Available via license: CC BY
Content may be subject to copyright.
Elizabeth Mullins
The Eusebian Apparatus in Irish Pocket
Gospel Books: Absence, Presence and Addition
Abstract: This paper focuses on the presence of the dif-
ferent elements of Eusebius’s system of gospel con cord-
ance in a series of pocket gospel books associated with
early medieval Ireland. It provides a brief overview of the
pocket gospel book series as a whole and discusses the
appearance of parts of the Eusebian system in the Book
of Armagh and in the MacDurnan gospels. The addition
of the Eusebian apparatus to the Book of Mulling is then
examined. The paper demonstrates how the way the
apparatus was included in Mulling echoes the close atten-
tion to the series that is evident in contemporary Hiberno-
Latin texts. It highlights how the marginal references to
the Eusebian system included in these gospels provide
hitherto neglected evidence for their transmission and for
the ways that the gospel may have been read in the Irish
medieval context.
This paper focuses on evidence of the Eusebian apparatus
in a series of gospel books, the so-called pocket gospels,
which are closely associated with medieval Ireland. While
Carl Nordenfalk, whom this volume honours, commented
on aspects of these books, particularly their author por-
traits, they did not feature in his scholarship on the
Eusebian system.¹ The reasons for this are clear. Pocket
gospel books are generally distinguished by the absence
of prefa tory material; the two instances of Eusebian tables
which feature in the series, one part of an original book and
the second an addition, are without significant decoration.
While thus falling outside the ambit of Nordenfalk’s schol-
arship, evidence of the Eusebian apparatus in these books
has much to reveal about the ways in which the system
was transmitted and used in eighth- and ninth-century
Ireland. This paper will initially provide a brief introduc-
tion to the pocket gospel book series. It will discuss the
This paper is based on my contribution to the conference ‘80 years
since Nordenfalk’ which took place in Hamburg in May 2018. I would
like to thank Bruno Reudenbach, Hanna Wimmer and Alessandro Bausi
for the invitation to speak at this event. I am grateful to Hugh Houghton,
Martin MacNamara, Dáibhí Ó Cróinín, Terence O’Reilly and Dermot
Roantree for their help at various stages in the research. I would like
to thank Tadgh Ó hAnnracháin and Catherine Cox, colleagues in the
School of History, for their practical support. UCD College of Arts and
Humanities Research Fund provided financial support for purchase of
the images and rights associated with this paper.
tables and the general preface in the Book of Armagh,
the marginal notation in the MacDurnan gospels and the
traces of the Eusebian system that remain in the text of
many of the other books. The addition of the Eusebian
apparatus to the Book of Mulling is then examined. There
is discussion of the model which underlies the apparatus
in Mulling and consideration of the reasons the Eusebian
system was added to this manuscript within a hundred or
so years of its original production.
1 The Irish Pocket Gospel Book
corpus
The Irish pocket gospel book group forms a relatively dis-
tinct set of manuscripts within the larger corpus of Insular
gospel books. The existence of the group was highlighted
as early as 1956 by Patrick McGurk in a seminal article
in Sacris Erudiri 8.² In this article and subsequent work
McGurk identified eight books which directly belong to the
pocket gospel book tradition.³ These are listed in Table 1.
The pocket gospel books are distinguished firstly
by their size. The smallest of these books is the Cadmug
Gospels, the largest is the Book of Armagh, with the Book
of Mulling coming somewhere in the middle. Another dis-
tinctive feature of the pocket gospel books is their codico-
logy. Frequently, each gospel is found either on a separate
set of quires or on a single large quire. The Book of Mulling
provides an example of the former practice, with the origi-
nal manuscript containing Matthew’s gospel on 22 leaves,
Mark on 18, Luke on 30 and John on 14. The later additi-
onal material is included on a single quire at the opening
of the book.
1Nordenfalk 1977, 126.
2McGurk 1956; McGurk 1987. For a more recent discussion of the
pocket gospel book series, see Meehan 2015.
3McGurk 1987, 166–167.
4McGurk 1961, 83; Houghton 2016, 227. A full digital edition of
the Book of Mulling is available at https://digitalcollections.tcd.ie/
home/index.php?folder_id=1648&pidtopage=MS60_001&entry_
point=1 (last accessed 13/04/2020).
Open Access. © Elizabeth Mullins, published by De Gruyter. This work is licensed under the Creative Commons Attribution-NonCom-
mercial-NoDerivatives . License.
https://doi.org/./-
48 Elizabeth Mullins
¹ ¹¹ ¹²
Pocket gospel books, in general, simply contain the gospel
texts without any paratextual material. The books present
the gospels in the Vulgate order, Matthew, Mark, Luke
and John. In common with many Insular gospel books,
the type of gospel text contained in these books is mixed.
Mulling, the most mixed of all the texts, has been seen to
drawn on a Vulgate text, an Irish mixed text and an Old
Latin text.¹³ Some passages, particularly towards the start
of Luke’s gospel (Chapters 4–9), have a much higher con-
centration of Old Latin readings. Research carried out on
the Old Latin text of Luke’s gospel has seen it as relating,
in particular, to the type of Old Latin text found in r1 Usse-
rianus, the earliest surviving gospel book in Irish script,
together with a text of the Old Latin b (Veronensis) kind.
Pocket gospel books are illustrated in a recognizably
Insular manner. While some of the books are incomplete,
each gospel seems to have opened with an enlarged incipit
preceded by the standing portrait of the relevant evange-
list or evangelist symbol. In Mulling, three of these evange-
lists and incipit pages survive. A variation on this pattern
is evident in two books. In the MacDurnan Gospels, a four
symbols page is included before the opening of Matthew’s
5CLA II, 267, https://elmss.nuigalway.ie/catalogue/586 (last acces-
sed 13/04/2020); McGurk 1961,78.
6CLA II, 275, https://elmss.nuigalway.ie/catalogue/594 (last acces-
sed 13/04/2020); McGurk 1961, 82–83.
