The work list

Retired by the control rather than carried out, and the distinction is deliberate. Images 652–662 were read (December 1787 to July 1791, no Llerena). Images 663–950 were never read and now will not be: the book is in the full-text index and Sancho is absent from it under all three spellings. Done is the nearest of the five states to what actually happened — the question this row existed to answer is answered, by cheaper means. The caveat survives: machine transcription of eighteenth-century hands is imperfect, so an index absence is weaker than a page absence, and if anything later contradicts this the 290 images are still there.

Run and passed, 15 September 2026. «Capellán del Regimiento Infantería de Buenos Ayres» returns 22 results from a collection titled Montevideo. Military Award Records 1812–1814, with snippets reading «…de Berroeta Capellán del Regimiento…» and years 1789, 1790, 1791, 1792 — the chaplain and the exact years of the book. The regiment's book is in the full-text index, so the absence of any military collection from the 68 results for «Sancho Llerena» is evidence rather than a gap in coverage. Two limits kept open below.

The sweep of 15 September returned 68 results against far fewer in this archive's earlier attempts, and only about fifty titles and three snippets were actually read before the block returned. Results 51–68 were never displayed at all. Re-run at 100 per page, extract every ark, and resolve each through the tile service's manifest to a film and image — that route keeps working while the main site is blocked.

Both opened. Montevideo Catedral, Baptism Records 1825–1828 at dgs:007713376 images 515 and 601 — Francisco Juan Ramón (22 March 1826) and Lucinda Gregoria del Corazón de Jesús (12 May 1827), both grandchildren of Sancho through Dominga. Cordón, Religious Marriage Records 1825–1916 at dgs:007713391 image 1669 — Justa Pastora's marriage of 3 November 1830. All three acts turned out to be already held, which is what made the parent-pair query a coverage check rather than a discovery.

A film this archive had never named until 15 September. It holds the 1789 godparent act at image 551 and the 1811 baptism of Luis at image 2410 — the best document there is on Sancho's parents. Two images of it have been read and nothing else. It spans at least 1785 to 1811, which is the whole of Sancho's Montevideo life.

Read at magnification: the castrense man is "Manuel Sánchez, Sargento de 1ª Clase del Regimiento, natural de la Ciudad de Córdova". The Canelones petitioner is "natural del Reino de Andalucía". They do not match. A man from the Andalusian city would be called of Andalucía — what the Canelones clerk writes, and what Sancho's own act means by Córdova la Llana — while an unqualified la Ciudad de Córdova in a Montevideo register most naturally means Córdoba del Tucumán, named in full a few leaves earlier in the same book. So the objection this session raised against its own argument is withdrawn, with the caveat that removing a counter-example is not producing evidence.

DGS 7713386 is not it. This archive called it «Índices de bautismos, matrimonios y defunciones, 1726–1800» and called opening it the cheapest test in the archive; images 300 and 1500 are cathedral burial registers of 1887 and 1905. Two samples of 2,244 do not prove the index absent from that film, but the attribution is disproved and is logged in corrections. The index may exist under another DGS — find it the way the Canelones one was found, from an image's own Information panel rather than from a catalogue line.

«Córdova la Llana» is the city of Córdoba in Andalusia; no municipality of that name exists because it is an epithet, and the qualifier distinguishes it from Córdoba del Tucumán. So Sancho was born there about 1767 and his parents Félix Llerena and Rosalía Aguilar were natives of the same city. Nobody has yet looked at a single Córdoba parish register. That is now the largest unopened door in this archive, and it is in Spain rather than Uruguay.

Resolved 15 September 2026, and there was never a contradiction. The J193 says Anton Armando Lerena died 3 April 2003; estate 15127/2013 was advertised on page 68 of the gazette of 16 August 2013. This archive's row held the gazette year and the page number, and this session read them as a death year and an age. gendatabase was right. One man. Logged in corrections, along with the discovery that the whole column was mislabelled.

The 1994 death notice of Ricardo Juan Carlos Lerena (b. 1920, m. 1941 Doreen May Chappell) from gendatabase.com — the source David found, which this archive had never tried. A South African death notice names the deceased's parents, which would put in a modern official document what this archive currently infers: that the Cape Lerenas descend from Pablo Armando. This session does not create accounts and does not pay for things, so it stops here. Logged in errands.

Materially advanced 17 September 2026 without touching the paywall — data/1999-deaths-tested.tsv. The artefact reading is dead for three of the five. The GEDCOM gives Pierre Frederick and both sons an exact shared day, 12 November 1999, and the Government Gazette — independent of it — advertised Pierre's estate ten weeks later. A default year does not behave that way. Two of the five remain open and rest on a single index row each: Gayle Lerena and Catherine Tertia Karon. And one of those was never actually tested: the gazette sweep was keyed on the surname LERENA and the index carries Catherine under KARON, so reporting her absence as a null would have been a straight error — sweep KARON too. Gayle's absence is doubly worthless: the Rhena May Lerena control shows this source misses real adult deaths, and a wife married in community of property shares one joint estate wound up in her husband's name, so the expected number of separate notices is exactly the one already found. Still needs the death notices, and row 17 governs publication either way. ANSWERED 20 September 2026: IT IS A CATASTROPHE, NOT AN ARTEFACT. The Johannesburg cemetery register — free, and independent of the commercial index — gives Pierre Frederick and Richard both dead on 12 November 1999, Gayle and Jeffrey buried with them on 22 November, and all four sharing a burial registration of 18 November 1999 with four consecutive record identifiers: one family, entered in a single act of registration. Gayle is the one the gazette could not reach, and the refusal to publish her 1999 as fact was correct until now. Doreen May died 17 November 1999, five days after them. Only Catherine Tertia Karon remains untested, because she is indexed under KARON. AND THE FIFTH NAME IS NOW TESTED TOO. Catherine Tertia Karon was swept under KARON, the surname she is actually indexed under: 28 Karon records in South Africa (so the surname is there), 281 Catherine Karons worldwide, every one American or Belgian. She is not found — and she is absent from the very Johannesburg register that holds the other four. That is not proof she did not die in 1999; she may be buried in another province, and FamilySearch's South African civil death collection stops at 1967. But it is the first evidence pointing away from her being part of the same event, and it is recorded as a lean rather than a finding.

Written up in notes/FOR-THE-ESTATE.md, 15 September 2026, where a sibling session will find it, with working code and the traps spelled out. Covers dgs:{film9}.{film9}_{image5}, the manifest.json ark round-trip, why it matters — the tile host is not behind the Error 15 block that takes down /search/ and ark: pages together — and tools/fsimage.js with its three traps.

Written up in notes/FOR-THE-ESTATE.md, 15 September 2026, where a sibling session will find it, with working code and the traps spelled out. Both cases set out in full: 7713373, where the label was wrong and the substance held, and 7713386, which is simply not the index it was called. Plus the film-gap lesson — three probes are not a survey — and the one-call manifest.json test that settles any of it.

Founded on a mistake and kept rather than deleted. On 15 September this session was asked whether outstanding work was tracked centrally, found that it was not, and wrote data/open-threads.tsv — the ninth session to invent its own answer, which is precisely how this estate acquired a Chip meaning two things and one chart drawn under three names. The right thing already existed in the kit, written first in the Defranceschi archive. The content of that file has been folded into this list and the file retired. The lesson is the one the kit's own check states: a convention that lives in a document drifts, because nothing is checking it.

Found 15 September 2026 after the archive had spent weeks recording it as probably not digitised. DGS 8166578, image 2687 — «En ocho días del Mes de Enero de 1791 años… Sancho Llerena, natural de Córdova la Llana, hijo legítimo de Félix Llerena y de Rosalía Aguilar, naturales de dicha Córdova… con María Fernández, natural de las Asturias, hija legítima de Rodrigo Fernández». The year is in arabic numerals. It settles Córdoba against Granada, corroborates Apolant's date, confirms his parents independently of the 1811 baptism, and corrects the trees on the bride's father.

Finished 15 September 2026, with a negative. DGS 8166573 does not cover January 1791: Volumen 6 runs to November 1790 and ends at image 1459, Volumen 7 opens at 1460 and its first dated case is March 1791. Sancho married on 8 January. The whole of 1791 in that series is one expediente belonging to somebody else. A seam in the filing, not a failed search — and the caveat is recorded: the filming card reads «hojas sueltas», so a stray sheet could sit elsewhere in Volumen 7. It also produced the one hard fact about the man himself: at image 2167 he is sworn as a witness and could not sign his name.

Materially reduced, 15 September 2026. David chose to re-source rather than argue fair citation, and the South African Government Gazette — free and official — carried more than the index did. The three sons of the 1950 estate are now off the commercial source entirely: Ricardo Juan Carlos with exact dates, Nuno Fernando with exact dates and a wife nobody knew of, Rogue Luis Armando with his identity number. What is still gendatabase's alone is the later generation — Doreen May, Roberto Juan Roque, Brigitte Zaage, Anton Armando, Patricia MacEwan, Jeanette Riekstins, Belinda Michael, Paul Raymond, Rhena May, and the five deaths of 1999. Their gazette notices exist and are located but unread, so the decision is narrower than it was but not gone. The attribution has still not been stripped. MATERIALLY REDUCED AGAIN, 20 September 2026. The five 1999 deaths and Doreen May — the last facts in this archive that came only from the commercial index — are now carried by the free Johannesburg cemetery register with exact dates and record identifiers. The decision this row asks for is now about very little.

Both closed, 15 September 2026. Limit one: six arks from the 22 results were resolved through the tile service's manifest — five return DGS 7713373, images 657, 667, 676, 716 and 831, and image 657 is a page this session had already read by hand. So the collection labelled Military Award Records 1812–1814 is the regiment's book, whatever its label says. The sixth returns DGS 7713382 image 2736, the Montevideo marriage register. Limit two: all three spellings swept — «Sancho Llerena» 68, «Sancho Lerena» 143, «Sancho Lorena» — and not one hit comes from DGS 7713373.

Four of the five done, and the fifth is a real null. Read off the gazette: Anton Armando (b. 1952, d. 3 April 2003), Roberto Juan Roque (b. 20 March 1944, d. 20 February 2022), Brigitte (b. 16 October 1951, d. 18 May 2023), Belinda Maud (b. 16 January 1957, d. 2024 — the printed day is impossible and is flagged, not smoothed), and Rogue Luis Armando in full: born 6 December 1905, died 29 January 1973, which makes him the eldest of the three sons by fifteen years. Doreen May could not be re-sourced: her estate number returns only two J187s, neither carrying a date. Still open and moved to its own row. Also still unchased: the 1986 notice for Lerena Racing Stables, which belongs on the horses page.

Two leftovers from the gazette work. Doreen May is the only person in this family who could not be taken off the commercial index: estate 16409/00 returns two J187 accounts and no J193, so gendatabase's born 1921, died 1999 stands as their fact. Her notice may sit in an unindexed or badly scanned gazette of 2000. And the 1986 notice for Lerena Racing Stables (Government Gazette, 8 August 1986, no. 10383, page 7) has been located and never opened — it belongs beside the gazette's own description of Robert Paul Lerena as a racehorse trainer. DOREEN IS FREE-SOURCED, 20 September 2026. She was the last person in this family whose dates rested on the commercial index — born 1921, died 1999 stood only as their fact. The Johannesburg cemetery register gives died 17 November 1999, buried 20 November 1999 (ark 1:1:6NWQ-VLD4). The J193 is no longer needed to get her off that index. The Lerena Racing Stables half of this row is unaffected and still open.

Answered 15 September 2026, and the answer is two-sided. The bad half: Córdoba's parish registers are not on FamilySearch. The Research Wiki's Church Records section for the province lists no FamilySearch collection at all — the registers are with the Diocese of Córdoba. Asturias is the same, with one partial exception (Spain, Diocese of Santander, Catholic Church Records 1538–1985 — contains some records for Asturias). Full text for Andalusia is nobility and notarial only. So Sancho's baptism is not sitting on a film waiting to be addressed by number; that route is a letter to a diocesan archive, not a search. The good half, which corrects this session's own pessimism: the provincial wiki pages are not the whole picture. A record search for Llerena in Spain 1730–1800 returns 228 results from national collections the Córdoba page never mentions — Spain, Catholic Church Records 1307–2005; Spain, Baptisms 1502–1940; Spain, Marriages 1565–1950; and Spain, Granada, Catholic Pre-Marriage Investigation Files, 1556–1899 — the same record type walked at Canelones this morning, indexed, for Andalusia, in the right century. The eighteenth century is reachable by index even where it is not reachable by image.

The single most valuable document not yet looked for. The act of 1791 puts his birth in the city of Córdoba; the trees say about 1767 and that is still only a tree. A baptism would give the date, the parish, and his grandparents — a fifth generation, where this archive currently stops at four. Note the act gives no age, so 1767 is unverified and the search window should be generous, say 1760–1775.

The 1791 act says both were naturales de dicha Córdova, which is the first evidence placing the generation above Sancho anywhere at all. Their marriage act would give their parents and their own birthplaces. One tree claims Félix was born about 1725 and married about 1753, every value "about" and none sourced — treat it as a search hint, not a fact.

Never searched, in any form. Two independent parish acts name them — the 1791 marriage and the 1811 baptism of their grandson Luis — and both say naturales de las Asturias. The 1791 act adds that they were vecinos de Santa Lucía in Uruguay, and the 1812 padrón puts Rodrigo, 63, asturiano, living in Sancho's own household. So they emigrated and lived with the family: their Asturian parish is findable and would carry María Fernández's baptism too.

Run 15 September 2026. «Llerena y de Rosalía Aguilar» returns zero; «Félix Llerena» returns only modern Latin America; «Rosalía Aguilar» returns forty collections of which two are in Spain, both Jaén notarial of 1829–37. On its own that would be worthless — so the control was run first. The corpus that is indexed for Andalusia is nobility and hidalguía files, chivalric orders and Madrid protocols, searchable back to 1489 — exactly where an armigerous couple would appear. They do not. Weak evidence that Félix and Rosalía were not gentry with a Chancillería file, which sits with the 1811 act calling Sancho Don while the 1798 deposition says he could not write his name.

Searched 15 September 2026; a controlled null. No indexed death or burial exists for him in Uruguay under any spelling — across 401 loose results he appears only as father, father of the bride, spouse or other, never as the principal of a death. The control passed: a surname-Rodríguez search for Montevideo deaths 1828–1838 returns church burials from Nuestra Señora del Carmen, Cordón — the exact parish where Justa Pastora married in 1830 and where Dominga lived — dated 1832, 1833 and 1838. So burials in his family's own parish, in the decade he is supposed to have died, are indexed, and he is not among them. What that does not mean: the index is name-based and a mangled surname is invisible to it, and the Cordón is one parish of several. He is not in the index; that is not the same as not in the registers. The trees' 1832 remains a tree. 11 OCTOBER 2026: THE DEATH NULL STANDS, AND SANCHO GAINED A RECORD ANYWAY. This row's own sentence — “He is not in the index; that is not the same as not in the registers” — turned out to be the exact description of a second instrument. A full-text search on “Sancho Llerena” returns four pages, and one of them is the marriage investigation file of Florentino Gómez at the Vicaría de Canelones, where Sancho gives sworn testimony in his own voice — the first time this archive has heard him as anything but somebody's father. He says he had spent some years “en la línea de Vitoria”. NONE OF THIS IS HIS DEATH, which is still not indexed and still not found, so this row remains closed as a controlled null on the question it asked. But the reasoning behind that null was narrower than it knew.

Same search, same control, same result. No indexed death or burial for María Fernández, wife of Sancho Llerena, in Uruguay 1810–1870. She returns only in records already held. One gain: FamilySearch's index gives her Parents Rosa Fernández, Rodrigo Fernández — a third independent witness to what the 1811 baptism and the 1791 marriage act both say. Three sources, one couple, no disagreement.

The archive has them piecemeal — Pedro Antonio 1793, Agustina Ramona 1795, Fidel about 1799, Justa Pastora 1800, José Loreto 1806, Luis 1811 and others — assembled from scattered searches rather than from a systematic read. Canelones baptisms 1791–1815 are filmed and the images are now addressable. Reading them straight through would produce a closed sibling list instead of an open one, and the consistency checker already reports that its sibling tests reach only 106 of 277 people. RUN AS A COVERAGE CHECK, 22 September 2026, and it passed. Parent-pair query on Sancho Llerena × María Fernández, Uruguay, against a live control (González in Uruguay = 479,660 minutes earlier). Six records, and this archive already held every one: Luis Escolástico's baptism of 10 February 1811; Justa Pastora as bride, 3 November 1830; Dominga Josefa's christening of 15 May 1804; and three entries where a child of theirs appears as a parent on somebody else's act — Dominga Llarena, Dominga Lorena and Avelino Lerena, the last two spelling the father LERENA where the first spells him LLERENA. So the keyed index contributes nothing this archive is missing, which is the same answer the Canelones phrase-search gave and is worth as much. The children it does NOT reach — Juan Agustín, Pedro Antonio, Agustina Ramona, Fidel, José Loreto — are known from the padrón of 1812 and from acts found by full text, and their baptisms are not keyed. The list is therefore as complete as this instrument can make it, and the remaining work is register-walking, not searching. Getting there also exposed an instrument fault — see false zero seventeen.

Traced. His marriage was already held — Vicenta Vega, natural de Cerro Largo, daughter of Don Romualdo and Doña Andrea Maestre — but his children were not: the indexes give Manuel Antonio and Vicenta Margarita, and a Pedro Antonio Natalio baptised at the Cordón on 12 December 1828, born 1 December. One entry names his mother as María Fernández, consistent with him being Sancho's son; another garbles his father as Francisco Llorena, which is why these are recorded as candidates and not attached to the line.

The archive quotes it — Sancho, andaluz, 45; María Fernández, 36; Rodrigo Fernández, 63, asturiano, in the household — but from a published transcription. Now that images are addressable by number, read the entry itself: padrones often carry occupation, household composition and enslaved people in columns transcriptions drop. Sancho's age of 45 in 1812 implies birth about 1767, which is currently the only support for that year. READ 20 September 2026, and the row's premise was wrong in the archive's favour. The page is on FamilySearch, in a collection nobody here had opened — Uruguay, Census Records, 1726–1858 — so it could have been read from a desk at any point. Index ark 1:1:61LW-R54J ('Sancho Llerena, masculino, 45, Andaluzia'), image ark 3:1:3QS7-99ZR-NKNR. The transcription was good: Sancho, María Fernández and all six children read exactly as quoted. What it did not carry: four named people directly above him under household 59 — Juana 41, Juan 'negro esclavo', Gabriel, María — one explicitly enslaved, and Juana is the name of the enslaved woman the 1812 baptism places in Sancho's household. Whose household those four are in is NOT resolvable at this resolution and is not asserted. See the new row.

The row assumed eleven errands sat on David. They did not — all eleven were already owned by me, this archive. That is the third time today this session asserted something about its own files without checking them. But it exposed a real fault underneath: by the estate's own division — errands.tsv is work for a person, the work list is work for a session — eleven rows of the archive's own machine work were sitting in the wrong register. All eleven migrated: closed in errands.tsv with their detail kept verbatim, and seven new work-list rows opened for the ones that had no counterpart. What remains in errands is what it should be — institutions written to, and four things genuinely waiting on David. Postscript: the migration was too blunt and drift.py refused it. Setting all eleven to done claimed work was finished because its index entry had moved — and one of them says in its own detail that the register has never been turned. All eleven are back at ready, each now carrying a pointer to the work list, which is where their state is maintained. The checker was right and the migration was wrong.

Logged as an errand for David when catalogue search was returning Something Went Wrong all day. The route that works now: open any image from the right parish and read its Information panel, which prints the catalogue link and that reel's item-by-item contents. The archive already holds one dispensation from this series — the 1911 Lerena–Cardozo — so an anchor exists. ANSWERED 17 September 2026, sideways. A full-text search aimed at the parents' names returned a hit inside this very register, and the viewer's own header gives it: DGS 007713367, item 13, 247 images. Found while looking for something else and not found by searching the catalogue, which was still erroring.

Full text has already shown this archive two Cordón collections it has never opened — Baptism Records 1828–1836 and Religious Marriage Records 1825–1916. The Cordón is where the Montevideo branch surfaces in the 1830s, and the 1831 Rivera godfather act sits there. Fold the errand into that work rather than leaving it addressed to David.

A wholly new person, produced this afternoon by the gazette: the surviving spouse of Nuno Fernando Lerena, who died 1 July 1971 at 24 Leiden Road, Gerdview, Germiston. Nobody in this archive knew he was married. Her identity number is partly legible in the notice; her own estate, if she has one, would be findable on gazettes.africa and would name children — grandchildren of Pablo Armando nobody has counted.

Directly tests one of this archive's declared gaps — the eight years between Northam and the Cape. Pablo Armando was born at Rosario in 1882; his eldest son was born 6 December 1905. If that birth was not in South Africa it locates the family in those missing years, and Argentina is the obvious first place to look — the archive already has Argentine baptism sweeps to extend. SHARPENED 20 September 2026, and the register's own answer is an inference. Roque was baptised 12 June 1911 at Corpus Christi, Wynberg — aged five and a half, and three months before his parents married on 3 September 1911, with his mother entered as Mary Septima olim Taylor. So 'Cape Town' as his birthplace is taken from where he was baptised six years later, not from anything that states where he was born. The pattern is legitimation. The question is still open and is now better posed.

The 1950 estate makes her Executrix Testamentary, gives her a half-share of the residue, and describes her as "no relation" — while listing her address as No. 7 Forth Road, Newlands, the same address as the deceased. A woman who is not a relative, lives with him, administers his estate and inherits a quarter of it is not a minor figure. gazettes.africa and the deceased-estates index will both carry her.

The gazette pipeline works and cost nothing; it has so far been pointed only at Lerena. The other two South African lines in this archive are Mary Septima Taylor, Pablo Armando's wife, and the Booyzens who appear in the DNA as Cape Afrikaners. Estate notices would give both families exact dates and surviving spouses. Search surname-first — modern notices print LERENA, Anton Armando, not the other way round.

One firm at 107 Oxford Road, Rosebank acted for Doreen May in 2001, Anton Armando in 2013, Roberto Juan Roque in 2022 and Belinda Maud in 2024 — a quarter-century of one family's estates. That is not a coincidence and it is a way of establishing which Lerenas belong to this family without guesswork. It is also a live firm, which makes it a David question rather than a search, if it comes to contact.

Done — tools/check-errands.py, wired into verify.sh. It fails the build when an errand whose status is still open carries a detail that announces the errand is finished. It found a false positive on its very first run — an errand saying the work "cannot be closed online at all", which is about feasibility, not about the errand — so the patterns were tightened to phrases that can only mean this errand is over. A bare closed is useless here, and a noisy check gets switched off. Then proved both ways: planting the real case (the 1791 marriage errand set back to ready) makes it exit 1 and name the row; restoring it returns to green. 40 errands checked, none finished-but-open.

Distinct from the estate-wide warning below it. Two of two Montevideo attributions failed inspection today — DGS 7713373 was mislabelled but right, DGS 7713386 was simply wrong. Every other DGS in this archive that came from reading a catalogue rather than opening the film is now suspect. They are cheap to test: one manifest.json call resolves any image to its true film and image number. AUDITED 20 September 2026 - see /method/ and data/dgs-audit.tsv. All 58 DGS numbers tested at three images each. 57 resolve; 0 not found. The suspicion is cleared. Done with FamilySearch signed out, because thumb_p200.jpg is public — and note a correction to this archive's own method note: manifest.json is not public, it answers 401 to an anonymous request. Two traps recorded, both of which would have produced false corrections: 401 means a restricted image, not a wrong film (one film gives 401 at image 1 and 200 at image 100), and a first pass that matched film as well as DGS threw six false failures that are old microfilm roll numbers, correctly labelled. One open item: DGS 008166571 gave no 200 anywhere, so it is restricted rather than proven.

Audited, and the suspicion was right. checkliving.py finds living people from an explicit flag — living, presumedLiving, alive, isLiving, or conf: "living". people.json carries none of them, so the green tick was passing over an empty set. The kit's own docstring says that is "correct rather than a gap" for an archive that genuinely has no living people — which had never been established here. So it was established. Of 278 people, three are born 1930 or later with no death recorded: Rieta Maria (1936) and Antoinette Septima Lynette (1943) are both over eighty and presumed deceased by the rule the checker itself applies; Anton Lerena, born 1952, was not — and he is the one person the archive was publishing who might have been alive. He is not: the J193 read today gives his death as 3 April 2003, which also upgrades him from named in the tree, not corroborated by a document. The gate now passes over a set that has been counted.

Written up in notes/FOR-THE-ESTATE.md, 15 September 2026, where a sibling session will find it, with working code and the traps spelled out. With the actual welded line quoted, how it was caught — two gazettes disagreeing about one estate number — and the whole free pipeline, including the predictable archive.gazettes.africa slug.

Written up in notes/FOR-THE-ESTATE.md, 15 September 2026, where a sibling session will find it, with working code and the traps spelled out. With both worked examples: died 1994, filed 1996 and died 2003, filed 2013 — the second of which cost this archive several hours believing two records described two men.

A standing constraint, recorded as a decision rather than left implicit. gendatabase.com's full records, and the South African death notices behind them, sit behind a subscription. This session does not sign up and does not pay. The consequence is visible and deliberate: Doreen May Lerena is the one person in this family who could not be taken off a commercial index, and her dates remain that source's rather than this archive's. The alternative — re-sourcing from the free official gazette — was tried first and worked for everyone else.

Also standing. FamilySearch's security layer blocked this session repeatedly today. The archive does not switch browsers, use another network, or otherwise get around it. The consequences are accepted and written down instead: the Berroeta control waited hours, the variant sweeps are still unrun, and every zero returned by a blocked page was discarded rather than recorded as a null — which is the rule that matters, because a blocked page and an empty result look identical.

Row 1 was running this morning and is now parked. Reading the regiment's book by hand was the right plan only while nobody knew whether it was indexed — the Berroeta control showed that it is, in one search, and a day of work evaporated. Kept rather than deleted because machine transcription of eighteenth-century hands is imperfect, so an index absence is weaker than a page absence and the hand-read may yet be needed if the two limits in row 18 do not close cleanly.

Opened 15 September 2026. It is DGS 8166573, images 72–478 — the same film as the informaciones, in a separate block, and this archive already held an ark into it without ever resolving it. Pedro Antonio's baptism is at image 274 and has now been read from the image rather than a transcript: "Pedro Antonio de la SS.ma Trinidad", born 31 July 1793, godparents Don Antonio Alemán and his wife. Opening it also disproved this session's own claim that the film held nothing below image 1057 — see corrections.

A Montevideo marriage register covering 1791, never opened here, surfaced while resolving the Berroeta hits. Sancho married at Canelones, so this is not where his act is — but it is where the marriages of his comrades, his witnesses and his wife's Asturian family would be, and DGS 7713382 is now anchored at image 2736, which is a way in.

1811–1813, 1831–1833 and 1840–1842, all carrying hits under one spelling or another, none opened. They cover the generation after Sancho — the years when this family moves from Canelones into Montevideo and when Luis, Fidel and Pedro Antonio are raising children. The 1811 volume also brackets the baptism of Luis this archive already holds, at DGS 7713375 image 2410.

