The Welty Family

Tuscarawas marriages 1808–1844, both surname blocks read at the page: Weldy — Daniel and Ann Bollman 1826, George and Sarah Altman 1816, Jacob and Elizabeth Butt 1810, John and Hannah Allman 1812, John and Anne Kitch 1825, Philip and Sarah Overholt 1816; Welty from 1829 onward. No Michael under any rung, in either spelling, anywhere in the book. The register demonstrably reaches into his years — the 1810 Butt marriage is the thirty‑ninth return in Book 1.

Then the door that opened behind it. Tuscarawas was erected out of Muskingum in 1808, and Muskingum’s Book 1 covers 1804–1818: all twenty letter pages parsed, 927 marriages, 104 of them before the erection. Zero on the ladder as groom or as bride, with the alphabetical slot plainly occupied on both sides — two Wellers stand immediately before Wells and Welsh. Zero Harbaugh and zero Bierbauer as well. And the control did not fire. A planted cohort of families known to be on Tuscarawas ground returns one hit in 927, so this book may simply not reach the upper valley: use it to find, not to exclude. What is left is Pennsylvania before about 1807, where no county register existed to record him.

Since the chart’s own spelling is unattested on every ground it names, two surnames that are heavily attested there are recorded as candidate readings of a handwritten word. The first is his own township in his own window: Philip and Casper Bierbauer stand in the Dover Township returns every year from 1779 to 1783, and the printed spelling changes annually — Berbower, Beerbrower, Beerbower, Bierbauer, Bierbrouer, Bierbrauer, Bierbach. Seven forms, five years, one source. The second is Harbaugh, whose equation with Horbach is printed in the county’s own history, and who is dense on every ground he ever stood on: a Casper Harbough is enumerated on the same printed page of the 1790 Bedford census as “Welty, Micael”.

Neither is adopted, neither is merged, and no Catharine has been found in either family. They are candidate readings of a word, not candidate people. The base rate for this kind of slip is inside the same document: the chart already writes another husband’s name wrong elsewhere.

The 1816 duplicate of resident proprietors’ lands in Tuscarawas County, a fourteen‑page transcription of the whole county, carries the row “Weldy, Michael | 100 | 3 | 8 | 11 | NW qr … Abraham Feney … $2.25”, the last name being the column for the original proprietor of the tract. Abraham Feney is Abraham Forney — the man the 1884 county history says came from Somerset County about 1809, entered a farm on Crooked Run, and sold it to Michael Welty.

So the sale is on or before 1816, and the tract is named for the first time: the north‑west quarter of Section 11, Township 8, Range 3, one hundred acres, the same Section 11 that carries the estate farm. A Jacob Weldy holds 150 acres a page later. This is a transcription and not the film, and the man taxed in 1816 is presumptively the son — Michael Sr died in September 1815, and an estate can be taxed to an heir. No identity is assigned.

Michael Welty heads a household on leaf n178 of the Tuscarawas roll, in the run Frederick Reinhart – Abraham Overholt – John Collins – Joseph Shrock – Michael Welty – Andrew Critz – Peter Walters, with Abraham Welty on the same leaf, Jacob Welty on the next, and a John Witty heading a house in Sugar Creek. Abraham Overholt standing four names away is the family Philip of 1789 married into in 1816.

The row itself was then read. These schedules were copied onto ruled paper with no printed column headings, so the leading narrow cell cannot be told from the name column’s margin and two alignments one column apart are possible. What survives both alignments is the column that matters: the sixth male column, “of 45 and upwards”, falls on a dash either way. There is no man of forty‑five or over in that house. Whoever he is, he was not born before 1775 — an exclusion, not an identification, and the full bracket profile waits on reconciling the page against its own totals and the county recapitulation.

On the page facing the Barber entry: case A 731, estate of Richard T. Bunnell, administrator John Welty, 1844. A probate index keyed on the dead man can never surface a Welty who appears as administrator, security or appraiser. The docket now has a film, a manuscript index and a page‑to‑image formula fitted at three points, so the four ladder pages the printed index points at — including a Wilty Abraham Sr. standing in the same county’s probate record — are one session’s reading rather than a search.

Six index rows, three people, eighteen years, all on the bare stem: Sarah Welt in May, June and July 1768, the first of the three indexed Wielt; Joseph Welt and Susanna, intention 16 March 1769 and marriage that 11 May; Ann Welt, 15 June 1786. They surface under four separate rungs and are the same six rows under each, so the ladder is expanding rather than finding four families. Not one leaf was opened — that is the next step and it is one session’s work. The interest here is the confession rather than the county: the Edenkoben line is Reformed, the Manheim–Ohio family is Brethren and Mennonite, and a Barbara Welty stands on the 1770 Coventry Brethren roll in this same county. A Welt family inside a monthly meeting is a fourth setting and is unaccounted for. The Chester Wildy family of the 1740s is named as the firewall to test first, and is not asserted.

The same afternoon’s other result is bibliographic, and it is worth more than the search zero beside it: for a German Elizabeth born in Chester County in 1728 there is no register in which she could stand. Vincent Reformed’s baptisms begin in 1733; Brownback’s pre‑1832 register is lost rather than unread; and the Coventry Brethren baptise believers, so that congregation has no infant record class at all, by doctrine and not by loss. Only the Quaker registers reach the year, and their 1728 Chester cohort of that given name is twelve records, six surnames and exactly one birth — an Elizabeth Allen with no date, no place and no parents carried in the index — while none of the six surnames occurs anywhere in this project’s papers. The place field on that collection is blank on a large share of its rows, so the honest statement is a zero among rows the index placed in Chester. The Chester birthplace this file once carried for Michael was retired long ago and none of this reopens it.

From the 1907 county history, in the account of the pastor’s second ministry: “During this period the stone church, built in the time of Rev. Wirtz, was destroyed by fire on July 4, 1797, and all the records were burnt save one book.” The 1886 history narrates the same building sequence independently. The one book is the 1744 volume — the volume every negative this file has recorded at that congregation was measured against.

This is print rather than manuscript, and a 1907 statement about a 1797 event is secondary by construction. What it changes is not the silence but its meaning: not “he was not baptised at this church” but “he is not in the one book of this church that came through a fire.” Where the leaves of 1746 to 1761 went is untested. They may have been in a second book that burned; they may be in the pastor’s own private pastoral record of 1744 to 1769, which survives, is reachable, and has never been read. That is now the most valuable unopened object on this ground.

Image 10 recto, in the run that opens 2 August 1744: an infant “Cath: Elisabeth”, parents “Michel Wel…” and Margaretha Barbara, witnesses Johannes Wolff and Catharina. The surname was first read as Welsch and that reading is withdrawn here rather than left in a footnote: there is no s‑c‑h on the page.

What it is not, each excluded against a control in the same hand: it is not a ‑tz name, because the clerk’s own ‑tz stands eleven lines above in a witness’s name and is a tall crossed t joined to a looped z, and this word has no crossbar; it is not the Weldy of the 1754 list three leaves later, which has an unmistakable dotted descender; and it is not the Wolff standing in its own column. The stem and a medial looped d are verified at the image and the terminal letter is not settled, so the reading is probable and not proven — one word, on one leaf, that a better raster would decide. If it holds, the register’s “no Welty before 1763” was never a fact about arrival; it was an artefact of a transcription that never covered these leaves.

The enlistment tables in the fifth series are printed sideways on a portrait page, so the published text layer renders them as noise: over the band worked here it turns the column head “Remarks” into a string of characters running backwards, and not one leaf returns a legible column head. Every machine sweep that counted this band as searched was reading rubbish.

Sixty‑eight leaves were pulled at native resolution, rotated three ways and read again; thirty‑five came back by rotation and thirty‑three read best unrotated, being the horizontal muster rolls interleaved with them. Over the forty‑six leaves recovered legibly, no name on the ladder appears in the place‑of‑birth column. Twenty‑two did not recover and are listed individually, so the re‑run is a list rather than a search. This is machine‑read and not eyes‑on, and the zero covers only what came back readable.

The entry below quoted the Dover class list of 1781 as “Philip Welte 666 / Philip Welte Jun 016 / Michel Welte 082…”. The first figure is 066. At magnification it carries the same leading zero every other entry on that leaf uses, and a 666 standing beside neighbours of sixteen and eighty‑two never sat right. The entry has been corrected in place, and so has every working file that repeated it.

It matters more than one digit usually does. At 666 the elder Philip read as a substantial holder standing over the rest of the column; at 066 he sits in the ordinary run, and the largest figure among the five Weltes belongs to Michel, at 082. Nothing about the junior changes — he is still the smallest of the five, still in the young‑man band. What changes is the shape of the group.

And the correction paid for itself. The same bundle holds a second copy of that column, written out for the collector in a different hand. It reads Philip at 64 and Jacob at 42, against the leaf’s 066 and 041 — agreement within a unit or two, on a copy whose figures run a shade low throughout. Against 666 they agreed with nothing at all. A misread number is not only wrong; it hides the corroboration that would have caught it.

A 1936 surname index to the standard Frederick County history is free and open. Read at native raster, one line of it prints: “Welty, 324, 460, 546, 560, 631, 723, 733, 1083, 1118, 1358, 1405, 1445, 1446”, with “Weltner, 8, 86, 479” on the line above.

That same page’s machine text deletes the word Welty entirely and reattaches all thirteen page numbers to Weltner. Six controls on the same page fired correctly. Nothing in the search would have shown it; only the image caught it. This is the third distinct way this week that a page has read as empty while carrying the name — after the sideways tables and the ruling line at Coventry — and it is the worst of the three, because an index is exactly the instrument a researcher trusts to be complete.

Ten of the thirteen pages have since been read in the free 1910 edition of the history itself, and they are the already‑known nineteenth‑century Maryland families. The one entry that looked eighteenth‑century is a mirage: the text layer reads “Trucks, George Welty, John Wimmer” in the printed 1790 census, and the government’s own edition shows the true run is “Trucks, Geo. / Welty, John / Wimmer, Abraham” — a given name shifted one surname to the right across three entries. There is no George Welty there, which confirms rather than disturbs the reading this file already had. Three of the thirteen pages remain unread.

Every register entry this file holds for the Württemberg village came through one subscription; the other one — the one used all summer for the Palatinate — holds the same parish as twenty‑four volumes of its own, and not one had ever been opened here. They are now enumerated with their identifiers. The volume opened this evening is not what its catalogue title suggests: its first leaf reads “Communicanten Buch”, signed by the pastor and dated 23 June 1721, with the parish’s cost of paper and binding noted underneath. Its first roll carries the two men of the godparent family that locks this line into that village. It is a presence instrument covering years the event registers do not — and it ends in 1748, six years before the crossing, so it cannot name the son this file is looking for. The long hunt for a soul‑register at that village also has an answer: there is no volume of that name there. The neighbouring parishes have them.

One caution belongs in public, because anyone working these registers will meet it. The same subscription offers on‑demand machine transcription. A credit was spent on a two‑column list of about two hundred names, and it returned fluent, grammatical German prose that is not on the page — whole invented sentences, with real surnames from the leaf scattered through them. A typed film target card on an adjacent image transcribed flawlessly. On a page of running prose it may well be useful; on a list, it produces text that reads like a transcription and is not one, and a zero taken from it is worth nothing.

No record was opened for this one. Three readings already in this file, each verified at its own image and never before set side by side, were simply put in a row. On the Dover return of 8 December 1774 a Jacob Welty is rated as a single freeman — which puts him at twenty‑one by then, born about 1753 or earlier. The Michael of Dover Township is measurably not in that block, and the block is known to rate single sons against their listed fathers, because it does so for another family on the same leaf. Michael’s earliest appearance anywhere is March 1778. And the 1800 census in the west puts Jacob’s household head in the forty‑five‑and‑upward bracket, born before 1755.

So on the earlier of the two instruments the Welty freeman is the older man, and it is Jacob. Several rows in this file still frame him as the younger brother on the strength of a birth year that came from a 1946 magazine query rather than from a record, and that year has now been contradicted twice. It also re‑reads the tax geometry in the plainer direction: Jacob taking eighty and then a hundred acres while Michael holds forty‑four as a tenant and is landless by 1786 looks like ordinary elder‑son succession rather than an oddity. The caution stated with it: absence from a freemen block means not rateable in this township, which under‑age satisfies but so do not resident and bound elsewhere. Nothing here touches whose sons they were.

Five Weltes stand consecutively in one column of the Dover class list of 1781: “Philip Welte 066 / Philip Welte Jun 016 / Michel Welte 082 / Jacob Welte 041 / George Welte 064”. The suffix is not an artefact of the order — this clerk is known to mark senior and junior when two men of one name stand in one township, and he leaves the others unmarked. At the smallest valuation of the five, the junior is in the young‑man band.

He has no roster row, no identifier and no place on any candidate list on any tab. The question was asked in terms in July — whether the second Philip in that township is a son or a brother — and it was never run. The unexciting answer is the leading one: a Philip Jacob born 1759 who married at York in June 1780 and whose father’s household was in Dover — that father stands fifth in this same column — would be a newly married twenty‑two‑year‑old in his father’s township, and he belongs to a branch already firewalled from this one. There are four men of that name already kept carefully apart in these files. A fifth will not be created on a valuation and a suffix.

Four entries in the same burial volume, three householders, read the same evening. 5 October 1782: “Joh: Conrad, des hiesigen Bürgers und Metzger‑Meisters, Johann Jacob Wälder, ehel. Söhnlein von 9 Monathen” — a master butcher’s infant son. 16 September 1783: “M. Margaretha, H. Philipp Jacob Wälder, Gerichts‑Verwandten und Schaffneris dahier, Ehefrau, in dem 39ten Jahr”, so his wife was born about 1744. October 1784: a son Johann Philipp of the same man, dead at seventeen. 8 March 1785: an infant daughter of a Gottfried Wälder. A Gerichtsverwandter sits on the village court, and with the honorific and a second office this is an officeholding household rather than a cottager’s.

The surname took two plates and the scribe’s own hand as the control. The viewer cannot resolve the word at all, so native leaves were bought for the first two entries. On the first, the second letter carries a superscript of two short parallel strokes — and the same scribe on the same leaf writes his u‑umlaut as two round dots and his o‑umlaut as a broad connected zigzag, while an unambiguous a‑umlaut elsewhere on the leaf takes the two strokes exactly. Thirteen months later the same hand writes the name again, and the fourth letter is a plainly looped d with no crossbar.

No connection to the Edenkoben line is claimed, nothing is merged, and nobody is added to the tree. What makes it worth publishing is that this is the first household of that name located in the home parish itself, in the right century, carrying the compound given name this file has to watch. And a caution belongs beside it: six leaves earlier in the same volume stands Philipp Jacob Walter, “der alte Eisenmüller”, buried in November 1776 in his sixty‑second year — same parish, same compound given name, a milling trade, and a different word on the page: a bare round a with nothing above it and a plainly crossed t. The two are distinguishable at the letter and must not be run together. What will decide the Wälder family is not more burials but the marriage register: a wife dead in her thirty‑ninth year in 1783 and a son born about 1767 put that marriage around 1764 to 1766, and it would name both fathers.

Every machine sweep this project has run has ranked its results by nearness to the word Michael. That is a filter shaped like the answer, and it cannot see a woman, a widow, or a son named as somebody’s child. A different triage was built — three axes, none of them a given name: is a feminine word or the word widow nearby, is a kinship formula nearby, and does the surrounding text have the shape of a list. All seven controls fired, and the three quarry‑shaped records already known sorted straight to the top, where the old method scores all three at nothing.

It paid out once, in a printed transcription of the surviving Lancaster County tax lists: “… henry musselman / herman long / john cauffman / hans newcomer / widow welty / joseph cochenour / frederick hoffman …” — Manor Township, 1751, a taxable householder in her own right, and the only name on the ladder in the whole surviving 1751 corpus for that county. The volume’s own printed index carries her too, alphabetised between Weller and Wendling, which is a check no machine misreading passes.

The grade is a text layer over a printed transcription and the manuscript behind it has not been found; the word widow may be the collector’s convenience rather than a probate fact, and about forty such entries stand in the volume. What the pass also surfaced is bigger than the entry: the editor’s own preface says the county’s tax records for roughly forty years in the middle of the century have never been published — a whole body of assessments outside every printed corpus swept here to date.

The next step everyone agreed on was the Philadelphia Orphans’ Court, on the chance that a guardianship names the blacksmith’s children. The films this file has been citing for it cover 1763 to 1775, which is correct for the row that first named them and eight years too late for a 1755 death — and a note written ninety minutes earlier the same afternoon had quoted them for exactly that. Two films that do cover 1755 were named for the first time, one of them a volume of estate papers for 1749–1757 never mentioned here before. Both were checked and both are behind an access padlock; no name search reaches them, and a control confirms the search itself is scoped away from those films rather than merely empty.

The remarried widow’s new surname was then measured rather than searched for. In five hundred and eighteen million characters of printed corridor literature it occurs exactly once, and the once is this project’s own marriage entry. It is absent from the full‑text corpus of Pennsylvania records before 1790 in thirteen spellings. Whatever became of that household, it left no printed trace — so the question of who raised the blacksmith’s daughters has to be asked of manuscripts, not of books.

The cheapest remaining door was costed instead of merely named. The county’s Register of Wills states in its own words that it is “the only repository in possession of original Philadelphia Administration files”; the request is a letter with a twenty‑dollar fee and a return envelope, and this file already holds everything the form asks for — the name, the date, the file number and the book and folio. Before spending it there is a free test worth twenty minutes: an earlier administration book from the same office is in print, and reading a dozen of its entries will show whether entries of this class name a widow at all.

The 1770 census had a companion volume for Maryland, Virginia and the Carolinas, written in 1772, and it reaches the exact ground on which the Manheim‑to‑Frederick group of these families is documented. It names and dates five Frederick County congregations — Conococheague, begun 1743; Monocacy, 1749; Middlecreek, 1747; Antietam, 1750; Pipe Creek, 1751 — with their ministers, and then simply stops. There are no member rolls in it at all. That is a property of the genre, verified at two independent witnesses, and the compiler says why in his own voice twenty years later, of a different congregation: “I am sorry I could not come at more of this good man’s history (for these people keep no records).” A congregational historian said the same thing of Coventry in 1898, and a Maryland historian said it again in 1936. Three independent witnesses across a hundred and sixty years agree that the early books of this denomination do not survive.

What the volume does give is a cohort walking the corridor this file has been arguing about. The minister at Middlecreek in Maryland in 1772 is the same man the Pennsylvania volume records leaving Codorus in York County — “this Tanner left them to go to Monocacy” — the Pipe Creek minister shares a surname with a woman standing a few names from Barbara Welty on the Coventry roll, and the man named superintendent of all the Maryland congregations later moves them south to the Pee Dee in Carolina. Nobody on the ladder is named in any of it. It is the road, described by someone who walked it in 1772.

The denomination’s western Pennsylvania history was swept alongside it — a million characters, forty‑one spellings, and a clean zero — over a book that opens: “First Settlers, 1762. Among the first settlers that came from the East across the Alleghany Mountains into Somerset County, about 1762 or 1763, was a colony of Brethren, who settled in the northern part of what was then called The Glades.” That is precisely the ground the Milford man stands on, and the narrative of it carries no name on the ladder. It is a narrative and not a membership register, which is the reopener: the congregations’ own rolls, if any survive, have not been looked for.

The Milford list did not arrive alone. The same printed volume holds twelve complete Bedford County township militia lists taken in January and February 1789 — effectively a name‑by‑name canvass of the county fourteen months before the 1790 enumeration, and a body of evidence this project had assumed did not exist, because a standing note said the printed Bedford series stopped in 1784. It does not.

The sweep that found them ran across eight hundred and twenty‑seven volumes tied to the counties this family moves through, four hundred and seven million characters, on the full ladder of spellings. Two positive controls fired first. Across the whole of it there is one name on the ladder in the Bedford section, and it is the Milford man. Two others surfaced elsewhere and belong to nobody yet: a Henry Welty on a Franklin County company of October 1786, and a John Welty on a York company returned in April 1786.

It also corrected an old negative of this file’s own. A search made in July for a particular York cohort — Bierbower, Messerly, Ramsey, Feyser — had concluded they were absent, having looked in the wrong volume of the right series. They are there, on a Third Battalion return of 18 November 1788, and “Jacob Welte” stands in the third class among them.

The full‑text search of an entire open book archive was run on the ladder for the first time, without county or date restriction. After filtering to pre‑1930 imprints, thirty‑one volumes survive and all but two were already in this file. The two are both worth having.

The first prints the Milford Township tax list of 1796 in full — “… Jacob Wimer, John Wagoner, James Wilson, Michael Wagoner, Thomas Wilson, Abraham Whipky, Jacob Weldy, Michael Weldy, George Wimer …”, closing with the collector’s name and the sum collected. That list had been held here only from an online transcription; it now has an independent printed witness from 1884. The same book carries a second, earlier list of the district’s settlers on which Jacob stands without Michael, and that asymmetry is not explained.

The second is a warning rather than a find. An 1894 history of the National Road lists the wagoners by their county of origin, and under “The following were of Washington county, Maryland” stands a Michael Welty. He is not the Michael of these pages, whose death in Ohio in 1815 predates the traffic described; he is filed as a decoy so that the next researcher who meets him does not have to work it out again. The Somerset County list on the same page carries no name on the ladder.

The printed abstract of the Land Office’s Old Rights for Philadelphia County R–Y, all of Bucks and all of Chester is set as landscape tables rotated ninety degrees, which is why no text search has ever touched it. All hundred and eighty leaves were pulled, turned, and read out. The result is zero on the ladder in all three counties — and because a zero of that shape deserves the working, the W‑runs of each county were transcribed in full instead of merely swept: Whitpain, Whartnaby, Watkins, Wood, Woodmanne, Willcox, Worley on the Chester side; Wate, Walton, Welch, Waddy, Watts on the Philadelphia; Wally, Waly, Waddy, Waite, Weight on the Bucks.

One near‑form was chased to the image at three times magnification and refused: “Reynolds & Rawson Wolly”, a joint entry of the twelfth month 1685–6, one letter from a spelling this family is genuinely recorded under, and a Chester man of 1685 who can be nobody’s ancestor here. The more useful measurement is the section’s own date profile: of two and a half thousand year‑tokens across it, ninety‑four in every hundred fall between 1680 and 1719, and fewer than sixty fall after 1730. The series closes before this family reaches Pennsylvania at all. What would reopen it is the Philadelphia A–Q run, which is not in this volume, and the true terminal date of the manuscript series, which has been taken from the print rather than from the archive’s own description.

A per‑leaf quality screen was built over the whole printed archive series and it turns a suspicion into a number: two thousand eight hundred and thirty‑one leaves, in thirty‑three of eighty‑three volumes, have never been readable by any machine — sideways tables whose text layer is gibberish. Thirty‑one volumes are clean. Every flagged leaf was then pulled, rotated, re‑read and swept on the full ladder, and the sweep is silent: no new name anywhere in it. Four candidates were chased to the image and died there — two are Kelly, one is Welter, one is Weltig.

