The work list

Done 15 September. Thirteen annual volumes, 55 images, scanned by cropping the surname column from every page and tiling them eight to a sheet. Two new children — Pasquale Falco (1838) and Matteo Falco (1842) — five empty years, four years already held, and one year that is not a negative: the 1834 index is a single leaf covering A to C and marked END. This did not close Antonia di Guida's death or Pasqua Falco, which is what it was picked up for. Neither is in the stretch, so both move to the register itself.

Done 15 September. All three magnified and settled. Two are real children and are published: a second Giuseppe Cioffi (5 Dec 1830) and a second Mariantonia Cioffi (3 Dec 1832), which brings that household to six children buried in nine years, none known to have reached adulthood. The third is a conflict left open on purpose: the index says «Morgillo Raffaela, 9 Marzo 1832», the act already held says «Raffaele Morgillo, 22, contadino, 9 May 1832». Two children nine weeks apart, or one death read two ways. Reading atto 14 of 1832 against the held act settles it — that is now the work, and it is small.

DONE, and the answer is that the record is out of reach — which is an answer. 1865 (`an_ua14367`) read end to end: 147 acts, no Falco death. The Antenati Arienzo death series stops there. VINCENZO FALCO was ALIVE on 16 December 1864, named without «fu» in his son Luigi's death act, a *colono* still on strada Camellara at seventy-two. So his death is 1866 or later, in the post-Unification *Stato civile italiano* this archive has never located for Arienzo. The months of searching 1844–1865 were not wasted: they are what proves it. GIUSEPPE FALCO of Antonia di Guida is in the same position — alive November 1849, and in none of 1850–1859 or 1865. See rows 39 and 40, which now share one answer.

Answered 15 September: the register is the finding aid. The «Matrimoni, indice» series stops at 1849, but the Matrimoni registers run on at `an_ua14213` + (year − 1809) — 28 to 37 acts a year, each an «atto di solenne promessa» naming both spouses, both sets of parents, and the parish celebration date in the margin. Small enough to read outright. The first volume opened produced Sara Falco's marriage (14 Feb 1855, to Antonio Mauro, fruttajolo). Nunziata is not in it, which is now its own question on the searched page. Antonio Falco's 1869 marriage is beyond this series, which ends with the Restoration in 1865.

The lead is confirmed and the baptism is still unread. The cached index leaf (`3Q9M-CSD3-1325-9`) reads «Sara Falco fol. 47» — an alphabetical index by FORENAME, so she sits in the S run beside Salvatore, Sabbatino and Saverio. Reading folio 47 needs the image, which depends on row 18. The search for her turned up something else entirely: SARA RUGGIERO, her grandmother, as MATRINA on ten pages across six Falco households — published, at transcript strength. Eight of the nine 'Sara' hits in the harvest were the grandmother, not the granddaughter.

Done 15 September — two men. The 1817 birth register holds both: Arcangelo Crisci 52, strada Camellara, husband of Sara Ruggiero; and Arcangelo Crisci 30, strada Capodiconca, husband of Maria Martone. Twenty-two years, two streets, two wives. Andreana's father is not Anna Falco's husband. Both ages were already in `data/sweep-people.tsv` and had been for weeks. A smaller question opened: Anna's husband is son of «fu Nicola, e Chiara Majone», and the younger Arcangelo was married to Maria Martone in 1817 — so either he remarried within months or there is a third man of the name.

Done 15 September — and the answer is no. Nunziata Falco is not in the Arienzo marriage registers for 1854, 1855 or 1856, all read act by act. She is held here from a parish record as marrying «13 [Nov] 1855», the month in brackets. Either that date is wrong or she married at her husband's comune. Nothing connects her to Giuseppe Nunzio beyond the shared root of the name, and this archive is not going to build on that. The sweep found other things instead: Pasqua Falco's marriage (9 Sep 1856), which vindicated a refusal; Giuseppe Falco of strada Costa remarrying in 1854 with his parents named; and Gelsomina Vigliotta dead before December 1854, stated twice.

Done. 192 entries read end to end; nineteen were still asserting something a later finding had overturned, eleven of them still labelled `open`. Each now carries the entry that replaced it, named and linked, above its own text. The structural cause was on the page in plain sight: a withdrawal was published as a new entry at the top, and the page sorts by date, so it sat sixty rows from the claim it killed. Three further faults fell out — a forty-volume negative run against a retired surname, a duplicate discovery of an entry on this very page, and the page printing its own markdown for a week.

1819 is done — 137 acts, act by act, because the volume has no annual tavola (both ends checked; the twenty-one-row list on the last leaf is a supplementary list, not the year's index). Three Falco births, all on strada Camellara: Domenico Clemente (18 Mar), Crescenzo — a birth date for a man held by his marriages alone (11 Oct), and Luigi's civil act (21 Dec). Rosaria Falco is not in it either, which closes the last year her birth act could have been in. Next: 1820 (`an_ua14417`, 143 images) — still hunting Maria Rosa Falco, absent from 1817, 1818 and 1819. Then 1829 onward and 1845–1860. 16 September — the tavole are NOT there for the years that matter. 1829 (`an_ua14426`, 114 images, 107 acts) has no tavola at either end: image 2 is act 1, image 113 the closing certificate, image 114 the Procuratore's. 1837 (`an_ua14434`, 146 images) likewise — image 2 is act 1. The archival tag «*THIS INDEX HAS BEEN MICROFILMED ALREADY*» on the 1829 endpapers refers to a separately filmed index, not to anything bound in. So 1829 onward is an act-by-act read at roughly one act per leaf, and the one-image-per-year economy that made 1828, 1848 and 1859 cheap does not apply. 1830 is 149 images, 1837 is 146. Budget accordingly; the sheets must be three columns wide to carry a negative, not four. ## WHAT THE TAVOLE ACTUALLY ARE, VOLUME BY VOLUME — 21 September 2026 Row 86 opened the index leaf of every birth register it read, and most of them cannot do the job this row wants: - Complete and surname-ordered, with both parents named: 1826, 1827, 1828. These three are the only ones that answer «did a Falco father register a child this year?» outright. - A blank printed form, never filled in: 1829. - Filled but stops at C: 1832 — eighteen entries for a hundred and ten acts. - Under the archivist's card «THIS INDEX HAS BEEN MICROFILMED ALREADY»: 1816, 1823, 1824, 1825, 1830 — and 1846, which this row already knew. Six volumes, one policy. - Only its last page filmed, with the child's-name column lost in the gutter: 1819. - Alphabetical by the child's FORENAME, so a surname cannot be looked up in it at all: 1818. And no tavola can ever answer for a FALCO MOTHER, because her child is filed under her husband's surname. All nine children row 86 recovered came that way. So this row's instrument is weaker than it looks even where it survives, and the act-by-act read is not a fallback but the primary method. --- ## THE «NEXT» IN THIS ROW HAS BEEN WRONG FOR THREE DAYS — 21 September 2026 Halfway up this row stands the sentence **«*Next: 1820 … Then 1829 onward and 1845–1860*». 1820, 1829, 1830, 1831, 1832, 1833, 1834, 1835, 1836, 1837, 1838, 1845, 1846, 1847 and 1848 have all been read since it was written, most of them by rows 85 and 86 and at both bands. A reader following this row's own instruction would have re-read fifteen volumes. `tools/staleness_report.py` did not catch it**, and the reason is worth more than the fix: the tool looks for phrases like *still to read* and *what remains*, and this row phrases its remainder as «Next:» — a word so ordinary the pattern list had left it out. It has been added. ### The Arienzo BIRTH registers that are genuinely unread, as of 21 September 2026 Counted from the coverage rows on What has been read, not from memory: - 1809–1815 — seven volumes, the earliest the civil registers reach. *`an_ua14406`–`an_ua14412`.* - 1839–1844 — six volumes, the gap between row 85's reading and 1845. *`an_ua14436`–`an_ua14441`.* - 1849–1865 — seventeen volumes. *`an_ua14446`–`an_ua14462`.* Thirty volumes. Every year from 1816 to 1838 and from 1845 to 1848 now carries a coverage row. And the cheap instrument is gone for all thirty, because the tavola survey above found only 1826, 1827 and 1828 usable, and none of those is in this list. Thirty act-by-act reads at two bands each, which is sixty passes of roughly 130 images. *That is the honest cost, and it is why this row should say what it costs rather than say «Next».* 1813 is read, 22 September 2026 — `an_ua14410`, 53 acts, the first of the 1809–1815 volumes this archive has ever opened. One Falco: ANTONIO, born 5 January 1813 to Giuseppe Falco and Gelsomina Vigliotta of strada Costa, four years earlier than anything that household held. The unread list for this decade is now 1809, 1810, 1811, 1812, 1814, 1815 — six volumes. And the band for all of them is `--band 0.13 0.25 --half LR --cols 3 --width 2600`: the Napoleonic format runs as continuous prose with the father, the mother and the child's name in one block, so one band reads the whole act and the two-band rule of the 1820s does not apply. *Nine images to a sheet; at four columns the names are gone.* 1812 is read too, the same day — `an_ua14409`, 93 images, no Falco in the year at all. The unread list for the decade is now 1809, 1810, 1811, 1814, 1815. And 1812's tavola is BLOCKED — «*Tavola Annuale Alfabetica de' Nati … Anno 1812*» at image 92, alphabetical by the child's forename, under the card «*THIS INDEX HAS BEEN MICROFILMED ALREADY*» with only its first page filmed. That makes seven: 1812, 1816, 1823, 1824, 1825, 1830 and 1846. --- ## THREE ARRICALE FROM THE 1858–1859 MARRIAGES — one is a candidate sister, and it is NOT claimed Found 23 September 2026 while re-reading the Arienzo marriage registers for 1858 and 1859 — a sweep this archive had already done on 21 September and recorded in `searched.json`, which the re-reading did not know. *The earlier pass was hunting Falco and did not record these.* | act | person | parents | |---|---|---| | 1858 act 6 | LUCIA ARRICALE, 22 (b. c.1836) | **Gaetano Arricale *defunto* and Maria Rivetti | | 1858 act 25 | — | Pellegrina Arricale, the bride's mother | | 1859 act 10 | ANTONIA ARRICALE, 27 (b. c.1832) | FRANCESCO ARRICALE** *(living)* and ANGELA PIGNATELLI | ANTONIA IS A CANDIDATE, NOT A SISTER. This archive's GIOVANNA ARRICALE was born 22 December 1848 to Francesco Arricale and PASQUA FALCO. Antonia's father is also a Francesco Arricale — but her mother is ANGELA PIGNATELLI, new to this archive, and sixteen years separate the two daughters. *A Francesco Arricale with a first wife Angela Pignatelli and a second wife Pasqua Falco would explain it exactly. So would two men of one name, in a town where this archive has repeatedly proved that a name identifies nobody.* Nothing joins them and nothing is claimed — but Antonia's own birth act, about 1832, would name her mother and settle it, and that is a cheaper target than sweeping fifteen years of tavole. *All three Arricale, and Angela Pignatelli, were run through `tools/isitnew.py` before being written down. All are new; «Antonia Arricale» matched only two parish machine-readings of 1792 and 1794, a different woman.* --- ## THE ARRICALE CANDIDATE IS CLOSED — 23 September 2026, and it cost one leaf The 1859 marriage register gave a candidate sister; the 1859 PROCESSETTI settled it in one image. Arienzo, Matrimoni 1859, act 10 — ANTONIA ARRICALE, twenty-seven, marries ANGELO MAGLIATO of San Felice, «*figlia di FRANCESCO ARRICALE … e di ANGELA PIGNATELLI*». *Same surname as generation five's wife, and a father of the same forename.* The marriage dossier holds her own birth act. `an_ua14206` — Arienzo PROCESSETTI 1859, 350 images, a volume nothing in this archive had ever mentioned — bundle 10 runs images 123–133, and image 127 is a certified copy of Arienzo Nati 1832, atto 37: > «*L'anno mille ottocento TRENTADUE, il dì NOVE del mese di APRILE … è comparsa CATARINA RUGGIERO, di anni SESSANTOTTO, di professione LEVATRICE … quale ci à presentato una FEMINA … essa è nata da ANGELA PIGNATELLI, di anni TRENTA … domiciliata ivi strada CAPODICONCA, e da FRANCESCO ARRICALE, MARITO DELLA STESSA, di anni TRENTA, di professione CONTADINO*» and the baptism column names her: «*il Sagramento del battesimo è stato amministrato AD ANTONIA ARRICALE*». ### TWO MEN OF ONE NAME ON ONE STREET, TWENTY-TWO YEARS APART | | Antonia's father | Giovanna's father | |---|---|---| | act | Nati 1832, atto 37 | Nati 1848, atto 114 | | stated age | TRENTA → b. c.1802 | VENTIQUATTRO → b. c.1824 | | wife | Angela Pignatelli, 30 | Pasqua Falco, 25 | | street | CAPODICONCA | CAPODICONCA | | trade | contadino | contadino | A man born in 1824 was eight years old in 1832. ANTONIA ARRICALE IS NOT GIOVANNA ARRICALE'S SISTER. SHE MAY BE HER AUNT. The elder Francesco is exactly the right age to be the younger's father. *That is a candidate and it is not claimed* — what would settle it is the younger Francesco's own birth act, about 1824, naming his parents. → [[wrong-negatives]] AND BOTH HER PARENTS WERE ALIVE ON 3 APRIL 1859 — image 128 of the same bundle is their consent, and they appeared in person, which dates them by a document rather than by the absence of «fu». The households are published: «Francesco Arricale & Angela Pignatelli» and «Angelo Magliato & Antonia Arricale». *No kinship to this line is claimed for either.* METHOD NOTE, AND IT IS THE CHEAP PART. This row proposes sweeping birth *tavole* across fifteen years. A marriage dossier holds the bride's birth act already, quoted in full with both parents and their ages — so for any Arricale who married between 1809 and 1865, the processetti answer the parentage question in one image instead of a volume. `an_ua14149` (1809) … `an_ua14211` (1865), bundle *n* = marriage act *n*. → [[processetti-method]]

The 1825 index leaf is covered by an inserted sheet in the filming held here, and the photographer's own card lies across the frame reading *THIS INDEX HAS BEEN MICROFILMED ALREADY*. So it exists. The method that found Gioacchino's 1779 baptism — regexing already-cached transcripts for a duplicate shoot — is the way in. 16 September — IT IS NOT JUST 1825. The 1846 birth volume (`an_ua14443`, img 109) carries the same card, in the same hand: «*THIS INDEX HAS BEEN MICROFILMED ALREADY*», laid across the tavola leaf, with the whole index showing through mirrored from the verso and nothing readable. Two volumes twenty-one years apart, one policy. So this row is not about recovering a single leaf — the Arienzo birth tavole were filmed as a separate series and that series has never been located. Finding it would hand this archive an every-year index to the births, which is the instrument it has been sweeping ninety acts at a time to do without. Raise the priority accordingly. 17 September — ONE PLACE RULED OUT, AND IT WAS WORTH RULING OUT. The `an_ua140xx` ark block, which sits just below the marriage-index block and looked like a candidate for a separate index filming, is not the tavole. It is the «Diversi» series — «*Atti per le nascite avvenuti in un viaggio di mare, Per le Ricognizioni, Adozioni, Persone defunte e fuori del proprio domicilio, e Morti-Nati*» — and it turned out to be a seam of its own. See the new row. The birth tavole are still unlocated. 21 September 2026 — 1825 NO LONGER NEEDS IT, AND THE LEAF IS CONFIRMED PRESENT. The 1825 volume has been read act by act at both bands, so the Falco question it was wanted for is answered without it. And the leaf is where this row says: image 143, the filled tavola with a sheet laid across it, only «*Provincia di Terra di Lavoro, Circondario d'Arienzo, Tavola Alfabetica annuale de' Nati*» showing above the card. *A separate, entirely blank printed tavola form sits at image 142 and is not the same leaf — anyone re-checking this should not mistake one for the other.* The 1824 volume carries the same card, across the index on image 154: three volumes now — 1824, 1825, 1846 — one policy, one missing series.

Generation four's marriage act names «*Josaphat Ruggiero, parroco di Sant'Agnese*» and makes the bride his parishioner; the civil act, found the same day, says she was born and still domiciled at San Felice a Cancello. This archive has been calling Sant'Agnese one of Arienzo's three parishes on the strength of the film the act sits in — and a film label covers a whole reel. One rate window: re-read the act for its own words for the place. 17 September — EVIDENCE, NOT AN ANSWER. Two marriages on ONE leaf of the Sant'Andrea register, June 1820, each marry a bride the writer calls «*meam Parochianam*» to a groom «*ex Parrocchia SANCTAE AGNETIS*» — Gennaro Solli to Angela Grassia on 31 May, and Angelo Clemente Frasca to Angela Rosa Crisci on 1 June. Two men from Sant'Agnese marrying into Sant'Andrea inside two days is the behaviour of a parish next door, not one in another town. It is still not the act saying where Sant'Agnese IS, which is what this row asked for — but it is the first evidence that does not come from a film label. **Both leaves also date the diocese: «*decreto Episcopalis Curiae SANCTAE AGATHAE GOTHORUM*» — Arienzo answered to Sant'Agata de' Goti**, which is where a diocesan archive would hold what the parish does not.

DONE 16 Sep 2026 — THE REGISTER ANSWERS IT ITSELF, AND THE ANSWER IS THAT IT IS NOT CONTINUOUS. Test 1, the numbering gap: there is none. Acts 78 to 91 run unbroken, read from the image. And the volume ends at 91, not ninety — this archive's own statement «*all ninety acts*» was one act short. Test 2, a supplementary list at the back: cannot be run on this filming. There is a tavola leaf after the last act (`an_ua14343` img 55 left), but it is photographed with its ink almost wholly lost and the facing page's writing transferred onto it in mirror. Three order-numbers survive and no names. Not read, and not recorded as absent. Test 3 is the one that answers the question. THE DEATH DATES HAVE AN ELEVEN-DAY HOLE, AND THE FALCO BURIAL IS INSIDE IT. Read act by act from the image, the deaths run: act 84 — 19 November · act 85 — 23 November · act 86 — 26 November · act 87 — 7 DECEMBER. The comune registered nothing between 26 November and 7 December 1841. The parish buried JOANNES FALCO on 27 November. And the hole is not special, which is the real finding. A fortnight earlier the register does it again: act 80 is a death of 30 October and act 81 a death of 13 November — fourteen days. Two silences of a fortnight and eleven days in one autumn, in a register that otherwise enters a death every two or three days that season. This register was not capturing every death at Arienzo, and its own date sequence proves it without any inference from the parish book. So the 1841 civil nil for Falco is not evidence of anything — and neither is any other nil taken from this volume without checking the dates either side of it. Rows 12 and 56 are the same question and both close here.

Done — and the row was measuring the wrong thing, which is the finding. «Promote the harvested files into households» assumed the backlog was people without a household. It is not: 764 of the 1,087 valle-caudina rows have no household, and they should not have one — they are the Falco of Paolisi and Moiano, whom this archive has never claimed as kin. Giving them a household would be a claim, not a promotion. Measured properly across every research file: 23,741 named rows, of which all but 128 already reach the register and the search index. The harvests are not stranded. But THREE files reached the site not at all — `falco-baptisms-parsed.tsv` (48), `arricale-from-transcripts.tsv` (18) and `falco-parentage.tsv` (22). Eighty-eight rows, invisible and unfindable for as long as they have existed, including the Arienzo parish baptisms whose blank names were row 57's subject. All three are now in `build_register.py`, marked as MACHINE READINGS of mangled Latin so nobody mistakes a parsed spelling for a transcription.

Moiano is the second stronghold of both Zampiello (104 records) and Annecchino (215). Paolisi is where Francesco Falco came from in 1714, and where the Giovanna Falco found this week married a Lettieri. Neither town has ever been opened. COUNTED, 15 September: `data/antenati-valle-caudina.tsv` holds 406 Falco at Paolisi and 310 at Moiano, plus 215 Annecchino and 104 Zampiello at Moiano. All of them reach the register and are searchable; none is claimed as kin. The row is not about making them reachable — it is about deciding whether any of them belongs to this family, which is a reading job on 1,087 rows and has never been started.

Two things sit in it. Matteo Falco's own marriage, which would name his father and break the line's frontier; and the Maria Falco × Francesco Cioffi marriage, which the civil series cannot hold because the Arienzo civil registers begin in 1809 and that marriage is older than they are.

DONE 17 Sep 2026 — and the two failed passes failed because this row had the wrong leaf. «*[Anno] D[omi]ni 1818, die vero septima 7 Junii, R[everendu]s D. BLASIUS MANSIONARIUS ARRICALE Substitutus baptizavit infantem natam ex JOSEPHO FALCO et ANTONIA DI GUIDA conjugibus hujus Parochiae, cui dedit nomen MAGDALENA. Matrina fuit DOROTHEA D'ADDIO. HYACINTHUS Archip. MORGILLO*» Same day as her civil birth act, and the godmother is the same Dorotea d'Addio who stood at Gioacchino Falco's font in 1779 — one woman at this family's baptisms across forty years. THE ARK IN THIS ROW IS THE INDEX LEAF, NOT THE ACT. `3Q9M-CSD3-1381-R` carries the margin label «Maddalena Falco», which is what both rendering passes went hunting and found. The act is on `3Q9M-CSD3-13LG-7`. In this parish register the year's index and the year's acts are on different leaves, and a search that lands on the label will never reach the entry however many tiles it renders. Found by full-text rather than by rendering, with the quoted phrase «Maddalena Falco» — two hits in the whole baptism collection, one the label and one the act. Machine transcription; the image is not yet read.

Both were found here and neither is peculiar to this archive. First: the person-graph builder cleans names to keep old URLs alive, and that cleaning strips brackets before comparing them — which cost this archive an entire generation of its own direct line, because `line.json` spelt a wife «Adriana (Andreana) Crisci» and every act spells her ANDREANA. Second: the search index indexed `404.astro`, which Astro emits as `404.html` and never as a page, so the archive check failed on a row pointing at a page that is never built. Any archive sharing these tools has both. A THIRD, FOUND 15 SEPTEMBER: `WorkList.astro`'s markdown renderer drops code spans and mis-pairs bold. Its `md()` handles `&`, `<`, links, `bold` and «guillemets» — but not backticks, so every ark number in a note (`an_ua14362`) renders with its backticks showing. This page has forty-six of them. **And its bold is `\*\*([^*]+)\*\*` — the greedy no-nested-italic form. On «San Felice a Cancello, *Matrimoni* 1844», the common shape in these archives, it mis-pairs across the inner italic and nests the tags wrongly. Both are fixed in this archive's own `site/src/lib/md.js` — non-greedy `[\s\S]+?` bold, then italics, then code — and the fix is four lines. Every sibling archive using WorkList with ark numbers in its notes has this. 17 September — MEASURED, TESTED, AND ONE SENTENCE ABOVE IS WRONG. The blast radius, on this page. Of 70 rows, 42 contain backticks — 148 code spans, and the built page renders 296 literal backtick characters and not one `<code>` tag. There are also 57 single-asterisk italics, and the kit's renderer has no italic rule at all, so every one of those shows its asterisks too. The second bug is worse than this row said.** This row claimed the greedy `\*\*([^*]+)\*\*` «*mis-pairs across the inner italic and nests the tags wrongly*». It does not mis-pair. It cannot match at all, because `[^*]` refuses to cross the inner `*` — so the whole span renders as literal asterisks and the bold simply does not happen. Three spans on this page are in that state. Run against both renderers rather than reasoned about: ``` IN ANTENATI `an_ua14362`, 55 images kit ANTENATI `an_ua14362`, 55 images new ANTENATI <code>an_ua14362</code>, 55 images IN **San Felice a Cancello, *Matrimoni* 1844 kit San Felice a Cancello, *Matrimoni* 1844 new <strong>San Felice a Cancello, <em>Matrimoni</em> 1844</strong> ``` THE PATCH, PROVEN.** Replace the bold line in `kit/components/WorkList.astro` and add two rules after it — the order matters, bold before italics before code: ```js .replace(/\*\*([\s\S]+?)\*\*/g, "<strong>$1</strong>") .replace(/\*([^*\n]+)\*/g, "<em>$1</em>") .replace(/`([^`]+)`/g, "<code>$1</code>") ``` That is this archive's `site/src/lib/md.js` exactly, and the three cases above are its test. NOT APPLIED FROM HERE, DELIBERATELY. `@daviddef/archive-kit` is a separate repository pinned by commit in `site/package.json`, another session owns the shared kit, and every one of the seven archives renders its work list through this component. The patch is four lines and it is written out above so whoever owns the kit can apply it without rediscovering any of this.

ONE SIGN-IN UNBLOCKS THIS, and the reason changed on 16 September. The note used to say the blocker was the desktop Browser pane not being displayed. It is not that. FindMyPast signs in fine in the browser context this session drives — twelve surnames were swept through it today. FamilySearch specifically has no session there: `/en/tree/find` redirects to `/en/search/tree/`, `/en/home/` to `/en/australia/`, and every image ark bounces to `ident.familysearch.org`. Those are the signed-out paths. A tab is parked on the target ark; signing in *in that tab* — not a new window — puts the session where this session can use it. Credentials are David's to enter and will not be entered here. THE TARGET: `3:1:3Q9M-CSD3-13CP-B`, whose snippet opens «*Raphael Falco Obiit die 26 Februarii 18[8]1*». RAFFAELE FALCO, GENERATION FOUR, HAS NO DEATH IN THIS ARCHIVE. Born 10 May 1818, he would be sixty-two in February 1881, and this is the only candidate ever seen. The OCR gives the year as «1811» twice while the harvest's own metadata says 1881 — and eight and one are the two digits this engine confuses most. If 1811, a stranger born before generation four. If 1881, very likely him. One image decides it. Also waiting on the same sign-in: row 5 (Sara Falco's parish baptism, folio 47 of `3Q9M-CSD3-1325-9`), row 16 (Maddalena Falco's baptism, `3Q9M-CSD3-1381-R`), and 43 parish-baptism snippets whose child's name falls outside the 260-character window. --- ## READ — 27 September 2026. THE YEAR IS 1811, AND THE HOPED-FOR READING IS DEAD David signed in to FamilySearch and the ark opened. `3:1:3Q9M-CSD3-13CP-B` was pulled at 3,840 px and read. > «*Anno Dni 1811, die vero 16 Februarii, R[everen]dus Marsius d'Simeon d'Ambrosio Subdiaconus baptizavit infantem natum ex JOSEPHO FALCO, et ANTONIA DI GUIDA conjugibus hujus Parochiae, cui dedit nomen RAPHAEL. Matrina fuit SARA RUGGIERO. Hyacinthus Archip[resbyte]r Morgillo*» — margin: «*Raphael Falco. obiit die 26 Februarii 1811*». THE OCR WAS RIGHT AND THE HARVEST'S METADATA WAS WRONG. This row hoped for 1881, which would have made it RAFFAELE FALCO OF GENERATION FOUR, born 1818 and dead at sixty-two with no death in this archive. It is 1811. *Generation four's Raffaele still has no death, and this candidate is spent — it must not be chased again.* → [[wrong-negatives]] WHAT IT IS INSTEAD IS A RECORD THIS ARCHIVE DID NOT HOLD. The baptism of the infant Raffaele of Giuseppe Falco and Antonia di Guida, whose civil birth (16 Feb 1811) and civil death (26 Feb 1811) were already read from the image. Born, baptised and buried inside eleven days, baptised the same day he was born. New: the godmother SARA RUGGIERO. AND THE METHOD IS NOW REUSABLE. The tiles are auth-protected — a plain request returns 401 — but with the browser session they resolve at `…/deepzoomcloud/dz/v1/3:1:<ark>/image_files/<level>/<col>_<row>.jpg`, level 12, a 15×11 grid, 3,840 × 2,816. *Two further arks were read the same hour on the same method.* Rows 5 and 16 and the forty-three parish snippets are no longer blocked by anything but time.

J25 1971/9830 and J25 1972/4964 — Brisbane immigration case files, paper only, not digitised. Files of this era normally carry the sponsor, the addresses, the occupations and often photographs, and 1972/4964 is already marked Open. There are also three NAA passenger images whose transcripts give a name, a year and nothing else. Anyone can order copies.

Several Caserta record images return *«This record can only be displayed on certain accounts»* — the index and machine transcripts are open, the scans are not. A visit to a FamilySearch centre or an affiliate library unlocks them. Separately, the Museo Alfonsiano stands in the old bishop's palace beside Sant'Andrea, the church this family was baptised in for two centuries, and may hold or know of parish material that was never filmed. That needs a letter or a visit.

Arpaia births 1895 and Forchia births 1897 are not digitised — first inferred from ark arithmetic, since the shelf-marks run consecutively and skip those years, and since confirmed. They are what still blocks Filomena Annecchino's birth. Ordering or reading them at the Archivio di Stato di Benevento is the only route.

Marriage supplement files: birth extracts, parental consents, death certificates of parents already dead. One file can carry two generations backwards. None have been read. This is the largest unworked seam in the project and it needs time rather than a decision. Re-counted 15 September: 378 volumes across SIX towns, not 258 across four. San Felice a Cancello was not surveyed until 13 September and carries 57 on its own; Arienzo holds 43 with zero name-indexed records. Three towns are still completely unexamined — Moiano, Paolisi and San Felice a Cancello, 199 volumes between them — and this row now covers exactly those three, so retiring one lowers the backlog by one. --- ## 21 September 2026 — THE ARIENZO SERIES IS NOT BLOCKED, AND THIS ROW'S COUNT FOR IT IS WRONG This row says Arienzo holds 43 volumes. It holds 63, and `data/antenati-processetti.tsv` has said 63 since the survey was made — the two have disagreed without anyone comparing them. And they are on Antenati, free, and now reachable by ark: `an_ua14149` (1809) running to `an_ua14211` (1865), one per year with 1855 and 1858 doubled, `an_ua14212` catalogued but not digitised. Two bundles were read out today — 1860 for work-list row 77, which gave four generations of Annecchino, and 1845 for row 73, which proved an identification by act number. Both of those rows opened with the word «order». Neither needed to. THE STATE AND OWNER OF THIS ROW ARE LEFT ALONE, ON PURPOSE. What it describes is genuinely large — three towns and 199 volumes still completely unexamined — and the blocker for those is time, which is David's to allocate. But for Arienzo the blocker was never access; it was that the archive held a count instead of an ark, and a count cannot be opened. *If this row is re-scoped, the Arienzo 63 belong on this archive's side of the line.*

Partly unblocked on 16 September, and the row was overstating it. MyHeritage is NOT blocked — this archive already harvests it: `data/myheritage-facts.tsv` holds 427 rows, and the family tree itself lives there. FindMyPast is now open on a one-month subscription David started on 16 September; that work is row 65, running. What is genuinely left is ANCESTRY, for the Brisbane generations' passenger lists and naturalisation files. No subscription, and nothing here can move it. Before buying one, read row 65's result: FindMyPast's Australian holding for this family turned out to be six passenger records and the Italian sets are the FamilySearch index re-hosted. Ancestry may be the same trade twice, and the thing actually worth paying for is probably the NAA files in row 19, which anyone can order for a copying fee.

Matches could confirm or break the Annecchino descent, and could find living cousins in Campania. No register can do that. The family already has tests in play; the matches need working through.

A decision, recorded so nobody spends a day rediscovering it. This family has no English and no Habsburg generations. All three are wrong for it, and saying so on What has been read is worth more than searching them would have been.

Founded on a mistake, kept rather than deleted. A day went into reading the 1816 and 1817 birth registers and into deriving the trap that these forms carry a *printed* year the clerk strikes and corrects by hand. Both were already written up in `notes/1817-birth-register-progress.md`, which says *«Result: no Falco at all»* and, in its second line, *«every act is pre-printed sedici and corrected by hand to diciassette»*. The negatives stand and were already this archive's negatives. The standing rule now is to read `notes/QUEUE.md` and the 83 per-year progress files before opening any register.

STRUCK — founded on a mistake, and kept here rather than deleted. The audit read the corrections page's prose, saw the retired surname «Cossi», and concluded forty volumes had been searched for the wrong man. They were searched for the bride. The *What has been read* record lists, year by year, where no FALCO married at Arienzo, and reports thirteen Falco marriages found — a reader in the F section of a tavola sees a Maria Falco whoever she marries. The negative stands and the marriage is pre-1809, as `notes/QUEUE.md` already said, with the surname already correct. What was real was a wrong name in the prose, now fixed.

Done. CIOFFI. Arienzo *Morti 1843* atto 43, read at magnification: «*LUCA CIOFFI di anni trentatré, Sartore … e GIOVANNI CIOFFI di anni quarantatré, contadino*», both of strada Camellara. The *ff* ligature — two tall ascenders and a long descender — settles it, exactly as the withdrawn correction's own method note described before reading it backwards. The withdrawn correction had the AGES right and the household file did not (it carried 48 for Giovanni, the act says 43). Two records of one act, each holding half of it, disagreeing for two days. Both now fixed.

192 entries took one pass and produced five findings. The page is appended to daily and read back never — which is how nineteen entries came to be standing after the finding that killed them. Recurring, not one-off, because the failure is structural: the page grows at the top and nobody walks down it. ## THE SECOND PASS — 326 ENTRIES, THREE FINDINGS — 21 September 2026 Walked from the top, looking for entries a later finding has killed or answered. Three, and they are the three shapes this row predicted: 1. AN ENTRY THAT WAS ANSWERED AND NEVER TOLD. «*Sara Ruggiero stood godmother to six Falco households*» (15 September) ends on «*That is not a neighbour*» and leaves the reach unexplained. The 1816 birth register, read today, calls her a LEVATRICE. A midwife stands godmother to the children she delivers. The entry needed an answer, not a withdrawal, and now has one. 2. AN ENTRY WHOSE EVIDENCE IS FALSE WHILE ITS CONCLUSION HOLDS. «*The elder Giovanni Falco is not in the birth registers of 1816 or 1817*» asserts «not one Falco in the entire year» for 1817 and «one Falco act only» for 1816. 1817 holds two and 1816 holds two, and one of each was cited in this archive's own household file on the day the entry was written. The conclusion — no Giovanni — survives; the argument «*a year that empty is itself a fact about the volume*» does not. 3. AN ENTRY A LATER ACT COMPLICATES. The Cospi rename of this morning now has an 1818 act, read in two hands, spelling the man COSSI. The name is not changed again — the 1821 leaf says Cospi in four places — but the corrections page now carries the complication instead of only the household file. All three were appended to, not rewritten. *The next audit is due at 500 entries; the page is growing at roughly twenty a day.*

Done: `tools/isitnew.py`. One command, every corpus, before the write-up — corrections, work list, both household files, people, register, the searched page and all eighty-odd progress notes. Each argument is searched independently and it prints where the term already appears. Tested against the failure it exists for. `python3 tools/isitnew.py "FALCO CARMELA"` returns hits in the corrections, the household file, people AND the register — the 1859 duplication would have been stopped before a word of it was written. Deliberately not a build gate. It cannot know whether a hit means the finding is old; it is a lookup, not a judgement. What it removes is the excuse: a hit is not proof the finding is old and a miss is not proof it is new — it means nobody can say they did not look.

