FROM THE HIGHER ATMOSPHERE. 473 
cular to the surface of a cubical canister containing either warm 
or iced water, are the most intense, and the rest appear to di- 
minish in the ratio of the sine of obliquity. Hence the action 
exerted on the sentient ball of the pyroscope, is always pro- 
portional to the visual angle subtended by the propellent sur- 
face. 
Nor are those i impressions necessarily propagated in straight 
lines. If connected rings of pasteboard be fashioned into a 
sort of cornucopia, its mouth being directed towards the fire, 
notwithstanding the twisted form of the passage, a very consi- 
derable action will be indicated on the ball of the pyroscope, 
presented at the narrow end. 
From all these combined observations it follows, that the 
portion of heat or cold, of which the discharge depends on the 
quality of the surface, is propagated by the vehicle of its gaseous 
medium, though not by any actual streaming of fluid matter. 
No alternative then remains, but to admit that the impressions 
of heat or cold are conveyed through the air with a spreading 
and progressive tendency, in the same, manner as the pulses of 
sound. The aerial medium, by a series of internal oscillations, 
successively transfers its charge, and delivers an impression at 
the end of the chain of communication, of the same kind pre- 
cisely as it had received at the beginning. 
This rapid transmission of Heat to a distance from its source, 
has been hastily termed radiant, and various inaccurate concep- 
tions are entertained concerning its mode and extent of action. 
1. In the first place, then, the process now described never ob- 
tains, unless where a difference of temperature occurs ; and 
it only contributes, along with other active causes, to restore 
the equilibrium of heat. 2. But, in the next place, it has in 
every case a subordinate share only in the diffusion of heat. 
When the air is perfectly still, and of the ordinary density, the 
tide of heat vibrated from a vitreous surface amounts scarcely 
Vou. VIII. P. II. 30 to 
