On the morning of October 12, 1905, a tunnel worker named Richard Seddon was standing eighty feet beneath the surface of the Hudson River when the compressed air around him changed. It was not the usual fluctuation of the compressors. It was something else. The compressed air had found a path.

Seddon was working inside the bore of the North River Tunnel, a twenty-three-foot-diameter tube being driven through the riverbed. Above him ran the Hudson, its depth set by the tide and current of that particular morning. Above the river was the surface, and above the surface the sky over New York and New Jersey. None of it was visible from where he stood.

What was visible was the face of the tunnel ahead, the massive shield pressed against the silt at the working edge, the rings of cast-iron lining receding behind him toward the shaft, and the arc lights the construction crews had hung along the ceiling—enough to work by, but not enough to make the tunnel look like anything other than what it was. The air pressure inside was thirty-six pounds per square inch above atmospheric. It had to be, at that depth. The Hudson exerts a hydrostatic pressure of roughly thirty-four pounds per square inch on any cavity opened in the silt beneath it.

The compressed air matched that pressure and slightly exceeded it. That margin was what kept the river out. If the pressure failed, the river flowed in. Every man in the tunnel knew this.

It was not a theoretical awareness; it was the nature of the work itself. What Seddon felt was the pressure changing. In a compressed-air tunnel, a blowout occurs when the air pressure at the working face forces a channel through the material above the tunnel roof, escaping toward the surface. In the North River Tunnels, that material was river silt, which is not consistent in density or texture.

In some places it has the porosity that allows a pressure channel to form. When the channel reached the surface of the Hudson, it blew. Witnesses on work boats above described the sound as like a cannon shot. A column of water, mud, and compressed air burst through the river surface.

Seddon was pulled toward the breach by the pressure differential as the air released. He was carried through it. He was thrown from the surface of the Hudson River into the open air, eighty feet above the tunnel where he had been standing. A work boat pulled him from the river.

He was taken to a recompression chamber on the surface. He received treatment. He was discharged. He returned to work.

His name appears in the engineering reports filed by the Construction Department of the Pennsylvania Railroad in October 1905. It appears in the newspaper accounts that covered the incident the following morning. It appears nowhere else that is publicly available. The tunnel he was working in when the blowout threw him into the river is still there.

It runs beneath the Hudson on the same alignment that was driven in 1905. It carries trains into the station built above it. The passengers on those trains do not know his name. Most of them do not know the tunnel has a history at all.

They know they are going to New York, and that the train takes them there through a tunnel. This is the story of how the Pennsylvania Railroad built Pennsylvania Station in New York City. To understand what the construction of Pennsylvania Station required, you have to understand the situation of Manhattan before it existed. Manhattan is an island, thirteen miles long and two miles wide at its widest point.

On its western side runs the Hudson, three-quarters of a mile across in most places. On its eastern side runs the East River, which is not a river but a tidal strait connecting the Upper Bay to Long Island Sound. The island’s rocky foundation is Manhattan schist, one of the hardest metamorphic formations in the northeastern United States, which is why the towers of Lower and Midtown Manhattan rise as high as they do. The schist is close to the surface at the tip of the island and in Midtown, and drops far down in the intervening area, which is why the tall buildings cluster where they do.

By 1900, Manhattan had two million people. It held the banks, the stock exchange, the insurance companies, the wholesale markets, the newspaper offices, the law firms, the shipping agents, and the import houses that made it the center of American commerce in a way no other city could match. A businessman who needed to reach Manhattan had no equivalent alternative. There was no way to make the city come to him.

The Pennsylvania Railroad was in 1900 the largest railroad in the world by most measures: revenue, employees, miles of track, freight tonnage, passenger volume. It tied Philadelphia to Pittsburgh, to Chicago, to Washington, to a hundred medium-sized cities, with the discipline and financial power of an institution that believed precisely that it was essential infrastructure. It was called the standard railroad of the world. The name was not modest.

