Integration of result lines 8–13 from workflow wf_70287d01-764 (2026-07-21): the date-pin-sweep frontier agent + five "r2:" verification agents. Round policy applied: any "CONFIRMED" resting solely on Find A Grave contributor data is downgraded to PROBABLE here. Ground truth checked against 2026-07-20-wilson-tree-v0.md, 2026-07-21-lineage-research-partA/B.md, 2026-07-21-sibling-round-partC/D.md.
Date-pin sweep — 5 estimated-date targets (frontier, archive.org-only discipline)
Verdict: Two major date-pin wins (Thomas Angell + wife Alice via Austin p.4 page-image; Lucy Burt/James A. Olney via Oneida Co cemetery transcription), one primary-source birth pin (Richard Burt Sr), two honest dead ends (Goodspeed, Gibby).
Key facts:
- Austin 1887 p.4, visually verified: "'{THOMAS, b. 1618 / m. / d. 1694. {ALICE, b. / d. 1695.'" Will "dated 1685, May 23, proved 1694, Jun 18". Thomas came on the Lyon (left Bristol Dec 1630, arrived Boston Feb 5, 1631), then "He and four others came with Roger Williams" to Providence 1636 — CONFIRMED. Retires the weakest-class estimates (~1626–~1688 / ~1630–~1692).
- Alice's maiden name blank in Austin — consistent with holding the Ashton attribution at PLAUSIBLE.
- Oneida Co history (1878) p.595, Westernville Presbyterian churchyard: "James Olney, died Feb. 22, 1862, aged 78" / "Lucy, his wife, died June 4, 1876, aged 86" — PROBABLE (OCR; page-image recheck timed out). Matches partA's FaG/Wager dates exactly. Maiden name not stated in this record.
- Sylvester p.413 verbatim: Richard Burt Sr "born April 17, 1745," RI native, 1770 → Columbia Co, 1790 → Northumberland, eleven children (only John named) — CONFIRMED.
- Elizabeth Goodspeed: NO Goodspeed entry anywhere in Austin (two independent whole-book methods); Olney entry has no Nathan — couple outside Austin's scope. CONFIRMED-absent; estimates stand.
- Nantucket VR (to 1850): no Hiram/Sarah Mary Salisbury, zero Olney entries — scope predates 1856 birth, uninformative.
Parents found: Richard Burt Sr as Lucy's father — PROBABLE (surname/decade/geography; the passage doesn't enumerate Lucy).
Top leads: (1) 1815 Westernville marriage record (Lucy's maiden name); (2) Northumberland/Bacon Hill burial compilation for Richard's 1825 death; (3) remaining ~15 Arnold RI VR volumes for Hiram Salisbury.
Dead ends: Rowena Rowland Gibby — zero archive.org coverage of Swain Co NC (both Gibby genealogy books unrelated; Arthur's WNC history zero hits); needs FamilySearch/Ancestry-class tools. Quarantine: Sylvester's "Lucinda Harris m. Jay Olney of Oneida Co" = near-miss namesake trap.
r2: James Williams, merchant tailor of London — proposed father of Roger Williams
Verdict: CONFIRMED the will corpus verbatim (NEHGR 43:290–303, archive.org direct fetch) and EXTENDED it (found Alice's 1634 will + Margery Pate's 1617 will independently). The founder-identification itself stays PROBABLE — Waters's 1889 circumstantial argument, never document-anchored.
Key facts:
- Will dated 7 Sept 1620, proved 19 Nov 1621 (Commissary Court of London Vol. 24 fol. 50): names "my loving wife Alice," sons Sydrack, "my sons Roger and Robert Williams," daughter Catharine (m. Ralph Wightman), overseer "my brother in law Roger Pemberton" — CONFIRMED.
- Margery Pate's 1617 will independently names Roger, Robert, Katherine as children of James & Alice of "St. Pulkers" (St Sepulchre) — pre-1620 corroboration — CONFIRMED.
