1816.) Meteorological Table. 365 
duced more than a few degrees below its mean temperature, yet it 
is obvious that the diminution must continue till the surface again 
approaches the temperature of the interior. ‘This equilibrium will 
take place towards the middle of March, as the surface of the 
earth is generally within a few degrees of the mean annual tempe- 
rature. It is found accordingly, as might be expected, that the 
temperature of the interior as indicated by the pump water is 
actually a minimum about the middle of March. From that period 
it gradually increases, and appears to reach its mean about the 
middle of June. The rain and dews, however, of the succeeding 
months, being still at a comparatively high temperature, communi- 
cate additional caloric to the strata beneath, and must continue to 
do so till the surface of the ground again descend towards the tem- 
perature to which the interior has risen. This point, for the reason 
already mentioned, will be a few degrees above the mean, and of 
course the equilibrium ought to take place about the beginning of 
October, as the temperature of the atmosphere is then generally 
within a few degrees of the mean annual temperature. It is obvious, 
however, that the ground, to the depth of several feet, may, from 
the accumulation of the sun’s rays, be preserved at a higher tempe- 
rature than the mean, even after that of the atmosphere has sunk 
considerably lower. This will happen perhaps to a certain extent 
every year, but especially in warm and dry seasons. Making an 
allowance, therefore, for this circumstance, the equilibrium be- 
tween the surface and the interior may be expected to take place 
about the end of October, which agrees exactly with observation, 
the pump water being then a maximum. From this period the 
temperature decreases, and reaches its medium again towards the 
end of December. Besides the coincidence between the above 
theory and the general results of observation, it is further confirmed 
by several circumstances of a more particular nature. During the 
whole of the month of August (1815), when the rain that fell was 
wholly consumed by evaporation, and of course could contribute 
nothing to springs, the temperature of the well was stationary at 
46°8°. In the course of September, and the first 15 days of Octo- 
ber, when the quantity of rain considerably exceeded the evapora- 
tion, the temperature rose to 47°8°, being 1° in an interval of about 
50 days; and between the 15th and 25th of October, during which 
time there fell upwards of two inches of rain, with little evapora- 
tion, it rose to 48°8°, being 1° in about one-fifth of the preceding 
interval. It may be inferred that, had the last fall of rain taken 
place souner, the pump water would have sooner reached its 
maximum. But might not the truth of the whole theory be put to 
the test of experiment, by noting the temperature and quantity of 
rain, ascertaining as nearly as possible the density and depth of the 
strata through which it passes before being conveyed to the pump, 
and determining by calculation the quantity of calorie which it 
ought to abstract from, or communicate to, these strata? And if 
