The microscopic hairs that let corals breathe 94 and why heat makes the pump fail

TITLE: The Tiny Pump That Lets Corals Breathe  and Why Heat Makes It Suffocate Them STANDFIRST: Reef corals beat microscopic hairs to spin millimetre-scale vortices that pull oxygen through still water; lab work shows the pump speeds up as seas warm but demand rises faster, thickening a hypoxic layer and collapsing above about 37C  a suffocation pathway that may precede bleaching, though it has so far been seen only in the dark.

Corals that cannot move beat microscopic hairs to pull oxygen toward their tissue, spinning corkscrew vortices about a millimetre across. In lab experiments on the reef-builder Porites lutea, that ventilator ran faster as water warmed  then began recirculating oxygen-depleted water and, above about 37C, lost coordination and collapsed, leaving tissue anoxic and killing the fragments by 39-41C in darkness. The finding reframes why a few degrees of warming can be lethal even before corals turn white, and why sheltered, low-flow reefs facing increasingly oxygen-poor seas may be most vulnerable. Whether the collapse actually triggers bleaching in sunlit reefs, or is simply a downstream symptom of heat stress, has not yet been tested with light, reactive oxygen and algal counts measured together.

A sessile animal inside still water

A reef coral is a thin veneer of animal tissue over limestone. It cannot swim to seek oxygen. At its surface, friction slows seawater to a near-standstill, forming a thin, slow-moving water layer  the concentration boundary layer  where oxygen and waste must cross. By diffusion alone, oxygen takes about four minutes to travel one millimetre, too slow to meet demand, especially at night when the symbiotic algae living inside coral tissue stop photosynthesizing and stop making oxygen.

Corals solve this with coral cilia  microscopic hair-like organelles, 10 to 15 microns long, that cover the surface and beat in coordinated waves. The beating is organized in hexagonal units that streamline motion, sweeping mucus and particles and, crucially, stirring the boundary layer.

Until about 2014, those hairs were thought to be mere mucus brooms. Work led by Orr Shapiro, then at the Massachusetts Institute of Technology and the Weizmann Institute and now at the Volcani Institute in Israel, showed they do something more active: they generate vortical flows that locally enhance mass transfer. Subsequent work refined the geometry and showed that skeletal architecture and tissue are functionally integrated, and that morphology tunes the flow  branched forms redistribute food, whip-like black corals ventilate and even drive internal flow through gastrovascular channels.

The new study, published in May 2026 in Science Advances by Cesar Pacherres and Michael Khl at the University of Copenhagen with colleagues at the Leibniz Institute for Baltic Sea Research Warnemnde and the University of Melbourne, closes the loop to heat tolerance. Using high-speed imaging in a microscope chamber and a technique called SensPIV  imaging that tracks both water flow and oxygen at once using fluorescent, oxygen-sensitive nanoparticles  the team mapped flow and oxygen fields while warming corals in darkness.

At the control temperature of 27C, cilia beat at about 20.8 to 21.6 Hz and remained stable for 900 minutes. The vortices they spun were 1,000 to 1,200 microns long and 400 to 900 microns high, with flow speeds of 400 to 600 microns per second. That advection dominated diffusion by more than a hundred-fold  a Sherwood number, the ratio of advective to diffusive transport, of about 137 at 27C, peaking near 366 at 32C  delivering roughly 400 times more oxygen than diffusion alone. By day the same mixing can redistribute photosynthetically produced, hyperoxic water; by night it is the main oxygen supply when the algae are off.

Warming speeds the pump  and demand speeds faster

Warming does two opposing things: it lowers how much oxygen seawater can hold, and it raises how fast corals burn oxygen.

From 27C to 37C, cilia beat faster by about 1 Hz per degree, reaching about 30.5 Hz at 37C, a temperature coefficient  Q10, the factor by which a biological rate rises with 10C warming  of about 1.51. The vertical component of the flow rose to about 601 microns per second at 35C, with a Q10 of about 1.48, before falling to about 400 at 37C and near zero at 39C as vortices disappeared.

Respiration, however, accelerates with a Q10 of 2 to 3  doubling to tripling with 10C warming, compared with about 50% faster for the cilia. That mismatch is the core of the paradox.

Initially the faster beating helped. Net advective oxygen flux toward tissue, measured about 400 microns above the surface, rose from about 0.8 nanomoles per square centimetre per second at 27C to about 1.4 at 32C. But because respiration was rising faster, the vortices increasingly recirculated water already stripped of oxygen within the boundary layer. The low-oxygen zone thickened to about 1,000 microns at 27C and about 1,200 microns at 32 to 35C, exposing tissue to transient hypoxia despite stronger stirring. Flux then fell to about 1.1 at 35C, 0.7 at 37C and -0.04 at 39C, while the Sherwood number fell to about 128 at 37C and about 3 at 39C, where diffusion again dominated.

A mechanistic model reproduced the pattern. At 27C, active cilia raised tissue oxygen by about 26% compared with diffusion alone, with tissue well above 50 micromolar in both cases. At 35C, diffusion alone left tissue near anoxia  below 5 micromolar  with about 15% of tissue volume anoxic, while active cilia still maintained oxygenation, boosting tissue oxygen by about 65%. Beyond about 36 to 37C, anoxic regions formed regardless of whether cilia were active. The model also captured a feedback: enhanced ventilation sustains higher respiration, which depletes local oxygen faster.

