TWO PALAEOLOGAN MOSAIC ICONS

IN THE
DUMBARTON OAKS COLLECTION
P ORTABLE mosaic icons1 are among the rarest and most precious
objects of Byzantine art. This is the result not only of the ravages of
time; the chances of survival were better for mosaic icons than, for
instance, for jewelry and goldsmith's work which were so often destroyed
because of their material value; better also than for painted icons, because of
the greater durability of the medium. That we have so few of them is in great
part due to the fact that they were from the beginning a rare and costly species:
the entire production must have been very small.
In assessing the number of portable mosaic icons which have come down to
us, care must be taken to eliminate all mosaic icons or fragments thereof which
were originally wall mosaics and which became "portable" only when detached
from the wall, mostly in recent times. Quite a few of these have, at one time
or another, been mistaken for portable mosaics; e. g., the mosaic fragments
of the Virgin in the Museums of Palermo (from C a l a t a m a ~ r o and
) ~ Cortona
(from the demolished church of S. Andrea).3
An interesting species in themselves are the templorn mosaics, some of which
are still in situ-like those of the Porta Panagia at Trikkala, Thessaly4-while
others were removed from their original context-like those of the Xeno-
phontos monastery of Mt. A t h ~ sAll . ~ these mosaics share their chief charac-
teristics, especially the setting bed of mortar, with "normal" wall mosaics.
As a matter of fact, the nature of the setting bed seems to be the only safe
criterion for distinguishing portable mosaics, in the specific meaning of the
term, from fragments of wall mosaics: genuine portable mosaics are, without
exception, set in wax or resin on a wooden backing. The size of the cubes,
on the other hand, cannot be relied upon for defining this group: there are
portable mosaics with tesserae as large as those of normal wall mosaics,
whereas some wall mosaics show exceptionally small cubes, especially in the
faces of the figures.
While the size of the cubes does not serve to distinguish genuine portable
mosaics from fragments of wall mosaics, it does help to establish another
1 The following study is a combination of two lectures which were delivered at Dumbarton Oaks in
1951 and 1958 respectively. The first, about portable mosaic icons in general, appears here greatly
reduced: the second, on the two Dumbarton Oaks panels, is somewhat amplified. During the preparation
of the second lecture I enjoyed the kind help and advice of the Director and the scholars of Dumbarton
Oaks, especially of Professor Sirarpie Der Nersessian who conducted the Symposium of 1958.
The author is preparing a corpus of Byzantine portable mosaics. For valuable help in this enter-
prise he should like t o thank Mr. Boris Ermoloff of Paris and Prof. D. Talbot Rice of Edinburgh. A t
present, apart from handbooks and special publications, the following general treatments of this subject
may be consulted: E. Miintz, "Les mosaiques portatives," Bulletin monumental, LII (Caen, 1886),
p. 223ff.; D. T. Rice, "New Light on Byzantine Portative Mosaics," Apollo, XVIII (1933))p. 265ff.;
S. Bettini, "Appunti per lo studio dei mosaici portatili bizantini," Felix Ravenna, XLVI (1938-41))
p. 7ff.; 0. Demus, "Byzantinische Mosaikminiaturen," Phaidros, I11 (1947)) p. ~ g o f f .
0. Demus, The Mosaics of Norman Sicily (London, 1949)) pp. 189, 191, 311.
See A. Bernardini and A. Castri, Cortona, Guida Turistica (Arezzo, 1951)) p. 13.
S. Bettini, op. cit., p. 37, with bibliography and illustrations.
Ibid., p. 37.
90 OTTO D E M U S
distinction, namely that between two more or less clearly definable groups of
portable mosaics proper. The real distinction, it is true, is one of the size of
the icon itself rather than of single tesserae; but the size of the tesserae is
usually determined by the size of the entire panel-large icons using large,
small icons small, sometimes minute, tesserae. In any case, an analysis of the
measurements of all known portable mosaics reveals that there are G o clearly
distinct groups: a group of large icons which range in size from 23 by 34 to
62 by 95 centimeters, and a group of small icons which measure from 6 by 10
to 18 by 26 centimeters. The gap between the two groups is in fact much greater
than would appear from their respective dimensions, since the icons of the
"large" group represent without exception half-figures, while the larger icons
of the "small" group contain many-figured compositions. Thus, the scale of
the figures in the "large" icons is at least ten (sometimes even twenty) times
that of the figures in the "small" icons. I t is natural that minute tesserae were
used for composing the tiny figures of the "small" group.
Since we are dealing in this paper with icons of the "small" group, a few
words will suffice to describe the species of "large" portable icons. All tenor
so icons of this group, which are all that have come down to us, are half-figure
versions of greatly venerated prototypes, most of them bearing the distinctive
N N
names of the icons from which they are derived, like IC XC 0 EAEHM ON
(Berlin) or M"P &Y H ElllCKE't'IC (Athens). Actually, they are nothing but
mosaic reproductions of painted icons, and, as far as we know, were regarded,
treated, and used exactly like large-scale icons in painting. I t seems that they
were destined solely for ecclesiastic use, to be hung on the walls of a church or
to be displayed on tables (proskynetaria, analogia).6 The arrangement of the
tesserae in the earlier examples of these large mosaic icons, as, e. g., the Hode-
-
getria from the Pammakaristos church, now in the Patriarchate of Constantin-
ople,' an eleventh-century icon measuring 60 by 85 cms., is exactly the same
as that of wall mosaics-the nearest parallels to the Pammakaristos Hodegetria
being among the mosaics of St. Sophia in Kiev.8 True, the tesserae of the
Constantinople Virgin are somewhat smaller, but their arrangement in form-
defining, curvilinear rows of cubes, full of plastic tension, is the same as in
Kiev or in other eleventh-century wall mosaics. At a later date, however,
specific decorative effects made their appearance, effects which are foreign to
monumental art. A good example of this decorative style is another Hodegetria,
that of the Monastery of St. Catherine on Mount Sinai which belongs to the
thirteenth c e n t ~ r yThe . ~ golden highlights of the drapery (chrysographia), the
ornamentation of the background with rinceaux and other motifs, and the
complicated finesse of a technique employing infinitesimal cubes place the
See the narrative of Anthony of Kovgorod, in J. P. Richter, Quellen der byzantinischen Kunst-
geschichte (Vienna, 1897), p. 60.
' G. A. Sotiriou, '"H E ~ K C ~ V ~ a p p ~ [ ~ a p ~ m onu p, ' 'm ~ T?S
~ h 'A~aGqpias'ASqvQv, V I I I (1933),
p. 359ff.; A. M. Schneider, Byzanz, Istanbuler Forschungen, V I I I (1936))p. 41, pl. 7.
8 0. Powstenko, The Cathedral of St. Sofihia i n Kiev (New York, 1gj4), pl. 45.
G. and M. Sotiriou, Ic6nes du Mont Sinai (Athens, 1g5g), p. 8j, pl. 71 ; 0. Demus, "Die Entstehung
des Palaeologenstils in der Malerei," Berichte zum X I . Internat. Byz. Kongvess, IV, 2 (Munich, 1958),
P. 55.
 T W O P A L A E O L O G A N MOSAIC I C O N S 91
Sinai icon in a class entirely distinct from that of the Constantinople Virgin.
Both form and technique differ widely from those of wall mosaic; the artist
seems rather to have aimed at producing the combined effects of panel painting,
book illumination, and enamel.
This mixture of technical styles and consequent wealth of effects is not only
indicative of an experimental period, the formative phase of a new style, the
Palaeologan; it also foreshadows the distinctive style of the second group of
mosaic icons, that of "small" portable mosaics which may be properly called
miniature mosaics. The Sinai icon is, in fact, nearest to these miniature mosaics
in both style and technique as well as in its dimensions; measuring as it does 23
by 34 cms., it is the smallest known icon of the "large" group, the only mosaic
icon that can be said to occupy an intermediary position between the two
groups. Its position in time is also somewhat intermediary: belonging to the
thirteenth century, it stands midway between the period in which most of the
icons of the "large" group originated-the eleventh century-and that which
saw the flowering of the miniature mosaics, the fourteenth century.
This does not mean, of course, that the Sinai icon precedes in date all portable
mosaics of the "small" group. A few of the latter are somewhat earlier and
there is at least one miniature mosaic which is considerably older than the
Hodegetria: the tiny panel, in the Sinai Monastery, with the half figure of
St. Demetrius,lowhich displays the style of the first half of the twelfth century,
but may be somewhat later. If not the earliest, this icon is at least one of the
oldest of the "small" group. There is, as yet, little in this panel to indicate the
direction which the stylistic development of the species was to take in the
following century. Apart from the pattern of cubes depicting the Saint's coat
of mail, the technique is almost that of wall mosaic (excepting, of course, the
minute size of the tesserae), and, were it not for the almost fragile delicacy of
the elongated face and narrow shoulders, the figure might be said to exhibit a
monumental style. To see in these characteristics and in the decorative pattern
of the armour an influence from the realm of book illumination is, perhaps, not
quite justified; but such influences certainly made themselves felt in the first
half of the thirteenth century, and it was in the course of that century that
the art of miniature mosaic developed its own technique, its own style.
In the most characteristic specimens of this group of minute mosaics, the
tesserae, about half a millimeter square (or even smaller), are set so close
together that the interstices are scarcely visible. The tesserae are mostly of
enamel paste; certain colors, however, are rendered by semiprecious stones
like lapis lazuli and malachite. The gold and silver cubes consist, in some cases,
not of glass with metal foil, but of solid metal. The background is rarely made
up of golden tesserae alone: checkered patterns of three or four colors frame
or fill part of both the ground and the haloes. Everything contributes towards a
colorful, rich, and precious effect.
An analysis of the iconographic data, not only of the extant miniature
mosaics but also of those that are adequately described in literary sources,
G. and M. Sotiriou, Icdnes, p. 84, pl. 70.
92 OTTO DEMUS
shows a preponderance of single figures, either full-length or busts. The single
figure most often represented is that of the Virgin, followed, in the order of
the frequency of their representation, by Christ, St. Nicholas, John the Baptist,
Michael, Demetrius, Theodore, and George. Other single figures represented are
St. Anne, Basil, Daniel, John Chrysostom, John the Evangelist, Peter and
Samuel. Thus, the Saints of the Empire (Nicholas), of Constantinople (John
the Baptist), and of the court (Michael, Demetrius, Theodore, George) are the
subjects most frequently represented.
As for representations of scenes (we know of about a dozen icons of this
kind), with one exception these are restricted to the great feasts of the
Church-the exception being the Forty Martyrs of Sebaste, which is not really
a scene but a collective portrait of saints.
Thus the subject matter represented in miniature mosaics presents no
exceptional features; on the contrary, only the most widely venerated subjects
are depicted. Nor is there anything in the iconographic treatment to suggest
that any of these icons originated in a provincial atmosphere: their icono-
graphical purism points rather to Constantinople itself or, at least, to one of
the great centers of Byzantine art in the thirteenth and fourteenth centuries.
As befits the small scale of these icons, they do not represent the "great,"
"public," or monumental themes, as, e. g. the Pantokrator, the enthroned
Virgin, etc., but rather the standing Saviour, the Eleousa, or the Glyko-
philousa; the saints are introduced as patrons, in frontal half figures, or as
intercessors, standing in a three-quarter posture, to transmit the prayers of
the worshippers to Christ. In short, these icons belong, thematically as well
as artistically, to an intimate form of art, destined for the chosen few, for a
select upper class.
There are other features which substantiate the conviction that the miniature
mosaics are, indeed, products of a court art, that the whole genre belongs to
the aulic sphere; such as, for instance, the frames of the icons, most of which,
judging from preserved examples, seem to have been made of chased silver.
Several of the frames that have come down to us contain figural representations,
usually full figures or busts of the apostles and other saints;ll in the case of
the Crucifixion icon of Vatopedi the frame is decorated with a complete cycle
of the twelve great church~festivals.12The frame of another Vatopedi icon,
with the standing figure of St. Anne, presents a full devotional programme, with
busts and full figures of saints, and with the Hetoimasia worshipped by angels.13
Some of these icons have especially interesting frames: the frame of the icon
of St. John the Evangelist in the Great Lavra of Mt. Athos,14 is decorated
11 See the icons of Esphigmenou, Lavra, Patmos, and Vatopedi (St. Anne). On Byzantine silver
frames in general see A. Bank in Vizantiiskii Vremennik, XI11 (1957)) p. 211ff.
l a W. Felicetti-Liebenfels, Geschichte der byzantinischen Ikonenmalerei (Olten-Lausanne, 1956), pl.
75: p. 66. The sequence of the feasts is here somewhat confused; moreover, the Annunciation occurs
twlce, while the Crucifixion is missing.
l 3 0. Wnlff and M. Alpatov, Denkmaler der Ikonenmalerei (Hellerau, 1gz5),p. 56, fig. 18 ;W. Felicetti-
Liebenfels, ofi. cit., p. 64, pl. 74.
) , 114ff.,
14 N. P. Kondakov, P a m i a t n i k i khristianskogo iskusstva n u Afone (St. Petersburg, I ~ O Z p.

