The Seed in the Substrate

The plant’s move along the time axis — a whole coherence-match wrapped, switched off, and handed forward

A seed is a plant that has stopped. Inside a millimetre of coat sits a complete miniature organism — a root tip, a stem, one or two leaves, and the shoot apical meristem that will build every organ the plant ever makes — packed against a bank of stored carbon and then dried to under a tenth its weight in water, at which point it stops respiring, stops transcribing, stops repairing, and stops being measurably alive. It can stay that way for a season, or for two thousand years. Add water and the whole thing lights back up.

That is the whole chapter. The seed is the plant’s move along the time axis: an entire nested coherence-match wrapped until its dynamics stop, held with maintenance switched off, and re-ignited across a boundary when the medium returns. The rest of the Plants section climbs space — from the chloroplast at \mum to the forest at 10^4 m. The seed climbs the other axis, and it does so by the framework’s one load-bearing memory law: persistence is boundary-wrapping depth, read in time as ring-down lifetime. The difference the substrate makes here is that the law stops being an intuition about aromatic stacks and becomes an ordering over an already-catalogued natural experiment of tens of thousands of species — one that biology has measured, filed, and never read as a single axis.

And the seed pays the framework back. The hardest thing the substrate picture asks of a reader is to accept that a medium can hold enormous organised energy and a complete pattern while presenting, to every instrument, as nothing. The seed is that claim you can hold in your hand.

The Anatomy: Three Wraps and a Bank

Strip a seed and you find the same architecture the cellular walk has met at every scale, drawn once more at the scale of a package.

• The embryo — radicle (root tip), hypocotyl, one or two cotyledons, and the shoot apical meristem. This is the cell-modon stack intact: membranes, nuclei, mitochondria (as promitochondria, stalled mid-assembly), plastids (as proplastids), cytoskeleton. Nothing is dismantled. Everything is stopped.
• The bank — endosperm, perisperm, or fattened cotyledons, holding starch, oil, and storage protein. The energy that will pay for re-ignition, banked before the power was cut.
• The coat (testa, often with a pericarp outside it) — a maternally-built, frequently lignified, sometimes water-impermeable shell. This is a third wrap outside the plasma membrane and the cell wall, and the plant cell already carries two.

The framework reads the seed as the cell modon with one more wrap and its maintenance switched off. The chloroplast earned its third membrane because the input direction demanded a steeper coherence isolation; the seed earns its extra coat for the same structural reason at a different job — not isolating a splitter from chemistry but isolating a whole stopped organism from a medium that would otherwise restart it at the wrong moment.

The coat is a wrap in the paper’s strict sense, because it is gated, not merely closed. Physically dormant seeds — most legumes, Malvaceae, Convolvulaceae — are sealed against liquid water entirely, and open through a single anatomically specialised water gap: the lens in a Fabaceae testa, the chalazal plug elsewhere, a millimetre-scale valve that pops on a specific cue (fire, a hot–cold cycle, freeze–thaw, gut passage) and not otherwise. That is the regulated jet at the seed’s boundary — the same architecture as the nuclear pore, the plasmodesma, the crista junction and the bordered pit — run at the outermost wrap a plant ever builds, and gating the one substance that turns the whole capsule back on.

The Glass: The Tun State, Made Ordinary

The cells chapter introduced a third state between coherent-alive and decoherent-dead — the tardigrade’s tun: metabolism below 0.01\% of baseline, water expelled, the interior vitrified into a biological glass by trehalose and intrinsically disordered CAHS proteins, the nested modon stack held frozen and re-ignitable. It called the tardigrade “biology’s proof that when maintenance stops, sufficient wrapping is enough.”

It is a proof, and it is the wrong lead witness. The tardigrade is the extreme case; the seed is the common one — the ordinary reproductive mode of the great majority of the roughly 350,000 seed plants, executed on every roadside, at a planetary scale, every year. And the chemistry is a startlingly close convergence on the same solution:

	Tardigrade tun	Orthodox seed
Water removed to	\sim 1–5\%	\sim 2–10\%
Glass former (sugar)	trehalose	sucrose + raffinose-family oligosaccharides
Glass former (protein)	CAHS/SAHS — intrinsically disordered	LEA (late embryogenesis abundant) — intrinsically disordered
Nucleic-acid protection	Dsup on the DNA \pi-stack	chromatin compaction, HSPs, DNA-repair enzymes staged for imbibition
Metabolism	undetectable	undetectable
Documented hold	\sim 30 y frozen	\sim 2{,}000 y air-dry

