The preceding seven sections developed the ether framework from foundations through gravity, cosmology, quantum mechanics, and entanglement. This section turns to the empirical programme: what the framework predicts, how those predictions interconnect, what experiments could test them, and what outcomes would falsify them.

We depart from the conventional approach of listing predictions in a table. The ether framework's empirical strength lies not in any single prediction but in the interconnection between predictions: a small set of fundamental parameters, many of which are determined by existing observations, generates consequences across domains that are completely unrelated in standard physics. This section makes that interconnection explicit.

9.1 The Parameter Landscape

9.1.1 Fundamental Parameters of the Ether

The superfluid ether model developed in Sections 3–7 introduces five fundamental parameters. These are material properties of the ether — the quantities that, in principle, a complete theory of ether microphysics would derive but which, at the present stage, must be determined empirically.

Parameter 1: Ether quantum mass $m_e$. The mass of the individual bosonic quanta whose Bose–Einstein condensation constitutes the superfluid ether (Section 4.2.3a). This parameter sets the mass scale for all gravitational and cosmological ether phenomena. Berezhiani and Khoury [71, 72] estimate $m_e \sim 1$–$2$ eV/$c^2$ from the requirement that the superfluid phonon force reproduces MOND phenomenology at galaxy scales; we adopt $m_e = 1$ eV/$c^2$ as the fiducial value. The ether quantum mass determines, inter alia, the BEC critical temperature ((4.78)) and hence the galaxy/cluster transition scale (Section 4.2.7).

Parameter 2: Chemical potential $\hat{\mu}$. The chemical potential of the ether condensate (Section 4.2.3a), defined through the superfluid Lagrangian $\mathcal{L}_e = P(X)$ with $X = \hat{\mu} - m_e\Phi_{\text{grav}} - \hbar^2(\nabla\theta)^2/(2m_e)$ ((4.25)). The chemical potential sets the condensate's energy scale and determines both the phonon sound speed $c_s = \sqrt{\hat{\mu}/m_e}$ and the healing length $\xi = \hbar/\sqrt{2m_e\hat{\mu}}$. It is the single parameter most tightly constrained by cosmological observations, as shown below.

Parameter 3: Interaction coupling $g_{\text{int}}$. The two-body contact interaction strength between ether quanta, related to the $s$-wave scattering length by $g_{\text{int}} = 4\pi\hbar^2 a_s/m_e$ (Section 4.3.8). This determines the equilibrium condensate number density $n_0 = \hat{\mu}/g_{\text{int}}$ and hence the cosmological ether mass density $\rho_e = m_e n_0$.

Parameter 4: Baryon–ether coupling. The strength of the direct coupling between baryonic matter and the ether condensate phase $\theta$, parameterised by the combination $\alpha_\Lambda\Lambda/M_{\text{Pl}}$ in the Berezhiani–Khoury Lagrangian ((4.72)). This coupling determines the MOND acceleration scale $a_0$ through the relation $a_0 = \alpha_\Lambda^3\Lambda^2/(M_{\text{Pl}}\hat{\mu})$ ((4.61)). We treat this combination as a single effective parameter, since the individual factors $\alpha_\Lambda$ and $\Lambda$ enter only in this specific combination in all observable predictions.

Parameter 5: Transverse microstructure scale $\ell_e$. The length scale of the ether's internal structure as experienced by electromagnetic (transverse) modes (Section 3.8). This parameter is logically independent of the gravitational-sector parameters ($m_e$, $\hat{\mu}$, $g_{\text{int}}$, baryon–ether coupling), which govern the longitudinal (phonon) sector. The healing length $\xi$ provides the UV cutoff for phonon modes; $\ell_e$ provides the UV cutoff for EM modes. As discussed in Section 6.6.3, the relationship between these two scales is an open problem — they are governed by different physical mechanisms (condensate coherence for $\xi$; transverse microstructure for $\ell_e$) and need not be comparable. The modified dispersion relation for light ((3.46)) depends on $\ell_e$, which may be as small as the Planck length $\ell_P \approx 1.6 \times 10^{-35}$ m.

9.1.2 Derived Quantities

All other quantities appearing in the monograph's predictions are functions of the five fundamental parameters:

Derived quantity	Expression	Equation	Value (fiducial)
Healing length $\xi$	$\hbar/\sqrt{2m_e\hat{\mu}}$	(4.103)	$7.9\;\mu$m
Sound speed $c_s$	$\sqrt{\hat{\mu}/m_e}$	(4.109)	$5.3 \times 10^6$ m/s
Condensate density $n_0$	$\hat{\mu}/g_{\text{int}}$	(4.124)	$1.25 \times 10^9$ m$^{-3}$
BEC critical temperature $T_c$	$2\pi\hbar^2(n_s/\zeta(3/2))^{2/3}/(m_e k_B)$	(4.78)	$\sigma_c \approx 500$ km/s
MOND scale $a_0$	$\alpha_\Lambda^3\Lambda^2/(M_{\text{Pl}}\hat{\mu})$	(4.61)	$1.2 \times 10^{-10}$ m/s$^2$
Dark energy density $\rho_\Lambda$	$0.0146\,m_e^{3/2}\hat{\mu}^{5/2}/\hbar^3$	(4.122)	$5.4 \times 10^{-10}$ J/m$^3$
Dark matter density $\rho_{\text{DM}}$	$m_e\hat{\mu}c^2/g_{\text{int}}$	(4.96, 4.132)	$2.25 \times 10^{-10}$ J/m$^3$
Nelson diffusion $D$	$\hbar/(2m)$	(7.12)	[particle-dependent]
Dispersion correction $\xi_2$	$-1/12$ (lattice model)	(3.46)	$-0.083$

The diffusion coefficient $D = \hbar/(2m)$ is notable: it contains no ether parameters at all. It is determined entirely by the ZPF spectrum and the particle mass (Section 7.3), making the Nelson–SED bridge and its consequence — the Schrödinger equation — completely independent of the ether's material properties.

9.1.3 Observational Constraints

Three well-measured cosmological quantities constrain three independent combinations of the gravitational-sector parameters.

Constraint I: Dark energy density. The observed $\rho_\Lambda^{\text{obs}} = (6.36 \pm 0.07) \times 10^{-10}$ J/m$^3$ [7] fixes:

$$m_e^{3/2}\hat{\mu}^{5/2} = \frac{\hbar^3\rho_\Lambda^{\text{obs}}}{0.0146} = (2.49 \pm 0.03) \times 10^{-99}\;\text{kg}^{3/2}\,\text{J}^{5/2} \tag{9.1}$$

This is the most precise constraint, determined by Planck satellite measurements to $\sim 1\%$.

Constraint II: Dark matter density. The observed $\rho_{\text{DM}}^{\text{obs}} = \Omega_{\text{DM}}\rho_{\text{crit}}c^2 \approx 2.25 \times 10^{-10}$ J/m$^3$ [7] fixes:

$$\frac{m_e\hat{\mu}}{g_{\text{int}}} = \frac{\rho_{\text{DM}}^{\text{obs}}}{c^2} = 2.50 \times 10^{-27}\;\text{kg/m}^3 \tag{9.2}$$

This constrains the combination $m_e\hat{\mu}/g_{\text{int}}$, linking the condensate density to the interaction strength.

Constraint III: MOND acceleration scale. The observed $a_0 = (1.20 \pm 0.02) \times 10^{-10}$ m/s$^2$ from the Radial Acceleration Relation [60] fixes:

$$\frac{\alpha_\Lambda^3\Lambda^2}{M_{\text{Pl}}\hat{\mu}} = a_0^{\text{obs}} \tag{9.3}$$

This determines the baryon–ether coupling in terms of $\hat{\mu}$ (and hence, via Constraint I, in terms of $m_e$).

Net result. The gravitational sector has four fundamental parameters ($m_e$, $\hat{\mu}$, $g_{\text{int}}$, baryon–ether coupling) and three observational constraints (9.1–9.3). This leaves one free parameter in the gravitational sector — which we may take to be $m_e$.

For any chosen $m_e$, the chemical potential is determined by Constraint I:

$$\hat{\mu}(m_e) = \left(\frac{\hbar^3\rho_\Lambda^{\text{obs}}}{0.0146\;m_e^{3/2}}\right)^{2/5} \tag{9.4}$$

The interaction coupling follows from Constraint II, and the baryon–ether coupling from Constraint III. All derived quantities — $\xi$, $c_s$, $n_0$, $T_c$ — become functions of $m_e$ alone.

The transverse microstructure scale $\ell_e$ is unconstrained by gravitational-sector observations, adding a second free parameter for the electromagnetic sector.

9.1.4 The Effective Parameter Count

We summarise the parameter economy:

	Fundamental parameters	Observational constraints	Free parameters
Gravitational sector	4 ($m_e$, $\hat{\mu}$, $g_{\text{int}}$, coupling)	3 ($\rho_\Lambda$, $\rho_{\text{DM}}$, $a_0$)	1 ($m_e$)
EM sector	1 ($\ell_e$)	0	1 ($\ell_e$)
Total	5	3	2

Comparison with $\Lambda$CDM. The standard cosmological model has six free parameters ($H_0$, $\Omega_b h^2$, $\Omega_c h^2$, $\tau$, $n_s$, $A_s$) determined by fits to the CMB power spectrum, baryon acoustic oscillations, Type Ia supernovae, and other datasets. These six parameters describe a narrower phenomenological domain than the ether framework addresses: $\Lambda$CDM says nothing about the MOND acceleration scale, the RAR functional form, the equation of state $w = -1$ (which it treats as an input, not an output), the vacuum energy scale (which it cannot predict), the Bullet Cluster transition, the quantum-classical bridge, or entanglement. The ether framework addresses all of these from five fundamental parameters.

A fully fair comparison requires the ether framework to predict the CMB power spectrum — which demands ether cosmological perturbation theory that has not yet been developed (see Section 11). Until this is done, the parameter-count comparison must be qualified: the ether framework is more economical for the phenomena it addresses, but it does not yet address all the phenomena that $\Lambda$CDM covers.

9.1.5 The Prediction Web

The interconnection between parameters, observational constraints, and predictions is the ether framework's central empirical argument. We represent this structure explicitly.

Tier 0: Structural predictions (no parameters). These predictions follow from the framework's mathematical structure and are independent of all parameter values:

• $w = -1$ for dark energy (Theorem 4.2: consequence of Lorentz-invariant ZPF spectrum)
• Functional form of the RAR: $g[1 - e^{-\sqrt{g/a_0}}] = g_N$ ((4.59): derived from superfluid phase transition physics; $a_0$ enters as a scale but the shape is predicted)
• MOND-to-CDM transition with increasing system temperature (two-fluid model: Section 4.2.7)
• Correlation between cluster collision velocity and dark matter–baryon offset (Section 4.2.7f)
• Schrödinger equation from ether dynamics, with $D = \hbar/(2m)$ (Theorem 7.1: no ether parameters appear)
• Entanglement of ZPF-driven oscillators for any nonzero coupling (Theorem 8.3)
• No-signalling despite ether non-locality (Proposition 8.3)
• Infinite ZPF correlation length at $T = 0$ (power-law $r^{-2}$ decay: Section 8.6)

These structural results cannot be adjusted by tuning parameters. They stand or fall with the framework itself.

Tier 1: One-parameter family (gravitational sector). Once $m_e$ is specified, the following predictions are fully determined:

• Sub-millimetre gravity: Yukawa deviation with range $\xi(m_e)$ and $\alpha_\xi \sim \mathcal{O}(1)$ ((4.173))
• Ether sound speed: $c_s(m_e) = \sqrt{\hat{\mu}(m_e)/m_e}$ (testable via gravitational wave dispersion)
• BEC critical temperature: $T_c(m_e)$, equivalently critical velocity dispersion $\sigma_c(m_e)$ (determines the precise galaxy/cluster transition)
• Phonon ZPF cutoff frequency: $\omega_{\max}^{\text{phonon}} = c_s/\xi = \hat{\mu}/\hbar$ (infrared, Section 4.3.4)

For the fiducial $m_e = 1$ eV: $\xi = 7.9\;\mu$m, $c_s = 5.3 \times 10^6$ m/s ($0.018\,c$), $\sigma_c \approx 500$ km/s.

