The next series of experiments related to the generation of “air” by the action of acids on metals. _Aqua-regia_ and gold, _aqua-regia_ and antimony, _aquafortis_ and iron, dilute oil-of-vitriol and iron, yielded gases which contracted on standing in contact with water. This, in the case of the oxides of nitrogen, is to be ascribed to their reacting with the oxygen of the air accidentally present in the receiver; but in the last case Hales noticed that the gas absorbed in cold weather was re-evolved on rise of temperature, as one would expect with hydrogen. These experiments led him to investigate the action of certain mixtures on ordinary air. Thus a mixture of spirits of hartshorn (or ammonia) with iron filings absorbed 1½ cubic inches of air, and one with copper filings, twice as much. Further, a mixture of iron filings and brimstone absorbed in two days no less than 19 cubic inches of air. But it is disappointing to find that, in spite of all the experimental facts which Hales accumulated, he was unable to make use of them. The prejudice in favour of the unity and identity of all these “airs” was too great for him to overcome. True, he sometimes theorises a little, as for example when he remarks (p. 285):--“If fire was a particular kind of body inherent in sulphur (_i.e._ combustible matter of all kinds), as _Mr. Homberg_, _Mr. Lemery_, and some others imagine, then such sulphureous bodies, when ignited, should rarefy and dilute all the circumambient air; whereas it is found by many of the preceding experiments, that acid sulphureous fuel constantly attracts and condenses a considerable part of the circumambient elastick air: an argument that there is no fire endued with peculiar properties inherent in sulphur; and also that the heat of fire consists principally in the brisk vibrating action and re-action between the elastick repelling air and the strongly attracting acid sulphur, which sulphur in its Analysis is found to contain an inflammable oil, an acid salt, a very fix’d earth, and a little metal.” Enough has now been said to give a fair idea of Stephen Hales’ researches. It will suffice if his conclusions be stated in his own words (p. 314):--