Collapse above about 37C  and why that number is not fixed

Above about 37C, the coordinated beating broke down, particularly at the tips and middles of cilia, as shown by rising variability in beating frequency. Vortices shrank to 600 to 700 microns high at 37C and were absent by 39C. The boundary layer thinned to about 1,000 microns at 37C and about 600 microns at 39C, but remained hypoxic. Diffusive flux rose slightly to about 0.015 at 39C as the thinner layer allowed more exchange, but after the buffering capacity was lost. By 41C cilia showed little to no movement and mortality reached 100% in the 24-hour dark experiments.

That collapse temperature coincided with field thresholds that commonly trigger bleaching and mortality, and the authors propose ciliary beating  specifically loss of rhythmicity  as a potential early indicator before visible paling, a suggestion noted by Rachel Alderdice, a marine biologist at the University of Konstanz who studies coral stress biomarkers and was not involved in the work.

But the threshold is fragile. In the model, varying the half-saturation constant for oxygen uptake, Km, from 0.1 to 5 micromolar  a range spanning intrinsic mitochondrial values to apparent tissue values that integrate transport limitations  shifts the predicted anoxia threshold by about 3C. Varying respiration Q10 from 2 to 3 shifts it by about 4C, from about 38C to about 34C. The experiments themselves were acute, one-way ramps from 27C to 41C at 1C per hour for 24 hours in darkness, under low laminar flow of 1 mm per second typical of sheltered reefs, with three biological replicates of a single species, Porites lutea, without recovery, without light, and without direct measurement of algal expulsion or photophysiology alongside cilia. Daytime hyperoxia dynamics were modeled, not measured. Higher turbulence thins the boundary layer but does not eliminate the near-surface microenvironment shaped by cilia, and the one-way ramp may accentuate abruptness compared with natural fluctuating regimes  though the authors note rapid multi-degree spikes over hours do occur during severe marine heatwaves, with surface water near Florida recently approaching 38C.

Why cilia beat faster in warmer water remains unresolved  possibilities include reduced seawater viscosity making water thinner, nerve-net sensing of heat or oxygen triggering molecular motors, or both.

What this reframes about bleaching  and what it does not replace

The established, dominant explanation for coral bleaching remains heat- and light-driven oxidative stress originating in the symbiont photosystem. Under elevated temperature and high light, damage to the D1 protein in photosystem II, limitation of the Calvin-cycle carbon-fixation sink, and damage to thylakoid membranes cause excess electrons to reduce oxygen to reactive oxygen species  singlet oxygen, superoxide, hydrogen peroxide  overwhelming symbiont and host antioxidant systems including superoxide dismutase and ascorbate peroxidase. Those species diffuse into host tissue and trigger host innate-immune and cell-death pathways involving NF-kB, nitric oxide, apoptosis and autophagy, leading to expulsion of the algae and the white skeleton showing through.

The ciliary work adds a parallel, oxygen-centric pathway rather than replacing that one. Microenvironmental hypoxia by night and hyperoxia by day can destabilize tissue homeostasis  constraining ion regulation and protein synthesis, impeding proton removal from the calcifying space and pH regulation for skeleton deposition, and increasing reactive oxygen species  potentially triggering bleaching or direct mortality without bleaching. Corals can die from heat without bleaching, and patchy bleaching that correlates with low-flow, low-oxygen microenvironments, as seen in work associated with David Suggett at King Abdullah University of Science and Technology, fits the microenvironment view.

Viewed together, bleaching is multi-causal and context-dependent, with reactive oxygen, nitric oxide, immune dysregulation, nutrient imbalance and symbiont identity all modulating thresholds. The ciliary ventilation mechanism is best understood as an additive, low-flow amplifier  most relevant in sheltered lagoons, at night, and where ambient oxygen is already low  rather than a new universal cause. The competing interpretation remains credible: that ciliary failure and micro-scale hypoxia observed in dark Porites lutea may be an epiphenomenon or downstream consequence of generalized tissue stress and energy failure, not the primary driver that precedes bleaching. That view is supported by the model sensitivity and the lack of light experiments.

The synergy with ocean deoxygenation sharpens the stakes. In one global survey using autonomous sensors at 32 representative reef sites, hypoxia  defined as at or below about 40% air saturation, about 2 milligrams per litre or about 61 micromoles per litre  was reported as already pervasive, with the authors reporting 84% of reefs experiencing weak to moderate hypoxia and 13% severe hypoxia. Under warming scenarios, models project more hypoxic observations by 2100, ranging from about 13 to 42% more under low-emissions pathways to about 97 to 287% more under high-emissions pathways. Separately, short-term assays on Acropora found deoxygenated seawater at about 2 milligrams per litre lowered the thermal bleaching threshold by up to 1C by bleaching index or 0.4C by photosynthetic efficiency, with RNA sequencing showing activation of hypoxia-inducible factor responses alongside distinct deoxygenation signatures in photo-reception, redox status and immunity.

For now, the practical implication is not a field-ready forecast. Tests that would resolve causation include light-exposed experiments with simultaneous reactive oxygen, photophysiology and symbiont counts; bleaching without heat, for example with red light, to see if ciliary failure occurs without thermal stress; replication across species, morphologies and thermally adapted populations; and field validation linking rhythmicity loss to in situ bleaching during the next marine heatwave. The authors note conservation angles such as identifying resilient species or morphologies and flow-aware restoration, but emphasize that emissions reductions remain essential and that thresholds are not yet generalizable.

What the vortices do establish is that corals are not passive recipients of whatever water passes by. They actively ventilate a millimetre-thin envelope that determines whether oxygen arrives. Warming makes them ventilate harder, until the same stirring traps the very depletion it tries to relieve  and then the ventilator itself stalls.