Tier 2: Independent parameter (EM sector). The modified dispersion relation for light ((3.46)) depends on $\ell_e$ independently of the gravitational sector:

• Energy-dependent photon time delay: $\Delta t \propto \ell_e^2 E^2 d$ ((3.48))
• If $\ell_e \sim \ell_P$: $\Delta t \sim 10^{-9}$ s for 100 GeV photons at $z \sim 1$ (marginally detectable with CTA)
• Current constraint: $\ell_e < 5.7 \times 10^{-13}$ m ((3.50))

Tier 3: Predictions requiring further theoretical work. These are well-posed within the framework but depend on solving open problems identified in Sections 7–7:

• Discrete Bell–CHSH violation from SED detection model (Problem 7.1: requires solving the nonlinear detection problem)
• Thermal Bell degradation functional form at intermediate temperatures (Prediction 7.1: requires Problem 7.1 for precise numerical coefficients)
• Excited state spectrum from nonlinear SED (Section 7.5: requires numerical simulation programme)
• Spin from ether vorticity (Section 7.6: highly speculative)

9.1.6 Cross-Prediction Constraints

The most powerful empirical feature of the ether framework is that a single measurement of $m_e$ would simultaneously determine predictions across unrelated physical domains. We trace three cross-prediction chains:

Chain A: Sub-millimetre gravity $\leftrightarrow$ sound speed. A measurement of the Yukawa range $\xi$ in a sub-millimetre gravity experiment determines $m_e\hat{\mu}$ via $\xi = \hbar/\sqrt{2m_e\hat{\mu}}$. Combined with Constraint I (which fixes $m_e^{3/2}\hat{\mu}^{5/2}$), this uniquely determines both $m_e$ and $\hat{\mu}$ individually. The sound speed $c_s = \sqrt{\hat{\mu}/m_e}$ is then predicted with no remaining freedom.

Explicitly: if an experiment measures $\xi = \xi_{\text{meas}}$, then $m_e\hat{\mu} = \hbar^2/(2\xi_{\text{meas}}^2)$, which combined with $m_e^{3/2}\hat{\mu}^{5/2} = C_\Lambda$ (from (9.1)) gives:

$$m_e = \left(\frac{C_\Lambda^2 \cdot 4\xi_{\text{meas}}^4}{\hbar^4}\right)^{1/4}, \qquad \hat{\mu} = \frac{\hbar^2}{2\xi_{\text{meas}}^2\,m_e} \tag{9.5}$$

Every entry in the Tier 1 prediction list is then fixed.

Chain B: Galaxy/cluster transition $\leftrightarrow$ sub-millimetre gravity. The BEC critical temperature $T_c$ (equivalently, the velocity dispersion $\sigma_c$ at which the superfluid-to-normal transition occurs) depends on $m_e$ and $n_0$, both of which are determined once $m_e$ is known (via Constraints I and II). A precise measurement of the transition scale — for instance, by systematic study of the mass discrepancy in galaxy groups spanning $\sigma = 200$–$800$ km/s — would constrain $m_e$, which simultaneously predicts $\xi$.

Chain C: Cosmic coincidence as consistency check. The ratio $\Omega_\Lambda/\Omega_{\text{DM}}$ is not a free prediction — it is constructed from the same observations used to fix the parameters (Constraints I and II). However, the fact that a single medium can produce both dark energy (via phonon ZPF) and dark matter (via condensate mass) with the observed densities is a non-trivial consistency requirement. In particular, the ratio depends on $m_e$ through (4.166):

$$\frac{\Omega_\Lambda}{\Omega_{\text{DM}}} = \frac{0.184\,a_s\,c_s^3}{c^4\,\hbar} \tag{9.6}$$

For arbitrary $m_e$, $\hat{\mu}$, and $a_s$ satisfying Constraints I–II, this ratio need not be close to the observed 2.65. That it is — for the same parameter range independently motivated by the MOND phenomenology — is a structural success of the unification. What would be a genuine zero-free-parameter prediction is the equation of state $w = -1$, which follows from Theorem 4.2 regardless of parameter values.

9.1.7 What Is Not Connected

Honesty requires identifying gaps in the prediction web.

The EM–gravitational disconnect. The transverse microstructure scale $\ell_e$ has no known relationship to the gravitational-sector parameters ($m_e$, $\hat{\mu}$, $g_{\text{int}}$). The modified EM dispersion relation (Section 3.8) is therefore an independent prediction that cannot be cross-checked against sub-millimetre gravity or dark sector measurements. This is a genuine gap in the framework's unification programme. A complete theory of ether microphysics would determine $\ell_e$ from the same condensate physics that determines $\xi$, but this has not been achieved.

The EM cutoff problem. As noted in Section 6.6.3, the phonon UV cutoff ($\omega_{\max}^{\text{phonon}} \sim 4 \times 10^{13}$ rad/s, infrared) is far below atomic frequencies ($\omega_{\text{atom}} \sim 10^{16}$–$10^{17}$ rad/s). The SED results of Sections 6–7 require the EM ZPF to extend to frequencies well above atomic scales, which means the EM cutoff must be governed by different physics than the phonon cutoff. The precise EM cutoff mechanism is an open problem. Until it is resolved, the SED programme operates with the standard assumption that the EM ZPF extends to arbitrarily high frequencies — an assumption that is internally consistent but not yet derived from the ether's microphysics.

CMB perturbation theory. The ether framework reproduces the homogeneous Friedmann cosmology (Section 4.1) but does not yet predict the CMB temperature and polarisation power spectra. This requires developing the theory of perturbations in the superfluid ether, accounting for the two-fluid (superfluid + normal) dynamics, the phonon-mediated force, and the coupling to baryonic matter. Until this is done, the framework cannot be compared to $\Lambda$CDM on its home ground — the six $\Lambda$CDM parameters are fitted to the CMB, and the ether framework offers no alternative fit.

Section 5 partially addresses the electromagnetic constitutive structure by deriving the plasma dielectric tensor (Theorem 5.1) and identifying the plasma frequency as the ether's low-frequency EM cutoff. However, the relationship between $\omega_p$ (a property of the charge population) and $\ell_e$ (a property of the ether's intrinsic transverse microstructure) remains unknown.

This is the most significant empirical gap in the programme. We identify it as the highest priority for future theoretical work (Section 11).

---

9.2 Structural Predictions: Results Independent of Parameter Values

The predictions in this subsection follow from the mathematical structure of the ether framework — from symmetries, phase transitions, and derivation chains — and are independent of the values of all five fundamental parameters. They cannot be adjusted by tuning $m_e$, $\hat{\mu}$, $g_{\text{int}}$, the baryon–ether coupling, or $\ell_e$. They stand or fall with the framework itself.

9.2.1 The Dark Energy Equation of State: $w = -1$

The prediction. The ether framework requires the dark energy equation of state to be $w = -1$ exactly, up to corrections of order $(\xi/R_H)^2 \sim 10^{-62}$.

Derivation chain. The argument proceeds in three steps, each proved in the indicated section:

(i) The superfluid ether's phonon zero-point fluctuations constitute the dark energy (Section 4.3.4). The phonon ZPF energy density is given by the exact integral over the Bogoliubov spectrum ((4.114)):

$$\rho_{\text{ZPF}} = \frac{\hat{\mu}}{8\pi^2\xi^3}\int_0^1 q^3\sqrt{q^2+2}\,dq = \frac{0.0146\;m_e^{3/2}\hat{\mu}^{5/2}}{\hbar^3} \tag{4.122}$$

The numerical coefficient is exact (evaluated by substitution in (4.116)).

(ii) At wavelengths $\lambda \gg \xi$, the Bogoliubov spectrum reduces to the acoustic (linear) dispersion $\omega = c_s k$ ((4.109)). In this regime, the zero-point energy per mode is $\frac{1}{2}\hbar\omega \propto \omega$, and the spectral energy density takes the form $\rho(\omega) \propto \omega^3$ — the unique Lorentz-invariant spectrum (Theorem 4.2, proved in Section 4.3.6).

(iii) A Lorentz-invariant energy density has a stress-energy tensor proportional to the metric: $T_{\mu\nu}^{\text{ZPF}} = -\rho_{\text{ZPF}}\,g_{\mu\nu}$ ((4.170)). In the rest frame with signature $(-,+,+,+)$:

$$T_{00} = +\rho_{\text{ZPF}}, \qquad T_{ij} = -\rho_{\text{ZPF}}\,\delta_{ij} \tag{4.141}$$

giving pressure $p = -\rho_{\text{ZPF}}c^2$, hence $w = p/(\rho c^2) = -1$ ((4.143)). $\square$

The only point where the Bogoliubov spectrum deviates from exact Lorentz invariance is near the cutoff $k\xi \sim 1$, where the dispersion bends from linear to quadratic ((4.110)). This introduces a correction ((4.144)):

$$|1+w| \sim \left(\frac{\xi}{R_H}\right)^2 \sim \left(\frac{10^{-5}\;\text{m}}{10^{26}\;\text{m}}\right)^2 \sim 10^{-62}$$

which is unobservably small for any conceivable experiment.

What is structural. The prediction $w = -1$ depends on no parameter values. It follows entirely from:

• The ether being a superfluid condensate (any condensate has a phonon spectrum)
• The phonon spectrum being approximately linear at long wavelengths (Goldstone's theorem: the broken $U(1)$ symmetry guarantees a gapless linear mode [84])
• Linear dispersion in three spatial dimensions producing a Lorentz-invariant $\omega^3$ spectrum (Theorem 4.2: the proof is purely kinematic)

The value of $\rho_\Lambda$ depends on $m_e$ and $\hat{\mu}$; the equation of state $w = -1$ does not.

Comparison with observation. The Planck satellite, combined with baryon acoustic oscillation and Type Ia supernova data, constrains the dark energy equation of state to $w = -1.03 \pm 0.03$ [7] — consistent with the ether prediction at the $1\sigma$ level.

The DESI caveat. The Dark Energy Spectroscopic Instrument (DESI) DR2 data, combined with CMB and supernova observations, show a 2.8–4.2$\sigma$ preference (depending on the supernova dataset) for a time-varying equation of state with $w_0 > -1$ and $w_a < 0$ in the CPL parameterisation $w(z) = w_0 + w_a z/(1+z)$ [133, 134]. However, when fitted with a constant $w$ (the $w$CDM model), the DESI data remain fully consistent with $w = -1$ [133]. The tension is specifically with time-varying dark energy, not with a cosmological constant.

In the ether framework, $w$ is constant by construction: the Bogoliubov spectrum is a property of the condensate ground state and does not evolve with redshift (so long as the condensate parameters $m_e$ and $\hat{\mu}$ are cosmological constants, which is the simplest assumption). If the DESI hint of time-varying $w$ is confirmed at high significance, this would present a challenge for the ether framework as currently formulated. Possible resolutions within the framework — such as cosmological evolution of $\hat{\mu}$ or a non-equilibrium condensate — are speculative and would require substantial theoretical development. We regard this as a genuine point of vulnerability.

Falsification condition. If future measurements establish $|1 + w| > 10^{-2}$ with high confidence (say, $>5\sigma$) for a constant $w$, the ether prediction is falsified. Time variation of $w$ is a separate question (see above).

9.2.2 The Radial Acceleration Relation: Functional Form

The prediction. The ether framework predicts a specific functional relationship between the observed gravitational acceleration $g$ and the Newtonian (baryonic) acceleration $g_N$ in galactic systems:

$$g\!\left[1 - \exp\!\left(-\sqrt{g/a_0}\right)\right] = g_N \tag{4.59}$$

The acceleration scale $a_0$ is a parameter (fixed by Constraint III, (9.3)); the functional form — the specific interpolating function $\mu_e(x) = 1 - e^{-\sqrt{x}}$ — is a structural prediction.

Derivation chain. The derivation proceeds through four stages (Section 4.2.3):

(i) The ether is a gravitational dielectric medium whose response to gravitational fields is described by the modified Poisson equation (Theorem 4.1):

$$\nabla\cdot\left[\mu_e(|\mathbf{g}|/a_0)\,\mathbf{g}\right] = -4\pi G\rho_m \tag{4.14c}$$

This structure follows from any medium with local, isotropic gravitational self-interaction — it is not specific to the superfluid model.

