I have also collated the thirteenth-century _Parisinus lat. 7993_, Heinsius' _codex Regius_. At ix 46 _P_ offers the correct _cernet_ for _credet_; _cernet_ is also the reading of _M_ after correction by a later hand and of the thirteenth-century _Gothanus membr. II 121_. At vi 7 _P_ alone of collated manuscripts agrees with _C_ in reading _praestat_ for the correct _perstat_. _P_ agrees with _L_ in reading _niuibus_ for the other manuscripts' _nubibus_ at v 5, _adeptum_ for _ademptum_ at vi 49, _signare_ for _signate_ at xv 11, and in the orthography _puplicus_ for _publicus_ at ix 48, ix 102, xiii 5, and xiv 16. The manuscript has many corruptions: a few examples are i 30 _igne_ for _imbre_, ii 18 _supremo_ for _suppresso_, iv 6 _pace_ for _parte_, vi 34 _uirtus_ for _uirus_, vii 15 _piacula_ for _pericula_, ix 42 _praeterea_ for _praetextam_, x 63 _in harena_ for _marina_, xiv 39 _conuiuia_ for _conuicia_, and xvi 24 _sacri_ for _scripti_. However, _P_ has no unique variants with any probability of correctness. To have given a full report of _P_ would have involved a considerable expansion of an already long apparatus, and I have cited the manuscript only occasionally, where a reading is only weakly attested by the other manuscripts.

Titles

_MF_ and _B2H2I2T2_ usually supply titles for the poems. As will be seen from the apparatus, there is considerable variation among the titles, and there is no reason to suppose that they form an authentic part of the transmitted text.

The manuscript authority for the text of _Ex Ponto_ IV

By and large the manuscripts of the fourth book of the _Ex Ponto_ offer a remarkably uniform text of the poems, and one which, considering the late date of the manuscripts, is in surprisingly good condition. I believe that all the manuscripts, with the exception of _G_, are descended from a single archetype. _B_ and _C_ are the best witnesses to the text of the archetype, although the other, more heavily contaminated and interpolated manuscripts are indispensable, since they correct the peculiar errors of _B_ and _C_.

The present edition

The apparatus of this edition is intended to be a full report of _BCMFHILT_ and of the fragmentary _G_; some reports are also given of _P_. It includes corrections by original and by later hands.

When no manuscripts are specified for the lemma in an entry, the lemma is the reading for those manuscripts not otherwise specified. For instance, the entry

deductum carmen] carmen deductum _M_

indicates that _deductum carmen_ is the reading of _BCFHILT_, while _carmen deductum_ is the reading of _M_.

I have from time to time cited from earlier editions readings of manuscripts which I have not collated. To make it clear that I have not personally verified these readings, I have added in parentheses after the citation the name of the editor whose report I am using. Professor R. J. Tarrant has inspected some nine manuscripts to see what readings they offered in some particularly vexed portions of the poems; I have similarly indicated when I am obliged to him for information on a manuscript.

The _excerpta Scaligeri_ mentioned at xiii 27 I know of through Heinsius' notes as printed in Burman's edition; according to M. D. Reeve (_RhM_ CXVII [1974] 163), the original excerpts are still extant in Diez 8 2560, a copy of the _editio Gryphiana_ of 1546. Reeve also gives identifications of certain of Heinsius' manuscripts; when citing Heinsius' codices, I give the modern name when the manuscript has been identified and is still extant.

The greater number of the manuscripts dealt with have been corrected, some heavily. In my apparatus _B1_ means "the original hand in _B_" and _B2_ means "a correcting hand in B". _B2ul_ indicates that the reading of _B2_ is clearly marked as a variant reading. _B2gl_ indicates that the entry is marked in the manuscript as a gloss; _B2(gl)_ indicates a gloss not marked as such. I have reported glosses where they contribute to the understanding of a textual problem.

If different correctors have been at work in different passages, both are called _B2_. If a later hand has made a correction after _B2_, the later hand is called _B3_. When I place _B1_ in an entry but do not report _B2_, it can be assumed that _B2_ has the lemma as its reading.

Sometimes a corrector has altered the original text so much (without however erasing it entirely) that only the altered reading can be made out. In such cases I have used the siglum _B2c_. Where a corrector has inserted or altered only certain letters of a word, I have indicated this in the HTML version of this edition by underlining the letters involved. In the Text version, these letters are capitalized.

Where the correction is apparently by the original scribe, _Bac_ indicates the original reading, and _Bpc_ the correction.

The asterisk is used to indicate illegible letters, and the solidus (/) erasures.

