ; at 2^ in the af- 
ternoon, 5 5°. 4; and at 11^ in the evening, 50°, (erroneously 46°. 4, or 8° Centig. 
in the original). In calculating by the duration, we have 
50°,9 the mean for 10^ of interval, = 509°.0 
52.7 9 =474.3 
48.2 5 = 241.0 
The true mean of which is 51°.0. The mean of the three observations, as com- 
monly taken, is 50°. 6. If we stop at the two extreme temperatures, we shall have 
for their half sum 50°. 45 H. 
VOL, IIT. NO. 5. JULY 18^0. 
B 