PI. 34.
 TWO PALAEOLOGAN MOSAIC ICONS 93
with ten, perhaps earlier, enamel medallions, showing the Hetoimasia (in the
upper center) and nine busts, representing (in the bottom row) St. John the
Baptist with his parents and (above) six Saints bearing the name of John.
A similar arrangement, with portraits of the synonymoi of the Forerunner
surrounding an icon of the Birth of the Baptist, is described by Manuel
Philes,15 which suggests that the frames of precious icons conformed to certain
types.
Another most interesting frame is that of an icon of St. Demetrius in Sasso-
ferrato. Its decoration, with emblems of the Palaeologi and with an inscription,
the historical significance of which has been studied by the late A. A. Vasiliev,
suggests that this icon was produced at Thessalonica, most probably for a
member of the Palaeologan dynasty.16
The Sassoferrato icon is not the only one that can be traced back to imperial
ownership or patronage. The St. John mosaic in Lavra,f7 the St. Anne in
Vatopedi,18 the Virgin in Sta Maria della Salute in Venice,lg and the two
panels representing the twelve great feasts in the Opera del Duomo of Flo-
rence20-all, according to tradition, were imperial gifts. In some cases the
representations themselves suggest that the panels were commissioned by, or
made as gifts for, emperors. George, Theodore, Demetrius, and Michael, - holy
warriors and special protectors of the emperors-would, of course, have been
their favorites. I t is noteworthy that the court poet Manuel Philes chose several
icons of St. George and St. Michael as subjects of his poetic descriptions
( ( ~ K ~ P & o E I s ) ; another poet of the time, Markos Eugenikos, described icons of
St. Demetrius.21 It is, perhaps, especially significant that the archangel
Michael, the holy namesake of Michael VIII Palaeologus who reconquered
Constantinople in 1261, is found six times among the seventy odd mosaic
icons which we know through both extant works and literary sources. The
prophet Daniel, too, who is occasionally represented on portable mosaics and
whose icons Manuel Philes describes in his poems, has a certain connection
with Michael VIII Palaeologus. That Emperor was called a "new Daniel" and
he applied to himself, with the help of puns on his own name as well as on that
of his dynasty, the prophecies of Daniel VII : 9 and X : 13, 21, in which the

