The British Train Line Bait That Made German Supply Locomotives Flip at High Speed

It is the 3rd of November 1942 and somewhere along the Lyon to Marseille rail line in occupied France, a man named Henri is lying face down in wet grass beside a set of track he has walked past 40 times without stopping. Tonight he stops. He reaches into the lining of his coat. His name appears in no Allied after-action report.

What appears in those reports is the result of what he does next. The night is cold. The rails are slick with November damp and smell of iron and machine oil. A freight locomotive passed this way 40 minutes ago heading south. Its boiler pushing warm air across the embankment as it went. Henri felt it from 40 m.

He has timed the patrol cycle to the minute. He has 11 minutes left. What he is holding fits in one fist. It weighs 340 g, less than a tin of boot polish. It was manufactured in a workshop in Welwyn Garden City, Hertfordshire and smuggled into France inside the false bottom of a violin case. It cost the British government under two shillings to produce in industrial quantities.

Over the next 14 months, devices exactly like this one would contribute to the destruction or severe delay of more than 200 Axis rail supply movements across occupied Europe. More disruption per unit cost than any Allied bombing campaign of the same period. Henri presses it against the rail web, listens for a moment he cannot hear, and gets back to his feet.

The problem facing the British railway sabotage effort in 1941 was not a shortage of courage. It was arithmetic. The German rail network supplying the Western and Mediterranean fronts comprised more than 80,000 km of active line. SOE had fewer than 300 trained agents operating across all of occupied Europe by the middle of that year.

Conventional sabotage, cutting lines, destroying bridges, derailing single trains by hand-placed obstacles, required exposure times of between 6 and 20 minutes per operation. The Gestapo had reduced average agent life expectancy in France to 4 months by autumn 1941. The mathematics made sustained conventional sabotage a method of destroying agents faster than it destroyed trains.

The locomotives themselves were a specific problem the RAF could not solve. Bombing the Henschel works at Kassel, which produced a significant proportion of German freight locomotives, had been attempted repeatedly from 1940 onward. Factory production continued within weeks of each raid. Bombers sent to hit marshalling yards at Ham, at Cologne, at Metz, found that German repair teams could restore primary lines within 36 hours and secondary lines within 5 days.

At the Rommel supply peak in the autumn of 1941, an estimated 40 trains per day were running the North Africa supply chain through Italian ports. The loss of 12 trains to bombing in a given fortnight barely registered. The locomotive was the problem, not the track, not the bridge, not the signal box, the locomotive.

A modern German freight engine of the Baureihe 52 class, the Kriegslokomotive, the war locomotive, weighed 87 tons, developed 1,500 horsepower, and could haul 1,800 tons of cargo at 80 km/h. Building one required 26 tons of steel, 900 man-hours of skilled labor, and a minimum of 6 weeks from foundry to first run.

Destroying one by conventional means, derailment by obstacle, demolition charge at the chassis, required so much preparation, so much explosives weight, and so much exposure time that it had been done successfully by SOE agents fewer than 20 times in 2 years. There were thousands of these machines. There seemed to be no answer that didn’t get agents killed faster than trains.

That specific answer was already being worked on in a requisitioned country house in Station Road, Welwyn Garden City, at an organization that most people who worked there were instructed to describe as a research station for the study of drainage problems. The organization was MD1. Its full designation was Military Intelligence Research Department 1, and it operated under the directorship of Major Millis Jefferis, a Royal Engineers officer of such peculiar ingenuity that Winston Churchill had personally ensured

his protection from the normal military administrative machine. Churchill called Jefferis “My Department of Wild Talents.” The men who worked for Jefferis called their output, collectively, Winston’s Toys. The official designation for one of these toys, the device Henry carries in his coat, was the locomotive fog signal Mark II.

Though the name was later changed to the track sabotage device, rail pattern, and later still was known informally, in almost every SOE circuit that used it, as the clam rail. It worked on a principle that a clever 15-year-old could grasp in 30 seconds. A standard British fog signal, the type used by railway workers to warn locomotive drivers of obstruction, the type that any rail worker across Europe would recognize on site.

Detonated when a train wheel ran over it. The sound was a sharp crack, audible at some distance, used since the 1840s as an acoustic warning system. The clam rail used that recognition. It looked to a German sentry with a torch like a standard fog signal fixed to the rail. It was in fact nothing of the kind. The clam rail consisted of an outer casing of pressed steel, 118 mm long and 62 mm wide, weighing 340 g complete.

