Codex Amiatinus (Amiatino 1)

Introduction: A Book That Was Not What It Seemed

For the greater part of two centuries, the largest and most imposing Bible to survive from the early Middle Ages was catalogued as an Italian book. It stood in the Biblioteca Medicea Laurenziana in Florence, where it had arrived from a suppressed Tuscan abbey in the 1780s, and its uncial script, its Mediterranean-looking miniatures, and its exceptionally pure Hieronymian text all seemed to place it securely in sixth-century Italy, most probably in the south. Angelo Maria Bandini, the Laurenziana’s librarian, catalogued it as such. Constantin von Tischendorf, who in the mid-nineteenth century published its New Testament and thereby made it famous among biblical scholars, accepted the attribution. The manuscript’s own dedication page appeared to confirm it, naming as donor a certain Peter, abbot “from the far borders of the Lombards.”

The attribution was wrong. In 1886–1888 Giovanni Battista de Rossi proposed, and Fenton John Anthony Hort demonstrated, that the dedication had been tampered with, that beneath the alterations lay the name of Ceolfrith, abbot of the Northumbrian twin monastery of Wearmouth-Jarrow, and that the codex was therefore an Anglo-Saxon book of the years around 700, made in the far north of Britain and carried south as a gift for the tomb of St Peter. The recognition ranks among the most consequential single acts of manuscript detection in the history of the discipline. It transformed the Codex Amiatinus from an anonymous late-antique survival into the most eloquent surviving witness to the intellectual ambitions of the monastery in which Bede spent his life.

This essay attempts a synthetic account of the manuscript on four planes that scholarship has come to treat as inseparable. The first is material: the Amiatinus as a physical object of extraordinary scale, whose production required resources, planning, and a trained workforce on a level almost without parallel in its period, and whose parchment has recently become the subject of fresh scientific dispute. The second is documentary: the manuscript’s relationship to two near-contemporary Northumbrian texts, Bede¹’s Historia abbatum and the anonymous Vita Ceolfridi, which between them describe its commissioning, its purpose, and its departure for Rome, and which make the Amiatinus one of the very few early medieval books whose origins are narrated by identifiable contemporaries. The third is artistic and intellectual: the programme of images and diagrams occupying its opening quire, which has generated a large and unresolved literature concerning its models, its designer, and its theological argument. The fourth is textual: the codex’s standing as the earliest complete Latin Vulgate in existence and, in consequence, as a foundational witness for every critical edition of Jerome’s Bible from Tischendorf to Stuttgart.

Throughout, an effort is made to distinguish three categories of statement that are too often conflated in general accounts: what the primary sources actually say, what specialists have inferred from them, and what remains genuinely contested. A number of claims that circulate widely in reference works and popular writing prove, on inspection, to be unverifiable or simply mistaken; three of these are addressed explicitly below, since correcting them is part of describing the manuscript accurately. Where the evidence runs out, this essay says so rather than filling the gap.

The Object: Scale, Script, and the Discipline of the Page

The Codex Amiatinus is now Florence, Biblioteca Medicea Laurenziana, MS Amiatino 1. Its leaves measure approximately 505 × 340 mm, with a written area in the region of 360–375 × 260 mm; the book-block is about 260 mm thick and the whole volume weighs some thirty-four kilograms. Hort’s response, recorded in the late nineteenth century and quoted ever since, was that the codex is so impressive as to fill the beholder with a feeling akin to awe. The observation is not merely rhetorical. Very few books of comparable bulk survive from the entire first millennium, and none earlier that contains the whole of Scripture in Jerome’s version.

The number of folios is given variously in the literature. The most careful recent statements, following Celia Chazelle’s monograph, count 1,029; older catalogue descriptions, including that of the World Digital Library, record 1,030; the figure of 1,040 that circulates in some reference sources appears to be erroneous and is internally inconsistent with the other measurements usually given alongside it. The prudent course is to report the range rather than a spurious precision, since the discrepancy turns partly on how the disordered opening quire is counted. The gatherings are quaternions, that is quires of four bifolia, in the standard late-antique and early medieval fashion.

The layout is one of the manuscript’s most telling features. The text is written in two columns of some forty-three to forty-four lines, and it is arranged per cola et commata: that is, the lines are broken according to sense-units rather than filled to the margin, so that the reader’s eye and voice are guided through the syntactic articulation of the sentence. This system, which Jerome himself commended for the prophets and which Cassiodorus discussed at length, is a scribal technology of comprehension and of public reading. Its adoption in the Amiatinus is a deliberate archaism, or rather a deliberate Romanism: it aligns the book with the practice of the late-antique Mediterranean rather than with anything native to Britain or Ireland. It is also expensive, since sense-broken lines waste a good deal of parchment. In a codex already vast, the choice compounds the extravagance.

The main script is uncial, executed with a control and consistency that long deceived palaeographers into an Italian attribution. E. A. Lowe catalogued the manuscript in Codices Latini Antiquiores (III.299) and devoted a separate study to its uncial, which remains a point of departure. The number of scribes has been debated. The most frequently cited modern count, associated with Richard Marsden’s work and repeated by subsequent scholars including Peter Darby, is seven hands; earlier writers sometimes proposed nine. Recent work has pressed the question further: Emily Cardwell, writing in Early Medieval Europe in 2025, notes that beyond the seven hands in the Amiatinus itself, an eighth and a ninth can be distinguished in the surviving leaves of the sister pandects, with a possible tenth in a related fragment at Hereford. The implication is significant. Wearmouth-Jarrow was not a house with one or two gifted calligraphers but a community that had trained a substantial cohort of scribes to a common and very high standard in a script that was not the local vernacular hand.

Malcolm Parkes, in his 1982 Jarrow Lecture on the scriptorium, argued that the community developed its own graded system of uncial, including a working variety he called “Capitular Uncial,” used for subsidiary elements of the page. The observation matters because it points to institutional method rather than individual virtuosity: a house that could distinguish grades of script for distinct textual functions, and could teach that distinction, was operating a scriptorium in the fullest organisational sense. Display scripts in the Amiatinus include rustic capitals and higher grades of uncial for headings; the purple contents page presents the titles of the biblical books in uncial against the dyed ground, a conscious evocation of imperial and late-antique luxury bookmaking.

The Cost of a Pandect: Parchment, Economy, and a Recent Reappraisal

The most immediately arresting fact about the Amiatinus is not its script but its raw material. A book of over a thousand large folios consumes an enormous quantity of skin, and since Ceolfrith commissioned three such books, the total demand was formidable.

The classic study is Richard Gameson’s article “The Cost of the Codex Amiatinus,” published in Notes and Queries in 1992. Building on figures proposed by Rupert Bruce-Mitford, Gameson calculated that the three pandects together required something in the region of 1,545 skins, that is roughly 515 per volume. This is the figure the British Library used in its public presentation of the manuscript in 2018, and it is the one most often repeated.

Gameson’s argument, however, was not that the enterprise was ruinously expensive but very nearly the opposite. He characterised parchment as a valued and valuable but not essentially prohibitive commodity, and he made the arithmetic of production concrete. If the work is spread over the period from Ceolfrith’s accession as sole abbot in 688 to his departure in 716, the average annual requirement is only around fifty-five skins; even compressed into a single decade, the figure is about 165 a year.

Against a community holding at least 150 hides of land by 716, and drawing on food-renders and tithes as well as its own husbandry, such quantities were manageable. Gameson further stressed that a slaughtered animal yielded far more than its pelt — meat, horn, glue, bone, fertiliser — so that the “cost” of parchment cannot be reckoned as though skins were produced for their own sake. By way of comparison, he noted that the Amiatinus consumed at least seventeen times the resources of the St Petersburg Bede. He also recorded that the surviving leaves of the companion pandects measure approximately 480 × 355 mm with a written area of 360 × 255 mm, closely comparable to the Amiatinus and confirming that the three books were conceived as a matched set.

Two corrections must be entered here, both of which bear directly on how the manuscript is described in general literature.

The first concerns a story repeated in innumerable accounts: that in 692 King Aldfrith of Northumbria granted the monastery land sufficient to support two hundred head of cattle, specifically in order to supply vellum for the pandects. This claim cannot be verified in the primary sources. What Bede’s Historia abbatum (ch. 15) and the anonymous Vita Ceolfridi actually record is a transaction of a different kind: Aldfrith granted eight hides of land near the River Fresca in exchange for a codex of cosmographers of wondrous workmanship (Cosmographiorum codice mirandi operis) that Benedict Biscop had purchased in Rome, and this land was subsequently exchanged, under King Osred, for twenty hides at Sambuce. The transaction is evidence of the extraordinary value attached to books at Wearmouth-Jarrow, and of the monastery’s landed wealth, but it says nothing about cattle or vellum. The “two hundred head of cattle in 692” appears to be a modern paraphrase that has hardened into a fact through repetition. It should not be reproduced.

