1887 .] Mr John Aitken on Thermometer Screens. 
81 
by the Stevenson screen on eight days was one degree or more above 
the maximum given by screen C. The maximum difference for the 
month was 1 ° *75. The screens gave the same maximum on eleven 
days, while the mean maximum temperature for the month w r as 36° ‘78 
by the Stevenson, and 36° ‘25 by the C screen ; that is, the C screen 
gave an average maximum temperature of fully half a degree below 
the Stevenson. Of course, the average error is determined very much 
by the number of bright days in the month. Taking the average 
error for the fine days of the month, it was about 1°*4, and that would 
have been the error if the month had been bright throughout. The 
difference for January 1887 promises to be very small. Owing to 
the dull and clouded weather, the screens have read exactly alike on 
almost every day of this month. The observations for December 
and January are not entered in the plates. 
Returning now to the consideration of why the result got by the 
different screens at Granton differ so much from those obtained here, 
I think I have taken every precaution to ensure the correctness of 
my results; and yet we find, even so late as the middle of November, 
that the Stevenson screen with open bottom gave higher readings 
than the closed one ; also that the thermometer with its bulb 
sheathed in silver gave, as in the previous trials, readings much 
lower than either of them, and yet the observations made at Granton 
show no such differences ; how then are we to account for the 
difference in the results obtained at the two places ? 
The first thing that suggests itself as a possible cause of the 
differences is the condition of the louvres in the two screens. Was 
the one dirtier than the other, or were the absorbing powers of 
the paints on the two screens different? We have seen that, on a 
not very trying day, in September, the absorbing powers of the 
two screens used in my last trials caused a difference of about 0 o- 5. 
On a bright day, such as those on which the Granton trials were 
made, this difference would be greater ; and if the closed screen w^as 
in the Granton trials the better absorber of the two, this might 
have neutralised any advantage arising from its being closed. 
With regard to the explanation of the high readings given by 
the silvered bulb at Granton, I have great difficulty. No doubt, 
any imperfection in the cleanness of the silver would increase its 
absorbing powers and so raise the temperature, but doubtless care 
vol. xiv. 7 / 9/87 
F 
