Quire 6: Herbal B · f41–f48
Click a leaf to turn it over (recto ↔ verso).
| Page | Subject | Language class | Scribe hand |
|---|---|---|---|
| f44r | herbal | A | 1 |
| f44v | herbal | A | 1 |
| f45r | herbal | A | 1 |
| f45v | herbal | A | 1 |
Bifolio 4 of 4 in this quire
The codex as bound today: nested quires, foliated by a later hand. Lost leaves (f12, f59–64, f74, f91–92, f97–98, f109–110) are simply absent, so their neighbours now face each other. The quire numbers were added only after the bifolios had already been mixed (Currier A and B bifolios alternate inside Quires 5 and 6, each one uniform in itself). They do not prove the original order.
Each bifolio was written and read as a self-contained folded sheet: Ar → Av | Br → Bv (e.g. 10r, 10v, 15r, 15v). Evidence: LSA text similarity is highest between the halves of one sheet and lowest between pages that face each other today; drawings cross the gutter of conjoint pages (78v|81r, 10v|15r, 33v|40r, 58v|65r, 76v|83r); apart from 57|66 and 85|86, all four pages of a bifolio are by one scribe. The first fifteenth-century binding would itself have been the misbinding. Sequences are proposed only for Quires 13 and 20 and are explicitly unverified; elsewhere the current order is kept.
Critique: Timm (2026) confirms the elevated similarity within physical bifolios but considers the reading-order conclusion underdetermined and criticises missing permutation controls for similarity-maximising resequencing.
The script runs line by line from left to right: paragraphs are left-aligned with a short final line, often open with a gallows glyph, and individual glyphs are strongly bound to the line start or line end.
An n-gram perplexity test finds a small but very consistent advantage for reading the EVA sequence right to left (n = 2–4). The author still holds the writing direction to be left to right and cannot rule out a cipher or scribal artefact; the Hebrew control gives mixed results. For a Hebrew substrate hypothesis this is an open question, not evidence.