STOP Early Journal Content on JSTOR, Free to Anyone in the World This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in the world by JSTOR. Known as the Early Journal Content, this set of works include research articles, news, letters, and other writings published in more than 200 of the oldest leading academic journals. The works date from the mid-seventeenth to the early twentieth centuries. We encourage people to read and share the Early Journal Content openly and to tell others that this resource exists. People may post this content online or redistribute in any way for non-commercial purposes. Read more about Early Journal Content at http://about.jstor.org/participate-jstor/individuals/early- journal-content . JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people discover, use, and build upon a wide range of content through a powerful research and teaching platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please contact support@jstor.org. Classical Philology Vol. II April, igOf Nc ON A BODLEIAN COPY OF PLINY'S LETTERS By Elmer Truesdell Merrill The book which is the subject of this article was bought at auction in Oxford, in the year 1708, by Thomas Hearne, the learned editor of the edition of Pliny's Letters published at Oxford in 1703. The purchase he himself records in an autograph note on the title-page. Hearne fully recognized the source and impor- tance of the MS readings of the book, and collated them with care, as is shown by two copies of his 1703 edition, now in the Bodleian, in which these readings are inserted by his hand, apparently in preparation for an enlarged edition, which was published at Amsterdam in 1709, if bibliographies can be trusted. I have not been able thus far to find a copy. But Hearne's work seems to have been disregarded by recent editors, and the book rested forgotten in the Bodleian until Mr. E. G. Hardy called attention to it in the Journal of Philology, Vol. XVII (1888), and repeated part of his statement in the intro- duction to his edition of Pliny's Correspondence with Trajan, which appeared a few months later. The article from the Journal of Philology has now been reprinted in his Studies in Roman History. Mr. Hardy appears to have overlooked Hearne's actual use of the book, though of course he notices his purchase of it. In the course of a summer's work in England in the collection of further material for my forthcoming critical edition of Pliny's Letters, I have recently had occasion to study this volume with [Classical Philology II, April, 1907] 129 130 Elmer Truesdell Merrill some degree of care, and have arrived at conclusions in certain respects quite different from those reached by Mr. Hardy. These conclusions, and the grounds for them, it is my purpose to set forth here, with grateful recognition of the obligation we are under to Mr. Hardy for calling renewed attention to the book, now so many years ago, and for emphasizing again its importance. Mr. Hardy judged it proved that the volume under considera- tion belonged to Aldus himself, was emended throughout in his hand from certain sources mentioned in the preface to his edition of 1508, and served as the copy from which the printer set up the type of that edition in its first form. I, on the other hand, trust that I can show good reason for believing that the volume never was in the possession of Aldus, but was the property, and, in cer- tain of its MS parts, the work, of quite a different man ; and that its importance for the reconstitution of the text of Pliny's Letters is all the greater on this account. The volume concerned has the Bodleian shelf-mark "Auct. L. 4. 3," and is cased in a Bodleian binding of only some half a century ago. Three individual books are, however, included within its covers. The first of these is the edition of the nine books of Pliny's Letters published at Bologna in 1498 by Beroaldus; the second is the "Epistole 46 nuper reperte" of the correspondence with Trajan published by Avantius in 1502; the third is the "Opus- culum eruditum" of Beroaldus himself, published at Bologna in 1497. The three books are of the same format, and of the same general character of paper and type, and apparently were included in the same binding long before the present covers were affixed; for a wormhole of antiquarian character has pierced the last pages of part 2 and the first of part 3 in accurate prolongation. But part 3 was added to parts 1 and 2 at a later date than that of the union of these two parts. For parts 1 and 2 have the folios numbered consecutively in the same ink, and apparently in the same hand, that has added many of the marginal and interlinear annotations that are later to be discussed, and furthermore, several of these annotations are cross-references by the number of the folio, and the hand is substantially contemporaneous with the printed text. Part 3, on the other hand, while it shows a num- On a Bodleian Copy of Pliny's Letters 131 bering of the folios consecutive with that of parts 1 and 2, yet exhibits in this numbering a different hand and ink. It is of no further interest for our discussion. Some facts concerning the early editions of Pliny's Letters must be borne in mind, in order to comprehend the bearing of the later discussion. The first edition (Keil's p) was like the large majority of the MSS of the Letters that date from the fifteenth century in containing eight books only, the eighth book being omitted without sign or remark, and the ninth numbered as the eighth. But an edition published without indication of place, date, or printer's name, suddenly added to the already well-known eight books a ninth. This edition is now generally ascribed by experts in such matters to the Roman printing-house of a certain Schurener from Boppard, and is believed to have been published about the year 1474. This Roman edition (Keil's r) was sub- stantially the basis of a number of other editions that appeared during the next quarter of a century. Of these later editors Pomponius Laetus appears to have used also the Vatican codex 3864 (Keil's V), which contains Books i— iv only, and he may have had access also to the same MS that r used, or to one like it. The other editors of this period either repeated their predecessors, or emended them from some of the then common eight-book MSS, or from conjecture. Roughly speaking, the MS authority of r in its first eight books was not substantially different from some of the eight-book MSS (all of the fifteenth century) still extant. But no such MSS known to me have any trace of more than eight books. A Prag MS, to be sure, that was followed by Titze in his edition printed at Leipzig in 1823, and another, and better, copy from the same original, which copy (of the early fifteenth century) I have recently found in the British Museum (Harleianus 2497, olim Cusanus), contain some letters of the ninth book intermingled with others of the seventh and eighth books, and numbered as Book viii ; but these codices, lacking in completeness and confused in the order of the letters, are poor relations of the M-family, and certainly have no connection with r. I have found no trace of any such MSS in Italy. 