My I can assign a date was in 1954. I was three months shy of my fourth birthday, and the event has stuck clearly in my mind for all these years.

Back then, there was a regional railroad in north Georgia named the Gainesville Midland. It was a small operation, always strapped for cash, and it was probably the last railroad in Georgia to run steam locomotives. The flagship of the line was a decapod, a heavy freight engine having ten driver wheels, number GM207, named “the Russian.” It was so named because it was built by the Baldwin Locomotive Works in Eddystone, Pennsylvania, in 1916, under contract with Czar Nicolas II, Emperor and Autocrat of All the Russias. It was ready to ship in 1917, but Nicolas was under severe stress at the time, and payment was not forthcoming. Finding the Russian government completely collapsed, Baldwin sold its entire inventory of oddly specified 2-10-0 decapods at auction in an attempt to recover manufacturing costs. The Gainesville Midland wound up with three of them, and Baldwin was happy to readjust the gauge for the light tracks in Georgia.

The Russian looked incomprehensibly huge to me. How could something so big, so massive, move at all? It blotted out the sun when it passed, throwing black soot high into the crisp autumn air and causing the ground to move under my feet like a Japanese earthquake. The boiler sat so high, I could see daylight through the spokes in the drive wheels as it thundered by at top speed, making 35 miles per hour pulling a mixed string of five cars. On the downhill, you could outrun it on a bicycle. It ran back and forth, between Athens and Gainesville, roughly alongside the Winder Highway.

One Sunday afternoon we were at my grandparents’ house in Hoschton, Georgia. The town used to be on the Gainesville Midland line, but the tracks had been torn up in 1947 when the route was cancelled. The train station was still there, empty of purpose like other buildings in the hamlet that had seen more prosperous times. It was a slow day.

It had been raining constantly for the past week, and everything in Georgia was soaking wet, including the fellow who came to the door with urgent news. “Colonel!” he cried. “The Midland done wrecked!”

Granddaddy dropped his New York Times and rose to his feet. “Wrecked?”

“Yes, sir! It’s the Russian. She’s off the rails, up yonder, nearly t’ Gainesville.” He pointed vaguely west.

This was no time to be sitting around listening to the house settle. We piled into the Studebaker and hot-wheeled it up the road to a wide spot that no longer exists, called Candler, just south of Gainesville. You could feel the spectacle growing as we approached. Cars were parked or abandoned off the road. First a few, then clumps, then seemingly every car in the world. People were walking, jogging, and sprinting, all in one direction, pointing and shouting. We pulled off and started walking.

After trudging about a hundred miles we reached a sharp turn, where the tracks veered off to the left, and there it was, lying on its side, wheels in the air, like a dead dinosaur. The heavy Russian had taken the turn too fast, and the red clay under the tracks, saturated with water, just slid out from under it. I could swear the thing was still breathing. Periodically you could hear steam burbling somewhere deep inside its enormous body. People were just standing there in awe of the spectacle, uncountable hundreds, quietly staring and whispering to each other. Someone said that the engineer had to be cut out of the wreckage with an acetylene torch. I stood on my tiptoes and tried to see the twisted wreck of the cab. It was too far away, down a

hill.2

I learned something that day, and it had nothing to do with going too fast around a curve: there’s a great deal of entertainment value in a train wreck. Even the aftermath of a crash, with the engine upside down and cars scattered all over the place, is surprisingly theatrical—a tragedy in hot steel, plowed mud, and scattered coal. There was sport in just analyzing the disaster, thinking what could have happened, back-tracing the last moments of the engine’s life, and imagining it digging the long trench as its energy dissipated into the ground. If it were roped off, you could sell tickets.

As is almost always the case, I was not the first to think of this. In 1896 a passenger agent for the Missouri, Kansas & Texas Railway (Katy), William “Bill” Crush, came up with a brilliant publicity stunt that would drum up passenger business. Being a natural-born salesman, he was able to convince his boss that they should stage a head-on collision between two locomotives. With a little advertising, it would attract thousands of people! There would be no charge to see the crash, but they could sell train tickets to bring people to the event. At two dollars per roundtrip ticket, they would not only gain publicity for their railroad, they would clear a profit as well.

