was afterwards carefully revised for me by M. Ch. Kohler with the MS. itself. He was also good enough to revise in the same way the section relating to the contents. I believe that by the aid of his keen eye and minute patience a high degree of accuracy has been secured; but I know by experience that absolute correctness, if not unattainable, is hardly ever attained. In fact, in some cases the distinction of the first and second hands is likely always to remain a problem. However, I believe that inaccuracies which may still exist are of very slight moment, and not such as will affect the use of the text for critical purposes. I shall (it is unnecessary to say) be much obliged to any scholar who will point out real blunders or imperfections of this or any other kind.

§ 6. Character of the text of St. Matthew. Before proceeding to discuss this point it will be desirable to state in what sense the terms 'Old-Latin' and 'Italian' and other cognate phrases are used. I have called this series 'Old-Latin Biblical texts,' meaning thereby that the books included in it are texts current before, or independent of, St. Jerome's revision of the New Testament, and re-translation of the Old at the end of the fourth century. Some scholars have adopted the term ' ante-Hieronymian,' practically in the same sense; but it appears awkward and cumbersome, and Old-Latin' is now in process of general adoption. Since the Council of Trent Hieronymian texts have usually been called 'Vulgate,' though the word was used in a different sense by St. Jerome 1, and even by Martianay 2. Old-Latin texts then, as distinguished from Vulgate, mean all early Latin versions of the Bible which are not Hieronymian, of whatever date the MSS. may be which contain them, or in whatever country they were current. They are divided by Dr. Hort3 (whose authority is very weighty) into three groups: first, 'African,' agreeing generally with the quotations of Tertullian and Cyprian ; secondly, European,' probably based on a distinct version or versions current in Western Europe, and especially in North Italy; and, thirdly,

1 See Westcott, Vulgate, p. 1689 foll. St. Jerome by Vulgata editio' means the LXX in its uncorrected form and the old Latin translation of it. The Council of Trent call Jerome's version 'vetus et vulgata editio,' and thus stamped the modern usage of the term.

2 Cp. above, p. xix.

3 Westcott and Hort's Greek Testament, Introduction, pp. 78-84, Macmillan, 1881.

'Italian' (or 'Italic ') agreeing to a great extent with the text which is found in many of St. Augustine's writings, and probably of the type which he refers to in his single laudatory notice (de doctrina Christiana, ii. 15). These last are evidently due,' says Dr. Hort (p. 79, § 110), 'to various revisions of the European text, made partly to bring it into accord with such Greek MSS. as chanced to be available, partly to give the Latinity a smoother and more customary aspect.'

To the first or African class he assigns the Palatine Gospels of Vienna (e), and the Bobbio Gospels of Turin (k), the latter of which I hope to re-edit in this series-ten palimpsest leaves of the Acts 1 (Paris, Lat. 6400 G=h), and two of the Apocalypse (same MS.).

In the European class he puts down ten Gospels, which are presumably the following: 1. Vercellensis (a), 2. Fragmenta Curiensia (a, ed. Ranke), 3. Veronensis (b), 4. Colbertinus (c), 5. Corbeiensis Parisiensis (ff or ff, Lat. 17225), 6. Claromontanus Vaticanus ( ed. Mai), 7. Vindobonensis (¿), 8. Saretianus or Sarzannensis (j, still unedited), 9. Sangallensis (), and 10. Dublinensis (r, ed. T. K. Abbott). We are also fortunate in possessing in the Gigas Holmiensis, recently edited by J. Belsheim, complete 'European' texts of the Acts and Apocalypse (g or y); with fragments of the Acts (g—at Milan-and s). The Epistle of St. James in the Corbey MS. at St. Petersburg (see p. xxii. note 2) is also conjecturally assigned to this class.

In the Italian group Dr. Hort classes two MSS. of the Gospels, Brixianus (ƒ) and Monacensis (9)-the latter to be edited, I hope, in this series. In the Catholic Epistles he conjecturally places with these the Frisingen fragments of the Epistles of 1 and 2 Peter and 1 John (9), and more decidedly the other Frisingen fragments of the Pauline Epistles, all of which were published by Ziegler (r, r), as well as the Gottvicensis (fragg. Rom. Gal. r) published by Rönsch.

It will be observed that our text, as well as several others usually classed as old Latin, such as Corbeiensis Petropolitanus (fi), Sangermanensis 2 (g=Paris Lat. 13169 )2, Rhedigerianus (1) are omitted from this enumeration. Dr. Hort speaks rather cautiously of this class, but concludes that they are certainly in most cases and not improbably in

1 Olim Reg. 5367. Cp. Sabatier on Acts iii. 2 foll.

* I am inclined to accept Dr. Hort's judgment for these two MSS., especially the latter.

all 'monuments of the process ... by which old-Latin readings, chiefly European, but in a few cases African, found their way into texts fundamentally Hieronymic' (ib. p. 82, § 114). In other words, he inclines to regard all these as mixed texts, i. e. texts with a Vulgate base, into which Old-Latin readings have been introduced. He is not here considering the bilingual or Graeco-Latin texts which belong to a different class, nor does he express any opinion as to the British' texts, except so far as the Dublin r is included in his European class. But he probably would class these all as mixed.

