There are half-a-dozen ancient manuscripts which are the foundation of our understanding of the text of the New Testament writings. Among these stands the copy known since the sixteenth century as Codex Bezae Cantabrigiensis. Any manuscript which has survived from antiquity is a marvel for this reason alone, and as we explore its pages, we have a rare opportunity to explore a little of the written culture of late antique Christianity. Although in the past century some remarkable papyrus manuscripts have been recovered from the sands of Egypt, their discovery has in general served more to highlight the significance of the parchment manuscripts than to diminish it.
Among this group, Codex Bezae occupies a unique place for several reasons. In the first place, as a bilingual manuscript, with a Greek text and a Latin version on facing pages, it provides a valuable insight into the reception of the Gospels and Acts in the western Christian tradition. The Latin version it contains is one of the small handful of manuscripts which are the most important witnesses to the development of a Latin version before Jerome's famous Vulgate of 382. Secondly, it provides a strikingly different form of text to that preserved in almost every other manuscript, and to the printed Greek text and the translations derived from it. These differences consist in the Gospels in frequent harmonisation of the text and in Acts in a free restyling of the text found best represented by Codex Vaticanus and reproduced in English translations.
The manuscript is the work of a single scribe, one trained primarily to copy Latin texts. Its present contents are the Gospels of Matthew, John, Luke and Mark, a single page of the last verses of 3 John (in Latin only) and the Acts of the Apostles. The only book that is complete is the Gospel of Luke, since there are pages missing from all the others. It is possible that between Mark and 3 John the manuscript originally contained Revelation and the rest of the Epistles of John. The Gospels are in the so-called Western order, with the two who were apostles first, followed by the two who were companions of the apostles.
The manuscript is best dated to the end of the fourth or the beginning of the fifth century. Many places have been proposed for its place of origin, including southern France, Africa, Egypt and Palestine. I have proposed Berytus (Beirut). There were a number of correctors and annotators working in the first centuries of its existence. The first strong evidence for the manuscript's history is replacement leaves for missing portions of Matthew, John and Mark. The style of writing and the use of blue ink provide a very strong case that these pages were written in Lyons in the ninth century. At this period Lyon was an important centre for the dissemination of ancient works in the west.
It is probable that the Codex Bezae remained there, in the Monastery of St Irenaeus, until the sixteenth century. It was apparently taken over the Alps to the Council of Trent in 1546. Its textual significance was already recognised, since it was one of the manuscripts whose readings was cited in the first edition of the Greek New Testament to include such information, made by Robert Stephanus in Paris in 1550. Then after the sacking of Lyons in the religious wars it came into the hands of the Reformer Theodore de Bèze, Calvin’s successor at Geneva. The first part of its name is derived from the Latin form of his name, Beza. In 1581, Beza presented the manuscript to Cambridge University. This is the origin of the second part of its name, Cantabrigiensis.
A printed transcription of the manuscript (using a font imitating the shape of the characters) was published by the University Press in 1793. A more accurate transcription, with the corrections and annotations fully detailed, was made by F.H. Scrivener and published by Deighton Bell in 1864. A facsimile edition was published by the University Press in 1899.
Of the many distinctive readings of the manuscript, the following deserve special mention:
It is the oldest manuscript to contain the story of the adulterous woman (John 7.53-8.11). It is on Folios 133v to 135.
The genealogy of Jesus in Luke's Gospel is arranged in reverse order so as to conform more closely with that in Matthew. It is on Folios 195v to 197.
There is a story about Jesus found in no other manuscript (the story of the man working on the Sabbath, placed after Luke 6.4). It is on Folios 205v and 206.
It is the oldest manuscript to contain the longer ending of Mark (16.9-20). The last pages of Mark are missing, so all that remains is the Greek text of verses 9-15. What follows is text supplied in the ninth century. It is on Folio 347v.
In Acts, when the angel delivers Peter from prison the detail is added that they go into the street down seven steps (Acts 12.10). It is on Folios 463v-464, eleven lines from the bottom of the page.
Professor David Parker
Edward Cadbury Professor of Theology and Director of the Institute for Textual Scholarship and Electronic Editing
University of Birmingham
March, 2012
Volume I: ii + 157 + ii Leaf height: 260 mm, width: 212-228 mm.
Volume II: ii + 258 + ii Leaf height: 260 mm, width: 218-225 mm.
