← The land

Story · 1777–1901 · v2.11

From Petersburg to Klaarstroom, a history since 1777

The poort opens in 1858, twenty-seven erven are laid out on a loan farm of 1777, and they are sold without water rights because the buyers meant to live off the traffic and not off the land

The loan farm De Claar Stroom (Klaarstroom No. 178), granted 1777, is the ground Klaarstroom village stands on, not to be confused with Klaarstroom No. 184, which is Christian Hatting's grant 19 km west. Meiringspoort opened in 1858; twenty-seven erven were laid out on Petrus Oosthuizen's land and the village was called Petersburg for him. The erven were sold WITHOUT water rights. Who Petrus was is still unknown, and the one testable fact about him is a brother named Ockert.

Two farms with one name

There are two Klaarstrooms in the Prince Albert registration division and confusing them is easy.

One thing to know before the rest of it. Almost everything below rests on a single tertiary source, a Karoo Cameo written for a development foundation rather than an archive. It is used because it is what exists, every claim taken from it is marked, and the deeds have not been read.

Klaarstroom No. 184, "Klaarstroom A", is 2,228 ha of what the 1833 diagram calls Government ground, surveyed for Christian Hatting Senior in August 1833 and titled on 15 April 1841 under deed Bf.W.Q.5-17, diagram 354/1833. It lies about 19 km west of the village.

De Claar Stroom, Klaarstroom No. 178, this farm, is 7,259 ha and is the ground the village actually stands on. It is easy to miss precisely because the village carries the name: for most purposes "Klaarstroom" means the settlement, and the farm underneath it goes unmentioned. Middelwater's eastern neighbour, named "Klaarstroom" on the 1833 diagram, is this ground.

The distinction matters beyond bookkeeping. Hatting's 184 is a grant of Government ground to the elder Christiaan Hatting. 178 is a loan farm, older, vaguer, and the kind of holding the whole district was built out of.

The loan farm

The loan farm De Claar Stroom was granted in 1777, clear stream, which is what the place is: the water coming down off the Swartberg at the northern mouth of the poort. It is why there is a settlement here at all, and why everything in this history gathers around this basin.

Between the 1777 loan and the 1841 grant one holder is now on record by name: the Beaufort opgaafrol of 1822 carries De Claar Stroom in its leenings-plaats column against Hermanus Carel du Preez [SRC-558] — not an Oosthuizen, which is worth knowing exactly because the village years make this feel like Oosthuizen ground from the beginning. It was not; the family's title here comes with the grant era.

The cameo describes an extent that is worth pausing on: the loan farm "stretched from the modern farm called Middelwater, at the present-day tarred turn-off to Prince Albert, to Vrolikheid, 10 km down the gravel road to Willowmore". If that is right, De Claar Stroom is the parent ground of both Middelwater No. 169 and Vrolikheid No. 177, the two farms this family's story begins on.

That is a large claim resting on a small source, and it is not yet verified. Loan-farm extents were described by walking, not surveying, and a loan place was a grazing right over a radius rather than a boundary. What can be said is that the three farms are contiguous and that the 1841 quitrent conversion titled them separately.

Meiringspoort, and why a town appeared

Nothing was going to be built here until the mountain could be crossed.

Meiringspoort opened on 3 March 1858: twenty-two kilometres, twenty-five river crossings, built in 223 working days. It is named for Petrus Johannes Meiring, the Klein Karoo farmer who found the route with two friends, Gerolm Marincowitz and a Worcester man named Benecke, and then campaigned twenty years for an official road. It opened with three hundred mounted men and fifty coaches riding through to De Rust; the champagne broke on the third attempt.

By June 1870 a million kilograms of wool had gone through the poort in a single season. What that traffic was worth to the village showed the day it stopped: a flood closed the poort in 1885, stranded fifty transport wagons at Klaarstroom with their loads spread across the veld, and it took Thomas Bain until 1886 to repair the road and reopen it.

The traffic is the town's whole reason for being, and it explains the one detail in this story that could not be invented.

