OF GASES AT HIGH EXHAUSTIONS. 
403 
655. The pressure and the logarithmic decrement columns are obtained in the fol¬ 
lowing way Observations are taken first at frequent intervals during exhaustion, 
sufficient time being left between the successive experiments to allow the slight 
heating due to internal friction, and the cooling due to rarefaction to become equalised. 
Air is then slowly admitted through the drying tubes, and another series is taken. 
When several hundreds of such determinations are accumulated, they are collected 
into groups, from which the mean pressure and mean logarithmic decrement for the 
group have been taken. 
In this table I have also given the measurements of the repulsion exerted on 
the blackened end of the mica plate by a candle-flame placed 500 millims. off. The 
repulsion due to radiation commences just at about the same degree of exhaustion 
where the viscosity begins to decline rapidly, and it principally comes in at the 
exhaustion above 1000 M. The observations are thus conducted :—The torsion fibre 
and plate being at rest, the shutter obscuring the candle is suddenly removed, and the 
light allowed to shine on the blackened mica for a series of seven or eight oscillations. 
The radiant force deflecting the mica (not including the torsion) is by no means con¬ 
stant, but mounts up during a very appreciable time. The motion of the mica is one 
of swinging about a point which itself changes more or less slowly with the time. 
This force begins from nothing, rapidly increases, and attains nearly its full value in a 
second or two after the candle is let on. The force depends on the heating of the 
surface of the swinging body by the previous action of the candle. 
The path of the index ray is of the following character :— 
Starting from zero, the mechanical force drives the index to division 88 ; thence it 
swings back to 29 ; the next swing brings it to 83, and so on. The problem is, from 
the varying arcs of oscillation, to find the true zero, or the point at which the index 
ray would rest supposing the force of repulsion to be constant and inertia absent. 
656. Professor Stokes, to whom I am greatly indebted for assistance and advice on 
the mathematics of this subject, has provided a method for calculating the zero for 
the swing produced by the candle-flame. Register the stopping-points when the 
candle is let on in successive swings ; and in an adjacent experiment find the 
logarithmic decrements, without the candle. 
Let l be the logarithmic decrement, N the number whose logarithm is l. Divide the 
successive arcs by J +N, and apply the results to the readings of the second ends of 
3 G 
MDGCCLXXXI. 
