As the extant Greek manuscripts of the New Testament supply both the most copious and the purest sources of Textual Criticism, we propose to present to the reader some account of their peculiarities in regard to material, form, style of writing, date and contents, before we enter into details respecting individual copies, under the several subdivisions to which it is usual to refer them.

1. The subject of the present section has been systematically discussed in the “Palaeographia Graeca” (Paris, 1708, folio) of Bernard de Montfaucon [1655-174116], the most illustrious member of the learned Society of the Benedictines of St. Maur. This truly great work, although its materials are rather too exclusively drawn from manuscripts deposited in French libraries, and its many illustrative facsimiles are somewhat rudely engraved, still maintains a high authority on all points relating to Greek manuscripts, even after more recent discoveries, especially among the papyri of Egypt and Herculaneum, have necessarily modified not a few of its statements. The four splendid volumes of M. J. B. Silvestre's “PalÉographie Universelle” (Paris, 1839-41, &c. folio) afford us no less than 300 plates of the Greek writing of various ages, sumptuously executed; though the accompanying letter-press descriptions, by F. and A. Champollion Fils, seem in this branch of the subject a little disappointing; nor are the valuable notes appended to his translation of their work by Sir Frederick Madden (London, 2 vols. 1850, 8vo) sufficiently numerous or elaborate to supply the Champollions' defects. Much, however, may also be learnt from the “Herculanensium [pg 022] voluminum quae supersunt” (Naples, 10 tom. 1793-1850, fol.); from Mr. Babington's three volumes of papyrus fragments of Hyperides, respectively published in 1850, 1853 and 1858; and especially from the Prolegomena to Tischendorf's editions of the Codices Ephraemi (1843), Friderico-Augustanus (1846), Claromontanus (1852), Sinaiticus (1862), Vaticanus (1867), and those other like publications (e.g. Monumenta sacra inedita 1846-1870, and Anecdota sacra et profana 1855) which have rendered his name perhaps the very highest among scholars in this department of sacred literature. What I have been able to add from my own observation, has been gathered from the study of Biblical manuscripts now in England. To these sources of information may now be added Professor Wattenbach's “Anleitung zur griechischen Palaeographie” second edition, Leipsic, 1877, Gardthausen's “Griechische Palaeographie,” Leipsic, 1879; Dr. C. R. Gregory's “Prolegomena” to the eighth edition of Tischendorf, and especially the publication of “The Palaeographical Society Greek Testament,” Parts I and II, Leipsic, 1884, 1891, “Facsimiles of Manuscripts and Inscriptions” edited by E. A. Bond and E. M. Thompson, Parts I-XII, London, 1873-82, and a Manual on “Greek and Latin Palaeography” from the hands of Mr. E. Maunde Thompson, of which the proof-sheets have been most kindly placed by the accomplished author at the disposal of the editor of this work, and have furnished to this chapter many elements of enrichment. It may be added, that since manuscripts have been photographed, all other facsimiles have been put in the shade: and in this edition references as a rule will be given only to photographed copies.

2. The materials on which writing has been impressed at different periods and stages of civilization are the following:—Leaves, bark, especially of the lime (liber), linen, clay and pottery, wall-spaces, metals, lead, bronze, wood, waxen and other tablets, papyrus, skins, parchment and vellum, and from an early date amongst the Chinese, and in the West after the capture of Samarcand by the Arabs in a.d. 704, paper manufactured from fibrous substances17. The most ancient manuscripts of the New Testament now existing are composed of vellum or parchment (membrana), the term vellum being [pg 023] strictly applied to the delicate skins of very young calves, and parchment to the integuments of sheep and goats, though the terms are as a rule employed convertibly. The word parchment seems to be a corruption of charta pergamena, a name first given to skins prepared by some improved process for Eumenes, king of Pergamum, about b.c. 150. In judging of the date of a manuscript on skins, attention must be paid to the quality of the material, the oldest being almost invariably written on the thinnest and whitest vellum that could be procured; while manuscripts of later ages, being usually composed of parchment, are thick, discoloured, and coarsely grained. Thus the Codex Sinaiticus of the fourth century is made of the finest skins of antelopes, the leaves being so large, that a single animal would furnish only two (Tischendorf, Cod. Frid.-August. Prolegomena, §. 1). Its contemporary, the far-famed Codex Vaticanus, challenges universal admiration for the beauty of its vellum: every visitor at the British Museum can observe the excellence of that of the Codex Alexandrinus of the fifth century: that of the Codex Claromontanus of the sixth century is even more remarkable: the material of those purple-dyed fragments of the Gospels which Tischendorf denominates N, also of the sixth century, is so subtle and delicate, that some persons have mistaken the leaves preserved in England (Brit. Mus. Cotton, Titus C xv) for Egyptian papyrus. Paper made of cotton18 (charta bombycina, called also charta Damascena from its place of manufacture) may have been fabricated in the ninth19 or tenth century, and linen paper (charta proper) as early as 1242 a.d.; but they were seldom used for Biblical manuscripts sooner than the thirteenth, and had not entirely displaced parchment at the era of the invention of printing, about a.d. 1450. Lost portions of parchment or vellum manuscripts are often supplied in paper by some later hand; [pg 024] but the Codex Leicestrensis of the fourteenth century is composed of a mixture of inferior vellum and worse paper, regularly arranged in the proportion of two parchment to three paper leaves, recurring alternately throughout the whole volume. Like it, in the mixture of parchment and paper, are codd. 233 and Brit. Mus. Harl. 3,161—the latter however not being a New Testament MS.

3. Although parchment was in occasional, if not familiar, use at the period when the New Testament was written (τὰ βιβλία, μάλιστα τὰς μεμβράνας 2 Tim. iv. 13), yet the more perishable papyrus of Egypt was chiefly employed for ordinary purposes. This vegetable production had been used for literary purposes from the earliest times. “Papyrus rolls are represented on the sculptured walls of Egyptian temples.” The oldest roll now extant is the papyrus Prisse at Paris, which dates from 2500 b.c., or even earlier, unless those which have been lately discovered by Mr. Flinders Petrie reach as far, or even farther, back20. The ordinary name applied in Greek to this material was χάρτης (2 John 12), though Herodotus terms it βύβλος (ii. 100, v. 58), and in Latin 4. Parchment is said to have been introduced at Rome not long after its employment by Attalus. Nevertheless, if it had been in constant and ordinary use under the first Emperors, we can hardly suppose that specimens of secular writing would have failed to come down to us. Its increased growth and prevalence about synchronize with the rise of Constantinopolitan influence. It may readily be imagined that vellum (especially that fine sort by praiseworthy custom required for copies of Holy Scripture) could never have been otherwise than scarce and dear. Hence arose, at a very early period of the Christian era, the practice and almost the necessity of erasing ancient writing from skins, in order to make room for works in which the living generation felt more interest, especially when clean vellum failed the scribe towards the end of his task. This process of destruction, however, was seldom so fully carried out, but that the strokes of the elder hand might still be traced, more or less completely, under the more modern writing. Such manuscripts are called codices rescripti or palimpsests (παλίμψηστα22), and several of the most precious monuments of sacred learning are of this description. The Codex Ephraemi at Paris contains large fragments both of the Old and New Testament under the later Greek works of St. Ephraem the Syrian: and the Codex Nitriensis, more recently disinterred from a monastery in the Egyptian desert and brought to the British Museum, comprises a portion of St. Luke's Gospel, nearly obliterated, and covered over by a Syriac treatise of Severus of Antioch against Grammaticus, comparatively of no value whatever. It will be easily believed that the collating or transcribing of palimpsests has cost much toil and patience to those whose loving zeal has led them to the attempt: and after all the true readings will be sometimes (not often) rather uncertain, [pg 026] even though chemical mixtures (of which “the most harmless is probably hydrosulphuret of ammonia”) have recently been applied with much success to restore the faded lines and letters of these venerable records.