When reporting variants, I have tried to indicate the spellings actually found in the manuscripts, but since mediaeval spellings do not in themselves constitute variant readings, they have not usually been reported when the text is not otherwise disturbed. I have been more generous with proper names, but have often excluded confusions of _ae_, _oe_, and _e_, of _i_ and _y_, of _ph_ and _f_, of _c_ and _t_, the doubling of consonants, and the loss or addition of the aspirate.

The apparatus is intended to include a comprehensive listing of all conjectures proposed. When the author of a conjecture is not a previous editor of the poems, I have given a reference either to the publication where the emendation was first proposed, or to the earliest edition I have consulted which reports the emendation. Conjectures of Bentley are from Hedicke's _Studia Bentleiana_. Conjectures of Professor R. J.

Tarrant, Professor J. N. Grant, and Professor C. P. Jones were communicated to me by their authors.

Printed editions

The first editions of the works of Ovid were printed in 1471 by Balthesar Azoguidus at Bologna and by Conradus Sweynheym and Arnoldus Pannartz at Rome. The Bologna edition was edited by Franc. Puteolanus, and the Rome edition by J. Andreas de Buxis. Lenz's edition gives numerous readings from both editions; to judge from his reports, their texts of the _Ex Ponto_ were derived from late manuscripts of no great value. The Roman edition, however, contained the elegant correction of _iactate_ to _laxate_ at ix 73.

For my knowledge of other early editions of the _Ex Ponto_ I have relied upon Burman's large variorum edition of the complete works of Ovid, published at Amsterdam in 1727. The edition contains notes of various editors of the sixteenth and early seventeenth centuries, among them Merula, Naugerius, Ciofanus, Fabricius, and Micyllus. Although I have occasionally quoted from these notes, they are in general of surprisingly little use, containing for the most part unlikely variant readings from unnamed manuscripts and explanations of passages not really in need of elucidation.

The principal event in the history of the editing of the _Ex Ponto_ was the appearance at Amsterdam in 1652 of Nicolaus Heinsius' edition of Ovid. Heinsius took full advantage of the opportunity his travels as a diplomat gave him of searching out manuscripts, thereby gaining a direct knowledge of the manuscripts of the poems which has never since been equalled[11]. Heinsius also possessed an unrivalled felicity in conjectural emendation. Some of his conjectures are unnecessary alterations of a text that was in fact sound, some of his necessary conjectures are trivial, and are already found in late manuscripts of the poems or could have been made by critics of less outstanding capacities; but many are alterations which are subtle and yet necessary to restore sense or Latinity. The present edition returns to the text many conjectures and preferred readings of Heinsius that were ejected by editors of the nineteenth and twentieth centuries.

[Footnote 11: In recent years much progress has been made in identifying the manuscripts Heinsius used. See the monograph of Munari and the articles of Reeve and Lenz listed in the bibliography.]

The edition of Heinsius formed the basis of all editions published during the two centuries that followed. Of these editions the most important was the 1727 variorum edition of Burman already referred to.

It is from the copy of that edition at the University of Toronto Library that I have obtained my knowledge of Heinsius' notes. Burman was apparently the first editor to make use of _F_. On occasion he differs from Heinsius in his choice of readings. At xvi 44 he made the convincing conjecture _Maxime_ (codd _maxima_), subsequently confirmed by _B_ and _C_. His notes are informative; and my note on x 37-38 in particular is greatly indebted to him.

For poem x Burman reproduced some notes from an anthology of Latin verse for use at Eton, produced by an anonymous editor in 1705[12].

[Footnote 12: _Electa minora ex Ovidio, Tibullo et Propertio_, London, 1705. The book was reprinted as late as 1860 (_Brit. Mus. Gen.

Catalogue_, vol. 177, col. 470). I quote some of the notes on x in the commentary and apparatus.]

In 1772 Theophilus Harles published at Erlangen his edition of the _Tristia_ and _Ex Ponto_ 'ex recensione Petri Burmanni'. Harles was the first editor to make use of _B_. In the introduction to his edition Harles relates how he wrote von Oeffele, librarian to the Elector of Bavaria, asking if there was any manuscript in the Elector's library that might be helpful in preparing his edition, and thereby learned of the existence of _B_. It is clear from Harles' introduction that he fully appreciated the manuscript's importance; and in his notes he gives many of its readings, pointing out where it confirmed suggestions of Heinsius and Burman. However, his text is simply reprinted from Burman's variorum edition.

