language menu item language menu item language menu item language menu item

Amy Myshrall and Timothy Brown doing transcription work at St Catherine's Monastery, Sinai.
Amy Myshrall and Timothy Brown doing transcription work at St Catherine's Monastery, Sinai.

Download the full image.

The transcription

Navigating the transcription

There are two ways of going to a particular place in the codex. The first is by means of ‘choose a passage’, by book, chapter and verse, which will take you to that place in the image and/or transcription. The second is by ‘choose a page’ which relates to the structure of the book. Codex Sinaiticus, like every properly constructed codex is divided into quires, consisting usually of four sheets folded in half. Each of these quires has a number, provided by the original production team. Each of the resultant eight leaves of each quire is called a folio, and has a number between one and eight. The right hand page is called the recto and is indicated with the abbreviation r, the reverse of it (the next left hand page) is the verso, indicated with v. Use the drop down menus to select the quire and folio that you want, and the white arrow to go to it. The next set of arrows allow you to go to the first, previous, next or final page.

Each page also has another identifier, its folio number in the library where it is held. This is given with the location in the pop-up box at the bottom of the image window.

The text

The transcription presents the text of the manuscript as it was created by the production team, including the layout, and indicates all subsequent corrections.

The transcription may be viewed in two different ways: ‘view by verse’ gives the text of each book, verse by verse; ‘view by page’ gives the text with its layout in the manuscript, with the page, column and line breaks and with most marginal material in place.

The ‘view by page’ transcription is chosen as the default, because the presentation is based upon the concept of being able to compare the image with the transcription.

Both transcription views present corrected text in the same way. Where a correction has been made, its presence is shown by blue text. If the correction adds previously absent text, a blue T indicates where it is provided. Running the cursor over the blue text or blue T opens a box in which the different stages of the text are indicated.

It may be helpful to know that there are four types of correction:

  • addition of text
  • deletion of text
  • substitution of one wording for another (changes in spelling technically belong here, although they may also be taken as a separate class of generally less significant readings)
  • changing the order of two or more words

More than one of these types can occur at any one place, and sometimes two or more correctors have made one or more types of change. The great majority of corrections in Codex Sinaiticus are changes to the spelling.

The correctors are given the following indicators.

Production of the manuscript

This consists of copying by one of the scribes, and revision, by the scribe or by someone else.

* the text first written by the scribe (information about the scribe of any particular page is provided in the pop-up box at the bottom of the window)
S1 a correction made in the production process, as part of the revision of the text after it had been copied, or a correction by the scribe in the copying process. These cannot always be distinguished
A Scribe A
B Scribe B. As a result of this transcription evidence has emerged that this scribe’s pages may be the work of two scribes:
B1 responsible for all the other work attributed to B
B2 copied the Minor Prophets and Hermas
D Scribe D

It is sometimes possible confidently to attribute an S1 correction to one of the scribes, and thus A, B and D appear as correctors. There are good arguments in favour of presenting the corrected text as the text viewed on screen, with the first hand reading placed in the pop-up box. However, there are some serious presentational difficulties with this in many places, particularly when this ‘final text’ is written between columns or in the top or bottom margin. Our approach has been to prefer * as the main text, and then S1 as part of the correction process, apart from some very exceptional places where the * reading is erased to such an extent that it is no longer legible. In these places the main text presents the text of S1. These are for the most part places where the scribe altered the main text as he went along.


This consists principally of the ‘c’ group, correctors who revised the manuscript rather extensively between the fifth and seventh centuries. They are grouped into four sub-series, namely

  • ca
  • cb, which is further divided between
    • cb1
    • cb2
    • cb3
    • cb is used when the correction cannot be more accurately assigned
  • cc
  • cc*, works only in Revelation
  • cpamph
  • c indicates a change which can be attributed to the ‘c group of correctors, but not to one of the separate hands within it
  • d, a hand who rewrote faded portions of text, occasionally providing corrections (many examples in Isaiah)
  • e, a hand which made a few corrections in Proverbs, Matthew, 1 Timothy and Acts
  • corr indicates a change which cannot be attributed to a particular corrector (in principal it might include corrections made in the production of the manuscript as well as later changes)
  • corr indicates a change which cannot be attributed to a particular corrector (in principal it might include corrections made in the production of the manuscript as well as later changes)

Where more than one corrector has worked on a correction, the correctors are listed within the app tags in order in which they are listed above.

Unreadable text

Unreadable text is indicated by square brackets with a space between them, as [ ]. This indicates that there is a space where one or more letters is illegible through washing out or erasure.

Corrections across a page break

There were technical problems in indicating a correction across a line break. Thus has been resolved by showing the first hand text in blue on both pages, but placing the text of the correction on the second page. See for example Quire 80, Folio 6r-6v, where both the original and the corrected text span both pages.

Corrections across a modern chapter or verse break

There are also technical reasons why a single correction in the manuscript spanning a chapter or verse break has to be presented as a separate correction for each verse concerned. See for example Quire 36, Folio 3r, Column 3, Lines 3-4 (2 Esdras 20.12-13).


One of the most striking aspects of the manuscript is the more extensive corrections and additions, such as the colophons to Esther and 2 Esdras, and many lengthy additions to the text in the top and bottom margins. We have therefore provided these in the main text view, in slightly smaller font and in blue. Marginal additions also appear in the pop-up boxs at the appropriate place in the text. It is not possible for technical reasons to do this with interlinear and intercolumnar additions, or those in the side margins. These appear only in the appropriate pop-up box.

