Home > About the Academy > Biographical memoirs

BIOGRAPHICAL MEMOIRS

Edward George Bowen 1911-1991

Introduction Early years The war years Ground radar Airborne radar The Tizard mission The Australian years The radiophysics laboratory, Sydney Post war research Navigational aids Cloud seeding and rainfall The radiotelescope at Parkes The Anglo-Australian telescope Sport Personal Honours and awards References

By R. Hanbury Brown, Harry C. Minnett and Frederick W.G. White

This memoir was originally published in Historical Records of Australian Science, vol.9, no.2, 1992. It also appeared in Biographical Memoirs of Fellows of the Royal Society of London, 1992.

Introduction

Edward George Bowen was one of the most dynamic and influential of the wartime generation of British physicists. Having completed his doctorate under Professor E.V. Appleton at King's College, London, he was recruited by Robert Watson-Watt in 1935 and played an important part in the early development of radar in Britain. He went to the United States with the Tizard Mission in 1940 and helped to initiate the tremendous enterprise that marked the evolution of microwave radar as a fighting weapon in the war. He was invited to join the CSIR(O) in Australia in 1943 and became the Chief of the Division of Radiophysics in Sydney. There he encouraged the greatest research effort that emerged from the war – the new science of radioastronomy – and brought about the construction of the 210ft radio telescope at Parkes, New South Wales. Following the initiation of cloud and rain physics by Langmuir and Schaefer in the United States, he mounted a remarkable effort to improve the rainfall in dry Australia which began in 1947 and continued after he retired in 1971. Throughout his Australian career, he remained a devoted Welshman, rejoicing in the name of 'Taffy'. He had a strong and independent view of his science which occasionally involved conflicting views with others, but this was balanced by an enthusiastic and engaging manner which won him many friends.

Early years

Edward George Bowen was born on 14 January 1911 in the village of Cockett near Swansea, Wales, to George Bowen and Ellen Ann (née Owen). He was the youngest of four children: Gwladys, Richard, Olwen and Edward George. Both their grandfathers had served apprenticeships on clipper ships sailing around the Horn to Pacific ports in Chile and Peru, there to load ore for the busy refineries of Swansea. George Bowen himself was a steelworker in a Swansea tinplate works, where he folded and flattened red-hot plates into the thin sheet steel needed, a task which required considerable skill and strength. He satisfied a love of music as the organist in the Congregational Chapel in nearby Sketty.

Edward Bowen had a keen mind and, at an early age, developed a lively interest in radio and also in sport, particularly cricket. At the primary school in Sketty, he won a scholarship in 1922 to the Municipal Secondary School in central Swansea. His senior years there coincided with the onset of bleak economic times in South Wales, but fortunately he was successful in again winning a scholarship which enabled him to enter Swansea University College. At first Edward's intention was to concentrate on chemistry, his top subject, but he soon changed to physics and related subjects, a decision he never regretted. He graduated with a First-Class Honours degree in 1930 and went on to post-graduate research on X-rays and the structure of alloys under the direction of the Senior Lecturer, Dr W. Morris Jones, and Professor J.V. Evans, an excellent teacher and physicist. This work earned him an MSc in 1931.

At the University he had met his future wife Enid Vesta Williams from nearby Neath, who graduated in geology and became a science teacher. They were later to marry (in 1938) and bring up a family of three sons: Edward, David and John.

The war years

Ground radar

It was Professor E.J. Evans who, recognising Bowen's intense interest in radio, arranged for him to take a PhD in the Physics Department of King's College (London University) under the direction of Professor E.V. Appleton. As part of his research, Bowen spent a large part of 1933 and 1934 working with a cathode-ray direction finder at the Radio Research Station at Slough and it was there that he was noticed by R.A. Watson-Watt and so came to play an important part in the early history of radar.

The first significant event in that early history was the proposal by H.E. Wimperis, then Director of Research at the Air Ministry, that a Committee for the Scientific Study of Air Defence should be established under the chairmanship of H.T. Tizard. Prior to the first meeting of that committee on 28 January 1935, Wimperis enquired from the Superintendent of the Radio Research Station (Watson-Watt) whether it would be possible to incapacitate an enemy aircraft or its crew by an intense beam of radio waves, or in more popular language by a 'death ray'. In two memoranda Watson-Watt showed that such a 'death ray' was impracticable, but made the immensely valuable suggestion that radio waves might be used to detect, rather than destroy, enemy aircraft.

Following a successful demonstration in February 1935 of the reflection of radio waves by an aircraft, the development of radar went ahead, and on 13 May 1935 a team of five people set out from Slough for Orfordness. Their ostensible purpose was to do ionospheric research but their real purpose was kept secret: it was to set up an experimental ground radar.

Bowen, now aged 24, was one of that team; he had been recruited by Watson-Watt as a Junior Scientific Officer. While the two senior members (A.F. Wilkins and L.H. Bainbridge-Bell) took care of the antennas and the receiver, Bowen's job was to assemble a transmitter from a miscellaneous collection of parts collected together in a hurry at Slough. Before the end of May he had the transmitter working, and by using 5,000 volts on the anodes of a pair of NT46 valves he persuaded them to produce a power output of about 20 kilowatts at 6MHz with a pulse width of 25 microseconds. In the course of the next few months he increased the anode voltage to over 10,000 volts, well beyond the rated limits, and managed to raise the pulse power to over 100 kilowatts.

The first detection of an aircraft, so Bowen (1987) claims, was made on 17 June 1935 when a clear radar echo was detected from a Scapa flying boat at a range of 17 miles. This was only the beginning; many improvements, such as shorter working wavelengths, larger antennas, greater transmitter power, and systems for measuring the height and direction of the target were soon introduced, ,and by early 1936 aircraft were being detected at ranges of up to 100 miles.

The success of this work prompted the decision to start work on a chain of radar stations (CH) to give warning of enemy aircraft approaching the coast, and in December 1935 the funds were made available for five stations covering the approaches to London. This ambitious project made it urgently necessary to enlarge the small team at Orfordness and to establish the programme on a more suitable site. The Air Ministry bought a large and isolated country house, Bawdsey Manor, into which the original team, including Bowen, moved in March 1936.

