40 Prof. Marcet on the Variations of the Temperature 
for example, in my observation of the 20th September, made 
at ten o’clock in the evening in cloudy weather. In this ob- 
servation, the thermometer, at 4 feet from the ground, stood 
at 13°.85, and that at the height of 108 feet, at 13°.80. The 
mean of the first 50 feet was 13°.81, and the mean of the last. 
50, 13°76, presenting, as will be perceived, a scarcely sensible 
variation. It is to be observed that, at the time of this ob- 
servation, the ethrioscope indicated zero, thus shewing that 
the earth received as much heat from the upper strata of the 
atmosphere as it sent thither by means of radiation. The 
same thing happened in my observations of the 21st and 22d 
September, taken at half-past nine in the evening. At this time 
the temperature was uniform throughout the whole length of 
the mast ; the ethrioscope indicating in the first case 0°.5, and 
in the second 1° of coldness in the focal ball. We see, there- 
fore, that when the sky is overcast, the different beds of the 
atmosphere may sometimes be found at the same temperature 
some hours after sunset. But I repeat that I have always 
found a sensible difference, at least during fine weather, when 
the observation was made at the moment of setting, always 
excepting in the case of violent winds, particularly if coming 
from the north-east: in the latter case, I have oftener than 
once found a uniform temperature throughout the whole length 
of the mast, even when the sky was not very much obscured. 
2p Question.—To determine the times of the day at which the in- 
crease of temperature as we ascend commences. Is it constant, 
or has it a tendency to augment during the night ? 
The result obtained by my observations on this subject is, 
that, when the weather is clear and calm, the increase of tem- 
perature begins to be perceptible about half an hour or an 
hour before sunset. If the weather be cloudy, it most fre- 
quently does not become sensible but at the moment of set- 
ting. When the sky is clear, the increase usually attains its 
maximum at the time of setting or a little after it; from this 
time it remains nearly stationary, when the weather does not 
vary. If, however, an abundant dew fall, I have remarked 
that the difference of temperature has a tendency sometimes 
to diminish. It was thus, for example, that, on the 17th of 
