He Found a Boiler Twice the Size of a Cabin – And Fit His Whole Life Inside 

He Found a Boiler Twice the Size of a Cabin – And Fit His Whole Life Inside 

There’s a question that nobody ever thinks to ask about the men who disappeared into the American road during the 1930s. Not where they went, not how they ate, or what they stole, or whether they made it back. The question is simpler than that, and it has been bothering me since I first came across Virgil Marsh’s name in a water stained section foreman’s log from Powell County, Montana, dated November 1,935.

The question is, where did they sleep when it got cold enough to kill? Most of them didn’t have an answer. They burned through their firewood before January, or they huddled in a box car siding until a bull found them and ran them off into a night that was 30 below. Or they made it to a Salvation Army cot in Missoula and spent three weeks coughing out whatever they’d breathed on the wrong freight car.

 A few of them didn’t make it at all. The Drummond section foreman, a man named Gerald Picket, whose handwriting was small and cramped, and who apparently kept a log of everything, noted two deaths that winter along the Northern Pacific’s Montana Division west of Deer Lodge. One man in early December, another one in January, found in a culvert near Baremouth.

Virgil Marsh was not one of them. What Virgil did instead is one of those things that sounds invented when you first hear it. and then starts to seem obvious the longer you think about it. He found an abandoned horizontal return tubular boiler in the Garnet Range foothills east of Drummond and he moved inside it and he stayed there for the whole of the winter of 1935 to 1936 which happened to be the coldest winter in recorded American history.

 He burned less than a third of a cord of wood doing it. I want to be careful here about what I’m claiming and what I’m not. There is no memoir. Virgil didn’t write one, or if he did, it hasn’t survived. What we have is Pickicket’s log, a few entries in the Drummond Merkantile Ledger from October and November of 1,935.

two paragraphs in a Granite County Historical Society newsletter from 1,962 based on the recollections of a man named Earl Bodin who spent part of that winter trapping near Quartz Flats and a physical description of the boiler itself from a 1,941 Forest Service survey of abandoned mill equipment in the Garnet Range.

 I’ve tried to put those pieces together honestly where I’m guessing I’ll tell you I’m guessing but the core of this the man the boiler the winter is real Virgil Marsh was 39 years old in the fall of 1 935 he had been on the road since the spring of 1,931 when the box factory in Moskegan Michigan where he’d worked since he was 19 closed its doors on a Wednesday morning with no notice at He came in at 6 the way he always did and found a padlock on the loading dock gate and a single piece of paper nailed to it.

 He didn’t save the paper. By the time he got back to the rooming house where he’d lived for 11 years, there was already a line of men outside the employment office three blocks over. He waited in that line until 2:00 in the afternoon. Then he went back to the rooming house, packed what he could carry in a canvas sack, and left.

 He was not particularly tall, 5′ n maybe, and lean in the way that men who work with their hands get lean not from not eating, but from spending it as fast as it comes in. Every calorie going somewhere. His hair was a dull reddish copper, thinning at the crown in a way that had started in his early 30s, and had gotten worse every year since, three weeks worth of beard the color of wet sand.

 He had a habit, people who knew him said, of tilting his head slightly to the left when he was thinking, the way a dog does when it hears an unfamiliar sound. His hands were large relative to his wrists, the hands of a man who had handled rough lumber for 20 years, permanently thickened at the knuckles, with a deep scar across the right palm from a bandsaw incident in 1927 that had never healed quite flat.

 What he wore in October of 1935 was this, a salt and pepper wool machino, red and black plaid, so faded the colors had blurred to something that was neither one exactly. Under it, a hickory stripe union shirt, the kind with vertical black and cream stripes that had been the standard railroad and mill workers shirt since before he was born, worn through at both elbows and patched with scraps of desses, a different fabric entirely, brown duck canvas from an old coat, so the patches showed as mismatched squares of a darker color.

Oshkosh Begosh bib overalls in 12 ounce denim that had been washed so many times the indigo was almost entirely gone leaving something closer to gray. The left buckle had cracked at the pivot and was held closed with a six penny nail bent double and threaded through both holes.