The 1951 gazette calls Robert Paul Lerena a racehorse trainer at Newlands; a 1986 gazette lists P. R. Lerena Racing Stables, P.O. Box 425, Randburg as an employer. Randburg is where Ricardo Juan Carlos lived and died, which makes P. R. most likely one of his sons — Paul Raymond is the obvious reading of the initials and is only a reading. Establish it from documents, and note that this family's racing connection may run to living people, who stay out of this archive. ATTEMPTED 20 September 2026 and NOT settled. The way to prove it is Paul Raymond's own estate notice, Government Gazette 22032, 9 February 2001, page 35, which would give his address. gazettes.africa's search works (control: 42 documents for Lerena) but its document viewer is behind a bot challenge, which was not worked around. A name search for him returned nothing and that is not a null — OCR of a name is exactly what a search misses. Untested, not absent.

Read, and the row's own premise was wrong. All three are acts this archive already had: Dominga Josefa's baptism of 15 May 1804 (godparents Joseph Ponteceli and Agueda Rivero), Francisco Juan Ramón, 22 March 1826, and Lucinda Gregoria del Corazón de Jesús, 12 May 1827 — the last two in fuller words here than the index gives, with the Capitán, the Mayor Don Joaquín Ruiz de Carvallo and a General as godfather. This session inferred unread act from unread collection, and that does not follow. So the parent-pair query is a coverage check, not a discovery engine — and it passed cleanly, which says the coverage of Sancho and María's children is better than assumed. Logged in corrections.

A method, not a lookup, and it generalises. Searching two parents as one phrase only matches when a document names both — so it finds children, children's marriages and grandchildren in one query, and it does not drown in the 190,000 results that the surname alone returns. Worth running for Félix Llerena and Rosalía Aguilar in Spain, for Rodrigo and Rosa Fernández in Asturias, and for every couple in the register where the consistency checker currently reports no reachable siblings. RUN 21 September 2026 across the register's twenty couples, and it earns its place. Searching two parents as ONE PAIR surfaces children that either name alone buries in a crowd. Pedro Llerena and Sofía Clavijo: six records where this archive held one child — Isabel, Celina, Telma and Tomasa Cantidia added, four of them indexed under the compound LLERENA CLAVIJO, which names both parents in the surname itself. Mauricio Lerena and Casimira Montaño: twelve records, one explicit — Cipriano Lerena Montaña, whose 1928 Flores death registration names both parents — and four candidates left unentered because their own entries do not name the parents. Luis Lerena × Justiniana Lenguas returns only what is held; Francisco Llerena × Rafaela Daraz returns only Baldomero. And the method has a trap in it: Bernabé × Casimira Peláez read (0) and then 22, on a couple whose children are in this register.

Two results, both negative, and the second is the useful one. First: the expediente has its own record (ark 1:1:XPG2-RYP) and it is Image Unavailable, with no image ark exposed and a completely empty collection wiki page — so no film number can be recovered and the film-and-image trick has nothing to bite on. Its index entry is thinner than the marriage index already held. Second: this session's own lead was an over-read. The 1754 bride Ana María de Aguilar is of La Malá — La Malahá, in Granada province, which this archive had already recorded; our Rosalía Aguilar is of Córdoba. Only the groom is Cordoban, and Aguilar is among the commonest surnames in Andalusia. An earlier session had already run the decisive test — a structured search for a son Sancho de Lerena y Aguilar, returning a controlled No Results Found — and was right to close the file. Logged in corrections.

Searched, and the answer is a real absence — after two fake ones were thrown away. Exact-surname queries returned zero twice, and both were discarded: the same exact-match shape cannot retrieve the 1754 act that is known to be in that collection, because the indexed surname is the compound Lerena y Armenta. A null from a query that cannot find what is there is not a null. The working shape — non-exact surname, exact given names, Spain, 1735–1785 — passes its control, returning the Granada Pre-Marriage collection and the full 1754 record for the known couple. Run for Félix Llerena and Rosalía Aguilar it returns 305 results and not one Andalusian collection — nothing from Granada, Córdoba, Sevilla, Jaén or Málaga. Their marriage is not indexed by FamilySearch anywhere in Andalusia in the fifty years that must contain it. With Córdoba's parish registers not filmed either, the searchable routes are closed and what remains is the diocesan letter. The 1791 act stays the only document this archive has that names them.

The mapping row above establishes that Córdoba's parish registers are not digitised by FamilySearch. The originals are at the Archivo General del Obispado de Córdoba. What is wanted is narrow and citable: the baptism of Sancho Llerena, about 1767, son of Félix Llerena and Rosalía Aguilar; and their marriage, about 1750–1766. Both parents are stated naturales de dicha Córdova on the Canelones marriage act of 8 January 1791, which is the citation to quote. Left to David because it is correspondence with an institution, and because the shared note both archives keep says to search the sent mail before writing to any archive.

SHARPENED 27 September 2026, and this row is now cheap and decisive. The original scope was 1826-1845. A fifteen-month window has been added and it is the one that matters: SEPTEMBER 1876 to DECEMBER 1877. ELISA VILLADEMOROS died in it. The Cordón act of 26 December 1877 calls her (hoy finada) and her daughter Jacinta Mercedes Gregoria was born 19 September 1876. She is not in any indexed corpus - she died before Uruguay had civil registration, which begins 1879 - so the burial is in these books and nowhere else. Both her children were baptised at the Cordón. Her burial would name her parents, and it dates the widowhood this archive used to find Alejandro M. Lerena's parentage. The route the null leaves open. Sancho and María are absent from the burial index, but the index is name-based and an eighteenth-century surname mangled by a transcriber is invisible to it. Nuestra Señora del Carmen, Cordón is the family's own parish — Justa Pastora married there in 1830, the 1831 godfather act is there, Dominga lived there — and its burials of 1832–38 are demonstrably filmed. The Cordón marriage register is already anchored at dgs:007713391 image 1669, so the burial volume is findable from the same reel's item boards. Do the Iglesia Matriz next, and Santa Lucía after that, where María's parents were vecinos.

SECOND CHOICE FOR THE SAME FIFTEEN MONTHS, added 27 September 2026. If the Cordón (row 56) does not hold Elisa Villademoros's burial of 1876-77, the Iglesia Matriz is next - it is where her kinswoman Celedonia's daughter was baptised in 1872, with Elisa herself standing godmother. The Cordón is one parish of several and Sancho may not be in it. The Catedral de la Inmaculada Concepción — the Iglesia Matriz — is where he stood godfather in 1789 and where his son Luis was baptised in 1811, so it is at least as likely a resting place. Its baptism films are already anchored (DGS 7713375, images 551 and 2410), which is the way into the same reel's burial volumes.

The third candidate parish, and the one where he married, where five of his children were baptised and where he swore his 1798 deposition. DGS 8166573 is already mapped — images 72–478 are the baptisms and 1057–3295 the informaciones — so the burial volume is likely on a neighbouring reel of the same diocesan filming, findable from the item boards.

The 1791 marriage act says Rodrigo and Rosa Fernández were vecinos de Santa Lucía. A couple often ends up buried where the wife's people are, and this archive has never opened a Santa Lucía register of any kind. It is also where María Fernández herself may be, and she is missing on the same terms.

Tested, and they did not. Montevideo church burials are indexed continuously across the period: 1832, 1833, 1838 in the first window, 33 indexed burials for one surname in 1839–1848 including 1841 and 1842, and 1852 in the next — all from Uruguay, Catholic Church Records at Nuestra Señora del Carmen, Cordón and Montevideo. The siege of 1843–51 did not destroy them. So Sancho's absence from the burial index is not explained by register loss, which strengthens the negative and justifies walking the images. One honest limit: indexing density is not the same as register survival — the 1850s window returns 7 where the 1840s returns 33, and that could be thinner indexing rather than fewer burials.

Done, and it corrects this session twice over. Fourteen records name him; only four require him to have been standing there — godfather in 1789, sworn witness in 1798, padrino at the 1804 confirmations, and enumerated in the 1812 padrón at calle San Juan No 60, household 59, as Sancho, andaluz, 45. A census enumerates the living, so 1812 is the last document in the world that shows him alive — not 1798, as this row itself claimed when it was written. Everything after is father, grandfather or ancestor, up to 1850, and a dead man can be all three. So the unaccounted gap before the trees' 1832 is twenty years. Written up in data/sancho-last-alive.tsv; the misstatement is in corrections.

Done — data/film-register.tsv, 51 films, built from resolved arks rather than catalogue lines. Nine are identified in detail, including the two-block structure of 8166573 and the correction that 7713386 is not the 1726–1800 index. The other 42 are listed as not yet identified, which is the honest state and a queue in itself. The archive had named perhaps nine films before today.

Done — 78 arks, 51 films. 73 resolved to a film-and-image, two are restricted (HTTP 403 — access-denied, not absent), two return no DeepZoom id, and two are Argentine enrolment cards on a different id scheme. Written to data/ark-to-film.tsv. Every one of them can now be reopened through the tile service even while www is behind Error 15.

Done — tools/fsimage.js. The mosaic reader, the contact sheet, the ark resolver and the extent prober, with the three traps written as comments where they will be read: never cache-bust a failed tile, wait after naturalWidth, and never batch the view and the screenshot. Plus the rule that is not about code — do not read names at fit-width — with the four misreadings of 15 September named in full.

It was already published, and this row was wrong to assume otherwise — the same error as assuming three read images were unread. The method page has carried A null is not a finding until a control says so since 14 September. So it was strengthened instead of duplicated, with the four times the rule decided a question on 15 September: once it saved three hundred images of hand-reading, twice it voided nulls that looked clean, and once — only once — it confirmed a negative worth keeping, which is Sancho's missing burial.

Baptised at Canelones in 1795 with Don Fidel Rey as godfather — which is why her brother is called Fidel. She has never been followed past her own baptism. ATTEMPTED 17 September 2026 and NOT completed — the FamilySearch session expired mid-search and this session does not sign in. The search to re-run is Agustina LERENA exact, Uruguay, birth 1790–1805, then the same under LORENA and LLERENA. Untested, not null. ANSWERED 20 September 2026, and it caught a trap on the way. Agustina Ramona Lerena is absent from Uruguay's indexed records. The first pass — exact Agustina, then Llerena and Lorena — returned three zeros, and those zeros were not safe: FamilySearch's exact forename matches the whole given-name string, so a woman indexed as Agustina Ramona would never have been returned. Re-run in the safe shape (fuzzy forename, exact surname) it returns 35 records and not one of them is an Agustina — they are fuzzy hits on Augusto, Agustín and A. That is the null, and it is a real one. See /method/.

Traced, and the yield is thin — reported as thin. One index cluster gives Fidel Lerena with Víctor Lerena and María Rueda, which is the Montevideo branch this archive already reached through Víctor, with the relationships garbled by indexing. It does not establish that this Fidel is Sancho's son born about 1799, which stays open. A second entry gives a Fidel Lerena with a child "Feliz Llama" — tantalising, since it could be a mangled Félix Llerena, and a mangled index entry is not a person.

The eighth child, baptised 10 December 1806, found only by the variant sweep on Sancho Lorena. Never followed. TESTED 17 September 2026 AND STILL NOTHING, which is now a tested null rather than an untried row. Searched as José Loreto LERENA (birth 1800–15) and, because the archive found him only through the variant sweep on Sancho Lorena, as Loreto LORENA exact. Neither returns him. He is not in FamilySearch's indexed Uruguayan records after his own baptism of 10 December 1806. The routes left are the ones an index cannot reach: the Cordón and Matriz registers walked by image (rows 33, 56, 57), or a death under a mangled spelling. RE-TESTED IN A SECOND INSTRUMENT, 11 OCTOBER 2026, AND HE IS STILL NOT THERE. The same day a full-text search found a marriage this archive had twice declared unfindable without an image walk, this row was re-run against the FULL-TEXT CORPUS, which is built by reading page images rather than by transcribing entries. “Loreto Lerena” as a phrase returns three hits and not one is Uruguayan — Guanajuato, Jalisco and a Madrid draft register. THE CONTROL WAS RUN AFTER THE NULL, not before: “Ambrosia Traibel” returns three pages including Montevideo religious marriage and baptism records, so the instrument demonstrably reaches Uruguayan parish registers. SO THE NULL NOW HOLDS IN TWO INDEPENDENT INDEXES — the record index and the full text — and that is a much stronger statement than this row could make before. The routes left are unchanged: the Cordón and Matriz registers walked by image, or a death under a spelling no index and no OCR has recognised.

Traced, and it opened a branch this archive has never named. The indexes give Avelino Eduardo Lerena married to Celedonia Villademoros, with children recorded under the compound Lerena Villademoros — Pedro Gualberto, and Avelino Pedro Joaquín Silvestre — plus an Avelino Lerena born 1869, died 5 January 1914 at Montevideo, son of Celedonia. The archive knows Avelino as the husband of Ambrosia Traibel, the Lerena Traibel branch it has carried since its first week, and has never heard of Villademoros. Two readings, and this archive is not choosing: either Avelino Lerena Fernández married twice, or Avelino Eduardo is a son of his and the Villademoros children are Sancho's great-grandchildren. A note already here — "she does not fit the Avelino this archive knows" — shows the two-Avelino problem was already suspected. Recorded unattached. DECIDED 20 September 2026: it is the SON. Two parish acts give Avelino Eduardo's own father as Avelino with Ambrosia Traibel named on the record — the christenings of Avelino Pedro Joaquín Silvestre at Unión on 5 March 1866 and of Pedro Gualberto at the Cordón on 15 January 1870. So Avelino Eduardo joins the children of that marriage and the Villademoros children are Sancho's great-grandchildren. Four people added to the register; published at /uruguay/#s-villademoros.

Luis Escolástico, baptised 10 February 1811 — the act that gives Sancho's parents and birthplaces. His marriage to Carolina Oliva is known and his children are not. Currently an errand addressed to David; it is machine work now. ANSWERED 21 September 2026, and it is a NULL. A parent-pair query on LUIS LERENA with a mother CAROLINA, Uruguay, returns ZERO — read twice, because the same session had (0) turn into 22 an hour earlier. The control is the same shape on Luis Lerena × Justiniana Lenguas, which returns the two children this archive holds. So no indexed child of Luis Lerena and a Carolina exists in the Uruguayan corpus. If that first marriage produced children they are not in the index, and the parish books are where to look. AND THAT NULL IS NOW SUSPECT — 21 September 2026, later the same day. The FamilySearch record search was caught returning a clean, stable 0 with “No Results Found” for everything, including surname González in Uruguay and the very query that had returned ten records that morning. A zero recorded from this instrument on this date proves nothing about this couple. The null above must be re-run when the service is known good — with a control whose answer cannot be zero — before it is believed. See data/false-zeros.tsv, entry sixteen. RE-TESTED AND THE NULL STANDS — later on 21 September. Run again in a browser context proved live in the same minutes by two controls: González in Uruguay 479,662 and Avelino × Ambrosia 10, plus a positive on the same query shape — Luis × Justiniana returns 4, among them Alberto Otario, christening 7 October 1850, which is Alberto Octavio's own act. So the shape reaches this exact father, and for Carolina it returns nothing. Luis Lerena and Carolina Oribe have no indexed children. Her husband's other marriage does. That is now a real null with a real control, and the row is closed.

Two are known — Francisco Juan Ramón (1826) and Lucinda Gregoria (1827) — and a third, Domingo, is recorded in 1840, thirteen years later, so this archive has already noted there are almost certainly others between. The Cordón and cathedral baptism volumes now identified are where they will be. ADVANCED 21 September 2026 — and it gave a DEATH rather than a child. A query on the MOTHER (Justa Lerena, Uruguay) returns seven, and one settles a person this archive already held: JUAN BARROS LERENA, died 3 May 1915 at Montevideo, aged eighty-three, registered the next day, certificate 261. The arithmetic is exact — a man born 27 November 1831 is eighty-three on that day — and the registration names his mother JUSTA LERENA and his father "MANUEL C BARROS", the Manuel Zeferino Barros of Pernambuco already here. It also names a wife nobody had: EMILIA ANTELO. Still open: Domingo, recorded in 1840 and never followed. AND A THIRD CHILD, 21 September 2026 — found because the index date was impossible. HILARIO BARROS LLERENA, christened at Canelones, indexed 19 April 1726 — the collection's first year, the same artefact that put his cousin Carlos Ambrosio Florencio in 1726 earlier the same day. The plate gives the real year: "El día DIEZ Y NUEVE DE ABRIL del año de mil ochocientos t[reinta y cinco] … por nombre HILARIO, hijo legítimo de Man[uel Zeferino Barros] natural de PERNAMBUCO, y de JUSTA LLERENA … padrinos PEDRO ANTONIO LLERENA y VICENTA VEGA." So Justa Pastora has three sons by Manuel Zeferino Barros at roughly two-year intervals — Juan Facundo Frutuoso (1831), Aurelio Eusebio (1833), Hilario (1835) — and at every font the godparents are her own brothers and sisters except the first, where the President of the Republic stood. Pedro Antonio and Vicenta Vega have now been found standing at two nephews' fonts seventeen years apart. His birth date is left unresolved: the index says 16 March 1835 and the act's own phrase appears to end del mismo, which would mean April. Read at four times magnification and still not settled.

The 1812 padrón puts four enslaved people in his household. Juana and María Mercedes are named in acts as negra esclava de Don Sancho Lerena y de Doña María Fernández. The castrense register read this morning is full of the baptisms of enslaved children belonging to the regiment's officers. They are in this archive as circumstances of somebody else's life. They were people, their names survive, and the records that name them are the same records this archive reads for its own purposes. CLOSED 20 September 2026, and it was most of the way done already - the row had simply never been marked. JUANA and MARIA MERCEDES are both in the register as people in their own right, with their own pages. What was missing is that the act names the relation — hija de Juana — and the archive had never recorded it, so mother and daughter stood on this site as strangers. The edge is in now, both ways, and both women have evidence rows carrying the act itself (ark 3:1:3Q9M-C9BW-X757-D) so it reaches their pages. The 1812 padrón counted four enslaved people in that household and named none of them; this act names two. The other two remain unnamed and that is recorded as a fact about the record, not a gap to be filled by guessing.

Rogue Luis Armando was born 6 December 1905 and this archive cannot place the family between Northam and the Cape. Pablo Armando was born at Rosario in 1882. If the eldest son was born in Argentina, that closes a declared gap — and the archive already has Argentine baptism sweeps to extend.

He is at Newlands by 1950 and a racehorse trainer; the archive has a 1940 passport and an 1882 Rosario birth. Nothing connects the two ends. Passenger lists, immigration and naturalisation records are the obvious route. CORRECTED 16 September 2026. This row used to end "and the 1908 enrolment card puts him still in Argentina." THAT IS FALSE, and it was pointing the search the wrong way. The card's Oficina enroladora reads CAPETOWN , with Región Militar and Distrito Militar both blank — which is the signature of a man registering at his own CONSULATE because he was already abroad. Far from placing him in Argentina, the card is evidence he had already left it. And the 1908 date is itself unsourced: the card carries no year but the class of 1882, sits in a collection catalogued 1911–1936, and the passport's page 6 gives a Cape Town consular libreta of 1915. See /enrolment-card/.

AND ONE ROUTE IS NOW CLOSED, 9 October 2026. The FamilySearch full-text corpus for South Africa has been worked to exhaustion for this surname — nine hits under the Africa facet, every one opened or excluded with a stated reason. It holds no Cape civil registration at all, and nothing resembling an arrival, immigration or naturalisation record. So this row cannot be answered by full text and the live routes remain passenger lists (BT 27 outward from the UK, indexed from 1890), the consular file behind the 1940 passport, and the 1915 Cape Town libreta of row 101. See data/south-africa-fulltext-closed.tsv.

Closed 9 October 2026, and it had been done since 16 September. The row survived three weeks of work asking for a document the archive already held, and its premise was false besides.

The record is in data/marriage-1911.tsv: Duplicate Original Register no. 7271 (M.L.), entry 160, Wynberg, St Dominic's, 3 September 1911 — read from the family's own papers on 16 September, and the witnesses read off the plate at full resolution on 20 September. He is 29, bachelor, horse-dealer, of Diep River, married after banns by Bertram W. Glynn, R.C. priest, signing "Pablo Armando Lerena" — not Robert Paul.

AND THIS ROW ASKED IT TO DO SOMETHING IT CANNOT DO. It said a South African marriage record "would give his stated age, occupation, residence and father's name — and his father is the whole question of this archive". The Cape marriage register of 1911 has no column for the fathers of the parties. Name, age, race, condition, profession, residence, banns, consent — and nothing else. So the absence of Juan Carlos Lerena from it is a fact about the form, not about the family.

What it did buy: he was a Catholic, married in a Catholic church — the first religious affiliation this archive has for him, and consistent with an Argentine origin; horse-dealer in 1911, nine years earlier than any other attestation of the trade; and the family settled at the Cape with banns called, which with Roque's birth in December 1905 puts them there from 1905 at the latest.

Why it stayed open: nothing checks the work list against data/. This was the fourth time in one day that the archive did not know what it held. See data/corrections.tsv and the retired claims in tools/check-superseded.py, which now fails the build if this premise is asserted again.

The liquidation account names Dichmont & Dichmont of 6 Darling Street as attorneys, Standard Trust Limited, United Fish Canners shares, and creditors including a claim for carrots. The gazette pipeline is free and already built. Those names place him in a business world the archive has not described.

The wiki says Spain, Diocese of Santander, Catholic Church Records 1538–1985 contains some records for Asturias — the only FamilySearch church collection that touches the Fernández homeland at all. Rodrigo and Rosa Fernández have never been searched in Spain in any form. Partial coverage of a neighbouring diocese is still the only free route.

FamilySearch is not the only place Spanish parish registers live. PARES (the national archives portal) and the diocesan portals of Córdoba and Oviedo hold and increasingly digitise what FamilySearch does not. This archive has never looked at either. The mapping row established what FamilySearch does not have; it did not establish that nobody has it. UPGRADED 17 September 2026, and it now has a TARGET. This row was generic. The 1812 Montevideo act has since named a Spanish town: VALDEMORO, in Castilla la Nueva — historically Archdiocese of Toledo, today Diocese of Getafe. FamilySearch holds no parish registers for it (checked, controlled — see row 141), so PARES and the diocesan archive are now the ONLY route. Look for: the baptism of Ramón Serena, about 1775–85, and the marriage of Salvador Serena and Ana de Cubas before it. This is Ramón's house, not Pablo Armando's — work it because it is a real Spanish origin the archive can actually name, not because it is his.

SHARPENED 9 October 2026. The target is unchanged — the baptism of Ramón Serena/Lereña, about 1775–85, and the marriage of Salvador Serena and Ana de Cubas — but the man is no longer known only from a grandson's act. His son Alejandro's marriage act of 16 September 1825 has been read off the plate and writes him "Ramón Lereña", married to "María Josefa González", with a son born at la Villa del Rosario. So a Spanish search now has a documented spelling in a priest's hand and a firm River Plate anchor date to work back from.

Catalogue search has returned Something Went Wrong all day, but an image's own Information panel prints the catalogue link and that reel's item-by-item contents — which is how the Canelones reels were mapped. Do the same from any Spanish image to reach the Córdoba and Oviedo catalogue entries directly, and see whether anything is filmed that the wiki does not list.

Done — data/sancho-children-reconciled.tsv, eight children, one list. The reconciliation found the fault it was written to find: sancho-children-complete.tsv holds four children and is called complete, while the archive names eight across its other files. Nothing in it is wrong; it is partial, and its name made the partiality invisible. Superseded with a banner rather than deleted. The new list carries, for each child, what the act actually says, its tier, and the film and image — four of the eight now point at an image read this week, and two are honestly marked not yet resolved. It also reconciles the count: the 1812 padrón returns six children where eight are named, which is consistent once Luis (1811) and Avelino (1815) are set aside and means one or more of the earlier children had already died.

consistency.py reports that its sibling tests reach 106 of 277 people, and prints the honest warning that a pass here is not a pass over the register. Today's parent-pair query is a cheap way to lift that: for any couple in the register, one phrase search finds their children, their children's marriages and their grandchildren at once. MOVED 21 September 2026, and by fixing the cause rather than the measure. The sibling tests reach people through a STRUCTURED PARENT, and 226 of 293 people had no edge at all — every connection was in prose. Twenty-three were wired from notes that state their own subject's parentage. Structured parents 36 → 60; coverage 106/277 → 114/292; islands 226 → 193. The remaining gap is the same shape: 43 parent names in notes resolve to no single register row, and a name that cannot be resolved is left alone rather than attached to the nearest thing that looks like it.

The Sancho search exposed a gap in the archive's own shape: it records when a person is named, not when they were last demonstrably alive. A man named as a grandfather may have been twenty years dead. One field, computed where possible from the record type, would stop that confusion for 278 people rather than one. BUILT 20 September 2026 - tools/last-alive.py, rendered on /register/. Computed, never typed, and it does not guess — dates in this archive live in prose, and prose carries other people's dates too. Two tiers: DEATH, an explicit date on the person's own row, and DEATH BY, a ceiling from a record that only exists because someone died, which is never treated as the date — Anton Armando died April 2003 and was advertised in 2013. The answer is stark: 20 of 278 are known to have died at all. The other 258 carry a birth estimate and nothing more. That is the finding, not a failure of the tool — and the register's inconsistency is exactly why: it writes 'd. 15 Nov 1935' for one person and 'SA death cert 1950' for the next. Ricardo Juan Carlos stays birth-only on purpose: his source is his father's 1950 estate, where being named means he was alive, not dead. Runs in the gate chain so it cannot go stale.

Migrated from errands, where it was misfiled. The task is to turn the register — page it — not to write to anybody, so it was never a person's errand. This archive already holds one dispensation from the series, the 1911 Lerena–Cardozo, and the register spans the generation the Montevideo branch is thinnest on.

SUBSTANTIALLY OVERTAKEN, 9–10 October 2026, AND THE ROW'S PREMISE HAS CHANGED. This row says the archive "already holds one dispensation from the series, the 1911 Lerena–Cardozo". It now holds thirteen. Nine ledger entries were read by full text on 9 October (data/montevideo-dispensations-fulltext.tsv), and three full EXPEDIENTES besides — 1854, 1855 and 1858 — one of which carries a kinship diagram the Cura Vicario drew himself.

AND THE SERIES IS NOT WHAT THIS ROW ASSUMES. The volumes of 1885–1913 are the Vicar General's summary ledger, one sentence per couple, naming no parents. The volumes of the 1850s in the same collection are full expedientes — petition, interrogatory, sworn depositions naming four generations. See data/corrections.tsv.

WHAT IS ACTUALLY LEFT: the decades nobody has touched — roughly 1860–1885 and 1913–1932 — and they should be worked as EXPEDIENTES or LEDGER according to date, not as one undifferentiated register.

Migrated from errands. Gilberto appears in the 1895 census household and is one of the few Argentine Lerenas this archive can place. Forward from 1895 is civil registration and later censuses.

