Sacred trails


Peter Harbison argues that the influences of continental frescoes could lead to a later dating of the Book of Kells

Two scenes in the Book of Kells ‒ Christ tempted on the Temple and Christ praying on the Mount of Olives ‒ mirror the revival of Late Antique narrative cycles under the Carolingian emperor Louis the Pious (814-840). Comparison with the subject-matter of his frescoes at Ingelheim on the Rhine (c.826) and those at Müstair in Switzerland (c.820-840) could favour a date in the late 820s for the Book of Kells. Its compositions for the pair of pictures may have been modelled on frescoes or mosaics on the Continent which may have provided the idea for the large full-page reproductions of narrative Gospel scenes in the Book, contrasting with the smaller scale usual in manuscripts of the period.

Artistic background
Among the rare early Irish manuscript examples of narrative illustrations from the Gospels, the Book of Kells distinguishes itself by having two of the largest (Figs 1 and 4). These are Christ on top of the Temple being tempted by the devil (fol. 202v), and what is wrongly called ‘The Arrest of Christ’, but is, more correctly (as the accompanying text tells us),(1) Christ on the Mount of Olives (fol. 114r), flanked by what we may presume are two apostles ‒ and not soldiers. These pages were both discussed in an article by the author over thirty years ago,(2) and this brief contribution may be regarded as a partially repetitive extension which provides new approaches to re-affirm much the same conclusions.

The pair of Kells pictures belong to a long tradition of Bible illustrations which brings us back some sixteen hundred years or more. Among the earliest surviving above-ground examples are small-size carvings on the fifth-century wooden door of the church of Santa Sabina in Rome.(3) But, around the same time, large-scale biblical pictures formed part of ecclesiastical fresco or mosaic cycles, such as those in an unidentified, yet probably Roman, church now lost, which were described by the Latin writer Prudentius in his Dittochaeum of around 400 A.D.(4) The text of his quatrains may have been modelled on tituli, brief poetic summaries of the contents of each picture for those whose Latin was superior to their knowledge of biblical iconography. Comparable cycles are found in the Roman basilica of Santa Maria Maggiore(5) in the fifth century, and in Ravenna in the sixth.(6)

These pictures on the walls of churches had to be large enough to be seen from below by the praying faithful, but by the sixth century the fashion was also being brought down to earth and practised on a much smaller scale in manuscripts such as the Cambridge Gospel fragment brought to England by St. Augustine of Canterbury,(7) where New Testament Passion scenes were enclosed by a dozen square frames all on a single page. The same system was adopted by ivory carvers associated with the emperor Charlemagne (771-814), at first on the cover of the Dagulf Psalter of c. 785,(8) and then on three other ivories ‒ one in the Bodleian Library in Oxford,(9) the Harrach diptych now in Cologne(10) and the Narbonne Cathedral plaque (11) (which, however, omits the frames) ‒ all recently ascribed by Kahsnitz(12) to Charlemagne’s Court School of around 800. Many of the scenes reproduced on them seem to copy Late Antique models of a few centuries earlier, though not reproducing any of the subjects in the Kells Codex, and the Nativity and Magi scenes on the Lorsch ivories (shared by the Vatican and the Victoria and Albert Museum in London(13)) are subordinate to the much larger individual figures above them so beloved of the emperor’s artists and craftsmen. With the exception of the Childhood of Christ and a few other rarer scenes on the ivories, it would seem that Charlemagne did not greatly favour the illustration of narrative scenes from the Gospels, particularly in the manuscripts produced during his reign. One of the few large New Testament scenes surviving from his time is the Transfiguration, Annunciation and Madonna mosaic above the arch in the Roman church of SS. Nereo and Achilleo(14) commissioned by his subservient ally, pope Leo III (795 – 816).