 Under the bibs, a dufold union suit in off-white wool, knit, repaired at the knees with a needle and thread he carried in a tin tobacco box. His broans were cracked leather with toe caps, resold once with a chunk of conveyor belting he’d gotten from a rubber plant in Akran in 1933. The right heel rettacked with roofing nails that had started to pull loose.

 He wore one gray wool sock and one brown one, not indifference, just arithmetic. He carried everything else in a horseshoe bed roll, a single army surplus blanket rolled around a square of oil cloth tied with a length of bailing wire and slung over his left shoulder. Clipped to the outside a Hills brother’s coffee can with a bailing wire bail.

 In the coffee can a stub of tallow candle, a Prince Albert tobacco tin that held matches, and a fish hook and 16 in of linen thread, a folding three-bladed Camila Stockman knife, and a partial sack of Bull Durham. in his right breast pocket, a card with his name on it, not an identification card, just a calling card from the Moskegan Benevolent Association that he’d found on a sidewalk years before and kept because it was the only piece of paper he owned with his name on it, which he could not entirely explain even to himself. He had ridden the

Northern Pacific West out of Billings in mid-occtober in the empty shell of a box car that smelled of coal dust and something organic and sweet he couldn’t identify. He’d dropped off at Drummond because the train slowed to walking speed on the grade there, and because he’d heard from a man in Billings that there was a section foreman near Drummond who sometimes hired a day laborer to split wood before winter.

 He found Picket, and Picket gave him two days of work at 30 cents a day, and that was how his name ended up in the log. He asked Picket at the end of the second day if there was any place a man could haul up for the winter in the hills east of town. Picket said there was an old stamp mill up a logging spur about a mile and a half, but there wasn’t much left of it.

 The buildings had fallen in years ago. And what the buildings left behind was a pile of rotten timber, some rusted iron pipe, and a boiler. He said the word boiler, the way you’d say the name of something that had stopped being interesting. Just a boiler still there, he said because nobody had ever come up with a reason to haul it out.

 Virgil asked how big it was. Pickicket thought about it. Twice the size of a cabin, he said. Maybe bigger. Virgil left before sunrise the next morning. The logging spur ran northeast from the northern Pacific mainline, following a creek drainage up into the Garnet Range foothills. In October, it was still passable on foot.

 Barely the ruts were deep and had dried at odd angles, and in the low spots there was ice in the mornings that hadn’t melted by midday. He found the mill site in about an hour and 40 minutes of walking, in a clearing where the ground was still littered with the gray, punky remains of the mill building’s floor, whole sections of it collapsed into the shallow crawl space below, like the ribs of something dead.

scattered across the clearing the rusted flanges of pipe fittings, a tangle of steam lines, a broken flywheel half buried in the earth, and then the boiler. It was horizontal, sitting on a low brick saddle that had settled unevenly on the downhill side, so the whole thing caned maybe 3° to the south. The shell was mild steel, rolled and riveted, roughly 5 ft in diameter and 18 ft long, big enough that when Virgil walked up beside it and put his hand flat on the cold metal, the curve of the shell was subtle enough that he could

almost forget it was curved at all. The outside was brown and orange with surface rust, but when he scraped at it with his knife blade, the metal underneath was solid. old boilers like this. This one had been manufactured sometime around 1,92 based on the partial maker’s plate still visible on the front head.

 Though the name of the manufacturer was gone, were built with steel that had been designed to hold 80 lb of pressure per square in. And even at the end of their working life, they were dramatically overbuilt relative to anything a man sleeping inside one could ask of them. There was a manhole in the upper part of the rear head, oval 12 in x 16 in with a cast iron ring bolted through the steel.

 The cover plate was still in place, held by six bolts, five of which were rusted so thoroughly into their threads that they wouldn’t move. He worked at them for 2 hours with a piece of iron pipe he found in the rubble as a cheater bar. Five of them he got loose. The sixth one he sheared off.

 He pried the cover up with the pipe, and it swung on its one intact bolt, and he put his head through the opening into the dark interior of the boiler. The inside smelled of old water and rust, and something else he couldn’t name, a dry mineral smell, like the inside of an old well. The fire tubes were still in place, a bundle of 50 or 62in steel tubes running the length of the shell from head to head, the lower half of the cross-section packed solid with them.