In the 19th century, rail travel was the premier form of ground transportation, and just about everybody spent time in a railcar, gazing out the window as the rural terrain sped by or sleeping in the sitting position. Steam trains were large, heavy, fearsome beasts, breathing fire and looking dangerous. Some people were excited by the technical advances that had made this mass transportation possible, and some were terrified of it. There were too many newspaper stories every day about train wrecks. It seemed that engines were always blowing up for no obvious reason, crashing into each other, tilting off the rails, or plunging off a trestle into a gorge. There were citizens who could not be forced onto a train at gunpoint. Engineers blamed impossible schedules and poorly maintained tracks. Conductors blamed engineers. Railyard workers blamed brakemen, and railroad owners blamed the newspapers for lurid prose.

In 1891, a particularly bad year, 7,029 Americans lost their lives in railroad accidents. There were only about 64.4 million Americans at the time, so that makes the fatality rate 1.4 times that of automobile travel in 2011. The idea of staging a train wreck in 1896 was a superb piece of psychology. Instead of assuring passengers that all trains were safe and nobody could get hurt, show them the worst that could possibly happen. Let them feel the heat blast, the steam escape, and the ground-trembling thud. Allow them to get as close as they dared, and, what was most essential, let them see it coming. There would be no buried dread of the random, completely unexpected accident. The fear of the unknown would be replaced by the excitement of expectation.

A bare patch of ground outside the city limits of Waco, Texas, was staked out, and a set of temporary tracks was laid. Two obsolete 4-4-0 American pattern locomotives, looking like Civil War relics, were purchased and dolled up. One was painted green with red trim, and the other was painted red with green trim. Boxcars were added, with advertising for the Oriental Hotel in Dallas and the Ringling Brothers Circus painted on the sides. Tents were erected. A temporary restaurant was built, as well as a jail, and a 2,100-foot-long platform was banged together to give people a place to stand and watch the show. Eight tank cars filled with water were brought in to prevent spectator dehydration.

The event was scheduled for September 15, and by then the crowd had grown to over 40,000 souls. As an afterthought, Bill Crush was asked, “Is this safe? Them old boilers ain’t gonna explode, are they?”

Since the invention of high-pressure steam earlier that century, boiler explosions had become the number one fear of everyone participating in the steam-power revolution. Boiler explosions had been killing anyone standing near an over-pressurized locomotive since 1831. Steam carried a lot of pent-up energy. It wasn’t just the immediate fire under the boiler that was the problem, it was the heat energy built up and stored in the steel vessel that was so dangerous. A steam explosion could happen at any time, out of the blue, without a hint of warning. A boiler would disintegrate, sending hot, knife-like pieces ripping mercilessly through a crowd. It was not the sort of publicity that a railroad ever needed.

“Naw,” said Bill, patting the still-sticky paint. “These old engines are tough. It’s just going to make a big noise and crush it like a tomato can. No blow-up. I’m sure.” Of all the employees in the Katy, Bill Crush probably knew the least about steam and mechanical stress.

The afternoon was getting hot, and the crowd was growing restless. Two hundred men were hired to control the mob, but it was beginning to get out of hand. The two engineers were ready at the throttles, the boilers were redlined, and the steam relief valves had sprung open and were blowing mist. Crush rode out in front of the crowd on a borrowed white horse, raised his hat high, let it hang for a moment, then dropped it. The crowd went wild, and the engineers jerked their throttles full open. C. E. Stanton in the green engine and Charles Cain in the red one coolly waited for 12 puffs from the cylinders and bailed out, with the lightly loaded engines gaining speed. People pushed and shoved for an unobstructed view.

