Finding an American Father From a Base, a First Name and a Year
Almost everyone who writes to us about this apologizes for how little they have. A first name, sometimes only a nickname. An installation. A stretch of months. That is not a hopeless starting set for the archive work. It is a workable one, because the American armed forces did not document themselves person by person – they documented themselves by unit and by station, and a unit is something you can pin down. This page shows how to turn a place and a period into a finite list of men, what the 1973 St. Louis fire did and did not take, what the archives will hand a person with no standing at all, and where DNA takes over.
One thing to know first: a first name or nickname with a base and a year is not something we can run a search on. It gives a private records search no person to look for, and the archive and DNA work below has to come first. We work subjects inside the United States only, so this has to be a man now living in the United States. Once the unit work, a records request or a DNA match has given you his full name, a search for a current address becomes possible.
The Short Version
Work the station, not the surname. Establish which units held that installation during those months and the population of possible men stops being everyone and becomes a roster. The fire that everybody warns you about is not the obstacle it sounds like. Army personnel files for discharges between 1912 and 1960 were eighty percent destroyed in July 1973, but the unit-level daily reports for almost exactly that period were carried out of the building and survive – the record of the outfit outlived the record of the man. You do not need standing to learn where he served. The National Archives publishes a list of fifteen items releasable from a non-archival file to anybody who asks, and assignments and geographical locations are on it. After sixty-two years from separation the whole file opens. DNA is what closes most of these cases now, and it identifies a family before it identifies a person – which is why the unit work still matters, since it tells you which men in that family were actually there. Citizenship is a different question with its own deadlines, and it belongs to an immigration attorney rather than to a records search.
Watch: Turning a Station Into a Roster
Why a base and a year are better raw material than a common surname, and what the archives do with them.
A short overview
The Outfit Was Written Down Every Single Day
Individual servicemen were recorded through the organizations they belonged to. That is why a place and a period are worth more here than a surname.
Consider what a common surname is actually worth on its own. There were tens of thousands of Americans in Europe in any given post-war year and tens of thousands more across the Pacific, and a nationwide index of men by name is a haystack with no edges. Now consider what a station and a date range are worth. At any moment, a finite number of named organizations held that installation: a squadron, a battalion, a hospital, a supply depot, a military police detachment. Each of those organizations produced paper on a fixed schedule, and each entry on that paper carried names.
The Army’s daily accounting document is the morning report. Air Force units kept the same kind of record. These were organizational rather than personal – they belong to the unit, not to a man – and that distinction is the whole reason they are useful to somebody who cannot name a man. The National Archives treats them as the material from which an individual’s service can be put back together: describing the aftermath of the 1973 disaster, it records that morning reports and similar organizational holdings “have proved vital in efforts to reconstruct basic service information for requestors.” If a federal agency reconstructs a soldier out of unit paper when his own file is gone, so can a researcher who never had his file to begin with.
What the archives ask for tells you what matters
The clearest signal comes from what the National Personnel Records Center demands before it will look. Its published conditions for access to morning reports and unit rosters ask for “the exact unit of assignment” – the page gives its own worked example, “Company A, 1st Battalion, 116th Infantry Regiment” – along with a description of the action and an approximate date, and states plainly that without that kind of information the Center may be unable to perform a search. On the separate page covering ordinary records requests, the archives note that for records caught in the fire, “additional information, such as place of discharge; last assigned unit; and place of entry into service may be useful.”
How you are actually allowed to read them
The access rules for these reports catch people out, because they run opposite to the way the rest of the system works. Army morning reports and unit rosters dated 1912 to 1959, and all Air Force morning reports, are archival records – accessioned into the National Archives and open to the public. But the archives state flatly that written requests for archival morning reports and unit rosters are not accepted. There are two ways in and only two: visit the Archival Research Room at the National Personnel Records Center in St. Louis in person, or employ an independent researcher to examine them on your behalf. Reviewing a record in that room costs nothing; copies are charged at the agency’s published fee schedule.
