Open Questions
Everything this archive knows it does not know, with the specific record set that would answer each. Nothing here is rhetorical — every line names somewhere to look.
Three letters
Western Cape Archives — sent 10 Sep 2026
Three documents have been identified that are the likeliest in existence to break this archive's oldest wall, and none is digitised:
| MOOC 7/1/104: 42 | James Mountjoy's will, 1818 — joint with Anna Catharina Slier, made a year after their marriage |
| MOOC 7/1/104: 43 | the same pair again, 1827, the year he died |
| CO 3972: 17 | a Colonial Office memorial, 1834 — "the widow of J Mountjoy, requesting remission of a fine". A petition in her own voice. |
A Cape will of 1818 will often state where the testator came from, and a memorial to the Colonial Office very often does. The letter also asks whether a liquidation account survives for estate 8907, which would name the second wife.
Sent 10 September 2026 to archives.clientservices@westerncape.gov.za, the address on the Western Cape Government's own service page. Delivery confirmed — no bounce. Awaiting a reply.
National Archives, Pretoria — ANSWERED 15 Sep 2026
Six Transvaal estate files, and a queue that has stopped being a screen problem. The films that carry them refuse to serve images: film 007830063 and every film after it in the collection returns 403, and those are exactly the ones running to 1946 and later.
| MHG 1095/48 | George Downing Mountjoy Booyzen, d. 1947 |
| MHG 3734/46 | Catherine Mary Sophia Daschner, born Booyzen — should name her parents |
| MHG 769/52 | John Barry Booyzen, who signed the 1905 form |
| MHG 5155/51 | Gert Jacobus Booyzen, b. 19 Jan 1870 |
| MHG 10/57 | Henry James Booyzen |
| MHG 52653 | Jacobus Nichols Booyzen, 1923 — the one that cannot even be found on film |
Sent 13 September 2026 to enquiries2@dsac.gov.za, the research address on the National Archives' own contact page, copied to the Head of Client Services. It asks for the death notice from each of the six ahead of anything else, on the ground that six death notices beat one complete file. ANSWERED ON 15 SEPTEMBER, by Ms Mavis Xaba, Archivist Assistant in the Reading Room: “All the estate files you are requesting are bound files. Unfortunately, we are therefore unable to scan these files and provide complete digital copies by email.” That is not a refusal — it is a physical fact about the record, and it retires an assumption this archive had been making for weeks, that the obstacle was willingness or staff time. She corrected none of the six references, including MHG 52653, and put a private researcher who specialises in estate files on copy. That researcher has been written to and asked what she charges and whether the work is worth doing; engaging her is the user’s decision and the user’s money. The work list →
SANDF Documentation Centre — half delivered
Correction This archive recorded the pension enquiry as sent on 9 September 2026. It was sent, and one of the two addresses bounced:
550 5.7.1 <sandfdoc@mweb.co.za>: Relay access denied
That is the address published almost everywhere for the Centre — the National Archives directory, the university guides, the genealogy sites — and it refuses mail at the server. The enquiry did reach archive@dod.mil.za, which has not replied.
"Sent" and "delivered" are not the same word, and this archive used the wrong one. What changed →
The queue
0 Who were Gerthardus Lodewikus Booysen's parents? Blocking
- What is known
- STILL BLOCKING, and the named lead is now exhausted at the Cape depot. The queue pointed at the estate papers of his wife MAGDELENA ELISABETH DE BEER, d. 1843. The Cape Town Archives Repository has been searched under DE BEER + MAGDELENA, BEER + MAGDELENA + ELISABETH, BOOYSEN + MAGDELENA, BOOIJSEN + MAGDELENA, BOOYSEN + GERTHARDUS, BOOYSEN + GERHARDUS, BOOYZEN + GERHARDUS and BOOYSEN + LODEWIKUS/LODEWICUS. There is no estate file for her, and none for him. One near-miss worth recording and NOT adopting: KAB NCD 2/1: 49, a notarial will of 1793 for 'JAN BOOYSEN, MARRIED TO THE LATE MAGDALENA DE BEER' — wrong husband's name, and a wife already dead fifty years before ours. It shows Booysen–de Beer was not a unique pairing, which is a warning about the tree rather than a lead. WHAT HAS CHANGED, 10 September 2026: the 1876 Rustenburg Sterfkennis of his son now names BOTH of them in line 3 — "Gerthardus Lödewikus Booijsen" and "Magdalena de Beer". That does not answer this question, which is about HIS parents, and the entry stays blocking. But generation seven is no longer an assertion of the family tree, the 1793 Jan Booysen will can be set aside rather than merely resisted, and the search has a new country to run in: the Transvaal probate series, where this family turns out to have been living from at least the 1870s.