His papal successor was Paschal I, from whose reign (817-824) we have a small enamelled cross decorated entirely on one face by Childhood of Christ panels, together with a silver and gold cross-shaped reliquary with further Christ scenes, both preserved in the Museo Cristiano in the Vatican.(15) Roughly contemporary is the Carmen Paschale of Sedulius manuscript from Liège, now in Antwerp,(16) which shows traces of copying an Early Christian New Testament cycle available in England in the eighth century. These may both be taken as symptomatic of a Carolingian-period revival of narrative biblical scenes which takes place during the reign of Charlemagne’s son, Louis the Pious (814-840). Fine examples of the new trend are found in the miniature illustrations in three remarkable northern French Psalters bearing the names of Corbie (c.820(17)), Utrecht(18) and Stuttgart,(19) the latter two being normally dated to around 830. These by their very nature illustrate largely Old Testament material, though they do also occasionally stray into the New.

In church decoration, the revival of the old Roman style is best seen in the large-scale fresco cycles reproduced on the walls of churches after Charlemagne’s death.(20) Louis the Pious gave it full expression himself in the frescoes he commissioned for his palace at Ingelheim on the Rhine. No trace survives of the paintings, but we know of their contents through an extensive laudatory poem by Ermoldus Nigellus(21) which helps to date them roughly to the year 826, though some think that they may be as early as 819 or before. Equally lost are the frescoes in a church in St. Gall painted under abbot Hartmut around 870(22) which, again, we know about only through a literary description preserved in a manuscript in Zurich.(23) Some recognisable fragments survive in the north Italian church of San Salvatore in Brescia,(24) and in the southern Italian monastery of San Vincenzo al Volturno painted during the abbacy of Epiphanius (824-842). Florentine Mütherich(25) drew attention to the descriptive details of a New Testament picture cycle recorded in both Greek and Latin before 850 by an Irish monk in an unidentified Italian location and which she presumed had been copied from a manuscript, but the possibility of a fresco source should also be kept open. However, by far the best preserved Carolingian frescoes survive in the church of St. Johann in Müstair in the eastern Swiss canton of Graubünden,(26) the most recent research suggesting a date for them of around 820 to 840.(27) Here the walls are covered with a grid of framed scenes, which give extensive coverage to both Old and New Testament material. It is, I believe, significant that it is in these Roman and Carolingian frescoes, either lost or extant, and stretching over almost half a millennium, that we may find some of the closest reflections of our two narrative pages in the Book of Kells, though first looking briefly at the manuscript and ivory evidence.

Temptation of Christ on the Temple
The fourth chapters of the Gospels of both Matthew (1-11) and Luke (1-13) describe three separate temptations of Christ by the devil after forty days of fasting ‒ in the wilderness, on the pinnacle of the temple and on a high mountain. Other than the Kells depiction of Christ on the Temple (Fig 1), the Temptation makes its first surviving manuscript appearance on fol. 107 of the Stuttgart Psalter(28) (Fig 2),where Jesus is shown on the mountain, tempted on the summit by a black, winged and naked devil bearing a fire-hook (for hell-fire). He is also shown again half way down one side of the mountain being dismissed by Christ, as angels come up the other side to minister to the Saviour. Probably some decades later, around 850-855, we find all three temptations illustrated together in and around the initial letter D on fol. 41 of the Sacramentary of Drogo of Metz, now in the Bibliothèque Nationale in Paris.(29) The centre of the letter is occupied by Christ on the edge of the temple roof gesticulating to a very human-looking stick-holding devil below, clad in a lappeted garment contrasting with the grotesque naked creatures in the Stuttgart Psalter. A very contrasting ivory of the same period preserved in the Stadt- und Universitӓtsbibliothek in Frankfurt(30) (Fig. 3) shows a well-dressed but shoeless devil in a very much human form, holding a stick and gesturing towards an elegantly -robed and sandaled Christ who stands on the opposite side of a tree, probably to be understood as being in the wilderness. The Stuttgart devils, though on the mountain rather than the temple, are the nearest approximation to the Kells devil, also black, naked and holding the scarcely-visible fire-hook, but otherwise none of these manuscript or ivory delineations offers a satisfactory match for the Kells temple depiction, nor are they even likely to be any earlier than the Irish Temptation.