132 Elmer Teuesdell Merrill What r's source was cannot be at present determined. It has utterly disappeared from human ken. It was certainly not M (the codex Mediceus, XL VII. 36 of the Laurentian Library), nor any defective member of the M-family ; for the differences of reading between r and M are too great to be accounted for on this theory. And furthermore, it is tolerably certain that M was brought to Italy in 1508, was not known to editors before that time, and was used by none before the second edition of Catanaeus, in 1518. Again, the ninth book of r is the eighth of M, and of the first Aldine edition (Keil's a), published in 1508. The reasons for trusting the arrangement in Ma, as above that in r and the other editions before Aldus, I need not here discuss. The eighth book is complete in M. But the book is not complete in r. It lacks 8. 3-18. 11, the sentence at the former point running nauium patiens eodem quo emerat, without remark or sign of any omis- sion. Now in the edition of Beroaldus, a copy of which forms the first part of the volume I am especially discussing, this same sentence reads nauium patiens hie mutilus Codex, eodem quo emerat. Just before hie a MS hand has added a cross + in the text, and repeated it in the margin, from which latter place a line with an angle at its middle point, where the word ua cat sits astride of it, is drawn in the margin to the bottom of the page. Then are inserted eight leaves of paper similar in general char- acter to that used for the printed book, and on these a manuscript hand, substantially contemporaneous with the date of the book, has written out the missing letters, and, for good measure, to make the last written page end with the text where the printed page that next follows begins, has copied also the last part of the page on which the lacuna occurred, thus duplicating 18. 11-20. 1. (But the copy was not from the text of Beroaldus himself, but from a MS that differed decidedly from his text. Mr. Hardy does not mention this interesting and important duplication, and, to judge from his statement, did not notice it.) Furthermore, the Beroaldus text thus made complete has many variant readings indicated in the margin, and between the lines of the text. These variants extend over both the printed portions and the written supplement, and are most of them apparently in On a Bodleian Copy of Pliny's Letters 133 a single hand, and in the same quality of ink, though I suspect they were not all written in a single sitting, or connected series of sittings. Mr. Hardy says that "throughout these nine books there are 155 variant readings inserted .... in the margin." I have not verified his count of the readings "inserted in the margin," for there is no possible reason for drawing any distinc- tion between these marginal corrections and those made between the lines of the text in the same hand and ink, which he did not think it necessary to reckon. But as to the total number, inter- linear and marginal together, it is well over five hundred, and that without counting mere corrections of spelling, which are not infrequent. Besides these variants there are very numerous marginal lem- mata of the sort common in such works that have been used by renaissance scholars. Some of these lemmata are in the same hand with the most of the variants. Others I suspect to have been written by the writer of the variants, but at another time, and in a more careless and easy manner. Others yet are in dif- ferent hands, of which three or four can possibly be distinguished, but unfruitfully. To turn now to part 2 of the volume under discussion: the first public appearance, after the Middle Ages, of Pliny's corre- spondence with Trajan was when Hieronymus Avantius, of Verona, printed, in 1502, "C. Plinii Iunioris ad Traianum PJpistole. 46. nuper reperte cum eiusdem responsis," explaining briefly in his preface, "Petri Leandri industria ex Gallia Plinii iunioris ad Traianum epistolas licet mancas deprauatasque habuimus : quas pro uirili mea castigatas impressorum beneflcio emittendas cen- SMt'mMS." The letters have both subject-headings and numbers, beginning with " C. De lacu Nicomedensium Epistola. XXVII." (i. e., Ep. 41, according to Keil's enumeration), and ending with "C De Diplomatibus. LXXIII." They include all the letters from 41 to 121, as now usually numbered, and in their order as in Keil's edition. No explanation is vouchsafed by Avantius as to the whereabouts of the missing letters 1-26 (i. e., 1-40, according to the enumera- tion now prevalent), though other things, I may stop long enough 134 Elmer Teuesdell Merrill to remark, make it likely that the aforesaid Petrus Leander, who- ever he may have been, or, at any rate, the unknown copyist, having access to some of the letters, had access to all, and copied all ; and that in some way the sheets containing letters 1-26 were lost before they came into the hands of Avantius. Whether Bero- aldus and Catanaeus, who issued the same letters in 1502 (a few months after Avantius) and 1506 respectively, had access to any- thing more than the printed text of Avantius is a question I need not discuss here. At all events their MS authority probably did not go back of the truncated copy that was the source of the edition of Avantius. But in 1508 Aldus issued the complete works of Pliny as we now have them. The only question concerning his sources that is of present importance for