the reciprocal tendencies of the particles of the ball and of the earth are as truly opposite, as if the laws of gravitation had, at the moment at which the rebound begins, been reversed through the whole system of the universe. It is, indeed, scarcely possible to imagine a more striking proof of the danger of extending, with too great certainty, a general law, than this instant conversion of _attraction_ into _repulsion_, without the addition of any new bodies, without any change in the nature of the bodies themselves, and a change of their circumstances so very slight, as to be absolutely indistinguishable, but for the opposite motions that result from it, with a change of their circumstances. After observing the gravity of bodies, at all heights of our atmosphere, and extending our survey through the wide spaces of our solar system,--computing the tendency of the planets to the sun, and their disturbing forces, as they operate on each other,--and finding the resulting motions exactly to correspond with those which we had predicted by theory;--in these circumstances, after an examination so extensive, if we had affirmed, as an universal law of matter, that, at all distances, bodies tend toward each other, we should have considered the wideness of the induction, as justifying the affirmation; and yet, even in this case, we find, on the surface of our earth, in the mutual shocks of bodies, and in their very rest, sufficient evidence, that, in making the universal affirmation, we should have reasoned falsely. There is no theory, then, which, if applied to the explanation of _new_ phenomena, is not, to a certain degree, conjectural; because it must proceed on the supposition, that what was true in certain circumstances, is true also in circumstances that have not been observed. It admits of certainty, only when it is applied to the very substances observed,--in the very circumstances observed,--in which case, it may be strictly said to be nothing more than the application of a general term to the particulars, which we have before agreed to comprehend in it. Whatever is more than this is truly hypothetical,--the difference being, that we commonly give the name of _hypothesis_ to cases, in which we suppose the intervention of some substance, of the existence of which, as present in the phenomenon, we have no direct proof, or of some additional quality of a substance before unobserved,--and the name of _theory_ to cases, which do not suppose the existence of any substance, that is not actually observed, or of any quality that has not been actually observed, but merely the continuance, in certain _new_ circumstances, of tendencies observed in other circumstances. Thus, if a planet were discovered revolving in the space which separates the orbits of any two planets at present known, were we to suppose of matter, in this new situation, that it would be subject to the same exact law of gravitation, to which