Vier-en-twintig maande

[6 Aug 2026] Discovery item MIG-04, and the window is tighter than the report's 'thirty months': from the George remarriage of 9 Feb 1858 to the Middelwater death of 18 Feb 1860 is almost exactly two years, and every item in between is an event row with its own id.

THE ORDERING IS THE FINDING. The remarriage happens THREE WEEKS BEFORE the pass opened, not after - so the pass did not create the traffic over the Swartberg, it ratified it. A man crossed on the old road to marry, and the road arrived three weeks late for that particular wedding.

ADRIANA CECILIA'S DEATH WAS NOT AN EVENT ROW. The date had been on her person row from the beginning, but with no row in `events` the site's own timeline carried her husband's death and not hers - the matriarch of the whole line missing from the chronology the site is built on. EVT-000291 added.

The page's spine carries [TO VERIFY]: the 3 March 1858 opening rests on a tertiary account (LEAD-0107) and the six-days-later line depends on it. The sequence does not.

[12 Aug 2026] BOLD PASS, both languages. 0 and said it still had lots of bold words and phrases, on a file the voice check had scored 0/3/0 and closed. Rule 4 counts bold carrying a whole SENTENCE and is blind to bolded PHRASES, so a speckled page reads as clean. checks.py now measures DENSITY - bolded words as a share of the page - against the S035 benchmark's 1.1%, plus an en/af asymmetry check, and this story was on that list. Out: dates, extents, survey records, deed and folio numbers, place names and, in S005, whole bolded sentences. STORY-VOICE.md names those exactly, and the benchmark bolds NOTHING in its narrative. NO WORD, FACT, HEDGE, [TO VERIFY], CITATION OR LINK CHANGED - the edit removed emphasis markers. verify_af passing.

[12 Aug 2026] VOICE REWRITE to STORY-VOICE.md, both languages. en 2/0/0 -> 0/0/0, af 5/0/0 -> 0/0/0 (narration/bold-sentence/punchline).

THE CLOSING BLOCK WAS A BUG REPORT IN THE PROSE and is removed under 10. It read: 'Every row in the table is an event in this model with its own id, and can be read on the timeline. Both deaths are there. That is worth saying because for a long time only his was: her date sat on her person row where a chronology built out of events could not see it, so the timeline of the two years that turned on both funerals carried one of them.' That is what the model used to hold and which pass fixed it - process archaeology exactly, and it is MOVED HERE rather than deleted, which is what the notes column is for. The fix itself stands: EVT-000291 carries Adriana Cecilia's death and the table links it.

THE SITE EXPLAINING ITS OWN SCOPE RULE TO THE READER. 'That is why 1859 is this project's hinge... the rules this site works by turn on it: witnesses are collected from occasions up to 1859 and after that only from people already known here' was 6's witness boundary quoted at somebody who came to read about two funerals. The observation underneath is real and is kept in the story's own terms: the two people the district descends from were gone inside twelve months of each other, and 1859 is where the history turns.

'What this is, and what it is not' -> 'The end of the frontier', a heading that tells. The two cautions become a 'What has not been worked out' panel, both intact, including the [TO VERIFY - LEAD-0107] on the 3 March 1858 opening date and the statement that the sequence does not depend on it. 'one of the best documents this project holds' -> 'one of the fullest documents gathered here'; 'this project has never checked it' -> 'has never been checked here'; 'the model's own dates' -> 'the dated records'.

FLAGGED, NOT FIXED - AN AGE THAT MAY BE OUT BY ONE. The page says Adriana Cecilia 'was sixty-six'. PER-000002 carries birth_edtf 1792 with no month, and its notes give a BAPTISM of 25 Nov 1792. A birth in late 1792 against a death on 9 Mar 1858 makes her 65, not 66; only a birth before 9 March 1792 makes 66, and a ten-month gap to baptism is unlikely. Not changed, because the record carries a year and not a date and this is an inference rather than an arithmetic error. S023 states no age at death, so the two stories do not conflict.

ALSO FLAGGED: PAR-000123, Modderdrift Remainder No. 106, is in the frontmatter and named nowhere in the prose. verify_af passing.

[13 Aug 2026] AN AGE CORRECTED, from the register itself. 'She was sixty-six' -> 'sixty-five', and 'had married at fifteen' -> 'at fourteen'. This was flagged here on 12 Aug as an inference the record could not settle, because PER-000002 carries birth_edtf 1792 with no month; the image settles it. S023 corrected in the same pass and the two pages still agree. Nothing else on this page changes: the sequence, the fourteen children, the thirty-two years and the [TO VERIFY - LEAD-0107] on the Meiringspoort opening date all stand.