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. THE MANUSCRIPTS OF PROPERTIUS By B. L. Ullman Propertius has been a very unfortunate poet in numerous ways. Many of his readers have failed to understand either the man or his poetry. He has been acclaimed one of the greatest poetic geniuses of all time, and he has received that most stinging of all criticisms, "the silence." Controversies of various sorts concerning him and his work have been among the fiercest and most prolonged, even if not the most important, in the field of classical philology. After a century's debate, there are still two factions on the question of the division of his poems into four or five books. After more than four centuries of contention about his birthplace, rivaling the ancient controversy over Homer, there seems at last to be a majority agree- ment. But it is the text of the poems with which we are concerned in this paper. Propertius has had no Ellis, as had Catullus, to pub- lish readings from a large number of MSS so that scholars could form their opinions about them; no Hiller, as had Tibullus, or Schwabe, as had Catullus, to offer a conservative yet progressive text until the MS problem could finally be settled. The very latest history of Latin literature, the revision of Teuffel by Kroll and Skutsch, 1 contains this statement: "Die einzige kaum verfalschte Ueberlieferung bietet der cod. Neapolitanus, jetzt in Wolfenbuttel, Gud. 224, s. XII/XIII? .... Die ubrigen Hss entstammen der Humanistenzeit und sind von Baehrens uber- schatzt worden." I have no quarrel with those who consider the Neapolitanus the best MS of all, but it is not true that all the other MSS belong to the humanistic period, in the ordinary sense of that term: there is one other (incomplete) MS in existence antedating the earliest humanist. The opinion expressed in Teuffel is but an index of the uncertainty which is prevalent concerning the MSS of Propertius. This attitude is largely due to the work of Baehrens, whose critical edition (1880), though long out of print, has not yet • Vol. II (1910), p. 92. [Classical Philology VI, July, 1911] 282 The Manusckipts of Pbopeetius 283 found a successor. His judgment concerning the age and value of N, the oldest MS, has not been accepted by most scholars, and for that reason some of the other MSS which he used have been under- valued. He himself hurt his own case by publishing a hasty collation of F, a difficult MS to collate on account of its numerous corrections by later hands. In this paper I shall confine myself to a discussion of the following points : the history of the oldest MSS, the relation of A and F, and the views presented in two recent articles on the MSS of Propertius. Perhaps it is well to state on what material I, base my conclusions. It consists of photographs, practically complete, of N A F, a collation of F made from the MS itself, and scattered readings from numerous other MSS. After an examination of a large number of MSS, per- haps a hundred or more, I find that none antedates the fifteenth century except the three just mentioned. I hope to examine care- fully most of the existing MSS, gathering material for what may some day become a critical edition. The collation of a number of MSS of Tibullus for a similar purpose has delayed the work on Propertius. The oldest MS of Propertius known today is Wolfenbuttel 224 Gud., called N(eapolitanus) because it was once in the church of S. Giovanni a Carbonara at Naples. A controversy long raged concerning the age of N. Baehrens and others thought that it was a contaminated MS of the fifteenth century, but the judgment of paleographers is now unanimous in favor of the late twelfth or the early thirteenth century. It may be interesting to note that Traube in 1906 expressed to me an opinion in entire accord with this. James 1 stated that he was confident that the manuscript was written in the latter part of the twelfth century and that it was "cer- tainly not of Italian origin." He likewise found on a fly-leaf the word Goeric, the name of a saint whose cult was localized around Metz (where he was bishop in the seventh century), Toul, Verdun, and Epinal (where he was buried). James's conclusion that N was once in some library in or near Metz is highly probable, in fact, almost certain. One point of interest in this connection has not been noticed in N, as far as I know. The only heading in N is "Incipit Propertius" (italics indicate abbreviation) by an early i Class. Rev. XVII (1903), p. 462. 284 B. L. Ullman hand, though not by either of the text hands. But the early hand wrote "Pertius" and a later hand added the sign for "pro" in the ample space left. The same hand also worked over the curved stroke of the following "p," to make it more like his own. Likewise he worked over the first "i" of "incipit" into a capital letter. This hand was very likely that of an Italian humanist. Hence we have a right to expect that a Propertius is hidden under the name of the well-known Persius in some mediaeval catalogue. It is even possible that a Propertius MS is today masquerading as a Persius in some library. There was a Persius at Verdun in 1070 and another at Schlettstadt (not far from Epinal, near the Rhine) in 1296. 1 At this point we may leave N temporarily. The next oldest MS is Leyden, Voss. Lat. in Oct. 38, known as A. As to its age and origin James 2 says that it is "certainly not Italian in origin, but most probably French. I should place it early in the fourteenth century: I could even conceive it to have been written in the last years of the thirteenth." E. Maunde Thompson has stated 3 that "the Vossianus was written about 1300 and just as likely before it as after it." I may add that I agree entirely with James that the MS was written in France, as I happen to be somewhat familiar with that character- istic script. Now about the year 1250 Richard de Fournival of Amiens evolved an ideal classification system for a library and entered therein titles taken directly from the MSS. He includes "Propertii Aurelii Naute liber monobiblos." 