EDITED BY F. W. HALL, M.A. AND R. HACKFORTH, M.A. BOARD OF MANAGEMENT: Prof. J. F. DOBSON, M.A., Chairman. Prof, A. C. CLARK, Litt.D., F.B.A., Hon. Treasurer. L. WHIBLEY, M.A., Hon. Secretary. Prof. R. S. CONWAY , , E. A. BARBER, M.A With the Chairman, Trea- Litt.D., F.B.A, suver, and Secretary, repre- Representing the Oxford Philological Society. Prof. H. WILLIAMSON, | Srting the Council of the | 1s ROBERTSON, M.A., M.A, — Representing the Cambridge Philological Society. With the co-operation of Prof. WM. GARDNER HALE, University or Cuicaco; Prof. MAURICE HUTTON, Toronto; Prof. T. G. TUCKER, Litt.D., Metsourne. VOLUME XxIl PUBLISHED FOR THE CLASSICAL ASSOCIATION BY JOHN MURRAY, ALBEMARLE STREET, LONDON, W. AND G. E. STECHERT & CO., 31-33, EAST toro STREET, NEW YORK 1928 Prosoc The V ‘ Imita Lu Notes Horac Miscel The The S| The Ilapact The S! Callim: Nicand Notes | Summa Lz esese eee eeer seers eer er e Se PRINTED IN GREAT BRITAIN BY BILLING AND SONS, LTD., GUILDFORD AND ESHER > ¥ . : at 4 ap , | O. I. PAGE Prosody and Method II. A. E. Housman ‘ ‘ I The Vatican Plato. L.A. Post ‘Imitation’ in Plato's Republic. J. Tare. 4% + 16 Lucan, Statius, Juvenal in Early Centuries. H. J. BHOMSON . : ; - = Notes on Plotinus (II.). J. H.Stezman.. The Hymn to Hermes. T. L. Acar. Hedonism in Plato’s Protagovas. R.HackrortH . More Facts about Oldest Latin MSS. E. A, Lowe. Summanes of Periodicals : No. 2. Horace and the Moral Function of Poetry. J. Tate The Egquttes Illustves. H. Hitt . The Spondeion Scale. R. P. WINNINGTON-INGRAM . The Pertplus Maris Evythraa. M. P. CHARLESWORTH The Shorter Glosses of Placidus. J. W. Piriz -Callimachus and Others. J. U. Power . Nicander’s Signature, Questions without Answers. E. Loser . Notes on Festus and Terence. W.M. Linpsay _. Summaries of Periodtcals : . é TABLE OF CONTENTS Nos. 38, 4. The Parmenides of Plato and the Origin of the Neoplatonic ‘One.’ E.R. Dopps 129 The Seventh and Eighth Platonic Epistles. J. HaARwarp 143 : Alexander of Abonuteichos. A. D. Nock. 160 eS ; a The Pseudo-Aristotelian Problems: Their Nature and Composition. E. S. The Historia Augusta: Its Date and Purpose. Norman H. Baynes. . 166 Tacitus, Htstores I. 13. H.C. NuTTING . Zu P. Oxy. 414. S. Luria ; . 176 Martial V. xvii. 4. W.M. Linpsay . , . Miscellanea II. T. W. ALLEN . . 203 Karl Marx on Greek Atomism. CyriL BAILEY 205 OYAON KEKAHTONTE>. R. McKenzie . 206 Aquilo, the Black Wind. R. McKenzie . . 207 Summamnies of Peniodicals : Language : oe PA ; Bor a The CLASSICAL QUARTERLY and CLASSICAL REVIEW are the Organs of the Classical Association. The QUARTERLY is pemnenneg in January, April, and October, the last issue being a double number; the REVIEW in February, May, July, September, November, and December. THE CLASSICAL UARTERLY EDITED BY _F. W. HALL, M.A., St. John’s College, Oxford. R. HACKFORTH, M.A., Sidney Sussex College, Cambridge. ue) Board of Management ; ake Prof. J. F. DOBSON, M.A., Chairman. Prof, A. 0. OLARK, Litt.D., P.B.A., Hon. Treasurer. L, WHIBLEY, M.A. Hon, Secretary, With the cea © vea- | J. BELL, M. Prof, RB. 8. CONWAY. surer, and S. ie Philological Society. Litt.b., P.B.a. Serer ing fhe Council the | 8. ROBERTSON, M. Prof. H. WILLIAMSON, M.A. } Classical Association Representing the PCat Philological Society. With the co-operation of Prof. Wm. GARDNER HALE, University of Chicago; Prof. Maurice Hutton, Toronto; Prof. T. G. Tucker, Litt.D., Melbourne, Vol. XXII. JANUARY, 10928 Il. A. E. Housman THe VaTican PLATO L. A. Post ‘ImiTaTion’ 1n Piato’s Repuszic Jj. Tate Lucan, Stratus, JUVENAL in Earty C H. J. THomson Nores on Piotinus (I1.). J. H. Tue Hymn to Hermes . .T. L. Acar HEDONISM IN Prato’s Proracoras . . R. Hackrortu More Facts asovut Orpest Latin MSS. E. A. Lowe SuMMARIES OF PERIODICALS Lonpon: JOHN MURRAY, ALBEMARLE STREET, W. New Yor«: G. E. STECHERT & CO., 31-33, EAST torn STREET. Price for single numbers, 4/- net, or 4/2 post free. Yearly Subscription 16/- post free ; U.S.A. $4. Comb Yearly Su oo for the CLASSICAL QUARTERLY and the C ICAL REVIEW, 25/- post free; U.S.A. $6.25. | | | THE CLASSICAL ASSOCIATION ; THE objects of the Classical Association are to promote the development an maintain the well-being of classical studies, and in particular (a) to impress upon public opinion the claim of such studies to an eminent place in the national scheme of education; (0) to improve the practice of classical teaching; (c) to encourage investigation and call attention to new discoveries; (d) to create opportunities for intercourse among lovers of classical learning. Membership of the Association is open te men and women alike. The annual subscription is 5s. (life composition, £3 15s.), and there is an entrance fee of §s. (not charged to Libraries). Members receive a copy of the annual Proceedings of the Association and, on payment of 2s. 6d., of The Year's Work in Classical Studies (both post free). They may also obtain the Classical Review and Classical uarterly at reduced prices, provided that the subscriptions be paid before anuary 31st in each year. Subscriptions sent in later than that date must at the rates offered to the general public. Inquiries and applications for membership should be addressed to the Hon. Treasurer, Mr. H. F. Hose, Dulwich College, S.E.21; or to either of the Hon. Secretaries, Dr. E. Norman Gardiner, 186, Woodstock Road, Oxford, and Miss E. C. Gedge, Crayford Rectory, Kent; or to the Hon. Secretary of any of the district Branches—viz., Mr. F. H. Sandbach, The University, Manchester; Dr. G. A. Auden, Education Office, Margaret Street, Birmingham; Mr. R. B. Onians, The University, Liverpool; Professor Granger, University College, Nottingham; Mr. H. A. Pitman, 4, Prince’s Buildings, Bristol; Mr. Basil Anderton, The Public Library, New Bridge Street, Newcastle-upon-Tyne; Miss E. J. Vaughan, Westfield College, Hampstead; Mr. L. J. D. Richardson, Fernshaw, Hilary Road, Llanishen, Cardiff; Mr. E. E. Bibby, The University, Leeds; Mrs. Jackson, The Cottage, Cordwell Valley, Holmesfield, Nr. Sheffield; Mr. E. D. T. Jenkins, University College, Aberystwyth ; Miss M. E. Lees, University College, Bangor; Miss M. K. Wakefield, Maidstone Girls’ Grammar School, Kent; Mr. S. Price, 33, Walton Well Road, Oxford; Mr. J. M. Gaussen, 16, Lewes Crescent, Brighton ; Mr. G. W. Dyson, University College, Southampton; Mr. F. Fletcher, University College, Exeter; Mr. C. W. Baty, 49, Waterloo Road, Bedford; and Mr. J. E. Scott, Gonville and Caius College, Cambridge. | The Society for the Promotion of Roman Studies 19 BLOOMSBURY SQUARE, LONDON, W.C. 1 President: G. MACDONALD, C.B., LL.D., F.B.A. Malan Society embraces the history, art, and archaeology of Rome, Italy, and the Roman Empire, down to about A.D. 700. In connexion with the Hellenic Society, the Roman Society maintains a joint library of Greek and Roman art, archaeology, and history, and a collection of lantern slides and photographs. Members are entitled to borrow books and slides, and these can be sent through the — Communications respecting books and slides should be made to the Librarian at 19 Bloomsbury Square. Afternoon meetings for the reading and discussion of papers are held at Burlington House, Piccadilly, W.1. Notices of these are sent to all members. Annual Subscription one guinea. Life membership composition for persons over fifty years of age ten guineas. For others fifteen guineas. There is at present no entrance fee. Persons desirous of joining the Society are asked to communicate with the Secretary, Miss M. V. Taylor, at the Haverfield Library, Ashmolean Museum, Oxford. _ ee THE JOURNAL OF ROMAN STUDIES (11 X 7% inches) is published by the Society in half-yearly parts, and sent post free to all members. Vols. I.-XI. include papers by English, American, Danish, Russian, French, and Italian scholars, and, amongst others: ANDERSON, J. G. C. Festivals of Mén Askaénos. HAVERFIELD, F., and STUART-JONES, H. Romano- ASHBY, THOMAS. Roman Malta. British Art. ASHBY, THOMAS, and RAL FELL. The Via Flaminia. MACDONALD, G._ The Corbridge Gold Coins. The Bodleian MS. of Pirro Ligorio. MACDONALD, G. The Agricolan Occupation of North BAGNANI, GILBERT. The Subterranean Basilica at Porta Britain. : iore. MACKAIL, J. W. Virgil and Roman Studies. BON!, Comm. Recent Discoveries on the Palatine RAMSAY, Sir W. M. Colonia Caesarea. Hill REID, Prof. J.S. Roman Public Law. Bury, J. B. The Notitia Dignitatum. REID, Prof. J.S. Problems of the Second Punic War. CALDER, W. M. Roman Inscriptions in Asia Minor. REtNACH, SALOMON. A Bronze Statuette of Zeus. » CHERBSMAN, G.L. The Family of the Caristanii. RUSHFORTH. G. MCN. Magister Gregorius de Mira- — COLLINGWOOD, R. G. Hadrian's Wall; a History of bilibus Urbis Romee. i the Problem. STRONG, Mrs.S. A. The Archaeological Exhibition of Cur_s, J. Roman and Native Remains in Caledonia. 1911 at Rome. Dessau, Prof. A Roman Senator under Domitian and | VAN BUREN, A. W. Varro's Aviary at Casinum. Trajen. VAN BurREN, A. W. The Ara Pacis ESDAILE, Mrs. A Statue in the Palazzo Barberini. WARDE FOWLER, W. The original meaning of the word Harpy, E. G. Notable ludicie Populi on Capital * Sacer.” Charges. WARDE FOWLER, W. ~Mundus Patet.” . HAVBERFIELD, F. Roman London. WHEELER, R. E. MORTIMER. Roman Colchester, Vol. XII. isin active preparation. | : ¢ Ic outset its irre! uel gna stagnum illegitin like ns Lucian porrigit ning of obey th though before I the cou! authorit Hexame gn mach irrt man He the Prisciar staten THE CLASSICAL QUARTERLY JANUARY, 1928. PROSODY AND METHOD II. THE METRICAL PROPERTIES OF GN. I CHOOSE the word metrical rather than prosodical, to make it plain at the outset that I am not concerned with the rule in Priscian—not of Priscian, for its irrelevance is sufficient proof of that—G.L.K. II p. 82 7-9 ‘ gnus quoque uel gna uel gnum terminantia longam habent uocalem paenultimam, ut regnum stagnum benignus malignus abiegnus priuignus Pelignus’, still less with the illegitimate inference sometimes drawn from it, that this pair of consonants, like ns and uf, lengthened a short vowel whenever they followed it. The present dispute is not about vowels but about syllables; the power of gn to make position, as they say, like st or x, and so to impede the flow of verse. Sixty years ago it might have been thought that this question was settled. Lucian Mueller de v. m. pp. 315 ff. said of gu in the body of a word ‘ semper porrigitur dactylicis correpta uocalis, si sequatur gv’, and of gu at the begin- ning of a word ‘ gw succedens longam efficit priorem’. Editors did not always obey this rule, but most of them obeyed it when they remembered it. Now, though some obey it still, others not only disobey it but try to overturn it; and before I go on to set forth the facts of the case and, as I hope, to reason upon them, I will give some examples of the strategy and tactics of the rebels. I take for a starting-point H. Magnus’s notes on Ouid. met. VI 468, where he prints with most MSS ad mandata Prégnes adding ‘de positione gn cf. Comm. Cat. p. 193 sq. Friedrich’, and on X 531, where he prints piscosamqué Gnidon adding ‘cf. ad Catullum 36 13 Friedrich’. Since it thus appears that the countrymen of Lachmann and Meineke regard Mr Gustav Friedrich as an authority on prosody, to that fount of doctrine let us go. In his note on Catull. 36 13 Ancona Gnidumque (alternatively described as Comm. Cat. p. 193 sq.) Mr Friedrich has the following remarks. Das iberlieferte Guidum muss dem C, gelassen werden, ebenso wie guatis in dem Hexameter 64, 298 inde pater diuom sancta cum coniuge gnatisque. Wenn man glaubt, gn mache im Unterschied von den andern Fallen der positio debilis stets Position, so irrt man. \He then rightly observes that the statement which I just now quoted from Priscian applies ostensibly only to the terminations guus gna gnum and is not ‘a statement that gu makes position, and he proceeds thus : Ob aber die Romer ignts oder ignis sprachen, ist nicht sicher. Jedenfalis finden B wir bei Ter Eunuch, 777 quid ignaue und C I G 1, 1060 KOFNITOY. NO. I. VOL. XXII. A ¥ 2 A. E. HOUSMAN This is very incautious, and damaging to his own case. If the vowel in ignis was short, the fact that in verse the syllable is always long must be due to gn always making position. The irrelevance of citing quid ignare from a scenic poet will be best exposed by citing tb: euenit from Plaut. merc. 774- Natiirlich machte gm noch viel weniger Position in Fallen wie c. efigr. 1076, wo sich unter lauter prosodisch tadellosen Versen der Hexameter findet : paene immatura morte eveptam stht gnatam, I transcribe a few of these verses, that the reader may admire with Mr Friedrich their ‘ quite faultless prosody ’. nam postquam et monumento hoc condecorauit. That is an hexameter. heic sica quas rapit mortis acerba quis. That is a pentameter. heu quantum mater in biciociost. That is the pentameter belonging to Mr Friedrich’s hexameter. Ebenso 986, 11 quem wos imspicere et uestris ostendeve gnatets. Yes, in a piece which begins with the hexameter gui legis et flores utae car- mina pia. Selbstverstandlich wird von der Regel Priscians auch das aus dem Griechischen stammende cygnus nicht getroffen, noch weniger Progne. Mart. 9, 103, 2 quae capta est alio nuda Lacaena cygno. So ist einhellig iberliefert,— False: the oldest MS (E saec. X) and the family (vy) to which it belongs give cycno, as Lachmann at Lucr. III 7 required. But Mr Friedrich when he edited Catullus had never seen any edition of Martial but Friedlaender’s, and did not know that Friedlaender’s apparatus criticus was a snare and a delusion. — und so schreibt Friedlander, indem er anfihrt: Fragm. Bobiense (G L 7, p. 539 K): mn rv liquidae uocantur, quia in metro saepe defictant, ut—nuda Lacaena cygno. This passage, G. L. K. VII pp. 538 f., from cod. Vindob. 16 (saec. VIII- IX), which Mr Friedrich copies out of Friedlaender with all Friedlaender’s errors and omissions, really runs thus: ‘/ m m r liquidae uocantur, quia in metro saepe deficiunt (not -ant): /, ut neue flagella, m, ut distincta smaragdo, n, ut nuda Lacena cigno (not Lacaena cygno ), r, ut ne desere frater’. All therefore that the grammarian attests in the word is the : it is only the MS which attests g for c, and the MS attests equally ¢ for y, which nobody, not even Mr Friedrich, will accept, because nobody has a motive for accepting it. He then transcribes from Friedlaender, without attempting to verify them the following references : Auson Petror which ing tw V cygnur let us magne A false Tollius instead imperf not tha with tl refutins fresh as figment reading century and the Last cc pure in gave Pr if not a no reco! the hum TO, 531 | ebensow the two discours Sen. Hor. sat. Not Meineke, that they satires to Es verste diuom sanc —this se my enqui The < big « | | } . wel in ust be ignare Plaut. 176, wo mmatura ith Mr yiae car- chischen uae capta belongs when he er’s, and jelusion. G kL & ena cygno. ce. VIII- laender’s , quia in wagdo, n, therefore {S which even Mr ‘ify them || my enquiry. PROSODY AND METHOD II. 3 Auson. epp. 20, 8 quid vefert cornix non ideo ante cygnum; Mosella 316 mdgnetis ; Petron. c, 131 Progne ; Val. Fl. 3, 34 Prognesson ; which Friedlaender himself had similarly borrowed, verifying none and falsify- ing two, from Markland’s note on Stat. silu. V 3 222. What Ausonius wrote is no matter, and anyone who pleases may add cygnum 410 8 to nosceré yv@O 320 I9 and his many other misdemeanours; but let us do what neither Friedlaender nor Friedrich has done and try to find mdagnetis in Mosell. 316. That verse is this: spirat enim tecti testudine chorus achates. A false reference, you may think. No, nothing so venial as a false reference. Tollius, and I daresay other editors of the 17% and 18% centuries, printed, instead of chorus achates, the conjecture of that great scholar and critic but imperfect metrist Gronouius, ueva Magnetis; Markland, whose prosody was not that of a Bentley or a Dawes, accepted this without suspicion ; Friedlaender, with the negligence natural to one engaged in so pleasant and easy a task as refuting Lachmann, copied it cheerfully down; and here it starts up again as fresh as a daisy in Mr Friedrich’s note, which I will now pursue to its further figments. ‘Petron. c. 131 Prdégne’. Progne is, as you might expect, the reading of most of Petronius’ MSS, which date from the 128 to the 15 century; but the best MS, which is of the rr‘ or rot®, had originally Procne, and the c was later altered into g. There is a lesson for those who can learn. Last comes ‘ Val. Fl. 3, 34 Prégnesson’. This, so far as I can discover, is a pure invention of Friedlaender’s. Markland, from whom he took the reference, gave Procnesson (as he gave Procne); that is the reading of V, which virtually if not absolutely is our sole authority for the text of Valerius; and I can find no record of Prognesson as the reading of any MS or any edition. ‘Unter diesen Umstanden’ proceeds Mr Friedrich, without suspecting the humour of the phrase, ‘ist auch an dem einhellig iiberlieferten Ovid met. 10, 531 piscosamque Gnidon grauidamque Amathunta metallis nicht zu andern; ebensowenig an 6, 468 ad mandata Prognes et agit sua uota sub illa’: these are the two verses on which Magnus wrote the notes which are the text of my discourse. Mr Friedrich then adds Hor. carm.I 15 17 spiculé Gnosit and Sen. H. f. 18 serté Gnosiacae, and continues ‘ Ebenso hat Kiessling mit Recht Hor. sat. 2, 5, 28 uinet uter locuples sine gnatis, inprobus, ultro’. Not merely Kiessling but the best prosodists among Horace’s editors, Meineke, Haupt, and Lucian Mueller, retain sin? gnatis; and there is no doubt that they are right. But when Mr Friedrich proceeds to argue from Horace’s satires to Catullus,— Es versteht sich von selbst, dass unter diesen Umstanden auch 64, 298 inde pater diuom sancta cum coniuge gnatisque nicht zu andern ist, —this sentence, which concludes his note, will serve very well to inaugurate The enquiry has two parts and the examples fall into two classes, 4 Fs J , 4 A. E. HOUSMAN according as the correption occurs within the body of one word or at the contact of two. The latter case, where a short final vowel is followed by an initial gn and the syllable remains short, is the more credible and shall be considered first. And first of all I will take the only native Latin word which enters into the discussion!: the gnatus which has just now cropped up for the fourth time. When Horace’s MSS give siné gnatis at serm. II 5 28, there is no shadow of a reason to doubt that Horace wrote it. Horace the satirist disregarded position even in pairs of consonants where the second is not a liquid: he wrote fornicé stantem, uelatumqué stola, saepé stilum, quid sctlicet, malé stultitia, mtht Stertinius, even fastidiré strabonem and praemid scribae.*, That was the tradition of satire, as may be seen in Lucilius; and if the correption were ascribed by their MSS to Lucretius or Propertius, who wrote mollia strata and bené sponde- bant and much else of the sort, it would be no less credible. Nay, since gnatus and gnata are the regular forms of the substantive in the MSS of Horace’s satires, the solitary exception should probably be corrected, and Aulidé gnatam restored with Keller and Mueller in II 3 199.° Now turn to the alleged example in Catullus, 64 298, inde pater diuum sancta cum coniugé gnatisque aduenit, and turn also from Horace’s practice to his. Of all Latin poets Catullus is the most sensitive to position: he nowhere disregards it except in the traditional and unavoidable uwndé Scamandri; and not only has he poté stolidum, modo scurra, nefaria scripta, gelida stabula, nulla spes, but even, in one metre, Propontida trucemue, impotentia freta, and ulttma Britannia. That is one test, a general one, by which to judge the likelihood of contug? gnatisque ; and here is another, a particular one. 1 Terentianus Maurus 894 has quandé ‘ Gnaeum ’ enuntio, but that, as Lucian Mueller says de r. m. p. 386 ed. 2, ‘exemplorum libertate excusabitur’, like 887 quandd pvhunv atqué dictitatque Mvec@éa. There is however one more Latin word which I feel bound to mention, because it has been unaccountably overlooked by the re- actionaries: unaccountably, for it is Magnus’s own name. The paroxytone accent of Mdyvos ought to be a sign (though in fact it cannot be trusted) that the vowel itself was short; and that gx did not necessarily even lengthen it by position is proved by that sacred and indefeasible thing, the authority of MSS. I find two ex- amples in Catullus, 90 1-3, nascatur mdgnus ex Gelli matrisque nefando coniugio et discat Persicum haruspicium nam mdgnus ex matre et gnato gignatur oportet, and one in my favourite hexameter ciris 374, inde mdgno geminat Idui frigidula sacra. 2 To show the prevalent ignorance of this subject among scholars who set up for metrists, Vollmer, citing these examples on p. 344 of his The usual spelling of the substantive in his MSS Horace under the head ad positionem notanda, adds ?tibi triplex and does not add uinosdé glomus nor precé blandus; and Jebb at Soph. O.C, 996, volunteering a needless note on the regular scansion mepiBréras, adduces -epidpouh and mweprypagpd as relevant to the question, and cannot even keep to himself the opinion that reprypdgecs is ‘ambiguous’ in Ar. pac. 879 odros, ri mwepi- ypagdes; 7d deiv’, eis Did they never hear of a book entitled Miscellanea Critica and published in 1745? 3 When Mr Friedrich says of II 5 28 ‘auch klange sine natis (ne-na) viel weniger gut’, he dis- plays that enviable fineness of ear which is peculiar to bad metrists. Neither Horace, who wrote not only sine nascitur serm. I 3 68 and bene nata carm. IV 4 36 but even sine neruis serm. II 1 2, nor any other Latin poet aspired to the fastidious elegance of Mr Friedrich. It would take a long time to count the examples in Ovid of bene natus, ovdine ovigine sanguine natus, Agamemnone Amazone Amythaone Andraemone A polline Ixione Polypemone natus. is witl gnata word 324 posses case, a this: t less cc gnatus, the in commc more 1 centuri and at and if . faces t natos, a Iv ‘ placité credis ’, prove a All them is the Gre habitua the MS. will wa an unu without But thc Magnus I w with as and Qui in the 1. is twice G.L.K. one whi cannot | right sp combina evidence But a | ‘ | | } if fig r at the d by an shall be ‘d which » for the shadow egarded he wrote tia, ‘radition ‘ibed by & sponde- e gnatus Torace’s gnatam tullus is in the tolidum, metre, ne test, nd here 1is MSS notanda, os glomus 0.C, 996, le regular jpouh and nd cannot wepiypagers s, Th wept- ney never vitica and 28 ‘auch it’, he dis- which is race, who 3 and bene rm. II 2, fastidious ake a long bene natus, e Amazone Polypemone PROSODY AND METHOD II. 5 is with initial g: setting this verse aside there are ten examples of gnatus or gnata to eight of natus or nata. But, except in this single place, wherever the word follows a short vowel it is spelt without the g: 62 22 auellere natam, 64. 324 carissime nato, 401 funera nati, 403 impia nato, 67 26 semine natus. What possessed Catullus not to write also coniuge natisque here? When they realise that ‘es versteht sich von selbst’ does not quite suit the case, and begin casting about for something else to say, they will probably say this: that scribes were not likely to change the commoner form natis into the less common. But, not to mention that in Catullus the commoner form is gnatus, this argument, though often used, has little force. Scribes, unlike the imaginary scribe of scholars who do not read MSS, frequently alter commoner forms to rarer; but for the present occasion one example will be more than sufficient. The second-best MS of Ovid’s heroides, G, is two centuries older than the principal MSS of Catullus and incomparably better ; and at XII 187 it gives vespicé gnatos. The best MS, P, has lost many pages, and if it had lost this page we should have this correption brandished in our faces together with Catull. 