Towards the end of 1935 Watson-Watt decided that when the move to Bawdsey Manor took place Wilkins would take responsibility for the chain of radar stations and that Bowen, at his own request, would tackle the highly speculative – and at that time unique – venture of putting radar in an aircraft. As part of the deal, Bowen was to remain responsible for his original transmitter which would be left at Orfordness, unused but untouched, while a new transmitter was constructed at Bawdsey Manor.

In the event it proved a wise decision to leave Bowen's old transmitter untouched. The first major Air Exercise to demonstrate the use of radar in air defence was held in September 1936 using large numbers of aircraft and the new radar station at Bawdsey Manor. It was watched not only by members of Tizard's committee but also by important members of the Air Ministry and the RAF notably the Commander in Chief of Fighter Command, Sir Hugh Dowding. The first day of the Exercise was an absolute shambles; the incoming aircraft were not detected until they were so close to the coast that their engines could be heard – a sound locator would have done just as well. Urgent enquiries showed that the new transmitter at the Manor was not putting out enough power.

Bowen helped to save the day in two ways. Before a disgruntled Sir Hugh Dowding returned to London, Bowen gave him an impromptu demonstration of an experimental radar, built as part of the airborne radar programme, which was detecting the aircraft engaged in the Exercise at ranges of up to 50 miles. This, so Bowen (1987) tells us cheered Dowding up immensely. Bowen then travelled to Orford with one assistant (A.G. Touch) and, working all night, made his original transmitter work satisfactorily in time for the Air Exercise on the morning of the second day. The old transmitter at Orford held the fort until the new transmitter at Bawdsey was put right.

The rest of the Exercise went reasonably well and the plans for the construction of a chain of coastal stations survived; which was just as well, otherwise they might not have been ready to play a vital part in the Battle of Britain four years later.

Airborne radar

The problems of installing a radar in an aircraft which Bowen faced in the spring of 1936 were, to put it mildly, challenging. The principal application envisaged for airborne radar was night interception and at that stage the principal problems were not operational but technical; it is easy to underestimate how difficult they looked in 1936. The most obvious difficulty was to reduce the size and weight of the equipment; the existing ground radars would fill a small house, weighed several tons and took many kilowatts of power. Bowen decided that a viable airborne radar should not exceed 200 lbs in weight, 8 cubic feet in volume and 500 watts in power consumption and that, to reduce the aerodynamic drag of the antenna, the operating wavelength would have to be about one metre – a very short wavelength in those days.

These targets were very difficult to meet. In those days most radio components were large, heavy and unsuitable for use in the extremes of vibration, temperature and atmospheric pressure met with in military aircraft. The aircraft power supply was DC, variable in voltage and very limited in capacity. There were a number of other troublesome problems; for example, there were no solid dielectric cables to connect the radar equipment to the antennas. But the greatest difficulty of all was to generate enough power at short wavelengths in a transmitter that could be carried in an aircraft.

Over the next few years, Bowen and his group tackled and solved most of these problems. To take two important examples, in 1938, with the help of Metropolitan Vickers, he solved the problem of the power supply in aircraft by introducing an engine-driven alternator which gave an 80 volt, 1,000 Hz, voltage-stabilized supply. In 1939 he encouraged ICI to produce the first radio-frequency cables with solid polythene dielectric, a most important advance.

Faced with the difficulty of fitting a sufficiently powerful transmitter into an aircraft, Bowen's first move was to leave it on the ground and carry only the receiver and indicator in the air. He erected a powerful (30 Kw) 6.7 metre transmitter on the roof of Bawdsey Manor and installed a receiver and cathode-ray indicator in a Heyford aircraft with a simple half-wave dipole strung between the wheels. Flying from Martlesham Heath in the autumn of 1936, he detected aircraft at ranges of up to 12 miles.

This hybrid system had the advantages that the transmitter could be large and powerful and that, unlike later metre-wave airborne radars, its maximum range was not limited to the height of the aircraft above the ground; nevertheless it had the obvious limitation that the range of the target aircraft, as seen by the fighter, was only correct when the fighter was in a direct line between the transmitter and the target. Although Bowen argued hard for its further development, he failed to persuade Watson-Watt; so he dropped it and pressed on with the construction of a complete airborne radar.

In early 1937 he acquired some Western Electric 316A valves that were capable of delivering a pulse power of a few hundred watts at a wavelength of about one metre. A complete radar system, using these valves, was built at a wavelength of 1.25m and installed in an Anson. On 17 August 1937 it was tested in the air by two of Bowen's group, A.G. Touch and K.A. Wood; although they detected no aircraft, they obtained clear echoes from ships off the coast at Felixstowe at ranges of two to three miles. Following this flight the performance was greatly improved by increasing the wavelength to 1.5m, which subsequently became the standard wavelength for metre-wave airborne radar.

In September 1937, hearing that an exercise was planned during which Coastal Command would search for the Fleet, Bowen gave a dramatic, uninvited, demonstration of the application of radar to aerial reconnaissance. Together with KA. Wood he used the experimental 1.5m radar to search for the Fleet in the North Sea under conditions of low visibility and, much to the astonishment of the Navy and Coastal Command, he found the aircraft carrier Courageous, the battleship Rodney and the cruiser Southampton. It was during this flight that they detected radar echoes from the aircraft of Courageous – the first detection of an aircraft by a complete airborne radar. This demonstration was, so Bowen (1987) tells us, 'a landmark in the history of airborne radar'; it was followed by many demonstrations to senior officers of the RAF.

The airborne radar group now had two major projects, the detection of ships (ASV – Air to Surface Vessels) and the interception of aircraft (AI – Aircraft Interception). Although there were many other applications in Bowen's lively and fertile mind, there was never enough time to explore them properly. He did, however, manage to experiment briefly with the use of airborne radar to detect features on the ground such as towns and coastlines, to detect falling bombs in a scheme to attack bomber aircraft from above, and as an aid to navigation in which the contours of the ground beneath an aircraft were compared with a map.

ASV (Air to Surface Vessels)

During 1938, most of the work of Bowen's small group was devoted to the development of improved components for both AI and ASV, and to the design of a practical system of ASV. The principal question was whether the radar should scan the sea for ships by looking forward, sideways or all round. The three modes required different antennas and displays.

