1887.] Mr John Aitken on Thermometer Screens. 
65 
situation exposed to walls and other large surfaces that get highly- 
heated in the sunshine. 
Owing to the lateness of the season, no trials with this screen 
have been made under severe test conditions. In the beginning of 
October it was fitted up on the lawn, and near it was placed a 
horizontal sunshade under which was placed the fine-bulbed ther- 
mometer with its silver sheath, and a considerable number of com- 
parative readings were taken. This was done on a number of days 
and in different conditions of weather, and the screen has proved itself 
to be considerably in advance of all the others. Its readings being 
quite as good as those given by the fine silvered bulb, I shall not 
attempt to say which gives the lowest readings, as the conditions 
of the trials have not been sufficiently severe or varied enough to 
bring out any decided difference ; but under all conditions yet tried 
the screen was quite as low as the silvered bulb. It is unnecessary 
here to give any detailed account of these trials by themselves, but 
I shall presently give some comparative readings taken with this 
screen, with the Stevenson screen with the bottom open, and with 
it closed, also with the silvered bulb, showing that the new screen 
gives much lower and more correct readings than either of the 
other screens, and, as the circulation through it is very free, these 
lower temperatures cannot be due to the high inertia of the new 
screen. 
Trial of Screens. 
In the first and second parts of this communication the results 
were given of some trials of different methods of protecting the 
thermometer against the effects of radiation. The conclusions 
arrived at were, that the most correct readings were given by the 
thermometer placed in a strong current of air produced by a suction 
fan, or by a fine-bulbed thermometer exposed under a sunshade 
with its bulb protected by a silver sheath. Readings taken by 
these two methods agreed very well with each other, and either of 
them was taken as a standard of temperature with which to com- 
pare the readings given by the different screens. Compared with 
these standards, the ordinary Stevenson screen was found to give 
readings of from 1 0, 3 to 2° ‘8 too high, when there was much radia- 
tion and little wind. Further, it was found that when the Steven- 
VOL. XIV. 24/8/87 E 
