520 The Philosophy of the Radiometer [October, 
spheric pressure. The absorbing face of the disc becomes 
heated above the temperature of the opposite face, the film 
of air in contact with the warmer face rises, leaving a 
relatively vacuous space in front. This produces a rush of 
air from back to front which qarries the radiometer vane 
with it. When the exhaustion of the radiometer is carried 
so far that the residual air is only just sufficiently dense to 
neutralise the direCt repulsion of radiation, the neutral point 
is reached. When exhaustion is carried beyond this, repul- 
sion predominates. 
Taking Mr. Crookes’s estimate of the mechanical energy 
of solar radiation at 32 grains per square foot, 2 cwts. per 
acre, 57 tons per square mile, &c., and accepting these as 
they are offered, i.e ., merely as provisional and approximate 
estimates, we are led to a cosmical inference of the highest 
importance, one that must materially modify our interpreta- 
tions of some of the grandest phenomena of the universe. 
Although the estimated sunlight pressure upon the earth, 
the three thousand millions of tons, is too small a fraction of 
the earth’s total weight to effeCt an easily measurable 
increase of the length of our year, the case is quite otherwise 
with the asteroids and the zones of meteoric matter revolving 
around the sun. 
The mechanical repulsion of radiation is a superficial 
aCtion, and must, therefore, vary with the amount of surface 
exposed, while that of gravitation varies with the mass. 
Thus the ratio of radiant repulsion to the attraction of 
gravitation goes on increasing with the subdivision of masses, 
and becomes an important fraction in the case of the smaller 
bodies of the solar system. A zone of meteorites travelling 
around the sun would be broken up, sifted, and sorted, into 
different orbits according to their diameters, if this superficial 
repulsion operated against gravitation without any compen- 
sating agency. Gravitation would be opposed in various 
degrees, neutralised, and even reversed, in the case of cosmic 
dust. Comets presenting so large a surface in proportion to 
their mass, would either be driven away altogether or 
forced to move in orbits utterly disobedient to present calcu- 
lations. This would occur if the interplanetary spaces were 
as nearly vacuous as the torsion instrument with which 
Mr. Crookes made his measurements. 
Regarding the properties of our atmosphere only in the 
light of experimental data, irrespective of imaginary mole- 
cules, and their supposed gyrations or oscillations, we see 
at once that an inter-planetary or inter-stellar vacuum 
must aCt like a Sprengel pump upon our atmosphere, upon 
