IN METALLIC TUBES. 49 



a day reduced to 1.25. In another tube, two ounce mea- 

 i'ures were in the lame time reduced to 0,45, both com- 

 pletely phlogirticated. 



Four ounce meafures of nitrous air were reduced 

 in this procels to two completely phlogifticated ; where- 

 as, in any other procefs, only one-fourth of phlogifti- 

 cated air can be found in any given quantity of nitrous 

 air. 



Air naturally contained in clear water is fomething 

 purer than common air ; but air produced by expofing 

 metallic tubes filled with water to a moderate heat, {o as 

 to be kept fome time in the ftate oi Jlcam^ is always lefs 

 pure than atmofpherical air. There muft, therefore, be 

 a produ£lion of phlogifticated air in this cafe alfo. 



Having filled a Jilvcr tube with water, and kept it 

 fufpended over the fire a whole day, I found the air 

 within it of the flandard of 1.25, -when the air expelled 

 from the fame water was of the ftandard of i.o. 

 Ufing a tube of lead^ in the fame manner, the air 

 within it was of the ftandard of 1.6. In both thefe ca- 

 fes, therefore, there muft have been a produdion of 

 phlogifticated air, and probably from the phlogifton of 

 the metals. 



P. S. Since 1 wrote the preceding account I have found 

 that inflammable air heated in a gun barrel is ^o far from 

 approaching to the ftate of phlogifticated air, that, 

 when it is fired together with dephlogifticated air, the 

 diminution is greater than with the original inflamma- 

 ble air. This 1 tried twice, keeping the gun barrel in 

 a red heat the whole day, and not examining the air till 

 the next morning. This difference between the effedt 

 of copper or filver, and ot iron, on inflammable air, 

 in the fame degree of heat, is not a little remarkable. 



To the account of thefe experiments 1 Ihall add, that 

 pure phlogifticated air may be procured in the eafiefl: 



Vol. V. G and 