A birth act can carry a marriage, and this archive only learned that today. Act 33 of 1819 has the registrar's later note across its foot: «*Con atto 6 agosto 1856 atto 21 in S. Maria a Vico il Falco sposò Domenica Vigliotti*» — a wedding in another comune, recorded nowhere an Arienzo marriage register would show it. Work-list row 7 closed with Nunziata Falco absent from the Arienzo marriage registers of 1854, 1855 and 1856, all read act by act. That absence now has a second possible explanation and a way to test it. First find her birth act, then read its margin. PROMOTED by the pattern found on 15 September. The annotation practice appears to begin between 1845 and 1856 — absent on marriages of 1844 and 1845, present on 1856 and 1880. Nunziata married on 13 November 1855, which is exactly on that boundary. Her margin is now the most informative one in the archive either way: it either gives her marriage, or it dates the practice to after November 1855. A SECOND ROUTE, 16 September. The marriage volumes carry a NOTIFICAZIONI register bound in after the acts — banns, published where the couple lives even when the wedding is celebrated elsewhere. That is this row's exact shape. See the new banns row. ## 21 September — BOTH CHEAP ROUTES ARE CLOSED, AND THE ROW IS NOW HONESTLY PRICED The NOTIFICAZIONI route is gone for this year. `an_ua14259` (Arienzo, Matrimoni 1855) holds thirty images and the twenty-eight acts fill images 2–30. Nothing is bound in after them. `tools/antenati_holdings.py Arienzo` confirms the town has only Nati and Matrimoni series on Antenati — 58 years each, 1703 and 1809–1865 — with no separate notificazioni series to look in. And the TAVOLA route is gone for the decade she was born in. The birth volumes of the 1830s do not carry a usable index: 1831 has no tavola at all (`an_ua14428` ends with act 246, two blanks and the Procuratore's certificate) and 1836's tavola is printed and left entirely blank (`an_ua14433` img 127 right, ruled columns, every cell empty). This archive has read tavole for 1824, 1825 and 1828; they stop. On /searched as a null. WHAT THE ROW ACTUALLY COSTS NOW. Her window is 1820–1823, 1829, 1831–1833 and 1835–1838 — the years between her documented siblings that no birth register has been read for. Each volume is 100 to 150 images, swept at the declarant band for VINCENZO FALCO as father, which is the method row 63 used on the whole of `an_ua14428`. Call it eleven volumes and about 1,300 images. There is no shortcut and the row should stop implying there is one. The two likeliest years first. Her siblings run 1816, 1818, 1819, 1824, 1826, 1828, 1830, 1834 — roughly two-year intervals — and Andreana Crisci, born 30 November 1796, was forty in 1836. 1832 and 1836 are where a child fits the pattern, and a bride of nineteen or twenty-three in November 1855 fits either. ## 1836 IS READ OUT AND SHE IS NOT IN IT — 21 September 2026 `an_ua14433`, 128 images, 125 acts, every one swept at the declarant band. NO FALCO. Vincenzo Falco fathered no child registered at Arienzo in 1836, and neither did any other Falco. One of the two likeliest years is spent. The method is cheaper than this row feared and should be stated for the rest. These volumes run one act per opening, on the RIGHT page, so sweeping the right halves alone reads the whole year — five contact sheets for 128 images. The band at 0.36–0.51 of the page height carries the declarant's name, age, trade and street in one strip, which is enough to reject an act without opening it. The row's estimate of «1,300 images» stands, but it is about fifty contact sheets, not 1,300 reads. Remaining: 1820–1823, 1829, 1831–1833, 1835, 1837, 1838 — with 1832 the other year where a child fits the sibling spacing. ## 1832 IS READ OUT TOO — AND BOTH BEST YEARS ARE SPENT `an_ua14429`, 117 images, every act at the parents band. NO NUNZIATA. Vincenzo Falco and Andreana Crisci have no child registered at Arienzo in 1832, as they have none in 1836. The two years where a child fits their sibling spacing are both gone. THE 1836 SWEEP HAD TO BE DONE TWICE AND THE FIRST ATTEMPT IS ON /corrections. These acts begin on a right page and the parents are named at the top of the NEXT LEFT page; the right page names only the declarant, who in 1832 is very often the LEVATRICE. A declarant sweep silently drops every father whose child was presented by the midwife, and it produces a clean, complete-looking, wrong negative. Use band 0.02–0.21, half L, three tiles at 2,400 pixels. And the tavola is not the shortcut: 1831 has none, 1836's is printed and blank, and 1832's is filled and stops at C — the same shape as the 1834 death index, which is «a single leaf covering A to C and marked END». An Arienzo index that stops at C is a pattern. Remaining: 1820–1823, 1829, 1831, 1833, 1835, 1837, 1838 — seven volumes at about seven contact sheets each. The year paid for itself anyway. Three Falco births, one of them the first child ever recorded for Aniello Crisci and Giovanna Falco, and one — MARIA ROSA FALCO of MICHELANGELO Falco — that bears directly on the Michele/Michelangelo distinction drawn on 20 September. Both on /corrections. ## 1835 TOO — AND THE GROOM IS THE CHEAPER QUESTION `an_ua14432`, 118 images, read out at the parents band. No Falco, no Crisci. No tavola either. Three years spent: 1832, 1835, 1836. Remaining: 1820–1823, 1829, 1831, 1833, 1837, 1838 — nine volumes, about sixty contact sheets. AND THIS ROW HAS BEEN ASKING THE EXPENSIVE QUESTION. This archive holds Nunziata's marriage of 13 November 1855 from a machine transcription that captured only her side — «*Nuntiatam Falco filiam Vincentii, et Andreanae Crisci*» — and has never recorded whom she married. The groom's name is one image away, and it would open a whole household to search instead of nine registers. It is blocked, and not by credentials. A logged-in FamilySearch session is open in the browser pane; the viewer loads with its signed-in controls. What is missing is a DISPLAYED pane — screenshots return «*the Browser pane is not displayed, so the page is not compositing frames*», and this archive's FamilySearch reading method is a render-and-screenshot loop. On /searched, with the three things the attempt did establish: the DeepZoom image is 2394 × 1831, `/das/v2/<ark>/name?namespace=apid` gives the internal asset name, and there is no full-text-by-ark endpoint — only the search, whose results carry the whole page's `textDocument`. ## 1837 TOO — FOUR YEARS SPENT, AND THE TAVOLA QUESTION IS SETTLED `an_ua14434`, 146 images. No Nunziata — and one Falco birth: RAFFAELE FALCO, act 108, 31 August 1837 at six in the morning, to FRANCESCANTONIO FALCO, thirty-one, *contadino* of strada Camellara, and RAFFAELA FALCO, twenty-six, baptised at Sant'Andrea the same day. A sixth child for the first-cousin marriage. Years read out for this couple: 1832, 1835, 1836, 1837. Remaining: 1820–1823, 1829, 1831, 1833, 1838. AND THE TAVOLA IS NOW SETTLED AS USELESS FOR THIS FAMILY. 1837's exists, is filled, and ends at C — image 145 is a duplicate of 144 with the Archivio's own conservation slip lying across the page, and 146 closes the volume. With 1832's doing the same, 1831 having none and 1836's blank — and the 1834 death index being «a single leaf covering A to C and marked END» — an Arienzo index of this period that reaches C has not reached F. For a family called FALCO that is the whole of the difference, and it is why these years cost a sweep each. ## 1838 — FIVE YEARS SPENT, AND A NIL THAT DATES SOMEBODY ELSE `an_ua14435`, 120 images. No Falco at all, no Crisci. Years read out at the parents band: 1832, 1835, 1836, 1837, 1838. Remaining: 1820–1823, 1829, 1831, 1833. The 1838 nil is worth more than the search that produced it. This archive holds PASQUALE FALCO, dead 30 October 1838, a child of the first-cousin marriage — and he is not in that year's birth register, so he was not born in the year he died. His brother Raffaele was born 31 August 1837. And two Crisci mothers appear, neither ours: ALESSANDRA CRISCI at image 4 — a dated sighting of a woman whose parentage is an open question here — and an ANNA CRISCI of strada Capodiconca, married to Simone Martone, at image 26. ## 1833 AND 1831 — SEVEN YEARS SPENT, AND THE RE-READ PROVED THE METHOD Years read out at the parents band: 1831, 1832, 1833, 1835, 1836, 1837, 1838. No Nunziata in any of them. Remaining: 1820, 1821, 1822, 1823, 1829. Two more children for FRANCESCO CIOFFI and MARIA FALCO of strada Camellara — a household that had buried five: ANTONIA, 8 August 1831 (act 71) and CLEMENTE, 23 August 1833 (act 85), both baptised at Sant'Andrea within days. THE 1831 RE-READ IS THE ARGUMENT FOR THE BAND, MADE ON THE ONE VOLUME THAT HAD BEEN DONE THE OTHER WAY. This volume was swept for row 63 on the declarant line. The declarant of act 71 is the father, a CIOFFI — so a declarant sweep sees no Falco in that act at all; the Falco is the mother, on the facing leaf. The parents band found her. *Every volume swept on the declarant line is owed a re-read.* And the 1831 act settles the surname for a third time: *C-i-o-ff-i*, plainly, with the i between the C and the o. The 1833 act of the same household is in a worse hand where the two cannot be separated, and it was left alone. ## 1829 — EIGHT YEARS SPENT, FOUR LEFT `an_ua14426`, 114 images. No Nunziata — and a third child for Francesco Cioffi and Maria Falco: GIUSEPPE, 5 October 1829, act 74, who is probably the Giuseppe Cioffi «the second» dead 5 December 1830 at fourteen months. Not claimed. Read out at the parents band: 1829, 1831, 1832, 1833, 1835, 1836, 1837, 1838. REMAINING: 1820, 1821, 1822, 1823. The contact sheet lied twice in this one volume. Image 18 read «Maria Falco» at tile size and the act says «MARIA DELLE FAVE»; image 77 read «Rosa Falco» and the act says «MARIA FALCO». Every candidate was opened, which is the only reason the first did not enter the archive and the second did not enter it wrong. ## 1821 — NINE YEARS SPENT, THREE LEFT `an_ua14418`, 117 images. No Nunziata — and ANNA MARIA TERESA COSPI, 16 August 1821, act 77, the first child, first address and first age for a couple this archive knew from a death act and one index line. It also renamed the household: Domenico COSPI, not Crisci, on a third document. On /corrections. Read out: 1821, 1829, 1831, 1832, 1833, 1835, 1836, 1837, 1838. REMAINING: 1820, 1822, 1823. ## 1822 — TEN YEARS SPENT, AND THE SWEEP PAID FOR ITSELF `an_ua14419`, 141 images. No Nunziata — and three Falco births, one of them the best thing found in this row: ANGELA MARIA FALCO, 19 January 1822, act 8, daughter of «GIUSEPPE FALCO FU MATTEO, forty, *MASSARO*, of strada Camellara» and ANTONIA DI GUIDA, forty — a granddaughter of the founder, in an act that names him, filling the gap between Crescenzo (1819) and Clemente (1824). On /corrections. SALVATORE DIGLIO, 3 September, act 78, to Alfonso Diglio and Maria Falco, thirty-one. ANTONIA CIOFFI, 30 September, act 89, to Francesco Cioffi and Maria Falco, twenty-four — two Maria Falco of one street, twenty-seven days apart, separated by their husbands. The second dates the Antonia Cioffi who died 20 June 1823. Read out: 1821, 1822, 1829, 1831, 1832, 1833, 1835, 1836, 1837, 1838. REMAINING: 1820, 1823. --- # ALL TWELVE YEARS ARE READ OUT, AND SHE IS IN NONE OF THEM 1820, 1821, 1822, 1823, 1829, 1831, 1832, 1833, 1835, 1836, 1837, 1838 — every act of every year, read at the band that names both parents. About 1,450 images. VINCENZO FALCO and ANDREANA CRISCI have no child registered at Arienzo in any of the years between their documented ones. So Nunziata Falco was not born in the gaps. What is left, and this row is closed pointing at it: - She was born in a year this archive has already read and the sweep that read it was looking for something else — the 1824–1828 and 1830 and 1834 volumes were read for *other* children. - Or she was not born at Arienzo, which would make her the only one of her siblings who was not. - Or the marriage of 13 November 1855 is not hers. This archive has her from a machine transcription of a parish act and has never recorded whom she married. That is now the cheapest question and the row that should follow this one: the groom's name is one image away, blocked only by a browser pane that will not display. ## WHAT THE TWELVE YEARS DID GIVE — nine children and a generation named ANGELA MARIA FALCO, 19 January 1822 — a granddaughter of the founder, in an act that calls her father «GIUSEPPE FALCO FU MATTEO», *massaro*, of strada Camellara. FORTUNATO CRISCI, 22 January 1820, who lived two days — and whose act gives ARCANGELO CRISCI as fifty, agreeing to the year with the 1814 act's forty-four. Three children for FRANCESCO CIOFFI and MARIA FALCO of strada Camellara — Antonia (1822), Giuseppe (1829), Antonia the second (1831), Clemente (1833) — a household this archive had only ever watched bury its children. Two of the four date deaths it already held. ANNA MARIA TERESA COSPI, 16 August 1821, the first child, first address and first age for a couple known from a death act — and it renamed the household. MARIA CARMELA FALCO (1832) and RAFFAELE FALCO (1837) for the first-cousin marriage; MARIA ROSA FALCO (1832) for MICHELANGELO Falco, bearing on the Michele/Michelangelo distinction; PASQUALINA CRISCI (1832), the first child ever recorded for Aniello Crisci and Giovanna Falco; SALVATORE DIGLIO (1822); and GIUDITTA (1820) and ANGELA MORGILLO (1823) for Carlantonio Morgillo and Rosa Falco. ## POSTSCRIPT, 21 September 2026 — 1832 WAS NOT SPENT This row argued from sibling spacing that 1832 and 1836 were the two years where a child of Vincenzo Falco and Andreana Crisci would fit, read both, and reported both empty. The declarant-band re-sweep of work-list row 85 has since found MARIA CARMELA FALCO, act 84 of 1832, in the volume this row read — invisible to it because the left page says «*e da lui dichiarante*» and never writes the father's name. The row's own conclusion is unchanged and if anything firmer: the child in the 1832 gap is not Nunziata. What changes is the argument that got here — the gap is now occupied, so sibling spacing no longer points anywhere, and Nunziata's birth remains in no Arienzo register this archive has read.

Done, and the answer is PROVEN. Arienzo *Matrimoni* 1845, act 3: «*CRESCENZO FALCO maggiore di anni venticinque, nato in Arienzo … figlio di Giuseppe Falco e di Antonia di Guida*», promise 2 February, church 3 February at Sant'Andrea Apostolo. Twenty-five on 2 February 1845 brackets his birth to 3 Feb 1819 – 2 Feb 1820, and the civil act of 11 October 1819 sits inside it. Three sources, one man. The row paid for itself twice. The same act gives ANTONIA DI GUIDA alive on 2 February 1845 — twenty-five years later than the only other dated document this archive has for her, on a form that marks the bride's two dead parents explicitly. Her death window moves accordingly.

Done. David said take it. The bar now plots the group's own `reachCount` instead of `n`, so a family filed under a reach is measured by that reach: Crisci 10, Falcone 9, Biondo 7, Militano 6, Ellevsen 4, D'Arcy 3, Zavaglia 3, Defranceski 2, Mazza 2, and one each for Zaino, Abatiello, Halliday, Isla, Reid — 51, and every living branch is now visible instead of drawn as zero. The unit label «*people this archive names*» went with it: that was the honest name for the wrong number. Committed together with the Falco session's uncommitted chart work, which these three files were: the shared `marriedin.js` rule, the chart section on the front page and the marriages-page refactor onto the same lib. The fix could not be separated from it.

DONE — sixteen years, 1850 to 1865, every act read. The series runs to `an_ua14269`, not to 1858; that earlier claim was four rate-limit refusals read as absences and is withdrawn. TWO FALCO MARRIAGES IN SIXTEEN YEARS. 1850 act 6: MICHELE FALCO, *vedovo di Antonia Viglioti*, marrying Giovanna d'Ambrogio — merging two households and naming his parents. 1861 act 16: STANISLAO FALCO, 24, PASTORE, son of Vincenzo Falco and Andreana Crisci — a child of generation three this archive did not have — marrying Filomena di Lucia. Nil years: 1851, 1852 (two bound registers), 1853, 1857, 1858, 1859, 1860, 1862, 1863, 1864, 1865. 1854–1856 were read earlier and gave three Falco weddings. And the year's best find was not a Falco: 1860 act 3 is the ANNECCHINO–CARFORA marriage that gives that line a generation and puts the Carfora at Arienzo. What remains from this row is the banns. The 1861 volume binds a notificazioni register after its acts; see the banns row.

Row 9 names 1818–1820, 1829 onward and 1845–1860. These three fall between and are named nowhere. 1822, 1823, 1824, 1825 and 1828 are read; 1821 (`an_ua14418`), 1826 (`an_ua14423`) and 1827 (`an_ua14424`) are not. Generation three was bearing children across the whole span — Luigi 1819, then Giuseppe Nunzio and Clemente in 1824, Angelandrea 1826, Sara 1828 — so a gap year is a missing child, not a quiet one. Angelandrea's own act of 12 September 1826 was found from elsewhere, which means the 1826 volume has been opened at one page and never read. ## 1821 IS OFF THIS ROW — 21 September 2026 1821 (`an_ua14418`) has now been read twice, at the parents band and at the declarant band, and it is on /searched with both readings. It gave a child — ROSARIA FALCO, 15 March 1821 — and a corrected household name. The row's remaining volumes are 1826 (`an_ua14423`, 116 images) and 1827 (`an_ua14424`, 157 images), and they are being read now, at BOTH bands from the start. And this row's premise needs a correction of its own. It says «*1822, 1823, 1824, 1825 and 1828 are read*». 1822 and 1823 are, and say so on /searched. 1824, 1825 and 1828 have no coverage row at all — what this archive has from them is single acts cited in households (Giuseppe Nunzio, atto 31 of 1824; Sara, atto 118 of 1828), which is a volume *opened*, not a volume *read*. 1819 and 1830 are in the same state. See row 86. ## 1826 AND 1827 READ OUT — AND WITH THEM 1824 AND 1825 — 21 September 2026 All four volumes are now on /searched, read end to end at both bands. What they gave: - 1826: GIUSEPPE CIOFFI, 9 March, to Francesco Cioffi and MARIA FALCO of strada Sant'Andrea — new, and it dates a death act this archive already held. Angelandrea Falco's act confirmed. A second ANDREANA CRISCI, thirty-one, of strada Capodiconca. - 1827: MARIA CARMELA MORGILLO, 20 December, to Carlantonio Morgillo and ROSA FALCO — a granddaughter of the founders, new, and it dates the household's move to strada Camellara. Maddalena Falco's act read at last, and its number corrected from 52 to 57. - 1824: three Falco, all already held. Its tavola is under the archivist's card, as 1823's is. - 1825: two Falco, both already held; Clementina Diglio's act number corrected from 72 to 62. THIS ROW IS DONE. What remains of the birth series is on row 86: 1819, 1828, 1830 have no coverage row at all, and 1816, 1817, 1818 have been read at the parents band only.

DONE for the window it names. 1845, 1846, 1847 and 1848 all read — there is no Falco marriage in the 1849 index, so 1845–1849 is finished. What the registers gave that the indexes could not. ANGELANDREA FALCO is «MINORE» in 1847 and ANDREA FALCO «maggiore, di anni ventiquattro» in 1846 — majority being twenty-one, that settles them as two men four years apart, which the index could never have done. ANDREANA CRISCI is a FILATRICE, her first recorded occupation against five recorded ages. CAMILLA CARBONE gets the forename the Airola household's bracketed name was waiting for. And Vincenzo Falco and Andreana Crisci are named as living in 1846 and 1847, filling the middle of the 1844–1861 gap. Nine of the thirteen index marriages remain, all of them before 1845 — 1813, 1814, 1815, 1816, 1818, 1819, 1821, 1824, 1825, 1827, 1828, 1829, 1832, 1834, 1837, 1839. That is a separate row's worth of volumes, not this one's.

STRUCK — founded on a mistake, and kept rather than deleted. The row existed because I read «S[?]basia a Vico» in a margin as Santa Maria a Vico. That register excludes it: `an_ua55736`, act 21 is Don Domenico de Lucia, promised 14 June 1856, and the year has no act of 6 August at all. The marriage is real and the comune is not established. A Falco married a Domenica Vigliotti on 6 August 1856, act 21, somewhere «…a Vico». Finding where is a live question; it is just not this row's answer.

Done — and it cost one request, now that `tools/antenati_holdings.py` exists. Santa Maria a Vico on Antenati: Nati 1811, 1813–14, 1817–1865 (first ark `an_ua55866`); Matrimoni 1809–11, 1813–14, 1817–1865 (`an_ua55691`); Morti 1809–14, 1817–1864 (`an_ua55780`). 1812, 1815 and 1816 are missing from all three. The year filter resolves a year straight to an ark: `?localita=…&tipologia=…&anno=1856` leaves `+id:"55736"` in the page's Solr comment. That is how `an_ua55736` was found without any arithmetic through the gaps. The row was raised to reach a marriage that turns out not to be here — see the struck row — but the enumeration stands and the comune is now open for any other purpose.

She is alive on 2 February 1845, stated on a form that marks the dead explicitly. That is twenty-five years later than the only other dated document this archive had for her. The existing negative was built when her window opened in 1819: 229 Arienzo pages of the parish death register, searched on «Guida». The Arienzo CIVIL death registers are read act by act for 1844–1858 and 1859 — so if she died at Arienzo before 1860 she should already have been seen, and she was not. That is now a real conflict rather than a gap, and the candidates are: she died 1860–1865 (row 3's volumes), she died after 1865 where Antenati stops, she died at Santa Maria a Vico or another comune, or she was missed in a sweep that was not looking for her. Re-read the 1845–1858 F and G leaves with her name as the target. 16 September: the series is exhausted. 1865 is now read — 147 acts, no Falco. Antenati's Arienzo deaths stop at 1865, and Vincenzo Falco was demonstrably alive in December 1864, which shows the series simply does not reach this generation's deaths. The remaining years of 1860–1863 are still worth reading for other Falco, but the answer for these two is almost certainly the same: 1866 or later, in a series this archive has not located. Locating it is now the higher-value task. 20 September — THE «DIVERSI» REGISTER IS READ OUT, AND THE ANSWER IS NO. Work-list row 70 has now read all 45 volumes, 1812–1865, of the Arienzo register of people who died away from home — the one instrument that could have held a death this town never recorded. Every act from 1845 to 1865 is read and its dead are named, and there is no FALCO and no DI GUIDA among them. The only unnamed acts left in the series are four, all before 1845, so none can be this person. So the third explanation is spent. A death missing from the Arienzo register is not explained here by the person having died elsewhere. What remains is 1866 or later, in the post-Unification registers this archive has still not located — or a comune whose own register was never transmitted to Arienzo at all. --- ## 21 September 2026 — THE MARRIAGE REGISTERS OF 1861–1865 ARE READ, AND THEY DO NOT ANSWER THIS ROW EITHER A child's marriage act names both parents and marks the dead ones, which makes it the instrument this row has been missing: it dates a parent without needing their death act. So the Arienzo marriage registers for 1861, 1862, 1863, 1864 and 1865 were read act by act — five years nobody here had opened, ending at `an_ua14269`, the last one Antenati holds. Two Falco acts in five years. Neither names Giuseppe Falco or Antonia di Guida. 1865 act 3 is ANGELA FALCO, a daughter of MATTEO FALCO and ALESSANDRA CRISCI and new to this archive; 1861's is Stanislao Falco, already held. 1862, 1863 and 1864 hold no Falco at all. WHAT IS STILL WORTH DOING, AND IT IS THE CHEAPEST THING LEFT ON THIS ROW. The Arienzo marriage registers for 1857–1860 — four volumes, `an_ua14261`–`an_ua14264` — have never been swept for Falco. *(1855 and 1856 were read too: both their Falco acts were already held, and the sweep turned up GIACOMO ANNECCHINO, a brother of Sabbatino, instead.)* *(1850 to 1854 were read on 21 September. 1853 holds no Falco; 1854 holds two — Giuseppe Falco's remarriage, already held, and FRANCESCO FALCO's, act 20, which gave this archive a second son of Raffaele Falco and Chiara Rivetti. 1852 holds no Falco; 1850's is Michele Falco's, already held; and 1851's is LUIGI FALCO's, 13 August, act 14 — a marriage in generation three's own household that this archive had only from a death act. Neither Giuseppe Falco nor Antonia di Guida is named in any of the three.)* This archive read the marriage *indexes* for 1809–1849 and then stopped, because the index series itself stops at 1849. Eleven years of marriages, unlooked at, in exactly the window where these two people's children were marrying. One sheet of fourteen acts per two images. --- ## THE WINDOW IS TWENTY-SEVEN MONTHS — 21 September 2026 Two of her daughters married, twenty-seven months apart, and the second act calls her dead. - 15 November 1857, act 26 — ANGELA MARIA FALCO: «*figlia di GIUSEPPE FALCO, di professione contadino, domiciliato ivi, e di ANTONIA GUIDA, domiciliata ivi*». Unmarked. - 14 February 1860, act 8 — MADDALENA FALCO: «*… e di ANTONIA GUIDA DEFUNTA, domiciliata ivi*». ANTONIA DI GUIDA DIED BETWEEN 15 NOVEMBER 1857 AND 14 FEBRUARY 1860. The old window opened in February 1845 and ran to 1865; it was twenty years wide and it is now twenty-seven months. ### And the registers cover every month of it The Arienzo civil deaths are read act by act for 1844–1858 and for 1859. The 1860 volume — `an_ua14362` — had never been opened, so it was, and its acts from 1 January to 5 March 1860 were read out: fourteen dead, every one named. No Falco. No Guida. SO THE CONFLICT THIS ROW WAS RAISED ABOUT IS NOW SHARP INSTEAD OF VAGUE. She died inside a window that is entirely covered, and she is not in it. ### The likeliest explanation, and how to test it in one pass The sweeps were not looking for her, and this register hides a married woman behind her husband. The 1860 death acts enter women as «*MARIA ROSA SARTORE, moglie di GAETANO SASSONE*», «*GIACINTA FERRAJOLO, vedova di ANDREA DI GENNARO*». A pass that reads the first name on the line and moves on will see «Antonia Guida» — but a pass that indexed her under her husband would file her as a FALCO, and a pass hunting FALCO would file her under GUIDA. She can fall between two sweeps that were each complete. Re-read the Arienzo death registers for 1858 and 1859 — `an_ua14360` and `an_ua14361` — at the deceased's own line, for ANTONIA GUIDA, for «moglie di GIUSEPPE FALCO», and for DI GUIDA. Two volumes. *If she is not there either, she died outside Arienzo, and that is a different row.* ### The test was run the same hour, and she is not there The 1858 and 1859 tavole and the 1860 acts to 5 March were read for HER, not for Falco: - 1858 G: «*GUIDA ALFONSO*» (act 1) and «*GUIDA PASQUALE*» (act 62). Neither is her. - 1859 G: Gagliardi, Gesuè, Grasso, Giglio. No Guida at all. - 1860, acts 1–14: fourteen dead, named from the image. No Guida, no Falco. - *And the F sections: 1858 has one FERRIELLO; 1859 has three Falco, all children of a MICHELE, all published here on 14 September.* SHE DIED BETWEEN 15 NOVEMBER 1857 AND 14 FEBRUARY 1860 AND SHE IS NOT IN THE ARIENZO DEATH REGISTER FOR ANY MONTH OF IT. This row's four candidate explanations are down to one. She did not die 1860–1865 — her daughter's act of February 1860 calls her dead. She did not die after 1865. She was not «missed in a sweep», because the sweep has now been run for her own name on the years that matter. What is left is that she died somewhere other than Arienzo — at a married daughter's house, at Santa Maria a Vico where her son Giuseppe's family was living, or in any of the comuni her children married into: San Felice a Cancello, Liveri, Forchia. That is a different search and it should be a different row. *The Arienzo branch of this question is finished.* --- ## THIS ROW WAS STALE, AND IT COST A DUPLICATED SWEEP — 23 September 2026 The paragraph above says the Arienzo marriage registers for 1857–1860 «have never been swept for Falco». They had been — on 21 September 2026, two days before this row was last edited. `site/src/data/searched.json` records all four volumes, `an_ua14261`–`an_ua14264`, «*every act read at the name band*», and states in terms: «*1858 AND 1859 HOLD NO FALCO AT ALL — thirty-nine and twenty-eight images, every act read*». Two files of this archive disagreed about whether four volumes had been opened, and nothing in the nineteen-step chain compares them. On 23 September I believed this row, re-read `an_ua14262` and `an_ua14263` end to end, and reproduced the earlier nil exactly. [[archive-contradicts-itself]] THE NIL IS CONFIRMED AND NOW CARRIES AN ACT COUNT. 1858: images 3–38, THIRTY-SIX acts. 1859: images 3–27, TWENTY-FIVE acts. Sixty-one acts — no FALCO as groom, as bride, or as a parent of either, which is the reading this row needs, because a child's marriage act is what dates a parent. So neither Giuseppe Falco nor Antonia di Guida is named in 1858 or 1859, and no window moves. AND THE EARLIER SWEEP'S BAND IS NOW VINDICATED. It recorded «read at the name band» without saying which half — and this archive's standing hazard is that a band on one half reads half the acts *consistently*, so nothing looks missing. In these volumes an act BEGINS on the RIGHT page, which carries the date, both spouses and all four parents; the LEFT page of the next image holds the *previous* act's witnesses. `image = act + 2`. The wrong half holds no parents at all, so a nil from it would have been obvious — the earlier reading stands. [[half-page-sweeps]] THE RULE THIS PRODUCED. *Before sweeping any volume, grep `searched.json` for its ark.* The work list records what someone thought still needed reading; the coverage data records what has actually been read, and they drift. One `grep an_ua14262 site/src/data/searched.json` would have saved the pass.

Alive on 2 February 1845 and again on 26 November 1849, when his son the deacon died and the act named him without «fu», a *colono*. He would be about seventy. The Arienzo civil death registers are read for 1850–1858 and 1859 and he is not in them. Same shape as his wife's problem, and probably the same answer — so test them together. He is the founder's son and one of the best-documented men in the archive; his death should not be missing. 16 September: the series is exhausted. 1865 is now read — 147 acts, no Falco. Antenati's Arienzo deaths stop at 1865, and Vincenzo Falco was demonstrably alive in December 1864, which shows the series simply does not reach this generation's deaths. The remaining years of 1860–1863 are still worth reading for other Falco, but the answer for these two is almost certainly the same: 1866 or later, in a series this archive has not located. Locating it is now the higher-value task. 20 September — THE «DIVERSI» REGISTER IS READ OUT, AND THE ANSWER IS NO. Work-list row 70 has now read all 45 volumes, 1812–1865, of the Arienzo register of people who died away from home — the one instrument that could have held a death this town never recorded. Every act from 1845 to 1865 is read and its dead are named, and there is no FALCO and no DI GUIDA among them. The only unnamed acts left in the series are four, all before 1845, so none can be this person. So the third explanation is spent. A death missing from the Arienzo register is not explained here by the person having died elsewhere. What remains is 1866 or later, in the post-Unification registers this archive has still not located — or a comune whose own register was never transmitted to Arienzo at all. --- ## 21 September 2026 — THE MARRIAGE REGISTERS OF 1861–1865 ARE READ, AND THEY DO NOT ANSWER THIS ROW EITHER A child's marriage act names both parents and marks the dead ones, which makes it the instrument this row has been missing: it dates a parent without needing their death act. So the Arienzo marriage registers for 1861, 1862, 1863, 1864 and 1865 were read act by act — five years nobody here had opened, ending at `an_ua14269`, the last one Antenati holds. Two Falco acts in five years. Neither names Giuseppe Falco or Antonia di Guida. 1865 act 3 is ANGELA FALCO, a daughter of MATTEO FALCO and ALESSANDRA CRISCI and new to this archive; 1861's is Stanislao Falco, already held. 1862, 1863 and 1864 hold no Falco at all. WHAT IS STILL WORTH DOING, AND IT IS THE CHEAPEST THING LEFT ON THIS ROW. The Arienzo marriage registers for 1857–1860 — four volumes, `an_ua14261`–`an_ua14264` — have never been swept for Falco. *(1855 and 1856 were read too: both their Falco acts were already held, and the sweep turned up GIACOMO ANNECCHINO, a brother of Sabbatino, instead.)* *(1850 to 1854 were read on 21 September. 1853 holds no Falco; 1854 holds two — Giuseppe Falco's remarriage, already held, and FRANCESCO FALCO's, act 20, which gave this archive a second son of Raffaele Falco and Chiara Rivetti. 1852 holds no Falco; 1850's is Michele Falco's, already held; and 1851's is LUIGI FALCO's, 13 August, act 14 — a marriage in generation three's own household that this archive had only from a death act. Neither Giuseppe Falco nor Antonia di Guida is named in any of the three.)* This archive read the marriage *indexes* for 1809–1849 and then stopped, because the index series itself stops at 1849. Eleven years of marriages, unlooked at, in exactly the window where these two people's children were marrying. One sheet of fourteen acts per two images. --- ## HE WAS ALIVE ON 14 FEBRUARY 1860 — 21 September 2026 Ten years and three months later than anything this archive held for him. His daughter Maddalena's marriage act, Arienzo Matrimoni 1860 act 8: «*figlia di GIUSEPPE FALCO, di professione contadino, domiciliato ivi, e di ANTONIA GUIDA DEFUNTA, domiciliata ivi*». The clerk wrote «defunta» for the mother eight words later and nothing for him. He is also unmarked in his daughter Angela Maria's act of 15 November 1857. Born about 1780, he is about eighty in 1860 — the founder's son, outliving his wife. WHAT THIS DOES TO THE ROW. The window was *«after 26 November 1849»*. It is now *«after 14 February 1860»*, and the Arienzo civil death registers run out in 1865. Five years of them remain between his last sighting and the end of the series: 1860 (read to 5 March on 21 September), 1861, 1862, 1863, 1864, 1865 — and 1861–1865 were read out earlier for other reasons, with no Falco of his generation. So this row's own answer — «1866 or later, in a series this archive has no access to» — has gone from likely to very hard to avoid. *What is still worth doing is the rest of the 1860 death volume, which is forty-five unread images.* ### 1860, 1861 and 1862 read out — 22 September 2026 Three volumes, 208 images, and none of them had a coverage row; `an_ua14364` had never been opened here. No Giuseppe Falco in any of them. But the years are not empty. 1862 gives PASQUALE FALCO, twenty-four, of strada Camellara, a fifth child for «Michele Falco & Antonia Vigliotta»; and MARIA FALCO, wife of FRANCESCO CIOFFI, sixty-seven, dead 1 May 1862 — a woman this archive followed through eight children's acts without a date at either end. *Her act names her parents as a third Falco–Chiara Rivetti couple; see row 87.* 1863 and 1864 are what is left, `an_ua14365` and `an_ua14366`, with 1865 already read. If he is in neither, the whole Antenati death series 1844–1865 is closed against him and this row's answer — 1866 or later — is the only one standing. ### 1863 and 1864 read — THE SERIES IS CLOSED AGAINST HIM With these two, every Arienzo civil death register from 1844 to 1865 has been read for FALCO. 1863 holds no Falco at all; 1864 holds only Luigi Falco, dead 16 December, which this archive already had. So: GIUSEPPE FALCO was alive on 14 February 1860 and he is not in any Arienzo death register between that date and the end of the series in 1865. Antenati has no Arienzo civil deaths after 1865. This row's answer was «1866 or later, in a series this archive has no access to». It is now the only answer left, and it has been tested rather than assumed. *What could still reach him: the parish register, which is not indexed by surname the machine can read; the Arienzo processetti of 1866 onward, which do not exist on Antenati; or a death in another comune, as his wife's now appears to have been.* Row 39's conclusion and this one are now the same shape, reached independently: neither of them died where this archive can see. --- ## THIS ROW WAS STALE, AND IT COST A DUPLICATED SWEEP — 23 September 2026 The paragraph above says the Arienzo marriage registers for 1857–1860 «have never been swept for Falco». They had been — on 21 September 2026, two days before this row was last edited. `site/src/data/searched.json` records all four volumes, `an_ua14261`–`an_ua14264`, «*every act read at the name band*», and states in terms: «*1858 AND 1859 HOLD NO FALCO AT ALL — thirty-nine and twenty-eight images, every act read*». Two files of this archive disagreed about whether four volumes had been opened, and nothing in the nineteen-step chain compares them. On 23 September I believed this row, re-read `an_ua14262` and `an_ua14263` end to end, and reproduced the earlier nil exactly. [[archive-contradicts-itself]] THE NIL IS CONFIRMED AND NOW CARRIES AN ACT COUNT. 1858: images 3–38, THIRTY-SIX acts. 1859: images 3–27, TWENTY-FIVE acts. Sixty-one acts — no FALCO as groom, as bride, or as a parent of either, which is the reading this row needs, because a child's marriage act is what dates a parent. So neither Giuseppe Falco nor Antonia di Guida is named in 1858 or 1859, and no window moves. AND THE EARLIER SWEEP'S BAND IS NOW VINDICATED. It recorded «read at the name band» without saying which half — and this archive's standing hazard is that a band on one half reads half the acts *consistently*, so nothing looks missing. In these volumes an act BEGINS on the RIGHT page, which carries the date, both spouses and all four parents; the LEFT page of the next image holds the *previous* act's witnesses. `image = act + 2`. The wrong half holds no parents at all, so a nil from it would have been obvious — the earlier reading stands. [[half-page-sweeps]] THE RULE THIS PRODUCED. *Before sweeping any volume, grep `searched.json` for its ark.* The work list records what someone thought still needed reading; the coverage data records what has actually been read, and they drift. One `grep an_ua14262 site/src/data/searched.json` would have saved the pass.

