In 1805, he extended his atomic hypothesis to the explanation of chemical phenomena. His first postulate was, of course, that all matter is made up of small particles; these he found possessed the power of attracting (and holding) other particles. He therefore concluded that these invisible particles never subdivide in taking part in chemical changes, and that all atoms of any one element must be alike. But the atoms of the different elements vary in weight, form and combining power. He established the rule, already assumed by William Higgins (1789), that different atoms tend to combine in the proportion of atom to atom. When a compound was composed of two elements only, it was presumed to be binary; that is, since water, for example, was known to be composed of hydrogen and oxygen, it was supposed by Dalton that it _must_ consist of one atom of each of these elements; as was also assumed in the case of ammonia, which he knew to be a combination of hydrogen and nitrogen. Now we know that one atom of hydrogen combined with one atom of oxygen forms, not water, but oxide of hydrogen; while the combining of _two_ atoms of hydrogen with one atom of oxygen forms water. Dalton established the fact that elements combine only in _definite proportions_--e.g., that oxygen and hydrogen will combine only in the _proportion_ of 8 to 1. Had Dalton known that a molecule of water contains _two_ atoms of hydrogen, and _one_ of oxygen, he would have known that _one_ atom of oxygen must weigh _sixteen_ times--instead of eight times--as much as one atom of hydrogen. _One part_ of hydrogen _by weight_ does combine with _eight parts by weight_ of oxygen, but this does not prove that the portion of hydrogen in water contains only one atom. Since it contains two atoms, instead of one, then, in order to preserve the relative combining (or “equivalent”) weights of the two substances, we must assume that the _atomic_ weight of oxygen is 16, for the _combining_ weights of elements represent the relative (not actual) weights of the atoms. And each element has its own fixed combining weight, ascertained by experiment. Water, then, is not a mere mixture or combination of hydrogen and oxygen, in the sense that a pound of hydrogen and eight _and a half_ pounds of oxygen will, when exploded, produce 9½ pounds of water. Combination would indeed be effected, but there would be a residue in the container of just one-half pound of oxygen. The mass of oxygen that has combined will weigh exactly eight times as much as the hydrogen; and now we know that the water formed will contain exactly twice as many atoms of hydrogen as there are oxygen atoms present in the combination, uniting not only “one by one,” but, in this case, two to one. What the law of definite proportions proves is that chemical combinations always take place _between atoms_.