THE MANUSCRIPTS OF APPIAN

IT is now well established that the Roman History of Appian was originally embraced in twenty-four books which he divided in the following order:²⁹ --

ROMAN AFFAIRS

ROMAN AFFAIRS

This so voluminous work early met a fate corresponding to its bulk. Undoubtedly the Byzantine age, impatient of reading and transcribing books in general, and especially of an author not very remarkable for art or genius, and who had severed the connection of the events themselves by a bad plan, preferred an immediate enjoyment of selections to the trouble of continuous reading. Their rules of dismembering an author were two, by one of which they extracted certain special passages from the whole work and brought them together according to their resemblance of subject; by the other they selected those entire books which seemed to them more important than the rest, and these alone were circulated by copying. To this rule, often salutary and oftener pernicious, are due whatever of the fragments of Appian were collected in the Constantinian extracts, and it is truly astonishing that the compilers of these extracts were acquainted with, or gave attention to, only one volume of his works, the one containing the first nine books. On the other hand, it seems to have been in consequence of the importance of the subjects and of the multitude of readers thence arising that, besides the Preface and the Celtic epitome, the Spanish, Hannibalic, Punic, Illyrian, Syrian, and Mithridatic histories, and the five books of the Civil Wars, escaped entire the slothfulness of those ages.³⁸

The remains of the whole work, which in one way and another were preserved, became known late, and little by little, to learned men, and still later began to receive attention from them. The beginnings and gradual progress of Appianic criticism have already been described in great part by Schweighuser in his edition of Appian (vol. i. p. iii seq.) and in his opusc. acad. (vol. ii. p. 97), but it will be pertinent to give a new sketch of it, which, while more brief, will in several particulars be more correct; especially since it will naturally introduce us to the contributions which are peculiar to the present edition.

In the early times of the revival of learning it was not seldom the fortune of Greek writers that before they were presented in their own tongue, or could be, they were brought out clad in Latin, or something like Latin, garb. This fate overtook Appian also. He found a translator in Peter Candidus December,³⁹ master of correspondence to Pope Nicholas V., not a new hand at the translation of Greek books, but always a rude one. About the middle of the fifteenth century he made a translation of those works which were then ascribed to Appian, which translation, both in manuscript (of which a number of copies still remain) and if print, was usually embraced in two parts, thus: --

II

II

I

I

That this translation of Candidus was not only uncouth, but in many places utterly misleading, is the unanimous voice of scholars. It has, nevertheless, been universally esteemed till now, and reckoned among the sources for the emendation of Appian, because it was believed to have been based upon a good copy that was afterwards lost. Although this general view is correct, it will be possible to define it with more precision than has yet been given to it, and for this reason I shall return to it later.

Although this work of Candidus, such as it was, was frequently reprinted from the year 1472 onward, the Greek text remained unpublished until Charles Stephen, the uncle of Henry Stephen, published the Preface, the Celtic epitome, the Punic wars, the Illyrian fragment, the Syrian [Parthian] and Mithridatic wars, and the five books of the Civil Wars, arranged in that order, at Paris, in the year 1551. The sources and critical method of this edition were first made plain by Schweighuser in his preface (p. vi. seq.), showing that it had proceeded from the two Paris MSS., 1681 and 1682 (designated by Schweighuser as Reg. A. and B., and by myself as a. and b.), and still remaining in Paris; but that it did not follow the text of these MSS. closely, and did not observe the same order of arrangement of the books. For the order of arrangement in these "Royal" MSS., as Schweighuser styled them, was the following: --
Preface
Celtic Epitome
Syrian
Punic
Illyrian fragment
[Parthian]
Mithridatic
Civil Wars, V.

