Our sense of smell is affected by extremely minute traces of gases and vapours--traces so small as to be unrecognisable by any other method of perception, direct or indirect. A piece of musk retains its fragrant odour for years, and the most delicate balance fails to detect any appreciable loss of weight in it. We are capable of smelling gases only: liquids and solids, if introduced into the nostrils, irritate the olfactory nerves, but do not stimulate them so as to incite the sense of smell; yet the admixture of a minute trace of some odorous vapour with air appears entirely to change its properties. The effect of inhaling such air, although sometimes pleasant, is very different from the sensation produced by pure inodorous air, and such admixtures were in olden times naturally taken to be air modified in its properties. But such modifications are obviously almost infinite in number, for varieties of scent are excessively numerous; and it was therefore perhaps deemed useless to attempt to investigate such a substance as air, whose properties could change in so inexplicable and mysterious a manner. Owing, therefore, to its elusive and, as it were, semi-spiritual properties, and to its unexpected changes of character, it was long before its true nature was discovered. It had not escaped observation that “air” obtained by distilling animal and vegetable matter, or by the action of acids on iron and zinc, differed from ordinary air by being inflammable; but such “airs” were regarded as atmospheric air, modified in some manner, as it is modified when perfumed. And “airs” escaping from fermenting liquids, or produced by the action of acids on carbonates, were neglected. For long no attempt was made to catch them; and the frothing and bubbling were regarded as a species of boiling, as is still seen in the use of our word “fermentation” (_fervere_, to boil).