IN TPIE PRESENCE OF DUST-FREE AIR AND OTHER GASES. 
271 
which the plate was fixed. The reading telescope was then fixed some distance oft", 
on a level with the surface of the water in the inner vessel. 
Some water was now allowed to escape through the tube C, and the level of the 
water read on the glass scale by means of the telescope. The tap T was then closed, 
expansion made by pulling the cord which opens communication with the vacuum 
vessel F, and the effect on the contents of the expansion chamber A noted. The 
new level of the water in the inner vessel was again read by the telescope, and the 
air made to contract to its former volume by opening the tap T. The same expan¬ 
sion could be repeated as often as was wished, or the air could be expanded to a 
greater extent by first running out a little water through C. If it was desired to 
try the effect of a smaller expansion the tap T was only slightly opened, and was 
closed before the water in B had quite returned to its original level. Then the 
pinch-cock on the tube C was opened for a moment while the end of the tube was 
dipped into a beaker of water. 
To find the volumes corresponding to the various readings, the bell-jar, with its 
inner vessel, was removed from the glass plate after every series of observations, and 
fixed in an inverted position so that the water could be poured into the inner vessel. 
The whole arrangement was then adjusted so that the ground surface of the rim of 
the bell-jar was level. The weight of water which had to be poured in to fill the 
inner vessel up to the various readings on the scale was then determined, the 
telescope being fixed in exactly the same relative position as in the expansion 
experiments. 
(jenern] AccoTint of the Phenomena Observed. 
The air was generally admitted into the apparatus in the way already described, 
and, therefore, without any attempt to remove dust by filtering. Ptepeated expansion 
of saturated air, as Aitken has shown, removes all “ dust ” particles, and this method 
was generally employed in these experiments. 
The first expansion made, whether large or small in amount, unless the air had 
been allowed to stand for many hours in the apparatus, always produced a fog. This 
was allowed to settle as completely as possible before allowing the air to contract to 
its original volume. In this way a considerable proportion of the dust was removed, 
the particles being carried down by the drops which condensed upon them into the 
water below. 
When this process was several times repeated, the resulting fog became by degrees 
coarser-grained, the drops being both fewer and larger, and therefore, falling more 
quickly. The fog passed at length into a fine rain. When this stage was reached 
one more expansion vvas generally sufficient to remove the remainder of the dust 
particles, and any further expansion, unless it exceeded the limit spoken of below, was 
without visible effect. 
