In the sixteenth chapter of the Gospel of Mark, between verses 8 and 9, the English Standard Version has a prominent heading-note: “(SOME OF THE EARLIEST MANUSCRIPTS DO NOT INCLUDE 16:9-20).” Some other versions contain a similar note at this point, and most of them are similarly vague. The ESV also has a long footnote which begins, “Some manuscripts end the book with 16:8; others include verses 9–20 immediately after verse 8.” This is spectacularly inaccurate: the number of Greek manuscripts which include Mark 16:9-20 is over 1,600; the number of Greek manuscripts in which the text of Mark clearly ends at the end of verse 8, followed by nothing but the closing-title, is exactly two.
Among major translations, only the New King James Version uses precise language in its note about Mark 16:9-20: “Verses 9-20 are bracketed in NU-Text as not original. They are lacking in Codex Sinaiticus and Codex Vaticanus, although nearly all other manuscripts of Mark contain them.” (“NU” in this footnote stands for the Nestle-Aland/United Bible Societies compilation, which very heavily favors the Alexandrian Text.)
Even the NKJV’s footnote, however, tells only part of the story. The patristic writer Irenaeus, bishop of Lyons (or Lyon – ancient Lugdunum), writing around 184, specifically quoted Mark 16:19 in Against Heresies, Book Three, chapter 10. This shows that the passage was in the text of the Gospel of Mark in manuscripts used by Irenaeus over a century before either Vaticanus or Sinaiticus were made. Irenaeus’ contemporary Tatian utilized Mark 16:9-20 in his Diatessaron, too, and Justin Martyr almost certainly borrowed language from Mark 16:20 in chapter 45 of his First Apology, around 160. Granting that detailed information cannot be expected of footnotes, it is unacceptable to put fourth-century manuscripts in the spotlight while keeping second-century patristic evidence in the shadows.
But today, the early and widespread patristic evidence that supports Mark 16:9-20 is not my focus. Instead, I want to look at the unusual features that we find in Codex Sinaiticus with a focused lens, setting aside the warped lens through which two Greek manuscripts appear as “Some” and 1,600 manuscripts appear as “others.”
The first thing to notice about the pages of Codex Sinaiticus on which the Gospel of Mark concludes is that they are replacement-pages. They were not written by the same copyist who wrote the text on the surrounding pages. Instead, the text on these four pages was written by the diorthotes, or proof-reader, of the manuscript, while it was still in production.
|Columns 9, 10, 11, and 12 of the cancel-sheet
in Codex Sinaiticus.
In the replacement-pages, however, the text of Mark 16:8 does not end near the bottom of column nine. It ends in column ten – even though the diorthotes probably wrote 84 fewer letters of Mark than what the main copyist had written on the replaced pages. This raises a question: why did the diorthotes make the text of Mark 16 extend into column 10, instead of finishing it in column nine?
|A comparison of the rate of letters-per-column
shows a significant decrease in columns 5-9,
and a drastic increase in columns 11-16.
To answer that question, we must examine the columns of Luke 1:1-76 that the diorthotes wrote on the replacement-pages. A simple count of the number of letters in each column of text demonstrated the extensive variation in the rate of letters-per-column.
Before proceeding further, it may be worth pointing out the following: Mark 16:9-20 contains 971 letters (depending on textual variants). Even if the main copyist had accidentally skipped the same 106 letters that the diorthotes skipped in
15:47-16:1, the remaining 886 letters would not
fit into the remaining space after 16:8 (which would have a normal capacity of
662 letters) in columns nine and ten.
Thus, whatever motivated the diorthotes
to replace the four pages that the main copyist produced, it was not because those pages contained Mark
(If one wanted to write Mark 14:64-16:20 in columns 1-10, the resultant average rate of letters per column in columns 1-10 would jump to 667 letters per column. Such script-compression is technically possible: an average rate of letters per column of 673 is observed in columns 11-16 in the cancel-sheet. But the main copyist, unlike the diorthotes, had no reason to suddenly compress his script.)
Now let’s turn our attention to the columns in the replacement-pages that contain text from the Gospel of Luke. If the main copyist had accidentally repeated a large chunk of text, and the diorthotes made the replacement-pages in order to remove the repeated lines, this would require the corrector to fill the space with fewer letters than the original pages had contained. But what we see in columns 11-16 is a staggering increase in the rate of letters-per-line. Instead of 635 letters per column, we see here in Luke an average rate of 691 letters per column. If we work from the premise that the text of Luke began at the top of column 11 in the replaced pages, then the diorthotes made the replacement-pages in order to correct a large omission that the main copyist had committed.