The design of the forward-looking mode was technically the simplest and was fairly well established by mid-1938; as we shall see later, this was the first form of ASV to be adopted by the RAF.

To test the sideways mode, Bowen had two six-element Yagi antennas fitted to an Anson so as to project a beam at right angles to the direction of flight. Using a photographic recorder to record the returns from objects scanned by this beam, he demonstrated the system to the Services by showing them 'radar pictures' of ships of the Home Fleet as they passed from Spithead to Portland in May 1938.

To test the all-round-looking mode, Bowen arranged to fit a rotating dipole to an Anson and to display the signals on a cathode-ray tube using what is now called a B-scan. Although one of his group (P.L. Waters) made this system work, its maximum range was unsatisfactory, probably due to losses in the rotating joint.

Following extensive tests and demonstrations, in which Bowen played a major part, it was eventually decided that the first ASV system in service would be forward-looking. In this system the power from the aircraft transmitter was radiated in a wide beam forward, and the returns from the target were received on two simple antennas mounted on each side of the aircraft to give overlapping beams in the forward direction. The receiver was connected in rapid sequence to these two antennas by a fast rotating switch, and the signals were displayed as deflections to the right and left of a vertical timebase on a single cathode-ray tube.

The first installation of ASV MkI was made in a Hudson aircraft in December 1939. It could detect a 10,000 tonne ship at a range of about 20 miles and coastlines at 30 to 40 miles. About 300 sets were made and were fitted in Hudsons and Sunderlands of Coastal Command. In practice its main use was to aid navigation, not to find enemy shipping; it helped patrols to rendezvous with convoys (The Battle of the Atlantic [1946]), provided navigational assistance by detecting coastlines and, more popularly, helped aircraft to return to base using transponder beacons.

When war was declared in September 1939, Bawdsey Research Station, now called the Air Ministry Research Establishment (AMRE) was 'evacuated' to Dundee and the airborne group to Perth aerodrome. Bowen was then faced with the extremely awkward problem of carrying on the development of airborne radar at an aerodrome which had neither laboratory space nor adequate hangars. That did not last long; in late October his group was moved to 32 Maintenance Unit at St Athan which, although it had adequate hangars, was too far from Dundee and a most unsuitable place in which to do laboratory work. However the main task of Bowen's group at St Athan was to help the RAF to fit radars to their aircraft, and in doing this they were entirely successful; within a few months, aircraft were being fitted with AI or ASV at the rate of about one per day. It is very much to Bowen's credit that this was achieved in such difficult circumstances.

One of the first things that Bowen did at St Athan, in response to an urgent enquiry from Admiral Somerville, was to try to detect a submarine by radar. In the first week of December 1939, Bowen and I (RHB) carried out flight trials using ASV MkI in a Hudson to look for submarine L27 in the Solent. On the first flight at 1,000 feet we detected the submarine in a fairly rough sea at a range of 3 miles; on a subsequent flight at 6,000 feet, with a calmer sea, we detected it at a range of up to 6 miles. In our report on these trials we pointed out that although these ranges were short, they had been obtained with simple dipole antennas and could be doubled by using high gain directional antennas in a sideways-looking system.

Following these results it was agreed to introduce a sideways-looking system (Long Range ASV, LRASV) for anti-submarine patrol and, as a start, Bowen arranged that a Whitley should be fitted with high gain directional antennas. The first Whitleys with LRASV went into service at Aldergrove in December 1940. At a height of 2,000 feet they could detect coastlines at about 60 miles, a 10,000 tonne ship at 40 miles, a destroyer at 20 miles and a submarine at 8 miles; at 5,000 feet their range on a submarine increased to between 10 and 15 miles.

The engineering of the equipment, to make it more rugged and reliable, was carried out at the Royal Aircraft Establishment (RAE) under the supervision of a senior member of Bowen's original group (A.G. Touch). The set which they developed (ASV MkII) was produced in far greater numbers than MkI; in the UK alone, 6,000 sets were made, and many thousands were produced in the USA and Canada. It was fitted to patrol and reconnaissance aircraft all over the world and used in anti-submarine patrols, anti-shipping strikes, convoy escort and many other duties. Its principal value was in the first phase of the Battle of the Atlantic when the Germans were using the captured French ports to give their U-boats easy access to the Atlantic. In April 1941 Coastal Command was operating anti-submarine patrols with about 110 aircraft fitted with ASV MkII, and the use of radar by these aircraft increased the daylight sightings of submarines significantly. More importantly, it made it possible to attack submarines at night as they travelled on the surface; in 1941-1942 over 90 per cent of night attacks were made as the result of ASV contact. However very few of these attacks were lethal until the introduction in mid-1942 of a powerful searchlight (Leigh Light) that illuminated the submarine. The combination of this light with ASV MkII was so effective that the submarines tended to submerge by night and surface by day, thereby increasing their destruction by daytime patrols. This satisfactory state of affairs lasted for a few months until the Germans introduced a listening receiver – Metox – which warned the submarine of the approach of an ASV-equipped aircraft so that it could dive. As far as metre-wave ASV was concerned the introduction of this listening device marked the end of the first phase of he Battle of the Atlantic; the second phase was taken up by centimetre-wave radar.

AI (Air Interception)

Most of the early development of AI and ASV was done in parallel and they were able to share many of the same components. Nevertheless AI had its own peculiar problems; for example, the components had to work at higher altitudes, making it, among other things, more difficult to design a transmitter. Also AI is more complicated than ASV because it must guide the fighter to its target in three dimensions, and the range and relative direction of a fast moving target must be presented to the operator immediately and simply. Furthermore it must bring the night fighter so close to the target that the pilot can identify it visually before opening fire.

In early 1939 Bowen, together with his group, decided that the first AI radar would measure the relative direction of the target by four antennas mounted on the fighter; two 'azimuth antennas' would give overlapping lobes in the azimuth plane and two 'elevation antennas' would give overlapping lobes in elevation. Failing to devise a simple display on a single cathode-ray tube, Bowen had to accept that there must be two tubes and a separate radar operator. One tube would display the signals from the elevation antennas and the other the signals from the azimuth antennas. As in ASV, the signals would be distributed by a fast rotating switch. Following some tests on night vision made at the R.A.E. Bowen decided that AI must have a minimum range of 1,000 feet.

