at Low Pressures. 
253 
uniform column can only be obtained with the very smallest 
current that will maintain the discharge. At any pressure below 
about 2 mms. as the current is increased from the smallest 
possible steady value the column is at first uniform, but after a 
time begins to develope faint striae, which rapidly become more 
and more distinct as the current rises in value. At pressures of 
several millimetres the current begins to oscillate when it is 
increased much, but striae do not usually appear. 
The electrodes E and E' although made as small as possible 
nearly always produced some disturbance of the luminosity of the 
column. Usually they appear surrounded by a faint dark space, 
and followed on the side nearest the anode by a faint stria. This 
effect was generally greatest at pressures near one millimetre. 
At higher or lower pressures it was possible to nearly get rid of 
it by carefully adjusting the current. If this disturbance was 
very marked the results obtained did not agree very well with 
those obtained when it was not present, but when it was slight 
they were in good agreement. It will be seen in some of the 
tables of results given below that the discrepancies which the 
observations show from the regular laws which they nearly follow 
are rather greater at pressures near one millimetre than at higher 
or lower pressures. This is no doubt due to this disturbance 
of the discharge by the small electrodes being especially marked 
at these pressures. When an observation was being taken if this 
disturbance was slight, then the results were always sure to be in 
good accord with the others of the same kind. All the obser- 
vations given in this paper were taken when the disturbance in 
question was very small. 
The application of the magnetic field always produces a 
transverse motion of the positive column so that it becomes 
brighter along one side of the tube than along the other. This 
shift is in the direction in which a flexible conductor carrying 
a current moves. The electrodes E and E' were sufficiently near 
the axis of the discharge to be still well immersed in the bright 
part of the column after the field was applied. 
This concentration of the discharge on one side by the field is 
theoretically a necessary consequence of the Hall effect. 
Table I. shows the variation of the P.D. between the electrodes 
E and E' with the angle between EE' and a plane perpendicular 
to AB. Evidently if the fall of potential along the discharge 
is uniform this P.D. should vary as the sine of the angle in 
question. The results given in Table I. show that this is approxi- 
mately the case. 
