Behind both stands the U.S. Board of Land Commissioners (created by the Act of March 3, 1851), whose determinations were appealed through the U.S. District Courts for the Northern and Southern Districts of California and, in some instances, the U.S. Supreme Court.

Derivation. Each claim is keyed by its district-and-index docket number in the form ND nnn (Northern District) or SD nnn (Southern District). In the Hoffman table these are split across district and index_no, alongside dc_no (the District Court number), claimant, rancho, grantee, governor, grant_date, county, leagues, acres, and outcome. A grant's popup summarizes this as, for example, "U.S. land case SD 063, Confirmed." Outcomes on the map are normalized to Confirmed / Rejected / (unstated).

Known limits and caveats.

• Hoffman reports the Northern District fully and the Southern District only in part; Southern cases are therefore thinner in the printed volume and are supplemented from the Bancroft files and the finding aid.
• Outcome language is not binary. The raw Hoffman table records 545 claims as confirmed or patented and 239 as rejected, plus roughly 29 in other postures (discontinued, dismissed, or "reversed and remanded," several of which were ultimately confirmed or abandoned on remand). The map's three-way Confirmed/Rejected/blank coloring collapses that nuance; the underlying outcome string preserves it.
• A confirmed claim is not automatically a patented one. "Confirmed" means the title was validated; the patent (and its final survey acreage) sometimes followed years later or not at all. Treat "Confirmed" as an adjudication result, not proof of a recorded patent.
• The digicoll record links point to the archival case file, not to a transcription. Some files are partially digitized.

3. Diseños (the manuscript grant maps)

Source. The Bancroft Library, "Maps of Private Land Grant Cases" (Land Case Maps) collection, UC Berkeley, digitized at digicoll.lib.berkeley.edu and discoverable through Calisphere / OAC. Every image in gallery/disenos-data.json carries the identical, explicit credit line:

"The Bancroft Library, University of California, Berkeley (Land Case Maps). Digitized via digicoll.lib.berkeley.edu. Public domain."

Derivation. The gallery holds 1,870 images spanning 1,396 distinct archival records (source_url), organized into 38 county groups. Each item stores its Bancroft source_url, a catalog title, a working caption, and, where written, a scholarly headnote. Diseños are joined to a rancho by the digicoll record id: a grant's disenos array lists the images whose Bancroft record matches that grant, which is why a single grant (e.g. San Gabriel) can surface several sheets. County placement, per the collection's own note, "is from the catalog title or a strict match to the state's land-grant dataset; ambiguous names are left unplaced."

Known limits and caveats.

• Rights posture: attribution. The imagery is presented as public domain with the Bancroft credit required; reuse should carry the credit line above and link to the archival original.
• Titles are catalog titles, not transcriptions, and the gallery's own note calls its labels "working descriptions." Diseño place-names are frequently misspelled or generic in the catalog, so title-based county placement is deliberately conservative and some sheets are left unplaced rather than guessed.
• Not every grant has a diseño, and not every diseño maps to a grant. The join is only as good as the record-id match; absence of a diseño on a card means no confident match, not necessarily that none survives.
• Some records comprise multiple sheets (suffixed -1, -2), which is why image count (1,870) exceeds record count (1,396).

4. Governors and granting authority

Source. The granting governor named in the definitive grant, taken from Hoffman's report of the grant (the governor field) and, where Hoffman is silent or the grant passed through more than one administration, reconciled against the standard gubernatorial chronology.

Derivation. Each grant is assigned to a governor via governor_name, and its map color via the era layer. The assignment is tenure-checked: the grant date is matched to the governor in office on that date, which is why the field sometimes records a chain (for example "Nicolas Gutierrez; Alvarado," "Alvarado (ratified by Micheltorena)," or "Micheltorena and J. A. Sutter" for grants Sutter issued under delegated authority) rather than a single name.

Known limits and caveats.

• Spanish-era concessions frequently name no governor in the surviving record; these are coded generically ("a Spanish governor") and colored in the Spanish layer.
• Interim and contested administrations (Gutiérrez, Chico, Castro, 1835–36) are grouped into one interim layer for legibility even though the underlying governor_name keeps the specific official.
• A small number of grants (unattributed layer) have a governor still to be confirmed; these are shown in neutral gray and should be treated as provisional.
• The acting gobernador and the juez or prefect who actually signed a concession were not always the same person; where the record distinguishes them the field notes it (e.g. "Jose E. Estrada (Prefect); Jose Jimeno"), but the era coloring follows the governor.

5. Acreage

Source and an important honest caveat. There are two different acreage figures in play, and the portal does not always make the distinction visible.