The first complete installation of AI was flown in a Fairey Battle on 9 June 1939. It gave a maximum range of 12,000 feet with a Harrow as a target; the minimum range was about 1,000 feet and in mock interceptions the display seemed easy to use. A week or so later Bowen gave the Commander in Chief of Fighter Command (Sir Hugh Dowding) a successful demonstration; within a few weeks the airborne group was committed to fitting AI into 30 Blenheims for trials by 25 Squadron at Northolt.

This programme at Northolt was premature; not only was the AI equipment inadequately engineered, no proper provision was made for training and maintenance. Nevertheless the trials did expose one important fact; they showed that in order to make a successful interception with AI, it was essential to control the path of the fighter with a precision that could not be achieved by the existing system of fighter control. In November 1939 I pointed this out in a memorandum written to Bowen from Northolt and suggested that a special radar should be developed for fighter control. Bowen immediately forwarded my memorandum to the Superintendent of AMRE at Dundee, adding the excellent suggestion that the solution to the problem was to build a radar with a narrow rotating beam, a 'Radio Lighthouse'. Bowen had in fact suggested such a radar in July 1938, but not specifically for the control of night fighters. Unhappily that suggestion was turned down and Sir Robert Watson-Watt (1957) tells us that the failure to follow it up may have been one of his greatest mistakes in the development of radar. However, this time it was followed up; a radar with a narrow rotating beam and plan-position-indicator was developed by AMRE and the first Ground Control Radar (GCI) was delivered to the RAF in October 1940.

While at St Athan, Bowen's group developed an improved version of AI (MkIII) and helped to fit it into the Blenheims of various night-fighter squadrons. Although the maximum range of AI MkIII was regarded as adequate, its minimum range was about 1,000 feet and this became the subject of considerable friction between Bowen and the main establishment at Dundee. AI was not proving successful in the hands of the RAF and the Superintendent (A.P. Rowe) and his Deputy (W.B. Lewis) had been persuaded that this was largely due to the minimum range being too great; it must, they insisted, be reduced as a matter of urgency. Bowen disagreed profoundly; he was convinced that the minimum range had nothing to do with the shortcomings of AI in service and that the 1,000 feet achieved by AI MkIII was adequate operationally. To Bowen's intense annoyance Lewis, acting on a request from Rowe, started a programme of experimental work on AI MkIII at Dundee (Lovell 1988). Fortunately one of the things Lewis did was to enlist the help of EMI; in due course A.D. Blumlein and his colleagues at EMI produced an excellent transmitter modulator that reduced the minimum range to 500 feet and was subsequently incorporated in AI MkIV.

Nevertheless it is likely that in this controversy Bowen was right. The principal technical defect in AI MkIII was later shown by tests at Fighter Interception Unit to be a squint in the antenna system of the Blenheim which was cured by changing from horizontal to vertical polarization (Hanbury Brown 1991). The minimum range was probably not a serious defect; in an assessment of the combat reports of fighter pilots in 1940 and 1941, Bowen found that the median range at which enemy aircraft were tracked by A1 and then seen visually was between 1,200 and 1,500 feet (Lovell 1988). Furthermore the subsequent success of AI in 1941 suggests that two operational factors contributed to the failure of AI MkIII, the inadequate speed and armament of the Blenheim and the absence of GCI (Ground Control Radar).

This controversy about minimum range is only worth mentioning because it greatly exacerbated the strained relations between Bowen and A.P. Rowe, which had never been good ever since Rowe had succeeded Watson-Watt as Superintendent at Bawdsey in 1938 and appointed Lewis as his Deputy. As his senior staff Rowe had inherited Wilkins and Bowen, old colleagues of Watson-Watt who had pioneered radar, and as Bowen (1987) tells us, Rowe 'never came to terms with them'. The separation between Dundee and St Athan was a further strain and the friction with Lewis over the question of minimum range was, from Bowen's point of view, the last straw. When in May 1940 the main radar establishment (AMRE) moved from Dundee to Worth Matravers and the airborne group left St Athan to rejoin them, Bowen ceased to take an active part in their work and, as we shall see, he was soon to leave for the USA. [AMRE became TRE (Telecommunications Research Establishment) in November 1940.]

The final engineering of AI MkIII was undertaken at the RAE and introduced into service as AI MkIV in Blenheims and Beaufighters in the autumn of 1940. AI MkIV did everything which Bowen had originally visualized for a metre-wave AI set. It was the last and vital link in an elaborate and successful system of hunting enemy bombers that involved the coastal CH stations, inland GCI radars, radar beacons and transponders, VHF radio and AI-equipped Beaufighters with their Hispano cannons. In the hands of a skilled crew, AI MkIV was remarkably effective, and in the heavy night raids of 1941 the AI-equipped fighter proved to be the principal weapon of air defence at night (Douglas 1948); thus in May 1941 over 100 enemy aircraft were definitely shot down at night using AI compared with 30 by anti-aircraft guns.

The Tizard mission

In August 1940 Bowen left the UK as one of seven members of a Mission, led by Sir Henry Tizard, to disclose recent British technical advances to the USA and Canada. Bowen's job was to tell them all about British radar He took with him not only information on all existing and projected equipment, but also an early sample of the cavity magnetron, the essential and highly secret key to the development of centimetre-wave radar that had just been invented by J.T. Randall and H.A.H. Boot at Birmingham University.

Following discussions with the Tizard Mission, the US made the important decision that the development of metre-wavelength radar should be the responsibility of the Armed Services, and that the development at centimetre wavelengths should be the responsibility of a special Microwave Committee of which Dr Alfred Loomis was appointed Chairman.

As far as metre-wave radar was concerned, Bowen, together with other members of the Mission, visited the various laboratories of the Armed Services telling them about developments in the UK; in particular he told them about airborne radar and arranged for demonstrations of ASV MkII, AI Mk IV and IFF (Identification Friend or Foe) equipment in the air. However most of his considerable energy and enthusiasm was devoted to helping them develop centimetre-wave radar. Ever since the days of Bawdsey Manor he had urged that work should be done on shorter and shorter wavelengths so that radars could use narrow beams; an airborne radar, for example, might use a narrow beam to eliminate the returns from the ground that limited the maximum range of AI at metre-waves.

