About this edition

The edition

The Greek New Testament According to the Byzantine Text was edited and prepared by J. Jeffrey Dodson and Dwayne Green for Guardian Press, with the Guardian Press Editorial Committee. It runs to 752 pages.

The main text is the Robinson-Pierpont New Testament in the Original Greek: Byzantine Textform 2026. Every change from the earlier 2005 and 2018 editions is listed in Appendix A.

Two sets of footnotes sit under the text on each page.

• The Edition Apparatus compares the main text with eleven printed editions: Hodges & Farstad, Pickering, the Antoniades Patriarchal Text, the Editio Critica Maior, Nestle-Aland 27 and 28, the SBL Greek New Testament, Tyndale House, Westcott & Hort, Scrivener, and Stephanus. Minor differences such as accents, punctuation, movable nu, final sigma and spacing in compound words are not recorded.
• The Manuscript Apparatus gives the manuscript evidence for fifty significant variants chosen by the editorial committee. For each one it lists every papyrus and uncial extant at that point (ninth century and earlier), plus select minuscules through the tenth century.

Four appendices document the edition changes, corrections to the comparison editions, alternate readings and long variants, and a dated list of every manuscript cited.

The whole edition is in the public domain as of 16 August 2026. You may copy, print, quote, adapt and share it, even commercially, without permission or fee.

How this website was made

The book is the authority. This website is a set of tools for exploring it, and every page here was built from the printed edition.

Fifty variants

The readings, the manuscripts cited for each reading, and the edition notes come from the Manuscript Apparatus and Edition Apparatus. The short plain-English summary on each card was written for this website and is not from the book; where it quotes English, it uses the wording of the English text described below. Each variant also shows the English of its verse, except where the edition prints no text for that verse. A box in the variant window turns the English off and on. The timeline places each manuscript by the century given in Appendix D.

Compare editions and Explore variants

Every note in the Edition Apparatus was read from the book’s PDF by software and stored as data: the book, the page and verse, the lemma, which editions share the main text, which read differently, and which print the reading in brackets or parentheses. Each note was also labeled by what the difference is (words omitted or added, text replaced, word order, spelling, brackets, and so on). Those labels were assigned by rules that read the note’s wording. A single note can have more than one label, so the categories overlap, and a few notes may be labeled imperfectly.

The Compare editions page draws its pictures from this data. In the map of chapters, each square counts the notes in one chapter, and the shading runs from the lightest fifth of chapters to the darkest fifth. A switch above the map changes it to count only the notes marked as changing the English; beneath it the page gives the range, across the eleven editions, of the share of each edition’s differences that change the English. In the comparison chart, each bar shows the share of a part’s notes in which one edition reads differently from the main text. In “Standing alone,” each row counts the notes in which one edition, or the main text, has a reading that no other edition shares.

Each reading in a note is given one main label, and a reading of any kind can also be marked “has bracketed text.” In order of testing, the main label is: the reading reads the same as the base text (the same words in the same order), or only the brackets differ (the same words in the same order, but some of its words or letters are in brackets); it differs only by elision or crasis; it has a different spelling (the same words in the same order, but at least one word spelled differently); it has a different word order (the same words, rearranged); it omits the base reading (all of its words); it omits words (some of them, in one place or several, with the rest in the same order); it adds words (in one place or several, with the base words in the same order); or it replaces the base text (everything else). If some words of a reading of any kind are in brackets, the reading also shows “has bracketed text.” A reading that is only a bracketed list of editions is labeled for its brackets: single brackets, double brackets, or chevrons around the reading. A few readings that are described in English, such as moving a passage to another place, are labeled Other. Two words count as the same word when they have the same Strong’s number and part of speech and are spelled alike, apart from elision and a movable nu or final sigma that the reading marks in parentheses, as in εστι(ν). Accents and breathings are compared only when both words show them. A word with the same Strong’s number and parsing but a different spelling counts as a spelling difference. A reading’s words are also counted: it shortens the text when it has fewer words than the main text at that point, and lengthens it when it has more.