5. We need say but little of a practice which St. Jerome23 and others speak of as prevalent towards the end of the fourth century, that of dyeing the vellum purple, and of stamping rather than writing the letters in silver and gold. The Cotton fragment of the Gospels, mentioned above (p. 23), is one of the few remaining copies of this kind, as are the newly discovered Codex Rossanensis and the Codex Beratinus, and it is not unlikely that the great Dublin palimpsest of St. Matthew owes its present wretched discoloration to some such dye. But, as Davidson sensibly observes, “the value of a manuscript does not depend on such things” (Biblical Criticism, vol. ii. p. 264). We care for them only as they serve to indicate the reverence paid to the Scriptures by men of old. The style, however, of the pictures, illustrations, arabesques and initial ornaments that prevail in later copies from the eighth century downwards, whose colours and gilding are sometimes as fresh and bright as if laid on but yesterday24, will not only interest the student by tending to throw light on mediaeval art and habits and modes of thought, but will often fix the date of the books which contain them with a precision otherwise quite beyond our reach.

6. The ink found upon ancient manuscripts is of various colours25. Black ink, the ordinary writing fluid of centuries (μέλαν, atramentum, &c.) differs in tint at various periods and in different countries. In early MSS. it is either pure black or slightly brown; in the Middle Ages it varies a good deal according to age and locality. In Italy and Southern Europe it is generally blacker than in the North, in France and Flanders [pg 027] it is generally darker than in England; a Spanish MS. of the fourteenth or fifteenth century may usually be recognized by the peculiar blackness of the ink. Deterioration is observable in the course of time. The ink of the fifteenth century particularly is often of a faded grey colour. Inks of green, yellow, and other colours, are also found, but generally only for ornamental purposes. Red, either in the form of a pigment or fluid ink, is of very ancient and common use, being seen even in early Egyptian papyri. Gold was also used as a writing fluid at a very early period. Purple-stained vellum MSS. were usually written upon in gold or silver letters, and ordinary white vellum MSS. were also written in gold, particularly in the ninth and tenth centuries, in the reigns of the Carlovingian kings. Gold writing as a practice died out in the thirteenth century: and writing in silver appears to have ceased contemporaneously with the disuse of stained vellum. The ancients used the liquid of cuttle-fish. Pliny mentions soot and gum as the ingredients of writing-ink. Other later authors add gall-apples: metallic infusions at an early period, and vitriol in the Middle Ages were also employed.

7. While papyrus remained in common use, the chief instrument employed was a reed (κάλαμος 3 John ver. 13, canna), such as are common in the East at present: a few existing manuscripts (e.g. the Codd. Leicestrensis and Lambeth 1350) appear to have been thus written. Yet the firmness and regularity of the strokes, which often remain impressed on the vellum or paper after the ink has utterly gone, seem to prove that in the great majority of cases the stilus made of iron, bronze, or other metal, or ivory or bone, sharp at one end to scratch the letters, and furnished with a knob or flat head at the other for purposes of erasure, had not gone wholly out of use. We must add to our list of Writing materials a bodkin or needle (acus), by means of which and a ruler the blank leaf was carefully divided, generally on the outer side of the skin, into columns and lines, whose regularity much enhances the beauty of our best copies. The vestiges of such points and marks may yet be seen deeply indented on the surface of nearly all manuscripts, those on one side of each leaf being usually sufficiently visible to guide the scribe when he came to write on the reverse. The quill pen [pg 028] probably came into use with vellum, for which it is obviously suited. The first notices of it occur in a story respecting Theodoric the Ostrogoth, and in a passage of Isidore's “Origines”26 (vi. 13).

8. Little need be said respecting the form of manuscripts, which in this particular (codices) much resemble printed books. A few are in large folio; the greater part in small folio or quarto, the prevailing shape being a quarto (quaternio or quire) whose height but little exceeds its breadth; some are in octavo, a not inconsiderable number smaller still: and quires of three sheets or six leaves, and five sheets or ten leaves (Cod. Vaticanus), are to be met with. In some copies the sheets have marks in the lower margin of their first or last pages, like the signatures of a modern volume, the folio at intervals of two, the quarto at intervals of four leaves, as in the Codex Bezae of the Gospels and Acts (D), and the Codex Augiensis of St. Paul's Epistles (F). Not to speak at present of those manuscripts which have a Latin translation in a column parallel to the Greek, as the Codex Bezae, the Codex Laudianus of the Acts, and the Codices Claromontanus and Augiensis of St. Paul, many copies of every age have two Greek columns on each page; of these the Codex Alexandrinus is the oldest: the Codex Vaticanus has three columns on a page, the Codex Sinaiticus four. The unique arrangement27 of these last two has been urged as an argument [pg 029] for their higher antiquity, as if they were designed to imitate rolled books, whose several skins or leaves were fastened together lengthwise, so that their contents always appeared in parallel columns; they were kept in scrolls which were unrolled at one end for reading, and when read rolled up at the other. This fashion prevails in the papyrus fragments yet remaining, and in the most venerated copies of the Old Testament preserved in Jewish synagogues.

9. We now approach a more important question, the style of writing adopted in manuscripts, and the shapes of the several letters. These varied widely in different ages, and form the simplest and surest criteria for approximating to the date of the documents themselves. Greek characters are properly divided into “majuscules” and “minuscules,” or by a subdivision of the former, into Capitals, which are generally of a square kind, fitted for inscriptions on stones like Ε; Uncials, or large letters28, and a modification of Capitals, with a free introduction of curves, and better suited for writing, like Ε; and Cursives, or small letters, adapted for the running hand. Uncial manuscripts were written in what have frequently been regarded as capital letters, formed separately, having no connexion with each other, and (in the earlier specimens) without any space between the words, the marks of punctuation being few: the cursive or running hand comprising letters more easily and rapidly made, those in the same word being usually joined together, with a complete system of punctuation not widely removed from that of printed books. Speaking generally, and limiting our statement to Greek manuscripts of the New Testament, Uncial letters or the Literary or Book-hand prevailed from the fourth to the tenth, or (in the case of liturgical books) as late as the eleventh century; Cursive letters were employed as early as the ninth or tenth century, and continued in use until the invention of [pg 030] printing superseded the humble labours of the scribe. But cursive writing existed before the Christian era: and it seems impossible to suppose that so very convenient a form of penmanship could have fallen into abeyance in ordinary life, although few documents have come down to us to demonstrate the truth of this supposition.

