There is a name in the history of Latin manuscripts that points at two different books, and the two have been confused with one another for long enough that the confusion now travels on its own. The name is the Barberini Gospels. One of the objects that answers to it is in the Vatican: a small parchment book of the four Gospels in Latin, copied in the British Isles in the second half of the eighth century by four scribes, ornamented with a canon table, four evangelist portraits and fifteen decorated initials, and carrying at its end a prayer – ora pro Uuigbaldo, pray for Wigbald. The other object is supposed to be in Vienna: a Greek-Latin diglot of the Gospels whose Latin column preserves an Old Latin text, copied by a scribe whom its colophon calls a Galician, and which by common account carries the earliest Latin evidence for the closing chapter of the Gospel of John. They share a name. They share nothing else. This essay treats the Vatican book as the titular object, sets it beside the other, and asks how a shelfmark and a collector’s name come to detach from an object and reattach to the wrong one.
Which Barberini Gospels?
The titular object is Vatican, Biblioteca Apostolica Vaticana, Barberini lat. 570. It is listed in the Codices Latini Antiquiores as I 63, and the Vatican Library has digitised it in full. Houghton’s catalogue of the Latin New Testament manuscripts lists it under the two names it goes by in the literature – Barberini Gospels, Wigbald Gospels – and records Fischer’s figure for the agreement of its text with the Vulgate as Ev 86.4 per cent, ranging from 75.6 per cent in Mark 24 to 92.8 per cent in Matthew 13. It is a late eighth-century Hiberno-Saxon book: four scribes, Latin only, Insular in its ornament, with the figures in the evangelist portraits showing a strong Italian-Byzantine influence. It was in the Barberini library by 1679 and has been in the Vatican since 1902.
The name itself has a history that helps explain the collision. The shelfmark carries a collector’s name, and the book passed into the Vatican Library with a collection rather than as a single acquisition, arriving there in 1902. But by then the manuscript had been described and discussed for a century and a half, and a book with a memorable descriptive title takes that title from the literature rather than from the shelf. The Vatican’s digitisation is the place at which the collision can be seen directly, because it shows the object; and the object is a Latin book with a canon table, four evangelist portraits and fifteen decorated initials, with no Greek in it at all.
The other object is Vienna, Österreichische Nationalbibliothek, Cod. 48. The description usually attached to it is a Greek-Latin diglot of the Gospels with an Old Latin text; a colophon naming the scribe as a Galician; by common account the earliest evidence for the Latin translation of John 21, including the addition about the second disciple and the Talitha cumi pericope; and a marginal note in Greek, eis ten diakonistan, “into the deaconess”, preserving a trace of the deaconess tradition of the Latin West. Not one of those features belongs to the Vatican book. A Latin-only Insular gospel book cannot have an Old Latin text in a column, because it has no Greek column at all. It cannot carry a Greek marginal note about the deaconess, because it contains no Greek whatever. And a scribe who calls himself a Galician is not one of four hands working in a British and Irish scriptorium.
It is worth seeing what the standard finding aid says, because it says something quietly devastating. The Vetus Latina’s register of Old Latin Gospel manuscripts lists, among its Vienna holdings, a single folio of John copied in the sixth or seventh century in North Italy or France – Vienna, ÖNB lat. 502, fol. II, entered as VL 25 – and fragments of Matthew from a fifteenth-century Italian volume, Vienna, ÖNB lat. 563, foll. 122-77, entered as VL 43. Neither is a Cod. 48, and neither is a ninth-century bilingual. The register of Latin New Testament manuscripts proper lists no Vienna manuscript at all, and the institution records that aggregate the Austrian library’s holdings in bulk, of the kind Biblissima has built for Vienna, are not the place where such a codex appears either. The description of the Vienna diglot is therefore not merely unverified; it is a description of something that the registers which exist precisely to catalogue such things do not have.
