214 
MR. A. Ll. hughes ON THE 
Let us consider whether reflection of electrons is adequate to explain the part AB 
in curve I., fig. 7. Let w be the solid angle subtended by the disc at any part of the 
case in front of the disc. In the expression for the number of the electrons which get 
back to the disc there will be a factor wl'Zir. Tliis is roughly '1 for tViis apparatus. 
The other factor is less than 1, otherwise it would mean complete reflection of 
electrons. The ratio of AB to AO is therefore much bigger than can be accounted 
for by reflection of electrons. The earth’s magnetic field is more likely to account for 
this efiect. It will be noticed tliat the maximum velocity is apparently reduced by 
the magnetic field. Tliis is due to the fact that the paths of the electrons are so 
curved that none of them strike the case normally. As long as a few electrons impinge 
on the case normally the correct maximum velocity will be given. This was shown 
in the following experiment. Two coils were arranged on opposite sides of the 
apparatus to assist or to oppose the earth’s field. Light of wave-length X 2537 caused 
the plate to charge up to ‘70 volt. This remained the same for all magnetic fields 
from 0 to 2 gauss. With greater fields the maximum velocity began to diminish 
slightly. This shows that, for velocities of '7 volt (and above), the efiect of the 
earth’s magnetic field (‘5 gauss) can be ignored. 
Column IV. shows that the presence of enough air to allow the majority of the 
electrons to collide with air molecules modifies the velocity distribution curves and 
reduces the maximum velocity from 1'93 volts to 1'86 volts. On opening communica¬ 
tion with the cooled charcoal tube, the emission velocity rose to U93 volts in five 
minutes. It would take considerably longer for the pressure of tlie residual gas to 
reach its lowest value. 
It therefore appears that reflection of electrons, the earth’s magnetic field, and the 
degree of vacuum used in these experiments have no appreciable efiect on the 
maximum emission velocities. 
There remains to consider the efiect due to reflected light. In column I., Table I., 
