Keep Source Notes Attached While a Research Bible Reflows

A note number is a position, not an identity. That single distinction is the whole problem. When you type a superscript 3 beside a claim and type “3.” at the end of the page, you have told the reader where to look, not what they will find. The number is a set of directions. Revise the document and the directions go stale while the page still looks tidy.

For a research bible that gets moved around — passages relocated, notes inserted, unsupported references cut — typed numbers and page-end note text are the wrong mechanism. Use the application’s real endnote references so a marker and its note body travel as one object. Then decide what the numbering counts, keep enough source detail inside each note that it can be identified without its number, and test the whole thing in a copy before you trust it.

One boundary before the mechanics. A working endnote proves a marker and a note are connected. It proves nothing about whether the note supports the claim. That question is editorial, it is answered by reading the source, and no improvement to your file structure touches it.

Begin with checked claims and identifiable notes

Do not start by replacing the mechanism. Start by reading what you already have.

Inventory the document by hand. List every superscript you typed and every note you typed, and match each pair to the claim it was meant to serve. This is the last moment the old arrangement is legible, and it is also the moment you find out whether your notes were ever correctly matched. Confirm that each claim’s source decision is settled — that you know which source you are citing and are not about to rewrite or drop the claim. Wiring a live reference to a note you are deleting next week is wasted work and, worse, gives the note a permanence it hasn’t earned.

For each surviving note, ask whether it can identify its source on its own. A note reading “3. Archive” or “3. See above” has no meaning apart from its position. It cannot survive renumbering, because renumbering is exactly the process of removing the thing it depended on. Keep author, work, locator — page, chapter, box, date, whatever the source's own structure supplies — and enough about where the source lives that another editor could go and find it. A number is a document reference. It is not a source identity and it is not a shelf mark.

There is a second reason to do this reading first. When notes are typed at page ends, a reflow breaks them visibly: the note strands on page 2 while its claim moves to page 1, and you see it. That visible breakage is an accidental audit. Once the notes are native and move correctly, that audit disappears. Nothing will look wrong anymore, including the claim you matched to the wrong source three drafts ago. Do the editorial pass now, while the document is still noisy enough to argue with you.

Insert native references and choose the numbering scope

InDesign’s endnote feature creates a reference associated with text and manages its numbering and its endnote frame. What you get when you place a reference is two related things: a marker in the flow of the story, and a note body that lives in the endnote area rather than at the foot of the page.

Find the command yourself rather than copying a path from an old screenshot. In recent versions endnote creation sits under the Type menu, and the bulk of the settings sit in the endnote options; the exact wording and the depth of the submenu shift between releases. Read your own menu and your own options panel. If your version and mine disagree about a label, yours is correct for your document.

The important thing to understand is what the marker is. A native reference is a real object, not a styled character. A superscript 3 formatted with a character style, sitting in ordinary body text, is still just a character. It has no relationship to any note. You can usually tell the two apart by behaviour: move a native note and the number changes by itself; move a typed digit and it changes only if you retype it. Delete a native note body through the application’s own workflow and the marker responds; a typed digit has no idea the note under it is gone.

Before you insert anything, decide what the numbering counts, and check what the scope control is currently set to. Endnotes can be numbered within a story — where a story is one continuous flow of text through linked frames — or across a larger document that contains several stories. Whether you want one continuous sequence or a restart per story depends on how the bible is actually assembled. A single flowing story almost certainly wants one sequence. A bible built from several separate stories, or assembled from multiple documents, may want the wider scope so that “note 12” means the same thing wherever a reader encounters it.

The decision is not reversible for free. Change the scope and every number downstream changes with it. So make the choice deliberately, write it down, and treat it as part of the bible’s specification rather than a personal setting. And check the endnote frame: endnotes collect into a frame you place in the document, not at the bottom of each page. If the bible’s current design genuinely depends on notes at the foot of every page, that is a footnotes question, not an endnotes one — a separate feature with separate behaviour, worth confirming in the same help area before you commit.

Test reflow by moving, inserting, and removing references

Now build the test, and build it in a duplicate. Keep the untouched original so that a failed experiment costs you nothing but the time you spent.

The fixture below is invented. It exists to test linkage, and nothing in it should be treated as research. It is a three-page passage with five claims and five notes; every source is fabricated for the walkthrough.

Laid out as typed notes, the fixture reads: page 1 holds claims A and B with notes 1 and 2; page 2 holds claim C with note 3; page 3 holds claims D and E with notes 4 and 5.

Now perform three operations, and after each one inspect the claim, the marker, the note body, the numbering, and any overset text.

Move a cited passage. Lift the passage containing claim C to the top of page 1. With native references, the anchor moves and the note body follows it into the endnote frame; C renumbers from 3 to 1, A and B each shift down by one, and D and E keep their numbers. In a version with typed superscripts, something different happens: the digit 3 travels with the words because it is only a character sitting in the text, while the note text “3.” stays where you left it at the end of page 2. Page 1 now shows the numbers 3, 1, 2 in reading order against notes 1 and 2 at its foot. A reader following the 3 on page 1 finds no note on that page; the note text is stranded on page 2. That is the failure the whole method is designed to eliminate, and it is worth producing once so you can recognise it.