- What has been tried
- CONFIRMED A SECOND TIME, from an older source. The Master of the Supreme Court's own index of death notices 1834–1912 has been read on film at the “Beer, de” pages: the Magdalena de Beers in it are Magda Catha du Plessis (1872), Magda de Beer (1865), Magda Vermeulen (1883), Magda Eliz Troskie (1863), Magda Jacomina Nel (1850) and Magda Greef (1839/40). No Magdelena Elisabeth, and nothing in 1843. So the absence is not an artefact of the modern NAAIRS index. What is left is the published Suid-Afrikaanse Geslagsregisters Booysen genealogy, and Dutch Reformed baptism registers for Graaff-Reinet and Paarl 1740–1790 read page by page rather than by index — his son was baptised at Graaff-Reinet in 1788. And now, newly available: the Transvaal Master’s series, where the same son died in 1876 — Weeskamer records, Rustenburg district registers, and the estates of the seven newly-named siblings. AND ON 13 SEPTEMBER, FROM WIKITREE, THE FIRST HARD POINTER THIS QUESTION HAS EVER HAD: its profile for Petrus Jacobus cites HEESE, SUID-AFRIKAANSE GESLAGSREGISTERS, DEEL 1, PAGE 331. This page has named that book as the route left since the beginning and marked it Not yet followed, without a page number to follow it to. There is one now. The same profile also makes the father JAN BOOYSEN rather than Gerthardus Lodewikus — against line 3 of the 1876 Sterfkennis, which this archive has read — and that matters here because it is the SAME Jan as the 1793 notarial will KAB NCD 2/1: 49, the near-miss this question declined to adopt. A document outranks a tree and nothing has been changed. But the 1793 will is no longer a curiosity, and the disagreement now has a printed book and an exact page to be settled against. AND AN HOUR LATER, TWO MORE PROFILES SETTLED IT — IN THIS ARCHIVE'S FAVOUR. Jan Booysen (Boijens-9), baptised Drakenstein 29 January 1741, married Magdalena de Beer on 21 January 1770 AND MARRIED AGAIN ON 15 DECEMBER 1793. So his de Beer wife was dead by December 1793; ours died in 1843, fifty years apart, which is exactly the objection this entry raised when it declined the 1793 will. THE REASONING WAS RIGHT, and the will is now positively identified as HIS — a man remarrying in December 1793 is a man making a will that year. The tree's error is explained too: its own de Beer profile (De_Beer-638) carries the codes "b6c2 (and b6c6)", two baptisms nine years apart (1745 Tulbagh, 1754 Stellenbosch), and a sibling list holding BOTH a Magdalena and a Magdalena Elizabeth. Two women merged into one profile is what put Jan over our Petrus Jacobus. ONE GAIN: if ours is the second of the two, her parents are ZACCHARIAS DE BEER and ANNA PRETORIUS — the first candidate this question has ever had for her side. A lead, needing the codes checked in print. AND ON 14 SEPTEMBER THE BOOK WAS READ, AND IT DID NOT GO THIS ARCHIVE'S WAY. de Villiers 1893 is freely downloadable; page 34 marks only the 1754 Magdalena “zie Booijsen”, which is this archive's own birth year for her, so she is ONE woman and not two — the paragraph above is withdrawn and logged on /corrections/. Page 78 then gives her husband as JAN BOOIJSEN, baptised 29 January 1741, and their SEVENTH CHILD as “Petrus Jacobus, gedoopt 21 September 1788, gehuwd met Anna Maria van der Merwe”, whose first child is “David Schalk, gedoopt 4 Januari 1807”. FOUR FIELDS AGREE WITH THIS ARCHIVE AND THE FIFTH IS THE FATHER. There is exactly one Petrus Jacobus of 21 September 1788 in the whole volume and exactly one David Schalk of 4 January 1807, and the word Lodewikus appears nowhere in it. AND ON 14 SEPTEMBER A SENIOR WIKITREE MEMBER OFFERED A CANDIDATE FOR THE NAME. Riël Smit: the only Gerhardus Booysen in DE VILLIERS & PAMA (1981) is b3c1d1 GERHARDUS LORET BOOIJSEN, baptised September 1770 at Drakenstein, and “maybe the Loret changed to Lodevicus over time”. This archive had already met that man in the 1893 text and not known what to do with him. He does not solve the question — he is b3c1d1, a son of BOOIJ and not of JAN, which makes him Petrus Jacobus's first cousin rather than his father, and he was eighteen in 1788. But he is the first plausible source for the NAME on the 1876 notice, and the likeliest reading is now that the informant reached for a remembered family name and hung it on the wrong man. SO THIS QUESTION HAS CHANGED SHAPE: it is no longer “who were Gerthardus Lodewikus's parents” but “DID HE EXIST?” He is in this archive on one line of one 1876 death notice, in the field that asks a family for the names of grandparents dead seventy years. Nothing has been rewritten. But the notice is no longer the only evidence in the room, and the next step is the Graaff-Reinet baptism register for 21 September 1788, read directly. AND ON 15 SEPTEMBER THE LAST ESCAPE ROUTE CLOSED AND THE MECHANISM WAS FOUND. The standing objection was that Pama’s 1981 revision might have moved him, in which case the death notice would stand unopposed. RIËL SMIT, WHO OWNS THAT EDITION: “Yes, de Villiers & Pama (1981) still place Petrus Jacobus (b3c4d7) b. 21 Sept 1788 under JAN b3c4.” PIETER VAN ZYL searched for a marriage between a Gerhardus Lodewicus and a Magdalena de Beer AND FOUND NONE, noted that Jan had been dead more than forty years when Petrus Jacobus died, and remembered from the notice itself that a CHILD of Petrus Jacobus carried the name. HE IS RIGHT AND THE CHILD IS NINE LINES LOWER ON THE SAME SHEET: line 12 lists eight children and the fifth is JAN GERTHARDUS LODEWIKUS BOOIJSEN. The name in the parents’ box is the back half of a son’s name in one clerk’s hand, and the half that was dropped is JAN. The man who signed was Hendrik Stefanus, the seventh child, born about 1828, writing the name of a grandfather dead before he was ten. THE SPINE HAS NOT BEEN MOVED AND THAT IS DELIBERATE: everything against Gerthardus Lodewikus is secondary — one printed genealogy agreeing with its own earlier edition is one source agreeing with itself — and what he has is a primary document signed by a son. Jan Booijsen is entered as inferred, not documented. AND THE SCREEN ROUTE TO THE BAPTISM IS NOW KNOWN TO BE SHUT: the FamilySearch Dutch Reformed collection holds 54 registers for Paarl (Drakenstein), 48 for Graaff-Reinet and 55 for Stellenbosch, and SORTED EARLIEST FIRST THE FIRST ROW IN EVERY TOWN IS 1843. It is the post-1842 church archive series and it does not reach the eighteenth century at all.
- What would settle it
- THE BAPTISM ENTRY OF 21 SEPTEMBER 1788 ITSELF, AT CAPE TOWN. De Villiers is not a source; he is an INDEX to one, compiled from the doopregister. That entry names the father in a contemporary hand, written within days of the birth, by somebody with no reason to be confused. The originals are at the Western Cape Archives, 72 Roeland Street; Drakenstein first, since b3c4 Jan was a burgher there, then Graaff-Reinet and Stellenbosch. NOT REACHABLE THROUGH FAMILYSEARCH — that collection begins in 1843.
1 Is there a Booysen on a Cape Rebel roll? — NO. BUT THREE WERE TRANSVAAL PRISONERS OF WAR, WHICH IS A DIFFERENT QUESTION THIS ARCHIVE NEVER ASKED. High
- What is known
- The Anglo-Boer War Museum's register of 12,792 Cape Rebels holds 26 Booysens, every one from the Northern Cape or Griqualand West. 'Booyzen' returns nothing in any of the fifteen registers. THE SURNAME BARRY HAS NOW BEEN SEARCHED, as this entry said it should be, and it returns four: BARRY, FRED J., plaasboer of IVANHOE, district BARKLEY-OOS/WODEHOUSE, TRIED JUNE 1901 — the same district and the same month as Daniel Willem Schoeman of Caerlaverock, in whose house the family married five weeks later; BARRY, RICHARD PETER, plaasboer of JERPOINT — and Richard Peter Barry is Kathleen Barry's brother, the R. P. Barry who signed his mother's 1915 death notice at Manorowen, Barkly East; plus Adolf Barry of Keimoes and Thomas Andrew Barry of Oudtshoorn, neither relevant. Identity is not proven — the register gives no birth dates or parents — but the claim that this family was rebel country now has two men of its own surname in its own district.
