Florida Keys • Engineering • Disaster • Remembrance
Research reviewed 2026-09-07 · About 95 minutes · 16 original visuals
Florida Keys • Engineering • Disaster • Remembrance
Where the evidence stands: The hurricane and the failed evacuation are documented. This investigation examines how the railway was built, why the camp evacuation failed, what survivors and rescuers recorded, and which questions about responsibility and casualty accounting remain unresolved.
The watercolors are modern interpretations. Evidence graphics distinguish reported observations from later estimates; neither artwork nor a precise-looking number is a substitute for an archival record.
Explore the investigation
- A light on the water
- An island looking outward
- The fortune behind the line
- Before the rails
- A railway built by boats
- The men beneath the achievement
- The storms before Labor Day
- Arches, girders, and the ground below
- A terminus before the terminus
- January 22, 1912
- The city at the end of the track
- Across the window, beyond the harbor
- A different road, a different workforce
- Messages falling from the sky
- The storm the instruments could not fully see
- The house Russell trusted
- A train going toward the storm
- Fifty-five minutes on Long Key
- Ships carried across the reef
- The morning the aircraft could leave
- The search beyond the camps
- The numbers after the names
- Who had been responsible?
- A map over a crypt
- Roy Anderson’s name
- The road on the old piers
- Reading the surviving landscape
- What remained of the promise
- Sources and reading notes
1. A light on the water
On Long Key, the wind gave J. E. Duane a little time.
For hours on September 2, 1935, it had battered the fishing camp where he served as manager and a cooperative observer for the United States Weather Bureau. Timbers had been flying through the air. The water had risen toward the railroad grade. Buildings that had seemed substantial enough in ordinary weather were coming apart. Duane and the people sheltering with him moved from one structure to another, looking for something that would hold.
Then the wind diminished.
In the account later published by the Weather Bureau, Duane put the beginning of that respite at 9:20 in the evening. He could hear sounds that the wind had overwhelmed. To the north, stars were visible. Twenty people gathered in another cottage. The lull offered a chance to move, but it did not mean the hurricane had passed. A different wind would follow when the other side of the storm arrived.
During the interval, Duane directed his flashlight toward the sea. The water was rising from the ocean side of the camp. He described what appeared to be high walls of water. Although he was only about sixty feet from the cottage entrance, he could not get back before the water reached his waist. Soon the cottage lifted from its foundations and floated.
These details survive because Duane recorded them. They require no imagined conversation, no account of what he must have thought when his light found the water. The distance to the door, the movement of the cottage, the direction of the returning wind are enough. They describe a place in which the familiar relationship between land and sea had temporarily ceased to exist. Weather Bureau, McDonald’s 1935 report
Twenty-three years earlier, the Florida East Coast Railway had announced the completion of a route through this same island chain. Its promotional booklet described concrete viaducts, steel bridges, dredged embankments, a new terminal at Key West, and a triumph over the natural forces that separated the islands. Engineers had found a way to carry a train where passengers previously needed a boat. Henry Flagler, who had financed the undertaking, lived long enough to reach Key West aboard the first ceremonial train in January 1912.
The booklet made a larger claim than the fact of completion. It presented the railway as a durable conquest of storm and sea. Its construction section explained how protective slopes, elevated roadbed, and surrounding vegetation would keep the line secure. The engineering descriptions were specific. The confidence was almost absolute. Florida East Coast Railway, opening booklet, 1912
By September 1935, the line had become part of ordinary island life. It carried people and freight. Its stations and communications connected small settlements with the mainland. When a hurricane threatened the veterans working in highway camps, a train was the means chosen to bring them out.
The rescue train reached the Upper Keys. It did not complete the evacuation. The sea swept its cars from the track.
That image has come to dominate the railway’s history: a stranded locomotive, displaced cars, the work of years broken in a night. Yet beginning and ending with the wreckage makes the people between those two moments difficult to see. There were surveyors working before a continuous route existed, laborers living on floating quarters, engineers trying to establish foundations in tidal channels, island residents whose communities predated the railway, and veterans building a different transportation project during the Depression.
Their histories converged in the hurricane. They did not begin there.
To understand what happened on Labor Day, it is necessary to go back beyond the warnings, beyond the evacuation order, beyond even Flagler’s decision to build. Before the railway could become a route to safety, it had to become a route at all. And before a financier could imagine bringing the world to Key West, the island already had a world of its own.
2. An island looking outward
The railway company’s 1912 booklet offered its readers a flattering picture of progress moving south. Cities, hotels, farms, and stations appeared along the line as evidence of Flagler’s enterprise. At the end stood Key West, ready to enter a new age of commercial importance.
There was a problem with this arrangement of the past. Key West had not been waiting without a history for a train to arrive.
It was a port. Its relationships extended across the water, toward Cuba and the Bahamas, into Gulf shipping, and along the Atlantic coast. The sea that appeared to a mainland railway promoter as an obstacle was also the medium through which islanders worked, traded, traveled, and communicated. A place could be distant from the American railway network without being remote from commerce.
Jefferson Beale Browne’s 1912 history of Key West provides a contemporary counterweight to the railway’s account. His book devotes separate chapters to commercial activity, cigar manufacturing, wrecking, population, hurricanes, and the railway itself. Browne was a civic advocate rather than a detached modern historian; his judgments require examination. But the range of subjects in his account establishes something essential. The railway entered an existing, complicated city. It did not create life on an empty island. Browne, Key West: The Old and the New, 1912
The geography encouraged more than one kind of economy. A useful harbor could serve ships. Nearby reefs could endanger them. The same maritime position that offered access to international traffic generated work recovering vessels and cargoes in trouble. Cigar manufacturing linked the island’s commercial life to Cuba. Government and naval interests added another reason for Americans on the mainland to care about a city near the entrance to the Gulf.
In the railway’s language, Key West’s importance was about to become greater because of the Panama Canal. The United States had committed itself to the canal project before construction of the extension began, and the prospective canal changed the way promoters described the island. Key West would become a point where ocean routes met a railway running north through Florida. Cargo need not remain aboard a ship for the whole journey to a more distant American port. Passengers, mail, and goods could transfer between water and rail.
This was a proposition about the value of time. Shortening one part of a journey might make an entire route more attractive. A railway did not have to replace every ship to benefit from maritime trade. It could connect with ships, deliver to them, receive from them, and turn a harbor into the end of a land route.
The opening booklet repeatedly linked the proposed commercial advantage with national defense. It presented Key West as a strategic station from which the United States could protect Gulf approaches and the eastern entrance to the canal. In that formulation, rails would carry the supplies required by ships. The railway was imagined as both a business and an instrument of national power. These were the company’s arguments for the investment; they should not be mistaken for an audited account of the profits eventually earned. FEC, 1912 opening booklet, sections on the Keys and the extension
There was another kind of distance to overcome. A line on a map could connect Key West to the mainland in an instant. A working railway required a continuous physical route through the islands. Between Homestead and the terminal lay shallow water, swampy ground, limestone, tidal channels, and stretches where a foundation had to be created before track could be laid.
For a passenger, the future journey would be a succession of views. For an engineer, it was a succession of different problems.
Nor were all the Keys geologically identical. NOAA describes the Upper Keys as remnants of ancient coral reefs and the Lower Keys as limestone derived from old sand bars. Their low elevations were part of a much older history of changing sea levels. They were not simply pieces of an otherwise continuous mainland waiting to be stitched together. NOAA, Florida Keys habitats
The distinction matters when looking at the art in this collection. A broad blue sea, a low green island, and an elegant bridge can communicate the railway’s setting. They cannot establish the shape of a particular channel, the depth below a specific pier, or the site of a family’s house. Those details belong to maps, engineering records, and identified photographs.
In 1904, the surveyors had to encounter that actual landscape. They could not build the promotional view. They had to find a route across the ground and water that existed.
3. The fortune behind the line
Henry Morrison Flagler was born in New York in 1830. By the time the first train reached Key West, he had passed his eighty-second birthday. The railway across the Keys was a late undertaking in a life already connected to one of the largest concentrations of industrial wealth in the United States.
He had made his fortune through Standard Oil. The National Endowment for the Humanities identifies him as a founding partner of the company and places his Florida work within the larger history of Gilded Age capitalism: railways, extensive land development, luxury hotels, and the construction of Whitehall, his Palm Beach residence. Oil and Florida were not separate biographies. The success of the first supplied resources for the second. NEH, the Henry Morrison Flagler Museum
The extension’s opening booklet preferred a particular version of this story. Flagler appeared as a constructive genius, a man whose imagination became bridges and whose persistence overcame doubt. Business struggle and industrial concentration gave way to a portrait of personal energy. The document was written to celebrate him, not to investigate the sources or consequences of his power.
Its account is still useful when read for what it reveals. Flagler’s Florida system developed through a connection between destinations and transportation. A hotel needed guests. Guests needed a practical means of reaching it. A railway made a destination accessible, while the destination helped generate traffic for the railway. Investment in one encouraged investment in the other.
At St. Augustine, the company’s history placed the beginning of the Ponce de Leon hotel project in 1885. The surrounding transportation system was not yet a single uniform route. Different lines operated with different track gauges, and the St. Johns River presented another interruption. The booklet described purchases, conversions to standard gauge, and a bridge that made through service possible.
Those details explain more than a sequence of acquisitions. A railway network becomes useful through compatibility. A traveler should not have to abandon one vehicle, cross a river separately, and find another train simply because two lines use different arrangements. Freight benefits from the same continuity. The practical power of consolidation is visible in the elimination of interruptions.
The company traced through service to West Palm Beach to April 1894 and regular service to Miami to April 1896. Hotels followed the expanding line or helped justify its extension. The geographical center of the enterprise moved south. Homestead became a further stage, rather than the final destination. FEC, 1912 opening booklet, “The Development of the East Coast”
It is tempting to make this progression look inevitable. Once the line reached Miami, the next stop in a short history is Key West. But the two undertakings differed fundamentally. Moving down the peninsula was not the same as crossing the island chain. The extension required a construction fleet, marine foundations, temporary work routes, imported supplies, and arrangements for workers who could not simply walk into a nearby town at the end of a shift.
It also required a willingness to invest before the railway could produce the traffic promised for it. Money committed to a half-finished bridge was not yet earning the revenue of a completed route. A terminal at the far end could not provide the full benefit imagined for it while gaps remained in the line. The project’s parts depended on one another, and that dependence increased the importance of sustained financing.
In the company’s version of the decision, Flagler asked Joseph R. Parrott whether the railway could be built. Parrott answered that it could. Flagler told him to proceed. The exchange is often retold because it reduces an enormous undertaking to a few words between two men.
It should be attributed to the corporate account. There is no need to supply a room, a gesture, a pause, or a tone of voice that the source does not establish. Even if the exchange accurately preserves the decision, it cannot explain the construction by itself.
A financier could authorize an expenditure. He could not, through authorization alone, place a pier on sound material below a current. Parrott could direct an organization, but direction had to become work performed by people far from the conference in which the project was approved.
The railway would require many forms of expertise. Some would be visible in the finished bridges. Others would leave little trace: a supply delivery that arrived before a crew ran out of water, a repair to a dredge, a calculation checked before concrete was placed, the work of keeping accounts for an enterprise scattered across land and sea.
The famous name belonged to Flagler. The capacity to finish belonged to an organization. Its first task was to turn a southward intention into surveyed lines, construction plans, and a sequence of operations that could actually be carried out.
4. Before the rails
The company’s construction summary dated surveys south of Homestead to January 1904. Construction began in April 1905. Between those dates lay work that passengers would never see once the railway was complete.
Surveying was an act of selection. The islands provided possible ground, but not an automatic railway alignment. A route had to accommodate channels, shorelines, curves, grades, access for materials, and the transition from one type of structure to another. Choosing where to cross water influenced the kind of bridge required. Choosing an embankment rather than an open span changed the quantity and type of work.
These are basic engineering relationships, not a reconstruction of an individual surveyor’s day. They explain why the preliminary phase mattered. A finished railway looks like one continuous object. Its design is a collection of decisions whose effects spread forward through construction and operation.