On they came, blowing dark clouds of smoke and setting off emergency signal torpedoes placed all along the track. Bang. Bang. Bang bang bang. Faster and faster, reaching a combined collision speed of 100 miles per hour. The official event photographer, J. C. Deane, tripped his high-speed shutter just as the two cowcatchers met. The two old engines, weighing about 35 tons each, suddenly occupied the same spot on the track. There was a terrific sound of crashing, bending metal as the two locomotives melted together, lifted their front trucks off the track, and seemed to hang for an instant. The wooden cars behind splintered and crushed

as the two trains telescoped together.

Then, something bad happened. At least one of the boilers exploded with a heavy roar, sending a rain of jagged metal into the crowd. The first casualty was Deane, the photographer, stationed closest to the crash point. A piece of hot locomotive hit him in the face, cleaned out an eye-socket, and left a bolt and washer embedded in his forehead. He spun around to face the audience and went limp. Louis Bergstrom, also on the photography team, was cold-cocked by a flying plank. Ernest Darnall, a boy with a rare viewing opportunity sitting high in a tree, caught a heavy iron hook trailing a length of chain right between the eyes, splitting his skull down the middle. DeWitt Barnes, in a dignified standing position between his wife and another woman, was killed instantly by an unidentified fragment. People in the front row were scalded, screaming, and dripping blood. In all, three people were killed on the spot and six were very seriously injured. A Civil War veteran was visibly shaken, saying that it reminded him of seeing a line of men dropped by a Yankee rifle volley.

Instant tragedy, however, did not dampen the crowd’s enthusiasm. They rushed the scene by the thousands in an incoming wave, poring over the wreckage to pick up or wrench loose the largest pieces they could carry. Many palms were singed as people pounced on bolts, rivets, bits of boiler tubes, and all manner of unidentifiable relics. To appease grieving families, Bill Crush was immediately and visibly fired from his job at the Katy. He was quietly re-hired the next day. From that day forward, the Katy Railroad flourished, and the many who had decided not to go to the event regretted the decision for the rest of their lives, as the stories of “The Crash at Crush” were told over and over in song, ragtime march, musical play, and Sports Illustrated.

Bill Crush wasn’t even the first to think of this. Incredibly, there were four independently staged engine head-butts in September 1896. None was as spectacular as Crush’s 100-mile — per-hour boiler bust, but the clustering indicates an unfulfilled need in the human psyche, peaking in 1896. Just outside Denver on September 30, two old narrow-gauge 2-6-0 Union Pacific and Denver & Gulf engines were smushed together for a crowd as a fund-raiser for the

Democratic Party.3 The crash made a lot of smoke and noise, but the engines were so feeble, the railroad was able to rebuild them and put them back into service.

On September 18 at the county fair in Sioux City, Iowa, two ancient Mason Bogey engines were smashed together to a cheering mob. In Des Moines at the State Fair on September 9, just six days before Crush’s spectacle, “Head-On Joe” Connolly arranged the collision of two really old 4-6-0 engines bought as junk from the Des Moines Northern & Western Railroad. The teeming masses numbered 70,000, and the gate receipts exceeded $10,000. That was a lot of money in 1896. Connolly was more adept at staging a crash than was Crush, and he knew to avoid a steam explosion. He had nothing to worry about. The elderly, arthritic engines were leaking steam at every joint. One was able to make 10 miles per hour, and the other 20. They hit at almost the right spot in front of the stands, there were the obligatory smoke and noise, and parts cartwheeled through the air, but the crowd was slightly disappointed. Still, they swarmed over the heap of steaming wreckage and carried off everything that was loose. Connolly returned home with $3,538.

Head-On Joe went on to make a career of locomotive crashing, eventually boasting that he had staged 73 wrecks, without killing a single spectator. He put together shows from