our discussion is the statement in his preface that Iohannes Iucundus (Giovanni Giocondo, of Verona, a well-known priest, scholar, and architect, who went to Paris in 1499 to build the Pont Notre Dftme for Louis XII) brought him seven copies of Pliny's Letters, some in manuscript, and some in print, indeed, but collated with manuscripts. Of these the most remarkable was a complete copy made "in Gallia" by Iucundus himself of a wonderful MS, which two years later was itself brought from the same country by the Venetian ambassador, and also given to Aldus. This MS Aldus describes as written in characters so different from the current Italian (adeo diuersis a nostris characteribus) that only with long practice could it be read, very correct in text, and so ancient that Aldus can well believe it to have been written in Pliny's own age! Any fair interpretation of the further statements in the Preface indicates that this marvel contained ten complete books, the nine books in due order, and the correspondence with Trajan as the tenth. The extreme age that Aldus was disposed to attribute to the MS will of course occasion no wonder in the minds of those who are familiar with the vague notions on such matters that prevailed among scholars before the study of paleography had been developed into some- what of a science. The MS may have been written in one of the so-called "national" hands — Lombardic, Visigothic, or Merovin- gian. But if it were in a "gothic" hand of the twelfth or On a Bodleian Copy op Pliny's Letters 135 thirteenth centuries, it might have appeared sufficiently grotesque and illegible to a reader accustomed for the most part to the exceedingly clear Italian book hands of the fifteenth century. Some allowance must doubtless also be made for rhetorical exaggeration oil the part of Aldus. This wonderful MS, like so many others, appears to have van- ished from earth. Early editors saw no especial reason for pre- serving what was to them but copy for their own better printed texts. Possibly some leaves of it may be lying hid in old bind- ings ; possibly they went to cover preserve-jars, or tennis-racquets ; possibly into some final dust-heap. At any rate the MS is gone; the copy by Iucundus is gone; the copy of the correspondence with Trajan that Avantius owed to Petrus Leander is gone; if other editors had any other copies of Book x, in whole or in part, they are gone too. I must not take space to argue the matter here, but the very great probability is that the manuscript copied (or merely collated, for the letters already extant?) in France by Iucundus, and later brought thence to Aldus by Moncenigo, the ambassador, was the only complete copy that had survived the Middle Ages, the only one of its time that contained Books i-x. Now in the Bodleian copy of the edition of Avantius, which is bound up with the Beroaldus edition of which I have been speak- ing, there are bound in, immediately after the preface, nine leaves of paper, similar in general character to those inserted in the Bero- aldus. On these are written out, in the same hand that copied the missing letters of Book viii (ix) in Beroaldus, the miss- ing letters of the correspondence with Trajan, from . IIII . , to . XXVI . , as they are duly numbered here (4-40 by our numbering). Another leaf, on which presumably letters 1-3 B stood, has been cut out before the nine. This mutilation was perpetrated in the early years of the book, for the numbering of the folios disregards it, and this numbering was done, as I have already said, by the same early hand that has added most of the variant readings. Futhermore, this same hand that inserted over five hundred correc- tions in the text (and MS supplement) of Beroaldus, has likewise inserted many similar variants in the text of Avantius, and some 136 Elmer Tbubsdbll Mbkeill also in this MS supplement. (A few corrections were also made in the readings of the MS supplements of both the Beroaldus and the Avantius by the hand that wrote them). Furthermore, the same hand that has thus copiously interpolated both the Beroaldus and the Avantius has copied into the margin of the Beroaldus at the proper place (but, as the readings show, not from Aldus, nor from M) the letter (ix. 16) that is missing in all the editions of Pliny's letters before Aldus, as in all extant MSS except M, and at the end of the text of Avantius has added the following note : "he plinii iunioris epistole ex uetustissimo exemplari parisiensi et restitute et emendate sunt opera et industria ioannis iucundi prestantissimi architecti hominis imprimis antiquariiP Now Mr. Hardy observed that a very large majority of the cor- rections made in the margin of both the Beroaldus and the Avantius (it will be remembered that for some mysterious reason he does not count in the interlinear corrections made by the same hand, and apparently at the same time), and a large majority of the readings of the two MS supplements, agree with the readings of Aldus in 1508, and clearly are not derived from any other edition. He also noted that the same hand (apparently) had made numer- ous corrections of punctuation in the two books, many of which agree with Aldus. He also observed one other point which seemed to him to clinch the argument; and his conclusion from all the premises was that this book (Beroaldus and Avantius) was cor- rected in Aldus 1 own hand from the copy of the Parisian codex made and brought to him by Iucundus, and later perhaps from the codex itself ; that Aldus had the missing letters copied out from the same transcript and inserted; and that the book thus constituted and before us was the printer's copy from which Aldus had the first proof of his edition set up, and struck off. (He thus agrees with Hearne, who accepted all these readings as coming from the Paris codex of Iucundus, but goes beyond him in ascribing the book in its present shape to the correcting hand of Aldus.) I do not think it difficult to show that Mr. Hardy's conclu- sions, so far as they regard the relation of Aldus to the book, are quite untenable, and in attempting to show this I will treat of his clinching point first; and throughout this article I will On a Bodleian Copy of Pliny's Lettees 137 indicate the MS supplements to both Beroaldus and Avantius (i. e., viii. 8. 3-18. 