Class A — the ledger

Every class-A reading from variae-lectiones-transcript.md (apparatus pages [i]–vi, printed pp. 6–242), resolved against our own transcripts.

⚠ The rule this ledger exists to enforce

NOTHING HERE HAS BEEN APPLIED, AND NOTHING HERE MAY BE APPLIED SILENTLY.

Our transcripts record the 1853 plate. Class A records the manuscript behind it. They are two witnesses and the edition needs both. The 1853 says class A is pro veris habenda — "to be held as the true ones" — and it may well be right, but "the editor of our base text thinks his own text is wrong here" is an argument to be made in an apparatus, not a licence to overwrite a transcript. Any change to a transcript is a separate, argued commit, and it is Wilson's decision, not a session's.

Status values: applied is never used and should stay never used. Everything is recorded (located in our files, difference confirmed), recorded — not locatable (the apparatus's line number does not resolve), or structural (not a word-change but a layout or content claim).

1. Readings that change the WORDS — the ones that would alter our text

printed	our transcript reads	apograph reads	our file	note
6. 15	`Μνήσθητι` present	dele	front	a whole word dropped
6. 20	`τοῦ ἁγίου φροντιστηρίου τούτου`	`τῆς ἁγίας μονῆς ταύτης`	front	Latin sides with the plate (Collegii)
6. 22	`δοξάζοντι`	`δοξάσαντι`	front	present → aorist
12. 16	`φροντιστήριον`	`συνέδριον`	front	the same substitution as 6. 20, different word
12. 26	`τῶν νῦν` + `Βασιλικόν`	`τῶν τανῦν`, dele `Βασιλικόν`	front	Latin sides with the plate (Regium)
12. 27, 28	`ἐλεημοσύνης Βασιλικῆς`	dele	front	second deletion of a royal reference
12. 29	`Φροντιστήρια`	`Συνέδρια`	front	third of the φροντιστήριον set
16. 8	`διάγειν`	`παρὰ τοῦ Κυρίου εἰσελθεῖν αἰτησώμεθα`	front	a word replaced by a whole clause
16. 14, 18, 21, 25; 18. 3, 9	`Παράσχου, Κύριε`	`Παρὰ τοῦ Κυρίου αἰτησώμεθα` / `Παράσχου, Κύριε`	front	six places — the MS has the deacon's bidding and the people's response; the plate prints only the response
18. 21	—	adde `τῶν ἁμαρτώλων πρώτῳ`	front	a clause the plate lacks
26. 26	`ὁ Θεὸς` present	not in MS	front	Latin sides with the plate (Deus)
34. 27	`Τὰ ἔργα`	`τὸ πλάσμα`	day1	⚠ Latin sides with the APOGRAPH (Figmenta) — see `THE-LATIN-QUESTION.md`
34. 27–28	punct. after `παριδῇς`; `ὁμοίωμα`	dele punct.; `ὁμοίωμα` + point	day1
42. 4 sqq.	`αὐτοῖς`	`αὐτῶν`	day1
50. 18	`προσδοκίας`	`ἐλπίδος`	day1	Latin sides with the plate (exspectatione)
52. 16	—	insere `καὶ ἐσθήσω`	day1
52. ult.	`ἃ μέλλει ἀποθανεῖν`	dele	day1
62. 25	`ἀπελπισμένων`	`ἀπηλπισμένων` (plate perperam)	day1	the 1853 calls its own reading wrong
100. 7	`Βουλευτὰς`	`συμβουλευτάς`	day3	Latin Consiliarios leans to the apograph
100. 10	`[στρατιώτας]` — the 1853's own conjecture	`βιαστάς`	day3	⚠ see §3
102. penult.	`ὑπεριδὼν`	dele	day3
104. 3	`οὐχ ἑκὼν`	dele	day3
104. 6	`[οὐκ εἰς τέλος]` — the 1853's own conjecture	`οὐκ ἀεί`	day3	⚠ see §3
106. 20	—	adde `ὁ σεισμός`	day4
114. 17	—	`ἀπαρχή`	day4
124. 17, 19	punct. after `γενοῦ`; `αὐτοῦ`	dele punct.; `αὑτοῦ`	day4	breathing change
136. 24	`οὗτος`	dele	day5
138. 11	`ὑπὸ`	dele	day5
152. 8	Greek	⚠ Hebrew `ואתה תעל משחת חיי`	day6	see §2
158. ult.		`ἐκκόψῃς`	day6
162. 16	`ὑπερέχουσα`	`ὑπερβάλλουσα`	day6	Latin inconclusive (exsuperat is the Vulgate lemma)
162. antepen.		`τιμίαν`	day6
170. 16	`ἀνομίας`	`ἁμαρτίας`	day7	Latin sides with the plate (iniquitates)
174. 13	`ἀπόλυσαι`	`ἀπόλουσαι`	day7	⚠ Latin sides with the APOGRAPH (Ablue)
188. 18	`ἀδελφὰς`	`ἀδελφίδας`	day7
190. 13	`ἀνόμων`	`ἀνομιῶν`	day7
194. 2		`ἐπιστροφήν` + full stop	deprecation
194. antepen.	`ὡς εἰκὸς`	`[κατ᾽ ἴσον]` uncinis inclusum	deprecation	the MS itself brackets it
202. 15	`ῥῦσαί με`	⚠ Hebrew `הצילני`	deprecation	see §2
208. 5	`[σου]`	deleatur	deprecation
208. 15, 16	`εἰσελεύσονται` / `ἀπολαύσονται`	`εἰσέλθουσιν` / `ἀπολαύσουσι`	deprecation
224. 8	`ἑσπερινή`	`δειλινή`	evening	Latin inconclusive (Vulgate lemma; and the plate's Latin renders `ΔΕΙΛΙΝΗ` as vespertina on the same spread)
242. 6	words the 1853 bracketed as its own conjecture	dele uncinas; `τῆς` before `διανοίας`	evening	⚠ see §3

