686 Proceedings of the Royal Society 
ordinary conditions for testing the temperature of the air. The 
thermometer was placed under a sunshade, similar to that shown 
in fig. 4, hut without louvre box at bottom, or draught tube at top. 
The stem of the thermometer passed through the sunshade, the 
silvered bulb was freely exposed under the shade, while the scale 
projected through the top of the shade for convenience in reading. 
The bulb was thus protected only from the direct rays of the sun, 
and from the greater part of the sky radiation. The fan apparatus 
was used as a standard as in the test with the screens. It must be 
confessed that the result surprised me not a little. That the readings 
of the silvered thermometer would be nearly as low as those of the 
fan thermometer, I quite expected ; but that they should be 
almost always lower, was somewhat astonishing, d.'here can be no 
doubt but that it does read lower, but how much it is very difficult 
to say. I have made many and continuous observations with it in 
different conditions of weather, but it always read lower than the 
thermometer in fan draught, when the silver coating was at all in 
good order, the weather bright, and the slightest air of wind moving. 
I have watched the two thermometers for hours under trying con- 
ditions from calmness and brilliancy of sun, and yet the result was 
always the same. The difficulty of saying how much lower the 
exposed silvered thermometer was than the fan thermometer, arises 
from the extreme sensitiveness of the fine-bulbed silvered thermo- 
meter to rapid changes of temperature. Sometimes it rises higher 
than the other by a fraction of a degree, but only for a very short 
time, and at other times it falls much lower, the changes taking 
place quickly. The thermometer in the fan, on the other hand, 
follows these changes more slowly, and never goes to the same 
extremes. The quickness of this instrument is most interesting, 
and reveals to us a most curious fluctuating state of the temperature 
of our atmosphere. The mercury in it is in a constant state of ebb 
and flow. That these ups and downs really indicate changes in the 
temperature of the air, or rather differences in temperature in 
different parts of the passing air, and are not due either to changes 
in radiation or to variations in the cooling effect of the wind, is I 
think indicated by the fact, that it was always possible, by watching 
the silvered thermometer, to tell whether the thermometer in the 
fan draught was going to rise or to fall, as all its indications 