What does exist, and fits much of that description, is at St Gall. The Codex Sangallensis 48 – GA 037 in the Gregory-Aland sequence, VL 27 with the siglum delta in the Vetus Latina – is a ninth-century Greek-Latin interlinear diglot of the four Gospels in the order Matthew, Mark, Luke and John, copied at St Gall in an Irish hand, and, on the catalogue’s own account, written by Irish monks probably in the cloister of Bobbio about 850. It is a hundred and ninety-nine folios, a little over twenty-two by eighteen centimetres, in a single column of nineteen to twenty-seven lines, Irish minuscule, black ink with rubrication in red and yellow, with a Latin preface and capitula to Matthew in the opening pages and John 19:17-35 missing. Its Latin stands on its own line above almost every Greek word. Houghton’s catalogue calls it a sister manuscript of the Codex Boernerianus, the Pauline diglot at Dresden, and its Latin column is described as based on a form of the Vulgate close to that of the Egerton Gospels but conformed to the grammar of the Greek, preserving little Old Latin evidence.
So the choice is plain. This essay treats the Vatican book as the titular object, and takes the other two as things the name has been made to say. The Saint Gall diglot is named because it is the manuscript that genuinely matches the description of a ninth-century Greek-Latin Gospel diglot with a Latin column. The Vienna manuscript is discussed below because it is the object that has absorbed the name, and because the way its description has been repeated and re-repeated, without the check that would have dissolved it, is the point of the whole exercise.
The Wigbald Gospels as an Object of Insular Craft
The Vatican manuscript does not stand alone. Houghton places it in a group, and the group is more informative than the individual. The scriptorium at Echternach – an abbey in Germany founded by the English monk Willibrord in 698 – may not have been responsible for the Echternach Gospels themselves, but it did produce several members of a family of gospel books, the best known of which is the Codex Aureus Holmiensis, copied in southern England around the year 775. The Echternach-related manuscripts are the Augsburg Gospels, the Trier Gospels and the Maaseyck Gospels, all from the first half of the eighth century. To these may be added the Barberini Gospels, and possibly also the Fleury Gospels, the Ghent Livinus Gospels and the Douai manuscript. The origin of the Barberini book is not known; that it belongs to this orbit is.
The best known member of that group, the Codex Aureus Holmiensis, shows how thoroughly Insular the ornament could be and how thoroughly the figures could be imported. It has lavishly coloured polychrome canon tables and evangelist portraits, and the manuscript is named after the gold ink used on the purple leaves that alternate with pages of white parchment; on many pages letters or rows of letters are picked out in silver or red ink to form patterns. One of the finest of the golden pages carries the chi-rho monogram, and that page alone attests the Insular origin of a book which could at first glance have belonged anywhere in western Europe. The Barberini Gospels is a smaller and plainer relative, but it belongs to the same world of production, and the judgement made about it – Insular ornament, Italian-Byzantine figures – is the judgement made about the whole orbit.
The group has a textual signature as well as a geographical one. One feature shared by many of these witnesses is the lengthy interpolation at Matthew 20:28 of a text resembling Luke 14:8-10, a reading present in many Old Latin codices as well as in some Vulgate gospel books. The oldest surviving representative of the group appears to be the sixth-century Burchard Gospels, which was in Northumbria by the end of the seventh century and was corrected at an early stage, with many of the non-standard readings erased. The point to hold on to is that this is a marker of a school and a text-type, not of a country of origin: the same reading runs through manuscripts made in England and on the Continent, and it survives in books whose physical home was decided by accident rather than by lineage.
The book itself is a work of the Insular art of the eighth century, and its material is as important to it as its text. It is written on parchment, and it opens with the canon table and its carpet page, followed by the four full-page evangelist portraits and the closing miniatures; the ornament runs in interlaced borders and large decorated initials, and the ornamental vocabulary is thoroughly Insular while the figures themselves show the strong Italian-Byzantine influence that the whole Anglo-Saxon world in that period was absorbing. Four hands did the work: at least two in a Northumbrian habit, and at least one Mercian. The record of Insular manuscripts is honest about what can be said here – the origin of the book is not known – and that honesty is part of its subject.
Set beside this, the text is a Vulgate text. Fischer’s figure of Ev 86.4 per cent, with its fall to 75.6 per cent in Mark 24, describes a book that agrees with the received Latin Bible for the best part of its length and departs from it in identifiable and traceable ways. That is the profile of a gospel book made for use in the Latin Church, not of a witness to the Latin text of the third or fourth century, and not of a bilingual. Whatever the Barberini Gospels are, they are not the manuscript with the Old Latin John 21 in them; and the text of the Vatican book, with the Old Latin element that runs through the Echternach-related group, is precisely the Old Latin element that people have been reading into the wrong object.