4 The heading of A is "Incipit monobiblos propertii aurelii naute." I hope to show later that all the MSS having this heading are derived from A. It is tempting to suppose that A, a MS written in France, probably in the thirteenth century, is the one referred to by Richard. I see no objection to this hypothesis, for I think that A may have been written as early as 1250. If A is not the MS mentioned by Richard it is very likely a copy of that MS. It has been suggested that Richard did not own 1 Manitius Philologisches aus alien Bibliothekskatalogen, Rhein. Mus. 47 (Ergan- zungsheft), 1892, pp. 53-54. Thinking that the Verdun catalogue might have been misread, I examined it, but it has Persius (not Pertius). I have not, however, examined the Schlettstadt catalogue. 2 hoc. cit. » Class. Rev. IX (1895), p. 184, n. * Manitius op. cit., p. 31. The Manuscripts of Propebtius 285 all the MSS he names, but copied the titles of books in various libraries, 1 so that it is not certain that the Propertius was ever at Amiens. On the bottom of the first page of A is the name P. Petauius. Paul Petau lived at Orleans and Paris in the sixteenth and seven- teenth centuries. He and Bongars bought a number of MSS from Pierre Daniel of Orleans, who in turn had got them from the monas- tery of St. Benedict at Fleury. 2 It is altogether probable, then, that A came from Fleury, whence came also the famous Berne MSS of Horace and Valerius Maximus, though it is possible that it is not one of the MSS that Petau bought from Pierre Daniel. Petau's MSS went to his son Alexander, who sold them to Vossius. From the latter they passed into the possession of the University Library at Leyden. The MS now ends with ii. 1. 63, at the end of the second quaternio, but was once complete, as is shown by the fact that the "signature" (the first two words of the next quaternio) is found on the bottom of the last page. There is no trace of Propertius MSS in Italy during the Middle Ages. The humanists of the first half of the fourteenth century, too, with the sole exception of Petrarch, were unacquainted with our author. The case of Catullus is strikingly different; his poems, a copy of which Petrarch obtained in Verona, were known to Bencius Alexandrinus, Hieremias de Montagnone, Guglielmo da Pastrengo, the author of the Verona Flores, Benvenuto de Campe- sanis, and others, but Petrarch alone knew Propertius. It is alto- gether likely, then, that Petrarch obtained his copy outside of Italy. Since he spent many years in France, one naturally thinks first of that country as the source of his MS. A was in France at that time : did Petrarch obtain a copy of A ? In order to answer this question we must see what evidence there s for identifying Petrarch's copy. De Nolhac made the entirely unwarranted suggestion that N was owned by Petrarch, 3 and this suggestion has often been repeated. So for example, Sandys 4 1 Manitiua op. tit., p. 2. ' Cf., e.g., Valerius Maximus, ed. Kempf (1854), p. 80. 3 Most recently in Pitrarque et V Humanism*, ed. 2, I, 172. * History of Classical Scholarship I, 621. 286 B. L. Ullman speaks of N as a " MS known to Petrarch." It is clear that Petrarch's exemplar belonged to the A F group and it has been generally con- ceded that it was the archetype of F — which surely was not N. In the margin of his Virgil Petrarch comments on Servius' note in regard to the shortened vowel in words like caue, uale as follows: "Sic utitur Propertius 1. 2° e. 11, et sepe alibi"; he refers to ii. 13, 14 cauS. According to the division of poems in F, this comes in the eleventh poem, as Petrarch has it, but in N it comes in the tenth. Therefore Petrarch was not using N. Furthermore A shows a mar- ginal numbering of poems by an early hand, and F has a few indica- tions of such a condition in its archetype. 1 There is another kind of evidence. Most of Petrarch's MSS passed into the Visconti library at Pa via, an inventory of which, made in 1426, is still pre- served. In it we find the notice: 2 "Monobiblos Propersii Aurelii Naute uolumen paruum in assidibus et fundo corii albi in carminibus. Incipit Cinthia prima suis, et finitur ossa uehuntur aquis." If this were a Virgil or a Cicero, we could not be so sure that it was Petrarch's MS, but considering the great scarcity of Propertius MSS at that time we can confidently assume that it was his. The title agrees with that of A F (except for the misspelling propersii). The read- ings uehuntur aquis are found together only in F, of the MSS reported by Baehrens.* The history of the MS cannot be traced far after 1426. In a catalogue of the Pavia library made in 1459 appears a Propertius, undoubtedly our MS. It disappeared between 1459 and 1518, for in the latter year an inventory of the library after its removal to Blois was compiled, and our MS does not appear therein. 4 Lastly, we can gain an idea of Petrarch's Propertius from another source. In 1375 Coluccio Salutati writes to Caspar of Verona "Si prece uel precio Propertium de bibliotheca illius celeberrimi uiri, 1 iii. 2, title Ad librum secundum (sc. capitulum). iii. 3, title Tertium (sc. capitulum) somnium property, iii. 4, title De triumpho quartum capitulum. iii. 11, title De imperijs feminarum contra (contra is a corruption of 9", exactly the form regularly found in A). * D'Adda Indagini storiche .... sulla Libreria Visconteo-Sforzesca del Castello di Pavia (1875), No. 48. I have examined the original MS catalogue at the Brera in Milan and found d' Adda's report correct. 