If you want the Voynich Manuscript to be tidy, do not begin with Quire 8.
Quire 8 is where several of the manuscript’s favourite classifications collide.
Herbal page.
Cosmological-looking diagram.
Text-only page.
Starred paragraphs.
Key-like character sequences.
Later-looking cleartext.
Currier A.
Currier B.
Pages that refuse to fit either cleanly.
Different proposed scribal hands.
And then, precisely where we would most like continuity, six folios are gone.
Quire 8 is not merely an awkward gathering.
It is a controlled demolition of the idea that one simple partition explains Voynich.
In Quire 8, physical structure, illustration type, textual regime and proposed hand refuse to collapse into one neat history.
Quick Read
- Quire 8 probably originally contained five normal bifolia, or ten folios.
- The three central bifolia are missing, removing folios 59–64.
- The surviving folios are 57, 58, 65 and 66.
- f57r is a herbal page with Currier-B text and is assigned to Hand 5 in Lisa Fagin Davis’s model.
- f57v is a circular/cosmological-looking diagram containing rare characters and the famous fourfold 17-unit sequence; Currier classified its text as B, while René Zandbergen’s finer-grained classification treats it differently, and Davis assigns it to Hand 1.
- f58r and f58v are text-heavy pages marked by stars; their text is Currier A and Davis assigns them to Hand 3, a combination that complicates simple “hand = Currier language = page type” equations.
- f65r and f65v return to herbal imagery but contain little text; both are assigned to Hand 3 in the five-hand model, while their textual classification is sparse or uncertain.
- f66r is text-heavy, Currier B, Hand 5, and also contains columns of unusual single-character/word sequences plus mixed later-looking writing near a human figure.
- f66v is herbal again, Currier B under the classic classification, assigned to Hand 5, and carries the gathering mark.
- The surviving quire therefore combines at least three very different page architectures and several textual/palaeographic behaviours.
- The six missing central folios prevent us from knowing whether the surviving extremes were once connected by a smooth transition or by something entirely different.
- Quire 8 is best treated as a transition-zone laboratory, not as proof of one hidden story.
What a Quire Actually Preserves
A quire is not a chapter.
It is a physical gathering.
Several folded sheets are nested and sewn together.
That distinction matters because a quire can preserve production relationships more directly than a modern thematic heading does.
A folio can be “herbal” because of what we think the picture resembles.
It belongs to Quire 8 because of the physical construction of the codex.
The second relationship is material.
The first is classificatory.
Quire 8 becomes important when those classifications stop agreeing.
See The Bifolium as a Production Unit.
The Missing Middle: Six Folios Are Gone
René Zandbergen’s codicological reconstruction describes Quire 8 as a gathering that probably once contained five normal bifolia.
Only the two outer bifolia survive.
The three central bifolia are missing.
That removes:
- f59;
- f60;
- f61;
- f62;
- f63;
- f64.
This is not a small lacuna.
More than half the original gathering is absent.
That gives us a strange observational problem.
The surviving outer leaves look heterogeneous.
But we do not know what the missing middle looked like.
Perhaps it contained a smooth progression between the surviving page types.
Perhaps it contained more of the same mixture.
Perhaps it contained a completely different programme.
The missing leaves increase uncertainty.
They do not permit us to invent the transition we would like.
Quire 8 is a transition zone with its bridge partly missing.
The general rule remains the one established in The Missing Leaves and Quire Reconstruction: absence constrains inference but cannot provide positive evidence for desired content.
f57r: A Herbal Page That Already Refuses Simplicity
f57r looks, at first, like familiar Voynich territory.
A large plant fills much of the page. Two paragraphs of text occupy the remaining space, with the second carefully avoiding the plant.
Classic visual classification: herbal.
Classic Currier classification: B.
Davis five-hand model: Hand 5.
Already, however, Zandbergen notes that the two paragraphs have quite different textual properties.
That is an important warning.
A page-level label can hide paragraph-level variation.
If we assign one “language” or one text state to the entire folio, we may average away a real internal transition.
f57r therefore gives us the first Quire 8 lesson:
the page may be too large a unit for some textual classifications.
f57v: Same Folio, Different World
Turn the leaf over.
