Tracing a Family Through Federal Land Patents and Homestead Case Files
When the United States handed a piece of ground to a private person for the first time, it issued a patent and it opened a file. The patent is indexed online, free, one row per parcel: a name, an acreage, a date, an authority, and a legal description written in townships and ranges. The file is not indexed at all. It sits in Washington, it can run to dozens of pages, and it holds the things the row does not — sworn proof of who lived on that ground, what they built, who swore they saw it, and where they were born. Almost everyone finds the row. Very few people ever order the file.
What this page settles
There are two records here, held by two different institutions, and confusing them wastes months. The federal government disposed of the public domain through the General Land Office. Its patent index is searchable free at the Bureau of Land Management, and it will tell you that a named person received a named acreage on a named date under a named act. Behind each of those rows is a land entry case file, and that is where the family history lives. Proof of residence. A description of the house, the well, the broken sod, the fence. Sworn statements from neighbors who had to testify they watched the claim being lived on. Evidence of citizenship for an immigrant claimant. Those files are at the National Archives in Washington, they are not name-indexed for most of the period, and you order them one at a time. Two warnings before you start. First, twenty states never had a federal public domain at all, so if your family farmed in Virginia, Texas or Kentucky this entire method is unavailable and the research belongs in a state archive. Second, this is a nineteenth and early-twentieth-century record. It can prove a generation and place a family on a specific square mile. It is not a current address, and no technique on this page turns it into one.
Watch: the row in the index versus the file in Washington
What a federal patent record contains, what the case file behind it contains, and which states the whole method simply does not reach.
Watch the walkthrough
Two Custodians, One Piece of Ground
The county courthouse is not where this record lives, and knowing that on day one saves a wasted trip.
Most people who go looking for an ancestor’s land start at a county office — a Recorder of Deeds, a Register of Deeds, a County Clerk depending on the state — because that is where land records are supposed to be. For land that was bought and sold between private parties, that is exactly right. For the moment the ground stopped being federal and became somebody’s farm, it is the wrong building entirely.
The distinction is one of law rather than filing convenience. A county recorder maintains the chain of title: private owner to private owner, deed after deed, in a book indexed by grantor and grantee. That chain has to start somewhere, and in more than half the country it starts with a sovereign act — the United States conveying land out of the public domain to a private person for the first time. The instrument that does it is a patent, and it was issued by the General Land Office in Washington, not recorded by a county.
So the two records answer two different questions. The county tells you who owned the parcel in 1911 and who they sold it to in 1923. The federal record tells you who got it out of the government’s hands to begin with, and under which of the more than forty separate acts of Congress that governed land entry. Both are real. Only one of them is in the courthouse.
Where the federal side actually sits today
The General Land Office no longer exists as such; its functions and its paper passed to the Bureau of Land Management, and the historical case files went to the National Archives. That split matters because it decides who you write to.
The Bureau of Land Management runs the public search. Its own land records service page describes the General Land Office Records website as covering federal land conveyance records for the public land states, offering images of more than five million federal land title records issued since 1820, along with survey plats and field notes reaching back to 1810. The Bureau also states plainly on that page that the collection, extensive as it is, does not contain every federal title record ever issued. That sentence is worth taking seriously; a blank result is not proof of nothing.
The National Archives holds the files. Its land entry case files page puts the scale at over ten million individual land transactions in its custody, covering entries in all thirty public land states, and describes those case files as the land records generally of most interest to genealogists.
There is one wrinkle that catches researchers who need the tract books — the volumes arranged not by name but by township, range and section, which record every entry attempted on a given piece of ground. The Archives explains the division: for the western states the tract books are at the National Archives Building in Washington and at some field offices, while for the eastern public land states the Bureau of Land Management holds both the tract books and the patents, through its Eastern States office in Washington, DC. If your ancestor entered land in Ohio or Mississippi, the tract book you want is at a different agency from the one that holds an Idaho tract book.
Why this is not the same trace as a county land search
A present-day question about a parcel — who holds it now, who is paying the taxes, which company sits between the family and the title — is county work and mapping work, and it is a genuinely different discipline from archival research. Where a reader arrives wanting the modern answer rather than the historical one, the practical route is set out in our piece on tracing the present holder of a family farm. The federal patent is the very first link in that chain, not a substitute for it.
