In a recent instalment of the Transactions of the Cambridge Philological Society (vol. IV, Part I) Dr Postgate has issued a pamphlet of eighty-two pages ‘on certain manuscripts of Propertius’. It includes a collation and discussion of cod. Holkhamicus 333, henceforth to be called L, a MS written in 1421 and containing nearly two-thirds of the author, from ii xxi 3 to the end; excerpts and briefer notices of several other codices; and a disputation on the value and relationship of Propertius' MSS in general. My name is scattered through the treatise, and I hasten to acknowledge the invariable benignity with which Dr Postgate reproves me, sometimes for doing what I have not done, and sometimes for doing what it was my bounden duty to do.

To make clear what follows, let me premise that hitherto the only MSS of Propertius which count for anything have been NAFDV with f and v the correctors of F and V: AF form one family (which Dr Postgate calls ϕ), DV another (which he calls Δ); the agreement of these two families is signified by the letter O ; N has something in common with each family but also something derived from neither, and this third element appears too in f and v.

All Ovid's works, except the amatory poems, are now equipped with a decent apparatus criticus. The apparatus to the amatory poems is no more decent than themselves: the three chief MSS containing them were collated by Keil in 1851: his collations were lent to three editors in succession, Merkel Riese and Ehwald, and remain unpublished to this day; for let no one fancy that what stands on pp. xiv–xvi and xx–xxii of Merkel's preface is anything but a string of excerpts. But Korn in the ex Ponto, Korn and Mr Riese in the metamorphoses, Mr Riese and Merkel in the fasti, Mr Ellis in the Ibis, Mr Owen in the tristia, Mr Kunz in the medicamina, Mr Sedlmayer in the heroides, Mr de Vries in the Sappho, have furnished full and exact collations of the principal MSS. Nothing is now lacking but an editor. But Nicolaus Heinsius is dead and buried; and Ovid, in spite of all this new material, is perhaps in a worse condition than he was two hundred years ago.

Merkel and his followers accomplish this result, not merely by depraving the text with a number of bad readings drawn from good MSS, but by two other methods, both efficacious: they expel the emendations of Heinsius, and they insert their own.

In 65 G and the old editors have petit altera et altera habebit which is un-Ovidian in metre and makes nothing fit to be called sense (‘my sister asks and my sister shall have’). It is altered by some to at altera habebat (‘my sister asks the boon but it was mine to give’), by others to at alter habebit (‘but another, i.e. Iason, will have it’): these changes mend nothing but the metre.

All this while the reading of P is alter petit alter habebit. This was commended long ago by Salmasius at Iul. Capit. Maximin. I ‘barbaro etiam patre et matre genitus, quorum alter e Gothis, alter ex Alanis genitus esse perhibetur’, and more lately by Mr Birt in the Goettingische gelehrte Anzeigen for 1882 p. 854 who says ‘bei der sentenziösen Form der Rede musste hier für altera petit nothwendig alter petit eintreten’; and it is printed by the three last editors Messrs Sedlmayer Ehwald and Palmer. The grammar is no doubt correct enough, but the sense is every whit as foolish as before. When you have said that A asks help for B you never add that the asker is one person and the recipient will be another: that is said already, and more than that.

Verse 38 is so given by P and G and most MSS: V (saec. XII) has ‘prima mihi uultus’. The sense is poor, and the repetition of ‘uultus’, first plural, then singular, is poorer. ∥ To all intents and purposes the verse has already been emended by Mr Palmer: ‘prima mihi uulnus nuntia fama tulit’: see the metaphor of the next distich. But in writing mihi Mr Palmer abandons better MSS for a worse, and in writing tulit he abandons all MSS: I would sooner follow them where they agree and desert them where they differ:

‘tui’ depends on ‘nuntia’. I suppose the archetype had

and fuit and mihi are alternative corrections of this manifest error.

From 38 to 145 all good ancient MSS fail us and leave us to the mercies of the 15 th century. Accordingly the very next verse is corrupt, oporteat is not even grammar; the oportuit actum or oportet ab arcu of Heinsius has no sufficient sense; Bentley rightly expels the couplet and proposes Apollinis for oporteat, but I think the original form of the interpolation can be recovered with less ado:

This is now the vulgate, since Madvig in 1831 expelled the gloss Sex. Clodio. Mr A. C. Clark however proposes further to expel duce and then to write cui mortuo unus instead of qui mortuus uno: another editor adopts the proposal, and I see in the March number of this Review, p. 119, that Mr S. G. Owen approves it.

Between qui mortuus uno and cui mortuo unus, so far as authority goes, there is nothing to choose. The MSS split their votes: qui mortuo unus H, cui mortuus uno E, cum mortuus uno T. The exchange of qui and cui is quite common; quite common too is metathesis of inflexion, not only in this simple form, Stat. silu. iii i 18 angusto bis seni, angusti bis seno, Aesch. supp. 373 〚369〛 ἀστοις … τωνδ∈, ἀστων … τοιδ∈, but also in stranger fashions, Ovid am. ii 5 27 Phoebo … Dianam, Phoebum … Dianae, Eur. Hipp. 331 αἰσχρων ἐσθλων αἰσχρά. The choice of reading therefore will depend on other considerations.

