is assigned to this MS by Baiter, Lahmeyer and others. My judgment upon the existing evidence is that their estimate is exaggerated. P has peculiar errors of its own, and there is a most singular agreement in many places between it and Halm's D ('codex Vindobonensis saec. XV') which any one who reads Halm's critical notes will see to be so grossly corrupt as to be practically worthless.

Of M we have still less information,—nothing indeed beyond the readings Baiter gives from a collation by Karl Halm. While I have carefully weighed the MSS evidence on every point, even the minutest, the general result has been that when any two of these three MSS, viz. M G P, are agreed, I have adopted the reading, while I have seldom found it necessary to take a reading given by any one of them, when unsupported by other evidence. Very rarely have I had to depart from the consensus of the three. These three MSS give an unusually sound

basis for the text of the Laelius.

Baiter's text is denoted by B, Halm's by H, Lahmeyer's by L (I have used the third edition), Nauck's by N (seventh edition). The readings I have selected here for comment are chiefly those which illustrate points of grammar, syntax, or orthography.

§ 2.

1. 14. fere multis: some scholars have wished to strike out multis as a gloss. Cic. certainly would not qualify multis by the addition of fere, but fere is to be taken with tum, not with multis. The writer of one of Halm's MSS (D) felt the difficulty, and changed multis into omnibus.

1. 15. utebare: some MSS have utebaris, but although in the second person singular of the present indicative deponent and passive Cicero commonly uses the form in -ris, he oftener has the form in re in the second person singular of the imperfect (indicative and subjunctive) and of the future passive.

1. 19. querella: so B but H L N querela; P has quaerella (Mommsen p. 596) but in § 35 querela; see n. [C. F. W. Müller follows P in both places.]

1. 20.

§ 3.

Laeli: I have everywhere written single i in the genitives of nouns whose nominatives end in ius, ium, from a conviction that Cicero so wrote. See a good deal of the evidence in Neue, Formenlehre, 1,

85-94, ed. 2. [C. F. W. Müller writes -i throughout.]

1. II.

§ 4.

ceteros: so rightly spelt: it is high time the spelling caeterus was banished from modern texts, with coelum and other like enormities.

§ 5.

1. 28. te ipse: so PLN; but H B with G and some inferior MSS have te ipsum. Phrases like te ipse are so much commoner in Cicero than those like te ipsum, and the former are so easily and so often corrupted into the latter, that I have followed the reading of P. (C. F. W. Müller ipse.]

§ 6.

1. 3. habebat et multa: I have inserted et on my own conjecture. Without it Cato is one of the subjects to the verb putabatur; with it, is subject to the verb habebat below, propterea repeating quia, owing to the length of the sentence. With the old reading a long stop was needed after habebat, another after ferebantur, and the transitions were most awkward.

§ 7.

1. 8. volgus: so B, but H L N vulgus. It is almost certain that Cicero wrote volgus, volt and the like.

1. 9. Graecia reliqua: so H with G and other MSS; B L N reliqua Graecia with PD only.

1. 18.

affuisti: so L N rightly (the Latins objected to df, bf) but H B adfuisti.

§ 9.

1. 3. Gallum: all the MSS here have Gaium; so in 21, 1. 14 Gaios, but in 101, 1. 20 gallum. Since Galus is occasionally found for Gallus, Mommsen and after him B L N write the name with one / in all three places.

§ 10.

1. 7. vestrum: omitted by B L N with P D only.

§ 11.

1. 21. consul: bracketed by B H L, though in all MSS. Cf. however my n. On this principle many other words in the dialogue would have to be bracketed or ejected; e.g. sapientem in 7, 1. 8; sapientes in 18,

1. 12.

1. 27. omnis: I have written everywhere i not e in the accusative plural masc. and fem. of nouns whose genitive plural ends in -ium. Though Cicero's usage may have varied, he probably wrote the -i in the vast majority of instances.

§ 12.

1. 32. etiam nunc: H commends without adopting the conjecture of Victorius, tunc. See, however, my n.

§ 13.

1. 18. qui: Putsche in Philologus XII, p. 300 proposes cui after Gulielmius, which is adopted by H B L, B also taking a suggestion of H to change ut in to uti (utei). It is inexplicable to me why all these scholars should substitute by conjecture the very rare (if not unparalleled) ellipse of videbatur for the very common ellipse of the verbum dicendi. They should at least have gone on to insert videbatur after semper, as Kayser suggested. [C. F. W. M. marks qui as corrupt.]

§ 16.

1. 21. quaeruntur: B L quaeritur with P only. [C. F. W. M. quaeruntur.]

1. 23. mihi vero erit gratum: H B omit erit gratum (after Beier) though the words are in all MSS. The omission is groundless, though the elliptic answer mihi vero would be Ciceronian enough; cf. Acad. 1, 14; Off. 3, 35. [C. F. W. M. keeps erit gratum.]

1. 24. antevertit: edd. antevortit, a form which was distinctly archaic in Cicero's time, and used by Sallust on that ground.

§ 20.

1. 33. duos: L N duo with P only. I believe, however, that the weight of MSS evidence is in favour of duos as the Ciceronian form; inscriptions certainly point that way.