 They hadn’t been cleaned when the mill shut down. The inside of the tubes was coated with old soot and scale, and between them, where smaller pieces of debris had accumulated over the years, there was a gritty gray powder that he supposed was the remains of whatever combustion products had been left when the fire went out for the last time.

 He stayed in the clearing for the rest of that day, and made a list in his head of what needed doing. The biggest problem was the fire tubes. The lower half of the boiler’s cross-section was almost entirely occupied by them, which meant the usable interior space, the space above the tube bundle where the steam had once collected, was roughly a half cylinder, about 5 ft wide and 3 ft tall, running the full 18 ft length.

That was not nothing. That was, in fact, more habitable space than a lot of men on the road had managed in worse winters. But 3 ft of headroom meant he couldn’t sit upright, which meant lying or crawling, which for 18 ft of Montana winter was not what he had in mind. So the tubes had to come out, or most of them did.

 He went back to Drummond that evening and spent 40 cents of his 60 cents, the two days wages from Pickicket, on a cold chisel and a twob hammer at the Drummond merkantile. That’s in the ledger. One cold chisel, 20 cents. One two-lb hammer, 20 cents. Virgil Marsh, the 18th of October, 1935. He went back up the spur the next morning at first light.

 Pulling fire tubes out of an HRT boiler is the kind of job that looks worse from a distance than it actually is, and also looks easier than it actually is. And the real experience of it is somewhere off to the side of both of those impressions in a place that is mostly just slow. The tubes were rolled into the tube sheets at both ends, expanded into the holes by a tube roller when the boiler was built, so that the metal of the tube wall had been pressed outward into the sheet and formed a seal. What Virgil had to do was

break that seal. He used the cold chisel to work around the inside face of each tube end, driving the chisel between the tube wall and the sheet, and then pulling the tube free with the pipe he’d found. Some of them came loose with one or two blows. Some of them didn’t want to come at all and had to be worked from both ends in succession before they let go.

 A few of them he left in place when they were too rusted or too stubborn, and the cold chisel handle was raising a blister on his right palm that was close to breaking. He worked at it for 5 days. Five days of climbing in and out of the manhole, working by the thin light that came through the opening and the even thinner light that filtered through the open tube holes in the heads, coughing the dry rust and scale out of his lungs at the end of each day.

 By the end of the fifth day, he had pulled 43 tubes. He stacked them outside in two rows along the south side of the shell, and they were useful later, more useful than he expected, actually. What he had now inside the boiler was a space roughly 5 ft in diameter and 18 ft long. Interrupted by the 17 tubes he’d left in place along the bottom where the chisel hadn’t been able to get purchase.

 The floor of this space, if you could call it a floor, was the curved interior bottom of the shell, which was a problem because it meant nothing sat flat without being wedged. He fixed that by dragging in a section of the old mill floor, two planks still nailed to a joist, which he levered up from the pile of rot in the clearing, and laying them across the lower tubes, so they bridged the curve, and made something like a level surface.

 Not perfectly level, close enough, the manhole in the top of the rear head was still the only way in and out, which wasn’t workable for a winter’s living. He needed a door in the side. He spent two days on this, using the cold chisel and hammer to score a rectangle about 18 in wide and 24 in tall in the shell plate near the front head.

 Then working around the scored line again and again, deepening the cut each pass. Steel plate 38 of an inch thick doesn’t want to yield to a cold chisel at hand swing velocities. He broke through at the corners first and then connected the brakes. And the piece of plate he cut out he kept. It became the door.

 Hung on two short sections of fire tube threaded through holes he drilled with a hand brace he borrowed from Picket for two days in exchange for half a day of splitting wood. He was inside by the first week of November before the first serious snow. The question of heat was simpler than it sounds and also harder. The shell of the boiler was mild steel with no insulation value.

