690 Proceedings of the Royal Society 
tlie true temperature of the air. The result of the preceding 
experiments seems to indicate that observations taken in this way 
may not be correct, owing to a fundamental error in the construction 
of the apparatus. We have seen that radiant heat is absorbed in 
the thickness of the walls of the bulb, and at the surface of the 
mercury as well as at the surface of the bulb, and this internally 
absorbed heat must be conducted outwards to the surface before it 
can be carried away by the air through which the thermometer is 
rushing in its circular flight. There will evidently therefore be 
a heating effect due to radiation, which it is impossible to keep down 
by the cooling effect of the air, and the only question is, what is the 
amount of this heating ? Following up the same line of experiment 
as was previously employed to prevent the internal absorption of 
heat by the bulb, and to reduce the surface absorption to a 
minimum, two similar thermometers were selected, the bulb of 
one being coated with silver and the other kept clean. These 
thermometers were firmly tied together and slung with as great 
rapidity as was thought safe. The readings of these thermometers 
were then compared with each other, and with the fan arrangement. 
The result of a number of trials was, that when slung in sunshine 
the silvered bulb always read at least one degree lower than the 
clean bulb. Compared with the thermometer in the fan draught, 
the silvered thermometer was a small fraction of a degree higher, 
while the clean bulb was higher by more than 1”. These figures 
of course varied with the heat of the sunshine. 
These experiments show us that the ordinary sling thermometer 
in a bright day is more than V too high ; probably it is 1 to 2° too 
high in bright weather, as we must remember that the fan indica- 
tions are not quite down to the true temperature of the air. These 
experiments also show us that the silvered sling is nearer the truth 
by at least 1° than the thermometers generally used. Other 
experiments made at the same time show that slinging the silvered 
thermometer has no advantage over simply placing it in shade, even 
if there is only a slight air of wind moving ; this proves that it 
takes all the cooling effect of rushing through the air to counteract 
the heating effect of the sun’s rays. Instead of slinging the two 
thermometers, they were hung up under a sunshade on a day when 
a little wind was blowing : the readings then obtained agreed very 
