No. 1

On a Modification of Diffusion Which Must Occur on the Surface of a Mixed Gaseous Cosmical Body.

Supposing the entire surface to be at the same temperature; if in their excursions two molecules were both moving directly outwards from the surface, then the velocity being proportional to the square root of the mass, the lighter molecules will travel further against gravitation than the heavier. Again, when on their return they meet with other molecules, the mean direction of their new motion will be less towards the centre than before the encounter; whilst the mean direction of the heavy molecules will be more towards the centre of the mass than before, and generally, supposing originally an equal distribution of various molecules at the surface, the tendency of the movements is to produce such a distribution of mass that as a final result, although a few of the lighter molecules may probably be found at the centre of the mass, yet the greater number of the light molecules will be on the outside and the heavy on the inside.

This fact directly leads to several interesting cosmical speculations, of which the following are illustrations:—Supposing two bodies to come into collision by mutual attraction from infinite distance. After impact it may be shown that in order for any one single particle to be again projected into space, so as to escape the bodies' attraction, any particle at mean position must have twice the velocity or four times the energy, of the velocity of mean square during impact. The whole of the impacting molar ^* velocity will not go to produce velocity of molecule, as may be found by comparing the molar velocity necessary to produce any definite temperature with the molecular velocity as calculated for that temperature, therefore the actual velocity of mean square of the page 4 molecules will be less than the mean molar velocity, so if the temperature at impact be assumed to be approximately uniform throughout, the heat motions of any of the molecules will be too small to enable them to leave the mass in opposition to the attraction of gravitation. In such a collision, at impact the temperature of masses of different elements will of course be proportional to the square root of their atomic weight, because, as we have assumed, the mean square of the velocity of different elements coming into collision with equal velocities will be the tame. But the temperature of a gaseous mass must generally become uniform. When it is so, the velocity of hydrogen may frequently be many times as great as the initial velocity, and hence any particles of hydrogen that come to the surface with this velocity would be carried away into space. Thus hydrogen must almost certainly be diffused through space. In the following letters I have called this action "selective escape."

Again this selective process must also be at work in the sun, and hence probably the reason why hydrogen forms so large a part of its surface. On this principle the assumption that the line 1474 is of an element lighter than hydrogen is shewn to be very probably true. It may not be hopeless, when the real temperature of the photosphere, and the distance to which it extends, is quite known that the approximate atomic weight of this element might be ascertained. If hydrogen is diffused through space, then this element may be still more largely diffused. It is not impossible that some of this element may be in some of the meteorites, at all events in the analysis of these bodies this should be thought of.