W. E. Weber's text of _Ex Ponto_ IV in his 1833 _Corpus Poetarum Latinorum_ is in effect a reprint of the Heinsius-Burman vulgate, except that at viii 59 he prints the manuscripts' incorrect accusative form _Gigantes_ (Heinsius _Gigantas_). But this fidelity to the vulgate text seems not to have been the editor's intention: in his introduction he speaks of 'Heinsianae emendationes felices saepe, superuacuae saepius ... quarum emendationum partem Mitscherlichius eiecit [Gottingen, 1796; I have not seen the edition], maiorem eiicere Iahnius coepit [Leipzig, 1828: the part of the edition containing the _Ex Ponto_ was never published]. dicendum tamen, etiamnunc passim haud paucas fortasse latere Heinsii et aliorum correctiones minus necessarias in uerbis Ouidianis, quas accuratior codicum inter se comparatio, opus sane immensi laboris, extrudet'. It would be understandable enough if Weber, faced with the labour of editing the entire corpus of Latin poetry, found himself unable to effect a radical revision of the text of the _Ex Ponto_.

In 1853 there appeared at Leipzig the third volume of Rudolf Merkel's first Teubner edition of the works of Ovid, containing his text of the _Ex Ponto_. The part of Merkel's introduction dealing with the _Ex Ponto_ is entirely concerned with describing the appearance, orthography, and readings of the ninth-century _Hamburgensis scrin. 52 F_. The manuscript ends, however, at III ii 67, and Merkel says nothing of the basis for his text of the later poems, which in general is the Heinsius-Burman vulgate.

In 1868 B. G. Teubner published at Leipzig Otto Korn's separate edition of the _Ex Ponto_. Korn's apparatus is the first to have a modern appearance; but this appearance is deceptive, for of the twenty sigla Korn uses, ten are for individual or several manuscripts collated by Heinsius, and only five are for manuscripts collated by Korn himself.

The edition is important, since Korn was the first editor to make substantial use of _B_ in constituting his text. Usually he printed the text of _B_ in preference to the vulgate: 'Ceterum eas partes in quibus _A_ caremus, [Greek: b] [=_B_] libri uestigia secutus restitui, prorsus neglectis recentiorum exemplarium elegantiis, quorum ad normam N.

Heinsius, cuius in tertio quartoque libro R. Merkelius assecla est, textum conformauit' (xv).

There was some reason to review critically the vulgate established by Heinsius and Burman. Even Heinsius was capable of error; examples of this in _Ex Ponto_ IV include his preference for the inelegant _idem_ for _ille_ at iii 17, for the impossible _ullo_ instead of the better attested _nullo_ at v 15, and for the obvious interpolation _domitam ... ab Hercule_ at xvi 19 instead of _domito ... ab Hectore_. His most pervasive fault is a partiality for elegant but unnecessary emendation: often he is guilty of rewriting passages which are in themselves perfectly sound. A typical instance is vii 30: Heinsius' _globos_ is elegant enough, but there is no reason to suspect the transmitted _uiros_.

Some of the readings proposed or preferred by Heinsius had been unnecessary or wrong, but many had been necessary to make sense of the text; and Korn is often guilty of damaging the text by excluding readings not found in _B_. The supreme example of this is his restoration of the manuscripts' reading _iactate_ for _laxate_ at ix 73.

Korn used the collation of _B_ by Harles, which had errors and omissions (in his preface Harles had warned that his report might contain errors[13]), so that at i 9 Korn prints _in istis_ and at x 83 _perstas_, without noting in his apparatus that _B_'s false readings were _ab istis_ and _praestas_ respectively. He was aware that at xi 21 _B_ read _mihi_, but printed _tibi_ nonetheless, although Burman had already explained why _mihi_ was the correct reading.

[Footnote 13: 'Diligenter autem et religiose tractaui codicem et singulas epistolas bis, et in locis uexatis saepius contuli. Neque tamen, quae hominum est imbecillitas, aciem oculorum quaedam effugisse, negabo' (xi-xii).]

A curious feature of Korn's edition is its dual apparatus: below the report of manuscript variants is a listing of passages where his text differs from those of Heinsius and Merkel: 'Lectiones discrepantes editionum Heinsii et Merkelii adposui, ut et quantopere Ouidius Heinsianus a genuina forma discrepet dilucide perspiciatur, et quibus locis a Merkelio discesserim facilius adpareat' (xxxii). Korn ejects such obviously correct readings as _leuastis_ at vi 44 and _laxate_ at ix 73; in each instance the true reading is printed in large type at the bottom of the page. In addition, Korn rather unfairly included as different readings what were in fact only spellings which did not conform to the purified orthography then coming into use. _Cymba_ does not differ from _cumba_ (viii 28), nor is _Danubium_ a variant for _Danuuium_ (ix 80), nor again is _Vlysses_ different from _Vlixes_ (x 9). Finally, the second apparatus at several points misrepresents what Heinsius actually thought.