64 298. But it contains this page, and it gives natos, and so do the other MSS. I will end this section with an act of showy generosity. In Sen. Tro. 246-8, ‘placita nunc subito improbas | Priamique natam Pelei nato ferum | mactare credis’, the best MS, E, has patriamqué gnatam, part of which at least should prove acceptable to some. All the other words involved in the controversy are Greek; and none of them is spelt in Greek with yr. The Latin g, in every one of them, represents the Greek «; and in every one of them we know that the Romans, often if not habitually, represented the Greek « by c. When therefore, to take an instance, the MSS of Ovid at met. X 531 present piscosamqué Gnidon, a rational enquirer will want to know two things: why should Ovid choose to write Gnidon, with an unusual correption, when he indisputably might have written Cnidon without one; and what authority have medieval MSS in this detail of spelling. But those are just the questions which enquirers like Messrs Friedrich and Magnus never ask. I will take this name Kvidos first. The verses in which Gnidus is found with a short vowel before it and defended are Catull. 36 13 Ancona Gnidumque and Ouid. met. X 531 piscosamqué Gnidon. The oldest of Catullus’ MSS are late in the 14‘" century, the oldest of Ovid’s are of the 11th. Ovid’s verse however is twice cited by Priscian, whose oldest MSS are of the g‘4 century, and at G.L.K. II p. 10 they too give Gnidon; but at p. 52 they give Cnidon, except one which has Gnidum. What Priscian himself wrote in these places we cannot learn from the context; but we do know what he regarded as the right spelling of the name, for on p. 42, where he is enumerating the various | combinations of consonants, he has ‘cn, Cnidus, .. . gn, Gnaeus’, a piece of evidence which the thes. ling. Lat. does not record. But why do not the allies cite a well-known verse where Gnidus follows a Wigs te f ate wie é | 6 A. E. HOUSMAN short vowel in MSS much older and better than Catullus’ or Ovid’s? For a very good reason. The verse is Hor. carm. I 30 1, where regina Gnidt is given not merely by the vulgar crowd of Horace’s MSS but by some of the most important, ACAI, the oldest of which is dated about A.D. goo. But unluckily a still more important company, BRd¢y, the oldest of which is dated about A.D. 800, give Cnidi, and so too the verse is cited in the cod. Par. 7530 of Seru. de metr. Hor. G.L.K. IV p. 469, which is of the 8 century, not to mention the cod. Pal. 1753 of Mar. Vict. de metr. Hor. ib. VI p. 175, which is of the 9%. If we had MSS of Catullus and Ovid as old and good as Horace’s, do Messrs Friedrich and Co. think that they, unlike his, would all give Gnidus, or that, like his, they would lean to the side of Cuidus? The answer is that they think neither; they think nothing. There is no room left for thought in minds inhabited by a passion for false quantities. Apart from the metrical witness borne by the three poets, it can be shown that in their times the regular if not the only spelling was Cutd-. I have already quoted Priscian, whose testimony is borne out by Martianus Capella 249 ‘C ... consonantes quasdam praecedit, LTRMN, ut... Cnidus’. There exists on the other hand what looks at first like an earlier testimony to Guidus: Ter. Maur. 890-2, tertia uda (N) sic uidetur posse G mutam sequl, Graia uerba quando in usum sermo noster suscipit, Gnosios si dicere arcus insulam aut Gutdum uelis. But if the context is carefully read it will appear that Terentianus (of whom we have no MS, only editions from 1497 onward) must have written C mutam and Cnosios and Cnidum; for he goes on to say that the praenomen written Cn. is pronounced Gnaeus, digresses to the similar case of C(azus) and to amurca pronounced amurga, and then proceeds, 902 f., G tamen, mox tertia uda, r a r i u s componitur, quando guatum nominare dicere aut gnarum uoles. The oldest MS in which I find Gntdus is the cod. Moneus (saec. V-VI) at Plin. 2.4. XIII 59. In the twelve inscriptions cited by the thes. ling. Lat., one of which, C.J.L. IV 5533, is from Pompei, and another, X 1403 g 1 34, from Herculaneum, the spelling is always with C, Cutdus Cnidius Cnidia, except that in one, IX 4929, the letter is doubtful.’ 1 Before quitting Catullus I will let Mr Fried- rich’s competence to handle this matter be seen in the light of his dealings with 64 36 and 75 and his display of knowledge and consistency there. At 75 he prints, like most editors, ‘iniusti regis Gortynia tecta’ where the MSS have cortinia, but he thinks that the change needs justification and justifies it thus: ‘die Schreibart Gortynia wird durch Claudian (VJ cons. Hon. 634) gegen Ciris 114 Cortynius und Stat. Theb. 4, 530 arbiter Cortynius geschiizt’. He has found these three examples in Baehrens’s commentary and has searched no further; it does not occur to this student of orthography to enquire whether Virgil has the word, and how his MSS spell it. And then he proceeds ‘G fiir griechisches K entsprach dem Genius der lateinischen Sprache; vgl. zu 11, 6’ (Sagas for Zdxas): he does not know that the Greek word is Toprivws. Now come to 36, where he prints like everyone else ‘ Crannonis que domos’, and writes no note. But instead of C the MSS give G, in accordance with the principle F inhosy as old date, and C no. 1& later ; CNOS gradu Ter. } adject usually Ir Cnosiz : in the sordido edition back a Keller’ whole, substit: vanishe VE: Ovid, § puellae lucida. ( Th postero theless and the short vi the evi which I That is unimpa: and tho earlier t hood. | that G for lateinische better MS does he n Gnidum a Create no afford Mr 1 In ca Hi 3 | at | | ? Fora + is given the most aluckily a ed about . 7530 of y, not to which is Horace’s, e Gnidus, sr is that lought in be shown I have ipella 249 ’. There Gnidus: of whom C mutam ritten Cn. O amurca V-VI) at Lat., one 34, from cept that ry and has ccur to this ether Virgil ell it, And K entsprach he; vgl. zu t know that come to 36, yannonts que stead of C he principle PROSODY AND METHOD II. 7 From Caria let us now cross over to Crete, and see if Cnosus will be less inhospitable to the letter G than Cnidus was. It is; for the spelling Gnosus is as old as the reign of Augustus, and appears upon two Cnosian coins of that date, Svoronos no. 190 GNOS, 191 GNO. But Cuosus was still the usual form, and C(olonia) I(ulia) N(obilis) CN(osus) or C(nosus) appears not only on no. 180, which is earlier, but on 181-3, 188, 189, 192-4, some of which are later; and we also find COL. I. N. CN. in C.J.L. III suppl. 12041 and C. I. N. CNOS. in 14377, of the time of Claudius. Afterwards the true form was gradually superseded by that which the vulgar found easier to pronounce. Ter. Maur. 892 I have already dealt with; but in Virgil’s text, where the adjective occurs six times, the majority of the capital MSS (saec. IV-VI) are usually for Gnosius. So now for the verses where a short vowel precedes. In Hor. carm. I 15 17 it has long been customary to print calami spicula Cnosit ; for Horace the lyrist has other principles than Horace the satirist, and in the odes there is no saepé stilum nor siné gnatis but on the contrary II 18 28 sordidosque natos and III 5 42 paruosque natos. But when I saw that Kiessling’s edition was to be revised by Mr R. Heinze I knew that we should have Gnosiz back again, and also palis in art. 65; and we have. Gnosit is the reading of all Keller’s MSS excepting one; but that one is R, probably the best MS on the whole, and certainly much the best in orthography. R has Cnosii; but R? substitutes the vulgar form, and in the next copy the true spelling would have vanished away.! Naturally then it has vanished away from the later and inferior MSS of Ovid, Seneca, and Statius: met. IX 669 proximd Gnosiaco, H. f. 18 ‘mundus puellae serté Gnostacae gerit’ (but fert. a nobis lac e cod. opt.), stlu. V I 232 lucidé. Gnosis. The opinion that Seneca could write serté Gnosiacae is not in itself pre- posterous. Though he writes frigidd spatio and undiqué scopuli, he does never- theless write enod? Zephyris, tranquilla Zephyri, tramité zonas, and antré Zethi, and the form Gnosiacus did exist in his day. The question is whether after a short vowel he would prefer it to the also existing form Cnosiacus, and whether the evidence of MSS no earlier than the 13% century (for the spelling to which E points is Nosiacae, not uncommon in MSS) is any proof that he did. That is a question which can easily be answered, and answered right, by any unimpassioned enquirer. Ovid was a more scrupulous metrist than Seneca, and though the MSS of the metamorphoses are older than Seneca’s they are not earlier than the 11‘ century; so that Gnostaco in IX 669 has no more likeli- hood. Jucidé Gnosis in Stat. silu. V 1 232 has even less. The siluae have | Gnidum at 36 13? | create no false quantity and therefore would that G for Greek K ‘entsprach dem Genius der and this is the usual spelling of Kpdyos in lateinischen Sprache’, and so do the older and Latin MSS; but Mr Heinze does not accept better MSS of Cic. de or. II 352. Why then it, because it does not charm his ear with any does he not print Grannonis at 64 36 as he prints__ illegitimate correption. And at carm.I 3 6 he- Because Gyrannonis would prints the Vergilium of R, not the Virgilium of R? and the other MSS, because the false spelling does not redeem itself by being also a false quantity. afford Mr Friedrich no voluptuous sensation. 1 In carm, I 21 8 the MSS have uwiridis Gragi, ; i a ‘ud ated as if er tg 8 A. E. HOUSMAN survived only in one MS of the 15 century, and Statius was a poet acutely sensitive to position. His only violation of it is the unavoidable ign’ zmaragdos of Theb. II 276, and he lengthens a short final before st ib. VI 551 agilé studium. But stop: I am overlooking the march of modern progress. P and some other MSS represent this same poet as having written ‘fas et mihi spernere Phoebum’ in Theb. VII 733, and the latest editors believe them. When zeal for a best MS joins its voice to partiality for bad versification, neither the whispers of reason nor the wails of Statius can be heard. The next stage in the gratification of appetite will be to print nihil transit amantes with P in II 335 and mithél foedere rupto in 3309. Here ends the tale of supposed correptions before initial gn, and we come to the still more difficult correption in the body of aword. And first in cygnus. Horace, I hardly know why, is still allowed to write ‘donatura cycni, si libeat, sonum’ in carm. IV 3 20, though C is the only one of Keller’s MSS which preserves the spelling, and all the rest have cygni or cigni with the required false quantity. But in Mart. IX 103 2 we have already seen Fried- laender and Mr Friedrich upholding cygno against the incredulity of Lachmann. I said above that half of Martial’s MSS give ‘cycno, and that G.L.K. VII Pp. 539 I gives not Lacaena cygno, which these metrists are resolved to write, but Lacena cigno, which they dare not But I now go further, and I raise the question whether cygnus is a form which Martial had ever seen or a sound which he had ever heard. Descend to the 4 and 5‘ centuries, to the capital MSS of Virgil: cycnus occurs there nine times and the proper name Cycnus once; and this is the invariable spelling except that M in three places has cygnus. Again in the palimpsest of Fronto we find cycnus twice and cygnus never. From inscriptions the thes. ling. Lat. cites no example of Cygnus or cygnus either, but one of cycnus, six of Cycnus, one of Cicnus, one of the adjective cycneus, and three of the derivative cycnion. To say that cygnus may have been the vulgar pronunciation in Martial’s time, as Gnosus doubtless was, is to over- look the fact that the vulgar did not use the word. This name for the swan was exclusively literary and almost exclusively poetical. From all prose earlier than Fronto there are cited only nine examples, one of which does not count, for Mela I 110 is explaining why the town Cycnus was called so, and six of which are in Pliny’s natural history. The vulgar called the bird olor, and when they began to try at its Greek name they pronounced it cicinus, preserving the c, so that it came into Italian as cecino and cecero. 1 I note here, because Lucian Mueller de r. m. p. 385 ed. 2 overlooks it, that Martial does in one particular carry correption further than any other Latin poet of classical times. To neglect posi- tion before initial gi is quite regular; but Martial, sound pagan though he was, indulged in a licence which Mueller confines to Christians, and left a short vowel short at the junction of the compound andglypta IV 39 8. For this he had the authority of Theocr. epigr. 4 2 (anth. Pal. IX 437) dprtyAugés ; but it is not to be inferred that either Theocritus or Martial would have shortened the first syllable of Aglauros. When Mueller adds ‘at non credibilest auctorem perui- gilii Veneris posuisse peruiglanda pro ditrochaeo [46], sed unice amplectemur quod habes libro Salmasiano peruiclanda’, he disquiets himself in vain: the writer of the peru. Ven. allows spondees in the fifth foot of his tetrameter, and anapaests too. ‘ procn | gnost < met | in fa we te ] anth. Plau ill Proc fi i b 5 4 Prog f : Proc . vice 2 « > spe prod sing prim are 1 no o 3 KVQaT the evid ; | wor : l a poet 1avoidable before st modern ig written st editors ty for bad is can be nihil we come In cygnus. cycnt, si er’s MSS with the en Fried- ichmann. to write, raise the a sound » to the er name ee places id cygnus yenus or adjective ave been to over- he swan prose does not illed so, the bird t cicinus, e inferred ould have s. When rem perui- litrochaeo ibes libro himself in spondees anapaests PROSODY AND METHOD II. 9 The next case, and the last, is Prdgne, cited from Ouid. met. VI 468 mandata Prognes and Petr. 131 8 7 urbana Prégne. The latter, as I have signified already, is not a prudent citation, and a true statement of the facts, ‘ procne L, procne mutatum in progne B, progne ceteri,’ teaches the same lesson as ‘ cnosit R, gnosit R? ceteri’ at Hor. carm. 115 17. The MSS which give Progne in Ouid. met. VI are of the 11‘ century or later, and so are those which give that form in fast. II 629 and 855, trist. V 1 60, and ex Pont. III 1 119. But in the fasts we have also a MS of the rot® century, and it gives prone and prome, and at ex Pont. III 1 119 we have a MS of the gt, and it gives Procne. The oldest MS in which I have found Progne is the cod. Salmasianus (saec. VII-VIII) at anth. Lat. Ries. 27 in the title of the poem. The absurd and unmetrical verse Plaut. rud. 604 ‘natas ex Philomela atque ex Progne esse hirundines’ is illegible in the Ambrosian palimpsest, and in that MS it was shorter; and if Plautus himself named the woman he called her neither Progne nor Procne but Procina. The earliest MS witness is in Verg. georg. IV 15, and it is instructive to those who admit instruction: all the capital MSS, MPR, have Procne, but g for c has already made its appearance in the gt century cod. Bern. 184. As for inscriptions, in the C.J.L. I find three examples (II 5392, III 2935, IX 405”) of Procne, one (IX 649) of Prochne, one (V 1323) of Procine; no trace of Progne unless in VI 135, where the reports of witnesses vary between Caectlia Procne and Cecilia Progne. The first business of those who pretend that Ovid and Petronius shortened the first syllable of Progne is to show reason for supposing that in their lifetimes Progne was even an existing form.’ But what I am arraigning is not misjudgment in any particular case but a vice of method. All the words cited to prove correption before gn, except where the citations are false, have one common feature: they are alternative spellings of words which were also spelt without the g. If there could be produced from classical poetry a single instance of Polygnotus or andégnostes, a single instance like the Epignomus of Plaut. Stich. 464, it would constitute prima facie evidence and require consideration. But the instances produced are these: the Latin word gnatus, also spelt natus, and the Greek words Guidus, Gnosus, cygnus, Progne, also, and originally, spelt Cnidus, Cnosus, cycnus, Procne: no others. These weigh no more than catéyvawe from Eur. Tro. 1252 when xvantw exists, or the paorevwv and dumdaxwv and potiretur and déprehendo which MSS proffer to all who crave them. They constitute evidence only for the intellectually or morally deficient; for those who do not know what evidence is, or for those who have set before themselves an end which they are determined to reach by fair means or foul. A competent enquirer and an upright man will ask two questions. First, did scribes introduce g into such words, or did they not; and the answer is that they did. Second, what motive had poets for spelling such words with g, when to do so involved a scansion good MSS do, in Hor. carm. II 4 6, but un- 1 I have already exposed their fiction that fortunately more and better MSS read Tecmessae. Valerius Flaccus has Prégnesson in III 35. They would dearly love to read Tégmessae, as many sh f : ey ; | ie Io PROSODY AND METHOD II not elsewhere admitted ; and the oply person who can answer that question is Echo. If words which exist in alternative forms may be cited from MSS as evidence in matters of prosody, Mr Friedrich must be taken to task for corrupting with unnecessary alteration two verses of Catullus, 63 8 and 9g, in both of which the consonants mp fail to make position and the first syllable of tympanum is short. niueis citata cepit manibus leue tympanum tympanum tubam Cybeles tu mater initia. This scansion recurs in Varro, Non. p. 49, tibi tympana non inanis sonitus matris deum, and in Maecenas, G.L.K. VI p. 262, ades et sonante tjmpano quate flexibile caput ; and at Oxy. pap. XIII p. 43 I collected from Greek poetry no fewer than five examples of tiymavov. There is therefore much more authority for tympanum than for either cygnus or Prégne, and it surprises me that Mr Friedrich should accept Scaliger’s conjecture typanum; or rather it would surprise me if one could expect to find even consistency, the humblest of the virtues, keeping house with this confederacy. But evidence for them is good or bad according as it may serve their turn: a short vowel before mp is not a false quantity on which they happen to have fixed their affections, and so they have no motive for believing that a vulgar scribal error is anything other than what it is. A. E. HOusMAN. 7 invest wrong accou listed that i It ac it is 7 Jorda A. the indep that | lacun hand is a | missi gap. 1 that the | why founc MS. | ( instal toget differ but t Vi.a text A an ing 2 it im) 1 R) 2H 3 Cl from t | | \ ; 3 | \ estion {SS as isk for dg, in syllable in five banum should if one 2eping ording ity on notive [AN. THE VATICAN PLATO. THE Plato MS. designated by Bekker as 2. and by Burnet as O escaped the investigation of editors of the text of Plato for nearly a century, because it was wrongly cited by Bekker as Vat. 796. Finally, in 1908 Rabe! published an account of the missing MS., which he had discovered in the Vatican library listed as Vat. gr. 1. Until its rediscovery the opinion of Jordan? prevailed that it was a comparatively late MS., copied from A (Parisinus graecus, 1807). It actually belongs, however, to the late ninth or early tenth century—i.e. it is of almost equal age with A and B. The study of its relation to A has hitherto remained where it was when Jordan’s article was written. There is an excellent account of the situation in A. C. Clark, The Descent of Manuscripts, Oxford, 1918, pp. 395-398. In spite of the prevailing view that O was dependent upon A, Burnet nevertheless cited it independently in the apparatus to his text. In fact there is conclusive evidence that O is not copied from A throughout, for O contains in the text a long lacuna in the fifth book of the Laws (745a-c) that was supplied in A by a later hand (Burnet’s A*). There can be no doubt about the relative dates, for there is a lacuna in the sixth book (783b-d), which occurs in both MSS., and the missing matter is supplied in both by the same corrector who filled the earlier gap in A, as the minute identity of the handwriting proves. Clark suggests that O was not copied immediately from A, but from a derived MS. in which the lacuna had been filled up.* This hypothesis does not, however, explain why in the first five books of the Laws O presents many good readings not found in A, sometimes not in any other MS., while in the remainder of the MS. it hardly deviates from A. On this point the following statistics are instructive: I have noted all instances in Burnet’s apparatus for the Laws where A, L, and O are cited together, and have noted the agreement of O with A and L respectively in the different books. This method affords only the roughest kind of a comparison, but the discrepancy is so great that it can hardly be accidental. For Books VI. and VII. I also give revised figures (6 R, 7 R), obtained by examining the text of O for all cases in those books where Burnet’s apparatus indicates that A and O disagree. The revision consists in correcting errors and in disregard- ing all cases where O agrees with A? or where erasure in A or in O makes it impossible to determine whether the MSS. originally agreed or not. | 1 Rheinisches Museum LXIII. (1908), pp. 235-8. contains only Laws, Epinomis, Epistles, Defini- 2 Hermes XII. (1877), pp. 161-172. tions, and Spuria (all but a few lines of the 3 Clark makes the mistake of citing evidence A-xiochus are missing). from the Republic with reference toO. The MS. ok. Bi iy ; ¥ ‘ axd 5 at = y 12 L. A. POST Books of the Laws... ws £39 3/1 4 | 7/7R! 10} 11 12 A and O agree againstL ... | 11 | 3 | 5 | 16] 20 | 25 | 25 | 21 | 25 | 30 | 26 | 22 Land O agree againstA ...