Insert an earlier note. Add a new claim X with its own note ahead of C. The native sequence becomes X=1, C=2, A=3, B=4, D=5, E=6, with nothing retyped. With typed numbers you must now edit every following superscript and every following note, and the moment you miss one you have either two notes sharing a number or a silent gap.

Remove a reference. Delete claim D and its reference, and E moves up one place. Here, be careful: confirm in your own version whether removing the marker removes the note body or leaves it behind, and whether the application offers a separate command for deleting a note outright. The behaviour is worth verifying rather than assuming, because an orphaned note body that no marker points to is exactly the kind of quiet mess that this arrangement is supposed to prevent.

After the three operations the order is X, C, A, B, E, numbered 1 through 5. Claim C has been note 3, then note 1, then note 2. Its source never changed: it was Harbour Records, Box 12, Ledger 3 at the start and it is the same ledger at the end. If you had verified by counting numbers — five markers, five notes, looks right — you would have confirmed the appearance and not the connection. Verify by following claim C to its note and checking the source identity inside it. The number is the thing that moved; the source is the thing that must not.

Verify the edition the reader will actually receive

The file that passes these tests is not the file your reader gets. Two more inspections stand between them.

First, look at the endnote frame in the final document. Every note the reader needs should be visible in it. Overset endnote text — notes pushed beyond the frame — still exists in the file and is simply absent from the page and from the export. A numbering sequence can be perfect while a note has quietly gone missing from view.

Then export the reading edition you actually intend to deliver and open it in the viewer you expect the reader to use. Go straight to a claim you moved during the test and follow its reference. Check three separate things, because an export can preserve one without the others. Check the numbering: are the references sequential and sensible in the exported form? Check the note text: is it present, complete and legible? Check navigation: does the reference actually take the reader to its note, and back? A print-oriented export can carry correct numbers and complete note text with no live link at all. An interactive export can carry links that the reader’s viewer ignores. Neither outcome is a fault in your document, but both are facts about the delivery, and the reader needs to know which one they are holding.

If a format cannot carry working navigation, say so plainly rather than implying that every reference will respond to a click. Correct numbering, readable note text and working navigation are three claims, and you have to check each.

What the notes are for

Aim to end with notes that sit against the claims they were written for, in the edition the reader receives, in every format you ship. The point of the whole exercise is that the connection is a property of the document rather than of your memory about the document.

Keep the failure test, though — the typed superscript, the stranded page-end note, the reader sent to the wrong place. It is the best argument you have for why the styled character that merely looks like a reference is not one.

And keep the other distinction intact too. A technically flawless reference and an editorially adequate citation are not the same achievement. The machinery guarantees that the marker and the note move together. It will do that just as faithfully for a note that never supported its claim in the first place. Native endnotes make the document quiet. Quiet is not the same as correct, and the part that checks correctness is still you, reading the source.

Frequently asked questions

Why is a typed superscript number a problem in a research bible that gets reflowed?

A note number is a position, not an identity. Typed superscripts and page-end notes tell the reader where to look, so when passages move, the directions go stale while the page still looks tidy. A typed note can strand on another page, and a reader following the marker can be sent to the wrong place. The suggested mechanism is the application’s real endnote references, so a marker and its note body travel as one object.

What should I do before converting notes to native references?

Start by reading what you already have. Inventory by hand every typed superscript and typed note, and match each pair to the claim it was meant to serve. Confirm that each claim’s source decision is settled, because wiring a live reference to a note you may delete next week is wasted work. For each surviving note, make sure it can identify its source without its number, keeping author, work, locator and enough detail about where the source lives. A note like 3. Archive or 3. See above has no meaning apart from its position.

What does a working endnote reference prove, and what does it not prove?

A working endnote proves that a marker and a note are connected. It proves nothing about whether the note supports the claim. That question is editorial, answered by reading the source, and no file-structure improvement touches it.

What numbering scope must I decide, and can I change it freely?

Endnotes can be numbered within a story, where a story is one continuous flow through linked frames, or across a larger document containing several stories. A single flowing story almost certainly wants one sequence; a bible built from separate stories or multiple documents may want the wider scope. The decision is not reversible for free: change the scope and every downstream number changes with it, so choose deliberately and treat it as part of the bible’s specification. Also check the endnote frame, because endnotes collect into a frame you place, not at the foot of each page. If the design depends on foot-of-page notes, that is a footnotes question.

How should I test reflow and what should I check in the reader’s edition?

Build the test in a duplicate, keeping the untouched original. Move a cited passage, insert an earlier note, and remove a reference, checking the claim, marker, note body, numbering and overset text after each operation. Then inspect the endnote frame in the final document, because overset endnote text still exists in the file but is absent from the page and export. Export the reading edition you intend to deliver and open it in the viewer you expect the reader to use. Check three separate things: numbering, note text presence and legibility, and navigation to the note and back. A print-oriented export can carry correct numbers and complete note text with no live link, and an interactive export can carry links the reader’s viewer ignores, so say plainly if a format cannot carry working navigation.

Further reading

• One TV Bible File or an InDesign Book? Coordinate Chapters Without Breaking Their Layouts — Read before fixing your numbering scope: it sets out how one file and a multi-chapter Book differ, and why that structure governs what note numbers count across the bible.
• Lay Out a Commercial Treatment in InDesign Without Losing Copy Between Pages — Explains overset text in threaded frames: why words can sit in the file yet never reach a page, and how unconnected frames hide missing copy instead.