Jump to ContentJump to Main Navigation
The Latin New TestamentA Guide to its Early History, Texts, and Manuscripts$

H. A. G. Houghton

Print publication date: 2016

Print ISBN-13: 9780198744733

Published to Oxford Scholarship Online: March 2016

DOI: 10.1093/acprof:oso/9780198744733.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (oxford.universitypressscholarship.com). (c) Copyright Oxford University Press, 2021. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use. date: 10 May 2021

Features of Latin New Testament Manuscripts

Features of Latin New Testament Manuscripts

(p.187) 9 Features of Latin New Testament Manuscripts
The Latin New Testament

H.A.G. Houghton

Oxford University Press

Abstract and Keywords

This is an introduction to the characteristics of manuscripts of the Latin New Testament and their description. The first section covers writing materials and the layout of each page, including the per cola et commata sense-line format. The second deals with scripts, abbreviations, and punctuation. The section on ‘Contents and Paratext’ includes a discussion of the order of New Testament books and details of the most common biblical prologues and chapter titles (capitula). There is an explanation of the Eusebian apparatus for identifying material found in more than one Gospel and the canons and concordances for the Pauline Epistles. The final section treats decoration and illustrations, including evangelist portraits and symbols, ‘carpet pages’, and the chi-rho page typical of Insular manuscripts.

Keywords:   Latin, New Testament, Vetus Latina, Vulgate, manuscripts, codices, Eusebian apparatus, evangelist symbols, biblical paratext, mise en page

Manuscripts of the New Testament constitute a significant proportion of the surviving Latin codices from Antiquity and the Middle Ages. The last century of studies in palaeography and codicology has resulted in a much better appreciation of these documents as artefacts. Among the chief contributors have been E.A. Lowe, editor of Codices Latini Antiquiores (CLA), which features images of all manuscripts copied before the ninth century, Bernhard Bischoff, whose publications include a standard introduction to Latin palaeography, Leonard Boyle, and Patrick McGurk.1

A) Material and Format

There are no surviving papyrus manuscripts of the Latin New Testament. Even though VL 23 and 85 have a papyrus shelfmark at their holding institution because of their Egyptian provenance, both are fragments of parchment.2 All the early witnesses, beginning with VL 1, are written on parchment. Paper was not used in the first thousand years of Latin biblical manuscripts: the only item in the Catalogue in Chapter 10 which is written on paper is VL 58, produced in Bohemia in the late fourteenth century. A relatively high proportion of early Latin gospel books are written on purple parchment (VL 2, 4, 10, 17, 22, 22A; see Image 4). Although this practice was scorned by Jerome (see page 45 above), it leaves its traces in Vulgate manuscripts: VL 15, a particularly lavish production, has alternating leaves of plain and purple parchment, while some Theodulf Bibles (e.g. VgSe Θ‎) have the Gospels on purple parchment and plain (p.188) for the rest of the New Testament. Others have significant pages painted purple or deep blue to mimic the effect of the ancient sheets which were dyed purple with the pigment of the murex shell. These include Codex Amiatinus (Vg A, fol. 3–4 of the first quire) and Codex Cavensis (Vg C), while Codex Ingolstadiensis (VgOe I) has painted purple panels.

A single leaf of parchment is also known as a folio. The front, which is on the right-hand side of an ‘opening’ of two sides, is called the recto, and the back the verso. One face is the outer side of the animal skin, the hair side, on which follicles can still sometimes be seen; the flesh side is smoother and tends to hold the ink better over time. Openings are normally arranged so that both visible pages are the same side of the skin. Vellum is a type of parchment which has been prepared in such a way as to make hair and flesh sides indistinguishable.3 Folios are normally only numbered on the recto, with the suffixes –r and –v used to indicate the front or back. If each side is given its own number, locations are referred to as ‘pages’ rather than ‘folios’. A sheet of parchment folded in half to make two folios (four sides) is called a bifolium. Manuscripts are made up of gatherings of several folded sheets sewn together along the spine. These are also called quires. The standard quire consists of four folded sheets (a quaternion), making up eight folia (sixteen pages), although the number can vary. The ‘ruling pattern’ consists of holes pricked into the quire to mark the space for the text to be copied, with lines scored between them using a dry point and straight edge (such holes are visible on Image 11 between the second and third columns). The size of the written area of the page (the ‘text block’), which may be in multiple columns, is often given in manuscript descriptions because margins may have perished or been trimmed each time the manuscript was rebound. Insular manuscripts follow a different practice for the arrangement of hair and flesh sides in a quire, allowing the two to face each other with a hair side on the first page. They also have a distinctive ruling pattern, and tend to be written not on sheepskin but on calfskin, which is more suitable for colour painting.4

Black or brownish ink is the norm, with red ink (rubrics) used in moderation from the fourth century onwards.5 In VL 3 and 5 the first three lines of each gospel are written in red, which is also used for book titles in VL 5. The fifth-century VL 8 is the earliest Latin gospel book with the first line of each chapter written in red. In his preface to the Vulgate, Jerome specifies that the Eusebian apparatus should be written in two colours, implying that minium (red) for the canon numbers was universally available alongside black.6 On purple parchment, silver replaces black as the standard ink, while gold corresponds to red. VL 2 has the opening lines of each gospel in gold, while (p.189) VL 4 uses gold for the first page of each book: both manuscripts have gold nomina sacra (described below). The use of gold ink on plain parchment is comparatively rare and restricted to luxury productions of the Carolingian and Ottonian scriptoria.7 Other colours are used for the decoration of initials and pictures (see ‘Decoration’ below): the late sixth-century Gospels of St Augustine (VgO X) offer an exceptionally early instance of full-page colour illustrations; VL 13, from around the same time, is the oldest manuscript with a pre-Vulgate text to feature multiple colours used to decorate capital letters. The use of red letters for the words of Christ is not attested in the first millennium: the earliest example in Chapter 10 is the Passion narratives in VL 62 from the eleventh century.8 The use of green or blue ink for the opening lines of biblical books is not widespread until Atlantic Bibles.

