The decision made in Washington determined what shape the hole between the two oceans would take — and how many people would pay for it.
In January 1906, a document landed on President Theodore Roosevelt’s desk that was meant to settle the argument. The international advisory engineering panel — the Board of Consulting Engineers — had delivered its verdict after months of work: the canal should be built at sea level, without locks. Eight members of the board voted for that option. Five against. It seemed that the democracy of engineering reason had worked: the majority preferred the design that looked more elegant. A ship enters the canal at one end and leaves at the other, no lifts, no chambers, no gates as tall as a five-story building. Pure geometry: cut through the isthmus from coast to coast, the way a knife cuts through butter. That was precisely how Ferdinand de Lesseps had triumphantly cut through the Isthmus of Suez — and precisely how he had tried to proceed in Panama, before the tropics devoured his company, his reputation, and the money of tens of thousands of French investors.
And yet the final word did not belong to the board’s academics. John Frank Stevens, the project’s chief engineer, was no academic, no armchair strategist, but a man who had already spent six months living in Panama’s jungle and knew the local earth, the local water, and the local rain. Stevens flatly rejected the majority recommendation. His counterproposal sounded audacious: dam the valley of the Chagres River with a gigantic barrier, create an artificial lake 26 meters above sea level — at the time, the largest man-made lake in the world — and raise ships to it by locks. Roosevelt, having heard both sides, took Stevens’s side. The president, known for his love of grand gestures, chose the option that looked more modest. Roosevelt trusted the men standing knee-deep in mud.
To understand Stevens’s logic, you have to understand the Chagres River. It crosses the future canal route and behaves as tropical rivers do: calm for most of the year, then, in the rainy season, a furious muddy surge capable of raising the water level by ten meters or more in flood. The Isthmus of Panama receives about 2500 millimeters of rainfall a year, with the bulk of it crashing down over six or seven months. The Chagres becomes a weapon in earnest: its floods swept away bridges and inundated construction sites even under the French.
A sea-level canal meant dealing with the Chagres another way. The river would have to be diverted — a separate drainage channel built so that floodwaters could not burst into the navigable trench. That was a colossal additional task: essentially, building a second canal, only for the water interfering with the first one. Stevens’s lock design solved the problem elegantly: Gatun Dam turned the Chagres from an enemy into a resource. The river filled the lake, the lake fed the locks, the locks passed the ships through. Floods were absorbed by the vast sheet of water — the lake worked like a gigantic buffer, like a bathtub into which you can pour as much as you like until it overflows. And to keep it from overflowing — a spillway. Put bluntly, Stevens proposed not to wage war on geography, but to come to terms with it on its own terms.
There was a second argument, purely arithmetic. A sea-level canal required cutting through the isthmus down to the zero mark — to its full depth, along its entire length. A lock canal allowed ships to travel across the surface of Gatun Lake, with deep excavation needed only at the narrowest and highest point — around the mountain ridge near Culebra. The difference in earthmoving volume was not a matter of percentages, but multiples.
Even with the lock design, the Culebra Cut became the hardest stretch of the project. The Americans removed more than 76 million cubic meters of earth from it, using about 27,000 tons of dynamite. The figures sound abstract until you picture the railway trains hauling away rock without pause — hundreds of trains a day, Bucyrus steam shovels working in shifts, tracks being relaid as the cut deepened.
But the main problem was not the volume; it was the nature of the ground. The walls of the cut had no desire to stay where they were. In 1907, one of the most famous landslides occurred — near Cucaracha: 400,000 cubic meters of earth slid down, filling what the workers had just excavated. Imagine it: you spend weeks deepening a trench, and then overnight the mountain takes back what was taken from it. It was not a one-off event. Landslides continued throughout construction and — crucially — after the canal opened in 1914. The geology of the Isthmus of Panama includes layers of clay that lose stability when saturated with water. Tropical downpours provide that saturation with predictable regularity.
For a sea-level canal, the problem would have been many times worse. The cut would have had to go dozens of meters deeper — down to zero relative to the ocean. The deeper the trench, the higher and less stable its walls. The higher the walls, the greater the mass of earth that can slide down. The greater the mass, the wider the cut must be made to reduce the slope angle. And the wider the cut, the more earth must be removed. A vicious circle in which every step downward creates two steps outward. Engineers in 1906 already understood this mechanism; the French experience of twenty years earlier lay before their eyes — thousands of de Lesseps’s workers digging while Culebra filled up again.
There is a temptation to tell the story of the Panama Canal as the story of medicine’s triumph over the tropics. And that thread is real, not invented. William Gorgas, the project’s sanitary officer, launched an unprecedented campaign: draining swamps, eliminating standing water, spreading oil films over puddles, mosquito netting, quarantine measures. By the time full-scale construction began, yellow fever had been virtually defeated and malaria mortality had dropped sharply. Without this, no canal — neither a lock canal nor a sea-level one — could have been built. The French lost, by various estimates, between twenty and twenty-two thousand workers, and a substantial share of those deaths came from tropical diseases.
