Questions that have been answered

83 questions this project set itself and has since answered. Each one carries what was expected, what was actually found, and what settled it. Some were answered by proving the expectation wrong.

1. ✓ Morgen or hectares? extent_text may be inconsistent LEAD-0007

   Are the figures in parcels.extent_text really morgen, or are some of them hectares mislabelled as morgen?

   What would settle it: Reading the figure and its unit directly off the diagrams for the nine flagged parcels.

   What settled it: Found 29 Jul 2026 by cross-checking the cadastre against extent_text. Per section 10, reported and not silently corrected: the fix is to read the diagrams, not to pick a factor. Until it is settled, treat extent_text as the historic record AS TRANSCRIBED and sg_area_ha as the modern measured area, and do not present either as confirming the other. | CLOSED 1 Aug 2026, AND THE ANSWER IS THAT THE UNIT IS MORGEN. Tested rather than assumed: for every one of the ten parcels whose stated extent still disagrees with its cadastre polygon, reading the figure as HECTARES makes the fit WORSE, not better. Oorlogs Kloof 16,124 against a ratio of 0.42 as morgen and 0.36 as hectares; Bloemendaal 0.52 against 0.45; Witberg 0.60 against 0.52. If any figure had been hectares mislabelled, its ratio would have moved towards 1 and none does. So extent_text is trustworthy as to unit. THE RESIDUAL DISAGREEMENT IS A DIFFERENT QUESTION and it is subdivision: the historic figure is the whole original farm and the polygon matched is one modern portion of it, which is the Modderdrift pattern written up on PAR-000099. One outlier does not fit that: Swanepoel's Aanleg is 2,000 morgen against a cadastre polygon of 11,464 ha, six times LARGER, which reads as a match to a consolidated modern farm and is worth its own look. Recorded on the parcel rather than reopened here.

   Middelwater · Groot Rivier (Varsche Fontein No. 27 Ptn 2) · Plaat Fontein (Varsche Fontein No. 27 Ptn 3) · Kopjes Kraal · Kromrivier (Kopjes Kraal No. 123 Ptn 1) · Brand Kraal (Rondabel No. 124 Ptn 3)

   opened 2026-07-29

2. ✓ Which Rondabel - Tarka (east) or Traka (Koup)? LEAD-0014

   Do the 1818 du Plessis and 1823 de Bruin inventories ('leeningsplaats Rondabel aan de Tarca / in Tarka') describe the Prince Albert Rondabel No. 124 (PAR-000009), or a namesake loan farm in the eastern-Cape Tarka?

   What we expected: Genuinely uncertain. EAST: the 'Tarka' reading; the 600 Rd sheep debt to a Grahamstown butcher (1823); the 1815 Graaff-Reinet baptism. KOUP: Tarca/Traka spelling drift is normal; Zacharias de Beer (the Kweekvallei / Prince Albert family) as a debtor; and Graaff-Reinet was the whole Koup's church before 1820 (PLC-000013), so the baptism does not discriminate.

   What would settle it: Placing Greyling's or van Wyk's veldkornetskap, or a loan-farm register entry locating the leeningsplaats Rondabel.

   What settled it: Opened 30 Jul 2026, when FS LW51-CXR was identified as the 1823 widow. The risk was first flagged in the DOC-0033 transcription analysis as a same-name trap. Cross-bearing with LEAD-0013 (what 'district Albert' means). | 30 Jul 2026, later: Frans, on reading the full TANAP transcription of MOOC8/38.12, is DOUBTFUL of the de Bruin connection to PAR-000009 and is gathering more evidence. Additional east-leaning signals from the transcription: sons-in-law Krugel (Andszoon), Grobbelaar (Nicolaaszoon) and Jordaan (Gerritzoon) - three checkable marriage locations; witness George Aldrich, an English name; creditor wed. Theunis de Bruin. Treat the loan-farm era of the farm page as DISPUTED pending resolution. Note the person identification (PER-000049 = FS LW51-CXR) and the second-cousinship stand independently of this lead - only the farm link is in question. | 30 Jul 2026, MOOC8/44.48 (SRC-128) SHIFTS THIS LEAD MATERIALLY: when a clerk meant the Koup farm he wrote 'the place Rondavel ... field cornetcy of Hermanus van der Westhuisen in the Traka district of Beaufort' - farm, ward, region and district all named. The 1818/1823 inventories name none of these, say 'Tarka', and sold sheep to Grahamstown. The Koup Rondabel also now HAS its own documented loan-farm household (the Venters, OWN-000064) with a very different material profile (707 sheep vs de Bruin's 1,540, no slaves vs three). Working position strengthens: de Bruin/du Plessis PROBABLY an eastern namesake farm - still awaiting Frans's further evidence before OWN-000025 and the 1818/1823 rows are detached from PAR-000009. | ANSWERED 31 Jul 2026, from the day's own imports. The question assumed Traka might be an EASTERN namesake. It is not: Rondabel's own 1833 diagram (DOC-0023) places it in the FIELD CORNETCY OF TRAKA, and so do Vrolikheid, Oorlogs Kloof and Klipfontein on theirs. Traka was the ward these farms sat in, later renamed Kredouw. So the 1818 du Plessis and 1823 de Bruin inventories - 'leeningsplaats Rondabel aan de Tarca / in Tarka' - describe OUR Rondabel. Which means the 1823 widow Aletta Johanna Oosthuizen, second cousin of the root ancestor, was on this ground before the family reached Middelwater. | CORROBORATED, AND THE GEOGRAPHY UNDER THE NAME, 8 Aug 2026. This lead was answered on 31 Jul from the field cornetcy on Rondabel's own 1833 diagram. Three sheets that arrived in the two days since name Traka independently, from a document series that had nothing to do with the original question, and they turn an administrative label into a place on the ground. * DOORN KRAAL No. 116 (DOC-0248, diagram 1154/1863; DOC-0250, diagram 2103/1903) place that farm in the FIELD CORNETCY OF TRAKA, division of Beaufort. A different farm from any in the 31 Jul evidence, forty and eighty years apart, and it widens the ward: Traka held Doorn Kraal as well as Rondabel, Vrolikheid, Oorlogs Kloof and Klipfontein. * LEEUKRAAL No. 24 (DOC-0251, L.G. 8870/47) has the consolidated figure running down the middle of the TRAKA RIVER. The ward is named for a river, and the river is on a surveyed diagram. * TRAKAS KUILEN No. 15 (DOC-0255, SG 1283/1955) is a FARM named for the kuile of that river - the pools. PLC-000043 held the name from 7 Aug, read off the Leeukraal sheet, before this project had a row for the farm at all; PAR-000180 is the farm. So Traka is a river, the pools in it, a farm named for those pools, and the field cornetcy the whole basin was administered as - later renamed Kredouw. Nothing here reopens the lead. It removes the last of the reason to doubt it: an eastern Tarka namesake would have to explain why four separate surveyors, between 1833 and 1955, wrote Traka onto diagrams of ground in this basin. ONE MORE THING THE DOORN KRAAL SHEETS SAY, and it belongs beside LEAD-0013: both give the division as 'Beaufort, now Prince Albert'. That is the same administrative history that decides who counts as part of this project - Beaufort West was the district the Koup was governed from until Prince Albert was founded, which is why the scope rule on completeness.html treats a Beaufort West record as this region and not as somewhere a family came from.

   Rondabel · Doorn Kraal No. 116 · Leeukraal No. 24 (consolidated) · Trakas Kuilen No. 15 · Trakas Kuilen No. 15 Ptn 1

   opened 2026-07-30

3. ✓ match_cadastre.py no longer reproduces the geometry on disk LEAD-0017

   Re-running match_cadastre.py against the current CSVs changes 11 parcels and does not reproduce what is in parcels.csv. Which is right?

   What we expected: Some centroids were hand-entered rather than produced by the script - the Rhemhoogte and Oorlogs Kloof group carry 4-decimal values where the script writes 6, which is the signature of a hand entry.

   What would settle it: Deciding per parcel whether the script's match or the hand entry is correct, then either fixing the script's name-matching or recording the hand values as deliberate.

   What settled it: | CLOSED 1 Aug 2026. match_cadastre.py is idempotent again: a re-run against the current CSVs changes nothing, verified by diffing parcels.csv before and after. The eleven drifting parcels were symptoms of three faults fixed today. (1) The script read <repo>/geo/ while fetch_cadastre.py wrote data/geo/, so it was matching against a stale three-division pull. (2) It derived farm and portion by parsing a CSG_link URL and skipped any feature without one, which silently dropped every national polygon; it reads the 21-digit key now. (3) A hand-set key that named no feature fell back to the farm NUMBER and re-derived a wrong answer, which is what overwrote Oorlogskloof twice; a parcel that has been looked at and has no correct polygon now carries the NO-POLYGON sentinel and is left alone.

   Rietfontein · Faans Kraal (Vrolikheid No. 177 Ptn 4) · Rietfontein (Zeekoegat district) · Swanepoel's Aanleg (Swanepoelsaanleg) · Oorlogs Kloof No. 172 (Roode Els Bosch and Oorlogs Kloof) · Oorlogs Kloof No. 175 (1940 consolidation)

   opened 2026-07-30

4. ✓ Which congregation baptised Jan de Bruijn in 1769? LEAD-0030

   The register page (SRC-168) gives the entry in full but not the church. If it is Roodezand/Tulbagh, the de Bruin family is a Tulbagh family and the Vrolykheid name-transfer of LEAD-0010 has a documented carrier. If it is Drakenstein or Swellendam, it does not.

   What we expected: Genuinely open, and the surnames cut both ways. The opening is dominated by van der Merwe, du Plessis, du Preez, Viljoen, Joubert and Cronje, which is Drakenstein/Paarl Huguenot stock; but it also carries Venter, van Jaarsveld, Prinsloo, van Rensburg, Potgieter, Steijn and Liebenberg, who are trekboer families baptising at whatever congregation they could reach - and in 1769 the choice inland was Roodezand (Tulbagh) or Swellendam, since Graaff-Reinet was not founded until 1786. Note against the Tulbagh hypothesis: Jan Louis Venter (PER-000120) is recorded as born in the DRAKENSTEIN.

   What would settle it: The waypoint. This is a thirty-second lookup, not a research task.

   What settled it: Also on the same opening and worth chasing separately: entry 16, Jan Adriaan Venter x Cornelia Smit baptising a daughter Cornelia on 2 April 1769, witnesses Albert van Jaarsvelt and Hester Venter; and entry 25, Anna Martha Venter with Pieter Venter as witness. Whichever congregation this is, the de Bruijns and the Venters were in it together in 1769. | ANSWERED 31 Jul 2026, same day it was opened. The register is the TULBAGH ('t Land van Waveren) doopboek, begun 1743 - Jan de Bruin was baptised at Tulbagh, 16 April 1769. SRC-169.

   Jan de Bruin · Hendrik de Bruijn · Anna Loots · Rondabel · Vrolikheid · Vrolykheid No. 61 (Tulbagh)

   opened 2026-07-31

5. ✓ Two Oosthuizen-Venter marriages in eighteen months, on the Olifants River LEAD-0034

   Ockert Almero married Adriana Cecilia Venter in Oct 1807; by April 1809 his sister Cornelia Petronella was married to Petrus Jacobus Venter. Were the two Venters siblings, and is this a double alliance between two families about to move east together?

   What we expected: A double alliance. Two siblings of one family marrying two of another inside two years is the classic Cape pattern, and it is the strongest form of the 'relinking' this project measured on 30 July - which found only two relinked marriages in the whole model because so few people have recorded ancestry. If Petrus Jacobus Venter turns out to be a brother of Adriana Cecilia, this is a third, and the earliest. Adriana Cecilia's father is Jan Louis Venter (PER-000120), b.1754 in the Drakenstein.

   What would settle it: Petrus Jacobus Venter's parents. If they are Jan Louis Venter and Johanna Jacomina Smit, it is a double alliance and the case is made.

   What settled it: Also in the same inventory: a 'Pieter Corn:s Oosthuizen Piet:r z:' owing 16:2 - a cousin line not otherwise in the model. | [17 Sep 2026, lead review] RESOLVED AND CLOSED: A DOUBLE ALLIANCE. De Villiers (SRC-656, Deel 3, VENTER) lists among the children of JAN LOUIS VENTER (PER-000120) and JOHANNA JACOMINA SMIT (PER-000121): '4 Petrus Jacobus, gedoopt 19 Maart 1786, gehuwd in October 1807 met Cornelia Petronella Oosthuijzen' and '9 Adriana Cecilia, gedoopt 25 November 1792, gehuwd met Ockert Almaro Oosthuijzen'; and the OOSTHUIJZEN family (Deel 2) marries Cornelia Petronella to Petrus Jacobus Venter and Ockert Almaro to Adriana Cecilia 'in October 1807'. So a brother and sister married a brother and sister, apparently in the same month, eighteen months before the 1809 inventory names Cornelia Petronella as already married. The tree already had PER-000589's parents as PER-000120 x PER-000121; that parentage row is now supported by De Villiers as well. A register entry would still be the primary proof [TO VERIFY at Swellendam, October 1807]. S046 tells the two weddings.

   Ockert Almero Oosthuizen · Adriana Cecilia Venter · Cornelia Petronella Oosthuizen · Petrus Jacobus Venter · Jan Louis Venter · Ockert Jacobus Oosthuizen (b.1755) · De Modder Fonteyn (aan d'Olifants rivier)

   opened 2026-07-31

6. ✓ Going live: the decisions that have to be made before anything is published LEAD-0050

   What must be settled before this site goes on a public domain?

   What we expected: Four are decisions, not work: (1) the FAMILY EXEMPTION - 16 living people including two minors - must come off for a public tier; (2) FAMILYSEARCH-DERIVED SCANS (24 sources) probably cannot be republished under their terms, so publish citation + transcription + ark link and NOT the image; (3) LEAD-0046, the 20 raster scans that never went through the document date gate; (4) the Afrikaans-first question, which is a translation job across 17 stories, not a switch.

   What would settle it: on each, then a build with PUBLIC_RELEASE=True that passes the leak tests.

   What settled it: A SHARED PASSWORD IS NOT A PRIVACY CONTROL AND MUST NOT BE TREATED AS ONE. It keeps material out of search engines and away from casual visitors. It does not survive being forwarded, and it is not a lawful basis for publishing personal information about living people. So the tiering is: PUBLIC = the history; GATED = the working apparatus (notes, leads, open questions) about people who are DEAD; LOCAL ONLY = living people. The password moves the second tier, never the third. | CLOSED 1 Aug 2026, the day the site went live. What was decided: the site ships as TWO builds from one tree, `build_site.py --public` for the public one and `build_site.py` for the full one, with the difference being exactly the 13 living family members. Frans chose the hybrid and accepted the risk in those words. The public build is at mense-van-die-koup.vercel.app. The private build is NOT deployed and the runbook says why: Vercel password protection has to be switched on BEFORE any real content goes up, or the full site including living people sits on a public URL for the length of that gap. This lead's own warning stands and is repeated in DEPLOY.md, that a shared password is not a privacy control. The remaining launch work is the Afrikaans translation, which is tracked as work rather than as a decision.

   opened 2026-08-01

7. ✓ Where is Kriedouw, and were the Oosthuizens there in July 1821? LEAD-0052

   The Beaufort West register puts the direct-ancestor couple's son born 'aan KRIEDOUW' on 26 July 1821. Most entries on the page say the same. Where is it, and how does it sit against the belief that the family established at Middelwater about 1821?

   What we expected: That Kriedouw is a ward or a district the minister was covering rather than the family's own farm - it recurs on most entries on the page, which is not how a farm name behaves. If so it dates their REGION in July 1821 and not their address. EITHER WAY IT IS EVIDENCE OF A KIND THIS PROJECT RARELY GETS. Unlike the tree's place strings, which 5 treats as hostile because they were auto-geocoded, this is a PRIMARY record written by a clerk who was there. It still needs resolving to a gazetteer before it earns a map pin - fail closed - but it is a claim from the right kind of source.

   What would settle it: A gazetteer or a map fixing Kriedouw, or an opgaafrol entry putting the couple on a named farm in 1820-22.

   What settled it: Priority 1 because EVT-000002 dates the Oosthuizen establishment at Middelwater to about 1821 with a tilde, and this is a dated primary record of the same family in the same year at a named place. It either corroborates the arrival, refines it, or moves it. | [1 Aug 2026, re-read of DOC-0068 and DOC-0072] *** CLOSED, AND ANSWERED EXACTLY AS PREDICTED. *** This lead guessed that Kriedouw was 'a ward or a district the minister was covering rather than the family's own farm'. It is. Both 1934 diagrams of Vrolykheid's portions carry the printed words 'situate in the Field Cornetcy of', with TRAKA STRUCK OUT AND KREDOUW WRITTEN IN BY HAND. KREDOUW IS A FIELD CORNETCY - a veldkornetskap - in the Prince Albert division. SO WHAT THE 1821 BAPTISM ACTUALLY SAYS is that the direct-ancestor couple were IN THE KREDOUW WARD in July 1821, not on a farm called Kriedouw. It dates their REGION, which is still worth having: it is a primary record of where they were in the very year the family is thought to have established at Middelwater, and Vrolykheid - which is in that ward - adjoins Middelwater. It does NOT give an address and it earns no map pin. Worth keeping in view: the ward NAME changed. The surveyor struck out Traka twice in 1934, so Traka and Kredouw were both current at different times for overlapping ground, which is exactly the sort of thing that makes nineteenth-century place references slippery.

   Cornelis Andries Oosthuizen · Johanna Machteld Petronella Strydom

   opened 2026-08-01

8. ✓ Where is Drooge Kloof, the house Frans's father grew up in? LEAD-0064

   Drooge Kloof is a werf name in family use - Frans's father grew up in the house - and it is NOT a farm name in the Prince Albert cadastre. It appears twice in the paper, on two DIFFERENT pieces of ground: as the Government ground annexed into Middelwater No. 169 by the 1833 diagram, and as a struck-through 'Droege Kloof' against Portion 3 of Bloemendal Annex No. 164 on its 1913 verso. Which one is the family's?

   What we expected: The Middelwater one, on the balance of what else is known: this family's ground is Middelwater, and A.J.H. Oosthuizen gave 'Droogekloof, Prince Albert' as his residence in 1922, four years before the Bloemendal Annex portion was transferred away in 1926. But that is an argument, not evidence, and the two are 25 km apart.

   What would settle it: Frans saying which farm the house is on, or a coordinate.

   What settled it: The general point is worth more than this one farm: THE CADASTRE RECORDS TITLE, NOT WHAT PEOPLE CALL THE GROUND THEY LIVE ON. Vischat on Rondabel is the same class (LEAD-0018), and there will be more. Such names need a location from family knowledge and should never be matched to a cadastral polygon by name. CLOSED 2 Aug 2026, SAME DAY IT WAS OPENED, and it should never have been opened. Frans had already pinned the house on 31 July: PLC-000035 'Droge Kloof (werf)' sits inside Middelwater, verified by point-in-polygon. The model also already carried the sharper version of this question as LEAD-0023, which asks whether MIDDELWATER PORTION 3 was known as Droge Kloof. This lead asked a question the record had already answered, opened without reading places.csv or the open leads first. The STATE move in LOOP.md exists precisely to stop that. Kept rather than deleted so the duplicate is on the record. [7 Sep 2026] DROOGEKLOOF IS NOW IN PRINT AS A BIRTHPLACE, WHICH THIS LEAD NEVER HAD. SRC-172, Helena Marincowitz, *Prince Albert and the Anglo-Boer War (1899-1902)*, Fransie Pienaar Museum, 1999 p.38: 'Almero was born at Droogekloof in the Prince Albert district.' That is PER-000168, born 28 June 1912, the brother of Frans's grandfather. The house Frans's father grew up in was already his father's uncle's birthplace forty years earlier. IT CHANGES NOTHING ABOUT THE RULE AND EVERYTHING ABOUT THE CONFIDENCE. A werf name with no cadastral polygon has now been used, without comment, as an address by a published local history and by the family itself across three generations. This is what 5's second rule means when it says the absence of a name from P_FARMNAME is evidence of nothing.

   Almero Jacobus Hitge Oosthuizen · Drooge Kloof · Middelwater · Bloemendal Annex No. 164

   opened 2026-08-02

9. ✓ The opgaafrolle of Drakenstein and Tulbagh: were these families neighbours for a century before the Koup? LEAD-0077

   Do the annual tax and census rolls put the Oosthuizen, Venter, Theron, Joubert, Strydom, de Beer and van Zyl households in the same districts in the same years, before any of them reached the Karoo?

   What we expected: Yes, and the tree already shows the shape. Counting EVERY direct ancestor of the anchor who has a birth year and a Cape birthplace gives a clean three-stage migration: CAPE TOWN in the 1650s, DRAKENSTEIN AND PAARL from 1675 to 1750 - 48 of 80 ancestors born 1700-1724 were born there - then TULBAGH, the Land van Waveren, from 1750 to 1825, then GRAAFF-REINET and the Koup. THREE OVERLAPS FALL OUT OF IT. At Drakenstein c.1690-1760 the Oosthuizen stamvader took land in 1691 on the same settlement terms offered to the Huguenots arriving in those very years - and the Jouberts (Pierre b. La Motte-du-Caire) and Therons (Jacques b. Nimes) ARE those Huguenots. At Tulbagh c.1750-1825 the same cohort moves together; this project had already half-noticed it in the note on Matthys Andreas de Beer, which calls the de Beer presence at Tulbagh in the same decades as the Oosthuizens "the evidence for the friends-before-the-move question" - the aggregate says that was not two families but the whole group. At Graaff-Reinet c.1780-1820 Adriana Cecilia Venter (b. 1792) and Johannes Adriaan Oosthuizen (b. 1816) are both born, one generation before the Koup. So Ockert Almero and Adriana Cecilia are not two strangers meeting on a loan farm. They are the fourth generation of two families who had been neighbours at Drakenstein, then Tulbagh, then Graaff-Reinet.

   What would settle it: The same households on the same roll in the same year. That converts an aggregate pattern into named neighbours.

   What settled it: Opened: trace the key families from arrival in ZA to the Koup, to see whether they knew each other. Full working in REVIEW-arrival-to-the-koup.md. | [4 Aug 2026] THE TULBAGH STAGE IS NO LONGER ONLY AN INFERENCE FROM PLACE STRINGS. Boeseken and Cairns put PIETER JANSZ SWANEPOEL on a loan place in the Land van Waveren from 1709 and grant him Nieuwmunster in freehold in 1716, with his own census return, his one slave, his corn and his stolen sheep. That is a named ancestor, on a named farm, in the right valley, at the right date - documentary support for the second overlap rather than a statistical shape. The book also names Willem Wouter Venter paying other farmers' rent across a wide area, and a Marthinus Oosthuijsen in 1738, though neither is yet identified with a person here. | [22 Sep 2026] ANSWERED FROM THE ROLLS THIS PROJECT ALREADY HOLDS (data/opgaaf_rows.csv), and the district is SWELLENDAM, not Drakenstein or Tulbagh. On the Swellendam rolls of 1751 (SRC-488), 1760 (SRC-489) and 1770 (SRC-490) the Oosthuizen, Strydom and Joubert households stand on the same pages: in 1760 folio 6 carries Matthijs Strijdom (line 13), Jacobus Oosthuijsen x Dina Brits (15, PER-000230), Jan Oosthuijsen de oude (16), Jan d'jonge (17), Willem Maarten (18) and the widow of Johannes Strijdom (23); in 1770 folio 2 has Gideon Joubert Jansz on line 8 and Jan Oosthuijsen de jonge on line 9. A Matthys de Beer x Maria van Staden (1751, f.6) and Therons (Jaques 1751; the widow of Jacobus Theron 1760) are on the same rolls. The rolls list households by field cornetcy, so the same page means the same ward, which is the strongest 'neighbours' the rolls can show without a farm column [TO VERIFY that page order follows the ground]. NOT ON THESE ROLLS: Venter and van Zyl (a first pass matched 'van Deventer' as Venter; it is not). The Drakenstein roll of 1751 (SRC-487) has Jouberts, Strydoms and a van Zyl but no Oosthuizen. So the answer is yes for Oosthuizen, Strydom and Joubert, three generations before the 1808 Oosthuizen-Strydom marriage (UNI-000031), and the Venter join belongs to the Graaff-Reinet side (LEAD-0078).

   opened 2026-08-04

10. ✓ Which DISTRICT was the Koup in before 1818? The opgaafrol recommendation depends on it and it is not established LEAD-0079

    Before the Beaufort district existed, was the Gouph administered from Graaff-Reinet or from Stellenbosch (later Tulbagh)?

    What we expected: THIS IS FRANS'S CHALLENGE AND IT IS A GOOD ONE - "why would prins albert people be in graaff reinet opgaafrol. that correct?" The honest answer is that it is NOT established, and my recommendation of a Graaff-Reinet 1800 roll rested on a reason I stated imprecisely. FIRST, THE ANACHRONISM. There were no "Prince Albert people" in 1800. The village was proclaimed in 1845 and named in 1846; the congregation dates from 1842. PLC-000015 already says so: a record placing anyone at Prince Albert before 1842 is a later normalisation, not an address. SECOND, WHAT OUR OWN DOCUMENTS SAY. Every dated district statement on Koup ground from 1829 onward reads BEAUFORT - the 1829 Venter estate at Rondabel (field cornetcy of Hermanus van der Westhuisen, Traka), and the whole 1833-34 Stretch campaign (Traka and De Beer cornetcies). Beaufort district was established in 1818, so those tell us nothing about 1800. THIRD, THE DOCUMENT THAT LOOKED LIKE EVIDENCE AND IS NOT. DOC-0036 is an 1830 inventory taken "at the place Spreeuwfontein, fieldcornetcy Winterveld, district of Graaff Reinet" - which would put Koup ground under Graaff-Reinet. But the row already carries a same-name caution: it is PROBABLY NOT our Spreeufontein No. 26. Using it would be exactly the namesake error this project keeps hitting. It is set aside. FOURTH, WHAT NARROWS IT. The 1808 boundary proclamation reproduced in Appendix II of the Tulbagh book (SRC-259) has the new Tulbagh district running east into the Great Karoo "tot aan de Leeuwen Rivier of de uiterste grenzen der Drostdij Graaf-Reinet". So from 1804 the Karoo was split between Tulbagh and Graaff-Reinet at roughly that line, and before 1804 Tulbagh's share was STELLENBOSCH. The 1800 candidates are therefore STELLENBOSCH or GRAAFF-REINET, and which one depends on where the Koup sits relative to the Leeuwen River. Not guessed at here.

    What would settle it: Either the 1818 Beaufort proclamation, or simply finding the family: search BOTH the Stellenbosch and Graaff-Reinet 1800 rolls rather than assuming one.

    What settled it: Opened 4 Aug 2026 on Frans's challenge to LEAD-0078. | [4 Aug 2026] PARTLY OVERTAKEN. This lead asks which district the KOUP was in before 1818. A separate question turns out to be answerable now and matters more for the opgaafrol: WHERE THE FAMILIES THEMSELVES WERE. The Oosthuizens were on the Olifants River at Citrusdal until at least 1809 (MOOC8/28.63, INV-0005) - see the correction on LEAD-0078. So for a 1800 roll the district that matters for the Oosthuizen line is STELLENBOSCH, not Graaff-Reinet, and not because of where the Koup was administered from but because the family had not moved yet. The Koup district question remains open and still needs the 1818 Beaufort proclamation. | [9 Aug 2026] LEAD-0078 IS BLOCKED ON THIS ONE. That lead wants the Graaff-Reinet c.1800 opgaafrol ordered, and which roll to order is exactly what this question decides. Answer this first; ordering the wrong district's roll costs an archive request and a wait. | [SRC-606, 2 Sep 2026] Opgaafrollen 1787-1847: FamilySearch catalogues 221850 and 17063. Unindexed on NAAIRS. | [22 Sep 2026] ANSWERED from the Records of the Cape Colony (SRC-662). On 15 July 1804 General Janssens named the district between the provisional line, the northern boundary and the Gamka river TULBAGH (vol. V p. 114), and a paper placing a farm 'in de Coup onder 't district van Tulbagh' (MOOC8/45.72, SRC-655) fits that window. The Beaufort subdrostdy of 1818 took the Koup, its western line running to where the Dwyka cuts through the Ghoup's Tafelberg (vol. XII pp. 61-62). So for 1804-1818 the Koup's roll is TULBAGH's. Before 1804 the ground west of the Gamka fell to STELLENBOSCH, out of which Tulbagh was made [TO VERIFY by the pre-1804 rolls]; it is not Graaff-Reinet's. For the Oosthuizen line the question matters less than it seemed: until the 1820s the household was on the southern Olifants, in Swellendam and from 1811 George (SRC-497, SRC-564), not in the Koup at all. LEAD-0078 is unblocked.

    opened 2026-08-04

11. ✓ Why 15 April 1841? The project's headline question has never been a lead LEAD-0091

    Was the day a batch signature clearing a quitrent conversion backlog, a settlement scheme, or routine administration?

    What we expected: A BACKLOG CLEARED BY BATCH SIGNATURE, on four grounds already in this model: every parcel was already occupied and the diagrams say so in their own words ("The LOAN FARM Middle Water...", "The Loan Farm Vrolykheid with the annexed Government ground"); the surveys ripened for years first, 1833 to 1841 here and 1825 to 1845 for Bloemendaal; the field cornet Christian Hatting is himself a grantee, which a settlement scheme does not do; and the folios are not ordered by anything but the clerk's stack, with 1839 sheets sitting at 5-25 and 5-27 inside the 1833 run. NOTE ALSO that S007's premise needs correcting: it is not twelve or thirteen farms to as many surnames but at least TWENTY parcels, across volumes 3, 5 and 6 plus a P.A.Q. series, and the surnames repeat - two Oosthuizens, two van Veuren/van Vuurens.

    What would settle it: The register. If many farms across the district share the date and the underlying entries are spread over years, it is a signing session and the question is answered.