(ii) The ether is a superfluid condensate with three-body equation of state $P(X) = \frac{2\alpha_3}{3}X^{3/2}$ ((4.28)). The phonon-mediated force on baryonic matter, computed from the Euler–Lagrange equation for the condensate phase ((4.33)), produces the deep-MOND acceleration $g \approx \sqrt{a_0 g_N}$ in the weak-field regime ((4.53), following Berezhiani and Khoury [71]).

(iii) The superfluid-to-normal phase transition, governed by the Landau criterion for superfluid stability, determines the condensate fraction as a function of local gravitational acceleration (Section 4.2.3c). The fraction of ether remaining superfluid at acceleration $g$ is:

$$f_c(g) = 1 - \exp\!\left(-\sqrt{a_0/g}\right) \tag{4.56}$$

This functional form arises because the ether flow velocity scales as $\sqrt{g\,r}$ (from the Painlevé–Gullstrand identification, Section 3.5), and the relevant energy per ether quantum scales with velocity, giving a disruption energy $E_{\text{grav}} \propto \sqrt{g}$. The Boltzmann factor $\exp(-E_{\text{cond}}/E_{\text{grav}}) = \exp(-\sqrt{a_0/g})$ then determines the normal fraction.

(iv) Combining the MOND force (weighted by the superfluid fraction) with Newtonian gravity yields the gravitational permittivity:

$$\mu_e(g/a_0) = 1 - \exp\!\left(-\sqrt{g/a_0}\right) \tag{4.58}$$

and the full relation (4.59). $\square$

What is structural. The functional form $\mu_e(x) = 1 - e^{-\sqrt{x}}$ is determined by the superfluid phase transition physics. It satisfies both limiting constraints derived in Section 4.2.3:

$$\mu_e(x) \to 1 \quad \text{as } x \to \infty \qquad \text{(Newtonian limit)} \tag{4.19}$$ $$\mu_e(x) \to \sqrt{x} \quad \text{as } x \to 0 \qquad \text{(MOND limit: flat rotation curves)} \tag{4.20}$$

The specific shape of the transition between these limits — the $\exp(-\sqrt{x})$ form — is the prediction. Different ether microphysics could satisfy (4.15a–b) with different interpolating functions; the superfluid condensate model selects this particular one.

Comparison with observation. McGaugh, Lelli and Schombert [60] measured the radial acceleration relation using 2693 data points in 153 late-type galaxies from the SPARC (Spitzer Photometry and Accurate Rotation Curves) database. Lelli et al. [135] extended the sample to 240 galaxies spanning nine decades in stellar mass, including late-type galaxies, early-type galaxies, and dwarf spheroidals.

The empirical fitting function adopted by McGaugh et al. [60] is:

$$g_{\text{obs}} = \frac{g_{\text{bar}}}{1 - \exp\!\left(-\sqrt{g_{\text{bar}}/g_\dagger}\right)} \tag{9.7}$$

with best-fit $g_\dagger = (1.20 \pm 0.02) \times 10^{-10}$ m/s$^2$. This is precisely (4.60) — the explicit approximation to our implicit relation (4.59), obtained by replacing $g \to g_N$ in the exponential argument. The identification is $g_\dagger = a_0$.

The agreement is exact at the level of the functional form: the empirical fitting function that best describes 2693 independent data points across 153 galaxies, spanning five decades in baryonic mass, is the same function that the superfluid ether model derives from condensate phase transition physics.

Quantitatively:

• The observed rms scatter about the RAR is 0.13 dex [60], largely driven by observational uncertainties (distance, inclination, stellar mass-to-light ratio).
• Li et al. [136] fitted the RAR to 175 individual SPARC galaxies using Markov Chain Monte Carlo, marginalising over stellar mass-to-light ratio, distance, and inclination, obtaining residual scatter of only 0.057 dex (~13%).
• There is no credible evidence for galaxy-to-galaxy variation in the critical acceleration scale $g_\dagger$ [136] — consistent with $a_0$ being a universal constant determined by ether parameters.

Comparison with alternative interpolating functions. The MOND literature employs several interpolating functions satisfying (4.15a–b):

Function	$\mu(x)$	Provenance
Simple [137]	$x/(1+x)$	Famaey & Binney (2005)
Standard [138]	$x/\sqrt{1+x^2}$	Milgrom (1983)
Ether (this work)	$1 - e^{-\sqrt{x}}$	Derived from superfluid phase transition

All three satisfy the asymptotic constraints and produce acceptable fits to most rotation curves individually. The differences appear primarily in the transition region $g \sim a_0$ (where $x \sim 1$). For $x = 1$:

$$\mu_{\text{simple}}(1) = 0.500, \qquad \mu_{\text{standard}}(1) = 0.707, \qquad \mu_{\text{ether}}(1) = 0.632$$

The ether function is intermediate between the simple and standard functions. Li et al. [136] found that the RAR functional form (9.7) — identical to the ether prediction — provides excellent fits with astrophysically reasonable parameters, though the precision is not yet sufficient to decisively discriminate between interpolation functions.

What $\Lambda$CDM predicts. The standard cosmological model does not predict the RAR. In $\Lambda$CDM, the relationship between baryonic and total acceleration depends on the dark matter halo profile, which varies with halo mass, concentration, and assembly history. Reproducing the observed tightness of the RAR requires either fine-tuned feedback prescriptions [139] or a conspiracy between baryonic and dark matter distributions that has no natural explanation within the framework. The existence of a universal, tight RAR with the specific functional form (4.59) is a prediction of the ether framework and a puzzle for $\Lambda$CDM.

Falsification condition. If future observations — particularly from 21-cm surveys (SKA), spatially resolved kinematics from IFU surveys (MUSE, ALMA), or from the ultra-low-surface-brightness regime — establish that:

(a) the RAR has significant intrinsic scatter beyond observational uncertainties (breaking universality), or

(b) the functional form is better described by a different interpolating function that cannot be derived from any superfluid phase transition (e.g., a broken power law with a sharp transition),

then the specific superfluid ether model is falsified. The general gravitational dielectric framework (Theorem 4.1) would survive scenario (b) but not (a), since any single-function $\mu_e$ predicts a universal RAR.

9.2.3 The Galaxy–Cluster Phase Transition

The prediction. The ether framework predicts that the dark sector exhibits qualitatively different behaviour at galaxy scales and cluster scales:

• Galaxy scales ($\sigma \lesssim 300$ km/s): superfluid ether → phonon-mediated MOND enhancement → tight RAR, flat rotation curves, baryonic Tully–Fisher relation
• Cluster scales ($\sigma \gtrsim 800$ km/s): normal ether → no MOND enhancement → CDM-like behaviour, Bullet Cluster phenomenology, cluster mass-to-light ratios $M/M_b \sim 5$–$10$

The transition between these regimes is a thermodynamic phase transition of the ether condensate, not an imposed boundary.

Derivation chain. In Landau's two-fluid model (Section 4.2.7b), a superfluid at finite temperature consists of two interpenetrating components:

$$\rho_e = \rho_s + \rho_n \tag{4.76}$$

The superfluid fraction depends on the effective temperature through the condensate fraction ((4.77)):

$$\frac{\rho_n}{\rho_e} = f_n(T) = (T/T_c)^\alpha \qquad \text{for } 0 < T < T_c$$

A virialised gravitational system with velocity dispersion $\sigma$ has effective temperature $T_{\text{eff}} = m_e\sigma^2/k_B$ ((4.81)). The BEC critical temperature is given by the standard formula ((4.78)).

The superfluid fraction $\rho_s/\rho_e = 1 - (T_{\text{eff}}/T_c)^\alpha$ determines whether the phonon-mediated MOND force is operative. The prediction is that systems with $T_{\text{eff}}/T_c \ll 1$ (galaxies) exhibit MOND phenomenology, while systems with $T_{\text{eff}}/T_c \gg 1$ (clusters) exhibit CDM-like behaviour.

What is structural. Three qualitative features are parameter-independent:

(i) Existence of the transition. Any BEC has a critical temperature. Any superfluid ether will therefore exhibit a phase transition between MOND-like and CDM-like regimes as the system's effective temperature increases. This is a thermodynamic necessity, not a model choice.

(ii) Direction of the transition. Larger, hotter systems (higher $\sigma$) are driven into the normal phase, suppressing the MOND enhancement. This means galaxy-scale MOND and cluster-scale CDM, not the reverse. The direction is fixed by the Landau criterion.

(iii) Smoothness. The transition is continuous (second-order phase transition for an ideal BEC), not a sharp cutoff. Systems near $T_{\text{eff}} \sim T_c$ — galaxy groups with $\sigma \sim 400$–$600$ km/s — should show intermediate behaviour: partial MOND enhancement and partial CDM-like mass discrepancy.

What depends on parameters. The precise transition scale $\sigma_c$ depends on $m_e$ (through $T_c$). For the fiducial $m_e = 1$ eV, $\sigma_c \approx 500$ km/s (Section 4.2.7b). The constraint from requiring that galaxies are superfluid and clusters are normal gives $\sigma_c \approx 300$–$800$ km/s ((4.96)), which is satisfied by the fiducial value.

Observational signatures. The phase transition structure explains several otherwise puzzling observations:

(a) Why MOND underestimates cluster masses. Milgrom's formula applied to galaxy clusters predicts total-to-baryonic mass ratios of $\sim 2$–$3$, whereas observation requires $\sim 5$–$10$ [77]. In the ether framework, the discrepancy arises because the normal ether component — which gravitates like CDM — is not captured by the MOND enhancement formula. The correct cluster mass is ((4.94)):

$$\frac{M_{\text{total}}}{M_b} \approx 1 + \frac{\Omega_e}{\Omega_b} \approx 1 + \frac{0.26}{0.05} = 6.2 \tag{4.95}$$

consistent with observed values [82].

(b) The Bullet Cluster. The observed separation of lensing peaks from X-ray emission in the Bullet Cluster (1E 0657-558) is reproduced by the two-fluid model: the normal ether ($\sim 83\%$ of total mass) is collisionless and co-locates with the galaxies, while the intracluster gas ($\sim 15\%$) piles up at the collision centre (Section 4.2.7d). The predicted lensing peak ratio $\kappa_{\text{lobe}}/\kappa_{\text{centre}} \approx 5.7$ ((4.91)) is within a factor of 2 of the observed ratio of $\sim 8$–$10$ [64, 80], with the discrepancy attributable to simplifications in the mass distribution model.

(c) The Abell 520 anomaly. The "dark core" in Abell 520 — a mass concentration coincident with gas, anomalous for collisionless CDM [81] — is naturally explained if this slower collision ($T_{\text{eff}} \sim T_c$) retains a higher superfluid fraction, causing part of the ether to track the gravitational potential (including the gas component) rather than passing through collisionlessly (Section 4.2.7f).

What $\Lambda$CDM predicts. Standard CDM has no phase transition. Dark matter is collisionless at all scales, and there is no mechanism to produce MOND-like behaviour at galaxy scales. The RAR tightness must be explained by baryonic feedback processes, which require fine-tuning and do not naturally produce a universal relation. Conversely, pure MOND has no mechanism to produce CDM-like behaviour at cluster scales. The ether framework is, to our knowledge, the only single-substance model that captures both regimes through one physical mechanism.

9.2.4 Cluster Merger Dynamics: Collision Velocity vs. Lensing Offset

The prediction. The two-fluid ether model predicts a correlation between the collision velocity of merging galaxy clusters and the spatial offset between weak lensing peaks and X-ray emission peaks: higher-velocity collisions should produce larger offsets.

Physical basis. The superfluid fraction $\rho_s/\rho_e$ at the time of collision depends on the pre-collision effective temperature of each subcluster. The collision velocity provides additional kinetic energy, further heating the ether into the normal phase. The relevant parameter is $T_{\text{eff}}/T_c$, which increases with both the pre-collision velocity dispersion $\sigma$ and the collision velocity $v_{\text{coll}}$.

For high-velocity collisions ($v_{\text{coll}} \gg \sigma_c$, as in the Bullet Cluster with $v_{\text{coll}} \approx 4700$ km/s): the ether is driven deep into the normal phase. The normal component is collisionless and passes through, producing large lensing–X-ray offsets. This is what is observed [64].

For low-velocity collisions ($v_{\text{coll}} \lesssim \sigma_c$): a significant superfluid fraction remains. The superfluid component responds to the total gravitational potential, including the gas, and does not separate from it during the collision. The result: smaller (or zero) lensing–X-ray offsets, and possible "dark cores" coincident with the gas concentration.