Inside the casing sat two chambers. The first contained 60 g of Nobel 808 plastic explosive, a substance with the color and consistency of marzipan and a scent, disconcertingly, of almonds. The second chamber contained a delay fuse capsule of the type already in use across SOE’s standard toolkit, adjustable between 15 minutes and 8 hours using a simple rotary collar.

A magnet on the underside, 23 kg of pull force for a device the size of a tobacco tin, bonded the unit to the rail web or, more critically, to the underside of a locomotive’s axle housing or bogie frame. This was the key difference from every previous attempt. You did not need to destroy the track. You attach the device to the locomotive itself.

The train carried its own destruction with it. The first prototype test at Welwyn in March 1942 failed because the delay fuse proved too sensitive to vibration from the moving train. The device detonated 340 m from the test siding before the locomotive had reached operational speed, blowing a section of axle casing on an old LMS tank engine and bringing the test to a halt.

Jefferis’ team, specifically a civilian engineer named Stuart McCrae, who had been a motoring journalist before the war and who kept a running notebook of failures that he described characteristically as useful data, modified the fuse housing with an inner rubber damping sleeve in April. The second test on May 9th, 1942, worked perfectly.

The locomotive ran 11 km before the fuse initiated. The resulting axle explosion derailed all four driving wheels and buckled the bogie frame beyond field repair. If this story is new to you, a quick subscribe means you will never miss another one like it. SOE’s F section began receiving clam rails in quantity by July 1942. The distribution network ran through the Spanish border crossing at Andorra and through Lysander drops to reception committees in the Dordogne, Corrèze, and Haute-Loire departments.

The first confirmed operational use came on the 17th of August, 1942, when an agent operating within the Car Cerci in southern France attached three clam rails to a Bow Rai 52 freight locomotive at Avignon marshalling yard. The locomotive departed at 0340 hours carrying 42 wagons of artillery ammunition bound for the Italian port of Taranto.

It derailed at kilometer marker 63 on the Aix-en-Provence line at 0517 hours. Surviving German field reports, declassified from the Bundesarchiv in 1978, record the derailment as having destroyed the locomotive, 17 wagons, and blocked the main line for 31 hours. The report describes the cause as a mechanical failure of the the assembly and does not mention sabotage.

That last fact was the clam rail’s secondary weapon. Because it attached to the locomotive and detonated kilometers from any station or yard, the resulting crash resembled to exhausted German track engineers working at night with inadequate tools, a catastrophic mechanical failure. Through the autumn and winter of 1942, the Sicherheitsdienst’s rail security reports show a rising pattern of investigations into what German engineers were classifying as bogie fatigue failures on Baureihe 52 locomotives.

The investigation consumed man-hours. It prompted the withdrawal of 17 locomotives for emergency inspection at Metz depot in November 1942. Those 17 locomotives spent a combined 340 days out of service. Not one of them had a mechanical fault. Henry on the embankment outside Lyon that November night is placing his device for reasons he does not fully understand.

His SOE controller has given him a railway worker’s lamp, a false identity card from a forged set printed by the SOE’s forgery section in Hertfordshire, and four clam rails in a tool bag marked with a French national rail company stencil. He knows the device works. He does not know that a version of this moment, an agent, wet grass, a cold rail, a 2-minute exposure, is happening in at least four other locations across occupied France tonight.

The exact count of clam rail deployments across occupied Europe between August 1942 and September 1944 remains genuinely unknown. SOE records were deliberately and systematically incomplete, a security precaution that also makes post-war accounting impossible. Declassified files from 1971 suggest a minimum of 900 devices were delivered to French, Belgian, Dutch, and Norwegian resistance networks.

The Belgian Group G networks own post-war accounting, submitted to the Belgian government in 1946, claims 34 confirmed locomotive derailments attributable to rail placed devices of the clam type in the 12 months ending June 1944. The German response came in two stages, and both were, by surviving accounts, deeply frustrated.

The first stage was the deployment of additional track patrols, an instruction issued by Wehrmacht rail security command in February 1943, requiring double patrol coverage of all primary supply lines in occupied France. This instruction required the reallocation of approximately 3,400 security personnel from other duties.

The second stage, initiated after the Sicherheitsdienst finally identified the device type from a failed deployment recovered intact near Liège in April 1943, was a program of locomotive undercarriage inspection at all major marshalling yards. The program, which required trained examiners to check by hand torch the underside of every locomotive departing major yards, added between 45 minutes and 2 hours to each locomotive’s departure preparation time.

