It is 0:47 on the morning of 25th of July 1943 and somewhere over the North Sea, Joan Curran is not on a bomber. She is at her desk in the Telecommunications Research Establishment in Malvern, England and she has been asleep for several hours. But above Hamburg, 746 aircraft of RAF Bomber Command are carrying a secret that she spent two years designing.
Bundles of black paper strips coated in aluminum foil, each strip cut to exactly 27 cm in length and 2 cm wide packed into parcels of 2,000. At precisely 0:52, the leading aircraft of the bomber stream crosses the German coast and the first wireless operator tears open the first parcel and pushes it through a flare chute into the freezing air above the Reich.
The strips tumble outwards caught in the slipstream and bloom into a shimmering metallic cloud that hangs and drifts and to the eye of a Würzburg radar operator 90 km below looks exactly like an aircraft. Then another parcel follows, then another. Within 4 minutes, the sky above Hamburg is filling with ghost airplanes, hundreds of them, spreading and multiplying faster than any controller can track.
The radar screens across the whole of Luftgau 11 light up as though a thousand bombers are approaching from every direction at once. For 3 hours and 11 minutes, 63 flak batteries across Hamburg, some 450 guns in total, will fire into clouds, into false signals, into nothing. The anti-aircraft shells will arc upward at coordinates that contain only falling aluminum foil.
Joan Curran’s strips will save more than 100 aircraft that night. The Germans will call it the catastrophe of Hamburg. The RAF will call it Operation Gomorrah. But before any of that, there was a physicist who could see what a radio wave could not. Joan Elizabeth Josephine Curran was born in Swansea in 1916, the daughter of a school teacher.
She read physics at Girton College, Cambridge at a time when women could attend lectures and sit examinations, but were not permitted to receive degrees. Cambridge would not formally grant degrees to women until 1948. She graduated in 1938 with results that would have earned a first-class degree had the university been willing to award one.
And she joined the Cavendish Laboratory as a research assistant. When the war came, she was recruited into the Air Defence Research and Development Establishment, which soon merged into the larger Telecommunications Research Establishment at Worth Matravers in Dorset, later relocated to Malvern in Worcestershire. By 1941, she was working directly on one of the most consequential problems in British air power, how to blind the German radar network.
The Kammhuber Line, named after the Luftwaffe General Josef Kammhuber who designed it, was by 1942 one of the most sophisticated air defense systems ever constructed. It ran from Denmark south through the Netherlands, Belgium, and into France, a chain of overlapping defensive zones, each roughly 32 km wide and 20 km deep. Each zone contained a Freya long-range search radar with a range of approximately 160 km, two Würzburg precision radars with ranges of 50 km, a searchlight battery, and a single night fighter directed by ground controllers
who tracked both the German interceptor and the incoming British aircraft simultaneously on their radar screens. When the radar return of the interceptor and the bomber overlapped on the cathode ray tube, the controller knew the pilot was in firing range. By 1942, this system was destroying British bombers at a rate that threatened to end the strategic bombing campaign entirely.
In the summer of that year, RAF Bomber Command was losing approximately 5% of its aircraft on every raid over the Ruhr Valley, a rate that compounded over 20 raids would statistically eliminate an entire aircrew without them completing their tour of duty. Amongst senior RAF planners, there was genuine fear that the losses were unsustainable.
The solution in principle was not new. The idea of dropping metallic strips to confuse radar had been floating through British scientific circles since at least 1937, when radar pioneer Robert Watson-Watt had noted in a memorandum that a cloud of metal foil would reflect radio waves in a manner indistinguishable from a large aircraft.
By 1941, the Germans had a similar idea under development, which they called Düppel. But the theoretical insight and the practical weapon are entirely different things, and it was Curran who bridged the gap. Working through 1941 and 1942, she determined the precise physical dimensions that would make a strip of aluminum-coated paper resonate at the frequency of the Würzburg radar, 570 MHz, giving a wavelength of approximately 53 cm.
The optimal strip length was half that wavelength, 27 cm. The optimal width was 2 cm, narrow enough to drift slowly in the air and maintain a persistent cloud. She then calculated how many strips would be needed per parcel and how frequently parcels would need to be ejected from the aircraft to maintain a continuous curtain of false returns.
The answer was one parcel every minute per aircraft. Each parcel containing 2,000 strips weighing a total of 1 lb. The Telecommunications Research Establishment code named the project Window. What followed was one of the stranger delays in British military history. By April 1942, Window had been tested successfully in trials over the Dorset coast.