That is the shape of the Vatican book: a product of the eighth-century Insular trade in texts and in styles, with a Vulgate text carrying a recognisable Old Latin element, and a decoration that mixes two visual languages. It is an impressive object, and it is emphatically not a bilingual manuscript. Everything that has been said about Greek columns, interlinear Latin and the Latin of John 21 belongs to a different book – and, as the registers show, possibly to no book at all.
Uigbald: The Plea in the Colophon
The manuscript ends with a prayer for a named man: ora pro Uuigbaldo, pray for Wigbald. The transcription is standard and the spelling is worth keeping in mind, because the Latin writes the name with a double u – Uuigbald – while the form normally given in the literature is Wigbald. It is a short text, one line at the end of a book, and it is the only place in the volume where a person other than the four scribes is named.
What the colophon establishes is narrow but real. There was a Wigbald connected with this book in some way that somebody thought worth mentioning at the end of it, and somebody – a scribe, an owner, a reader – asked that prayers be said for him. What it does not establish is nearly everything else. It does not say whether Wigbald copied the book, paid for it, owned it, was its patron, or merely read it in the same community that read it before. It does not give a date, and it does not give a place. It does not say whether the name refers to one man or to a family of them.
It is worth noticing how the two objects answer to the same phrase. The Vienna description offers a colophon that names the scribe’s province; the Vatican manuscript offers a colophon that asks for prayers for a man. Both are one-line testimonies at the end of a book, and both have been made to carry far more than one line can bear. A name of origin has been treated as a solution to the question of the manuscript’s origin; a name in a prayer has been treated as a pointer to a commentary on the Octateuch. In each case the evidence is a scrap, and in each case the scrap has been stretched to cover a construction.
There has been a natural identification to make. A Wigbold wrote a set of Quaestiones in Octateuchum, a commentary on the first seven books of the Bible. Whether the Wigbald of the colophon is that Wigbold is unknown, and it should stay unknown. The reason is structural rather than incidental: a name in a colophon is a single string of letters, and two names that look alike are not two pieces of evidence. They are one piece of evidence and a conjecture. To make the identification sound would be to give the manuscript a biography, and the biography would be an invention.
It is worth being clear about this because a colophon is exactly the kind of scrap that manuscript studies runs on. Shelfmarks, colophons, catchwords, the stains on a leaf where a thumb held the book open, the ghost of writing on the page opposite: these are thin, and the whole discipline is built on them, because they are all we have. What distinguishes a sound inference from an unsound one is not the quantity of evidence but the refusal to add to it. A colophon gives you a name and a prayer. Everything else – the man’s trade, his country, his date, his link to the four scribes and the canon table and the Italian-Byzantine figures – is hypothesis, and the hypothesis has to be kept visibly separate from the evidence or it will be quoted back as evidence in a decade.
The comparison with the rest of this manuscript’s apparatus is instructive. The canon table, the evangelist portraits, the four hands of at least two different regional habits, the Italian-Byzantine manner of the figures: none of these is a document that names a place. Each is an argument that has to be constructed, from the way a book was made, and the argument can be checked by anyone who can see the book. The colophon is the only place where a person is named, and it is precisely there that the temptation to read more than is there is strongest – because a name looks like an answer.
Vienna Cod. 48 and the Earliest Latin John
The Vienna manuscript belongs to a class of book that the ninth century made and that later readers have sometimes found hard to imagine. A Greek-Latin diglot puts two texts in one volume, usually with the Greek on the right and the Latin on the left, or with the Latin written on a line of its own above the Greek and aligned word by word with it. The Codex Sangallensis 48 is the classic instance of the second arrangement in the West: the Latin stands on its own line above almost every Greek word, sometimes with a second and even a third equivalent set above the first where the scribe wanted to offer alternatives. The Vetus Latina’s register for the Latin text of John records the same feature in the technical vocabulary of the editors, where two Latin alternatives for one Greek word are separated by uel, “or”, and where the recorder has had to expand abbreviations for est, uel, autem and enim because the shorthand of the scribe makes them ambiguous. A diglot is a working instrument: it is a book in which two languages are read against each other, and the second language is there for use, not for display.
