3«2 Dr. Austin on the 
other exceed their attra&ion to fire, they will not unite. Even 
when they are combined in the form of volatile alkali, if heat 
be applied, they immediately recede from each other, and the 
alkali is decompofed. When they are not in an aeriform ftate 
their attraction to each other is greater, on account of the 
proximity of their parts ; it is then fuperior to their attraction 
to fire, and therefore they combine; but when their particles 
have receded from each other, as in the aeriform ftate, their 
attraction to each other is fo diminifhed by the diftance of 
their parts, that their attraction to fire, which is uniform, 
prevails, and keeps them in a feparate ftate. The fpecific gra- 
vity of inflammable air being eleven times lefs than that of phlo- 
giflicated air, the diftance of its particles muft be greater than 
the diftance of the particles of phlogifticated air in the propor- 
tion of \/ 1 1 to i, if the elementary particles of the two airs 
be of equal magnitude; and its effeCt, on this account, in 
diminifhing attraction muft be greater than that of phlogifti- 
catcd air, in the proportion of, or more probably as the fquares 
of, thofe numbers. 
Whether it be admitted, that fire thus combines with other 
fubftances, and is feparated from them by their mutual attrac- 
tions, according to the general law, is not further material to 
the prefent enquiry, than as it accounts for -a circumftance 
which feems to be eflablifhed by the following experiments ; 
namely, that the combination of the phlogifticated and inflam- 
mable airs, and the formation of volatile alkali, depends 
chiefly, if not altogether, on the approximation of the parts 
of inflammable air, when phlogifticated air is prefen ted to 
them. 
Into a cylindrical glafs tube, filled with, and inverted in, 
quickfilver, I introduced fome phlogifticated air, and after- 
wards fome iron filings moiftened with diftilled water. By 
this 