Petersburg

Soon after the poort opened the ground was surveyed and twenty-seven erven measured out. The first sold on 4 August 1860.

The cameo is explicit that the ground was Petrus Oosthuizen's, and that the village was named Petersburg, partly for the several Pieters with a hand in it, but "named mainly as a tribute to Petrus Oosthuizen", who "had seen the need for a settlement and gone out of his way to take the first steps". And "the fledgling town kept Petrus Oosthuizen and his brother, Ockert, very busy ensuring that all their stipulations were adhered to".

The diary of Almero Oosthuizen of Droogekloof, born 1878, adds the one inside view of the village this record has, in the years it still carried the old name — he writes it Pietersburg: a hotel and one business, the building where mnr. L. van Zyl later kept his shop, owned by R. [?Faesie] & Kie under a manager Bernhardt — and the hotel lost its licence when ds. A. G. du Toit became predikant at Prince Albert [SRC-586].

The erven were sold without riparian rights, dry erven. In a Karoo village that is close to unthinkable, and it is the detail that proves the town was a creature of the road rather than of the land: the buyers meant to live off the traffic through the poort, not off water they would never get. It also tells you the seller understood exactly what he was keeping.

Petrus Oosthuizen

He is the man this village is named for and he cannot yet be placed in the family.

The cameo gives him no dates whatever. A birth about 1798, a death on 2 January 1876 aged 77, two wives and twenty-one children are sometimes attached to him, and every one of those facts belongs to Petrus Johannes Meiring, five pages away in the same document. Anyone reading the cameo straight through will make the same join. What is actually left is a name, an ownership, and a brother called Ockert.

The brother is the best lead in it. Johannes Petrus Oosthuizen (b.1798) is the obvious candidate on the name, and he fails on the only testable point the record offers: his six recorded siblings include no Ockert. Petrus-and-Ockert as a brother pair is a strong signature in a family that names by rule, and it is what the next search should be built on, not the name alone, which is never enough on its own. See LEAD-0029.

A second candidate now passes that test. Pieter Marthinus Oosthuijzen, born in 1796 or 1797, was twelve when his father died at de Modder Fonteyn on the Olifants River and was one of the seven children placed with his eldest brother.

Both halves of that are now on paper. The Roodezand register enters him under 1796: 't kind PIETER MARTINUS, ged: den 27 Maart, geb: den 9 Junij [17]95, ouders OKKERT OOSTHUIJSEN, SUSANNA CATRINA VAN TONDER. [SRC-518] And the George roll of 1813 carries him beside his eldest brother's household as a single man, four years after the estate, in the same block as two other brothers. [SRC-523] So the placement held and he was still there at about eighteen.

His brothers include an Ockert Almero, who is very likely the "Okert Oosthuizen" the neighbouring farm Middelwater was surveyed for in the same month of 1833. Piet, Pieter and Petrus are one name. That household is told in From Modder Fonteyn to the Koup.

Neither row carries a date the other can be tested against, so nothing is joined. It is a candidate that survives its first test, which the other one did not. See LEAD-0215. not yet verified

His estate file in the Master's series would settle the dates and the parentage in one sheet. It has not been pulled.

The commando

On 1 February 1901 Gideon Scheepers rode into Klaarstroom with Fouché and some three hundred men, led in by Lieutenant Fanie Swanepoel, and raided the post office and the government stores. The cameo adds that Deneys Reitz was with them and describes it in Commando. He does not. The memoir has now been read against the claim [SRC-437]: it does not name Klaarstroom, Prince Albert, Meiringspoort or the Kredouw anywhere in its 624,000 characters, and Reitz never rode with Scheepers - the three times Scheepers appears, Reitz is learning at second hand, from captured court-martial papers and from talk, that the man had already been taken and shot. Reitz did cross the Swartberg, but southward towards Oudtshoorn with Smuts, and in October 1901, eight months after this raid. Helena Marincowitz puts that crossing on Vrolikheid or Oorlogskloof, from a farm two ridges away and with a different commando behind him. [SRC-172] The next day Lieutenant-Colonel R. K. Parke sent two patrols and five howitzers up the Kredouw pass from Prince Albert to drive them out.

