Tabby's star (KIC 8462852)Evidence
Evidence · Tabby's star (KIC 8462852)
Gaia DR3 rules out a blend as the cause of the post-1969 offset
Gaia DR3 sources within 15 arcsec of the star sum to 1.2 per cent of its G-band light (the brightest is the bound M dwarf at 1.8 arcsec, G = 17.19), and within 30 arcsec to 6.5 per cent, so no blend on the coarse post-1962 plates can produce a 9 per cent change; with the matched sample re-screened on Gaia (39 stars), the target remains first on gap offset (z = 1.8) and at the 97th percentile on slope.
What it means
Key questionWhat causes the slow fading of the star?
Most likely: a large, diffuse amount of circumstellar dust, more extended than the clumps that make the dips
Most likelyunchanged
Openunchanged
Answer this evidence bears onThe century fade is a calibration artefact of plate photometry.
Counts against2 for, 3 against
Disputedunchanged
Next steps 1
1
Full per-plate differential with the Starglass platephot endpoint (needs a free API key): for each of the ~900 plates, pull every source within 10 arcmin and fit a plate zero-point from colour-matched stars within 5 arcmin of the target rather than 30 arcmin, then a hierarchical model with one zero-point per plate and one slope per star. Removes the position-dependent part of the plate systematics that the current 30-arcmin differential cannot.
~12 h
—not judged
Ranked by how likely each is to move the key question, for the time it takes
How it was found 1 run
Result. No blend: the target's neighbours within the plate resolution are 1.2 per cent of its light, far below the 9 per cent a blend would need. With the Gaia screen the matched scatter is 0.077 and the target stays first or second on every statistic at z about 1.8. The blend explanation for the post-1969 offset is excluded; the ranking result of Session 1 stands..
Target neighbour flux fraction; target percentile and z among Gaia-screened matched stars on slope, differential slope and gap offset{'tight_n': 17, 'tight_z': 1.85, 'matched_n': 39, 'gap_offset_z': 1.76, 'target_z_wls': 1.77, 'differential_z': 1.84, 'matched_wls_sd': 0.077, 'differential_err': 0.021, 'matched_wls_mean': 0.03, 'gap_offset_target': 0.088, 'differential_slope': 0.131, 'brightest_within_15as': 'G 17.19 at 1.8 arcsec (the bound M dwarf)', 'gap_offset_matched_sd': 0.042, 'gap_offset_percentile': 1, 'target_percentile_wls': 0.974, 'differential_percentile': 1, 'gap_offset_matched_mean': 0.015, 'matched_dropped_by_gaia': 18, 'target_neighbour_flux_frac_15as': 0.012, 'target_neighbour_flux_frac_30as': 0.065}No blend: the target's neighbours within the plate resolution are 1.2 per cent of its light, far below the 9 per cent a blend would need. With the Gaia screen the matched scatter is 0.077 and the target stays first or second on every statistic at z about 1.8. The blend explanation for the post-1969 offset is excluded; the ranking result of Session 1 stands.
DASCH DR7 light curves of KIC 8462852 and 184 comparison stars within 30 arcminGaia DR3 sources within 30 arcsec of KIC 8462852 and of the 184 DASCH comparison starscode/gaia_screen.pycode/analyze_dasch.py4 outputs
Notes from the run
analyze_dasch.py gained a SCREEN=gaia switch that replaces the APASS neighbour screen with data/gaia_crowded.csv; otherwise unchanged from Session 1 (sha256 b1e712bc... vs 8e38ee9e...). The ESA Gaia archive answered one query and then stalled; VizieR was refused by the network; the MAST copy of Gaia DR3 was used.
Sources 2
Cite it
Mysteries Department. (2026). Gaia DR3 rules out a blend as the cause of the post-1969 offset [Evidence EVD0DEC1E1MET8K]. https://mysteriesdept.com/r/EVD0DEC1E1MET8K
@misc{mysteriesdept_evd0dec1e1met8k,
title = {Gaia DR3 rules out a blend as the cause of the post-1969 offset},
author = {{Mysteries Department}},
year = {2026},
note = {Evidence, catalog key EVD0DEC1E1MET8K},
url = {https://mysteriesdept.com/r/EVD0DEC1E1MET8K}
}