11, and x. 4-40) by I, and the variants in interlinear space or margins (so far as they are by the hand that wrote also ix. 16, and the note concerning Iucundus at the end of the Avantius) by i. Mr. Hardy says, as to his clinching point: "A curious mis- print in Aldus, otherwise inexplicable, receives immediate and con- vincing explanation by turning to a marginal reading . . . . , and supposing that the Bodleian copy was before the printer of the Aldine edition." The passage referred to is in viii. 6. 10, where in i reads, cut nulla re, and Aldus reads, cut mulla re. This error of Aldus is clearly due, says Mr. Hardy, to a misreading of the margi- nal note to the Beroaldus ; for in the lack of a caret, and of a dot to the i, the printer read in as m, and supposed it to be a correction of the n below, and so printed it, giving us mulla instead of in nulla. Now as to the point that Aldus' mulla is "otherwise inexpli- cable," it is reasonable to say that the edition of Aldus is by no means free from evident typographical errors, and that in con- sideration of the character of the ordinary close running hand of that day (or of this, for that matter), and of the small Italic type used in the Aldine edition, no mistake could be much more easy for a printer to make (even if this m was not in the n-compartment of his case), and for the proof-reader to overlook, than an m for an n. What I should think "inexplicable," and impossible, is that Aldus should deliberately adopt the unintelligible mulla for the nulla to which he changed it in his second edition ! But, apart from this, Mr. Hardy is wrong in saying that there is no dot over the i in the marginal reading. It is there, plain as day, and, though not so large as an ocean, large enough to be perfectly distinct. It is no accident ; it is not a fly-speck ; no recent enemy hath done this. To make sure on these points I examined it critically under a microscope. There is no mistake about it ; it was put there when the word was written, and the pretty coincidence that Mr. Hardy's fancy built upon does not exist. As the reading is one of interest, I must stop to explain it, premising that I believe, though not altogether agreeing with 138 Elmek Truesdell Meebill Mr. Hearne and Mr. Hardy, that Ii go back to the Paris codex discovered by Iucundus. That codex, or the copy of it, read cui nulla re in this place — wrongly, of course, for cui in nulla re, as is clearly shown by Pliny's quotation of the phrase twice over (but once with a transfer of the negative) just below, in viii. 6. 12, where the preposition incontestably stands (as it does also in viii. 6. 10 in codex M). But Aldus, following his copy, wrote, and meant to print, cui nulla re here, as he finally succeeded in doing in his second edition ; and the barbarous mulla of his first edition must have irked him. The copyist i also wrote truly from his archetype cui nulla re, but saw at once, from his grammatical knowledge, or from the two quotations of the phrase just below, that the reading certainly ought to be otherwise, and consequently he inserted in above. This I suspect to be the true explanation of the matter, rather than that the reading of the codex was doubtful. But to take up the next point, vcrrepov -rrpdrepov — the matter of punctuation : the punctuation of Aldus is very generous in commas, but is much better than that of the editions before him, which, after all, are not much worse in point of punctuation than MS that I have seen scholars trust to the enlightening labors of the type-setter nowadays. But there were brave men before Agamemnon, and doubtless good punctuators before Aldus, as after him. It is certainly not impossible, nor a great strain upon the imagination or the experience, to suppose that the owner of a book then, as he corrected the readings of his classic text to perfect it for his own use, so might also correct such misleading punctuation as that of Beroaldus and Avantius. What evidence should we find sufficient in the matter of punctuation, considered by itself, to indicate that this book was that put by Aldus before his printer? I should think that in such an uncertain matter anything short of pretty close identity of punctuation would fall short of the quality of evidence. Mr. Hardy contents himself with a general affirmation. I felt bound to count instances, taking passages here and there at random as fair samples. It is difficult to give precise figures, owing to differences of reading that might call for differ- ences in punctuation, but one or two of my results may be given On a Bodleian Copy of Pliny's Letters 139 as specimens ; and in them I have treated as cases of agreement in punctuation every instance where Aldus and Beroaldus (or Ii) put a mark in the same place, whether the marks were the same or not. This was to give Mr. Hardy's argument its best possible chance. If I had insisted on counting as agreements only instances of agreement in the mark itself as well as in the place of it, the case against Mr. Hardy's position would have been stronger. Here is the result in one long letter (vi. 16) : Punctuation of Beroaldus agreeing with Aldus, 127 times. Punctuation in Beroaldus, but none in Aldus, 11 times. Punctuation in Aldus, but none in Beroaldus, 61 times. Punctuation corrected by i agreeing with Aldus, 18 times. Punctuation corrected by i not agreeing with Aldus, 4 times. Of course it will be remembered that the uncorrected punctu- ation in the print of Beroaldus was (on Mr. Hardy's theory) as much a part of Aldus' directions to his printer as the corrected punctuation by i was ; but Mr. Hardy entirely ignores this consid- eration, and insists on counting only the agreement of i's punctu- ation with that of Aldus. But I venture to submit, on the basis of the showing above, that, though i was doubtless a better punctu- ator than Beroaldus, if Aldus, with his training, professing to prepare this printed text as copy for his printer, in punctuation as in readings, had to make over seventy corrections of punctuation in his proof of a single letter, he must have been more careless about details of "copy" than so skilled an editor is likely to be, and more haughtily oblivious to the cost of proof -corrections than most editors are likely to become. Examination of different parts