Voynich Missing Leaves and Quire Reconstruction

The Voynich Manuscript we study is not the Voynich Manuscript that once existed.

That sentence should sit beside every theory built from page order.

Leaves are missing.

Whole bifolia are missing.

Two leaves appear to have been deliberately cut out after binding.

Entire gatherings once known to later numberers have disappeared.

The present binding is not the first physical state we can safely assume.

And the sequence a modern reader scrolls through on a screen is therefore a surviving arrangement, not a guaranteed transcript of the maker’s intended intellectual journey.

This is not a minor technical problem for librarians.

It changes what counts as evidence.

If two page families appear to meet, perhaps something once stood between them.

If a sequence appears incomplete, perhaps it literally is.

If the zodiac lacks two signs, missing material becomes a serious candidate explanation.

If the final starred section contains 324 surviving markers, the missing central bifolium tells us not to treat 324 as an original designed total.

If a grand theory requires the manuscript to run neatly from plant → vessel → body → zodiac → recipe, lost leaves and uncertain order become part of the burden of proof.

A missing page is not permission to invent what was there. It is a reason to lower confidence in every theory that depends on what was beside it.

This article is about learning to read absence as evidence without turning absence into imagination.

Quick Read

One-sentence answer: the surviving Voynich Manuscript contains 102 parchment folios but preserves an old foliation running to 116, allowing at least fourteen missing folios to be identified; the pattern of stubs, skipped numbers, quire marks and missing bifolia shows that losses occurred at different times and forces every theory of page sequence, section boundaries and whole-book function to distinguish the surviving order from the original or intended one.

Yale’s Beinecke catalogue records 102 surviving parchment folios plus later paper material.
The old foliation runs to 116, with numbers skipped where leaves are now absent.
Fourteen folios are missing under the conventional foliation: f12; f59–64; f74; f91–92; f97–98; and f109–110.
Folio 12 and folio 74 were apparently cut out individually; stubs or cut traces survive.
The other losses occur as complete bifolia or missing gatherings rather than isolated leaves.
Quire 8 has lost three bifolia from its centre—folios 59 through 64.
Later quire-numbering conventions identify now-missing gatherings 16 and 18; the skipped folio numbers imply that the missing leaves were still present when old numbering systems were established.
The final starred-text gathering has lost its central bifolium, folios 109–110.
Quire and folio numbers were added after the manuscript’s principal writing and illustration, and the two numbering systems appear to belong to different hands or moments.
The current order is not guaranteed to equal original production order, intended reading order or conceptual order.
Some physical evidence indicates that production was organised substantially at bifolium level.
A bifolium is one sheet folded to make two leaves; pages far apart in modern reading order can therefore belong to the same original skin.
Missing leaves can remove transitions, labels, diagrams, headings, entries, cross-references or ordinary content; their exact content cannot be recovered merely from the gap.
Codicology can reconstruct constraints around the loss even when semantics cannot be reconstructed.

The central lesson is simple:

the surviving manuscript is evidence of a larger object.

A Book Is Built From Sheets, Not From Page Numbers

Modern reading begins with page numbers.

Codicology begins underneath them.

The basic physical unit is often a bifolium: one sheet folded down the middle to produce two leaves and four writing surfaces.

Several bifolia can be nested into a quire or gathering.

Gatherings are sewn into a codex.

This creates a counterintuitive fact.

The first and last leaves of one quire may be halves of the same sheet.

Two consecutive folio numbers may belong to different sheets.

A missing half of one bifolium can leave its partner behind elsewhere in the gathering.

So when researchers say that “f12 is missing”, the loss is not only one numbered page location.

It is a wound in a physical construction.

The position of the stub, conjugate leaf and sewing can tell us how the wound happened.

That is why the codex can remember a missing leaf even after its text has vanished.

The Old Numbers Are Fossils of a Fuller Book

The manuscript’s folio numbering is later than the main text.

That might make it sound less important.

In fact, the later numbers preserve evidence about an earlier physical state.

The numbers increase through the manuscript and skip the numbers assigned to leaves no longer present.