Quantitative framework. For a cluster merger at collision velocity $v_{\text{coll}}$ between subclusters with pre-collision velocity dispersions $\sigma_1$ and $\sigma_2$, the effective post-collision temperature is approximately:

$$T_{\text{eff}}^{(\text{post})} \sim m_e\!\left(\sigma_{\text{pre}}^2 + \frac{v_{\text{coll}}^2}{3}\right)/k_B \tag{9.8}$$

where $\sigma_{\text{pre}} = \max(\sigma_1, \sigma_2)$ and the factor $1/3$ accounts for the three-dimensional equipartition of the collision kinetic energy. The superfluid fraction in the post-collision state is:

$$\frac{\rho_s}{\rho_e}\bigg|_{\text{post}} \approx \left[1 - \left(\frac{T_{\text{eff}}^{(\text{post})}}{T_c}\right)^\alpha\right]_+ \tag{9.9}$$

where $[\cdot]_+ = \max(\cdot, 0)$. For the Bullet Cluster ($\sigma_{\text{pre}} \sim 1200$ km/s, $v_{\text{coll}} \sim 4700$ km/s), (9.8) gives $T_{\text{eff}}^{(\text{post})}/T_c \gg 1$, so $\rho_s/\rho_e \approx 0$ and the offset is maximal. For Abell 520 (lower $\sigma_{\text{pre}}$, lower $v_{\text{coll}}$), $T_{\text{eff}}^{(\text{post})}/T_c$ is lower, and a non-negligible superfluid fraction produces a reduced offset.

The observable. Define the lensing–X-ray offset $\Delta_{\text{LX}}$ as the angular separation between the dominant weak lensing peak and the nearest X-ray emission peak, normalised by the cluster virial radius. The prediction is:

$$\Delta_{\text{LX}} \propto 1 - \frac{\rho_s}{\rho_e}\bigg|_{\text{post}} \tag{9.10}$$

A systematic survey of cluster mergers spanning a range of collision velocities would test whether $\Delta_{\text{LX}}$ correlates with $v_{\text{coll}}$ (or equivalently, with the pre-collision temperature $T_{\text{eff}}/T_c$) as (9.10) predicts.

Current observational status. The available sample of well-studied cluster mergers is small: the Bullet Cluster [64], Abell 520 [81], the Musket Ball Cluster [140], MACS J0025.4–1222 [141], and a handful of others. The existing data are qualitatively consistent with the prediction — the Bullet Cluster (high $v_{\text{coll}}$) shows large offsets, while Abell 520 (lower $v_{\text{coll}}$) shows a dark core — but the sample is too small for a quantitative test.

Future tests. The eROSITA All-Sky Survey, the Euclid space mission, and the Vera C. Rubin Observatory Legacy Survey of Space and Time (LSST) are expected to identify thousands of merging clusters with both X-ray and weak lensing data. A sample of $\sim 50$–$100$ well-characterised mergers, with independent estimates of $v_{\text{coll}}$ from shock Mach numbers and subcluster separations, would be sufficient to test the predicted $\Delta_{\text{LX}}$–$v_{\text{coll}}$ correlation at high significance.

What is structural vs. parametric. The existence of the correlation is structural: any superfluid ether model predicts that higher collision velocities drive more ether into the normal phase, increasing the lensing offset. The quantitative relationship between $\Delta_{\text{LX}}$ and $v_{\text{coll}}$ depends on $T_c$ (hence $m_e$), making the precise correlation curve a Tier 1 prediction.

What $\Lambda$CDM and MOND predict. In $\Lambda$CDM, dark matter is always collisionless; $\Delta_{\text{LX}}$ depends only on the geometry and mass ratio of the merger, not on a phase transition. There is no predicted correlation between collision velocity and the nature of the dark matter–baryon coupling. In MOND without dark matter, large lensing offsets in high-velocity mergers are difficult to explain at all, since there is no collisionless dark component to separate from the gas. The ether prediction is therefore discriminating: a positive correlation between $v_{\text{coll}}$ and $\Delta_{\text{LX}}$ of the form (9.10) would support the ether model, while the absence of such a correlation (with $\Delta_{\text{LX}}$ depending only on geometry) would support $\Lambda$CDM.

9.3 One-Parameter Predictions: The $m_e$-Dependent Programme

The structural predictions of Section 9.2 are parameter-independent: they hold for any values of the five ether constants. This section addresses the complementary class — predictions that become fully determined once a single parameter, the ether quantum mass $m_e$, is specified. These are the framework's most experimentally accessible predictions, and the cross-constraints between them constitute its strongest empirical argument.

9.3.1 The Constraint Chain: From $m_e$ to Observables

We first derive the algebraic chain that maps $m_e$ to all gravitational-sector observables. This chain uses the three observational constraints ((9.1)–(9.3)) to eliminate the other three gravitational-sector parameters ($\hat{\mu}$, $g_{\text{int}}$, baryon–ether coupling), leaving predictions as explicit functions of $m_e$ alone.

Step 1: Chemical potential $\hat{\mu}(m_e)$. Constraint I ((9.1)) fixes the combination $m_e^{3/2}\hat{\mu}^{5/2}$:

$$m_e^{3/2}\hat{\mu}^{5/2} = \frac{\hbar^3\rho_\Lambda}{C_{\text{num}}} \equiv \mathcal{C}_\Lambda \tag{9.11}$$

where $C_{\text{num}} = 0.0146$ is the exact numerical coefficient from the Bogoliubov integral ((4.120)) and $\rho_\Lambda = (6.36 \pm 0.07) \times 10^{-10}$ J/m$^3$ [7]. Solving for $\hat{\mu}$:

$$\hat{\mu}(m_e) = \left(\frac{\mathcal{C}_\Lambda}{m_e^{3/2}}\right)^{2/5} = \mathcal{C}_\Lambda^{2/5}\,m_e^{-3/5} \tag{9.12}$$

Evaluating the constant:

$$\mathcal{C}_\Lambda = \frac{(1.0546 \times 10^{-34})^3 \times 5.36 \times 10^{-10}}{0.0146} = 4.306 \times 10^{-110}\;\text{kg}^{3/2}\,\text{J}^{5/2} \tag{9.13}$$

For the fiducial $m_e = 1$ eV/$c^2 = 1.783 \times 10^{-36}$ kg:

$$\hat{\mu}(1\;\text{eV}) = \left(\frac{4.306 \times 10^{-110}}{(1.783 \times 10^{-36})^{3/2}}\right)^{2/5} = (1.809 \times 10^{-56})^{0.4} = 5.046 \times 10^{-23}\;\text{J} = 0.315\;\text{meV} \tag{9.14}$$

Step 2: Healing length $\xi(m_e)$. From the Gross–Pitaevskii equation (Section 4.3.4):

$$\xi(m_e) = \frac{\hbar}{\sqrt{2m_e\hat{\mu}(m_e)}} = \frac{\hbar}{\sqrt{2}\;\mathcal{C}_\Lambda^{1/5}}\;m_e^{-1/5} \tag{9.15}$$

The scaling $\xi \propto m_e^{-1/5}$ follows from substituting (9.12): the exponent of $m_e$ in $m_e\hat{\mu} = \mathcal{C}_\Lambda^{2/5}\,m_e^{2/5}$ is $2/5$, hence $\xi \propto m_e^{-1/5}$.

Numerically:

$$\xi(1\;\text{eV}) = \frac{1.055 \times 10^{-34}}{\sqrt{2 \times 1.783 \times 10^{-36} \times 5.046 \times 10^{-23}}} = \frac{1.055 \times 10^{-34}}{1.342 \times 10^{-29}} = 7.86\;\mu\text{m} \tag{9.16}$$

Step 3: Sound speed $c_s(m_e)$. From the phonon dispersion relation ((4.109)):

$$c_s(m_e) = \sqrt{\frac{\hat{\mu}(m_e)}{m_e}} = \mathcal{C}_\Lambda^{1/5}\,m_e^{-4/5} \tag{9.17}$$

The exponent $-4/5$ follows from $\hat{\mu}/m_e \propto m_e^{-3/5}/m_e = m_e^{-8/5}$, hence $c_s \propto m_e^{-4/5}$.

Numerically:

$$c_s(1\;\text{eV}) = \sqrt{\frac{5.046 \times 10^{-23}}{1.783 \times 10^{-36}}} = \sqrt{2.831 \times 10^{13}} = 5.321 \times 10^6\;\text{m/s} = 0.0177\,c \tag{9.18}$$

Step 4: Number density $n_0$ and interaction coupling $g_{\text{int}}(m_e)$. Constraint II ((9.2)) gives the cosmological mean ether number density:

$$n_0(m_e) = \frac{\Omega_{\text{DM}}\,\rho_{\text{crit}}}{m_e} = \frac{2.218 \times 10^{-27}}{m_e}\;\text{m}^{-3} \tag{9.19}$$

where $\Omega_{\text{DM}}\rho_{\text{crit}} = 0.26 \times 8.53 \times 10^{-27} = 2.218 \times 10^{-27}$ kg/m$^3$. The interaction coupling follows:

$$g_{\text{int}}(m_e) = \frac{\hat{\mu}}{n_0} = \frac{\mathcal{C}_\Lambda^{2/5}}{\Omega_{\text{DM}}\rho_{\text{crit}}}\;m_e^{2/5} \tag{9.20}$$

Step 5: Scattering length $a_s(m_e)$. From $g_{\text{int}} = 4\pi\hbar^2 a_s/m_e$:

$$a_s(m_e) = \frac{g_{\text{int}}\,m_e}{4\pi\hbar^2} = \frac{\mathcal{C}_\Lambda^{2/5}}{4\pi\hbar^2\,\Omega_{\text{DM}}\rho_{\text{crit}}}\;m_e^{7/5} \tag{9.21}$$

The steep scaling $a_s \propto m_e^{7/5}$ means that the scattering length is highly sensitive to $m_e$: a factor-of-10 change in $m_e$ produces a factor-of-250 change in $a_s$.

Numerically:

$$a_s(1\;\text{eV}) = \frac{4.056 \times 10^{-32} \times 1.783 \times 10^{-36}}{4\pi \times (1.055 \times 10^{-34})^2} = \frac{7.231 \times 10^{-68}}{1.398 \times 10^{-67}} = 0.517\;\text{m} \tag{9.22}$$

Step 6: BEC critical temperature $T_c$. The standard BEC formula ((4.78)) gives:

$$k_B T_c = \frac{2\pi\hbar^2}{m_e}\!\left(\frac{n_s}{\zeta(3/2)}\right)^{2/3} \tag{9.23}$$

where $\zeta(3/2) = 2.612$ and $n_s$ is the local ether number density at the point where the phase transition is being evaluated.

This introduces an important subtlety. The cosmological constraints ((9.1)–(9.2)) fix the mean ether density $n_0 = \Omega_{\text{DM}}\rho_{\text{crit}}/m_e$. But the BEC critical temperature within a gravitationally bound system depends on the local density, which is enhanced by the overdensity factor $\delta$:

$$n_s^{(\text{local})} = \delta\,n_0(m_e) \tag{9.24}$$

where $\delta = \rho_{\text{local}}/\bar{\rho}_e$ is the ratio of local to mean ether density. For a virialized halo, $\delta$ ranges from $\sim 200$ at the virial radius to $\sim 10^4$–$10^5$ in the core, depending on the density profile.

The critical velocity dispersion — defined by $k_B T_c \equiv m_e\sigma_c^2$ — is therefore:

$$\sigma_c(m_e, \delta) = \frac{\hbar\sqrt{2\pi}}{m_e}\!\left(\frac{\delta\,\Omega_{\text{DM}}\rho_{\text{crit}}}{m_e\,\zeta(3/2)}\right)^{1/3} = \frac{\hbar\sqrt{2\pi}\,\delta^{1/3}}{m_e^{4/3}}\!\left(\frac{\Omega_{\text{DM}}\rho_{\text{crit}}}{\zeta(3/2)}\right)^{1/3} \tag{9.25}$$

At the cosmological mean density ($\delta = 1$), this gives $\sigma_c \propto m_e^{-4/3}$. The Berezhiani–Khoury estimate $\sigma_c \approx 500$ km/s for $m_e = 1$ eV [71] corresponds to a local overdensity $\delta \approx 80$, consistent with the mean density within the inner $\sim 10$ kpc of a typical galactic halo.