The second correction is more recent and more technical. The assumption that all three pandects were written on calfskin has been challenged by Jiří Vnouček, whose study of the parchment appeared in the Journal of Paper Conservation in 2020. On the basis of visual and microscopic examination combined with experimental reconstruction, Vnouček argues that the text-block of the Codex Amiatinus is composed of goat and sheep skins, whereas the surviving leaves of the two sister bibles are calf.

If sustained, this finding complicates the traditional narrative in two ways. It disturbs the tidy assumption that the three volumes were produced from a single, uniform supply, and it unsettles the broader association of Insular high-grade bookmaking with calfskin. The conclusion is not yet universally accepted and would benefit from confirmation by peptide mass fingerprinting (ZooMS), which can identify species from collagen without damaging the leaf. Until such analysis is published, the responsible formulation is that the species identification is contested, with recent specialist opinion favouring sheep and goat for the Amiatinus itself.

The manuscript no longer retains an early binding; it was removed from its nineteenth-century covers in 1999 during the preparation of the electronic facsimile. Its condition is famously good, the parchment retaining a freshness that has itself contributed to the impression of a much later date. Between 1999 and 2001 a campaign of non-destructive analysis — X-ray fluorescence, particle-induced X-ray emission, Raman spectroscopy, ultraviolet examination — was undertaken, in part to characterise the pigments and in part to study the offsets, the ghostly impressions left by painted surfaces on facing leaves, which provide evidence for the original arrangement of the disordered opening quire.

Wearmouth-Jarrow: A Mediterranean Monastery in Northumbria

The Amiatinus is unintelligible apart from the institution that produced it. Benedict Biscop founded the monastery of St Peter at Monkwearmouth around 673–674, on land granted by King Ecgfrith of Northumbria, and the sister house of St Paul at Jarrow in 681–682 on a second grant. The Jarrow dedication stone, still preserved, records the church’s dedication on 23 April in the fifteenth year of Ecgfrith and the fourth year of Ceolfrith’s abbacy, and is among the earliest dated inscriptions from Anglo-Saxon England.

Stone, Parchment, and the Authority of a Date

The inscription illustrated here is the dedication stone of the church of St Paul at Jarrow, Tyne and Wear, where it is set into the fabric above the chancel arch.

The inscription is cut in two registers, divided by a horizontal band, and reads:

✠ DEDICATIO BASILICAE SCI PAVLI VIIII KL MAI ANNO XV ECFRIDI REG

CEOLFRIDI ABB EIVSDEMQ ECCLES DO AVCTORE CONDITORIS ANNO IIII

Expanded, with abbreviations resolved:

Dedicatio basilicae Sancti Pauli VIIII Kalendas Maias, anno XV Ecfridi regis, Ceolfridi abbatis eiusdemque ecclesiae Deo auctore conditoris anno IIII.

“The dedication of the basilica of St Paul on the ninth day before the Kalends of May, in the fifteenth year of King Ecgfrith, and in the fourth year of Ceolfrith, abbot and, under God’s guidance, founder of the same church.”

The Jarrow stone is therefore among the very small number of securely dated inscriptions from Anglo-Saxon England, and one of the earliest. Its evidential value for the chronology of early Northumbrian building, for the career of Ceolfrith, and for the palaeography of Insular display lettering is correspondingly high. Elisabeth Okasha’s Hand-List of Anglo-Saxon Non-Runic Inscriptions (Cambridge, 1971) and Rosemary Cramp’s volume on County Durham and Northumberland in the Corpus of Anglo-Saxon Stone Sculpture (Oxford, 1984) remain the standard points of reference for its description and context.

Biscop travelled repeatedly to Rome and Gaul, returning with books, relics, panel paintings, liturgical practice, and craftsmen; the glaziers and masons he brought from the Continent gave his churches a fabric unlike anything then being built in Northumbria. The library he assembled, and which Ceolfrith afterwards doubled, was the material condition of everything Bede subsequently wrote. It is important to grasp what this meant in practice. A Northumbrian monk of the late seventh century, writing at the northern edge of the known world, had access to a working collection of patristic, exegetical, grammatical, and computistical texts that would have been respectable in Rome itself. The Amiatinus is the visible face of that acquisition programme: not merely a book made at Wearmouth-Jarrow but a book made possible by the library, and made in emulation of it.

The monastery’s cultural orientation was self-consciously Roman. This is the quality that scholarship has come to label Romanitas, and it operated at every level of the Amiatinus: in the choice of uncial rather than an Insular minuscule or half-uncial; in the per cola et commata layout; in the late-antique idiom of the miniatures, which are almost entirely free of the interlace, spirals, and zoomorphic ornament that define the Lindisfarne Gospels or the Book of Durrow; in the selection of a text as close to Jerome’s as could be obtained; and, most explicitly of all, in the destination of the third pandect, which was to be laid at the tomb of St Peter. The contrast with the Insular manuscripts produced in the same generation and within a few miles is one of the most instructive juxtapositions in early medieval art. It is not a contrast between provincial and metropolitan, or between crude and accomplished; it is a contrast between two fully realised visual cultures, one drawing on native and Irish ornamental traditions and the other on Mediterranean models, coexisting in the same kingdom and sometimes in the same generation.

Ceolfrith himself is the pivot. He had accompanied Biscop, knew Rome, and on becoming sole abbot in 688 pursued an expansionist policy for the library and the scriptorium. Ian Wood’s Jarrow Lecture of 1995, The Most Holy Abbot Ceolfrid, remains the standard treatment of his career and of the way the community subsequently remembered him. Bede, who entered the monastery as a boy and remained there until his death, is the other presence, and the question of his personal involvement in the pandect project is among the most debated issues in the modern literature. It is taken up below.

The Primary Testimony: Three Pandects and a Journey

The Codex Amiatinus belongs to a small class of medieval books whose making is described by near-contemporary written sources. Two Northumbrian texts, both produced within a generation of the events, record the commissioning of the pandects. They were most recently edited by Christopher Grocock and Ian Wood as Abbots of Wearmouth and Jarrow in the Oxford Medieval Texts series (2013), which supersedes earlier editions and should be the basis of any citation.

Bede’s Historia abbatum, chapter 15, records that Ceolfrith arranged that tres pandectes nouae translationis, ad unum uetustae translationis quem de Roma adtulerat, ipse super adiungeret — “three pandects of the new translation, in addition to the one of the old translation which he had brought from Rome, he should himself add over and above.” Two points deserve emphasis. First, “pandect” (pandectes) denotes a complete Bible in a single volume, a format that was rare in this period and would remain so until the Carolingian ninth century. Second, Bede distinguishes explicitly between the noua translatio, Jerome’s version, and the uetus translatio, the Old Latin, and states that a book of the older kind had been brought from Rome. That older pandect is the key to the Cassiodoran question discussed below.

The anonymous Vita Ceolfridi, chapter 20, supplies the purpose:

ita ut inter alia tres pandectes faceret describi, quorum duo per totidem sua monasteria posuit in ecclesiis… tertium autem Romam profecturus donum beato Petro apostolorum principi offerre decreuit

“so that among other things he had three pandects copied, of which he placed two in the churches of his two monasteries… but the third he decided to offer as a gift to the blessed Peter, prince of the apostles, when he set out for Rome.”

Chapter 37 of the same text describes the gift as a pandect

interpretatione beati Hieronimi presbiteri ex Hebreo et Greco fonte transfusus,

“translated by the blessed Jerome the priest into Latin from the Hebrew and Greek source.”

Bede repeats the notice in De temporum ratione, chapter 66:

misit ecclesiae sancti petri pandectem a beato hieronimo in latinum ex hebreo uel graeco fonte translatum.

The consistency of these statements across three independent notices, two of them by Bede, is unusual and valuable. It establishes that three complete single-volume Bibles were made at Wearmouth-Jarrow; that two remained in the monastery’s churches; that the third was destined for Rome; and that the text was understood to be Jerome’s.

Ceolfrith departed Northumbria on 4 June 716, carrying the gift. He did not reach Rome. He died at Langres in Burgundy on 25 September 716, and his companions continued the journey with the book. The intended recipient was Pope Gregory II, and the intended resting place the shrine of St Peter. Whether the codex reached either is not documented, a silence to which we shall return.

The date of production can therefore be bracketed but not fixed. The secure terminus is “before June 716.” The commonly cited range is c. 688–716, taking Ceolfrith’s accession as sole abbot as the starting point; some scholars narrow to c. 692–716 or propose a date around 700 on palaeographic and stylistic grounds. No internal evidence permits greater precision, and confident single-year datings should be treated with suspicion.

On Bede’s involvement, the position of the field has shifted decisively over the last three decades without arriving at certainty. Richard Marsden’s article “Manus Bedae: Bede’s Contribution to Ceolfrith’s Bibles” (Anglo-Saxon England, 1998) argued for his active editorial participation, and Paul Meyvaert went further, associating Bede with the capitula lectionum of the Old and New Testaments (Revue Bénédictine, 1995) and with the conception of the prefatory material. The hypothesis is now widely accepted and is intrinsically plausible: Bede was resident, learned, deeply engaged with the biblical text, and precisely the kind of person to whom such work would fall. But it remains an inference from textual and palaeographic evidence, not a documented fact. No source names Bede in connection with the pandects. The essay that treats his authorship of the Amiatinus programme as established has overstepped the evidence, however attractive the conclusion.