With that premise in place, it looks as if a section of text consisting of over 330 letters was absent from the main copyist’s text of Luke 1:1-56. Probably the main copyist accidentally skipped from either the beginning of Luke 1:34 to the beginning of
1:38 (losing, in the process, 311 letters). (An alternative is that he skipped from the
beginning of Luke 1:5 to the beginning of Luke 1:8, thus losing 319 letters,
but this would almost require that he was not thinking at all about what he was
writing.) Without those 311 letters, the
text of Luke 1:1-56 that is on the replacement-pages consists of 3,835 letters
occupying six columns, which yields 639 letters per column – well within the
copyist’s natural range of variation.
What if, instead, the main copyist began the text of Luke 1 at the top of column 10? In that case, it would appear that the main copyist accidentally repeated a large portion of text in Luke 1. If we add to 4,146 letters an additional 311 letters, caused by the repetition of verses 34-37, we reach a total of 4,475 letters occupying seven columns of text. Divided into seven, this yields (again) 639 letters per column – well within the copyist’s natural range of variation.
|The replacement-pages consist of a single
sheet of parchment, folded in the middle.
So: while we can discern that the creation of the replacement-places was due to a problem in the text of Luke 1:1-56a, a definitive reconstruction of the format of the text on the replaced pages is not easy to make, because a reconstruction involving an omission in a text in which Luke 1 began at the top of column 11, and a reconstruction involving a repetition in a text in which Luke 1 began at the top of column 10, are both feasible. These competing possibilities, however, should not obscure the observation that in neither reconstruction is Mark 16:9-20 present on the replaced pages.
The remarkable range of variation in the diorthotes’ rate of letters per column in the cancel-sheet tells a little story about how the text on these pages was written – and this may imply something interesting about the ending of the Gospel of Mark. Instead of beginning at Mk.
the diorthotes realized that the main
challenge he faced would be to format the text in such a way that the final
line of the final column ended exactly where final line of the final column of
the replaced pages ended. With this in
mind, he began the replacement-pages by writing Luke 1:1 at the top of column
11. (This was a practical precautionary
step, inasmuch as in the event that his attempt was unsuccessful, he would have
thus saved himself the trouble of writing out the text of Mark 14:54-16:8 only
to have to start the whole thing over.)
Only after he had carefully succeeded in cramming Luke 1:1-56a into six columns did the diorthotes begin to write Mark 14:54b at the top of column 1. He wrote columns 1-3 without any unusual deviation from the usual rate of letters per column (635, 650, and 639). In column 4, he accidentally reverted to the use of the lettering-compression he had used when writing the text of Luke 1:1-56; this is why there are 707 letters in column 4. Then, realizing what he had done, he compensated by slightly stretching out his lettering in columns 5, 6, 7, and 8. But after accidentally skipping most of Mark 16:1, he still did not have enough text to reach column 10, even writing at a rate of 600 letters per column (30 letters less than usual).
The diorthotes could have simply written the rest of chapter 16, up to verse 8, in his normal lettering, and thus finished Mark in column 9, with a blank column between the end of Mark and the beginning of Luke. But he made a conscious decision not to do that. Instead, he stretched out his lettering even more, so as to write only 552 letters in column 9. Thus he had 37 letters remaining to place in column 10.
With all these things in the equation, let’s again approach the question: why didn’t the diorthotes finish Mark 16:8 in column 9, and thus leave a blank column before the beginning of Luke? Why did he stretch out his lettering (and write Jesus’ name in Mark 16:6 in its full, uncontracted form) so as to make his lettering reach the tenth column?
Conceivably, a strong sense of aesthetics motivated the diorthotes to avoid leaving blank columns between books in the same genre (genres such as Poetry, Minor Prophets, Gospels, Epistles). In Codex Sinaiticus, a book usually begins at the top of the column which immediately follows the previous book, unless a new genre is being introduced. Four columns (a single page) are blank after the Gospel of John. Six columns separate the end of Philemon from the beginning of Acts. A blank column separates the end of Acts and the beginning of James.
This pattern, however, was not kept with complete consistency. There is no blank column between Jude and Revelation, and there is no blank column between Revelation and the Epistle of Barnabas, although there is a blank column between the end of the Epistle of Barnabas and the beginning of the Shepherd of Hermas. And in the Old Testament portion of Codex Sinaiticus, after the end of the book of Judith, the first column of the next page is blank, followed by the beginning of First Maccabees at the top of column 2.