15 Manuelis Philae carmina, ed. E. Miller, I (Paris, 1855)~p. 58, no. CXXXIII.
16 S. Bettini, 09.cit., p. ~ g f fA.
. ;A. Vasiliev, "The Historical Significance of the Mosaic of St. De-
metrius at Sassoferrato," Dumbarton Oaks Papers, 5 (1950), p. 31 ff., with bibiography.
17 N. P. Kondakov, Pamiatniki, op. cit., p. 114f.Tradition connects this icon with John Tzimiskes,
a date which is, of course, much too early.
l8 Ibid., p. 113.
19 S. Bettini, op. cit., p.off.; 0. Demus, "Die Entstehung", p. 16, note 70.
20 Reproduced in color by A. Grabar, La peinture byzantine (Geneva, 1953)~p. 191. For the story
of the dedication see A. F. Gori, Thesaurus veterum diptychorum I11 (Florence, 1759), p. 320.An icon
(or a pair of icons ? ) with the twelve feasts, described by Manuel Philes (ed. E. Miller op. cit., p. 9,
no. XXIV), was dedicated by John Kanabes.
21 Manuelis P h i l a e carmina, I, pp. 36, 46f., 317f., 357f., 457, 460; 11, pp. 202, 287f., 415. On
ekphraseis in general see A. Muiioz, "Alcune fonti letterarie per la storia dell'arte bizantina," N. Bol-
lettino di Archeologia Cristiana, X (1go4), p. 221ff.; idem, "Descrizioni di opere d'arte in un poeta
bizantino del secolo XIV (Manuel Philes)," Repertorium fur Kunstwissenschaft, XXVII (1go4), p. 309ff. ;
idem, "Le Ekphraseis nellaletteratura bizantina e i lor0 rapporti con l'arte figurata," Recueil Kondakov
(Prague, 1926), p. 139.
 94 OTTO D E M U S
VrraAalbs TGV fiyepGv and "Prince Michael" appear as the saviors of the Chosen
People. 22
These considerations lead us to believe that miniature mosaics were a specific
genus of imperial art. Unfortunately, the written sources do not tell us anything
further of the ri3le that portable mosaic icons played in the imperial household.
There is, for instance, no justification for identifying them with the precious
"ergomoukia" which, according to Constantine Porphyrogenitus, were kept,
together with the crown insignia, in the Pentapyrgion, a five-turreted cupboard
in the Imperial Palace.23This strange word has defied translation, and Labarte's
suggestion that it is a scribe's error for "ergomouzakia," and means "portable
mosaics," is entirely unf0unded.2~
Kondakov proposed a somewhat different interpretation of this puzzling
word: according to him "ergomoukia" ("une abbreviation de Ergomouzakia")
were mosaics made of glass which, therefore, resembled enamels-a supposition
no more tenable than L a b a r t e ' ~I.t~is~ much more likely that the "ergomoukia"
were embroideries; according to Constantine Porphyrogenitus they were hung
in the P e n t a p y r g i ~ nwhereas
,~~ mosaic icons would, in all likelihood, have been
stored lying in their wooden boxes, one of which is (or was) preserved in
Sassof err at^.^'
A few other alleged references to miniature mosaics should also be discounted,
so that we have, in fact, very few specific references to them in contemporary
sources. The most interesting occur in the works of Markos Eugenikos and
Manuel Philes who have left us several poetic descriptions of such icons, one
or two of which may refer to works that have actually survived.28The wording
of these descriptions is a strong indication that the miniature mosaic icons were
considered as precious gifts and used mainly for private worship.
Not a single mosaic icon is dated. The frame of one, the Eleousa of Sta Maria
della Salute invenice, bears, it is true, an inscription on its reverse, giving the
date 1115, the name of the artist, and other details, and this has been taken
seriously by a number of authors.29But the frame is of the fifteenth century
2 2 See A. Martini, "iblanuelis Philae carmina inedita," Atti della R. Accademia di Archeologza, Letteve
e Belle Arti, XX (Supplement, goo), p. 46; Manuelis Philae carmina, ed. E. Miller, I, p. 50, etc. For
the identification of Michael V I I I with Daniel see C. Sathas in Revue arche'ologique, I (1877), p. 99.
23 J. Ebersolt, Le Grand Palais de Constantinople et le Livre des Ce're'monies (Paris, I ~ I O ) ,p. 82.
24 J. Labarte, Histoire des arts industviels, I1 (Paris, 1864), p. 47.
25 N. P. Kondakov, Histoire et monuments des e'maux byzantins (Frankfurt am Main, 18gz), p. 102.
Another Greek term, aapohla, was also taken to mean mosaic icons: W. Nissen, Die Diataxis des
Michael Atfaliates con 1077 (Jena, 1894). The correct interpretation of this word, which meant icons
of copper, has been given by S. Vryonis, "The Will of a Provincial Magnate, Eustathius Boilas (105g),"
Dumbarton Oaks Papers, 11 (1957), p. 268.
26 De Cerimoniis, Bonn ed., I, p. 582.
2 7 L. Serra, L'Arte nelle Marche, I (Pesaro, 1929), p. 344.
28 See, e. g., Manuelis P h i l a e carmina, I, p. g ; A. Muiioz, "Descrizioni," p. 3goff.
29 SO S. Bettini, op. cit., p. off.; D. T. Rice, OF. cit., p. 265, as against V. N. Lazarev, "Byzantine
Icons in the Fourteenth and Fifteenth Centuries," T h e Buvlington Magazine, LXXI (1937), p. 250,
who rightly regards the inscription as apocryphal.
If there is anything at all genuine in the inscription, it may refer t o the year 1315, not 1x15; the
name Manuel (allegedly Manuel I, whose dates, however, do not correspond t o 111j,since he reigned
from 1143 t o 1180) may originally have designated the despot Manuel Palaeologus. But, as the entire
inscription is an almost incredible farrago of nonsense, it is more likely that it was made up in Venice
without any basis whatsoever.
 T W O P A L A E O L O G A N MOSAIC I C O N S 95
and the inscription is a naive and highly phantastic invention of that time;
moreover, the style of the icon is indubitably Palaeologan. The fanciful in-
scription on the reliquary in which the icon of the Saviour in Sta Maria in
Campitelli, in Rome, is now encased is equally valueless.30 This inscription
claims that the reliquary was the altare viaticum of St. Gregory of Nazianzus;
in addition, it gives the name of a goldsmith, "GG," that is George or Gregory,
who has been tentatively identified with an artist of that name mentioned in
an inscription of 1117. However, the identification is purely conjectural and
the mosaic panel itself seems to have become a part of the reliquary only in
the seventeenth century. Its style points to the middle of the thirteenth century.
Other traditions, too, like the one that connects the mosaic icon of St. John
the Evangelist in Lavra with the Emperor John Tzimiskes at the end of the
tenth century, have to be discounted as having no basis whatever.31 I t is only
in the fourteenth century that we are on firmer ground: the Sassoferrato icon
can be connected with one of the Palaeologan Emperors, possibly Michael VIII
or Andronicus 11,32 and the two magnificent mosaics with the twelve feasts, in
F l ~ r e n c emust
, ~ ~ have been made before 1394) when they were given to S. Gio-
vanni by a Venetian-born lady, the widow of a cubicularius of the Emperor
John Cantacuzenus. As it is most likely that she or her husband acquired the
two panels before the downfall of their patron in 1354155, this date would appear
to be a safe terminus ante quem for the mosaics-a terminus which is not very
helpful since the style of the two mosaics places them, in any case, in the first
half of the fourteenth century.
The time of John Cantacuzenus seems, generally speaking, to mark the end
of the period in which portable mosaic icons were made: about the middle of
the fourteenth century Cantacuzenus ordered the icons throughout the Empire
to be stripped of their jewels, thringia (i. e. silver mountings), and frames so
that the precious metals might be melted down for coinage.34
M?e have no such convenient record to connect with the beginning of the art
of miniature mosaic so that we are thrown back on stylistic dating. Judging
from the approximate dates that can be assigned to the extant portable mosaics
of the "small" variety, the majority of these seems to have originated in the
two generations from about 1260 to 1320; a few are earlier, one (the St. De-
metrius of Sinai) even belongs to the twelfth century. As in the case of some
other branches of Byzantine art, e. g. sepulchral sculpture35and painted icons,
it was during the first half of the thirteenth century that the art form of
miniature mosaic was fully realized; and as in those, the patronage of western
barons and clerics may have played a part not only in keeping the technique
alive, but in elaborating this specific form of art.
30 A. Colasanti, "Reliquiari medioevaliin chiese Romane," Dedalo, XI11 (1933) p. 288, fig. on p. 293;
A. Valente, "Intorno ad un orafo del secolo XII," Bollettino d'arte, XXXI (1937-38), p. 261 ff.
31 K.P. Kondakov, Pamiatniki, p. 114.
32 See supra, note 16.
33 See supra, note 20.
34 J. Ebersolt, Les arts somptuaires de Byzance (Paris, 1923)) p. 108 (after Nicephorus Gregoras,
Bonn ed., 11, p. 748).
3"hus A. Grabar in a lecture held in Dumbarton Oaks in May 1958.
96 OTTO DEMUS
With a life-span of scarcely more than a century, the art of miniature mosaic
was one of the most short-lived art forms of Byzantium. This factor as well as
the costliness of the materials and the complexity of the techniq~e,3~ makes it
certain that the production was not a very large one. Of course, only a part of
this production is known to us from extant works and from data preserved in
written sources; among the latter, a few inventories are the most valuable. The
richest collections of portable mosaic icons, both large and small, that were ever
formed were those of Pope Paul I1 and of Lorenzo Medici (il Magnifico), the
inventories of which have come down to us. The difficulty is, however, that
the term "mosaic" was often used loosely, being also applied to enamel. Of
several icons mentioned in these and other inventories (e.g. that donated to
St. Peter's by Cardinal Bessarion) it is not even certain that they were of
mosaic and not painted;37 nor do we know in all cases of reliably reported
portable mosaics whether these belonged to the large or to the small variety.
However, it may be said with all due caution that we have data concerning
thirty to forty miniature mosaics that have perished and that twenty-nine are
now extant; eight of these are in Italy, five in Russia, four on Mt. Athos;
whereas other countries possess only one or two such icons. All mosaic icons
are at present in the possession of churches or museums so that, unless
unknown ones are discovered, it is hardly likely that any private or public
collection will henceforth be able to acquire any of these precious panels.
All the more remarkable, therefore, is the acquisition of two of the finest
miniature mosaics by the Dumbarton Oaks Collection, in 1947 and 1954.
I t needed luck and acumen, assisted by generosity, to secure for the Col-
lection these two icons which are, perhaps, the last of their kind to change
hands.

THE ICON OF THE FORTY MARTYRS OF S E B A S T E ~ ~

(Figures 1-3)
The wooden panel which has lost its metal frame measures 16 by 22 cms.
The mosaic surface has suffered seriously; there are a great many cracks which
crisscross the entire area but which are most noticeable in the upper third
of the area. In this part there are also two large gaps and a number of
smaller ones. One of the two large gaps is on the left-hand side, in the gold
ground above the heads of the figures: it affects the upper contours of four
36 The claim of the apocryphal inscription on the back of the Sta Maria della Salute icon (see supra,
note 29) that its author, "Quidam Theodorus Constantinopolitanus," spent twenty years making
this and another mosaic of the same size is, of course, just as much nonsense as the affirmation that
the artist was created baron by the Emperor for this achievement. The latter motif might be derived
from the story told by the Continuator of Theophanes (Bonn ed., p. 452) about Theodoros Belonas.
37 E. Muntz, "Les mosa'iquesportatives," p. 229ff. ; idem, Les arts ri la Cour des Papes, I1 (Paris, 1879)~
pp. 142, 201 ff.; E. Muntz and A. L. Frothingham, I1 Tesoro della Basilica d i S. Pietro in Vaticano dal
X I I I a1 X V secolo (Rome, 1883)) p. 111ff.
38 Handbook of the Dumbarton Oaks Collection (Washington, D. C., 1955))no. 290, p. 147; 0. Demus,
"An unknown Mosaic Icon of the Palaeologan Epoch," Byzantina Metabyzantina, I/I (New York,
19461,P. 107ff.
 T W O P A L A E O L O G A N MOSAIC I C O N S
heads and the part of the inscription between the i- and the second C of 01
AT[! 01 TECICAP A KO NTA. The damage on the right-hand side is more serious:
its area is about one-ninth of the entire surface and extends over a part of the
composition, blotting out one of the most important iconographic details of the
scene. However, the main group of the Martyrs, that is thirty-nine figures or
heads, is very well preserved, with only a few tesserae missing where the cracks
intersect one another. The Martyrs stand in three rows, the foremost one
consisting of full figures, the middle one of busts, and the upper one of hardly
more than heads. The ice on which the figures stand is olive green, grey, and
white, with brownish rocks in the foreground; the loin cloths are shaded in
subdued colors: ochre, grey, pink, emerald green, olive, blue, and white, with
the most vivid tones concentrated in the figure (third from the left) of the
youthful saint who upholds the collapsing body of his neighbor. The flesh tones
run the whole gamut from grey to pink, ochre, brown, and olive in most delicate
shades, with white highlights emphasizing the plastic modelling. The color of
the hair varies from white to grey and blue, from yellow to ochre and from
brown to black. The inscription is traced in black, the diadems which descend
on the Martyrs have red and blue stones in gold settings outlined in black, the
heavenly arc in the middle of the upper margin is divided into blue, grey, and
white concentric rings. The color effect of the whole panel is rather delicate and
subdued, with golden-brown and silver-grey as the resulting general tones.
The provenience of the icon cannot be traced back very far: it is said to have
entered the Segredakis collection from the possession of a Greek refugee from
Asia Minor; from Segredakis it passed to Danos from whom it was acquired
by Hayford Peirce (1931) ; after the latter's death it was given in his memory
to the Dumbarton Oaks Collection in 1947.
The subject matter of the icon confirms our observations concerning the
aulic character of the iconography of the entire species of miniature mosaic.
The veneration of the Forty Martyrs, while popular and widespread enough,
had its center at the court of Constantinople and among the upper ranks of the
Byzantine army. The fact that legend made the martyrs of Sebaste (or Sebastia)
soldiers of the Roman army, who suffered martyrdom for their faith under
Maximianus or Licinius, caused them to be venerated as warrior saints.
An ivory triptych, in Leningrad, accordingly presents them accompanied by
the greatest military saints of Byzantium, SS. George, Demetrius, the two
Theodores, and others39 (fig. 13). Nine shrines of the Forty Martyrs are known
to have existed in Constantinople and its environs alone, some of them inti-
mately connected with the court .40
The veneration of the steadfast soldiers and of their relics began in very
early times, perhaps immediately after their martyrdom. Oddly enough, the
martyrs themselves seem to have, quite consciously, provided a firm basis for
the cult of their relics; at least they were credited with having done so at an
39 A. Goldschmidt and K. Weitzmann, Die byzantinischen Elfenbeznskulfituren des X . - X I I I . Jahr-
hunderts, I1 (Berlin, 1930/34), p. 27, pl. 3. Concerning warrior saints on ivory triptychs see E. Kanto-
rowicz "Ivories and Litanies," Journal of the Warburg and Courtauld Institutes, V (194z), p. 56ff.
40 R. Janin, "Les Cglises byzantines des saints militaires," Echos d'orient, XXXIV (1935))p. 64ff.