Joseph Carroll Meredith became chief constructing engineer on July 22, 1904, according to the opening booklet. He had been working on docks at Tampico, Mexico. That background was appropriate to a project in which building a railway required much of the experience associated with harbors and marine works.
The booklet credits Meredith with extensive planning well ahead of construction. It also records that his plans were not a substitute for judgment in the field. Unexpected conditions continued to arise. The organization needed people who could respond when the actual work differed from the conditions anticipated on paper. FEC, 1912 opening booklet, Meredith and construction sections
The distance from Homestead to the Key West terminal was given as 128.4 miles. That number describes a defined segment, not every mile of the Florida East Coast Railway and not the distance of a modern highway journey from Miami. The distinction is small enough to disappear in a caption and large enough to confuse an entire map.
The construction summary separated the permanent bridging into concrete arch viaducts and steel bridges on concrete piers. It also described long stretches of embankment. The railway did not leap from island to island on a uniform procession of arches. In shallow areas, dredges and excavators built roadbed. At deeper or otherwise demanding crossings, different bridge forms were used.
This variety is one reason a single generalized painting cannot stand in for every location. Long Key’s arches and Bahia Honda’s through trusses are not interchangeable. A train in one setting would be supported by a different structural system from a train in the other. A documentary that reused the same attractive bridge image for both would lose a central part of the story: the builders adapted the railway to different conditions.
For the construction crews, temporary works were as necessary as permanent ones. Materials had to reach the point where permanent construction was under way. A temporary track or trestle might serve that purpose and then disappear from the operating landscape. A period photograph can therefore show rails beside or below a structure without depicting the final passenger route.
The company’s booklet includes precisely this kind of distinction in its illustrations. A photograph associated with Knights Key shows a temporary connection to the dock in relation to the main line. Another caption identifies temporary track used during concrete work. Reading those captions changes what the pictures mean.
Without them, the viewer might assume that every visible rail belonged to the final alignment. With them, the photograph becomes evidence of a railway building another railway: a provisional network supporting the creation of a permanent one.
The physical setting imposed a similar double meaning on the sea. It was the obstacle the extension sought to cross, but it was also the route by which much of the extension could be built. Barges could approach work sites that had no rail connection yet. Dredges could create embankment using material taken from nearby water. Boats could move people and supplies between separated sections of work.
The builders were not simply advancing one railhead south in a neat line. The project depended on coordinated activity across multiple locations. Some crews could be working on foundations while others formed embankments or prepared a terminal. Each part had its own immediate difficulty. Together, they had to become a continuous route.
In the corporate booklet, the complexity is submerged beneath the language of triumph. In the construction photographs and data, it returns. A barge, a temporary trestle, an unfinished arch, a pile driver: these are the objects that make the achievement comprehensible. They show the railway before it could conceal the enormous effort required to produce its apparent simplicity.
5. A railway built by boats
One of the most revealing pictures in the opening booklet is not a bridge. It is a water barge.
The railway’s promoters wanted readers to admire the great structures. The barge points to a quieter requirement. People working among the Keys needed fresh water. So did the larger construction system in which they worked. Before the completed railway could carry ordinary traffic, a supply network had to support the men and machines creating it.
The booklet identified fresh water reaching the line near Everglade station through Manatee Creek as a principal source of supply. It also pictured a floating camp, or quarterboat, and a land camp. Together these images describe the practical arrangements behind the spectacle of building at sea. The workforce had to be housed, fed, moved, and supplied in places where the finished railway did not yet exist. FEC, 1912 opening booklet, water supply and construction photographs
The phrase “railway construction” can suggest a work train carrying rails, ties, tools, and men to the end of an existing track. That was only part of this undertaking. The company described a fleet ranging from small powered launches to large dredges. Different vessels performed different tasks. Some transported. Some excavated. Some provided living space. Their usefulness depended on the same water that made the railway difficult to build.
This is the first of several reversals in the extension’s history. Boats built the route intended to reduce dependence on boats. Decades later, when the route failed in the hurricane, boats and aircraft would again become essential to reaching the islands.
The construction fleet had needs of its own. A vessel transporting material was an asset that required care, protection, and coordination. The booklet acknowledged that managing the fleet was itself a substantial undertaking. A project scattered across channels could not be reduced to the amount of steel or concrete used. Its success depended on getting the right material to the right place while that place was still accessible and ready to receive it.
Imagine the difference between counting a load at a mainland depot and using it at an unfinished bridge. The count establishes possession. It does not establish delivery. Delivery does not establish that the receiving crew has everything else necessary to use the load. Construction progress depends on the whole sequence. This is an explanation of the logistics, not an invented account of a particular shipment.
At Marathon, the company showed girders being transferred from railway cars to barges. The caption described the settlement as the center of work to the south. The picture makes the connection between land transport and marine construction explicit. A girder could travel part of its journey by rail and the next part by water before becoming a fixed element in the route that would carry later trains.
Materials did not all arrive as finished components. The roadbed itself was created in part from local limestone and dredged material. The company reported that dredges and traveling excavators built forty-nine miles of line through shallow water openings and low, swampy Keys. Where necessary, limestone was blasted underwater so that it could be used in the embankment.
The construction summary also described marine marl used to protect exposed slopes. Dredges removed it from deposits near the line. Steel dump cars carried it to places where it was needed. When deposited, the material spread into a flat slope; the booklet asserted that exposure to air hardened it into a protective covering.
Those statements tell us both how the company said it built and what it believed the resulting protection could accomplish. The process description and the promise of performance belong to different categories of evidence. The first can be compared with photographs, engineering records, and surviving works. The second must also be tested against what happened when later storms arrived.
The water barge belongs in that same careful reading. It does not prove that supplies were always adequate. A picture of a quarterboat does not establish that the people aboard it were safe. An official account can document that arrangements existed while presenting their effectiveness more favorably than workers would have described it.
Still, these unglamorous objects are indispensable to the history. Remove the barges, the dredges, the temporary tracks, and the camps, and the railway becomes an impossible act of will. Put them back, and it becomes a human undertaking: costly, organized, vulnerable, dependent on repeated tasks, and carried forward by people whose names rarely appear in the large lettering of the celebration.
6. The men beneath the achievement
The railway company could name its executives and engineers. Its laborers usually appeared in groups.
The opening booklet identified workers from Sweden and Norway, Spaniards who arrived by way of Cuba, African Americans, Bahamians and other Caribbean workers, and divers from Greece. The language it used to describe them reflected the racial hierarchy of its time. The information beneath that language is nevertheless significant. The extension was built by an international workforce, including people whose routes to the Keys crossed the same seas the railway was supposed to overcome.
Some brought specialized skills. Divers worked in a part of the project that the passenger above the water could never see. Foremen coordinated labor. Other men handled the repeated physical tasks by which material became roadbed and structures. The distinction between skilled and unskilled labor, so often made in administrative records, can hide how much practical judgment was required around heavy loads, moving equipment, water, and unstable work surfaces.
The company insisted that it took exceptional care of its people. It described good food, appropriate housing, arrangements for illness and injury, and a railway hospital in Miami prepared for the construction service. It compared sickness favorably with the United States Army and characterized loss of life as small for a project of its difficulty.
Those are employer claims. They deserve attribution, especially because the same publication was designed to convince readers that the undertaking had been managed with extraordinary competence. A celebration of management is not an independent inspection of labor conditions. FEC, 1912 opening booklet, “The Actual Work”
Browne, writing in the same year and praising Flagler with equal enthusiasm, left a harsher account of the physical cost. He described drowning, crushing injuries, dynamite accidents, exposure, and the violence of hurricanes. His language was highly rhetorical, but his celebration did not make the workforce’s suffering disappear. Instead, he treated suffering as evidence of the magnitude of the achievement.
That was another way workers could be absorbed into the story of a great man. In one version, they had been exceptionally well cared for. In the other, they had endured extraordinary hardship so that the enterprise could succeed. Neither version necessarily gave them the opportunity to explain the undertaking in their own words. Browne, Key West, railway chapter, pp. 194–196
The absence of individual voices must not be repaired with fictional ones. There is no responsible reason to invent a laborer’s letter home, place unrecorded words in a foreman’s mouth, or assign a particular emotion to a face in a construction photograph. The surviving record already offers a way to restore the workers to the center: describe the work accurately, identify the limits of the company’s account, and refuse to make human loss a decorative measure of ambition.
A quarterboat illustrates the stakes. As an arrangement for housing, it brought men close to work that might otherwise be difficult to reach. Proximity reduced daily transportation problems. But floating quarters were still vessels in a hurricane region. Their usefulness in ordinary conditions and their vulnerability in extreme conditions were parts of the same design choice.
The long, low geography also limited the meaning of shelter. A structure could protect against rain without being secure against a great storm. Land could stand above an ordinary tide without remaining above storm-driven water. An island that provided room for a camp might offer little opportunity to move to higher ground.
Such distinctions became terribly important during construction storms. They would become important again in 1935, when a different employer housed a different workforce along the same chain.
The two groups should not be merged. Flagler’s construction workers built the railway in the years before 1912. The veterans present in the relief camps in September 1935 were working on a highway. Some popular accounts describe the veterans as if they had been caught while completing Flagler’s line. That error erases more than two decades and obscures the institutions responsible for their circumstances.
The connection between the two workforces is geographical and historical, not organizational identity. Both undertakings placed labor in an exposed marine landscape. Both depended on transportation and shelter arrangements. Both would later be remembered through a mixture of engineering pride and disaster.
But the men who worked on the original extension also deserve a history that reaches beyond catastrophe. They made a continuous railway possible. Their labor passed into concrete, steel, dredged fill, and a route that other people would use without having to know how difficult each part had been.
7. The storms before Labor Day
The hurricane of 1935 was not the railway’s first encounter with violent weather.
Construction extended across years in which storms struck the Keys. The October 1906 hurricane killed railway workers and damaged the undertaking. Further hurricanes came in 1909 and 1910. The later National Register documentation treats these storms as part of the building history, while the company’s own opening account acknowledges repeated attacks by tropical weather. National Register, Overseas Highway and Railway Bridges amendment
It would be easy to give the 1906 disaster a single large number and move on. The surviving summaries do not make that responsible. Accounts differ over the number of workers exposed, the occupants and losses associated with individual floating quarters, and the scope of the total being counted. A figure for one vessel cannot be silently converted into a figure for the entire railway project.
The essential event is not in doubt: a storm interrupted construction and cost lives. What requires care is the accounting. A missing person, a recovered body, a man reported aboard a particular vessel, and an employee recorded elsewhere in the project are not automatically the same category. Disaster records are often assembled under conditions that make mistakes and later revision unavoidable.
For the company, continuing after a storm demonstrated resolve. Damage could be repaired. Work could resume. Each recovery became part of the story of a railway that would ultimately overcome the sea.
For a worker, the event had another scale. A wrecked camp was a place where people had slept. A lost construction vessel was also a setting in which particular men had been working. The company’s ability to rebuild did not reverse those losses.
Browne’s account of Key West’s October 1909 hurricane shows how deeply a storm could reach into the city’s ordinary systems. He described damaged factories, churches, docks, power-related facilities, and vessels. A coal hoist fell. Boats broke from moorings. A schooner carrying sand and gravel for bridge work was driven away and sank; its crew, he reported, survived after clinging to the rigging and being brought in by another vessel.
The episode joins the railway project to the surrounding port economy. Material intended for construction could become storm debris before it reached the work. The harbor was not simply a supply point outside the railway’s risk. It was part of the same exposed system. Browne, Key West, hurricane chapter, pp. 156–161
His account of 1910 is revealing for a different reason. It describes instruments failing while conditions worsened. Wires connected to the anemometer were torn down. The rain gauge was carried away. Later figures therefore included estimates rather than a complete series of uninterrupted measurements.
There is a tendency to treat old weather numbers as if they were produced by a modern automated station that operated successfully throughout the event. The physical history of the instrument matters. If the device failed before the worst conditions, its maximum reading may describe the strongest wind it recorded, not the strongest wind that occurred.