That means a later foliator knew enough about the physical sequence to number material that subsequently vanished—or worked while it was still there.

Quire 8 is particularly revealing.

The surviving structure does not simply look like a standard gathering missing one obvious sheet.

The foliation jumps from 58 to 65.

Six consecutive folio numbers are gone.

That is three bifolia.

The numbering therefore preserves a memory of a substantially larger quire.

This is a beautiful example of metadata becoming archaeological evidence.

A later numberer accidentally tells us something about a version of the manuscript we can no longer see.

Fourteen Missing Folios Are Not Fourteen Identical Events

The missing material falls into different physical classes.

That matters because different loss patterns imply different histories.

Single-leaf removals

F12 and f74 appear to have been cut out individually. Their partner halves survive and physical traces remain.

Missing central bifolia

F59–64 represent three bifolia missing from the centre of Quire 8.

Lost gatherings or bifolia in the later sequence

The gaps around f91–92 and f97–98 align with missing quire material under the traditional quire-numbering reconstruction.

Missing final central bifolium

F109–110 are absent from the centre of the final star-marked gathering.

These patterns should not be collapsed into “fourteen pages fell out”.

Some leaves were removed.

Some sheets disappeared as physical pairs.

Some losses may belong to entire gathering-level events.

Different physical histories create different semantic uncertainties.

Folio 12 Was Cut Out

The missing f12 is not merely inferred from a number gap.

The remaining physical structure indicates an individual leaf was removed.

This raises a question that cannot currently be answered:

Why would somebody cut out one leaf?

Possible reasons are almost unlimited.

a desirable illustration;
damage;
censorship;
reuse of parchment;
a collector removing a specimen;
accidental destruction followed by trimming;
unknown later intervention.

None is established.

The cut proves removal more strongly than motive.

This is another example of claim control.

Physical damage can establish what happened to the object without telling us why a person did it.

Folio 74 Was Also Cut Out—and It Sits Beside Quire 13

F74 is especially tempting because of where the gap sits.

Immediately after the zodiac sequence and before the human-figure material, a leaf is missing.

It is extraordinarily easy to imagine what we want f74 to have contained.

A bridge diagram.

A title.

An anatomical transition.

The missing Capricorn and Aquarius.

A key connecting zodiac and baths.

All are narratively satisfying.

None follows merely from the location of the loss.

The physical evidence tells us that a leaf was removed.

Its lost content remains unknown.

This may be the single best page in the manuscript for teaching one rule:

The more useful a missing page would be to your theory, the more careful you should be not to write it yourself.

The Six Missing Folios in Quire 8 Are a Larger Structural Loss

F59 through f64 represent six missing leaves—three complete bifolia.

This is not a tiny interruption.

A six-folio loss removes twelve page surfaces.

Twelve surfaces can carry a substantial textual or visual module.

The remaining quire therefore cannot be interpreted as though its present transition structure were complete.

What was lost?

We do not know.

But the number of missing surfaces constrains the possibilities.

It could contain multiple entries.

Several plant pages.

Text-only matter.

A diagram sequence.

A mixture.

The responsible conclusion is not a reconstruction of content.

It is recognition that any model of Quire 8 has a large unobserved region in its centre.

The Missing Quires Tell Us the Numberers Saw More Than We Do

The manuscript’s old quire marks traditionally run through twenty, while gatherings identified as 16 and 18 are now absent.

The missing quire numbers align with gaps in the old foliation.

This is one of the strongest reasons to think the numbering systems were applied when now-lost material still existed or was otherwise fully accounted for.

A modern reader sees absence.

A later medieval or early modern numberer apparently saw a continuous enough physical structure to number across it.

The numbers therefore become witnesses.

Not witnesses to meaning.

Witnesses to existence.

Something was there.

We know less about what it said than about the fact that the codex once included it.

F109–110 Warn Us Not to Worship the Number 324

The final star-marked section contains 324 surviving marginal stars.

Numbers attract symbolism.

324 can be factored.

Compared with calendars.

Connected to astronomical schemes.

Made to look intentional.

The missing central bifolium immediately destroys the strongest version of that temptation.

The original total was almost certainly larger if the missing sheet carried the same kind of material.

