¹ Here following the lead of Peterson in his edition of Book X (1891), Introduction, lxviii, where some of H's readings are given.

² This word appears to have been corrected from -verim, and is presumably the end of the preceding word optaverim, in the middle of which the Bernensis must have begun when the scribe of Bg started to copy it. Another transcriber of the Bernensis, as we shall see, preferred to start at nee. Full marks, then, to Spalding (ad loc.): ‘Melius absit haec vox. An irrepsit per errorem ex praecedentibus ultimis: optaverim ?’

³ See my previous article, p. 169.

¹ For example, in 2. 13 this manuscript stands with A against the Bernensis and all other manuscripts I have seen, except at times the D'Orville, in the following readings (I have marked with asterisks what appear to be the correct alternatives): 3 forte for fortasse*; 8 variare (variari*); 9 affectum* (factum); 11 mutant enim* (mutant etiam); 12 habet et in (habet in*); omission of puto (incorrect); 16 et si rarum (et stratum*).

² And correctly so. The word interrupts the balance of the sentence; and at so early a stage Quintilian surely strove to be elegant. In any case, H does not read instituti, but institi, which a late corrector has emended; moreover, it has institi again instead of infiniti just afterwards: the two errors are presumably connected.

³ I remain doubtful about the true reading here. Halm deleted in while retaining the ablative eloquentia. Schöll (Rh. Mus. xl (1885), 321 n. 1) deleted both words, perhaps preferably.Google Scholar

¹ G is often preferred to what A reads after correction, and rightly; for the corrections in A, made not long after the manuscript was written, are often wrong, though by no means never right; they seem to consist both of corrections from an ancestor, for they can fill gaps, and of independent conjectures. But that is a different matter.

¹ But hi is odd after corpora, despite Colson, and the trouble may be deeper rooted.

² For G having trouble with abbreviations, cf. 6. 3. 104 ‘si constat rebus’ A: ‘sicut ista rebus’ G. Did an intervening manuscript read ‘sicstat rebus’ ?

³ For the similarity see the plates of both in Châtelain, E., Paléogr. des class, lat.,Google Scholar plate CLXXIX, and the remarks of Lehmann, P., Philologus lxxxix (1934), 357.Google Scholar

⁴ ‘A direct or indirect transcript of the cod. Bernensis’, Sandys, J. E., History of Classical Scholarship, i. 631Google Scholar n. 1. Petrarch's corrections are not very numerous or very helpful. One agreeably demonstrates the hazards of conjecture. 11. 1. 18 ‘quae egerat in consulatu’: so b (A is here not available); in om. B A₁K; consula A₁K: consul Petrarch.

¹ Here with the meaning, apparently peculiar to Quintilian, of ‘at the same time’; though Bonnell does not classify this passage (or 4. 2. 29: 5. 13. 26) under that heading (Lexicon Quintilianeum, p. 717).Google Scholar

² Though here the correction futurus is easy enough to have occurred independently in the other manuscripts. The same may be true of some of the other passages classified above under (a).