With remarkable speed the Microwave Committee set up a special laboratory, the Radiation Laboratory at MIT, for the development of centimetre-wave radar, and Bowen collaborated closely with them on their programme. His advice was particularly valuable in the early stages; for example, he wrote the first draft specification for the development of their 10cm AI.

So successful was the programme at the Radiation Laboratory that the first experimental airborne 10cm radar was tested in a Douglas B18, with Bowen on board, on 27 March 1941, only seven months after the Tizard Mission had arrived in the USA. Their first 10cm AI (SCR720), accompanied by Bowen, was demonstrated in the UK in August 1941 and later became known as AI Mk IX.

In the course of the next year the Radiation Laboratory grew in size and soon became the most important and productive radar laboratory in the USA; by the end of the war the staff numbered about 4,000.

The Tizard Mission, in which Bowen played such a large part, was highly successful. It drew the attention of the Americans to the importance of radar as a weapon of war, introduced them to airborne radar, accelerated the development of centimetre-wave radar by giving them the cavity magnetron and, owing much to Bowen, helped them to set up the highly successful Radiation Laboratory.

The Australian years

The radiophysics laboratory, Sydney

In the closing months of 1943 one of us (White) was in the USA and, when visiting the Radiation Laboratory at MIT, met Bowen again for the first time since King's College, London. Bowen seemed to be at a loose end. His work in the USA was virtually finished and the invasion of Europe by the Allies was imminent. In Australia, the Radiophysics Laboratory was still hard at work helping the Australian and American forces in the Pacific. It was proposed to Sir David Rivett Chief Executive Officer of CSIR, that an offer be made to Bowen to come to Australia to join the Radiophysics Laboratory. Rivett agreed and Bowen arrived in Sydney on 1 January 1944. In his book Radar Days Bowen tells how he consulted Tizard and received the reply: 'They seem to need help in Australia. Go there my man.' He flew by US Air Force through Hawaii, Canton Island and Noumea to Sydney, a route well known to many Australians.

When Bowen arrived in Sydney, security conditions on radar information were gradually being lifted. CSIR was planning the return to peacetime work and within a year Fred White, who had been Chief of the Radiophysics Laboratory, had joined the Executive Committee in Melbourne. This was a period of great change; the Japanese surrendered in August 1945 after the atomic bombs had been dropped, and all hostilities in the Pacific ceased. In May 1946, when John Briton who had succeeded White returned to industry, Bowen was appointed Chief of the Division of Radiophysics. One of his first actions was to organise and edit A Text Book of Radar, a collective work by the staff of the Laboratory.

Radar was still unknown to most Australians and Bowen could now talk freely about the exciting secret effort that had helped to win the war for Britain and her allies. His first paper in Australia was a general account of 'Radar in War' (Aust. Jour. of Science, 1945, 8, 33-37) in which he spoke with personal authority of the way the Royal Navy had frustrated the U-boat attack on civilian shipping. He drew a moral from the extraordinary assimilation of civilian scientists, 'in grey bags and green jackets', by the fighting forces of England, in contrast to the rigid military control of scientific warfare by the Germans and the Japanese. The 'boffin' was everywhere in evidence and accepted amongst the military men. Bowen addressed the Institution of Radio Engineers on the historical development of radar, its military uses and its potential peacetime applications to civil aviation, marine navigation and surveying.

Post war research

With the cessation of the war, the skilled staff of the Division began to look around for work of interest to themselves and of importance to Australia. The professional staff of Radiophysics had grown to 66 by 1945, with several important newcomers recruited from the British and Australian services and Bowen was conscious of his responsibility to them. Two lines of research grew up naturally and became the predominant interests of the Division: radioastronomy and cloud and rain physics. The first grew out of the curiosity of J.L. Pawsey who repeated the observations of J.S. Hey in England on the jamming of radar receivers by radiation from the sun. Research on cloud and rain physics was started by Bowen in 1946 when I. Langmuir and V. Schaefer in the USA reported that rain could be induced by seeding clouds with dry ice. These two programmes absorbed the attention of a considerable proportion of the staff until Bowen himself retired from CSIRO in 1971.

Navigational aids

Bowen had also undertaken two other research activities. These were the pulse method of acceleration of elementary particles, with Pulley and Gooden, and more extensive work on air navigation with V.D. Burgmann. The latter resulted in the Distance Measuring Equipment (DME) that was ultimately adopted for all civil aircraft flying in Australia on internal routes.

Cloud seeding and rainfall

While many reacted cautiously to the 1946 claims by Langmuir and Schaefer that clouds could be made to rain by creating ice crystals in them, Bowen immediately saw the potential importance of the technique for dry Australia. Within months, two members of his staff had investigated the work and, on their return, had carried out a trial in eastern New South Wales using RAAF aircraft. Success was immediate. When seeded with dry ice the selected cloud reacted with spectacular changes of shape and heavy rainfall. This striking result held such promise that a systematic programme of cloud seeding was set up in February 1947 and continued for the next twenty-four years.

As little was known about the properties of clouds in Australia or the mechanisms of rainfall, Bowen initiated a vigorous research programme of cloud studies. This included not only the effects of adding ice crystals to cold clouds, but also the effect of spraying water into warm clouds which are responsible for much of the rainfall in the warmer parts of Australia. Bowen took part in the latter work himself and during 1950-1955 published papers on the theory of coalescent rainfall and directed experimental trials.

The difficulty with both these methods of stimulating rainfall was that only a few clouds could be treated on any one day and large amounts of dry ice or water were required. This limitation was overcome by the discovery, again in the USA, that tiny quantities of silver iodide smoke could be used as a seeding agent. Unlike many of his contemporaries, Bowen saw the potential for seeding large areas from the air using silver iodide burners mounted on an aircraft.

The first experiments with this method were made in 1955 over the Snowy Mountains in south-eastern Australia. The first two years were so successful, with an estimated rainfall increase of 25%, that several more regions were quickly selected. There the early indications were also successful, but in many subsequent years all areas showed a gradual decay of the induced rainfall with time. Most people would have become discouraged by such a result and given up. Bowen, however, proposed a simple explanation, based on the idea that a persistence phenomenon in the seeding process had confused the statistical analysis. Although this concept failed to win much support at the time, Bowen insisted that the next experiment (in Tasmania) should use target and control areas rather than two randomly-seeded areas, which was the method most susceptible to persistence effects. Moreover there was to be a gap of one year between seeded years.