Browne also described the difficulty of interpreting the storm’s movement from available reports. Threatening weather persisted even when an account of the cyclone’s position appeared reassuring. Experienced residents continued preparations. His narrative emphasizes local judgment, but it should not be turned into a simple contest in which islanders always understood storms and official observers never did.
The experience of 1935 would be more complicated. Some residents recognized danger, secured boats and houses, and believed their preparations sufficient. Some correctly judged the veterans’ temporary camps more vulnerable than their own homes. A person could take a hurricane seriously and still underestimate the water it would bring.
The earlier storms nevertheless remove one possible misunderstanding. Hurricanes were not an unknown category of danger suddenly introduced to the Keys on Labor Day. Engineers, railway managers, island residents, and weather observers had histories of dealing with them.
What those histories could not guarantee was that the next storm would resemble the last one in the ways that mattered most. A structure’s survival in an earlier hurricane was evidence of survival under those earlier conditions. It was not proof of invulnerability to every future combination of wind, water, waves, and debris.
When the opening booklet later described the railway as effectively storm-proof, it transformed a record of persistence into a promise about the future. The bridge photographs made the promise persuasive. The concrete looked permanent. In the years ahead, many of those substantial structures would indeed survive. The route as a working system required more than their survival alone.
8. Arches, girders, and the ground below
At Long Key, the railway acquired one of its defining forms: a long procession of concrete arches across water.
An arch is visually satisfying because its load seems to resolve into the supports at either end. Repeated over a long distance, it gives the impression of a rhythm imposed on the sea. The company’s photographers and writers recognized that power. The Long Key Viaduct became a principal image of the undertaking, a structure that could make the route’s audacity understandable at a glance.
But the visible arch was only part of the engineering. Weight had to pass from the track and structure through the supports to foundations capable of carrying it. The water below was not an empty decorative space. Currents, depth, bottom conditions, construction access, and the forces expected during storms all mattered to what could be built.
Concrete and steel also performed differently within the overall route. The opening booklet’s construction summary anticipated 17.2 miles of permanent bridge work: 11.1 miles of concrete arch viaducts and 6.1 miles of steel bridging on concrete piers. These are figures from the company’s 1912 statement, whose wording included work still to be completed. They should be labeled that way rather than presented as a modern survey of every surviving structure.
The longest crossing, between Knights Key and Little Duck Key, exceeded seven miles including approaches. At Bahia Honda, the company described through steel trusses, with spans reaching 243 feet. It reported that some pier foundations there extended thirty feet below tide level. FEC, 1912 opening booklet, “Construction Data”
The differences were practical responses to different crossings. A through truss places its principal structural framework alongside and above the level through which the train passes. That produces a very different view from a train riding above low arches. Later highway conversion changed some bridge profiles again. A photograph of a road carried atop a former railway truss does not show the arrangement through which a train traveled in 1912.
The National Register inventories are valuable precisely because they separate these layers. They identify original structures, later additions, bridge types, and the effects of conversion. They allow a surviving bridge to be read as a structure with a history, rather than as a perfectly preserved instant from opening day. Original National Register nomination
The construction process had its own temporary architecture. Forms gave wet concrete the shape it needed while it gained strength. Work platforms and temporary tracks made assembly possible. Floating equipment brought material within reach. A completed arch concealed the formwork that had supported its creation, just as the finished route concealed the temporary network that had served it.
The booklet includes a photograph explicitly showing arches with forms still in place. It is a useful corrective to the illusion that monumental structures arise as complete objects. Concrete had to be placed. Components had to be aligned. Work proceeded through stages in which the eventual structure was incomplete and could not yet perform its intended task.
There was an economic judgment beneath the choice of each construction method. An open bridge and a filled crossing require different quantities of material, different kinds of labor, and different maintenance. They also leave different paths for water. The company’s use of embankments in shallow areas was not merely a way to avoid drawing more bridges on the map. It shaped the physical relationship between the railway and the surrounding sea.
After the 1935 hurricane, a survivor would argue that some fills had made flooding worse. His testimony deserves attention, but it does not by itself measure the hydraulic effect of an embankment. To establish that effect precisely would require the pre-storm geometry, water levels, openings, timing, and flow conditions. An eyewitness could observe water rising and an embankment breaking without being able to isolate every cause of the inundation.
The larger point remains visible without overstating the evidence. The railway did not float above the Keys without changing them. It added solid material, narrowed or crossed waterways, created new land at its terminal, and placed a continuous transportation system in a landscape shaped by moving water.
The great bridges are therefore both engineering achievements and entries in an environmental history. They show what the builders could accomplish with the materials and knowledge available. The gaps between them, the fills approaching them, and the ground supporting them explain how that achievement functioned—and where its vulnerabilities could lie.
9. A terminus before the terminus
By February 1908, trains could reach Knights Key Dock.
The final destination remained Key West. Knights Key was an intermediate solution, a place where the growing railway met vessels serving Key West and Havana while construction continued farther south. The opening booklet identified February 5 as the beginning of that terminal arrangement. Its photographs show the dock and the relationship between its temporary rail connection and the eventual main line. FEC, 1912 opening booklet, Knights Key Dock and route description
This stage is easily lost in a story told as a seven-year wait for a single triumphant arrival. Parts of the extension could become useful before the whole route was complete. Rail and steamship services could meet at a temporary endpoint. The railway’s practical geography moved as construction advanced.
For management, an intermediate terminus offered a way to connect the unfinished project to existing traffic. For passengers, it meant a transfer rather than an uninterrupted rail journey to Key West. For construction crews south of the dock, the work remained unfinished regardless of how impressive the line already appeared to visitors.
Marathon, on Key Vaca, served as a major center for work in that southern portion. The company’s photograph of girders moving from cars to barges belongs to this stage of the story. It shows the railway supplying the marine operations that would extend it.
The remaining crossings could not be treated as a minor final stretch. They included some of the undertaking’s most demanding and memorable structures. A nearly completed route could still be commercially incomplete if its remaining gaps prevented the journey it had been designed to provide.
In April 1909, the project lost Meredith.
The company recorded his sudden death on April 20. He had directed construction from its beginning south of Homestead. His first assistant, William J. Krome, succeeded him. The opening booklet described plans already laid and unforeseen problems still to be solved, a combination that captures the nature of a large undertaking at the moment of a leadership change.
Meredith’s work could continue because it had become more than a set of instructions held in one person’s mind. Plans, divisions of responsibility, experienced colleagues, and established procedures gave the project continuity. Yet the company’s account also made clear that Krome had to exercise judgment. Completing a plan was not the same as mechanically following it under conditions that never changed.
The booklet named the division engineers P. L. Wilson, C. S. Coe, G. R. Smiley, and Ernest Cotton. It identified R. W. Carter as bridge engineer, E. H. Sheeran as general foreman, and B. A. Deal as auditor of construction. These names widen the familiar portrait of Flagler and his railway. They show a differentiated organization in which engineering, supervision, and accounting all had to work together.
The company erected a memorial to Meredith in Miami. Its inscription associated his memory with skill, fidelity, and devotion to the extension. The memorial also records a limit on the familiar opening-day narrative. One of the principal men responsible for creating the route did not live to see its ceremonial completion.
His absence is more historically grounded than the sentimental dialogue sometimes supplied for Flagler’s arrival. It needs no embellishment. The first train to Key West would cross works designed and begun under an engineer who was no longer there.
The project continued through the storms of 1909 and 1910. Meanwhile, the terminal at Key West required its own transformation of the waterfront. The company reported creating 134 acres by hydraulic filling along the island’s northern side. Its 1912 construction summary described a concrete pier being built into Man-of-War Harbor, 1,700 feet long and 134 feet wide, with slips dredged through rock.
Those figures reveal that the final station was not simply a platform placed at the end of the track. The commercial future promoted for Key West required facilities through which passengers and cargo could move between railway and harbor. The extension’s endpoint was itself a substantial construction project.
The route thus developed from both practical necessity and the larger ambition that had justified it. Temporary arrangements carried traffic while permanent ones took shape. Boats met trains at Knights Key. Barges carried bridge material south. At Key West, dredges created land for a railway that had not yet arrived.
10. January 22, 1912
The arrival has a date that can be anchored to more than a later anniversary story: January 22, 1912.
The State Archives of Florida preserves photographs cataloged to the event. One records Flagler disembarking. Another records the crowds at the arrival. A third identifies a parade of Key West police on Duval Street. They place the celebration in a city, among people, rather than in the abstraction of a completed line on a railway map. Florida Memory: Flagler arriving; arrival crowds; Duval Street parade
Flagler was eighty-two. The distance between the beginning of his working life and this arrival had been measured in more than years. He had participated in the growth of an industrial fortune and spent decades extending a transportation and hotel system down Florida’s east coast. Now a train could carry him to a city that had previously remained beyond the railway’s reach.
That achievement did not require the company’s more extravagant claims to be extraordinary. A continuous route had been constructed across the Keys. The work had survived organizational changes, storms, difficult logistics, and the death of its first chief constructing engineer. The train arriving at Key West was evidence that the essential route worked.
The public occasion also served a commercial purpose. A railway wanted attention, traffic, and confidence. The opening publication turned the construction into a national story of ingenuity. It invited readers to regard the route as a marvel they should experience and as an investment in the country’s future.
The photographs and the booklet therefore perform different tasks. The photographs establish particular visible moments. The booklet explains how the company wanted the event understood. Neither supplies unlimited access to the private feelings of everyone present.
An often-repeated deathbed-like remark attributed to Flagler at the opening is not needed here. Without a securely established contemporary source for the exact words, it would add apparent intimacy at the expense of evidence. His age, presence, and achievement are sufficient. So is the knowledge that Meredith was absent.
For Key West, the event opened a new connection without closing the old ones. The city remained an island port. A train could bring passengers who would continue by ship. Goods could travel between railway and waterfront. The route changed the sequence of movement rather than abolishing the sea.
The date also belongs in the wider chronology of the Panama Canal. The railway opened before the canal began operating in 1914. The company’s arguments about canal traffic were anticipatory. Its promises were being made about a commercial future still taking shape. It would be misleading to write that the extension had been built to serve an already established stream of canal traffic.
Some construction language in the opening booklet was itself future-facing. The data section described what the system would contain when completed and work being done at the terminal. A ceremonial opening can mark the beginning of service without meaning that every element of a large transportation enterprise has reached its final form.
This is not a reason to diminish the occasion. It is a way to understand it. Railways are working systems, subject to additions, replacement, maintenance, and changing demand. The first train begins a new phase. It does not freeze the infrastructure into the exact condition in which it will remain forever.
The opening also changed the meaning of the construction landscape. Temporary routes and work camps could recede from view. What had been a sequence of difficult jobs became, for a passenger, a journey. A bridge could be experienced in minutes after requiring years of planning, supply, and work.
The water beyond a carriage window supplied the spectacle. The organization beneath the journey supplied its reliability. A passenger did not need to understand concrete forms, dredged marl, or a pier’s foundation to cross the channel. That freedom from having to know was part of what the builders had created.
In the years that followed, the extension would have to do something less dramatic than opening and more demanding in its repetition. It would have to operate. Trains would need a continuous track, bridges would need maintenance, cargo would need handling, and connections would need to be made. The wonder would have to become a service people could use.
11. The city at the end of the track
In Browne’s account of Key West, two men worked a small boat over the sponge grounds. One sculled. The other used a long pole with a three-pronged hook to bring up the catch. A larger sponging vessel carried a small boat for each pair of men.
The description is a useful companion to the railway’s images of giant structures. It records another way of working with the water, one dependent on knowledge, coordination, and a market far beyond the place where the catch was taken. Browne also described competition from diving operations based at Tarpon Springs, which he associated with a shift in the industry away from Key West.
