Wind

The Train

The way home from Kitty Hawk took days — the sailboat across the sound, then the long rails north and west, Norfolk falling away, the Virginia hills, Ohio flattening out the windows. Tradition puts what follows in a railway carriage somewhere on that line, and whether or not the carriage is right, the time is: there was nothing to do on the way home in August of 1901 but sit with what had just happened. Somewhere in there, Wilbur Wright told his brother that man would not fly for a thousand years.

He wasn't being poetic. He was being arithmetical.

The brother was Orville, four years younger, the other half of the Wright Cycle Company of Dayton, Ohio. Between them they had no degrees, no funding, no patron and no standing in any field but bicycles. Neither had finished a diploma. What they had was a shop that made its own frames, a business quiet enough in autumn to leave the back room free, and a habit of not believing things they hadn't measured.

They had come to it the way most people came to it in the 1890s, which is to say through a death. Otto Lilienthal had spent years throwing himself off a hill in Germany in gliders of his own design, and had published the numbers he took from it — the first real aerodynamic tables anyone had. In 1896 a gust stalled him at fifty feet and broke his back. The news reached Dayton, and it turned an idle childhood interest into work. Lilienthal had shown a man could fly a curved surface. He had also shown what it cost to be wrong about one.

So the brothers built to his numbers. In 1900 they built a glider to his numbers and it came up short. In 1901 they built the largest glider anyone had ever attempted — twenty-two feet of wing, some three hundred square feet of surface — and hauled it to the Outer Banks, and it came up short in the same direction by about the same amount.

Not catastrophically. Badly. It produced roughly a third of the lift the tables said it should. It wallowed. It needed an absurd angle to hold the air at all. Once it slid sideways out of the sky and dropped Wilbur into the sand hard enough to blacken his eye and bruise his nose.

The weeks around it were miserable in ways that would have finished less stubborn people. They arrived in July into a mosquito plague that has become legendary in the literature — Orville wrote home that the insects came in a cloud thick enough to dim the sun, and the brothers ended up lying under blankets in July heat, sweating, bitten through the wool. Storms took the tent. The sand got into everything.

They also had company, which turned out to be instructive. Octave Chanute — the elder statesman of American aeronautics, a retired railroad engineer who had become the field's clearing house and the Wrights' most valuable correspondent — came down to observe, and sent two other experimenters with him. Edward Huffaker brought a glider built of paper tubes. The Outer Banks damp got into the tubes and the machine came apart before it ever flew. George Spratt, a young doctor with a sharp instinct for aerodynamic questions, arrived and became a lifelong correspondent.

So for a few weeks in 1901 the Kill Devil Hills camp held about as much of the world's practical flying expertise as existed anywhere on earth. And none of it could explain why the machine wouldn't lift.

That is a specific and useful kind of despair. It is one thing to fail alone. It is another to fail in front of the best-informed people in your field and watch them fail to account for it either.

What nobody in that camp said out loud — because it was very nearly unsayable — was this. Lilienthal's tables were not one source among several. They were the closest thing aviation had to physics. Chanute used them. Samuel Langley, Secretary of the Smithsonian Institution, was using them to spend the largest research grant ever awarded an American scientist. Every serious experimenter in the world worked from those numbers, and had for a decade, and the numbers had been measured by the one man who had actually flown.

To say the machine was wrong was ordinary. Machines were wrong all the time. To say the tables were wrong was to say that the entire discipline had been building on a bad foundation for ten years without noticing — that Lilienthal had been wrong, that Chanute had been wrong, that the Smithsonian was spending a fortune on wrong, and that the two people who had worked it out repaired bicycles in Ohio.

They broke camp. The machine was crated, the tent came down, and they went back to Dayton having spent their own money to establish that the thing could not be done.

Wilbur reached the number on the train, and he reached it by elimination. Not the season. Not the site. Not the workmanship, and not the span. Every candidate had been struck out, and what was left was the possibility that the knowledge did not yet exist — that somebody would have to build it from nothing, and that it would take longer than any life had room for.

He gave it a number, the way he gave everything a number.

A thousand years.

He was wrong by a margin that is difficult to overstate. Within about six weeks he had arrived at something considerably more useful and much more arrogant: that the tables were wrong, that nobody in the world actually knew the numbers, and that he and his brother were going to have to go and find out for themselves.

