The Map
One map and four layers. Where are the record books and how far have we got with them; where can this archive put a named person; where is the family actually in the ground; and where does a record merely name somebody of the name on one day. Switch any of them on or off. A place that answers more than one question is still one dot, and it wears the colour of the rarest answer it has — so with everything on, the records, the graves and the people stand out of the shelf rather than hiding under it.
Red through amber to green is the shelf, and red means nobody has opened it Violet and pink is our people Blue is the ground
1245 + 165 + 69 markers are 1394 places. 96 of them answer more than one question and 20 answer all three: Rijeka, Ližnjan, Pazin, Poreč, Omiš, Split, Pula, Rovinj, Bakar, Kaštelir, Sovinjak and Tar.
Rijeka is the case for merging. On one page it now says: sixty-nine register volumes catalogued and not one of them opened; ninety-six people the archive can put here; and twenty-three burials. Those were three separate facts on three separate pages, and the third of them — that we have twenty-three burials in a town whose registers we have never read — is only visible when they sit together.
The layers change the colour. They never change the panel. Click any marker, with any combination switched on, and it tells you everything it has: the registers, the people, the graves and the records together.
The fourth layer is a different kind of claim, and it is drawn that way. The shelf, our people and the ground are all statements about where this family is. Findmypast is not: a census line, a landing, a marriage licence is a record of somebody of the name at a place on one day. 156 such records stand on 31 places, in amber, and the panel calls them what they are — named in a record here, not who the archive puts here. A 1913 arrival at Ellis Island must never be allowed to look like a home. A further 21 records name only a state or a country and get no dot at all, for the same reason.
A place that answers two questions is still one dot. Drawing it twice is exactly what merging the three maps was for. It takes the colour of the rarest question it answers among the layers that are on — the ground before our people, our people before the shelf. There are 1,235 shelf places, 167 with people and 61 with graves, so the rare thing is the thing worth seeing and the common thing is the ground it sits on. Rijeka answers all three and wears the burial colour; Gologorica answers two and wears the line's; Senjska Draga answers one and is sand.
Every view of this map is a link. The layers, the basemap, whatever is typed in the find box, which filters are switched off and where the map is looking all live in the address bar, so a view can be sent to somebody rather than only remembered for you. Every parish where a book has actually been opened is six places and one link; so is the ground alone, on satellite. The button says Copy link to this view.
It was a switch for a day — one question at a time, and asking a second put the first away. David: «can our map rather layer the shelf, our people and the ground?»
The old addresses still work. /research-map/, /atlas/ and /graves-map/ all land here.
And there was a map before any of these. The Google My Map this family drew over years, imported whole from its KML and kept exactly as it was drawn — pins the Atlas has no place for, the movements David traced by hand, and the links that have since died. It is not a fourth layer, because it is not this archive's reading of anything: it is the map that came first, kept as a record of what was already known before any of this was built.
The six colours here are not coverage states. They say how the archive knows a place — the documented line, Carnia, the Istrian sweep, the Senj and Karlobag side, the diaspora, and other families of the name. Weight on this map is not kinship: a parish can hold forty people of the surname and be connected to this family by nothing whatever.
32 of these towns also carry a shelf, and 20 carry all three questions at once. They are one marker, not two: click it and it tells you the registers, the people and the graves together.
Nothing here is geocoded twice. The three maps this page replaced each resolved their own place names; they now share one set of places, and the coordinates come from a single gazetteer committed to the repository rather than from three different files in three different states.
Living people are on this map by name and by town, and by nothing else.
Where this map comes from, and what it is worth
So it was exported and rebuilt here. That export is where the trouble starts. The KML holds 1,709 placemarks, but only 247 of them carry coordinates — and of its 1,196 record-book places, only ninety do. The rest are address pins, which Google geocodes when it draws the map and not before. As a backup, the export is about eight per cent of a map.