3 They occur also in Postgate's L. 4 The Blois catalogue is now published in Omont Anciens inventaires et catalogue* de la bibliotheque nationale, t. 1, 1908. I consulted the original MS at Vienna. The Manuscripts of Peopektius 287 Petrarce inquam, .... haberi posse confidis, uel ut meus sit, uel ut exemplari queat, deprecor ut procures." He evidently did not get it at once, for in 1379 he writes to Lombardo a Serico, who had charge of Petrarch's library after the latter's death : " De Propertio autem nihil aliud, nisi quod scribis expecto, ut uidelicet exemplatum habeam." 1 From this letter we see that Lombardo was intending to send a copy to Coluccio. As Coluccio quotes Propertius in 1381, we may assume that Lombardo did not delay very long. In 1392-94 Coluccio quotes Propertius again. Now, F (Florence, Laurentianus 36. 49) is the MS that he owned (he signed his name in it), and it agrees in text with these quotations. 2 Yet it is generally spoken of as being written in the fifteenth century ever since Baehrens assigned it to that period. A curious and interesting paper could be written on the misstatements which have been made about this MS. But the one that takes the palm is the following by Plessis: 3 "C'est pourquoi je suis 6galement porte" a croire que le Laurentianus, qu'il (Baehrens) dit 6crit tout a fait au debut du quinzieme siecle, pourrait bien, lui aussi, etre d'une epoque un peu plus r^cente, et se rapprocher plut6t de 1450 que de 1400." Immediately afterward Plessis mentions the autograph of Coluccio appearing in the MS. Though Coluccio died in 1406, the MS in which his autograph appears was written nearer 1450 than 1400! But I have discovered decisive proof that F is the MS sent to Coluccio by Lombardo. In the letter to Lombardo just mentioned, Coluccio speaks of a MS of Petrarch's "de viris illustribus" which Lombardo is having copied for him. The MS which was sent to him is now in the Vatican, Ottob. 1883. Most of it was copied by a scribe, but Lombardo's own supplement to Petrarch's unfinished work was written by himself and he also served as technical corrector throughout. At the end is the date : ' ' transcriptus MCCCLXXX die XV nouenbris," followed by the autograph signature of Coluccio. On examining this MS in 1907 for quite another purpose, I thought I saw a resemblance between the hand of Lombardo and one of the i De Nolhac Petrarque I, 89, 90. • Sabbadini also pointed this out, but he has been the only one. Studi ital. di fit. class. VII (1899), 106. 'Etudes critiques sur Properce (1884), p. 23. 288 B. L. Ullman marginal hands in F, which I had examined in the autumn of 1906. In the summer of 1910 I had an opportunity of comparing photo- graphs of Ottob. 1883 with F itself, in Florence, and found not only that the second hand in F was that of Lombardo but also that the copyist of F was the very scribe who copied the most of Ottob. 1883. In this identification I am supported by the most excellent authority of Professor Rostagno of the Laurentian Library. The third hand in both MSS is that of Coluccio. The Propertius, then, was written between 1379 and 1381, probably in 1380. These identifications are of considerable importance for the text. Baehrens lumped the work of Lombardo, Coluccio, and one or more correctors of the fifteenth century under the one term F 2 . It is especially important to dis- tinguish the work of Lombardo, for his changes are those of a proof- reader correcting from the original. Undoubtedly Coluccio never saw the original, for it went from Padua to Pavia, as we saw above. It will, however, be interesting to examine Coluccio's readings. I had intended to publish a new collation of F (considering only F 1 , F 2 , F 3 ) in connection with this paper, but lack of time makes it necessary to postpone this until a later date, though I trust it will not be long deferred. To come back to our former question: Was Petrarch's MS a copy of A ? The history of the MSS themselves offers no objection to an affirmative answer. A was probably in Fleury, Orleans, Paris, or Amiens in Petrarch's time. We know that he was in Paris, and it is quite probable that he passed through Fleury and Orleans on his way to that city. He is also likely to have touched Amiens on his way from Paris to Belgium in 1333. 1 On these journeys he was ever on the alert for MSS, and whenever he saw a monastery that looked old (such a one as that at Fleury) he turned aside to investigate the library. 2 Since we have shown that Petrarch's MS was the archetype of F it will be necessary to show by internal evidence that F really is a granddaughter of A before we can state that Petrarch's copy was taken from A. That such is actually the relation between A and F is what I believe. But before discussing this question let us look into some other matters. ' De Nolhao PUrarque I, 38. * Ibid., p. 40. The Manuscripts of Peopertius 289 Did A have any influence on the text-tradition except through Petrarch's copy ? It does not seem likely that it did. Since it was written in France, and since presumably it was found there by Paul Petau, it is scarcely likely that it was ever out of the coun- try before Petau's time. Nor is it likely that humanists after Petrarch's time copied it, for their own copies (descended from Petrarch's) would be too similar to A to make it worth while, even if they did chance to find it. Unless evidence to the contrary be shown, it is therefore likely that A was not utilized after Petrarch's time. A MS of interest for the history of the text of Propertius is Paris 7989 (which we will call P), containing, in addition to Tibullus, Catullus, and other writings, the only MS text of Petronius' Cena Trimalchionis. The MS has the date November 20, 1423, after the Catullus. Since Poggio in his letters mentions two different MSS of Petronius at about this time, it is very likely that he had the Cena, and that the Paris MS is to be closely connected with him or Niccoli. 