ANSWERED 27 September 2026. His parents are AVELINO LERENA and AMBROSIA TRAIBEL. The Cordon christening of his daughter Jacinta Mercedes Gregoria, 26 December 1877, read off film 007713388 image 02455: "hija legitima de D. ALEJANDRO M. LERENA y de D.a ELISA [VILLADEMOROS] (hoy finada), orientales. Abuelos paternos D. AVELINO y D.a AMBROSIA TRAIBEL". So the Alejandro M. Lerena of army drawer Cajon 985 is a LERENA TRAIBEL - a grandson of Sancho Llerena. THE ROUTE MATTERS MORE THAN THE ANSWER. Not found by searching for his parents, which this archive had done for thirteen days. Found by searching for his FIRST WIFE - he is viudo in 1886, so one had to exist - and the record was indexed under "Alejandro M. SERENA", with the grandmother as "Ambrosia TRECIBEL". A search under LERENA would never have returned it. Two further gains: Elisa Villademoros is (hoy finada) on that act and her daughter was born 19 September 1876, so she died inside those fifteen months - the widowhood that the 1886 and 1888 acts record. And a 1872 act at the Iglesia Matriz names the godparents as "ALEJANDRO M. LERENA y ELISA VILLADEMOROS", which is what led here. WHAT IS LEFT IS A MERGE, NOT A QUESTION: this register still holds Alejandro LERENA TRAIBEL, baptised at La Mauricia 7 January 1846, a son of the same couple with the same forename and an overlapping birth window. Almost certainly one man. See data/_person-pairs.txt.

Migrated from errands. Luis Lerena married Justiniana Lenguas in 1842, and a general of that name would be in the published military and political record of the period — the easiest external corroboration available for that marriage.

Worked 10 October 2026. The row asked for this archive's compound-surname technique to be crossed with the dispensation series. It was, and the crossing itself returned almost nothing: of the compounds tried, only "Lerena Cardozo" reached a dispensation, and it was the 1911 Lerena–Cardozo this archive already held.

What the compounds found instead was four acts that rebuilt three generations. "Lerena Peláez" returns four documents in the entire corpus and every one was opened:

• Antonina Isabel Lerena de Peláez, dead at ninety-two in 1911, hija legítima de Don Ramón Lerena y de Doña María González — confirming from the civil register the apex of the kinship diagram a priest drew in 1858, and naming her four surviving children.
• Alejandrina Lerena de Peláez, d. 1926 — the woman this archive could only infer the day before.
• Amador Lerena Peláez, d. Durazno 1937 — a second child for the couple dispensed in 1894.
• Luis and Aurora, confirmed at Trinidad, and a new Margarita Lerena.

AND THE TECHNIQUE'S FAILURE MODE WAS FOUND AT THE SAME TIME. Two Durazno birth registers of 1915 matched "Lerena Cardozo" and are not about a Lerena at all: Pedro S. Lerena signed for an illiterate declarant and Saturnino Cardozo witnessed the same act, so the surnames fall adjacent in the signature block. A compound search cannot tell one person's two surnames from two people's one surname each — the Trove column-of-type trap in a new place. See data/false-zeros.tsv.

WHAT IS LEFT, and it is a different row's work: the compounds of the Montevideo trunk — Lerena Lenguas, Lerena Salvanach, Lerena Villademoros — were not run against the dispensations here, because the Trinidad side paid so heavily that it consumed the session. They are cheap and they are still open.

Migrated from errands. An open identity question with candidates already assembled in data/josefa-lerena-candidates.tsv. The act itself decides it.

Two migrated errands, both external sites rather than correspondence. ALEX is to be retried when its full-text search works. Sporting Life matters more than it looks: the gazette now calls Robert Paul Lerena a racehorse trainer and a 1986 notice names Lerena Racing Stables, so an Argentine racing column of 1905–1920 is a plausible place to find him before South Africa — which bears on the eight unaccounted years.

Built 16 September 2026, after David pointed at the Luwinski archive's home page, where the line of record and the line the DNA points to run side by side. The Lerena equivalent is now the second section of the front page: The line of record (Rosario, Wynberg, Newlands, Germiston — a document behind every step) beside The line this archive infers (Sancho Llerena and María Fernández, married at Canelones 8 January 1791; their eight children; the Montevideo trunk; Luis Lerena Lenguas). The point of the drawing is the row where they fail to meet — Pablo Armando's own 1908 enrolment card names Juan Carlos Lerena and María Lerena, and no Juan Carlos Lerena appears in any indexed record in Argentina or Uruguay. That row is dashed in both lanes. Checked rendered at desktop and mobile; it stacks and does not overflow.

Raised by David, 16 September 2026, looking at Nuno Fernando's chart: "2 juan carlos lerena's and 2 mario lerenas and mary septima twice." Two faults, both in the kit's drawing rather than in this archive's data. First, a husband or wife was always drawn as a chip beside the person, with no check that they were already on the chart in their own right — and every married pair of ancestors is two blood relatives, so each got a box AND a chip of the other. 13 relatives came out in 22 boxes A chip is now drawn only for somebody the chart does not already hold; where both halves are on it, they are seated together and the marriage is a tie between them. The test is the slug, never the name. Second, the register deliberately keeps a row for the mangled machine-index reading of Pablo Armando, so the false trail stays visible — and the sibling pass handed that row its own person as a brother. An alias now folds onto the person it is a spelling of. Fixed in the kit (bloodline.js, bloodline.client.js, commits 7f776b3 and 45b37a7) and in tools/build-people.py; kit repinned. Checked against all six archives that draw this chart: the graph is byte-identical for five, and Lerena loses exactly the one alias.

This was never a Lerena-only fault — it was worst in the biggest families. Counted across every chart in every archive, 218,856 duplicate boxes were being drawn. Mazza's chart for Frank Mazza drew 232 of its 530 people twice; Falco's for Costanza Abatiello, 193 of 453; Blazevic's for Petar Blazevic, 24 of 92. They are all still pinned to a kit from before 45b37a7 and will keep drawing the doubled chart until each repins. Not this archive's to do — recorded here because it was found here, per the rule about things belonging to another archive.

Done 16 September 2026, at David's request. 11,581 MyHeritage matches (11,114 with trees) and 1,500 at 23andMe. Surname probes with four controls; a full country census; the one South American tree opened and read. Lerena appears in no match's tree except her own son's and nephew's; Lenguas and Traibel, the two rare Montevideo compounds, return nothing at all Written up on /dna/, with data/dna-match-probes.tsv, dna-match-corpus.json and dna-23andme.json.

This is the single most valuable unbought thing in this archive 23andMe will say which PARENT each slice of ancestry came from, without needing a parent to have tested. It would put the Italian 8.0% on her father's side, where Pablo Armando is, or her mother's, where the Cape is That is the question /dna/ argues around for two thousand words, and one screen answers it. It also settles the Cape trace for the Booyzen archive at the same time. This session does not buy sources, so it stops here.

Done 16 September 2026, from AutoClusters rather than raw match lists. The aunt is Cape and nothing else, so a cluster of the niece's with no counterpart on the aunt's belongs to her father. Run naively it says five of fourteen clusters are paternal, and that is wrong The aunt's chart begins at 75.3 cM and the niece's at 50.6, so four of those five have no member that could ever have appeared on it. Only 15 of her 107 clustered matches clear the floor. Done properly: six clusters confirmed Cape, two paternal candidates, six untestable. One candidate is a false positive — its own Theory of Family Relativity names Kruger, Venter and Nel. The other stands: cluster 14, three matches, 61-116 cM, all BAILEY trees

David generated them on 15 September 2026 and they were read on the 16th — for the granddaughter, for her nephew (the only other tested Lerena) and for her son. Her fourteen clusters are exhaustively Cape Afrikaner and British Isles; not one Spanish, Italian, Argentine or Uruguayan surname appears on her chart The son's chart is the control that makes this readable: it carries two plainly Croatian clusters for HIS father, so the tool does form a block for a foreign paternal line when cousins from it have tested. It found his father's line. It did not find hers. The nephew's chart was the surprise — nineteen clusters, almost all Ashkenazi, which is his mother's family; he shares exactly three clustered matches with his aunt and all three sit in her Cape cluster. Written up on /dna/ with site/src/data/dna-autoclusters.json.

Done the same afternoon, at David's request. New page /dna/ in TheBooyzen, linked under Evidence, with site/src/data/dna.json. Carries the aunt as the instrument, the cousinhood counted (581 of 1,500 relatives name a South African grandparent, 319 name all four), the surname census, the 500 theories, and a trace of Angolan-Congolese and Northern Indian-Pakistani ancestry that archive had not written down before - the ordinary genetic signature of the eighteenth-century Cape. That archive's kit was repinned to 45b37a7 at the same time, clearing its own gate's advisory.

Raised 16 September 2026 by the aunt-subtraction. Cluster 14 on the granddaughter's AutoClusters is three matches, 61 to 116 cM, all in Bailey trees, and it is the ONE cluster on her chart that is demonstrably paternal — its 116 cM member clears the aunt's floor comfortably and is absent from her chart. Mary Septima Taylor's mother was MARY ANN HELLEN BRAYLEY of Bideford, Devon rootschat-assessment.tsv says in terms that Brayley is the one this archive could not push past, and wikitree-searches.tsv records Brayley queries coming back swamped by Bradley, Bayley, Burley, Barley. A 116 cM paternal cluster of Bailey trees is the best lead that wall has ever had Bailey is not Brayley — open the three trees, find the common couple, and see whether it lands at Bideford or Northam. Also worth the same treatment: cluster 12, Jullien and Walker — WALKER runs through the Taylor line as a given name (Margaret Hellen Walker Taylor, Agnes Walker Taylor), which normally means a Walker grandmother.

16 September 2026. David asked whether this archive had searched his folder Genealogy & Family History/Lerena. It had not. Out of it in twenty minutes: the 1911 marriage register (which was on THIS LIST as a thing to go and find), the death notices of Rogue Luis Armando and Nuno Fernando (which had just been ranked first and second on a R2,850 pull list for Pretoria), and six pages of the 1940 passport this archive had transcribed without turning over. Published at /family-papers/. A rule is now on /method/: before any list of documents to acquire, list the documents already held

The Lerena folder still holds eight FamilySearch record-images that have not been opened — 31843_233558__0002-00065.jpg, six record-image_3Q9M-C913- and two record-image_939D- — plus Lerena 2.pdf, a hand-drawn descendant chart, and two photographs (Mary Septima Burial, Rosaline Wilhelmina Forbes Lerena). And there is a whole TAYLOR folder, and a BARRY & BOOYZEN folder, neither of which this archive has ever looked at On today's evidence that is where the next finding is, not in an archive.

Page 6 of the passport names the proof of nationality he actually produced: a libreta de enrolamiento issued at the Cape Town consulate, dated 1915. The enrolment record this archive holds is the ficha of 1908. So there are two Argentine enrolment events seven years apart, both filed at the Cape, and only one of them has ever been looked for Add it to the open errand with the Archivo General del Ejercito, and to the Cancilleria errand — a consular libreta should leave a trace in the consulate's own returns.

Two signatures, something like J. M. Wathington and A. H. Blackalock, and this archive will not print a guess as a reading. Witnesses at a 1911 Catholic wedding are usually family or close friends, and this couple had neither parent present If one of them is a Lerena or a Taylor relative it is the only social connection to him in the Cape before 1911. Needs the original image at full resolution, not the screen copy. READ 20 September 2026 at full resolution (image ark 3:1:3Q9M-C9BP-73LY-N, 2693×1757). First witness: J. M. Wathington — the -ington ending is unambiguous and the earlier guess was right. Second witness: the earlier guess 'A. H. Blackalock' is wrong in its ending. The initials A. H. are clear and the surname ends -bach, not -lock; best reading 'A. H. Blackabach', with the middle explicitly not secure. Neither is a Lerena or a Taylor, so the row's hope — that a witness would be the couple's only social connection in the Cape — is not met by either name. What the page did give: the officiant is Bertram W. Glynn, R.C. Priest, the same priest who baptised Roque at Corpus Christi Wynberg on 12 June 1911 — one priest, both acts, three months apart.

16 September 2026. A filename sweep across ~/Documents — something this archive had never once run — found Photos from Email/, holding Pablo Armando's own death notice, KAB MOOC 6/9/17017 which was on open request at the Western Cape Archives. Item 3: Father, NOT KNOWN. Mother, NOT KNOWN Also the baptisms of Roque and Nuno from Corpus Christi Wynberg, and the 1911 marriage. Published at /family-papers/; errand closed.

The Devon family remembered his surname as MARINO, not Lerena From Sue Rowbotham's Three Corners of the Land: Mary Septima went out "to work for a widower by the name of Robert (or Roberto) MARINO". The three sons she names are exactly right, so the memory is anchored; it is the surname that drifted. Marino is one of the commonest surnames in Italy, and two DNA panels independently put Italian ancestry in his granddaughter. Run: CEMLA and the Argentine passenger lists under MARINO; Rosario 1882 baptisms under MARINO; Cape shipping and directories under MARINO. It may be his mother's surname, a name he used, or a mishearing — all three are searchable and none has been tried.

Roque's baptism names José Bretón and María Luisa Bretón as godfather and godmother, both marked (proxy). They are the first Hispanic-named associates of Pablo Armando ever found in the Cape, and proxy is what a parish writes when the real sponsors are overseas Look for Bretón in the Corpus Christi Wynberg registers, in Cape directories, and in Argentine records beside Lerena. If they are family, they are the closest thing to a relative he ever produced. TESTED 20 September 2026 and it is a NULL — which narrows the lead rather than killing it. Surname BRETON exact with spouse surname LERENA exact, all countries: zero. Control in the identical form, SALVANACH with spouse LERENA: 19. So no Bretón married into this family — if they are connected to Pablo Armando it is as friends, compatriots or employers, not kin. Look for them beside Lerena in Argentine records, not in marriages.

From the cemetery office at Newlands, via Cheryl's sister Patricia: the unmarked grave carrying only the word LERENA holds three people — Pablo Armando, Mary Septima, and a baby girl, CAREEN MARIA LERENA, born and died 1947 This archive has never heard of her. He was sixty-five and twelve years a widower. Cheryl's own reading is that Rose Lilian Theys was the mother; that is a granddaughter's inference, not a record. A Cape birth and death registration of 1947 would settle it. Note his death notice says he had no children of any other marriage. DATE FOUND, 20 September 2026. South Africa, Cape, Cemetery Records: Careen Maria Lerena, died 27 June 1947 (ark 1:1:QP7Z-K7B8). The row had only born and died 1947; here is the day. Whether she is the third person in the Newlands grave is still to be settled from the grave itself. THE REGISTER WAS READ ON 8 OCTOBER 2026, from a page David supplied, and the grave is at MAITLAND. Maitland Road Cemetery, Record of Interments, page 4706, entry 423: surname LERENA, FE, age 2 WKS, late residence ending '...Victoria R[oa]d', allotment ROMAN CATHOLIC, Cons. No. 1, grave 7719. That is Mary Septima Taylor's grave. This archive already cited her burial as 'Maitland Road Cemetery interment register p.3215 entry 163, Nov 1935 ... Cemetery no 1, grave 7719A' - same cemetery, same allotment, same number. WHAT IS STILL OPEN, and it is the part this row was always about: nobody has named her parents. The interment entry gives no father, no mother and no informant - only an age of two weeks and a late residence on a Victoria Road. A Cape birth and death registration of 1947 is still the record that would settle it, and Cheryl's reading that Rose Lilian Theys was the mother is still a granddaughter's inference. AND THE FORENAME IS NOT SETTLED EITHER. The scan is about 1,100 pixels across a whole opening; at that resolution Careen and Carmen cannot be told apart, and Carmen is by far the commoner name in this family - it runs to Patricia Carmen in 1958. The reading this archive already carried is kept rather than quietly improved into the likelier guess. TWO FACTS ADDED 8 October 2026, pulling opposite ways. FOR her being his: she was buried in grave 7719 B, inside the plot the register calls 'In Plot of R. P. Lerina' - his own grave, between his wife and himself. AGAINST the standing guess about her mother: her late residence is 42 Victoria Road, Bellville, and Pablo Armando and his executrix Rose Lilian Theys both lived at 7 Forth Road, Newlands, a different suburb. Cheryl's reading that Theys was the mother is still a granddaughter's inference, and it now has an address to explain. AND THE NAME IS SETTLED: Careen, not Carmen - the clerk's pointed r (as in Frederick, two lines below), not his three-arch m (as in mary, one line below). A Cape birth and death registration of 1947 is still the record that would name her parents, and the Bellville address is now the place to look.

Sue Rowbotham to Cheryl, 2 July 2020: a cutting was, or is, held by the Taylor/Braddick family in north Devon reporting that Roberto was fined £5 for shooting at two native boys who were trying to steal — a horse? — from him. Neither Sue nor Cheryl could place it. A five-pound fine means a magistrate's court, a date and a name in print Cape newspapers of the period are digitised. It would be the earliest independent description of him in South Africa outside a register.

Sue Rowbotham finds no sign of her in England after the April 1901 census, where she is a 16-year-old pupil-teacher at Northam. Her mother died 1900, her father November 1902. Roque was born at Cape Town 25 November 1905. So she sailed between 1901 and early 1905, orphaned, aged 17 to 20 And she had a reason to choose the Cape: her brother Horace Archibald 'Archie' Taylor was posted to South Africa on 21 October 1899 and served through the Second Boer War to 15 September 1902, with the Queen's South Africa Medal, Cape Colony clasp. A nineteen-year-old orphan does not sail to the Cape at random.

The cheapest open question in this archive. Two photographs of a Buenos Aires baptism were filed in David's own Roberto Paul Lerena Documents folder, unexplained. Somebody was pursuing the idea that Pablo Armando was really Roberto Juan Roque Llorens, born 1872 — and the two South African death records both make him mid-seventies, which fits 1872 and not 1882. Against it: his own passport and his own enrolment card, both in his hand, both saying Rosario 1882, and parents who do not match at all Look for Llorens after 1872 in Buenos Aires — a marriage, an enrolment, a death. If he lived and died in Argentina the question closes in an afternoon. Written up at /family-papers/.

The death certificate names the intended place of burial as Maitland Cemetery no 1 The family were told by the office at Newlands that he, Mary Septima and the infant Careen Maria lie there. THE REGISTER HAS NOW BEEN CHECKED - 8 October 2026 - and it agrees with the certificate, not the family. This row used to end 'neither has been checked against a burial register', and that is no longer true. Two of the three are at Maitland Road Cemetery, Cons. No. 1: Mary Septima in grave 7719A (p.3215 entry 163, Nov 1935) and the infant in grave 7719 (p.4706 entry 423, 1947). Three independent things now name Maitland - his own certificate, her interment and the child's. WHAT IS STILL NOT PROVED, AND IT IS EXACTLY WHAT THIS ROW ASKS: his own entry. He died in March 1950 and the opening read is 1947, so it cannot carry him. The record that closes this row is the Maitland Road Cemetery Record of Interments for 1950 - and if he is in 7719 with the other two, that one page settles the grave as well as the cemetery. CLOSED 8 October 2026. The 1950 volume was found and read: Maitland Road Cemetery, Record of Interments, page 339, entry 327 - LERENA, Robert Paul, M.E., 76, of Newlands, 7 Forth Road, Roman Catholic allotment, Cemetery No. 1, grave 7719 C, depth 8 feet, position Bott, officiating Father Plotton, undertaker Goodall & Wms - and the last column reads 'In Plot of R. P. Lerina' THE PLOT WAS HIS OWN. Grave 7719 holds three: A Mary Septima (1935), B Careen Maria (1947), C himself (1950). The death certificate's intended place of burial turned out to be the actual one. WHY IT HAD NEVER BEEN FOUND: FamilySearch indexes him as ROBERT PAUL LESINA - the seventh machine misreading of this surname on record here. He was found by not searching the surname at all: given name 'Robert Paul', death year 1950, collection only, 663 results, first row.

Belongs to the Booyzen archive, found here. A sweep of David's Documents folder on 16 September 2026 found the death notice of GEORGE DOWNING MOUNTJOY BOOYZEN, estate 1095/48, already photographed and in the house It was one of six Booyzen estates Anne was booked to pull at Pretoria on 22 September. Its companion, 2945/44, that archive has already read.

WHAT HAPPENED: THE WHOLE PULL WAS CALLED OFF ON 20 SEPTEMBER, AND THAT STANDS. Anne was told to stop rather than told about one duplicated document. David was asked whether to un-cancel and said no. This row is closed as a decision, not as an oversight — nobody should re-raise it.

WHAT WAS NOT OBTAINED, recorded so the gap is known rather than rediscovered: the will and liquidation account of 1095/48, which no one holds — only its death notice is in the house — and five entire estates, none of which is in the folder: 3734/46, 769/52, 5155/51, 10/57 and 52653. 52653 is written without a year suffix, unlike the rest, so it may belong to a different reference series; that was never checked.

AND THE ROW ITSELF WAS PART OF THE CAUSE. Its headline read "tell Anne before Tuesday, its death notice is already in the house", which scanned quickly is an instruction to call something off; every limit lived in this note, where a person reading a work list does not look. A row that needs its note read to avoid being actioned wrongly is a badly written row. It also said "the other four" and listed five. Both fixed 20 September, and logged in corrections.tsv.

AND THE NAME IN THIS ROW MAY BE WRONG, flagged 8 October 2026. David sent the death notice again that day. Mountjoy is clear and so is Booyzen; the middle name reads more like DOWNEY than Downing. NOT CORRECTED HERE, for two reasons. The photograph is about 768 pixels for a whole foolscap sheet, which is the same resolution that could not tell Careen from Carmen that morning - and this is the Booyzen archive's person, not this one's. A Lerena session should not quietly rewrite another archive's name off a thumbnail. The original is in the house and settles it in a second.

First run used 2018-07 Defranceski.ged, found on the disk. David: "Do not use the 2018 gedcom. That is an old backup only. And is incredibly wrong and out of date." He was right and everything from it is withdrawn. Re-derived from the current export of 15 September 2026 - 15,643 individuals against the backup's 8,286 The Brayley ascent now runs eleven generations to Matthew Reed, born 1643 at Braunton, with parishes and dates throughout, and HELEN WALKER of Haddington is found five generations up, which explains the Walker given name and the DNA match sharing Temple AND Walker

A media object of that title is attached to Mary Septima Taylor in the tree. It is her mother's certificate — MARY ANN HELLEN TAYLOR, born BRAYLEY, died 1900 — the death that orphaned Mary Septima and set up the emigration. This archive does not hold it and has never seen it. rootschat-assessment.tsv says Brayley is the line this archive could not push past, and on the same day the DNA subtraction produced a paternal cluster of Bailey trees at 116 cM. Three separate roads now point at Brayley. The file is not on the disk — the GEDCOM's media references point at MyHeritage's servers — so it needs the live tree, which David is logged into.

A media object titled "Lerena Grave" hangs on Nuno Fernando in the tree. The family were told Newlands; the 1950 death certificate names Maitland Cemetery no 1 as the intended place of burial. A photograph of the headstone or plot may settle which, before anybody writes to a cemetery office OVERTAKEN BY THE REGISTER, 8 October 2026. The interment entries put Mary Septima and the infant at Maitland Road Cemetery, Cons. No. 1, graves 7719A and 7719, which answers the Maitland-or-Newlands question without the photograph. The photograph is still worth having for a different reason: the family's account came with a picture of a grave, so there is an image of something, and whether it shows plot 7719 at Maitland or a plot at Newlands is now a question the archive can actually test. AND NOW THE PHOTOGRAPH HAS SOMETHING TO BE CHECKED AGAINST, 8 October 2026. All three burials are in grave 7719, Maitland, in a plot the register names as R. P. Lerina's own. The eGGSA kerb at Maitland reading LERENA and nothing else is now almost certainly this plot - a surname with no forename is what a plot holding three people gets. What a photograph would add is the section and row markers, which would tie the kerb to 7719 by position rather than by inference.

The tree attaches record-image_3Q9M-C913-K98Q-J to Roberto Paul. The folder on disk holds K941, K94R, K983-D, K983-R, K98S-V and K9H3-D — not K98Q-J. One more page of MOOC 6/9/17017, already captured by somebody, sitting on MyHeritage. Cheaper than the Cape Town archives.

The Brayley reading was tested on 16 September and did not hold The real Brayley matches were found — one Australian at 23.0 cM sharing BRAYLEY and BAKER, one Englishman surnamed Brayley at 16.2 cM — and neither is in cluster 14. But the three Bailey trees hold two people and four people, so their absence proves nothing either. The link is UNTESTED, not refuted She is the largest paternal match on the whole chart and her tree has two people in it. One message asking for two generations of her grandparents' names settles in a week what no amount of surname probing can. This session does not message DNA matches on anyone's behalf, so it stops here.

Attempted 16 September and blocked by the application, not by permissions: the tree view is a canvas app carrying 8,286 people and it froze the browser renderer three times. The three wanted objects are "Mary Ann Taylor Death Certificate" (her mother's, d. 21 April 1900), "Lerena Grave", and record-image_3Q9M-C913-K98Q-J A lighter route than the tree view is needed — the person page reached directly by its MyHeritage id, or Family Tree Builder on the desktop, where the media are stored against the person rather than rendered in a canvas.

Founded on a mistake and struck rather than deleted The Dapto, Wollongong birth came from the 2018 backup. The current tree has Paul Raymond Lerena, born 11 March 1951 at Johannesburg, and no Australian place anywhere in that generation. This archive had already flagged the entry as impossible because his sister was born in Johannesburg six months later - the flag was right and the cause was the backup being wrong.

The current tree has been half-corrected and it needs finishing. His birth now reads Rosario, 22 February 1882 - his own passport's date and town. But his parents still read "Roberto Llorens" and "Francisca Fernandez Lerena", which came from the unsolicited Buenos Aires baptism of 1872 that David himself says is likely no relation. The birth was fixed and the parentage was left behind And the mother's surname is spliced: the Buenos Aires act says FRANCISCA FERNANDEZ, not Fernandez Lerena. Fix it in the tree before the next export, or it will be mined again as evidence by somebody who does not know where it came from.

MEASURED 16 September 2026 — and the first version of this row, written the same day, said these were departments the archive had NEVER SEARCHED. That was wrong about FLORES and too strong about the other three. Retracted; see /corrections/. Flores has a dedicated study — trinidad-flores-kindred.tsv, read 13 September, titled "Trinidad, department of Flores — the founding nest of the surname", with eleven children, the Basque mothers Beloqui, Echeverry and Pelaez, and the Lerena–Pelaez consanguinity dispensations in lerena-pelaez-trinidad.tsv. The Bernabé Lerena and Casimira Pelaez family I reported as a find on 16 September was already in it. WHAT SURVIVES IS THE SCALE, WHICH IS NEW AND WHICH IS THE USEFUL PART. Flores 608, San José 315, Florida 231, Colonia 189 = 1,343 exact LERENA records. CANELONES, worked hardest of all, returns 191. So the archive has read a nest at Trinidad and there is a department of 608 behind it; and San José, Florida and Colonia have no dedicated study at all, only passing mentions, and hold another 735 between them. The picture is not wrong. It is thin in places nobody had measured. And a real warning stands: the tree's Gabino of Flores is given as a son of Ramón Llerena Cubas and Margarita Leal — not Sancho's line — so the interior may hold a different Lerena family entirely. LARGELY ANSWERED 17 September 2026, and not in the way the row expected. Three of the four departments are accounted for by the children of ONE COUPLE — Mauricio Lerena and Casimira Montaño — San José de Mayo, Trinidad in Flores, and Florida, with Salto and Montevideo besides. They were never four leads; they are one family spreading across the interior in a single generation. The tree's own San José person, Demetria Clara Lerena Montaño, is identified as their daughter Clara. Only ROSARIO in COLONIA is still unaccounted for. MARKED DONE 20 September 2026. The row as written — four departments, never searched — is answered, and answered against itself: three of the four turned out to be one family spreading in a single generation, and the 'never searched' claim was retracted to /corrections/. One remainder survives and it is too specific to hide inside this row, so it has its own: Rosario, in Colonia.