The throughput reduction at Metz yard alone, calculated from surviving German operations logs, was approximately 17% for the 3 months following the inspection order. The American OSS developed a broadly comparable device designated the Firefly in early 1943. The Firefly used a similar magnetic attachment principle, but a larger explosive charge.

120 g against the Clam rail’s 60. And proved significantly less reliable in vibration testing. Of 200 Firefly devices delivered to French resistance networks in the summer of 1943, declassified OSS records indicate 34 confirmed successful deployments against an estimated 41% failure rate. The Clam rail’s recorded failure rate from the same period was approximately 12%.

The German Abwehr, for its part, produced no equivalent device for use by stay-behind networks. Their rail sabotage doctrine relied on demolition charges requiring substantially greater quantities of explosive and substantially greater exposure time. The Soviet NKVD, according to documents released from Russian state archives in 1993, obtained three intact Clam rails through an intelligence exchange in Tehran in late 1943, and produced a reverse-engineered version designated the ZhD-3.

How many were produced and whether they were operationally deployed remains a subject of historical dispute. The material effect of the Clam rail on the Allied campaign has been carefully separated from post-war enthusiasm by historians including M.R.D. Foot, whose 1966 official history of SOE in France remains the foundational text.

Foot was precise in his accounting. The Clam rail was not a war-winning weapon. It was a sustained irritant of the highest order. Its genius was economic. The cost to disruption ratio was extraordinary, and the secondary effect of consuming German engineering, security, and administrative resources was arguably as significant as the primary effect of destroying locomotives.

A single clam rail costing under two shillings produced in the optimal case a locomotive out of service for weeks, a blocked main line for days, and a German security investigation lasting months. The psychological record is harder to quantify, but it exists. Letters recovered from German rail security personnel in France, held at the Imperial War Museum, London, in their Second World War Occupation Records Collection, describe a specific anxiety, not of attack, but of uncertainty.

The mechanical failure deception meant German soldiers did not know whether a derailed train had been sabotaged or had simply broken down. One letter from an Oberst in the Wehrmacht rail security command, written in January 1943, describes his men as unable to distinguish between enemy action and the ordinary failures of machinery, which has the effect of making the men suspicious of everything and confident of nothing.

He was describing, without using the word, demoralization. Original clam rail devices are held in the collections of the Imperial War Museum, London, in the SOE Special Devices Collection, and in the collection of the National Army Museum in Chelsea. MD1’s Welwyn facility, or the site of it, now houses a private technology company.

Stuart Macrae’s notebook, in which he recorded both the failures and the successes of the development program, was donated to the Imperial War Museum by his family in 1983. The entry for May 9th, 1942, the date of the successful second test, reads in its entirety, “Ran 11K. The didn’t. Modern railway sabotage countermeasures in several European nations include routine undercarriage scanning equipment at major terminals, a direct descendant of the inspection regime the Wehrmacht was forced to introduce in 1943.

The connection is not always acknowledged. The problem it answers has not changed. Henri finishes the job in 90 seconds. He has practiced it until he can do it by touch in total darkness until the click of the magnet meeting the rail steel is the only thing he can hear. He picks up his lamp and his tool bag, walks back along the embankment to the maintenance access track, and does not look back.

His patrol window has 6 minutes remaining. He has already walked 40 m. The Baureihe 52 locomotive he has attached the device to departs Perrache Station Lyon at 05:40 hours on November 4th, 1942, hauling 38 wagons of Wehrmacht motor fuel. The fuse is set for 4 hours. By kilometer marker 71 on the Valence line, in the gray light of a November morning, the bogie frame of the right-hand driving axle disintegrates.

The locomotive goes left. Then it goes very fast. Then it stops on its side in a ravine 12 m from the track with 38 wagons piled in behind it. The German field report filed that afternoon attributes the derailment to bogie fatigue. The investigating engineer notes that the fracture pattern is consistent with metal fatigue from extended service.

He recommends the locomotive’s maintenance records be reviewed. Henri is eating breakfast in a safe house in the third arrondissement by the time the report is written. He does not know the result. He will not know it for months. He knows only that he placed it, walked away, and did not look back. In the logic of the clam rail, that was enough.

You did not need to see the crash. You needed only to place the device and trust the physics and the mathematics and the two shillings worth of marzipan smelling explosive inside a tin the size of a tobacco box. The device weighed less than a boot. It derailed the war.

