air was another. Cavendish was not content with noting that this latter air went pop when ignited; he reported careful measurements showing that it was 8700 times lighter than water and capable of holding ‘1/9 its weight of moisture’.
This kind of detail reveals the way Cavendish thought about experiments. His laboratory, housed within the grounds of his ample townhouse in Great Marlborough Street, near Piccadilly in London, was filled with measuring devices. The caricature presented by Wilson, and more or less uncritically repeated ever since, shows Cavendish as a calculating machine, obsessed with quantification; but the fact was that he understood this was now the only reliable way to do science. We’ve seen that van Helmont recognized the value of measurement in the seventeenth century; but Cavendish’s vision penetrated further than that. He understood the meaning of accuracy and precision, and realised that all experiments have a finite and unavoidable margin of error. He estimated the accuracy of his determinations, making distinctions between the errors introduced by the experimenter and the limitations of the instrumentation. To reduce such sources of error, he would repeat experiments and take averages of the results. And he would quote numerical results only to the appropriate number of significant figures. The great French scientist Pierre-Simon Laplace, who pioneered statistical techniques for handling errors in experiment (and of whom more later), remarked to Blagden that Cavendish’s work was conducted with the ‘precision and finesse that distinguish that excellent physicist’. This is arguably Cavendish’s greatest contribution to experimental science: an attention to numerical detail that keeps the experimenters’ claims in proportion to what their methods justify.
And numbers have power. By putting numbers on the low weight of this vapour relative to common air, Cavendish excited speculations about whether it might enable a man to ‘fly’ by means of the buoyancy of a balloon filled with it. And so it did: the physicist Jacques Charles took to the air in 1783 in Paris, prompting Antoine Lavoisier to scale up his method of producing ‘inflammable air’ while Joseph Banks covered up his nationalistic chagrin with sniffy remarks about the flighty French.
In the early 1780s, Cavendish decided to explore ‘the diminution which common air is well known to suffer by all the various ways in which it is phlogisticated’. In other words, he was keen to examine the process that Warltire and Priestley had described, in which common air is reduced in volume by igniting it with inflammable air (which might or might not be phlogiston itself). Thus he was not, in a sense, proposing to do anything new; rather, he saw that sometimes an experiment yields its secrets only when you start to look at the details. ‘As the experiment seemed likely to throw great light on the subject I had in view’, he explained in the report of his studies, presented to the Royal Society in 1784, ‘I thought it well worth examining more closely’.
Anatomy of an explosion
Like the others before him, Cavendish made inflammable air by dissolving zinc or iron with acids, and he set off the detonation with a spark. ‘The bulk of the air remaining after the explosion’, he wrote,
is then very little more than four-fifths of the common air employed; so that as common air cannot be reduced to a much less bulk than that by any method of phlogistication, we may safely conclude that when they are mixed in this proportion, and exploded, almost all the inflammable air, and about one-fifth part of the common air, lose their elasticity, and are condensed into the dew which lines the glass.
Every detail was carefully checked out; nothing was taken for granted. The dew, he said ‘had no taste nor smell, and . . . left no sensible sediment when evaporated to dryness; neither did it yield any pungent smell during the evaporation; in short, it seemed pure water’. In some experiments he noticed that the explosion produced a little ‘sooty matter’, but he concluded that this was probably a residue from the putty (‘luting’) with which the glass apparatus was sealed; and indeed ‘in another experiment, in which it was contrived so that the luting should not be much heated, scarce any sooty tinge could be perceived’.
Was the dew truly pure water? Cavendish found in some initial experiments that it was in fact slightly acidic, and he spent long hours tracking down where the acid came from. Although he did not put it quite this way himself, the acidity stems from reactions between oxygen in the air and a little of the nitrogen that makes up the ‘inert’ four-fifths of the remaining gas, creating nitrogen oxides, which are acidic when dissolved in water. Such pursuit of anomalies was one reason why Cavendish was so slow to publish his findings, which he did some three years after the experiments were begun. But the fact is that Cavendish was in no hurry in any case. For him, publication was not the objective, and he seems blithely unconcerned about securing any claims to priority. He seems to have adopted the approach advocated by his colleague William Heberden, who said that the happiest writer wrote ‘always with a view to publishing, though without ever doing so’.
So what was this experiment telling him? In retrospect it seems obvious: inflammable air and the ‘active’ constituent of common air – or hydrogen and oxygen, as Lavoisier was already calling these substances – unite to form water. But Cavendish was at that stage still in thrall to the phlogiston theory, and so things were by no means so clear to him. He ascertained that the lost one-fifth of common air could be identified as the ‘dephlogisticated air’ that Priestley had described: indeed, when he used this air instead of common air, it was all used up if ignited with twice its volume of inflammable air. But this interpretation meant that phlogiston had to appear somewhere in the balance. Cavendish concluded that dephlogisticated air was ‘in reality nothing but dephlogisticated water, or water deprived of its phlogiston’. In other words, water was not being made from two constitutive parts but was appearing through the combination of phlogiston-poor water with the phlogiston contained in inflammable air. Alternatively, if the inflammable air were not phlogiston itself, then it was ‘phlogisticated water’, or ‘water united to phlogiston’. The phlogiston effectively cancelled out:
[water – phlogiston] + [water + phlogiston] = water
How we are to understand Cavendish’s conclusions has been a matter of great debate, because to some extent the issue of whether or not he made a genuine ‘discovery’ about the nature of water hinges on it. The truth is that there is nothing in what Cavendish wrote about his experiment that indicates unambiguously that he questioned the elemental status of water. That is to say, it remains unclear whether he decided that water somehow pre-existed in his airs and was simply being condensed in the explosion (which is pretty evidently what Priestley believed) or whether he had some inkling that water was being created from its constituents in a chemical process. Traditional historical accounts of Cavendish’s experiment tend to imply that he made more or less the correct interpretation, even if he couched it in the archaic terms of phlogiston theory. But historian of science David Philip Miller has argued fairly persuasively that Cavendish’s thoughts were closer to Priestley’s. In any event, for an explicit and decisive statement of water’s compound nature, we must look across the English Channel.
A new kind of chemistry
In Paris, Antoine Lavoisier was on the same path: familiar with Macquer’s work, he too was looking more closely at what happened when the two airs were united. But he had a different hypothesis. In the mid-1770s he had concluded that Priestley’s dephlogisticated air was in fact a substance in its own right: an element, which he proposed to call oxygen. The name means ‘acid-former’, for Lavoisier had the (misguided) notion that this element was the ‘principle of acidity’, the substance that creates all acids.
Cavendish knew of Lavoisier’s oxygen, but he did not much care for it. He pointed out, quite correctly, that there was at least one acid – marine acid, now called hydrochloric acid – that did not appear to contain this putative element. (Lavoisier admitted in 1783 that there were some difficulties in that regard which he was still working on.) But while some of Cavendish’s contemporaries, Priestley in particular, were trenchantly opposed to Lavoisier’s theory because of an innate conservatism, Cavendish was more pragmatic – he argued simply