    What settled it: | [8 Aug 2026] THE ANSWER IS LEANING HARD TOWARDS 'QUITRENT CONVERSION', and it comes from three Stretch diagrams opened for the first time. Every one of them shows a LOAN FARM ALREADY OCCUPIED, surveyed in the 1830s, and titled on 15 April 1841: * SG 350/1833 (DOC-0189) - 'the loan Farm KLAARE STROOM including the annexed Government ground', 6498 morgen, surveyed August 1833 FOR THE WIDOW DU PRE [PER-001602]. Title P.A.Q.5-8, 15 April 1841. * SG 361/1833 (DOC-0196) - 'the Loan Farm DE RIET KUYL including Government ground annexed', 14 379 morgen, September 1833, FOR HERMANUS STEIN [PER-001603]. Title Bf.W.Q.3-35, 15 April 1841. * SG 522/1839 (DOC-0220) - 'the FARM WATERFALL', 11 583 morgen 300 roods, 1839, later Bokfontein. Title Bf.W.Q.5-25, 15 April 1841. So the day was NOT the creation of farms and NOT a settlement scheme. The ground was already held, already occupied and already measured - two of the three surveys are eight years older than the title. What 15 April 1841 did was convert existing loan places into quitrent, in a batch, on paper. That is administration catching up with occupation. IT ALSO MOVES THE FARMS' OWN START DATES. This project has treated 15 April 1841 as the beginning of these farms; on the evidence it is the beginning of their TITLE. The Widow Du Pre held Klaarstroom in 1833 and a widow implies a husband before her, so the occupation runs back further still and is not dated here. AND THREE DIFFERENT WARDS ARE NAMED across the three sheets - C. Halling's Zwarte Berg, Traka, and M de Beer's - which bears on LEAD-0066. M de Beer is the same field cornet who took Swanepoel's beacon declaration on Vlakfontein in July 1833. STILL OPEN: whether a proclamation, a Cape Town circular or a land commission sits behind the date. The diagrams say what happened, not why it happened on that Thursday. [TO VERIFY] against the quitrent registers and the Government Gazette for early 1841. | [8 Aug 2026, later] A FOURTH FARM ON THE SAME DAY. SG 523/1839 (DOC-0217), 'the Farm RIET FONTEIN', 13 290 morgen, surveyed by Stretch in 1839 in the field cornetcy of M. de Beer, is titled Bf.W Q5-27 of 15 APRIL 1841. Same pattern as the other three: the ground was surveyed two years before the title and the sheet describes an existing farm, not a new one. Two of the four - Waterfall/Bokfontein and Riet Fontein - sit in the SAME ward, M de Beer's, which makes the batch look administrative rather than geographic. | [22 Sep 2026] ANSWERED ON THE MODEL'S OWN DOCUMENTS: A SIGNING SESSION, NOT A SCHEME. Twenty parcels here carry a title of 15 April 1841, on 21 diagrams. ALL 21 are Charles Lennox Stretch's surveys, and their survey dates are spread over six years - 17 in 1833, 2 in 1834, 2 in 1839. Their title references run as one series through the Beaufort West quitrent register volume 5 (folios 1 to 32 among them: Vrolikheid 5-1/5-7/5-13/5-18, Middelwater 5-7, Vrisch Gewagt 5-11, Bokfontein 5-25, Riet Fontein 5-27, Oorlogs Kloof 5-30, Uitnood 5-32) with Klein Sleutlefontein at 6-5 and Rietkuil at 3-35. So the day is when a backlog of surveyed loan farms was converted to quitrent title together - the ground had been measured years before, and the farms were already occupied. [TO VERIFY, and worth having for S007: the Government Gazette of April 1841 and the entry dates in Bf.W.Q. 5, which would show the batch from the register's side.]

    De Claar Stroom (Klaarstroom No. 178) · Rietkuil No. 119 (Kouka) · Bokfontein No. 94 (formerly Waterfall) · Rietfontein No. 12

    opened 2026-08-05

12. ✓ Is the stamvader in this model twice - PER-000296 and PER-000559? LEAD-0113

    PER-000296 Johannes "Jan" Oosthuijsz (stamvader) and PER-000559 Johannes Oosthuizen share an identical death date, share a wife, and one of them is the recorded father of the other's son. Are they one man, and if so which FamilySearch profile survives - LCJN-657 or GW32-G7M?

    What we expected: One man. Identical death ../1730-11-04, 1672 is one of the two birth years PER-000296 already carries, both married to PER-000297, and PER-000559 was created on 31 Jul 2026 by the kin_sweep one-degree pull - the pull that matched on tree id alone and duplicated people carrying no tree id. The stamvader carries no tree id. LCJN-657 should survive: it is the id FamilySearch itself puts on Jacobus's father in the family view, and the id this project reached independently.

    What would settle it: Either FamilySearch merges them, or one profile turns out to hold a source the other does not and they are two men.

    What settled it: UNI-000220 (PER-000559 + PER-000297) is the duplicate of UNI-000038 and should go with the merge. merge_person.py re-points references, so run it rather than editing by hand. | CLOSED 6 Aug 2026, SETTLED UPSTREAM AND BEFORE THIS PROJECT ASKED. Frans sent the FamilySearch page for GW32-G7M: 'This person was deleted by merge. Surviving person: Johannes Oosthuijsz. ID numbers: GW32-G7M and LCJN-657.' Merged there on 19 October 2024, by another researcher, more than a year before this model duplicated him. The surviving id is LCJN-657 - the id PER-000296 already carried and reached on its own evidence. PER-000559 merged into PER-000296. THREE THINGS CAME ACROSS WITH IT AND ONLY ONE WAS WANTED: the tree id, which the stamvader lacked and now has; the retired FamilySearch id, kept as a door; and the GRIJPSKERKE PARENTAGE, which this project has weighed and REFUSED - that row is marked confidence=excluded, not deleted. UNI-000220 became an exact duplicate of UNI-000038 and was retired in known_issues.txt. Two checks added the same day: one couple recorded as two unions, and a merge that retired a FamilySearch id the survivor does not carry.

    Johannes "Jan" Oosthuijsz (stamvader) · Jacobus Oosthuizen (b.1722) · Johanna Martens · Johannes Oosthuizen (1648-1682) [REJECTED]

    opened 2026-08-06

13. ✓ Import the nine Cape descent chains - they are Frans's own ancestry, generation 9 to 13 LEAD-0116

    Eight of S021's ten have no recorded child in this model, so the site names a descent it cannot walk. The private tree holds every chain. Importing them means about fifty new person rows across the de Wet, Mocke, Duminy, Snyman, van Dyk, Janse van Vuuren, Smuts and du Preez lines. Is that in scope?

    What we expected: IN SCOPE, and the earlier doubt was based on a misreading. 6 takes 'Frans's direct ancestors to generation 13 together with their siblings and spouses', and these ten sit at generations 9 to 13 of exactly that pedigree. The chains run through the maternal and married-in families rather than the Oosthuizen line, which is what made them look outward at first glance. They are not outward. They are the other strands of the same braid.

    What would settle it: Frans deciding. A middle course exists: import only the chains that reach families the site actually writes about, and leave the rest as a stated gap on the page.

    What settled it: The two that DO walk are Catharina van Malabar, five generations to PER-000001, and Louis van Bengale, whose line runs two generations to Clara Herbst and stops. Louis is also the only one of the ten with support from outside the family tree - Boeseken and Cairns on Tulbagh - so his chain is the best candidate if only one is done. | [6 Aug 2026, SUPERSEDING THIS LEAD'S OWN SCOPE DOUBT] Raised to priority 1. The 9a objection does not hold: following a NEIGHBOUR's family outward is out of scope, and this is Frans's own ancestry through his mother's side and the families that married in. About fifty rows, kinship only, no places per 5. The one judgement left is generation 13 as the honest limit, and these chains sit inside it. | [13 Aug 2026] DECIDED: IMPORT ALL NINE. Scheduled as a research pass. | [13 Aug 2026, CLOSED - THE IMPORT WAS ALREADY DONE] Executing the decision found the work complete: the fifty thin rows entered people.csv on 6 Aug (the PER-001449 to PER-001499 block) and every later pass built on them. Verified against the parentage graph this day: NINE of the ten chains walk end to end - eight reach Frans (PER-000438), Catharina van Malabar reaches Ockert Almero and Adriana Cecilia (PER-000001) - and the checks watch *a page claims a descent the parentage table cannot walk* fires zero times. The tenth, Maria van Bengale (b.1638), stops at Marij Zacharias BY DECISION, not by gap: the excluded parentage row on PER-001474 records the tree contradicting itself on that branch (LEAD-0118), and re-importing the same edge would overturn an exclusion. gen_people_tree.py --chains now re-derives all ten from the private tree and emits nothing new, which is the closure test and can be re-run. What remains is not import: familysearch_id is blank on the imported block - the research pass half of the decision - and LEAD-0118 still holds the tenth chain.

    Catharina van Malabar · Angela van Bengale · Elisabeth Rebecca van Bengalen · Louis van Bengale · Souwa van Madagascar · Claesje van Angola

    opened 2026-08-06

14. ✓ Who was Jan Nijs's mother? Twenty-one years say Jacob had two wives LEAD-0118

    Where does the Maria van Bengale (b.1638) descent actually run? The tree gives Marij Zacharias born 1660, married December 1672, with a son born about 1671. Which date is wrong, or is a generation missing between them?

    What we expected: TWO WIVES, and Jan by the first. It is the most economical reading on the table and it needs no date to be wrong. Jacob Jansz de Nijs was born about 1650: he is 21 at Jan's birth about 1671 and 42 at Maria's baptism in January 1692, and TWENTY-ONE YEARS separate those two children. One wife could just span it - she would have to be born about 1651-1655 and be near forty in 1692 - but two marriages is the ordinary shape of a gap like that, and the second wife is the one the 1692 register names, Janna Marij Bousijn. On that reading she is simply not Jan's mother, and Jan's mother is a first wife who is not in any record this project has seen.

    What would settle it: A baptism naming Jan de Nys's mother, or the 1672 marriage entry itself. Failing those, any record naming Jacob Jansz de Nijs's wife before 1692 - if she is not Marij Zacharias, this descent is not this family's.

    What settled it: Note what this does NOT put in doubt: Maria van Bengale's own dates, 1638 to 21 May 1665, are the sharpest of any of S021's ten and match the tree exactly. It is the generation below her that will not add up. | [6 Aug 2026, LATER THE SAME DAY] THE ANSWER IS PROBABLY NOT A DATE AT ALL. The Amsterdam baptism of 9 January 1692 (SRC-298) names the mother of Maria de Nijs as JANNA MARIJ BOUSIJN, wife of Jacob de Nijs - and FamilySearch has that index record ATTACHED TO TWO PROFILES AT ONCE, Johanna Maria Boesijns and Marij Zacharias (LJR9-7H8), the Cape-born daughter of Maria van Bengale. Bousijn is not Zacharias and Amsterdam is not the Cape. TWO WOMEN HAVE BEEN RECORDED AS ONE, upstream, and the splice came into the private tree with everything else. That explains an arithmetic no single wrong date could: a bride of twelve, a son born before the wedding, and a grandmother of twelve, all at once. THE CONSEQUENCE FOR THIS SITE, STATED PLAINLY: if the splice is real, MARIA VAN BENGALE (b.1638) MAY NOT BE AN ANCESTOR OF THIS FAMILY AT ALL, and one of S021's ten falls away. The parentage link is already excluded, so the model asserts nothing - but S021 still lists her, and it should keep listing her with the doubt attached rather than quietly dropping her. Deciding on this evidence alone would be doing the same thing that caused it. | A SECOND WIFE WAS PUT TO THE TREE AND THE TREE HOLDS ONLY ONE. Jacob Jansz De Nijs, born about 1650, has exactly one family in the private tree: Marij Zacharias, married 4 December 1672, one child, Jan De Nys born about 1671. No Janna Marij Bousijn appears anywhere in it, under that spelling or as Boesijns. THE QUESTION IS STILL THE RIGHT ONE, and it splits the problem in two. A second wife would explain the 1692 Amsterdam baptism cleanly - a man of forty-two baptising a daughter by a later wife, with FamilySearch simply attaching the record to the wrong one of the two - and BOTH WIVES BEING CALLED MARIJ is exactly the trap this project already knows about, the same one behind the 'one person married to two people of the same name' check. But it does not touch the arithmetic, because the arithmetic fails at the FIRST marriage, not the second: Marij Zacharias born 1660 is twelve in December 1672 and eleven when Jan is born. SO THE SHARPER VERSION OF THE SAME IDEA IS THE ONE TO TEST: that the 1672 marriage and Jan's birth belong to a first wife who was NOT Marij Zacharias, and that the Cape-born Marij Zacharias has been attached to this family on the strength of a shared given name. The tree's own places, hostile as they are, point that way - Amsterdam, Duesseldorf, Mauritius in 1695, Cape Town by 1699 - which is a Dutch family moving east, with one Cape-born woman spliced into it. | TWO CAPE DOCUMENTS NOW SIT AROUND THE GAP WITHOUT CLOSING IT, which is worth saying precisely rather than letting them look like progress on the question. SRC-299, the estate inventory of 22 May 1713, names Jan Nijs's WIFE (Lijsbet Janse de Rode) and his three CHILDREN. SRC-300, a baptism of 3 April 1701, names him with a wife Elisabeth. Both are about the generation BELOW him and the one BESIDE him. LEAD-0118 is about the generation ABOVE him, and neither document touches it. WHAT THEY DO CHANGE: Jan is now a documented Cape burger, dead by May 1713, with a wife and children in the record - so the question is no longer whether he existed or where he lived, it is only who his mother was. That is a narrower and better question than the one this lead opened with. The record to look for is his own baptism, or a marriage entry for him naming his parents. He was born about 1671, so a Cape baptism would be in the same register series as SRC-300 and roughly thirty years earlier in it. | [6 Aug 2026, FRANS] 'there is 20 years between jan de nys and mary de nys. jan may have had a different mother.' TWENTY-ONE YEARS, and it is the best explanation yet because it is the only one that requires nothing to be wrong. IT ALSO ESCAPES THE VICE THAT DEFEATED EVERY DATE FIX. The arithmetic is caught between two links: Maria van Bengale born 1638 is a comfortable 22 at Marij Zacharias's birth in 1660, but Marij is then 11 at Jan's birth and 12 at the 1672 marriage. Move Marij's birth back to about 1650 to rescue the marriage and Maria van Bengale becomes 12 at HER daughter's birth. No single correction can satisfy both ends. Two wives dissolves it instead of solving it: Marij Zacharias keeps her 1660 birth and her mother, and simply is not Jan's mother. AND THAT IS THE STING FOR THIS SITE. The descent S021 claims runs Maria van Bengale -> Marij Zacharias -> JAN -> Anna Johanna and onward. Jan is the link. If Jan's mother is a first wife who is not Marij Zacharias, the chain breaks at exactly that point and MARIA VAN BENGALE IS NOT AN ANCESTOR OF THIS FAMILY - and she comes off the ten. The cleanest explanation is also the one that costs the most. WHAT WOULD SETTLE IT, in order of how easy it is to get: a marriage entry for Jacob before 1671 naming the bride; Jan's own baptism, about 1671, naming his mother; or a second Amsterdam or Cape marriage for Jacob to Janna Marij Bousijn, which would date the second union and confirm the first was somebody else's. | THE FAMILYSEARCH FAMILY VIEW SETTLES THE SHAPE OF IT, whatever is still open about the names. Jacob Jansz De Nijs (LJR9-74B) and Marij Zacharias (LJR9-7H8) are shown married 4 DECEMBER 1672 AT MAURITIUS with THREE children: JAN 1671, MARY 1691, MARIA 1692. A son a year BEFORE the wedding and two daughters twenty years AFTER him. FamilySearch is carrying the same impossibility this project's own check found, and it is carrying it on the page. The two daughters are also almost certainly ONE PERSON TWICE: Mary De Nys 1691 is 9VZF-514 and Maria de Nijs 1692 is 9VZF-S14 - the same id but for a 5 against an S - and 1692 is the Amsterdam baptism, SRC-298. SO THE READING THAT FITS EVERY DOCUMENT AT ONCE: Jacob's LATER wife bore the child baptised at Amsterdam in January 1692, and the register names her JANNA MARIJ BOUSIJN. Whether 'Marij Zacharias' on FamilySearch IS that Amsterdam woman under a wrong surname, or a separate Cape woman spliced in beside her, the consequence for this site is the same: JAN, BORN 1671, IS BY AN EARLIER WIFE WHO IS IN NONE OF THESE RECORDS. His mother is unknown, and the line from Maria van Bengale runs through him. A SEPARATE AND CLEANER RESULT CAME OUT OF THE SAME SCREEN, and it is worth not losing: Jan's OWN marriage is fully corroborated. Lijsbet Jansz Fockese de Roode, 8 August 1694 at Mauritius, four children, three of whom are exactly the three the 1713 inventory names. Everything one generation below Jan is now solid; everything one generation above him is now openly unknown. | CLOSED 6 Aug 2026, AND NOT BY A DATE. A published Cape genealogy [SRC-301] gives him outright: 'Jan Nys, ook genoem Jan de Rye. v. DUSSELDORF. trou Mauritius (as SOLDAAT) 8.8.1694 met Lysbeth Fockese Roda. In 1696 is hy boswagter op Mauritius en in 1699 burger aan die Kaap. oorlede 1713.' He was a German who came out as a VOC soldier. He has no Cape parents at all, so the question this lead asked - who was his mother - has no answer at the Cape and never did. THE ANSWER WAS SITTING IN A PLACE STRING THE WHOLE TIME. The tree gives his birth as Duesseldorf, and this project discounted it because 5 makes the family-tree's places hostile evidence. Here the PLACE was right and the PARENTAGE was wrong, which is the reverse of the failure that rule was written for. Worth remembering before the next place string is thrown away: a place that does not fit the family may be telling you the family is wrong rather than the place. The line of reasoning that got here belongs to Frans and is worth recording in order: two wives, from a twenty-one year gap between children; then the FamilySearch view showing a son born a year before his parents' wedding; then the biography. Each step made the next one visible. Nothing in this project's own checks could have reached it - they can tell you an arithmetic refuses, not that a man arrived by ship.

    Maria van Bengale (b.1638) · Marij Zacharias · Jan Nys van Dusseldorf

    opened 2026-08-06

15. ✓ The eight other children on Schalk Willem van der Merwe's 1866 death notice have no rows LEAD-0144

    Death notice 2786 names all nine children in order: Jacobus Adriaan, Carel Aaron, Maria Catharina Elizabeth, Schalk Willem Jacobus, Anna Jacoba Louw, Johanna Christina Adriana, Jacobus Nel, Francois Jacobus and Elizabeth Susanna Louw. Only the first is in this model. Should the other eight be here?

    What we expected: THEY QUALIFY AND THEY WERE NOT ADDED, WHICH IS A DECISION FOR FRANS. 6 puts the siblings of a direct ancestor in scope. Against that, R3 says depth is free and breadth is not, and eight rows carrying a name and nothing else is the thin-record shape this project has been moving away from. The transcription holds every name, so nothing is lost by waiting.

    What would settle it: Frans saying whether siblings of a direct ancestor get rows when a document names them all.

    What settled it: Opened 9 Aug 2026. Deliberately not acted on: people are not added in bulk without a decision first. | [11 Aug 2026, SRC-388] THE FAMILYSEARCH SIDE OF THE SAME LIST IS NOW HELD, WITH IDS, and the two sources disagree on the count. Death Notice 2786 names NINE children; FamilySearch gives TWELVE. The three FamilySearch has and the notice does not are Jacobus Adriaan 1814-1868 (LZ8C-1C5), Elsje Johanna Jacoba 1815-1860 (L5RN-X1X) and Petrus Johannes Jacobus 1825- (LZ8C-1NS). One of those is very likely a duplicate of the direct ancestor (FIX-0005); the other two may simply have died before their father without issue, which is why a death notice would leave them out - the form asks who the heirs are, not who was ever born. SO THE NOTICE IS THE BETTER LIST OF HEIRS AND THE WORSE LIST OF CHILDREN, and the ids are what would make the import cheap if it is wanted: eight rows, all with FamilySearch ids and both dates, all deceased long before 1936. Still not acted on. | CLOSED 11 Aug 2026 on a decision to import the siblings, each required to carry a FamilySearch id because the family-tree is old. All eight are now rows, each with a FamilySearch id, together with the other three children of Carel Aaron's first marriage, the two of his second, and two generations of van Heerdens with their siblings - 41 people in all. SRC-391. THE ANSWER TO THE LEAD'S OWN QUESTION, "should the other eight be here?", turned out to be yes for a reason the lead did not anticipate: the point of the import is not the people, it is the IDS. The family-tree this project was built from is a stale copy and its ids cannot be re-checked against the live tree. An FS id can.

    opened 2026-08-09

16. ✓ The Calvinia baptism register names a woman dead ten years as the mother of Frans Jacobus Wilhelmus van der Merwe LEAD-0152

    Janet Melville reads the Calvinia baptism register as giving Francois, son of Schalk Willem Burger van der Merwe and Anna Maria Visagie. Anna Maria Visagie died in 1853 and he was born in 1863. Which is wrong: the register, the reading of it, or her death year?

    What we expected: THE ENTRY IS PROBABLY BEING READ OFF THE WRONG LINE, or the compiler carried the first wife's name forward from an earlier child. Melville herself resolved it that way when Frans put the dates to her - "hy is van die 2de vrou" - and she is the one who read the register. The alternative, that Anna Maria Visagie died later than 1853, is possible and nothing here tests it.

    What would settle it: The 1863 baptism entry, read directly, with the mother's name as written.

    What settled it: Opened 10 Aug 2026 from the Melville correspondence. | CLOSED 11 Aug 2026 by reading the register itself, SRC-378. Entry 62 gives "Kind van: Schalk Willem Burger v.d. Merwe en ELSJE JOHANNA V.D. MERWE". Anna Maria Visagie is not in the entry. The answer was the one this lead predicted - the name was taken off the wrong part of the page - and the mechanism is now visible: TWO VISAGIE WOMEN STAND IN THE WITNESS COLUMN OF THIS SAME ENTRY, paired with two van der Merwe men. The register also reads FRANS JACOBUS WILHELMUS, not Francois, so the second half of SRC-373's central statement does not survive either. Nothing in the model changed as a result, which is the point: it was already right, and it is now right on evidence instead of on a compiler's conclusion.

    opened 2026-08-10

17. ✓ Which of the two Karoo Hoogland farms called Biesies is the family's? LEAD-0156

    Irish 1996 keys collecting locality 384 to 'Biesies' and gives coordinates for other sites in the same survey but not in this part. Where is his numbered locality list, and what coordinate does it give 384?

    What we expected: THE NORTHERN ONE, 21.10474 E / 30.86073 S. It is 18 km from Irish's Daskop, which he places in the Williston district, and his Biesies is two localities later in the same collecting run. It also lies north-east of Williston towards Boesmanland, which is how Melville describes the farm. The southern candidate, 21.1 E / 31.61667 S, is south-east of Williston and away from Bushmanland. This is a convergence of two independent sources and not a reading.

    What would settle it: Frans saying which side of Williston the farm lies, or Irish's numbered locality gazetteer, entry 384.

    What settled it: Opened 10 Aug 2026. Note the paper is CC-licensed through Sabinet and the survey ran 1988-1994. | [10 Aug 2026] The Esri place gazetteer behind the CSG viewer returns four farms and settlements called Biesies in Namakwa District Municipality; two are in Karoo Hoogland and are the candidates. No pin has been written for either. | [11 Aug 2026] A THIRD LINE NOW POINTS AT THE NORTHERN CANDIDATE, and it is Frans's own: "where is biesies - must be close to brandvlei". Measured, that discriminates cleanly. THE NORTHERN CANDIDATE IS 74 km FROM BRANDVLEI; the southern is 141 km. Note it does NOT discriminate on Williston, where the southern is actually the closer of the two at 34 km against 56 km, so Brandvlei is the useful test and Williston is not. THAT MAKES THREE INDEPENDENT CONVERGENCES on 21.10474 E / 30.86073 S: Irish's Daskop 18 km away in the same collecting run, Melville's description of the farm as north-east of Williston towards Bushmanland, and the family's own sense of where it sits relative to Brandvlei. Family knowledge is primary in this record and is not hedged. IT IS STILL NOT A PIN. Three convergences on a coordinate are not a cadastral match, and this project has spent the week watching farm-name matches fail. What would settle it is Irish's numbered locality list, or the farm on the CSG cadastre for the Karoo Hoogland municipality. | [11 Aug 2026, SRC-413] A THIRD CANDIDATE, AND THE FIRST ON PAPER. Frans fetched Dgm. B 459/1870 for the Crown Land farm BISSIES EN ANTEEL KOLK No. 91, division of FRASERBURG, field-cornetcy 3, surveyed 1870. The face carries a BISSIES DAM with a dwelling beside it. The two candidates this lead has been weighing are coordinates off an insect-collecting list and a voting-station schedule; this one is a titled farm with the name on it. IT CANNOT YET BE TESTED against the proximity-to-Brandvlei test that separated the other two, because it has no coordinate: not in the Calvinia cadastre held here, and the CSG national service returns nothing for BISSIES or ANTEEL. Fraserburg town is 180 km from Brandvlei and would argue against it; the old division ran far north of its town and field-cornetcy 3 may be that end. THE CHEAP WAY TO PLACE IT is the diagram's own boundary list: RIET POORT, WATER KLOOF, ZAND PUTS, KWAAI PUTS, DASSIE KLOOF and VAN RENENS PLAATS all adjoin it, and finding any one of them fixes the rest. That is the next step and it does not need another fetch. | [13 Aug 2026] DECIDED [FAMILY]: THE NORTHERN CANDIDATE. 10474 E 30.86073 S that Melville's description and Irish's collecting run already pointed at. Coordinate written to the place row on that basis; Irish's gazetteer entry 384 remains the outside corroboration to fetch. | [14 Aug 2026] THE COORDINATE WRITTEN ON 13 AUG HAS BEEN REMOVED, on Frans's own instruction: "lets not add landmarks we dont know". So this lead is back to what it was - two candidates 85 km apart and no pin - and it matters more now, not less, because Biesies has a parcel row of its own (PAR-000272) sitting in the farm list with no location. Irish's numbered gazetteer, entry 384, is still the outside source that would settle it. | *** [6 Sep 2026] CLOSED. 3"S 21 08'04.7"E (-30.857583, 21.134639). It is 2.9 km from this lead's predicted point and confirms the reasoning built from Irish, Melville and Frans's own 'close to Brandvlei' test. PAR-000272 and PLC-000036 now carry the coordinate and the pin is back on the map.

    Frans Jacobus Wilhelmus van der Merwe

    opened 2026-08-10

18. ✓ Brasse Fontein 371 is pinned 110 km from a farm its own diagram says it adjoins LEAD-0173

    The 1878 diagram gives Brasse Fontein 371's southern boundary as "Riet Kolk of Oude Muur" and places it in the field-cornetcy of Achter Hantam. The Calvinia cadastre's parcel 371 sits 110 km east of Oude Muur 619. Which is wrong: the pin, or the assumption that the boundary means the family's Oude Muur?

    What we expected: THE PIN IS PROBABLY WRONG. Achter Hantam is the Hantam, and the polygon is out in the country east of it. A cadastre matched on a bare farm NUMBER is exactly the failure this project has hit repeatedly - it withdrew a Matjiesfontein pin for the same reason a week ago. IF THE PIN IS WRONG AND THE FARM ADJOINS OUDE MUUR, IT MATTERS A GREAT DEAL: Christina Elizabeth Nel died at Brasfontein in 1895 and her husband was recorded on Oude Muur in 1897, and the two would be neighbouring ground rather than a hundred kilometres apart.

    What would settle it: Any one of the five named neighbours located in the cadastre.

    What settled it: Opened 11 Aug 2026 on reading the diagram, an hour after the pin was added. | CLOSED 11 Aug 2026, WITHIN THE HOUR, AND THE PIN WAS RIGHT. The lead predicted the pin was wrong because Achter Hantam is the Hantam. It is not: the five farms the diagram names as neighbours all sit within 14 km of the pinned polygon and all of them are ~100 km from Oude Muur 619. RIET KOLK No. 385 is the "Riet Kolk of Oude Muur" of the boundary - A SECOND FARM OF THAT NAME. So Brasse Fontein does not adjoin the family's Oude Muur, Christina Elizabeth Nel died 100 km east of where her husband was living two years later, and the project now knows of THREE pieces of ground called Oude Muur. Checking the neighbours cost one query and it was the right check.

    Brasse Fontein No. 371 · Oude Muur No. 619

    opened 2026-08-11

19. ✓ Which Jacobus Johannes Oosthuizen was Vrolykheid surveyed for in 1833, and what does "As. son" mean? LEAD-0213

    Diagram B355/1833 reads "Surveyed for Jacobus Johannes Oosthuizen As. son". Whose son, and which Jacobus Johannes?

    What we expected: PER-000208 IS BACK IN FRONT, and the only thing still standing against him is the patronymic. | SRC-454 puts him and Anna Magdalena Strydom AT ONDER TRAKA on 5 November 1818, with his brother CORNELIS ANDRIES witnessing - and Traka is the field cornetcy the Vrolykheid diagram names. The 'his life never touched this district' objection, which I raised this morning off the 1867 death notice, is withdrawn. | So the three-brothers picture Frans proposed now has documented people on the ground: Ockert Almero on Middelwater, Cornelis Andries on Middelwater from 1818/1821, Pieter Marthinus as the Petrus of De Claar Stroom, and Jacobus Johannes at onder Traka. | WHAT IS STILL UNEXPLAINED IS THE ONE WORD. His father was OCKERT JACOBUS, not an A-name. Either the 1963 typist misread the 1833 hand - 'Os.' for Ockert's son is one glyph from 'As.', and would have been written precisely to tell him from his first cousin JACOBUS JOHANNES OCKERT OOSTHUIZEN (b.1789), son of MARTHINUS - or the applicant is a man we have not found. THE FIRST IS NOW THE SIMPLER EXPLANATION, and it is a hypothesis about a typescript, not a reading of the manuscript. It must not be written on any page until the original is seen.

    What would settle it: THE DEATH NOTICE OF JACOBUS JOHANNES OOSTHUIZEN, who died 29 October 1867 - it names the father and usually the farm, and it should sit in the same FamilySearch collection as SRC-077, 'Cape, Probate Records of the Master of the High Court, 1822-1990'. That is now the cheapest test and it is cheaper than the deeds office. | Then the ORIGINAL 1833 MANUSCRIPT diagram at the SG office, because what this project holds is a 1963 typed copy and the disputed word is the typist's. | Then title deed Bf.W.Q.5-1[?] of 15 April 1841.

