What changed

This page is the archive's working record. Each entry is one research pass: the sources opened, what came out of them, and which pages of this site changed as a result. Where a pass produced an error that was later corrected, it says so and links to the correction rather than quietly editing the history.

Back to Britain — and an audit that says the archive had drifted

• Audited the register by nationality of source: 95 Australian records against 70 British. The archive had drifted to the wrong side of the world and nobody had measured it.
• Named the five TNA RG 4 registers that would hold Hannah's baptism, replacing a vague instruction to write to Bristol.
• Found a Sinegar brewer's household at St Philip and St Jacob with a daughter Lydia — candidate kin, labelled as such.
• Established that William the weaver's baptism is in no transcribed Staffordshire register at all, and that Madeley has no Bishop's Transcripts in FreeREG.
• Searched Denbighshire for John Maddock Jones of Ruabon. Well covered, and he is not in it.
• Read the Gazette for Joseph, Constantine and George Pitt D'Arcy. Nothing, with an OCR caveat.
• Was blocked by Cloudflare at irishgenealogy.ie — three Irish lines, one free database, and no way in from here.

The Gazette, the Birches, and a jurisdiction still shut

• Found ten London Gazette notices naming Robert D'Arcy between 1793 and 1819, and read one in full.
• Identified the Birch household: Margaret and Elizabeth Ann were daughters of John Birch and Annie Reeves.
• Traced nine Wakefield grandchildren through the Queensland index.
• Tried the NSW BDM index and was stopped by a bot check — recorded as an errand rather than a null.
• Corrected a stale blocked entry: Museums of History NSW is live and the State Archives catalogue is reachable.

The Sneyd chain closes, and four other lines tested

• Reconstituted all 551 Staffordshire Sneyd records into 94 households.
• Found William Sneyd the weaver's family at Madeley — thirteen children, 1769 to 1794.
• Corrected the tree: their Samuel is baptised 1781, not 1769.
• Ran the Queensland immigration index for every other family in the archive. Only the Wakefields came assisted.
• Exhausted Bristol's four transcribed registers. Neither family is in them.
• Searched Gloucestershire for Richard Wakefield and found one weak candidate.
• Searched Queensland for the three eldest Wakefield daughters. They did not sail and never followed.
• Tested the Somerset-side-of-Bristol hypothesis: Bedminster and Long Ashton are covered and empty; Temple, Redcliffe and Brislington are not transcribed.
• Searched Hampshire and Kent for the other eight D'Arcy children. Two of ten are in the registers.

Chatham — the parish the Corps lived in

• Searched Chatham St Mary the Virgin, where Robert D'Arcy served and died, and which nobody had ever tried.
• Found seven D'Arcy records including three burials the archive has only ever known second-hand.
• Read Robert's burial: age 76 — the primary source behind the tree's “1751”, seen for the first time.
• Found Jean Ward's burial, at sixty-six, and Constantine's, entered with his rank.
• Found George Pitt D'Arcy in a parish register as “Bt: Major 39th Foot”, christening a daughter eight days before the England sailed.
• Searched six more counties around Bristol for Hannah Saniger, with controls. Nothing.

The ship: General Hewitt, Brisbane, 16 December 1854

• Found Queensland's assisted-immigration index published as open data after ArchivesSearch itself refused to run a search.
• Found the whole family on one voyage and downloaded the register scan, DR38490.
• Read the register: James a mason, Hiram a plumber, and Miriam — entered as Marion — shipped at fourteen as a domestic servant.
• Established that the register does NOT give Hannah's own county: hers is a ditto of her husband's.
• Searched Hanley — the parish the tree names and nobody had tried — and found Samuel Charles Sneyd's baptism with both parents.
• Followed up the Madeley marriage of William Sneyd and Mary Blackbourne, which had been in the results all day.
• Corrected this morning's overstatement that the Sneyd line was unevidenced. Two of its four links are documented.

The county reconstituted, and the Bristol migration

• Opened all 556 Gloucestershire Saniger records individually instead of searching them by name.
• Assembled 81 households from the 196 baptisms that name a father, splitting same-named couples wherever 25 years separate consecutive baptisms.
• Asked all 81 whether any could be Hannah's. None can.
• Published /sanigers: the households, the trades, seventeen spellings rather than eleven, and the 29 Sanigers in Bristol on census night 1841.