The Erased Dedication: How the Manuscript Was Recovered

A Palimpsest of Intentions

The opening leaf of the Codex Amiatinus carries no image, no diagram, and no biblical text. It carries twelve lines of verse beneath a drawn arch, set in a field of blank parchment so generous that the written area occupies barely a third of the page. The Roman numeral I in the upper left corner belongs to the refoliation of 1876 and is a modern intrusion. Everything else on the leaf is early eighth century, with one crucial exception, and that exception is the reason this is the most consequential single page in the manuscript.

The text reads:

✠ CENOBIUM AD EXIMII MERITO VENERABILE SALVATORIS QVEM CAPVT ECCLESIAE DEDICAT ALTA FIDES PETRVS LANGOBARDORVM EXTREMIS DE FINIB· ABBAS DEVOTI AFFECTVS PIGNORA MITTO MEI MEQVE MEOSQ· OPTANS TANTI INTER GAVDIA PATRIS IN CAELIS MEMOREM SEMPER HABERE LOCVM

Set as verse, these twelve written lines resolve into six, three elegiac couplets:

Cenobium ad eximii merito venerabile Salvatoris quem caput ecclesiae dedicat alta fides, Petrus Langobardorum extremis de finib(us) abbas devoti affectus pignora mitto mei, meque meosque optans tanti inter gaudia patris in caelis memorem semper habere locum.

“To the community venerable by the merit of the exalted Saviour, whom lofty faith dedicates head of the Church, I, Peter, abbot from the far borders of the Lombards, send these pledges of my devoted affection, desiring that I and mine may always have a place of remembrance in heaven amid the joys of so great a father.”

Three stretches of text are written in a visibly different ink, a warm reddish brown against the dense blackish brown of the remainder, and in a hand whose letterforms sit less comfortably in the line. They are: NOBIVM in the first line, following an original CE; SALVATORIS in the second; and PETRVS LANGOBARDORVM in the fifth. These are precisely the passages that Angelo Maria Bandini, cataloguing the manuscript at the Biblioteca Medicea Laurenziana in the eighteenth century, identified as later substitutions written over erasures.

The page thus displays its own falsification. Whoever made the alteration did not attempt to match the original ink, and the result is that the dedication announces, to any attentive eye, that it has been tampered with. Bandini saw this; so can anyone looking at a good photograph. What Bandini could not do was recover what had stood there before.

His observations were nonetheless acute and remain the foundation of all subsequent discussion. He noted that cenobium had replaced a shorter word. He noted that in the second line the final letters of Salvatoris stood on parchment that had not been scraped, so that ten letters had been fitted into a space cleared for five. He noted that the metre was broken. And he noted the internal contradiction of the sense: the phrase caput ecclesiae, head of the Church, is the standard designation of St Peter, not of the Saviour, and it sits awkwardly in a dedication addressed to a monastery of the Saviour. The document was arguing against itself.

The resolution came in the 1880s. Giovanni Battista de Rossi proposed the connection between this tampered dedication and the pandect that Bede records Abbot Ceolfrith of Wearmouth-Jarrow carrying toward Rome in 716; Fenton John Anthony Hort confirmed it by locating, in the anonymous Vita Ceolfridi, a transcription of the original verses:

CORPUS AD EXIMII MERITO VENERABILE PETRI DEDICAT ECCLESIAE QUEM CAPUT ALTA FIDES CEOLFRIDUS ANGLORUM EXTIMIS DE FINIBUS ABBAS…

“To the body of Peter, rightly to be venerated for his surpassing merit, whom lofty faith [acknowledges as] head of the Church, Ceolfrith, abbot from the furthest borders of the English.” The first word was Corpus, not Bandini’s Culmen; and Samuel Berger subsequently corrected an early reading of Britonum to Anglorum. Once the two states of the text were set side by side, the mechanism of the alteration became transparent. A later hand had converted a dedication to the body of Peter at Rome into a dedication to a monastery of the Saviour, and had replaced the name and origin of the Anglo-Saxon abbot with those of a Lombard one.

The consequences were immediate and total. A manuscript universally assigned to sixth-century Italy was revealed as Northumbrian, made at Ceolfrith’s command within Bede’s lifetime and almost certainly within his sight. Every subsequent discussion of Insular art, of the Anglo-Saxon library, and of the transmission of the Vulgate had to be revised accordingly.

Michael Gorman’s long article “The Codex Amiatinus: A Guide to the Legends and Bibliography” (Studi Medievali, 2003) is indispensable here, both for its account of the discovery and for its patient dismantling of the garbled versions of the story that circulate in secondary literature. The episode is worth dwelling on as a methodological parable. The attribution to Italy had rested on genuine and reasonable evidence: the script, the pictures, the text-type, the provenance were all consistent with it. What overturned it was not a new theory but the collision of physical inspection with a neglected text. The erasures had been visible on the page for a thousand years; the Vita Ceolfridi had been in print. The discovery consisted in bringing the two together.

Set beside the page, the mechanics of the substitution become transparent, and they are governed by a single constraint: the forger had to fill the space he had cleared, and in each case he overfilled it.

Original	Substituted	Letters
CORPVS	CENOBIVM	6 → 8
PETRI	SALVATORIS	5 → 10
CEOLFRIDVS ANGLORVM	PETRVS LANGOBARDORVM	19 → 20

Bandini’s observation about Salvatoris is exactly borne out: five letters replaced by ten, the overflow running onto unerased parchment. And because every substitution is metrically longer than what it displaced, the hexameters collapse. Anglorum and Langobardorum are not interchangeable in a dactylic line, and the alteration destroyed the verse it was meant to preserve. The forger was concerned with the sense of the dedication, not with its prosody.

Two smaller discrepancies between the manuscript and the Vita’s transcription deserve note and should be checked against the critical edition of Grocock and Wood (Abbots of Wearmouth and Jarrow, Oxford Medieval Texts, 2013). The manuscript reads EXTREMIS where the Vita has extimis; and the second verse appears in the manuscript as quem caput ecclesiae dedicat alta fides, whereas the Vita transmits dedicat ecclesiae quem caput alta fides. Neither difference affects the substance of the identification, and both may well be variants in the transmission of the Vita rather than evidence of further alteration; but a scholarly discussion should not silently harmonise them.

What was changed is as revealing as how. The original dedicated the book ad corpus Petri, to the body of Peter, that is to the physical relic in the shrine at Rome. The alteration redirects it ad cenobium Salvatoris, to a community of the Saviour. The shift is from relic to institution, from a specific tomb to a house, and it is the change that converted a Northumbrian gift to the apostolic see into the benefaction of a Lombard abbot to his own monastery, San Salvatore on Monte Amiata.

The retention of caput ecclesiae is the alteration’s undoing, and it is worth asking why the phrase was left standing. The likeliest explanation is economy: erasure is laborious and conspicuous, and the forger changed only what named the wrong people and the wrong place. He appears not to have registered that the epithet itself was Petrine. In consequence the page preserves, in a phrase he thought innocuous, the theological signature of the original dedication — the assertion, central to Wearmouth-Jarrow’s entire cultural programme, that Peter is head of the Church and that a monastery at the edge of the world belongs to the communion he heads.

The Sister Bibles: Fragments and What They Prove

Bede and the Vita Ceolfridi say there were three pandects. Two were placed in the churches of Wearmouth and Jarrow, and both were subsequently lost, presumably in the Viking depredations of the ninth century or in the long attrition of the English Reformation. They are not, however, entirely gone. A small number of leaves survive, and they carry an evidential weight out of all proportion to their bulk.

The Greenwell Leaf (British Library, Add MS 37777) was identified by the Reverend William Greenwell, who acquired it in Newcastle in the early 1880s. The Middleton Leaves (British Library, Add MS 45025) comprise eleven leaves containing portions of 3 and 4 Kings. The Bankes Leaf (British Library, Loan MS 81) came to light in 1982 at Kingston Lacy in Dorset, where it had been used to wrap Willoughby family documents relating to a sale of Dorset lands in 1585; it contains text from Ecclesiasticus. Lowe classified the group as CLA II.177.

Palaeographic analysis shows that the Greenwell and Middleton leaves are the work of one scribe, copying Kings, and the Bankes leaf of another, copying Ecclesiasticus. Their dimensions, as noted above, closely match those of the Amiatinus. Taken together the fragments do three things. They corroborate the written testimony that three matched pandects were produced, converting Bede’s statement from assertion into demonstrated fact. They extend the roster of identifiable Wearmouth-Jarrow scribes beyond the hands visible in the Amiatinus. And, in Vnouček’s analysis, they introduce the material distinction discussed earlier: the fragments are calf, the Amiatinus apparently not.

The circumstances of the Bankes leaf’s survival are worth a moment’s reflection. That a bifolium of one of the greatest books ever made in Anglo-Saxon England should have spent four centuries as a document wrapper in a Dorset country house is a reminder of how much of what we possess is accidental, and of how much has been lost without leaving even that trace.