‘When a scholar of A. E. Housman's eminence has disposed of much of his work in periodicals, the convenience of scholars no less than respect for his memory commonly demands the publication of his Collected Papers.’ These are the opening words of A. S. F. Gow's biographical sketch, published in 1936, the year of Housman's death. The reason why the re-publication of Housman's papers has been delayed for over thirty-five years is to be found in a clause of his will:

For undertaking the publication of these papers in defiance of Housman's express wish we offer no apology, but we offer our reasons, and we offer them with all brevity. It is proper that Housman's friends and scholars nearer to him in time should have bound themselves by this prohibition; but posterity cannot be bound for ever. Housman's reputation is secure; and if he feared that the re-publication of earlier judgements later recanted might impair that reputation, or that the absence of an explicit recantation might be interpreted as a continued endorsement, then his fears were unfounded. And even if such fears had been well founded, littera scripta manet, nescit uox missa reuerti.

‘I am not a bird’ said the Irishman ‘to be in two places at once’; and it is another injustice to his distressful country that we call this speech a bull. The bird which is in two places at once is the Virgilian swan. Aeneas is bidden to behold these fowl alighting or alighted on the earth, and with the same breath is told that even as they are sporting together in the zenith, so are his scattered ships united in the harbour or the harbour-mouth. To evade the contradiction they propose to give the perfects in v. 398 the force which the perfect has for instance in ‘fuimus Troes’, and to interpret the line ‘and have ceased from circling the sky and from singing’. This interpretation is so obscure to its own inventors that they cannot agree when it was that the circling and singing took place, Ladewig and Wagner putting it before the swoop of the eagle, Weickert and Forbiger afterwards.

I said on p. 16 of this vol. of the Journal of Philology 〚this edition p. 40〛 that I feared some of the corrections there proposed had been forestalled by others. I am to blame that this is true of a larger number than one could wish, mainly through trusting to my memory of Burmann's notes instead of giving them a fresh perusal. I now make restitution: 1 xx 24 sacram Rutgersius, II ix 7 uisuram (uisurum is an error) Paley, xxviii 62 punctuated so by Postgate, xxxiv 12 posses tun and 40 irato Heinsius, III viii 12 haec Liuineius, xvi 21 cursus Markland, xvii 24 carpta Heinsius, xviii 21 manet Palmer, xxii 15 siqua et Heinsius, IV ii 12 credis id Postgate, vii 23 eunti Reland. The three living scholars will, I hope, accept my apologies.

Further, the following proposals have more or less in common with my own, and ought to be mentioned: I iii 37 nempe ibi Burmann, II viii 30 Teucros Passeratius, ix 12 apposito…Simoente Guietus, x 2 campum Maeonio and xxi 12 excepta Aesonia est Heinsius, III vi 28 exsuccis unguibus Burmann.

Let me here subjoin a few conjectures accidentally omitted from the paper of which I speak: II i 53 an in me for siue, II xxxii 9 quid iubet for cum uidet, III i 32 terra for Troia, III xix 17 more parentis for tempore matris, IV iv 83 ascensum monstrat dubio for mons erat ascensu dubiusy IV viii 13 fuerunt for fuerint.

We have already in this country a very good school edition of the Iphigenia in Tauris; but Prof. Flagg's book is also a good one and will be found useful even here, designed as it is for younger students than Mr England's, students who have hitherto read no Greek verse but Homer and are brought to a standstill by each fresh crasis or Doricism. There is therefore none of that ∥ critical matter which occupies Mr England's footnotes, but only an incomplete list of deviations from MS authority at the end of the volume. Mr Flagg has constructed his text with common sense, but yet I think that for boys Mr England's is the better. For scholars of course neither of the two is satisfactory, inasmuch as both are full of conjectures which have no probability at all; but for boys the matter stands otherwise. We cannot always give them what Euripides wrote, for the simple reason that we have it not: the next best thing then is to give them what he might have written, grammar and good sense; and this is what Mr England's freely amended text attempts to do. Mr Flagg's more conservative recension no doubt comes quite as near as Mr England's to what Euripides wrote, but it contains much more which Euripides not only did not write but could not; and the learner's conception of Greek is impaired accordingly.

I see no hope of completing a presentable commentary on Propertius within the next ten years; but in the mean time I trust that the following list of corrections may be found of service to scholars. For my own sake too I have some desire to put my conjectures on record, as I am for ever seeing them forestalled by other students: Mr Konrad Rossberg in vol. 127 of Fleckeisen's annual has bereft me of no less than nine. True, it is agreeable enough to have one's results confirmed by a scholar who stands next to Mr Baehrens and Mr Palmer at the head of living Propertian critics; but I should like to retain something of my own. As many readers are apt to fancy that the textual critic proposes alterations out of pure gaiety of heart and not because the vulgate wants altering, I have added an examination in detail of the first elegy; ‘ne mea dona tibi studio disposta fideli, Intellecta prius quam sint, contempta relinquas’. I employ Mr Baehrens' MSS and notation.