DONE. He was married — to CAROLINA CIMMINO. His death act, read from the image (it was a FamilySearch index stub): «*è morto nella sua casa LUIGI FALCO = MARITO DI CAROLINA CIMMINO — di anni quarantuno, di professione industriante, domiciliato Contrada Capodiconca*». He had left strada Camellara, and his stated age is wrong by three — forty-one for a man five days short of forty-five. And the act answered a far older question in its parentage clause: his parents are given with domiciles and no «fu», in a register that marks four other dead parents on the same opening. Vincenzo Falco and Andreana Crisci were alive on 16 December 1864. What remains of this row: Carolina Cimmino. The Arienzo marriage record is complete 1809–1858 and holds no Falco–Cimmino, so they married after 1858 or not at Arienzo.

She died 15 March 1819, eleven hours after her brother, as a daughter of Giuseppe Falco and Gelsomina Vigliotta. The Arienzo civil birth registers for 1816, 1817, 1818 and 1819 are all read end to end and she is in none of them. A child dead in March 1819 was born in one of those years. So either the act was never drawn, or she was not born at Arienzo — and this household is the one that turns up at strada Costa, then strada Camellara, then Santa Maria a Vico. Test the parish baptisms first, then Santa Maria a Vico. 17 September — THE PARISH TEST WAS RUN AND IT CANNOT PRODUCE A NEGATIVE. Here is the proof. FamilySearch's full text of the Arienzo parish baptisms was searched for the Latin forms the clerk actually writes. **«*ex Josepho Falco*» returns TWO pages in the entire collection — for a household with at least eight children — and Rosaria is not among them. That is not evidence of absence, because the machine cannot read the surname. Measured on the same collection, same place filter: | spelling | pages | |---|---| | `Falco` | 361 | | `Falo` | 126 | | `Falio` | 92 | | `Zalco` | 7 | | `Salco` | 6 | | `Talco` | 3 | Two hundred and eighteen pages render the name in a form a search for «Falco» never sees — about a third of the corpus.** And one index leaf writes the SAME entry twice, once as «*Joseph Zalco fol: 185*» and once as «*Joseph Falco fol: 185*». A surname phrase search here is a finding aid, never a negative. What the test did turn up: Maddalena Falco's baptism of 7 June 1818 (row 16), the day of Matteo's 1807 baptism, and Domenico Clemente's baptism of 18 March 1819 — three days after this household buried Domenico and Rosaria eleven hours apart. Rosaria's own act is still not found, and Santa Maria a Vico is still untried. ## A SECOND ROSARIA IS FOUND, AND THIS ROW IS STILL OPEN — 21 September 2026 ROSARIA FALCO, born 15 March 1821, act 22 of `an_ua14418`, to Giuseppe Falco and Gelsomina Vigliotta — found on the declarant-band re-sweep of row 85, because the mother is a Vigliotta and the parents band never saw her. She is not the Rosaria this row is about. That one died on 15 March 1819; this one is born on 15 March 1821, the same day of the same month, two years later. *The coincidence is recorded and not explained.* What it does change: the first Rosaria is now certainly a distinct child rather than a possible mis-dating, because the household demonstrably reused her name — and her birth act is still in none of 1816, 1817, 1818 or 1819. The row stands as written. ## CONFIRMED FROM THE IMAGES, AT BOTH BANDS — 21 September 2026 Work-list row 86 read 1816, 1817, 1818 and 1819 end to end, at the combined band those prose-format volumes allow — declarant and mother on one half-page. Every Falco in all four years is now listed on /searched, and there is no Rosaria among them. The nearest thing is her own household's DOMENICO FALCO, 25 November 1816, act 116, to Giuseppe Falco and Gelsomina Vigliotta of strada COSTA. So the negative is now as strong as this archive can make it: four consecutive years read act by act, at the band that finds a father who declares his own child — and that band did overturn two other years' nils in the same pass, so it is not a band that fails quietly. *Either the act was never drawn, or she was not born at Arienzo.*

A household created on 3 February 1845 and completely unexamined. He was a *colono* of strada Sant'Andrea, twenty-five; she was twenty-three and an orphan. Their children would be in the Arienzo birth registers from late 1845 — which row 9 reaches but has never opened. The tree already carries a «Crescenzo Falco» with a wife «Maria Concetta Morgillo» and a son Giovanni, unsourced. That is either this marriage under a different forename for the bride, or a second couple. A single birth act would decide it. 16 September — 1846 READ END TO END, AND CRESCENZO IS NOT IN IT. All 110 images of `an_ua14443` swept on the «*è comparso*» band, plus the last two months of 1845 (`an_ua14442`, images 118–135), which is the window a first child of a February 1845 wedding would fall in. Three Falco fathers in the two sweeps, none of them Crescenzo: MICHELE FALCO, 38, strada Camellara (December 1845), FRANCESCO FALCO, 42, strada Camellara (January 1846) and MATTEO FALCO, 38, colono, strada Camellara (September 1846). Same caveat as every sweep of this kind: the declarant line was read, and a large minority of these acts are declared by a levatrice — Leontina Martone, Rosa Morgillo, Felicia Martone recur — whose acts name the father lower down. This does not close 1846; it makes Crescenzo unlikely in it. And there is no index to check the sweep against. The 1845 tavola is printed and left blank. The 1846 tavola exists but the camera skipped it, laying the operator's own card across the frame: «*THIS INDEX HAS BEEN MICROFILMED ALREADY*» — the same card row 10 found over the 1825 leaf. So it is not one lost index, it is a policy. Next: 1847 onward, and the tree's rival «Crescenzo Falco × Maria Concetta Morgillo with a son Giovanni» is still undecided. ## FOUR CHILDREN, AND THE ROW'S OWN QUESTION ANSWERED — 21 September 2026 MARIA FELICE FALCO, born 21 November 1845 — act 111, images 116–117. *The earlier sweep began at 118.* Nine months and nineteen days after the wedding. And the margin of her act carries her own marriage — 3 February 1873, to PASQUALE QUOTOLO, at Arienzo — out of the years whose registers are lost. GENNARO FALCO, born 1 June 1848 (act 42) and dead on the 3rd, at two days; GIUSEPPE FALCO, dead 2 June 1848, aged one. Brothers, a day apart, declared by the same two neighbours of strada Camellara. THE TREE'S RIVAL IS SETTLED. «*Maria Concetta Morgillo*» is almost certainly this woman: the marriage act calls her ANGELA MARIA and every act of her children calls her plain MARIA, twenty-three in 1845 and twenty-five in 1848 — the age she married at. *The tree's «son Giovanni» is still unattested.* AND THE ARCHIVE ALREADY HAD TWO OF THESE CHILDREN. `notes/1844-1847-death-indexes.md` recorded both 1848 deaths from the image in an earlier session, under the heading «*CRESCENZO FALCO × MARIA MORGILLO is a household this archive has never held*». It is this one. The note reached no data file. Why the earlier sweeps missed them, precisely. They read the declarant line. Maria Felice's act sits two images before the window; Gennaro's was presented by a midwife, whose name is the only one on that line. *The parents band is the instrument; the declarant band was never going to find either.* STILL OPEN: GIUSEPPE FALCO's BIRTH ACT. One year old on 2 June 1848 puts it between 3 June 1846 and 2 June 1847. 1846 is now read at the parents band and holds no Maria Morgillo; 1847 is swept and part-read, and its *tavola* is under the archivist's card — a seventh volume with that card, after 1816, 1823, 1824, 1825, 1830 and 1846. *And three Falco mothers were seen in 1846 and not chased: RAFFAELA FALCO, 34 (image 8), GIOVANNA FALCO, 40 (image 34) and FRANCESCA FALCO, 34 (image 46).* ## AND THE PERSON-GRAPH BUILDER REFUSED THE FIRST VERSION OF THIS NICOLA COSSI and ANDREA DIGLIO, the two neighbours who declared both June 1848 deaths, were first written into this household as members with the role *declarant*. `tools/build_people.py` refused the build: «*SELF-CHECK FAILED — 2 parent AND child of one person; 2 somebody is their own ancestor*». The builder merges people by name across households — it says so itself, «*19 people were one person recorded in two houses*» — so a Nicola Cossi who is somebody's father elsewhere in the archive and the Nicola Cossi who declared a neighbour's child's death became one man, and the graph closed a cycle. A NEIGHBOUR WHO DECLARES IS NOT A MEMBER OF THE HOUSEHOLD. The two men are named, aged, placed and quoted inside the death acts' own evidence, which is where they belong; the rows were withdrawn. *Nothing was lost, and the refusal was right.* ## GIUSEPPE'S BIRTH ACT IS NOT IN THE WINDOW — AND THE SEARCH FOUND GENERATION FOUR INSTEAD «Di anni UNO» on 2 June 1848 puts his birth between 3 June 1846 and 2 June 1847. Both volumes have now been read end to end at the PARENTS band and neither holds a MARIA MORGILLO. Three things remain possible and this row prefers none: the stated age is wrong, which these clerks do constantly; the act was never drawn; or he was not born at Arienzo. AND THE 1847 VOLUME GAVE SOMETHING ELSE ENTIRELY. Act 92, 20 November 1847: VINCENZA FALCO, to RAFFAELE FALCO, thirty, and ANGELA ROSA MORGILLO, twenty-nine, of strada Camellara. That is GENERATION FOUR, and this archive had no household for it — Raffaele appeared only as a child in his father's house, and their earliest documented child was generation five's Carmine Antonio of 1850. The tree named Vincenza, unsourced; here is her act. On /corrections. THIS ROW IS DONE — four children, the household identified, the 1818 date corrected. *What is left is one act that may not exist, and it is recorded as a named absence rather than a task.*

Done. The household exists. GENNARO MORGILLO, *colono*, and MARIA FELICIA SANSONE, *contadina* — both dead before 2 February 1845, stated in their daughter's marriage act, which marks the dead explicitly. Their daughter ANGELA MARIA MORGILLO, twenty-three and an orphan, married Crescenzo Falco. SANSONE had never appeared in this archive at all. Not claimed as kin; recorded because the Falco married the Morgillo repeatedly and this archive still cannot tell those Morgillo houses apart.

DONE 17 Sep 2026 — she has a wedding, a father, a death window and two more children. THE MARRIAGE. «*Anno D[omi]ni … 1793, die vero vigesima sexta 26 Martii … in matrimonio conjunxit nondum uxoratos ARCHANGELUM CRISCI FILIUM NICOLAI, et SARAM RUGGIERO FILIAM PASCHALIS, utrosque hujus Archipresbyteralis [Ecclesiae] S. Andreae*» — Arienzo, 26 March 1793, before D. Joseph Crisci, coadjutor, by decree of the episcopal curia of Sant'Agata de' Goti. He was about twenty-eight, she about twenty-one, and Andreana Crisci — generation three's wife — was baptised three and a half years later. HER FATHER: PASQUALE RUGGIERO, living in March 1793. His is the name this row existed to find. And the groom's father, NICOLA CRISCI, a generation this archive did not have. A DEATH WINDOW OF THREE YEARS, where there was none. She bore a son in 1817 aged forty-five, and by 1 June 1820 her daughter's marriage act writes «*q[uonda]m Sara Ruggiero*». Dead between 1817 and June 1820, aged about forty-five to forty-eight. AND TWO CHILDREN. ANGELA ROSA CRISCI, married 1 June 1820 to Angelo Clemente Frasca of the parish of Sant'Agnese; and LORENZO CRISCI, married 25 March 1846 to Maria Angelica Ruotolo, daughter of Antonio and Maria Giovanna CARFORA. THE IDENTIFICATION RESTS ON THE COUPLE, NOT THE FORENAME — this archive holds a second Arcangelo Crisci, the one who married Anna Falco, and it is the pairing with Sara Ruggiero that separates them. Every act above is a machine transcription of the parish hand and the images are not yet read, though the 1793 act is attested twice on its own leaf, by the year's index line and by the act.

Done, and it produced three things. 1. Eight citations pointed at the wrong act — the whole harvest-numbering problem, found because the first act opened was not the act its citation named. Corrected; see row 58 for the 161 never followed. 2. A marriage nobody would have looked for. PASQUA FALCO married PASQUALE CARFORA on 30 December 1880, written across the foot of her own birth act of 1823. Her second marriage, at fifty-seven, and her first dated sighting after 1856. 3. THE INSTRUMENT HAS A DATE. Four acts, sorted by the MARRIAGE: 1844 Raffaele — none. 1845 Crescenzo — none. 1856 Domenico Clemente — annotated. 1880 Pasqua — annotated. The birth volumes span 1818–1823 and do not predict it; the marriage dates do. The claim published an hour ago that Luigi Falco's blank margin was «a real negative» is corrected — the control had not been run. It now means only that he did not marry after about 1855 where Arienzo was told.

DONE 16 Sep 2026 — «ivi» is the COMUNE, and the finding that rested on it is withdrawn. The test this row asked for was in three acts already rendered for row 60. The proof, three times, in two different printed forms. The 1824 birth act atto 73 names two witnesses in one sentence: «*Tomaso Esposito … domiciliato ivi strada Camellara*» and «*Antonio Bernardo … domiciliato ivi strada Terra Murata*» — one word, two streets. Crescenzo Falco's own 1845 marriage act does it again: groom «*domiciliato ivi strada Sant'Andrea*», bride «*domiciliata ivi strada Cappuccini*». And the 1828 death act atto 54 uses «ivi» for the first person it names, where the only antecedent is the comune. A transcription error was hiding inside it. The 1845 act reads «domiciliato ivi strada Sant'Andrea»; this archive wrote «domiciliato in strada Sant'Andrea». Both files are corrected. Consequence: a bare «domiciliato ivi» states the comune and gives no street. So the 1845 act does NOT place Giuseppe Falco off strada Camellara, and the household may never have left. The correction supersedes the published entry; the rest of that finding — Crescenzo's birth date, Antonia di Guida alive in 1845 — is untouched, because it turns on the form marking other parents dead and not these two. Written into the method page as the sixth thing the Arienzo registers taught this archive to distrust, at `/method#ivi`.

Done, and the hypothesis it was raised to test is dead. The same list is at the back of the 1818 volume too — same header «*Nomi, e cognomi de' genitori | Giorno della nascita | Osservazioni*», facing the same 1849 Procuratore del Re certificate about counting sheets and removing the blank ones. The «Osservazioni» column carries a roman numeral: VIII in 1818, XI in 1819. It is NOT a list of births missing from the register. Its second row in 1818 is «*[Vince]nzo FALCO ed Andreana CRISCI — 10 Maggio 1818*» — which is Raffaele Falco, generation four, act 47, folio 149, sitting in the main sequence at image 42R where it always was. So it cannot explain Rosaria Falco's absence, which was the reason for asking. What it IS remains open — the roman numeral and the certificate about removed leaves suggest a reconciliation list for renumbered folios, but that is a guess and is recorded as one. What is certain is that it is USABLE. Parents and a date, one page, at the back of a volume — and in a year with no annual tavola, such as 1819, it is the only index that exists. Check the last leaf of every birth volume.

Done, and it is on /method. VINCENZO FALCO, baptised 1 February 1792, is the only person in the archive with a proven birth and a long run of civil acts. Six sightings, twelve years: +10.8, +0.7, −1.9, +0.9, −2.6, −0.9. Five of six within three years — so a stated age here is good to about two or three years and no better, which is now published as the working figure. The 1816 outlier is kept: both spouses inflated by ten years and five in one act, with two readings offered and neither chosen. A limit came with it. It could only be done for one man, because the archive stores its facts as prose and the ages sit inside quotations usually attached to somebody else. Any «across all people» question is currently unanswerable here — the same root cause as row 55.

Done, and it found a live conflation instead. Grouping each household's children by forename turned up the real necronyms — DOMENICO CLEMENTE born three days after Domenico was buried, Giuseppe and Mariantonia Cioffi each given again, Maddalena the second, Matteo the second, the first Raffaele of 1816 — and three pairs that were not necronyms at all but one child spelled two ways. `francesco-cossi` and `francesco-cioffi` were the same man with two pages, and four children of an established CIOFFI house still carried COSSI. Merged and corrected; 632 people where there were 634. The finding that matters: in a house that buried six children in nine years a repeated name is the ordinary case, not the anomaly — which is exactly why the two living Giovanni were so hard to credit and took weeks.

Done, and it is in the build chain as `check:corrections`. `tools/check_corrections.py` holds two rules. First, a table of RETIRED READINGS. One row per surname a correction has killed, naming the household it was killed in. The gate fails if any PERSON in that household still carries it — while leaving the ACT's own wording alone, because «*GIUSEPPA COSSI, di mesi otto*» is what the clerk actually wrote and must stay quoted. Second, no child of one household may appear under two surnames. Restricted to `role == child` deliberately: a household legitimately holds Sara Ruggiero the matrina and Sara Falco her granddaughter, and a gate that flags those is a gate people learn to ignore. Tested against the bug it was written for. Reintroducing «Antonia Cossi» and «Giuseppa Cossi» into the household file makes it fail with three findings and exit 1; the corrected file passes. It would have caught all three of this week's conflations, each of which was found by accident.

Done. The front-page chart was the only place it had gone wrong, and the reason is worth keeping. Everything else that displays a total counts a source that already includes the living: the front-page Numbers block counts `people.json`, which holds all 106 living people with pages of their own; `/marriages/` counts named and withheld separately and says so in its dek; `register.length` counts acts, not people. The chart was the one total built from a field that means «publishable». `n` is `len(g["people"])` in `build_marriedin.py`, and the living never reach that list. The lesson generalises past this archive: a count is safe when it is drawn from the population, and dangerous when it is drawn from what the archive is allowed to print about the population.

Done. It was on twenty-three more pages. The built HTML — the only place the fault shows — carried raw markdown on /worklist (46 backticks), /register (34), /searched (28), /households and /people/crescenzo-falco (13 bold each) and nineteen /who/ pages printing a plate filename in backticks. Fixed by one renderer rather than five. `site/src/lib/md.js` now holds it, `/corrections/` was moved off its private copy onto the shared one, and `/households/`, `/register/`, `/searched/`, `/people/[slug]` and `/who/[slug]` were put on it. Escape first, then render, so no markup inside a transcription can get through. /worklist is the exception and it is not this archive's to fix — see row 17.

Done — and the answer is no, it does not. /marriages/ solved this problem and the chart then reintroduced it. That page buckets its families three ways: `families` where `n >= 3`, `thin` where `0 < n < 3`, and `held` where `n === 0` — and a held family is rendered with its withheld count instead of a bar, under a sentence that says it outright: «*Where a family appears with no named people at all, every person the archive holds for that surname is living.*» So the older page knew. The shared rule was extracted into `marriedin.js` precisely so the two could not drift — and the chart drifted anyway, by taking `n` from the shared filing and leaving behind the treatment of zero that was never in the lib. Extracting a rule moves the rule, not the knowledge around it.

Done, and the answer is that the automated version cannot be written — for the same reason row 49's could not. A regex over the evidence prose returns 268 name-shapes, and almost all are noise: Latin declensions of people the archive already holds (*Mattheo Falco*, *Vincentium Falco*, *Franciscae Crisci*) and sentence fragments (*comparso Giuseppe Falco*, *two Giuseppe Falco*, *not Cossi*). The archive stores narrative, and narrative does not yield a list. What the check DID establish is the shape of the leak, and it is not where the row assumed. Of the candidates, forty reach the register and are searchable while never becoming a person; a handful reach neither. The register is the safety net and it usually holds. Seven people were promoted by hand — the ones introduced in the last day and verified individually: Gennaro Morgillo, Maria Felicia Sansone, Angela Maria Morgillo, Domenica Vigliotti, Pasquale Carfora, and the strada Cossi household of Francesco Cossi and Teresa Morgillo, which exists to hold the two surnames apart. 638 people.

DONE 16 Sep 2026 — THE REGISTER ANSWERS IT ITSELF, AND THE ANSWER IS THAT IT IS NOT CONTINUOUS. Test 1, the numbering gap: there is none. Acts 78 to 91 run unbroken, read from the image. And the volume ends at 91, not ninety — this archive's own statement «*all ninety acts*» was one act short. Test 2, a supplementary list at the back: cannot be run on this filming. There is a tavola leaf after the last act (`an_ua14343` img 55 left), but it is photographed with its ink almost wholly lost and the facing page's writing transferred onto it in mirror. Three order-numbers survive and no names. Not read, and not recorded as absent. Test 3 is the one that answers the question. THE DEATH DATES HAVE AN ELEVEN-DAY HOLE, AND THE FALCO BURIAL IS INSIDE IT. Read act by act from the image, the deaths run: act 84 — 19 November · act 85 — 23 November · act 86 — 26 November · act 87 — 7 DECEMBER. The comune registered nothing between 26 November and 7 December 1841. The parish buried JOANNES FALCO on 27 November. And the hole is not special, which is the real finding. A fortnight earlier the register does it again: act 80 is a death of 30 October and act 81 a death of 13 November — fourteen days. Two silences of a fortnight and eleven days in one autumn, in a register that otherwise enters a death every two or three days that season. This register was not capturing every death at Arienzo, and its own date sequence proves it without any inference from the parish book. So the 1841 civil nil for Falco is not evidence of anything — and neither is any other nil taken from this volume without checking the dates either side of it. Rows 12 and 56 are the same question and both close here.

Done — and the row's own premise was wrong, which is the finding. There is nothing to re-parse. Every stored transcript is a SNIPPET of at most 260 characters, and the median length is identical for rows that have a child's name (260) and rows that do not (260). The child column is blank not because a parser gave up but because the 260-character window happens not to contain the name. Forty-three of the forty-four blanks are like this. A margin-position extractor was written and tried: eight of forty-four snippets open with a name-shape, and only two are real — «Cresantios Falco», already promoted today, and «Raphael Falco Obiit die 26 Februarii 1881», which is a death annotation rather than a baptism and is now a lead on row 18. So the queue is a fetch problem, not a parsing problem. Forty-three half-answers, each needing the full page or the image. And the Crescenzo case is explained properly now: his snippet did contain the name, in the margin position at the very start, which the parser never looked at — it was not defeated by the Latin.