Answered 16 September 2026 by reading every citation on every Lerena, Taylor and Brayley person in the export. Félix's Andalusia is sourced to "Carvalho Calcagno Web Site" and "Familia Martinez Camino" — two other people's MyHeritage trees So the agreement this archive noted, "from a direction that owes the archive nothing", owes it nothing because it is two strangers' trees, which may well have copied the same reasoning from the same place. No support. And the wider answer: of 244 people on this line in the tree, SIX cite a record, 94 cite only other trees, and 144 cite nothing at all Written up at /family-papers/ with data/gedcom-graded.tsv.

Worked 17 September 2026, and the free sources took it as far as they go — see data/1999-deaths-tested.tsv. NAAIRS was 503 all day (genuinely down, not blocked), so the death registration this row asks for was not reachable. What the free Government Gazette did settle: Pierre Frederick's estate was advertised 28 January 2000 (GG 20821 p.41, Transvaal; re-advertised GG 20926, 3 March 2000) — ten weeks after the claimed 12 November 1999, from a source independent of both the GEDCOM and the index. That kills the artefact reading for him: a padded year does not reproduce as an exact day in a separate source and then land ten weeks before a real government notice. The two boys are uncorroborated and always would have been — minors with no assets get no J193. A control decides how the absences read: Rhena May Lerena, died 1 April 1998 on cemetery records, has no gazette estate at all inside a 1994–2023 sweep — so absence here is evidence of nothing, and the lag from death to advertisement runs from ten weeks to ten years. Retry NAAIRS for the registration. CLOSED 20 September 2026. The registration this row asked for is in the Johannesburg cemetery register: Pierre Frederick Lerena, died 12 November 1999 (ark 1:1:6X6S-M484) and Richard Lerena, died 12 November 1999 (1:1:6X6S-M48C) — the GEDCOM's date confirmed to the day by an independent source, with NAAIRS still never consulted.

Horace Archibald Taylor born at 23 Geneva Place, Bideford; Lucy Elizabeth at North Street, Northam; Agnes Walker at 2 Fore Street, Northam. And 2 Fore Street is a few doors from 26 Fore Street, where Mary Septima stayed with her sister Edie on her last visit home in 1935 An address is what a census is searched on, and this archive has the 1871, 1891, 1901, 1911 and 1921 censuses and the 1939 Register already worked. These three should place the household in each of them without a name search. TESTED 17 September 2026. 2 FORE STREET, NORTHAM: CONFIRMED — the 1891 return (piece 1787, ED 4, folio 42, schedule 2) gives the household's address as "2, Fore Street", so the tree's birth address for Agnes is the family's real address two years later. 23 GENEVA PLACE and NORTH STREET: NOT TESTABLE FREE — they need the 1871 and 1881 censuses and FreeCEN holds neither for Bideford or Northam (1851, 1861, 1891 only). AND THE ROW'S METHOD DOES NOT WORK ON THE FREE CENSUS AT ALL: FreeCEN HAS NO ADDRESS FIELD. Its form searches names, years, place, occupation, sex, marital status — an address can only be READ OFF a return a NAME search found , which is how Fore Street was confirmed. Needs findmypast or Ancestry. See data/devon-addresses.tsv. REOPENED 17 September 2026, and my own verdict corrected. The 'not testable free' above tested the census route and nothing else. The UK National Archives had never been searched in this project at all, and Discovery shows Geneva Place, Bideford is a council school — North Devon Record Office 629C holds its log books and ADMISSION REGISTERS, 1885–1949, with Kew's ED 21/26168 and ED 70/427 alongside. Admission registers name the child, the date of birth, the father and the address. That cannot reach Horace, born 1872 and too old for them, but Lucy Elizabeth (1881), Mary Septima (1884) and Agnes Walker (1889) were all of school age inside the run.

PART DONE, 16 September 2026. The ascent was published unsourced below 1843. It is not any more at its lower end. FreeBMD — the free GRO index — was made to work (the search button is an <input type=image>, so form.submit() silently re-renders the form; it has to be clicked) and four claims were put to it. John Brailey's death: CONFIRMED — Deaths, Jun quarter 1851, BRAYLEY, John, St Thomas, 10/177, and the Devon County Asylum at Exminster is in St Thomas district. A search under BRAILEY finds nothing in 1851 at all — the spelling matters. The 1870 marriage: CONFIRMED — BRAYLEY, Mary Ann H, Bideford, 5b/794, double-keyed; and TAYLOR, William is on the same page , which is the join that attaches this whole line to the Taylors. Her 1900 death: CONFIRMED, and it pays a dividend — TAYLOR, Mary Ann H, age 56, Bideford, 5b/363. Age 56 on 21 April 1900 is exactly a birth of 20 May 1843 , which corroborates the birth year from a second direction and kills the 2018 backup's 1844. Her 1843 birth registration: NOT FOUND, control passed — but registration was not compulsory until 1875, so that null is weak. See data/brayley-bmd.tsv. WHAT IS STILL OPEN: everything above generation 3. Westleigh, September 1799 — Robert Brayley born the 21st, dead the 23rd, his mother Ann Snow dead the same day — is still the densest single page, and it needs the parish register, not the index. FamilySearch record search was blocked all evening and FreeREG does not hold Westleigh; see data/devon-tooling.tsv.

This is now the cheapest decisive document in the archive, and it was not identifiable before tonight. The GRO index puts Mary Ann H Brayley and William Taylor on the same page, Bideford, Mar quarter 1870 — but the page carries four names (the others are Huxtable Elizabeth and Johns Thomas), which is two marriages, and the index does not say which pairs with which. More importantly: a marriage certificate names BOTH FATHERS. Mary Ann Hellen's father is the entire question at the foot of the Brayley ascent — the tree says John Brailey and cites a stranger's web site. One certificate confirms or destroys him, and fixes the Taylor pairing in the same line. Ordered from the GRO at gro.gov.uk (PDF, about £3 for the record-only option where available; £12.50 for a certificate).

PART DONE the same evening it was written, 16 September 2026. The three 1900 deaths were run and all three are in the index, all three in BIDEFORD, all within nine months of each other: Georgina Brayley, age 85, Mar quarter, 5b/421 — the Scotswoman, and 85 is exactly right for a birth of 23 August 1814 at Haddington ; Mary Ann H Taylor, age 56, Jun quarter, 5b/363; and John Walker Brayley, age 53, Dec quarter, 5b/305. His age is the one thing tonight that did NOT match — 53 at 22 October 1900 means born about 1847, and the tree says 22 June 1845. Two years out, in a family whose other two ages were exact to the month , which points at the tree rather than the informant. And neither he nor his sister has a registered birth at all — no John Walker Brayley 1843–1850 anywhere, no Mary Brayley or Brailey in Jun 1843. That is a household that did not register births , which was lawful before 1875 — so the place to look is the parish baptism register, which is where "St Paul's" and "St Bartholomew's" pointed all along. THEN THE NINE TAYLOR CHILDREN, IN ONE SEARCH — births, surname Taylor, Bideford district, 1870–1890. NINE OF NINE, every one in the right quarter , including Mary Septima, Dec quarter 1884, 5b/506 — Pablo Armando's wife now has her own birth registration. The tree's odd-looking pair of November 1870 births turns out to be right: both registered that quarter on adjacent pages, 489 and 490, so they are twins. But the names are the find, not the dates: Isabella TEMPLE (1874), Margaret Hellen W[ALKER] (1876) and Agnes WALKER (1889). Those are the surnames of Georgina Temple and of Helen Walker, who five generations up cited nothing at all. She is still not on a record — but three of her descendants were registered with her surname by a clerk, and that is. AND A THIRD SPELLING TRAP, LOGGED BECAUSE IT NEARLY LANDED. A county-wide search of all 129 Brayleys in FreeCEN's Devon for 1861 found none of this family, with a clean control — 53 Taylors at Bideford the same year. It was one sentence away from being recorded as "the family left Bideford between 1851 and 1870". Searched again as BRAILEY, every one of them came back, on the same street as in 1851. A controlled null is only ever as good as the spelling it was run under. FOURTH PASS, 17 September 2026 — the sibling DEATHS, which were the unfinished half. HORACE ARCHIBALD: Deaths, March quarter 1934 — TAYLOR, Horace A, AGE 61, BIRMINGHAM. A death on 18 January falls in the March quarter and a man born April 1872 is exactly 61 that January. The district is the surprise — Birmingham, not Devon — but he was a FOOTMAN at nineteen in 1891, and servants went where the places were. WILLIAM JOHN, the twin: Deaths, June quarter 1951 — TAYLOR, William J, AGE 80, TOTNES. Quarter, age, initial and county all right. Of the twins born in one quarter of 1870, one died in Devon at eighty and the other went to Cape Town. AND A METHOD RULE THAT NEARLY COST BOTH: the first search for each man returned NOTHING. "Horace Archibald" and "William John" both give no rows; both men are there under Horace A and William J. The GRO index abbreviates second forenames, so a FULL-forename search is WEAKER than a first-forename one. Search the first forename and filter by AGE. Controls held both times — 34 Horace Taylors in the decade, 205 William Taylors in 1951 at >99%. The coverage was never the problem; the query was. AND THE SISTERS, 17 September — which needed a different method and cracked two of them. A married woman vanishes from the death index under her maiden name and this archive holds no husbands. That looked like a wall; it is not, because the MARRIAGE index still holds her maiden name. KATHLEEN GEORGINA — never married: Taylor, Kathleen G, age 69, Mar q 1940, Bideford. Born and died in the same town. ISABELLA TEMPLE — married Sep q 1900 at Bideford, 5b/901, a page carrying four names, so the index could not say which man was hers. Her death settles it: DALLING, Isabella T, age 79, Dec q 1953, Bideford. She married HENRY DALLING — disambiguated by her own death fifty-three years later. LUCY ELIZABETH — married Jun q 1909 at Totnes, 5b/373, a page carrying only two names: William Norrish SWEET and herself. Died Sweet, Lucy E, age 75, Jun q 1957, Plymouth. Two husbands and two married surnames the archive did not have. And Totnes is not random — her brother William John died at Totnes in 1951. THE METHOD, for Irene, Margaret Hellen Walker and Agnes: death under Taylor → if absent, that is a finding; marriage under Taylor → take vol/page → search the page with no surname → the husband appears → death under his surname, confirmed on age. If the page holds four names, the death resolves the pairing. COMPLETED 17 September 2026. EIGHT OF THE NINE now have a marriage, a death, or both. Irene → Henry Albert BURTON (Bideford 1900, 5b/900; d. Barnstaple 1969, aged 90). Agnes Walker → Alfred S WOOD (Devonport 1918, 5b/588; d. Cheltenham 1974, date of birth 3 August 1889 — her birthday to the day). Margaret Hellen Walker married at St Pancras, Dec q 1903, 1b/43 — husband unresolved, either Charles William Bell or Arthur Wallis, and with no death date there is nothing to disambiguate on. And a detail only visible with the pages side by side: IRENE MARRIED ON 5b/900 AND ISABELLA ON 5b/901 — adjacent pages, same quarter, same district. Two sisters married at Bideford within days of each other in autumn 1900, the year their mother died. FOUR RULES ABOUT THIS INDEX ARE NOW IN brayley-bmd.tsv: search the first forename only; a woman missing from the death index has married and the MARRIAGE index still holds her maiden name; when a marriage page holds four names the DEATH resolves the pairing; and the index changes shape twice — from 1912 it names the spouse, from ~1969 it gives a date of birth instead of an age.

DONE the same evening it was opened, 16 September 2026, and it gave more than it was asked for. FreeCEN holds Northam for 1891. The household is at 2 FORE STREET, NORTHAM — piece 1787, ED 4, folio 42, page 1, schedule 2. See data/1891-northam-census.tsv. WILLIAM TAYLOR, head, 65, "Cab Driver / ret'd Army Srgt", born Worlingworth, Suffolk — which also solves a puzzle filed as a data error: the tree attaches "Sergeant-Major in the army" to the infant who died at nine months in 1825, and the trade belongs to this man, on a census return, in his own house. MARY A. H. TAYLOR, wife, 47, born ST BARTHOLOMEWS, LONDON (CITY) — and that is the birthplace FreeBMD could not find a registration for. Three independent records now agree on 1843 and not one of them is her birth certificate. WILLIAM J. and KATHLEEN G., both 20, both born Bideford: the twins are confirmed for the third time. WHAT IT DID NOT SETTLE: there is no Edith, so that candidate is weakened — but MARY SEPTIMA, aged 6, is missing from her own parents' household too , and a county-wide check of all 368 Devon Taylors born 1882–1886 does not find her. A census cannot tell a child who was elsewhere from a child who never was , and a child who died before 1891 would fit every fact. The count still demands a seventh daughter. Row 128. MARKED DONE 20 September 2026 — it had been finished for four days and left open. The census work is complete and the one question it could not settle, the seventh daughter, is row 128's and is already written there. A row whose remainder lives in another row is finished, not open.

Two questions that did not exist this morning, both raised by records rather than by trees. (1) Mary Septima Taylor, born Northam, registered Dec quarter 1884, is not in her parents' house at 2 Fore Street on 5 April 1891 , and is not among the 368 Taylors born 1882–1886 in FreeCEN's Devon. The null is soft — FreeCEN's Devon 1891 is transcribed parish by parish and is not complete — but Northam itself IS complete, which is what makes her absence from it real. Her elder sister Isabella Temple, 16, is absent too, and at that age service explains it; at six it does not. (2) The tree says ten children and names nine, and SEPTIMA MEANS SEVENTH while the known daughters make her sixth. THE CHEAPEST TEST OF BOTH: the 1901 census at Northam , where she would be 16 — FreeCEN's place list should be checked for Northam 1901. And for the tenth child, FreeBMD deaths in the Bideford district 1871–1890 under an infant age , because a child who died before 1891 would be in the birth index, counted among the ten, and in no census at all. UPDATE, SAME EVENING. The first suggestion above is dead: FreeCEN lists Northam for 1851, 1861 and 1891 only, so there is no 1901 there. But the same place list held Westleigh for 1851, and chasing that instead found the whole family at LOWER GUNSTONE, BIDEFORD on 30 March 1851 — see data/1851-bideford-brayley.tsv. It answered three other open questions and raised one new person. John Walker Brayley is aged 5 there, which fits the tree's 1845 and not the 1900 death index's implied 1847, so that discrepancy resolves in the tree's favour; Georgina Margaret Brayley is 8 months old, so born about July 1850, and the tree's "1851" is a year late ; and there is an ISABELLA BRAYLEY, 10, born London (St Andrew), whom this archive did not have at all. The two questions in this row are still open. SECOND UPDATE. The obvious hypothesis was tested and it FAILS. Georgina Temple was alive and in Northam in 1891 — found only by searching BRAYLEY, the fourth spelling trap of the evening — and she is ALONE in the ALMS HOUSE (piece 1787, ED 6, folio 72, schedule 40), "Living On Means", aged 77. Mary Septima was not with her grandmother. She is not at 2 Fore Street, not in the alms house, and not among the 368 Devon Taylors born 1882–1886. Where she was at six years old is a real open question. WORKED 17 September 2026 AND HANDED OVER, because it is structurally beyond this session. The infant-death search this row asked for was run: FreeBMD deaths, Taylor, Bideford district, 1881–1891 — 37 entries, fourteen of them aged 0–3. An ETHEL Taylor, born Bideford Dec quarter 1882, died aged two, fits the gap exactly — born between Lucy Elizabeth and Mary Septima, dead before any census could hold her. And she cannot be attached to this family, and neither can Edith, and neither can the other five infant girls. THE REASON IS STRUCTURAL: the FreeBMD birth index carries NO mother's maiden name and the death index carries NO parents, so there is no field anywhere in FreeBMD capable of joining a Bideford Taylor child to THIS household. More searching there will not help. WHAT SETTLES IT IN ONE SEARCH, AND IT IS YOURS: the GRO's own index at gro.gov.uk is NOT FreeBMD and gives the MOTHER'S MAIDEN NAME on every birth from 1837. Search births, surname TAYLOR, mother's maiden name BRAYLEY, district BIDEFORD, 1870–1890. That returns every child of William Taylor and Mary Ann Hellen Brayley by definition, with no inference at all — the tenth included, whoever she was. It needs a free GRO account, which this session does not create. The second half of the row — where Mary Septima was at six in 1891 — remains open and is not answerable from FreeCEN, whose Devon 1891 is transcribed parish by parish.

ANSWERED 8 OCTOBER 2026. David logged this session in to gro.gov.uk and the search this row named was run: TAYLOR births, mother's maiden surname BRAYLEY, 1870-1893, both sexes, ten windows. NINE CHILDREN, ALL AT BIDEFORD, AND NOTHING MISSING: Kathleen Georgina (1870 D, 05B/490), William John (1870 D, 05B/489), Horace Archibald (1872 J), Isabella Temple (1874 D), Margaret Hellen Walker (1876 D), Irene (1879 M), Lucy Elizabeth (1881 S), Mary Septima (1884 D, 05B/506) and Agnes Walker (1889 S).

THE TREE'S TEN IS WRONG, EDITH AND ETHEL ARE NOT THIS COUPLE'S, AND IRENE HAD NO TWIN. And Mary Septima is the sixth daughter and the eighth child — the name matches neither count. See /corrections/.

Found on the 1851 census at Lower Gunstone, Bideford: BRAYLEY, Isabella, granddaughter, 10, born London (City), ST ANDREW — an elder sister of Mary Ann Hellen Brayley, born about 1841, who appears nowhere in the current tree. She is worth chasing for two reasons beyond simply being a person. (1) Mary Ann Hellen named her second daughter ISABELLA TEMPLE — registered at Bideford, December quarter 1874 — so she named a child for this sister and gave her their mother's maiden name too. (2) Sue Rowbotham's family history has a chapter called "The mystery of the missing husband and daughter". The missing husband is now firmly John Brailey, alive in the asylum on census night and dead five weeks later. Whether Isabella is the missing daughter is NOT established and is written here only so it can be tested, not assumed. Where to look: FreeBMD births, Brayley or Brailey, London, 1840–1842 — St Andrew Holborn is the likely parish; then her marriage or death. UPDATE, SAME EVENING: SHE IS CONFIRMED. The 1861 census at Lower Gunstone, Bideford has BRAILEY, Isabella, daughter, unmarried, 20, NEEDLE WOMAN, born London (City), St Andrews — the same birthplace as 1851, ten years apart, so she is a real person and not a transcription ghost. See data/1861-bideford-brailey.tsv. What is still open is what became of her: a marriage or a death, and whether she is Sue Rowbotham's missing daughter — which remains untested and must not be assumed. ANSWERED 17 September 2026, and the answer is a controlled silence. No marriage under Brailey or Brayley in Devon 1861–1890; none anywhere in England and Wales 1858–1895; and no death under either spelling 1861–1925. Control: Brayley marriages with no forename in 1870 alone return FOURTEEN nationally — the surname is thickly indexed and the searches could have found her. She did not marry in England and did not die in England. Either she emigrated, or she married or died under one of this surname's manglings. And a candidate, offered as a candidate: of John Brailey and Georgina Temple's four children, Mary Ann Hellen, John Walker and Georgina Margaret all have documented later lives. Isabella alone has nothing after 1861, which makes her the best candidate for Sue Rowbotham's "missing daughter" — an argument from absence, the weakest kind, and recorded as untested. See data/brayley-two-questions.tsv.

Found in the 1891 census at Tower Street, Northam — piece 1787, ED 3, folio 36, schedule 107 — BRAYLEY, Georgina, servant, single, 20, General Servant, born Middlesex, LONDON KILBURN, in the household of Thomas Lock, draper. The pattern is the family's exactly: Mary Ann Hellen herself was a "House Servant" in a china dealer's house at Bideford at eighteen. And the forename is the family's — Georgina Temple's name is carried by Georgina Margaret (1850) and Kathleen Georgina (1870). The obvious reading is a granddaughter of Georgina Temple, most likely a daughter of JOHN WALKER BRAYLEY , who was born in London, was 26 in 1871, and whose birthplace the records give as St Pancras. NONE OF IT IS PROVEN. The census states no relationship to anybody in this family. What would settle it: her birth in the index, London, about 1871, under Brayley or Brailey; and whether John Walker Brayley married at all — the archive has no marriage for him, and he died at Bideford in 1900. REFUTED 17 September 2026. John Walker Brayley married in the December quarter of 1876, at Bideford, 5b/978 — five years AFTER a child born at Kilburn in 1871. She cannot be his daughter by that marriage. The reasoning that produced her — a Brayley of the right age, in service at Northam, carrying the family's forename — was a resemblance and nothing more, and she is not adopted. What the same search bought: page 5b/978 carries Bonning Elizabeth, Brayley John Walker, Paddon Elizabeth, Richards John — two marriages. So John Walker married an ELIZABETH, either Bonning or Paddon , and the index cannot say which, exactly as on his sister's 5b/794 six years before. The certificate would settle it and would name his father.

DONE 7 October 2026, and deliberately NOT by replacing the prose. The page is 1,429 lines and thirty-three sections, several of them tables of many searches. Flattening that into five fields would have thrown away the reasoning, which on this page is the asset - it is the archive's record of how it was wrong. So data/searched-index.tsv holds thirty rows that INDEX the sections, each linking to the prose that explains it, and the sections stay where they were. The count the page now leads with: fourteen of the thirty returned NOTHING. That is the number the kit's own docstring says a register exists to show - an archive that records only its hits is keeping a highlight reel. The outcome words are this archive's own - found, nothing, mixed, blocked, answered, unresolved, abandoned - because the component insists on that and says flattening four archives' vocabularies to a shared five would be inventing a finding. Three sections are deliberately not rows: Why a page of failures, What is on this page and What the failures add up to are the page's argument about itself, and giving them an outcome would be a category error. when is coarse where the prose is coarse: a bare year means the section does not date itself, and guessing a day to fill a column is how a register starts lying.

DONE 7 October 2026. 188 lines of prose became nine rows in data/open-questions.tsv, rendered through the kit's Questions component: seven open questions and two blocked routes, each with what is known, what has been tried and the specific record that would answer it. The page now filters by state and is countable by a gate. THE RISK WAS LOSING THE NEGATIVE EVIDENCE, and it was checked phrase by phrase against the build. The long panel under question 1 - the Lerena-Lerena compound test, the forty-three transcription doublings, the Carmen Lerena act off film 007713392 image 2511, and the sentence saying the test lowers the probability without settling it - is carried in full. Four questions have no tried field, and that is accurate rather than missing: the original prose listed only what to do about them. One trap for whoever edits that file next, written into its header: the href COLUMN is passed through u() and gets the /TheLerena base, while links inside the HTML fields are raw and must carry it themselves. Writing both alike produced six /TheLerena/TheLerena/ links and the link gate caught every one.

Found while making the record links on /enrolment-card/ prominent. The markup writes <div class="grid two"> in two places, and the stylesheet has .grid { display:grid; gap:14px } and NO .two rule at all — checked against every loaded rule, including inside media queries. So both of them have always been a single column, at every width , including the front page's "What this archive has established" cards. Worked around locally rather than fixed properly: the enrolment-card grid now carries an inline grid-template-columns:repeat(auto-fit,minmax(290px,1fr)), which is responsive and self-contained. The real fix belongs in the shared archive-kit stylesheet , which means the kit changes and six archives repin — see row 92, which already wants a repin for the /bloodline/ fix. Do them together, and drop the inline style when the rule exists. PARTLY OVERTAKEN, 17 September 2026. Repinning the kit to 8e565d9 to fix the bloodline chart also brought in whatever else had landed there, and the archive still has its own inline grid-template-columns on the enrolment-card grid. Check whether a .two rule now exists upstream; if it does, drop the inline style. If it still does not, the fix and the repin can go together next time. DONE 17 September 2026, and NOT by changing the kit. The repin to 8e565d9 did not bring a .two rule — checked. And by then the inline workaround had been written into FIVE pages , which is well past the point where it should be a rule. So .grid.two now lives in this archive's own site/public/styles.css, beside the .grid rule it was always missing from , as repeat(auto-fit,minmax(290px,1fr)) — auto-fit rather than a hard 1fr 1fr, so it collapses on a phone with no media query. The five inline overrides are removed. No kit change and no repin of six archives, which is why this was the right place to fix it. Verified: every .grid two on /uruguay/, /argentina/, /enrolment-card/ and the front page now renders two columns.

Opened 16 September 2026 out of David's question, "do we have a Juan Carlos Lerena / Lerina in Italian records?" The direct answer is no — no Giovanni Carlo or Gian Carlo under either spelling, controlled. The useful answer is that LERENA and LERINA are Italian surnames and they sit in PIEDMONT: 66 exact Lerena records in Italy clustered in the Diocese of Torino, and 114 exact Lerina, Torino again plus Ivrea. See data/italy-lerena-lerina.tsv. And that is the same corner the DNA points at. The archive's own DNA page reasons that Cheryl's 11.1% Italian, with three non-Italian grandparents, implies about 44% in Pablo Armando — one Italian parent — and that the split is NORTHERN-dominant, 6.1 against 3.6. Two independent things, an admixture estimate and a surname distribution, now point at northern Italy, and neither of them names anybody. WHAT TO DO WITH IT: the Torino and Ivrea civil registration and diocesan records are indexed on FamilySearch and are searchable by parent name — so look for a LERINA or LERENA male born about 1840–1862 in Piedmont who then DISAPPEARS from the Italian record , and cross him against the Argentine arrivals. The archive already holds cemla-lerina.tsv: a Lerina family of four landing at Buenos Aires from GENOVA on 15 March 1889 — seven years too late to be his parents arriving, but the right surname on the right route , and Genova is the port Piedmont emigrates through. Also worth re-running CEMLA for LERENA and LERINA before 1882, which is the arrival that would matter. ANTENATI CHECKED, 16 September 2026, and it cannot answer this. The Italian State Archives' own portal returns one Lerena in its whole national name index — Cesare Lerena, death 1915, BRESCIA — and eight Lerina, most of them the different surname Di Lerina. The control destroys that null: Antenati holds 2,621 registers for Torino, but a name search there for ROSSI restricted to Torino returns 102 people called Charles, Marguerite and Jacinte, in 1805–1811 — the Torino name index is NAPOLEONIC ONLY, kept in French, and does not reach the 1840s–1880s at all. So Antenati's images exist for Piedmont and are not name-searchable for the years that matter — the same lesson as the Rosario typescript and the FreeREG Devon parishes, that the null belongs to the finding aid and not to the past. The searchable collection for the right years is FamilySearch's "Italy, Torino, Torino, Civil Registration (Tribunale), 1866-1899" , which is where the 66 and 114 came from. Antenati's 2,621 Torino registers are the fallback: browsable by comune and year, image by image, the way Rosario and Canelones were read. WORKED IN DETAIL 16 September 2026, and the conclusion is negative in a useful way. 49 Lerina records in and around Torino — Ivrea, Montaldo Roero, Costigliole — families that stay put for sixty years , with two Carlo Lerinas acting as fathers and neither a candidate for anything. THE ROUTE IS STRUCTURALLY INCAPABLE of answering the question: the Turin civil series begins in 1866, so a man born in the 1840s–50s has no birth record in it at all; and an emigrant leaves no later Italian record by construction , so "a son who vanishes" cannot be told from an unindexed death or a move. Enumerating Piedmontese Lerinas maps the people who STAYED, which is the complement of what is wanted. Do not spend a week on it.

