Spatial Fluctuation of the Hubble "Constant"
| Scientific Paper | |
|---|---|
| Title | Spatial Fluctuation of the Hubble "Constant" |
| Read in full | Link to paper |
| Author(s) | Jean Pierre Vigier, Toivo Jaakkola, Jean-Claude Pecker |
| Keywords | redshift asymmetry non-Doppler intergalactic |
| Published | 1990 |
| Journal | Apeiron |
| Volume | 1 |
| Number | 6 |
| No. of pages | 11 |
| Pages | 12-16 |
Read the full paper here
Abstract
Six samples of objects have been analyzed in order to check whether the redshift asymmetry discovered by Rubin et al. is a general effect. The results for all samples are consistent with the existence of the asymmetry. Its average magnitude is = 1300 + 210 km/s. The asymmetry vanishes at large distances. Various interpretations of the effect have been discussed. Arguments are given in favour of a non-Doppler redshift occurring in the intergalactic space within the Local Supergalaxy and other concentrations of galaxies.
This article was written in 1975, and has remained unpublished until now. It is published here in the hope that it will stimulate fresh thinking on the controversial issue of non-velocity "cosmological" redshifts.
Overview
This is a short observational paper by Jean-Claude Pecker, Jean Pierre Vigier and Toivo Jaakkola, written in 1975 and published without revision in Apeiron fifteen years later. It asks whether the hemispheric redshift asymmetry that Rubin, Ford and Rubin reported in 1973 for Sc I galaxies is a general property of the sky or a peculiarity of that one sample. The authors assemble six independent samples — 268 objects in all, spanning Sc I galaxies, first-ranked cluster galaxies, supernovae, Seyfert-like galaxies, Markarian galaxies and absorption-line compact galaxies — and find the asymmetry in every one of them.
Their conclusion is that the Hubble Constant is not constant across the sky at moderate distances. The measured effect corresponds to a Hubble parameter about 24 per cent higher in one hemisphere than in the other, and it fades out at large distances, so it is a local phenomenon. After eliminating eight alternative explanations they settle on the ninth: an excess non-velocity Redshift acquired by light on its path through the Local Supergalaxy and other dense concentrations of galaxies, which lie predominantly in the high-redshift hemisphere. The closing claim goes further than the data — that redshift is "an effect bound to the presence of matter" rather than a property of expanding space, and that cosmological redshifts generally might be understood on the same footing "within the framework of a no-expansion cosmology."
The argument
The measurement
Rubin, Ford and Rubin (1973) had found that Sc I galaxies in the narrow magnitude band 14.0 ≤ m ≤ 15.0 — hence at roughly the same distance — had mean velocity 4966 ± 122 km/s in one region of the sky and 6431 ± 160 km/s in another, the two regions separated by a great circle.
The authors' statistic is the "Hubble modulus", HM = log V − 0.2m, which fixes an object's position in the magnitude–redshift diagram; the anisotropy is then Δ(HM) = ⟨HM⟩II − ⟨HM⟩I. Galactic absorption is corrected by A = 0.25 cosec |b|, and objects at galactic latitude below 20° are excluded. The six samples give Δ(HM) between +0.054 and +0.204, all positive, with five of the six significant at the 2σ–4σ level. Weighting by (NI + NII)/2 and inversely by variance yields
- ⟨Δ(HM)⟩ = +0.097 ± 0.016, ⟨ΔV⟩ = +1300 ± 210 km/s, ⟨ΔH⟩ = +24 ± 4 per cent
which the authors describe as a 6σ result, while cautioning that inhomogeneity in the upper magnitude limits and overlap between samples would make an appropriate value somewhat smaller. In concrete terms: if H = 75 km/s/Mpc holds in region I, then H = 93 km/s/Mpc in region II over the distance interval studied. A second key property is that the asymmetry vanishes at large distances — visible in the distant cluster sample and at magnitudes fainter than those tabulated — which the authors regard as cosmologically reassuring, since isotropy on the metagalactic scale is preserved.
Eliminating the alternatives
The paper's method is elimination. Nine possible causes are listed and eight rejected.
Statistical fluctuation (a) is excluded by the chance probability. A selection effect (b) is hard to construct, since the two regions were observed mostly by the same observers with the same instruments. Absorbing clouds (c) would produce an effect that persists at large distances, contradicting the observed local character; moreover, mapping Holmberg's (1974) regions of exceptionally high and low absorption onto the sky shows no correspondence between them and the sign of the Hubble-modulus residuals. A difference in absolute magnitude between the regions (d) would require ΔM = −0.5 mag holding across two nearly hemispherical volumes more than two hundred megaparsecs across, and holding simultaneously for supernovae and for several different classes of galaxy — "very improbable". Motion of the Local Group or of the Local Supergalaxy toward region I (e) is dismissed in a single sentence: it "is ruled out if the usual interpretation of the 2.7 K background radiation is adopted." A large-scale velocity perturbation (f) fails on the same geometrical grounds as (c), and is further argued against by the fact that the most prominent nearby mass concentrations lie in region II, which should retard rather than accelerate expansion there. A general anisotropic expansion of the Metagalaxy (g) is ruled out by the effect's local character, and intrinsic redshift in region II galaxies (h) by the arguments used against (d).