And finally he showed that their goodness, or rather their badness, was about equal, and that their age and origin were about the same, i.e. the beginning of the sixteenth or end of the fifteenth century. That all this was true I perceived as soon as I had these Paris MSS. in my own hands, and also that that distinguished man had left nothing undone to add to the accuracy of his collation.⁴⁰

This editio princeps⁴¹ was followed by a separate edition of the Spanish and Hannibalic books which Henry Stephen obtained on the occasion of his journey to Italy, from Arnold Arlen, and which he published, together with some fragments of Ctesias, Agatharcides, and Memnon, at Geneva in 1557. That the copy of these books of Appian was very faulty Henry Stephen himself lamented, and his edition testifies.⁴²

Although the Stephens had limited their labors to the books of Appian which had been preserved entire, a little later the fragments which had been taken from the first nine books and thrown together in the store of the Constantinian collections began to find editors. Of these, F. Orsini published at Antwerp, in 1582, the one entitled De Legationibus, he relying upon the Vatican Greek MS. No. 1418, and the Neapolitan III. B. 15, but reproducing the text with too little accuracy.

Following this came a second edition of Henry Stephen, published at Geneva in 1592, to which he added the Spanish and Hannibalic books recently issued by himself, but to this edition he applied no other aids than a talent often happy but often wide of the mark. So it came about that in this edition also he was able to give of the Illyrica (which Candidus had entire in his Greek copy) only the fragment preserved in the Royal MSS. Seven years later David Hoeschelius found in a MS. of Appian then at Augsburg (mentioned on page 67 of Reiser's "catalogue of the Augsburg library," published at that place in 1675), but now of the Munich library, Gr. 374, the Greek text of the Illyrica entire, and gave it to the light at Augsburg in a separate book. After this Henry de Valois, a man of rare learning and industry, merited well of Appian by bringing out, in 1634, a considerable number of extracts from Appian preserved under the Constantinian title De Virtutibus et Vitiis from a codex belonging to Peiresc. For these reasons little praise is due to Alex. Tollius, who superintended a republication at Amsterdam, in 1670, of the relics of Appian's work; for besides being scarcely acquainted with Greek he had no knowledge of his predecessors who were his own superiors, and so it came about that the passages published by Orsini and Hoeschelius are not contained in that edition.

Such were the labors of the men, learned and unlearned, who contributed to the study of Appian before Schweighuser. He alone was of more service than all the rest. Certainly when he found that, for revising those books of Appian which remained entire -- for it was plain that the condition of the excerpta was deplorable and almost beyond hope of amendment--it was impossible to rest satisfied with those Royal MSS. of Charles Stephen or the Italian one of Henry Stephen, he resolved first to strengthen the very basis of his emendations. Nor was this a vain attempt, for in addition to some other small helps, he obtained this threefold apparatus for emendation:
Vatican Gr. 141.
Preface, 12th Century
Celtic epitome, 12th Century
Spanish, 11th Century
Hannibalic, 11th Century
Punic, 11th Century

Laurentian LXX, 26, 15th Century.
Spanish
Hannibalic

Augsburg class⁴³ (0), 15th Century.
Preface
Punic
Syrian
[Parthian]
Mithridatic
Civil Wars, V.
Illyrica entire

These MSS., to which were added the Royal MSS. (class i) newly collated by himself, and the translation of Candidus, he so used that he brought in requisition for the Preface the Vatican MS. 141, and those of classes O and i; for the Celtic epitome, Vat. 141 and i; for the Spanish and Hannibalic books, Vat. 141 and the Laurentian edition of H. Stephen; for the Punic wars, Vat. 141, O and i; for the Illyrica, O, the fragment in the Royal MSS., the Leyden codex, and the translation of Gradius,⁴⁴ and finally for the Mithridatic and the remaining books, O and i; but in such a way as to give a preference over the Royal MSS. whenever possible, to class O, which he saw was much superior. Finally he made constant use of Candidus, who, in common with O, had the Illyrica entire, but regarding him as a source of much less authority than O.