It is, however, not really clear whether or not we come any closer to the Kells picture when considering fresco representations of the Temptation, because none of these survives to give us a satisfactory comparison. It is widely accepted that one of the three vanished panels on the south wall at Müstair located between The Baptism of Christ and Jesus expelling the money-changers from the temple is likely to have portrayed a Temptation,(31) but which of the three temptations it may have depicted, we know not. The description by Ermoldus Nigellus(32) of the corresponding panel at Ingelheim tells of

How Christ as a man endured such long fasts
and skilfully laid low his tempter,
and a few lines later
took away the weapons of the demons and expelled him from afar.
(Translation by Peter Godman)

The fasting could suggest that the wilderness may have been meant, whereas the weapons could be compared to the fire-hooks of the mountain Temptation in the Stuttgart Psalter. Even if it is unclear which of the three temptations was depicted at Ingelheim, Ermoldus Nigellus at least makes it clear that a Temptation did feature there among the frescoes of around 826. Almost fifty years later, it recurred in the St. Gall frescoes,(33) where it was described as follows :

Christ rejected the vain temptation of the devil
And despised his pomps as nonsense

again without us learning from the description which of the three temptations was represented.

Christ on the Mount of Olives
The link between frescoes and the Book of Kells is, however, strengthened when we look at Christ on the Mount of Olives (Fig 4), a subject which ‒ unlike the Temptation ‒ does not seem to feature in either Carolingian manuscripts or ivories which survive. In his Dittochaeum, Prudentius(34) has a quatrain which might pertain :

Christ, from high on the summit of the olive-bearing mount
Returned to the Father, marking out the footprints of peace
Very fertile moisture flows from the eternal boughs
Which represents the poured-out gift of anointing oil to the lands of the earth.
(Translation Angela Malthouse)

While the last two lines can help to interpret the symbolism of the potted plants in the Kells picture, the words ‘Returned to the Father’ could hint that Prudentius was describing The Agony in the Garden, found also in the fourth-century ivory Lipsanothek in Brescia,(35) where Christ stands alone between two trees.

One mosaic (Fig 5) which Gertrud Schiller interpreted as The Agony in the Garden is found decorating a wall of the church of S. Apollinare Nuovo in Ravenna,(36) dating from the sixth century. It shows the nimbed Christ standing with half-raised fore-arms half-way up the mount, flanked by trees, and standing above eleven apostles, six on one side and five on the other. They all seem to be wide awake (nessun dorma!) which would argue against The Agony in the Garden but in favour of a different subject, Christ praying on the Mount of Olives, an identification supported by the fact that Christ is not seen to be suffering in any way.

The same may be said of a damaged panel in the church at Müstair,(37) of which enough fortunately survives (Fig 6) for us to see the nimbed Christ figure with head tilted towards his right, flanked by a tree on one side, and apparently having an unknown number of apostles beside him ‒ probably akin to the Ravenna mosaic ‒ but more than the pair on the Kells page, who may, however, represent pars pro toto.(38) The large scale of such a picture may have inspired the Kells painter to devote a whole full page to the reproduction of this narrative subject, almost uniquely in manuscripts of the period. The same could equally be said about the Temptation picture ‒ the scarcity of surviving fresco parallels helping to explain why scholars have never sought in them any possible origin for the compositions of the Kells pages.

The Ingelheim frescoes do not seem to have featured the Mount of Olives, or, at least, Ermoldus Nigellus makes no mention of it, and its absence from manuscripts and ivories surviving from the first half of the ninth century could favour the Kells painter having known a fresco or mosaic representation of the subject of the Ravenna or Müstair variety. Given that we have no knowledge of extensive narrative cycles at the time in the churches of Ireland or Britain, the considerable possibility of the Kells narrative subjects having been modelled on continental church frescoes (or even mosaics) could also imply that their painter may have travelled abroad in search of inspiration. But also given that these Kells pages fit in so well with the Carolingian Renaissance in reviving narrative cycle depictions starting with the Vatican cross and reliquary of 817-824, and continuing with the Antwerp Carmen Paschale of after 814, the Corbie Psalter of c.820 and the Ingelheim frescoes of around 826, we must seriously consider dating the Book of Kells to no earlier than the 820s, and possibly even towards their end. In this context, it should not be overlooked that Bernard Meehan noted a remarkable similarity between some Kells profile heads and one in an Armagh Priscian manuscript now in the library in the university of Leiden which can be precisely dated to the year 838.(39) Such considerations would, if anything, argue more in favour of Kells being the main location of the manuscript’s scriptorium, rather than Iona, which had suffered badly from Viking raids early in the ninth century, causing its monks to move their headquarters to Kells in 807, and finish a new church there by 814. What still remains a complete mystery, however, is why the scriptorium chose to confine itself to these two comparatively rare narrative scenes rather than depicting more of the better-known subjects.

