At length, after baffling delays only too readily accounted for by the public calamities of the Papal state, the concluding volume of this sumptuous and important work was published late in 1881. Sergius had now retired through failing eyesight, and his place was taken by “Henricus Canonicus Fabiani,” Cozza (who is now Abbot of the Grotta Ferrata at Tusculum near Frascati, the chief seat of the monks of the Greek order of St. Basil) still holding the second place. From the laudatory tone in which the latter is spoken of (p. xiv), it would seem that the Preface was written by his new colleague, who acknowledges the help of U. Ubaldi and the Basilian monk Ant. Rocchi, all three “adjutoribus et administris miratis equidem se tantis viris adjutores et successores datos” (p. xv). This Preface consists of twenty-two pages, and contains almost nothing that is interesting to the critic, much that displays superficial and newly-acquired acquaintance with the whole subject. Fabiani assigns the end of the fourth century as the date of the manuscript, regarding it as only a few years older than the Sinaitic copy[158], whose discovery he [pg 119] hails without a vestige of ungenerous jealousy: “Quorum tale est demum par, ut potius liber Vaticanus gaudere debeat quod tam sui similem invenerit fratrem, quam expavescere quod aemulum” (p. viii). Since that time a splendid edition has been issued of the New Testament in 1889, and the Old in 1890, under the care of the Abbate Cozza-Luzi, in which the whole is beautifully exhibited in photograph: so that all students can now examine for themselves the readings and characteristics of this celebrated manuscript with all but the advantage which is given in an examination of the original vellum itself (Novum Testamentum e Codd. Vat. 1209, &c. Rom. 1889, 4to): and gratitude is due from all textual scholars to the authorities of the Vatican.
Those who agree the most unreservedly respecting the age of the Codex Vaticanus, vary widely in their estimate of its critical value. By some it has been held in such undue esteem that its readings, if probable in themselves, and supported (or even though not supported) by two or three other copies and versions, have been accepted in preference to the united testimony of all [pg 120] authorities besides: while others, admitting the interest due to age, have spoken of its text as one of the most vicious extant. Without anticipating what must be discussed hereafter we may say at once, that, while we accord to Cod. B at least as much weight as to any single document in existence, we ought never to forget that it is but one out of many, several of them being nearly (and one quite) as old, and in other respects not less worthy of confidence than itself. One marked feature, characteristic of this copy, is the great number of its omissions, which has induced Dr. Dobbin to speak of it as presenting “an abbreviated text of the New Testament:” and certainly the facts he states on this point are startling enough[159]. He calculates that Codex B leaves out words or whole clauses no less than 330 times in Matthew, 365 in Mark, 439 in Luke, 357 in John, 384 in the Acts, 681 in the surviving Epistles; or 2,556 times in all. That no small proportion of these are mere oversights of the scribe seems evident from the circumstance that this same scribe has repeatedly written words and clauses twice over, a class of mistakes which Mai and the collators have seldom thought fit to notice, inasmuch as the false addition has not been retraced by the second hand, but which by no means enhances our estimate of the care employed in copying this venerable record of primitive Christianity[160]. Hug and others have referred the origin of Codex B to Egypt, but (unlike in this respect to Codex A) its history does not confirm their conjecture, and the argument derived from orthography or grammatical forms, is now well understood to be but slight and ambiguous[161]. Dr. Hort, on no very substantial [pg 121] grounds, is “inclined to surmise that B and א were both written in the West, probably at Rome” (Introduction, pp. 265-7).