Two lineages separated by a billion years of evolution, one an animal and one a plant, reached the same architecture: remove the medium, replace it with a hydrogen-bonded glass built from a non-reducing sugar and a disordered protein, and the whole nested boundary stack is immobilised in register. LEA_4 proteins from Arabidopsis have been shown to stabilise dry sucrose glasses directly, and tardigrade CAHS proteins vitrify like trehalose and act synergistically with it.¹ The framework’s reading is the one the stamp-lifetime chapter already stated and the epigenetics chapter already tabulated for chromatin: life has found many ways to wrap substrate energy into long-lasting topologies, and wraps more heavily when the state must last longer. Two kingdoms converging on one glass, when neither shares the chemistry that got them there, is what a substrate-set principle with chemistry-side implementations looks like from the outside.

Lay the seed’s own strategies on that one axis and the biology’s hierarchy of persistence falls out as a wrapping ladder, exactly as the epigenetic marks did:

Wrap	What it immobilises	Hold it buys
Hydrated embryo, no glass (recalcitrant)	nothing beyond the cell’s own membranes	weeks–months
Partial drying, incomplete glass (intermediate)	some boundaries, not all	months–a few years
Full cytoplasmic glass (orthodox)	every organelle boundary at once	years–decades
+ impermeable coat (physical dormancy)	the embryo against the medium itself	decades–centuries
+ cold, dry, dark burial or store	the capsule against its environment	centuries–millennia

Each rung adds a boundary layer between the coherence-match and the thing that would disturb it, and each rung buys roughly an order of magnitude. The documented tail is real: a date palm seed from Masada, radiocarbon-dated to \sim\!1{,}990 years, was germinated in 2005 and grew into the tree named Methuselah, with six more Judean Desert seeds germinated since; a directly-dated sacred-lotus fruit from Xipaozi germinated at \sim\!1{,}300 years.²

One frequently-cited record has to be handed back, and handing it back sharpens the claim. The \sim\!31{,}800-year Silene stenophylla from Siberian permafrost (Yashina et al., PNAS 2012) was not a germinated seed: whole fertile plants were raised by in vitro clonal micropropagation from the placental tissue of immature fruits recovered from ground-squirrel burrows at -7\,^\circC. That is a beautiful result and it belongs to a different rung — the wrap there was supplied externally, by never-thawed permafrost, not built by the organism. The framework should not borrow it: the whole content of the seed’s claim is that the capsule builds its own wrap and carries it into an environment that offers none. The date palm’s two millennia at ambient Dead Sea temperatures is the datum that actually tests the law.

The Wrap-Tension Signature: Where the Plant’s Pressure Ladder Ends

The Plants section has been reading pressures as wrap-tension signatures — the boundary gradient every modon keeps at its edge, in whatever currency the local chemistry supplies. Turgor runs at \sim\!+1 MPa, phloem at \sim\!+1 to +2 MPa, xylem at \sim\!-1 to -3 MPa with the cohesion ceiling near -3 MPa. The dry seed extends that ladder by two orders of magnitude, and it is the chapter’s one computed result.

A seed in equilibrium with air at relative humidity h sits at the water potential the Kelvin relation fixes, \psi = \frac{RT}{V_w}\,\ln h , with RT/V_w \approx 135 MPa at 20\,^\circC. An orthodox seed dried for storage to \sim\!15\% RH sits at \psi \approx -256 MPa. This is the steepest boundary gradient anywhere in a plant — some four orders of magnitude beyond the wet soil (\sim\!-0.03 MPa) it will imbibe from — and the framework reads it as exactly what the wrapping ladder predicts: the deepest wrap carries the steepest tension across it.

The striking part is what stays fixed. Ellis, Hong, Roberts & Tao found that the otherwise log-linear seed-viability relation between longevity and moisture content breaks at a low-moisture limit below which further drying buys nothing — and that this critical moisture content varies nearly threefold across species, from 2.0\% in groundnut through onion (3.5\%), sugar beet (4.5\%), barley (4.6\%), chickpea and wheat (5.3\%) to cowpea (5.6\%), tracking oil content. But the equilibrium relative humidity at those breaks varies only from 9.9\% (onion, sugar beet) to 11.5\% (wheat), mean 10.5\% (s.e. 0.2).³ Put through the Kelvin relation at 20\,^\circC, that band is

\psi_\text{crit} \;=\; -304 \;\text{MPa} \quad (\,-312\ \text{to}\ -292\ \text{MPa}\,),

a \pm 3\% spread about the mean against a \pm 41\% spread in the mass fraction — the potential is roughly twelve times tighter than the water content.

