202 
Proceedings of the Royal Society 
the same effect being produced as in one of the curves just discussed 
- — but by almost opposite causes. In the forenoon the weather was 
dull and wet, the sky being overcast with a pallium of cumulus 
clouds. At 13 hrs. the clouds cleared away, the sun shining brightly. 
Observations were begun at 11 hrs., when the Stevenson screens 
read lower than the standard ; the o rain water drying off the louvre 
boards probably cooling them, and so reducing the temperature 
of the air passing over them. About 11*40 the curves cross and 
assume their usual positions. At 13 hrs. a maximum occurs, and the 
effect of the clear sky overhead and sunshine is at once seen by the 
standard crossing the others and giving the highest reading. At 
14 hrs. we have the sky again overcast, and the lines rearrange them- 
selves as before at a minimum point. At 15 hrs. another maximum 
comes in, the clouds breaking overhead and sun gleaming occasion- 
ally, and here again the standard rises — higher than the improved 
Stevenson, and coinciding with the old. After 18 hrs. the clouds 
cleared away almost entirely, but as the boxes had not been heated 
during the day to any considerable extent they give readings slightly 
lower than the standard, but gradually approach it as the tempera- 
ture falls, and ultimately cross and take up their usual positions. 
The readings of the thermometer with the silvered bulb are 
extremely good at first, the radiation effects being all reduced to a 
minimum ; and the delicacy of the instrument remains almost un- 
affected. One great advantage of this instrument is that it can be 
much more carefully read than those in ordinary use, as the radiation 
from the body of the observer is almost without effect. But the 
silver becomes so much tarnished in less than a week as to be 
almost entirely ineffective, the readings being ultimately the highest 
throughout. Very great care is necessary in repolishing the 
thimbles, as they are liable to be scratched, -which seriously impairs 
their reflective powers. It is intended to try experiments with 
nickel-plated thimbles, as it has been found that nickel stands 
exposure to the weather for a longer time without tarnishing. 
The readings of the black bulb arrangement are on the whole very 
satisfactory. They are as a rule lower than those of the standard 
apparatus, sometimes so much so that one is led to suspect that the 
black bulb has too much of its surface covered. At other times 
curious points occur, which are not indicated by any of the other 