| 10] 8 | 19 | 22 | 28}; o| 6; 3; 8] 2] 8] § Aand LagreeagainstO 4}; 0] 8] 3 | 3/ of of 8] 2) gi 4] of 3 In the first five books O agrees with A 25 times and with L 87 times; in the last seven books O agrees with A 165 times and with L 40 times. It is apparent that O’s relation to A is not the same in the first five books of the Laws as in the last seven. If we take the revised figures for VI. and VII. O’s agreement with A is almost complete; there are only 3 cases where O agrees with L against A to 45 where it agrees with A against L. There is another possibility, that the scribe of O may have begun to copy A somewhere in the middle of his MS. That this was actually the case is amply proved by a comparison of O with the facsimile of A published by the authorities of the Bibliothéque Nationale. This comparison is not difficult since the Vatican library possesses such a facsimile. For the first five books of the Laws O has the title in large uncials, [IAdtwvos Nouv 4G, etc. Above this is drawn an ornamental line. There are no colophons. In the case of Book I. there are added to the title the words 4 epi vopobecias. Except in the case of Book I. O begins the text with a large ornamental letter. The titles and colophons as given in A were, however, inserted in O in the margin by a later hand. At the beginning of Book I. we find in the upper margin I]Adtwvos vopobecia, the form of title found in A. In the upper margin of the page in which Book I. ends in O we find voyuer 4 vopobecias vopor 4 vouobeciat 8, which are identical in form with A’s colo- phon and title at this point. At the beginning of Books III. IV., and V. A’s titles appear in appropriate places in the margin or between the lines of O. At the beginning of Book VI. we find A’s title standing originally in O in the same form as in A, except that in O the numeral is at the right of the title; in A it appears as usual in the left-hand margin, at the beginning of the text. For A’s ornamental crosses O substitutes in each case four dots. O for the first time at the end of Book V. inserts in the body of the MS. a colophon identical with A’s even to the ornamental sign which A regularly employs at the end of a book. O employs it here only. At the beginning of the text of Book VI. O forgets his ornamental letter. He remembers it in Books VII., VIII, IX., and XI., but copies A in Books X. and XII. Except for a slip on O’s part in omitting NOMOI in the title of VII., the titles of O and A agree from this point on. With Book VIII. O adopted A’s practice of placing the numeral in the left margin at the begin- ning of the text. O’s copying of A’s title design is particularly noticeable at the beginning and end of the Definitions and at the beginning of the Eryxias, where there are special peculiarities. In the Epistles A was misled by the quo- tation in of the Fj title g: in all Ep. 8 T for sor convin occurs from t the inl been f where stop a origin: instan puncti the m A: 6@ A: Ta 746b Tooe C few re not to N like ve O doe adopt O wit point and C Bernt | ‘7 Befor midd. capit line is | ae ) agree for e Ke ch te +f | | > : | 26 | 22 imes ; in mes. It s of the VII. O’s ) agrees to copy is | by the difficult e books &, etc. In the oecias. imental in O in e upper In the obecias s colo- and V. e lines in O in e title; e text. for the lophon nploys inning bers 1 XII. I., the lopted begin- ceable yyxtas, quo- ~~ ~ — | 745¢ 7. THE VATICAN PLATO 13 tation from Dionysius’ letter (Ep. 7, 339b 8), and left space for the filling in of a title. At the beginning of Ep. 8 no superscription appears, and the Epistles are not numbered by the first hand after Ep. 7. In fact A in his title gives the number of the Epistles as twelve instead of thirteen. O follows A in all these peculiarities, though a later hand supplied a superscription for Ep. 8 and numerals (wrongly) for the remaining letters. The demonstration that somewhere in Book V. of the Laws the scribe of O for some reason turned from his original model and began to copy A is thus far convincing. Certain considerations make it possible to show that the break occurs after the word yiyveoOa: (746b 7), and that the scribe began to copy A from the words pndev aroneizrery (Plate 4). There is a change in the shade of the ink at this point. Before this point is reached the punctuation of O has been frequently corrected to agree with A. Stops are inserted in darker ink where the original scribe had left no break between words; for instance, the stop after éxdotows (746a 3). Furthermore there are stops in O, placed there originally, as the spacing between the words shows, which are not in A; for instance, the colon after Sesxcvivta (746b 7). From this point on, however, the punctuation was originally that of A, as is shown by the ample breaks left for the marks of punctuation. The readings of the text confirm this result : 745d 4, mabe. A: O. d6, A: poprapr.O. 8, dwdexa Oévtas A: d@dexa ‘ut apparet,’ pr.O. e 2, Ta dwdexa THs TuHpaTa A: Ta THs .. . S@dexa Ty. O. 746 a 8, oxedn A: Exes O. 746b 1, wavta AapBavew A: O. b 6, rovde A: rode O. From this point there is minute agreement between A and O. The few readings that seem to indicate a disagreement will be found in most cases not to be original in one or the other of the manuscripts. Moreover, it is A’s custom to connect the two parts of a compound word like vouoBérns by means of a curved hyphen below the letters (e.g. O and ®). O does not do this, except very rarely for special reasons, until 746c, when he adopts the practice from A. This mark, for instance, is found in A but not in O with girAdripos (744e 3), vouobérns 4), (746b 3). Up to this point it is hardly found at all in O, except to correct a mistake in spacing. A and O both have the mark in voyoecias (746c 7), evOutropiay (7474 4), vopo- Gerntéov (747d 5). O fails to copy A in this respect in vowodérny (746c 5) and (747b 8). There is another change in O's practice caused by his copying of A. | Before the end of the fifth book, O several times begins a new paragraph in the middle of a long speech. The beginning of the paragraph is marked by a ' capital letter placed further to the left than the ordinary line. The previous line is left partially blank. This occurs at 695d 6, 698e 5, 716d 4, 730b 1, and It does not occur in the sixth book or later, as it does not in A. A careful examination of the MS. after 746b 7 shows a much closer | agreement between O and A than is indicated by Burnet’s apparatus. Note _ for example the following: 747e 4, O had originally tots, which has been I : a 4 j i | 14 L. A. POST corrected to rovs; 752a 2, yap is a marginal addition in O; 752e 1, the correc- tion in A is early and was copied in O. The same explanation holds for the many other cases where O agrees with A®. Erasures explain other apparent discrepancies. The rest are mostly accounted for by O’s rather frequent slips of omission. In 760b 6, O has plainly ¢uAdpyous ; 763e 6, A and O both had originally décav ; 773c 6, here and elsewhere, where an omission in A is not supplied by A*, the omission also occurs in O. The omission is frequently supplied in O by the same hand (A®*) asin A, though often by what appears to be another hand. I doubt the possibility of distinguishing the hands with certainty, and infer that A* had helpers in his work of revision. 792b 4, O's reading is wrongly reported, being really the same as A’s. 794d 7, 804a 4, the discrepancy is due to correction by the second hand in O. 802c 4, O and probably A read originally waca araxrws. The only real divergences of O from A or A? in Books VI. and VII., apart from omissions, that are correctly indicated in Burnet’s apparatus, occur in the following places: 768d 7, 798a 4, 802e 4, 816b 4, 820e 11, 821e 3. They are all of them slight errors. On the other hand, before 746b 8 the original reading of O differed from that of A much more frequently than the apparatus indicates. This part of O abounds in erasures and corrections, evidently made in order to bring O into conformity with A. This is not surprising, since A was doubtless employed by the original scribe in his revision of the whole of O, not merely of the part actually copied from A. In the first book, for instance, there are nine cases of such correction, where O had elision between words and A does not, and seven cases where A has elision but O did not. Numerous omissions of single words are similarly repaired. The demonstration that the Vatican Plato (Vat. gr. 1) is a copy of Parisinus graecus 1807 from Laws 5, 746 b 8 tothe end is complete. A few remarks in regard to the various hands of the MS. may be of interest. The hand of the text has been recognized as that of two other MSS., MS. grec 451 of the Bibliothéque Nationale, containing works of St. Clement and others, and the Lucian MS. of the British Museum, Harley MS. 5694 (see E. M. Thompson, An Introduction to Greek and Latin Paleography, p. 222 and p. 227). The former of these contains a subscription stating that it was written by the hand of Baanes, notary of Arethas, Archbishop of Caesarea in Cappadocia, in the year 914, i.e. the well-known Arethas of Patras. Two pages are repro- duced in facsimile in Omont, Facsimilés des Manuscrits Grecs Datés de la Bibliothéque Nationale. Whether in the library of Arethas or elsewhere, O was being copied by Baanes from some other MS., when the acquisition of A or some other cause led the scribe to turn to it from his original source. A and — O either remained for some time in the same library or else they were brought together once more some time after the copying of O from A, for the scholia of | O are identical with those of A, and, as has been noted, lacunae are supplied in both MSS. in the same hand, notably at Laws 6, 783b 2-d 3. This hand is termed A* by Burnet and assigned tentatively to the twelfth centt corre of th mist he w is an say t schol his ce type addex I can late. copy! He < minu book hand: centu than in an proje later revis¢ and \ centu by O' on f. introc that Kal Te sO sl Exan exam Corre I books collat | photc readil the J deter valua br 4 : > 3 A : : a4 | e correc- s for the apparent ent slips both had A is not equently »pears to ids with 2b 4, O's 4a 4, the , O and es of O correctly , 798a 4, ed from art of O into loyed by the part cases of 1d seven le words copy of A few t. The TeC 451 others, E. M. 227). tten by yadocia, repro- s de la was f A or A and 1olia of upplied twelfth | | THE VATICAN PLATO 15 century. In O it will be convenient to call it O%, leaving the term O* for corrections in the text. Now O also wrote the scholia and added the names of the dtalogorum personae in O. He probably had helpers, but there is no mistaking his cursive minuscule script in the scholia at 865b (f. 95 v.), where he wishes to distinguish verse quotation by a different style of writing. There is another example of the same thing in Laws 12, 950e (f. r30v.). I venture to say that except for these scholia no one would guess that the corrector and the scholiast were identical. It is instructive to note that the same scribe had at his command several styles of writing corresponding to the different founts of type used in modern printing. The question arises when the scholia were added in O and when the revision of O (with additions to A) was carried out. I cannot answer this question, but I am sure that the twelfth century is too late. A priori there is a strong presumption in favour of a date soon after the copying of the text. I suggest that O* (A*) is much earlier than was supposed. He could write uncials or a minuscule book hand or his own cursive minuscules as he pleased. Since dated examples of script are regularly in book hand, I do not see how one is to determine O*s date by his informal handwriting. Certainly O*s uncials and book hand would pass as tenth century, for Rabe assigns some examples of them to that period. Other hands than O*’s may be distinguished in the margin of O, but not with certainty, and in any case it is probable that the different scribes were carrying out a single project of revision. The rather shaky writing of the marginalia on f. 135 and later seems to be that of a novice taking up the work of O® for a few pages. The early date of O* is further attested by a comparison with the later reviser O*. This is the eleventh-century corrector who is called M* by Rabe, and who cites the authority of the patriarch’s book. If he is really eleventh century, O® cannot be twelfth, for O* has occasionally corrected scholia written by O8 and has written his own marginalia according to the space left by O°, e.g. on f. 100 v. O* has very infrequently added scholia of his own. He has introduced a few emendations into the text, while placing a note in the margin that all the copies have another reading. We owe to him, for instance, the wal ta for cata of Ep. 7, 346a 7. His dddrq@ for adAo in Laws 3, 700b 7 was not so successful. At the end of Laws 8, he supplies an omitted colophon. Examples of marginal signs added by him are frequent. There are good examples on f. too r. Of later hands than O* there are very few traces. Corrections in the text are numerous. Their date is usually uncertain. From the facts noted in this article we may conclude that for the first five books of the Laws O is of independent value. It deserves to be carefully collated. For this purpose it is necessary to study the original, not a photographic copy, because of the numerous instances where the original reading of O has been erased and the reading of A inserted. For the rest of the Laws, the Epinomis, the Letters, and the Spuria O may be of use in determining the original reading of A. Thecorrections and marginalia are also valuable, but O is certainly copied from A. L. A. Post. 4 | he re : | es ‘IMITATION’ IN PLATO’S REPUBLIC. I. It has become a standing reproach upon Plato’s treatment of poetry in the Republic that he forgets or misrepresents in the tenth book what he said in the third.* According to the earlier discussion, poetry is required to perform important services in the ideal state; its subject-matter will make the young familiar with true doctrines (376 sqq.); its style will reflect the qualities proper to the character of guardian, and therefore—by the principle of imitation—induce and confirm such qualities in the souls of young and old (392c sqq.). Poetry, like all other forms of art, must train the young to love and resemble the beauty of truth (7@ xaX@ Aoyw, 401d), so that when reason develops, they will recognize it as an old friend (402). Now the tenth book attacks all ‘imitative’ poetry (607 psuntixn, 595a) as remote from truth, and excludes it from the state because of its pernicious influence. The commentators say that all poetry is imitative according to Plato, so that the tenth book contradicts the third; Plato has forgotten the valuable services which poetry can render. But the tenth book expressly exempts from condemnation hymns to the gods and panegyrics on good men (607a). Even these, however, are imitative;* for they imitate or represent the characters and actions of gods and men. The strange result is that the tenth book not only contradicts the third ; it also contradicts itself. It is quite clear that in the earlier discussion Plato regards as ‘ imitation’ (uiunors and cognates) the poetry which he admits as well as that which he condemns. For example, in 398b he retains the poet who will imitate (jpoiro) the style of the virtuous man; in 399a the harmony in songs is to imitate (utpnoatto) the tones and actions of brave men; in 401a grace, rhythm, and harmony in every kind of production are allied to, and imitations (usypara) of, the good and temperate character. But perhaps there is a sense in which the poetry which he admits is not imitative. Otherwise it is hard to understand how he can have retained in 607a the kind of thing which he has just excluded. Some interpreters, indeed, think that since he retained some poetry, though all is imitative, he made an unconscious distinction between a true and a false kind of imitation.* And it would seem that the true kind of imitation is in some sense non-imitative; for at the beginning of the tenth book he says that his previous decision was to 2°See e.g. Adam on 6072 4. 3 See Greene (0p. cit.), pp. 67-68, especially p. 68, n. 2; and cf. p. 34. 1 The clearest statement of the charge is by W. C. Greene in his valuable study, ‘ Plato’s View of Poetry’ (Harvard Studies in Class. Phil. XXIX.), p. 54. \ | exclud in whi Bi conclu: that in the go Homer possibl was ur nothings It seen the ear the enc sion in Tk 392¢c. (I) (2) (3) To ma if the imitatic Be to be c senses and ‘ ye Im sympat! the first charact is dang built (3950 sc Im the gua their the sam of men fee A and i / ? . | | | in the said in portant iar with to the uce and try, like eauty of cognize as rnicious ding to tten the xpressly n good tate or esult is itself. itation’ hich he pot otTo) imitate im, and sis not ined in indeed, 1ade an And it ive; for was to especially \ | 17 exclude so much of poetry as is imitative (ro wapadéyecOas bon pipntixyn). This certainly looks as if not all poetry is imitative in the sense in which the word is here used. But the commentators say that this statement regarding his previous conclusion is a mistake.’ The third book, they say, did not exclude the poetry that imitates a good model. But this accusation is surely very loosely put; the goodness of the model will not satisfy Plato’s earlier requirements. Homer, for example, was censured for a passage where he imitated the best possible model, Zeus himself (388c wsunoacOar). It is true that the imitation was unlike (dvoyoiws).2 But this is no support to the interpreters who say nothing of likeness or unlikeness, and merely require the model to be good. It seems then that it is not Plato but the commentators who have forgotten the earlier discussion. Fortified by this consideration, and also by a remark at the end of 595c tor etc.), let us recall the discus- sion in Book III. The question of the poetic style which is to be admitted begins with 392c. Three styles are first distinguished : (1) Simple narrative, where the poet speaks in his own person, as in dithyrambic poetry. (2) Narrative by means of imitation, where the poet speaks in character, as in dramatic. (3) A style compounded of (1) and (2), as in epic. J. TATE To make oneself like another, either in voice or appearance, is imitation ; if the poet nowhere concealed his own personality, there would be no imitation in his work (393c). Before deciding which style is to be admitted, an important question has to be considered: are the guardians to be imitative (394e)? Here the two senses of the term ‘imitative’ clearly appear. For the answer is both ‘no’ and ‘ yes.’ Imitation in one sense is forbidden; for it is harmful to identify oneself sympathetically (whether as poet, actor, or audience) with other people. In the first lace, such imitation would destroy the single-mindedness which must characteiize the guardian (394e, 395a). And, in the second place, as imitation is dangerously apt to become reality, an inappropriate character would be built uy by imitating, in body, voice or thought, inappropriate qualities (3950 sqq.). Imitation in a second sense is permitted; indeed, it is recommended. It the gua’dians imitate, they must imitate from childhood the qualities proper to their oy cupation, such as courage, purity, temperance (395c). Formally this is the same kind of imitation as the first, for it involves speaking in the character . of men who are courageous, etc. But really it is very different; for the 2 Cf. 377e, where the comparison of such poets with bad painters makes its first appearance, fee Adam on 505a 3, followed by Greene, so and 54. | ‘7h — _ Mis as ad - | : B . 18 J. TATE guardians who practise it will be imitating their own ideal character, not characters utterly alien from their own. It involves not the suppression but the development of the personality. Owing to these considerations the three styles described above are modified so as to appear in the following form: (1) The non-imitative style, which is the style of the virtuous man, the man of measure (396bcd). Such a man will refuse, as a rule, to imitate (i.e., to assume the character of) baser men; nor will he be inclined to imitate a virtuous man who has been thrown off his balance by some misfortune. He will tell his story for the most part in his own person ; that is, his style will in the main consist of simple narrative. But he will imitate (a) good men like himself (with the exception already mentioned), (b) baser men occasionally for the sake of amusement, (c) baser men when they happen to perform a virtuous action. Accordingly this style will resemble the epic in containing both simple narrative and dramatic or imitative elements. But—in contrast with Homer—it will have very little imitation (396e) and the imitation which it contains will be ethically good. Its characteristic is the restriction of the imitative element ; hence it is still regarded as the non-imitative style. (2) The imitative style, which is natural to the man of opposite character. The more worthless this poet is, the readier he will be to imitate every kind of speech and action, and indeed every kind of noise, from thunder to the bleating of sheep (397a). His style will either be wholly imitative or contain but a small portion of direct narrative (397b). As the non-imitative element is reduced to a minimum, this is the imitative style. (3) A style compounded of (1) and (2). Which of these three styles is to beadmitted? The answer is that we shall admit only the unmixed style which imitates the virtuous man (Tov tod axparov, 397d); in other words, (1) the non-imitative style, which nevertheless contains such kinds of imitation as the virtuous poet will not disdain to practise. Plato could scarcely have made it clearer that the style which is non-imitative in the first sense is yet imitative in the second sense, the sense in which the guardians are to be imitative. It is now possible to set right the error of the commentators who: ay that in the third book Plato would have admitted imitative poetry where tie model imitated was good. Not only should the model be, as a rule, god; the important point is that the poet must himself be good, and understand the principles of goodness (396b 6 Kadds Kayabds ; 396c 6 pérpios 4oode the artist must have 4o2bc the truly perczives the ideas and their images). | | The poetry which Plato admits as non-imitative may then be imitative or dramatic in some passages, as when the poet speaks in the character of a good man like himself. But from 401-402 it appears that it will also be im ‘tative 1 306c (otk éxi rg roratTy give the impression that is used ir 5” and 5¢ sense with reserve as a kind of metapho”. conde: canno charac imitati sense sometl the pu leaves the gu reality 1 On4 tion here than the in the te ideal sta best and essential] by the i unlike th ina Im1té and | Kal = : of th th “ suppl paneg 1 we I thing n charac wheth < tal : | = i racter, not ression but above are s man, the nitate (i.e., » imitate a tune. He tyle will in d men like sionally for perform a containing in contrast yn which it tion of the le. character. ery kind of he bleating put a small reduced to at we shall 1 (Tov Tov ative style, $ poet will sar that the the second : ay that in tie model good; the rétand the pios ; rcsives the mitative or ‘of a good > im ‘tative ised ir . apho” ‘IMITATION’ IN PLATO’S REPUBLIC 19 in a slightly extended sense; in its beauty, rhythm, and harmony it is an imitation or expression of the good and temperate character proper to the poet and to those for whom he writes (401a awppovds te Kal ayaBod HOous Te kal pipnpata); or, to put the same thing in more general terms, it is imitative because it imitates the ideal world which the philosopher strives to imitate and resemble in his own person (401bc, 402bc, 500c).} The conclusion then is that Plato is quite right in stating at the beginning of the tenth book that he had previously decided to exclude the imitative branch of poetry; ‘imitative’ must, of course, be understood as meaning the style of the unphilosophic, unvirtuous poet, as explained in 397abc. IT. This result means that the first half of the tenth book is to be read as supplementary to, and in the light of, the third. There are at least two reasons why this way of regarding the tenth book is the most easy and natural. In the. first place Plato himself in 595a refers to his previous conclusion (which, in spite of the commentators, he has correctly stated) as something which he is now about to elucidate further. In the second place hymns to the gods and panegyrics on good men are retained in 607a without any explanation ; which, if we do Plato the justice of requiring an explanation, proves the necessity of seeking it in the earlier discussion. } | It is only necessary now to show that the tenth book really brings out some- thing which is implicit in the earlier doctrine. The imitative style condemned in the third book is that of the unworthy poet who is always putting on a character alien tohis own. In so far as he assumes the character of another— whether a shoemaker or a man of virtue or wisdom—he is imitative in the condemnatory sense.” Having no knowledge he does not see, and therefore cannot represent, the ideal forms which are in some way immanent in human character. His work will consist merely of words and actions which are an imitation (1) of a virtuous man or a shoemaker (2) who is himself in some sense an imitation of reality—the ideas (3). That is to say, he produces something which is thrice (inclusively) removed from truth. This is exactly the purport of Plato’s attack upon imitative poetry in the tenth book. He leaves it to be understood that the poet who is imitative in the sense in which the guardians are permitted to be imitative will produce a direct copy of reality ; he will be like the painter who uses the ‘divine paradigm’ (cf. 500-501 1 On 401-402 cf, Greene, pp. 37-38. Theimita- like the truth as possible (cf. 382d, ad init.). tion here mentioned is ‘clearly something more 2 It should be observed that what is formally than the literal kind of imitation that isimplied simple narrative may really be imitation in this in the tenth book.’ Cf. also Laws 816-817. The _ sense; for the poet may speak as though he knew ideal state is an imitation (817b piunors) of the the truth about subjects like philosophy—i.e, he best and noblest life, and the truest tragedy is will be putting on the character of a philosopher. essentially the same thing. Imitationas practised If he is not really a philosopher, this is imitation by the ideal poet will not produce something in the bad sense . unlike the original (cf. 388c), but something as 2 j | a “a 4 | | 20 J. TATE to which I shall refer below); not like the painter who is content to hold the mirror up to nature. In 596 sqq. Plato uses this analogy between imitative poetry and imitative (or realistic) painting. If such a painter makes a lifelike copy of (say) a shoe- maker, it does not follow that he knows the principles of shoemaking (598bc, 600e, 601a). Similarly, if such a poet represents—to the satisfaction of those as ignorant as himself—a virtuous man or one gifted with any quality or art, it must not be thought (as most men think, 598e) that he himself knows the truth about virtue and whatever arts he happens to represent. The mere imitator, whether poet or painter, produces what is thrice removed from truth—an imitation (picture or poem) of an imitation (the object as perceived by the senses, not as perceived by the rational part of the soul) of reality (the idea). The object, whether a bed or a man, is not perfectly real (597a); on one side it is real, in so far as it partakes in, or imitates, the real; on the other side it belongs to the world of multiplicity and change and therefore partakes in non- being. It is precisely the latter side that the art of imitation represents; it lays hold upon a partial and unsubstantial aspect of the object (598b); the copy is, then, a further distortion of what is already a distortion of reality. All this being true of the realistic painter (who aims merely at holding the mirror up to nature) is equally true of the imitative poet; for it is already clear from the third book that, having no knowledge of reality, he is an imitator in the same sense. The point is that the tragedian, in assuming a character alien from his own, is not merely open to the ethical criticisms of Book III; he is necessarily remote from truth, because he can copy only appearances, the words and actions of such characters (603c); he does not express, for he does not understand, the principles which underlie those appearances and which constitute reality. Both imitative painter and imitative poet represent only the externals, not the inner meaning of that which they imitate. This interpretation assumes that the analogy of poetry with painting is not meant as an argument, but simply as an illustration. The commentators who refuse to consider that the tenth book is to be read in the light of the third think that the passage is meant as a proof, and try to discredit it by alleging discrepancies in the analogy. Adam, for example (on 598b 11), says that the inference from painting to all imitative art is hardly to be justified. But there is no inference; and there are no discrepancies. From the instance of the imitative painter it is clear that remoteness from truth is a characteristic of imitation. Nowthe tragedian being an imitator in the same sense (as explained in Book III) is similarly remote from truth. This transition from painting to poetry gives rise to an important difficulty (597e). How do we know, if we leave the third book out of account, that the tragedian is an imitator in this sense? Thus Greene (op. cit. p. 53) complains that the question is begged in the assumption ‘that the mere definition of imitation in general will cover the aim of poetry.’ Of course, it covers only the I ha thirs imit spea hanc Plat poet! art fr imita char: realit 501, the (500d First as the justici chara amon; godlik T the ge becaus requir It kinds 1 597¢ reality tragedia: reality.’ All wh All tra There That is (which illicit s tributed who aré not seer first pre tion, so 4: ‘ : : ‘ old the imitative a shoe- (598bc, of those or art, it the truth imitator, ruth—an d by the he idea). ne side it ar side it ; in non- sents; it 8b); the lity. All 1e mirror lear from or in the lien from I; he is nces, the r he does id which only the ing is not itors who the third alleging ; that the But there ce of the eristic of explained difficulty , that the omplains inition of vers only reality.” Formally this is : ‘IMITATION’ IN PLATO’S REPUBLIC ai the aim of imitative poetry as described in the earlier aot by which, as I have shown, the assumption is fully justified. The conclusion must be that the tenth book is entirely consistent with the third in its treatment of imitation, and merely illustrates the relation of imitative poetry to truth which is implicit in the earlier discussion. The distinction between the good and bad senses of imitation is, broadly speaking, a distinction between two kinds of artist, the ignorant on the one hand and the enlightened onthe other. Here we have the key to understanding Plato’s attitude towards poetry and art in general. We have seen Plato using (in Rep. 596 sq.) an analogy between imitative poetry and imitative painting, in order to illustrate the remoteness of imitative artfromtruth. But there is an art which is not imitative in this sense, though imitative in another sense; an art which does not merely copy external characteristics of particular objects, but has some regard for the world of reality. The clearest account of this genuine kind of imitation is in Rep. 500- 501, where Plato compares genuine painting with genuine statecraft. The philosophers who desire to introduce the truths of the ideal world into the public and private lives of men will be no mean artificers of all virtue (500d sqq.) They will be, as it were, ‘painters using the divine paradigm.’ First they will outline the form of the constitution upon a clean canvas; then as they fill in the details they will turn their gaze now upon the ideal forms of justice, beauty and the like, now to the embodiment of these forms in human character. On the latter side of their work they will be guided by instances among mankind of that quality of soul which even Homer called godly and godlike. This sustained metaphor may be used to shed a light on the procedure of the genuine artist in every kind of production. We are justified in so using it because in 401-402 painting and poetry and all the arts and crafts are alike required to express the essence of beauty in all their works.’ It follows that there are two possible kinds of artist, and two corresponding kinds of imitation and art in general. One kind consists of lovers of beauty 1 597e runs thus: ‘Do you call the third from This would save the ‘syllogism,’ but still leaves reality an imitator?’ ‘Certainly.’ ‘Then the us with the formidable charge of ‘ begging the tragedian, being an imitator, will be third from question’ in the second premiss, 2 On the artistic expression of the ideal in All who are third from reality are imitators ; sensible form (which Plato sometimes calls ‘ imi- All tragedians are imitators ; tation’ in a good sense of the word) see further Therefore, all tragedians are third from reality. Greene, of. cit., pp. 66 sqq. On p. 41 Greene That is to say, if the passage be an argument rightly says that the theory of ideas ‘does not (which I deny), the ‘argument’ is formally an make art impossible; it actually explains that illicit syllogism, the middle term being undis- which is valuable in art.’ I cannot therefore tributed. It is strange that the commentators understand why Greene thinks that all poetry is who are anxious to discover discrepancies have condemned in the tenth book, and that the reason not seen thisone! However, we must take the for this ‘change of spirit’ is the discussion of the first premiss as stating a characteristic of imita- theory of ideas which comes between the earlier tion, so that the predicate is really distributed. and later discussions of poetry in the Republic. te gore 4 22 J. TATE and wisdom, who have some knowledge of the ideal world. Their work is really beautiful and harmonious, so that by the principle of imitation it infects others with harmoniousness of character (Rep. 522a). Though it does not impart scientific knowledge (ibid.), it is produced in the light of such knowledge (402bc). Thus poetry will be true as well as beautiful, so as to make the young familiar with right opinions (376e sqq.). Such art will be an imitation or expression of the reality of truth and beauty. But it would appear that in Plato’s view no extant poetry belonged to this class. The ideal state must in its own interests call a new race of poets into being. The other kind consists of imitation in the condemnatory sense. It includes, for example, the poets who copy only external characteristics of men, etc., the poets who do not know, but pose as knowing. It includes the painters who merely represent the appearance of particular objects of sense. Homer and the tragedians belong to this class. They are of the ignorant many (6014), and express, not the truth about virtue, etc., but the vague notions which prevail among the ignorant many (600e, 602b). If they had knowledge, they would not be imitative in this sense. But they have not even right opinion (534a, 602a). This second class will be utterly excluded from the state. Numerous passages in other dialogues are explained by, and in turn confirm, this interpretation of Plato’s attitude towards art. According to the Symposium (210-212), for example, it is only by communion with the essential beauty—the idea beheld by the eye of the mind alone—that a man will be able to produce really beautiful things, and not mere imitations of particular beautiful objects; for he has hold not of an image but of a reality (22a). In the Gorgias (502-503) orators are divided into two classes as above. The genuine orator, concerned to instruct his audience, had yet to be born—like the genuine poet in the Republic. The Phaedrus makes the same distinction in respect of poets and orators. The genuine poet or orator will base his compo- sitions on a knowledge of the truth (277-8). If by poet we mean a mere imitative trafficker in words—one who has not knowledge—then the genuine poet will be greater than poet ; he will be a philosopher (278cd). Similarly in the myth which distinguishes between different classes of men according to the degree in which their souls have beheld truth before falling to earth, the genuine poet is surely in the first class, which consists of lovers of beauty and of wisdom, men full of culture (music) and love (248d). Poets and other imitative artists appear in the sixth class (248e). Their inspiration is very low 1 This is the point of the additional arguments have been no mere man of words, but an excel- to prove Homer's ignorance in 598c-Gooe—a _ lent artificer of all virtue (cf. 500d). It may be passage which commentators (e.g. Adamon 599b) worth observing, in support of my interpretation, take as meaning that Plato would rather be that Proclus (In Platonis Remp. 405) sees clearly Achilles, the man of action, than Homer, the that if Homer contemplated the truth, he was man of words. Certainly Plato would rather be not imitative in this sense. For imitation, as Achilles if Achilles possessed the true heroic the word is used in Book X., is the companion of character—virtue, wisdom, self-control, etc.— ignorance (e.g. 598d dvemriwrnuoctvn and pulunors and were not as Homer represents him. But are coupled together). if Homer had been really wise, he too would in val accorc one wl stand Ii greate than imitat only tl The t kind c regarc under: world. recog! just a: mirro! kings | A com submi T copyir imitat merely words injuric T analog poetry mere sensib which 1 The placed whom t 60re), s value o deal ful but its right approve but whi knowlec right in denying 3 4 7 + ‘ 4 work is t infects Hoes not howledge ne young ‘ation or that in must in snse. It of men, painters Homer y (6014), s which ge, they opinion > in turn x to the sssential be able irticular za). In e. The like the in compo- a mere yenuine larly in x to the th, the ity and 1 other ery low an excel- t may be yretation, Clearly he was ation, as anion of d ulunocs ‘IMITATION’ IN PLATO’S REPUBLIC 23 in value if compared with the creative love of the philosophers.’ Again, according to the Laws (667d-669b) there are evidently two kinds of poetry: one which asks to be judged by ‘pleasure and false opinion,’ and one which can stand the test of comparison with the truth. It is clear, then, from these and other passages,” that there is a far greater degree of consistency in Plato’s utterances regarding poetry and art than commentators have hitherto allowed. To sum up, there are two forms of imitation according to Plato. The first is the merely imitative, which imitates only the apparent nature of things which are apparent to the senses (Rep. 598b). The tragic poet who assumes a character alien to his own is doing the same kind of thing as the painter who represents the appearance of objects without regard to their nature or meaning. They are both imitating what they do not understand. The second form of imitation is that which imitates the ideal world. This form can be achieved only by the man of understanding, who can recognize both the ideas in themselves and their images in the sensible world, just as one learns to read both letters and the images of letters distorted in a mirror or in water (Rep. 402bc, 500d sqq.). Thus the poets must, like the kings (499bc), become philosophers; or else philosophers must become poets. A compromise is, of course, possible, if the poets who are not philosophers will submit to the censorship of the philosophers (Laws 719, 816-7, etc.). The first kind of imitation is imitation in the literal sense—the mere copying of sensible objects. The poetry which is not also philosophy is imitative in this sense; when it depicts an Asclepius, for example, it copies merely the words of an Asclepius* without knowing the truth which the words express. This mere poetry is always and necessarily false and injurious.* The second kind of imitation is not imitative in the literal but in an analogical sense. Accordingly it is easy to understand why Plato regarded poetry of this kind as non-imitative (in the Republic) in contradistinction to mere poetry. For ‘imitation’ when applied to the expression of the ideas in sensible form is only an unsatisfactory metaphor, as are all the other words which Plato used to denote the same relation.®° J. TATE. 1 The fact that the imitative poets are here allowed them ‘genius and inspiration’; but we placed one class higher than the artisans, to should also remember the low estimate which whom they are inferior in Rep. X. (cf. Adam on - 60re), seems due to a more lenient view of the ' value of inspiration. I have no space here to deal fully with Plato’s treatment of inspiration ; but its highest gift is not knowledge, but only right opinion, which enables Plato indeed to approve of many passages in Homer (Rep. 383a), but which is a poor thing when compared with knowledge (Meno 98-99). Adam (on Rep. 598e) is right in asking us to remember that Plato, while denying knowledge to Homer and kindred poets, Plato placed on inspiration without knowledge. 2 One might quote in support even the early Charmides (173): there are pretenders, deceiving and self-deceived, in every art and craft; only if wisdom rules shall we have the true physician, etc. 3 Rep. pdbvov larpixay Adywvr. 4 Cf. Rep. 387b roinrixwrepa, Tocotry adxoveréov. 5 uébekis, wapovola, etc., see Greene, p. 66. | Be rig ‘ . = LUCAN, STATIUS, AND JUVENAL IN THE EARLY CENTURIES. THE histories of literature tell us that these three poets were out of favour with scholars during the second and third centuries and the first half of the fourth. Lucan and Statius certainly had a vogue in the first: Suetonius studied Lucan at school (Vita Lucant ‘ Poemata eius etiam praelegi memini ’), and Statius (Theb. XII. 814-5) can say to his book: ‘Iam te magnanimus dignatur noscere Caesar, Itala iam studio discit memoratque iuuentus.’ But the revulsion of taste in the second century towards older writers no doubt affected them as well as Juvenal, who was nearer to it in time. The question is to what extent they were neglected and how long this neglect continued. Wessner (in Teuffel III. 431, 2 and 3; cf. Schanz, II. 2, pp. 175 and 211) regards them as having been totally eclipsed so far as schools and works of learning were concerned, and attributes their restoration to Servius, who often quotes them in his commentary on Virgil. The matter is of importance in connection with scholia on Virgil and their relation to Donatus; Wessner repeats his statement (particularly as to Statius) in Corp. Gloss. Lat. I., p. 378. If these poets had been banished and it was Servius who brought them back, then, of course, a scholium on Virgil (or Terence), which quotes one of them, cannot be as old as Donatus. I wish to examine the evidence for this opinion. The first suggestion seems to have been made by Kiessling, and this was taken up by R. (Halfpap) Klotz in two publications, Quaestiones Seruianae (Greifswald, 1882) and De Scholits Statianis (Treptow, 1895). I shall take his arguments one by one. I. Jerome is the first writer who mentions a commentary on Lucan: adu. Rufin. I. 16, ‘ Puto quod puer legeris Aspri in Vergilium et Sallustium commen- tarios, Vulcatii in orationes Ciceronis, Victorini in dialogos eius, et in Terentii comoedias praeceptoris mei Donati, aeque in Vergilium, et aliorum in alios, Plautum uidelicet, Lucretium, Flaccum, Persium atque Lucanum.’ He does not mention a commentary on Statius (or on Juvenal). Therefore, says Klotz, Statius was not read in schools at that time. And therefore also, Wessner would add, Donatus did not quote Statius in his commentary on Virgil. But can we assume that Jerome is giving a full list, that ‘aliorum in alios’ is completely specified in what follows? There is nothing in the context to indicate that. And if it were so, did the commentator on Virgil confine his illustrations to authors who were read in schools? Were, for example, the minor they o scholia hithert some ¢ on Ae quia u Comm numer conue! metho imitar school owed | the la Lucar 2 since Pauli argue but ti At all thoug 3 if the 4 ofad in the survi' is ac first ¢ to go there date ! Juve are— P | at } 3 * “ W es: S of favour alf of the Suetonius nemini ’), no doubt question yntinued. und 211) works of ho often ‘tance in Wessner » Pp. 378. m back, of them, for this his was rulanae take his n: adu. mmen- Terentii 1 alios, le does Klotz, /essner But ios’ is ext to ine his le, the H. J. THOMSON 25 minor playwrights of the Republic read in the schools ot Servius’ time? Yet they often appear in his commentary. The miserably fragmentary Verona scholia on Virgil quote no fewer than twenty-six Latin authors. Lines or phrases hitherto unknown of minor poets have been found in glosses which come from some commentary on Virgil... As for Lucan, we have a statement by Servius on Aeneid I. 382, ‘Lucanus namque ideo in numero poetarum esse non meruit quia uidetur historiam composuisse, non poema,’ which is paralleled in the Commenta Bernensia on Lucan I, 1, ‘ Lucanus dicitur a plerisque non esse in numero poetarum quia omnino historiam sequitur, quod poeticae arti non conuenit.’ This appears to be only a characterization of Lucan’s subject and method, just as Quintilian speaks of him as ‘ magis oratoribus quam poetis imitandus’; not Servius’ way of saying ‘ Lucan was not recognized in the school list till I admitted him.’ And in any case, as we know now, Servius owed so much to Donatus that this statement is at least as likely to belong to the latter as to the former. There is nothing in Jerome’s words to suggest that Lucan was a recent addition to the school authors. 2. Statius and Juvenal are never quoted by Festus. How should they be, since Festus was in the main merely epitomizing Verrius Flaccus? Festus’ Pauli does indeed once quote Lucan (31, 14 L), and Klotz feels it necessary to argue that this quotation must have been introduced by Paulus, not by Festus; but the only ground for doing so is the very theory which he wishes to prove. At all events, we should not expect Festus to quote Lucan or Statius or Juvenal, though a stray quotation need not cause any surprise. 3. These poets are not quoted by Gellius. But should we not be surprised if they were quoted by a writer of Gellius’ tastes ? 4. Nor by Nonius Marcellus: but he again was concerned with literature of a date earlier than theirs; and, as it is quite uncertain whether Nonius lived in the fourth century, he is no sure instance of the spirit of the second still surviving in that age. 5. Nor by the grammarians Charisius and Diomedes: but Charisius’ work is a compilation derived from earlier grammatical writings, some of them of the first century ; and Diomedes is closely related to him. 6. Nor in Probus’ commentary on Virgil. We have little enough of that to go upon, and in what we have many more Greek than Latin authors are cited ; there is no mention of Plautus nor Terence nor Horace. And think of Probus’ date! He could have quoted Lucan, but hardly Statius, and certainly not Juvenal. 7. Nor in the Verona scholia on Virgil: but how fragmentary these scholia are—often the merest scraps ! In all this there is little ground on which to base a statement so sweeping as Wessner’s. On the other hand, the Scholia Pithoeana on Juvenal are held by Leo to be built upon a commentary of the second century (Hermes, 1909, p. 603). 1 See Amer. Journ. Phil. 43, 353; and cf. Medieval Latin Glossaries (Longmans, 1925), pp. Mountford, Quotations from Classical Authors in 127 sqq., Nos. 250, 253 ?, 258. | | : ON = i | ; 5 er ~ ~ 26 H. J. THOMSON One of the main reasons why it has been thought that the basis of Lactantius Placidus’ scholia on Statius cannot be older than the end of the fourth century is simply that no commentary on Statius is mentioned in the list quoted from Jerome. Another is that it is only from Servius’ time that Statius is quoted by commentators. What pre-Servian commentaries can we be said to possess for comparison? Even if it were granted that scholars of the earlier fourth century neglected these three poets, there is no evidence to show that Donatus did not quote them, whereas what we know of the relation between Donatus and Servius (along with the Daniel additions) makes it prima facie likely that he did. Indeed Klotz does not go quite so far. He argues from Servius on Aeneid XII. 365' that Donatus did not quote Statius on this line; and certainly it looks as if he did not, though, remembering the character of his commentary, we can hardly be sure. On the other hand, it seems clear from the note that Donatus did base his opinion on a quotation of Lucan, as Klotz admits. Lastly, there are only a few a places where the Daniel additions to Servius mention these poets; but as there ave a few, Klotz and Wessner must cast suspicion upon them, though apart from their theory these passages seem no more open to suspicion than many another. In one place they have Thilo with them. He inclines to think (praef. pp. xxxiv-v) that the D note (quoting Juvenal) on Aeneid II. 445 really belongs to Servius, and has dropped by accident out of the MSS. of the ‘vulgate’ Servius. His reason is that similar remarks on the declension of domus are found in Servius’ commentary on Donatus’ grammar and in two fifth- century grammarians, Cledonius and Pompeius, who took much material from Servius on Virgil, whereas in other places where he has compared them Cledonius does not agree with D. Buton socommonplace a topic, with which every teacher must have dealt in the same way, arguments of this kind have little force. The worst that Wessner has to say about these passages in D is not very bad. He mentions as one of the marks that distinguish D from S ‘die geflissentliche Nichtbeachtung der von S. bevorzugten drei Dichter,’ and adds ‘die vereinzelten Zitate ...sind eben in ihrer Vereinzelung nicht unverdachtig.’ But in the light of all recent investigations the scarcity of references to these poets in the D notes does not in itself imply anything more than that practically all the references to them in Donatus had already been taken over by Servius. And if I am right in thinking that Donatus’ original commentary was not the immediate source of D, that only strengthens the present argument ; for if the references to Lucan, Statius, and Juvenal had been added to Donatus by the person who went over his work and produced the version from which D was taken, because they had come into fashion in the interval since Donatus wrote, there would surely have been more of them in our D notes. An apparently stronger argument for Wessner’s view might be 1 ‘Sane sciendum hoc loco errasse Donatum, que certum est systolen fecisse Lucanum... qui dicit ‘‘ Edonii’’ legendum, ut ‘‘do"’ breuis Statius (cf. Theb. V. 78) et Vergilium et artem sit, secundum Lucanum qui dicit (I. 675) secutus ait ‘ tristius Edonas hiemes Hebrumque ‘* Edonis Ogygio decurrit plena Lyaeo:’’ nam- niualem,’’ non “ Edonias."’’ LUCA founded on Teren loses muc comparat source, tl almost al excerptor Terence, for the ill proved. BaNnGo 5 : 3 antius entury d from by ess for entury lid not ervius ie did. XII. oks as ve can onatus there these them, Dicion nes to I. 445 of the on of fifth- from them which have D is om and nicht ty of more been ginal ; the been | the 1 the m in it be LUCAN, STATIUS, AND JUVENAL IN EARLY CENTURIES 27 founded on the want of references in the excerpt version of Donatus’ commentary on Terence. (Lucan is quoted on Eunuchus 348.) But on consideration it loses much of its force. The range of quotation in the excerpt commentary is comparatively limited. Virgil (with Terence himself) is by far the commonest source, then Cicero and Sallust ; a few other authors are quoted occasionally, almost all of them pre-Augustans. The limitation may not be the work of the excerptor; but even if Donatus found these few authors sufficient for his Terence, it does not by any means follow that he would be content with them for the illustration of Virgil. On the whole I submit that the Klotz-Wessner theory is far from being re: H. J. THOMSON. Bancor, NortH WALES. y 2 9) 4 arte ae if NOTES ON PLOTINUS (II.). Ennead 11, 2. 1 (Volkmann’s edition, p. 131, 1. 8). As a reason for the cosmic revolution Plotinus advances the curious argument that ‘the soul of the universe being omnipresent permits the heavens part by part to be omnipresent as far as they can; and this possibility they achieve by passing through and traversing every part.’ The text continues éorn peév yap, ef mov eotaca hy EADdv viv raed épierac: ‘for if the soul were stationary in any place, the heavenly fire would have become stationary there; but in fact, since the soul is wholly self- possessed, the fire strives towards every part.’ No translator ventures to render the words wacd éotw airs correctly, because in the context they obviously make no sense. Ficinus gives ‘cum anima totam se per totum fundat,’ and his successors have something similar. A very slight change restores the meaning: raed éore rdvry (or rdévrn, to employ the usual spelling), ‘since the soul. is al] everywhere.’ For mdvtn equivalent to ravraxov in Plotinus see VI. 8. 16 (p. 500, 1. 6). 2. 3 (p. 133, 1. 8): 6 S€ vovs yap Kai mepi avtdv ydp. We have just been told that the phase of the soul concerned with sensa- tion takes its good from above, and as that good is everywhere, scours the universe in pursuit of it. Spirit, however, clearly does not ‘move thus.’ It moves ‘ about itself,’ not towards an outside good. Hence oirw must be wrong, but Kirchhoff’s ovrw is equally wrong, being contradicted by xcveira: in the next clause. Read there- fore otrw, vos moves indeed, but as it moves about its own centre it may also be said to rest; cp. V. 1. 4 (p. 166, 1. 19). In the same way, Plotinus adds, the revolving universe both moves and rests, év tw xeveio Oar 7d pévery Exov, as he reiterates in IV. 4. 32 ad fin. Enn. II. 4. 14 (p. 164, 1. 10): pévroe 7G dopioty Kal Kai TH Ay TavTov, was Ere of Adyot. No good sense can be extracted from these words as they stand. Most of the proposed renderings imply ro dépicrov eiva, ‘ because something, privation for example, is indeterminate.’ The meaning required is ‘if essential indetermination .. . is identical with matter,’ and the familiar Aris- totelian Greek for the essence of a thing is 7d «Zvac with the dative, an idiom which occurs on the next page of Plotinus: 17d avrd obv 7d drecpov Kai 7d areipy (p. 165, l. 13; seealso 1. 16). For dopiorw elvac read, therefore, 7d dopioryw efvar. That the ‘essence’ of the three attributes mentioned is intended, is proved by the emphatic repetition of efvat, Aristotelian idiom is not a strong point with editors of Plotinus. In. II. 1. 6 (p. 125, 1. 26) they have failed to correct 7d ri Fv efvax vp where the dative yp is clearly necessary. Eun. 4. 16 (p. 166, 1. 18): ov« dv dpa avry erepov dv Tw Tod dvTos. ‘Matter therefore is not Being, because other as compared with (‘ different ¢o’ in fact) the beauty of Being’ is the meaning required, but no editor has had the courage to alter rpds 7@ KaA@ to pds 7d kaAdv, which must surely be what Plotinus wrote. Enn. Il. 9. 11 (p. 200, 1. 19): wAdopaTe Tas Toinots; TovTi pév mpwotov, Se per’ éxeivo, dAXr’ ws ELovoias A€yovres. Editors have been much puzzled by this passage, but if aAAws be read for dAX’ ws, most of its difficulty dis- The Gnostic doctrine that the sensible world is created by an évvénpa is appears. ‘ condemne necessary how does else after argument Adyov whether grammati Enn. rational ? TEWS EVTTOF dependin; with a d believe, « TEPUTTACE regarded alteratior lV. 8. 4 | Plotinus Korot a quite re oTacis ¢ Beors, priate in Enn GAAnAa refer. | l. 4 En Deis offended ‘ unless En ei TUTroL last wo! objects: rept 6 is obj regard | the imp En impress impres: continu rd would | = 4, Li * ve G * = = 4 > cosmic universe r as they ry part.’ Tard leavenly olly self- ider the nake no ccessors Tt For 1 sensa- niverse ‘ about shhoff’s 1 there- it may ds, the iterates amroiw 1 these y eivat; quired r Aris- which p. 165, iat the phatic ytinus. dative OVTOS. to’ in urage Ti much y dis- is J. H. SLEEMAN 29 condemned by Plotinus on the ground that an évvénua cannot have the active powers necessary to a creative agent. ‘ Not only is the whole thing a fiction,’ he says, ‘ but how does the creation proceed? First, they reply, it creates this, then something else after it, merely talking arbitrarily ’ (see p. 201, ll. 3 ff. for a development of this argument). A precisely similar use of dAAws is found in II. 7. 3 (p. 180, 1. 30) rav Abyov rovrov, ei pH GAAws Gorep Opicpds SnAwrtixds Tod ti Td tpaypa, I doubt whether any further alteration of the passage is warranted. Its abruptness and grammatical incoherence are well within the capacity of Plotinus. Enn. IV. 3. 18 (p. 30, 1. 22). Ifsouls yonder never reason, how can they still be rational? To this question Plotinus replies 7 Sivavras, tis dv, evropnaa (C and A pr. m.: evropior the rest). drav evTopags Seems unmeaning, and there can be no doubt that depending upon Svvavra: is correct, the sense being ‘ possibly when souls have met with a difficulty they can solve it by examination.’ The word repiwrdceuws is, I believe, corrupt. In IV. 4. 25 (p. 74, l. 3) we find oodpaiver Oar xai yeverOar yvpov wepurtdcers av Tis Kal mepreAxvopors THs Oeiro, ‘smelling and tasting may be regarded as distractions of the soul,’ where sense and balance almost compel the alteration of repurdces to repurrdoces (distractiones ductusque, Ficinus). So in IV. 8. 4 (p. 147, L 27) 75 xaOéxacrov pera mepiotdcews Stovxet (Of the fallen soul) pera wepuordcews Should probably be read (distractus, Ficinus), In IV. 3. 18 therefore Plotinus may well have written érav <év> Siacxorotca, Srav mepioracGaow would do equally well, for which cf. Phrynichus tives Eri TOU év doxodig wavy KiBdjAws, though as a fact quite respectable authors use the word thus. Neither the ordinary sense of zepi- oragis ‘ circumstance,’ especially,‘ critical circumstance,’ nor the special meaning of d:dBeoxs, ‘ disposition,’ given it by the Sceptics (Sext. Emp., Pyrrh. I. 100) is appro- priate in any of these three passages of the Fourth Ennead. Enn. IV. 5. 3 (p. 106, 1. 1): rod (wou dvatpeiras rpds Kal mpds &AAnAa tov pepov. There is nothing in the context to which avra can conveniently refer. Read, therefore, aird, ‘the sympathy of the living whole with itself,’ and cp. l. 4 (Gov 7d Tay EavTy. Enn. IV. 5. 5 (p. 109, 1. 1): py anp wAnyeis ev TH Taxeig ovvddw avTay Kal éEwoOeis eSuxe TH péxpis Grwv. It is strange that editors have not been offended by the unmeaning wAjfas, for which wAyydés must obviously be read ; ‘unless the air forced out struck upon successive layers of air as far as our ears.’ Enn. IV. 6. 2 (p. 116, 1. 13). The visible and the audible can be discerned ov« riot GAN’ ef py pdt weivers Evepyeras O Everor The last words Mr. MacKenna translates by ‘ there can be only acts of that in which the objects become known.’ But ‘acts of’ should be évépyesas followed bya genitive, and wept 6 is an impossible attraction for wepi éxeivo ». Moreover, Plotinus is arguing that the objects are not im the mind at all. I therefore suggest wepi 6 éreo, ‘acts in regard to what comes to the mind.” We are told below (I. 24) that the mind reads the impressions on the air when they have come near enough to be perceived. Enn. IV. 6. 3 (p. 119, 1. 16). Plotinus is contending that sensations are not impressions (rior), and that therefore memory cannot consist in the retention of impressions which ex hypothess were never made. Good memory and quick wit, he continues, do not as a rule go together. éxwAve tov éxovra reovefias dvaywdoxev Ta Keipeva, ovdé Tov TavTy prevTa THY TOU Kal 7d advvapiav ‘and yet (if memory were the retention of rvmo) there would be nothing to prevent persons with the highest qualities of mind from reading ie. = 30 J. H. SLEEMAN the stored-up impressions, nor one inclined that way (i.e. an intelligent man) from being incapable of receiving and retaining an impression.’ There is clearly some- thing wrong with the last clause. The meaning which we want is ‘nor on the ee of tuo: need an intelligent man be incapable, etc.’ Read, therefore, ové for ovdé. V. 1.2 (p. 163, 1. 3): jovxos Hovyos Kai Kai airds odpavds dpeivwv, Plotinus is asking us to imagine a universe asleep before the world-soul dawns upon it. The only plausible correction of the absurd dyeivwy is that of Professor Dodds (C.Q. XVI., p. 96), dua pévwy, which seems to me nearly, but not quite, right. dua is not wanted. The universe is just ‘waiting’; and one of Plotinus’ favourite words for ‘ waiting’ is dvapévwv, which I feel sure should be read here. Cp. VI. 8. 9 (p. 489, 1. 23): dvapévew Sei ra GAA ti wore adrois 6 BaciXreds pavein. Enn. V. 1. 5 (p. 167, 1. 12): pev Suds olov AapBavo- pevn, 0 Se aprOpds o avris Kai tov évds Exacros oluv poppwhévros yevopevors ciéeow é€v avrg. Plotinus is describing the generation of vois, which having an element of plurality in it is an dpcOyés, and as mathematical dpOyoé are derived from the mathematical unit, so is it derived from the transcendental One. He continues (p. 167, 1. 11), ‘number, so-called, in yonder world and the Dyad are Reason- Principles and Intelligence’—that is to say, Intelligence and its contents are separable into these elements of form and matter respectively. Then comes the Greek quoted above. The words rw ofov iroxepévp have neither sense nor con- struction, and should be emended to v@ ofov tmoxeiyevov, Volkmann’s change of to éxaorov is wrong—Plotinus may, however, have written ékaocrov—and poppwévros agrees with vod understood. Thus we get, ‘ but while the Dyad taken as substrate for Intelligence is undetermined, the number arising from the Dyad and the One is in each case a Form, Intelligence being, so to speak, shaped by the Forms arising in it.’ This doctrine that in vovs we can distinguish a material from a formal element, the former being the Indeterminate Dyad, Plato’s dmeipov, or ‘great and small’ (cp. Aristotle, Met. M. 1081a 14), the latter the One, has already appeared in V. 4. 2 (p. 204, 1. 13)—according to Porphyry an earlier tract. Enn. V. 1. 5 (p. 167, 1. 18): €ore yap vénows dpacis audw rd év. Plotinus has just said that vovs is shaped (a) by the One, (d) by itself, like vision in act, which is similarly determined by subject and object. The words dydw rd év apparently conceal a statement that subject and object are identical. A comparison with IV. 6. 2 (p. 117, 1. 3), yuxn ... Opa... ra &v, suggests that Plotinus may have written dpoca adudw ta év here. Another possibility is to read dudw te &v as a Separate clause. Enn. V. 2. 1 (p. 176, 1. 20): «ios rovro avrov, dorep aitd mpdrepov mpoéxee. Kirchhoff for aird substitutes 73 rpd, but it is a much easier cure to read atrov mpdtepov. rovro, the Sivayis roAAy poured forth by the Divine Intelligence, is Soul, and the passage means, ‘this, too, is a form proceeding from Intelligence, just as that was poured forth by its own prior, the One.’ Enn. V. 3. 4 (p. 182, 1. 9): Eavrods TH ToLovTw Ta GAAa paOeiv. ws must have dropped out before ra aAAa; ‘and in virtue of such a vision (the vision of vods) we know ourselves, so that we understand all else as well.’ Enn. V. 3. 10 (p. 192, 1. 30): mpd yap vonoa: drep eotiv The Intellectual Principle, says Plotinus, is dual, not simple, yet it will not need to be curious about itself, as it has nothing to learn by the intellective act. Then follows the reason quoted above: ‘for before (it is) Intelligence, it can have intellection of what neverth of vow: that th En éyévero Mr. Ma which ; a mista became im unity Possibl En you €v T avrov. I have being tl mare pd Then if constru En gramm: V. 5.1 wrote substitv En before « Good r have us Td ovv 7 used of in the v En: Aéyous ; Cp. IV. Eni Kal Ta K Td ingenioi change creature which t! to the / p. 207, | to colou (So Mr. e.g. VI. see II. ¥, | J aa 4 in f 4 : nan) from rly some- rr on the fore, ovd’ ds ovpavds rorld-soul ; that of but not 1 one of be read AapBavo- yevopevors an ved from ontinues Reason- ents are mes the 10r con- ange of ‘ov—and d taken yad and > Forms a formal eat and eared in Plotinus which arently with >lotinus to read pOTEpov to read Divine ig from vision . Lhe l to be follows lection NOTES ON PLOTINUS (II.) 31 of what it is.’ But there is no sense in saying that what is not yet intelligence can nevertheless exercise it; cp. p. 193, 1. 20, vots otrw vojoas; p. 195, 1. 17, . . mpd Tov vonwat voyoe (an absurd supposition). Read, therefore, rod in place of vow: ‘ for before exercising intellection it has its own content.’ Plotinus means that the self-contemplation of Spirit involves no discovery of new knowledge. Enn. V. 3. 11 (p. 193, 1. 17): otros obv 6 ewéBare pév exeivy, AaBdv éyévero vous, det évdiaBépevos Kai yevdpevos Kai vovs Kai ovoia Kal vénows Gre évdnoe. Mr. MacKenna takes évésaéuevos to mean ‘absorbed’ in the One; but the word, which apparently does not occur elsewhere, is probably corrupt. May it not be a mistake for év dvo Oéuevos? ‘This Intelligence concentrated upon the One, and became Intelligence by attaining it, and by eternally forming and becoming a duality in unity, is Intelligence and Essence,’ etc. Cp. p. 192, 1. 27, d«xdoe yap avrd éavrd, Possibly we should read <1r> dre événoe and take the nouns with yevdpevos. Eun. V. 3. 13 (p. 196, 1. 4): 7d eréxewva Kal éréxewa TOU TEeuvordrov év Tots mace povov GAnGes Svopa] dv [avTod odre Ti TOV Svopa. avrov. Miller omits everything from év rots to dAAo 71; Volkmann the words which I have bracketed. Volkmann’s reading involves damnable heresy. So far from being the only true év, the One is not év at all. voisis dv. So read vot é rois Tac. povov dv, omitting ov« dvoya as a blundering anticipation of ovre dvopa. Then if ovre dAAo te is read for avrov aAAo tT, the sentence is correct in meaning and construction. Enn. V. 3. 17 (p. 201, 1. 26): 8 Kai trav . . . yévorro. The grammar of this can be corrected by writing «dv for xaé. On the other hand, in V. 5. 1 (p. 206, 1. 6) there is a superfluous dv. We need not suppose that Plotinus wrote Se dpa ... dv wore, when can be so easily substituted for yd’ av. Eun, V. 6. 5 (p. 226, 1. 9): 7 yap voeiv od rparov . . . dAAd Sedrepov Kal yevopevov, before éxivnoe, but the word should be kept, and iréornee written for iréorn: ‘for the Good made it subsist, and moved it when created towards itself.’ Plotinus could not have used the aorist iréorn, ‘ came to exist,’ of the Good; cp. VI. 8. 10 (p. 492, 1. 6), 7d obv trocrdcews Tas Gv tm’ Vd’ however, és used of the Good, e.g. p. 493, 1. 13; p. 496, 1. 26; but there is all the difference in the world between the two tenses. V. 7. 3 (p. 229, 1. 29): mas obv roddGy Sidipwv Siaddpovs rods Adyous; Why ‘many’ twins? zoAAdv is almost certainly a blunder for zaidwv. Cp. IV. 7. 5 (p. 125, 1. 31), dééupa yevvnpara, Eun. V. 8. 2 (p. 232, 1. 17): ri KdAXOs eoriv év ; ov yap Td Kal Ta KaTapynvia, GAAG Kai xpoa GAAn TovTwv Kal OvdeY TL 7 olov 7} meptéxov amAovv Tt, ofa An. Professor Dodds has emended this passage by ingenious additions, but I doubt whether any alteration is necessary except the change of tAn to tAnv. The substance of the argument is that beauty in living creatures, being form, cannot be due to matter. 1a xarayjvia—and the blood of which they are the product—can have nothing to do with beauty, because, according to the Aristotelian doctrine, which Plotinus accepts in the clearest terms (II. 9. 12, p. 201, ll. 14.sqq.), they are ‘matter.’ For the explanation of beauty we must look to colour and form distinct from these, ‘or there is nothing unless sheer ugliness.’ (So Mr. MacKenna, rightly, I believe, as the ellipse of dAAo is constant in Plotinus— e.g. VI. 1. 29, 30, p. 297, l. 22, p. 298, 1. 23. ‘Sheer ugliness’ is of course matter; see II. 4. 10 and 16.) Or, again, beauty may be described as what embraces ere 32 J. H. SLEEMAN (i.e. form) a simple something as matter. For matter as simple see II. 4. 8, p. 156, 1, 31, Set 52 adriv pi civOecrov Kai & te TH airns pice: yap TavTWY If readers doubt whether Plotinus or anybody else could have written passages like the above, let them remember Longinus’ complaints of the faultiness of his text of Plotinus, though that text, according to Porphyry, was a faithful copy of the master’s own manuscript (Porphyry, Vita Plotsns, 19, 20). Enn. V. 8. 10 (p. 243, l. 32): radra obv épav 6 Zebs xal ef Tis cvveparTis Td teAevtaiov 7d KdAAos Kai Tod éxet. This sentence has not, and probably never had, a main verb. dpév is for épav éore by an ellipse common in Plotinus. If for épapmevov we read épav pévov we get a reasonable sense : ‘ Zeus therefore beholds this vision as well as any one of us who shares his love, beholding at last the beauty abiding entire on all things, and winning part in the beauty there. For beauty ‘abiding’ yonder cp. VIII. 8. 1, p. 231, lL. 18, GAN’ éxeivo pev peéver, A few lines below in the passage ov yap ére 7d pév ew, adrd Oewpevov éfw, 7d pav is the thing seen, 7d aird Ocdpevov is that which sees it. I see no reason to alter the reading of the MSS. to o pev «fw, rd 8 ad ew, omitting Gewpevov, with Miiller, or to commit the solecism of taking Oedpuevov as passive with Volkmann. Enn, VI. 1. 11 (p. 277, 1. 14). Rough and smooth, rare and dense, cannot rightly be called qualities, od 5) tais dm’ Kal eyyis Td pavov kal 7d Kal Tpax’Tns Kal ov dvwpyadias Kai «i kal rovtwv, ovdév Kwrvec ws mod The insertions which I have made in brackets seem necessary to the sense of the passage, and bring it into accord with Categories 10a 21, where Aristotle gives the precise explanation of these attributes, which Plotinus here rejects. And Plotinus rejects it because it involves an atomistic theory of matter, as is clear from Simplicius’ commentary on this passage of the Categories. If the Aristotelian explanation were right, Plotinus implies, he would have no objection to describing roughness, smoothness, etc., as qualities, but for him they are, in general, not qualities at all, What, then, are they? The answer must be that, where they are specific differences, they are ‘ rather actualities, reason-principles or parts thereof, indicating essence’ (VI. 1. 10, p. 274, l. 27; cp. p. 277, 1. 10); for a specific difference is not properly a ‘quality’ in Plotinus’ sense of the word. Where, however, they are external and more or less accidental attributes, obviously due to ‘evenness and unevenness of position,’ Plotinus is prepared to regard them as qualities. So Iamblichus (in Simplicius loc. cit.) distinguishes two kinds of density and rarity—one év Trois 6 maperre kat ovowwdées cots, like the rarity of air and fire, the other érixrnrov xai like that of a sponge. Enn, V1.1. 24 (p. 291, 1. 13): év kai ov... . dvaxeio Oat, xa@jcG0. So Volkmann prints, assuming a lacuna. But dv is surely a blunder for ofov, which, if written with the tachygraphic sign for -ov, might easily have been misread. ‘To take next the category of xeicGau, it too (like that of éxyev) applies only in a few cases, e.g. dvaxeio Oa,’ is exactly the sense required. Enn. V1. 1. 30 (p. 299, 1. 1). Plotinus is arguing that acting and being acted upon cannot be reduced to the Stoic category of wws éxov. émei ovd’ 6 roy Tws Exwv, GAAG padrdAdv tws rowdy 7 GAws ovK dv, GAAG pdvov* Kal 6 Tada Ov Tws Exwv, tacxwv ovtws. For the unintelligible ovx ov I suggest ov mws; ‘the agent is not disposed in a particular way, he is rather acting in a particular way, or indeed not in a particular way, but acting simply.’ ovtrws, the last word of pO En TOUTO, less Aey into them to be guide Ens 7} a first du distingu: troBeBA for Bn Eun 7d pey & passage at least : prove it beauty | because at least to A«y is said tc wpoiwras Enn will sho must gra THE | mig > : 2 3 4 we , 4 156, yap passages his text y of the VEPATT?) This Te by an asonable ares his ing part 1, 18, Gewpevov see no omitting ive with , cannot > eyyds is isertions nd bring ation of cause it mentary Plotinus —etc., as hen, are * rather » P- 274; lity’ in or less sition,’ mplicius de? GAov i Ban, nder for ve been lies only g acted "WS EXWV, wv, suggest ng in a the last NOTES ON PLOTINUS (Il.) 33 word of the extract, should be emended to drAds, récywv arAds thus balancing povov. Eun. VI. 2. 11 (p. 314, 1. 20): ra rais réxvais, ait) Exdotn mpds rovro, Kal’ Sivara, ws Sivatas éxeiva, otrws Miiller alters the meaning- less Aeyee to woret. The right reading, I believe, is aye, which could be corrupted into Aéyes more easily than vovet; ‘as for the products of the arts, each art gusdes them to this (the One or perfection) according to its capacity and so far as they can be guided.’ Enn. V1. 3. 10 (p. 339, 1. 