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 RELATION BETWEEN CODICES B AND F OF PLINY'S LETTERS' By Frank Egleston Robbins The exact relationship between codex Beluacensis and codex Florentinus of the letters of Pliny is not yet settled, although even before the recovery of B the close connection of the two MSS was undisputed. The purpose of this paper is to examine the evidence and to decide the question as definitely as possible. B and F agree very closely in contents. The content of B at the present time is: i. 1 — ii. 4. 2 extiterim; ii. 12. 3 praebere — iii. 5. 20 futura; iii. 11. 9 amicos — iv. 25. 5 vale; iv. 27. 1 — v. 6. 32 pererrat. The MS breaks off suddenly and without any warning at the bottom of the verso of fol. 18, its last leaf. The first two lacunae, as has been shown, are due to the removal, in each case, of two leaves. In F the only large lacuna is iv. 25. 5 vale — iv. 27. 1, and the MS extends in the first hand through v. 6, ending on the sixth line from the bottom of fol. 77 recto; the rest, comprising v. 7 — v. 8. 4 natura curiosi, was written much later, apparently by the same scribe who corrected the Greek quotations. If, then, v. 6 was originally com- pleted in B, on an extra leaf stitched on after fol. 18, as seems very- probable (although the last page of B in its present state seems from its rubbed and blurred condition to have been on the outside for a long time), the original content of B and F was the same. In any case the loss of iv. 26 in both points strongly to their close con- nection, and their remarkable similarity throughout the text is further evidence in the same direction. To illustrate this point we may cite: 2 3. 7: curatius, D; accuratius, BF; cura maiore, M V. 5. 8: ilia popularia ilia paucorum quid cubicula, M V; ilia popine quid euripus quid cubicula, B F. 5. 17: numquam tuum desinet esse si semel coeperit, BFD a; num- quam desinet esse si semel coeperit tuum, MV. 1 1 am indebted to Professor Elmer Truesdell Merrill, of the University of Chicago, for much material and friendly suggestion in the preparation of this and the following paper. The system of nomenclature of the principal MSS, and the abbreviations there- for, are those adopted by him in his Selected Letters of the Younger Pliny (Macmillan, 19031. 2 Citations of the text are made by page and line in Keil's larger edition (1870). [Classical Philology V, October, 1910] 467 468 Fbank Egleston Bobbins Although in general B and F are much alike, there are many points of difference which at once furnish us the bulk of the evidence upon the questions under discussion, and allow us to form a 'con- ception of the methods followed in the copying of each MS and of the character of its scribe. The scribe of F seems to have been not unwilling to emend portions of the text which he failed to under- stand, or where he thought a small change or addition made an improvement; he was rather keen, but not very brilliant. On the other hand, the scribe of B evidently presents a faithful copy of his exemplar, and, it will be seen from various citations, probably repro- duces in certain cases even its minute peculiarities. As an illustra- tion of the character of the transmission of F, see for example 67. 14, where F reads dormiebatque non minimum, inserting non against the authority of all the other MSS. The scribe must have reasoned that if the elder Pliny awoke quasi alio die, his nap was not particularly short. Again, in 104. 17, F reads conscribitur, while BMVDa correctly read constringitur; F's reading must be an attempt to better the text. There are besides about thirty instances where in F a short word, like inquit, sed, enim, et, ut, -que or quidem, has been added. These additions are not necessary, but are precisely the sort of glosses with which the scholars of the Middle Ages often delighted to cover their classical texts. A fair example is 8. 1, nondum ab exilio uenerat BMVD; nondum enim ab exilio uenerat F (enim above the line, first hand). The character of the copyist of B is best illustrated by contrast- ing him with the copyist of F in cases where the former was content to copy what stood in his exemplar, even though difficult to under- stand, while the latter would attempt an emendation; as in 19. 3, qui studia ut non simul et nos amet B (non is added by the second hand, approximately contemporary with the first); qui nos amet, ut studia non simulet, F. 1 The readings so far cited are unimportant for the solution of our problem. The last reading of F, for instance, could be based as well upon the text of B itself as upon that of any ancestor of B. The two MSS are indeed so nearly alike that clear evidence, based on 1 o reads the same as F. There is, however, absolutely no proof that F was affected by the "eight-book" MSS, though many of them show interpolations from it. Relation between Codd. B and F of Pliny's Letters 469 striking differences, is very rare, and the weighing of testimony is a delicate matter. In the first place, we shall examine evidence bearing upon the question whether F is a direct descendant of B itself. It is obvious that the best proof that this is not the case is to point out matter in F which cannot be derived from B but must have been taken from a MS earlier than B of which B is a descendant. In 4. 