    What settled it: The sheet is 1833, not 1933 - it is C. Lennox Stretch's survey of August 1833, titled 15 April 1841. Priority 1: it is the founding document of one of the five written-up farms. | TESTED AGAINST THE 'jo and dcb descendants' EXPORT, 17 Aug 2026, AND IT DOES NOT CONTAIN THE MAN. That file is a MacFamilyTree descendancy of Jacobus Oosthuizen (b.1722, PER-000230) and Dina Carolina Brits - 277 people, six generations. It holds exactly TWO men called Jacobus Johannes Oosthuizen, and both are sons of Ockert (1755): the 1785/1784 one (PER-000208) and Daniel Jacobus Johannes (1807, PER-000150, aged 26 in 1833). NOT ONE Oosthuizen anywhere in the descendancy has a father whose given name begins with A. So the export narrows the field rather than filling it, and it does not touch the reading. | It also carries 'Ockert Almaro' for the 1755 man, which is the same retro-application Frans doubts, arriving from a DIFFERENT compiler - evidence of how widely the assumption travelled, not evidence for the name. | A SECOND READING OF THE ABBREVIATION IS NOW ON THE TABLE and it is not settled from a photograph: 'As.' may be a contraction of the given name (Dutch scribal practice keeps the first and last letters, which fits Andries but not Almero - that would contract to Ao.), or it may be a plain possessive, 'A's son', in which case the sheet gives only the INITIAL and Andries, Adriaan, Abraham and Almero are all still open. 5a forbids picking the plausible one. The sheet itself needs re-reading at full resolution before either reading is written down as what it says. | CANDIDATE PROPOSED AND FAILED ON DATES: Jacobus Johannes L5D6-B22 = PER-000362, born 10 April 1830, three years old at the survey. His father Cornelis Andries GCZ4-2WS = PER-000015 (1792-1860) IS the right family - a brother of the root Ockert Almero, holding MIDDELWATER from 1821~ to 188X, 14 km from Vrolikheid and surveyed by Stretch in the same month of 1833. But he married on 11 April 1813, so no child of his was adult in 1833: the eldest is Ockert Andreas (8 Apr 1814), aged 19, and the name is wrong. | WHAT THIS LEAVES: 'Andries' is now a documented patronymic in exactly the right household, on the right ground, in the right decade - and there is no Jacobus Johannes son of any Andries adult in 1833 in either the 44k tree or the descendancy export. Either a son born 1810-1814 is missing from the tree (FamilySearch omits children silently - 5a-i), or the Andries meant is an older man this project has never seen. | A SIDE FINDING WORTH KEEPING: Cornelis Andries's FIRST SON is OCKERT Andreas, and the vernoemingspatroon names a first son for his father's father. That is independent support for the 1755 man having been plain OCKERT, from a different branch than the 1786 baptism. Hypothesis-strength only, per 5b. | [17 Aug 2026] THE 1809 ESTATE OF THE FATHER IS THE STRONGEST DOCUMENT IN THIS ARGUMENT AND THIS PROJECT ALREADY HELD IT. MOOC8/28.63 (INV-0005, DOC-0065, SRC-092), taken 23 April 1809 at de Modder Fonteyn on the Olifants river, lists ELEVEN HEIRS WITH THEIR AGES: Jacobus Johannes 25, Ockert Almoro 23, ... Pieter Marthinus 12. The seven minors are bracketed 'alle woonachtig bij denzelver broeder Jacobus Joh:s Oosthuijsen' and were placed with him on his own promise to raise and school them. SO THE THREE MEN OF THE THREE FARMS GREW UP IN ONE HOUSEHOLD - the eldest brother's - from 1809. Three adjoining farms taken up by three brothers who were raised together is a far better explanation than three unrelated Oosthuizens. | IT ALSO SETTLES THE FATHER'S NAME INDEPENDENTLY: the deceased is styled 'Ockert Oosthuijzen' throughout, while his son signs 'Ocker Almero Oosthuysen'. The 1755 man was NOT an Almero, from a document rather than an inference. | WHAT IT DOES NOT DO IS NAME THE PATRONYMIC. If these are Ockert's sons the 1833 sheet should read 'Os. son' and not 'As.', and that is a prediction about a 1963 typist's reading of an 1833 hand - testable at the SG office, not from here. Do not write it down as what the sheet says. | NO MOOC8 INVENTORY IS HELD FOR ANY OF THE THREE BROTHERS THEMSELVES. Only the father's. | [17 Aug 2026, SRC-451 - THE DEATH NOTICE ARRIVED, AND IT ARGUES AGAINST THE MAN IT WAS MEANT TO CONFIRM]. Frans found death notice 3999 1/2. It is internally coherent and matches PER-000208 at every point, so there is no confusion of two men INSIDE the record. What it does is account for his whole life somewhere else: born Division of SWELLENDAM about 19 Nov 1784; at MODDER FONTEYN on the Olifants river in April 1809 for his father's estate, with seven minor siblings living in his household; married at UITENHAGE 1808 and at COLESBERG 1847; died 29 Oct 1867 on the farm GROOTFONTEIN, division Colesberg, holding four oxen, two bales of wool and HALF SHARE OF A CROP - a share-farmer's estate, not a landholder's. | AND IT NAMES HIS FATHER: OCKERT JACOBUS OOSTHUIZEN. Not an A-name. So the patronymic still does not fit him, and now his geography does not either. | IT ALSO WEAKENS THE FAMILY TRADITION THAT CARRIED HIM. Fred van Zyl (SRC-119) has his son Ockert Jacobus 'gebore en afkomstig van die plaas Vrolikheid' in 1809 - but the estate of 23 April 1809 puts this household on the OLIFANTS RIVER, four hundred kilometres west. Both can only be true if the family crossed the Colony inside a year. The tradition is not discarded; it is now carrying weight a document contradicts, and it should be tested against the son's baptism entry. | SO THE ANSWER TO FRANS'S QUESTION - a different Jacobus Johannes, or this one moving north? - IS PROBABLY THE FIRST. This man moved north-EAST, to Colesberg, and the Koup is neither. | [17 Aug 2026] THE COUSIN POOL WAS NEVER SEARCHED, AND AN EARLIER NOTE MISSED IT. That note has the 'jo and dcb descendants' export ruling out an A-named father. IT DOES NOT, because it is not the descendancy its filename claims: the full tree gives Jacobus Oosthuizen (b.1722) and Dina Carolina Brits THIRTEEN children, and the export contains ONE of them - Ockert (1755). Twelve branches of first cousins were simply absent from the file. | SEARCHING ALL THIRTEEN CHANGES THE ANSWER TWICE. There IS an uncle whose name begins with A: ADRIAAN OOSTHUIZEN, b. 25 Dec 1774, d. about 1805, seven children - PER-000245 here, with PER-000666 already tombstoned onto it. And his son JACOBUS OOSTHUIZEN, b.1797, d. 25 Aug 1865, was THIRTY-SIX in August 1833. 'Jacobus Oosthuizen, Adriaan's son' is exactly what the DIAGRAM'S BODY TEXT says - 'Surveyed for Jacobus Oosthuizen', no Johannes, no patronymic - with the signature block supplying the patronymic. Adriaan is also the one A-name whose Dutch patronymic form is literally 'Adriaans', which abbreviates to As. | BUT HE FAILS ON PLACE, WHICH IS THE THIRD LEG 9 REQUIRES. Born Overberg, married Susara Maria Elizabeth Olivier there in 1828, children born Overberg and 'Opzoek', died Overberg 1865. Nothing puts him within 300 km of Vrolykheid. Recorded as the best-shaped candidate found, NOT as an identification. | AND ACROSS ALL THIRTEEN BRANCHES THERE IS NO JACOBUS JOHANNES WITH AN A-NAMED FATHER AT ALL. The only grandsons of the 1722 couple carrying both names are the two sons of Ockert and Jacobus Johannes Ockert Oosthuizen (b.1789), whose father was MARTHINUS. So the man may sit outside this whole descendancy, and more exports of it will not find him - the 1841 title deed and the Beaufort West registers will. | [17 Aug 2026, LATE] SEQUENCE OF THE DAY, KEPT BECAUSE THE ORDER MATTERS: the descendancy export appeared to rule out an A-father (it did not - it held one branch of thirteen); the 1867 death notice appeared to rule out PER-000208 on geography (it did not - it only showed where he ENDED); and the 1818 baptism put him in Traka after all. Three documents arrived in four hours and the answer moved twice. Nothing here is settled by argument; it is settled by the 1833 manuscript and the 1841 deed, and both are still unread. | [17 Aug 2026] THE MERGE HAPPENED ANYWAY, and this lead is deliberately NOT closed with it. THE WORD IS STILL NOT EXPLAINED. His father was Ockert; the sheet says As. son; the sheet is a 1963 typescript. Reading the 1833 manuscript at the Surveyor-General or the title deed remains the only thing that settles it, and if either names an A-father this merge is the thing to undo. | *** [17 Aug 2026] THE PATRONYMIC HABIT IS NOW DOCUMENTED IN THIS FAMILY, IN THIS DECADE, ON ONE PAGE. *** The George opgaafrol of 1816 (SRC-459) writes three of its six Oosthuizen households with the father's name attached to the surname: "Jacob Oosthuysen Ock[?]", "Marth: Pet: Jac[s] Oosthuizen" and "Joh: Pet: Jac[s] Oosthuizen". A clerk enrolling six Oosthuizen households in one veldcornetcy could not tell them apart any other way. | THIS DOES NOT READ THE WORD ON THE DIAGRAM and must not be written as though it did. What it does is remove the last thing that made the patronymic on B355/1833 look odd - Stretch wrote "Senior" on every other sheet that month and a patronymic on that one, and the answer is that a patronymic is what these people were called locally. | AND IT LEANS THE READING, WITHOUT SETTLING IT: the eldest brother is entered here as Oosthuysen Ock-something, which is the "Os. son" the 1833 sheet would carry if the 1963 typist read an O as an A. The final letters are illegible in the screenshot. Full resolution of film 008145125 image 45 is now on the list of cheap tests, alongside the 1833 manuscript and the 1841 deed - it cannot settle the diagram, but it can show what form the family used. | *** [16 Sep 2026] THE BEAUFORT ROLL OF EARLY 1824 PUTS VROLYKHEID AGAINST AN 'OCKERT ZOON'. *** In J. Erasmus's Traka ward (SRC-634, film 008145115 image 48, line 1) the first household is 'Jacobus Johannes Oosthuisen, Ockert Zoon' with Johanna Magdalena Strydom, 3 sons and 1 daughter at home, and the facing leaf (image 49) writes 'Vondeling 1/2 / Vrolykheid' in the loan-place column on that same first row. Line 5 is Cornelis Andries Oosthuisen 'Ockert Zoon' (PER-000015), with 'Vondeling 1/2' level with his row. So in 1824 the man holding Vrolykheid as a loan place was a Jacobus Johannes whose father was Ockert, nine years before the survey. THIS STILL DOES NOT READ THE WORD 'As.' ON THE 1963 SHEET, and it does not say which of the Ockert-son Jacobus Johannes rows he is: PER-000208 married ANNA Magdalena Strydom in 1808, and this roll writes JOHANNA Magdalena. What it removes is the need for an A-named father. | RESOLVED 17 Sep 2026 ON WHICH MAN; THE WORD 'As.' IS SPLIT OFF AS LEAD-0354. THE HOLDER OF THE LOAN FARM VROLYKHEID IS JACOBUS JOHANNES OOSTHUIZEN (PER-000208, KFRS-SCK), eldest son of Ockert Oosthuizen (b. 1755) and husband of Anna Magdalena Strydom (PER-000209). Two tax rolls two years apart, read against the page, say so between them and neither needed the other: (1) The Beaufort roll of 1822 (SRC-558, image 454-455, line 16) enrols JACOBUS JOHANNES OOSTHUIZEN EN ANNA MAGDALENA STRYDOM with VROLYKHEID in the leenings-plaats column, inside an unbroken block with his brothers CORNELIS ANDRIES (line 19) and ROELOF (line 22) and the Strydom parents-in-law. Wife's name and pairing are exact to PER-000208 x PER-000209, married Uitenhage 6 March 1808, and De Villiers (SRC-656) marries the Jacobus Johannes of Ockert's family to Anna Magdalena Strijdom. (2) The Beaufort roll of early 1824 (SRC-634, image 48-49, line 1) enrols him as 'Jacobus Johannes Oosthuisen, OCKERT ZOON' with 'Vondeling 1/2 / Vrolykheid' on his row, and the brothers again as 'Ockert Zoon'. Its wife's name 'Johanna Magdalena' is one clerk's slip against the 1822 roll, the 1808 marriage and the 1818 Traka baptism (SRC-454), which all have Anna; the household of 3 sons and 1 daughter is consistent with his children then living. NO SECOND JACOBUS JOHANNES OOSTHUIZEN stands in the Traka ward in 1822 or 1824, and the cousin candidates this lead raised are ruled out on place or age: Jacobus son of Adriaan (b. 1797) lived and died in the Overberg; Jacobus Johannes son of Cornelis Andries (PER-000362) was three in 1833. SO THE ANSWER TO 'WHOSE SON' IS OCKERT'S, from the 1824 roll itself, and the merge of 17 Aug (PER-000043 onto PER-000208, LEAD-0215) stands. ONE STEP IS CONTINUITY, NOT A RECORD: nothing read here covers 1825-1832, so that the 1833 applicant is the 1824 holder rests on the loan farm staying with its holder until surveyed for him - the ordinary course, and nothing points elsewhere. WHAT STAYS OPEN, as LEAD-0354: the typed '1963 copy' reads 'As. son', which does not fit Ockert; the Surveyor-General's esio index holds only that copy (SCAN484/485, checked 17 Sep 2026), so the 1833 manuscript and the 1841 deed are the only places the word can be read.

    Jacobus Johannes Oosthuizen (b.1784) · Ockert Jacobus Oosthuizen (b.1755) · Ockert Almero Oosthuizen · Pieter Marthinus Oosthuijzen · Cornelis Andries Oosthuizen · Jacobus Johannes Oosthuizen

    opened 2026-08-17

20. ✓ Were the three farms of the 1833 survey season taken up by three brothers? LEAD-0215

    Middelwater was surveyed for "Okert Oosthuizen" (PER-000054), Vrolykheid for "Jacobus Johannes Oosthuizen As. son" (PER-000043), and De Claar Stroom between them belonged to "Petrus Oosthuizen" (PER-000505). Are these three men Ockert Almero (PER-000001), Jacobus Johannes (PER-000208) and Pieter Marthinus (PER-000146), three of the eighteen children of Ockert Jacobus Oosthuizen (PER-000118)?

    What we expected: STRENGTHENED 17 Aug 2026 BY AN ABSENCE THAT TURNS OUT TO BE AN ARGUMENT. Frans: the cameo's "Petrus Oosthuizen and his brother, Ockert" describes 1858-1860, and by then Jacobus Johannes had been gone from this district for more than ten years - Hantam in 1847, Colesberg from at least 1846 to his death in 1867. A source about 1858 naming only two brothers is what the three-brothers reading PREDICTS. | PROBABLE, and it is now the frame the whole 1833 season is read through. | THE HOUSEHOLD IS DOCUMENTED: their father died at de Modder Fonteyn on the Olifants river in April 1809 (MOOC8/28.63), and the SEVEN MINORS - Pieter Marthinus among them, aged 12 - were placed with their eldest brother JACOBUS JOHANNES, aged 25, on his own promise to raise and school them. These men grew up in one house. | AND TWO OF THEM ARE DOCUMENTED IN TRAKA BEFORE THE SURVEY: Jacobus Johannes and Anna Magdalena Strydom baptise a daughter "aan onder Traca" on 5 November 1818 (SRC-454), with their brother CORNELIS ANDRIES witnessing - the same Cornelis Andries this model anchors on MIDDELWATER from about 1821. So the family was on this ground fifteen years before Stretch measured it. | A FOURTH BROTHER MAY BELONG IN THE PICTURE: Cornelis Andries holds Middelwater 1821-188X in this model, which sits awkwardly beside PER-000054 holding it 1833-1841 - two rows, one farm, overlapping. That is itself worth resolving.

    What would settle it: A deed or a register naming any one of the three with a patronymic or a father, on the farm. Failing that, an estate file for Pieter Marthinus that names De Claar Stroom.

    What settled it: NOTHING HAS BEEN MERGED. All three placeholders stand as their own rows, and the identifications are recorded on both sides as proposals. 9 wants name AND date AND place, and PER-000505 and PER-000146 between them have no date at all. | [17 Aug 2026] A DATABLE CONSTRAINT FALLS OUT OF THE SAME OBSERVATION AND IT SHOULD BE TESTED. Ockert Almero died 26 February 1859, in the village and not on a farm. The town was surveyed from 1858 and the erven sold on 4 August 1860. So he can be the Ockert of the laying out and cannot be the Ockert of the sale. Either the cameo compresses a period into one sentence - it is tertiary and carries no dates anywhere - or the later Ockert is somebody else. I proposed the nephew OCKERT ANDREAS OOSTHUIJZEN (PER-000355) and WITHDREW IT the same day: he has no tie to any ground here and the tree has him dying at Caledon. Of the ten Ockert Oosthuizens alive in 1860, only one has a land tie in this district at all, and he is a grandson aged twenty-four. Compression in a dateless tertiary source is the likelier explanation. Worth settling because it also tells us whether the family's hold on De Claar Stroom passed to the next generation. [TO VERIFY] | *** [17 Aug 2026] DECIDED AND CLOSED BY FRANS. All three merged. *** PER-000054 into PER-000001 (Ockert Almero, Middelwater), PER-000043 into PER-000208 (Jacobus Johannes, Vrolykheid), PER-000505 into PER-000146 (Pieter Marthinus, De Claar Stroom). The placeholders are tombstoned, not deleted, so the reading can be audited and reversed. | WHAT THE SITE GAINED: the three farms now name men with dates, parents, siblings and a documented household instead of three floruits. WHAT IT GAVE UP: the page no longer distinguishes what the diagram SAYS from who we think it was, and that distinction is now carried by the prose and by the [TO VERIFY] rather than by two rows. | LEAD-0213 STAYS OPEN. The Vrolykheid patronymic is still unexplained and the merge does not explain it. If the 1841 deed or the 1833 manuscript names a different man, this merge is what has to be undone, and the tombstones are how. | *** [28 Aug 2026] THE MIDDLE FARM DOES NOT HOLD, AND TWO DOCUMENTS SAY SO INDEPENDENTLY. *** The 1822 Beaufort roll (SRC-559) puts HERMANUS CAREL DU PREEZ (PER-001911) at line 3 with KLAARE STROOM in the leenings-plaats column, and SG 350/1833 (DOC-0189) reads 'Surveyed for THE WIDOW DU PRE' (PER-001602). So De Claar Stroom was du Preez ground in 1822 and a du Pre widow's in 1833, in the very season Stretch surveyed Middelwater for Okert and Vrolykheid for Jacobus Johannes. The only Oosthuizen hold on it in this model is OWN-000082, PIETER MARTHINUS (PER-000146) in 1858-1860, from the Klaarstroom cameo, which is TERTIARY and dateless elsewhere. *** THE THREE-BROTHERS-ON-THREE-FARMS READING IS THEREFORE TRUE FOR TWO FARMS IN 1833 AND FOR THE THIRD ONLY A GENERATION LATER. *** That is not a small distinction: the strength of the 1833 reading was that the three men held three adjoining farms in one survey season, and the middle farm was somebody else's that season. It does not touch Middelwater or Vrolykheid, and it does not touch the 1858-60 Petersburg story. What it removes is the mutual support the three identifications were lending each other. WHAT WOULD STILL RESCUE IT: a transfer from du Pre to an Oosthuizen between 1833 and 1858; or evidence that the family occupied the ground without holding the loan, which is exactly what the 1818-1821 record shows them doing elsewhere. See LEAD-0324 on the name PIETERSDORP.

    Ockert Almero Oosthuizen · Jacobus Johannes Oosthuizen (b.1784) · Pieter Marthinus Oosthuijzen · Cornelis Andries Oosthuizen · Ockert Jacobus Oosthuizen (b.1755) · Middelwater · Vrolikheid

    opened 2026-08-17

21. ✓ The van der Merwe road needs no spine, and here is what it needs instead LEAD-0245

    S044 tells one line's road - how the Oosthuizens came to the Koup. The van der Merwe line has no equivalent, and the prequel rule now covers Calvinia and the Hantam. What is the van der Merwe road, and where does it start?

    What we expected: A road running Hantam/Calvinia to the Koup, with a Swellendam stage in between: folio 1 of the 1770 roll has JACOBA VAN EERDEN, widow of GERRIT VAN DER MERWE D'OUDE, with four sons and four daughters counted to her and two grown sons - Schalk Willem and Hendrik, both GERRITSZ - entered alone beside her, and folio 2 opens with a third, Pieter Schalk Gerritsz.

    What would settle it: A descent line from one of the 1770 Swellendam van der Merwe households to a van der Merwe who holds ground in the Klaarstroom basin, sourced at every step. Absent that, the piece is about the ground and not about descent.

    What settled it: STORY-VOICE.md binds if this is written. S035 is the benchmark. | *** CLOSED 19 Aug 2026, AND THE PREMISE WAS WRONG. *** This lead asked for a companion to S044. S035 already carries the whole road in one piece, Paarl 1735 to Vredendal about 1950, and S044 exists only because the OOSTHUIZEN road was scattered over five stories. A spine over a single story retells it. | The lead also said folio 1 of the 1770 roll 'gives it a first paragraph'. IT DOES NOT, not yet: those van der Merwes are entered as Gerritsz and this line's Schalk Willem has no father on record, so the connection is open in both directions and asserting it would have been a match on a name. That question is now LEAD-0250 and it is stated as open in S035 v2.3.

    opened 2026-08-19

22. ✓ The eGGSA caption harvest missed four graves on one album, and they may not be the only ones LEAD-0253

    grave_photos holds 11 captions for CEM-522 (Wolver Fontein 137, Ladismith) and the live album carries at least 15 named ones. The four missing are all CRAFFORD. Is this album an exception, or does the harvest under-read albums generally?

    What we expected: Either the four were added to the album after the harvest ran, or fetch_eggsa_captions.py stops early on some page shape. THE SECOND WOULD MATTER A GREAT DEAL: the harvest is 20,949 captions over 473 burial grounds and a systematic shortfall would be invisible, because nothing compares a ground's caption count against the album it came from.

    What would settle it: A re-run that returns 15 for this album proves the album changed; one that returns 11 proves the harvester is at fault and the whole set needs re-taking.

    What settled it: The four: CRAFFORD Anna Margarita nee van Tonder 1890-1949; CRAFFORD Christoffel Frederick 1856-1945 & Elsie Maria v.d. Vyver 1857-1919; CRAFFORD Isak 1877-1938 & Anna van Eck 1882-1953; CRAFFORD Jan Abraham 1884-1956. Recorded in CEM-522's notes because grave_photos is keyed on the photo url and those were not captured. | *** ANSWERED 20 Aug 2026, AND IT WAS THE HARVESTER. *** Not the album. eGGSA renders a grave with ONE photograph as `image-thumb` and a grave with SEVERAL as `album-thumb`, a sub-album holding them; the regex matched only the first. So every grave carrying more than one photograph was invisible across every ground walked. | A FULL RE-WALK OF ALL 597 GROUNDS RECOVERED 1,093 CAPTIONS, 23,591 rows to 24,684 - a 4.6% shortfall, silent since the harvest began. Worst affected included the Prince Albert NG Kerk cemetery (21) and the Calvinia main cemetery (25). 129 grounds gained at least one. | A SECOND FAULT SAT BEHIND THE FIRST AND IS THE MORE IMPORTANT ONE: the 'already harvested; don't re-walk' guard meant a fix to the harvester could never reach a ground already walked, so the blindness was permanent by construction. That is why --rewalk exists now. | THE CHECK THIS LEAD ASKED FOR IS STILL MISSING and is now LEAD-0254: nothing compares a ground's caption count against the live album, because that needs the network, and checks.py runs offline.

    Wolvefontein No. 137 (Ladismith)

    opened 2026-08-19

23. ✓ The line moves EAST across three generations - measuring it, and finding where it turns LEAD-0260

    THE TULBAGH-CITRUSDAL SEQUENCE IS WITHDRAWN AS A NAMED HYPOTHESIS at his instruction - it was a reading of three Boland baptisms, and the 1786 one turned out to be PAARL, the family's ancestral church, not Roodezand (SRC-076). What replaces it is his actual claim, which is testable: does the line move gradually east and north-east?

    What we expected: UNDECIDED, DELIBERATELY. Two readings fit and this project holds nothing that separates them. (A) THEY MOVED, west to east: on or near the BOLAND Olifants around 1802, which Swartland serves on the road to Piketberg and Clanwilliam, and in the Swellendam district by 1803. That would revive the Citrusdal identification the project withdrew on 18 August. (B) THEY TRAVELLED. A twenty-one-month gap between Susanna Johanna's birth and her baptism, and a seven-month gap for Daniel Jacobus Johannes, say this was a household a long way from any church, baptising in batches wherever it went - and a van Tonder godfather at Graaff-Reinet (LEAD-0258) gives one such journey its own reason. WHAT IS NOT IN DOUBT is the 1808 residence: only that entry carries a 'woonende aan' clause, and A.A. Faure took the 1809 estate as Landdrost of SWELLENDAM, which is territorial.

    What would settle it: Any roll, muster or tax list that counts the household in a named field cornetcy in 1801-1803. A baptism says where the church was; a roll says where the home was.