7CLA II, 276, https://elmss.nuigalway.ie/catalogue/595 (last acces-
sed 13/04/2020); McGurk 1961, 83.
8CLA II, 277, https://elmss.nuigalway.ie/catalogue/596 (last acces-
sed 13/04/2020); McGurk 1961, 84.
9CLA VIII, 1198, https://elmss.nuigalway.ie/catalogue/1674 (last ac-
ces sed 13/04/2020); McGurk 1961, 67–68.
10CLA II, 179, See https://elmss.nuigalway.ie/catalogue/494 (last
acces sed 13/04/2020); Mc Gurk 1961, 32–33.
11See the catalogue entry at http://archives.lambethpalacelibrary.
org.uk/CalmView (last accessed 13/04/2020).
12CLA II, 270, https://elmss.nuigalway.ie/catalogue/589 (last ac-
cessed 13/04/2020).
13For this and what follows see Doyle 1973, 177–200; Houghton
2016, 76, 227.
Manuscript Name Date Size (mm)
Dublin, Royal Irish Academy, D.II. (fols –)Stowe St John VIII–IX ×
Dublin, Trinity College Library, Book of Dimma VIII ×
Dublin, Trinity College Library, Book of Mulling VIII–IX ×
Dublin, Trinity College Library, , fols –Mulling Fragments VIII–IX c. ×
Fulda, Landesbibliothek, Bonif. Cadmug Gospels VIII–IX ×
London, British Library, Add. Irish Gospels VIII–IX ×
London, Lambeth Palace Library, MacDurnan Gospels IX ×
Dublin, Trinity College Library, Book of Armagh New Testament c. ×
Tab. 1: The Irish Pocket Gospel Book Corpus.
gospel, providing a frontispiece for Matthew 1:1–1:17, while
the Matthew evangelist portrait is included subsequently,
facing the opening of Matthew 1:18.¹ The gospel book
contained in the New Testament Book of Armagh features
a four-symbols page before the Gospel of Matthew and
an evangelist symbol before or as part of the first page of
the Gospels of Mark, Luke and John, with the symbols for
the final two gospels incorporating allusions to the other
evangelical beasts.¹ All of these illustrative features are
present in larger gospel books associated with Ireland at
the time, such as the Books of Durrow and Kells, the Ech-
ternach Gospels and the Irish Gospel Book, St Gall, Stifts-
bibliothek, Cod. Sang. 51.¹
Art historical research on these images, by Jennifer
O’Reilly, Bernard Meehan and others, has demonstrated
the sophistication of these books’ visual exegesis and how
they reflect aspects of patristic and Insular textual com-
mentary on the harmony and distinctiveness of the four-
fold text.¹ Art-historical appreciation of these books has
been supported by contemporary scientific work, which
has highlighted, for example in the case of Mulling, the
deliberate combination and juxtaposition of opaque and
translucent pigments on illuminated pages.¹
14Alexander 1978, 86–87. For discussion of the significance of Mt
1,18 in the Insular context see, for example, O’Reilly 1998.
15Alexander 1978, 76–77. For digital reproduction of the Book of
Armagh see https://digitalcollections.tcd.ie/home/index.php#folder_
id=26&pidtopage=MS52_01&entry_point=1 (last accessed 30/08/2018).
16Digital edition of the Book of Kells is available at https://digi
talcollections.tcd.ie/home/index.php?DRIS_ID=MS58_003v (last ac-
ces sed 13/04/2020); Digital Edition of the Book of Durrow is avail-
able at https://digitalcollections.tcd.ie/home/index.php?folder_id=
1685&pidtopage=MS55_001&entry_point=1#folder_id=1845&pidto
page=MS57_178&entry_point=1 (last accessed 13/04/2020). Digital Ed -
ition of the Echternach Gospels is available at https://gallica.bnf.fr/
ark:/12148/btv1b530193948 (last accessed 13/04/2020); digital edition
of St Gall, Stiftsbibliothek, Cod. Sang. 51 is available at http://www.
e-codices.unifr.ch/de/list/one/csg/0051 (last accessed 13/04/2020).
17O’Reilly 1998a; Meehan 2014.
18Bioletti/Smith 2017, 123.
The Eusebian Apparatus in Irish Pocket Gospel Books 49
While the pocket gospel books reflect some of the illus-
trations of the Insular tradition, they omit others, such
as carpet pages, and, of course, Eusebian canon tables.
When canon tables appear in larger Insular gospel books,
they have been seen to reinforce themes of evangelical
harmony and distinctiveness. This is dramatically exem-
plified by the Christological symbolism and shape-shift-
ing evangelical beasts decorating many of the tables in the
Book of Kells.¹ Similarly, recent research on the Lindis-
farne Gospels has demonstrated how the restrained deco-
ration of its canon series is based on a complex numerical
and ornamental scheme alluding to the harmony of the
Eusebian text.² In addition to these examples of arcaded
series, larger Insular gospel books also bear witness to
a second contemporary tradition of non-architectural
tables. This tradition is evident in the final two tables
included in the Book of Kells, and in the Durrow and Ech-
ternach manuscripts (Figs 1–2). It reappears in the set of
tables spread across fifteen pages that are included before
the Hiberno-Latin commentary on Matthew in Vienna,
Österreichische Nationalbibliothek, Cod. 940 (Fig. 3).²¹
While pocket gospel books generally omit the apparatus,
it is this second tradition which is reflected in the two
examples of tables which survive in the pocket gospel
book series, in the New Testament portion of the Book of
Armagh (Fig. 4) and in the Book of Mulling (Fig. 5).
2 The Eusebian apparatus in the
Pocket Gospel Book corpus
The unframed tables in the Book of Armagh are the only
instance of the inclusion of this part of the apparatus in
the original planning of a member of the pocket gospel
book group.²² The gospel sections in Armagh begin with
one general preface, Novum Opus. This is succeeded by
a set of unframed canons that spread across seven pages
and chapter lists for each of the four gospels. Each of
the gospels is then introduced by an appropriate preface
19O’Reilly 1998, 71; Mullins 2001, 188–198; See also Endres 2017;
Neuman de Vegvar 2007.