The Wrap-Tension Ladder — and What Is Invariant at Its End

The dry seed ends the plant's ladder of boundary pressures two decades past the xylem — and where drying stops buying life, the tension holds fixed while the water content scatters with chemistry.

push (positive pressure) pull (negative potential) chemistry's variable (water content)

The plant’s wrap-tension ladder, and what is invariant at its end. Left: the pressures the Plants section has read as boundary signatures, on a log scale — turgor and phloem pushing at \sim\!+1 MPa, xylem pulling to its \sim\!-3 MPa cohesion ceiling, and the dry orthodox seed two orders of magnitude below all of them. Right: the same seven species at the point where drying stops extending life. Their critical moisture contents scatter across \pm41\% of the mean, tracking oil content (grey); their critical water potentials, computed from the measured equilibrium humidities through \psi = (RT/V_w)\ln h, collapse onto a \pm3\% band at -304 MPa (gold). Chemistry sets how much water a given seed holds at a given tension; the substrate reading is that the boundary answers to the tension, not the mass fraction.

This is the ladder chapter’s division of labour in a place nobody was looking for it. Chemistry sets the boundary conditions — a lipid-rich seed simply holds less water at a given humidity, so its critical mass fraction is lower. The substrate supplies the principle — what determines whether the glass has formed and the wrap has closed is the tension across the boundary, which is why the invariant lives in \psi and the variance lives in \% water. The honest caveat is Ellis & Hong’s own: the limit is temperature-sensitive, so the band is not a universal constant but a well-defined isotherm, and the seven species are a small sample of the thousands whose longevity Kew has on file. The framework’s ask is that the sample be extended — the prediction below.

Orthodox and Recalcitrant: The Wrap Closed, or Never Closed

Seed biology’s own primary division is the sharpest test the framework’s wrapping law has ever been handed, because the division is about whether the wrap can close.

• Orthodox seeds tolerate drying to \sim\!5\% water, vitrify, and store for years to millennia. This is the great majority of species and essentially all of agriculture.
• Recalcitrant seeds die if dried below \sim\!20–30\% water. They never enter the glass. Avocado, mango, cacao, coconut, oak acorns, most mangroves, a large share of tropical rainforest trees.
• Intermediate seeds — coffee, papaya, some citrus — dry part way and then fail. About 2\% of Kew’s Seed Information Database.

Read through the framework these are not three kinds of seed. They are three depths of wrap, and the longevity ordering is the wrapping ordering, exactly as the epigenetic marks order from acetyl to heterochromatin. A recalcitrant seed is alive and leaking — a live coherence-match that must be continuously re-pumped, so its lifetime is a live ring-down measured in weeks. An orthodox seed is frozen-coherent — the wrap is closed, maintenance is off, and the ring-down is measured in decades. There is no third mechanism; there is one law and three positions on it.

And the distribution is the retrodiction. Recalcitrance is not scattered phylogenetically — it concentrates almost exactly where waiting is never required. Roughly 8\% of seed plants globally are desiccation-sensitive, but the figure climbs to \sim\!19\% of the flora of aseasonal tropical moist forest and \sim\!47\% of its trees and shrubs, and falls toward zero in deserts, grasslands, and boreal and temperate seasonal systems.⁴ The framework’s reading inverts the usual framing. Recalcitrance is not a failure to evolve desiccation tolerance; it is the absence of a boundary the environment ever asked for. Wrapping is expensive — LEA and oligosaccharide synthesis, a whole maturation-drying programme, a delayed and hazardous re-ignition — and a wet aseasonal forest never charges for skipping it. The rule is the sign rule in its plain form: name what the seed must survive, and its wrapping depth is fixed before you measure it.

The Latch, Kept Dry: Dormancy as a Coherence-Match in Time

Wrapping tells you how long a seed can wait. It says nothing about when it stops. That is dormancy, and dormancy is the latch chapter’s apparatus run in a new place — which is the connection the section was missing.

