The equivalence of the thermic, mechanical and chemical forces has been demonstrated; the only question hereafter will be to select from among such forces the one that must be adopted as the sole principle, in order to account for all the phenomena that depend upon these causes of various orders. But in the present state of science, it is not yet possible to explain completely by heat or electricity, taken isolatedly, all the effects dependent upon the causes just mentioned. We must confine ourselves for the present to a study of the relations that exist between the principal natural forces--affinity, molecular forces, heat, electricity, and light. But from the mutual dependence of such forces, it is admitted that, in every natural phenomenon, there is a more or less apparent simultaneous concurrence of these causes. In order to explain electric or magnetic phenomena, and also those of crystallization, it is admitted that the atoms of which bodies are composed are surrounded, each of them, with a sort of atmosphere formed of electric currents, owing to which these atoms are attracted or repelled on certain sides, and produce those varied effects that we observe under different circumstances. According to this theory, then, atoms would be small electro-magnets behaving like genuine magnets. Entirely free in gases, but less so in liquids and still less so in solids, they are nevertheless capable of arranging themselves and of becoming polarized in a regular order, special to each kind of atom, in order to produce crystals of geometrical form characteristic of each species. Thus, as Mr. Saigey remarks in "Physique Moderne" (p. 181): "So long as the atmospheres of the molecules do not touch each other, no trace of cohesion manifests itself; but as soon as they come together force is born. We understand why the temperatures of fusion and solidification are fixed for the same body. Such effects occur at the precise moment at which these atmospheres, which are variable with the temperature, have reached the desired diameter." [Illustration: Figs. 1., 2., and 3.] Although the phenomenon of crystallization does not essentially depend upon temperature, but rather upon the relative quantity of liquid that holds the substance in solution, it will be conceived that a moment will arrive when, the liquid having evaporated, the atmospheres will be close enough to each other to attract each other and become polarized and symmetrically juxtaposed, and, in a word, to crystallize.