2. ⚠⚠ THE HEADLINE: THE 1853 DROPPED HEBREW THAT STOOD IN ANDREWES' TEXT

This overturns a standing correction in our own map, and the correction has now been reversed in STRUCTURE.md.

STRUCTURE.md carried, and the project memory repeated, the line: "CORRECTION: NO Hebrew in the main text at 198/200/212 (confirmed by eye) — the Hebrew is Variae Lectiones §A apparatus only."

Half right, and the wrong half is the important half. It is true that no Hebrew is printed at those places in the 1853 — our eyes were not deceived. But class A is "variæ lectiones quæ in textu ipso occurrunt" — readings that occur in the text itself, held to be the true ones. So what the note actually records is that the Hebrew stands in the manuscript's text and the 1853 edition left it out.

The Hebrew class A restores, in full:

printed	what the MS has	gloss
38. 3	`אדאג` in the margin opposite	Heb. for `μεριμνῶ` — "I am anxious"
76. 4	adde `יצר טוב` after `Φαντασίαν καλὴν`	the rabbinic yetzer ha-tov, the good inclination
86 (ult., penult.), 88	Voces Hebrææ Græcis respondentes in altero στοίχῳ subjiciuntur	a Hebrew column set under the Greek
92. 10, 13, 16	`תהום` · `תוהו` · `בוהו` before `αἱ ἄβυσσοι` · `γῆ` · `τὰ χλοερὰ`	Genesis i. 2 — the deep, and tohu wa-bohu
96. 25–31	a full Greek–Hebrew bilingual column (below)	the vocabulary of sin
152. 8	`ואתה תעל משחת חיי` in place of the Greek	Jonah ii. 7, under the MS's own heading `Τοῦ Ἰωνᾶ`
192. 17, 18	Hebraice leguntur; the Greek is at the foot of the page	the plate has inverted the MS's priority
198 (before line 1)	`אל תשחית` / `הצילני` ‖ `תעל משחת` / `חיי`	glossed by the 1853 itself: ne perdas / libera me ‖ subleva de corruptione / vitam meam
200 (after line 8)	`אל תשחית` / `הצילני`
202. 15	`הצילני` in place of `ῥῦσαί με`
210 (before last line)	eadem Hebraice — Ps. lxxxvi. 11
212. 1	`יראתי כסלתי` after `ὁ φόβος μου`