What every pin does carry is a FamilySearch browse waypoint — 1,194 of them — and a waypoint is the key that opens a parish's shelf. That turned out to be the last thing his map was needed for. Each confession root in the collection lists its own parishes, so the list is now walked from FamilySearch's own tree instead: 1,333 shelves, against his 1,288, and Ždala through Zvoneće that his export never had because the root returns exactly one thousand and stops.
His map still supplies what the tree does not: the confessions, drawn as he separated them, and ninety coordinates. The rest of the coordinates came from the GeoNames country dumps for Croatia and its seven neighbours — 1245 of 1286 places placed, without a single geocoding call.
41 places are still unplaced — they are mostly compound parish names like Split-Lučac or Sv. Jakov-Šiljevica that no gazetteer carries as one string.
That is the honest picture and it is meant to be uncomfortable. 1252 places catalogued and untouched · 23 researched without their registers being opened · 6 with a volume of their own read · 0 finished · 18 volumes of 7771 read one by one.
Teal is new, and it exists because gold was lying. Twenty-two markers were gold and twenty of them had never had a single register volume opened. They were gold because the search register NAMES them — a local history of Žminj; a migration paper that lists Žminj among the towns the Carni settled at. Reading a book about a parish and reading the parish's own registers are two different acts, and a scale whose legend says «how far we have got with the shelf» must not paint the first in the colour of the second. Logged as a correction.
This is the third scale this map has worn. It carried three coverage states until the cataloguing finished and turned 1,219 of them the same colour; then it shaded by depth — one or two volumes against ten or more — which looked better and said less. A parish with forty volumes nobody has opened and a parish with two nobody has opened are in the same state. Depth now lives in the panel, in words, next to the sentence that says whether anybody has read it.
«Read right through» can only be claimed where this archive knows which volumes were read — which today is Antenati alone, because its catalogue carries a per-volume flag. Everywhere else a reading is recorded against the place, not the book, so the best that can honestly be said is «started». That is on the work list.
And the map cannot see what is filed abroad. A Croatian town's registers sit wherever the state that kept them filed them: Čakovec is under «Csáktornya», in a Hungarian county, inside a collection named for Slovenia; five Croatian Istrian parishes are inside an Italian one. Croatia, filed under other countries lists what the walk of all 3,488 FamilySearch collections turned up.
Grey is «in no catalogue at all», and every grey marker on it is now a checked negative. They are places this archive walked the FamilySearch catalogue for and found nothing — Grožnjan, Lindar, Motovun, Višnjan, Villa Santina and Pirovac. A checked negative is worth more than a blank.
Two of them used to be grey for a worse reason. Brekinjska and Hrastovac were Evangelical, and this archive had written that it «never found the Evangelical root in FamilySearch's waypoint tree» — the shelf was not walked, not proved empty. The root was found on 15 September and it was never hidden: the collection's own waypoint list names it — Evangelical (Evangelička crkva), one of ten confession roots. It holds six parishes and seven volumes, Brekinjska and Hrastovac among them, and all seven are now openable from this map. The lesson is the honest one — the archive had not looked in the place the collection tells you to look.
The first cut of this map used three colours — not walked, catalogued, read. It worked for about a day, and then the cataloguing finished and 1,219 of 1,239 places went gold. A map where everything is one colour is not a map, so the shading was re-cut to carry something that still varies.
Researched, books unopened means a row in the search register names the place as a source or a destination — and no volume of its registers has been read. That flag is derived by matching place names against the register's own source and destination fields — never against its prose, which mentions places for all sorts of reasons. A handful of names that are also ordinary Croatian words or surnames are excluded by hand. It is approximate, and it will be wrong in both directions until every row is tagged properly.
Seven providers, and they are not the same kind of thing. FamilySearch — Croatia — 1204 of 1204 places walked, 5299 volumes · FamilySearch — Udine — 26 of 26 places walked, 642 volumes · Pazin State Archive — 14 of 14 places walked, 1151 volumes · Antenati — 3 of 3 places walked, 74 volumes · FamilySearch — Pola & Trieste — 9 of 9 places walked, 36 volumes · Matricula — the Littoral — 3 of 3 places walked, 9 volumes · FamilySearch — Međimurje — 46 of 46 places walked, 365 volumes · FamilySearch — Delnice — 1 of 1 places walked, 6 volumes · FamilySearch catalogue — 75 of 75 places walked, 189 volumes.
