Professor Thomas "Tommy" Gold, who has died aged 84, was the initiator, the pragmatist and the persuader among the trio of young Cambridge scientists who turned cosmology upside down in the 1950s by proposing their controversial and comforting "steady state" hypothesis of the universe. This held centre stage for several years, with Fred Hoyle as its underpinning cosmological philosopher, Hermann Bondi in mathematical support, and Tommy Gold as its extrovert propagandist.

Gold, some years younger than Hoyle, was the natural athlete of the trio, both academically and physically. He could leap easily from engineering to physiology, from physiology to cosmology and on to almost any other speciality. Closed academic cliques feared him. Throughout his life he would dive into new territory to open up problems unseen by others - in biophysics, astrophysics, space engineering, or geophysics.

Controversy followed him everywhere. Possessing profound scientific intuition and open-minded rigour, he usually ended up challenging the cherished assumptions of others and, to the discomfiture of the scientific establishment, often found them wanting. His stature and influence were international.

The "steady state" trio were regarded as mavericks in the 1950s although, among other things, Bondi later became chief scientific adviser to the Ministry of Defence. As a group they first worked together on Admiralty radar research in 1942. Before this, however, Gold had met and befriended Bondi in the internment camps in Britain and Canada where both had ended up - with many other highly expert and loyal academic refugees from Hitler - as "enemy aliens" during the 1940 panic about fifth columnists.

When internment came, Gold was studying engineering at Trinity College Cambridge, while Bondi was doing mathematics and physics. Both came from Vienna. Gold's athletic physique matched his academic agility. He was a good climber and an expert skier (as probably befits anyone who came to Cambridge via education at Zuoz College, Switzerland), but naturally practical and highly gifted in other kinds of sport. He was therefore exactly the kind of well set up, handsome young man who, at Cambridge, was able to live life to the full.

Internment hit him hard. He was enjoying himself and well on his way to his tripos. But a darker reason for distress was that, with difficulty, he had accepted that his whole academic future lay, not in his homeland Austria, nor in Germany, but in Britain or the United States. To be locked up by the hand that seemed to be protecting him was unexpected and disconcerting. However, like many who were swept indiscriminately into camps, he believed that the mistake would soon be rectified.

In the event this took over 15 months and turned out to be a wearing and dispiriting process. But on the first night of internment, in a bare army barracks in Bury St Edmunds, he met Hermann Bondi, his fellow Austrian and Trinity student. They had not met before, even though their parents had known each other in Vienna. Their friendship changed their lives.

In the camps Bondi kept spirits up by giving ad-hoc lectures on various aspects of mathematics, and Gold soon realised that his own mathematics were unimpressive. At first infuriated by Bondi's natural skills - and his ability to dismiss Gold's hard-worked and mountainous calculations at a glance by spotting errors of scale - Gold grappled seriously with his limitations and won.

His personal philosophy was that intelligence is not specific. If you are very good at one thing, he would say, then you can be very good at everything else. Like the rest of Gold's life, this triumph over mathematics demonstrated that, in his case, the philosophy was true.

Released from internment, he took his degree and, at the request of Hoyle and Bondi and with (eventual) official approval, joined them in secret Admiralty research into problems of radar ground clutter. In the way of wartime boffins, they worked as a group, generally in remote locations, and largely fending for themselves.

For two years they shared a farm cottage near Dunsfold, Surrey, where Gold naturally took command of practical things, like finding a daily and cooking. Hoyle visited during the week and the radar work was demanding. But off-duty hours were dominated by intense and wide-ranging scientific discussion.

Gold emerged from the cold comfort of this extended wartime seminar aware of a host of new problems in astrophysics and cosmology and much better equipped to investigate them. It turned out that the electron dynamics of the magnetron, at the heart of radar, has similarities to the dynamics of stellar accretion. Hence it related to the theory of matter dispersed throughout space, to gravitational accretion and to hypotheses put forward before the war by Hoyle and Raymond Lyttleton. But it was Gold who first suggested that, whatever the turbulence and violence of galaxies or stellar systems, the energy balance of the universe would remain stable if matter were being continuously created and destroyed in equal amounts.

It was many years before this comforting and rather God-like idea succumbed to the Big Bang, although the steady-state theory was still reverberating gently in 1980, when Cornell University held a world level symposium in Gold's honour, the contributions to which were later published as a collective festschrift.

In the introduction to the book, Professor Edwin Saltpeter, who was studying electrodynamics at Cambridge in the late 1940s, recalls that at this time Gold had switched from the Cavendish Laboratory to the Medical Research Council's physiology laboratory, where he was working on a resonance hypothesis for human hearing.