The earliest type of Latin gospel book had roughly square pages, with the text in a single column (e.g. VL 1; Image 2).9 This is seen in fifth-century manuscripts with other parts of the New Testament (e.g. VL 5, 55, and 75; Image 3). Two-column manuscripts could also be square, such as VL 4, 8, 16, and 22 (see Image 5), although the ‘portrait’ format which became standard was already present in the fourth century (VL 2 and 3; see Image 4). The use of enlarged letters (litterae notabiliores) for the first letter of each column, regardless of where it occurs in the word, is a very ancient practice.10 Double columns are the norm for pandects such as Vg A and G, Tours Bibles, Atlantic Bibles, and Paris Bibles. Nevertheless, single-column biblical manuscripts remain common, especially in the Insular world and places which experienced Irish influence: VL 14, 20, 24, 28, 29, 30, 44, 47, and 48 are all written in long lines (see Image 8), as are the Book of Durrow (VgS D) and VgOe E, L, Q, and R; the biblical books produced by Winithar at St Gall in the eighth century are also single columns (VgSaprc S). Miniature codices are usually too small to have multiple columns. The oldest examples for the Latin New Testament are VL 33 (from sixth-century Italy) and VgOe S (seventh-century Northumbria). This is also true of the Irish pocket gospel books from the ninth to eleventh centuries (Image 12).11 The long, narrow format of Vg F is unusual, if not unique (Image 7). Three columns tend to be characteristic of Spanish tradition, probably starting with the seventh-century pandects. They are found in early Theodulf Bibles, as well as the large Spanish codices of the ninth century onwards. The fifth-century VL 45, from North Italy, is a remarkable and early exception. In the later period, manuscripts of the Glossa ordinaria tend to have (p.190) three columns, with the biblical text in the central column. This is also true of certain commentaries, such as the twelfth-century VL 63. The expansiveness of the margins is an indicator of the luxury of the production, although in the case of larger Bibles the size of the sheet was dependent on the parchment available.

B) Script, Abbreviations, and Punctuation

The script of Latin New Testament manuscripts conforms to broader trends in book production. The only examples of biblical manuscripts in rustic capitals (also known as capitalis) are both from Egypt (VL 23 and 85), although this style continues to be used for headings in later manuscripts as part of a hierarchy of scripts (e.g. the explicits in Image 10).12 The majority of early New Testament manuscripts are written in capital letters known as ‘uncial’, based on Jerome’s reference in his Preface to Job (HI Jb H):

ueteres libros…uncialibus ut uulgo aiunt litteris

Old books…in inch-high/uncial letters, as they commonly say.13

Italian uncial is the standard (Image 7), and served as the model for the script developed in Wearmouth–Jarrow for the great Northumbrian manuscripts (Image 9). A less common form is b-d uncial with a distinctive form of these two letters (as in VL 5, Image 3). Hammond Bammel has suggested that half-uncial script was developed in the early fifth century in the circles of Rufinus of Aquileia for scholarly productions.14 Examples in biblical manuscripts are rare, although they include VL 53 and VgSe S (Image 6). Many later codices, however, are written in Irish or Anglo-Saxon half-uncial (e.g. VL 14, Image 8). ‘Majuscule’ is a general term for capital letters. Minuscule (lower-case) script was developed in northern France in the late eighth century. The most common variety is named after Charlemagne (‘Caroline minuscule’ or ‘Carolingian minuscule’), and adopted in the Tours Bibles (Image 10). Again, local forms developed, such as Insular minuscule in Britain and Ireland (Image 12), Visigothic minuscule in Spain (Image 11), and Alemannic minuscule in Germany and Switzerland, as well as versions peculiar to specific monasteries such as Luxeuil.15 The adoption of Gothic script, also known as textura or textualis, which is found in most Romanesque and Paris Bibles (Image 13), (p.191) seems to have been popularized by the Glossa ordinaria.16 This provided the model for the typeface of the first printed Bibles of Gutenberg.

Even before the development of minuscule, capital letters were still sometimes joined together as ‘ligatures’, especially at line ends in order to save space: the most common are –nt and –unt, found in the earliest manuscripts (e.g. VL 1 and 2: see the end of line 9 on Image 2 and the last characters of column 1 on Image 4). Later copyists often displayed their virtuosity by using a number of scripts. In a few manuscripts, normally of Insular origin, some Latin characters are transliterated into Greek as a mark of learning (e.g. VL 35 and 61, but also VL 13).

Abbreviations are rare in the earliest Latin manuscripts, with the exception of nomina sacra. These ‘sacred names’ correspond to Christian practice in Greek, where divine titles (such as God, Father, Lord, Jesus, Christ, and Spirit) are abbreviated as two- or three-letter forms with a line above. Latin Bibles follow suit for the words deus (DS), iesus (IHS), christus (XPS), spiritus (SPS), and dominus (DNS), although not pater; the final letters vary according to grammatical case. For Iesus and Christus, the first two letters of the Greek word were borrowed with the Latin termination, giving forms such as IHU or XPM. A second layer of Greek nomina sacra (including Jerusalem, Israel, heaven, and mother) is less well attested in Latin: forms such as isrl (for Israhel) tend to occur in manuscripts with many other abbreviations. Conversely, Latin writers appear to have given sanctus (‘holy’) the status of a nomen sacrum from a relatively early stage (SCS), possibly because it was frequently found in collocation with the other terms: noster too is often abbreviated in the context of other nomina sacra (NR). Some fluidity is seen in the oldest Latin nomina sacra. VL 1 features a range of peculiar forms with varying combinations such as HIS for Iesus (Image 2, line 11), DOM for dominus, and a monogram based on the chi-rho symbol (☧) for Christus. Other early manuscripts have DMS for dominus (e.g. VL 75); VL 5 has DMS in the Gospels, but DNS in Acts. Rufinus of Aquileia appears to have introduced IS rather than IHS in his works: he also preferred DOM for dominus.17 Around the eighth century, Irish scribes restored the Greek lunate sigma in the nominative of the Greek-based forms, reading IHC and XPC: this practice was also adopted in Carolingian scriptoria.18 VL 11 is unusual in using XRS (also present in VgSp s). The origin of the practice of nomina sacra is still a matter of debate, and it is intriguing that there is not more explicit discussion of it by Latin authors. Differences in the forms of Latin nomina sacra have also been taken as indications of geographical origin, since they occur in inscriptions carved into burial stones or buildings.19

(p.192) The commonest abbreviations in uncial script are the use of a superline for n or m, often at the end of lines (e.g. Image 2, lines 5 and 11). These are sometimes distinguished by the placing of one or two dots below the line when m is intended (VL 4, cover; Vg F, Image 7, lines 18 and 30). Traube suggests that this practice, like the nomina sacra, was borrowed from Greek biblical tradition.20 The ending –bus is abbreviated as – and –que as – in early manuscripts (e.g. Image 7, line 22). A system soon develops for reasons of economy, with words such as dix(it), dic(it), q(uo)d, n(on), and –er(unt) all having the letters in parentheses replaced by an overline. Later innovations include the creation of symbols for words and syllables such as per, pro, quia, quod, quam, the ampersand (&) for et, and so on.21 Many Latin abbreviations appear to have originated in Ireland. The symbols Features of Latin New Testament Manuscripts for autem and ⧺ for enim are peculiar to Insular manuscripts (Features of Latin New Testament Manuscripts features in Image 12, lines 2, 10, and 16). Others were adopted in continental minuscule scripts, including con– represented as ɔ–, eius as ɘ, et as Features of Latin New Testament Manuscripts, and est may be abbreviated as or ÷ (Image 12 has ‬ in line 3 onwards, and a form of ÷ in line 12, 21, and 22).22 The use of ·· for est originated in Tours.23 Some of these standard abbreviations are comparable to Latin shorthand, or ‘Tironian notes’, which is occasionally seen in marginal comments (e.g. VL 7).24

Sound changes are also reflected in spelling conventions, with the use of the digraphs æ and œ and the e-caudata (‘e with a tail’, ę). The latter is found in place of ae in the eighth century, becoming very common in the tenth and eleventh centuries before being replaced by e alone in the twelfth century.25 Later manuscripts tend to assimilate consonants, as in app– and suff– instead of adp– and subf– respectively. The development of consonantal u and i to v and j respectively takes place at different times: the former is relatively early, as seen in betacism (b for u) in VL 5, whereas the latter is largely a late mediaeval innovation.

