but not in any other Greek manuscript.
Contents: Introduction * Chapters and Verses * kefalaia, titloi * The Divisions in Vaticanus * The Ammonian Sections and Eusebian Canon Tables * The Euthalian Apparatus * Andreas's Divisions * Stichoi and Stichometry * Table Summarizing the Various Divisions * Order and Arrangement of Books
Historically, the New Testament has been divided and organized in many ways. Some divisions, such as our modern chapters and verses, are merely cataloguing schemes, used to find passages quickly. Others, such as the Eusebian apparatus, served scholarly purposes. This document will briefly outline some of the methods used over the centuries and preserved in the manuscripts. In addition, it will describe some of the more common marginalia found in the manuscripts.
This is followed by a description of some of the order in which books occur in the New Testament.
We may first dispose with the modern scheme of divisions.
The modern division of the Bible into chapters is believed to have been the work of Stephen Langton, the famous Archbishop of Canterbury (1207-1228) during the reign of the English King John. This system of chapters is found in many Latin Bibles, but only a few of the most recent Greek manuscripts; it has no historical significance.
Our modern verses have even less importance; they were devised by Robert Estienne (Stephanus) for his edition of the Textus Receptus, and have survived in printed editions ever since. They do not, however, occur in the manuscripts.
The kefalaia or Major Headings, the ancient equivalent of our modern chapters, are the most widespread form of organization in the ancient gospel manuscripts. Their exact date is not known; they have been ascribed to such worthies as Tatian. Their absence from the Codices Vaticanus and Sinaiticus, however, argues against such an early date. We first find them in the Codices Alexandrinus and Ephraemi of the fifth century (in the gospels; for the other books see the sections on the Euthalian Apparatus and Andreas's Divisions). It will be noted that the kefalaia constitute a series of numbers which restart with each book, but not with the first word of the book. In Matthew, for instance, the first entry coicides with 2:1; in Mark, the first notation occurs at 1:23; and similarly throughout. The locations of the kefalaia are noted (with italic Arabic numerals) in the margins of the Nestle-Aland editions, and so are readily accessible today.
Corresponding to the major kefalaia are the titloi or Titles. These are simply short summaries of the actions which happen in each section. Tables of titloi are often found at the beginnings of the gospels, and the headings themselves may appear at the heads of pages or the margins of manuscripts. The titles usually take the form "peri (something)," e.g. "About the Wedding at Cana."
We noted above that Vaticanus does not use the
kefalaia. Instead
it has its own system of chapter numeration -- in places two of them.
The system in the gospels is rather less orderly than the kefalaia,
as the sections vary greatly in length (some as short as a sentence, others many
paragraphs long). These numbers were written in red, though the chapter divisions
in the other part of the New Testament are in ordinay ink. The divisions in the
gospels are also found in
but not in any other Greek manuscript.
In the Acts, Vaticanus has two systems of division, of different ages and
independent of each other. The first-written of these was also available to
the scribes of Sinaiticus, as it also has some of these numbers (up to Acts 15:40,
where the numbering in 
breaks off).
In Paul we also find two unique systems of numbering. The older system has interesting trait that the entire corpus was numbered consecutively. This also reveals the interesting fact that, although Hebrews follows 2 Thessalonians in Vaticanus, the numbering is derived from a manuscript in which Hebrews followed Galatians (this follows since Galatians ends with §58, while Hebrews starts with §59; Vaticanus breaks off in Hebrews in the middle of §64, and we find §70 as the first entry in Ephesians).
In the Catholic Epistles we yet again find two systems of numbers, with the interesting feature that 2 Peter is not numbered. Presumably it was not regarded as canonical when the system was devised.
The sections described above were simply that: Sections. Ways of finding things. They had no other purpose, and little real value.
Not so the Eusebian apparatus, which was an early (and amazingly good) cross-referencing scheme for the Gospels.
The system had its roots in the work of one Ammonius of Alexandria, who some time in the second century arranged a sort of partial gospel harmony, taking the text of Matthew as his base and paralleling it with sections of the other gospels. Each section was numbered, and the numbers are referred to as the Ammonian Sections. (Confusingly, the Ammonian Sections are sometimes referred to as kefalaia. This usage is to be avoided. Not only is it confusing, but the Ammonian Sections average much shorter than the kefalaia -- e.g. in Matthew there are 355 sections but only 68 kefalaia.)
