and the finning Converfton of Water into Air. 4j ., 
“ s ftron S l y inflammable as ever by the tube communicatin'’ 1 
with the infide of it, while the water rofe within the receiver’ 
and even covered the retort, which was fixed at the very top 
of it, fa that hardly any of the inflammable air remained 
witnin it. In like manner nitrous air pafled through the retort 
■unchanged. 
, From thefe experiments it is impoffible not to infer, that the 
clay of the earthen retort, being thus heated, deftroys for a 
time the aerial form of whatever air is expofed to the outfide 
ol it ; which aerial form it recovers after it has been tranfmit- 
ted in combination from one part of the clay to another, till it 
has leached the infide of the retort, while the water is drawn 
through it in the contrary direction. 
Had this hypothecs been propoied a priori, it would, I doubt 
not, have been thought more extraordinary than the converfion of 
water into air. I propofe to make many other experiments in 
the profecution of thefe ; but till I have an opportunity of 
doing this, I fhall not trouble the Society with any conjectures 
that have occurred to me on the lubject. 
The great difficulty with refpeft to the experiment with the 
.unis is, that the water ffiould pafs through the retort one way, 
and the air the other, and yet that the air ffiould not be able 
to pafs without the water. It is alfo not a little extraordinary, 
that the weight of the air and that of the water ffiould be fo 
nearly equal. 
In the lafl place I muffi obferve, that there is nothing in this 
experiment that contradicts the idea of the converlion of water 
into air, though it does not prove it : for ftill the experiment 
with the tobacco-pipe, in which the fleam is made red-hot 
(whereas in that with the lens it is only of a boiling heat) 
cannot be explained fo well on any other hypothecs anv more 
LI 1 z than 
