But it does not seem evident that an attractive medium would not also interfere with the passage of bodies through it, though not in the same way as a repulsive one; and the oscillation through its centre necessary for its preservation complicates the theory somewhat. Also, any marked accumulation of a powerfully acting medium round the various celestial bodies would cause, if varied in any way by their changes of relative position, perturbations in their movements. The very fact, however, that its own action was so energetic might make the disturbance in its arrangement produced by other masses small, especially if it penetrates those masses, as is probably generally maintained. The subject is, of course, one of great difficulty, and objections readily suggest themselves to any hypothesis regarding it; still, it would appear that on some accounts it might be better, instead of assuming the medium to be wholly or predominantly attractive or repulsive, to suppose it to have the two forces equally balanced in its constitution; and if it be, like other matter, grouped in molecules, the balance would naturally exist in each molecule, making it inert at any but very small distances, and exerting at these very small distances a force the character of which would vary according to the direction. We have said that the discussion of the exterior action of the molecules--that is, of their action on each other, or on exterior points in general--would be exceedingly complicated. The only way in which it seems practicable is that in which the mutual actions of the planets have been investigated, namely, a development of the force in the form of a series; but this cannot be done advantageously unless the distances between the molecules are considerably greater than the molecular diameters. If, however, we make the development of the ratio of the attraction (or repulsion) exerted by the vertices of a regular polyhedron in the direction of its centre, to what it would exert if concentrated at that centre, in a series of the powers of the ratio of the molecular radius to the distance of the point acted on from the centre, it will be found that the coefficients of the first and second powers vanish in all cases; and that in all, except that of the tetrahedron, those of all the odd powers also disappear, as well as that of the fourth in the dodecahedron and icosahedron. If, then, the absolute attractive or repulsive power of any envelope is very nearly compensated by that of an opposite character prevailing in the rest of the molecule, (as seems probable,) the whole series can be reduced, at any distance which is very great compared with the molecular diameter, to two terms--one a constant with a very small value, and the other containing the third, fourth, or sixth power of the small quantity which the ratio of the diameter to the distance has now become. This should have a negative multiplier, in order that the force should become zero; and