7
98 OTTO D E M U S
early period. In their so-called ' 'testament" (diatheke), a most curious and
valuable document of the early Church which has bein favorably evaluated
by modern critics,*l they are said to have implored their relatives and fellow
Christians to collect their remains after their death, and above all, not to
separate them from each other, but to bury them together in a certain place,
since they wished to remain together in death as they were in their last struggle,
their tryGv. This moving document purporting to have been penned by three
of their number, is signed by all forty. Whether genuine or not, the diatheke
is in any case a contemporary document and presents, if not the thoughts of
the martyrs themselves, at least the ideas of their friends and fellow Christians.
The main idea is that these forty persons of very different ages (a fact which
speaks against their having been soldiers) who came from different parts of the
Empire, had in their last hours found a new cumulative personality which they
wanted to preserve in death and in after-life. In this they succeeded: they
became and remained for all time, the Forty Martyrs, the Hagioi Tessarakonta.
However, their wish, alleged or real, to remain together in the physical sense,
was not granted: in a time of growing relic worship, this was hardly to be
expected. From their first resting place, Sareim, a village near Sebaste, some
of the relics were taken to Jerusalem, others to Constantinople, to Rome, and
elsewhere; to-day, they are widely di~persed.~2
Not long after the martyrdom, some of the relics came into the possession
of a Cappadocian lady, Emmeleia, who had a shrine built for them on her
domain, near Caesarea. The dedication festival of this church was attended by
two of her sons; the younger, Gregory, was not at first overly willing to be
present and would have preferred to shirk the long walk to the church, but
a dream he had during the night before the festival made him change his mind.
He dreamed that he was threatened and all but beaten by a number of soldiers
whom he finally recognized as the Forty Martyrs. His respect for these energetic
saints, enforced in so drastic a manner, persisted throughout the life of Gregory,
later surnamed "of Nyssa" : he has left us three homilies in honor of the Forty,
in one of which (the third) he describes his dream.# These homilies and the
twentieth homily of Gregory's older brother Basi144 seem to be the earliest
extant sources of a legend which had grown within the two generations between
the martyrdom of the Forty and the date of the sermons, the seventies of
the fourth century. The central motif that underlies these sermons is the pre-
servation of the number forty, that is, of the cumulative personality of the
Hagioi Tessarakonta. St. Basil's homily is nothing more than an elaboration
41 G. N. Bonwetsch, "Das Testament der 40 hlartyrer," Studien zur Geschichte der Theologie u n d
Kzrche, I/I (Leipzig, 1897), p. 71 ff. Cf. also the following note.
4 2 The legend in A A S S , X Mart., p. 12ff.; 0. v. Gebhardt, Acta m a r t y r u m selecta (Berlin, 1902),
p. 171ff.; W. Weik, "Die syrische Legende der 40 Martyrer von Sebaste," BZ, XXI ( I ~ I Z ) ,p. 76ff.;
P. Franchi de Cavalieri, "Note Agiografiche," Studi e Testi, XX/3 (19og), p. 64ff.; idem, "I Santi
Quaranta Martiri di Sebastia," Studi e Testi, XLIX/7 (1928),p. 155ff.; H. Delehaye, Les origines d u
culte des martyrs, Subsidia Hagiographica, XX2 (Brussels, 1933).- I should like to thank Prof. S. Der
Nersessian for her advice on iconographic questions.
4 3 Migne, PG, XLVI, cols. 749ff., 757ff., 773ff. The first homily is but the beginning of the second.
The dream is related in the third homily, ibid., col. 785.
f i Migne, PG, XXXI, col. 508ff. St. Basil knew the diatheke.
 T W O P A L A E O L O G A N MOSAIC I C O N S 99
of this motif; indeed, the whole legend is understandable only from this point
of ~ i e w . ~ 5 account begins with the professio of forty soldiers of the "Ful-
The
minata" legion which was stationed in Sebaste (Sebastia). They repeat their
profession of faith at the tribunal of the Governor, are thrown into prison and
condemned to die of cold, exposed on a frozen lake in or near the town. A
well-heated bath is kept open on the shore to receive those who might recant.
While they are slowly freezing to death, thirty-nine golden diadems are seen
to descend on the "phalanx" of the Forty. The guardian of the bath house,
as he is counting these crowns and wondering why there are only thirty-nine
of them -one short of the number of confessors-sees one of the Forty break
the ranks and run for the refuge, on entering which he is immediately dissolved
into air. This so impresses the heathen guardian that he throws off his clothes
and joins the martyrs on the ice in order to restore the original number of
forty. The next day the limbs of those who were not yet dead are broken (the
crurifragium) and the bodies loaded onto a cart and taken to be burnt. The
youngest of the Forty, Melitoj46who is still alive, is left behind by the execu-
tioners who take pity on him. But his mother, who had watched the martyrdom
from nearby, not wishing to see him cheated of his martyr's crown, lifts him
up in her arms and runs with him after the cart. So he too is burnt, and all the
ashes are thrown into a river by order of the Governor who wants to make
sure that no relics are left, but his plan is frustrated and the relics are miracu-
lously recovered from the waters.
This is not the place to analyze the various versions and motifs of this legend.
As we have already pointed out, it seems to be mainly concerned with the
preservation of the full number of forty, which accounts for the story of the
guardian and the stark episode of Melito's mother. The substitution of the
faint-hearted martyr by the guardian was soon compared to the substitution
of Judas by Matthias: this parallel is already drawn in St. Basil's sermon.
As a matter of fact, all the essential motifs are already present in the sermons
of the Cappadocian Fathers; they are, indeed, mentioned there in such an
allusive form that they must have been widely known in the second half of
the fourth century. We do not know, however, at what period the legend was
illustrated in a full cycle of pictures. The earliest examples of this type that
have come down to us are contained in a group of Psalters with marginal
illustrations, the oldest of which was written in the second half of the eleventh
century. This does not, of course, exclude the possibility that the narrative
cycle of the legend could have originated at a much earlier period. The chief
manuscript of this group, the Psalter of the British Museum, Add. MS 19352,
(figs. 4 a, b) may even give us a clue as to where to look for the origin of the
45 The earliest sources after the diatheke and the homilies of St. Gregory and St. Basil are several
hymns and a homily by Ephraem Syrus (C. Assemani, S. E p h r a e m S y r i opera o m n i a , V [Rome, 17432,
p. 341ff.), a Latin sermon by Gaudentius of Brescia (PL, XX, col. 964ff.), the Greek Acts (AASS,
X. Mart., p. rzff.), a Synaxarion (H. Delehaye, S y n a x a r i u m Ecclesiae Constantinopolitanae [Brussels,
19021, col. y ~ f f . )and , a hymn by Romanos (K. Krumbacher, "Miscellen zu Romanos," Abhandl.
der K. B a y r . A k a d . der Wissenschaften, I. K1. XXIV/3 [Munich, 19071, p. 16ff.).
48 There is a discrepancy here between the legendand the diatheke, inwhich thename of the youngest
member is Eunoicus.
100 OTTO D E M U S
cycle. The Psalter was illuminated in 1066 in the Studios Monastery, by a
scribe named Theodore, a native of Caesarea where, it may be remembered,
the mother of St. Basil and St. Gregory built a shrine for some of the relics
of the Forty. If we also consider the fact that the miniatures of the Theodore
Psalter are the most faithful illustrations we have of the narrative contained
in St. Gregory's sermons, it is, perhaps, not too fanciful to suggest that the
miniatures were inspired by an early narrative cycle that existed in the church
of Caesarea. The transmission of this cycle from early prototypes down to
the eleventh century may have been effected by means of illuminated manu-
scripts of the sermons of Basil or Gregory, or else through illustrated copies
of Symeon Metaphrastes' Lives.47
There is one more element which leads us back, by a different route, to the
two Cappadocians. In the group of Psalters I have mentioned the scenes from
the legend of the Forty serve to illustrate Psalm 66 :12 : "We went through
fire and water but thou broughtest us into a wealthy place." This refers to
the posthumous fate of the Forty, namely to the Governor's attempt to
destroy their relics by fire and water rather than to their actual martyrdom.
Now, this rather artificial connection is not a late invention, since it is al-
ready found in the twentieth homily of St. Basil who may very well have
invented it .48
In the London Psalter the cycle unfolds itself in the margins of two pages,
81 and 81", beginning with the tribunal and ending with the salvaging of the
relics. I t is closely followed by the Barberini Psalter49 (fol. 103~)and by the
Russian Psalterof 1397 (fol. 86).50Since the burning of the bodies is not repre-
sented in these manuscripts, the cycle must be regarded as an abbreviated
version of a fuller one, which must, therefore, have existed at an earlier date.
An even more abridged redaction is contained in the Hamilton Psalter, of the
thirteenth century, at Berlin151which has only two scenes, namely the group
of the Martyrs on the lake and the collecting of the relics.
The narrative cycle which has found its finest literary expression in a hymn
by R ~ m a n u sis, ~not
~ alone in pointing to Asia Minor as the place of origin of
the iconography of the Forty Martyrs; another mode of representation, com-
pletely opposed to the scenic cycle, seems also to have been prevalent in that
region, namely the representation of the Saints as forty half figures clad in
patrician robes and enclosed in medallions. This manner of representing the
Forty Martyrs by "pseudo-portraits" identified by the names given them in