The strips worked exactly as Curran had calculated. They saturated Würzburg radar screens with ghost returns that functionally impossible to distinguish from real aircraft. The air staff was presented with a weapon that could, in the words of one assessment, reduce Bomber Command’s losses by a third at a stroke. And then nothing happened for 15 months.
The reason was a strategic paralysis rooted in a very specific British anxiety. If Window was released over Germany, the Germans would immediately understand the principle, manufacture their own version, and deploy it against the Chain Home radar network defending Britain against the Luftwaffe. Given that British bombers were already attacking Germany, whilst Germany was not currently mounting large-scale raids on Britain, the calculus seemed to some deeply unfavorable.
Churchill himself reviewed the question multiple times. The Air Ministry debated it in committee. The Chiefs of Staff disagreed. Window sat in a drawer whilst British aircrew continued to die over the Ruhr at a rate of roughly one crew per day. If this deep dive into history is interesting to you, consider subscribing. It helps more than you might imagine.
The decision that finally unlocked Window came not from a strategic reassessment, but from a change in circumstances. By mid-1943, the Luftwaffe’s capacity to mount major offensive raids on Britain had been dramatically reduced. The Eastern Front was consuming aircraft and crews at a terrible rate. The North African campaign had ended in Axis defeat in May 1943 and Allied air superiority in the Mediterranean was being firmly established.
The threat of a German Window-enabled raid on British cities had not vanished, but it had receded sufficiently for Churchill to approve deployment on July 15th, 1943. Bomber Command was given 10 days to prepare. The first operational use would be the opening raid of Operation Gomorrah, the destruction of Hamburg. Hamburg in July 1943 was Germany’s second largest city and its most important port, home to 1.
8 million people and more than 3,000 industrial enterprises, including the Blohm und Voss shipyard where U-boats were built at a rate of approximately 30 per month. The city’s air defenses were amongst the strongest in the Reich. 54 heavy flak batteries equipped with 88 mm and 105 mm guns, 26 light flak batteries with 20 mm and 37 mm weapons, 22 searchlight batteries each with 150 cm carbon arc lamps, and three night fighter airfields within 80 km, Stade, Wittmundhafen, and Lüneburg, each housing between 24 and 36 Messerschmitt
Bf 110 or Junkers Ju 88 nightfighters. The standard German prediction for an RAF raid of 700 aircraft against Hamburg without countermeasures would have been approximately 50 to 60 aircraft lost, a catastrophic 7 to 8% attrition rate. Harris’s staff had factored in a figure close to that when planning the operation.
They did not yet know what Window would actually do. At 047 on 25th of July, the first bombers crossed the German coast near Sylt. By 052, the leading wave was releasing Window at the prescribed rate, one parcel per minute per aircraft, 2,000 strips per parcel, tumbling at a drift rate of approximately 400 m per minute in the upper air.
Within 7 minutes, Freya operators at Stade and Cuxhaven were reporting something wrong with their screens. The returns were multiplying, not in the orderly, trackable manner of a real bomber stream, but explosively, unpredictably, filling the display with returns that crawled and split and merged in patterns that had no aerodynamic explanation.
At the Würzburg stations of the 3rd Flak Corps, operators attempted to lock onto individual returns and found that each target they selected seemed immediately to multiply into six or eight. The standard ground control intercept procedure, in which a controller tracked both nightfighter and bomber on separate Würzburg screens and talked the pilot onto the target, required stable, consistent radar returns.
What the controllers were now seeing was not that. Oberleutnant Hans Soonline, a ground controller at Stade, would later record in his incident report that at 1:03 he had simultaneously 12 separate bomber returns within a 3-km radius, none of which corresponded to any aircraft his night fighter pilot could visually locate.
He described his cathode ray tube as looking like a badly stirred pot of soup. Across Hamburg, the flak batteries received firing data that was in effect random. The Würzburg-Riese, the giant Würzburg radar, had a theoretical accuracy of plus or minus 50 m in range and 0.1° in azimuth under normal conditions. Under window, it was producing firing solutions that varied by 2 to 3 km.
The 88-mm shells, each costing approximately 40 Reichsmarks and capable of a lethal blast radius of 10 m, were detonating at altitudes and positions where there were no aircraft. 450 guns fired for 3 hours and 11 minutes. German records from the 3rd Flak Korps indicate that 2,400 heavy-caliber rounds were expended during the raid.