Besides the broad and palpable distinction between uncial and cursive letters, persons who have had much experience in the study of manuscripts are able to distinguish those of either class from one another in respect of style and character; so that the period at which each was written can be determined within certain inconsiderable limits. After the tenth century many manuscripts bear dates, and such become standards to which we can refer others resembling them which are undated. But since the earliest dated Biblical manuscript yet discovered (Cursive Evan. 481, see below Chap. VII) bears the date May 7, a.d. 835, we must resort to other means for estimating the age of more venerable, and therefore more important, copies. By studying the style and shape of the letters on Greek inscriptions, Montfaucon was led to conclude that the more simple, upright, and regular the form of uncial letters; the less flourish or ornament they exhibit; the nearer their breadth is equal to their height; so much the more ancient they ought to be considered. These results have been signally confirmed by the subsequent discovery of Greek papyri in Egyptian tombs especially in the third century before the Christian era; and yet further from numerous fragments of Philodemus, of Epicurus, and other philosophers, which were buried in the ruins of Herculaneum in a.d. 79 (“Fragmenta Herculanensia,” Walter Scott). The evidence of these papyri, indeed, is even more weighty than that of inscriptions, inasmuch as workers in stone, as has been remarked, were often compelled to prefer straight lines, as better adapted to the hardness of their material, where writings on papyrus or vellum would naturally flow into curves.

10. While we freely grant that a certain tact, the fruit of study and minute observation, can alone make us capable of forming a trustworthy opinion on the age of manuscripts; it is worth while to point out the principles on which a true [pg 031] judgement must be grounded, and to submit to the reader a few leading facts, which his own research may hereafter enable him to apply and to extend.

The first three plates at the beginning of this volume represent the Greek alphabet, as found in the seven following monuments:

(1) The celebrated Rosetta stone, discovered near that place during the French occupation of Egypt in 1799, and now in the British Museum. This most important inscription, which in the hands of Young and Champollion has proved the key to the mysteries of Egyptian hieroglyphics, records events of no intrinsic consequence that occurred b.c. 196, in the reign of Ptolemy V Epiphanes. It is written in the three several forms of hieroglyphics, of the demotic or common characters of the country, and of Greek Capitals, which last may represent the lapidary style of the second century before our era. The words are undivided, without breathings, accents, or marks of punctuation, and the uncial letters (excepting [symbol like capital Roman I] for zeta) approach very nearly to our modern capital type. In shape they are simple, perhaps a little rude; rather square than oblong: and as the carver on this hard black stone was obliged to avoid curve lines whenever he could, the forms of Ε, Ξ and Σ differ considerably from the specimens we shall produce from documents described on soft materials. Plate I. No. (1).

(2) The Codex Friderico-Augustanus of the fourth century, published in lithographed facsimile in 1846, contains on forty-three leaves fragments of the Septuagint version, chiefly from 1 Chronicles and Jeremiah, with Nehemiah and Esther complete, in oblong folio, with four columns on each page. The plates are so carefully executed that the very form of the ancient letters and the colour of the ink are represented to us by Tischendorf, who discovered it in the East. In 1859 the same indefatigable scholar brought to Europe the remainder of this manuscript, which seems as old as the fourth century, anterior (as he thinks) to the Codex Vaticanus itself, and published it in 1862, in facsimile type cast for the purpose, 4 tom., with twenty pages lithographed or photographed, at the expense of the Emperor Alexander II of Russia, to whom the original had been presented. This book, which Tischendorf calls Codex Sinaiticus, contains, besides much more of the Septuagint, the whole New Testament [pg 032] with Barnabas' Epistle and a part of Hermas' Shepherd annexed. As a kind of avant-courier to his great work he had previously put forth a tract entitled “Notitia Editionis Codicis Bibliorum Sinaitici Auspiciis Imperatoris Alexandri II susceptae” (Leipsic, 1860). Of this most valuable manuscript a complete account will be given in the opening of the fourth chapter, under the appellation of Aleph (א), assigned to it by Tischendorf, in the exercise of his right as its discoverer. Plate I. No. 2.

(3) Codex Alexandrinus of the fifth century (A). Plate I.

(4) Codex Purpureus Cotton.: N of the Gospels, of the sixth century. Plate II.

(5) Codex Nitriensis Rescriptus, R of the Gospels, of the sixth century. Plate II.

(6) Codex Dublinensis Rescriptus, Z of the Gospels, of the sixth century. Plate III.

(7) Evangelistarium Harleian. 5598, dated a.d. 995. Plate III.

The leading features of these manuscripts will be described in the fourth and fifth chapters. At present we wish to compare them with each other for the purpose of tracing, as closely as we may, the different styles and fashions of uncial letters which prevailed from the fourth to the tenth or eleventh century of the Christian era. The varying appearance of cursive manuscripts cannot so well be seen by exhibiting their alphabets, for since each letter is for the most part joined to the others in the same word, connected passages alone will afford us a correct notion of their character and general features. For the moment we are considering the uncials only.

If the Rosetta stone, by its necessary avoiding of curve lines, gives only a notion of the manner adopted on stone and not in common writing, it resembles our earliest uncials at least in one respect, that the letters, being as broad as they are high, are all capable of being included within circumscribed squares. Indeed, yet earlier inscriptions are found almost totally destitute of curves, even Ο and Θ being represented by simple squares, with or without a bisecting horizontal line (see theta, p. 35)29. [pg 033] The Herculanean papyri, however (a specimen of which we have given in Plate iv. No. 10), are much better suited than inscriptions can be for comparison with our earliest copies of Scripture30. Nothing can well be conceived more elegant than these simply-formed graceful little letters (somewhat diminished in size perhaps by the effects of heat) running across the volume, thirty-nine lines in a column, without capitals or breaks between the words. There are scarcely any stops, no breathings, accents, or marks of any kind; only that >, < or [right-pointing triangle] are now and then found at the end of a line, to fill up the space, or to join a word or syllable with what follows. A very few abbreviations occur, such as [symbol like Pi with Rho] in the first line of our specimen, taken from Philodemus περὶ κακιῶν (Hercul. Volum. Tom. iii. Col. xx. ll. 6-15), the very manuscript to which Tischendorf compared his Cod. Friderico-Augustanus (Proleg. § 11). The papyri, buried for so many ages from a.d. 79 downwards, may probably be a century older still, since Philodemus the Epicurean was the contemporary and almost the friend of Cicero31. Hence from three to four hundred years must have elapsed betwixt the date of the Herculanean rolls and that of our earliest Biblical manuscripts. Yet the fashion of writing changed but little during the interval, far less in every respect than in the four centuries which next followed, wherein the plain, firm, upright and square uncials were giving place to the compressed, oblong, ornamented, or even sloping forms which predominate from the seventh or eighth century downwards. While advising the reader to exercise his skill on facsimiles of entire passages, especially in contrasting the lines from Philodemus (No. 10) with those from the oldest uncials of the New Testament (Nos. 11-14; 17; 18; 20; 24); we purpose to examine the several alphabets (Nos. 1-7) letter by letter, pointing out to the student those variations in shape which palaeographers have judged the safest criteria of their relative ages. Alpha, delta, theta, xi, pi, omega, are among the best tests for this purpose.