of the book showed results concerning punctuation like the above. Most of the corrected punctuations of i are those any intelligent reader might make, and about three-fourths of these agree with Aldus. But it is to me impossible, in the face of the large proportion of discrepancies existing in the total punctuation between the two books, to see any evidence from punctuation that Aldus could have given this book as copy to his printer. The evidence from punctuation appears to be entirely against Mr. Hardy's conclusion, rather than for it. 140 Elmer Truesdell Merrill The final point, 'On^piKw, is concerned with the matter of the readings. Mr. Hardy remarks, as I have said, that a large majority (his figures are untrustworthy, and need not be repeated) of the marginal readings of i (he disregards those that are interlinear) appear in Aldus, and thinks this, in connection with the other evidence that I have already discussed, is sufficient to prove that this book was the printer's copy of Aldus. But, as in the case of the punctuation, it should hardly need remark that the entire text, as corrected, and not the corrections alone, formed the printer's copy. A few specimens of collation will show how far this text of Beroaldus, after correction by i, is removed from the text of Aldus; they will also be of importance for reference later in my more constructive argument. I give the references by page and line in Keil's edition of 1870; the reading before the bracket is that of Aldus, after it that of Beroaldus, as corrected by i, where- ever such corrections occur: p. 3 1 PLINII SECVNDI EPISTOLARVM LIBER PRI- MVS] C. CAECILII PLINII SECVNDI EPIST. LIBER PRIMVS 6 C. PLINIVS SECVNDVS SEPTITIO SVO SALV- TEM] C. Plin. Septicio. S. 7 accuratius] cura maiore 10 uenerat] uenerat 12 Vale] om. 14 C. PLINIVS ARRIANO SVO SAL.] C. Pli. Arriano S. p. 4 1 et legas et relegas] et legas ante] unij 2 £^] stilo 3 Caluu] om. 4 uim tantorum uirorum] uim tantam uerborum pauci quos aequas assequi possunt] 0i,fj.evoio-iv to ocriov vTov] vofiov /xev^iTov add. in lacuna i 33 occurrere] occurre This complete list of the variation within a short space of text will serve for a sample. That the text of Aldus and of Ii are on the whole so much alike here (as elsewhere in Book x) will of course occasion no surprise, when it is recalled that there is no evidence of — no probability of — the existence at the end of the fifteenth century of any other source for Book x than that MS discovered at Paris by Iucundus. An examination of even these variants in x. 4-10 will be enough to indicate that Ii was not copied from Aldus ; it might also indicate that Ii was not the printer's copy for Aldus. For on the supposition that i is his hand, why did he not correct his copy into still further agreement with his own ideas before sending it to the printer ? And is it conceivable that he would finally deliberately reject the reasonable quadragies of I for the unreasonable quadringenties, which has caused many critics to stumble? That he would reject the intelligible aduerte of I for the unintelligible conjecture of his own (as it must under this theory be accounted), adit te? That he would finally think eum scilicet a more reasonable emendation than his earlier esse eum for etsi eum ? That he would actually substitute the ungram- matical Kal. Septembris for Kal. Septembribus f The only rational On a Bodleian Copy of Pliny's Lettebs 147 answer to these questions appears to me to be in the negative. (I have omitted the striking instance of the reading locationem\ locatio cum, because of the possible doubt that the correction was made by i.) Similar questions may be asked concerning x. 41-121, where the readings of Avantius (A) are by no means all corrected by i where the Aldine edition (a) differs from them. Still less are all the important readings of i in agreement with Aldus. Under the latter head the following list will furnish some basis for judgment : heae A hae a eae i quern Aa quam i quibus ilium intelligis A qui pusillum intelligit a quibus ilium intelligi i conferre Aa conferrem i si ingrata A si enim grata a sint rata i emendatae Aa emendata i huic A hunc a hinc i mitteretur dispiciebam A mitteret ut dispicias a mitteret dis- picias i 27 putauerimus Aa si putauerimus i (26 si om. Aai) plures enim et quanto infirmiores erunt idem fiduciam A plu- res enim tanto magis eadem requirent quanto infirmiores erunt tibi earn fiduciam a plures enim et quanto infirmiores erunt idem petent fiduciam i 300. 18 senator Aa senatores i in eadem positam A in eodem opere positam a in eodem posi- tam i 304. 22 continerent Aa continebis i 307. 10 cohaerenti Aa coharentia i 308. 16 passumque uenire uictimarum cuius A passimque uenire uicti- mas quarum a pastumque uenire uictimarum cuius i 310. 3 diem quae in tutela A diem in quem tutela a diem quo in te tutela i 311. 27 iniuriam Aa in iniuriam i 312. 1 boyali et ecclesia A et bule et ecclesia a bule et ecclesia i 313. 6 scilicet Aa sed i Now these readings of i that I have quoted are supposed to be taken by Aldus from the copy of the Paris codex by Iucundus, if not from the codex itself. Moreover, they are all reasonable, 285. 6 290. 19 291. 18 21 292. 1 295. 6 296. 3 299. 15 148 Elmer Truesdell Merrill and some, at least, of them surely right which in the 1508 Aldus are surely wrong. Is it conceivable that, having once set down for printing these reasonable MS readings from Iucundus, Aldus should have changed them in his proof to the no better, and often impossible, readings that appear in his published text ? It is not sufficient to say that he may have corrected his proof-sheets from the original MS, having had access only to the copy by Iucundus up to the time when the proof was struck off; for when the entire mass of readings in i-x is reviewed, it appears impossible to believe that many of the readings of Ii that are rejected by Aldus, agreeing as they do with the MS tradition that he usually follows, did not stand in the Paris MS. They were rejected by him for other reasons, and too often in favor of elaborate and arbitrary improvements of his own. Even the list from x. 41-121, that I have just given, shows something of this tendency. And it is one thing to know that Aldus doubtless did not accept all of the excellent readings of