146 MR. C. E. S. PHILLIPS OX THE ACTIOX OF MAGNETISED ELECTRODES 
liy the sudden production of a magnetic field at the centre of the bulb, the following 
experiment was then arranged :—The apparatus shoA^ui in fig. 9 was intended to 
answer two purposes. In the first place, it would test whether the negative ions 
tended to concentrate and be drawn from the ends of the bulb in towards the central 
portion, and also, if that action were prevented, whether the ball A would lose a 
positive charge when the magnets were excited. 
Fig. 
9. 
Two thin copper cones Cj Co were soldered, one upon the pointed end of each of the 
magnetic electrodes M^Mo. They were fitted into the glass tube as sho^sm, and the 
pressure lowered until a discharge j^assed between the cones witli difiicult}”. When 
Ml was connected with the negative terminal of the induction coil and Mo to the 
positive, and the discharge had been started and stoj^ped, the magnetisation of the 
electrodes gave rise to a luminous ring behind and close to the cone C 3 . The reverse 
conditions produced a ring l^ehind the cone Ci. Subsequently to an oscillatory dis- 
cliarge through the tube a luminous ring appeared simultaneously behind each cone 
immediately the magnets were excited. In all these cases the luminosity visil)le 
1 )etween the apices of the cones was slight, hut it was not to he expected that all 
the negative ions could he prevented from concentrating at the centre of the tube by 
tliis device. A was then positively electrified, and the electrodes magnetised. It 
])ecame nuly jmrtially diselectrified—a very different result from that obtained in 
previous experiments where no cones were attached to the electrodes. This result 
was repeated a great many times, and, although in each case the diselectrification 
was enfeebled by tlie introduction of the cones, a further experiment was made with 
a view to completely preventing diselectrification of the hall A while the conditions 
remained suitable for the magnets to act iq^on the electrified gas particles as before. 
The cones were removed from the electrodes. An aluminium cylinder C was then 
slipped in V)efoi'e the tube was sealed up, and arranged so as to he capable of sliding 
along the fine wire as shown in fig. 10 . The hall xV remained as before. As a 
trial experiment the pressure within the tube was lowered and A positively electrified. 
Wlien the magnets were excited A was completely diselectrified. The cylinder C 
was then tapped into the position indicated by the dotted line and the wire 
connected to earth. A was again positively electrified, but when the electrodes MjM^ 
were magnetised there was no diselectrification of the Ixall Avhatever. A proof tliat 
