34 Joty—On the Conservation of Mass. 
a large area, the surface of which, so far as it lacks symmetry, is varying with 
each experiment, the idea of a directed acceleration arising from this source must, 
I think, be excluded. What error there is, supposing it appreciable, which is very 
doubtful, must be irregular in its effects. 
Again there is nothing in the hanging of the paper cylinder containing the 
vessels to determine in one direction more than in another any resultant force 
arising from the effects of temperature; we therefore do not expect that in 
successive experiments it will take the same direction, and change according 
as the observations are effected at noon or midnight. 
No error from the momentary heating of the radiator when melting the plug 
could be detected, although repeated trials were made. One of these is given in 
detail further on. The isolation of the radiator in the cell was, however, found by 
preliminary experiments to be essential to its successful working. 
It is probable that there is some adjustment of the point of centre of gravity 
of the vessels during the thirty seconds occupied by the liquid in descending to 
the lower vessel. This will doubtless give rise to a corresponding shift of the 
beam about its centre of suspension. But I have not been able to detect any 
acceleration of the masses arising from thissource. Similarly the kinetic effect of 
the falling liquid is balanced by internal reactions, leaving no external resultant. 
A slight tremor or vibration of the light-spot may sometimes be observed to attend 
the descent of the liquid. ‘These mechanical effects are again not a source of 
trouble so far as I have observed. Like thermal effects, they would, if existent, 
be indefinite as regards direction. In the earlier days of the experiments, not 
being quite satisfied that residual effects from such sources might not exist, | made 
experiments in which the conditions were specially arranged to favour the develop- 
ment of mechanical effects, the vessels being hung at a considerable angle out of 
the vertical. I could not detect the smallest effect. In these experiments, the 
upper vessel was filled with 30 grammes of distilled water, and the lower one with 
the same weight of sand. 
There remains still a mechanical effect of assured existence and of directed 
nature—that arising from the fact that the liquid leaving the upper vessel 
possesses a velocity greater by a very little than the lower vessel into which it 
falls, being further from the centre of the earth. As might be at once anticipated, 
the effect is very small, and, in fact, outside the limits of present observation. 
Assuming that one gramme of liquid falls per second through the height of 2 cms., 
and taking the radius of the Earth as 64 x10’ ems., and the diurnal velocity as 
3 x 10° cms. per second, we find that the excess velocity which the liquid possesses 
amounts to 10“ ems. per second. A force amounting to about the ten-thousandth 
part of a dyne acts on the vessel. This force would not be determinable with the 
present arrangements. It will always act in the one direction—that is, from west 
to east. 