The Shop

They found out in the back room of a bicycle store at 1127 West Third Street in Dayton, Ohio, which turns out to be exactly the right place.

The Wright Cycle Company was a real business, not a hobby. The brothers sold other people's bicycles and built their own — the Van Cleve at $65, the St. Clair at $42.50 — and the shop that supported that was a precision metal shop. There was a lathe. There was a drill press. There was brazing equipment for frame joints, and there were truing stands where a wheel could be brought round to within a thread's width by adjusting the tension in thirty-six spokes.

All of it was driven by a single one-cylinder gas engine that Orville had built himself, turning an overhead line shaft, throwing power down to each machine by leather belt. When they eventually needed a wind tunnel, they didn't need a power plant. They already had one running.

They had a mechanic too. Charlie Taylor had come to work for them that summer at eighteen dollars a week, a machinist who could hold a tolerance by feel and who would, two years later, build them an aluminum engine in six weeks because nobody would sell them one light enough.

And they had the bicycle itself, which mattered more than any of the tools.

Everyone else trying to fly in 1901 was trying to build a machine that would stay upright by itself. Stability was the goal — a craft so inherently balanced that the operator could largely leave it alone. It was an entirely reasonable idea and it was the wrong one. The Wrights, who had spent a decade around a vehicle that falls over the instant you stop working it, had a different instinct. A bicycle is unstable. That's fine. You don't solve it with geometry; you solve it with a rider. What an aircraft needed wasn't balance. It needed control.

Which is to say: they came at the hardest engineering problem of the age with no degrees, no laboratory, no institutional funding and no patron — and with one genuinely superior idea that they got from selling bicycles.

They had one other advantage, and it's easy to miss because it looks like a personality flaw. The Wright brothers argued. Constantly, loudly, for hours, in a small shop, about everything. Charlie Taylor found it unnerving at first and then stopped noticing. Orville later described a pattern that ought to be taught in every engineering school: they would take opposite sides of a question and go at it until, somewhere in the third or fourth hour, each of them would find himself arguing the position he had started out attacking — and only then would they feel they understood the problem well enough to decide.

That is not a temperament. That is a method. Two people, no hierarchy, no reputation at stake with anyone outside the room, systematically trying to demolish each other's reasoning. It is roughly the opposite of how a well-funded institution with a distinguished secretary at the top of it arrives at conclusions, and over the next three months it was going to matter enormously.

For comparison, Langley had received $50,000 from the War Department's Board of Ordnance and Fortification, plus roughly $20,000 more from the Smithsonian. It was the largest sum ever committed to a scientific project by the United States government at the time. The Wright brothers would spend, across the entire arc of the thing, well under a thousand dollars of bicycle money.

A Third Wheel

The first experiment was almost comically crude.

They took a bicycle wheel, laid it flat, and mounted it horizontally on the front of a bicycle so it could spin freely like a turntable. On the rim they fixed two small surfaces standing upright: a flat plate on one side, and on the other a curved airfoil shaped to Lilienthal's own numbers. The two were positioned so that if Lilienthal's tables were correct, the forces would cancel and the wheel would sit still.

Then somebody rode it, hard, down West Third Street.

The wheel turned. It shouldn't have. Whatever balance of lift against drag the published data promised, moving air in Dayton, Ohio disagreed with it. It was not a precise result — there was a rider's body in the airflow, an uneven street, a wind that came and went — and the brothers knew it proved nothing except the one thing that mattered. Something in the accepted science was off, and it was off by enough to see from a bicycle.

That was the permission slip. In October they stopped riding and started building.

The Box

The first tunnel was a throwaway — an old starch box with one end open and a fan pointed into it, put together in an afternoon to see whether the idea would work at all. It did, badly, and it taught them what to build next.

The real one was a wooden box roughly six feet long with a square cross-section sixteen inches on a side, mounted on a stand in the back room, with a viewing window of glass set into the top so you could watch what was happening inside. At one end sat a two-bladed fan belted to the overhead line shaft, driven by that same one-cylinder engine. It pushed air down the length of the box at twenty-five, maybe thirty miles an hour — a steady, indoor, repeatable wind, in a country where the wind is never any of those things.