1 One thing has not been pointed out: the MS was evidently written by a scholar, not an ordinary scribe. This is shown by the nature of the corrections, variants, and notes by the first hand. I have not yet been able to identify the hand, although I have enlisted the aid of Professor Sabbadini, who is the court of last appeal on such matters. The Propertius is closely related to F, but later corrections and vari- ants were added by the first hand, apparently from N. For example : iii. 11. 35, Hec tibi F, Res tibi L, Tres ubi N(DV), Re tibi P, i. m. Tres ubi P 1 . Now the presence of the Cena, as said above, closely associates this MS with Poggio or Niccoli. It is tempting to suppose that the presence of readings from N means that Poggio brought N to Italy from Germany. That it was in Italy in the fifteenth century is certain, for it contains writing by fifteenth-century Italian hands. 2 In May, 1423, a few months before the copyist of P entered the date in his MS, Poggio returned to Rome from England by way of Ger- many and the Rhine. We have seen that N was in some town on or » A. C. Clark Class. Rev. XXII (1908), 178. Both my friend Dr. E. T. Sage and I had arrived at the same conclusion several years before Clark's article appeared. 2 Cf. James op. cit. 290 B. L. Ullman near the Rhine. Why may Poggio not have found N, just as he found a MS of Petronius at Cologne ? 1 In 1427 Poggio writes to Niccoli that he cannot send him his Propertius (which Niccoli had requested) because some person — who it was he could not remember — had borrowed it. Two weeks later he writes briefly that he has recovered the MS and that he will send it on at once. 2 Why did Niccoli want Poggio's Propertius? Coluccio's copy (F) was accessible to him in the library of Cosimo de' Medici (whose name appears in F), whose every book was as accessible to him as his own, and it had probably passed through his hands before reaching Cosimo. Petrarch's MS was in Pavia, as we saw above. It is possible that he may have wanted Poggio's MS merely for the latter's emendations, but it is much more likely that it had come from an entirely different source from the one to which he had access. Niccoli was constantly pestering Poggio, while the latter was abroad, to look for MSS, and if Poggio had told him that he had brought a MS from Germany he would be sure to want it. Old and foreign MSS were Niccoli's chief delight. In 1484 N seems (still?) to be at Rome, in the possession of Bernardino Valla, for in that year Politian saw a "codex vetustus" of Propertius in Valla's possession which had the reading ocno in iv. 3. 21. Only N has that reading (Baehrens wrongly reports F: it has oeno). Sabbadini has shown 3 that Politian was a good judge of MSS and did not use the terms vetustus, antiquissimus, etc., as recklessly as most of his contemporaries. It seems altogether likely, therefore, that Valla's codex was N. The claim has been made that N once belonged to the humanist Giannozzo Manetti. That claim is based on certain scarcely legible letters on a fly-leaf of N which have been variously read as Manetti, Mometti, Momett - , Minetti, etc. An argument against this ownership is the fact that most of Manetti's books found their way to Padua, thence to the 'For the time being, I am disregarding the possibility, suggested by Postgate (Trans. Cambridge Phil. Soc. IV), that L (dated 1421) was copied from a MS in which readings from N had been entered. ! Poggii Epistularum lib. Ill, ep. XII, XIII, ed. Tonelli. As Tonelli's book is not accessible to me I quote from Hochart De V authenticity des annates et des histoires de Tacite (1889) 291-92. 8 Le Scoperte dei codici latini e greci ne' secoli XIV e XV, p. 169. The Manuscripts of Pbopertius 291 Palatine library in Heidelberg, and finally to the Vatican at Rome together with the rest of the Palatine MSS. A consideration of the history of the MSS, as outlined above, in connection with internal evidence, leads me to believe that all our MSS are descended from A and N, directly or indirectly; that only Petrarch's lost MS was copied from A; that this lost MS was the archetype of F and other MSS, and that F was the archetype of many more. It is very likely in itself that Coluccio's MS should, through Niccoli and the Florentine circle, become the ancestor of a long line of MSS. Petrarch's MS, too, would be copied by, or for, the group of scholars at Milan and Pavia. When N was brought down from Germany it was perhaps copied to some extent, but more often, apparently, its new readings were entered on the descendants of Petrarch's MS and F. We have seen that the external history of the MSS makes it possible — perhaps probable — that Petrarch's MS was copied from A, and that, therefore, F is a granddaughter of A. Let us now con- sider what light the texts of the MSS throw on this point. Ever since Baehrens published his collation of A and F, the close relation existing between these two MSS has been recognized, though, as far as I am aware, they have always been regarded as of independent authority. As A is the older it cannot be descended from F. F cannot be a direct copy of A, as will be seen from the readings quoted below. There are a large number of readings in F which are corruptions of those in A, and some of them are due to a script like that of A, but that is not decisive proof. On the other hand there are certain corruptions in A which do not appear in F. The number is not so large as