- What would settle it
- Birth dates or parents for the two men, to close the identification: Cape Rebel trial files for Barkly East and Wodehouse, June 1901; and the published three-volume Rebel Record. Also JERPOINT — an Irish abbey name, joining Manorowen and Caerlaverock as farm names in this district, and the first Irish place name found near the Barrys.
2 Who was “D. W. Schoeman” of Caerlaverock? High
- What is known
- ANSWERED IN PART. The Cape Rebel register holds DANIEL WILLEM SCHOEMAN, plaasboer, of CAERLAVEROCK, district Wodehouse–Barkly East, TRIED AS A CAPE REBEL IN JUNE 1901 — the month before the family married in his house. A second rebel, Christiaan F. Klopper, farmed the same place and was tried the same month. What remains open is whether he was also the COMMANDANT D. Schoeman of Labuschagne's Nek: the record's Commando field is blank, and there is a cluster of D. W. Schoemans in the same division (Wolwekloof, Waschbank, both tried 1 May 1901).
- What would settle it
- The Boer Officers register (514 records; a surname search reports 2 Schoemans, but the filtered list would not display); the Barkly East treason files for a sentence; Cape deeds for Caerlaverock; and the family of Ruben Schoeman of Barkly Oost.
3 Were the women and children in a camp? — SEARCHED, AND NOT FOUND. Medium
- What is known
- The Anglo-Boer War Museum's fifteen registers contain no camp list at all, so this entry named the wrong source. The right one is the University of Cape Town's British Concentration Camps of the South African War 1900-1902 database, and it has now been searched. BOOYZEN returns 5, and every one is Transvaal — Gert Jacobus, aged 2, died Johannesburg camp, farm Rietvallei, district Marico; Gertruida Magdalena, aged 19, died Vereeniging camp, farm Rietvlei, district Heidelberg; Jan Gerhardus Lodewicus, aged 2. None from the Cape. AND ON 15 SEPTEMBER THE THREE WERE READ ON THE DATABASE’S OWN PAGES, WHICH DO NOT SAY WHAT THIS ENTRY SAYS. Gert Jacobus and Gertruida Magdalena are each recorded “died in camp: YES”, with farms — Rietvallei, district Marico, and Rietvlei, district Heidelberg — and Gertruida carries the note “or Irene RC”, so even her camp is uncertain. JAN GERHARDUS LODEWICUS IS RECORDED “DIED IN CAMP: NO”, WITH NO CAMP AND NO FARM AT ALL. Listing all three together as camp deaths was this archive’s compression, not the database’s. The source is named on every page: RS 25 TRANSVAAL DEATH LISTS, National Archives, Pretoria — which is exactly why the caveat below is right, and the cause of death on all three is one word, ZIEKTE. BARRY returns 0 in the entire database. Booysen with Catherine, Kathleen, John Barry, or Willem Hermanus all return 0, as does Booyzen with Catharina. So no member of this family has been found in a camp. IMPORTANT CAVEAT: that database is built largely from camp DEATH lists and surviving registers, so it records who died far better than who was interned and walked out. Absence here is much weaker evidence than presence would have been, and this claim is NOT disproved — only unsupported.
- What would settle it
- Cape camp registers proper rather than death lists — Aliwal North is the camp nearest Barkly East. Also the Cape Colony refugee and 'undesirables' returns, which handled rebel families in the Cape districts and are not in the UCT database.
4 Is the letter in the 1879 signature a z or an s? Closed
- What is known
- LARGELY EMPTIED, and by an answer rather than a failure. The archive scan of the 1851 Graaff-Reinet baptism register shows the father written unambiguously as BOOYZEN — and, three lines below in the same hand and ink, the godparents written twice as PIETERZEN. The Pietersens changed no name: -zen is simply how Dutch writes the patronymic ending, and this clerk wrote it that way for everyone. A z-shaped letter in a Cape Dutch hand is therefore not evidence of intent, and the 1879 signature can no longer carry the weight this archive put on it. What the question was really asking — when did the z start? — is not answerable from letterforms at all.
- What would settle it
- Nothing further in the handwriting. The change itself now rests on the 1905 sheet (two spellings of one name on one page), the tree's recorded 'Former name: Booysen', and the distribution data.
5 Was John Augustus Barry Irish, English, or born at sea? — ANSWERED. IRELAND. Answered
- What is known
- ANSWERED, 13 September 2026, by a document twelve years older than the one this archive had been arguing from. The 1903 REGISTRATION of his death — entry 74, district Barkly East, made by his son RICHARD BARRY, who was present when he died — enters question 5 as “IRISHMAN BORN IN IRELAND”. It also gives his age as eighty-four years two months, his trade as CARPENTER, his residence and place of death as MANOR OWEN Ward IV, his death as 25 AUGUST 1903, and the number of his children as ELEVEN. The 1915 death notice that typed “at sea, exact whereabouts unknown, British” was sworn by a different son twelve years after the fact. Where the two disagree this archive now follows 1903: it is closer to the man, its informant was in the room, and it is a statutory registration rather than an estate paper. THE OLD TEXT OF THIS ENTRY FOLLOWS. PARTLY ANSWERED. The 1915 death notice is now read: it types his birthplace as “At sea (exact whereabouts unknown), British” and his mother as “born Neagle”. What remains is the other two documents, and the question of who his parents actually were.
- What would settle it
- The two 1903 death certificates named in the tree note; Cape Mounted Rifles records at Johnson’s Leap; Cork Nagle/Neagle baptisms around 1819; passenger lists 1819–1820. Also new: why a farm in Barkly East is called MANOROWEN, a village in Pembrokeshire, Wales.
6 Break the John Barry (1796–1853) conflation. — BROKEN, 14 September 2026. THE YORKSHIRE HALF DOES NOT EXIST, 16 September 2026. Answered
- What is known
- One profile holds a Yorkshire farmer of Bilsdale and a man who died at de Bruins Poort in the Cape.