Parisinus ; it is quite another thing to believe that he rejected so many of these excellent readings after actually putting them into type. Mr. Hardy did not hesitate to enunciate his belief that these notes of i are in the handwriting of Aldus himself, but he appa- rently did not think it necessary to apply the final test of con- fronting the one with the other, at least in photographic facsimile. My argument appears to me perfectly conclusive against Mr. Hardy's supposition without any resort of mine to the handwriting test, though my memory (which is, to be sure, very fallible) assures me that specimens of Aldus' handwriting which I have seen in Italy are widely different from the hand i. In the three great libraries of England I have been unable to find any specimens of the writing of Aldus, and at this distance from Italy I am com- pelled to let this desired comparison go. But another I have made, and shall mention further on (p. 155) . [Since writing the above I have been able to compare the writing of i with a facsimile of a letter of Aldus in the Stadtbibliothek of Mainz. As I had thought, the two hands are about as different as two hands could possibly be.] Thus I have summed up the matter concerning each individual On a Bodleian Copy of Pliny's Letters 149 part of the volume under consideration, and without finding sub- stantiated the alleged evidence that this volume was the " printer's copy" for even the "first proof" of the 1508 edition of Aldus; finding, indeed, that all the evidence is the other way. I trust, then, that the notion advanced by Mr. Hardy may be finally set down as untenable. Who, then, did write these notes, and who owned this book at that time ? I naturally thought first of Beroaldus, between whose edition of x. 41-121, published in 1502, and this corrected book, there are a number of striking points of agreement. There is no copy of this Beroaldus in the Cambridge University, Bodleian, or British Museum libraries, and I have therefore been unable at the present writing to determine the matter personally, but if the report in Keil's apparatus criticus is sufficiently complete, none of the readings of Beroaldus depend on anything more than (1) the edition of Avantius, published a few months earlier; (2) possibly the same copy from the Paris codex that Avantius used, or a copy of it (Avantius, I suspect, did not reproduce it with sufficient faithfulness), and (3) acute conjecture. Of course Beroaldus could not have had this particular book before his edi- tion of x. 41-121 was issued, or he would have printed all the letters instead of merely those that Avantius had already published. But, further than this, he could not have had a source that gave readings like these of i, or he would not have wandered so far from the mark as he seems to do in, for example, the following instances : 301. 18 in aedem positam A in eodem opere positam a in eodem posi- tam i in aede positam Ber. 303. 27 debeo quam ea quae speret instructum A debeo quam abunde ea quae speret instructum ai 1 debeo: % fabium ualentem ualde probo instructum i 2 debeo quem ad ea quae speret instructum Ber. 304. 22 continerent Aa continebis i continerem Ber. 307. 5 extruere A instruere ai excutere Ber. 315. 21 dandamque quod .... debebantur A dandumque quod .... debebantur a dandumque quod .... debebatur i dan- daque quae .... debebatur Ber. 316. 5 peregerant A perceperant ai peregerat Ber. 150 Elmer Truesdell Merrill Prom these and similar instances it is evident that Beroaldus had no other sources than Avantius and perhaps Leander's copy. He was emending on the basis of these. I naturally thought next of Catanaeus. There are also striking agreements between Catanaeus and i, and a discussion of these I must reserve for another occasion. But Catanaeus, too, printed, from Books viii and x, in his 1506 edition only the same letters as Avantius and Beroaldus before him, and so could not have had this particular book as his basis at that time ; and there is so much disagreement between him and i on the whole that the theory that he was the owner and user of the book had to be abandoned. After other early editions had been fruitlessly compared, only one man known to me seemed to remain as a possible proprietor and sponsor, Gruillaume Bude (Gruilielmus Budaeus), of Paris, the prodigiously learned reviver of Greek letters in France. His overwhelming Annotationes in XXIV. Libros Pandectarum ("doctis, Iuppiter, et laboriosis!") were first printed in 1508, the same year that saw the publication of the first Aldine edition. The Aldine edition appeared in November, and the preface to the Annotationes is dated: Parisiis pridie Nonas Nouembris. There were also no relations, so far as I know, existing between Aldus and Budaeus. It is impossible that the edition of Aldus should have been used by Budaeus in the preparation of his Annotationes, nor was there at that time any printed text that contained viii. 8. 3-18. 11, nor any that contained x. 1-40. Yet Budaeus, who quotes copiously from Pliny's Letters, quotes from these parts then unprinted and generally unknown, as well as from other parts of the Letters ("Idem libro nono ad Aristonem, sed in epistola manu scripta quae uulgo non reperitur, id et aliae non paucae;" u Plin. igitur ad prosocerum libro nono epistolarum .... Verum haec epistola et aliae non paucae in codicibus impressis non leguntur j" "Idem ad Traianum epist. 4, quae uulgo non extat;" u Traianus ad Plinium epistola septima, quae uulgo non reperitur;'''' u Plinius ad Traianum in libris inte- gris" ) . He is also aware that the printed editions and the com- mon MSS inverted the proper order of Books viii and ix ("Apud eundem ad Oeminum lib. IX., qui tamen in antiquis octauus On a Bodleian Copy of Pliny's Letters 151 inscribitur /' "ut Plinius -VIII- epist." quoting from viii. 2. 