The RAF lost 12 aircraft that night, 1.5% of the force. Normal predicted losses without window would have been at minimum 50 aircraft, likely closer to 60. The night fighter wing based at Lüneburg, JG 300, launched 22 sorties. They claimed three kills, all in the visual search phase after their ground control had broken down.
Without a functioning radar intercept, the pilots were navigating by searchlight and visual contact in a sky crowded with fires and smoke, conditions in which the German advantage in night fighter technology was effectively nullified. The immediate German response was confusion, and then within 48 hours, alarm.
At Luftflotte Reich headquarters in Berlin, General Hans Jürgen Stumpff received reports from Hamburg that described the failure of the entire radar-based defense network. His staff correctly identified the cause within hours. The metallic cloud phenomenon had been theoretically studied by German scientists as early as 1939, but countermeasures took time that Hamburg did not have.
Bomber Command returned on the night of July 27th, again on July 29th, and again on August 2nd, each time deploying window in the same systematic manner. By the second raid, German flak gunners had been instructed to switch to barrage fire, preset altitude zones rather than radar-guided tracking, but barrage fire against an unconfirmed target dispersed over kilometers of sky is in effect firing blind.
The second raid was the worst. 787 aircraft released incendiaries into a city that had been dry for weeks in a summer heat wave. The resulting firestorm, fed by temperatures that reached 800° C in some streets, destroyed 16,000 buildings, killed approximately 37,000 people, and created a column of superheated air that rose to 6 km and generated its own weather system, a self-sustaining cyclone of fire that consumed 21 sq km of the city center in less than 3 hours.
The term firestorm, Feuersturm, entered the German language that night. Bomber Command lost 17 aircraft on the second raid. The window strips were already falling before the bombers reached the coast. Within a week of the first Hamburg raid, German scientists at Telefunken had begun work on countermeasures. The most practical response was to retrain the Würzburg operators to distinguish the movement signature of window strips, which drifted slowly with the wind, from that of a real aircraft, which maintained a consistent course and speed. This took
time and required experienced operators. A second approach involved modifying the radar to operate at higher frequencies, outside the resonant range of 27 cm strips. But, frequency modification required hardware changes at every Würzburg installation in the Reich, more than 1,500 stations by late 1943, and manufacturing capacity was already overstretched.
The Allies, for their part, responded to the threat of German window by simply keeping pace. As German radar frequencies changed, British scientists adjusted strip dimensions to match. It became a running technical duel that would last for the remainder of the war. Joan Curran continued working at the Telecommunications Research Establishment throughout, developing refinements to window, and moving on to related work on air- borne radar jamming.
She received no public recognition during the war. Her name did not appear in press accounts of the Hamburg raids. The legacy of window extended far beyond Hamburg. In the D-Day landings of June the 6th, 1944, Operation Taxable and Operation Glimmer used window dropped by specially equipped aircraft to simulate two phantom invasion fleets, one approaching the Pas de Calais, one heading for a point north of the actual landing beaches in Normandy, convincing German radar operators that the main assault was not where it actually was. The
deception contributed to Hitler’s decision to hold the 15th Panzer Army in reserve around Calais for weeks after the real landings, waiting for an invasion that never came. In the Pacific theater, American forces adopted a version of window they called chaff and deployed it extensively from mid-1944 onwards.
The underlying technology, strips of metal resonant at radar frequencies, remained in active military use through the Korean War, the Vietnam War, and beyond. Modern military aircraft still carry chaff dispensers based on the same physical principle that Curran derived in a laboratory in Dorset in 1941. It is the morning of 26th of July, 1943, and Joan Curran reads the first preliminary report from the Hamburg operation at her desk in Malvern.
The numbers she reads are 12 aircraft lost from a force of 746. She knows what the normal number should have been. She does not record her reaction in any document that survives. She was not in the habit of recording reactions. What she does, by the account of colleagues, is set the report down, mark several figures in pencil, and return to work.
Somewhere over the North Sea, 2,000 tons of aluminum-coated paper are drifting slowly down through the summer darkness, catching the first light of dawn on their metallic surfaces, flickering like a million tiny mirrors above a city that is already burning. The radar screens across Hamburg are still full of ghost returns, echoes of nothing, the lingering signatures of strips that have done their work. The guns have stopped.
The sky they aimed at was always empty.