Volume I:
Volume II:
There are leaf signatures in Latin letters (for quires) and Arabic numerals (for bifolios) in pencil in the lower right-hand corner of the rectos, by H.L. Pink. The letter sequence for quires runs, in Volume I: a-i, k-n, p-u, x, y*; Volume II: z, A-I, K-T, V, X, X*, Gg-Ii, Kk, Mm-Ss.
Asterisks denote leaves inserted during the 9th century to supply missing text (see Provenance for further details). A sequence of signatures, numbered by folio rather than leaf but preceded by a letter indicating the quire, appears on ff. 4-8, in Latin letters and Arabic numerals, apart from f. 8r, where the number is in Roman numerals.There are quire signatures in Greek numerals in brown ink at the end of quires, in the lower right-hand corner of the versos, probably original to the production of the manuscript. They are accompanied by a horizontal line above, turning into a downward diagonal to the left. This is consistent with Latin practice, as is the positioning of the signatures. Most of them have been lost to cropping or damage. They survive on Quires 30 (λβʹ), 34 (λϛʹ), 36 (ληʹ), 37 (λθʹ), 39 (μαʹ), 40 (μβʹ), 44 (νδʹ), 46 (νϛʹ), 48-51 (νθʹ-ξβʹ) and partially on Quires 14 (ιϛʹ?), 23, 38 (μʹ) and 47 (νηʹ).
The evidence of these signatures supports the conjecture reflected in the foliation that one original quire has been lost after Quire 2 and another after Quire 12, since apparently by Quire 14 and certainly by Quire 30 they are larger than the count of surviving quires by two. They also corroborate the conjecture that eight full quires were lost after the incomplete Quire 42, since by 44 they are larger than the count of surviving quires by ten, and one quire after Quire 46, since apparently by Quire 47, and certainly by Quire 48, they are larger by eleven.
There is another set of quire signatures in Arabic numerals in brown ink at the beginning of each quire, in the lower left-hand corner of the rectos. Those on Quires 22-23, 37 and 43 have apparently been lost to damage, suggesting that they were added before this damage was incurred. No signature has been placed on Quire 45. This sequence includes the replacement Quire 20 as well as the original quires; the number 21 appears on the last folio of this quire. When the sequence resumes after those lost to damage, on Quire 24, it bears the number 23, and this sequence is thereafter smaller than the count of surviving quires by one. However, the number 38 is skipped, so from Quire 39 they again correspond to the actual count. The signatures of Quires 42 and 53 appear on the second folio rather than the first.
A third set of quire signatures appears in Greek numerals in pencil at the beginning of each quire, in the lower right-hand corner of the recto.These take into account lost quires, corresponding to the original numbering rather than the count of those currently surviving. No signature of this sequence appears on Quire 53.
No catchwordsQuarter tawed pigskin over oak boards, in two volumes. Sandy and Elizabeth Cockerell, 1965 (stamp on skin on the inside of the right-hand boards).
Volume I Binding height: 279 mm, width: 250 mm, depth: 52 mm.
Volume II Binding height: 279 mm, width: 252 mm, depth: 71 mm.
A former binding, dating to the 19th-century, is preserved separately: pasteboards covered in light brown leather, with blind-tooled panel and scrollwork, with metal clasp, metal plates at the corners and another in the centre engraved with the coat of arms of the University of Cambridge (on the front) and 'Codex Bezae Cantab.' (on the rear) (?pewter, all now tarnished).
A report by Sandy Cockerell - entitled 'Repairing and Rebinding the Codex Bezae' - is kept separately, along with several black-and-white photographs of the manuscript and the rebinding work in progress. Dated November 1965, but beginning with a description of the condition of the manuscript dated January 1962, the report suggests that the project to conserve and rebind Codex Bezae took place over several years. According to Cockerell:
"The manuscript had been removed from the 19th century binding at some time and in removing it a great deal of damage had been done to the folds of the vellum leaves, particularly to the outside fold of each gathering."
It is not clear when this removal took place, but other notes imply that this occurred during the Librarianship of Henry Bradshaw (b. 1831, d. 1886, Librarian 1867-1886): specifically, a collation diagram by Bradshaw's hand, which describes the structure of the quires and concludes 'all in box Aug 25' and 'together with Bezae's letter + seal Sept. 3' (no year is given). This sheet then continued to be used by subsequent members of Library staff as a means of checking that all of the leaves were present: it is signed 'all here' by Francis Jenkinson on 30 September 1898, and by H.L. Pink on 14 December 1949, 5 December 1952 and 26 January 1962.
