A working reference · updated as we read
This page exists because we could not find it anywhere else. It is the practical knowledge we use daily: which layers of record exist, what each one actually tells you, why one village's books end up in two different buildings, and why a thirty-year hole in a parish series is normal rather than a sign that you searched badly. It is written for people doing this themselves as much as for our own clients. Where we state a fact we say where it came from.
Start here
Most people know only the first one and give up when it runs out. The second layer is where the real depth usually is, and it is the one that survives when churches burn.
Births, marriages and deaths, kept by the parish. One entry names the child, both parents, the father's own patronymic, the godparents, the village, and often the priest and clerk. This is the layer with names in it — and the layer most often lost.
Tax censuses of households. They give the whole family in one place: every person, their age, and the household's number. Because they repeat every fifteen to twenty years, they let you watch one household age, split and move — and they survive in state archives when church books do not.
Court files, land and cadastral surveys, school lists, army records, and in the twentieth century the files of whatever state was in charge. Uneven, scattered, and often the only place a person appears after the church series stops.
If the family emigrated, the single most valuable document is usually not in Europe at all. A ship manifest, a naturalization file, a draft registration card or a census line will often name the exact village of origin — the one the family stopped saying out loud two generations ago. Without that name, searching a European archive is close to guessing; with it, the search becomes finite. We start there more often than we start in Europe.
Layer two, in detail
A revision list is a tax document, which is exactly why it is useful: the state had a reason to be accurate about who was alive and how old they were.
The lists were compiled at irregular intervals through the first half of the nineteenth century. For the households we have worked through, the usable series runs 1811, 1816, 1834, 1850, 1858. Each entry gives a household number, every male in it with his age, and — from the later revisions — the women too. Crucially, each list also records what happened since the previous one: who died, in which year, who was taken as a recruit, who moved away.
That last detail is what makes the layer powerful. A parish death register may be lost, but the revision list still says the man died in 1821. And because the lists repeat, ages can be cross-checked: sixteen years between two revisions should make a man sixteen years older. When they don't agree — and they often don't by a year or two — that is worth noting rather than smoothing over.
One household we traced carries the number 78 in the revisions of 1816 and 1834, then 102 in 1850 and 1858. The head of household appears in 1816 aged about forty-nine and is recorded as having died in 1821. His son appears at thirty in 1834, at forty-six in 1850, and as a widow's household by 1858. Four documents, one family, and a death date for a man whose parish burial entry does not survive.
Sources: regional state archive, fond 37, inventory 2, files 156 (1816), 175 (1834), 186 (1850), 195 (1858).
Why your search failed
This is the single most common reason a search stalls, and it has nothing to do with how hard you looked.
The borders moved and the paperwork did not follow. A parish that sat in one province in 1860 may sit in a different country in 1930 and a third administrative unit today. When archives were reorganised, a single parish's registers were often divided: one stretch of years went to one regional repository, another stretch to a diocesan or provincial one, hundreds of kilometres away. A series that looks like it simply ends in 1868 may continue in another building from 1900, catalogued separately and never cross-referenced.
The catalogue does not use the name your family uses. A village can appear under its Polish form, its Ukrainian form, its Russian-era spelling, and a Latin transliteration of any of them — and the archive's finding aid will have picked exactly one. Searching the family's own spelling returns nothing at all. We have worked with a village that appears as six different strings across the documents and the catalogue, one of them written by the parish priest's own hand in 1784.
The film and the book are numbered differently. Much of this material was microfilmed in the twentieth century and is now available as images. An image number is not a page number and not a folio number: a single reel can carry nine hundred images spanning several villages and several years. Working out which stretch of images covers which village, before reading anything, saves more time than any other single habit.
The part nobody writes down
A missing decade is not a failure of research. It is a fact about the archive, and it belongs in the result.
In one parish we worked through, the church series available in one archive stops at 1868, and the series in the second archive begins at 1900. Between them lies a thirty-one year hole — 1869 to 1899 — and into that hole falls a birth and a marriage we needed. No amount of further searching in those two repositories will produce them. What can be done is narrow and specific: check whether a duplicate copy was deposited elsewhere, check whether a later document restates the missing facts, and file a formal request for anything not yet filmed.
We write gaps like this into the books we produce, in plain words. It is tempting not to: a clean line of descent reads better than a line with a hole in it. But a family history is checked eventually — usually by a grandchild with more patience than sense — and a gap that was hidden becomes a reason to doubt everything else on the page.
Where two people share a surname and a village but no document links them, we say so and leave them unlinked. In one case a family with an almost identical surname turned out to belong to a different district entirely; they stayed out of the tree. A matching surname is not kinship. Kinship is a chain of household numbers, ages that agree across revisions, and patronymics that carry from one generation to the next.
Practical
One surname we work with appears in thirty-five documented forms across Cyrillic, Polish and Latin script, plus clerks' and indexers' errors.
Any serious search has to run all the plausible forms in parallel, and then some implausible ones: an indexer working fast produces variants no linguist would predict. The hardest case we have hit was a clerk who ran out of room on a line and broke the surname across two lines with a hyphen. No index search will ever return that entry. It was found by a person reading the whole page — which is, in the end, what this work is.
Method
If a finding cannot be re-found by a stranger, it is not a finding yet.
We publish this because the standard is not widely applied outside professional work, and because it is the thing that separates a family tree from a family legend. Anyone is welcome to use it, cite it, or argue with it.
Tell us the surname, the village as your family says it, and one exact date. We will tell you which of the problems above you are actually facing — and whether it can be solved.
Write to us