Roughly a century later, Eusebius of Cæsarea (the famous church historian) hit on a scheme to dramatically improve the Ammonian apparatus, by allowing any section of any gospel to serve as the basis point while still letting the reader look up parallels. Starting from the Ammonian divisions (which he may have modified somewhat), he created a set of lookup tables (to use a modern computer term) for finding cross-references. To each Ammonian number, he affixed a canon table number, showing the table in which the reader was to look for the cross-references. The contents of the tables were as follows:
The Eusebian system is not perfect; apart from occasional imperfections in the parallels, it was much easier to look up passages from Matthew than the other gospels (since the sections had to be listed in the order they occurred in one gospel, and Matthew was the chosen one). They were, however, compact (much more compact than our modern system of parallels), and they worked. They worked well enough that they were found in most later gospel manuscripts, and are even found in the modern Nestle-Aland margin (though with the section numbers transcribed into Arabic numerals and the canon numbers, perversely I think, converted to Roman numerals in the modern style -- i.e. IV for IIII and IX for VIIII). An example of its use is shown below, based on the opening sections of Matthew.
| Modern Verse | Ancient Canon | Modern Canon | Equivalent |
| Matt. 1:1 (through 1:16) | a G | 1 III | Item 1, found in Table III: Matthew #1 = Luke #14 (Luke 3:23f.) = John #1 (John 1:1f.) |
| Matt. 1:17 | b I | 2 X | Item 2, found in Table X: Table X means no parallels |
| Matt. 1:18 | g E | 3 V | Item 3, found in Table V: Matthew #3 = Luke #2 (Luke 1:35f.) |
| Matt. 1:19 | d I | 4 X | Item 4, found in Table X: Table X means no parallels |
Most manuscripts with the canon numbers naturally also included the canon tables, as well as Eusebius's Letter to Carpianus which explained the system, but this was by no means universal.
There are some variations in the canon system (in some cases, such as the ending of Mark, caused by variations in the text); the Nestle-Aland apparatus shows the variations found in many earlier editions of the canon tables (though manuscripts are not cited).
Finally, we should point out that the Eusebian apparatus did not always list actual parallels as we would understand the term; some items were linked only by theme (as witness the first example above: The genealogy of Jesus in Matthew is quite properly linked with the genealogy in Luke -- but also to the hymn to the incarnate Word in John).
Historical Note: Some have suspected that the Ammonian Sections did not exist prior to Eusebius's work. In support they urge the fact that the first manuscript to contain either (Alexandrinus) has both. (The numbers are also found in Sinaiticus, but from a later hand. N S F have them from the first hand, but they were added later in Bezae). We should note, however, that a significant number of manuscripts exist with the sections but no canon numbers or tables. In some cases this may mean that the manuscript was never truly finished (the canon numbers were usually added after the manuscript was completed, as they were usually written in colour; Eusebius had preferred that they be written in red. Also, some manuscripts listed the actual parallels at the bottom of the page, but this was easier done after the manuscript was finished). However, it seems more likely that the canons and sections truly were separate entities.
The most important supplements to the Acts, Paul, and the Catholic Epistles are associated with the name of Euthalius (or Evagrius). Who Euthalius was we do not know, nor can we even fix his dates (suggestions range from the fourth to the seventh centuries, though the fourth century is the usually accepted date, and he is sometimes described as Bishop of Sulci). Euthalius prepared an edition of the Acts and Epistles in sense-lines (this survives in manuscripts such as Hp; see the section on Stichoi and Stichometry).
In addition to his text, which occurs only in a few manuscripts, Euthalius compiled various helps for the reader; these are much more commonly found. Working, seemingly, from an earlier edition (Mill conjectured that it was that of Theodore of Mopsuestia, whose work was officially unacceptable due to his alleged unorthodoxy), Euthalius produces a system of sections and titles for Paul (similar to the kefalaia system in the gospels), and later extended it to the Acts and Catholic Epistles (these perhaps based on the work of Pamphilius).
Euthalius also organized the Old Testament quotation in the various Pauline Epistles, numbering and cataloguing them.
Finally, Euthalius is credited with the prologues and/or subscriptions to the various Epistles found in many manuscripts. This is, however, less certain -- and, as Scrivener remarks, the prologues "do no credit to the care or skill of their author," for they are patently inaccurate.
In the Apocalypse, the leading system of divisions is that of Andreas of Cæsarea, who lived in the sixth century and wrote the commentary that is found in so many of the Apocalypse manuscripts. Andreas's divisions are highly artificial (and not very well preserved, as the variations in the Nestle margin will show). Andreas arbitrarily divided the book into 24 sections (logoi); this seems to have been inspired by the 24 elders of Rev. 4:4. Each section was subdivided into three kefalaia (these inspired, apparently, by body, soul, and spirit). Thus there are 72 divisions in all in the Apocalypse, which the Nestle text numbers continuously though they are properly divided into groups of three.