PETER HARBISON is Honorary Academic Editor with the Royal Irish Academy


1 For the most recent view on the identification see Daniel McCarthy, ‘The illustration and text on the Book of Kells, Folio 114rv’, Studies in Iconography 35 ( 2014), 1-38. See also P. Harbison, ‘Christ praying on the Mount of Olives ‒ not the Arrest ‒ in the Book of Kells’, Archaeology Ireland 25 (2011), 9-12.
2 Peter Harbison, ‘Three miniatures in the Book of Kells’, Royal Irish Academy Proceedings 85 C (1985), 181-94.
3 A number of individual panels of the door are illustrated in Gertrud Schiller, Iconography of Christian Art (2 vols., London 1971-72).
4 J.-P. Migne (ed.), Patrologiae Cursus Completus, Series Latina Prior LX ( Paris, 1862), 90-111.
5 C. Pietrangeli, Santa Maria Maggiore in Roma ( Florence, 1988).
6 Giuseppe Bovini, Ravenna und seine Mosaiken (Munich, 1962); Giuseppe Cortesi, Sant’Apollinare Nuovo in Ravenna (Ravenna, 1975).
7 Francis Wormald, The miniatures in the Gospels of St. Augustine, Corpus Christi College MS 286 (Cambridge, 1954). See also Schiller, Iconography of Christian art (London, 1972), Vol. 2, Fig. 11.
8 Adolph Goldschmidt. Die Elfenbeinskulpturen aus der Zeit der karolingischen und sӓchsischen Kaiser VIII.-XI. Jahrhundert (4 vols., Berlin 1914; reprinted Oxford 1969), Vol. 1, 9, No. 3, with Taf. III.
9 Goldschmidt, Elfenbeinskulpturen, Vol. 1, 10-11, No. 5, with Taf. III.
10 Goldschmidt, Elfenbeinskulpturen, Vol. 1, 15, No. 18 with Taf. X.
11 Goldschmidt, Elfenbeinskulpturen, Vol. 1, 20, No. 31, with Taf. XV.
12 Rainer Kahsnitz, ‘ “Die Elfenbeinskulpturen der Ada Gruppe”. Hundert Jahre nach Adolph Goldschmidt. Versuch einer Bilanz der Forschung zu den Elfenbeinen Goldschmidt 1-39’, Zeitschrift des deutschen Vereins für Kunstwissenschaft 64 (2010), 9-72. My thanks to Charles Little for this reference.
13 Goldschmidt, Elfenbeinskulpturen, Vol. 1, 13-14, Nos. 13 and 14 with Taf. VII-VIII.
14 Schiller, Iconography of Christian art, Vol. I, 149 with Fig. 406.
15 César García de Castro Valdes (ed.), Signum Salutis. Cruces de orfebrería de los Siglos V al XII (Asturias [Oviedo], 2008, 228- 38. My thanks to Virginia Teahan for a copy of this work. See also Schiller, Iconography of Christian art, Vol. 1, 26 with Fig. 54, where a potential Palestinian origin for the enamelled cross is discussed.
16 Plantin-Moretus Museum, Ms. M 17.4. J.J.G. Alexander, Insular manuscripts 6th to the 9th century (London, 1978), 83, No. 65.
17 This date was suggested to me by Bernhard Bischoff in correspondence many years ago. See Bernard Meehan, ‘The Book of Kells and the Corbie Psalter (with a note on Harley 2788)’, in Toby Barnard, Dáibhí Ó Cróinín and Katharine Simms (eds), ‘A Miracle of Learning’, Studies in manuscripts and Irish learning. Essays in honour of William O’Sullivan (Aldershot and Brookfield [Vermont], 1998), 29-39, where a date of c.800 is suggested.
18 The Corbie (under the name Amiens) and Utrecht Psalters were discussed together in Wolfgang Braunfels, Die Welt der Karolinger und ihre Kunst (Munich, 1968), 156-179. The Corbie Psalter is preserved as Ms. 18 in the Bibliothèque Municipale in Amiens, and the Utrecht Psalter as Ms. 484 in the University Library in Utrecht.