 He’d known that from the first morning pressing his hand to the outside plate the metal was exactly as cold as the air. Whatever heat went inside, the boiler would conduct straight through the walls and be gone. That was the physics of it. And there was nothing to do about the physics directly. What he could do was slow it down.

 He spent 3 days pulling the original lagging off the outside of the shell. The lagging, the outer insulation that had been applied when the boiler was in service was mostly rotted canvas and compressed magnesia powder, crumbling when he touched it. But underneath the outer canvas there were sections of what appeared to be a speestous blanket gray and matted still intact.

 He didn’t know what it was exactly, only that it was dense and didn’t burn and felt nothing like the cold metal beneath it. He pulled it off in strips and pushed it inside through the door opening, packing it against the curved walls of the interior and wedging it in place with sections of fire tube propped against it.

 Where the asbestous blanket ran out, he used hay. He found a collapsed stack of it in the remains of what might have been a stable structure at the south edge of the clearing, still dry in its center, where the collapsed roof had sealed it against the weather. He dragged it inside in arm loads and packed it against the walls over the asbestous, building up a layer that was 3 to 4 in thick around most of the interior circumference.

 He lined the floor planks with more hay, packed tight, and then laid his oil cloth over that, and his blanket over the oil cloth, and had something that could reasonably be called a bed at one end of the available space. At the other end of the space, near the door, he set a small stove. Not the stove he would have chosen.

 He would have chosen a round oak parlor stove, the small one, which he’d seen sold for $4 at a hardware store in Billings. What he had instead was a salvaged kerosene tin, a square 5gallon can with the top cut open, in which he punched holes with a nail for draft and cut a door with the cold chisel. The stove pipe problem took him almost a full day to solve.

 He ended up threading it up through one of the remaining fire tube holes in the top head, a 2-in hole, and the pipe was 2 in, and it was tight enough that he had to grease the joint with a piece of tallow candle. But it held, and it drew, and he understood after the first fire that the draft through that hole would be everything.

Too small, and the fire would smoke him out. The right size and the combustion gases went up and out, and the heat stayed. He cut the draft hole at the base of the tin stove to match. The thing he had not fully worked out in advance, and he would tell Earl Bodin this directly. In the winter of 1,935, when Bodin came past the mill site on his trap line, and Virgil invited him in, out of the cold for coffee was the air supply.

 [snorts] Combustion uses air. A man breathing uses air. In a space that size with that many tons of steel around it and two openings sealed against the cold, the oxygen would deplete and the carbon dioxide would build and you would go to sleep in the middle of the night and not wake up and nobody would find you until spring. Virgil had figured out most of the solution empirically.

 By the time he talked to Bodin, he kept a 2-in gap at the bottom of his makeshift door. always, summer and winter, no exceptions, plugged with a rolled piece of burlap that he pulled partway out at night. He also kept one of the lower fire tube holes open, plugged with a twist of hay that he’d pull free when he banked the stove for the night, which let air in along the floor.

 The cold air came in low, the combustion gases went out the stove pipe high, and the circulation worked. He was not entirely sure why it worked as well as it did. He’d been honest about that with Bodin, and Bodí had apparently been honest about it in the 1962 interview because the newsletter account preserves the uncertainty.

 He said he wasn’t certain of all the reasons, just that he’d found the combination by trial and error and didn’t mess with it after that. That’s one of the things I find believable about all of this. A man who actually did it says he’s not sure exactly why it worked. That sounds right to me. November passed. He was out of the boiler most days, cutting wood, lodgepole pine from the standing dead timber above the clearing, split with a maul he’d made by drilling a handle hole through a fist-sized piece of iron pipe fitting and wedging a split pine branch

into it. Not elegant, functional. He stacked the wood against the south side of the boiler shell under an overhanging section of the old mill roof he’d propped up on two fire tubes driven into the ground as posts. So the wood was under cover and the wind came from the northwest and didn’t reach it. He had 3/4 of a cord split by the end of November.

 He thought that was more than he’d need. He was not wrong, but not in the way he expected. The first week of December brought a cold snap that dropped the temperature at Drummond to minus17° F on the morning of December 9th. The Northern Pacific recorded it at the station. Virgil knew the temperature at the mill site was colder than Drummond because the clearing sat in a bowl in the terrain where cold air cooled on still nights.