• The Hoffman table (hoffman-claims.json) carries the legal acreage: the claimed leagues and the patented / final-survey acres as reported in 1862 (for example, Las Mariposas at 44,386.83 acres, La Habra at 6,698.57). This is the figure a researcher should cite for the size of a patent.
• The mapped features (ranchos-experimental.json) carry a different acres value, and testing shows it does not match the Hoffman patent acreage (0 of 109 cross-checked features agreed). The feature-level rank_note labels it "by area," and the values track the GIS area of the reconstructed ECAI polygon, not the patent. For fragmentary features (small mission tracts) the two diverge sharply.

Derivation. On a grant's map card, the acreage and the size-rank ("361st of 493 by area") are computed from the reconstructed boundary's area. The Hoffman patent/claimed acreage lives in the separate land-case table.

Known limits and caveats.

• These two numbers answer different questions. "By area" is how large the reconstructed polygon is; the Hoffman figure is what the U.S. survey certified. Where they differ, the Hoffman patent acreage is the citable legal figure; the polygon area is a cartographic byproduct and should be read as approximate.
• Recommendation for the site: label the map figure explicitly as "reconstructed area" and expose the Hoffman patent acres alongside it, so the two are never mistaken for one another. As currently built, a card that shows only the "by area" number understates or overstates the patent for any grant whose reconstruction is imperfect.
• Rejected and never-patented grants have no patent acreage at all; any acreage shown for them is polygon-derived and nominal.

6. The universe of claims: why the counts differ

These figures come from different documents and count different objects. They are not revisions of one number and must not be conflated or summed.

| Count | Figure | What it actually is | Source |

|---|---|---|---|

| Finding-aid case files | ~857 | Case files enumerated by the Bancroft land-case finding aid. Includes duplicates, town-lot and mission preemption claims, and non-rancho matters. Reported figure, external to our datasets; treat as approximate. | Bancroft / OAC finding aid |

| Hoffman claims | 813 | Rows in hoffman-claims.json: the private land claims Hoffman reported (545 confirmed/patented, 239 rejected, ~29 other). One claim per docket line. | Hoffman 1862 (verified in file) |

| Distinct grants | ~800 | Distinct underlying grants. Fewer than claims because one grant could generate several claims (and a few claims covered fractions of a grant). Approximate; not a stored field. | Derived estimate |

| Register rows | 673 | Rows in register-data.json: 535 mapped + 138 unmapped. The portal's working register of grant-level entries. | register-data.json (verified) |

| Index rows | 662 | Rows in ranchos-index.json (updated 2026-08-24): 524 mapped, 138 unmapped. A slightly earlier build of the same register. | ranchos-index.json (verified) |

| Mapped grants | 535 | Features in ranchos-experimental.json that carry a boundary/point on the map (511 ranchos, 13 ex-mission, 7 pueblo, 4 presidio). | ranchos-experimental.json (verified) |

Version drift is real and worth stating openly. The three "mapped" numbers you will see on the portal (510, 524, 535) are the same quantity at three build dates, not three different things. The current data is 535 (the experimental layer and the register agree). The 524 in the index and the 510 still printed in the map's abstract and legend note are stale text from earlier builds; the legend line "510 of 813 claims mapped" should be refreshed to 535. Similarly the register's 673 rows supersede the index's 662. When the site quotes a headline figure it should quote the register/experimental pair (673 total, 535 mapped) and update the older strings to match.

Two ratios worth keeping straight: 535 of the ~800 grants are drawn (roughly two-thirds), and separately 1,870 diseño images across 1,396 records are catalogued. Those describe different collections and are unrelated to each other.

7. Standard reference works to consult alongside this portal

A researcher tracing a specific rancho should treat the portal as a finding aid and confirm particulars against the printed authorities. Full citations:

• Hoffman, Ogden. Reports of Land Cases Determined in the United States District Court for the Northern District of California, June Term, 1853, to June Term, 1858, Inclusive. San Francisco: Numa Hubert, 1862. (The adjudication of record; the portal's outcome and party data derive from it. Reprinted Yosemite Collections / Rio Grande Press, 1975.)
• Bowman, J. N. Index of the Spanish-Mexican Private Land Grant Records and Cases of California. Berkeley: Bancroft Library, University of California, 1958. (Typescript. The concordance of grant, claimant, docket, and archival file; indispensable for reconciling the ND/SD numbering used here.)
• Bancroft Library. Finding Aid to the Documents Pertaining to the Adjudication of Private Land Claims in California, circa 1852–1892 (BANC MSS Land Case Files). Online Archive of California. (The catalogue of the federal case files; the portal's per-case scan links and General Land Office patent numbers derive from it.)
• United States Supreme Court. United States Reports. Library of Congress digital collection. (The linked official opinion texts for claims decided by the Supreme Court.)
• Cowan, Robert G. Ranchos of California: A List of Spanish Concessions, 1775–1822, and Mexican Grants, 1822–1846. Fresno: Academy Library Guild, 1956. (Grant-by-grant list with grantee, date, and county.)
• Robinson, W. W. Land in California: The Story of Mission Lands, Ranchos, Squatters, Mining Claims, Railroad Grants, Land Scrip, Homesteads. Berkeley: University of California Press, 1948. (The narrative history of the tenure system and the Land Commission process.)
• Perez, Cris. Grants of Land in California Made by Spanish or Mexican Authorities [Ranchos of California]. Sacramento: Boundary Determination Office, California State Lands Commission, 1982. (County-organized grant list with patent data; useful for acreage cross-checks.)
• Shumway, Burgess McK. California Ranchos: Patented Private Land Grants Listed by County. Edited by Michael and Mary Burgess. San Bernardino: Borgo Press, 1988. (Patented grants arranged by present-day county.)
• Beck, Warren A., and Ynez D. Haase. Historical Atlas of California. Norman: University of Oklahoma Press, 1974. (Cartographic context, including a rancho plate; good for orienting reconstructed boundaries against the standard atlas.)