The seed occupies three distinct states, and biology names all three:

1. Dry and vitrified. The whole match wrapped, no metabolism. This is the tun state, and it maps onto the latch chapter’s flash: written once, held with the power off.
2. Imbibed but dormant. Water is back, the cell is running, and germination still does not happen — because the gene node is held closed. This is the latch proper: the live loop restored with the detent still set.
3. Germinating. The node slides open.

Between (2) and (3) sits an ABA/GA balance that biology has mapped in detail — abscisic acid holding the node closed, gibberellin sliding it open, with DOG1 (DELAY OF GERMINATION 1) setting how deep the detent is set in the first place. The framework adds nothing to that chemistry. What it adds is a reading of the trigger, and the reading has teeth because it explains the field’s long-standing difficulty with single-cue models.

A dormant seed does not test a threshold. It tests a trajectory. Physiological dormancy in temperate species is typically broken by cold-moist stratification — weeks at \sim\!4\,^\circC — and in many species by a sequence: warm, then cold, then warm again; or a specific number of chilling units accumulated before a rising temperature. Seeds buried in a soil seed bank undergo annual dormancy cycling, their germinable window opening and closing with the season while they stay in the ground for years. None of that is a scalar comparison. It is a match against a long vector in time — the seed carries a template of what its year should look like and fires on the overlap, the same coherence-match operation the framework reads at the Mediator’s nose, the aromatic pocket, and the cortical column, run here on a coordinate whose unit is a month. That is why a rising thermometer in November does not germinate a seed that would germinate at the same temperature in March: the scalar matches and the vector does not.

The dry state is not idle either, and this is the chapter’s most framework-native observation. Dry after-ripening — the slow, well-documented loss of dormancy that happens in a seed with no metabolism, over weeks to years, at rates set by storage moisture and temperature — is the latch’s own ring-down inside the glass. The seed is not merely waiting; it is counting time by leaking, and the leak rate is set by how tight the wrap is. Which raises an immediate consequence the framework insists on and the field keeps in separate literatures: after-ripening (useful, dormancy release) and ageing (damage, viability loss) are one leak read on two dials. Both run in the glass, both accelerate with moisture and temperature, and the framework says they must scale together with the same molecular-mobility parameter. Seed priming — hydrating to Phase II and re-drying — is the same coin spent deliberately: it buys faster, more uniform germination and it costs storage life, which is what partially unwrapping a frozen hold must do.

And the latch is written by the mother. The depth of dormancy a seed carries is set substantially by the temperature and photoperiod the maternal plant experienced during seed maturation, transduced through FLC and FT in the silique and read out at DOG1 — with a VEL3 histone-deacetylase complex establishing a maternal epigenetic state that controls progeny dormancy, and non-canonical RNA-directed DNA methylation contributing.⁵ This is precisely the transaction the epigenetics chapter flagged and declined to overclaim: a mark laid down in one generation, surviving the reset, and doing a job in the next. Plants are where it is least contested, and the seed is where it happens. The framework’s contribution is not to assert the channel — the biology already has it — but to hand it the chapter’s ordering law: whichever maternal marks reach furthest into the progeny should be the most heavily wrapped ones, methylation and compacted chromatin over acetylation, on the same axis the latch chapter drew for the marks themselves.

The Dynamic: Re-Ignition Across a Boundary

The user of this framework should expect the interesting physics to be at the boundary, and it is. Germination is triphasic, and the framework reads each phase as a distinct boundary operation.

Phase I — imbibition. Water rushes in, driven by the \sim\!-300 MPa matric potential against soil at \sim\!-0.03 MPa. This is pure physics; a dead seed imbibes as fast as a live one. It is the medium returning to a structure that was built to be re-wetted.

And here is the tell. A dry seed dropped into cold liquid water dies. Imbibitional chilling injury is one of agriculture’s oldest practical hazards: the membranes are in the gel phase, the boundary is crossed faster than it can re-form into the liquid-crystalline state, solutes pour out, and the cell is lost — the damage is done in the first two minutes, and it is abolished by letting the tissue take up water as vapour first, to 35–50\% moisture, before liquid ever arrives.⁶

The framework’s reading is its own decoherence claim, run as farm practice: a frozen coherence-match must be re-ignited slower than its boundaries can re-match, or it is destroyed at the crossing. The cells chapter already says that a nested boundary either carries coherent state across or scatters it, and that scattering at cell scale is the substrate definition of cell death. Imbibitional injury is that sentence with a stopwatch on it. What is lost is not information — the pattern is intact — but the ability of each wrap to re-match its neighbour while the medium arrives. This also makes the framework’s sharpest cross-kingdom prediction, below: if the law is the law, the tardigrade must show the same rate dependence on rehydration, and for the same reason.