FamilySearch is broad and shallow: a title, a year span, and nothing name-indexed. Antenati is the opposite shape again — two comuni, fifty-nine volumes, and every one of them a book to be turned by hand, because the Italian portal publishes no index of names at all. Its arks sit on a different host from everything else here, so the map sends you to the archive's own list rather than dressing them up as FamilySearch links. Pazin is narrow and deep, and it is the only one that admits what it has not done — 154 of its volumes are marked «not digitised». That column means «not digitised by Pazin» and nothing more: checked against the films, 116 of the 154 have been photographed by FamilySearch anyway, and only 38 are books that two catalogues agree nobody has filmed. Every one of the 7771 catalogued volumes that FamilySearch has filmed is now openable from this map — the ark harvest finished on 15 September, 5,482 of 5,482, with nothing left outstanding.
This paragraph said something else for about an hour. It reported that thirty-four of the forty Kvarner parishes had never been walked — and they had been. The builder was reading the wrong field: Kvarner keeps its volumes under a field called films, like Dalmatia, and carries a stale books title-list on six of its forty parishes. Reading books invented a gap that did not exist and under-counted the archive by thirty-four parishes and 201 volumes. Logged as a correction.
339 volumes are listed by one holder, from before 1890, that the other has not filmed. Those are the ones worth caring about — books on a shelf with no photograph of them anywhere. Pazin's own list marks 154 volumes «not digitised», of which 38 are filmed by nobody else either — the short list actually worth posting, set out place by place and holder by holder on the undigitised.
448 more fall after 1890. They are not a gap. FamilySearch withholds recent registers for privacy and the other holder lists them because it holds the paper. Counting those as missing film would have turned a real finding of 339 volumes into a fake one of 787.
The test is deliberately strict about what counts as a shelf. The library catalogue is not one. Its records are bibliographic lines that can span three centuries — «Matična knjiga, 1568-1917» is a single row standing for a parish's whole holdings — and letting it into the comparison inflated this test to seventy-eight places and 1,572 volumes, almost all of it a catalogue entry being read as a claim about a particular year. It is excluded, and the number came back down.
And strict about what counts as cover. An index is not a register and an allegato is not a register; a marriage index for 1815–1830 does not make the marriage book for those years present. Only births, marriages and deaths are compared, and only against their own kind.
| Place | Filmed nowhere, before 1890 | Held back for privacy, after 1890 |
|---|---|---|
| Barban | Births 1629–1714 (5) · Deaths 1691–1714 (2) · Marriages 1669–1689 (1) · Marriages 1692–1749 (3) · Marriages 1793–1799 (1) | Births 1923–1999 (26) · Deaths 1917–1979 (24) · Marriages 1920–1979 (24) |
| Novigrad | Births 1662–1759 (3) · Births 1825–1843 (3) · Deaths 1825–1832 (1) · Marriages 1667–1809 (3) · Marriages 1825–1830 (1) | — |