9): am’ GAAns dpxjs troBeBAnxvias Siadopday cvvGérov 7 Toros 7) poppais roeioGar, Plotinus is speaking of the logical d:atpeors of bodies. In a first dva/peors we may divide simple bodies, then after mixing simple bodies ‘ we may distinguish the resulting composites on some other principle, either by place or form.’ troBeBAnkvias is accepted by all editors, but gives no sense. It is surely an error for troBeBynxvias, ‘subordinate,’ ‘secondary.’ Cp. III. 5. 9 (p. 279, 1. 12) 7d imoBeBnkds vod. V1.6. 7 (p. 406, 1. 7): rurrovpeBa rpds robs reOavpaxdras éx TOV 7d péyeDos adrov Kai 7d Epwrt rpds (adrdv K) xai trav dAAwv Tov eis Yuxnv épwra dia tHv Kai TH KaTd Te Mpoiwrat. This obscure passage occurs in a description of the avrofwov or intelligible world. Probably rq «is yuxiv épwre and wpowwoGa: should be substituted for the reading of the MSS., when at least a tolerable sense results: ‘ to those who marvel at this intelligible world we prove its existence from what participates in it, and (we prove) its greatness and beauty by the love of soul towards it and by the love of all other things for soul, because their nature is such as we have described, and because they possess soul, or at least are in some respect assimilated to it.’ 6a rv rovatrnv diow looks back to Acyouévyn iors jv Kai id’ Fs Exarra yevvaras (p. 406, |. 1), which, like soul, is said to ‘imitate’ vovs. If with Kirchhoff we omit 7 before xard, rq éxew xa? dre wpoiwrat can be read. Enn. V1.6. 10 (p. 410, 1. 10): roivuv Secxréoy. A reference to the context will show that the meaning required is not ‘we must prove universally,’ but ‘we must grant universally.’ Read, therefore, Sexréov for decxréov, J. H. Steeman. Tue Royat Hottoway 7 ‘ Ak A a he | fy = a “A 4 +a ae ‘ ap ‘as THE HYMN TO HERMES. In my last paper, C.Q., January, 1927, I stated briefly that v. 349 should stand in immediate sequence to v. 351. Few, I think, will refuse assent to this. But more than this may be learnt from the line in its new position: wédwp’ ws Tis Spvoi Baivor. We have already had in 225 this rota réAwpa where the better reading 7éAwp has been shown to be more in accordance with both usage and the sense of the passage. Here there is a further objection to the adjective which cannot be urged against it in 225. The word dpaws shows in its Homeric use indisputable traces of an initial digamma: E 425 ends with yeipa apaujv, 2 411 and Y 37 show similarly pwovro dpacoi. Nothing like this e¢ tis dpacyot can be found in Homer. The change here required is obvious: Toia, wéAwp ws et Te Spvai Baivor. ‘Such tracks (ixvia), as if some monster walked on slender saplings.’ The greatest possible emphasis is laid on zéAwp by its position, as is well known. It is some ultra- human creature, neither fish, flesh, fowl, nor anything normal. 357 Kat 7d peév Eva, 7d EvOa. The curious verb (restored by Ilgen from éarip MDL and zaddynoe M) is probably right ; but the whole line can hardly be regarded as genuine. The inter- polation was doubtlessly intended to summarize all that Hermes is said to have done after penning up the stolen beasts, but utterly fails to convey any distinct meaning. Nothing can be made of ‘and had juggled this side and that of the road.’ Omit this peculiar statement and Apollo’s speech gives all the facts he had any knowledge of, everything to which he could bear witness. Again, the line is unepic. The verb is anything but Homeric, nor is 6800 73 pév éva. 7d 8’ évOa to be paralleled in the old epic, v. note on 226. In the next line peAaivy éorxws recalls A 47 0 & nie vuxri and A 606 6 épeuvy éovxws said of Apollo and Heracles respectively. It is fairly certain, especially if we compare the vuxti oy dréAavros vrwria said of Hector (M 463), that the expression describes in Homer the grim look of wrath and berserk fury, cf. Shakespeare, Henry V. iii. 1. 8: Disguise fair nature with ill-favour’d rage : Then lend the eye a terrible aspect. Euripides, Ovest. 428 iSeiv tpeis vuxti Kopas. It would be a defiance of all reason to apply vuxri éovxds in this sense to the infant Hermes. No one would believe for a moment that the writer of the Hymn intended such an absurdity. Some have taken the desperate course of explaining the phrase as equivalent to ‘invisible’; but ‘like night’ cannot bear this sense; for night and darkness are visible enough. The approach of night might indeed be invisible: this would avail nothing here. Hermes does not fade away gradually. He eclipses himself suddenly and his disappearance is instantaneous. We are inevitably forced to the conclusion that the poet did not write the vuxri éorxws of the tradition. It reasona very pa in a que Dis ye it is \ that in original both lin matter | (w 334) (corr. de aeOAa). here. I this case The invariab that em probabl; First a¢ has give In tory of the he The justificat [veov y’ p On) 7’ oide, must be ATPEKEWS | Eith The iam more tha normal f emphasis At tl Suppose and the : Se f i aft ‘as < ; | ay ote ould stand But more » has been passage. ainst it in an initial VTO apasoi, required > greatest me ultra- M) is ‘he inter- lave done meaning. this ledge of, ov Td pev A 606 certain, 163), that fury, cf. ie infant intended e phrase ght and this eclipses y forced T. L. AGAR 35 It is well known that Zenodotus read vuxri éAvo Geis in A 47, and it is not un- reasonable to suppose that he had epic authority for the expression perhaps from this very passage where it certainly would be well placed. More than this cannot be said in a question so problematic. 364 ovde ke pnvicatp’, od’ Gv pyvutpov dpodunv. Dismissing the unepic avrov (359) which seems to have displaced a primitive rév ye it is worthy of note that in 264 we find ov& dv unvicaw’, and it is difficult to believe that in the second appearance of this notable line the obsolete xe has expelled an original dv. It is worth observing that ové’ av must be maintained before pijvvrpov in both lines, and this seems to countenance the dv objected to in 264. The truth of the matter is dv pavutpov apoiunv should read du pyvutpov dpoipny, i.e. dvapoipny, just as (w 334) opp’ av EAoiunv Sapa is properly aveAoiuny, cf. 736 fo’ dvedAdvres (corr. deOAa to’) dvéAnrar, WY 823 io’ (corr. dveAéoOax deOda). There is a possibility that dveAotunv rather than dvapoiunv should be read here. In € 297 dyorro has a variant dpocro, It seems hardly necessary, however, in this case to alter the traditional verb. 365 To ap’ ws eimav Kar’ ap’ éfero PoiBos The encroachment of dpa here has often been suspected; but the suspicion has invariably fallen, I believe, on the wrong one. The most objectionable of the two is that embedded in the compound verb. Here and elsewhere the true reading was probably and properly eirav adap First adap has been taken for dd’ dp’ or az’ dp’, and then the incorrect preposition has given way to the usual xard. In 367 dei~aro 5¢ Kpoviwva, ‘he saluted the son of Cronos,’ is a far more satisfac- tory rendering than the usual ‘he pointed at.’ Turning towards with slight inclination of the head even without our wave of the hand would justify de‘£aro. 371 onpepov véov The tribrach in the fourth place and y 367 ov te véov ye would afford a fair justification for véov y’ here, even without the concurrence of no small MS. authority [véov y’ p (praeter A. Q.): y’ add. D m.p.] Allen. On the other hand, we have an impermissible overcrowding of particles in 376 ra 8é 7’ ofde, where Barnes has vainly proposed ra 8 ofde, and either this or 7d dé ofda must be correct, for there is no sense here in a generalizing re. In 380 1d de 7’ dtpexéws the 7’ is different, being ro: the personal pronoun. 379 ws ovK oixad’ Pdas, ds «inv. Either ds od of Bots or ws Boas ov of oixad’ éhago’ would be metrical. The iambus in the fourth place is not to be attributed to the early epic composer any more than the x of ovx in the first. In the first alternative reading oi occupies its normal position; in the second the placing of das is legitimate in early epic when emphasis on the word is intended, v. Homenica a 37. 382 Kai ce PiAG Kal Kai avros, Ws ovK péyav 8 OpKov At this stage in the development of the story it is quite out of the question to suppose that the little hero would express the least respect for his accuser, Apollo, and the word ézifoua: is too close in sense to the preceding aidéouar to be acceptable . t 4 uf i 4 i i 36 T. L. AGAR as correct even if we disregard the digamma of ofo@a, as we ought not todo. A verb in absolute opposition and contradiction to ¢:A@ is required. I suggest therefore, Kai Tovrov arifw, ‘and I have no respect for the prosecutor.’ A still stronger and more objectionable word may have been used here, arérrvoa, but it deviates further from the tradition than ari{w, however much support it may derive from an ungraceful national usage in the Orient. I have elsewhere shown conclusively that éridafoua: is an error for érdjopar, ‘1 will find for myself.’ 385 Kai rov TovTy Ticw Tore mov (Hermann) for zor’ though S. and A. say it is not required. Still ‘and methinks’ is very suitable for Hermes to say here, and I find wov is adopted by Wolf, Baumeister, Gemoll, Abel, and Evelyn-White. I note, however, that in his latest edition of this Hymn (Scrip. Class. Bibl. Oxon.) Mr. Allen reads wort before vnAéa, adding in a note ‘cf, p 191’; but in the line to which reference is made ori is a preposition. It cannot be so here, and it is useless to follow even the Moscow MS., unless it affords a tolerable sense. The meaning usually given to gwpyv here ‘ inquisition,’ ‘investigation,’ ‘ search,’ though derived from Hesychius ‘ dwpav 8? rhv épevvay’ is unusual. It seems to me that it would be better and more humorous to make it bear its ordinary sense and refer to the jest of Zeus at the opening of the court. DoiPe, ravrynv pevorxéa Anid’ ; Hermes would naturally lay stress on the ill-treatment he had suffered, and adopting the hint unintentionally given him by Zeus he may well threaten to punish Apollo some day for his ‘cruel kidnapping.’ The next line also supports this view of the matter. Hermes claims sympathy because of his physical helplessness. 388 Kai 73 ordpyavov ex’ wrévy OVS GreBadde, Almost to a certainty rd ordpyavoy is a later correction of an original 6v ordp- yavov. There can, however, be no doubt (v. note on 306 ordpyavov dud’ dows écApeévov and 331) that éx’ wAévy is erroneous. The correction is fortunately easy. Kai Ov ordpyavov elyev érwpaddy ovd' aréBadde. 392 “Eppnv dudxropov It is rather surprising that the two divinities should be ordered to search for the cattle, when one of them had already done so and the other knew perfectly well where they were hidden, as the next words of the speaker plainly state. The true reading may be (nAevew or (nAov éxew ‘to esteem one another,’ ‘to congratulate,’ cf. Aesch. Prom. 330 {nA@ éxrds aitias xvpeis. Eurip. Medea 60. In 393 read x@pov pév for tov x@pov v. Homerica « 55. 394 dréxpuye ipOipa Kdpnva. Instead of expanding the traditional 8 ar’ into 5) adr’, it would be better in every way to read dréxpupe Bowv xdpnva. S. and A. refer to Leaf on A 340 where aére is rightly said to mean ‘ hereafter’: the two editors translate here ‘where have you hidden now?’ This is scarcely possible. With raira all difficulties disappear. The properly In 1 right. | combina pace Mr, already | philolog Its M. att cf. E 27 at any r cannot ‘ where t the trad Finally COWS We would b Cf. 402 Rez The edvvao 2 In ¢ If the < qualifies 409 paragra] Hermes Apollo a reason f friendlin In original (414) is (v. Hym with wh: read het Hermes Apollo and 150 line vp of Typh "9 BS: AS fi ~ Fag > j & aig 47 ng a lo. A verb erefore, bjectionable e tradition ional usage Still and by Wolf, 1 his latest sfore vnXéa, e mort isa MS., search,’ ems to me sense and fered, and to punish nis view of ye Ov ordp- up’ easy. ch for the yell where e reading Aesch. ad x@pov better in sreafter ’: scarcely HYMN TO HERMES 37 400 Ommov 5) Ta xpypar’ arddXero vuKtds év Spy. The second foot is merely a trochee. A short open vowel in thesis cannot properly be lengthened before yp at the beginning of another word. In the best MS., the Moscow, éxov opens the line, and Ilgen’s érmov is probably right. If so, 7x is a gloss. Conversely od might be said to be a gloss on jjxs. The combination 7» xov (Fick) has no literary standing, and it would be more than rash, pace Mr, Allen to give it admission. Let it be Eretrian and remain so. Our line is already in sufficiently evil case without being further disfigured by this abnormal philological freak of one untrustworthy stone-cutter. It seems fairly certain both from the evidence of the MSS. (xpyjyara rurdAAero M. aritadAXAero vulg. avriréAXero T.), and from unquestionable Homeric usage, cf. E 271, 2 280, etc., that drdéAXero should be drirdAAero. Consequently xpypara, at any rate in the plural, cannot be maintained. If the singular be right, and this cannot be regarded as certain, we might either read ‘where the great stock (of cattle) was fed,’ or more explicitly, but with less regard to the tradition, péya xphp’ arirddXero, Finally it is at least remarkable that vuxrds év py should be the feeding-time. The cows were placed in the cave for concealment, and even in the daytime the cave would be dark as night ev Oppv7. Cf. 402 és paos Kdpynva, viz. dppvns. 403 44 * In the following table an asterisk preceding a press mark indicates that the MS. is palimpsest, in whole Or 1n part. E. A. LOWE F.S. means on the flesh side of the leaf. R.=red ; Rt.=right; b.-ll, =bounding-lines. H.S. means on the hair side of the leaf, No. Press mark. Contents. Script. Date. rue rene Running titles. 48 | * St. Gall 1394 (pp. 7-49) VIRGIL Square; IV Smaller square +248 (pp. 196-212)+275 | (Primary Script) | capitals capitals 49 | Vatic. 3867 (St. Denis) ... VERGILIUS Rustic | IV/V — Smaller Rustic RoOMANUS capitals 5° | * Verona XL (38)... _.... VIRGIL Rustic | IV/V — Smaller Rustic (?) (Primary Script) 51 | * Milan, Ambr. G. 82 sup. PLAUTUS Rustic | IV/V Smaller Rustic (Bobbio) | (Primary Script) 52 | Vatic. 3226 ++ «| TERENTIUS | Rustic | IV/V | None | None |Full size _ red BEMBINUS Rustic 53 | * Milan, Ambr. R. 57 sup. CICERO Rustic Vv Smaller Rustic (Bobbio) | (Primary Script) 54 | Bamberg, Class. 35a... Livy Uncial | Vex ~- — Fourth Decade (Frag. ) | 55 | ** London, Add. 17212 LICINIANUS Uncial V -- — Smaller uncial (S. Maria Deipara, Nit- (Frag. ) rian Desert) 56 | Vatic., Urbin. 1154 Uncial | Vex 57 | Paris 9643 (Lyons) ...| Copex THEop. | Uncial| VI! > 58 | Milan, Ambr. C. 105 inf. HEGESIPPUS Half- vii ~ — |Smaller half-un- (Bobbio) uncial cial on F.S. only 59 | Cassel, Theol. F. 65 HEGESIPPUS Half- VI2 — — |Smaller half-un- (Fulda) uncial cial 60 | * Milan, Ambr. E.147sup.| SymmMacuvus Half- | VI ex =f +f | Smaller half-un- + Vatic. 5750 (Bobbio) | (Primary script)} uncial cial 61 | Vercelli, Cathedr. (Cod.| Gospzts(a) | Uncial [V(circa} —- ~ or — | Smaller uncial Vercellensis) (antehieron.) 371) 62 | Turin G. VII. 15 (Bobbio)| Gospets(k) | Uncial|} IV/V — — | Smaller uncial (antehieron. ) 63 | Stuttgart (Const.-Weing. PROPHETS Uncial V None§ | None§ | Smaller uncial Fragm.) ¢ (antehieron.) 64 | * Wurzburg Mp.th.f.64a} Pentatevcu | Uncial V -- — | Smaller uncial and PrRopu. (antehieron.) (Primary Script) 65 | St. Gall 1394\| (pp. 50-89)| Gospers(n) | Uncial| V = — |Small uncial, (antehieron.) chapter number in R. Greek letter 66 | Berlin, K.B. Th. lat. fol. KINGS Uncial Vv : Smaller uncial 4857 (Frag.) + Quedlinburg frag. t Teste O. Seeck. ¢ Complete press-mark : Stuttgart Landesbibl. H.B. II. 20, 54; VII. 1, 8, 12, 25, 28, 29, 30, 39, 45: XI. 30; XIV. 14, 15+5 leaves and a strip in the Hofbibl. + Darmstadt 895, 896+ Fulda Aart, 11, 13, 15 MO. In the colun ordinar\ millime In the ‘ Qui Quire mark Small R. | cial letter F (added lat Rom. nui H.S. Probab! Rom. nu Rom. nt 4 ; | } “ce 4 ; } J | | | Rom, nw Rom, nu | on F.S | Rom, nw i om, nu | Rom. nu | wee | | A | é | ; St. Pa +St MORE FACTS ABOUT OUR OLDEST LATIN MANUSCRIPTS 45 mpsest, in whole | In the column marked ‘ Measurements’ the figures are always given in the order — by width ; those in ordinary type indicate page size, those in bold faced type the written space; all measurements are in e of the leaf, millimetres. In the ‘ Quire marks’ column the examples given are typical, and are taken from the MS. for illustration. Measure- No. of No. of unning titles. Quire marks. cane, queen, | aes. Colophons. Nomina Sacra.| Miscellaneous remarks. No. laller square Circa I 19 =| Twolines of capi- Very wide margins 48 apitals 225 X 270 tals ialler Rustic | 333 X 332 I 18 | Rustic,occasional Red for first line of | 49 | 240 X 250 red books aller Rustic (?) 260 x 205 I 13 Sloping b-d_ uncial | 50 130 X 170 marginalia aller Rustic Small R. un- | 262 x 241 I 19 =| Rustic 51 cialletter F.S.} 175 x 175 | | (added later) ll size red Rom. num. | 160x 185 I 25 Rustic Double bounding-lines | 52 ustic H.S. 113 X 132 Ruled on F.S. aller Rustic 275 * 255 3 24 53 290 x 250 3 35 Wide margins 54 197 X 205 aller uncial | 235 x 172 2 24 55 158 x 158 Rom. num, | 265 x 160 I 25 Uncial, alternate Single bounding-lines. | 56 F,S. 155 X 105 red lines Wide margins Rom. num, | 280x220 I 30 Chapter headings 57 215 X 195 -in Rustic aller half-un- Rom. num. | 298 x 245 2 25 Single bounding-lines | 58 11 on F.S. only on F.S. 210 180 wler half-un. Rom, num, | 280X215 I 36 Rustic b=bus Single bounding-lines, | 59 ¥ on F.S. 238 x 173 Ruling on H.S. ler half-un- Circa I 28 Capitals begin page| 60 v1 295 X 255 even in mid-word 223 x 185 Ruling on F.S. ler uncial ‘Rom, num, | 255 x 160 2 24 | Uncial, alternate | ‘dms’, dne | Citations indented 61 : 170 X 105 red lines dei ller uncial Rom. num. | 190x 160 I 14 Uncial, whole dom Single b.-ll. Ruling} 62 H.S. 135 X 120 age; lines: dei both hair and flesh. lack, red, red, Peculiar abbrevia- black tion of Nomina Sacra ller uncial Probably 280 x 250 3 23 Uncial, alternate dms De luxe MS.: no} 63 Rom. num.§ | 173 x 185 red lines at on abbreviation except =, Nomina Sacra. New books begin with R. lines ller uncial | 275 X 227 2 25 dms Twelve-leaved quires. | 64 165 X 150 Very wide margins 111 uncial, Rom. num. | 300x260 2 24 Uncial, very Single 65 pter number 215 X 195 simple. NoEx- dei Citations marked by .. Greek letter plicit or Incipit - >in left margin | opposite each line ler uncial 2 26 -dns° First three lines of first | 66 — column of each page eo in R. 24+St. Paul in Carinthia XXV. a/11 and a/16 (single leaves). § Teste P. Lehmann. 29, 30, 39, 45: || +St. Gall 172 (pp. 257 sq.) +Stadtbibl. Vad. 70 (f. 278) + Chur 2 leaves, 11, 13, 15, Teste Schultze. if : ihe, 46 E. A. LOWE No. Press mark. Contents. Script. Date. we are Running titles. 67 | Cambridge, University; Gospets and b-d Vv — — | Smaller uncial NN. II. 41 (Cod. Bezae) Acts (d) Uncial (antehieron. ) 68 | * St. Gallg12 JEREMIAH Uncial None (Primary Script) 69 | * Munich 6225 (ff. 76-115) | PENTATEUCH Uncial | Vex. ? (Freising) (antehieron.) (Frag.) (Primary Script) 70 | Vatic. 7223 (Part I., ff. 1-| Matruew(h) | Uncial | V ex. ane — Smaller uncial 66) (antehieron.) 71 | St. Gall GOSPELS Half- | V ex. — Smaller half-un- uncial cial 72 | Paris 10439 (Chartres) ... | Joun (ch. i-vi= | Uncial | V ex. _— > antehieron.) Often 73 | * Autun 21 GosPELs Uncial | V/VI _ — | Smaller uncial + Paris N.A. 1628 (ff. | (Primary Script) 5-14) 74 | Sarezzano,§ now at GosPELs Uncial | V/VI - — | Smaller uncial Florence (antehieron. ) (circa (Frag.) 500) 75 | Lyons 425 (351) PSALTER Uncial | V/VI = w + Paris, N.A. 1585 ‘ 76 | Paris 11947 (St. Germain)| Psarreriom j Uncial | V/VI _ ARGENTEUM 77 | Werona VI (6) GosPELs Uncial | V/VI — | Smaller uncial (Cod. Purpureus)| (antehieron.) 78 | Verona 1V (4) (ff. 3-5)... WispDoM Half- | V/VI — Smaller half-un- uncial cial 79 |* Vienna 16 (Part II.= Acts Half- | V/VI — | None ff. 42*, 43-56, 71-75)now; (antehieron.) | uncial at Naples (Bobbio) —— PETER (Primary Script) 80 | Fulda, Bonif. 1 ? GosPEL-Har- | Uncial | Ante ian None | Small Rustic (Cod. Fuldensis) | mony anp N.T. 546 “° 81 | London, Harley 1775 ...| Gospets(z) | Uncial VI | Small Rustic, not on all pages 82 | Orleans 19 (16) (ff. 11-18) | Ecctestasticus| Uncial | VI? s SonG OF SONGS MOR t +St. Gall 414+ Stadtbibl. 292+ St. Paul in Carinthia XXV. d. 65+ Zirich C 79b+ Ziir. Staatsarchiv. Quire marks. Gk. num. wit stroke above on H.S. or vil F.S, Rom, num -XVI- F.S. Rom. num, Rom. num Rom. num F.S. F.S. No traces Rom. num Rom, num Rom. nur F.S. a | 2 = F S ‘ * | | | 3 3 | | < ‘ Running titles. aller uncial ~ wller uncial half-un- ller uncial ller uncial ler uncial er half-un- Rustic Rustic, not pages ———— faatsarchiv. MORE FACTS ABOUT OUR OLDEST LATIN MANUSCRIPTS 47 Quire marks. aa a = Colophons. Nomina Sacra. Miscellaneous remarks. No. Gk. num. with | 260 x 220 I 33 Uncial; Matthew; qdmsand | Single b.-ll. Citations | 67 stroke above | 185 145 and John alter- — — not specially indi- ‘| on HS. nate lines red;| ns, dei cated. First three Luke alternate lines of each book red words red Circa I 15 dmi Capitals begin each} 68 200 X 200 page 141 X 102 2 25-7 | Big + red denn Columns begin with | 69 234 X 195 and black on capitals. R. some deo lines in Exod., Levit., stm and Deut. -W- or | 240x200 2 23 | Uncial ; alternate _ Single bounding-lines. | 70 vu F.S, 185 X 145 red lines Purplish ink for first line of new chapter preceded by number in Rustic Rom, num. | 230X175 2 25 Uncial dni Double bounding-lines | 7; 155 X 140 -XVI- 72 X 56 I II oo Single bounding-lines | 72 F.S. 47 X 34 dno 7 Rom. num.f | 295X150 2 24 Rustic; red and — Citations indented. t black f{ dns First line of chapters 73 R.t Rom. num, | 300x250 > 16 Apparently none; — Writing between two 180 X 175 but chapter dms ruled lines si headings marked| in margin by Greek uncial letter with stroke above it Rom. num, | 290X 25 I 13 — Writing between two Ss. 205 X dms and ruled lines. Double} 79 dns, deus | bounding-lines. Nomina Sacra in red 275X215 I 18 Double bounding-lines. FS pr 180 dmsand | Gold for Nomina| 7° dns, deus Sacra No traces 270 X 230 > 18 Uncial dms. dns, | Writing between two 165 X 150 Siete ruled lines. Single | 77 dei bounding lines 293 X 173 I 30 dns Written space very g 230 X 120 oblong 7 Rom. num. 229X175 I 25 Uncial dns d, Ss, T occasionally 170 X 125 (or 24) uncial 79 Rom, num. | 290x 140 I 35 | Rustic dine Single bounding-lines. | 80 190 X 80 Citations marked by very =< inleft margin oppo- oblong site each line. Syllabi- fication corrected to Greek, Anglo-Saxon marginalia Rom. num. Circa I 25 Square capitals; dns Citations marked by> | 81 180 xX 120 alternate red in left margin appo- 130 X 80 lines st site each line, Initial line or lines of new books R. (Recon- 2 30 Rustic Ruling on hair side. | 82 structed) Rubricated headings circa in Song of Songs 186 X 130 t Teste A. Royet. § Teste Amelli. red 3 zs x } 48 E. A. LOWE | MOR! No. Press mark. Contents. Script. Date. nara Cues Running titles, Quire marks. 83 | Vatic. 7223 (Part I].=ff.; Marx, Luxe, | Uncial| VI? Smaller uncial OF XXX 67-end) JouN F.S. 84 Verona I. oes GREEK-LATIN Uncial PSALTER q F.S. 85 | Milan, Ambros. C. 39 inf.| Gospzts(M) | Uncial| VI? — — | Smaller uncial —_ FS. 86 | Verona IV. (4) (frag. [= LEVIT. XVI. Half- | VI ex. - \ XIII ff. 1-2) DEUT. Xxx. uncial FS. 87 | Munich (Part I. PAULINE Uncial | VI ex, — _ Exists, on F,S§, + Gittweig (Freising) once | ably eg. 88 Cambridge, C.C.C. 286§ | Gospzts (X) | Uncial/ VI ex. _ None | Smaller uncial Rem. num. (Canterbury) preced. by Q on F.S, 89 | Orleans 19 (16) (ff. 19-25)!|! PROPHETS Uncial | VI/VII| + _ { 90 | Wiirzburg Mp. Theol.Q.| Sr. Kizian’s | Uncial | VI/VII} — Smaller uncial g vil Ia GOSPELS q under Rt. col 91 | * Milan, Ambros. G. 82 KINGS Half- | VI/VII sup. (Bobbio) (Secondary uncial Script) 92 | London, Add. 40165 _.... CyPRIAN Uncial| IV — — | Smaller uncial (Frag.) 93 | Brescia, Quirin. H. vi. 119 CyYPRIAN Uncial Smaller uncial (Frag.) 94 | Turin F. iv. 27 CYPRIAN Uncial Vv _ — | Smaller uncial + Milan D 519 inf. (fly- leaves) + Vatic. 