Army morning reports and unit rosters dated 1960 to 1974 work the other way round. Those are non-archival, so they are obtained by a written request, signed in cursive and dated within the past year, sent by mail or fax to the Center – and this is where the exact unit of assignment must appear. For anyone searching from outside the United States, that difference decides the whole plan: the older material needs somebody physically in Missouri, and the newer material needs a correctly filled form.
Read those two requirements together and the sequence of a sensible search falls out on its own. Nobody at the archives can find your father from a base and a year. But a base and a year are exactly what you need to derive the unit, and the unit is the key that opens the door the archives are holding shut. The order of operations is the finding, not a trick.
Building the roster of units
Establishing which organizations held an installation in a given period is library and archive work rather than database work. Installation histories, unit lineage and honors records, station lists, and the published administrative histories of the commands that governed a theater will each carry part of it. The Library of Congress maintains a research guide to locating military unit records that is a fair map of the terrain and costs nothing to read. Two service history organizations hold much of the primary material and are worth naming even though we cannot link to them here: the Air Force Historical Research Agency at Maxwell Air Force Base in Alabama, which holds unit histories for the air arm, and the U.S. Army Heritage and Education Center at Carlisle, Pennsylvania. Both take written research inquiries.
What you are trying to produce at the end of this stage is not a name. It is a short, defensible list: these four or five organizations were at that station in those months, here is the citation for each, and here is where their rosters and daily reports now live. Everything after this is comparatively mechanical.
Where the base itself is the only certainty
Sometimes the mother remembered a place but not a period, or a period but not a place, and sometimes the place was recorded in a hospital record in a language that transliterated it oddly. Work outward. Local civil registration in the town near the installation, church baptismal records, the maternity hospital’s own admission books, and any welfare or municipal file opened at the time can each fix the date within a month or two, and a two-month window is usually enough to reduce a station’s unit list to something manageable. This is the same kind of patient lateral work described in our guide to breaking through a genealogy brick wall, applied to a military rather than a civil record set.
What Burned in St. Louis, and What Somebody Carried Out
The fire is real and the losses are enormous. It also spared, almost exactly, the records this particular search depends on.
Anybody who researches this topic for an afternoon runs into the 1973 fire and often stops there, having concluded the trail is ash. The facts are worse than most people realize and the consequence is better. Shortly after midnight on 12 July 1973, fire broke out on the sixth floor of the National Personnel Records Center on Page Boulevard in St. Louis. It burned out of control for twenty-two hours, took the participation of forty-two fire districts, and was not declared out until 16 July. The National Archives account of the fire puts the destruction at approximately sixteen to eighteen million Official Military Personnel Files.
The losses were not spread evenly. The archives publish them as a short table, and it is worth having the exact boundaries rather than a general impression:
- Army – personnel discharged 1 November 1912 to 1 January 1960: eighty percent destroyed.
- Air Force – personnel discharged 25 September 1947 to 1 January 1964, with surnames falling alphabetically after Hubbard, James E.: seventy-five percent destroyed.
Two things follow that people rarely take in. The first is that the table names only those two branches and those two windows. A father who served in the Navy, the Marine Corps or the Coast Guard is not in the loss at all, and neither is an Air Force man whose surname sorts before that alphabetical break. The second is that there was no backup of any kind. The archives state that no duplicate copies were maintained, no microfilm was produced and no indexes existed beforehand, which is why a complete list of what was lost has never existed. Nobody can tell you with certainty that your father’s file burned. They can only tell you the odds for his branch and his discharge window.
The detail that changes the search
Now the part that almost no page about this subject mentions. While the building was still burning, staff pulled certain holdings out for safekeeping. Among them, by the archives’ own account, were the Center’s operating records, a computer index to a major portion of its holdings, and more than one hundred thousand reels of morning reports – Army covering 1912 to 1959, Air Force covering 1947 to 1959. The account then explains why that mattered so much: officials determined that the fire damage “had been the worst among the Army and Air Force records for this same time period.”
Set those two spans beside each other. The individual files destroyed run 1912 to 1960 for the Army; the unit-level daily reports rescued run 1912 to 1959. The overlap is nearly total. In the exact window where a father’s own file is most likely to be gone, the record of the organization he belonged to is most likely to survive. That is not a consolation prize, it is the reason a unit-first search is the right method for precisely the era most of these births fall into.