Bradshaw's notes indicate that Codex Bezae was kept disbound in a box and this seems to have been the case until Cockerell's rebinding work. Another note, written by H.R. Creswick (1902-1988, Librarian 1949-1967), states '4 Red boxes inspected 9 Dec. 1949 HRC & HLP [i.e. H.L. Pink] & JCTO [i.e. J.C.T. Oates]. All the Manuscript is in good state'. An index card bears the notes by several members of staff between 4 June 1904 and 16 May 1963, recording the removal of selected leaves for display in the Library's show cases. In some cases, the display was just for a single day, notes regarding their removal and return beearing the same date, however in other instances it appears that leaves from Codex Bezae remained on view for extended periods: a note by H.R. Creswick concerning unspecified folios states 'In show case since 1926 at least, returned to box 9 May 1934 (Librarian's order to clear show cases)' (presumably in preparation for the re-location of the Library to the new building on West Road).
20th-century foliation:
Volume I
[i-ii] + 3-6, 7*, 8-16, 25-104, 113-168, 169*-175* + [iii-iv]
Volume II
[v-vi] + 177-347, 348*, 415-446, 455-502, 504-508, 510 + [vii-viii]
Both volumes are foliated in Arabic numerals, in pencil, recto, upper right, by H.L. Pink (Assistant Under-Librarian 1947-1960 and Head of the Department of Manuscripts, 1960-1970).
Folio numbers 1-2, 17-24, 105-112, 176, 349-414, 447-454, 503, 509 were assigned to leaves no longer present in the volume, according to the interpretation of the manuscript's collation (as per historic foliation practice at Cambridge University Library).
Asterisks denote leaves inserted during the 9th century to supply missing text (see Provenance for further details).
Early modern foliation:
Volume I
1-156
Volume II
157-270, 270 [bis], 271-312, 314-393, 393 [bis], 394-413
There is an earlier foliation in Arabic numerals, in black ink, recto, upper right.
This foliation does not count the mutilated leaf later numbered as f. 96. The numbers 270 and 393 have been erroneously used twice, and the number 313 skipped.
An early hand has included marginal Ammonian section numbers accompanying the Greek text, and running heads indicating the current book in both Greek and Latin in the upper margin, on the single horizontal ruling line besides the text lines. Occasional Greek lection notes, including some marginal ἀρχή and τέλος marks, have been added by different hands, including one set in fomal majuscule script and others in a very crude hand. There are numerous corrections of the main text by several different hands.
Later Greek hands, one using an informal minuscule script and a crude and eccentric majuscule script have added notes in margins of the Greek text regarding the content of the page on which they appear. The latter hand has also added προσερμηνείαι (notes to be used in divination) in a crude Greek majuscule script in the lower margins of the Greek text of the Gospel of Mark. Where these are added to Gospel manuscripts, they usually appear in the Gospel of John, but that of Mark here appears last in the sequence, the position usually occupied by John.
The beginning of modern chapters has been marked with their number in Roman numerals in the margin of the Latin text by different hands, whille running chapter numbers in Roman numerals appear in the upper right-hand corner of recto folios from f. 69r onwards. These numbers also appear in the upper left-hand corner of the verso in the same portion of the manuscript, except in the latter part of Matthew (ff. 83v103v), where this position is occupied by Greek κεφάλαιον (chapter) numbers.
There is a Latin note on f. 73r regarding the names of the Magi and the thieves crucified with Jesus. Latin notes in ink in different hands and modern English notes in pencil have been added regarding lost folios and quires.
Perhaps at the monastery of St Irenaeus in Lyon, France from the ninth century. Leaves inserted in the ninth century to supply missing text bear Lyonese characteristics, notably the use of blueish ink. The text also appears to have influenced the martyrology written by Ado of Lyons in 850-60, and it is possible that Florus of Lyon was responsible for it coming to this location.
Taken to the Council of Trent by Guillaume du Prat, bishop of Clermont, where some of its Greek readings were cited as evidence relevant to matters then under debate.
According to a note by Theodore Beza, the manuscript was discovered in the monastery of St Irenaeus in 1562 when the city was sacked by the Huguenot army during the religious wars.Editions:
Manuscript descriptions:
Secondary literature:
Under the 'More' menu you can find metadata about the item, and information about sharing this image.
Volume I: ii + 157 + ii Leaf height: 260 mm, width: 212-228 mm.