The bilingual column at printed 96. 25–31 — seven Greek names for sin, each with its Hebrew:

φαντασίαν יצר · σφάλμα שגגה · ἐνοχὴν πταῖσμα אשם · ἁμαρτίαν חטאה · παράβασιν פשע · ἀδικίαν עון · βδέλυγμα תועבה

That is the standard Hebrew vocabulary of sin — yetzer, shegagah, asham, chata'ah, pesha, avon, to'evah — set out as a lexicon. It is not in the printed book at all.

⚠ Consequences, all of them for Wilson, none of them for a session:

The edition's claim to carry four scripts is understated, not overstated. Andrewes' Hebrew was systematic where the 1853 shows it as occasional.
The יצר at 96. 25 and the יצר טוב at 76. 4 mean Andrewes was reading his own φαντασία through the rabbinic doctrine of the two inclinations. That is a substantive theological finding about the Preces, not a textual footnote.
Whether to restore any of it is a decision about what edition this is — a faithful reprint of the 1853, or a critical text of Andrewes. The project has never had to answer that question before, because until now the two coincided.
✅ All Hebrew here was RE-READ at native resolution on the Toronto scan, 2026-08-10, and every form is confirmed — nothing changed. The look-alike sorts (ד/ר, ב/כ, ה/ח) are cleanly separable there. אדאג at 38. 3, which had been secure on sense but not on its sorts, is now secure on both. ⚠ Still unread because nothing sets them: the אל תשחית / הצילני cluster at 198–202 and יראתי כסלתי at 212. 1. See SOURCE-IMAGES.md.

3. ⚠ The 1853 prints its own conjectures, and the manuscript overrules them

Three places where class A is not correcting a misreading but withdrawing an editorial invention:

100. 10 — the plate has [στρατιώτας], marked ex conjectura nostra; the MS reads βιαστάς. Soldiers for violent men.
104. 6 — the plate has [οὐκ εἰς τέλος], ex conjectura nostra; the MS reads οὐκ ἀεί.
242. 6 — dele uncinas quibus has voces e conjectura nostra additas notavimus, and read τῆς before διανοίας.
202. 3 &c., 9, 13 — ⚠ the strongest of the set: Notæ numerales non leguntur; ex Vers. Angl. a nobis desumpta sunt. The numerals in our transcript of printed 202 are not in the manuscript at all — the 1853 editor took them from an English version. Anything our files infer from that numbering is inference from an editor's scaffolding.
204. 17–23 — and the MS's own numbers, where it has them, run 2, 1, 4, 3, 5, 6, 7, not in sequence.

To the project's credit, the discipline already in force handles all of this correctly: the transcripts record brackets as printed and never silently mend. But [στρατιώτας] and [οὐκ εἰς τέλος] were transcribed as the 1853's supplied readings, and are now known to be wrong supplied readings. That is worth a line in the introduction.

4. ⚠ Class A and class C collide — and the collision is NOT resolved here

At eight lines a class-A reading and a Drake marginal address the same words:

printed	class A (the MS's text)	class C (Drake, in ora libri)
34. 27	`τὰ ἔργα` → `τὸ πλάσμα`	`τὰ ἔργα] τὸ πλάσμα (τὰ ἔργα superscriptum est)`
34. 28	`ὁμοίωμα` + point	`ὁμοίωμα] ὁμοίωσιν`
50. 18	`προσδοκίας` → `ἐλπίδος`	`προσδοκίας.`
52. ult.	dele `ἃ μέλλει ἀποθανεῖν`	`μέλλει ἀποθανεῖν.`
104. 3	dele `οὐχ ἑκὼν`	`οὐχ ἑκὼν.`
162. 16	`ὑπερέχουσα` → `ὑπερβάλλουσα`	`ὑπερέχουσα in marg. legitur`
170. 16	`ἀνομίας` → `ἁμαρτίας`	`ἁμαρτίας] ἀνομίας`
190. 13	`ἀνόμων` → `ἀνομιῶν`	`ἀνόμων] in marg. tantum`
224. 8	`ἑσπερινὴ` → `δειλινή`	`δειλινή] ἑσπερινή`

The tempting theory — that the printed 1853 text has in places adopted Drake's marginal reading rather than the manuscript's, so that our transcript at those lines is a later hand, not Andrewes — is attractive, partly supported, and not established. It cannot be settled from the printed apparatus alone, because the direction of class C's ] is ambiguous: Drake wrote in the margin of the apograph, so his lemma should be the manuscript's reading, yet at 34. 27 and 224. 8 the lemma is what the printed text has, which points the other way.