What the Center built afterwards
Reconstruction has been going on ever since. Roughly six and a half million burned and water-damaged files were recovered and indexed into what the Center calls its B registry, with an R registry created in April 1974 for records rebuilt from other sources. Those other sources are worth knowing about because several are searchable by an outsider: Veterans Administration claims files, individual state records, Multiple Name Pay Vouchers from the Adjutant General’s Office, Selective Service System registration records, pay records from the Government Accounting Office, medical records from military hospitals, entrance and separation x-rays, and organizational records. A man whose personnel file no longer exists may still be documented across half a dozen of those. Selective Service registration in particular ties a name to a birth date, a physical description and a home address, and it survived independently of St. Louis.
Sixty-Two Years After He Left the Service, the File Opens to Anyone
A rolling threshold decides whether you are asking for a restricted federal record or reading a public one.
Every military personnel file lives on one side or the other of a single line, and which side it is on governs everything about how you get at it. The National Archives explain the rule on their page on public access to personnel files: a file is accessioned into the National Archives and becomes archival sixty-two years after the servicemember’s separation from the military, and that is a rolling date – the current year minus sixty-two. Take the year you are reading this, subtract sixty-two, and any separation on or before the result is archival.
An archival file is a record of the National Archives, open to the general public. No relationship to the veteran is required and no signature from anyone is required; you order copies and pay the archives’ published copying fee. That single sentence resolves a great many searches. A father who was demobilized after the Second World War, or who left the service during the occupation years or in the early Korean-war period, is almost certainly on the open side of the line already.
Below sixty-two years, the file is a Federal record and access narrows sharply. Only the veteran, the next of kin, or an authorized representative may obtain it, and the archives define the next of kin precisely as the un-remarried widow or widower, son, daughter, father, mother, brother or sister of the deceased veteran.
A correction worth making, because it is usually stated wrongly
You will read in a great many places that an unacknowledged child born abroad should simply ask for the file as next of kin, on the strength of son and daughter appearing on that list. Read the archives’ sentence again, all the way to the end. The category is defined as those relations of the deceased veteran, and the archives make that condition operational rather than decorative: a next-of-kin requester must supply proof that the veteran has died — a death certificate, a letter from the funeral home, or a published obituary — before the authorization carries any weight at all. So if your father is living, or if you do not yet know whether he is, there is no next-of-kin route to a restricted file, and documenting the relationship will not open one. While he is alive the file is his: he may authorize its release, and an authorized representative acting for him may obtain it, and neither of those is something you can do on your own initiative. That is worth saying plainly because the wrong version of it sends people into a long documentary project aimed at a door that is not there.
What is open to you either way is the release list in the next section, which requires no relationship, no standing and no signature, and which is smaller than a file and more useful for this particular search than it sounds. And if he has already died, the next-of-kin category does become real to you — at which point the proof of death is itself a research task before it is a paperwork one, and the scenario card further down on a father who has already died is where that work starts.
One practical note for the non-archival case. Requests must be signed in cursive and dated within the past year, and the archives will want his complete name as used in service, his service number or Social Security number if you have one, branch, dates of service, and date and place of birth. Very few searchers can supply that set at the start. It is the finishing move, not the opening one.
Fifteen Things the Public May Be Told, With No Standing At All
Even on a restricted file, and even with no relationship established, one of these fifteen items is the one you are actually after.
This is the part of federal records practice that goes unmentioned on almost every page about GI parentage, and it is the part that most often gets a search moving. The National Archives publish a list of what may be released from a non-archival personnel file to a member of the general public without the veteran’s authorization and without any next-of-kin authorization, on the basis that disclosing it is not an unwarranted invasion of privacy. Their explanation of how the Freedom of Information Act applies to personnel files gives fifteen items:
- Name
- Service number
- Dates of service
- Branch of service
- Final duty status
- Final rank
- Salary
- Assignments and geographical locations
- Source of commission
- Military education level
- Promotion sequence number
- Awards and decorations, eligibility only rather than the medals themselves
- Photograph
- Transcript of courts-martial trials
- Place of entrance and separation
Count them and there are fifteen, three of which – salary, source of commission and promotion sequence number – carry an asterisk on the archives’ page marking them as rarely available in the holdings the Center actually maintains. If the veteran is deceased, three further items become releasable: place of birth, the date and geographical location of death, and place of burial.