Made executable: `tools/citation_check.py`. 207 citations, and 134 are the SOLE support for their fact — that is the priority order the row asked for and could not previously produce. Eight are already corrected. It prints the render command for each, so verifying one is a copy and a glance — does the act number match, does the date, is the name there. `--volume an_uaNNNNN` does a whole volume in one sitting, which is the cheap way to spend it: `an_ua14319` (deaths 1817) alone holds 22 citations, 12 of them sole support. The top of the list is misleading in one way worth knowing: the acts read from the image in this session sort first because nothing else supports them, and they are verified by construction. Start at `an_ua14319` and work the old volumes. --- STARTED 20 September — two volumes verified, 35 citations, TWO ERRORS. `an_ua14319` (deaths 1817) — 22 citations across five images, ALL CORRECT. Including the act this archive is built on: «*RAFFAELE FALCO figlio di VINCENZO FALCO e ANDREANA CRISCI*», declared 28 October 1817 by Giuseppe Falco, 37, massaro di campagna, strada Camellara, «PARENTI». The name reads ANDREANA, not Adriana, which matters given the name-cleaning bug in row 17. One suspicion of mine was wrong: «Michedalena Rivetti» looked like it should be «Michela Maddalena», and the act wraps the word across a line break as «MICHE-/DALENA». The archive's transcription is right and needs no change. `an_ua14318` (deaths 1816) — 13 citations across three images, and two of the three are wrong. 1. A citation one leaf early. Donna Francesca Rivetti's death was cited at img7L act12. Image 7 left carries atto 10, the death of a Rosa Cimmino. Act 12 is on img8L and says exactly what this archive said it says. The fact was right; the leaf reference was wrong — the failure mode this row was written to catch. 2. A month and a mother. GAETANA LAUDATO — «*the deepest this archive reaches*», born about 1738 — was published as dying 9 March 1816 with her mother as «*Giovanna —*», surname unread. The act reads «a' detto dì NOVE del mese corrente di APRILE» and «figlia delli furono GIACINTO LAUDATO e VITTORIA RUOTOLO». A month out, and a mother who was never unreadable. Both fixed. Method note for whoever continues: the 22 citations of 1817 sit on five images and the 13 of 1816 on three. Render the distinct images, not the citations — the volume is a few renders, not twenty-two. Remaining: `an_ua14416` (13), `an_ua14325` (12), `an_ua14321` (12), `an_ua14320` (12), `an_ua14345` (11) and the long tail. FOUR VOLUMES DONE — 59 citations, FIVE ERRORS, and every error is a DATE or a LEAF, never a fact. | volume | citations | images | result | |---|---|---|---| | `an_ua14319` deaths 1817 | 22 | 5 | all correct | | `an_ua14318` deaths 1816 | 13 | 3 | 2 wrong | | `an_ua14320` deaths 1818 | 12 | 4 | all correct | | `an_ua14321` deaths 1819 | 12 | 6 | 2 wrong | 1819's two are both off by one day, and both were confirmed at magnification before being touched: - SUOR MARIA MADDALENA RIVETTI — published 30 March 1819; the act reads «*a' dì TRENTUNO del mese di Marzo*» and repeats «*a' detto dì trentuno … ad ore ventidue*». 31 March. - CARLANTONIO MORGILLO, declarant — published 20 October 1819; the act reads «*a' dì VENTUNO del mese di Ottobre*». 21 October. His age series is unaffected. THE PATTERN ACROSS ALL FIVE ERRORS IS WORTH MORE THAN THE FIXES. Not one was a misread NAME, a wrong PARENT or an invented RELATIONSHIP. Every one was a date word or a leaf number — «trenta» for «trentuno», «venti» for «ventuno», «marzo» for «aprile», img7L for img8L. The archive reads its people well and its numbers carelessly, and the small words are exactly what no reader downstream can catch. 1819 also confirmed the archive's most quoted detail to the hour: DOMENICO FALCO died «*ad ore dodici*» on 15 March and ROSARIA FALCO «*ad ore ventitré*» the same day — eleven hours apart, as published. Remaining: `an_ua14416` (13), `an_ua14325` (12), `an_ua14345` (11) and the tail. `an_ua14416` (births 1819) — 13 citations, the archive's four cornerstone birth acts, and ONE error that could have cost a family. Verified exactly: CRESCENZO FALCO, born 11 October 1819, «*è comparso GIUSEPPE FALCO, di anni quaranta, contadino … strada Camellara … da ANTONIA DI GUIDA sua moglie legittima, d'anni trentuno*». DOMENICO CLEMENTE FALCO, 18 March 1819, Giuseppe Falco 38, Gelsomina Vigliotta 35 — and the margin that founded this archive's whole margin method is there and reads as published: «*Con atto 6 Agosto 1856 atto 21 in S. Maria a Vico il Falco sposò DOMENICA VIGLIOTTI*». THE ERROR: Luigi Falco's birth was cited as act 133. It is act 134. Image 71 left is headed «*Numero d'ordine CENTOTRENTAQUATTRO*». And act 133, on the facing leaf, is a different family that looks like this one: «*è comparso VINCENZO MAURO, di anni ventisette, viaticale, strada Sant'Andrea … da ANDREANA CARFORA sua moglie legittima, d'anni ventisei, un Maschio*» — same day, same clerk, a Vincenzo and an Andreana in each. Anyone chasing act 133 would have found a plausible wrong family. The leaf reference was right and the number was not. ONE FLAGGED AND DELIBERATELY NOT CHANGED. Act 22 is cited at 19 February 1819 and the date word looks like «*diciotto*» at ordinary magnification. At the leaf's full resolution it is too faint to separate diciotto from diciannove, so it stays as published and is recorded here as unresolved. A date this archive cannot read is not a date it may change. Running total: five volumes, 72 citations, six errors — every one a number, none a person. `an_ua14325` (deaths 1823) — 12 citations, ONE error, and the «nipote» holds. Act 101 of 9 October 1823 verified word for word: PAOLO RIVETTI, fifty, husband of Angela Espositella, of strada Fontana, son of Pasquale Rivetti defunto and Giovanna Perrotta — declared by GIOVANNI RIVETTI, twenty-three, POSSIDENTE, and the relationship word, magnified as this archive's rule requires, is «Nipote». The error is an age: Giuseppe Falco fu Matteo is published at forty-three on 20 June 1823 and the act reads «quarantadue». The clerk's own «tre», three lines below on the same leaf, is a different shape. `an_ua14345` (deaths 1843) — 11 citations, five acts, THREE errors and all three are dates. Act 17 is 10 APRIL 1843, not 10 June: the month is written «Aprile» on both of the act's date lines. Act 29's death is 26 July, not 27 — the archive had taken the day the act was drawn. Act 33 is 4 AUGUST, not 14: «il dì quattro del mese di Agosto», and act 32 beside it is 3 August while act 35 is 10 August, so 14 is arithmetically impossible. AND THE VOLUME GAVE UP SOMETHING THE ARCHIVE ALREADY OWNED. Act 33 is GENNARO MORGILLO, colono, widower of a Felicia, son of Angelo Morgillo of the Monticello. Crescenzo Falco married ANGELA MARIA MORGILLO eighteen months later, daughter of a GENNARO MORGILLO DEFUNTO, colono, and of MARIA FELICIA SANSONE defunta. Two households in this archive, never joined, and the evidence line on one of them called the act «the closest the record comes to the Falco-Morgillo marriage that was still thirty years away». It was eighteen months away, and the archive held the marriage. Published as PROBABLE, with the *processetto* named as what would settle it — work-list row 73. Running total: seven volumes, 95 citations, ten errors. Not one is a name, a parent or a relationship. Every single one is a number. `an_ua14328` (deaths 1826) — 9 citations, three acts, and THE RULE THIS ROW HAD STARTED TO BELIEVE IS BROKEN. Act 73 of 28 October is clean: Carlantonio Morgillo and Giuseppe Falco, both forty-six, both «vicino della Defunta» — the first act in the series where a Falco and a Morgillo stand side by side, and it verifies exactly. Act 22 of 25 April is right in everything that matters, including the silence: Pasquale Falco and Chiara Rivetti are named with no «defunto» on a leaf where this scribe does mark the dead. Its one error is an age — Angelo Sansone is forty-six, not forty-five. Act 56 of 9 August gave up TWO MISREAD NAMES, the first of this whole row. The dead child is GIUSEPPE CIOFFI, not Cossi, and his father is CESARE CIOFFI, not Gaspare Cossi. So the finding this row had been repeating — «every error is a number, never a person» — is false, and it is recorded here as false rather than quietly dropped. Eight volumes, 104 citations, thirteen errors: eleven numbers and two names. The same act states a kinship the archive was not holding: «*SALVATORE SPOSITO … ZIO DEL DEFUNTO*», the mother's brother, magnified and confirmed. The 1826 reading log had it all along. And Angelo Sansone's surname settled a question two volumes back. «Sanſone», in this hand, on this street, is the control that turned the 1843 «Sompone» into SANSONE — so the wife of the Gennaro Morgillo who died in 1843 is the MARIA FELICIA SANSONE of the 1845 marriage, and the identification of Crescenzo Falco's father-in-law now agrees on five things instead of four. `an_ua14344` (deaths 1842) and `an_ua14342` (deaths 1840) — 15 citations, seven acts, and the two worst errors of the whole row. 1842 is nearly clean. Act 2 (Luigi Cioffi, one day old, son of Raffaele Cioffi and FRANCESCA FALCO — her name plain on the leaf), act 15 (Don Angelo Scarpati, seventy-six, ARCIPRETE) and act 80 (Raffaele Cioffi declaring at thirty-six) all verify word for word. Act 55's first declarant is not a Cioffi: the Giovanni Cioffi two lines below him carries the double-f ligature and his name does not. It reads GASPARO, which is an Arienzo surname — probable, not settled. 1840 gave up two errors this archive should be ashamed of. One: a quotation that is not in the act. «*È morta nella sua propria casa MARTA CIOFFI D'UN MESE*» was published in quotation marks. The act says «d'anni SETTE». She was seven years old, not one month. Both declarants were mis-aged with her — Clemente Cioffi is thirty-six, Virgilio Cioffi twenty-three. A wrong number can be suspected. A quotation cannot. Two: a person who never existed, and a problem invented to go with her. «Donna SERAFINA Cimmino» was published as the mother of the archpriest's sister, with a note calling the 1840/1842 difference «*THE ONE DISCREPANCY, published rather than resolved… either a clerk misread a forename, or Stefano Scarpati married twice*». Both acts say FAUSTINA. The misreading was this archive's; the discrepancy was manufactured out of it; and publishing a doubt is only honest when the doubt is real. It does buy something back — the archpriest and the sorora now agree on both parents, not one. And the age series that was held up as proof of a careful clerk runs backwards. GIOVANNI CIOFFI was published at «FORTY in 1840, FORTY-SEVEN in August 1842, FORTY-EIGHT in September 1843 — this clerk got one right». The 1840 act gives him no age at all, and the 1843 act says forty-three, which this archive had read correctly on 15 September and corrected in one file only. Forty-seven, then forty-three: he got younger. Ten volumes, 119 citations, eighteen errors. Fourteen numbers, three names, and one sentence that was never written. `an_ua14330` (deaths 1828) — 8 citations, four acts, five errors and one act that is perfect. Act 54 is the one that matters and it verifies word for word. «*Sono comparsi GIUSEPPE FALCO, di anni cinquantasei, di professione POSSIDENTE, domiciliato in strada Camellara, PARENTE DELLA DEFUNTA*» — both the trade and the kinship word magnified. And the act says it twice: the second declarant, CESARE CIOFFI, thirty-one, of the same street, is «Parente della Defunta» too, which this archive had not recorded. Act 11 verifies in everything except the word the household is named after. Twenty-six months, 14 March, Antonio Morgillo 27 and Simone Morgillo 24 both «parente della defunta» — all right. The mother is VINCENZA PEPE, not Vincenza Vece. Act 14 and act 56 gave up four numbers and a forename. The dead boy of May is PASQUALE, not Raffaele. Vincenzo Morgillo is twenty-seven, not twenty-nine. Nicola Morgillo is thirty-two, not thirty-four — and the act cited for his age does not age him at all, exactly as with Giovanni Cioffi's «forty in 1840» two volumes ago. Twice in one morning, a number was attached to a man from an act that never states it. Vincenzo's wife is FELICELLA Rivetti, not Felicetta. And the retired reading «Cossi» was still standing in the prose of act 54's evidence, months after the household was corrected to Cioffi and a gate was written to hold it. `check_corrections.py` reads person fields, not sentences — and this archive keeps most of what it knows in sentences. Eleven volumes, 127 citations, twenty-three errors. `an_ua14428` (births 1831) — 9 citations, three acts, NOT ONE ERROR. Aniello Carfora thirty-one and Orsola Porrino thirty on strada Camellara; Raffaele Carfora thirty-seven and Francesca Cioffi twenty-eight; Antonio Carfora twenty-six and Giuseppa Perreca twenty-three on strada Camellara. Every age, every name, every date, both baptism columns: all of it holds at full resolution. The first volume of this row to come back clean. It gave up three things the archive had not written down. The acts have numbers — 6, 76 and 85 — and none was recorded. The parish in the second column is SANT'ANDREA APOSTOLO DI ARIENZO. And RAFFAELE CARFORA is on STRADA SIGNORINDICO, a street this archive's register already knows twenty-three times and which the household's evidence had left off: he is not on strada Camellara, where the other two Carfora are. Two Carfora houses on the Falco street and a third elsewhere. AND IT NEARLY PRODUCED A FALSE CORRECTION — MINE. Read down as far as the usual band goes, the baptism column looked like nothing but the printed formula about the parish returning the *notamento*, and the archive's quotation «*il Sacramento del Battesimo è stato amministrato a ANTONIO CARFORA nel giorno dieci Gennajo corrente anno*» looked constructed out of the column's printed heading. Three quotations were about to be withdrawn as fabrications. They are all there, lower down the same column, past where the band stopped. A band is not a page, and a false correction impugns work that was right — which this archive already knows and had to relearn inside an audit whose whole purpose is not to do that. Twelve volumes, 136 citations, twenty-three errors. `an_ua14249` (marriages 1845) — 7 citations, all on one act, and the act is the spine of generation five. Act 3 of 2 February 1845 verifies almost entirely: Domenico Zinelli Sindaco, Arienzo, distretto di Caserta, ore ventuno; Crescenzo Falco colono «domiciliato ivi strada Sant'Andrea»; GIUSEPPE FALCO and ANTONIA DI GUIDA both named without «defunto» in an act that writes it twice for the bride's parents; the church celebration on 3 February before LUIGI VERDICCHIO and NUNZIO RIVETTI. The one error is a letter, and it undoes a claim. The bride's mother is MARIA FELICIA SANSONE, not Sanzone — a long ſ, not a z, proved against «Crescenzo» four lines above, whose z has no ascender. So the surname this archive announced as one that «HAD NEVER APPEARED IN THIS ARCHIVE AT ALL» is one it was already holding: ANGELO SANSONE, contadino of strada Camellara, declared Maria Falco's death in 1826. The same wrong letter had been read three ways in three files — Sompone, Sanzone, and nothing. AND A SECOND FALSE CORRECTION WAS AVOIDED IN THE SAME ACT. The groom is «maggiore di anni venticinque», which elsewhere is a statement of legal majority and not of age — on that reading the act bounds his birth on one side only and this archive's «it dates his birth» is false. The bride, four lines below, is «maggiore di anni VENTITRE», and no age of majority is twenty-three. «Maggiore» is a status, the number is the age, the inference stands. Two near-misses in two volumes; both controls were inside the document. Thirteen volumes, 143 citations, twenty-four errors. `an_ua14322` (deaths 1820) — 7 citations, two acts, one wrong number. ANNA FALCO's death act verifies word for word, and it is one of the load-bearing acts of the whole archive: twenty-five, wife of Arcangelo Crisci, «*figlia di PASCALE FALCO … e di CHIARA RIVETTI*», neither parent marked «defunto», declared by two Martone of strada Camellara. The error is on the other act: Fortunato Crisci's death is act 6, not act 5 — and the trap is the volume's own layout, because the stamped figure in the margin is the FOLIO. Folio 5 carries act 6; folio 25 carries act 47. `an_ua14421` (births 1824) — 7 citations, three acts, NOT ONE ERROR. Even the folio stamps this archive recorded — 139, 134, 181 — are right. The second clean volume of the row. But it had been read three times and not read. Act 26 was cited here for the child, for the father and for the date, and the same sentence says «*che lo stesso è nato da ANTONIA DI GUIDA sua moglie legittima, di anni QUARANTA*». This archive held no age at all for her — generation four's wife, whose death has never been found and whose last dated sighting is 1845. Forty in March 1824, so born about 1784, and about sixty-one at her son's wedding. That is the finding of these two volumes, and it is not an error. A citation verified is not a document read, and the commonest thing this row has turned up is not a wrong word but an unread one. Fifteen volumes, 157 citations, twenty-five errors. `an_ua14264` (marriages 1860) — 6 citations, one act, and it is the best thing this row has found. The act is the marriage of SABBATINO ANNECCHINO of Forchia to MARIA ANTONIA CARFORA of Arienzo, 21 January 1860 — the spine of the Annecchino line this family descends from. A LIVING MAN HAD BEEN BURIED BY A QUOTATION. This archive published «*fu CARMINE ANNECCHINO … fu ROSA CEGLIA … fu [?]ERIO CARFORA … fu PASQUA MORGILLO*», four «fu» inside quotation marks. The act uses «fu» nowhere. It writes «defunto» and «defunta» after the name, and it writes them three times, not four: the bride's father is «*di professione contadino, domiciliato ivi*» and nothing else. He was alive. The act says so again in its own list of documents — item 6 is «*Nel CONSENSO DEL PADRE della stessa, dato nell'atto*». AND THE NAME WAS NEVER ILLEGIBLE. IT WAS SPLIT. This archive recorded that «*the word reads «[L/t]erio» at four times magnification and is left unread rather than guessed*». The word is two half-words: the line ends «*figlia di Va*» and the next begins «*lerio Carfora*». «VALERIO CARFORA» — and no Italian forename ends in *-lerio* except that one. THAT CONCLUSION WAS WITHDRAWN ON 21 SEPTEMBER 2026. The fragment is «*cesco*», not «*lerio*», and the man is FRANCESCO CARFORA; his own written consent of 1 January 1860 settles it. The elimination was sound and the four letters it ran on were not. See the corrections page, and work-list row 63. A good rule was applied to the wrong object: the archive magnified the second half of a word and called the whole of it unreadable. The act also lists the eight documents it was built on, and item 4 is a GENERATION: «*Nell'atto di morte dell'AVO PATERNO dello stesso*» — the groom's paternal grandfather, the father of Carmine Annecchino, a man this archive has never named. That is work-list row 77. Also: the act is of the twenty-first, not the twentieth; the groom writes SABBATINO with two b; and the witness is CRISCI, not Crispi. Sixteen volumes, 163 citations, twenty-nine errors. `an_ua14323` (deaths 1821) — 6 citations, one act, one error. Carmine Rivetti, fifty-three, husband of Giustina Morgillo, «*figlio di NICOLA RIVETTI Defunto … e di ANNA DI RUGGIO Defunta*» — the act that gave this archive its best candidate for Chiara Rivetti's parents — verifies exactly. The error is the second declarant's patronymic: GIOVANNI MORGILLO fu PASCALE, not «fu Raffaele». And the act is the proof of the rule this archive broke eleven volumes later: «fu» here is a patronymic on a living man, while the death mark in the filiation clause is «Defunto» AFTER the name — written twice on this very leaf. `an_ua14327` (deaths 1825) — 6 citations, three acts, NOT ONE ERROR. Maddalena Falco at seven; Giuseppa Cioffi at eight months; CLEMENTE FALCO at fifteen months, and the arithmetic closes to the day against his birth act of 11 March 1824. The third clean volume. It gives up an act number nobody recorded — Clemente's death is atto 34 — and CESARE CIOFFI, thirty, of strada Camellara, which independently corroborates the «Cesare Cioffi» read out of the 1826 volume this morning. And it hands over a second demonstration that the ages are guessed, sharper than the archive's own. ANIELLO DIGLIO is forty-six on 2 April 1825 and FIFTY on 19 June 1825 — four years in seventy-eight days, beside Giuseppe Falco's five years in four days. Two men, two pairs of acts, one volume. Eighteen volumes, 175 citations, thirty errors. `an_ua14071` (Diversi 1825) — 9 citations, one act, and the act had two men standing in its first line that nobody had written down. Everything this archive took off the leaf about the dead man verifies: CRESCENZO CRISCI, fifty-three, colono, husband of ANGELA QUOTOLO, dead in the Masseria di Domenico de Luca at the luogo FABRICA, son of ALESSANDRO CRISCI Defunto and CARMINA SORRINO, leaving Aniello fifteen, Sante twelve, Carmina ten — all exact. The declarants are his BROTHERS. «*sono comparsi DOMENICO CRISCI GERMANO DEL DEFUNTO, d'anni quarantanove, colono … e GABRIELE CRISCI GERMANO*». «Germano» is full brother, magnified before it was written down. So Alessandro Crisci and Carmina Sorrino had at least three sons, and this archive had one of them. The act was read for the family it buried and not for the family it named. The youngest child is MARIA, not «Maria Rosa». The clerk closes the list by giving the last child the family surname — «*e Maria Crisci d'anni cinque*» — and the surname was read as a forename. A child filed under a name she never had is a child no index will return. AND GABRIELE'S AGE IS NOT PUBLISHED, BECAUSE IT CANNOT BE READ. His line runs into the binding, which on this filming is in shadow. A first pass over this page had «d'anni trentuno» for him; that was the reader's eye filling a gap and it is not in the image. The near-miss is recorded because the gutter shadow is exactly where this archive will invent something if it is not careful. Nineteen volumes, 184 citations, thirty-two errors. `an_ua14326` (deaths 1824) — 5 citations, two acts, one unresolved word improved. The act this archive once had to correct verifies completely: «*PASCALE FALCO, di anni cinquantacinque, COLONO … ZIO DEL DEFUNTO, e GIUSEPPE FALCO, di anni quarantadue, colono … ZIO DEL DEFUNTO*» — both «zio», both magnified, both right — and so does the child they buried, GIUSEPPE CIOFFI, five years old, son of Francesco Cioffi and Maria Falco. Donn'Anna Quotolo's mother is NOT «d'Isteria»: the word ends -ustina with no r, and reads most likely DI GIUSTINA. The negative is settled; the positive is not. `an_ua14334` (deaths 1832) — 5 citations and FIVE ERRORS. The worst volume of the row, and the only one where every citation failed. Both image references point at other families: image 20 left is act 30, Michele Morgillo son of Nicola Morgillo and Elisabetta Crisci; image 48 right is act 79, DON FRANCESCANTONIO GESSI, seventy-eight, PROPRIETARIO. Act 32 is on image 21 left; Carmina Falco is act 81 on image 49 right. And the readings were a decade and a name out. GIOVANNA FALCO is SESSANTA, not settanta — «*ſeſſanta*», two long ſ and no crossed tt — so born about 1772, not 1762. Her husband is ANTONIO LETTIERI, not «Antonio Sett[embre]». Her mother is VINCENZA SPOSITO, not Esposito. The evidence for this volume is five terse unformatted lines with no quotation and no act cited in the original. It was harvested and never read, and it is the only volume in twenty-one where that shows in every single line. Twenty-one volumes, 194 citations, thirty-eight errors. `an_ua14336` (deaths 1834) — 5 citations, FOUR of them wrong, and one of them is withdrawn outright. The act that matters most is right. PASCALE FALCO, seventy-four, husband of CHIARA RIVETTI, son of MATTEO FALCO Defunto and FRANCESCA CRISCI Defunta, dead on 1 October 1834 — act 69, verified word for word. Only the half of the leaf was missing from the citation, and image 43 LEFT is a Nicola Rivetti a reader would have landed on. MARIA ROSA FALCO WAS TWENTY-TWO MONTHS OLD, NOT SIXTEEN YEARS. «*di mesi ventidue*», magnified. Everything anyone might have built on «sedici» — a birth about 1818, a working girl, a marriageable daughter — was wrong in the direction that invites false matches. Her mother, ANTONIA VIGLIOTTA, had never been recorded, and she opens a real question against the ANTONIA MIGLIORE married to a Michele Falco in 1842. The infant PASCALE FALCO of December 1834 is act 89 on image 54 right, not image 53 right — which is Maria Michele Sperato — and his mother ANDREANA CRISCI was in the act and not on the page. AND ONE CITATION IS WITHDRAWN. «Anna Falco, eight days old, daughter of MICHELANGELO FALCO, 7 March 1834» cites image 10 left. Image 10 left is MARIA PORRINO, eight days old, dead on the 8th of March. Five leaves around it were opened and hold no Falco; MICHELANGELO FALCO is a forename that appears nowhere else in this archive; and she is entered twice, so she has been counted twice in every Falco total this archive publishes. The person is not deleted and the citation is — row 78. Twenty-two volumes, 199 citations, forty-two errors. `an_ua14254` (marriages 1850) — 5 citations, one act, ALL CORRECT. The fourth clean volume. Act 6 of 12 February 1850: MICHELE FALCO, forty-three, colono of strada Camellara, «figlio di RAFFAELE FALCO DEFUNTO … e di CHIARA RIVETTI DEFUNTA, di professione FILATRICE», VEDOVO DI ANTONIA VIGLIOTI, marrying GIOVANNA D'AMBROGIO. Every word held except one, and it took a second pass at 3,200 pixels to see it: her age is «TRENOVE» — *trentanove*, thirty-nine — not «diciannove». The initial is a crossed ascender, a t; this hand's d, in «CHIARA RIVETTI DEFUNTA» four lines above, is a closed bowl. The act corroborates it in its own words: she is «MAGGIORE», of age, and no maternal consent is taken. Born about 1811, not 1831. The tension was legible without the image — «maggiore» and «nineteen» stood in the same sentence of this note for a day. It gave up the bride's parents, whom this archive had never recorded — DOMENICO D'AMBROGIO defunto and CARMINA MANNA, filatrice, of strada Sant'Andrea — with the sindaco Arcangelo de Fusco, the parish Sant'Andrea Apostolo, and the witnesses Alfonso Verdicchio and a Clemente Mor[gillo]. AND IT CLOSED A QUESTION THIS ROW HAD OPENED THREE HOURS EARLIER. «*vedovo di ANTONIA VIGLIOTI*» sent me back to the 1842 act, where the mother had been published as «ANTONIA MIGLIORE». Cropped and set beside the 1834 act's «Antonia Vigliotta» in the same hand, it is the same word — same capital V, same «gli» with its tall l, same «-otta». Migliore has no V and no double t. One wife, one surname, three acts — 1834, 1842, 1850 — and this archive had her in two households under two names. They are merged. Twenty-three volumes, 204 citations, forty-three errors. `an_ua14422` (births 1825) — 5 citations, two acts, ALL CORRECT. The fifth clean volume. It verifies this archive's own standing demonstration at both ends: GIUSEPPE FALCO is «di anni QUARANTACINQUE, di professione CONTADINO» on 6 April 1825, against the forty and MASSARO of 2 April in the death register. Five years and a trade in four days. It also gives ANTONIA DI GUIDA at FORTY in April 1825, thirteen months after the forty of March 1824 — two independent statements agreeing, for a woman this archive had no age for at all this morning. `an_ua14338` (deaths 1836) — 5 citations and SEVEN errors, and the citations have a fingerprint. They are written p11, p17, p29, p32 — a numbering used nowhere else here — and the true leaves are images 12, 18, 30 and 33, all RIGHT halves. Every one is exactly one higher. A constant offset means the numbers were copied from a page counter and never checked against an image; the evidence written from them is five terse lines with no quotation and no act number, the same hand found in the 1832 and 1834 volumes. Two contact-sheet sweeps found all four Falco in two requests. Then: Margarita Falco's husband is DOMENICO COSPI, not the bracketed guess «Domenico [Crisci]», and her mother MARIANGELA IULIUCCIO, not «Maria Angela Ulivieccio». PASQUALINA FALCO is SIX months old, not ten. Donna Raffaela de Falco's mother is DONNA MADDALENA SICILIANO — the first name this archive has for the wife of Don Francesco Falco the proprietario. And naming two mothers dissolved two households. SALVATORE FALCO's mother is MARIANTONIA VIGLIOTTA, so he belongs to the Michele Falco of strada Camellara; PASQUALINA FALCO's is RAFFAELA FALCO, so she is a sister of Carmina. Two «unjoined» Falco households in this volume were not families. They were lines nobody had read. Twenty-five volumes, 214 citations, fifty errors. `an_ua14420` (births 1823) and `an_ua14445` (births 1848) — 10 citations, two acts, NO ERRORS. The sixth and seventh clean volumes. Pasqua Falco's birth is ATTO 38, 30 March 1823, beginning on image 40 RIGHT: Giuseppe Falco forty-two and massaro di campagna of strada Camellara, GELSOMINA VIGLIOTTA thirty-eight, before Vincenzo Morgillo Sindaco. And the margin is there and reads as published — the 1880 marriage to Pasquale Carfora, annotated at Caserta in 1881. One more point for the age series: forty-two in March 1823, against forty-eight in March 1824 and forty on 2 April 1825. Giovanna Arricale's birth is ATTO 114, 22 December 1848, image 115 RIGHT: Pasqua Falco twenty-five, Francesco Arricale twenty-four, declared by ROSA MORGILLO, forty-eight, LEVATRICE — a midwife, not a parent. AND HER BIRTH ACT CARRIES HER MARRIAGE IN THE MARGIN, WHICH NOBODY HAD TAKEN FROM IT: «*A 18 Novembre 1872 … sposò GIOVANNA ARRICOLA con CARMINE ANTONIO FALCO*». This archive held that date from one document, the 1897 dossier of their son. It now holds it from two that never met. The act had been opened, the parents read off it, and the margin left — the same failure as the «zio» of 1826 and the two brothers of the 1825 Diversi: the page was read for what it was wanted for. Twenty-seven volumes, 224 citations, fifty errors. Four more volumes — `an_ua14366` (deaths 1864), `an_ua14250` (marriages 1846), `an_ua14265` (marriages 1861), `an_ua14443` (births 1846) — 16 citations, two errors, and one real find. `an_ua14265` is clean, including the part this archive did the hard way: STANISLAO FALCO, twenty-four, PASTORE, son of Vincenzo Falco and Andreana Crisci, marrying FILOMENA DI LUCIA, twenty-three — with her father's forename recovered from the notificazioni bound into the same volume because the marriage act itself is illegible there. Verified, and the method with it. `an_ua14366` verifies in everything it is relied on for and its act number is the leaf's. «*LUIGI FALCO = marito di CAROLINA CIMMINO … di anni QUARANTUNO, di professione INDUSTRIANTE, domiciliato Contrada Capodiconca, figlio di VINCENZO … alla CAMELLARA*» — all of it right. But the act is «Num. d'ordine 91», not 47; 47 is stamped in the margin, and the facing page is 92. Third volume in this row where a stamp was taken for an act number, and in each one the stamp counts something different. `an_ua14250` gives ONOFRIO BUONO a trade and a death: «*figlia di ONOFRIO BUONO Defunto, di professione BARBIERE*». A barber, already dead in May 1846, and his wife is ANGELICA Imbimbo, not «Angela». AND `an_ua14443` HAS A SECOND MARGIN NOBODY HAD READ. Domenica Falco's birth act carries her 1874 marriage down the left edge — quoted here for weeks — and, in the right-hand column, «*A 25 Marzo 1881 … sposò DOMENICA FALCO con PASQUALE SACCAVINO*». She married twice, and Pasqualantonio Guida died between 1874 and 1881. One birth act, two marriages, thirty-five years past the year Antenati stops filming the series — and the fourth time today that a page was read for what it was wanted for and the rest of it left. Thirty-one volumes, 240 citations, fifty-two errors. `an_ua14413` (births 1816) and `an_ua14414` (births 1817) — 8 citations, NO ERRORS, and four things nobody had taken off the pages. The ninth and tenth clean volumes. The unread child has a name. This archive entered a person as «[son] Crisci», 1817, and wrote honestly that «*the child's given name is further down the act and was NOT REACHED in this pass*». It is act 63, and the name is LORENZO ANTONIO — born 1 September 1817, a brother for ANDREANA CRISCI. The promise sat on the page while the answer sat four inches lower on the same image. And a childless household has a child. Carlantonio Morgillo and Rosa Falco had no children here at all; act 69 of 22 September 1817 gives GIOVANNA MORGILLO, and her mother ROSA FALCO at thirty-two, born about 1785. The citation had been made for her husband's age and the act read no further. Two cautions out of 1816. VINCENZO FALCO is THIRTY-FIVE in April 1816 — a birth about 1781, eleven years before the baptism of 1 February 1792 this archive read from the parish book, and two years older than the thirty-three he gives in 1824. The baptism dates him; the register does not. And GIUSEPPE FALCO and GELSOMINA VIGLIOTTA are on STRADA COSTA in 1816 and on STRADA CAMELLARA by 1823, ages agreeing across both acts — so from 1823 the street stops telling that Giuseppe Falco from the one who married Antonia di Guida. This archive leans on strada Camellara constantly. *[Corrected 21 September 2026: the date is 1819, not 1823. Act 33 of the 1819 birth register — held here already — has the same Giuseppe Falco, thirty-eight, massaro di campagna, on strada CAMELLARA on 18 March 1819, seven months before the other Giuseppe Falco declares a birth on the same street. The street has never been a discriminator in any year this archive reads.]* Thirty-three volumes, 248 citations, fifty-two errors — and the commonest finding is no longer an error at all. It is a page that was opened, read correctly, and stopped. `an_ua14332` (deaths 1830) and `an_ua14333` (deaths 1831) — 6 citations, two errors, and both are addresses. Rosa Falco is on image 37 RIGHT, not 36R — which is Costanza Majone, sixty, dead in the Ave Gratia Plena. Anna Falco is on image 20 RIGHT, not 19R — which is Carmina Maccaniello, ten months. Two sweeps found both acts in two requests. Both readings are exactly right. Rosa Falco's act verifies word for word including the part the whole line hangs on — «*figlia di MATTEO FALCO Defunto … e di FRANCESCA CRISCI Defunta*» — and the mother's surname passes this archive's own three-way test at magnification: narrow shoulder, no round o, no double-f ascender. A daughter of the founders, confirmed from the image. AND THE 1831 ACT SHARPENS TODAY'S OPEN QUESTION. This archive holds two Anna Falco, three years apart, each «figlia di MICHELANGELO FALCO», with two different mothers — 1831 MARIANTONIA VIGLIOTTA, 1834 RAFFAELA FALCO. Both mothers have now been magnified and neither reading is in doubt. A remarriage between them is excluded if that Mariantonia Vigliotta is the woman still bearing children to a Michele Falco in 1834, 1836 and 1842 whom he survived in 1850. So either there are two Michelangelo Falco at Arienzo, or one of the mothers belongs to a man of the same name. One more for the age series: Rosa Falco is forty-two in 1830 and thirty-two in 1817 — ten years in thirteen. Thirty-five volumes, 254 citations, fifty-four errors. `an_ua14415` (births 1818) and `an_ua14423` (births 1826) — 6 citations, NO ERRORS. The eleventh and twelfth clean volumes. Act 47 of 10 May 1818 (Raffaele Falco), act 57 of 7 June 1818 (Maddalena Falco) and act 69 of 12 September 1826 (Angelandrea Falco) all verify word for word — and both 1818 acts carry the clerk's own struck «sedici» with «diciotto» written above, the habit found in 1817 too. TOGETHER THEY CONVICT THE 1816 VOLUME. Vincenzo Falco and Andreana Crisci are thirty-five and twenty-five in April 1816, twenty-seven and twenty in May 1818, thirty-three and twenty-seven in 1824 and thirty-two and twenty-eight in 1826. Their baptisms — 1 February 1792 and 30 November 1796 — agree with 1818, 1824 and 1826 to within two years, and with 1816 not at all: eleven years out on him, six on her. It is not that the ages drift. It is that one volume is wrong. AND ONE OF TODAY'S OWN FINDINGS IS CORRECTED. This row published Antonia di Guida at forty in 1824 and forty in 1825 and called them «two independent statements agreeing». Act 57 gives her THIRTY in June 1818 — ten years in six. They were two guesses, not two witnesses, and calling them corroboration dressed them as something they were not. Also new: the mother declared the 1826 birth herself — «*La stessa ha in oltre dichiarato*», with the father named only as «marito della stessa» — and the witness is CLEMENTE SANSONE, twenty-six, the surname settled from three other acts today. Thirty-seven volumes, 260 citations, fifty-four errors. `an_ua14329` (deaths 1827) — 3 citations, TWO READ AND ONE NOT REACHED. Act 30 verifies: GIUSEPPE FALCO, forty, COLONO, «VICINO DEL DEFUNTO» — and the act names a kinsman this archive had not recorded, «*e CESARE CIOFFI, di anni trenta … PARENTE DEL DEFUNTO*». The same Cesare Cioffi is «Parente della Defunta» in 1828 for another child of the same couple. So he is kin to Francesco Cioffi's household, which is the nearest this archive has come to placing him. Act 52: MATTEO FALCO THE POSSIDENTE IS TWENTY-THREE, NOT TWENTY-EIGHT. «*Ventitre*» — the i dotted, the tre crossed. Born about 1804, not 1799, which moves him out of the generation this archive had put him in. And «Cossi» was found standing in a THIRD place — in the prose of this 1827 act, months after the household was corrected to CIOFFI and a gate written to hold it. ACT 84 WAS NOT REACHED AND IS NOT CLAIMED AS VERIFIED. Antenati rate-limited after nine renders and five backoffs; a 403 means wait, never missing, and an unread citation is recorded as unread. It returns to the queue. Thirty-eight volumes, 262 citations read, fifty-six errors. AND THE 1827 CITATION THAT COULD NOT BE REACHED IS NOT A RATE LIMIT. IT IS IMPOSSIBLE. «img47L atto84» — the volume has 42 images and 68 acts. Earlier attempts on that leaf returned 403, which this archive's rule reads as «wait, never missing», and the rule was right and not enough: both things were true at once, and a citation can hide behind a 403 indefinitely. Three sweeps across images 20–41 found no priest; the volume closes with three December deaths and none of them is his. The citation is withdrawn and the man is not — his sister's act of 1843 names the same parents. Row 79. `an_ua14425` (births 1828) — 3 citations, one error, and the best age in the archive. «ANDREANA CRISPI» is CRISCI — a narrow r, a long ſ, no p-bowl — and «Crispi» appears nowhere else here. Everything else is exact, including VINCENZO FALCO at THIRTY-SIX in December 1828 against a baptism of 1 February 1792: the only stated age in this household's series that matches the register exactly. The mother declared the birth herself, as in 1826. `an_ua14260` (marriages 1856) — 3 citations, CLEAN, and three things added. ANGELO MAJONE is forty (b. c. 1816), which this archive had not recorded. PASQUA FALCO is thirty-three, against a birth act of 30 March 1823 — exact. And GIUSEPPE FALCO IS NOT MARKED DEAD in an act that writes «defunto» three times over for the other parents: he was living on 9 September 1856, twenty-one months later than this archive's previous latest sighting of him. Forty-five volumes, 268 citations read, fifty-seven errors. `an_ua14127` (marriage index 1828) — 2 citations, and it ANSWERS THE BIGGEST QUESTION THIS ROW OPENED. «*10. FALCO MICHELANGELO | Arienzo | FALCO PASQUALE, e RIVETTI CHIARA* … *VIGLIOTTI ANTONIA | ivi | VIGLIOTTI FRANCESCO, e REPETELLI VITTORIA* … Xbre 1828». MICHELANGELO IS A SON OF PASQUALE; MICHELE IS A SON OF RAFFAELE — magnified on both sides. Two men. And both married an ANTONIA VIGLIOTTI, which is not a misreading but two marriage records, and is what had the households tangled. Michelangelo's died between 1831 and 1834 and he married RAFFAELA FALCO; Michele's lived until before 1850. The index had been cited on this site and never opened. It also gives her parents — FRANCESCO VIGLIOTTI and VITTORIA REPETELLI — and it caught MATTEO FALCO standing in both households at once, the third double-counted child found today. Forty-six volumes, 270 citations read, fifty-seven errors. --- AND THE ROW FOUND ITS LARGEST ERROR IN A VOLUME IT WAS NOT SENT TO. Verifying the 1814 marriage index (`an_ua14113` img 3 atto 16) turned up a live contradiction rather than a bad citation: `households.json` gave the marriage as 15 October 1814 from that index, while `line.json`, `plates.json` and `processetti.json` gave «act 26 of 15 December 1814». Both were published. Nothing in the archive compares them. The civil register settles it: `an_ua14218` img 9, left page, ACT 16, «a' dì QUINDICI 15 del mese di OTTOBRE». The volume runs its even acts down the left page and its odd acts down the right: act 14 is 2 October, act 15 is 6 October, act 17 is 2 December on the facing page, and the book stops at 17. There is no act 26 in it, and act 16 cannot be in December. An earlier session read this same image and took three elements of one line wrong — the act number, the month and the groom's age (23, not 21). What the wrong number had cost, in order: a search for «act 26's supporting file» reported on *Searched* as blocked; a *do-not-search-again* entry in the queue; and, when the dossier was found anyway at bundle 16, a false general rule — that processetti bundles run in publication order and so cannot share the marriage's number — which then had to be corrected once for a wrong illustration before being withdrawn entirely today. Bundle 16 is act 16; bundle 17 (Crisci × Morgillo) is act 17. Four errors from a 1 read as a 2. Re-reading the dossier's own consent acts moved them three months: they are 8 SEPTEMBER 1814, registered at Arienzo 12 September — not 8 December — and both were taken by Felice Morgillo, not by a second notary «Salvio Morgillo» who does not exist. The bride's consent is act 86, not act 26. The sequence now closes: consent 8 September, banns Sunday 11 and Sunday 18 September, marriage 15 October. And the act carries the only autograph in this line. «*quindi si è firmato soltanto il detto PASQUALE FALCO, avendo la detta CHIARA RIVETTI dichiarato di NON SAPERE SCRIVERE*» — and below it, in a slow unjoined hand, «PASCALE FALCO». The gate that could not see this now can. `tools/check_corrections.py` holds a list of retired PHRASES as well as retired surnames, and refuses the build if one is stated as fact anywhere in `site/src/data` or `data` outside guillemets — which is how this archive writes a reading it has withdrawn. Verified by reintroducing «act 26 of 15 December 1814» into `plates.json`: exit 1. Running tally: forty-seven volumes, 272 citations read, sixty-one errors. --- ## ROW 58 IS FINISHED — 221 citations, 36 volumes, every one checked against its image The last five volumes, seven citations, and two errors. | volume | act | verdict | |---|---|---| | `an_ua14251` marriages 1847 | 3 | correct — Angelandrea Falco «MINORE», of strada Camellara, son of Vincenzo Falco *colono* and Andreana Crisci, both alive | | `an_ua14252` marriages 1848 | 20 | correct — Francesco Falco, son of Pietro Falco *defunto* and Camilla Carbone | | `an_ua14324` deaths 1822 | 65 | correct, but the street is CAPODICONCA, not «Capodiuomo» | | `an_ua14331` deaths 1829 | 9 | correct word for word — Luca Crisci, sixty, husband of Antonia Morgillo, son of Gioacchino Crisci and Francesca Falco | | `an_ua14335` deaths 1833 | 16 | act, date and person correct; the image is 10, not 9 | And the errors of this pass were not in the sources. Every one of the five was this archive misreading its own image: a street, an image number, a digit in an index, an age of nineteen that was thirty-nine, and — this morning — an act number, a month and an age all in one line. The one structural lesson worth keeping. In these registers even acts run down the left page of each opening and odd acts down the right. That fact proved the 1814 marriage act's month without reading a letter of it, and it found the 1833 act two pages from where the citation put it. Sweep the act-number band of a whole volume before magnifying anything; it costs one call and it tests the number, the date and the volume's extent at once. Final tally: fifty-one volumes, 279 citations read, sixty-six errors — and not one of them a person wrongly identified.

A dated marriage with both names, known only from a margin. «*A 30 Dic.re 1880 l'Uf.le dello Stato Civile d'Arienzo sposò PASQUA FALCO con PASQUALE CARFORA*». The register act would confirm the chain from a second document: it names Pasquale Carfora's parents, and it should give Pasqua as vedova di Angelo Majone, which is the reading this archive is currently inferring rather than reading. First check whether Antenati holds Arienzo marriages that late at all — the death series stops at 1865 and the marriage series has never been enumerated past it. If not, collection 2043630 covers 1866–1929 and has already been mined for other years. RELOCATED, 16 September. It is NOT in the Antenati marriage series — that series ends at 1858, twenty-two years before this wedding. So the act is in the post-1865 Stato civile italiano for Arienzo, which this archive has never located for this comune, or in collection 2043630 (1866–1929), already mined for other years. Start with 2043630, since it is open and indexed. 16 September, second relocation — THE OTHER ROUTE IS MEASURED AND CLOSED TOO. FamilySearch collection `2043630` was swept twice, on Falco and on Carfora — 1,503 and 2,392 hits, 347 distinct Arienzo pages between them — and the 1880 act is quoted on none of them. The Arienzo material in that collection begins at 1891 whatever surname is run: Falco, Carfora and Majone all return the same span, 1891–1910. So the act is not filmed there, and no processetto of those twenty years reproduces it. One near miss, recorded as a near miss. A PASQUALE CARFORA, 43, agricoltore, witnesses an Arienzo birth in 1896 — born about 1853, which would make him twenty-seven at that wedding. That is a coincidence of name and age and this archive does not treat it as an identification. What is left. The comune of Arienzo's own registers, which no filming this archive can reach holds for 1866–1890; the Tribunale volumes for those years, if they were ever filmed; or a processetto belonging to a CHILD of Pasqua Falco and Pasquale Carfora, who would have had to produce their parents' marriage extract. That last one is testable inside the seam already open — see the mining row.

DONE 16 Sep 2026 — and the surname was the least of it. All four death acts re-rendered at the leaf's maximum resolution and read line by line. Every «COSSI» in them was this archive's misreading, not the register's spelling: the child, the father, and in 1828 the second declarant and the officiating eletto are all CIOFFI, by the dotted i, the round o and the tall ff ligature. The «*GIUSEPPE CRISCI(?)*» question mark of 1824 is answered — CIOFFI. The find is one word on the 1824 act. Both Falco declarants are entered «ZIO DEL DEFUNTO», which this archive transcribed «vicino del defunto». The 1823 and 1825 acts of the same household, in the same hand, write the six-letter «vicino» and are the control. So the register names the kinship four years earlier than the 1828 «parente» this archive had rested on, and in a stronger word — and it does not settle Maria Falco's parentage, because «zio» read strictly makes her a daughter of Matteo Falco and her stated age of twenty-five in 1824 forbids it. Recorded, refused, published as both. Also at the foot of the 1824 leaf: «PASCALE FALCO», his own signature, with the formula naming his brother Giuseppe as the declarant who could not write. The image numbers in all four citations were correct.

