In addition to these real sources of knowledge, the fabricated codices of Bosius were till quite lately believed in, and carefully regarded in the arrangement of the text. Just as Henri Estienne (the famous Stephanus) vitiated the criticism of Euripides by recommending his own usually excellent conjectures by the authority of imaginary mss, so this other and almost equally able Frenchman, Simeon Du Bos, a native of Limoges, born 1535, imposed on the most learned men of three centuries with his imaginary Decurtatus and Crusellinus, and his pretended or falsified citations from the really existing Z. Even Orelli was deceived by the imposture. Indeed the great critic of Zurich would probably have left little to be added by his successors had he been aware of the fictitious character of the codices of Bosius. And yet Bosius' own account of the manner in which he gained possession of his vetustissimi codices might have excited suspicion. His Decurtatus (commonly quoted as Y) he obtained from a private soldier who had rescued it in the sack of a monastery, in which it had been deposited. Of his Crusellinus (X) he does not tell his readers the source in his edition published at Limoges, 1580, but he gives the following rather vague details:-adiutus sum praeterea codice quodam excusso Lugduni qui olim fuerat Petri Cruselli, medici apud nostrates celeberrimi; ad cuius libri oras doctus ille vir varias lectiones appinxerat, a se, ut ipse dicebat, diligentissime et summa fide e retustissimo et castigatissimo libro Norioduni descriptas. The imposture, however, escaped detection for nearly three hundred years, and it was not until the year 1855 that Maurice Haupt discovered that no such mss as the Decurtatus (Y) and Crusellinus (X) of Bosius ever existed. The discovery of Haupt acquired the certainty of a demonstration when Mommsen found that a ms deposited in Paris contained the rough draft of Bosius' notes for the last seven books of the Epp. ad Att. On comparing these with the published commentary of Bosius, Mommsen found that Bosius had frequently ascribed one reading to the mss in his first draft, and another in the published commentary. In each case he recommended his own conjecture by the authority of the fabricated ms; and in some cases he changed his view of a passage in the time intervening between the first draft and the ultimate publication, and accordingly changed his account of the reading of his ms. For instance, in Ep. ad Att. x. 6, 2,"Bosius in his published edition reads De Quincto filio fit a me sedulo; on which he states that his Codex Decurtatus has de Q. F., and his Crusellinus, de Q. filio. In his unpublished première ébauche, found by Mommsen, he had given de Q. frat. as the reading of the Decurtatus, adding 'Victorius legit de Q. filio, quam scripturam in meis non reperio.' Baiter certainly transcends even the asperities of the Brunckian age, but does not (at least in his own opinion) transcend that emphasis of expression which is warranted by the case, when after narrating the circumstances just referred to he adds, 'Bosium cito scelus suum morte luisse a latronibus trucidatum.'* To the above sources of information may be added (10) A (Codex Antonianus) and (11) F (Codex Faërninus), in so far as their readings are reported by Malaspina; but these mss must be viewed with some suspicion. We cannot be sure that we have not in Malaspina something of the Bosius, whom he rivals in the brilliancy of his conjectures. The title of the work of Malaspina (which is extremely rare) is, Malaspinae emendationes et suspiciones in epistolas ad Atticum, Brutum, et Quintum fr.; it was published in Venice in 1563, 4. (12). The most ancient editions are the editio Romana (R), published at Rome in 1470, and the ed. Iensoniana, published in Venice It is a strange coincidence that these two Frenchmen, Estienne and Du Bosboth scholars of the most profound learning, and of almost unsurpassed brilliancy in emendation should have stooped to a fraud to establish conjectures, which in some cases were so good that they must have been almost universally accepted as certain conjectural emendations, if candidly put forth as such. It is curious, too, to observe the different way in which scholars have received the disclosure of each of these impostures. Since Haupt and Mommsen proclaimed the disingenuousness of Bosius, the subsequent editors have vied with each other to swell the chorus of obloquy, and none so poor to do him reverence.' We hear of nothing but 'fraus et fallacia,' 'mendacium fraudulenti hominis,'audacia et perfidia'; and Baiter goes so far as to say 'furca igitu expellendae quotquot lectiones a sola Bosianorum codicum auctoritate pendent.' But Stephanus has been more fortunate. Kirchhoff has proved the non-existence of his veteres codices Italici as clearly as Haupt has proved the non-existence of the X and Y of Bosius. Yet Dindorf has not a word to say about his imposture; and Mr. Paley charitably gives Béλos on Bacch. 25, as Stephanus' correction of μéλos of the mss, without hinting that he alleged for it substantive ms authority; and on v. 1060 the same justly valued editor puts forward о μóеwv as the conjecture of Musgrave for boo vov, adding that Elmsley tells us that Porson approved oτоι μółwv-as if Musgrave would have proposed, or Porson approved, such a reading, if they had known that półw was nothing more than a guess of Stephanus; that the ms to which he ascribed it never existed; and that all his statements about it, and his quotations from it, are mere fictions. in the same year. These are founded on M, R giving generally the reading a prima manu, while I, as a rule, presents the marginal or superscribed corrections. For other ancient editions, see list appended to Adnotationes Criticae. The estimate of the value of the mss of the letters ad Familiares put forward by Orelli in his masterly preface to his third volume (2nd ed., Zurich, 1845) has been largely modified by recent discoveries, as I have already shown. His estimate of the relative value of the mss of