That last detail is particularly instructive, because the diorthotes served as copyist for the book of Tobit and Judith. Apparently, another copyist (Scribe A) had finished a section which concluded at the end of Esther in the second column of a page. The same copyist had also made a section containing First Maccabees, beginning in the second column of a page. The diorthotes faced the task of writing the contents of Tobit and Judith in another section to be placed between the two already-written sections, beginning where the other copyist had left off. After completing Tobit and most of Judith, he realized (as Dirk Jongkind has noted) that he didn’t have enough text to reach the column next to the beginning of First Maccabees. For this reason, he resorted to stretching out his lettering (in much the same method that is seen in Mark 16:2-7) and wrote one or two fewer lines per column. His efforts, however, were still not sufficient, and that is why a blank column precedes First Maccabees. It is a “seam,” so to speak – merely a side-effect of a quirk that occurred in the production of the manuscript.
Besides a desire to insert blank space only between books of different genres, something else seems to have motivated the diorthotes to take drastic action to avoid leaving a blank column between Mark and Luke: a determination to avoid leaving a feature which could be considered memorial-space for the absent twelve verses. Occasionally, copyists encountered readings in their exemplars which lacked material that the copyists recollected from another manuscript they had encountered; they copied the text of the exemplar but left blank space – memorial-space – to express the thought that something was missing.
Perhaps the best-known example of this occurs in Codex Vaticanus. In this manuscript, the copyists never left blank space between books, except between the Old and New Testament, and at two production-seams (once where the format shifts from three-columns-per-page to two-columns-per-page, leaving a large gap, and once (at the end of Tobit) where the work of one copyist ends and another copyist’s work begins), and in one other place: the end of Mark, where, after the copyist wrote 16:8, the closing-title appears below it, and the next column is blank. In addition, if a copyist were to erase the closing-title, and insert Mark 16:9-20 after 16:8 using the same script-compression technique seen in Sinaiticus in Luke 1:1-56, all twelve verses would neatly fit into the remaining space.
|A digital image of the last page of Mark in Codex
Vaticanus, with memorial-space for verses 9-20,
is at the Vatican Library's website.
A second example occurs in Codex Regius (L, 019) after John 7:52, where the copyist left some blank space, including an entire blank column, before writing
8:12. The copyist thus
expressed his awareness of the existence of John 7:53-8:11.
This blank space in Codex L is not nearly sufficient for the absent 12
verses (or even for 8:3-11), but it nevertheless demonstrates the copyist’s
knowledge of the existence of the passage.
(The apparatus in the 28th edition of the Nestle-Aland Novum Testamentum Graece lists “L*vid”
as a witness for the inclusion of John 7:53-8:11, and “Lc” as a witness for
non-inclusion, even though it is obvious that the manuscript never contained
The diorthotes of Codex Sinaiticus may have realized that if he did not stretch his lettering so as to be able to put some text in column 10, he would run the risk that the resultant blank column would be interpreted as memorial-space.” The possibility cannot be ruled out that his exemplar concluded Mark’s text in column 9 and left column 10 as memorial-space in a format similar to that of Codex Vaticanus. The diorthotes of Sinaiticus apparently staunchly rejected Mark 16:9-20: not only did he stretch the text of Mark 16:2-8, which prevented any future readers from interpreting a blank column as memorial-space, but he practically turned his arabesque-design following Mark 16:8 – a design which was usually much less ornate – into a fence, emphatically spread across the column, before the closing-title. (I recommend using the zoom-feature at the Codex Sinaiticus website to see this feature in detail.)
The diorthotes’ embellished arabesque was noticed in the 1800’s by several researchers, including John Gwynn and George Salmon. The arabesque-designs that the diorthotes drew at the end of Tobit, at the end of Judith, and at the end of First Thessalonians (where he had made another cancel-sheet) are much simpler than the one that follows Mark 16:8. Salmon drew the conclusion that “The scribe who recopied the leaf betrays that he had his mind full of the thought that the Gospel must be made to end with efobounto gar, and took pains that no one should add more.”
|George Salmon (1819-1904) -
New Testament scholar,
mathematician, and chess player.
In 1883, John Gwynn wrote, “As regards the omission of the verses of S. Mk. xvi. 9-20, it is not correct to assert that Cod. À betrays no sign of consciousness of their existence. For the last line of ver. 8, containing only the letters TOGAR, has the rest of the space (more than half the width of the column) filled up with a minute and elaborate “arabesque” executed with the pen in ink and vermilion, nothing like which occurs anywhere else in the whole MS. (O.T. or N.T.).”
