In the fall of 1887, a man stood on a wooden platform one hundred and sixty feet above the Hudson River in Dutchess County, New York, driving the pin that would connect two halves of a cantilever structure reaching toward each other from opposite directions over the fourth river span. The platform had no railing on the river side. It did not need a railing to perform its function. The man driving the pin needed to reach the bolt hole, and he was on the correct side of the platform to do so.

The river below him moved toward the sea at the speed the tide ran that evening. In the fall, the Hudson runs with the tide in both directions at Poughkeepsie. At the bridge site the river lay two hundred and twelve feet below the rail level he was helping to build. A fall from the work platform to the water would not have been two hundred and twelve feet.

The distance would have been exactly the height at which he stood above the surface. When the Poughkeepsie railroad bridge opened to traffic on January 1, 1889, it was the highest railroad crossing of the Hudson River and among the longest railroad bridges in the United States. It crossed the Hudson through five main cantilever spans between six stone piers, with approach bridges of wrought-iron trusses extending the structure from the river spans to the western shore at Highland in Ulster County and the eastern shore at Poughkeepsie in Dutchess County. The total length of the crossing from abutment to abutment was 6,767 feet.

The rail deck stood two hundred and twelve feet above high water at the main channel span. The river beneath the middle span was sixty-eight feet deep in its natural state. There was a swing span, required by the Army Corps of Engineers as a condition for permitting the bridge, located in the third opening from the eastern shore. It was 525 feet long and rotated around a central pier to open a navigational channel for river traffic when the mechanism was operated.

The bridge was the solution to a problem that had constrained freight movement in the northeastern United States since the beginning of railroad construction. The men who built it, the compressed-air workers in the caissons beneath the river, and the iron crews on the structures above, were documented in the records of the Poughkeepsie Bridge Company to the extent that maintaining those records required. To understand the problem the Poughkeepsie Bridge was built to solve, one must understand the Hudson River as an obstacle. The Hudson runs north to south through the geographic center of New York State, separating New England from the continental interior.

At its southern end it empties into New York Harbor. The railroad lines of southern New England, those serving the industrial cities of Connecticut, Massachusetts, and Rhode Island, the textile mills of the Connecticut River valley, and the foundries and machine shops of Springfield, Providence, and Worcester, reached the Hudson at its tidewater mouth, which was also the mouth of New York Harbor, to move goods between the industrial interior of New England and the coal fields of Pennsylvania, between the mills of the Connecticut valley and the grain elevators of Chicago and the ironworks of Pittsburgh. Freight had to cross the Hudson. In 1870 it crossed by ferry.

In 1880 it crossed by ferry. The terminal facilities and freight transfer operations at New York Harbor were among the most congested points in the American railroad network. Every freight car moving from New England to the Pennsylvania coal fields or the industrial cities of Ohio had to pass through the harbor terminals and be maneuvered at least twice. It was loaded onto a car float, towed across the water, then unloaded from the float at a terminal on the opposite shore, re-sorted, and coupled into a departing train.

Time, damage, cost, and delay accumulated in a terminal system handling more cars than it was designed to absorb because the manufacturing economy of New England had outpaced the capacity of the ferry that served it. The bottleneck was structural. It was the physical consequence of the Hudson lying at a point where geography forced all east-west freight movement through a single river mouth over which no bridge could be built. The harbor was a seaport.

The lower Hudson was a navigation channel. A railroad bridge over the lower Hudson in the 1880s would have been a structure obstructing one of the busiest commercial waterways in North America. The Corps of Engineers would not have permitted it. The navigation requirements of the harbor would not have accommodated it.

The ferry crossing at New York Harbor remained the only option because no alternative existed below the Hudson Highlands. The Hudson narrows as one moves north. Above the harbor, above the wide expanse of the Tappan Zee, and through the Hudson Highlands at West Point, the river cuts through a mountain chain the glaciers had not removed. The valley walls descend to the water, and the river runs between hills of Precambrian gneiss and granite.

Above the Highlands, the valley opens again into the flat agricultural lands of the mid-Hudson, the river valley the Dutch had farmed in the seventeenth century and that Hudson Valley manor estates had occupied since before the Revolution. At Poughkeepsie, the river was wide, but not impossibly wide. Rock was available on both banks for pier foundations. The valley walls on both sides offered approach grades a railroad could climb without exceptional engineering.