Thirty States Had a Public Domain. Twenty Never Did.
The first question is not what your ancestor did. It is whether this record exists for the state they did it in.
This is the fact that should appear at the top of every page on this subject and almost always appears at the bottom of one. The federal land entry system only ever operated on land the United States held as public domain. Where the United States never held the land, there is no General Land Office record, there is no patent index row, and there is no case file to order. Not a thin one. None.
The National Archives publishes both lists. The thirty public land states are the eastern group — Alabama, Arkansas, Florida, Illinois, Indiana, Iowa, Louisiana, Michigan, Minnesota, Mississippi, Missouri, Ohio and Wisconsin — and the western group: Alaska, Arizona, California, Colorado, Idaho, Kansas, Montana, Nebraska, Nevada, New Mexico, North Dakota, Oklahoma, Oregon, South Dakota, Utah, Washington and Wyoming. Thirteen and seventeen. Thirty.
The other twenty were never part of the original public domain, and the Archives names them too: the thirteen original states — Connecticut, Delaware, Georgia, Maryland, Massachusetts, New Hampshire, New Jersey, New York, North Carolina, Pennsylvania, Rhode Island, South Carolina and Virginia — plus Hawaii, Kentucky, Maine, Tennessee, Texas, Vermont and West Virginia. The Archives’ own definition of federal public land is territories and later states that were not among the original thirteen colonies, or were at some point their own country, which is exactly why Texas sits on the wrong side of the line despite being the second-largest state in the union.
For those twenty, the Archives’ instruction is short and it is the right one: start with the state archives or the state historical society. Vermont, Kentucky, Tennessee, Maine and West Virginia each disposed of land under their own arrangements; Texas ran its own land office as a republic and then as a state. The records are frequently excellent. They are simply not federal, and nothing at the Bureau of Land Management or the National Archives will produce them.
Public land is not the same as land the government still owns
A confusion worth heading off. “Public land state” is a historical category about where the federal public domain once was, not a statement about what the government holds today. Illinois and Ohio are public land states with almost no federal land left in them; the entire point of the General Land Office was to give the public domain away. If you are researching an Illinois farm, you are in the right system even though nothing about modern Illinois looks federal.
Not every entry in a public land state is a homestead
The word “homestead” has become shorthand for the whole system, and it is misleading. The Archives warns about this directly in its own guidance: the term is widely used, but a claim on the same ground might have been made as a mining claim, a cash entry, a military bounty warrant, a forest reserve entry or a scrip location, and all of those were recorded separately from a homestead entry. Its ordering guidance lists the acts that generated the richest files — the Cash Act of 1820, the preemption acts of the 1830s and 1840s, the Donation Act of 1850, the Homestead Act of 1862, the mineral acts of the 1860s and 1870s, the Timber Culture Act of 1873, the Desert Land Act of 1877, and several military bounty land warrant acts of the 1840s and 1850s.
Over forty separate legislative acts were used to make entries on the public lands, and the Archives is explicit that what ends up in a case file depends largely on what the particular act required the entryman to prove. That is the single most useful sentence in the whole subject, and it is the reason two families who both “homesteaded in Nebraska” can receive files of wildly different thickness.
The Index Row: A Name, an Acreage, and a Square of Ground
What the free federal search gives you, why it looks disappointing, and why it is still the key to everything else.
Start at the General Land Office Records site, which is free and requires no account. You can search by patentee name, narrowing by state and county, or you can search by land description if you already know where the ground is. What comes back for a match is a single row and a scanned image of the patent itself.
That row is not a family history. It is a title record, and it was designed to answer a title question. What it reliably carries is the name of the person who received the patent, the acreage, the date the patent issued, the land office that issued it, the legal authority under which the entry was made, and the legal description of the parcel. The National Park Service’s guidance for the Homestead National Historical Park describes the same database in the same terms and adds the most useful framing of it: researchers use this site to confirm whether someone homesteaded or obtained land by some other means. Confirmation, not narrative.
The three things the index will not tell you
First, it does not name the family. A patent is issued to one person. A wife, children, parents and in-laws who lived on that quarter section for five years appear nowhere in the row, though several of them may appear in the file.