- What would settle it
- THE ENGLISH HALF HAS NOW BEEN SEARCHED AND IS EMPTY. FindMyPast’s Yorkshire Burials carries Bilsdale, St Hilda from the Ryedale Family History Society transcripts — 5,934 burials, with entries at 1849, 1851, 1857, 1859 and on through the 1860s, so January 1853 sits inside an indexed run. BARRYS AMONG THEM: ZERO. Barry baptisms at Bilsdale: zero. Barry burials elsewhere in Yorkshire: 771, so the surname is not invisible to the index. Barry burials anywhere in Yorkshire in 1853: one, at Sheffield St George Brook Hill, with no forename given. John Barry burials in Yorkshire, all years: 66 — the list runs 1837, then 1856, 1858, 1858, 1860. A MAN BURIED AT ST HILDA’S ON 3 JANUARY 1853 IS NOT IN AN INDEX OF 5,934 BILSDALE BURIALS. AND THE CENSUSES PRODUCE THE MAN, WHICH IS BETTER THAN A NEGATIVE. There are 50 Barrys in the North Riding in 1841 and 52 in 1851, and NOT ONE OF THE HUNDRED AND TWO IS IN THE HELMSLEY REGISTRATION DISTRICT, which is the district Bilsdale sits in — they are at Whitby, Scarborough, Brompton, Ruswarp, Sneaton, Allerston, Lythe, Fylingdales, Easingwold, Great Smeaton, Eston, Hinderwell and Filey. ACROSS BOTH CENSUSES THERE IS EXACTLY ONE JOHN BARRY BORN IN 1796 IN THE WHOLE NORTH RIDING, and he is a SHIP OWNER living at SHIP YARDS, RUSWARP, WHITBY — 1841 HO107/1265 book 11 folio 38 page 4 schedule 1116, aged 45, with Jno, Mary, Ann and Catharine; 1851 HO107/2374 folio 563 page 12 line 6 schedule 46, aged 55, married, born Whitby, with Mary, Catherine and Elizabeth. THE SAME HOUSE TEN YEARS APART. He is not a farmer, he is not of Bilsdale, there is no Barry at Holly Bower, and he was keeping house at Ruswarp in the spring of 1851. WHOEVER THE TREE MEANT, HE CANNOT ALSO HAVE BEEN BURIED AT ST HILDA’S ON 3 JANUARY 1853. What remains open is much smaller: whether the Whitby shipowner has any connection at all to the man who died at de Bruins Poort. Nothing found says he does.
7 Who was Petrus Jacobus Booysen's second wife? — ANSWERED. JOHANNA MARGARITHA PRETORIUS. Answered
- What is known
- ANSWERED, and by a christening rather than a death. The archive had spent this question on death notices: his own 1884 notice, which enters him as married twenty-one years after his first wife died but never asks her name; and his son Stephanus Francois's of 1940, where the widow wrote only 'oorlede' for both parents. The mistake was looking at deaths. A BAPTISM NAMES BOTH PARENTS — and the child whose birth date the 1884 notice gives to the day, GERT JACOBUS, BORN 19 JANUARY 1870, was baptised on 19 June 1870. The register reads: Oud: PETRUS JACOBUS BOOYZEN en JOHANNA MARGARITHA PRETORIUS. Read from the image, not the index. And the witnesses are worth as much as the answer: PETRUS JACOBUS BOOIJZEN JUNIOR, born 1839, and MAGDALENA JOHANNA FRANCINA BOOIJZEN, born 1853 — the third and sixth children on the 1863 death notice. Two children of the first marriage stood godparent to the first child of the second.
- What would settle it
- Her own dates and parentage; the marriage itself, now bounded to 1863–January 1870; and whether Izak (1864) is hers too. Note and do not chase: Pretorius is also the surname of Anna Johanna Maria Pretorius, who married into this family in 1924 — but it is among the commonest surnames in the Cape and a coincidence is not a link.
8 Extract what is left of the photographs and biographies. — DONE. Answered
- What is known
- COMPLETE, and this entry was carrying three stale claims of its own. It said George Kolbe's 1803 baptism was 'downloaded and unread' — it had in fact been read and published (born 7 Dec 1802, baptised 2 Feb 1803, son of John Gottlob and Mary, at St Anne's Soho). It listed a 'Johanna Catharina birth' as unread — that image is the 1889 Doop Register spread already transcribed, pages 212-213, with the Booyzen baptism on one page and the Mountjoy baptism on the other. It listed 'george death information' as unread — that is the 1947 death form already published. Of everything on the two profiles exactly one image was genuinely unread, and it has now been read: the 1891 baptism at Barkly Oost of CATHARINA MARIA SOPHIA BOOYSEN, born 22 April 1891, a daughter of Willem Hermanus and Kathleen new to this archive.
- What would settle it
- Nothing on these profiles. The remaining unextracted profiles are the ones this entry already names and none is a document set: Anna Pretorius, Margarietha Barry and Kathleen Barry hold photographs rather than records, and the five Kolbe children's death records are the only document group left.
9 Confirm the 1820 Settler voyage independently. — CONFIRMED, and it turned up a second lie. Answered
- What is known
- CHECKED against the settler roll for Owen's party. Every date the tree carried is right: Gravesend 3 December 1819, the Nautilus, Table Bay 17 March 1820, Algoa Bay 14 April 1820, fifty-two settlers. The roll adds an occupation nobody had — he is entered as a CHEMIST, not the preacher or farmer this archive knew. And it adds ages: 'KOLBE, George Augustus 20. Chemist. w Margaret 19.' He was born 7 December 1802 and the ship sailed four days before his seventeenth birthday; she was born 23 February 1803. They were SIXTEEN AND SIXTEEN. Set beside the 1819 marriage allegation, where both swore they were twenty-one at sixteen and fifteen, that is two age overstatements in one year on two official documents. Whether the second was theirs or the party organiser's cannot be told from the roll.
- What would settle it
- The Colonial Office settler papers themselves, rather than a compiled transcript — CO 48 series — which would show whether the ages were sworn or simply written down.
10 Did the boys work on the mines? — YES. ONE DID, AND THIS ARCHIVE HAD SAID THERE WAS NONE. Medium
- What is known
- The last line of the testimony, now tested as far as it can be from here. No Booyzen mining record appears anywhere in the national archives index, searched across ALL repositories for MINE, MINER and MYN. That index covers government files rather than private company employment registers, so the absence is weak on its own — but it is not on its own. Every occupation this archive has actually read for these men points somewhere else: Willem Hermanus is a boer in 1879 and a MASON in 1905; George Downing Mountjoy is a BOOMKWEKER, a tree nurseryman, in 1924 and a CLERK ON THE SOUTH AFRICAN RAILWAYS at his death in 1947; his grandfather was a farmer and his great-great-grandfather a saddler. Not one miner among them. THAT SENTENCE IS NOW WITHDRAWN. The Indwe marriage register of 21 September 1908 enters WILLIAM HERMANUS BOOYZEN, aged 22, jonkman, under Rang of Beroep as MIJNWERKER, living at DUGMORE MIJN — and his mother Kathleen Barry stood godmother at his son's font at Indwe three months later, which is how this archive knows whose son he is. So the last line of the testimony is supported, and mislocated: Indwe is a COAL town in the north-eastern Cape, this is 1908, and the testimony puts the mines after a train to Johannesburg that had plainly not yet been taken. The negative was built on the occupations this archive happened to have read, and it had not read this one because it did not know the man existed.