The city of Poughkeepsie on the eastern bank already lay on the main line of the New York Central and Hudson River Railroad. A crossing at Poughkeepsie would connect to existing railroad infrastructure on both sides of the river and would lie north of the Highlands, where the navigation requirements of the lower harbor did not apply with the same absolute force. The Poughkeepsie Bridge Company was organized to build precisely that crossing. It was incorporated by railroad interests that wanted an independent connection across the Hudson above the New York terminals, a route allowing freight to move directly between the New England connections and the main lines west and south of the Hudson without entering the New York harbor freight system at all.

The economic case was the avoidance of costs. Every ton of freight crossing at Poughkeepsie instead of the New York terminals saved the transfer fees, the time cost of the transfer operation, the congestion delays the terminals imposed on everything passing through, and the damage to goods from repeated handling. For freight moving between the manufacturing cities of Connecticut and the coal fields of Pennsylvania, the Poughkeepsie crossing removed an obstacle that had added cost to every ton since the beginning of the railroads. The company obtained its charter and appointed an engineering staff.

The survey located the crossing on the river at Poughkeepsie where geology and valley engineering permitted. Design work began. The engineering problem posed by the Hudson at Poughkeepsie differed in kind from those at the Susquehanna at Rockville or the Patapsco at Relay. The Susquehanna at Rockville was a wide, shallow, braided river with a bed of gravel and stone, easily accessible for building temporary cofferdams at working depths that horse-powered pumps could keep dry.

Forty-eight elliptical stone arches with seventy-foot spans, built in cofferdams that were dewatered between the falling spring flood and the rising autumn flood, were the structural solution for a crossing where materials, depth, span requirements, and construction method allowed the use of masonry. The Hudson at Poughkeepsie was not that river. The main channel at the crossing site reached depths of sixty-eight feet at ordinary stage. The river was tidal.

The tidal variation during a working day changed water levels in ways a construction plan had to absorb. The bearing rock beneath the river sediments lay at depths that temporary cofferdam construction with methods available in the mid-1880s could not practically reach, and the Corps of Engineers, in the bridge permit, required that the river remain navigable for commercial traffic for the entire construction period. Any temporary dam extending across part of the main channel during construction would have closed navigation. That was not permitted.

The structural form of the Poughkeepsie crossing therefore could not be a stone arch. A stone arch required temporary falsework, temporary guiding structures supporting the arch ring until the keystone was set and the mortar cured to bearing strength. Falsework in a tidal, navigable river sixty-eight feet deep was not the solution. The form had to be erectable from above without support from below.

The form that could do this, that could cross several hundred feet over a navigable channel and be built outward from pier tops with no support from the river surface, was the metal cantilever truss. The cantilever principle had been proven in long-span American railroad bridges during the 1870s and 1880s. The Kentucky River bridge of the Cincinnati Southern Railway, completed in 1877, had used cantilever construction to cross a gorge where falsework from below was impractical. The cantilever method allowed the truss to grow outward from each pier in both directions, balanced around the pier top with no support from below.

From each pier, truss arms extended outward over the adjacent spans. When arms from neighboring piers approached each other at mid-span, a shorter suspended truss was installed between them, hung from the ends of the arms by pin connections. The suspended span transferred its weight to the cantilever arms. The cantilever arms transferred that weight through compression in the top chord and tension in the bottom chord to the pier tops and down through the masonry to the foundations.

The load path ran continuously from the rail deck to the river bottom. The river remained open throughout construction. The navigation channel remained navigable. The entire bridge was assembled in the air above the waters it crossed.

Five main cantilever spans of this type would cross the Hudson between six stone piers. The pier foundations posed the construction problem that the cantilever truss design had avoided. The problem of how to found six stone piers in sixty-eight feet of tidal river water on a rock bottom that no cofferdam could reach. The solution was the pneumatic caisson.

Pneumatic caissons had been used in major American bridge foundations since the East River caissons of the Brooklyn Bridge in the early 1870s, and before that at the Eads Bridge in St. Louis beginning in 1869. The principle was the same at Poughkeepsie in 1886 as it had been at the Eads Bridge seventeen years earlier. A steel working chamber was set on the river bottom at the pier site.