4 7 H. Delehaye, Les lLgendes grecques des saints militaires (Paris, 1909) ; Bibliotheca Hagiographica
Graeca (2nd ed., Brussels, ~ g o g )pp.
, 168f., 290.
48 The motifs of fire and water and the reference to David (the Psalter) are also found in an Ekphrasis
(Cod. Marc. gr. 524) of a representation of the Forty Martyrs that existed in the propylaion of one of
their churches in Constantinople; cf. Nios 'EAAqvopvfipwv, VIII (191I), p. 126f.
48 Vat. Barb. 372 (cf. A. De Wald, in Hesperia, XI11 [1g44], p. 76ff.).
50 Leningrad, Public Libr., no. 1252, FO VI (cf. Istoriia Russkogo Iskusstva, ed. Akad. Nauk SSSR.,
I11 [Moscow, 19551, p. 94, with bibliography).
51 Berlin, Kupferstich Kabinett, no. 78AG (Hamilton I I ~ )fol., 130.
5 2 K. Krumbacher, "Miszellen." The hymn contains the motifs of fire, water, and the substitution
of Judas by Matthias; the words of St. Basil are in some cases reproduced literally.
 T W O P A L A E O L O G A N MOSAIC I C O N S 101
the Diatheke is found in five cave churches of Cappadocia, among them Toqale
(the New Church) and Tschaouch In.53 Some of these representations are as
early as the tenth century. Cappadocia is not, however, the only region in
which this mode of representing the Forty Martyrs occurs: other, though later,
examples are found in Sicily (the mosaics of M ~ n r e a l eand
) ~ ~in Russia.55
A similar method of representing the Martyrs' portraits is shown in Syriac
manuscripts of the thirteenth century, in which the medallions, arranged in
a honeycomb pattern, fill two facing pages56 (fig. 5 ) . These rows of medallions
are the exact pictorial counterpart of chanted litanies which must have been
current from comparatively early times, though their earliest extant example
is in a tenth-century manuscript (Paris, gr. 476).57The litany or hymn, as it
is called in the introductory verse, seems to have been divided into two halves
each containing the invocation of twenty saints, a bipartition that we shall
meet again in another group of representations of the Forty. As regards the
litany, only its first half appears to have been copied in the Paris manuscript.
The names of the saints are those of the Diatheke, while the ideas expressed
in each invocation have been developed from the names by association or
alliteration. The resulting effect is just as monotonous as that of a row of
medallions.
The main development of the iconographic theme of the Forty Martyrs is
not, however, connected either with this "portrait" type or with the "cyclic"
representation described above. What eventually became the "classic" mode
of depicting the Forty Martyrs lies midway between the two poles, the static
and the narrative. This "classic" type represents the martyrdom of the Forty
in one image, supplemented by marginal scenes which allude to the most im-
portant motif of the legend, that of the substitution of the deserter by the
proselyte.58 The earliest example of this type that has come down to us is found
in the apse of the Chapel of the Forty Martyrs at Sta Maria Antiqua in Rome,
of the seventh or eighth century59 (fig. 6). The attitudes of the Martyrs, with
their hands raised in orant gestures, conforms more to the earliest versions
of the Acts, in which the emphasis is laid on the triumph of the Saints who
praise God during their "agon," than to the evocative sermons of the two
Cappadocian at hers in which the suffering and the painful death of the
" 3. de Jerphanion, Les iglises rz~pestresdeCappadoce, I (Paris, 1925), pp. 314, 316, 529; 11, pp. 26,
158f.,167f.,276.
54 0. Demus, T h e Mosaics of N o r m a n Sicily, pp. 121, 200, 235.
55 Kiev, St. Sophia: 0. Powstenko, op. cit., p. 112; Nereditsa: V. K. Myasoyedov and N. Sychev,
Freski Spasa-Nereditsy (Leningrad, 1925), Index, p. 25.
58 Vat. Cod. syr. 559 (G.de Jerphanion, Les mzniatures d u manuscrit syriaque N . 559 de la Biblio-
thBque Vaticane [Vatican City, 19401, P. 88, pls. XI, XII), and London, Add. MS 7170 (ibid., p. 91,
fig. 38).
57 D. Amand, "Un court poBme en l'honneur des Quarante Martyrs de Skbaste," Scriptorium, I11

(19491, p. 52ff.
58 In accordance with the "complettierende Typus" of Wickhoff or the "simultaneous method" of
Weitzmann: cf. K. Weitzmann, Illustrations in Roll and Codex (Princeton, 1g47), P. 12ff.
59 J. Wilpert, Die romischen Mosaiken u n d Malereien der Bauten vom 4. bis 8. Jahrhundert, I1
(Freiburg i. B., 1916), p. 722; IV, pl. 199. The date is controversial, as is also that of the fresco in the
oratory of Sta Lucia, Syracuse: P. Orsi, Sicilia bizantinu, I (Rome, 1g42), p. 8off., pl. v. The latter
painting was brought to my attention by Dr. Hans Belting.
102 OTTO D E M U S
Martyrs are described in harrowing detail.60The figures are arranged in three
rows, in a regular pattern; a blue zone at their feet represents the ice of the
lake, while a celestial apparition at the top of the composition, with rays des-
cending on the "phalanx," imparts a hieratic note to the group. On the right is
the bath house, guarded by a seated figure with shield and lance; a man is
seen entering the door (the defecting member of the group), while the converted
guard joins the "phalanx" from the left.
The simplicity and relatively static character of this composition are in
accord with its early date. The main group of figures is little more than a
group portrait, as it was frequently represented in Roman art, from the "school
group" of the Capua Museum to the sacrificial scenes of Dura-Europo~.~l
Sta Maria Antiqua, however, contains a second representation of the Forty
which is even closer to its Roman prototypes, a "celestial group portrait" pure
and simple, which shows the Martyrs in a well-ordered group, all with haloes
and in patrician costumes, with Christ appearing above them in a medallion.
This representation is on one of the side walls of the chapel,62 at right angles
to the image of the "martyrdom" and, apparently, intended to be viewed
together with the latter. \Ve would probably be justified in interpreting it as
a depiction of the Martyrs after their "agon," as saints in heaven, still forming
their indivisible "phalanx. "
There was yet a third image of the Forty Martyrs in Sta Maria A n t i q ~ a , ~ ~
which seems to have been as unique as the one just described, but of which
only a small fragment is preserved. The part that is still visible shows two
figures clad in loin cloths and, above them, the remnants of a figure with a
crenellated crown, doubtless a personification of the city of Sebaste, conceived
in the Hellenistic manner. That the painter was a Greek seems certain because
of the Greek subscription containing a prayer.
All three compositions of Sta Maria Antiqua, ancient as they are, may be
derived from even earlier prototypes; but only one of the three iconographic
schemes lived on to become a dominant type: the one represented in the apse,
with the complementary side scenes. Of course, there were considerable com-
positional problenls involved in this rendering of the scene in the shape of a
horizontal rectangle : it was as difficult to arrange satisfactorily forty standing
figures, as it was to find a suitable place for the episode of the deserter and the
proselyte. In some cases the two scenes were simply omitted, as in a miniature
~ ~ 7) which, though belonging to the late
of the Moscow h f e n ~ l o g i u m(fig.
eleventh or early twelfth century, follows quite faithfully a late tenth-century
60 h combination of the two interpretations is found in an inscription in the Cypriote church of
Asinou, A.D. 1106: "It is flesh that here bears the winter's cold. Thou shalt hear the martyrs' sobs and
groaning. They are steadfast as they suffer under the sharpness of the frost; at the clouds they look
and not upon their pangs." (Archaeologia, LXXXIII [1933], p. 328f., pl. 9811);similarly an epigram
by Manuel Philes ed. E. Miller, I, p. 438.
The Capua mosaic was found in St. -4ngelo in Formis (ill. in Caserta and its Province [Caserta,
n. d.], p. 17). Dura, Temple of Bel, Sacrifice of the Tribune Julius Terentius, ca. A.D. 239.
6 2 J. Wilpert, op. cit., IV, pl. zoo.
63 Ibid., 11, p. 709; IV, pl. 177.
64 D. K. Trenev, Miniatzwes d u minologe grec d u XIe siBcle no. 183, de la BibliothBque Synodale de
Moscou (Moscow, 191I ) , pl. ~111136.
 T W O P A L A E O L O G A N MOSAIC I C O N S 103
prototype of the type of the Vatican Menologium. This simplified composition
is also found in other m a n u ~ c r i p t sin
; ~one
~ of these, the Menologium of St. Saba
in Jerusalem,BB the illuminator attempted to solve the compositional problem
by building up a symmetrical, bipartite composition with two bending or,
rather, collapsing figures on either side of the central group, and by enclosing
the entire mass of figures in a lunette formed by the shape of the frozen lake.
This device was also used in monumental painting; we may even go so far
as to say that it originated in wall painting or mosaic. In the considerably
later (middle of the fourteenth century) wall painting of Lesnovo, Serbia,67
(fig. 8) the scene is actually inscribed in a lunette and divided into two halves
by a double window. Such a division of the scene into two halves apparently
proved to be convenient: this is, perhaps, the reason for the wide dissem-
ination, from Cappadocia to Serbia, of an arrangement whereby the Martyrs
were disposed in two groups of twenty on either side of a barrel vault, with
the bath house at one end, as in %Ea, SerbiaB8(figs.9, a, b), or in the tympanum,
as in the church of the Forty Martyrs near Souvech in CappadociaB9 (fig. 10).
This type of representation, on either side of a barrel vault, does, indeed,
seem to have been the standard one in wall painting from the eleventh to the
first half of the fourteenth century.'O At this date, however, it was challenged
by an entirely different kind of image, shaped like an upright rectangle, in
the format of an icon. Its appearance in wall painting after about 1325 -a good
dated example (ca. 1340) is found in DeEani, Serbia7* (fig. 11)-is part of a
larger process that is characterized, among other things, by the intrusion of
"iconic" forms into monumental painting. There can hardly be any doubt
that the upright rectangular met hod of representing the Forty Martyrs was
developed in icon painting and carving as early as the tenth, perhaps even the
late ninth, century in Constantinople itself. The earliest examples that have
come down to us are two tenth-century ivory reliefs in Berlin and Leningrad72
(figs. 12, 13). Of course, the problem of arranging forty standing figures in
an upright rectangle is even more difficult than it is in a horizontal rectangle,
unless the artist were to pack the frame with heads neatly arranged in rows.
If, on the other hand, the figures were concentrated in the lower half of the
rectangle, the upper half was in danger of being left empty or, at least, of
being inadequately filled. To remedy this the scene of the bath house had to

