Take some germanium and two bits of goldfoil...

We have recently been bombarded with books and articles about the great figures of the mathematics behind modern computers: Charles Babbage, Alan Turing, John von Neumann and so on - a colourful lot providing a fascinating set of stories (as told, for instance, in George Dyson's Darwin Among the Machines.) This new book, Crystal Fire, argues that the revolution leading to the information age was ignited by quite a different set of events, centred on the invention of the semiconductor amplifier, dubbed the transistor. The authors, Lillian Hoddeson, the preeminent historian of semiconductor physics, and physics populariser Michael Riordan, focus on the immediate postwar decades but review the earlier history of semiconductor physics, describing the gradual infiltration of quantum theory into what had been primarily of interest to engineers and amateur radio enthusiasts.

Mathematical proofs of principles are all very well, but practical necessity is the true mother of invention. Turing's machines that broke the Enigma code involved rooms full of machinery and squadrons of young women assistants; von Neumann's Eniac computer and its compeers barely outperformed hand calculators at routine jobs such as ballistics and were massive and unreliable, especially because of the heat, power drain, and unreliability of the necessary banks of vacuum tubes and their incandescent cathodes. AT&T had 100,000,000 telephones, each of which had to be available through a massive network of switches to each other, as well as undersea cables and radio transmission links, all of which cried out for something smaller, more reliable and less wasteful of power than vacuum tubes. The fascinating story of how this need was satisfied and of the resulting technological explosion has never been told so accessibly as in Crystal Fire.

The central figures are Walter Brattain, John Bardeen and Bill Shockley, the three scientists at Bell Telephone Laboratories who, in December 1947, produced a remarkably effective solid-state amplifier from a tiny block of very pure germanium and two bits of gold foil. Nine years later, to the month, with their home bases now dispersed across the United States - Shockley in Silicon Valley, Bardeen at the University of Illinois about to solve the 46-year-old problem of superconductivity and Brattain still at Bell - the trio were reunited in Stockholm for a Nobel ceremony.

After briefly glancing at the moment of discovery the book takes us through the diverse childhoods of each of these three: Brattain's wild-west upbringing on a ranch, Shockley as the only son of an older mining entrepreneur and his intellectual bride and Bardeen as the remarkably bright and favoured son of a comfortable professorial family in Madison, Wisconsin. Following a run-through of the early development of the quantum theory of solids, we pick up on these three lives at a time when each is gravitating towards the study of electrons in metals and semiconductors. Brattain had been a member of the Bell staff since 1929, working successfully on early semiconductor rectifiers and taking advantage of slow Depression-era conditions to study quantum theory with colleagues. Bardeen and Shockley were students in two of the earliest centres of solid-state quantum physics, E. P. Wigner's group at Princeton and John Slater's at the Massachusetts Institute of Technology. Shockley was hired by Bell in 1936, an enlightened decision by M. J. Kelly, adding the first quantum theorist to the Bell staff just before the announcement of Bell's first Nobel prize, awarded in 1937 to C. J. Davisson, for decisive experimental proof of the wave nature of the electron.

Semiconductor physics was transformed during the war years by the massive use of semiconductor rectifiers in microwave radar systems, which led to impressive improvements in materials purity and in understanding, especially of the elemental semiconductors silicon and germanium (some key aspects of which, we are told, were not shared with Bell's wartime collaborators in the military and other civilian laboratories). But in the spring of 1945, as physicists elsewhere were frantically winding up preparations for the first atom bombs, Bell management under Kelly was preparing a reorganisation and redirection of effort towards basic research and, specifically, towards Shockley's and Kelly's pet project of a semiconductor amplifier. Bardeen was among the first postwar hirees, delighted to escape from his war work on antisubmarine measures. A unique constellation of stars was being rapidly assembled, made easier by the fact that most universities and industrial labs of the time were all but unaware of the strides the new discipline had taken. I would have liked some discussion here of the wider explosion in solid-state physics brought on by such developments as microwave electronics and neutron scattering.

In their first two years the little group assembled by Shockley worked cooperatively with remarkable dedication, culminating in the "Christmas present'' of 1947 with which Crystal Fire opens. Here, in the book's most revelatory and carefully researched passages, it is detailed how within less than a month Shockley had begun building the basis for a totally independent series of patents. Soon Bardeen and Brattain found themselves essentially excluded from the expanded groups working on newer and more practical versions of the original device. Shockley now becomes a sort of tragic hero. In the course of contributing massively and brilliantly to the outflowing of technology which first elaborated on and then transformed time and again the original device, Shockley's self-destructive demons repeatedly blocked his path to the recognition and riches he fully deserved. Shockley was frustrated at the degree of credit given to Bardeen and Brattain rather than to himself, and, as time went on, at the fact that the legally acceptable primary patent was in the name of Bardeen and Brattain and was not an earlier one of his. This was entirely correct - the new idea of carrier injection, which is the deep principle on which the first decade of transistors worked, was Brattain's discovery. Much later, Shockley's idea of the "field effect" became practical in a new system involving an "inversion layer'', but again this basic concept stemmed from early joint work. The corporate culture of Bell Laboratories, at least in that period, saw the managerial role as fostering and nurturing the creativity of staff; internally the successes of subordinates were credited to the manager. Managers did not compete with staff. This was not Shockley's way. After two years in his group, Bardeen pointedly resigned; and, when, perhaps even more damagingly, Shockley's cavalier treatment of Gordon Teal, the chemist who developed the first modern semiconductor crystal growing method, led Teal to leave for an obscure Texas firm, later Texas Instruments, it became clear that Shockley did not fit the Bell mould. In the meantime, for half a dozen years, Shockley's output "was nothing short of phenomenal'', and he monopolised the expanded experimental effort in the research department, working out his stream of ideas.