How One British Physicist Made Nazi Anti-Aircraft Guns Fire at Empty Sky for Three Hours
It is 0:47 on the morning of 25th of July 1943 and somewhere over the North Sea, Joan Curran is not on a bomber. She is at her desk in the Telecommunications Research Establishment in Malvern, England and she has been asleep for several hours. But above Hamburg, 746 aircraft of RAF Bomber Command are carrying a secret that she spent two years designing.
Bundles of black paper strips coated in aluminum foil, each strip cut to exactly 27 cm in length and 2 cm wide packed into parcels of 2,000. At precisely 0:52, the leading aircraft of the bomber stream crosses the German coast and the first wireless operator tears open the first parcel and pushes it through a flare chute into the freezing air above the Reich.
The strips tumble outwards caught in the slipstream and bloom into a shimmering metallic cloud that hangs and drifts and to the eye of a Würzburg radar operator 90 km below looks exactly like an aircraft. Then another parcel follows, then another. Within 4 minutes, the sky above Hamburg is filling with ghost airplanes, hundreds of them, spreading and multiplying faster than any controller can track.
The radar screens across the whole of Luftgau 11 light up as though a thousand bombers are approaching from every direction at once. For 3 hours and 11 minutes, 63 flak batteries across Hamburg, some 450 guns in total, will fire into clouds, into false signals, into nothing. The anti-aircraft shells will arc upward at coordinates that contain only falling aluminum foil.
Joan Curran’s strips will save more than 100 aircraft that night. The Germans will call it the catastrophe of Hamburg. The RAF will call it Operation Gomorrah. But before any of that, there was a physicist who could see what a radio wave could not. Joan Elizabeth Josephine Curran was born in Swansea in 1916, the daughter of a school teacher.
She read physics at Girton College, Cambridge at a time when women could attend lectures and sit examinations, but were not permitted to receive degrees. Cambridge would not formally grant degrees to women until 1948. She graduated in 1938 with results that would have earned a first-class degree had the university been willing to award one.
And she joined the Cavendish Laboratory as a research assistant. When the war came, she was recruited into the Air Defence Research and Development Establishment, which soon merged into the larger Telecommunications Research Establishment at Worth Matravers in Dorset, later relocated to Malvern in Worcestershire. By 1941, she was working directly on one of the most consequential problems in British air power, how to blind the German radar network.
The Kammhuber Line, named after the Luftwaffe General Josef Kammhuber who designed it, was by 1942 one of the most sophisticated air defense systems ever constructed. It ran from Denmark south through the Netherlands, Belgium, and into France, a chain of overlapping defensive zones, each roughly 32 km wide and 20 km deep. Each zone contained a Freya long-range search radar with a range of approximately 160 km, two Würzburg precision radars with ranges of 50 km, a searchlight battery, and a single night fighter directed by ground controllers
who tracked both the German interceptor and the incoming British aircraft simultaneously on their radar screens. When the radar return of the interceptor and the bomber overlapped on the cathode ray tube, the controller knew the pilot was in firing range. By 1942, this system was destroying British bombers at a rate that threatened to end the strategic bombing campaign entirely.
In the summer of that year, RAF Bomber Command was losing approximately 5% of its aircraft on every raid over the Ruhr Valley, a rate that compounded over 20 raids would statistically eliminate an entire aircrew without them completing their tour of duty. Amongst senior RAF planners, there was genuine fear that the losses were unsustainable.
The solution in principle was not new. The idea of dropping metallic strips to confuse radar had been floating through British scientific circles since at least 1937, when radar pioneer Robert Watson-Watt had noted in a memorandum that a cloud of metal foil would reflect radio waves in a manner indistinguishable from a large aircraft.
By 1941, the Germans had a similar idea under development, which they called Düppel. But the theoretical insight and the practical weapon are entirely different things, and it was Curran who bridged the gap. Working through 1941 and 1942, she determined the precise physical dimensions that would make a strip of aluminum-coated paper resonate at the frequency of the Würzburg radar, 570 MHz, giving a wavelength of approximately 53 cm.
The optimal strip length was half that wavelength, 27 cm. The optimal width was 2 cm, narrow enough to drift slowly in the air and maintain a persistent cloud. She then calculated how many strips would be needed per parcel and how frequently parcels would need to be ejected from the aircraft to maintain a continuous curtain of false returns.
