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. TABLES OF CONTENTS IN THE MSS OF PLINY'S LETTEKS By Frank Egleston Robbins Since the article by Stangl in Philologus (XLV), little attention has been paid to the indices of codex Beluacensis, a unique character- istic of this codex among the Plinian MSS. The following notes have been made as the result of a study of these indices, in the effort to supplement Stangl's work, and to gain an idea of the significance of the indices in the Plinian text-history. A complete text 1 of the indices of B follows : 2 C PLINI SECUN DI | EPISTULARUMLIBKINU | MERO DE- CEM:; | INCIP LIB • I • FELICITEB; | (thus far in large capitals of faded red) adsecundum adarrianum | adcaniniumrufum • adpompeia • | aduoconium • adcornelium taciturn — | adoctauum ruf um • adpompeium • | •r {2d A.) adminucium ■ adatticum • adfautuiustu • — | adcelest • ium tironem • ad- sossium ■ adiunium (n over a blur or blot) | adsepticium • aderucium • ad- cornelium • | adsuetonium • adromatium • adcornelium • | adplinium • : ci (gd k.) adca.-lium • adpompeium • adbaebium ^f ■ (by 2d h. over erasure of simi- lar mark) Frequenter quia tardiorem • | quid agit, (comma apparently by 2d h.) quantum copiaru • uidisti • ridebis • | uide // (erasure of two let- ters, possibly -bo) inquo • peroportune • mirumest • | siquando olim mihi • iacturam • magnu | petis • heus turn • amabam • estadhuccure • | scribis K ilsth.) te • municeps (-ps by 2d h. over erasure) freques • utanimitui • | diui- diam • consulis • tranquillus • | (Thus far in smaller capitals, with here and there an uncial form; some of the initials are in red.) [Book IT] Ad romanum post aliquot annos IN | Ad paulinu irascor nee liqu& michi || (there are two addresses and first lines to the line, separated by an irregular dividing line, thus making two columns of the whole; the writing is minuscule.) Adnepotem • magnis aeuum fama • | Ad galuina sipluribus pat tuus || Ad lupcu actionem • a&atefre- 1 Por this text I am indebted to Professor Elmer Truesdell Merrill, of the University of Chicago. For the description of Beluacensis (sometimes called Eiccardianus) see Stangl loc. cit.; L. Haret Revue Critique XVI; E. Chatelain Paliographie des classiques latins, Pt. 2, text to PI. CXLIII ; C. Paoli / codici Ashburnhamiani delta R. Biblioteca Medieeo-Laurenziana di Firenze I. 53 f. 2 The division by lines in the index of Book I, and the rudely effected division into double columns in the index of Book II, are indicated in the print. In the indices of Books III-V the columnizing is neatly carried out in the MS. [Classical Philolooy V, October, 1910] 476 Tables of Contents in Pliny's Letters 477 quent • | Adauitu • (au- over eras.) longu e altius rep& || Ad magnu • nereasenatuuestricio • | Ad caninium studes anpiscorisa || Ad apollinare • anxiume&inqu& ■ | Ad octauium • hominemtepatient || Adarrianu • •r AIHOYPriON illud qd | Ad pri-cu & {over eras, apparently of -m) tuoc- casionesoblig — || Ad maximu ■ ueruopinaris • | Adualerium quomo- doteu&eres || Ad annium • tuquidem f>c&era • | Adgallumirariscurme- laurent || Ad marcium • quid attemihi | Iucundius || Adceriale • horta- risutorationem • | Ad caluisium ■ assempara&accipe || [Book III] Adcaluisiumrufum • Aduibium maximum • Adcaerelliae hispullae- Adcaeciliummacrinu • Adbaebiummacrum • Adanniumseuerum Adcaniniumruf um • Adsu&on tranqui ■ Adcorneliumminicianu • Aduestric {erasure of -iu) spurinn Adiuliumgenitor • Adcatilinum seuer • Aduoconiumromanu • Adpatilium Adsiliumprocul • Adnepotem • Adiulium seruian • Aduirium seuerum ■ Adcaluisium rufum • Admaesiummaximum • Adcornelium priscum • [Book IV] i Adf abium • prosoc ii Adattium • clemen • iii Adadrianum • antoNIN iiii Adsosium • senec v Adiuliumsparsum vi Adiulium nason • vii Adcatium • lepidum viii Admatur • arrian viiii Adcornel • ursum • x Adstatium • sabinum xi Adcornel • minic xii Amatur • arrian • Nescio anullum Quodipse amicistuis • Cumpatremtuum • Quamuis&amici • Pergratumest michi Exhereditate quae Modonuntiatusest Facisadprocetera Possumiamperscrib (-i- long like I) Composuissemequaed Estomninoartemidori Veniam adc§nam Librum quonuper Rematrocem Petisutlibellostuos Adnotasseuideorfacta Recte omnia [dictaq Officiumconsulatus Adsumoteinconsiliu Meministinete u {1st h.) Adio ualeriumartial upispostlongum • regulus filium quodsimulatq • iterum uarisidium(co»T./r.