Summary of scaling laws. Collecting the power-law dependences:

$$\hat{\mu} \propto m_e^{-3/5}, \quad \xi \propto m_e^{-1/5}, \quad c_s \propto m_e^{-4/5}, \quad n_0 \propto m_e^{-1}, \quad a_s \propto m_e^{+7/5}, \quad \sigma_c \propto \delta^{1/3}\,m_e^{-4/3} \tag{9.26}$$

The prediction web is tightly constrained: $\xi$ depends weakly on $m_e$ ($\xi$ changes by only a factor of 1.6 as $m_e$ ranges from 0.3 to 3 eV), while $a_s$ and $\sigma_c$ depend strongly ($a_s$ changes by a factor of $\sim 250$, $\sigma_c$ by a factor of $\sim 22$ over the same range).

9.3.2 Sub-Millimetre Gravity: The Healing Length as Yukawa Range

The prediction. At distances $r \lesssim \xi$, the Bogoliubov dispersion ((4.108)) departs from the linear phonon regime ($\omega = c_s k$) and enters the free-particle regime ($\omega = \hbar k^2/(2m_e)$). Modes with $k > 1/\xi$ acquire an effective mass gap, and their exchange produces a Yukawa-type modification to the gravitational potential (Section 4.3.10):

$$V(r) = -\frac{Gm_1 m_2}{r}\!\left(1 + \alpha_\xi\,e^{-r/\xi}\right) \tag{4.170}$$

The Yukawa range is $\xi(m_e) = \hbar/\sqrt{2m_e\hat{\mu}(m_e)}$. The coupling $\alpha_\xi$ parameterises the ratio of the phonon-mediated gravitational interaction to the direct (metric) gravitational interaction at the healing length scale.

Derivation of $\alpha_\xi$. The coupling $\alpha_\xi$ is estimated as follows. At distances $r \gg \xi$, gravity is carried by long-wavelength phonon exchange, which reproduces Newtonian gravity in the PG framework (Theorem 3.2). At $r \lesssim \xi$, the phonon propagator acquires a mass $m_{\text{eff}} = \hbar/(c_s\xi) = \sqrt{2m_e\hat{\mu}}/c_s = \sqrt{2}\,m_e$ (using $c_s = \sqrt{\hat{\mu}/m_e}$). The massive propagator produces a Yukawa potential with range $\xi$ and coupling of order $G_{\text{phonon}}/G_{\text{Newton}} \sim \mathcal{O}(1)$, since the phonon-mediated interaction is the same interaction that produces Newtonian gravity at long distances. We therefore expect $\alpha_\xi \sim \mathcal{O}(1)$, though a first-principles derivation from the full Bogoliubov Green's function — which would fix $\alpha_\xi$ precisely — has not been completed. We identify this calculation as a priority for future work.

Current experimental constraints. The most sensitive tests of the gravitational inverse-square law at sub-millimetre scales come from torsion balance experiments. The Eöt-Wash group [70] constrains:

$$|\alpha_\xi| < 2.5 \;\;\text{at}\;\; \xi = 52\;\mu\text{m}, \qquad |\alpha_\xi| < 44 \;\;\text{at}\;\; \xi = 25\;\mu\text{m} \tag{4.171}$$

At the fiducial ether prediction $\xi = 7.9\;\mu$m, the current bound is approximately $|\alpha_\xi| < 10^4$, which does not constrain $\alpha_\xi \sim \mathcal{O}(1)$.

The experimental frontier is advancing. The CANNEX (Casimir And Non-Newtonian force EXperiment) collaboration [88] aims to probe Yukawa deviations down to $\xi \sim 1\;\mu$m with $|\alpha_\xi| \sim 1$ sensitivity. The IUPUI short-range gravity experiment [142] targets similar scales using a parallel-plate geometry optimised for sub-10 $\mu$m ranges.

$m_e$-dependence. The healing length varies slowly with $m_e$ (scaling as $m_e^{-1/5}$):

$m_e$ (eV)	$\hat{\mu}$ (meV)	$\xi$ ($\mu$m)	Current constraint on $|\alpha_\xi|$	Status
0.3	0.649	10.0	$\sim 10^3$	Unconstrained
0.5	0.477	9.0	$\sim 5 \times 10^3$	Unconstrained
1.0	0.315	7.9	$\sim 10^4$	Unconstrained
2.0	0.208	6.8	$\sim 5 \times 10^4$	Unconstrained
3.0	0.163	6.3	$\sim 10^5$	Unconstrained

For the entire range $m_e = 0.3$–$3$ eV, the prediction falls within $\xi = 6$–$10\;\mu$m — a narrow window that is technologically accessible with next-generation experiments but beyond current reach.

Falsification condition. If future experiments establish $|\alpha_\xi| < 1$ at all scales $\xi > 5\;\mu$m, the entire range $m_e = 0.3$–$3$ eV is excluded. The ether framework would survive only for $m_e > 3$ eV (pushing $\xi$ below 6 $\mu$m) or for $\alpha_\xi \ll 1$ (requiring the phonon-mediated interaction to be anomalously suppressed at short range, which would require explanation). If $|\alpha_\xi| < 0.01$ is established down to $\xi \sim 1\;\mu$m, the sub-millimetre gravity prediction is falsified for all astrophysically motivated values of $m_e$.

9.3.3 The Phase Transition Scale in Galaxy Groups

The prediction. The superfluid-to-normal phase transition (Section 4.2.7) produces a transition in dark-sector phenomenology from MOND-like (superfluid) to CDM-like (normal). The characteristic velocity dispersion at which this transition occurs is $\sigma_c(m_e, \delta)$ ((9.25)), which depends on both $m_e$ and the local ether overdensity $\delta$.

At the cosmological mean density ($\delta = 1$), (9.25) gives:

$$\sigma_c^{(\text{cosmo})}(m_e) = \frac{\hbar\sqrt{2\pi}}{m_e^{4/3}}\!\left(\frac{\Omega_{\text{DM}}\rho_{\text{crit}}}{\zeta(3/2)}\right)^{1/3} \tag{9.27}$$

$m_e$ (eV)	$\sigma_c^{(\text{cosmo})}$ (km/s)	$\sigma_c(\delta = 50)$ (km/s)	$\sigma_c(\delta = 100)$ (km/s)
0.3	577	2126	2677
0.5	292	1076	1355
1.0	116	427	538
2.0	46	170	214
3.0	27	99	124

The observational requirement that spiral galaxies ($\sigma \sim 100$–$300$ km/s) are predominantly superfluid while galaxy clusters ($\sigma \sim 800$–$1500$ km/s) are predominantly normal constrains the local critical dispersion to $\sigma_c^{(\text{local})} \sim 300$–$800$ km/s ((4.96)). From the table, this is satisfied for $m_e \sim 0.5$–$1.5$ eV at typical halo overdensities $\delta \sim 50$–$100$.

The observable. The mass discrepancy — defined as the ratio $M_{\text{dyn}}/M_{\text{bar}}$ of dynamical to baryonic mass — should vary systematically with velocity dispersion across the galaxy-to-cluster transition:

• For $\sigma \ll \sigma_c$: $M_{\text{dyn}}/M_{\text{bar}}$ follows the MOND prediction, $\sim (a_0 R/\sigma^2)^{1/2}$ in the deep-MOND regime
• For $\sigma \gg \sigma_c$: $M_{\text{dyn}}/M_{\text{bar}} \to 1 + \Omega_{\text{DM}}/\Omega_b \approx 6.2$ ((4.95))
• For $\sigma \sim \sigma_c$: intermediate values, with the transition governed by $(T_{\text{eff}}/T_c)^\alpha$ ((4.77))

Galaxy groups with $\sigma \sim 300$–$600$ km/s are the natural laboratory. A systematic study of the mass discrepancy as a function of $\sigma$ across $\sim 50$–$100$ groups would map the transition curve and constrain $\sigma_c^{(\text{local})}$, which — combined with a density model — determines $m_e$.

Existing data. Evidence for a MOND–CDM transition in galaxy groups is mixed. Several studies [143, 144] find that low-mass groups ($M \lesssim 10^{13}\,M_\odot$, $\sigma \lesssim 300$ km/s) follow the MOND prediction more closely than high-mass groups. However, the scatter is large, and the data do not yet cleanly map a transition curve. The eROSITA All-Sky Survey and the 4MOST spectroscopic survey, which will characterise thousands of galaxy groups, offer the prospect of a definitive measurement.

What $\Lambda$CDM predicts. In $\Lambda$CDM, there is no phase transition. The mass discrepancy $M_{\text{dyn}}/M_{\text{bar}}$ should approach the cosmic ratio $\Omega_m/\Omega_b \approx 6.2$ at all scales where dark matter dominates, with scatter determined by halo formation history. There is no predicted systematic transition from MOND-like to CDM-like behaviour as a function of $\sigma$. Observational discovery of such a transition would be strong evidence for a phase-transition model.

9.3.4 Modified Photon Dispersion: The EM Sector

The prediction. The ether's transverse microstructure at scale $\ell_e$ modifies the electromagnetic dispersion relation (Section 3.8.2):

$$\omega^2 = c^2 k^2\!\left(1 + \xi_2\,(k\ell_e)^2 + \mathcal{O}((k\ell_e)^4)\right) \tag{3.46}$$

where $\xi_2 = -1/12$ for the simplest lattice model. This produces an energy-dependent group velocity:

$$v_g = c\!\left(1 + \frac{3\xi_2}{2}\frac{E^2\ell_e^2}{\hbar^2 c^2}\right) \tag{3.47}$$

and a time delay between photons of energies $E_1$ and $E_2$ from a source at cosmological distance $d$:

$$\Delta t = \frac{3|\xi_2|\,d\,\ell_e^2}{2\hbar^2 c^3}\!\left(E_1^2 - E_2^2\right) \tag{3.48}$$

Parameter independence from the gravitational sector. The scale $\ell_e$ governs the UV structure of transverse (EM) modes, which propagate at speed $c$ in the ether metric. It is logically independent of the healing length $\xi$, which governs the UV structure of longitudinal (phonon) modes propagating at speed $c_s$. A measurement of $\ell_e$ provides no information about $m_e$, and vice versa.

This independence is both an honest gap and a falsifiable structural feature: the framework predicts that sub-millimetre gravity experiments (sensitive to $\xi$) and gamma-ray dispersion experiments (sensitive to $\ell_e$) probe different physical scales. If a future theory of ether microphysics predicts a relationship $\ell_e = f(\xi, m_e, \hat{\mu})$, this becomes an additional testable constraint.

Current constraints. The Fermi-LAT observation of GRB 090510 [51] and MAGIC observations of Mrk 501 [52] constrain the quadratic dispersion coefficient:

$$|\xi_2|\,\ell_e^2 < 3.2 \times 10^{-26}\;\text{m}^2 \tag{3.49}$$

For $\xi_2 = -1/12$ (the lattice model value):

$$\ell_e < 6.2 \times 10^{-13}\;\text{m} \tag{9.28}$$

This is $\sim 10^{22}$ times larger than the Planck length, leaving an enormous observational window.

Scaling with $\ell_e$. The time delay ((3.48)) scales as $\Delta t \propto \ell_e^2 E^2 d$. The observable signal therefore increases rapidly with $\ell_e$:

$\ell_e$	$|\xi_2|\,\ell_e^2$ (m$^2$)	Ratio to current bound	Status
$\ell_P = 1.6 \times 10^{-35}$ m	$2.1 \times 10^{-71}$	$6.7 \times 10^{-46}$	Undetectable
$10^{-20}$ m	$8.3 \times 10^{-42}$	$2.6 \times 10^{-16}$	Undetectable
$10^{-14}$ m	$8.3 \times 10^{-30}$	$2.6 \times 10^{-4}$	Near CTA threshold
$6.2 \times 10^{-13}$ m	$3.2 \times 10^{-26}$	$1.0$	Current bound

If $\ell_e$ is near the Planck scale, the prediction is not testable with any foreseeable technology — the signal is $\sim 10^{45}$ times below current sensitivity. The Cherenkov Telescope Array (CTA) [53] is projected to improve sensitivity by approximately one order of magnitude, reaching $|\xi_2|\,\ell_e^2 \sim 3 \times 10^{-27}$ m$^2$ and probing $\ell_e \sim 2 \times 10^{-13}$ m for $|\xi_2| \sim 0.1$. Significant detection would require $\ell_e \gtrsim 10^{-14}$ m — far above the Planck scale but far below nuclear scales.