The answer was one parcel every minute per aircraft. Each parcel containing 2,000 strips weighing a total of 1 lb. The Telecommunications Research Establishment code named the project Window. What followed was one of the stranger delays in British military history. By April 1942, Window had been tested successfully in trials over the Dorset coast.
The strips worked exactly as Curran had calculated. They saturated Würzburg radar screens with ghost returns that functionally impossible to distinguish from real aircraft. The air staff was presented with a weapon that could, in the words of one assessment, reduce Bomber Command’s losses by a third at a stroke. And then nothing happened for 15 months.
The reason was a strategic paralysis rooted in a very specific British anxiety. If Window was released over Germany, the Germans would immediately understand the principle, manufacture their own version, and deploy it against the Chain Home radar network defending Britain against the Luftwaffe. Given that British bombers were already attacking Germany, whilst Germany was not currently mounting large-scale raids on Britain, the calculus seemed to some deeply unfavorable.
Churchill himself reviewed the question multiple times. The Air Ministry debated it in committee. The Chiefs of Staff disagreed. Window sat in a drawer whilst British aircrew continued to die over the Ruhr at a rate of roughly one crew per day. If this deep dive into history is interesting to you, consider subscribing. It helps more than you might imagine.
The decision that finally unlocked Window came not from a strategic reassessment, but from a change in circumstances. By mid-1943, the Luftwaffe’s capacity to mount major offensive raids on Britain had been dramatically reduced. The Eastern Front was consuming aircraft and crews at a terrible rate. The North African campaign had ended in Axis defeat in May 1943 and Allied air superiority in the Mediterranean was being firmly established.
The threat of a German Window-enabled raid on British cities had not vanished, but it had receded sufficiently for Churchill to approve deployment on July 15th, 1943. Bomber Command was given 10 days to prepare. The first operational use would be the opening raid of Operation Gomorrah, the destruction of Hamburg. Hamburg in July 1943 was Germany’s second largest city and its most important port, home to 1.
8 million people and more than 3,000 industrial enterprises, including the Blohm und Voss shipyard where U-boats were built at a rate of approximately 30 per month. The city’s air defenses were amongst the strongest in the Reich. 54 heavy flak batteries equipped with 88 mm and 105 mm guns, 26 light flak batteries with 20 mm and 37 mm weapons, 22 searchlight batteries each with 150 cm carbon arc lamps, and three night fighter airfields within 80 km, Stade, Wittmundhafen, and Lüneburg, each housing between 24 and 36 Messerschmitt
Bf 110 or Junkers Ju 88 nightfighters. The standard German prediction for an RAF raid of 700 aircraft against Hamburg without countermeasures would have been approximately 50 to 60 aircraft lost, a catastrophic 7 to 8% attrition rate. Harris’s staff had factored in a figure close to that when planning the operation.
They did not yet know what Window would actually do. At 047 on 25th of July, the first bombers crossed the German coast near Sylt. By 052, the leading wave was releasing Window at the prescribed rate, one parcel per minute per aircraft, 2,000 strips per parcel, tumbling at a drift rate of approximately 400 m per minute in the upper air.
Within 7 minutes, Freya operators at Stade and Cuxhaven were reporting something wrong with their screens. The returns were multiplying, not in the orderly, trackable manner of a real bomber stream, but explosively, unpredictably, filling the display with returns that crawled and split and merged in patterns that had no aerodynamic explanation.
At the Würzburg stations of the 3rd Flak Corps, operators attempted to lock onto individual returns and found that each target they selected seemed immediately to multiply into six or eight. The standard ground control intercept procedure, in which a controller tracked both nightfighter and bomber on separate Würzburg screens and talked the pilot onto the target, required stable, consistent radar returns.
What the controllers were now seeing was not that. Oberleutnant Hans Soonline, a ground controller at Stade, would later record in his incident report that at 1:03 he had simultaneously 12 separate bomber returns within a 3-km radius, none of which corresponded to any aircraft his night fighter pilot could visually locate.
He described his cathode ray tube as looking like a badly stirred pot of soup. Across Hamburg, the flak batteries received firing data that was in effect random. The Würzburg-Riese, the giant Würzburg radar, had a theoretical accuracy of plus or minus 50 m in range and 0.1° in azimuth under normal conditions. Under window, it was producing firing solutions that varied by 2 to 3 km.
The 88-mm shells, each costing approximately 40 Reichsmarks and capable of a lethal blast radius of 10 m, were detonating at altitudes and positions where there were no aircraft. 450 guns fired for 3 hours and 11 minutes. German records from the 3rd Flak Korps indicate that 2,400 heavy-caliber rounds were expended during the raid.