Phase II — the plateau. Water uptake stalls; the machinery restarts. Two facts about the order of operations are exactly the framework’s:

• Translation before transcription. Seeds accumulate more than 10{,}000 stored “long-lived” mRNAs during maturation, and the first proteins of germination are translated from those, with de novo transcription not required for Phase I and early Phase II.⁷ The frozen coin is spent before any new coin is minted.
• Repair before run. ATP is essentially absent in the quiescent embryo; promitochondria must be structurally repaired and matured before oxidative phosphorylation can supply the germination budget, and DNA repair runs in parallel. The machine is re-matched before it is switched on.

Read against the ladder’s economy — the substrate sets the price list, chemistry sets the boundary conditions, and the breath is the coin — Phase II is the seed re-opening its accounts in the only order that works: repair the channels, spend the stored coin, and only then start minting.

Phase III — radicle emergence. Water uptake resumes, the endosperm cap is weakened enzymatically, and the radicle is pushed out by turgor — the plant cell’s wrap-tension signature doing mechanical work. Germination sensu stricto ends here; everything after is growth. The capsule has handed its match back to the live economy the Plants section spends five chapters describing.

The Seed at Both Poles

The seed sits at both poles of the ladder, and it splits them in a way no other member of the set does.

Lock — the individual. A vitrified seed is the most completely bound object biology builds. Every boundary in the nested stack is immobilised in register against every other, and the glass exists precisely to stop anything sliding relative to anything else. This is the lock pole taken to its limit: maximum binding, zero capacity to slide — which is exactly why it can hold, and exactly why nothing can happen in it. The health-is-the-capacity-to-slide reading is not violated but suspended; the seed has traded the capacity to slide for the ability to wait, and it buys the capacity back with water.

Anti-lock — the cohort. A population of seeds that all fired on the same trigger would be the seed-bank version of a metronomic heart, a market whose correlations have all gone to one, a stand of vessels sharing one cavitation threshold, or a forest wired into one over-connected web: efficient until the one shared trigger fires into a bad year, then a total recruitment failure. So genetically near-identical seeds from one mother are given different dormancy depths, and only a fraction germinates at each opportunity — a soil seed bank that spreads the cohort across years. Ecology has modelled this since Cohen (1966) and measured it in Sonoran Desert annuals, where germination fraction tracks environmental unpredictability across species and populations.⁸

So the seed realises both poles by level — the individual at the lock pole, the cohort at the anti-lock pole — and that is new to the paper’s set. The cortex slides by state, the eye and chloroplast split by subsystem, the symplast by callose gating, the vascular column by organ position, the meristem by developmental sequence, the mycorrhizal network by topology. In every one of those, both poles live inside one object. In the seed they live in different objects at different levels — one seed cannot hedge; only a cohort can. Bind the individual completely; spread the population deliberately.

Two disciplines, and the chapter should apply them to itself. First, this is a temporal anti-lock, and the paper already owns a temporal face of the gap — the periodical cicada’s primes. The seed bank is not a fourth face of the gap. The cicada is a genuine witness because its answer is an extremal one: whole-generation periods, gcd-minimised, 13 and 17 sitting exactly on the resonance floor. The seed bank reaches the pole by grading a continuous property — dormancy depth across a cohort — which is the same route the vascular system takes with cavitation thresholds and the forest takes with modularity. It is an application of the sign rule, not a fresh witness of the gap. Second, the bet-hedging ecology is textbook and sixty years old. What the framework adds is the unification — that the desert annual’s germination fraction, the heart’s variability, the trunk’s graded cavitation, and the forest’s modularity are one pole of one axis — and one sharpening the ecology does not currently make, which is that the load-bearing quantity should be the spread, not the mean.

Why the Seed Is How You See the Substrate

Now the analogy the chapter has been earning, stated as an analogy and fenced as one.

The substrate framework’s central and hardest claim is about the stealth vacuum: a medium of enormous organised energy, tiled at \xi \approx 97\;\mum, whose anti-phase breath cancels its own dipole so completely that the residual self-screens to \sim\!1\% within a single cell and a probe one envelope away sees nothing at all. The whole pattern is present. Every measurement returns empty. It expresses only where a boundary condition lets it.

A seed is that structure at a scale you can pick up. Its complete long vector is present — an organism’s worth of nested boundaries, every one intact. Its amplitude is zero: no respiration, no transcription, no signal any instrument reads as life. And it expresses only when a boundary condition changes — the medium returns and the whole pattern comes up at once. A full pattern held at zero amplitude, indistinguishable from inert, waiting on a boundary. That is the seed, and it is the sentence the substrate asks you to accept about empty space.

The framework should be explicit that this is pedagogy, not derivation. A seed’s silence is metabolic; the vacuum’s is interferometric; nothing about the one implies the other. But the paper has been claiming throughout that this shape of thing — a pattern kept whole by wrapping, silent because it is balanced or stopped, and recoverable across a boundary — is a substrate-level motif that recurs wherever the medium allows, and the seed is the motif’s most familiar instance. It is what the reach chapter called the frozen carrier rung, and the seed is a better witness for it than the tardigrade in one specific respect: the tardigrade’s tun is an emergency response to an environment that went wrong, while the seed’s hold is the designed, normal, planetary-scale reproductive mode of an entire kingdom, with a two-millennium documented datum at ambient temperature.

It also closes a loop the paper opened in cosmology, in the paper’s own words. The moraine crust is described as “the seed-bed of the round that follows,” and the surviving dense relics of the previous cycle as “the seeds, the bowling-pins for the structure that grows inside \mathcal{B}^0” (the universe that boils; cosmic web). The paper reached for the word before it had the chapter. The word is doing real work in both places, and it means the same thing in both: a wrapped remnant that carries structure intact across a discontinuity and organises what grows on the other side. The plant seed is that word’s literal case, the case where the wrap is measurable, the hold is dated, and the re-ignition can be watched on a windowsill.

Predictions and What Would Falsify

Six predictions extend the reading beyond the structural anchors.

1. The glass threshold is invariant in potential, not in water content. Extend Ellis et al.’s seven species across a much wider sample and a wide range of oil content: the critical moisture content should keep varying severalfold, while the equilibrium water potential at the break stays inside a narrow band near -300 MPa at 20\,^\circC (with a defined temperature dependence). Falsified if \psi_\text{crit} scatters as widely as the mass fraction does — which would say the boundary answers to composition rather than tension. This is a retrodiction on the seven species in hand and a genuine prediction on the rest; the Kew SID and the gene-bank literature supply the data.

2. Longevity orders on wrapping depth, and the ordering is callable before measurement. Across Kew’s database, storage life should sort by the completeness of the wrap — recalcitrant < intermediate < orthodox < orthodox with physical dormancy — and within orthodox species should track measured glass quality (molecular mobility, T_g, LEA and raffinose-family oligosaccharide content) more strongly than any single compositional variable such as seed mass, lipid fraction, or genome size. Falsified if longevity is fully predicted by composition with no residual on glass stability.

3. Desiccation sensitivity tracks whether the environment ever charges for waiting, not phylogeny. Recalcitrance should concentrate in aseasonal wet systems and vanish in seasonal and arid ones — already confirmed at the global scale, so a retrodiction there. The novel form is within-lineage: in clades containing both habits (Quercus, Acer, Coffea, the Fabaceae), the orthodox/recalcitrant split should track habitat seasonality rather than phylogenetic position, with independent losses of desiccation tolerance wherever a lineage entered an aseasonal wet forest. Falsified if the split is phylogenetically conserved and habitat-blind.

4. Re-ignition is rate-limited by boundary re-matching, across kingdoms. Imbibitional damage should scale with the rate of the boundary crossing rather than the total water taken up, and should be abolished by vapour-phase pre-humidification that lets each wrap re-match before bulk water arrives. That much is established for seeds and pollen. The framework’s prediction is that the tardigrade tun shows the same rate dependence on rehydration — controlled-humidity staged rehydration outperforming direct immersion, by the same margin and with the same temperature dependence — because the law is about the wrap, not the phylum. A tun that revives equally well from direct immersion at any rate would falsify the boundary-re-matching reading.

5. After-ripening and ageing are one leak on two dials. Dormancy release in the dry state and viability loss in the dry state should be governed by the same molecular-mobility parameter in the glass, so any treatment that changes one (storage moisture, temperature, oligosaccharide content, priming) should change the other by a proportional amount, with a single shared activation behaviour. The sharpest form: primed seeds should lose dormancy faster and lose viability faster in the same ratio. Falsified if the two rates can be decoupled — if a treatment accelerates after-ripening while leaving longevity untouched, the “one leak” reading is wrong and the two are separate chemistries.

6. The cohort’s spread of dormancy depth is the load-bearing quantity, not its mean. Bet-hedging theory and data currently centre on the mean germination fraction. The framework predicts that fitness in a variable environment tracks the within-cohort variance of dormancy depth independently of the mean, and that populations from more unpredictable environments carry wider spreads at matched means. A cohort engineered to a uniform trigger threshold at the optimal mean fraction should still underperform a spread one across a run of years. Falsified if fitness is fully explained by the mean with no benefit to the spread.

The picture is falsified if (a) the critical water potential scatters as widely as critical moisture content, (b) longevity carries no residual on glass quality once composition is controlled, (c) desiccation habit is habitat-blind within mixed clades, (d) rehydration rate does not matter for the tardigrade tun, (e) after-ripening and ageing rates decouple, or (f) germination-fraction spread confers nothing beyond its mean. It is supported, even partially, if any of the six orderings holds against existing data.

Honest Accounting

What is solid and not the framework’s: the anatomy, the glass, the LEA and CAHS convergence, the orthodox/recalcitrant/intermediate classification and its ecological distribution, the triphasic germination sequence, imbibitional injury, stored mRNAs, the maternal control of DOG1, and sixty years of bet-hedging ecology. Every fact in this chapter is somebody else’s measurement.

What the framework computes is one number: the critical water potential \psi_\text{crit} = -304\ (-312\ \text{to}\ -292) MPa, obtained by pushing Ellis et al.’s measured equilibrium humidities through the Kelvin relation. That is arithmetic on published data, not a derivation from the backbone — but the point of it is a framework claim: that the invariant sits in the tension and the variance sits in the chemistry, which is the ladder chapter’s division of labour showing up where nobody filed it.

What the framework adds is organisation, and it should be judged as organisation. That the orthodox/recalcitrant division is one wrapping axis rather than three seed types; that the seed’s three states map onto the latch chapter’s flash / held-node / open-node; that stratification is a long-vector match in time rather than a threshold; that after-ripening and ageing are one leak; that imbibitional injury is the decoherence claim with a stopwatch; that the individual and the cohort sit at opposite poles of one ladder. These are re-descriptions. They are faithful to the framework and they reorganise known biology, and predictions 1, 4, 5, and 6 are the places where the reorganisation is forced to pay — a number, a cross-kingdom rate law, a coupling the field currently keeps apart, and a statistic the ecology currently averages away.

What is not the framework’s to claim: the Silene permafrost result, handed back above; any implication that a seed’s metabolic silence is evidence about the vacuum’s interferometric silence, which is pedagogy and is flagged as such; and any suggestion that the substrate explains why a particular species is recalcitrant, which is ecology and chemistry doing their own work.

Putting the Section in Context

The Plants section climbed space: chloroplast at \mum, symplast at tissue scale, the Calvin loop at the chemistry rung, xylem and phloem at organism scale, tropisms pooling the channels without a brain, and the mycorrhizal network at forest scale. Every rung was a different size of the same architecture. The seed is the section’s other axis: the whole stack folded into a capsule, wrapped until its dynamics stop, and handed forward through a stretch of time the running plant could not have survived.

It is the cell modon with one more wrap and its maintenance switched off. Its glass is the tardigrade’s tun made ordinary, and its wrapping ladder — hydrated, dried, vitrified, coated, buried — is the stamp-lifetime law and the chromatin hierarchy written a third time, at a scale you can see. Its wrap-tension signature, \sim\!-300 MPa, ends the plant’s pressure ladder two orders of magnitude past the xylem, and holds nearly constant in potential across species whose water content varies threefold. Its dormancy is the latch, set by the mother and read as a match against the shape of a year. Its germination is a boundary crossing that fails if taken too fast, spends stored coin before minting new, and repairs its channels before it runs them. And it stands at both of the ladder’s poles at once, bound absolutely as an individual and deliberately spread as a cohort — the first member of the paper’s both-poles set to split them across two levels rather than inside one object.