 He didn’t have a thermometer, but he had a tin can of water that froze solid before the fire in the tin stove was fully out, which told him what he needed to know. He had a fire the size of two fists in the stove that night. That’s all. The asbestos and hay on the walls held enough of the heat that the inside of the boiler, by his estimate, never dropped below the mid30s.

 He woke up twice to feed the fire. He was not comfortable exactly, but he was not cold the way a man in a box car or a canvas tent or a half-faced camp shelter would have been cold that night. Earl Bodin came by on December 14th, following his trap line along the creek drainage. He saw smoke from the stovepipe, a thin gray thread coming straight up in the still cold air, and followed it to the clearing, not knowing what he’d find.

 He knocked on the boiler. He said later that the knocking sound, it made a bell-like ringing, then a longer resonance that rolled through the shell, was the strangest thing he’d ever heard in those hills. Virgil opened the door. Bodin said he stood there for a moment and looked at the interior, the haypacked walls, the plank floor, the small stove, the candle stub on a fire tube shelf, the horseshoe bed roll at the far end, and then said, “Well, I’ll be damned.” Virgil made coffee.

 He heated water in the hills. Brothers can over the stove, and Bodin had coffee and sat on the plank floor with his back against the haypacked wall, and they talked for most of an hour. Bodin’s camp was two miles up the drainage, a standard canvas wall tent with a small sheet ironed stove. He burned a lot of wood, he said more than he liked.

 Some nights, even with the stove going, the inside of the tent got cold enough that the water in his canteen froze. Virgil’s water in a tin can on the plank floor had not frozen in 3 weeks. January came. The cold that fell on the northern plains and the Montana high country in January and February of 1,936 was not like other cold.

 I mean that in a specific documented sense. The winter of 1935 to 1936 was the coldest winter in recorded American history. Nationwide, February 1,936, average 25.2 2° F, the lowest monthly average ever measured for the contiguous United States. On the 15th of February, 1936, partial North Dakota recorded -60° F.

 That is still the North Dakota state record. Still, Jordan, Montana, which is roughly 160 mi northeast of where Virgil Marsh was sleeping, hit -60° F the same week. Langden, North Dakota, stayed below 0 degrees Fahrenheit for 41 consecutive days, January 11th through February 20th. The mercury in a standard mercury thermometer freezes solid at my 38° F.

 Because mercury solidifies at my 37.89° F. At those temperatures, you need an alcohol thermometer just to take the reading because the standard instrument stops working. The section foreman at Drummond didn’t have an alcohol thermometer. He wrote in his log on February 15th. Thermometer went to limit. Then the next morning, still at limit, Virgil had no thermometer at all.

What he had was a routine. He woke each morning and fed the stove before he did anything else. Then he pulled on his macho and opened the door just a crack, just enough to feel the outside air and closed it again. he said in what Bodin recounted later, that he could tell roughly how cold it was by how fast the inside of his wrist got cold when he opened the door, and that on the mornings of the worst week in February that winter, the cold hit him so fast and so hard when he cracked the door that he thought for a moment something

had struck him. He pulled the door shut and went back to the stove. The stove, during those days, he kept burning all night rather than banking it. That meant waking every 3 to four hours. He burned more wood that week than in any other week of the winter, and still he burned less than Boddine burned in two nights in the canvas tent, according to what Boddine told him when he came by on February 21st after the worst of it had passed.

 Bodin said he’d nearly abandoned the trap line. He’d come close to walking out to the northern Pacific Maine and catching the first freight to Missoula. His tent stove had burned through its firebox. The metal had fatigued from the sustained extreme heat, and he’d spent two nights with almost nothing between him and the air, patching the hole with a piece of tin can.

 Virgil told him to come stay in the boiler. Bodin didn’t in the end. He was three miles up the drainage by then, and the thought of moving his camp in that cold was worse than the alternative. But he came down three more times before the thaw, and each time he came in and had coffee, and sat against the haypacked wall, and each time he left, he took a little more of what Virgil had figured out, back up the drainage with him.