When Schweighuser's materials had been augmented by the welcome addition of the notes of Musgrave and Reiske, in whatever way one man could earn distinction by the restoration of a text deformed by extraordinary corruption either by the choice of MSS. or by long-continued study of the writer, or finally by a genius for emendation -- he distinguished himself abundantly, and relieved his successors of the greater part of their labor. Yet even then it was impossible but that some faults should be left which it were desirable to remove; for besides the fact that not all the codices that have some weight were within his reach, and that the friends to whom he had entrusted the task of making collations did not, in all cases, perform that task with scrupulous fidelity, what was of more consequence, he did not himself arrive at a true estimate of the codices. Although this renowned work is lame in these particulars, no one after Schweighuser has concerned himself with the emendation of the very words of Appian on a settled plan and guided by new lights. For the editors who have succeeded him -- they are, beside Teucher, who did not distinguish himself, the Didot editor, Dbner, and Bekker -- have satisfied themselves with inserting in their editions a few small fragments of Appian embraced in the Constantinian title De Sententiis, and almost limited the remainder of their work to a repetition of Schweighulser's text, although Bekker must not be deprived of the praise due him for some notable emendations and for an improved punctuation of the whole.

We may now pass to our opinion of the relations of the codices and inquire whether perchance any connection or kinship exists between the five classes, or whether each one must be considered, as Schweighuser thought, to have its own authority and origin, distinct from the others. These classes, we repeat, were the following:--
I

Vat. 141

Vat. 141

Laur. LXX 26 and H. Steph. book 15th Century

Laur. LXX 26 and H. Steph. book 15th Century

August. class (O) 15th Century

August. class (O) 15th Century

Cand. (C) 15th Century

Cand. (C) 15th Century

Reg. class (i) 15th and 16th Centuries

Reg. class (i) 15th and 16th Centuries

As it was not credible that this large number of codices, for the most part carelessly examined, should not show distinct and decided marks of relationship of the MSS., it was the twofold duty of a new editor that, in selecting the codices, a sound rule should be observed, and that the account of those selected should be as exact as possible, especially of those that Schweighuser had not himself handled. Whence sprang the necessity of narrowing, as well as of enlarging, our critical apparatus. The necessity of narrowing it I could not but perceive clearly as soon as I touched the oldest one of all, the Vat. 141.⁴⁵ This codex, however, is not of a uniform type. Without doubt the first eight leaves, containing the Preface and the Celtic epitome, did not originally belong to this codex. They were written in the twelfth century, the others in the eleventh, and, what is of more importance, the quaternions are reckoned from the ninth folio on; and it is evident that the plan of these two small parts is not the same as of the others, i.e. of the Spanish, Hannibalic, and Punic books. For those books, therefore, I have set these codices in the lowest place of authority, although they are all derived from that very Vatican 141. It is perfectly evident from differences in the text that neither Candidus's Preface nor that of the Augsburg MS. go back to V, although i has come from that source. Therefore, in the study of these few pages, V and O C had to be compared. On the other hand, in the Celtic epitome, which is omitted in O, there was no doubt that C, no less than i, was of no authority as compared with V, and although it is doubtful from what source Candidus drew his knowledge of the epitome, I am inclined to believe that it was from i.

By this means a more certain judgment can be formed of the other three books which V contains. That the Laurentian codex lxx. 26, and that of H. Stephen of the Spanish and Hannibalic histories in their farthest origin⁴⁶ go back to that same V, or to some twin copy, my own collation of the Vatican proves, since the work of Spalletti, which Schweighuser was obliged to use, had been very negligently performed. Therefore, while both manuscripts had to be rejected, yet sometimes, as it happens, some reading from Stephen's text not attested by evidence, but conjectured with felicity, had to be adopted, although too often it was far from clear what the book had contained and what the editor had inserted by way of conjecture.⁴⁷