This experiment was a success but, Bowen having retired (1971), the result was not immediately attributed to the correctness of his persistence hypothesis. Some years later Bowen reopened the question and the outcome of the ensuing debate established the persistence phenomenon as a vital factor that must be taken into account when designing and analysing a seeding experiment. Subsequent work by E K Bigg has done much to explain the detailed mechanism of the phenomenon. With the continuing success of cloud seeding work by the Australian states of Tasmania and Victoria and the recognition of the role of persistence, there appears now to be a promising future for the rain making techniques that Bowen did so much to pioneer.

Bowen's remarkable energy and enthusiasm were evident also in other programmes. He was not afraid to speculate and presented his intuitive ideas with a persuasive and engaging optimism that was either inspiring or alarming to his colleagues, depending on their views of science. Two of his well known theories about periodic rainfall variations illustrate this.

The influence of meteor showers

>From the daily rainfall records for Sydney over the period 1859 to 1952 and for stations elsewhere in New South Wales and in other countries, Bowen found well defined peaks of rainfall in January and February. These anomalies he correlated with the passage of the Earth, 30 days earlier, through specific meteor streams that orbit the sun. He suggested that the smaller particles fell through the atmosphere to cloud level in 30 days, where they induced the observed rainfall.

The apparent physical implausibility of this hypothesis attracted a wave of criticism: the number of particles was insufficient, the fall time would not be fixed, and the particles would not form ice crystals. Even the reality of the anomalies was vigorously questioned, but independent analysis showed that they were statistically significant. But Bowen was not impressed by purely statistical arguments and insisted that his staff probe crucial aspects of his hypothesis by empirical tests in clouds. Whether he was right to invoke meteor showers to explain the rainfall anomalies and if so, how they influenced clouds after a fixed time interval, has yet to be demonstrated.

Lunar effects

In 1962, following a paper published in the USA, Bowen and Adderley showed that there were similar lunar effects in the monthly rainfall records for fifty New Zealand stations with comparable magnitude and closely related phase. The reality of the effect was beyond doubt. Independent frequency analysis revealed an amplitude variation of 20% and a periodicity of 29.5307 days. The mean period between full moons is 29.5306 days.

Bowen suggested that the Moon, revolving about the Earth, could modulate the amount of meteor dust reaching the Earth, and later showed that meteor rates in both the northern and southern hemispheres varied similarly with lunar phase. He argued that the Moon could intercept the particles or alternatively could deflect them because of electrostatic charges on the Moon and particles. Modern studies by his colleague, E.K. Bigg, however, suggest that the Moon's influence on rainfall is more likely to be caused by the lunar tides in the Earth's atmosphere.

The cloud and rain physics group, under Bowen's leadership, worked in a most stimulating environment. Even his more speculative ideas sometimes drove his critics to discover truths that would otherwise have remained hidden. Over twenty-four years, the group established a high international reputation with its achievements and an impressive number of sound scientific publications.

The radiotelescope at Parkes

In the first decade following the end of the war Radiophysics established an enviable reputation in the new science of radioastronomy. It was a time of exciting discoveries and innovative ideas, a time when a new observing system could be quickly tried out. The outstanding Australian successes in this period were recognised when URSI elected to hold its 10th General Assembly in Sydney in August 1952, the first meeting of an international scientific union ever held outside Europe or the USA. But by then the era of improvised equipment was drawing to a close and the era of big science was soon to begin.

Radioastronomy now needed aerial systems with much higher resolution and able to collect more of the extremely weak signals arriving at the Earth. One approach was to develop a very large parabolic-reflector aerial and as early as 1948 Bowen had been convinced that this was the best solution. Bernard Lovell at Manchester University in 1952 was the first to set off down this path. Bowen was very conscious that the British government had funded the project at a cost far beyond the resources of Radiophysics. Nevertheless he persisted and tried to find more economical designs, but none were quite satisfactory.

During visits to the USA, where he had made many influential contacts during the war, Dr Vannevar Bush (President, Carnegie Corporation) and Dr Alfred Loomis (Trustee, Carnegie Corporation and Rockefeller Foundation) revealed that it might be possible for Bowen to build a large radio telescope in Australia with financial help from the USA. In April 1954 the Trustees of the Carnegie Corporation of New York announced a grant of $250,000 to Australia for this purpose. This generosity was returned by Bowen, in part, over the next year by his help in establishing US radio astronomy: in January 1955, he arranged for John Bolton and Gordon Stanley to be seconded to the California Institute of Technology, a move that marked the beginning of the science in California.

Bowen organised a Technical Advisory Committee (TAC) in 1955 to advise on and specify the proposed design study for the Australian telescope. The committee included two structural experts, H.A. Wills of the Aeronautical Research Laboratories, Melbourne, and J.W. Roderick, head of the Civil Engineering School of the University of Sydney.

A highly significant development occurred in mid-year when Bowen had discussions with Barnes Wallis (later Sir Barnes Wallis FRS), the famous airship and aircraft designer at Vickers Armstrong, Weybridge. Wallis revealed some innovative ideas including a 'master equatorial' for controlling the movements in equatorial coordinates of the mounting, a concept which was to become a key feature of the Parkes Telescope. The outcome was that Freeman Fox and Partners (FF&P), London, the designers of the Sydney Harbour Bridge, were selected for the studies with advice from Barnes Wallis. Harry Minnett, from the Telescope Planning Committee, was appointed as CSIRO liaison officer and radio consultant to FF&P.

Bowen was forced to turn to a number of US funding organizations in the hope of supplementing the available funds. These overtures were successful for in December 1955 the Rockefeller Foundation contributed $250,000, with an important condition that the Australian government should match this sum as well as the amounts previously received. When approached by Sir Ian Clunies Ross, Chairman of CSIRO, the Prime Minister, Robert Menzies, agreed to this proposal and also to pay for the running costs of the complete installation.

In London the senior partner of FF&P was Ralph Freeman but the telescope project was directed by Gilbert Roberts, a brilliant if somewhat idiosyncratic engineer. Later both men were knighted and Roberts was also elected to the Royal Society. Roberts' first assistant in charge of the telescope team was Michael Jeffery, an outstanding structural engineer.