These details might be indefinitely added to by references to other codices and monuments of antiquity, but we have employed most of the principal copies of the Greek Testament, and have indicated to the student the chief points to which his attention should be drawn. Three leading principles have perhaps been sufficiently established by the foregoing examples:

First, that the uncials used in writing differ from the capitals cut in stone by the curved shapes which the writing hand naturally adopts34.

Secondly, that the upright uncials of square dimensions [pg 041] are more ancient than those which are narrow, oblong, or leaning35.

Thirdly, that the simpler and less elaborate the style of writing, the more remote is its probable date.

Copies of a later age occasionally aim at imitating the fashion of an earlier period, or possibly the style of the older book from which their text is drawn. But this anachronism of fashion may be detected, as well by other circumstances we are soon to mention, as from the air of constraint which pervades the whole manuscript: the rather as the scribe will now and then fall into the more familiar manner of his contemporaries; especially when writing those small letters which our Biblical manuscripts of all dates (even the most venerable) perpetually crowd into the ends of lines, in order to save space.

11. We do not intend to dwell much on the cursive handwriting. No books of the Greek Scriptures earlier than the ninth century in this style are now extant36, though it was prevalent long before in the intercourse of business or common life. The papyri of Hyperides (e.g. No. 9) and the Herculanean rolls, in a few places, show that the process had then commenced, for the letters of each word are often joined, and their shapes prove that swiftness of execution was more aimed at than distinctness. This is seen even more clearly in a petition to Ptolemy Philometor (b.c. 164) represented in the “PalÉographie Universelle” (No. 56). The same great work contains (No. 66) two really cursive charters of the Emperors [pg 042] Maurice (a.d. 600) and Heraclius (a.d. 616). Other instances of early cursive writing may be found in two Deeds of Sale, a.d. 616, and 599, a Manumission in 355, an Official Deed in 233, a Deed of Sale in 154, in Aristotle on the Constitution of Athens, about 100, in a Farm Account in 78-79, in a Receipt in a.d. 20, in the Casati contract in b.c. 114, in a Letter on Egyptian Contracts in 146, a Treasury Circular in 170, in a Steward's letter of the third century b.c., in various documents of the same century lying in the British Museum, at Paris, Berlin, Leyden, and elsewhere, of which the oldest, being amongst the papyri discovered by Dr. Flinders Petrie at Gurob is referred to b.c. 268, and the Leyden papyrus to 26037. Yet the earliest books of a later age known to be written in cursive letters are Cod. 481 (Scholz 461, dated a.d. 835) the Bodleian Euclid (dated a.d. 888) and the twenty-four dialogues of Plato in the same Library (dated a.d. 895)38. There is reason to believe, from the comparatively unformed character of the writing in them all, that Burney 19 in the British Museum (from which we have extracted the alphabet No. 8, Plate iii), and the minute, beautiful and important Codex l of the Gospels at Basle (of which see a facsimile No. 23), are but little later than the Oxford books, and may be referred to the tenth century. Books copied after the cursive hand had become regularly formed, in the eleventh, [pg 043] twelfth and thirteenth centuries, are hard to be distinguished by the mere handwriting, though they are often dated, or their age fixed by the material (see p. 23), or the style of their illuminations. Colbert. 2844, or 33 of the Gospels (facsim. No. 39), is attributed to the eleventh century, and Burney 21 (No. 15)39, is dated a.d. 1292, and afford good examples of their respective dates. Beta (l. 1 letter 4), when joined to other letters, is barely distinguishable from upsilon40; nu is even nearer to mu; the tall forms of eta and epsilon are very graceful, the whole style elegant and, after a little practice, easily read. Burney 22 (facsimile No. 37) is dated about the same time, a.d. 1319, and the four Biblical lines much resemble Burney 2141. In the fourteenth century a careless style came into fashion, of which Cod. Leicestrensis (No. 40) is an exaggerated instance, and during this century and the next our manuscripts, though not devoid of a certain beauty of appearance, are too full of arbitrary and elaborate contractions to be conveniently read. The formidable list of abbreviations and ligatures represented in Donaldson's Greek Grammar (p. 20, third edition)42 originated at this period in the perverse ingenuity of the Greek emigrants in the West of Europe, who subsisted by their skill as copyists; [pg 044] and these pretty puzzles (for such they now are to many a fair classical scholar), by being introduced into early printed books43, have largely helped to withdraw them from use in modern times.

12. We have now to describe the practice of Biblical manuscripts as regards the insertion of ι forming a diphthong with the long vowels eta and omega, also with alpha long, whether by being ascript, i.e. written by their side, or subscript, i.e. written under them. In the earliest inscriptions and in the papyri of Thebes ι ascript (the iota not smaller than other letters) is invariably found. In the petition to Ptolemy Philometor (above, p. 41) it occurs four times in the first line, three times in the third: in the fragments of Hyperides it is perpetually though not always read, even where (especially with verbs) it has no rightful place, e.g. ετωι και αντιβολωι (facsimile No. 9, ll. 3, 4) for αἰτῶ καὶ ἀντιβολῶ. A little before the Christian era it began to grow obsolete, probably from its being lost in pronunciation. In the Herculanean Philodemus (the possible limits of whose date are from b.c. 50 to a.d. 79) as in Evann. 556, 604 (Matt. ii. 12, 13), it is often dropped, though more usually written. In Codd. Vaticanus and Sinaiticus it is probably not found, and from this period it almost disappears from Biblical uncials44; in Cureton's Homer, of the fifth or perhaps of the sixth century, ι ascript is sometimes neglected, but usually inserted; sometimes also ι is placed above Η or Ω, an arrangement neither neat nor convenient. With the cursive character ι ascript came in again, as may be seen from the subscriptions in the Bodleian Euclid and Plato (p. 42, note 1). The semicursive fragment of St. Paul's Epistles in red letters (M of St. Paul, Plate xii No. 34), used for the binding of Harleian 5613, contains ι ascript twice, but I have tried in vain to verify Griesbach's statement (Symbol. Crit. ii. p. 166) that it has ι subscript “bis tantum aut ter.” I can find no such instance in [pg 045] these leaves. The cursive manuscripts, speaking generally, either entirely omit both forms, or, if they give either, far more often neglect than insert them. Cod. 1 of the Gospels exhibits the ascript ι. Of forty-three codices now in England which have been examined with a view to this matter, twelve have no vestige of either fashion, fifteen represent the ascript use, nine the subscript exclusively, while the few that remain exhibit both indifferently45. The earliest cursive copy ascertained to exhibit ι subscript is Matthaei's r (Apoc. 502 [x]), and after that the Cod. Ephesius (Evan. 71), dated a.d. 1160. The subscript ι came much into vogue during the fifteenth century, and thus was adopted in printed books.