-uum)nepotem • aeschinen aiunt tuscigrandine excussi saepetibidico gratularismichi causamperhosdies scribismichi sabinum audistineualminic amasegnatiummarcum 478 Feank Egleston Bobbins xiii Adcornelium taciturn xiiii Add • paternum • xv Adminic • fundan xvi Adualerium • paulinum xvii Adclusinium • gallum xviii Adarr • antoninum xviiii AdcalpurN' hispull' xx Adnouium • maximum xxi Aduelium cerialem • xxii Adsempron • rufum xxiii Adpompon • bassum xxiiii Adfabium valent xxv Admaesium maximum • xxvi Admgcil • Nepotem xxvii Adpompei • falconem xxviii Aduibium seuerum • xxviiii Adromat • firmum • xxx Adlicinium suram • [Book V] Adanniumseuerum • AdcalpurN' flaccum • Adtitium ariston • Adiuliumualerianum • Adnouiummaximum • Addomit • apollinar • Adcalpurnium rufum • Adtitinium caepiongm • Adsemproniumrufum • Adsueton • tranquillum • Adcalpurn • fabat • pros Adscaurum • recitaturus Adterentium scaurum Adpontium • allifan' Adarrium antonin' Adaefulanmarcellinum • Aduestric spurinna AdcalpurN' macrum Adualerium paulinum • Adcornelium ursum • Adpompemm • saturN • saluuminurbeuens tufortasse siquidomnino gaudemeo & admones & rogas quemadmodummag cumsitpi&atis • quidsenserim tristem&acerbum inter fuiprincipis magnamcepiuoluptat proxime cumapud scripseramtibi petisutlibellos tertius dies est herenniiseuerus heatu proxime attulit tibi • legatummichiobuenit accepipulcherrimos cumplurimaofficia resparua nuntiaturmichi amauicuram necheredeminstitui hicadfi • (much later hand) suades uthistoriam descenderaminbasilicam 1 1 1 1 1 1 1 1 1 1 (erasure) liberatandem praecepilitterastuas (-cepi corr. by 2d hand from -cipi) oratiunculam & turogas secesseram inmunic • (corr. by 2d hand from secesserim) cum uersus tuos tristissimus haectibi scio (started as sco-) quantoopere beneest michi uideo quam molliter iterum bythini uariaeadfecerunt Tables of Contents in Pliny's Letters 479 Stangl concluded that the indices of B were not compiled from that MS itself, but from an earlier MS of the same family. The evidence for this is obvious. In certain cases (as in i. 1, ii. 15, ii. 16, and v. 5) , the index and the context of B agree in readings not given in M and V; and on the other hand, some of the readings of the indices give evidence of greater antiquity than the context. 1 It may also be noticed, first, that while in the first two books the indices are irregu- larly arranged, in the later books they become better systematized; second, that their transmission has been notably careless; and third, that to all appearances they were for some time transmitted in the form of indices — a point made by Stangl. The first point is obvious. Not only were the tables for books i and ii in irregular form, but, especially in book i, the addresses and first lines are much less full than in the succeeding books. The second point is hardly less plain; to carelessness must be referred cases of merely orthographical error (see, for example, the address of i. 13), and the more aggravated cases, such as the addresses of i. 11, 12, 22, and iii. 3, and the first lines of i. 22, ii. 3, 7, iv. 28, 30. Certain of the variants which have a direct bearing on the text- history will be discussed later. The very frequency of careless mistakes in the indices goes far to corroborate the third point men- tioned above, namely, that the indices were independently copied for some time; for there is a greater chance for error in copying a list of unconnected words and phrases than in the transmission of a running text. Further corroboration of the third point is drawn from the omission of ii. 11 from the table of contents for book ii. The eleventh and twelfth letters are addressed both to the same person, and a copyist, working from a table of contents, not referring to the full MS, could easily copy the address of letter 11, and skip to the first line of 12 without noticing his omission. Some of the mistakes of the indices may be accounted for on the 1 The best evidence is the inclusion of the address and first line of iv. 26 in the index, though missing entirely from the context ; cf . likewise the numerous instances where double names occur in the indices and single ones in the context ; the first line of iv. 18, where B and F in their context omit magis, which occurs in the index ; also the difference of word-order in the first line of ii. 18 (the index has mihi lueundius, and the context iucundius mihi). 480 Feank Egleston Robbins hypothesis that in their original compilation, or at least very early, many of the proper names were written in abbreviated form, and that these were later incorrectly filled out by copyists. For example, i. 10 in the index reads adatticum. It should be ad Attium, of course; but if the copyist had had before him an abbreviated form (e.g., ad Atti.) it would be a very natural mistake to supply the well-known form Atticum. 