The screen also caught itself. Two known records of this family sit on pages whose published text contains not one occurrence of the name, and the same two printed pages score above the alarm threshold on one digitisation and below it on another, because one scan is coarser than the other. The threshold is a property of the photograph, not of the book — which means some of this file’s machine zeros are softer than they look, and re‑running them against a better scan is the reopener.

The cheap answer to all of it turned up the same morning, twice. The printed general indexes to the fifth and sixth series were located and read. They were compiled by hand from the printed pages, so a sideways table is no obstacle to them, and between them they print the complete footprint of the ladder across two whole series — forty‑seven entries, name and volume and page. Most are already in this file. Four have never been logged at all: a Witty, Philip on the page adjoining the depreciation‑pay entry that carries Michael; a Welty, Danl.; a third spelling and page for Anthony; and a Wildey, Philip. Those are pointers, not records. Nobody has opened any of the four.

A search handle inside a genealogical digital library was worked out and turned into a tool: it returns the full text of a page rather than a snippet, which meant a book that had previously been recovered thirty‑five pages at a time by hand came down complete — a hundred and seventy images of a hundred and seventy, in about four minutes.

The book is the 1941 typescript of the Falckner Swamp Reformed register, and it carries no name on the ladder anywhere in a quarter of a million characters, on thirty spellings, with no near‑miss. But its contents page is the finding. It reads “IV. Confirmations. (Not Copied) 1439 listed.” The transcriber counted the confirmands in the church book and declined to copy them — and a whole‑book census confirms it, the word appearing on three of a hundred and seventy pages, all of them front matter. So the fourteen hundred confirmations in the subscription index opened yesterday cannot have been keyed from this typescript, and for that one role the index is a second, independent witness to the church book after all. The correction made yesterday afternoon was right about the marriages and burials and wrong about the confirmations.

The same handle was turned on the Dover register the following half‑hour and stopped part‑way through by choice, at a hundred and sixty‑two of about a hundred and seventy‑four images. It is enough to answer a question standing since 9 July: this book carries no burials at all — baptismal language on a hundred and forty pages, confirmation matter on ten, communicant matter on two early leaves, and burial language on none. Its eight ladder occurrences are all entries already in this file, including the union‑church covenant of 30 May 1757 with George Weldy first on it.

The entry published here that morning ended on a new handle: “Philip Weldy (suspended)” on the 1823 communicant leaf, and the observation that a suspension is an act of the consistory, minuted with a cause and often a restoration — and that the congregation’s consistory book sat on the same film, located and never read. It was read the same afternoon.

The book has a thirty‑one year hole in exactly the wrong place. Its last dated minute is “The Bord of Elders, Council and Church wardens met at the Said Reformed Church on the 11th of January 1818 and Proceeded to Business”, signed by a president, a secretary and a treasurer; the next dated minute in the volume resumes in April 1849. The suspension of 1823 cannot be minuted here, because 1823 is not here.

Reading the rest of it whole was still worth the hour, and it answers the question the suspension raised. He is not restored at this charge. The communicant roll of 9 May 1824 runs about a hundred and seventy‑five names with no Weldy on it, and the roll of 1 April 1825 carries none either — and the same 1824 leaf shows the register does annotate discipline, printing two members as “(since excluded)”. The volume also gave up something nobody knew existed: the congregation’s charter of 16 October 1808, about a hundred and fifteen signatories read one at a time, with no name on the ladder among them. What would move any of this is the missing minute book for 1818–1849, which is a letter to a seminary archive rather than a search.

A lead had been carrying a promise for weeks: a bound run of 1777 York County township assessments, which would have given a Dover freemen block a year earlier than any yet read. The volume was opened at its title page and it is something else. The warrants inside run to a fixed formula — “York County ss. To [collector], Collector of [township] Township … Given under our Hands and Seals this 10th Day of May 1777” — charging a flat three pounds ten shillings a head. It is a bound volume of non‑associator fine rolls, not assessments, and a fine roll cannot deliver a freemen block, an acreage or a household.

All sixty‑two leaves were then opened rather than sampled, which took an hour and settled the shape of it. Twenty‑three townships are named. Four are not: Dover, Manchester, Codorus and Huntington. The temptation is to read that as four townships making no return — and the volume’s own closing leaf forbids it, in the clerk’s words: “This Dublicate of Fine is now Collecting, need no Transcribing.” This is a transcript book, and a transcript book skips what was already in a collector’s hands. Its gaps are evidence about the copying, not about the townships. Its own docket numbering runs to twenty‑four against twenty‑three named townships, so one numbered packet is unaccounted for, and that is the last place inside this volume a Dover roll could sit.

Mapping the whole five‑hundred‑image film off its own archival target cards then produced a caution worth more than the map. The cards are wrong in both directions. The card opening the 1778 volume lists six townships and omits Dover, which is in that volume twice; it omits Shrewsbury and Newberry, which are also there; and a leaf two hundred images past it is headed with a year three years later than the card claims. No card on this film can be used to rule anything out. The map did locate one thing worth the errand: a volume at images 141–216 whose card reads “Tax Records — Hellam, Newberry, Dover, Warrington, Monaghan — 1774–1775”, which is where the December 1774 Dover freemen list should be sitting.

A subscription index of the Falckner Swamp Reformed register, never used here before, was opened and worked facet by facet. It holds nearly two and a half thousand sponsor rows and one thousand three hundred and eighty‑seven confirmands, two hundred and fourteen of them inside the corridor window this project cares about. Ten wildcard sweeps across every spelling family returned not one name on the ladder in any role — not as child, parent, sponsor, bride, groom or confirmand. Five external controls were planted before the sweeps and all five fired, including a count control that matched a thin year an earlier reader had counted by eye.

The row that came out of it claimed something more than the zero: that the index was an independent keying of the church books, and that because its marriage and burial series show the same vacancies as the 1941 typescript, two independent copies agree the church book itself is blank in those years. Ninety minutes later the same session took that back. A catalogue lookup on the film number behind the index returns the identical title, compiler and volume span as the typescript — so the index is a keying of the copy, not of the original, and two witnesses collapse into one. The zero stands; the argument built on it does not.

The whole sponsor column of the printed New Hanover baptism table — both halves of it now, the whole table — was measured rather than searched, and the measurement is a fact about every sponsor‑anchored search this project has ever run. Of roughly two and a half thousand sponsor rows, about forty‑four in every hundred give no surname at all — some printing the couple simply as “X and wife”, more giving a bare given name and nothing else — and only about eleven in a hundred name a woman with a surname of her own. A search that keys on surnames is structurally blind to nearly half this column, and the blindness is in the book, not in the method.

That changes how one line in it reads. The compositor sets “Joh. Martin Miller and Anna Barbara” and then, on its own line, “Welte.” Because naming a woman in full is the minority form and “and wife” is the ordinary one, the likelier reading is that she is named in her own right rather than as the miller’s wife. It is a reading from typographic convention, not a statement in the record, and it is filed at that grade. Every line‑break in the whole table was then enumerated — two hundred and eighty‑seven of them, hyphenated and unhyphenated — and exactly one ever concealed a name on the ladder: this one. What would settle it is the original church books behind the print, where the women given here as “and wife” may be named.

The entry published here on 3 August reported a typescript burial section headed “DEATHS or BURIALS — (Page 21 of oldest register)” and named its reopener: page 21 of the oldest register, and the leaves around it, which the transcript proves exist and which nobody has looked at. The volume has now been walked end to end, and that sentence was wrong in its second half.

The filmed 1744 book is three sections and no more: baptisms at images 3–136, one continuous run — German script and numbered entries at 1762, a four‑column page at 1802, a full‑page display heading reading “1844 Baptisms” at image 127; marriages at 137–159, opening “AD 1755” and closing in English in 1853 with “Look in the New Church Book for the later Names”; and deaths at 160–162, beginning “Died — The 11th of June 1844”. There is no death‑format page anywhere between. The band that had been named as the last place one could hide was opened at its first leaf, and its first leaf is already 1802.

So the transcriber of the 1930s had a book this film does not carry. The negative is not “the page is not there” but “the book is not here” — and the honest consequence is that the only surviving route to an eighteenth‑century burial page for this charge runs through the original registers in institutional custody, at the Evangelical & Reformed Historical Society at Lancaster Theological Seminary or with the congregation itself. That is a letter or a telephone call, not a search.

A marriage recorded in this charge on 18 October 1827 — Henry Welty and Catharina Coleman — has been in the file without issue. The register carries their child: Elisabeth, born 10 May 1828, baptised 20 July 1828, parents Henry Welty & Catharina, in a four‑column German register ruled Kind geb. | Eltern | Zeugen.

The sponsor column is the point, and it is an answer rather than a silence. It reads “Parentes” — the parents stood for their own child. Naming in these families runs by godparent, which is the discriminator this project reaches for when it wants to tie one household to another; for this baby there is no godparent to reach for. It is worth recording that the register was asked and said so plainly.

One letterform flag, not a re‑grading: at high magnification the surname on that leaf carries two crossed ascenders between Wel‑ and ‑y, so it reads Weltty or Weldty rather than a plain Welty, while the index prints Welty twice. And a household this file did not have: a Philip Welti and Cath baptise a child born 14 November 1819 at the same charge. That leaf has not been opened.

The Land Office warrant registers are a county‑by‑county series and this project has worked several of them — York, Lancaster, Northampton, with Dauphin named. Chester had never been opened, and Chester is the county the family pedigree names for the birth this project is trying to date. It was not a decision anyone took; it was a gap.

It is now read. The whole W section — five interface pages, 154 numbered warrants running from 1733 to a warrant of 20 April 1903 — eyes‑on at 200 dpi, the crop bands deliberately overlapped so that no name is lost at a seam, because a name lost at a seam is exactly how this project has manufactured a false negative before. Zero on every rung of the ladder, and no bare stem of any kind. The section boundary was measured rather than taken from the interface label: the page before it is the V run, the page after opens the Y run, and one blank ruled leaf sits between.

Three near‑misses were magnified and ruled out rather than eyeballed — “Wertz, Jacob”, whose third letter is a short hooked r and not a tall looped l, which makes it the single line in the section most likely to be misfiled later; “Wholey, Timothy J.”, a round o; and “Welch, Sarah”, a clear ch ligature. Two English‑neighbourhood readings are firewalled so nobody re‑chases them: Wield, John, an ield order rather than eld, and two separate Robert Witt‑ entries on different leaves and different dates.

The county’s tax lists were already a whole‑corpus zero. Land now says what tax said. The reopener is structural rather than hopeful: a man who acquired by purchase or by patent is invisible to a warrant register, and two free series that would catch him — the patent index and the Chester old rights index — are the next places to look.

A reconciliation this project relied on had shown a church‑records collection to be complete on the ladder. It ran on two rungs, and both of them carry an E. The A‑vowel rung had never been read. It has now been read whole, one hundred and forty‑three records, partitioned by date and reconciling exactly — and it held a household.

At Moorestown, Moore Township, Northampton County: Philipp Waldy and Emma Margaretha, with Johann Nikolaus Waldy baptised 23 June 1782 and Emma Caterina Waldy baptised 11 October 1784; then Johannes Waldy confirmed in 1796 and Jacob Waldy in 1798, same congregation. The town lists make it a community rather than a couple: Philip, Georg and Johannes Waldy as three adult males in the residence list of 1791, and in 1792 a Joh Philip, a Joh Georg and a Philip, which reads as a father and a son sharing a name.

Moorestown is in Moore Township, and a Philip of that township has existed in this file only in land paper — warrants of 1774 and 1787, surveys of 1775 and 1786, a patent as late as 1823, and three tax years in the 1780s. He now has a wife, four children and two more adult men of the surname beside him. He is not merged with any other man of the name, and the firewalls between them stand. The same village turns up a household on a third rung baptising in the 1820s.

The lesson costs nothing to state and it is the reason the record existed to be found: the source did not change. The reach did.

Four German genealogy domains that this project would ordinarily query from a script were measured directly. All four return HTTP 200 — success — with a 4.4 kilobyte proof‑of‑work page reading “making sure you are not a bot”, and zero result rows, for any query. A script sees a successful request and an empty answer. It is a perfect forgery of a negative, and it is worth publishing for any researcher scripting against those hosts. A real browser passes it invisibly, with no puzzle and no human step.

The first thing checked was whether anything in this file rests on those hosts. Nothing does — the whole workbook returns no mention of any of the four. The next thing was a positive control before the sweep rather than after it: the surname Welde unqualified returns 223 entries, so the search works; Welde with Edenkoben returns a true zero, now measured instead of assumed.

The bycatch is the find. Maria Elisabetha Welde, born 1743 at Altdorf in the Pfalz, died 1817 at Böbingen, and Salome Welde, born 1750 at Altdorf — both inside the ring of parishes around Edenkoben this project has been working, neither with parents recorded, and both at submitted‑pedigree grade, which is the lowest grade this file recognises. They are a lead, not a record. But the ring was swept once already and returned nothing, and the reason is now visible: that sweep read Böbingen’s marriages, and these are births. The scope was right and the section was wrong. Altdorf itself keeps nothing before 1759, as this project reported in July — so whoever compiled these two read them somewhere else, most probably in the Böbingen book, which is one leaf each to check.

The entry published earlier today ended on a qualification: the Pennsylvania supply act for 1778 had not been read, and if its freeman clause were conjunctive — twenty‑one and six months clear of service — the Dover return would yield a hard floor of born on or before March 1757. The clause was read. It is conjunctive, word for word as expected. And it does not govern this list.

Two independent reasons, either sufficient. The act was passed 27 March 1778, and the warrant this return sits under is dated 26 March 1778 — the day before. And the decisive one, because it is on the page rather than in a book: that act charges a single freeman £3, and every freeman on the Dover list is charged thirty shillings. Thirty shillings is the rate set by the supplement of 13 October 1777, and that clause is disjunctive: “every single freeman, whether residing with his parents or elsewhere, who at the time of the assessment shall be of the age of twenty‑one years or hath been out of his apprenticeship or servitude for the space of six months … shall pay the sum of thirty shillings.”

So there is no floor. A man could be rated at under twenty‑one provided he was six months out of service, and the birth window stands unchanged at roughly December 1753 to 1755 — now for a read reason rather than an assumed one. One inference published this morning is withdrawn with it: under a disjunctive clause, being rated in March 1778 does not show a man was clear of an indenture. That is a withdrawal, not a claim that he was a servant; nothing here measures what the assessors actually did.

What the day gained is worth more than the floor would have been. A flat sum on a colonial list is a citation, not a number. Both flat rates in this one packet name the law that produced them, and neither had been looked up in five weeks of working the film. An undated list can often be dated and classified by its own rate — and a rate matching no act is itself a warning that a document has been mis‑classed.

The 26 March warrant and the packet wrapper were pulled at native resolution and read whole. The date is now read at the page rather than inferred — “the 26 Day of March Annoque Domini 1778”, in the commissioners’ own hand — which matters because the whole argument above turns on this warrant preceding the 1778 act by one day.

The measured negative comes first, because it is what was asked. The warrant recites no act. There is no “by virtue of an act of Assembly passed the ___ day of ___” clause anywhere in it, and the packet’s exterior docket is equally silent. That errand is closed as answered rather than as unrun — and it is normal rather than exceptional: these warrants are administrative forms, and the act lives in the commissioners’ instructions, not on the collector’s copy.

Which makes the arithmetic the primary tool rather than the fallback. The assessment columns are headed Rate and Tax’d, valuation and charge side by side, and the ratio is exactly five shillings in the pound, without exception, across sixteen consecutive entries: rate 4:10 charged 1‑2‑6, rate 2 charged 0‑10‑0, rate 9:10 charged 2‑7‑6, and so on down the column. Five shillings in the pound is a quarter of assessed valuation, which is a war‑supply rate and cannot be a county levy, since county levies are laid in pence in the pound. An undated, headless leaf whose freemen block has been lost can still be dated from any single householder’s two figures.

One thing on the page, recorded because it is on the page and not because anything is read into it: both Dover warrants and the wrapper were first made out to one man and readdressed to another, the strike visible in all three places — so the collector’s name is a late hand throughout this packet, and the man first appointed stands on the fine roll himself, struck.

The entry published here earlier today reported that no child of the Gültstein household is named for a godparent, and drew from it that a naming custom proved in one branch is not a property of the surname. A fourth and a fifth baptism have since been read at native resolution, and the fifth overturns the first half of that. Johann Martin, baptised 30 January 1748, son of Hans Jerg Welthe — written Baur, a farmer, an occupation this log has never had for him — and his godfather is Hans Martin Nueber. The child is named for him.

So the custom is not missing at Gültstein. It is intermittent — one baptism in four in this household, against five for five in the Edenkoben line. What survives is the practical form of the rule, and it is worth stating plainly: a godparent match may be treated as a weak positive, and a mismatch never as an exclusion. The fourth child, Catharina Elisabetha, baptised 2 May 1746, is new to this log as well, and the same sponsor pair stands at her baptism too — the sixth attestation of one couple, across both of the father’s marriages, from 1741 to 1750. That is the hardest internal lock this household has, and it does not depend on how any clerk spelled the surname.

Douglass and New Hanover in Montgomery County — the Falckner Swamp ground where a Welte blacksmith died in 1755 and a Wilte household in 1767 — had never been opened in the county’s own tax returns. They have now been read whole: Douglass, all 58 leaves across five books; New Hanover, all 66 leaves across seven books, 1785 to 1789; roughly 4,750 taxable entries, every W‑surname magnified individually and the complete W population of both townships written into the log. Not one entry on the ladder, in any year those books cover.

What makes it a measurement rather than a gap in the reading is that the neighbours are all still there. Yeager — Conrad the blacksmith in New Hanover, Conrad the farmer and cordwainer in Douglass — and Dachenbach, spelled six ways across the books, stand in every single one; so does Potts. Eighteen to twenty‑two years after the estate that named two of those surnames as appraisers, the appraisers’ families are on the rolls and the decedent’s is not. That agrees, from an entirely independent record class, with what the Falckner Swamp church register says by itself: entries on the ladder run there from 1752 and stop dead in 1767. Two archives, one answer. The bound is stated as well — nothing in this path runs later than 1789, and the federal tax of 1798 is still owed.

The 1755 administration on a JOHN WELTE of Hannover Township had only ever been read at a single leaf. The file was pulled instead, and it is five leaves, closed and complete: wrapper, administrator’s account, account docket, the administration bond itself — never seen before by this project — and the bond docket. The bond gives him a trade the log has never had: “JOHN WELTE BLACK‑SMITH deceased”, administration granted to Casper Singer, bound in one hundred pounds to the Register General on 13 February 1755. There is no inventory in the file. And the sureties are not Weltys — a fourth instance in one day of the assumption that the people standing round a German household are its kin.

Both that estate and the 1767 one now have their recorded book and page: Administration Book G folio 11 and Administration Book H page 41. Neither can be read. The bound administration books are not filmed at all — the catalogue was enumerated, all 350 film notes and 97 titles, and the indexes are filmed while the books they index are not; the printed abstract series runs A, C, D, E, F and stops one volume short of G. This project can learn that a bond was recorded, and where, and can never read it by that route.

The same ceiling stands over the Philadelphia Orphans’ Court, with a sharper edge. It has been carried as the next step in this log since 2 August, by four separate sittings. It is image‑restricted — every thumbnail padlocked, viewable only at a FamilySearch Center or an affiliate library — and it is not in the full‑text index either, which was measured rather than assumed. It was never a search anybody skipped. It is a record class that cannot be reached from a desk, and it is now written down as an errand, with the four films named in the order they should be used.

One smaller correction, recorded because it is a search key rather than a curiosity. The appraisers’ oath in the 1767 file had been read off a sideways scan and its names Anglicised. Turned upright and magnified, the clerk wrote “MARTHIN DACHAUBACH & GEORGE JAGER”, sworn before a Justice signing POTTS. A neighbourhood control aimed at Daughenback and Yeager would have been aimed at forms the manuscript does not use. The spelling rule this project applies to its own surname applies to the neighbours too — and it paid the same day: those are the surnames standing in the Douglass returns above.

A controlled test over one whole printed tax volume — one file, one thirty‑four‑rung ladder, three pattern shapes — turned up something nobody was looking for. The ladder found fifteen entries in that volume. The bare stems WELT, WELD, WILT and WILD, which the ladder does not carry at all, found nineteen more.

It has already cost something specific. “Weld, Jacob, weaver” stands in the 1773 Lancaster borough return between a potter and a shoemaker, one leaf before a Welty freeman entry this project transcribed, inside a neighbour run it had read and written out. He was not missed by carelessness: no pattern in use could match him. Nothing is claimed about who he was — only that he was invisible.

The stems are still not promoted, and York is the reason. On the 1783 Dover roll the assessor prints Wild John and Wild Paul immediately above Weldy Phillip and Weldy Jacob, and Paul’s own spelling runs Wilte, Wilt, Wilt, Wilt, Wild across five printed years. The clerk who wrote Weldy wrote Wild on the same page of the same list, so in York these are two families and the discriminator is the record’s own. Outside York it is untested. Two further whole volumes — Cumberland across six years, and six western counties — were swept on the full ladder and return zero, and they print WILTZ, WILLTS, WILTS and WILDT, four more spellings the ladder has never carried. Three of the four belong to a Michael Wilt of Greenwood Township, Cumberland County, who excludes himself by arithmetic: he is rated there in 1779, 1780, 1781 and 1782, the same four years the Michael of Dover Township is rated in York. One man cannot be assessed in two counties in one year.

A different way of testing the corridor: instead of searching for the family, search for its fellow passengers. Twenty‑five surnames distinctive enough to be unambiguous were lifted from the 1754 signed passenger leaf — Schnurenberger, Dockenwadel, Blumenschein, Feinnauer, Knobelich, Steidtle and the rest — and run, with Anglicised stems, across the whole printed Pennsylvania tax series: eleven counties, 1765 to 1788.

Sixteen of the twenty‑five return zero in every county. The remainder resolve to German surnames common enough to be there regardless, or — in one county — to a shipmate’s surname being used as another man’s given name. The shipload is essentially absent from the printed Pennsylvania tax record, and that is a measurement against the corridor model rather than for it: the Falckner Swamp grouping this project has been working is a cluster of the surname, not a cluster of that ship. And the asymmetry is worth ending on. In York County the ladder returns thirty‑seven tokens and the shipmates return none. The Michael of Dover Township sits in a county that, on this corpus, not one of his supposed shipmates reached.

Much of this project’s reasoning about German families in Pennsylvania leans on a premise: that the people standing round a household — godparents, sureties, witnesses — are its kin. Three independent tests landed on the same day and none of them supports it for this family.