Why the eighth item is the one that matters
Assignments and geographical locations. That is where he was posted and when, and it is releasable to a person who is, as far as the file knows, nobody at all. If you have narrowed a station and a period to a handful of candidate men, this is the mechanism that tests each candidate against the place and the date without asking anyone’s permission and without waiting sixty-two years. The photograph matters too, more than its position in the list suggests: a face from the right years, put in front of a mother or an aunt who saw him, has settled more of these cases than any document.
Two honest limits. First, this route needs a name – it is a test you apply to candidates, not a way of generating them. Second, everything not on that list stays closed on a non-archival file: the Privacy Act protects the rest, and no amount of explaining your circumstances changes it. The archives are not being obstructive; they are applying a statute that protects a living man’s file, including from his own children where the relationship is not on record.
Five Starting Points, and How Far Each One Reaches
Most searches use several. Knowing which one is doing the work stops you spending months on the wrong one.
| Starting point | What it can reach | Where it stops |
|---|---|---|
| A surname alone | Very little on its own. Useful only once a place and a period bound it | A nationwide name index has no edges. Common surnames make this worse, not better |
| Station and date range | The units present, then their rosters and daily reports. The method this page is built on | Needs archive work or a researcher on site. Slow, and it produces candidates rather than an answer |
| A DNA test | A family. Cousin clusters resolve to shared ancestors and then to a surname group | Rarely names one man by itself. Needs tree building, and needs somebody in the family to have tested |
| Records research on a named candidate | Where he lived after service, whether he is living, his current whereabouts, the relatives around himOur part | Cannot tell you whether he is your father, and cannot make him answer |
| Volunteer and veterans’ networks | Context no file holds – who else was in that outfit, what the station was like, occasionally a direct recognition | Informal and unverifiable. Treat anything it produces as a lead to be checked, never as a finding |
| The pattern across the table | The archives narrow the field, DNA narrows the family, and records research turns a name into a living address. No single row finishes the job, and any service claiming one row does is describing a case it has not had. | |
Our own row sits fourth because that is where it belongs in the sequence. Locating a named living person is the narrow, ordinary trade we practice, described in full on our skip tracing and public-records research page, and it is worth nothing at all until the archives or a DNA match have produced a name to work on. When the name arrives already attached to a service record, the follow-on work overlaps heavily with what we describe in our guide to locating someone who serves or served in the military.
A Test Gives You a Family Before It Gives You a Man
This is how most of these cases actually resolve now. It is also the part people misunderstand most consistently.
Consumer autosomal testing has changed this search more than any archival development of the last fifty years, and it works in a way that surprises people. Almost nobody in this situation tests and finds a father waiting in their match list. What comes back is a scatter of distant relatives – cousins several degrees out, most of whom have no idea who you are. That looks like failure on the first afternoon and it is not. It is the raw material.
The method that resolves it is tree building. You group the matches who are related to each other and not to your mother’s side, follow each group back to the couples they have in common, and watch surnames and localities repeat. The published case studies in this field follow that shape closely: one widely-cited account of an Amerasian search began with nothing closer than a fourth cousin and ended by identifying a common ancestral couple through the clustering of surnames tied to a single locality. The output of that work is not a father. The output is a family, and often quite a large one, containing many men.
Where the unit work rejoins the DNA work
Here is why the two halves of this page belong together. Once you have a family, you need to know which of its men were in the right place at the right time – and that is a question about assignments and stations, which is the question the archives will answer. Run the descendant lines forward, list the men of serving age in the relevant years, and test each against the station and the period. The genetic evidence supplies the candidates; the service record eliminates most of them. Neither instrument does the job alone, and a search that leans entirely on one of them usually stalls at exactly the point where the other one would have moved it.
