He seemed, therefore, to have a large experimental basis for his answer to the question, "What happens when a substance burns?" But at a later time it was proved that iron was also produced by heating the calx of iron with carbon. The antiphlogistic chemists regarded fixed air as composed of carbon and dephlogisticated air; the phlogisteans said it was a substance highly charged with phlogiston. The antiphlogistic school said that calx of iron is composed of iron and dephlogisticated air; the phlogisteans said it was iron deprived of its phlogiston. Here was surely an opportunity for a crucial experiment: when calx of iron is heated with carbon, and iron is produced, there must either be a production of fixed air (which is a non-inflammable gas, and forms a white solid substance when brought into contact with limewater), or there must be an outrush of phlogiston from the carbon. The experiment was tried: a gas was produced which had no action on limewater and which was very inflammable; what could this be but phlogiston, already recognized by this very property of extreme inflammability? Thus the phlogisteans appeared to triumph. But if we examine these experiments made by Priestley with the light thrown on them by subsequent research, we find that they bear the interpretation which he put on them only because they were not accurate; thus, two gases are inflammable, but it by no means follows that these gases are one and the same. We must have more accurate knowledge of the properties of these gases. The air around a burning body, such as iron, after a time loses the power of supporting combustion; but this is merely a qualitative fact. Accurately to trace the change in the properties of this air, it is absolutely necessary that exact measurements should be made; when this is done, we find that the volume of air diminishes during the combustion, that the burning body gains weight, and that this gain in weight is just equal to the loss in weight undergone by the air. When the inflammable gas produced by heating calx of iron with carbon was carefully and _quantitatively_ analyzed, it was found to consist of carbon and oxygen (dephlogisticated air), but to contain these substances in a proportion different from that in which they existed in fixed air. It was a new kind of air or gas; it was _not_ hydrogen. This account of Priestley's experiments and conclusions regarding combustion shows how easy it is in natural science to interpret experimental results, especially when these results are not very accurate, in accordance with a favourite theory; and it also illustrates one of the lessons so emphatically taught by all scientific study, viz. the necessity of suspending one's judgment until accurate measurements have been made, and the great wisdom of then judging cautiously. About 1779 Priestley left Lord Shelburne, and went as minister of a chapel to Birmingham, where he remained until 1791.