At Gültstein, the sponsor column of Hans Jerg’s four baptisms was read for the first time. The same couple — Hans Martin Nueber and Elisabetha, with a Mayer pair — stands at every one, across two marriages, from 1741 to 1750. Not one Welthe kinsman sponsors any of them, in a village where his father and brothers were living. In Pennsylvania, twenty surnames lifted from his Gültstein world were run against the entire New Hanover register: sixteen return zero, and Nueber — the tightest tie the German record shows — does not appear at all, in the baptisms or in sixteen confirmation classes spanning 1743 to 1825. And the 1767 bond names the widow’s co‑obligor: not a Douglass neighbour and not a Welte, but Christian Schneider, gentleman, of the City of Philadelphia — a professional surety, signing in his own hand while she signs by her mark.

A fourth casualty, from the same reading. The godfather‑naming custom documented five times over in the Edenkoben line does not operate at Gültstein: neither child there is named for a godparent. A naming custom proved in one branch is not a property of the surname — and namesake hunts aimed with it in the Württemberg thread have been aimed wrong.

The Gültstein baptism of a Johann Georg in August 1741 — the only candidate for the twenty‑two‑year‑old confirmed in Pennsylvania in 1764 — carries a death cross in the register margin. If he died as an infant the whole thread collapses. So the village’s burials were read entry by entry, both pages of every leaf, 1741 through 1748 entire. There is no burial for him, and no Welthe child is buried at Gültstein in any of those eight years.

One Welthe burial does exist in the run and it was not what anyone was looking for: 22 May 1748, “Joh: Michel Waelthe, todtgeb: töchterlin” — Johann Michel Welthe’s stillborn daughter, which puts him at Gültstein as a married householder six years before the crossing. That is what makes the zero worth logging: a sweep that demonstrably detected the surname when it was present is a different grade of negative from one that simply reports nothing.

And one near‑miss refused. March 1746 offers “Hß Waelther hier” — tempting at viewer scale. Cropped at full resolution and set directly beneath the 1748 Waelthe in the same clerk’s hand, the terminals do not match: the 1746 form carries a distinct Kurrent -er. It reads Walther, a different and common surname. No rung is claimed, and it is written down so the next reader does not claim one.

The first bounds a great deal of past work. In the printed Pennsylvania Archives tax volumes the rated taxables print surname first — “Walty, Dan’ll” — but each township’s list ends with a freemen section that inverts the order: “Jacob Walty.” A search pattern shaped Surname, Given therefore matches no freeman anywhere in the series. A freeman is a landless adult male over twenty‑one — a son of the house, or a young immigrant — which is precisely the demographic every unresolved question here turns on. Two sweeps of this log claim forty‑three volumes between them and do not record whether their pattern was bare. Their coverage is now stated as rated taxables only until somebody measures it.

The second is a rule quoted from the archive itself: a county’s warrant register begins when the county does, and earlier records were never transferred from the parent county. So this project’s finding that no Welty warrant exists in the York register is correct and correctly bounded — it measures the register, which starts in 1749, not the land. Everything before that lies in Lancaster’s volume, which is exactly where the four warrantees above were sitting.

George Shram of Manchester Township, yeoman — father of the Anna Maria who married into the Manchester Welty household — made his will on 3 October 1772, and letters testamentary were granted on 10 November 1772. He died inside those five weeks; this project previously had only “before 1790.” His widow Elizabeth takes a third of the moveables and the plantation “during the term of her widowhood” until the youngest child turns twenty‑one — and she was still a widow eighteen years later. He devises the plantation to “my nine children … share and share alike”, which confirms in the testator’s own words the nine‑share arithmetic that had been inferred from a 1790 quit‑claim. He signed by mark, as his daughter still did in 1790. And a long‑standing puzzle dissolves: the Welshhans who turns up as a court‑appointed guardian in that 1790 deed is there because Joseph Welshhance had been George Shram’s own executor eighteen years earlier.

The printed Berks tax volume yields ten candidate tokens on this family’s spelling ladder. Read at the leaf, eight of them are not on the ladder at all. The Rockland Township cluster — a Michael, a Jacob, a Peter and a Gottlieb, the exact shape a migration corridor would predict — is the WEIDA family. The Cumru man who prints as Welth is a Weith, one man across three spellings. The remaining two are real, and they are the one Berks man this project already had: George Welde of Oley in 1767, George Welte cooper of Oley in 1768.

The earlier Berks measurement therefore holds, and is upgraded from machine grade to printed‑page grade. What is worth keeping is the warning attached to it. A machine sweep trusting the text layer of that volume would have reported a Welty colony in Rockland Township — and it would have looked exactly like the thing the model predicts. A false positive that confirms your model is the most dangerous kind there is.

The man printed immediately below the Weltys in both years is a Michael Wilhelm — sixty acres in 1781, forty in 1782 — and this page has already flagged him as the likeliest way a second Michael Welty could get invented by a tired reader. At the image he turns out to be a worse trap than that, and for a different reason. The 1781 page is broken type. The surname sets as “Wiinilm” or “Wilnilm” — letters physically damaged in the 1898 printing, not mangled by a scanner, as this project had recorded. The 1782 page sets the same man cleanly as WILLIAM, Michael, 40.

So one volume gives one man two surnames in two consecutive years, one of them not a word in any language. And with the letter-grouping established above, the other half of his story goes too: he was never standing next to the Weltys in any sense that means anything. He was standing next to them in the alphabet.

“WELTE, George — Officials — Dover Twp — March 1775 — supervisor of highways,” docket book 1‑3, page 261. In Dover Township before 1801 that index holds exactly two entries anywhere on the ladder: this one and Philip Welty in 1786. It is not a thin corpus — the same index carries 342 Dover rows across 133 surnames for 1775–1800 — so the scarcity is a real measurement rather than an artefact of coverage.

The temptation here is obvious and is being refused. A George Weldy baptised three children at Strayer’s Reformed in Dover in 1763, 1764 and 1767, and that last baptism is currently his final dated record; if the two are one man, that bracket moves eight years. But the index gives no age, no father and no residence beyond the township, George is not a rare name, and the entry sits in tension with a “Walty widow” who first appears on the Dover provincial tax of the very same year. A serving supervisor of highways and a widow in one township in one year are not obviously one household. The entry is filed unassigned, with the test written beside it: the 1775 Dover tax, read against this.

The other Dover entry is “WELTY, Philip — Tavern — Dover Twp — July 1786,” docket book 14, page 124. That classification contradicts a reading published on this page: an entry that walked back an older tavernkeeper claim, on the grounds that the July 1786 licence list, read at the image, holds no Dover township and no Philip.

Both halves can be checked, and the index says that reading was right about the page and wrong about the list. Page 123 does hold no Dover — confirmed. But the list continues across the page turn. Reconstructed township by township out of the index, page 123 carries 20 entries in 6 townships and page 124 carries 63 in 18, all 83 classified as tavern licences at the July 1786 sessions — and Dover, with Philip on it, is on the second page. The count even reconciles with the roughly one hundred men the original reading tallied, once the two townships whose queries overflow the server are added back.

It is left standing as a contradiction on purpose. A licence would be the first occupational record this project holds for him, and he appears in no other year, in a township whose licence names otherwise recur year on year. That is exactly the sort of claim that should not be settled by an index against a page somebody read with their own eyes. The manuscript has not been re-read, so nothing is re-graded; what is written down instead is the ten-minute test — does a running heading carry over onto page 124, is there a fresh heading, are the entries numbered?

Both new rungs — the one that gains a letter at the head, and the one found last week with a two-letter tail — were typed into two large record collections for the first time, across the eighteenth century and the nineteenth. In the eighteenth every new shape returns zero, against controls that reproduced this project’s own earlier counts exactly, so the zero is a measurement rather than a shrug. Neither new rung reaches the arrival window in those corpora.

In the nineteenth, one of them returns its first record anywhere: William Henrey Whelty, born 4 November 1842, baptised 10 October 1846 at Grace Evangelical Lutheran, Rocky Hill, Frederick County, Maryland. The same man is carried as William H. Whelty in 1900 and as William Welty in 1870 — one household lettered both ways across fifty years. Nothing is merged and no relationship is claimed. What it does is corroborate a map label from an entirely independent chain of records, which is what a single-document reading needs before anyone leans on it.

The standard wording in these books is “Letters Testamentary in common Form” — which is exactly what stands at Peter Welty’s own memorandum in the previous volume, and at Bartholomew Maul’s beside it. The Register wrote the longer phrase, “with a copy of an English Translation of a Will in the German Language annexed,” only where the will really was in German. Two testators carry it in 1764: Wampffler and Henry Worth.

Two consequences, and the second is the useful one. Where a York testator wrote in German, the will book can only ever hold an English translation of him — the German original stayed with the estate packet, so no will book will give such a man his own orthography. And the phrase is a searchable marker: it finds the German-language testators of a county in one pass. It also sharpens an errand already on the list, since a will that reached the Register untranslated is precisely what a volume of abstracts of unrecorded wills would hold. Peter Welty’s, for the record, was not one of them.

Book B carries its own front index, and its W leaf holds no name on the ladder — ten entries, all enumerated in the log, Welshantz through Wilkeston, with the facing leaf clean. That measures the index, not the book, and the proof of the limit is sitting on page 26: Abraham Weldie is in the volume and not in its index, because the index lists testators and he is an executor. A whole class of Welty appearance — witness, executor, guardian, adjoiner — is invisible to every index of this kind, which is the same lesson the orphans’ court compilers taught this project four days ago in their own introduction.

Worth recording because it unblocks work rather than closing it. The search that found this household two days ago also established that FamilySearch’s full-text service holds only the two Montgomery County deed volumes and no will book at all. That negative was true and it stopped nothing: the Register of Wills’ own filmed every‑name index produced the will in about twenty minutes, on the first spelling tried. Its arrangement is worth knowing for any county — surname initial plus given‑name initial, each block running chronologically from 1784, so every eighteenth‑century entry sits on the first leaf of its block. When the machine index comes back empty, go to the archive’s own finding aid.

With one caution attached, because it cost an hour. That index’s page column was read through the viewer as 347, which is a different family entirely; pulled at full resolution it reads 327, the first page of the will. A screen‑resolution crop is enough to read a clause and is not enough to read a number or a surname. The finding aid was right and the reading of it was wrong — the second time in two days this project has had to move the blame back onto itself.

The obvious guess about a German weaver in that corner of the county is Germantown, and it does not survive. The only printed Philadelphia County tax lists that exist — 1769, 1774 and 1779 — were swept whole on the ladder and then every W‑surname in the two townships that matter was enumerated by hand. Germantown 1779: four hundred and seventy‑five entries, thirty‑two W‑surnames, not one on the ladder. Springfield in any printed year: none either. And the scope limit matters as much as the zero — Germantown appears exactly once in the whole volume, so this measures one year, not the township.

One entry was tempting and is refused: “Michael Wells” in Springfield in 1779, two years before Weltz bought there, in a county whose transcripts render this very surname as Wells, Wills and Wetts. The base rate says no. Wells is ordinary in this corpus, twenty‑one entries across fourteen wards and townships, and one township carries a “Michael Wells, Jun’r,” which entails a senior. Then the printed page settled it outright: a Michael Wills and a Michael Wells stand twelve lines apart on the same leaf, in the same township, in the same year. Two ordinary men of two ordinary names. Nothing merged, and the man still has no documented existence before 24 July 1781.

Read whole, the eighteenth‑century Welty deed record of York County is essentially one township, and it is not Dover. Eight instruments between 1760 and 1799, nearly all Manheim: Abraham buying from Adam Simon in 1760 and from John Huntzacher in 1763; John buying from Daniel Reisher in 1764 and again in 1768; “Weltee Peter Est to John Weltee” in June 1764, the same instrument standing on both sides of the index; John selling to Michael Newman in 1771; a John in Heidelberg and Manheim in 1799.

So while the Dover Weltes leave no deed at all, Peter Welty of Manheim and his two sons are a continuously land‑transacting York family across four decades, on the county’s own record. Nothing is merged — but the contrast between the two clusters is now measured rather than assumed.

The Dover Weltys run Philip, Michel, Jacob and George; the three Welty men on a 1754 ship run Michel, Jacob and Hans Jerg. It is the kind of echo that feels like evidence. So it was measured against its own county: twenty‑eight thousand two hundred name‑lines parsed out of the printed York County tax returns for 1779 to 1783, across seven thousand three hundred surnames.

Jacob is the second commonest given name in the county at 7.2 per cent, George the third at 5.7, Michael the eighth at 3.0, Philip the sixteenth. Thirty‑five York surnames carry a Michael and a Jacob and a George; thirteen carry all four names — Gross, Hoffman, King, Little, Long, Meyer, Miller, Snyder, Wagner, Welty, Ziegler among them. And Gross is the surname standing two lines above the Weltes on the 1781 militia class leaf. The echo sits at the base rate and carries no weight toward kinship in either direction. It neither supports nor damages the ship‑list question, which is precisely what a base rate is for.

Set down as a warning, because it would be a spectacular false positive if anybody met it cold. On the neighbouring reel the transcript reads: “Guardians may be appointed to take care of the Estate of Michael Welt aged eleven years, Catharine West aged thirteen years and Mary Well aged eight years, the minor children of Adam West … the Court appoint Christian Keller and Adam Welt Guardians.”

The answer is inside the sentence. The three children are stated to be the children of Adam West; the second child is West in the same line and the same hand; the third is Well. One surname is coming out of the transcriber four different ways in a single sentence, and the one the entry supplies as the father’s is West. Michael Welt is Michael West. It is the same discipline this project applies to handwriting — find the control letterform before you name the letter — turned on a machine instead of a clerk. It cuts both ways, too: if a transcript can invent a ladder name, it can also bury a real one under a token no ladder would ever contain.

Also on the record because the rule exists for exactly this. The sweep of the 1782–1803 reel returned a cluster of entries around the death of a John Welty of Manheim, and they were drafted up as a discovery. They were not a discovery. This project had already read both leaves in July, by a different route, and logged them under a different series on a different tab — the capture files were sitting on disk with the matching archive references in their filenames while the row was being written. A new tool re-finds old evidence, and it feels exactly like finding it. The check costs one command: search the log for the record’s archive reference before writing the row, because that string is the one thing identical across tabs, series and sessions.

Three things did survive the dedupe. The docket sweep is a different kind of evidence from the index read it duplicates, and corroborates it from underneath. Two leaves are genuinely new, and they are new for precisely the reason the compilers warned about — a John Welty settling an estate indexed under the decedent’s surname, invisible to any index read. And a correction against this project’s own earlier row: the minor choosing his guardian in 1795 was read then as “aged nineteen(?)” with an honest question mark. At magnification the clerk wrote “aged sixteen.” He was born about 1778–79, not 1776.

On the 1781 class leaf the assessor wrote “Philip Welte” and, further down, “Philip Welte Jun.” When two men of one name stood in one township, this clerk marked it — and the church clerks of the same community had the same habit, writing “(smith)” and “(of Geo)” beside duplicate names. “Michel Welte” carries no mark of any kind. That is weak evidence and it is real: the man writing the list appears to have known only one.

Written down here so that nobody has to re-derive it: a Michael William — Wilhelm — farms directly beside the Dover Weltys in print, sixty acres in 1781 immediately after Philip Welty’s hundred and ten, forty acres in 1782 immediately after Philip’s eighty and Jacob’s eighty. His line is badly garbled in the scanned text layer. One tired pass turns Michael Wilhelm into a second Michael Welty, and he is the reason the whole population had to be read by eye rather than by machine alone.

In the 1783 Dover list of single freemen, Peter Messerly stands two lines from Michael Welty, with an Abraham Messerly beside him. The following year Peter Messerly married a Christina Welty. This is the first record putting the Messerly household on Dover ground, and it strengthens — without proving — the reading that the Christina who married him was a daughter of the Dover Weltys.

Two more lines off the same roll, carried at no weight. In 1780 Dover prints “Wallick, John, 80 / Welty, Philip, 100 / Wallick, Philip, 80” in a run, which gives a Welty–Wallick marriage known here only from family tradition a candidate family on Dover ground for the first time. And Heidelberg’s 1780 return carries “Welty, Jacob” and “Welty, John, 150” adjacent — a line an earlier tally of that township does not include. Manheim adjoins Heidelberg and its own John Welty was assessed at about a hundred and fifty acres two years later, so the two may yet be one man sitting on a township line. Nothing is merged.

The 1897 letter puts the burying ground “a mile west of New Philadelphia, at Blicktown.” Three whole township plates — Goshen, Dover and York — were swept tile by tile. Blicktown does not appear. Neither does Blick. Neither does Pleasant Hill.

The useful part is structural. At New Philadelphia’s latitude the Goshen plate does not map the west bank at all — the township boundary follows the river, and everything beyond it is left white, carrying only the italic words “North West Quarter.” A burying ground a mile due west of the town could not have appeared there, and its absence there measures nothing. The plates that do cover that ground are Dover’s south-east corner and York’s north-east.

Two things on the Dover plate are the right shape and are named by nobody: a church symbol, a building with a cross, standing within J. Knisley’s hundred and fifty-five acres a mile or two west-south-west of the town; and a west-bank hamlet at Dover’s extreme south-eastern tip, between the canal and the river at the mouth of a small run, whose printed labels include “Sch House” and “S. Mill.” A schoolhouse on the west bank at the mouth of a run is the shape the letter describes. It is a shape, and not an identification.

In Class No 3 the five Welte lines sit between two Rothrauffs — one above them, a John Ruthrauf four lines below. Three years later the Michael of Dover Township married Christina Ruthrauff. It is exactly the kind of adjacency that reads like a finding and may be nothing at all: whether these class lists were ordered by neighbourhood or copied off a roll already sorted some other way has never been tested, and a York tax list of 1787 that once looked adjacency‑rich turned out to be alphabetized. The observation is on the record and no weight is put on it until the ordering itself is tested.

The Michael of Dover Township has long been carried as born about 1757 in Chester County. That figure comes from a compiled pedigree, and the only primary age statement this project holds contradicts it: the 1800 census bracket puts the head of his household at 45 or over, with the 26–44 column empty — born by about 1755. The public pages were corrected two days ago; the research log itself still read 1757 until tonight, and both now carry the pedigree‑derived flag on the birth and on the filiation, which no record anywhere states.

There is no birth or baptism record for him in any register this project has opened, and his first appearance in any record at all is a Dover single men’s list of 1779 — twenty‑two blank years in front of it. One class of record aims straight at that blank and has never been swept anywhere in this log: York County’s orphans’ court and guardianship dockets, 1755–1775. Guardianships state parentage. That is the next thing to be done.

The sponsor table was then read on from that leaf — twelve exposures, nine distinct openings, about 210 baptisms, births from 1826 to March 1836, every line, all three columns, on the full spelling ladder and on the tail of each name as well as its head. Zero. So the household above has no second child at this charge; the January 1825 couple have no child here in eleven years of marriage; and a Philip who married in 1812 has none in this run either. Three measurements, not three assumptions — and one of them now has the census explanation above. The reading stopped at the twelfth exposure; the leaves past it are still dark.

And in Tioga County, a second silence closes for a reason worth knowing. A household of these names appears in a Lutheran register there as a communicant without ever being baptised, married or confirmed at that church — the profile of someone who arrived holding a transfer certificate from another congregation, which would name where they came from. The congregation’s own constitution, adopted 30 May 1847, closes that road: at this church confirmation itself is the mechanism of reception into membership. There is no separate register of members received on certificate, and there never was one to find. The same eight leaves, however, gave something better than the thing being looked for — a burial register that writes each person’s birthplace by township and county.

Reading the typed marriage column leaf by leaf turned up something the eye only catches when the leaves are consecutive: no marriage at all between August 1789 and October 1793. The register runs continuously either side of it. That gap sits on the death window of Philip Jacob Welty, which makes it worth knowing whether it is the transcriber’s or the church’s. It was tested the same afternoon; the answer is above.

All three read at the image, none attached to anybody, all in the same typed run. “Mr 17 1812 — Philip WELTE — Catharine KNAB” (page 328) — Welte is the spelling under which the largest historical‑society index files the York Manchester family. “My 15 1775 — Jacob WELTY — Anna Maria SCHRAMM” (page 309), which does real work: this register holds two marriages of a Jacob to an Anna Maria, 1775 and 1785, and the open question published here two nights ago was which of them fathered the four children baptized 1784–91. The earlier bride now has a maiden surname, and the 1785 entry sits a few unopened leaves back. And “Mr 26 1775 — George GAUF — Elizabeth WELTY” (page 309) — the earliest of the women of this name marrying out, into a surname already standing in this corridor.

Two more things came off the same leaves. A measured gap: no marriage is transcribed between 19 July 1825 and 22 April 1827, twenty‑one months in the middle of the window that matters most, and the manuscript leaves that would fill it are named in the log. And a Wild family is at this charge after all, spelled Wildt and Wilt — five marriages between 1806 and 1827. Not one is a Jacob or a George, and all of them are a generation after the 1780 marriage of Philip Jacob Welty to Anna Maria Wild. Nothing merged. The 1780 marriage itself should be sitting in this same typed run, ten or so images earlier, unopened.

The second leaf of the typescript is the register’s own title page, dated 1754, and it describes its contents in one clause that changes how every York entry on this site should be read: the baptisms are of children “welche … in diesen, u. denen nächst umliegenten reformirten Gemeinden getauft worden sind” — in these, and the next surrounding Reformed congregations. This site has said for a week that a York entry carries a four‑way uncertainty about where the event happened. It is not four‑way and closed. It is four‑way and open‑ended, and the book said so on page one — which is exactly what the burial parentheses show from the other end of the volume.

The same page gives the four points in the clerk’s own spelling — Jork, Croisses Creek, Catores, Pramothien — and a 1763 burial names a man of “Crays Grigg.” A search built on modern place spellings walks past all of them. It also announces the sponsor column in so many words, “die Kinder mit Eltern und Tauffzeugen,” eighteen years before the earliest baptism this project has read out of it. One conflict is flagged and not resolved: the leaf signs itself “Catores 13 April 1754,” while this log has carried the charge as beginning 13 April 1744. One digit, in a transcriber’s reading of another man’s hand; the manuscript leaf settles it and has not been examined.

And a leaf this project’s own triage called a section divider is a second title page: word for word the first one, with “VOLUME III” in place of “VOLUME I.” So the reel carries the first and third volumes of the 1935 transcript, and the missing Volume II is a numbered part of that transcript rather than a lost successor register out in the world. Its baptisms of the 1870s run in three columns with no sponsor column, unlike the first volume’s four — which is why the 1762 burials sit in the late volume as back‑matter.

The reel segment that produced the Tioga household published here two nights ago holds a fourth filmed item, and it is a different church: “Zion Evangelical Lutheran Church, Sebring, Tioga County, Pennsylvania — parish registers, 1836–1870.” Two Lutheran congregations in one county, filmed back to back, under a browse label that names neither of them. Its surname index has never been read by anybody; read whole, letter by letter, it holds one W headword — “Welte: 4” — and no V, U or A section at all.