Massachusetts to California, mostly at state fairs but anywhere people would gather and pay to see two trains smash together. The city with the most staged crashes was San Antonio, Texas, with four. New York City, Milwaukee, and Des Moines had three each. His biggest audience was at the Brighton Beach Racetrack, New York, on July 4, 1911, where 162,000 people paid at the gate to see two old 4-4-0 engines kill each other. There were imitators, of course, but Head-On Joe had it down to a science. He knew that he had to have at least 1,800 feet of track, or the engines could not make enough speed for a proper spectacle. A track length of 4,000 feet was optimal, as the engines could accelerate to a combined speed of 45 miles per hour. That was fast enough to tear up the machinery and make the tender ride up over the cab without a boiler explosion. It took a mile of track to make 65 miles per hour combined, but that was too fast. Boiler explosions were fine, but you had to have the onlookers so far away, they couldn’t see anything. They wanted to be close enough to feel the collision, to hear the iron screaming in agony, and smell the hot metal, without being maimed. The locomotives had to be inexpensive and junky, without being undersized or wheezy. To wreck two nice-looking passenger engines seemed extravagant and in bad taste. To bury two old freight haulers in a moment of glory seemed merciful. Sometimes the engines looked hesitant as they tried to accelerate toward oblivion. Sometimes they looked angry, like pit bulls, not really knowing why they had to kill the other engine, but up to the task and really getting into it. It was art, in a machine-age sort of way.

At 73 years old, Head-On Joe’s last staged train wreck was back in Des Moines, on August 27, 1932, at the State Fair. A matched pair of 4-6-0s, just retired from the Chicago, Milwaukee, St. Paul & Pacific, faced off on the field. Both were freshly painted, and they were named “Roosevelt” and “Hoover.” Roosevelt was aimed east, toward Washington, D. C. A respectable mob of 45,000 came to see them on their last trip. After a short but suspense-filled run, the engines met, with the drama intensified by a box of dynamite tied to the pilot on each participant and fire-starters in the trailing passenger coaches. Hoover’s boiler exploded, rudely injuring two spectators with hurled shrapnel. There would be no lawsuits. They were, after all, standing near where they knew there was going to be a train wreck. What did they expect? Connolly collected his $4,000 and quietly faded away to his home town in Colo, Iowa. When he died in 1948, a brass locomotive bell was found on the family estate, possibly the only souvenir he had kept from the destruction of 146 train engines.

The last staged train wreck in the United States was probably the one near Magnolia, Illinois, on June 30, 1935. Two 2-6-0s from the Mineral Point & Northern, the 50 and 51, were supposed to meet on a bridge going a combined 50 miles per hour, but they missed the point, impacting instead in an open field at a fraction of the required speed. Coal flew vertically out of the 51’s tender and a puff of smoke rose, but the damage was so slight and the spectacle was so pitiful, it didn’t make the morning paper. The age of the staged train wrecks ended with a whimper. A creative plan to replace them with airplanes crashing into each other in mid-air did not materialize.

The need to see train engines crash together may have played out in the 1930s, but the specter of exploding locomotives would affect engineering for generations. Even today, in the 21st century, most of the safety design effort in a nuclear power plant is devoted to preventing a steam catastrophe. A nuclear plant is, after all, just another steam engine, heating water to a temperature beyond the boiling point and using the resulting vapor to rotate a shaft. The main difference between a nuclear generating station and its equivalent 100 years ago is that disintegrating uranium has replaced burning coal as the source of heat.

Numerous substitutes for steam as the prime mover in a power plant have been tried, but nothing has proven more reliable, efficient, or economical than boiling water. The task of converting heat into electrical current is not straightforward, but using steam as the transfer medium means that a large-output plant can be compact, and the working fluid is neither toxic nor flammable. Sitting on a small plot of land next to a river, a four-boiler steam plant can light up everything for a hundred miles, and if it is nuclear-powered then there is not even a pile of coal cinders and a mile-long line of rail cars waiting to be unloaded. Still, there is the fear of a steam explosion, something that impressed itself on both the public and the technical acolytes long ago.

In the early years of nuclear power development, in the technology scramble after World War II, early experiments and some small disasters pointed out the dangers of a runaway nuclear reaction. In practice, it was possible to increase the power output of a nuclear reactor not as a gradual heat transfer, like boiling water on the stove to make tea, but as a step function, or an abrupt increase in the blink of an eye. If you were standing near such an occurrence, you died, and it had the potential of flashing water directly and promptly into steam. The possibility of a runaway reaction and a resulting steam explosion was seen as the most critical safety concern in nuclear power development. If only this worst possible accident could be designed out of nuclear reactor plants, then everything else would be taken care of. All we had to do was keep the steam from exploding, and nuclear power would be stable enough to unleash on a safety­conscious public.