Therefore 324 is a count of surviving markers.

It is not securely the total chosen by the original maker.

This is a general lesson in damaged datasets.

Do not build symbolic meaning from a total until you know the dataset is complete.

The manuscript tells us that this dataset is not.

The Missing Zodiac Signs Are a Different Kind of Absence

Capricorn and Aquarius are absent from the surviving zodiac sequence.

Does that prove two zodiac folios are missing?

No.

It creates a hypothesis.

The manuscript also contains known physical losses near the transition toward Quire 13, including f74.

This makes loss more plausible than it would be in a physically complete codex.

But the arithmetic is not simple.

Aries and Taurus are split across multiple diagrams.

One missing leaf gives two surfaces.

The original zodiac architecture may not have been one sign per leaf.

Therefore we cannot reconstruct the exact lost zodiac content from the missing signs alone.

Semantic absence and physical absence can support one another.

They are not interchangeable evidence.

Present Page Order Is Not Automatically Original Order

Loss is only one reason sequence is difficult.

The manuscript has been bound and handled over centuries.

The present sewing is old, but scholarship has not established that it is the first binding.

Several arguments suggest the current folio order may differ in places from the originally planned order.

This means four orders must remain separate:

production order — when sheets were written or drawn;
assembly order — how bifolia were nested into gatherings;
binding order — how gatherings were sewn at a particular stage;
reading order — how the intended user navigated the information.

They may overlap.

They do not have to be identical.

A reference manuscript can be manufactured in batches and read non-linearly.

A compiler can gather earlier units into a later order.

A rebinder can change adjacency without changing the content on each leaf.

Any theory that reads the codex as one continuous modern textbook therefore has to earn that assumption.

Bifolium Consistency Is One of the Strongest Production Clues

Several textual and palaeographic patterns tend to remain consistent within individual bifolia.

Currier’s A/B observations are especially important here.

Later handwriting classifications also correlate substantially with physical sheets.

This suggests that the manuscript was produced to a significant degree sheet by sheet.

That changes how missing material should be imagined.

A lost bifolium is not merely two random leaves.

It may represent one coherent production batch.

Its two halves may have shared:

scribe or hand;
textual regime;
source material;
illustration family;
production stage.

We cannot recover those properties for a missing sheet directly.

But neighbouring physical structure can sometimes constrain what kind of production environment the missing sheet belonged to.

That is more disciplined than inventing its semantic content.

A Stub Is More Valuable Than a Guess

A small strip of parchment left near the binding can look unimportant.

It may tell us more than a thousand speculative sentences.

A stub can indicate:

a leaf once existed;
it was removed after folding;
its partner leaf survives;
the loss occurred after a particular stage of assembly.

None of this reveals the lost text.

But it transforms the state of the claim from:

maybe something is missing

to:

a physical leaf was removed here.

Good manuscript research often advances through exactly these apparently boring upgrades.

The story becomes less dramatic and more constrained.

Foldouts Make Collation Harder Because One Leaf Can Behave Like Several Pages

The Voynich Manuscript contains multiple double, triple and larger foldout leaves.

One famous large sheet around f85–86 carries the Rosettes material.

Foldouts break ordinary page intuition.

A single physical leaf may have several visible panels.

Writing can cross a fold.

Numbering may refer to the folded leaf rather than every exposed panel.

A digital scan necessarily turns one three-dimensional handling sequence into several flat image files.

This makes the old foliation and modern screen sequence even less equivalent to physical manufacture.

Collation is difficult not because scholars are careless.

The object itself has unusual geometry.

Any reconstruction must therefore begin from parchment topology rather than treating the manuscript as 200 ordinary rectangular slides.

Missing Leaves Can Destroy False Section Boundaries

Modern descriptions divide Voynich into visual sections.

Herbal.

Zodiac.

Biological or balneological.

Cosmological.

Pharmaceutical.

Recipes.

These labels are useful for navigation.

But if material is missing near a transition, the sharpness of the boundary may be partly an artefact of survival.

Imagine a twelve-page transitional module has vanished.

Two surviving families now appear to switch abruptly.