    What settled it: Opened 20 Aug 2026. Supersedes nothing; LEAD-0220 remains the farm question, this is the household question. | *** [same day, AND IT MOVED] SRC-518 puts a fifth baptism at ROODEZAND (TULBAGH) in March 1796 - the congregation this project named in advance, in LEAD-0257, as the one that would point at the Boland Olifants, because it served that valley until Clanwilliam opened in 1826. With Swartland 1802 that is TWO of five in the north-west. Reading (A), that the household moved west to east, is now the better-supported of the two rather than the weaker. It is still not established. THE DECIDING DOCUMENT IS UNCHANGED AND UNREAD: the 1786 baptism of the root ancestor, SRC-076, whose ark is recorded and whose congregation nobody has looked at. If that is Roodezand too, the household was western from at least 1786 to 1802 and the Citrusdal identification comes back on documentary grounds. | *** [same day, AND IT MOVED BACK] SRC-519 is the sixth baptism, at PAARL in October 1799 - one month after SRC-515 has both parents standing IN PERSON as witnesses at Graaff-Reinet. Those two cannot both be journeys. The Paarl godparents are Gert and Adriaen Oosthuijsen and the child is named for them, so he was presented by uncles. *** THE LESSON IS THE LEAD'S CENTRAL PROBLEM: A BAPTISM PLACE IS WHERE THE KIN WERE, NOT WHERE THE HOUSEHOLD WAS. *** Reading (A), a west-to-east move, was called better-supported an hour before this page arrived; that judgement is withdrawn. Six fonts in four congregation-regions over twelve years, with the mother's people standing in the east and at Roodezand and the father's at Paarl, is what a widely-connected trekboer family looks like however it was living. The residence question is back to where SRC-514 and the Swellendam landdrost put it, and the 1786 baptism is still the document that would break the tie. | *** [20 Aug 2026, CLOSED THE SAME DAY IT WAS OPENED - BY A DOCUMENT ALREADY IN THIS REPO.] *** The question was whether the household lived in the north-west, as Roodezand 1796, Paarl 1799 and Swartland 1802 suggested, or in the south-east. A BAPTISM CANNOT ANSWER IT AND A TAX ROLL CAN, because a man is counted in one district and one only. | SRC-497, THE 1805 SWELLENDAM OPGAAFROL J 318, ENTRY 784: 'Oosthuizen / OCKERT - SUSANNA CATHARINA VAN TONDEREN - 1|1|6|4'. The household is counted at SWELLENDAM. His sons are on the same page - 788 Jac: Joh: Ockertsz and 789 Ockt: Almero - and some twenty-two Oosthuizen households run from entry 769 to 790. It has been in the repo since 19 August and nobody put it against the baptisms until today. | TWO INDEPENDENT OFFICIALS AGREE: A.A. Faure was landdrost of SWELLENDAM 1798-1810 and took the 1809 estate at de Modder Fonteyn; A. Stockenstrom, who signs the 1802 estate, was secretary to Faure at Swellendam from March 1796. Jurisdiction is territorial. | *** SO READING (B) IS RIGHT: THEY TRAVELLED. *** The household was a Swellendam household from at least 1796 to 1809, and the six baptisms map its KIN, not its address - which is exactly what SRC-519 forced this lead to concede an hour earlier, when a five-year-old turned out to have been carried to the Paarl font by his uncles while his parents stood in a church at Graaff-Reinet. | AND THE CITRUSDAL READING IS OUT ON POSITIVE EVIDENCE AT LAST. The Boland Olifants lay in the Tulbagh and later Clanwilliam districts and never in Swellendam's. Until today it was excluded by an inference from a signature that the inventory transcript itself called 'not proof'; it is now excluded by a tax roll that counts the man somewhere else. LEAD-0220 keeps the farm question. | *** AND FRANS CLOSED THE BACK DOOR THE SAME HOUR. *** He read the 1804 DRAKENSTEIN roll and found NO OOSTHUIZEN in it (SRC-520). Drakenstein is where a family on the Boland Olifants would have been counted. Add this project's own 1751 Drakenstein roll - 402 households, no Oosthuizen - and the family was never a Drakenstein family at all, in 1751 or in 1804, while the Swellendam rolls carry three of them in 1751, five in 1760, nine in 1770 and twenty-two in 1805. THREE LEGS NOW: present at Swellendam 1805, absent from Drakenstein 1804, and the landdrost who took the 1809 estate was Swellendam's. | *** [20 Aug 2026, REOPENED - FRANS PROPOSES A SEQUENCE THAT FITS BETTER THAN EITHER READING THIS LEAD HELD.] *** His words: 'unlikely they move from tulbagh to de rust, back to swellendam, then to the koup... more likely tulbagh, citrusdal, swellendam, koup.' | *** WHY IT IS BETTER: IT EXPLAINS THE WESTERN BAPTISMS INSTEAD OF EXPLAINING THEM AWAY. *** Three of the six are Boland - ROODEZAND 1796, PAARL 1799, SWARTLAND 1802 - and this project had reduced all three to 'the kin network', which is true but weak. On Frans's sequence they are simply where the household lived. ROODEZAND IS THE POINT: it is the congregation that served the OLIFANTS VALLEY, Citrusdal included, until Clanwilliam opened in 1826. A family at Citrusdal in 1796 baptises at Roodezand. Ours did. | AND IT RESCUES THE FAMILY MEMORY WITHOUT REQUIRING IT TO BE THE 1809 FARM. The Citrusdal tradition would be true of where they CAME FROM rather than where he died - which is how family memory usually survives, attached to the ground a generation remembers rather than to a last address. | AND IT KEEPS EVERYTHING THE 1805 ROLL SAYS. If they were in the Swellendam district by 1804, then the roll's 'aan d' Olyphants Rivier, 43 uuren' is the SOUTHERN Olifants, and the family had lived on one Olifants and moved to another - which is not a coincidence so much as trekboers recognising the same kind of ground. | *** THE ONE OBSTACLE, AND IT IS REAL. *** The 1802 estate of the PARENTS, MOOC8/24.27 (SRC-093), closes 'Mij present: A: Stockenstrom, Secret:s', and Stockenstrom was secretary to Landdrost Faure of SWELLENDAM from March 1796. That puts Jacobus d'oude and Dina Britz under Swellendam in 1802, not the Boland. It does not settle where their son Ockert was - a man's parents' estate is administered where THEY sorted - but it is evidence against the whole family being west in 1802, and the Swartland baptism is the same year. | *** THE TEST IS CLEAN AND NEITHER SIDE CAN DODGE IT: A ROLL OF THE 1790s. *** Ockert is counted in ONE district each year. If he is in a DRAKENSTEIN or TULBAGH roll of the 1790s, Frans is right and this project has had the migration wrong. If he is in a SWELLENDAM roll of the 1790s, he never left that district and the western baptisms are travel after all. SRC-481 indexes seventy Swellendam rolls; the Drakenstein series is at Den Haag in the same run as SRC-487. Frans has already checked 1804 Drakenstein and found no Oosthuizen - on his own sequence they had left by then, so 1804 does not test it. THE 1790s DO. | *** [20 Aug 2026] FRANS TESTS A VARIANT: OCKERT STAYS AT CITRUSDAL, THE ELDEST SON MOVES TO LADISMITH, THE SON'S SONS GO TO THE KOUP - AND HE VISITS. THE SPECIFIC VERSION DOES NOT FIT, BUT IT BROKE OPEN THE ARGUMENT THIS LEAD WAS LEANING ON. *** | WHY THE SPECIFIC VERSION FAILS, on this project's own rows. (1) THE ELDEST SON WENT EAST, NOT TO LADISMITH. PER-000208's documented life is already written down as Swellendam 1784 - married Uitenhage 1808 - Modder Fonteyn April 1809 - married again Colesberg 1847 - died 1867, and his own note says in as many words that 'it never passes through the Koup'. Nothing puts him in Ladismith. (2) THE KOUP LINE IS THE SECOND SON, not the eldest: Ockert Almero (PER-000001), through Graaff-Reinet 1817 and Beaufort West 1821 to Middelwater. (3) A TAX ROLL COUNTS RESIDENCE, NOT VISITS, and the 43-uuren Olifants farm is recorded against Ockert's own household as land 'welke HIJ bezit'. | *** BUT HERE IS THE CRACK, AND IT IS MINE. *** I have been arguing that a SWELLENDAM roll counting him proves the farm was in the SWELLENDAM district, so Citrusdal is out. THE ROLL'S OWN COLUMN HEADER SAYS OTHERWISE. SRC-521 gives the 1808 wording: the distance is from 'de drostdij of Hoofdplaats WAAR ONDER 'T LAND GELEGEN' - the drostdy UNDER WHICH THE LAND LIES. And SRC-459 describes the same block on the 1816 George roll as giving the farms a man holds, how far they lie from the drostdy, AND UNDER WHICH LANDDROSTDIJ THEY SORT. *** A COLUMN THAT ASKS WHICH LANDDROSTDIJ THE LAND SORTS UNDER EXISTS PRECISELY BECAUSE A MAN ENROLLED IN ONE DISTRICT COULD HOLD LAND IN ANOTHER. *** The argument here assumed the case the form was designed to record. | AND THE NUMBER ITSELF NOW LOOKS LIKE AN OUTLIER RATHER THAN A LONG DRIVE. Everything else on that leaf is 14, 15, 17, 18 or 30 uuren. MODDERFONTEIN AT 43 IS THIRTEEN CLEAR OF THE NEXT HIGHEST. A farm sitting outside the district would look exactly like that. | *** SO THE DECIDING THING IS A COLUMN NOBODY HAS READ. *** Frans has read the river and the hours off the far-right block. If it carries a third element naming the LANDDROSTDIJ - and the 1816 roll's does - that column says outright whether Modderfonteyn sorted under SWELLENDAM or under TULBAGH. IF IT SAYS TULBAGH, THE CITRUSDAL READING IS RIGHT AND THIS PROJECT HAS HAD IT WRONG SINCE 18 AUGUST. That is the ask, and it is one glance at the same opening. | *** [20 Aug 2026] THE FOSTER PROMISE MAKES THE FAMILY ONE TRACKABLE UNIT, AND IN 1816 THAT UNIT IS IN THE GEORGE DISTRICT. *** [FAMILY] Remember, the eldest son took the youngest 7 into foster. That is the strongest argument yet in this whole question and it had been sitting in two documents this project already holds. | THE 1809 INVENTORY (SRC-092) brackets heirs 5 to 11 with 'ALLE WOONACHTIG BIJ DENZELVER BROEDER JACOBUS JOH:S OOSTHUIJSEN' and records his promise to keep and educate the seven minors until their majority. He also bought de Modder Fonteyn out of the estate. So from April 1809 the seven youngest are ONE HOUSEHOLD, on that farm, under a legal undertaking. | THE 1816 GEORGE ROLL (SRC-459) THEN FINDS THAT HOUSEHOLD, AND THREE OF THE SEVEN AS NEW HOUSEHOLDERS BESIDE IT, IN A SINGLE VELDCORNETCY: (1) 'Jacob Oosthuysen Ock[?]' x 'Anna Magd: Strydom', 1|1|4|2 - PER-000208 and PER-000209, the couple who married 6 March 1808; (2) 'Piet: Marth: Oosthuyzen' - heir 7, twelve in 1809, nineteen and on his own here; (3) 'Corn: And[s] Oosthuyzen' x 'Joh: Petron: Strydom' - heir 5; and further down (4) 'Gerrit Adr: Oosthuyzen' x 'Anna Cath: Schoeman' - heir 6. A sister married to a Venter is on the same page. | *** THAT IS THE ARGUMENT. *** George was cut out of Swellendam on 23 April 1811. IF DE MODDER FONTEYN LAY ON THE SOUTHERN OLIFANTS, THIS HOUSEHOLD NEVER MOVED - the boundary moved around it, which is exactly what SRC-486 predicted from Faure's signature alone. IF IT LAY AT CITRUSDAL, then a man with seven minors he had undertaken to raise moved them some 500 km east between 1809 and 1816, and happened to settle in the very district that had just been carved out of the one he left. | THE COUNTER, STATED FAIRLY: this family did move long distances in these years. Ockert Almero is at Graaff-Reinet in 1817 and Beaufort West in 1821, and is NOT among the George households above. Trekboers moved. So this is a strong probability and not a proof. | AND THE ONE UNREAD THING IS STILL THE SAME ONE: the veldcornetcy holding these four is the one IMMEDIATELY AFTER Veldcornetschap Boven Olyphantsrivier, and its name is cut off in the photographs. THAT NAME IS THE FAMILY'S ADDRESS IN 1816. Film 008145125, image 45, left margin. | *** [20 Aug 2026] THE LAST ADDRESS IS SETTLED - SRC-522 - AND THIS LEAD KEEPS ONLY THE EARLIER YEARS. *** The 1811 George roll puts the foster household on the Olifants inside the district carved out of Swellendam that April, which closes the river question for 1805-1816. WHAT THIS LEAD STILL HOLDS IS FRANS'S SEQUENCE FOR THE 1790s: Roodezand 1796, Paarl 1799 and Swartland 1802 are Boland, there is no roll for this household between 1770 and 1805, and 'Tulbagh, Citrusdal, Swellendam, Koup' would explain all three and the family's Citrusdal memory besides. A ROLL OF THE 1790s STILL DECIDES THAT, and it is now a question about ORIGIN rather than about where he died. | [20 Aug 2026] MITCHELL'S CHAPTERS 1-4 DATE THE BOLAND OLIFANTS: sustained colonial occupation from 1725, seventy years before our family could have been there. Our own sweep dates the southern Olifants to one surveyed farm by 1802 and a quitrent conversion in 1832. So the two readings of the 1790s mean different things - settled country versus frontier - and both remain coherent. THE 1790s ROLL STILL DECIDES IT. | *** [20 Aug 2026] THE 1786 ANCHOR IS PAARL, NOT ROODEZAND. *** A Roodezand reading had been put under the Tulbagh-Citrusdal sequence off the surname profile of the page. The film's item table says PAARL G3/3/4 (SRC-076), and the surname reading is withdrawn. | THE SEQUENCE OF CONGREGATIONS IS NOW: PAARL 1786 - ROODEZAND 1796 - GRAAFF-REINET and PAARL 1799 - SWARTLAND 1802 - SWELLENDAM 1804 and 1808. Paarl twice, thirteen years apart, at the church where the grandfather was baptised in 1722. THAT READS AS A FAMILY TIE RATHER THAN A RESIDENCE, which is the distinction this whole day turned on. The 1790s roll still decides where they LIVED. | *** [20 Aug 2026] MEASURED ON THE RESIDENCE EVIDENCE ONLY - ROLLS, ESTATES AND LAND, NO BAPTISMS - AND IT HOLDS. *** Baptisms are excluded on the day's own lesson: they record a church, not a home. | THE SPINE, and every leg is a tax roll, an estate taken on the ground, or a survey: | 1722 DRAKENSTEIN | 1751, 1760, 1770 the SWELLENDAM rolls - 140 km ESE (103 deg) | 1805 MODDERFONTEYN 'aan d'Olyphants Rivier', 43 uuren out, and the 1809 estate taken on that farm - 225 km ENE (76 deg) | 1811, 1813, 1816 the GEORGE rolls, Veldkornetskap Oliphants Rivier - 18 km E (86 deg), and that leg is the BOUNDARY MOVING, not the family | 1821 BEAUFORT WEST on the roll - 134 km NNW (343 deg) | 1833 MIDDELWATER in the Koup - 109 km SSW (208 deg). | *** THE ANSWER: DRAKENSTEIN 1722 TO THE OLIFANTS 1811 IS 374 km ON A BEARING OF 87 DEGREES - ALMOST DUE EAST, ACROSS NINETY YEARS AND THREE GENERATIONS. *** Over the whole arc to Middelwater in 1833 the net displacement is 282 km EAST against 58 km NORTH in 111 years. Frans's 'gradual move east, north-east' is not an impression; it is the measurement. | *** AND IT SHOWS WHERE THE PATTERN BREAKS, WHICH IS THE USEFUL PART. *** The eastward run STOPS at the Olifants. The next leg is NORTH to Beaufort West and then back SOUTH-WEST into the Koup - the family goes ROUND the Swartberg, not over it, and arrives in the Koup from the north. That is the leg Meiringspoort cannot explain, being fifty years too late (EVT-000209), and it is why 9a was qualified today. | ONE THING THE MEASUREMENT CANNOT SEE, and it is worth saying: the 1751-1770 point is the SWELLENDAM DROSTDY used as a proxy, because a roll gives a district and not a farm. The household could have been anywhere in a district that then reached to the Gamtoos. If it was already at the eastern end in 1770, the eastward move happened a generation earlier and faster than these legs suggest. THE 1790s ROLL WOULD SHOW IT. | *** [20 Aug 2026, second pass] THE BAPTISMS ARE TIER B AND THE MEASUREMENT SAYS SO. *** [FAMILY] Baptism does not mean residence - could be far from residence, though it implies district i guess. Measured on the six registers he supplied the same day, plus SRC-076: Ockert Jacobus and Susanna Catharina van Tonder took SEVEN children to FIVE churches spanning 538 km - Paarl 1786, Paarl 1794, Roodezand 1796, Graaff-Reinet 1799, Swartland 1802, Swellendam 1804, Swellendam 1808. *** GRAAFF-REINET 1799 TO SWARTLAND 1802 IS 558 km APART IN THREE YEARS. *** Read as residence that column has the household crossing the Colony and back between confinements, so the caution is not theoretical. | WHAT THE COLUMN TRACKS IS KIN, NOT DISTANCE: Paarl is the family's own church (the stamvader's son baptised there 1722); Roodezand is where the grandfather married in 1749; at Graaff-Reinet the witness is ROELOF VAN TONDER, the mother's family, and the child is named Roelof for him. | AND THE TWO HALVES BEHAVE DIFFERENTLY. From 1804 the baptisms AGREE with the rolls - Swellendam twice, the district church for the Olifants until George took that ground in 1811, and the 1808 clerk wrote 'woonende aan Oliphants Rivier' against the entry (SRC-514) because the distance was worth noting. Before 1804 they scatter, and that is exactly the period the household cannot otherwise be placed. THE DISTRICT SIGNAL IS REAL BUT SWITCHES ON ONLY ONCE THE FAMILY IS SETTLED, which is the half where it is least needed. Recorded as a standing rule in 5, third place-data rule.

    De Modder Fonteyn (aan d'Olifants rivier)

    opened 2026-08-20

24. ✓ The Rondabel du Plessis ARE kin - Martha and Daniel are first cousins LEAD-0272

    MARTHA DU PLESSIS (PER-000750, 1730-1796) married Petrus Venter on 14 October 1753 and their son is JAN LOUIS VENTER (PER-000120, b.1754) - father of the sixteen Venter children, of whom Adriana Cecilia married the root Oosthuizen and Maria Catharina held Rondabel by 1829. So du Plessis is a DIRECT ANCESTRAL SURNAME one generation above the Venter bloc, and that part is not in doubt. | Two other du Plessis touch the same farm. IGNATIUS MICHAEL DU PLESSIS (PER-000050, d.1818) held HALF THE OPSTAL of a loan farm called Rondabel. CECILIA MARIA DU PLESSIS (PER-001346, 1857-1895) died on Rondabel and is buried there (GRV-0021, DOC-0333). Is either of them kin to Martha, and is Ignatius Michael's Rondabel even the same farm?

    What we expected: Ignatius Michael's farm is probably NOT ours. Cecilia Maria may well descend from the same du Plessis stock as Martha, but she has no parents recorded at all, so nothing can be said yet.

    What would settle it: A shared father and mother for Daniel and Martha du Plessis.

    What settled it: Opened 23 Aug 2026 on Frans's question about further du Plessis connections in the Rondabel and early-Koup context. The answer is yes and the best one was not the farm: MARTHA DU PLESSIS IS AN ANCESTOR, sitting directly above the Venter sibship of STO-0046. | *** [23 Aug 2026] THE QUESTION IS NOW SHARP AND SMALL. *** Cecilia Maria du Plessis's pedigree runs back through Johannes Gysbertus (1826-1913), Daniel Jacobus (1802-1879), Jan Gysbert (1781-1814) to DANIEL DU PLESSIS (PER-001895, 1732-1795). | MARTHA DU PLESSIS (PER-000750) WAS BORN IN 1730 AND DIED IN 1796. Daniel was born in 1732 and died in 1795. Two years apart at birth, one at death. Martha married Petrus Venter in 1753 and is the grandmother of the sixteen Venter children of STO-0046 - among them Maria Catharina, who held RONDABEL by 1829, and Adriana Cecilia, the root matriarch. | SO IF DANIEL AND MARTHA ARE SIBLINGS, the du Plessis woman buried on Rondabel in 1895 is kin to the du Plessis woman whose granddaughter held it in 1829, and the farm's chain becomes a family returning to its own ground across four generations. | IT IS NOT ESTABLISHED AND MUST NOT BE WRITTEN AS THOUGH IT WERE. Neither Daniel nor Martha has a recorded parent here, du Plessis is among the commonest surnames at the Cape, and two people born two years apart is a coincidence that costs nothing to produce. THE TEST IS ONE PAIR OF PARENTS. FamilySearch should have both. | AND THE 1818 RONDABEL IS STILL PROBABLY A DIFFERENT FARM - MOOC8/64.12a places it ON THE TARKA, 380 km east. That half of the old question is unaffected by any of this. | *** [23 Aug 2026] CLOSED, NEGATIVE. *** FamilySearch gives Daniel du Plessis (LC5W-J51, 1732-1795) the parents Charl Prieur du Plessis (1688-1736) and Cecilia van Maarseveen (1692-1762), married at Stellenbosch 12 June 1712, and lists THIRTEEN children in full: Charl 1716, Pieter 1716, Elizabeth 1716, Magdalena 1717, Johannes Jean 1720, Elisabeth 1722, Catharina 1723, Charel 1724, Philippus 1727, Jacob 1730, DANIEL 1732, Jacobus 1734, Anna Magdalena 1736. NO MARTHA. The 1730 place is Jacob, and the list runs to its end. | SO THE ANSWER IS NO, and the near-identical dates were the coincidence this lead said they might be. Two du Plessis born two years apart in a colony full of du Plessis is not a relationship. Recorded as a closed negative rather than quietly dropped, because a lead that dissolves is worth as much as one that lands - the alternative was a satisfying story about a family returning to its own ground, and it would have been wrong. | WHAT REMAINS OPEN AND IS NOW ITS OWN QUESTION: MARTHA DU PLESSIS (PER-000750, 1730-1796) still has no parents in this model. She is the grandmother of the root matriarch and of the Venter woman who held Rondabel in 1829, so her own line is worth having on its own account and not as a test of anything. | The 1818 Rondabel of Ignatius Michael du Plessis is unaffected and still probably a different farm - MOOC8/64.12a places it ON THE TARKA. | *** [23 Aug 2026, REOPENED AND ANSWERED THE OTHER WAY WITHIN THE HOUR] THE NEGATIVE ABOVE IS RIGHT ABOUT SIBLINGS AND WRONG ABOUT THE QUESTION. *** A third chart gives MARTHA DU PLESSIS's father as JEAN LOUIS DU PLESSIS (PER-001909, 1691-1768), and Jean Louis is a son of JEAN PRIEUR DU PLESSIS and MARIE MADELEINE MENANTEAU - the same couple who are Charl Prieur's parents. *** SO CHARL PRIEUR AND JEAN LOUIS ARE BROTHERS, AND MARTHA AND DANIEL ARE FIRST COUSINS. *** | THE SHARED ANCESTOR IS JEAN PRIEUR DU PLESSIS (1638-1708), the Huguenot: 4x-great-grandfather of CECILIA MARIA DU PLESSIS who died on Rondabel in 1895, and 2x-great-grandfather of both ADRIANA CECILIA VENTER, the root matriarch, and MARIA CATHARINA VENTER, who held Rondabel by 1829. The du Plessis buried on that farm and the du Plessis whose granddaughter held it descend from one man. | THE LESSON IS ABOUT THE TEST, NOT THE DATA. I asked whether they were SIBLINGS, got a correct no from a complete sibling list, and closed the whole question on it. A negative result is only ever as wide as the question asked of it, and the next relationship out was one step away. The right test was always 'do they share an ancestor', not 'do they share parents'. | Still true and unaffected: the 1818 Rondabel of Ignatius Michael du Plessis is a different farm, ON THE TARKA. | *** [23 Aug 2026, THIRD PASS - AND THE HEADLINE WAS OVERSTATED] HE IS RIGHT AND THE CLAIM NEEDS SPLITTING IN TWO. *** | WHAT IS NARROW AND REAL: Martha du Plessis and Daniel du Plessis are FIRST COUSINS. Their fathers, Jean Louis and Charl Prieur, are brothers. That is a specific documented relationship two generations wide and it stands. | WHAT IS NEARLY VACUOUS: 'they share Jean Prieur du Plessis'. JEAN PRIEUR DU PLESSIS IS THE DU PLESSIS STAMVADER AT THE CAPE - the Huguenot who landed in 1688 - so essentially every du Plessis in this country descends from him. Saying two du Plessis share him says little more than that they are both du Plessis, which the surname already said. Presenting it as a finding dressed a tautology as a discovery. | THE TEST THAT MATTERS is not whether a common ancestor EXISTS but HOW CLOSE he is, and four to six generations back at the Cape is not close. First cousins in 1730 is meaningful; a shared 1688 progenitor in 1857 is background. | THIS IS THE SAME ERROR AS THE ONE ABOVE, WEARING THE OPPOSITE HAT. First I asked too narrow a question and closed on a correct negative; then I widened until the answer was guaranteed. Neither is the right width. The lead stays closed and the record keeps all three passes.

    Rondabel

    opened 2026-08-23

25. ✓ The du Plessis of Reim Hoogte No. 167 - answered: they are Cecilia Maria's own family LEAD-0275

    The farm cemetery on REIM HOOGTE No. 167 (CEM-025) holds six du Plessis, and three of them are now named: JOHANNES GYSBERTUS (1826-1913) and JOHANNA ELIZABETH VAN ROOYEN (1831-1913), the parents of Cecilia Maria du Plessis who is buried on Rondabel; and JACOBUS JOHANNES PETRUS (1869-1941), who bought Rossel Erf on that farm in 1908. Is he their son and her brother? And who are ENA (1907-1926), GERTRUIDA ALIDA (1882-1970) and SUSANNA FREDERIKA (1920-1945)?

    What we expected: Brother, and the others his wife and descendants: Gertruida Alida (b.1882) reads as his wife, Ena and Susanna Frederika as daughters or granddaughters.

    What would settle it: A register or a FamilySearch parentage for Jacobus Johannes Petrus du Plessis b. 21 June 1869. The stones have been read and cannot answer it.

    What settled it: Opened 23 Aug 2026. Note the namesake hazard that nearly cost a wrong link an hour earlier: REIM HOOGTE No. 167 and RHEMHOOGTE No. 125 are different farms 12 km apart and both have burial grounds. CEM-005, CEM-025 and CEM-026 are on 167; CEM-027 is on 125. | *** [23 Aug 2026] THE FIRST STONE IS READ AND IT CONFIRMS THE NAME IN FULL. *** GRV-0711: 'JACOBUS JOHANNES PETRUS DU PLESSIS / GEB. 21 JUNIE 1869 / GES. 20 JUNIE 1941'. The initials on deed 3908 are no longer initials. | THE OTHER FIVE STONES IN THE PLOT ARE STILL UNREAD, and the photographs exist - IMG_6381 is his and IMG_6382 is Johannes Gysbertus's, so the sequence is on the eGGSA album. A stone often names a parent or a spouse in its dedication line, which is what would settle whether he is Cecilia Maria's brother. Reading the remaining images is the cheapest way to close this lead. | *** [23 Aug 2026] ALL TEN PHOTOGRAPHS READ, AND THE ANSWER IS THAT THE STONES CANNOT GIVE IT. *** Every marked grave at this ground now carries an exact date and a transcribed inscription: Johannes Gijsbertus 29 Juli 1826 - 25 Nov 1913; Johanna Elizabeth [geb. van Rooijen] 20 Nov 1831 - 25 Aug 1913; Jacobus Johannes Petrus 21 Junie 1869 - 20 Junie 1941; Gertruida Alida 15 Aug 1882 - 30 Sept 1970; Ena 9 Mei 1907 - 4 Julie 1926; Susanna Frederika 9 Jan 1920 - 30 Nov 1945; and Hendrik Johannes Neethling, August 1795 - Mei 1863. | *** NOT ONE STONE NAMES A PARENT, A SPOUSE OR ANY RELATIONSHIP. *** They give a name, two dates and a verse, and nothing else. The Neethling stone of 1863 is the only one in the plot that states a relation at all - 'ter gedachtenis aan hunnen geliefden vader door zyne dankbare kinderen' - and it names no child. So the question this lead was opened on, whether Jacobus Johannes Petrus is Cecilia Maria's brother, is NOT settled by the stones and must be settled from a register or from FamilySearch. | TWO GRAVES HAVE NO NAME AND NEVER WILL FROM THE PHOTOGRAPHS. GRV-1464 is a large gothic marble headstone whose whole inscription panel has SPALLED AWAY - the most ambitious monument in the plot, closest in style to the 1863 Neethling stone, which suggests an early burial and nothing more. GRV-1465 is an unmarked mound outlined in rough stones. Both are now rows, because a grave that cannot be read is still a grave. | ONE THING THE READING DID GIVE, and it is worth keeping: THE LANGUAGE TURNS IN THIS PLOT BETWEEN 1913 AND 1926. Neethling 1863 and both du Plessis stones of 1913 are in DUTCH - 'geb.', 'gest.', 'zoek in 's Hoogsten lof uw' lust'. Ena's stone of 1926 and everything after it is in AFRIKAANS - 'ter gedagtenis', 'oorl.', 'Rus'. One family burial ground datable by its own spelling. | AND THE COUPLE DIED THREE MONTHS APART IN 1913, BOTH ON A 25th. | *** [23 Aug 2026] CLOSED, AND THE ANSWER IS YES. *** A FamilySearch family view gives JOHANNES GYSBERTUS DU PLESSIS (1826-1913, KC34-PVM) and JOHANNA ELIZABETH VAN ROOYEN (1831-1913, L631-CW5), married 27 NOVEMBER 1848 AT GEORGE, with NINETEEN children. Two of them are ours: CECILIA MARIA 1857-1895 (GHW4-KDS) and JACOBUS JOHANNES PETRUS 1869-1941 (GHLG-66D), whose years match his Reim Hoogte headstone to the day. *** SO THE MAN WHO BOUGHT ROSSEL ERF ON REIM HOOGTE IN 1908 AND LIES ON THAT FARM WAS HER BROTHER. *** | Their parents confirm the fan chart exactly: Daniel Jacobus du Plessis 1802-1879 (LC5W-JYL) x Susanna Johanna Catharina van der Westhuizen 1803-1852 (9WVN-24K), married 4 AUGUST 1822 AT GEORGE. BOTH MARRIAGES AT GEORGE - the district the Olifants prequel ground sat in, so this family was George-district before it was Prince Albert. | WHAT IS STILL OPEN, and it is smaller: GERTRUIDA ALIDA du Plessis (1882-1970) is NOT among the nineteen, so she is most likely Jacobus Johannes Petrus's WIFE rather than a sibling; ENA (1907-1926) and SUSANNA FREDERIKA (1920-1945) would then be his children or grandchildren. Untested, and the same family view would settle it. | The family-tree fetched to answer this held a DIFFERENT Johannes Gysbertus du Plessis and a different Cecilia Maria - see FIX-0015.

    Reim Hoogte No. 167 · Rossel Erf (Reim Hoogte No. 167 Ptn 1)

    opened 2026-08-23

26. ✓ Anna Magdalena CARELSE or CALITZ - who was Johanna Machteld Petronella Strydom's mother? LEAD-0283

    Her baptism register of 13 Jan 1793 names the mother ANNA MAGDALENA CARELSE, and FamilySearch's index of the same entry agrees. This model has ANNA MAGDALENA CALITZ (PER-000338, 1771-1836, FS LZP9-TX6) on a [TREE] link. Which is right, and is PER-000338 even the same woman?

    What we expected: The register is likelier to be right than the tree. CARELSE appears TWICE in the one entry - as the mother and as the witness PETERNELLA CARELSE - and the child's own third name is Peternella, which is the vernoemingspatroon pointing at the mother's kin. Calitz and Carelse are not variants of each other, so this is probably not a spelling question but a wrong-woman question.

    What would settle it: The marriage record, or the mother's name as written on two or three more of the children's baptisms.

    What settled it: Also carries a DATE TENSION on PER-000138: the model has her born 1792-01-05 from the tree, the register baptises her 13 Jan 1793, and infants on those same pages are baptised days after birth. A birth of 5 Jan 1793 would fit exactly. THE RECORD GIVES NO BIRTH DATE, so nothing is changed. | Not overwritten anywhere - 10, report what looks wrong with the evidence and let Frans decide when it is a question of identity rather than proof. | *** [23 Aug 2026, THE SAME EVENING] CLOSED, AND THE MODEL WAS RIGHT. *** SRC-556, Anna Magdalena Strydom's own baptism of 14 November 1790: mother ANNA MAGDALENA CALES, witnessed by JOHANNES MATTEWIS CALES and JOHANNES MATTEWIS CALES DE JONGE. CALES IS THE REGISTER'S FORM OF CALITZ, and this model already held two men of that name, PER-000785 (b.1748) and PER-000316. A senior and a junior on both sides is not a coincidence. PER-000338, Anna Magdalena Calitz, stands. | THE LESSON IS ABOUT AGREEMENT, NOT ABOUT THE SURNAME. The reading of the 1793 entry here and FAMILYSEARCH'S INDEX of the same entry both gave CARELSE, and the name appeared twice on that page, which is what made it look solid. Two independent readings of ONE HAND agreeing is not two pieces of evidence - it is one, read twice. A SECOND DOCUMENT settled it in a sentence. | [23 Aug 2026, A THIRD CONFIRMATION] SRC-558, the 1822 roll, line 17: 'DANIEL JACOBUS STRYDOM EN ANNA MAGDALENA CALITS'. Calits, Cales (SRC-556), Calitz (the model) - one family. The CARELSE that this project's reading and FamilySearch's index both produced from SRC-554 is a misreading, now beaten by two independent documents.

    opened 2026-08-23

27. ✓ Roelof Oosthuisen and Hester Maria Strydom in the Beaufort West register, 1838 LEAD-0288

    A ROELOF OOSTHUISEN stands witness twice in the 1838 Beaufort West baptisms (BWB-212-01 and BWB-214-16), the second time beside a HESTER MARIA STRYDOM. This model has ROELOF ADOLPH OOSTHUIZEN (PER-000147, 1799-1883), a son of the 1755 Ockert, married to HESTER MARIA CATHARINA STRYDOM (PER-000591, b.1802). Same two people?

    What we expected: Probably yes. The given names match on both sides, the pairing is unusual, and the district is right - the family is enrolled at Beaufort West from 1821 and Roelof Adolph was 39 in 1838. A Hester Maria Catarina Strydom also witnesses on the same page as the first sighting, BWB-212-02, on the same day.

    What would settle it: One entry with Roelof Oosthuizen as the father and Hester Maria Catharina Strydom as the mother.

    What settled it: Also relevant: BWB-214-02 has a JAN LOIS VINTER married to a Susanna Johanna Strydom with an ANNA MAGDELENA OOSTHUISEN standing - the three in-law surnames of the 1809 heirs on one entry. And the C. Hatting circle of BWB-215-10, 220-11, 223-08, 231-10 and 274-07 is a separate and denser group. None of it is matched. | *** [23 Aug 2026, HOURS AFTER IT WAS OPENED] CLOSED BY A ROLL. *** SRC-558, the 1822 Beaufort roll, line 22: 'ROELOF OOSTHUIZEN EN HESTER MARIA CATHARINA STRYDOM' - the two named as a household, not as two witnesses on one page. PER-000147 and PER-000591. The lead was opened on BWB-212-01 and BWB-214-16, where the same two names stand in 1838 witness lists; the roll says outright what those only suggested. THE SHAPE WAS RIGHT AND THE SHAPE IS NOT PROOF - what closed it was a second KIND of document.

    opened 2026-08-23

28. ✓ Which country is the Veldkornetskap Boven Olyphants Rivier - Cango, Attaquaskloof, or the Langkloof? LEAD-0296

    SRC-459 reads the 1816 George roll as putting the subtotal 'VELDCORNETSCHAP BOVEN OLYPHANTSRIVIER' above the Oosthuizen block and concludes they are in the NEXT ward. Image 181 of the 1813 roll carries 'BOVEN OLYPHANS RIVIER' on the carried-forward line at the head of the page the four brothers close. Which is right?

    What we expected: Probably that they ARE in Boven Olyphants Rivier and the 1816 reading needs revisiting. On image 181 the words stand in the MANS NAAMEN column against 'P:r Transport' and a carried-forward total, which reads as a ward label carried across a page break rather than as a closing subtotal. But SRC-459's reading was made from the arithmetic of the running totals, which is a careful argument and not a glance.

    What would settle it: The ward heading immediately before and immediately after the Oosthuizen block, in both rolls.