⚠ Do not build an argument on these collisions until the manuscript itself has been seen. They are recorded here as a question, in the same spirit in which the volume's cruces are left as cruces. The brief's standing instruction — "class C is evidence about Drake, not about Andrewes; never apply it to the text" — is unaffected and still governs.

Worth noting for whoever does see the MS: Drake was not a mere normaliser. He collates against the Wechel edition (pp. 4. 20; 170. 28), cites Aquila, Theodotion and Symmachus at p. 134. 23–24, and traces the Prayer of Manasses to the Apostolic Constitutions II. 22 apud Zonaram (Paris, 1618). The brief's "later hand with a normalising agenda" is accurate about his aim and understates his competence.

5. Structural claims — no words change, but the page does

10. 3 – 14. 11 — indicem quendam precum constituere videntur: this stretch is an index of prayers, marked with the MS's dot-symbols, pointing to pp. 186, 190, 198, 208, 214, 220, and fortasse 234 or 236. Our transcripts render it as text. If it is an index it is apparatus, and Prototype B would set it differently.
32. 27 — after ἐποίησα a ruled line, quasi spatium ad enumeranda peccata reservatum esset — a blank left for the reader to enumerate his own sins. A deliberate silence in the book, and our transcript does not show it.
30. 19, 22 — α΄, β΄ in the margin: the MS numbers parts with Greek letters, per the preface's principle iii. Our files use arabic throughout.
62. 21–22 · 70. 12 · 74. 6–20 / 76. 5 · 82. 27 · 116. 1–4 · 156. 12 · 158. 4 · 174. 5–6 · 182. 11–15 · 186. 11–12 · 204. 1–2 — column and line re-arrangements; several say two facing sets are meant to answer one another invicem, which bears directly on the mirror's page-fit.
82. 27 – 84. 7 · 186. 14–16, 21 sqq. · 188. 1–3 · 190. 26 — ⚠ manu posteriori addita sunt: the intercessions naming τῶν Παναγίων Βερκίγγ and Οὐϊντονίας (Winchester — Andrewes' own see) are a later hand in the manuscript. Whether they are Andrewes adding to his own book or someone else adding to it is exactly the sort of question the introduction should raise and not answer.
70. 19 · 72. 5, 13, 17 · 94 · 96. 9, 11 · 106–110 · 130–134 · 150–152 · 164. ult. · 170. 2 · 218. 6, 12, 13 — words to be set litteris grandioribus / majusculis: the MS's own display hierarchy (Τοῦ Μωϋσέως, Τοῦ Ἰώβ, Τῆς Χαναναίας, Τοῦ Μυριοταλάντου, Τοῦ Τελωνοῦ, Τοῦ Ἀσώτου, the prophets, Τοῦ Δαβὶδ, Τοῦ Σαλομῶν, Τοῦ Μανασσῆ). These are speakers' names — the Publican, the Prodigal, the Canaanite woman, the Debtor of Ten Thousand Talents — and the manuscript displays them. The 1853 flattens them into the text. For a mirror edition this is a live typographic decision.
128. 10 · 136. 4, 18 — cross-references to an English version in Andrewes' Works, vol. vii, pp. 290, 293. A witness the project has not consulted.

6. Not locatable

None. Every class-A citation resolves to a page held in part1/, which covers printed 1–263 continuously. Line numbers were not individually verified against our line counts — the apparatus counts lines of the 1853's Greek column, and our transcripts preserve those lines, but a line-by-line audit is a separate job and is not claimed here.