Brace

Stake a young tree crooked, or just let the wind lean it, and given time it straightens itself back up. That much is old news to any gardener. What's new is how far a tree can do this without any of the senses you'd assume it needs. A September 2026 study put bent poplar saplings on a clinostat — a platform that turns slowly on its own axis inside a sphere lit equally from every direction at once — specifically to take gravity's fixed "down" and the sun's fixed "over there" both off the table. The saplings still curved themselves back toward straight. Whatever was doing the correcting, it wasn't reading the world. It was reading the plant's own shape.

That sense is called proprioception — the same word used for how you know where your own elbow is with your eyes shut — and in a stem it isn't metaphorical. The tool it drives is real carpentry: a wedge of tension wood, laid down along whichever flank of the bend is convex (the outer, longer side), built from cellulose-rich fibers that shrink lengthwise as they mature. A shortening flank on one side of an unshortened one is exactly what bends a rod, and it bends this one the right way: back toward straight, because the convex flank is by definition the one that needs to get shorter to undo the curve. The paper's own comparison is to a muscle, and it names the sharper version of that too — because the flank that's under tension depends only on which way the current bend goes, a stem bent one way and then the other grows tension wood first on one side and then the other, the wood equivalent of a biceps and triceps working in opposition. Except a muscle relaxes; a wedge of tension wood, once grown, stays.

Bend the sapling below with the slider — standing in for the wind, a stake, or anything else that could put a curve in a real one, since none of that matters to a mechanism that only ever asks "how curved am I right now?" Then press “let a week pass”. Each week, one flank grows a little reaction wood (drawn as a wedge along the trunk, and in the small cross-section on the right), and the lean eases back by a fraction — not a fixed number of degrees, so a sharp bend corrects faster in absolute terms but takes more weeks to finish, the same shape a real decay curve has. Bend it the other way at any point and the flank that had been idle takes its turn.

Which flank, and which wood, depends on the sapling. Poplar is an angiosperm and grows tension wood on the convex (outer) flank — the amber wedge, cellulose-rich fibers that shrink as they mature and pull that side shorter. Pine is a conifer, and conifers run the opposite trick: compression wood on the concave (inner) flank — the rust wedge, tracheids with thicker, more rounded walls and lignin enriched in a different monolignol than normal wood, laid down by spatially localized laccase activity, that lengthen and push that side longer instead. Different flank, different chemistry, opposite verb — pull versus push — but the same outcome: the bend eases back toward straight either way. Switch the sapling below and bend it the same amount; the wedge grows on the other side.

Honest gaps. A real bent stem carries curvature all along its length, graded and continuous; this room hinges the whole bend at one point and puts the growth in one lens-shaped zone for the sake of a legible drawing, not because the biology is lumpy that way. A week here is one click; the paper's own weeks are weeks — xylem drying and lignifying on a real clock this slider skips past entirely. The 30%-a-week correction and the wedge's own size cap are chosen to be visible in a handful of clicks, not measured off the paper — and the pine sapling reuses those exact same numbers rather than a rate measured for conifers, since the room's point is the flank flipping, not a claim that poplar and pine correct at the same speed in the wild. The color split is stylized too: real compression wood is darker than normal wood, but nowhere near as saturated as this room's rust, and only shows clearly in a fresh, well-lit cross-section, not to the naked eye on a standing tree.

No date, no plant.js, no rng() — every shape here is pure geometry from the slider and two counters kept only for this visit, the same footing /plumb, /roots and /veins already stand on. Unlike /plumb, which answers to gravity, and /tip, which answers to light, nothing drawn here answers to either — that absence is the entire point of the clinostat this room is built from.

Sources: Caulus, A. et al., “Proprioception drives tension wood formation for autotropic straightening and postural control in trees,” New Phytologist (2026), INRAE & Université Clermont Auvergne, on young poplars; see also the INRAE research summary. For tension and compression wood generally, Groover, A., “Gravitropisms and reaction woods of forest trees — evolution, functions and mechanisms,” New Phytologist 211(3):790–802, 2016. For the pine sapling's own mechanism, Hiraide, H. et al., “Localised laccase activity modulates distribution of lignin polymers in gymnosperm compression wood,” New Phytologist 230(6):2186–2199, 2021 — this room's third paper from the same journal.