The engineering that went into nothing happening is the part people skip. Raw air from a fan is a rotating, gusting mess, and a mess would have produced numbers as useless as the ones they were trying to replace. So they fitted straightening vanes to comb the swirl out of the flow. They leveled the whole apparatus with a spirit level, because a tunnel out of true would bias every reading in the same direction and they would never see it. They tested it, adjusted it, and tested it again until the wind coming past the model was as close to uniform and honest as a bicycle shop could make it.

Wilbur later described what they were after with characteristic flatness: they wanted to know what a surface did, and they wanted to know it in a way that would still be true tomorrow.

The Balance

The trap that caught everybody before them was this. To measure lift, the obvious move is to put a wing in moving air and measure the force on it in pounds. But converting that force into a usable coefficient — a number you could design a wing with — required knowing exactly how much pressure a given wind exerts on a given area. That value was called Smeaton's coefficient, it had been in use since the eighteenth century, and it appeared in every calculation anyone made.

Nobody actually knew it. It was a number everyone had inherited.

Which meant that any absolute measurement the Wrights made would be contaminated by the same uncertainty that was already poisoning everyone else's. They would have spent a winter producing a fresh set of wrong tables.

So they didn't measure force. They measured comparison.

The instruments they built — assembled out of hacksaw blades and bicycle spoke wire, because that is what a bicycle shop has lying around — were balances in the true sense. A model airfoil was mounted on one side. On the other was a reference surface of known area, sitting in the same airstream, at the same instant, feeling the same wind. The apparatus was free to rotate, and it would settle at whatever angle made the two forces equal. You didn't read a scale. You read a pointer against a marked arc.

Think about what that eliminates. Both surfaces are in the same air, so whatever the air is actually doing — however fast the fan is really turning, whatever the true value of Smeaton's coefficient is, whatever the air pressure and temperature in a Dayton back room in November happen to be — it acts on both sides equally, and it cancels out of the result. What survives is a pure ratio: how this shape performs relative to a known shape. Clean. Reproducible. Free of every assumption they had reason to distrust.

The second instrument was harder and cleverer. Lift is a single force in a single direction; drag has to be separated out from it while the model is being pushed around by both at once. Their drift balance was arranged so that the airfoil's resistance was weighed against the lift it was producing at that same moment — so the pointer gave them not two separate mystery numbers but the ratio between them, which is the only quantity that actually tells you whether a wing is any good. A shape that lifts enormously while dragging enormously is useless. The ratio is the wing.

Both instruments read out the same way: a pointer swinging against a scribed arc, settled at the angle where the forces balanced. No spring to calibrate, no scale to trust, no constant to look up. You opened the glass lid, changed the model, closed it, waited for the pointer to steady, and wrote down a number that meant something.

With those two readings you can characterize a wing completely.

This is why the Wright brothers get the credit rather than any of the dozen serious people working the same problem. They didn't invent a better wing that autumn. They invented a way of knowing — an instrument designed around the specific admission that they did not trust the constants, and therefore built to not need them.

Two Hundred Wings

Then they did the tedious part, which took two months and is the reason it worked.

They didn't measure force. They measured comparison.

The models were made from strips of twenty-gauge sheet steel, a thirty-second of an inch thick, cut and hammered and filed and soldered into shape by hand, each one small enough to sit in a sixteen-inch box. The count is disputed — a hundred models, perhaps two hundred — made across September to December of 1901.

Squares. Rectangles. Ellipses. Wings with the curve deepest at the front, at the middle, at the back. Wings with barely any curve at all and wings arched like a drawn bow. Single surfaces, then pairs stacked as biplanes at varying gaps to see how much each stole from the other. Long thin wings and short broad ones.

Out of that survey they pulled roughly thirty-eight shapes for systematic testing, and here they did the thing that separates an experiment from a fishing trip: they changed one variable at a time. Same wind, same mount, same procedure, same everything — vary the camber alone, or the aspect ratio alone, or the angle alone — and run each surface through some forty-five different angles of attack, writing every reading into a notebook.

The division of labor was the obvious one and it held for two months. One brother worked the tunnel and called the readings; the other sat with the notebook and wrote them down, then they swapped so that neither man's eye became the instrument. When a number looked wrong they ran it again. When two runs disagreed they ran a third. They were not gathering data to support a hypothesis — they had no hypothesis they trusted, which is precisely what made the exercise honest.