a E. g. Messina, Cod. S. Salvatore no. 27 (cf. Ch. Diehl in Me'langes d'archLologie et d'histoire de
I'Ecole Frangaise de Rome, V I I I [1888], p. 320), and Dionysiou, Cod. 50 (Sp.Lambros, Catalogue, I,
P. 322).
66 Jerusalem, Greek Patriarchate, no. 208, Menaion of St. Sabas (A.Baumstark in Oriens Chris-
tianus, Ser. III/I [1g26], p. 70, pl. 11).
67 N. Okunev, "Lesnovo," L'art byzantin chez les Slaves, I (Paris, 1g30), p. qgff., pl. XXXVIII.
G. Millet and A. Frolow, La peinture du moyen-dge en Yougoslavze, I (Paris, 1g54), pl. 60/2,3.
G. de Jerphanion, Les e'glises, op.cit., pl. 16114.In Syracuse the Forty Martyrs are divided into
four groups of ten each. See supra, note 59.
' 0 Ohrid (north chapel), Studenica, SopoEani, Gradac: cf. G. Millet and A. Frolow, La peinture,
op. cit., I , 37/4, 91/2; 11, 28, 29; 62/4.
' 1 V. R. Petkovi6 and Dj. BoSkovi6, Manastir DeZani, I1 (Belgrade, 1941)~pl. CXXI/I.
7 2 Berlin (A. Goldschmidt and K. Weitzmann, op. cit., 11, p. 27, pl. 3, no. 10); Leningrad (ibid.,

11, P. 3, no. 9).

 104 OTTO D E M U S
be moved up and the remaining area filled with a heavenly apparition that
took up more space than the Hand of God which appears, e.g., in the Moscow
Menologium. To solve these problems the carvers of the two ivories (or their
prototype) seem to have borrowed or to have been inspired by an existing
compositional scheme that contained similar constituent parts, namely an
agitated group of figures and a celestial apparition above-an arrangement
that was commonly used for the Ascension of Christ. Very rarely in the course
of the Byzantine period do we encounter the creation of an entirely new
compositional pattern; usually a time-honored scheme which contained similar
features was taken as a point of departure, and only gradually did the new
scheme free itself from the parent form. The compositional scheme of the
Ascension also left space for the insertion of the bath house between the
"phalanx" of the Martyrs and the enthroned Pantokrator with adoring Angels.
The guardian who joins the phalanx in the place of the deserter may be identi-
fied, in the Berlin ivory, with the figure in the right lower corner, an old man
who is shedding his garments as he approaches the group.
Not only the general con~position,but also the figural design of the Berlin
ivory, including every attitude, every movement, and every facial type, is
found again in a silverpoint drawing, parts of which are preserved in a Vatican
manuscript, Cod. Barb. lat. 1 4 4 ~(fig.
~ 14). The drawing must originally have
formed an upright rectangle measuring ca. 58 by 44 cm. ; of this only the lower
half is preserved, showing the greater part of the "phalanx." Unfortunately,
however, a large area of this very delicate drawing was written over, in the
fourteenth or fifteenth century, so that only the first five figures from the left
are distinctly visible. While, as I have said, it tallies in every detail with the
Berlin ivory, the drawing is more elaborate, especially as regards hair, beards,
and drapery. Thus, it is hardly likely that it was copied from the Berlin or
any other ivory; it seems rather to have been derived from a painted icon of
about the same size as the drawing. There is, of course, no possibility of estab-
lishing the date of this presumed icon; the drawing itself, which may be con-
siderably later than its prototype, seems to belong to the thirteenth century.
The monuments studied so far make it possible to reconstruct the missing
parts of the Dumbarton Oaks icon (figs. 1-3). In the three rows which make
up the phalanx, including the collapsing figure in the foreground, thirty-nine
figures or heads can be discerned, which means that two are, at present,
missing: the deserter and the proselyte. A trace of the first, a part of his
loincloth, can be seen in the upper row near the right edge; he was, apparently,
represented in the act of breaking away in order to enter the bath house which
must have been immediately above, where there is now a large gap filled with
gilt wax. There is also a remnant of the guardian: his arm and hand, lifted in
7 3 W. F. Volbach, "Le miniature del Codice Vat. Pal. lat. 1071 'De Arte Venandi cum Avibus,'"
Estr. dai Rendiconti della Pont. Accad. Romana d i Archeologia, X V (1g3g), p. 28, figs. 23, 24 (cf. H.
Swarzenski in T h e Avt Bulletin [1g42], p. 298); B. Degenhart, "Autonome Zeichnungen bei mittel-
p. 98, fig, 2 0 . Accord-
alterlichen Kiinstlern," Miinchener Jahrbuch der bildenden K u n s t , N . F., I (1950)~
ing to Degenhart there was an intermediary link between the original and the drawing; the latter was
traced from the intermediary version.
 T W O P A L A E O L O G A N MOSAIC I C O N S
prayer, can be seen at the left edge of the gap, above the last letter of the in-
scription. He must have been represented kneeling, with his back to the bath,
looking up towards the celestial apparition. Thus, the entire scene of desertion
and substitution was moved to the upper right side, as in the ivories; the
inscription and the forty descending crowns were arranged asymmetrically,
so as to leave sufficient space for the supplementary scene.
In the composition of the icon the effects of lateral compression can be felt
quite clearly, and rather painfully. The artist seems to have followed a model
of oblong format and, as is usual with compositional re-arrangements, to have
started translating his model into an upright rectangle beginning from the
left. The sequence of the figures in the left half is almost normal and organic,
every figure being afforded sufficient space. From the middle axis on, however,
the arrangement begins to be crowded, and at the right edge the rows of
figures, especially the front row, are almost curved back with overcrowding.
Furthermore, some of the busts in the second and third rows seem to have been
conceived without regard to the remainder of their bodies. Thus, the entire
composition is little more than an assemblage of heads, of character types,
rather than a "phalanx" of figures. Only in the front row (and in a few half
figures of the second row) has there been an attempt to go beyond this j uxta-
position of stock portrait heads.
A very similar composition, with similar faults and virtues, came to light
recently only to be lost again, in an annex to the Euphemia Martyrion in
Constantinople; to all appearances it was a work of the fourteenth century.74
The main difference seems to have consisted in the fact that the figures were
more widely spaced and, consequently, their movements were more clearly
differentiated.
On the whole, it appears that the artist's most difficult problem was not the
general arrangement of the composition-a problem that he did his best to
solve after the pattern of familiar iconographic schemes -but the differentiation
of forty standing figures, and the convincing depiction of their suffering; for
this is what the artist of the Dumbarton Oaks icon set out to do, just as if he
had wished to illustrate one of the sermons of the Cappadocian Fathers. Every
shade of psychological reaction is depicted here, from stoicism to despair; some
of the bodies are rendered as if contracted by the cold; one of the Martyrs is
breaking down; another, somewhat left of center, has collapsed and is held by
one of his younger comrades. Nevertheless, considerable restraint is shown in
portraying the suffering'--more mental than physical -of the Martyrs of the
Dumbarton Oaks icon, and this is especially evident when their attitudes are
compared with those of the Berlin ivory. The mosaic offers a convincing and
dignified representation of anguish, despair, and gloom, quite different from
the grotesque contortions found in the Berlin ivory.
74 A. M. Schneider, "Grabung im Bereich des Euphemia Martyrions zu Konstantinopel," Archae-
ologischer Anzeiger (1g43),p. 280, fig. 16. Schneider's assertion that the fresco was almost monochrome
should be taken c u m grano salis. Prof. P. Underwood tells me that the painting was already badly
weathered a t the time of its discovery; it was certainly very much damaged and faded when I saw i t
in 1951.
 OTTO D E M U S
Among the many renderings of this subject there are a few that are even
more turbulent than the Berlin ivory; one of the most agitated compositions
is that of the now destroyed wall painting of VodoEa in Macedonia, of the
eleventh ( ? ) century, only part of which is known through an unsatisfactory
(fig. 15). I t has been claimed, on the one hand, that the agitated realism
of movement, gestures, and grimaces, as exemplified by the VodoEa fresco,
was inspired by Hellenistic models; on the other hand, it is held by a number
of scholars that the spirit of this fresco is typically Macedonian and contrary
to the classical spirit of Constantinople. There may be some truth in both of
these points of view, insofar as, generally speaking, Macedonian and metro-
politan artists, while using the same late antique models, gave them different
interpretations. However, before we analyse the character of the antique
prototypes used by the mosaicist of the Dumbarton Oaks panel, and the use
which he made of these prototypes, we should take another factor into account.
This, "the middle-Byzantine" factor, as we may call it, affects the attitudes
and movements of the individual figures of the Dumbarton Oaks icon.76 The
closest parallels to these attitudes and to the general "mood" of the group of
Saints can be found in representations which, at first, seem far removed from
the iconographic context to which the Forty Martyrs belong. I t is almost with
a shock that one realizes that these parallels are found in the groups of damned
and tormented souls which usually occur in representations of the Last Judge-
ment. The mosaic of Torcello, though somewhat provincial (especially in the
lower parts of the huge composition) and all but ruined by bad r e ~ t o r a t i o n , ~ ~
(fig. 16) shows, for example, the man clutching his face with both hands
who recurs in the Dumbarton Oaks mosaic as the third figure from the right in
the middle row. Both mosaics also have in common the man with one hand
raised to his head, the other to his breast -he is the second from the right in
both instances. In another representation of the "worm that sleepeth not,"
in DeEani, Serbia (about 1 3 4 0 ) , ~(fig.
~ 17) the same gesture is found once more,
as well as that of the man with the folded arms. The impressive representation
of the Last Judgement in the Parecclesion of the Kariye Camii79(fig. 18) is
steeped in the same atmosphere of gloom and despair that is so characteristic
of the Dumbarton Oaks icon. The astonishing iconographic relationship be-
tween the two themes, the martyrdom of the Forty and the torment of damned
souls in the Last Judgement is finally brought home by a most characteristic
detail: the fresco of the Last Judgement in the Kariye (fig. 19) and that of