But the Pennsylvania Railroad stopped at the Hudson. Its main station in the New York area was at Exchange Place in Jersey City, on the New Jersey shore. Passengers arrived at Exchange Place on Pennsylvania trains from Philadelphia, Washington, Pittsburgh, Chicago, and everywhere else the railroad reached. They stepped off the trains.

They walked to the ferry slips. They boarded ferries that crossed the Hudson to the station piers on the west side of Manhattan. The trip took fifteen to twenty minutes in clear weather. In fog, it took longer.

In winter ice, longer. In the traffic of a port handling a large share of the nation’s imports and exports, slower still. It was always the last obstacle between the passenger and the city. The New York Central did not have this problem.

The New York Central came down the east bank of the Hudson from Albany and entered Manhattan at 97th Street on a track along the edge of the island. Its tracks ran into Grand Central Terminal at 42nd Street and put its passengers onto the street in Midtown. The tracks had been elevated at first, then put below grade at the city’s demand, but they were inside Manhattan. The Pennsylvania Railroad was not.

Alexander Johnston Cassatt became president of the Pennsylvania Railroad in June 1899. He was fifty-nine years old. He had retired from the railroad once, in 1882, and spent seventeen years managing his racing stables in Pennsylvania and traveling in Europe. In Europe, he saw electric railways operating in tunnels beneath Paris and London, penetrating cities that steam trains could not enter because of smoke, noise, and the danger of the grade crossings steam operations produced.

He saw that the technical problem of bringing a railroad into a crowded city had been partially solved by the Europeans. He returned to the Pennsylvania Railroad in 1899 convinced that the tunnel approach could work under the Hudson, and that the railroad’s long-term competitive position required it. The New York Central was inside Manhattan. The Pennsylvania Railroad was across the river.

That was not a sustainable arrangement. In October 1901, Cassatt announced the project. A new passenger station between 31st and 33rd Streets in Midtown Manhattan. Tunnels under the Hudson connecting it to New Jersey and the Pennsylvania main line.

Tunnels under the East River connecting it to Long Island and the New Haven Railroad connection into Boston. The whole to be designed by McKim, Mead & White and built at the expense of the Pennsylvania Railroad. The engineers’ initial cost estimate was about $40 million. The project eventually cost $114 million.

The first problem was the Hudson. The river at the crossing point is between three-quarters of a mile and a full mile wide. The water depth in the main channel ranges from fifty to seventy feet. Beneath the water lies a bed of river silt and clay, unconsolidated, water-saturated, running in places thirty to forty feet deep before reaching bedrock.

The bedrock was where the tunnel had to go. Driving through the silt to reach the rock was the greatest engineering challenge. Tunneling through silt is nothing like tunneling through rock. Rock holds its shape when cut.

You can drill and blast in the schist, and the face stays stable while you remove the broken material and install the lining. Silt does not hold its shape. Silt is a saturated material that wants to flow into any cavity you open in it, and the pressure of the river above pushes it to flow faster. You cannot blast silt.

You cannot dig it by conventional means. You have to hold the river back while you advance into the material. The only way to keep the river out of a tunnel cavity sixty feet beneath the water in 1904 was compressed air. The engineering solution was the tunneling shield: a cylindrical steel device twenty-three feet in diameter, with a cutting edge at its front face and a set of hydraulic jacks at its rear bearing against the completed lining behind it.

Workers at the face of the shield dug the silt by hand or with small pneumatic tools, removing the material and loading it into muck cars that ran on rails back to the main shaft. Once the face was cleared, the jacks pushed the shield forward into the excavated space. Behind the advancing shield, crews erected rings of cast-iron lining, bolting them together to form the completed tunnel wall. Each ring was about eighteen inches wide.

Each advance of the shield allowed one or two rings to be set. The tunnel grew ring by ring. There were four North River tubes. Each ran from a deep shaft in the Palisades at Weehawken on the New Jersey side, under the Hudson, to the tunnel entrances beneath Manhattan.

