Working note · from our own desk
Most family lines in Central and Eastern Europe run out of parish registers somewhere in the eighteenth century. What comes before that is not silence — it is tax. Landowners were assessed household by household, and those assessments were printed in the nineteenth century in scholarly volumes that now sit in digital libraries as scanned images with no searchable text. This is a working note on how we get into them: the exact mechanics, the traps, and what such an entry actually tells you about a family that has not yet acquired a surname.
The mechanics
Volumes of this kind are usually served as DjVu, and the URL you are given returns a file of a few hundred bytes. It looks like a failed download. It is not: it is an indirect DjVu — an index that points at one file per page, each stored separately.
So the first move is to list the pages the index refers to, then fetch them. With the DjVuLibre tools that is two commands: djvused to print the file list, and any parallel fetch to pull them down. A 485-page volume came to under five megabytes and under a minute for us. Keep the index file beside the pages or nothing will open.
Run the text extractor over the volume and count the characters. If it returns zero, there is no text layer: the book is photographs of paper, and no amount of searching will find your village. This check costs one command and saves an hour of wrong assumptions. Ours returned zero.
Without a text layer you need another way in, and the nineteenth century provides it. Serious editions of this material carry a place index at the back. Render the last dozen pages as images, read the index with your eyes, find the printed page number for your place, then convert that to an image number — the offset between printed page and scan is constant once you have calibrated it on any one page. Two renders, and you are looking at the right spread.
The entry
The entries are terse and formulaic, and that is what makes them readable once you know the pattern. A typical line names the holder, then counts what is taxable in each settlement: houses on the market square, poor cottages, gardeners with and without land, mill wheels of each type, a craftsman, a priest.
That is not a list of names. In 1583 the people in these villages are counted, not named — surnames as we use them are still forming, and the register records the lord and the taxable units, not the households' own identities. Anyone promising you a named sixteenth-century peasant ancestor from a document like this is selling something.
Method
Two reasons, and neither is generosity.
The first is that a finding nobody can re-find is not a finding. We cite volume, section, printed page and scan, so that anyone — a client, a client's sceptical grandchild, another researcher — can open the same page and disagree with us if we are wrong.
The second is that this is the part of the work that does not scale by buying a subscription. Databases have swallowed the easy records. What is left is scattered across national libraries in formats that were never meant to be searched, in languages and scripts that were never meant to be machine-read. Getting into them is a craft, and we would rather be judged on the craft than on a marketing claim.
If you are stuck at the point where the parish registers stop, this is usually the layer underneath. Tell us the place as your family says it and we will tell you whether such a volume covers it.
Send us the surname, the village as your family pronounces it, and one exact date. We will tell you which layer of record your search has actually run into — and whether it can be got past.
Write to us