The original book may have blended them gradually.

Or the missing module may have been completely different.

We do not know.

The correct conclusion is not that every section boundary is false.

It is that survival can sharpen boundaries the maker did not intend.

Missing Leaves Can Also Hide Real Boundaries

The opposite is equally possible.

A missing leaf may have contained a heading.

A title page.

A table.

A transition diagram.

A blank separator.

If such a leaf vanished, two originally distinct modules could now look continuous.

Therefore damage can both create false discontinuity and erase real discontinuity.

This is why missing content cannot be repaired by whichever narrative makes the surviving sequence look nicest.

Loss increases uncertainty in both directions.

The Grand Medical Pipeline Has a Missing-Data Problem

One attractive whole-book theory reads the sections operationally:

identify plant → select part → prepare → treat body → time astrologically → record recipe.

The sequence looks especially persuasive when the codex is viewed from first surviving page to last.

But missing leaves create three problems.

1. Transition uncertainty

What appears to be one module moving directly into another may once have contained additional material.

2. Order uncertainty

Present assembly may not equal original intended operational order.

3. Sampling uncertainty

The surviving pages may overrepresent or underrepresent certain page types because whole units were lost.

The result is not that the medical-compendium hypothesis becomes impossible.

It means page order alone cannot prove the pipeline.

Independent cross-references must do more of the work.

A Missing Page Can Never Be Used as Positive Evidence for Its Imagined Content

This rule sounds obvious and is violated constantly.

Suppose a theory requires a table linking zodiac signs to treatments.

There is a missing leaf between zodiac and Quire 13.

The theorist says:

The missing page probably contained the table.

That sentence does not strengthen the theory.

It protects the theory from missing evidence by moving the evidence into an inaccessible page.

The proper inference is weaker:

Because a page is missing, we cannot exclude the possibility that a transition or cross-reference once existed there.

That is uncertainty.

It is not positive support.

The difference is one of the most important habits in serious historical reasoning.

Can We Reconstruct Lost Content From Conjugate Leaves?

Sometimes the surviving half of a bifolium can constrain the production environment of the lost half.

If the surviving partner belongs to one handwriting regime, one visual family and one textual cluster, perhaps the missing half shared some of those properties.

That can be a useful probabilistic inference.

But the inference has limits.

One bifolium can contain different content on its two halves.

A writer can switch topic.

A foldout can have different page roles across one sheet.

And exceptional mixed-hand cases exist in the manuscript.

Therefore the conjugate leaf can constrain a lost page’s production context more safely than its semantics.

“Probably written in the same broad production state” can be reasonable.

“Contained the missing Aquarius diagram” requires much more.

Can Ink Transfer or Damage Reveal Lost Pages?

In some manuscript and document contexts, offset ink, stains or impressions can preserve traces of lost adjacency.

This is an attractive forensic possibility.

But the Voynich Manuscript has been rebound, handled, folded and stored for centuries.

A stain may come from a later cover.

Damage may postdate original adjacency.

Any claimed transfer trace therefore needs conservation-grade physical analysis.

It cannot be inferred casually from a digital image.

This is an important boundary for reader-facing research.

High-resolution scans are extraordinary.

They do not transform every discoloration into a forensic witness.

The Later Binding Is Its Own Layer of History

Yale’s older catalogue describes the present vellum binding as eighteenth or nineteenth century.

Conservation work indicates the surviving sewing itself may be much older, possibly medieval, while not proving that it was the first binding state.

This is exactly why “the binding” should not be treated as one event.

A codex can be sewn, repaired, resewn, rebound and recased.

Each intervention can preserve some earlier structure and change another part.

The current cover belongs to one historical layer.

The quire marks belong to another.

The foliation belongs to another.

The principal text and drawings belong to earlier layers.

Reading the object well means refusing to flatten those layers into one manufacture date.

Foliation Is Not Pagination

A folio number identifies a leaf.

Each leaf has a recto and verso.

So f12 would have had f12r and f12v.

When six folios are missing, twelve page surfaces are lost.

This matters because casual statements such as “fourteen pages are missing” undercount the physical loss.