C. Codex Ephraemi, No. 9, in the Royal Library of Paris, is a most valuable palimpsest containing portions of the Septuagint version of the Old Testament on sixty-four leaves, and fragments of every part of the New on 145 leaves, amounting on the whole to less than two-thirds of the volume[162]. This manuscript seems to have been brought from the East by Andrew John Lascar [d. 1535], a learned Greek patronized by Lorenzo de' Medici; it once belonged to Cardinal Nicolas Ridolphi of that family, was brought into France by Queen Catherine de' Medici of evil memory, and so passed into the Royal Library at Paris[163]. The ancient writing is barely legible, having been almost removed about the twelfth century to receive some Greek works of St. Ephraem, the great Syrian Father [299-378]. A chemical preparation applied at the instance of Fleck in 1834, though it revived much that was before illegible, has defaced the vellum with stains of various colours, from green and blue to black and brown. The older writing was first noticed by Peter Allix [pg 122] nearly two centuries ago; various readings extracted from it were communicated by Boivin to Kuster, who published them (under the notation of Paris 9) in his edition of Mill's N. T., 1710. A complete collation of the New Testament was first made in 1716 by Wetstein, then very young, for Bentley's projected edition, for which labour (as he records the fact himself) he paid Wetstein £50. This collation Wetstein of course used for his own Greek Testament of 1751-2, and though several persons subsequently examined the manuscript, and so became aware that more might be gathered from it, it was not until 1843 that Tischendorf brought out at Leipsic his full and noble edition of the New Testament portion; the Old Testament he published in 1845. Although Tischendorf complains of the typographical errors made in his absence in the former of these two volumes, and has corrected them in the other, they probably comprise by far the most masterly production of this nature up to that date published; it is said too that none but those who have seen Codex C can appreciate the difficulty of deciphering some parts of it[164], in fact, whatever is not patent at first sight. The Prolegomena are especially valuable; the uncial type does not aim at being an imitation, but the facsimile faithfully represents the original, even to the present colour of the ink. In shape Codex C is about the size of Cod. A, but not quite so tall; its vellum is hardly so fine as that of Cod. A and a few others, yet sufficiently good. In this copy there is but one column in a page, which contains from forty to forty-six lines (usually forty-one), the characters being a little larger than those of either A or B, and somewhat more elaborate[165]. Thus the points at the ends of sigma, epsilon, and especially of the horizontal line of tau are more decided than in Codex A; delta, though not so fully formed as in later books, is less simple than in A, the strokes being of less equal thickness, and the base more [pg 123] ornamented. On the other hand, alpha and pi are nearer the model of Codex B. Iota and upsilon, which in Cod. A and many other copies have two dots over them when they commence a syllable, and are sometimes found with one dot, have here a small straight line in their place (see p. [36]). There are no breathings or accents by the first hand: the apostrophus is found but rarely, chiefly with Proper names, as δαδ᾽ The uncial writing is continuous; the punctuation of Cod. C, like that of A and B, consisting only of a single point, mostly but not always put level with the top of the preceding letter; wherever such a point was employed, a space of one letter broad was usually left vacant: these points are most common in the later books of the N. T. The κεφάλαια are not placed in the upper margin of the page as in Cod. A, but a list of their τίτλοι preceded each Gospel: the so-called Ammonian sections stand in the margin, but not at present the Eusebian canons; though, since lines of the text written in vermilion have been thoroughly washed out, the canons (for which that colour was commonly employed) may easily have shared the same fate (see p. [61]). There is no trace of chapters in the Acts, Epistles, or Apocalypse, and both the titles and subscriptions to the various books are very simple. Capital letters are used quite as freely as in Cod. A, both at the commencement of the (Ammonian) sections, and in many other places. All these circumstances taken together indicate for Cod. C as early a date as the fifth century, though there is no sufficient cause for deeming it at all older than Cod. A. Alexandria has been assigned as its native country, for the very insufficient reasons stated when we were describing A and B. It is carefully transcribed, and of its great critical value there is no doubt; its text seems to stand nearly midway between A and B, somewhat inclining to the latter. Two correctors have been very busily at work on Cod. C, greatly to the perplexity of the critical collator: they are respectively indicated by Tischendorf as C**, C***. The earliest, or the second hand, may have been of the sixth century, and his corrections are for some cause regarded by Dr. Hort as almost equally valuable for critical purposes with the manuscript itself: the second corrector, or the third hand, is perhaps of the ninth century, and he revised such portions as were adapted to ecclesiastical use, inserting many accents, the rough breathing, and some vocal [pg 124] notes. By him or more probably by a fourth hand (who did not change the text, but added some liturgical directions in the margin) small crosses were interpolated as stops, agreeably to the fashion of their times.
D of the Gospels and Acts, Codex Bezae Graeco-Latinus, belongs to the University Library at Cambridge, where the open volume is conspicuously exhibited to visitors in the New Building (Nn. ii. 41). It was presented to the University in 1581 by Theodore Beza, for whom and his master Calvin the heads of that learned body then cherished a veneration which already boded ill for the peace of the English Church[166]. Between the Gospels (whose order was spoken of above, pp. [72-4]) and the Acts, the Catholic Epistles once stood, of which only a few verses remain in the Latin translation (3 John ver. 11-15), followed by the words “epistulae Johannis III explicit, incipit actus apostolorum,” as if St. Jude's Epistle were displaced or wanting. There are not a few hiatus both in the Greek and Latin texts[167]. The contents of this remarkable document were partially made known by numerous extracts from it, under the designation of β´ in the margin of Robert Stephen's Greek Testament of 1550, whose account of it is that it was collated for him in Italy by his friends (τὸ δὲ β´ ἐστὶ τὸ ἐν Ἰταλίᾳ ὑπὸ τῶν ἡμετέρων ἀντιβληθὲν φίλων. Epistle to the Reader)[168]. It is not very easy to reconcile this statement with Beza's account prefixed [pg 125] to the manuscript and still extant in his own cramped handwriting, wherein he alleges that he obtained the volume in 1562 from the monastery of St. Irenaeus at Lyons (“oriente ibi civili bello”), where it had long lain buried (“postquam ibi in pulvere diu jacuisset”). This great city, it must be remembered, was sacked in that very year by the infamous Des Adrets, whom it suited to espouse for a while the cause of the Huguenots; and we can hardly doubt that some one who had shared in the plunder of the abbey[169] conveyed this portion of it to Beza, whose influence at that juncture was paramount among the French Reformed[170].
