For some considerable time after this, Cavendish appears to have laid aside Chemistry for other departments of physics. In 1771 he published an elaborate paper on the theory of the principal phenomena of electricity; and in 1776 appeared the curious and interesting account of his attempts to imitate the effects of the torpedo, by an apparatus constructed in imitation of the living fish, and placed in connection with a frictional electrical machine and a Leyden battery. In this imitation he succeeded so well, that all doubts were removed as to the identity of the torpedinal benumbing power with common electricity. In 1776 Cavendish was selected by the Royal Society, in whose 'Transactions' all his previous papers had been published, to describe the various meteorological instruments which were made use of in their apartments; and the succeeding year to this marks the period when he commenced his most important chemical researches, entitled 'Experiments on Air,' which were carried on with frequent and sometimes long interruptions until 1788, no part of them, however, having been published before the year 1783. They led to the discovery of the constant quantitative composition of the atmosphere, the compound nature of water, and the composition of nitric acid. To solve the important problems, whether the atmosphere is constant in its composition, and if so, what is its composition? Cavendish experimented in 1781 for some sixty successive days, making many hundred analyses of air. The honour of the discovery of the compound nature of water, by which perhaps his name has become most famous, is also claimed by James Watt. Cavendish, however, seems at all events entitled to the honour of having first supplied the data on which that discovery was founded, whilst Watt appears to have supplied the conclusion. Between the years 1783 and 1788, Cavendish published his papers on 'Heat,' and his 'Experiments on Air;' the former are three in number, and relate chiefly to the phenomena of congelation, and embody some of the results of experiments made as early as the year 1764. The first of these papers refer to quicksilver, demonstrating the true freezing-point of this metal to be 39° or 40° below zero, while the second and third refer to the freezing of the mineral acids and of alcohol. His experiments on air, which led to the important results already referred to, supplied materials for four papers, besides leading to the observation of many phenomena which were never made public. With the last of these papers published in 1788, Cavendish closed his chemical researches, his remaining publications referring to meteorology and astronomy.