5 6 Observations on

which a candle burns in it quite naturally,and sometimes with a much enlarged flame,and at last becomes mere phlogisticated air.In all these processes I took it for granted,(not having examined the air except whenit was completely or, at least, very nearlyreduced to one of the two states above-mentioned) that the aproximation to itsfinal state of phlogisticated air was equable,so that as soon as it began to be diminished,it also began to lose its power of affectingcommon air. I find, however, that, withrespect to several of the causes of diminu-tion, and perhaps all of them, the airpasses very suddenly from the state in whichit is perfect nitrous air, to the state above-mentioned ; but that the term at whichthis change takes place is various, as some-times two thirds, and sometimes fourteenfifteenths of any quantity on which the ex-periment is made, will have disappearedbefore any sensible change can be observedin the remainder. I have even sometimesbeen inclined to think that its power of

affecting