The three basic questions that Bowen had posed for the consultants were: compensated or rigid reflector structure; altazimuth or equatorial mounting; telescope cost as a function of reflector size for both mountings. As Wallis had remarked, the design of a giant radio telescope to the precision required was a venture into the unknown. It was not expected that Bowen's questions would be easily settled.

The structural aspects of the study proceeded satisfactorily. A small, very rigid central hub supporting the reflector structure was adopted to encourage symmetrical deflection patterns. For a rigid steel reflector, these were found to be so promising that the investigation of complicated servo-compensated aluminium structures was ultimately abandoned as unnecessary. Roberts and Wallis intuitively preferred an altazimuth mounting because of its structural simplicity compared with an equatorial, and a compact and extremely rigid design was evolved. However, a thorough study of the feasibility and cost of the Wallis master equatorial concept and the altazimuth servo drive system would clearly be crucial to the mounting decision. Unfortunately it proved very difficult initially to interest competent firms in this task.

In October 1956, however, Grubb Parsons Ltd. agreed to develop and cost a master equatorial system. They also suggested an important innovation for sensing the error between the pointing directions of the master unit and the slave reflector axis. This idea was based on proven auto-guidance technology and was a significant advance on the untried mechanical and hydraulic system in the Wallis proposal. By that time also Minnett had proposed a servo system that avoided the stability problems arising from structural resonances, and had shown that it could accurately track astronomical sources under dynamic wind loads. These ideas were adopted by Metropolitan Vickers, who had agreed to develop and cost the drive system. FF&P were confident by early 1957 that an altazimuth mounting was the best solution.

The design study report was completed by November and Bowen asked the TAC to critically review its recommendations. After discussions with Minnett and Roberts in Sydney, the Committee agreed that the feasibility of the telescope had been established and that the design was an excellent one. From the cost-size data, a diameter of 210 ft. (64 m) was chosen early in 1958 to match the available funds. Bowen's foresight in setting up and carefully organizing the design study was a major factor in this result and avoided many pitfalls.

Following completion of the detailed design, Bowen insisted on international tenders. MAN (Maschinenfabrik Augsberg Nurnberg AG) in West Germany was successful, with Metropolitan Vickers as contractor for the servo drive systems. The offer by Askania Werke of West Berlin was accepted as sub-contractor for the master equatorial control system. The MAN contract was finalized in July 1959. By his vigorous participation in the tendering process and contract negotiations, Bowen achieved a significant improvement in earlier estimates of the completion date and cost. Some additional funding was still needed, however, and he approached the Rockefeller Foundation again. Early in December it generously approved a further $130,000 which was matched by the Australian government.

MAN proceeded with great vigour. The construction of the base tower at Parkes started in September 1959 and a trial assembly in Germany of the mounting and servo drives took place in May 1960. On-site construction of the telescope commenced in September 1960, with Jeffery as resident engineer for FF&P. That it was completed closely to schedule was a tribute not only to MAN and to FF&P's careful design work and supervision, but also to Bowen's energetic efforts throughout the project.

On 31 October 1961 the Governor General, Lord de Lisle, was invited by the CSIRO Chairman, Dr F W G White, to perform the opening ceremony; Bowen followed with a speech of thanks. The occasion was a grand affair in spite of the unusually high wind. The ceremony was attended by a large assembly of Radiophysics staff, Chiefs of CSIRO Divisions, academics, industrialists and local people.

Bowen was delighted with the performance of the new instrument. In 1963 he wrote 'It is clear from the figures that the telescope is one of superlative performance and provides both surface and pointing accuracy which is approximately double that called for in the original specifications'. The Parkes Telescope also proved timely for the US space programme. Bowen received a NASA grant for Minnett to participate in studies at the Jet Propulsion Laboratory in California for the design of a 210 ft. instrument for communicating with very distant space probes. Many of the Parkes features, including the drive and control concepts, were adopted.

John Bolton, the first Director of Parkes, initiated an intensive survey to detect radio sources and eventually listed many thousands, including many quasars. Detailed studies of hydrogen line emissions at 21cm. wavelength helped to reveal for the first time the spiral structure of our galaxy. The versatility of the instrument made possible a variety of other investigations including: ionized interstellar hydrogen, supernova remnants, polarization and magnetic fields, the discovery of new pulsars, the study of the Magellanic Clouds and remote galaxies. During the first twenty-five years of operation, over 1,000 research papers were published.

The telescope played a vital role in NASA's Apollo moon landing programme and through it the world's television audiences saw Man's first steps on the Moon. For the European Space Agency's Giotto mission to Halley's Comet, Parkes was the prime receiving centre. The telescope was linked to the NASA station at Tidbinbilla to boost the signal during the successful flight of Voyager II past Jupiter, Uranus and finally Neptune, then the most distant planet of the solar system.

Over more than a quarter century, the achievements of the Parkes Telescope have more than justified the very great efforts necessary to bring it into being. Now the major element in the Australia Telescope National Facility, it is destined to continue its scientific contributions well into the next century. There could be no more enduring monument to the vision, tenacity and energy of 'Taffy' Bowen.

The Anglo-Australian telescope

In the first decades after the War, there was much discussion about the need for a large optical telescope in the southern hemisphere. The matter was taken up formally by the Royal Society of London and the Australian Academy of Science on a joint basis towards the end of 1963. Their discussions were lengthy, and at the end of June 1965 submissions for the construction of a 160-inch (3.8m) Anglo-Australian telescope were presented to both governments. A long delay then ensued.

In the months that followed, the Australian government was non-committal on the Anglo-Australian proposal despite British pressure. A firm commitment had been delivered by W.L. Francis, the Secretary of the Science Research Council, that Britain would fund half the cost of designing and building a 3.8 m telescope.

The Australian Academy of Science asked L.G.H. Huxley and Bowen to seek an interview with the responsible Minister, Senator Gorton. Once a few matters had been clarified, they found the Minister was very much in favour of the Anglo-Australian proposal. In May he announced the agreement of the two governments to build a 3.8m optical telescope on Siding Spring Mountain near Coonabarabran, New South Wales, the site of an Australian National University observatory.

