have not yet been detected. However, this is no final argument against their existence as new and striking discoveries are being made every year. These longitudinal waves force us to conclude that the ether is at least slightly compressible. If compressible we are inclined, from our knowledge of matter to think of the ether as discrete or molecular. If discrete the particles are elastic and we have to postulate a second ether to explain the elasticity of the molecules of the first. This destroys the simplicity of the ether.
J. J. Thomson in his Silliman Lectures (page 160) enunciates a view somewhat similar to that given by Brush. In place of longitudinal waves he used short ether-pulses something like the Röntgen and gamma rays are supposed to be. This view presents the aberrational difficulty, as do other views which attempt to explain gravitation by a type of radiation. On this view any change in gravitation would be propagated with the velocity of light; and certain phenomena in astronomy require the gravitational effect to be propagated much faster than light.
Near the beginning of the paper we showed that the gravitative attraction between small charged bodies is very small compared with the electrical effects. Now if we assume that the linear equations of the ether are only approximate we may account for this relatively small gravitational effect from terms involving differentials of second or higher orders. Larmor opposes this view because the introduction of higher order term not only disturbs the simplicity of the ether scheme, but also leads to optical dispersion in free space. If such dispersion exists it must be very minute, as bodies emerging from eclipses show no appreciable change in color. However, when we compare the gravitational force with the electrical force we see the former is so very small that the higher order term introduced to account for gravitation might give us a dispersion too minute for observation. But, as usual, we meet insurmountable difficulties. On this view the velocity of transmission would be of the same order as that of light; and while we do not know whether the speed is finite at all, we do know that if it is finite it greatly exceeds that of light.
Hence at present gravitation seems to be precluded from the electromagnetic scheme, owing to speed of propagation. Since in the last ten years we have tried to unify, it is unfortunate that it rebels against admission into our scheme. Many have pronounced it a mode of activity in the ether not specified but entirely different from the electromagnetic mode.
Replace Brush's long waves or Thomson's short electromagnetic pulse by rapidly moving corpuscles and we have Le Sage's theory. According to Le Sage space is filled with minute particles, ultramundane corpuscles, moving rapidly in all directions; hence a single body experi-
- See "Ether and Matter," p. 187.