75 K. Miyatev, "Les Quarante Martyrs, fragment de fresque de VodoEa, Mac6doinelHL'art byzantin

chez les Slaves, I (1930)~ . , X, 2 ; G. Millet and A. Frolow, op. cit., I, pl. 14, 5.
p. ~ o ~ f fpl.
78 For a different interpretation suggesting a derivation from classical prototypes, see Prof. Weitz-
mann's paper in this volume of the Dumbarton Oaks Papers, p. 66.
77 0. Demus, "Studies among the Torcello mosaics," The Burlington Magazine, LXXXII (1g43),
p. 136ff.; LXXXIV (1944),p. 41 ff.; LXXXV (1944),p. 195ff. The mosaic should be dated in the first
half of the twelfth century.
78 V. R. Petkovi6 and Dj. BoSkovi6, 09. cit., pl. c c ~ x x x I, .
70 P. A. Underwood, "Third Preliminary Report on the Restoration of the Frescoes in the Kariye
Camii a t Istanbul by the Byzantine Institute, 1956," Dumbarton Oaks Papers, 12 (1958), p. 235ff.,
figs. 20-22.
 T W O PALAEOLOGAN MOSAIC I C O N S 107
the Forty Martyrs in &a, Serbia (fig. gb) have a very odd figure in common,
that of a man seen from the rear with head thrown back.80
The theme of the Last Judgement is not, however, the only one that shows
close contacts with that of the Forty Martyrs: the figures of contrite monks,
doing penance by exposing themselves to cold, who appear in the illustrations
of the Penitential Canon affixed to John Climacus' Heavenly Ladder, as shown
in the Vatican Manuscript, Cod. gr. 1754 (fol. 13"),belong to the same cate-
gorys1 (fig. 20). The caption above the illustration reads: "These torture them-
selves in the cold, and shiver, stiffened with frost." The figures are depicted in
the same poses as the Forty Martyrs or the damned souls: cold and despair
furnish the common denominator. Byzantine art seized on such common
denominators, in the representation of single figures as well as of complex
scenes, and applied them in different contexts.
Within a given sphere-in this case, that of agony, anguish, and cold-figure
types were almost interchangeable; they could be used equally for martyrs,
damned souls or penitents. I t was left to the individual artist to impart to
each of these interchangeable figures a special character to fit the context. But
the adherence to types of this generalized nature is one of the things that gave
Byzantine art its grandiose homogeneity.
Of course, not all the attitudes and gestures of the Forty Martyrs in the
Dumbarton Oaks icon were drawn from this sphere. Others are derived from
New Testament iconography. For instance, some of the freezing figures,
especially those hugging themselves with both arms, may have been taken
from representations of the B a p t i ~ m another,
;~~ most characteristic figure,
which occurs in many pictures of the Forty Martyrs, can also be traced to a
New Testament scene: the sagging body of the Martyr upheld by the youthful
figure in the foreground is an adaptation of the dead body of Christ, held by
the Virgin in the Descent from the Cross.83In the Dumbarton Oaks icon there
is only one figure in this posture, whereas in the Jerusalem Menologiums4there
are two. The Berlin ivory and its relatives (figs. 12-14) contain a group of two
figures, one resting its head on the breast of another, which recalls the Christ-
St. John group of the Last Supper. Later representations of the Forty Martyrs
abound with highly agitated and strangely contorted figures. One very charac-
teristic, if somewhat extreme, example is the collapsing man seen from the
rear with his head turned upside down, whom we have already met in ZiEas5
(fig. 9) ; others of a similar kind are seen, for instance, in a sixteenth-century
so Ibid., fig. 21,and G. Millet and A. Frolow, op. cit., I, pl. 60/2,3.
J. R. Martin, The Illustration of the Heavenly Ladder of John Climacus (Princeton, 1954),p. 138ff.,
fig. 266. Further illustrations of the theme are enumerated ibid., p. 139. Professor Martin considers
i t possible that the miniature painter was inspired by a representation of the Forty Martyrs; both
types, however, are more likely to have been derived from representations of the Last Judgement.
E. g. Brussels, Coll. Stoclet, fourteenth-century icon (V. N. Lazarev, Istoriya vizantiiskoi Zivopisi,
I [Moscow, 19471, pl. XLII); frescoes of Aquileia, crypt (La Basilica di Aquileja [Bologna, 19331,
pl. LXXI).
83 FO; example, Paris, Bibl. Nat. cod. gr. 543 (G. Millet, Recherches sur l'iconographie de ~'gvan~ile,
aux X I Ve, X Ve et X V I e sidcles [Paris, 19161,fig. 168).
84 See supra, note 66.
See supra, note 80.
 OTTO DEMUS
icon of the Byzantine Museum of Athenss6 (fig. 21) in which three figures have
already broken down and are sitting on the ground and others are collapsing,
while the guard's extended arm is clasped by two of the martyrs as a sign of
welcome. One could hardly imagine a stronger contrast than that between this
lively work of popular art and the noble restraint of the Dumbarton Oaks
mosaic.
Many of the facial types of the portable mosaic appear, like the postures of
the figures, to have been borrowed from the realm of religious iconography.
Some of the Martyrs' heads resemble the usual types of the evangelists. The
features of St. Luke can be found at least three times with slight variations;
some of the bald heads with long beards recall St. John, while those with long
grey hair and beard are reminiscent of St. matt he^.^' A number of figures can
be compared to one or the other of the prophets, while others are modelled
after representations of saints. An especially close affinity as regards facial
types exists between the Dumbarton Oaks icon and the mosaics and frescoes
of the Kariye Camii: the fifth head from the right in the bottom row is, for
instance, very similar to the head of St. Elpidiphoros in the Constantinopolitan
mosaics.88 Others, like the youthful heads in the bottom row of the icon can
be found once more in the frescoes of the Kariye Parecclesion, particularly
in the hell-fire group of the Last Judgementsg (fig. 18). In these frescoes we
also find the same type of shaggy bearded head, e.g., the first on the right in
the middle row.
Ultimately, however, many of the facial types found in our icon as well as
in the mosaics and frescoes of the Kariye Camii can be traced not to models
belonging to religious art, but to antique prototypes. The profile head in the
top row of the icon, for instance, is reminiscent of that of Alexander in the
famous Pompeian mosaic ; a head similar to the fifth from the left in the top
row and the fourth in the bottom row can be found in the Centaur mosaic from
Hadrian's villa, now at Berlin;gothe fourth from the left in the top row may
be derived from a Hellenistic portrait like that of Dioscorides (= Krateuas) in
the Vienna Herbal,g1and the youthful heads in the first row are also as Hellen-
istic as can be.
But not only were some of the facial types derived from antique models;
several attitudes, too, are essentially Greek -the contraposto of the first figure
on the left for instance. No less Hellenistic is the manner in which the flesh is
modelled, partly by continuous shading, partly by glittering white highlights

