The preparation and development of the edition are described in the Introduction to the printed volume. What follows is a reduced version of the section describing the process by which the transcriptions were made.

We divide the majuscules into two categories:

Category 1. Those which are fragmentary and difficult to read. These have been transcribed with the highest possible level of representation of all that is visible. In practice, this will vary depending on the parts of a leaf that are extant (e.g. lack of margins to show original line breaks, presence of page numbers or running titles, uncertainty over use of nomina sacra in small fragments).
 
Category 2. Those which are either complete or extensive or, even if they are fragmentary, are easy to read. What is described as 'standard formatting' in the descriptions of individual transcriptions below applies to manuscripts in this category. It consists of the following:

1. Layout: division of text into folios (with numbers found in the MS given), columns and lines.
2. Punctuation: upper-case letters indicating enlarged initial letters.
3. Orthography: as in the manuscript transcribed, with superline for nu when the MS uses it, other abbreviations indicated with brackets (e.g. κ(αι) indicates use of the καί compendium), and nomina sacra indicated by a superline over the second letter of the contraction. Note that we use this superline as an indication of the nomen sacrum, and not as a reproduction of precisely what is in the manuscript. Thus it is present even when it is omitted by the copyist. This represents a difference from our practice in the papyrus volume. The change has been made in order to allow the program to recognise all readings in this category.
4. Original text and subsequent corrections are all shown, with the following sigla:


*	the text as first written by the original hand
C*	a correction made by the original hand
C	a correction by another hand. C1 or C2 is used where more than one correction is made

Other sigla are indicated in the descriptions of individual transcriptions. Where the first hand (or a correction) is illegible, we enter the number or probable number of letters. For example, see the reading of 02 at 4.19.
Note that in the printed transcriptions, corrections to manuscripts are given in the footnotes, the corrected text being in italics.
5. Lacunae: missing text is indicated by these signs:


lac	parchment is missing and the text is lost (this includes places where the parchment is present, but the text cannot be read); where the parchment is lost, and has been replaced, the text has been transcribed in a separate file
[____c]	restoration of missing text is given in brackets. Sometimes we give a number or range of numbers in square brackets
________c̣	dots under letters indicate that the restoration of the text is uncertain

6. Superscription and subscription are provided when present.

The transcriptions were made as follows

1. An electronic transcription was made of each witness. These were tagged according to TEI guidelines. We followed what we call a modular principle. We have certainly not provided transcriptions that are exhaustive with regard to layout, but they all include folio, column and line layout. Of course the original text and subsequent corrections are included. Some include indented or hanging lines, and other features as detailed in the paragraphs on the individual manuscripts below. One great benefit of electronic transcriptions is that they can be developed and improved by successive teams of scholars.

2. Once a transcription was complete, each individual transcription was automatically collated against a base text, using Collate.

3. The collation of two independent transcriptions of a witness led to the verification of the discrepancies, and the revision of one of them into a textually accurate file.

Once these files were available, the critical apparatus was compiled, in the following stages:

1. An initial collation in Collate produced a raw apparatus, in which every difference between the witnesses was included, and with the variants displayed as the machine found them, not necessarily as a philologist would choose them to appear. The next two stages resolve these problems.

2. 'Regularisation' is the process of suppressing differences between witnesses which were not to be included in the apparatus. Part of this was done by 'Fuzzy Match', making superline for final nu equivalent to final nu, equating upper- and lower-case letters, and so forth. We have regularised this apparatus very lightly. A heavy regularisation would exclude all orthographical variants. Since full information about such details is an important justification for the present volume, they are retained.

2. The second part of this stage of the process is 'Set Variants'. This determines the length and thus the content of variant readings. The application of these tools has the result that a single transcription may lead to somewhat different presentations in two collations, depending on the readings of other witnesses cited and the editor's decisions. Users may therefore expect to find the readings of witnesses being presented differently in this edition and in the Byzantine text of John, although the same transcriptions were used by each.

Once all this has been done, the text is output as XML into Anastasia, software devised for publishing Collate editions. The tagging is used to determine the format of the transcriptions.

< Title         The Base Text >