- What would settle it
- Private records rather than public ones: Witwatersrand mining company employment registers, and Johannesburg street directories 1905–1935 for Booyzen at Forest Hill and Mayfair. Also the five Transvaal estate files now identified, which will state each man's occupation at death. And now, first: Dugmore colliery employment records if any survive, and the Indwe registers read forward from 1910 for the rest of this man's life.
11 Find the farm — and test the burning. AND NOW A LAND REGISTER THAT IS NAME-INDEXED, WITH BOOYZEN IN IT. High
- What is known
- STILL UNTESTED, and two blind alleys closed now. He is a BOER (farmer, of Lady Grey) at his marriage in 1879 and a MASON at his death in 1905; something took the land away in between. The Anglo-Boer War Museum's Farm List was searched for Caerlaverock, Manorowen and Wolwekloof and returned nothing — but that register turns out to be an ORANGE FREE STATE LAND ROLL (farm number, morgen, square roods, owners), covering only Winburg, Fauresmith, Vrede, Rouxville, Bethlehem, Bloemfontein, Hoopstad, Kroonstad and Philippolis. A control search on 'Dordrecht' returns 670 Cape Rebels and 0 farms. The negative result is VOID, not informative.
- What would settle it
- 13 September 2026: the Cape series is identified. It is the WAR LOSSES COMPENSATION COMMISSION, Cape Town Archives Repository, 1902 — the Cape's equivalent of the Transvaal CJC. KAB returns 454 documents for COMPENSATION and WAR, and 679 for LOSSES. BOOYZEN returns 39 documents in KAB overall and NONE of them in combination with LOSSES or COMPENSATION; BOOYSEN returns 2,370 overall and exactly one in combination, a WORKMEN'S Compensation Act claim at King William's Town in 1947. BUT THE ZERO IS WEAK AND THE CONTROL SAYS SO: SMIT, SCHOEMAN and BARRY also return nothing under LOSSES, and the two BOTHA hits turn out to be a 1935 locust and arsenical-poisoning stock claim from an entirely different series. The Commission's files are catalogued by sub-commissioner and volume, not by claimant, so a name search cannot reach them. What is left is the Commission's own claim registers, in the Reading Room at Roeland Street, which no search from here can substitute for. Also still untouched: Cape deeds and quitrent records for Barkly East, Lady Grey and Wodehouse, and Cape district and refugee-camp records.
12 Who were James Montjoy's parents? — NOBODY AT THE CAPE KNEW, AND NOW WE KNOW WHY: HE ARRIVED IN UNIFORM. High
- What is known
- READ. KAB MOOC 6/9/21 ref 4393 was located by interpolating reference numbers across film 007844936 and found at image 339 of 2,223. (THIS ARCHIVE PUBLISHED 338 AND 338 IS SOMEBODY ELSE'S FAMILY — corrected 9 October 2026 after both frames were opened: 338 is the death notice of CHRISTOFFEL HENDRIK OLIVIER, trader, who died at his own house in the town of Graaff-Reinet on 15 June 1840, and 340 is WILLEM JACOBUS VAN DER MERWE of Bouwbosch Fontein. The record ark 1:1:QP4Z-WK68 opens at 339.) James Montjoy, saddler, born in 'Europe' about 1782, died at Cradock in September 1827 leaving a house and erf worth £90 and four children under nine. Line 3, Names of the Parents of the Deceased: NOT KNOWN — written thirteen years after his death by a relative who knew his trade, his property and his wife's first husband. It also settles two things: our tree's 1827 death date is RIGHT (the 1840 file date is the filing year, and the contradiction was this archive's error), and Anna Catharina was indeed the widow of a Pietersen, confirming the half-brother inference from the 1851 font.
- What has been tried
- NOTHING FURTHER INSIDE SOUTH AFRICA, and the 1817 marriage register now proves it rather than assuming it. That register HAS been found and read: he married Anna Catharina Slier, widow of Harmanus Pieterzen, at Graaff-Reinet on 9 March 1817. Cape entries of that date routinely record where a party came from, and three couples on that very page are entered 'beide van Uitenhage' — but against James Montjoy there is a ruled dash and nothing. The blank is not an artefact of the 1840 notice being written thirteen years after his death; it was already there at his wedding while he stood in the room. THREE DOCUMENTS HAVE NOW BEEN LOCATED THAT THIS ARCHIVE HAD MISSED, because it filtered a search and wrote the filter up as the whole search: KAB MOOC 7/1/104: 42 — JAMES MOUNTJOY'S WILL of 1818, joint with ANNA CATHARINA SLIER, HIS WIFE, made a year after their marriage; KAB MOOC 7/1/104: 43 — the same pair again in 1827, the year he died; and KAB CO 3972: 17 — a Colonial Office memorial of 1834 from THE WIDOW OF J MOUNTJOY requesting remission of a fine, which is a petition in her own voice seven years after his death. A will of 1818 may state where the testator came from, and a memorial to the Colonial Office very often does. These are now the most likely documents in existence to break this wall, and none of them is digitised. As for the English end: FindMyPast returns 12,984 for the surname and birth year alone, and “a saddler in the Cape by March 1817” will not narrow that — the way in is a Cape arrival record, not an English birth record. AND ON 15 SEPTEMBER THE ARRIVAL RECORD WAS NAMED, BY A MAN WHO SLEPT IN HIS HOUSE. John Montgomery, founder of Burgersdorp, in The Friend of the Free State of 25 November 1869, recalling his arrival at Cradock in January 1822 when the village held about six people: “Messrs. SOLOMON and S.M. DE BEER, the Rev. Mr. EVANS, Mr. P. GIE, JAMES MOUNTJOY (A SADDLER, WHO WAS DISCHARGED FROM THE 21ST LIGHT DRAGOONS), the two Messrs. ZIERVOGEL, Mr. HEILBERG, Fieldcornet B.J. VORSTER, and Gert VAN DEVENTER. After some trouble I obtained lodgings for the night from Mr. MOUNTJOY.” THERE IS NO ROOM IN A VILLAGE OF SIX FOR TWO SADDLERS OF ONE NAME. The 21st Light Dragoons were the cavalry of the Cape garrison from the 1806 expedition. He did not arrive as a settler and no passenger list holds him, which is exactly why the 1817 marriage register rules a line through the column and leaves it blank and why his death notice says NOT KNOWN — the colony had no way to know. THE ARMY DID. This entry is no longer blocked on a missing kind of document; it is waiting on a named one, in a building in London.