Fourteen folios means twenty-eight ordinary recto/verso surfaces, though foldout geometry elsewhere can make the concept of “page surface” even more complicated.

Precise vocabulary protects quantitative reasoning.

Leaf.

Page.

Bifolium.

Quire.

They are not interchangeable terms.

When the book itself is the evidence, vocabulary about the book is part of the method.

What Missing Leaves Can Tell Us

The surviving codex is incomplete.
Some losses occurred after foliation or quire numbering.
Some leaves were deliberately removed rather than merely falling free.
Several losses occurred in complete physical bifolia.
Some gatherings were once larger than their surviving state.
The final starred section originally contained more material.
Modern section transitions may not reproduce original transitions.
Current total counts can be survivor counts rather than designed totals.
Physical partner leaves can constrain production context around a loss.
Page-order theories need stronger evidence than simple adjacency.

This is genuine knowledge created by absence.

What Missing Leaves Cannot Tell Us Alone

The exact words on a missing page.
The precise image once drawn there.
Why an individual leaf was cut out.
Whether the missing zodiac signs occupied f74.
Whether a lost page contained a section title.
Whether a missing bifolium connected two modern visual sections semantically.
Whether a lost final record was a recipe, index or note.
The original total number of starred entries.
The complete original reading order.
The complete original conceptual architecture.

The boundary matters.

Codicology can prove incompleteness more readily than it can resurrect content.

A Better Reconstruction Ladder

1. Establish the physical unit

Leaf, bifolium, foldout or entire quire?

2. Establish the loss mechanism

Cut stub, missing conjugate sheet, missing gathering or uncertain loss?

3. Use old numbering

What did the foliation and quire marks know about the fuller manuscript?

4. Reconstruct only topology

Where did the physical unit sit? What leaf was conjugate? Which gathering contained it?

5. Add production constraints

What hand, Currier regime or visual family characterises the surviving physical neighbours?

6. Lower confidence in sequence claims

Do not repair the gap by narrative instinct.

7. State semantic possibilities separately

A candidate reconstruction can be discussed without entering the “known” column.

This ladder lets physical evidence do its real job without forcing it to become a lost-page generator.

What Would Count as a Real Reconstruction Breakthrough?

A dramatic breakthrough would not necessarily require finding the missing leaf itself.

Imagine a contemporary copy of one lost folio is discovered in another library.

Or an old photograph made before a twentieth-century loss comes to light.

Or a binding fragment preserves transferred text that can be physically demonstrated to come from a lost facing surface.

Or a fifteenth-century inventory describes a diagram whose wording and physical position match one missing unit.

Or conservation imaging reveals enough surviving stub content to assign a missing leaf to a particular visual or textual family.

Each would add new positive evidence.

That is the crucial difference.

A real reconstruction breakthrough adds information from the world.

It does not merely make a missing page more convenient for a theory.

Primary School: Build a Book and Remove a Sheet

Take three sheets of paper.

Fold them in half.

Nest them into a small booklet.

Number the leaves.

Now remove the middle sheet.

Ask the child:

How many leaves disappeared?
Which page numbers are missing?
Can you tell that something was removed?
Can you tell what words were on it?

The final question is the important one.

You can reconstruct the physical loss more confidently than the lost story.

That is codicology in a child’s hands.

Lower Secondary: Separate Evidence From Reconstruction

Give students three columns:

Observed: f74 is missing and cut traces survive.
Inferred: the leaf existed after the bifolium had been assembled.
Imagined: f74 contained Aquarius.

Ask them to place statements into the correct column.

This builds a durable mental habit.

Inference is legitimate.

It simply needs to remain visibly different from observation.

The exercise transfers directly to science, archaeology and news analysis.

Upper Secondary: Missing Data Changes the Claim

Give a student a dataset of 100 measurements.

Then reveal that 20 values are missing—not randomly, but from one entire time interval.

Ask whether the same trend claim remains safe.

This is analogous to a missing quire.

Missingness has structure.

A single random leaf loss differs from losing an entire production batch.

Students learn a sophisticated statistical idea:

how data went missing affects what can safely be inferred from what remains.