It was cold. The shop was heated by a stove, and the tunnel needed the fan running, which meant a November draft moving steadily through the room they were standing in. Their sister Katharine, keeping the house on Hawthorn Street for the three of them and their father the bishop, reported that the boys had essentially disappeared into the back of the store. Orville wrote that they had become so absorbed they had nearly forgotten the bicycle business existed — awkward, since the bicycle business was the only thing paying for any of it, and the selling season was coming.

There is no drama in this stretch of the story and that is the point. It is eight or ten weeks of two men in overcoats in an unheated back room, cutting steel the thickness of a fingernail into little curved shapes, putting them in a box, watching a pointer, and writing the number down. Nobody was watching. Nothing was at stake except being right.

By December they had the first reliable table of lift and drag coefficients ever produced. Not the best of the era. The first that was actually right.

The Number

The twist — and they clearly did not expect it — was that Otto Lilienthal's data was, broadly speaking, fine. The man had done careful work. His coefficients described his own airfoil section reasonably well, and the Wrights ended up saying so in public.

The error was upstream of him, in the constant everybody multiplied his numbers by. Smeaton's coefficient had been carried forward as 0.005. The Wrights' comparative measurements said it was closer to 0.0033.

That is roughly a forty percent overstatement of how much force moving air actually delivers. Every wing designed anywhere in the world for the better part of a century had been sized against a number that flattered it. This is why the Wrights' 1900 and 1901 gliders underperformed: not because their construction was poor or their airfoil was wrong, but because the arithmetic told them a given wing area would carry far more than it could. They had been building to a promise the air had never made.

Two other findings came out of the same notebooks, and both went straight into the next machine.

The first was aspect ratio. Long, narrow wings dramatically outperform short, broad ones of the same area — an effect nobody had quantified. Their 1901 glider had a wing about three times as long as it was deep. The 1902 machine would be more than six.

The second was camber. The efficient sections were thin, with the deepest point of the curve well forward, and much shallower than the pronounced arc Lilienthal had favored. They also caught something genuinely dangerous that nobody had documented: on a deeply cambered surface, the center of pressure can reverse direction as the angle of attack changes — it walks forward as you'd expect, and then at a certain point it turns around and runs backward, which will pitch a glider into the ground. It had almost certainly been killing people. The Wrights found it in a wooden box with a glass lid and designed around it.

What It Bought

They built the 1902 glider from their own tables and took it to Kill Devil Hills in September.

It was a different order of machine. Long thin wings, shallow camber, a fixed vertical tail that they discovered was causing a nasty spin and rebuilt — reportedly at Orville's suggestion, over a sleepless night — into a movable rudder linked to the wing-warping. That linkage is the third axis. With it, the 1902 glider became the first aircraft in history that a pilot could control in pitch, roll and yaw together.

They made seven hundred glides that autumn by the low count, a thousand by the high. They took turns. They got good at it — long, controlled, banking flights that they could put down where they intended, over and over, until flying stopped being an event and became a skill.

Everything after that is engineering, and it went the way engineering goes when you have real numbers. They needed an engine nobody would sell them light enough, so Charlie Taylor built one in the shop in about six weeks, out of aluminum, without a formal drawing.

Then they went looking for propeller theory and discovered, to their genuine astonishment, that there wasn't any. Marine engineers had been making ship's screws for the better part of a century entirely by trial, rule of thumb and inherited practice, and none of it transferred to a thin blade moving through air. So the brothers had the argument again — this one apparently ran for weeks and got loud — and came out the other side with the insight that a propeller is not a screw at all. It is a wing, traveling in a spiral. Which meant the tunnel data they already had described it.

They designed a pair from their own tables and got somewhere around seventy percent efficiency on the first attempt. That is a figure a modern designer would sign off on. They arrived at it in a bicycle shop, from first principles, because they had spent a winter establishing what the air actually does.

On October 7 and again on December 8 of 1903, Samuel Langley's Aerodrome — the machine with the government money — was catapulted off a houseboat on the Potomac and fell into the river both times. Newspapers had a wonderful time with it.

Nine days after the second one, on a cold Thursday morning at Kill Devil Hills, Orville Wright flew a hundred and twenty feet in twelve seconds.

Plenty of capable people were building flying machines in 1901, several of them better funded, better connected and better educated than two bicycle mechanics from Dayton. What none of them did was stop — come home from a failed summer, look at the most respected data in the field, and check the numbers nobody had checked. They didn't out-imagine anyone. They out-measured everyone.

The flight was a consequence.