Second, and this one costs people real time: the index only covers claims that were completed. The Archives states it twice, on two different pages, because it matters so much. The site indexes patented or completed land entries and does not index cancelled or relinquished ones. So an ancestor who filed, broke sod for three years, lost the crop and walked away leaves no trace in the index at all — and yet, as the Archives points out, an unpatented case file can hold as much information as a patented one. Absence from the index is a statement about how the claim ended, not about whether the person was there. When that is the situation, the route is to contact the National Archives facility holding records for that state rather than to conclude the family was somewhere else.
Third, it is not complete even for patents. The Bureau of Land Management says so on its own service page.
Reading the legal description, because you will need to quote it back
The strings that make people give up look like this: the southeast quarter of Section 12, Township 4 North, Range 9 West. They are not a code. They are a coordinate system, and it is older than the Homestead Act by three-quarters of a century.
The National Archives’ account of the origin is precise. The Land Ordinance of 1785 imposed a standardized system of federal land surveys to replace the practice of stepping off plots from landmarks, which had produced overlapping claims and constant border disputes. Using astronomical starting points, territory was divided before settlement into six-mile squares called townships, and each township was divided into thirty-six sections of one square mile — 640 acres — apiece. The National Park Service describes the same rectangular system as mandated by Congress in 1785, with each piece of land identified by numbered section, township and range.
Read from the inside out. The section is your square mile, numbered one to thirty-six within its township. The township and range numbers say which six-mile square you are in, counted north or south of a baseline and east or west of a principal meridian. The quarter and half fractions in front cut the section down: a quarter section of a 640-acre section is 160 acres, which is precisely the maximum a Homestead Act claim could cover, and that is not a coincidence. The Act was written to the survey.
Learning to read it is not optional, and here is why. The National Park Service explains that General Land Office clerks were focused on which tracts of public land had been claimed, not on the names of the people claiming them. The system was filed by ground. Which means that for a large part of the period, the legal description is not a detail you record for tidiness — it is the retrieval key, and without it the file cannot be pulled.
One practical note the Park Service adds: if you do not have the legal description, a county recorder of deeds where the land sat can often supply it, because it follows the parcel into every subsequent deed. That is the one place where the county office genuinely helps with a federal search, and it is also why a nineteenth-century signature on a recorded conveyance is sometimes the thread worth pulling — a problem we take up separately in the note on working outward from a signature on a deed.
The Case File: What Stayed in Washington
Twenty-four pages for one Dakota farm, and the reason so few researchers ever see them.
To see the size of the gap, look at what the National Archives itself has published. Its sample homestead file for Charles Ingalls — the father of Laura Ingalls Wilder — is presented as twenty-four separate page images, one thumbnail after another, with a separately catalogued item titled “Witness for Ingalls Homestead” listed alongside it. Twenty-four pages, plus a witness record, behind what the patent index shows as one line.
That is the whole argument of this page in a single example, and it is free to look at before you spend anything.
What the paperwork was for
The richness is not an accident of archiving. It is a consequence of what the law demanded. The Archives’ account of the Homestead Act of 1862 lays out a three-part process: file an application, improve the land, and file for deed of title. Any citizen or intended citizen who had never borne arms against the United States government could claim up to 160 acres of surveyed government land. For the next five years the General Land Office looked for a good-faith effort — meaning the homestead was the claimant’s primary residence and that improvements were being made on it. After five years the homesteader could file for the patent by submitting proof of residency and the required improvements to a local land office, which forwarded the paperwork to Washington with a final certificate of eligibility, where the case file was examined.
Every one of those steps generated a document, and every document had to convince a clerk who had never been to Dakota Territory. That is why the file describes the house and the well and the acres broken. It was evidence.
What a file can contain
The Archives sets out the range on its ordering guidance, and it is deliberately worded as a range rather than a promise. Case files can attest to the entryman’s age, place of birth, citizenship, military service, literacy and economic status, and sometimes carry similar information about family members. They provide evidence of first title, and occasionally of land use, rights of way, survey lines, crops and improvements, archaeological imprints and conflicting claims. The National Park Service, describing the same files, says they describe the improvements made to the property and include many details of a homesteader’s life.
Two of those items are worth pulling out.
The witnesses. Proof of residence required people who had watched it happen, and the neighbors who signed those statements were rarely strangers. Brothers-in-law, cousins, the family two claims over whose daughter later married into yours. A witness list is a documented association at a fixed date, and associations are frequently the thing that breaks a stalled line — which is also the logic behind working sideways to the descendants of an ancestor’s sibling rather than pushing straight back up a direct line that has stopped.
