694 
Proceedings of the Royal Society 
111 workiag with these differential thermometers, it is not neces- 
sary to screen them from the diffused radiation. If they are simply 
shaded from the sun, it is enough. The principal point to be 
attended to is, that they are placed where there is as free a circula- 
tion of air as possible, and not near any surface on which the air 
might be heated before coming into contact with them. When 
permanently fitted up, the screen required for them should be all 
above the bulbs, and may be of the simplest description sufficient 
to protect them from the weather. Two pieces of wood fixed 
parallel to each other, and with an air space between them, to 
prevent heat descending to the under side of the screen, is quite 
sufficient, and only requires to be securely fixed in a horizontal 
position. The thermometers are placed with their bulbs projecting 
a short distance below the screen, and the scale above it. One of 
the principal things to be aimed at is to secure a free circulation of 
air, and to prevent heated air coming to the thermometers, the screen 
having nothing to do with the diffused radiation, as it is welcomed 
and allowed for by the instruments. 
This differential arrangement may be used for maximum and 
minimum registering thermometers. So far as the experiments go, 
they seem to give truer readings than most forms of screens ; and 
as the air can circulate freely over the bulbs without being heated 
or cooled on louvre boards, the thermometers follow the changes of 
temperature ciuickly. Further, as the maximum error of a clean glass 
bulbed thermometer under those simple screens is scarcely ever 2'', 
and is more generally only about 1°, the amount of correction 
necessary is not great, and can be easily made. 
Night Temperatures. 
But little has been done in observations of night temperatures ; 
no trials of the minimum screen have as yet been made. A few 
trials, however, with the silvered bulb were made on the 19th of 
the month, when the sky was clear and air calm and chilly. The 
experiments were conducted in the following manner : — The two 
thermometers used in the sling experiments, the one with bulb 
silvered and the other clean, were tied firmly together, and prepared 
for sling observations. The thermometers were first slung, and 
their readings taken ; then they were hung up freely exposed to the 