one would imagine from Baehrens' critical apparatus, because many of the readings there reported as being in F are really due to later hands, while the original text agrees with A. The following are indications of the descent of F from A. In A the book-heading and title, in red ink by the first hand, is "Incipit monobiblos propertii aurelii naute ad tullum," and a second hand (not much later) added in the margin the words "heroys prima" (in black ink). This hand wrote notes in the margin in various other places, and none of them has the appearance of being drawn from the 292 B. L. Ullman archetype of A. 1 F includes the words " heroys prima" in the head- ing by the first hand. The marginal hand in A likewise numbered the poems. F shows traces of this numbering, as we saw above. The following cases show how A was misunderstood by the arche- type of F: a. i. 3. 33: ocellos A N edd., but in A the word can easily be taken for ocellis, as Baehrens wrongly took it; ocellis F. b. i. 5. 3: A has a word probably meant for meos, but it is very difficult to make out; F omits, leaving a space. c. i. 6. 27: Mult A Multi N edd. Multum F (an attempt to make sense out of the queer form Mult). d. i. 9. 11: munermi A; F omits, leaving a space. The copyist of the archetype of F could not bring himself to copy the senseless munermi. e. i. 9. 19: tigres AN edd., but in A the word is easily taken for tigies; tigies F al'tigres F 2 . /. i. 13. 9 : illar ( = rum) A illarum N edd. Maris F. g. i. 13. 13: aug 2 e ( = augure) A augure N edd. augere F. h. i. 14. 19: arabium A N edd., but in A the word can easily be taken for arabtum (as Baehrens wrongly took it) ; arabtum F. i. i. 20. 21: houm ( = heroum) A heroum N edd. hominum F. j. ii. 1. 31: c (=cum) A cum N edd. est F. k. ii. 1. 53: circeo A N edd., but in A the word closely resembles cureo; cureo F. Particularly convincing are examples a, e, h, k. Examples /, g, i, j show errors due to misreading of abbreviations uncommon in northern Italy in the fourteenth century. Still more convincing is the following formidable list, based, it will be remembered, on a re-examination of A and F, and on photo- graphs of these MSS which are before me at this moment. In all of these cases A 1 has corrected an error of his, but so carelessly (generally without erasure) that, without close scrutiny, one would fail to notice the correction at all, or else would mistake the scribe's intention. In fact, in examples m, p, q Baehrens reports the original reading for A, not having noticed the correction. In q he adds "o uix ab e discernenda." I submitted photographs of several of these 1 Some of the marginalia are: i. 3. 37: " consumpsti pro consumpsisti " ; i. 3. 43: "Femina cum sola, male cogitat ut dicit seneca in prouerbiis"; i. 22. 3: "perusium mater ciuitas propertii"; i. 22. 10: "nota unde propertius. Umbria est tertia pars lombardie montuose." The Manuscripts of Propebtius 293 passages to a person unacquainted with the relations of the MSS or their readings, and was confirmed in my interpretation of A's cor- rections. 1 I. i. 1. 19: deducto A F deducte A 1 N edd. m. i. 6. 19: securas A F secures A 1 N edd. n. i. 13. 10: uoces A F uices A 1 N edd. o. i. 20. 39: descerpens A 1 (dot faint), descerpens F decerpens N edd. p. ii. 1. 41: pueniunt A F pueniunt A 1 conwniunt N edd. q. ii. 1. 60: philliridos A phillirydos F phillirides A 1 N edd. For the change in example p, which at first glance seems to consist of the addition of a line almost parallel to the down-stroke of the p, one should compare a similar correction in cognita i. 7. 14. The abbreviation for con in this MS has far less sweep to the left than in most MSS (cf. e.g. in i. 1. 21). Let no one think that the first read- ings of A were in its archetype, and that F is independently descended from that archetype; for there is this unanswerable objection: A has been corrected by the first hand in a great many places, but only where the correction has been so poorly made as to deceive even a modern reader does F have the uncorrected reading. Very inter- esting also are the following cases: r. i. 4. 10: etas ( ?) A e = at A 1 , but can now easily be taken for erat; eat N edd. erat F. s. i. 15. 30: uicis A uices A 1 N edd., but, as corrected in A, the word looks very much like was; uias F. t. i. 21 : No division A F ^ in margin A 1 new poem N edd. u. ii. 1. 56: funera funera A F; in A a deleting line was run through the second funera by a later hand. Either the archetype of F overlooked this line, or else it was inserted in A after F's archetype was copied. The following curious readings also support my contention. A 1 changes correct readings to incorrect. In one case, however, the supposedly incorrect reading is supported by N and, therefore, should probably not be included in the list (example y). The only plausible explanation for these cases is that A's archetype had the 1 Another example should probably be added to this list. I have no photograph of A showing i. 17. 19, at which line Baehrens reports peperissent AF sepelissent A» (N edd.). Probably the original writing is still very clear. The fact that Baehrens noticed it points to that, for he fails to report most of the original readings where A has been corrected. 294 B. L. Ullman readings to which A corrected, and that A unconsciously emended as he copied, and afterward conscientiously "corrected." The most mechanical scribe might unconsciously emend timporibus to temporibus (example w), but the other changes presuppose that the copyist was following the sense of the passage and was perhaps some- what of a scholar (Richard de Fournival himself ? The mere fact that a work so little known at that time should be copied at all betokens intelligent interest.). All these emendations could very easily be made by an intelligent reader. v. i. 1. 