Cassiodorus, Vivarium, and the Question of Models

Behind the Amiatinus stands, at some distance and through a thickening haze of scholarly dispute, the figure of Cassiodorus (c. 490–585), Roman senator, minister to Theoderic, and founder of the monastery of Vivarium in Calabria.

In his Institutiones — the standard edition remains that of R. A. B. Mynors (Oxford, 1937; second edition 1961) — Cassiodorus describes three biblical formats produced at Vivarium: a nine-volume Old Latin edition, the novem codices; a large single-volume pandect that he called the Codex Grandior (Institutiones I.14.2); and a smaller pandect written in minute script (I.12.3). None survives. The Codex Grandior in particular has become one of the ghosts that haunt early medieval book history.

The link to Northumbria is furnished by Bede himself, who refers in De tabernaculo and De templo to a pandect associated with Cassiodorus, evidently in the monastery’s possession. Read together with the Historia abbatum’s mention of an Old Latin pandect brought from Rome, the natural inference is that the Codex Grandior, or a Vivarium book very like it, had reached Wearmouth-Jarrow and was available to the makers of the Amiatinus. The inference is widely accepted, though it remains an inference: Bede does not say that the book he used was the Codex Grandior, and the identification depends on matching his descriptions against Cassiodorus’s.

Paul Meyvaert’s article “Bede, Cassiodorus, and the Codex Amiatinus,” published in Speculum in 1996, reoriented the whole discussion and remains the point of departure. Two of his conclusions are especially consequential. The first is that Bede did indeed possess and use Cassiodorus’s Codex Grandior, an Old Latin pandect, at Jarrow. The second, more surprising, is that Bede did not know the Institutiones and was unaware that Cassiodorus had founded Vivarium at all. If Meyvaert is right, the Northumbrians had the book without the framework: they inherited a Cassiodoran object and its contents, including apparently its prefatory apparatus, but not Cassiodorus’s own account of what he had made or why. This has direct implications for how the Amiatinus’s opening quire should be read, since it means that whatever intellectual structure that quire possesses cannot have been derived from Cassiodorus’s own programmatic statement of it.

Meyvaert’s further contention is that the diagrams and images of the Amiatinus reflect Cassiodoran models available at Wearmouth-Jarrow but were mediated, selected, and to a significant extent redesigned there, with Bede as the likely designing intelligence. The alternative position, associated above all with Karen Corsano’s article “The First Quire of the Codex Amiatinus and the Institutiones of Cassiodorus” (Scriptorium, 1987), argues for a closer and more direct dependence on Cassiodoran material. Meyvaert’s later work qualified Corsano’s conclusions; the debate is not settled, and any account that presents the Amiatinus quire as a simple copy of the Codex Grandior’s opening, or alternatively as a wholly Northumbrian invention, is flattening a genuine disagreement.

Two further contributions refine the picture. Steven Wander and others have shown that the Tabernacle and Ezra illustrations correspond more closely to Cassiodorus’s Latin Josephus, the Jewish Antiquities, than to the biblical text itself — a finding that both strengthens the Vivarium connection and complicates any simple derivation from a Bible. And Thomas O’Loughlin, writing in the Revue Bénédictine in 2014, demonstrated that the nine-volume arrangement displayed in the Ezra miniature’s bookcase follows Augustine’s canon list in De doctrina christiana II.8.13, which Cassiodorus had himself adopted. The Amiatinus’s opening quire is thus layered: an Augustinian scheme, transmitted through Cassiodorus, realised in a Northumbrian book.

The Prefatory Quire: An Argument in Images and Diagrams

The opening quire of the Codex Amiatinus is among the most intensively studied gatherings in medieval art history. It is also, inconveniently, disordered. The Roman-numeral foliation reflects a rearrangement of 1876, and there is no confidence that either the current or the previous sequence preserves the original. The offset analysis undertaken in 1999–2001 was designed in part to address this, and the question remains open. Any interpretation of the quire as a sequential argument must therefore carry a caveat about the sequence.

Its contents are as follows. The dedicatory verses stand under an arch (fol. I/1v). A bifolium opening presents a diagram of the Tabernacle of Moses (ff. II/2v–III/7r). A purple page carries a prologue to the Old and New Testaments beneath arches (fol. IV/3r), and its verso a purple contents page listing the books of the Bible (fol. IV/3v). There follows the framed portrait of a seated scribe, conventionally identified as Ezra (fol. V/4r). A cruciform diagram sets out excerpts from Jerome’s Epistle 53 in interlinked medallions (fol. VII/6v). Three further diagrams present schematic divisions of Scripture: the division associated with Hilary and Epiphanius, that of Augustine’s De doctrina christiana, and that of Jerome and the Hebrew canon (fols. VI/5r, VII/6r, and 8r). At the head of the New Testament, far into the volume, stands the Maiestas Domini (fol. 796v), followed by seven pages of canon tables (ff. 798r–801r).

The Restorer of Scripture

The most celebrated page shows a venerable figure seated on a low stool, writing in a large open codex on his knees. He wears the jewelled breastplate and dress of a Jewish high priest. Behind him stands an open armarium, a book-cupboard, whose shelves display nine bound volumes with their covers visible — among the earliest surviving depictions in Western art of both a bookcase and a set of bookbindings. The volumes are shown recumbent, laid flat on their shelves, not ranged upright as in later libraries; the fore-edge of the lowest volume is presented to the viewer, and the covers of the others are seen face-on. Their tooled red bindings are the earliest visual evidence we possess for what a decorated codex cover of the late antique or early medieval Mediterranean looked like. That the cupboard is given a pediment, converting it into a shrine or aedicula, makes an argument about the status of its contents: Scripture is housed as a cult object is housed.

The number nine is not decorative. As written before, Thomas O’Loughlin demonstrated (Revue Bénédictine, 2014) that the nine-volume arrangement follows the canon list set out by Augustine in De doctrina christiana II.8.13, a scheme that Cassiodorus subsequently adopted. The literature records that the volumes carry inscribed titles corresponding to these divisions — the Octateuch and books of Law, Kings and Chronicles, the historical books, the Psalms, the Solomonic books, the Prophets, the four Gospels, the Apostolic Epistles, and Acts with the Apocalypse. These inscriptions are not legible at the resolution of the present reproduction and should be verified against the facsimile or the original before being quoted.

The image therefore does something conceptually unusual. It depicts, inside a Bible, the library of which that Bible is the compression: nine volumes on shelves, and the single pandect in the reader’s hands that contains all of them. The relationship between the two formats is the visual argument of the page.

Above the frame runs a hexameter couplet:

Codicibus sacris hostili clade perustis Esdra Deo fervens hoc reparavit opus

“When the sacred books had been burnt in the enemy’s destruction, Ezra, burning with zeal for God, restored this work.”

The allusion is to the legend, developed in 4 Esdras 14, according to which Ezra restored the Scriptures destroyed at the fall of Jerusalem. The volumes in the armarium are labelled with the divisions of Scripture: OCT[ateuchus] LIB. LEG.; REG. PAR. L. VI; HIST. LIB. VIII; PSAL. LIB. I; SAL. LIB. V; PROP. L. XVI; EVANGL. L. IIII; EPIST. AP. XXI; ACT. AP. APOC. — the nine-codex scheme identified by O’Loughlin as Augustinian by way of Cassiodorus.

The identification of the figure is contested, and the dispute is more than antiquarian. Rupert Bruce-Mitford, whose Jarrow Lecture of 1967 on the art of the manuscript remains foundational, held that the image essentially reproduced an author-portrait of Cassiodorus taken over from the Codex Grandior: the nine volumes on the shelves are, on this reading, the novem codices of Vivarium, and the seated scholar is their maker, relabelled as Ezra when the image was copied.

Meyvaert dissented sharply in a second major article, “The Date of Bede’s In Ezram and his Image of Ezra in the Codex Amiatinus” (Speculum, 2005). He found implausible the notion that the aged senator Cassiodorus would have had himself portrayed as a working scribe, seated at labour with a pen. He argued instead that the figure was conceived and designed as Ezra at Wearmouth-Jarrow; that the miniature was newly composed rather than transcribed; that Bede was closely involved; and that Bede probably wrote the couplet himself. He connected the image to Bede’s commentary In Ezram et Neemiam, in which the restoration of Scripture after destruction carries a rich typological charge.

The field has not simply adopted either position. Lawrence Nees, among others, has found Meyvaert’s positive reconstruction only partly persuasive while agreeing that the Cassiodorus self-portrait hypothesis does not convince. Celia Chazelle, in an article of 2003 in Early Medieval Europe on the first quire and the evidence of the manuscript’s Roman destination, read the Ezra portrait as functioning in relation to the papal office: an image of the restorer and custodian of Scripture, presented to the pope who was to receive the book. This reading has the merit of tying the picture to the demonstrable purpose of the volume, and it exemplifies the shift in scholarship from asking what the Amiatinus copies to asking what it argues.

