488 ON IMPRESSIONS OF COLD 
the liquor will indicate merely the difference of opposite ef- 
fects, and will rise or fall according as the impressions of cold 
or heat sent from the sky chance to predominate. If the light 
reflected down from the heavens be profuse, it will excite 
more heat than the simultaneous frigorific impressions can de- 
stroy. Under a canopy of fleecy clouds, a considerable excess 
of heat is hence excited on the black ball ; but when the sky is 
clear, the influence of cold generally prevails, increasing as; 
the sun declines. 
The question is now to discover the cause of the phenomena 
which have been thus revealed. I have already stated, that 
different experiments appeared to concur in indicating the cold- 
ness of the superior atmosphere as the source of those effects. 
Since pulses are darted from such various surfaces, and since 
the softness of the external coat, and its tendency to fluidity, 
seem vastly to augment their power ; may they not likewise be 
excited from a boundary of air itself? This extension of a 
great principle in the economy of Nature, has never yet been 
surmised ; nor can it be readily brought to the test of direct 
experiment, since a body of air, whether hotter or colder 
than the general medium, would evidently not remain station- 
ary, but continually rise or fall. 
I sought accordingly to examine the effect of directing the 
zethrioscope to a hot stream of ascending air. I placed on bricks 
before that instrument, the lower part being screened, a large ~ 
mass of iron carried from the fire, at almost a red heat. 
The zthrioscope then gave impressions of heat or cold, ac- 
cording as its aperture was without or within the warm cur- 
rent, or was affected by the anterior or the posterior boundary. 
But this experiment proved very troublesome, and occasionally 
turned out alittle unsatisfactory. Another experiment per- 
formed out of doors, to try the action of hot smoke raised from 
a 
omy 
i—_ 
ve 
Ss EY 
aver 
4S" 
ine saith 