The chamber was sealed and pressurized with compressed air pumped from the surface. The compressed air, at a pressure sufficient to balance the hydrostatic pressure of the water above, kept river water from entering the working space inside the chamber. Workers entered the working chamber through an airlock, a small compartment between the pressurized interior and the atmospheric surface that allowed pressure to be equalized in both directions, and worked in the chamber at the base of the pier being built. They excavated riverbed material by hand with picks, shovels, and handcarts, and removed it through a separate muck outlet that allowed excavated material to be raised to the surface under pressure.

As excavation deepened, the pier structure above the caisson forced it down under its own weight toward the founding level. When the working face of the caisson reached the bearing stratum, the rock beneath the river sediments, the foundation excavation was complete. The working chamber was filled with concrete. The pier shaft was built upward from it in masonry.

The foundation was thus created. The men who worked in the caissons at Poughkeepsie in 1886 worked in compressed air at pressures demanded by the water depths above the caisson roof. At the depths encountered in the main channel, the air pressure in the caissons reached levels at which medical references in the 1880s were beginning to study the physiology of nitrogen absorption under elevated pressure, and its connection to the condition caisson workers called the bends. The relationship between decompression sickness and the formation of nitrogen bubbles in the blood and tissues was not yet understood in the mechanical terms that later physiological research would provide.

What was understood, empirically and imperfectly, was that men who left the airlock too quickly, reducing the pressure on their bodies too rapidly upon exiting the working chamber, developed symptoms ranging from severe joint pain that could bend a man’s body, to paralysis, and even death. What was understood was that the required decompression time, a specified duration of slow pressure reduction in the airlock during exit, reduced the incidence of the condition but did not eliminate it. Whether the decompression time set by the construction management was the same time the men actually spent in the airlock was the gap between specification and application that existed at every pneumatic caisson project of the period. The men working in the compressed-air chambers at Poughkeepsie were paid by the unit of work completed where piece-rate systems applied, or by the day at a premium wage above surface labor rates that the nature of compressed-air work commanded.

This premium wage was the recognition of risk in the terms the labor market of 1886 used to recognize risk. A man who spent the required time in the airlock decompressing was a man spending that time without working, and therefore without earning. The man who could not afford to spend time without earning money was the man who moved through the airlock faster than the prescribed rate. The consequences of that choice fell unevenly across the workforce, concentrated among the men under the greatest economic pressure to accept the risk, and were recorded in the construction’s medical records to whatever extent record-keeping required.

The excavation work in the caissons of the Eads Bridge in St. Louis between 1869 and 1874 had killed twelve men from decompression sickness and left many more permanently disabled. Poughkeepsie’s caisson excavation proceeded with better empirical understanding than the Eads Bridge foundations, but under the same fundamental conditions: compressed air, excavation under pressure, the airlock as the only link between the working environment and the surface, and workers making decisions about decompression time in the chamber that construction management could not fully determine or enforce. Men emerged from the Poughkeepsie airlocks bent over by the condition called caisson disease in the construction records to the extent that 1886 record-keeping required for documenting an injured worker.

Many of them never appeared in the records at all. The workers who performed the caisson work at Poughkeepsie came from the class of men available to the heavy construction labor market of the Hudson Valley in the mid-1880s. Irish immigration in the 1840s and 1850s had left a labor concentration in the Hudson Valley and the cities of the Northeast that, by the 1880s, had partly settled into the trades and political structures of port cities while still partly working in itinerant construction labor. Ireland-born workers and American-born sons of Irish immigrants supplied the general labor and a share of the skilled excavation and construction trades in the caisson foundations at Poughkeepsie.

Italian immigration into the northeastern construction labor market was growing during the decade but had not yet reached the scale it would attain in the 1890s. The caisson workers of Poughkeepsie in 1886 were the generation before that shift, the end of the Irish construction workforce and the beginning of the Italian workforce, working in compressed-air chambers beneath a Hudson River crossing that neither group would have chosen to build had the wages not been what they were. The compressed-air wages at Poughkeepsie were the wages set by the Hudson Valley construction labor market of 1886 for men willing to work beneath the river in conditions that the medical understanding of the era could not fully explain and engineering management could not fully control. They were higher than surface labor rates.