Two tubes were assigned to eastbound trains, two to westbound. All four were driven simultaneously to meet the schedule the Pennsylvania Railroad had committed to, and which investors expected to be kept. The men who worked at the shield faces, the muckers, and the ring-erection crews inside the compressed-air tunnels were called sandhogs. It is the name the tunnel workers gave themselves, and the name the construction industry used for them.

It describes the men who worked in the compressed-air environment of tunnels and caissons on American construction projects from the 1870s onward. The men who worked at depth on the Brooklyn Bridge and died of decompression sickness before anyone understood what was killing them. Their successors in the trade inherited the skills and the knowledge of how to survive, paid for in lives by the generation before. By 1904, the core of New York’s sandhog community was skilled Irishmen.

Irish immigrants and their sons had flocked to tunnel work in the 1870s and 1880s, when underground projects in New York and Boston created a market for men willing to do this kind of labor. They stayed in it because the pay was higher than surface construction, and because sandhog skills were not easily transferable to other industries that had no use for them. The men who developed those skills followed compressed-air work wherever it existed. They worked on the East River gas tunnel in the 1890s.

They worked on the Manhattan Rapid Transit tunnels from 1900 onward, and on the Hudson & Manhattan Railroad’s tubes under the North River beginning in 1904. When the Pennsylvania Railroad’s tunnels began on the same stretch of river in the same year, the sandhog workforce was the same pool of men, split between competing projects running simultaneously in the same river. They worked alongside Italian immigrants. Italian immigration to the United States had been large and continuous since the 1880s, and by 1904 Italian-born men formed a substantial part of the construction workforce in every major city in the Northeast.

They dug. They removed muck. They handled the lining rings. They operated the muck cars.

They did the physical labor the tunnel advance required, work that did not demand the specific compressed-air expertise of the old sandhogs. They were paid the wages paid to immigrant construction workers in New York in the first decade of the twentieth century: about a dollar and a half a day for ordinary labor, with a modest premium for compressed-air work. Whether that premium adequately compensated for the physiological risks of working under pressure is a question the payroll records do not answer. The records were written by the men who set those wages.

The physiology of compressed-air work is straightforward. The consequences of getting it wrong are not. At the pressures required in the North River Tunnels, between twenty and thirty-six pounds per square inch above atmospheric, the nitrogen that makes up about seventy-eight percent of the air the men breathed dissolves into the blood and tissues at a rate proportional to pressure and exposure time. The longer a man stayed in the compressed air, and the higher the pressure, the more nitrogen dissolved.

When the worker returned to atmospheric pressure, that nitrogen had to come out of solution. It leaves slowly through the lungs as the blood carries it from the tissues to the respiratory surface. If the ascent is too fast, the nitrogen comes out of solution faster than the lungs can eliminate it. It forms bubbles.

The bubbles form in the joints and surrounding tissues, and their pain is what gave the condition its common name: the bends, because the pain forced men to bend over. The knees, hips, and shoulders were common sites. But bubbles can form anywhere the blood carries dissolved nitrogen. In the spinal cord, they caused paralysis.

In the cerebral blood vessels, strokes. In the coronary vessels, heart attacks. A man suffering acute decompression sickness on leaving a North River tunnel in 1905 needed immediate recompression in a surface chamber, followed by a slow, controlled decompression over hours. Not everyone reached a chamber in time.

Not all chambers were staffed adequately at all times. Not all workers who showed symptoms were immediately recognized as suffering from a distinct medical condition, because the symptoms could resemble other illnesses. And a worker who wanted to get home after a compressed-air shift did not always mention that his knee had been hurting since he left the lock. The decompression protocols of 1904 were better than those of 1870, when the Brooklyn Bridge caisson workers died without anyone understanding the mechanism that was killing them.

They were less precise than the navy tables published in 1908, which established the staged decompression times required for various depths and exposure durations, and which became the standard for compressed-air work for decades. The men building the North River Tunnels in 1904 and 1905 worked under the protocols available to them. Those protocols were administered by contractors whose first commitment was the schedule, and whose commitment to the workers was secondary, in an ordering that contained no ambiguity. The injuries are in the records available to researchers.

