VOL. LXXII.] PHILOSOPHICAL TRANSACTIONS. 251 



after the operation had continued 10 minutes, near $ of the quantity had dis- 

 appeared; the remainder became fit for respiration, and yet was weakly inflam- 

 mable. By further agitation it was diminished half, and then admitted a candle 

 to burn in it, though feebly; but, on continuing the agitation a little longer, it 

 came to extinguish a candle. On this I shall remark, first, that it clearly follows, 

 from this experiment, that if the external respirable air had no access to the inside 

 of the jar, half nearly of the inflammable air was converted into, or consisted 

 of respirable air, since such quantity of air was found in it after the operation. 

 Now it is absolutely impossible that either could happen; for inflammable air 

 could not be converted into half or even -^ or J- of its volume of respirable air, as 

 even ^ of respirable air contains more matter than 4 times its bulk of inflam- 

 mable air; it is then evident, that the external air must have had access to it. 

 Secondly, I agitated about half a pint of inflammable air, obtained from iron 

 and previously passed through lime-water and kept over mercury, in about 12 

 times its bulk of water, out of which its air had been boiled in a glass bottle 

 closed with a glass-stopper. The agitation continued at several times at least 2 

 hours. A large quantity of the air was indeed absorbed, as appeared by opening 

 the bottle in water; but the remainder appeared, by the nitrous test, as noxious, 

 and was also found to be as inflammable as at first. Even Dr. Priestley attests, 

 that inflammable air, which had been united to water for one month, was after- 

 wards as inflammable as ever. 3 pr. iQj . 



The true explanation of the first experiment appears, therefore, to be the 

 following: first, water easily imbibes inflammable air, but does not combine with 

 it; for after it has imbibed -f T of it, its taste is no way altered, as Dr. Priestley 

 has observed. 1 pr. J 96. Water also easily imbibes common air: therefore 

 when inflammable air is agitated in water, having a communication with the 

 atmosphere, the inflammable air must necessarily be diminished by reason of its 

 absorption, and the part so absorbed immediately escapes out of the water into 

 the atmosphere, as is evident by the smell which is perceived when the quantity 

 of inflammable air is considerable. This escape gives room for the further 

 absorption of the inflammable air which then escapes in the same manner. In 

 the mean time the common air under the jar rises into it, as appears by the 

 direct experiments both of Dr. Priestley and Mr. Fontana; and hence the air in 

 the jar must appear by the nitrous test slightly phlogisticated and respirable; but 

 a further agitation will decompose the common air, as we shall soon see, and 

 then a candle will be extinguished. The same process takes place when inflam- 

 mable air stands long in water whose surface is exposed to the atmosphere. 



Another experiment of the same tendency, but seemingly more decisive, is to 

 be found in the 4th vol. of Dr. Priestley's Observations, p. 368. There it is re- 

 lated, that a portion of inflammable air, inclosed in a glass tube hermetically 



k k 2 



