I 166 ] 
of longer continuance, did by no means reftore this 
kind of air : for when I have expofed the phials which 
contained it a whole night, in which the froft was 
very intenfe ; and alfo when I kept it furrounded with 
a mixture of fnow and fait, I found it, in all re- 
foedis, the fame as before. 
It is alfo advanced, in the fame Memoir, p. 41. 
that heat only, as the reverfe of cold, renders air 
unfit for candles burning in it. But I repeated the 
experiment of the Count for that purpofe, without 
finding any fuch effedt from it. I alio remember that, 
many years ago, I filled an exhauded receiver with 
air, that had pafl'ed through a glafs tube made 
red-hot, and found that a candle would burn in it 
perfectly well. Alfo, rarefaction by the air-pump 
does not injure air in the lead: degree. 
Though this experiment failed, I flatter myfelf 
that I have accidentally hit upon a method of re- 
ftoring air which has been .injured by the burning 
of candles, and that I have difcovered at lead; one 
of the reftoratives which nature employs for this 
purpofe. It is vegetation. In what manner this pro- 
cefs in nature operates, to produce fo remarkable an 
effedt, I do not pretend to have difcovered; but a 
number of fadts declare in favour of this hypothefis. 
I fhall introduce my account of them, by reciting 
iome of the'obfervations which I made on the grow- 
ing of plants in confined air, which led to this dii- 
covery. 
One might have imagined that, fince common 
air is neceffary to vegetable, as well as to animal 
fife, both plants and animals had affedled it in the 
fame manner, and I own I had that expectation, 
when 
