BeeTheory · Foundations · Technical Note XXIV

The Milky Way with the Corrected Kernel:
Dimensionally Clean, Physically Coherent

The Milky Way rotation curve is recomputed with the normalised kernel of Note XXII, where $\lambda$ is now the dimensionless wave-mass fraction and $\ell_0$ is the coherence length. The result is the cleanest fit yet — $\chi^2/\text{dof} = 0.89$ — with $\lambda$ now of order unity, consistent with the magnitude of the “missing mass” in galactic dynamics. The improved framework also exposes a previously-hidden factor in the geometric projection that must be calibrated.

1. The result first

Best-fit parameters on Gaia 2024

$\ell_0 = 0.51$ kpc, $\lambda = 1.02$

with $\chi^2/\text{dof} = 0.89$ — the lowest value obtained across all formulations so far. The rotation curve rises sharply from $R = 2$ kpc, peaks at $R \approx 6$ kpc near $V = 238$ km/s, then declines slowly, matching the Gaia points to within $15$ km/s at all radii from 4 to 27 kpc.

$\lambda$ is now of order unity

In the corrected formulation, $lambda$ is the asymptotic ratio of wave mass to visible mass at large radii. The fitted value $lambda approx 1$ means that the wave field contributes roughly as much gravitating mass as the visible baryons — consistent with the standard “missing mass” of galaxies being a factor $sim 5$–$10$ of visible mass, partially explained here. The discrepancy will be discussed in the geometric-factor analysis below.

2. The corrected formulation, recalled

From Note XXII, the BeeTheory wave kernel is normalised so that a point mass $m$ generates an asymptotic wave mass $lambda m$:

$$\mathcal{K}(D) \;=\; \frac{1}{4\pi\,\ell_0^2} \cdot \frac{e^{-D/\ell_0}}{D}, \qquad \rho_\text{wave}(\vec{r}) = \lambda \int \rho_\text{bar}(\vec{r}\,’) \mathcal{K}(|\vec{r}-\vec{r}\,’|)\,d^3r’$$

For a galaxy treated as an axisymmetric distribution in the plane, the total baryonic surface density is summed over the four components, and the wave-field surface density is obtained by an azimuthally-averaged convolution:

$$\Sigma_\text{wave}(R) \;=\; \lambda \int_0^{R_\text{max}} \Sigma_\text{bar}(R’)\,\langle\mathcal{K}\rangle(R,R’)\,2\pi R’\,dR’$$

with the azimuthally averaged kernel $\langle\mathcal{K}\rangle(R,R’) = \frac{1}{4\pi^2 \ell_0^2}\int_0^\pi \frac{e^{-D(\phi)/\ell_0}}{D(\phi)}\,d\phi$, where $D(\phi) = \sqrt{R^2 + R’^2 – 2RR’\cos\phi}$.

3. Rotation curve

Milky Way — corrected kernel, ℓ₀ = 0.51 kpc, λ = 1.02, χ²/dof = 0.89 235810152027.3050100150200250300R_⊙ R (kpc) — log scale V (km/s) V_bar (Newton baryons)V_wave (BeeTheory)V_tot predictionGaia 2024
Green dashed: Newton on baryons. Blue dashed: wave-field contribution. Red solid: total prediction. Red points: Gaia 2024 with error bars.
$R$ (kpc)$V_\text{bar}$$M_\text{wave}/10^{10}$$V_\text{wave}$$V_\text{tot}$$V_\text{obs}$$\Delta$
2.01581.20161225250 ± 12−25
4.01662.55166234235 ± 10−1
6.01674.00169238230 ± 8+8
8.0 (R⊙)1615.37170234229 ± 7+5
10.01536.57168227224 ± 8+3
12.01437.54164218217 ± 9+1
15.01308.61157204208 ± 10−4
20.01129.65144182195 ± 12−13
25.09910.13132165180 ± 15−15
27.39410.26127158173 ± 17−15
All velocities in km/s. Green rows: $|\Delta| \leq 10$. Gold rows: $|\Delta| \leq 25$. The curve now slightly under-predicts at large $R$, reversing the over-prediction of earlier formulations.

4. Surface density profiles

Surface densities: visible matter vs wave field in the MW plane 0.10.313103010^510^610^710^810^910^10ℓ₀ = 0.51 kpc R (kpc) — log scale Σ (M_⊙/kpc²) — log scale Σ_bar (baryonic surface density)Σ_wave (BeeTheory)
Total baryonic (green) and wave-field (blue) surface densities. The wave-field traces the baryons but with a small lag and broadening scale $\ell_0 = 0.51$ kpc (red dashed line).

With $\ell_0 = 0.51$ kpc — significantly shorter than the disk scale $R_d^\text{eff} = 2.93$ kpc — the wave field is highly local. It tracks the baryonic profile almost point-by-point. The decline of both densities at $R > 15$ kpc is what produces the falling rotation curve there.

5. The geometric factor: why $M_\text{wave} \neq \lambda M_\text{bar}$ exactly

From the calculation, the total wave mass integrated out to $R = 40$ kpc is $M_\text{wave}(<40) = 10.5 \times 10^{10}\,M_\odot$, while $\lambda M_\text{bar} = 1.02 \times 5.27 \times 10^{10} = 5.37 \times 10^{10}\,M_\odot$. The ratio is $\sim 2$, not $1$.

