208 
MESSRS. W. SPOTTISWOODE AND J. FLETCHER MOULTON 
When the air-spark was lengthened, not only did the black bands get broader, 
but, associated with the intensely bright bands, there seemed to be one or more other 
bright bands, much fainter in character, and apparently narrower. Whether they 
were in fact narrower it is difficult to decide, as their greater faintness would perhaps 
give them the appearance of bemg so ; but if it is the fact that they were narrower, 
it is possible that this points to the main bright lines being broader than the image 
of the slit, which would show that the principal discharge occupies some little time; 
on the other hand, the apparent width of the bright bands may be due to irradiation. 
The subsidiary discharges may perhaps arise from the existence of return currents, 
such as are evidenced by the appearance of negative discharge at the positive pole, 
when an air-spark of considerable size is introduced, and also by the double effects which 
tubes so constantly show when we try to produce the relief or non-relief-effects. 
A metal ring was now put round the tube about the middle of its length, and the 
slit was so placed that the ring passed across it about its centre. This ring was then 
joined successively to the air-spark and non-air-spark terminals, and to earth, and the 
slit examined in each case. There was not in any instance found to be any crookedness 
in the bright lines, although when the air-spark was in the positive the slit sometimes 
included the bright termination of the positive column and a section of the hollow 
luminous cone which surrounds it. These observations show that there is no want of 
simultaneity between the discharges caused by the relief or the non-relief-effect and 
the ordinary discharge in the tube; and as we know that the pulse which causes the 
former passes along the outside of the tube with the velocity of electrical currents in 
a conductor, this experiment would seem to demonstrate the existence of a velocity in 
the discharge within the tube of a similarly high order. 
In order further to test the possibility of single pulses giving rise to the effects of 
which we have been speaking, the following experiment (to which reference has already 
been made) was tried. The terminals of a tube were connected with the outside and 
the inside of a small Leyden jar. The poles of the secondary circuit of a coil were 
placed so that the discharge from the coil charged the jar by leaping over intervals of 
considerable size (about a quarter of an inch for the negative pole and three-quarters 
of an inch for the positive pole), so that the make-current was excluded. When the 
coil was worked there appeared a brilliant discharge caused by the jar periodically 
discharging itself through the tube. A slit was placed on the tube and the luminous 
column was examined by the revolving mirror, and it was found that the discharge 
was quite instantaneous, and that usually it was not followed by the next at any 
regular interval, but that occasionally it was multiple. The discharge was then tested 
for both relief and non-relief-effects, and notwithstanding the very large quantity that 
passed at each discharge, it gave them very markedly. The contact breaker was then 
worked by hand so as only to give single flashes. These were tested for sensitiveness 
and were found to be perfectly sensitive. Thus it appears from experiment that the 
whole of the relief and non-relief-effects are completed within each single pulsation, 