They were the rate the contractor agreed to pay and the men agreed to accept. The difference between the agreed wage and the wage actually received, after the contractor deducted the cost of housing, provisions, and any advances against future earnings, was a difference a worker could verify only if he could read the contractor’s books, which most of the tunnel diggers at Poughkeepsie in 1886 could not. The iron crews that erected the cantilever trusses over the Hudson differed in composition from the diggers in the riverbed. They were specialists in a trade that the expansion of American railroading had created continuous demand for since the beginning of the great iron bridge construction in the decades after the Civil War.

Bridge crews traveled from one major structure to another across the railroad network as work required. They went where the iron was being raised. The Poughkeepsie crossing in 1886 was where the iron was being raised, and the men capable of working at heights on iron structures came to it as skilled tradesmen follow workplaces through the internal communication that carries news of coming jobs and their wages. The riveting gangs that closed the connections of the Poughkeepsie cantilever trusses worked in four-man teams.

The heater kept the rivets in a portable furnace at a temperature that made the iron workable. If too cold, the rivet hardened before its head was fully driven. If too hot, the metal became soft at the moment it needed to be tight. The catcher received the hot rivet thrown by the heater using tongs or a metal can, placed it in the hole, and stepped aside.

The holder-on held a bucking bar against the manufactured head of the rivet on the back side of the connection, applying counterpressure that allowed the head to be formed properly against the other side, while the riveter drove the head with a hand hammer, or a pneumatic hammer if pneumatic equipment was used, drawing compressed air from a hose connected to a compressor on the pier top or on a boat in the river below. This process required the coordination of four men on a platform that might consist of wooden planks tied to structural members with ropes, secured by nothing more than the strength of the rope lashings themselves. The platform one hundred and sixty feet above the Hudson in the fall of 1887 was sufficient to support the riveting gang, their tools, and the weight of the work they performed. The word sufficient was what the construction schedule required and the project budget provided.

The railroad on the river side of the platform was not part of the equipment required to close the pin connections in the fourth cantilever arm above the Hudson. The man driving the pin on the river side of the staging drove it there because that was where the connection point was located. He had been making such connections at extreme heights on iron structures throughout the time he had spent in the iron bridge building trade. The risks were not unknown to him.

This was the essence of the work. The fabricated truss members that the iron crews erected over the Hudson were produced in a specialized bridge plant and reached the construction site via the existing railroad network. Chord sections, web members, pin plates, and lattice connections arrived at the site by train or barge, were lifted from the delivery point by construction derricks located atop the pier caps, and were placed into position for the connection crews by men operating the derrick lines and guiding the swing of the loads. A chord section weighing several tons swinging from the head of a construction derrick two hundred feet above the Hudson was not a load that could be controlled the way a hand tool could be controlled.

It was a load that could be managed. The difference between controllable and manageable was what the men operating the derricks and the men receiving the loads on the staging understood through their daily experience and expressed in the habits of movement and positioning that allowed them to complete the work without being struck by the loads they handled. These habits were not written in any records. No law of New York State or regulation of the federal government in 1886 required them to be.

They were professional practices passed down within the crews from the men who had learned them from the men who came before. The five cantilever openings grew outward from the tops of the six piers over the five river bays through 1886 and into the 1887 construction season. As the cantilever arms from adjacent piers reached toward each other at mid-span, the erection engineers managed the sequence of member installation to control the stresses within the structure that was not yet complete. A cantilever arm weighing several hundred tons of steel and extending over the Hudson without a matching arm on the approach side would have imposed moments on the pier foundations that those foundations were designed to resist only in the completed structure, not in a structure still under construction.

The erection sequence was the engineering document that turned design intent into physical reality in construction. The foremen executing that sequence above the pier caps read the erection drawings produced by the engineering staff and made daily decisions about which member to install next based on the requirements of the drawings and what construction conditions permitted. The approach bridges extending the structure from the river spans to both shores were built on stone piers and iron trestles mounted along the valley walls at each bank. The western approach climbed the hills of Ulster County from river level to the bridge deck height on iron trestles rising from the hillsides below rail level.