Done, and the forename list is derived rather than written. `build_register.py` now harvests the forenames this archive has actually seen — the first word of every person filed under a well-attested surname — so it grows with the data instead of being a list somebody must remember to extend. Two rules. If every word of a name is a known forename, the surname was never recorded: «*Angela Rosa*», «*Carolina Rosa*», «*Antonia Rosa, defunta*» → unknown. And if the LAST word is a well-attested forename while the first is not a surname this archive knows: «*Gaetano Antonio*», «*Joanna Maria*» — neither GAETANO nor JOANNA appears anywhere in this corpus paired with a surname, so no threshold could have reached them. Checked safe before shipping: no word in this archive is both a frequent forename AND an attested surname, so the rule cannot eat a real surname. FALCO is unchanged at 1,478 across the change. Five bad rows to zero.

DONE 16 Sep 2026 — the registers are gone, and their contents are not. Where they are not. Antenati stops at 1865 for Arienzo in every series; that was already confirmed from the year facet. Added today: «Italy, Campania, Births, from 1825 to 2000» (`M9J1-SZX`), a FamilySearch collection this archive had never searched, holds NO Arienzo pages at all — 563 hits swept for Falco, every one placed in a neighbouring comune. A clean nil, measured from `recordPlace` on all 563 rather than from the hit count. Where they are. FamilySearch collection `2043630`, «Caserta, Santa Maria Capua Vetere, Civil Registration (Tribunale), 1866–1929». Its Arienzo material is marriage supplements, proclamations and bann supplements, 1891–1910 — *processetti*, and a processetto binds in certified copies of the civil acts of the couple and their dead parents. So the lost registers of 1866–1890 survive inside the surviving registers of 1891–1910. One name — Falco — returns 1,503 hits, 202 distinct Arienzo pages, and seventeen certified extracts of Arienzo acts of 1871, 1872, 1873, 1874, 1875, 1878, 1879, 1880, 1881, 1883, 1887 and 1889. And they land on tree names that had no documents. `MATTEO FALCO × ALESSANDRA CRISCI` (m. 24 Jan 1834) stands in the tree with nine children and `records: 0` against every one. Three of the nine now have an act, and the two sources are independent: the extracts name Filomena Falco FU MATTEO as wife of Antonio Crisci, Vincenza Falco FU MATTEO as wife of Carmine Morgillo, and Giuseppe Falco, 46, as husband of Nicoletta Nuzzo — which is exactly whom the tree had married them to. Matteo Falco was therefore dead before 11 August 1872. Nothing has entered a household. Every line is a machine transcription of a handwritten copy, two removes from the register — a candidate, exactly as a contact sheet is. The arks are citable and the images are unread. Method, every extract and every ark are in `notes/arienzo-1866-1890-via-processetti.md`; the reading is the new row. Rows 39 and 40 stay open. Neither Vincenzo Falco of Andreana Crisci nor Giuseppe Falco of Antonia di Guida is in the seventeen — but the instrument that would hold them is now identified and only one surname has been run through it.

16 Sep 2026 — THE CARFORA ARE AN ARIENZO FAMILY, AND THEY LIVED ON STRADA CAMELLARA. The whole of `an_ua14428` (Arienzo, Nati 1831) was swept on the declarant line and three separate Carfora households appear in that single year, all read from the image: - ANIELLO CARFORA, 31, contadino, STRADA CAMELLARA × Orsola Porrino, 30 — son Antonio, 10 January 1831 (img 7r-8l). - RAFFAELE CARFORA, 37, contadino × Francesca Cioffi, 28 — son Pietro, 12 September 1831 (img 78r-79l). - ANTONIO CARFORA, 26, contadino, STRADA CAMELLARA × Giuseppa Perreca, 23 — daughter MARIA ROSA, 4 November 1831 (img 87r-88l). Two Carfora houses on the Falco's own street. The family this archive assumed belonged to Forchia was living fifty yards away the whole time. MARIA ANTONIA IS NOT AMONG THEM, AND THIS IS NOT YET A NEGATIVE. The November girl is MARIA ROSA, read at magnification, and none of the three wives is PASQUA MORGILLO. But the sweep read the «*è comparso*» line only, and a large minority of acts in this volume are declared by a LEVATRICE — Rosa della Selva, Caterina Ruggiero, Gelsomina Ruggiero appear again and again — whose acts name the father further down the page. The sweep is triage, exactly as a contact sheet is, and it cannot close 1831. Next, and cheaper than another birth volume. She was twenty-nine on 20 January 1860, so 1829–1832. The marriage TAVOLE give both spouses and BOTH sets of parents on one or two leaves a year — `an_ua14108` + (year − 1809), three or four images each. 1830 was read and holds no Carfora groom (Carfora appear there only as mothers — Marta Carfora, Anna Carfora). 1829 was read rotated, as triage, and no Carfora groom was seen. 1826, 1827 and 1828 are rendered and NOT read. The format changes every year and some are written sideways. Then the 1860 processetto, which would carry her birth extract outright — the same instrument that row 62 has just proved reaches years the registers no longer do. --- ## AND THERE IS A GENTRY CARFORA, A GENERATION EARLIER — 20 September 2026 Found in the half of the 1815 marriage register nobody had opened (row 81), not by looking for Carfora at all. Act 20 of `an_ua14219`, 18 November 1815: > «*è comparso il Signor PASQUALE CARFORA di anni quaranta‑, di professione MEDICO CERUSICO, domiciliato nella comune di SANTA MARIA A VICO, circondario di Arienzo, figlio legittimo delli fu Signori PAOLO CARFORA ed ANNA SABATELLI, morti in detta comune, il primo a sei Novembre dell'anno millesettecentottantadue, e la seconda a' dì ventidue Gennajo dell'anno millesettecentottantanove, come consta dagli atti di morte a noi esibiti*» — marrying «*la Signora ALESSANDRA DE ANGELIS, di professione GENTILDONNA, nativa di NAPOLI, domiciliata in questa suddetta comune Corpo, strada Terra Murata, figlia maggiore delli Signori FABIO DE ANGELIS, di anni settantaquattro, di professione gentiluomo … e della Signora ALIMPIA GUARDIA*». This is the earliest dated Carfora couple this archive holds by twenty years — Paolo Carfora dead in 1782 — and it is a different social register from the three *contadino* households of strada Camellara found in the 1831 births. A *medico cerusico* marrying a Neapolitan gentlewoman is not the same world as Aniello Carfora the contadino, and nothing joins them. What it changes for this row. The row was raised because the archive did not know there were Carfora at Arienzo to look for. There are now two kinds, and the search for Mariantonia Carfora — generation seven's probable grandmother — has to say which it is looking in. The 1860 marriage act that names her gives no trade for her people; the contadino houses of strada Camellara remain the likelier ground, and the gentry line is recorded here so that a later reader does not mistake one for the other. Neither is claimed. --- ## ANSWERED — 21 September 2026. FRANCESCO CARFORA AND PASQUA MORGILLO, OF STRADA SIGNORINDICO. Maria Antonia Carfora was born at Arienzo on 30 December 1830, act 142 of `an_ua14427`, at seven in the morning, in her father's own house on strada SIGNORINDICO — not on strada Camellara, where this row spent a week looking. Baptised at Sant'Andrea Apostolo the same day. Father FRANCESCO CARFORA, twenty-eight, contadino; mother PASQUA MORGILLO, twenty-six. Witnesses GENNARO MORGILLO, thirty-three, of strada Cappellotto della Porta di Sopra, and CRESCENZO CRISCI, thirty-five, of strada Crisci. She is twenty-nine on 21 January 1860 and she had turned twenty-nine three weeks before. ### The three marriage tavole this row left rendered and unread They were read, and the 1828 one carried the answer. - 1826 (`an_ua14125`, 18 acts): CARFORA GIOVANNI of Santa Maria a Vico, son of *Carfora Matteo e Crisci Diamante*, × QUOTOLO LUCIA of Arienzo, 29 June 1826. No Valerio and no Francesco. - 1827 (`an_ua14126`, 20 acts): act 6 CARFORA VINCENZO of San Felice; act 14 CARFORA LORENZA as bride, father *Carfora Angelo*; and act 16, 19 August 1827, CARFORA ANTONIO × PERRECA MARIA GIUSEPPA — the strada Camellara household this archive holds from one 1831 birth act, with both sets of parents. *The index is not the act and act 16 has NOT been opened*; nothing from it has been written into that household. - 1828 (`an_ua14127`): act 11 — CARFORA FRANCESCO (father CARFORA MATTEO, mother ZIMBARDO GRAZIA) × MORGILLO PASQUA (father MORGILLO GIOVANNI, mother CRISCI DOMENICA). The act behind it was opened — `an_ua14232` img 12R, foglio 214, 29 June 1828 — and it gives the street, the ages, and «*vedovo di MARTA CRISCI*» in a clause the printed form does not assign cleanly to either man. The church celebration followed on 30 June 1828 at Sant'Andrea Apostolo. ### The forename was this archive's, not the clerk's The groom is FRANCESCO; the 1860 marriage act had been read as «VALERIO». The disagreement was treated as a fault in the reading rather than in the find, and two further documents settled it — the birth act of 1830, which names him twice, and his own consent of 1 January 1860, recovered from the 1860 processetti (row 77). See the corrections page. ### The nil this row can now publish She is not in the 1831 births. `an_ua14428` was re-read end to end at the parents band, all 100 images, hunting PASQUA MORGILLO specifically — the first pass swept only the declarant line, which is why this row could not close before. No Pasqua Morgillo in the year, and no Carfora father named Francesco or Valerio. The sweep did find three more Carfora mothers at Arienzo in that single year — ELISABETTA CARFORA, thirty-six; ANNA CARFORA, thirty; CARMELA CARFORA, thirty, of strada Santa Lucia — beside the Giuseppa Perreca act already held. Six Carfora households in one year of one register. ### What is left, and what it is worth The gentry Carfora of Santa Maria a Vico stay unconnected. Pasquale Carfora *medico cerusico*, son of Paolo Carfora dead in 1782, is a different social register from a contadino of strada Signorindico, and nothing joins them. Act 16 of `an_ua14231` was read the same day and is written into the Antonio Carfora × Giuseppa Perreca household: Pasquale Carfora and Maria Carbona on one side, Nicola Perreca and Orsola Morgillo on the other, and the discovery that the two witnesses this row kept meeting are the parish's standing pair. Still owed, and small: Marta Grifo's death act, act 33 of the 1827 Arienzo death register, already read here as «*Marta Grifo, 30, moglie di Francesco Carfora*» — if she is the wife the 1828 act calls Marta Crisci, one of the two surnames is wrong, and that is a reading to magnify, not a conclusion to draw.

Done, and the answer is that they do not exist for the year that needed them. The notificazioni register is bound into the 1861 volume only — `an_ua14265`, 76 images for 29 acts, the rest banns. 1855 (`14259`, 30 images) and 1856 (`14260`, 57 images) are acts to the end: image 24 of 1855 is act 22, image 46 of 1856 is act 20. No banns register in either. 1859 (28 images/25 acts), 1862 (30/27), 1863 (24/21), 1864 (28/~25) and 1865 (39/~36) have no room for one. So the practice appears with the 1861 volume and not before — which is the year Italy was unified, and a plausible reason. It cannot reach NUNZIATA FALCO's marriage of 13 November 1855. What the 1861 banns DID give: both sets of parents in a clerk's cleaner hand, which supplied LUIGI DI LUCIA where the marriage act's own writing would not resolve him. Read them for every year from 1861 on when a name will not come off an act.

David opened a one-month full subscription on 16 September and asked whether FindMyPast had been run through. It had — twice, both signed in (9 and 13 September), which found the six passenger-list records, established the Italian sets as the IGI re-hosted, and returned zero Annecchino. But those covered six surnames only. The branch surnames were never searched. TWELVE NOW SWEPT (Australasia, variants on, page one only — totals not captured): Falcone 21 · Biondo 21 · Mazza 21 · Zavaglia 21 · Carfora 21 · Cimmino 20 · Zaino 20 · Crisci 20 · Militano 20 · Vigliotta 9 · Abatiello 5, every row a UNITED STATES WW2 record mis-tagged into the Australasia filter · ARRICALE 0 — «No results found», the one TRUE NIL here. ONE QUEENSLAND FIND: CONCETTA MILITANO on the Australia Electoral Rolls at INNISFAIL — 1939, 1941, 1949, 1959. A lead, not a link: Innisfail is North Queensland, fourteen hundred kilometres from Brisbane, and matches the Maione cluster at Winton and Charters Towers that this archive already logged as unconnected. Militano is a living branch surname, so the living-people rule binds anything published. Five record sets found that the September note's «holdings worth remembering» does not list: NSW Deceased Estate Files 1880–1923; Australia Parish Deaths & Burials 1814–2014; Victoria Wills & Probate; Victoria Petty Sessions Registers; Britain & Ireland Incoming Passenger Lists. REMAINING: page two onward for every surname above; and Isla, Reid, Halliday, Ellevsen, Dreghorn, Morgillo, Diglio, Rivetti, Cioffi. The domain is `findmypast.co.uk` — see `notes/findmypast-2026-09-16-method.md`; `.com.au` silently returns a sign-up page.

DONE 16 Sep 2026 — 273 lines of prose are now 61 rows, and the conversion caught three stale sentences. `site/src/data/searched.json`, rendered through `kit/components/Searched.astro`. 61 searches, 20 of them nulls, and the page now says so in a line a reader can check rather than a claim buried in prose. THE OUTCOME VOCABULARY IS THIS ARCHIVE'S AND IT IS SIX WORDS, because six different things happen to a search here. The kit is explicit that flattening these to a shared five would be inventing a finding, and it would be: - found (37) — something new came back - no Falco (10) — the volume was read end to end and the family is not in it - not there (7) — the one record sought is missing from a source that holds others - nothing (3) — the source holds nothing for this family at all - blocked (3) — the film is obstructed or the volume was never digitised, which is not a negative, it is an obstacle - not searched (1) — declined deliberately, reasoning in the row The nulls are «no Falco», «not there» and «nothing» — twenty of sixty-one, and that is the number the page exists to publish. AND THE CONVERSION DID WHAT AN AUDIT DOES. Three sentences on this page were still carrying the reading corrected this morning: the 1824 entry said Pasquale Falco was «*declaring a neighbour's death*» when the register says ZIO; the 1825 entry said «*no act ever says they are kin*» when one does; and the 1828 entry called «parente» «*the word it had waited six years for*» when it was the second and the weaker. A fourth file carrying a retired reading, found only because the page was being touched for another reason. Two implementation notes for the sibling archives. The kit takes `what` and `got` as HTML the archive has already rendered, so they go through this archive's `md()` before being handed over — escaping them turns every `` into literal asterisks. And `src` is** escaped, so markdown must be stripped from it, not left in.

DONE 16 Sep 2026 — 162 lines of prose are now 37 questions, and two of them had to be re-answered on the way. `site/src/data/questions.json`, rendered through `kit/components/Questions.astro`. 37 questions: 20 open, 6 blocked, 11 answered. The answered ones stay on the page on purpose — the reasoning that settles a question is worth as much as the answer, and the filter keeps them out of the way of anyone working the queue. THE STATES ARE THIS ARCHIVE'S OWN THREE. *Open* — nobody has answered it. *Blocked* — a person has to do something a machine cannot: order a file, walk into a library, write a letter. A blocked question is not an unanswered one, and the old page kept them in separate sections rather than letting a reader see them together. *Answered* — settled, and left standing. AND THE EXTRACTION FOUND TWO WRONG ANSWERS, which is the second time today that converting a page audited it. «Is the third Maria Falco kin?» was answered «*Closed by the fifth act*» — the 1828 «parente». It was closed by the SECOND act. The act of 25 March 1824 enters both Falco declarants as «ZIO DEL DEFUNTO», four years earlier and a stronger word, and this archive had transcribed it «vicino». The entry also said «Francesco COSSI» and «Cesare Cossi» — both are CIOFFI. Rewritten, with the reason the pedigree is still refused. «TWO sons called Giovanni» ended on a puzzle: the 1841 civil register held no Falco, «*the days on either side of the 27th are accounted for*», so «*it is now more likely that the 1841 transcript is wrong*». The days are not accounted for. Acts 86 and 87 record deaths of 26 November and 7 December — eleven days with nothing registered. The puzzle dissolves and the two-Giovanni reading stands unopposed. Same implementation note as row 66: every field arrives at the kit as HTML the archive has already rendered, so all four go through `md()` first.

DONE 16 Sep 2026 — and the versions this row named were mostly wrong. The row proposed «checkout@v5, configure-pages@v6, setup-node@v5, upload-artifact@v5», written from the deprecation annotation rather than from the registry. Asked the GitHub API for the latest release of each action instead of guessing, which is the same discipline as reading a register's own tavola rather than its catalogue label: | action | was | now | row's guess | |---|---|---|---| | `actions/checkout` | v4 | v7.0.1 | v5 | | `actions/setup-node` | v4 | v5 | v5 ✓ | | `actions/configure-pages` | v5 | v6 | v6 ✓ | | `actions/upload-pages-artifact` | v3 | v5 | named as *upload-artifact*, which this workflow does not call directly | | `actions/deploy-pages` | v4 | v5 | not named at all | Release notes read before bumping. Every major here is a Node-runtime and dependency upgrade; none changes an input this workflow passes — `checkout` and `configure-pages` are called bare, `setup-node` with `node-version`, `upload-pages-artifact` with `path`, `deploy-pages` with an id. `node-version: 22` is left alone: that is the Node the site builds on, not the Node the actions run on, and 22 is current LTS. Two actions the row never mentioned were also on old majors — `upload-pages-artifact@v3` and `deploy-pages@v4`. A deprecation annotation lists what the runner noticed, not what is out of date.

Row 62 found the seam; this row works it. FamilySearch collection `2043630` holds 202 distinct Arienzo pages naming a Falco, 1891–1910, and their certified extracts quote Arienzo civil acts from years Antenati does not hold. One pass on one surname recovered seventeen. Three jobs, in order. 1. Read the images. Every extract in `notes/arienzo-1866-1890-via-processetti.md` carries its ark. A machine transcription is a candidate and nothing more — the OCR renders Arienzo as «Chienno», «Criengo» and «Artento» on the same pages. Nothing enters a household until the image is read. Start with the three that touch Matteo Falco × Alessandra Crisci, whose nine children have no documents at all: `3:1:3QS7-897N-5R1` (Filomena, 1872), `3:1:3QS7-897N-16SB` (Vincenza, 1874), `3:1:3QS7-897N-P3N5` (Giuseppe & Nicoletta Nuzzo, 1887). 2. Sweep the other surnames. Only `q.text=Falco` has been run. Crisci, Guida, Arricale, Annecchino, Carfora, Cimmino, Morgillo, Zampiello, Rivetti, Cioffi have not. 3. Sweep the wording, not the name — `q.text=Estratto` enumerates every processetto in the collection whoever it belongs to, which is how the 1866–1890 acts get counted rather than sampled. The prize. Rows 39 and 40 want the deaths of Giuseppe Falco of Antonia di Guida and Vincenzo Falco of Andreana Crisci, both after 1865 and both unfound. A grandchild marrying in the 1890s would have had to produce exactly those extracts. FIRST RESULTS, 16 September. A second sweep on `q.text=Carfora` brings the corpus to 347 distinct Arienzo pages, and three things fell out of it: A FOURTH child of MATTEO FALCO × ALESSANDRA CRISCI. The 1908 proclamation names «*FERRARA AGNESE, 25, figlia di GIOVANNI, 60, colono, e figlia di MARIA FALCO, contadina, residente in Arienzo*» — and the tree holds Giovanni Ferrara married to Maria Falco (Ferrara), daughter of Matteo and Alessandra, with `records: 0` against both. Both alive at Arienzo in 1908, and a granddaughter the tree does not have. THE CORSO CAUDINO HOUSEHOLD CLOSES ON ITSELF. Vincenzo Falco × Angela Pellone, of Corso Caudino 80 — a birth in 1880, his death in 1889 — now has a daughter: «*Falco PASQUALINA, 21, figlia del fu VINCENZO … e di PELLONE ANGELA*», marrying Crisci Raffaele in 1899. Three acts, one household, two separate sweeps. A THIRD HOUSE NUMBER ON STRADA CAMELLARA. A 1906 supplement quotes an Arienzo birth act of about 1879: «*è comparso PASQUALE FALCO, contadino, di anni trentanove … nella casa posta in VIA CUMELLARA AL NUMERO UNDICI, da MARIA AMALIA GUIDA*». That is this archive's own household, and the age agrees with his death act — but he dies at number FIVE, and the Morgillo are at number NINE in 1881. Three numbers where the archive had one, and no reason yet to think they are stable. All of it is still machine transcription. The images are unread.

A WHOLE SERIES THIS ARCHIVE HAD NEVER OPENED, found on 17 September while looking for something else. Its title page: «*Atti per le nascite avvenuti in un viaggio di mare, Per le Ricognizioni, Adozioni, PERSONE DEFUNTE E FUORI DEL PROPRIO DOMICILIO, e Morti-Nati*». When an Arienzo person died elsewhere, the comune where they died drew the act and sent a certified copy to Arienzo, which copied it in whole — age, trade, street, both parents, whether they were dead, and the surviving children with their ages. So a death that is not in the Arienzo death register may be here, because the person did not die at Arienzo. That is exactly the shape of rows 39 and 40. GIUSEPPE FALCO of Antonia di Guida is alive in November 1849 and in no volume since; ANTONIA DI GUIDA is alive in February 1845 and in none either. Read 1849 onward first. The series is small. Forty years — 1703, 1812–1813, 1820, 1822–1835, 1837–1840, 1842–1850, 1853, 1855–1860, 1864–1865 — at one to eight images each, so the whole thing is about the size of a single year of deaths. THE ARK ARITHMETIC IS NOT LINEAR AND MUST NOT BE GUESSED. `an_ua14056`–`14062` run 1835, 1837, 1839, 1840, 1846, 1848, 1849 — ascending with gaps — but `an_ua14071` is 1825 and `an_ua14095` is 1848, so at least two sub-series are interleaved in the range. Probe each ark and read the breadcrumb. Antenati rate-limits hard: seven probes 0.8 s apart earned a 403 on the eighth. 1825 is read (`an_ua14071`, two acts): a colono who died at a masseria in the Nola district, and a woman of fifty-five who died in the Real Stabilimento delle Donne Matte at Aversa with her parents left blank. Method, arks and both transcriptions: `notes/arienzo-diversi-series.md`. --- 17 SEPTEMBER — THE MAP IS COMPLETE AND THE READING HAS STARTED. All 45 volumes located, 1812–1865, ark by year, in five requests — and the table published here yesterday was wrong. It listed `an_ua14056`–`14062` as Diversi, from probing arks and reading the breadcrumb year; those arks are a different series and the assumption was mine. Do not probe arks. `/search-registry/?localita=Arienzo&tipologia=Diversi` renders `<li class="search-item" data-id="NNNNN">` with `Registro: YYYY` beside it — ten a page, `&s_page=N` to continue. (`&anno=YYYY` breaks the page entirely.) The real block is `an_ua14064`–`14106`, plus `an_ua57637`–`57638` for 1813. A METHOD TRAP, FOUND THE HARD WAY. Each act is headed «*Atto di morte fuori domicilio di [NAME]*» and in 1849–1850 that heading is the perfect sweep band. From 1856 the clerk leaves the printed heading blank and writes only the body, so the heading sweep reported «no acts» for a volume holding five. Sweep `--band 0.06 0.30 --half LR`. The reading of 1855 as empty is void and must be redone. READ SO FAR. 1849 — three acts, all «morte fuori domicilio», no Falco. 1850 — «*ESTRATTO DI MORTE*» from the Comune di AIROLA: «*è morto nella strada Mercato NICOLA D'ADDIO, nato in ARIENZO, di anni ventisei, Bracciale, domiciliato in Airola DI PASSAGGIO, figlio di PIETRO, Bracciale, domiciliato in Arienzo, e fu FLAVIA TANOLLA*». Exactly the record this series exists for — a man of Arienzo who died in the next town and is invisible to every Arienzo death sweep. 1853 — two deaths transmitted from NAPLES, one from the Sezione Montecalvario, plus local acts for Carmela de Roggio and Antonio Rivetti. 1856, 1857 — acts present, bodies unread. STILL TO READ: 1855, 1858–1865, the bodies of 1853/1856/1857, and the whole of 1812–1848. Neither GIUSEPPE FALCO nor ANTONIA DI GUIDA has appeared yet, and nothing about rows 39 and 40 may be concluded until the series is finished. --- 17 SEPTEMBER, SECOND PASS — AND THE SERIES IS NOT WHAT ITS NAME SAYS. `localita=Arienzo` RETURNS VOLUMES FROM SEVERAL COMUNI. More than one comune in this district has «Arienzo» in its name, and the holdings list matches all of them. Read from the manuscripts: `an_ua14064` and `14069` and `14071` are «*Comune di CORPO D'ARIENZO*»; **`an_ua57637` and `57638` (1813) are «*Comune di SAN FELICE D'ARIENZO*»; `an_ua14068` (1824) is «*Comune di SAN CASALE DI ARIENZO*». The printed title page gives the CIRCONDARIO — Arienzo — at the top, and that is the district, not the town. The comune is the line below it. A sweep that reads the top of the page will call every volume Arienzo and be wrong about a third of them, and this archive has a standing refusal to claim the San Felice Falco as kin. Read the comune before attributing anything. READ SO FAR — no Falco, no di Guida, anywhere. 1812 two death acts, both CRISCI, natives of Arienzo: MARCO CRISCI received from the Sindaco of ACERRA, and a second received from MONTESARCHIO. 1820 one stillbirth. 1822, 1823, 1824 (`14069`), 1825 (`14070`) — no acts at all, title page and ministerial certificate only. 1824 (`14068`, San Casale) one act. 1825 (`14071`) the two already published. 1849 three acts. 1850 Nicola d'Addio, died at Airola. 1853 two deaths from Naples. STILL TO READ: `an_ua14072`–`14095` — twenty-four volumes, 1826 to 1848, about 100 images — then 1855, 1858, 1859, 1860, 1864, 1865, and the bodies of 1853, 1856 and 1857. Roughly half the series. Nothing about rows 39 and 40 may be concluded until it is finished. --- FINISHED 20 September 2026 — 45 volumes, 1812 to 1865, and there is no Falco in any of them. Every act from 1845 to 1865 has been read and its dead named — which is the window rows 39 and 40 need, since Antonia di Guida is alive on 2 February 1845 and Giuseppe Falco on 26 November 1849. Neither died away from Arienzo in a way this register recorded. Three acts in the whole series remain unnamed — two in 1835, one in 1839, one in 1840 — and all are before 1845, so none can be either of them. What the series turned out to be. About thirty-five deaths away from home for Corpo d'Arienzo, mapping where an Arienzo labourer went and did not come back from: the Ospedale degli Incurabili at Naples (5), Sant'Agostino degli Scalzi at Aversa (3), the Real Stabilimento de' Matti and its women's side at Aversa (6), Acerra (3), Caserta (2), Airola (2), a masseria in the Nola district, Montesarchio, Santa Maria Maggiore, the Ospedale della Trinità, and — once — LIPARI. The four that are worth a page of their own: - PASQUALE PORRINO, 52, married, a labourer born at Arienzo, dead 31 October 1864 in the hospital of the DOMICILIATI COATTI on LIPARI — internal exile, three years after Unification. - CARLO BALLETTA, soldier of the 1° Reggimento Dragoni, dead in the Ospedale della Trinità on 3 March 1855; his colonel wrote asking the sindaco to tell the family to send someone for one ducat and thirty-five grana, the proceeds of his kit. - ANGELO IMBIMBO, 60, dead in the Aversa madhouse on 30 November 1855 — and the 1845 volume has FRANCESCO IMBIMBO, «figlio d'ANGELO», dead in the same asylum on 4 November 1844. Father and son, eleven years apart. - PELLEGRINO MORGILLO**, 26, «*vedovo di MARIANTONIA D'ADDIO*», dead at Aversa in 1828 — a year after the Giudicato Regio's letter about his wife's body found buried in a wood. Two surnames are left deliberately unread, both flagged rather than guessed: the RI[?]ETTI of January 1845 («*figlia del fu MATTEO*») and the [T/D]IGLI of December 1855. Neither is Falco. Both are worth settling, because Rivetti and Diglio are families this archive holds. Method, every volume, every act and every ark: `notes/arienzo-diversi-series.md`.

Found while auditing for the Mazza fault. `site/src/data/processetti.json` is read by no build script and no page — `processetti.astro` carries its content inline instead. Its own note says «*Written from data/antenati-processetti.tsv; the key is the town and does not move*», and that TSV is read by nothing either. It is not the Mazza fault: the key is the town, not a person, so no one is being told they are undocumented. But it is a file the archive believes is live and is not, and its row for Arienzo carries a finding — the 1814 dossier at bundle 16, images 265–274 — that would silently rot. Either wire it into `processetti.astro` or delete it and move the note into the page. Leaving a data file that nothing reads is how the next person comes to trust a stale number. DONE 20 Sep 2026 — wired in, not deleted, and it turned out to be the only honest copy of a number this archive had written four ways. `processetti.json` now feeds `processetti.astro`: the page derives every volume count from the six rows and renders them as a table. The four totals were all arithmetically correct and all counted a different set of towns, none of them labelled: | number | what it actually counts | |---|---| | 378 | all six towns | | 321 | the five of the Valle Caudina — excludes San Felice a Cancello | | 258 | the four in Benevento — excludes Arienzo as well | | 320 | 321 less the single bundle that has been read | So there was nothing to correct arithmetically and everything to correct in the labelling, and the page now says which towns it means every time it quotes a total. One statement was simply wrong and is fixed: */searched* said the processetti were «*258 volumes across four towns, of which none have been read*». One bundle has been read — Arienzo 1814, bundle 16 — and Arienzo is not among the four. It now reads 378 across six, of which exactly one bundle has been read, and links to the table.

The only record in the whole of the UK National Archives that touches this family's town. «*Special Investigation Branch investigation: alleged murder of Herron FG, Private, 1583838, Argyle and Sutherland Highlanders, by two Italian civilians at ARIENZO, Italy, 11 March 1945*». 1–31 March 1945, open since 25 August 2017, held at Kew, not digitised. IT IS A LEAD ABOUT ARIENZO, NOT ABOUT THE FALCO, and it is recorded as that. No Falco is named in the catalogue description and none may be assumed. What makes it worth ordering is that this archive holds almost nothing from Arienzo in the 1940s — the civil registers stop at 1865, the processetti at 1865, and the Australian records begin in 1964. A military police file names its accused and its witnesses, and its witnesses would be townspeople. It needs a person. Kew will copy it on request, or it can be read in the reading room. That is why this row is blocked rather than next.

The marriage is act 3 of 2 February 1845, Arienzo — `an_ua14249` img4R. Its *processetto* must contain a death certificate for GENNARO MORGILLO, because the bride is described as an orphan of both parents, and a Neapolitan marriage dossier of this date proves that with the civil death act itself. That certificate will quote an act number. If it is 33 of 1843, the man who died a widower on strada Camellara on 4 August 1843 — son of Angelo Morgillo, colono of the Monticello de' Cappuccini — is Crescenzo Falco's father-in-law, and the Falco and the Morgillo of the hill are joined by a document instead of by four agreeing details. It would also read a surname this archive cannot. The 1843 act calls his wife «Felicia S___»; the archive published «Sompone» and the 1845 marriage gives «SANSONE». At the leaf's full resolution — 1,622 pixels across a half page, which is where this filming stops — the word will not separate. The processetto would carry her death act too. And it is the same bundle that would answer row 40's other question, since a 1845 dossier for an Arienzo bride quotes the birth act as well: Angela Maria Morgillo is twenty-three in February 1845, so born about 1822, in a birth register this archive has not read. --- ## ANSWERED — 21 September 2026. IT IS ACT 33, AND NOTHING HAD TO BE ORDERED. This row opens with the word «Order». It was the second row in one day to be wrong about that. The Arienzo *processetti* for 1845 are `an_ua14190`, 432 images, free on Antenati; bundle 3 is images 27–39; and it took four ark probes and two contact sheets to get there. See row 77 for the same lesson arriving first. ### The question this row was raised to ask > *«If it is 33 of 1843, the man who died a widower on strada Camellara on 4 August 1843 … is Crescenzo Falco's father-in-law.»* «MUNICIPALITÀ DI ARIENZO — Estratto di Morte del 1843 — Num. d'ordine TRENTATRE», issued «*per uso di matrimonio*» on 22 November 1844, filed twice in the bundle at images 31 and 32. PROVED. The Falco of strada Camellara and the Morgillo of the Monticello de' Cappuccini are joined by a document. ### The second question, and it was already answered elsewhere This row wanted the processetto to read a surname the 1843 filming cannot resolve. **It does read it — «*vedovo di FELICIA SANSONE*» — but the better witness is the bride's own birth act, which the dossier also named: Arienzo, Nati 1822, act 66**, where the 1822 clerk splits the word at the syllable, «*MARIA FELICE SAN=/SONE*». Three documents in three hands. «Sompone» is finished. ### The third question, thrown in > *«a 1845 dossier for an Arienzo bride quotes the birth act as well: Angela Maria Morgillo is twenty-three in February 1845, so born about 1822.»* Born 2 August 1822, act 66, at seventeen hours, in her father's house on strada VALLONE — read from the register, not from the dossier's copy. Which makes her twenty-TWO when she married, and the dossier's copy of the promise says «*ventidue*» where the register's own act 3 says «*ventitre*». *Two copies of one act, one year apart on the same woman.* ### What came out that nobody asked for GIUSEPPE FALCO WAS A COLONO, and both he and ANTONIA DI GUIDA are unmarked — alive — on 2 February 1845, in an act that marks the bride's parents dead in the next sentence. The banns were posted on Sunday 24 November 1844, ten weeks before the wedding. Angelo Morgillo had a wife, «*MARIA G———*», named in the death extract and in the register act, and her surname will not resolve. And the register row for act 33 was wrong in four fields at once — the date, the first declarant's age, the second declarant's surname, and the wife's. All four came from the sweep that found the act. See the corrections page.