The plant disappears.
In its place is one of the manuscript’s most controlled circular diagrams.
Four concentric bands carry circular writing.
Four text items radiate outward.
Four human figures sit near the centre.
Rare character forms appear.
One ring contains a 17-unit sequence repeated four times with only small deviations.
This page has already earned its own specialist owner: f57v: The Seventeen-Sign Ring.
What matters here is its position inside Quire 8.
The recto is a herbal page assigned to Hand 5.
The verso is a circular diagram assigned to Hand 1 in Davis’s model.
Currier classified f57v as B, while Zandbergen’s later descriptive system gives the page a different placement within a more graded A/B/C-style landscape.
This is not a problem to be cleaned up.
It is evidence that physical folio, hand assignment, page type and text classification are independent axes.
One Leaf Can Carry More Than One Production History
f57 is a single physical leaf.
Yet under the five-hand model, recto and verso belong to different hands.
That possibility changes how we imagine manuscript making.
A bifolium or folio may pass between workers.
Different page programmes may be completed at different stages.
A diagram specialist may work on one side.
A herbal scribe may work on the other.
Or the discrete hand assignments may be over-partitioning continuous variation.
Either way, Quire 8 forces the hand model to meet physical codicology.
f58r: A Text Page Wearing Stars Before Quire 20
Now turn to f58r.
The page is almost all text.
Three stars sit near the starts of three paragraphs, with six, seven and eight points.
The visual resemblance to the later starred pages of Quire 20 is immediately interesting.
Yet f58r is classified as Currier A.
Davis assigns it to Hand 3.
Quire 20, by contrast, is dominated by B-like text in the conventional account.
This means a star-marked page layout is not owned exclusively by one Currier regime.
That is an important anti-collision principle inside the manuscript itself.
Same visual document device does not force same textual regime.
f58v: Still A, Still Hand 3, Still Unusual
f58v continues the text-heavy environment.
One six-pointed star sits above the first paragraph.
Four paragraphs occupy the page.
The longest fills about half the page, and one final line is centred.
The writing is notably neat.
Classic Currier: A.
Davis: Hand 3.
The centred line connects Quire 8 to the larger question of paragraph-final registers discussed in Where Does a Sentence End?.
Again, one gathering contains features normally discussed in several separate Voynich research compartments.
Then the Manuscript Disappears for Six Folios
After f58, the numbered sequence jumps to f65.
What did the missing centre contain?
We do not know.
This is exactly where narrative instinct becomes dangerous.
Because f57 contains herbal + circular material and f58 contains star-marked text, a reader may want the missing pages to complete a transition into the later surviving f65/f66 material.
Perhaps they did.
Perhaps not.
The missing middle should be encoded in our model as:
unknown six-folio interval.
Not:
the place where our preferred transition probably happened.
f65r: A Huge Plant With Almost No Text
When the surviving gathering resumes at f65r, we are back in a herbal-looking world.
A large plant fills the page.
Its roots and frilled leaves have visual similarities to f57r.
But the text almost disappears.
Only two or perhaps three words form a plant label.
There is not enough normal running text for a secure classic Currier assignment.
Davis assigns the writing to Hand 3.
This page is a useful reminder that textual classification has a sample-size problem.
A page can be visually obvious and textually underdetermined.
Do not force Currier A/B where the evidence is only a couple of tokens.
f65v: Herbal Image, Centred Paragraphs
f65v continues the plant programme but changes the text layout again.
Two very short paragraphs sit in the lower part of the page around the plant stem.
Unusually, the paragraphs are centred as a whole.
Davis assigns the page to Hand 3.
Zandbergen’s finer classification treats the page as neither a straightforward classic A nor a straightforward classic B case.
The point is not which letter we assign it.
The point is that the surviving quire contains transitional behaviour that a binary map can conceal.
f66r: Text, Key-Like Sequences and a Human Figure
f66r changes the page again.
It is predominantly text-only.
Five paragraphs contain 32 lines of running text.