13. Breathings (spiritus) and accents46 were not applied systematically to Greek Texts before the seventh century. But a practice prevailed in that and the succeeding century of inserting them in older manuscripts, where they were absent prim manu. That such was done in many instances (e.g. in Codd. Vatican. and Coislin. 202 or H of St. Paul) appears clearly from the fact that the passages which the scribe who retouched the old letters for any cause left unaltered, are destitute of these marks, though they appear in all other places. Cod. א exhibits breathings, apparently by the original scribe, in Tobit vi. 9; Gal. v. 21 only. The case of Cod. Alexandrinus is less easy. Though the rest of the book has neither breathings (except a few here and there) nor accents, the first four lines of each column of the book of Genesis (see facsimile No. 12), which are written in red, are fully furnished with them. These marks Baber, who edited the Old Testament portion of Cod. A, pronounced to be by a second hand (Notae, p. 1); Sir Frederick Madden, a more competent judge, declares them the work of the original scribe (Madden's Silvestre, Vol. i. p. 194, note), and after repeated examination we know not how to dissent from his view47. So too in the Sarravian Pentateuch of the fifth century [pg 046] we read ΤΟΝΥΝ [with several symbols written over the Υ] (Lev. xi. 7) by the first hand. The Cureton palimpsest of Homer also has them, though they are occasionally obliterated, and some few are evidently inserted by a corrector; the case is nearly so with the Milan Homer edited by Mai; and the same must be stated of the Vienna Dioscorides (Silvestre, No. 62), whose date is fixed by internal evidence to about a.d. 500. In the papyrus fragment of the Psalms, now in the British Museum, the accents are very accurate, and the work of the original scribe. These facts, and others like these, may make us hesitate to adopt the notion generally received among scholars on the authority of Montfaucon (Palaeogr. Graec. p. 33), that breathings and accents were not introduced prim manu before the seventh or eighth century; although up to that period, no doubt, they were placed very incorrectly, and often omitted altogether. The breathings are much the more ancient and important of the two. The spiritus lenis indeed may be a mere invention of the Alexandrian grammarians of the second or third century before Christ, but the spiritus asper is in fact the substitute for a real letter (H) which appears on the oldest inscriptions; its original shape being the first half of the H ([symbol like the left half of a Roman capital letter H]), of which the second half was subsequently adopted for the lenis ([symbol like the right half of a Roman capital H]). This form is sometimes found in manuscripts of about the eleventh century (e.g. Lebanon, B.M. Addit. 11300 or kscr, and usually in Lambeth 1178 or dscr) ed. of 1550, but even in the Cod. Alexandrinus the comma and inverted comma are several times substituted to represent the lenis and asper respectively (facsimile No. 12): and at a later period this last was the ordinary, though not quite the invariable, mode of expressing the breathings. Aristophanes of Byzantium (keeper of the famous Library at Alexandria under Ptolemy Euergetes, about b.c. 240), though probably not the inventor of the Greek accents, was the first to arrange them in a system. Accentuation must have been a welcome aid to those who employed Greek as a learned, though not as their vernacular tongue, and is so convenient and suggestive that no modern scholar can afford to dispense with its familiar use: yet not being, like the rough breathing, an essential portion of the language, it was but slowly brought into general vogue. It would seem that in Augustine's age [354-430] the distinction between the smooth and rough breathing in the manuscripts was just such a point as a careful reader would [pg 047] mark, a hasty one overlook48. Hence it is not surprising that though these marks are entirely absent both from the Theban and Herculanean papyri, a few breathings are apparently by the first hand in Cod. Borgianus or T (Tischendorf, N. T. 1859, Proleg. p. cxxxi). One rough breathing is just visible in that early palimpsest of St. John's Gospel, Ib or Nb. Such as appear, together with some accents, in the Coislin Octateuch of the sixth or seventh century, may not the less be prim manu because many pages are destitute of them; those of Cod. Claromontanus, which were once deemed original, are now pronounced by its editor Tischendorf to be a later addition. Cod. N, the purple fragment so often spoken of already, exhibits prim manu over certain vowels a kind of smooth breathing or slight acute accent, sometimes little larger than a point, but inserted on no intelligible principle, so far as we can see, and far oftener omitted entirely. All copies of Scripture which have not been specified, down to the end of the seventh century, are quite destitute of breathings and accents. An important manuscript of the eighth or ninth century, Cod. L or Paris 62 of the Gospels, has them for the most part, but not always; though often in the wrong place, and at times in utter defiance of all grammatical rules. Cod. B of the Apocalypse, however, though of the same age, has breathings and accents as constantly and correctly as most. Codices of the ninth century, with the exception of three written in the West of Europe (Codd. Augiensis or Paul F, Sangallensis or Δ of the Gospels, and Boernerianus or Paul G, which will be particularly described afterwards), are all accompanied with these marks in full, though often set down without any precise rule, so far as our experience has enabled us to observe. The uncial Evangelistaria (e.g. Arundel 547; Parham 18; Harleian 5598), especially, are much addicted to prefixing the spiritus asper improperly; chiefly, perhaps, to words beginning with H, so that documents of that age are but slender authorities on such points. Of the cursives the general tendency is to be more and more accurate as regards the accentuation, [pg 048] the later the date: but this is only a general rule, as some that are early are as careful, and certain of the latest as negligent, as can well be imagined. All of them are partial to placing accents or breathings over both parts of a word compounded with a preposition (e.g. ἐπὶσυνάξαι), and on the other hand often drop them between a preposition and its case (e.g. ἐπάροτρον).

