the middle of a line in any extant document earlier than Codex Beza (D) of the beginning of the sixth century, and even then so rarely and timidly that some persons have denied their existence altogether. But, in truth, the only writings which resemble Codd. B and in this respect are older than Codex D can possibly be, and must, with them, be referred to the fourth century at latest. Such are the Egyptian papyri in the British Museum and elsewhere; the Herculanean rolls named above; the few leaves of the Colbert Pentateuch; four palimpsest leaves in the British Museum (Add. 17,136), containing sixteen verses of S. John; the Vatican fragment of Dio Cassius (but here again the case is not quite clear); to which Tischendorf (Prol. p. xix.) now adds 'item fragmenta Euripidis Claromontana et fragmenta Menandri Porfiriana.'

Another considerable note of very great age consists in the scarceness and extreme simplicity of the marks of punctuation. A simple point will be seen to occur about twenty-seven times on our specimen page, and even this is rejected by J. L. Hug (whose treatise De Antiquitate Vaticani Codicis Commentatio, first established the right of our manuscript to the paramount consideration now granted to it), for he bluntly declares that it 'manifeste nullam habuit interpunctionem,' the stops we now perceive having been made by a later hand. Now although Hug had facilities for examining the document when it lay in Paris in 1809, the spoil of war, such as have fallen to no more recent critic whose judgment can be relied on, we believe he was mistaken in this instance. The stop we do find is not such as a later pen would have introduced, but one simple dot, (in some rare cases, or at the end of books, two, e.g. :) like the Greek colon, abreast of the top of the preceding letter. Such a point even Hug admits to be original in 2 Cor. iii. 15, Keira (a passage of which Tischendorf gives that interesting fac-simile from a tracing made by him in 1843), wherein the clause having been written in error twice over, the modern tracer retouches it but once. Tischendorf (Proleg. p. xxi.) records for the benefit of future collectors more than twenty places in the two earlier Gospels in which such original points may even now be detected, and any one at all acquainted with the oldest manuscripts now in England (e. g. Codd. A, D) will be aware that these marks of punctuation, the dialysis (..) over I and Y, and the apostrophus ('), are almost illegible in many pages where the letters are yet clear and almost fresh. Now in all these respects the Sinai manuscript is an exact counterpart of that in the Vatican: the dialysis is sometimes written, sometimes not-the apostrophus is common after or final-the single point is almost the only sign of punctuation; it occurs far more rarely in some parts of the

copy than at others (e. g. less frequently in the Gospels than in the Epistles), and its place is often supplied by a simple blank interval, left often, but by no means invariably, to distinguish one paragraph from another.

The same inference, leading to a conviction of their remote antiquity, may be drawn from marking the abridgments employed in Codd. B and . The most conspicuous on our specimen page is the small horizontal line set over and a little to the right of the last letter in a line to indicate v; this occurs about eighteen times on the page, as well when v is in the middle, (e. g. col. 2, 1. 18, Evтoλai) as at the end of a word. A few names also in constant use are frequently (not always) abridged by writing only their first and last letters, and placing over them a horizontal line; such are θεοσ, ιησους, χριστος, and their oblique cases on this page, and κuptor elsewhere. Tischendorf notes, that πνα, ισλ, ανοσ, δαδ, for πνευμα, ισραηλ, ανθρωποσ, δαυιδ are also found in Codex B, but more rarely than in Codex N, while πατηρ, μητηρ, υιοσ, σωτηρ, ιερουσαλημ, ουρανοσ, are seldom if ever shortened in the Vatican New Testament: he cannot answer for the Old Testament. All these words, however, are abridged thus in Codex (e. g. unTepa thirteen times out of the twenty-three places in which it is found), so that as far as this test will carry us, Codex B being considerably more free from the compendia scribendi of more recent copies, may be regarded as somewhat the older of the pair; an inference which will be confirmed in the sequel by other arguments.