The island’s future was therefore not determined by transportation alone. Methods of production, business competition, access to resources, and the location of markets could change the fortunes of an industry. A new railway entered those relationships. It did not control them all. Browne, Key West, material development chapter, pp. 103–111
The same chapter describes a factory producing canned turtle soup and another manufacturing cigar boxes. The turtle business, in Browne’s telling, was constrained by the supply of animals suitable for its product. That detail belongs to an extractive economy that treated marine wildlife as commercial material. It should not be softened into a picturesque culinary anecdote or translated into a recommendation for present-day consumption.
The cigar-box factory offers a different view of the connection between production and transport. Browne described an earlier operation that closed after difficulties with freight rates and transportation. He then described a new reinforced-concrete plant at Caroline and William streets, which began work in September 1911. Its stated capacity was seven thousand boxes a day—a contemporary claim about potential output, not proof that it produced that many every working day.
Boxes were a supporting industry. Their usefulness depended on cigar manufacturing, and cigar manufacturing depended on materials, labor, markets, and transport. The railway did not arrive in a city of isolated occupations. It arrived in a web of businesses serving one another.
Browne’s account of express service makes the value of small logistical improvements especially clear. He described an operator who consolidated small packages in New York into a larger case for shipment to Key West. By spreading a minimum freight charge across multiple customers, the operator could offer a service at a lower per-package cost. That was an island solution to a transportation problem long before the through train.
The city also possessed ways to repair ships. Browne contrasted earlier methods of heaving a vessel over beside a wharf with the development of marine railways, which hauled vessels out for work. These railways were not Flagler’s passenger line. They were inclined hauling facilities, another reminder that the vocabulary of transportation can hide important differences if used carelessly.
At the waterfront, government performed its own work. Browne reproduced a description of the customs building, completed and occupied in 1891. Postal and customs services occupied the first floor; court facilities occupied the second; lighthouse officials and other government offices were above. The arrangement placed communication, commerce, law, and maritime administration within the same building. Browne, Key West, commercial chapter, pp. 99–102
At Key West, the first train joined a city already organized around such movements. People brought experience from Cuba, the Bahamas, the mainland, and the sea. Businesses had succeeded and failed before the extension. Storms had damaged buildings and vessels. Residents had rebuilt.
The railway made a new promise to this existing city: a more continuous connection to the mainland. The promise was powerful partly because islanders already understood the value of dependable transport. They had been managing its absence, interruptions, costs, and opportunities for generations.
12. Across the window, beyond the harbor
The company’s route description carried the reader south through a sequence of recognizable places: Key Largo, Tavernier, Plantation Key, the Matecumbes, Long Key, Grassy Key, Key Vaca, and the crossings farther down the chain.
It paused at the fishing camp on Long Key. In 1912, that camp appeared in the promotional geography as a winter destination for people who wanted water, climate, and fishing. Its presence is important because it gives the later hurricane account a life before the storm. Duane’s camp was not merely the setting of a dramatic survival story. Long Key had been advertised as a place to visit, to stay, and to enjoy the sea. FEC, 1912 opening booklet, southbound route description
The railway’s passenger experience depended on an unusual division of attention. Outside lay the water. Below and ahead lay a carefully maintained route that the passenger might scarcely consider. The train made it possible to look at the sea without having to navigate it.
A journey across an open bridge could make that separation especially striking. The water appeared on both sides while the carriage continued along a fixed line. On land, the view changed again: vegetation, settlement, stations, and the next crossing. The route alternated between enclosure and exposure.
Those are features of the documented geography, not the diary of a particular passenger. There is no basis here for inventing the smell of a specific carriage, the meal served on an unverified train, or a conversation overheard at a station. The route itself supplies enough to understand why travelers found it memorable.
At its southern end, the line connected with shipping. The Florida East Coast Railway’s company history records three freight-car ferries carrying boxcars and refrigerator cars between Key West and Havana until the 1935 hurricane. Railway cars could be transferred aboard vessels rather than all their contents being unloaded into a separate cargo-handling system. FEC, freight-car ferry history
The arrangement fulfilled the basic idea of a junction between rail and sea, though it does not establish that every commercial expectation expressed in 1912 was realized. A service’s existence, its traffic volume, and its profitability are separate questions. Exact answers about the latter two require operating and financial records, not simply a history of the ferry fleet.
The distinction also prevents the hurricane from becoming an overly convenient explanation for every difficulty the railway faced. The storm caused physical destruction. It did not invent the economic conditions of the Depression or erase the need to consider costs and expected traffic in deciding what to rebuild.
An operating railway is more than its celebrated spans. Track must remain aligned. Approaches must remain intact. The route must connect continuously through every essential point. A large bridge can survive while the service using it becomes impossible because its track or an adjoining embankment has been removed.
For a comparison of maritime abandonment and incomplete records, read our Mary Celeste investigation.
That fact explains the apparent contradiction in the post-1935 landscape. Substantial remnants could remain standing even though trains no longer ran to Key West. The continued existence of concrete or steel did not mean the line could simply resume service the next morning.
The passage of years also changed the public’s relationship with the route. A marvel becomes familiar through use. Familiarity can be beneficial: it reflects a system repeatedly doing what it was built to do. It can also make dependencies less visible. A train arriving as expected does not call attention to the maintenance and communications that made the arrival possible.
By 1935, the railway had a role the opening booklet had not centered: it could serve as an evacuation route for a vulnerable population working on another project. The veterans in the highway camps did not need a sightseeing journey. They needed the line’s practical ability to carry many people away from low ground before weather made movement impossible.
The problem was not merely whether a locomotive and cars existed. A usable evacuation depended on ordering the train, assembling it, staffing it, moving it along an exposed route, reaching people who needed it, boarding them, and returning through conditions that were worsening.
Each part took time. The same continuity that made the railway useful in ordinary service became a demanding requirement under a closing deadline. A train could not choose an alternate street around a broken section. Its path to the camps was also the path it needed to bring people back.
13. A different road, a different workforce
The men in the veterans’ camps were not finishing Flagler’s railroad.
They were working on the Overseas Highway under a relief program during the Depression. The Federal Emergency Relief Administration employed veterans in camps along the Keys. The distinction is documented in NOAA’s retrospective and in the records of the disaster’s investigation. NOAA, 1935 hurricane retrospective
The confusion is understandable at the level of a photograph. A railway, a highway project, and the same narrow island chain occupy one broad setting. A rescue train appears in the story. A viewer may assume that the men needing rescue were railway workers.
But the institutions and dates are central. The railway had opened in 1912. The camps destroyed in September 1935 belonged to a later public work program. Flagler had died long before the Labor Day hurricane. Responsibility for the camps cannot be assigned through the name attached to the old railway.
The highway represented another attempt to make the Keys more continuously accessible. Where a railroad carries people in trains managed as a service, a road permits a different form of movement. Its usefulness depends on completed bridges and road sections, but the traveler’s vehicle can be individually owned. The eventual highway would inherit parts of the railway’s infrastructure, yet in 1935 that inheritance had not taken its final form.
John A. Russell lived between the camp locations. In his sworn statement after the hurricane, he identified himself as Islamorada’s postmaster and a lime grower. He was forty-seven and said he had lived in Florida all his life. He described camps above and below his community, on Windley Key and Lower Matecumbe.
He also described changes in the officials responsible for the project. Personnel turned over; by the time he became familiar with one, another could be in charge. That is Russell’s account, not a complete administrative roster. It is nevertheless a revealing description of how a local resident experienced an outside institution operating nearby. Russell sworn statement, September 14, 1935, questions 1–6
His most serious concern was physical. He believed the temporary camp buildings offered much less protection than the houses in which established residents intended to remain. He described Camp 1 as low and exposed to the sea. He said its huts stood on piers without adequate anchoring and that he had earlier warned that they should be tied down.
The statement was taken after the catastrophe, when questions of prudence and responsibility had become urgent. That timing does not make it worthless. It means the reader should distinguish a witness’s recollection of a prior warning from an independently documented inspection made before the storm.
Russell’s reasoning was specific. He did not say that every person on every island faced identical conditions. He believed the men in temporary quarters were differently situated from residents in more substantial houses. That distinction shaped his opinion that the camps should have been evacuated earlier.
The veterans themselves also should not be reduced to a single anonymous type. They were former servicemen, workers, and individuals with separate histories. Their presence in a relief program describes an employment circumstance. It does not establish a common motive, political identity, state of mind, or personal character.
Some period descriptions were openly judgmental about them. Russell, while advocating their evacuation, also expressed harsh opinions about drinking in the camps. Those opinions belong to the witness and the social world in which he spoke. They are not a census of behavior, and they do not diminish the obligation to account for the men’s safety.
The holiday affected who remained in the area. Russell recalled that some men had gone to town over the pay period and holiday interval. His recollection cannot provide an exact final camp population. It does show why a payroll list and a count of people physically present during the hurricane might differ.
That distinction would matter when investigators tried to identify the dead and missing. A name on an employment record did not necessarily mean a man had been in the camp when the water came. Absence from a camp roster did not necessarily mean a civilian or former serviceman had escaped exposure elsewhere.
Before those accounting problems came the immediate question: when should the men be moved? On Sunday, September 1, the storm was already serious enough for Coast Guard personnel to carry warnings to people who might not receive them any other way.
14. Messages falling from the sky
At the Miami Coast Guard Air Station, the warning became an object someone could hold.
On the morning of September 1, personnel received a forecast describing strengthening winds, heavy squalls, and possible hurricane conditions in the Florida Straits. According to the station’s subsequent operational narrative, they considered the information serious enough to warrant immediate action. They prepared message blocks containing the warning and a request that recipients pass it to nearby vessels.
Lieutenant W. L. Clemmer took off to reach boats south of Miami, with particular attention to people beyond ordinary radio, telegraph, or telephone communication. When he returned around midday, he reported that the supply of message blocks had been insufficient. More were prepared for an afternoon flight.
The problem was simple and consequential. A forecast available at an office could not help someone on a boat or at an island picnic unless it reached that person. The Coast Guard was trying to carry information across the last gap in the warning system. Miami Air Station operational narrative, September 1935
When the regulation message blocks ran out, the station improvised. It used paraffin-coated containers with lengths of tape attached to make them easier to see. A hurricane warning became paper inside a small protected container, dropped from an aircraft toward people below.
The detail is vivid because it is operational, not decorative. It reveals the limits of communication in 1935 and the practical ingenuity used to work around them. There was no assumption that every person near the coast possessed a receiver or would hear a broadcast at the right moment.
The report says Clemmer’s flights paid attention to Labor Day picnic parties. It later recorded that members of a party on Indian Key came to the station to thank personnel for the warning. In the station’s account, most had left after receiving it, while two men who remained were lost.
That is an institutional report of the outcome, not an independently reconstructed biography of each person in the party. Its significance is narrower and strong: the station documented a warning delivery that recipients later associated with their decision to leave.
The episode complicates any statement that the hurricane arrived wholly without warning. Warnings existed. They traveled through several channels. Some people acted on them. The harder questions concern what each warning said, when particular decision-makers received it, how they understood the geographical threat, and whether they acted early enough for their circumstances.
A small party able to depart in its own boat faced a different logistical problem from several camps requiring organized transport. A warning adequate to prompt one household to secure property might not provide enough time to assemble a train and remove hundreds of people from low islands.
The Coast Guard’s own situation also changed as the storm approached. After distributing warnings, the station had to secure its facilities and aircraft. Personnel who might later conduct rescue flights first had to protect the equipment that would make those flights possible.
Its narrative records strong winds, rising water, a failure of city power, and later loss of telephone communications. The station used its own generating equipment. The rescue organization was not observing the hurricane from outside its effects. It was operating within a broader region of dangerous weather and damaged infrastructure.
This is another reason to avoid a simple comparison between the time someone first needed help and the time an aircraft arrived. Readiness, weather, access, and the survival of facilities all affected what could be done. The fact that rescue could not immediately reach a place does not settle whether an earlier evacuation should have occurred.
Russell would make that distinction explicitly. He criticized the timing of evacuation. He did not criticize the later efforts to bring aid through conditions he considered extraordinarily hazardous.
For the men in the camps, however, such distinctions could not stop the approaching storm. A warning had to become a movement. A movement required transport. Transport required time and a route still usable in both directions.