The point can be made more sharply by looking at the register entry for the Saint Gall diglot itself. The Vetus Latina records it as the Codex Interlinearis Sangallensis, entered under the siglum delta, gives the date as 960/970 in St Gall, notes that it is lacunose in John 19:17-35, and observes that the Latin version is written above a Greek text which is available as witness 037 in the edition of the majuscule manuscripts of John. The Gospel register gives the same manuscript a slightly different date, 860/70, and the two dates differ by a century, which is a small illustration of how much precision a catalogue entry can carry. It is also a reminder that what a register can give you is a shelfmark, a date, a description and a note of what is missing – and that anything more specific than that has to be argued for.
What is usually said about the Vienna Cod. 48 is this. It is a Greek-Latin diglot whose Latin column preserves an Old Latin text of the Gospels. Its colophon names the scribe as a Galician. It is by common account among the earliest, and generally taken to be the earliest, evidence for the Latin translation of John 21, including the addition about the second disciple and the Talitha cumi pericope of John 21:1-14, the catch of fish at the sea of Tiberias that ends with the dead man raised and the Aramaic words spoken by Jesus. And it carries a marginal note in Greek, eis ten diakonistan, “into the deaconess”, which preserves a trace of the deaconess tradition of the Latin West.
Two qualifications have to be attached to that paragraph, and both are worth making explicitly because the paragraph is usually quoted without either of them. The first concerns the note. The reading is standardly cited for this manuscript, and the deaconess tradition it would attest is a real feature of the Latin West, but the exact wording of the note and the folio on which it stands are not established here, and it is given as a commonly reported note rather than as a verified transcription. The second concerns the whole description, and it is the more serious one: the Vetus Latina registers, which are the standard finding aids for Old Latin Gospel witnesses and for the manuscripts of the Latin New Testament, list no Vienna Cod. 48 at all. Their Vienna holdings are a sixth- or seventh-century folio of John and a set of fifteenth-century Matthew fragments, neither of which is a ninth-century diglot. The register of Latin New Testament manuscripts proper has no Vienna entry.
So the reader who wants to check is in an awkward position. The description is fluent, repeated, and internally coherent; the standard register that would confirm or dissolve it has nothing to say. That is the condition in which a description of this kind usually turns out to be a composite – assembled from the properties of a genuine manuscript and of another, and then attached to a shelfmark. The Saint Gall diglot is the manuscript that fits the shape of the description, with the important qualification that Houghton reads its Latin column as an interlinear gloss built on a Vulgate form rather than as a pure Old Latin witness, so that the label “Old Latin text” needs qualifying even there. The Vienna shelfmark has not been shown to belong to any of it.
John 21 is a place where the difference would matter, which is why the claim has been made so often and so confidently. The Latin text of John is one of the most contested in the New Testament, and the closing chapter is one of its most contested parts. Houghton notes, in passing on his way to larger matters, that in the early history of the Vulgate the reading discentium, “the disciples”, at John 21:12 was replaced by discumbentium, “those reclining”, under the influence of the context, and that the variant was at the centre of long-running discussions about the relative primacy of the Greek and Latin text. Houghton’s Latin New Testament devotes attention to the Latin of this chapter, and the point of that attention is that the Latin text there is not simply a translation but a text with a history of its own. An early Latin witness to John 21 would therefore be a genuinely valuable object; a Latin witness whose text is largely a gloss on the Greek would be a different and lesser thing. Both have been called the Vienna Cod. 48.
The deaconess note is the most attractive of the Vienna claims and the one that most deserves the caution. If the manuscript really carries a note reading eis ten diakonistan, it preserves a memory of the deaconess in a text of the Gospels, and the deaconess tradition of the Latin West is one of the more distinctive cultural-religious traces that the early Latin text of the New Testament carries with it. It is exactly the kind of marginal note that gets quoted for years without anyone checking the folio, because a note in a manuscript kept in another country is very easy to describe and very hard to verify. The deaconess is one of the first things a reader of the early Latin West expects to meet. That expectation is part of what has built the Vienna description, and it should be read as part of the evidence rather than as a substitute for it.
Western Affinities and What Western Means
The Vienna Latin is also said to carry Western affinities, and the phrase needs unpacking before it can be used. In the study of the Latin New Testament, Western is a text-type label rather than a geographical one. It designates a body of readings that run through the Latin tradition broadly and early, associated with the category the Vetus Latina calls text-type F and with the form of the Latin Bible that circulated in the West before the revision associated with Jerome. That category is not a guess about a workshop. Houghton’s description of one of its representatives notes that the text of text-type F has similarities with quotations in Pope Innocent I, with Chromatius of Aquileia, and with Jerome’s commentary on Zechariah, and is therefore likely to derive from Italy shortly before the year 400. A manuscript whose Latin falls into that stream is telling us about the history of a text, not about the latitude of the desk on which it was written.