The citizenship evidence. Because the Act required a citizen or an intended citizen, an immigrant claimant had to show where they stood, and that is why naturalization material turns up in these files at all. Treat it as a pointer rather than the whole answer: it tells you a court somewhere handled that person’s citizenship and roughly when. Naturalization is a separate record system with its own rules about which court held what and when a single federal form finally arrived, and that is properly its own subject rather than a paragraph here.
Why a five-year file can be six months thin
Here is the disappointment nobody warns about, and it has a specific cause. The Homestead Act had a second route to title. The Archives describes it plainly: title could also be acquired after a six-month residency and trivial improvements, provided the claimant paid the government $1.25 an acre. That is commutation, and a great many claimants took it — sometimes because they had the cash, sometimes because five years on that particular ground was not survivable.
A commuted claim still produces a patent and still shows up in the index looking like a homestead. What it does not produce is five years of proof. No long residency affidavit, far less testimony, a much thinner file. If you order a case file and receive eight pages of transactional paperwork, the likeliest explanation is not that the Archives sent the wrong file. It is that your ancestor bought the ground out.
One more variation, and it is a useful one for Civil War descendants: the Archives notes that after the war Union soldiers could deduct the time they served from the residency requirement. A three-year veteran therefore reached patent faster, and the deduction itself is a documented statement about service that sits inside a land record.
The scale of the thing
For context on how many families this touches: by 1934, according to the Archives, more than 1.6 million homestead applications had been processed and more than 270 million acres — ten percent of all United States land — had passed into private hands under the Act. The Federal Land Policy and Management Act of 1976 repealed the Homestead Act in the forty-eight contiguous states while granting a ten-year extension for claims in Alaska. So the record runs from 1863 to the middle of the 1980s at its outer edge, and the great bulk of it sits between the Civil War and the First World War.
What the Archives Needs Before It Can Pull Your File
The prerequisites change with the era and, for the earlier period, with the state. This is where most requests stall.
Requests go in on NATF Form 84 or through the Archives’ online ordering system. What you have to supply first is not one list; it depends on when the entry was made, because the General Land Office reorganized its filing in July 1908. Before that date, files were kept by state, land office, type of entry and final certificate number. From July 1908 the Office abandoned state-by-state numbering and put every general land entry into one series ordered by serial patent number — the serial patent files at the National Archives Building run from roughly 167 to 1,242,597. The table below sets out what each route needs, drawn from the Archives’ own accessing land entry records guidance and from the Park Service’s state-by-state notes.
| Which route you are on | What you must supply | Where that information comes from |
|---|---|---|
| Entry from July 1908 onward | Name of the entryman and the serial patent number. Name indexes exist for all the public domain states for this period, so the approximate date of entry and the state are usually enough to get started.Easiest route | The serial patent number is generally the “Accession Nr.” field in the General Land Office database, with an issue date of July 1908 or later. |
| Pre-July 1908, in a state with a name index | Name of the homesteader, the state, and the approximate date of entry. The Park Service names these states as Alabama, Alaska, Arizona, Florida, Louisiana, Nevada and Utah. | A name index at the Archives does the lookup, so the legal description is not required first. |
| Pre-July 1908, western state with no name index | Name, state, approximate date of entry, and either a legal description of the land or the land office name plus the land entry file number. Applies to California, Colorado, Idaho, Kansas, Montana, Nebraska, New Mexico, North Dakota, Oklahoma, Oregon, South Dakota, Washington and Wyoming. | From the General Land Office database — except for Kansas, Nebraska and Oklahoma, where the Park Service says that route does not supply it. |
| Pre-July 1908, eastern state with no name index | Name, state, approximate date of entry, the file number, and the name of the land office that issued the file. Applies to Arkansas, Illinois, Indiana, Iowa, Louisiana, Michigan, Minnesota, Mississippi, Ohio and Wisconsin. | From the General Land Office database — except for Iowa, which the Park Service excludes. |
| Any pre-1908 general entry, as the Archives states the minimum | State, land office, type of land entry, and final certificate number — sometimes called the file or document number. | Type of entry is generally the “Authority” field in the database; the certificate number is generally the “Document Nr.” field. |
| A surrendered military bounty land warrant | The year of the congressional act authorizing the warrant, the acreage granted, and the warrant number. Warrants were typically issued in forty-acre increments. | The authorizing act is generally listed in the “Authority” field. |
| A cancelled or relinquished claim | Not orderable through the index at all, because the index does not cover it. Contact the National Archives facility holding records for that state. | Land office records, tract books, and the Archives’ regional name indexes to cancelled, rejected and relinquished files for particular offices. |
Two states the Park Service’s own lists handle badly
Read those two pre-1908 rows against each other and something does not add up, so we are going to say it plainly rather than let you discover it halfway through an order. Louisiana appears in both. The Park Service names it among the seven states for which a name index exists, and it names it again among the eastern states for which no name index exists and a file number and land office are therefore required. Those cannot both describe the same claim. We have reproduced both lists exactly as the Park Service publishes them, because inventing a tidier version of a federal agency’s list is a worse error than reporting the one it actually prints.