While it is quite plain, therefore, that the Spanish and the Hannibalic books should be edited from Vat. 141 alone, the same ease and clearness of proof exist also respecting the Punic book. For example, in this Vatican book chapters 56-59 of the Punic wars have disappeared in a vast lacuna, and the same lacuna exists in all the other codices (i and O and C), and could not be filled except by inserting that scrap embraced in the Constantinian excerpta of the Embassies of the Nations to the Romans. But this lacuna in V. 141, much the oldest of all the MSS., as I have said, is not as Schweighuser thought (iii. 426), primitive. It did not proceed from its original, but was due to accident and chance, for while the fifteenth quaternion ought to have four pairs of leaves in the following order: [figure in text: Diagram of leaves in fifteenth quaternion]

the middle one (4 and 5) has fallen out by accident, as will immediately appear to any one examining it with his own eyes, as Schweighuser did not. Since, then, it is quite certain that this lacuna had its origin in V. itself and not in its original, the Punic books of all the remaining codices which have this lacuna must have been transcribed from this very same Vatican, or, to speak more accurately, from copies derived in the course of time from it. O, C, and i therefore cannot have any superiority in the Punic wars over V. itself. What they have peculiar to themselves and differing from V. are blemishes, or interpolations, or corrections of errors, usually slight ones; for he would greatly err who should attribute the not very easy amendment of certain passages to some original copy, rather than to the happy genius of certain learned copyists of whom there was no lack. By virtue of these emendations, therefore, and not of the codices themselves, it has come to pass that in some places mention of copies of the Punic wars has been found setting them almost above V. in point of honor.

Although in the Spanish, Hannibalic, and Punic histories there is no manuscript authority before Vat. 141, another help exists in the compilations of Constantine and the themes drawn from them by Suidas -- an assistance in some degree for the Spanish and Punic entire books, and the sole help (except the fragments preserved in Bekker's Anecdota) for the other first six books, i.e. the Kings, Italic, Samnite, Celtic, Sicilian, and Macedonian. I have deemed it not the least part of my duty to give an account of these compilations somewhat more copious and certain in this edition than has been done before. To touch each one briefly--there are two parts of the title treating of the Embassies, one written "concerning ambassadors (not embassies) of the Romans to the nations"; the other "concerning ambassadors of the nations to the Romans." After Orsini had published that part from the Vatican Gr. codex 1418 and the Neapolitan III. B 15 with such carelessness as I have described above, Schweighuser acquired the Munich codex 185 -- which he calls Bav., and I call M -- containing the Embassies of the nations to the Romans, and from it he extracted quite accurately not only the fragments of the first six books, but he added also, most advantageously, scattered scraps from the Spanish and Punic histories evidently neglected by Orsini. That the account of this part, by far the most important of all, might be established with certainty, I myself collated afresh this M 185, and I obtained by the courtesy of Alfred Eberhard, a new collation of the Neapolitan MS., from which the carelessness and almost fraudulent license of Orsini fully appeared; and finally I examined the third, the one considered the best of all, the Ambrosian, N 135 sup.⁴⁸

Of the second and much shorter part of this title, which contains the Embassies of the Romans to the nations, I have myself afresh gone over the codex Vat. 1418, and also the Munich codex 267,⁴⁹ and I have added what few fragments exist in either, extracted from the Spanish and Punic histories.

Since all the careful study that rests on manuscripts has been already applied to the fragments of Appian in the compilation entitled De Legationibus -- and it is hardly possible for farther help to be found -- it has been possible to accomplish scarcely anything in respect of the other compilations which have to do with this writer. For in the book of Peiresc containing the copies of "Virtues and Vices," although H. de Valois bestowed admirable and most scrupulous work upon it, nevertheless he also, as it were, knowingly and designedly passed by everything that he found that had been previously published. That these passages were not few I inferred from the preface of M. Gros's edition of Dio Cassius, vol. i. p. lvii, and so I applied by letter to Julius Wollenberg, by whom I knew that this whole codex had been freshly examined; but in vain, for that distinguished man had died a little before. So I was obliged reluctantly to rest satisfied with the very few excerpta that de Valois himself had extracted from it at the end of his book, p. 125.

