He stops with the pick still in his hands and looks up. Seventy-seven feet above his head, running the same direction he is walking, sits a section of track he graded himself five weeks earlier. Same rail, same grade-stake pattern he drove himself. Same mountain.
He is standing at the bottom of a hole his crew is about to turn into a tunnel, and the tunnel is going to run directly underneath the railroad he already built. Nobody has explained this to him in a way that makes it make sense, because there isn’t a way. You lay track to go up a mountain. You do not lay track that ends up crossing over itself, stacked on the same hillside like a coiled rope, in a spot where a mile of steel wire wouldn’t span the space between the upper rail and the lower one as the crow flies.
But that is what is happening in the Tehachapi Mountains of Southern California in the winter of 1876. And the man with the pick is one of roughly three thousand Chinese laborers who are going to make it happen with hand tools, black powder, and eighteen tunnels cut through granite that does not want to be cut. The problem, stated plainly, is this: how do you make a train climb a mountain by crossing over itself, cut by hand? Not with modern earthmoving equipment.
Not with dynamite, which did not yet exist commercially in a stable form for this kind of work. They used an early nitroglycerin-based blasting compound, kegs of it, hauled up a mountain by horse. Not with any way to know, standing in a hole in 1875, whether the tunnel you were boring was actually going to line up eleven months later with a piece of railroad being built by a different crew a third of a mile away and seventy-seven feet higher. You are cutting toward a target you cannot see, through rock you cannot always predict, to build a shape—a spiral, a loop, a helix, call it what you want—that has to close on itself with enough precision that two rails meet at the right elevation, the right curve, the right grade.
Or none of it works, and the railroad from San Francisco to Los Angeles does not get built. Before any of that makes sense as an engineering answer, you have to understand what the mountain was doing to the railroad, because the loop wasn’t an idea some engineer got excited about. It was what was left after every other option failed. Trains in 1875 were limited by adhesion—how much pull the locomotive’s driving wheels could put on the rail before they just spun in place.
Adhesion falls apart on a steep grade. A freight locomotive of the era could manage a slope of maybe one to one and a half percent rise for every one hundred feet of travel and still keep a train moving at a working pace with a full load behind it. Push past two percent and you’re not hauling freight anymore; you’re fighting the mountain and mostly losing. And you’re also creating a second problem nobody likes to talk about as much as the first: getting back down.
A loaded train coming down a slope steeper than it was built to climb doesn’t need better engines. It needs better brakes than 1876 had. If it doesn’t have them, gravity wins, and men die in the wreckage. The Southern Pacific’s engineers, William Hood chief among them, with the construction run in the field by James Strobridge—the same man who had pushed the Central Pacific’s Chinese crews over the Sierra Nevada a few years earlier—had a fixed budget of elevation to spend and a mountain that demanded more than the budget allowed.
Between the town of Caliente and the summit at Tehachapi, the railroad needed to climb roughly two thousand feet over about twenty-eight miles of line. Do that math on a straight run and you are well over the grade a locomotive can manage. So the line doesn’t run straight. It runs long, doubling back on itself, threading canyon walls at an angle steep enough to be walkable but shallow enough for a train, averaging around two percent over the full twenty-eight miles.
Even that wasn’t going to be enough in one particular stretch where a hillside called Wong sat in exactly the wrong place, at exactly the wrong angle, to let the track keep climbing without either blasting straight through the hill’s full mass or finding some other way to buy elevation. The other way was the loop. Instead of going through the hill, the track was going to go around it—a full 360-degree turn, gaining height the entire way, so that when the rail came back to roughly the same compass heading it started on, it would be seventy-seven feet higher than where it began. Still climbing at a manageable, steady two percent the whole time.
A circle a little over a thousand feet across, a track length of about 3,800 feet to complete the turn. Which meant, and this is the detail that makes people’s stomachs drop when they finally picture it, any train longer than 3,800 feet is still on the lower part of the loop when its front end reaches the upper part, seventy-seven feet overhead, and the whole train passes underneath itself. That’s not poetry. That’s geometry forced on a nineteenth-century railroad by a hillside that would not move.