The maiden name, settled five times over

• Searched the Queensland death index by MOTHER's name rather than by the deceased — the search this archive should have run a week ago.
• Found five of Hannah's children's death registrations, each naming her, across 41 years and four spellings.
• Found Hannah's own registration, naming her father John Saniger — entered by the clerk as “John Jenniger”.
• Found James Wakefield's registration, naming HIS father Richard Wakefield: a new generation.
• Found Miriam's marriage to William Hartley Sneyd, 15 November 1859, and corrected her age from twenty-one to nineteen.
• Added 14 records and 9 recorded relationships to the register.
• Asked Madeley for the two Sneyd baptisms the tree actually names — 1746 and 1769 — and found neither, in a parish transcribed continuously with 110 Sneyd records.
• Searched FreeBMD for Lyddya Sinegar's death in England and found nothing under any spelling.
• Recorded that the 1861 and 1871 censuses cannot help: the family had emigrated by 1857.

Portsea, Madeley, and three counties that hold nothing

• Found Joseph D'Arcy baptised at Portsea St Mary on 19 March 1780, father Robert, mother Jean — the first record of Robert D'Arcy in an English parish register this archive has ever had.
• Read the Madeley baptism that killed the Keele graft, and then found a William son of William Sneyd buried there in September 1744.
• Established that FreeREG holds nothing for County Limerick, nothing for South Leith and nothing for St Quivox — three of this archive's oldest dead ends are dead ends in the database, not in the record.
• Tested and rejected the Berkley-in-Somerset hypothesis for Hannah Saniger.

The Saniger chain, tested at last — and it breaks

• Checked FreeREG's coverage for Bristol BEFORE searching it. Four registers, 1,551 records, and St Paul is not among them — so the Wakefield join was never testable there and any null would have been a lie.
• Found that Dursley St James (1577–1951) and Cam St George (1568–1939) cover the whole Saniger chain, which /berkeley had implied was untestable.
• Searched the county: 494 Saniger records 1670–1830, and three baptisms matching the tree's dates almost exactly, all at Dursley.
• Opened all three. The fathers are Abraham, John and Samuel — three unrelated men. The chain is a coincidence of dates.
• Established that no Saniger baptism exists in Gloucestershire 1801–1809, in a window that is well covered.
• Established that in 139 Gloucestershire records spanning 1790–1900, no Saniger ever marries a Wakefield.
• Found Miriam Wakefield's birth registration in FreeBMD: June quarter 1840, Bristol, volume 11, page 183 — the only one in England in five years.
• Control-tested the FreeBMD harness (21,248 Wakefield marriages, 1838–1842) before reporting any absence from it.
• Rewrote /berkeley around the result and added the GRO certificate to /errands as the cheapest decisive document on the site.
• Then found the household in the 1841 census and published /drivers-fields — James and Hannah Wakefield at St Paul, Bristol, with Miriam aged one, and a seventy-year-old Sinegar widow living with them.

The DNA, and the branding brought into line

• Published /dna — one living descendant's autosomal estimate against the tree's own prediction, computed twice, once with both grafts cut.
• Corrected an error in that analysis within the hour: the first version put the predicted Scottish share at 0.02% and announced the DNA was contradicting the tree. It was not. Recorded as error nine.
• Brought the homepage onto the .hero banner every sibling archive uses. The rules had been in this site's stylesheet, unused, since it was built.
• Built /search over the whole archive, with the eleven spellings of Saniger folded together.
• Added /searched, /changes and /errands — publishing work that had been sitting in notes/ and requests/ on disk.

Berkeley, and the audit that moved the floor

• Searched the Gloucestershire parish registers instead of another family's website.
• Published /berkeley: 44 Sainger entries, three Williams, one good candidate and a father the archive never had.
• Caught a false story before publishing it as fact, and left it on the page as a worked example.
• Audited every 'defended' ancestor and published the result even though it was bad.

The worst error, and the register

• Told the family their own founding brief was wrong. It was not; I had read the root person's spouse instead of her parents. It stood on the site for a day.
• Published /register — every named individual the archive has met anywhere, with the source beside each one.
• Put a pedigree chart on all 473 person pages, with the border encoding how each relationship is known.

The inquest nobody knew about

• Read QSA coronial file DR103140 in full and published it plainly at /inquest-1902.
• Published /finding-arthur from the Red Cross file.
• Rewrote /great-war around four men instead of one.

If you find something this page does not record, that is a gap in the record-keeping and worth telling me about. What has been read → · What was taken back →