Page 4: “Friedrich WELTE / Marianna nee Miller”, son Wilhelm Konrad, born 7 March 1838, sponsors Konrad Gundlach and wife. That is a fifth child for the Friedrich Welte household — landing in the one four‑year gap its Salem baptisms leave — in a second congregation, and it is godfather‑naming again, this family’s documented custom, now a third documented instance. The household is still not merged with the York Frederick Weldy who married a Mary Miller in 1825, and the test is still the 1850 census of that county, which now has a five‑child household to check against.

One caution for anyone searching that database by county. It files these Tioga baptisms under a place in Somerset County, two hundred miles away — the browse label of the first item on the reel, applied to a different congregation in a different county. And its place filter reaches only about a third of the records in a result set: filtering by county silently discards the rest. A negative taken from that filter is not a negative.

One more leaf of the York typescript, printed page 299: “Members received into the German Reformed Congregation on CERTIFICATE” — and every entry names the body the person came from: the Presbyterian church at Sinking Valley in Huntingdon County, a Methodist church in Baltimore, the German Reformed congregations of Lancaster, Chambersburg and Frederick City in Maryland. One names where a woman went to: “dismissed at her request to the German Reformed Congregation at Gettysburg.”

No name from this family is on the leaf. It is recorded because of what the class of record is. This project’s standing difficulty is people appearing in one county and vanishing from another with no document connecting the two — and a certificate of transfer is the ordinary church record that states both ends. The three Weltys received as adults at the Tioga church in May 1857 came from somewhere; this is the kind of book that would say where.

Four more households come out of the same book, none of them attached to anybody on this tree and all of them firewalled from the York and Somerset lines on purpose. John WELTI and Elisabeth Lenhard, a son Calvinus born 2 December 1836 — a Reformed given name in a Lutheran book. Philip J. Weldy’s four children, three of them baptized together on Christmas Day 1852, which is the signature either of a household newly arrived in a congregation or of a congregation newly supplied with a pastor, and neither reading is adopted. Jacob WELTE and his wife Elisabetha standing sponsors on 26 May 1833 at Ulrich Furrer’s daughter Elisabeth — a man who appears in this book only as a godparent, and would be invisible to any search that indexes parents. And on 17 May 1857, William, Sophia and Sarah WELTY confirmed together, Sophia baptized the same day with no parents named and standing as her own sponsor. That is an adult reception: those three came from somewhere else, together, and that somewhere is worth chasing.

One candidate presents itself and is not being taken. A Frederick Weldy married a Mary Miller at York on 6 January 1825 — found the same night, in the entry below this one. The Tioga Friedrich’s wife is Mary Ann Miller, and his first recorded child arrives in July 1833. Given name matches, wife’s surname matches on both halves. It is also exactly the shape of coincidence that has burned this log before, and a Frederick married to a Miller is not a rare event in this population. Not merged, not adopted — and the test is cheap and named: the 1850 Tioga County census, which gives every household member’s age and birthplace, and in which a Frederick with a wife Mary Ann and children born 1833, 1835 and 1841 either exists or does not.

One caution for anyone reading indexed entries off this volume: it is not a single‑point charge. Page 13 records the “beginning of the ministry of the third congregations at Blockhouse on Aug. 24, 1841.” A place stamped onto one of these baptisms is the book’s place, not the entry’s.

The ninety leaves in the middle of the 1744 volume, never opened by anyone here, were triaged this evening. Image 46 is 1802, headed “Kind | 1802 | Eltern | Taufzeug” — so the manuscript carries a sponsor column of its own, not only the typescript whose sponsor column turned up earlier tonight. Its facing page records the pastor’s removal to Fredericktown in Maryland, last baptism 17 October 1802, and his successor’s first, 13 May 1804.

And at image 81 a new section opens: “Record of Baptisms”, beginning 27 February 1821, in English, sponsors written inline — and the clerk names the preaching point in the right margin. “Creutz creek” twice, “Cadorus creek”, “Hanover”. This register served four congregations from 1744, which is why every York entry on this site carries a four‑way uncertainty about where the event actually happened. From 1821 the clerk resolves it himself, entry by entry. The section runs to about image 91 — and it covers precisely the window of the two marriages above. No name in it has been read yet; this is thumbnail grade, and so is everything else in those ninety leaves.

One measurement to close the night’s bookkeeping, because it changes what “searched” means here. The Pennsylvania church‑records collection this project leans on was swept for the whole eighteenth century and closed months ago. Its nineteenth century has never been touched. Counted for 1800 to 1849 alone it holds roughly three hundred records under one spelling group, two hundred and eighty under a second and a hundred and thirteen under a third — and standing in that third group is a “George Washington Velti”, who is the eldest child of the Tioga household at the top of this entry. About seven hundred records of the right shape, in fifty years, in the half nobody has swept.

The baptism section was then read entry by entry, every column, from January 1783 to late 1795 — roughly two hundred and ninety baptisms. Exactly one ladder household appears as parents (Jacob Welty and Anna Maria, above) and exactly one ladder man appears as a sponsor (Jacob Weldy, above). That is the whole ladder presence of thirteen years at this charge.

Two absences are worth more than they look, because both are people this project has watched marry in this very book. Peter Messerle married Christina Welty here in April 1784 and baptized no child here in the eleven years that followed. Peter Becker married Anna Maria Weld here in September 1787 and baptized no child here in the eight years that followed. Neither couple can be run forward through their own children out of this register, which was the obvious next move for both. The reopeners are named: this is the Reformed charge, and the Lutheran congregation at York has its own book; so do Codorus and Bermudian once the four‑point charge divided; and this register’s baptisms run well before 1783 and well after 1795, both ends unread. Michael and Christina, married here in May 1784, likewise baptize nothing here — which is a fact about this register and not about them, and sits comfortably with the household being in Somerset County by June 1786.

Two days ago this site made much of the heading above the 1754 crossing: the clerk at the Philadelphia court house describing the ship Richard & Mary’s passengers as “Inhabitants of the Dukedom of Wirtemberg,” with the three men on this ladder signing immediately below it. That heading has now been tested the only way it should have been tested in the first place: all five hundred and twenty‑three list headings in the printed volume were scored, 1727 to 1775, for whether they name a territory at all.

The field is period‑bound, and that is the finding. Every list from 1731 to 1741 carries a territory clause. Then: none at all in 1742–44 and 1748; eighty‑two per cent of them in 1749; one out of thirteen in 1750, none out of twelve in 1751, none out of fourteen in 1752, none out of eighteen in 1753; eighty‑eight per cent in 1754; and essentially none for the twenty years after. It is a clerical habit that switches on and off by season.

Half of the earlier claim survives and half does not. What stands: within 1754 the clause is not boilerplate — fifteen of seventeen lists that year carry one, they name different territories, and this is still the only 1754 arrival described as Württemberg and nothing else. That comparison is unharmed and remains the best American‑side datum the Württemberg thread has. What does not stand: the idea that a silent heading says anything about its shipload. The Royal Union’s silence in 1750 — Philip Jacob’s own crossing — is one of thirteen silences that year. It is a clerk’s habit, not a statement about any ship, and no split origin should be read out of it in either direction.

The same sweep produced a structural fact that constrains every future attempt at these three years. Strassburger printed three classes of list where they survived — captain’s list, oath list, signature list. For the whole of 1750, 1751 and 1752 — the Royal Union, both Neptunes, the Janet, the Forest, the Snow Ketty — only one of the three survives at all. There is no second exemplar to compare, for any ladder arrival in those years, and there never will be from this book.

Two ships named Neptune qualified their passengers at Philadelphia on consecutive days in September 1751, under different captains, and the running head at the top of the printed page says only “Neptune 1751,” which is true of both and identifies neither. This log picked the wrong one. The Johannes Welte it has carried since July is on Captain John Mason’s list of 24 September, not Captain James Wier’s of the 23rd — established by walking back to the bracketed heading itself, which begins at the foot of a page two leaves earlier. There is no ladder name anywhere on Wier’s list.

That matters because a resolution rested on it. This man was identified in July with a Catholic John Welty of Maryland on the reasoning that Wier’s Neptune carried a particular cohort. He was not on Wier’s Neptune. The identification is not disproved — the two ships sailed the same route a day apart — but its stated ground has moved, the cohort that has to be checked is now a different and much larger one, and the man goes back to unassigned. This is the third time a running head in this one volume has mis‑assigned a ship in this log. A running head names the ship a page mostly belongs to. Only the bracketed heading settles a boundary.

And because the whole arrival inventory was rebuilt rather than spot‑checked, something new fell out of it. On the ship Sarah, 20 September 1764, “Johannes Welte” and “Johan Philipp Welde” stand consecutively, both signing their own names — the fourth pair of ladder men to cross together, and a crossing this project had never looked at. Nothing attaches to anybody. The reason it was never looked at is the part worth recording: the arrival inventory stopped at 1754 out of habit, not evidence.

One more thing opened while the volume was in hand. Strassburger’s third volume is not only an index — it prints the Philadelphia arrival lists of 1785 to 1808, which are the only arrival class in this corpus that names women. Its own index already shows a ladder woman on one of them. That volume is free, it needs no login, and this project has never opened it.

The last entry named a cheap next step: an administration bond sits within a few leaves of the account and names its sureties, who in a German settlement are overwhelmingly kin. Both have been read, and Michael Welty is not a surety on either. The 1755 bond for John Welty of Hannover Township, Casper Singer administrator, has Michael Walther and Adam Mayer — a name close enough to raise the pulse and, on checking, the New Hanover congregation’s own schoolmaster. The 1767 bond for the Douglass George was entered by Anna Barbara Wilte, “widow and relict,” signing by her mark.

The files gave up two things anyway. The 1755 one calls its decedent “John Welte, black smith” — a trade this project did not have, and the kind of detail that makes a man findable in records that never mention his family. And across the two files the same clerks spell one surname Welte, Welty, Welly, Wilte and Willy, now attested inside a probate file rather than guessed at.

The Falckner Swamp Lutheran register was opened here days ago, but its baptisms stayed dark: a hundred and forty‑six pages of wide table set sideways on the page, whose published machine text is pure noise. Every leaf was pulled at full resolution, rotated, read again by machine, swept against twenty‑five spellings, then every hit read with human eyes. It is the section with the sponsors in it.

Anna Barbara Welte stands godmother in February 1757 — the first woman of that name in this corridor, ten years before the widow of the same name on the 1767 bond — and her surname is invisible to any text search because it wraps onto its own line. George Welte and his wife stand godparents to all three children of the Johann Jacob Welte household, and at the third, in 1766, the printer names her: “and wife Barbara” — the discriminator the Douglass estate has been waiting for, and still not merged, because a sponsor line carries no township and no age. A Jacob Welte stands alone for the son of Thomas and Margareth West in 1762, and this log holds a Margareth Weld who married a Thomas Weston in 1761: on the face of it, a sister.

One entry does something quieter and better. A household baptizes a daughter as Johannes Welde and Christina in May 1753 and another as Welty and Christina in January 1754 — one printer, one book, one household. The equation Welde = Welty is now attested inside a source instead of assumed by us. And the negative: across the whole table, all twenty‑five spellings, no Michael on this ladder anywhere in that register — which names its own reopener on the same page, since all twenty‑five spellings begin with a W.

The other congregation in that township — the Reformed one, same families — was swept whole: a hundred and seventy pages, twenty‑eight spellings, marriages, burials and early baptisms read line by line. Not one name on the ladder. Ordinarily that closes a lead. It does not, because of what the reading measured: the marriage register in this copy is blank from February 1755 to March 1769, and the burial register from September 1755 to February 1772. The marriage this project most wants from that corridor, and the burial that would go with it, fall inside those holes — they were never going to be there, whatever had been searched for, which is a different statement from “they are not there.”

What is clean is the stretch the copy covers: the blacksmith who died by February 1755 is not among the eleven burials recorded in those years. One positive fell out — the administrator of that estate buried his wife in this Reformed congregation in January 1752. And the transcript caught itself being incomplete, in print, on its own contents page: “IV. Confirmations. (Not Copied) 1439 listed.”

The 1754 crossing has been read here name by name for weeks. Nobody had read the sentence above the names. On 30 September 1754 the clerk at the Philadelphia court house wrote that the foreigners from the ship Richard & Mary were “Inhabitants of the Dukedom of Wirtemberg” — and the three men on this ladder sign immediately below it. The other 1754 arrivals are described by the same clerk as Lorain, Alsace, Franconia, Swissers, and four mixed descriptions: this is the only arrival of that year described as Württemberg and nothing else. The first corroboration of the Württemberg case from the American end — and it describes a shipload, not a man. Scoped the following night: across all five hundred and twenty‑three headings the clause proves period‑bound, so the comparison holds within 1754 and a silent heading in other years says nothing at all.

Yesterday this project admitted that a Pennsylvania negative broke when re‑run across the whole spelling ladder. This morning asked the duller question: not was it wrong but what did it actually type. Of the four men it claimed to cover, it queried three given names — Michael, Michel, Jacob — and exactly one surname unattached to a given name: Welthe. George was never typed at all. And the one spelling that ran unqualified is the one this family stopped using the moment it landed.

It was run again properly — eighteen spellings, none tied to a given name, across thirteen Pennsylvania collections for the whole eighteenth century, with the given name read off each hit afterwards. It produced three records, all below. And a caution worth passing on: in an eighteenth‑century Pennsylvania deed, “Welthe” is almost always the word Commonwealth, and “Wealty” is the same word again. The one rung the old search ran unqualified is a rung whose Pennsylvania hits are legal boilerplate.

The 1754 shipload has been traced before by the rarest surname on its passenger list, and that calibration pointed along the Berks–Montgomery–Philadelphia corridor. Berks itself had never been searched. Its printed tax lists for 1767, 1768, 1779, 1780, 1781, 1784 and 1785 hold one ladder man: George Welde in the Oley return of 1767 and “Welte, Geo., cooper” in Oley in 1768 — no acres, no stock, both years, with the same block of neighbours holding its order across the two lists, which is what makes them one man. A separate Peter Weld appears in Cumru. There is no Michael and no Jacob of any spelling in Berks across all seven years, and the Oley man is gone by 1779.

The Oley taxables of 1734 and 1759 carry nobody on the ladder, so this George arrives between 1759 and 1767 — and the reason he leaves no church paper is structural: Oley had one German congregation and it had no settled pastor until 1771. He is taxed in exactly the pastorless stretch. One trap fell out of the same volume: the scan turns printed ei into el with great consistency, producing four handsome false hits for a family spelled Welda, including a Michael. At the image every one is Weida, a different Berks family.

A compiled reference records a two‑hundred‑acre donation‑land grant to a “Michael Welty” in December 1786, never traced to source. The source has now been read — the state archives printed the entitlement roll itself in 1896 — and there is no Welty on it under any spelling, the whole W run read leaf by leaf at full resolution. What the roll does hold is “Welsh, Michael”, and in the compiled book a Michael Welsh with the same rank and the same two hundred acres stands four lines above the Michael Welty entry. Not stated as proof of a transposition, but precisely the shape one leaves. The limit: the eyes‑on read covers the letter W, and a clerk who wrote this name with an F or a V sits elsewhere in the alphabet.

Getting there required breaking the volume open. That register is set as wide tables turned sideways on the page, so its published machine text is upside‑down gibberish while the horizontal running heads read perfectly — the section looks searchable and is not. Any keyword search of it, anywhere, returns a false negative, and nothing in the result tells you so.

The best thing the morning turned up is not a person but a proof about names. In one company of the third battalion of Northampton County militia, under one captain, across the battalion’s own returns of 1780–82, one private is written down under four spellings in five successive returns: Philip Weldy, Weldy, Welday, Wealty, Welty. A church index adds Welti and Waldy for the same household in the same township — six in all. What locks it is not the name but the neighbours: Sheneperger, Walker, Cromly, Diel, Sigll, Rence, travelling with him in every list under the same officers, their own spellings wobbling just as freely.

The spelling “Welday” has been treated for months as the marker of a distinct branch. It carries no such information; any argument leaning on it is void, and the question has to be settled on records, residence and DNA. The same sweep demoted a rung the ladder had carried since a 1938 list: “Felty” in these records is not a surname at all but the Pennsylvania‑German short form of the given name Valentin.

A volume of the printed state archives that a previous sweep listed as unread — then quietly reasoned about as though it had been read — contains “George Welty,” paid a six‑pound enlistment bounty as a soldier of the 2d Pennsylvania Regiment of the Continental Line in 1781, read at the image with the governing heading walked back four pages. That alone undoes an earlier conclusion here that the Continental‑service question was closed.

The regiment’s rolls were then searched for him and he is in none of them. The obvious reading — a bounty taken and a service never rendered — was written into the log this morning and withdrawn again the same morning, because it was tested: all hundred and thirty‑seven men on that account were matched against the regiment’s printed roster, and only about three in five appear on it, roughly a quarter not at all. Absence from that roster is ordinary and cannot carry an inference about anybody. The facts stand as read; the story draped over them does not.

The same leaf settles a smaller question. In the heading and the Michael Welty line the clerk writes Welty — a looped l, then a t with a crossbar running out to the right. Four lines down the other column, paying for the burial, the same hand writes “the funeralls of John Welly” — two ascenders of equal height, no crossbar, identical to the double l he has just written in Campbell. One clerk, one page, one man, two spellings. This project has never had that attestation inside a single document before; the “Welly” rung had rested on machine misreadings and one printed entry that turned out at the image to say Kelly. The consequence is practical: a search for Welty finds the heading and misses the funeral line, and a search for Welly finds the reverse.

George Welty of Douglass Township, Philadelphia County, dead by 21 October 1767, when two neighbours appraised his goods. The inventory is the man: apparel at nine shillings and ninepence, a hammer and anvil, an axe and drawknife, scythes, two casks, a pail, a tub, a reel, a spinning wheel, old books, a bed at fifteen shillings. No land, no stock. A cooper’s and jobbing tradesman’s kit — and the Oley man in the Berks list of the following year is entered as “Welte, Geo., cooper.” One year and one county apart, and therefore two different men, since the Oley one is alive and assessed in 1768.

George Weldy of Berks County, dead intestate by 30 June 1794, administration granted at Reading to his son John, the widow Dorothy having renounced her right, the decedent described as “late a Soldier in the Service of the United States of America.” The first ladder man ever placed in Berks with a family attached — and the widow’s name is a genuine discriminator: a Georg Welde and his wife Maria Dorothea baptize children at Reading in 1771 and Philadelphia in 1772, a couple this project has held for weeks with nothing to test them against. Still not merged, but the question has moved from is there any link to is this the same Dorothy.

Then the arithmetic, which cuts against the tidy version. The Johann Georg this project reads as the third man on the 1754 list was baptized in 1710. A soldier of the United States dying in 1794 was not born in 1710. The 1710 baptism and this administration cannot both describe one person — so either the Berks soldier is not the Oley cooper, or the Oley cooper is not the man off the ship. Saying which is not possible today, and saying that plainly is worth more than an identification would have been.

Set the day’s finds on a map and they stop being a scatter. Inside about ten miles and twelve years, on the Philadelphia–Berks line, this log now holds: John Welty, tenant, of Hannover Township, dead in 1755 or 1756; Michael Welty, his named debtor, in 1756; Johann Jacob Welte and his wife Maria Margareth, baptizing children in the New Hanover congregation to 1763; Johannes Welty and his wife Christina in that same congregation; George Welty of Douglass Township, landless, dead in October 1767; and George Welde, later Welte the cooper, in Oley across the county line in 1767 and 1768.

Six exposures, one settlement, one generation — in the corridor a calibration off the ship’s own passenger list had already pointed at. Not one is attached to a person on this tree, and several are firewalled from each other on purpose. What the pattern changes is the method: this is now one problem, not six, and chasing each George and each Michael separately is the wrong shape of work. It also makes the next step obvious and free — two of these estates were administered through the same Philadelphia office, and an administration bond names its sureties, who in a German settlement are overwhelmingly kin. (Both bonds were read that evening. He is not on either.)

One more pass at the Gültstein baptism register, testing the question the evening left open: could the Michael Welthe who married there in 1732 have fathered a son Michael early enough to be the man who later appears in Dover Township? The book was read line by line, 1739 to the middle of 1745, every father’s name down the parentage column of every page — and in those years there is no son Michael to any Welthe couple. Three baptisms this project did not have turned up instead.

Anna Maria, baptized 8 October 1741, daughter of Michel Welthe — citizen, day‑labourer — and his wife Catharina: the first child recorded to that couple since a son Jacob in 1736, so that five‑year gap is now known to be a real silence rather than an unread stretch of pages. The other two belong to Hans Jerg Welthe and Maria Agnes — a son Johann Georg in August 1741, a daughter Anna Maria in February 1744 — with one godparent couple at both fonts, the sort of cross‑lock that keeps two entries from drifting into two different families.

Two things follow, neither a conclusion. The window this test still has to cover has narrowed to mid‑1745 to November 1747, and those pages are not read. And both 1741 infants carry the register’s mark for a child who died, which promotes a nicety into a real next step: the village’s funeral book for the 1730s and 1740s has never been opened, and a series of infant burials would account for this family’s spacing better than any amount of inference. The parish’s family‑register volume was searched the same session and the household’s page was not found. It remains the single most valuable unread thing in that village.

A week ago the German origin of this line was a family legend with a wrong year stapled to it. Tonight there is a house at Gültstein holding three brothers of exactly the three remembered names, whose record in that village ends the way an emigrating family’s does — graves for everyone who stayed, and none for the three whose names turn up on a 1754 passenger list. That is a hypothesis with its remaining tests written down. On the American side the Dover Michael went from a man known by a marriage and an estate to a man standing on the record year over year, with a column of five Weltes, one under another, on a single 1781 Dover leaf that nobody had read.

And here is the part worth being honest about. The line’s one unproven rung is more open tonight than it was on Monday, not less. Reading the 1800 census at the manuscript removed the birth year that had been quietly doing the work of ruling out the 1754 crossing — it came from a compiled pedigree, not a record. So three readings sit live at once: that this Michael was born in Pennsylvania to the Dover Philip Jacob, which is where the tree still draws him; that he was himself aboard the ship in 1754; or that he was the son of a man who was. Each has a test attached in the log, none is settled, and the tree will keep drawing that link dashed until a record settles it. A genealogy that only ever narrows is a genealogy that has stopped checking itself.

This has been an unusually heavy stretch — long daily runs, whole films read end to end, registers worked line by line — and that pace has a bill attached. The German church‑register subscription has lapsed; the American ones have roughly a month left. So the heavy research pauses for a while. Not stopping — pausing, with the log, the leads and the unread page numbers written down precisely so the next run starts where this one stopped. Everything already found stays where it is: free, unwalled, public domain.