Gold's approach to research was awe inspiring, says Saltpeter. It somehow managed to combine three incompatible ingredients: willingness to question any basic principle; the application of an engineer's ability to analyse complex systems; and an interest in detailed evidence even if it were not quantitative. "Are servomechanisms generally important in physiology?" Gold asked. "Were Beethoven's musicality and his particular kind of deafness related to the small change from a sharply resonant amplifier to an oscillator?" It is now recognised that feedback mechanisms play an important role and, although Tommy Gold's papers are seldom read, physiologists pay more attention to such mechanisms than they did. They may owe far more to Gold's two years of research than they realise.

In the 1950s, Gold switched back to astronomy, becoming chief assistant at the Royal Greenwich Observatory, where he raised a host of uncomfortable questions about stellar dynamics and produced a complex mathematical model, which became known as the "Gold-Hoyle hot universe".

Although superseded, this is still highly relevant to some aspects of modern x-ray astronomy. Yet, even though highly productive, Gold felt limited and out of place in the narrow and introspective academic confines of Britain. He therefore sought wider horizons. In 1956, he was offered and took the chair of astronomy at Harvard and never looked back.

He made an extraordinary series of contributions across the spectrum of planetary and astronomical sciences, being swept on to various US national committees and becoming a much sought-after NASA consultant. In 1959, he took the directorship of a new centre for radio-physics and space research at Cornell University, a context within which his extrovert originality had great freedom and where he remained for the rest of his life, becoming emeritus in 1981.

One of the most dramatic demonstrations of his genius was the speed and rigour with which, in 1968-69, he showed that the "pulsars", just discovered by the radio astronomers Antony Hewish and Jocelyn Bell, working under Sir Martin Ryle in Cambridge, must contain rotating neutron stars. This revealed huge new vistas of possibility, for if neutron stars exist in a galaxy, then, as Dennis Sciama later wrote, it is only a short step to accepting that black holes also exist. Gold opened the door for Hawking.

He also generated many controversies. In the 60s, on the run-up to the manned space programme and a possible lunar landing, there was much confused debate about the nature of the surface of the moon. Was it hard rock or was there a deep layer of fine dust? If the moon lander and its astronauts had to cope with dust layers that were metres thick, then designers needed to know, and know quickly.

By making use of evidence from microimpacts, moon cratering, electrostatic fields, and various other tools, Gold made a prediction the astronaut's boots would sink in no more than three centimetres. Within the range of possibilities, this turned out to be very close to the truth. But his critical and popular approach had infuriated other experts. He spoke and wrote about "moon dust" instead of "the lunar regolith", and quickly came under attack for being a centimetre or two wrong.

Then, in the late 70s and early 80s, when the world was taking serious stock of its energy resources, Gold pointed out that some old, deep and theoretically exhausted gas boreholes were still producing methane at a low but constant rate. Isotopic dating suggested that a large proportion of this gas was very old.

Gold suggested that we might be seeing primeval methane, trapped during the formation of the planet, but continuously rising from the deep interior of the earth. His calculations suggested that the volume might be prodigious and hence of extreme importance. Further, this rising gas could be routed to - and trapped in - major fault structures, and therefore a factor that could both trigger earthquakes and render them predictable.

These hypotheses, cutting directly across the received wisdoms of narrow fields of science in which Gold had no recognised expertise, infuriated some. Small, deep, experimental boreholes, put down in the 80s by the Swedish government to test Gold's deep gas hypothesis, yielded only a small volume of gas, but it seemed to be ancient methane and it continues to flow. Gold later altered his hypothesis to propose a "deep, hot biosphere" of methane-producing organisms and has been proved resoundingly right.

In the 80s and 90s Gold became increasingly disenchanted with the structure, achievements and scientific credibility of NASA. He drew sharp attention to the loss of high quality scientists and engineers, to the incestuous nature of many scientific contracts, to the financial pressures being used to suppress criticism and to the tragic technical mismanagement of some major projects, like the Hubble telescope. He was very unhappy and unable to answer his own question: is it possible that institutional corruption has become so pervasive that NASA can no longer hold together a body of scientists and engineers of the calibre required for their ambitious plans? To be sure, NASA's sails were trimmed. But only time will bring an answer.

Throughout his life, Gold was boisterous, amusing and direct. He was also one of the most entertaining academic speakers of his time, on either side of the Atlantic. Quite apart from his rapier-like critical wit, he was able to keep lay listeners fascinated and delighted by the unexpected mathematical elegance of everyday things, like the number of dimples in a golf ball.

Not surprisingly, he was also a great family man. He married twice: to Merle Tuberg when at Cambridge in 1947, by whom he had three daughters; and to Carvel Beyer in 1972, by whom he had one daughter. Cosmology may be full of eternal question marks, he once said, but life is here and now. That was Tommy Gold.

· Thomas Gold, scientist, born May 22 1920; died June 22 2004

· This obituary has been revised and updated since the writer's own death