No other reason comes to mind to explain why the diorthotes enlarged and embellished his arabesque-design here, and only here. Salmon’s deduction appears to be correct: the mind of the diorthotes was full of the thought that Mark should end at the end of 16:8. This implies that the diorthotes was aware of at least one other way in which the Gospel of Mark concluded.
Which ending did the diorthotes reject: verses 9-20, or the Shorter Ending, or both? When we consider that Codex Sinaiticus was almost certainly produced at
Caesarea, the answer is clear: the diorthotes was aware of, and rejected,
Now let’s pause and consider something that was written by Eusebius, who was bishop of
Caesarea in the early 300’s.
In his composition Ad Marinum,
Eusebius addressed a question: “How do
you harmonize Matthew’s statement that Jesus’ resurrection was “late on the
Sabbath” with Mark’s statement that it occurred “early in the morning on the
first day of the week”?
In the course of his answer, Eusebius mentioned that there were two ways to resolve the perceived discrepancy: one person might say that the passage in Mark (beginning at 16:9) is not in every manuscript, or is not in the accurate manuscripts, or is hardly found in any of them, or is present in some copies but not in all of them, and is therefore superfluous, especially considering that it might seem to contradict the other accounts. But – Eusebius continued – someone else, reluctant to dismiss anything he finds written in the Gospels, may accept both accounts instead of picking and choosing between them. Granting this premise, the way to resolve the perceived difficulty is to simply read the phrase in Mark with a comma: as “Having risen, early in the morning on the first day of the week He appeared to Mary Magdalene.” This is in agreement with what John says. The meaning is not that Christ’s resurrection was “early in the morning,” but that this is the time when He appeared to Mary, afterwards.
Twice more in Ad Marinum, Eusebius utilizes Mark 16:9. At one point Eusebius mentions a theory that there were two women named Mary Magdalene, and points out that one of them would be “the one of whom it is stated in Mark, in some copies, that he had cast seven demons out of her.” In the course of answering another question, Eusebius again mentions the theory that there were two Mary Magdalenes, and mentions that the Mary Magdalene mentioned by John would be the same person from whom, according to Mark, he had cast out seven demons. In this third utilization of Mark 16:9, Eusebius did not bother to mention anything about manuscripts.
While what Eusebius says in Ad Marinum throws a hot informative light upon the various wax commentaries which misrepresent Eusebius’ statements about the ending of Mark, his comments are instructive for the question at hand because of what he does not say. Eusebius displays no awareness whatsoever of the existence of the Shorter Ending.
If we take the evidence that Codex Sinaiticus was produced at
Caesarea c. 350 (perhaps under the supervision of Acacius)
alongside the evidence that the Shorter Ending was not known at Caesarea in the
early 300’s, then we may conclude that the ending of Mark 16 rejected by the diorthotes of Codex Sinaiticus was verses
This would be consistent with what may be deduced from a comparison of the treatment given to Mark 16:9-20 in Ad Marinum and in the Eusebian Canons. When Eusebius wrote Ad Marinum, he was comfortable with the inclusion of the passage and went through two verbose paragraphs to explain to Marinus how Mark 16:9 should be read and how the opening sentence should be pronounced. By the time he developed the Eusebian Canons and Sections, though, Eusebius had decided not to include the passage. It would not be surprising if Acacius, bishop of
Caesarea from 339 to 365, inherited the latter view and enforced
it when he oversaw the production of new parchment manuscripts, including Codex
Sinaiticus, based on old papyrus copies which were wearing out. (Jerome, in Lives of Illustrious Men – see chapters 98 and 113 – and in Epistle 141, Ad Marcellam), mentions that
Acacius and his successor Euzious
engaged in this enterprise. Jerome did
not specify that Acacius and Euzois preserved the texts of exemplars of books
of the Bible, but it seems highly probable, and would explain the use of a
Western copy as a secondary exemplar in John 1:1-7:38 of Codex Sinaiticus – and
not just any Western copy, but one with some affinities to the text used by the Gnostic heretic Heracleon, which had been cited by Origen in his response
The thing to see here, regarding the ending of Mark, is that when we take a close look at the two Greek manuscripts in which the text of Mark clearly stops at 16:8, with nothing afterwards except the closing-title, one of them (Vaticanus) expresses the copyist’s awareness of the absent 12 verses by the addition of memorial-space, and in the other one (Sinaiticus), the last chapter of Mark is written on replacement-pages by a copyist who probably indicates his own awareness of, and rejection of, the absent 12 verses via his script-expansion (avoiding a blank column) and arabesque-enhancement.
These extra details should be kept in mind when reading Bible-footnotes about Mark 16:9-20 which frame the manuscript-evidence in vague terms without mentioning the patristic evidence.