Several of the most promising things left are free, online and open to anyone. If you have an evening and a tolerance for eighteenth‑century handwriting: the York County township assessment film (FamilySearch 008350826, 793 images, home‑viewable), the film that produced the 1781 Dover leaf, with many Welte‑bearing leaves located but not read closely and the Manchester ones barely touched. The ship‑list facsimiles — Strassburger’s Pennsylvania German Pioneers volume II, holding the actual signatures, including two same‑day exemplars of the 1754 trio; and volume III, which prints the 1785–1808 arrival lists, the only Philadelphia arrival class that names women, and which this project has still never opened. The Gültstein registers — the baptisms from mid‑1745 to November 1747, the funeral book for roughly 1736 to 1748, and the parish family‑register volume. The Pennsylvania Donation Land patents (state archives, free, no login). The printed Pennsylvania Archives depreciation lists on archive.org, where mapping the block headings would turn every bare name into a name with a county attached.

And newspapers, Bibles and drawers. The York papers of 1789–1798 have never been browsed for the estate notice that would be the first direct record of Philip Jacob’s death. And the thing no archive holds: a family Bible, a German Familienregister, a fraktur baptismal certificate, a hand‑copied sheet of names in somebody’s attic. Those have decided more questions on this site than any subscription has.

With the township question settled, the film that raised it was worked to its end — every Welte‑candidate leaf on all 793 exposures opened at the image. The Dover Michael now stands on the record year over year: the 1781 class list places Michel Welte in the township’s third class, inside the five‑man Welte block; the same year’s property valuation lists him again among the single men, a Jacob of the name beside him; in 1782 he occupies a forty‑four‑acre farm; in 1784 he is on the township assessment once more — and on 18 May 1784 he marries at York. One township, one name, four documents, straight into the marriage record.

The same read‑out killed two lookalikes at full resolution — a promising 1785 “Welty” trio that resolved into the Holly family, and a “Willy Michael” that resolved into a Worley — and opened a thread deliberately not attached to anyone: a Jacob in Manchester Township, present every year from 1781 to 1785 among Manchester’s own community of names. A Jacob also recurs beside the Dover Michael. Whether either is the same man, a brother, or another family is a question for records, not name‑matching. One limit holds through the whole run: not one of these documents states an age.

A living cousin, reading along, wrote to say he had found the 1754 trio himself — and to propose that the third signature reads not “Hans Jerg Welte” but a different name altogether. Run down at the facsimile, it paid twice. The check surfaced something never catalogued here: the volume holds a second, same‑day exemplar of all three signatures, so every reading of these hands can now be tested against two independent specimens. And the proposed reading does not survive: at both exemplars the letterforms rule it out, and the ship list’s clerk — a fourth, independent hand — wrote the man down as “Hans Jurg.” Hans Jerg Welte stands. But one instinct in that letter was exactly right in a way its writer could not have known: reading “Hans Jurg” as a Johann Georg is precisely what a German marriage register proved the same afternoon, an ocean away.

The same volume holds List 218 B: the ship Richard and Mary, 30 September 1754. In the left column, one directly above the other, stand Michael Weldta and Jacob Weld[t], each in his own hand, neither a mark. The men behind the legend now have faces. What those faces cannot do is prove kinship — and that was tested rather than assumed. Nine unrelated German immigrants sampled off the same page share the same loops, the same rounded W, the same slant: the letterforms are the curriculum, not the person. Sharper still: the 1754 Michael’s hand resembles the 1750 Philip Jacob’s at least as closely as the 1786 signature does, and nobody claims those two were kin. The ruling is now in the log: letterform resemblance is not evidence of identity or kinship in this corpus, in either direction.

The family papers hand the 1754 trio a tidy dispersal: Jacob to York, Hans Jerg to Hagerstown, Maryland. Maryland had never been swept. It has now been — the 1776 state census and 1778 oath rolls read eyes‑on at their index leaves, the 1783 Washington County tax, the 1790 federal census, the county histories, and the full‑text corpus for the whole eighteenth century, across every spelling on the ladder. There is no Michael of any spelling anywhere in Maryland in the window, and — the sharper result — no Hans Jerg at Hagerstown either: not one George of the name in Washington County until an 1808 bond surety, half a century after the crossing. The trio’s “known” dispersal was never known. Maryland’s real eighteenth‑century Weltys are there — Johns, Jacobs, a Christian, a Frederick, in traceable and separate streams — and none of them is any of these men.

One more correction, freely admitted: yesterday this project recorded that the 1754 shipmate “leaves no post‑arrival trail under his own spelling in Pennsylvania.” That negative rested on four keyword forms. Re‑run today across the full eighteen‑rung ladder, it broke — and what broke it is the day’s last find. On a militia‑class assessment film this project had never worked stands a 1781 York County leaf carrying, in one column: Philip Welte Jun · Michel Welte · Jacob Welte · George Welte, read at the image at high zoom. Which township it belongs to is now the live question, and it cuts both ways cleanly: the warrant sheets ahead of it name Heidelberg Township, a different Welty family’s country — but the neighbours on the leaf, the Quickels, Ramseys, Reisingers and a Wilt, are the documented Dover and Manchester community. Until the township is established, nothing on that leaf attaches to anyone.

The 1906 print gives the discharge under “Michael Willty.” The manuscript does not. Read at full resolution the clerk’s surname carries a plain double t — Weltty — and the vowel is the ordinary ambiguity of the hand, not an i. So one rung of this project’s spelling ladder was never a rung at all: it is a nineteenth‑century editor tidying a soldier’s name. The ladder has found real men, but it is worth knowing which of its rungs were cut by clerks and which by typesetters.

The state settled this man’s militia service with certificate 12,706, issued 26 October 1786 for £10.10.0. The assignment is dated 22 April 1786 — six months earlier. Whoever collected that certificate, it was very probably not the soldier: he had already sold the claim. The interest register in Harrisburg records who was paid on each certificate, and the leaf is known — volume B, folio 95. Whether Jacob Sittler’s name stands there is a specific question with a specific page number attached, which is the most useful shape a research question can take.

On the other side of the project, a subscription ran out and two loose ends were pulled in before it did. The Edenkoben Reformed baptism register has been searched for the baptism of the Croissant daughter who married into this family in 1738; the last unswept leaves were read this morning, and she is not in that book between 1713 and 1723. Nothing after that can be her either — a baptism in 1724 or later would make her a bride under fifteen. If she was baptized at Edenkoben at all, it happened before 1713. Separately, the Lutheran marriage register at Gommersheim gave up its remaining years — 1736 to 1743, read entry by entry — and returned nothing under any spelling. With that, the ring of parishes around Edenkoben is finished. Two candidate books remain, both long shots.

Two certificates, both read as page images. On the first, Capt. Peter Ford certifies over his own signature that a soldier “as Sarved is two month in Camp Securtey in my Companey in the Forth class of York County Militia” — and that soldier is himself a private on Ford’s July–September 1781 pay roll, so the class statement and the roll describe one body of men. On the second, a man declares: “I Joseph Bigler of Dover Township York County have in the 4th class Sarved a touer in the Militia in Camp Security … for Casper Spahr of the Same place” — with the clerk’s own footnote underneath putting Spahr in the 4th as well. The class that guarded the stockade that summer is the class in which the only documented Dover Township men appear. That is an alignment and not a proof, and it rests on a single document naming two men — but the objection it answers was the last structural one standing, and what replaced the objection leans the other way.

A Michael Wellty appears on Ford’s pay roll. A Michael Willty has his own discharge certificate, dated at Camp Security on 20 September 1781 and signed by Lieut. John Mappin. These have been carried as two records in two spellings. Mappin signs twelve discharges in the alphabetical section and every one falls between the 11th and the 29th of that September — and four of the men he releases (Bowman, Philip Henry, Ramsey, and McWhirter, discharged “as Orderly Serj.”) are privates on Ford’s roll. Mappin was letting Ford’s company go as its two months ran out. So it is one service by one man, enrolled and paid at one end and discharged on his own certificate twelve days after the roll closed. The weight of the evidence does not double. The record beneath it gets a good deal harder.

Pennsylvania settled its militia debts through a loan raised in 1784 and 1785, and the Comptroller General issued each veteran a numbered certificate. Three York County Weltys hold one. John, certificate 12,665, issued 26 October 1786. Michael, certificate 12,706, the same day, for £10.10.0. Philip, certificate 13,453, issued 30 December 1787, for the same sum. That figure is not arbitrary: a receipt in the same archive itemises it as “Five Pounds for Two Months Melitia Duty and Five Pounds Ten Shillings state Bounty” — the settlement for exactly one two‑month tour. Each card names the leaf of the interest register the entry was copied from: folios 94, 95 and 117 of volume B. That ledger is still in Harrisburg. This trail had been written off a week ago as unrunnable, on the correct finding that a particular printed volume could not carry it — which was true, and beside the point. It runs through the comptroller’s own records, and it now has three numbers and three folios attached to it.

Two cards away in the same drawer, under an identical name, sits a Philip Welty of Northampton County — John Ritter’s company, under Lieut. Col. Christian Shouse, 1st class, entered 16 September, duty recorded as “actual service on the frontiers.” Wrong county, wrong war — frontier defence, not the prisoner guard — wrong officers, wrong class. He is filed with the decoys, where he will be waiting for the next researcher who searches Philip Welty, Pennsylvania militia, 1781 and takes the first hit.

A lead had been open for weeks on a reasonable premise: York County had no newspaper of its own until 1789, so the county Lieutenant would have published his militia delinquents — the men who failed to turn out for their class — in the Philadelphia press, where a name might sit that no muster roll preserved. The Philadelphia papers of those years are digitised and searchable, and the phrase to look for is a fixed one. Across every location, 1779 to 1788, “list of delinquents” returns nothing at all. “Delinquents in the militia” returns two, and both are the same thing: the text of the 1787 militia act, which directs that delinquent returns go from the commanding officer to the county Lieutenant. It was a filing, not a notice. Nobody printed it because printing it was never what the law asked for. The premise was wrong, and the search that proved it wrong is worth more than a hundred empty queries against a surname — a negative that rests on what a document type was survives bad transcription, and a negative that rests on not finding a name does not. Three archive volumes long carried as “missing” closed out the same afternoon: two of them turn out to be forfeited‑estate books that never held a muster roll, and the third covers 1783 to 1790, which is after the war it was wanted for.

Back in the Palatinate, in the Edenkoben Reformed baptism register, folio 346: “d. 26 [Sept] nat. / 29 bapt. Hanss Jacob Waeldy, & ux. Anna Cath. filius. Testes: H. Georg Wolfg[ang] Georg Schaefer, Kirchenschuldiener, & ux. Anna Elis. m. inf. Georg Wolfgang.” Born 26 September, baptized 29 September 1716. The surname carries both its marks plainly — the umlaut over the a, the diaeresis over the y — so there is no reading to argue about. The man standing as godfather is the village Kirchenschuldiener, the church schoolmaster, and the child is given his name. This family named its children after their sponsors and not after their fathers, which is now documented at every baptism in the sibship, and it is the reason a namesake hunt in these registers has to chase godparents. The entry had been inferred before, from a bracketed note in a sibling list. It is now an entry: a date of birth, a date of baptism, a leaf, and two witnesses.

Three weeks ago a reading of this register reported a faded August 1714 baptism for a daughter, Anna Barbara, with a godmother from Neustadt. A later sweep went back through 1714 entry by entry, found nothing, and concluded that no such baptism existed and that the girl’s year had only ever been back‑calculated from her age at confirmation. Both accounts could not stand. The page was opened again and read at full magnification, and the entry is there: “Aug. d. 4 nat. / d. 6 bapt. Hanss Jacob Waeldj, & ux. Anna Cath. filia” — born 4 August, baptized 6 August 1714, the godmother an Anna Barbara of Neustadt, the child named for her. It is the final line of the leaf, squeezed in under the last full entry in a smaller and lighter hand, on a lower edge that has browned with age. At the magnification a page‑by‑page sweep normally runs at, it reads as staining rather than as writing. The first reading was right and the second was wrong, and the correction has been written into the record beside it rather than quietly replacing it. The rule it leaves behind is a cheap one and will be followed from here: read the last line of every page deliberately and separately — a scribe filling the foot of a leaf writes small, and that is exactly where an eye running down a column has already stopped paying attention. With the entry recovered, the Edenkoben children of this household are complete and every one of them is dated: 1710, 1713, 1714, 1716, 1719. The intervals are ordinary and there is no gap left for a sixth to be hiding in.

Last week a Pennsylvania newspaper trail argued that the Stark County Philip Welty (b. 1789) served in Col. Rees Hill’s regiment — the thousand‑man Erie detachment raised in April 1813. This week the federal index answered: a card exists, and it reads “Welty, Philip / Hill’s Regiment Pennsylvania Militia / Private.” A Philip Welty demonstrably served in that regiment. What the card does not carry is an age or a residence, so the identity is not closed — the muster rolls behind it, held at the National Archives and orderable for a fee, would give an enrollment date and place, and that is the test. Two New York decoys were disarmed in the same sweep: a Philip Welty of Hopkins’ New York militia appears in both the service and the pension indexes as one man, and he drew a survivor’s pension — something a man buried in Stark County in 1848 could not have done.

The state grave‑registration card for the same Philip, paywalled and unread for weeks, opened: died 11 September 1848, born 1789, Welty Cemetery, Sugar Creek Township, Stark County — Lot 21, Grave 3, upright marker. It carries the War of 1812 service as a bare claim, Private, and then the line that says as much as any of it: “Regt. unknown.” The 1930s compiler who filled the card in could not name the unit either.

The last leg was a clean sweep of both federal 1812 indexes — service and pension applications, every spelling on the ladder. The result is a negative worth as much as a find: apart from Philip and the Welday cousins, no federal service or pension record exists for any of Michael’s other sons — not Jacob, George, Michael Jr, Samuel, Daniel, nor Henry — under Welty, Weldy or Welday. One card stays open on the table: a John Welty of Bay’s 3rd Ohio Militia, no rank, no residence, unidentified.

Her sister’s turn went the other way. Every baptism from 1713 to August 1716 was read, entry by entry, and there is no Wäldi child but Johannes in any of them — so Anna Barbara’s “b. 1714” is derivative, back‑calculated from her confirmation age rather than read off a page, and her real baptism sits later in the book than anyone had looked. Two decoys were guarded on the way past (a Hans Jacob Helm in 1714, a Hans Jacob Noll in 1716 — neither a Wäldi), and two Croissant daughters were dated as bycatch, which matters because the same stretch of images is where a Croissant baptism has been wanted for months. One pass now serves both hunts.

Weidenthal — a Reformed parish twenty kilometres over the mountains into the Pfälzerwald, serving a dozen forest hamlets, and the remotest place a couple from the Edenkoben plain would plausibly have married — was swept end to end: the single October 1724 entry and then every marriage from 1737 to January 1759. Groom, bride, and both named fathers on each. Zero. With that, every accessible Reformed marriage register in the ring around Edenkoben is negative for the years either side of the 1750 crossing. The same day, the earliest Edenkoben baptism book was opened at its 1691 war‑year pages — sparse, half‑blank openings from the Nine Years’ War — looking for the baptism of the man at the trunk. Not there, and most likely destroyed with the years around it, so his birth stays an inference drawn from his own children’s dates.

A roster row had been fusing two men: a real son of the Dover household, born about 1823, and a Jacob Welty born 16 January 1817 whose death, burial and marriage had drifted onto him from a lookalike. The marriage register image settles it. The man who married Barbara Miller in September 1839 is the 1817 Jacob — son of the Tuscarawas Mennonite bishop Abraham Welty, of the Manheim‑and‑Ohio line, living in Dover Township, which is the whole trap: a man of one Welty family keeping house in another family’s township. The arithmetic agrees — he was twenty‑two at the wedding, while the other Jacob was sixteen and had been a ward of the court a decade earlier. The borrowed dates come off the row.

The Canton newspaper was swept systematically for the first time, every spelling, 1815 through the 1830s, and a business cluster came out of it that no one had been looking for. A distiller advertising in September 1816 gives his address as “in Pike… two miles south of Welty’s” — a Welty place‑name already functioning as a landmark. In April 1817 a storekeeper’s last notice calls in debts owed to “my stores, either in Canton or the one lately at Welty’s Mill”: a mill named for a Welty, prominent enough to host a store, and shut by that spring. In 1819 a Jacob Welty leaves his vendue notes and accounts with a collector and disappears from the paper — the paper trail of a man leaving the county. And in the county’s own expense list for January 1819, printed three weeks running, a Philip Weldy draws a chain‑carrier’s fee for survey work: the earliest civic record of a Philip anywhere in the Canton orbit. Which Welty family ran that mill is unassigned, deliberately — it is not being hung on a line until a deed or a tax duplicate says so.

The immigrant of 1750 married somewhere, and not at Edenkoben — that register has been negative for years. The working logic says a Reformed groom marries in his bride’s church, so the target became every Reformed marriage register within reach of the Edenkoben plain, read for the years either side of his August 1750 crossing. Two days took four of them. Hassloch, the largest Reformed parish left in the ring, 1739–1750: zero. Gommersheim, 1744–1752: zero — and the entry standing three days after his Philadelphia qualification, the one that would have been perfect, is a Sieber. Böbingen with its Duttweiler filial, two ring parishes in a single book, 1736–1750: fifteen marriages in fifteen years and not a Welty among them — though two grooms named Philipp Jacob turned up in 1736 alone, a Fuchs and a Frey, both read at full zoom and both firewalled before they could become somebody’s ancestor. Meckenheim, 1739–1748: zero, in a register the pastor himself had to reconstruct after the old one was lost.

Two of those books are physically misbound — leaves out of folio order, marriages spliced among burials — so the sweeps had to be navigated by the year headers written on the pages rather than by image number, and that trap is now written down where the next person to open them will find it. A fifth parish, Altdorf, closed itself: nothing survives there before 1759, and its Protestants of the 1740s appear in the Böbingen book anyway, which the same pass had just read.

An old typescript names an Ensign Frederick Spaar in the same York County company as the Michael of this line — a claim that has never been tested against a record. Spaar left a federal pension file after all: rejected, twenty‑one pages, read cover to cover. Bound into it is his own militia officers’ return, certified at Harrisburg in 1834 as a true extract from the state files, and it names the unit outright: Ensign, 4th Company, 1st Battalion, York County Militia, under Capt. Christian Lowman, commissioned September 1777 and still serving at the April 1778 return. A neighbour’s deposition then walks him out of York County to western Virginia about 1809, which is why he vanishes from Pennsylvania records. The other four men named beside him left no federal file at all.

The same session put the question to the federal catalogue directly: is there any Revolutionary War record for Michael Welty — a compiled service record, a militia return, anything — that the subscription sites have not been showing? Every spelling, forward and reversed. The hits are a general store on the historic register, a land‑patent index, two censuses and a stack of twentieth‑century files; the variants return nothing. And the negative holds up, because the catalogue demonstrably does index these records by name — a Welty of a New York regiment surfaces from the same search, filed at exactly the level a Michael would be. So the absence is real and not a gap in coverage. Michael’s entire Revolutionary footprint remains two Pennsylvania state pay records — a militia pay certificate and a depreciation‑pay roll. One untried test is left, and it is a paid one: a typescript roster in Washington that reportedly names Michael and a Jacob in one company, which can now be checked against Capt. Lowman’s 4th Company by name.

For weeks the Place‑by‑Place timeline did something the rest of the site had quietly outgrown: it marched a single household — the Edenkoben line — from the Palatinate to Michigan, while every other family sat in a static drawer at the foot of the page, outside of time altogether. That split undercut the page’s own argument, which is that these unrelated families kept piling up in the same few square miles. So it was rebuilt into one braided narrative: every family now enters at the moment it actually appears in the record, and the eras where several arrive at once — colonial York, frontier Ohio — read as the pile‑ups they were, not as a footnote. The page now opens where the paper trail does, on Anthoni Wälti, christened at Lenk in 1568 — the oldest dated person anywhere on the site. Two new chapters gather the southern and valley lines and close the page on the DNA, tying each family to its own kit and haplogroup.

The all‑families tree used to load with its top two generations already unfolded across every branch — a wall of names before you had chosen anything to look at. It now opens fully collapsed: just each family’s name and a line of description, quiet until you pick one. Click a family to meet its founders; click a founder to walk down the generations. The “Key” legend and the “how to read the sources” panel came off the top of the page in the same pass — the proof and DNA badges still sit on every individual name, where they earn their keep, but now the reading starts with the families themselves instead of the instructions.

The timeline and people‑map pages had accumulated a layer of scaffolding around their two interactive maps — an intro paragraph, a row of stat chips, an eleven‑colour legend, a proof key, a quotation banner at the foot of the page. On inspection the maps explain themselves the moment you touch them, so all of it came off across a few careful passes, leaving one clean line: a hint, the German map, a hint, the American map. Every dot still scrolls you to that family’s fullest paragraph in the timeline. 380 people placed, none dropped.

One group on the tree had been carrying its deep origin as its name — “Swiss” — which hung a sixteenth‑century Emmental valley over a family whose real American story runs through Manheim Township in York County and out to Fairfield and Putnam Counties in Ohio (the branch that eventually gave the world the writer Eudora Welty). The public label now names those American places — where the records are, and where a descendant would recognise their own people — while the Swiss I2b ancestry stays on the page as ancestry, not as the family’s name. Manheim with one n, the York County township — not Mannheim in Baden, the different place the sources love to confuse it with.

The front page used to open on a number — “Eight Welty families” — and the number would not hold still: as lines were added, split, and relabelled, eight and nine were already contradicting each other across the site. The count was never the point, so it is gone. The hero, the tree’s introduction, and the page metadata now say “a cluster of” and let the family headers and filter chips do the actual enumerating. The tallies that are true — the count of documented people, the running lead and backup figures — stay, because those are honesty rather than branding. A family tree should not have to amend a headline number every time the evidence moves an inch.

For a week the single best shot at a stated Pennsylvania birthplace for the Cross Creek Welday brothers has been Henry Welday’s own 1871 pension declaration — found last session, unread. Today the file was read cover to cover, all twenty‑seven pages eyes‑on. The answer is a clean miss: the 1871 form has no birthplace field, and nothing else in the file states one. What the file gives instead is the man himself. His own narrative of the war: drafted at Smithfield, Jefferson County, into a rifle company; marched Cleveland to Fort Meigs to Detroit to Mackinac Island, where they “engaged with the enemy and were defeated” — the failed assault of August 1814 — and discharged at Chagrin Harbor. His age, seventy‑nine in March 1871, squares with the grave card’s 4 August 1791. And bound into the file sits a prize the county books had only whispered: a certified copy of his marriage record — Nancy Smith, 12 December 1816, Jefferson County, Ohio — image‑verified, upgrading a date that had stood at index grade. Nancy’s widowhood dates her end too: dropped from the rolls, 4 February 1883. The birthplace question survives Henry with exactly one documentary shot left — his two bounty‑land application files at the National Archives, unread, offline, orderable.