7: defic ano A deficit annis A 1 F deficit anno N edd. w. i. 3. 22: temporibus A 1 timporibus F temporibus N edd. x. i. 20. 39: ungui A N F ungui A 1 . y. ii. 1. 47: uni A Baehrens uno A 1 N F. Let us not forget to consider the evidence of the opposition. I will say at once that I expect to use even this evidence to support my case. Differences in spelling may be ignored. Anyone who has become acquainted with the ways of humanistic MSS (such as F) knows that most arguments based on orthography are unsound; that a scribe ordinarily would no more imitate the spelling of his archetype than he would its script. The dropping or adding of h, the interchange of i and y, etc., are due to the caprice or training of the copyist. In addition to orthographical differences easily recognized as such, I believe that we may also classify the following readings under that head: i. 1. 26: Qerite A Querite F N; i. 4. 28: quiquam A quicquam F N; i. 11. 26, ii. 1. 15: quiquid A quicquid F N; i. 12. 19: dissistere A N, desistere F. The list of passages in which F is apparently superior to A is not nearly so long as one would infer from Baehrens' apparatus. 1 a. i. 2. 15: leucippis F N edd. leucupis A. &. i. 3. 15: Subiecto F N edd. Cubiecto A. c. i. 5. 27: tupotero F turn potero F 2 N edd. cum p. A. d. i. 6. 4: ( = gwod?) A. s. i. 17. 4: q,FN (edd.) gwe A. i. i. 18. 15: furor F N edd. fluor A. m. i. 20. 17: pagasse F Pagasae edd. pegase A N. ». i. 20. 24: sepositi F N edd. se posiii A. w. ii. 1. 7: sparsos F N edd. sj)arsos A. a;, ii. 1. 62: patriis F N patruis A. Let us examine this long list. Example I should be canceled because i. 12. 9 begins a new quaternio and the "signature" at the bottom of the preceding page has the correct fuimus. The arche- type of F could have taken his reading from this source. Example w is of no consequence. Even an ignorant scribe would scarcely have reproduced the reading of A. Example h should surely be stricken out, for a scribe would be very apt to substitute the more familiar nautas for the rare nauitas. Examples /, i, prove little by themselves; the right readings could be due to easy mistakes. The authority of A N shows that no value is to be attached to the reading of F in example u. We could cut down the list somewhat more, but there will remain a few readings that can be satisfactorily accounted for in only one way — unless we abandon our position that F is a granddaughter of A, and the evidence presented above will not permit us to do that. The way out of the difficulty is to assume that the correct readings were the result of emendation in the arche- type of F. In every one of the examples quoted it is quite easy to arrive at the reading of N by emending A. When we remember that the archetype of F belonged to Petrarch, the first of the manu- 296 B. L. Ullman script-emending humanists, this assumption becomes most probable. The following facts also tend to show that he emended his Propertius (in margin or text) : Lombardo a Serico was the technical corrector (F 2 ) of F. There is good reason to believe that everything written by him in F is derived from the archetype. He adds comments which no doubt originated with Petrarch; e.g., at iv. 5. 53: "proprie ais, in exitium propera callida lena." F 2 also adds variants which seem to be the result of emendation. So i. 21. 6 ruis A F tuis N edd. al' tuis F 2 . i. 12. 19 is interesting : Mi neque A N edd. Minimeque F al' me neque F 2 . The reading of F 2 may be explained as a mistake, but a far more satisfactory explanation is that F's archetype had Mi neque and Petrarch wrote me above (a little to the right, out of the way of the capital M) as an emendation for mi. F took the whole thing for minimeque, and F 2 entered Petrarch's reading, as he usually did. Example u above is perhaps to be similarly explained : Petrarch's MS had pegase and he wrote pro (in abbreviation) a above, rather carelessly perhaps, and this was interpreted by F as pera to take the place of pe. From Petrarch's remark about Pro- pertius, written in the margin of his Virgil (quoted above, p. 286), we see that he scanned his Propertius, at least in part. If he scanned he would be apt to emend. Furthermore, in this same Virgil Petrarch quotes iii. 9. 1. We might expect him to have read rather carefully, and perhaps emended, a poem from which he quotes. There is a trace of his work in the perplexing v. 8. N F have iugo, F 1 (sic) has I'rogo in the margin. Rogo (from rogus) seems to be an emenda- tion to fit flamma. From the list given above we notice that most of the cases in which F has a superior reading are confined to a few poems: i. 8, 12, and 15. Now Petrarch did not count Propertius among his favorite authors; probably, therefore, he read the elegies hastily, stopping only where he was particularly interested — as he himself tells us he did in the case of certain other authors. 1 This accounts for the many emenda- tions in poems 8, 12, and 15 of the first book. In addition to examples j and k there are two other probable emendations of Petrarch's in poem 12: me neque of vs. 19, which we have just discussed, and in the same verse, the reading desistere, which was probably intro ' De Nolhac Petrarque I, 188. The Manuscripts of Propertitjs 297 duced by Petrarch, even if it is only an orthographical variation, as stated above. There is other evidence that Petrarch paid special attention to the poems mentioned. In six cases A repeats a word (in a somewhat different form in two of these cases). 