10959) (Bobbio) 95 | Turin G. v. 37 (Bobbio) CyYPRIAN Uncial _ a Smaller uncial Rom. num. F.S. 96 | Orleans 192 (169) (f. 1) CyYPRIAN Uncial — Smaller uncial (Fleury) 97 | Bologna, Univ. 7or (509); Lactantivus {| Uncial; V? — — | Smaller uncial Rom, num. 98 | Orleans 192 (169) (ff.40-41)| Lactantivs | Uncial Smaller uncial (Fleury) 99 | * Paris 13246 (ff. 296-300") AMBROSE Half- Vv — — (Primary Script.)| uncial t Teste A. De Bruyne. ¢ Teste Pal. Soc., Series I., plate 33. § Teste E. C. Hoskyns. || +The Hague, Museum Meermanno-Westreenianum 108 1, ff. 26-30, 31-2. i see f ite j We j J Running titles, MORE FACTS ABOUT OUR OLDEST LATIN MANUSCRIPTS —§ 49 Measure- No. of Quire marks. Colophons. Nomina Sacra,| Miscellaneous remarks. | No. aller uncial “11° Or XXX | 240X200 2 23 Uncial Single bounding-lines. | 83 g 185 145 Red: first line of chap- F.S. terand first three lines of new books 270 X 200 I 26 Uncial, Double bounding-lines.| 84 q -W 210 X 140 ably added later Greek written with + Latin letters F.S. aller uncial = 275X170 I 25 Curious capitals, dns Single bounding-lines. | 85 XXX 185 X 111 alternate red = lines F.S. 356 290 I 32 86 F.S. 260 x 245 sts, on F.S.9) pom num. | 260x 175 I 32 dni Citationsindented;and | 87 aly above 210 X 150 when one begins in Greek num. mid-line < is put be- 111 fore first word (cf. 6.8 T Lyons 483). First three lines of epistles R ier uncial Rom, num. | 255x195 2 25 | Rustic, alternate Double b.-1I. 88 preced. by Q | 184x130 red lines Ruling F.S. on F.S, 355 X 245 2 38 Double bounding-lines | 89 285 X 220 ler uncial g vil 206 X 172 2 26 | Mixed capitals; Red used at beginning | go q Xxil1 150 X 125 some of chapters under Rt. col. 263 x 243 I 19 =| Uncial dns Citations indented. | gr 165 X 170 Running foot rubric ller uncial 305 X 225 4 33 dms Single b.-ll. ; wide mar- 92 (Recon- gins. Citations in red, structed) indented 220 X 200 ller uncial 280 X 210 2 27 dmsand | Single b.-ll. Titles| 93 written — indented. Chapters space al- dom numbered at left in most sq. red Greek uncial ller uncial 275 X 230 2 23 | Uncial, very sim- dmi Single b.-ll.; ruled line | 94 165 X 165 ple cuts through letters. Very wide margins. Bright red at begin- ning of books Her uncial Rom. num. | 200x165 2 20 | Uncial, alternate dms Double bounding-lines.| 95 oS. 130X 112 red lines deo Citations indented ller uncial 2 26 ann African 96 190 X 155 ller uncial Rom, num. | 290x 255 2 33 | Uncial, whole dno Single bounding-lines. | 97 .S. 230 X 200 page, alternate dei Citations indented red lines lier uncial 270x200| 2 ge 98 223 X 185 I circa dns Citations indented 99° 225 x 85TT 30-32TT . C. Hoskyns. 7 Teste G. Mercati. tt Teste A. Wilmart. an ae: a § ‘ E. A. LOWE No. Press mark. Contents. Script. Date. ee aeay Running titles, 100} Paris 2630 (St. Denis) ... HIvLary Uncial| V _— — | Smaller uncial tor} Paris 152 (ff 9-16) (Lyons) HILary Uncial y3 _ None | Smaller uncial 102} Verona XIV (12) HILary Uncial | V ex. Smaller uncial 103 | Paris 8907 (Chartres) ... Hivary, Uncial | V/VI Smaller uncial AMBROSE, ETC. 104 | Lyons 452 (381) HILARY Uncial | V/VI — — | Smaller uncial + Paris, N.A. 1593 105| Rome, Basilica D. 182 HILARY Half- | VI in. —_— — (Cagliari) uncial | (ante 509/10) 106} Vienna 2160 * HIrLary Half- VI — | Smaller half-un- + St. Florian uncial cial + Vatic. Barberini, s.n. 107} *St. Gall 722... HILary Half- VI + Apparently none uncial 108 | Paris, N.A. 1592 (Tours) HILARY Uncial | VI ex. > — | Small Rustic 109; Oxford, Bodleian EvsEBIUS- Uncial | post -- — | Column headings, Auct. T. II. 26 JEROME 442 small uncial 110| Paris 6400B ¢ (Fleury) ... JEROME Uncial — | None 111| Paris 2235 JEROME Uncial| VI — 112} Bamberg B. IV. 21... JEROME, Half- VI 7 None | Small uncial AUGUSTINE uncial 113 | Orleans 192 (169) (ff. 29-30) PsEvupDo- Uncial | VI? = None (Fleury) ? JEROME 114| Orleans 192 (169) (ff. 21-7) JEROME Uncial | VI/VII = _ (Fleury) ? 115| Verona XXII (20) _ JEROME Half- | VI ex. + — |Smaller half-un- uncial rare cial 116| Orleans 192 (169) (ff. 32-33)| AUGUSTINE /Quarter-| V _— — None + Paris 13368 (f. 256) uncial +N.A. 2199 (f. 1.) (Fleury) Orleans 192 (169) (f. 38)...| AuGustinz | Uncial/| VI in. t (Ff, 1-8, 285-90) + Leyden Voss. Lat. 4 4 None Rom. num, Rom, num. F.S. A, B, C, or XI etc, 9 XXXIIII g liu ull or XVI Uncial then minusc, letters ‘A etc. F.S. Rom. num. Letter middle lower margin (ff. 167-72) or xxr F.S. ; He ps ee a Quire marks. F. S i F, Ss. ef \3 F.S | | | : | F.S . . Ss. d i é H.S. A | | use MORE FACTS ABOUT OUR OLDEST LATIN MANUSCRIPTS 51 inning titles. Quire marks, | Measere- | No. of | To. of Colophons. Nomina Sacra. | Miscellaneous remarks. | No. ler uncial Rom. num. | Big 4° 2 28 pers alternate dns Single b.-ll. — 100 oe x red lines comes in centre for hori- om zontal ruling. Cita- tions indented. First line of new book red ler uncial 272 X 230 2 2 Uncial dns Ruling on hair side, | ror indented ler uncial 2 21 Uncial, alternate dns Single bounding-lines, | 102 | S um. red lines Citations indented. ee De luxe MS. ler uncial 0X 260 2 2 | Uncial, alternate dnm Single b.-ll. ; citations | 103 xX 210 red lines ; alter- | indented. Three red nate R. letters | (4™S Once) | lines begin books in Feliciter ler uncial A. B,C,or | 290x250 2 2 Uncial. Both Finit dns Double bounding-lines.| 104 q XXXIIII 270 X 205 I 30 Uncial and half- 105 210X 155 uncial lier half-un- g Int 287 x 204 1 27-34 | Half-uncial Citations marked by > | 106 , 210 X 130 average in left margin oppo- 30 site each line rently none —— 255 x 162 1 27 Half-uncial. Finit Single bounding-lines | 107 F.S 195 X 120 and Explicit ll Rustic Uncial then | 280x250 I 22 | Rustic dni Single ne 108 minusc, 180 X 175 First line of new letters book red imn headings, =. Several — MS. required special | 109 all uncial were, ce)| Tuling which is on ete. H.S.; groove faces F.S. groove. tm (=tamen) t (=tur) in contem- porary marginalia Rom. num, | 205x170 | Several] 26 110 F.S. 150 X 145 Letter middle | Square 4°| 1! 36 "— are capitals dni Citations marked by> | 111 lower margin | 208 x 170 (hand A), Rustic or — in left margin F.S (hand B), some- opposite each line. my times with alter- ere and there half- nate red lines uncial letters Il uncial 290 X 210 z 34-5 | Capitals and un- daa Single bounding-lines | 112 233 X 160 cial I 27 Single bounding-lines | 113 213 X 130 ? 290 X 200 I 35 Ruling on hair side 114 260 X 190+ HS. lier half-un- = |asexr7| 1 25 Single bounding-lines | 115 or xxP | 150 ked b 6 itations mar 196 4 in left margin o ppo- i8oX 100 site each line, Inter- esting marginalia 230X 160+ I 25 117 195 X 160+ en Voss. Lat. 4’ (ff. 167-72) + Vatic., Regin. 1709. } at 52 E. A. LOWE No. Press mark. Contents. Script. Date. an — Running titles. Rome, Vitt. Em. 2094) AvGusSTINE Uncial| VI — Smaller uncial (Sessor. 13) (Nonantola) 119} Lyons 478 (408) ... .. | AUGUSTINE Uncial| VI > 120| Paris 13367 (Corbie) ...| AUGUSTINE Half- VI — | None uncial 121| *Rome, Vitt. Em. 2099; AvuGustTINE /|Half(})-| (Sessor.55)(Nonantola)| (Secondary uncial Script) 122 | Turin, G. V. 26 (Bobbio) | AvuGusTINE Half- | VI3 uncial 123| Milan, Ambr. G. 58 sup. | AvuGUSTINE Uncial |} VI? = (ff. 74-5) (Bobbio) 124 | Vatic., Palat. 210 AUGUSTINE Uncial | VI/VII;| = 125 | Paris, 12214 AUGUSTINE Half- | VI/VII| Unc.g| — | Rustic on FS. + Leningrad Q. I. 4 uncial hand only (Corbie) g-unc. None 126| Paris 9533 oon ...| AUGUSTINE Half- | VI/VII ae — | Cursive add., but uncial probably temporary 127| Wiirzburg Mp. Th. Q. 3 Priscirrian | Uncial| V/VI | (Hand| A) (Hand| 3) 128| Vienna 847 RoFINvs Uncial | V/VI — | Occasional title (see colophons) 129| Paris 8084 (ff. 1-155) ...| Propentivus Rustic | VI Smaller Rustic 527 130| Ravennas.n.§... nae AMBROSE Half- | VI in. = — |Small half-uncial} uncial on F.S, only —— ¢ Possibly a later edition, but was doubtless originally there, too. MOR Quire marks. xxvii F.S. In lower left corner page of eac quire F.S. Rom. num, +abbrev. for coxtuli F.S g+ Rom. num., then +unc. letters g ul or /q xu F.S. Rom, num, Rom. num, F.S. None visible Rom. num.f_ F.S. q il F.S, (first part) q° XIII" (second part) letters 7 ty | | ts: i 4 | | F, S. | : 3 | 3 i gt 23 ; | j AS ‘ “ | | | : 4 nning titles. — ler uncial ic on FS. ive add., but bably con- porary sional _ title colophons) ler Rustic half-uncial only MORE FACTS ABOUT OUR OLDEST LATIN MANUSCRIPTS 53 Quire marks. Colophons, Nomina Sacra.| Miscellaneous remarks. | No. g XXVIII F.S. | 330X225 I 32 Rustic, alternate | Nom, Sac. | Single bounding-lines. | 118 255 X 175 lines regular. No; Citations marked by otherabbre-| £ or >> in left mar- viation gin opposite eachline. First three lines of books and first line of chapters red In lower left | 260x215 I 28 | Uncial ; Finit for dae First two lines of new | 119 corner of first | 198 x 182 explicit book in red page of each quire F.S. Rom. num. | 245x195 I 25 | Uncial and half- dni Single bounding-lines, | 120 +abbrev. | 167x147 uncial ; Explicit Citations usually in- for contuli F.S. black, Incipit dented, and always red marked by € in left margin opposite each line. First line of book in R. g+Rom. 280 x 200 I 43 | Uncial, alternate ni First line of books red | 121 num., then | 225x170 lines red +unc. letters F.S. q ul or 215X143 I 21-22 | Uncial o i Single bounding-lines | 122 F.S. | 175X110 (=nostro, -t), cf. Lyons 483 I 29 Uncial, alternate 123 200 X 160 lines red Rom. num, | 240x 180 I 33 | Some in Rustic Citations from earlier | 124 FS. 195 X 135 work of St. Augustine in half-uncial Rom. num. | 260x220 2 29 Rustic and square dni Single bounding-lines. | 125 F.S. 197 X 165 capitals; Expii- First line of new book cit and Finit in red uncial. Prob- ably copied from un- cial exemplar. Con- temporary slanting uncial and half-un- cial marginalia, with ancient abbreviations None visible | 260x175 I 26 Half-uncial, al- dns Single bounding-lines | 126 190 X 130 ternate redlines| —pys on F.S. Citations marked by £¢€ in left margin opposite each line. Probably of Visigothic origin Rom. num.f | 215x157 I 2x Rusticoruncial;! gnsand_ Citations indented. | 127 S. 175 X 105 Finit for explicit = Red for first three lines of books : 192X172 I 14-15 | None: rubrics and Writing between two/| 128 Il 125 X 125 Incipit in top ruled lines. First two T margin uncial lines of MS. and first FS line of new sections | bi in red q il 302 x 280} I 20 Double bounding-lines | 129 F.S. 205 X 185 (first part) (present 2 28 Bold half-uncial dns Ruling on H.S. Cita- | 130 q° XIII" size) tions in smaller half- (second part) | 300x 225 uncial and indented letters 205 X 175 ‘ t Teste Pal. Soc., Series I., plate 29. § Teste G. Mercati, j | | | | | | | | | | | | | | if ape, x 54 E. A. LOWE No. Press mark. Contents. Script. | . Date. Cuts Cutaes Running titles. 131 | Verona XXXVIII (36) ... SULPICIUS Half- | A.D. 517 +. Smaller _half-un- SEVERUS uncial also cial 132| *Autun 24 CASSIAN Half- — | None + Paris, N.A. 1629 (ff.; (Secondary uncial 17-20) Script) 133| Lyons 483 (413) ... ORIGEN Half- VI — —_ uncial 134| Paris 12097 (Corbie) _... CANONS Half- VI + None CATALOGUE OF | uncial | (537- | also Pores 555) — 135| * Verona L V (53) ... | DpascaLia Half- Vi APOST. uncial (Primary Script) 136} Florence, Laur. 65.1... Orosivs Uncial}| — | Smaller uncial 137 | Paris (ff. 1-23) EvcHERiIvs | Uncial} VI = None + (ff. 53-65) CoSMOGRAPHIA 138| Vatic. 5704 ..| EPIPHANIUS | Uncial| VI2 — | None 139| Paris :0593 ‘ BAsILIvs, Uncial | None Rusticon first and RvuFINUsS last pages of quire only 140| Monte Cassino C L (346)| AmsBrosiaSTER | Half- | ante - None uncial | A.D, 57° 141| Vatic.3375f =... Evcippius Half- | VIlex.| None uncial 142| Milan, Ambros. H. 78sup. AMBROSE Half- | Vi ex — None + Turin G. V. uncial (Bobbio) 143| Milan, Ambros. O. 210 VIGILIUS Half- | VI/VII; — — | None sup. (ff. 1-46) (Bobbio) | TapsENsis, etc. | uncial > t Teste E, Carusi. MORE Quire marks. Right side, but g almost in centre F.S. Rom. num, F.S. q g XXXVIIII g Rom. num. None visible q XXXUIIII Right corner, but spread out /vi F\S. Rom. num, 4 | 3. | } i | | . Ss. F.S } i F. S. | = | | | Rom. num F.S. — re nning titles. er half-un- uncial on first and pages of only MORE FACTS ABOUT OUR OLDEST LATIN MANUSCRIPTS 55 Colophons. Nomina Sacra.} Miscellaneous remarks. | No. = 255 X 220 I 20 | Half-uncial, some Ruling on F.S. 131 g x1 | 200X 190 red; Finit and = Explicit Right side, but q almost in centre F.S. Rom. num. | 241XI5I1 I 21 Half-uncial 132 F.S. 175 X 100 290 X 225 I 21 Half-uncial; Finit} _o Single bounding-lines, | 133 g XXXVIIII 200 X 160 and Explicit dms n (cf. at beginning and ; Turin, end of citation ; 33 if G. V. 26) a second citation is =iesus occurs on same page (cf. Munich 6436). Contemporary slant- ing mixed uncial and half-uncial margi- nalia g i111 280 x 210 I 33 None dno Single bounding lines. | 134 -"y 235 X 180 interesting | ‘Islands’ of uncial F.S. abbrev. in the text 255X155 I 35 135 205 X 125 Rom. num. | 243X215 I 25 |Square capitals, scts Capitals often begin a | 136 F.S. 175 X 160 alternate red page, even in mid- lines sentence. Single bounding-lines, First line of new book in red. Marginal rub- rics and sloping b-d uncial marginalia None visible | 245 x 180} I 31 Red Rustic ; occa- dni Single bounding-lines | 137 190 X 145 sional red Rustic Ruling F.S. headings g XXXUIIII 280 X 220 I 23 Uncial mixed with dno Single bounding-lines | 138 Right corner, | 250x 160 Rustic, alternate but spread out red lines F.S. jvi 260 x 180 I 28 Square caps, dni Single bounding-lines, | 139 F.S 202 X 132 sometimes alter-| (=mnon) First lineof books red. nate lines red,; Add. in Rustic rubrics where sometimes all uncial running title should be. Interesting an- cient abbreviation in marginalia Rom. num. | 299x215 I 30 ‘Argumentum'’ dns Citations indented and | 140 240 X 175 in Rustic marked by~ in left margin opposite each line. Has slanting uncial marginalia Rom. num, | 343x250 I 38-40 | None Some contemporary | 141 F., 240 X 190 slanting uncial mar- ginalia g XXX! 242 X 203 I 23 | Rustic; Finit for dns Single bounding-lines. | 142 F.S. 173 X 153 explicit Citations marked by -~in left margin oppo- site each line. Many marginalia, with an- cient abbreviations QO hu 247 X 197 I 31-3. | Ornamental Rus- dns 143 205 X 165 tic (not oldest type) or half- uncial t The page-size of Paris 4808 is 260 x 190. i if | i f 4 iW 56 E. A. LOWE MORE No. Press mark. Contents. Script. | Date. | Omitted | Omitted | Running titles. | Quire marks. 144| Verona XXXVII (35) ... CLEMENT Half- VI2 — None | A line ruled andj Rom. num., uncial traces visible then letters smal] F.S. f-uncial 145| Verona XXXIX (37) ...| Cassioporus | Uncial| VI/VII} + + Rt. F.S. 146) Milan, Ambros. D.36sup.;| Propgntivs | Uncial | VI/VII = — | Cut away, if any _— (Bobbio) = va F.S. 147| Vercelli, Cathedr. 158 ... CLEMENT Uncial | VI/VII; + — | Smaller uncial = F.S. 148| Turin, Archivio di Stato | Lacranrivs,etc.} Uncial | VI/VII; — g J. b. II. 27 (Bobbio) F.S. 149/| Vatic. Regin. (ff.| Forcenrivs | Uncial | VI/VII — — Rom, num. 1’-98") (Fleury) F.S. 150 | Vatic. 267 (ff.| Half- | VI/VII = None Rom. num, g9-end) (Fleury) uncial — F.S. I other e of the : THe Larce LETTER AT THE BEGINNING OF EacH PaGE. fact th: Before I remark upon the data contained in the above tables, I wish to make a came | few additions that have since come to my notice of other MSS. which exhibit the capital feature of a large letter at the beginning of each page, and to point out the practical therefo value it may have to an editor of an ancient text. Through the kindness of Professor from ™ F. C. Burkitt I am in a position to add St. Gall 912,1 the Old-Latin version of this cu Jeremiah (No. 68),and Munich 6225, an Old-Latin version of the Pentateuch (No. 69). edly in With these two should be put Milan, Ambr. E. 147 sup., Symmachus, in half-uncial norma! (No. 60). In these three MSS. the use of the capital letter at the beginning of the page is still a living practice. In the following MSS. its presence is due either to inadvertence or to faithful copying from a much older exemplar: I refer to the Florence Orosius (No. 136) and the Turin Lactantius (No. 148), which show this feature here T and there; to the Regina MS. of Fulgentius (No. 149), which has it frequently in Latin the uncial section but not in the half-uncial; and to three other MSS. not included Christ: in the list, since they are more recent than the year 600—namely, the famous fragme Escorial MS. of St. Augustine’s de baptismo (uncial, saec. vii), which has it on many by an pages ; the Paris MS. 12205 of the Regulae monasticae, etc., also in uncial of the seventh scribe century, where it occurs only in certain quires (foll. 77-92” and 143-157); and the usage Boulogne MS. 47 (42) of Jerome on Matthew, written in eighth-century pre-Caroline : fhe minuscule of a type used in Northern France. I do not think there can be any 2 Bi, 1 References to literature and to published Abhandlungen I, 161-261,and in Miscellanea Ehrie 3 Ff. facsimiles of most of the MSS. mentioned in this XV. 34-6r. ‘ E. article will be found in Traube’s Vorlesungen und Ff iy it away, if any taller uncial MORE FACTS ABOUT OUR OLDEST LATIN MANUSCRIPTS Running titles. line ruled and races visible: robably smal] jalf-uncial Quire marks. 5 Colophons. Nomina Sacra.| Miscellaneous remarks. | No. Rom, num., | 324 x 246 2 25 | Half-uncial Single bounding-lines | 144 then letters | 200x180 F.S. = 280 X 225 I 25 | Big uncial, alter- Single bounding-lines | 145 qv 215 X 175 nate red lines, toward Rt. pyramidal F.S. on 225X145 I 28 | Mixed capitals Double bounding-lines | 146 = VIII 185 X 110 F.S. - 250 X 215 I 23 Uncial, alternate dns First three lines of | 147 = 200 X 170 red lines books in red. Visi- xi gothic marginalia q UllI 232 X 210 I 24 | Uncial, red Here and there capi- | 148 F.S. 205 X 185 tals begin a page even in mid-sentence. Single b.-ll. First line or first few words of books in red Rom. num, | 270x190 I 23 Capitals (hardly Many pages begin with F.S. 200 X 135 Rustic, perhaps capitals. Single|149 Square), alter- bounding-lines. << nate red lines to left of citations Rom. num, | 270X190 I 33-4 | Rustic, sometimes Single bounding-lines | 150 civca alternate red 205 X 140 lines make a bit the ‘actical ofessor sion of 10. 69). -uncial of the her to orence e here ntly in cluded amous many sventh id the roline e any 1 Ehrle other explanation of the presence of the large letter in the Boulogne St. Jerome. One of the scribes faithfully reproduced his original page by page,’ as is shown by the fact that some pages are overcrowded,’ while on others he plainly spread out as he came to the end.* A corrector carefully cancelled‘ or erased’ the meaningless capital at the beginning of a page and put a smaller one in its place. This MS., therefore, is as good—if it be a correct copy—as the fifth- or sixth-century exemplar from which it was copied.® It is not likely that there existed a middle member with this curious feature. Any MSS., therefore, in which capital letters rise up unexpect- edly in the text doubtless go back to a very old archetype in which this feature was normal, and their value as witnesses must be assessed accordingly. II. OMISSION OF # AND m AT END oF LINE. The practice of omitting m at the end of a line to save space is not indigenous in Latin MSS., but is an importation of a Greek custom which goes back to pre- Christian times.? There is no sign of the practice in the oldest Latin papyri fragments that have come down tous. The earliest dated Latin example is furnished by an inscription of the time of Pope Damasus (366-384) eagraved by Philocalus, scribe and stone-cutter. What we find in this inscription doubtless reflects the usage of his scriptorium. Traube suggested that the custom was introduced into 1 He wrote ff. 1-67. ¢ C. H. Turner used this MS. for the purpose 2 E.g. f. 21°. of comparing it with the Worcester leaves in 3 Ff. 26, 30, 51 et passim. eighth-century uncial. Cf. Early Worcester * E.g. ff. 26", 30”, 51”, etc. MSS., p. xv, Oxford, 1916. Ff. 5%, 6", 12, 14”, etc. 7 L. Traube, Nomina Sacra, p. 241. f : ‘ ae ‘ , ive 58 E. A. LOWE Latin centres in the course of translating Scripture from Greek into Latin, which seems a reasonable hypothesis.? When a Greek scribe came to the end of a line and had to write the letter N to finish the word or syllable, he made use of the device of indicating that letter by a horizontal stroke instead, in order not to go beyond the written space. In Latin, however, the analogous letter is not N but M. This explains why the Latin practice is apparently at first confined to the m, the omission of » being a later extension.? The data furnished in our present list bear out in a general way the observations made on the first forty-seven MSS. An examination of these additional hundred odd MSS. makes it perfectly clear that the oldest way of marking the omission of m or is by the simple stroke. Yet it cannot be denied that the use of a stroke with a point beneath it is also of very great antiquity, since it is found in such MSS. as the Codex Vercellensis (No. 61) and the Brescia Cyprian (No. 93). It thus unfortunately becomes useless as a dating criterion. Another fact which emerges with some definiteness is that the habit of making the distinction between m and » grows more frequent the nearer we approach the year 600. The earliest instances of the distinc- tion are furnished by the Vatican Virgil in square capitals and the Palatine Livy (Nos. 1 and ro of our first list). The practice may be expressed statistically thus: Out of seventy-two MSS. from the fourth to the end of the fifth century only nine make the distinction, whereas out of forty-four MSS. of the sixth century twenty-one make the distinction. In the earlier MSS., one in every seven, roughly speaking, may be said to make the distinction ; in the later, it is one in every two. A few MSS. are in a class by themselves as regards their manner of distinguishing m from m. In the Codex Fuldensis (No. 80) we find two consecutive dots under the stroke for m and a mere stroke for m, In the Lyons Psalter, however, m is indicated by the horizontal stroke over two dots underneath each other, the ” by a stroke over a single dot. Of the peculiar style employed in the Wiirzburg Jerome (No. 40) mention was made in the previous article. Our table shows cases in which different forms of the m-stroke occur in different parts of the MS. Thus in Milan, Ambr. C. 39 inf. (No. 85) three varieties of the stroke occur: the ordinary horizontal line, the stroke over a dot, and the stroke to the left of a dot. In Verona XXXVIII (36) (No. 131) of the year 517 we find + ~- —, and the same in Verona XXXIX. (No. 145). Two modes of indicating omitted m deserve special attention: (1) The use of a line surmounted by a point for m, a usage which was destined to become characteristic of Visigothic MSS. —this is found once in the Pauline Epistles, Munich 6436 (No. 87) which has other Spanish symptoms, in the Verona Sulpicius Severus (No. 131) of the year 517, in the Ravenna Ambrose (No. 130) also in sixth-century half-uncial, and in the Paris MS. 9533 of Augustine (No. 126) which has unmistakable Visigothic features ; (2) the other is the use of a sign like our division sign (+)—this is found in the Verona MSS. XXXVIIL, XXII., and XXXIX. (Nos. 131, 115, 145), and in the Paris MSS. 12097 and 12214 (Nos. 134 and 125), both apparently of French origin, and in St. Gall 722 (No. 107) of Hilary (all of these six MSS. being of the sixth century). Nearly all MSS. mark the omission of m by the single stroke. In a few, however, the stroke above a dot is found, the oldest example being the Codex Vercellensis (No. 61). Many scribes carefully avoid the omission of and prefer to insert it outside the space allotted for text even though they make regular use of the m-stroke. On the whole, it must be admitted that the practice with regard to m and m omission signs has proven less helpful as a dating criterion than had been anticipated. 