    What settled it: SRC-525 names a VELDKORNETSKAP OLIPHANTS RIVIER in volume J 91 and SRC-459 names a VELDCORNETSCHAP BOVEN OLYPHANTSRIVIER in J 97. Whether those are one ward or two is part of the same question. | *** [23 Aug 2026] FRANS AGREES, AND NAMES THE COUNTRY: 'yes, seems so - Cango'. *** So the working reading is that the Oosthuizens ARE in the Veldkornetskap BOVEN OLYPHANTS RIVIER in 1813, and that its ground is the CANGO. | WHAT SUPPORTS IT. The ward's own places column names that country: image 166 carries 'PAARDE FONTEIN BY DE CANGOO BERG', and image 182 carries 'WILLEM HOUT RIVIER / CAMMANIE BERG' at 40 uur, 'WILGE HOUT RIVIER / CAMMANIE BERG' at 41 and 'DE JAGERS RIVIER' at 40 - the Cango and the Kammanassie, the two ranges either side of the upper Olifants (SRC-561, SRC-563). And it agrees with the strongest independent evidence this project has: SRC-549 puts Faure's party at d'OUDE MURAGIE 'op 't eynde van de Cango' on 19 April 1809 and at de Modder Fonteyn four days later, about 32 km apart. THE ROLL AND THE LANDDROST'S CIRCUIT POINT AT THE SAME COUNTRY. | *** WHAT DOES NOT FIT YET, AND IT IS NOT SMALL. *** SRC-525 matched the farm list of the Veldkornetskap Oliphants Rivier against the Surveyor-General's gazetteer and found SEVEN UNIONDALE farms - Keurfontein 45, Misgund 25, Gorgida 14, Buffels Klip 63, Wilge River 46, Schoongezigt 52, Wilgehoute Rivier 69 - which is the LANGKLOOF, east of Oudtshoorn, not the Cango. A ward that runs from the Cango berg to the Langkloof is a long ward. Either it genuinely spanned that arc, or 'Oliphants Rivier' (J 91) and 'Boven Olyphants Rivier' (J 90, J 97) are TWO DIFFERENT WARDS - which is the second half of this lead's question and is still open. | AND IT BEARS DIRECTLY ON WANHOOP: SRC-563 reads 'WANHOOP EN MISGUNT' at 55 hours on image 182, and MISGUND is one of the Uniondale seven. If that row is the Oosthuizen row (LEAD-0295) then the family's ground is Langkloof, not Cango, and the two halves of this note pull apart. THE ALIGNMENT DECIDES IT. | *** [23 Aug 2026] FRANS: 'either that or attaquas kloof. stil workign out'. SO THERE ARE THREE CANDIDATES, NOT TWO, AND EACH HAS REAL EVIDENCE. *** | (1) THE CANGO. The ward's own places column names it - 'PAARDE FONTEIN BY DE CANGOO BERG' on image 166, 'CAMMANIE BERG' twice and 'DE JAGERS RIVIER' on image 182 - and SRC-549 puts Faure's party at d'Oude Muragie 'op 't eynde van de Cango' four days and about 32 km from de Modder Fonteyn in April 1809. | (2) THE ATTAQUASKLOOF. PLC-000084 already exists for it, and it is a NAMED FIELD CORNETCY OF THE GEORGE DISTRICT - diagram 219/1821 says so outright, 'the fieldcornetcy Attaquas kloof district of George' (SRC-469). And SRC-467's Finding 4 puts OOSTHUYSENS IN IT FROM ABOUT 1782: MOOC8/50.101a, the estate of MARTHINUS OOSTHUYSEN JANSZ on the loan place WELBEDAGT, 'GELEGEN INDE ATTAQUASCLOOF AAN DE OLIPHANTS RIVIER', with three Oosthuysen men appraising it on the spot in May 1792. That is a documented Oosthuysen presence in that kloof a generation before 1813. | (3) THE LANGKLOOF. SRC-525 matched seven of the Veldkornetskap Oliphants Rivier farms to UNIONDALE - Keurfontein 45, Misgund 25, Gorgida 14, Buffels Klip 63, Wilge River 46, Schoongezigt 52, Wilgehoute Rivier 69 - and SRC-563 reads WANHOOP EN MISGUNT on image 182. | *** AND THE ROLL ITSELF CARRIES THE DISCRIMINATOR: THE HOURS COLUMN. *** Distances from the George drostdy read so far: DE MIST KRAAL 32 | WELGEVONDE OLYPHANTS RIVIER 36 | ZOODE WAL / OLS KRAAL / ONSPAENG 37-39 | DE JAGERS RIVIER 40 | WILLEM HOUT RIVIER / CAMMANIE BERG 40 | WILGE HOUT RIVIER / CAMMANIE BERG 41 | KRUYS RIVIER AAN OLYPHANTS RIVIER 44 | WANHOOP EN MISGUNT 55. *** A SPREAD OF 32 TO 55 HOURS IS A LONG WARD, AND WANHOOP IS THE FARTHEST THING ON IT. *** That is consistent with the Langkloof being the ward's far end and the Cango/Kammanassie its near end. | *** WHICH MAKES THE COLUMN CALIBRATABLE, AND THAT IS THE METHOD WORTH TRYING. *** Misgund 25 and the other six Uniondale farms have coordinates in the Surveyor-General's gazetteer (SRC-510). If Wanhoop en Misgunt is 55 hours and its position is known, hours-per-km on this ward's roads can be solved - and then EVERY OTHER FARM'S HOURS BECOMES A RADIUS FROM GEORGE. Including, if the alignment in LEAD-0295 lands on the Oosthuizen row, this family's own. LEAD-0297. | *** [23 Aug 2026] CLOSED. THE ANSWER IS DE CANGO, AND THE DOUBT THIS LEAD WAS OPENED ON WAS WRONG. *** So a ward name on a 'P:r Transport' line is the CLOSING subtotal of the ward just finished, carried over - not a heading for what follows. 'BOVEN OLIPHANS RIVIER' at the head of image 181 therefore CLOSES that ward, and the households below it - the four Oosthuizen brothers among them - open the next one, which image 185 names 'd CANGO'. SRC-564. | SRC-459 WAS RIGHT ALL ALONG and its own sentence was the warning: 'That distinction is the difference between a finding and a confident wrong answer, and it is legible only from the arithmetic.' A ward name at the top of a page was read as a heading, and a P1 was opened against a careful argument. The correction cost one sentence from Frans. | THE THREE CANDIDATES RESOLVE: THE CANGO, which Frans named on instinct and the structure now proves. ATTAQUAS KLOOF is a real George field cornetcy and Oosthuysens did hold ground in it in 1792 (SRC-467) - but on this roll it is a SEPARATE AND LATER WARD, closing at image 192, eleven pages on. The LANGKLOOF question survives only as the Wanhoop/Misgunt anomaly, which is LEAD-0295's alignment. SETTLED 23 Aug 2026 by the inbox batch. Four SG diagrams of 1832 on adjoining Uniondale ground write the ward three ways - 'boven Oliphants Rivier' (DOC-0339, DOC-0341), 'upper Oliphantsriver' (DOC-0340) and 'Olyphantsrivier' (DOC-0343). Same year, same district, neighbouring farms: one ward, written loosely. Recorded as PLC-000089.

    opened 2026-08-23

29. ✓ Our portion numbers do not match the SG on Vrolykheid and Spreeufontein LEAD-0008

    Do our portion numbers for Vrolykheid and Spreeufontein correspond to the Surveyor-General portion numbers, or are they our own sequence?

    What would settle it: The portion number printed on each diagram, against the SG portion number in the cadastre.

    What settled it: Consequence for now: those parcels are LEFT UNMATCHED rather than given a polygon that is probably wrong. Better a hole in the map than a farm drawn in the wrong place. | 29 Jul 2026, verification pass: VROLYKHEID HALF LARGELY RESOLVED - the mismatch was ours, not the SG's. DOC-0028's verso had been transcribed with Lot B and Marinco swapped; corrected (Marinco=177/3, Lot B=177/5) the cadastre fits: Ptn 1 within 2.5%, Ptn 2 within 2.2%, Ptn 3 within 1.1%, Ptn 5 within 0.9%, Ptn 6 (Lot A, 15226 sq ft) vs 0.2 ha. ONLY FRANS KRAAL (177/4, 213.5677mg = 182.9 ha vs modern Ptn 4 = 1,776.0 ha) still fails - left unmatched. Spreeufontein half of this lead unchanged. | [1 Aug 2026] *** THE VROLYKHEID RESIDUE IS RESOLVED, AND BY ARITHMETIC RATHER THAN OPINION. *** This lead's last outstanding failure was Ptn 4: '213.5677mg = 182.9 ha vs modern Ptn 4 = 1,776.0 ha'. Re-reading the 1934 diagram today gives the extent as 2113.5677 MORGEN, not 213.5677 - a dropped leading digit. 2113.5677 mg x 0.856532 = 1,810 ha against the cadastre's 1,776.0 ha: a 2% difference, which is the same order as Ptn 1 (2.5%), Ptn 2 (2.2%), Ptn 3 (1.1%) and Ptn 5 (0.9%). The old figure was out by a factor of ten. So Ptn 4 was never the odd one out; the number was. CAUTION, AND IT IS THE INTERESTING PART: the SAME 213.5677 appears in DOC-0028's verso deduction table, which was itself carefully re-read on 29 July. So either the 1833 verso also carries the dropped digit, or one of the two readings is wrong. THREE THINGS NOW AGREE ON 2113 - today's reading of the 1934 sheet, the modern cadastre area, and the pattern of the other five portions - and one thing says 213. The verso is worth a third look. Ptn 4 can now be matched to the cadastre; the Spreeufontein half of this lead is untouched. | [17 Sep 2026, lead review] RESOLVED ON THE SPREEUFONTEIN HALF, SO THE LEAD CLOSES. Our portion numbers ARE the Surveyor-General's: the 1898 deduction table (DOC-0015) numbers Lot A = Ptn 3, Lot B = Ptn 4, Lot C = Ptn 5, and the provincial cadastre (layer 2, queried 17 Sep 2026) agrees on area once later subdivisions are taken out: 3/26 = 5,230.5 ha against Lot A 6,103 mg 122 sr (5,227 ha, 0.1%); 4/26 = 5,240.5 ha against Lot B 6,098 mg 243 sr (5,223 ha, 0.3%); 5/26 = 5.2 ha against Lot C 4 mg 480 sr (3.9 ha); RE/1/26 = 5,200.4 ha against Ptn 1's 18,309 mg less the three lots (6,104 mg, 5,228 ha, 0.5%). The one wrong thing was OURS: PAR-000022 (Lot A) had been pinned to RE/3/26 (3,224.6 ha), the remainder after 7/26 (2,005.9 ha) was cut from it; re-pinned to 3/26. Not placed: Ptn 2 (190 mg) and 8/26 (7,246.5 ha), and the 1903 remainder Klein Waterval, cancelled in 1991 into Farm 215.

    Lot B (Vrolikheid No. 177 Ptn 5) · Faans Kraal (Vrolikheid No. 177 Ptn 4) · Marinco (Vrolikheid No. 177 Ptn 3) · Vredendal (Spreeufontein No. 26 Ptn 1) · Lot A (Spreeufontein No. 26 Ptn 3)

    opened 2026-07-29

30. ✓ Where exactly is Vischat / 'fisgat', and what is its real name? LEAD-0018

    Vischat is where Frans's grandfather Hendrik Lodewyk Malherbe Oosthuizen farmed until 1979 and where his father CJJ grew up, yet the model has no ground for it: no extent, no geometry, no ownership row, no document. Which parcel does it actually sit on, and is the name Vischgat/Visgat?

    What we expected: A werf on one of the farms the family held rather than a farm in its own right. The farm_no cell reads '124/142' = Rondable + Frans Kraal, and Tokkie Marincowitz - who BOUGHT Vischat - farmed Rondabel until 2003, so Rondable No.124 is at least as likely as the 'in Frans Kraal' note suggests. The name is probably VISCHGAT ('fish pool').

    What would settle it: The purchase deed - it will name the parcel Vischat forms part of. Failing that, a topo sheet showing the werf, or CJJ simply saying which farm it was on.

    What settled it: RESOLVED 31 Jul 2026. 19"S 22°37'05.06"E, which point-in-polygon puts inside PAR-000009 RONDABEL, 5.2 km from the Rondawel werf. The lead's own prediction was right - 'Rondable No.124 is at least as likely as the in-Frans-Kraal note suggests'. Recorded as PLC-000034. What is NOT resolved is the spelling and the name's origin, which is now LEAD-0024.

    Hendrik Lodewyk Malherbe Oosthuizen · Christiaan Johannes Joubert Oosthuizen · Gert Annis Oosthuizen Marincowitz · Vischat · Faans Kraal (Vrolikheid No. 177 Ptn 4) · Rondabel · Vrolikheid

    opened 2026-07-30

31. ✓ Wilgemond's diagram number: 150/1833, or 350 mis-transcribed in 1952? LEAD-0025

    Every other farm in Stretch's August 1833 run carries a diagram number between 348 and 369, and their quitrent folios rise with them. Wilgemond is numbered 150/1833 yet holds folio 5-9, which falls exactly inside that run. Is the number wrong?

    What we expected: Probably 350. The sheet we hold is a RECOPY made 15.9.1952 (no. 1841/52), so the typed number and the handwritten one are not independent readings - both would carry a single 1952 slip. With 350 the series reads 348->5-6 (Klein Valie), 349->5-7 (Middelwater), 350->5-9 (Wilgemond), 351->5-10, 354->5-17, 355->5-18 (Vrolikheid), 356->5-30, 366->5-3? (Rondabel), 369->5-43 - monotonic throughout. With 150 the correlation collapses, and with it the argument that fixes Vrolykheid's disputed folio at 18.

    What would settle it: The original diagram, or the SG's index of 1833 Prince Albert diagrams.

    What settled it: Do not amend the Vrolykheid or Rondabel postscripts on this yet. Both are marked [TO VERIFY] already and both are still the better reading; this lead records the one contrary data point so it is not lost. | [17 Sep 2026, lead review] CLOSED ON THIS LEAD'S OWN TEST. It said: 'if 350 is already taken by another farm, the hypothesis dies.' 350/1833 is Stretch's diagram of the loan farm KLAARE STROOM, De Claar Stroom No. 178, held as DOC-0189 since 8 Aug 2026. So Wilgemond's 150/1833 is not a miscopied 350, and the folio 5-9 sitting inside the run stays unexplained but is no longer evidence of a wrong number.

    Charles Lennox Stretch · Wilgemond No. 179 · Middelwater · Klein Valie No. 182

    opened 2026-07-31

32. ✓ Are the two 1839 Ockert Almeros one man, or did Ockert Almero and Adriana Cecilia really have two? LEAD-0033

    PER-000133 (FS GVZF-ZHH, b. 8 Jul 1839, baptised 20 Aug 1839) and PER-000364 (FS G6ZT-D65, b. 1839) are both Ockert Almero Oosthuizen and both children of Ockert Almero and Adriana Cecilia. Different family-tree records, different FamilySearch ids. One man or two?

    What we expected: One man, recorded twice. Two sons of the same name in the same year is only possible as twins, and the vernoemingspatroon's name-reuse rule requires the first child to have DIED before the name is used again - which cannot happen inside one year. PER-000133 carries a baptism date of 20 Aug 1839 and PER-000364 carries nothing but the year, which is the usual shape of a duplicate rather than of a twin.

    What would settle it: The register page. Twins would be entered together, as the 1825 Swellendam page shows.

    What settled it: Detector gap this exposed: fs_corrections.py finds our-duplicates by a shared FamilySearch id and familysearch-duplicates by identical name AND identical birth string. This pair has neither - different ids, and '1839' against '1839-07-08'. It slipped through both. The precision-aware comparison already written for merge_person.refines() would catch it. | ACTED ON 5 Aug 2026, AND SAY WHICH TEST DID IT. PER-000364 was merged into PER-000133 through merge_person.py. What settled it was NOT the register page this lead asked for - that has still not been read. It was the sibship arithmetic: Ockert Almero and Adriana Cecilia carried sixteen children in parentage, 5b's independently tested count is fourteen, and SRC-077 names thirteen surviving in 1859. Dropping this pair and the Susanna Catharina pair (PER-000461 into PER-000123) gives fourteen. The merge brought a wife and twelve children onto PER-000133, which contradicts Pottas's 'unmarried' and sharpens the SRC-077 omission - both written onto that row rather than resolved. THE REGISTER PAGE FOR AUGUST 1839 IS STILL WORTH READING, and now has a second question to answer: whose children are PER-001088 to PER-001099.

    Ockert Almero Oosthuizen (b.1839) · Ockert Almero Oosthuizen · Adriana Cecilia Venter

    opened 2026-07-31

33. ✓ Duplicate candidates from the one-degree pull, settled on FamilySearch ids LEAD-0035

    38 people added on 31 Jul 2026 share a name AND a birth YEAR with somebody already in the model, and 32 share a name with no date at all. Are they the same people?

    What we expected: Most of the year-matches ARE duplicates and most of the undated ones are not - a thin tree record with no date is usually a different person of a repeated name rather than a second copy.

    What would settle it: A full birth or baptism date on either side of a pair. Name plus YEAR is not enough here and must not be treated as enough.

    What settled it: CLOSED 31 Jul 2026, the same day it was opened, and closed by the right key. s that are unique - so you should be able to see when it is duplicate or different." He was right and the key was already in the record: MacFamilyTree writes the FamilySearch person id into a custom _FID tag and 44,141 of the tree's 44,143 individuals carry one. The sqlite export drops that column, which is the whole reason the pull matched on tree id alone and duplicated people who had no tree id. | 853 FamilySearch ids stamped onto the model from the family-tree; 55 pairs then proved to be one person recorded twice and were merged. That is EIGHT TIMES what name-and-date matching had found, and it caught pairs no name test could: Eugene against Eugenius, 'Caroline Mircle' against 'Caroline Merle', 'Pieter Cornelis Oosthuizen' against 'Pieter Cornelis Oosthuisen'. | The other direction matters as much: of 350 same-name pairs among live people, 329 are now PROVEN DISTINCT because both sides carry a FamilySearch id and the ids differ. Those are no longer candidates for anything. 21 remain undecidable, all because one side or the other has no FamilySearch id at all - see LEAD-0036.

    Ockert Almero Oosthuizen · Anna Helena Roscher · Daniel Johannes Oosthuizen (b.1811) · Anna Magdalena Oosthuizen (b.1818) · Jacobus Johannis Oosthuÿsen (b.1849) · Charles Frederick Marincowitz (b.1878)

    opened 2026-07-31

34. ✓ The surname column was blank on 62% of people, fixed, and the lesson LEAD-0042

    796 of 1,270 live people have no surname recorded. What can safely be counted by surname, and what cannot?

    What we expected: Nothing aggregate. The gap is not random - it is worst on the imported one-degree rows, which is precisely where a family tally would draw its numbers from.

    What would settle it: Populating it from the family-tree's own SURN field via kin_sweep.

    What settled it: CLOSED 31 Jul 2026, same day. FIXED: kin_sweep --fill-names reads the family-tree's own SURN and GIVN tags, which the sqlite export drops, and filled 848 surnames, 191 given names, 118 dates and 472 line assignments. Coverage 38% -> 99%. The 8 rows still without a surname are two entities and six single-name people, and every one is correct. | AND THE STING IN IT: the 'Botha x Oosthuizen, 5 times' finding that this lead was opened to warn against IS REAL. Recomputed on complete data it is still the commonest cross-surname pairing in the district. The artefact was my CORRECTION of it, not the original - I recomputed on 'recorded surnames only', which was the same broken 38%, and declared a true finding false. A partial dataset does not only produce false positives; it produces confident false NEGATIVES, and those are worse because nobody goes back to check them. | THE REAL LESSON, now in AUDIT.md as stage 1a: every check in this project looked for data that was WRONG and none looked for data that was ABSENT.

    opened 2026-07-31

35. ✓ Did Ds Adriaan Hofmeyr travel to Mashonaland in 1897, after he had left Prince Albert? LEAD-0044

    SRC-068 has Hofmeyr serving Prince Albert 1883-1895 and then joining Jan Luttig and Cecil John Rhodes on a Mashonaland expedition in 1897, two years after he left the town. The organ that ended up at Zeekoegat came out of that trip.

    What we expected: Both can be true, he need not have been minister here to travel, but the model dates the organ gift 'about 1897' on the strength of this passage, and if the trip is misdated the organ date moves with it.

    What would settle it: A dated account of the N'gami expedition.

    What settled it: [17 Sep 2026, lead review] ANSWERED AS FAR AS PUBLISHED ACCOUNTS GO, AND CLOSED. Afrikaans Wikipedia's article on Adriaan Hofmeyr (13 April 1854 - 1 May 1937) dates his Mashonaland journey to 1891: a four-man church commission left on 31 March 1891, reached Fort Salisbury (then 400 to 500 inhabitants) by early July, visited the Zimbabwe ruins and was back by 10 August 1891; his colleague G. Stegmann died on the way home. His congregations: Willowmore 1881-83, PRINCE ALBERT 1883-95, Wynberg 1895-99, Kuruman 1928-33. So he went while still minister at Prince Albert, and the cameo's '1897, after he had left' (SRC-068) is most likely a misdating of the 1891 journey. Neither source puts Rhodes or Jan Luttig in the 1891 party; the invitation story is the cameo's alone [TO VERIFY]. The organ's route to Zeekoegat is not touched. Tertiary sources (af.wikipedia.org/wiki/Adriaan_Hofmeyr; gemeentegeskiedenis.co.za gives the same congregations and no journey).

    ds Adriaan Hofmeyr · Jan Luttig · Cecil John Rhodes

    opened 2026-08-01

36. ✓ Images of documents are not date-gated the way PDFs are - should they be? LEAD-0046

    A PDF of a document is published only if the document is dated before 1936, because no text check in this build can see inside one. A JPEG rendered from a .tif or .jpg scan is exactly as opaque, and has never been put through that gate. 20 published renderings are of documents that are undated or dated 1936 or later. Tighten the gate, or narrow it?

    What we expected: Neither extreme. Most of the 20 are SG diagrams with no date FIELD rather than no date - the diagram year is often in the filename or the deed reference, so filling those in removes most of the problem without withholding anything. The real decision is the handful genuinely dated after 1936, of which DOC-0017 (a 1939 death notice, quoted in S002) is the clear case: the subject died in 1939 and is certainly beyond privacy, but the mechanical rule cannot know that.

    What would settle it: on whether a document's date should govern its IMAGE as well as its PDF - and, if yes, dates filled in for the diagrams that merely lack the field.

    What settled it: FOUND BY THE TEST WRITTEN FOR A DIFFERENT BUG. L15 was added because the new PDF-preview feature shipped a page-one JPEG of every WITHHELD PDF, including two 2013 deeds searches naming current registered owners. That is fixed. The test then reported this, which is older and wider and is NOT being changed silently, because tightening it would pull evidence off live pages. | CLOSED 1 Aug 2026. The answer was yes, and the hole was bigger than the lead said. derive_images() in build_site.py was a THIRD consumer of sources/ with no gate on it at all, so every raster went to the site regardless of date while copy_scans() and derive_pdf_previews() both checked. It now calls the same scan_is_publishable(). 40 rasters are withheld as a result: 9 diagrams dated 1937 to 1990, a 1939 and an undated death notice, an undated plot plan, and 10 FamilySearch-sourced images that were being republished against Frans's cite-and-link decision. Two other defects fell out of it. _doc_year() took the MINIMUM of every date on a row, so a document whose dates straddled 1936 published on its oldest one; it now takes the latest CONTENT date and treats image_edtf as the copy date rather than content. And a verso, which carries no date of its own, now inherits its front's, so the backs of thirteen diagrams are not collateral. Caught by DOC-0157, a 1939 partition diagram that came in through the inbox and published silently. The leak test now FAILS on a raster instead of reporting one.

    opened 2026-08-01

37. ✓ Were Jacobus and Wynand van Zyl twins, born 6 December 1809? LEAD-0048

    The van Zyl family history says Jacobus and Wynand were TWINS, both born 6/12/1809. The model has Jacobus at 20 Sep 1809 and Wynand at 1808. Which is right?

    What we expected: The baptism register decides it and nothing else will. Twins are unusually easy to confirm - two entries on one day, often on one line - so this is a cheap question with a clean answer. Worth noting the stakes: if they were twins, then TWO TWIN BROTHERS MARRIED TWO SISTERS off one Karoo farm, which is a considerably better fact than two brothers doing so.

    What would settle it: A baptism entry for either brother giving a date, or both on one day.

    What settled it: Also unresolved on the same rows: PER-001263 and PER-001264 look like each other's duplicate - two Wynand van Zyls, both 1808, both recorded as spouse of Johanna Jacomina Oosthuizen. Settle the twin question and the merge probably settles with it. | [1 Aug 2026, SRC-180 - the baptism register] CLOSED, AND ANSWERED NO, ON THE SAME DAY IT WAS OPENED. The register has both entries consecutively on page 148: Wijnand 'geb: den 3 Junij 1808' and Jacobus 'geb: den 20 September 1809', same parents, BOTH BAPTISED 6 DECEMBER 1809. They were baptised together and born fifteen months apart. The family history read the shared baptism date as a shared birthday. THE LESSON IS THE USEFUL PART, and it is the ordinary shape of this error: a family history records a DATE it found attached to two people and infers a RELATIONSHIP from it. So the answer to 'were two twin brothers married to two sisters' is no - but two brothers were, and that was never in doubt. Also worth noting the model needed no correction: both dates were already right.

    Jacobus Stephanus van Zyl · Wynand van Zyl

    opened 2026-08-01

38. ✓ Two Prince Albert parcels numbered 172, and a Haggas the cadastre extract cannot see LEAD-0053

    What is the 1,420 ha parcel also numbered 172 at (-33.3165, 22.8003), and where exactly is Haggas No. 145?

    What we expected: On 172: probably a detached piece of the same farm, or a portion the extract has typed as a farm. Both features carry RD=PRINCE ALBERT and the SAME CSG link, so they are not two different registration divisions - which makes this a different bug from Haggas, and a subtler one. On HAGGAS: it is NOT IN THE CADASTRE FILES AT ALL - no feature named Haggas or Haggis anywhere, and farm 145 in the Prince Albert extract is ROODE PUNT WES at lon 21.65, nowhere near. But the official CSG viewer plainly shows a 143/144/145/147/148 block immediately east and south-east of Oorlogs Kloof 175, exactly where both 1833 and 1863 diagrams put Haggis. THE EXTRACT SIMPLY DOES NOT REACH THAT FAR EAST. That is a gap in our download, not in the record.

    What would settle it: A cadastre feature named Haggas with a PRCL_KEY, touching Oorlogs Kloof on its east or south.

    What settled it: Raised by Frans spotting a hole in the map between Vrolikheid and Oorlogs Kloof, and by his screenshot of the official CSG viewer showing RE/172 hard against 1/177. The Oorlogskloof half is FIXED; this lead is the residue. | CLOSED 1 Aug 2026. Haggas, Karee Rivier and Spitskop are all drawn, from the Council for Geoscience's national cadastre (layers 9 and 8), which serves the Eastern Cape and therefore Willowmore. fetch_cadastre.py now falls back to it for divisions the Western Cape service cannot serve, and normalises MAJ_REGION/TAG_VALUE into the RD_name/P_FARMNAME shape every consumer already expects.

    Oorlogs Kloof No. 172 (Roode Els Bosch and Oorlogs Kloof) · Haggas No. 145

    opened 2026-08-01

39. ✓ Two Rietfontein rows: one farm, or one farm and a phantom? LEAD-0058

    Should PAR-000010 "Rietfontein" and PAR-000052 "Rietfontein (Zeekoegat district)" be one row, and does PAR-000010 describe a farm that ever existed?

    What we expected: PAR-000010 is probably a phantom. Its only support is one unsourced legacy line saying a portion of Middelwater No. 169 became Rietfontein, and the 1883 verso (DOC-0039) lists no such deduction from 169.

    What would settle it: Either a Middelwater deduction table naming a Rietfontein, which would make PAR-000010 real and separate, or the absence of one across the full set, which would retire it into PAR-000052.

    What settled it: Raised by , spotting that both rows showed the same index number 38/56. That part is now fixed: both were wearing Ptn 38 of Rietfontein No. 56 and neither is pinned any more. Merging or retiring a parcel id is a decision for Frans, not a cleanup, because ids are permanent. | [CLOSED 9 Aug 2026] FOLDED INTO LEAD-0080, which asks the same Rietfontein question as one of its three parts and carries the other two with it. Nothing is lost: the detail from here, that both rows were wearing Ptn 38 of Rietfontein No. 56 and that neither is pinned any more, is carried across. Merging or retiring a parcel id remains a decision for Frans.

    Rietfontein · Rietfontein (Zeekoegat district)

    opened 2026-08-01

40. ✓ 1,302 cadastre keys carry more than one polygon, of different sizes LEAD-0060

    Is PRCL_KEY safe to use as a unique identifier in the Western Cape extract, and if not, what is?

    What we expected: No, it is not unique. It looks like the extract carries both a farm-level and a portion-level feature under one key in some cases, and genuine duplicates in others.

    What would settle it: Either a rule that picks the right feature deterministically (PRCL_TYPE F vs FP already does most of it), or confirmation from the Surveyor-General that the extract has duplicates.

    What settled it: Found during the audit of 1 Aug 2026. | [17 Sep 2026, lead review] ANSWERED AND CLOSED. Re-counted over geo/*.geojson: 1,412 keys carry two or more features of different area. In 204 the features differ in WSTATUS - a current parcel (C) beside a superseded one (H). In the other 1,208 status and PRCL_TYPE are the same, and the features are PIECES OF ONE PARCEL: adjoining or detached parts (a road, a river, a sliver), e.g. W053C009000000000177000011 is two touching polygons of 249.4 and 1,811.1 ha. So PRCL_KEY IS the parcel identifier; it is not a feature identifier. The rule, which match_cadastre.py already applies: skip WSTATUS='H' unless a parcel is explicitly pinned to it (Modderdrift Ptn 4), and take ALL remaining features of the key as one MultiPolygon, summing their areas.

    Modderdrift No. 106 · Rondabel No. 124 Remainder

    opened 2026-08-01

41. ✓ A township in East London had been entered as a Karoo farm LEAD-0061

    Was the parcel created from DOC-0032 a real farm, and how many others rest on a document nobody opened?

    What we expected: Retire it. Nothing references it and the document it rests on is about ground 500 km away.

    What would settle it: Ids are permanent, so retiring means marking the row, not deleting it.

    What settled it: Raised 1 Aug 2026, on an audit of whether all the source files had been mined. Two other things fell out of the same audit: DOC-0003's entire extraction was the word "consolidation" for a diagram that names two grantees and two grant dates, and the Klein Sleutlefontein scans were on disk under DOC-0091/0092, ids belonging to two estate inventories. | CLOSED 1 Aug 2026, the same day. The parcel is retired: the row is removed from parcels.csv and the id is never reused, which is what retiring means here. Nothing referenced it - no ownership, no neighbours, no story links, no parcel links, no residences, all verified zero before removing it - and the full record of what the sheet actually is lives on DOC-0032 and SRC-045, which are corrected rather than deleted because the document is real, it is simply about East London. PROOF THE CHECK WORKS: build_land_db.py refused the first rebuild after the row went, because this lead still named PAR-000037 in three of its own fields. A retired id cannot be left pointed at.

    opened 2026-08-01

42. ✓ "[?Hoeagus]" is HAGGAS No. 145 - RESOLVED by the widow's own death notice LEAD-0072

    What is the place written as the residence in field 8 of death notice No. 682 of 1863, and where is it?

    What we expected: A werf or farm in the District of Prince Albert. The word is not confidently read - recorded as "[?Hoeagus]" - and it matches NO farm name in the Prince Albert or Beaufort West cadastre, which was checked name by name including every name beginning with H. That absence is not evidence against it. This project's second place rule says the cadastre records TITLE, not what people call the ground they live on, and Drooge Kloof and Vischat are already on the books as real places with no polygon. A man can die at a werf that never reached a diagram.

    What would settle it: A clearer image of field 8, or the estate inventory naming where his stock and ground lay.

    What settled it: CLOSED 4 Aug 2026. The farm is HAGGAS No. 145 (PAR-000094), surveyed 1833 for P. VAN DER WESTHUIZEN, which is the dead woman's own surname. Recorded as a RESIDENCE, not ownership, at Frans's explicit caution: "it says they lived there, not necessarily owned it." One discrepancy stays open - the cadastre puts farm 145 in the Willowmore division while both notices say Prince Albert - and is noted on the parcel rather than argued away.

    Cornelis Petrus Janse van Vuuren · Wilgemond No. 179

    opened 2026-08-03

43. ✓ Three farms carry the same name twice or three times, and ten story links are held back because of it LEAD-0080

    Which Rietfontein, which De Claar Stroom, and is Sleutelfontein (Groot and Klein) a real parcel or a composite of two?