Nor did my efforts succeed better in the extracts De Sententiis, although there were so few from Appian among them that the damage is rather slight. The things found by A. Mai in cod. Vat. Gr. 73 and inserted in his new collection of ancient writers, v. ii. p. 367, are so affected by Mai's doctoring that I have scarcely been able to gather anything, or to trust Mai for anything except the trifles which Herwerden has ingeniously extracted from him.

Such, then, is the critical apparatus of both Schweighuser and myself for reviewing the remains of the first nine books. For the remaining whole books, the Illyrica, -- which seems to have been rather early torn from the Macedonian and should accordingly be connected with the latter part of that book, -- the Syriaca, the Mithridatica, and the five books of the Civil Wars, there were, as I have said above, three kinds of codices employed by Schweighuser; namely, the Augsburg, class O, the translation of Candidus (C), and the Royal i. He gave the foremost place to O, constantly referring to C and i, but not following them except in case of necessity. The painstaking man was quite correct in this preference, for there is scarcely a chapter in which the supreme merit of O does not outshine the rags and tatters of C and i. Yet it will not be amiss to define a little more precisely than he has done the distinctions of value between the three classes, especially as it is not possible to dispense entirely with C or with i.

O and C betray in two ways a certain primitive relationship, by having in common the Illyrica entire and by the order of arrangement of the books. (I will speak of the Celtic epitome later.) While O and C eventually go back to the same fountain, yet they deviated early, as is shown by certain lacuna which each of them supplies for the other and by the general differences of the text. This difference, even if it must be attributed to the inexperience of the translator himself, nevertheless remains always very great, and its nature is such that one ought for the most part to follow O and reject C. What almost every page of Schweighuser and of this edition teaches, it seems unnecessary to confirm by examples.

The general merit of O being sufficiently accepted, the relations of the several codices belonging to that class must be explained. Three sound ones belong there, namely: --
A. Munich Gr. 374, formerly Augsburg, on cotton paper, 382 small leaves, fifteenth century, written by many hands nicely indeed, but mingled with very many errors of orthography. Many correcting hands appear.
B. Venice, St. Mark's library, 387, paper, 390 square pages written in the year 1441 by a certain Gedeo. Described in the catalogues of Zanetti and Morelli. Cited by Schweighuser as Ven.
V. Vatican Gr. 134, cotton paper, square form, 318 pages written by many hands in the fourteenth and fifteenth centuries, for the most part carefully. The Appiana are written on pages 125-318; the previous ones are occupied by the Roman Antiquities of Dionysius of Halicarnassus, Books VI.-X., after which, on page 124, come some iambic verses of J. Eugenicus to the emperor John Paleologus. The Appiana are corrected only by the first hand. This is cited by Schweighuser in the Mithridatica as Vat.

Additional to these, containing the Civil Wars, ii. 149-154 and iv. 1-5 2:
F. Paris, Reg. 1672, parchment, largest size, 944 pages. Compilations of Appian begin on page 937r, written at the beginning of the fourteenth century. The first part, embracing Plutarchea, I had attributed to the thirteenth century, but more correctly now M. Treu attributes this also to the fourteenth. I have myself collated it.
E. Paris, Reg. 1642, paper, of various sizes, fifteenth century. The Appiana are transcribed from a copy very similar to the preceding and by a poor use of the codex. Schweighuser collated it with sufficient care. It is Reg. C. in his list. I afterwards reread it myself. Each of these two codices was at Constantinople until the year 1687. See Bernhardy's History of Greek Literature, 1 4, p. 744.

