Joty—On the Conservation of Mass. 35 
As before stated, the chief source of annoyance is the want of thermal 
equilibrium of the air within the case. I hope to be able to reduce the effects of 
this in future experiments, not only by carrying out the experiments on a smaller 
scale, but by removing a large part of the air from around the beam and its load, 
using an air-tight vessel of much smaller size than the present one. Although not 
hoping to carry the vacuum so far as to effect the viscosity of the contained air, a 
reduction of its mass, and corresponding reduction of kinetic effects arising from 
draughts, would, I anticipate, greatly improve the conditions of experiment. 
It was at first intended to study the effects, if any, of the reactions by the 
perturbations of the curve of vibrations as determined by the motion of the 
light-spot. It was soon found—as already stated—that when the vibrations got 
sufficiently small in amplitude to give the requisite sensibility, they were subject to 
irregularities which rendered any reliance on the form of the curve unsatisfactory. 
In fact, the vibrations have become aperiodic before the conditions suitable to 
experiment are attained. Frequently, as long as a couple of hours of patient 
observation had to be put in before a motion sufficiently slow, and at the same 
time sufficiently regular, was reached. It is very questionable if so feeble a 
constraint from the suspension was altogether desirable. However, it is to be 
remarked that a suspension sufficiently rigid to control the effects arising from 
draughts would itself prove a serious restraint on the motion of the beam. I 
therefore resolved to try and seize favourable moments when a slow and apparently 
steady motion was in progress, and then seck to determine if the effect of the 
experiment was to produce any perturbation on this motion. While I believe I 
have in this way probably got as much out of the apparatus as is possible, a study 
of the results which follow will show how unsatisfactory it is when very small 
perturbations come to be sought for. At the instant of experiment, we are, in 
fact, not aware of the tendency of the motion beyond a short distance. The 
torque of the suspension may be dismissed. It will not act as a sudden interrup- 
tion to the steadiness of the motion. The difficulty is with air-currents within 
the case. All that can be done is to compare the curves of experiment with those 
obtained under precisely similar conditions, when a known force was applied to 
therload. It is evident that a change of velocity so great as that obtained under 
the force of 0:04 dynes, as before referred to, would be conspicuous among the 
perturbations affecting the motion of the beam before and during the experiments. 
In interpreting the tables and curves which refer to the inertia experiments, 
it has to be borne in mind that the vessels are hung on the south extremity of the 
beam. The scale receiving the light-spot lies to the west side of the beam; 
and the spot is at 0 when the beam is due north and south. The scale being 
divided and numbered to either side of 0, it is necessary to distinguish these 
numbers in order to know where on the scale the spot stands at any reading. 
