return. But the molecules are continually changing their velocity and may sometimes attain a speed which is seven times as great as the average. Suppose, therefore, that a certain quantity of hydrogen were diffused through our air, every now and then a molecule of hydrogen in its wanderings would attain the upper limit of our atmosphere, and then it would occasionally happen that with its proper speed it would cross out into space beyond the region in which its movements would be interfered with by the collisions between other atmospheric molecules. If the attraction of the earth was sufficient to recall it, then, of course, it would duly fall back, and in the case of the more sluggishly moving atmospheric gases the velocity seems always small enough to permit the recall to be made. But it happens in the case of hydrogen that the velocity with which its molecules are occasionally animated rises beyond the speed which can be controlled by terrestrial gravity. The consequence is that every now and then a molecule of hydrogen succeeds in bolting away from the earth altogether, and escaping into open space. Thus it appears that every molecule of free hydrogen which happened to be present in an atmosphere like ours, would have an unstable connection with the earth, for wherever in the vicissitudes of things it happened to reach the very uppermost strata it would be liable to escape altogether. In the course of countless ages it would thus come to pass that the particles of hydrogen would all effect their departure, and thus the fact that there is at present no free hydrogen in the air over our heads may be accounted for.
If the mass of the earth were very much larger than it is, then the velocities with which the molecules of hydro-