Two practical realities. Testing at more than one company is not optional advice, it is arithmetic – the databases do not share, and the cousin who cracks the case is in whichever one you skipped. And a match who does not reply has not refused you. Most people check those inboxes once a year at best, and many tested for a company’s ancestry percentages and never looked again. Locating a current address for someone who matched you and went silent is ordinary records work, which we set out in our page on reaching a DNA match who has not answered your messages. If the strongest match turns out to be a half-sibling rather than a cousin, the situation changes considerably, and our guide to identifying a half-sibling through a DNA match deals with that particular pattern.
What a test cannot settle
A match confirms a biological relationship. It does not confirm anybody’s account of how that relationship came about, and it says nothing about what anyone knew or intended at the time. It also cannot compel a living man to test. If he declines, and he is entitled to decline, the genetic case may still be built through his relatives – siblings, cousins, and his other children if they have tested and consent – but that is a slower and more delicate road, and it involves people who did not ask to be part of it.
Citizenship Is a Separate Question With Its Own Clock
We name it because it comes up in nearly every one of these inquiries, and then we hand it to somebody qualified to answer it.
A person who discovers that her father was an American almost always arrives at the same second question, and it deserves a straight answer rather than an evasion. A child born abroad to a United States citizen parent may have acquired citizenship at birth. Whether any particular person did is governed by federal statute, turns on the date of birth, and is not a matter a records-research firm has any business deciding. What we can usefully do is show you the shape of the rule so that the conversation you have with an attorney starts in the right place.
The controlling provision for a person born outside a marriage is 8 U.S.C. 1409. Subsection (a) sets four requirements, all of which must be satisfied: a blood relationship between the person and the father established by clear and convincing evidence; the father holding United States nationality at the time of the person’s birth; the father, unless deceased, having agreed in writing to provide financial support until the person reaches eighteen; and, while the person is under eighteen, one of three things – legitimation under the law of the person’s residence or domicile, the father acknowledging paternity in writing under oath, or paternity established by adjudication of a competent court.
Read the fourth requirement carefully, because it is where most hopes founder. Every route in it has a deadline that ran while the person was a child. An adult who locates her father at fifty cannot go back and have him acknowledge her at fifteen. The versions of the rule in force for earlier births work the same way: subsection (b) covers births between January 1941 and December 1952 and turns on paternity being established by legitimation while the person was under twenty-one, and the USCIS Policy Manual chapter on citizens at birth sets out, era by era, the corresponding requirements for births between December 1952 and November 1986 – legitimation before the age of twenty-one and before November 1986, the person unmarried at legitimation, and a physical-presence requirement on the father of ten years in the United States, five of them after his fourteenth birthday.
What this means in practice is that a DNA test does not create a claim. Genetic evidence can go a long way toward the first requirement, the blood relationship. It does nothing whatever about the third and fourth, which are legal acts with statutory timing. Anyone telling you that a modern test converts into a passport is describing a system that does not exist.
That is as far as we go, and we go that far only so that nobody spends a year chasing the wrong thing. There are other immigration routes that have nothing to do with acquisition at birth, and there have been specific statutory programs for particular groups and particular conflicts. Whether any of them touches your situation is a question for a licensed immigration attorney or an accredited representative, working from your actual dates and documents. What you have just read is general information and not legal advice, and nobody on this team is an attorney; the statute is quoted here so that you can read the primary text yourself and take it to someone qualified to apply it to you.
He May Have a Family Who Has Never Heard of You
The hardest part of this search happens after the finding, and almost nothing published about it prepares anyone for that.
Suppose it works. You have a name, an address, and a man in his eighties in a town in Ohio. Before you write anything, consider who else is in that house and in that town: very often a wife of fifty years, adult children, grandchildren, a church, a circle of friends – none of whom know you exist, and several of whom will learn something about their own family from your letter. A first contact in this situation is not a private message between two people. It is an event in somebody else’s household.
Say nothing about him that you do not know. This is the single most important line on the page. You do not know that he was told about the pregnancy. You do not know that he left, or was posted away with three days’ notice, or wrote letters that were never delivered, or was forbidden by a commanding officer, or believed a marriage was impossible, or simply was never informed at all. Some of these men spent decades not knowing. Others were told and made a choice. A page written by strangers cannot tell you which, and a first letter that assumes the worst forecloses the only conversation you wanted.