The messages falling from Clemmer’s aircraft on Sunday were part of the available warning history. By Monday, the practical interval in which larger movements could be completed was closing. In the weather offices, the task was to establish where the storm was going. Along the Keys, people had to decide what to do before that knowledge became certain.
15. The storm the instruments could not fully see
The modern image of a hurricane is usually an image from above: a spiral, an eye, bands of cloud turning over the ocean. In 1935, forecasters did not possess that view.
They assembled an understanding of the storm from reports. Wind directions, pressures, observations from islands and vessels, and communications between offices supplied pieces of a moving system. The pieces were unevenly distributed. A ship that avoided the storm could protect the people aboard while also ceasing to provide observations near its center.
NOAA’s retrospective describes precisely that problem in the forecasting of the Labor Day hurricane. The Weather Bureau had reorganized hurricane warning responsibilities in 1935. Jacksonville was among the centers involved, with Grady Norton and Gordon Dunn. Reports from the Bahamas helped identify the developing circulation, but the storm’s position and future movement remained difficult to establish. NOAA, 1935 warning and reconnaissance retrospective
The contemporary Weather Bureau account by W. F. McDonald traced a definite depression near Long Island in the southeastern Bahamas on August 31. It placed hurricane development near southern Andros Island on September 1. By the time the storm crossed the Keys, it had become extraordinarily intense while retaining a compact center.
Compactness mattered because it made the most destructive conditions geographically concentrated. The weather at a more distant observation point could not be assumed to represent the storm’s core. Two people at different places along the island chain could experience different winds, water movements, and intervals of relative calm.
McDonald’s report acknowledged difficulty reconciling some local observations with the storm’s track. It described efforts to obtain and test instruments associated with exceptionally low pressure readings. That admission is valuable. A scientific record can be both the best available account and a record in the process of correction. McDonald, Monthly Weather Review, September 1935
Later estimates should therefore not be silently inserted into an eyewitness’s observations. The widely reported 892-millibar minimum and an estimated maximum sustained wind of about 185 miles per hour belong to the hurricane’s reconstructed history. Duane’s individual barometer readings describe what he reported at his location with his instrument. They are not interchangeable measurements.
The distinction between sustained wind and gusts also matters. McDonald’s contemporary discussion included estimates of very high winds and still higher gusts. Modern classifications use specified definitions. A spectacular number from a period narrative should not be placed beside a modern sustained-wind estimate without explaining the difference.
The most consequential danger on the Keys was not captured by wind alone. McDonald emphasized the depth and moving force of the storm tide. Water could rise over land, carry debris, undermine embankments, and move structures. Waves riding on that elevated water added further force. An intact roof was not enough if the water below could lift or break the building.
A reconnaissance flight became part of the effort to locate the storm. NOAA describes Captain Leonard Povey, an American serving in Cuba’s military aviation organization, flying out in an open-cockpit Curtiss Hawk II after concern about the cyclone’s position. He observed the storm and its movement without penetrating its cloud mass.
The distinction between observing a hurricane from an aircraft and flying into its eye is important. Povey’s flight should not be illustrated as a modern instrumented hurricane-hunter mission or described as if it carried the capabilities of later reconnaissance aircraft. Its historical significance lies in an early effort to use an airplane to improve knowledge of a dangerous storm.
The information still had to return to forecasters and then reach people making decisions. Observation, interpretation, warning, and action were linked stages, not one instantaneous event. A better position report could improve a forecast without restoring hours already lost.
At Long Key, the barometer continued to fall. In Islamorada, Russell and his neighbors were securing property. The camps remained vulnerable. The railway offered a possible way out, but a possible route was not yet a completed evacuation.
The approaching hurricane would soon destroy some of the very systems through which it was being understood. Telegraph and telephone lines could fail. Instruments could be lost. Observers could become survivors struggling to remain alive. The record of the night would be assembled afterward from what they remembered, preserved, measured, or found.
16. The house Russell trusted
John Russell had built a house he believed could withstand a great deal.
He described heavy timber, bolts, cement, and rods anchoring it to the ground. He had lived with hurricanes as a recognized part of island life. His family’s experience reached further back. When he explained his decision to remain, he did not describe indifference to weather. He described confidence based on preparation and the storms his people had already endured.
On Monday morning, he secured his property. Neighbors were doing the same. Boats had been moved into creeks. Russell studied the barometer and received weather information through radio and railway communications. The operator near the post office was part of the network through which warnings reached Islamorada.
His sworn testimony preserves the coexistence of two beliefs that can appear contradictory after a disaster. He knew a dangerous storm might arrive. He also believed his household could weather it. Russell statement, questions 7–18 and 35
That is an important distinction in the history of risk. To recognize a hazard is not necessarily to understand its eventual magnitude. To prepare is not necessarily to prepare for the conditions that will actually occur. Experience can sharpen judgment and at the same time supply a reassuring comparison that proves inadequate.
Russell did not extend his confidence to the camps. In his judgment, temporary buildings on low, exposed ground placed the veterans in a different situation. He told investigators that he would have moved them on Sunday if he had been responsible for them.
His statement was emphatic about the opportunity for earlier action. It was also precise about what he did not know. Asked whether the official in charge had received weather reports, Russell said he could not say. He had not seen him during the day. He and his neighbors had been occupied with their own preparations.
That answer should remain beside the criticism. It prevents a witness’s local knowledge from expanding into omniscience. Russell knew what he had observed and done. He had opinions about prudence. He did not have direct knowledge of every communication within the relief administration.
The house’s failure changed the meaning of all the work he had put into it. In his account, the wind alone did not destroy it. Water and wreckage broke it apart. He was injured in the side. His wife and three daughters died. One of his children survived.
The testimony presents these losses with the compressed language of an investigation. A question asks about injury. The answer reports a wound and the destruction of a family. The form continues to the arrival of relief. The reader is left with the distance between the scale of the loss and the small space allotted to it on the page.
Russell’s account also contains an observation about the railroad embankments. He believed they had held water against the land before breaking. He described a rapid change when the fills gave way and argued that more open channels would have reduced the loss of life.
That interpretation belongs in the history because a survivor made it under oath and because it directed attention to the built landscape. It should not be converted into a proven engineering conclusion without additional evidence. Water levels in a hurricane are affected by several interacting forces. The timing and geometry necessary to determine the contribution of each fill are not established by Russell’s statement alone.
Nor should the interpretation be discarded merely because it was not a technical model. Russell had been in the water. He reported a change he experienced. The responsible distinction is between the observation and the causal explanation he attached to it.
His testimony has another unusually clear division. He criticized the evacuation but praised the efforts to reach survivors. He believed the conditions after the storm made early access extremely dangerous. When asked whether he had criticism of the aid efforts, he said he did not.
It is possible to lose that complexity by choosing only the most accusatory lines. It is also possible to erase his criticism by quoting only his praise of rescuers. The whole statement resists both treatments. It describes a man who believed a preventable exposure had occurred and who respected the people trying to help afterward.
He spoke from Jackson Memorial Hospital in Miami on September 14, less than two weeks after the hurricane. Investigators D. W. Kennamer and H. W. Farmer asked the questions. The transcript now available through the Historical Preservation Society of the Upper Keys was prepared from archival material by Mary Lou Wilkinson in 2010.
Between Russell’s voice and a reader today lie the original investigation, the preservation of its records, and the later act of transcription. The chain is visible. It is not perfect access to the night, but it is enough to recover a particular person from the impersonal phrase “local residents.”
17. A train going toward the storm
The rescue train had to travel into the region from which it was supposed to bring people out.
That fact defines its predicament more clearly than any dramatic label. The railway was a narrow, continuous route across low islands and exposed crossings. Debris could obstruct it. Water could undermine it. Conditions could worsen between the outward journey and the return.
The relief administration requested railway transport to evacuate the work camps. The train was assembled and sent south. Later accounts describe delays associated with assembling equipment and crew during a holiday and with obstacles encountered along the route. The precise sequence of requests, dispatch decisions, and train movements should be reconstructed from the original orders and testimony, rather than forced into a single clock from conflicting summaries. NOAA, evacuation account
What the sources establish is the outcome. The train reached Islamorada on Upper Matecumbe. It did not carry out the intended evacuation of the camps farther along the route. The storm tide swept the cars from the rails, while the locomotive remained.
McDonald’s Weather Bureau report described an eleven-car train. The Coast Guard’s later aerial narrative described cars broken and displaced beside the line. The famous photographs show why the wreck became such a powerful public image: an object designed to move people along a fixed route had itself been moved out of that route. McDonald, 1935 report; Coast Guard operational narrative
The people aboard and the people in the camps must be kept separate in the accounting. NOAA’s retrospective reports that those on the train survived. That does not mean the train rescued the population it had been sent to evacuate. Conversely, the heavy deaths in the camps do not mean that everyone pictured near the wrecked train had died aboard it.
Even an official preservation document can repeat a faulty disaster narrative. The Florida Keys Memorial nomination contains a description of train deaths that conflicts with the NOAA account. The nomination remains useful for the monument’s architecture, but its institutional status does not make every historical sentence equally reliable.
The rescue train is therefore both an event and a warning about historical evidence. The visual impression of disaster is not a passenger list. The fact that a train was wrecked does not establish its occupants’ fate. A correct caption matters because an incorrect one can be repeated beneath an image for decades.
The Carroll A. Deering investigation similarly tests what a striking wreck photograph can—and cannot—establish about the people aboard.
For those who never reached the train, the consequences were immediate. An evacuation that had not happened before the worst weather could no longer be improvised simply by deciding that it should occur. Camps, houses, bridges, and communications were failing together.
The historical question of an earlier departure is unavoidable. Russell believed the veterans should have been moved on Sunday. That opinion has force because it accounts for their exposed quarters and the time needed to remove them. But a precise counterfactual—how many would have survived if a particular order had been issued at a particular hour—requires more evidence than a later narrative can honestly supply.
Earlier action would have provided a larger margin. Whether a given alternative would have succeeded depends on the route’s condition, the train’s capacity and movements, boarding arrangements, and the storm’s progress. Those conditions should be investigated, not invented to make a conclusion feel complete.
The failure also cannot be assigned solely to the last people visible in the sequence. A train crew confronting debris and rising water inherited decisions made before the locomotive moved. Evacuation planning, the location of camps, the standard for ordering departure, and the availability of transport belonged to a wider system.
That does not absolve any particular actor. It defines the questions that an investigation needs to ask. Responsibility is not determined by choosing the most memorable photograph and finding someone standing nearest it.
By the time the storm tide overtook the train, the railway’s two histories had converged. The great construction project of the early twentieth century had become an attempted rescue instrument in a Depression-era emergency. Its usefulness depended on the survival of a route whose builders had once advertised exceptional protection against storms.
The locomotive remained on the line. Around it, the continuity on which the evacuation depended was gone.
18. Fifty-five minutes on Long Key
Duane’s account begins the afternoon’s sequence with a falling barometer, heavy rain squalls, a high tide, and a sea swell.
At two o’clock, he placed the wind in the north or north-northeast. An hour later, he noticed large waves arriving from the southeast, in a direction that did not simply match the local wind. By four, the wind was increasing and the pressure dropping steadily. By five, he described hurricane force.
The value of the sequence is that it preserves change. A single maximum wind estimate or a single low pressure cannot show how an observer moved from concern to danger. Duane’s entries show conditions accumulating: water, wind, rain, falling pressure, then damage to shelter. Duane’s observations as published in McDonald’s report
At six, he recorded 28.04 inches on the barometer and water rising on the northern side of the island. At 6:45, the reading was 27.90. The wind had shifted toward the northwest. A large timber traveled from another part of the camp and struck his house. His account gave its approximate dimensions and described a near miss to three people.
There is no reason to convert those dimensions into an exact aerodynamic calculation. The report does not establish the beam’s full trajectory, mass, or flight conditions. Its evidentiary value lies in the observed effect: material from one structure had become a threat to another, well before the storm was finished.
