days of printing editors have successively been trying to find the 'missing word': e.g. calamitas, damnum, pernicies, clades, pestis, malum, have all been suggested; and in the year of grace 1911 H. J. Mueller solemnly and finally accepts for his revision of Weissenborn's text malum from M. Mueller. Our British Museum Harl. 2781 (Luchs' 8) long ago (in the fifteenth century) rightly omitted quam. But Luchs never examined ẞ in Book XXVI.

VII. c. 17. 12-c. 18. 1. In secundo interregno orta contentio est quod etc.... Cum intercedendo tribuni nihil aliud quam ut differrent comitia ualuissent, duo patricii consules creati sunt, C. Sulpicius Peticus tertium M. Valerius Publicola eodemque die magistratum inierunt CCCC anno quam urbs Romana condita erat, XXXV quam a Gallis reciperata, ablato post XI annum a plebe consulatu patricii consules ambo ex interregno magistratum iniere C. Sulpicius Peticus tertium M. Valerius Publicola.

So all the MSS., save that, as Glareanus saw, they (or the Symmachi ?) have corrupted XI into IX, a common fault. Peticus and Publicola were no doubt most honourable men; but what did they do that Livy should record precisely the same harmless facts about them twice over in seven lines? To ask the question is to answer it. Livy's sentence ended at a plebe consulatu; all that follows is merely a marginal summary, highly correct, which has climbed into the text and been kept there with pathetic patience for at least ten centuriesmaintained partly by the prudence of Gruter, who brought down his chopper1 between cc. 17 and 18 at the word inierunt; but chiefly, no doubt, by the reverent not to say somnolent eye with which such names are read. If any reader needs evidence of the frequency with which these summaries are foisted on Livy, let him refer to our note on 2. 36. 3 (Vorm. M), where he will find one example in full and references to nine others. In the Preface to our forthcoming edition of Books VI.-X. Professor Walters has collected a mass of examples from the part of M written mainly by the scribe whom we call Tertius (see our Preface to Books I.-V., § 9).

VII. 24. 4. consul uolnere alligato reuectus ad prima signa.

So rightly all editors, with P2 and, as we now find, ODAF3. But it is worth while to note the beautiful shape in which the earlier manuscripts give the participles, because it shows how close those MSS. stand to the undivided text of their uncial predecessors: alligatore uectus' M, 'alligator euectus' PFHT. What meaning, if any, did they attach to this alligator ?

BOOK VIII.

VIII. 7. 16-19. In the brief and terrible speech of Titus Manlius there are in the current text two corruptions which appreciably diminish both the fineness of delineation of the father's character and the tense feeling embodied in an

1 See our Preface to Lib. I.-V. § 5. p. x. The edition of 1612 there mentioned is in the

possession of Professor Walters.

utterance which has not a single superfluous word. It is best to consider it in full; we print what we believe Livy wrote:

Quandoque inquit, 'tu, T. Manli, neque imperium consulare neque malestatem patriam ueritus, aduersus edictum nostrum extra ordinem in hostem pugnasti et, quantum in te fuit, disciplinam militarem, qua stetit ad hane diem Romana res, soluisti meque in eam necessitatem adduxisti ut aut rei publicae mihi aut mei obliuiscendum sit, nos potius nostro delicto plectemur quam res publica tanto suo damno nostra peccata luat; triste exemplam sed in posterum salubre iuuentuti erimus. Me quidem cum ingenita caritas liberum tum specimen istud uirtutis deceptum uana imagine decoris in te mouet; sed cum aut morte tua sancienda sint consulum imperia aut impunitate in perpetuum abroganda, nec te quidem, si quid in te nostri sanguinis est, recusare censeam quin disciplinam militarem culpa tua prolapsam poena restituas―i, lictor, deliga ad palum.'

lu 16 Manlius states the conflict between his public and his private duty, one of which must be forgotten'; and the single word in which he expresses the second claim is mei, deliberately chosen by Livy, we believe, to tepresent on the one hand the completeness of the sacrifice which the consul makes to his country of his own paternal affection, and on the other the ogotism deeply inherent in his character-the whole of Ibsen's Brand in one speech. For what Manlius has no time to consider, earning by his blindness. the tolong detestation of all young men (c. 12. 1), is the right of his boyish sou as a separate person to an equitable judgment in the light of all the circumstances, and the loss to his country which the extinction of such individual promise would involve. Infelix, utcumque ferent ea facta minores.

