Dufau

The same name, spelt two ways

One systematic substitution — s for z, w for v, ch for h — where BOTH spellings are independently attested surnames. Kosina and Kozina are one man in two clerks' handwriting. Stronger than sounding alike, weaker than somebody having checked.

Duffau

What Wikidata says this name is

Copied from Wikidata, not worked out here. It is crowd-sourced and wrong in places, which is why the item is linked below rather than the claim simply repeated.

• French the language of the name

  Which is not where a family is from. A Polish surname can be borne by people whose forebears never saw Poland, and often is.

Q23793509 on Wikidata ↗ — the item every line above came from.

It is also the name of a place

This is as close to an origin as this map will go. A surname that is also a place name often came from it — and often did not: plenty are occupations or patronymics that happen to collide with a place, and plenty of places are named after a family rather than the other way round. It is a fact with a source, which is more than an etymology page offers.

• Duffau United States of America a spelling of the name

Where it is borne, and where a record names it

Country by country

Each country below can carry up to three different kinds of evidence, and none of them is the same claim. A register is a head-count — a state counted these people. A record is one document naming the surname on that ground, narrower evidence and often much older. And a volume overlapping these dates is a document from the right place and years — never a document known to name this surname. None of the three ever means the name exists in that country and nowhere else, or that it exists nowhere but here.

When it was being born, in France

INSEE counts births by surname and decade from the état civil, 1891 to 2000 — the only dated evidence in this dataset. A name that starts at zero and climbs arrived; a name that falls to nothing left, died out, or emigrated, and the decade it stops is the decade to stop searching. It counts births registered in France, so a family that emigrated in 1900 simply stops appearing, which is information rather than a defect.

• 1891–1900 286
• 1901–1910 413
• 1911–1920 361
• 1921–1930 484
• 1931–1940 371
• 1941–1950 478
• 1951–1960 508
• 1961–1970 462
• 1971–1980 398
• 1981–1990 341
• 1991–2000 315

Whose archive held that ground, and when

Every surname site can tell you a name is common somewhere. This one also knows which filing system held that ground in which years — and those are the two halves of one question: my people were called this, they were there, so who has the paper, and for which century. A name in Croatia before 1881 may sit in a military series and not a county one; a name in the Netherlands before 1811 has no civil record at all, only a baptism.

This is about the ground, not about your family. A register counting a name in a country says nothing about which village. And these are the pieces of a country this atlas has drawn, which in a large one may be a corner of it — Acadia reaches into Maine and is not a statement about the United States, and Bohemia's ground crosses into Poland. Each region's own page says which ground it covers. Where a country is missing below, nothing covers it yet.

No record region is drawn yet for Spain — which is a gap in this atlas and not a statement about that country.