The deaths are in those records too. Their exact numbers are not in general circulation. What is in general circulation is the tunnel on the Manhattan side of the river. The problem on that side was different.

The station required a site. Cassatt and McKim selected the city blocks between 31st and 33rd Streets and from Seventh to Ninth Avenues: eight city blocks in the middle of Manhattan. In 1902, those blocks contained tenements, stables, small factories, commercial properties, and at least one slaughterhouse. They held the lives and businesses of a great many people who were not part of the Pennsylvania Railroad.

The railroad spent nearly $30 million acquiring those properties, through negotiation where negotiation produced agreement, and through condemnation proceedings where it did not. Once the properties were acquired and the site cleared, excavation began. The station tracks were to lie roughly thirty feet below street level. The foundations had to reach bedrock.

The bedrock was Manhattan schist, and in this part of Midtown it lay at varying depths below the surface. Reaching it required removing it. Removing it required blasting. Blasting rock in the middle of a city that kept operating on all sides of the construction fence was a controlled demolition operation across eight city blocks.

Three and a half million cubic yards of earth and rock were removed from the Pennsylvania Station site between 1904 and 1908. The excavation crews used steam shovels for the softer material and dynamite for the schist. Horses and wagons carried the muck from the site to barges that hauled it away. The men who loaded the wagons, drove them, and cleared the blasted rock from the pit after each blast were mostly Italian immigrants.

Men who followed New York’s expanding construction economy from job to job as work became available. They worked in the pit while the city moved around them in the streets at the edge of the construction fence. They breathed the dust of blasted schist and the cold smoke of the steam shovels. They worked in conditions that would have shut down any modern construction site under safety regulations that did not exist in 1905.

The men injured during the excavation are recorded in the construction companies’ accident records. They appear in no other document that has been widely available. Charles Follen McKim was fifty-five years old when the commission was given to his firm in 1902. His career had been spent building in the Beaux-Arts tradition, the American interpretation of French academic architecture that by 1902 had become the preferred language of major American institutional buildings.

He had designed the Boston Public Library. He had designed the second Madison Square Garden. He had designed enough massive permanent structures that American institutions called his firm first when they wanted something that would stand for a hundred years. For Penn Station, he looked to Rome, specifically to the Baths of Caracalla, a third-century imperial bath complex that covered twenty-seven acres of the Caelian Hill and could hold 1,600 bathers at once in a complex of pools and halls, roofed in concrete vaults at a scale unprecedented in Western architecture until the nineteenth century.

McKim took the proportional system of the Caracalla halls and applied it to the design of a railroad waiting room in the middle of Manhattan. The main waiting room of Pennsylvania Station was 314 feet long, 108 feet wide, and 150 feet from the floor to the top of the coffered vault ceilings. Light filtered through iron-framed windows set into the curves of the vaults and through the glass roof of the train shed below, falling from above in a quality of light unlike any other interior space in New York. The exterior was built of pink Milford granite quarried in Massachusetts.

The interior surfaces were clad in travertine marble. The columns of the exterior colonnade were thirty-five feet high, Doric, carved from the same granite. The quarrymen who cut that stone are recorded in the stone contractors’ payroll records. They appear nowhere else.

The ironworkers who erected the glass-and-steel train shed roof worked from scaffolding at heights sufficient to kill any man who fell. The train shed was structural ironwork of the same class as bridge work, done by the same community of men who did that work on other sites. Some of them fell. The falls are recorded in the accident reports.

The men who fell from the iron structures of the Penn Station train shed roof are not on the structure they were building when they fell. Alexander Cassatt died on December 28, 1906, at his home in Philadelphia. He was sixty-six. The tuberculosis that had been consuming him during the final years of the project had at last taken his ability to work, and then finished the rest.