 

 

 

 

The British Train Line Bait That Made German Supply Locomotives Flip at High Speed

 

It is the 3rd of November 1942 and somewhere along the Lyon to Marseille rail line in occupied France, a man named Henri is lying face down in wet grass beside a set of track he has walked past 40 times without stopping. Tonight he stops. He reaches into the lining of his coat. His name appears in no Allied after-action report.

What appears in those reports is the result of what he does next. The night is cold. The rails are slick with November damp and smell of iron and machine oil. A freight locomotive passed this way 40 minutes ago heading south. Its boiler pushing warm air across the embankment as it went. Henri felt it from 40 m.

He has timed the patrol cycle to the minute. He has 11 minutes left. What he is holding fits in one fist. It weighs 340 g, less than a tin of boot polish. It was manufactured in a workshop in Welwyn Garden City, Hertfordshire and smuggled into France inside the false bottom of a violin case. It cost the British government under two shillings to produce in industrial quantities.

Over the next 14 months, devices exactly like this one would contribute to the destruction or severe delay of more than 200 Axis rail supply movements across occupied Europe. More disruption per unit cost than any Allied bombing campaign of the same period. Henri presses it against the rail web, listens for a moment he cannot hear, and gets back to his feet.

The problem facing the British railway sabotage effort in 1941 was not a shortage of courage. It was arithmetic. The German rail network supplying the Western and Mediterranean fronts comprised more than 80,000 km of active line. SOE had fewer than 300 trained agents operating across all of occupied Europe by the middle of that year.

Conventional sabotage, cutting lines, destroying bridges, derailing single trains by hand-placed obstacles, required exposure times of between 6 and 20 minutes per operation. The Gestapo had reduced average agent life expectancy in France to 4 months by autumn 1941. The mathematics made sustained conventional sabotage a method of destroying agents faster than it destroyed trains.

The locomotives themselves were a specific problem the RAF could not solve. Bombing the Henschel works at Kassel, which produced a significant proportion of German freight locomotives, had been attempted repeatedly from 1940 onward. Factory production continued within weeks of each raid. Bombers sent to hit marshalling yards at Ham, at Cologne, at Metz, found that German repair teams could restore primary lines within 36 hours and secondary lines within 5 days.

At the Rommel supply peak in the autumn of 1941, an estimated 40 trains per day were running the North Africa supply chain through Italian ports. The loss of 12 trains to bombing in a given fortnight barely registered. The locomotive was the problem, not the track, not the bridge, not the signal box, the locomotive.

A modern German freight engine of the Baureihe 52 class, the Kriegslokomotive, the war locomotive, weighed 87 tons, developed 1,500 horsepower, and could haul 1,800 tons of cargo at 80 km/h. Building one required 26 tons of steel, 900 man-hours of skilled labor, and a minimum of 6 weeks from foundry to first run.

Destroying one by conventional means, derailment by obstacle, demolition charge at the chassis, required so much preparation, so much explosives weight, and so much exposure time that it had been done successfully by SOE agents fewer than 20 times in 2 years. There were thousands of these machines. There seemed to be no answer that didn’t get agents killed faster than trains.

That specific answer was already being worked on in a requisitioned country house in Station Road, Welwyn Garden City, at an organization that most people who worked there were instructed to describe as a research station for the study of drainage problems. The organization was MD1. Its full designation was Military Intelligence Research Department 1, and it operated under the directorship of Major Millis Jefferis, a Royal Engineers officer of such peculiar ingenuity that Winston Churchill had personally ensured

his protection from the normal military administrative machine. Churchill called Jefferis “My Department of Wild Talents.” The men who worked for Jefferis called their output, collectively, Winston’s Toys. The official designation for one of these toys, the device Henry carries in his coat, was the locomotive fog signal Mark II.

Though the name was later changed to the track sabotage device, rail pattern, and later still was known informally, in almost every SOE circuit that used it, as the clam rail. It worked on a principle that a clever 15-year-old could grasp in 30 seconds. A standard British fog signal, the type used by railway workers to warn locomotive drivers of obstruction, the type that any rail worker across Europe would recognize on site.

Detonated when a train wheel ran over it. The sound was a sharp crack, audible at some distance, used since the 1840s as an acoustic warning system. The clam rail used that recognition. It looked to a German sentry with a torch like a standard fog signal fixed to the rail. It was in fact nothing of the kind. The clam rail consisted of an outer casing of pressed steel, 118 mm long and 62 mm wide, weighing 340 g complete.