The “Show comparison” button under each reading in Explore variants draws on the same data. It pairs each word of the reading with a word of the main text, keeping the words in order. Words spelled the same are paired first; then a word that is spelled differently is paired with a leftover word of the main text that has the same Strong’s number and sits in the same stretch of text. When a reading has the same word both inside and outside brackets, a main-text word pairs with the one bracketed the same way. It shows each pair with its dictionary form, gloss and parsing from the Robinson-Pierpont analysis. Each pair is marked with a number, the word’s place in the main text, so the pairing does not depend on color. A shade marks a difference: words that moved, and words spelled or parsed differently from their partners, are shaded, and words that are unchanged and in place are left plain. When the two words of a pair are spelled differently, the fewest letters that differ are highlighted and underlined with a wavy line, so the difference can be seen at a glance: the letters left unmarked are the same letters in the same order in both words. A pair that is the same dictionary word in a different form, such as a present and an aorist participle, is not treated as a spelling difference: its letters are not marked, and the difference in parsing is highlighted instead. An ellipsis in the main text, which stands for words the note does not quote, is filled in with the actual words of the verse, in a red bracketed span, and so is an ellipsis in a reading, so that the reading’s words are paired with real words and do not look like additions. When a stretch of the main text is replaced by the same number of different words, the words are paired one to one, in order, and marked as a different word. When the replacement has a different number of words, a word is paired with a word of the other side only if the two share more than a third of their letters, in the same order; the words left over are marked as added or omitted. When one word of each side is left over and both are the same kind of word, such as two verbs, they are paired as a replacement even if they stand in different places, and so are a single negative word on each side, such as ουδε and ουδενι. A crasis, where two words in one text are written as one word in the other, is shown as a group. A word with no partner is marked with a plus sign if the reading adds it, and with a dashed outline if the reading omits it. The pairing is made by software and has not been checked word by word.

In the chart of kinds of difference, the eleven editions are ranked by how often their readings shorten the text or lengthen it. This is a count of words only: a reading that has fewer words than the main text at that point shortens it, whether it omits words or replaces them with fewer, and a reading with more words lengthens it. Each edition’s two figures are its shares of its differences that change the number of words. Readings with the same number of words, such as differences of spelling or word order, elision, and swapping one word for another, are not charted; the kind filter in Explore variants finds them, along with shortening and lengthening.

The section on RP^mg counts the notes where the Robinson-Pierpont edition offers an alternate Byzantine reading. Its first chart ranks the books by the number of such notes for every 100 verses, so that long and short books can be compared fairly; verses are counted as this edition prints them, and a verse number with no text of its own is not counted. Its introduction also says how many of these notes change the English. Its second chart ranks the editions by how often they read the alternate reading. “Byzantine” there means Hodges-Farstad, Pickering and Antoniades, “critical” means ECM (where it exists), NA27, NA28, SBL, Tyndale House and Westcott-Hort, and “Textus Receptus” means Scrivener and Stephanus. The list of thirty notes gathers the alternate readings that a critical edition shares and no other Byzantine or Textus Receptus edition does, and the page number in each row opens that page of the edition. It is a pattern found in the data, not a judgment about which reading is right.

