104 ME. P. V. BEVAN ON THE COMBINATION OF HYHEOGEN 
Wilson/'' so tJiat a sudden expaiisio]! of adjustable amount could be produced, as 
often as was desired, in the same sample of the gas mixture. Great difficulty was 
found t(.) arise fi'oni the action of chlorine on tlie indiai'ubber stopper used for stopping 
the descent ot the piston. Several artifices were tried to overcome this difficulty, 
which was never entirely removed. By covering 
the stopper with a thin coating of a mixture of 
solid paraffin and vaseline, if care was used in 
filling the bulb used, so that as little chlorine as 
possible should come in contact with the stojDper, 
a single stopper could be made to hold a few days. 
The expansion bulb, cylinder and piston, are 
represented in fig. 16. The rest of the ajDparatus 
was the same as in Wilson’s experiments, and 
requires no further descrijition. The method of 
reading the amount of the exjiansion was to 
measure the distance the piston descended in any 
experiment. After a series of experiments, the 
instrument was calibrated by attaching to one of 
the exits from the bulb a manometer, so that the 
distance travelled over by the piston could be 
expressed in terms of the change of pressure 
produced in the expansion bulb. The only part 
of the apparatus on which the chlorine could act 
was the stopper on which the j)iston falls; the 
taps to the exits were lubricated v ith water, and 
one tube leading into the bulb was sealed to the 
delivery tube of the generating and washing- 
apparatus, which was the same as described on 
P 2 L 78 and 79, 
The experiments with this a^Dparatus showed that when light is allowed to fall on 
a mixture of hydrogen and chlorine, a nucleus-forming substance is produced, which 
causes condensation on a certain definite supersaturation. This condensation apiiears 
when an exjiansion is made greater than a certain definite quantity, and no 
condensation a2)23ears with the same expansion unless the gas has been illuminated. 
With chlorine alone the condensation apjieared when the chlorine had been 
illuminated, and it was absent when the cldorine had not been illuminated. To 
observe the cloud or the absence of the cloud in non-illuminated U'as, the imao-e of a 
bat’s-wing gas flame was foi'ined in the bulb, the light j)i’eviously having passed 
through yello\v glass. The cloud produced in gas that had been illuminated was 
easily seen when illuminated by the bat’s-wing flame, so that the difierence between 
* ‘ Cambridge Phil. Soc. Proc.,’ vol. 9, p. 333. 