Missouri appears in neither. It is a public land state — the Archives names it in the eastern group of thirteen — but the Park Service’s pre-1908 breakdown does not place it in the name-index list or in the eastern no-index list, so a Missouri researcher reading that page alone falls between the two and is given no route at all.
What to do in either case is the same, and it is already in the table’s last general row: fall back to the minimum the Archives itself states for any pre-1908 entry — the state, the land office, the type of land entry, and the final certificate number. That set satisfies the heavier of the two Louisiana routes, so it works whichever one turns out to apply, and it is a complete request for Missouri regardless of which list Missouri should have been on. All four of those values can generally be read off a single hit in the General Land Office database. If a name index does exist for your state, you will simply have supplied more than was needed, which costs you a database lookup and nothing else.
Three things to expect once the request is in
It is slow, and the Archives says so. The Park Service’s ordering guidance asks researchers to allow several months between ordering and receiving reproductions of the case files. Plan the rest of your research around that rather than waiting on it. For questions rather than orders, the Archives publishes a reference address for land entry files at Archives 1 and asks for ten to twelve business days for a reply, given the volume.
There is a fee, and we are not going to quote it. Reproduction charges change, and a stale figure on a page like this is worse than no figure, so check the current amount at the point of ordering rather than budgeting from something you read here.
Terminology has its own publication. The Archives’ long-standing background guide, General Information Leaflet 67, Research in the Land Entry Files of the General Land Office (Record Group 49), has been superseded by Reference Information Paper 114. If an archivist or an older guide refers you to GIL 67, that is the document that replaced it. We name it rather than link it, because the address we tried for it does not currently resolve and we will not publish a link we could not open.
Some of it is already digitized, unevenly. Selected homestead records for claims completed by 1908 in a set of states have been digitized and are searchable through subscription genealogy databases and free of charge at Homestead National Historical Park in Beatrice, Nebraska. Which states are covered has changed as the project has progressed, so check the Park Service’s current requesting-records page rather than trusting any list, including one written today.
Six Correct Records Read the Wrong Way
None of these is an obscure edge case. Every one of them is a record behaving exactly as designed.
Concluding the family was elsewhere because the index is silent
Cancelled and relinquished entries are not in the patent index. A claim abandoned in year four leaves a file at the Archives and nothing online. The silence describes how the claim ended, not where the family lived.
Searching a state that never had a public domain
Virginia, Texas, Kentucky, Tennessee, Maine, Vermont, West Virginia, Hawaii and the rest of the original thirteen produce nothing here, ever. That research belongs in a state archive, and knowing it on day one is worth a month.
Assuming every claim was a homestead claim
The same ground could be taken by cash entry, mineral claim, timber culture, desert land, scrip or a surrendered bounty warrant, each filed separately. Searching only homestead authority hides the entry that actually happened.
Discarding the legal description as a formality
For most of the pre-1908 period the files were arranged by ground, not by name. The township, range and section string is the retrieval key. Copy it exactly, fractions and all, the first time you see it.
Reading a thin file as a filing error
Commutation gave title after six months and a cash payment. A short file is usually a bought-out claim, and pre-1840s entries can hold nothing but a name, a location, an acreage, a price and a date. The file is the file.