The honest summary is that we do not know who the figure was originally meant to be, that the balance of recent specialist opinion favours a deliberate Ezra designed in Northumbria rather than a relabelled Cassiodorus, and that both positions depend on reconstructing a lost book.

The Camp of Israel

The Tabernacle diagram spreads across a bifolium and presents the Mosaic sanctuary in plan, with its courts, furnishings, and the disposition of the tribes. Its interest lies partly in its relationship to Bede’s own exegetical writing: De tabernaculo is precisely a commentary on the meaning of the sanctuary, and the coincidence of a major Bedan treatise with a major Amiatinus image is one of the strongest circumstantial arguments for his involvement in the book. As noted, Wander’s work indicates that the diagram’s details track Cassiodorus’s Josephus rather than Exodus, which places it in the Vivarium orbit.

The three schematic divisions of Scripture — after Hilary and Epiphanius, after Augustine, after Jerome — do something unusual and characteristically Northumbrian. Rather than presenting one canon as authoritative, they set out three competing ways of organising the biblical books side by side. The effect is to display the structure of Scripture as itself an object of learning, and to place the reader in a scholarly rather than merely devotional relation to the volume in hand. Carol Farr’s study of the diagram pages (1999) reads them as evidence for the shape of learning at Wearmouth-Jarrow; Jennifer O’Reilly’s work on the library of Scripture, and Chazelle’s on the biblical diagrams and the vision of God, develop the theological implications; Nees has considered them within the broader problem of form and function in early medieval illustrated Bibles.

The Maiestas Domini at the head of the New Testament shows an enthroned figure flanked by two angels and surrounded by the four evangelists with their symbols. Peter Darby’s article of 2017 in Speculum analyses the page in relation to Jerome’s Gospel prefaces, arguing that its composition fuses Old Testament prophetic vision with New Testament fulfilment and that its details answer to Jerome’s prefatory statements about the evangelists. The reading exemplifies the current tendency to treat the Amiatinus images as exegetically motivated rather than decoratively or merely reproductively conceived.

A detailed analysis

The right-hand leaf carries two numerals in its upper corner, an Arabic 7 and a Roman III, and the diagram is variously cited as ff. 2v–3r and as ff. II/2v–III/7r. The discrepancy is not trivial. The two systems belong to different reconstructions of a quire whose original order was disturbed, refoliated in 1876, and remains unresolved; in a regular quaternion, folios 2 and 7 are the conjoint halves of a single bifolium, which is not the same thing as a facing opening. Anyone citing this image should give both numerals and should treat the physical relationship of the drawing to its quire as an open question, to be pursued through Chazelle’s monograph of 2019 and through the offset analysis undertaken on the manuscript between 1999 and 2001.

What is beyond doubt is that the composition was planned as a single field spanning the gutter. The green ground, the surrounding arcade, and the framing inscriptions all continue across the join without interruption. Whoever designed this page was working to a double-page format, which places a considerable demand on planning and on the accurate registration of two separately handled surfaces.

The most immediately arresting feature of the page is that it is oriented by four Greek terms, written in Latin letters in large display capitals along the four sides of the green field:

• ARCTOS (Ἄρκτος, the Bear) — north, at the top
• DYSIS (δύσις, setting) — west, at the left
• ANATOL[E] (ἀνατολή, rising) — east, at the right
• MESEMBRIA (μεσημβρία, midday) — south, at the foot

The choice is significant on two counts. First, it establishes the orientation of the whole plan: the sanctuary occupies the western half of the enclosure and its entrance faces east, exactly as prescribed in Exodus, with the altar of burnt offering standing in the court before it. Second, and more consequentially for the history of the page, the use of Greek cardinal terminology in a Latin book made in Northumbria is one of the clearest indications available that the diagram descends from a Mediterranean model rather than having been devised locally. No Northumbrian would reach for ἀνατολή to label the east side of a drawing unless he were following something. This is an inference rather than a demonstration, but it is a well-founded one, and it converges with the other evidence for the Cassiodoran affiliations of the opening quire.

The green field represents the court of the Tabernacle. It is bounded on all four sides by an arcade of columns with moulded capitals and bases, between which runs a red-orange field strewn with lozenges. This is the enclosure of Exodus 27:9–18, the hangings of fine twined linen suspended from pillars. On the east side, at the right, one bay of the arcade is filled with a magenta and green panel distinct in colour from the rest: the screen of the gate of the court (Exodus 27:16), specified as a hanging of blue, purple, and scarlet, and here marked out by its difference of colour and its position.

Within the court, at the left, stands the Tabernacle proper, drawn as a rectangle of concentric coloured bands: an outer band of black hatching, then orange-red, then white, then magenta, enclosing a gold-tan band and a green interior. The bands are most economically explained as the four coverings prescribed in Exodus 26 — the fine linen with cherubim, the curtains of goats’ hair, the rams’ skins dyed red, and the outermost covering of skins — presented in plan rather than in elevation, with the gilded boards of the frame as the inner band. This reading is the standard one and is consistent with the text, though a plan cannot render a set of superimposed coverings unambiguously and the correspondence should not be pressed into false precision.

The interior is divided by a vertical partition, the veil, and its two chambers are labelled. In the western chamber, SCA SCORVM (sancta sanctorum, the Holy of Holies) contains the ARCA TEST[AMENTI], the Ark of the Covenant, drawn as a gabled chest. In the eastern chamber stand the ALTAR[E] THIM[IAMATIS], the altar of incense, the MENSA, the table of the shewbread, and the CAND[ELABRVM], the seven-branched lampstand, this last rendered as a golden arc of lamps.

Outside the sanctuary, to the east, two further objects are labelled in vertical lettering. The ALTARE HOLOCAVSTI, the altar of burnt offering, is drawn as a dark grated structure from which red flames rise. Below it, the LABRVM, the bronze laver, appears as a large vessel with scrolled volute handles. Between these, in red, stands the inscription MOSES ET AARON, marking the station assigned to Moses, Aaron, and his sons on the east side of the tabernacle in Numbers 3:38.

The numbers: a verifiable diagram

What distinguishes this page from a merely illustrative plan is that it is a data diagram. Around the arcade, outside the court, the twelve tribes are named with their censused strengths, and within the court the three Levitical clans are named with theirs. Every one of these figures can be checked against the biblical text, and every one of them is correct.

The three Levitical clans appear inside the enclosure, each on the side where Numbers 3 stations it:

Inscription	Reading	Numbers 3	Position
FILII MERARI VI CC	6,200	Merari, 6,200	north
FILII GERSON VII D	7,500	Gershon, 7,500	west
FILII CAATH VIII DC	8,600	Kohath, 8,600	south

The twelve tribes are arranged outside the court in the four camps of Numbers 2, each with the figure given in the census of Numbers 1:

Side	Inscriptions	Readings	Numbers 1
North (top)	ASER XLI D; DAN LXII DC; NEPHTHALIM LIII CCCC	41,500; 62,700; 53,400	Asher 41,500; Dan 62,700; Naphtali 53,400
South (foot)	GAD XLV DCL; RUBEN XLVI D; SIMEON LVIIII CCC	45,650; 46,500; 59,300	Gad 45,650; Reuben 46,500; Simeon 59,300
West (left)	BENIAMIN XXXV CCCC; EFRAIM XL D; MANASSE XXXII CC	35,400; 40,500; 32,200	Benjamin 35,400; Ephraim 40,500; Manasseh 32,200
East (right)	ISSACHAR LIIII CCCC; IUDAS LXXIIII DC; ZABULON LVII CCCC	54,400; 74,600; 57,400	Issachar 54,400; Judah 74,600; Zebulun 57,400

The agreement is complete, in the figures and in the tribal disposition alike. Two observations follow.

The first is notational. The Roman numerals must be read with the leading element denoting thousands: XLI D yields not forty-one plus five hundred but 41,500. The convention is confirmed by the correspondence with Numbers across all twelve entries, and it is worth noting because it is the kind of detail that a careless copyist would have destroyed. The abbreviation preceding each figure, a barred N, is most plausibly n(umero).

The second is evidential. Whoever transmitted this diagram to Wearmouth-Jarrow, and whoever executed it there, preserved a substantial body of numerical data without corruption across at least one act of copying and probably several. Numbers are the most fragile element of any manuscript tradition. Their integrity here is a small but genuine index of the quality of the exemplar and of the care of the Northumbrian scriptorium.

The correspondence with Numbers is close, but it does not exhaust the question of sources. Steven Wander has argued that details of the Tabernacle and Ezra images in the Amiatinus answer more closely to Cassiodorus’s Latin version of Josephus, the Jewish Antiquities, than to the biblical text itself. The two observations are compatible: the numerical and positional armature is biblical, while certain features of the furnishings and their rendering may descend from a Josephan intermediary. Distinguishing the two strands requires detailed comparison against the Latin Josephus and is not something a photograph can settle.