He had held the project together through cost overruns, engineering complications, negotiations with the city, and board members who questioned the financial logic. He had committed the resources of the Pennsylvania Railroad to a project whose total cost exceeded the original estimates by an amount that would have stopped the project had it been known in 1901. The project did not stop. His successors recognized what had been built and why it had to be completed.

The money was spent. The tunnels were nearly done. The station structure was rising. The logic was as sound in 1907 as it had been in 1901.

The men who followed him finished what he had started because they saw that he had been right. He never saw it completed. The East River tunnels connected the station to Long Island and the New Haven Railroad connection at Sunnyside Yard in Queens. Four tubes under the East River, as there were four under the Hudson, two in each direction, but the conditions were different.

Under the East River, the tunnel route ran mainly through rock, the same Manhattan schist the island rests on. Hard-rock tunneling differs from silt tunneling in most respects. In hard rock, you drill and blast rather than using a shield and mucking. The drill crews went in first, advancing along the rock face with pneumatic drills, boring the pattern of holes that would be filled with explosive charges.

The powder crews placed the charges and set them off, the blast inside the enclosed tunnel something the men felt in their bodies as much as heard. The mucking crews removed the broken rock. The concrete crews followed behind. Then the drill crews returned and the cycle repeated.

The East River tunnels were wet. Immersed tunnels under water are always wet, because seepage through cracks in the rock is continuous, and pumps ran constantly to control it. The men worked in ankle-deep water in the low spots, with a steady drip from the tunnel roof above them. Two headings advanced from opposite ends of each tube, one from the Manhattan shaft and one from the Queens shaft, and they had to meet in the middle under the river with an accuracy of no more than a few inches in all three dimensions.

The survey work that controlled that accuracy was done with 1905 instruments: the transit, the level, and the steel tape, used inside the tunnel to define the line and grade each heading had to follow. When the headings met, they met exactly where the survey calculations had predicted. The mathematics was correct. The construction followed it.

On November 27, 1910, Pennsylvania Station opened for service. President William Howard Taft was present. He had come from Washington on a Pennsylvania Railroad train that arrived through one of the North River tubes, stepped onto a platform below the concourse, walked up through the concourse and the main waiting room, and delivered a speech about what the Pennsylvania Railroad had accomplished. The governor of New York was there.

The railroad officials were there. The chief engineers were there. The architects were there. The opening was covered by newspapers across the country.

The photographs show the main waiting room with its vaulted ceiling, the light filtering through the iron-framed windows in the corners of the vaults, the scale of the interior space, three hundred feet long and one hundred fifty feet high to the coffered ceiling. The people in their 1910 clothing stand inside it looking small, the way people always look small in photographs of vast spaces. The photographs do not show the excavation workers. The men who drove the North River Tunnels through the river silt beneath the Hudson in 1904, 1905, and 1906 were not in New York as a group on November 27, 1910.

They had moved through the construction economy of the Northeast from project to project as jobs began and ended, because the work is the point, and when one job is finished another begins somewhere else. A tunnel worker who finished his last shift in a North River tunnel in 1907 did not stay in New York for three years waiting for the opening ceremony. He went where the next compressed-air job was. The Italian workers who excavated the station site were in whatever city the construction economy had carried them to by 1910.

The quarrymen from Massachusetts were in their quarries. The ironworkers who built the train shed roof were working on whatever steel structure was rising somewhere else. The ribbon was cut for the people whose names appear in the engineering reports and the ceremony programs. The men whose names appear in the payroll records and the accident records had already gone.

That is how construction work goes. Pennsylvania Station served its passengers for fifty-three years. In that time, it handled over one hundred million people. The Pennsylvania Railroad’s New York traffic moved through the North River and East River tunnels, through the platforms beneath the concourse, through the waiting room, and out onto Seventh Avenue and into the city.