In the left margin are two columns: one containing short Voynich words, another containing numerous single characters.
Near the lower-left human figure is a mixture of Voynich writing and clearer alphabetic-looking material that has attracted competing readings.
Classic Currier: B.
Davis: Hand 5.
The single-character columns belong to the broader Key-Like Sequences problem.
The mixed alphabetic writing belongs to Marginalia and the Later Hands.
Quire 8 owns the fact that these systems meet on one surviving page inside a gathering already rich in transitions.
f66v: Herbal Again, With the Gathering Mark
The final surviving side of Quire 8 returns to a large plant drawing.
Three paragraphs sit above it.
The classic Currier classification is B.
Davis assigns Hand 5.
The quire mark appears in the lower-right corner.
That quire mark is physically mundane and intellectually precious.
It anchors this extraordinary mixture inside one actual gathering.
Without codicology, we might treat these pages as unrelated anomalies scattered across the book.
The binding structure tells us they survived together.
The Quire 8 Matrix
Put the surviving pages side by side conceptually.
- f57r: herbal + B + Hand 5 + internally different paragraphs.
- f57v: circular diagram + rare/key-like sequence + classic B / alternative transitional classification + Hand 1.
- f58r: text-only + stars + A + Hand 3.
- f58v: text-only + star + centred final line + A + Hand 3.
- f65r: herbal + almost no text + Hand 3.
- f65v: herbal + centred short paragraphs + transitional/uncertain text class + Hand 3.
- f66r: text-heavy + key-like columns + later-looking cleartext + B + Hand 5.
- f66v: herbal + B + Hand 5 + gathering mark.
No single column explains the others.
Herbal does not equal one hand.
Starred layout does not equal one Currier regime.
One folio does not equal one hand under the five-hand model.
One gathering does not equal one page type.
That is the Quire 8 result.
Why Quire 8 Is Dangerous for One-Cause Theories
Suppose someone proposes:
Currier A and B are simply two scribes.
Quire 8 asks what to do with Hand 3 writing A-like material and Hand 1 on f57v while its recto is assigned to Hand 5.
Suppose instead:
Currier A and B are simply two topics.
Quire 8 asks why herbal pages, text-only pages and a circular diagram cross those categories messily.
Suppose:
Each quire is one production batch with one system.
Quire 8 asks why the surviving batch is so heterogeneous.
A strong theory can still explain the quire.
It simply cannot pretend the quire is simpler than the object.
Transition Zone or Accident of Survival?
The phrase transition zone is useful but needs care.
It can mean two different things.
Historical transition
The manuscript’s production system actually changed during this gathering: new scribe, new textual regime, new page function, new convention.
Survival transition
The surviving outer leaves happen to preserve different material because the missing centre removed the pages that once made the organisation obvious.
Those histories make different predictions.
A historical transition might produce gradual changes in glyph forms, token families or layout conventions across the original sequence.
A survival accident might create abrupt apparent discontinuities that disappear if the missing content were restored.
We cannot currently choose confidently.
The distinction should remain explicit.
Quire 8 and the Currier Continuum Question
Classic Currier A/B classification is enormously useful because the difference is real.
But later researchers have repeatedly encountered pages that look intermediate or mixed under particular features.
Zandbergen’s own page/folio/bifolium labels use additional categories to capture some of this structure.
Recent quantitative work in 2026 has also argued that the A/B distinction may be a low-resolution projection of a more structured generative system rather than a single primitive binary.
That is a preprint result and should be treated as a current model, not settled fact.
Quire 8 is exactly the sort of material such models must explain.
The question is no longer merely:
Is this page A or B?
It becomes:
Which underlying features make this page more A-like in one dimension and more B-like in another?
Quire 8 and the Scribe Debate
Under Davis’s five-hand model, Quire 8 contains writing assigned to Hands 1, 3 and 5.
If those assignments represent three historical people, the quire is direct evidence of collaborative production.
If some assignments instead represent continuous variation or task-conditioned writing by fewer people, Quire 8 is equally valuable because the page-type changes give us a natural test of that alternative.
Does handwriting change because the person changed?