| Prelog | Births 1674–1800 (2) · Births 1804–1890 (4) · Deaths 1841–1890 (1) · Marriages 1674–1749 (1) · Marriages 1841–1890 (1) | Deaths 1907–1918 (1) |
| Štrigova | Births 1685–1881 (7) · Deaths 1730–1799 (2) · Deaths 1848–1849 (1) · Deaths 1858–1894 (1) · Marriages 1794–1857 (2) | Births 1895–1904 (6) |
| Vodnjan | Births 1559–1759 (4) · Deaths 1573–1704 (1) · Deaths 1714–1812 (3) · Marriages 1559–1813 (3) · Marriages 1885–1969 (37) | Births 1899–1975 (28) · Deaths 1923–1956 (26) · Deaths 1959–1969 (1) |
| Donji Vidovec | Births 1671–1892 (4) · Deaths 1760–1892 (2) · Marriages 1671–1753 (1) · Marriages 1760–1892 (2) | Deaths 1895–1906 (2) · Marriages 1917 (1) |
| Ližnjan | Births 1683–1745 (1) · Births 1884–1993 (4) · Deaths 1682–1812 (2) · Marriages 1680–1812 (1) | Deaths 1901–1952 (4) · Deaths 1959–1993 (1) · Marriages 1947–1952 (2) |
| Belica | Births 1848–1849 (1) · Deaths 1848–1849 (1) · Marriages 1794–1849 (2) | Deaths 1916–1925 (1) |
| Čakovec | Births 1748–1890 (5) · Deaths 1748–1890 (6) · Marriages 1748–1890 (4) | Births 1895–1906 (8) · Deaths 1895–1918 (10) · Marriages 1895–1906 (3) |
| Dekanovec | Births 1790–1857 (2) · Deaths 1789–1869 (2) · Marriages 1790–1857 (2) | Births 1895–1906 (5) · Deaths 1895–1906 (3) · Marriages 1895–1906 (3) |
| Delnice | Births 1858–1879 (1) · Deaths 1858–1889 (1) · Marriages 1858–1890 (1) | — |
| Donja Dubrava | Births 1790–1883 (4) · Deaths 1790–1883 (2) · Marriages 1790–1857 (1) | Births 1895–1917 (5) · Deaths 1895–1901 (2) · Marriages 1895–1917 (3) |
| Draškovec | Births 1789–1878 (2) · Deaths 1789–1878 (2) · Marriages 1789–1878 (2) | Births 1895–1904 (3) · Births 1907–1909 (1) · Deaths 1895–1906 (2) · Marriages 1895–1906 (1) |
| Goričan | Births 1789–1881 (2) · Deaths 1790–1881 (2) · Marriages 1790–1881 (2) | Births 1912–1917 (1) · Deaths 1916–1917 (1) |
| Gornji Mihaljevec | Births 1790–1893 (4) · Deaths 1790–1872 (3) · Marriages 1790–1857 (1) | Births 1899–1902 (1) |
| Kotoriba | Births 1789–1885 (3) · Deaths 1789–1857 (2) · Marriages 1789–1857 (2) | Births 1903–1906 (2) · Deaths 1907–1919 (1) |
| Križevci | Births 1689–1897 (9) · Deaths 1752–1861 (5) · Marriages 1702–1857 (6) | Births 1898–1905 (2) |
| Legrad | Births 1849 (1) · Deaths 1849 (1) · Marriages 1849 (1) | Births 1895–1906 (3) |
| Nedelišće | Births 1669–1857 (3) · Deaths 1741–1857 (1) · Marriages 1669–1857 (3) | Births 1905–1906 (2) |
| Podturen | Births 1770–1891 (3) · Deaths 1770–1894 (2) · Marriages 1770–1857 (1) | Marriages 1895–1906 (3) |
| Materada, Umag | Deaths 1835–1941 (2) · Marriages 1835–1941 (2) | — |
| Poreč (Poreč) | Births 1889–1954 (28) · Marriages 1888–1968 (34) | Births 1958–1963 (2) · Deaths 1924–1979 (31) |
| Vižinada | Births 1663–1732 (1) · Births 1887–1994 (28) | Deaths 1911–1980 (28) · Marriages 1924–1980 (26) |
| Pazin | Births 1883–1966 (35) | Deaths 1910–1985 (34) · Marriages 1913–1947 (31) · Marriages 1949–1964 (7) |
| Svetvinčenat | Deaths 1668–1738 (1) | Births 1894–1986 (32) · Deaths 1918–1987 (21) · Marriages 1924–1930 (7) · Marriages 1932–1937 (6) · Marriages 1942–1969 (7) |
| Mursko Središće | — | Births 1906 (1) · Marriages 1895–1917 (3) |
Volume counts in brackets. A span is merged from consecutive volumes, so «Births 1629–1714 (5)» is five books covering that stretch with nothing of their kind on film across it.