The earliest New Testament manuscripts are written with no breaks between words (scriptio continua) and straight left and right margins (see Image 2).26 The beginning of sense units on a new line, together with ekthesis (letters projecting into the margin), and the use of outsize or capital letters (known as litterae notabiliores, ‘more prominent letters’) constitute the most basic form of (p.193) punctuation. Bilingual codices favour short sense lines, in order to keep the two languages in correspondence (e.g. VL 5, 50, and 75; see Image 3). In his preface to Isaiah (HI pr Is), Jerome stated that he had introduced this format, which he describes as per cola et commata, into his translations of the prophetic books.27 In Codex Amiatinus (Vg A; Image 9), it has been extended to the entire Bible, with each short sense unit marked by ekthesis and little in the way of other punctuation. This type of line division, however, is relatively inefficient as it leaves a lot of unused parchment. The most common layout in biblical manuscripts is to have longer sections with a higher proportion of words split between lines, and straight margins on each side. This can be seen in the earliest surviving manuscript of Jerome’s text of the Gospels (VgSe S; Image 6). In Insular manuscripts, text is often copied in blank space left at the end of the previous section or on the following line after the opening part of a new section. These are called ‘runovers’, and are separated from the rest of the text by diagonal lines or other markings.28

The full system of punctuation developed by grammarians in the fourth century consists of dots (distinctiones) at three different heights relative to the line: a high point (distinctio), middle point (media distinctio), and low point (subdistinctio).29 This is largely absent from contemporary biblical manuscripts, in which blank spaces of differing lengths are sometimes used to indicate breaks (e.g. Images 2 and 5). Regular gaps between words appeared to have been introduced by Irish scribes in the seventh century (Image 8). The separation of words by dots (interpunction) is occasionally found as an intermediate stage, as in VL 11A. Insular copyists employed a number of other devices to assist readers of Latin texts, including a ‘diminuendo’, capital letters in decreasing size, at the beginning of a new section. They often used combinations of dots to signify pauses, in which the number of marks was equivalent to the significance of the break, although such patterns merge into decorative devices. Carolingian scriptoria developed a simplified form of the earlier system of punctuation, comprising a binary distinction between high points at the end of sentences and medial points for intermediate pauses. However, the late eighth century also saw the adoption of different symbols, later known as positurae, for liturgical readings. These are made up of dots and lines, with the punctus uersus (;) at the end of a sentence, the punctus eleuatus (.’) in the middle of a sentence, and a new symbol, the punctus interrogatiuus (.~) corresponding to the modern question mark. From the tenth to the twelfth centuries, the punctus flexus (.u) was inserted for minor pauses in the middle of sense-units. In many cases, the punctuation in surviving New (p.194) Testament manuscripts has been added or revised by later hands, frequently involving the introduction of a more recent system into an older manuscript (e.g. VL 8).

Quotations, normally from the Old Testament, may be indicated in a number of ways. Some New Testament codices use indentation (e.g. VL 8, Image 5, column 4) or rubrication (e.g. VL 75). A marginal symbol known as a diple, initially shaped like a horizontal arrow-head (>), is often placed at the first line or beside each line of a quotation. In later manuscripts this becomes simply a short wavy line. Diplai are often used in copies of biblical commentaries to indicate the scriptural lemma.30

C) Contents and Paratext

Manuscripts of the entire Bible, or even just the New Testament, are unusual before the ninth century. Constraints of space, as well as custom, meant that books generally circulated in subgroups. The illustration of the bookcase in Codex Amiatinus has the New Testament divided into three volumes: the Gospels, the Epistles, and Acts and Revelation.31 The two principal collections in surviving manuscripts are the Gospels and the Pauline Epistles. Figures from CLA show that forty-one surviving manuscripts of the Gospels and eleven of the Pauline Epistles were copied before 600, while 119 gospel manuscripts and twenty-five of the Pauline Epistles are preserved from 600 to 800. The figure for the rest of the New Testament (Acts, Catholic Epistles, and Revelation) is just five and seven respectively.32

Exceptions to the standard groupings are rare: VgSpc K consists of both the Pauline and the Catholic Epistles, as does VgSl Q; VL 66 combines the Letter of James with patristic texts (cf. VgSc R, with all the Catholic Epistles); VL 7 includes the Shepherd of Hermas after the New Testament, while VL 51, VL 61, and the Theodulf Bibles also have non-biblical material. Some manuscripts consist of single books, John and Revelation being the most common (e.g. VL 33, 47, VgOe S, and VgO Σ‎). Lectionaries provide only the passages read in the liturgy: these may be confined to the Gospels (sometimes called an ‘evangelistary’) or the Epistles (e.g. VL 87), but the majority contain extracts from throughout the New Testament.33 Irish (p.195) Pocket Gospels often have an abbreviated text.34 It is, of course, impossible to be sure of the contents of manuscripts which are now fragmentary, although details such as page numbers or quire signatures may offer a guide: for example, Codex Bezae cannot have contained the entire New Testament although a substantial portion is missing between the end of Mark and 3 John, which may have been sufficient to contain Revelation and the other Johannine Epistles.

The order of the constituent subgroups in complete New Testaments and lists of canonical books varies.35 The modern sequence, with Paul following Acts (eapcr), is found in Codex Amiatinus (Vg A; also VL 54, 95, Vg Φ‎V). A number of manuscripts place the Pauline Epistles after Revelation (eacrp; e.g. VL 6, 7, 51; VgS Φ‎T, s). Tours Bibles tend to have Paul between the Catholic Epistles and Revelation (eacpr; VgS Φ‎E, Φ‎G), an order which persists into the thirteenth century (VgO W). In Theodulf Bibles and the Spanish tradition, Paul follows the Gospels and Acts is located between the Catholic Epistles and Revelation (epcar; VL 67, 91, 92, 109; VgO C, T, Θ‎). VL 61 places Paul second and Acts at the end (epcra). The pseudonymous Letter to the Laodiceans is found in several New Testament manuscripts (e.g. Vg F, VL 6, 51, 58, 61, 62, 91, 109); 3 Corinthians, by contrast, is very poorly attested (VgS B and the table of contents in VL 86). A few biblical codices are ordered according to the sequence of liturgical lections for the year, such as VL 135 and VgS B.36 Very occasionally, the Gospels are omitted from later New Testament manuscripts because of the use of a separate gospel book for the liturgy.