1 This supposition is supported by the fact that many of the names in the tables of contents actually stand in abbreviated form. Words other than proper names are also frequently abbrevi- ated, but because they stand at the end of the several citations of first lines. Let us now turn to the question with which this paper is chiefly concerned — the distribution of double names in the addresses in the chief Pliny MSS as compared with that in the indices of B. Even in Kortte's edition many double names were given on the sole author- ity of the index of B, and this in fact is the most fruitful source for such addresses. On the other hand, the Medicean and the Vatican MSS are practically without double names in the addresses. The practice of these MSS is to give the cognomina (sometimes the nomina). None of the MSS seem to have transmitted praenomina. 2 The general facts concerning the distribution of double names in the addresses are presented in the following table, covering the chief MSS and Aldus: 3 Book I Book II Book III Book IV Book V Book VI Book VII Book VIII Book IX Totals 5- 20 20 20 16 18 24 1 1 20 19 11 a 2 2 10 11 21 28 13 6 9 8 30 19 5 21 34 33 24 40 71 20 F, including margin 49 29 M 3 V 3 35 37 Number of letters in 247 1 See also i. 12, iii. 12, iv. 1, 12. 2 For this reason we must reject Stangl's conjecture that Patilium (cf. the index, iii. 14) is a corruption for P. Atilium. 3 In this table account is taken only of genuine double names, not of forms like Fabatus prosocer. Tables of Contents in Pliny's Lettees 481 Two general observations may be made. First, in book i the context of B, F, and a agree exactly in reporting double names in the addresses. Secondly, book ii seems either to have lost all its double names in the addresses early, or never to have had them. The only double name in this book is Pavlino Nepoti (read by M V in the second letter) ; but this is without doubt due to a confusion, for the third letter is addressed Nepoti. The early editors intro- duced double names in this and three other addresses in book ii. In ii. 2 Laetus and Beroaldus followed V, which the former was the first to use; but Catanaeus emended to Valerio Paulino, evidently drawing this from iv. 16. The other instances in these three editors are Voconio Romano, ii. 1; Caninio Rufo, ii. 8; Cornelio Prisco, ii.13. These are all well-known correspondents of Pliny, and their full names occur in the "eight-book" MSS in i. 5, i. 3, and iii. 21 respec- tively, whence the editors easily could, and doubtless did, take them. The question now arises, was the B F family originally supplied with double names in the addresses throughout the letters? It may be seen from the table that in book i, where the index has single names, B (context), F, and a agree in reporting double ones, and that all alike have none in book ii; moreover in books iii-v the indices and F have double names, but B (context) and a have not. It has already been remarked that, when the MSS give single names, they regularly give cognomina. Now in book i the index reports five of the twenty double names found in the addresses of the context of B, in F, and in a, but in every c ther of the fifteen instances where it fails so to report, the nomen is given. I take this as a plain indica- tion that the exemplar from which the indices were originally com- piled had double names in at least these 20 letters of book i, and that the scribe failed to copy them; for if there had been single names only in the exemplar, it is likaly that most of these would have been cognomina. 1 The indices of the first two books are more careless in their arrangement than those of the later books, and give evidence that the copyist had not yet settled upon a regular form; it may well be that when he copied the index of book i he had not yet decided to copy double names. Of course the same may i It will be seen that the scribe of F, under similar circumstances, did precisely the same thing ; see below (p. 484). 