Such are the class of effects on which the modern discovery of the composition of water has been based. The merit of that discovery has been shared between the celebrated English chemist, Cavendish, and the not less celebrated French chemist, Lavoisier, the chief merit, however, being ascribed to the former.

We shall now briefly state the facts which led to this discovery, with their dates, which will necessarily show the share which Watt had in it.

When pure hydrogen gas is burned in an atmosphere of common air, the process which takes place is now known to be nothing more than the chemical combination of the hydrogen with eight times its own weight of oxygen taken from the atmosphere, and the product of the combustion is a quantity of water nine times the weight of the hydrogen consumed. In the year 1776, Macquer, a well-known chemist of that day, having held a saucer of white porcelain over a flame of hydrogen which was burning at the mouth of a bottle, observed that no smoke was produced and no soot deposited on the saucer. On the other hand, he found that after the lapse of some time drops of a clear pellucid liquid were perceptible on the saucer: this liquid he submitted to analysis, and found it to be pure water. Macquer mentioned this fact without comment or inference. It did not occur to him that the water thus produced upon the saucer was a substance which contained the hydrogen, which disappeared upon combustion from the bottle.

On the 18th of April, 1781, Mr. Warltire addressed a letter to Dr. Priestley, dated Birmingham, which letter is published in Dr. Priestley's Experiments on Air, printed at Birmingham in 1781, in which Warltire informs Priestley that he had fired a mixture of hydrogen and common air in close glass vessels, and that, although previously to firing the mixture the vessels were clean and dry, a dewy deposit was [Pg306] observed afterwards on their sides. In fact, water was present which was not present before.

The mixture was in this case fired by passing an electric spark through the vessel; and it is now known that the effect produced was the combination of the hydrogen, which formed part of the mixture of airs in the vessel with the oxygen, which also formed part of the same mixture.

It appears, from expressions in Warltire's letter, that the same experiment had been previously made by Priestley, and the same result observed by him.

The inference deduced from this by Warltire, and apparently acquiesced in by Priestley, was, that whenever hydrogen was fired in atmospheric air, the moisture, which is always more or less sustained in the latter, was deposited; but neither of these chemists perceived the real cause of the production of the water.

In the beginning of 1783, and not later than the 21st of April, this experiment of Warltire and Priestley was repeated by Cavendish, with this difference, that, instead of exploding the mixture of hydrogen and common air, Cavendish exploded a mixture of hydrogen and oxygen. He observed that water was present after the explosion, but inferred nothing.

In a published paper dated April, 1783, Priestley announced a further and most important result of his experiments. This was, that in examining the weight of water produced by the explosion of a mixture of oxygen and hydrogen, that weight was found to be precisely equal to the sum of the weights of the two gases, which disappeared in the process.

Immediately on observing this, Priestley, being then, as has been already stated, Watt's near neighbour, communicated to the latter what he had observed; upon which Watt immediately, viz. by a letter dated the 26th of the same month, declared that the inevitable consequence which followed from Priestley's observations was, that water was a substance compounded of oxygen and hydrogen deprived of [Pg307] a quantity of heat which was previously latent in them.[27] The letter containing this inference was communicated immediately by Priestley to Sir Joseph Banks, then President of the Royal Society, to be laid before that body; and it is accordingly printed with its proper date in the 74th volume of the Philosophical Transactions.