The station was used the way Cassatt had intended it, as a terminal that delivered Pennsylvania Railroad passengers directly into Manhattan, and the volume of that traffic justified its construction cost. The waiting room was a space that New Yorkers used the way public spaces of a certain quality and scale are used: not merely as a passage, but as a place, a room in the city where you could remain for a while before having to go somewhere else. The light from the arched windows, the mass of the ceiling, the quality of the stone. New Yorkers do not as a rule sentimentalize public places, but Penn Station bred in its users something close to attachment, which is not the same as sentimentality.

They were attached to it because it was good, and good spaces are not common in cities. In 1955, the Pennsylvania Railroad began to lose money. The automobile had taken the short-distance travelers. The airlines had taken the long-distance travelers.

The freight revenues that had subsidized passenger operations were under pressure from the trucking industry. Railroad costs were fixed. Revenues were shrinking. The real estate occupied by Penn Station in the middle of Manhattan was worth enormous sums to any developer who wanted to build a skyscraper in Midtown.

The air rights above the station’s low profile were worth more. The Uris Building Corporation proposed building Madison Square Garden and an office tower on those air rights. The Pennsylvania Railroad was to receive the sale proceeds while retaining the underground station and the tunnels. In July 1961, the Pennsylvania Railroad announced the agreement.

The public response was protest of a kind that urban architectural decisions rarely provoke. Ada Louise Huxtable wrote in the New York Times about what was planned. Architects organized demonstrations. The Municipal Art Society ran protest advertisements.

The users of Penn Station recognized that the vaulted waiting room was to be demolished for money the railroad needed, and they objected in numbers beyond what the company had expected. It did not change the outcome. The railroad needed the money. On October 28, 1963, the demolition crews began work.

The pink Milford granite columns of the exterior colonnade fell. The vaulted ceiling of the main waiting room fell. The iron-and-glass train shed roof fell. The travertine marble was hauled to the New Jersey Meadowlands, where pieces of it are still buried in the ground if you know where to look.

When the last trace of the above-ground structure was gone, Madison Square Garden was built on the site. The office tower went up beside it. At the base of both, below grade, is the present-day Penn Station. Huxtable wrote at the time: “We will probably be judged not by the monuments we build but by those we have destroyed.

”

The demolition of Penn Station led within two years to the creation of the New York City Landmarks Preservation Commission in 1965, a direct institutional response to what happened on 32nd Street. That commission gave New York the legal mechanism to designate buildings as landmarks and prevent their demolition without formal review. The mechanism that protects Grand Central Terminal from Penn Station’s fate exists because Penn Station was already destroyed. The agency that preserves New York’s architecture was founded on the ruins of the building that Richard Seddon’s tunnel had served.

The men who built what was demolished had no comparable institution protecting the record of their work. If you stand on a platform of today’s Penn Station and look at the tunnel portal at the end of the track, you are looking at the mouth of one of the four North River tubes. The tube is twenty-three feet in diameter. It was driven through the silt of the Hudson in 1904 and 1905 by men working inside a shield under compressed air at pressures that dissolved nitrogen into their blood, requiring controlled decompression when they came out.

The cast-iron lining rings those men erected behind the advancing shield are still in place. They have held the river for 120 years, and they still hold it now. The station above you has a low ceiling and fluorescent lighting, and it serves its purpose. It handles more passengers than any other rail facility in North America.

The people in it are going somewhere. They stand on the platforms and board the trains, then the trains enter the tunnel portals and go under the Hudson and out to New Jersey. They do not think about what the tunnel cost to build or who built it. The vaulted waiting room is gone.

The light that filtered through the iron-framed windows in the corners of the vaults is gone. The pink granite columns are gone. The travertine marble is buried in the ground in New Jersey. The tunnels remain.

Richard Seddon went back to work after the blowout that threw him from the surface of the Hudson in October 1905. His name is in the engineering reports. It is in the newspaper account of the accident the following morning. It is in no other document that has been widely available.

The tunnel he was building when the blowout happened is still there, carrying the trains of a railroad that was merged and renamed and reorganized to become the system that serves the station that replaced the one these tunnels were built to feed. The work he did still stands. His name is not on it.

The tunnels are still there.