Field notes on things that run themselves
A Thread of Water That Isn’t Allowed to Snap
Nothing inside a tree pumps. There is no muscle, no valve, no chamber squeezing anywhere between a redwood’s roots and the needles two hundred feet above them — yet water gets there, against gravity, every day, for centuries. In 1894 an Irish botanist named Henry Dixon and a physicist named John Joly proposed an explanation strange enough that it took decades to be widely accepted: the water isn’t pushed up from below. It’s pulled from the top, by evaporation, through a single unbroken thread of liquid running root to leaf — held together by nothing but water’s own grip on itself, stretched tighter than a steel cable, one stray bubble from snapping.
The pull starts at the top, in a leaf’s mesophyll cells, where water retreats into the curved corners of cell-wall pores only nanometers across as it evaporates into drier air. That curve does the work: the tighter it bends, the harder it pulls, by simple surface tension — a pore just ten nanometers wide can generate a negative pressure beyond fourteen megapascals, well over a hundred times atmospheric pressure, running in reverse. The pull doesn’t stay at the leaf: water molecules cling to each other through hydrogen bonds — the same stickiness that lets a water strider stand on a pond — so the tug transmits backward through the whole column, down through stem and root into the soil. The tree becomes one long, unbroken straw, sipping only at the top.
Everyone already owns a mental model for pulling water upward — a straw — and it’s the wrong one here. A straw, or an old suction pump, drops the pressure at the top toward a vacuum and lets atmospheric pressure shove water up from below; that caps out at about ten and a third meters, the most the atmosphere alone can push. A redwood exceeds that tenfold, because it isn’t sucking in the straw sense — it’s pulling on a continuous cohesive thread from above, a force never capped by how hard the atmosphere can push. Using a pressure chamber, researchers have measured real, sustained tensions inside living trees running from about negative half a megapascal to negative eight, deeper in drought — genuine negative pressure, well past where a vacuum pump gives up.
The tension never rests, and it’s never far from failure. Water’s theoretical tensile strength is enormous, on the order of 130 megapascals, but real sap carries dissolved gas and touches imperfect surfaces, either of which can nucleate a bubble long before that ceiling. When one does, the tension collapses instantly: liquid flashes to vapor, the column snaps in a real, sometimes audible click, and that conduit stops carrying water for good. This publication keeps finding the same bargain — a candle’s flame (No. 1), a power grid holding one note by matching supply to demand instant to instant (No. 30), a mitochondrion’s proton gradient with no reservoir, just a rate (No. 36) — and a tree pays it the same way: never staking everything on one line. Xylem runs as thousands of parallel, threadlike conduits instead of one wide pipe, so a snapped one goes quiet while its neighbors keep pulling.
What happens next is one of the field’s genuinely open questions. A snapped conduit is supposed to be finished — refilling it while the vessel next door stays under deep tension seems to violate the same physics that broke it. Yet some studies report exactly that: living cells around the xylem apparently pump a vessel back full while neighbors stay tensioned, a process researchers call “novel refilling.” Other studies, using different instruments on the same trees, see far less of it, or argue the signal is partly a measurement artifact. This publication says so plainly rather than picking a side: the pull itself is real and about a hundred and thirty years settled. Exactly how, or how often, a tree repairs its own break is not.
In 2008, engineers built a version of this system with no plant in it. A microfluidic “synthetic tree” — a porous hydrogel disc standing in for the leaf, a channel for the trunk, a reservoir for the soil — evaporated water from its own nanoscale pores exactly as a leaf does, pulling the water beneath it into sustained tensions around negative one megapascal, with nothing biological doing any of the work. No cells, no evolution, just geometry and water’s own cohesion, running the identical trick — the same proof this publication keeps returning to: No. 36’s chloroplasts, soaked in acid in total darkness, made ATP with no light at all. Strip a living system down to bare physics, and if the physics alone reproduces the behavior, you’ve found the actual mechanism.
The mechanism itself sets the ceiling. As a tree gets taller, gravity and friction add to the tension its leaves must generate just to keep pulling, until the leaves’ own water stress starts limiting the growth and photosynthesis the tree needs to grow taller — it chokes on its own plumbing before anything structural gives way. A widely cited 2004 study modeling this in California’s coast redwoods put that ceiling at roughly 122 to 130 meters. The tallest tree ever accurately measured, a redwood named Hyperion, stands about 116 meters — taller than the Statue of Liberty, closing in on that predicted limit, and, as far as anyone can tell, still climbing: one evaporating needle at a time, on a thread of water it can never once let go slack.
One loop I’m watching
Next: why the rainbow you’re looking at right now isn’t the same rainbow the person standing next to you sees — a shape with no fixed location and no material substance at all, remade instant to instant out of whichever raindrops happen to sit at exactly the right angle between the sun and your own two eyes, and gone the moment you turn your head.
Tip: the ← and → arrow keys move between issues.
Get the next issue the moment it’s live — subscribe by email or follow by RSS.
New to The Standing Wave? Start here →