The excess-redshift interpretation
What remains is (i): an excess non-velocity redshift acquired between source and observer. Two facts recommend it. First, region II contains the central region of the Local Supergalaxy including the Virgo cluster, the Coma cluster with its large-scale extensions, and the Hercules supergalaxy. Second, the photon–boson interaction theory developed by Merat, Pecker and Vigier (1974) predicts precisely that photons are redshifted when crossing luminous concentrations of mass.
The interpretation is also self-consistent with the fading of the effect: a fixed excess of Δz = 0.003–0.006 would be undetectable in the magnitude–redshift diagram beyond z ≈ 0.04–0.05 given the observational scatter, which is what Figure 2 is drawn to show. And it accounts for the long-standing discrepancy in determinations of H itself — low values near 50 km/s/Mpc from distant objects, high values near 100 from nearby objects mostly belonging to the Local Supergalaxy. Distance-scale calibration differences explain part of that, but an excess ΔH of order 30 km/s/Mpc remains within the Local Supergalaxy. The alternative — that the Local Supergalaxy expands faster than the sparser Metagalaxy around it — the authors call "very unphysical", since the mass concentration should retard expansion, and it does not account for the Rubin–Ford asymmetry either.
The authors then place the result in a wider pattern: excess redshifts have been reported in clusters, groups and pairs of galaxies, in single galaxies, in quasars and even in stars, and "the common feature in these observations is that excess redshifts appear in objects of higher-than-average compactness." The Local Supergalaxy, being denser than the Metagalaxy average, fits the pattern. Hence, they argue, redshift is not a property of space stemming from the Big Bang and independent of matter, but an effect bound to matter.
Assessment
The paper's arithmetic is correct, and checkably so. The tabulated sample sizes sum to exactly the 268 objects claimed. Recomputing the weighted means from the table with the stated weights — proportional to (NI + NII)/2 and inversely proportional to the variances — reproduces ⟨Δ(HM)⟩ = 0.098 and ⟨ΔV⟩ ≈ 1280 km/s, matching the quoted +0.097 ± 0.016 and +1300 ± 210 km/s. The internal conversions are also right: since log H differs from HM only by a constant, ΔH/H = 100.097 − 1 = 25 per cent, as against the quoted 24 ± 4 per cent; 75 × 1.24 = 93 km/s/Mpc as stated; Δ(HM) = 0.097 at fixed velocity corresponds to Δm = −0.097/0.2 = −0.49, the "−0m.5" quoted for case (d); ΔV = 1300 km/s gives Δz = 1300/c = 0.0043, inside the stated range 0.003–0.006; and 0.097/0.016 is indeed close to 6σ. Whatever else is at issue, the statistics here were done properly, and the survey of six independent samples rather than one is exactly the right way to test whether Rubin's result was a fluke.
The elimination argument is also, in structure, good practice — nine candidate causes listed, each addressed. And several of the eliminations remain persuasive. The absorption test using Holmberg's high- and low-absorption regions is a genuine check rather than an assertion, and the objection to case (d) — that the same luminosity offset would have to hold for supernovae and for four different classes of galaxy simultaneously — has real force.
The weight of the paper nevertheless rests on the one elimination that is given no argument at all. Case (e), motion of the Local Group toward region I, is disposed of in a single clause invoking "the usual interpretation of the 2.7 K background radiation." The measurement has since gone the other way. The dipole anisotropy of the Cosmic Microwave Background, measured from the late 1970s onward and now known to great precision, shows the Local Group moving at 627 ± 22 km/s toward Hydra–Centaurus. An observer moving at velocity v produces a first-order hemispheric redshift asymmetry of exactly 2v between the approaching and receding hemispheres. The authors' own measured ΔV of 1300 ± 210 km/s implies v ≈ 650 km/s — within the errors, the Local Group's peculiar velocity. That is not a minor coincidence: the single explanation the paper dismisses without discussion predicts the observed amplitude to within a few per cent. It also predicts, correctly, that the effect is constant in km/s and therefore fades in the magnitude–redshift diagram at large distance — the very property the authors treat as evidence for their own model. The direction of the Rubin–Ford effect was debated for years and does not align perfectly with the CMB dipole, so the match is not exact; but bulk motion of the observer, together with Virgocentric infall, is the standard modern reading of these data, and the paper as published in 1990 shows no awareness of the intervening measurements or of the Great Attractor work of Dressler and colleagues from 1987.
The proposed mechanism has its own difficulties. Photon–boson scattering that removes energy from photons along the line of sight is a variety of Tired Light, and inherits the standard objections: an inelastic interaction with the medium blurs images of distant sources, which is not observed, and distorts the spectrum of the background radiation, whereas the COBE FIRAS measurement finds the CMB to be a blackbody to better than one part in 104. The paper offers no cross-section, no predicted dependence of the redshift excess on the actual column density of matter along particular lines of sight, and no test that would distinguish its mechanism from an observer velocity — which is what a decisive test would have to look like, since the two hypotheses predict the same amplitude and the same distance behaviour. The stated correlation with "compactness" across objects as different as stars, single galaxies, quasars and a supergalaxy is asserted rather than quantified.
Finally, the closing inference is much larger than the evidence. Establishing that some redshifts are non-velocity in origin would not establish that the systematic Hubble redshift is, and the leap to "a no-expansion cosmology" is made in a sentence. The (1 + z) stretching of Type Ia supernova light curves, measured since, is a direct test of expansion that no static model with a redshifting medium reproduces. What survives from the paper is a careful and correctly computed demonstration that the redshift–magnitude relation is anisotropic at moderate distances — a real result, and one the standard model accommodates as the observer's own motion.