- Alice's 1634 will: "my son Roger Williams now beyond the seas," contingent to "his wife and to his daughter, if they survive" — matches Roger Williams's 1630 emigration + wife Mary + daughter Mary (b. Aug 1633; supports the tree's 1633-over-Austin's-1638 reading) — CONFIRMED text, PROBABLE identification.
- Ralph Wightman's 1628 will independently names "my brother Sidrach, my brother Roger and my brother Robert Williams" — third independent document, same sibling set — CONFIRMED.
- Rival Cornwall theory (Guild 1889: son of William Willyams of Roseworthy) refuted in the same exchange — Cornwall family had no brother Robert, but RW wrote of "Mine own Brother Mr. Robert Williams School-Master in Newport" — QUARANTINE the Cornwall line.
- James's own parents: explicitly unknown (Waters: an apprenticeship record would disclose them) — open wall.
Parents found: none (for James himself).
Top leads: (1) Merchant Taylors' Company apprenticeship records (James's father — Waters's own flagged unlock); (2) trace Robert Williams of Newport in RI records to upgrade the link toward CONFIRMED; (3) better scan of Austin's garbled Williams entry.
Dead ends: No burial record for James; no digitized St Sepulchre registers on archive.org.
r2: Alice (Pemberton?) Williams — proposed mother of Roger Williams
Verdict: CONFIRMED the same will corpus from the mother's side; held the Pemberton maiden name at inference (POSSIBLE→ arguably PROBABLE via the double will-language anchor); parents proposed at POSSIBLE. Also caught a serious tooling false-negative.
Key facts:
- Alice, widow, St Sepulchre-without-Newgate: will 1 Aug 1634, proved 26 Jan 1634 O.S. (early 1635 N.S.) → died latter half of 1634; asked burial at St Sepulchre's — CONFIRMED.
- Roger Pemberton of St Alban's, Esq. (will 1624, proved 5 Dec 1627): "Item I give and bequeath unto my cosen and Godsonne Roger Williams the some of ten pounds" — nephew-language; James had called him "brother in law" → Alice probably née Pemberton, of St Alban's, Herts — CONFIRMED texts, inference for the maiden name.
- St Albans register (begins 1558): "Alice Pemberton, daughter of Robert, 18 February 1564[/5]" baptism — plausible-but-unconfirmed match — POSSIBLE. St-Albans-pre-1600 note: the Pembertons are a St Alban's family — the same town the Olney/Ashton party left in 1635. Two independent lines of the tree now trace to one small Hertfordshire town.
- CONTAMINATION (three-way): an unrelated Roger Williams, parson of St Albans (d. by 1626), had a first wife Alice née ASHETON (m. 1583/4) — do not conflate with our London couple NOR with the tree's Marie Ashton St Albans family. Flagged by Waters himself.
- METHOD: WebFetch's summarizer falsely reported these names absent from the large NEHGR djvu.txt — caught only by direct curl + grep. Treat WebFetch negatives on multi-hundred-KB texts as unreliable.
Parents found: Robert Pemberton of St Alban's + Catherine (surname illegible) — both POSSIBLE (Visitation of Hertfordshire 1634 via Waters; two undisambiguated "Robert Pemerton" burials 1560/1578). Off-viz per policy.
Top leads: (1) Harleian Visitation of Hertfordshire 1634 directly (does it list a daughter Alice?); (2) later Waters Gleanings volumes for the promised Pemberton follow-ups; (3) St Sepulchre registers for the marriage + Alice's burial.
Dead ends: No document states "Pemberton" as her maiden name; no marriage record; burial unconfirmed.
r2: Almira Oakford (Green-Wood Sec 81 Lot 1540)
Verdict: CONFIRMED the Oakford trio's vitals via detailed FaG memorial pulls (→ PROBABLE per round policy); DOWNGRADED the Almira→Josephine parent link to POSSIBLE — no document names Josephine as their child; extended the Lambert side.