The research log had been generated by a script that never wrote the key the work list points at. `tools/build_researchlog.py` produced rows without a `key` column while `checkcovers.py` resolves every «covers: log:NAME» against exactly that column. So the file on disk carried keys that nothing could have put there, and the moment it was regenerated all thirty coverage links went blank and the build refused. The generator is fixed; the key is the filename stem, lowercased, which is how the committed file had always spelt it. What the fix exposed is larger than the fix. The log held 40 rows. There are 87 files in `notes/`. Forty-seven of this archive's own working files — the death of Matteo Falco in 1802, the parish marriage indexes, the processetti mining, every FindMyPast harvest, the Zampiello of Arpaia — had never reached the research-log page at all, because the log had not been regenerated since they were written. Forty-four of them now count as outstanding and none of them has a row. Most are not outstanding at all: they are finished findings, and the generator calls a file incomplete unless it contains the literal word COMPLETE, which only the per-volume sweeps use. `unreconciled` is set to 44 deliberately — the gate's own ratchet — and the number is printed on every build so it cannot be forgotten. The work is to walk the forty-four, attach each to the row that owns it or mark the file finished in its own words, and take the number down. It is the Mazza fault in a second place. A generator quietly stopped writing a column a consumer keys on, and nothing refused until a gate was pointed at it. ## DONE, 20 September 2026 — and the backlog was never forty-four The account was measuring the wrong thing. «Research files not finished» was pointed at `not complete`, and `complete` answers a different question: was this REGISTER read end to end. A FINDING file has no reason to contain the word COMPLETE, so all fifty-three of them — the death of Matteo Falco, the parish marriage indexes, every FindMyPast harvest, the Zampiello of Arpaia — were being counted as outstanding work. Forty-four closed pieces of work sat in a backlog they never belonged in. `build_researchlog.py` now writes `outstanding` separately, and the definition is a real convention and not a guess: a file is outstanding when it is a per-volume SWEEP that has not finished. This archive names those `*-progress.md` — thirty-four of the eighty-seven files — and nothing else. The number fell from forty-four to one, and the one was real: the Arienzo civil marriage register for 1815, images 17 to 30, never opened — about twelve marriages, half the year, in the densest kind of volume this archive reads. That is now row 81. `unreconciled` is 0. The work list accounts for 26 of 26 outstanding items: twenty-three unfinished register sweeps and three towns whose processetti are unexamined. The ratchet can only tighten from here. The lesson is the one this whole day has been about. The gate was right that something was uncounted; it was the archive's own declaration that told it what to count, and the declaration was wrong. A number that is too big is not a safe error — it hides the one item that matters inside forty-three that do not.

This archive keeps the same facts twice — `data/households.tsv` and `site/src/data/households.json` — and edits both by hand. `check_release.py` refuses when they disagree about who is in which household, keyed on «household, person, event, date». It has never compared the evidence text, and on 20 September 2026 thirty-seven rows were found to disagree about what the document actually says. Three of them mattered and are already fixed. The TSV held «*Magdalenae RIVETTA* — READ FROM THE IMAGE 14 Sep 2026» while the JSON, which the site is built from, still said «Pivera» — an OCR guess this archive had disproved six days earlier against a control letter in the same pen. It also held the note that Clemente Falco's parish entry says 1855 where the civil act says 1853. The rest is drift and has to be walked. A person can hold several rows under one date, so the mismatch cannot be resolved by a script without choosing between two readings — which is a judgement, not a merge. Nothing here may be bulk-copied in either direction. Then the gate. Once the two files agree, `check_release.py` should compare the evidence as well as the names, so the next correction that stops halfway is refused instead of published. The whole point of keeping the file twice is lost if only one copy is ever read. ## DONE, 20 September 2026 — sixty rows, nine recoveries, and the gate now reads the evidence Sixty rows disagreed about what the document says, not thirty-seven: the number had grown all day because this row's own verification work was going into the JSON and being mirrored only narrowly. Thirty-six were the JSON ahead of the TSV — today's readings — and were mirrored forward only where the JSON provably contains the TSV's own text, tested by normalised substring, never by length. NINE WERE THE TSV HOLDING SOMETHING THE SITE HAD NEVER SHOWN. Two of them were live over-claims: GIUSEPPE DI GUIDA and GIUSEPPE MORGILLO were published as «so alive» on the «quondam» signal this archive disproved on 14 September. The downgrade reached the TSV and stopped. The archive stopped believing it six days before it stopped saying it. The other seven: the identification of the fourth Maria Falco; that her household lasted at most a few years; Suor Maria Maddalena Rivetti as the third sign of a Rivetti house of standing; why the 1841 machine-transcript row is deliberately left standing beside the act it duplicates; the refusal that Vigliotti and Vigliotta are not established as one family; that Domenico Clemente Falco's margin is where this archive's margin method came from; and the *muratore* note on Filomena di Lucia's father. The two files now agree row for row, and `check_release.py` compares the EVIDENCE as well as the names — on whitespace-flattened text, allowing several entries under one key, because a person legitimately appears twice under one date. Verified by reintroducing the «Pivera» fault: the gate refuses it with exit 1. It prints the warning it has to as well — do not bulk-copy, the newer text is not always the longer one — which is how three corrected readings were nearly destroyed this morning.

`sitemap.py` decides whether a page noindexes itself by reading its first 4,000 bytes. `checkarchive.py` then refuses the build if the sitemap lists any page whose whole file contains such a tag. So a noindex tag below the fold is invisible to the first and fatal to the second, and the build fails with a message that points at the page rather than at the disagreement. That is exactly what happened on 20 September 2026, twice, when a commit message quoting a robots tag was rendered into the body of `/changes`. The archive's own bug is fixed — `build_changes.py` escapes what it writes — but the disagreement is still there for the next archive. Both files are in `@daviddef/archive-kit`, which another session owns. Reported, not patched. The fix is one line: have `sitemap.py` scan the whole file, or have both agree on the same window.

The marriage act of 21 January 1860 lists the eight documents it was built on, and four of them are papers this archive does not have. **Item 4 — «*Nell'atto di morte dell'AVO PATERNO dello stesso*». The paternal grandfather of SABBATINO ANNECCHINO, that is the father of CARMINE ANNECCHINO of Forchia. A whole generation, named in a document that was read aloud at Arienzo and filed. This archive has never had a name for him. Item 5 is MARIA ANTONIA CARFORA's birth act — twenty-nine in January 1860, born at Arienzo about 1831, in a birth register nobody here has read. Item 7 is PASQUA MORGILLO's death act, which dates the end of a Carfora–Morgillo marriage at Arienzo. Item 6 is the bride's father's consent — this row called him «VALERIO CARFORA», and the forename was wrong; he is FRANCESCO CARFORA, and the consent paper is what proved it — and it will be signed or marked — the only sample of him this archive could ever have. Where it is.** Arienzo *processetti*, 1860, in the Antenati series this archive has surveyed but barely opened — one bundle read out of three hundred and twenty. Work-list rows 22 and 69 own the series; this row owns the year. Why this year before the others. It is the only bundle this archive can name a specific missing generation inside. --- ## READ — 21 September 2026. IT WAS NEVER BLOCKED. This row said «order the processetto», and the bundle was on Antenati the whole time, free. The Arienzo *processetti* for 1860 are `an_ua14207`, 479 images, and the archive's own holdings table already recorded that Arienzo has 63 processetti volumes covering 1809–1865 — it had the count and not the arks. The arks run `an_ua14149` (1809) upward, one per year with a few years doubled, and they were recovered in four probes. Finding the bundle. Marriage act 3 of 1860 ⇒ bundle 3. Each dossier opens with the printed «*ESTRATTO della solenne promessa*» and runs about thirteen images: bundle 1 at image 2, bundle 2 at 15, bundle 3 at images 25–42, bundle 4 at 43. Nothing had to be ordered and nothing had to be guessed. ### What the four items held Item 4 — the paternal grandfather, and a generation past him. Forchia, Morti 1826, act 8, 14 June 1826: «*è morto nella sua propria casa ANTONIO ANNECCHINO, nato in FORCHIA, di anni settantatre, di professione bracciale, domiciliato ivi, figlio di DOMENICO ANNECCHINO defunto … e di ELISABETTA D'ONOFRIO defunta*». Born about 1753 — and his own parents were dead before 1826, which puts DOMENICO ANNECCHINO in the 1710s or 1720s. The dossier holds two copies of this extract, on images 30 and 31, word for word the same. Item 5 — Mariantonia Carfora's birth extract, which gave the date (30 December 1830) and the act number (142). The certified copy was not treated as the record: the Arienzo birth register of 1830 was then opened at act 142 and read, and it is the original that this archive cites. See work-list row 63. Item 6 — the father's consent, and it demolished a published correction. «*si è presentato FRANCESCO CARFORA, contadino di detto comune, il quale presta il suo pieno consenso alla sua figlia MARIA ANTONIA CARFORA*», Arienzo, 1 January 1860, before the same Sindaco Sargenti who wrote the marriage act. The forename this archive had published from the marriage act was wrong, and this paper says so in a clear hand. He signed nothing — the act ends «*dovendo la parte apposta non sapere scrivere*». There is no sample of him, and there never will be. Item 7 — NOT FOUND. «*Nell'atto di morte della madre della stessa*» — Pasqua Morgillo's death act — is not in bundle 3. Images 25 to 42 hold four death extracts and every one is from Forchia: Carmine Annecchino 1838, Rosa Ceglia 1852, and Antonio Annecchino 1826 twice. The Arienzo paper the act says was read is not filed with the others. This is recorded as a nil so nobody re-reads the bundle for it; where it went is not known. ### Two documents the row never asked for Item 1 — the groom's birth act. Forchia, Nati 1825, act 5, 20 February 1825: presented by CARMINE ANNECCHINO, forty, bracciale, contrada detta li PRETI, the mother «*ROSA CEGLIA sua moglie legittima, di anni trenta*», baptised the same day at San Nicola, the child named SABATINO. Items 2 and 3 — the groom's parents' deaths. CARMINE ANNECCHINO, Forchia act 20, 8 August 1838, fifty-two, colono of the Masseria de' Preti, son of Antonio Annecchino and Maria Guida. ROSA CEGLIA, Forchia act 4, 4 April 1852, sixty-two, ***filatrice*, born at ARPAIA, daughter of Antonio Ceglia and Grazia Vitagliano — a surname that had never appeared in this archive in any file. Four generations of Annecchino where there was one, three new households, and a correction. The lesson is the row's own first sentence. It said «order», and the reason it said «order» is that the archive knew the series by a count in a survey table rather than by an ark, and a count cannot be opened. Rows 22 and 69 own the rest of the series; 319 bundles are still shut.**

Two rows on this site record an ANNA FALCO, eight days old, daughter of MICHELANGELO FALCO, dead on 7 March 1834, both citing `an_ua14336` image 10 left. Image 10 left is MARIA PORRINO — eight days old, daughter of Angelo Porrino and Pasqua Rivetti, dead on 8 March 1834. Five leaves around it were opened on 20 September 2026 and none holds her. The only act of 7 March on those leaves is Francesco Liparulo's. And MICHELANGELO FALCO is a forename that appears nowhere else in this archive. What to do. The 1834 death register has its own tavola alfabetica; read it for FALCO, which is four images against ninety. If an Anna Falco of March 1834 is in it, the act number is there and this closes in an hour. If she is not, she is a mis-indexed Maria Porrino and both rows are retired — with the reasoning published, not deleted. Why it matters beyond one child. She is entered twice, which means she has been counted twice in every total this archive publishes about the Falco of Arienzo. ## DONE, 20 September 2026 — and the index route was a dead end **The 1834 death register has no *tavola* bound in: its fifty-seven images end with the last act and a blank form. The separate index volume, `an_ua14296`, is ONE IMAGE, and that image covers surnames A and C only — the years on either side have six images and five. The Arienzo death index for 1834 survives as a single leaf, and F is not on it. That is a hole in the record, and it is published rather than left to be rediscovered. So she was found the other way. A contact-sheet sweep of the volume's name lines — thirty-two tiles across sixteen leaves in one request — put her on image 11 left, and the leaf was then read at full resolution. ANNA FALCO IS REAL, AND ALMOST NOTHING ABOUT HER WAS RIGHT. Act 16, 16 March 1834, not the 7th. Eight days old, so born about 8 March. Her mother is RAFFAELA FALCO — a Falco married to a Falco, never recorded here — not Mariantonia Vigliotta, so she is not the child of the household she was filed under. Declared by Giuseppe Carfora, forty and Tomaso Sposito, sixty, both of strada Camellara. And one more age, for the series. Tomaso Sposito is forty-six in June 1824, fifty in April 1826, and sixty in March 1834**.

The citation «`an_ua14329` img47L atto84» is impossible. Arienzo Morti 1827 has 42 images and 68 acts; neither the leaf nor the act exists, and three contact sheets across images 20–41 found no priest. The reading itself is detailed enough to have come from an act: «*D. GIUSEPPE C—, d'anni cinquantasei, nato in Arienzo, di professione SACERDOTE, figlio di ANDREA —, possidente, e di ORSOLA ZIMBARDO, defunti*». It is not a guess and it is corroborated from the other end: his sister CELIA CIOFFI's act of 3 June 1843 gives the same father, the same trade and the same mother. Where to look. The date «4 December 1827» may itself be the error — try the 1826, 1828 and 1829 death registers, and the death indexes for those years, for a Giuseppe Cioffi aged about fifty-six, sacerdote. If he is not in any of them, the reading came from a machine index and must be marked as one, not retired: the parents are confirmed by a second act. ## FOUR YEARS EXCLUDED, 20 September 2026 The death indexes for 1826, 1827, 1828 and 1829 were read — `an_ua14288` to `an_ua14291`, nine leaves — and the C block was opened in every one. There is no Giuseppe Cioffi, Crisci or Cossi described as SACERDOTE in any of the four years. The only religious in the stretch is Crisci Suor Maria Teresa, a *monaca*, in 1827. So the year is wrong by more than two, or the reading never came from a death act. The second is now the likelier, and the same four indexes showed why: the archive's «Gaspare Cossi» of 1826 was assembled out of an index and a machine transcript, with the act itself saying neither. Still to try: 1830–1835, and the parish burial register, where a priest would be entered with his title. He is not at risk — his sister Celia's act of 1843 names the same parents — it is his own document that is lost. ## 21 September — THE ROW'S REMAINING ACTION WAS ALREADY DONE, AND NOBODY NOTICED This row said «Still to try: 1830–1835». Every one of those years was already read, and the row that says so is on *What has been read*: «*ANTENATI «Morti, indice» — thirteen annual volumes, 55 images, every page's surname column cropped and tiled*», covering 1831–1843, beside the earlier sweep of 1824–1830. The Arienzo death indexes are read continuously from 1824 to 1843. So the negative is much wider than this row believed, and it is bounded by two holes rather than by a year. Don Giuseppe Cioffi — a C — is in none of the twenty years, and the C leaf was opened in each. The two holes are: - 1834, whose index is a single leaf covering A to C and marked END. C is inside that range, so 1834 is covered for this man, not a hole. - 1836, which does not exist in the index series at all. That leaves exactly one unexamined year before 1844: 1836. Its death register is `an_ua14338`, 51 images, read for FALCO and not for a priest. And the row's own criterion is now nearly met. It said: «*If he is not in any of them, the reading came from a machine index and must be marked as one*». He is in none of 1824–1843 but 1836. The reading is therefore recorded as PROBABLY MACHINE-DERIVED, pending 1836 — not retired, because his parents are independently confirmed by his sister Celia Cioffi's act of 3 June 1843. The finding behind the finding is about this archive, not about the priest. A work-list row named volumes to read while the page that records what has been read already said they were done. Nothing compares a row's «still to try» against `searched.json`, and this is the second time in two days that a stale instruction has survived the work that answered it. ## CLOSED, 21 September 2026 — 1836 IS READ, AND HE IS NOT IN IT `an_ua14338`, all 51 images, swept at the band that carries the dead person's name, age and trade. That is the one year between 1824 and 1843 with no index in the series, and it was the only thing left. The year buries EIGHT religious and not one of them is he: Don Gennaro Esposito, *sacerdote*, ninety; Don Alessandro Bruno, *sacerdote*, twenty-seven, son of Don Carlantonio Bruno and Donna Teresa Migliore; Don Francesco de Gentij, *sacerdote*, forty-six, son of Don Mario de Gentij, *possidente*; a Diglio *sacerdote*, son of Pasquale Diglio and Angela Migliore; Don Antonio Cimmino, *seminarista*, twenty-four; Suor Fedele Ruggiero, *monaca*, sixty-three; Donna Carlotta Rauti, *monaca*, seventy-three; and Donna Raffaela de' Falco, *monaca*, sixty-seven — the gentry nun this archive already holds, her act confirmed here in passing. There is no Giuseppe Cioffi, Crisci or Cossi, sacerdote, aged about fifty-six, in any Arienzo death index or register from 1824 to 1843. By this row's own criterion the reading is now recorded as MACHINE-DERIVED. It is not retired: his parents are independently confirmed by his sister Celia Cioffi's act of 3 June 1843, so the man is not in doubt — only his own document is lost, and it was never where the citation said. The sweep paid for itself twice over: act 50 of the same volume is PASQUALE DI GUIDA, a brother of generation four's mother-in-law.

**Found in the 1829 *tavola alfabetica* while looking for something else.** «*9. C[?]ffi LUCA … [figlio di] GIOACCHINO e FALCO FRANCESCA … 16 [Gennaio 1829]*». The parents are the find. «GIOACCHINO» with «FALCO FRANCESCA» is a distinctive pair and this archive holds exactly one couple of that description — GIOACCHINO CIOFFI × FRANCESCA FALCO, known until now from a single document, the 1843 death act of their son ANIELLO, seventy-six. This would be a second child, and the household's second document in a century of looking. What the index cannot give. The dead man's surname is in the Crisci/Cossi/Cioffi trap, and an index is written faster than an act. It gives no age, so whether Luca is an infant or a grown son of a woman bearing children in the 1760s is unknown — and that decides whether he is a brother of Aniello or something else entirely. Act 9 of Arienzo Morti 1829 settles both in one line. The volume is `an_ua14331`. ## DONE, 20 September 2026 — and it was three findings, not one Act 9 is on image 7 RIGHT, found by sweeping the act numbers across images 4–12. «*è morto nella sua propria casa LUCA CRISCI, di anni SESSANTA, marito di ANTONIA MORGILLO, nato in Arienzo, di professione contadino … figlio di GIOACCHINO CRISCI Defunto … e di FRANCESCA FALCO Defunta*», 16 January 1829, at twenty-four hours. 1. He is ANIELLO's brother. Sixty in 1829 gives about 1769; Aniello, seventy-six in July 1843, about 1767. Two years apart, the same two parents. The household goes from one child to two and from one document to two. 2. THE SURNAME IS CRISCI, NOT CIOFFI — and so is the household's. Both acts were magnified, including the 1843 one this archive had never magnified, having taken the name from a transcription. A narrow r, no round o, no double-f ascender. 3. Both parents are «Defunto» and «Defunta» in January 1829, so Francesca Falco died before then, not merely before 1843 — a window closed by fourteen years. And Luca had a wife, ANTONIA MORGILLO. The index produced the candidate and the act produced the record, which is the order this archive insists on. The index could not spell his surname; the act could.

The only register sweep on this archive's shelf with no row against it, found by reconciling the research log — work-list row 74. `an_ua14219`, thirty images, and the volume holds EXACTLY TWENTY-FIVE acts — fixed by the sindaco's own closing certificate of 31 December 1815, which names act 25 as «*l'ultimo celebrato sul presente Registro*», and corroborated by the divider card in the processetti volume: «*Anno 1815 / Arienzo (66) / dal n° 1 al n° 25*». Images 2–16 were read at survey resolution, roughly acts 1 to 13. Images 17 to 30 were never opened — about twelve marriages, half the year. The read half already produced a household: SEBASTIANO FALCO × COLOMBA MONTEFUSCO of San Felice a Cancello, *sapunari* — soap-makers — whose son LUIGI, eighteen, married at Arienzo as a minor with both parents consenting in person. That is one of the five strands on the places figure. What the unread half holds is unknown, and a marriage register is the densest thing in this archive: both spouses, both sets of parents, ages, trades and streets, on one leaf. Survey resolution is not a reading. Even acts 1–13 should be re-read at full magnification before anything in them is quoted — the bride of the Falco act is still «*[?]ietta RENCHIELLO(?)*» and her mother «*[An]gela [?]*». --- ## DONE, 20 September 2026 — and there is no Falco in the unread half All ten acts read at 3,200 pixels. Not twelve: **image *N* right holds act *N−1*, so images 17–30 are acts 16 to 25, images 24 and 25 are duplicate photographs of openings already filmed as 22 and 23, image 29 is the closing certificate and image 30 the endboard. The header of this row said «roughly acts 1 to 13» for images 2–16; they are acts 1 to 15. Loffredo, Morgillo, Verdicchio, Grossi, Carfora, Morgillo, Ferriello, Arrigale, Marletta, Buono — and no Falco. The year's only Falco remains the soap-maker's son of San Felice a Cancello at act 8, published and still not claimed. The roster of all ten, with both sets of parents, ages, trades and streets, is in `notes/1815-marriage-register-progress.md`. Two things the unread half gave. Act 20 is a gentry Carfora marriage** — *Signor PASQUALE CARFORA, medico cerusico* of Santa Maria a Vico × *Signora Alessandra de Angelis, gentildonna, of Naples* — naming PAOLO CARFORA and ANNA SABATELLI dead on 6 November 1782 and 22 January 1789. That is material for row 63, which exists because this archive did not know there were Carfora at Arienzo. And act 21 gives ELISABETTA COSSI, daughter of the late MARINO COSSI — the documented Cossi household this archive has needed while correcting *Cossi* to *Cioffi* on strada Camellara: written twice in one hand as C + round o + double long ſ + i, with no *i* after the C and no narrow *r*. One age flagged and deliberately not changed. Act 18 makes Anna Martone twenty-nine with a father of forty, and calls her minore. The likeliest reading is *diciannove*, but the word was not magnified, so it stands as written. What remains, and it is now the only thing: acts 1–15 were read at survey resolution, which this archive does not count as a reading. The bride of the Falco act is still «*[?]ietta RENCHIELLO(?)*» and her mother «*[An]gela [?]*». Raised as its own row rather than left inside a finished one.

Images 2–16 of `an_ua14219` were read at SURVEY resolution only, and this archive's standing rule is that survey resolution is not a reading. Row 81 finished acts 16–25 at 3,200 pixels and found ten acts' worth of names the survey had not given; the same is owed to acts 1–15. The specific debt is inside the one act that matters. Act 8 is the Falco of San Felice a Cancello — LUIGI FALCO, eighteen, *sapunaro*, son of SEBASTIANO FALCO and COLOMBA MONTEFUSCO — and the bride is still unread: «*[?]ietta RENCHIELLO(?)*», of strada del Borgo, with her mother «*[An]gela [?]*», fifty-six. A household published with an unreadable spouse is a household that cannot be searched for anywhere else. Cheap: fifteen acts, one render each, using the layout row 81 established — image *N* right holds act *N−1*, each act starting on a right page and running onto the next left page. --- ## ACT 8 IS DONE, AND IT WAS WORTH DOING FIRST The bride is read. «*è comparsa egualmente STELLA PETRILLO di anni diciotto, domiciliata in questa suddetta comune, strada del Borgo, di professione filatrice, MINORE, assistita da BRIGIDA COSSI sua madre di anni cinquantasei, di professione contadina, domiciliata colla figlia, essendo morto il padre SABBATINO PETRILLO*» — and the groom is «LUIGI FALCO di anni diciotto compiti, di professione SAPUNARO, domiciliato nel comune de' SEI CASALI, strada Croce di San Felice, MINORE», with SEBASTIANO FALCO of the same trade and COLOMBA MONTEFUSCO, fifty-five, *filatrice*. 19 Mar 1815, before Daniele Loria, sindaco. Three corrections came out of one act. Her forename: the act says STELLA, the index says GIULIA, and this archive had «Giustina», which is neither — the parents are identical in both documents so the woman is not in doubt, only her name. Her mother: BRIGIDA, not the «[An]gela [?]» the note carried. And her surname went the other way: at survey width the act reads «Vecchiello», which is what this pass was about to publish, and at twelve times it is PETRILLO. *A word that changes when you magnify it was never read.* And the surname test this archive relies on was narrowed. «COSSI» here descends exactly as a double f would, so the descender does not separate it from «Cioffi» in this hand. The letter count does: *Cioffi* carries an i between the C and the o and *Cossi* does not. Three instances in this one volume say so. On /corrections as a method entry. Fourteen acts remain — 1 to 7 and 9 to 15. --- ## DONE — ALL FIFTEEN ACTS READ AT 3,200 PIXELS, 21 September 2026 The volume is now read end to end, both halves, every act at full resolution. The roster with both sets of parents, ages, trades and streets is in `notes/1815-marriage-register-progress.md`. IMAGE 14 IS A DUPLICATE PHOTOGRAPH OF IMAGE 13 — the same act 13 filmed twice. That is the extra leaf that made image *N* right hold act *N* up to image 13 and act *N−1* from image 15 on. With images 24 and 25 duplicating 22 and 23, the volume contains three duplicate photographs. AND THE FALCO ACT IS 19 MARCH, NOT 19 MAY. The date had stood as «19 [May] 1815», bracketed; hours earlier today the bracket was removed and «19 May» published as fact, in the entry that read the bride's name off the act for the first time. The word is MARZO, and the register's own order proves it without reading the word at all: act 7 is 16 March, act 9 is 21 March, act 10 is 6 April, and the volume runs in strict date order from 4 February to 2 December. A 19 May at act 8 is impossible. On /corrections. Five COSSI in one volume in one hand — Brigida, Chiara, Faustina, Elisabetta and Marino — every one written *C + round o + double long ſ + i* with no i between the C and the o. The surname is settled and the strada Camellara correction was about one household. Three things for other rows. DOROTEA CARFORA, 49, a *bracciale*'s wife of strada Rosciano, born about 1766 — a third social register of Carfora for row 63, joined to neither the gentry surgeon of act 20 nor the *contadini* of 1831. MARGARITA DI GUIDA, dead at Arienzo 15 October 1814, named in act 1 — a lead for rows 39 and 40, claimed as nothing. And two strada Camellara households, the Melillo and the Verdicchio shoemakers. One caution recorded rather than resolved: at contact-sheet resolution the 1815 marriage index appeared to number Michele Monaco 13 where the register numbers him 15. The index was not read at magnification, so nothing is claimed.

Row 79 told this archive to read six volumes that `searched.json` already recorded as read — the Arienzo death indexes for 1830–1835, part of a continuous sweep of 1824 to 1843. The row would have sent a reader to open all six for nothing, and it had been standing since the row was raised. Neither file knows the other exists. A work-list row's «still to try» is a claim about what has NOT been done; `searched.json` is the register of what HAS been. Nothing compares them, in either direction — so a row can ask for finished work, and a finished search can go unrecorded in the row that wanted it. What to build. A gate is probably not writable: the two files name sources in prose, not by key, and a regular expression over «1830–1835» against «1831–1843» is a range-parsing problem with a long tail. What IS writable is a report — pull every «still to try», «next» and «remains» sentence out of the open rows, pull every `src` and `when` out of `searched.json`, and print them side by side by ark and by year for a human to read in one pass. One reading, not a build step. Why it is worth a row of its own. This is the third instance in three days of the archive's own files disagreeing while both are right about their own contents: the two households files disagreeing about evidence text, the corrections page standing after the findings that overturned it, and now the work list against the coverage page. At this size the archive's largest source of wasted work is no longer the sources — it is not knowing what it already holds. --- ## BUILT AND RUN — `tools/staleness_report.py` Sixteen hits across twenty-three open rows, and every one of them is row 79. Nothing false anywhere else. The report pulls every «still to try / what remains / never opened / next» sentence out of the open rows, takes the years and the SERIES out of each, and prints it beside any coverage row covering the same years in the same series. It is a report and not a gate, on purpose. `searched.json` has a structured `when` and names the ark; a work-list row names years in prose, inside an argument, and often names no ark at all — row 79's six volumes exist only as the string «1830–1835». Matching those is a judgement about whether two prose descriptions mean the same shelf, and a row may be re-reading a volume at higher resolution on purpose (row 82 is doing exactly that to `an_ua14219`). A gate here would either cry wolf or miss the case it was written for. Matching on the year alone does not work, and two things had to be added before the output was readable. 1. The SERIES has to match as well as the year. Arienzo has deaths, births, marriages and an index series for each, in the same years. «1830–1835 read» answers «1830–1835 still to try» only if both mean the same shelf — so an index is kept apart from the register it indexes. 2. «ANTENATI» contains «nati». Every row in both files names Antenati, so on the first run every row was classified as a BIRTH row and the report was all noise. Six of the twenty-two hits died with that one substitution. And a third, which is about honesty rather than noise. A coverage row spanning a century — a commercial index, a whole parish harvest — overlaps every question ever asked and answers none of them. Those are marked WIDE and hidden by default rather than dropped, because «FindMyPast covers 1806–1900» is true and useless, and a report that silently discarded it would be hiding a row rather than ranking it. `--all` shows them. Ark matching, which was the obvious approach, is nearly worthless here: only four arks are named both in an open row and on the coverage page, and all four are explainable. The miss that cost six volumes named no ark at all.

This archive writes both, heavily, in the same files and the same years: 228 CAPODIUOMO against 707 CAPODICONCA in `data/people-register.tsv`, `data/sweep-people.tsv` and the published register. The Capodiuomo rows are concentrated and they have a shape. Ten years of death-register progress notes carry them — 1817 (11), 1818 (9), 1819 (5), 1820 (8), 1821 (10), 1822 (14), 1823 (25), 1824 (9), 1825 (1) — and then almost nothing until a single instance each in 1837 and 1838. That is the fingerprint of a reading habit in one stretch of work, not of a street that existed for eight years and stopped. What is actually established. Two acts were read at 3,000 pixels on 20 September 2026 — `an_ua14324` img34r act 65 and `an_ua14325` img55r act 51 — and both read CAPODICONCA, the word appearing three times in the first. Nothing else has been magnified. The correction published that day over-generalised from those two and has been amended. And the negative evidence points one way. The Arienzo BIRTH registers for 1832 and 1836, read out act by act on 21 September, give strada Capodiconca constantly and strada Capodiuomo not once. How to settle it cheaply. Take five Capodiuomo rows from different years — 1817, 1819, 1821, 1823, 1837 — render each act at 3,000 px and magnify the word. If all five read Capodiconca the street is one and 228 rows need correcting; if any reads Capodiuomo, Arienzo had two streets and both corrections stand as written. Five images decide a word that appears nine hundred times. --- ## SETTLED THE SAME DAY — SIX ACTS, AND IT IS CAPODICONCA Five images were budgeted; six were read, chosen to span both ends and the middle of the run. 1817 act 11, 1819 act 23, 1821 act 15, 1822 act 65, 1823 act 51, 1824 act 5 — every one CAPODICONCA, several twice in one act, and the 1824 act in a hand so clear it needs no magnification at all. 219 rows corrected in `data/sweep-people.tsv`, the published register and nine years of notes. Eleven left alone: the 1837 and 1838 rows, because that hand is too rapid to decide at this scan's limit. A word this archive cannot read is not a word it may change. The distribution was the tell and it is worth keeping. 1817–1825 held 219 of the 228, then nothing for eleven years, then nine stragglers. A name that appears heavily for one continuous stretch of work and then stops is a reading habit, not a fact about the town. Checking the *distribution* of a doubtful reading across the corpus cost one command and pointed at the answer before any image was opened.