This session cannot run CEMLA. The buscador is an iframe onto search.cemla.com, and that form carries CaptchaCode and LBD_VCID_SampleCaptcha — it is behind a CAPTCHA, and this session does not complete CAPTCHAs. So it is yours. SETTLE THE COVERAGE FIRST, because it may make the whole search pointless. The earliest arrival this archive holds under any spelling is the Lerina family of 15 March 1889, from Genova. CEMLA's arrivals database is generally given as beginning in 1882 — and if that is right, then an arrival before February 1882, which is the only one that could carry Pablo Armando's parents into the country before his birth, is invisible to it by construction, and finding nothing would prove nothing. IF IT DOES REACH BACK FURTHER, the searches are: LERENA, LERINA, LARENA and LORENA, arrivals before 1882, and note the port — Genova is how Piedmont emigrates , and Piedmont is where both the surname and the northern-Italian DNA now point (row 134, data/italy-lerena-lerina.tsv).

REWRITTEN the same day it was opened, because the control demoted half of it. THE STRONG HALF STANDS: there was no LERENA and no LERINA household, under any spelling, anywhere in SANTA FE PROVINCE in the 1895 National Census. Controlled: a Gonzalez search returns 88,739 records with the 1895 census reaching San Cristóbal, Alberdi/Rosario, San Jerónimo and Iriondo, so the index is dense and covers Rosario itself. For a whole surname to be absent from a province indexed that thickly, it has to be genuinely absent — and that agrees with the Rosario baptism sweep from an entirely different record type. THE WEAK HALF IS WITHDRAWN AS EVIDENCE. This archive said earlier today that Pablo Armando's own absence from the 1895 census at thirteen suggested he had already left. It suggests nothing of the kind. One person can always be missed or mangled, and this person's name demonstrably mangles: the Argentine army indexed his own card as "ARMAND LORENA", born "Rosario de Lerma, SALTA". A boy who becomes Armand Lorena on one index can be anything on another. SO THE HONEST POSITION IS THAT NOTHING DATES HIS DEPARTURE: born Rosario February 1882, son born Cape Town November 1905, twenty-three years with no record inside them. The live routes are the ones hypothesis 2 already names — UK outward passenger lists (BT 27), indexed from 1890 — and the consular file behind the 1940 passport. And note the archive already reasoned that the South African War shipped some 360,000 horses in, Argentina among the sources, when he was 17 to 20 and a horseman.

This came out of measuring the four departments, and it is worth more than any of the names that measurement failed to turn up. The archive's Uruguayan baptism layer, uruguay-baptisms-lerena.tsv, is a FINDMYPAST FREE RESULT LIST — its own header says "33 index rows... transcripts are behind findmypast's paywall and were NOT opened" — giving names, years and parents' forenames only. Twenty-two rows: eleven Trinidad, four Montevideo, three San José de Mayo. Beside it, free, sit 1,343 exact LERENA records across the same four departments with parents in full and frequently a View original document link. For Trinidad the 608 split roughly half Catholic Church Records 1726–2000, half Civil Registration 1879–2020, and only a couple from Uruguay Baptisms 1750–1900 — which is the index findmypast was charging to transcribe. Fair to the earlier work: the register does cite Catholic Church Records for particular people, so the collection was known — it was used person by person, never swept. THE JOB: take Flores, San José, Florida and Colonia one collection at a time, harvest every exact LERENA with its full parentage, and build the Trinidad kindred upward. That is the difference between knowing Bernabé married a Casimira Pelaez and being able to put a generation above either of them — which is the only way the Trinidad nest will ever be joined to Sancho, or shown to be a separate family. RUN ONCE, 17 September 2026, AND IT WORKED THE FIRST TIME. Searching by parent instead of by person turned up MAURICIO LERENA and CASIMIRA MONTAÑO , a couple with at least seven children across San José de Mayo, Trinidad, Florida, Salto and Montevideo. One of them is CIPRIANO LERENA MONTAÑA, and Cipriano is one of the six Trinidad fathers — so the archive now has a generation above the nest for the first time. Another is CLARA LERENA, b. 1848, d. Montevideo 3 March 1925 certificate 47, which is almost certainly the tree's "Demetria Clara Lerena Montaño" — and the record explains the compound name, because her mother was Montaño. See data/mauricio-casimira-montano.tsv. Graded honestly: Cipriano's card says Image Unavailable, so it is an INDEX and not an act. The method is proved and the sweep is barely begun — Alejandro Lerena, Emilio Lerena and Alejandro F. Lerena all appear as fathers in the same window and none has been worked.

UPDATED 9 October 2026, AND THE SWEEP'S OWN TARGET LIST IS NOW HALF SPENT. This row ended "Alejandro Lerena, Emilio Lerena and Alejandro F. Lerena all appear as fathers in the same window and none has been worked". Alejandro Lerena has been worked, and at plate level. There are two of him — father and son — and both are now in the register with acts: the elder married Margarita Fernández Álvarez at San José on 16 September 1825, the younger remarried as a widower at Trinidad in December 1854. The elder's children are Socorro, Alejandro and Bernabé, each from a separate document. EMILIO and ALEJANDRO F. REMAIN UNWORKED and are what is left of this row. See data/trinidad-founding-kindred.tsv and data/colla-trinidad-bridge.tsv. BOTH REMAINING MEN WORKED, 11 OCTOBER 2026 — AND NEITHER WAS ONE MAN. EMILIO. There are at least three. One is the Spanish immigrant from Logroño already in this register, arrived in Argentina 1930, irrelevant to the sweep. One is EMILIO ISIDORO LERENA RECUERO, born 2 January 1906 at Trinidad, son of Florencio P. Lerena Fernández and Juana Recuero — so the sweep's 'Emilio' is a son of that generation, not a father of it. And one is EMILIO LERENA × PAULA CASARES, father of Emilio Delfino and Clotilde Reina, who cannot be the 1906 man because his son has a marriage registration at Flores in 1931. That third man is the father row 137 met. He is attested and not placed: the index-card series carries no birth dates and no ages, which is exactly why these men could not be separated before. ALEJANDRO F. Worked as far as the index allows, and the result is thin but real: he is the husband of NATALIA LERENA and the father of AURORA LERENA (b. c.1868), named on Aurora's children's Trinidad registrations — a fourth Lerena-married-a-Lerena couple. No record of his own has been found: a forename sweep of Alejandro returns 533 Uruguayan hits and he is not among them. He exists only as a named parent. WHETHER HE IS A SON OF ALEJANDRO LERENA THE ELDER × MARGARITA FERNÁNDEZ ÁLVAREZ IS A HYPOTHESIS AND IS ARGUED, NOT ASSERTED, in data/_person-pairs.txt — the 'F.' would then be Fernández, and this branch does produce Lerena Fernández compounds. ROW CLOSED ON ITS OWN TERMS — its stated remainder was these two men — but the larger sweep it describes, four departments one collection at a time, is NOT done.

ANSWERED 17 September 2026, in two searches, and the answer is a SEPARATION. MAURICIO LERENA, born at COLLA, married Casimira Montaño at San José de Mayo on 3 February 1837 — father Ramón, mother María Gonzales (ark 1:1:F21B-SZ5). And his brother: ALEJANDRO LERENA, born at ROSARIO, married Margarita Fernández Álvarez at San José on 16 September 1825 — same parents, Ramón Lerena and María Josefa Gonsales (ark 1:1:F21Y-JRV). That marriage also closes a circuit the archive opened on 14 September: the-three-alejandros.tsv had an "Alejandro Lerena of the Trinidad orbit" with Severa Fernández and a Bárbara Lerena and could not place him — the Fernández is his wife and the Bárbara is his mother-in-law, Bárbara Sotura. THE STRUCTURE: Ramón Lerena × María Josefa González of Rosario/Colla → Alejandro, Mauricio, Manuel Lorenzo (b.1812 Unión) and Luís Manuel (b.1820) → and TWO OF THE SIX TRINIDAD FATHERS, Bernabé and Cipriano, WHO ARE FIRST COUSINS. AND THE ANSWER: RAMÓN IS NOT SANCHO'S SON. Alejandro married in 1825, so was born about 1800–05, putting Ramón's birth about 1775–85. Sancho was baptised about 1767 and married in January 1791, his first child born 1793. A man born in 1775–85 cannot be the son of a man who married in 1791. Ramón is Sancho's contemporary, about ten years his junior. So there are TWO documented Uruguayan Lerena families, not one. See data/ramon-lerena-rosario.tsv.

This is now the best-defined question in the whole Uruguayan file, and it turns on a single phrase. Ramón Lerena married María Josefa González and was fathering children from about 1800 (Alejandro, b. Rosario) to 1820 (Luís Manuel). So the marriage falls about 1795–1810, most likely at ROSARIO in COLONIA or at SAN JOSÉ. Uruguayan marriage acts state where the groom is NATURAL DE, and that is the whole question. Sancho Llerena was natural de Córdova la Llana. If Ramón is Córdoban too, the two families are brothers or cousins and the surname entered the Banda Oriental ONCE. If Ramón was born in the Banda Oriental, they are two separate arrivals and may not be related at all. Where to look: Uruguay Catholic Church Records 1726–2000 and Uruguay Marriages, free and indexed, searching Ramón Lerena as groom; and the Rosario (Colonia) parish registers, which this archive has never opened. And a loose end worth picking up in the same pass: the current tree gives Gabino of Flores as a son of "RAMÓN LLERENA CUBAS and Margarita Leal" — a different wife. Either Ramón married twice, or there are two Ramóns, and the tree may be carrying a version of this same man. ADVANCED 17 September 2026, and it went one rung FURTHER than asked. The marriage was not found; Ramón's PARENTS were: SALVADOR GERENA and ANA DE CUBAS , named on the baptism of his son Manuel Lorenzo at Unión, Montevideo, 15 November 1812 (ark 1:1:6D7Z-BD56) — because Montevideo acts name grandparents as a matter of course. THE IMAGE IS AVAILABLE AND HAS NOT BEEN READ. That also explains the tree's "Ramón Llerena CUBAS": Cubas is his mother's surname. Second time in two days a record has explained a compound name the tree merely carried. AND IT SETTLES HALF THE QUESTION: SANCHO AND RAMÓN ARE NOT BROTHERS. Sancho's father was Félix, Ramón's was Salvador. They may still be cousins, and the one thing that keeps it worth chasing is the SPELLING: both fathers are written GERENA in Montevideo acts a year apart — 1811 and 1812. A scribal habit of one place and time, not two families choosing the same variant by chance. A resemblance, not evidence. WHAT REMAINS: read the 1812 act, which may say where Ramón was natural de and close it in one reading. CLOSED 17 September 2026 by row 140. The marriage was never found, and it did not need to be: the 1812 act states Ramón is natural de VALDEMORO en CASTILLA LA NUEVA. Not Córdoba. The two houses are two Spanish emigrations, and the question this row existed to ask is answered.

READ 17 September 2026, and it closed the question in five words. DGS 007713375, image 02668 — which is the film already sitting on this work list as row 5. "…hijo legítimo de RAMON SERENA natural de VALDEMORO en CASTILLA LA NUEVA, y de María Josefa González natural de la Ysla de Canarias, LANZAROTE. Abuelos paternos Salvador Serena y Ana de Cubas; id. maternos Marcial de León González è Ysavel Lorenza." VALDEMORO IS 27 KM SOUTH OF MADRID. SANCHO WAS NATURAL DE CÓRDOVA LA LLANA, IN ANDALUSIA, 500 KM AWAY. So the surname entered the Banda Oriental twice, from opposite ends of Spain — two emigrations, not one kindred. And the Gerena/Gerena coincidence, which this archive flagged as "a reason to look and not a reason to believe", is discharged: the act writes SERENA, and the two men were baptising children at the SAME FONT one year apart. The resemblance was real and it meant nothing. The act also beat the index nine facts to four: the father's birthplace, the mother's birthplace, the real maternal grandmother (Ysavel Lorenza, whom the index never named and for whom the child Manuel Lorenzo is probably named), and the church — the index says "Unión" and the act says the Iglesia Matriz. See data/1812-union-act.tsv and /uruguay/.

What the 1812 act bought in exchange for closing the join. An hour ago Ramón's house had no origin at all; it now has two, both precise. ONE: VALDEMORO, in Castilla la Nueva, 27 km south of Madrid. Salvador Serena and Ana de Cubas were married in Spain and Ramón was baptised there about 1775–85 — both findable, because Valdemoro is one parish and not a province. FamilySearch holds Spanish diocesan records for Toledo and Madrid; PARES and the Spanish diocesan portals are row 78. TWO, AND NOBODY SUSPECTED IT: MARÍA JOSEFA GONZÁLEZ IS A CANARIA, FROM LANZAROTE. Her parents are Marcial de León González and Ysavel Lorenza. The Canary Islands fed a migration stream of their own into the River Plate and no part of this archive has ever looked at it. A CAUTION BEFORE ANYONE SPENDS A WEEK: none of this is Pablo Armando's line. It is Ramón's house, which the same act has just proved is NOT Sancho's, and neither house has yet produced a Juan Carlos. Work it because it is real and cheap, not because it is his. TESTED 17 September 2026 AND BLOCKED, with a control. FamilySearch's catalogue holds exactly ONE item for Valdemoro and it is a CENSUS ; a catalogue search for church records there returns "No Results Found" . The control is the census: the place resolves and the search works, so the null is about the holdings. FAMILYSEARCH HAS NO VALDEMORO PARISH REGISTERS. And a warning repeated because it has now misled twice: an indexed search for SERENA filtered to "Valdemoro, Madrid, Spain" returns 79 hits and not one of them is Valdemoro — Madrid city, Toledo, Murcia, Almería, Huesca. The place filter silently widens to the province, exactly as it did for Santa Fe. Check the place field on the rows. THE ROUTE IS DIOCESAN, NOT FILMED: Valdemoro is historically Archdiocese of Toledo, since 1991 Diocese of Getafe, and PARES is the state portal. That is row 78 — which until today said only "try PARES and the Spanish diocesan portals" and had no place to try them on. It now has one, and a precise target: Ramón's baptism about 1775–85, and the marriage of Salvador Serena and Ana de Cubas before it.

The Mazza fault was looked for here and found in a different shape. Mazza kept evidence in files its person-builder never read. This archive had no person-keyed evidence file AT ALL: build-people.py read lerena-register.tsv and relations.tsv and nothing else, so a whole session of records lived on topic pages while the people they were about had pages saying "No relationships are recorded for this person." AND THE GATE FOUND A PRE-EXISTING ONE ON ITS FIRST RUN, which is the real vindication. /people/carlos-lerena-salvanach/ carried that sentence while relations.tsv held five edges for him — father Gilberto, mother Julia Salvanach, wives Rosa Maria Eppens and Adela Camiglia, son Avelino Carlos. The cause is on his own page: the row was corrected on 14 September from "Retiro baptism" to "información matrimonial", and the rename fixed the register and orphaned every edge. relations.tsv does if not subj: continue — it drops unknown names in silence. All five repointed and recovered. THE FIX: data/person-evidence.tsv (person-keyed, same name convention as relations.tsv); build-people.py reads it and raises on an unknown name rather than skipping; the person page renders What has been read about this person and no longer prints the false claim when evidence exists. THE GATE: tools/check-evidence.py, wired into verify.sh and into npm run build so CI runs it. It refuses if a named person is not in the register, if their page does not carry their evidence, if the page still prints the false claim, or if relations.tsv names anyone the register lacks. Verified by reintroducing the bug three ways — build stops reading the file, template guard removed, edge repointed at an unknown name — each exits 1.

Discovery has ADM 157/197/182, folios 182–186: John Walker Brayley, born Middlesex, attested for the Royal Marines in 1861 aged 18, discharged 1868 as an Invalid. This archive has its John Walker Brayley born 1845 at St Paul's, London — which is Middlesex — and has already logged that his death index gives age 53 in the December quarter of 1900, implying about 1847: a two-year gap left open. Aged 18 in 1861 implies about 1843. Three sources, three different years. An attestation paper states date and place of birth outright and would settle it. Not asserted to be him — that is what the document is for.

Wills name relationships, which is exactly what the Brayley ascent is short of. Kew: IR 26/337/185, abstract of the will of George Brayley, Gentleman of Northam, 29 May 1800. North Devon Record Office: B127-8/126/1-3, Emily Kate Brayley of Northam, will 1935, proved 1943; B127-8/139/1-12, Edith Mary Brayley of Northam, will 1943, proved 1948. The last two are within living memory of Mary Septima Taylor and may name her generation directly.

The route row 123 did not have. Log books and admission registers, 1885–1949. Registers of this period give child, date of birth, father's name and address — which would test the family's Bideford and Northam addresses without a census, and may catch Lucy Elizabeth, Mary Septima and Agnes Walker Taylor. Not online; needs an order or a visit.

Discovery shows a Brayley family of standing in these two parishes across 1767 to 1861 — George Brayley master of the brig Cato at Appledore in 1767, George Brayley of Bideford gent in leases of 1769, 1831, 1844 and 1861, George Brayley of Orchard Hill in Northam in 1827, and B194/277/1-5, valuations of G.B. Brayley's property in Northam, Bideford and Barnstaple in 1915 and 1929. Whether any of them is kin to Mary Ann Hellen Brayley (1843–1900) is untested and is deliberately not assumed. The wills in the row above are the way in.

Settled 17 September 2026. David signed the archive into a TNA account and ordered the record himself; it was free. And it does more than nail the rejection down — it replaces the argument. The rejection had rested on a census age. The document gives two harder reasons. First, the form asks for 'Names in full' and the answer is plain 'William Taylor' — ours is William John Taylor, and the Navy had no reason to drop a middle name it had just asked for. Second, and it is conclusive: on census night, 5 April 1891, this man was borne on HMS Impregnable at Devonport (joined 20 May 1890, left 4 October 1892) while William John Taylor was enumerated at home in Northam, aged 20. One man cannot be in two places on one night. Two men, proved. The near miss is closed rather than merely argued.

If the Taylor men went to sea it was not the Royal Navy — that is now a tested null. The merchant service remains open, but crew lists are indexed by SHIP, not by man, so there is no name search to run. A Taylor family of Appledore, in Northam parish, appears on the Ulelia (ON 74429) — T Taylor mate b.1846, J W Taylor master b.1888, T G Taylor b.1893, previous ship Annie of Bideford — joined to this family by nothing at all. Work out what a ship-by-ship search would actually cost in hours before committing to it. COSTED AND CLOSED 20 September 2026 — it costs nothing, because it must not be started. The 1891 census already answered it, and had been in this archive since 16 September: William John, 20, is a CAB DRIVER like his father; Horace Archibald, 19, is a FOOTMAN in domestic service. Neither went to sea, which also gives the ADM 188 null its reason. For the record, the search would have meant several thousand annual agreements across a hundred-odd Bideford and Appledore vessels, held in three institutions — and it could not have been done by name in any case, because no central register of merchant seamen survives between 1857 and 1918, which is exactly when either man would have sailed. See /taylor/.

TNA digital downloads are licensed for private study or research for a non-commercial purpose and for education. This archive is a public website. The line taken, and to keep taking: cite the reference, transcribe the facts, do not republish the scan. Facts are not copyright; the image is licensed. Anything wider goes to image-library@nationalarchives.gov.uk. Note also that download links expire 30 days from the order, so anything wanted long-term must be read and transcribed inside that window — ADM 188/195/140021 was ordered 17 September 2026.

Tested 17 September 2026, and the control was run first. The 1908 card is the only document naming Pablo Armando's parents. Searched: father exactly Juan Carlos Lerena with mother's surname Lerena — 0; the same without the mother — 0; Juan Carlos Lerena as a named person in any role, country or collection — 1, and it is an infant, Lerena Caraballo, died 21 February 1983 at Montevideo aged 0, a century adrift. Control, same form, Gilberto Lerena and Julia Salvanach: 11. The form works; the zeroes are the corpus. Limits stated on /direct-line/: this tests the exact string and indexed records only. Taken with the 1882 Rosario walk — 750 consecutive acts, no Lerena — the parents' generation is not in any indexed corpus this archive can reach.

Walked 17 September 2026 — all three, same day. See /argentina/ and data/three-routes-walked.tsv. None finds his parents. Route 2 was dead before it started and should not have been offered: in February 1882 Rosario had one parish. Route 1 is real but FamilySearch has no Santa Fe civil registration collection at all — Buenos Aires and Catamarca do, Santa Fe does not — so it needs an approach to the province. Route 3 gave two false positives and one prize: the film number for the Montevideo dispensations, which closes row 32.

Drafted 17 September 2026 and ready to send — the Spanish text is on /argentina/. The ask was rebuilt before drafting, because the original premise was wrong: civil registration in Santa Fe begins 1886, so there is no 1882 birth act and asking for one would have spent the letter. The strongest ask is now the parents' DEATH acts — Juan Carlos and María would have died after 1886, so their deaths would be registered, and an Argentine acta de defunción names the parents and spouse of the deceased. That reaches the generation above the wall from records that should exist. Also asks for a late birth inscription (he held an Argentine passport and was enrolled in 1908; both needed proof of birth, and no baptism exists) and, failing everything, a certificado negativo — which for this archive is a document, not a disappointment. Registro Civil, San Luis 2950, CP 3000, Santa Fe (verified on the province's own site). Try santafe.gob.ar/rcdigital first; its form would not load here, so whether it accepts a requester with no Argentine DNI is untested.

A full-text hit turned up 'Don Juan Carlos González Lerena' in the Montevideo dispensations, 8 November 1920. He is not ours. But González Lerena appears in neither compound-surname-sweep.tsv nor lerena-lerena-compound-test.tsv, so the compound sweep is incomplete by at least one variant — and compound forms are how this surname hides. SWEPT 20 September 2026, and it was not one variant — it was a family. 'Gonzalez Lerena' exact returns 52 records, headed by ESTEFANÍA LERENA, wife of Servando González, with at least six children across San José, Montevideo and Colonia, 1854–1931. None of it was in either compound file. She is attached to nothing and is recorded unplaced. It also finishes the false positive: Juan Carlos González Lerena, b. 1899, married Concepción Petrona Urbano Orio in 1922, son of Leopoldo — Estefanía's grandson, and seventeen years younger than Pablo Armando. See /uruguay/.

The indexed Santa Fe church collection holds exactly three Lerena entries in 783,111 records, and one is new to this archive: Armando Lerena, baptised 1917 at Nuestra Señora del Carmen, Santa Fe city, son of Casimiro Lerena and Amelia Lepiani. It carries the name Armando, which is suggestive and nothing more — 1917 is a generation past Pablo Armando and Santa Fe city is 170 km from Rosario. Recorded unplaced; test it, do not adopt it. TESTED 20 September 2026, and it is not a Santa Fe household. Casimiro Lerena and Amelia Lepiani — indexed once as Lapriani — have Armando Lerena in the Santa Fe church records and Ventura Amelia Argentina Lerena in the Entre Ríos records. So the family is littoral, across two provinces, and Entre Ríos is also where Gilberto and Julia Salvañach's 1885 child sits — the province is worth mapping in its own right. Attached to Pablo Armando by nothing whatever. The only thing that ever recommended this household was the forename Armando, which is not evidence, and it has not been treated as any.

Found by running a control rather than by searching for them, and all three are snippet only — see /south-africa/ and data/gazette-three-new.tsv. Transvaal Provincial Gazette 4793 (11 Dec 1991, p.75) and 4796 (18 Dec 1991, p.50), a two-week township advertisement naming Ricardo Juan Carlos Lerena in his lifetime — the archive holds no property fact for him, and Wysigingskema 2595 is the searchable handle. And Legal Notices B 18373 (24 Oct 1997, p.165), ABSA Bank Beperk v Lerena in the Witwatersrand Local Division — the snippet cuts off before the forenames, so which Lerena is unknown. All need one page-view each, once the viewer is reachable.

The last remainder of row 120. Three of the four departments measured there resolved to the children of one couple, Mauricio Lerena and Casimira Montaño, spreading across the interior in a single generation — San José de Mayo, Trinidad in Flores, and Florida, with Salto and Montevideo besides. Colonia did not. It holds 189 exact LERENA records and has no dedicated study, only passing mentions. And the standing warning applies to it hardest: the tree's Gabino of Flores is given as a son of Ramón Llerena Cubas and Margarita Leal — not Sancho's line — so the interior may hold a different Lerena family entirely, and Colonia is where that would show. Needs a FamilySearch session. PART-ANSWERED 20 September 2026, and the method was the problem. Maria Bernarda González Lerena married at Campana, Colonia in 1931 — so part of Colonia's Lerena presence is a compound family that an exact-surname count cannot see. The 189 figure was measured by exact surname and therefore under-counts. Before anyone concludes Colonia is thin, re-measure it including compound forms. CLOSED 20 September 2026, AND THE ROW'S PREMISE WAS WRONG - my own row, logged as a correction. Rosario-Colla was never unaccounted for: it is where the family came from. The archive's own notes already read Ramón Lerena × María Josefa González, of Rosario / Colla, fathering children c.1800-1820 and from Rosario-Colla across San José, Flores, Florida, Salto — and the 1837 San José marriage gives Mauricio's birthplace as COLLA outright. What was genuinely missing was that none of them were people here: Mauricio, his father Ramón of Colla, his mother María Josefa González and his wife Casimira Montaño have now been added to the register with their edges, from ark 1:1:F21B-SZ5, tiered INDEXED not image-read. Also new: Vitalino Lerena, born San Pedro, COLONIA in 1870, son of Vitalino Lerena and Andrea Echeverry — a documented link between Colonia and the Trinidad/Flores kindred.