1 In only three of these cases does F follow A; two of the three cases in which F has the correct (single) form are in poem 8, the other is in poem 15. In the list given, Petrarch's emendation generally agrees with N. There is no significance in this, of course, as the list is made up of passages in which F agrees with the commonly accepted text based largely on N. But in i. 8. 17 and i. 12. 19 Petrarch's emendations (I think enough evidence has been adduced to show that they were his) disagree with N. No doubt many of the readings of F differing from A are due to Petrarch's attempts at emendation. Apparently, then, F is of no value for those portions of the text which A still offers us. The remaining problems, which I hope to take up later, are to restore Petrarch's MS from its existing descend- ants, especially F, in order to supply a substitute for the missing part of A, and, finally, to decide on the comparative value of A (with its substitute) and N. Two recent articles raise several interesting questions. The first of these, by Th. Simar, deals with the MSS of Propertius in the Vatican. 2 His method is a highly commendable one: he aims to connect the Vatican MSS with various humanists on the basis of handwriting and other indications. But in the few points in which I have tested his work I have found it unreliable to such an extraordi- nary degree that I am forced to refuse to accept any of his results until they can be verified. Simar asserts that the sign n "est 1'indice le plus sur de la prove- A nance napolitaine," decides that a number of MSS containing it were copied in Naples, and concludes that they all are descended from Vat. 3273, which also contains the sign and which, on the authority of Fulvio Orsini, was written by Panormita. But I doubt whether this sign is confined to Neapolitan MSS. Furthermore the authority >i. 5. 6; 7.18; 8.7; 8.37; 15.31; ii. 1. 56. » Th. Simar Les MSS du Properce du Vatican, Musee Beige XIII (1909), 80. 298 B. L. Ullman of Fulvio Orsini is now largely discredited, and it is very doubtful if Panormita wrote Vat. 3273. A remarkable statement of Simar's is the following: "Un des sigles les plus frequents, ce sont les 3 points en triangle .-. qui indiquent un scribe napolitain." I need give no proofs of the falsity of this statement; anyone who has handled MSS knows that the sign in question is common in all kinds of fifteenth century MSS. On the basis of other signs he attributes a number of MSS to Pomponius Leto and his circle. 1 But it is his assertions about Vat. 1612 and Barb. viii. 58 that have made me skeptical about all his arguments. I have examined these MSS and found his statements entirely unwarranted. Of Barb, viii. 58 he says: "Par hasard, j'ai pu identifier l'ecriture du scribe qui est, je pense, le poete Joannes Jovianus Pontanus." Vat. 1612 he maintains is by the same hand. How any person, with or without a knowledge of paleography, could make such a statement, if he had the two MSS before him at the same time, I cannot understand. I can only believe that Simar was merely trusting to his memory. A comparison with Soldati's 2 facsimile of Pontano's handwriting clearly shows that neither Barb. viii. 58 nor Vat. 1612 was written by Pontano. The reasons that led Simar astray in regard to Vat. 1612 are these. The MS has the following colophon: "Scripsi ego [Gaspar Manius Romanus] et ultro et jocunde Anno MCCCCLXXX Kl's Decembris fauente mihi clarissimo atque sanctissimo uate [Petreio ac Pomponio omnium hominum sapientissimo] principe meo B. M. cuius etiam habui emendatissimum exemplar eius propria manu confectum." The words in brackets are in rasura. Simar thought that he could read Jacobus under the first erasure. After this subscription there follows a poem of four lines (Qui nostras placido, etc.) in the margin of which Simar thought he saw the words Jacobus Pontanus. He suggests that Joannes Jovianus Pontanus may have signed his father's name, which was Jacobus (but why should he do so?), or that Jacobus may really have been his name. 1 Simar says that Vat. 3211, presumably a MS of Propertius, was inaccessible to him because it was "en restauration." Since I had no record of this MS, I investi- gated in June, 1910, and found that it was a MS of Michelangelo. I may also note here that Simar's report that Cappon. 196 and Vat. 1611 contain Catullus is untrue. ! Pontani Carmina (ed. Soldati, 1902) . The Manuscripts of Propertius 299 But the facts are against this. Under Gaspar I distinctly read Johannes. The second word cannot be made out. The third word was clearly Hyspalen. The subscription is followed by the words Vale Tridentine, later erased, then the poem mentioned above. In the margin of verse 2 the name A. Tridentoni has been erased and in the margin of verse 4 Jacobus Bonon {not Pontanus) has been erased. There remains absolutely nothing, then, to connect this MS with Pontano. Furthermore, I am confident that I have discovered the MS that he did copy; but I must leave a discussion of this matter for another time. The connections of Vat. 1612 can, however, be traced. After the poem mentioned there follows another, beginning Incipit a teneris, in the maTgin of which the words Nicolaus Vulpes Vicentinus. A. Tridentoni have been erased. Jacobus Bononiensis, therefore, sent the one poem to A. Tridenton, and Nicolaus Vulpes Vicentinus sent the other to the same man. Both poems are found also in a Ravenna MS, Class. 277, but the "Incipit" poem follows Tibullus (signed Nicolaus Vulpes Vicentinus), while the "Qui placido" poem (unsigned) follows the Propertius. 1 The MS is dated 1459 and has the scribe's signature: ^airr^a-Tov 1,TreWaiov. In the last word we see the name of the town Spello, as we do also in the Hyspalensis of the Vatican MS. It should be remembered that the people of Spello (Hispellum) took a particular interest in Propertius, claiming him as one of their own. There is still a third MS belonging to the series, Goettingen Philol. 