    What we expected: A cross-reference audit on 4 Aug 2026 found 22 farms named in a story's prose with no link to that story. Twelve were added. TEN WERE DELIBERATELY NOT ADDED, because the name in the prose matches more than one parcel row and guessing which one would be the namesake error this project keeps making. RIETFONTEIN, THREE ROWS: PAR-000010 (called a subdivision of Middelwater, its own note says "[TO VERIFY - probably wrong]"), PAR-000052 (Zeekoegat district, granted c.1858 to J.S. de Villiers), and PAR-000156 (Riet Fontein No. 12, established from its own diagrams on 4 Aug 2026). Two stories mention a Rietfontein - the gold rush and the Middelwater history - and neither can be linked until it is known which. LEAD-0058 already asks half of this. DE CLAAR STROOM, TWO ROWS: PAR-000004, the 1777 loan farm and parent of Middelwater and Klaarstroom, and PAR-000097, Klaarstroom No. 178. Four stories name it. The two are genuinely different things - a parent loan farm and a modern farm - so each mention needs reading, not a bulk link. LEAD-0039 is the same question. SLEUTELFONTEIN: PAR-000098 is "Sleutelfontein (Groot and Klein)", described in its own notes as TWO farms, while PAR-000071 and PAR-000072 hold Groot Sluitel Fontein No. 126 and its portion separately. A row that is two farms cannot be linked as one place.

    What would settle it: For each mention, which parcel the sentence means. For Sleutelfontein, whether PAR-000098 should exist at all given PAR-000071/072.

    What settled it: [4 Aug 2026] THE SLEUTELFONTEIN THIRD IS ANSWERED, by Frans and [FAMILY]: "Klein and Groot sleutel fonteint had various spellings over the years. sleutel, sluitel, etc but seems it was always 2 distinct farms - at least since surveyd." So PAR-000098 stays what it was already made on 3 Aug - a COLLECTIVE REFERENCE resolving to PAR-000071 (Groot, No. 126) and PAR-000073 (Klein, No. 127) - and STO-0020 now links all three: the collective name the source used, and the two farms it means. THE LEAD REMAINS OPEN for the other two thirds, Rietfontein (three rows) and De Claar Stroom (two), which are untouched by this. Note for anyone reading the two rows: KLEIN SLEUTELFONTEIN IS THE LARGER FARM, 14 398,3 ha against 4 199,1 - so the names cannot be used to guess which row a document means. | [4 Aug 2026] THE DE CLAAR STROOM THIRD IS ANSWERED, by [FAMILY] De claar stroom 178 and the other overlapping farm is the same. PAR-000004 (the 1777 loan farm) and PAR-000097 (Klaarstroom No. 178) are one piece of ground recorded under two epochs, not two farms. PLK-000053 changed from 'subdivision' to 'same_as' - a farm cannot be a subdivision of itself - and both parcel rows now say so. Checked before acting: no polygon really overlaps 178 apart from neighbours' shared edges, so the overlap is between the RECORDS. ONLY RIETFONTEIN REMAINS OPEN on this lead: three rows, PAR-000010, PAR-000052 and PAR-000156, with two stories that cannot be linked until it is known which one they mean. | [9 Aug 2026] ABSORBED LEAD-0058, the Rietfontein third of this question asked on its own. Carried across: Frans spotted on 1 Aug that PAR-000010 and PAR-000052 both showed index number 38/56, so both were wearing Ptn 38 of Rietfontein No. 56; neither is pinned any more. The open part is whether PAR-000010 describes a farm that ever existed, or is a phantom of the same name. Note that farm 12 is a THIRD Rietfontein and is its own question, LEAD-0085. | [13 Aug 2026] CLOSED - ALL THREE THIRDS ANSWERED. Sleutelfontein by Frans on 4 Aug (two distinct farms; PAR-000098 a collective reference). Rietfontein by Frans on 13 Aug: the S011 subdivision sentence means PAR-000052 - the Zeekoegat farm - not Riet Fontein No. 12; S002 was already linked to PAR-000052. De Claar Stroom by Frans on 13 Aug: the S011 chain row means PAR-000004 - the 1777 parent loan farm. Links written; whether PAR-000052 and PAR-000156 are one farm remains LEAD-0058 and is NOT decided here.

    Rietfontein · Rietfontein (Zeekoegat district) · Rietfontein No. 12 · De Claar Stroom (the 1777 loan farm) · De Claar Stroom (Klaarstroom No. 178) · Sleutelfontein (Groot and Klein)

    opened 2026-08-04

44. ✓ Four pairs of live person rows look like one person each, and they are withholding links LEAD-0082

    Are PER-001263/PER-001264 (Wynand van Zyl), PER-000562/563, PER-000646/001087 and PER-000956/957 each one person or two?

    What we expected: ONE PERSON IN EACH PAIR, and the duplication imported faithfully from the private tree rather than created here. Each pair shares a name and an exact birth date while holding TWO family-tree ids and TWO FamilySearch ids, which is the signature of a tree that has the person twice. The clearest is WYNAND VAN ZYL, PER-001263 and PER-001264, both b. 3 June 1808 and both recorded as spouse of Johanna Jacomina Oosthuizen - Ockert Almero and Adriana Cecilia's eldest daughter, so this is not a peripheral row.

    What would settle it: Whether the two FamilySearch records are the same person. If they are, merge on FS first, then here with merge_person.py, tombstoning the loser as status=duplicate - never deleting, ids are permanent.

    What settled it: Opened 4 Aug 2026 alongside REVIEW-witnesses.md, which holds 23 witness links awaiting the same kind of decision. 5 Aug 2026: a pass over parentage found 15 FURTHER probable duplicate pairs of the same shape - siblings sharing a canonical first given name and born within a year of each other, each side carrying a DIFFERENT familysearch_id and tree id, none flagged status='duplicate'. Examples: PER-000342/PER-000141 (Susanna Elizabeth, b.1784), PER-000611/PER-000612 (Martha Jacoba Venter, both d.1822), PER-000251/PER-000673 (Dirk Daniel Joubert, both b.1874-01-19), PER-000391/PER-000392 (Aletta Francina Theron, both b.1860-07-29), PER-000828/PER-000830 (Gert/Gerrit Adriaan van Zyl, both b.1846-04-03). ONE PAIR MUST NOT BE MERGED: PER-000628 and PER-000630, both Johanna Jacomina van Zyl, are TWO CHILDREN - the Beaufort West register holds both baptisms, born 1841-04-21 and 1842-10-28 to the same parents, and the model already carries PER-000628 d. ../1842-10-28. That is 5b's died-young-name-reused case documented from a primary source rather than inferred, and it is the reason this whole class needs a register and not a script. | [22 Sep 2026] ALL SETTLED. The four pairs in the title were merged in earlier passes (PER-001264, 000563, 001087 and 000957 are tombstones). Of the five further pairs this lead names, two were already merged (PER-000342, 000673) and three are merged today, each on same parents plus an exact birth date or a shared death year: Aletta Francina Theron (PER-000391 into 000392), Gert/Gerrit Adriaan van Zyl (PER-000830 into 000828) and Martha Jacoba Venter (PER-000611 into 000612). Each pair carried two FamilySearch ids, so FamilySearch itself very likely holds duplicates of these three [for [FAMILY] GD9F-1SP / G7MZ-7J5, LHZQ-V15 / GJB6-7CP, L1K1-VG1 / GDQG-1GF]. PER-000628 and PER-000630 remain two children, as this lead says. New candidates are listed by checks.py in PEOPLE-POSSIBLE-DUPLICATES.md, which is the live tracker.

    opened 2026-08-04

45. ✓ Four portions have a polygon in geo/ that match_cadastre.py still will not pick up LEAD-0084

    Why does the matcher miss PAR-000023, PAR-000058, PAR-000069 and PAR-000144 when their exact PRCL_KEYs are present?

    What we expected: A KEYING FAULT IN THE MATCHER, not a data fault. The two causes found on 4 Aug were both in the data - a blank plc_id, and an sg_farm_key of NO-POLYGON asserting an absence that was untrue - and fixing those placed six parcels. These four survive that fix: their division, farm number and portion number are all filled and the key each one implies is present in geo/. So the remaining suspect is how match_cadastre.py builds or normalises its key for a portion.

    What would settle it: Printing the matcher's computed key beside the cadastre key for one of these four.

    What settled it: Listed in known_issues.txt so the build is not blocked while they stand. | *** [4 Aug 2026] CLOSED, AND THE PREDICTION IN THIS LEAD WAS WRONG. *** It said the cause was "a keying fault in the matcher, not a data fault". It is neither. All five polygons carry WSTATUS='H' - SUPERSEDED - and match_cadastre.py keeps superseded features out of the name and farm-number indexes deliberately, holding them by LPI alone so that only an explicit PIN can reach one. The matcher was doing exactly what it was built to do, and the loader says so in a comment written on 2 Aug. WHAT WAS ACTUALLY REQUIRED was a judgement per parcel: is the superseded shape the right historical ground? Judged on extent, the same test that justified the Modderdrift Portion 4 pin. THREE PINNED: Spreeufontein Portion 4 (5 223,1 ha against 5 240,5, 0,3%), Annex Modderdrift (2 036,0 against 2 043,4, 0,4%) and Rhemhoogte Portion 6 (4 728,2 against 4 781,2, 1,1%). TWO REFUSED: Botterkraal, where our 3 418 morgen is 2 927,6 ha against the polygon's 1 653,8 - 77% apart, not the same ground - and Rhemhoogte Portion 7 at 4,65 ha against 5,6, which is 20% and no corroboration even though the absolute gap is under a hectare. Those two stay unplaced, which is honest; a wrong pin would not be. THE CHECK THAT RAISED THIS WAS ALSO WRONG and is fixed: it now ignores WSTATUS='H' features, so it no longer accuses the matcher of missing what it is designed to skip.

    Lot B (Spreeufontein No. 26 Ptn 4) · Botterkraal (Jansenskraal No. 104 Ptn 2) · Annex Modderdrift (Rhemhoogte No. 125 Ptn 3) · Swartskraal (Rhemhoogte No. 125 Ptn 6)

    opened 2026-08-04

46. ✓ Willem Venter and Anna Christina Oosthuizen stood as witnesses together four times over twenty years LEAD-0087

    Who were they, and were they a married couple? Four joint appearances between 1839 and 1859 make them the most active godparent pair in the register.

    What we expected: A MARRIED COUPLE, on the pattern the register itself shows. The only other repeated pair in 291 witness rows is Johanna Jacomina Oosthuizen with Wynand van Zyl - and those two are known to have married at Beaufort West on 1 February 1829. Couples stand together at the font; that is what the pairing means. So Willem Venter and Anna Christina Oosthuizen appearing together on 7 Oct 1839, 25 Dec 1848, 7 Jul 1850 and 2 Oct 1859 most likely means a Venter married an Oosthuizen, in a district where four such marriages are already recorded.

    What would settle it: A marriage entry naming both. Failing that, a baptism where one of them is a parent rather than a witness.

    What settled it: This is the second of three; the first is the forge on PER-000118. | [22 Sep 2026] ANSWERED FROM THE REGISTER ALREADY TRANSCRIBED: they were husband and wife, the parents in BWB-238-09, BWB-263-05 and BWB-284-11 (three daughters, about 1840-1843, the first named Anna Christina for her mother). Recorded as UNI-000365. She is PER-001665 (created by the witness sweep of 9 Aug, so this lead's 'in no model row' is out of date); the husband is very probably PER-000522, Willem Venter b. 1819, son of Jan Adriaan Venter of Rondavel - see the union's note for why, and the marriage entry is what would make it certain. Her own parents are not yet known: the private tree has no Anna Christina Oosthuizen.

    Willem Venter

    opened 2026-08-04

47. ✓ Reopen: is the 1818/1823 Rondabel of the inventories ours or an eastern namesake? LEAD-0092

    Were the field cornets Stephanus van Wyk (1818) and Abraham Greyling (1815) Koup wards or eastern Graaff-Reinet wards?

    What we expected: EAST, on the evidence LEAD-0014 itself assembled before it was closed. MOOC8/44.48 shows that when a clerk meant our farm he wrote farm, ward, region AND district - "the place Rondavel ... field cornetcy of Hermanus van der Westhuisen in the Traka district of Beaufort". The 1818 and 1823 inventories name none of those, say Tarka, and the de Bruin estate sold its sheep to a GRAHAMSTOWN butcher.

    What would settle it: A ward list for Graaff-Reinet and for Beaufort in those years.

    What settled it: [22 Sep 2026] ANSWERED: AN EASTERN NAMESAKE. THE 1818 FIELD CORNET IS PLACED, and he is on the EASTERN Tarka. The Records of the Cape Colony (SRC-662) carry 'S. J. van Wyk, Fieldcornet' reporting to the Landdrost of Graaff-Reinet from 1816 to 1823: on commando out of Cradock in November 1816, with provisional field cornets under him writing from the Tarka, Baviaans River and the Brak River in 1822-1823 (all printed in vol. XXXI, pp. 1-44, among the 1827 enclosures on commandos), and 'Mrs. Van Wyk' of the Tarka losing cattle in 1825. So the 'leeningsplaats genaamd de Rondabel gelegen aan de Tarca' of the 1818 du Plessis inventory, taken in Stephanus van Wyk's ward, is the eastern Rondabel on the Tarka river, not ours on the Traka. [Abraham Greyling (1815) is not placed by these volumes: a Greyling family signs a frontier memorial under field cornet C. Meyer, which points east too but does not settle it.] S010 already calls these 'probably a namesake farm', which this confirms for 1818. LEAD-0014's Traka reading of this project's own Rondabel (PAR-000009) is unaffected: it rests on the diagrams, not on these inventories.

    Rondabel

    opened 2026-08-05

48. ✓ Nine diagram sheets named in the model that no longer exist on disk LEAD-0095

    Re-pull nine Surveyor-General sheets whose archive references this model records but whose images were never kept. What do their versos carry?

    What we expected: MOSTLY VERSOS WITH DEDUCTION TABLES, and therefore mostly PORTION NAMES. That is what the second sheet of a diagram usually is in this collection, and 8 says so in as many words - 'the back of the sheet is where the deduction table lives, which is where the portion names and the later history live, so that is exactly the half worth losing'. The Klein Valie No. 182 batch of 5 Aug 2026 is the worked example of the payoff: nine portions, eight of them named, every name read off a sheet. Four of the seven affected rows are farms whose portions this model still holds unnamed or unnumbered.

    What would settle it: Re-pulling the nine files. Each then gets its OWN DOC- row per 8, never a second filename on an existing row.

    What settled it: The rule this breaks is 8, 'A SHEET IS A ROW', which was itself written on 3 Aug 2026 after the same mistake was made and corrected that morning. These seven rows pre-date the rule or slipped past it. CLOSED 5 Aug 2026, the same day: Frans re-pulled all nine sheets from the Surveyor-General. Each now has its own DOC- row (DOC-0238 to DOC-0246) and the seven parent rows carry a single filename again. THE VERSOS WERE WORTH IT - between them they carry the full eight-portion deduction table for Minnies Kraal No. 112, six named portions of Rondabel, and the one line that resolves the biggest extent anomaly in the model (see LEAD-0093).

    Farm 215 · Minnies Kraal No. 112 · Groot Sleutelfontein (Rhemhoogte No. 125 Ptn 2) · Modderdrift No. 239 · Swartskraal (Rhemhoogte No. 125 Ptn 5) · Rondawel (Rondabel No. 124 Ptn 1)

    opened 2026-08-05

49. ✓ Whose grave is on Vredendal, 24 July 1886? LEAD-0120

    A slate headstone lying flat on Vredendal is cut for a burial of 24 July 1886. The surname reads as Botha and is not certain, the given name and the age are illegible on the photograph. Who is buried there, and is it one grave or a family plot?

    What we expected: Not an ancestor, on present evidence - no Botha holds Vredendal in this model, and the van der Merwes come to it in the 1950s. More likely an earlier occupier of the ground or a neighbour buried where he died. That would make it the kind of grave this project should hold anyway: the people on the land, not only the family.

    What would settle it: A readable surname and age from the stone, or a death notice for a Botha at Spreeufontein in July 1886.

    What settled it: The same photograph places Vredendal inside Spreeufontein Portion 1, which had been an open question on PAR-000006. That part is recorded there at medium confidence. A second geotagged photograph, at the werf, would firm both up at once. | [CLOSED 9 Aug 2026] ANSWERED. The stone is GRV-0025 and the burial is ENGELA SUSARA NIEHAUS, born BOTES (PER-001526), who died on 24 July 1886 aged 49 years 2 months 22 days. The inscription reads in full and is transcribed on STO-0033. THE SURNAME THAT 'READ AS BOTHA' IS BOTES, which is worth keeping: the two are one letter apart on a weathered slate and the district holds both families. It is also NOT one grave but a plot of at least five, with two Botma stones of 1910 and 1939 beside it and a fifth name on the eGGSA list this project has not seen. The 'medium confidence' on the portion, mentioned in the note above, is also gone:

    Vredendal (Spreeufontein No. 26 Ptn 1)

    opened 2026-08-06

50. ✓ Three Botes dead in one year at Saairivier, 1904 LEAD-0132

    Barend Stefanus Botes (b.1882), Hendrik Willem Botes (b.1883) and Gertruida Elezabeth Jacoba Botes born Niehaus (b.1862) all died in 1904 and lie in one small ground on Blaauw Draay 3 near Leeu-Gamka. What killed three of one household in one year, and are they the Botes and Niehaus of Varsche Fontein and Vredendal?

    What we expected: A mother and two grown sons. The ages fit - she was 42 and they were 21 and 22 - and a single year points at disease or an accident rather than ordinary mortality.

    What would settle it: Three death notices, which would give parents, cause and estate.

    What settled it: Raised 7 Aug 2026 with the Leeu Gamka sweep to keep grounds where people relate. | [7 Aug 2026] ANSWERED FROM FAMILYSEARCH. She is Gertruida Elizabeth Jacoba Niehaus, KLYD-KST, born Prince Albert 8 May 1862, married Andreas Stefanus Botes there on 13 December 1880, died 22 April 1904 - and her parents are Hendrik Willem Storm Niehaus [PER-001527] and Engela Susara Botes [PER-001526], the couple of the Vredendal stone. The two young men beside her are her sons, and one of them carries her father's names. WHAT KILLED THREE OF THEM IN 1904 IS STILL OPEN.

    opened 2026-08-07

51. ✓ The Calvinia cemetery is photographed, and the grave of Schalk Willem Burger van der Merwe is in it LEAD-0153

    Frans photographed the Calvinia cemetery in August 2013 beside the grave of SWB van der Merwe (1833-1904). Janet Melville reported the cemetery had by then been shot for eGGSA. What do the stones say, and which of this line lie there?

    What we expected: SEVERAL OF THIS LINE SHOULD BE THERE. Jacobus Adriaan van der Merwe died at Calvinia in 1855 and was buried there; Schalk Willem Burger died on Oude Muur in 1904 and his stone is in the town cemetery. A photographed cemetery is the cheapest gravestone evidence this project can get, and eGGSA is already a source it uses.

    What would settle it: Transcriptions of the van der Merwe stones in the Calvinia cemetery.

    What settled it: Opened 10 Aug 2026 from the Melville correspondence. | CLOSED 11 Aug 2026. The stone is photographed and indexed: "MERWE VD Schalk W.B. 1833-1904" in the Calvinia main cemetery, eGGSA, and the identification rests on the initials, the surname and both years against a baptism register and a death registration. GRV row added. THE WIDER QUESTION THE LEAD ASKED - which of this line lie in that cemetery - is answered differently than expected: Calvinia's main cemetery holds over a hundred van der Merwe stones and only ONE of them is a person this model holds. The others are a district full of the surname. Two more direct ancestors were found in the same sweep, but in other towns: PER-000762 at Brandvlei and PER-000219 with his wife at Williston. SRC-389.

    opened 2026-08-10

52. ✓ Welbedacht is a family farm this project has never entered LEAD-0169

    Jacobus Alewyn van der Merwe was known as "Kwaai Koos Welbedacht". His daughter was born there in 1835 and married there in 1854, and his wife died there in 1855. Which Welbedacht, and can it be matched to the cadastre?

    What we expected: IT IS IN THE CALVINIA DIVISION AND THERE ARE AT LEAST TWO CANDIDATES. The CSG name index, asked on 11 Aug 2026 for a different reason, returned WELBEDAGT No. 555 and WELBEDACHTE No. 611, both CALVINIA, as well as WELBEDACHT No. 537 and No. 139 in CLANWILLIAM. The 1854 marriage register puts the wedding "op de plaats Welbedacht" in the PAROCHIE VAN CALVINIA, DISTRIKT VAN CLANWILLIAM, which in 1854 is not a contradiction - Calvinia was a parish inside Clanwilliam - so the district on the sheet does not choose between them. THE OTHER BRANCH'S GROUND SHOULD SETTLE IT BY PROXIMITY. Oude Muur is at -31.333, 19.291 and Tygerhoek at -31.464, 19.707. Jacobus Alewyn is Jacobus Adriaan's brother, so Welbedacht should be in the same country, and the two Calvinia candidates can be measured against it the way farm 601 and 791 were for LEAD-0151.

    What would settle it: A diagram or deed naming a van der Merwe on Welbedagt No. 555.

    What settled it: Opened 11 Aug 2026 from the FamilySearch note "Bekend as Kwaai Koos Welbedacht" and the 1854 marriage register. | [11 Aug 2026] PART-ANSWERED THE SAME HOUR, from the cadastre already on disk. The Calvinia division holds exactly ONE farm whose name begins WELBED - no. 555, WELBEDAGT, at -31.3032, 19.9192 - and it is 26.9 km from Tygerhoek and 59.7 km from Oude Muur, so it sits in the same country as the rest of this branch. It is now PAR-000199 with its cadastral polygon and is drawn on the map. WHAT IS STILL OPEN is the half that matters: no document puts a van der Merwe on that title. One candidate in the right division at the right distance is a good identification and it is not proof, and this project has spent the day watching farm-name matches fail. The next step is the CSG document index for Calvinia 555. | [CLOSED 11 Aug 2026, SRC-417/DOC-0291] WELBEDACHT IS WELBEDAGT No. 555, CALVINIA, and it is PAR-000199, which was already in this model with a matched polygon at -31.3032, 19.9192. The 1823 manuscript names it in words: the loan place Welbedagt, 4031 morgen 188 square roods, measured for CAREL AARON VAN DER MERWE in erfpacht on 2 June 1823. So the farm behind the byname "Kwaai Koos Welbedacht" was his FATHER'S ground before it was his.

    opened 2026-08-11

53. ✓ Is GW32-G7M the Oosthuizen stamvader this model already holds, or an extra generation? LEAD-0171

    The son's family-tree runs the Oosthuizen father-line Jacobus 1722-1802 -> Johannes Oosthuizen 1672-1730 (GW32-G7M) -> Johannes Oosthuizen 1648-1682. This model's stamvader is Johannes "Jan" Oosthuijsz, LCJN-657, born 1665. Same man, or a generation this project lacks?

    What we expected: PROBABLY THE SAME MAN, seven years apart on a birth year, which is ordinary for this period. Against that, the family-tree disagrees with this model on the stamvader's own birth as well - 1675 against 1665 - so it has him twice at two dates, which is what a duplicate looks like. This model's row rests on SRC-101 and SRC-102, Amsterdam and Leiden baptisms.

    What would settle it: Whether the two ids are one person on FamilySearch, or two.

    What settled it: Opened 11 Aug 2026 during the six-person import from the family-tree. | CLOSED THE SAME HOUR, WITHDRAWN AS A QUESTION, AND WRONGLY OPENED. This project settled the question long ago and the answer was sitting in the model: GW32-G7M is PER-000559, ALREADY TOMBSTONED as status='duplicate' of PER-000296, and PER-000296 carries the birth as the EDTF range 1665/1672, which holds BOTH readings on purpose, with the same death of 4 November 1730. Same man, one id retired, no missing generation. THE MISTAKE WAS THE SAME ONE MADE TWICE EARLIER ON 11 AUGUST with FIX-0007 and FIX-0008: the comparison excluded status='duplicate' rows, so a tombstoned id looked like an absent one. CHECK TOMBSTONES BEFORE CONCLUDING SOMETHING IS MISSING - that is now three for three in one day.

    Johannes "Jan" Oosthuijsz (stamvader)

    opened 2026-08-11

54. ✓ Read the diagrams for Brasse Fontein 371 and Welbedagt 555 LEAD-0172

    Five SG diagrams are in the inbox unread: 10051901 and 10051902 for farm 371 in two parts, 10052409 for 371, and 10053943 and 10053944 for Welbedagt 555. Do any of them name a van Wyk, a Nel or a van der Merwe on the title or a boundary?

    What we expected: BOTH FARMS ARE IDENTIFIED BY INFERENCE AND NEITHER IS PROVEN. Brasse Fontein 371 is where Christina Elizabeth Nel died and Welbedagt 555 is the only farm of that name in the division and Jacobus Alewyn van der Merwe was known as "Kwaai Koos Welbedacht". A diagram naming the family on either would turn an inference into a fact, and this project has watched farm-name matches fail all week.

    What would settle it: A van Wyk on the title of 371, or a van der Merwe on 555.

    What settled it: Opened 11 Aug 2026. The diagrams arrived while the data was being written and have not been read. | [CLOSED 11 Aug 2026, SRC-415 to SRC-418] ALL FOUR SHEETS READ, AND THE ANSWER IS SPLIT. WELBEDAGT 555: yes, and better than asked - the 1823 diagram was measured FOR CAREL AARON VAN DER MERWE himself. BRASSE FONTEIN 371: a van der Merwe is on a title, but in 1915, not 1890 - Portion 2 (Brassefontein West) was transferred to JACOBUS A. VAN DER MERWE by deed 8683 on 30 December 1915, twenty years after Christina Elizabeth Nel died there. AND THE GRANTEE IS NOT A VAN WYK. Both portions read "granted to M.J. VAN DYK December 12th 1890", and the initial letter matches the D of December on the same typed line. The family report of a van Wyk grant is NOT confirmed by these sheets; it may still be true of the Remainder, which neither sheet describes. No Nel appears on either.

    Brasse Fontein No. 371 · Welbedagt No. 555

    opened 2026-08-11

55. ✓ Which Bloemendal did Tokkie own, and when did he hold it and Tabaklande? LEAD-0187

    [FAMILY] Tokkie also owned Bloemendal and Tabaklande. Tabaklande is unambiguous - farm 164 portion 2, now PAR-000201. BLOEMENDAL IS NOT: this district has three farms of that name and a fourth thing called Bloemendal inside one of them. Bloemendal Annex No. 164 (PAR-000128), Bloemendaal No. 165 (PAR-000093), Bloemendal A No. 166 (PAR-000129), and the 2023 valuation roll also names farm 165 portion 2 'Bloemendal' at 15.4 ha. Which one is it, and what are the dates of both holdings?

    What we expected: Tabaklande is a portion of Bloemendal Annex No. 164, so 'Bloemendal and Tabaklande' most naturally means the Annex and a piece of it. Against that, Bloemendaal No. 165 is the larger farm and the plainer referent of the bare name. Suggestive, not evidence: this project has been caught before by a name that looked like it could only mean one thing.

    What would settle it: Frans naming which Bloemendal, or a deeds report showing the transfers in and out.

    What settled it: Opened 13 Aug 2026. The ownership row OWN-000104 records the Tabaklande holding at LOW confidence with no dates, because family knowledge gave the fact and not the years; nothing here invents a range. NO OWNERSHIP ROW WAS CREATED FOR BLOEMENDAL AT ALL, because a span has to point at one parcel and there are three candidates - recording it against the likeliest would have manufactured a fact out of a guess. | [13 Aug 2026] OWN-000105 now records the holding against Bloemendal Annex No. 164 (PAR-000128), because Tabaklande is that farm's portion 2 and he named the two together - a pairing, not a document. The lead stays OPEN on the question it was opened for: which of the three Bloemendals. A deeds search on farms 164, 165 and 166 settles it, and so does one word from Frans. | [14 Aug 2026] CLOSED, AND THE ANSWER WAS NOT THE ONE THIS LEAD'S OWN HYPOTHESIS PREFERRED. Frans gave a coordinate, -33.273998, 22.375727; it falls inside BLOEMENDAL A No. 166 (PAR-000129), and he confirmed: "that is the farm Tokkie had... he had Tabaklande too". So the two holdings are two separate farms 1.9 km apart, not a farm and its own portion, and the pairing argument that put OWN-000105 on Bloemendal Annex No. 164 was wrong. Worth keeping as a worked example: three farms of one name inside 2 km, and no amount of reasoning about which reading was more natural could separate them - a point on the ground did it in one move. What is still open is not this lead: the DATES of both holdings, which a WinDeed search on farms 166 and 164 portion 2 would give. | [14 Aug 2026] THE ANSWER RECORDED YESTERDAY WAS WRONG, AND SO WAS THE MORAL DRAWN FROM IT. 274613 / 22.388809 - and it falls inside BLOEMENDAAL No. 165 (PAR-000093), 1.22 km from the point that closed this lead and inside a different farm. The holding moves there. Note what this does to the sentence written above, that "a point on the ground did it in one move": the point did move it, twice, and the second move was to a third farm. The lead stays CLOSED on which Bloemendal, now farm 165, and the DATES of both holdings remain open - a WinDeed search on farm 165 and on farm 164 portion 2 is the way to settle them.

    Gert Annis Oosthuizen Marincowitz · Tabaklande (Bloemendal Annex No. 164 Ptn 2) · Bloemendal Annex No. 164 · Bloemendaal No. 165 · Bloemendal A No. 166

    opened 2026-08-13

56. ✓ The Modderfontein cluster in Riversdale - is it the Swellendam ground the family came from? LEAD-0222

    SRC-451, the 1867 death notice, says Jacobus Johannes Oosthuizen was born in the DIVISION OF SWELLENDAM about 19 November 1784. The cadastre holds a dense MODDERFONTEIN cluster in RIVERSDALE - farms 417, 569, 570, 579, 670, 685 and 367, all within two kilometres of -34.26, 21.40 - and Riversdale was carved out of Swellendam in 1838. Is that the family's earlier ground, and is it the same name travelling with them?

    What we expected: WORTH TESTING AND EASY TO OVERREAD. What supports it: the death notice puts the eldest brother's birth in Swellendam division in 1784; farm 417 is cut into some forty portions, which is what an old loan place looks like after two centuries; and a family carrying a farm name to new ground is documented in this project already - SRC-183 has Jan Adriaan van Zijl born at Middelwater in the Koup and dying at a Klaarstroom in the Marico. | WHAT CUTS AGAINST IT: the Riversdale farms drain to the Goukou and the Kafferkuils and there is no Olifants within a hundred kilometres, so this is NOT the farm the 1809 inventory was taken at. "Modderfontein" is also among the commonest farm names in the country - the same query returns 19 farms and 128 portions across the province - so the name alone establishes nothing. 9: never match on a name alone.

    What would settle it: An opgaafrol or a loan-farm register entry naming an Oosthuizen at a Swellendam Modderfontein before 1809. Absent that, the name is a coincidence with a good story attached.

    What settled it: The 1809 estate is the fixed point and it says Olifants river. Nothing here touches that. | Note also that the family's traceable path runs Swellendam (b.1784) - Modder Fonteyn on the Olifants (1809) - Graaff-Reinet (1808-1814) - George (1816) - Traka (1818) - Beaufort West (1819). Four districts in a decade. Whatever else it was, this was not a settled household. | *** [17 Aug 2026] CLOSED THE SAME DAY, BY THE DIAGRAM ITSELF. *** G. 592/1831 for Modderfontein No. 417. It confirms the premise - the farm is titled in 1831 as SITUATED IN THE DISTRICT OF SWELLENDAM, under deed SW.Q.7-23 - and it refutes the identification, because its north-west boundary runs on the ZOETEMELKS RIVIER, which drains to the Goukou and the sea. The 1809 inventory says d'Olifants rivier. NOT THIS FARM. | What survives is the geography lesson, and it is the one the whole Swellendam argument needs: a farm in the modern RIVERSDALE division was SWELLENDAM district ground in 1831, deed prefix and all. The modern division name is not the historical district. | Also kept: every neighbour on that sheet is still "the Land forming the Loan Place" in 1831 - Brakke kuil, Hectorskral, Melkhoutefontyn, Kranzfontyn, Melkhoute Kral, Plattekop - so the loan-farm rent rolls run back from 1831 and would name holders. If the family's Modder Fonteyn was a loan place too, that is the series it sits in.

