This history is built almost entirely out of two kinds of record: what the family kept, and what the offices around them kept without meaning to. The first kind is the smaller and the dearer — a Bible carried out of Switzerland in 1710 with its fly-leaf filling up, one keeper at a time, in a formal French no one in the line could read within three generations; the collateral family Bibles that caught the births and deaths the principal one missed; the workbooks, charts, and half-remembered stories that carried the keeping on after the Bible itself was gone. The second kind is the public record that four centuries of governments generated around ordinary people without intending to preserve them for me — a herald's visitation pedigree signed in a Hertfordshire field in 1572, a will proved in Chowan County, a land grant entered in a Guilford County deed book, an administratrix's appointment in a Lunenburg order book, a Mississippi slave schedule, a census marshal's page, a certificate of death filled out by a country registrar in 1920.
One rule governs all three volumes, and it is the rule the family itself seems to have held: set the known thing down, in the place kept for it, and do not invent to fill the silence. Every date in these pages that could be checked against a calendar has been checked. Where two records disagree — and they disagree about the day the last of these women died, and the year one man died, and a great deal in between — both readings stand on the page rather than the tidier one alone. Where the record simply stops, as it stops on where one man died and when, the silence is left open, because closing it would mean writing fiction and calling it history. And where the record was not merely empty but made empty — where a Mississippi probate book that should name three enslaved people names no one, because the estate that would have listed them was never opened — the book says so as plainly as it can, because a silence that was made is a different thing from a silence that merely fell.
What follows is the consolidated working bibliography for all three volumes, organized by volume and, within each, by the kind of source, so that a reader who wants to know what a given fact rests on can find the instrument and weigh it. Each volume's list follows the tier convention used in that volume's research; the conventions differ in their letters but agree in their spirit — a primary document of the period outranks a secondary work about it, which outranks a family tradition, and every entry says which it is. The open questions are gathered at the end of each volume's list, in the numbered form they carried through the years of the work, so that whoever takes this line up next will know exactly where the digging was left off.
— DLS