The other leg of the day went at the York Pennsylvania Herald, 1789–1798 — the paper that should hold Philip Jacob’s death or estate notice, if one was ever printed. The machine search says zero, so the approach went sideways: targeted probes, every hit read eyes‑on. “Late of Dover” returned fourteen hits and all fourteen are now identified — a 1792 estate (George March’s, with its executors) and the Orphans‑Court dockets of 1798 — which proves the point that matters: the paper demonstrably carried exactly this class of notice. A Welty estate would have had a channel. One trap was disarmed on the way: an April 1792 vendue advertisement under the will of “Philip Jacob, dec., of Paradise Township” — a man whose surname was Jacob. Any index hit for a “Philip Jacob will, York, 1792” is this man, and he is now firewalled on the log. And the bycatch was worth the trip: the Carlisle Gazette of September 1793 shows a “Welty” sitting as a member of the Pennsylvania House that session — an unsorted man, awaiting a name from the member rolls. What remains is the honest slog: an eyes‑on browse of the 1789–91 legal columns, some four hundred pages — until then the Herald negative stands at OCR grade only.

The first systematic War of 1812 sweep — service indexes, pension indexes, the files themselves — put paper under old assumptions and pulled some away. Philip (b. 1789) has his Ohio grave card saying it plainly: “War Served In: War of 1812” — his first official soldier’s standing, unit still to name. On the Cross Creek side, Jacob Welday Jr.’s pension file was read — Steubenville to Canton to Lower Sandusky in the panic winter of 1813, then fifty years of documented Jefferson County residence — and a third serving brother, Abram, surfaced in the cards. The negatives hold their own weight: no service card for any Welty Jacob, George or Michael; Sarah never applied for her bounty land; and the “John M. Welty” whose signature haunts dozens of files is a Pension Office clerk, not kin.

A long‑standing hypothesis for Elizabeth’s missing marriage — New York — was closed with an index built for exactly this: the colonial province’s complete marriage‑license index, 1664–1783, which double‑lists both parties. The search tool was validated first (703 Smiths), then the Wel‑ pages were read eyes‑on: Welch, an unrelated Weldrige, Welles, Wells, Welsh — no Welty‑plausible bride or groom in a hundred and nineteen years. Elizabeth was not married under a New York license. One route in one colony — but a whole shelf of “maybe New York” is now off the table.

At the German end, a door everyone had walked past was opened at last. The Edenkoben Lutheran marriage register, 1735–1756, had been dismissed years‑deep on a confessional assumption — a Reformed family wouldn’t be in it — without anyone reading it. Read now, entry by entry in the old Kurrent hand: zero Wäldi. And the zero means something, because the same pastor demonstrably recorded cross‑confessional weddings — one couple married “in der reformirten Kirch” lands in his book anyway. The mixed‑confession escape hatch is closed. Then the first parish of the surrounding ring: Neustadt an der Weinstraße, Reformed, 1738–1751 — chosen first because the family’s own baptisms twice name Neustadt godparents — also zero. Two more “Philipp Jacob” decoys were firewalled on the way, and a misbound 1761 folio was caught sitting inside the 1746 run. The point of the ring: under the old custom the wedding happens in the bride’s parish — so an Edenkoben zero is exactly what a real marriage elsewhere would look like. The front is officially re‑opened, parish by parish, against a clock: the archive pass lapses at the end of July.

The same register tried to hand over more, and some of it was refused. Where the book says Samuel married a Sophia Lower in 1832 and the censuses say his wife was Susannah, the conflict is logged as a conflict — not overwritten in either direction. And the book’s twenty‑five unassigned Welty‑variant marriages, 1821–1856 — prime candidates for the daughters of the next generation — go on the log as a lead with a standing rule attached: they get assigned by census and probate, never by name‑match. One flat negative earns its row too: Barbara — no marriage, any spelling, county or statewide.

An early family email listed four “Welty landmarks” that have floated through every conversation since; all four are now formally sorted. The Hicksville doctor: real kin, long since proven — grandson of Henry (b. 1790). The Wheeling distillery (“C. Welty & Bro.”): a Baden‑born Roman Catholic family arrived about the 1850s — it cannot descend from any colonial Welty line, and is firewalled. Welty’s Mill at Waynesboro: still not an Edenkoben line, and the founder’s given name is a genuine open‑web negative — every published sentence echoes one source; the 1798 federal tax, which enumerates mills by owner, is the lookup if it’s ever needed. And Welty’s General Store of Dubois, Wyoming: honestly unsorted — founded 1889 by a Texas‑born son of a frontier doctor who drowned in the Horse Creek flood of 1919, and whose parentage rests on one orderable document: his Wyoming death certificate. Three closed, one parked with its next step named — and none of them left to soak up another session’s attention.

The Welday question — whether the Cross Creek settler was Michael’s brother — has always rested on a twelve‑marker whisper. The FTDNA Welty project’s public results table allows something better: reading genetic distance as topology. Against the Edenkoben‑anchored kit, the Welday‑line kit #346970 sits at a genetic distance of 1 in 25 markers — measurably tighter than the documented uncle‑line kits, which sit at 3 to 5. On paper pedigrees of similar depth, that ordering is what brotherhood would look like — corroboration, not proof, and the log grades it exactly so. The same table surrendered a mystery: an unlabeled kit at genetic distance zero, pedigree unknown — now its own unsorted lead. The pending Big‑Y remains the real arbiter; until it lands, the yardstick is what we have, and it points the right way.

A survey of every surviving Welty family tradition — the 1897 letter, the compiled heirloom chart, a 1902 physician’s autobiography, the old typed genealogy — found they all agree on one thing, and it is wrong: each names a mythical immigrant “Michael Welty Sr.,” and not one of them names Philip Jacob. Grandfather Michael swallowed his own father in every retelling — the same telescoping, three independent witnesses. The consequence is a program: if no memory preserves Philip Jacob, only a contemporaneous register — a family Bible — still could. The hunt opened with its biggest index: the DAR’s Bible‑record files, eighty‑four thousand records, swept under every spelling to a graded negative. The negative is the map’s first edge; the county Bible files and manuscript collections are next.

The open web keeps offering two ready‑made “matches” for Michael’s sons, and both are now formally firewalled. The Jacob Welty of Sandyville who married an Elizabeth Holm belongs to the Manchester I1 family — not the Edenkoben Jacob (b. 1788), whose Elizabeth was a Butt. The John Welty of Pickaway County who married Barbara Wissler is not the Edenkoben John (b. 1786), whose wife was Hannah Allman. Same names, same state, same decade — different families. The tree stays clean because the log now says so in writing.

The morning’s pass through the compiled birth records produced a genuine start: Humphrey’s Chester County volume carries a “Weldy, Maria Elisabetha, circa 1752, father Georg” credited to a Brownback’s Reformed register beginning in 1753 — a book the log had concluded, two weeks ago, does not exist before 1832. If real, it was a live venue for Michael’s missing ~1757 baptism. The source was run to ground the same day: Humphrey’s register is Hinke’s eleven‑page typescript at Franklin & Marshall College — and its baptisms cover only 1753–54 and 1765–67. Michael, born about 1757, falls squarely in the gap. The venue reopened and re‑closed inside a single session — but the Georg entry itself stands, and its likeliest identity is George Wolfgang in the window between his October 1752 landing and his first York record: which would place both Edenkoben brothers’ households in Chester County in the early 1750s.

That first York record is itself new this noon. In the personal register of Jacob Lischy, the circuit‑riding Reformed pastor — a book begun in 1744, twenty years before the Strayer’s register opens — stands Johan Görg Weldy, baptized 14 December 1755, parents Görg & Anna Maria: the earliest church record of George Wolfgang’s household in York County, eighteen months before his 1757 covenant signature, and a son named for his father in the family’s own godfather‑naming custom. The same compilation block repaired three corrupt readings from an older extract — two Anna Maria baptisms restored to their true dates, a phantom “Margaret” retired, and a swapped column in a 1789 entry set right.

The other half of the itinerant‑pastors program came back all zeros — and usefully so. Schlatter kept no private baptismal register: his own journal shows him entering acts into each congregation’s book, and the one private record‑book he did keep — marriages, from his later years — was last seen in an aged schoolmaster’s hands before 1857 and is presumed lost. The Muhlenberg journals, the Hallesche Nachrichten, Daniel Schumacher’s register of 1754–74, the Tulpehocken corpus, and the Moravian registers were swept whole against the full spelling ladder: zero Welty variants, 1750 to 1771. And the Germantown Reformed baptisms, 1753–1780 — the one venue whose pastor bracketed the family’s ~1754 marriage window — came back zero as well. The gap in the family’s church paper remains — but the map of where it cannot be filled is now nearly complete, and every zero is on the log so no one sweeps these books twice.

The Pennsylvania Land Office’s warrant register for York County — all nine manuscript index pages of W surnames, 1733 to 1957 — was read at full resolution: not one Welty warrant, under any spelling, in Dover Township, ever. So Philip’s 110 acres, Michael’s 44, the dead John’s 50 — the whole 1781–86 tax ladder read yesterday — were improved farms held under other men’s earlier warrants, and their boundaries live in the York County deed books, not at the Land Office. The negative redirects the entire Dover‑farm hunt down the deeds route. The county’s only variant‑ladder warrantees, all four of them, are now logged — and one correction made eyes‑on: the Abraham Welty warrant of 1761, long carried as Manchester, sits on the old Adam Simon tract in Shrewsbury Township, down in the Barrens — a southern‑York man, not Manchester stock.

The register’s bycatch paid twice. A Jacob Welty of the Manchester line had the old Ludwig Fritlein tract — 166 acres — resurveyed in his own name on 30 October 1815, “at the request of the owner”: a fixed, mappable anchor for the branch, complete with a chain of named neighbors — Wogan, Day, Lichtenberger, Huber, Keller. Then, hunting Dover deeds, the full‑text index surfaced a will of George Welty of Manchester Township naming sons “Jacob Wilty, Philip Welty, Henry Welty and George Welty” — three spellings of one surname in a single clause — and an estate distribution that pays money to George Wogan and Frederick Day’s estate: the very neighbors on Jacob’s 1815 survey. Will, farm, and account interlock on the same block of ground. Whose estate it is exactly — the old George, or Jacob himself — waits on an eyes‑on read of the pages.

All thirty‑three Welty patents in Ohio were pulled from the federal General Land Office database in one exhaustive pass. John’s 1816 patent on Section 10 — the “John Welty from the President of the United States” line in the old abstract ledger — is confirmed from the federal side; a Michael Welty patent of 1825 sits in the same section as John’s 1824 entry, and can only be Michael Jr., ten years after his father’s death. Michael Sr. himself has no patent at all — consistent with an estate farm that was bought, not granted. And one long‑standing assumption failed its check: Abraham Forney, the man on the other side of the 1818 heirs’ deeds, holds exactly one Ohio patent — in Wayne County, nowhere near the estate. Who patented Michael’s home tracts is now an open question, and Deed Book 3, page 51 — the offline errand — remains the page that decides it.

Across the ocean, the deep‑origin question got its first systematic Swiss survey. In sixteenth‑century Zürich, Welti was still a living given name — a diminutive of Walter — so the surname sprang up independently in many places: there is no single “Zürich Welti clan” to descend from. Some 256 Welti and Wälti grooms in the canton’s early marriage books cluster hardest at Adliswil and the Sihl‑valley villages, with the city of Zürich itself a documented blank — no old burgher family of the name at all. The spelling itself splits by region: Welti in the Aargau and Zürich homes, Wälti in the Bern‑Emmental villages that cradle the Swiss I2b line. And one fix landed on the DNA map: the FTDNA project kit anchored to a Hans Heinrich Welti, b. 1586 “Zurzach” belongs to the old county of Baden — not canton Zürich at all. For the Edenkoben apex, the survey’s verdict is measured: Zürich was the single largest Swiss source of the great resettlement into the Palatinate — some ten thousand Swiss by 1720 — so a Zürich origin is geographically plausible, and so far entirely unevidenced. The checks that could evidence it are offline, and they are on the list.

Michael’s Dover years were read straight off the manuscript tax lists, 1781 through 1786, one page at a time. In 1781 a Philip Welty holds 110 acres while two young men — Michael and a brother Jacob — appear as landless single freemen on the very same list (the Jacob the modern index had missed entirely). By about 1782 Michael holds 44 acres; by 1783, that same 44, plus an entry for “John Wilty, deceased — 50 acres,” the farm appearing only in death. By 1786 Michael is landless again — most likely a farm sold in preparation for the move west that lands him in Donegal by 1798. And the surname bends down the years in the clerks’ own hands — Wilty, Welty, Watty, Witty, four spellings of one family across five lists, with “Watty” and “Wilty” now added to the permanent sweep.

A second pass through the lineage‑society files put more flesh on the John Welty b. 1760 of Cumberland Township — the Dover‑farm patriot untangled two days ago. A DAR application fixes his death to 30 December 1837, names his wife plainly as Catherine Weaver (independent of the Weaver land corroboration found last week), and — new to the log — gives him a second documented son beside Henry: a Jacob, who married Sophia Walter. His estate papers wait at Gettysburg, on the Adams County side of the 1800 line; that is the next leg.

The trail of the Donegal tax list’s mysterious “Welty John Junr” (found two days ago, hiding index‑blind at entry #205) ran straight into a primary record: a deed of 30 December 1796, in which John Harbaugh of Bullskin Township, Fayette County, conveys a tract bounded “by lands of John Welty.” So a John Welty held Bullskin land a full two years before the 1798 Donegal entry, one township over — and his 1800 census household (a couple in their prime, four children all under ten) reads like a family begun about 1790. By 1810, no Welty of any spelling remains in Fayette County; where he went next is an open door.

Then a typescript in the Ohio Genealogical Society’s First Families application files laid the whole puzzle out in one sweep. It walks through the family of John Welty Sr. of Manheim Township, York County — the Mennonite elder of Bair’s Meeting House near Hanover, whose 1793 will the log already holds at image grade — and disperses his children westward: son Peter buys the Hunsaker homestead in Springhill Township, Fayette County, in September 1796; the youngest, Joseph, is beside him there by 1800; and John — the eleventh child — “and family lived in Bullskin Township in Fayette County in 1800.” The father’s estate had just distributed a £1,532 balance in June 1796 — the exact funding window for those land entries. And the old will styles the father “John Welty Senr” — which would make his son, wherever he sat, a John Junr. The three “Welly” households of Fayette County, a riddle on the log since the first census sweeps, now have candidate names, parents, and a reason to be there. A typescript is a lead, not a proof — but its child list matches the will’s heir clause one for one, in order.

That one‑for‑one match untangled the will itself. A knotted clause had been read as giving daughter Eve a husband named Peter Weldy — an oddity flagged on the log as a decoy risk. The typescript’s numbering shows the simpler truth: Peter is the second child — the same Peter who turns up in Springhill Township — and Eve’s husband was a Golding. Above the whole generation, the typescript adds a name the log has never held: the father of the two Manheim patriarchs was a Peter Welty, whose sons Abraham and John married two Cocghnower sisters. Every claim of it is derivative until a register or deed says otherwise — but the elder generation of this family now has a shape, and legs to verify.

The same typescript carries a challenge. It makes Jacob Welday of Cross Creek — the Jefferson County settler whose line the log has long weighed as a possible brother of Michael — the tenth child of the Manheim John Sr., born about 1770, “Welday” being a spelling he adopted in Ohio. Its list of his ten children meshes, name for name and date for date, with the Welday family the log has already documented from deeds and estates — so the typescript is unquestionably talking about the right Jacob. Which leaves exactly one question: whose son was he? A 1938 pension‑inquiry letter already whispered “father = John.” The DNA hint that started the brotherhood theory rests on twelve markers. One proper Y‑test on a Welday‑line male now decides it cleanly: one haplogroup makes him Michael’s brother, the other makes him Mennonite Manheim stock — and either answer settles a corner of the tree.

The Jacob Welty who let his mail pile up at the Canton post office in 1834 now has a shape: his 1830 census page, read eyes‑on, shows a man and wife both born in the 1780s with five children born 1810–1825 — and a sweep of the 1820 census finds no Welty of any spelling in Stark County at all. He arrived in the 1820s, and by 1840 he was gone: a decade’s residence, bracketed by absences. He is provably not Michael’s son Jacob, who kept his own household in Tuscarawas the same census night. Who he was stays honestly open on the log — as does the John Welty of the 1835 letter list, who matches no Stark County household in either census.

And the John Welty b. 1760 of the old SAR application — the Dover‑farm patriot whose parentage graft the log untangled two days ago — turns out to have the quietest biography of all: the same Cumberland Township farm in the 1798 tax, and the censuses of 1800, 1810, 1820, and 1830, until his death in 1837. What moved was the county line: Cumberland Township left York for the new Adams County in 1800. The practical yield: his 1837 estate papers — the next leg — will be filed at Gettysburg, where a York‑only search would have come back a false negative. His 1789 marriage, though, is not where it should be: the Trinity York register’s Welty weddings stop in 1785, so the hunt moves to the Lutheran and Adams‑side registers.

The September 1815 inventory of Michael’s own goods — long held only at transcription grade — was read eyes‑on in the administration volume, appraiser Conrad Roth’s name plain on the page. Beside it, a negative worth nearly as much: the county’s typed index to probate files, read entry by entry through the W’s, shows no Michael Welty under any spelling. His 1815 administration never entered the numbered‑file series at all, so there is no loose estate file left to order. What survives of his probate is what the log already holds — plus one last untried vein, the pre‑1817 Common Pleas journals, readable only on film at a FamilySearch center.

A record class the project had never once touched: newspapers. The first sweep of the Library of Congress’s digitized pages found the German‑language Vaterlandsfreund of Canton, Ohio — printed in Fraktur type that defeats OCR, so the pages were read eyes‑on. In the Canton post office’s unclaimed‑letters lists of 1834–35 stand a Jacob Welty and a John Welty — and, nearby, a Jacob Routhrauff, the in‑law surname. Which Jacob and which John is deliberately left open on the log while the candidates are sorted against the census. The same day closed one hope for good: the Tuscarawas Chronicle, the local paper that should have printed the 1828 partition notice enumerating Michael’s heirs, was never digitized — and no 1828–29 issue survives in any catalogued collection. That notice is not late; it is gone. The court record stays primary.

The Sons of the American Revolution membership files carry exactly twelve applications with a Welty or Weldy rung, and all twelve are now triaged. Two findings stand out. The parentage that many online trees hand the patriot John Welty (b. 1760) — son of a John Jacob and a wife “Braiff” — traces to a single 1908 application and no further back; the 1921 application through the same patriot names no parents at all, but carries a state‑library certification of his actual militia units — a unit‑level firewall against ever confusing his service with Michael’s. And Michael himself is a clean negative in both DAR and SAR: in over a century of lineage filings, no descendant has ever filed on him. One more paper in the margin: an Ohio grave‑registration card gives Philip (b. 1789) his first official standing as a soldier — Private, buried at the Welty Cemetery.

The Manchester branch — the I1 fatherline, a separate people from the Edenkoben line — got a full day of its own. In the morning, the nine children of Philip (b. 1780) and Elizabeth Knaub were worked through one by one; in the evening, an old knot untied. George (d. 1819) left four sons by two wives, and which boy belonged to which mother had long resisted sorting. Now it is sorted: Jacob (1803) and Philip (1805) by Magdalena Rupert; Henry (1809) and George (1810) by Susanna Bethman, married November 1807. The youngest, George b. 1810, was then walked end to end — his 1850 Manchester household, his 1860 move north to Fairview, his death in January 1864, and an estate administered by John Lichtenberger, the man who in 1850 had been his next‑door neighbor. And a widow’s petition of 1837 put a number on Philip’s family: nine children at his death, two of them under seven — a figure that quietly implies at least one child the record hasn’t yet named.

The quiet days ended where Michael Welty’s story ends — with the land he left behind on Crooked Run. An 1837–38 Abstract of Titles ledger for Tuscarawas County, read at full zoom, lays out the recorded title chain of his estate farm instrument by instrument, each with its deed book volume and page: the 1818 conveyances between the Welty heirs and Abraham Forney for the two home tracts, and the chain’s later links into the 1830s. The ledger is an index, not the deeds themselves — the pre-1830 deed books aren’t yet readable online — but for the first time we know exactly which pages to pull, and the first of them (Deed Book 3, page 51) may name the heirs one by one.

The final decree that closed Michael’s estate had been known only as a bare line — “decree, 1829.” Today the recorded case itself was read at zoom: the commissioners report that the farm “cannot be divided without prejudice to or spoiling the whole” and appraise it — subject to the widow’s dower — at $465 for the two parcels. Then a human detail the summary never carried: six heirs were paid their shares directly, but Henry (b. 1790, by then away in Fayette County, Pennsylvania) and David never appeared — their money was left waiting with the clerk of court.

Independent corroboration, from print: the 1884 History of Tuscarawas County records that Abraham Forney “entered a farm on Crooked Run, which he sold to Michael Welty, and in 1814 came to Franklin” — putting Michael’s purchase of the estate farm at 1814 or earlier, from a named seller, in a book written while memories were still alive. The same page adds a surprise: a second Michael Welty tract, the northwest quarter of Section 13, co-owned with Peter Walter — land that never appeared in the estate partition and now needs an explanation of its own.

A recorded 1862 partition case — John Welty v. Christena Welty et al. — turns out to be the probate of Michael’s son Jacob (b. 1788), who died about 1861 still holding his “Horsefield tract” in Dover Township. The petition enumerates his family stem by stem: widow Christena with her dower, and the children and grandchildren named heir by heir. That list matters beyond Jacob himself, because several of these names — Elijah, Wesley, Lucinda among them — have long hung on the tree under his brother John (b. 1786), on the strength of an old family chart. This record says they belong under Jacob instead. The tree has not been moved yet — a faint line or two in the record still wants a higher-zoom re-read first — but the evidence is now on the log, waiting.

Page 4 of the “Welty Black Book” — the hand-copied family register whose page 3 the project already holds — was found attached to a FamilySearch profile, after being believed unavailable. In the register’s own voice it completes John Jacob’s child list with his wife Christina Broff: “my Son John Henry” born 4 November 1764 — the future founder of Weltytown — Frederick in 1767, Margaret in 1770; and then, plainly, Christina’s own birth year, 1730. That date is family-sourced now, independent of a debunked index match this project discarded weeks ago, and it supports the reading that Christina was a younger second wife whose children arrived long after the first set. One caution stays firmly in place: whether this John Jacob is the same man as the Edenkoben-born John Jacob of 1710 remains an open question, and the register alone doesn’t settle it.

Since a week of quiet accumulation is hard to picture, here is the shape of it. The family tree now carries 444 people across many distinct lines, and every one of them wears a grade rather than a bare claim: 38 proven outright, 314 documented, 32 resting only on an index awaiting an image, 24 still openly marked as family lore, and 13 flagged as working hypotheses the record hasn’t yet settled. The direct spine runs seven generations deep — from the Edenkoben roots down to the edge of the living — and the deepest root on the whole chart, on the separate Swiss line, reaches a 1568 ancestor in the Simmental.