Since these divisions were not invented until the sixth century, it
will be evident that none of our oldest manuscripts (P47,
, A, C)
contain them. Andreas summarized his sections, but since the number of
divisions was arbitrarily set, it will be observed that these sections
do not really accord with the logic of the book's arrangement.
Greek sticos means literally "line" (with many of the same extensions the English word has, e.g. a rank of soldiers or a line of a poem). In literary circles, however, it had a more specific meaning: The standard Homeric line of fifteen to sixteen syllables (about 35-50 letters). (This line is also sometimes called an epos, but this usage was in disuetitude by New Testament times.) This "standard line" came to have important implications. Seemingly by the fourth century, the notion of stichometry, or measurement by lines, was in existence (although it is officially credited to Callimachus c. 260 B.C.E.).
Stichometry had several uses for scribes and their patrons. It was the ancient equivalent of a "word count," used to determine what a scribe should be paid for a particular work. It could also be used to determine if a manuscript had been copied fully and correctly. And it could even be used as an approximate way to find quotations in a text. Thus it became standard practice to determine the number of stichoi in works that were regularly copied.
Stichometry seems to have been applied to the New Testament fairly early; Eusebius quoted Origen as commenting on the stichometry of various books. By the fourth century, we find Euthalius/Evagrius preparing an edition of the Acts and Epistles based on stichographic principles (although sense, rather than syllable count, had some part in the Euthalian edition; not all the lines are exactly one Homeric stichos long. Thus these books are properly arranged in cola et commata, rather than stichometrically). A stichometric edition of the Gospels is also known, though its compiler is not.
Relatively few New Testament manuscripts were copied in stichoi; sense-lines used too much expensive writing material. Still, there are books arranged in sense-lines (e.g. Dea, Dp, Hp. In addition, Fp and Gp seem -- based on the size and arrangement of letters -- to derive from an original in stichoi, though the lineation has not been preserved directly; the same is true of D). But the rarity of these manuscripts means that the stichometry of the New Testament was not well-known; although manuscripts beginning with P46 include stichometric information (usually in colophons), the figures quoted often vary significantly. The most common stichometry of the Gospels, according to Kirsopp Lake (K. Lake, The Text of the New Testament sixth edition revised by Silva New, p. 61), "gives 2600 [lines] for Mt., 1600 for Mc., 2800 for Lc., and 2300 for Jo.; but these are probably corruptions of 2560, 1616, 2750, and 2024 respectively, which are found in several MSS., and imply the presence of xvi.9-20 in Mark, and the omission of vii.53-viii.12 in John" (Lake does not, however, offer an explanation for this supposed "corruption." Also, Scrivener gives 2740 rather than 2750 as the number of lines in Luke). The table at the end of this document summarizes various stichometries, including the "common" one, the partial one in P46, and the early but rather defective one found in Codex Claromontanus (Dp; note the absence of four of the Pauline Epistles, although the omission of Philippians and the Thessalonian letters, at least, are likely accidental). In addition to the canonical works, the Claromontanus canon lists four extra-canonical works, Barnabas (850 lines), Hermas (4000 lines), Acts of Paul (3560 lines), and the Revelation of Peter (270 lines). The Revelation to John is listed among these semi-canonical works, as is, amazingly, the Acts of the Apostles.
The following table (adapted with some additions from Scrivener, A Plain Introduction to the Criticism of the New Testament, fourth edition, p. 68) summarizes the number and extent of the various divisions of the New Testament.