And so it was. With testing, accident simulations, well-thought-out engineering effort, and unusually robust building standards, the possibility of an explosive steam release was forcibly eliminated from nuclear power plants. In 56 years of commercial nuclear power generation in

the United States, there has never been a steam explosion, and not one life has been lost.4

No dreaded boilers coming apart, ripping holes in buildings and sending shrapnel into the crowd to worry about, but everything else in the history of nuclear accidents has happened for what seem to be the most insignificant, unpredictable reasons, much to the consternation of engineers everywhere. Entire reactor plants, billions of dollars of investment, have been wrecked because a valve stuck open or an operator turned a switch handle the wrong way. Some water gets into a diesel engine cooling pump, and six reactors are wiped out. Imagine the frustration of having built an industry having the thickest concrete, the best steel, meticulously inspected welds, with every conceivable problem or failure having a written procedure to cover it, and then watch as three levels of backup fail one at a time and the core melts. Obviously, the machinery was more sensitive to simple error than anyone could have thought, and thicker concrete is needed.

All the issues to be addressed concerning accident avoidance are not technical. Some are deeply philosophical. It is painful to notice, but some of the worst nuclear accidents were caused by reactor operator errors in which an automatic safety system was overridden by a thinking human being. Should we turn over the operation of nuclear power plants to machines? Would this eliminate the strongest aspect of human control, which is the ability to synthesize solutions to problems that were never anticipated? The machine thinks in rigid, prescribed patterns, but in dealing with a cascade of problems with alarms going off all over the place, has this proven to be the better mode of thought? Should operators be taught to think like machines, or should they be encouraged to be creative? Study the history of nuclear disasters, and you will have this subject to ponder.

There is also the elephant in the room: ionizing radiation. Nuclear engineers are acutely aware of this elephant and have designed it out of the way. Concrete thickness helps a lot to keep radiation away from all workers at the plant and certainly out of the public. The human fear of radiation is special and pervasive. As you will see, it originates in the initial shock of discovery, when we were introduced to the unsettling concept of death by an invisible, undetectable phenomenon. We have never quite gotten over it, and, in fact, all the fear of a steam explosion is not connected to the problem of hurtling chunks of metal or the burning sensation, but directly to the problem of radiation dispersal into the public. Steam, when it escapes in an unplanned incident at the reactor plant, takes with it pieces of the hot nuclear fuel. It floats in the air and blows with the wind, transporting with it the dissolved, highly radioactive results of nuclear fission. This undesirable process is at the root of accident avoidance in the nuclear power industry. Employee safety is, of course, very important, but public safety is even more so. To keep the industry alive, thriving, and growing, it is imperative that the general population not feel threatened by it.

Feeling threatened is not the same as being threatened, but the difference gets lost. The danger from low levels of radiation is quite low, as expressed as morbidity statistics or probabilities, but there is an unfortunate lack of connection to probability in the average person. Low probabilities are a particular problem of perception. If they were not, then nobody would play the lottery and the gambling industry would collapse. The impression of radiation, and even the science, can get lost in the numbers. In reading these chronicles of nuclear incidents big and small, I hope that you can develop a sense for the origins and the realities of our collective dread of radioactivity. Will this universal feeling prevent the full acceptance of nuclear power? Will we develop a radioactivity vaccine, or will we gradually evolve into a race that can withstand it? Perhaps.

There is also the problem of the long-term radiation hazard. People do not mind a deadly threat so much if it leaves quickly, like an oil refinery going up in a fireball or a train-load of chlorine gas tankers crashed on the other side of town. For some reason, a cache of thousands of rusting, leaking poisonous nerve-gas cylinders in Aniston, Alabama, does not scare anyone, but the suggestion of fission products stored a mile underground at Yucca Mountain, Nevada, causes great concern.