Volume II: ii + 258 + ii Leaf height: 260 mm, width: 218-225 mm.
Volume I:
Volume II:
There are leaf signatures in Latin letters (for quires) and Arabic numerals (for bifolios) in pencil in the lower right-hand corner of the rectos, by H.L. Pink. The letter sequence for quires runs, in Volume I: a-i, k-n, p-u, x, y*; Volume II: z, A-I, K-T, V, X, X*, Gg-Ii, Kk, Mm-Ss.
Asterisks denote leaves inserted during the 9th century to supply missing text (see Provenance for further details). A sequence of signatures, numbered by folio rather than leaf but preceded by a letter indicating the quire, appears on ff. 4-8, in Latin letters and Arabic numerals, apart from f. 8r, where the number is in Roman numerals.There are quire signatures in Greek numerals in brown ink at the end of quires, in the lower right-hand corner of the versos, probably original to the production of the manuscript. They are accompanied by a horizontal line above, turning into a downward diagonal to the left. This is consistent with Latin practice, as is the positioning of the signatures. Most of them have been lost to cropping or damage. They survive on Quires 30 (λβʹ), 34 (λϛʹ), 36 (ληʹ), 37 (λθʹ), 39 (μαʹ), 40 (μβʹ), 44 (νδʹ), 46 (νϛʹ), 48-51 (νθʹ-ξβʹ) and partially on Quires 14 (ιϛʹ?), 23, 38 (μʹ) and 47 (νηʹ).
The evidence of these signatures supports the conjecture reflected in the foliation that one original quire has been lost after Quire 2 and another after Quire 12, since apparently by Quire 14 and certainly by Quire 30 they are larger than the count of surviving quires by two. They also corroborate the conjecture that eight full quires were lost after the incomplete Quire 42, since by 44 they are larger than the count of surviving quires by ten, and one quire after Quire 46, since apparently by Quire 47, and certainly by Quire 48, they are larger by eleven.
There is another set of quire signatures in Arabic numerals in brown ink at the beginning of each quire, in the lower left-hand corner of the rectos. Those on Quires 22-23, 37 and 43 have apparently been lost to damage, suggesting that they were added before this damage was incurred. No signature has been placed on Quire 45. This sequence includes the replacement Quire 20 as well as the original quires; the number 21 appears on the last folio of this quire. When the sequence resumes after those lost to damage, on Quire 24, it bears the number 23, and this sequence is thereafter smaller than the count of surviving quires by one. However, the number 38 is skipped, so from Quire 39 they again correspond to the actual count. The signatures of Quires 42 and 53 appear on the second folio rather than the first.
A third set of quire signatures appears in Greek numerals in pencil at the beginning of each quire, in the lower right-hand corner of the recto.These take into account lost quires, corresponding to the original numbering rather than the count of those currently surviving. No signature of this sequence appears on Quire 53.
No catchwordsQuarter tawed pigskin over oak boards, in two volumes. Sandy and Elizabeth Cockerell, 1965 (stamp on skin on the inside of the right-hand boards).
Volume I Binding height: 279 mm, width: 250 mm, depth: 52 mm.
Volume II Binding height: 279 mm, width: 252 mm, depth: 71 mm.
A former binding, dating to the 19th-century, is preserved separately: pasteboards covered in light brown leather, with blind-tooled panel and scrollwork, with metal clasp, metal plates at the corners and another in the centre engraved with the coat of arms of the University of Cambridge (on the front) and 'Codex Bezae Cantab.' (on the rear) (?pewter, all now tarnished).
A report by Sandy Cockerell - entitled 'Repairing and Rebinding the Codex Bezae' - is kept separately, along with several black-and-white photographs of the manuscript and the rebinding work in progress. Dated November 1965, but beginning with a description of the condition of the manuscript dated January 1962, the report suggests that the project to conserve and rebind Codex Bezae took place over several years. According to Cockerell:
"The manuscript had been removed from the 19th century binding at some time and in removing it a great deal of damage had been done to the folds of the vellum leaves, particularly to the outside fold of each gathering."
It is not clear when this removal took place, but other notes imply that this occurred during the Librarianship of Henry Bradshaw (b. 1831, d. 1886, Librarian 1867-1886): specifically, a collation diagram by Bradshaw's hand, which describes the structure of the quires and concludes 'all in box Aug 25' and 'together with Bezae's letter + seal Sept. 3' (no year is given). This sheet then continued to be used by subsequent members of Library staff as a means of checking that all of the leaves were present: it is signed 'all here' by Francis Jenkinson on 30 September 1898, and by H.L. Pink on 14 December 1949, 5 December 1952 and 26 January 1962.