Because the text function changed?
Because the writing happened at a different time?
Because the same person wrote circular text differently from herbal prose?
Quire 8 gives all of those hypotheses somewhere to collide.
Quire 8 and the Exemplar Problem
A heterogeneous gathering could arise if several source exemplars were copied into one codex.
One source supplies herbal material.
Another supplies circular diagrams.
Another supplies record-like or starred text.
Different scribes or production moments could then preserve traces of those source boundaries.
This is historically plausible.
It is not proven by heterogeneity alone.
A multi-exemplar theory should predict recurring packages of:
- textual features;
- illustration style;
- layout;
- scribal habits;
- physical production choices.
See The Exemplar Problem.
The f57 ↔ f65 Visual Echo
Zandbergen notes visual similarities between the plants on f57r and f65r, especially the bulbous roots and frilled/fingered leaves.
This is useful because the pages sit on the surviving outer structure of the same damaged gathering.
Could they belong to one visual subprogramme?
Possibly.
But the textual situations differ sharply.
f57r has substantial B-like prose.
f65r has barely enough writing to classify.
This is a perfect example of why picture family and text regime must be separate variables.
Internal image similarity is worth tracking.
It should not be asked to supply missing text statistics.
Star Markers Before the Starred Section
f58r and f58v are especially important because star markers appear long before the final Quire 20 sequence in present foliation.
This tells us the star-as-organisational-device is not exclusive to the final section.
That immediately weakens a simplistic equation:
star marker = recipe section symbol.
The marker may instead belong to a broader document grammar.
It could signal entry, paragraph class, emphasis, addressability or another function used in more than one textual context.
The Starred Pages article owns Quire 20.
Quire 8 owns the evidence that similar star devices occur elsewhere.
A Gathering Can Contain Several Document Roles
This may be the deepest Quire 8 lesson.
We should not assume one physical gathering equals one intellectual genre.
Historical codices can be copied in batches, reorganised, rebound or compiled from heterogeneous source material.
Even a deliberately planned manuscript can group different page functions together.
Therefore Quire 8 heterogeneity does not automatically imply disorder.
It may represent:
- a planned transition;
- a mixed reference gathering;
- several production episodes;
- multiple source exemplars;
- a later rearrangement;
- or an original structure obscured by the loss of its centre.
The surviving evidence does not yet choose among them.
Why Whole-Manuscript Averages Can Fail Here
Suppose we pool every Quire 8 token.
The result mixes:
- A-like text;
- B-like text;
- rare circular sequence material;
- labels;
- single-character lists;
- pages with almost no text.
The average may describe no actual page well.
Quire 8 therefore argues for hierarchical analysis.
Token → line → paragraph → locus → page → folio → bifolium → quire → manuscript.
At each scale, ask which variables are homogeneous enough to pool.
This protects real structure from being dissolved by convenience.
A Better Quire 8 Research Matrix
For each surviving locus, record independently:
- physical folio;
- recto/verso;
- bifolium;
- visual page type;
- text locus type;
- classic Currier class;
- continuous/text-feature scores rather than only binary class;
- hand assignment under competing models;
- line and paragraph position;
- rare-character rate;
- star-marker presence;
- image/text interaction;
- later or extraneous writing;
- confidence in every field.
Then ask which variables actually travel together.
That is much stronger than deciding in advance that “Quire 8 is where A becomes B” and forcing every feature into that sentence.
What Would Count as a Real Transition?
A genuine production transition should generate ordered change.
For example:
- glyph forms drift gradually;
- one token family declines as another rises;
- layout conventions change in a consistent direction;
- hand features shift together;
- physical bifolium order matches the change;
- the same transition recurs in another independent production sequence.
Without ordered evidence, “transition” can become a story told after observing heterogeneity.
Quire 8 deserves the stronger test.
What Quire 8 Does Not Prove
- It does not prove Currier A evolved into B here.
- It does not prove three historical scribes worked on the quire.
- It does not prove the missing folios contained transition material.
- It does not prove starred text is a recipe register.
- It does not prove the f57/f65 plants are the same species.