Key facts:
- Almira 1835–Jun 1870, John M. Oakford 1826–1898 (d. Brooklyn), Jane E. Oakford 1855–1880 (only documented child, died unmarried) — all Sec 81/Lot 1540 — FaG #58813546/#58813550/#58813549 — PROBABLE (FaG network; exact plot match to partC).
- Josephine "Josie" (Oakford) Lambert FaG #166493900: maiden name Oakford, m. John James Lambert 1874 — the marriage-year claim confirmed on its own terms.
- BUT her 1885 death record gives parents' BIRTHPLACES only (both "New York City") — not names. She's in the Lambert plot (Sec K/17634); the Oakford plot's one documented child is Jane E., not Josephine. Link = surname + era + plot-city only — POSSIBLE, matching partC's standing assessment.
- John James Lambert burial registry names parents "Jeremiah" Lambert + "Vleit Van Dyke" — independent corroboration of the tree's Jeremiah Lambert + Garetta Van Dyke Vliet — PROBABLE.
- Anita B. Lambert (1874–1882) + Walter Rittenhouse Lambert (1875–1943) confirmed as Josephine's children (already in tree).
- Wider Green-Wood Oakfords, quarantined: John D. Oakford (1833–1893, broker, 86 Broad St, m. Helen Maria Hastings Anthony) in Sec 113/16938 = a DIFFERENT man from John M. (corroborates partC's separation); William W. Oakford mason cluster Sec 62/5115. Elizabeth Oakford (d. 1895, Vernon Square Cem) ruled out.
Parents found: none for Almira (maiden name still unknown).
Top leads: (1) John M. Oakford's Kings Co probate 1898 — would name heirs, settling the Josephine question outright; (2) Jane E. Oakford's 1880 death record (names Almira's maiden name); (3) Almira's 1855 NY census hint (paywalled).
Dead ends: Ancestry/FamilySearch paywalls; Green-Wood 1901 index is unsearchable map plates.
r2: Richard Burt Sr (re-verification + family reconstruction)
Verdict: CONFIRMED the Sylvester p.413 quote by independent byte-range fetch (also resolving printed-page-413 vs archive-leaf-615); EXTENDED with wife + 6 of 11 children incl. two direct-line date-pins; effectively REDID the failed Sarah-Burt-Salisbury r2 stub from the sibling round. Note: this agent's dispatch premise ("presumed father of 5 date-pin targets") was garbled — only Lucy was ever claimed as his child; the agent itself correctly rejected the rest.
Key facts:
- Richard Burt Sr: b. Apr 17, 1745 (primary, Sylvester); d. 29 Oct 1825, Bacon Hill Cemetery, Northumberland — FaG 26437606 — PROBABLE; corroborates the tree's previously unverified "1825". FaG birthplace "Bristol Co MA" vs Sylvester "Rhode Island" — plausibly reconciled by the 1747 boundary transfer; unresolved.
- Wife: Lucy ("Airies" — spelling suspect, partA suggested Ayres/Eayres/Ayer) Burt, 1758–1826, Bacon Hill — PROBABLE (upgrade candidate from POSSIBLE: FaG network + the Sally-Harris cross-validation below).
- Cross-validation: FaG's Sarah "Sally" Burt Harris (1777–1811, m. Wm Harris Jr Apr 1 1798) matches Sylvester p.412-413 EXACTLY — genuine two-source check on the FaG tree.
- Lucy Burt Olney, FaG 17474054: 8 May 1790 – 4 Jun 1876, explicitly daughter of Richard + Lucy Airies Burt, m. James A. Olney — PROBABLE. Retires the viz estimate (~1794–~1856). Poetic detail: born 1790, the year the family moved to Northumberland.