Most Old Latin gospel books have the evangelists in the sequence Matthew–John–Luke–Mark, sometimes described as the ‘Western’ order. There is evidence for a variety of early sequences: VL 1 has Mark followed by Matthew, while Matthew–Mark–John–Luke was found in the ancestor of VL 5 and Mommsen’s stichometry.37 Jerome explicitly mentions in his preface to the Gospels that he put the books into the same order as Greek (Matthew–Mark–Luke–John) in order to include the Eusebian apparatus. Several sequences are found for the Pauline Epistles, with or without Hebrews or Laodiceans.38 The principal variation involves the placing of Colossians:



VL 75


VL 86 (also Marius Victorinus and Ambrosiaster).


Pelagius and some Vulgate manuscripts


VL 7, 54, 78, Vg A, many Vulgate manuscripts


VL 61, Vg F, VgOp M

Eph–Phil–12Thess–Col … Hbr–Laod

VL 51, 109


Spanish Vulgate manuscripts (Vg C, Peregrinus).

Eph–Phil–Col–12Thess (Laod Hbr)

VL 77

Eph–Phil–Col–12Thess Hbr–Laod

VgOa U


VgSp S

The pre-Vulgate tradition of the Catholic Epistles put the two Epistles of Peter first, although no Old Latin manuscripts with all seven letters survive. Augustine’s Speculum and the stichometric list of VL 75 have the order Peter–James–John–Jude, while Peter–John–James–Jude is supported by the list in the Breviarium Hipponense adopted by the Council of Carthage in 397 and seems to underlie VL 55 and the Pseudo-Augustine Speculum.39

The first line of paratextual material in biblical codices is the title of each book, given as an incipit before the text and an explicit after the final verse. Early Latin gospel manuscripts and authors borrow the Greek word for ‘according to’, cata (e.g. VL 1, 2, Cyprian). Although secundum is soon established as the standard form, cata continues to be found in later manuscripts such as the ninth-century VL 61 and certain sequences of chapter titles (see below). The names of the evangelists are fairly consistent, although Lucanus is found instead of Lucas in some Old Latin witnesses (e.g. VL 3, 8, 16, 21, Cyprian and Fortunatianus of Aquileia). The Pauline Epistles, along with the latter Epistles of Peter and John, often have eiusdem (‘from the same person’) rather than repeating the author’s name. As noted in Chapter 8, the Catholic Epistles tend to be designated as the Epistulae Canonicae in Latin manuscripts, while Revelation is universally referred to as Apocalypsis.40 An abbreviated form of the title normally appears at the top of each page, known as the ‘running title’. A few Insular manuscripts have O Emmanuel as the running title in place of the name of the evangelist (Cadmug Gospels and the Liber Commonei). Additional material is frequently added to the explicit. In VL 1 and VL 13, it is recorded that Mark ends ‘happily’ (explicit feliciter; in VL 13 this is written in Greek letters; see also Philippians in VL 89), while VL 14 adds the word for ‘peace’ (pax) three times.41 In the Pauline Epistles, there is a tradition in both Greek and Latin of adding details of the location at which the (p.197) letter was believed to have been written (e.g. VL 75). This is also found in the Marcionite prologues.42 A colophon is often found at the end of a volume, in which the name of the copyist may be mentioned: a number of manuscripts are known by the name of their copyist, especially in Irish and Spanish tradition.43

Biblical prologues are generally identified by the numbers assigned by Stegmüller.44 Jerome’s Preface to the Gospels, a dedicatory letter to Pope Damasus also known as the Epistula ad Damasum, is found in most copies of the Vulgate Gospels.45 It is often identified by its opening words, Nouum opus (HI Ev [S 595]), and accompanied by Plures fuisse, the preface to Jerome’s commentary on Matthew (HI Mt [S 596]) or, later, Sciendum etiam [S 601]. Some codices with Northumbrian connections have a Latin translation of Eusebius of Caesarea’s Letter to Carpianus, beginning Ammonius quidam ([S 581], e.g. VL 9A and VgOe Y). VL 10 has a unique preface to the Gospels from a Gothic milieu discussing the problems of biblical translation (Sanctus Petrus apostolus).46

Prologues to the individual gospels are rarely present in Old Latin tradition, but the two series commonly found in Vulgate manuscripts date back to the fourth century.47 The most widespread are known as the ‘Monarchian’ prologues (PROL Mo), because they were originally believed to derive from Monarchian circles in Rome in the second or third century. Subsequent scholarship has suggested that they are Spanish in origin, possibly from the circles of Priscillian. Their original sequence matches that of Old Latin gospel codices. Each gives brief details about each evangelist, beginning as follows:

  • Matthaeus ex Iudaea sicut in ordine primus ponitur… [S 590, 591]
  • Marcus euangelista dei et Petri in baptismate filius… [S 607]
  • Lucas Syrus natione Antiochensis arte medicus… [S 620]
  • Hic est Iohannes euangelista unus ex discipulis dei… [S 624]

They appear to have been composed in response to a set of ‘anti-Marcionite’ prologues (PROL Pa), written in the middle of the fourth century, which only survive for the latter three gospels:

  • Marcus adseruit qui colobodactylus est… [S 604]
  • Est quidem Lucas Antiochensis Syrus… [S 612]
  • Euangelium Iohannis manifestatum est… [S 623]

The ‘anti-Marcionite’ prologue to Luke is the only one to be included in VL 8. Alternatives to these two sets of prologues are only found in a handful of Irish and Spanish manuscripts, and are late derivatives. Other prefatory material sometimes found before each gospel includes a list of Old Testament citations made by the evangelist (PROL Te [S 10236]), and Jerome’s glossary of Hebrew words in that book (HI nom [S 10235]). An abbreviated form of the latter is first attested in VL 14, but clearly predates the seventh century and is found in a number of Insular manuscripts, including VgOe Q. The list was considerably expanded for the Paris Bibles in the thirteenth century.48 Brief biographies from Jerome’s De uiris illustribus (HI ill) or Isidore’s De ortu et obitu patrum (IS ptr) are sometimes used to introduce each author. In Insular manuscripts, the first seventeen verses of Matthew are reinterpreted as a prologue, with a decorated chi-rho page beginning Matthew 1:18; this is also the case to a lesser extent at Luke 1:5 (e.g. VL 11A and 29).