482 Frank Egleston Bobbins have been true in the second book, and double names may have existed in the archetype of the indices; but the agreement between all the MSS is so close that the safer conclusion is that they had entirely disappeared from this book before the indices were compiled. In books iii-v the indices are quite fully supplied with double names. The MS from which they were compiled, therefore, had double names in these books, and the natural inference is that it had double names in the addresses of books vi-ix as well (for it is probable that this MS was not mutilated, as B and F are). Double names, then, existed in the addresses in the early Plinian MSS; they have disap- peared almost entirely in the M V family, and probably in the " eight- book" family as well, 1 but were kept longer in the B F family, and particularly in the indices which went with the MSS of that family, and which formed, as Stangl pointed out, a practically independent tradition. Another question to be answered is, did indices, like the indices of B, exist in other MSS? I am convinced that they did, but only in the B F family. I can find absolutely no evidence for their exist- ence in M, V, or in the "eight-book" tradition. 2 In F, however, there are traces of the influence of indices, which in two or three places are quite convincing. For the purposes of this paper it may be assumed without argument that B and F are closely akin, both descendants of a mutilated MS extending through v. 6, omitting iv. 26, and that Aldus' Paris manuscript was of the same family, but of an earlier branch, which left the main stem before the mutilation seen in B and F took place. Now, in iii. 18, B (context) and a read 1 In iii. 18 both o and u read curio smero in the address. Now this is an erroneous reading and can definitely be traced to the rubricator who worked upon F itself ; see above (p. 474). This is certain evidence of the influence of P. Besides this, there are the suspicious facts that no double names occur in the addresses in o and u which do not occur in F also, and that all the double names in addresses are confined to the first four books. Hence it is quite likely that the double names in the addresses in the "eight-book" MSS are derived from F or its descendants. 2 Since this article was sent to the printer Professor Merrill has informed me that in St. Mark's Library at Venice Lat. class, xi cod. xxxvii (chart, saec. xv) contains, beginning on f . 94, the Breuis Adnotatio de Duobus Pliniis, then, in columnar form, an index of the addresses of Plin. Ep. i (omitting, as in cod. Dresdensis T> 166, letters 8, 12, 23, 24, and giving the nomina), followed by the text of Book i, and that by a simi- lar index to Book ii, beyond which the copy of Pliny was never carried. The text of Book i is that of Dresdensis, and the index appears to have been made from the imme- diate archetype of the fragmentary copy. Tables op Contents in Pliny's Letters 483 Seruo, the index of B reads aduirium seuerum, and F has curio seuero, which is the mistake of the rubricator who inserted the addresses; for in the margin the note left for his guidance reads uirio seuero, and is still legible. This error therefore occurred in F itself. With- out a doubt Aldus' Parisinus read Seruo here, for this reading occurs in no known edition or MS available to him (he did not know B), and it is moreover a most unlikely emendation for seuero, which Aldus must have seen in the earlier editions, or in the copies of F, and in the "eight-book" MSS current in Italy in his day. Seruo, therefore, was indubitably the reading of the common ancestor of B and Parisinus, and therefore likewise of the common ancestor of F and B. We must therefore conclude, in the absence of any other possible source, that F took uirio seuero from indices like those of B. If we assume what seems likely, that F is not a copy of B (the evidence for this will be found in the preceding paper), this means that in the immediate archetype of F, as well as in that of B, there existed such tables of contents, all drawn from the same compilation. In iv. 29, F's margin bears romatio firmo (agreeing with the index of B, q.v.), but the letter is actually addressed in the context Romano, and at the top of the page in the side margin, in the first hand, we find Plinius Romano suo Firmo Sat. Probably the marginal note came from the index, but Romano from the context, of F's parent (for the context of B reads Romanio). If the additions and substitutions from the indices had already been made in the parent of F, we should hardly expect to find such a note as this, but one reading or the other would be given without comment. On the other hand, the presence of this note is very natural, if the copyist of F was working with constant reference to the indices, and in this instance changed his mind as to the form to be used for the address of the letter. I think, therefore, that the scribe