20Pulliam 2017.
21Mullins 2014. Digital edition is available at http://digital.onb.
ac.at/RepViewer/viewer.faces?doc=DTL_5897339&order=1&view=
SINGLE (last accessed 13/04/2020).
22The gospel sections in the Book of Armagh are part of a manu-
script containing a New Testament and texts relating to St Patrick.
The fact that the gospel sections in this book are on separate quires
led McGurk to consider them in the context of the pocket gospel book
series. On this see McGurk 1987, 166–167.
(argumentum) and by a set of Hebrew names. McGurk con-
nected Armagh’s prefatory series to that in a number of
other gospel books with Irish associations including the
Books of Durrow, Kells, and Mulling and the Echternach
Gospels. This connection was based on several common
features: confining the general prefaces to Novum Opus,
including unframed canons and the bunching of prefaces
normally dispersed before the individual gospels, in the
case of Armagh its chapter lists, at the beginning or end
of the manuscript. In addition to these features, the type
of Hebrew name list and argumentum included in Armagh
belong textually to a set which shows Irish influence. ²³
While the Book of Armagh includes the first two elements
of the Eusebian apparatus, it does not feature the third
element, the marginal notation, nor are the Eusebian sec-
tions consistently marked out within the text. This reflects
the reception of the system in some of the larger Insular
gospel books such as, for example, the Book of Kells,
which probably also had Novum Opus and includes a set
of tables, but famously features marginal notation only on
one double opening in John’s Gospel (fols 292v–293r).
The other pocket gospel book that features an element
of the Eusebian system as part of its original design is
the MacDurnan Gospels, thought to have been written in
Armagh in the second half of the ninth century.² Although
not containing any prefatory material, the entire gospel
text is beautifully laid out and consistently gives its Euse-
bian sections separate paragraphs, though, as McGurk
notes, chapters are ignored, and verses only occasionally
indicated.² The paragraphs corresponding to the Euse-
bian sections are marked with large initials and are accom-
panied by the appropriate marginal notation (Fig.6). The
general appearance and occasional framing of this nota-
tion by an orange cartouche is reminiscent of that of the
Echternach gospels.² Although the individual marginal
entries in MacDurnan differ from Echternach in the order
in which the canon and section numbers are written,
further research is needed to explore their respective nota-
tion, particularly in the light of the already established link
between their gospel texts.²
The MacDurnan Gospels include the extended system
of marginal notation, an early ‘improvement’ in the trans-
23McGurk 1987, 170–172.
24See Alexander 1978, 86–87. An earlier date has more recently
been proposed by Farr 2011.
25McGurk 1987, 173.
26Full digital reproduction of the Echternach Gospels is available
at https://gallica.bnf.fr/ark:/12148/btv1b530193948 (last accessed
13/04/2020).
27McNamara 1990, 102–111. This is the subject of ongoing work by
the author.
50 Elizabeth Mullins
Fig. 1: Dublin, Trinity College Library, MS 58, Book of Kells, fol. 6r. Canon X Lk, Canon X Jn. Image reproduced courtesy of The Board of Trinity
College Dublin.
The Eusebian Apparatus in Irish Pocket Gospel Books 51
Fig. 2: Dublin, Trinity College Library, MS 57, Book of Durrow, fol. 9v. Canons V–VIII. Image reproduced courtesy of The Board of Trinity
College Dublin.
52 Elizabeth Mullins
Fig. 3: Vienna, Österreichische Nationalbibliothek, Cod. 940, fol. 8r. Canon VII, Canon VIII, Canon VIIII. Image reproduced courtesy of
Vienna, Österreichische Nationalbibliothek.
The Eusebian Apparatus in Irish Pocket Gospel Books 53
Fig. 4: Dublin, Trinity College Library, MS 52, Book of Armagh, fol. 28r. Canons V–VIII. Image reproduced courtesy of The Board of Trinity
College Dublin.
54 Elizabeth Mullins
Fig. 5: Dublin, Trinity College Library, MS 60, Book of Mulling, fol. 10v. Canons V–VIII. Image reproduced courtesy of The Board of Trinity
College Dublin.
The Eusebian Apparatus in Irish Pocket Gospel Books 55
Fig. 6: London, Lambeth Palace Library, 1370, MacDurnan Gospels, fol. 19v. Marginal Entries for Matthew Sections LXXI–LXXIII. Image repro-
duced courtesy of Lambeth Palace Library.
56 Elizabeth Mullins
mission of the Latin version of the Eusebian concordance.
This notation, which is common in Insular books featur-
ing the series, contained a reference not just to the number
of the section and canon to which a passage belonged, as
Eusebius and subsequently Jerome advised, but also a ref-
erence or references to the relevant parallel section(s) in
the other gospels.² Including the parallel section(s) in this
way, as Tom O’Loughlin has pointed out, obviated the need
for readers of the gospel to go back to the tables included
at the start of the book.² MacDurnan’s complete omission
of the tables has taken this process a step further. Does the
inclusion of marginal notation as a standalone element in
this book indicate a lack of under standing of the apparatus
or an in-depth familiarity with the system? While it is dif-
ficult to answer this, it is worth noting in this context that
another contemporary Irish manuscript, St Gall, Stiftsbib-
liothek, Cod. Sang. 60, also includes marginal notation as a
standalone feature. This Gospel of John, which was copied
in Ireland around 800, has no prefatory material but fully
extended Eusebian marginal references throughout.³
A different reflection of the same practice occurs in
the Hiberno-Latin commentary on Luke in Vienna, Öster-
reichische Nationalbibliothek, Cod. 997. The presence of
marginal references to the Eusebian sections in this com-
mentary manuscript was first noted by Bernhard Bischoff
in his original article ‘Wendepunkte in der Geschichte
der latein ischen Exegese im Frühmittelalter’, but the ref-
erences were omitted from the subsequent edition of the
commentary by Joseph Kelly.³¹ Unlike Vienna 940, men-
tioned above, which includes both Novum Opus and a set
of canons before a commentary on Matthew that is struc-
tured according to Eusebian divisions, this commentary
on Luke omits the first two parts of the apparatus but has
79 instances of marginal notation running alongside two-
thirds of the text (Fig.7). This notation accompanies the
text in the sense that it directly relates to the passage from
Luke’s Gospel that is being commented on at the time. The
evidence this provides of marginal notation being used
independently of the other parts of the system as a tool of
exegesis and cross-referencing is also worth considering
in the context of MacDurnan.