In this book we will delve into the history of engineering failures, the problems of pushing into the unknown, and bad luck in nuclear research, weapons, and the power industry. When you see it all in one place, neatly arranged, patterns seem to appear. The hidden, underlying problems may come into focus. Have we been concentrating all effort in the wrong place? Can nuclear power be saved from itself, or will there always be another problem to be solved? Will nuclear fission and its long-term waste destroy civilization, or will it make civilization possible?

Some of these disasters you have heard about over and over. Some you have never heard of. In all of them, there are lessons to be learned, and sometimes the lessons require multiple examples before the reality sinks in. In my quest to examine these incidents, I was dismayed to find that what I thought I knew, what I had learned in the classroom, read in textbooks, and heard from survivors could be inaccurate. A certain mythology had taken over in both the public and the professional perceptions of what really happened. To set the record straight, or at least straighter than it was, I had to find and study buried and forgotten original reports and first-hand accounts. With declassification at the federal level, ever-increasing digitization of old documents, and improvements in archiving and searching, it is now easier to see what really

happened.5

So here, Gentle Reader, is your book of train wrecks, disguised as something in keeping with our 21st century anxieties. In this age, in which we strive for better sources of electrical and motive energy, there exists a deep fear of nuclear power, which makes accounts of its worst moments of destruction that much more important. The purpose of this book is not to convince

you that nuclear power is unsafe beyond reason, or that it will lead to the destruction of

civilization. On the contrary, I hope to demonstrate that nuclear power is even safer than transportation by steam and may be one of the key things that will allow life on Earth to keep progressing; but please form your own conclusions. The purpose is to make you aware of the myriad ways that mankind can screw up a fine idea while trying to implement it. Don’t be

alarmed. This is the raw, sometimes disturbing side of engineering, about which much of

humanity has been kept unaware. You cannot be harmed by just reading about it.

That story of the latest nuclear catastrophe, the destruction of the Fukushima Daiichi plant in Japan, will be held until near the end. We are going to start slowly, with the first known incident of radiation poisoning. It happened before the discovery of radiation, before the term was coined, back when we were blissfully ignorant of the invisible forces of the atomic nucleus.

2 I’m not sure what happened to GM207. I’ve found GM206, GM208, and GM209, all resting comfortably in display settings. GM208 is in Winder, GM209 is

in Gainesville, and GM206 is somewhere in North Carolina at a railway museum. All are Russian-pattern 2-10-0 locomotives, but only GM206, built by Alco-Brooks, is said to have been built for Russian export in 1918. I swear the wrecked engine was called “The Russian,” but the story is hazy and I don’t even know why such a tiny railroad needed so many engines.

3 For those who may wonder, “2-6-0” is the standard way of specifying a steam locomotive configuration. This particular engine has two wheels on the

pilot truck, six steam-driven wheels, and no trailing truck behind the drivers.

4 I have to be careful here not to fall into the usual pro-nuclear trap of overstating a concept. Several people have been killed in nuclear industry accidents,

and many of those incidents will be discussed here. The worst nuclear accident in American history was a steam explosion, but it was a military reactor. So far, every death that can be positively linked to nuclear activity has been of military personnel, government workers in the atomic bomb industry or a civilian working in fuel reprocessing. Nobody has died because he or she was working in a commercial nuclear power plant in the United States. The Soviet Union is another matter.

5 A good example of this enhanced document availability is in my search for the original report, “The Accident to the NRX Reactor on December 12, 1952,

DR-32,” by W. B. Lewis. This was a very important accident. It was the world’s first core meltdown, and it happened at the Chalk River facility in Canada. I had heard about it many times, but I wanted the raw document. It seemed that every nuclear data repository I could think of, even in Canada, had an abstract of the paper, but not the paper itself. After a lot of digging I found it. The Russians had it, possibly recovered from the old KGB archives. This turned out to be a gold mine of information, including such things as accounts of the “Castor and Pollux” vertical assembly machines used in the development of the French atomic bombs.