Behind this stands the larger question of the Codex Grandior, the great pandect made at Cassiodorus’s monastery of Vivarium in Calabria and generally thought to have reached Wearmouth-Jarrow. Bede refers in De tabernaculo and De templo to a pandect associated with Cassiodorus, and the Historia abbatum records that Ceolfrith possessed an Old Latin pandect brought from Rome. The standard hypothesis is that the prefatory diagrams of the Amiatinus derive, directly or at some remove, from that book. The hypothesis cannot be proved, because the Codex Grandior is lost; every argument about what the Amiatinus copies is an argument about a book no living scholar has seen.

Paul Meyvaert’s reconstruction of 1996 concluded that Bede used the Cassiodoran pandect but did not know the Institutiones and was unaware that Cassiodorus had founded Vivarium. If so, the Northumbrians inherited the object without its author’s account of it, and the intellectual organisation of the opening quire was reconstructed, or constructed anew, at Jarrow. Karen Corsano argued in 1987 for a closer dependence on Cassiodoran material; Meyvaert’s later work qualified her conclusions. The debate remains open, and any presentation of this page as a straightforward copy, or as a Northumbrian invention, misrepresents the state of the question.

The coincidence between this diagram and Bede’s De tabernaculo is the strongest circumstantial argument for his involvement in the design of the first quire. De tabernaculo is precisely a commentary on the meaning of the Mosaic sanctuary, read allegorically as a figure of the Church; and here, at the front of the greatest Bible his monastery produced, stands a full double-page plan of that sanctuary with its tribes numbered and stationed. The inference is attractive and widely accepted, but it remains an inference: no source names Bede in connection with the Amiatinus.

What the page does, in itself, is convert exegesis into layout. The Tabernacle is not depicted as a scene or a narrative but as a plan with data attached, a structure the eye can enter and traverse. The viewer is positioned above the camp of Israel, able to see at once where each tribe lies, how many men it musters, which Levitical clan guards which side, and what furniture stands in which chamber. This is the visual logic that governs the other diagrams of the first quire, where the divisions of the biblical canon after Hilary and Epiphanius, after Augustine, and after Jerome are set out side by side as competing schemes rather than as a single authority. Carol Farr’s study of the diagram pages (1999) reads them as evidence for the shape of learning at Wearmouth-Jarrow; Jennifer O’Reilly and Celia Chazelle have developed their theological implications; Lawrence Nees has situated them within the wider problem of form and function in early medieval illustrated Bibles.

The common element is that these pages address a reader who is expected to study rather than merely to venerate. A Bible that opens with a measured plan of the Tabernacle and three rival tabulations of the canon is a Bible that presumes a scholarly community. That presumption, as much as the uncial script or the Roman layout of the text, is what the Amiatinus asserts about the house that made it.

The parchment is cockled and stained, particularly along the outer margins and at the foot, and the green ground has darkened and abraded unevenly, with losses visible in the arcade and around the sanctuary. Pricking holes survive in the margins, evidence of the ruling procedure. As with every other leaf of this manuscript, statements about pigments should rest on the published non-destructive analysis rather than on visual assessment of a reproduction.

The Hinge of the Pandect

Of the three major paintings in the Codex Amiatinus, this is the only one that does not belong to the prefatory quire. The Ezra miniature and the Tabernacle bifolium stand at the front of the volume, among the diagrams and canon divisions that constitute its scholarly apparatus. The Maiestas Domini stands some eight hundred folios into the book, at the head of the New Testament, and is followed by seven pages of canon tables.

This placement is the first thing to be said about the image, because it determines everything else. The Amiatinus is a pandect, a complete Bible in a single volume, a format so rare in its period as to constitute an argument in itself. In such a book the junction between the Testaments is the principal structural event, and it is precisely there that this painting has been placed. The reader who has traversed the whole of the Old Testament arrives at a full-page theophany before encountering the Gospels. The image is the hinge on which the volume turns, and its subject is the figure in whom, on the exegetical reading the manuscript everywhere presupposes, the two Testaments are joined.

The painted field is bounded by a rectangular frame of orange-red set with small dark and gilded lozenges, a pattern used elsewhere in the manuscript and derived from the vocabulary of inlaid and jewelled work rather than from any Insular ornamental repertoire.

Within, the ground is composed of horizontal bands of pale green, salmon-pink, and green, resolving at the foot into a broad band of ochre. This striated treatment is the late antique convention for landscape and sky, a compressed atmospheric recession familiar from Mediterranean wall painting and mosaic, and it establishes the register of the page immediately: nothing here derives from the ornamental idiom of the Lindisfarne Gospels or the Book of Durrow.

At the centre hangs a great disc, occupying rather more than half the height of the field. It is built of concentric rings, reading from the circumference inward: an olive-gold band studded with lozenges; a broad red band similarly studded; a thin yellow ring; a band of alternating rectangular blocks in red, green, salmon, and teal; a second thin yellow ring; and then the interior, a field of deep blue-green strewn with white points, shading through concentric zones toward a paler green at the base.

Within this disc Christ sits enthroned. He is bearded, frontal, and haloed in gold; the halo’s design cannot be determined with confidence at this resolution, and whether it carries a cross should be verified against the original. He wears an orange-red tunic beneath a purple mantle that falls in heavy folds across his lap. In his left hand he holds a closed codex with a gilded cover; his right is raised, fingers extended, in the gesture that late antique art uses indifferently for speech and for blessing. His bare feet rest on a golden footstool.

Two angels flank him within the disc, robed in gold-ochre, winged, haloed, and turned inward toward the throne, each holding a slender vertical object.

At the four corners of the rectangular field stand four haloed men, barefoot on the ground band, each holding a book. Between and beside them appear four winged creatures: at the top, a winged human figure in flight toward the upper left, and an eagle toward the upper right; at the foot, a winged lion at the lower left and a dark winged quadruped, the ox, at the lower right.

The four and the four

The pairing follows the Western identification established by Jerome and standard in Latin tradition thereafter: Matthew the man, Mark the lion, Luke the calf or ox, John the eagle. The living creatures themselves derive from the visions of Ezekiel 1 and Revelation 4, where four beings of these forms surround the divine throne; their application to the four evangelists is exegetical rather than scriptural, and Jerome’s formulation of it in his prefatory writings on the Gospels is the channel through which it entered Latin bookmaking.

The distinctive feature of this page is that the evangelists and their symbols appear as separate figures. In much later medieval practice one finds either the four symbols alone, or evangelist portraits each accompanied by its own creature within a single frame. Here the four men stand as full-length human figures at the corners, and the four creatures occupy the intervals between them, flying inward toward the disc. The arrangement doubles the witness: the evangelists as historical authors holding their books, and the creatures as the visionary forms under which the prophets beheld them.

This is the point at which Peter Darby’s study of the page becomes indispensable. In an article of 2017 in Speculum, Darby analysed the Maiestas Domini in relation to Jerome’s Gospel prefaces, arguing that the composition effects a fusion of Old Testament prophecy and New Testament fulfilment, and that its details answer to Jerome’s prefatory statements about the evangelists. The reading exemplifies the direction of recent scholarship on the Amiatinus, which has moved from asking what the images copy toward asking what they argue.

The disc: clipeus, firmament, and rainbow

The circular medallion carries at least three overlapping associations, and it is not necessary to choose among them.

Formally, it belongs to the tradition of the imago clipeata, the portrait set within a shield or roundel, a device of Roman commemorative and imperial art that Christian usage adapted for images of Christ and the saints. The heavy jewelled rings that constitute the disc’s outer zones are consistent with this reading: they present the enthroned figure as though set in a precious frame or medallion.

Cosmologically, the interior reads as the heavens. The deep blue-green field strewn with white points is the starry firmament, and the concentric shading suggests the layered spheres of heaven. Christ is enthroned not in a landscape but above and within the cosmos, which is the standard theological content of the Maiestas type.

A third possibility deserves mention as a possibility rather than a conclusion. The band of alternating red, green, salmon, and teal blocks may be intended as the rainbow that Revelation 4:3 places about the throne, described there as emerald in appearance. The predominance of green in the band, and the fact that the same chapter of Revelation supplies the four living creatures who appear at the corners of this very page, make the suggestion attractive. It should nonetheless be marked as an interpretation offered here rather than a reading established in the literature, and anyone wishing to use it should test it against Darby and against the wider iconographic scholarship on the Maiestas.

The two flanking angels are a separate problem. They are not the living creatures, which appear outside the disc, and they are not obviously the seraphim of Isaiah 6, who are described as six-winged. Their identity within the composition is a legitimate question and one this analysis does not attempt to settle. It is precisely the kind of detail on which the relationship of the Amiatinus to its lost models might be tested.

Style, hand, and the question of execution

A comparison between this page and the Ezra miniature is instructive, and it can be made from reproductions without hazarding claims that require the original.

The Ezra painting attempts modelling: the drapery is shaded, the seated body has volume, the furniture is foreshortened, and there is an effort at cast shadow. The Maiestas is markedly more linear. Drapery here is indicated by parallel striations of contrasting colour rather than by modelled transitions; the figures are elongated and flattened; the evangelists’ feet sit awkwardly on the ground band; and the angels within the disc are drawn with a confident but essentially graphic economy.