Inside the casing sat two chambers. The first contained 60 g of Nobel 808 plastic explosive, a substance with the color and consistency of marzipan and a scent, disconcertingly, of almonds. The second chamber contained a delay fuse capsule of the type already in use across SOE’s standard toolkit, adjustable between 15 minutes and 8 hours using a simple rotary collar.

A magnet on the underside, 23 kg of pull force for a device the size of a tobacco tin, bonded the unit to the rail web or, more critically, to the underside of a locomotive’s axle housing or bogie frame. This was the key difference from every previous attempt. You did not need to destroy the track. You attach the device to the locomotive itself.

The train carried its own destruction with it. The first prototype test at Welwyn in March 1942 failed because the delay fuse proved too sensitive to vibration from the moving train. The device detonated 340 m from the test siding before the locomotive had reached operational speed, blowing a section of axle casing on an old LMS tank engine and bringing the test to a halt.

Jefferis’ team, specifically a civilian engineer named Stuart McCrae, who had been a motoring journalist before the war and who kept a running notebook of failures that he described characteristically as useful data, modified the fuse housing with an inner rubber damping sleeve in April. The second test on May 9th, 1942, worked perfectly.

The locomotive ran 11 km before the fuse initiated. The resulting axle explosion derailed all four driving wheels and buckled the bogie frame beyond field repair. If this story is new to you, a quick subscribe means you will never miss another one like it. SOE’s F section began receiving clam rails in quantity by July 1942. The distribution network ran through the Spanish border crossing at Andorra and through Lysander drops to reception committees in the Dordogne, Corrèze, and Haute-Loire departments.

The first confirmed operational use came on the 17th of August, 1942, when an agent operating within the Car Cerci in southern France attached three clam rails to a Bow Rai 52 freight locomotive at Avignon marshalling yard. The locomotive departed at 0340 hours carrying 42 wagons of artillery ammunition bound for the Italian port of Taranto.

It derailed at kilometer marker 63 on the Aix-en-Provence line at 0517 hours. Surviving German field reports, declassified from the Bundesarchiv in 1978, record the derailment as having destroyed the locomotive, 17 wagons, and blocked the main line for 31 hours. The report describes the cause as a mechanical failure of the the assembly and does not mention sabotage.

That last fact was the clam rail’s secondary weapon. Because it attached to the locomotive and detonated kilometers from any station or yard, the resulting crash resembled to exhausted German track engineers working at night with inadequate tools, a catastrophic mechanical failure. Through the autumn and winter of 1942, the Sicherheitsdienst’s rail security reports show a rising pattern of investigations into what German engineers were classifying as bogie fatigue failures on Baureihe 52 locomotives.

The investigation consumed man-hours. It prompted the withdrawal of 17 locomotives for emergency inspection at Metz depot in November 1942. Those 17 locomotives spent a combined 340 days out of service. Not one of them had a mechanical fault. Henry on the embankment outside Lyon that November night is placing his device for reasons he does not fully understand.

His SOE controller has given him a railway worker’s lamp, a false identity card from a forged set printed by the SOE’s forgery section in Hertfordshire, and four clam rails in a tool bag marked with a French national rail company stencil. He knows the device works. He does not know that a version of this moment, an agent, wet grass, a cold rail, a 2-minute exposure, is happening in at least four other locations across occupied France tonight.

The exact count of clam rail deployments across occupied Europe between August 1942 and September 1944 remains genuinely unknown. SOE records were deliberately and systematically incomplete, a security precaution that also makes post-war accounting impossible. Declassified files from 1971 suggest a minimum of 900 devices were delivered to French, Belgian, Dutch, and Norwegian resistance networks.

The Belgian Group G networks own post-war accounting, submitted to the Belgian government in 1946, claims 34 confirmed locomotive derailments attributable to rail placed devices of the clam type in the 12 months ending June 1944. The German response came in two stages, and both were, by surviving accounts, deeply frustrated.

The first stage was the deployment of additional track patrols, an instruction issued by Wehrmacht rail security command in February 1943, requiring double patrol coverage of all primary supply lines in occupied France. This instruction required the reallocation of approximately 3,400 security personnel from other duties.

The second stage, initiated after the Sicherheitsdienst finally identified the device type from a failed deployment recovered intact near Liège in April 1943, was a program of locomotive undercarriage inspection at all major marshalling yards. The program, which required trained examiners to check by hand torch the underside of every locomotive departing major yards, added between 45 minutes and 2 hours to each locomotive’s departure preparation time.