● read — somebody here opened it and went through it end to end. There are two such volumes outside Antenati: Gračišće's baptism book 1667–1745, read image by image off Pazin's microfilm, and Donji Proložac's deaths 1837–1859, swept page by page.
◐ part read — opened, and some of it read: a stretch of years, a run of images, an index followed. Svetvinčenat's 1663–1708 and one Vodnjan reel.
○ never opened — this archive keeps a per-volume record for this parish, and this volume is not in it. Nobody here has read it. It is a checked negative about our work, and about nothing else: the book is filmed, it is linkable from this panel, and for all this archive knows somebody read it yesterday in Pazin.
· not written down — the archive keeps no per-volume record for this parish at all, so it cannot say either way. Most of this map is in this state.
The difference between the last two is the difference between a negative and a silence, and it is the whole reason there are four marks rather than three.
And the honest limit of «never opened». It is only as good as the ledger behind it. A volume somebody read before that ledger existed, and never wrote down, would show as never opened — so the state means «not recorded as read», and the archive's answer to that is to record readings the day they happen rather than the week after.
Pola is this family's ground and Trieste and Gorizia are not, so the provider is drawn as what it is — a shelf that mostly overlaps nothing else, in the way Antenati and the Delnice films already do. They do not open from a marker, because the panel builds FamilySearch addresses and these are not FamilySearch; they open from here.
- Pula · Pola
- Trieste · Triest
- Trieste · Triest
- Gorizia · Görz
They are small Croatian parishes — Trojstvo, Viduševac, Taborsko, Sobotica, Prvić Šepurina, Juraj u Trnju, Hrnetić, Vlaška Kapela — and a handful of institutions that are not settlements at all: Bolnica Dubrovnik, Nahodište Dubrovnik and Nahodište Split, which are a hospital and two foundling houses and should in the end hang on their towns.
What was tried. GeoNames files a village too small for a settlement record under its church or its railway stop instead, and reading those classes as well as the populated places rescued four — Mađarevo is a railway stop, Mandalina a railway halt, Komarevo a locality. A stream or a hill of the same name is not the village and is deliberately not used: the Glogovnica in the dumps is a river.
The rest need a Croatian source and a person, not another pass of the same script. They are listed by the place audit, which runs the two tests that found six parishes standing in the wrong country.
GAP — no marriages register listed anywhere, 6 years. Nobody has a book of that kind for those years. Not Pazin, not FamilySearch, not Antenati, not anyone this archive walks. The book is missing, or it never existed. This is the row that should make you unhappy; it is tinted for exactly that reason. (It said «no register filmed» until 16 September, which was the wrong word and made it look like the next one.)
NOT FILMED ANYWHERE. The opposite. The book exists — a catalogue names it, a shelf holds it — and nobody has photographed it. Nothing is lost; it is simply not readable from a desk. These are set out one by one on the undigitised, and they are the ones worth writing a letter about.
A volume with no link at all. A third thing again, and the commonest. Gračišće is the example: thirty-two volumes, every one of them listed by the Pazin State Archive, which publishes a register list and no images — and FamilySearch's Croatian collection does not hold the parish. So there is no address to link to and there never was one to harvest. The panel now says so in those words. Where a place's volumes do come from a filmed collection and still show no link, that is the other case: an address this archive has not fetched yet, and the panel says that instead.
So: a gap is a hole in the record. Not filmed is a book waiting for a camera. No link is a book waiting for a reader to travel, or an errand this archive has not run.
176 of the 1245 places on the map have a real hole in at least one register series — 308 of them in all. 559 have none: every year between their first volume and their last is covered by something.
A one- or two-year hole is usually not a hole. A volume titled «1716–1816» sitting beside one that ends in 1714 leaves 1715 uncovered on paper and almost certainly not in fact, so those are marked as a labelling artefact rather than cried over. The exception is a parish whose volumes each cover a single year — the Napoleonic comuni are all like this — where a missing year is a missing book. That rule is why Mione's lost 1807 shows on the map as three gaps and not as three shrugs.