Now, to drop further discussion of E, F, Schweighuser saw that A, B, V, were transcripts of the same original, but he thought that the greatest confidence in the matter of copying should be placed in A. The conscientious man examined this whole MS. himself. Of the Venetian one he had only the collation of Blessing, and finally it must be that he had nothing of Vat. 134 except the Mithridatica carelessly treated by Spalletti. On the other hand, what is not difficult to understand at present is that this very Vatican MS. has been most carefully, and the Aug. and Venet. have been rather less carefully, written. So I have collated the whole of it and have also rexamined the Aug. collation of Schweighuser; and finally, for the Venetian, which for lack of time I have not been able to examine, I have contented myself with drawing from Schweighuser's⁵⁰ materials.

So much for the first group of codices. Proceeding to the remainder -- that there can be no doubt that the merit of Candidus is much inferior to O we have already signified. Nevertheless it will be worth while to explore a little more carefully the sources and as it were the incunabula of the translation made by him. From what has been said above it is clear that he used some Greek copy which belonged by contents and order of books to class O. As he worked at Rome and from Roman copies, one of which (Vat. 134) belongs to class O, I had for some time (while I was first collating that book) the same belief as Schweighuser, that I had discovered the copy used by Candidus. But that was an error. Besides other things less important, not only did numerous readings at variance with the Vat. oppose this conclusion, but certain lacun in the Vat. were filled by Candidus. On the other hand, the copy of Candidus could not have been of the class Reg., for, to say nothing of all other matters, the very order of the books agreeing with O excludes the Reg. unless perchance the Celtic epitome was drawn from it by him (see below). So, as far as I can see, only two assumptions are possible. Either Candidus had that Vat. 134, and besides it other books belonging to class Reg. and free from those lacun, or he used a book similar to Vat. 134, but not that book itself. And for some time I was induced to adopt the former assumption by an unpublished memorandum in the Medicean collection concerning the translation itself, which was kindly communicated to me by neas, Count de Piccolomini, professor at Pisa, and which I afterwards examined in person.⁵¹ It is as follows:--

"per procuratione Nicholas V.

"Dear son. Health and apostolic benediction. We learn that in the Florentine library of St. Mark, which by your diligence you have made illustrious with Latin and Greek books, two volumes are to be found written in Greek by Appian of Alexandria, on one of which is the inscription: 'Third book of Appian concerning Roman affairs,' and on the other 'Appian concerning Italian history.' Since we have received from elsewhere books in Greek of this same Appian and we greatly desire to see in Latin what that man considered worthy of the memory of posterity, we have ordered that they be translated. But since a truer understanding and a more faithful interpretation can be obtained from the reading of several volumes than from the inspection of one (since what is wanting in one the other may supply), we have thought to beseech your nobility to send us those books as quickly as possible. Immediately after using them we will return them to you to be put back in their own place. Since you have always been most compliant to our wishes, we are sure from your kindness that you will gratify us in this matter also.

"Given at St. Peter's in Rome under the seal of the fisherman, the 7th day of December, 1450, and the fourth year of our pontificate. P. CANDIDUS.

"To our dear son Cosmo de' Medici, citizen of Florence."

Since, then, this letter written by Candidus with his own hand was certainly very important, I began at once to inquire what these two Appian codices could be which Nicholas desired to be supplied to himself by Cosmo. And now, since there are three altogether in the Laurentian library, it was manifest that MS. lxx. 26, containing the Spanish and Hannibalic histories could not be taken into the reckoning, since Candidus did not have those books. Therefore it was easier to find the "third book of Appian concerning Roman affairs"; for, in fact, codex lxx. 33, belonging to the family Reg. and containing the three last books of the Civil Wars has, in an ancient hand, this writing on the blank leaf before the first page: "Third Book of Appian concerning Roman affairs." Besides this, Nicholas asked for another volume inscribed "Appian concerning Italian history." This one may have been the Laurentian codex lxx. 5, this also belonging to the family Reg. and containing all that they contain, to which, in fact, the title "Appian's Italian history" was prefixed.⁵² Whence it might seem quite certain that besides the Roman books, others also and those belonging to class Reg. were employed by Candidus in his translation. Without doubt they were sent to Rome. But that the good man did not everywhere compare them with his own copy, but thought it sufficient to trust to one codex, the tenor and form of his whole translation convince me.⁵³ To embrace the matter in one word, not only are there in the readings given by Candidus very many which differ equally from Reg. and Vat. 134, but also he has peculiar lacun of his own which do not exist in either of them, of which a most remarkable example is Civil Wars, iii. 22, which he has omitted entirely.⁵⁴