the letters to Atticus, supplemented or rather corrected by the discovery of Haupt, remains unshaken. But there has been one distinguished recusant. Wesenberg, in his editions of the Epp. ad Fam. and Epp. ad Att., published by Teubner in the years 1872, 1873, and in his Emendationes Alterae (1873), holds a view which, if he could recommend it to editors, would revolutionise the criticism of these letters. The mss on which we must depend for our knowledge of the letters to Atticus, Quintus, and Brutus, are thus placed in the order of merit by Baiter and all post-Hauptian editors (1). M. (3). W. (4). Zl (i. e. Z as reported by Lambinus, not Bosius). Wesenberg thus estimates the respective values of the mss(1). Z and v. c. (vetus codex' Lambini). (2). C. and W. (3). A and F. This startling revolt from received opinion he does not defend, hoc meum iudicium (he writes, Epist. vol. ii. 1872) accuratius hic confirmare necesse vix est: satis, opinor, erit lectores si qui talia curabunt ad varietatem scripturae textui a me subiectam reiecisse. Afterwards, in the preface to his Emendationes Alterae, he addresses himself with a little more energy to the task of proving that A and F are more valuable authorities than M. His method is straightforward; he gives a large number of cases where both A and F present (in his judgment) a better reading than M; a very large number of places where either A or F is superior to M; and a large number where A and F are inferior; and his conclusion is, Apparet paucis locis, mea quidem sententia, constare Antonianum et Faërninum Mediceo postponendos, permulto pluribus illos huic anteponendos, aut certe dubium inter eos esse certamen. Meo igitur iure rideor mihi duos illos codices vel cum aliis conspirantes vel solos multis locis secutus, ubi Baiterus in Medicei scriptura acquieverat (Praef. p. iv.). On this whole method of ratiocination it may be at once observed, that it does not really tend toward the proving of the proposition which it professes to establish. Though one were to agree with Wesenberg in every instance in his estimate of the relative values of F, A, and M, we could not be certain that the readings of F and A were not the conjectures of Malaspina or other scholars. On the same method an irrefragable case might be made for the superiority of X and Y, mss which never existed. Elmsley's Bacchae would contrast very favourably with the Codex Palatinus, but his readings have no authority but what they derive from the arguments by which he supports them. However, I will accept the method of Wesenberg, and examine a few of his loci; to examine all would be tedious and fruitless :(a) Att. ii. 1, 5, deducerem is M; deduceremus FA; evidently an editorial correction. (b) Att. ii. 22, 1, where the emended reading is opes eorum et exercitus, M gives opes et vim exercitus, FA opes eorem et vim exercitus. Cicero probably did write eorum, but this is an obvious correction: the expunging of vim is not so obvious, and accordingly tim of M is preserved. ... (c) Att. iii. 23, 1, attulisse scribis M, attulisse perscribis FA. This seems to me very like a conjecture. Cicero had written a few lines before eaque • perscribis. The commonest error of an unscientific editor, or a copyist who assumes to himself the function of an editor, is to introduce into his text such uniformity of diction as this. Baiter is too much disposed to this course: see note on Ep. ix. (Att. i. 4, 3), ea nondum vidi: because nondum occurs here, Baiter insists on correcting non to nondum in nos ea non vidimus in the preceding letter, as if because Cicero says, 'I have not yet seen them' in one place, he should not be permitted. to say, 'I have not seen them,' in another. (d) Att. v. 4, 4, vellet iam M; velis et iam FA; vel etiam is probably the true reading, which M very nearly preserves, though the words are wrongly divided. FA on the other hand have recourse to an obvious but unsatisfactory conjecture. (e) Att. v. 21, 4, fore venturum M, esse venturum AF; the latter is certainly right; in my mind a certain correction of Malaspina. (f) Ibid. 6, sed mirifice abstinentem M; sed mir. etiam abs. AF; another conjecture, in this instance wrong, in my opinion. (g) Ibid. 7, erogabatur M; erogabitur AF. Here again erogabitur is certainly right; but it is significant that this reading, as well as the two foregoing and (a), is quoted as the reading of his pretended mss by Bosius. (h) Att. vi. 7, 2, nisi eum graviter aegrum esse reliquissem M: esse is omitted in FA; what could be a more obvious correction? Yet, as is often the case, the obvious cure is wrong. Issi for esse, the brilliant emendation of Manutius, is now generally accepted. These passages, which are taken almost at random, being selected chiefly for their brevity, will serve for a sample of the readings adduced to show the superiority of FA over M as ms authority. To me they seem only to prove that FA have in each case reproduced the very obvious correction of an obvious blunder in M, which correction is sometimes right, sometimes wrong, but always the first that would occur to any editor or copyist who assumed the functions of an editor. And it is in my mind significant that in many cases this correction is the same as that quoted from his fictitious mss by Bosius. The same may be said of the places in which M is excelled by either A or F, for instance: (a) Att. ii. 6, 1, Romae fuisse M; Romae me fuisse A; and the Bosian Y. (b) Att. ii. 7, 3, quod potest M; quoad potes A; an easy but wrong conjecture, arising from ignorance of the impersonal use of potest, which is so common in the letters as quite to form a feature of their style: see note on Ep. xxxvi. (Att. ii. 9, 1) on the words citius omnino quam potuit, where many instances of this usage are quoted to resist the conjecture oportuit, which in that accepted by Wesenberg. passage is (c) Att. iii. 8, 2, reiectus ab illo M; reiectus ab Ilio F. This is a case which in one respect closely resembles the following: (d) In Att. i. 17, 9 M presents Asiani, while A is said to have |