Remark on the sign. For $\xi_2 < 0$ (as in the lattice model), higher-energy photons travel slower — a subluminal dispersion. This is the natural expectation for a medium with discrete structure (the lattice has a maximum frequency $\omega_{\max} = 2c/\ell_e$, and group velocity decreases toward this cutoff). A measurement of the sign of the time delay (whether high-energy photons arrive first or last) would test the lattice model's prediction $\xi_2 < 0$.

Falsification condition. If CTA or future gamma-ray observatories detect energy-dependent time delays with $\xi_1 \neq 0$ (linear, CPT-violating dispersion), this would indicate a parity-asymmetric microstructure not predicted by the simplest ether models. If quadratic dispersion is detected with $\xi_2 > 0$ (superluminal), the lattice-type microstructure is excluded.

9.3.5 The Cross-Prediction Web: Quantitative Tables

The central empirical argument of the ether framework is that a single measurement — of $\xi$ in a sub-millimetre gravity experiment, or of $\sigma_c$ in a galaxy-group survey — would simultaneously determine predictions across all gravitational-sector observables with no remaining freedom. We now make this argument quantitative.

Table 9.1: Cross-predictions as functions of $m_e$. All quantities derived from the algebraic chain of Section 9.3.1. Observational constraints I and II are exactly satisfied for every row; the verification column confirms $\rho_{\text{ZPF}}/\rho_\Lambda^{\text{obs}} = 1.000$ and $m_e n_0/\rho_{\text{DM}} = 1.000$ to machine precision.

$m_e$ (eV)	$\hat{\mu}$ (meV)	$\xi$ ($\mu$m)	$c_s$ ($10^6$ m/s)	$c_s/c$	$n_0$ ($10^9$ m$^{-3}$)	$a_s$ (m)	$\sigma_c^{(\delta=80)}$ (km/s)
0.3	0.649	10.0	13.94	0.0465	4.15	0.096	2484
0.5	0.477	9.0	9.26	0.0309	2.49	0.196	1257
1.0	0.315	7.9	5.32	0.0177	1.24	0.517	499
2.0	0.208	6.8	3.06	0.0102	0.62	1.365	198
3.0	0.163	6.3	2.21	0.0074	0.41	2.409	115

The fiducial row ($m_e = 1.0$ eV, boldface) reproduces all values derived in Section 4.3.8–4.3.10. The column $\sigma_c^{(\delta=80)}$ uses the Berezhiani–Khoury reference overdensity $\delta = 80$ (corresponding to a typical galactic core); as demonstrated in (9.25), $\sigma_c$ scales as $\delta^{1/3}$.

Table 9.2: If $\xi$ is measured. Suppose a sub-millimetre gravity experiment detects a Yukawa deviation at range $\xi_{\text{meas}}$. The constraint chain (9.15) combined with Constraint I (9.11) uniquely determines $m_e$ and $\hat{\mu}$. Explicitly: from $\xi = \hbar/\sqrt{2m_e\hat{\mu}}$, the product $m_e\hat{\mu} = \hbar^2/(2\xi^2)$. Combined with $m_e^{3/2}\hat{\mu}^{5/2} = \mathcal{C}_\Lambda$, we eliminate $\hat{\mu}$:

$$m_e\hat{\mu} = \frac{\hbar^2}{2\xi^2}, \qquad \hat{\mu} = \frac{\hbar^2}{2\xi^2 m_e} \tag{9.29}$$

Substituting into $m_e^{3/2}\hat{\mu}^{5/2} = \mathcal{C}_\Lambda$:

$$m_e^{3/2}\!\left(\frac{\hbar^2}{2\xi^2 m_e}\right)^{5/2} = \mathcal{C}_\Lambda \tag{9.30}$$ $$\frac{\hbar^5}{2^{5/2}\xi^5}\;m_e^{3/2-5/2} = \mathcal{C}_\Lambda \tag{9.31}$$ $$m_e^{-1} = \frac{\mathcal{C}_\Lambda \cdot 2^{5/2}\xi^5}{\hbar^5} \tag{9.32}$$ $$\boxed{m_e(\xi) = \frac{\hbar^5}{\mathcal{C}_\Lambda \cdot 4\sqrt{2}\;\xi^5}} \tag{9.33}$$

The steep scaling $m_e \propto \xi^{-5}$ means that the healing length is a highly sensitive probe of $m_e$. Once $m_e$ is determined, $\hat{\mu}$ follows from (9.29), and all other quantities from the chain (9.17)–(9.25).

$\xi_{\text{meas}}$ ($\mu$m)	$m_e$ (eV)	$\hat{\mu}$ (meV)	$c_s$ ($10^6$ m/s)	$\sigma_c^{(\delta=80)}$ (km/s)
6	3.86	0.140	1.80	82
8	0.92	0.332	5.70	560
10	0.30	0.648	13.93	2480
12	0.12	1.120	28.9	8360
15	0.04	2.187	70.5	37000

The table reveals a critical observational window: if $\xi$ is measured to be $\sim 7$–$9\;\mu$m, the resulting $m_e$ is $\sim 0.5$–$2$ eV, and the critical dispersion $\sigma_c^{(\delta=80)}$ falls in the astrophysically required range $\sim 200$–$1300$ km/s. If $\xi > 10\;\mu$m, the implied $m_e < 0.3$ eV gives $c_s > 0.05c$ and $\sigma_c^{(\delta=80)} > 2400$ km/s — even galaxy clusters would retain significant superfluid fraction, contradicting observation. If $\xi < 6\;\mu$m, the implied $m_e > 4$ eV gives $\sigma_c^{(\delta=80)} < 100$ km/s — even dwarf galaxies would be in the normal phase, eliminating the MOND phenomenology.

The viable window is $\xi \approx 6$–$10\;\mu$m, equivalently $m_e \approx 0.3$–$4$ eV, with the observationally preferred range $\xi \approx 7$–$9\;\mu$m ($m_e \approx 0.5$–$2$ eV). This is a strong, falsifiable constraint: the framework permits only a narrow range of $\xi$ values, determined jointly by cosmological observations and galactic dynamics.

Table 9.3: Summary of observational constraints on $m_e$.

Constraint	Observable	Required range	Implied $m_e$ range
Galaxies are superfluid	$\sigma_c^{(\text{local})} > 300$ km/s	$T_c$ high enough for spirals	$m_e \lesssim 2$ eV
Clusters are normal	$\sigma_c^{(\text{local})} < 1500$ km/s	$T_c$ low enough for clusters	$m_e \gtrsim 0.3$ eV
Berezhiani–Khoury MOND	Phonon force matches $a_0$	$m_e \sim 1$–$2$ eV [71]	$m_e \sim 1$–$2$ eV
Sub-mm gravity	$\xi$ within experimental reach	$\xi \sim 6$–$10\;\mu$m	$m_e \sim 0.3$–$4$ eV
Combined	All of the above	Intersection	$m_e \approx 0.5$–$2$ eV

The convergence of independent constraints to $m_e \approx 0.5$–$2$ eV is a non-trivial consistency check. The ether framework does not have the freedom to choose $m_e$ independently for each prediction — a single value must satisfy all constraints simultaneously. That such a value exists, and falls in a narrow range, is itself an empirical success.

9.3.6 Hierarchy of Sensitivity

Not all observables are equally useful for determining $m_e$. The scaling laws (9.26) define a hierarchy of experimental sensitivity:

Observable	Scaling	Sensitivity to factor-of-2 change in $m_e$
$\xi$	$m_e^{-1/5}$	13% change
$\hat{\mu}$	$m_e^{-3/5}$	34% change
$c_s$	$m_e^{-4/5}$	43% change
$\sigma_c$	$m_e^{-4/3}$	factor 2.5 change
$a_s$	$m_e^{+7/5}$	factor 2.6 change

The healing length $\xi$ is the least sensitive to $m_e$ — it varies by only a factor of $\sim 1.6$ across the entire viable range $m_e = 0.3$–$3$ eV. This is both a strength (the prediction is robust) and a weakness (a measurement of $\xi$ provides limited precision on $m_e$ unless the experimental uncertainty is small). Conversely, $\sigma_c$ and $a_s$ are highly sensitive, making galaxy-group surveys potentially the most informative probe — though the additional dependence of $\sigma_c$ on $\delta$ ((9.25)) complicates the interpretation.

The optimal experimental strategy combines a sub-millimetre gravity measurement of $\xi$ (robust, weakly dependent on $m_e$) with a galaxy-group measurement of the transition scale (strongly dependent on $m_e$, moderately model-dependent through $\delta$). Consistency between these two independent determinations of $m_e$ would constitute strong evidence for the ether framework.

9.4 Discrimination, Experimental Programme, and Falsification

Sections 8.1–9.3 developed the ether framework's predictions in three tiers: structural (parameter-free), one-parameter ($m_e$-dependent), and two-parameter (adding $\ell_e$). This section performs three tasks that the preceding sections deliberately deferred. First, we present the thermal Bell degradation — the one prediction that directly confronts quantum foundations rather than cosmology. Second, we construct a discrimination matrix comparing the ether framework against $\Lambda$CDM and MOND across all testable predictions. Third, we rank the experimental programme by feasibility and discriminating power.

9.4.1 Thermal Bell Degradation: The Quantum Foundations Test

The predictions of Sections 8.2–9.3 concern the gravitational and cosmological sectors: dark energy, dark matter, galactic dynamics, sub-millimetre gravity. These are domains where the ether framework offers an alternative explanation for phenomena that $\Lambda$CDM and MOND also address (with varying success). The thermal Bell prediction is qualitatively different: it is a prediction about quantum mechanics itself, where the ether framework disagrees with standard quantum theory.

The prediction. In a Bell test using entangled particles at frequency $\omega$ and ambient temperature $T$, the CHSH parameter degrades as (Theorem 8.8, corrected):

$$|S(T)| = \frac{2\sqrt{2}}{(1 + 2n_{\text{th}}(\omega, T))^2} \tag{9.34}$$

where $n_{\text{th}}(\omega, T) = (e^{\hbar\omega/(k_BT)} - 1)^{-1}$ is the Bose–Einstein thermal occupation number at the entangled pair's frequency.

Bell violation ($|S| > 2$) persists only for $T < T_{\text{crit}}(\omega)$, where:

$$T_{\text{crit}}(\omega) = \frac{\hbar\omega}{k_B\ln\!\left(1 + \frac{2}{2^{1/4} - 1}\right)} = \frac{\hbar\omega}{2.449\,k_B} \tag{9.35}$$

Derivation summary. The argument (Section 8.7, revised) proceeds as follows:

(i) At $T = 0$, the Nelson–SED bridge (Theorem 7.1) guarantees that the ether reproduces all quantum measurement statistics, including $|S| = 2\sqrt{2}$ for the singlet state (Theorem 8.5). The mechanism is the nonlocal osmotic coupling in configuration space: the osmotic velocity $u_A = -D\cot(\phi_B - \phi_A)$ for the singlet state depends on the remote particle's polarisation ((8.58)).

(ii) At $T > 0$, each detector's electromagnetic mode contains both ZPF photons (carrying entanglement, with occupation $n_{\text{ZPF}} = 1/2$) and thermal photons (local noise, with occupation $n_{\text{th}}$). The detector cannot distinguish them. The probability that a detection event is signal-triggered (governed by Nelson dynamics) rather than thermal (random $\pm 1$) is:

$$1 - p = \frac{n_{\text{ZPF}}}{n_{\text{ZPF}} + n_{\text{th}}} = \frac{1}{1 + 2n_{\text{th}}} \tag{9.36}$$

(iii) The thermal fields at Alice's and Bob's detectors are statistically independent for macroscopic separations $d \gg \xi_{\text{th}} = \hbar c/(k_BT)$ (Theorem 8.7: the thermal correlations decay exponentially on scale $\xi_{\text{th}}$, while the ZPF correlations decay as a power law). Therefore, the signal-vs-thermal discrimination at each detector is independent. The coincidence correlation carries the product of both signal fractions:

$$E(\Delta; T) = (1 - p)^2\,E_{\text{singlet}}(\Delta) = \frac{-\cos(2\Delta)}{(1 + 2n_{\text{th}})^2} \tag{9.37}$$

The CHSH combination gives (9.34). The exponent 2 (not 1) arises from the independence of the two detectors' thermal environments.