By seven, the group was in the main lodge. It too was shaking and being damaged by flying timber. The water had reached the level of the railway on the camp’s north side. The Weather Bureau editor associated that water with the shallow expanse of Florida Bay under the drive of northerly winds.
At nine, Duane still saw no relief. Twenty minutes later, the wind diminished. He read 27.22 inches. The group moved again, this time to the last cottage he believed might survive the returning wind.
His report put twenty people inside it. He timed the lull at fifty-five minutes. The sky was clear toward the north, but he did not describe an absolute calm. A light breeze remained.
The interval was dangerous because the sea did not share the wind’s temporary restraint. About halfway through the lull, water began rising rapidly from the ocean side. Duane directed his flashlight outward, saw the advancing water, and had to race back. The cottage lifted and floated.
The scene is often simplified into an eye passing over an island, followed by the second half of a hurricane. The observations are more particular. Water approached from different sides at different stages. The wind shifted. The refuge changed from a fixed building to a floating structure before the next violent phase began.
At 10:10, Duane recorded 27.02 inches and a wind beginning from the south-southwest. Five minutes later, the returning blast struck with full force. The house was breaking up. He glanced at the instrument, which he reported at 26.98, lost it in the water, and was blown outside.
He became caught in damaged coconut fronds and held on. An object struck him. He lost consciousness.
The next surviving entry presents a difficulty that should not be concealed. The NOAA transcription labels his return to consciousness on September 3 as 2:25 p.m., then continues with observations about hurricane winds lasting until five in the morning and gales afterward. The page reproduces a sequence that needs checking against the original printed report before treating every time label as settled.
What does not depend on resolving that time label is the central account of survival. Duane reported regaining consciousness in a tree, roughly twenty feet above the ground. The water had receded. He reported that the cottage had been carried back onto the island and that the people inside were safe.
His survival does not make the scene representative of everyone nearby. The course of a floating structure, a piece of debris, or a person caught in a tree could turn on conditions too local and contingent to reconstruct completely. A survivor’s account tells us what happened to that survivor, not what must have happened to everyone exposed to the same hurricane.
Nor should his pressure readings be smoothed into a perfect modern chart. The readings are reported observations, with the limitations of an instrument used during the destruction of its surroundings. A chart should display the recorded points and explain gaps. Connecting them with a line can aid reading, but it should not imply continuous measurement between entries.
The record is powerful precisely because it remains human. Duane continued observing while trying to find shelter. The instrument stayed in the story until it could no longer be held. At that point the observer became a body in the water, and the record stopped until he survived to resume it.
Compare this documented weather record with the disputed storm narrative in our Flannan Isles investigation.
19. Ships carried across the reef
The hurricane’s force was not confined to land or railway.
McDonald’s report described the Danish motorship Leise Maersk being carried across Alligator Reef and grounding several miles beyond it. The engine room flooded, the vessel was disabled, and the superstructure was wrecked. The report stated that no crew members were lost and that the ship was salvaged on September 20.
Nearby, the American steamship Dixie grounded on French Reef. McDonald reported no loss of life aboard and recorded its refloating on September 19, followed by towing to New York. The tanker Pueblo lost control for hours and was carried around the storm’s circulation before narrowly avoiding another reef. McDonald, 1935 report, vessel accounts
The ships supply a different measure of the event. A railway car displaced from its track shows one relationship between moving water and a heavy object. A vessel carried across a reef shows another. Familiar boundaries—shore, navigable channel, reef, anchorage—were altered by the combination of elevated water, waves, and wind.
The reported tracks should not be turned into exact animated paths without the underlying logs and charts. A captain’s description of being carried inland over a reef establishes a remarkable displacement. It does not necessarily provide every position needed to draw a precise curve minute by minute.
The Dixie also appears in the Coast Guard’s operational narrative. Initial information placed the vessel near Carysfort Reef. Later aerial observation located it on French Reef. The difference shows a developing rescue picture being corrected as personnel acquired direct knowledge.
An editor could silently replace the first location with the second and produce a cleaner account. A better history preserves the distinction where it matters. Rescuers acted on reports of varying quality. What they knew at one moment was not necessarily what later observation established.
The report describes requests for assistance arriving while aircraft remained unable to leave the hangar. The existence of a vessel in distress did not remove the wind preventing a safe flight. Personnel could prepare equipment and pass information, but they could not make the storm cease at the station or over the route.
At sea, survival also varied. The absence of deaths on the named ships does not lessen the disaster on the islands. It shows why a hurricane’s casualty history cannot be read simply from the apparent scale of wreckage. Some heavily damaged vessels retained enough refuge or good fortune to keep people alive. Some buildings on land did not.
In the Keys, the 1935 hurricane was a single weather system acting on very different settings. A fishing camp, a veterans’ hut, a grounded steamer, a lighthouse, an embankment, and a locomotive each responded according to its position and physical circumstances.
McDonald’s report attempted to connect those effects to the storm’s movement. It distinguished wind damage from the additional destruction caused by tidal inundation. Without the water, he argued, the damage would have been severe but less complete.
The railway’s failure is most clearly understood within that larger picture. The storm did not merely blow against an engineered object until it broke. It changed the surface through which the route ran. It displaced material, eroded support, carried wreckage, and cut the connections on which transportation depended.
The ships stranded beyond their usual waters remained visible after the storm. So did the train. They became landmarks in a landscape whose ordinary markers had been removed. To the rescuers approaching by water or air, those large objects helped reveal the scale of what had happened. They could not yet reveal how many people were still alive among the smaller wreckage.
20. The morning the aircraft could leave
At the Miami Air Station, the hangar door opened at 3:30 on the morning of September 4.
The operational narrative recorded that the wind had moderated sufficiently for personnel to clear the ramp and check it. Aircraft had been inspected while they were grounded. Now the preparations could become flights.
At daylight, Lieutenant Olsen departed in amphibian C.G. 133 on a survey and relief mission. The party included a sanitary engineer and a news representative. The aircraft’s identifier is firm in the transcript. The aircraft-type suggestions inserted by the later editor are not part of the original narrative and should not be treated as confirmed technical metadata. Miami Air Station operational narrative, September 4
From above, the crew saw a broad zone of destruction. The report described demolished buildings, damaged vegetation, scattered lumber marking the camps, and rails displaced from their roadbed. At Islamorada, it described the locomotive still on the track and battered cars beside it.
The description of the camps is particularly stark. A place previously occupied by working men was recognizable through the remnants caught in vegetation and the fragments of overturned buildings. A camp’s administrative identity survived more clearly than its physical form.
On Upper Matecumbe, the party encountered survivors and bodies among the wreckage. Doctors and relief workers were already providing aid. The aircraft’s return to Miami at 9:45 brought information that could help direct the larger response.
This flight was both observation and intervention. Reports sent during the mission and after its return helped officials understand where damage and need were concentrated. An aerial survey could make a broken region legible more quickly than travel along a road whose bridges and fills had failed.
Another aircraft, C.G. 255, covered areas of the mainland and adjacent Keys. The narrative describes stops near stranded boats and wrecked houseboats, searches for people believed to have been in exposed places, and the removal of survivors who needed assistance.
At Upper Matecumbe, the pilot landed in shallow water over mud. The pilot and mechanic waded ashore. The report describes injured men, women, and children, with volunteer workers, doctors, and Coast Guard personnel caring for them among bodies and wreckage.
The pilot judged the need urgent enough to carry sixteen women, children, and injured men. The report explicitly described the plane as overloaded for the departure and the takeoff as difficult. One of those transported later died in Miami.
This is a recorded emergency judgment. It should not be romanticized into a general principle that ignoring operating limits is admirable. The historical fact is that the pilot made a particular decision under severe conditions, and the station recorded both the difficulty and the result.
Even arrival in Miami did not finish the movement. Strong winds prevented the plane from being brought up the ramp. Survivors were transferred by the station’s crash boat. Rescue was a sequence of passages: from damaged land to aircraft, from aircraft to boat, and from boat to medical care.
The report contains many such practical details. They resist the tendency to describe relief as a single arrival of help. A rescuer needed to reach a survivor, assess the immediate problem, find transport, negotiate the damaged environment, and hand the person into another part of the response.
The railway had once made the journey across the Keys seem continuous. The hurricane restored all the interruptions. A washed-out creek crossing could become a major barrier. A landing place could be too shallow or obstructed. A usable aircraft could still be held by weather. A person who survived the storm could remain in danger from injury and isolation.
Russell understood those difficulties. In his hospital testimony, he recalled an aircraft overhead and the arrival of help. He emphasized how hazardous the conditions had been. His criticism of the failure to evacuate did not become criticism of people attempting to reach the islands afterward.
The distinction matters for the rescuers as well. Their courage cannot retrospectively make the evacuation adequate. Their work addressed a catastrophe that had already occurred. It deserves recognition on its own terms, without being made an argument that earlier decisions no longer require scrutiny.
By the end of the first flights, the outside world had a clearer picture. The next phase would bring officials, further searches, medical transport, and the recovery of the dead. The aircrews would continue to cross a region in which some homes could no longer be located because even their wreckage had disappeared.
21. The search beyond the camps
On September 6, C.G. 255 carried a man searching for his daughter and son-in-law.
The Coast Guard narrative identified him as J. G. Stoddard. The couple had been staying at his home on Middle Cape in the Cape Sable region. The aircraft went to look for them. The report records a negative result: the area had been swept clear, and even wreckage from the house could not be found.
That flight places another family within the hurricane’s history. The veterans’ camps and the wrecked rescue train understandably drew national attention. The storm also reached people whose exposure had nothing to do with either project. A private visit, a fishing trip, an established home, or an ordinary day’s work could place someone within the same field of destruction. Coast Guard operational narrative, September 6
The absence of visible wreckage was itself a problem for rescuers. A recognizable building can help locate a search. When the building is gone, the search depends on geography, prior knowledge, reports, and whatever traces remain. An aerial crew could scan an area without being able to prove the fate of every person expected to have been there.
On that same mission, the aircraft was hailed by a rowboat near Flamingo. Tom Conroy, a survivor needing medical attention, was taken aboard and brought to Miami. One flight could contain both the failure to find a missing family and the successful removal of someone still alive.
The operational report does not arrange those outcomes into a consoling balance. It records them as tasks and results. The distinction is useful. A rescue does not offset a death in any meaningful personal accounting. It is a life preserved amid other losses that remain losses.
Other flights carried officials. Governor David Sholtz and George Ijams, a senior Veterans’ Administration representative, surveyed the stricken area and went ashore. The report associated their visit with the state of relief work and the handling of bodies.
The record becomes difficult reading at this point. Bodies were decomposing in the heat. Some remained in water or wreckage. Officials considered cremation necessary under the public-health assumptions and practical pressures of the time. Describing those assumptions is different from endorsing every medical justification offered in the report.
For a history of the hurricane, the consequences are central. Decisions made in the immediate aftermath affected whether remains could be individually identified, returned to relatives, or buried separately. The physical destruction caused by the storm was followed by administrative choices that shaped remembrance for decades.
On September 7, C.G. 133 assisted in searches for bodies in Florida Bay. The report recorded twelve located and their positions communicated to personnel responsible for recovery. The sea had carried the catastrophe beyond the places where the victims had been living.
This widened the search area and complicated identification. A person found far from a camp could not be assigned to it merely by proximity. A body recovered later might revise an earlier missing-person entry. A personal item might help establish identity, but an absence of such an item did not establish that no identity existed to recover.
The air station’s narrative concludes its rescue account with names of people transported. Among them are John A. Russell and his son Bernard. Families appear together; some individuals are named alone; two seriously injured men were recorded without identification.
The list has a different character from the familiar disaster total. It records people moving through a service: picked up, carried, delivered to care. Its incompleteness is visible. It is not a complete list of survivors, and it should never be presented as one. But it restores the scale of an individual transport to a story dominated by aggregate loss.
The same records show how the response became increasingly administrative. Flights carried relief officials, investigators, and people organizing the disposition of remains. The mission changed from finding and removing survivors to documenting damage, recovering bodies, and accounting for what had happened.
