relating to Aif and Water, 2:9 7 



air, and alfo the addition of weight gained by the iron, are 

 much more eafily afcertained in thefe experiments than the 

 quantity of water expended in them, on account of the great 

 length of the veffels ufed in the procefs, and the different quan- 

 tities that may perhaps be retained in the worm of the tub, 

 though I did not fail to ufe all the precautions that I could think 

 of to guard againft any variation on thefe accounts. 



Of the many experiments which I made with iron^ I {hall 

 content myfelf with reciting the following refults. With the 

 addition of 267 grains to a quantity of iron, and the lofs of 

 336 grains of water, I procured 840 ounce meafures of in- 

 flammable air ; and with the addition of 140 grains to another 

 quantity of iron, and the confumption of 253 grains of water, 

 I got 420 ounce m^eafures of air*. 



The inflammable air produced in this manner is of the 

 lighteft kind, and free from that very qff'enjive fmell which is 

 generally occafioned by the rapid folution of metals in oil of 

 vitriol, and it is extricated in as little time in this way as it is 

 poffible to do it by any mode of folution. On this account it 

 occurred to me, that it mufl be by much the cheapefl method 

 that has yet been ufed of filling balloo?7S with the lighted: inflam- 

 mable air. For this purpofe it will be proper to make ufe of 

 cafl-iron cylinders of a confiderable length, and about three or 

 four inches, or perhaps more, in diameter. Though the iron 

 tube itfelf will contribute to the production of air, ^nd there- 

 fore may become unfit for the purpofe in time ; yet, for any 



* If the perfeft accuracy of the former of thefe experiments may be depended 

 on (and it may always be prefumed, that thofe in which little water is expended 

 are preferable to thofe in which more is confumed) the ^vater that neceflarily enters 

 into this kind of inflammable air is about equal in weight to t\\Q phlogijion that is 

 in it. But I propofe to give more particular attention to this fubje^l. 



Vol, LXXV. Q^q thing 