1

From what we stated respecting the missing portions of Codex Vaticanus (p. 411), the reader may have inferred that the sacred books stand in a different order from that of our modern Bibles. In all our most ancient Greek manuscripts except Codex Beza (which runs SS. Matthew, John, Luke, Mark), the order of the Gospels is the same as ours; but in only a very few do the Pauline Epistles precede the Acts. Such however is the arrangement of Codex Sinaiticus, which in this particular is at variance with Codex Vaticanus, wherein the Acts and seven Catholic Epistles immediately follow S. John's Gospel. In the order of the Pauline Epistles Codd. Vaticanus (with one exception, to be mentioned hereafter), Sinaiticus, Alexandrinus, and Ephraemi, (B, N, A, C) agree in setting the Epistle to the Hebrews immediately after 2 Thess. and before the four private or Pastoral Epistles. The Apocalypse invariably stands last,

1 The only other known Codices which resemble that of Sinai in this respect is the celebrated Leicester MS. (Evan. 69); the Dublin Cod. Montfort (Evan. 61); Faber's of Deventer (van. 90, whose Gospels stand SS. John, Luke, Matthew, Mark); and the Bodleian Canonici 34: all very modern MSS.

excepting that in Oxon. Wake 34 it precedes the Pauline Epistles, and in Vatican. 2,080 (Evan. 175) it seems to intervene between the Acts and the Catholic Epistles.

Judging from one other peculiarity of Codex Vaticanus, we might be led to infer that it is prior in date to Codex Sinaiticus. The latter, in common with the Codex Alexandrinus and most other manuscripts, exhibits in its margin, written in their appropriate red ink or paint, the canons of harmony throughout the Gospels, which Eusebius, the Church historian, about A.D. 330, devised, or perhaps only accommodated to the previously existing sections of Ammonius. Now these canons are completely wanting in Codex Vaticanus, which divides the Gospels into 464 chapters, an arrangement long believed to be unique, but which Dr. Tregelles in 1858 detected in a palimpsest fragment of S. Luke, of about the eighth century, the property of the Bible Society, and since published by him under the name of Codex Zacynthius (E) from its having been brought from Zante. In the Acts and Epistles, moreover, Codex Vaticanus contains two sets of chapters, one more ancient, apparently original, in which the Pauline Epistles are treated as one continuous work;1 the other by a later pen; and from these independent notations we gather two remarkable facts: (1) that in the older exemplar from which Codex B was taken the Epistle to the Hebrews, which now follows 2 Thess., and is on the same leaf with it, must have been placed after that to the Galatians,2 as it is even now in Zoega's Thebaic version; and (2) that there must be some connexion between Codex B and of a closer nature than exists between other documents, inasmuch as the smaller and more recent chapters of Codex B are placed (with some slight changes) in the margin of Codex &, throughout the first half of the Acts, by a hand not much posterior to that of the original scribe: the incomplete notation in Codex looking older than the complete one in Codex B. We shall hereafter see that the same conclusion with regard to the intimate relation between the two Codices is irresistibly forced upon us by coincidences in their readings too frequent and too remarkable to be imputed to accident.

Yet we must profess ourselves to have been rather startled than persuaded by the arguments alleged by Tischendorf to show

1 Tischendorf finds the same arrangement in an Arabic MS. of S. Paul's Epistles, dated A.D. 892, which he took to S. Petersburg.

2 The proof is simple enough, but quite conclusive. The Epistle to the Galatians in Codex B ends with section 58, that to the Ephesians that follows it begins with section 70, and the notation goes on regularly to section 93, in which 2 Thess. ends: then succeeds the Epistle to the Hebrews, beginning with section 59, and ending abruptly at ch. ix. 12, in the middle of section 64.

that Codices Vaticanus and Sinaiticus are to some extent the work of the same pen (Proleg. pp. xxi.—xxiii.) There is certainly no antecedent impossibility in this supposition; nay more, in later times, about the end of the thirteenth century, the same professional writer, Theodore of Hagios Petros, in the Morea (Delitzsch, Abhandlungen, pp. 217-9), affixes his name and date to no less than seven known copies of the New Testament or parts of it,1 each differing from the other not inconsiderably in regard to the text it follows. Yet we must say that the reasons produced so far by Tischendorf seem quite insufficient to prove his point. When he published Codex Sinaiticus in 1862, he declared his opinion that it was written by four several scribes, to one of whom, indicated by D., he refers the books of Tobit, Judith, three and a half leaves of 4 Maccabees, six leaves of the New Testament, and possibly the first thirty-two lines of the Apocalypse. This person he now believes to have transcribed the New Testament in Codex Vaticanus; respecting its Old Testament portion he cannot speak positively, though here, too, he finds traces of two other hands. Now, it must be known to all our readers that few things are harder to decide upon than identity or difference of handwriting-not to mention that the formal style maintained in these old Codices admits of much less individuality of character than the free running hand of common life. Hence we are by no means surprised that other scholars, scarcely less competent than Tischendorf, whom he permitted to examine the manuscript while it was in his possession, have doubted or even rejected his decision in this respect. Since Codex is now unfortunately at S. Petersburg, a city too remote to be visited by the few who have a right to speak authoritatively on such matters, scholars have been content to let the question rest for the present, the rather as up to this time it has seemed to have no practical bearing whatever on textual criticism. Tischendorf himself could not have seen Codex N for the space of four years or more when he was once more set face to face with its rival in February, 1866; and in nothing surely would the best memory be more likely to deceive us, than in pronouncing, after such an interval, upon the sameness of handwriting between two documents which now, alas! never can be brought together.