- What would settle it
- WO 97 AT THE NATIONAL ARCHIVES, KEW — Royal Hospital Chelsea soldiers’ documents, arranged by regiment for discharges before 1855, giving a man’s PARISH AND COUNTY OF BIRTH, his age at enlistment, his trade and his service. With it WO 12, the muster books and pay lists, which show when he stopped appearing and where the regiment was. Neither is digitised for this regiment. The three Cape documents named above — the 1818 and 1827 wills and the 1834 Colonial Office memorial — are still worth having and are still not digitised, but THEY ARE NO LONGER THE ONLY ROUTE, and they were never the likeliest one.
13 Which two of John Barry's eleven children are still unnamed? — ANSWERED, 16 SEPTEMBER 2026. MARGRITH LOUISA MARIA (1852) AND WILLIAM FREDERICK THOMAS (1861). Answered
- What is known
- Three documents count this family and no two agree. The 1903 death registration says ELEVEN children. John Augustus Barry's own death notice names EIGHT. His wife's, estate 1291, names NINE — the ninth being WILLIAM, who appears on nothing else. Two have never been identified. Every earlier attempt failed for a structural reason worth naming: you cannot search an index for a person whose name you do not know. On 13 September the question was re-put as a RELATIONSHIP — every Barry child anywhere in South Africa with a KOLBE mother, no forename given — and returned one name this archive did not hold: HENRY JAONAS BARRY, born 13 July 1848, baptised at TWENTY YEARS OLD on 13 November 1868, double-indexed to Cape Town and to Tulbagh. Two things pull against each other on it. FOR a new child: John Augustus II was born 13 March 1847 off a register, so James Henry Barry's “about 1847” — an age worked back from his 1939 death — cannot stand, and 13 July 1848 is exactly where a next child belongs. AGAINST: this family reuses forenames relentlessly, and JONAS appears nowhere among the Barrys or Kolbes in this archive's 3,653 distinct names. AND THE INDEX ROUTE IS NOW CLOSED, DEMONSTRABLY: sweeping every spelling the index gives the mother — Kolbe, Kolber, Kolbee, Colbe, Colby, Colbee — returns seven baptisms and nothing new, while HENDRIK NICHOLAAS, RICHARD PETER, MARGARET MARY and KATHLEEN MARY, all four named on their parents' own death notices, appear in it nowhere. An index that cannot produce four children known to exist will not produce two that are unknown. What is left is the registers read across the gaps — and the book has now been found and measured. BURGERSDORP, BAPTISM RECORDS 1852–1866, image group 008039088, 534 images, covers 1854 to 1858 entire. William Fredrik's 1853 entry is at image 54 and the 1866 entries are at image 528, so about 36 images a year, putting the gap at roughly images 70 to 250. There is no name index at the back. That is a sequential browse of some 150 images, not a search, and it has not been done. RESOLVED 14 SEPTEMBER, AGAINST THE CANDIDATE: Henry Jaonas Barry IS James Henry Barry, already in this archive. One baptism indexed twice — 1868 on one entry, 1848 on the other — same birth date, same parents, both attached in the FamilySearch tree to James Henry Barry 1848-1939. The register is titled Graaff-Reinet Baptism Records 1846-1851, which cannot hold an 1868 entry. So TWO children are still missing, not one, and this question is back where it was — but James Henry now has an exact birth of 13 July 1848, a baptism of 13 November 1848, and Graaff-Reinet rather than the Cape Town and Tulbagh the index gives.
- What would settle it
- ANSWERED. The sign-in came back on 16 September and the question fell in under an hour — not to a name search, which this entry correctly said was closed, but to THE RELATIONSHIP SEARCH: father John Barry, mother a Kolbe, child's own name left blank. Nine baptisms come back for the couple, and TWO OF THEM WERE NOT IN THIS ARCHIVE. MARGRITH LOUISA MARIA, baptised at BURGERSDORP on 5 APRIL 1852 — indexed twice, as “Margrith Louisa Maria” and “Margreth Louisa Mary”, AND WITH NO SURNAME OF HER OWN EITHER TIME, which is the blank Name column that hid her sister of 1855. And WILLIAM FREDERICK THOMAS BARRY, born 29 SEPTEMBER 1861 and baptised at GRAAFF-REINET — invisible for the OPPOSITE reason, because he carries the same name as the brother baptised in 1853 and a name search returns that one and stops. NINE BAPTISMS PLUS HENDRIK NICHOLAAS AND RICHARD PETER MAKES ELEVEN, which is exactly what the 1903 registration says. Margaret Mary of the death notices is almost certainly Margrith Louisa Maria — it is her mother's name and there is no other candidate. And two Williams eight years apart means the first died. THIS ENTRY'S OWN REASONING WAS RIGHT AND ITS CONCLUSION WAS TOO NARROW: an index that cannot produce four children known to exist will not produce two unknown ones BY NAME. It produced them both by relationship. The old text follows. The baptism register behind the 1868 entry, which needs the record identifier FamilySearch refused on 13 September with HTTP 403. Read it and it either names John Augustus Barry and Margaretha Kolbe or it does not, and either way the question moves. The second missing child has no candidate at all; the same relationship search, run on the father instead of the mother, has not been tried.