133 of the 278 people in this register carry two or more forenames — very nearly half — and FamilySearch's exact forename search matches the whole given-name string, so an exact search on the first forename alone would never have found them. Any null recorded that way is unsafe. None could be identified by grep, because the searches were written down in prose without naming the parameter, which is the lesson underneath this row: record the query, not just the answer. Re-run suspects in the safe shape — fuzzy forename, exact surname — which does match per token. RUN 21 September 2026, and it found something better than a bad null. The archive's own blast-radius note says no affected null could be identified by grep, because the searches were written in prose without naming the parameter — so the sharpest available version was run instead: the CENTRAL null, re-done in the fuzzy-forename / exact-surname shape. Fuzzy 'Juan' + exact LERENA returns 29 distinct given-name forms, including Juan Carlos Esteban Lerena, Juan a. Lerena and Juan P. Lerena — the initial and three-forename shapes a whole-string exact match cannot reach, which is exactly what the trap predicted. None is the father: the Esteban man married in 1932. But the re-run falsified two published zeroes — see corrections.tsv. Juan Carlos Lerena is now an indexed person, as the father on his own son's military record.

Found while running a control. Marriage of Gilberto Lerena and Julia Salvañach, 11 August 1877, San Francisco de Asís, Ciudad Vieja, Montevideo. Index ark 1:1:X9PV-HKCR; image ark 3:1:3Q9M-C9BW-N74V-N. The index gives the bride's parents-in-law — Gilberto's parents — as 'Luís Lenguas' and 'Justiniana Lenguas', almost certainly a machine garble of Luis LERENA and Justiniana LENGUAS, which is exactly the couple this archive assigns him at tree tier only. A parish register would make it documented. Do not promote it off the index: the page says indexed by a computer, and the same entry gives the bride a father called Cristóbal Lerena, which cannot be right for a Salvañach marrying a Lerena. Open the image. IMAGE READ 20 September 2026, and it does document them. The act: "contrajo Don GILBERTO LERENA, Oriental, soltero, hijo legítimo de Don LUIS y de Doña JUSTINIANA LENGUAS" — and the padrinos are "Don LUIS LERENA y Doña ÁNGELA ÁLVAREZ", so Don Luis is named twice in the one act, as father and as padrino, which is the document corroborating itself. Gilberto's parentage moves from tree to documented, and he becomes a documented brother of Don Luis Eugenio Lerena Lenguas, whose parentage on the same couple rests on Montevideo act 258 of 1864. And the index was wrong exactly where predicted: Cristóbal is a Salvañach, the bride's own father — the surname Lerena was carried across from the padrino line. There is no Cristóbal Lerena, and none was added.

Surfaced twice in one session and never looked at. Casimiro Lerena and Amelia Lepiani have a daughter in the Entre Ríos church records while their son is in Santa Fe; and Gilberto Lerena and Julia Salvañach have a child recorded at Uruguay, Entre Ríos, Argentina in 1885. Two unrelated Lerena households in the same province, and this archive has no Entre Ríos study at all — only Rosario, Buenos Aires and the Uruguayan departments. Entre Ríos sits between the two, on the river, which is exactly the route this family used. THREE REASONS NOW, NOT ONE — 11 October 2026. (1) VICENTE LERENA DÍAZ keeps the civil register of Pueblo Brugo, departamento de Paraná, in 1938. (2) ANTONIO LERENA DÍAZ married at Crespo, departamento de Paraná, in 1924 — so two Lerena Díaz in one department, probably brothers. (3) VICENTE ANTONIO LARENA is at Gualeguaychu in the same province, and this archive can neither merge him with the first nor rule him out. Entre Ríos is no longer a blank on the map; it is three questions.

Recorded because it cost a wrong page today. The viewer displayed "Image 33 of 250, item 9" for ark 3:1:3Q9M-C9BW-N74V-N, and image 33 of that film is a baptism register, not the marriage. The ark's real position is image 01765, and the only way to get it is to ask the manifest: GET /service/records/storage/deepzoomcloud/dz/v1/{ark-without-prefix}/manifest.json returns alternateId. The number on screen is the position within the waypoint item, not the film image. Also worth having: dist.jpg gives the whole page at full resolution — 4467×3476 here — so no tile assembly was needed at all.

tools/drift.py flags it: the register now calls Mauricio "the patriarch this archive had researched but never made a person" and no single page says so to a reader. He has a person page, but the finding — that one couple at Rosario-Colla accounts for San José, Flores and Florida — lives in a register note and in row 120. It belongs on /uruguay/ beside the González Lerena section, which is the same shape of finding. DONE the same session it was raised. /uruguay/ now carries Rosario-Colla — not a gap, an origin, with all four people and the 1837 act.

The blind spot it had was real and is fixed: drift.py read only the .astro sources, so it could see a finding a page had typed and was blind to one a page rendered. That put it in direct conflict with check-onsite.py, which requires every data file to be imported by a page — onsite said render it, drift said no page says it, and drift was wrong every time. It now reads the built pages, excluding /people/, /register/, /who/ and /search/, which re-print the whole register and would make any finding pass trivially. What is NOT verified: deleting the /uruguay/ section that carries the Mauricio finding did not make the check report it again, even though no other non-dump page and no .astro source holds those words. So the fix is sound but the catch is unproven, and a gate whose catch is unproven is a gate nobody should trust. Find out why — most likely the rarest-word selection shifts as the corpus grows. VERIFIED BOTH WAYS, 20 September 2026 - and my own test was broken first. The first attempt rebuilt with npm run build >/dev/null 2>&1 and never checked the exit code — the very trap verify.sh carries a warning about. Re-run with the status checked, the build was fine and the catch still did not fire, so the fault was real. What defeated it was /worklist/, and the reason goes to the heart of how this archive works: findings are written into row notes as a matter of habit, and /worklist/ re-prints every note in full — so logging a finding counted as publishing it, which is exactly the gap this check exists to find. /register/, /people/, /who/, /search/, /worklist/, /changes/, /corrections/ and /searched/ are now all excluded: they are ledgers of the work, not places a reader meets the story. Proved by deleting the /uruguay/ section — the check reports the finding — and restoring it — the check goes quiet.

This archive states, in data and on its pages, that the 1812 padrón 'records four enslaved people in Sancho's household and gives none of them a name'. The image shows four named people immediately above him — Juana 41, Juan 'negro esclavo', Gabriel, María — under margin number 59, with Sancho's bracket beginning on the next line and the family above them a Sargento Veterano de Caballería and his wife Sabina Balmaceda. So the four may be that sergeant's, or Sancho's, and dist.jpg at 3793×2909 does not settle it. It matters because the baptism of 26 September 1812 names JUANA as an enslaved woman of Sancho Lerena and María Fernández — a Juana of 41 at the edge of his household in the same year is a convergence, not a coincidence to be shrugged at. Needs the bracket at tile resolution, or the household numbering read across the whole opening. If it resolves to Sancho, the archive's 'gives none of them a name' is wrong and two more people become nameable. RESOLVED THE SAME DAY, AND THE ANSWER IS THE OTHER DIRECTION. The four above Sancho are house 59 and not his. His four are BELOW his six children: ANTONIO — written out as negro esclavo — JOSÉ, JUANA and TERESA, listed after the youngest child Dominga and before RODRIGO FERNÁNDEZ, the father-in-law, whose presence at the end is what proves the household boundary. So the archive's 'gives none of them a name' is wrong and is logged as a serious correction. Antonio, José and Teresa are now in the register; Juana is corroborated inside the household by a second document of the same year. The Juana aged 41 in house 59 is a different woman.

Found in Uruguay, Census Records, 1726–1858, a collection this archive had never opened. María Llerena, femenino, 60 years, birthplace ESPAÑA — ark 1:1:6BT5-QNYT, image apid TH-1942-36766-24983-3. Sancho was andaluz, 45 in 1812, so a María born in Spain about 1752 stands in his own generation or the one above. She is on a different film section from Sancho, so probably a different census year. Not claimed as kin — she is not attached to anything. To place her, find her line on the page: it is a nominal list with name, age, nationality, occupation, and it is rotated 90° — transform: rotate(-90deg) with mild contrast(1.5); heavy contrast destroys the pencil. A household number in the left margin will give her household. FOUND ON THE PAGE, 20 September 2026. She is the first name on it: 'Casa 2. María Llerena — 60 años — Española', heading her own household, with Manuel, 20, 'Africana' and Rosa, 12, 'Oriental' under her, both with a ditto where a surname would go. A Spanish-born Llerena woman keeping her own house in Montevideo. Added to the register as HIGH INTEREST and attached to Sancho by nothing. The census year is still not established, so her sixty cannot be turned into a birth year — that date, at the top of her own page, is now the one thing worth doing on her.

The third of three Llerenas in Uruguay, Census Records, 1726–1858 — ark 1:1:6BR7-8DPW — and the index gives him no age, no birthplace and no household. This archive holds two candidates: Luis, Sancho's son baptised 1811, and Don Luis Lerena (1811–1878), husband of Justiniana Lenguas and now the documented father of Gilberto. Which, or neither, is unknown. The page will say more than the index does. READ 20 September 2026, and the page is not a census. It is a padrón of the owners of carretillas, carretas y coches in a Montevideo district, and Luis Llerena is entry 10, with entry 11 marked ídem — how these lists repeat an owner for a second vehicle. The year is not secure: the heading ends de 18__ and reads as 1841, 1842 or 1846, so no year is claimed. He is still not identified — the page gives no age, no household and no parents, so neither candidate can be chosen. The general lesson is the bigger one: this collection is not all censuses, and the index does not say which entries are which. See /uruguay/.

The collection spans earlier than anything else worked here in Uruguay — the archive's Uruguayan work starts at the Canelones baptisms of 1791, and this reaches back to 1726. Sancho arrived from Andalusia and his parentage is the wall row 22 is stuck on. Census records of the eighteenth century are exactly where an immigrant first appears, with an age and a stated birthplace. Sweep it for Lerena, Llerena, Serena and Gerena across the whole span, not just 1812. SWEPT, all seven spellings, 20 September 2026 - see /uruguay/. SERENA 10 (two complete households: Don Luis's c.1858 and Pedro Antonio's c.1833), LLERENA 3 (Sancho, plus María 60 born España and Luis, both unplaced), LERENA 27 fuzzy and mostly noise — Sancho is not among them, because he is indexed Llerena — LORENA 6, and GERENA, LERINA and LARENA all zero, tested against controls of 10 and 3 on the identical form. It did not reach the generation above Sancho, which was the row's hope; what it reached was two generations of his own descendants and one new candidate for a daughter nobody has ever followed.

The c.1858 census gives Don Luis Lerena's household with ages — see /uruguay/. Three things need settling. (1) María Serena, 12, implying a birth about 1846, is nobody this archive holds, and she would sit exactly between Luis Eugenio (1845) and Alberto Octavio (1850). (2) Justiniana Serena, 10 implies 1848, but the archive's Justiniana Mauricia Maxima is b. 30 July 1843, which gives 15 — either the age is wrong, there are two Justianas, or the archive's date is. Recorded as a contradiction, not forced. (3) Josefa Serena, 7 implies 1851 and may be the 'Josefa in Gilberto's 1895 household (b. c.1852)' already here — one year apart, not merged. Also absent from the page: Alberto Octavio, b. 1850, who should be about 8 and is not there. WORKED 20 September 2026 and it cannot be closed by searching. Three children match: Luis E. (1845), Gualberto/Gilberto, Erna/Ema (1857). Three in the census are not in this register — María c.1847, Justiniana c.1849, Josefa c.1852. Two in the register are not in the house — Justiniana Mauricia Maxima, who would be 15, and Alberto Octavio, who would be 8, with no eight-year-old on the page at all. The Justiniana conflict has two readings: a five-year age error, or two daughters of the same name with the elder dead — which would also explain Alberto Octavio. Neither is asserted. No new people were added, deliberately: if the ages are what is wrong, adding them creates duplicates of people already here, and this archive has spent this week unpicking duplicates rather than making them. Blocked on reading the San Francisco de Asís baptism register straight through — which is where the archive's existing exact dates came from, so the register is already known to it.

New to this archive, added to the register. Entered at 34 in household 97 of the c.1858 census, a year older than Justiniana Lenguas herself and living among her children. A brother of hers is the obvious reading and is not asserted — the census gives names, ages and nationality and no relationships at all. The Lenguas are already the best-documented side of this marriage: General Don Pedro Lenguas, Don Antonio, D. Juan, D. Felipe and Doña Sixta are all in the register. Telesforo is not, and should be findable among them. TESTED 20 September 2026 and he is a NULL. Surname LENGUAS exact, Uruguay: 561 records — the surname is well represented and the form works — and Telesforo is in none of them. Telesforo Lenguas across all countries returns 43, every one Colombian, Venezuelan or Mexican. He exists for this archive in one census line. A brother of Justiniana remains the obvious reading and remains unasserted. Blocked on the same parish register as row 167 — not on another search. Noted in passing from the control: Luis Pedro Lenguas and Antonia Veiga of Montevideo had a daughter María Ema Lenguas, b. 1896, and this archive's María Ema Lerena Lenguas was born 1857 — the same two names, forty years apart, in a Lenguas house. Claimed as nothing.

Dominga Lorena, 28, born Montevideo, ark 1:1:6BV9-FR8D, on a page headed AÑO DE 1836 — DEPARTAMENTO DE MONTEVIDEO (image apid TH-1951-36766-25725-94). Sancho's youngest child was Dominga, aged 7 in the 1812 padrón, so born about 1805 — which gives 31 in 1836, not 28. Three years out on a census age, which is what a householder said rather than what a register holds. A candidate, not a finding, and not attached to Sancho. What makes her worth the chase: this archive has never followed Dominga past the 1812 padrón at all — she is one of the six children with no later record of any kind. Her household on the page will say whether she is married, and to whom. FOUND ON THE PAGE, 20 September 2026 - household 114. She is the wife of D. Antonio R. Carballo, hacendado, both entered casado, with two unmarried daughters, Lucinda and Dominga Carballo — a daughter named for her mother. And the page writes SERENA where the index says LORENA, the third such discrepancy found in one day. Still not attached to Sancho: 28 in 1836 implies 1808 against his Dominga's c.1805. What would settle it is the couple's marriage act, which would name her parents — a far smaller search than the one that found her. Added to the register as HIGH INTEREST. CORRECTED 20 September 2026. She is Dominga Josefa Llerena, b. 13 May 1804, proved on 14 September from the Catedral act of 1827. My 'candidate, not a finding' framing was wrong and I added a duplicate person before checking — removed, evidence folded in, logged as a serious correction. The census age of 28 against a baptism of 1804 is four years out, which is a census age.

The one search that would settle whether Sancho's youngest daughter has been found. They are married by the 1836 Montevideo census, with a daughter old enough to be entered soltera, so the marriage is likely Montevideo, c.1820–1832. A Uruguayan marriage act names both parties' parents outright — as the 1877 Gilberto act did this week. Search CARBALLO with spouse surname SERENA, LERENA, LLERENA and LORENA in Uruguay, Catholic Church Records and Uruguay, Marriages 1840–1900, and remember the lesson of this whole sweep: the spelling on the page is not the spelling in the index. RUN 20 September 2026, and the country is probably wrong. SERENA exact with spouse CARBALLO, Uruguay returns exactly one hit — Dominga as mother of the bride in her daughter's act — against a control of 769 on the loosened form. Their own marriage is not in Uruguay's indexed records. And the reason may be plain: their daughter JOSEFA CARBALLO was born and lived in RIO DE JANEIRO, marrying Miguel Carlos Lemus of Rio at San Francisco de Asís on 24 July 1862 (ark 1:1:X9PN-PYNH). The family was in Brazil. Blocked here until either Brazilian registers or pre-1826 Montevideo registers are reachable — the search was right and the country was wrong, which is worth more than sweeping the same corpus again.

RE-SCOPED 20 September 2026 after a correction against myself. The identification was not new: this archive established on 14 September that Dominga Josefa Llerena married Capitán Antonio Ruiz [Rodríguez] de Carvallo, from the Catedral act of 12 May 1827. What is new is Brazil. Their daughter Josefa Carballo was born and lived in Rio de Janeiro and married Miguel Carlos Lemus of Rio at San Francisco de Asís on 24 July 1862. No part of this archive has ever looked at Brazilian records. A Uruguayan family with a daughter born in Rio is the first reason it has had to.

'Casa 2. María Llerena — 60 años — Española' sits on an undated page of Uruguay, Census Records, 1726–1858 (image ark 3:1:3QSQ-G9ZR-NYGP). The collection spans 132 years and the index gives no year for this sheet. At an 1830s census she was born in the 1770s — Sancho's own decade; at an 1850s one, the 1790s. The whole weight of her rests on which. The heading sits sideways at the top of the sheet above her entry; the page needs rotate(-90deg) and a mild contrast(1.5) — heavy contrast blows the pencil out to white. Neighbouring images on the same film will carry the same heading if hers is torn. CHASED AS FAR AS THE FILM GOES, 20 September 2026, and the year is not on it. Her page is film 007499043, image 02771. The filming target board at image 02725 identifies the volume: Archivo General de la Nación, Montevideo — 'PADRONES DE MONTEVIDEO Y EXTRAMUROS, LIBRO 256', volumen 15, filmed by the Sociedad Genealógica de Utah on 5 April 2006. The board at 02724 closes the previous item, Libro 107. The board gives the FILMING date, not the record date, and her own sheet is headed by a street, not a year. So the question changes rather than fails: it is no longer when is this census but what does the AGN's own inventory say Libro 256 covers — which is a question with an address on it, and the same letter that goes to Santa Fe could ask Montevideo. Her age of sixty means nothing until it is answered.

The DGS audit of 17 September tested 58 film numbers and found them all sound. This one was not among them, because nothing here had ever opened the Uruguayan census collection. It carries AGN Padrones de Montevideo y Extramuros, Libros 107 and 256 at least, and holds María Llerena (image 02771). Add it to film-register.tsv with its item boundaries — 02724 closes Libro 107, 02725 opens Libro 256 — so the next person can navigate it by item instead of by probe. ADDED 20 September 2026, with more than the row asked for. The film runs 1 to 3208 (last image found by binary probe on the public thumbnail) and its leader gives the old microfilm number, 1263678. It is rollo 1 of project URUG NA01S, filmed at the Archivo General de la Nación in March–April 2006, and it carries an entire reel of 'Padrones de Montevideo y Extramuros' in numbered volúmenes: volumen 1 = Libros 2 y 334 (board at image 00004), Libro 107 closing at 02724, volumen 15 = Libro 256 opening at 02725 and holding María Llerena at 02771. The remaining boundaries are unmapped and the row below says how to find them.

3,208 images of AGN Padrones de Montevideo y Extramuros, and this archive has looked at five of them. Four volúmenes are identified (1 = Libros 2 y 334; one ending in Libro 107; 15 = Libro 256); the rest are unmapped. Find the boundaries with thumbnail contact sheets — the target boards show as dark frames among faint pages, and two sheets found 02724/02725 after a page-by-page hunt found nothing. Why it matters: the padrones are enumerations of the living with ages and birthplaces, they run earlier than any Uruguayan work here, and this one reel has already produced Sancho's household with four enslaved people named, two Lerena households with ages, and a Spanish-born Llerena woman heading her own house. Map it and the sweep can be run against structure instead of against luck. MAPPED 20 September 2026, and not with contact sheets. Each thumbnail was drawn to a canvas and its mean brightness measured: target boards read 16–30, ordinary pages 140–156, nothing in between. Scanned at stride 2 — sufficient, because a boundary is a fin/start pair of consecutive images so one is always odd. Eighteen boundaries found at images 4, 91, 125–127, 315, 563–565, 787, 959, 1073, 1215, 1315, 1475–1477, 1573–1575, 1847, 1961–1963, 2027, 2155, 2425, 2724–2725 — 1,604 thumbnails, three calls, no session, no screenshots. Three of the eighteen items are named; the rest are one image load each. See /uruguay/.

The boundaries are mapped; the volumes are not named. Three of eighteen are known — volumen 1 = Libros 2 y 334 (image 00004), Libro 107 closing at 02724, volumen 15 = Libro 256 at 02725. Each remaining board carries its own volumen and Libro numbers and costs one dist.jpg load: images 91, 125, 315, 563, 787, 959, 1073, 1215, 1315, 1475, 1573, 1847, 1961, 2027, 2155, 2425. Named, the reel becomes navigable by volume — and the Libro numbers are what the AGN's own inventory indexes, so they are also what dates each one. RE-SCOPED 20 September 2026, because the first attempt was wrong. The dark frames are not all target boards. Image 00127, predicted to be a title board, is a black book cover; image 00091 is a FIN DEL ÍTEM card carrying no title. Only 4, 2724 and 2725 are confirmed by eye. So each of the remaining fifteen costs one dist.jpg look to say what it is — board, cover, or blank — before any of them can be named. Loading four full-size boards into one page does not work: they are 1 MB each and none rendered. DONE 20 September 2026, and cheaply. All 41 dark frames were rendered as upscaled thumbnails — thumb_p200 at 194px with heavy contrast — which makes the big handwritten volumen numeral legible, and that was all that was needed. Two screenshots instead of fifteen full-image loads. The reel holds fifteen volumes, title boards at images 4, 126, 315, 563, 788, 959, 1215, 1315, 1475, 1574, 1847, 1962, 2155, 2425, 2725, each preceded by the fin del ítem card of the volume before. Twelve other dark frames are covers, spines or blanks. And the rule I withdrew was right: the pair is FIN(n) then title(n+1) — I had tested it on image 127, which is a third frame sitting after the 125/126 pair, not part of it. Libros known for volumes 1, 12, 14 and 15.

The reel is mapped and four volumes have their Libro numbers — 1 = Libros 2 y 334, 12 = Libros 465 y 255, 14 = Libro 107, 15 = Libro 256. The other eleven title boards (images 315, 563, 788, 959, 1215, 1315, 1475, 1574, 1847, 2155, 2425) carry theirs too, but the Libro line needs the full dist.jpg — the upscaled thumbnail gives the volumen numeral and not much else. Why bother: the Libro number is what the AGN's own inventory indexes, so it is what dates each padrón — and an undated padrón is why María Llerena's age of sixty currently means nothing. FOURTEEN OF FIFTEEN READ, 20 September 2026. Two full boards per page render; four do not — so they were done a pair at a time, and where one failed to load it was carried into the next pair. 1 = Libros 2 y 334; 3 = 444, 247, 249; 4 = 250, 251, 252; 5 = 252; 6 = 253; 7 = 261; 8 = 262; 9 = 464; 10 = 147; 11 = 148; 12 = 465 y 255; 13 = 149; 14 = 107; 15 = 256. Volumen 14's number came off its FIN board at 2724, its own title board having refused to render twice. Only volumen 2 is outstanding — board at image 00126. Note that Libro 252 appears on both the volumen 4 and volumen 5 boards, recorded as read rather than tidied: a volume may end and begin mid-Libro. COMPLETED: volumen 2 is LIBRO 246 , read from its board at image 00126. All fifteen volumes now carry their AGN shelfmark. And the numbers fall in groups: 246, 247, 249, 250, 251, 252, 253, 255, 256, 261, 262 run nearly consecutively across volumes 2–8, 12 and 15; 147, 148, 149 sit together in volumes 10, 11 and 13; 334, 444, 464, 465 are scattered; 2 and 107 stand alone. So one reel carries at least four separate runs of the Padrones series, filmed out of order — which matters, because it means a neighbouring image is not a neighbouring year.

tools/check-duplicates.py was written after this archive added a duplicate of Dominga Josefa Llerena — one hour after refusing to add duplicates on principle. It flags two register rows sharing a forename and a canonical surname (all seven spellings fold to one) plus something only the same person would share: an ark, three uncommon terms, or birth years within six. Verified by re-adding the Dominga duplicate: it names the pair and exits 1. 133 pairs already existed and are baselined in data/_person-pairs.txt as UNREVIEWED — being listed does not mean compared. Several look genuinely suspect: Bartolomé Llerena vs Bartolomé Llerena – Córdoba origin, both b.1824; Felisa (arr. 1911) vs Felisa (arr. 1923), b.1893/1891; Emilia LERENA vs Emilia LERENA Vega; Carolina LERENA vs Carolina LERENA (de Martinez). Work down the list and put the reason after a # on each line. DONE 20 September 2026, the day after the gate was written. All 123 surviving pairs carry a reason on their line. Seven merges came out of it — three Carlos rows were one man, two Avelino Carlos, two Rosa María Eppens, and four of Gilberto's children were held twice because the enumerator wrote initials and the priest wrote names. Every one of those identifications was already stated in the archive's own notes and never acted on. One SPLIT too: merging Carlos exposed relations.tsv giving him a wife the register had explicitly ruled out. And two merges were decided by reading the census image — the ages on lines 10-13 are 8, 6, 5, 2, not the '6, 6, 5, 2' this archive had transcribed. Register 292 → 286, then 293 with the Villademoros branch.

Nine children where this archive held six, and four of them were sitting in its own files undated. A parent-pair search — Avelino as father, Ambrosia Traibel as mother — returns ten records. Three of them are dated 1896 and 1919, and Avelino died in 1890. That impossibility was the finding. The film is a register of certified copies issued in 1896, and the indexer takes the year off the certificate and the day and month off the act inside it. Read off film 008166435, image 02850: Pedro Alberto Ysabelino Lerena, born 16 March 1852, baptised at the Aguada on 21 November 1852 — not 1896 — libro 1.º de Bautismos, folio 135, godparents Pedro Antonio Lerena and Vicenta Vega, his own uncle and aunt, and a man born in 1793 who could not stand at a font in 1896. He is the Alberto who died at Montevideo on 1 March 1921. The same opening carries Josefa's own baptism, which gives her a day at last: born 10 February 1851, and christened Josefina. From the indexes: Juan Arturo born 16 February 1861, Mercedes Micaela born 8 May 1863 — with Sancho and María Fernández named as her grandparents, a third independent attestation of that parentage — and María Adelina, who married Carlos Alberto Fein and died in 1946. The tell that catches this trap is cheap: two children of one couple indexed with the same birth date. See /method/ and data/1852-aguada-union-certificates.tsv.

A woman this archive had written off is ours, and she has eight brothers and sisters. On 14 September Zelmira Carlota Lerena Larriera, born in the Departamento de San José on 4 November 1880, was tested and closed: "she is NOT ours". The row named the one thing that would overturn it — who her father Alberto Lerena was — guessed Alberto Octavio Lerena Lenguas, and refused to draw the line on a forename and an age. It is the other grandfather. Alberto's Montevideo death act No. 246 of 2 March 1921 names his wife María Inés Larriera and his parents Avelino Lerena and Ambrosia Treibel — so her father is Pedro Alberto Ysabelino Lerena Traibel of the Aguada, and Zelmira is a great-granddaughter of Sancho. Asking for Alberto and María Inés as a pair then returned a whole household born in San José between 1879 and 1898. And the image was opened rather than left as an errand, which is where the real find was: the act counts nueve hijos and names them, so there is a ninth child — Eleiza or Eloiza, read at three times magnification and still not securely read — whom no index has matched. Eight index entries look like eight confirmations and are one indexing project; one death act is a different mouth. See /uruguay/ and data/alberto-lerena-larriera.tsv.