I i. Between 11 and 12 are lost two such verses as these: ‘multaque desertis fleuerat arboribus, ∣ et modo submissa casses ceruice ferebat’. ∥

I i 23 tune ego crediderim uobis et sidera et amnes ∣ posse Cytinaeis ducere carminibus] et manes et sidera uobis.

Thanks to Wecklein it is at length possible to study Aeschylus in comfort. Next to an accurate collation of the cardinal MSS, a complete register of the conjectures of critics is the student's prime requisite. Nothing short of a complete register will serve: no man can be trusted to sift good from bad: some editors do not know a correction when they see one, others through childish jealousy of this scholar or that ignore his discoveries, the most candid and the soundest judgment is human and errs. The time lost, the tissues wasted, in doing anew the brainwork done before by others, and all for lack of a book like Wecklein's Appendix, are in our brief irreparable life disheartening to think of.

In the ensuing pages I have not set down all or nearly all the corrections which I imagine myself to have made in the Agamemnon: I know how easily one is satisfied with one's own conjectures. I have arraigned the MSS only where their delinquencies can be made as clear as daylight, and I have proposed only corrections which I think may possibly convince others as well as myself. For instance, however confident I may feel that in v. 17 Aeschylus wrote not ὔπνου but πόνου, still I have to own that the former can by hook or by crook be defended, and that the indications which suggest the latter are not decisive; so I leave the reader in peace.

If I recur to this passage it is not so much out of inordinate affection for my own conjectures as because the discussion started by Dr Postgate on p. 275 〚CR 3 (1889), 275〛 may prove interesting to pursue. The ‘white liver’ of the coward as well as the ‘black heart’ of the traitor was present to my mind when I objected to ‘intus palleat’, but I think there is a difference. Cowardice and treachery are qualities, and inherent in the man: the bloodlessness attributed to the craven is with him from his birth and, to be prosaic, may be expected to reveal itself at a post-mortem examination; and so too the traitor's black heart. But I think it is otherwise when one has to speak not of a quality but of an emotion, as here in Persius of guilty fears: emotions may be held to cause by their presence some such internal disorder as the flight of blood from an inward part; but they come and go, and they all depart with life. The difference is of this sort: in the darkness of night a white rose may be called a white rose still, but can a face in the darkness of night be said to turn pale?

24 and 25 are found neither in P nor in G nor in more than a few of the other MSS.

The archetype itself contained many interpolated verses, which appear accordingly in P and G and all the rest. But some of the later MSS proffer new interpolations, from which P and G and many of the others are free. I here enquire whether, in spite of this fact, any of the later MSS preserve genuine verses which have been omitted by P and G.

Some of the inserted lines betray their spuriousness plainly in language or metre, as v 26 ‘est in qua nostri littera scripta memor’ and IV 132 sq. ‘Saturnus periit, perierunt et sua regna: ∣ sub Ioue nunc mundus; iura Iouis sequere’: such as these I leave alone. Nor shall I here discuss the couplets with which many MSS have filled up real or imaginary gaps at the opening of certain epistles. But I shall examine five places in the body of the poems where later MSS offer verses which are missing from the oldest.

First VIII (Hermione) 19 sqq.

The eyes of a homicide transformed into a wolf may well be said to flagrare, ardere, scintillare or the like, but ‘lucent’ is quite inadequate and should I suspect be replaced by ‘lurent’. The adjective ‘luridus’ is of course common, the substantive ‘luror’ very rare, the verb ‘lureo’ unknown to lexicographers: once however it has been preserved, and that once in the metamorphoses. At 11 776 our MSS give ‘nusquam recta acies, liuent rubigine dentes’; but this verse is found by Mr Ellis in a Bodleian codex of excerpts (MS Digby 65, saec. XII) with ‘lurent’ for ‘liuent’, and not only is it intrinsically improbable that the unique word should be a corruption of the common one, but ‘lurent’ is supported by Hor. carm. iv 13 10 ‘luridi dentes’. Again in Plaut. Menaech. 828 ‘uiden tu illi oculos uirere? ut uiridis exoritur colos’ cet., where ‘uirere’ cannot well be right with ‘uiridis’ following and one MS gives ‘iurere’, Ritschl has conjectured ‘lurere’. Once more, in Sen. Herc. fur. 766 sq. Charon is thus described: ‘inpexa pendet barba, deformem sinum ∣ nodus coercet, concauae lucent genae’ (for ‘lucent’ one family of MSS has ‘squalent’, but that comes from ‘squalidus’ in 765): here again the verb is miserably weak, especially if compared with Virgil's ‘stant lumina flamma’, and here too I propose ‘lurent’.