4, B and F omit nam uim, but between the lines -F has siue materiam. F cannot have taken siue materiam either from B or from a descendant of B; F, therefore, was copied from a MS earlier than B, from which already the words nam uim had disappeared, although the gloss on those words — siue materiam — was retained. One other passage points clearly to the same conclusion, that F is not a descendant of B: 102. 5, sed prius accipe causas rogandi deinde ipsum quod peto proxime cum B F a; deinde ipsum quod peto, om. M V. These words are plainly written in B, but the copyist of F first wrote proxime cum immediately following causas rogandi, later erasing it and substituting deinde ipsum quod peto. The traces of erasure are visible in the MS. Now, deinde ipsum quod peto, which M and V omit, is most probably a gloss, for as it stands in the context it is tautological and unnecessary; but it is just the sort of a gloss that an ingenious copyist would make. If F is a descend- ant of B there is no good explanation why proxime cum was first copied and then erased; it is best explained on the hypothesis that there was a gloss in the margin of the common source of B and F, which the copyist of B incorporated in the text, and which the copyist of F failed to notice at first. We must therefore conclude that B is not the ancestor of F. The evidence, to be sure, is very slight, but the resemblance between the two MSS is so close that a large number of striking differences is not to be looked for. Now if F is not a descendant of B, the pos- sible relationships between them may be represented as follows: X X X X /\ /\ / \ / \ 7 \ / y y z / \ B 7 \ B \ B 7 \ / \ B \ F F F P Fie. 1 Fig. 2 Fig. 3 Fig. i 470 Prank Egleston Bobbins Because of the close similarity of the two MSS it is quite out of the question to imagine any more complicated arrangement possible. We shall first examine evidence for and against Fig. 1, which assumes that F's immediate ancestor is more remote in the Plinian stemma than that of B. There are many cases where B preserves a peculiarity of x (see Fig. 1), -while in F a correction has been made. So in 120. 19, mea suscepisse B F (corr. first hand from me usu cepisse, B), it is possible, of course, that both y and B preserve the true reading, me usu cepisse, and a correction already made in x, while F copied merely the correction. In 84. 25, B reads a enim aduertienim (a enim corr. by second hand); F has aduerti enim. The reading of F is better, but it is likely that B preserves something from x which F has lost; if Fig. 1 be accepted, this must have been preserved in both B and y. Again in 126. 2 F has ibi, and B, sed ubi (line under sed and perhaps under first stroke of u). Again B has apparently preserved the text of x, correction and all; F has made the indicated correction. Likewise in 100. 2, B has a casto puroque corpore; F reads a casto corpore puroque. Here B seems to preserve dots that occurred in the archetype, and F interpreted them to call for a change of word-order. In 91. 21 B has hominem neromanu; F reads the same, but corrected by the addition of a vinculum to read hominemne romanum. B again is nearer the text of x. In 10. 19 B .men reads certdndum habent; F, certam ndu habent. Certandum habent is the correct text, and evidently this, incorrectly glossed, was the reading of x. It must be granted that B keeps the text of x, correc- tions and all, more faithfully than F. None of the instances cited above, taken by itself, definitely shows that y (Fig. 1) did not exist; but taken collectively, they make it veiy improbable that it did; for otherwise it is neces- sary to suppose that both y and B perpetuated the same peculiar and often illogical reading. The fact that B is certainly two hundred years older than F likewise makes Fig. 1 very improbable. There are, to be sure, certain readings that seem to favor the scheme, but they are explicable otherwise. For example, 106. 