John L Cavin

AUDITED 2 Aug 2026 - PARENTS DETACHED AS IMPOSSIBLE; CORRECT PARENTS UNKNOWN. This person, recorded b. 1720, was linked as the son of [3029] John Cavins (b. 1700) and [3030] Mary Chism (b. 1710) - a mother aged TEN. The link (ChildTable RecID 2366 to F730) has been deleted; it already carried ProofFather = ProofMother = 3 (disputed). THE DEEPER PROBLEM, recorded so it is not re-created: [3030] is not a documented person. The upstream source (a CAVIN family sheet) gives the 1700-born man as literally '___ CAVIN' - no given name, no wife - and gives the wife only as 'MARY CAVINS', a MARRIED name with no maiden name, no birth date and no parents. The surname 'Chism', the 1710 birth and her parents are all downstream accretions. Diagnosis: a single garbled string, 'Mary Hamblin Prewitts Chism', was split into TWO women in TWO generations of this tree - [3022] Mary Prewitts (wife) and [3030] Mary Chism (mother). They share a death year AND a death place (1793, Leesburg). They are one entry, duplicated. [3029]'s given name 'John' should also be treated as invented. WHY NOTHING WAS SEARCHED FURTHER, and it is a finished claim rather than a failure: FamilySearch full-text returns 0 for both 'John Cavan' and 'John Cavin', but the control 'Loudoun County' returns 245 pages - and reading the DATES of those holdings shows Loudoun coverage runs 1762-1765 and then jumps to 1813. THERE IS NO LOUDOUN VOLUME IN THE FULL-TEXT CORPUS BETWEEN ABOUT 1766 AND 1812, which is exactly the window in which this man died (~1787). He is not absent; he is unsearchable there. NEXT DOCUMENT: Loudoun County, Va. will and deed books 1780-1795, browsed IMAGE BY IMAGE via the FamilySearch catalog - not by keyword search. A 1787 John Cavin will naming a wife is the one document that would decide whether anything can be re-attached. See ancestry/carpenter-cavins-prewitt-2026-08-02.md and db-write-queue.md.