Flight Mechanics & Origins

CONTESTED Trees-down versus ground-up

For most of a century, the origin of bird flight was framed as a choice between two stories. Trees-down: a small climbing animal leaps, glides, and gradually gains control until gliding becomes flying. Ground-up: a fast bipedal runner flaps, gains lift, and gradually leaves the ground. Whole careers were spent on which one won.

It's worth knowing that framing, because you'll meet it everywhere — but it's better read as historiography than as a live either/or. The modern synthesis treats the two as not mutually exclusive, and shifts the emphasis onto flapping-based precursors: what a partly-developed wing was doing, in a real animal, before it could carry that animal through the air. That question turns out to be answerable in ways the old debate never was, because living birds still do the intermediate things.

BRACKETED WAIR — wing-assisted incline running

Ken Dial's work with chukar partridges is the cleanest experimental handle anyone has on incipient wings. Chukar chicks that cannot yet fly still flap while running up steep surfaces, and the flapping isn't producing lift in the usual sense — it generates force directed into the substrate, pressing the feet down and improving hindlimb traction. The bird effectively climbs a slope it could not otherwise climb, using wings that are nowhere near flight-capable. As the wings develop, the same behaviour grades continuously into flight.

That's the value of it: a functional, measurable use for a partial wing, in a living animal, at every stage of development — no gap to explain away, no useless intermediate.

The tier matters here. WAIR is directly observed in chukars. Applying it to non-avian theropods is a behavioural model carried across from living birds, which is a bracketed inference, not a fossil observation. It shows that a partial wing can pay for itself. It does not show that any particular extinct animal used it that way.

DIRECT The semilunate carpal — a flight stroke that predates flight

The wrist that lets a bird fold its wing against its body, and that guides the sweeping arc of the flight stroke, depends on a half-moon-shaped carpal bone. That bone originated in maniraptoran theropods well before anything in the lineage could fly. So did the furcula — the fused clavicles that act as a spring and strut in the modern flight apparatus.

This is exaptation in its clearest form: the machinery of the flight stroke was assembled, in skeletons, for other purposes, and was standing ready when flight became possible. Both are direct skeletal evidence — bones in hand, not inferences about soft tissue.

DIRECT Microraptor and the four-winged plan

The morphology is not in question: long asymmetric flight feathers on the forelimbs and on the hindlimbs, preserved on the animal. Four aerofoils on a dromaeosaur.

What the animal did with them is a second, separate claim. Wind-tunnel work and aerodynamic modelling support gliding and parachuting capability with the hindwings acting as aerofoils. That inference is bracketed: it comes from physical and computational models of a reconstructed animal, not from watching one. So the split is clean — four wings, directly preserved; gliding, well supported but inferred.

Microraptor is a reminder that the paravian radiation was not a single approach march toward the modern bird wing. It was a set of experiments, and several of them look nothing like where the lineage ended up. Sapeornis, a long-winged basal avialan above Archaeopteryx grade, is another of them.

DIRECT The membranous wing: a flight experiment that failed

Two scansoriopterygids — Yi qi and Ambopteryx longibrachium — did something no other dinosaur is known to have done. Instead of building a wing out of feathers, they built one out of skin: a membrane stretched from the arm and supported by a long rod-like styliform element, structurally closer to a bat than to a bird. In Ambopteryx both the membrane and the styliform element are preserved, which is what confirmed that Yi qi was not a one-off oddity or a misreading.

Aerodynamic analysis finds these animals were poor gliders and incapable of powered flight. That's the interesting part. This is a second, independent architecture for getting airborne inside the dinosaur lineage — attempted, built in bone and skin, and gone. Flight in theropods wasn't a single inevitable outcome; it was tried more than once, and at least one of the attempts didn't work.

DIRECT Flightlessness, over and over

Flight is expensive, and lineages give it up whenever the bill stops being worth paying. After the K-Pg boundary, flightlessness re-evolved many separate times — the ratites are the familiar example, and they are not the only one. This is worth holding alongside the origin story: the capacity that took tens of millions of years to assemble gets abandoned readily, and the fact that a bird can't fly tells you nothing about whether its ancestors could.

INFERRED The Anchiornis moult signal

One recent result is worth flagging without overselling. Moult patterns described in Anchiornis have been reported as resembling those of secondarily flightless birds (Kiat and colleagues, 2025). Read that carefully: a resemblance in moult strategy is what's reported. It is not a demonstration that Anchiornis was secondarily flightless, and treating it as one would be reading past the evidence. We note it because it's the kind of observation that could matter a great deal if it holds up, and because the honest version of the sentence is narrower than the exciting one.