Is our St. Matthew then based on a Vulgate text? This is a nice question, and one that I will not undertake to determine ex cathedra. My readers are probably aware that there is considerable variety of opinion on such points, and that Dr. Ziegler, for instance, speaks of Brixianus as 'mit Bestandtheilen der Vulgata untermischt,' while Drs. Westcott and Hort (with others) call it, as I believe rightly, 'Italian 1.'

Without therefore presuming to dogmatize I would offer the following remarks: That there is a mixture of a certain kind, in parts at least of our MS., is clear from the emphatic words of the editor already noticed (fol. 69 above, pp. x, xiv). They can only mean that he took different classes of MSS. and made an eclectic text by excerpting now a piece from one and now a piece from another. That he supposed himself to be generally following St. Jerome is also clear from the same subscription and from that on fol. 187, while the fact that he gives the Eusebio-Ammonian sections and canons is in itself sufficient, I imagine, to prove the use of a Hieronymian MS. of some kind in the Gospels.

The process of mixture has also been here and there so careless that we have sometimes double renderings, sometimes inconsistencies and inequalities, on the very surface of the text. These 'conflations' are occasionally striking-though a common phenomenon in Old-Latin MSS. Thus the same Greek word is twice rendered in v. 11 (Beati estis cum uos odio habuerint homines, maledicent, et persequentur), and vi. 2

1 Die Lat. Bibelübersetzungen von Hieronymus, etc., p. 108, München, 1879. I have not examined it thoroughly enough to assert that it contains no Vulgate readings. Mr. J. Belsheim in 1878 published the Stockholm codex aureus (Christiania, P. T. Mallings) as ante-Hieronymian, but he would probably now acquiesce in the general judgment which classes it as 'mixed,' that is based on a Hieronymian text with many old readings.

(in synagogis et in plateis, et in uicis). In xxv. 6 there is a twofold rendering (ecce sponsus est, uenit), representing a variation in the Greek text, and in iii. 5 there is a grammatical inequality arising from the mixture of two Latin texts (tunc exiebant ad eum ex Hierusolymis et omnis Iudaea, a combination of readings of Vercellensis and Brixianus).

Neither of these cases however necessarily implies the use of a Hieronymian MS. and the first pro tanto excludes it, inasmuch as St. Jerome there reads 'Beati estis cum maledixerint uobis et persecuti uos fuerint,' using a third phrase for the one Greek word.

In order therefore to ascertain the character of the text more decidedly it is necessary to compare it with the different types of text, Vulgate, Italian, and European, to which it has most general similarity. This I have done throughout a number of chapters in different parts of the book, and more thoroughly and minutely in the Fifth Chapter. The results of this comparison may be stated numerically.

In chapter v. there are forty-eight verses. In eight verses (3, 6, 8, 10, 26, 27, 35, 36) the three types of text coincide with G in the three representative MSS. which I compared with our book, viz. Amiatinus for the Vulgate, Brixianus for the Italian, and Veronensis (and once or twice Colbertinus) for the European. In the remaining forty verses there are at least seventy-six places in which G varies from one or other or all of the three1. These may be thus divided :—

Peculiar readings, in which G stands alone

Readings in which it has a distinctly Vulgate type

Neutral readings in which it agrees with the Vulgate and Italian type

18

3

2

27

ΙΟ

It will be seen, then, that the possible Old-Latin element is extremely large, that is to say, about ninety-six per cent. (only three readings in seventy-five being distinctly Vulgate), while the distinctly Old-Latin

1 I believe that I omitted to count one or two 'neutral' readings.

e

element is about seventy-three per cent.; and the distinctly European element is far larger than either of the other two. The proportions probably vary in other chapters. In some, such as the first and the fourth, there are I believe no distinctly Vulgate readings, while in others there are more than here. But I think that if the whole book were examined the general result would be found the same, with probably a larger proportion of distinctly Italian readings than there happens to be in this chapter. I conclude from this examination that the basis of our book was not a Hieronymian text, but a mixture of the Italian and European texts, which was corrected occasionally by the Vulgate, but has a large peculiar element, perhaps drawn from several MSS.

A few specimens of the distinctly Vulgate readings are enough to prove that our book was corrected from St. Jerome's text. Such are

ii. 16. quoniam inlusus esset instead of quod delusus esset, brix.; quoniam delusus est, ueron. colb.

23. per prophetas instead of per prophetam, O. L.

v. 42. Omni qui petit a te da ei, a conflation of the Vulgate Qui petit a te da ei with the Old-Latin Omni petenti te da or da ei.

44. his qui oderunt uos instead of odiunt, O. L.

xv. 8. populus hic labiis me honorat against plebs haec labiis me honorat, uerc. colb.; plebs haec labiis me diligit, uer. ; ad propinquat mihi populus hic ore suo et labiis me honorat, brix.

xv. 17. secreto1 against in uia seorsum, uerc. colb.; om. corb. 1. 2 ueron.; seorsum in uia, brix. clar.

xxi. 7. et eum desuper sedere fecerunt with g, against et sedebat super eum, uerc. (ut uidetur) uer. corb. 1. pal. et sedit super eum, brix. mon. clar. ib. 37. filium suum against filium suum unicum, uerc. uer. corb. 1. 2 clar.; filium suum unigenitum, brix.

1 Bianchini quotes S. Hilary (in Matt. col. 708 e) as reading secreto, but it appears from Sabatier that it is the reading only of some (six) MSS. of that father. It is probably a Vulgate reading interpolated into his text.