The difference is real and observable. What it means is less certain. It may indicate different hands at work on the manuscript’s paintings; it may indicate different models, of different date and quality, behind the two images; it may indicate the same painter working in different registers for different purposes. The literature on the number of painters is less developed than that on the number of scribes, and any conclusion drawn here should be checked against the specialist studies, beginning with Bruce-Mitford’s Jarrow Lecture of 1967 and Chazelle’s monograph of 2019.

What is not seriously in dispute is that the models were Mediterranean and late antique. The striated sky, the clipeus, the enthroned frontal Christ, the barefoot standing apostolic figures, the jewelled frame: none of this has an Insular source. The question that has divided scholarship is whether Mediterranean painters executed the pages in Northumbria, as Per Jonas Nordhagen proposed in his Jarrow Lecture of 1977, or whether Northumbrian artists copied imported models. The prevailing view now favours the latter, on evidence that is material rather than stylistic: the identification of orpiment, an arsenic sulphide yellow characteristic of Insular manuscript production rather than of Italian workshops, indicates that the brush was held by someone working with the materials and habits of a Northumbrian scriptorium. Nothing about pigments should be asserted from a photograph, and the published non-destructive analysis conducted on the manuscript between 1999 and 2001 is the proper authority.

The image and the argument of the book

Placed where it is, this painting performs a function that the two prefatory images cannot.

The opening quire addresses a reader who is about to study: it supplies a plan of the Tabernacle with the tribes numbered and stationed, three competing tabulations of the biblical canon, and a portrait of the restorer of Scripture at work with a library open behind him. Its mode is diagrammatic and scholarly. The Maiestas addresses the same reader at the moment of transition, and its mode is visionary. Where the Tabernacle diagram sets out the sanctuary of the old covenant as measured data, this page presents the enthroned Lord of the new, surrounded by the creatures under whose forms the prophets saw him and by the four men who wrote him down.

The pairing is not accidental in a volume of this ambition. Bede’s own exegetical writing, in De tabernaculo and De templo, is built precisely on the reading of the old sanctuary as a figure of the Church; and the interpretation of the four living creatures as the four evangelists is the standard Latin means of binding the prophetic vision to the Gospel text. Whether Bede designed the Amiatinus programme is a question that cannot be settled — no source names him in connection with the manuscript, and the attribution remains a well-supported inference rather than a documented fact. But the intellectual world in which these images make sense is unmistakably his.

The leaf is generally well preserved, with the painted surface intact and the colours strong, though there is abrasion along the lower edge of the disc and around the frame at the upper right, and a repaired area at the top of the leaf. The outer edge carries a pronounced semicircular indentation, which may be an original flaw in the skin worked around by the parchmenter or a later loss; direct examination would be needed to distinguish these. Faint traces of writing are visible in the outer margin, most plausibly show-through or offset rather than deliberate annotation. Offsets of this kind are not merely damage: pigment and ink impressions transferred between leaves were among the phenomena analysed in the 1999–2001 campaign, precisely because they carry information about the manuscript’s structure.

Who Painted the Amiatinus? Style, Pigment, and the Limits of Connoisseurship

That the miniatures descend from late-antique Mediterranean models is not in serious dispute. What is disputed is the hand that executed them. Two positions have divided the literature.

The first, argued most forcefully by Per Jonas Nordhagen in his Jarrow Lecture of 1977 on the Amiatinus and the Byzantine element in the Northumbrian renaissance, holds that an Italo-Byzantine painter was at work in Northumbria. The proposal has an obvious attraction. Biscop demonstrably imported craftsmen from the Continent for building and glazing; the Amiatinus miniatures display modelling, foreshortening, and an approach to drapery that has no Insular parallel; and the simplest explanation of a Mediterranean picture is a Mediterranean painter.

The second position, now generally preferred, holds that Northumbrian artists copied imported models. The decisive evidence is material rather than stylistic. Analysis of the manuscript’s palette has identified orpiment, a bright yellow arsenic sulphide, in the painting. Orpiment is characteristic of Insular manuscript production and is not the pigment an Italian workshop of the period would have reached for. Its presence indicates that whoever wielded the brush was working with the materials and habits of a Northumbrian scriptorium. The argument has been pressed against Nordhagen in the review literature surrounding the volume Imaging the Early Medieval Bible, and it is difficult to answer.

Supporting this, research on pigments in the earliest Northumbrian manuscripts, conducted at Durham, has observed that the Ezra image — with its attempt at modelling, shading, and the suggestion of volume — represents a rare and revealing case of a Northumbrian artist reaching for semi-naturalistic painting. The hesitations and inconsistencies of the result read less like the work of a trained Mediterranean painter than like a highly skilled copyist negotiating an unfamiliar visual grammar.

The episode has a wider methodological moral. Style, in the absence of corroboration, misled scholarship about the Amiatinus for two centuries: the pictures looked Italian, and so the book was Italian. Style is now again the principal ground on which the painters’ origin is argued, and again it is a material rather than a formal analysis that adjudicates. The Amiatinus is a standing argument for the integration of scientific examination into art-historical method.

The Text: The Amiatinus as Witness to Jerome

If the images have attracted art historians, the text has for longer attracted philologists, and it is as a textual witness that the Amiatinus first became internationally famous.

The manuscript is the earliest surviving complete Latin Vulgate; only the incomplete León palimpsest is older, and it is not complete. Under the siglum A (or am) the Amiatinus is a primary witness in every major critical enterprise: Tischendorf, who published its New Testament in 1850 and again in 1854; the great Oxford edition of Wordsworth and White; the Benedictine (Roman) Vulgate; and the Stuttgart Biblia Sacra iuxta Vulgatam versionem. Hugh Houghton’s The Latin New Testament (Oxford, 2016) catalogues it among the foremost witnesses.

Its text is generally reckoned an exceptionally pure representative of Jerome’s version, and its affinities point to Italy, and specifically to southern Italy. The likeliest explanation is the books Biscop brought from Rome, though the possibility that exemplars came by way of Hadrian, the abbot from near Naples who accompanied Theodore of Tarsus to Canterbury, has also been canvassed. Richard Marsden’s The Text of the Old Testament in Anglo-Saxon England (Cambridge, 1995), especially its long treatment of the Ceolfrith Bibles, is the fundamental study of these relationships and of the editorial work that lay behind the Northumbrian text.

The quality is not uniform across the volume, and this is itself informative about the making of the book. Samuel Berger noted the inferiority of Wisdom and Ecclesiasticus; Tischendorf that of Maccabees. Ezra–Nehemiah, Samuel, Kings, and Chronicles are each presented as a single book rather than divided. Baruch is absent, following Jerome’s own exclusion. The picture is of a text assembled from exemplars of differing quality, edited with care where care was possible, and reproduced faithfully where it was not.

A correction: the Psalter

One point requires explicit correction, since it is misstated in a great deal of general literature. The Amiatinus Psalter is not the Gallican version. It is Jerome’s third rendering, the Psalterium iuxta Hebraeos, translated from the Hebrew. This distinguishes the manuscript from the overwhelming majority of later Vulgate Bibles, in which the Gallican psalter displaced the Hebraicum from the ninth century onward, and from the Roman psalter that was then standard in English liturgical use.

The Amiatinus is not, however, a first-rate witness to the iuxta Hebraeos; its psalms are commonly judged inferior. A further detail complicates the picture attractively: the manuscript carries the so-called “Columba” series of psalm headings, also found in the Cathach of St Columba, which indicates that an Irish psalter lay somewhere behind the Amiatinus text, even though its readings diverge in many places from the best Irish witnesses. Here, in the most Roman of Insular books, is an Irish thread. Pierre-Maurice Bogaert’s chapter on the Latin Bible c. 600–900 in The New Cambridge History of the Bible (2013) sets out the three psalter families and the Amiatinus’s position among them.

Later textual use

The manuscript’s authority was recognised long before modern scholarship. In 1587 it was carried to Rome for consultation during the Sixtine revision of the Vulgate under Pope Sixtus V. After the pope’s death on 25 August 1590, a prohibition on its sale was issued on 5 September of the same year, and it was returned to Monte Amiata. The episode is a striking illustration of the codex’s standing: a book then believed to be a sixth-century Italian Bible was treated as an authority against which the printed Vulgate might be corrected. That it was in fact Northumbrian did not diminish the soundness of the instinct, since its text is indeed among the closest to Jerome’s that survives.

Monte Amiata: A Long Silence and a Second Identity

Between Ceolfrith’s death at Langres in September 716 and the manuscript’s appearance at the abbey of San Salvatore on Monte Amiata in Tuscany, the record is essentially blank. This is the largest single gap in the codex’s history and it has not been closed.

