Record-correction · Fluid dynamics

The Air That Got There First

Ask the internet how a wing makes lift and you will mostly be told a tidy story that is wrong. Here is the air, told the truth.

The story goes like this. A wing's top is curved, so the air crossing it has farther to travel than the air underneath. To arrive at the trailing edge at the same moment as its partner below, the top air must speed up. Faster air has lower pressure (Bernoulli), so the top is sucked upward, and the wing flies. You have seen it in museums, in textbooks, in the first page of search results.

The hidden hinge of that story is the phrase "at the same moment." It assumes two parcels of air that split at the front of the wing must rejoin at the back — the equal-transit-time assumption. Nobody ever says why they would. And in fact they don't. The air going over the top reaches the trailing edge first, with time to spare, and never meets its partner again.

Below is a real wing in a real flow — the exact, closed-form solution for an idealised airfoil, the same one the verifier checks. Press Release dye and watch a straight ribbon of smoke split at the nose. Equal transit time predicts the two halves arrive together. Watch what the air actually does.

The dye is not animated to a script — each speck is carried by the exact velocity of the flow at the point where it sits, step by step. The ribbon tears because the flow tears it. Drag the angle of attack: more angle, more circulation, more lift, and a bigger head start for the top. (Turn the angle negative and the whole thing flips — the lift points down, and now the under-side air wins the race.)

The two parcels never reunite. Whatever lifts the wing, it cannot be a rejoining that never happens.

So what does hold the wing up?

Lift is real, and — this is the part the myth gets backwards by trying to be simple — the two famous explanations of it are not rivals. They are the same flow, counted two ways.

• Newton's way. The wing throws a great sheet of air downward. Push air down, and by the third law the air pushes the wing up. That downward deflection — the downwash — is the lift, plainly.
• Bernoulli's way. The same flow is faster over the top (you can read the peak speeds off the panel above) and slower below; faster flow means lower pressure, so the surface pressures, added up, push the wing up by exactly the same amount.

The verifier computes the lift both ways — once from the circulation (Kutta–Joukowski, L = ρ·U·Γ) and once by integrating the Bernoulli pressure over the whole surface — and they agree to better than a thousandth of a percent. Newton and Bernoulli are two languages for one fact.

What's actually false in the schoolbook version is narrower, and more interesting, than "Bernoulli is wrong." It's the reason given for why the top is faster. The top is not faster because it has a longer path to cover in a fixed time. It is faster because the air has circulation around the wing — a net tendency to flow forward under the belly and back over the top — and that circulation is set by a single, hard geometric fact: air cannot whip around the infinitely sharp trailing edge. It would need infinite speed to do so, so nature picks the one amount of circulation that lets the flow leave the sharp edge smoothly. That is the Kutta condition, and it is the real author of the speed difference — and of the lift.

What this model leaves out (named, not buried)

This is potential flow: two-dimensional, inviscid (no viscosity), incompressible, steady. It is a real and load-bearing model — it is how the lift of a wing is actually computed to first order — but it is a model:

• Viscosity is what sets up the circulation in the first place. A truly frictionless fluid would never start the flow turning; in real air, a "starting vortex" is shed from the trailing edge as the wing accelerates, and its equal-and-opposite reaction is the circulation you see here. The Kutta condition is the fingerprint that viscosity leaves on an otherwise inviscid picture.
• No stall. Crank the angle of attack to 14° here and the lift keeps rising forever. A real wing's boundary layer separates somewhere around 15°, the flow detaches, and lift collapses. This model cannot see that — it has no boundary layer to lose.
• Two dimensions only. A real, finite wing also sheds vortices off its tips, adding "induced" drag the 2-D model has none of.

None of these rescue the equal-transit story. In the real, viscous flow the top air still reaches the trailing edge well before the bottom air — that is exactly what the smoke-tunnel experiments show. The myth isn't a simplification of the truth. It's a different, false claim that happens to reach a true conclusion (there is lift) by a route that doesn't exist.

Flight was never a debt the air had to repay by the back edge. It is the wing leaning on a turning river of air, and the river — over the top — getting there first.