← Project Scroll · About · Discoveries
Standards
How we research. Every clue is true™.
Every statement on Project Scroll comes from a record you can check, and every claim of something new has been searched for in print first. These are the rules we work by.
- Findings in the register
- 183
- Research notes
- 12
- Permanent DOIs
- 6
- Stories, every one sourced
- 65
The record comes first
We work from original records, read on the page wherever a scan exists. Each story lists its sources in a numbered list at the foot, with where the original is held and a link to it; numbers in the text point to that list.
County histories
The histories of 1875, 1882 and 1906, and their like across the valley, read for what they say and who told them.
Newspapers
Weeklies, dailies and the state-news columns, read on the page image, with the date of the issue and the page.
Journals and laws
Legislative journals, session laws, government registers and gazetteers: what was voted, printed and listed, and when.
Land, court and maps
Plats, deeds, land entries, court records, atlases and fire-insurance maps, read against each other.
Letters and diaries
The people of the time in their own words: letters, diaries, travellers’ accounts and family papers.
Pictures
Postcards, photographs and engravings, credited to the library or archive that holds them, with their rights.
Scrollsigns: what the labels mean
Each paragraph of a research story carries a label, a Scrollsign™, so you can see at once what the record says and what is our reading of it. A news story uses four.
- Documented
A source says it, and the source is cited.
For example The village plat was recorded on 16 February 1848.
- Reading
Our interpretation of the sources, kept apart from the facts.
For example The creek already had the name, so it was the obvious one.
- Open
A question the records do not yet answer.
For example Who named the village Dowagiac?
- New finding
Something we found in the record that no earlier publication we could find has.
For example The name is in print in 1833, five years before the first gazetteer.
The examples come from How Dowagiac got its name. Our research reports and notes use the full set of nine Scrollsigns, from Document, the strongest, down to Hypothesis, which fold into these four in a news story; all nine are set out on the Scrollsigns™ page.
Eight tests before we call it new
We call a finding new only when every one of these holds.
- We have seen the record.The words are read on the page image, not only in machine-read text.
- It is fully cited.Title, date, printed page, who holds the original, and a link.
- Each witness counts once.A later history that reprints an earlier one is not a second source.
- We searched for it in print first.Google Books and Scholar, the Internet Archive and HathiTrust, the Library of Congress’s newspapers and the state newspaper collections, the standard histories, the societies that keep the subject, Wikipedia and its references, and the journals of the field. What we searched is listed with every finding.
- Conflicting sources are shown.Where the records disagree, we say so and cite them all.
- Dates and context are read correctly.The date of the record, the date of the event and the date of any reprint are kept apart, and a joke is read as a joke.
- The claim is no larger than the evidence.We write “the earliest we have found,” never “the first ever.”
- A second reader checks it.Every quotation, page and date is checked against the original before it is published.
Three kinds of finding
- New findingNot in any earlier publication a careful search could find.New to print as far as we can tell, not a claim that nobody ever knew.
- New documentationSomeone said it before us, but without a source.We supply the record, and credit them by name.
- RecoveredAlready in an old printed book, but missing from every modern account.We bring it back into the story, and say where it was.
There are 183 findings in the discoveries register, each with the record it came from and what we searched.
From the record to the page
Every story goes the same way.
- Find the recordIn a library, an archive or a scan: the oldest witness we can reach.
- Read it on the pageEvery quotation is taken from the page image, not from machine-read text.
- Search print for itBefore anything is called new, we look for it in the books, papers and journals listed with the finding.
- Cite it and mark itFull citation in the numbered sources; every paragraph labelled for what it is.
- A second reader checksQuotations, pages and dates against the originals, then it goes up.
Research notes and permanent identifiers
The most significant findings are set out as research notes: short, dated papers giving the record, where it is held, what it changes, its limits and what was searched. Each has a fixed address, a PDF, and a DOI, a permanent identifier issued through Zenodo, the open research repository run by CERN, so that it can be cited and found for good.
- 2026-15The forgotten revival of E Clampus Vitus, 1915–1920: the Marysville boosters’ lodges, numbered one to eight, and how the war ended themDOI to come
- 2026-13E Clampus Vitus before the California Legislature: the Senate debate at Benicia struck from its journal (March 1853), and the Assembly’s dollar-a-member resolution (April 1855)DOI to come
- 2026-11Older than “about 1839”: the Treats, Henry Dewey and the red-chalk warning on the road to Dewey’s Lake, Silver Creek Township, MichiganDOI to come
- 2026-09Three names at once: Crum’s Point, Crumtown and Crumstown, St. Joseph County, Indiana, 1875–1903DOI to come
- 2026-08Three sawmills on Potato Creek, 1863: the mills on the ground of Potato Creek State Park, IndianaDOI to come
- 2026-07A museum before the museum: the Cass County Pioneer Society’s collection, 1874–1899DOI to come
- 2026-06Seven elk for the King of Italy: James E. Bonine’s own account of 1875doi:10.5281/zenodo.23638846
- 2026-05The Kentucky raiders’ bail and wagon: what the Licking Valley Register reported in September 1847doi:10.5281/zenodo.23631269
- 2026-04The earliest L in Kalamazoo yet found: Schoolcraft’s “Kikalemazo” of 1820, and the likely atlas of 1823doi:10.5281/zenodo.23631264
- 2026-03Ephraim Bee in the newspapers, 1855–1860: the remembered founder of E Clampus Vitus in records of his own daydoi:10.5281/zenodo.23631262
- 2026-02The Cassopolis lodge of 1846: the earliest lodge of E Clampus Vitus yet found with its members nameddoi:10.5281/zenodo.23631260
- 2026-01The earliest dated mention of E Clampus Vitus yet found: Connersville, Indiana, 27 February 1846doi:10.5281/zenodo.23631258
Corrections, versions and other people’s work
Corrections
When we get something wrong, we correct it on the page the same day, and the correction is noted on the page itself. If someone shows an earlier publication of one of our findings, we credit them by name and change the finding’s standing.
Versions
A research note is never rewritten in place. It gets a new version, and the old one stays on record.
Contributed articles and collaborations
We publish articles by readers, families and fellow researchers under their own names, and work with historical societies, museums and libraries on joint research. It is all held to the same standard, and credited by name.
Citing our work
Every story and note has a “Cite this” button that gives a full citation. For a finding with a research note, cite the note and its DOI. Research is by Michael Urbanski for Project Scroll; his ORCID iD is 0009-0006-8183-5828.
Tell us
Know of an earlier publication of anything we call new, hold a record that adds to a story, or found an error? We will check it, credit you, and say what changed.
These rules follow the Board for Certification of Genealogists’ Genealogical Proof Standard and the American Historical Association’s Statement on Standards of Professional Conduct, which asks historians to document their findings so that others can “retrace the steps in an argument,” and to be “explicit, thorough, and generous in acknowledging one’s intellectual debts.”