JC and Adult Readers: Treat Survival as a Sampling Process

At a higher level, the surviving manuscript is a sample generated by historical preservation.

The sample is not random.

Leaves attractive enough to cut out may disappear selectively.

Central bifolia may detach more easily under certain binding histories.

Whole gatherings can be lost.

Damaged foldouts can be especially vulnerable.

Therefore frequencies in the surviving codex may differ from frequencies in the original object.

This does not make statistics useless.

It makes the preservation process another variable.

A theory about original section proportions should acknowledge that survivor bias may exist.

The manuscript is both dataset and archaeological remnant.

A Parent and Teacher Guide

The missing leaves are one of the best ways to teach children how to reason with incomplete evidence.

Ask what is physically known. Number gap, cut stub, missing bifolium, lost quire.
Ask what the physical evidence implies. Was the leaf removed individually or lost with its partner?
Ask what remains unknowable. Exact text and images.
Do not reward convenient reconstructions. A missing page does not become whatever the theory needs.
Lower confidence in adjacency arguments. Lost pages can remove transitions.
Distinguish surviving totals from original totals. 324 stars are surviving markers, not securely the original count.
Use absence as a constraint, not as a canvas.

This is a rare form of intellectual maturity:

learning that “we cannot know this from the surviving evidence” is sometimes the highest-quality answer available.

Reader Checklist: Before You Build a Voynich Theory From Page Order

Are any folios missing near the proposed transition?
Is the loss one leaf, one bifolium or a larger gathering?
Does a physical stub survive?
Are skipped folio or quire numbers preserving an earlier state?
Are you treating present binding order as original reading order?
Does the proposed sequence survive bifolium structure?
Could a missing unit have altered the apparent section boundary?
Are you using imagined missing content as positive support?
Is a numerical total based on a complete surviving dataset?
Could rebinding have changed adjacency?
Does an independent textual or visual cross-reference support the proposed order?
What part of the argument remains true even if the missing content was completely different from what you imagine?

The final question is powerful.

A robust theory should not collapse simply because an unknowable page refuses to cooperate.

Frequently Asked Questions

How many Voynich folios are missing?

Fourteen under the old foliation: f12, f59–64, f74, f91–92, f97–98 and f109–110.

How many folios survive?

Yale’s catalogue describes 102 surviving parchment folios, with complex foldout leaves among them.

Were all missing leaves torn out?

No. F12 and f74 appear to have been cut out individually. Other losses are complete bifolia or missing gathering-level material.

Why do we know leaves are missing?

Skipped old folio numbers, quire marks, surviving stubs, conjugate-leaf structure and formal collation all preserve evidence of lost material.

Was the current page order the original order?

Not established. Several physical and historical considerations make it unsafe to treat present order as guaranteed original production or reading order.

Did f74 contain the missing zodiac signs?

Unknown. Its position makes zodiac-related reconstruction tempting, but no surviving evidence establishes the lost content.

Does the missing final bifolium mean there were more than 324 stars?

Probably, if the missing f109–110 carried the same star-marked record architecture. The important safe statement is that 324 is the surviving count, not securely the original total.

Can missing content be reconstructed?

Physical position and production context can sometimes be reconstructed probabilistically. Exact text and images cannot currently be recovered from the gaps alone.

Why does this matter for decipherment?

Because theories often depend on page sequence, section transitions, counts and adjacency. Missing material changes the reliability of all four.

Research and Further Reading

The Final Idea

The lost leaves of the Voynich Manuscript teach a difficult lesson.

Absence is not nothing.

A skipped number is evidence.

A stub is evidence.

A missing conjugate sheet is evidence.

A gap in a gathering is evidence.

But none of them is a photograph of what disappeared.

That distinction is the discipline.

We can know that a page existed without knowing what it said.

We can know that the surviving sequence is incomplete without rebuilding the missing sequence in our preferred image.

We can let the hole remain a hole.

And sometimes that is exactly what world-class research looks like:

not filling the blank until reality gives us something with which to fill it.

Voynich | Everything eduKate Knows and Tested | The Missing Leaves and Quire Reconstruction