19 Florentine Mütherich, Der Stuttgarter Bilderpsalter, Bibl. Fol. 23, Württembergisches Landesbibliothek (2 Vols., Stuttgart 1968).
20 For a recent overview of medieval frescoes see M. Exner, ‘Wandmalerei’, in Reallexikon der germanischen Altertumskunde 33 (Berlin/New York, 2006) , 220-31.
21 Ernst Dümmler (ed.), Monumenta Germaniae Historica, Poetae Latini Aevi Carolini (Berlin, 1884), Vol. II, 63-5. See also Peter Godman, Poetry of the Carolingian Renaissance (London, 1985), 250-8, and Walther Lammers, ‘Ein karolingisches Bildprogramm in der Aula Regia von Ingelheim’, in Festschrift für Hermann Heimpel, Verӧffentlichungen des Max-Planck-Instituts für Geschichte (Gӧttingen, 1972), 226-89.
22 Dümmler,Poetae Latini, 480. See also Hans Rudolf Sennhauser, Das Münster des Abtes Gozbert (816-837) und seine Ausmalung unter Hartmut (St. Gall, 1968).
23 Landesbibliothek, Zurich, Ms. C 48, fol. 48v-50r.
24 Barbara B. Anderson, The Frescoes of San Salvatore in Brescia, Doctoral Dissertation Berkeley 1976 (Michigan, 1977). ‘La Chiesa di San Salvatore in Brescia’, Atti del VIII Congresso di studi sull’arte dell Alto Medioevo (2 vols., Milan, 1962) Vol. 2, 7-205. Adriano Peroni, ‘San Salvatore in Brescia: un ciclo pittorico altomedievale revisitato’, Arte Medievale 1 (1983), 53-80. Werner Jacobsen, ‘San Salvatore in Brescia’, in Katharina Bierbrauer, Peter K. Klein and Willibald Sauerlӓnder (eds), Studien zur mittelalterlichen Kunst 800-1200, Festschrift für Florentine Mütherich zum 70. Geburtstag (Munich, 1985), 75-86.
25 ‘Das Verzeichnis eines griechischen Bilderzyklus in den St. Galler Codex 48’, Dumbarton Oaks Papers 41 (1987), 415-23.
26 Jürg Goll, Matthias Exner and Susanne Hirsch, Müstair. Die mittelalterlichen Wandbilder in der Klosterkirche, Müstair/Zurich 2007. See also Jürg Goll, ‘Les peintures murales romanes de Müstair (Suisse)’, in La peinture murale à l’époque romane, Les Cahiers de Saint-Michel de Cuxa XLVII (2016), Actes des XLVIIes Journées romanes de Cuxa, 6-11 juillet 2015, 175-92.
27 Goll et al.,Müstair, 107-9.
28 Schiller, Iconography of Christian art, Vol. I, 144 with Fig. 389.
29 Ms. lat. 9428. Schiller, Iconography of Christian art, Vol. 1, 143 with Fig. 390.
30 Goldschmidt, Elfenbeinskulpturen, Vol. 1, 42-43, No. 75. Schiller, Iconography of Christian art, Vol. 1, 143 with Fig. 391.
31 Goll et al., Müstair 95-6.
32 Godman, Poetry, 252-3, lines 233-8.
33 Sennhauser, Das Münster, 15.
34 Migne, Patrologiae Cursus Completus, 110, No. 44.
35 Schiller, Iconography of Christian art, Vol. 2, 48 with Fig. 10.
36 Schiller, Iconography of Christian art, Vol. 2, 48 with Fig. 141. Reproduced in colour in Cortesi, Sant’Apollinare Nuovo, page 62.
37 Goll et. al., Müstair, 48, No. 61k.
38 This Müstair panel adjoins one featuring The Arrest of Christ, showing that the Kells picture does not represent The Arrest!
39 Ms. BPL.67, fol. 34r. Bernard Meehan, The Book of Kells (London, 2012), 109 and 240, Fig. 238.