16, omi ope omi labore omi gra, F; omni ope omni labore gratia, B a; omni ope labore Relation between Codd. B and F of Pliny's Letters 471 gratia, M V. Pliny undoubtedly wrote the words as they appear in F, and at first sight it seems that F preserves the earlier text. But the agreement of Aldus with B is significant, for Aldus' Parisinus was undoubtedly akin to B and F, and being a ten-book MS, as he tells us, was from a source older than the mutilated MS which was the source of both B and F. Aldus undoubtedly followed his MS in this passage, for he would not have retained a poorer text without MS authority, when he could have borrowed the reading of F from the early copies or editions; he did not use B. The mutilated ancestor of B and F, then, read as Aldus does, and the reading of F is only a simple and necessary emendation, natural enough to anyone familiar with Pliny's style. If Fig. 1 be rejected it follows logically that Fig. 3 is just as untenable. The same passages may be quoted, with the following additions : 35. 5: lectioinsubitis, B (corr. from subditis in first hand) lectio in sub/itis (erasure of d), F. 85.3: Icetioris, BF; 11.4: desidioso (apparently corrected by second hand to desidiose), B; desidioso corr. to desidiose, or vice versa, F. In 85. 3 both B and F have preserved a digraph, rare in both MSS, and in 11. 4 and 35. 5 the archetype of each seems to have been similarly corrected. Only by the barest coincidence could these peculiarities have been preserved in both lines of Fig. 3. It is therefore probable that either Fig. 2 or Fig. 4 represents the true situation. In either case the relationship is so close that it is a delicate matter to come to a decision as to which figure repre- sents the truth. The chief argument in favor of assuming a step between F and x (that is, in favor of Fig. 2) is drawn from the number of emendations in F, the character of which has been indicated above. It may be argued that these are too many to be the work of a single copyist, who was not professedly editing the text. Moreover, some of these emendations are incorporated in the text — e.g., see the interpolation of inquiens, 7. 22, in F, where B reads parce (om. inquiens) cui ego, as though the parent MS had presented the same appearance, although we know that x, as the parent of B, could not have read thus. F also contains interpolations above the line. But after all, is it 472 Frank Egleston Bobbins unlikely that these interpolations are the work of the scribe of F? In one case at least I am convinced that such an addition can be certainly ascribed to him. From this it would appear that the scribe of F had a taste for such correction, and it is easier to postu- late that one man of such tastes revised the MS than to assume that there were two or more revisions of the same character. There is absolutely no evidence, moreover, that necessitates the supposition of an intervening step y, and, on the contrary, all the facts can be reconciled to the theory that B and F are sister MSS. The latter theory becomes all the more probable when we bear in mind that we know absolutely nothing of the history of the parent of B, after the copying of B, and that successive owners may very well have inserted many of the above-mentioned interpolations. Further- more, there is no evidence from the " eight-book " MSS — which were affected by the B F tradition — for the intermediate step. Let us now examine the evidence. In the address of iv. 29 F reads Romano, but in the margin the rubricator's guide read romatio firmo, while at the top of the page the first hand wrote Plinius Romano suo Firmo sat. Now Romano seems to have been the reading of the B F family in general in this place; it is found in a, and B has Romanio, clearly a corruption of the same reading; but in the indices of B, and there alone, ad romat. firmum is read. In the following paper I have shown from this and similar passages that F had access to indices like those of B but not those of B itself (for F is not the descendant of B). In this instance the copyist of F, after deciding to emend in accordance with the reading of the indices, changed his mind, and wrote at the top of the page the text which he finally favored, although the rubricator wrote Romano, not Romano Firmo. Now if all this were not due to the copyist of F himself, it is extremely likely that the note at the top of the page would not appear, but that one reading or the other would be given without comment, or with comment in such form that we might be sure that both readings appeared in the text which the scribe was copying. The instance is surely enough to show that the scribe of F was capable of such emendation as we have been discussing, and that probably many of the emendations noticed are due to him. Relation between Codd. B and F of Pliny's Letters 473 On the negative side, the evidence is drawn from the relations between F and the " eight-book " MSS. 1 It is evident that in certain places this tradition has suffered correction from a MS very like B and F. They are not copies of either, for there are traces in them of an independent tradition earlier than B or F. For example, iv. 26, which is omitted in B and F, is found complete in o and u and is without doubt not an interpolation from any other family, as a few readings from this letter will show: 114.7: recognoscendos emendandosque oua; legendos recognoscen- dos, MV. 114. 