126 REPORTS ON THE STATE OF SCIENCE. 
The vapours of halogens can also produce high electric fields at any 
other point of the discharge, where there is a constriction of the path, 
and produce positive rays there. 
If one fill a tube such as is illustrated in fig. 2 with H., evacuate it 
and pass a current from i, the anode, to the cathode x, then cathode 
rays go out from the cathode, and striction-cathode rays out of the tube 
1, 2, at 2, but no positive rays are to be found in the tube. If we 
now introduce any halogen vapour into the tube, not only does the fall 
at the anode rise, but a high fall is also created near 1, and hydrogen- 
anode rays are emitted from A, and striction-anode from A:. 
If we have helium instead of H, in the tube together with a halogen, 
helium rays are produced. 
Last winter I endeavoured to ascertain the cause of the remarkable 
power of the halogen vapours to increase the fall at a constriction of 
the path of discharge; and I found that the following hypothesis is 
supported by experiments of which time does not allow me to give 
an account: the halogens are, in accordance with their electro-negative 
character, much more inclined than other gases to form negative ions, 
Fie. 2. 
or, what is the same, to absorb negative electrons; and where there is 
a decrease of electrons one always finds an increase in the fall. 
If, now, the tendency of the halogens to form negative ions is the 
explanation of the high fall at the anode, then when a perforated anode 
is used those negative ions should become accelerated and appear behind 
the anode as negatively charged atomic rays, in the same manner as 
positively charged rays (the canal rays) pass through a perforated 
cathode. 
In a tube like the one shown in fig. 3, I tried to find these rays. 
K is an aluminium cathode; into the big tube leads another glass tube, 
B. In the end of this tube there is fused the anode a, a platinum 
plate which has a central slit. co is a cylinder of wire gauze, D a 
metallic diaphragm, c and pD are in metallic connection with each 
other and form a closed conductor inside which there can be no electric 
field. They can be connected by a wire with the pole of an electric 
machine. F is a fluorescent screen. 
If there is a mixture of H. and I. vapour in the tube, there is, 
in a suitable vacuum, an anode fall of about 4,000 volts; from a the 
H, anode rays (already mentioned) enter into the big vessel. In the 
opposite direction, towards F, there pass not only negative electron- 