Explore variants lists the notes that match your choices. Each choice appears as a button above the results, and clicking it removes that choice. The number beside each choice in the lists is how many notes you would see if you picked it and kept your other choices. If nothing matches, the page offers the choices you could remove to see notes. Each note lists its readings first, in the order the printed edition gives them, and below them the main text of its verse in Greek (labeled “Main text”), with the words the note is about shaded (where the main text of a note has an ellipsis, the words it stands for get a lighter shade), and under that the English of the verse (labeled “ASVBT”); neither of these two lines is a variant reading; the Greek or English words you searched for are highlighted. One button copies the reference and the verse, with the Greek and the English that are showing and without the shading or highlighting; another, “Copy this note,” copies the reference, the main text and each reading with its edition labels, as plain text for pasting. An English word or phrase is searched for in the verse and in the note’s translation hints. A switch above the results hides the English for readers who would rather not see it. The same choice is offered in the variant window on the Variants and Manuscripts pages, the two switches always agree, and the choice is remembered on that device. A second switch above the results hides the Greek in the same way. Point at a word of the Greek, or tap it on a phone, and a small box gives its Strong’s number, its dictionary form, a short gloss, and its parsing with the code spelled out in English. The word you are pointing at turns blue and is underlined, with a pale blue box behind it, so it is not mistaken for the red shading, which marks only the words the note covers. The word data is loaded the first time you do this. This works on the verse, on the main-text words of each note, on the readings of each note, and on the readings in the variant window. The box describes the word as it stands in the text of the edition that prints it, and a reading is analyzed from one of the comparison editions that has that reading. When no edition prints the whole reading, the analysis is the most common one for that word in the same verse. When a word can be analyzed in more than one way, the box lists the possibilities. Where a note refers to Appendix C, those words are a link that opens the first page of that appendix in the page viewer. The box “Effect on the English” shows only the notes where a variant changes the English, or only those where it does not. The marks come from comparing each note with the footnotes of an English translation of this Greek text, together with a later review of the remaining notes for readings that clearly add, drop, or replace words. That translation footnotes only the variants that make a difference to its wording, so a note is marked “Changes the English” when the translation has a footnote at the same place. The marks are that translator’s judgment. Roughly three notes in ten are marked “Changes the English”; the rest are differences the translation does not footnote. The results can be listed in the order of the edition, or with the notes where the most editions differ from the main text first. The page also remembers your last search, but only on your own device; nothing about it is sent anywhere.

The English text on this website is the Boyd ASVBT, a form of the American Standard Version made to follow the Byzantine text, and it is in the public domain. It was matched verse for verse to this edition: every verse the edition prints has its English, and the number of verses is the same in both. The English is shown for the reader’s convenience. It is not part of the printed edition, and a translation cannot show every difference between the Greek readings in the notes. To help with this, every note in Explore variants that changes the English has a collapsed “Translation hints” link, and so do many notes where the English stays the same, in which case the hint says so. A hint gives approximate English wording for the changed words, in the style of the ASVBT. The hints were written for this website, are not part of the ASVBT or the printed edition, and are only approximate.

To check the extraction, the number of times each edition’s abbreviation appears on every page was counted a second, independent way and compared with the stored data. The two agreed on every page that carries notes. That check proves the editions were captured correctly. It does not prove every label is right, so please use the report form if you find one that is not.

Page numbers on this website are the page numbers printed in the book.

Manuscripts

The Gregory-Aland number, letter, contents and date of each manuscript come from Appendix D, which takes its dates from the Kurzgefaßte Liste of the Institute for New Testament Textual Research (INTF) in Münster. The names and the places where the manuscripts are kept were compiled from public lists of New Testament manuscripts, and each card links to the manuscript’s Liste entry so you can check it against the official record. The “follows the main text” figure counts, among the fifty variants, how often a manuscript’s own reading matches this edition’s main text. It leaves out readings supplied only by a later corrector. For a manuscript cited in only a few places, the percentage can be misleading.

Fonts and design

The site sets its text in EB Garamond and its Greek in GFS Didot. Both are free fonts released under the SIL Open Font License, and they are served from this site, so no font service sees your visit.

Visitor counts

The site uses a privacy-friendly counter to tally page views and clicks on the download links. It sets no cookies and does not build profiles of visitors. The counts are private to the people who run the site.

Corrections

No edition is free of mistakes. Known corrections to the printed book are listed on the Errata page, and the same page has a form for reporting a new one, whether in the book or on this website.

Use it without a connection

This website can be added to the home screen of a phone or tablet, where it opens like an app and keeps working with no internet connection. The pages, charts, search and manuscript information are stored on the device the first time you visit it. Whenever you are online, you get the newest version.

iPhone and iPad: open the site in Safari, tap the Share button, and choose “Add to Home Screen.” Android: open the site in Chrome, open the menu, and choose “Install app” or “Add to Home screen.”

The page images of the edition are large, about 76 MB in all. Pages you view are saved as you read. To have the whole edition available offline, save it here once, preferably on Wi-Fi.