Elegant Solutions. Philip Ball

even exist without experimentation. The Classical Greek philosophers such as Aristotle and Plato were scientists to the extent that they believed things happened because of natural mechanisms, not through the whimsy of the gods; and they were determined to root out these causes using reason and logic. But on the whole, theirs was the logic of abstract thought, which can never get you very far – because our intuitions about nature are seldom reliable. (You need only read Plato’s recommendations for making colours by mixing to realise that he probably never picked up a paint brush.) It wasn’t until Greek thought mingled with Middle Eastern artistry in the crucible of polyglot Alexandria that proto-scientists came to appreciate the value of experiment. This philosophical melting pot of Hellenistic Greece produced some of the finest ancient experimentalists, such as Hero and Archimedes.

      We have to be careful, however, what we understand by ‘experiment’ here. Today scientists use a well-designed experiment to probe and perhaps to falsify a theory, or to enable them to choose between different theoretical interpretations. Yet, until the Renaissance, it was extremely rare that an experiment would be conducted to test an idea: it was simply a way of demonstrating that you were right.

      Even so, there was a difference between actually doing the experiment and just talking about it. The Arabic alchemists of the ninth and tenth centuries appreciated that experimental science must inevitably be a quantitative science. In contrast to the qualitative theories of Aristotle, they gathered knowledge by weighing and measuring, using sophisticated instrumentation: balances, rulers and so on. Experiment creates a demand for instruments, but by the same token, instruments make new experiments inevitable.

      So the fact is that, for the Whiggish historian of science (and many practising scientists fall into this category), experiment has roots every bit as disreputable as Garcia implies. Experimental science was not a part of the natural philosophy studied in the medieval and Renaissance universities; it was something done by alchemists and magicians, by the mystical adepts of the Neoplatonic tradition. Medical doctors studied anatomy from books, and the cutting up of bodies was left to unlettered surgeons (which was why the mistakes of the Classical writers persisted for so long). Useful materials like dyes, alkalis and soap were manufactured by artisans and tradesmen. Scholars who, like the thirteenth-century Franciscan monk Roger Bacon, showed an appetite for experiment were inevitably labelled as wizards.

      Roger’s namesake Francis, nearly four centuries later, professed contempt for alchemists, magicians and their ilk. They had, he said in the Novum Organum, so far exerted ‘faint efforts’ that had met with ‘meagre success.’ But that was not because natural magic itself was a pile of nonsense; rather, it was because its proponents were hitherto mostly fools and charlatans. Henry points out that Bacon, like most of his contemporaries, ‘willingly accepted [that] astrology, natural magic and alchemy were noble and worthwhile pursuits even though, in practice, they were full of error and futility’.

      Well, so what? Why should it matter that the ‘father of experimental science’ drew inspiration from an occult tradition that today plays no role in science? Why should we care that the very concept of a scientific experiment has roots in magic and practical ‘arts’?

      I believe that bearing this in mind should help prevent us from being too narrow-minded about what we imagine an ‘experiment’ is. To define it as an enquiry into nature would be to impose a modern definition that denies a great deal of the genealogy of experimental science. I would argue that, at all times before the twentieth century, experimentation was closely linked to techne, to applied science and to the skills of the fabricator and the artisan. This perspective, moreover, is particularly important within the context of chemical science, because that discipline has a history that is in many ways quite distinct from the history of physics or biology (with their origins in natural history and the observation of nature). Some areas of what we would now deem to be the sciences of the material world, such as metallurgy, have only rather recently established firm connections with the ‘fundamental’ sciences on which we now consider them to be based. Likewise, there has been a convergence between some strands of applied chemistry – the manufacture of dyes and pigments, and of soaps and detergents, and the brewing of beverages – with ‘academic’ science only since the nineteenth century, and that was itself largely driven by the demands of industry for more reliable and versatile methods of synthesis, rather than because academia decided for itself that these ‘arts’ were worthy of intellectual effort.

      This is perhaps why chemistry is so conspicuously absent from some recent books both about the history of science and about its future prospects: it does not ‘fit’ today’s modish model of what science is. The truly bizarre result is that we now have an image of ‘science’ that is largely at odds with the way it is actually practised. Philosopher of science Joachim Schummer has estimated that there are more – many more – scientific papers published in chemistry than in any other scientific discipline. ‘Thus’, he says,

      if we want to know what our actual sciences are about, we should – from a quantitative point of view – first and foremost turn our attention to chemistry. Or, to put it in different terms, philosophies of the natural sciences that neglect chemistry should arouse our strongest suspicion.

      What’s more, the overwhelming majority of those papers report the results of experiments. ‘Chemistry’, says Schummer, ‘has always been the laboratory science per se,’

      such that still in the 19th century the term ‘laboratory’ denoted a place for experimental research in which chemical operations were performed. The chemical laboratory became the model for all the other laboratory sciences when they replaced ‘thought experiments’ by real experiments. Although chemistry is no longer the only experimental science, it is by far the biggest one and historically the model for all others. Thus, if we want to know what scientists mean by ‘experiment’, chemical papers are the right point to start with.

      Schummer points out that roughly a third of all scientists worldwide are engaged not in the experimental testing of theories, but in producing (and characterizing) new substances – in chemical synthesis. Chemistry, as the eminent French chemist Marcelin Berthelot recognized, creates its own object: it is not necessarily an inquiry into nature, but sets synthetic goals that are shaped by the considerations of the engineer, in particular by the issues of function and design. Synthetic chemistry has its own aesthetic: the ‘unnatural’ molecules that chemists try to make, while bounded by practical issues such as stability and synthetic accessibility, are ultimately no less a ‘designed’ product than motor vehicles or buildings, and as such their structure is not inevitable. This brings an added dimension to the notion of a ‘beautiful’ experiment in chemistry: the beauty need not lie in the conception or the execution, but in the product.

      Thus, it seems to me that any attempt to discuss ‘beautiful experiments’, not just in chemistry but in the whole of science, becomes a skewed endeavour if it neglects that aspect of experimental science engaged in techne, in a tradition allied to the arts and crafts, a tradition of making useful and marvellous things – the tradition, indeed, that Francis Bacon drew upon in setting out his ground-breaking plan for giving science a logical and organized structure.

      And just what is beautiful?

      Good question. Happily, the places and people and things that we find beautiful are many and varied – which means that the selection of experimental examples in this book can never be more than arbitrary. I was heartened by the fact that after I had drawn up a shortlist for the challenge set by the Royal Society of Chemistry – to identify the ten ‘most beautiful’ experiments in chemistry – I discovered that the American Chemical Society (ACS) had already conducted the same exercise a year previously, and had come to many of the same conclusions as mine. In late 2002 the ACS canvassed its members to submit proposals for the list, and the shortlist of 25 was then assessed and ranked by a panel of chemists and science historians whose combined authority exposes my own list as the scribblings of a rank amateur. At that point, perhaps, I should have just jettisoned my own efforts and adopted the ACS ‘top ten’.

      But of course, there were parts of that list with which I didn’t agree at all. While encouraged by the coincidences with my own choices, I was also stimulated to defend the differences. I ended