Treating the patentee as the only person on the land
One name goes on the patent. A spouse, adult children, in-laws and hired hands may all have lived on that quarter section, and several of them are likely to be named inside the file as witnesses or as the household proved up on.
How We Run a File That Starts With a Township and a Range
Four stages. The order is not stylistic; each one produces what the next one needs to exist.
Establish that the state and the era can answer at all
Public land state or not, and roughly when. If the ground is in one of the twenty states outside the public domain, we say so before anything is billed and point the research at the state archive instead of taking a fee to confirm an absence.
Pin the parcel, not just the person
Name searches come first, but the deliverable from this stage is the legal description — section, township, range, meridian — because for most of the period that is what the file is filed under. Where the index is silent we work the county deed chain backward to recover it.
Order the case file with the right key for its era
Serial patent number after July 1908; state, land office, entry type and certificate number before it, or the state-specific substitute the Archives accepts. A request sent with the wrong key comes back unfilled, and the wait is measured in months either way.
Read the witnesses as hard as the claimant
Every name in the proof testimony is a documented association at a fixed date and place. We index them, check them against the surrounding tract entries, and treat the neighbors as leads rather than as background.
What This Record Cannot Do, and What We Will Not Do With It
Historical research that sometimes ends at a living person’s door needs its rules stated at the front.
Say the limiting thing first: an 1878 patent is a genealogical lead, not a current address. It places a named person on a named square of ground at a named date more than a century ago. Everyone who signed those papers is long dead. Where the work ends in a present-day locate, that is separate work running through entirely different records, which is what our records-based locate practice is for — and it is governed by different rules, which we would rather set out here than bury.
We are a public records research firm and we are not private investigators. Nobody on this team holds a private investigator’s license, and we never let an archivist, a county recorder or a reference librarian believe otherwise. We read records, we file written requests under our own name, and we say who instructed us. Every file opens with a purpose the law permits, written down before the first search runs — an estate that has to identify its heirs, a court that requires evidence of a diligent search, a quiet-title action that needs the chain established, a family reconstructing itself. Wanting a stranger’s whereabouts is not on that list, and we would rather decline at the first email than take the retainer.
We do not pretext. Nobody from this firm telephones the National Archives claiming to be a descendant they are not, lets a convenient wrong assumption sit, or misrepresents who is paying for a request in order to get a file pulled. On land research the temptation is genuinely small, and the principle is not: a reference archivist who knows exactly what we are doing helps us, and one who has been misled remembers. Nor do we go after the contents of private financial accounts — balances, statements, account numbers — which sit behind a bank’s door for good reason and are typically obtained by exactly the pretexting we just refused. A nineteenth-century land entry file is the opposite kind of record: created by the government, open to anyone who asks, cited openly.
Where the Fair Credit Reporting Act draws the line across this work
People Locator Skip Tracing is not a consumer reporting agency and nothing we deliver is a consumer report. Do not use anything on this page, or anything we produce, to decide on a tenancy or a rental application, on employment or a promotion, on credit, on insurance, or on a professional license or benefit. The Fair Credit Reporting Act governs those decisions, and they belong to a screening firm running the consent, disclosure, adverse-action and dispute process that statute requires. We do not operate that process. Where a request turns out to be one of those decisions dressed as family history, we decline it rather than repackage it.
The safety decline, and why land research invites it
Where the person at the far end of a land file appears to have put themselves out of reach deliberately, that file stops here. Ancestral ground is one of the warmest covers available for reaching someone who does not want to be reached. It sounds harmless. It supplies a sympathetic reason for wanting an address. And it is unusually hard for an outsider to test, because the paper trail genuinely exists and genuinely leads somewhere. So a land inquiry that turns into a request for a living person’s current whereabouts gets more scrutiny from us at intake, not less. Domestic violence; an order of protection; a pattern that reads as stalking; a custody fight being conducted somewhere other than in front of a judge; a relative who has already stated, unambiguously, that no contact is what they want — any one of those and the work ends there, and it does not resume because a more carefully drafted version of the same request turns up a fortnight later. An address is not a thing a relative is owed. People settle where they settle, and the distance some of them keep has been weighed a good deal more carefully than the request to close it.