The label has a demonstrable history. A familiar example of it in Latin New Testament studies is the text of Acts in Codex Bezae, where the amount of additional material amounts to a separate recension around a third longer than the standard text, and which is often described simply as the Western text. That description does not mean the manuscript is western in origin, and nobody who uses the phrase thinks it does. It means that this text stands in a stream with a lineage of its own, and that the stream is old.
Three readings are commonly cited as the Western features of the Vienna diglot: the blood of Jesus at Matthew 27:25, where the received Latin has only his blood and the Western form names him outright, sanguis Iesu; the longer reading of Luke 9:56, which expands the command of the Lord to save men’s lives rather than destroy them; and the longer form of Luke 24:12, which adds the return to the sepulchre to the account of the young man in white. These are standard markers of the Western stream in the Latin Gospels, and they are the readings to be checked in the manuscript and in its edition. They are offered here as commonly cited, not as established contents of the Vienna codex, because the check is exactly what has been missing.
If they are present, the consequences are considerable. A Latin text of the Gospels carrying the blood of Jesus at Matthew 27:25, the longer reading at Luke 9:56 and the longer form at Luke 24:12 would belong to the stream of the Latin text that is demonstrably older than the version associated with Jerome, and the diglot would then be a witness of some standing to the Latin text of the Gospels in the early medieval West – and, if the Old Latin character of its column were also established, a witness to a Latin text running parallel to the Vulgate rather than derived from it. That would make it a first-rank manuscript. It would also make it, by any measure, an Insular or Iberian or Frankish or Irish book, and the colophon’s geography would have to be reconciled with a text whose ancestry reaches back into Italy before the year 400.
It is worth comparing that claim with the fingerprints of a manuscript whose affinities are not in doubt. The group to which the Barberini Gospels belongs has a signature of its own: the lengthy interpolation at Matthew 20:28 of a text resembling Luke 14:8-10, present in many Old Latin codices and in some Vulgate gospel books, and shared by the group back to the sixth-century Burchard Gospels, which was in Northumbria by the end of the seventh century. That is an Old Latin element running through a Vulgate-shaped book, and it is the sort of thing that marks a school. The Western readings are a different kind of marker altogether: they mark a text-type. Two manuscripts can both be described as showing Western or pre-Vulgate traits and carry completely different fingerprints, which is one of the reasons that a label of that kind travels so easily from one book to another and settles so comfortably in the wrong place.
The Scribe from Galicia
The most distinctive single feature of the Vienna manuscript, on every account, is its colophon. It names the scribe as a Galician. That is a claim about origin, and it is a claim made by the manuscript about itself, which is a different kind of authority from a modern catalogue entry. What it licenses is the statement that the scribe was Spanish or Portuguese, or Galician in the medieval sense – a designation of origin that says rather less about a place than it appears to, and that does not by itself name a scriptorium. What it does not license is the word Mozarabic, which is an inference about a scriptorium tradition and a religious and cultural milieu, not the colophon’s word, and which flatters the manuscript with a specificity the evidence does not supply.
The route such a scribe implies is the interesting part. A Galician writing a Greek-Latin diglot, whose Latin column is said to preserve an Old Latin text and to carry the Western character of the Latin West, is evidence for the Latin text of the Gospels travelling through the north-west of the Iberian peninsula in the early medieval period, and for a scriptorium there that was bilingual and in contact with the older Latin tradition. If the Latin of the column were indeed an Old Latin text, the combination would be remarkable, because it would put the oldest recoverable Latin of the Gospels in a ninth-century Galician book and make the diglot a point on the map of the Latin text’s survival rather than merely of its transmission. The traffic is more interesting than the man: a name in a colophon becomes interesting when it is the visible edge of something larger.
The alternatives have to be weighed. There are proposals that the scribe was of a West German or Irish milieu rather than an Iberian one, and these have to be flagged rather than dismissed, because the question of a scribe’s origin is precisely the question on which the scripts of the early Middle Ages are least reliable. Script is a better indicator of a school than of a person. The same book production trained scribes who went on to write in Italy, in France and in the British Isles; the same habits of penmanship and abbreviation were picked up by men who had never seen the place where they were learned. That is why a colophon’s statement of origin has to be weighed against the script rather than dismissed by it, and why the Galician identification, which is the older and the better-attested reading, is the one that should stand unless there is positive evidence against it.
