687 
of Edinlurgli, Session 1883-84. 
were predicted by the silvered thermometer. The changes in the 
fan thermometer are slow compared with the other, owing to the 
shape of the hulh, and to time being required to alter the temperature 
of the air passages, which radiate heat to the thermometer. So 
nearly as I was able to judge from observations made when the 
temperature was nearly steady, the average reading of the silvered 
bulb was about a quarter of a degree below the fan readings. It 
might sometimes be less when the air was calm. 
The silvered thermometer scarcely rose one degree above the 
standard when hung in sunshine during a fine calm day, while 
the radiation thermometer similarly exposed showed a temperature 
of 131°, or 65° above the temperature of the air. 
As a great deal of radiant heat is absorbed by the glass of the 
bulb of an ordinary thermometer, not only at the surface, but all 
through the body of the glass, it seems probable that part of the 
lower temperature indicated by the small bulbed thermometers is 
probably due to this cause ; the small bulbs having a less thick- 
ness of glass than the large ones, as bulbs with thin walls will 
absorb less radiant heat than those with thick ones, and will also 
conduct the absorbed heat more quickly outwards to be carried away 
by the air. 
I have frequently referred to the constant fluctuations in tlie 
temperature of our atmosphere. Now, to prevent any mistake in 
this matter, it will be as well for me to state more clearly the 
amount of these fluctuations. With a thermometer having a large 
bulb, these changes are little noticed ; but after one becomes ac- 
customed to the careful reading of thermometers with medium-sized 
bulbs, they become very evident. In the fan apparatus the changes 
on certain days would amount to half a degree more or less almost 
every minute, and in the fine-bulbed thermometer they often 
amounted to a degree or more in the same time. These changes 
take place whether the wind circulation is strong or slight. 
It may be as well for us to consider here why the thermometer 
in the fan draught should read higher than the silvered thermo- 
meter. No doubt the silvered bulb will be a little higher than the 
temperature of the air, but why should the fan give a higher 
reading still '? So far as I understand it, part of this higher read- 
ing given by the fan apparatus is due to heat conducted inwards 