Primum quaeritur (PROL Paul 1 [S 670]), an introduction to the entire Pauline corpus, is found before the Epistle to the Romans in the Vulgate. This was the work of the translator responsible for the revision of these letters, and probably the rest of the New Testament. The argumentum for Romans (PROL Rm Arg [S 674]) appears to be from the same pen. The most common set of prologues for the individual Pauline Epistles by far are the Marcionite prologues (PROL Ma [S 677, etc.]), preserved only in Latin tradition. There has been considerable debate as to whether these derive from Marcion’s collection of selected letters of Paul, suggested by the identification of Ephesians as Laodiceans as well as the correspondence between the original set of seven prologues and the unusual order of Marcion’s collection reported by Tertullian (Galatians, Corinthians, Romans, Thessalonians, Ephesians/Laodiceans, Colossians, and Philippians). The most recent evaluation puts forward a strong case in favour of their Marcionite origin.49 They are attested in Marius Victorinus and Latin manuscripts from the fourth century onwards in a revised form which includes additional prologues in order to provide one for each canonical Epistle. These were also adopted as the standard prologues to Paul in Paris Bibles.50 Alternative sets of Pauline prologues include one attributed to Pelagius (PROL Pel), one of French and two of Italian origin (PROL Fr, PROL It 1 and 2), and a series composed by Bede (BED pr/PROL Eln). In addition, lists are sometimes provided of the quotations from the Old Testament in each Epistle.

The most common prologues to Acts and Revelation, composed before the sixth century, draw on the ‘Monarchian’ prologues to Luke and John: Lucas (p.199) natione syrus cuius laus in euangelio canitur (PROL Act Mo; S 640) and Iohannes apostolus et euangelista (PROL Apc Mo; S 834/5). Other prologues are found before Acts, as well as lists of names, places, and biblical citations. The principal prologue to the Catholic Epistles is a later addition, not by the translator, which includes a polemic about the omission of the Johannine Comma: Non ita (or idem) ordo est apud graecos (PROL cath; S 809). Like the ‘Monarchian’ prologues to Acts and Revelation, this is first attested in Vg F. Individual letters are often prefaced by the biographical material from Jerome or Isidore mentioned above. Alternatives to the ‘Monarchian’ prologue of Revelation are a preface of Spanish origin, Isidore’s prologue, Jerome’s biography of the evangelist, or even an extract from Jerome’s Epistula 53 to Paulinus of Nola stating that Revelation has as many signs as words (tot habet sacramenta quot uerba): this letter is used as a preface for the Bible in Vg A and Tours Bibles.

The creation of sets of chapter titles (capitula) is already attested in the third century: one series for the Gospels predates Cyprian (KA Cy), while Donatists created their own sets for the other books of the New Testament.51 In manuscripts, they are introduced by a variety of terms including breues, breuiarium, tituli, elenchus, breues causae, capitulatio, capitula lectionum. No fewer than thirteen types are attested for each Gospel, of which around half appear to derive from Old Latin texts. Their chronology has already been discussed in Chapter 8. The titles were linked to numbered divisions in the text, although the two sometimes become separated: marginal chapter numbers in some manuscripts do not correspond to the initial titles (e.g. VL 9A). The length of the divisions varies considerably: Matthew, for example, ranges from twenty-eight to one hundred chapters. VL 8 has one of the earliest full sets (Image 5), but unexpected numbers in the margin or text of other Old Latin manuscripts may indicate the presence of chapters in their exemplar (e.g. Matthew 28:1 in VL 3). In Vulgate tradition, the capitula normally come between the prologue and the biblical text (e.g. Image 10), although in certain manuscripts the capitula for a group of books are placed at the beginning of the collection (e.g. VL 7). The use of Greek numerals for chapters is an indication of North Italian origin.52 One series of capitula has been identified as the work of Bede.53 In the early Tours Bibles and Paris Bibles, the series of capitula varies from manuscript to manuscript although a degree of standardization emerges later on. Latin capitula do not correspond to the divisions or titles of the Greek kephalaia, although one set of capitula for Hebrews is a translation of Euthalian material (KA Heb Z).

(p.200) The Eusebian apparatus included by Jerome is a system for identifying shared material between the four Gospels. It consists of two parts: an initial set of ten canon tables showing the different types of agreement, and a series of numbers in the margin of the biblical text. Each Gospel is divided into sequentially numbered ‘Eusebian sections’ (or sometimes ‘Ammonian sections’), the usual figures being 355 in Matthew, 233 in Mark, 342 in Luke, and 232 in John. The ten ‘Eusebian canons’ cover the various permutations of overlap: Canon I lists the sections which contain material paralleled in all four Gospels (see Image 15), Canons II–IV similarities between three Gospels, Canons V–IX agreements in two Gospels only, and Canon X, split into four, material unique to each Gospel.54 Jerome states that the section numbers and canon numbers in the margin should be written in different colours:

hic nigro colore praescriptus sub se habet alium ex minio numerum discolorem, qui ad decem usque procedens indicat, prior numerus in quo sit canone requirendus.

The first number, written in a dark colour, has another number underneath it in a different colour from red pigment, which goes up to ten and indicates in which canon table the first number should be sought.

Despite the reference only to a single section number, from a very early stage the numbers of the parallel sections from the other gospels were also added in the margins (e.g. VL 23, VgSe S; see Image 6). This is clearly secondary to Jerome as it obviates the need for the initial canon tables. Nevertheless, they were generally retained and become ever more decorated. In fact, the Eusebian apparatus is far better attested in Latin manuscripts than in Greek tradition.55 In Greek canon tables, the numbers are usually given in groups of four, whereas in their Latin counterparts they tend to be in groups of five, or with no divisions at all: VL 7, with groups of four, is an exception.56 Nordenfalk identifies three different types of Latin canon tables: the first, ‘shorter Latin series’, occupies twelve pages, while the second and third both have sixteen pages.57

Alternatives to Jerome’s implementation of the Eusebian apparatus suggest that he was not the only scholar to transfer it to Latin. In VL 10 (which preserves the Old Latin order of the Gospels) and VL 11, the section numbers are written alongside the text and the corresponding passages in the other gospels are identified by number in four decorative arches at the bottom of each page. In VL 5, the section numbers but not the canon numbers are written alongside the Greek text. Further evidence of pre-Vulgate use of the (p.201)

Features of Latin New Testament Manuscripts

Image 15 Canon Tables in the Lindisfarne Gospels: VgO Y (London, British Library, MS Cotton Nero D.IV, folio 10r), a page from the initial canon tables. © The British Library Board. Also available at <http://www.bl.uk/manuscripts/Viewer.aspx?ref=cotton_ms_nero_d_iv_f010r>

(p.202) Eusebian divisions is provided by the four manuscripts which include the opening words of each section in the initial canon tables with an Old Latin text (VL 10, 39, 40, and 46).58 Errors in the numeration of the Eusebian apparatus may often serve as markers of genealogical relationship: for example, the exemplar of Jerome’s Vulgate appears to have placed several section numbers one verse early.59 The Eusebian section numbers were also used to identify gospel passages in lists of liturgical lections, such as the Capitulare euangeliorum de circulo anni (described below).

