The almost constant relation between the volumes of oxygen and nitrogen in air made it appear not unlikely, in the opinion of some, that air was a compound, and not a mixture; for the law of combination in definite proportions had by this time been enunciated by Professor Thomas Thomson, Dalton’s intimate friend. But between the numbers 21 and 79 there exists no such simple ratio; and, moreover, on artificially producing air by mixing oxygen and nitrogen, there are none of the usual phenomena which characterise the formation of a compound: there is no rise or fall of temperature, nor does the product differ in any way in properties from the constituents. And in 1846 Bunsen showed that the proportion between oxygen and nitrogen is not a constant one, but that the oxygen varies between 20·97 and 20·84; the experimental error did not exceed 0·03 volume, while the difference found amounted to 0·13 volume. Regnault, Angus Smith, A. R. Leeds, and von Jolly confirmed these results at later dates, from analyses of air collected from all parts of the world. That air contains ammonia was first observed by Scheele. He found that the stopper of a bottle containing muriatic acid, when exposed to air became covered with a film or deposit which he recognised to be sal ammoniac, or ammonium chloride. The amount of ammonia in atmospheric air is, however, exceedingly small, and it is best detected in rain-water, which dissolves it; thus the air is considerably poorer in ammonia after a shower. The ammonia, small though its proportion is, plays a great part, although not an exclusive one, in yielding to plants their supply of nitrogen. The rain, percolating through the soil, leaves the ammonia behind, in some form of combination; and it is then attacked by the nitrifying ferments and converted into nitrates, from which the plants derive the nitrogen which forms part of their substance, in combination with carbon, oxygen, and hydrogen.