Bradshaw's notes indicate that Codex Bezae was kept disbound in a box and this seems to have been the case until Cockerell's rebinding work. Another note, written by H.R. Creswick (1902-1988, Librarian 1949-1967), states '4 Red boxes inspected 9 Dec. 1949 HRC & HLP [i.e. H.L. Pink] & JCTO [i.e. J.C.T. Oates]. All the Manuscript is in good state'. An index card bears the notes by several members of staff between 4 June 1904 and 16 May 1963, recording the removal of selected leaves for display in the Library's show cases. In some cases, the display was just for a single day, notes regarding their removal and return beearing the same date, however in other instances it appears that leaves from Codex Bezae remained on view for extended periods: a note by H.R. Creswick concerning unspecified folios states 'In show case since 1926 at least, returned to box 9 May 1934 (Librarian's order to clear show cases)' (presumably in preparation for the re-location of the Library to the new building on West Road).
20th-century foliation:
Volume I
[i-ii] + 3-6, 7*, 8-16, 25-104, 113-168, 169*-175* + [iii-iv]
Volume II
[v-vi] + 177-347, 348*, 415-446, 455-502, 504-508, 510 + [vii-viii]
Both volumes are foliated in Arabic numerals, in pencil, recto, upper right, by H.L. Pink (Assistant Under-Librarian 1947-1960 and Head of the Department of Manuscripts, 1960-1970).
Folio numbers 1-2, 17-24, 105-112, 176, 349-414, 447-454, 503, 509 were assigned to leaves no longer present in the volume, according to the interpretation of the manuscript's collation (as per historic foliation practice at Cambridge University Library).
Asterisks denote leaves inserted during the 9th century to supply missing text (see Provenance for further details).
Early modern foliation:
Volume I
1-156
Volume II
157-270, 270 [bis], 271-312, 314-393, 393 [bis], 394-413
There is an earlier foliation in Arabic numerals, in black ink, recto, upper right.
This foliation does not count the mutilated leaf later numbered as f. 96. The numbers 270 and 393 have been erroneously used twice, and the number 313 skipped.
An early hand has included marginal Ammonian section numbers accompanying the Greek text, and running heads indicating the current book in both Greek and Latin in the upper margin, on the single horizontal ruling line besides the text lines. Occasional Greek lection notes, including some marginal ἀρχή and τέλος marks, have been added by different hands, including one set in fomal majuscule script and others in a very crude hand. There are numerous corrections of the main text by several different hands.
Later Greek hands, one using an informal minuscule script and a crude and eccentric majuscule script have added notes in margins of the Greek text regarding the content of the page on which they appear. The latter hand has also added προσερμηνείαι (notes to be used in divination) in a crude Greek majuscule script in the lower margins of the Greek text of the Gospel of Mark. Where these are added to Gospel manuscripts, they usually appear in the Gospel of John, but that of Mark here appears last in the sequence, the position usually occupied by John.
The beginning of modern chapters has been marked with their number in Roman numerals in the margin of the Latin text by different hands, whille running chapter numbers in Roman numerals appear in the upper right-hand corner of recto folios from f. 69r onwards. These numbers also appear in the upper left-hand corner of the verso in the same portion of the manuscript, except in the latter part of Matthew (ff. 83v103v), where this position is occupied by Greek κεφάλαιον (chapter) numbers.
There is a Latin note on f. 73r regarding the names of the Magi and the thieves crucified with Jesus. Latin notes in ink in different hands and modern English notes in pencil have been added regarding lost folios and quires.
Perhaps at the monastery of St Irenaeus in Lyon, France from the ninth century. Leaves inserted in the ninth century to supply missing text bear Lyonese characteristics, notably the use of blueish ink. The text also appears to have influenced the martyrology written by Ado of Lyons in 850-60, and it is possible that Florus of Lyon was responsible for it coming to this location.
Taken to the Council of Trent by Guillaume du Prat, bishop of Clermont, where some of its Greek readings were cited as evidence relevant to matters then under debate.
According to a note by Theodore Beza, the manuscript was discovered in the monastery of St Irenaeus in 1562 when the city was sacked by the Huguenot army during the religious wars.Editions:
Manuscript descriptions:
Secondary literature: