# DISCOVERIES — matter_life.md (D026–D050)

*Materials, molecular systems, cells, and the physical Earth. Each item names a blocker of a specific
kind, and prefers the blocker "two measurements that were never taken on the same sample" over
"the instrument is expensive".*

### D026 — Is fault friction set by contact area or by asperity aging?
**Question.** In water-saturated crustal gouge, does the velocity dependence of friction (a−b) track the measured real area of contact, or the time asperities have spent in contact?
**Why unresolved.** Concept blocker: two laws fit the same steady-state friction coefficient but predict opposite transients — the contact-area law makes friction depend on the instantaneous area, the aging law on log(time in contact), and at seismic slip rates the two diverge by how much strength is lost per decade of slip speed. Instrument blocker: friction is measured in rock-mechanics presses, real contact area in tribology rigs on dry, low-pressure contacts; the two have essentially never been measured on the same gouge at the same pressure, pore-fluid chemistry and sliding rate.
**Decisive measurement.** Transparent sapphire anvil plus optical interference to count load-bearing junctions, simultaneous with friction, on the same calcite or quartz gouge at hydrothermal pressure and controlled pore chemistry, through a step change in sliding rate.
**Counts as an answer if.** The measured a−b is predicted by the measured area alone, or by contact time alone, without a fitted mixture of the two.
**Why it matters.** Earthquake nucleation models, afterslip decay and induced-seismicity forecasts all use a−b as an input that is currently inferred from friction alone, with the microscopic mechanism left undetermined.

### D027 — Is the asthenosphere creeping by dislocation or diffusion creep?
**Question.** At the megapascal-scale shear stresses of the asthenosphere, does olivine deform with a stress exponent near 3 (dislocation creep) or near 1 (diffusion creep)?
**Why unresolved.** Assumption blocker: laboratory flow laws are fitted at strain rates ten-odd orders of magnitude above the mantle's, and the extrapolation is exponential in the assumed stress exponent, so the answer is decided by a parameter the experiment cannot reach. The two mechanisms never coexist cleanly in a single run, so no lab test isolates which one dominates at low stress. Seismology measures attenuation at exactly mantle strain amplitudes but reports Q, not mechanism.
**Decisive measurement.** On one olivine aggregate at a fixed pressure and temperature, measure seismic-frequency attenuation (forced-oscillation, ~1 Hz) and steady-state strain rate in the same apparatus, then test whether the Q predicted by the dislocation-creep law with n≈3 matches; repeat at several stresses to see the exponent emerge.
**Counts as an answer if.** Attenuation and creep rate are jointly fit by a single microphysical model, fixing n at mantle stress, or the joint fit fails and a third mechanism must be named.
**Why it matters.** Mantle viscosity, post-glacial rebound timescales and the longevity of plate boundaries all hang on an extrapolation that has never been anchored at the relevant stress.

### D028 — Does glass toughness track medium-range order?
**Question.** Does the fracture toughness of a silicate glass vary monotonically with a single measurable parameter of its medium-range order, at fixed composition?
**Why unresolved.** Concept blocker: "medium-range order" has no agreed scalar — ring-size distribution, first sharp diffraction peak position, and Raman D1/D2 ratios are used interchangeably and do not co-vary. Access blocker: toughness is measured on centimetre specimens, order on powders or fibres, and no one has built a compositionally fixed series in which order alone varies.
**Decisive measurement.** A series of one glass composition quenched from different fictive temperatures and densified at different pressures, each measured for K_IC by double-cantilever and for neutron structure factor plus ring statistics; test whether toughness collapses onto one ordering parameter across the series, and whether the correlation survives when fictive temperature is the only variable.
**Counts as an answer if.** A single ordering parameter predicts K_IC within scatter across all preparations, or the series shows toughness varying while every ordering metric stays fixed.
**Why it matters.** A structural predictor of toughness would let glass be designed rather than screened, and would settle whether toughness is a network property or a flaw-population property.

### D029 — At a crack tip, does hydrogen reach the decohesion threshold?
**Question.** At an arrested crack in steel, is the local hydrogen concentration at failure the value predicted for atomic decohesion, or far below it?
**Why unresolved.** Instrument blocker: hydrogen is invisible to most microanalysis and mobile even at room temperature, so a nm-scale composition map of a loaded crack tip requires cryogenic atom-probe tomography on a specimen that must also be held under load — the two requirements are physically incompatible in current holders. Meanwhile the plasticity-based alternative predicts failure at concentrations well under the decohesion threshold, so both mechanisms fit bulk embrittlement data.
**Decisive measurement.** Cryo-transfer atom probe on a crack tip arrested under load in a charged specimen, plus nanoindentation of the same volume to get local flow stress; compare the measured hydrogen excess with the threshold from first-principles decohesion for that grain boundary.
**Counts as an answer if.** The measured tip concentration sits above the computed decohesion threshold (decohesion) or demonstrably below it while the local yield stress has dropped (hydrogen-enhanced plasticity), at a stated confidence.
**Why it matters.** Hydrogen embrittlement limits pipeline and storage steels; a mechanism decides whether the fix is traps, coatings, or microstructure.

### D030 — Why does strength stop rising below ten nanometres?
**Question.** Below roughly ten nanometres grain size, does the strength of a metal stop increasing because grain boundaries begin to slide?
**Why unresolved.** Instrument blocker: sliding displacement at an individual boundary under load is a nanometre measurement inside a bulk specimen, while in-situ transmission electron microscopy offers the resolution only in foils whose stress state differs from the bulk. The competing explanation — that grains rotate and coalesce, or that diffusion-accommodated sliding takes over — predicts the same plateau.
**Decisive measurement.** Strain-rate-jump tests on bulk samples spanning grain sizes from 100 nm down to a few nm, extracting the strain-rate sensitivity m: dislocation glide gives m near 0.02, boundary sliding with diffusional accommodation gives m near 0.5–1. Pair each with synchrotron peak-shape analysis to count boundaries that actually carried sliding.
**Counts as an answer if.** m switches at the same grain size where the strength plateau begins, and the boundary population that slid is identified.
**Why it matters.** Nanocrystalline alloys are chosen for strength; knowing the mechanism sets a hard ceiling on grain refinement and tells you which solute slows the loss.

### D031 — Do pits nucleate at predictable grain boundaries?
**Question.** Does passivity breakdown on a stainless steel initiate preferentially at boundaries of a specific coincidence-site character, identifiable from their measured segregation?
**Why unresolved.** Access blocker: pitting is a rare stochastic event at the nanometre scale, while the standard tool for boundary character, electron backscatter diffraction, maps micrometres — no one has registered a nucleated pit back onto the exact boundary that hosted it at nm resolution. Assumption blocker: chromium depletion is assumed to be the controlling variable, but molybdenum and nitrogen segregate differently and no experiment has measured local composition at a boundary known to have pitted.
**Decisive measurement.** Mark individual pits by potentiostatic nucleation, then lift out the marked site by focused-ion-beam and run atom probe on the boundary that hosted it, with the crystallography from the same lift-out; repeat over tens of pits to build a boundary-character histogram.
**Counts as an answer if.** The pitted boundaries fall into a character class whose measured depletion exceeds a threshold, or they are statistically indistinguishable from the general boundary population.
**Why it matters.** If pitting is boundary-selective, grain-boundary engineering becomes a design variable instead of a post-hoc explanation.

### D032 — Where do shear bands nucleate in a metallic glass?
**Question.** Do shear bands in a bulk metallic glass nucleate at the regions where string-like cooperative motion has already been measured, or at sites fixed by quench-in residual stress?
**Why unresolved.** Instrument blocker: string-like motion is observed at a free surface by X-ray photon correlation spectroscopy or dynamic force microscopy over seconds, while shear-band nucleation is a nanosecond event in the bulk; nobody has located both in one specimen under one stress. Concept blocker: the "flow defect" picture predicts nucleation in fast regions, but residual stress from casting also concentrates there, so the correlation is ambiguous unless stress is mapped independently.
**Decisive measurement.** One sample: map fast dynamics by XPCS under load, map residual stress by focused-ion-beam digital image correlation or synchrotron strain scanning, then load to the first acoustic emission and locate that event; check which map it lands in.
**Counts as an answer if.** The first nucleation sites coincide with the dynamic map after the stress map is regressed out, or the stress map alone accounts for them.
**Why it matters.** Ductility of metallic glasses is set by whether bands multiply or run away; a nucleation criterion is the missing input.

### D033 — Does iron's spin transition change mantle heat flow?
**Question.** Across the spin crossover of iron in bridgmanite, does the lattice thermal conductivity change by an amount that matters for lower-mantle heat transport?
**Why unresolved.** Instrument blocker: spin state is measured by Mössbauer or X-ray emission spectroscopy, thermal conductivity by the periodic-heating (Ångström) method in a diamond-anvil cell — two different experiments, usually different samples, and both are at their limit above 100 GPa. Concept blocker: the crossover changes ionic radius and therefore phonon scattering, but the sign of the effect is not fixed by any simple argument.
**Decisive measurement.** A single diamond-anvil experiment measuring spin state and thermal conductivity in situ at the same pressure and temperature while crossing the transition, on the same grain, so that the conductivity step can be attributed to the spin change rather than to a change of sample.
**Counts as an answer if.** Conductivity shows a resolvable step or kink at the crossover pressure that disappears in an aluminium-free or low-iron analogue.
**Why it matters.** Lower-mantle heat flow sets the core-mantle boundary temperature and the age of the inner core; a spin-driven conductivity change would revise both.

### D034 — Is the 410-km discontinuity sharp because of water?
**Question.** Is the sharpness of the 410-km seismic discontinuity controlled by the water content of wadsleyite?
**Why unresolved.** Access blocker: seismic sharpness is a reflectivity inference, while water content is measured in quenched multi-anvil products or in diamond inclusions — never in the same material at the pressure and temperature where the discontinuity exists. Concept blocker: water both depresses the transition temperature and widens the two-phase loop, so it can sharpen or blur depending on the loop geometry, and the loop width has never been measured as a function of water at pressure.
**Decisive measurement.** In-situ multi-anvil X-ray diffraction on olivine–wadsleyite mixtures with controlled water contents at 13–14 GPa, measuring the loop width and the Clapeyron slope; compare with the transition width inferred from seismological reflectivity beneath regions of differing water flux.
**Counts as an answer if.** One water content reproduces the observed width and the observed topography of the discontinuity, or the observed width requires water contents outside what mineral physics allows.
**Why it matters.** The transition zone is the largest candidate water reservoir in the planet; a seismic handle on its water content would constrain the total.

### D035 — Is inner-core anisotropy hcp-iron's single-crystal elasticity?
**Question.** Does the elastic anisotropy of hexagonal close-packed iron measured on single crystals at inner-core conditions account for the seismic anisotropy of the inner core?
**Why unresolved.** Instrument blocker: making a single-crystal hcp-iron grain in a diamond-anvil cell is possible, but knowing its crystallographic orientation at 330 GPa while measuring direction-dependent sound velocities requires indexing a micrometre grain through an anvil — the orientation determination, not the scattering, is the limit. Assumption blocker: seismic anisotropy models are non-unique (uniform vs. depth-varying, cylindrical vs. hemispherical), so a single predicted ratio does not uniquely test the hypothesis.
**Decisive measurement.** Inelastic X-ray scattering or Brillouin scattering on an indexed single hcp grain at inner-core pressure and temperature, giving v_P along the basal and prismatic directions; compare the ratio and its temperature dependence with the anisotropy pattern inferred from body-wave travel times through the inner core.
**Counts as an answer if.** The predicted fast direction and the magnitude of anisotropy match the seismological pattern within uncertainty, or the predicted magnitude falls short by a stated factor.
**Why it matters.** Inner-core anisotropy is the main evidence for its texture and hence for its growth history; a mineral-physics calibration would turn a pattern into a history.

### D036 — Is there a molten layer at the base of the mantle?
**Question.** Is the several-hundred-kilometre layer above the core–mantle boundary a silicate melt, as its electrical conductivity would reveal?
**Why unresolved.** Access blocker: the layer cannot be sampled. The available probe is indirect — the M2 ocean tide drives a magnetic field whose amplitude and phase depend on mantle conductivity at depth — but the predicted contribution of a thin highly conductive basal layer sits near the current uncertainty of that signal, and no one has jointly inverted the tidal magnetic data with seafloor observatory geomagnetic responses at the required precision.
**Decisive measurement.** Joint inversion of satellite and seafloor geomagnetic tidal and secular-variation signals for a conductivity profile, and comparison with a profile computed from measured melt-bearing silicate conductivities at core-mantle boundary conditions.
**Counts as an answer if.** The inversion requires a conductive layer whose thickness and conductivity match a melt fraction that is independently allowed by mineral physics, or it excludes such a layer above a stated bound.
**Why it matters.** A basal magma ocean changes the chemistry of the core-mantle boundary, the heat flux into the core, and the interpretation of ultralow-velocity zones.

### D037 — Does the quasi-liquid layer depend on what it sits on?
**Question.** Does the thickness of the quasi-liquid layer on ice depend on the lattice mismatch between the ice and its substrate?
**Why unresolved.** Instrument blocker: the layer is one to ten nanometres thick, exists only at a buried interface, and the surface-sensitive tools — sum-frequency generation, ellipsometry — were built for the free ice surface. A buried interface needs a substrate transparent in the relevant infrared range and thin enough for the probe to reach, so almost all existing measurements are on ice/vacuum.
**Decisive measurement.** Sum-frequency generation and X-ray reflectivity through thin ice films grown on substrates spanning a range of lattice mismatch (sapphire, magnesium oxide, platinum, graphene), at fixed temperature just below melting, extracting layer thickness as a function of misfit and of film thickness.
**Counts as an answer if.** Thickness varies systematically with lattice mismatch at fixed temperature and humidity, or it is invariant and the free-surface value is recovered for every substrate.
**Why it matters.** Premelting controls ice friction, sintering, frost heave and the adhesion of ice to solids; a substrate dependence means the layer is a strained film, not a liquid.

### D038 — Does the Kelvin equation survive below ten nanometres?
**Question.** In pores narrower than about ten nanometres, does the chemical potential of confined water still follow the Kelvin–Laplace relation?
**Why unresolved.** Concept blocker: the Kelvin equation assumes the pore is large compared with the adsorbed film and that the interface curvature is the only variable; below ten nanometres the film is a large fraction of the radius, and the correction is unmeasured rather than small. Access blocker: sorption isotherms on real porous media convolve pore size with pore connectivity, so the pore-size dependence is inferred from a model rather than measured.
**Decisive measurement.** Water sorption isotherms at several temperatures on a family of ordered mesoporous silicas with narrow, independently characterised pore sizes spanning three to ten nanometres, extracting the temperature dependence of the filling potential and comparing it with the Kelvin prediction at each size.
**Counts as an answer if.** A systematic deviation appears below a definable pore diameter, with a scaling that identifies the missing term (film thickness, disjoining pressure, or pore-mouth cavitation).
**Why it matters.** Soil water retention, shale sealing and cement durability all extrapolate a macroscale relation into exactly this range.

### D039 — Does solvent reorganization energy predict electron-transfer rates?
**Question.** For a fixed donor–acceptor pair, is the electron-transfer rate predicted by the solvent reorganization energy measured independently, across a wide solvent series?
**Why unresolved.** Concept blocker: in practice the reorganization energy is extracted by fitting the rate to Marcus theory, which makes the test circular — the parameter is defined by the data it is supposed to predict. An independent value requires measuring the solvent's response to an instantaneous charge change, which needs a probe whose charge displacement is known and whose spectral dynamics can be timed.
**Decisive measurement.** One donor–acceptor pair, ten solvents spanning dielectric constant and relaxation time: measure the rate by transient absorption and the solvent reorganization energy from the temperature dependence of a charge-transfer band or from time-resolved Stokes-shift dynamics of the same pair, then compare without fitting.
**Counts as an answer if.** The measured rate is predicted from the independently measured reorganization energy within a stated factor across the series, or it fails in the inverted region where Marcus theory is most specific.
**Why it matters.** If the prediction holds, solvent choice becomes a design parameter for photoredox and redox catalysis; if it fails, the continuum solvent model is wrong in a way that is worth naming.

### D040 — Is an anomalous kinetic isotope effect tunnelling or a barrier shift?
**Question.** For an enzymatic carbon–hydrogen transfer showing a large, curved kinetic isotope effect, is the cause hydrogen tunnelling or a change in the free-energy barrier between isotopes?
**Why unresolved.** Concept blocker: both produce curved Arrhenius plots and swollen isotope effects, so the standard diagnostics cannot separate them. Instrument blocker: a direct tunnelling signature is the nuclear momentum distribution, available from neutron Compton scattering, but that requires gram-scale samples and is almost never performed on the same enzyme preparation whose kinetics were measured.
**Decisive measurement.** Neutron Compton scattering on an enzyme with hydrogen and with deuterium at the transferring position, in turnover conditions, extracting the mean kinetic energy of the transferring nucleus; combine with the isotope effect and its temperature dependence on the same preparation.
**Counts as an answer if.** The momentum distribution yields a tunnelling fraction that quantitatively accounts for the observed isotope effect and its curvature, or it shows no excess and the barrier-shift model is forced.
**Why it matters.** Enzyme design and the interpretation of isotope effects in mechanism all rest on which of these is operating.

### D041 — Does proton transfer go through a measurable intermediate?
**Question.** Does the excess proton in water pass through a distinct, short-lived presolvation intermediate whose lifetime can be measured independently of the transport coefficient?
**Why unresolved.** Instrument blocker: the intermediate is predicted to live for tens of femtoseconds and its vibrational signature overlaps the broad OH stretch, so it is inferred from simulations and from fits to conductivity. Access blocker: ultrafast two-dimensional infrared and impedance/conductivity measurements are made in different cells at different concentrations, so the two reported times have never been compared under identical conditions.
**Decisive measurement.** Polarization-controlled two-dimensional infrared spectroscopy of the excess proton bands in the same sample, temperature and concentration as a four-electrode conductivity measurement, extracting the spectral diffusion time and the structural interconversion time and comparing each with the proton mobility.
**Counts as an answer if.** The spectroscopic interconversion time, converted through a step length, reproduces the measured mobility across a temperature series, or the two disagree by a factor that names a second transport channel.
**Why it matters.** Proton transport sets fuel-cell and biological membrane performance; a measured intermediate would replace a fitted Grotthuss picture with a kinetic scheme.

### D042 — Does rate follow the active site or the spectator?
**Question.** Under reaction conditions on a stepped metal surface, does the catalytic rate track the coverage of the adsorbed intermediate at step sites or the coverage of an abundant spectator on terraces?
**Why unresolved.** Access blocker: operando infrared spectroscopy reports coverage averaged over all sites, and resolving step from terrace requires either a vicinal single crystal with a known step density or isotopic labelling that separates the two bands — neither is routinely combined with a simultaneous rate measurement at elevated pressure. Assumption blocker: microkinetic models assume the most abundant species is the relevant one, which is exactly what a site-resolved measurement would test.
**Decisive measurement.** Infrared bands of carbon monoxide resolved by site using a stepped platinum crystal and isotopically labelled doses, with the reaction rate measured by mass spectrometry in the same cell at one bar; vary step density and temperature and test which coverage the rate follows.
**Counts as an answer if.** The rate is proportional to step-site coverage alone across the series, or to total coverage, or the correlation breaks in a way that identifies a third species.
**Why it matters.** Site-counting is how catalyst nanostructure is optimised; the assumption that terraces only dilute is currently untested in operando.

### D043 — Is upconversion limited by measured exciton diffusion?
**Question.** Is the triplet–triplet annihilation upconversion yield in a solid limited by the triplet diffusion length measured independently of the yield?
**Why unresolved.** Concept blocker: the standard extraction fits the yield itself with a diffusion length, so the parameter is not a prediction. An independent measurement needs the yield as a function of a geometric length the experimenter sets — a spacer between the sensitizer and the annihilator layer — which requires bilayer or core–shell architectures with controlled thickness and clean interfaces.
**Decisive measurement.** Bilayer structures in which the sensitizer and annihilator are separated by a spacer of varied thickness, with upconversion quantum yield measured against thickness; extract the diffusion length from the decay of yield with distance and compare with the value obtained by fitting the neat-film yield curves.
**Counts as an answer if.** The geometric and fitted lengths agree, or the discrepancy identifies a loss channel (self-quenching, interfacial transfer, or trap density) that the fitted model omitted.
**Why it matters.** Upconversion is limited by exciton transport in every solid-state implementation; knowing the real length sets the maximum usable film thickness.

### D044 — Does a protein lengthen radical-pair spin coherence?
**Question.** At room temperature, is the spin coherence time of a photo-generated radical pair longer inside a protein than for the same cofactors free in solution?
**Why unresolved.** Instrument blocker: transient electron paramagnetic resonance gives ensemble coherence times but the broad lines of a protein-bound pair make the dephasing time hard to separate from hyperfine inhomogeneity; single-spin techniques with the required sensitivity have not been applied to a protein-bound pair. Concept blocker: slower motion in a protein reduces some dephasing channels while added nuclear spins accelerate others, so the sign is not obvious.
**Decisive measurement.** Transient EPR with nutation and electron spin echo envelope modulation on a flavin–tryptophan pair in a purified protein and on the same cofactors in solution, at matched temperature and viscosity, extracting the phase memory time from the same pulse sequence.
**Counts as an answer if.** The protein-bound coherence time differs from solution by a resolvable factor and that factor tracks a controlled change in protein rigidity.
**Why it matters.** Spin coherence is the resource in any radical-pair-based sensing or spin-chemical yield effect; knowing where it is longest decides which environment to build devices in.

### D045 — Does a channel read the membrane or the cortex?
**Question.** In a living cell, is the open probability of a mechanosensitive ion channel set by bilayer tension or by the stiffness of the actin cortex beneath it?
**Why unresolved.** Access blocker: patch clamp sets and reads membrane tension, but the cortex stays attached and its own tension cannot be independently clamped; tether pulling measures tension of the membrane-cytoskeleton composite. Concept blocker: both routes predict channels opening on stretch, and the standard experiment — applying pressure — changes both at once.
**Decisive measurement.** In one cell, measure membrane tension by tether pulling and channel open probability by cell-attached patch simultaneously, then partially depolymerise actin and repeat at matched tension by adjusting pipette suction; the dissociation of the two curves identifies the transducer.
**Counts as an answer if.** Open probability collapses at matched tension once the cortex is removed (cortex-reader), or it is unchanged (bilayer-reader), with tension measured rather than assumed.
**Why it matters.** Every statement about cellular mechanosensing inherits this ambiguity; the answer decides whether to target lipids or the cytoskeleton.

### D046 — Below what size does cytoplasmic viscosity stop being a constant?
**Question.** Is the viscosity of the cytosol independent of probe size down to molecular dimensions, or does it fall below a crossover length set by the crowding mesh?
**Why unresolved.** Access blocker: bulk methods average over the probe's whole trajectory — fluorescence recovery after photobleaching gives one number per bleach spot, fluorescence correlation spectroscopy one diffusion coefficient per probe size — and no experiment has run a size ladder of inert tracers in the same cell under the same conditions. Concept blocker: the obstruction model and the glassy-arrest model predict different size dependences, both consistent with any single-size measurement.
**Decisive measurement.** Microinject or genetically tag a ladder of inert tracers spanning one to fifty nanometres into one cell, measure diffusion for each by correlation spectroscopy at matched temperature, and plot diffusion against probe radius divided by the measured mesh size.
**Counts as an answer if.** A crossover radius appears, and it moves when the mesh is changed by osmotic compression or ribosome depletion.
**Why it matters.** Reaction rates, transport limits and the meaning of "viscosity" in every cytoplasmic model depend on whether one number suffices.