An index is not cover, and neither is an allegato. They are counted separately at the foot of each panel, because a marriage index for 1815–1830 does not make the marriage book for those years present.
Roman Catholic — 997 · Orthodox — 240 · Civil — 53 · Military — 17 · Jewish — 16 · Greek Catholic — 15 · Reformed — 11 · Evangelical — 6
Croatian registers are not all Catholic and a family that looks absent from one set can be sitting in another. The 240 Orthodox parishes in particular are a body this archive has never touched.
It is kept in the data, unpublished, and the useful question it answers is narrow: is anybody else working our parishes?
Our people: reading the second question
Nine places carry the whole documented descent, and seven of them sit inside a hundred kilometres of each other. The family did not move much until 1939, and then it moved to three continents in a decade.
Rijeka, Pula and Ližnjan carry far more of the surname than Gračišće ever did — and none of them is connected to this family by any document. Weight on the map is not kinship.
Between Mione in 1679 and Gologorica in 1711 there is nothing but a house with a date on it. That two-hundred-kilometre jump is the one thing this archive cannot yet prove.
1031 placements across 165 towns, and 143 of those towns carry at least one named person.
Living people are on this map by name and by town, and by nothing else. The rule until 15 September 2026 was no place of residence, which kept every living relative off the map entirely. It is now name and town, never a date — and all 26 living people this archive names are placed.
8 of them are at Johannesburg and 5 at Brisbane, and two are at both. That is the shape of it: the family left Senj for the Rand, and one household of the three went on to Australia — the aunt and her husband, the uncle's widow and his daughters, and the cousins are all still in South Africa. Not one of them has a date on this map, and none ever will.
The ground: the third question, and the finding in it
That is not a quirk of the data. It is the emigration, drawn: the people who stayed are in parish books and have no findable grave; the people who left are under stones and were never in a Croatian register. The register entries run 1619 to 1997 and are almost all Istrian and Kvarner — Rijeka, Ližnjan, Poreč, Pula, Kaštelir. The headstones run 1844 to 2024 and are almost all somewhere else — Connecticut, Rome, Queens, Broken Hill, Ticino.
The green colour in that legend has no members. It is left there because the day a place turns green — a grave found in a parish this archive has read — is the day one of these two halves joins the other.
4 burials have no place to pin. Thirty-six are entered as «Croatia, parish not named» and four as «not stated», read out of indexes that record a death without recording where. They are counted on the burials page and deliberately absent here: a map that guessed at a parish would be worse than a map with a hole in it.
The twelve heaviest places in the ground
Where most of the family lies, by how many burials this archive can name there.
| Place | Known from | Burials | Years |
|---|---|---|---|
| Rijeka | 23 register | 23 | 1717–1899 |
| Svetvinčenat | 14 register | 14 | 1808–1897 |
| Pula | 13 register | 13 | 1659–1997 |
| Torrington, Litchfield County, Connecticut, USA | 13 headstones | 13 | 1925–2020 |
| Ližnjan | 10 register | 10 | 1843–1891 |
| Poreč | 8 register | 8 | 1849–1879 |
| Rome, Lazio, Italy | 8 headstones | 8 | 1925–1996 |
| Draguč | 6 register | 6 | 1819–1855 |
| Santa Maria di Sala, Venezia, Veneto, Italy | 6 headstones | 6 | 1926–1961 |
| Rovinj | 5 register | 5 | 1619–1864 |
| Giubiasco, Ticino, Switzerland | 4 headstones | 4 | 1956–1991 |
| Kaštelir | 4 register | 4 | 1858–1892 |
Rebuilt from the KML export of Genealogy Resources — Croatia & Surrounds, kept whole at data/genealogy-resources-croatia.kml. Coordinates from the GeoNames dumps for HR, BA, SI, RS, HU, IT, AT and ME, used under CC BY 4.0 — 1,123 of the 1,286 places here are placed by it. Volume inventories from this archive's own walks of the FamilySearch waypoints. Nothing on this map is a finding — it is a record of what exists and how far we have got.