Apart from Armagh and MacDurnan, in the other six
members of the pocket gospel book series the only traces
28This extended notation is found for example in the Lindisfarne
Gospels, the Book of Durrow and the Codex Amiatinus. On this see
O’Loughlin 1999.
29O’Loughlin 1999, 5–6; O’Loughlin 2010, 17–20.
30McGurk 1961, 98; Houghton 2016, 78–79. Digital reproduction of St
Gall, Stiftsbibliothek, Cod. Sang. 60 http://www.e-codices.unifr.ch/
de/csg/0060/bindingA/0/Sequence-263 (last accessed 13/04/2020).
31Bischoff 1954, 263; Commentarium in Lucam 1974.
of the Eusebian system that remain are within the gospel
text, primarily in the decoration of the first letter of par-
ticular Eusebian sections. Most of the books (and this,
again, reflects the tradition in the larger books) have a
relatively inconsistent approach to this—some section
initials are marked but these are not the only passages
with enlarged initials.³² Much of the explanation for this
inconsistent pattern lies in the kinds of models used for
these manuscripts and the extent to which they reflected
Vulgate or Old Latin section divisions. This fact is borne
out by the text in the Book of Mulling. Work carried out by
Lawlor at the end of the nineteenth century demonstrated
how in its original conception, Mulling contained traces of
the Eusebian system only in the parts of its gospel which
were closest to the Vulgate, as in the gospel of Mark, while
the divisions of St Matthew and St Luke had nothing to
do with the Eusebian sections, because they feature a
stronger Old Latin element, and John presents a mixed
pattern.³³
3 The additions to the book of
Mulling
Sometime after its original conception, the Eusebian
apparatus was attached to the Book of Mulling as part of
a prefatory series. This series contains many of the fea-
tures of the texts that are included in the Book of Armagh
and the other Irish manuscripts mentioned earlier in this
paper. Mulling has just one general preface, Novum Opus.
This is followed by the prefaces (argumenta) for the four
gospels bunched together as a group. Canon tables in
non-architectural frames follow, with plain red vertical
lines dividing the tables of numbers (Fig.5).
Mulling’s canon tables are based on a series which was
unusually distributed across 15 pages. Mulling includes 14
of these, but is missing the final page for the end of Canon
X Jn.³ Mulling’s unusual page distribution is paralleled
only in the set of tables in Vienna 940. There are differences
between the series in the two manuscripts, however, both
in their table text and internal distribution. A collation of
the Mulling tables with the Vulgate shows a relatively clean
text in the Durrow/Amiatinus tradition.³
32See O’Loughlin 2007.
33Lawlor 1897, 37.
34See Lawlor 1897, 8–9 for a discussion of the contents of the first
quire of the manuscript. In addition to the final page of the canon
series, the second half of the Novum Opus preface and the first part of
the argumentum for Matthew are also missing.
35The Mulling tables were collated with the edition which appears
The Eusebian Apparatus in Irish Pocket Gospel Books 57
Fig. 7: Vienna, Österreichische Nationalbibliothek, Cod. 997, fol. 27v. Marginal Entries for Luke Sections XXXVI–XXXVII. Image reproduced
courtesy of Vienna, Österreichische Nationalbibliothek.
58 Elizabeth Mullins
Even if the Mulling series does not appear as clearly laid
out as Vienna 940, minimal errors are included, and these
are often picked up on. One such correction is evident on
the final line of Canon II on fol. 8r, where the parallel for
Luke’s gospel is corrected from 339 to 338. Similarly, on
fol. 9r, Canon IIII, there is a correction from 77 to 67 in the
entry for Mark in the fourth parallel.³
One major change in the Mulling series is in Canon
VIII, where the order of the evangelists in the table is
reversed from the standard form of Luke Mark to Mark
Luke (Fig. 5). Neglected to this point by scholars, this
reversal is also evident in the series in Durrow (fol. 9v)
and Kells (fol. 5v) (Fig.2).³ Interestingly, it also appears
in the canon series in the fragmentary Anglo-Saxon Bible
Royal I. E. VI.³ In this manuscript, unlike in Durrow, it
also appears in the listing of the canons in the Novum
Opus preface (fol. 2v). While the reversal is not discussed
in most of the contemporary Hiberno-Latin commen-
tary literature, it is a feature of the way that the table is
described in Ailerán’s well-known poem on the evangeli-
cal canons.³ The stanza on the eighth table (‘In the eighth
now a lion’s cub/ brings forth the words of God, and a calf/
whose apostolic number is computed together/ with Paul
added as a colleague’) places the symbol for Mark, the
lion, before that of Luke’s calf. This is the only instance in
Ailerán’s poem where the order of the canons is reversed,
also in keeping with the gospel books mentioned above. It
is interesting to note that while Ailerán’s poem appears as
a preface in several gospel books with Irish associations,
such as the Augsburg Gospels and Poitiers 17, their series
do not reflect the order of Canon VIII, as described in the
poem and presented in Mulling. It is also worth noting
that the reversal of Canon VIII as listed in Ailerán’s poem
is not a feature of the description of Canon VIII in the other
poems on the canons included in De Bruyne’s Préfaces de
la Bible Latine.¹
in Biblia Sacra Iuxta Vulgatam Versionem, ed. Robertus Weder, Stutt-
gart: Deutsche Bibelgesellschaft, 1983.