but this dramatic intensity passed the comprehension of our scribes, and, unhappily, of our modern editors, though the latter might have judged more soundly it they had had the manuscript evidence fully before them. In the content text atter wei is inserted meorumque, though this is given by no good excript, only by Up and FDA with dett. aliquot, i.e. there is only one www to pectable authority, D3. What have the rest? meorumue H, bout my M. COLDL, and therefore quite certainly in the Nicomachean archewww gly cow. Now we may agree with H and D3 that the asyndeton is quite out of place here; but what is its origin? Simply that the scribo or reader thought he could improve on Livy by remedying ou of a reference to Manlius' son and the rest of his kin, an omission kle, the scribe, had the heart to feel but not the wit to interpret. He cow wrote morio above mei, and his faithful successors copied both copy a tuto the text, as they have done in hundreds of other cases.

can glad of such famous scenes
Handal for his first great
14), the interview of
Nucleus (4. 49.4), and the
his daughter and to

case will be well seen if it be compared with the speech put into Verginius' lips by no less an orator than Lord Macaulay in his Lay of Virginia.

to mention no others. The § 26.
Lovg'a tostratat in the last

See the Preface to Vol. I. of our edition,

In § 18 an emendation of Jakob Gronov has been rather hastily adopted by all subsequent editors at the climax of the scene. All the MSS. (MPFUþOHTDLA) read nec te quidem, i.e. et ne te quidem (as often e.g. Cic. de Sen. 9. 27); this makes excellent sense if we connect censeam with the preceding sint as equally dependent upon cum, and gives a dramatic change at the end which represents Manlius as turning from his son to the lictor with the fatal command. That the decision should be made known by the command itself and not previously announced to his son is surely demanded by every sentiment, ancient and modern; and it is surely not less certain that Livy would put the close of such a story not as a short isolated sentence, but as the climax of a period. The change of address is parallel to that in the familiar narrative in the Synoptics of the Healing of the Paralytic: 'but that ye may know that the Son of Man hath power on earth to forgive sins,'-He saith unto the sick of the palsy, 'I say unto thee, Arise.'

Yet to avoid this absolutely natural break, Jacob Gronov alters nec into ne, and puts a full-stop at restituas. This makes the main verb of the whole stately period nothing better than the feeble censeam, which, since we are now to regard it as independent, can mean nothing but I should suppose, I am inclined to think,' either faltering or ironical, and in either case totally repugnant to the situation and to the character of Manlius.

VIII. 8. 3-8. LIVY'S ACCOUNT OF THE ROMAN ARMY.

The text according to the MSS., if we disregard minor matters, runs thus :

[3] Clipeis antea Romani usi sunt; dein, postquam stipendiarii facti sunt, scuta pro clipeis fecere; et quod antea phalanges similes Macedonicis, hoc postea manipulatim structa acies coepit esse: postremo in plures ordines instruebantur. [4] ordo sexagenos milites, duos centuriones, uexillarium unum habebat. [5] prima acies hastati erant, manipuli quindecim, distantes inter se modicum spatium. manipulus leues uicenos milites, aliam turbam scutatorum habebat; leues autem, qui hastam tantum gaesaque gererent, uocabantur. [6] haec prima frons in acie florem iuuenum pubescentium ad militiam habebat. robustior inde aetas totidem manipulorum, quibus principibus est nomen, hos sequebantur, scutati omnes, insignibus maxime armis. [7] hoc triginta manipulorum agmen antepilanos appellabant, quia sub signis iam alii quindecim ordines locabantur, ex quibus ordo unusquisque tres partes habebat-earum unamquamque primum pilum uocabant. tribus ex uexillis constabat. [8] uexillum centum octoginta sex homines erant. primum uexillum triarios ducebat, ueteranum militem spectatae uirtutis, secundum rorarios, minus roboris aetate factisque, tertium accensos, minimae fiduciae manum: eo et in postremam aciem reiciebantur.

In the last two sections Livy is represented as saying (1) that each of the three parts of each of the fifteen ordines was called primus pilus, so that each legion had 45 'first companies'; (2) that each of the ordines had, on a strict

interpretation 9 uexilla, or by straining the order and supplying ordo, not pars as the subject of constabat, 3 uexilla; and (3) even on the latter hypothesis, that each ordo contained (3 × 186=558) men, which, with 15 ordines, gives a modest total of 8370 men for this part of the legion alone; whereas in § 14 we learn that scribebantur . . . legiones quinis milibus peditum.

The passage may almost be called a locus desperatus. Every modern editor rejects at least primum in § 7 and uexillum in § 8. Weissenborn further deletes the whole of § 4 (ordo . . . habebat); Mommsen, drastic as usual, cuts out three sentences bodily (§ 7 earum. homines erant § 8); and neither troubles to hint at any reason which could prompt anyone to invent these complex statements which no ingenuity can harmonise even with what is told us in the rest of the chapter, let alone what we know from other places and sources. Even more prudent critics like Madvig and Luterbacher, who are content to excise only the two words in §§ 7 and 8, have first to suppose not merely ignorance but an extraordinary confusion on Livy's part, so that what he first called manipulatim he then described as changed in plures ordines; and then proceeded to use ordo, the very next word, to mean manipulus; and then in § 7 used ordo in a totally different sense ;-and they have then to assume two senseless interpolations on the part of some early scribe or scribes which had the extraordinary fortune of being both planted into the text of the archetype. The wilder excisions are merely signals of distress from benighted folk

crying for the light,

And with no language but a cry.'