Beza in his editions of the Greek Testament published in 1582, 1589, and 1598, made some occasional references to the readings of his manuscript. Archbishop Whitgift borrowed it from Cambridge in 1583, and caused a poor transcript to be made of its Greek text, which he bequeathed to Trinity College (whereof he had been Master), in whose Library it still remains (B. x. 3).
Patrick Young, of whom we have heard in connexion with Cod. A (p. [103] and note 1), sent extracts from Cod. D to the brothers Dupuy at Paris, through whom they reached Morinus and Steph. Curcellaeus. An unusually full collation was made for Walton's Polyglott (Tom. vi, Num. xvi, 1657) by pious Archbishop Ussher, who devoted to these studies the doleful leisure of his latter years. Mill collated and Wetstein transcribed (1716) this document for their great editions of the Greek Testament, but they both did their work carelessly; and though Bentley was allowed to keep it at home for seven years, his notices of its readings, as represented by Mr. Ellis (Bentleii Critica Sacra, pp. 2-26), or preserved in Stephen's N. T. of 1549 (Trin. Coll. B. xvii. 4), were put to no practical use. The best collation by far was made about 1732 by John Dickinson of St. John's College for John Jackson of Leicester, with whose other books it came into Jesus College Library (O. θ. 2), where it has lain neglected. But a manuscript replete as this is with variations from the sacred text beyond all other example could be adequately represented only by being published in full; a design entrusted by the University of Cambridge to Dr. Thomas Kipling, Senior Wrangler in 1768 and afterwards Dean of Peterborough [d. 1822], whose “Codex Theodori Bezae Cantabrigiensis” 1793, 2 vols. fol. (in type imitating the original handwriting much more closely than in Cod. A and the rest), is a not unfaithful transcript of the text[171], [pg 127] though the Prolegomena too plainly testify to the editor's pitiable ignorance of sacred criticism, while his habit of placing the readings of the several later hands (very loosely distinguished from each other) in the text, and those of the first hand in the notes (a defect we have also noted in the Roman editions of Cod. B), renders his volumes very inconvenient for use. Let Kipling be praised for the care and exact diligence his work evinces, but Herbert Marsh [1757-1839] was of all Cambridge men of that period the only one known to be competent for such a task. In 1864 the present writer was aided by the Syndics of the Cambridge Press in publishing an edition of Codex Bezae in common type, illustrated by a copious Introduction and critical notes, to which work the reader is referred for fuller information respecting this manuscript.
The Codex Bezae is a quarto volume 10 inches high by 8 broad, with one column on a page, the Greek text and its Latin version being parallel, the Greek on the left, or verso of each leaf, and the Latin on the right, opposite to it, on the recto of the next. Notwithstanding the Alexandrian forms that abound in it as much as in any other copy, and which have been held by some to prove the Egyptian origin of Codd. ABC, the fact of its having a Latin version sufficiently attests its Western origin. The vellum is not quite equal in fineness to that of a few others. There are thirty-three lines in every page, and these of unequal length, as this manuscript is arranged in στίχοι, being the earliest in date that is so (see p. [53]). The Latin is placed in the same line and as nearly as possible in the same order as the corresponding Greek. It has not the larger κεφάλαια or Eusebian canons, but only the so-called Ammonian sections, often incorrectly placed, and obviously in a later hand of about the ninth century. The original absence of these divisions is no proof that the book was not at first intended for ecclesiastical use (as some have stated), inasmuch as the sections and canons were constructed for a very different purpose (see above, pp. [59-63]), but is another argument for its being copied in the West, perhaps not far from the place where it rested so long. Other proofs of its Occidental, perhaps of its Gallican origin, especially that derived from the style of the Latin version, are [pg 128] collected in Scrivener's edition (Introd. pp. xxxi, xl-xlv). The characters are of the same size as in C, larger on the whole than in AB, but betray a later age than any of these, although the Latin as well as the Greek is written continuously, excepting that in the titles and subscriptions of the several books (as in Codd. DH of St. Paul) the words are separated. This copy has paragraph divisions of unequal length peculiar to itself[172]. They are indicated by placing the initial letter out in the margin, that letter being usually of the same size with the rest, though sometimes a little larger. Cod. D appears to be the earliest which exhibits larger letters after a pause in the middle of a line; but these are not very frequent. Instances of each case may be noticed in our facsimile (No. [42]), wherein the shapes of kappa, rho and phi, as indicated before (pp. [32], note 1, [37], [39]), are very observable. The Greek and Latin writing on the opposite pages are much like each other in appearance, the Latin letters being