36There are slightly larger number of errors in the listing of num-
bers for Mt X.
37This reversal in the canon series has not been noted in scholar-
ship I have been able to access to this point. It is the subject of current
work by the author.
38Royal I. E. VI is available as a digital edition at http://www.bl.uk/
manuscripts/Viewer.aspx?ref=royal_ms_1_e_vi_fs001ar (last acces-
sed 13/04/2020).
39Ailerán’s poem has been reprinted many times. The earliest edi-
tion is in De Bruyne 1920, 185. See also Howlett 1996, 12–16. For the
way the order of the tables is explained by Hiberno-Latin commenta-
tors see Mullins 2014, 331–332.
40McGurk 1961, 64–66, 68–70.
41See De Bruyne 1920, 186.
In keeping with the level of care which is given in Mulling
to creating an accurate canon table text, the marginal
notation inserted alongside its gospel text is relatively con-
sistent. Because, as mentioned earlier, the manuscript’s
main text originally reflected only partially the opening
of Eusebian sections by the use of enlarged initials, the
scribe who added in the tables also strove to both adapt
the punctuation and enlarge the initials at the start of sec-
tions when this was absent from the original. McGurk pro-
vides a good example of this practice on fol. 18r column 2
line 25 which corresponds to Matthew Section 68 Canon
V, where two dots and a comma are additions and the fol-
lowing Et has been inked over for emphasis (Fig.8). ²
In addition to the level of care with which the gospel
text has been made to accommodate the apparatus,
the most striking thing about the system in Mulling is
the fact that the manuscript’s marginal references are
simple ones, without the parallel sections in the other
gospels. Preliminary comparison of the Mulling notation
with that of other contemporary gospel books reveals
some interesting results.³ Of the roughly 200 early Latin
gospel books which are extant, around 115 have no nota-
tion at all. This is largely to be explained by the fact that
Old Latin gospel books generally were without paratex-
tual material. At other times, as with a manuscript like
Armagh, the manuscript has parts of the apparatus but
no notation. There are over 70 manuscripts among the
set examined which contain the extended marginal nota-
tion, described above in the context of the MacDurnan
Gospels. The oldest example of this notation is that in
St Gall 1395, which has been discussed by O’Loughlin in
relation to the notation in the Book of Durrow. Another
early witness to this extended notation is to be found
in Aberdeen, University Library, Papyrus 2a, a papyrus
fragment of John’s Gospel written in rustic capitals in the
fifth century.
42See McGurk 1987, 173 and plates 3 and 9.
43The set of early Latin gospel books which forms the comparison
point here is based on those listed in Hugh Houghton’s Latin New
Testament and McGurk’s Latin Gospel Books with reference to Codi-
ces Latini Antiquiores. It has been difficult to determine absolutely
exact figures for the survey because of the limited availability of some
of these manuscripts as complete digital editions. While this is the
subject of ongoing work by the author, the overall results presented
here give a valid impression of the relatively limited transmission of
the simple system of notation.
44O’Loughlin 1999. For St Gall, Stiftsbiliothek, 1395 see CLA VII, 984,
https://elmss.nuigalway.ie/catalogue/157 (last accessed 13/04/2020).
A digital edition is available at http://www.e-codices.unifr.ch/de/
list/one/csg/1395 (last accessed 13/04/2020).
45See CLA II, 118, https://elmss.nuigalway.ie/catalogue/431 (last ac-
cessed 13/04/2020); Houghton 2016, 223; McGurk 1961, 24.
The Eusebian Apparatus in Irish Pocket Gospel Books 59
Fig. 8: Dublin, Trinity College Library, MS 60, Book of Mulling, fol. 18r. Marginal Entries for Matthew Sections LXIII–LXVIII. Image reproduced
courtesy of The Board of Trinity College Dublin.
60 Elizabeth Mullins
¹ ² ³
Surprisingly, of the 200 manuscripts examined during
this research, there are only eight clear examples of the
simple version of the notation, either in parts or through-
out the manuscript, which have been identified. These
manuscripts are listed in Table 2.
These manuscripts witness the reception of the simple
form of the Eusebian marginal notation in different ways.
There is a spectrum in this, from the notation in the Vero-
nensis, an original late antique manuscript containing an
Old Latin gospel text to which the Eusebian system was
added by a corrector, to later manuscripts which reflect
late antique models. An example of the latter is found in
the Codex Bigotianus made in southern England in the
eighth century, which reproduces the cola et commata
46CLA IV, 481, https://elmss.nuigalway.ie/catalogue/828 (last ac-
cessed 13/04/2020); McGurk 1961, 93; Houghton 2016, 212.
47Houghton 2016, 213–214; full digital edition from microfilm is av-
ailable at https://gallica.bnf.fr/ark:/12148/btv1b9065958t (last acces-
sed 13/04/2020).
48CLA 11, 1605, https://elmss.nuigalway.ie/catalogue/359 (last ac-
cessed 13/04/2020); McGurk 1961, 101; Houghton 2016, 215–216; Alex-
ander 1978, 64. See selected digital reproductions at http://www.helsin-
ki.fi/varieng/series/volumes/09/bleskina/ (last accessed 13/04/2020).
49Houghton 2016, 225. Full digital edition available at https://gallica.
bnf.fr/ark:/12148/btv1b10500013s/f33.image (last accessed 13/04/2020).
50CLA V, 684, https://elmss.nuigalway.ie/catalogue/1073 (last
ac cessed 13/04/2020); McGurk 1961, 64; Houghton, 2016, 225-226.
Full digital edition available at https://gallica.bnf.fr/ark:/12148/btv
1b8423842n (last ac ces sed 13/04/2020).
51See note 4.
52CLA V, 526, https://elmss.nuigalway.ie/catalogue/883 (last access-
ed 13/04/2020); McGurk 1961, 59-60; Houghton 2016, 267. Full digital
edition available at https://gallica.bnf.fr/ark:/12148/btv1b8492142v
(last accessed 13/04/2020).