Behind that tree sits the master research log it’s all generated from: 21 tabs and roughly 1,480 rows, every find a typed, numbered entry. Two ancestors have soaked up the most ink — the Michael Welty workup at 171 rows and Philip Jacob at 167. The log now holds 32 settled conclusions, 27 leads still open, and 166 leads closed and archived — because a dead end written down is worth as much here as a discovery, and keeps the next session from re-walking it. The whole thing has been copied to a dated safety backup 27 times, and the surname has been swept, every search, across eight spellings — Welty, Weldy, Welti, Welte, Welde, Wildy, Wäldi, Wäldy.

The public side you’re reading has its own tally: 81 record scans pulled down at full resolution, 41 people with a record card of their own, 35 pieces of evidence linked to the pages beneath them, and about 90 family photographs catalogued.

The family’s oldest legend — a brother who settled near Gettysburg — has a real family behind it, and the evening gave that family its own group and then filled it in. This is the Roman‑Catholic Welty clan of Eppingen in Baden, whose immigrant John Welty (b. 1722) crossed in 1751, lived in York County, then settled the Taneytown–Emmitsburg district of Maryland, ten miles south of Gettysburg, and died there at 94 — a documented Revolutionary patriot, and a wholly separate line from the Edenkoben one. Starting from the immigrant and his eight children, the branch was built out generation by generation — grandchildren from an 1834 deed and an 1836 will, a census slice, then the nine children of James M. Welty and Ellen Hobbs, and finally twenty‑five great‑grandchildren harvested from the published record. Sixty people, six generations, including the DNA‑confirmed rung that ties this branch to the author Rita Welty Bourke’s family. The seventh generation, being likely still living, was deliberately left off.

One node on the tree had been carrying a marriage it didn’t deserve. A 1783 York wedding — John Welty and Margaret Ilgenfritz — had been pinned under the Edenkoben Philip Jacob on nothing more than proximity: a John, in Dover, in the same church and decade. But this John’s line tests R1a, which cannot share a father with the Edenkoben R1b line, so the two can’t both be true. Doing the work rather than leaning on assumption, the evening established him for what he actually was: John (Nicolaus) Welty Sr., a German‑born Continental Army soldier of the German Regiment, a documented Revolutionary patriot (DAR A203144, pension S39882) who married Margaret, followed the Great Wagon Road down into Virginia and Tennessee, and founded a line of his own. It even corrected the log’s own error — the “George” and “Peter” once read as his forebears are in fact his son and grandson. He is now his own group, and the Edenkoben Philip Jacob is freed of a marriage that was never his.

The evening closed by seating two further families, each honestly qualified. The Goochland County, Virginia Weldys trace on paper to a Revolutionary soldier, George Weldy — but the two DNA kits that should agree don’t quite, a discrepancy that hints at a break somewhere in one branch. Rather than smooth it over, the tree seats the family with the discrepancy flagged, not silently merged. And the Saanen Wälti — a Swiss clan of haplogroup G, a genuinely separate people unrelated to any other Welty line on the chart — got its spine from a 1568 ancestor in the Simmental down to the immigrant who settled at Upper Sandusky, Ohio, with the deep Swiss rungs marked plainly as parish lore. Eleven people each, both clearly labelled as their own thing.

The son who carried the line into Ohio began the day with no page of his own and ended it with five. His baptism at Center Lutheran Church, Somerset County, on 2 January 1791 — which newly fixes Michael’s household in Somerset in 1790–91, between York and the move to Ohio — names Michael and Christina as his parents, the plain proof of the father-to-son join. Beside it now hang the family’s own Bible-style register, one page that records his 1812 marriage to Mary Byerly and the births of all nine of their children; his place on the Connellsville town council across the 1810s and 1820s; the appraisement of his estate at his death in 1842, with his widow sworn as executrix; and the partition suit that, eight years on, names his heirs one by one. A whole life, from the font to the division of the estate, in a single afternoon’s gathering.

With the spine imaged, the afternoon turned to breadth — the generation of Michael’s grandchildren who went west to Indiana, until now documented but deliberately left off the tree pending proof. That proof having landed, the households opened all at once: Samuel’s twelve children and Daniel’s ten, read straight off their 1850 and 1860 Owen County census pages, and then Michael Jr.’s eight in Adams County. Thirty new people in one afternoon; the tree grew from 314 names to 344. Where an age disagreed between one census and the next, or a name sat only on an index and not yet on a manuscript, the row says so plainly rather than smoothing it over.

Not every family that spells the name this way belongs on the tree, and part of the afternoon was spent proving one of them off it. Three “Weldy” men in an 1887 Adams County biographical volume had looked, from a distance, like Michael Jr.’s people. Read closely, they are the Fairfield County, Ohio Weldy family — a separate, Swiss-derived line the log had already flagged as a trap, sharing the spelling and nothing else. There is no Michael Welty among them at all. Recording the mismatch is what keeps the near-namesakes from drifting quietly into the family and blurring it.

A long-standing puzzle on a collateral line — why Jacob Welday, Sr. left no probate estate to trace — resolved itself in a single 1824 deed. He had not died intestate and unrecorded; he had, while living, deeded himself into his son Isaac’s care in exchange for maintenance, with his wife Mary named first, another son a witness, and his daughters accounted for. The absence of an estate was never a gap in the record — it was the record, doing exactly what it should. The keystone that explains the silence, and redirects the search from probate to the deed books.

One page in this pass had to be taken down and re-hung. A September-1709 Edenkoben marriage had been pinned to John Jacob (born 1710) as though it recorded his bride’s father — but read slowly at the page, it is nothing of the sort. The entry is Hans Jacob Wäldi’s own marriage, the miller of Durchsteinbach, “surviving son of the late Johann Georg Wäldi, citizen of Bischheim.” So a single record does double duty: it is the father’s marriage and the only life-mention the grandfather ever earned — and it plainly cannot belong to a man not yet born. It now hangs where it belongs. The quiet kind of fix that a hurried pass would have carried forward as fact.

The immigrant’s own life is now imaged from one end to the other. To his 1784 marriage and his Revolutionary depreciation pay, this pass added the inventory of his estate, taken at his death in September 1815 in Tuscarawas County, Ohio — a public probate page, read and set on the tree. Around it came the marriages of three of his daughters: Elizabeth to George Gauf, Christina to Peter Messerle, and Catharine to Jacob Boehm — the last of these the real double Welty wedding of 10 May 1785, which is very likely where that stray “10 May” once wandered onto Michael’s own 1784 marriage date and had to be chased back off.

A quieter kind of progress, and a visible one. For the first four generations of the family, the actual documents behind the names have been gathered and pinned to the tree itself: open a person and, where one survives, the original page now appears beside them — Philip Jacob’s 1719 baptism in the Edenkoben register, his brothers John Jacob’s and Georg Wolfgang’s, and a growing set besides. Around twenty images across fourteen people so far. The aim is simple: the tree should not merely assert a fact but hand you the paper it rests on.

Reading Daniel Welty’s 1850 household in Owen County, Indiana at first hand turned up more than expected. Daniel — a proven son of Michael — is there with his wife Louisa and their children, and a small correction rides along: the census calls him a house carpenter, not the farmer an earlier compilation had made him. But the striking thing sits in the very next dwelling. A Samuel Welty, about thirty-nine, born in Pennsylvania, keeps house there with his wife Susannah and eight Ohio-born children. A Samuel is exactly the son the family’s memory attaches to Michael yet never documents as a grown man — and here one stands, Pennsylvania-born, the right age, settled next door to a proven brother along the same Ohio-to-Indiana road. It is a strong case, and it is written down as exactly that: strong, but not yet proof. No record welds this Samuel to Michael, and a Samuel-and-Sophia marriage in the right Ohio township carries the wrong wife’s name. So he is logged as a promising candidate, his eight children with him, and the record that would settle it — an Ohio probate, an 1840 census, a second marriage — named as the next thing to find.

A loose thread worth pulling, and worth cutting. A twenty-six-page Revolutionary War pension file for a “John Welty” kept surfacing in searches for this family — but read through, it belongs to a John Welty of Maryland (survivor pension S.39882), whose file names a comrade from a German regiment raised in Maryland and Pennsylvania. He is not the York County John, not the Dover John, not the Swiss John. The reason it kept turning up at all is a small modern artifact: a later page carries reader-added notes — “Jacob Weldy,” “George Weldy” — tags someone attached online, not anything the record itself says. Filed as a decoy, so the next searcher need not chase it twice.

Last night’s note ended on a promise: that the easy, open web was nearly worked out, and the next real gains sat behind logins and in courthouse films — the Ohio wills and marriage books, a Civil War service card, the headstone pages that kept refusing to open. This afternoon, roughly four o’clock, that was exactly the ground worked — one signed-in session, one record pulled at a time — and nearly every door that had been shut the night before swung open.

The Weltytown grandson lost in the war — William Bierer Welty — had been carried on the strength of a county history; today his own service record was read. The Pennsylvania Civil War card file, reached through his memorial, gives the plain facts: a resident of Latrobe, he enlisted there on 21 August 1861, mustered in as a private in Company D of the 11th Pennsylvania Infantry, and was killed at Antietam on 17 September 1862. With the signed-in grave pages open, the household around him fell into place as well: his father Daniel Welty (b. 1806) and Barbara Bierer with all ten of their children, each with dates and married names; and a neighboring branch — John Welty (b. 1793) and his wife Rebecca Lose, her name recovered at last — whose daughter Elizabeth, only guessed at last night, is now confirmed. One further thread runs bright through this line: a grandson, Welty McCullogh, who went to Washington as a United States Congressman from Pennsylvania in 1887.

The open question about the Welday branch — where exactly it joins the immigrant family, and which Jacob is its father — sent the search into the Ohio marriage books, and two records did come up: Jacob Welday’s marriage to Polly (Mary) Lockart in 1816 and Elizabeth Welday’s to George Thomas in 1807, both in Jefferson County. They confirm the family was settled there early, which is worth having. But they cannot answer the parentage question, and for a structural reason worth writing down: Ohio married couples by license and return, not by bond, so these registers name only the groom, the bride and the date — there are no bondsmen or witnesses to trace back to a father. The Ohio marriage books, in other words, simply cannot settle which Jacob this branch descends from; that answer will have to come from Pennsylvania records or from the Welday estates. A door opened, and honestly showed there was nothing behind it.

The grave pages that sat behind a security wall all last evening opened on the first try once signed in. In the Welty Cemetery at Brewster, Sugar Creek Township, the memorials for Philip Welty (d. 1848) and his wife Sarah were read, along with their son Isaac’s, which lists both as his parents. They add no new fact to what the cemetery transcription already gave — but they do something quietly satisfying: Philip’s stone records his age at death as 59 years, 4 months, 9 days, which counts back to exactly 2 May 1789 — the very birthdate the old family pedigree carried, now independently confirmed by the stone.

A bit of upkeep, not a discovery: the site moved to free, static hosting served straight from its open-source repository, so it can’t be paused or walled off and every version is preserved in public. The research doesn’t change — this just makes sure the record stays reachable, for good.

The depreciation-pay roll named a few nights ago has now been read at the page itself, at full zoom. Under the heading “Depreciation Pay,” page 481, three names stand in a row — John, Michael and Philip Welty — each a private drawing back-pay for Revolutionary service. That much is now confirmed beyond the index. What the page does not do is name a county: the printed genealogy that filed these men under the “York County militia” put a label on them the manuscript never carries. So the service stands as first-hand fact, the county tag is set back down as unproven, and the three names sitting together are not assumed to be one household. A clean line between what the record says and what a later book added.

The 1850 and 1860 censuses of Tuscarawas County, Ohio were pulled and read image by image, and they put two families in order. In Dover Township, Jacob Welty (born about 1788) and his wife Christina kept house with a son, Finley — and the same couple, the same Finley, appear together again ten years on. That settles a point the notes had gotten wrong: Finley had been filed under the wrong man; the two censuses running back to back show him plainly as Jacob’s son, and the earlier reading is corrected. A short way off, George Welty (born about 1793) and his wife Susanna sat with their children Sarah, George and John — two of whom the family had carried only as names in old family lore, now found living in their father’s household in the federal count.

Among the many children of Henry Welty and Mary Byerly, two daughters gained proper marriage records this pass. Susan Welty married Walter Carr in Stark County, Ohio, in March 1843 — a courthouse register line that anchors her to Stark County and quietly confirms that the Welty sisters set out from the same corner of Ohio before turning up, married, in neighboring Indiana towns a migration later. Lucinda Welty’s marriage to Hosea Stockwell in 1846, until now known only from a memorial, was found in the marriage book itself and upgraded to a record. And one hoped-for match — a Welty–Hoopingarner marriage — was searched both directions across the likely years and simply is not there: a clean, useful nothing.

A smaller correction was begun in this spirit and finished by nightfall. A wife recorded in the tree as Mary Lobengier had been attached to one of two same-era brothers, David and Samuel, and a first pass this morning was not yet sure which. The fuller evidence that evening settled it — a printed county genealogy places her with Samuel — and the corrected reading, together with what it means for his brother David, is written up in the night’s entry above. The quiet kind of fix that keeps a growing tree honest.

The morning’s open question — whether the Ruthrauff bride was Michael Welty’s first wife or his second — now has an answer, and it came from four tax returns read one after another at full zoom. Dover Township assessed Michael as a single man in 1779, again in 1780, and again in 1781, each time on the single-men’s list with no land of his own; by 1782 he holds 44 acres on the main roll, and the 1783 septennial count carries him at the same 44. He married Christina Ruthrauff in May 1784. A man taxed as a single freeman into 1781 cannot have married anyone in 1754 — so the old pedigree’s “married Elizabeth, 1754” belongs to a different Michael entirely, and Christina stands as the first wife, not the second. The reversal that yesterday was logged only as a hypothesis is now the plain reading of the record. The same volume gave up one more page: the Pennsylvania Archives’ Revolutionary depreciation-pay roll, where three Welty men — John, Michael and Philip — stand on the one page, each marked “private.” The family’s “Lieutenant” tradition finds no support, but three Weltys drawing a soldier’s back-pay for the Revolution do.

Michael’s son Philip (born 1789) carried the line into Stark County, Ohio, and this pass found where he rests. The usual grave pages would not open all evening — they sat behind a security wall — so the answer came instead from a published book of cemetery inscriptions for Stark County, read through its full text. In the Welty Cemetery in Sugar Creek Township — a small ground on Welty Street, a little over a mile west of the highway — stand the stones of Philip and his wife Sarah, worn past legible dates but named plainly, beside their son Isaac. And a 1916 county history supplied an origin the family had never had in writing: Philip, it says, “fought in the War of 1812, received a grant of land for his services, and located it in Sugar Creek” — very likely the very ground the cemetery still occupies. That account is a century-old secondhand sketch rather than a federal record, so the 1812 service is carried as reported, not proven; but it points a clear way toward the land warrant that would settle it. One more name surfaced on the same page — a Susan Welty, born 1823 — noted and set aside until an original stone or register can say whose daughter she was.

A step sideways from the direct line sits the Gauf family — George Gauf married Philip Jacob’s daughter Elizabeth — and its records had gone stubbornly quiet: search after search for the elder Philip Gauf’s 1777 will returned nothing. This pass found why. In three York County deeds the family is written “Philip Gauff, by the name of Philip Kauff” — an alias no index search under “Gauf” would ever surface. With the right spelling the deeds opened: one recites a son’s will — “Philip the younger,” holder of the 213-acre home plantation — leaving a yearly annuity “unto my mother Christina, as my father’s will says,” which fixes him firmly as the elder man’s son. The elder’s own will is still not among the digitized books, so the question that began the hunt — does it name a son George? — waits on the will book itself; but a family that kept slipping out of the record now has a second name to be searched under.

Christina Welty’s mother was Barbara Ruthrauff of Newberry Township, and her estate came up for settlement in 1802. The hope going in was that Michael Welty might appear on it — as administrator or bondsman — which would have been the first legal document to tie him directly to his wife’s family. The record was pulled and read: administration went to a Jacob Miller, and no Welty appears anywhere on the account. It is a negative, but a clean and useful one — it corroborates the Ruthrauff parentage the church register already gave, and it closes a door so no one need open it again. Who Jacob Miller was to the Ruthrauffs is a new small question, set down for later.

Further back, at the immigrant generation, John Jacob Welty’s wife is given across the internet three different maiden names — Broff, Braff, and “Bruss of Solingen,” the last stated with confident detail: a christening at Wald, near Solingen, in 1730. Run to ground, the whole structure rests on a single index card attached to a family tree in 2014 — a bare first-name-and-date match to an unrelated German infant, with no record of any kind connecting her to Pennsylvania. A sweep of the ship lists and the local church books turned up no Broff, Braff or Bruss family at all. The name the family has carried, in other words, is unproven, and the tidy “Solingen” origin is a modern invention rather than a record. What her name truly was remains open — the likeliest real candidate is an anglicized misreading of a common Lancaster surname — and it is filed just so, precisely so no one builds further on sand.

A small confirmation to close the night’s records. The date of Michael Welty’s death — narrowed a few days ago to “on or before 7 September 1815,” against the pedigree’s unsupported “9 September” — was this time checked against the manuscript itself: the Ohio appraisers’ book, in the original hand, opening “the goods and property of Michael Weldy, dec’d, Sept 1815.” A court cannot appraise a living man’s estate; the date now rests on the page, not on an index.

And a closing word, since the night’s work is done. Thank you — to the family who have been reading these entries as they appear, following the hunt night after night, and caring about people four and five generations back. Writing the search down in the open only means something because there are people on the other end of it who want to know. Your interest is the reason it is worth doing carefully. More soon.

A Ruthrauff household turned up at the family’s own Dover church: Johannes Rohtrauf and his wife An Maria baptized a daughter at Strayer’s in June 1785, one year after Michael Welty married Christina Ruthrauff — and on the same day as baptisms in the Messerle and Lauer families, three households of the kin web at the font together. Following that thread into the 1925 Ruthrauff family genealogy produced its chapter on the marriage: the Ruthrauff who married Michael Welty of New Philadelphia was, by that book’s account, the only daughter of Johannes Ruthrauff and Anna Barbara (Hoffman) — “wedded without consent.” The book guesses her name as Barbara, but the guess traces to an 1834 deed belonging to a later Michael and his wife Barbara; the 1784 church register supplies the name, Christina. The chapter’s own citation then led to a page image: the 1779 Dover Township tax return, and on its single-men list “Michael Welty” — his earliest known record anywhere — sharing the township with Welty Philip at 100 acres, Welty George at 120 (the acreage family tradition always gave him and no deed ever showed; here he is assessed on exactly it), John Rudrauff at 195, George Gauff at 150, and Daniel Messerly. The whole kin web, one list, five years before the wedding. And a quiet consequence: single in 1779, reportedly still single in 1781, married in 1784 — the old family letters’ claim that the Ruthrauff wife was Michael’s second is likely reversed. If Christina was the first wife, the line of descent this project follows runs through her — and through Johannes Ruthrauff and Anna Barbara Hoffman behind her. That reading is logged as a hypothesis; the 1780–83 returns are queued to test it.

The tax film that yielded the 1771 Manchester list has now given up the next chapter. The Dover provincial tax list warranted 8 December 1774, read at full zoom, contains the entire December-1771 Manchester community — the same Quickels, Rudys, Walks and Wolfs — running in one continuous list under Dover: between 1771 and 1774 the family and its neighbors stayed put, and the township label around them changed. In the W-block: Welty Philip, Welty George — and “Walty Widow.” The John who was assessed beside them in 1771 is gone; John died between 11 December 1771 and 8 December 1774, and a widow now heads the household. She persists: the printed 1780 return carries “Welty, Widow,” with 80 acres, landed on a par with George beside her — a six-year widow household the record had never registered. That she is John’s widow is the natural reading, and it is logged as a hypothesis; who John himself was remains open. The same sweep hardened the edges of the 1771 picture: Dover’s own 1771 list survives only A through O — there is no W-section to hold anyone — and a county-wide read of all seven surviving 1771 township lists found Weld-form taxpayers in Manchester and nowhere else. The 1771 “Welty Paul” was re-read at zoom and stands exactly as written, but the 1774–1780 runs show a Wilt/Wild Paul in his place and no Welty Paul anywhere — the entry is now judged a probable clerk’s slip rather than a fourth Welty household.

The Dover home-church register has now been eye-read across its whole early run, 1745–1786, with surname-blind sweeps over all 174 images. The published extracts the project once leaned on did not survive the comparison: of the twelve Welty entries audited against the typescript, only one stands unchanged — the rest carried swapped columns, grafted dates, a child recorded as a wife, sponsors promoted to parents — each now corrected against the page. The page that refused to render two nights ago finally rendered, and its “Georg Gauf & Elisabet” proved to be the known 1782 baptism itself — the daughter whose sponsor was Christina Welty — not an earlier child. With that page read and every candidate venue now searched, the working hypothesis is that the 1782 child was the couple’s firstborn. The elder Gauf household also gave up its children — seven baptisms, 1750–1764, among them Magdalena (1753), the future wife of the orphans’ guardian Jacob Lauer — and no Georg among them, which keeps George’s parentage a hypothesis for the queued will-pull to decide. Oldest of all: the union-church covenant of 30 May 1757, by which the Reformed and Lutheran members agreed to share the land and churchyard — and the first signature on it is Görg Weldy, with Philliph Gauff and Daniel Messerly signing the same page: the earliest church act by anyone of the name in Dover, years before the families intermarried.

A negative worth keeping. The hope that Philip Jacob’s blank years (1754–1771) might surface in French & Indian War supply paper — the army contracted York County wagons in 1758 — has been run to the end of print: all nineteen transcript volumes of the Bouquet papers plus the letterpress edition, swept surname-blind, hold no Welty under any spelling, and the contracting letters name the contractors, never the wagoners — the wagoner name-rolls were simply never printed. The one record series that could still cover those years, the provincial rent rolls (three York County volumes, roughly 1749–1776), exists only on paper in Harrisburg — queued for a reading room, not a browser.

A side session that ran without access to the research log came back with three “new” leads. Checked against the log, two were old acquaintances — one a question answered in full last week, the other a bounty-land warrant already pulled and shown to belong to another man entirely. The method note is filed: memory lives in the log, and leads get checked before they get chased. The third lead survives review, and it is a good one: an Ohio county history reports a “Welty Church” (United Brethren) with its own burying ground in Sugar Creek Township, Stark County — squarely in the orbit of Michael’s son Philip (b. 1789), whose son Isaac lies in the Welty Cemetery at Brewster. Whether Philip himself is buried on that ground is now on the hunt list.

Last night’s suspect for the origin of “married Elizabeth, 1754” has been checked against the manuscript. The marriage is real, and it sits in the same Edenkoben Reformed register series that holds the family’s own baptisms: on 9 January 1754 the pastor married Philipp Jacob Doll to Maria Elisabetha, daughter of Lorentz Neu. The entry was read at full zoom, and the groom’s surname is unambiguous. Just as telling is where the 1754 date doesn’t appear: a sweep of the public trees and databases found it nowhere — the one public tree that names the wife offers a different, equally uncited guess (“about 1751”). The date survives in exactly one private pedigree, and the question for its keeper is now sharp: does a record stand behind it, or an Edenkoben index hit that belongs to another man?