The throughput reduction at Metz yard alone, calculated from surviving German operations logs, was approximately 17% for the 3 months following the inspection order. The American OSS developed a broadly comparable device designated the Firefly in early 1943. The Firefly used a similar magnetic attachment principle, but a larger explosive charge.

120 g against the Clam rail’s 60. And proved significantly less reliable in vibration testing. Of 200 Firefly devices delivered to French resistance networks in the summer of 1943, declassified OSS records indicate 34 confirmed successful deployments against an estimated 41% failure rate. The Clam rail’s recorded failure rate from the same period was approximately 12%.

The German Abwehr, for its part, produced no equivalent device for use by stay-behind networks. Their rail sabotage doctrine relied on demolition charges requiring substantially greater quantities of explosive and substantially greater exposure time. The Soviet NKVD, according to documents released from Russian state archives in 1993, obtained three intact Clam rails through an intelligence exchange in Tehran in late 1943, and produced a reverse-engineered version designated the ZhD-3.

How many were produced and whether they were operationally deployed remains a subject of historical dispute. The material effect of the Clam rail on the Allied campaign has been carefully separated from post-war enthusiasm by historians including M.R.D. Foot, whose 1966 official history of SOE in France remains the foundational text.

Foot was precise in his accounting. The Clam rail was not a war-winning weapon. It was a sustained irritant of the highest order. Its genius was economic. The cost to disruption ratio was extraordinary, and the secondary effect of consuming German engineering, security, and administrative resources was arguably as significant as the primary effect of destroying locomotives.

A single clam rail costing under two shillings produced in the optimal case a locomotive out of service for weeks, a blocked main line for days, and a German security investigation lasting months. The psychological record is harder to quantify, but it exists. Letters recovered from German rail security personnel in France, held at the Imperial War Museum, London, in their Second World War Occupation Records Collection, describe a specific anxiety, not of attack, but of uncertainty.

The mechanical failure deception meant German soldiers did not know whether a derailed train had been sabotaged or had simply broken down. One letter from an Oberst in the Wehrmacht rail security command, written in January 1943, describes his men as unable to distinguish between enemy action and the ordinary failures of machinery, which has the effect of making the men suspicious of everything and confident of nothing.

He was describing, without using the word, demoralization. Original clam rail devices are held in the collections of the Imperial War Museum, London, in the SOE Special Devices Collection, and in the collection of the National Army Museum in Chelsea. MD1’s Welwyn facility, or the site of it, now houses a private technology company.

Stuart Macrae’s notebook, in which he recorded both the failures and the successes of the development program, was donated to the Imperial War Museum by his family in 1983. The entry for May 9th, 1942, the date of the successful second test, reads in its entirety, “Ran 11K. The didn’t. Modern railway sabotage countermeasures in several European nations include routine undercarriage scanning equipment at major terminals, a direct descendant of the inspection regime the Wehrmacht was forced to introduce in 1943.

The connection is not always acknowledged. The problem it answers has not changed. Henri finishes the job in 90 seconds. He has practiced it until he can do it by touch in total darkness until the click of the magnet meeting the rail steel is the only thing he can hear. He picks up his lamp and his tool bag, walks back along the embankment to the maintenance access track, and does not look back.

His patrol window has 6 minutes remaining. He has already walked 40 m. The Baureihe 52 locomotive he has attached the device to departs Perrache Station Lyon at 05:40 hours on November 4th, 1942, hauling 38 wagons of Wehrmacht motor fuel. The fuse is set for 4 hours. By kilometer marker 71 on the Valence line, in the gray light of a November morning, the bogie frame of the right-hand driving axle disintegrates.

The locomotive goes left. Then it goes very fast. Then it stops on its side in a ravine 12 m from the track with 38 wagons piled in behind it. The German field report filed that afternoon attributes the derailment to bogie fatigue. The investigating engineer notes that the fracture pattern is consistent with metal fatigue from extended service.

He recommends the locomotive’s maintenance records be reviewed. Henri is eating breakfast in a safe house in the third arrondissement by the time the report is written. He does not know the result. He will not know it for months. He knows only that he placed it, walked away, and did not look back. In the logic of the clam rail, that was enough.

You did not need to see the crash. You needed only to place the device and trust the physics and the mathematics and the two shillings worth of marzipan smelling explosive inside a tin the size of a tobacco box. The device weighed less than a boot. It derailed the war.