Yet on the circumstantial evidence of signs still visible in the manuscripts but hitherto entirely unnoted, we believe that the trouble has sprung from a transposition (of 57 letters, i.e. of three lines) in the archetype, due to an obvious cause and paralleled by at least a score of other examples. If this be granted us, with one other change (of u to a in one word) long since conjectured by Lipsius, we think we can show that Livy's account differed from what Polybius' description of the legion of the Second Century B.C. renders probable for that of the Fourth only in two points, both of a perfectly simple character, and long since pointed out. In one of these Livy may, in the other he must, have been misled, misinterpreting, we may guess, a too compendious statement in his authority. The result was that he could not make his account complete; but so far as it goes, he is, if our hypothesis be granted, quite intelligible and consistent with himself, and honestly silent only about details in which his predecessors seemed to him obscure. This feature in Livy's method ought to be generally recognised. Examples are 2. 11. 9 (where see our note); 3. 24. 8-9 (extremum anni); 2. 39. 9; 2. 51. 1; 5. 19. I (these three with iam, see the note on 2. 39. 9 in Professor Conway's edition of that Book); 7. 25. 10 (redierat, where the pluperfect covers a gap in the narrative, see Weissenb. ad loc.); 21. 31. 7 (arbiter factus, without any statement from whom the invitation came).

We believe that the words of § 4 ordo . . . habebat which Weissenborn rightly despaired of interpreting where they stand, belong to § 8 after constabat, to which ordo is the subject; and that the word uexillum, which is unintelligible following constabat, should become the subject of habebat, the last word of our transferred clause; and that ordo has only one meaning, namely one of the fifteen segments of the third division of the army, segments separated from one another by lines at right angles to the front. With these changes the passage runs thus:

[3] Clipeis antea Romani usi sunt; dein, postquam stipendiarii facti sunt, scuta pro clipeis fecere; et quod antea phalanges similes Macedonicis, hoc postea manipulatim structa acies coepit esse: postremo in plures ordines instruebantur. [5] prima acies hastati erant, manipuli quindecim, distantes inter se modicum spatium. manipulus leues uicenos milites, aliam turbam scutatorum habebat; leues autem, qui hastam tantum gaesaque gererent, uocabantur. [6] haec prima frons in acie florem iuuenum pubescentium ad militiam habebat. robustior inde aetas totidem manipulorum, quibus principibus est nomen, hos sequebantur, scutati omnes, insignibus maxime armis. [7] hoc triginta manipulorum agmen antepilanos appellabant, quia sub signis iam alii quindecim ordines locabantur, ex quibus ordo unusquisque tres partes habebat -earum unamquamque primam pilum uocabant. [8] tribus ex uexillis constabat ordo; sexagenos milites, duos centuriones, uexillarium unum habebat uexillum; centum octoginta sex homines erant. primum uexillum triarios ducebat, ueteranum militem spectatae uirtutis, secundum rorarios, minus roboris aetate factisque, tertium accensos, minimae fiduciae manum: eo et in postremam aciem reiciebantur.

If the passage had run thus in the MSS. we venture to think that such difficulties as remain would have caused but little trouble. Before discussing these it will be best to note what causes could have led to such a transposition and what evidence the MSS. contain in favour of supposing that it has taken place. But we had better deal at once with one apparently fatal objection to the new arrangement which will probably have caught the reader's eye. In § 8 if each uexillum has 60 men, 2 centurions and 1 uexillarius, why is the total for 3 uexilla only 186, not 189? Because the uexillarius was one of the 60 men, the 2 centurions being chosen before the uexillum was formed (Polyb. 6.24. 3), the uexillarii by the centurions themselves afterwards (id. ib. § 6 EK TŵV KAтaλeiñoμévwv). Livy may1 have thought this detail too well known to need any explanation to his readers; it was apparent to any child who had ever looked on at a levy in the Campus Martius while the centurions were helping the military tribunes to pick their men: μετὰ τῶν ταξιάρχων διεῖλον τὰς ἡλικίας (οἱ χιλίαρχοι), as Polybius (§ 3) explains.

We suppose that the beginning of § 8 stood thus in the uncial archetype

1 Professor Postgate shakes his head, and would prefer to alter the VI. to IX.