Why an intermediary is so often the right call
In adoption practice, contact through a neutral third party is standard rather than exotic, and this situation has the same structure with an extra decade or four of accumulated life around it. An intermediary – a search professional, a counselor experienced in reunion work, a family member on his side who already knows, occasionally an attorney – can make a private approach that gives him room to react without an audience.
What that buys is real. It lets him absorb the news before deciding what to tell his wife. It avoids a letter being opened by somebody else at the kitchen table. It gives a man with a health condition or a cognitive decline a route to a considered answer rather than a shock. And if the answer is no, it lets that be delivered quietly, without the household ever having to discuss it. Families who go direct sometimes do fine. Families who go direct and get it wrong rarely get a second attempt.
He is under no obligation, and that has to be genuinely accepted
A located man does not have to acknowledge you, does not have to take a test, does not have to reply and does not have to explain. There is no legal duty and, sixty years on, no realistic mechanism to create one. Wanting an answer for your entire life does not generate an obligation in him, however unfair that feels – and it is unfair. The searchers who come out of this best are the ones who decided in advance what they would do with a refusal, and who wrote a first letter that could be read comfortably by a man who intended to say no.
It is worth saying that a refusal is not always final. People change their minds after a year, after an illness, after a spouse dies, after a grandchild asks a question. Several of the outcomes we have watched arrived long after the first attempt appeared to have failed. What kills those second chances is pressure – repeat letters, arriving at the door, approaching his children behind his back, or anything that turns a request into a campaign.
One more asymmetry to hold on to. His other children are not adversaries. They are usually the people with the most to process and the least warning, and in the cases that end well they are frequently the ones who eventually open the door. If you reach them, reach them as relatives rather than as a claim.
Six Circumstances That Redirect This Search
Each of these changes which records apply, what we can do, or whether we take the case.
Only a nickname survives
Chuck, Red, Tex and Sarge are not names but they are filters. Applied against a unit roster of a few hundred men, a nickname plus a rank plus a rough age narrows a list far more than it looks like it should.
He was Navy, Marines or Coast Guard
Then the 1973 loss table does not touch him at all, and his personnel file is far more likely to be intact. Ship and shore-station records also run on different series to the Army and Air Force material.
He was a civilian contractor, not a serviceman
Base construction, engineering and support work employed large numbers of American civilians overseas. Their records sit in an entirely different system, and a search built on unit rosters will never find them.
He has already died
Common, and not the end. His file opens further, his obituary names survivors, probate identifies heirs, and his children and siblings are the people who can answer the question and take a test if they choose to.
Your mother does not want him found
She may have reasons she has never shared, and they may be about her safety rather than her privacy. That conflict deserves resolving before a search runs, not after a letter has landed.
Someone is asking on behalf of the child
A grandchild or a nephew researching for an elderly parent is normal and welcome. We will still want to know that the person the answer is for actually wants it, because some do not.
How We Work a Case Like This One
Four stages. The last one is us handing over and getting out of the way.
Everything the family remembers, written down
The place, the years, the fragment of a name, the branch if anyone recalls a uniform, the mother’s own documents, and the reason you are looking, in your own words. The purpose is recorded at intake, and if the honest answer is that the archives cannot carry your facts, you will hear that.
Station and period become a list of units
This stage is archive work, and it is yours or an archive researcher’s rather than ours: installation and command histories, published lineage material and archive finding aids, worked until the organizations present in those months can be named and cited. The sections above show where each record sits. It is the stage that converts an unbounded search into a bounded one, and the stage most searches skip.
A full name tested against the record
Once rosters, daily reports, a records request or a DNA match have given you a full name, we check it against the civil trail after service, for a man now living in the United States. We corroborate identity from independent sources before an address reaches you, because a wrong name here means writing to an innocent household.
You decide what happens next
You receive the person, the address history, and the source behind every finding, generally within 24 hours once a solid identification exists. We do not telephone him, we do not write to him, and we do not tell anyone in his household why somebody was looking. If you want an intermediary, we will say so plainly, and you choose one.