- Sarah Burt Salisbury, FaG 27002467: 25 Jul 1811 – 30 May 1840, "Daughter of John and Sarah Burt," m. Hiram Salisbury 27 Feb 1834; children incl. Sarah Salisbury Olney (1835–1924) — pins Sarah Mary Salisbury's dates and gives Sarah Burt NEW parents: John Burt (1775–1871) + Sarah Olney (dau. of Stephen Olney of Saratoga, m. 31 Dec 1807) — PROBABLE. So Sarah Burt = Richard's granddaughter via John. Plus sibling-exchange corroboration: John's son Stephen O. Burt m. Rebecca Salisbury (Hiram's sister) 1836.
- CONFLICT logged, unresolved: Sylvester says John Burt had TWO children (unnamed eldest daughter + youngest son Stephen O., b. 1808); FaG lists FIVE incl. Sarah b. 1811 (later than the "youngest" son) — flagged, not merged.
- Contamination: a second in-book "Lucy Burt" (wife of James Burt, mother of Stephen J., p.511) — do not conflate with Lucy Burt Olney.
Parents found: none for Richard himself (Austin has no Burt entry — too late for its scope).
Top leads: (1) "History of Northumberland" + Saratoga NYGenWeb pages cited on FaG (may name the 5 undocumented children); (2) remaining ~17 unreviewed Bacon Hill Burt burials; (3) RI town/church records for Richard's origins.
Dead ends: Angell/Goodspeed/Gibby links to Burt: zero (as expected — garbled premise). FaG parent-name search endpoint is JS-only via proxy — use name+year queries instead.
r2: John [M./D.?] Oakford — NYC Death Cert #16353 pull attempt
Verdict: DOWNGRADED — cert #16353 unpulled (every open route walled: NYC portal 403/JS, FamilySearch/Ancestry login, GGG/ItalianGen JS forms, no death-index mirror on archive.org); the round-1 "John D. Oakford, broker, Oakford & Brewster" identity claim independently NOT corroborated. One headline claim is WRONG: "Find A Grave has zero Oakford memorials in NY" is a search-shell false negative — the sibling r2 agent pulled the actual memorials the same night (and partC counted 8 Oakford burials at Green-Wood). Treat that specific negative as void; the rest of the agent's work stands.
Key facts:
- Verbatim directory finds (via the reliable fulltext/inside.php API): zero "Oakford" in Trow's NYC 1863 + 1865 and in Wilson's Business Directory 1874; zero "Oakford & Brewster" anywhere on archive.org — the brokerage-firm claim has no searchable footprint — CONFIRMED negatives.
- "Oakford John, police, h 104 Perry" (Trow's 1876 p.1063) = a Manhattan policeman, same-name trap — QUARANTINE.
- William W. Oakford, mason, Brooklyn, continuous 1858–1871 (329 Bridge → 505 Fulton av → 393 Hudson av) — the only other era-and-area Oakford found; possible relative, no link established — POSSIBLE.
- Coverage gap: Manhattan directories 1866–1868 (the exact broker-claim years) have no OCR index on archive.org; 1867 missing entirely — claim neither confirmed nor refuted.
- No Oakford in Commercial & Financial Chronicle issues bracketing the 19 Jun 1898 death — soft negative.
- METHOD (round-critical): fetching a large
_djvu.txtand summarizing silently truncates to ~25-30 pages → false not-found on alphabetical directories. Use archive.org's search-inside API (fulltext/inside.php) instead. Earlier directory negatives produced the raw-djvu way should be treated as suspect.
Parents found: none (John's parentage: zero information anywhere — wall holds).
Top leads: (1) NYC Municipal Archives paid order for cert #16353 (founder hands or Ray browser session); (2) Green-Wood paid/in-person lot lookup for Sec 81/1540; (3) GGG/ItalianGen free NYC death indexes once a JS-capable browser session is available; (4) Manhattan Surrogate's Court 1898 probate index.
Dead ends: 1874 Lambert-Oakford marriage (archive.org NYC marriage mirror starts 1908); pre-1880 Kings Co deaths (state registration began 1880).