1948

The four-month gap is explained Solved

This archive could not see how Ivone was a spinster on 2 February 1948 and divorced on 19 June. TPD 534/1948 is the answer: her husband sued her. Johannes Antonie Oelofse de Witt versus Yvonne (born de Figueiredo), divorce, brought in the Transvaal Provincial Division and evidently undefended — which is how a marriage of four months was over in four months.

She married a miner in February, he divorced her by June, and she married Kurt seventeen days after that.

And she was not only the defendant New

WLD 2545/1950 — Ivone Luwenski (formerly de Witt, born Figueiredo) versus Kurt. She brought it. Restitution of conjugal rights is the action a deserted spouse brings to compel the other to return, and in South African practice it was the necessary first step before a divorce for desertion.

So this is not one man disposing of a wife:

1948 — he files to end it, weeks after the wedding, before the child is born.
1950 — she files, to make him come back.
1951 — he files again, twice, and this time it ends.

Three years, three actions, one child. Whatever else it was, she fought it.

Read a third time, from the film itself Verified

The 1948 register had been read twice, both times from a photograph attached to the family tree — a copy of a copy. This archive has already been caught once treating a transcription as a document, so on 15 September it was fetched from the microfilm: digital film 008109682, frame 2047 — the index entry and the page image, in South Africa, Civil Marriage Records, 1801–1974. FamilySearch's own pages would not render — the session had been rate-limited to about one response every thirty-five seconds, which the site's front end shows as a hang — so the page was rebuilt locally from its 108 deep-zoom tiles and every cell magnified four to five times, with the sideways columns rotated upright.

One correction, and it is a real one. The address this archive published — 108 Dunstan Street, Braamfontein — is wrong. The film reads 108 Dunottar Street, Sydenham, Johannesburg. Rotated upright at 4.7× the two are not close: Dunottar carries the double t of the Scottish castle name, and the second line is a short word beginning Syd where Braamfontein is long and begins Br. Dunottar Street, Sydenham 2192 is a real street and still is. Braamfontein is the city end of town; Sydenham in 1948 was one of Johannesburg's Jewish northern suburbs. It has been corrected in five places.

Everything the argument rests on holds. Checked cell by cell: 858/1948; Kurt, 36, Berlin, Germany, Bachelor, shop-keeper; Ivone, 24, Portuguese East Africa, Divorcee, supervisor; special licence 164531 of 9 June; and column (10) — read at 5.2×, where the word is four letters — With A.N.C., not Without.

The back of the sheet was never photographed Settled 15 September

Every one of these forms ends with the printed instruction [P.T.O. — and this archive has never seen the other side of one. On 15 September that was settled properly, across three separate films rather than assumed from one:

Fronts only, on all three. It is systematic to how this collection was filmed, not a quirk of one reel, and no amount of patience with FamilySearch will produce the reverses. Whatever is written on the back of any of these registers survives only in the original bound books — which is what request 15 asks Home Affairs for.

Two things the tree copies did not carry New

It was a magistrate's wedding, and nobody was told. The sheet is stamped Magistrate's Office, Johannesburg twice, and the two witnesses — E. W. Coetzee and P. H. Els — are the office's own people, not family or friends. A special licence taken out ten days earlier means no banns were called: nothing was announced anywhere. Two strangers off the corridor signed for them.

“Order Filed” is a physical document, in a second place. A divorcée could not remarry in 1948 without producing the decree, so the remark means the order in TPD 534/1948 was handed in on the day and filed with Johannesburg entry 858. The court file is already on the Pretoria request; this says a copy also sits behind the marriage entry at Home Affairs. Two custodies for one document that nobody has read.

Her stated age is not evidence Rewritten 15 September, later

This panel used to be headed “She aged two years in four months”, and it set the 22 on the Germiston register of February 1948 against the 24 on the Johannesburg register of that June, and said one of them must be wrong. A third reading changes the shape of it, and the old panel was the wrong reading.

The 1953 figure comes from the Salkow register, opened for the first time on 15 September and read at seven times magnification: the second digit has the flat top bar and lower bowl of a five, not the double curve of a three. It is 25.

So this is not a jump. It is a stated age drifting steadily younger, and against her real birth date not one of the three is right. Her own birth registration, supplied by the family in September 2026, gives 4 March 1922 — confirmed independently by her death certificate. She was 25 at Germiston, 26 in Johannesburg that June, and 30 when she married Salkow, and she gave 22, 24 and 25.

This panel’s own rule is what saved it. It said that any inference resting on an age Ivone gave a clerk had to be re-checked against a birth record rather than against another register — and when the birth record arrived it destroyed all three figures, including the one this panel had called “the only one that fits”. That entry fitted a birth year that was itself wrong by two years. Two wrong numbers agreeing is not corroboration.

The consequence is a rule, not a curiosity. Any inference in this archive that rests on an age Ivone gave a clerk has to be re-checked against a birth record rather than against another register. Kept and rewritten rather than deleted, because the first version of this panel was published and people may have read it.

A warning about reading backwards Method

Everything on this page was in the record before the DNA was looked at, and none of it was thought to mean anything. That should make anyone careful. A timeline read in the light of a conclusion will always seem to support it.

So: the dates here are documented and the inference is labelled as an inference, and they are kept apart on purpose. If the Y-DNA came back Ashkenazi, this page would stay exactly as it is and the conclusion would go — a marriage in trouble from the first month, and nothing more.

It came back on 14 September 2026, and it is not Ashkenazi. Derrick's paternal haplogroup is R-S11477 — R1b, branch U106/Z381, the North Sea lineage that peaks in the Netherlands and England. The page still stands on its documents and the conclusion still stands apart from them, exactly as it did before. But the test this page named as the one that could take the conclusion away has been run, and it did not take it away. What it proves and what it doesn't →