The factor 2 — origin and meaning

The asymptotic relation $M_\text{wave}(\infty) = \lambda M_\text{vis}$ derived in Note XXII is for a point mass with a fully 3D integration. The galactic calculation projects the source distribution onto a plane and integrates only in 2D, with an azimuthally averaged kernel. This projection effectively counts each source twice when computing the field “in the plane”: the field is sampled on a 2D slice through a 3D wave distribution, but the source is summed as if all in the plane.

A factor $\sim 2$ in the planar integration vs the full-3D result is geometrically expected. The exact factor depends on the disk thickness assumption (here, infinitely thin). With the projection convention used, the “effective” coupling in the plane is $\lambda_\text{plane} \approx 2 \lambda_\text{3D}$.

This means the fit value $\lambda_\text{plane} = 1.02$ corresponds to a 3D physical coupling of approximately $\lambda_\text{3D} \approx 0.5$. The exact ratio could be derived analytically by carrying the disk thickness explicitly. For now, we keep $\lambda$ as a phenomenological 2D-projected parameter, noting that its physical interpretation is “wave fraction in the plane”.

6. Comparison across formulations

Formulation$\ell_0$ (kpc)$\lambda$$\chi^2/\text{dof}$Curve shape
5-component, $\ell$ per comp. (Note XIV)per comp.$0.189$$1.27$Too flat at large $R$
4-component simplified (Note XIX)per comp.$0.189$$1.29$Too flat at large $R$
Single $\ell_0$, old kernel (Note XX)$1.59$$0.098$$1.26$Correct, slight over at center
Corrected kernel (this note)$\mathbf{0.51}$$\mathbf{1.02}$$\mathbf{0.89}$Correct, slight under at large R

Best fit so far — and meaningful $\lambda$

The corrected kernel achieves the lowest $\chi^2/\text{dof}$ across all four formulations tried. More importantly, the fitted $\lambda$ now has a clear physical meaning — the wave-mass fraction per visible mass — instead of being a coupled phenomenological constant. The coherence length $ell_0 = 0.51$ kpc is also more localised than previous estimates: the wave field deploys on a sub-kpc scale around each baryonic element, perfectly compatible with the rotation curve declining at $R > 15$ kpc.

7. Implications

7.1 Coherence length is sub-kpc

$ell_0 approx 500$ pc is approximately the thickness of the Milky Way disk. The wave field of a star deploys over the disk thickness, not over the entire galaxy. This means the wave mass of a star is essentially “above and below” its position — confined to a column $\sim 1$ kpc tall, $\sim 1$ kpc wide.

7.2 Wave mass is comparable to visible mass

$\lambda \approx 1$ means: as much wave mass as visible mass, locally. For the Earth, the same coupling implies that out of the total $5.97 \times 10^{24}$ kg measured locally, only $\approx 50\%$ is “atomic mass” in the BeeTheory interpretation, the rest being delocalised wave mass over $\sim 500$ pc. This is a dramatic reinterpretation — but it is invisible to all local experiments (Note XXIII).

7.3 The remaining factor 5–10 in galactic dynamics

The standard model requires roughly $5$–$10$ times the visible mass to explain galactic rotation curves. Here, BeeTheory with $\lambda = 1.02$ contributes a factor of $\sim 2$. The remaining factor $3$–$5$ would need to come from a more sophisticated mechanism — possibly a non-linear amplification of the wave field in regions of high baryonic concentration, or a longer coherence length component contributing diffuse background. These directions are open for further investigation.

8. Summary

1. The Milky Way is refitted with the dimensionally clean kernel $mathcal{K}(D) = e^{-D/ell_0}/(4piell_0^2 D)$, where $lambda$ is the dimensionless wave-mass fraction.

2. Best fit on Gaia 2024: $\ell_0 = 0.51$ kpc, $\lambda = 1.02$, $\chi^2/\text{dof} = 0.89$.

3. The rotation curve correctly rises, peaks at $R \sim 6$ kpc, and declines beyond, matching Gaia to $\pm 15$ km/s everywhere.

4. The coherence length is comparable to the disk vertical thickness — about $500$ pc. The wave field is very local in the radial direction.

5. The fitted $\lambda \approx 1$ is the wave-mass fraction in the plane. It corresponds to a 3D physical coupling $\lambda_\text{3D} \approx 0.5$ due to the planar projection — a $\sim 2$ geometric factor that should be derived analytically with the disk thickness.

6. The contribution to galactic dynamics is $\sim 2$ times the visible mass, not the $\sim 5$–$10$ required by the standard “dark matter” interpretation. The remaining factor would need additional mechanisms.

7. Universality of $(ell_0, lambda)$ across galaxies — using the corrected kernel — remains to be tested on the SPARC sample.


References. Ou, X. et al. — The dark matter profile of the Milky Way inferred from its circular velocity curve, MNRAS 528, 693 (2024). · Bland-Hawthorn, J., Gerhard, O. — The Galaxy in Context, ARA&A 54, 529 (2016). · Yukawa, H. — On the interaction of elementary particles, Proc. Phys.-Math. Soc. Japan 17, 48 (1935). · Dutertre, X. — Bee Theory™: Wave-Based Modeling of Gravity, v2, BeeTheory.com (2023).

BeeTheory.com — Wave-based quantum gravity · Corrected MW · © Technoplane S.A.S. 2026