The eastern approach descended from the bridge deck to the railroad connections in the city of Poughkeepsie through shorter truss bridges carried on stone piers. The approach structures were not the real engineering challenge of crossing the river. They were the work that had to be done before the river spans became useful, the connections that turned the 6,767-foot structure into a railroad crossing rather than a freestanding bridge over the Hudson. The swing span in the third opening from the eastern shore was an acknowledgment built into the structure that the river it crossed had commercial users whose interests predated the railroad.

The steamboat trade of the Hudson between New York City and Albany had operated since Robert Fulton’s first commercial voyage in 1807. The brickworks of the Hudson Valley had been shipping their products from river wharves since before the era of canals. The ice harvest that supplied Manhattan with refrigeration before mechanical systems replaced it was taken from the winter ice of the Hudson above the Highlands. The Poughkeepsie swing span extended 525 feet of bridge that rotated around its center pier to open a navigation passage when river traffic required it.

The right of river navigation through the passage was older than the bridge itself. The swing span was the form that right took in response to the federal permit requirement. The bridge opened to railroad traffic on January 1, 1889. Its first revenue trains were the through freights between the New England railroad connections and the main lines west and south of the Hudson that the Poughkeepsie Bridge Company had been organized to serve.

The route ran from the railroad systems of Connecticut and Massachusetts through connecting lines to the eastern end of the bridge, across the Hudson at Poughkeepsie 212 feet above the river, and from the western end via the Central New England and Western Railroad to connections with the Erie Railroad and other western lines. The bottleneck was not eliminated. The New York harbor terminal system continued to handle shipments that could not reach the Poughkeepsie crossing economically or for which connecting routes made the Poughkeepsie route impractical. But the shipments the route served, and for which economic feasibility favored it, crossed the Hudson at Poughkeepsie instead of passing through the ferry terminals, and the costs saved were the basis of the revenue the bridge company had projected.

The New York, New Haven and Hartford Railroad acquired the Central New England Railway in 1904, and with it the trackage rights and connection arrangements that had used the Poughkeepsie crossing since its opening. Under New Haven control, the bridge became part of the through freight route the New Haven used to move coal from the Pennsylvania fields to the industrial market of New England. Coal from the Pennsylvania mining regions moved west and north through Pennsylvania Railroad connections, crossed the Hudson at Poughkeepsie, and reached New England through the New Haven’s connections in Connecticut. The coal traffic was always the foundation on which the viability of a through freight route for a Hudson River crossing north of New York rested.

Coal continued to move across this bridge 212 feet above the river for decades. The Penn Central acquired the New Haven’s properties in 1968. Penn Central went bankrupt in 1970. It was the largest corporate failure in American history at that time.

The Poughkeepsie bridge and the through freight route it enabled passed, after the Penn Central bankruptcy, to the Conrail operating entity in 1976. The bridge had carried railroad traffic for eighty-seven years. The cantilever arms, pinned together by ironworkers above the Hudson in 1886 and 1887, had carried commercial freight trains since January 1889, through the full development of railroad motive power, from the steam locomotives of the New Haven and its predecessors to the diesel power that replaced steam during the 1950s. The pier foundations, sunk by caisson workers into the bedrock of the Hudson riverbed in 1886, had remained continuously in the river since they were poured.

On May 8, 1974, a fire broke out in the middle section of the Poughkeepsie bridge. Welding operations on the bridge deck, performed by a maintenance contractor, ignited the creosote-soaked timber decking and walkways in the center spans. The fire spread through the combustible material in the decking of the main spans over the river before it was brought under control. The steel structural elements of the cantilever trusses did not burn.

Steel does not burn. But steel, at the sustained temperatures of a creosote-and-timber fire, loses a significant portion of its structural strength. The load-bearing and tensile capacity of a steel member heated to 1974 fire temperatures and then cooled is not the same as that of a member that has never been heated. The post-fire inspection concluded that the thermal damage to the structural steel in the affected spans was sufficient to require major repairs before the bridge could safely return to commercial service.

The cost of repair compared to the revenue the through freight route generated in 1974 was the calculation Conrail made. The calculation did not favor repair. The bridge was closed to rail traffic and remained closed. The structure stood over the Hudson for the next thirty-five years carrying no traffic at all.