- It does not prove several source manuscripts were combined.
- It does not prove the present quire structure preserves the original conceptual chapter structure.
What It Does Give Us
- A physically coherent gathering with extraordinary surviving heterogeneity.
- A major six-folio lacuna whose location matters.
- A direct demonstration that page type, Currier regime and hand model are non-identical axes.
- A folio whose recto and verso receive different hand assignments in the five-hand model.
- Starred A-text pages outside Quire 20.
- Rare circular/key-like material integrated with a major diagram.
- A page combining running text, single-character lists and later-looking cleartext.
- A compact laboratory for testing layered production hypotheses.
Primary School: Sort the Same Eight Cards Three Ways
Take eight cards.
Give each a colour, shape and number.
Sort by colour.
Then by shape.
Then by number.
The groups change because the classification question changes.
Quire 8 is the same lesson with pages: physical gathering, visual type, textual regime and hand are different sorting rules.
Secondary School: Remove the Middle Six Cards
Arrange ten cards in a sequence that gradually changes colour and shape.
Remove the middle six.
Now ask students to reconstruct the transition from the four survivors.
Several histories fit.
This is why Quire 8’s missing centre must increase uncertainty rather than provide a blank canvas.
JC and Adult Readers: Model Quire 8 as a Multilevel Dataset
Do not assign one label per page and stop.
Represent each locus with independent features, then test clustering at paragraph, page and bifolium scale.
Ask whether Currier-like variables align more strongly with:
- hand;
- page function;
- physical folio;
- visual regime;
- or local textual neighbourhood.
Quire 8 is valuable because competing explanations make different cross-axis predictions.
Reader Checklist: Before You Tell a Story About Quire 8
- Which exact surviving folios are being discussed?
- Which six folios are missing?
- Is the claim page-level, paragraph-level or bifolium-level?
- Which Currier classification system is being used?
- Is the page confidently classifiable from enough text?
- Which hand model is being used?
- Would the conclusion survive a different hand model?
- Is visual page type being confused with textual regime?
- Are stars being treated as one universal symbol?
- Are f57r and f57v being assumed to have one writer because they share one leaf?
- Is the missing centre being used as imagined evidence?
- Does the proposed transition make a prediction elsewhere?
Frequently Asked Questions
What folios survive from Quire 8?
Folios 57, 58, 65 and 66 survive. Folios 59–64—the three central bifolia—are missing.
Is Quire 8 Currier A or B?
Both A- and B-classified material occurs among the surviving pages, while some sparsely written or transitional cases resist a simple page-level binary. That heterogeneity is one reason the quire is important.
How many scribes wrote Quire 8?
In Lisa Fagin Davis’s five-hand model, surviving Quire 8 writing is assigned across Hands 1, 3 and 5. The broader five-scribe model is currently debated, so this should be treated as a palaeographic hypothesis to test rather than a settled headcount.
Why is f57v so important?
It combines a circular diagram, rare characters, aligned start markers and a fourfold 17-unit sequence. Its hand and text classifications also interact unusually with the herbal recto of the same folio.
Does Quire 8 prove the manuscript was assembled from several books?
No. Heterogeneity is compatible with multiple exemplars or production episodes, but it can also arise inside one deliberately varied codex. Independent evidence is needed.
Research Foundations
- René Zandbergen — Quire 8 folio-by-folio documentation, including physical reconstruction, page types, Currier labels and hand assignments.
- Prescott Currier — Some Important New Statistical Findings, the historical A/B and hand observations.
- Lisa Fagin Davis — How Many Glyphs and How Many Scribes?
- Christophe Parisel — A Quantitative Confirmation of the Currier Language Distinction (2026 preprint), useful as a current quantitative model rather than settled consensus.
The Final Idea
Quire 8 is valuable because it refuses the answer before the question has earned it.
It is one physical gathering.
It is not one visual genre.
It is not one textual regime.
Under one influential model, it is not one hand.
And more than half of its original leaves are missing.
Quire 8 teaches us that the manuscript’s real architecture may live in the intersections among classifications, not inside any single classification by itself.