Those activities overlapped. There was no single hour when rescue ended everywhere and investigation began. The storm had affected a broad, broken geography. Information arrived unevenly. Different organizations had different tasks and access.
The railway’s absence remained part of every task. A line built to remove the interruptions of island travel had been interrupted itself. The response had to work through boats, aircraft, damaged road sections, and surviving local knowledge. The old maritime landscape reasserted its practical demands just when the need to cross it was greatest.
22. The numbers after the names
McDonald’s 1935 report gave a Red Cross estimate of 409 dead or missing in the Keys. It divided that total into 244 known dead and 165 missing.
The arithmetic is clear. The status of the number is equally important: it was an estimate furnished at a particular stage of the aftermath. It was not a declaration that no later recovery or correction would change the accounting. McDonald, 1935 report, casualty paragraph
Later sources give other totals. The Library of Congress description accompanying a rescue-train photograph refers to at least 485 deaths and about 260 veterans. A NOAA anniversary narrative uses another overall estimate. A Veterans Affairs commemorative article concentrates on veterans described as buried at Woodlawn and others reported missing. These statements differ in date, purpose, and category. Library of Congress, rescue-train record; VA, hurricane veterans remembrance
The responsible broad statement is that the hurricane killed hundreds of people, including a large number of veterans and local civilians. A more exact statement should identify the source and the population being counted. It should not borrow a number from one record and attach the definition from another.
There are several reasons the accounting could change. Men might be on a payroll but away from camp. Bodies might be recovered after an initial report. Identifications might be tentative, confirmed, or corrected. A civilian could also be a veteran without being enrolled in a veterans’ work camp. A list of camp employees therefore cannot automatically serve as a list of all veterans who died in the storm.
The physical treatment of remains added further difficulty. Some were buried; some were cremated; some were not individually recoverable. A burial total measures a disposition, not necessarily the entire loss of life. A memorial total may represent a defined group being commemorated rather than every victim of the event.
Errors also enter through transcription. A later table can contain a subtotal that does not equal its rows, or a date that is plainly inconsistent with the event. Correct arithmetic is necessary, but silently repairing a transcription does not prove what the original record said. The correction should be checked against the underlying document before it becomes an authoritative total.
This is why the charts above identify their categories. A chart labeled simply “Deaths” can conceal whether a value means confirmed dead, dead and missing, veterans in particular camps, civilians in a defined region, or an estimated total for the hurricane’s whole path.
Precision in these distinctions is a form of respect. A person is not a flexible unit that can be shifted between categories to make an infographic more dramatic. Nor should uncertainty be used to imply that nothing can be known. The records establish a catastrophe and preserve many identities even when a single reconciled total remains difficult.
The changing numbers also reveal the labor of the aftermath. Someone had to compare rosters, seek relatives, examine recovered remains, correct names, and distinguish the missing from people found elsewhere. That work was less visible than the rescue train and continued after the most dramatic photographs had been taken.
For families, an accounting category could have immediate consequences. Confirmation might end a search. An uncertain identification might leave questions unresolved. A record of burial or cremation could determine where relatives would go to mourn, or whether they would have an individual grave at all.
The hurricane thus produced two kinds of absence. It removed people from homes, camps, and communities. It also left gaps in the record by which the living tried to locate and identify them. The first could not be reversed. The second could sometimes be reduced through patient work, but never by replacing uncertainty with an attractive number.
The history must hold both scales together: the hundreds lost and the individual names recovered from the files. Without the total, the magnitude becomes difficult to grasp. Without the names, the magnitude becomes impersonal. Neither alone is an adequate account of what the water took.
23. Who had been responsible?
The inquiry began while relief and recovery were still under way.
A National Archives guide identifies investigative files concerning the Labor Day hurricane within the records of the Veterans’ Administration. It describes a requested joint investigation involving the veterans’ agency and the relief administration and notes that the investigators reached divergent conclusions. The files include an index of veterans, a reminder that the investigation involved both administrative responsibility and the fate of particular men. National Archives, guide to the hurricane investigation records
The disagreement matters. A statement that “the government investigated” can suggest a single body reaching a single settled conclusion. In this case, different officials and investigators approached the event with different responsibilities, information, and interpretations.
The fundamental questions were concrete. Why had the camps been placed where they were? What protection did their buildings offer? Who had responsibility for deciding to evacuate? What warnings had reached those people? When was transport requested? What prevented or delayed the movement? What could reasonably have been done before conditions became impassable?
The answers should not be inferred from the catastrophe alone. A terrible outcome is not, by itself, a complete explanation of the decisions that preceded it. At the same time, the exceptional severity of a storm does not automatically answer questions about avoidable exposure to it.
Russell’s sworn statement entered that field of inquiry. He described what he had observed, what he believed should have happened, and what he did not know. The investigators’ questions repeatedly used the language of prudence. They sought to establish whether warning signs would have prompted a reasonable person to move the men earlier.
That testimony has more value when read as testimony rather than as a verdict. Russell’s opinion that the camps should have been cleared on Sunday is evidence of a knowledgeable local resident’s judgment. It does not establish that a court adjudicated negligence against a named official.
Public criticism also came from outside the agencies. Ernest Hemingway visited the damaged area and wrote angrily about the veterans’ fate. NOAA’s retrospective records his criticism of the Roosevelt administration. His presence gave the disaster another route into national attention, but the fame of a writer does not make his account a substitute for dispatch records, camp files, or sworn evidence. NOAA, aftermath and public criticism
Congress subsequently held hearings. The verified catalog identifies Florida Hurricane Disaster, published by the Government Printing Office in 1936, as hearings before the House Committee on World War Veterans’ Legislation concerning H.R. 9486. The proposed legislation addressed relief for widows, children, and dependent parents of veterans who died in the hurricane at Windley Island and the Matecumbe Keys. HathiTrust catalog, 1936 hearings
The title itself reveals a distinction between investigating responsibility and providing relief. Congress could consider assistance to dependents while also examining how the deaths occurred. Those are related purposes, but they are not identical. A payment mechanism is not necessarily a judicial finding of liability, and a hearing is not the same procedure as a civil trial.
The catalog establishes the volume’s existence, scope, and publication details. It does not grant this article direct access to every page of testimony. Where the original hearing pages have not been reviewed, they are not quoted or paraphrased as if they had been. The verified survivor and operational transcripts provide the detailed evidence used here.
The limitation leaves some questions unresolved, particularly a complete minute-by-minute evacuation chronology and a full comparison of the agencies’ final positions. It also prevents a false conclusion. The article can describe a failed evacuation, documented criticism, and disagreement in the investigative record without pretending to have adjudicated every individual’s responsibility.
The distinction is especially important on a legal website. Historical language can drift into legal language without acquiring the evidence or procedure that gives a legal finding its meaning. “Negligence” in a witness’s answer, an investigator’s analysis, a newspaper’s accusation, and a court’s judgment are different uses.
The work of historical judgment remains necessary. Temporary camps on low islands, a known hurricane environment, warnings before landfall, and an evacuation that did not occur in time are not trivial facts. They demand explanation. Careful qualification should make that explanation stronger by showing exactly where the evidence supports it.
The storm moved on. The argument over what people had done, and what they should have done, remained. Decades later, another legal dispute would arise from the same deaths. Its question would be different: who could ask the government to place an individual veteran’s name on a memorial marker?
24. A map over a crypt
The Florida Keys Memorial in Islamorada was dedicated on November 14, 1937.
Its design translated the hurricane into stone and ceramic. The National Register nomination describes broad steps, a coral-limestone shaft, and a tile map of the Keys above a crypt. The relief imagery evokes palms bent by wind and moving water. The map gives geographical form to a catastrophe whose victims had been scattered across land and sea. Florida Keys Memorial, National Register nomination, architectural description
The architectural survey measured the tiled feature at twenty-two by ten feet and described an eight-compartment crypt beneath it. Those are observations about the monument as documented by the survey. They are more securely supported by that source than its conflicting account of the rescue train.
The distinction illustrates why a source should be evaluated claim by claim. A preservation nomination may contain excellent descriptions of materials, dimensions, and design while repeating an error in an associated historical narrative. The useful evidence does not have to be discarded. The error does not have to be inherited.
The memorial’s imagery is collective. Islands, waves, palms, and stone can represent a disaster affecting many people. A map is particularly apt because the storm had broken the connections between places. It can reassemble the chain in a single view.
But collective remembrance and individual identification answer different needs. A monument can acknowledge a community’s loss without giving every victim a separate visible name. For a family, a fellow veteran, or a later researcher, that absence can become the next problem to address.
The handling of remains after the storm made individual commemoration more difficult. Cremated remains of veterans and civilians were commingled. A common place of burial could not necessarily be divided into identifiable individual graves. The distinction between a grave marker and a memorial marker would eventually become legally significant.
The nomination records the dedication ceremony and identifies a young survivor, Fay Marie Parker, as the person who unveiled the monument. That detail places a living witness at the moment when the disaster became a formal public memorial. It does not tell us what she thought or felt during the ceremony, and there is no need to invent it.
The event occurred a little more than two years after the hurricane. For those attending, the loss was recent. The destroyed houses, camps, and railway were not distant historical subjects. People had known the dead. Some had survived the same water. Others had worked in rescue and recovery.
The monument also stood beside a changing transportation corridor. The highway would carry people through the Keys, bringing new visitors past a memorial to a disaster that had helped end rail service. Travel and remembrance occupied the same landscape.
That placement created an enduring tension. A visitor could see the monument during a holiday journey through beautiful surroundings. The memorial asked the visitor to recognize that the same islands had been a place of mass death. The sea did not need to look threatening for the history to remain present.
Public monuments often compress complicated events into a date and an image. Compression makes remembrance accessible. It also creates the need for fuller histories that explain the people, institutions, and circumstances behind the inscription.
Here, the full account must distinguish the original railroad workers from the highway veterans, the people aboard the rescue train from those who never reached it, and the engineering achievement from the administrative decisions of 1935. Without those distinctions, even sincere remembrance can preserve a distorted story.
The watercolor illustrations in this article must also be read with their limits intact. They may help a reader imagine a low island, a work camp, a railway window, or a flooded route. They must remain identified as modern interpretations. They are not photographs of named victims, authenticated final moments, or measured reconstructions of the disaster.
The map over the crypt has another lesson for historical storytelling. A place can contain more than one layer of evidence. The surface offers an image of the islands. Beneath it are remains that cannot all be separated. Around it is a road carrying new movement through the chain. The historian’s task is to keep those layers visible, rather than allowing the most beautiful one to conceal the others.
25. Roy Anderson’s name
Roy H. Anderson served in the United States Army from April 1918 to March 1919. His wife died in 1930. They had no children.
Those facts appear in a court opinion issued in 2022. They explain why a hurricane death from 1935 could still produce a dispute nearly nine decades later. Anderson had been assigned to a work camp on Lower Matecumbe. According to the opinion’s account of the record, his body was found on October 10, 1936. His cremated remains were commingled with others and interred at the Florida Keys Memorial. No separate marker had been provided for him.
In August 2017, Richard C. Bareford applied for a government-furnished memorial marker on Anderson’s behalf. The dispute that followed concerned whether Bareford was an eligible applicant. Bareford v. McDonough, February 28, 2022
The legal distinction was between a burial marker for an identifiable grave and a memorial marker where remains were unavailable or not individually identifiable. The regulations discussed in the opinion allowed a broader group of applicants for the first category than for the second. The memorial-marker provision limited applicants to family members.
That distinction had a particularly sharp effect in a case like Anderson’s. The absence of an identifiable individual grave was a consequence of the hurricane and its aftermath. The absence of a spouse or children made a family-only application rule another obstacle to individual remembrance.
The question before the Court of Appeals for Veterans Claims was not whether Anderson was an eligible veteran. The parties did not dispute that point. The question concerned the agency’s restriction on who could request the marker.
The court did not accept the broadest argument that Congress had unambiguously permitted anyone to apply. It concluded that the statute left a gap concerning applicants. But it held that the particular restriction on memorial-marker applicants was arbitrary and capricious. The court set aside that regulation, vacated the Board’s denial, and remanded the matter.