Further annotation

Three medieval writers wrote notes in the margin. They were classified as e and f by earlier editors. We have retained the names of two of them (Dionysios and Hilarion). The third group is more complex. It consists of two notes and four corrections. The two notes (Quire 43 Folio 1r top and Quire 68 Folios 1v-2r bottom) are both by someone called Theophylact, but are unlikely to be by the same hand. Of the corrections, those in Quire 65 Folio 3v Column 1 Line 36 and Quire 75 Folio 3r Column 4 Line 33 are probably by the same hand as the Theophylact of Quire 69 Folios 1v-2r. The two other may be by two different hands: The insertion of ?e and addition of an accent on the following o at Quire 86 Folio 2v Column 3 Line 2 looks somewhat different, and the change on Quire 86 Folio 8v Column 1 Line 33, which is partially a rough retracing of letters in the manuscript, looks unlike the neat corrections at Quire 65 Folio 3v and Quire 75 Folio 3r. We have retained the designation ‘e’ for all these corections.

Arabic glosses

The Arabic glosses are presented in notes (indicated by the § symbol). The glosses were written without pointing, but this has been included in the transcription, along with a translation. Two Arabic hands have been identified, and are named A and B.

Presentation of symbols

The difference between a manuscript, where the scribe can produce features with endless variation, and a web presentation, which is limited both by what is available in any single font and the resources involved in programming, has led to various pragmatic decisions. There are arguably a considerable variety of paragraphi. For example, there may be two different types in Jeremiah, which may or may not have an undetermined difference in function. After some hesitation, we showed this by presenting the horizontal line paragraphi in Jeremiah (they start on Q46-8v) differently. It is possible that other distinctions could be made which we have not included.

Punctuation is also less flexible. The position of the high, mid or low point can vary on the manuscript page, so that frequently the transcriber has to make a judgement as to whether a point is, for example more high or mid. Some arrangements of dots defy typography, and we have adopted a symbolic presentation.

Diaeresis (a double dot over iota or upsilon, usually when they are in the initial position) is sometimes written without lifting the pen, almost as a straight line. It is always presented as e.g. ?¨.

Overlines and underlines are another difficult issue. The transcription ignores all lines which we decided are decorative. Superlines over numbers (e.g. in chapter numbers and the Eusebian apparatus) are retained as necessary to indicate what they are. Superlines over nomina sacra (words such as ?¯s¯?¯ for ?s?a??) are provided, and generally the extent of the superline is approximately recorded. Superlines in place of final nu are also recorded.

Rubrication is retained in the obvious way, with the text in red (see the Psalter). Where material written in red has been corrected, blue takes precedence (see the beginning of Psalm 151).

How the transcriptions were made

Different parts were composed in different ways.

The New Testament is based upon transcriptions made in the Institute for New Testament Textual Research, in Münster in Westfalia, Germany, for ‘NTTranscripts’ and the ‘Digital Nestle-Aland’ (see Certain features needed to be added, namely more precise information about individual corrections, and metatextual elements such as the Eusebian Apparatus in the Gospels, running titles, quire signatures and hanging lines, as well as bringing the transcriptions into conformity with those of the Old Testament. The resulting transcriptions were then reviewed by staff in the Münster Institute, and a final version agreed. The transcription of the Gospel of John has another layer of preparation. It is based on the one made for the International Greek New Testament Project, and available at This version had been revised by T.C. Skeat, and had also been revised in collaboration with the Münster Institute.

Behind all these transcriptions lay the principal of double independent transcription of each book. Each transcription is made by adapting an existing electronic version of a printed text as the base, one of them being also tagged to present the layout of the manuscript. The two transcriptions are then compared automatically using the Collate programme (see, and the differences resolved, the transcription containing the layout information being used for the final version. This methodology ensures a high level of textual accuracy.

Hermas and Barnabas and the majority of the Old Testament were produced by the team in the way described in the previous paragraph. There are two exceptions. One is that the extreme complexity of Isaiah led to our abandoning the principal of two independent transcriptions about half-way through the book, and resorting instead to a careful collation of the first transcription against the images. Second, the Sinai fragments were transcribed afresh rather than by adapting existing base text.

The source of the transcription was in the first place the new digital images. This has led to more secure and detailed information at many places, since the images generally show detail more clearly than the manuscript itself does. But it was still necessary to consult the manuscript in places where problems remained, and the transcribers made the following visits.

  • St Petersburg (May, 2008), where verification of the extremely difficult fragment of Judith was essential.
  • Leipzig (April and June, 2008), for verification of several readings, including the colophons of Esther and 2 Esdras.
  • St Catherine’s (June, 2008), for transcription of the material, especially the fragments, from the New Finds. Initial transcriptions of some of these had been made from provisional images. The conservation and imaging according to the Project standards prior to this visit made a full transcription possible. For further details, see the section “The fragmentary pages”.

Technically, the transcriptions were made as plain text files, using tags compliant with the TEI. These were then converted into xml. Some final editing was carried out on these files. which were then delivered to the web site developers.

The alignment

When the ‘Standard Light’ images and transcription are selected, the two are aligned. Point the mouse over any word in the image, and it and its representation in the transcription will be highlighted. The same will happen if the mouse is pointed at a word in the transcription. This was made possible by software developed in ITSEE for this project.

Sometimes in the fragmentary leaves, very little of some words remains. Use this tool to see exactly where the remains of the letters are.

Note: The alignment system does not include any marginalia (quire signatures, glosses, additions and titles) in the alignment system. This includes colophons of books.