53CLA VI, 730, https://elmss.nuigalway.ie/catalogue/1135; McGurk
1961, 53. Digital reproductions are available at https://bvmm.irht.cnrs.
fr/consult/consult.php?reproductionId=843 (last accessed 13/04/2020).
54One addition to this set which could be considered is Durham,
Cathedral Library, A. II. 10. This fragmentary manuscript includes
section numbers for Mark but no reference to the canon to which
these sections belong. See CLA II, 147, https://elmss.nuigalway.ie/
cat alogue/460 (last accessed 13/04/2020); Houghton 2016, 221.
layout, script and Eusebian sections of a much earlier
Italian Vulgate (Fig.9). Similarly, the Codex Sangerma-
nensis primus, copied in Saint-Germain-des-Prés in 810,
has been seen to rely on a pandect assembled in Rome in
the fifth century. Other manuscripts in the group reflect
a mishmash of models, Vulgate and Old Latin, some of
which had full marginal notation, others the simple
version, and others no notation at all. This mishmash of
different models is evident in the Leningrad Gospels from
southern England and the Codex Sangermanensis secun-
dus, which is thought to have been copied in Brittany in
the tenth century. The Avranches fragments present
an inconsistent approach: Avranches 71 and 48, which
contain a folio from Mark’s gospel and some fragments
from John respectively, have the initial for the evangelist
and section number with no canon reference, although
there may be some fading, particularly in the John frag-
ment; Avranches 66, which includes a folio from Luke’s
gospel, has the initial for the evangelist, section number
and canon reference in red on at least three occasions.
The St Gatien Gospels, copied in Brittany around the year
800, is perhaps closest to the kind of book that Mulling
may have relied on. While it is laid out as a block as
opposed to Mulling’s columns, it reflects many Insular
features and consistently includes the simple apparatus,
although it omits the first two elements of the Eusebian
system (Fig.10).
55The Leningrad Gospels uses simple notation in Mark, Luke and
John’s gospel. Matthew’s gospel begins with extended notation, but
this stops completely after Matthew 16. For discussion see Houghton
2010, 114. The Codex Sangermanensis secundus has no notation in
Mark’s gospel, simple notation in Matthew and Luke’s gospels and
extended notation in John.
56These include the references for Luke Sections 31–37. See the re-
production https://bvmm.irht.cnrs.fr/consult/consult.php?VUE_ID=
327527 (last accessed 13/04/2020).
57McGurk 1987, 175-176. On its decorative similarities to Insular
books see Alexander 1978, 78-79. He sees it as a ‘copy of a very sump-
tuous, probably Irish, Gospel book’.
Verona, Biblioteca Capitolare, VI Codex Veronensis, Gospel Book Vex, Italy
Paris, Bibliothèque nationale de France, latin Codex Sangermanensis primus, Bible c., St Germain-des-Prés
St Petersburg, National Library, F.v. I. Leningrad Gospels VIIIex, England
Paris, Bibliothèque nationale de France, latin Codex Sangermanensis secundus, Gospel Book X, Brittany
Paris, Bibliothèque nationale de France, nouv. acq. lat. St Gatien Gospels c., Brittany
Dublin, Trinity College Library, Book of Mulling, Gospel Book VIII–IX, Ireland
Paris Bibliothèque nationale de France, latin Codex Bigotianus, Gospel Book VIIIex , Southern England
Avranches, Bibliothèque Municipale, (fols I–II) +
(fols I–II) + (fols A-B) +Leningrad O.v.I..
Gospel Book Fragments VIII, Northumbria?
Tab. 2: Latin Gospel Books with Simple Eusebian Notation.
The Eusebian Apparatus in Irish Pocket Gospel Books 61
Fig. 9: Paris, Bibliothèque nationale de France, latin 281 + latin 298, Codex Bigotianus, fol. 12v. Marginal Entries for Matthew Sections XIII,
XIIII, XV. Image reproduced courtesy of the Bibliothèque nationale de France.
62 Elizabeth Mullins
Fig. 10: Paris, Bibliothèque nationale de France, latin 1587, St Gatien Gospels, fol. 7v. Marginal Entries for Matthew Sections LV–LXII. Image
reproduced courtesy of the Bibliothèque nationale de France.
The Eusebian Apparatus in Irish Pocket Gospel Books 63
The Insular connections of many of the manuscripts listed
here is significant. It supports a point made by Houghton
in relation to transmission of the gospel text in the Insular
world, where the move from Old Latin to Vulgate was the
result of a gradual accretion of readings. It seems likely
that the simple notation may have accompanied either a
stage in the original Old Latin or, more likely, early Vulgate
transmission of the gospels to Ireland and that the nota-
tion in Mulling is reflecting this. Indeed, it is worth con-
sidering whether the simple notation present in Mulling
is another feature of the ancient imported archetype,
posited by McGurk, which restricted its general prefaces
to Novum Opus and non-architectural tables.
Another perspective on the kind of marginal notation
used in early medieval Ireland is provided by the discus-
sion of the tables in the Hiberno-Latin commentary mate-
rial. While these texts generally avoid commenting on
the practical use of the system, their faithfulness to the
description of the series in Jerome’s Novum Opus preface
and in Isidore’s Etymologiae means that only the simple,
rather than extended, notation is the basis for the com-
mentary. In Vienna 940, also, interestingly, the uncial
titles which break up the commentary on Matthew do not
provide the parallel references in the other gospels, but
note purely that a particular section belongs to a particu-
lar canon, something else which suggests the use of the
simple system. It is difficult to draw firm conclusions in
relation to this, however. As the discussion of MacDurnan
demonstrated, extended marginal notation is present in
gospel books associated with medieval Ireland at this
time. Matthew Crawford has also noted the explicit dis-
cussion of the extended notation in Sedulius Scottus’
commentaries on the apparatus, although these are gen-
erally seen as later in date than the Hiberno-Latin texts
referred to by Bischoff, and Sedulius himself acknowl-
edges that the extended notation is not what is advised
by Jerome.¹
Why was the Eusebian apparatus added to the Book
of Mulling after it was created? Contemporary evidence for
the addition of prefatory material is provided by a manu-
script like the Burchard Gospels, which presents a sixth
century Italian text with the addition of prefatory mate-
rial in either a Northumbrian or Continental context in
the seventh or eighth centuries.² This general point aside
and turning specifically to Mulling, it is worth noting
58Houghton 2016, 74–77.