These are the holdings of the 2022 decision, described as historical legal developments rather than instructions for filing a current benefits claim. The distinction matters because the law and its administration can change. The opinion also used a framework for reviewing agency interpretations associated with its own time. There is no need to treat every part of that framework as a statement of present law to explain what happened in Anderson’s case.
An appeal reached the Federal Circuit. Its order dated April 29, 2024, dismissed the appeal by agreement, with each side bearing its own costs. That disposition should not be described as a separate merits opinion affirming every statement in the lower court’s reasoning. Federal Circuit, dismissal order, No. 2023-1189
The case did not adjudicate negligence in the evacuation of 1935. It did not award damages for the hurricane deaths. It concerned the ability to request a memorial marker. Conflating it with a judgment about the original disaster would misrepresent both the litigation and the history.
Its importance is nevertheless substantial. The hurricane had left Anderson without an individually marked grave. Decades later, an administrative rule determined who could ask the government to address that absence. A legal challenge brought the rule under review.
The chain from storm to courtroom is unusually clear: death, recovery, commingled remains, collective memorial, application for individual recognition, denial, appeal, and judicial decision. Each stage preserved or created records that make the history more specific.
Anderson is no longer only one unit within an uncertain total. The opinion gives his name, service dates, family circumstances, camp assignment, and the history of his remains. It demonstrates how legal records can recover personal detail even when they are created for a dispute far removed in time from the original event.
The railway’s story began with a celebrated individual whose name became inseparable from an achievement. Anderson’s later story concerned a different problem: how to ensure that one less famous individual did not remain inseparable from an anonymous collective loss.
26. The road on the old piers
By July 1938, a National Park Service official could describe a continuous drive from Miami to Key West.
C. R. Vinten’s article, “A Highway Over the Sea,” appeared in the first volume of the Park Service’s Regional Review. It reported that the state had purchased the former railway right-of-way and bridges for $640,000 as part of a larger Public Works Administration loan. New road construction and adapted bridge infrastructure had closed a gap previously crossed by railway and ferry. Vinten, “A Highway Over the Sea,” July 1938
The article is contemporary evidence that the highway was already open to vehicular traffic at that point. It also reveals the continuity beneath the change. The railway did not vanish without leaving usable structures. Its route and bridges became part of a different transportation system.
That outcome should not be described as simply paving the rails. Converting infrastructure designed for trains to carry road traffic involved different requirements. Road width, deck arrangement, approaches, and traffic movement had to be addressed. The National Register documentation describes adaptations to the old bridges, including the distinctive treatment of Bahia Honda, where the road was carried above the through trusses rather than through them at the former railway level. National Register, original bridge nomination
The difference is visible in the bridge’s layered silhouette. What appears in a later photograph is not necessarily what Flagler’s train crossed. A structure can retain its earlier foundations and major steelwork while presenting a substantially altered route to the next generation of travelers.
Vinten described work by the Civilian Conservation Corps as a camp established quarters on West Summerland Key. The setting introduced conservation workers accustomed to larger land areas to a narrow island chain. The article’s interest was not merely how to move cars across water. It was how to manage the landscape that the road would make accessible.
In the opening railway booklet, accessibility had largely been treated as a triumph in itself. By 1938, Vinten explicitly anticipated the pressures of increased private development. The road would attract visitors. Visitors would create incentives for building. Without planning, the qualities drawing them south could be diminished by the development intended to serve them.
The proposed responses included land acquisition, broader rights-of-way, scenic easements, and the considered placement of visitor facilities. Vinten described cooperation among state agencies and the National Park Service. He placed conservation ahead of indiscriminate physical development.
This is not a modern environmental argument retrospectively assigned to 1938. It appears in the contemporary article. The highway’s promise and its potential consequences were being discussed together as the route opened.
The shift from railway to road also changed how people encountered the Keys. A train carried travelers according to a service and its stops. A road distributed decisions among motorists, businesses, public agencies, and property owners along the corridor. That difference does not make one mode inherently virtuous and the other destructive. It changes the pattern through which access can generate development.
The old infrastructure made the transition possible, but public investment and new work were necessary to realize it. The successor route was not a private railway continuing under another name. It was a new arrangement of ownership, financing, construction, and use.
The victims’ memorial stood within that changing landscape. Motorists could travel along structures inherited from the railway and pass a monument to the hurricane that had ended its service. The road made the story more accessible even as it altered the setting in which the story had occurred.
The freight-car ferry operation found another geography as well. The FEC’s company history records its move from Key West to Port Everglades after the hurricane and later to Palm Beach. The end of rail service to Key West did not mean every maritime connection associated with the company ended at the same moment. Some functions moved elsewhere. FEC, freight-car ferry history
That is a more accurate ending for a transportation system than total disappearance. Some parts were destroyed. Some were abandoned. Some were adapted. Some business moved. The railway as a continuous passenger and freight route to Key West was over, while its material and organizational consequences continued.
For a traveler looking from the later highway toward an old span, the relationship can be difficult to untangle. The bridge may appear as a relic standing beside its replacement. In fact, it may have carried more than one kind of traffic, undergone several alterations, and survived because it remained useful long after its first purpose ended.
27. Reading the surviving landscape
The surviving structures do not form a single intact scene from 1912.
They are remnants of construction, operation, storm damage, highway conversion, maintenance, replacement, and preservation. A concrete arch can belong to one phase while its deck or approach belongs to another. A modern road may run near an old alignment without reproducing every historical curve and crossing.
The National Register’s later amendment to the Overseas Highway and Railway Bridges district expanded the documented collection of structures. Its inventory describes multiple bridge types and their relationship to the original route. It is a record of the infrastructure as surveyed at a particular time, not a guarantee of current condition or public access. National Register amendment, Overseas Highway and Railway Bridges
That qualification matters when an article includes a map. An approximate regional point can orient a reader to the Florida Keys. It cannot establish a safe place to park, a publicly accessible route onto an old bridge, or the precise site of a vanished camp. Those require different information.
For regional orientation, the NOAA Florida Keys habitat overview helps explain the island setting. It does not establish an authenticated coordinate for the rescue train, Duane’s cottage, a veterans’ hut, or an individual recovery location. Historical uncertainty should not disappear merely because a digital map can display a precise-looking pin.
A more exact historical map would need several separate layers: the railway’s operating alignment, the highway and ferry arrangement before the hurricane, the camp locations, the storm’s reconstructed track, the surviving structures, and changes to shorelines and fills. Each layer has its own date and source.
Combining them carelessly can produce an attractive fiction. A modern shoreline placed beneath a 1935 storm track may be useful for broad orientation, but it does not prove the exact inundation of the earlier landscape. A current highway centerline is not automatically the former railway track. A symbol for a camp can imply a known footprint when only its general location is established.
The same discipline applies to illustrations. The pile-driving watercolor evokes a documented construction activity. The survey image communicates the challenge of low islands and channels. The opening celebration interprets an event whose date and public character are documented. The camp painting depicts temporary highway-worker quarters, not the original railway’s labor camp.
Those paintings can make the article inviting without claiming the authority of an archival photograph. Captions identify them as interpretations. Where a structure’s exact arrangement is not established, the image should not pretend to settle it.
The landscape also preserves different kinds of absence. A bridge standing without rails is a visible absence of service. A camp no longer standing requires a record to make its former presence legible. A family home swept away may survive only in testimony, photographs, property records, and memory.
The distinction helps explain why preservation cannot focus only on the largest structures. The monumental bridges are compelling. They are also the parts most likely to survive long enough to become monuments. Smaller, more vulnerable buildings can disappear precisely because they were less able to withstand the forces that killed the people inside them.
If the history follows only what still stands, it risks becoming a history of the strongest objects rather than of everyone who lived around them. The records of Russell, the veterans, the rescuers, and the missing families counter that bias.
These are not obstacles to compelling history. They are the means by which a compelling account remains trustworthy. A reader can be drawn into a place while still knowing what is observed, what is documented, what is interpreted, and what remains unknown.
The old railway’s surviving spans invite that kind of attention. Their beauty is real. So is their engineering significance. But the most complete view includes the water between the piers, the workers who made them, the people who relied on them, and the storm that revealed how much a route depended on what stood beyond the most impressive bridge.
28. What remained of the promise
In 1912, the railway company described an achievement that would endure through the ages. Its writers imagined the viaducts standing firm against tropical storms and the extension binding the islands to the mainland with a permanence that seemed almost beyond question.
Some of the structures did endure. That is part of the truth. The surviving arches and steelwork are not merely illustrations of misplaced confidence. They are evidence of substantial engineering and of materials that could be adapted to later purposes.
The promise failed at the scale of the whole route. A continuous railway could be lost even where great portions remained. Track, embankments, approaches, communications, equipment, finance, and management all belonged to its ability to function. The hurricane struck that connected system, not just the most photogenic bridges.
The human promise was more difficult. The railway had helped make the islands accessible. In 1935, accessibility became a matter of life and death. A train could potentially remove people from the camps, but only if the evacuation became action while the route still permitted it.
The surviving accounts do not offer a single convenient moral. Flagler’s ambition produced a remarkable railway. The labor behind it carried hardship and loss. Island residents possessed knowledge of hurricanes and could still be overwhelmed. Warnings existed, yet the camps were not emptied in time. Rescue personnel acted courageously in the aftermath, while questions about earlier decisions remained.
These facts belong together because they describe the same history. Separating them into a celebration of engineering and an unrelated disaster story makes both accounts less complete.
Duane’s flashlight gives the story one scale: a short distance between a man and a cottage as water rose around him. Russell’s testimony gives it another: a house anchored with care, a family lost, and an insistence that men in more vulnerable quarters should have been moved. The Coast Guard’s reports add the practical scale of rescue flights, shallow landings, damaged routes, and people carried to medical care.
The opening booklet contributes the scale of an enterprise. It names structures, distances, materials, and administrators. Browne’s city history restores a society at the terminus. The memorial and Anderson’s case show that the aftermath continued beyond the clearing of wreckage or the opening of a new road.
None of those sources can do the work of all the others. Together, read with their limits intact, they make a more complete record than the familiar photograph of the train alone.
The route’s final transformation is visible in the highway. What had been built for locomotives became part of a road for motorists. The new movement carried people past the old works and through a landscape that remained beautiful, commercially valuable, and exposed to the sea.
Beauty did not refute the danger. Danger did not erase the lives people made there. The Keys were never only a paradise awaiting a railway, nor only a disaster waiting for a hurricane. They were inhabited places, joined to larger histories through work, trade, weather, government, and the choices people made under conditions they could not fully control.
For the Carolina and Lake Norman readers who encounter this story through the LKN history collection, the distance to the Keys need not make it remote. It asks recognizable questions about ambitious construction, the treatment of workers, the use of local knowledge, public responsibility, and the preservation of names. Those connections are stronger than an invented claim that every event happened close to home.
The train that reached Key West in 1912 demonstrated that the islands could be joined by rail. The hurricane of 1935 demonstrated that a connection could be broken. The records left by builders, survivors, rescuers, and investigators make another connection possible: between the people who lived through those events and the readers willing to follow their histories beyond the legend.
Sources and reading notes
Weather Bureau, McDonald’s 1935 report
Florida East Coast Railway, opening booklet, 1912
Browne, Key West: The Old and the New, 1912
NEH, the Henry Morrison Flagler Museum
National Register, Overseas Highway and Railway Bridges amendment
Original National Register nomination
Florida Memory: Flagler arriving
FEC, freight-car ferry history
NOAA, 1935 hurricane retrospective
Russell sworn statement, September 14, 1935, questions 1–6
Miami Air Station operational narrative, September 1935
Library of Congress, rescue-train record
VA, hurricane veterans remembrance
National Archives, guide to the hurricane investigation records
HathiTrust catalog, 1936 hearings
Florida Keys Memorial, National Register nomination, architectural description
Bareford v. McDonough, February 28, 2022
Leave a Reply