The paper’s own cosmology already calls the moraine “the seed-bed” and its relics “the seeds.” That was the right word, borrowed early. Here it is at home: a wrapped remnant carrying a whole pattern intact across a gap that would have destroyed the running version of it, and organising what grows on the far side. The plant did not invent the move. It found the one the substrate already offers — wrap the match deeply enough and it does not have to stay awake to survive.

Footnotes

1. Boothby, T.C. et al., “Tardigrades use intrinsically disordered proteins to survive desiccation,” Mol. Cell 65, 975–984 (2017); Hibshman, J.D. & Goldstein, B. et al., “Trehalose and tardigrade CAHS proteins work synergistically to promote desiccation tolerance,” Commun. Biol. 5, 1046 (2022); Bremer, A. et al., “Stabilization of dry sucrose glasses by four LEA_4 proteins from Arabidopsis thaliana,” Biomolecules 11, 615 (2021); Hincha, D.K. & Thalhammer, A., “LEA proteins: IDPs with versatile functions in cellular dehydration tolerance,” Biochem. Soc. Trans. 40, 1000–1003 (2012).

2. Sallon, S. et al., “Germination, genetics, and growth of an ancient date seed,” Science 320, 1464 (2008); Sallon, S. et al., “Origins and insights into the historic Judean date palm based on genetic analysis of germinated ancient seeds and morphometric studies,” Sci. Adv. 6, eaax0384 (2020); Shen-Miller, J. et al., “Long-living lotus: germination and soil \gamma-irradiation of centuries-old fruits,” Am. J. Bot. 89, 236–247 (2002).

3. Ellis, R.H., Hong, T.D., Roberts, E.H. & Tao, K.-L., “Low moisture content limits to relations between seed longevity and moisture,” Ann. Bot. 65, 493–504 (1990); Ellis, R.H. & Hong, T.D., “Temperature sensitivity of the low-moisture-content limit to negative seed longevity–moisture content relationships in hermetic storage,” Ann. Bot. 97, 785–791 (2006).

4. Tweddle, J.C., Dickie, J.B., Baskin, C.C. & Baskin, J.M., “Ecological aspects of seed desiccation sensitivity,” J. Ecol. 91, 294–304 (2003); Wyse, S.V. & Dickie, J.B., “Predicting the global incidence of seed desiccation sensitivity,” J. Ecol. 105, 1082–1093 (2017); Daws, M.I. et al., Ann. Bot. 97, 667–674 (2006).

5. Chiang, G.C.K. et al., “DOG1 expression is predicted by the seed-maturation environment and contributes to geographical variation in germination in Arabidopsis,” Mol. Ecol. 20, 3336–3349 (2011); Chen, M. et al., “Maternal temperature history activates Flowering Locus T in fruits to control progeny dormancy according to time of year,” PNAS 111, 18787–18792 (2014); Kendall, S.L. et al., Plant Cell 23, 2568–2580 (2011); Zhao, M. et al., “A VEL3 histone deacetylase complex establishes a maternal epigenetic state controlling progeny seed dormancy,” Nat. Commun. 14, 2220 (2023); Footitt, S. et al., New Phytol. 225, 2035–2047 (2020).

6. Crowe, J.H., Hoekstra, F.A. & Crowe, L.M., “Membrane phase transitions are responsible for imbibitional damage in dry pollen,” PNAS 86, 520–523 (1989); Hoekstra, F.A., Crowe, J.H. & Crowe, L.M., “Effect of imbibition temperature on leakage and germination of desiccation-tolerant pollen,” Plant Physiol. 99, 1234–1239 (1992); Bewley, J.D. et al., Seeds: Physiology of Development, Germination and Dormancy, 3rd ed. (Springer, 2013), ch. 4.

7. Sano, N. et al., “Staying alive: molecular aspects of seed longevity,” Plant Cell Physiol. 57, 660–674 (2016); Bai, B. et al., “Extensive translational regulation during seed germination revealed by polysomal profiling,” New Phytol. 214, 233–244 (2017); Sajeev, N. et al., “Seeds: a unique system to study translational regulation,” Trends Plant Sci. 24, 487–495 (2019).

8. Cohen, D., “Optimizing reproduction in a randomly varying environment,” J. Theor. Biol. 12, 119–129 (1966); Venable, D.L., “Bet hedging in a guild of desert annuals,” Ecology 88, 1086–1090 (2007); Clauss, M.J. & Venable, D.L., “Seed germination in desert annuals: an empirical test of adaptive bet hedging,” Am. Nat. 155, 168–186 (2000).