The gap, stated plainly. In the Arienzo birth registers an act opens on a RIGHT page and the parents are named at the top of the NEXT LEFT page — but only when somebody other than the father presents the child. When the father declares, the left page says «*da lui dichiarante*» and his name appears only on the right page. A sweep of the parents band therefore reads every mother and misses a father whose wife is not herself a Falco. How it was found. Two 1836 death acts give SALVATORE FALCO at fourteen months and PASQUALINA FALCO at sixteen — births in mid-1835 — and 1835 had just been published here as holding no Falco. Salvatore's mother is a VIGLIOTTA: if his father declared, the parents band could never have seen him. Already closed: 1834, 1835 and 1836 have now been read at both bands. All three still hold no Falco, so the two children remain unaccounted for — but the nils are now sound. Still owed a second band: 1820–1823, 1829, 1831, 1832, 1833, 1837, 1838. 1820 is exempt — each act fits on one page there, so the single band caught both parents. Roughly sixty contact sheets at `--band 0.34 0.19 --half R --cols 4 --width 1600`. Why it is worth the second pass. Every nil this row touches was published as a negative, and a negative is the most expensive thing this archive can get wrong — nobody re-checks one. Two of them are already known to be incomplete in principle; the rest are untested. ## VINDICATED ON THE FIRST VOLUME IT TOUCHED — 21 September 2026 1821, declarant band, act 22: ROSARIA FALCO, born 15 March 1821 to GIUSEPPE FALCO, thirty-eight, and GELSOMINA VIGLIOTTA, thirty-seven, of strada Camellara. The mother is a Vigliotta, so the parents band could never have seen her — and that band is what this archive had published 1821 on. It also broke a discriminator. This household is the one the archive places on strada COSTA; here it is on strada CAMELLARA. On /corrections. ## 1822 AND 1823 READ OUT — 21 September 2026 Both years hold exactly one Falco declarant each, and the archive already had both. 1822 act 8 is GIUSEPPE FALCO FU MATTEO, presenting Angela Maria; 1823 act 38 is GIUSEPPE FALCO, forty-two, massaro di campagna of strada Camellara, presenting Pasqua. 278 images read; nothing new. The second band is still paying, but not in new children. It has now produced three corrections to work already published: an 1821 contact-sheet candidate that was a COSCI, not a Falco; a 1823 coverage row that had left Pasqua Falco's birth act out of its own year; and Pasqua's birth place, which said *strada Costa* while the act quoted inside the same row said *strada Camellara*. Re-reading an act this archive has already read is not wasted work. ## 1829 READ OUT, AND 1831 SHOULD NEVER HAVE BEEN ON THIS LIST — 21 September 2026 1829: no Falco declarant at all. The year is carried by midwives — ROSA DELLA SELVA, seventy-eight, and CATERINA RUGGIERO, seventy-one — so in a year like this the parents band is the stronger instrument and the declarant band returns the same two names over and over. **Its *tavola* is a blank printed form: 1829 has no index and never had one, which makes the act-by-act pass the only instrument there is. 1831 was already done at both bands and this row said otherwise.** The coverage page's own first sentence reads «*RE-READ at the PARENTS band; the earlier pass swept the DECLARANT line*». A hundred images were read a third time on this row's instruction. `tools/staleness_report.py` exists to catch exactly that and printed nothing, because «*still owed*» was not a trigger phrase and «*remaining*» is not «*remains*». The tool has been widened; the episode is on /corrections. ## AND THEN IT FOUND A CHILD OF THE DIRECT LINE — 1832, act 84 MARIA CARMELA FALCO, born 10 October 1832, in her father's own house on strada Camellara, to VINCENZO FALCO, forty, and ANDREANA CRISCI, thirty-six. This archive had published that this couple had no child in 1832, after reading the whole volume at the parents band. The act states the failure outright. The left page reads «*nata da ANDREANA CRISCI sua moglie legittima … E DA LUI DICHIARANTE, di anni come sopra, di professione come sopra, domiciliato come sopra*». The father's name is nowhere on the page this archive reads for parents. No index rescues it either: the 1832 *tavola* stops at C. Row 85 has now paid twice in one day — Rosaria Falco in 1821, Maria Carmela Falco in 1832 — and the second one is generation three. Every year still on this list was published as a negative on the band that cannot see this. 1833: no Falco declarant, in a year carried by the two midwives again. One candidate opened and excluded — VINCENZO CALCAGNO, forty, contadino, at image 109, which is what Vincenzo Falco's age and trade would have been. ## THE LIST IS FINISHED — AND 1838 WAS THE WORST OF THEM 1837: one Falco declarant, FRANCESCANTONIO FALCO at act 108, which is the Raffaele Falco this archive already held. Nothing new. 1838 — the year this archive had published as «NO FALCO AT ALL» — holds TWO Falco births. Act 79, 9 September: STANISLAO FALCO, to VINCENZO FALCO, forty-six, and ANDREANA CRISCI, forty, of strada Camellara — a son of generation three whose birth act had never been found; the archive knew him only from his wedding of 1861. And act 77, 4 September: MARINO ALESSANDRO FALCO, to MICHELANGELO FALCO, thirty-four, and MARIA MIGLIORE, thirty-five, of strada Sant'Andrea — a third wife for Michelangelo Falco in seven years, on a street neither of the others is on. **Both left pages say «*e da lui dichiarante*» and never write the father's name. The old row even reported «no Crisci» in a year where Andreana Crisci is named in full on one of them. ## WHAT THE ROW PRODUCED, IN ONE DAY Nine volumes re-read; four Falco births recovered — ROSARIA (1821), MARIA CARMELA (1832), STANISLAO and MARINO ALESSANDRO (both 1838) — three of them into households this archive already had, two of them into the direct line. Three published statements corrected besides: a coverage row that had left Pasqua Falco's birth act out of its own year, her street, and an 1821 contact-sheet candidate that was a COSCI. And one process fix: `tools/staleness_report.py` could not see this row's own «still owed» and now can. Done: 1821, 1822, 1823, 1829, 1832, 1833, 1837, 1838 — and 1831, which never needed it. Exempt: 1820. NOTHING REMAINS ON THIS ROW. ## WHAT IT LEAVES BEHIND The two bands are not alternatives and this archive should stop treating a single-band sweep as a reading. Every Arienzo birth register read at one band only is a nil of unknown strength. The nine years on this row were the ones already suspected; the rest of the series has never been tested.**

Row 85 has just shown what a one-band reading is worth. Nine volumes already published as read were re-read at the declarant band and gave four Falco births, two of them into the direct line, one of them in a year whose outcome was «no Falco». The nine years on that row were the ones already suspected. The rest of the series has never been tested. The rule, in the act's own words. When the father presents his own child the left page says «*è nato da [MOTHER] sua moglie legittima … E DA LUI DICHIARANTE, di anni come sopra, di professione come sopra, domiciliato come sopra*» — and never writes his name. A parents-band sweep sees the mother's surname and nothing else. So a Falco father whose wife is not a Falco is invisible at any resolution. TWO GROUPS OF YEARS. (a) Read at the parents band only — the nil is of unknown strength: 1816 (`an_ua14413`), 1817 (`an_ua14414`), 1818 (`an_ua14415`), 1834 (`an_ua14431` — already done at both), 1835 (`an_ua14432` — done), 1836 (`an_ua14433` — done). *So the real remainder here is 1816, 1817 and 1818*, and 1817's outcome is «no Falco», the same kind of statement that failed in 1838. (b) No coverage row exists at all: 1819, 1824, 1825, 1826, 1827, 1828, 1830. Single acts from 1824, 1825, 1826 and 1828 are cited in households, which means those volumes were *opened at one page*. Nobody can say whether they were read, and that is precisely the state /searched exists to prevent. 1826 and 1827 are row 35's and are being read now. Cost. About 1,300 images across ten volumes, which is roughly fifty-five contact sheets at `--band 0.34 0.19 --half R --cols 4 --width 1600`, plus the parents band for the years that have never had one. Row 85 took one day for nine volumes. Why it is worth it, stated plainly. Generation three bore children across this whole span — Luigi 1819, Giuseppe Nunzio 1824, Angelandrea 1826, Sara 1828, Pasquale 1830, Maria Carmela 1832, Pasquale the second 1834, Stanislao 1838, Antonio 1843 — and two of those, Maria Carmela and Stanislao, were found only yesterday, in volumes this archive had called read. A gap year in that series is a missing child until a reading at both bands says otherwise. ## THE THIRD INSTRUMENT, AND WHAT IT MAKES REDUNDANT — 21 September 2026 Reading 1826 and 1827 produced a rule that halves this row. There are three ways into one of these volumes and they do not overlap the way they look as though they do: - The TAVOLA indexes the child's surname, with both parents named. So a complete tavola answers the question «did a FALCO FATHER register a child this year?» outright — the child is a Falco and is filed under F. - The PARENTS band is the only instrument that finds a FALCO MOTHER, because her child is filed in the tavola under her husband's surname. That is how MARIA FALCO's son of 1826 and ROSA FALCO's daughter of 1827 were missed for years. - The DECLARANT band finds a Falco father whose left page says «*e da lui dichiarante*». It is redundant wherever the tavola is complete, and it is the expensive one. So the cheap route is TAVOLA + PARENTS BAND, and the declarant band only where the tavola fails. And it fails often enough to matter: 1829's tavola is a blank printed form, 1832's stops at C, 1823's is under a paper slip, and 1838's leaf was never filmed past the recto. 1826's and 1827's are complete through the whole alphabet and both were checked against the page-break rule. Check the tavola first. It is one opening and it decides how much of the volume has to be read. ## FOUR VOLUMES OFF THE LIST — 21 September 2026 1824, 1825, 1826 and 1827 are read end to end at both bands and are on /searched. Two new children came out of them — GIUSEPPE CIOFFI (1826) and MARIA CARMELA MORGILLO (1827), both through Falco mothers, which is the half of the problem the tavole cannot answer — together with a second Andreana Crisci, two corrected act numbers and a death act that had been harvested without being held. REMAINING: 1819, 1828 and 1830, which have no coverage row at all; and 1816, 1817 and 1818, read at the parents band only. 1817's outcome is «no Falco» — the same kind of statement that failed in 1838. ## 1828 DONE, AND THE TAVOLA RULE PAID FOR ITSELF MARIANTONIA CIOFFI, born 8 March 1828 to Francesco Cioffi and MARIA FALCO — her death act, already held, says seven months old on 23 October, and she was seven months and fifteen days. A life closed from both ends. **The 1828 *tavola* is complete and its F group holds one entry, Sara Falco at act 118, so the declarant band was swept and deliberately not read: an index answers for Falco FATHERS and cannot answer for Falco MOTHERS, and act 28 is a mother. That is the rule this row set out, working. REMAINING: 1819 (different format — prose, two acts an opening, one band catches both parents) and 1830; then 1816, 1817 and 1818, read at the parents band only. ## 1819 READ OUT — TWO NEW CHILDREN, BOTH THROUGH FALCO MOTHERS ALESSANDRA DIGLIO, 27 April 1819 (act 52), to Alfonso Diglio and MARIA FALCO — and her father signed his own name where both witnesses could not write. GIUSEPPE CIOFFI, 7 June 1819 (act 68), to Francesco Cioffi and MARIA FALCO — a child this archive held only as a death, 25 March 1824 aged five; born 7 June 1819 he was five years and nine months. Three more Falco acts confirmed from the image — Domenico Clemente (act 33), Crescenzo (act 99), Luigi (act 134) — and with them the proof that two Giuseppe Falco were on strada Camellara seven months apart in 1819. That corrects a date this archive had published twice. The 1819 volume is in the older Napoleonic format — running prose, two acts to an opening — so one band catches declarant and mother together**: `--band 0.17 0.28 --half LR --cols 4 --width 2400`, six tiles to a sheet. Its *tavola* survives only as its last page, numbered XI, with the child's-name column lost in the gutter. REMAINING: 1830, which has no coverage row at all; and 1816, 1817 and 1818, read at the parents band only. ## 1830 READ OUT — AND IT IS THE CASE FOR RUNNING BOTH BANDS, IN ONE VOLUME Two Falco children born on 24 March 1830, to two households on strada Camellara, and this archive had neither act. PASQUALE FALCO (act 42, six in the morning) to Vincenzo Falco and Andreana Crisci — the civil act for a child known only from a machine-read baptism and a death. ANNA FALCO (act 43, eleven) to Michelangelo Falco and Antonia Vigliotta — new, and her death act's «*di mesi TREDICI*» closes to ten days. EACH BAND FOUND ONE AND NEITHER WOULD HAVE FOUND BOTH. Act 42 says «*e da lui dichiarante*» and names no Falco on the parents page. Act 43 was presented by a midwife and names no Falco on the declarant page. And the 1830 tavola is under the archivist's card, so there was no third route. REMAINING ON THIS ROW: 1816, 1817 and 1818, read at the PARENTS band only. 1817's outcome is «no Falco» — the same kind of statement that failed in 1838. ## 1816 READ OUT — NO NEW FALCO, AND THE BEST FIND OF THE ROW SARA RUGGIERO — generation three's mother-in-law — WAS A MIDWIFE. Act 29 of 1816: «*è comparsa SARA RUGGIERO, di anni QUARANTACINQUE, di professione LEVATRICE, domiciliata in detto Comune, strada CAMELLARA*». This archive holds a Sara Ruggiero, forty-five, of strada Camellara, in the 1817 register as Arcangelo Crisci's wife. And it explains the ten pages. «*Matrina fuit SARA RUGGIERO*» stands on ten pages of the parish baptism harvest, godmother across six Falco households, and the archive had no account of it. A levatrice stands godmother to the children she delivers. *Three Ruggiero women delivered this town — Sara, Caterina and Gelsomina — across twenty years.* The 1816 tavola is under the archivist's card as well, which makes six volumes: 1816, 1823, 1824, 1825, 1830, 1846. Work-list row 10 is now about a policy, not a leaf. REMAINING ON THIS ROW: 1817 and 1818. Both are in the same prose format as 1816 and 1819, so one band does declarant and mother together — `--band 0.17 0.28 --half LR --cols 4 --width 2400`, and the sweeps are already rendered. 1817's outcome is «no Falco», the same kind of statement that failed in 1838. ## 1817 READ OUT — THE SECOND «NO FALCO» OVERTURNED IN ONE DAY SALVATORE DIGLIO, 15 May 1817 (act 49), to Alfonso Diglio and MARIA FALCO of strada Camellara — new, and the first of two Salvatore in that house. And GIOVANNA MORGILLO, 22 September 1817 (act 69), to Carlantonio Morgillo and ROSA FALCO, a daughter of the founders — which this archive has cited from this very volume, image 36 left, since before this week. The coverage page and the household file were contradicting each other about the year. Three di Guida acts came with it, and one closes a prediction: MADDALENA DI GUIDA, 25 May 1817, to ANGELANDREA DI GUIDA DI GIUSEPPE — the archive had predicted she was «*named for her grandmother Maddalena Rivetta*» and here is the birth, with «di Giuseppe» corroborating the parish filiation from a civil document five months before her father's death. PASQUALE DI GUIDA is forty-two and a MASSARO DI CAMPAGNA in June 1817, which dates his unread baptism year to about 1775. REMAINING ON THIS ROW: 1818 alone — 85 images, same prose format, sweep already rendered. ## 1818 READ OUT — THE ROW IS FINISHED MARIA CARMELA, 18 January 1818, to DOMENICO COSSI of strada Cossi and MARGARITA FALCO, forty — new, and the earliest act that household has. Raffaele Falco (act 47) and Maddalena Falco (June) confirmed. The volume holds the year twice — a printed slip declares the 1817 register stock «*vale pel corrente Esercizio 1818*» — so act 9 could be read in two hands. Both say COSSI, where this archive's household is called COSPI and an 1819 index makes the son a CRISCI. Three spellings, three documents, one household; published as a question, and the household is not renamed. ## WHAT THE WHOLE ROW PRODUCED Ten volumes read end to end — 1816, 1817, 1818, 1819, 1824, 1825, 1826, 1827, 1828, 1830 — and nine children recovered: Giuseppe Cioffi (1819), Alessandra Diglio (1819), Salvatore Diglio (1817), Maria Carmela of Margarita Falco (1818), Giuseppe Cioffi the second (1826), Maria Carmela Morgillo (1827), Mariantonia Cioffi (1828), Pasquale Falco and Anna Falco (both 24 March 1830). **Every one of them came through a FALCO MOTHER or a «*da lui dichiarante*», and no index could have reached any of them. TWO PUBLISHED NILS OVERTURNED — 1817 and 1838, both «no Falco», the strongest negative this archive publishes. Three di Guida acts recovered, one of them closing a prediction the archive had made from a name. SARA RUGGIERO found to be a midwife, which explains ten parish pages the archive could not account for. Two act numbers corrected by the register's own order. And six volumes found to carry the same archivist's card across their index — 1816, 1823, 1824, 1825, 1830, 1846 — which turns work-list row 10 from a leaf into a policy. NOTHING REMAINS ON THIS ROW.**

Raised 21 September 2026, from act 20 of the Arienzo marriage register for 1854 — a volume nobody here had opened. The act says: «*FRANCESCO FALCO maggiore, di anni quarantotto, nato in Arienzo, di professione contadino, domiciliato strada CAMELLARA, figlio del fu RAFFAELE … e della fu CHIARA RIVETTI … VEDOVO DI RAFFAELA FALCO*». This archive says: FRANCESCANTONIO FALCO, married 16 July 1832, entered as a child of PASQUALE FALCO and Chiara Rivetti — and his wife is a RAFFAELA FALCO, the two of them being the parents of PASQUALINA FALCO, dead 1836. One wife's name, two fathers. Either there are two men, or one of the filiations is wrong. IT IS THE SAME SHAPE AS THE MICHELE-AND-MICHELANGELO PROBLEM, WHICH WAS SETTLED ON 20 SEPTEMBER AND WENT THE OTHER WAY. There, two names that looked like one man were two — MICHELANGELO a son of Pasquale, MICHELE a son of Raffaele — and both married an Antonia Vigliotti. Here the pattern repeats to the letter: two forms of one forename, two fathers of those same two names, and wives who share a name. *Which is a reason to expect two men and not a reason to conclude it.* How to settle it, cheapest first. 1. Read the 1832 marriage act. This archive has Francescantonio's marriage from an index, and an index is not an act. Arienzo Matrimoni 1832 = `an_ua14236`. The act will name his father. 2. Pasqualina Falco's death act of 1836 — `an_ua14338` — already read here for Margarita Falco. It names her parents; whether it says Francesco or Francescantonio is the second discriminator. 3. Raffaela Falco's own death act. Francesco is her widower by August 1854 and the Arienzo death registers 1844–1858 are read out — so if she died in that window she has already been seen, and if she has not, she died before 1844. Nothing is to be merged until one of the three is read. Both readings stand in the household file, each beside the other. --- ## TEST 1 RUN — 22 September 2026. THE INDEX WAS RIGHT, AND THAT MAKES IT WORSE Arienzo, Matrimoni 1832, act 8 (`an_ua14236` img 9R), read from the act: > «*sono comparsi nella Casa Comunale FRANCESCANTONIO FALCO, maggiore, di anni ventisei, nato in Arienzo, di professione contadino, domiciliato ivi strada CAMELLARA, figlio di PASQUALE FALCO —, di professione contadino, domiciliato ivi, e di CHIARA RIVETTI, di professione contadina, domiciliata ivi; e RAFFAELA FALCO, MINORE, di anni diciannove … figlia di GIUSEPPE FALCO, di professione COLONO, domiciliato ivi, e di ANTONINA DI GUIDA, domiciliata ivi*» PASQUALE. The clerk drew a dash where «defunto» goes and left it empty. ### The two acts now line up field for field | | 1832, act 8 | 1854, act 20 | |---|---|---| | name | Francescantonio Falco | Francesco Falco | | age | 26 → b. c.1806 | 48 → b. c.1806 | | trade | contadino | contadino | | street | strada Camellara | strada Camellara | | father | PASQUALE, *alive* | fu RAFFAELE | | mother | CHIARA RIVETTI, *alive* | fu CHIARA RIVETTI | | wife | RAFFAELA FALCO | *widower of* RAFFAELA FALCO | Everything agrees except the father — and neither act contradicts itself. Pasquale Falco died 1 October 1834, twenty-seven months after being called living in 1832. Raffaele Falco was dead before February 1850. Both filiations are internally sound. ### What that leaves Either two men of one name, born the same year, on the same street, each with a mother called CHIARA RIVETTI and a wife called RAFFAELA FALCO — which is the Michele-and-Michelangelo answer, and it was the right answer there — or one clerk wrote one wrong forename. AND A THIRD HUSBAND HAS TURNED UP. This archive also holds «Michelangelo Falco & Raffaela Falco», from a death act of 16 March 1834. Three men, three households, one wife's name. *Michelangelo is a son of Pasquale. Francescantonio is a son of Pasquale. Francesco is a son of Raffaele.* The overlap is no longer a coincidence to explain away; it is the shape of the problem. ### The next test, and it is cheaper than any of the three first listed RAFFAELA FALCO WAS NINETEEN AND A MINOR IN JULY 1832, SO BORN ABOUT 1813, AND SHE IS A DAUGHTER OF GIUSEPPE FALCO AND ANTONIA DI GUIDA. Her birth act should be in the Arienzo births of 1812–1814 — `an_ua14409`–`an_ua14411` — three volumes this archive has never opened, and they are on the 1809–1815 list row 9 has just been corrected to name. If there is ONE Raffaela Falco of that generation, the three households are one and two of the three husbands are spellings. *If there are two, the Michele answer applies again.* Nothing is merged until that act is read. ### 1813 read out the same day — she is not in it `an_ua14410`, 53 acts, every one read. One Falco in the year, and it is the *other* Giuseppe Falco: act 2 of 5 January 1813, «*GIUSEPPE FALCO, trentatré, contadino, strada COSTA … da GELSOMINA VIGLIOTTA … un maschio … ANTONIO*». No Raffaela Falco. No Antonia di Guida. Nineteen in July 1832 means born between July 1812 and July 1813, so the window is now 1812 alone — `an_ua14409`, 93 images — and its second half first. *1814 (`an_ua14411`) is outside the window unless the clerk rounded her age down, which these clerks do.* ### 1812 read out too — and there is no Falco in it at all `an_ua14409`, 93 images, every act read. Not one Falco in the year, as father, mother or child. With 1813 done the same day, the whole of her stated birth window is covered and she is absent from it. So one of three things. Her age was rounded — *the same day's reading produced a groom called «thirty-two» whose birth act makes him thirty-one* — which puts 1811 (`an_ua14408`) and 1814 (`an_ua14411`) next. Or she was not born at Arienzo. Or her birth was never registered. The 1812 tavola cannot shortcut it: image 92 is «*Tavola Annuale Alfabetica de' Nati … Anno 1812*», alphabetical by the child's forename, and the archivist's card «THIS INDEX HAS BEEN MICROFILMED ALREADY» covers all but its first page. 1812 joins the blocked-index list. ### A THIRD CHIARA RIVETTI — 22 September 2026 MARIA FALCO, wife of Francesco Cioffi of strada Camellara, died 1 May 1862, and her act (`an_ua14364` img 19R, act 33) calls her «*figlia del fu GIUSEPPE, di professione contadino … e della fu CHIARA RIVETTI*». So the tally is now: | husband | wife | how this archive got him | |---|---|---| | PASQUALE FALCO | Chiara Rivetti | a household of long standing | | RAFFAELE FALCO | Chiara Rivetti | separated from Pasquale on 20 Sep 2026, by magnifying one capital R | | GIUSEPPE FALCO | Chiara Rivetti | this act, read at 3× on 22 Sep 2026 | Three forenames, one wife's name, one street. And the same pattern already sits at the head of this row for MICHELE / MICHELANGELO and FRANCESCO / FRANCESCANTONIO, each pair splitting between a father Pasquale and a father Raffaele. THAT IS NOW A PATTERN AND NOT THREE COINCIDENCES. Either the Falco of strada Camellara really did run three brothers who each married a Chiara Rivetti — which is not credible — or CHIARA RIVETTI is one woman and the clerks of Arienzo wrote her husband's forename three different ways across forty years. *If that is so, then Michele and Michelangelo are one man, Francesco and Francescantonio are one man, and the correction of 20 September went the wrong way.* Nothing is merged on this reasoning. What would settle it is Chiara Rivetti's own death act, which names her husband once and for all — and the Arienzo deaths are read act by act for 1844–1862 with no Chiara Rivetti in them, so she died before 1844, in the years this archive has read only in part. --- ## MICHELANGELO'S OWN ACT, AND THE CASE FOR ONE MAN — 22 September 2026 Arienzo, Matrimoni 1828, act 17, 6 December (`an_ua14232` img 18R), read from the act: > «*sono comparsi nella Casa Comunale MICHELANGELO FALCO, MINORE, di anni ventiquattro, nato in Arienzo, di professione contadino, domiciliato ivi strada CAMELLARA, figlio di PASQUALE FALCO, di professione contadino, domiciliato ivi, e di CHIARA RI=/VETTI, domiciliata ivi; ed ANTONIA VIGLIOTTA, MAGGIORE, di anni ventisette … figlia di FRANCESCO VIGLIOTTA DEFUNTO … e di VITTORIA REPETELLA*» ### The two acts, field by field | | 1828, act 17 | 1850, act 6 | |---|---|---| | name | MICHELANGELO Falco, *minore* | MICHELE Falco | | age | 24 → b. c.1804 | 43 → b. c.1807 | | trade | contadino | colono | | street | strada Camellara | strada Camellara | | father | PASQUALE, *alive* | fu RAFFAELE | | mother | CHIARA RIVETTI, *alive* | fu CHIARA RIVETTI | | wife | ANTONIA VIGLIOTTA | *widower of* ANTONIA VIGLIOTTA, remarrying | ### Three things this archive already held, which nobody had put together One. Both parents die between the two acts, in the right order. PASQUALE FALCO died 1 October 1834; CHIARA RIVETTI died 13 September 1843. In February 1850 both would be marked dead, and the 1850 act marks both dead. Its only disagreement with 1828 is the father's forename. Two. There is exactly ONE Chiara Rivetti death in the Arienzo registers, and she is Pasquale's widow. The civil deaths are now read act by act for 1816 to 1865 — fifty consecutive years. The single act is 13 September 1843: «*CHIARA RIVETTI VEDOVA DI PASQUALE FALCO, nata in Arienzo, di anni OTTANTA*». If a second Chiara Rivetti had been married to a Raffaele Falco and died before February 1850, she should be in those fifty years. She is not. Three. The parish says the same. The Arienzo baptisms hold «MICHELE ANGELO ANTONIO FALCO, baptised 25 September 1808, mother CHIARA RIVETTI» — *a machine transcription, unverified, and named as such.* A child christened «Michele Angelo Antonio» can be called MICHELE by one clerk and MICHELANGELO by another for the rest of his life. ### What this does to the correction of 20 September That correction split the two men on one magnified capital R, and it was careful and well argued. The weight has now moved the other way, and it has moved on records this archive already held. *Either two men of one christening name, born three years apart on one street, each with a mother called Chiara Rivetti and a wife called Antonia Vigliotta — or one man and one clerk's slip.* NOTHING IS MERGED. Both households stand, each pointing at the other. What would settle it is his own death act: he was alive on 7 February 1862, declaring his son Pasquale's death, and the Arienzo deaths are read to 1865 with no Michele or Michelangelo Falco in them — so he died after 1865, where this archive cannot follow, which is exactly where Giuseppe Falco and Antonia di Guida went too. *The next best is the 1808 baptism read from the image, which names the father; it is a FamilySearch full-text page and sits behind row 18.* --- ## HALF OF THIS ROW IS SETTLED — 22 September 2026 FRANCESCO FALCO of 1854 IS FRANCESCANTONIO FALCO of 1832. The chain is three records this archive already held: 1. 6 Jul 1832, from the act — RAFFAELA FALCO, of Giuseppe Falco and Antonia di Guida, marries FRANCESCANTONIO FALCO, «*figlio di PASQUALE FALCO … e di CHIARA RIVETTI*», strada Camellara. 2. 20 Apr 1854, cited on the register page — «*RAFFAELA FALCO, 34, contadina, strada CAMELLARA, MOGLIE DI FRANCESCO FALCO*», parents Giuseppe Falco and Antonia di Guida. 3. Aug 1854, four months later — a FRANCESCO FALCO of strada Camellara marries again as «*VEDOVO DI RAFFAELA FALCO*». So the 1854 act's «figlio del fu RAFFAELE» is a clerk's error, and the household «Raffaele Falco & Chiara Rivetti», created here on 21 September, rests on it. *It is left standing with the correction attached.* See the corrections page. ### What is left Only the MICHELE / MICHELANGELO half, and it now leans the same way: Michelangelo's own act says PASQUALE; the parish christened a son of that couple MICHELE ANGELO ANTONIO in 1808; both men are of strada Camellara, both married an ANTONIA VIGLIOTTA, and there is one Chiara Rivetti death in fifty years of registers, «*vedova di PASQUALE FALCO*». NOTHING IS MERGED. What would settle it: Michele/Michelangelo's own death act — he was alive 7 February 1862 and is in no Arienzo register to 1865, so he died after the series ends; or the 1808 baptism read from the image, which names the father and sits behind row 18. AND ONE LEAF IS OWED. Raffaela Falco's death act of 20 April 1854 (`an_ua14356`) is a citation in this archive's register and has never been re-read at full resolution. It should be — it is now load-bearing, and today's whole lesson is that a swept or cited row becomes evidence the moment something is built on it. --- ## THE OWED LEAF IS READ — 23 September 2026 `an_ua14356` img 24, act 34, at 3,200 px. *Everything this row built on the citation survives the image.* > «*… sono comparsi MATTEO C[RISCI], di anni SESSANTASEI … e CLEMENTE C[RISCI], di anni QUARANTA … domiciliati strada PORTA DI SOPRA, i quali han dichiarato, che nel giorno VENTI … di APRILE … alle ore UNDICI, è morta nella sua propria casa RAFFAELA FALCO, MOGLIE DI FRANCESCO FALCO, nata in Arienzo, di anni TRENTAQUATTRO, … strada CAMEL[LARA], figlia di GIUSEPPE FALCO … domiciliato ivi, e di ANTONIA GUIDA, domiciliata ivi*» «MOGLIE DI FRANCESCO FALCO» is written out in full and is not «Francescantonio». The half of this row settled on 22 September stood on a register citation; it now stands on the act. AND THE QUESTION THIS ROW'S OWN RECORDS ASKED IS ANSWERED. Both household entries instructed that the act be checked for whether it marks ANTONIA DI GUIDA living too. It does, identically — no «fu», no «defunta», a domicile written out, in the clause after her husband's. *The act does not distinguish them.* It changes no date, because she is already held living on 15 November 1857; what it removes is an asymmetry in this archive, where one parent was read as living from an act that says the same of both. Carried as a weak inference, as the civil series requires. New: two declarants — MATTEO, 66, and CLEMENTE, 40, *regnicoli* of strada PORTA DI SOPRA, neither the husband nor a Falco. Surname faint, reading as CRISCI, not claimed. Their shared trade is recorded UNREAD. THIS FILMING STOPS AT ~1,800 PIXELS A PAGE. The opening is 3,659 × 2,763 native; above ~3,200 px the server returns 403 «pixel area exceeds the maximum threshold», and a IIIF *region* request at native scale — which that cap does not apply to — resolves no further. *The word was magnified seven times and left unread rather than guessed.* What remains on this row is the MICHELE / MICHELANGELO half alone, which this act does not touch. --- ## A FOURTH CHIARA RIVETTI — 23 September 2026, and it weakens this row's central argument Arienzo, Matrimoni 1858, act 23, 26 August (`an_ua14262` img 25), read from the image: > «*comparsi nella Casa Comunale SAVERIO GUIDA, di anni ventisei, nato in Arienzo, di professione contadino, figlio di LUIGI GUIDA DEFUNTO, di professione agrimensore, domiciliato ivi, e di CHIARA RIVETTI, domiciliata ivi*» THE SURNAME WAS PROVED, NOT ASSUMED. In this hand «Rivetti» opens with a stroke readable as a capital V. The control is in the same volume: act 6 writes «Lucia Rivetti» and «Maria Rivetti» in the same pen with the identical initial, and writes «Arricale» with a plainly different A. SHE IS NOT THIS LINE'S CHIARA RIVETTI, AND THE PROOF NEEDS NO HANDWRITING AND NO «FU». Pasquale Falco's widow died 13 September 1843, «*di anni OTTANTA*» — born about 1763. Saverio Guida was twenty-six in August 1858, so born about 1832. *A woman born in 1763 does not bear a child in 1832.* WHAT IT DOES TO THE ARGUMENT ABOVE. This row reasons that three Falco men each recorded with a wife «Chiara Rivetti» is not credible, so she is one woman and the clerks wrote her husband's forename three ways — and it leans on there being exactly one Chiara Rivetti death in fifty years of Arienzo registers. Here is a Chiara Rivetti alive at Arienzo in 1858, married to a man of a wholly different surname, who is ALSO absent from those fifty years. *So the absence of a second Chiara Rivetti death does not show there was only one Chiara Rivetti — it shows the death series does not reach them.* The name is commoner at Arienzo than this row assumed, and the case for collapsing MICHELE into MICHELANGELO — which would reverse the correction of 20 September — is correspondingly weaker. NOTHING IS MERGED AND NOTHING IS UNMERGED ON THIS. It is recorded as evidence against an argument, not as an argument of its own. The Michele / Michelangelo half of this row stays open, and what would settle it is unchanged. *Found while re-reading 1858 and 1859 — a sweep this archive had already done and recorded, which is how it came to be read a second time. See rows 39 and 40.* --- ## THE 1808 BAPTISM IS READ FROM THE IMAGE — 27 September 2026 This row's third argument has rested on a machine transcription since it was written, and said so. It no longer does. `3:1:3Q9M-CSD3-13FJ-4`, pulled at 3,840 px: > «*Anno Dni 1808, die vero 25 Septembris, R[everen]dus D. JOSEPH RIVETTA de nostri licentia baptizavit infantem natum ex PASCHALE FALCO, et CLARA RIVETTA conjugibus hujus Parochiae, cui dedit nomen MICHAEL ANGELUS ANTONIUS. Matrina fuit FORTUNATA VILLANO.*» A child christened «MICHAEL ANGELUS ANTONIUS» to PASQUALE FALCO and CHIARA RIVETTI — which is what his own marriage act of 1828 says, in Italian, twenty years later. A boy of that christening can be written MICHELE by one clerk and MICHELANGELO by another for the rest of his life. WHAT THIS DOES AND DOES NOT SETTLE. It removes the weakness this row named in itself: the evidence is now read, not transcribed by a machine that has been caught turning 1802 into 1902. It does not merge anybody. The 1850 act still calls a Michele Falco «*figlio del fu RAFFAELE*», and this archive does not merge people on a name — *the fourth Chiara Rivetti found on 23 September is the standing reason why not*. The balance of the evidence has moved and the refusal has not. Both households stand, each pointing at the other, and what would settle it is still his own death act — after 1865, where Antenati stops.

Raised 22 September 2026, out of an audit of whether the site had kept up with the research. `tools/build_marriedin.py` builds from the MyHeritage tree and joins to this archive on the tree's own id — correctly, because 103 of the 294 people on that page share a name with somebody else on it, so a name join would point a third of the links at a guess. The consequence is that a person documented only from an act is not on the page at all. Twelve of the thirteen people this archive marks «married in» are missing, among them ANGELA CARFORA, who married Giuseppe Falco in December 1854 and whose whole filiation is read from the act; VINCENZO ABATIELLO, who married Angela Falco in 1865; CORNELIA SABATINO, who married Francesco Falco in 1854; and FILOMENA DI LUCIA, who married Stanislao Falco in 1861. What was fixed on the day, and what was not. Where the tree is blank and the archive is not, the build now uses the archive's own dates — twelve fields, four births and eight deaths. Mariantonia Carfora had no date span on that page while her birth act of 30 December 1830 and her marriage act of 21 January 1860 were both published and both cited on her own person page. *That fixed the dates of the people who are on the page; it did not put the missing ones on it.* Why it was not done in the same pass. Every surname on that page carries a reach — whether it carries Falco blood, married one, is such a family, or married into one of those — and that is computed by walking the tree's topology. A person with no tree row has no topology to walk, so admitting them means deciding what «reach» means for somebody the tree has never heard of. That is a design question, not a patch, and the page now states the gap in its own words rather than hiding it. The honest interim. The page says how many it cannot see, and the number is derived from the household file, so it cannot drift away from the truth while nobody is looking — which is the failure this whole audit was about.

The 1809–1815 death volumes are the earliest Arienzo holds, and until 21 September 2026 this archive had never opened any of them. Read out on both halves, 22 September 2026: 1809 (73 images, four Falco acts), 1810 (48 images, one Falco), 1811 (52 images, one Falco), 1812 (46 images, one Falco). Still owed: 1813 (`an_ua14315`, 49 images), 1814 (`an_ua14316`, 61), 1815 (`an_ua14317`, 49). Each volume needs its own layout check before the band is set. These four consecutive years use four different page layouts and three different act formulae. 1809 and 1810 are Napoleonic («*sono comparsi avanti il sottoscritto Sindaco*», deceased at 40–60% of the page); 1811 is the restoration form («*Avanti di noi incaricato del registro*», deceased at 44–58%, and it counts and names the surviving children); 1812 is a third form («*L'Anno milleottocento dodici*», deceased at 32–47%) that gives both parents of the deceased in every act. CLOSED 22 September 2026. All seven volumes of 1809–1815 are read out on BOTH halves. 1813, 1814 and 1815 were read the same day: they gave six Falco women married into other Arienzo houses, a son of Giuseppe Falco and Antonia di Guida who was not in this archive, Giuseppe Falco's own stated age, and two siblings for Antonia di Guida with both their parents named.