What standard quantum mechanics predicts. In QM, Bell violations arise from the entangled state itself and do not depend on the ZPF or any ambient medium. Thermal decoherence degrades the visibility through Lindblad dynamics [135]:

$$|S_{\text{QM}}(T)| = 2\sqrt{2}\exp\!\left(-\gamma_0\tau\,n_{\text{th}}(\omega, T)\right) \tag{9.38}$$

where $\gamma_0\tau$ is an implementation-specific decoherence parameter. The QM prediction contains a free parameter ($\gamma_0\tau$); the ether prediction (9.34) does not. We stress that the standard QM prediction is implementation-dependent: the Lindblad master equation gives exponential decay for Markovian environments (the generic case for superconducting circuits coupled to thermal baths [135]), but non-Markovian environments can produce non-exponential decay. The discriminating test is therefore the parameter-free ratio ((9.40) below), which distinguishes the ether's algebraic form from any monotonically decaying QM prediction, not merely the Markovian one.

The three discriminating signatures. The two predictions differ in experimentally measurable ways:

(a) Functional form. The ether prediction (9.34) is algebraic: $|S| \propto (1 + 2n_{\text{th}})^{-2}$. The QM prediction (9.38) is exponential: $|S| \propto e^{-\gamma_0\tau n_{\text{th}}}$. At high temperatures ($n_{\text{th}} \gg 1$):

$$|S_{\text{ether}}| \sim \frac{\sqrt{2}}{2n_{\text{th}}^2} \propto T^{-2}, \qquad |S_{\text{QM}}| \sim 2\sqrt{2}\,e^{-\gamma_0\tau n_{\text{th}}} \;\to\; 0\;\text{exponentially} \tag{9.39}$$

The ether prediction has a power-law tail; the QM prediction vanishes exponentially. At $T \approx 5\,T_{\text{crit}}$ for 10 GHz microwaves, $|S_{\text{ether}}|/|S_{\text{QM}}| \approx 21$ — easily measurable.

(b) Parameter-free ratio test. The ratio of CHSH values at two temperatures is:

$$R_{\text{ether}}(T_1, T_2) = \left(\frac{1 + 2n_{\text{th}}(T_2)}{1 + 2n_{\text{th}}(T_1)}\right)^{\!2} \tag{9.40}$$

This prediction has no free parameters: it depends only on $\omega$, $T_1$, and $T_2$, all of which are directly measured. The QM prediction $R_{\text{QM}} = \exp(\gamma_0\tau[n_{\text{th}}(T_2) - n_{\text{th}}(T_1)])$ contains the implementation-dependent parameter $\gamma_0\tau$. A single measurement of $R$ at two temperatures discriminates the two frameworks without fitting.

Example. For 10 GHz microwaves ($T_{\text{crit}} = 0.196$ K) at $T_1 = 0.10$ K and $T_2 = 0.50$ K:

$$R_{\text{ether}} = \left(\frac{1 + 2 \times 0.621}{1 + 2 \times 0.008}\right)^{\!2} = 4.86 \tag{9.41}$$

(c) Critical temperature. The ether predicts a sharp frequency-dependent threshold (9.35) below which Bell violation is impossible regardless of experimental isolation. For microwaves at 10 GHz: $T_{\text{crit}} = 0.196$ K. At 5 GHz: $T_{\text{crit}} = 0.098$ K. Standard QM predicts no such threshold — a sufficiently isolated system violates Bell inequalities at any temperature.

Experimental requirements. The optimal platform is a superconducting microwave circuit in a dilution refrigerator, performing a Bell test at variable temperature. The protocol:

(i) Prepare entangled microwave photons at frequency $\omega/(2\pi) = 5$–$50$ GHz using a Josephson parametric amplifier.

(ii) Measure the CHSH parameter $|S(T)|$ at a series of temperatures from $T_{\text{crit}}/20$ (deep in the quantum regime) to $\sim 5\,T_{\text{crit}}$ (deep in the thermal regime).

(iii) Fit the resulting $|S(T)|$ curve to both the ether prediction (9.34) and the QM prediction (9.38). The ether prediction has zero free parameters; the QM prediction has one ($\gamma_0\tau$).

(iv) Compute the ratio $R(T_1, T_2)$ at two temperatures separated by $\gtrsim T_{\text{crit}}$ and compare against the parameter-free prediction (9.40).

Storz et al. [132] have already demonstrated a loophole-free Bell test with superconducting circuits at $\sim 20$ mK, achieving $|S| = 2.0747 \pm 0.0033$. The required modification is a temperature sweep — technically straightforward with existing dilution refrigerator technology, which can stabilise temperatures from 10 mK to $> 1$ K.

Predicted $|S(T)|$ curve for a 10 GHz Bell test:

$T$ (K)	$T/T_{\text{crit}}$	$n_{\text{th}}$	$|S_{\text{ether}}|$	$|S_{\text{QM}}|^{*}$	Violates Bell?
0.010	0.05	$< 10^{-4}$	2.828	2.828	Yes
0.050	0.26	0.0001	2.828	2.828	Yes
0.100	0.51	0.008	2.737	2.744	Yes
0.150	0.77	0.043	2.403	2.421	Yes
0.196	1.00	0.095	2.000	2.000	Marginal
0.250	1.28	0.172	1.567	1.507	No
0.300	1.53	0.253	1.247	1.119	No
0.500	2.55	0.621	0.563	0.291	No
1.000	5.10	1.624	0.157	0.007	No

${}^{*}$QM column uses $\gamma_0\tau = 3.66$, normalised to give $|S_{\text{QM}}(T_{\text{crit}})| = 2$.

Tier classification. This prediction occupies a unique position in the tier structure:

• The existence of thermal degradation (qualitative) is Tier 0: it follows from the framework's structure (the ether is a physical medium; thermal noise exists in any medium).
• The functional form $(1 + 2n_{\text{th}})^{-2}$ and the critical temperature $T_{\text{crit}} = \hbar\omega/(2.449\,k_B)$ are Tier 0: they depend on no ether parameters — only on $\omega$ and $T$, which are experimentally controlled.
• The prediction is entirely independent of $m_e$, $\hat{\mu}$, $g_{\text{int}}$, $\ell_e$, and the baryon–ether coupling. It tests the quantum-mechanical core of the framework (the Nelson–SED bridge, Theorem 7.1) rather than the gravitational sector.

Why this prediction is fundamental. Every other prediction in this section concerns the gravitational or cosmological behaviour of the ether — domains where our knowledge of dark matter and dark energy is incomplete, and alternative explanations are plausible. The thermal Bell prediction confronts quantum mechanics on its own ground: the correlations between entangled particles, where standard QM has been tested with extraordinary precision. A confirmation of the ether prediction would require revision of the foundations of quantum theory. A falsification would disprove the specific thermal depolarisation mechanism (Theorem 8.8) while leaving the gravitational-sector predictions intact — since those depend on different aspects of the framework (the condensate equation of state, not the Nelson detection model).

---

9.4.2 Discrimination Matrix

The following table compares the ether framework against $\Lambda$CDM and MOND across all testable predictions developed in Sections 8.1–9.4.1. For each observable, we indicate: the ether prediction (with the derivation source), the $\Lambda$CDM prediction, the MOND prediction, and the discriminating power — whether a measurement of that observable could distinguish the three frameworks.

Table 9.4: Complete Discrimination Matrix

#	Observable	Ether prediction	$\Lambda$CDM	MOND	Discriminates?
Structural (Tier 0)
1	Dark energy EoS $w$	$-1$ exactly (Thm 4.2)	$-1$ (input)	No prediction	Ether ≈ $\Lambda$CDM; neither falsified
2	RAR functional form	$\mu_e = 1 - e^{-\sqrt{x}}$ ((4.58))	No specific form	Various $\mu(x)$ (postulated)	Ether vs MOND: shape test
3	RAR intrinsic scatter	$< 0.05$ dex (superfluid uniformity)	0.11 dex (simulated)	$< 0.05$ dex (expected)	Ether ≈ MOND $\neq$ $\Lambda$CDM
4	Galaxy$\to$cluster transition	Phase transition at $\sigma_c(m_e, \delta)$	No transition; CDM at all scales	No transition; MOND at all scales (fails)	Three-way
5	Cluster $M/M_b$ ratio	$\approx 6.2$ (normal-phase CDM)	$\approx 6.2$ (CDM halos)	$\approx 2$–$3$ (fails without DM)	Ether ≈ $\Lambda$CDM $\neq$ MOND
6	Merger offset $\Delta_{\text{LX}}$ vs $v_{\text{coll}}$	Correlated (phase transition, (9.10))	Uncorrelated (always collisionless)	No DM component (fails)	Three-way
7	Bell violation at $T = 0$	$|S| = 2\sqrt{2}$ (Thm 7.5)	$|S| = 2\sqrt{2}$	N/A	Agrees with QM; not discriminating
8	Bell vs temperature	$|S| = 2\sqrt{2}/(1+2n_{\text{th}})^2$ (Thm 7.8)	No degradation$^{\dagger}$	N/A	Ether vs QM
One-parameter (Tier 1)
9	Sub-mm gravity: Yukawa range	$\xi(m_e) \sim 6$–$10\;\mu$m ((9.15))	No deviation	No prediction	Ether-unique
10	Sub-mm gravity: coupling	$\alpha_\xi \sim \mathcal{O}(1)$ (estimated)	$\alpha = 0$	No prediction	Ether-unique
11	BEC critical velocity $\sigma_c$	$\sigma_c(m_e, \delta)$, (9.22)	No analogue	No analogue	Ether-unique
12	Ether sound speed $c_s$	$c_s(m_e)$, (9.17)	N/A	N/A	Ether-unique
Two-parameter (Tier 2)
13	Photon dispersion	$\Delta t \propto E^2\ell_e^2$ ((3.48))	No dispersion	No prediction	Ether-unique
Open (Tier 3)
14	CMB power spectrum	Not yet derived	6-parameter fit	Not competitive	Gap in ether programme

$^{\dagger}$ Standard QM predicts decoherence-induced degradation ((9.38)), but the functional form differs from the ether prediction: exponential vs algebraic.

Reading the matrix. The discrimination matrix reveals three classes of observables:

Class A: Ether-unique predictions (rows 8–13). These are predictions that only the ether framework makes: sub-millimetre Yukawa deviations, a BEC phase transition scale, ether sound speed, photon dispersion, and thermal Bell degradation. Positive detection of any of these would provide strong evidence for the ether framework, since neither $\Lambda$CDM nor MOND predicts them. Current experimental bounds are consistent with the ether predictions but have not yet reached the sensitivity needed for detection (Section 9.3).

Class B: Three-way discriminators (rows 4, 6). These observables — the galaxy-to-cluster transition and the merger offset correlation — yield different predictions from all three frameworks. A systematic survey of galaxy groups spanning the transition region, or a large sample of cluster mergers with lensing and X-ray data, would distinguish all three.

Class C: Partial discriminators (rows 1–3, 5, 7). These observables distinguish some but not all framework pairs. The dark energy equation of state (row 1) does not currently discriminate ether from $\Lambda$CDM (both predict $w = -1$), though the DESI hint of time-varying $w$ could become relevant. The RAR scatter (row 3) discriminates $\Lambda$CDM from both ether and MOND but cannot distinguish the latter two.

The gap (row 14). The CMB power spectrum is the one domain where $\Lambda$CDM has overwhelming empirical success and the ether framework has no prediction. This is the most significant vulnerability of the ether programme (Section 9.1.7).

---

9.4.3 Experimental Roadmap

We rank the experimental tests by three criteria: (a) feasibility — can the measurement be performed with current or near-term technology?; (b) discriminating power — how many frameworks does the result distinguish?; (c) fundamental impact — does a positive result challenge established physics?

Priority 1: Thermal Bell degradation (microwave circuits)

Criterion	Assessment
Feasibility	High. Superconducting Bell tests already demonstrated at 20 mK [132]. Temperature sweep requires only standard dilution refrigerator capability.
Discriminating power	Maximum. Distinguishes ether from standard QM — the only prediction in the programme that does so.
Fundamental impact	Transformative. Confirmation would require revision of quantum foundations.
Timeline	1–3 years with existing infrastructure.