The men doing the paying started with a survey. Before a single pick touched the ground, somebody had to walk that hillside with a transit and a chain and figure out exactly where the circle needed to sit, exactly how tight the curve could be without exceeding what a steam locomotive’s wheelbase could tolerate, and exactly where eighteen tunnels needed to be bored through solid and decomposed granite to keep the grade from ever exceeding that same two percent. It is the part of the story nobody photographs, and it is the part that had to be right before anything else could be. A survey error of a few feet, uncorrected, compounds over 3,800 feet of curve into a mismatch that no amount of pick-and-shovel labor fixes later.
The surveyors worked with transit instruments, plumb bobs, and iron survey chains sixty-six feet long, walking the hill in sections, driving wooden grade stakes every one hundred feet that told the crews behind them exactly how much dirt to cut or fill to hold that two percent line. Those stakes were the only communication, most days, between the engineer’s math and the laborer’s shovel. Get the stake wrong and you don’t find out for months—not until two separate crews digging toward each other from opposite ends of a tunnel, deep inside the hill, either meet in the dark or don’t. That is the first danger, and it is the one that shapes everything after it.
The grade could not be built straight, could not be built through, and had almost no margin for error, because a train that cannot hold two percent on this stretch is a train that cannot cross the Tehachapi Mountains at all. The answer the surveyors gave was the loop itself, trading distance for grade, spending 3,800 feet of curved track to buy seventy-seven feet of rise the mountain would not give up any other way. Then came the tunnels, and this is where three thousand men with hand tools met granite that had been sitting undisturbed since before there were mountains to name. Eighteen tunnels along the Tehachapi line, several of them right in the loop itself.
Tunnel 9 sits at the bottom of the spiral, the ninth bore counting up from Bakersfield, and it is the one a long train passes over itself above, seventy-seven feet up, if you’re standing at the mouth of it and looking straight up the hillside. A crew working a tunnel face in 1875 had one method and no shortcuts. Drill a hole into the rock by hand, using a steel drill bit and a four-pound single-jack hammer, or on a bigger face, one man holding and turning the bit while a second and third swung eight- and ten-pound sledges in rotation, striking it every second or two, turning the bit a quarter turn between each blow so it bit fresh rock instead of grinding the same groove. A skilled three-man team could drill a hole eighteen to twenty-four inches deep into hard granite in about twenty minutes of that rhythm.
And then they’d do it again and again until a tunnel face had a dozen or more holes drilled into it in a pattern the powder monkey had already worked out. Center holes angled to break the rock inward, edge holes to define the tunnel shape. All of them packed with kegs of the nitroglycerin compound the crews called Hercules powder, tamped carefully with a copper tamping bar, because iron against that compound could spark, and there would be no warning after that. Southern Pacific’s own construction accounts for the pass put the powder consumption at more than six hundred kegs a week during the heaviest tunneling.
That tells you better than any adjective could how much of this mountain simply had to be turned to gravel and hauled out, one cartload at a time, before a train could ever run through it. Here is what a man’s hands knew that no engineer’s report writes down. You could feel a bad hole starting if the drill bit chattered instead of biting clean. If the dust coming back up the hole ran gray and dry instead of the wet gray paste of good granite being cut, the rock ahead had a seam in it, a fracture plane, a soft pocket, water working through a crack you couldn’t see.
A crew that ignored that feeling and packed the charge anyway could blow the face wrong, sending the break sideways into unsupported rock instead of straight back the way the pattern intended. An uncontrolled break like that could bring down more of the ceiling than the shoring was built to hold. Tunnel collapse was not a rare event on granite work of this kind in this period. It was an accepted cost, priced into the schedule the way weather was.
It is the second danger on this mountain, arriving exactly where you’d expect it: underground, in the dark, with a granite roof over men’s heads that had never once, in all of geological time, needed to hold itself up over an empty space before. The answer to that danger was timber—heavy pine sets, upright posts, and crossbeams wedged tight against the freshly blasted rock within hours of a shot going off, before the loosened material had time to work and shift and find the path of least resistance downward. Shoring crews followed the blasting crews at a pace measured in feet per day, not tunnel lengths, because you did not leave open rock unsupported longer than you had to. And even with that discipline, granite that looked solid could hide decomposed pockets, rock gone soft and crumbling from water and time working through it for longer than anyone standing there had been alive.
A decomposed section could let go without the courtesy of warning cracks. The company’s own construction reports and the newspapers covering the Tehachapi work in 1875 and 1876 both record tunnel accidents along this stretch. Neither source gives a clean, agreed-upon count. Chinese laborers on Southern Pacific and Central Pacific work in this era were frequently paid through headmen and labor contractors, moved between crews as jobs opened and closed, and buried or shipped home through channels the company’s own payroll never had to track.