### D047 — Can mitochondrial membrane potential change without matrix pH?
**Question.** Can the electrical and chemical components of the mitochondrial protonmotive force vary independently, or does the membrane potential change only with a compensating pH change?
**Why unresolved.** Access blocker: potential-sensitive dyes bind the membrane and perturb the quantity they report, while pH sensors are compartmentalised and their calibration inside the matrix is uncertain; measuring both in the same organelle, at the same time, without one perturbing the other, is not established practice. Concept blocker: chemiosmotic coupling predicts a trade-off only at steady state, so transient uncoupling would look like independence.
**Decisive measurement.** A mitochondrion carrying both a genetically encoded pH sensor and a potential sensor, with two-photon calibration of each in situ, measured through a substrate titration that changes flux; test whether the two quantities move independently on the same timescale.
**Counts as an answer if.** Potential changes by a stated amount with matrix pH held fixed, or the two are locked within measurement error across the titration.
**Why it matters.** Mitochondrial dysfunction is read out through potential alone; if pH moves independently, that readout is ambiguous.

### D048 — Is spindle length a force balance or a microtubule count?
**Question.** Is the length of a mitotic spindle set by a mechanical force balance across sliding microtubules, or by the number of microtubules available?
**Why unresolved.** Instrument blocker: the force on a single overlapping antiparallel bundle cannot be measured inside a spindle without a probe that disturbs it; cantilever measurements give whole-spindle stall force, while electron tomography gives microtubule counts only in fixed cells, so the two are never from the same spindle. Assumption blocker: models of spindle scaling treat microtubule number and sliding force as interchangeable because both grow with spindle size, and no experiment has varied one at fixed size.
**Decisive measurement.** Measure stall force with a calibrated cantilever against a spindle while imaging the same spindle by lattice light-sheet to count antiparallel overlaps, then titrate a kinesin-5 inhibitor to change sliding without changing microtubule number; test which variable predicts length.
**Counts as an answer if.** Length tracks the measured force balance at constant microtubule number, or tracks number at constant force, across the titration.
**Why it matters.** Spindle size scaling and chromosome segregation errors both hinge on the control variable; the two models prescribe different interventions.

### D049 — Do ER sheets and tubules follow curvature proteins or lumenal pressure?
**Question.** Is the sheet-to-tubule ratio of the endoplasmic reticulum set by the density of curvature-inducing membrane proteins or by the osmotic pressure of the lumen?
**Why unresolved.** Concept blocker: both are physically sufficient — protein scaffolds impose spontaneous curvature, and lumenal pressure inflates sheets against the same bending rigidity — and the standard perturbations (overexpression, osmotic shock) change both. Access blocker: lumenal osmotic pressure cannot be measured directly in a living cell, only inferred from volume.
**Decisive measurement.** Measure the sheet-to-tubule ratio by super-resolution in cells whose lumenal osmolarity is clamped by calibrated permeant solutes while curvature-protein density is held fixed by inducible expression, and separately vary protein density at clamped osmolarity; the two response slopes separate the contributions.
**Counts as an answer if.** Morphology is invariant to lumenal osmotic pressure at fixed protein density, or scales with pressure and not with protein density, with the other variable measured rather than assumed.
**Why it matters.** ER morphology is used as a readout of secretory load and of disease states; which variable it actually reports is undetermined.

### D050 — Is D″ anisotropy the texture of post-perovskite?
**Question.** Does the seismic anisotropy of the D″ layer match the crystallographic texture measured on post-perovskite deformed at core–mantle boundary conditions?
**Why unresolved.** Access blocker: texture must be measured on an aggregate sheared at roughly 120 GPa and 2500 K, which needs multigrain X-ray diffraction in a diamond-anvil cell with a controlled shear geometry; seismology supplies a path-integrated anisotropy with trade-offs against isotropic velocity and layer thickness. No one has propagated a measured texture through a realistic flow field to a predicted splitting pattern.
**Decisive measurement.** Shear a post-perovskite aggregate in a diamond-anvil cell, measure its texture by multigrain diffraction, compute the aggregate elastic tensor, and forward-model SKS splitting and its azimuthal dependence for a plausible subduction-driven flow; compare with observations beneath a well-studied region.
**Counts as an answer if.** The predicted fast direction and magnitude match the observed splitting including its azimuthal pattern, or the measured texture cannot produce the observed magnitude.
**Why it matters.** D″ anisotropy is the main window on lowermost-mantle flow; without a texture calibration, every flow inference is model-dependent.