Two situations come up often enough to name. The first is the heirship approach: someone reads a patent, concludes that a mineral interest or a sliver of surface was never properly conveyed away, and now wants to stand on the porch of a family that has never heard the name. Our answer does not vary. We will write on your behalf, put your own contact details in the letter, and leave the next move entirely with whoever opens it. What we will not supply is a street number so that first contact can be an unannounced knock. A genuine claim on that ground is still a genuine claim two weeks later; it does not need surprise to work. The second is the reverse, and it is worth knowing about from the other side: unconveyed mineral and surface interests attract approaches from strangers, some of them legitimate and some of them not. If someone has contacted you about an interest inherited from an ancestor’s patent, ask them in writing for the legal description, the county and the instrument they are relying on, and read anything before you sign it. What happens when a severed interest has genuinely gone unclaimed is a separate problem we set out in the piece on identifying the current holder of a mineral interest.
Everything here is general information about federal records and is not legal advice. Whether a patent still means anything for a parcel today, who inherited an interest, whether a chain of title is good, and what a particular court will accept as proof of descent are questions for an attorney licensed in the relevant jurisdiction. That includes the popular claim that locating or filing an old patent changes a present-day owner’s obligations — we take no position on it, we do not advise on it, and it is not what this research is for. We find the document and show the citation. The legal conclusion belongs to somebody else.
Who Arrives Holding a Township and a Range
Six recurring instructing parties, wanting six different things out of the same square mile.
Pioneer lineage applicants
Needing one generation documented to a standard a membership committee will actually accept.
Quiet-title attorneys
Establishing the first link in a chain of title that has to be proved back to the sovereign.
Mineral and royalty analysts
Tracing a severed interest back to the patent that first put the acreage in private hands.
Ranch succession advisors
Reconstructing how a spread was assembled from separate entries before it was ever a single title.
Boundary surveyors
Wanting the original survey plat and field notes behind a line that two neighbors now disagree about.
Museum and park archivists
Answering the same claim question from visitors and wanting an answer that holds up in writing.
The mineral analysts and the quiet-title attorneys arrive from the opposite direction to everybody else, and it changes the whole shape of the job. They do not have a name and want a place. They have a place — a precise one, already written as a legal description — and want the names attached to it across a century and a half. For them the tract book matters more than the patent index, because it records every entry attempted on that ground rather than only the one that succeeded, and the failed attempts are frequently where a family that “was never there” turns out to have been there for four years. It is one of the few historical searches that runs cleanly from ground to person, which is also why it so often surfaces a surname the present-day family never knew about.
Every finding leaves here with the ground it came from
Nothing goes out of this office as a bare assertion. A land finding travels with the state, the land office, the authority under which the entry was made, the accession or certificate number, and the full legal description down to the fraction of the section — enough that you or your attorney can order the identical file and read it without us. Where the honest result is that the state never held public domain, or that the claim was relinquished and the index will therefore never carry it, that goes out as a documented negative finding with citations of its own rather than as a file we keep describing as active. A locate that never produces a person is never billed.
Questions About Patents, Case Files and Public-Land States
What are federal public lands, and which states are public land states?
Federal public land means territories and later states that were not among the original thirteen colonies, or that were at some point their own country, where distribution of the land was controlled by the federal government through the Interior Department, the General Land Office and later the Bureau of Land Management. There were thirty such states. The eastern ones are Alabama, Arkansas, Florida, Illinois, Indiana, Iowa, Louisiana, Michigan, Minnesota, Mississippi, Missouri, Ohio and Wisconsin; the western ones are Alaska, Arizona, California, Colorado, Idaho, Kansas, Montana, Nebraska, Nevada, New Mexico, North Dakota, Oklahoma, Oregon, South Dakota, Utah, Washington and Wyoming. Note that this is a historical category, not a description of what the government still owns — Ohio is a public land state with almost no federal land left in it.
How do I find records for a state that was never public land?
Start with that state’s archives or historical society, because the federal system has nothing for you. Twenty states were never part of the public domain: the thirteen original states — Connecticut, Delaware, Georgia, Maryland, Massachusetts, New Hampshire, New Jersey, New York, North Carolina, Pennsylvania, Rhode Island, South Carolina and Virginia — plus Hawaii, Kentucky, Maine, Tennessee, Texas, Vermont and West Virginia. Texas is the one that surprises people most, and the reason is that it was an independent republic before statehood and kept control of its own land. These states often have superb early land records. They are simply held by the state rather than by the Bureau of Land Management or the National Archives.
