oS ee 
ON THE MONTHLY ISOTHERMAL LINES OF THE GLOBE. 89 
be occupied entirely by sea: we should in such case have two lines of maxi- 
mum temperature on the mean of the whole year; but not in the separate 
portions of the year ; for the summer heat of the northern zone of land would 
be simultaneous with the winter of the southern zone, while the temperature 
of the equatorial sea would always be intermediate between the two. 
Hence we see how little one is justified in deriving from the distribution 
of the mean annual temperature conclusions respecting its distribution in 
the separate portions of the year: it might even be asserted, on the contrary, 
that the annual isothermal lines become first elucidated by the consideration 
of the monthly isothermals ; and that for this reason all attempts to refer 
their form directly to the configuration of the continents have proved 
unsuccessful. 
If we divide the globe at the meridian of Ferro, and compute the tempera- 
ture of the parallels east and west of that meridian at every ten degrees of 
latitude, we find (with the exception of the latitudes of 70°) the eastern half, 
which has the largest mass of land, colder than the western half, the difference 
diminishing constantly as the equator is approached. 
Within the tropics the diminution of temperature in going northward takes 
place with great regularity. On the eastern side it is represented exactly 
between 0° and 30° by the equation 
t,=21 cos 22, 
t being in degrees of Reaumur, and zx being the latitude; and on the 
western side it is represented very approximately between 0° and 40° by 
t =21.4cos (22—7). 
No formula has been found applicable to all latitudes ; in latitude 30°-40° the 
deviation is always considerable. The reason is easily seen; onthe Ameri- 
can side the Gulf-stream flowing from America to the Azores, and in Asia 
the lofty mountains and table-lands rising from the lowlands of the Ganges, 
cause a sudden break in the progression of temperature. As a general for- 
mula for the equator and the higher latitudes, 
1,=24.5 + 45.5 cos?x 
does best ; for the lower 5 re 
t,==24+45 costa 
is still nearer. According to this the temperature of the pole is 242° of 
Reaumur below the freezing-point. 
For the eastern half of the southern hemisphere the formula suits 
ty =5 + 26°2 cos? (w— 5). 
For the polar regions there remains an uncertainty, which however is of 
less consequence, when the question respects the determination of the mean 
temperature of an entire hemisphere. We obtain an approximate determina- 
tion by calculating the mean temperatures of the zones 0 and 10, 10 and 20, 
and so forth, applying the observational values directly as far as observations 
suffice, and employing for the highest latitudes the value given by interpola- 
tion. Admitting these determinations to be only a first approximation, they 
still appear less uncertain than the wholly arbitrary method hitherto em- 
ployed, of proceeding along a given meridian and deducing therefrom the 
mean temperature of the pole; the values may be improved subsequently 
by combining the temperatures of the eastern and western hemispheres into 
awhole by means of Bessel’s formula, and the form of the function being left 
indeterminate, by the addition of members the observational values will be re- 
produced as nearly as possible. 
