The following emendations will not be out Avoidance of of place in illustrating the carrying out of this dittography.

The passage is an illustration of the use of Xaipw with accusative. In the lacuna Heath proposed ἐλθόντα, Cobet σωθέντα. But neither of these words could well have been omitted. Apply the law of the accidental omission of similar words, and read:—

ἔτ ̓ ὄντα τόν τε μιαρὸν ἐξολωλότα.

'I'm glad, Hercules, you're alive, and the rascal slain.” ἔτ ̓ ὄντα fell out before τόν τε.

Eur. Frag. 254.

This fragment should run thus:

ἐκ τῶν δικαίων γὰρ νόμοι τ' αὐξήματα
μεγαλὰ φέρουσι πάντα δ ̓ ἀνθρώποις τάδε.
τάδ' ἐστὶ χρήματ ̓ ἤν τις εὐσεβῇ θεόν.

For é in the first line, the MSS. have ei, and
Táde is left out in verse three before the follow-

ing τάδε. From justice law is strengthened,
and justice is everything to man: justice is
money, if a man be pious.' The repetition of
Táde is in accordance with a universal custom,
by which a word is repeated from the end

Avoidance of of a preceding line to emphasise a climax. dittography. Thus Hor. Ep. 1. xi. 30:—

Quod petis hic est:

Est Ulubris, animus si te non deficit aequus.

Eur. Frag. 652.

Πόλλ' ἐλπίδες ψεύδουσι καὶ ἄλογοι βροτούς.

καὶ λόγοι is proposed by Dindorf; but λόγοι has no business here. If the form aλóλoyos can exist I would read:

Πόλλ' ἐλπίδες ψεύδουσι καλόλογοι βροτούς. 'Fine talking hopes' (castles in the air). According to the law, καλόλογοι became καλογοι= καὶ ἄλογοι.

Tacitus, Annals, i. 51.

Incessitque itineri et proelio.' I cannot but think that this is too pregnant a construction even for Tacitus. 'He advanced [prepared alike] for marching and fighting.' Orelli defends it by Ann. xiii. 40: qui viae pariter ac pugnae composuerat exercitum.' But this is nothing like so strong an expression. Substitute in' or 'ad' with the accusative for the datives in the latter passage, and there is nothing unusual to strike a reader. Make the same substitution in the first passage, and the construction is nearly as harsh as before the substitution. I cannot help believing, inasmuch as the word pars begins the next sentence, that a contracted form of paratus, resembling pars, has dropped out. Read: 'Incessitque itineri et proelio paratus. Pars,' etc.

So Curtius iii. 8, itineri simul paratus et Avoidance of proelio.' dittography.

I will conclude this subject with an application of this principle which may fail to convince the reader, although it has convinced me.

Propertius, III. xxxi. 5.

Hic equidem Phoebo visus mihi pulchrior ipso
Marmoreus tacita carmen hiare lyra.

The poet is describing a statue of Apollo.
The reading above has two glaring faults:—
(1) Equidem is properly only used with the
first person; (2) Marmoreus has nothing to
agree with.
with. Read:-

HiC PHOEBUS Phoebo visus mihi pulchrior ipso
Marmoreus tacita carmen hiare lyra.

'Here a marble Apollo, more beautiful it seemed to me than Apollo himself, oped his his lips, accompanying his silent lyre.'

Phoebus was lost before Phoebo, and equidem was just the word that a half learned scribe would select to make up a line with.

This edition being in the main critical, I Alterations have never shrunk from altering the text of the text. where an emendation appeared necessary. In

doing so I have bound myself by three conditions:-(1) to avoid needless1 alterations: (2) to adhere as closely as possible to the best

'Madvig's condemnation of causeless alterations is pithy and just: 'coniectu

ris non necessariis, id est,
malis,' Adv. Lat., p. 45.

Alterations MSS. (3) to take care that my emendation of the text. should be in keeping with Ovidian usage. While adhering to these three rules, I have, I flatter myself, been fortunate enough to relieve the text of several barbarisms.

The following is a complete list of deviations from Merkel's text, which are either proposed for the first time in this edition, or are defended on original grounds:

i. 1. haec for hanc.

i. 40. vigil for dolo.

ii. 100. negante data for negata meo.
iv. 86. militia for materia.

vi. 54. nauta-fui for causa-fuit.
vi. 55. iuvi for vidi.

vi. 100. cavet for favet.

vi. 118. dotales for res tales.

vi. 131. hanc, hanc, for hanc tamen.

vii.

45. quid non censeris for quod non verearis. vii. 71. ut tum for totum.

vii. 152. remque, or iamque, for hancque.
viii. 120. se for sic.

xii. 123. mersisset for misisset.

xii. 149. Cum clamore Pheres for cum minor

e pueris.

xii. 170. Et-abit for nec-habet.

xiii. 110. muta querella latens for multa querella tuis.

xiii. 122. refecta for referre.

xiv.

42. plena soporis for vina soporis.

1

See Corrigenda.

The majority of these readings approach Alterations more closely than those hitherto adopted to of the text. the best MS., and nearly all are easily deducible from it, according to established critical rules while in those instances where an arbitrary change has seemed necessary, that change has been as slight as possible, as, for instance, in the substitution cavet for favet, vi. 100. The only case in which I can be charged with audacity is in my conjecture on xii. 149. I have, however, introduced it into the text, which I should not have ventured to do had any reading previously suggested appeared even tolerable.

The above, with one or two other suggestions of less importance, constitute the sum of what I have been able to do for the text of the Heroides. I am indebted to Mr. Tyrrell for hac instead or hinc, in i. 103: and in vi. 156 will be found an an emendation of Lindemann's, which I looked upon as certain, until I saw Madvig's defence of the MS. reading, which is, however, substantially the same as regards meaning.

These are all the points of difference from Merkel's text due to modern scholars. Those derived from Heinsius, and the ancient commentators, are pointed out in the notes.

As Professor Madvig in his Adversaria Graeca, Madvig's published in 1871, had anticipated me in a very

See infra, p. xxxv.

emendations.