MR. W. CROOKES ON REPULSION RESULTING FROM RADIATION. 
133 
Exhaustion. Grain. 
2237' M, force of molecular pressure ='000126 
1316' M, 
424- M, 
259' M, 
153' M, 
94- M, 
647 M, 
32-9 M, 
26'0 M, 
20-0 M, 
13'9 M, 
121 M, 
9-3 M, 
91 M, 
6*0 M, 
1-3 M, 
07 M, 
= •000206 
= ■000368 
= •000511 
= •000718 
= •000987 
= •001086 
= •001140 
= •001076 
= •000987 
= •000727 
= •000646 
= •000682 
= •000619 
= •000520 
= •000269 
= •000224 
481. In fig. 33 these results are plotted as a curve, taking the abscissae in units of 
the hundred-thousandth of a grain, and the ordinates in millionths of an atmosphere. 
The maximum action would not be far from 40 millionths, increasing slowly up to that 
exhaustion, and diminishing suddenly after that point. 
482. In the Proceedings of the Royal Society for November 16, 1876,'“' I gave a 
similar curve of the variation of the force of repulsion in air at different degrees 
of exhaustion. The source of radiation was a candle, and the body repelled was 
blackened mica. In that curve the maximum action is very near 40 millionths 
of an atmosphere, rising slowly and sinking rapidly. 
483. Another similar curve, obtained in quite a different way, is given in the 
fifth part of this research (par. 334).t The instrument here experimented with 
was a cup-shaped aluminium radiometer, and the source of light was a candle. In 
this case also the maximum action occurred close upon an exhaustion of 40 mil¬ 
lionths of an atmosphere, increasing slowly and dying away rapidly. 
484. In the same paper (par. 383)! I have given a third curve of the variation 
of the action of a candle on a blacked mica surface in an air vacuum. Observations 
are wanting between 59 and 14 millionths of an atmosphere, but by continuing the 
curve passing through the other points, it is seen that here, likewise, the maximum 
would not be very far from 40 millionths of an atmosphere. 
485. These agreements must be more than accidental coincidences. Care must, 
however, be taken not to consider them as expressions of more than a partial law, 
for in the diagram containing the curve last referred to (483, 383), is another curve 
* Yol. xxv. p. 305. 
t Phil. Trans., 1878, part I. (The Bakerian Lecture), p. 301. 
X Loc. cit., p. 316 
