AIR BY WATER. 23 



than fevera] of thofe that were in the jars. And this is 

 the circumftance which makes the experiment of fuch 

 difficult folution. I always, however, found that when 

 common air was fubjeded to this experiment, the de- 

 phloglfticated part of it was abforbed in the firfl: place. 

 For whenever 1 examined the air it was always found to 

 be more and more phlogifticated, till at laft it was whol- 

 ly fo ; and this was generally the cafe when about three- 

 fourths of the quantity remained unabforbed. 



Ten ounce meafures of common air expofed to rain 

 water from the 28th of July to the 15th of Auguft, in 

 a glafs jar about ten inches in diameter, were reduced to 

 7 ounce meafures, completely phlogifticated, as was ano- 

 ther quantity of 20 ounce meafures, when it was reduc- 

 ed to 15. 



In order to afcertain what kind of air would be moft 

 afFeded in thefe circumftances, I expofed equal quanti- 

 ties of them in the fame manner on the 19th of Dec. 

 and obferved them all to be gradually diminilhed, till 

 July ift; when the dephogifticated air was reduced to a 

 very fmall bubble, and on July 6th the inflammable and 

 common air, and an equal mixture of common and ni- 

 trous air, were all wholly vanilhed. Nitrous air was al- 

 ways abforbed fooner than any other, till it was reduced 

 to the ftate of phlogifticated air, which, if the furface 

 expofed to the adion of the water was large, was foon 

 effeded. 



Thinking that the nearer the air on which this expe- 

 riment was made was to the common atmofphere, the 

 fooner this abforption would be effeded, and that the 

 farther it was from it the more time would be re- 

 quifite for it, I put a meafure of common air into a glafs 

 tube 5 feet in length, placed in a trough of water iB" 

 inches deep, fo that there were 6^ feet from the confined 

 air to the atmofphere. But being left in this fituation 

 from June 5th to July 28th, it was reduced to 0.8, com- 

 pletely 