What followed at Remhoogte is the war in miniature. The Boers caught the retreating Yeomanry and shot the horses from under them. A wounded lance-corporal sheltered at Wolwekloof; Johannes Kleu, a man of twenty-two or twenty-three, came upon him and was shot through the chest. Whether he was walking home to Aberdeen unarmed, as the cameo has it, or one of the commando, as Marincowitz and the eyewitness both have it, is not settled. not yet verified Scheepers's men slept that night at Bloemendal and slipped north four days later by Boesmanskloof, Sleutelfontein and Botterskraal to Zeekoegat.

The war story, The war reaches the Koup, carries the eGGSA record's different telling of the same chase: Prince Albert Mounted Troops surprised in the village rather than Yeomanry sent to clear it, and Kleu a man of the commando, aged twenty-two. The two accounts disagree and both stand, each said in full on its own page.

Four of those farms are here. The commando's line of retreat is a tour of the family's ground.

What this story is about

Sources

Everything this story rests on, with a link to the original wherever one exists. Where a citation is incomplete it says so, an incomplete citation is a research gap, not a formatting problem. If your work is here and the credit is wrong or thin, please tell us and it will be fixed.

  1. Surveyor-General diagram, 354/1833, August 1833 Surveyor-GeneralSG Dgm 354/1833
  2. Rose Willis and Doreen Atkinson, with Judy Maguire and Bruce Rubidge, "Klaarstroom: Ancient place of clear waters" (Karoo Cameos, 2021) Karoo Development Foundation, Karoo CameosKaroo Cameos, 2021view it online →
  3. Klaarstroom No.184 original diagram (1833) Surveyor-GeneralSG Dgm 354/1833
  4. Andrew Geddes Bain and Thomas Charles John Bain - Meiringspoort and the Swartberg Pass web (Swartberg Circle Route; SA Country Life; Wikipedia: Swartberg Pass)view it online →
  5. Helena Marincowitz, Prince Albert and the Anglo-Boer War (1899-1902), Fransie Pienaar Museum, Prince Albert, 1999 Fransie Pienaar Museum, Prince AlbertMarincowitz, Helena (1999): Prince Albert and the Anglo-Boer War (1899-1902). Prince Albert, Fransie Pienaar Museum. 51 pp. Read in full from pp.4-51; pp.1-3 (cover, title, contents) not photographed.
  6. Surveyor-General diagram, 350/1833, August 1833 Landmeter-Generaal, KaapstadNo. 350/1833; surveyed August 1833 by C. Lennox Stretch, Land Surveyor; copied from the diagram relating to Title Deed No. P.A.Q. 5-8 dated 15.4.1841; endorsed 27-4-1959; scale in Rhynland Roods
  7. Surveyor-General diagram, A.942/1858, February 1858 Landmeter-Generaal, KaapstadNo. A.942/1858; surveyed and drawn from actual admeasurement February 1858 by Frederick Wm. Barry, Sworn Surveyor; examiner L. Marquard; copied from the diagram relating to Transfer Deed No. 45 dated 3.6.1858 (Vol. 23); endorsed 19.3.1965; scale in Rhynland Roods
  8. Roodezand (Tulbagh) baptism register: the baptism of Pieter Marthinus, son of Okkert Oosthuijsen and Susanna Catrina van Tonder, born 9 June 1795, baptised 27 March 1796 FamilySearch, "South Africa, Dutch Reformed Church Registers (Cape Town Archives), 1660-1994". Project S-AF 00532, roll 87.Film #008121003, ITEM 3 = TULBAGH G4/1/1, BAPTISM, 1743-1800. Image 416 of 779; page 333, entry No. 30, year band 1796. The volume's title page is image 185: 'Het boek der Gedoopte kinderen De Kerkgemeente van 't Land van Waveren, sedert 't jaar 1743'. The film's item table also lists item 1 PAARL G3/4/2 death 1824-1859 (99 pp) and item 2 PAARL G3/6/3 death 1876-1883 (97 pp).
  9. GEORGE opgaafrol J 90, 1813 - the foster household again, with a third brother now enrolled in his own right FamilySearch, from the Cape Archives J-series. Filmed by the Genealogical Society, Salt Lake City, at Gov Archive's Cape Town RSA.Film # 008145124, images 165, 181 and 182 of 545. TARGET CARD: LOCALITY OF RECORD GEORGE / CAPE PROVINCE; TITLE OF RECORD OPGAAFROLLEN / CENSUS - TAX; VOLUME/S J 90; YEAR/S INCLUDED 1813. Filming block: OPERATOR C MARAIS; REDUCTION X 14; DATE FILMED 9 MAY 1978; FILM EMULSION NUMBER 5460-394-361; FILM UNIT SER. NO. 3385; PROJECT NUMBER S-AF 20 B; ROLL NUMBER 12. Same film and roll as SRC-522 (J 89, 1811), so this film carries at least two consecutive George volumes.

