GASE.S BY THE ACTION OF KONTGEN BAYS AND OTHER AGENTS. 
421 
with a hole somewhat smaller than the diameter of the tube at each end, so that a 
beam of light fi-om an arc lamp might pass along the axis of the tube. 
I'lie quartz lens through which the light entered had a focal length of about 
to centims., and the arc was generally placed at such a distance that the light was 
brought to a focus about the middle of the tube. A strong arc is necessary in these 
experiments. 
Filtered air was passed through the apparatus until all dust particles were 
removed. The presence or absence of dust particles was easily ascertained by 
allowing expansion to take place as described above, while the light from the arc 
traversed the tube, a sheet of mica being interposed at this stage to cut off the ultra¬ 
violet rays. When drops ceased to appear on 02)ening the stopcock T.,, a few minutes 
were generally allowed to elapse before exposing to the ultra-violet rays, to enable the 
air inside the tube to come to rest. To start the exposure to ultra-violet rays the 
mica was removed. 
Under the conditions described above, a bluish fog is seen in the tube in about 
two minutes, making its a})pearance first near the apex of the beam of light, and 
then extending both ways in the form of a double cone. That the fog when it first 
appears is confined to the path of the light is easily proved by displacing the tube 
slightly to one side; or better, by inserting the mica screen and moving the box with 
the tube fixed inside it nearer the arc, so that the luminous rays converge to a focus 
much nearer the far end of the tube ; or, without moving the tube, by inserting a 
glass lens just in front of the tube after the fog has appeared, so as to bring the 
luminous rays to a new focus. In each of these ways it is easy to prove that the fog 
does actually arise, not near the quartz nor the glass walls of the tube, but along 
the axis of the tube in the neighbourhood of the point where the light is most 
concentrated. It was found that the shape of the cloud was specially well defined 
when the water in the tube contained two or three per cent, of caustic potash. This 
prevented any dej^osit of fine drops on the inner surface of the quartz by keeping 
the inside of the tube not quite saturated. 
We thus obtain a further confirmation of the conclusion already drawn from the 
expansion experiments, that ultra-violet light produces nuclei throughout the volume 
of the moist air which it traverses, and not only at the surface of the quartz or tlie 
glass walls of the tube. 
On allowing the air to expand after the fog has a})peared, condensation takes place 
throughout tlie tube, showing that outside the part traversed by the strongest 
radiation, nuclei have been formed, but not large enough or numerous enough to form 
a visible fog. These may arise partly through some of the nuclei produced in the 
strongest part of the beam travelling into other parts of the tube; they may also 
have been produced by the action of ultra-violet rays, scattered by the cloud particles 
produced in the direct path of the light. These will scatter ultra-violet light even 
before they have grown large enougli to make themselves visible by scattering the 
