The Map

Only 122 of the Žubrinić records name a place at all, and they fall into six of them. Frontier registers identify a household by village and house number rather than by surname, so the house numbers are kept here: they are the closest thing these records have to a family identifier.

The Croatian record stops in 1888

Set the Croatian records against the Australian ones by decade and the two curves cross once, in the 1880s — and then the Croatian side simply ends.

A hundred births and seven deaths

The Croatian records are almost entirely baptisms. Across the whole surname there are about a hundred dated Croatian births and seven recorded deaths — and thirty-three of the births fall in the 1840s alone.

That is not what a family looks like. It is what one decade of one register, read hard and then abandoned, looks like. The Otočac death books have barely been opened.

The seven deaths, in full

Five of the seven are at Senj, and three of those five are children — aged eight, one and nought, in 1861, 1861 and 1863. They are from three different households. Whether that is a hospital, an epidemic, or simply where the town's register happens to have been read is not something this archive can say.

Five deaths nobody wrote down

A Catholic family of this period reused a given name when the child who carried it had died. So a sibling set holding two Franciscas is, by itself, a record of a death — even where no death entry has ever been read.

One household, that of Joannes Žubrinić and Maria Kostelac, shows it twice: eight recorded children between 1844 and 1853, of whom two names are used again — and their son Nicolaus, born 1853, is one of the three children who die at Senj. Of eight births, three deaths, and only one of the three written down anywhere.

Where the deaths were written down, and where they were not

Re-run 10 September 2026, after four sorting buckets were found being published as ancestors and removed. Every figure above is unchanged — the buckets were already excluded by this tool, which is why it reported them as excluded and the site build did not. The Blažević family count, left blank when this table was first drawn, is 63.

The zero is the interesting number. It does not mean the Papići lost no children. It means their losses were written down: Milka's sister Zora died in 1899, the year she was born, and her brother Stanko in 1903, aged one — both recorded as deaths, not merely implied by a reused name.

The Papići are a Senj family, and somebody read the Senj death register. The Blaževići and Žubrinići are village families whose death books have barely been opened — so their infant deaths survive only as a name used twice. The pattern in this table is a map of where the research went, not of where children died.

Checked 13 September 2026, and it cannot be closed from here. There is no village death register in this archive at all — not one page of Krmpote, Smokvica or Otočac. The only death records held are two pages of the Senj book and five gravestones. So the zero in the Papić row is an inference from an absence this archive cannot yet test, and it stays marked as one until somebody opens a village book.

And one page already argues against it. The Krmpote baptism register for 1894 carries, as numbered entry 8 with no name and no baptism, “12 veljače porodila je Vica Blažević … iz Starog sela k.b. 86 mrtvo žensko dite” — a Blažević stillbirth, recorded directly. Village losses were not systematically unwritten, so the inference in this table should be weaker than the zero makes it look. The Gospić letter now asks for a sample of any of them, any decade, with the undertaking that if the villages recorded their deaths as fully as Senj did, this finding is withdrawn. What has and has not been opened →

And it catches an error on its own

The same scan turned up a sixth repeated name that is not a reused one: Nicolaus (1834) and Nikola (1880), entered as brothers, forty-six years apart. No mother bears both. It is the same faulty attachment that put a woman born in 1835 forward as the mother of a son born in 1834. The correction →

Three further pairs share a name and a birth year — Catharina 1840 twice, Josephus 1845 twice, and Mihovil 1802 beside Michael 1811 with both deaths recorded. Those are almost certainly one child entered twice, from two passes over the same register under two spellings of the name.

Four surnames, four different shapes

Run the same reduction over each name and they do not look alike at all. One sits in two villages; one sits in three houses of a single village; one leaves for a single port on the other side of the world; one is simply everywhere. A surname's map is a fact about the family, and these four maps are four different kinds of family.

Two houses that only a daughter's wedding records

Everything above is reduced from the tree. But a marriage entry names the bride's parents, their occupation and their house — so a woman's wedding is often the only place a dead mother's maiden name and address were ever written down. Reading eleven collateral marriages produced nine addresses that appear nowhere in the export.

Neither can be added to the counts above without inventing a record, because the tool reads the export and these are not in it. They are set down here instead, as the two places to look next. Open questions →

How this was produced

tools/gazetteer.py reduces every place string to a village and, where the register gives one, a house number. It ran on Žubrinić alone until September 2026; it now runs on all four surnames, which is what makes the shapes below comparable. It also folds the tree's spellings of one village together — Otocac Croatia with Otočac, Kriviput with Krivi Put — without which a village looks half as populated as it is. tools/name_reuse.py finds repeated given names among siblings, folding the Latin and Croatian forms of a name together (Joannes = Ivan, Georgius = Juraj), excluding the tree's sixteen sorting buckets, and separating three cases: reuse after a gap, the same year twice, and a gap too wide for one mother. Both write their output to notes/. Method →

Klenovica 22 and Povile 22 are one house

Kalanj was missing from this gazetteer entirely until today — 216 records, the second-largest family in the archive, and the best-placed: 187 of them name a place, 87%, against Blažević's 77% and Žubrinić's 43%. Which makes it the right family to ask the house-number question of.

Eighty-five published people carry a Klenovica or Povile house number, and every one of those houses is entirely Kalanj — not a single house in either village is shared with another surname. That is already unlike Selce, where the whole question was whether two families shared one.

But the two largest looked like separate houses and are not:

No household moves house nine times and alternates back. Povile and Klenovica are neighbouring settlements on the same stretch of coast, and the registers — or the people transcribing them — used the two names for one place. Eleven couples in all have children recorded in both villages.

So house 22 holds 67 published people between 1840 and 1890, not the 43 this data appeared to show. It is the largest single household in the archive, comfortably ahead of Smokvica 114 and Krivi Put 136 — and it is the zadruga this archive has been describing, seen whole for the first time.

One caution. Luka's last two children are at Klenovica 36 and 37, in 1892 and 1894, after four at house 22. That reads as a real move to a new house, not a second name for an old one — so the argument above should not be stretched to say every Klenovica number is interchangeable. It is specifically 22 and 22.