of F has interpolated many double names into the addresses in F from indices which he found in his immediate archetype. The theory that F was interpolated from a set of indices is almost necessary to explain why F, which in general is not so faithful a copy of its immediate archetype as B seems to be, should be so much better furnished with double names in its addresses than the context of B; and likewise to explain why F is so startlingly like 484 Frank Eglbston Bobbins the index of B in these double names. The following collation, kindly furnished me by Professor Merrill, shows marginal notes in F which make this similarity all the more striking: iv. 17: | sinio ga F in margin; address omitted in context. iv. 24: fabio | ualenti, F in margin; Fabio, F context, iv. 25: mesio ma | ximo, F in margin ; Mesiono Mimo, F context, iv. 27 : popeio f . . | coni, F in margin; Pompeio, F context, iv. 29: romatio | firmo, F in margin, but in side margin, top of page, in first hand, Plinius Romano suo | firmo sat; Romano, F context, iv. 30: licinio | sure, F in margin; Licinio, F context, v. 1 : annio | . . . ero, F in margin ; Annio, F con- text; v. 2: calpu . . | nio flaco, F in margin; Calpurnio, F context. v. 3: titio | aristoni, F in margin; Titioni, F context, v. 4: iulio | uale- riano, F in margin; Iulio, F context, v. 5: nouio | maximo, F in mar- gin ; address omitted in context. These notes are the remnants of the guides left by the copyist for the benefit of the rubricator who inserted the addresses. The margins were afterward clipped, but the notes aside from this cur- tailment remain perfectly legible. In each of the above cases the marginal note is in substantial agreement with the indices; the rubricator, however, failed to copy the whole title furnished him, but inserted in the context only single names, 1 and, moreover, the nomina. In each of the above places the addresses in the context of B have the cognomen, which as has been pointed out is the regular practice of the Plinian MSS. Inasmuch as it is possible to check certain interpolations of addresses from the indices into F, as has been shown above, probably the readings from F just cited, and in fact most of the double names in the addresses of F after book i, are similarly interpolated. It is at any rate likely that the original extent of double names in the addresses of the context of both B and F was the same, and it is in keeping with the general character of the copying of F that there should be interpolation and emendation. The extent of double names in the addresses of the parent MS of F, then, was probably about the same as that in the context of B. If the immediate ancestors of F were supplied with indices, can the same assertion be made of Parisinus? The question is a more elusive 1 The scribe of the indices of B also gives many nomina, as single names in the index of Book i. This is, I think, a case exactly paralleled by the one under discussion. The scribe of the indices had full names before him, but copied only single ones. Tables of Contents in Pliny's Letters 485 one, inasmuch as we can judge of Parisinus in this respect only through Aldus. After book iv, Aldus did not print any double names what- ever in the addresses, although he might easily have supplied a considerable number by conjecture, as some of the other editors did. Furthermore, every double name which he prints in the first four books is found either in F, the copies of F, the early editions, or the "eight-book" MSS. It is hardly likely that the ten-book MS Parisinus contained double names in precisely those portions of the letters covered by B and F, and by a strange coincidence had lost every vestige of them in the addresses of the other books; and if he had had them in his MS, Aldus undoubtedly would have printed at least some of them. I believe, therefore, that the double names in the addresses of Aldus were printed by him on the authority, not of Parisinus, but of some of the MSS of the "eight-book" type, of the early editions, or of copies of F which were current in Italy in his day; that Parisinus had no indices; 1 and that double names had fallen out of the addresses in the context of Parisinus to a great extent, that is, that no more survived in Parisinus than in the con- text of B. None of the known MSS give double names in the addresses beyond the fifth book. In the early editions (those of Laetus, Beroaldus, and Catanaeus) are found the following: vi. 15: Voconio Romano, Laet. Ber. Cat.; vi, 21: Caninio Rufo, Laet. Ber. Cat. (so also in vii. 18, ix. 33, and viii. 4, with orthographi- cal variations); vii. 2: Fabio Iusto, Laet. Ber. Cat.; vii. 3: Cres- cenio Praesenti, Laet. Ber.; vii. 9: Cornelio Fusco, Laet. Ber. Cat.; vii. 17: Nonio Celeri, Laet. Ber. Cat.; vi. 23: Triario Traiano Laet., Ber. 1498 (but in edition of 1501, Triatio); vi. 34: Maximo Aphricano, Laet. Ber. Triario Traiano is probably a combination of two "eight-book" readings for the same name. Caninius Rufus, Voconius Romanus, and Fabius Iustus are addressed in i. 3, 5, and 11 respectively, and the double names here are clearly brought into the text from the previous letters. Nonio Celeri could have been taken from vi. 32. 