Protocol. Replicate the Storz et al. [132] Bell test at microwave frequencies (5–50 GHz) and measure $|S(T)|$ at 10–20 temperature points from 10 mK to 1 K. Fit the data to both (9.34) and (9.38). Compute the parameter-free ratio (9.40) at two well-separated temperatures. The measurement precision required is $\delta S \lesssim 0.01$, already achieved in [132].

Decision tree. If $|S(T)|$ follows the algebraic curve (9.34): strong evidence for the ether's thermal depolarisation mechanism. If $|S(T)|$ follows the exponential curve (9.38): Theorem 8.8 is falsified; the ether framework's quantum sector requires revision, though the gravitational sector is unaffected. If $|S(T)|$ shows no temperature dependence at all up to $\sim 1$ K (after correcting for decoherence): both the ether prediction and the standard QM decoherence model are inconsistent with the data, requiring new physics.

Priority 2: Sub-millimetre gravity (torsion balance / CANNEX)

Criterion	Assessment
Feasibility	Medium. Current Eöt-Wash experiments probe $\xi \sim 25$–$52\;\mu$m. The CANNEX experiment targets $\xi \sim 1\;\mu$m. The predicted range $\xi \sim 6$–$10\;\mu$m requires intermediate-scale experiments currently in development.
Discriminating power	High. A Yukawa detection uniquely supports the ether ($\Lambda$CDM and MOND predict no sub-mm deviations). A null result at $\alpha_\xi < 1$ for $\xi > 5\;\mu$m would exclude the entire viable $m_e$ range.
Fundamental impact	Major. Detection of sub-mm gravitational deviations would be a discovery of new physics at the interface of gravity and quantum mechanics.
Timeline	3–10 years. CANNEX commissioning expected $\sim$ 2027.

What would be measured. A Yukawa range $\xi = \xi_{\text{meas}}$ would determine $m_e$ through the steep scaling $m_e \propto \xi^{-5}$ (Table 9.2, Section 9.3.5). This single measurement would fix all Tier 1 predictions simultaneously: $c_s$, $\sigma_c$, $n_0$, $a_s$ (Table 9.1). Cross-checking these against independent observations (e.g., the galaxy-group transition scale) would provide a powerful consistency test.

Sensitivity requirements. For the fiducial $m_e = 1$ eV: $\xi = 7.9\;\mu$m, with $\alpha_\xi \sim \mathcal{O}(1)$ estimated. Current bounds at this range: $|\alpha_\xi| < 10^4$ (unconstrained). Required improvement: $\sim 4$ orders of magnitude in $\alpha_\xi$ at $\sim 8\;\mu$m, or extension of the probed range from 25 $\mu$m down to $\sim 8\;\mu$m at existing sensitivity.

Priority 3: Galaxy-group transition survey

Criterion	Assessment
Feasibility	High. Requires systematic spectroscopy and weak lensing of galaxy groups spanning $\sigma \sim 100$–$1000$ km/s. Data from SDSS, DESI, Euclid, and Rubin/LSST.
Discriminating power	Three-way. Ether predicts a phase transition at $\sigma_c$; $\Lambda$CDM predicts no transition; MOND predicts anomalous dynamics at all scales (but fails at cluster masses).
Fundamental impact	Significant. Detection of a phase transition in dark matter phenomenology would challenge the particle dark matter paradigm.
Timeline	2–5 years (survey data already accumulating).

Observable. The mass discrepancy $M_{\text{dyn}}/M_{\text{bar}}$ as a function of velocity dispersion $\sigma$ across the galaxy-to-cluster transition. The ether predicts a systematic increase from the MOND-like regime ($M_{\text{dyn}}/M_{\text{bar}} \sim \sqrt{a_0 R}/\sigma$) to the CDM-like regime ($M_{\text{dyn}}/M_{\text{bar}} \approx 6.2$) over a characteristic scale $\sigma_c(m_e, \delta)$ (Section 9.3.3). For the fiducial parameters at typical halo overdensity $\delta \sim 80$: $\sigma_c \sim 500$ km/s. The transition should be observable as a change in the slope of the $M_{\text{dyn}}/M_{\text{bar}}$–$\sigma$ relation, concentrated in the group regime ($\sigma \sim 200$–$800$ km/s).

Priority 4: Cluster merger survey

Criterion	Assessment
Feasibility	Medium. Requires simultaneous X-ray and weak lensing data for a large sample of merging clusters, plus estimates of collision velocities from shock Mach numbers. eROSITA + Euclid + Rubin will provide the data.
Discriminating power	Three-way. Tests the $\Delta_{\text{LX}}$–$v_{\text{coll}}$ correlation ((9.10)).
Fundamental impact	Significant. Would test the ether's two-fluid (superfluid + normal) model at cluster scales.
Timeline	5–10 years (awaiting next-generation survey data).

Priority 5: Photon dispersion (CTA / next-generation gamma-ray)

Criterion	Assessment
Feasibility	Low to medium. Current Fermi-LAT/MAGIC constraint: $|\xi_2|\ell_e^2 < 3.2 \times 10^{-26}$ m$^2$. CTA will improve by $\sim 10\times$. If $\ell_e \sim \ell_P$: signal is $\sim 10^{45}\times$ below current sensitivity (undetectable with foreseeable technology).
Discriminating power	Ether-unique (if detected).
Fundamental impact	Extraordinary. Detection of energy-dependent photon dispersion would be direct evidence for spacetime microstructure.
Timeline	5–15 years (CTA first light expected late 2020s). Practical detectability depends entirely on the unknown $\ell_e$.

Assessment. If $\ell_e \sim 10^{-14}$–$10^{-13}$ m: CTA could detect the predicted dispersion. If $\ell_e \sim \ell_P$: the prediction is correct but untestable. The EM–gravitational disconnect (Section 9.1.7) means no gravitational-sector measurement can constrain $\ell_e$. This prediction is therefore speculative in practice, despite being well-defined in principle.

---

9.4.4 Falsification Criteria

A framework that cannot be falsified cannot be scientific. We specify, for each class of prediction, the observations that would falsify the ether framework.

Global falsifiers (any one of these would refute the framework's foundations):

(F1) Dark energy equation of state. If a constant $w \neq -1$ is established at $> 5\sigma$ (i.e., $|1+w| > 0.01$ with high confidence), the superfluid phonon ZPF mechanism is falsified. This is because $w = -1$ follows from Theorem 4.2 (Lorentz invariance of the acoustic ZPF spectrum), which is a structural consequence of any superfluid condensate — no parameter adjustment can evade it.

(F2) Absence of the galaxy$\to$cluster transition. If the mass discrepancy $M_{\text{dyn}}/M_{\text{bar}}$ shows no systematic $\sigma$-dependence across the group regime ($\sigma = 200$–$800$ km/s) — i.e., if the dark matter phenomenology is scale-invariant — the BEC phase transition mechanism is falsified. This is because the transition is a mathematical consequence of Bose–Einstein condensation: any bosonic medium has a critical temperature, and any system above that temperature is in the normal (non-superfluid) phase.

(F3) RAR functional form. If the RAR is measured with sufficient precision ($\lesssim 0.02$ dex scatter) and the data favour a different interpolating function — for instance, the "simple" $\mu(x) = x/(1+x)$ over the ether's $\mu_e(x) = 1 - e^{-\sqrt{x}}$ — at high significance ($> 5\sigma$ in the residuals), the specific superfluid phase-transition derivation of $\mu_e$ is falsified. The current data (SPARC, [60, 135]) do not yet discriminate at this level.

Sector-specific falsifiers (these would refute specific sectors while leaving others intact):

(F4) Thermal Bell: exponential degradation. If $|S(T)|$ in a microwave Bell test follows the exponential form (9.38) rather than the algebraic form (9.34), with the ratio test (9.40) ruling out the ether prediction at $> 3\sigma$, the thermal depolarisation mechanism (Theorem 8.8) is falsified. The gravitational-sector predictions would be unaffected, since they depend on the condensate equation of state, not the Nelson detection model.

(F5) Thermal Bell: no degradation. If Bell violation persists with $|S| \approx 2\sqrt{2}$ at temperatures well above $T_{\text{crit}}(\omega)$ (e.g., at $T = 5\,T_{\text{crit}}$ with $|S| > 2.5$) after rigorous decoherence control, the entire thermal degradation mechanism is falsified. This would also challenge the specific model of how the ZPF mediates entanglement (Sections 7.5–8.7), though the Nelson–SED bridge (Theorem 7.1) would remain valid at $T = 0$.

(F6) Sub-mm gravity: null result. If sub-millimetre gravity experiments establish $|\alpha_\xi| < 1$ for all $\xi > 5\;\mu$m — i.e., no Yukawa deviation with coupling $\alpha \gtrsim 1$ at ranges above $5\;\mu$m — the entire viable $m_e$ window ($0.3$–$3$ eV) is excluded. The healing length prediction would be falsified. This would not automatically falsify the superfluid ether model (the Yukawa coupling $\alpha_\xi$ could be suppressed by a mechanism not yet identified), but it would remove the framework's most accessible experimental signature in the gravitational sector.

(F7) Photon dispersion: superluminal propagation. If high-energy photons are observed to arrive before low-energy photons from the same astrophysical event (i.e., if $\xi_2 > 0$, giving a negative time delay), the lattice-model prediction $\xi_2 = -1/12$ is falsified. The sign of $\xi_2$ is determined by the ether's UV dispersion relation and is a structural prediction of the lattice model (Section 3.8).

What would NOT falsify the framework:

• Detection of a particle dark matter candidate (e.g., at a collider or in a direct detection experiment) would not automatically falsify the ether framework, since the ether's "dark matter" is the condensate mass, not a particle species. However, if such a particle were shown to constitute $\Omega_{\text{DM}} = 0.26$ of the cosmological mass budget independently of the ether, the ether's dual role (DM + DE from a single medium) would become redundant.

• A non-detection of photon dispersion by CTA would not falsify the framework, since $\ell_e$ is unconstrained and could be as small as $\ell_P$ (Section 9.1.7).

• Standard QM decoherence in the Bell test (exponential degradation) would falsify Theorem 8.8 but not the ether framework's gravitational sector.

Summary of falsification hierarchy:

Test	Falsifies	Sections affected	Confidence required
$w \neq -1$ (constant)	Superfluid ZPF mechanism	4.3, 8.2.1	$> 5\sigma$
No group transition	BEC phase transition	4.2.7, 8.2.3, 8.3.3	$> 3\sigma$
Wrong RAR shape	Superfluid $\mu_e$ derivation	4.2.3, 8.2.2	$> 5\sigma$
Exponential Bell $S(T)$	Thermal depolarisation	7.7, 8.4.1	$> 3\sigma$
No Bell degradation	ZPF entanglement model	7.5–8.7	$> 3\sigma$
No sub-mm Yukawa ($\xi > 5\;\mu$m)	Healing length as Yukawa	4.3.8, 8.3.2	$> 3\sigma$ on $\alpha_\xi$
Superluminal $\gamma$-ray dispersion	Lattice microstructure model	3.8	Any detection

---

9.4.5 Summary: The Empirical Programme

The ether framework makes predictions in three domains — cosmology/gravity (Sections 8.2–9.3), quantum foundations (Section 9.4.1), and electromagnetic propagation (Section 9.3.4) — connected by a small number of fundamental parameters. The programme's empirical strength lies in the cross-prediction web (Section 9.1.5): a measurement of any single quantity (e.g., the Yukawa range $\xi$) determines predictions across all other gravitational-sector observables.

The discrimination matrix (Table 9.4) identifies the observables with the greatest discriminating power: the thermal Bell test (ether vs QM), sub-millimetre gravity (ether-unique), and the galaxy-group transition (three-way discriminator). The experimental roadmap (Section 9.4.3) prioritises the thermal Bell test — the only experiment that tests the quantum-mechanical core of the framework — followed by sub-millimetre gravity experiments that would determine $m_e$ and thereby fix all Tier 1 predictions.

The framework specifies explicit falsification criteria (Section 9.4.4). The most vulnerable prediction is $w = -1$ for dark energy: if the DESI hint of time-varying $w$ is confirmed, the framework faces a serious challenge. The most consequential test is the thermal Bell experiment: confirmation would challenge quantum foundations; falsification would constrain the ether's quantum sector while leaving the gravitational programme intact.

PART VI: EPISTEMOLOGY AND CONCLUSIONS