59McGurk 1990, 52–5 7.
60 For this and what follows see Mullins 2014.
61Crawford 2017, 80.
62McGurk 1961, 75–76.
that this prefatory material is not the only addition to the
manuscript. Possibly contemporary with the insertion
of the prefatory series, was the addition on fols 33v–34r
of a service for the visitation of the sick, while on fol.
94v, a liturgical text of 13 prayers in Latin and Irish was
inserted.³ This was followed by a circular drawing which
has most recently been interpreted by Dominique Bar-
bet-Massin as the visual representation of a daily prayer
sequence. The inclusion of this kind of material in the
book points perhaps towards a change in the book’s func-
tion, which allowed for a more personal and more schol-
arly use of the manuscript, a development which would
have lent itself to the introduction of the apparatus.
The broader monastic context needs also to be consid-
ered in the discussion of these additions. Barbet-Massin
in her work on the drawing and prayers in Mulling noted
the similarity between their content and that of the Fleury
prayer-book, a manuscript associated with Alcuin, which
shows a strong connection to Irish material. Barbet-Massin
attributes this Irish influence to Alcuin’s association with
the Irish monastic centre at Salzburg, particularly because
of his friendship with Arno—abbot there in the late eighth
century. The addition of the Eusebian apparatus to the
manuscript may reinforce the link between Mulling and
Salzburg made by Barbet-Massin in the context of the
prayers, for it is to the Salzburg/St Amand circle that many
of the Hiberno-Latin manuscripts, such as for example,
Vienna 940 are attributed. This manuscript, in an original
St Amand binding, includes as noted earlier the only other
set of non-architectural canon tables distributed across 15
pages, a discussion of the Eusebian system in its introduc-
tion, and continual reference to the apparatus throughout
its commentary text. The addition of the tables to Mulling
in the ninth century may thus reflect the kind of intense
interest in the series, which is evident in Vienna 940 and
other Hiberno-Latin manuscripts at this time.
4 Conclusion
Scholarship over the last two decades has created a
nuanced picture of the reception of the Eusebian appara-
tus in the Irish pocket gospel book context. The complete
absence of the apparatus and the inconsistency of atten-
tion to section initials within many of the gospel texts are
63Due to the faded quality of the script on fol. 94v scholars to this
point have not been able to precisely date this hand or to conclude
whether it is the same as that of the other additions. On this see Barbet-
Massin 2017, 161.
64For this and what follows see Barbet-Massin 2017.
64 Elizabeth Mullins
evidence of how the system was either not known or was
misunderstood at times. The careful addition of the three
elements of the Eusebian system to Mulling demonstrates,
however, that at other times there was a clear understand-
ing of the apparatus. This understanding is different to
that which appears in many of the larger contemporary
Insular gospel books, where the numerical tables are used
often as a moment for extended visual exegesis, some-
times, as in Kells, to the detriment of the system’s practi-
cal use. Mulling’s use of the apparatus has much more in
common with the way the system is treated in contempo-
rary Hiberno-Latin texts. This is reflected in, for example,
the care given to inserting correct entries to the apparatus
throughout the gospel’s text and margins, which parallels
the granular attention to the system, evident in a manu-
script like Vienna 940. Mulling’s use of simple rather than
extended marginal notation is in keeping with the sources
used for the discussion of the tables in the Hiberno-Latin
corpus of manuscripts. It also provides another piece of
evidence to support theories about the transmission of
the gospel text to early medieval Ireland. Although further
research is needed, the reversal of the order of the evange-
lists in Canon VIII noted here for the first time in Mulling,
Durrow and Kells and in Royal I E VI is a new piece of evi-
dence in the history of the transmission and development
of the Eusebian system in the Latin West. The extended
Eusebian notation included in the MacDurnan gospels
adds another layer to the discussion of the system in the
pocket gospel book corpus. MacDurnan’s approach, which
omits the first two elements of the apparatus but includes
the extended system of notation, leads to the interesting
possibility that this part of the apparatus may have been
used as a standalone exegetical tool. This is not something
which has been considered in scholarship up to this point
and it also requires further investigation, particularly in
relation to contemporary Hiberno-Latin exegetical mate-
rial. While there needs to be further research, it is clear
that this notation, both in its simple and extended form,
has a role to play in our understanding of the transmission
of the gospel text to both Ireland and the Continent in the
early medieval era.
References
Barbet-Massin, Dominique (2017), ‘The circular drawing in the Book
of Mulling and its relationship to the Fleury Prayer Book’, in
Rachel Moss, Felicity O’ Mahony and Jane Maxwell (eds), An
Insular Odyssey Manuscript Culture in Early Christian Ireland
and Beyond, Dublin: Four Courts Press, 159–174.
Bioletti, Susie and Allyson Smith (2017), ‘The pigments of early
Irish manuscripts’, in Rachel Moss, Felicity O’Mahony and
Jane Maxwell (eds), An Insular Odyssey Manuscript Culture in
Early Christian Ireland and Beyond, Dublin: Four Courts Press,
114–127.
Bischoff, Bernhard (1954), ‘Wendepunkte in der Geschichte der
lateinischen Exegese im Frühmittelalter’, Sacris Erudiri, 6,
189–281.
CLA II = Codices Latini Antiquiores. A Palaeographical Guide to Latin
Manuscripts Prior to the Ninth Century, edited by Elias Avery
Lowe: Great Britain and Ireland, Oxford 1935; 2nd edn 1972
(Codices Latini Antiquiores, II).