14. The punctuation in early times was very simple. In the papyri of Hyperides there are no stops at all, in the Herculanean rolls exceeding few: Codd. Sinaiticus and Vaticanus (the latter very rarely by the first hand) have a single point here and there on a level with the top of the letters, and occasionally a very small break in the continuous uncials, with or (as always in Cod. Ib of the sixth century) without the point, to denote a pause in the sense. Codd. A N have the same point a little oftener; in Codd. C Wa (Paris 314) Z and the Cotton Genesis the single point stands indiscriminately at the head, middle, or foot of the letters, while in E (Basil. A. N. iii. 12) of the Gospels and B of the Apocalypse, as in Cod. Marchalianus of the Prophets (sixth or seventh century), this change in the position of the point indicates a full-stop, half stop, or comma respectively. In Cod. L, of the same date as Codd. E and B (Apoc.), besides the full point we have the comma (::.) and semicolon (::), with a cross also for a stop. In Codd. Y Θa (of about the eighth century) the single point has its various powers as in Cod. E, &c., but besides this are double, treble, and in Cod. Y quadruple, points with different powers. In late uncials, especially Evangelistaria, the chief stop is a cross, often in red (e.g. Arund. 547); while in Harleian 5598 [symbol like three ovals] seems to be the note of interrogation49. When the continuous writing came to be broken up into separate words (of which Cod. Augiensis in the ninth century affords one of the earliest examples) the single point was intended to be placed after the last letter of each word, on a level with the middle of the letters. But even in this copy it is often omitted in parts, and in Codd. ΔG, written on the same plan, more frequently still. Our statements refer only to the Greek portions of these [pg 049] copies; the Latin semicolon (;) and the note of interrogation (?) occur in their Latin versions. The Greek interrogation (;) first occurs about the ninth century, and (,) used as a stop a little later. The Bodleian Genesis of this date, or a little earlier, uses (,) also as an interrogative: so in later times B-C. iii. 5 [xii], and Evan. 556 [xii]. In the earliest cursives the system of punctuation is much the same as that of printed books: the English colon (:) not being much used, but the upper single point in its stead50. In a few cursives (e.g. Gonville or 59 of the Gospels), this upper point, set in a larger space, stands also for a full stop: indeed (·) is the only stop found in Tischendorf's loti or 61 of the Acts (Brit. Mus. Add. 20,003): while (;) and (·) are often confused in 440 of the Gospels (Cantab. Mm. 6. 9). The English comma, placed above a letter, is used for the apostrophus, which occurs in the very oldest uncials, especially at the end of proper names, or to separate compounds (e.g. απ᾽ ορφανισθεντες in Cod. Clarom.), or when the word ends in ξ or ρ (e.g. σαρξ᾽ in Cod. B, θυγατηρ᾽ in Codd. Sinait. and A, χειρ᾽ in Cod. A, ὡσπερ᾽ in the Dioscorides, a.d. 500), or even to divide syllables (e.g. συριγ᾿γας in Cod. Frid.-August., πολ᾿λα, κατεστραμ᾿μενη, αναγ᾿γελι in Cod. Sinaiticus). In Cod. Z it is found only after αλλ and μεθ, but in Z's Isaiah it indicates other elisions (e.g. επ). This mark is more rare in Cod. Ephraemi than in some others, but is used more or less by all, and is found after εξ, or ουχ, and a few like words, even in the most recent cursives. In Cod. Bezae and others it assumes the shape of > rather than that of a comma.

15. Abbreviated words are perhaps least met with in Cod. Vatican., but even it has θσ, κσ, ισ, χσ, πνα for θεός, κύριος, ἰησοῦς, χριστός, πνεῦμα, &c. and their cases. [Transcriber's Note: In this e-book, underlines are used to represent the overlines used in the paper book.] The Cotton Genesis has θου ch. i. 27 by a later hand, but θεου ch. xli. 38. Besides these Codd. Sinaiticus, Alex., Ephraemi and the rest supply ανοσ, ουνοσ, πηρ (πρ Cod. Sarrav. Num. xii. 14, &c., πτηρ Cod. [pg 050] Rossanensis), μηρ, ιλημ or ιηλμ or ιλμ or ιηό (ιελμ Cod. Sarrav.), ιηϊ or ισλ or ισηλ, δαδ, and some of them σηρ for σωτήρ, υσ for υἱός, παρνξς [with a theta above the line] for παρθένος (Bodleian Genesis), σρσ for σταυρός: Cod. L has πνευ, and Cod. Vatican. in the Old Testament ανοσ and πρσ occasionally, ἐθν and ιλημ or ιλμ often51; Evan. 604 has σηρ for σωτήρ, and ἐθν for ἐθνῶν52. Cod. Bezae always writes at length ανθρωπος, μητηρ, υἱος, σωτηρ, οὐρανος, δαυειδ, ἰσραηλ, ἱερουσαλημ; but abridges the sacred names into χρσ, ιησ53 &c. and their cases, as very frequently, but by no means invariably, do the kindred Codd. Augiens., Sangall., and Boerner. Cod. Z seldom abridges, and all copies often set υἱος in full. A few dots sometimes supply the place of the line denoting abbreviation (e.g. θσ [with dots over the letters] Cotton Genesis, ανοσ [with dots over the letters] Colbert. Pentateuch). A straight line over the last letter of a line, sometimes over any vowel, indicates N (or also M in the Latin of Codd. Bezae and Claromont.) in all the Biblical uncials, but is placed only over numerals in the Herculanean rolls: κ [with tilde below and to the right], τ [with tilde below and to the right], and less often θ [with tilde below and to the right] for καί (see p. 16, note 1), -ται, -θαι are met with in Cod. Sinaiticus and all later except Cod. Z: [symbol somewhat like Arabic number 8] for ου chiefly in Codd. L, Augiensis, B of the Apocalypse, and the more recent uncials. Such compendia scribendi as [symbol like Pi with Rho] in the Herculanean rolls (above p. 33) occur mostly at the end of lines: that form, with ΜΥ [with small circle between the letters] (No. 11 a, l. 4), and a few more even in the Cod. Sinaiticus; in Cod. Sarrav. Μ [with small circle over the letter] stands for both μου and μοι; in Cureton's Homer we have Πς for πους, Σς for -σας and such like. In later books they are more numerous and complicated, particularly in cursive writing. The terminations [small raised circle] for ος, [horizontal line] for ν, [symbol like left single quotation mark] or [symbol like two left quotation marks] for ον, [symbol like right double quotation mark] for αις, [symbol like tilde] for ων or ω or ως, [symbol like raised small sigma] for ης, [symbol like raised small upsilon] for ου are familiar; besides others, peculiar to one or a few copies, e.g. τγ for ττ in Burney 19, and Burdett-Coutts i. 4, h for αυ, b for ερ, — for α, [symbol like horizontal line with circle at right edge] for αρ in the Emmanuel College copy of the Epistles (Paul 30, No. 33), and [symbol like colon] for α, [symbol like small raised capital sigma] or [symbol like raised small sigma] for αν, [symbol like square root symbol] for ας in Parham 17 of the Apocalypse. Other more rare abridgements are [symbol like two raised small sigmas] for εις in Wake 12, [square root sign] (Burdett-Coutts I. 4) or < or [raised small alpha] for εν, [symbol like two dots] for ι and [symbol like Hebrew letter Tet with two dots over it] for εσ (B-C. iii. 37), [symbol like two dots with comma below them] for εσ and ε for σε and [symbol like tilde over Rho over Tau] for τησ (B-C. [pg 051] ii. 26), [symbol] for ται and [symbol] for ωσ (B-C. iii. 42), [symbol] for ην (B-C. iii. 10), [symbol] for ισ and ὀ or [symbol] for ουν (B-C. iii. 41), [symbol] for ιν or ἐστι, [symbol] for αν, [symbol] or [symbol] for οις, [symbol] for ας, [symbol] or [symbol] for οις, [symbol] for τε or -τες or τθν or τον, [symbol] for ειν, [symbol] for ους or ως (Gale O. iv. 22). The mark > is not only met with in the Herculanean rolls, but in the Hyperides (facsimile 9, l. 6), in Codd. Vaticanus and Sinaiticus, the two Pentateuchs, Codd. Augiensis, Sangall. and Boernerianus, and seems merely designed to fill up vacant space, like the flourishes in a legal instrument54.

