Beaver Island ArchiveSearch

Dataset

Chronicling America coverage of Beaver Island and Strang

Also National Digital Newspaper Program coverage

Draft — sources verified, prose evolving

TypeDataset
ErasStrang era

Summary

This is the contemporary national newspaper record of Strang’s Beaver Island settlement, as held by this project via the Library of Congress’s Chronicling America / National Digital Newspaper Program (NDNP). Strang’s own newspaper, the Northern Islander, was never digitized into the NDNP and is not part of this collection at all – it survives instead through Internet Archive holdings (see data/books/internet_archive/NorthernIslander18501856/). What this dataset holds instead is the national press’s reaction to the island: other papers, in other cities, quoting, reporting and editorializing about Strang’s kingdom, concentrated in 1851-1856 and running to at least the early twentieth century. It carries one serious, well-documented contamination trap that any use of this data must guard against.

What the collection holds

Retrieval used the www.loc.gov search API (the legacy chroniclingamerica.loc.gov API is retired and returns HTTP 403) against a set of phrase queries, quoted exactly – an unquoted q=Beaver Island returns the entire 356,767-item collection, not a filtered result.1 Per-query page counts recorded in this project’s own retrieval notes:

Query Pages
"beaver island" "strang" 500 (capped)
"beaver island" mormons 500 (capped)
"king strang" 482
"james j. strang" 406
"mormon king" 357
"jesse strang" 188
"northern islander" 55
"kingdom of saint james" 0

After deduplication across queries, this project’s notes report 1,301 unique pages, concentrated in 1851-1856, from papers based in New York, Washington, Richmond, Baltimore, Saint Paul, Indianapolis, Kenosha and Montpelier – several within weeks of Strang’s 16 June 1856 assassination.1 Two of the eight queries hit a disclosed cap of 500 results (5 pages of 100), so this total is a documented floor rather than an exact count.

Separately, this project’s data-quality notes describe an initial full retrieval run that built a catalogue of 1,089 pages, which was then lost: a rerun launched during an loc.gov rate-limit block had all its search queries fail, and the retrieval script wrote the resulting empty result list over the good 1,089-entry catalogue (ocr_saved texts were unaffected, since OCR download is idempotent and skips existing files).2 As held on disk at the time of this article, chronicling_america_catalog.json records only 190 entries – the partial rebuild performed after that loss – while the ocr/ directory itself contains 1,121 saved page-text files, more than the current catalogue lists. This indicates the recovery process (re-fetching missing OCR text) has continued past the point the catalogue was last rebuilt, and that the catalogue undercounts what is actually held. Anyone querying this dataset by catalogue alone will undercount the newspaper pages actually present on disk; a directory listing of data/federal/chronicling_america/ocr/ is a more current inventory than chronicling_america_catalog.json as of this writing.

The critical trap: the other Jesse Strang

The query "jesse strang" alone, without a date floor, returns pages from as early as 1827 – fourteen years before James Jesse Strang (b. 1813) had done anything of note, and sixteen years before his 1843 conversion. These are not noise: they are consistent, substantial newspaper coverage of a real, unrelated case – the trial and execution of a different Jesse Strang, in Albany, New York, for the 1827 murder of John Whipple, a sensational story carried nationally at the time. This project’s own scan of the held OCR (using min_year=1800, no floor) finds 23 hits across 18 distinct pages dated 1827-1834 matching “Jesse Strang,” confirmed by direct reading to be uniformly about the Albany case; this project’s data- quality notes independently record the same trap as “46 pages dated 1827-1834” from a related but not identical query formulation.2 Confirmed directly from the OCR of sn82014309_1827-08-09_ed1_p2.txt:

“The trial of Jesse Strang for the murder of Mr. John Whipple, commenced in Albany on the 26th ult. and closed on the next day, by the jury’s returning a verdict of guilty…”3

Rule, restated here for this dataset specifically: apply a hard date floor of 1843 (the year of James Jesse Strang’s conversion) to any “jesse strang” hit before drawing a conclusion from this collection. Any unfiltered count or timeline built from this dataset that includes pre-1843 material will read the Albany murder trial as early national interest in Beaver Island – exactly backwards.