What can be said is limited. The altered dedication, naming the Lombard abbot Peter and rededicating the volume to the Saviour, is the earliest physical evidence of the manuscript’s Italian second life, and it establishes that the change was made by or for a community of the Saviour, which San Salvatore was. The earliest documentary attestation is an inventory of the abbey’s relics dated 1036, which describes an Old and New Testament written, it was believed, by the hand of the blessed Pope Gregory. The codex had, in other words, become a relic of Gregory the Great — a misattribution that is itself eloquent about how such objects were understood, and which incidentally guaranteed its preservation. In the eleventh century Bonizo of Sutri cited the modified dedication. In 1462 the manuscript was examined by Aeneas Silvius Piccolomini, by then Pope Pius II.

How it reached Monte Amiata is unknown. Both Langres and San Salvatore lay on or near the pilgrimage routes toward Rome, the Via Francigena and its feeders, and it is easy to construct plausible narratives: that the book reached Rome and was later given or removed to Tuscany; that Ceolfrith’s companions or their successors disposed of it en route; that it passed through Lombard hands. None of these is documented. The most that responsible scholarship can say is that the manuscript travelled south with Ceolfrith’s party in 716, that it is attested at San Salvatore by the eleventh century, and that the intervening three centuries are dark. Whether the gift ever reached the shrine of St Peter for which it was made is, on present evidence, unknowable.

The abbey of San Salvatore was suppressed in the reforms of Grand Duke Peter Leopold of Tuscany, and the manuscript passed to the Biblioteca Medicea Laurenziana in Florence. The date is given variously in the literature as 1782, 1785, and 1786; the World Digital Library catalogue specifies 1782, explicitly in connection with Peter Leopold’s suppression of the religious orders, while other specialist sources give the later years. The discrepancy is minor but should be reported as a range rather than resolved arbitrarily. At Florence, Bandini catalogued the codex and noticed the erasures whose significance would not be understood for another century.

In the modern era the manuscript has been made progressively more accessible. A complete facsimile on CD-ROM appeared as La Bibbia Amiatina: Riproduzione integrale su CD-ROM del manoscritto Firenze, Biblioteca Medicea Laurenziana, Amiatino 1, edited by Luigi G. G. Ricci, Lucia Castaldi, and Rosanna Minello (Florence, 2000). A printed facsimile followed from La Meta Editore in 2003. The Laurenziana’s own digital repository provides reproductions, and the catalogue description in Alidori and others, Bibbie miniate della Biblioteca Medicea Laurenziana di Firenze (2003), remains a standard point of reference.

In the autumn of 2018 the Codex Amiatinus returned to England. It was lent by the Laurenziana to the British Library for the exhibition Anglo-Saxon Kingdoms: Art, Word, War, which ran from 19 October 2018 to 19 February 2019. The British Library described the loan as the manuscript’s first return in over 1,300 years; press coverage, following the Guardian, gave the interval more precisely as 1,302 years. It was displayed alongside the St Cuthbert Gospel, itself a Wearmouth-Jarrow product and the earliest intact European book, together with the Lindisfarne Gospels, the Book of Durrow, the Echternach Gospels, and the Durham Cassiodorus, and near surviving leaves of one of its own sister pandects. The juxtaposition made visible, in a single room, the two visual cultures of Northumbrian Christianity and the material fact of Ceolfrith’s threefold commission.

Romanitas, the Pandect, and the Historiographical Lesson

Three broader questions remain, and each has generated substantial literature.

The meaning of Romanitas

The Amiatinus is the fullest material expression of Wearmouth-Jarrow’s alignment with Rome. That alignment operated in script, layout, art, text, and destination, and it was not incidental but constitutive. Jennifer O’Reilly’s article “‘All that Peter stands for’: The Romanitas of the Codex Amiatinus Reconsidered” (Proceedings of the British Academy, 2009) has done more than any other single study to show that this was a theological rather than merely a stylistic commitment. On her reading the codex articulates a claim about Peter, about the unity of the universal Church, and about the place of a monastery at the world’s edge within that unity. The gift is an argument as much as an offering: it asserts that Northumbria belongs to the Roman communion not as a periphery but as a full participant, capable of producing for the apostle’s tomb a book better than anything then available in Rome.

This should not be read as a repudiation of Insular tradition. The Lindisfarne Gospels were made in the same kingdom and within a few decades, and the Amiatinus itself carries, in its psalm headings, an Irish inheritance. What the contrast reveals is a culture with genuine stylistic choices available to it, in which the adoption of a Mediterranean idiom at Wearmouth-Jarrow was a decision with meaning precisely because alternatives existed.

The pandect as a format

Complete Bibles in one volume were rare in the early Middle Ages. The practical obstacles were formidable: the parchment, the scribal labour, the sheer physical unwieldiness of the result. Most communities owned Scripture in parts. Cassiodorus’s Codex Grandior was an exception, and Ceolfrith’s three pandects were another, on a scale that suggests deliberate emulation of the Vivarium precedent.

The format was revived in the Carolingian period, most famously at Tours under Alcuin and his successors, where one-volume Bibles were produced in quantity. It is tempting to draw a line from Wearmouth-Jarrow to Tours, particularly given Alcuin’s Northumbrian formation at York. The temptation should be resisted, or at least heavily qualified. The parallel between the two enterprises is one of type and ambition; a demonstrable textual or codicological descent from Ceolfrith’s pandects to Alcuin’s has not been established, and asserting one goes beyond the evidence. What can be said is that the Amiatinus stands near the beginning of the tradition of the complete one-volume Bible in the medieval West, and that it demonstrates the format’s feasibility a century before that tradition became general.

The historiographical lesson

Finally, the Amiatinus is a case study in the fallibility and the self-correcting power of scholarly method. For well over a century the book was securely misattributed by capable scholars using the best available criteria. The correction came from the convergence of physical examination — Bandini’s observation of the erasures — with textual recovery — Hort’s use of the Vita Ceolfridi — and from de Rossi’s willingness to entertain a conclusion that ran against the consensus. Isolated earlier dissent, such as Paul de Lagarde’s, had not been enough; the case required the two kinds of evidence together.

The lesson is not that stylistic attribution is worthless, which would be absurd, but that it is defeasible, and that the material and documentary record retains an authority that style cannot override. It is a lesson the discipline has had occasion to relearn, and one that the recent dispute over the manuscript’s parchment species illustrates in a new register: an assumption held for decades, and repeated in authoritative sources, has been challenged by direct examination of the object.

The State of the Question

Scholarship on the Codex Amiatinus has been unusually active over the past three decades, and the essay’s account should be read as a report on a moving field.

Celia Chazelle’s The Codex Amiatinus and its “Sister” Bibles: Scripture, Liturgy, and Art in the Milieu of the Venerable Bede (Leiden: Brill, Commentaria 10, 2019), running to over 650 pages, is the first comprehensive treatment of the manuscript and its lost companions and is now the indispensable synthesis. Meyvaert’s two great articles of 1996 and 2005 remain the fulcrum of debate about Cassiodorus and about Ezra. Marsden’s 1995 monograph governs the textual side, as Gorman’s 2003 bibliographic guide governs the historiographical. Darby’s 2017 article on the Maiestas Domini, Vnouček’s 2020 study of the parchment, and Cardwell’s 2025 work on the scribal milieu represent the most recent substantial contributions.

The principal open questions may be stated plainly. The original order of the first quire is unresolved, and with it the reading of the prefatory programme as a sequence. The identity of the seated scribe, and the degree of dependence of the Amiatinus images on the Codex Grandior, remain disputed between the positions of Bruce-Mitford, Corsano, and Meyvaert. Bede’s involvement is probable but undocumented. The species of the parchment is now contested and awaits biomolecular confirmation. The manuscript’s route from Langres to Monte Amiata is unknown and may remain so. Whether Ceolfrith’s gift ever lay at the tomb of St Peter is likewise unknown.

Conclusion

The Codex Amiatinus is a book made at the edge of the world to prove that its makers were not at the edge of anything. Its uncial, its sense-divided lines, its Mediterranean pictures, its Hieronymian text, and its destination all point the same way: toward Rome, and toward a claim of full participation in the Latin Christian inheritance. Ceolfrith commissioned three such volumes and set out with one of them in June 716 as an old man, dying on the road at Langres. His companions carried the book onward, and it survives, transformed by an eleventh-century erasure into the gift of a Lombard abbot to a Tuscan monastery, revered as a relic of Gregory the Great, consulted by papal revisers in 1587, transferred to Florence in the 1780s, and restored to its true makers by the philology of the 1880s.

What makes the manuscript inexhaustible is that each of its aspects illuminates the others. Its scale is the measure of the monastery’s wealth and organisation; its script is the measure of the scriptorium’s training; its text is the measure of its library; its pictures are the measure of its ambitions and of the models it had to hand; and its erased dedication is the measure of how completely an object can lose and then recover its history. It is the earliest complete Bible in Jerome’s Latin, one of the earliest depictions of a bookcase, one of the largest books of the first millennium, and one of the few whose commissioning is described by a writer who knew the man who commissioned it.

It is also, and finally, a warning against the confident reading of surfaces. For two hundred years the most eloquent surviving witness to Anglo-Saxon learning sat in Florence, catalogued as Italian, its dedication rewritten, its origin forgotten, and its authority undiminished. The book was always telling the truth about itself. It took a very long time for anyone to read it correctly.