| Book | Vaticanus | titloi | Euthalian Divisions | Ammonian Sections | Sticoi | Rhmata (from f13) | |||
| Older | Newer | Common | P46 | Claromontanus | |||||
| Matthew | 170 | -- | 68 | 355 | 2560 | 2600 | 2522 | ||
| Mark | 62 | -- | 48 | 236 | 1616 | 1600 | 1675 | ||
| Luke | 152 | -- | 83 | 342 | 2740 | 2900 | 3803 | ||
| John | 80 | -- | 18 | 232 | 2024 | 2000 | 1938 | ||
| Acts | 36 | 69 | 40 | 2524 | 2600 | ||||
| James | 9 | 5 | 6 | 242 | 220 | ||||
| 1 Peter | 8 | 3 | 8 | 236 | 200 | ||||
| 2 Peter | -- | 2 | 4 | 154 | 140 | ||||
| 1 John | 14 | 3 | 7 | 274 | 220 | ||||
| 2 John | 1 | 2 | 2 | 30 | 20 | ||||
| 3 John | 2 | -- | 3 | 32 | 20 | ||||
| Jude | 2 | -- | 4 | 68 | 60 | ||||
| Romans | 1-58 | 8 | 19 | 920 | 1000 | 1040 | |||
| 1 Cor. | 19 | 9 | 870 | 1060 | |||||
| 2 Cor. | 11 | 590 | 70 (=570?) | ||||||
| Gal. | 3 | 12 | 293 | 375 | 350 | ||||
| Eph. | 70-93 | 3 | 10 | 312 | 316 | 375 | |||
| Philip. | 2 | 7 | 208 | 225 (222?) | - | ||||
| Colos. | 3 | 10 | 208 | 251 | |||||
| 1 Th. | 2 | 7 | 193 | - | |||||
| 2 Th. | 2 | 6 | 106 | - | |||||
| 1 Ti. | (lost) | (lost) | 18 | 230 | 209(?) | ||||
| 2 Ti | 9 | 192 | 289(?) | ||||||
| Titus | 6 | 97/98 | 140 | ||||||
| Philem. | 2 | 38 | 50 | ||||||
| Hebrews | 59-64 (69) | 5+ | 22 | 703 | 700 | - | |||
| Apocalypse | (lost) | (lost) | 1800 | 1200 | |||||
In discussing the order of New Testament books, we should keep several points in mind. The first is that the books of the New Testament were canonized over a period of time, and the second is that the vast majority of surviving manuscripts contain only parts of the New Testament.
Taking the last point first, it's worth remembering that, until the era
of the minuscules, there is not one Bible which demonstrably contains exactly
and precisely our modern New Testament, even if one allows for damage to the
manuscripts. Of the five major uncials
( A B C
Y),
and A
contain all the books of the New Testament, but have extra books as well;
Y omits the Apocalypse;
B is defective for the latter part of Paul and may never have contained
the Apocalypse. C, based on the surviving leaves, contains only the books
we now think of as the New Testament -- but this cannot be proved; too many
leaves are missing. We cannot be sure that it did not contain other books as
well. C probably contained our present New Testament, but we dare not
be dogmatic.
Most Biblical manuscripts consist of only a subset of the New Testament. Normally one finds the books grouped into subsets: Gospels, Acts and Catholic Epistles (these two are very rarely separated, though there are a few exceptions), Paul, Apocalypse. This explains the common abbreviation "eapr" (or "eapcr") for the contents of the New Testament: e=gospels, a=Acts (plus Catholics), p=Paul, r=Apocalypse.
Almost every combination of these units is found. The majority of manuscripts are Gospels alone -- there are thousands of such manuscripts. The most next common is Acts (including Catholics) plus Paul; there are hundreds of books of this form. The Apocalypse very often stands alone (not infrequently with non-canonical works), though it might be attached almost anywhere. But we also find the following (based on the data in the first edition of the Kurzgefasste Liste; the list is neither complete nor guaranteed):
The order of these divisions is fairly standardized. The gospels are almost always the first thing in a codex (and at least some of the exceptions are the result of rebinding). Acts and Catholic Epistles generally precede Paul, though this is not universal. The Apocalypse is generally last.
For the order of books within the four sections, there is rather more variety. The most notable case of a "movable book" Hebrews, found at various places within the Pauline corpus. Usually it follows either 2 Thessalonians or Philemon, but it has occurred in many other places (as it followed Galatians in the ancestor of Vaticanus). The order of some of the other shorter books also varies, e.g. Philippians may swap with Colossians. The first four books (Romans, 1 Corinthians, 2 Corinthians, Galatians) almost always occur in that order. Other variations might possibly be scribal -- e.g. a scribe finished Ephesians, quit for the day, and accidentally copied Colossians next rather than Philippians, then went back and copied the other. There is no proof of this happening, but it is much more likely in Paul than any other section.
The Gospels almost always occur in the order Matthew-Mark-Luke-John. But there are exceptions, and most of them are early. The most common variation on this order is the so-called "Western" order, found in D, W, and probably P45: Matthew, John, Luke, Mark.
The Catholic Epistles probably show the most variation, especially in early manuscripts, since some of the books (James, 2 Peter, 2 John, 3 John, Jude) were of questionable canonicity. The Peshitta, for instance, includes only James, 1 Peter, and 1 John. It will be evident that the order of the books will be dependent upon which books are included.