The RAF lost 12 aircraft that night, 1.5% of the force. Normal predicted losses without window would have been at minimum 50 aircraft, likely closer to 60. The night fighter wing based at Lüneburg, JG 300, launched 22 sorties. They claimed three kills, all in the visual search phase after their ground control had broken down.
Without a functioning radar intercept, the pilots were navigating by searchlight and visual contact in a sky crowded with fires and smoke, conditions in which the German advantage in night fighter technology was effectively nullified. The immediate German response was confusion, and then within 48 hours, alarm.
At Luftflotte Reich headquarters in Berlin, General Hans Jürgen Stumpff received reports from Hamburg that described the failure of the entire radar-based defense network. His staff correctly identified the cause within hours. The metallic cloud phenomenon had been theoretically studied by German scientists as early as 1939, but countermeasures took time that Hamburg did not have.
Bomber Command returned on the night of July 27th, again on July 29th, and again on August 2nd, each time deploying window in the same systematic manner. By the second raid, German flak gunners had been instructed to switch to barrage fire, preset altitude zones rather than radar-guided tracking, but barrage fire against an unconfirmed target dispersed over kilometers of sky is in effect firing blind.
The second raid was the worst. 787 aircraft released incendiaries into a city that had been dry for weeks in a summer heat wave. The resulting firestorm, fed by temperatures that reached 800° C in some streets, destroyed 16,000 buildings, killed approximately 37,000 people, and created a column of superheated air that rose to 6 km and generated its own weather system, a self-sustaining cyclone of fire that consumed 21 sq km of the city center in less than 3 hours.
The term firestorm, Feuersturm, entered the German language that night. Bomber Command lost 17 aircraft on the second raid. The window strips were already falling before the bombers reached the coast. Within a week of the first Hamburg raid, German scientists at Telefunken had begun work on countermeasures. The most practical response was to retrain the Würzburg operators to distinguish the movement signature of window strips, which drifted slowly with the wind, from that of a real aircraft, which maintained a consistent course and speed. This took
time and required experienced operators. A second approach involved modifying the radar to operate at higher frequencies, outside the resonant range of 27 cm strips. But, frequency modification required hardware changes at every Würzburg installation in the Reich, more than 1,500 stations by late 1943, and manufacturing capacity was already overstretched.
The Allies, for their part, responded to the threat of German window by simply keeping pace. As German radar frequencies changed, British scientists adjusted strip dimensions to match. It became a running technical duel that would last for the remainder of the war. Joan Curran continued working at the Telecommunications Research Establishment throughout, developing refinements to window, and moving on to related work on air- borne radar jamming.
She received no public recognition during the war. Her name did not appear in press accounts of the Hamburg raids. The legacy of window extended far beyond Hamburg. In the D-Day landings of June the 6th, 1944, Operation Taxable and Operation Glimmer used window dropped by specially equipped aircraft to simulate two phantom invasion fleets, one approaching the Pas de Calais, one heading for a point north of the actual landing beaches in Normandy, convincing German radar operators that the main assault was not where it actually was. The
deception contributed to Hitler’s decision to hold the 15th Panzer Army in reserve around Calais for weeks after the real landings, waiting for an invasion that never came. In the Pacific theater, American forces adopted a version of window they called chaff and deployed it extensively from mid-1944 onwards.
The underlying technology, strips of metal resonant at radar frequencies, remained in active military use through the Korean War, the Vietnam War, and beyond. Modern military aircraft still carry chaff dispensers based on the same physical principle that Curran derived in a laboratory in Dorset in 1941. It is the morning of 26th of July, 1943, and Joan Curran reads the first preliminary report from the Hamburg operation at her desk in Malvern.
The numbers she reads are 12 aircraft lost from a force of 746. She knows what the normal number should have been. She does not record her reaction in any document that survives. She was not in the habit of recording reactions. What she does, by the account of colleagues, is set the report down, mark several figures in pencil, and return to work.
Somewhere over the North Sea, 2,000 tons of aluminum-coated paper are drifting slowly down through the summer darkness, catching the first light of dawn on their metallic surfaces, flickering like a million tiny mirrors above a city that is already burning. The radar screens across Hamburg are still full of ghost returns, echoes of nothing, the lingering signatures of strips that have done their work. The guns have stopped.
The sky they aimed at was always empty.