On the model of the Eusebian canons, Priscillian of Avila developed canons for the Pauline Epistles. These group together similar material from the fourteen epistles under ninety headings, preceded by a preface (Prologum subter adiectum…multis occupatus necessitatibus…corrigere mentes [PRIS can pr; S 672, 656]). Each canon is identified by a number preceded by the letter K, which may also appear in the margin of the biblical text.60 These are a staple of Spanish biblical manuscripts, but are by no means confined to them: they are also adopted in Theodulf Bibles and other codices. Priscillian may have been inspired by Jerome’s treatment of the Eusebian apparatus, as both refer to the use of contrasting colours of ink.61 An alternative concordance to the Pauline Epistles is the Concordia epistularum attributed to Pelagius, which groups the thirteen epistles under fifty-six headings (AN conc; S 646). This is found in Vg A, F, and a number of Tours Bibles and Atlantic Bibles. The headings of the two series are easily differentiated: Priscillian’s canons usually begin with Quia, and the titles of the Pelagian concordance titles all begin with De. Some Tours Bibles have the Pelagian concordance in a set of arcades similar to the Eusebian canon tables.

Several poems are included among the ancillary material of New Testament manuscripts. The four evangelists and their symbols are the subject of a number of short verses found at the beginning of the Gospels or alongside illustrations. Some extracts are taken from Sedulius’ Paschale Carmen (e.g. VL 62) or Juvencus (e.g. Cadmug Gospels). Three of the most common poems concern the content and division of the Eusebian canons; Aileran’s Quam in primo speciosa quadriga (AIL Eus; S 843), Alcuin’s Textum si cupies canonis (S 851), and the anonymous In primo certe canone (S 848). Although the other books of the New Testament have inspired less poetry, a particularly ingenious poet managed to conclude a poem on the Pauline Epistles (Uersus quot et (p.203) quibus missae sunt epistolae, S 681) with a mnemonic for all fourteen in a single quasi-hexameter line:

  • Ro cor bis gal ephe phi co te bis timo bis ti phil hebre.
  • Romans, Corinthians twice, Galatians;
  • Ephesians, Philippians, and Colossians;
  • Thessalonians twice, and Timothy’s two;
  • Titus, Philemon, and one to Hebrews.

Lectionary manuscripts are each affiliated to a particular liturgical tradition.62 Each extract is usually preceded by an indication of the appointed day for the reading and its source. The beginning of a lection is often adjusted to enable it to be read out of context: fratres (‘brethren’) is usually added as the first word in extracts from the Epistles, while in illis diebus (‘in those days’) or dixit Iesus discipulis suis (‘Jesus said to his disciples’) may preface readings from the Gospels. The influence of lectionaries can sometimes be seen in continuous-text manuscripts, with liturgical indications added in the margins by later hands (e.g. VL 13) or the incorporation of the initial formulae (e.g. VL 29). The reverential addition of dominus before Iesus (e.g. VL 11A) may also reflect lectionary influence. The Passion narratives are occasionally marked up for dramatic reading, for example by the insertion of a cross above the words of Jesus (e.g. VgOe R). Many gospel books have a table of liturgical readings, the Capitulare euangeliorum de circulo anni [S 852]. Beginning with the Christmas reading In natale domini and normally ending with the second week after Pentecost (Ebd. II post Pentecosten), these 192 entries enable the gospel book to be used as a lectionary.63 The individual readings are identified by the Eusebian section number in each gospel, with the opening and closing words provided in the table for precision. This table is widely attested in manuscripts from the ninth century onwards.

Various other miscellaneous material may be found in the margins or at the beginning or end of New Testament manuscripts. Proper nouns often identify the book’s owners, written in normal characters, runes (e.g. VL 9A, 65), or even Ogham (Stowe St John). Random letters or words are sometimes described in catalogues as ‘testing of the pen’ (probatio pennae), and could have been added at any point in the history of the manuscript. Missing text is sometimes noted by the letters hd (haec desunt), with the corresponding text added in the margin with the letters hp (haec pone) or sr (sequitur).64 In addition to the use of the margin by correctors, it is occasionally employed by copyists for noting alternative readings (e.g. VL 88, 91, 95, 109, etc., VgOe EP) (p.204) or, in later manuscripts, cross-references (VL 58 and the Paris Bibles; see Image 13). There is sometimes an indication that a manuscript has been compared with its exemplar in the scriptorium, with the word contuli or req(uisitum).65 Books may be followed by stichometric numbers indicating the total lines. These are found in Old Latin witnesses (e.g. VL 64) as well as several ninth-century productions (e.g. VL 7, VL 61, VgSp R, VgSp S, VgO L): the total lines for Romans in VL 7 can be seen in the last line of the first column in Image 16. Numerals were often garbled during transmission and do not necessarily relate to the preceding text: sometimes the stichoi appear to correspond to Greek rather than Latin texts.66

D) Decoration

The earliest Latin New Testament manuscripts have little in the way of decoration to the text, apart from lines or dashes surrounding the book titles (e.g. VL 1 and 5). From the sixth century, capital letters start to attract decoration either through the use of several colours of ink or the introduction of zoomorphic elements such as animal and bird heads or leaves (e.g. VL 13).67 Outsize initials afford the opportunity for more complex ornamentation, whether in the form of interlace and geometric patterns (seventh century onwards) or, eventually, human figures (e.g. VL 6). The latter develop from around the ninth century, as in the Moutier-Grandval Bible (VgS Φ‎G), with a head of Paul in the first initial of Romans, or the depiction of John on Patmos in Revelation in VgS I. They reach their apogee in Romanesque Bibles of the thirteenth and fourteenth centuries; the Paris Bibles produced for academic purposes are less lavish. The Eusebian apparatus also affords an opportunity for artistic creativity, either in the architectural structures enclosing the initial canon tables (see Image 15) or the arcades at the bottom of the page in VL 10 and 11.68

The style of the illustrations in a few manuscripts could derive from classical models. The Gospels of St Augustine (VgO X), produced in Italy in the sixth century, provide the earliest surviving full-page colour illustrations in a Latin New Testament manuscript. The two surviving pages feature twelve scenes from the Passion and twelve episodes from the Gospel according to Luke with a picture of the evangelist. Their depiction of events in the life of Christ is (p.205)

Features of Latin New Testament Manuscripts

Image 16 VL 7, Vg G: Codex Sangermanensis primus (Paris, Bibliothèque nationale de France, latin 11553, folio 164v), showing the end of Romans and beginning of 1 Corinthians. Also available at <http://gallica.bnf.fr/ark:/12148/btv1b9065916g/f158.item>

(p.206) unique in surviving Latin tradition, although there are contemporary parallels in Greek and Syriac manuscripts.69 The Trier Apocalypse (VgO Σ‎) has a series of seventy-four full-page colour images whose exemplar may go back to fifth-century Italy. A large number of single-volume illustrated Apocalypses were produced from the ninth century onwards, of which the Trier Apocalypse and Valenciennes, BM, 99, are the earliest.70 Bede records that Benedict Biscop brought back a series of paintings from Italy to Wearmouth–Jarrow in 675, including a cycle based on the Apocalypse, to hang in the monastery church.