Research notes

Written 31 Jul 2026 at Frans's prompt, "why is there no information in klaarstroom no 178? surely with all the source, you must be able to build a story?" The parcel had been added hours earlier and left bare while eleven events, seven sources and twelve people about this ground sat in the model unattached, because events reach a farm only when their text names its PAR id. Leans almost entirely on SRC-010, a Karoo Cameo, which is tertiary; the deeds have not been read and the story says so twice.

[31 Jul 2026] kind normalised 'farm history' -> 'farm_history'. Two spellings of one kind had split the farm histories across two groups on the stories index, showing 2 where there are 5. Underscore is the house form; the other kinds are single words so nothing else was exposed.

[5 Aug 2026, moved out of the prose] . Removed: 'It was in the model as a name in three boundary descriptions and nothing else until 31 July 2026, when it was finally given a parcel row. Before that, Middelwater's eastern neighbour Klaarstroom pointed at a place id, because there was no farm to point at.' That is this repository's data history. The page now says why the farm is easy to miss - the village carries its name.

[12 Aug 2026] VOICE REWRITE to STORY-VOICE.md, both languages. en 1/3/0 -> 0/0/0, af 3/3/0 -> 0/0/0 (narration/bold-sentence/punchline).

THE FLOOD WAS LINKED AND NEVER MENTIONED. EVT-000211 sat in the frontmatter and appeared nowhere in the prose. It is the best evidence the page has for its own argument: the town existed for the traffic through the poort, and when the 1885 flood closed the poort fifty transport wagons stood at Klaarstroom with their loads spread across the veld until Thomas Bain reopened the road in 1886. Now in both languages, sourced to the event row.

PROCESS ARCHAEOLOGY OUT, THE WARNING KEPT. The Petrus Oosthuizen section narrated its own correction - 'An earlier reading of this project's own, mine, gave him a birth about 1798... and all of it has been withdrawn'. Per 10 the page now states the settled picture: the cameo gives him no dates, and the 1798 birth, the death on 2 Jan 1876 aged 77, the two wives and the twenty-one children belong to Petrus Johannes MeiringM five pages away in the same document. THE WARNING IS NOT DROPPED, because it is about the SOURCE and not about us: anyone reading the cameo straight through will make the same join, and the page now says so. PER-000517 Meiring is linked at that point, which he was not before. Likewise 'was proposed as a match and has since been withdrawn' for PER-000513 -> 'is the obvious candidate on the name, and he fails on the only testable point the record offers: his six recorded siblings include no Ockert', which is what the record itself concludes.

'Two farms, one name, and the reason this page exists' -> 'Two farms with one name'. 'Who Petrus Oosthuizen was, which is not known' -> 'Petrus Oosthuizen', with the first line still saying he cannot be placed. 'What is missing' -> 'What has not been worked out' so the research tail renders as a panel and the prose ends on the commando's line of retreat. 'the Klaarstroom basin is the anchor of this project' -> 'why everything in this history gathers around this basin', which says the same thing without the project naming its own scope rule.

