Nothing has been said here regarding the rival claims of Watt to the discovery of the composition of water, and little need be said. The discovery was made by both in 1784, yet Cavendish visited Watt at Birmingham in 1785, and was apparently on the best of terms with him; and Watt, as proved by Cavendish’s diary, showed him many of his devices connected with the steam-engine. There can be no doubt that Watt had also discovered that when hydrogen and oxygen are exploded together water is the sole product, but he coupled the phenomenon with views involving the material nature of heat, or caloric, as it was then called, which Cavendish repudiated. Cavendish’s later work was carried out in a villa at Clapham, which was fitted as a laboratory, workshop, and observatory, but he had a town-house near the British Museum, at the corner of Gower Street and Montague Place. He had also a library in Dean Street, Soho, which was available for any scientific man who chose to present himself. So singular were Cavendish’s habits that when he wished a book he went to this house and borrowed it as from a public library, giving a receipt for it. Of all men, Cavendish was probably the most singular, but there can be no question of his extraordinary genius. [23] _Brit. Assoc. Report_, 1839, p. 64. [24] _Experiments and Observations on Different Kinds of Air_, vol. i. p. 181, and vol. ii. p. 238. Second edition, 1776. CHAPTER V THE DISCOVERY OF ARGON With the advent of Lavoisier’s system of representing the phenomena of combustion, and the expression in his terms of the various changes resulting in air when metals are oxidised, and when carbonaceous substances burn, the investigation of air was abandoned. It was no longer regarded as a mysterious element, possessed of “chaotic” properties, but was held to be a mixture of oxygen, nitrogen, and small quantities of carbon dioxide and water vapour, together with a trace of ammonia. More exact determinations of the proportion between its oxygen and so-called nitrogen than Cavendish had made by the nitric oxide method were carried out in 1804 by Gay-Lussac and Humboldt, by explosion with measured quantities of hydrogen, according to the method suggested by Volta; and they concluded, from a large number of analyses made on specimens collected in all weathers and from various localities, that 100 volumes of air contained 21 volumes of oxygen and 79 volumes of nitrogen. These experiments, too, led Gay-Lussac to the conviction that oxygen and hydrogen unite to form water in the exact proportion of one volume of the former to two volumes of the latter; and he published, some years later, accounts of numerous experiments of the same kind, as the result of which he found that, when two gases combine or react with each other, they do so in some simple number of volumes; for example, one to one, one to two, or one to three.