The elements of the problem are these:--The atoms[8] of several ponderable matters exist in combination: those which are combined having strong affinities, but having also affinities less strong for some of the surrounding atoms that are otherwise combined. The atoms thus united, and thus mixed among others with which they are capable of uniting, are exposed to the undulations of a medium that is so rare as to seem imponderable. These undulations are of numerous kinds: they differ greatly in their lengths, or in the frequency with which they recur at any given point. And under the influence of undulations of a certain frequency, some of these atoms are transferred from atoms for which they have a stronger affinity, to atoms for which they have a weaker affinity. That is to say, particular orders of waves of a relatively imponderable matter, remove particular atoms of ponderable matter from their attachments, and carry them within reach of other attachments. Now the discoveries of Bunsen and Kirchoff respecting the absorption of particular luminiferous undulations by the vapours of particular substances, joined with Prof. Tyndall's discoveries respecting the absorption of heat by gases, show very clearly that the atoms of each substance have a rate of vibration in harmony with ethereal waves of a certain length, or rapidity of recurrence. Every special kind of atom can be made to oscillate by a special order of ethereal waves, which are absorbed in producing its oscillations; and can by its oscillations generate this same order of ethereal waves. Whence it appears that immense as is the difference in density between ether and ponderable matter, the waves of the one can set the atoms of the other in motion, when the successive impacts of the waves are so timed as to correspond with the oscillations of the atoms. The effects of the waves are, in such case, cumulative; and each atom gradually acquires a momentum made up of countless infinitesimal momenta. Note, further, that unless the members of a chemically-compound molecule are so bound up as to be incapable of any relative movements (a supposition at variance with the conceptions of modern science) we must conceive them as severally able to vibrate in unison or harmony with those same classes of ethereal waves that affect them in their uncombined states. While the compound molecule as a whole will have some new rate of oscillation determined by its attributes as a whole; its components will retain their original rates of oscillation, subject only to modifications by mutual influence. Such being the circumstances of the case we may partially understand how the Sun's rays can effect chemical decompositions. If the members of a diatomic molecule stand so related to the undulations falling on them, that one is thrown into a state of increased oscillation and the other not; it is manifest that there must arise a tendency towards the dislocation of the two--a