Significance

This dataset is this project’s only holding of the external contemporary press record of the Strang period – independent of both the Strangite press (the Northern Islander itself) and later retrospective works like The Kingdom of Saint James (Quaife, 1930). It is essential for confirming dates (see the independent telegraph confirmation of the 16 June 1856 assassination date in Assassination of James Strang) and for gauging how the kingdom was perceived nationally, but it says nothing about the Northern Islander’s own reporting, which this project holds separately via Internet Archive.

Open questions

  • The discrepancy between the “~1,089 pages” catalogue figure (docs/data- quality.md) and the “1,301 unique pages” per-query estimate (docs/sources.md) is not resolved here; they may reflect different retrieval runs, different deduplication logic, or one being a floor and the other an estimate before the 500-page caps were understood. Neither figure has been reconciled against the 1,121 OCR files actually present on disk.
  • The full catalogue rebuild recommended in this project’s own notes (uv run python scripts/retrieve_loc_newspapers.py, once the loc.gov rate limit clears) has evidently run at least partially (more OCR files exist than the 190-entry catalogue), but has not been completed or reconciled as of this writing.
  • Whether any Chronicling America page independently corroborates a detail contested between the Daily Northern Islander, Quaife, and Elizabeth Whitney Williams (see Assassination of James Strang) beyond the ones already cited there has not been systematically checked here.

See also

Sources

Footnotes

  1. docs/sources.md, section on Chronicling America / national press coverage. 2

  2. docs/data-quality.md, sections “Newspaper search contamination: the other Jesse Strang” and “loc.gov rate limiting looks like missing data.” 2

  3. data/federal/chronicling_america/ocr/sn82014309_1827-08-09_ed1_p2.txt.

Sources

  1. Library of Congress, Chronicling America / National Digital Newspaper Program, search results for phrases including "beaver island" strang, "king strang", "james j. strang", "mormon king", "jesse strang", "northern islander" and "kingdom of saint james", retrieved via `www.loc.gov/collections/chronicling-america/?fo=json`.
    This project's own retrieval and catalogue of the collection, not the collection itself. Catalogue file `chronicling_america_catalog.json` currently holds 190 entries (a partial rebuild after a rate-limit incident overwrote a fuller ~1,089-entry catalogue -- see below); the `ocr/` directory on disk was independently counted at 1,121 saved page-text files as of this writing, more than the current catalogue records, indicating a partial OCR-recovery pass has run since the catalogue was last rebuilt.
    federal · chronicling_america
  2. Project research diary and data-quality notes documenting the retrieval, the loc.gov rate-limit incident, and the "other Jesse Strang" contamination.
    Records the per-query hit counts (`"king strang"` 482, `"james j. strang"` 406, `"mormon king"` 357, `"jesse strang"` 188, `"northern islander"` 55, `"kingdom of saint james"` 0) and a combined estimate of 1,301 unique pages after deduplication, concentrated 1851-1856. Two queries hit a disclosed 500-page retrieval cap, so the true total is a floor, not an exact count -- and does not itself match the "~1,089"-page catalogue figure recorded elsewhere in this project's own notes (see Open questions).
    docs · sources
  3. Project data-quality notes on the Jesse Strang / John Whipple newspaper-search contamination.
    Documents that the query `"jesse strang"` alone returns 46 pages dated 1827-1834, all reporting the unrelated Albany, New York trial and execution of a different Jesse Strang for the 1827 murder of John Whipple. Quotes the OCR directly: "The trial of Jesse Strang for the murder of Mr. John Whipple, commenced in Albany on the 26th ult. and closed on the next day, by the jury's returning a verdict of guilty..." (`sn82014309_1827-08-09_ed1_p2`).
    docs · data-quality