282 Mr Ritchie's Experiments and Observations 
glass, it is obvious that the heat communicated to the included 
air will radiate more rapidly from the half-blackened bulb than 
from the transparent one, and consequently the fluid in the stem 
ought to rise, which is directly contrary to the result of the ex- 
periment. 
Having thus shown that this experiment is incompatible with 
the hypothesis, that caloric passes through glass in the same 
manner as through an opaque substance, it will now be easy to 
prove, that a portion of caloric passes freely through the bulbs 
of the thermometer in the same manner as light. When both 
bulbs were transparent, the temperature of the included air 
would be raised, merely by the caloric absorbed by the sub- 
stance of the glass ; the portion which permeated the bulbs 
having no power to expand the air through which it was trans- 
mitted. Now, if the portion of caloric which freely permeated 
the interior hemisphere of one of the bulbs, be prevented from 
passing freely through the exterior hemisphere, an increase of 
temperature would evidently be the result. The coating of 
lamp-black completely intercepted this portion, which was ren- 
dered sensible by the additional temperature which it communi- 
cated to the included air. 
I was anxious to vary this experiment, by employing very 
thin plates of glass as screens ; but to procure plates of sufficient 
thinness, and perfectly transparent, appeared at first a very dif- 
ficult problem. At length it occurred to me to try portions of 
a large sphere, instead of plane surfaces. I therefore caused 
some very large globes to be blown extremely thin, with por- 
tions of which, used as screens, I performed the following expe- 
riments. 
2. Place a plate of polished metal, having an aperture about 
an inch in diameter, at a small distance above the differential 
thermometer, one of the bulbs of which being directly under the 
aperture. Blacken one of the surfaces of the plate of glass, by 
passing it rapidly through the flame of a candle, and place it 
directly above the aperture in the metallic plate. Place the 
heated ball directly above the aperture, and observe how far the 
liquid in the stem descends before it becomes stationary. Re- 
move the coating of lamp-black by a dry brush or feather, and 
the fluid will instantly begin to descend, and actually sink se- 