round and flowing, not square as in codices a little earlier in date, such as the Medicean and Vatican fragments of Virgil. This manuscript has been corrected, first by the original penman with a light stroke made by a pen nearly empty; after him by not less than eight or nine different revisers, some nearly coeval with the Codex itself, others not many centuries old. The changes they have made, especially when they employed a knife to scrape away the primitive reading, render too many places almost illegible. The first scribe often used a sponge to wash out his error before the ink was well dried in (see p. [27]). In addition to the single point about three-fourths of the height of a letter up, which often subdivides the στίχοι in both languages (facsimile, No. [42], l. 9) the coarse late hand which inserted the Ammonian sections placed double dots (:) after the numerals, and often inserted similar points in the text, before or over the first letter of a section. Each member of the genealogy in Luke iii forms a separate στίχος, as in Cod. B: quotations are indicated by throwing the commencement of the lines which contain them, both Greek and Latin, about an inch back or less [pg 129] (e.g. Matt. xxvi. 31; Mark i. 2, 3; Acts ii. 34, 35; iv. 25, 26). The first three lines of each book, in both languages, were written in bright red ink, which was also employed in the alternate lines of the subscriptions, and in other slight ornaments. The traces of the scribe's needle and lines (see p. [27]) are very visible, the margin ample, and the volume on the whole in good keeping, though its first extant page (Latin) is much decayed, and it is stained in parts by some chemical mixture that has been applied to it. The portions supplied by a later hand are of course in the uncial Greek and cursive Latin characters usual at the dates assigned to them. The liturgical notes in the margin of the Saturday and Sunday lessons (ανναγνοσμα is the form often used) are in thick letters, of a yet later date than the Ammonian sections. A few others for the great Feasts and Fast days occur; and, in a hand of about the twelfth century, lessons for the Festivals of St. George and St. Dionysius, the patron saints of England and France, as may be seen in the table of Menology.
The vellum employed for Codex Bezae is arranged in quires of four sheets (or eight leaves) each even throughout[173], the numeral signatures of which are set primâ manu so low down in the margin at the foot of the last page of each, that they are mostly cut off, in whole or partly, by the binder. Assuming that it ended with the Acts of the Apostles, it originally consisted of upwards of sixty-four (probably of sixty-seven) quires, of which the first, forty-fourth, and sixty-fourth, have each lost some leaves, the thirty-fourth is entire though containing but six leaves, while those signed Γ (3), ΙΔ (14), ΚΒ (22), ΜΕ (45), down to ΝΒ (52), ΝΖ (57), and all after ΞΔ (64), are wholly wanting. The result is that out of the 534 leaves it originally contained, only 406 now survive, about twelve of them being more or less mutilated. It is not easy to surmise what may have been written on the sixty-seven leaves that intervened between ΜΔ 5 and ΝΓ 1; the gap ends with 3 John ver. 11 [pg 130] (Greek), but the space is apparently too great for the Catholic Epistles alone, even though we suppose that Jude was inserted (as appears in some catalogues) otherwise than in the last place. The leaves added by later hands are nine in number. The Greek portion of the supplement to St. John (xviii. 14-xx. 13) much resembles in text the style of the original manuscript, and is often supported by Codd. אAB(C). The Latin of this portion is taken from the Vulgate version.
The internal character of the Codex Bezae is a most difficult and indeed an almost inexhaustible theme. No known manuscript contains so many bold and extensive interpolations (six hundred, it is said, in the Acts alone), countenanced, where they are not absolutely unsupported, chiefly by the Old Latin and the Curetonian version: its own parallel Latin translation is too servilely accommodated to the Greek text to be regarded as an independent authority, save where the corresponding Greek is lost.
This passage was penned by Dr. Scrivener before the publication of the highly ingenious treatise by Mr. Rendel Harris, entitled “A Study of the Codex Bezae” (1891), being the beginning of the second volume of the Cambridge “Texts and Studies.” Mr. Harris from curious internal evidence, such as the existence in the text of a vitiated rendering of a verse of Homer which bears signs of having been retranslated from a Latin translation, infers that the Greek has been made up from the Latin, and traces the latter to the second century. He shows its affinity with the text of Irenaeus, and discovers traces in it of Montanism. He opens up many points of interest for any one who would examine this “singular Codex”: but injustice must not be done to the fertile author by supposing that in what is evidently 'a Study' he concludes that he has settled all the numerous questions which he broaches. No one however can really investigate the Codex Bezae without studying this work, which will be found both instructive in the highest degree and amusing.