When Dalton was meditating on the laws of chemical combination, a French chemist, M. Proust, published analyses of metallic oxides, which proved that when a metal forms two oxides the amount of metal in each is a fixed quantity--that there is a sudden jump, as it were, from one oxide to another. We are sometimes told that from these experiments Proust would have recognized the law of multiple proportions had his analyses only been more accurate; but we know that Dalton's analyses were very inaccurate, and yet he not only recognized the law of multiple proportions, but propounded and established the atomic theory. Something more than a correct system of keeping books and balancing accounts is wanted in natural science. Dalton's experimental results would be the despair of a systematic analyst, but from these Dalton's genius evolved that splendid theory which has done so much to advance the exact investigation of natural phenomena. Probably no greater contrast could be found between methods of work, both leading to the establishment of scientific (that is, accurate and precise) results, than that which exists between the method of Dalton and the method pursued by Priestley. Priestley commenced his experiments with no particular aim in view; sometimes he wanted to amuse himself, sometimes he thought he might light upon a discovery of importance, sometimes his curiosity incited him to experiment. When he got facts he made no profound generalizations; he was content to interpret his results by the help of the prevailing theory of his time. But each new fact only spurred him on to make fresh incursions into the fields of Nature. Dalton thought much and deeply; his experimentally established facts were to him symbols of unseen powers. He used facts as Hobbes says the wise man uses words: they were his counters only, not his money.