But if Candidus cannot be supposed to have made his translation from Vat. 134 alone, nor from that amended by other books joined to it, it follows that he used one codex and that one of a lineage of books approaching O. But the Celtic epitome, the translation of which I have already said differs in no respect from Vat. 141, nor from the i class which were transcribed from it-- you will probably judge was known to Candidus from another copy of the i kind. For its very position seems to prevent us from believing that it could have been in the copy from which he drew the rest.

In using his work, I have followed in the main the example of Schweighuser; that is, in no place have I failed to inspect Candidus, and I have not put confidence in him except where O failed me. Wherever mention is made of him it refers to my copy published at Venice in 1477, except that in the Mithridatic wars, 11--121, it was better to rely upon that published by Michael Vascosan, Paris, 1538. The utility of the more recent editions is almost the same as of the most ancient, since they differ in no respect except in some trifles of orthography.

The argument concerning the codices has now resolved itself into this, that whereas before Schweighuser the i class held the sole place, it was by him justly consigned to the lowest place. Wherefore, not unwilling to spare my eyes and my strength in a paltry matter, I have acquiesced in the labors of my predecessor. Yet when by way of experiment I had examined one, the Laurentian, lxx. 33 (f), parchment, fifteenth century, 111 octavo pages, containing Civil Wars, iii.-v., I gained scarcely anything new that was not in Reg., and the same happened to Schweighuser in the case of a certain Breslau codex, which seems to be the first part of this Laurentian, since it does not contain those last three books of the Civil Wars.⁵⁵ That both of these are somewhat more carefully written than the Reg. is of no moment, since the same badness is common to all of them. So, as I did not care to accumulate rubbish, whatever other books of that class⁵⁶ I knew of, I rejected, and I have been satisfied with the Parisian a and b of which I have spoken above, and Schweighuser's report upon the Breslau MS., and my own test of the Laurentian. Let no one think, however, that this whole class should be despised, since some blemishes which have crept into O and C are not in it and some lacuna have been filled.

These materials which I have just enumerated have been employed in this new revision in such a way that, except manifest errors of book makers and orthographic minuti, all the apparatus has been referred to in proportion to the authority of the several classes of MSS. Upon the emendation, for which there was great room even after the collation of the best books, not only did I myself labor according to my powers, but August Nauck did splendid work. For while he, by a rare example of friendship, voluntarily shared with me the task of correcting the proofs, he wrote down a great many emendations such as might naturally come from a man of so great genius and learning. If I affirm that I am under the greatest obligations to the kindness of this illustrious scholar, I shall say too little. In respect of using the things found by him Nauck lays down the rule that whatever I adopt I shall adopt at my own peril, because they occurred to him while he was reading, since it has not been his fortune to have been deeply acquainted with the editor's author. So if from Nauck I have either accepted or adduced anything at variance with Appian's peculiarity of diction, the blame of the error must fall upon me. I have applied or sought to apply the same caution also to my own conjectures, for this writer employs so unusual and perverse a style of discourse that often you do not venture to decide for certain what is in accordance with his style and what is not.

It only remains that I testify, in this place also, my sense of obligation to Charles Halm, by whose kindness the Munich codices were sent to me at Leipzig, and to Alfred Eberhard, to whom I have said I am indebted for the collation of the Neapolitan MS.; and finally to Rudolph Schll, who kindly imparted to me some emendations of Nipperdey.

DORPAT, January 1, 1879.

BIBLIOGRAPHY

BIBLIOGRAPHY