Bell licensed their basic patents to a large number of other firms, and the process of technical diffusion that was to foster Texas Instruments, Sony, and Silicon Valley had begun. AT&T was large enough to profit enormously from any technological improvements others achieved, so long as it could retain the right to use whatever was discovered. As a Bell executive is quoted as saying: "When you cast bread upon the waters, sometimes it comes back angelcake.'' This was the period when, in the hands of Texas Instruments and a tiny Japanese company that renamed itself Sony, the moniker "transistor'' came into worldwide use to mean a pocket radio, an immediate hit around the world. The transistor radio created the first mass market tapped by the device, but not the last.

In 1955 Shockley himself joined the exodus from Bell. Financed by Arnold Beckman, he organised Shockley Semiconductors, the first of the Silicon Valley firms, in Palo Alto. Here he received news of the Nobel prize and was toasted by the crack team he had assembled; and here, less than a year later, most of the same team resigned to form the highly successful Fairchild Semiconductors - for much the same reasons (and a few extra) that had led to the problems at Bell: Shockley's refusal to accept the creative ideas of his brilliant coworkers.

In the final chapter, the authors bring us to the end of the 1950s, roughly the first decade of the transistor, with the appearance of the integrated circuit and the Mosfet, the field-effect device that powers today's computers. Gordon Moore proposed "Moore's Law'' of the continued exponential growth of anything connected with the semiconductor industry: number of transistors per chip, computing speed and power, you name it. The information age was upon us.

A brief epilogue discusses the subsequent careers of the three heroes. Shockley, after a disastrous car accident, took up the "race and intelligence'' cause for which he is probably best remembered, disrupting National Academy meetings with single-minded persistence, bending every ear he could reach, and publishing wherever he could. He took a professorship at Stanford, and eventually old colleagues found him a sinecure at Bell outside of the research department. John Bardeen went from strength to strength, earning a second Nobel prize and profiting from close ties with Xerox and Sony; he lived simply and continued active research until his death. Walter Brattain continued pure research on semiconductor surfaces at Bell until retirement, when he returned to his old college, Whitman.

The account of these early days, and of these three men in particular, can hardly be faulted. I was hired by Shockley in 1948 and so was among the first to experience the rapid slide from admiration and gratitude to the disillusion that was so crippling. I can only add that, from a fellow-theorist's point of view, Shockley had, during the first stages of a problem, the quickest, most brilliant and most incisive mind I have known. His scope was broad: he found and sold to the military the appropriate special applications of the first generations of transistors, and, among all the early semiconductor men, he was the one who foresaw their application to computers. But if his "try simplest cases'' mantra failed to work, that was it: when a problem required rethinking, when the first approach failed, he seemed incapable of shifting gears and trying again. In this he and Bardeen could hardly have differed more. It was not that John was a patient man, but he was creatively stubborn. He had, in addition to brilliance, the persistence and judgement to encourage others to help solve problems cooperatively and to recognise that when one line failed one had to look deeper.

Finally I would like to take exception to a brief remark in the epilogue. Bardeen's departing speech at Bell was called "Transistoritis'', and, if there is such a disease, Riordan and Hoddeson seem to have caught it. They conclude that "in the '60s Bell Laboratories began to lose its innovative edge". One may say that if this is so the news has not reached Stockholm; in both 1996 and 1997 one of the three Nobel winners in physics had carried out a substantial fraction of his rewarded research at Bell after 1970, unrelated to transistors.

The 1960s began with the launch of Telstar, the brainchild of John Pierce (who coined "transistor") and continued, inter alia, with the discoveries of the high-field superconducting magnet, the light-emitting diode, the Josephson effect, the semiconductor laser and the development of fibre-optic communication.

What happened at Bell was probably inevitable. It would have been more destructive for AT&T to try to keep hold of even an appreciable fraction of the gigantic industries it had spawned than to restrict itself to its primary job of communications. But the result was a weakness, not in innovation but in the ability to transfer and market technology, a hardening of commercial and managerial arteries that made it increasingly difficult to respond to changed conditions, both political and economic, in the 1980s. It is worth noting that long after near-meltdowns of many of its Silicon Valley competitors, Bell Labs is experiencing a rejuvenation, independent of the sclerotic AT&T, and is still hiring brilliant young scientists. To be fair, the telling of this later story is clearly more than one can expect in a book of finite weight.

Having been closely involved with Crystal Fire's three protagonists, I found the book especially hard to put down. But anyone who is curious about the origins of modern technology and likes a cracking good story should enjoy it just as much.

Philip W. Anderson, Nobel laureate, is professor of physics, Princeton University, US.