The two governments set up a Joint Policy Committee (JPC), pending the legal creation of a Board, to direct the design, construction and operation of the new telescope. Bowen with Professor O.J. Eggen and Mr K.A. Jones represented Australia and Sir Richard Woolley, Professor Hermann Bondi and Mr J.F. Hosie represented the United Kingdom. Shortly after the first meeting, Bondi accepted another post and was replaced by Professor F. Hoyle.

At the first JPC meeting in August 1967 in Canberra, it was decided to follow broadly the design of the 150-inch polar axis telescope then being planned for the Kitt Peak National Observatory (KPNO) in Arizona. Many of the new post-war technologies had been applied to radio telescopes and the time was ripe for changing some traditional optical telescope practices. R.O. Redman of the University of Cambridge and S.C.B. Gascoigne of Mount Stromlo Observatory had been appointed as permanent astronomical advisers to the project, as a link with potential users and with special responsibilities for the optics.

Redman and Gascoigne recommended that the prime focal length adopted by KPNO should be increased to f/3.3 and that a simplification should be made to the arrangements at the prime focus cage. The primary mirror blank, to be cast in a new material with a zero coefficient of temperature expansion, was ordered at once from the US supplier to take advantage of the discount offered by adding to the KPNO order. The nucleus of the Project Office was established by appointing Hermann Wehner (Mount Stromlo Observatory) and John Pope (Greenwich Observatory) with particular responsibilities for instrumentation design.

After the first meeting of the JPC, Bowen set out to implement a number of his proposals which had been agreed. He organized the secondment as Project Manager of M.H. Jeffery, chief assistant to Sir Gilbert Roberts at Freeman Fox and Partners, London, during the design of the Parkes Telescope and resident engineer during its construction. H.C. Minnett from Bowen's Radiophysics Division, together with a British counterpart, R.L. Ford of the Royal Radar Establishment, were appointed as consultants on drive and control. Bowen also recruited D. Cunliffe from the CSIRO Division of Mechanical Engineering as the Executive Officer of the Project Office.

As a result of Bowen's initiatives, Jeffery was able to attend the next JPC meeting in London in March 1968. Minnett and Ford, after investigations in the UK and USA at the end of 1967, had produced a drive and control report for the JPC recommending that the setting accuracy target should be 10 arcsec; that the telescope should be controlled by a computer system operating through servo drives; and that a modern auto-guidance device should be developed to relieve the astronomer of this chore. Bowen later proposed that Maston Beard should be seconded from Radiophysics for a major role in this work. He also supported a proposal by Minnett and Jeffery that traditional worm drives should be replaced by high-precision spur gearing with symmetrical anti-backlash drives as in radio telescope technology.

When Jeffrey died suddenly from a heart attack early in September 1969, Bowen's reaction was typically swift. Within days he had arranged for Minnett to be seconded to Canberra as Acting Project Manager and had obtained the agreement of Freeman Fox to make a study of a serious problem in the design of the declination bearings. The engineer selected was Colin R. Blackwell who had worked on the design studies for the Parkes Telescope.

At the August 1970 meeting of the JPC in London, Blackwell was able to recommend a satisfactory solution to the bearing problem. By then Freeman Fox's role had been expanded by Bowen and the Board to include responsibility for the supervision of the complete mounting contract on behalf of the Project Office. The AAT inter-government agreement specified that tenders had to be called on an international basis and Bowen was insistent on the observance of this proviso. In October 1970 the contracts for both the mounting and the drive and control system were awarded to a Japanese company that offered specially favourable terms designed to win the work.

In February 1971, following the passage of the necessary legislation through the Australian Parliament, the JPC was dissolved and its members were appointed to the AAT Board, with Bowen as Chairman and Hoyle as Deputy. The management and operation of the telescope now became a critical and divisive issue. It was not settled until April 1972 when Bowen supported the British stand for a Scientific Director responsible only to the Board. Within a year Bowen was appointed to the post of Science Counsellor at the Australian Embassy in Washington, D.C., and he therefore had to resign as Chairman of the Board. Hoyle took his place and Paul Wild was appointed as a new Australian representative.

Bowen had successfully guided the project through the complex years when the design of the telescope was evolving and had overcome other problems of great difficulty to arrive at last at a highly satisfactory result. In the words of Hoyle: 'there is no doubt that a large share of it (the credit) must go to Taffy Bowen. Without him the telescope would have been only a shadow of what it was eventually to become'.

The telescope was officially inaugurated on 16 October 1974 in the presence of H.R.H. Prince Charles. When operations commenced in 1975, the telescope was accepted as a technological tour de force. In the words of Gascoigne: 'The mounting and the optics were clearly of the highest standard, but what created the real impression was the computer control system, which was comprehensive, versatile and efficient to a degree beyond anything previously contemplated'.

Sport

Bowen had an enduring love of cricket, which he began playing while he was a youth in Wales. After playing for the South Wales League at Gormorton, he continued at King's College, London, and later at Felixstowe arid Martlesham. He continued his passion for cricket when he joined the Radiophysics Laboratory in Sydney.

He was also a keen sailor having started in England, but his main opportunity was in Sydney, where he became devoted to VJ's and Moths. He later bought a Yachting World boat that he raced in the Middle Harbour Yacht Club. About 1968 he was elected Rear Commodore of the Club. Later, as Science Counsellor in Washington, he lived at West River on Chesapeake Bay. There he sailed a 32 ft yacht named 'Sosie' about the Bay and with some success in local races.

Personal

Bowen's Division of Radiophysics was quite unlike others in CSIRO. It had been founded in 1939, in the utmost secrecy, to work on wartime radar for the fighting services. Several scientists spent the war in the fighting services and when demobilised came to Australia and joined in the remarkable post-war researches that Bowen headed. Two such men were John Paul Wild and John Bolton. the former, who succeeded Bowen as Chief of Division, has this interesting analysis of Bowen as his predecessor:

I was one of several young research scientists who joined the CSIR Radiophysics Laboratory in the early post-war years. The Chief, Taffy Bowen, was firmly in command: young, confident, cheerful and breezy, always optimistic and giving the impression that he knew exactly where he was going. He had supervised the transition of the laboratory from its wartime programme of military radar to its new peacetime policy.