462 
MR. W. CROOKES ON THE SUPPOSED 
the rotation changed from negative to positive; and when the vacuum was so good 
that the repulsion between the two heated bodies was at its maximum, then also the 
positive rotation was the strongest. It was impossible to resist the conclusion that 
the two sets of phenomena were due to the same cause, and that, as air-currents did 
not produce the old attractions of the magnesium pendulum, so likewise were thev 
equally inoperative in giving rise to the present rotations of the suspended cylinder. 
I will not give in a tabular form the observations taken with this apparatus, as more 
decided results were obtained with a modified form of apparatus which I will now 
describe, and it is not worth while to record observations beyond what are needed to 
prove the case under discussion. 
If the rotation is produced by a reaction between the suspended and fixed body, it 
follows that, were both free to move, each would rotate, but in opposite directions. A 
modification of the form of apparatus last used was therefore devised ; it is shown 
in fig. 5. It consists of a long glass tube, AB, having a bulb, C, blown near the 
lower part. At the top two narrower glass tubes, D, E, are blown on; these contain 
glass rods sealed to the tubes at the upper ends, but, in other respects, loose in the 
tubes. To the ends of these rods are attached two fine silk fibres, each having a 
cylinder of blackened ivory, 1, 2, suspended to it. F is a platinum spiral, equidistant 
from the two cylinders. The small tubes, D and E, are clamped at their upper ends 
by a brass band having a screw at one side. By tightening or liberating this screw 
the tubes are more or less inclined to one another, and the cylinders, 1, 2, can thus be 
adjusted to any desired distance apart. Exhaustion was effected through a lateral 
tube. Observations were taken at intervals during exhaustion, the source of heat 
being either a flask of hot water or a non-luminous gas-flame applied to the glass 
bulb at the place marked by an asterisk. In all cases when rotation was obtained 
the two cylinders moved in opposite directions. Thus, in air of ordinary density, 
cylinder No. 1 rotated counter-clockwise, while No. 2 rotated clockwise ; I shall 
designate this movement as negative, and the opposite rotations, where No. 1 rotated 
clockwise and No. 2 counter-clockwise, I shall call positive. 
The following Table exhibits, in a convenient form, the results obtained with this 
apparatus :— 
