76 _ M. POUILLET ON SOLAR HEAT, 
tain the real temperature of the air, and the relation of the cooling 
resulting from its contact to that resulting from the radiation— 
two data upon which it is impossible not to have some uncer- 
tainty. 
The actinometer is represented in the above figure : it is com- 
posed of four rings of two decimetres diameter, furnished with 
swans’-down, and resting one upon another, so that the down 
may not experience any compression: the swans’ skin itself 
forms the foundation of the circle of each of these rings. This 
system is inclosed in a first cylinder of silver plate c, enveloped 
also in swans’ skin, and contained in a larger cylinder cl. A 
thermometer rests on the centre of the upper down: the mar- 
gin d is of such a height that the thermometer can only face 
two-thirds of the hemisphere of the sky; this margin is per- 
forated with holes at the level of the down, in order that the cold 
air may flow regularly *. 
This apparatus is during the night exposed to the radiation of 
the sky, and its thermometer and an adjoining thermometer 
freely suspended in the air at four feet from the ground are ob- 
served hourly ; the zenithal temperature is deduced from the 
difference of these temperatures, or from the lowering of the 
actinometer; but for this the apparatus must have received a 
graduation, which we proceed to indicate. 
20. If the actinometer had an indefinite surface, and were 
maintained at a constant temperature in vacuo under an hemi- 
spherical inclosure, it would evidently take the temperature of 
the inclosure; on the contrary, with its real form, facing only 
two-thirds of the hemisphere, and enveloped in a stratum of air 
‘which warms it again, it must always remain at a higher tem- 
perature than that of the inclosure. The graduation is intended 
to determine how much it is heated, so that it is sufficient to 
know its temperature and that of the ambient air to deduce 
thence the temperature of the inclosure, with which it is in inter- 
change of radiant heat. It is obvious that there must exist a 
simple relation between the temperature of the inclosure and 
the lowering of the actinometer. To discover this relation, I 
composed an artificial sky with a zinc vessel a metre in dia- 
meter, supported at a height of two metres by three slender 
columns; this vessel, the bottom of which was blackened, was 
filled with a refrigerating mixture at — 20°, and the actinometer 
was placed vertically beneath at such distances that the central 
* See Note 2, -p. 89. 