8: yet he sometimes quotes Book viii as Book ix, as, e. g., above). Codex M, as I have said, was surely unknown to him. Keil quotes the readings of Budaeus a number of times, especially in the uaria lectio on the particular letters unprinted until the issue of the first Aldine edition in 1508. But Keil does not quote the many other passages drawn by Budaeus from Pliny's Letters. I have examined them all throughout all the works of Budaeus. The list is of course too long for rehearsal here. There is no especial consistency in the manner of quotation, and apparently no great worriment about text. Budaeus was quoting Pliny for the subject-matter, and generally concerns himself with text only when the subject-matter is clearly involved. Sometimes he con- denses a passage by the omission of words to him unimportant; sometimes he adds a word for clearness; sometimes he inverts the order of two successive words, when they might about as well stand in one order as in the other; usually his readings are those of the printed texts of his day (before Aldus), and especially of Beroaldus, but very commonly with the incorpora- tion of the readings of Ii, as in this volume, and that, too, in cases where Ii and a do not agree, as in the following examples, in which i may stand for either I or i (I should perhaps warn students that in the Basel edition [1557] of the Opera Omnia of Budaeus the quotations from Pliny have often been cor- rected to read like a; only the earlier editions show the text of Budaeus) : 86. 17 Maximo a Metio i Bud. (" Plinius in epistolis ad Metium scribens") 148. 15 obiit ille a obiit ilia (obiit corr. from abiit and ilia left un- changed i) i Bud. 204. 10 inimica a mimica i Bud. (" quo in loco ex antiquis codicibus mimica legimus ;" Keil also so reads, but does not cite Budaeus) 218. 14 uideo a uideor i Bud. 220. 9 ignorationem a ignorantiam i Bud. 221. 27 absoluerit a soluit i Bud. 222. 14 prement a periment i Bud. 223. 20 debuerim an abstinere quemadmodum abstinuit is qui a debuerim quemadmodum abstinuit is qui i Bud. 152 Elmeb Truesdell Merrill 271. 1 quadringenties a quadragies i Bud. 272. 15 iatraliptae a latraliptae i Bud. 295. 3 ad eosdem acheos A ad eosdem dein ad Achaeos a ad eosdem et acheos (Achaeos Bud.) i Bud. 314. 21 concedendum iussi inuitationes (immutationes A) Aa conce- denduin iussisti inuitationes i Bud. That is, the quotations of Pliny by Budaeus are precisely such as would be expected, if he had used such a book as this, Bero- aldus for Books i and ix, with this MS supplement for viii. 8. 3- 18. 11, and Avantius for x. 41-121, with this MS supplement for x. 1-40, the whole then corrected as i has corrected this book. Budaeus certainly used, then, such a text as this, and the ulti- mate source of his hitherto unknown readings is stated by himself : " Verum haec epistola et aliae non paucae in codicibus impressis non leguntur: nos integrum ferme Plinium habemus: primum apud parrhisios repertum opera Iucundi sacerdotis: hominis antiquarii Architectique famigerati.' 1 ' 1 (Annot. in Pandectas, f. 27r. I quote here and elsewhere from my own copy of the first edition of the Annotationes, published by Ascensius, at Paris, in 1508.) Compare the verbal form of this statement with that of the note at the end of the Avantius text, quoted on p. 136 ; to this comparison we shall return briefly later. It must be conceded, then, that Budaeus used a text like this volume now under consideration, and of the same source. But what proof that he used this particular volume ? I might say that the volume is marked in the margin throughout with the same ink as i by dashes, sometimes emphasized by a single dot, above or below, or, in especially important instances, by a dot on either side (-*-), opposite certain passages; and these passages are pre- vailingly such as concern matters of antiquities, precisely those that Budaeus cites in his Annotationes in Pandectas, and in his De Asse, and other passages of similar sort. But these passages might, of course, have been noted by another student of such matters. The letter viii. 14 has in the margin opposite its begin- ning the lemma in the hand of i, "De hire senatorio," and Bu- daeus refers to it under that title in Annot. in Pand. (1508 ) , f . 45v, "Idem libro nono ad Aristonem ... . de iure senatorio loquens On a Bodleian Copy of Pliny's Letters 153 ita inquit,^ etc. This, too, and other points like it, may be mere coincidences. Another lemma is more striking: in the margin opposite x. 96. 3 there is written in i's own hand: "0 miserum te plini et perditissimum;" and in commenting on this very sen- tence from Pliny's letter Budaeus says (Z)e Asse, Lib. iv, Basel ed., 1557, Vol. II, p. 193): "O miserum te Plini,'''' etc., while just above he has called Tacitus " perditissimus' 1 ' 1 because of his views on this same matter of the Christians. This also might possibly be a mere coincidence, but is hardly likely to appear so, when all the other evidence is taken into consideration. There is also a striking verbal similarity between the note at the end of the Avantius and that in which Budaeus acknowledges his obli- gations to Iucundus; I have already quoted the two on pp. 136 and 152. There is a similar verbal resemblance between the remark of Budaeus (Annot. in Pandectas, 1508, f. 145r), "apud eundem ad geminum libro nemo : qui tamen in antiquis octauus in- scribitur," and the note prefixed by i to Book ix in the Beroaldus, "in antiquis et emendatis exemplaribus hie liber octauus inscri- bitur." But more striking yet are three readings that are inserted by the same hand, I believe, that wrote the other notes that I have ascribed to i, but in a slightly paler ink, and, I suspect, at a later time. These are as follows: 68. 5 electorumque a eclectorumque i' 2 131. 23 sestertium sedecies contuli huic quadringentorum millium a sestertium sedecies i (but following, in the different ink, the comment, "hie locus corruptus sester ■ decies con- tuli huic • CCCC ■ millia") 279. 