    Ockert Jacobus Oosthuizen (b.1755) · Jacobus Johannes Oosthuizen (b.1784) · Ockert Almero Oosthuizen · Pieter Marthinus Oosthuijzen · Cornelis Andries Oosthuizen · De Modder Fonteyn (aan d'Olifants rivier)

    opened 2026-08-17

57. ✓ Entry 439 - did Ockert's brother Pieter Cornelis marry a Swanepoel? LEAD-0232

    The opgaafrol index (SRC-482) shows entry 439 as an Oosthuizen whose wife is M: SWANEPOEL. The given names read either 'P: Corn:' or 'P: Cornz'. If the first, this is PIETER CORNELIS OOSTHUIZEN (PER-000242, b.1767, d.25 Dec 1836), a son of Jacobus and a BROTHER of the Ockert who died at de Modder Fonteyn. Who is M. Swanepoel?

    What we expected: PER-000242 HAS NO MARRIAGE IN THIS MODEL, so there is nothing here to contradict and nothing to confirm it either. | IT WOULD BE A SECOND, INDEPENDENT OOSTHUIZEN-SWANEPOEL TIE, and that is why it is worth a row. The one this project already documents is AFFINAL and runs through the Venters: Pieter Johannes Swanepoel of Vlakfontein (PER-000083) married Adriana Cecilia Venter's niece, and 9a admits the eight Swanepoel parcels as a block on the strength of it. A direct marriage between an Oosthuizen brother and a Swanepoel woman would be a CONSANGUINEOUS route into the same family a generation earlier, and the Swanepoel pages currently say in terms that the tie is affinal and not consanguineous. If this holds, those pages need changing. | THE READING IS THE WHOLE QUESTION. 'P: Corn:' is a pair of given names and fits PER-000242 exactly. 'P: Cornz' is a given name and a patronymic and means a son of a Cornelis, which PER-000242 is not - he is a Jacobszoon. One flourish decides it. 5a. | No Swanepoel woman in this model fits: the candidates born in the right decades carry no dates at all, and Maria Sibilla Swanepoel (PER-001001, 1766-1843) is the only close-dated M.

    What would settle it: The qualifier on entry 439 read at a legible scale, plus a marriage record naming both parties.

    What settled it: | [23 Sep 2026, SRC-690] ANSWERED FROM THE ROLL ITSELF, read on screen. Entry 439 is 'P: Corn: Jun:' - Pieter Cornelis JUNIOR - and FamilySearch LH2T-JRD and the private tree have him: born 18 Feb 1788, son of Pieter Cornelis (PER-000242) and Petronella Hermina Kok, married Magdalena Susanna Swanepoel at Swellendam in October 1807. So Ockert's brother did not marry a Swanepoel; his son did, and the roll of 1808 catches the couple in their first year. Now PER-002084 x PER-002085, UNI-000368. Whether her father Pieter Swanepoel (x Magdalena Gous) is of the Vlakfontein line stays open on PER-002085; the Swanepoel pages' 'affinal' sentence stands until it is.

    Pieter Cornelis Oosthuizen · Jacobus Oosthuizen (b.1722) · Pieter Johannes Swanepoel (b6c5d7)

    opened 2026-08-18

58. ✓ Which cadastral polygons are Annex Uitkyk, Blaauwpunt, Voorzorg and Spitzkop? LEAD-0241

    SG diagram 2920 of 1911 names four farms bounding Uitkyk No. 12 in the Field Cornetcy of Cango East. All four are in the model as NO-POLYGON. Which modern farms are they?

    What we expected: THE OUDTSHOORN CADASTRE CANNOT BE SEARCHED BY NAME. Its P_FARMNAME column holds a numeric code - Uitkyk No. 12 is 6457 - so the usual route of matching a diagram's name to a polygon is closed for this division. That is worth knowing before anybody tries it again. | ADJACENCY GIVES FOUR CANDIDATES AND DOES NOT ASSIGN THEM. Sharing a boundary with Uitkyk No. 12: farm 11 to the W, farm 24 to the SW, farm 261 to the S, farm 13 to the E. The diagram names its neighbours NE, SE, SW and W-and-SW. | TWO LOOK PLAUSIBLE AND NEITHER IS ESTABLISHED: Spitzkop (W and SW) against farm 11, Voorzorg (SW) against farm 24. | THE DIAGRAM'S NORTH-EAST NEIGHBOUR HAS NO MODERN COUNTERPART AT ALL, which is the most interesting part. 'Annex' names a piece added to a parent farm, so Annex Uitkyk was very likely absorbed back into Uitkyk itself at some point in the 114 years between the survey and this cadastre - which would leave exactly the hole the adjacency search found.

    What would settle it: Any one of the four neighbours' own diagrams, or the 1879 regional sheet read at the Cango East.

    What settled it: Same division and field cornetcy as PAR-000285 Uitkyk No. 12. | *** 20 AUG 2026, TWO OF THE FOUR ANSWERED FROM A SHEET THAT WAS ALREADY ON DISK. *** SG 2440 OB sheet 2 (SRC-493, DOC-0318), the 1879 Plan of the Division of Oudtshoorn, read 20 Aug 2026 after being filed unread on 19 Aug. It had been filed on 19 Aug with the note 'Scanned UPSIDE DOWN; rotate 180 before reading. Not yet transcribed', and Frans had said of it 'the olifants 1879 file is a nice find. when you get to it.' | BLAAUWPUNT IS OUDTSHOORN NO. 13. The plan letters '12 Uitkyk Lot 40' and '13 Blauwpunt Lot 41' side by side, and the cadastre puts 13 seven and a half kilometres EAST of 12, which is the direction SG 2920/1911 gives for it. Pinned, 2,161.5 ha. | VOORZORG IS ON THE PLAN AND CARRIES NO NUMBER, while every neighbour does - so it was vacant ground in 1879, which is why the 1911 plan of the VACANT FARMS went out at all. Still unnumbered here. | ANNEX UITKYK AND SPITZKOP ARE ON NEITHER PLAN under those names. An annexe postdating an 1879 plan is unremarkable; Spitzkop is the odd one, because SRC-501 lists ANNEX-SPITZKOP as No. 10 and something must have been annexed to. | AND ONE NEGATIVE WORTH RECORDING: no MODDERFONTEIN appears anywhere in the F.C. Lower Oliphants River on this plan. That does not settle LEAD-0220 - the 1832 Uniondale diagram puts the 1809-era ground in the BOVEN Olifants field-cornetcy, which is a different unit - but it is one more place the name is not. | *** CLOSED 20 AUG 2026. THREE OF THE FOUR ARE NUMBERED AND PINNED, off the 1879 plan. *** BLAAUWPUNT = OUDTSHOORN NO. 13. VOORZORG = NO. 24. SPITZKOP = NO. 10. All three lettered with their numbers on SG 2440 OB sheet 2 and drawn on the map. | ANNEX UITKYK IS NOT ON EITHER PLAN under that name, and that is the residue. An annexe is ground added later, so its absence from an 1879 plan is ordinary; SG 2920/1911 names it as Uitkyk No. 12's north-eastern neighbour and that is still the only mention. Whoever picks it up should look at Oudtshoorn farm numbers adjoining 12 on its north-east. | A CORRECTION INSIDE THIS LEAD, MADE THE SAME HOUR: an earlier note here said Voorzorg carried no number on the plan and was therefore vacant ground in 1879. THAT WAS A CROP ARTEFACT - the frame cut the number away - and a wider tile shows '24 Voorzorg' plainly. The inference built on it was tidy and wrong.

    Uitkyk No. 12 (Oudtshoorn) · Annex Uitkyk (Oudtshoorn, Cango East) · Blaauwpunt No. 13 (Oudtshoorn) · Voorzorg No. 24 (Oudtshoorn) · Spitzkop No. 10 (Oudtshoorn)

    opened 2026-08-19

59. ✓ Who were Cecilia Maria du Plessis's parents - and was Rondabel her own family's ground first? LEAD-0267

    PER-001346, buried at Rondabel in 1895, has NO parents in this model. Two du Plessis holdings of Rondabel sit fifty years apart: Ignatius Michael du Plessis (PER-000050, d.1818) held it as a LOAN FARM before 1818 with Jan de Bruin (OWN-000025), and Cecilia Maria du Plessis married into the Theron family that was co-granted it in 1868. Is there any descent between them, or is it a coincidence of one of the commonest surnames at the Cape?

    What we expected: Most likely a coincidence. Ignatius Michael left the farm by 1823 and the model records no descent from him at all; du Plessis is everywhere in these registers. But her parents are unrecorded, so the question has never actually been asked.

    What would settle it: Her father's name, and whether he descends from Ignatius Michael du Plessis.

    What settled it: Opened 22 Aug 2026 out of the identification of GRV-1457. The 1895 death notice is the cheapest route and would close her parentage, the ambiguous death day on the stone (10 or 19 April) and possibly her place of death in one document. | *** [22 Aug 2026] THE DOCUMENT IS LOCATED. KAB, MOOC 6/9/340, REFERENCE 1007 - 'THERON, CECILIA MARIA. NEE DU PLESSIS. DEATH NOTICE.', 1895. SRC-542. *** Found the day after this lead named it, by the farm-name-not-surname method of LEAD-0264 applied to a person: query KAB on (THERON) AND (CECILIA). | The lead is NOT closed - an index entry is not a document. It stays open until somebody reads the notice at KAB or has a copy made. What it should then give: her PARENTS, which is this lead's question; the surviving spouse and children, which test the ten parentage rows already held; the place and date of death, which would say whether she died at Rondabel; and the DEATH DAY, which the headstone leaves ambiguous between 10 and 19 April 1895. | *** [22 Aug 2026] THE DOCUMENT ARRIVED AND IT DOES NOT ANSWER THIS LEAD'S QUESTION. *** DOC-0333 is a 'Form of Information of a Death' under Act No. 7 of 1894 - a CIVIL REGISTRATION form, and it HAS NO PARENTS FIELD. It asks residence, age, race, birthplace, conjugal condition, issue living and deceased, the spouse's age, occupation, date and place of death, place of burial, cause, the doctor and the informant. Her birthplace is given only as 'CAPE COLONY'. So her parentage is exactly as open as it was, and the du Plessis loan-farm question with it. | WHAT IT DID SETTLE, and it is a great deal: the death day, the residence, the burial, the cause, and nine children's ages. Recorded on PER-001346 and GRV-1457. | WHERE TO LOOK NEXT: this sheet is ONE item in MOOC 6/9/340; a Cape estate file usually holds more - a liquidation and distribution account naming heirs, and sometimes the older-style death notice that DOES ask for parents. Order the file rather than the sheet. Her marriage record to Petrus Jacobus Theron, c.1876, would name her father directly, and FamilySearch GHW4-KDS may already hold it. | *** [23 Aug 2026] ANSWERED. HER PARENTS ARE JOHANNES GYSBERTUS DU PLESSIS (1826-1913) AND JOHANNA ELIZABETH VAN ROOYEN (1831-1913). *** From a FamilySearch fan chart supplied by Frans, which carries her ancestry four generations back; 24 rows added, PER-001879 to PER-001902. | THE VERNOEMINGSPATROON CONFIRMS IT WITHOUT BEING ASKED: her son Johannes Gysbert Theron (PER-001169, 1883-1884) carries her father's name. | The lead's second question - whether the du Plessis who held a Rondabel before 1818 is any relation - is NOT answered and moves to LEAD-0272, which now has something much better to work with.

    Rondabel

    opened 2026-08-22

60. ✓ Martha du Plessis (1730-1796) - who were her parents? LEAD-0273

    She married Petrus Venter on 14 October 1753 and their son Jan Louis Venter fathered the sixteen Venter children of STO-0046, three of whom married Oosthuizens and one of whom held Rondabel by 1829. She has no parents recorded here.

    What we expected: A Cape du Plessis household of the 1720s; the surname arrived with the Huguenots in 1688 and had spread widely by then.

    What would settle it: Her father's name.

    What settled it: Split out of LEAD-0272 when that lead closed negative on 23 Aug 2026. | *** [23 Aug 2026] ANSWERED WITHIN THE HOUR OF BEING OPENED. *** Martha du Plessis's parents are JEAN LOUIS DU PLESSIS (1691-1768) and MAGDALENA ROUSSEAU (1695-1780), from a FamilySearch fan chart of her son Jan Louis Venter. Her father's parents are Jean Prieur du Plessis and Marie Madeleine Menanteau, which is what makes her Daniel du Plessis's first cousin - see LEAD-0272. | ONE DISCREPANCY LEFT OPEN: the chart gives her death as 1789 and this model has carried 1796. Not changed on a fan chart alone. [TO VERIFY]

    Rondabel

    opened 2026-08-23

61. ✓ Anna Magdalena Petronella Strydom in two registers, where the model has Johanna Machteld LEAD-0282

    Two baptism registers - Stellenbosch 1812 and onder Traca 1818 - name the woman standing with Cornelis Andries Oosthuizen as ANNA MAGDALENA Petronella Strydom. His wife in this model is PER-000138, JOHANNA MACHTELD Petronella Strydom. Same woman, or two?

    What we expected: Probably the same woman, but this is not a spelling variant of the kind 5b covers - Anna Magdalena is not an everyday form of Johanna Machteld.

    What would settle it: A register naming her at her own marriage or death.

    What settled it: WIT-0344. She is Frans's 4x-great-grandmother, so this is a direct-ancestor row, which is why it is P2 and not P3. | *** [23 Aug 2026] ANSWERED, BY HER OWN BAPTISM. *** SRC-554, supplied by Frans within the hour: Graaff-Reinet, 13 January 1793, "'t kind JOHANNA MAGTELT PETERNELLA, ouders Daniel Jacobus Strydom, Anna Magdalena Carelse". SHE IS JOHANNA MACHTELD PETRONELLA and the model's doopnaam was right. | AND THE CAUSE OF THE SLIP IS NOW VISIBLE RATHER THAN GUESSED: her elder sister is ANNA MAGDALENA STRYDOM (PER-000209), and at the 1818 onder Traca baptism that sister is THE MOTHER OF THE CHILD, written two lines above the witness list. Two Strydom sisters on one entry, one of them Anna Magdalena; a clerk carrying the name down is ordinary. The same shape holds at Stellenbosch in 1812. | THE SISTERS MARRIED BROTHERS - Anna Magdalena to Jacobus Johannes, Johanna Machteld Petronella to Cornelis Andries - which is why the two couples stand at each other's fonts six years and 500 km apart. | The baptism record raises a NEW question about the mother's surname, which is LEAD-0283 and not this lead.

    opened 2026-08-23

62. ✓ Which Jan Oosthuysen held Welbedagt to April 1778 - Johannes (b. about 1736), most likely, or his father Jan Hendrik (b. 1714)? LEAD-0362

    Appel (SRC-665, citing RLR 18 p. 19) has 'Jan Oosthuysen (jr.)' on Webedagt aan die Rietfontein for almost 25 years to April 1778, too poor to pay 13 years 11 months of arrears. Is 'jr.' the register's word or Appel's, and is the farm Welbedagt No. 150 (PAR-000277)?

    What we expected: [FAMILY] Johannes (b.. about 1736), farming with his brother Marthinus, who takes the place over in 1782 when Johannes dies.

    What would settle it: The register's own entry, or a roll on which one of the two Jans still heads a household of substance after April 1778.

    What settled it: [22 Sep 2026] Opened from the passage | [22 Sep 2026, FRANS] TWO READINGS, AND HIS IS THE LIKELIER. *** (1) FRANS'S READING, MOST LIKELY: 'Jan (jr.)' is JOHANNES OOSTHUISEN (PER-000739, LVZR-B4P, b. about 1736), who farmed Welbedagt with or beside his brother MARTHINUS (PER-001851). THE ARITHMETIC CARRIES IT: Marthinus's estate of 21 September 1792 owes 10 years 6 months of arrear recognitiepenninge, so his holding begins about MARCH 1782, and Johannes died in 1782 (FamilySearch LVZR-B4P). The farm passes in the year the brother dies, which is a family hand-over rather than a new tenant, and it leaves no gap between Appel's April 1778 and 1782. (2) THE OTHER READING: 'jr.' is their father JAN HENDRIK (PER-001828, b. 1714), the only Jan on the 1751 Swellendam roll and 'de oude' from 1760; a start about 1753 suits his age, and his widow's inventory of March 1780 (SRC-432) names no farm. It leaves 1778-1782 unaccounted for. AGAINST BOTH, ON ONE POINT: Appel's 'feitlik 'n kwarteeu' puts the start about 1753, when Johannes was 17. The register may well have opened under the father and carried the son, with the arrears running from about 1764. AND A NAME THAT CANNOT BE USED EITHER WAY: the inventory taken on the farm on 18 May 1792 (MOOC8/50.101b) names 'Jan Ooshuysen Jan zoon' among the three appraisers, but the sheet is SIGNED by Johannes Jacobus Oosthuysen. Johannes died in 1782, so that line is the clerk writing loosely, or a third Oosthuysen. WHAT SETTLES IT: RLR 18 p. 19 - whether the register says 'de jonge' or 'Jan zoon', and whose name stands on the entry after April 1778. | [22 Sep 2026, Appel's 'Jan Oosthuysen (jr.)' of Welbedagt, ruined in April 1778, is JOHANNES OOSTHUISEN (PER-000739, LVZR-B4P, b. about 1736), 'Jan d'jonge' of the Swellendam rolls, not his father Jan Hendrik (PER-001828, 'Jan d'oude'). THE CHAIN: married 1759, first child 1760; the arrears run from about 1764 (13 years 11 months before April 1778); payment suspended April 1778 rather than the place taken; he dies in 1782, and his brother MARTHINUS's own arrears begin about March 1782, so the farm passes inside the family in the year of the death. The father cannot be the man of 1778 and go on holding afterwards: he was dead by 31 March 1780 (SRC-432), and his widow's estate names no farm. STILL TO CONFIRM, and it is a confirmation rather than an open question: RLR 18 p. 19, whose name the register carries after April 1778. AND NOTE AN ABSENCE: the published MOOC8 holds no estate for a Jan or Johannes Oosthuizen around 1782, which fits a man who died with nothing to divide but proves nothing on its own.

    Welbedagt No. 150 (Attaquaskloof, Oudtshoorn)

    opened 2026-09-22

63. ✓ Seven people carry a hand-entered FamilySearch id that the family-tree disagrees with LEAD-0019

    Backfilling familysearch_id from the family-tree's _FID tag agreed with the hand-entered value on 103 of 110 people and disagreed on 7. Which id is current for each - or do both resolve to the same person after a merge?

    What we expected: Most likely FamilySearch merges: when two profiles are merged the losing id redirects to the survivor, so both ids can be 'right' and resolve to one person. Three of the seven family-tree ids share a GHF6- prefix (GHF6-4KQ, GHF6-HMP, GHF6-4KH), which is the signature of a duplicate cluster rather than seven independent errors. The alternative - that a person was matched to the wrong profile - is the one that would matter, and it cannot be ruled out from here.

    What would settle it: A redirect proves a merge and the survivor id is the one to keep. Two live, distinct profiles means one of them is the wrong person, and the vitals decide which.

    What settled it: Low priority - the site links to whichever id is recorded and a merged id still resolves. It becomes urgent only if one turns out to be a wrong-person match. | [CLOSED 9 Aug 2026] DUPLICATE OF LEAD-0036, which asks the same question over a superset of the same people. This lead's seven (PER-000056, 000134, 000158, 000383, 000443, 000455, 000470) are all nine of LEAD-0036's less PER-000211 and PER-000246. Same test, same fix, one opened hours after the other on 31 Jul 2026. Work it on LEAD-0036.

    Gerolm Samuel Marincowitz · Hendrik Johannes Oosthuizen · Roelof Daniel Jacobus Oosthuysen · Ockert Almero Oosthuizen · Adriana Cecilia Oosthuizen · John Peter Oosthuizen

    opened 2026-07-31

64. ✓ Who was 'The Ape', and who ended up with it? LEAD-0026

    Both the Wilgemond and Klein Valie surveys of 1833 annex a piece of Government ground called THE APE, and Wilgemond is bounded on the north by it. What is it, where is it, and is the name Die Aap?

    What we expected: A named piece of crown land on the Zwartberg foot between the two farms, probably 'Die Aap' - Karoo features called Aap are not uncommon. It was annexed to BOTH surveys, which suggests it was split between them rather than granted whole.

    What would settle it: Any map or deed naming it.

    What settled it: [17 Sep 2026, lead review] The esio farm-name gazetteer has no APE, THE APE, AAP or DIE AAP in Prince Albert or Beaufort West. The annexed Government ground was never registered under that name. | [17 Sep 2026, CORRECTION AND CLOSE] The note above ('never registered under that name') misses what the model already held. THE APE IS DE AAP: the Government ground annexed to the loan farm Klein Valie in Stretch's survey of 1833 and granted with it on Bf.W.Q.5-6 (PAR-000088, DOC-0119). The 1913 Venters Kloof Annex diagram (DOC-0133) draws the unit on its southern boundary as 'KLEINE VALLEY EN DE AAP', pencilled B.Q.5.6, and Rose Willis (SRC-010) has Gideons Hoop, Klein Valie's portion 182/4, as 'today the De Aap guest farm'. So the name was not lost: it rode with Klein Valie's title and survives on the ground. Why it was called The Ape is not recorded.

    Wilgemond No. 179 · Klein Valie No. 182

    opened 2026-07-31

65. ✓ Are A.J. de Villiers and A.I. de Villiers one surveyor or two? LEAD-0081

    Do the thirteen 1908-1924 diagrams signed "A.J. de Villiers" and "A.I. de Villiers" carry one man's signature, or two?

    What we expected: ONE MAN, on three grounds, none of them conclusive. The years interleave rather than separate - A.I. in 1908 and 1911, A.J. in 1910, 1916, 1923 and 1924 - so there is no clean handover of the kind a father-to-son succession usually leaves. Both work the same district. And I and J are barely distinguishable in the copperplate and early typescript of this period, which is a transcription risk on our side as much as a signature question on theirs.

    What would settle it: A full given name on any one of the thirteen sheets, or the SG admission register showing whether one or two de Villiers held a licence in the Cape in 1908-1924.

    What settled it: Opened 4 Aug 2026 from the new checks.py surveyor-spelling check, which also flagged J. Loubser/J.A. Loubser and van den Heever with and without his PLS number - both almost certainly one man each, and both harmless until something counts per-surveyor. | [22 Sep 2026] VERY PROBABLY ONE MAN, and the documents' 'A.I.' is a reading of 'A.J.'. Three things: (1) the same farm in the same months - 'A.J. de Villiers' surveys Klipgat No. 114 in September 1910 (DOC-0158) and 'A.I. de Villiers' its three portions in December 1911 (DOC-0193 to 0195), both 'Government Land Surveyor'; (2) the 'A.I.' sheets we hold are COPIES - DOC-0193 is stamped as copied in 1950, and the name on it is the copyist's transcription of a signature, '(Sgd) A.I de Villiers'; (3) in the hands of the period a capital I and J are made the same way, and Dutch practice treated them as one letter. The documents keep what each sheet says (documents.surveyor is a transcription and is not normalised). NOT FOUND: his full name. S2A3's de Villiers surveyor is Josias Eduard (1843-1898), not him. The Cape roll of admitted land surveyors would settle it [TO VERIFY].

    opened 2026-08-04

66. ✓ What should farm 12 be called on this site - Riet Fontein or Minnies Kraal? LEAD-0085

    The farm's own 1839 and 1868 diagrams title it RIET FONTEIN No 12; the modern cadastre and local usage call it MINNIES KRAAL. Which name should the map and the pages carry?

    What we expected: A RENAME WITH A DATE SOMEWHERE IN BETWEEN, which nobody here has yet found. The identification of the two as one farm is not in doubt - the 1868 diagram gives 15 167 morgen 520 sq roods = 12 991,0 ha against the cadastre's 12 957,1 ha for farm 12, 0,26% apart, and the 1839 diagram's boundary list names Antjes fontein and Riet Port, which are exactly the farms the polygon touches. What is missing is WHEN and BY WHAT INSTRUMENT the name changed.

    What would settle it: A dated document carrying the name Minnies Kraal for farm 12, or a deed recording the change.

    What settled it: Applies to the three portions too - PAR-000158/159/160 are currently named Minnies Kraal Remainder, Portion 1 and Portion 2, following the cadastre. 5 Aug 2026: the 1990 Farm 215 consolidation sheet (DOC-0238) names farm 12 RIETFONTEIN, twice, on its neighbour ring. That is a Surveyor-General sheet using the name 150 years after the 1839 diagram did, which narrows the question from 'which name is right' to 'when did local usage diverge from the sheets'. | CLOSED 10 Aug 2026. RIETFONTEIN, on the farm's own diagrams. SG 5774/58 of May 1958 [SRC-350] titles it 'Gedeelte 1 van die plaas RIETFONTEIN Nr 12' and SG 8388/1994 [SRC-352/353] titles it 'PORTION 2 OF THE FARM RIETFONTEIN NO. 12'. Two surveyors, thirty-six years apart, one name, and neither sheet says Minnies Kraal. The four farm-12 rows are renamed. The cadastre's MINNIES KRAAL for farm 12 is now itself the open question, LEAD-0147.

    Rietfontein No. 12 · Rietfontein No. 12 Remainder · Rietfontein No. 12 Ptn 1 · Rietfontein No. 12 Ptn 2

    opened 2026-08-04

67. ✓ A third of the marriages in this model are within one surname - but the sample chose itself LEAD-0088

    Is the endogamy real, or an artefact of collecting exactly the families that intermarried?

    What we expected: BOTH, and the interesting part is the second. Of 39 marriages where BOTH surnames fall in this project's core set, 13 - a third - are within one surname: Oosthuizen x Oosthuizen 4, Venter x Venter 4, Swanepoel x Swanepoel 2. The cross-weaving is as heavy: Oosthuizen x Venter 4, Marincowitz x Oosthuizen 3, Oosthuizen x Strydom 3, Oosthuizen x van Zyl 3. THE NUMBER MUST NOT BE PUBLISHED AS A DISTRICT FACT. The denominator is marriages where both partners already carry a surname this project collects - so it measures how tightly the web we chose is woven, not how often Karoo families married their own. That is the same selection artefact that killed the "twin peaks of land transfer" idea on 4 Aug: a histogram of what we hold, presented as a histogram of what happened.

    What would settle it: The same-surname rate across ALL marriages in the register for a decade, against ours. If the district rate is much lower, the endogamy is a real family trait.

    What settled it: [22 Sep 2026] ANSWERED: NOT A FAMILY TRAIT. The denominator this lead asked for is already transcribed: the Beaufort West baptism register 1819-1844 (SRC-200, data/transcriptions) holds 947 distinct couples, and 80 of them (8.4%) share a surname once particles and spelling are folded (Loots, van der Westhuizen, Kruger, Botes and Jacobs lead; Oosthuizen has 3). The model's own unions now run 22 of 346 (6.4%), not a third: the early figure came from a small sample that had collected the intermarried families first. The model figure is if anything high, because a few women are recorded under a married surname. The district rate and ours are the same order, so marrying a same-surname cousin was ordinary here, not a mark of these families. Method: first particle through the end of the name as the surname; ij/y and uis/uys/uiz folded.

    opened 2026-08-04

68. ✓ Okkert Jacobus Oosthuizen, dead in 1842 on Uitnood in the Swartberge LEAD-0134

    A stone on Uitnood 147, 23 km from Klaarstroom, reads OOSTHUISEN Okkert Jacobus and gives 1842 and no birth year. Who was he, and what was an Oosthuizen doing on that ground sixteen years before Ockert Almero and Adriana Cecilia died?

    What we expected: A brother or a cousin of Ockert Almero Oosthuizen. The name Ockert Jacobus recurs hard in this family - the model holds eleven of them - so he is very likely of the same stock.

    What would settle it: A death notice of 1842, which would give his age, his parents and his estate.

    What settled it: Raised 7 Aug 2026 from the Willowmore eGGSA sweep, run by radius from Klaarstroom. | [7 Aug 2026] CLOSED THE SAME DAY. The rest of the stone reads: died 9 October 1842, aged 2 years and 10 months. He is Ockert Jacobus Oosthuizen [PER-000848], born 28 November 1839, a son of Gert Adriaan Oosthuizen - a person already in this project whose death date was blank. The lead assumed an adult of Ockert Almero and Adriana Cecilia's generation and was wrong about that; the age on the stone is what settled it. WHAT REMAINS is the household: a child buried on Uitnood in 1842 means Gert Adriaan Oosthuizen was on or near that ground, 23 km from Klaarstroom, seventeen years before Ockert Almero and Adriana Cecilia died. That is worth its own question.

    opened 2026-08-07

69. ✓ Is Jacobus Johannes Gideon Nel the father of Christina Elizabeth Nel? LEAD-0159

    He stands at the 1887 baptism of her first recorded child, and her son born the next year is christened Jacobus Johannes Gideon. Is he her father?

    What we expected: VERY LIKELY, ON THE NAMING ORDER AND NOTHING ELSE. The Cape pattern gives the second son the mother's father's name, the witness is present at the elder sibling's font, and the family was of Brandvlei where she was living when she married in 1885. That is a pattern, not a parentage, and this record does not join people on a name.

    What would settle it: Her baptism entry, naming her parents.

    What settled it: Opened 11 Aug 2026 from the six register photographs | CLOSED 11 Aug 2026, ANSWERED YES, AND WITH MORE THAN WAS ASKED. The Calvinia baptism index for 1867 (SRC-399) and her 1895 death notice (SRC-400) both name her parents as JACOBUS JOHANNES GIDEON NEL and ANNA JACOBA VAN WYK. The lead reasoned from the Cape naming order alone and got it right; what it did not see is that BOTH of the 1887 witnesses were her parents, so two of the four names at that font are the child's maternal grandparents. Both are now rows here.

    opened 2026-08-11

70. ✓ Did Anna Maria Visagie belong to this Schalk Willem Burger van der Merwe at all? LEAD-0160

    She entered this model on Janet Melville's reading of the 1863 Calvinia baptism register. That register has now been read directly and does not mention her. Was she ever his wife?

    What we expected: SHE PROBABLY BELONGS TO A DIFFERENT MAN OF THE SAME NAME. Two things point that way. The reading that produced her is demonstrably wrong on the same page. And this Schalk Willem Burger was born on 23 February 1833, so a wife who died in 1853 would have married him before he was twenty-one. Melville's report is keyed to Van der Merwe [70897], a number this project cannot resolve, and the name recurs in the family. Against that, she gave a death year, which suggests she had a death notice in front of her for somebody.

    What would settle it: A death notice for Anna Maria Visagie naming her husband.

