The baptism this archive twice declared missing was indexed the whole time under SANIGOR — one of the very few spellings its own alias list did not carry. A control test proves a source covers a parish. It proves nothing about whether your search string can reach the record.

The line this site called untestable for months, tested. Three baptism dates match the family tree TO THE DAY across a hundred and thirty years — and the compiler of that tree had plainly seen the register I could not reach.

The oldest question here, answered by the next census along — and the answer was the one the tree gave at the start. Three days of my work went to a Somerset parish on the strength of one yes/no tick.

A Wiltshire newspaper of September 1805 calls him “son of Lieut. Colonel D’Arcy” — and Robert D’Arcy had been gazetted to exactly that rank six months earlier. The first record to make the join anything but an assertion.

The boy baptised at Portsea in 1780, followed to a Hampshire will: Royal Artillery, a Persian knighthood, and a death at Lymington in his sixty-ninth year.

The parish this archive called untranscribed holds her baptism, names both her parents, and took one search. Five generations up a Scottish line it had none of at breakfast.

Seventeen newspapers in three countries announced Major-General D'Arcy's death. Not one gives his first name, and not one names a single relative. That is a fact about the man, not a gap in the archive.

Most of what this archive failed to find, it failed to find in places that were never transcribed — or under spellings it never tried. Ten holdings, measured rather than assumed.

Sixteen people this archive can prove and the family tree has never held. Most of them are women, and most sit in the generation just above where the tree's compiler stopped.

Hannah's brother, living a mile from her in Bristol, with a daughter named Lydia after their mother and a daughter named Hannah after her.

The tree runs this line to a Yorkshire earldom. The tree also admits, in its own note, that the join is only 'thought to be' — and then contradicts itself about which earl. Set out in full.

The oldest claim in this archive is a yeoman born in 1654. Somebody finally went and read the parish register instead of another family's website — and found forty-four Saingers, three Williams, and a transcript that stops eighteen years too early.

The Royal Engineers lived at Chatham and so did this family, and nobody had read its register. It holds a general’s burial, his wife’s, his brother officer’s — and a christening eight days before a convict ship sailed.

All 556 Saniger records in Gloucestershire opened one by one and assembled back into 81 households, to answer one question: is Hannah in any of them? The answer given here was no, and it was wrong — she is at Berkeley, under a spelling the search never tried.

All 551 Staffordshire Sneyd records assembled into households, which finally produces the weaver the tree has been pointing at — and moves one of its dates by twelve years.

One page of the 1841 census holds the entire family this archive's oldest claim depends on — James and Hannah Wakefield at St Paul, Bristol, with Miriam aged one, and a seventy-year-old Sinegar widow nobody expected.

The tree reaches the kings of Denmark. This archive reaches a Gloucestershire yeoman in 1654. The difference is two grafts and about a hundred people — one unevidenced, one now disproved.

D'Arcy was coined twice: once in Normandy as a place, once in Ireland as a colour. The apostrophe is a hint and not a proof.

Eight wives of the direct line who bore sixty-seven children between them, and are recorded in a marriage line and a burial plot apiece. The most valuable document in this archive is one of their testimonies.

A major of the 39th Foot, his children born at Chatham in 1825 and Parramatta in 1827, and a regimental history that confirms both without ever having heard of him.

Four generations in one city: a boy born at Parramatta who went to the California diggings, a railway clerk found dead in his kitchen, a father dead at forty-three, and the widow who took in boarders.

An obituary from 1936 confirms five details of a eulogy delivered in 2010. Two other headlines look damning and say the opposite — the man in them is the witness, not the accused.

Four men out of one family: a sergeant who married nine weeks before he sailed; his wife's brother, killed at Pozières with no grave; her other brother, a stretcher-bearer with the Military Medal and Bar; and a cousin who came home on a hospital ship to die.

Six men were asked what happened to a private of the 9th Battalion at Pozières, and gave six answers. One of them, his closest friend, was still trying to find his grave thirteen months later, from a lecture room at New College, Oxford.

A coroner's file nobody in the family knew existed. It records how Ivy Sneyd's grandfather died in Victoria Park — and, in his widow's sworn evidence, it names his father and settles a generation the archive had been unable to prove.

An organist christened in Limerick Cathedral who emigrated in 1862 and, two years later, installed the first organ in All Saints' Wickham Terrace — and then played it.

Seven lines reached Australia seven different ways — a garrison posting, a convict ship, a free passage, a paid one, and three nobody has dated. Running the assisted-immigration index across every family returns the Wakefields and nobody else.

The archive knew this family reached Queensland. It did not know how, when, or on what. The passenger register gives all three — and a fourteen-year-old Miriam Wakefield shipped as a domestic servant.

Sneyd, Atkinson, Murdoch, Atwell, Blum, Keeling, Creech, Matson, Wakefield, Wright — out of Limerick, Monaghan, Cork, Ayrshire, Somerset, Bristol and Germany.

One box over the whole archive — people, register entries, places and every page. It knows that Saniger, Sainger and Sinnegar are the same family, and that D'Arcy is also Darcy.

Every source this archive has opened, including the ones that gave nothing. Four of the null results are load-bearing: they are why the Hornby claim is called unevidenced rather than disputed.

A living descendant tested. Cutting both rejected grafts — 98 people, thirteen generations, four royal houses — moves her predicted ancestry by 0.39 of a percentage point.

Lives against the state that governed them. In an archive built almost entirely out of government paper, the state is not background — it is the reason the records exist.

The jobs that need a card, a reading room or a reply to a letter. One has been written and sent: Gloucestershire Archives, for the Berkeley marriage of 1797 that would give Hannah's mother a surname.

The seventy-five people this archive met in documents and had nowhere to put: the constable who recovered a body, the six men who gave six accounts of one death, the convicts who broke open a major's crate.

Where this family lies — and two cemetery plot references that two different people are each supposed to occupy.

Every named individual this archive has met anywhere — 586 entries, with the source printed beside each one. Only 113 of them rest on a document. The other 473 come out of the family tree, and 98 of those hang above a graft the archive has already rejected.

Every person the archive can name and publish — 473 of them, most of whom never continued the line.

167 places, from Hornby Castle to Boulia. Read the Australian list in order and you can watch a family spread along the Queensland railway.

This is one of six, kept to the same method. The D'Arcy line married Defranceski; the Defranceski line had married Falco. All three meet in a living generation that none of the archives publishes.

56 mistakes this site published and then corrected, with the cause of each — and 13 inherited from the family tree. The largest was building three days of research on a single yes/no tick in a census.

What 'documented' means here, what 'family lore' means, and the rule that keeps living people out of the build entirely.