1. 3. nil unquam: five out of Halm's six Mss (including G) have quicquam only; the sixth has nihil only; so has P. Nil is wanted (see my n.), but the variants are best explained by supposing that Cicero wrote nil unquam.

1. 30. itaque non aqua etc.: Brieger, Beiträge zur Kritik einiger philosophischer Schriften des Cicero, Posen 1873, p. 7 proposes sweeping transpositions and other changes affecting this and the three succeeding sentences, on the very insufficient ground that the proverb applies to all friendships whereas Cicero professes to be speaking of perfect friendship.

pluribus locis: so H with G; BL N locis pluribus with P. Cf. 47, P. 43, 1. 4 multis locis.

§ 23.

1. 13. ne agri quidem: all MSS have nec, but nec quidem is a phrase not used by good writers; cf. n. on 30, p. 38, 1. 4.

1. 15. percipi: so all MSS, and the word makes very good sense, percipere being very commonly used by Cicero with the meaning 'to grasp' or 'to understand'. BL however follow Madvig (Opusc.

2, 279) in reading perspici, for which cf. 29, 1. 23. H approves Madvig's reading without adopting it. C. F. W. M. perspici; cf. however Hor. A. P. 335 dicta percipiant animi; Ter. Eun. 972 odium me percipit; Lucr. 3, 80.

§ 24.

1. 1. si quae: judging from the MSS evidence (some of which is in Neue 2, 233, 234, ed. 2) Cicero most probably wrote si quae and the like, not si qua and the like, both in the feminine singular and in the neutral plural.

§ 25.

1. 12. quid? amicitiam: edd. quid amicitiam? i.e. quid fuit amicitiam defendere? To avoid awkwardness, I have shifted the note of interrogation. For the form of expression cf. Acad. 2, 86 quid? hoc nonne videtur contra te valere? ib. 2, 81 quid? talpam num desiderare lumen putas? De fato 10 quid? Socraten nonne etc.? Examples might be multiplied to any extent. [C. F. W. M. has independently adopted the same punctuation.]

§ 26.

1. 22. quod quisque: all мss have quo; all quisque except P which has quis; L N accordingly write quod quis.

§ 32.

1. 13. ab his: so all Mss; H at ii.

1. 22. sintque: so all мSS; HBL however (after Beier) suntque, also est for sit.

1. 24.

concertatio: so G: HBLN certatio with P. On this passage Mr Shilleto (in MS note) compares Tac. Ann. 3, 55 nostra quoque aetas multa laudis et artium imitanda posteris tulit. nobis maiores: certamina ex honesto maneant.

Verum haec nos:

§ 33.

1. 9. deponerentur: B L N ponerentur with P only.

1. 32.

§ 36.

Vecellinum: H Viscellinum, but Mommsen on p. 598 discusses the name and arrives at this form.

§ 38.

1. 20. si simus: all Halm's Mss have sumus si, except E (codex Erfurtensis) which has simus si; so has P, and this reading (adopted by N) may very likely be right—'We should indeed be men of perfect wisdom, did the arrangement prove not to be faulty'. One point in favour of simus si is that when there is one apodosis with two protases, Cicero in the majority of instances places the apodosis between the protases. [C. F. W. M. si simus.]

1. 22.

by L N.

memoriam: P memoria, approved by Mommsen and adopted

§ 40.

1. 10. aliquantum: L with P only, on the ground that aliquantulum is foreign to Classical prose. It is now ejected from Div. 1, 73 where the inferior MSS alone have it, but there is still some evidence for it in De inv. 2, 29; Pro Quint. 15; (Cornificius) Ad Herenn. 4, 14.

§ 41.

1. 14. in P. Scipione: four of Halm's MSS (including G) have in p. nasicam Scipionem, the remaining two in scipionem; P has in scipione. It is not at all improbable that Cicero wrote in P. Nasica Scipione, or P. Nasica merely (as in Phil. 8, 13).

1. 18. proclivis: G H B proclivius; PLN proclivis (also four of Halm's MSS). Cf. 84, 1. 11 where three MSS have gravius for gravis. It has often been proposed to read proclivi, the adverb. (Cf. Tusc. 4, 42.)

§ 42.

1. 27. re publica: all MSS have the common abbreviation re p.: B H thinking the has come from the following p in peccantibus strike out publica, but the whole context shews them to be wrong.

§ 44.

1. 13.

verum: so MSS; edd. vero; cf. however my explanatory n.

§ 48.

1. 27. contrahat amicitiam: qy contrahatur amicitia?

1. 29. contigit: so B L with all the best Mss. H contingit.

1. 32.

1. 4.

§ 49.

animante: so B L with M P D; H animo autem with G.

§ 50.

et tam trahat: so MSS; H B L N et attrahat, needlessly.

§ 55.

1. 24. laborant: so I have written with all MSS; edd. laborent. I think laborant is used in its very common Ciceronian sense=. = solliciti sunt nor do they trouble themselves as to the person for whom etc.'.

§ 56.

13. facit: so MSS.; edd. faciat, to suit 59, 1. 21.