Set out to do. Of the open items after Session 1, run the four that bear on a key question and need only data already available: Gaia neighbour screen, colour-term stratification, transit-window table, dip-shape fit.
Found.
- No blend. Gaia DR3 sources within 15 arcsec of the star carry 1.2 per cent of its G light (the bound M dwarf at 1.8 arcsec is the brightest); a 9 per cent step cannot be a blend on the coarse dnb plates. With the matched sample re-screened on Gaia (39 stars) the target stays first or second on slope, differential slope and gap offset at z about 1.8.
- Not emulsion colour. On plates DASCH classes as good blue the offset is 0.097 mag (0.088 on all plates), above every one of 41 matched stars (z 1.9); the within-epoch dependence of residuals on colour term (0.48 mag per unit before 1953) would make the post-1969 points brighter, not fainter [@williams2025dasch]. The step is now either real or a systematic tied to the star's exact plate position.
- Transit shape. A limb-darkened planet transit fits the Sector 15 dip with reduced chi-square 1.04, beating a Gaussian dust profile by 196 in BIC and a free asymmetric trapezoid by 27; Rp/Rs 0.110, b 0.79, a/Rs 296, implying a stellar density of 460 kg per cubic metre at a thousand-day period against about 410 expected. Every deep Kepler dip has ingress faster than egress (ratios 0.57-0.77); the 2019 dip does not.
- Calendar. Kepler and TESS coverage (1,523 days) excludes 80 per cent of periods between 100 and 3,000 days for a strict transit and leaves the published windows 1, 3 and 5 [@madurga2026evidence]; windows 2 and 4 as printed are excluded by Kepler on 3-6 March 2011 and 21 April 2012 and are evidently typographical. Window 1 survives only because its March 2011 transit falls in the Q8/Q9 gap, and its 2022 return missed Sector 54 by eight days. Window 3's second return, 18-23 October 2026, had no space coverage. Next: 15-16 January and 28 February to 2 March 2027, then 17-23 February 2028. TESS covered under a tenth of the 782-day recurrence windows [@socas2026revisiting], so that model is untested.
Concluded. The century question narrows to a real ~0.07-0.09 mag drop in the 1950s-60s versus a position-specific plate systematic; the cross-archive step test and a per-plate differential within a few arcminutes are what remain. The dips question gains a dated, falsifiable prediction with shape evidence on its side.
Left open. Check AAVSO/ASAS-SN for 18-23 October 2026; observing request for the 2027 and 2028 windows; eccentric-orbit and ring fit to widen the windows honestly; cross-archive step test; platephot differential; Sector 75 from FFIs; the 2018-2026 ground record.