How do I find my ancestor’s homestead records?
Work out when and where they settled first — census work is the usual way in — then search the General Land Office Records site at the Bureau of Land Management for the name, narrowing by state. If you find them, copy the legal description, the authority, the land office and the document or accession number exactly, because those are the keys to the file. If you do not find them, do not conclude they were not there: the site indexes only patented and completed entries, so a cancelled or relinquished claim is invisible on it. In that case the next step is the National Archives facility that holds records for that state, and they will generally want the legal description in townships and ranges, though a town name may be enough for staff to work from.
What is actually inside a land entry case file?
It varies enormously by the act the entry was made under, and there were over forty of them. At the rich end, a homestead file can attest to the entryman’s age, place of birth, citizenship, military service, literacy and economic status, sometimes carrying the same for family members, alongside descriptions of the house, the well, the acreage broken, the crops and the fencing, plus sworn testimony from neighbors who watched the claim being lived on. At the thin end, an entry made before about 1840 may hold only the name, the land’s location and acreage, the price, and the date and place of entry. The National Archives publishes the Charles Ingalls homestead file as a worked example: twenty-four pages, with a separately catalogued witness record beside it.
How do I order copies of a land entry case file?
Through NATF Form 84, by mail or through the National Archives’ online ordering system. What you need in hand depends on the era. From July 1908 the files are one series ordered by serial patent number, so the entryman’s name and that number are the core requirement. Before July 1908 the files were arranged by state, land office, type of entry and final certificate number, and the Archives asks for all four — though seven states have name indexes for that period, where the name, the state and an approximate date of entry are enough. For a surrendered military bounty land warrant the requirements are different again: the year of the authorizing act, the acreage, and the warrant number. Expect to wait several months, and check the current reproduction fee at the point of ordering rather than relying on a figure published anywhere else.
How long did homesteading take, and why is my ancestor’s file so thin?
The headline route was five years. Any citizen or intended citizen who had never borne arms against the United States could claim up to 160 acres of surveyed government land, and for five years the General Land Office looked for good faith — that the homestead was the claimant’s primary residence and that improvements were being made. After that they could file for the patent by proving residency and improvements to a local land office. But there was a second route, and it explains most thin files: title could also be taken after six months’ residency and trivial improvements on payment of $1.25 an acre. A commuted claim looks like a homestead in the index and produces far less paper. Separately, after the Civil War, Union soldiers could deduct their service time from the residency requirement, which is itself a documented statement about service sitting inside a land record.
What is the difference between a land patent and a deed, and does an old patent mean anything today?
A patent is the instrument by which the United States conveyed land out of the public domain to a private owner for the first time. A deed is a later transfer between private parties, recorded at the county. So the patent starts the chain of title and the deeds continue it, which is why the two records live in two different buildings under two different custodians. As for what an old patent does for a present-day owner: that is a legal question about a specific parcel in a specific state, and there is a large amount of confident material online claiming that locating or filing one changes a modern owner’s obligations. We take no position on it and we do not advise on it. Our work is documentary — we retrieve the instrument, cite it, and hand the legal conclusion to an attorney licensed where the land sits.
Can you find the living family behind a nineteenth-century patent?
Sometimes, and only under conditions we state before starting. The archival half — establishing the entry, ordering the case file, reading the witnesses, building the descent — is measured in months, because the Archives asks researchers to allow that long for reproductions. Carrying a documented line forward to a living person is separate work with its own rules, and where we have a full name and a reasonable starting point that generally comes back within 24 hours. That stage runs on a lawful purpose stated up front, an approach that names us and names the client, and a contact the person can refuse without any friction at all. If the indications are that somebody has removed themselves on purpose, the file closes instead of producing an address, and redrafting the instruction does not reopen it.
Send Us the State and the Surname. We Will Tell You If the File Exists
Give us the family name, the state, and roughly when you think they took up the ground, and we will tell you three things before anybody commits to anything: whether that state was ever public domain at all, whether a patent turns up in the federal index for that name, and what the Archives will need before it can pull the case file behind it. If the answer is that your state never had a public domain and this whole route is a dead end, that is what we will say, and we would rather say it in the first message than the fourth. Tell us which piece of ground you are trying to trace and we will tell you where the paper is.
Open a records research file