For the manuscript maps specifically, cite the Bancroft collection directly: Maps of Private Land Grant Cases (Land Case Maps), The Bancroft Library, University of California, Berkeley, with the individual digicoll record number.

8. Names, spelling and dates

Policy. Personal and place names are given as the source writes them, and are not silently normalized. The same family therefore appears in more than one form across the register, and a search for one spelling will not always find the other. This is deliberate: regularizing the spelling would quietly overwrite what the record says, and the variants are themselves evidence of how these names were used and heard. Where two forms are known to be the same person, the grant page notes it rather than merging them.

Dates. A year alone means the source gives only a year. Where a grant passed through more than one administration, the year recorded is that of the definitive grant, not of an earlier provisional concession, and section 4 explains how the governor is then assigned.

9. How this was compiled

The register was assembled by one person, working from the digitized sources named above rather than from the original paper files, and using large language models as an assistant: to parse the case summaries into structured fields, to draft the per-grant prose, and to cross-check one source against another. Every field that carries a factual claim was then reviewed against a named source, and the sources are recorded per grant rather than in aggregate. The method is stated here because a reader is entitled to weigh it, and because the appropriate response to it is measurement, which is what section 10 reports.

Boundaries are the one place where geometry was produced rather than sourced. Mare Island was never surveyed as a Mexican grant, so it appears in no boundary set. Its polygon was traced from Ringgold’s 1850 U.S. Coast Survey chart of Vallejo and Mare Island Strait, georeferenced by the chart’s own scale bar and anchored to the historic Navy Yard, and it is deliberately narrower than the modern island because the island was enlarged by fill after 1853. It is an approximation drawn from a period chart, it is labelled as such on the grant page, and it should not be read as a survey.

10. Accuracy audit

A blind sample of 100 of the 672 records was drawn on 1 September 2026, stratified by granting era, using a fixed public seed, and committed to the public repository with its sha256 before any record was checked. Anyone can redraw it and confirm the ids match. Eight fields are graded per record against the land case file and the patent record: year, governor, grantee, land case number, outcome, patentee, patent date and General Land Office number.

Status: sample committed, verification in progress, 0 of 100 checked. The rate will be posted here per field and overall, together with every correction the audit turns up, and a partial audit will be reported as partial. The design and the grading rubric are public: see audit/README.md in the portal repository.

11. How to cite this portal

Cite the portal as a database, name the specific grant if you are citing one record, and give an access date, because the data is periodically rebuilt (as the count history in Section 6 shows).

Whole portal:

Coyne, Aodhan. Ranchos of Spanish and Mexican California: A Research Portal. ranchos.archivesofcalifornia.com. Accessed [date].

The register as a citable dataset:

Coyne, Aodhan. Ranchos of California: A Register of the Land Grants of Alta California. Version 1.0.0, Zenodo, 2026. doi:10.5281/zenodo.22185500

A single grant record:

"Rancho [Name]" (U.S. land case [ND/SD nnn]). In Aodhan Coyne, Ranchos of Spanish and Mexican California: A Research Portal, ranchos.archivesofcalifornia.com. Accessed [date].

When you rely on a datum the portal itself sourced, cite through to the original: Hoffman (1862) for a case outcome or party, the Bancroft land-case file (with its digicoll record) for archival detail, the ECAI/UCSD dataset for a boundary, and the Bancroft Land Case Maps (with credit) for a diseño image. The portal is the finding aid; the authorities above are the record.

Provenance of this page: every figure and quoted source line was read directly from the portal's own data files (ranchos-experimental.json, hoffman-claims.json, ranchos-index.json, disenos-data.json, and the register build) on 2026-08-26. Where this document quotes a source note, the quotation is verbatim from those files.