Fernando Moratorio with Josefa Lerena returns seven, adding Eduardo, Josefa (born 21 November 1884), María and Fernando Moratorio Lerena to the two girls baptised at the Unión in 1880. Two other pairs returned zero, and one of those zeros was a lie. Avelino Lerena with Celedonia Villademoros returned a clean 0 — rows rendered, page settled, read twice. Deleting the mother's surname returned seven, two of them christenings this archive has held since 20 September. That is how it was caught, and no control would have caught it: the control for this query shape had returned sixteen records in the same session, which is exactly what made the zero look trustworthy. Recorded as the fourteenth false zero. What it hid: Pedro Segundo Lerena Villademoros, dead at Montevideo 23 August 1936 aged 66, wife Paulina Prando Algar; an Avelino Lerena dead 5 January 1914 aged 45, wife Carmen Lorente, whose mother is written Villademoro without the s. It also weakens a null this archive published about the Villademoros marriage acts: the control used, Lerena × Juanico, tested the shape but not the surname, and Juanico has one spelling where Villademoros has four. The null is not withdrawn; it is now stated as the weaker claim it always was.

Three surnames turned out to be a machine's, so the question became how many others were. tools/transcript-only.py counts rows citing full-text transcript read with no image named: sixty-one, more than half the rows carrying a source at all. It ranks the unopened plates by how many people depend on each, because one image settles seven rows at once. The top plate was the 1831 Cordón baptism where the President of Uruguay stood godfather to Sancho's grandson — the most consequential proper noun here, and a machine had read it. Opened: film 007713388 image 00726. "padrinos Dn. FRUTUSO RIBERA, PRESIDENTE DE LA REPÚBLICA y su esposa Da. Bernardina Fragoso". Frutuso Ribera, with a B — the transcript had printed the history books' Fructuoso Rivera. The finding holds and is stronger for the check: the act states his office and names his wife, either of which identifies him alone. The same page spells the grandfather LLERENA with the double L. Both of the day's transcript failures are one failure pointing opposite ways — a machine invents a plausible name where it cannot read one, and writes the famous name over the page where it can. Seven rows moved off transcript onto plate; forty-nine remain.

Her surname was wrong, and she has a death. The 1892 marriage act was held from a machine transcript and the plate was never pulled. Film 007713383 image 02601, folio 340: "hija legítima de Don GERÓNIMO GAVAZZO y Doña CAROLINA SUÁREZ" — not Gavarro, not Juárez. Three register rows, a dispensation table, a working log and the last-attested-alive file carried the machine's reading. Corroborated by a second image forty-eight years later: her own Montevideo death registration of 20 June 1940, certificate 632, at R. Masini 3273, "de estado VIUDA", of a fractured neck of femur. So Alfredo Justiniano died between February 1892 and June 1940 — the first thing this archive has ever known about the end of his life. And the index of that death act is a catalogue of its own failures: it attaches three children to the couple, and there are none on the act — one is the registrar, whose name follows ante mí, and two bled across the column rule from the neighbouring entry, which concerns a family with English names. It gives her birth as 1934 for a woman who died in 1940, her nationality as mexicana, and her birthplace as Minteridco. Every field is wrong except the maternal grandparents — and that pair is the only reason she was found.

Sixty-one rows cited full-text transcript read and named no image — more than half the rows carrying a source. Twenty plates later the count is zero. tools/transcript-only.py ranked the unopened plates by how many people stood on each, because one image settles seven rows at once. What the sweep found: four claims about this family's spelling were the machine's, not the scribe's — Abelino Serena is LERENA, Sancho Lorena is LERENA, the LORENA of 1889 is a margin and not a sentence, and Fructuoso Rivera is the history books' spelling written over the page's own Frutuso Ribera. Five surnames were wrong: Gavarro→Gavazzo, Juárez→Suárez, Florencia→Florinda, Seguera→Segueira, and "Del Suando"→Zalsamendi. One baptism was in the wrong parish (the Catedral Basílica, not the Unión) and one citation pointed at the wrong film. A birthplace changed the family's shape — María Fernández is natural de Asturias, so both halves of Sancho's household were Spanish immigrants. And about twenty-five people standing on the acts had never been entered, among them Sancho's parents-in-law, two grandparents drawn at the apex of a priest's kinship diagram, a second child of Justa Pastora, and Adolfo Lerena. The sweep also confirmed more than it corrected: Lavalleja, Maggesi, Larrañaga and Rivera's office are all really on the plates; only Rivera's spelling was the machine's. See /method/.

Done, and the diagnosis needed correcting before the nulls did. The record search returned 0 for everything — Historical Record Search Results (0) with No Results Found — stable across two reads, still 0 after ninety seconds, including González in Uruguay. Full-text, asked the same question, redirected to the sign-in page, so this archive concluded the account was signed out and said David would have to sign in. That was wrong. A sibling archive ran both searches in the same minutes and got 3,652 and 3,967 with the account name in the nav bar: each session's in-app browser is its own browser with its own cookies, and one context had lost its auth and recovered on its own. In a fresh tab the González control returned 479,662 and Avelino × Ambrosia returned 10 again. All three held-over nulls were then re-run against that live control: Luis × Carolina = 0, Carlos Ambrosio × Luisa = 0, Juan Arturo × Isidora = 0. They stand, and row 70's null is now properly established rather than merely asserted. Rule kept: a control must be recent and arithmetically impossible to be zero. Rule added: before asking a person to fix something, open a second tab.

A CONTROLLED NULL, AND THE CONTROL WAS THE RIGHT KIND. Searched 27 September 2026 under LERENA and under SERENA - the spelling that had hidden the 1877 act - for a marriage to a Villademoros in Uruguay: nothing. The control passed in the same minutes: the Montevideo parish marriage of Luis Lerena and Julia Juanico, 19 December 1864 at San Francisco de Asís, returns with the groom's parents attached. So the instrument reaches 1860s Montevideo marriages and carries parents on them. The act is therefore not in the indexed corpus and must be walked by image. Their eldest known child was born 29 October 1872, so the marriage falls about 1868-1872, most likely at the Cordón. WHY IT IS STILL WORTH IT NOW THAT HIS PARENTS ARE KNOWN: it names them a second time, independently - and it gives his AGE, which is the one thing that separates the 1843 reading of the La Mauricia act from the 1845, and so the one thing that would finish the merge. SEARCHED AGAIN FROM THE WIFE'S SIDE, 11 OCTOBER 2026 — AND IT STILL IS NOT THERE. The same hour this archive found the 1897 Zalduondo marriage by searching the GROOM's rare surname bare instead of the bride's common one, that technique was turned on this row: VILLADEMOROS, bare, no given name, no date — 649 hits in Uruguay; then Elisa Villademoros — 33 hits, and every one is accounted for. Four carry the role Other, which is where this index hides marriage parties, and not one of them is a marriage: two are Angélica María's own baptism (the 1872 original at Unión and the 1919 certified copy, the forty-seven-year gap this archive already holds), one is an 1891 Montevideo christening where Elisa stands godmother, and the rest belong to other Villademoros households — Adolfo's of Rivera, Rafael and Concepción Estavillo's. SO THE NULL NOW HOLDS FROM BOTH SIDES OF THE COUPLE AND UNDER FOUR SPELLINGS (Lerena, Serena, Villademoros, Villademeros). The instrument is working — these searches returned 649 and 33, not zero. THE ACT IS NOT IN THE INDEXED CORPUS AND THE IMAGE WALK IS THE ONLY ROUTE LEFT, which is what this row already said; it now says it on twice the evidence. A SMALL THING FOUND ON THE WAY: that 1891 christening of a boy Pedro at Montevideo carries, among its mangled names, an 'Alejandro M. Pero' beside Elisa — almost certainly Alejandro M. Lerena standing godfather with his wife. Computer-indexed and not opened, so it is a lead and not a fact. FOUND 11 OCTOBER 2026 — 28 DECEMBER 1871, AT THE CORDÓN, AND NOT BY WALKING AN IMAGE. This row twice concluded the act 'is not in the indexed corpus and must be walked by image'. That was true and it was not the whole truth: the RECORD INDEX and the FULL-TEXT CORPUS are two different instruments, and the act is in the second. A single full-text search on the phrase “Elisa Villademoros” returned fifteen hits, one of them Nuestra Señora del Carmen La Mayor, Cordón — Religious Marriage Records 1866-1872, which is this row's own parish and window. THE ROW HAD NAMED THE RIGHT BOOK AND NEVER ASKED THE RIGHT INDEX. WHAT THE ACT GIVES: his full name — ALEJANDRO MÁXIMO, so the 'M.' is Máximo; that he was soltero and natural de Montevideo; his parents AVELINO and AMBROSIA TRAIBEL named a second time and independently; his mother standing as a witness, so alive in December 1871; and Elisa as hija de los finados D. PEDRO y Doña ANGÉLICA VILLADEMOROS. THAT LAST CLAUSE PAYS FOR THE WHOLE SEARCH. Celedonia Villademoros is the daughter of Pedro Villademoros and Angélica Méndez — so Elisa and Celedonia are SISTERS, which this archive has said 'probably' since September, and two Lerena brothers married two Villademoros sisters, making the Lerena Villademoros cousins double first cousins. It also rules the dead mother out as the 1873 godmother, forcing that woman to be a third sister. WHAT IT DOES NOT GIVE, AND THIS ROW WANTED IT MOST: HIS AGE. The act states status and birthplace and no number. The 1843-against-1845 question on the La Mauricia act is untouched and stays open.

Entered 27 September 2026 from the Iglesia Matriz act of 1872, and found the same day in the Uruguayan civil index under her married name, "Angélica Lerma de Zalduando", daughter of Celedonia Villademoros and a father the machine renders "Arabia Larena Lerena" - which is AVELINO LERENA. INDEX ONLY; no image opened, and "Lerma" is the index's word. The marriage itself has not been found. She is a granddaughter of Avelino and Ambrosia Traibel and a niece of Alejandro M. Lerena, who stood godfather at her baptism. ADVANCED 11 OCTOBER 2026, AND IT TURNED INTO SOMETHING ELSE. The Zalduando is 'EDUARDO ZALDUANDO POU', dead at Montevideo on 22 August 1913 aged fifty-two (so born about 1861), certificate 553, son of JUAN and DOLORES POU — which widows her at forty-one — with two children, EDUARDO AVELINO (b. c.1901 at La Paz in Canelones, m. María Ester Costa there 20 Dec 1927) and MARÍA MARGARITA LUISA. BUT EVERY ONE OF THOSE PEOPLE WAS ALREADY HELD, UNDER ZALSAMENDI, off the Pocitos plate of 15 August 1903 — same husband Eduardo, same father Juan, same mother Dolores Pou, same Lerena wife of Avelino and Celedonia, same daughter María Margarita. The duplicates gate caught it before it was committed. See data/zalsamendi-zalduando-one-household.tsv and /uruguay/#s-zalduondo. IT ALSO CLOSED A DIFFERENT OPEN DIFFICULTY FOR FREE: the 'AVELINO y CELIDONIA' of the 1903 act is AVELINO EDUARDO LERENA TRAIBEL × CELEDONIA VILLADEMOROS, so the Pocitos household joins to the Lerena Villademoros. STILL NOT CLOSED, AND TWO REASONS WHY. Her own marriage act is still not found; and the plate says the wife is AMELIA where the index says ANGÉLICA, which is now open question 10. NOTHING WAS MERGED. AND THEN THE PLATES OVERTURNED THE WRITE-UP, the same day. See row 198. The husband is EDUARDO ZALDUONDO, already in this register under a misread surname; Amelia Lerena is this woman and has been merged into her. THE ROW STAYS OPEN FOR ONE REASON ONLY: her marriage act has still not been found, which is now open question 10. CLOSED, 11 OCTOBER 2026 — THE MARRIAGE IS FOUND. Montevideo, 4th section, civil act No. 83 of 20 November 1897, folio 42, expediente 94, before the Juez de Paz Gallo Pareja; PLATE READ. It was found by searching the GROOM's surname instead of the bride's — four hours after this row recorded the act as missing. Angélica signed it herself. The act also gives Eduardo's exact birth (15 August 1860), Avelino Eduardo Lerena Traibel's death in Canelones in 1895, the Zalduondo parents' nationalities (Juan oriental, Dolores brasileña), and — by leaving 'Legitiman' blank — places all three children after November 1897.

From the 33 findmypast transcripts opened 27 September 2026. Four households this archive does not hold: Manuel x Juana FERREIROZ (Montevideo 1891), Manuel x Dolores ESPINDOLA (Montevideo 1884) - two Manuels or one with two wives, not established - Pedro A. x Manuela LEMOS (Montevideo 1883), and Fermín x Paulina BARRAGÁN (Trinidad 1877, with a second child María Regina in 1882 found on FamilySearch the same day). And three LERENA WOMEN appear as MOTHERS at San José - OLEGARIA (1862), ESTEFANÍA (1872) and FELICIA (1875). A surname search on the father's line cannot see them, which is why this archive has never counted that side. Each needs testing against the San José cluster. Tier: index of an index; no image opened. ANSWERED 11 OCTOBER 2026, AND TWO OF THE THREE WOMEN TURNED OUT TO HAVE HUSBANDS. FELICIA married at the Catedral Basílica San José on 19 December 1874, and her act names her parents as MAURICIO LERENA and CASIMIRA MONTAÑO — the couple married at Colla in 1837. So she is a second child of the Rosario-Colla house, alongside Manuel, and a woman married in December 1874 mothering a child at San José in 1875 is exactly what this row recorded without being able to explain. ESTEFANÍA married SERVANDO GONZÁLEZ; her children are indexed GONZÁLEZ LERENA — Servando Cecilio, Leopoldo, Francisco, Cesárea, María Bernarda and Estefanía Fausta Leopoldina. That is the compound-surname principle working exactly as this archive states it: a father-line search cannot see her. Whether this joins the existing González Lerena thread is NOT established. OLEGARIA REMAINS WHAT THIS ROW CALLED HER — one record in the whole corpus, mother of Juana de las Nieves Lerena, the child carrying her surname because no father was named. AND THE FOUR HOUSEHOLDS ARE ADVANCED TOO: Pedro A. × Manuela Lemos gains a second child, Felicia Clementina, b. c.1868, married 1886, fifteen years before Manuel Antonio; Manuel × Juana Ferreiroz is confirmed by their daughter Felicia, b. 6 May 1891, in two separate collections; Fermín × Paulina Barragán stands as recorded. A SUGGESTION MADE EARLIER THE SAME DAY WAS WITHDRAWN: that Juana Ferreiroz might be the Juana Galarza who married a Manuel Lerena in 1890. Two independent records name her Ferreiroz. All INDEX ONLY; no act opened.

Started 27 September 2026 and deliberately not finished. A single FamilySearch search on birthplace = Trinidad, Flores returned the CHILDREN OF THE CHILDREN - a generation findmypast's 33 transcripts do not reach. See data/trinidad-flores-sweep-2026-09-27.tsv. It found that CUSTODIO LERENA and EMA CARDOZO - the double first cousins of the 1911 dispensation - had a daughter in 1915, so the dispensation was granted. WHAT IS LEFT: the remaining result pages; four households new to this archive (Pedro Segundo x Petrona Fernández, Florencio x Juana Recuero, Gualberto x Rosaura Álvarez, Doroteo x Isidora Cardozo); a FOURTH Lerena woman recorded as a mother with no father named (Cayetana Faustina); and a fifth Alejandro - Alejandro Estéban, b. 1901, son of Gualberto - for data/the-three-alejandros.tsv to watch. Every row is INDEX ONLY and no image has been opened. FINISHED 11 OCTOBER 2026, AND THE ANSWER WAS IN THE MARRIAGES, NOT THE BAPTISMS. Pages two to four of the same search carry the marriage acts the September pass never reached, and they explain the 1911 dispensation outright. Three weddings in thirteen months: Bernabé Lerena × Florinda Cardozo on 22 March 1888; then, on one day, 25 April 1889 in one church, his sisters María × Eliseo Cardoso and Casimira × Juan Basilio Cardoso. A Lerena brother and his two sisters married a Cardozo sister and her two brothers. That is how double first cousins are made, and in 1911 Custodio Lerena needed dispensing to marry Ema Cardozo. The 1911 marriage itself is now found — 18 September 1911 — so the grant is on the act rather than inferred from a birth in 1915. Also dated: Socorro Lerena × Serapio Fernández, 15 July 1885. Also new: a second sister (Casimira), Alejandro F. Lerena × Natalia Lerena — a fourth Lerena-married-a-Lerena couple, and the first record to give row 137's 'Alejandro F.' a wife and a child — their daughter Aurora Lerena × Manuel Vaquero with five children, and Manuel Lerena, son of Mauricio and Casimira Montaño, married at Trinidad in April 1882, which is a second and independent thread from the Rosario-Colla house to Flores. Everything remains INDEX ONLY; not one act was opened. What is still not shown: which Cardozo fathered Ema — if Eliseo or Juan Basilio, the doubling is accounted for exactly.

data/tab-pull-list-2026-09-22.tsv and its JSON hold NINE ranked Transvaal Archives references with, for each, what it would answer — TPD 1367/1980 and TPD 1248/1979 for who Anton Armando and Ricardo Paul are, and the rest of MHG 1518/73 and MHG 6746/71 where only the death notice is held. That is a finished piece of work sitting in a data file with no row here, which is part of why this archive's own gate reports 20 outstanding items not yet reconciled.

TWO THINGS REPORTED BY THE BOOYZEN SESSION ON 27 SEPTEMBER, NEITHER VERIFIED HERE, both of which bear on ordering these. NARSSA's tariffs changed on 1 September 2026, so any cost worked out before that is stale. And the Pretoria reading room is reported closed, which would make a pull list an order-by-post or nothing at all. Recorded as reported rather than established, because this archive has not checked either and the difference matters to whoever acts on it.

Blocked on David because ordering costs money and the tariff is the thing that changed.

TWO FAULTS, BOTH INVISIBLE IN A GREEN BUILD. The chart adapter passed dates: p.dates and people.json has no dates field and never had one, so every self node carried a bare name while five sibling archives showed a lifespan. Undefined does not render as a gap a reader notices — it renders as nothing and looks deliberate. And 195 of 358 person pages drew NO CHART, the worst coverage of the seven archives, because the kit hides the figure when every tier is empty.

Both fixed 27 September. The chart draws for everybody via the kit's always, and the caption says which case applies from this archive's own status: 181 unplaced, 8 excluded, 6 connected and kinless. For the excluded eight an empty chart is the FINDING — Lola LERENA is Italian-born and the name is coincidental, Catherine BUYS is a false positive, Antonio LARENA of Cayasta is retracted.

AND THE FIRST FIX PUBLISHED A PLACEHOLDER AS A DATE. «-» is how this register writes «not known», on 166 of 357 rows, and it is truthy — so 163 charts read «b. -». The page's own field table had always filtered it and the adapter did not. Worse, the verification said «357 of 357 now show a lifespan» because it counted the PRESENCE of a dates span rather than whether it held a digit. Measured on the digits: 194 charts state a real date, 163 correctly state none, 0 placeholders.

Deaths were the other half. This register has no died column — deaths live in three separate files — so the kin gate had been green since it was written while examining nobody. 20 explicit deaths are now carried from the hand-reconciled last-attested-alive.tsv, tier DEATH only, never the three DEATH BY ceilings which say in their own text that they are a ceiling and not a date. See row 82 for what that does not fix: the 23 line-file people the register cannot hold under any spelling.

CORRECTED AND DONE, 7 October 2026. THE COUNT OF SEVENTEEN WAS WRONG AND IT WAS MY OWN ERROR. The test that produced it asked whether a register row contained every word of the string a regex had captured - and the regex over-captured, so "m. Capitan Don Antonio Ruiz de Carvallo and daughter of Sancho L" was searched for as if it were a name. Spouses that were in the register all along came back as missing. The real figures: four were already rows, three more resolved on reading, and EIGHT were genuinely absent. WIRED (7): Bernardina Fragoso to Fructuoso Rivera; Cipriana Sáenz Peña to Gilberto Lerena Salvanach the younger; Leonarda Correa to Nicolás Benteno; Dominga Carvallo to Capitán Antonio Ruiz de Carvallo; María Carlota Amadeo Lastra to the Buenos Aires Rafael Llerena, not the Córdoba one; Juana del Espíritu Santo to Don Domingo Rodríguez the elder, both naturales de la Colonia; and Carolina Oribe to the Luis Lerena who remarried in 1842 as viudo de Doña Carolina. ENTERED (8): Petrona López, Vicenta Mercedes Fuentes, Florencia Calixta Ramos, Antonio Medero, Rosario Moyano, Vicente Campellone, María Mercedes Rozzetti, Cristóbal Canale - each on the citation its citing row already carried, with no new research behind it and the row saying so. AND IT LEFT A SMALLER GAP BEHIND: five CHILDREN are named on those same rows and still have no entry - Félix Antonio and Aurora Eduarda Larena, Petrona Medero, Ana María Lerena Rozzetti and Susana Avelina Canale Lerena.

The tenth child. The tree says William Taylor and Mary Ann Hellen had ten children and names nine; Septima means seventh and Mary Septima is only the sixth daughter of those nine; and the birth index offers exactly one candidate in the gap between Lucy (1881) and Mary Septima (1884) — TAYLOR, Edith, Bideford, September quarter 1882, 5b/506.

THE ARGUMENT LOST A LEG ON 8 OCTOBER AND GAINED A BETTER ONE. Struck: "Mary Septima stayed with her sister EDIE in 1935", which was never independent — Edie is Isabella Temple Dalling, named as such in four other files and in the 1939 Register for that address. Added: the family group photograph shows eight females besides the infant, where the named family accounts for seven including the mother.

WHY A CENSUS CANNOT DO IT: Mary Septima is absent from the whole of FreeCEN's Devon in 1891, and the only Edith Taylor at Northam was born at Cardiff in 1886. The birth index gives no mother's maiden name until September 1911.

So it has to be a baptism register, which names the parents. Northam and Bideford, 1882–83. One line attaches her or detaches her. AND THE ALTERNATIVE MUST BE TESTED WITH IT: if any one of the eight women in the photograph is a grandmother, aunt, sister-in-law, visitor or servant, nine children satisfy the count and no Edith is needed — David's own reading, and the 1891 return for that house does record a visitor.

BOTH FREE ROUTES WERE TRIED ON 8 OCTOBER AND BOTH ARE CLOSED, with controls.

FreeREG — a coverage wall, not an absence. Its own contents pages give Northam, St Margaret, Parish Register: 12,120 records, 1653 to 1817 — it stops sixty-five years short. Bideford is listed as a place and holds no church, no register and no records at all. The search was run before that was known and returned 0; the contents page then explained the 0. Nobody should read that zero as evidence about the family.

FamilySearch — two false zeros in one sitting, both now in data/false-zeros.tsv. Its place parameter returned 0 for surname TAYLOR at Northam, Devon, 1860–1900 with no other constraint — impossible — while a looser place gave 6,941 results from Kent and Northamptonshire. The field was ignored when loose and fatal when tight. And a full-name search returned 0 for Horace Archibald, Mary Septima and Agnes Walker Taylor, all three of them on records here, because the index carries initials.

SO THE ROUTE IS NOW A PAID OR A POSTAL ONE, and it should be named rather than re-discovered: findmypast's Devon Baptisms, licensed from the Devon FHS, or Ancestry's Devon parish registers; failing those, the North Devon Record Office, which holds the Northam and Bideford originals. One line in a baptism register, which names the parents, attaches or detaches her.

ANSWERED THE SAME DAY, on findmypast's Devon Baptisms. The free routes were shut; the paid one was open. Northam, St Margaret, archive reference 1843A/PR/1/16 holds seven children of William Taylor and Mary Ann Helen/Ellen: Horace Archibald and Isabella Temple (1874), Margaret Helen Walker (1876), Irene (1879), Lucy Elizabeth (31 Aug 1881), Mary Septima (30 Oct 1884) and Agnes Walker (21 Aug 1889).

NO EDITH. Lucy and Mary Septima are three years and two months apart with nothing between them, and the candidate falls exactly in that gap. The only two Devon Ediths near the date are at Bideford with mothers Maria and Elizabeth — other William Taylors.

THE METHOD IS THE TRANSFERABLE PART: SEARCH ON THE MOTHER. One query on 'Mary Ann Helen' returns the whole sibling group in a single page, which is what makes an absence provable; searching child by child could only ever have failed to find her. And the control ran first — Agnes Walker Taylor, whose birth this archive had to the day and the address.

AND IT GAVE SOMETHING BETTER THAN IT TOOK: Mary Septima's own baptism, 30 October 1884, which this archive had never held — her parents named together in a parish register for the first time, twenty-seven years before her marriage at the Cape.

ANSWERED 8 OCTOBER 2026. David logged this session in to gro.gov.uk and the search this row named was run: TAYLOR births, mother's maiden surname BRAYLEY, 1870-1893, both sexes, ten windows. NINE CHILDREN, ALL AT BIDEFORD, AND NOTHING MISSING: Kathleen Georgina (1870 D, 05B/490), William John (1870 D, 05B/489), Horace Archibald (1872 J), Isabella Temple (1874 D), Margaret Hellen Walker (1876 D), Irene (1879 M), Lucy Elizabeth (1881 S), Mary Septima (1884 D, 05B/506) and Agnes Walker (1889 S).

THE TREE'S TEN IS WRONG, EDITH AND ETHEL ARE NOT THIS COUPLE'S, AND IRENE HAD NO TWIN. And Mary Septima is the sixth daughter and the eighth child — the name matches neither count. See /corrections/.

THIS ROW WAS WRITTEN WRONG AND IS CORRECTED. It first said "the GRO death index for Bideford, 1870–1884, under TAYLOR is the cheap test nobody has run." It was run on 17 September, row 128: FreeBMD deaths, TAYLOR, Bideford, 1881–1891 — 37 entries, fourteen of them aged 0–3. Written from memory while finishing a different piece of work, which is the second time in one day this archive has claimed something was undone when it was done.

What this row actually adds, which is the baptism register. Row 128 weighed two candidates for the missing daughter and could attach neither. Both are now harder.

EDITH, born Bideford Sep quarter 1882: not in the family's baptism register, where every named sibling is. Downgraded from strong to weak.

ETHEL, born Bideford Dec quarter 1882 and dead at Bideford aged two — the candidate row 128 called the better fit, because she would appear in the birth index and in no census. She is not in the register either. She would have been about two when she died, and this family baptised Lucy at a few weeks and Agnes at eighteen days. So either she is not theirs, or she went unbaptised for two years — a disjunction, not a refutation, but it is the first evidence that bears on her at all.

SO THE HOLE HAS A SHARPER SHAPE: a child of this couple born between 31 August 1881 and 30 October 1884 — the exact gap in the register — who left a birth registration and nothing else: no baptism, no census, no household entry.

AND THE ONE SEARCH THAT SETTLES IT IS STILL DAVID'S, as row 128 said in September. gro.gov.uk is not FreeBMD, and the difference is the whole point: it gives the mother's maiden name on every birth from 1837. The search is births, surname TAYLOR, mother's maiden name BRAYLEY, district BIDEFORD, 1870–1890 — which returns every child of this couple by definition, the tenth included. It needs a free GRO account, and this session does not create accounts.