Bold came off three links and off the farm name; the emphasis that stays is on 'village actually stands on' and 'nineteen portions', which are the two things a reader scans this page for. Every [TO VERIFY], the tertiary-source warning at the top, the 1,700 ha discrepancy and the LEAD-0029 pointer are unchanged. verify_af passing.

[14 Aug 2026, SRC-437] THE REITZ CLAIM IS SETTLED AND IT IS FALSE. The page now says so instead of hedging, in both languages. The [TO VERIFY] is removed because the document that was named as the settle has been read, which is the only thing that removes one.

[17 Aug 2026] MINOR, v2.3, both languages: a second candidate for Petrus - Pieter Marthinus OosthuijzenM PER-000146, who unlike PER-000513 does have a brother Ockert - recorded as a candidate and not joined. No restructuring; verify_af passing 1/1. Done as part of writing STO-0042, per STORY-VOICE.md - a story is not finished until the pages it links to tell the same history.

[17 Aug 2026] v2.4: the Petrus of De Claar Stroom is now Pieter Marthinus rather than a separate record. py. No fact changed; the men are the same men, and the hedge and the [TO VERIFY] stay because the diagrams still name nobody. LEAD-0213 remains open on the Vrolykheid patronymic.

[17 Aug 2026] MINOR, v2.5, both languages: the STORY-VOICE.md pass of 17 Aug - the caveat lead moved or removed where it was housekeeping, and citations rewritten so the source is a noun in the sentence rather than a bracket after the full stop. NO fact, hedge, number, quotation or [TO VERIFY] changed anywhere; verify_af passing.

A LIMIT WORTH RECORDING: the worklist was built from the ENGLISH files, so the Afrikaans was corrected only where the same passage happened to be edited. The Afrikaans side still has its own unattributed tags and needs the same pass. verify_af CANNOT catch this - it compares facts, hedges, numbers and ids, and attribution STYLE is none of those.

[25 Aug 2026, currency review] v2.6: the 1822 du Preez holding, the Petrus link corrected to his own row, the diary's Pietersburg inside view.

[27 Aug 2026] AFRIKAANS LANGUAGE PASS, v2.7: long sentences broken, tone particles and the gedagtestreep out of our own prose. NOT a content change; verify_af confirms every fact of the English twin survives and the English is untouched.

[27 Aug 2026] RESEARCH SWEEP, v2.8: BOTH HALVES OF THE PIETER MARTHINUS CANDIDACY ARE NOW ON PAPER. The Roodezand register enters him under 1796, born 9 June 1795, son of Okkert Oosthuijsen and Susanna Catrina van Tonder (SRC-518); and the George roll of 1813 carries him beside his eldest brother's household as a single man, four years after the estate, in the same block as two other brothers (SRC-523). So the placement held and he was still there at about eighteen. The candidacy is still a candidacy and the [TO VERIFY] stands.

v+ 28 Aug 2026: title and subtitle set in the title review. The title now carries the village's own name change, which the page's best material is about. *** [7 Sep 2026] v2.10, both languages. *** TWO CORRECTIONS from SRC-172. Reitz crossed in OCTOBER 1901, eight months after the raid, not after Scheepers's execution a year later. And Kleu's status is now flagged as disputed rather than asserted as an unarmed man walking home.

[17 Sep 2026] minor, FACT CORRECTED: Christian Hatting SeniorM (PER-000055) was very probably NOT the field cornet. The Beaufort roll of 1822 (SRC-559) lists the field cornet (line 1) and 'Christiaan HattingM d'oude' x Maria Christina VenterM (line 10) separately, and the roll of 1824 (SRC-634) names the field cornet 'Christiaan HattingM Diedriks Zoon' on Kleine Valey, the Christian Hatting JuniorM of Klein Valie. Senior is Christiaan HattingMh (KPSQ-FKH, 1772-1848), whose wife died on 'her own Venterskloof' in 1875 (SRC-660).

The research behind this page, kept separate from it: dated working notes, open questions, and what is not yet settled.