The first quire of Vg A (Codex Amiatinus) has two illustrations which have been connected with those of Cassiodorus’ sixth-century codex grandior. These comprise an illustration of the Tabernacle, across a single opening, and a copyist seated in front of a bookcase containing a nine-volume Bible. Cassiodorus’ Institutiones indicate that his pandect also featured an illustration of the Temple in Jerusalem, which is not present in Vg A. Instead, its other pages contain three diagrams of different systems of grouping the books of the Bible and a page with a painted series of roundels describing the contents of each book of the Pentateuch. The only other picture in Vg A is a magnificent full-colour image preceding the New Testament. A central roundel depicts Christ in majesty, flanked by two angels, against a background of dark blue concentric circles studded with white points, while in the corners of the rectangular frame are the four evangelists, each with their symbol.

The association of each Gospel writer with the four creatures listed in Revelation 4:7, a man, lion, ox and eagle, is found in numerous early Christian writers although the correspondences vary. The earliest surviving reference, in Irenaeus of Lyons, identifies Matthew with the man, John with the lion, Luke with the ox, and Mark with the eagle.71 This is the same sequence as the most common order of the Gospels in Old Latin codices, and also features in the seventh-century Book of Durrow (VgS D). Nevertheless, the standard sequence in Latin and Greek tradition identifies Mark with the lion and John with the eagle. Evangelist symbols are particularly favoured in Insular tradition. The Echternach Gospels (VgO EP) has only a full-page symbol before each Gospel rather than a human figure, while other gospel books with evangelist portraits also include a symbol page depicting all four creatures (e.g. VgS D, VgOe L, VgO Q). Full-page portraits of evangelists occur across all traditions of the New Testament, although allusion to the evangelist symbols (p.207) in these illustrations tends to be peculiar to Latin tradition.72 In Irish manuscripts, the evangelists are always presented facing the reader directly rather than in profile.73

Insular gospel books from the seventh century onwards have a particular set of illustrations, in addition to evangelist portraits and a four-symbols page. ‘Carpet pages’, entire pages with a colourful geometric design (VgSe D; VgOe L, Q, Y), are found either at the beginning of the manuscript or before each Gospel. The opening page for each book is highly stylized, with the initial letters woven into a full-page design and decorated with interlacing patterns. This is also the case for the chi-rho at Matthew 1:18, based on the nomen sacrum which begins the verse (XPI autem generatio).74 The illustration of Matthew sometimes appears opposite this page rather than at the beginning of the Gospel. At the end of VL 48, copied in Ireland at the end of the eighth century, are full-page colour pictures of the Crucifixion and Christ in Majesty (or possibly the Last Judgement). The Book of Mulling (VL 35) contains a line drawing at the back of the manuscript which may be a plan of the monastery.

A sequence of five images became standard for gospel books copied in Tours around 830. This consisted of the four evangelist portraits and an initial Christ in Majesty surrounded by the evangelists and their symbols, comparable to that described above in Codex Amiatinus. Later in the ninth century, two more illustrations were added at the beginning of Tours Gospels depicting the emperor on his throne and the twenty-four elders adoring the lamb (from Revelation 5:8). Two styles of decoration can be discerned in these manuscripts.75 The Moutier-Grandval Bible (VgS Φ‎G) is a lavish production, although it has no evangelist portraits. It features full-page colour pictures of Christ in Majesty before the Gospels and the book with seven seals flanked by evangelist symbols after the conclusion of Revelation, and decorated arcades for the canon tables and the concordance to the Epistles. Miniature illustrations are added in outsize capitals: in Acts, these include human heads and the Lukan evangelist symbol, while capitals in Romans contain Paul and a man with a sword. Initial illustrations of Christ in Majesty are also found in the eighth-century VL 39 and, more surprisingly, at the beginning of the Pauline commentary in VL 89. Certain Latin scholars in the time of Charlemagne, however, among them Theodulf and Claudius of Turin, were iconoclasts, disapproving of sculpture and human representation.

Spanish pandects are illustrated in a distinctive style.76 Chief among them is the tenth-century Codex Legionensis (VL 91 and also 133). In the Old Testament, one or more columns of the page are interrupted or displaced in (p.208) order to accommodate colourful representations of the biblical passage. These are less common in the New Testament, apart from Revelation, which seems to be a special case (e.g. VL 62). Illustrations were an integral part of the Commentary on the Apocalypse by the eighth-century Beatus of Liébana.77 The ever-increasing use of images in mediaeval picture Bibles, such as the Bible moralisée or the Biblia pauperum, some of which had Latin rather than or as well as vernacular text, is beyond the scope of the present volume.78 Even so, the full-page image of the devil in Codex Gigas (VL 51) led to its being known as the ‘Devil’s Bible’.

Bindings are also part of the decoration of manuscripts, although in most cases the original binding has perished or been replaced. One of the earliest to survive is the late seventh-century tooled leather cover of the Cuthbert Gospel (VgOe S). Jerome’s comment about ‘jewelled gospel books’ (see page 45) shows that the practice of creating luxurious bindings was already prevalent at the end of the fourth century. The magnificent casings of gospel books at the court of Charlemagne, and later the Ottonian dynasty, stand in the same tradition. The incorporation of carved ivory panels, a feature of luxury late antique bindings, was reintroduced in Carolingian times; decorative metalwork and precious stones are also found in particularly lavish productions. In Ireland, biblical manuscripts were often treated as relics and kept in a case, or shrine (e.g. VL 35).79


(1) Bischoff 1990; Boyle 1984; McGurk 1961a and 1994a. Reynolds & Wilson 2013 describes the transmission of Classical literature. For the Greek New Testament, see Parker 2008.

(2) For Latin Christian texts on papyrus, including an eighth-century codex of Augustine, see Bischoff 1990:8.

(3) Bischoff 1990:9.

(4) Bischoff 1990:9–10, 20–2; Gameson 1994:31; Parker 2008:60–1.

(5) For more on ink and writing materials, see Bischoff 1990:16–19.