8: suscipere libentius, oua ; libentius suscipere, M V. 114. 15: istis (istos, M) addere uelis, M V a ; uelis istis addere, ou . The last is particularly noteworthy. The following cases are like- wise significant, though not conclusive: 129. 17: gracili marmore continetur, om. F, add. Mou. 111. 21: catullo, om. BP, add. Mou. 111. 23: quo, om. BPa, add. Mou. 68. 7: potuisse se, om. BF, add. M Voua. Finally, these differences of reading indicate at least independence of B and F on the part of o u: ii. 15 (title): ualeriano, M VouD; ualerio, BF. ii. 16 (title): anniano MVDu; aniano, o; annio, BF. 112. 16: imperare, BF; impertire, Voa; impartire, u; impetire, M. 119. 10: equites romanos splendidos, BPa; splendidos equites romanos, Mou. Since the "eight-book" MSS display evidences of independent tradition, we may next ask to what MS we shall ascribe the BF influence. It may be stated at once that neither B nor a descendant of B is responsible, but that o and u bear a much greater resemblance to F. Were they affected, then, by F itself, or by an ancestor, or by a descendant of F? The text gives evidence as to this: 6. 21: cum meis, B a; cum eis, Fou. 7. 22: parce. cui ego, BM Va; puree inquiens. cui ego, Fou. In both these places F has the wrong reading, in the second instance by reason of interpolation. The readings of Fou in each case could not have existed in x, at least at the time when B was copied, but •Oodex Urbinas lat. 1153 (u) and Ottobonianus lat. 1965 (o) are selected as repre- sentatives of this class. See Merrill Selected Letters, Introduction. 474 Frank Egleston Kobbins they and many others like them 1 must have arisen either in F itself or between x and F. o and u then must have been influenced by F itself, by an ancestor of F less remote than x, or by a descendant of F. Fortunately this question can be satisfactorily answered. The address of iii. 18 reads as follows: seruo, B a (context) ; aduir- ium seuerum, B (index) ; curio seuero, F in context, but in margin, uirio seuero; curio seuero, ou; seuero, M V. The marginal reading of F was inserted as a rubricator's guide, and at the present time can easily be mistaken, as the rubricator did in fact mistake it, for curio seuero. This error was made in F and did not exist in the ancestors of F: therefore o and u must take curio seuero from F itself or from a descendant of F, and not from an ancestor of F. Exactly similar is iv. 28, where the title reads seuero, B (context); aduibium seuerum, B (index); iubio seuero, F in context, but in margin, uibio seuero; iulio seuero, u; iunio seuero, o. The readings of o and u are not identical with the erroneous iubio of F, but they must be corruptions or corrections of iubio rather than of the true reading. Certain additional facts, however, demand explanation. For instance, the heading of iv. 24 is fabio F; ualenti B (context), M V; fabio ualenti, ou; adfabium ualent, B (index), and it seems that o and u drew from a source older than F. This, however, cannot be true. With but few exceptions the MSS give the cognomen in addresses, when but a single name is used; but in certain cases and under certain circumstances the nomen is given — as here in F. Now the " eight-book " MSS in their own proper tradition apparently had few double names used in addresses, but took most of them from the BF family, and in iv. 24 it is altogether probable that the unadul- terated " eight-book " text would have ualenti, reading the cognomen like the other MSS. The scribe has simply combined his own reading ualenti with fabio, which he found in F or in a descendant of F. There is, therefore, apparently nothing in the relations between F and the " eight-book " MSS that makes necessary the intervening step y. J E.g., 8. 1: nondum ab, Ba ; nondum enim ab, Fou (enim above line in first hand, F); 8. 16: haesitabundus, BMV; haesitabundus inquit, Fou ; 8. 25: fortius amove, BMV a ; amove fortius, Fou. Relation between Codd. B and F of Pliny's Letters 475 The most reasonable conclusion, then, seems to me to be that B and F are sister MSS. Their scribes were puzzled by the same passages, where the appearance of their archetypes would seem to have been identical; 1 their lacunae agree; all apparent discrepancies are explicable; and in their text the two MSS agree to an astonishing extent. Besides this, there is no good evidence for any other state of affairs. Therefore, although it may be admitted that Fig. 2 is a possible explanation of their relationship, I am inclined to favor the simpler relationship indicated in Fig. 4. The University of Chicago November, 1909 1 See, e.g., 120. 19, 100. 2, 91. 21, 10. 19, 85. 3, 35. 5, 11. 4, 102. 5, all cited above.