Raised 22 September 2026, out of the discovery that two volumes published as «read in full» were half read. A band sweep of one half of each image reads half the acts consistently, so the output looks like a complete sweep and nothing appears to be missing. In 1809 and 1811 both Falco acts that had already been found happened to sit on right pages, so the method confirmed itself. The unread halves held four more Falco, one of them a couple a generation above this archive's head of line. The check is cheap: render one full page, read the act number on each half, and compare image *n*'s left-page number with image *n+1*'s. If they differ by two, there are two acts per image. What this row is for: every volume in `notes/` whose progress file records a `--half L` or `--half R` band needs that check run against it, and its coverage row amended or confirmed. Until then, every nil published from a half-page sweep is only as good as its volume's layout. [[wrong-negatives]]

Read out, both halves: 1812 (93 images, no Falco at all), 1813 (53 acts, one Falco — a child of the *other* Giuseppe Falco), 1814 (66 images — RAFFAELE FALCO, b. 26 May 1814, of Giuseppe Falco and Antonia di Guida), 1815 (63 images — MARIA ANTONIA FALCO, b. 17 December 1815, of the same couple, and a daughter of the other Giuseppe Falco). Still owed: 1809 (`an_ua14406`, 72 images), 1810 (`an_ua14407`, 67), 1811 (`an_ua14408`, 50). EVERY VOLUME NEEDS ITS OWN BAND, AND THE 1814 ONE PROVES WHY. A first attempt at `--band 0.33 0.32` on 1814 read the witnesses and the child's given name and missed both parents — it would have returned a clean, worthless nil. The forms differ year by year: - 1811 — the early restoration form, father («*del fu X*», with a patronymic) at ~11% of the page and the mother at ~26%: `--band 0.09 0.24 --half LR --cols 2 --width 2200`. *At width 2600 the Antenati tiler returns «Java heap space».* - 1814 — the printed form, declarant at ~26%, mother at ~38%: `--band 0.20 0.28 --half LR --cols 3 --width 2600`. - 1815 — declarant ~19%, mother ~27%, child's name ~33%: `--band 0.16 0.20 --half LR --cols 3 --width 2600`. Two acts to an image in every one of them. [[half-page-sweeps]] 23 September 2026 — 1811, 1810 AND 1815 CLOSED; 1809 IS STILL OWED AND IS THE LAST VOLUME OF THE DECADE. 1809 (`an_ua14406`, 72 images) is READ ONLY FOR IMAGES 45–50. The rest of the volume did not render: after a day of sweeping, Antenati began returning «*curl: (28) Operation timed out*» on most tiles and «*500 Internal Server Error — Java heap space*» on others. That is a rate limit, not a missing volume, and the twelve panels that did render produced a Falco act on the first sheet. [[antenati-method]] What images 45–50 gave: «*CARL'ANTONIO MORGILLO DEL FU CARL'ANTONIO … strada della MADDALENA … con ROSA FALCO sua Moglie*», son Raffaele Morgillo born 1 September 1809 — the earliest record of that household, and the act that proves her husband is one man under three forenames. Resume at image 3 with `--band 0.17 0.22 --half LR --cols 2 --width 2200` — the same band as 1810, whose printed form 1809 shares. CLOSED 23 September 2026. ALL SEVEN BIRTH VOLUMES OF 1809–1815 ARE READ OUT ON BOTH HALVES — and with row 89, so are all seven death volumes. The earliest decade of Arienzo civil registration is finished in both series. 1809 needed two passes. Antenati began timing out after a day of sweeping and only images 45–50 rendered; it recovered after a pause and the rest was read. *A timeout is a rate limit, not a missing volume.*

Raised 23 September 2026, after four household names were published twice in one night with every gate green. A gate now catches it — `tools/check_households.py` — but the gate treats a symptom. The diagnosis came from the session that maintains the shared kit, which ran the question across every archive in the estate that holds a households file: | archive | households | identity | |---|---|---| | D'Arcy | 199 | the couple — *(a, b, married, place)* | | Defranceski | 127 | `key` / `id` | | Mazza | 276 | `headId` | | this archive | 111 | `name` | Three of the four cannot have this bug, not because their scripts are more careful but because appending a duplicate would create a visibly duplicate id. Mazza has five households whose NAMES repeat — «Orazio Mazza» twice, «Maria Mazza» three times — and none whose `headId` repeats: five genuinely distinct people, held apart correctly, by an archive whose rule is that it never merges on a name. This archive merges households on a name while refusing to merge people on one. `H.append({...})` without a lookup, and `by.pop(name)` leaving the list entry behind, are both only possible because the index is keyed on a display string. WHAT A FIX WOULD LOOK LIKE, AND WHY IT IS NOT DONE HERE. 756 of 757 member rows already carry an `ark`. A household keyed on its father's and mother's arks, with the name kept as a label, would make the duplicate check *structural* rather than a gate, and would close the second check too — a member's back-reference would point at an id instead of a string that has to match character for character. It is a change to the shape of this archive's central file, touching 111 households, 757 rows, five builders and `households.tsv`. That is David's call, not a patch to make at the end of an unattended session. The gate holds the line until then. DONE 23 September 2026, on David's own ruling in this session — and not on the shape that was proposed, because that shape is not available. THE ARK-PAIR KEY CANNOT BE DERIVED FROM THIS DATA. «756 of 757 member rows carry an ark» is true and does not translate: a household key needs the father's ark and the mother's ark to be different, and 75 of 111 households cannot supply two distinct parent arks. The commonest reason is not a gap — it is the archive working correctly. A couple named together in one act share ONE ark, because they are one document: «*Nicola Rivetti & Anna di Ruggio*» are both `an_ua14345 img30r atto43`. And eleven households are a single person with no spouse at all — «*Tommaso Zampiello*», «*Gioacchino Falco (strada Camellara)*». WHAT WAS DONE INSTEAD: a frozen assigned id. Every household carries `id`; every member carries `hid`; `data/households.tsv` carries an hid column; `build_people.py` keys people on `id` instead of on the display string; `check_households.py` fails on a missing or duplicated id and on a member whose `hid` does not match its house; and `check_release.py` matches the two household files on the id rather than the name. THE NAME IS STILL PARSED, AND THAT IS CORRECT. «X & Y» is this archive's statement of who the parents are — it is *data*. What it is no longer is *identity*. THE REKEY MOVED NOTHING. Zero person slugs changed, zero relationship edges changed, zero records changed household. A rekey that alters behaviour is a rekey that was doing something else. WHAT IT BUYS, IN ONE EXAMPLE. `check_release.py` has carried a hand-written rename table — `{"Francesco Cossi & Maria Falco": "Francesco Cioffi & Maria Falco"}` — purely because a household was corrected in one file and not the other, and the gate would otherwise fail on the archive's own correction. With an id, a rename is one field and that table is dead weight. The «Michele Crisci → Nicola Crisci» correction that produced a duplicate household on 22 September could not have done so. AND THE NEW CHECK HAD THE NIGHT'S OWN FAULT IN ITS FIRST DRAFT: the id-resolution test in `check_release.py` printed «FAILED» and exited zero, because it set a flag no other line read. *A gate that reports and does not refuse is the same thing as no gate.* Fixed and tested by injecting an unresolvable hid.

Found 23 September 2026 by a date check, not by a name check — which is the only reason it was found at all. `/people/angela-maria-falco/` currently carries four parents: - from the registers — a daughter of GIUSEPPE FALCO and ANTONIA DI GUIDA, born 19 January 1822 by an act that names her grandfather, married 15 November 1857; - from the family tree (id 4512504) — born 1772, died 14 September 1850, daughter of FRANCESCO FALCO and MADDALENA SICILIATTO. Fifty years, different parents, one page. They were joined on «name alone», which is the weakest basis `build_people.py` accepts and — by construction — the basis it uses precisely when the years do NOT agree: `best == 0` means the two sides' years failed to overlap, and the merge proceeds anyway if each side has exactly one candidate. A GUARD WAS WRITTEN AND WITHDRAWN THE SAME HOUR. Refusing «name alone» wherever the years failed to overlap cost ten merges to fix this one, and nine of the ten were correct: Alessandra Crisci has a tree birth of 1817 and a register marriage of 1834 — *years that do not overlap and do not contradict either, because a birth year and a marriage year never overlap.* Absence of agreement is not disagreement, and the first version of the rule could not tell them apart. What a correct guard needs is to compare like with like — the household's BIRTH event against the tree's `b`, its DEATH event against the tree's `d` — and to refuse only when one of those pairs both exists and disagrees. *That is a careful piece of work on the archive's central builder and it was not done at the end of an unattended night.* Until then the build prints the conflict on every run — «*ACTS THAT CONTRADICT A TREE DATE — the act is used, the MERGE is suspect*» — so it cannot go quiet. The act's date is what the page shows, on the archive's own ranking of line over register over tree. The likely answer is that the tree's 1772 woman is not this archive's at all: a Francesco Falco and a Maddalena Siciliatto appear nowhere in the registers read so far. Unmerging is probably right, and proving it is the work. [[two-falco-families-arienzo]] DONE 23 September 2026, on David's own ruling in this session: unmerge AND fix the rule. The proof: every Siciliatto row in this archive is sourced «*the family tree (unverified unless a record is cited)*» — none in `households.tsv`, none in the notes — and the tree gives Francesco Falco's own parents as «*[Unconfirmed Family] Falco*». The 1772 woman rests on nothing this archive has read. She keeps a tree-only page, because an unverified person is not a deleted one. The rule now compares like with like and refuses exactly one merge, with its reason printed: «*angela maria falco — birth [1822] vs tree [1772]*». The fourteen other «name alone» merges stand, including Alessandra Crisci, whose 1817 birth and 1834 marriage the first attempt wrongly called a contradiction.

Found 23 September 2026 while checking a merge, not by looking for it. DOMENICA FALCO has three parents on her page: *Alessandra Crisci*, and Matteo Falco twice — `matteo-falco-2` from the tree and `matteo-falco-of-matteo-falco-alessandra-crisci` from the registers. They are one man held as two records, so every child of that couple shows him twice. Sixteen people are in that state. It is the Luigi Falco fault one level up: the name test needs exactly one candidate on each side, and a household head whose name this archive holds more than once never gets one. The parents-and-exact-date pass will not reach them. It requires the person to be a CHILD in a household with a dated birth or death event; a household HEAD is matched by the generation passes or not at all. What it would take: a pass for household heads on the same standard — *both spouses named, and a date that agrees* — refusing unless the pairing is unique in both directions. The couple identifies the generation on both sides, which is the argument the direct-line passes already use, so it is an extension of something published rather than a new claim. *It was not done on the night it was found, after a rule written in the same session cost ten good merges before it was reverted.* [[person-graph-method]] --- ## DONE 23 September 2026 — sixteen became one, and the one is on the record The head pass was written on the standard this row proposed and no wider: a household HEAD is paired with a tree person when his own name and his wife's both agree, and only when the pairing is unique in each direction. It made five merges. The parents-and-exact-date pass was extended from death dates to birth dates as well, keyed by kind so a birth is never matched against a death; it made four. Sixteen people had a parent listed twice. One does. *Filomena Annecchino*, whose Carmine Annecchino is two records the merge rule is right to refuse — the tree gives him no wife, so the couple cannot identify him — and that refusal is now published as a `same-as` verdict on her page rather than left to look like a bug. A PARTICLE NEARLY COST THE LARGEST GROUP. The first version stripped particles only at the START of a name, so «Antonia di Guida» was never normalised and the di Guida house went unpaired. Dropping *di, de, del, della, dello, dei, lo, la, d'* wherever they stand took the head merges from three to five. And the pass found what it was not looking for: three households for one couple. «Giuseppe di Guida & Maddalena Rivetta», «… & Michedalena Rivetti» and «Giuseppe di Guida (strada Maddalena)» are one house — Antonia di Guida's own parents' — and two Falco parent-name variants were folded with them. 108 households, from 111. WHAT IT COST, MEASURED AND NOT ASSUMED. Thirteen person pages merged away and every one redirects; the built estate went 10,862 pages to 10,863 and not one URL stopped resolving. The row floor fell — rows 293→285, spouse 577→561, grade 277 from 285 — and the fall was accounted to the act: eight spine rows across four family pages, each carrying two spouse marks and one grade mark, 8×2=16 and 8×1=8. Seven were the merged-away duplicates; the eighth was Angelo Andrea di Guida, who had been drawn on the Guida page twice. `place` rose 52→53 and `household` 987→993. *The floor was rewritten from a build made for the purpose, in a directory no other session writes to, and the counts were compared page by page against a clean build of the previous commit — not inferred from the data files, which count something else.* `check_households` now FAILS on a duplicated parent role regardless of qualifier. The trailing-qualifier exemption is this archive's way of telling two children of one name apart and stays for child roles; a house has one father and one mother, so on a parent role there is no exemption to claim.

Raised 23 September 2026, out of the Arricale work for row 9. She matters because of who she is. PASQUA FALCO is the mother of GIOVANNA ARRICALE, generation five's wife — so she is a FALCO woman standing one generation above a Falco marriage, and whether she attaches to this line is a question about the direct line itself, not about a neighbour. What this archive holds of her is two documents, both her daughter's. From Giovanna's birth act of 22 December 1848 (`an_ua14445` atto 114): «*la stessa è nata da PASQUA FALCO moglie di FRANCESCO ARRICALE, di anni VENTICINQUE, domiciliata STRADA CAPODICONCA*» — so born about 1823. Her own birth act has never been found, and the *Tavola Alfabetica* of Arienzo Nati 1823 has a paper slip lying across the column where FALCO falls — an obstruction, not a nil. SHE IS NOT THE OTHER PASQUA FALCO, AND THAT REFUSAL IS SOUND. The Pasqua Falco born 30 March 1823 to Giuseppe Falco and Gelsomina Vigliotta (atto 38) married ANGELO MAJONE on 9 September 1856 and is not described as a widow in that act — so she cannot be a woman whose husband Francesco Arricale was alive, and Angelo Majone is himself alive in 1865. Two women of one name and one birth year. ## THE NEW OBSERVATION THE MARRIAGE IS IN NONE OF THE ARIENZO MARRIAGE INDEXES. This archive read all 116 images, 1809–1849 (`an_ua14108`+, then `14107`+ from 1837; 1836 does not exist), and each line of that tavola gives both spouses, BOTH sets of parents and the church date. Giovanna was born in December 1848 to a woman already called «*moglie di*», so the marriage is 1848 or earlier and falls squarely inside the read range. So one of three things is true. 1. The index reading missed it. Possible — *the sort order of these tavole is not stable, early years by forename and the 1820s–30s by surname* — but forty volumes were read end to end and twelve Falco marriages came out of them. 2. They married at another comune. And this archive has already established the rule that decides which: the civil marriage act is registered at the BRIDE's comune — proved on Raffaele Falco's marriage of 1844, which is at Santa Maria a Vico and not at Arienzo. If the marriage is elsewhere, then PASQUA FALCO WAS NOT AN ARIENZO WOMAN — which would also explain why her birth is not in the Arienzo register, and would mean the paper slip across the 1823 tavola is hiding nothing. 3. She is in the index under ARRICALE and was read past. The tavola is alphabetical by one spouse; a FALCO bride filed under her husband's A would sit far from the F section that Falco work naturally reads. THE CHEAPEST TEST IS (3), AND IT IS ONE LEAF PER YEAR. Re-read the A stretch of the Arienzo marriage tavole for 1843–1848 — `an_ua14141`–`an_ua14146` on the post-1837 mapping — looking for ARRICALE, not for Falco. *One to four images a year; six years is about a dozen leaves.* If it is there, the line itself gives her parents and the question is closed without a birth act at all. IF IT IS NOT THERE, (2) BECOMES THE READING, and the next question is which comune — for which her daughter's or her own later documents are the lever. *This row exists because a nil in a fully-read finding aid is evidence, not a gap — provided the finding aid really was read the way the question needs. Reading it for FALCO is not the same as reading it for ARRICALE.* [[wrong-negatives]] --- ## TEST (3) RUN THE SAME DAY — the A stretch read for ARRICALE, and she is not in it One leaf per year, as this row said. The tavole are alphabetical by the GROOM's surname, two lines to an entry — groom, then bride — so an ARRICALE groom sits at the very front, far from the F section that Falco work reads. | year | ark | A stretch | ARRICALE / ARRIGALE groom | |---|---|---|---| | 1844 | `an_ua14142` | read | act 24 — ARRIGALE FRANCESCO × CRISPI CRESCENZA | | 1845 | `an_ua14143` | read | act 19 — Arrigale Angelo × Biondo Alessandra | | 1846 | `an_ua14144` | read | act 21 — Arrigale Vincenzo × Coppola Maria Michela | | 1847 | `an_ua14145` | read | act 20 — Arrigale Clemente × Ambrogio Giuseppa | | 1848 | `an_ua14146` | read | none — the list opens at *Ambrogio* | | 1843 | `an_ua14141` | NOT REACHED | the three leaves run C–Z; the A–B stretch is not in this filming | NOT ONE OF THEM MARRIED A FALCO. *The spelling in these tavole is ARRIGALE, which is the variant this archive already records as wandering between Arricale, Arrigale and Arricole in one decade — so the entries were there to be found and none of them is hers.* So explanation (3) is spent for 1844–1848. What remains is (1) the earlier reading missed it — now much less likely, since two independent passes over these years have not produced it — or (2) they married at another comune, which on this archive's own rule means PASQUA FALCO WAS NOT AN ARIENZO WOMAN. *That would also explain why her birth is not in the Arienzo register, and would mean the paper slip across the 1823 tavola is hiding nothing.* WHAT IS STILL OWED HERE: 1843's A–B stretch. The volume holds thirty-five acts («*La presente tavola contiene atti di matrimonj numero TRENTACINQUE. Arienzo li trentuno Decembre 1843*») across three leaves running C–Z. Either the year had no A or B groom, or a leaf is not filmed. *Until that is settled this row's negative covers 1844–1848 and not 1843.* ### AND THE PASS PAID FOR ITSELF ANYWAY «Francesco Arricale & Crescenza Crisci» now has a date. That household stood here with «—» for its event and «—» for its date; the 1844 tavola gives act 24. *The act itself is not yet read, and it is the act that will govern.* *Also seen and already held: FALCO ANGELANDREA, act 3 of 1847, of VINCENZO FALCO e ANDREANA CRISCI, marrying Maria Cimmino — the tavola's last column puts the CHURCH wedding at 4 May 1847, where this archive holds the civil promise of 29 March.*

The last impossible pair in the person graph, and this archive cannot fix it — the contradiction is inside MyHeritage. `/people/vittoria-cimmino/` carries born 1877 and died 7 February 1857. There is no merge to withdraw: both fields come from the tree, `treeId 4512393`, and no register act names her at all. SHE IS NOT WHO SHE WAS REPORTED TO BE. The finding reached this archive as «*the mother of Matteo Falco, the head of this archive's line*». She is not. Her son in the tree is `matteo-falco-3`, born 1807 and dead 7 April 1880 — and his father GIOACCHINO FALCO (b. 1775, d. 14 Feb 1851) is himself a son of MATTEO FALCO and FRANCESCA CRISCI, the founders. *So Vittoria Cimmino is the founders' daughter-in-law, and her son is their grandson, carrying his grandfather's name* — the necronym pattern this archive documents repeatedly. The head of the line is a different man, `matteo-falco`, born about 1744 and dead 18 July 1802. WHICH FIELD IS WRONG CAN BE SHOWN, AND IT IS THE BIRTH. She bore a son in 1807. A woman born in 1877 cannot. A death in 1857 is entirely ordinary for the wife of a man born in 1775 and dead in 1851. *So «1877» is the field that cannot stand, and «7 February 1857» is probably sound.* WHY IT IS WRONG IS OPEN, AND THERE ARE TWO GOOD READINGS. *(The second was put to this archive by the session that maintains the landing site, and it is at least as good as the first.)* 1. A mis-keyed digit — 1777 for 1877, which fits a wife of a man born 1775 exactly. 2. TWO WOMEN CONFLATED UPSTREAM — a *Vittoria Cimmino* of the 1770s and a *second Vittoria Cimmino born 1877*, merged into one profile in MyHeritage before this archive ever saw it. A woman of that name born at Arienzo in 1877 and marrying a Falco is entirely ordinary, and it would explain why the record carries a full birthplace and deathplace — both «*Arienzo, Caserta, Campania, Italia*» — rather than the bare year a typo usually leaves. THE EXPORT CANNOT TELL THEM APART. It holds one Vittoria Cimmino, `id 4512393`, carrying both dates; there is no second record to point at. *A single profile is what a typo looks like and it is also what an upstream merge looks like.* So the question is not which reading is right but which document would separate them — and that is the same document either way. NOTHING IS CHANGED ON THAT BASIS. This archive does not repair a source by guessing the digit it thinks was meant; 1777 is an inference, not a reading, and inventing it would be exactly the fault the corrections page exists to record. The pair stands as the tree gives it, flagged by `check:kin` on every build, and the ratchet is set to 1 so a second one fails. BLOCKED ON: a document. Any act naming VITTORIA CIMMINO, wife of Gioacchino Falco — her own death act, or her son Matteo's birth act of about 1807, which would name his mother and give her age. *Arienzo births for 1807 are before the civil series begins in 1809, so this is a parish question.*

Raised 27 September 2026, while giving every empty person-chart a stated reason. *(The blank charts were found by the session that maintains the landing site; this is what looking at them turned up.)* A marriage act names both spouses' parents, and this archive records them as members of the couple's household under roles like «father of the groom», «mother of the bride», «consenting father». `build_people.py` builds no edge from those roles, so eight people stand on their own page with no line to the child whose marriage names them. The clearest case is in the direct line, and it is BOTH parents. In the household of Vincenzo Falco and Andreana Crisci, the marriage act of 15 October 1814 names `pasquale-falco-of-vincenzo-falco-andreana-crisci-2-2` as «*father of the groom*» and `chiara-rivetti-of-vincenzo-falco-andreana-crisci` as «*mother of the groom*» — that is PASQUALE FALCO and CHIARA RIVETTI, generation two, named as the parents of VINCENZO FALCO, generation three. *Both stand on their own pages with no line to their son.* On the bride's side the same act gives `arcangelo-crisci-of-vincenzo-falco-andreana-crisci` as Andreana Crisci's consenting father and `sara-ruggiero-of-vincenzo-falco-andreana-crisci` as her mother. Note the slug: `…-2-2`. Generation two's Pasquale Falco already holds a record of his own; the marriage act's row is a SECOND record of the same man, which is precisely why drawing the edge is a merge and not a lookup. WHY IT IS NOT SIMPLY FIXED, AND WHY NOTHING WAS DRAWN TODAY. Building an edge from «father of the groom» means deciding which person on this site that father IS, and this archive already holds several Pasquale Falco. That is a merge, and merging on a name plus a role is exactly the fault corrected twice this month — Angela Maria Falco on the 23rd and Carmina and Pasqualina Falco this morning. *The role says a relationship exists; it does not say which record is the other end of it.* WHAT WOULD BE SAFE. The edge may be drawn when the parent named at the marriage is the same household member as the spouse's own recorded parent — that is, matched on the household's `hid` and the person string, not on a name across the archive. Where the couple's own household already names a father, the marriage act's «father of the groom» row is the same man and the edge is free. Where it does not, leave it undrawn and say so — which is what the page now does. *Until then these eight pages carry the caption «the act states a relationship that this chart does not yet draw», which is true, rather than «nothing is recorded», which would not be.*

Measured 27 September 2026, after the session maintaining the landing site asked what the estate's gates EXAMINE rather than what they are HANDED. `check:kin` refuses a person whose death precedes their birth. Every test it makes needs both a birth and a death, so it can only look at people who have both: | | people | | |---|---:|---| | both a birth and a death | 119 | *all `check:kin` can examine* | | a birth only | 178 | | | a death only | 118 | | | neither | 451 | | | total | 866 | | So «866 people · 1 impossible relationship · the ratchet is 1» is a statement about 119 people, not 866 — thirteen per cent. *The gate is not wrong and this archive is not faulty; it is what a parish reconstruction looks like when most people are named in a single act.* But the ratio was invisible in the output, and a ratio nobody prints is a ratio nobody checks. THE LARGER FACT IS ABOUT THE EVIDENCE, NOT THE GATE. 451 of 866 people here — fifty-two per cent — carry neither a birth year nor a death year. They are known from one act: a witness, a declarant, a grandparent in a «del fu» clause, a child named in a burial. *That is the true shape of this archive and it has never been stated on the site.* WHAT THIS ROW IS FOR. Two things, neither urgent: 1. Say the ratio wherever the number is quoted. `check:kin`'s own output will carry it after the next kit pin — the kit now prints «examined N of M». *This archive should not quote «866 people, 1 impossible» without the 119.* 2. Ask whether the other gates hide the same ratio. The date guard in `check_release.py` now prints what it examined and no longer only what it was handed. The rest have not been audited. --- ## AND IT IS NOT PECULIAR TO THIS ARCHIVE — measured across all seven, 27 September 2026 *(Measured by the session maintaining the landing site, after this row was written. Stated here because the row as first written implied the shape was this archive's own.)* | archive | rows | both | birth only | death only | no year | | |---|---:|---:|---:|---:|---:|---:| | Mazza | 592 | 65 | 145 | 3 | 379 | 64.0% | | Falco | 866 | 119 | 178 | 118 | 451 | 52.1% | | Lerena | 357 | 0 | 191 | 0 | 166 | 46.5% | | Booyzen | 354 | 113 | 100 | 18 | 123 | 34.7% | | Blazevic | 1833 | 492 | 741 | 16 | 584 | 31.9% | | Defranceski | 1445 | 427 | 545 | 54 | 419 | 29.0% | | D'Arcy | 243 | 129 | 40 | 4 | 70 | 28.8% | Between 29% and 64% in every archive in the estate, and no build had ever stated it. *So it is not a fault and it is not this archive's peculiarity: it is what a register of acts IS.* Falco sits second-highest, which is what a line reconstructed almost entirely from Arienzo parish and civil acts should look like — the more the evidence is single acts naming witnesses, declarants and grandparents, the larger the share of people with no year at all. The only check worth running on a census is that its parts sum to the whole: 119 + 178 + 118 + 451 = 866. *It does, here and in all seven.*

FOR DAVID, because this is his rule and his wife's family, and he knows the answer where this archive cannot. The living-people rule is the one this archive treats as absolute: a living person is a name and a relationship, never a date. `/method/` states it in those words. Two records sit at the edge of it. | slug | born | death recorded | classed as | |---|---|---|---| | `carmine-falco` | 1930 | none, in the tree or here | not living | | `domenico-prostamo` | 1930 | none, in the tree or here | not living | Both would be ninety-six. *The family tree's own flag means «no death date recorded», so the tree would call both of them living; this archive demoted them.* FOR CARMINE FALCO THE REASON IS WRITTEN IN THE CODE — `tools/build_living.py`: «*anyone in the ancestor chain is by definition not living*». He is generation eight, the direct line. That rule is plainly right for a line reconstructed from registers of the 1800s, and it becomes an assumption the moment the chain reaches the 1930s. *A direct ancestor is not dead because he is an ancestor; he is dead because the evidence says so, and here there is no evidence either way.* WHAT IS AND IS NOT AT RISK. Both pages are noindexed, so neither is offered to search engines, and `check_release` covers all 106 people the data flags living. But both men's birth dates are published on their own pages, which is the thing the rule forbids for a living person. *If either is alive, the archive is breaking its own rule — not through a fault in the machinery, but because a rule decided the question in advance.* NOTHING HAS BEEN CHANGED. This archive cannot resolve it: no act, no register and no tree field says whether either man is alive, and guessing in either direction is wrong. Flipping them to living would withdraw dates that may be perfectly proper to publish; leaving them is a bet that a rule of thumb held. *David can answer it in a sentence.* IF THE ANSWER IS THAT EITHER IS LIVING, the fix is not only the flag: the ancestor-chain filter in `build_living.py` should stop being an override and become a *tie-breaker of last resort*, applied only where no birth year places the person within a possible lifetime. A removal request is honoured within days, without argument — that promise is on `/method/` and it applies here first. --- ## A BRISBANE NEWSPAPER MAY ANSWER THIS, AND ITS TEXT IS WITHHELD — 27 September 2026 **Searching Trove for the exact phrase «*Carmine Falco*» returns two articles and they are the same story in two editions: > Telegraph (Brisbane, Qld.), 11 July 1963 — «*Man dies in fire at Spring Hill*»** > *article ids 295850296 and 295850513* The text cannot be read. Trove holds it inside copyright: «*This newspaper article is still within its copyright period and can't be displayed*». The search index found the phrase; the page will not show it. *So it is known that the words «Carmine Falco» stand in that column, and nothing more — whether as the man who died, a relative, a neighbour or a witness.* WHY IT BEARS ON THIS ROW. This row exists because `carmine-falco`, born 8 June 1930 with no recorded death, is published with his birth date on the strength of a rule that says a direct ancestor is not living. Spring Hill is inner Brisbane, this family was in Brisbane, and 1963 falls inside his life. *If that article reports his death, this row is answered and the archive gains a date it has never had.* NOTHING IS CLAIMED. A name in an index is not a death, the phrase may belong to another man, and this archive has spent the month refusing exactly this kind of inference. Two things would settle it, in order of cost: *David, who is this man's grand-son-in-law, may simply know*; and the Queensland death index for 1963, which this archive already queries and already holds as `data/qld-bdm.tsv`. --- ## ANSWERED, AND HALF OF THIS ROW WAS WRONG WHEN IT WAS WRITTEN — 27 September 2026 David answered in four words: «he died in 2026». *So `carmine-falco` is not living, the classification was right, and his birth date was properly published.* BUT THIS ROW SAID «no recorded death, in the tree or here», AND THAT WAS FALSE. `line.json` has carried «*died: 2026*» since 10 September 2026, with the note «*Family knowledge, given by David de Franceski… No death record has been read for him, and none is claimed*». *I checked `people.json` and the MyHeritage export and did not check the archive's own highest-ranked file.* THE REAL FAULT WAS UNDERNEATH, AND IT WAS STRUCTURAL. Nothing in `build_people.py` read `line.json`'s dates at all. The estate's ranking is line > register > tree, and the line file was the one source that could not reach a person page. So `/direct-line/` said he died in 2026 while `/people/carmine-falco/` said he had no death — the same site, two pages, two answers — and `check:kin` counted him among the birth-only. FIXED. `build_people.py` now fills a blank from `line.json` and prints every disagreement rather than silently choosing. It filled one date — *Carmine Falco d=2026* — and printed two standing disagreements it will not resolve on its own: Pasquale Falco b, line.json «24 Jul 1767» against «1765», and Vincenzo Falco b, «1 Feb 1792» against «1793». DOMENICO PROSTAMO IS STILL OPEN. The other man this row named — born 1930, no death, classed not living — was not asked about and is not answered. *He is not in the direct line, so `line.json` says nothing about him.*

Opened 27 September 2026, after David's sign-in. THIS ARCHIVE HAS PUBLISHED THAT THE PARISH INDEXES CANNOT REACH THE 1760s. That remains true of the indexes. It is not true of the images: FamilySearch's full-text search reads the handwriting on unindexed pages, and it returned PASCHALIS FALCO, baptised 24 July 1767, son of MATTEO FALCO and FRANCESCA CRISCI — *the earliest dated child of the head of this line, and three months from where this archive had him.* THE SWEEP, AND ITS DENOMINATOR. `f.collectionId=M9J1-7MY` (baptisms), `q.anyPlace=Arienzo`, `q.text=Falco`: 361 pages returned, 239 of them actually Arienzo — 66%. *That rate is far above the 24% this archive measured for «Guida» in the same collection, and it must be stated with any count taken from it.* → [[wrong-negatives]] WHAT THE 239 PAGES HOLD, unmined. A first pass for the pattern «*ex ‹forename› Falco et ‹mother›*» found fifteen distinct parent-pairs, including Mattheo Falco × Francisca Crisci on five separate leaves, Paschale Falco × Clara Rivetta in 1794 and 1795, and Josepho Falco × Gesummena Vigliotta — the other Falco family. None of the 239 has been read from the image except the 1767 act. WHY THE REST IS NOT SIMPLY HARVESTED. *The OCR is bad* — «Marcheo» for Mattheo, «Crisei» for Crisci, «Avili» where a mother's name should be — so a name-pattern sweep over it produces candidates and not records, which is the rule this archive already works to. Every one of the fifteen needs its image read before anything is claimed. WHAT WAS ALREADY TESTED. The dates this archive holds for the other founder-child baptisms — 4 Nov 1771, 25 Nov 1774, 2 Feb 1779, 5 Jan 1786 — all match the machine text. *One of five was wrong. That is the expected yield and the reason to keep going, not to distrust the lot.* Sara Falco's baptism is NOT reachable this way — row 5 wants folio 47, and the act names her «*nomen Sara*» with the surname coming from her father, so «Sara Falco» appears only in the index. Film 007959561, 2,246 images; folio 47 has to be walked to.

The cheapest outstanding question in the Australian end, and the item number has been in `notes/findings.md` since early September. NAA Brisbane series J25, file 1971/9830, item 5115541 — «*Falco, Carmine; wife Francesca [nee Zampiello]; and child Giuseppina [known as Josephine]*», 1957–1977. WHAT IT WOULD SETTLE. *Whether Carmine Falco was in Australia before the ss «Sydney» docked on 26 March 1964.* David states he was here in 1963. The file opening in 1957 and the Petrie Terrace address given on 2 September 1963 both agree with him; the passenger list does not contradict him, because a list of one voyage cannot. AND IT BEARS ON A NEWSPAPER REPORT. The Telegraph of 11 July 1963 names «*Mr. Carmine Falco*», a builder, in a fatal fire at Spring Hill — and David gives independently that Carmine was a builder who spent much of his time there. Trove withholds the text under copyright, so the article cannot be read; *the file can.* A sister file is already identified for the other brother: J25 1972/4964, item 5054689, «*Falco, Raffaele … wife Caterina … children Filomena and Antonio*», 1959–1979, marked Open. *NAA RecordSearch is a source this archive already works — see the Australian records method — and neither file has been ordered or read.* --- ## THE ARRIVAL IS FOUND WITHOUT THE PAPER FILE — 28 September 2026 J25 is paper only and cannot be read online, so the passenger index was searched instead. It answers the same question. NAA holds exactly TWO Carmine Falco arrivals: | date | port | ship | from | item | |---|---|---|---|---| | 7 Nov 1956 | Fremantle | CASTEL BIANCO | Genoa | 30139890 | | 26 Mar 1964 | Brisbane | SYDNEY | Naples | 12100311 | The 1964 voyage carries Carmine, Francesca and two further passengers of the surname together — the family. The 1956 voyage carries a Carmine Falco alone, and it is the only pre-1964 arrival of that name in the index. Everything that could not be reconciled now fits: the case file opens 1957, the Petrie Terrace address was given 2 September 1963, and David states he was here in 1963. *The husband came first and alone; the family followed seven years later.* WHAT IS LEFT, AND IT IS SMALL. NAA: K269, item 30139890 is Open and DIGITISED — twenty-nine pages, «*Incoming passenger list to Fremantle «Castel Bianco» arrived 7 November 1956*». Reading his entry would give his age and settle whether this is the man born 8 June 1930. *The images are behind a WAF that refuses a direct request; they render in the browser, so the page must be found by eye among the twenty-nine.* Nothing is claimed until that page is read. *This archive holds a Home Hill Falco cluster and a Mareeba Falco, and has proved repeatedly that the name identifies nobody.* --- ## THE 1956 LIST IS READ, AND IT CANNOT GIVE AN AGE — 28 September 2026 The form has no age column. NAA: K269 item 30139890 is a Commonwealth Department of Health, Quarantine Service «PASSENGER LIST», and its columns are: > No. · NAME · CLASS · PORT OF EMBARKATION · PORT OF INTENDED DEBARKATION · ADDRESS AT DESTINATION IN AUSTRALIA *«REPORT of Passengers on Board the ss. «CASTEL BIANCO» on the 7 NOV 1956 at FREMANTLE from GENOVA. Master GIUSEPPE MORTOLA, Medical Officer GUIDO D'AVANZO, Agents Melbourne Steamship Co. Ltd.»* — each page certified «*I hereby certify that the above is a true and complete list … on arrival at FREMANTLE*». So this document class can never confirm a birth year. *What it can give is the thing that would identify him just as well: his intended port and his address in Australia.* HIS ENTRY IS NOT YET LOCATED. The list runs to about 1,790 passengers over 29 pages, in alphabetical groups by destination, not one alphabet. Pages 5 to 13 are cleared: the Italian run occupies numbers 99–343 and its F section is 205 FAELLA, 206 FAELLA, 207 FARRARO, 208 FIORE, 209 FIORENTINO, 210 FLORIO, 211 FOLINO, 212 FORTUNA, 213 FRANCO, 214 FROCIONE, 215 FURNO, 216 FUSELLA, with a second short run at 217 FABIANO, 218–220 FARRUGIA. *No FALCO*, and FAELLA→FARRARO is exactly where he would sort. Numbers 1–49 and 50–99 are also cleared. What remains is pages 14–29 — later destination groups, one of which is Sydney at numbers ~1750. *It is a page-by-page hunt with no text layer to search.* A TRAP WORTH RECORDING. The rendered page is taller than the browser viewport, so a screenshot shows only the top ~28 of ~49 rows. *Three pages were 'checked' that way before it was noticed.* Scroll and take a second frame, always. --- ## THE HUNT IS OVER BEFORE IT STARTED — 29 September 2026 The index row links to the image, and it names the page. `ViewImage.aspx?B=30139890&``S=21` — his entry is on page 21 of the 29. *Pages 14–29 do not need sweeping; one page does.* The sequence number was in the link behind the row on 28 September and was missed because only the row's text was read. And the page still cannot be read, for a reason that is worth naming. The image is served, with `HTTP 200` and `image/jpeg` at 4316 × 3070 — and the transfer dies at 1,590,848 bytes with `net::ERR_QUIC_PROTOCOL_ERROR`, every time. *That is an HTTP/3 transport failure, not a block: the National Archives is sending the file and the connection is breaking.* Nothing about this is an access restriction, and it should not be recorded as one. *What the page would give is his intended port of debarkation and his address at destination in Australia — the form has no age column, so it can never give a birth year.*