16. Capital letters of a larger size than the rest at the beginning of clauses, &c. are freely met with in all documents excepting in the oldest papyri, the Herculanean rolls, Codd. Vaticanus, Sinaiticus, the Colbert Pentateuch, Isaiah in Cod. Z, and one or two fragments besides55. Their absence is a proof of high antiquity. Yet even in Codd. Vaticanus, Sinaiticus, and Sarravianus, which is the other part of the Colbert Pentateuch (in the first most frequently in the earlier portions of the Old Testament), the initial letter stands a little outside the line of writing after a break in the sense, whether the preceding line had been quite filled up or not. Such breaks occur more regularly in Codex Bezae, as will appear when we come to describe it56. Smaller capitals occur in the middle of lines in Codd. Bezae and Marchalianus, of the sixth and seventh centuries respectively. Moreover, all copies of whatever date are apt to crowd small [pg 052] letters into the end of a line to save room, and if these small letters preserve the form of the larger, it is reasonable to conclude that the scribe is writing in a natural hand, not an assumed one, and the argument for the antiquity of such a document, derived from the shape of its letters, thus becomes all the stronger. The continuous form of writing separate words must have prevailed in manuscripts long after it was obsolete in common life: Cod. Claromont., whose text is continuous even in its Latin version, divides the words in the inscriptions and subscriptions to the several books.

17. The stichometry of the sacred books has next to be considered. The Greeks and Romans measured the contents of their MSS. by lines, not only in poetry, but also artificially in prose for a standard line of fifteen or sixteen syllables, called by the earliest writers ἔπος, afterwards στίχος57. Not only do Athanasius [d. 373], Gregory Nyssen [d. 396], Epiphanius [d. 403], and Chrysostom [d. 407] inform us that in their time the Book of Psalms was already divided into στίχοι, while Jerome [d. 420?] testifies the same for the prophecies of Isaiah; but Origen also [d. 254] speaks of the second and third Epistles of St. John as both of them not exceeding one hundred στίχοι, of St. Paul's Epistles as consisting of few, St. John's first Epistle as of very few (Euseb. Hist. Eccles. vi. 25, cited by Tischendorf, Cod. Sinait., Proleg. p. xxi, note 2, 1863). Even the apocryphal letter of our Lord to Abgarus is described as ὀλιγοστίχου μέν, πολυδυνάμου δὲ ἐπιστολῆς (Euseb. H. E. i. 13): while Eustathius of Antioch in the fourth century reckoned 135 στίχοι between John viii. 59 and x. 41. More general is the use of the word in Ephraem the Syrian [d. 378], Ὅταν δὲ ἀναγινώσκῃς, ἐπιμελῶς καὶ ἐμπόνως ἀναγίνωσκε, ἐν πολλῇ καταστάσει διερχόμενος τὸν στίχον (tom. iii. 101). As regards the [pg 053] Psalms, we may see their arrangement for ourselves in Codd. Vaticanus and Sinaiticus, wherein, according to the true principles of Hebrew poetry, the verses do not correspond in metre or quantity of syllables, but in the parallelism or relationship subsisting between the several members of the same sentence or stanza58. Such στίχοι were therefore not “space-lines,” but “sense-lines.” It seems to have occurred to Euthalius, a deacon of Alexandria, as it did long afterwards to Bishop Jebb when he wrote his “Sacred Literature,” that a large portion of the New Testament might be divided into στίχοι on the same principles: and that even where that distribution should prove but artificial, it would guide the public reader in the management of his voice, and remove the necessity for an elaborate system of punctuation. Such, therefore, we conceive to be the use and design of stichometry, as applied to the Greek Testament by Euthalius59, whose edition of the Acts and Epistles was published a.d. 490. Who distributed the στίχοι of the Gospels (which are in truth better suited for such a process than the Epistles) does not appear. Although but few manuscripts now exist that are written στοιχηδόν or στιχηρῶς (a plan which consumed too much vellum to become general), we read in many copies, added usually to the subscription at the foot of each of the books of the New Testament, a calculation of the number of στίχοι it contained, the numbers being sufficiently unlike to show that the arrangement was not the same in all codices, yet near enough to prove that they were divided on the same principle60. In the few documents written στιχηρῶς that survive, the length of the clauses is very unequal; some (e.g. Cod. Bezae, see the description below [pg 054] and the facsimile, No. 42) containing as much in a line as might be conveniently read aloud in a breath, others (e.g. Cod. Laud. of the Acts, Plate x. No. 25) having only one or two words in a line. The Cod. Claromontanus (facsim. No. 41) in this respect lies between those extremes, and the fourth great example of this class (Cod. Coislin. 202, H of St. Paul), of the sixth century, has one of its few surviving pages (of sixteen lines each) arranged literatim as follows (1 Cor. x. 22, &c.): εσμεν " παντα μοι εξεστι " αλλ ου παντα συμφερει " παντα μοι εξεστιν " αλλ ου παντα οικοδομει " μηδεισ το εαυτου ζη " (ob necessitatem spatii) τειτω " αλλα το του ετερου " παν το εν μακελλω πω " (ob necessitatem) λουμενον " εσθιετε μηδεν ανα " κρινωντες δια την " συνειδησιν " του γαρ κυ η γη και το πλη " ρωμα αυτης. Other manuscripts written στιχηρῶς are Matthaei's V of the eighth century (though with verses like ours more than with ordinary στίχοι), Bengel's Uffenbach 3 of St. John (Evan. 101), Alter's Forlos. 29 (36 of the Apocalypse), and, as it would seem, the Cod. Sangallensis Δ. In Cod. Claromontanus there are scarcely any stops (the middle point being chiefly reserved to follow abridgements or numerals), the stichometry being of itself an elaborate scheme of punctuation; but the longer στίχοι of Cod. Bezae are often divided by a single point.

18. In using manuscripts of the Greek Testament, we must carefully note whether a reading is We have next to give some account of ancient divisions of the text, as found in manuscripts of the New Testament; and these must be carefully noted by the student, since few copies are without one or more of them.