1, where his full name is given. Cornelius 1 Additional (negative) evidence for this lies in the fact that Aldus in his preface does not describe Parisinus as having indices. There is an index in Aldus' edition of the letters, but it is a mere compilation from the text, and furnishes no evidence. 486 Fkank Egleston Bobbins Fuscus and Crescenius Praesens are not spoken of elsewhere in Pliny, but the former is a character well known to authors contemporary with Pliny, and the full name here is undoubtedly a conjecture of Laetus. Crescenio Praesenti must have been introduced by Laetus, but on what grounds I cannot at present state. He is otherwise unknown. In a copy of the Treviso edition of 1483, once owned by Laetus and annotated by his own hand, now the property of Professor E. T. Merrill, Crescennio is noted, undoubtedly in Laetus' own hand, for insertion before Praesenti, but Crescenio was printed by Laetus in 1490, and adopted by Beroaldus in both editions. Aph- ricano (vi. 34) is evidently taken, mistakenly, from Africanae (vi. 34. 3). The facts set down above may help in the reconstruction of the stemma of the B F family. The three readings cited (ii. 18, iii. 18, iv. 29) show that the compilation of the indices took place before the erroneous readings of B and Parisinus came into the text. The stemma, then, is somewhat as follows: P(arisinus) Here /3 represents the earliest Plinian MS of the B F P type, and z the MS from which the indices were compiled. This z was earlier than y, the latest common ancestor of B, F, and P. Aldus tells us that P contained ten books; 1 both y and z then may be assumed to be of that extent. Between y and x, however, the mutilation seen in B and F took place. Both y and x probably contained sets of indices, but between y and P they seem to have been lost. Double ! His statement is confirmed by the discovery of Budaeus' notes taken from Parisinus, in a copy of the letters in the Bodleian Library; see an article by E. T. Merrill in Classical Philology, II (1907). Tables of Contents in Pliny's Letters 487 names must have existed in the addresses in the B F branch as far back as z; but after that they dropped out to a great extent, so that B, F, and P, in their context, had them only sporadically, or, like B, only in book i. One is tempted to imagine that the lack of double names in the index of book i led to their preservation in the context of B, and that, conversely, the completeness of the indices in reporting double names in books iii-v caused their disappearance from the addresses in the context of those books. It seems probable that z had double names in the addresses in the context of books vi-ix, as well, and that the indices of those books were compiled in the fashion finally adopted in books iii-v, giving the full form of address in each case. There is no reason to suppose that the original form of the indices was very much different from that now seen in B, except that if they had originally been prefixed in a body to the whole work, and were not distributed before the appropriate books until later, this may explain why the indices were lost in P and else- where. But they merely have been disregarded, wherever they occurred. Stangl's conjecture of the eighth century as the date of our hypothetical z will fit as well as any other. B is of the ninth or tenth century, and despite Aldus' belief in the great antiquity of P, it is not necessary to suppose that that MS was much, if any, older than B. After the copying of z, however, the text of the B F family, which had previously been in closer agreement with the M V tradition, underwent certain changes, including variation of word-order (cf. index and context of B in the first line of ii. 18), the loss of certain words (cf., e.g., iv. 18), and corruptions (cf., e.g., iv. 4, iii. 18, iv. 29). The University op Chicago November, 1909