1 L. Traube, Nomina Sacra, p. 241. in Codices Graeci et Latini photographice depicti, 2 Traube, Hievonymi Chronicorum Codicis Flovia- Suppl. I. censis Fragmenta, p. vii, Leyden, 1902 ; published MORE The borne ou oldest ite running t clear is tl in the tex sixth cen useful in earliest fi uncial, tk an ancier (No. 112 one obse! the leav suggests which, a: In t square, ¢ sixth cer oldest M is nearly width—t square. items ar Codex F In) show a ] to the f rare. Three the Cic Livy Prophet two colt often in As (No. 72 three (1 preferer five, for the abo He only go place tl ; | = ay, / | i n, which er N to ter bya n Latin, practice ion.?2 rvations lred odd f morn point e Codex tunately h some more distinc- e Livy y thus: e make e make may be . are in In the stroke ) three and ar 517 les of ed by MSS. other 17, in 1 the ‘ures ; n the n the rigin, sixth . few, ‘odex er to f the and been picti, MORE FACTS ABOUT OUR OLDEST LATIN MANUSCRIPTS 59 III. RuNNING TITLES. The observations made under this head in the previous article (pp. 206-7) are borne out by the MSS. in the present list. Running titles are found even in the oldest items; and Dziatzko’s dictum remains untenable—that is, the absence of running titles cannot be used as a criterion of antiquity. What remains abundantly clear is that the use of a different type of script for the running title from that used in the text is first noticeable in MSS. of the end of the fifth or the beginning of the sixth century (Nos. 18, 19, 29, 80, 81). As corroborative evidence this fact should be useful in arriving at the date of a MS., since it furnishes a terminus a quo. The earliest five MSS. showing a different type of script in their titles are all written in uncial, the titles being in small Rustic—that is, in a type which already counted as an ancient script. Similarly, the half-uncial Jerome of the sixth century at Bamberg (No. 112) had its running titles in the older script, namely uncial. In a few MSS. one observes the interesting custom of placing the running titles on the flesh side of the leaves only (Nos. 58, 87, 125, 130). An explanation of this custom which suggests itself is that the running titles were entered as soon as the sheets were ruled, which, as far as my present data indicate, was normally done on the flesh side. IV. S1zZE AND DISPOSITION OF PAGE. In the former list of forty-seven MSS. six had a written space which is perfectly square, and in the present there are seven. All of these MSS. are earlier than the sixth century, thus confirming the impression gained from the previous list that the oldest MSS. preferred the square format. Of MSS. having a written surface which is neavly square—that is, with less than an inch difference between height and width—the present list has twenty-four, so that one out of every four MSS. is almost square. In the former list of forty odd MSS. prior to the year 600, eighteen such items are found. Two MSS. have a format which is distinctly oblong—namely, the Codex Fuldensis and the Verona fragment of the Book of Wisdom (Nos. 80 and 78). In the matter of dividing his writing space the earlier scribe may be said to show a preference for two columns. Barring poetry, most of our books that go back to the fifth century are in two columns. MSS. with more than two columns are rare. The Cyprian fragments in the British Museum have actually four columns. Three MSS. in our list have three columns to the page; two of these are classical— the Cicero in Rustic capitals from Bobbio (No. 53) and the Bamberg fragment of Livy (No. 54), and one is Biblical—the Constance-Weingarten fragments of the Prophets (No. 63). Our earliest copies of the Gospels are nearly always written in two columns, This is not the case with the other books of the Bible, which are very often in long lines. As to the number of lines on a page, the smallest found in our table is eleven (No. 72); two MSS. show thirteen (Nos. 50 and 75). The largest number is forty- three (No. 121). The great majority (namely, fifty-nine) of the MSS. show a preference for the twenties. The favourite number, to judge by our list, was twenty- five, found in fourteen MSS. out of the fifty-nine. To base any generalizations on the above data would be hazardous. V. QuiRE SIGNATURES. Here, too, the additional evidence furnished by our one hundred and three MSS. only goes to support the earlier observations. The ancient custom manifestly is to place the quire mark in the extreme right-hand corner of the lower margin of the - and a izon 1 2 60 E. A. LOWE last page. It is not till we get to the seventh and eighth centuries that the tendency to move the quire mark toward the centre appears, though exceptions occur—for instance, Paris 2235 (No. 111) has it in the middle. This is also the case in the still older MS. (not included in our list) of the Codex Carolinus of the Pauline Epistles in Gothic and Latin (Wolfenbittel Weiss. 64). For some unknown reason Lyons 478 (No. 119) has the quire marks on the left-hand corner of the first page, in which respect it resembles the Florentine Digests, to which attention was called in the first list (No. 25). Normally the quire signature consists of a Roman numeral, often under a gamma-shaped bracket. In many manuscripts the numeral is preceded by q (signifying quaternio), variously abbreviated. Although the use of Roman numerals continues to be the normal practice, the use of letters is found from the end of the fifth century onward (examples are furnished by No. 36 of the first list, and by Nos. 104, 108, III, 121, 130, 144 of the present list). A few MSS. show peculiar features. In the Bodleian Jerome (No. 109) the first quires are marked with Roman numerals, the subsequent ones with Greek; in the Pauline Epistles from Freising (No. 87) we find the odd practice of placing the Roman numeral above the Greek; in the extant leaves of the Codex Carolinus (mentioned above) the form of the letter € signing the quire is Gothic, not Latin; in the Codex Bezae (No. 67) Greek numerals alone are used, but they are placed, it should be noted, in the lower right- hand corner of the last page of the quire, which is normal in Latin but not in Greek MSS. _ In one sixth-century MS. of our list (No. 120) the quire mark is followed by the reversed letter c=con(tult). CoLOPHONS. The first fact that strikes one in scanning our list is that in our oldest MSS. the style of the colophon is very simple, with few or no ornamental features. The second fact is that the letters used, though generally of a larger size, are of the same script as the main text. During the first half of the sixth century the new practice comes in of using a more ancient type of script for the colophon than was used in the text. Perhaps the earliest dated example is the Basilican Hilary of the year 509-10, which has in its colophon both uncial and half-uncial lines. The Codex Fuldensis (No. 80), written before 546 in uncial, uses Rustic letters for the colophon ; the Codex Harleianus (No. 81) of the same age, also in uncial, uses square capitals for its colophons. The contemporary Bamberg Jerome (No. 112) has capitals and uncials in the colophons. In our oldest MSS. the words Explicit and Incipit are not used at all. It will be seen from our table that in a number of MSS. the word Finit is used instead of Explicit, a practice which may throw some light on the history of the text. Statistics on this point have been given by Professor Lindsay.? As regards the ornamental side of the colophon, a device which early became popular with scribes was the use of red in alternate lines or in alternate words, as well as in alternate strands of the rope-like pattern often enclosing the colophon. This practice goes back to very ancient times: it is found, for example, in the Codex Bezae (No. 67) and in the Constance-Weingarten fragments (No. 63). VII. NoMINA SACRA, In the column marked ‘Nomina Sacra’ I have noted chiefly the data for dominus, since there was a clear development in the manner of abbreviating this 1 The word contuli occurs in some other MSS. 2 W. M. Lindsay, Palacographia Latina II. at the ends of quires. It is generally written (1923), pp. 5-10. out; occasionally it occurs in Tironian symbols. MORE word, whic dominus We etc. The sacred wo: thus seem ‘Nomina forms of t were give Nomina Si With fifth centt exception Wiirzbur (No. 116) the left o century 2 Graeco-I indentati MS. in n means to the cited (No. 124 in half-t methods (No. 12¢ different other pe before a occurs Pauline citation before i Th deserve may be been de of the! touchst incomp outside direct i skin (s the rul Gospel and A (No. 1 J Mice | tendency cur—for n the still] Epistles yons 478 in which h the first ral, often ceded by numerals nd of the , and by peculiar h Roman Freising Greek; n of the 7) Greek er right- in Greek by [SS. the The he same Practice used in he year Codex lophon ; -apitals als and all. It instead e text. ecame rds, as ophon. Codex ta for this na II. MORE FACTS ABOUT OUR OLDEST LATIN MANUSCRIPTS 61 i rs it an important dating criterion. Where I fail to indicate how it taken for granted that the form used was dns, etc. The unabbreviated form of deus, surmounted by a stroke to call sn ia sacred word, is a feature found only in our oldest Biblical and sass ne = - thus seemed worth tabulating in this column. Here and there a few ot er s eres' a6 ‘Nomina Sacra’ have been included without the least attempt at giving = me forms of that word found in the MS. In my first list of forty-seven M — ee were given under this head, To fill the gap the reader is referred to Trau Nomina Sacra, pp. 178-185. a CITATIONS. With few exceptions the method of marking a citation in MSS. of the fourth = fifth centuries is by indentation, which varies in width from one to three pom a exceptions are the Fragmenta Sangallensia (No. 65) Codex n of the Gospe - e Wiirzburg Jerome (No. 40), and the St. Augustine in quarter-uncial from 0m (No. 116), in which citations are indicated by a small mark placed in - — - the left of each line. This is the method usually adopted by scribes of the a century and later. It is interesting to note that in the Codex Claromontanus, the Graeco-Latin MS. of the Pauline Epistles (No. 33), the rule is to mark pe 7 indentation, except in the Epistle to the Hebrews, which ae eo = of sa i ; i oyed vari wi re than one respect (cf. former article, p. 204). Scribes emp cal attention to a citation. In the London fourth-century i es are in red; in the uncial MS. of St. Augustine, Vatic. Pala . mel half-uncial, ew . use tw in half-uncial, they are in a smaller type of : ; at indentation and quotation marks—e.g., Paris 1 3 367 (No. 120) and Monte Cassino CL (No. 140), both of the sixth — rhe iffer i i f Origen (No. 133)—a . remarkable fo different style is found in Lyons 483 0 : . i i t off as in modern times by marks other peculiar features—where the citations are set off prone i lly single, but if a second citation before and after the quotation. These are usua i ks. In Munich 6436, the rs on the same page it has double quotation mar In I Pauline Epistles, probably of Spanish origin (No. 87), indentation is used : st citation begins in the body of a line a single quotation mark is placed imm ely before its first word. RULING. er of how sheets were ruled and how many were ruled at a time clearly attention than it has yet received. How useful may become as aids in dating MSS., especially MSS. from a single school, re . "y been demonstrated by Professor Rand.* It is also not unlikely that a careful stu J of the methods followed in different scriptoriums may furnish new and unsuspecte - touchstones for localizing MSS. Unfortunately my data on the subject of ruling are ically all our oldest MSS. the te; but such facts as I have show that in practi “aan at of each quire (that is, £1" and f.8 of the normal gathering) received mm direct impression of the stylus, and that these outside were In anumber of very ancien = , skin (see column marked Quire marks). ery a ing i ir si the Codex Bobiensis (No. 62), Oo the ruling is on the hair side. They are f supposedly African origin ; the Codex Bezae (No. 67), is as eccentric as its text ; the Ravenna (No. 1 30) in which other apparently un-Italian characteristics are noticeable; the 1 Palacographia Latina V. (1927), pp. 52-77: = te 62 MORE FACTS ABOUT OUR OLDEST LATIN MANUSCRIPTS Paris Hilary from Lyons (No. 1o1), probably a local : fragments of Ecclesiasticus and of (Nos. 82 and from centre situated outside the main current. o insure greater regularity in the size of letters, a fi i i de luxe, have ruling for the headline as well as for sclumunen Shaoaavan ie: zano and Verona purple Gospels, the Lyons Psalter, and the Vienna Rufinus (Nos 77) 75: 128). This feature, it may be mentioned in passing, is found in the entail illuminated Greek Codex Purpureus, the Vienna Genesis. Another fact which emerges is that the vertical bounding-lines which confine the written space were, in our abe MSS.., single lines. The use of double bounding-lines seems to be a later practice, The foregoing observations brief as the i i 1 : y are, may fairly be said to co the tentative conclusions reached in the former paper. I re not flatter me eo anything startlingly new has emerged from the facts presented ; certainly no dating criteria have been established capable of mechanical application. Yet the two articles will have been written to some purpose if any useful hints have been thrown out here ro — and if the data collected and set down in tabular form will make it easier for scholars to discern what some of the facts about manuscripts really are.! E. A. Lowe. OXFORD. 1 For the sake of convenience I repeat the list 26. Verona LX i : II. (60): Justin., Unci of forty-seven MSS. examined in my former 27. Paris 17225: article : 28. Na i i a 28. Naples (olim Vienn : i 1. Vatic. 3256: Virgil, Square capitals, s. V. s. V 2. Vatic. 3225: Virg., Rustic capitals, s. IV 2 i ,s. IV. 9. Trent (olim Vien : i 3. Florence Laur. 39. 1: Virg., Rustic, s. V. Ss. “ya criti: 4. Vatic. Pal. 1631: Virg., Rustic, s. V 30. Vatic ; i i i ‘ ic, s. V. ; - 5763 : Hiob, Judic., Uncial, s. V ex. 5. Naples IV.A8: Lucan, Rustic, s. IV. 31. Milan Ambr. C 73 inf.: ‘Lae fr Uncial 6. Vatic. 5750: Persius, Juvenal, Rustic, s. V. s. V ex. ke : ae 192: Sallust, Rustic, s. V. 32. Paris 6400 G: Act. and Apoc., Uncial, s. V. in Reg. 2077 : Cicero, Rustic, s. IV. 33. Paris Gr. 107: Ep. Pauli, Uncial, s, V//VI. 9. Turin AIL 2: Cicero, Rustic, s. IV. 34. Naples (olim Vienna 16): Apoc. Ep. Apost 10. Vatic. Pal. 24: Livy, Rustic, s. V. Uncial, s. V 7 11. Verona XL. (38): Livy, Uncial, s. V Leon Cath. 12. Paris 5730: Livy, Uncial,s. V. s. VII. 13. oe 10696 : Livy, Uncial, s. V. 36. Paris 10592: Cyprian, Uncial, s. V ex. 14. iar 15: Livy, Uncial, s. V. 37. St. Gall 213: Lactantius, Uncial,s. V. 15. Vatic. 5757: Cicero, Uncial, s. IV./V. 38. Verona XIII. (11): Hilar., Uncial, s. V ex. 16. Vienna 1a: Pliny, Hist. Nat., Uncial, s. V i i a A 39. St. Paul in Carin. XXV, 3. : 17. Rome, Vitt. Em. 2099 (sess. 55): Pliny, H.N., Uncial, s. VI in. si nai Uncial, s. V »& ¥. 40. Wirzburg th. q. 2: Hieron., Uncial, s. V 18. St. Paul in Carinthia XXV. 2. 36: Plin ierc cial, H.N., Uncial, s. V ex. VI ex 19. New York, Morgan M 462: Pliny, Epist., 2. Vero XX : i inci 4 VIII. (26): Augustin., Uncial, 20. a 5750 : Fronto, Uncial, s. ¥. 43. Petrograd Q. v. I. 3: Augustin., Uncial, s. V. 21. Paris 12161 : Asper, Uncial, s. V ex. 44. Paris 12634: Augustin., Uncial, s. VI./VII. 22. Vienna 1: Ulpian., Uncial, s. V. ex. 45. Vatic. 5766: Cassian, Uncial, s. VIII 23. Verona XV. (13) : Gaius, Uncial, s. V ex. 46. Verona LI. (49): Maximus Taur Uncial 24. Vatic. 5766: Frag. ante-Justin., Uncial, s. V. s, VI in, y 25. Flor. Laur. s.n.: Justinian, Uncial, s. VI. 47. St, Gall 912: Glossar., Uncial, s. VIII in Amer historical, the clause enforced | in Latin, si with verba ment of tl list of B.Gc. VII. date of tl represent | twenty-fiv assuming between t! been anti- the conte1 Aias, argu as a whol The Natu middle vc languages forms. Bulle Cons and Hofn to the que philologis showing unfounde: Herr The Quid ag: question some sug speaker’s In this u Nordic ve form or « the latte! in which or constr (H. C. I of which ‘tunc tin antequan malachite (b) ferrot | fe : Orleans. so come y copies. e Sarez- Nos. 74, famous. =merges r oldest ctice, oborate elf that dating he two thrown 1 make Latin OWE. s. VI. 4 Uncial, Uncial, V ex. Uncial, A post., Jncial, V ex. \bros., ¥. al, s, SUMMARIES OF PERIODICALS LITERATURE AND GENERAL. American Journal of Philology. XLVIII. 2. April, May, June, 1927. T. Frank, Naevius and Free Speech, while accepting the exile of Naevius as historical, argues that libel was not a ‘capital’ crime under the Twelve Tables, but the clause siue carmen condtdisset, etc., was a piece of party legislation temporarily enforced by some praetor’s edict. W. W. Kuk, Observations on the Indirect Volitive in Latin, suggests that the predominance of the subjunctive in ‘ volitive’ constructions with verba dicendi is due to (1) the frequency of parataxis, and (2) the special develop- ment of the negative clause in which the infinitive is practically unknown. Gives a list of examples of infinitives from Latin of various periods. S. G. Oliphant, Caesar B.G. VII. 41, 1: A Defence of the MSS., retains noctis with tribus horis, because at the date of the march (suggested to be the middle of May) nine night horae would represent about seven of our hours, ample time for an army to cover a distance of twenty-five Roman miles. C. Pharr, The Testtmony of Josephus to Christianity, assuming that the reference to Christ (XVIII. 63-64) is an interpolation (made between the time of Origen and that of Eusebius), and that the original must have been anti-Christian in tone, uses the nature of the context in an attempt to reconstruct the contents of the passage. S. E. Bassett, The Single Combat between Hector and Aias, argues that the passage H 1-312 is an integral and necessary part of the poem as a whole, and defends it in detail against the charge of lateness. E. F. Claflin, The Nature of the Latin Passive, seeks to show that ‘the Latin passive is essentially a middle voice’ by arguments derived from (1) the analogy of other Indo-European languages, (2) such phenomena as deponent verbs and the reflexive use of passive forms. Bulletin de la Soe. Ling. de Paris. XXVII. 2. Consists of book reviews. Meillet deals with the new edition, by Leumann and Hofmann, of Lateinische Grammatth (Stolz-Schmalz); special reference is made to the question of accent, the outstanding point at issue between French and German philologists. Lachmann’s law (dctus : dgo) does not apply in the group of words showing -ss-¢.-ds- (cf. sessus). The theory of a -dr-->-tv- change in Latin is unfounded. Hermathena. XLIV. 1926. The classical articles comprise: (1) Latin Questions of the Types Quid ago? and Quid agam? (C. Exon). A question of the type Quid ago? was a spontaneous question for advice; a question of the type Quid agam ? expressed a protest against some suggestion already conveyed either by someone present or (less often) by the speaker’s own thoughts, E. gives numerous examples to illustrate the difference. In this use of the indicative, among others, E. finds a survival of the primitive Nordic verb. He attributes the survival to the principle that whenever a grammatical form or construction takes the place of a previously existing form or construction the latter does not necessarily disappear altogether, but remains in every situation in which it serves the ends of speech as well as the more recently developed form or construction would have done. (2) Notes on the Historia Naturals of Pliny (H. C. Bailey). These notes are written to discuss passages, the subject-matter of which needs elucidation from the scientific point of view. Thus on XXXIII. 88, ‘tunc tinguitur chrysocolla alumine schisto et herba supra dicta (sc. luto) pingiturque antequam pingat,’ B. reports an experiment made by himself: ‘ Finely powdered malachite (bluish-green) was boiled with weld (lutum) and (a) aluminium sulphate, (b) ferrous sulphate, or (c) iron alum. With (a) a fine emerald-green powder, far Ss. PRE 64 SUMMARIES OF PERIODICALS superior in colour to the original malachite, was obtained, while with (d) and (c) the pigment was a dark bottle-green. (3) The Zulichemianus, Mazarinianus and other MSS, 7 of the Fasti of Ovid (E. A. Alton). The tendency of editors of the Fasti has been to con- centrate on the codd. Mallersdorfiensis (D), Reginensis (R), and Ursinianus (U), other readings which for the most are derived from the Zulich. and Maz. being somewhat under suspicion. The Zulichem readings were found by Heinsius in an interleaved 7 Aldine. He attached great importance to them, and was struck by their similarity | to those of the Mazarimianus. The Aldine used by Heinsius was rediscovered at Leyden in 1900, and the actual MS. from which the readings were taken is at Brussels. The part containing the Fast: has been assigned to the eleventh century. A. gives an account of this MS. and seeks to estimate its value. He holds that ‘ the claims of the Z tradition have been unduly depreciated in the past, and I may say merely on hearsay evidence.’ (4) The Medieval Commentators on Ovid’s Fasti (E. A. Alton). A. has no high opinion of these commentators, even of the highly esteemed work of Arnoul de Saint-Euverte, professor at Orleans in the twelfth century, of whose work he gives some specimens. There are, however, in the Brussels MS. (G) excerpts of a commentary by an unknown scholar who exhibits critical insight and method in his attempt to throw light on the dark places of the Fastit. His work, indeed, left its trace on certain MSS. of the poem, but before the end of the thirteenth century the frothy exegesis of Arnulphus and his followers had- prevailed. Neue Jahrbucher fur Wissenschaft und Jugendbildung. II. 6. 1926. O. Weinreich, Antikes Gottmenschentum. Discusses deification from Homeric to Christian times, with special emphasis on Pythagoras, Empedocles, Hellenistic kings, and Roman emperors. III. 1. 1927. F. Oertel, Die Soziale Frage im Altertum, Practical ancient socialism never looked beyond the redistribution of wealth among citizens: the slave revolts of the second and first centuries B.c. had no revolutionary communistic programme. III. 2. 1927. A. Busse, Kulturgeschichtliche Anschauungen in den Dvamen des Sophokles. Deals chiefly with Antigone 332 sqq. Sophocles accepted the sophistic evolutionary theory of material culture, but denounced its extension to ethics and religion. He was perhaps influenced by the philosopher Archelaus. H. Bulle, Antthe Technther und Architekten. An appreciative account of the late W. Sackur’s Vitrvuv und die Poliovketiker. Vitruv und die christliche Antthe, 1925. Among other things S. proved that the greater part of Athenaeus zepi pnxavnpdrwv is a genuine letter addressed to Marcellus, the captor of Syracuse, by a Greek engineer. III. 3. 1927. A. Tumarkin, Das Apollinische und das Dtonysische in dev griechischen Philosophie. Anaximander, Heraclitus, Protagoras, and Socrates are treated as Dionysiac; Pythagoras, the Eleatics, Plato, and Aristotle as Apolline. F. Petri, Die Wohil- fahrispflege des Augustus. Augustus’ distributions of corn and money, building activities, public shows, and the like, followed no reasoned programme of social reform, but were purely political attempts to popularize the Principate in the city of Rome. III. 4. 1927. F. Koepp, Das Problem dey dritten Dimension in dev griechtschen Flachenkunst (with 4 plates). Discusses the development of perspective, shading, etc., in ancient painting and relief. | | 4 - <7, | 3 | | : wah | |