When a man died in a Tehachapi tunnel, the fact of the collapse might make the paper. His name, in the overwhelming majority of cases, did not. A single blast typically advanced a tunnel face somewhere between eighteen inches and three feet, depending on how the rock broke. That meant a bore of four hundred feet—not unusual for the longer Tehachapi tunnels—could take somewhere in the neighborhood of two hundred separate shots to complete.
Each one followed by the same sequence: clear the smoke and dust, muck out the broken rock by hand shovel into a cart, haul the cart out on temporary rail laid as the tunnel advanced, shore the new section, drill the next round of holes, load the powder, shoot again. On a good day, working two faces of a single bore at once, a crew might advance the combined total by five or six feet. On a bad day—a bad hole, a jammed drill, a section of rock too hard for the pattern that had worked fine twenty feet back—advance could stop entirely while the crew reworked the plan. Multiply that rhythm across eighteen separate bores, several of them running simultaneously with different crews, and you start to understand why this section of railroad, only twenty-eight miles long, needed nearly two full years and three thousand men to finish.
The loop itself still had to close. This is where the third danger sits, and it is the quiet one, the one that doesn’t look dramatic in a photograph but that every other piece of this project depended on. The loop had to align over itself. Not approximately—exactly.
The upper leg of the circle, carried by a crew working one direction around the hill, and the lower leg, worked by a different crew going the other way or continuing on from the tunnels, had to arrive at the same final curve, the same grade, the same elevation math, so that when the two sections met, the rail simply continued. No jog, no mismatch in slope, no seam a locomotive’s wheels would feel as anything other than more track. A tunnel driven a few degrees off its surveyed line inside a hill doesn’t announce the error until the crews break through, and by then there is no cheap way to fix a misaligned bore that has already cost weeks of hand-drilled granite. So the surveyors rechecked their transit lines constantly, reset stakes after every significant cut or fill, and effectively reproved the circle’s geometry dozens of times over the length of the work, because the whole design’s only real safety margin was getting the math right the first time and confirming it relentlessly after.
It worked. Sometime in early 1876, the last tunnel through, the last curve graded to its stake line, the loop closed on itself the way the survey said it would. Seventy-seven feet of rise packed into 3,800 feet of curling track. Southern Pacific was ready for its fourth danger, the one every other danger on this mountain had been leading toward without anyone saying so out loud: the first train.
Nobody had run a loaded locomotive over a brand-new spiral cut through granite by hand labor before. Not on this railroad, not at this grade, not with tunnels this fresh. A working engine of the period, loaded with water and coal, could weigh somewhere in the neighborhood of sixty or seventy thousand pounds, bearing down through narrow steel wheels onto rail that had, in places, been laid over fill the crews themselves had cut and packed weeks earlier. If a section of grade had settled unevenly, if a curve was tighter than the surveyed radius by even a small margin, if a tunnel’s clearance was off by inches in the wrong direction, the first train through was how you found out—at speed, with a crew aboard, on a curve where there was no room to stop and check.
The Tehachapi loop was completed, and the line opened for through traffic on February 24, 1876. When that first train made the climb, rising at its steady two percent, passing over the ground it had already covered five minutes earlier, seventy-seven feet below its own rear cars in places, it proved every stake, every drilled hole, every timber set, every rechecked survey line correct in the only way that mattered. It stayed on the rail. Southern Pacific had its last link between San Francisco and Los Angeles, finished in under two years.
Started in 1874, finished in 1876. Twenty-eight miles of mountain crossed, with eighteen tunnels and this one impossible circle carrying it. The payroll patterns for Chinese railroad labor in this era are well enough documented across both Central Pacific and Southern Pacific records that a specific, honest shape can be put around one man’s month, even though the Tehachapi ledgers that would have carried his actual name have not survived in any archive. Call him Ah Lim, because it is a plausible period name and the story needs one.
