f 20 RA YS OF POSITIVE ELECTRICITY

presence of X3 after the mixed gases of which it was a con-
stituent had been subjected to any treatment was to Introduce
a small quantity of the mixed gases into the discharge tube,
take a positive ray photograph and estimate the brightness of
the X3 line. Before such experiment the discharge tube was
well washed out with oxygen and a test photograph taken to
make certain that no X3 was In the tube before the introduc-
tion of the gas which was to be tested. One property of this
gas, that of combining with mercury vapour when an electric
discharge passes through a mixture of these gases, has already
been mentioned. Another property was discovered accident-
ally : the mixed gases obtained by bombarding the potash were
drawn off day by day and stored up for further tests. It was
soon noticed that some of the samples kept much better than
others and It seemed possible that this difference might be
due to differences in the brightness of the light to which the
samples had been exposed. To test this a piece of magnesium
wire was burnt In front of a sample which was known to con-
tain a considerable quantity of X8, with the result that the X3
almost disappeared. The gas exposed to the light was a
mixture of hydrogen, oxygen and X3J If the oxygen is taken
out of the mixture by absorbing It with charcoal cooled with
liquid air, exposure to light produces no effect on the X3; the
conclusion we.draw Is that under the Influence of the light the
X3 combines with oxygen. If the mixture is kept In the
dark or if the oxygen Is taken out of it the X3 lasts for a long
time, certainly for several weeks. Again If a strong spark is
sent through the mixture containing oxygen so that a vigorous
explosion takes place the X3 disappears, presumably combin-
ing with the oxygen. If the oxygen is removed the mixture
of hydrogen and X3 will stand a good deal of sparking with-
out any considerable diminution in the amount of X8.

The fact that sparking with oxygen destroys the X3, makesograph showsth these gases monatomic