But — and this is crucial — Gorgas’s sanitary victory did not solve the engineering problems. It made construction possible, preserved the workforce, allowed plans to be made years ahead rather than only until the next epidemic. But the landslides at Culebra were not afraid of quinine. The floods of the Chagres did not respond to drainage ditches. Gorgas gave the Americans time and people; what to do with that time and those people was for the engineers to decide. What fascinates me is this double nature of Panama’s construction: at once a medical feat and a geological war, with neither victory guaranteeing the other.
Now for the most interesting part. Was a sea-level canal physically impossible? The honest answer is no. To declare it impossible is to ignore decades of later debate. In the 1930s, 1940s, and 1960s, various commissions and research groups returned to the idea of a lockless canal through Panama. In the 1960s, they even considered using nuclear explosions to excavate the earth — a project that, fortunately, remained on paper, but one that showed the engineering community regarded the task as fundamentally solvable with sufficient resources.
But “fundamentally solvable” and “solvable in 1906” are different things. Let’s carefully unfold the alternative. Suppose Roosevelt had listened to the board’s majority. What then?
First: the volume of excavation at Culebra would have increased severalfold. Instead of cutting through the ridge to the level of Gatun Lake — 26 meters above the sea — they would have had to dig down to zero. Every additional meter of depth meant exponential growth in landslide problems, which were already swallowing hundreds of thousands of cubic meters of earth that had been removed. The logistics of hauling away the spoil, already running at full capacity, would have required expansion on a multiple scale.
Second: the Chagres. Without Gatun Dam, the river would have had to be diverted into a separate channel or contained by a system of dams and interceptor canals along the entire route. Either solution would have been a separate engineering project, comparable in scale to the canal itself.
Third: time. The canal opened in 1914, ten years after American work began. A sea-level canal, by estimates at the time, would have required almost twice as long: figures ranged from fifteen and a half to twenty years. That meant opening no earlier than the mid-1920s, and allowing for landslides and floods — perhaps later. World War I began in 1914. A canal not ready by then meant an American fleet unable to move ships rapidly between the Atlantic and Pacific. The strategic consequences are hard to overstate.
Fourth: money. The lock canal cost about 375 million dollars at the prices of the day — a sum that had already caused scandals in Congress. A sea-level canal, by the same calculations, would have cost about 563 million dollars — almost twice as much. Political will is a resource no less finite than dynamite.
Had construction dragged on into the late 1910s or early 1920s, the consequences would have extended far beyond engineering. The Panama Canal reshaped global trade routes: the journey from New York to San Francisco shrank from more than 24,000 kilometers around Cape Horn to roughly 9,700. Every year of delay was a year in which cargo still traveled the old route, coal stations in southern South America remained in demand, and the Strait of Magellan retained its strategic importance.
The American presence in Panama would have looked different too. The Canal Zone, controlled by the United States, was created in 1903 under the Hay–Bunau-Varilla Treaty. The longer the construction lasted, the longer Panama bore the costs of the American presence without receiving the economic benefits of a functioning canal. The political tensions that were already mounting for decades and eventually led to the transfer of the canal to Panama in 1999 might have sharpened earlier and more brutally.
And there is one more turn that is rarely discussed: the lock canal created a limit on ship size. The Panamax standard emerged — the maximum dimensions of a ship capable of passing through the locks. For decades, that standard shaped the design of merchant and naval vessels around the world. A sea-level canal would not have created such a limit — or would have created another one, determined by the width and depth of the trench. The global merchant fleet might have developed along a different trajectory. Still, in 2016, new, expanded locks opened — the Neopanamax standard — and the history of limits continued, just with different numbers.
I’ll admit it: the deeper I go into this fork in the road, the less certain I am that only one of the options was right. Stevens and Roosevelt chose a route that could be completed with the technologies of 1906 — steam shovels, dynamite, railway trains, and the determination of fifty thousand workers. A sea-level canal might have been the better canal — for navigation, for the future, for ships of any size. But it required technologies that did not yet exist, patience Congress did not possess, and time that world politics would not allow.
Engineering maximalism is a seductive thing. Make it perfect, once and for all, without compromise. The Suez Canal — a straight sea-level trench — seemed proof that it could be done. But Suez runs through a flat sandy desert: no rivers, no rain, no landslides. Panama is an entirely different planet: a mountain ridge, clay soils, a river with the temperament of a berserker, and a climate in which water is not scarce but excessive. To transfer the Suez experience onto Panama’s geography is like building a house in a swamp from blueprints proven in the desert.
The history of megaprojects is full of such forks: the moment when you must choose between a beautiful solution and one that works. Sometimes they coincide. In Panama in 1906 — they did not. Roosevelt bet on what worked. The canal opened in August 1914, when war was already raging in Europe, and the first ships passed across Gatun Lake beneath the rumble of distant guns. Meanwhile, landslides in the Culebra Cut continued: one of them, in September 1915, closed the canal for almost seven months, and others came later. They were not defeated; people learned to live with them, widening the cut, reinforcing the slopes, accepting as a given that the mountain would keep crawling. A lockless canal would probably have turned this chronic problem into a catastrophic one — but it is impossible to say for certain, because the experiment was never carried out.