111. b. It has the poem about Tibullus at the end of that author's poems, and so reads explicit a teneris, instead of incipit. The poem is preceded by the name Nicolaus Vulpes Vicentinus, and is followed by a modification of the usual vita Tibulli. AH of this is in a different hand from the text. The other poem is not found in the MS. Its absence makes it more likely that the Goettingen MS, and not the Ravenna MS (as Sabbadini thought), belonged to Vulpes. 2 The Goettingen MS originally had the date Millesimo CCCCL Villi at the beginning, but the last three strokes were erased, leaving the date 1456. I cannot account for this. 1 Sabbadini Studi ital. di fil. cl. 7 (1899), p. 106. • It is true that in vs. 3 Goett. (and Vat.) has decet, whereas Eav. correctly has docet; but this may be a slip even on the part of the author. 300 B. L. Ullman Closely connected with Vat. 1612 is Barb. viii. 58 (which too Simar thought was written by Pontano) and Nap. iv. F 22, as is shown by a discussion of the name of Propertius' work which all three have in common. Nap. iv. F 22 was written in Bologna, July 17, 1465 by a "Vicarius potestatis." A Jacob of Bologna wrote one of the poems discussed above, and Nicolaus Vulpes was professor at Bologna for a number of years. 1 Bologna and Spello are the towns, then, that are connected with the MSS that we have been discussing. Far more pretentious than Simar's article is Richmond's, 2 as he has in mind the preparation of a critical edition. If he is correct in his main conclusions, he has done a most important piece of work, but I must confess that I have not been convinced by the evidence thus far submitted. Briefly, his main argument is that he has unearthed certain fifteenth-century MSS which ultimately go back to an earlier source than any existing MS. Much of his evidence is entirely subjective, depending on his view of the text in any given passage. He uses many readings of the new MSS to support an emendation of his own. This evidence I will not discuss, as such discussion would be entirely fruitless. But the evidence of a more objective nature calls for some criticisms. In the first place, Rich- mond lays great stress on the dislocations in his new MSS. He argues, plausibly enough, that only an old book suffers dislocations, that, therefore, his new MSS go back to an old source. But that is not necessarily true. Many MSS of Catullus show similar dislocations, MSS which I am convinced are not descended from an ancient arche- type, although I realize that Richmond would argue to the contrary. There is another explanation just as plausible. Perhaps the arche- typal MS had never been bound, and so was easily disarranged. Furthermore a book but a year old may fall to pieces if it is used much and hard, while another may, if undisturbed, last for centuries. Probably the archetype of these MSS was studied and copied a great deal and suffered much accordingly. Hence it is, too, that it was not preserved to us. Richmond thinks that c 1 (Voss. 81) was written by a scholar. I should rather say an ignoramus, judging from his text of Catullus, 1 Sabbadini op. cit. * O. L. Richmond, "Towards a Recension of Propertius," Journal of Phil. XXXI (1910), 162. The Mantjscbipts of Pkopertius 301 Tibullus, and Propertius. The word scholar is, however, an elastic one. It may be well to state that the text of Tibullus in this MS contains the known interpolations of certain humanists — a fact which throws a shadow of suspicion over the Propertius as well. Misapplied paleography, I am sorry to say, has affected Rich- mond's work, as it has that of so many other students of fifteenth- century MSS. On the authority of Bradshaw and Kuno Meyer he asserts that the use of finit for explicit shows that a MS was copied by an Irish scribe or was descended from an Irish MS. If this were true, all roads would lead to Ireland, for the use of finit is exceedingly common. It may be true that in Irish MSS finit is always, explicit never, used; it may even be true that the use of finit arose in Ireland, but its use in fifteenth-century MSS does not prove nor even suggest that these MSS are of Irish descent. I have incidentally noted finit in fourteenth- and fifteenth-century Italian MSS of Tibullus, Tacitus' Dialogus and Germania, Virgil's Culex, Cicero Ad Familiares, Persius, Catullus, Paulus Diaconus (Festus), and others, and I am certain that the list could be extended to include almost any Latin work. But I can present decisive evidence against the theory. Finit is often found in Italian MSS of works produced by Italian humanists of the fifteenth century. A few examples will suffice: 1 Paris. 5689 has "Dyodori siculi historiarum priscarum a Poggio in latinum traducti liber sextus et ultimus finit." Paris. 6310 (Aristotle Econ. and Pol. translated by Leonardo Bruno) has "Finit liber secundus," "finit liber octavus." Richmond's deductions from spelling are also valueless, in my opinion. Even if it is true that the spelling of his MSS shows that they are descended from an Irish ancestor, it is still possible that an Irishman copied the MS in Italy in the fifteenth century from an Italian archetype. But I doubt if even so much can be admitted. In the same way Richmond's arguments that the errors are most easily explained as being due to an Irish archetype do not convince. Rasce from nisee, canna from omne, which he cites as examples, are quite as easily explained from a humanistic archetype, if we assume that the copyist knew little or no Latin. University of Pittsburgh 1 They are taken from Mazzatinti La biblioteca dei re d'Aragona in Napoli, pp. 7, 37. Other examples may be found there and elsewhere.