15 Who were the Commandant's twelve children? — ONE NAMED FROM A DOCUMENT, AND TWO CANDIDATES SINCE High
- What is known
- The obituary of JOHN GERSON ELIEZAR KOLBE, dead at Leeuwberg on 3 April 1868, says he left “A WIDOW AND 12 CHILDREN”, and The Friend's version — which the Grahamstown reprint drops — counts them: “of the EIGHT SONS AND FOUR DAUGHTERS he has left behind him”. Only one daughter was married and the youngest child was an infant; TWO SONS HAD BEEN SERVING WITH HIM IN THE FIELD. NOT ONE OF THE TWELVE WAS NAMED IN IT. Exactly one has since been reached from a document: the Free State criminal court of 1869, reporting a horse theft at Springfield, ward Upper Modder-river, in October 1868, lists among the stolen animals “1 blue schimmel mare, with foal, the property of JOHN KOLBE, JUN.” — beside a black mare belonging to “ANNA JACOBA VAN DER MERWE, WIDOW OF LATE J.G.E. KOLBE”. A son, holding stock in his own right beside his mother six months after the death. A SEARCH ENGINE OFFERS THIRTY-ONE MORE, born between 1843 and 1869, which is exactly the window between the marriage and the death — AND EVERY ONE OF THE THIRTY-ONE IS TREE DATA, other people's trees indexed. Thirty-one plausible children do not become evidence by being plausible together, and none has been entered. 21 SEPTEMBER 2026 — TWO MORE CANDIDATES, AND NEITHER IS PROOF. South Africa, Civil Marriage Records, 1801–1974 holds ANNA JACOBA KOLBE × FRANCIS McCABE, 13 July 1871, and JAN HENDRIK KOLBE × ROSE MARIA McCABE, Bloemfontein, 4 June 1872 — two of the Landdrost's daughters marrying two Kolbes in ten months, in the town the Commandant's whole public life was attached to, four years after he died and while his widow was married to that same Landdrost. Jan Hendrik is given as 23, born about 1849, which sits correctly among children of a man who married in 1843. BUT THE INDEX CARRIES THE COUPLE AND NO PARENTS. It was read through this archive's own image-data route and the record holds nothing else, so the fathers are on the image and the image has not been read. Until somebody reads that handwriting, these two are Kolbes of the right name in the right town at the right moment and the link to him is borrowed from a user tree. ONE NAMED FROM A DOCUMENT; TWO WAITING ON ONE. VAB MHG M777, his widow's 1896 estate, would name all twelve outright.
- What would settle it
- THREE DOCUMENTS, NONE OF THEM ONLINE. VAB MHG K71, his own estate, and VAB MHG M777, his widow's — she died in 1896 under the name of her second husband, which is why she was hard to find. And now a third and possibly better one: the estate of FRANCIS McCABE, Landdrost of Bloemfontein, who married her about April 1871 and died on 23 February 1872 leaving “four sons and two daughters, BESIDES A LARGE FAMILY OF STEP-CHILDREN”. An executor listing step-children by name would answer this from the opposite side. Also: The Friend and the Grahamstown Journal read forward from 1868 for their marriages and deaths — the method that has already produced John Kolbe junior, a grandson killed by lightning in 1884, and a GERSOM ELIEZÄR KOLBE born in the year his father died.
14 Does Elizabetha van der Merwe exist anywhere but her son's death notice? — THREE ROUTES CLOSED. Medium
- What is known
- She is in this archive on the strength of ONE document: the 1908 death notice of her son, which names her as his mother. Nothing else — no marriage, no baptism of any child, no death, no estate. On 13 September the index was asked three separate ways and answered nothing three times. First her daughter by name: Elesabettha Maria Cornelya returns only her own persona on her mother's 1891 estate. Second her children as a class, any child born 1824–1840 to Petrus Jacobus Booijzen and a mother Elizabetha van der Merwe: other households, none of theirs. Third — and this was the real test — HER SON'S OWN BAPTISM, Hendrik Stef* born 1824–1834, which would have named both parents on one line. Not there. The only return was the 1850 Potchefstroom marriage this archive already has.
- What would settle it
- Not FamilySearch, which has now been asked the question from three directions. The Dutch Reformed registers for the district the couple were actually in during the 1820s and 1830s, read page by page rather than searched — and those are exactly the years and places where indexing is thinnest. This is a door question, not a screen question, and it belongs with the Pretoria and Cape Town letters.
Nothing at that state.
Held as data since 17 September 2026. It was written inline before that, as a list of tuples destructured into five names — and two of the sixteen carry a sixth field, what has been tried, which the destructure never reached. Those two are §0, this archive's oldest wall, and §12, James Montjoy's parents: the page printed their whole what-would-settle-it paragraph where the priority label goes, and showed what had been tried under the heading Where to look. The fields are named now, so a row can gain one without silently shifting the rest.
THE STATE OF A QUESTION IS SET BY HAND, ONE AT A TIME, AND NEVER READ OFF ITS OWN HEADING. Seven headings here announced an answer while the row still carried a priority. Deriving the state from the words would have marked all seven shut, and TWO OF THE SEVEN ARE NOT. §3 — were the women and children in a camp — reports a negative result from one database and names two unread sources that would settle it, the Cape camp registers proper and the Cape refugee and undesirables returns. §10 — did the boys work on the mines — corrects a claim this archive had made rather than closing the question, and still points at mining company registers, street directories, five Transvaal estate files and the Indwe registers. Both stay open at the priority they had. The five that are shut were shut by a document: §5 by the 1903 death registration entering IRISHMAN BORN IN IRELAND, §6 by Yorkshire burials returning zero Barrys at Bilsdale in an indexed run, §8 by there being nothing left on those profiles, §9 by the settler roll, §13 by a relationship search naming both missing children. ANSWERED means a document answered the question as asked; CLOSED, which only §4 carries, means the question itself went away.
Archives identified and not yet worked
NAAIRS — the National Automated Archival Information Retrieval System, South Africa. The single most important unopened door: it indexes Cape (KAB) and Transvaal (TAB) estate files and death notices, which is where a death notice for Willem Hermanus Booyzen would be found.
Cape Archives (KAB), Master's Office (MOOC) — one reference is already in hand: MOOC 6/9/35: 7698 (1845), the estate of George Augustus Kolbe. That one has since been read — it is published in full, and on 20 September it caught an error in a historian’s draft that cites it as his own first source. The notice → The rest of the Master's Office holdings are still unopened.
The Dutch Reformed Church registers for Graaff-Reinet, Cradock, Colesberg, Burgersdorp, Dordrecht and Barkly East — already the backbone of the tree's citations, but read only where FamilySearch has indexed them.
Anglo-Boer War records — camp registers, Cape Rebel registers, Special Tribunal files, the list of the 44 executed for high treason, and the National Scouts rolls that would test the "joiners" half of the story.
FindMyPast and FamilySearch — both hold transcripts already cited in the tree (one FindMyPast transcript reference, R_177176049779, is recorded against the 1812 birth) and neither has been systematically searched for this surname.
Who is in the "Kolbe–Downing" photograph?
A carte-de-visite captioned as George Augustus Kolbe and Margaret Downing cannot be them — the costume is c. 1875–1885, and they died in 1844 and 1867. The likeliest sitter is their son George Augustus Kolbe (b. 1836).