 By his third visit, Bodin had re-chanked his tent floor with a layer of hay and bark, the same way Virgil had done the boiler floor. He said it helped. The thaw came in the first week of March. The temperature at Drummond rose above freezing on the 4th of March, 1936 for the first time since December 12th. Virgil noted this because the drip of melt water from the overhanging mill roof onto the top of the boiler shell made a sound like someone knocking slow and irregular, and he lay in his bed roll, listening to it, and knew what it

meant. He was still there in April. He’d seen no reason to go. The wood he’d cut before winter was almost exactly a third gone. He’d burned roughly a quarter of a cord in what had been the coldest winter in a century. In a structure that was technically not a structure at all. He started thinking about summer, which was a thing he hadn’t done in several years.

He thought about whether the boiler would be too hot in July. He thought it probably would. He started thinking about where he’d go in May when the NP freights were easier to catch and the weather was something you didn’t have to fight every waking hour. He left in late April.

 The section foreman’s log has no entry for him after that. The Drummond merkantile ledger has him buying a sack of dried beans on the 22nd of April, 1936. And that’s the last record. what he left behind. The door he’d cut and hung still on its fire tube hinges. The haypacked walls slowly decomposing back into the earth.

 the plank floor, the stove pipe running up through the fire tube hole in the top head, still in place when the Forest Service surveyor came through in 1941 and noted one HRT boiler approximat in south side single stove pipe exiting rear upper head interior showed evidence of habitation quote evidence of habitation. That’s the phrase.

 The surveyor didn’t know who had been there or when. He noted the modification and moved on. His job was to assess the scrap value of abandoned industrial equipment and a modified boiler with holes cut in it was worth less than a stock one. He marked it for low priority and it dropped off the list and as far as I can tell it was never hauled out.

 Bodin in 1962 remembered the coffee. He remembered the way the interior of the boiler felt warm and close and smelling of hay and tallow, the candle throwing a yellow circle on the curved steel ceiling. He remembered that Virgil had a harmonica, a Honer Marine band in the key of C, and that he played it badly and with apparent enjoyment, and that the sound of it inside the boiler shell was something he’d never heard anywhere else, full and strange and larger than the instrument had any right to make it.

He didn’t remember Virgil’s last name. He called him Virgil, just Virgil. The last name came from the section Foreman’s log, Pickicket’s cramped handwriting. V Marsh Moskegan MI two days split wood 60 cents. That’s what we have. I’ve thought about what Virgil Marsh knew or figured out or stumbled into. He wasn’t an engineer.

 He had no training in thermodynamics and almost certainly had never heard the word. What he understood was the same thing a blacksmith’s son in 1876 understood. The same thing a Monomony logger in 1921 understood. the same thing. Anyone who’s ever been genuinely cold and survived, it eventually works out.

 The difference between stopping cold and slowing it down. Stopping cold is impossible. The physics won’t allow it. Not with the materials available to a man with 60 cents in his pocket and a Camila knife. Slowing it down is something else entirely. The boiler stopped the wind. Every layer of hay and asbestous blanket slowed the conduction.

 The small fire added heat at a rate the slow conduction could not drain away faster than it arrived, which meant the inside stayed warmer than the outside by a margin wide enough to matter. You put all of that together and you get a man sleeping through a menace 60° Fahrenheit Montana winter on a third of a cord of lodgepole pine.

 You also get a man whose name doesn’t appear in any history of the period. He’s not in the hobo memoirs. He’s not in the WPA oral history collection. He’s not in the Federal Writers Project records of depression era migratory labor. He’s in a water stained log, a hardware store ledger, two paragraphs in a local newsletter, and a government surveyor’s description of a boiler with holes cut in it.

 Bodin taught his son what he’d learned from the winter of 1,935 and what Virgil had shown him. His son used it in 1949 when he was caught overnight in the Clark Fork drainage in a January cold snap and built a sleeping shelter against the back wall of an abandoned mining cabin using bark and dead grass layered in the same way.

 He told Bodin he’d been fine. Not comfortable, he said, but fine, Bodin said. That’s what Virgil would have said,