The pier foundations remained in the river. The cantilever arms remained in the air. The swing span mechanism sat on its center pier without rotating. The tidal river continued to flow beneath the six stone piers in its natural cycle, the cycle it had followed before any crossing existed above it.

The caisson foundations that compressed-air workers had sunk to the bedrock of the riverbed in 1886 continued to support the pier shafts. The pier shafts continued to support the cantilever arms. Those arms continued to extend over the river. This structure was designed to last, and it had lasted.

It no longer carried trains. In 2009, the Poughkeepsie bridge was reopened as a walkway over the Hudson River State Historic Park. The damaged structural steel was repaired. The deck was rebuilt for pedestrian use.

The swing bridge was inspected, fixed in place, and preserved. The piers, founded by men who worked under compressed air beneath the Hudson in 1886, carried pedestrian traffic on the reopened walkway in the same way they had carried New Haven freight for eighty-five years before the fire. The foundations were sound. The pier proportions were correct.

The masonry had stood in the tidal river for 123 years and continued to perform the function assigned to it by the 1884 engineering design. Two million people have crossed the walkway over the Hudson since 2009. They cross 212 feet above the river on steel. The iron crews connected the bridge over the Hudson in 1886 and 1887.

The Hudson moves below them in the tidal cycle it has always followed. Looking south from the walkway on a clear day, the river disappears around the bends of the valley and continues toward the harbor. It extends eighty miles to the south. No freight train crosses at Poughkeepsie anymore.

The direct freight route for which the bridge was built no longer operates as the Poughkeepsie Bridge Company organized it. The goods the bridge carried between New England and Pennsylvania throughout eighty-five years of rail service now move by other means over other routes. There are interpretive signs on the walkway over the Hudson. They identify the bridge and its engineering history.

The signs record the length, the height, the opening date, and the significance of the cantilever construction. The Poughkeepsie Bridge Company and its engineering department are represented in the terms historical interpretation uses to represent them. The bridge is classified as an engineering achievement. The iron crews who walked the members of the cantilever arms two hundred feet above the Hudson in the construction seasons of 1886 and 1887 are not named on the signs.

The riveting gangs, the heaters, catchers, holder-ons, and riveters who closed the field connections in whatever weather the Hudson Valley experienced over two years of construction, are not named on the signs. The names of the men who worked in the pneumatic caisson chambers beneath the river surface, in the air pressure the water depth demanded, and who left the airlocks at rates the physiology of 1886 could not fully protect them from, do not appear on the interpretive panels. The compressed-air workers appear in the medical records of the construction period, to the extent record-keeping required for documenting a worker’s injury. Many of them are not in those records at all.

The payroll accounts of the Poughkeepsie Bridge Company document the labor costs of the caisson work, the ironwork, the masonry, and the approach bridges, to the extent that financial accounting in 1886 required. The workers appear in those accounts as part of the cost of the work. The iron crews appear at the crew level, where timekeeping tracked crew productivity. The individual members of those crews, the heater who held the proper temperature, the holder-on who braced the bucking bar against the rivet stem, the man who drove the pin on the river side of the work platform without a railing in the fall of 1887, appear in those records to the extent required to track the cost of the connection.

They do not appear to the extent required to track the identity of who did the work. The cantilever arms are still there. The piers stand in the river. The foundations are in the rock beneath the river.

The tidal Hudson continues to run through the space occupied by the six pier shafts in its channel. The swing bridge rests on its center pier the way it has rested since it last rotated, which was before 1974. The structure built by compressed-air workers and iron crews in 1886 and 1887 over the Hudson at Poughkeepsie now carries pedestrian traffic over the same river it was originally built to carry trains across. The record of how it got there exists in the engineering drawings, the cost accounts, and the corporate records of the company that built it, operated it, and ultimately closed it to rail traffic.

The engineering drawings are precise. The cost accounts are complete to the extent the accounting of that period required. The men who did the work exist in those records as costs, crew counts, daily wage rates, and total expenditures submitted by the contractor and paid by the bridge company. They do not exist in those records as individuals who worked in the pressure chambers, hoisted the iron sections, drove the rivets, and pinned the connections two hundred feet above the Hudson.

The work was done. The record of the work is the record the system required to be kept. The men exist in that record as the cost of the work. The bridge remains.

The river moves beneath it.