29 sacramento militar i nondum a sacramentum ita nondum I sacramento ita nondum i 2 Now Budaeus quotes and comments on each of these readings as follows: (Comm. Ling. Graecae, Opera Omnia, Basel, 1557, vol. II, p. 923) " Vnde eclecta apud Plinium iuniorem in epistola in qua de auunculo suo loquitur Hoc intentione tot uolumina peregit eclectorumque commentarios. sic enim legi debet.'''' — (De Asse, ibid., p. 53) "Illud Plinii ad Caluisium in quinto Epist. An cui de meo sedecies contuli, ita lego. An cui de meo sester- tium decies contuli, huic • CCCC • millia paulo amplius tertiam 154 Elmer Tkuesdell Merrill partem ex aduentitio denegemf etc. — [Annot. Eeliquae in Pand., 1551, p. 37) "Plin. ad Traianum de semis inter tyrones inuentis, Ipse enim dubito ob haec maxime: quod ut iam dixerant Sacra- mento, ita nondum distributi in numeros erant. Qui locus corrupte legitur in peruulgatis exemplar ibusP Now these phrases are those that imply not that Budaeus was here taking his reading from some MS authority, but that he was emending suo Marte; they are not, I think, found elsewhere than in i 2 and in Budaeus (though Sichardus later conjectured quad- ringenta millia) ; Budaeus gives them in his works published some years after the Annot. in Pand. (1508), in which he first quotes the readings of i (the Comm, Ling. Oraecae was published in 1529, the De Asse et Partibus Eius in 1514, and the Annota- tions Reliquae in 1526) . I take it, then, that most of the readings of i were inserted by Budaeus within a short time after he came into possession of them through the agency of Iucundus; but that these three readings (and, I suspect, some others) that are shown by their writing to have been inserted at a later time, were emendations of his own, added during his later study of the text. Another reading is of significance in the same direction. In viii. 5. 3 (212. 24) a reads aduocamenta, but i, auocamenta. The writing of i is in this instance not in the paler ink that, as I think, denotes a somewhat later date, but in the same ink and manner as the large majority of the annotations. It was therefore, we must conclude, written at substantially the same time with them. This reading also Budaeus quotes, as follows: u ero ergo suspensus . . . . et dies et satietas doloris inducit .... Quo in loco in omnibus exemplaribus aduocamenta corrupte pro auocamenta legitur ■ perperam • d ■ adiecta." The phraseology here also implies that Budaeus was giving, not the reading of a MS, but a conjectural emendation of his own; and the statement occurs, not in his later works, where are mentioned the emendations written in the paler ink in the Bodleian volume, but in his early work, the Annotationes in Pandectas (ed. 1508, f. 94r). Word, ink, writ- ing, and date of occurrence in the printed work of Budaeus all combine to indicate that this emendation was jotted down by him On a Bodleian Copy of Pliny's Letters 155 in his copy of Pliny's Letters at about the same time that he entered therein the mass of corrections from the MS discovered by Iucundus, that is, while he was yet preparing his Annota- tiones. The reading in codex M is also auocamenta, but, as I have already said, Budaeus could not have known that MS. It remains to show by comparison of handwritings that i is really the script of Budaeus. M. Henri Omont has published a list of the MSS now extant that are known to have been in the possession of Budaeus (in Georges Hermonyme de Sparte .... suivi cPune notice sur les collections de manuscrits de Jean et Guillaume Bude~, Extrait des Memoires et du Bulletin de la Societe de l'Histoire de Paris et de l'He-de-France, t. XII., 1885). Two of these MSS are in the Cambridge University Library, and of these one ("Kk. VI. 23. 2") contains glosses in what M. Omont declares to be Budaeus' own hand. I confronted these glosses with photographs from the Bodleian volume while that was also fresh in my mind's eye. The glosses are but seven in number, and are all in Greek ; but the hand is in size, in ink, in every detail, unmis- takably the same hand that wrote the Greek glosses in the Bodleian volume, which was, therefore, the hand of Budaeus. It is of course impossible to prove from this that the Latin hand i was also that of Budaeus, though the ink is identical, and considerations of spacing sometimes indicate also that both Latin and Greek glosses were written at the same time. I hope to be able later to confront the unquestionable Latin hand of Budaeus with the photographs of the Latin hand i; but my argument seems to me so conclusive as not to need further confirmation from the handwriting test. According to M. Omont (Zoc. cit.), the books of Budaeus were sold after his death in 1540, and came into the possession of President Francois de Saint-Andre, who bequeathed his library to the Jesuits of the College de Clermont. When the Jesuits were expelled from France, in 1595, these books were dispersed. The later history of the Bodleian volume, until it was bought by Thomas Hearne at auction in Oxford in 1708, 1 have been unable to trace. As to the two Cambridge MSS of Budaeus, they came into the possession of the University Library in 1715 by the gift of King George I, with the rest of the library of the deceased 156 Elmer Tbubsdell Merrill Bishop Moore of Ely. One of them ("Kk. VI. 23") has on the inside of the front cover the name J. B. Hautin, which stands also on other MSS in the collection. On "Kk. VI. 24" Hautin has written, "ea; bibliotheca Manteliana, Parisiis." This may pos- sibly furnish a clue to the fortunes of the Bodleian volume. Suppose that this Bodleian volume were, as Mr. Hardy believed it, the printer's copy for the 1508 edition of Aldus. It might then be held to outrank the text of Aldus as a witness to the readings of the lost Parisinus, if only we could be sure that the final read- ings of Aldus were not due to a correction of this combination of excerpts and transcripts by reference to the original MS. But we could not be sure of that; and therefore the book must have had much less certain value than Mr. Hardy imagined. As the case actually stands, however, it has far more value than Mr. Hardy imagined, even if not as much as Mr. Hearne believed (see his MS notes in the Bodleian copies of his edition of 1703). For Budaeus doubtless did not depend on Griocondo's copy of the Paris codex, but had the codex itself before him. This handexemplar of Budaeus, is therefore, so far as the corrections and MS supple- ments go, and the few emendations of Budaeus himself, and of others, can be sifted out from among them, a more immediate witness to the readings of the unique and ancient Parisinus, and a more trustworthy witness, than the "printer's copy" of Aldus could possibly have been. It remains to discuss, in the light of this newly gathered evidence, what the character of that lost Parisinus was, and what the trustworthiness of the first edition of Aldus as a witness to the original text of the Letters. This I hope to essay on another occasion. Tkinity College, Connecticut August 1, 1906