We know that eight parts by weight of oxygen combine with one part by weight of hydrogen to form water; but we do not know whether the molecule of water contains one atom of each element, or two atoms of hydrogen and one atom of oxygen, or some other combination of these atoms (see p. 131). But by vaporizing water and weighing the gas so produced, we find that water vapour is nine times heavier than hydrogen: now, 1: 9 = 2: 18, therefore the molecular weight of water gas is 18. Analysis tells us that eighteen parts by weight of water gas contain sixteen parts of oxygen and two parts of hydrogen; that is to say, we now know that in the molecule of water gas there are two atoms of hydrogen combined with sixteen parts by weight of oxygen. We now proceed to analyze and determine the molecular weights of as many gaseous compounds of oxygen as we can obtain. The outcome of all is that we have as yet failed to obtain any such compound in the molecule of which there are less than sixteen parts by weight of oxygen. In some of these molecules there are sixteen, in some thirty-two, in some forty-eight, in some sixty-four parts by weight of oxygen, but in none is there less than sixteen parts by weight of this element. Therefore we conclude that the atomic weight of oxygen is 16, because this is the smallest amount, referred to hydrogen taken as 1, which has hitherto been found in the molecule of any compound of oxygen. The whole of the work done since the publication of Dalton's "New System" has emphasized the importance of that chemist's remark, that no safe conclusion can be drawn as to the value of the atomic weight of an element except from a consideration of many compounds of that with other elements. But in Avogadro's law we have a far more accurate and trustworthy method for determining the molecular weights of compounds than any which Dalton was able to devise by his study of chemical combinations. We have thus got a clearer conception of "atom" than was generally possessed by chemists in the days of Dalton, and this we have gained by introducing the further conception of "molecule" as that of a quantity of matter different from, and yet similar to, the atom. The task now before us will for the most part consist in tracing the further development of the fundamental conception of Dalton, the conception, viz., of each chemical substance as built up of small parts possessing all the properties, other than the mass, of the whole; and--what we also owe to Dalton--the application of this conception to explain the facts of chemical combination. * * * * *