    What settled it: Opened 11 Aug 2026 from the six register photographs | CLOSED 11 Aug 2026, ANSWERED YES, and this lead was wrong. The Hantam marriage register (SRC-383) has the marriage on 13 January 1851 with his own father consenting by name. The argument this lead made - that a wife dead in 1853 would have meant marrying before he was twenty-one, and that she therefore probably belonged to a different man of the same name - was sound reasoning from the wrong premise. He married at seventeen. Worth keeping visible: the lead was opened the same morning the register arrived, and the register was three hours away.

    opened 2026-08-11

71. ✓ Two Visagie households, each with a van der Merwe mother, marry into this line LEAD-0161

    Izaak Heremias Visagie married Johanna Sophia van der Merwe in 1812 and Hendrik Johannes Visagie married Louisa Jacoba van der Merwe about 1828. One supplied a first wife to this line in 1851 and both supplied a witness at a font in 1863. How are the two Visagie men related, and where do their van der Merwe wives sit in this pedigree?

    What we expected: THE TWO VISAGIE MEN ARE VERY LIKELY BROTHERS - born 1792 and 1801, both in the same district, both marrying van der Merwe women within about sixteen years of each other, and their children christened across the same small set of names. The van der Merwe wives are the more useful question: if either is a daughter of Carel Aaron van der Merwe (PER-001639), then Schalk Willem Burger's first marriage in 1851 was to a first cousin, which is exactly the pattern his SECOND marriage follows. Nothing here tests that and it is not asserted.

    What would settle it: Parentages for Johanna Sophia van der Merwe 1796-1872 and Louisa Jacoba van der Merwe 1802-1845.

    What settled it: Opened 11 Aug 2026 from the FamilySearch pages | CLOSED 11 Aug 2026, THE SAME DAY IT WAS OPENED, and the prediction it made was right. It guessed that if either Visagie wife were a daughter of Carel Aaron van der Merwe then the 1851 marriage was a cousin marriage. BOTH ARE HIS DAUGHTERS. Johanna Sophia van der Merwe 1796-1872 (LWFG-THQ) and Louisa Jacoba van der Merwe 1802-1845 (LZ2Z-2XD) appear in Carel Aaron's own child list on FamilySearch under the very ids recorded from the Visagie side hours earlier (SRC-386, SRC-388). THAT IS AN IDENTIFICATION BY ID, NOT BY NAME, which is the only kind this project accepts here. So the two Visagie men married two sisters, and every Visagie standing beside this family - the first wife in 1851, both witnesses in 1863 - is a descendant of Carel Aaron coming back round. The remaining half of the question, how Izaak Heremias and Hendrik Johannes Visagie are related to each other, is NOT answered and is not pursued: under 9a that is their community, not this line's ancestry.

    opened 2026-08-11

72. ✓ Was Oupa Walters kin to Paul Kruger, or to Jan Smuts? LEAD-0164

    The 1987 memoir says Jacobus Abraham Walters (1877-1962) was a cousin of President Paul Kruger. Frans's own recollection is that the connection was to Jan Smuts. Which, and how?

    What we expected: BOTH SURNAMES ARE ALREADY IN HIS HOUSEHOLD AND THEY POINT DIFFERENT WAYS. His wife is MAGDALENA MARIA KRUGER, so a Kruger link runs through her and he is kin by marriage - which fits a memory that has drifted one step. His son born 1911 is JACOBUS ABRAHAM SMUTS WALTERS, and in this family a surname used as a given name means a surname in the ancestry, not a hero. So the memoir has the right family and possibly the wrong president, and the Smuts name has the better claim to be genealogical. GENERATION FAVOURS SMUTS TOO: a man born 1877 is a natural cousin of Smuts, born 1870, and an awkward one of Kruger, born 1825.

    What would settle it: A documented descent joining either family, or the missing first page of the memoir.

    What settled it: Opened 11 Aug 2026 from the 1987 memoir. | CLOSED 11 Aug 2026 by the missing first page of the typescript (SRC-395), which the memoir itself had pointed at. THE ANSWER IS BOTH, AND NEITHER IN THE WAY THE MEMOIR PUT IT. The KRUGER is Oupa Walters's WIFE's maiden name - "die Carnarvon Kruger's" - so a Kruger kinship runs through her and he is kin by marriage, which is what this project deduced from the model before the page arrived. The SMUTS is real too and older: her mother was born Smuts. But the Smuts in Jacobus Abraham SMUTS Walters's name is neither of those - GENERAL JAN SMUTS STOOD GODFATHER to him in 1911, after his parents had promised the unborn child to Smuts because they could not afford another. Frans's instinct that Smuts was the real connection was right; the mechanism is a godparent, not a cousin. What is still undocumented is whether either presidential kinship is genealogical at all. LEAD-0166.

    opened 2026-08-11

73. ✓ A leak test fired on my own prose, not on a broken gate LEAD-0177

    Why did the public build report blocked media as named in data.json when the export drops those rows on rights?

    What settled it: TWO RULES COME OUT OF THIS AND BOTH ARE CHEAP. Describe a withheld item in WORDS, never by id, in anything the public build emits - an id is a handle, and a published page naming the id of a withheld photograph tells a reader exactly what to ask for. And CHECK THE IMG STEM IS FREE before filing: media ids and image stems are numbered independently here and they had drifted apart, so the next free MED number was not the next free IMG number. The real cost was a wrong finding written down as a fact, which is the 5a note 5 mistake - an inference stated as a finding.

    opened 2026-08-11

74. ✓ Where does "van Helslandt" come from? - ANSWERED: a childless uncle by marriage, 1800 LEAD-0228

    The line reads, as far as it can be read from the crop: '191 | Helsland | Charles van | S:t Dumen[ij]'. Two names this project carries appear beside each other a century before it holds either at the Cape.

    What we expected: *** THE DATES RULE THE DUMINYS OUT AND LEAVE HELSLAND WIDE OPEN. *** | ON THE DUMINY SIDE, NO. The earliest Duminy here is ANTOINE LEBRE DUMINY, born 2 July 1691, who would be seventeen in 1708 - so the arithmetic alone does not exclude him. What excludes him is where the family was: his son FRANCOIS RENIER was born A FOUNDLING AT LORIENT IN BRITTANY on 4 October 1747 and is the one who came to the Cape, as S039 says. The Duminys were in France in 1708. A Dumeny in a Cape register of that year is not this family on anything this project holds. | ON THE HELSLAND SIDE, THIS IS THE BEST LEAD THERE HAS BEEN. 'van Helslandt' exists in this model ONLY as a middle name, ONLY on Duminys, and ONLY from 22 August 1818, when FRANCOIS JEAN VAN HELSLANDT DUMINY (PER-000567) is born. It then repeats for three generations - 1853, 1884, and again. THERE IS NO VAN HELSLAND SURNAME ANYWHERE IN THIS PROJECT. | AND IT DID NOT COME FROM HIS MOTHER: she is CATHARINA MARIA STADLER (PER-001479, b.1796). A middle name appearing from nowhere and then repeating for three generations usually comes from a godparent, a benefactor or a marriage a generation or two back, and a Van Helsland family at the Cape in 1708 is exactly the kind of source that would explain it. | THAT IS A HYPOTHESIS ABOUT A NAME, NOT A DESCENT, and the 110-year gap is the whole problem with it.

    What would settle it: A van Helsland named as a witness at the 1818 baptism, or as a spouse anywhere in the two preceding generations. Failing that, the name stays unexplained and the 1708 entry stays a coincidence of two surnames on one line.

    What settled it: S039 is the story this touches, and it currently says nothing before Antoine Lebre Duminy of 1691. | No source row is created for the register: what arrived is one cropped line with no title, no folio and no repository visible, and 5a-ii says the reading and the citation are two claims. Record it properly when the page is seen. | CLOSED 18 Aug 2026, the day it opened, by SRC-071 - a 21-page Duminy genealogy that has been in sources/genealogy/ since 29 July. It reads: 'Jeanne Francoise (Sanet) b. 17.12.1779 ... m. 26.4.1800 Charles Jean VAN HELSLANDT from Thourout (Torhout) in Vlaandere (b. 24.6.1764 d. 5.1827), captain of the Danish vessel Forsoget. No offspring.' She is PER-000815, sister of Benjamin Jacob Wilhelm Duminy, whose son born 1818 was named for her husband. Charles Jean van Helslandt is now PER-001860 and the marriage is UNI-000335. | THE DUMINY HALF OF THE PREDICTION STANDS: the family was in Brittany until Francois Renier first came to the Cape in 1771, so no Duminy is in a Cape register of 1708. | AND SO DOES THE HELSLAND HALF, differently from how it was predicted: the guess was a Cape Van Helsland family present from 1708 whose name reached the Duminys by an unfound marriage. The marriage is found and it is 1800, not the 1700s, and the man is a Flemish sea captain rather than a Cape householder. Charles Jean was thirty-five when he married and was born in 1764 - so if a Helsland does stand in a register of 1708 it is not him, and this project has nothing tying the two together. 9: never match on a name alone.

    Francois Jean van Helslandt Duminy · Antoine Lebre Duminy · François Renier Duminy

    opened 2026-08-18

75. ✓ Mev. Oosthuizen, killed September 1793 - CLOSED, the record is the north-west Cape LEAD-0180

    The Council of Policy record of 27 September 1793 names an OOSTHUIZEN woman, given only as 'Mev.', among three people killed in the Swellendam district. Who was she, and was she on this project's ground?

    What we expected: Probably not resolvable from the summary alone. 'Mev.' with no forename is how the clerk wrote it, and the depositions themselves - which would name her - are the part not transcribed here.

    What would settle it: A forename in the depositions, or a Swellendam estate record of 1793-94 for an Oosthuizen woman. Either would also settle the place question, which the summary cannot.

    What settled it: CLOSED THE SAME DAY IT WAS OPENED, on the Dutch transcription. Answering the question the lead asked - was she on this project's ground - NO, and the candidate work below is void. The killing was 'in het Gebergte tusschen Namaqualand en de Oliphantsrivier', with Picqueniersklooff named alongside: the north-west Cape, not the Koup. Swellendam is where the report was WRITTEN, not where the people died. And she is 'een Huisvrouw van Jemand genaamd Oosthuizen' - the record never held a forename, so no source could have supplied one. THE THREE TREE CANDIDATES ARE WITHDRAWN, not merely unproven. Johanna Jonker, Hester Gouws and Johanna Petronella Oosthuizen were ranked mainly on Swellendam and on death-date bands around 1793, and the district was the artefact. Johanna Jonker's 'died before 1794 at Swellendam' is now evidence of nothing in particular. Do not resurrect this matching. WHAT WENT WRONG IS WORTH MORE THAN THE LEAD WAS. Three separate index features each pointed the wrong way and agreed with each other: the locations array carried Coupasberg, Couga and Zwartenberg for an entry whose text contains none of them; the English summary said 'Swellendam district'; and the person index rendered an unnamed woman as 'OOSTHUIZEN, (Mev., Vermoorde)', which reads exactly like a named person. An index is a finding aid. The transcription is the evidence. This project already knew that - 5a note 5, an inference is not a finding - and it still took reading the Dutch to catch it. STATUS: closed. The out-of-scope reason is the whole of this note and the title; the status vocabulary here is open/closed/parked and does not need widening to carry it.

    opened 2026-08-12

76. ✓ LEAD-0324 LEAD-0324

    PIETERSDORP. Frans names De Claar Stroom / Klaarstroom as 'Pietersdorp', and the word appears nowhere in this repo. What ground does it name, and is it the werf?

    What settled it: Raised by Fransin passing, while checking which farms the Oosthuizen brothers held. Searched across parcels, places, events, prose and leads: NOT ONE OCCURRENCE. A name a family uses that the cadastre never recorded is exactly what 5-second-rule is about, so this is a research lead and not a doubt. *** AND IT MAY BEAR DIRECTLY ON LEAD-0215. *** That lead proposes that the middle farm of the three, De Claar Stroom, was PETRUS OOSTHUIZEN's (PER-000505), between Okert on Middelwater and Jacobus Johannes on Vrolykheid. A werf on that farm called PIETERSDORP would be the vernacular trace of exactly that man. THAT IS A HYPOTHESIS AND NOTHING MORE. It is built on one syllable, and 9 forbids matching on a name alone; the same reasoning would make any Pieter in the district a candidate. It is written down so it can be tested, not so it can be believed. What would settle it: Frans's own account of where he heard the name and what it points at; the Klaarstroom cameo (SRC-010); the 1858-60 village survey and erf sales; and the 350/1833 deduction table, which already renumbers a 'Lot a' of nine morgen at the werf. NOTE THE COMPLICATION THE DIAGRAM ITSELF RAISES: SG 350/1833 says De Claar Stroom was 'Surveyed for THE WIDOW DU PRE', not for any Oosthuizen. See LEAD-0215. | *** CLOSED THE SAME DAY BY FRANS: 'ek het bedoel PETERSBURG, wat Klaarstroom aan Petrus Oosthuizen verbind.' *** The name is not a werf name and not a lost one. It is the village laid out on De Claar Stroom after Meiringspoort opened, first sold 4 August 1860, already in this model as EVT-000206 and OWN-000082 from the Klaarstroom cameo (SRC-010). *** WHICH MEANS IT CANNOT RESCUE LEAD-0215. *** The hypothesis was that a vernacular name might put an Oosthuizen on the middle farm BEFORE 1833. Petersburg dates from 1858-60, a quarter-century after Stretch, and is exactly the evidence already counted. The middle farm is still du Preez ground in 1822 and a du Pre widow's in 1833. The lead is kept rather than deleted because the reasoning is the useful part: a name that would have been decisive if it were older, and was not.

77. ✓ Almero Oosthuizen, retired headmaster, born at Droogekloof [IDENTIFIED] LEAD-0346

    A published source names an Almero Oosthuizen born at Droogekloof in the Prince Albert district and living at Durbanville in 1999. Who is he, and is his Droogekloof the werf on Middelwater that Frans's father grew up on?

    What we expected: That it is the same Droogekloof - PLC-000035, the werf inside Middelwater - and that this man is a near relative of Frans's father rather than a stranger who shares two names.

    What would settle it: Frans naming him, or a birth or baptism record giving Droogekloof.

    What settled it: He is an informant in the book, so he was alive in 1999 and may be living. NOTHING ABOUT HIM GOES ON A PAGE until that is settled: 4 fails closed and no person row is created here. *** [7 Sep 2026] CLOSED WITHIN THE HOUR, BY FRANS. *** 'that almero ooshtuizen, was my grandfather, hennie's brotehr' and 'he is my father's uncle'. He is PER-000168, ALMERO JACOBUS HITGE OOSTHUIZEN, 28 June 1912 to 21 January 2001, youngest brother of Hendrik Lodewyk 'Hennie' Malherbe Oosthuizen (PER-000017), Frans's paternal grandfather. THE DATES CONFIRM IT AND EXCLUDE THE OTHER ALMERO. He was alive and retired when the book was published in 1999; his eldest brother Ockert Almero Oosthuizen (PER-000165) died in 1975. He is also dead, which THE LEAD PREDICTED HE MIGHT BE A NEAR RELATIVE RATHER THAN A STRANGER SHARING TWO NAMES. He is the anchor's own great-uncle, and it took one question to Frans rather than any record. WHAT IT LEAVES OPEN, AS TWO SMALLER THINGS: whether he was headmaster of Jan van Riebeeck and whether he lived at Pinelands. Frans recalls both and is explicit that he is not sure. A school's own records would settle the first in a sentence. [TO VERIFY] AND WHAT IT GAVE THAT NOBODY ASKED FOR: DROOGEKLOOF IN PRINT, AS A BIRTHPLACE, IN 1912. See LEAD-0064 and PLC-000035.

    Almero Jacobus Hitge Oosthuizen (b.1912) · Hendrik Lodewyk Malherbe Oosthuizen

    opened 2026-09-07

78. ✓ A duplicated burial in the Botma ground on Vredendal LEAD-0310

    Two grave rows stand for Johanna Maria Magdalena Botma in the Botma burial ground - a hand-made row from a family photograph and a promoted eGGSA caption. Which survives, and how is the merge recorded?

    What we expected: One burial, two rows: a hand-made row from a family photograph and a second promotion of the same eGGSA caption.

    What would settle it: on which id keeps the burial, with the other's external_url folded into it.

    What settled it: CLOSED 26 Aug 2026, same day, on evidence rather than a decision: both rows sit in CEM-003 and both trace to the SAME eGGSA album url, so it is provably one burial. GRV-0026 keeps it, with the photograph url, the dates and the maiden name folded in; GRV-0724 is retired and not reused.

    Johanna Maria Magdalena Bothma

    opened 2026-08-26

79. ✓ The benchmark story carries no source tags at all, and it is not alone LEAD-0315

    S035 names its evidence in the sentence, as STORY-VOICE.md asks, but carries no SRC- tags. How many stories do the same, and should a citation pass run across all of them?

    What we expected: MEASURED, NOT PREDICTED: FIFTEEN of forty-seven published stories cite nothing at all, neither an SRC- nor a DOC-, including the house benchmark S035 and S036, S040, S038, S039, S031, S033, S034, S027, S028, S024, S022, S017, S015 and S001.

    What would settle it: A decision on whether to run a citation pass across those fifteen. They are not unsourced: every one of them names its evidence in the sentence, which is the harder half. What is missing is the tag a reader can follow.

    What settled it: [27 Aug 2026] PASS RUN. Nine of the fifteen now carry their spine source: S031, S033, S034, S038, S022, S035, S049, S001, S017. *** AND THE PASS FOUND THAT NOT ALL FIFTEEN NEEDED ONE. *** S024 and S027 are SYNTHESIS stories that count across the whole corpus rather than resting on documents of their own, and both say so in their own words: 'Every figure here is counted from the deeds and diagrams on this site and can be recounted', and 'Every figure below is counted from the dates gathered here and can be recounted, and each says how many records it rests on'. That IS the citation for a counting story, and hanging an SRC- on it would be worse than leaving it, because it would point a reader at one document when the claim rests on all of them. They are recorded here as correct as they stand. Still open: S015, S036, S039, S040, and the remaining untagged sentences in S035. | *** CLOSED 27 Aug 2026. FIFTEEN DOWN TO THREE, AND THE THREE ARE CORRECT AS THEY STAND. *** Twelve stories now carry the tag a reader can follow. The three left are SYNTHESIS stories - S024, S027 and S040 - which count across the whole corpus rather than resting on documents of their own, and which say so in their own words. Hanging an SRC- on one of them would point a reader at a single document when the claim rests on all of them. | ONE STRUCTURAL FAULT SURFACED ON THE WAY: STO-0040 HAD NO STORY_LINKS AT ALL, so it appeared on no farm page and the currency sweep was blind to it - with no linked entities there is nothing to walk. Linked to the two parcels its own prose names. A check of every other published story found none below three links, so it was the only one.

    opened 2026-08-27

80. ✓ Which Spreeufontein did Stephanus Botma live on? LEAD-0320

    Frans believes Stephanus Botma (1762-1830, LH6F-TD5), father of the Lodewyk who held Spreeufontein, lived on Spreeuwfontein. Which one?

    What we expected: NOT the Prince Albert farm. A KLEYNSPREEUWFONTEIN and a GROOT SPREEUW stand in the LAINGSBURG registration division on the national cadastre, and Laingsburg sits between Worcester and the Roggeveld, which is exactly where this family was in those decades.

    What would settle it: A residence for Stephanus Botma in any record, or the farm's district on a deed.

    What settled it: [27 Aug 2026] THE DOCUMENT EXISTS AND IS NOW REFERENCED. Stephanus Botma's estate is vol. 111 no. 60 in the Orphan Chamber index, filed 1830 (SRC-598). That file would give his residence, which is the one thing that settles which Spreeufontein. Unread. | [SRC-599, same day] A THIRD FIX ON THE WESTERN SIDE. The Worcester register of 3 April 1821 enters Lodewyk as 'BURGER ALHIER', a burgher of Worcester, and his brother Johannes Paulus three weeks earlier the same way. With the 1834 Worcester marriage and the Roggeveld birth that December, the family is documented on the western side across the 1820s and 1830s. The Laingsburg Spreeufontein reading gets stronger; nothing yet proves it. | [SRC-600, same day] A FOURTH FIX, AND THE EARLIEST. Stephanus Botma and Huibrecht Cloete baptise a daughter at TULBAGH in 1794, with two of his brothers on the same opening. The trail now reads Tulbagh 1794, Worcester 1821 and 1834, Roggeveld December 1834, and the Prince Albert Spreeufontein only from about 1871. THE FAMILY WAS NOT IN THIS DISTRICT IN STEPHANUS'S LIFETIME, so the Spreeufontein of the memory is very unlikely to be ours. Kleynspreeuwfontein in the Laingsburg division remains the candidate; his 1830 estate, vol. 111 no. 60, would say. | *** CLOSED 28 Aug 2026 BY TWO DEATH NOTICES, AND THE ANSWER IS NO. *** SRC-601: Stephanus Botma's widow was born in the Klein Roggeveld and died in 1837 at Hartebeestfontein in the MIDDEL ROGGEVELD, in her son Jacob's house; nothing on that sheet touches Prince Albert. SRC-602: their son Lodewyk died on 30 June 1874 'at his residence Spreeuwfontein, Division of Prince Albert'. So the family's Spreeufontein is OURS and it is LODEWYK'S, reached in his own lifetime and not his father's. The Laingsburg Kleynspreeuwfontein reading is dropped. The trail stands as Tulbagh 1794 and 1802, Worcester 1821 and 1834, the Roggeveld to 1837, Prince Albert by 1871 at the latest. WHEN he came onto the farm is now LEAD-0321.

    Spreeuw Fontein & Toornits Kraal (Spreeufontein No. 26)

    opened 2026-08-27

81. ✓ Identify R du Plessis of Remhoogte and his daughter Miemie LEAD-0336

    Which du Plessis owned Reim Hoogte 167 in February 1901, and who was the ten-year-old Miemie whose memory carries the whole story? Neither is identified and both are one degree from the only war grave in this district.

    What we expected: That R du Plessis belongs to the family buried in CEM-025 - Johannes Gijsbertus du Plessis 1826-1913 and Johanna Elizabeth nee van Rooijen, and their son Jacobus Johannes Petrus 1869-1941 who took Rossel Erf in 1908. NEITHER MAN'S INITIAL IS R, so either the cameo's initial is wrong or a third man held the ground in 1901. Miemie, born about 1891 on her stated age, is likely a MARIA under 5b.

    What would settle it: A transfer or a valuation roll naming the holder of Reim Hoogte 167 in 1901.

    What settled it: Opened 3 Sep 2026 alongside LEAD-0335. *** [7 Sep 2026] CLOSED, BOTH HALVES, BY SRC-172, Helena Marincowitz, *Prince Albert and the Anglo-Boer War (1899-1902)*, Fransie Pienaar Museum, 1999. *** The father is ROELOF DU PLESSIS, named in his daughter's own words at p.34: 'Scheepers' commando came from Zeekoegat and asked my father (Roelof du Plessis) the way to Prince Albert.' PER-002050 carries the name now instead of an initial. The daughter is MIEMIE DE WIT, nee du Plessis, PER-002051. The same page gives her married name, her husband - Christiaan 'Kraffie' de Wit of Scholtzkloof, now PER-002069 - the year she told the story, 1977, her age then, 85, and her death the following year. WHAT THIS LEAD PREDICTED AND WHAT IT GOT. It predicted that R du Plessis belonged to the family buried at CEM-025, Johannes Gijsbertus du Plessis 1826-1913 and his son Jacobus Johannes Petrus 1869-1941. THAT IS STILL UNTESTED. A name and a farm are not a descent, and 9 forbids closing it on the strength of both being du Plessis on one farm. The lead is closed because it asked WHO, and it now has two names; the kinship is a different question and it is open. IT ALSO PREDICTED THE WRONG ROUTE. It looked for the answer in a deeds chain and a valuation roll. It was in a local history in the Fransie Pienaar Museum's own list, which this project had already recorded as SRC-172 and marked 'NOT YET READ' - and had then cited at third hand, through two works that both name it, for a month. The cheapest source was the one already on the shelf.

    Roelof du Plessis (Remhoogte) · Miemie de Wit (nee du Plessis) · Jacobus Johannes Petrus du Plessis · Johannes Gysbertus du Plessis · Johanna Elizabeth van Rooyen · Reim Hoogte No. 167 · Rossel Erf (Reim Hoogte No. 167 Ptn 1)

    opened 2026-09-03

82. ✓ Searching a noemnaam against display_name finds nobody LEAD-0314

    Closed on the day it was opened, on a false premise.

    What we expected: She is the spouse of an uncle in the family exemption, so the existing clause would admit her automatically the moment a row and a union exist, exactly as it admitted PER-001293.

    What would settle it: Frans's word that she should be in, and her full name.

    What settled it: *** CLOSED 27 Aug 2026, WRONG FROM THE START. *** This lead said Karien Oosthuizen had no person row and so could not be named. She has had one all along: PER-000031, CATHARINA VAN DER MERWE, doopnaam Catharina, NOEMNAAM KARIEN, born 1953, and she is one of the eighteen in the family exemption. The search went to display_name, which carries the FamilySearch form under 5a-i, and found nothing. The noemnaam column is exactly the field that answers this and was never queried. | THE SAME QUERY MADE A SECOND ERROR: PER-000018 was called an uncle. He is FRANS'S FATHER, which parentage.csv states plainly. Both people are named on STO-0037 now. | THE LESSON IS 9's, one field over: query the FORMAL name, and when a familiar name is given, search noemnaam as well as display_name. crosscheck.py should probably do both.

    Christiaan Johannes Joubert Oosthuizen · Varsche Fontein

    opened 2026-08-27

83. ✓ LEAD-0326 LEAD-0326

    Do the two du Preez branches in this district join, and where? The Klaare Stroom holder of 1822 and the du Preez family baptising at Prince Albert in 1873 are NOT the same line.

    What settled it: TESTED IN THE PRIVATE TREE, 28 Aug 2026, AND THEY DO NOT MEET WITHIN REACH. OURS (PER-001911): Hermanus Carel du Preez 1760-1839, son of JOHANNES LODEWICUS DU PREEZ (b. bef. 9 Oct 1729) and Maria Potgieter, on Klaare Stroom in 1822. THE OTHER (SRC-604, FS cards): Elizabeth Susanna Johanna du Preez 1854-1928 (G8M2-361) < Jacobus Lodewicus du Preez 1821 (L46V-82D) < ANDRIES LOURENS DU PREEZ c.1787-c.1850 (LC5C-LDF), married Margaretha Alberta Rautenbach at Swellendam on 4 Feb 1809 < PETRUS 'PIETER' DU PREEZ b. 29 Dec 1748 < JACOB DU PREEZ b. 1711 and Susanna Maria Theron. So one descends from a Johannes Lodewicus and the other from a Jacob, and the join is at least a generation above both. Both families are Tulbagh-born, which is suggestive and is not evidence. WHY IT MATTERS: the LODEWICUS name runs through both, and a careless match on it would put the wrong family on the ground the village of Klaarstroom stands on. Recorded so the two stay apart until a document joins them. | *** [FS, 28 Aug 2026] THE OTHER BRANCH IS NOW FULLY NAMED, AND IT STILL DOES NOT JOIN OURS. *** Andries Lourens du Preez (1787-1850, LC5C-LDF) is a son of PIETER DU PREEZ (1748-1829, LCVM-538) and JOHANNA SUSANNA DE BRUYN (1753-1834, LCVM-548), MARRIED AT TULBAGH ON 21 OCTOBER 1770. Ten children, of whom the eldest son is JACOBUS LODEWICUS DU PREEZ (1772-1851, 9CK8-YD3) and the youngest HERMANUS FRANCOIS DU PREEZ (b.1796, LC5C-5W3). *** NOTE THAT YOUNGEST NAME AND DO NOT TRIP ON IT: HERMANUS FRANCOIS, NOT HERMANUS CAREL. *** He is in the private tree, born 5 April 1796, and he is a third man whose name reads close enough to ours to be mistaken for him. | *** A TESTABLE JOIN, STATED AS A HYPOTHESIS AND NOTHING MORE. *** Pieter du Preez's eldest son is a JACOBUS Lodewicus; our man's father is a JOHANNES Lodewicus. Under the vernoemingspatroon an eldest son carries his father's father's name, so Pieter's own father should be a Jacobus - and the private tree does give him JACOB DU PREEZ (b.1711) and Susanna Maria Theron. IF a Jacobus Lodewicus and a Johannes Lodewicus were brothers, the two Klaarstroom-district du Preez branches are first cousins' lines and the join is about 1700. THE PATTERN GENERATES THE HYPOTHESIS AND CANNOT CONFIRM IT; a baptism entry for either father would. | Both families are Tulbagh. Andries Lourens married at SWELLENDAM on 4 February 1809, ten weeks before the Oosthuizen estate was taken on the Olifants by officials of that same drostdy - which is a coincidence of jurisdiction and is not evidence of anything. | *** ANSWERED AND CLOSED 28 Aug 2026. *** *** THE TWO DU PREEZ FAMILIES OF THIS DISTRICT ARE ONE FAMILY, AND THEY JOIN AT THE IMMIGRANT. *** PHILIPPE DES PREZ, born about 1681 at CALAIS and dead 21 May 1722, married ELIZABETH PREVOST (b. 31 Oct 1683), had five children of whom two carry these lines: PHILIPPE du Preez (1705-1765) > JOHANNES LODEWICUS (bef. 9 Oct 1729 - 24 May 1782) x Maria Potgieter > HERMANUS CAREL (1760-1839), on KLAARE STROOM in 1822 JACOB du Preez (1711-1749) > PIETER (1748-1829, LCVM-538) x Johanna Susanna de Bruyn > ANDRIES LOURENS (1787-1850, LC5C-LDF) > JACOBUS LODEWICUS (1821, L46V-82D) > ELIZABETH SUSANNA JOHANNA (1854-1928, G8M2-361), baptising at PRINCE ALBERT in 1873 (SRC-604) So Philippe and Jacob are brothers; Johannes Lodewicus and Pieter are first cousins; and HERMANUS CAREL AND ANDRIES LOURENS ARE SECOND COUSINS. The man who held the ground Klaarstroom stands on, and the family baptising in the village church two generations later, are one family four generations up. | *** THE VERNOEMINGSPATROON POINTED THE RIGHT WAY AND NAMED THE WRONG PAIR. *** I proposed that a Jacobus Lodewicus and a Johannes Lodewicus might be brothers. They are not: the brothers are PHILIPPE and JACOB, one generation higher, and no Lodewicus is involved in the join at all. The pattern was right that the lines meet and wrong about where, which is exactly the limit 5b sets on it - a lead to test, never a conclusion. | *** THIS IS TREE EVIDENCE AND NOTHING MORE. *** Every step above Hermanus Carel comes from the private tree, with no register read and no FamilySearch id below Pieter's generation. 10 ranks original > register > tree, and a compiled Huguenot descent is exactly the sort of thing that gets copied between compilers without being checked. [TO VERIFY] | *** AND THE TRAIL STOPS HERE BY RULE, NOT BY EXHAUSTION. *** Philippe des Prez's own father is given as Hercules des Prez (c.1645) and Cecile d'Atis. That is the head of a trail and 9a gives a head of trail a bounded set of records and no growth - the same treatment the Oosthuizen stamvader gets. NO PERSON ROWS have been created for any of these people. The question worth answering was whether this district's two du Preez households are one family, and it is answered.

Back to the open questions