(6) See page 200 below.

(7) See pages 80–1 and 103–4 above.

(8) Nevertheless, Charlemagne’s Coronation Gospels (see pages 81–2) have the words of Christ in uncial script and the rest of the text in minuscule (Berger 1893:275).

(9) McGurk 1994a:8–11 provides information on page size and shape, and the number of columns.

(10) See Lowe 1925.

(11) See further McGurk 1956; Farr 2011.

(12) See Bischoff 1990:57–61.

(13) Text in the Stuttgart Vulgate (Weber, Gryson, et al. 2007:732); examples in Bischoff 1990:66–72.

(14) Hammond Bammel 1984.

(15) See Bischoff 1990:83–127.

(16) See Bischoff 1990:127–36.

(17) Hammond Bammel 1979.

(18) Bischoff 1990:87 and 154; see also Horsley & Waterhouse 1984.

(19) Hälvä Nyberg 1988; Felle 2006.

(20) Traube 1907:241; Bischoff 1990:151.

(21) Handbooks such as Capelli 2011 give a full account, as do an increasing number of websites.

(22) For more on these abbreviations see Bischoff 1990:80–1, 86, and 119, as well as Lindsay 1910 and 1915 and Bains 1936.

(23) Bischoff 1990:178.

(24) These include ‬, known as the ‘Tironian et’; this should not be confused with the positura, also shaped like a ‘7’, for marking the end of paragraphs (Parkes 1992:22, 24, 43).

(25) Bischoff 1990:122.

(26) On the challenges of reading scriptio continua see Parkes 1992:67.

(27) See further Parkes 1992:15–16.

(28) See McGurk 1994a:12; Irish scribes call this practice ‘turn in the path’ or ‘head under wing’ (Brown 1980:77).

(29) Details in this paragraph are taken from Parkes 1992:13–40.

(30) See McGurk 1961b and Parkes 1992:22.

(31) See pages 58–9.

(32) This information, deriving from data in CLA, comes from McGurk 1994a:4. Fischer 1963b [1985:98–9] includes a list of biblical manuscripts in CLA: there are 108 from Italy, 67 from Germany, 47 from England, 47 from France, and 35 from Ireland. See also Parker 2008:63 and 75.

(33) See further page 203 below.

(34) See page 102.

(35) For lists of different sequences of biblical books, see Berger 1893:331–9 and the works of Bogaert and Frede mentioned below.

(36) See further Bogaert 2012:80 or Bogaert 2014a, which includes a list of thirty-seven such manuscripts from the ninth century onwards.

(37) Bogaert 1999 and 2013:513–14 gives details of other sequences of gospel books. Several different sequences of the evangelists are found in the opening sections of Fortunatianus of Aquileia’s commentary, but the chapter titles suggest the sequence Matthew–Luke–John–Mark.

(38) These are summarized in Frede 1964:161–2; see further Frede 1966–71:290–303.

(39) See Thiele 1956–69:4.

(40) See pages 176 and 181.

(41) For other Insular practices, see Bischoff 1990:44.

(42) e.g. VL 75; see Amphoux 2013.

(43) Some colophons are copied unaltered from exemplars: see pages 48 and 74 above.

(44) See Chapter 6 section j (page 138).

(45) A critical edition appears in the Stuttgart Vulgate (Weber, Gryson, et al. 2007:1515–16).

(46) See page 53.

(47) On both groups, see Regul 1969; Norelli 1990 reaffirms a Marcionite origin for the Monarchian prologues, although this is surprising given Marcion’s advocacy of a single Gospel.

(48) See McGurk 1994b, 1996; Szerwiniack 1994 and page 106 above.

(49) Jongkind 2015; I am grateful to the author for a pre-publication copy of this study, which includes a text and translation. The classic survey (which argues for a non-Marcionite origin) is Dahl 1978.

(50) See Light 1994:167.

(51) Technically, capitula refers to the divisions of the text and tituli to the initial set of titles, but it is customary to refer to the titles as capitula as well (cf. the siglum KA in the Vetus Latina Repertorium). For more on the divisions and history, see Bogaert 1982 and Houghton 2011.

(52) Fischer 1965 [1985:178]. The claim that capitula beginning with Ubi are of Spanish origin (Enciso 1954 and Vives 1957) is not borne out by further investigation.

(53) Meyvaert 1995; see page 70.

(54) The system does not provide for agreements between (a) Mark, Luke, and John or (b) John and Mark.

(55) See Thiele 1981; O’Loughlin 2010.

(56) See further McGurk 1993. Other examples of groups of four in canon tables are Erlangen, Universitätsbibliothek, 10 (page 295) and Munich, BSB, Clm 6212; Vg A switches from five to four.

(57) Nordenfalk 1938.

(58) For examples of this in other witnesses, see McGurk 1955:194.

(59) See the Stuttgart Vulgate for Matthew 123, 226, 231, and 343; Mark 4 and 185; Luke 2, 3, and 106 (and also the canon number for John 180).

(60) On the use of the marginal K to denote chapters in classical texts, see Parkes 1992:12 and 27; Parkes also notes the possible use of K to indicate corrections in manuscripts copied per cola et commata (1992:125 note 73).

(61) The text of PRIS can is given in the Oxford Vulgate (II.20–32).

(62) See above, pages 55, 89–90, and 97–9.

(63) For the Ordo lectionum per circulum anni, a list of lections for the entire year in early Roman office books, see Klauser 1972 and Bogaert 2012:72.

(64) Compare hd at the end of the penultimate line of Image 2. Other ways of indicating corrections are explained in Bischoff 1990:172.

(65) See page 27 and Bischoff 1990:43.

(66) Bogaert 2003:167–70; see also pages 21 and 27 above.

(67) Bischoff 1990:197, who notes that the decoration in VL 13 also includes imitations of jewellery.

(68) See pages 53 and 86–7.

(69) e.g. the Rossano Gospels (GA 042), also from the sixth century, written in Greek on purple parchment with scenes from the life of Christ or the Syriac Rabbula Gospels from 586: see Lowden 2012.

(70) Morgan 2012:409; 418–20.

(71) IR hae 3.11.8, also found in FO-A; see also AU Ev 1.6.9 where the Irenaean sequence is rejected in favour of Matthew–Mark–Luke–John; discussion in Bogaert 2001.

(72) Portraits are, very occasionally, found at the end of the Gospel, as in the Stowe St John, Autun, BM, 3, and possibly also VL 35: see further McGurk 1956:258–9 and 1961a:16.

(73) McGurk 1956:261.

(74) See also page 159 above, and McGurk 1956:257–8.

(75) See Bischoff 1990:209.

(76) See further Williams 1999.

(77) Bischoff 1990:211.

(78) See van Liere 2014:248–53.

(79) McGurk 1994a:20.