Two federal indexes for LEVI H. COOPER, Private, Company B, 15th Regiment New Jersey Infantry (Union) — the Index to Compiled Service Records (NARA M550) and the Organization Index to Pension Files (NARA T289), pension dated 9 March 1907, file 3263590.

Transcription and notes

Data from the documents [verbatim]

Index to Compiled Service Records — ark 1:1:FSTL-RJF

Organization Index to Pension Files — ark 1:1:NHHY-WNY

Is this our Levi? The honest answer is "probably, and it is not established"

What fits:

• The name with the middle initial H — [D.4], [D.470], [J.232], [E.11] and [E.12] all give him as Levi H. Cooper, and the initial is uncommon enough to matter.
• New Jersey, where he was living in his father's Independence Township household in 1860 ([D.470]) and where [J.232] says he was born.
• The timing of his children. [D.587]'s 1870 Illinois household has children born in New Jersey in 1861 and about 1865, then twins born in Illinois about 1868. A man serving in a New Jersey regiment until 1865 and emigrating westward afterwards fits that pattern exactly.
• The 1907 pension date. [J.232] says he "had been an invalid" in his last years and had come from Brown County, Kansas about 1908; he died in December 1909. An invalid's application in March 1907 is what one would expect.
• [D.587] already established that he is the only Levi Cooper indexed as born in New Jersey in either the 1870 or the 1880 census, and that no Levi Cooper at all was resident in New Jersey in 1880.

What is missing:

• Neither index gives an age, a birthplace, a residence or a next of kin. There is nothing in either record that could distinguish this man from another Levi H. Cooper.
• No image was seen. The service-record index says Image Unavailable, and the pension card itself is served through fold3.com.
• The catalog has no other evidence that Levi served at all — [J.232], his own death notice, does not mention military service, and a Grand Army of the Republic connection would be the ordinary thing for a Kansas veteran's obituary to name. That silence is worth weighing against the identification, not ignored.

⇒ Recorded as provisional. It is a strong lead, not a finding.

Why it is worth pursuing anyway

A Civil War pension file — as opposed to the index card — is among the richest genealogical documents in existence. An invalid's file typically contains the soldier's own affidavits, medical examinations, and, where a wife or children are claimed, marriage certificates, family Bible pages, and sworn statements from relatives naming dates and places. A widow's file goes further still.

For [p.1945] specifically, such a file could bear on four open questions at once:

1. His birth date, which varies across five records from 1830 to 1837.
2. His wives. The catalog has Eliza J. → Permelia → Sarah, with a maiden name for none of them that rests on a record. A pension application claiming a wife must prove the marriage.
3. His children, and whether [p.1969] Irvin was living in 1907.
4. Possibly his parents — which is the [p.113] question, and nothing else available touches it directly.

Next

1. Establish or refute the identification first, before ordering anything. The cheapest tests: (a) the 1890 Census of Union Veterans and Widows — searched 2026-09-05 for a Levi Cooper in Kansas and there is none indexed; (b) his Find A Grave memorial 89793085, whose stone reads only "LEVI H. COOPER / 1831 —— 1909" with no military marker ([G.528]); (c) a G.A.R. post record for Stafford or Brown County, Kansas.
2. The roster of Company B, 15th New Jersey Infantry — a published regimental roster would give the enlistment age and residence, which is exactly what the index cards lack. (New Jersey's adjutant-general's records are the obvious source; the catalog already holds Stryker for an earlier war.)
3. ORDERED 2026-09-07 — and the order supersedes step 1 rather than skipping it. The maintainer bought Gopher Records' "Civil War Soldier's Bundle" on their Labor Day sale — compiled service record + pension file + carded medical records, \$68, against \$140+ ordering the same things from NARA directly. The reason this is not a premature purchase: step 1 says establish the identification first, and the blocker is that neither index card gives an age, a birthplace or a residence. The compiled service record carries enlistment age and residence — so the bundle contains the test and the payoff, and ordering the pension file alone would have been the risky purchase. WHEN IT ARRIVES: READ THE SERVICE RECORD FIRST. A thick pension file is precisely what will tempt a reader past the test. If the enlistment age and town do not fit a man born about 1831 in Warren County, this is another family's soldier and NOTHING in the file may be written into [p.1945]. No order number was issued; PayPal was to be requested separately. Tracked at projects/pending-reminders.md — check 2026-09-14, escalate 2026-09-28.
4. The fold3 image of the pension card, which usually carries the application and certificate numbers and the names of any widow or minor claimants.