He is a composite, built from documented wage rates, documented crew structures, and documented working conditions on this exact stretch of railroad—not a specific man found in a specific record. Ah Lim is somewhere in his late twenties in 1875, part of a work gang of roughly twenty to thirty men organized under a headman. Chinese labor on Central Pacific and Southern Pacific work was almost always structured this way: contracted and managed through headmen who dealt with the company on the crew’s behalf, handled pay, and often handled food. Chinese crews commonly fed themselves on a diet the men chose and cooked themselves—dried fish, rice, vegetables, tea boiled and cooled for the drilling gangs to carry to the tunnel face—rather than the salt pork and bread the company issued to white laborers.
That single fact tells you something the wage numbers alone don’t. Even where pay was close to equal, the terms around it usually weren’t. Documented Chinese labor rates on Central Pacific work in the early 1870s sat around thirty dollars a month, and Southern Pacific’s Tehachapi work followed the same pattern: thirty dollars, paid out through the headman, for a laborer’s month of six days a week, ten to eleven hours a day, grading track or hauling spoil out of a tunnel bore by handcart. White laborers doing comparable unskilled work on the same railroads in this period were frequently paid a similar cash figure, but with board included at company expense, which meant the Chinese worker’s thirty dollars had to stretch further before it was worth the same thing in his pocket.
To give that number some weight: a loaf of bread in California in the mid-1870s ran somewhere around a nickel, sometimes less in a mining or rail camp with its own bakery. So thirty dollars was, in the crudest terms, six hundred loaves of bread for a month of ten-hour days in a granite tunnel with black powder going off down the drift. Ah Lim’s headman is the man who actually runs his working life. The headman negotiates with the company’s construction superintendent, assigns his gang to a tunnel face or a grading cut, keeps the crew’s pay accounts, and is the only person in this entire chain with a name reliably likely to survive in a company ledger, because he is the one the company paid directly and the one the company had a reason to write down.
A laborer like Ah Lim exists in the historical record, if he exists at all, as a headcount on a crew roster and a share of a lump payment. That is not speculation; it is how the documented system worked. He has a wife back in Guangdong Province, in most documented cases of this era, and sends money home through remittance networks organized by the same regional associations that recruited and organized much of this labor in the first place. A portion of that thirty dollars a month—sometimes a third or more of it—routed back across the Pacific to support a household he might not see again for years, if ever.
Some men in this labor system saved toward a return passage. Some never made it home. Some sent for family later. All three outcomes are documented broadly across Chinese railroad labor of this period; which one belonged to any specific man on the Tehachapi crews is mostly lost.
A day for Ah Lim, if it followed the documented pattern of a Chinese grading and tunnel gang on this railroad, started before sunrise with rice and tea over a fire. The crew’s own cook kept going and ran six days a week, ten to eleven hours a day, with the seventh day given over to rest, laundry, letters, and, where a large enough camp allowed it, a joss house or a shared shrine the men had built themselves out of scrap lumber. Sunday work stoppage on Chinese crews was common enough on Central Pacific and Southern Pacific lines to be documented across multiple contractors’ accounts, and it stands in fairly sharp contrast to the popular image of the era as one long unbroken grind. It wasn’t unbroken, though; eleven hours on a drilling gang, six days running, is not a schedule anyone would call gentle by any measure, then or now.
What is documented specifically for this stretch of railroad is the danger his work carried. Drilling and shooting granite by hand, hauling spoil out of a bore on a handcart over uneven temporary rail, working beneath timber sets that were only ever as good as the last man who wedged them. These were not survivable jobs in the way modern construction, with its inspections and its engineered shoring, tries to be survivable. A man in Ah Lim’s position spent his working days in a narrow granite corridor lit by candle or oil lamp, breathing rock dust and powder smoke that never fully cleared between shots.
Listening, because listening was part of the job—the way an old miner learns to hear a roof before it comes down, for the particular creak or trickle of falling grit that meant the ground above him had decided to move. That listening is the whole of what a headman’s crew relied on for warning most days, because there was no other kind. The accounting on what this cost is incomplete, and it is worth sitting with why rather than rushing past it. Southern Pacific’s construction records for the Tehachapi line note the scale of the labor force, the number of tunnels, the powder consumption, the schedule—company numbers kept because the company needed them for its own purposes.
What those same records mostly don’t note with any consistency is Chinese worker deaths and injuries as a distinct, counted category. Newspapers covering the pass work in 1875 and 1876 occasionally reported a tunnel accident, a cave-in, a blasting mishap, usually in a sentence or two, sometimes with a body count, and almost never with names. County records from this period and region are thin, and in places simply don’t survive. What that leaves is a range instead of a number.
