Memoirs presented to the Cambridge philosophical society on the occasion of the jubilee of Sir George Gabriel Stokes, bart., Hon. LL. D., Hon. SC. D., Lucasian professor.

142 MESSRS GLAZEBROOK AND LODGE, EXPERIMENTS ON THE OSCILLATORY
Meanwhile a small Voss machine, attached to the spark knobs, which formed the
terminal of a circuit containing the condenser and the coil, had been excited with its
knobs in contact. At a signal from the observer watching the disk they were drawn
apart, and one, two, three, or four sparks listened for inside the case. The machine
was then short-circuited again, and the lens slightly shifted a felt amount (which
could be done without opening the "camera") so as to bring the spark image a trifle
nearer the centre, and another ring of sparks was then taken; sometimes with the
conditions varied, sometimes with them just the same. Then a third, a fourth, and sometimes a fifth circle of sparks were also taken. The number of sparks which without
too much fear of unintelligible superposition could be taken in a single circle depended
partly on their strength. With a large condenser a single spark might overlap its own
record; with a very small condenser 6 or 8 sparks could be safely taken.
In practice either 4 or 5 was the commonest number, and though chance frequently
caused some overlap it was not usually difficult to disentangle the records when reading
the plate.
It was customary to get about 2 dozen sparks on a single plate, though sometimes
it would have been wiser to try for fewer. But a bad overlap after all is no worse
than if neither record had been attempted.
Lastly, a needle point was held on the still spinning plate near its middle so as
to centre it by a small circular scratch, and then the turbine was stopped, the room
darkened, and the plate removed.
An assistant, Mr Robinson, to whose careful manipulation we are much indebted,
then proceeded to develop the plate, sometimes using an intensifier when the markings
were too faint.
Meanwhile whatever conditions had to be varied were attended to, other measurements,
such as that of the self-induction of the coil, or the timing of fork, were made, and things
were got ready for another spin.
This process went on without interruption for some weeks, and a large number of
negatives were obtained. The plate at first used was the ordinary half-plate size, but in
order to permit larger circles, Mr Swan subsequently sent us square plates, 4 inches
square, and on these the final records were taken.
The spark-trace exhibited the alternate oscillations very distinctly: one end (probably
the cathode) being always brighter than the other, and this brighter end alternated
from side to side with every half-period. The beginning and end of each oscillation
though clear enough to ordinary vision became furry under magnification, and by far
the most definite things to set the crosswire on was a narrow bright radial line or sharp
spit, due evidently to the sparking of the knobs into one another: a phenomenon which
accompanied the main oscillations of the condenser and marked the beginning of each
electrical surge. These spits were so instantaneous that the rotation of the plate had
absolutely no effect on their sharpness. They were narrow lines no wider than the
crosswires.