426 Dr. Priestley’s Experiments relating to PhlogiJlon y 
cylindrical glafs tube, divided according to ounces and parts of 
ounces of water, it was not eafy to avoid an error of a few 
grains. At other times there was an error of a fmall magni- 
tude on the other fide. But it will appear hereafter, that more 
fleam muff have efcaped invifibly at the open mouth of the 
iron tube than I was aware of. 
That nothing could enter by the pores of the retort at the 
fame time that the water was making its efcape out of them, I 
thought I afeertained pretty fatisfaCtorily by immerfing the 
bulb of it in mercury, contained in an iron veflel. In thefe 
circumftances I obtained air as ufual, only the produce was not 
l'o rapid. I11 this way, however, I procured above an hundred 
ounce meafures of air from moiftened clay ; and I difeontinued 
the procefs without perceiving any termination of it. But the 
moment the retort was railed out of the mercury, it gave air 
three times as fall as it had done before. The quality of the 
air was the fame in both cafes, viz. a little worle than that of 
the atmofphere. 
I even colle&ed thirty ounce meafures of air when the bulb 
J 
of the fame retort was immerfed in hot linfeed oil, but the pro- 
duction of air gradually ceafed, and the next day I found the 
retort almoft full of the oil, which had foaked through it. 
Diflilllng this oil I get 300 ounce meafures of air, wholly in- 
flammable, except a very few ounce meafures at the lafl, which 
were only phlogifticated. 
Still hearing of many objections to the converlion of water 
into air, I now gave particular attention to an experiment of 
Mr. cavendish’s, concerning the re-converlion of air into 
water, by decompoling it in conjunction with inflammable 
air. And in the firfl: place, in order to be fure that the wa- 
ter I might find in the air was really a conftituent part of 
it 