Commentarium In Lucam (1974), Joseph Kelly (ed.), Corpus
Christianorum Series Latina 108 C, Turnhout: Brepols.
Crawford, Matthew R. (2017), ‘Scholarly practices: the Eusebian
Canon Tables in the Hiberno-Latin Tradition’, in Giles E. M.
Gaspar, Francis Watson and Matthew R. Crawford (eds),
Producing Christian Culture Medieval Exegesis and its
Interpretative Genres, New York: Routledge, 65–125
DeBruyne, Donatien (1920), Préfaces de la Bible Latine, Namur: A
Godenne; reprint (2015), Prefaces to the Latin Bible, Turnhout:
Brepols.
Doyle, P. (1973), ‘The Text of St. Luke’s Gospel in the Book of
Mulling’, Proceedings of the Royal Irish Academy, 73, 177–200.
Endres, William (2017), ‘Canon II in the Book of Kells’, in Rachel
Moss, Felicity O’Mahony and Jane Maxwell (eds), An Insular
Odyssey Manuscript Culture in Early Christian Ireland and
Beyond, Dublin: Four Courts Press, 213–230.
Evangeliorum quattuor Codex Durmachensis (1960), A. A. Luce,
George Simms, Peter Meyer, Ludwig Bieler (eds), New York: Urs
Graf.
Farr, Carol A. (2011), ‘Irish Pocket Gospels in Anglo-Saxon England’
in Jane Roberts and Leslie Webster (eds), Anglo Saxon Traces,
Tempe: Arizona Center for Medieval and Renaissance Studies,
87–100.
Houghton, H. A. G. (2010), ‘The St Petersburg Insular Gospels:
Another Old Latin Witness’, The Journal of Theological Studies
NS, 61, 1, 110–127.
Houghton, H. A. G. (2016), The Latin New Testament, Oxford: Oxford
University Press.
Howlett, David (1996), ‘Seven studies in seventh century texts’,
Peritia, 10, 1–70.
McNamara, Martin (1990), Studies on Texts of Early Irish Latin
Gospels (A.D. 600-1200), Dordrecht: Kluwer.
McGurk, Patrick (1956), ‘The Irish Pocket Gospel Book’, Sacris
Erudiri, 8,2, 249–269.
McGurk, Patrick (1961), Latin Gospel Books from A.D. 400 to A.D.
800, Paris: Les Publications de Scriptorium.
McGurk, Patrick (1987), ‘The Gospel Book in the Celtic Lands before
AD850: Contents and Arrangement’, in Próinséas Ní Chatháin
and Michael Richter (eds), Ireland and Christendom: The Bible
and the Missions, Stuttgart: Klett-Cotta, 165–189.
The Eusebian Apparatus in Irish Pocket Gospel Books 65
McGurk, Patrick (1990), ‘Texts at the opening of the Book’ in Peter
Fox (ed.), The Book of Kells, MS 58, Trinity College Library
Dublin: commentary, Lucern: Urs Graf, 52–57.
Meehan, Bernard (2014), ‘The Book of Dimma’ in Rachel Moss (ed.),
Art and architecture of Ireland, I: Medieval c. 400-c. 1600,
Dublin: Royal Irish Academy.
Meehan, Bernard (2015), ‘Irish pocket gospel books’, in Claire Breay
and Bernard Meehan (eds), The St Cuthbert Gospels: studies
in the Insular manuscript (BL, Additional MS 8900), London:
British Library, 83–102.
Mullins, Elizabeth (2001), ‘The Insular Reception of the Eusebian
Canon Tables: Exegesis and Iconography’, 2 vols, Unpublished
PhD Dissertation, University College Cork.
Mullins, Elizabeth (2014), ‘The Eusebian Canon Tables and
Hiberno-Latin Exegesis: The case of Vienna, Österreichische
Nationalbibliothek, lat. 940’, Sacris Erudiri, 53, 323–344.
Neuman de Vegvar, Carol (2007), ‘Remembering Jerusalem:
architecture and meaning in Insular canon table archives’,
in Rachel Moss (ed.), Making and Meaning in Insular Art
Proceedings of the fifth international conference on Insular art
held at Trinity College Dublin, 25–28 August 2005, Dublin: Four
Courts Press, 252–256.
Nordenfalk, Carl (1977), Celtic and Anglo-Saxon Painting Book
Illumination in the British Isles 600-800, New York: George
Braziller.
Lawlor, H. J. (1897), Chapters on the Book of Mulling, Edinburgh:
Douglas.
O’Loughlin, Thomas (1999), ‘The Eusebian Apparatus in some
Vulgate Gospel Books’, Peritia, 13, 1–92.
O’Loughlin, Thomas (2007), ’Division systems for the Gospels: the
case of the Stowe St John (Dublin, RIA, D.II.3)’, Scriptorium, 61,
150–164.
O’Loughlin, Thomas (2010), ‘Harmonizing the Truth: Eusebius and
the Problem of the Four Gospels’, Traditio, 65, 1–29.
O’Reilly, Jennifer (1998a), ‘Patristic and Insular traditions of the
Evangelists: exegesis and iconography’, in Anna Maria Luiselli
Fadda and Eamónn Ó. Carragáin (eds), Le isole Britanniche e
Roma in età romanobarbarica, Rome: Herder, 49–94.
O’Reilly, Jennifer (1998b), ‘Gospel Harmony and the Names of
Christ Insular Images of a Patristic Theme’, in J. L. Sharpe III
and K. van Kampen (eds), The Bible as Book: The Manuscript
Tradition, London: British Library and Oak Knoll Press, 73–88.
Pulliam, Heather (2017), ‘Painting by Numbers: The Art of the Canon
Tables’, in Richard Gameson (ed.), The Lindisfarne Gospels New
Perspectives, Library of the Written Word, 57, The Manuscript
World, 9, Leiden, Boston: Brill, 112–133.
The Book of Kells, MS 58, Trinity College Library Dublin, (1990),
Peter Fox (ed.) Lucern: Faksimile Verlag.