ANSWERED 8 OCTOBER 2026. David logged this session in to gro.gov.uk and the search this row named was run: TAYLOR births, mother's maiden surname BRAYLEY, 1870-1893, both sexes, ten windows. NINE CHILDREN, ALL AT BIDEFORD, AND NOTHING MISSING: Kathleen Georgina (1870 D, 05B/490), William John (1870 D, 05B/489), Horace Archibald (1872 J), Isabella Temple (1874 D), Margaret Hellen Walker (1876 D), Irene (1879 M), Lucy Elizabeth (1881 S), Mary Septima (1884 D, 05B/506) and Agnes Walker (1889 S).

THE TREE'S TEN IS WRONG, EDITH AND ETHEL ARE NOT THIS COUPLE'S, AND IRENE HAD NO TWIN. And Mary Septima is the sixth daughter and the eighth child — the name matches neither count. See /corrections/.

A route found on 9 October and not yet worked. FamilySearch's FULL-TEXT search reads the words on unindexed images, so it goes round the transcriber entirely. Uruguay is in the corpus — and so is Arquidiócesis de Montevideo, Marriage Dispensation Records 1856–1932, along with Flores and Trinidad baptisms and deaths, Montevideo burial registers 1886–2021, and the Catedral's own books.

WHY DISPENSATIONS: data/canelones-films.tsv already says it — "the pre-marriage investigation is worth MORE than the act. It routinely gives a Spanish groom's PARENTS." Parentage is the one thing this archive is short of at the Uruguayan end, and Trinidad in Flores is where it has a crowd of Lerenas and almost no parents.

AND FULL TEXT FINDS WHAT AN INDEX CANNOT: a man with a COMPOUND surname is filed under his father's name. Searching "Juan Carlos Lerena" returned exactly two hits in the whole corpus, one of them Don Juan Carlos GONZÁLEZ LERENA, dispensed from banns at Montevideo on 8 November 192-, entry N. 14610 — a man no surname index of LERENA would ever have surfaced. He is set aside (González from his father, so his children would be González, and Pablo Armando is a Lerena) but he is proof the route reaches people the indexes hide.

What to run: LERENA and LLERENA against the dispensation books; then the Trinidad and Flores registers; then the Montevideo burial registers for an infant Alejandro Lerena 1846–1860, which is the one record that would settle the merge this archive has refused twice.

WORKED 9 OCTOBER 2026, BOTH HALVES. South African half: Roque on a Union voters' roll (polling district 770) and Nuno in a land record carrying his exact birth date — and the claim that South Africa is absent from the corpus, corrected.

URUGUAYAN HALF: NINE DISPENSATION ENTRIES READ. Seven new people, Zelmira Carlota Lerena's own marriage found, and the reason for coming here corrected — this book names no parents; it is the Vicar General's summary ledger, not the expediente.

WHAT IT GAVE INSTEAD: THE DEGREE OF KINSHIP IN EACH MARRIAGE. Four Lerena-Peláez unions, two of them between cousins; two Lerena women of Trinidad dispensed in one sitting on 26 June 1885, one to a Fernández and one to a Peláez; and Juan Lerena dispensed twice over, in the second degree AND the third. A kindred marrying inside itself, with the Church recording how closely.

WHAT IS LEFT, and it is the thing that would name parents: the informaciones matrimoniales / expedientes behind these dispensations — a separate series, which data/canelones-films.tsv says routinely gives a groom's parents.

The ledger named no parents; the files behind it should. Nine Lerena dispensations were read on 9 October from the Arquidiócesis de Montevideo summary register. Each is a one-line entry: whom the Vicar General dispensed, from what impediment, and the bride's parish. The informaciones matrimoniales are a different series, and data/canelones-films.tsv records that they routinely give a Spanish groom's parents.

THE FOUR WORTH CHASING FIRST ARE THE CONSANGUINITY ONES, because a dispensation of kinship required the pedigree to be set out before it was granted: Antonina Lerena and Serapio Fernández (Trinidad, 26 Jun 1885, first cousins); María Lerena and Máximo Peláez (Trinidad, same day); Juan Lerena and Eduviges Peláez (Mercedes, 31 Mar 1894, second AND third degree); Zelmira Carlota Lerena and Adolfo Cordero (Metropolitana, 1 Feb 1901).

Zelmira's is the one to open first — she is already in this register, she was once written off as "not ours", and her husband was her blood relative.

WORKED 9 OCTOBER 2026, AND THE ANSWER CAME FROM A DIFFERENT SERIES. The expedientes were not needed. Zelmira's marriage ACT (La Unión 1902–1934) named four parents and moved her wedding from 1901 to 1904. And a Flores BIRTH ACT of 1897 named all four grandparents of a child of the 1885 couple, which is what a Uruguayan civil birth registration does and no dispensation ledger ever will.

IT EXPLAINED THE CONSANGUINITY. Serapio Fernández's mother was SOCORRO LERENA; Antonina's father was BERNABÉ LERENA. Lerena blood on both sides — which is what the Archbishop was dispensing. Bernabé and Casimira Peláez have ages at last (born about 1835 and 1841, at Cerro Colorado) and a fifth child.

AND IT REACHED BACK A GENERATION FURTHER: at Trinidad in 1880, María Peláez, thirty, daughter of Juan Peláez and Antonina Lerena — so a Lerena married a Peláez around 1845. That is why the couples of 1885, 1894 and 1913 were all cousins.

STILL NOT DONE: the acts for Juan Lerena × Eduviges Peláez (Mercedes, 1894) and Vitalino Gregorio Peláez × Florinda Juana Lerena (Trinidad, 1913). Same method: search the full text for the bride, and look for a civil BIRTH act of a child, which names all four grandparents.

Answered 9 October 2026, and by a document rather than an inference. The route this row proposed was Socorro's own marriage act. That is exactly what settled it — but it turned out to be an expediente, not an act, and it was found by full-text search on her name.

Arquidiócesis de Montevideo, Marriage Dispensation Records 1855–1857, images 26–27: the petition of Serapio Fernández at Santísima Trinidad in August 1855 to marry Doña Socorro Lerena, with a sworn witness naming four generations — she is a daughter of Don Alejandro Lerena and Doña Margarita Fernández and a granddaughter of Don Ramón Lerena and Doña María González.

And then the same collection gave Bernabé's own expediente, Santísima Trinidad, 15 February 1858, calling him 'hijo legítimo de Don Alejandro Lerena y Doña Margarita Fernández, naturales de San José' — the same parents — with a kinship diagram the Cura Vicario drew himself, putting Ramón Lerena and María González at the apex over Alejandro and Antonina.

SOCORRO AND BERNABÉ ARE BROTHER AND SISTER. Two expedientes, three years apart, one parish, two clerks. That closes the dispensation of 26 June 1885: Serapio Fernández the younger descends from Socorro, Antonina Lerena from Bernabé, so they are first cousins — exactly the second grade in the equal lateral line the Archbishop lifted.

And the row over-delivered. It was asked for one sibling link and returned the head of the kindred: Ramón Lerena and María González, with children Alejandro and Antonina, all naturales de San José. See data/trinidad-founding-kindred.tsv. What is still open is a different question — whether that Ramón is the Ramón Lerena of Rosario-Colla already in this register, whose son Mauricio married at San José de Mayo in 1837. Not merged.

Opened and closed the same afternoon, and the answer was in this archive's own files the whole time.

The township is GERDVIEW, GERMISTON. His death notice MHG 6746/71 — read on 16 September 2026 and sitting in data/death-notices-read.tsv — gives his residence as 24 Leiden Road, Gerdview, Germiston (Johannesburg struck out), and Cologne Road, where lot 245 lies, is a street in Gerdview, Germiston, postal code 1401. So he bought two stands in the township he lived and died in, from the Township Owners, on a transfer of 22 August 1953.

THE ROW WAS WRITTEN BEFORE THE ARCHIVE WAS CHECKED, AND IT NAMED THE VERY DOCUMENT THAT ANSWERS IT — "the 1971 death notice 6746/71, already cited in his register row, which should give his residence". It had been read three weeks earlier. The check took one grep.

AND THE STREET SCHEME CAUGHT A MISREADING. Gerdview's streets are named for Dutch and German cities — Leiden, Weert, Cologne. "London Street", which this archive had just published for lot 210, did not belong to that scheme, so the plate went back under a harder magnification: it reads LENDIN, with a plain e where London has an o and an n before the d where London has none. Withdrawn; see data/corrections.tsv.

WHAT IS LEFT, AND IT IS SMALL: lot 210's street is still not confidently read. Leiden is the likeliest intended name and the letterforms do not support it, so it stays unresolved rather than guessed. The Deeds Office township register behind Land Records: South Africa, Farm Records 1895–1955 would settle it.

AND ONE NEGATIVE WORTH KEEPING: that death notice leaves father and mother blank — both marked only "Deceased", with no names. The one document of Nuno's that could have named Pablo Armando does not.

THE WHOLE QUESTION TURNS ON TWO IMAGES AND NEITHER NEEDS A SEARCH. (1) DGS 008165206, IMAGE 283 — Montevideo deaths 1913, certificate 553, the death act of the man the index calls 'Eduardo Zalduando Pou'. It gives his surname and his parents in the register's own hand and should name his widow, which settles BOTH open points at once: ZALSAMENDI vs ZALDUANDO, and AMELIA vs ANGÉLICA. (2) FILM 007713400, IMAGE 02828, FOLIO 344 — the 1903 Pocitos baptism, already read once on 21 September 2026, to be re-read for the mother's forename alone now that there is a reason to doubt it. WHY THIS IS BLOCKED AND NOT MERELY UNDONE: on 11 October 2026 the index record page loaded and read normally, and following its own VIEW ORIGINAL DOCUMENT link REDIRECTED TO ident.familysearch.org — a sign-in. This archive holds that redirect as false-zero disguise seventeen and has already had to withdraw three published conclusions to it once, because the control had been run BEFORE the null instead of after. So the plates are recorded as unread, not empty, and no conclusion is drawn from the failure. David signs in; the session then uses it. DONE, 11 OCTOBER 2026, AFTER DAVID SIGNED THE SESSION IN. Both plates were read. The surname is ZALDUONDO on both and ZALSAMENDI was a misreading this archive had carried since 21 September — and had re-argued that same morning. The widow is ANGELICA, and because the 1913 act puts Maria Margarita inside the union with her while the 1903 act puts the same girl inside the union with “Amelia”, the two women are one and are now merged. Also gained: his birthplace (Montevideo), his pension, his cause of death, his address, both parents dead by 1913, a fourth name for the daughter (LUISA), and a third child, MARIA ELENA.

NOTICED 11 OCTOBER 2026, from a generated file that was right about its inputs and wrong about the people. The 1897 marriage act names JUAN ZALDUONDO, 80 and DOLORES POU, 58 as living parties, and their son's 1913 death act calls them fallecidos. So this archive knows they were alive on 20 November 1897 and dead by 22 August 1913 — a bracket of fifteen years. data/last-attested-alive.tsv files both as 'born_est only - a floor, not an attestation / BIRTH ONLY', because the tool reads the register's born_est column and nothing else. THE SAME IS TRUE OF EVERY WITNESS, GODPARENT AND NAMED PARENT IN THE REGISTER — Eugenio Marcos Abella, Maria Ester Costa, Agustin Riva and dozens more are all attested alive on a dated act and all filed as birth-only. WHY IT MATTERS: the point of a last-alive ledger is to stop a person being proposed for a record they cannot have been alive for, and a ledger that under-reports attestations cannot do that job. WHAT WOULD FIX IT: the register has no column for 'attested alive on', and the facts live in prose notes. Either a column is added — which is a change to the archive's central file and belongs to whoever owns it — or the tool learns to read dated acts out of data/relations.tsv, where the note already carries them. NOT ATTEMPTED HERE, because guessing dates out of English and Spanish prose is exactly the method tools/check-corrections.py says was tried and abandoned once already.

FOUND BY ACCIDENT ON 11 OCTOBER 2026, while running a control. Testing whether the full-text corpus reaches Argentina at all, a search for “Rosario de Santa Fe” returned 245 pages — and the Argentine material in them is not parish registers. It is Buenos Aires City: Probate Estate Case Files (1858, 1887, 1894 among the first page alone), Ship Inspection Files (1924), and notarial records. WHY THIS MATTERS TO THIS FAMILY. A probate file is the richest civil record there is: it names heirs, widows, children, addresses and property, and it names them in one place. This archive already has a Buenos Aires thread — the 1918 San José de Flores baptism of the Villademoros branch, Juan Carlos Llerena born there in 1910, Alejandrina Llerena standing godmother in 1910 — and it has never looked at a probate file for any of them. AND THE SHIP INSPECTION FILES ARE A SECOND SEAM: this family crossed the Plata repeatedly, and the archive's passenger-list work has all been CEMLA arrivals. An inspection file is the other side of the same voyage. HOW TO WORK IT: full-text phrase searches on surnames, since these files are indexed by case title (“Villarruel, Blas; Luna, Rosa”) and not by every person inside them — which is exactly why the record index cannot see into them and the full text can. NOT STARTED. STARTED THE SAME DAY IT WAS CREATED, and the first twenty results of seven hundred already carry three leads. See data/buenos-aires-fulltext-2026-10-11.tsv. (1) A RAMÓN LERENA SIGNING AS ALCALDE IN BUENOS AIRES IN 1800, inside a probate file. This register's Ramón Lerena of Rosario-Colla is recorded as flourishing c.1800-1820 and stands at the apex of the 1858 Trinidad kinship diagram. The date fits and the place does not — Buenos Aires is across the river from Colla — so this is a candidate and not a match. (2) A VICENTE LERENA DÍAZ keeping the civil register of Pueblo Brugo, Paraná, ENTRE RÍOS in 1938. The compound is the point: Arturo Lerena García married Amalia Díaz, and their children would be surnamed exactly that. Bears directly on row 159. (3) SPANISH-BORN LERENAS IN A BUENOS AIRES PARISH — San José de Flores 1909-1912 names D. Eloy Lerena, natural de España, Doña Felisa Lerena, natural de España, and D. Gregorio Lerena y Doña Rosa Defenari. This register holds a Felisa and a Gregorio from CEMLA passenger lists with no record on Argentine soil at all. This is the first chance to turn an arrival into a life. ALL THREE ARE OCR SNIPPETS WITH ELLIPSES AT BOTH ENDS. NO REGISTER ROWS WERE CREATED and none should be until the acts are opened. The 503 Buenos Aires LEGAL results were not filtered to on their own, and 680 of the 700 Argentine results are unread. WORK RESUMED AND STOPPED AGAIN THE SAME DAY — THE SESSION IS SIGNED OUT, AND THIS TIME COMPLETELY. The first frame attempted was the 1800 probate's next image, to read the alcalde's jurisdiction and settle whether that Ramón Lerena is this register's own. It redirected to ident.familysearch.org. So did the full-text search itself, which is the difference from this morning: then the index pages read normally while only the images were shut, and now nothing on FamilySearch is reachable. NOTHING WAS RECORDED FROM THAT STATE AND NO NULL WAS TAKEN — a signed-out session returns zero for everything, and this archive has already withdrawn three published conclusions to exactly that. WHAT IS WAITING, in order: (1) the 1800 probate, one frame of 249 read — an alcalde's jurisdiction is normally stated in full somewhere in a file like that, and it is the single highest-value question open; (2) 680 of the 700 Argentine results unread, 503 of them Buenos Aires legal files; (3) the Entre Ríos snippet — Vicente Lerena Díaz keeping the register at Pueblo Brugo in 1938 — still unopened. David signs in; this session then uses it. AND IT HAPPENED A SECOND TIME WITHIN THE HOUR — the sign-out is recurring, not a one-off. Three times on 11 October 2026 the session lost FamilySearch: once partially (index readable, images shut) and twice completely (search redirecting too). THE OPERATIONAL FACT FOR THE NEXT SESSION: a FamilySearch session here lasts roughly an hour. Plan the expensive reads first and re-check the instrument after each null rather than before. Nothing was read and no null was taken from either signed-out state. ALL THREE LEADS ARE NOW OPENED. The San José de Flores baptism, the 1800 probate (whose court order settled which Rosario), and now the Entre Ríos register — VICENTE LERENA DÍAZ is the Jefe del Registro Civil of Pueblo Brugo, a strong but unmerged candidate for a third son of Arturo Lerena García. Nine frames of the probate read of 249, and 680 of the 700 Argentine results still unread. The row stays running.

FOUND 11 OCTOBER 2026 while trying to use it. BranchRivers.astro is the drawing this archive wanted — several families of one surname, as rivers, with lanes, unplaced households and candidate-join bridges. It is better than the component written here instead. IT COULD NOT BE USED AS IT STANDS. At line 226 it renders <text ...>— the commonest patronymic in Italy</text> outside the LOOSE.map(), so it prints unconditionally — even with loose: []. On a LERENA page that sentence is simply false: the name is rare Spanish-Basque, and Forebears puts it at 1 in 3,431,048. THIS IS THE SAME FAULT THE COMPONENT'S OWN DOCSTRING RECORDS FIXING ONCE AND MISSING HERE: “every other archive that adopted the drawing got Defranceschi's words underneath it, describing rivers it does not have”. The caption was made a prop; this note was not. THE FIX IS ONE LINE — move it inside the map, or make it a looseNote prop defaulting to the present text so Defranceschi keeps its sentence and nobody else inherits it. MEANWHILE this archive wrote site/src/components/Lines.astro, which is simpler and has no bridges. WHEN THE KIT IS FIXED, THIS ARCHIVE SHOULD SWITCH TO IT AND DELETE ITS OWN — the kit's version supports candidate joins, which is exactly what the Sancho-to-Trinidad question will need the day any evidence appears.

FOUND 11 OCTOBER 2026 AND NEVER LOOKED AT. The Spanish consulate at Buenos Aires records a power of attorney of 14 November 1900 by Prudencio Benito and Juan José LERENA Y LERENA, bachelors of Navarrete, to their widowed mother Isabela Lerena y Basarán and their married brother José Dolores Lerena y Lerena. A Madrid pension file of 1896 shows that same brother signing as juez municipal suplente of Navarrete and keeper of its civil register, certifying from tomo 11, folio 191 of the Defunciones. SO THE BOOKS EXIST AND ARE DESCRIBED. This archive believes none of these people are Pablo Armando's — they are Riojan Spaniards and every lane here is Banda Oriental from the 1790s — but it is the only place on earth it can name where the doubled surname was settled local usage, and the marriage that produced Lerena y Lerena is documented only by its children. WHAT TO LOOK FOR: the marriage of Isabela Lerena y Basarán to a Lerena, and whether any son went to Rosario de Santa Fe.

THREE OF FIFTEEN WERE READ ON 11 OCTOBER 2026 AND TWO OF THE THREE PAID. The phrase mostly finds two Lerenas joined by and rather than the compound surname — and in this archive that is better, because two Lerenas at one font are usually kin. The Montevideo cathedral act gave the Larriera grandparents; the Cordón 1891-92 book moved Adolfo Lerena back twelve years and produced a new María Mercedes. STILL UNREAD: Cordón 1900-1902; Montevideo 1890-91 and 1901-02; Uruguay 1917-18 (snippet: Alejandro Lerena y Clorinda Lerena); and seven volumes of Trinidad, Flores, one of which reports 54 matches (snippet: padrinos Juan Lerena y Margarita Lerena). The Trinidad ones are the lowest priority — that town has had a sweep — and the Cordón and Montevideo ones the highest. DONE 12 OCTOBER 2026 — all fifteen volumes read. The phrase finds two Lerenas standing together, nearly always as godparents, and most pairs are people already held here. Three things came out of it worth more than the pairs: a “Montevideo” volume turned out to be the capital’s civil register transcribing a Trinidad parish certificate, so a place filter does not separate the two towns; the Trinidad sweep missed the SERENA spelling entirely, with four Serenas in one 1905-09 volume pair; and two more children** appeared for Fernando Moratorio and Josefa Lerena Traibel. Nine people entered, all index-level.

A SPANIARD IN BUENOS AIRES APPEARS IN THESE WITH HIS VILLAGE AND A CONSULAR ROLL NUMBER, which no Argentine civil act gives. The 1900-1901 volume produced the Navarrete household on 11 October 2026. A SECOND VOLUME, 1909, IS IN THE SAME RESULT SET AND IS UNREAD. Worth it for any Lerena who was a Spanish subject — and this archive holds a stream of them out of Logroño into Buenos Aires commerce: Martín at 14 in 1907, José M. at 16 in 1909, José at 16 in 1924, Emilio and José María in 1930.

A one-line snippet seen on 11 October 2026 while paging the Argentine full-text results: “lo hace el Doctor Lerena si la”. NOT OPENED. A Lerena with a professional title in the Buenos Aires courts in 1889 is worth a look, and the same court the following year heard Carlos A. Lerena sue the Escalada estate for a broker's commission. CAUTION BEFORE SPENDING TIME: in this corpus lerena is very often the transcriber's reading of herencia, so check the sentence before the name — see data/false-zeros.tsv. DONE 12 OCTOBER 2026, AND THE WARNING EARNED ITS PLACE. He is real: the Agente Fiscal's dictamen names him as the curador appointed in the Feliciano Aguirre probate, and objects that the court appointed him without consulting the Ministry or inviting the Spanish consul. The same passage spells him LERENA and LLERENA, twice each, in one hand. Doctor is again the style of an abogado. Three seams opened and not worked: the same Doctor Lerena in the Buenos Aires probates of 1890, 1894 and 1897; a physician of the name signing Montevideo death certificates in 1949 and 1967; and 'Señores Doctor Lerena' in the student files of the Universidad de Toledo, 1790s-1830s.

THE SERENA SWEEP IS DONE, AND ITS FIRST HEADLINE WAS WRONG. The Trinidad sweeps were run on LERENA; a sweep on SERENA found children for two couples this archive had spent weeks on and had no children for — Custodio Lerena with Ema Cardozo (1911) and Juan Bernardo Lerena Peláez with Eduviges Peláez (1894). But the 290-result count was the size of the surname in Uruguay, not of the seam: q.anyPlace ranks and does not filter, and the Flores set is about twenty-two records, all now read. WHAT IS ACTUALLY LEFT: the same sweep for PERENA, SEREÑA with a tilde and GERENA — all three documented at Trinidad in one decade, and the G-for-L known here since Sancho's baptism of 1811. DONE 12 OCTOBER 2026. PERENA gave 3 Flores records, GERENA 7, LLERENA 11; SEREÑA turned out not to be a separate seam at all — FamilySearch normalises the ñ, so it returns the identical 290 as SERENA. Between them they produced a fifth child for Cipriano, the first child ever found of Alejandro the younger, two daughters for Bernabé and Casimira eleven years earlier than anything held here, and the completion of the Fermín Lerena × Paulina Barragán household. Every set paged to where the matches stop.

FamilySearch has not filmed them and will not answer this one. The catalogue for Navarrete, Logrono returns a single subject — Census — and a single item, the 1752 Catastro; the full-text corpus holds no Navarrete volume at all. What to ask for: the marriage of a LERENA groom to a LERENA bride at Navarrete before about 1876 — the parents of Prudencio Benito, Juan Jose and Jose Dolores LERENA y LERENA, the bride Isabela LERENA y Basaran, widowed by November 1900 — and the three sons' baptisms. This is the only place this archive can name where the doubled surname was settled local usage, and open question 1 has called it worth a day since 11 October.

DGS 4386870 (vols 481–484) and DGS 4386703 (vol 485, Memoriales de seglares), filmed from the Archivo Historico Provincial de Logrono. Vols 483 and 485 are the household-by-household ones. It cannot name a Juan Carlos born in the 1840s; it answers the question underneath him — was LERENA a Navarrete surname at all before the 19th century, and was it already being doubled? No waypoint established, film numbers only, and nothing has been opened.

Seccion 18, Subdivision 34, Ciudad de Buenos Aires, with Emilia Lerena b.1886, also born in Spain (index arks MWW4-GMZ, MWW4-GMD). Almost every Argentine Lerena this archive has traced is Uruguayan-born, so a Spanish-born household in the city five years before the Navarrete brothers signed at the consulate is worth an hour. Index-level only — the census page itself has not been opened and nothing is tiered.

Resized downwards on 12 October, the same day it was raised. The 470 figure was the neighbourhood, not the parish: q.anyPlace=Cordón scopes the barrio, and 95 of the first 100 hits are parish volumes while pages 2 and 3 hold none. The correction. So this is an afternoon, not a campaign. Baptisms 1828–1923, marriages 1842–1931, burials and deaths 1839–1926. Densest: Religious Marriage Records 1916–1926 (6), the bundled Rite Registers / Marriage Dispensation Records / Religious Marriage Records 1895–1898 (5 — a parish dispensation book nobody here knew existed, three of its five unread), and the baptism registers of 1869–72, 1904–08 and 1908–10 (4 each). Still untouched: the Manuel Antonio Lerena recorded paraguayo.

Done the same evening it was raised. The act confirms the index: hijo legítimo de Don PEDRO A. LERENA, oriental, fallecido, y de Doña VICENTA DE LA VEGA, oriental, fallecida — so the Pedro who married Manuela Lemos is a son of that couple, and the question this register asked on 11 October is answered. It also names three children new to the archive — Antonio, Carlos Teodoro and Manuel Antonio — and a fourth son, the declarant. And it breaks a date: sixty-nine in 1896 means born about 1827, so he is six in the census this archive dates c.1833, not sixteen. [Re-dating household 229](#) is now its own row.

Settled at catalogue level. FamilySearch catalogue koha:1408224 lists film 7499042 item by item: item 1 = Padrones de habitantes, libro 263, 1843; item 2 = libro 264, 1850; item 3 = libro 257, 1852–53. Household 229 is in item 2, so the enumeration is the padrón of 1850 and c.1833 is withdrawn. Three birth years moved seventeen years later — Emilia c.1832, Pedro c.1834, Vicenta c.1838 — and the head, entered at forty, was born about 1810, so he is not Pedro Antonio Lerena born 1793. That holds under the 1843 book too. The old date had been inferred from the identification it was then used to support. The correction.

Estanislada 20, Sinforiano 19, Manuela 60 and Rosa 46 carry the LERENA surname by ditto at casa 229 of Calle Sarandí in the padrón of 1850; Manuela's patria is ÁFRICA. A census has no column for condition — a parish act does, and writes liberta, esclava de or parda libre outright. The Cordón's death register of 1839–1852 is already identified here and unread, and the Catedral covers the same years. Manuela and Rosa are both casada, so each has a husband who is not in that house.

The consul test of 12 October found no LERENA in any South African gazette before 1951, with both controls passing — coverage 1910–2026, and exequatur returning 644 documents with hits in 1914, 1932, 1933 and 1937. That null is on one spelling. This archive's own rule is that a null on one spelling is a null on a letter, and 1910s–40s gazette OCR is exactly where a surname breaks. Run SERENA, LLERENA, LORENA, LARENA, LEVENA, LERINA, GERENA, PERENA, looking only for pre-1951 hits. Filter results on the /akn/za link prefix — the site's country= parameter does not bind.

Gazettes 2594 of 30 December 1938 and 4389 of 26 May 1950 both surface on a search for Argentine Republic and consul, and both sit behind a Cloudflare bot check. This archive does not work around bot detection, so they are recorded as blocked rather than searched; an ordinary browser should open them. If either names the Argentine consul of the day and it is not Pablo Armando Lerena, open question 2 is answered outright — the most decisive thing left on it.