We must not, however, imagine that Tischendorf's notion that the penman of the Vatican New Testament also wrote a portion of the Sinai Bible rests only or mainly upon similarity

They are Cod. Havn. 1 (Evan. 234) A.D. 1278; Burney, 21 Brit. Mus. (Evan. r), A.D. 1292; Cod. Fabri (mentioned above, p. 12 note; Evan. 90), A.D. 1293; Biblio. Sussex. (Evan. q), A.D. 1295; Synax. Mosc. Syn. Typ. 26, A.D. 1295; Ven. Nan. (Evan. 412), A.D. 1301; Wake Oxon. (Evan. 74), undated.

of style. Were that the case, under the circumstances we have described, we should attach but little weight to his conjecture. There are people in England who take much interest in studies of this nature, and sincerely admire the German critic's energy, perseverance, sagacity, and skill, but are by no means disposed to regard him as impeccable. By these it was perceived, soon after the publication of Codex Sinaiticus, that the portions of its New Testament referred by him to a separate scribe, do indeed differ from the rest in certain small particulars of internal character, an argument all the more cogent as it had passed unnoticed, or at any rate unmentioned, by Tischendorf. Not only did the six leaves range themselves as they should do into three separate sheets,' but the text itself is far more free from itacisms, or cases of false spelling from the confusion of vowel sounds in popular discourse, than on the leaves which either precede or follow them. Nay more, such itacisms as do occur on the six leaves in question resemble those found in Codex Vaticanus, much more than we see elsewhere in Codex Sinaiticus. The latter, for example, often has in the place of the diphthong et, seldom et for, which variation is common enough in the Vatican manuscript; but the six leaves of the Sinai copy attributed by Tischendorf to the scribe of the other (as any one of the several Collations of Codex & that have been published will show for itself) contain et fort no less than forty-six times (e.g. γαλειλαια, ηλειας, μεικρων, πειλατοσ, φαρεισαιοσ), the converse, i for et (e.g. πελάγι, αγγιοισ, στάσι) thirty-two times. So too ιωανησ for ιωαννησ, which is not found elsewhere in Codex, but perpetually in Codex B (yet not in Luke i. 13, 60, 63, by the first hand), appears with the single v, Matt. xvi. 14, xvii. 1, 13; Luke i. 13, the only places in which it occurs in the course of the six leaves. It is also met with twice in the few lines which open the Apocalypse (ch. i. 1-5), which for other reasons have been attributed to the same scribe.

To these arguments, which of course derive their chief weight from their very minuteness and absence of design they indicate, Tischendorf adds others, more slight in themselves, though valuable as auxiliaries, so far as they are true in fact. The mark >, for instance, we learn from his evidence to be very frequent in all

1 Fol. 10 (Matt. xvi. 9-xviii. 12), and fol. 15 (Matt. xxiv. 36-xxvi. 6), form the 2d sheet of quire 74; foll. 28, 29 (Mark xiv. 54-Luke i. 56), the 4th sheet of quire 76; fol. 88 (1 Thess. ii. 14-v. 28), and fol. 91 (Heb. iv. 16-viii. 1), the 3d sheet of quire 84.

2 This last fact we should hardly have gathered from Tischendorf's language: 'Ut enim in universum Sinaiticus innumeris locis pro e præbet, multo rarius et pro, Vaticanus vero e pro 4, innumeris locis habet, multo rarius pro e, ita scriptura sex foliorum illorum ab omnibus ejus codicis reliquis discedens, ad maximam Vaticani similitudinem accedit.' (Proleg. p. xxii.)