Historians working from the broader pattern of railroad tunnel construction in this era—comparable powder use, comparable rock, comparable crew sizes on other Central Pacific and Southern Pacific projects where slightly better records exist—generally place the death toll on projects of this scale and duration somewhere in the dozens, not the hundreds. That is an estimate built from analogy, not a confirmed count for Tehachapi specifically. Anyone who gives an exact number for this stretch is telling you more confidence than the record supports. What can be said with more certainty is what the survivors got.
When the Tehachapi line was finished in February of 1876, Southern Pacific had a completed railroad, a construction budget closed out, and a workforce it began releasing or moving on to the next stretch of line. The company’s obligation to a crew member, once his section of grading or tunneling was done, generally ended with his final pay packet, handled through his headman. There was no pension for this work, no company memorial, no plaque with a laborer’s name on it. The plaques that do exist at the Tehachapi Loop today—and there are two of them at the overlook—credit William Hood by name and note the Chinese labor force only as a number: three thousand, doing the work.
To be fair to the record, it would be easy to overstate this. Southern Pacific’s construction accounts do preserve a fair amount about how the work was organized: the powder tonnage, the tunnel footages, the crew counts by section, the pace of the grading. Historians working from railroad company archives and from the broader documented pattern of Chinese labor on the Central Pacific have been able to reconstruct wage rates, contracting structures, and working conditions with real confidence. That reconstruction is what makes it possible to build Ah Lim’s month with the detail given.
What the record does not preserve, on this stretch specifically, is individual identity below the level of headman. That absence isn’t neutral. It is a choice made over and over across a hundred and fifty years by whoever decided what was worth writing down, and it has left a railroad that can be described in feet and dollars and kegs of powder far more precisely than the men who built it can be described in names. The Loop has been one of the most photographed places in American railroading for the better part of a century.
Train spotters have been coming to Wong for decades to catch a long freight crossing over itself, and it is a genuinely beautiful thing to watch happen. But almost never has a caption on one of those photographs named a single man who cut the tunnels or graded the curve. The photograph is always of the train. The mountain is the backdrop.
The three thousand men who made the shape the train is following are, at absolute best, a sentence in the corner: built by Chinese labor in 1876. Neither the famous aerial shots nor the train watcher’s dream put a name on the man who felt the drill bit chatter and knew, before anyone told him, that the rock ahead was about to go wrong. The Loop didn’t get replaced. The mountain never stopped needing that same trick.
Diesel locomotives got more powerful. Air brakes got dramatically better than anything available in 1876. Modern signaling let Southern Pacific’s successor railroads run trains far heavier and far more precisely spaced than anything that could have used this line in its first decades. None of that changed the fact that a train still has to gain seventy-seven feet of elevation across ground that only allows it if the track spirals over itself to do it.
The grade the surveyor staked in 1875 is still in use, within engineering tolerances. The tunnels bored by hand in granite—all eighteen of them—are still the tunnels the trains run through, reinforced and maintained across a century and a half of use but not replaced, not rerouted, not made obsolete by anything an engineer with better tools ever came up with. The geometry that made the Loop necessary in 1876 hasn’t changed an inch. Something like thirty-six freight trains a day still cross this same ground, still climb this same spiral, still, if they’re running long enough, pass over their own rear cars in the exact spot where a Chinese grader once stood in a half-dug tunnel and looked straight up at a track he’d laid himself, going the same direction, seventy-seven feet over his head.
You can stand at the overlook above Wong today, the way people have stood there for decades with cameras and lawn chairs and radio scanners tuned to the dispatcher’s channel, and you will very likely see it happen: the front of a long freight train appearing to chase its own tail, disappearing into Tunnel 9 while the rear cars are still climbing the curve above it. It is one of the strangest and most satisfying things you can watch a train do anywhere on Earth. And every single person who has ever photographed it has photographed a piece of engineering that three thousand men built with a hand drill, a sledgehammer, a keg of unstable powder, and a wage of thirty dollars a month, paid through a headman whose name is more likely to have survived than theirs. Under that mountain, in rock that is still holding the same tunnels a century and a half later, there is a hole that was the last one a specific man ever drilled.
Neither the company’s ledgers, nor the county’s inquest books, nor anyone now can tell you which one it was.