The 1754 marriage got a brute-force search: every published or indexed Pennsylvania register that could cover August 1750 through 1757 was swept surname-blind — every Philip married to every Elisabeth, some 330 index rows, plus full-text reads of the major printed sets and a line-by-line read of the First Reformed Lancaster register. Fourteen candidate couples emerged; none is the right one. The surviving registers have holes right at the window — First Reformed Lancaster’s marriage book starts only in 1752, Chester’s early records are lost, Philadelphia has a gap — so the marriage was likely never recorded, or sits in a book that no longer exists. The sweep did surface one suspect for the tradition itself: an Edenkoben marriage of 9 January 1754 — Philipp Jacob Doll to Maria Elisabetha Neu — just the hit a pedigree-builder searching Edenkoben indexes for “Philip Jacob” would find. The Edenkoben Philip Jacob had been in Pennsylvania since 1750, so it cannot be his, but it is the likely source of the uncited “married Elizabeth, 1754.” One lead survives the negative: a “Jacob Weld” naturalized in Lancaster County in 1761, which under the law of the day required seven years’ residence and a church’s sacramental certificate.

The Elizabeth (Welty) Gauf story gained its last chapter. The Quarter Sessions docket shows that in the same March 1792 court cycle as the guardianship petition, the widow Catharine Gauff was indicted for assault and bound to good behavior toward Elizabeth Gauff, the nine-year-old orphan; witnesses included the court-appointed guardian and Peter Messerle, the child’s uncle by a Welty marriage. Catharine pleaded guilty and was fined five shillings — and her treatment of the child is the likely reason the family went to court. The guardian, Jacob Lauer, had married a Gauff daughter in 1779, making him the orphans’ uncle by marriage. A generation up, Philip Gauf of Dover died testate in 1776/77, his estate handled by Daniel Messerle (Peter’s father), so the Gauf–Messerle bond predates every Welty marriage in the cluster; George fits as his of-age elder son, though that will-pull is queued and the reading stays a hypothesis. The hunt for Elizabeth’s burial closed without a stone: no York-town church kept burial registers across 1783–1790 — Trinity’s book has an eighty-one-year gap in that record class, Christ Lutheran’s earliest burials are three entries from 1801, and county cemetery indexes hold no early Gauf stones. Her burial is unrecoverable; her death window — after November 1782, before about 1790 — rests on the court reconstruction alone. The same night opened two doors: the Strayer’s register shows Jacob Lauer and his Gauff wife baptizing at that Dover church in 1786, and its 1777–1786 pages carry “Georg Gauf & Elisabet” as a couple — parents or sponsors, the page images wouldn’t render to say.

Last week’s two baptisms at Quickel’s church in Michael and Christina’s “silent years” — Catharina (March 1785) and Johannes (March 1786) — had been logged as children who died young, since neither appears in the estate papers. A re-analysis makes a tidier claim: they are likely the first two children on the family’s heirloom chart — Catherine (1785–1863), who married a Forney, and John (1786–1827), who married Hannah Aultman. The birth years match exactly; the chronology leaves no room for same-name doubles (Michael and Christina married in May 1784); and the pair’s absence from the estate records was already explained — John died the year before the land was partitioned, and Catherine appears to have been bought out rather than deeded land. It is logged as a hypothesis, and the tree stays unchanged until the merge is decided. (A bonus from the same pass: the chart’s daughter who “married a Harney” may be the same hard-to-read surname as Barney/Forney — possibly the same Catherine, written twice.)

That Manheim Township man supplied the day’s caution. His estate account — a substantial £1,532 — was settled by an executor “according to the said deceased’s Will,” contradicting an earlier finding that no John Welty will existed in the York index. A full-text search turned up the will in minutes: John Welty Sr. of Manheim, written 21 September 1793, naming his wife Eve and a dozen children — the entire elder generation of the Swiss-line family, in one document. None connect to the Edenkoben line, but the lesson holds: an index sweep can miss what the record plainly contains, and every “no such record” in the log now carries that caveat.

From the Trinity York register browse: baptisms to a “Jacob Welty & Anna Maria” — Johannes (1784), Johann Peter (1786), Henrich (1787), and, from an earlier read, Abraham (1791). Cross-checked against the 1826 estate petition listing the Manchester patriarch’s surviving heirs, the fit is clean: John and Henry appear as heirs with matching birth years, while Peter and Abraham are absent — both died young. One flag: a second Jacob-and-Anna-Maria couple married in the same church in 1785, so the identification is corroborated rather than closed.

The date every tree carries for Michael Welty’s death — 9 September 1815 — traces to one source: the same pedigree line whose marriage date the church register already disproved. Against it, the court recorded that administration of his estate opened 7 September 1815, and an estate cannot open before its owner dies. The family record now reads: Michael died on or before 7 September 1815. A small correction, made before the wrong date could fossilize further.

The same regimental pages held a hard story. At Antietam, on 17 September 1862, the colors of the 11th Pennsylvania were handed to William Welty of Company C, a Weltytown grandson, “who was killed a few minutes after it was put into his hands.” The flag passed to Frederick Welty of the same company, who was wounded and left it on the field. The roster confirms it: W. B. Welty, killed at Antietam Sept. 17, 1862 — dating a death the family record had left open.

With the first six generations standing on records, the project’s new goal is set: carry every line of the Edenkoben family — daughters’ lines included — complete through the seventh generation, each person resting on at least one independent record and every legend either upgraded or fenced. A scoreboard now tracks a children-search for every person in the first six generations. The tree crossed 298 people today.

The brother-candidate line got its paperwork upgraded at the source. The federal land office’s database confirms Jacob Welday Sr.’s Ohio patent — issued 10 August 1813 for the quarter-section at Cross Creek his family had entered under the old credit act — with son John patenting the adjacent quarter of the same section eight years later. A spelling-variant sweep found three of his sons as Ohio patentees. The Welday memorials confirm that Jacob Sr.’s own grave was lost to strip-mining; the oldest Welday stone belongs to his son. One memorial adds an independent source for his wife’s name (Mary Rubel), and another gives son Abraham a birthplace — Milford Township, Somerset County — exactly where Jacob Sr. and Michael sat side by side on the 1796 tax roll. A widely-copied online tree that grafts this branch onto the Manchester family is contradicted by the DNA and logged as a decoy. What would settle brother-or-cousin to Michael is a deeper Y-DNA test on a Welday-line volunteer.

The middle Dover daughter’s family firmed up. Peter Messerle, who married Christina Welty at Trinity York in 1784, has his own Dover baptism — 1760, son of Daniel Messerle — making the couple a local match, and two of their children (Magdalene, 1785, and John, 1793) joined the tree from the Strayer’s register. The same sessions opened the record hunt on her sister’s Gauf household and surfaced the 1803 “Elizabeth Gauff” marriage whose meaning — daughter, not widow — the Orphans’ Court work would settle two days later. All of it still rides, clearly flagged, on whether these Trinity brides were Philip Jacob’s daughters.

A long pass through the public family trees sorted every rival spelling of the name once and for all, and none of them is the Edenkoben line. Welti is Swiss; Welte is a separate German family; Wälti / Walti is entirely Canton Bern — the Emmental cradle of the unrelated Swiss branch; Welde is mostly sixteenth-century Norwich in England, with a distinct Baden family behind the rest; and Weldy is itself three unrelated families, one of them carrying a grafted “born 1614 in Switzerland” ancestor. The result is reassuring: the surname noise that has bred look-alike ancestors for years is now charted, so a stray “Welti” or “Weldy” in a record no longer has to be argued from scratch. One more origin tale — a “Peter Welty, Mennonite settler of 1727” — went into the myth ledger as belonging to none of these lines.

Some widely-shared online trees (on Geni and WikiTree) list a different father for the Edenkoben Michael Welty — John Jacob Welty, of the immigrant uncle’s line, rather than Philip Jacob. That rung is an easy place for trees to differ, so it’s worth setting gently alongside the family’s own papers. Those trees pair John Jacob with a wife, Christina Braff, married in 1757 — and a hand-copied Welty family register, kept in the first person, records that same couple: John Jacob Welty and Christina Broff, married 28 March 1757, followed by their children as they arrived — Ann Elizabeth (January 1758), John (April 1760), Christina (May 1762), and on. Michael isn’t among them, and the dates make it hard to fit him in: a couple who married in the spring of 1757, with their first child in January 1758, are unlikely to be the parents of a Michael already born around 1757. So the record leans back toward the Philip Jacob reading, and a Y-DNA test on the direct male line would settle the question for good.

The parentage at Michael’s generation still rests on DNA, and no record naming Philip Jacob’s children turned up this session — but the afternoon found the one documentary path never actually tried. An earlier York County search had come back empty, yet it had only covered administration bonds and partitions; the county’s Orphans’ Court dockets — a separate, image-only series — have never been read for a Welty. That matters because a person’s land could be sold by order of the court, with no will and no administration bond — which is exactly the shape of things here: Philip’s Dover house and two town lots left the family around 1788–1790 with no deed to explain it. If that sale went through the Orphans’ Court, the docket would name Philip Jacob and his heirs outright. A separate check of the published Adams County wills (1745–1800) came back cleanly empty — an honest negative pointing to a family already scattered west — so the docket browse is now the next real step.

One person on the six-generation chart had rested only on the Y-DNA pedigree: Isaac Welty, born 1820. His gravestone at Welty Cemetery in Stark County, Ohio — reading 22 June 1820 – 26 February 1894 — independently names his parents (Philip b. 1789 × Sarah Overholt), his spouse, siblings and children, matching the DNA line exactly and confirming the published 1903 family account. It also settles an old January-versus-June birthdate dispute in favor of June.

On the separate John Jacob Welde (b. 1710) line — the immigrant uncle’s family — the Weltytown / Greensburg thread firmed up. Lewis Welty (b. 1796) gained his wife, Susan Stanabaugh Wannamaker, and a death year of 1871; Daniel Welty (b. 1820) earned an independent gravestone at St. Clair Cemetery, Greensburg, that matches the Y-DNA pedigree exactly and adds a first wife, Mary Ann Drum, the pedigree had never recorded; and a new great-grandchild, Susanna b. 1820, surfaced in a church baptism. The remaining siblings of that generation have no free-record descendants yet — a later census project.

The 1902 obituary of Levi Welty — grandson of Michael, and the man credited with naming Cripple Creek, Colorado — was run down at last: the Rocky Mountain News carried it on 27 December 1902, confirming his age of 77, his survivors, and the Denver-1861 → Monument-ranch → Cripple Creek arc of his life. A cluster of 1868–1877 notices turned up alongside it, showing Levi as a well-known figure around Monument and Divide years before Cripple Creek made his name.

The same full-text move that recovered Michael’s children was pointed at the hypothesized daughters of his father, Philip Jacob of Dover — and it does not transfer, for a reason worth recording. Philip Jacob left no will, and his family had already scattered west before those daughters’ in-law estates arose, so no Welty was ever on hand to appear in the York County records the way Michael’s sons were in Ohio. The only eighteenth-century York-PA Welty records in the full-text corpus are the Dover baptisms already logged. No claim was upgraded here. The honest paths forward are an on-site York County orphans’-court docket search, or DNA from a daughter’s descendant — and the failed shortcut is logged so it isn’t re-run.

Five snapshots from the family’s own papers each carry a thread. Together they give a name for Michael’s burial place — “Crooked Run / Welty’s Graveyard,” matching the 1897 letter’s “Uncle Jake’s farm” — a location for Henry’s grave about ten miles north of Massillon, and a correction that sends Michael Jr. not to Ohio but to Owen / Clay Co., Indiana. Each is now its own line of inquiry.

The internet’s Welty lore was swept into one running ledger, each story paired with the fact that answers it: a Gettysburg “Swedish Welde,” a pair of eighteenth-century immigrant claims that belong to other families, a “Welty’s Mill” whiskey tale, the commercial coat-of-arms sellers, and an Ellis Island story that is really an 1851 arrival. Ten decoys logged with their counter-facts — and, notably, no “three brothers” or stowaway tale exists in any of it; that one appears to be the family’s own.

The Marian Welty Family Archive is now up as its own page — the hand-drawn lineage charts, the 1897 letter, the family registers and record copies, shown as-is with corrections noted beneath each. More of Marian’s collection will be added as it is re-photographed.

Henry’s son David Welty (1831–1924) followed the country west all the way to Escondido, California, where he lies in Oak Hill cemetery. His son Owen L. (“O.L.”) Welty turns out to be the recipient of the treasured 1897 family letter that carries so much of this family’s oral tradition — so we now know exactly who was being written to, and why a copy survived on the California side. Along the way, a widely-copied online memorial that had given David an impossible set of parents was corrected against the record. The Ohio-to-California branch is now traced end to end.

The Conewago (Quickel’s) baptismal register had been the wall behind several questions, its microfilm locked behind a library-only tier. The same register turns out to sit on a second archive, home-viewable — the way around the lock. Beyond Michael’s two children above, it confirmed two sons of the Manchester Philip (Johan Paul, 1794, and Johan Georg, 1797) exactly where an older list had placed them, while three other entries on that list proved absent here and must have been baptized elsewhere. A lead that had sat blocked for weeks is retired.

The Edenkoben register, read directly, dates Georg Wolfgang’s baptism to 29 September 1716 — the entry just below his fixes the month beyond doubt. The “26 August 1716” birth date copied across online trees has no support in the register and is set aside as index-lore. His death year remains the open question.

The sister-pair found earlier at Dover firmed up. Naming custom now points to Philip Jacob as their father: his wife was an Elizabeth and his mother an Anna Catharina, matching daughters Elizabeth and Catharina, whereas George Wolfgang’s known daughters carry an entirely different set of names. It stays a hypothesis — no record yet states the parentage outright — but the three candidate sisters are upgraded from a bare guess to a naming-corroborated set that leans one way.

A full-text sweep of York County records, 1783–1815, turns up John of Dover only as a taxpayer — no will, no estate administration, no orphans’-court file naming a wife or children. The free-record avenue for his family is therefore a closable negative: he is documented as a Dover taxable and by his 1783 marriage, and nothing more is coming from that quarter. What remains are image-only and on-site records.

Until now the Welty daughters were only names in a marriage cluster. A transcription of the Strayer’s / Salem Reformed baptisms at Dover puts two of them in the same records: Christine Welty stands as sponsor at her sister Elizabeth (Welty) Gauf’s child’s baptism in 1782, then marries Peter Messerle in 1784 — an internally consistent sister-pair at the family’s home church, the strongest daughter evidence yet. Which father they belong to — the Edenkoben line’s Philip Jacob, or the Manchester founder, whose household shared the same church — is left deliberately open, and the pair is logged as a hypothesis.

Philip Jacob’s house and two town lots at Dover were traced forward from his last tax returns of 1787–88. By 1789–94 no Welty appears on the Dover rolls, the 1790 census and 1793 county lists show no Welty household or widow there, and the property passed to a non-Welty buyer around 1788–90. The hoped-for first trace of his wife Elizabeth — as a widow holding the lot — does not materialize; that leg is closed negative. His death still brackets to about 1788–90.

Every Edenkoben Reformed marriage from 1739 to Philip’s August 1750 crossing was read entry-by-entry: no Philip Jacob Welde wedding. He married Elizabeth on the Pennsylvania side, ~1751–56 — redirecting the hunt to the German Reformed churches that served fresh Philadelphia arrivals (First Reformed Lancaster, the Goshenhoppen circuit).

The baptism book was read to its last leaf (“FINIS,” 1771): zero further Wäldi or Welde after 1752. That closes the hunt for late-born children of the vanished uncle Hans Philipp Wäldi — he leaves no trace in the register — and means the Edenkoben Reformed baptisms are now 100% swept.

Philip Jacob’s wife — “Elizabeth, m. 1754,” known only from a cousin’s pedigree — got her own research front (six new leads). First sweeps: every published Philadelphia marriage corpus (licenses, Gloria Dei, 1st Reformed) is negative, pointing the wedding to Chester Co.; a Chester in-law sweep found no father’s will but flushed out two flags — a “Weldy” taxpayer who proved to be Welsh (Obedia Wieldy of Haverford, decoy, killed same day) and a keeper: a 1804 Hagerstown MD will naming son-in-law Jacob Weldy — a documented Weldy family at Hagerstown, right where the 1897 letter’s legend put a “brother.”

The flagged “Philip Welty” entries at Bethlehem (1772) and Moore Twp (1781) — the one man who could have rewritten Philip Jacob’s blank years — are a persistent Moore Township farmer (40 ac in 1785, 60 ac in 1786, same neighbors both years, image-read) living there through the exact years Philip Jacob heads the Dover tax rolls. Two simultaneous households, two men: decoy, fenced.

GLO bycatch: a female “Catharine Walta” patented 80 acres in the estate farm’s own section in 1833 (“to HER heirs,” image-verified), with no 1830 household of her own — a fit for Michael’s widow Catharine living in a child’s household and buying land beside her dower third.

Jacob Welday’s 1800 Somerset household holds an unexplained second man 45+ — the first record-grade datum compatible with an aged Philip Jacob having gone west with his sons instead of dying ~1789 in Dover. Killer test: the 1810 follow-through. The same read pushes Jacob’s own birth to before 1755 (the old “b.1759” came from a 1946 query).

A sweep of every federal Rev-War collection finds no Michael and no Jacob Welty among the 13 exact-surname hits — consistent with York Co militia service living outside the federal rolls. The unpublished DAR GRC company roster remains the only known service paper for the pair.

A new cluster: John Welly of Bullskin Twp — adjacent to Donegal, ~10 mi from Michael’s 1797–1807 residence — is a live candidate for brother John of Dover tracking west; Joseph + Peter Welly of Springhill Twp are unsorted.

The Trinity York sibling-set — Elizabeth (×Gauf 1775), Christina (×Messerle 1784), Catharina (×Boehm 1785), plus John’s 1783 Ilgenfritz marriage — is now on the family tree as clearly-flagged hypotheses under Philip Jacob. Two date audits opened alongside: Michael’s “9 Sep 1815” death vs. an administration docket opened 7 Sep, and Philip Jacob’s death bracket now carries the 1800 fork above.

A full-text sweep of PA Archives vol. 21 surfaced a new 1783 Dover tax pairing — “Weldy Phillip 100” adjacent to “Weldy Jacob 100” (and 1781 prints two Philip entries). Fresh support for the farm-partition and brother readings.

The 1925 Ruthrauff book’s claim fails: the complete 1779 Dover tax list has no John Welty — the book misread the adjacent line “Wilte, John.” The real John of Dover (1786–88) is untouched, but 1790 shows two Johns in the future-Adams-Co grouping, and the Adams Weltys trace to Washington Co MD — so the Straban leg is now in doubt.

20 May 1753 turns out to be a Lutheran confirmation day — the Ancestry entries carrying that date as a “baptism” are confirmations of older children, and are now treated as suspect.

The vol. 21 hit-map cleanly separates the lines: a continuous “Welty, Jacob 100” in Manchester Twp 1779–82 (the Manchester I1 branch — George Wolfgang’s Edenkoben descendants), the Manheim Twp cluster of John + Peter verified (the separate Swiss family), and the Dover Weldys (the Edenkoben main line — the same family). The Taneytown/Emmitsburg MD Weltys — Catholic, from Eppingen not Edenkoben — are confirmed a separate family entirely.

All 20 sheets audited: numbering collisions fixed, the ghost 10-May wedding date purged everywhere, four new leads opened (highest-payoff: the New Hanover register test), and the trees regenerated.

“Jacob Welde” and “Geörg Welde,” both self-signed (facsimile-checked), arrived Philadelphia 16 Oct 1752 — now positively identified as the brothers George Wolfgang and John Jacob, who followed Philip across. Philip himself had crossed alone on the Royal Union in 1750 (“Phipps Welde”).

The register holds a tight cluster — likely a sibling set. New sisters surfaced: Christina × Messerle (1784), Catharina × Boehm (1785), plus Johannes × Ilgenfritz (1783).

An unpublished DAR GRC typescript roster (2nd Battalion, York Co., ~1780s) lists both Michael and Jacob Welty in one company — strong brother evidence. Photocopy ordered.

Dover Twp tax lists, 1782: Philip’s 100 acres shrinks to 80, a new adjacent “Jacob 80” appears, alongside Michael’s 44 — the classic signature of a father dividing land among sons.

Levi (1825–1902), a probable grandson of Michael, is credited with naming Cripple Creek. His son Alonzo ran a $100k livery business there.

The founder of Weltytown was John Henry Welty b. 1764 (wife Eva Catherine Stoner) — not the Edenkoben Michael, who is clean-negative in all three Ruff volumes 1782–1820.

The village pastor from 1681 was Johann Jacob Walth — the apex surname is also the pastor’s family. A kinship hypothesis is open.

The separate Swiss (I2b) family’s spine now runs York → Frederick MD (1766) → Rowan NC (1773) → KY → Missouri. All “Swiss origin” grafts onto the German line trace to Bishop Abraham of Dover Twp, Tuscarawas — Michael’s own township. Mystery of the bad grafts: solved.

A new FTDNA match between two living relatives connects the Manchester (I1) branch’s two halves internally, and a living cousin’s Big Y-700 order is confirmed by the project admin.

One of Henry’s sons, Dr. William Welty (1816–1905) of Wyandot County, Ohio, left a biographical sketch in a 1902 county history, and it matches the record at every turn: it names his father Henry and mother Mary Byerly, confirms the move from Fayette County, Pennsylvania to Stark County, Ohio, and repeats the old family count that the immigrant Michael married twice and had “seventeen children in all.” It also writes down the family’s Swiss-origin legend — an ancestor “from the river Aar, near Berne” — which we set aside: the documented origin is the Palatinate at Edenkoben, and the Y-DNA points to a different paternal line than the Bern Weltys the story names. Its value is as a witness that the Swiss tale was already being written into print by 1902.

Subscribed to Archion and found the correct register: Edenkoben Band 2 (Reformed), baptisms 1696–1738 — holding both John Jacob (bp. 1710) and Philip Jacob (bp. 1719).

FindAGrave proves his father William (b. 1816) was a son of Henry & Mary Byerly — making the Hicksville benefactor Henry’s grandson. William’s three pre-1905 children also grave-confirmed.

All ~33 patrilineage pages reviewed: Hans Jacob Waldi (b. ~1691, Edenkoben) named as father of John Jacob b. 1710 — and three different “Philip Jacob Weltys” at Dover cleanly split by haplogroup.

The FTDNA Welty Y-DNA project confirmed the Edenkoben branch map and provided the STR chart. A living cousin’s Y-67 kit traces to George b. 1823 — the direct line — making its Big Y upgrade the keystone test.