1. So far as we know at present, the oldest sections still extant are those of the Codex Vaticanus. These seem to have been formed for the purpose of reference, and a new one always commences where there is some break in the sense. Many, however, at least in the Gospels, consist of but one of our modern verses, and they are so unequal in length as to be rather inconvenient for actual use. In the four Gospels only the marginal numerals are in red, St. Matthew containing 170 of these divisions, St. Mark 62, St. Luke 152, St. John 80. In the Acts of the Apostles are two sets of sections, thirty-six longer and in an older hand, sixty-nine smaller and more recent63. Each of these also begins after a break in the sense, but they are quite independent of each other, as a larger section will sometimes commence in the middle of a smaller, the latter being in no wise a subdivision of the former. Thus the greater Γ opens Acts ii. 1, in the middle of the lesser β, which extends from Acts i. 15 to ii. 4. The first forty-two of the lesser chapters, down to Acts xv. 40, are found also with slight variations in the margin of Codex Sinaiticus, written by a very old hand. As in most manuscripts, so in Codex Vaticanus, the Catholic Epistles follow the Acts, and in them also and in St. Paul's Epistles there are two sets of sections, only that in the Epistles the older sections are the more numerous. The Pauline Epistles are reckoned throughout as one book in the [pg 057] elder notation, with however this remarkable peculiarity, that though in the Cod. Vatican. itself the Epistle to the Hebrews stands next after the second to the Thessalonians, and on the same leaf with it, the sections are arranged as if it stood between the Epistles to the Galatians and Ephesians. For whereas that to the Galatians ends with § 58, that to the Ephesians begins with § 70, and the numbers proceed regularly down to § 93, with which the second to the Thessalonians ends. The Epistle to the Hebrews which then follows opens with § 59; the last section extant (§ 64) begins at Heb. ix. 11, and the manuscript ends abruptly at καθα ver. 14. It plainly appears, then, that the sections of the Codex Vaticanus must have been copied from some yet older document, in which the Epistle to the Hebrews preceded that to the Ephesians. It will be found hereafter (vol. ii) that in the Thebaic version the Epistle to the Hebrews preceded that to the Galatians, instead of following it, as here. For a list of the more modern divisions in the Epistles, see the Table given below. The Vatican sections of the Gospels have also been discovered by Tregelles in one other copy, the palimpsest Codex Zacynthius of St. Luke (Ξ), which he published in 1861.

2. Hardly less ancient, and indeed ascribed by some to Tatian the Harmonist, the disciple of Justin Martyr, is the division of the Gospels into larger chapters or κεφάλαια majora64. It may be noticed that in none of the four Gospels does the first chapter stand at its commencement. In St. Matthew chapter A begins at chap. ii. verse 1, and has for its title περὶ τῶν μάγων: in St. Mark at chap. i. ver. 23 περὶ τοῧ δαιμονιζομένου: in St. Luke at chap. ii. ver. 1 περὶ τῆς ἀπογραφῆς: in St. John at chap. ii. ver. 1 περὶ τοῦ ἐν Κανᾶ γάμου. Mill accounts for this circumstance by supposing that in the first copies the titles at the head of each Gospel were reserved till last for more splendid illumination, and were thus eventually forgotten (Proleg. N. T. § 355); Griesbach holds, that the general inscriptions of each Gospel, Κατὰ Ματθαῖον, Κατὰ Μάρκον, &c., were regarded as the special titles of the first chapters also. On either supposition, however, it would [pg 058] be hard to explain how what was really the second chapter came to be numbered as the first; and it is worth notice that the same arrangement takes place in the κεφάλαια (though these are of a later date) of all the other books of the New Testament except the Acts, 2 Corinth., Ephes., 1 Thess., Hebrews, James, 1 and 2 Peter, 1 John, and the Apocalypse: e.g. the first chapter of the Epistle to the Romans opens ch. i. ver. 18 Πρῶτον μετὰ τὸ προοίμιον, περὶ κρίσεως τῆς κατὰ ἐθνῶν τῶν οὐ φυλασσόντων τὰ φυσικά. But the fact is that this arrangement, strange as it may seem, is conformable to the practice of the times when these divisions were finally settled. Both in the Institutes and in the Digest of Justinian the first paragraph is always cited as pr. (i.e. principium, προοίμιον, Preface), and what we should regard as the second paragraph is numbered as the first, and so on throughout the whole work65.

The τίτλοι in St. Matthew amount to sixty-eight, in St. Mark to forty-eight, in St. Luke to eighty-three, in St. John to eighteen. This mode of division, although not met with in the Vatican and Sinaitic manuscripts, is found in the Codices Alexandrinus and Ephraemi of the fifth century, and in the Codex Nitriensis of the sixth, each of which has tables of the τίτλοι prefixed to the several Gospels: but the Codices Alexandrinus, Rossanensis, and Dublinensis of St. Matthew, and that portion of the purple Cotton fragment which is in the Vatican, exhibit them in their usual position, at the top and bottom of the pages. Thus it appears that they were too generally diffused in the fifth century not to have originated at an earlier period; although we must concede that the κεφάλαιον spoken of by Clement of Alexandria (Stromat. i) when quoting Dan. xii. 12, or by Athanasius (contra Arium) on Act. ii, and the 3. The Ammonian Sections were not constructed, like the Vatican divisions and the τίτλοι, for the purpose of easy reference, or distributed like them according to the breaks in the sense, but for a wholly different purpose. So far as we can ascertain, the design of Tatian's Harmony was simply to present to Christian readers a single connected history of our Lord, by taking from the four Evangelists indifferently whatsoever best suited his purpose67. As this plan could scarcely be executed without omitting some portions of the sacred text, it is not surprising that Tatian, possibly without any evil intention, should have incurred the grave charge of mutilating Holy Scripture68. A more scholar-like and useful attempt was subsequently made by Ammonius of Alexandria, early in the third century [a.d. 220], who, by the side of St. Matthew's Gospel, which he selected as his standard, arranged in parallel columns, as it would seem, the corresponding passages of the other three Evangelists, so as to exhibit them all at once to the reader's eye; St. Matthew in his proper order, the rest as the necessity of abiding by St. Matthew's order prescribed. This is the account given by the celebrated Eusebius, Bishop of Caesarea, the Church [pg 060] historian, who in the fourth century, in his letter to Carpianus, described his own most ingenious system of Harmony, as founded on, or at least as suggested by, the labours of Ammonius69. It has been generally thought that the κεφάλαια, of which St. Matthew contains 355, St. Mark 23670, St. Luke 342, St. John 232, in all 1165, were made by Ammonius for the purpose of his work, and they have commonly received the name of the Ammonian sections: but this opinion was called in question by Bp. Lloyd (Nov. Test. Oxon. 1827, Monitum, pp. viii-xi), who strongly urges that, in his Epistle to Carpianus, Eusebius not only refrains from ascribing these numerical divisions to Ammonius (whose labours in this particular, as once seemed the case with Tatian's, must in that case be deemed to have perished utterly), but he almost implies that they had their origin at the same time with his own ten canons, with which they are so intimately connected71. That they were essential to Eusebius' scheme is plain enough; their place in Ammonius' parallel Harmony is not easily understood, unless indeed (what is nowhere stated, but rather the contrary) he did not set the passages from the other Gospels at full length by the side of St. Matthew's, but only these numerical references to them72.