There is a further reason for caution about the alternatives, which is that they are often proposed for manuscripts whose origin is otherwise secure, precisely in order to assimilate them to a larger pattern. If a book is one of a recognisable group, the natural scholarly question is which member of the group it belongs to, and the group is where the argument should begin. The Barberini Gospels belongs, on Houghton’s account, to the Echternach-related group of Insular gospel books, whatever the origin of its production; that is a stronger position than any hypothesis about a Galician scribe elsewhere.
There is also the wider point. When an object’s name and its shelfmark come apart, the name is usually what survives, because a name is memorable and a shelfmark is not. A collector’s library lends its name to what it holds, and the name then travels through the literature, shedding and picking up properties as it goes: a Greek-Latin diglot becomes a Gospel book with an Old Latin text; a prayer for a Wigbald becomes a pointer to a commentary on the Octateuch; a note about a deaconess becomes proof of the manuscript’s standing. The Barberini Gospels are two objects because two sets of properties were collected under one name. The cure for this is dull and has not changed since the first catalogues were compiled: give the shelfmark, describe the object, and let the name follow.
References
- Barberini lat. 570, Barberini Gospels, Wigbald Gospels, full digitisation — Biblioteca Apostolica Vaticana, DigiVatLib manuscript viewer, accessed 2 November 2026
- The Vatican Library — Biblioteca Apostolica Vaticana, library portal, accessed 2 November 2026
- The Latin New Testament: A Guide to its History, Texts, and Manuscripts — H. A. G. Houghton, Oxford University Press, 2023, open-access copy at the University of Birmingham PURE repository, accessed 2 November 2026
- The Latin New Testament: Appendix 3, Additional Gospel Manuscripts, entry for Barberini lat. 570 — H. A. G. Houghton, Oxford University Press, 2023, publisher landing page, accessed 2 November 2026
- The Latin New Testament: Appendix, catalogue of Latin New Testament manuscripts — H. A. G. Houghton, Oxford University Press, 2023, publisher landing page, accessed 2 November 2026
- The Latin New Testament, chapter 8: The Text of the Early Latin New Testament — H. A. G. Houghton, Oxford University Press, 2023, publisher landing page, accessed 2 November 2026
- Old Latin Gospel manuscripts, register including VL 27 Codex Sangallensis 48 — Vetus Latina Institute, University of Birmingham, accessed 2 November 2026
- Vetus Latina Iohannes: manuscripts — Vetus Latina Institute, University of Birmingham, accessed 2 November 2026
- Manuscripts of the Old Latin Bible, master register — Vetus Latina Institute, University of Birmingham, accessed 2 November 2026
- Concordance of sigla for Old Latin New Testament manuscripts — Vetus Latina Institute, University of Birmingham, accessed 2 November 2026
- Manuscript collections, Department of Manuscripts and Rare Books — Österreichische Nationalbibliothek, accessed 2 November 2026
- Österreichische Nationalbibliothek, English site — Österreichische Nationalbibliothek, accessed 2 November 2026
- ONB digital, digitised holdings — u00dsterreichische Nationalbibliothek, accessed 2 November 2026
- ONB catalogue search — u00dsterreichische Nationalbibliothek, accessed 2 November 2026
- Vienne, Österreichische Nationalbibliothek, institution record — Biblissima, accessed 2 November 2026
- Vienne, Österreichische Nationalbibliothek, Cod. Hist. gr. 48, Johannes Chortasmenus — Biblissima, accessed 2 November 2026
- Codex Delta (Cod. Sang. 48): the Gospels in Greek with an interlinear translation into Latin — e-codices, Universitätsbibliothek Freiburg, accessed 2 November 2026
- The Latin Text of John in the Saint Gall Bilingual Gospels (Codex Sangallensis 48) — H. A. G. Houghton, in At One Remove: The Text of the New Testament in Early Translations and Quotations, Gorgias Press, 2021, accessed 2 November 2026
- Codex Sangallensis (GA 037), manuscript record and digitised images — Center for the Study of New Testament Manuscripts, accessed 2 November 2026
- Manuscript search interface — Center for the Study of New Testament Manuscripts, accessed 2 November 2026