430 Proceedings of Royal Society of Edinburgh, [july 18 , 
On our solar radiation measurements, we however have no such 
check. The thermometer with blackened bulb in vacuo may indicate 
a very strong solar radiation, and yet we may feel it chilly ; or it 
may indicate a comparatively low radiation effect, and we may feel 
warm. This difference in the indications may be due to different 
causes, but no doubt wind is one of the most important. The wind 
has but little effect on the solar radiation thermometer, while it has 
a most powerful influence in checking the heating effect of the sun 
on our bodies. It seems therefore desirable that some other 
instrument be designed for the purpose of showing the radiation 
effect, as tempered by the wind, in order that our observatories may 
tell us something more definite about the climate of the place. 
The radiation thermometer described in a previous part of this 
communication is affected by the wind, and might be used for this 
purpose ; but owing to the absorbing surface being flat, it is neces- 
sary that it be kept always turned towards the sun ; it is not, 
therefore, suitable for ordinary work. In place of it, I have for 
some time used a large hollow sphere made of thin metal, and 
having a thermometer fitted to it, with its bulb in the centre. 
What size this ball ought to be has not yet been determined, 
but if observations are to be taken with an instrument of this 
kind, all that seems necessary is, that a uniform size of ball be 
adopted for all observatories, and that it be made of the same 
kind and thickness of metal. The size of the ball, if not too small, 
does not seem to affect the readings much. Readings have been 
taken here with two balls — one 15 cm. in diameter, and the other 
40 cm. Readings given by the 40 cm. ball will be given later on ; 
those given by the smaller ball were only 2° or 3° lower than the 
large one. 
These hollow balls have also been used for night radiation 
measurements. The readings obtained with them are not so valuable 
as those got with flat-surfaced instruments, as they do not get cooled 
so much. While the flat surface gets cooled 10° or 11° below the 
temperature of the air on a clear night, the large ball falls only 5° or 
6° and the small one not so much by 1° or 2°. When in use, the 
balls are fixed to a post at the same height as the screen, and at no 
great distance from it. This method of taking radiation tempera- 
tures at night seems to be better than the one in general use. The 