Korn's confusion on this point is understandable, since determining Heinsius' textual preferences is often more difficult than it might at first appear. Editions were published under his name which did not incorporate all his preferred readings[14]; even the lemmas to his notes are taken from the edition of Daniel Heinsius, and are not a guide to Heinsius' own view of the text, which can only be discovered by reading the actual notes[15]. A good example of this can be found at x 47. Here Heinsius' text reproduces the standard reading _Cratesque_. The lemma in his note is _Oratesque_, the reading of Daniel Heinsius' edition. In the note itself Heinsius indicates his preference for the conjecture _Calesque_, communicated to him by his friend Isaac Vossius. Here Korn, along with all modern editors, prints _Calesque_ in his text; he reports _Cratesque_ as Heinsius' reading.

[Footnote 14: A. Grafton has noted that Heinsius' publisher Elzevier seems to have been unwilling to alter the text as it already existed (_JRS_ LXVII [1977], 173). I owe my knowledge of Heinsius' editorial practices as here described to Professor R. J. Tarrant, who has examined the Harvard copies of the 1664 edition of Heinsius' text (without notes), the 1670 Leiden edition of Bernard Cnippingius, which reproduces Heinsius' notes, and the 1663 reprint of Daniel Heinsius' edition.]

[Footnote 15: Consequently any statements I make on Heinsius' editorial practices are based on explicit statements in his notes.]

Korn made one important conjecture in _Ex Ponto_ IV, printing _decretis_ at ix 44 for the manuscripts' _secretis_.

For the third volume of his complete edition of Ovid, published at Leipzig in 1874, Alexander Riese drew on Korn's edition, but was less radical in following the readings of _B_: 'nec eclecticam quam dicunt N.

Heinsii nec libri optimi rigide tenacem O. Kornii rationem ingressus mediam uiam tenere studui' (vii). Riese restores Heinsius' preferred reading in only about a quarter of the places where it was deserted by Korn; even so, no editor since has shown such independence in the selection of readings.

In 1881 there appeared at London a text of _Ex Ponto_ IV with accompanying commentary by W. H. Williams. The text, which Williams says is drawn from the "Oxford variorum edition of 1825", seems in general to be a reprint of the Heinsius-Burman vulgate with some readings drawn from Merkel's first edition. In spite of occasional conjectures and notes on variant readings, based on information drawn from Burman and Merkel, Williams is not generally concerned with the constitution of the text: his note on x 68 _curasque fefelli_ is 'so Tennyson in the "In Memoriam'". The commentary, which is about eighty pages long, consists largely of discussions of the cognates of various Latin words in other Indo-European languages, 'though the limits of the work preclude more than the _data_ from which a competent teacher can deduce the principles of comparative philology'. A typical note is that on i 11 _scribere_: 'from [root] skrabh = to dig, whence scrob-s and scrofa = 'the grubber,'

_i.e._ the pig; Grk. [Greek: grapho] by loss of sibilant and softening'. The edition has been only occasionally useful in editing the poems or writing the commentary.

In 1884 Merkel published his second edition of the poems of exile. In his previous edition he had in general followed Heinsius and Burman in the fourth book; in the new edition, without specifically saying so (although in his introduction he mentions the "codex Monacensis uetustior"), he generally alters his text so as to conform with _B_'s readings. He does not always desert his former text, rightly retaining _hanc_ at i 16, _quamlibet_ at iv 45, and _tempus curasque_ at x 67; he also keeps _lux_ at vi 9 and _domitam ... ab Hercule_ at xvi 19.

In his 1874 monograph _De codicibus duobus carminum Ouidianarum ex Ponto datorum Monacensibus_ Korn had made known the existence of _C_. S. G.

Owen's first edition of the _Ex Ponto_, printed in Postgate's _Corpus Poetarum Latinorum_ in 1894, was the first edition to report this manuscript as well as _B_. His text is unduly partial to the readings of _B_ and _C_, and his well-organized apparatus is so abbreviated as to be deceptive. It cannot be relied upon even for reports of _B_ and _C_. At ix 73 it gives no hint that for four centuries editors had read _laxate_; many of Heinsius' preferred readings are similarly consigned to oblivion. At vi 5-6 he reports Housman's ingenious repunctuation, presumably communicated to him by its author.