Analyfis of the Air * 233 
verted receiver zza a, (Fig. 3 5.) and with 
a fyphon I immediately drew the water up 
to z z : Then drawing out the fyphon, the 
water would defeend for a quarter of a mi- 
nute, and after that afeend, notwithftand- 
ingthe Candle continued burning, and heat- 
ing the air for near 3 minutes. It was ob- 
fervable in this Experiment, that the fur- 
face of the water ^ z did not afeend with 
an equal progreffion, but would be fome- 
times ftationary 5 and it would fometimes 
move with a flow, and fometimes with an 
accelerated motion but the denfer the fumes 
the fafter it afeended. As foon as the Can- 
dle was out, I marked the height of the 
water above z z, which difference was e- 
qual to the quantity of air, whofe elafti- 
cicy was deftroyed by the burning Candle. 
As the air cooled and condenfed in the re- 
ceiver, the water would continue rifing a- 
bovc that mark, not only till ail was cool, 
but for 20 or 30 hours after that, which 
height it kept, tho’ it flood many days 5 which 
fhews that the air did not recover the elafli- 
city which it had loft. 
The event was the fame, when for great- 
er accuracy I repeated this Experiment by 
lighting the Candle after it was placed un- 
2 der 