Where to look: the pencilled inscription on the card mount, at higher resolution than the tree holds; the photographer's imprint if the reverse survives; and the descendants who uploaded it. Also: who took it, and where.
Who are the Mountjoys of Barkly Oost? — ANSWERED
Found by accident, on the page opposite a Booyzen baptism: the Doop Register, Barkly Oost, 6 October 1889 records Hercules Johannes Mountjoy and Maria Sophia van Niekerk baptising a son, with Jakobus Jonathan Mountjoy and Maria Sophia Mountjoy among the witnesses.
This line descends from Johanna Catharina Mountjoy (1819–1863), of whom the archive has little more than a surname and a father who died in 1827. Here are three named Mountjoys in the right district, twenty-six years after her death.
Answered They are the descendants of Johanna Catharina Mountjoy's brother. George Mountjoy (b. 1823, Cradock), saddle maker, married Maria Sophia Botha at Cradock in 1848 and had thirteen children. Hercules Johannes (the father at the 1889 font) and James/Jakobus Jonathan (the witness, who died at Barkly East in 1937) are two of his sons. The child baptised that day, George Mountjoy (1889–1963), was a second cousin of the Booyzen daughter baptised in the same register three weeks earlier. The branch →
Now confirmed from originals The parentage no longer rests on a user-edited tree. The Dutch Reformed baptism register records Herculas Johannes Mountjoy, born 16 July 1866, baptised 29 July 1866, father George Mountjoy, mother Maria Sophia Botha. His 1941 death registration gives his age as 74 — a man born in July 1866 is 74 in May 1941. His marriage at Dordrecht on 8 October 1888, aged 22, is a year before the 1889 font. The chain closes.
James Jonathan is "James John." His estate is KAB MOOC 6/9/5047, ref 58379 (1938), and his wife's is the very next file, 58380 — Maria Louisa Bertha née Kumm. Their children's baptisms enter the father under both name forms and spell the mother's surname five ways (Kumm, Kuen, Kunn, Kum, Keem).
Hercules Johannes has no estate file at all. The Cape Town Archives were searched under MOUNTJOY (42 documents; all 22 dated 1930 or later read in full), MONTJOY (4), and MOUNTJOY combined with HERKLAAS, HERCULES, WODEHOUSE and BARKLY (0 each). The estates around him run 1938, 1939, 1943, 1946, 1948 — an unbroken sequence, so this is an absence rather than a gap in the index. The likeliest reading is an estate too small to report.
Tested again, and it holds The Cape probate collection has now been read for every indexed MOUNTJOY at once — 61 personas, the whole surname in the whole collection. Hercules Johannes returns exactly one hit, and it is his mother-in-law's. His name is written on her 1902 death notice, in the list of her children, and the index keys every name on a page. The absence is real, and this is what it looks like from the inside. Her notice →
What is left: the four small corrections for our own tree — Sleer/Slier vs Schleher (the 1863 death notice now settles this in favour of an l), the 1817 Graaff-Reinet marriage, Thomas Jacobus living to 1827, and the missing sibling Jacoba Catharina (1820–1842).
Why was Kathleen Barry a "Military Pensioner"?
Her 1940 death certificate records her, at eighty-two, as a Military Pensioner. Her husband had been dead thirty-five years and died of a stomach ulcer. 13 September 2026 — one candidate found and rejected the same day. A ninth son turned up in the Brixton burial register, RICHARD PETER BOOYZEN, dead at twenty-four on 25 October 1918, in the middle of an epidemic that killed thousands of South African servicemen and pensioned their dependants. His death registration says he was a MARRIED SAWYER of 18 Anderson Street, Turffontein, dead of influenza and pneumonia after eleven days. NOT A SOLDIER. The question stands, and is now sharper for it: the pension belongs to a son who SERVED AND SURVIVED. This archive guessed a son killed in the First World War, with a dependant's pension.
SEARCHED — and the guess is wrong. The South Africa War Graves Project holds no Booyzen among the First World War dead (its two Booyzens died in 1944 and 1993), and of thirteen Booysen deaths in 1914–18, not one matches a son by forename — and most are in the Native Labour Corps, Cape Coloured Labour Regiment and Military Labour Bureau, units not open to a family recorded as "European" on every certificate this archive has read. Decisively: all her identifiable sons survived — dying in 1932, 1952, 1956, 1966 and 1980.
So the pension belongs to a man who served and came home — a disabled son with his mother as dependant, or another entitlement altogether.
Where to look now: the SANDF Documentation Centre, Pretoria (Personnel Archives and Reserves, 20 Visagie Street), which holds Union Defence Force attestation papers, personnel files and pension files from 1912. None of it is online, and it needs a written request. Note also that the CWGC site is behind a bot check and was not searched, and that FamilySearch's South African Defence collection is browse-by-image and not name-indexed.
Request sent, 9 Sept 2026 A letter has gone to the Documentation Centre at both published addresses, asking three things: whether a pension file exists in her name; on whose service it rested; and whether any of her eight named sons appears in UDF attestation or personnel records for 1914–18 or 1939–45, and whether an attestation names his mother as next of kin. See requests/sandf-documentation-centre-pension.md.
The strength of the request is that it runs the search backwards. We have no service number for any son — but we know there was a pension, and we know the pensioner's name, her exact address and her date of death. The pension registers can be searched on those.
Did any of the Bond Street money reach the Cape?
Johan Gottlieb Kolbe's 1842 will divides his residue among his children, with the issue of any dead child taking that child's share, and a later clause names "my said children or grandchildren" outright. George Augustus Kolbe died eleven months before his father. On paper his share therefore fell to his children at Burgersdorp — including Margaretha Kolbe, who married John Augustus Barry.
Where to look: the PCC estate and accounting papers behind the 1846 probate (TNA, PROB series); the Cape estate files for Margaret Downing (estate 4094, 1868) and for the Barry marriage settlement; and any correspondence between Timothy Duke of Lombard Street and the Cape.
Test the deeper Kolbe and Downing pedigrees
The tree carries Kolbe back to 1699 in Württemberg and Downing back to 1683, through the surnames Jung, Larman, Trick, Heard, Hooper and King. None of it has been tested, and Margaret Downing's own death notice names neither of her parents.
Where to look: London parish registers for St Anne's Soho and St Leonard's Shoreditch; Devon parish registers for Trick, Heard and Hooper; Württemberg Lutheran church books for Kolbe and Jung.
It had been sitting in the family tree's own photo store, unread, the whole time. That is the lesson worth keeping: look at what you already have before you go looking for more.