French chemists had sought to apply the principle of _chemical affinity_ in explaining the apparent homogeneity of the atmosphere. They supposed that oxygen and nitrogen entered into chemical union, the one element dissolving the other. The resultant compound in turn dissolved water; hence the phenomena of evaporation. Dalton tried in vain to reconcile this supposition with his belief in the atomic nature of matter. He drew diagrams combining an atom of oxygen with an atom of nitrogen and an atom of aqueous vapor. The whole atmosphere could not consist of such groups of three because the watery particles were but a small portion of the total atmosphere. He made a diagram in which one atom of oxygen was combined with one atom of nitrogen, but in this case the oxygen was insufficient to satisfy all the nitrogen of the atmosphere. If the air was made up partly of pure nitrogen, partly of a compound of nitrogen and oxygen, and partly of a compound of nitrogen, oxygen, and aqueous vapor, then the triple compound, as heaviest, would collect toward the surface of the earth, and the double compound and the simple substance would form two strata above. If to the compounds heat were added in the hope of producing an unstratified mixture, the atmosphere would acquire the specific gravity of nitrogen gas. "In short," says Dalton, "I was obliged to abandon the hypothesis of the chemical constitution of the atmosphere altogether as irreconcilable to the phenomena." He had to return to the conception of the individual particles of oxygen, nitrogen, and water, each a center of repulsion. Still he could not explain why the oxygen did not gravitate to the lowest place, the nitrogen form a stratum above, and the aqueous vapor swim upon the top. In 1801, however, Dalton hit upon the idea that gases act as _vacua_ for one another, that it is only like particles which repel each other, atoms of oxygen repelling atoms of oxygen and atoms of nitrogen repelling atoms of nitrogen when these gases are intermingled in the atmosphere just as they would if existing in an unmixed state. "According to this, we were to suppose that atoms of one kind did _not_ repel the atoms of another kind, but only those of their own kind." A mixed atmosphere is as free from stratifications, as though it were really homogeneous.