G. A. Sotiriou, 'OGqyo~-roil BuLavr~voOMouoeiou 'ASqvGv, 2nd ed. (Athens, 1941), pl. 6.

07 Luke: first and tenth figures, top row; fourth figure, middle row. John: third figure, middle row.
Matthew: fifth figure, middle row; second figure, bottom row.
as Cf. Th. Schmit, "Kahrie Djami," Izvestiya Russk. Arkh. Instituta v Konstantinopole, XI (1906),
pl. XIX, 66; other types ibid., pl. XXII, 165; LXIV, 129; LXXXIV.
89 P. A. Underwood, Third Preliminary Report, figs. 21, 22.
O , pittura Ellenistico-Romana (Milan, 1g2g), pi. CLXXXV.
90 G. E. R ~ Z ZLa
91 Vienna, Nat. Libr., Cod. Med. graec. I , fol. 4: Dioskorides and Heuresis. illustrated in V. N.
Lazarev, Istoriia viz. iivopisi, 11, pl. 16; P. Buberl, "Die byzantinischen Handschriften, 11," Be-
schreibendes Verzeichnis der illuminierten Handschriften in ~sterreich, VIII (IV/I) (Leipzig, 1937)~
pl. 111.
 TWO PALAEOLOGAN MOSAIC ICONS
which illuminate the foreheads, the tips of the noses and the hair, and light
up the eyes. Some of the white highlights have a somewhat impasto quality:
the tiny white tesserae of which they consist project ever so slightly from the
otherwise smooth surface of the mosaic.92 The modelling of the drapery is
bolder than that of the bodies. Jagged patches of light and dark are interwoven
so as to form a lively pattern. Here, too, the artist's familiarity with Hellenistic
models of the illusionistic manner can be clearly felt.
The classicising nobility of the Dumbarton Oaks icon, the Hellenistic flavor
of its facial types and modelling, as well as its subdued and somewhat mono-
chrome coloring, enable us to classify this as a work of the so-called Palaeologan
renaissance, produced in Constantinople. As to types and style, a close parallel
is provided, as we have seen, by the mosaics and frescoes of the Kariye Camii
which date in all probability from the second decade of the fourteenth century.
Some slight differences of style, however, which can be detected by a detailed
analysis, seem to suggest that the icon should be dated two or three decades
earlier than the decoration of the Kariye. The feet, for instance, though already
"broken" at the ankles, have not yet the characteristic "flat-iron" shape of
the Kariye, the crania are not as bulbous, the hands not as prong-like. Neither
the composition as a whole nor the design of the bodies exhibits the calligraphic
linearism of the Kariye mosaics. In fact, the figures of the icon still betray the
experimental character of the art of the thirteenth century and its classical
trends. One feels that the painter of the icon -and the mosaicist was a painter
par excellence-regarded his Hellenistic models as something only just dis-
covered and therefore deeply exciting. In some cases he copied them so closely
that certain heads of the icon, if considered in isolation, would seem to belong
to the sixth century rather than to the late thirteenth. A case in point is the
sixth head from the left in the front row, the uppermost in the group of four:
the quick turning of the head and the flashing glance directed upward at an
oblique angle recall the illusionistic manner of the Vienna Genesis rather than
any mediaeval method of depicting the face of a saint. But, as has been pointed
out above, there are other faces in the icon which do not correspond to any
antique type and do not show an antique treatment; and it is this uneven
character, this medley of the antique in its various forms, both classical and
illusionist, and of the mediaeval, that distinguishes this work of an experimen-
tal, exploratory art from the perfectly balanced, homogeneous and assured art
of the Kariye. Compared to the Kariye mosaics with their deeply saturated
style, their insistent homophony of forms, their mannerist abbreviations, the
Dumbarton Oaks icon has no "style" at all: it is a document of an artist's
search rather than a product of skilled routine. Thus it is one of the small
number of monuments which illuminate for us the highly complex process
of the genesis of Palaeologan art .93
92 A similar technique is found in the mosaics of the Kariye Camii.
93 0. Demus, "Die Entstehung," passim.
 OTTO D E M U S

THE ICON O F ST. JOHN C H R Y S O S T O M ~ ~

(Figures 2 2 , 23)
The more recently acquired of the two Dumbarton Oaks icons represents a
half figure of St. John Chrysostom in a frontal attitude. The wooden panel on
which it is set measures 13 by 18 cm. When it was acquired, the icon was
encased in a narrow metal frame which, though not recent, can hardly be
regarded as the original one.95The mosaic surface is very well preserved except
for some cracks and a very few missing tesserae, mostly in the golden ground.
The interstices between the impasto highlights of the forehead are somewhat
discolored; otherwise, the colors are as bright today as when the icon was made.
The white folystavrion of the Saint is covered with red crosses, outlined in
gold; the omophorion is also white, but shot with gold and decorated with
large, dark blue crosses with gold contours. The sleeve of the undergarment
is red and gold. The golden gospel book has a red edge and is set with red, blue,
and green stones. The most outstanding feature of the icon is the large halo in
low relief, with an all-over pattern of small crosses in red, blue, and green
against a white and gold ground; it appears almost like a life belt or a cushion
in which the head is embedded. The subdued, flat colors of the head stand in
effective contrast to the gaily colored patterns of the nimbus and vestments.
The hair and beard are dull brown with a little pink, olive, and black; grey,
olive, light
-
brown, pink, and white make up the flesh tones -the white tesserae
projecting from the surface like highlights put on in impasto. The ground is
gold, the highly decorative inscription spelling the name of the Saint is black ;
the inner border is of the same colors as the halo.
The pedigree of the icon is the best imaginable. When it first became known,
it was in the monastery of Vatopedi on Mt. Athos. At the end of the last
century it was presented by the monastery to Count Nelidov, then Russian
ambassador at the Sublime Porte. With the Nelidov collection it came to
Paris; Dumbarton Oaks acquired it in 1954.
The great preacher and patriarch, whose portrait is the subject of the icon,
was certainly one of the most deeply venerated of the Fathers, although he
was neither a miracle worker nor a martyr in the strict sense of the terms. He
was, however, reckoned among the confessors, having died in exile in 407 in
Colchis, where he was banished by the Empress Eudoxia and her party.96The
94 Handbook (1g55), no. 291, p. 147. Bibliography (with the exception of general handbooks), in
chronological sequence: D. Ainalov, "MozaiEeskaia portativnaia ikona sv. Ioanna Zlatousta Vatoped.
monastyria," Vizantiiskii Vremennik, V I (18gg), P. 75ff.; N. P. Kondakov, Pamiatniki ( ~ g o z ) ,
p. 116ff.;A. Muiioz, L'art byzantin d l'exposition de Grottaferrata (Rome, 1go6), p. 170; J. Wuescher-
Becchi, "Saggio d'iconografia di San Giovanni Crisostomo," X p v a o m o p ~ ~ Sa t, u d i e ricerche (Rome,
1908), p. 1026; 0. Wulff and M. Alpatov, Denkmaler (1g25), pp. 61f., 296; V. N. Lazarev, Byzantine
Icons (1937), p. 250; S. Bettini, "Appunti" (1938), p. 17; V. N. Lazarev, Istorzia (1g47), p. 360;
W. Felicetti-Liebenfels, Geschichte, p. 65.
96 Illustrated in the Handbook of the Dumbarton Oaks Collection (rg55), p. 150.
96 The chief authorities for the life of St. John Chrysostom, Palladius, Socrates (lib.VI), Sozomen
(lib.VIII), Theodoret (lib.V), Isidore of Pelusium (Letters 11, q), are quoted by W. R. W. Stephens,
St. John Chrysostom. H i s Life and T i m e s (London, 1883); for his works and his veneration see Chr.
Baur, S. J e a n Chrysostome et ses oeuvres duns l'histoire littkraire (Louvain-Paris, 1907); H . Kellner,
"Die Verehrung des hl. Johannes Chrysostomus im Morgen- und Abendland," Xpvaomoyi~&, p. 1007ff.
 T W O P A L A E O L O G A N MOSAIC ICONS 111
Saint's claim to distinction is twofold: he was an intrepid fighter for the
freedom and integrity of the Church, and he was a brilliant orator. I t is
the latter quality for which he is especially remembered; it has earned him
the sobriquet of the Golden-Mouthed (used since the sixth century) and has
provided some of the stock motifs of the many epigrams that have been com-
posed to honor the Saint or to describe his images, especially in the middle and
late Byzantine periods. One of these ever-recurring conceits which have served
poets from classical times on, is the affected astonishment that the portrait of
the great preacher was obstinately Another common motif was that
St. John Chrysostom inherited the gift of words from St. Paul who had it
from C h r i ~ t .The
~ s most elaborate form of this idea can be found in later Byzan-
tine icons and wall paintings which show St. John writing at his desk with
St. Paul standing at his side and prompting him ; apostles, saints, and monks
surround the group.99This iconographic type was subsequently enriched with
additional motifs and developed into the nqyfi T ~ o So c p i a s , the earliest extant
example of which belongs to the late eleventh century.loOIn representations of
this subject the scroll on which the Saint is writing becomes a flowing fountain
of water from which monks and clerics are drinking.lOl A most interesting
variant of this allegorical theme is found in the painting at Lesnovo, ~ e r b i a
(middle of the fourteenth century), where the persons who are drawing and
drinking the life-giving water are explicitly described as painters in an in-
scription which reads :Oi L o y p a c p o t ptpo13[V]TE ~ f i vTEXV~TT)V $QIV uai KEPCTVUVTE (sic).Io2
Thus, St. John Chrysostom had become the preceptor of painters.
This was not, however, his only function in the thirteenth and fourteenth
centuries, a period when the combative patriarch and his writings gained
especial importance. After the Latins had taken Constantinople in 1204, they
not only transferred the greater part of his relics from the Church of the Holy
Apostles to St. Peter's in Rome,lo3but also made Chrysostom the patron of
the Latin Patriarchate of Constantinople, a choice which was suggested by the
fact that Chrysostom himself as well as his followers had solicited the help of
Rome, and appealed to the Pope's judgement in their struggle against Theo-
philus. I t was this fact as much as Chrysostom's endeavors to preserve the
97 Manuel Philes, ed. Miller, I, p. 58f. (no.CXXXV).
@a Cf. the inscription on fol. 7 of Ms. gr. 224 of the Bibl. Nat., Paris (H.Omont, Miniatures des plus
anciens manuscrits grecs de la BibliothBque Nationale [Paris, 19291, pl. CI) and in Cod. Jerusalem,
St. Saba 33 (XI saec.):
X p ~ m o im
l 6 p a W ~ ~ V K~bE llairhov m 6 p a
Z ~ 6 p a86 llairhov -rb Xpvooo~6povo ~ 6 p a .
See also A. Xyngopoulos, "Aylos 'IwCrvvqs 6 Xpvo6o-ropog "llqyfi T?S Eoqias," ' E ~ q p ~ p I'sA p x a ~ o -
h o y t ~ f i ,1942/44 (19481, P. I ff.
SQ Xyngopoulos, op. czt., figs. 1-6: icon in the Loverdos collection; Cod. Vat. gr. 766; Cod. Athen.
gr. 7 : fresco in Chilandari.
- l b o Homilies of St. John Chrysostom, Milan, Ambros. A. 172: (Xyngopoulos, op. cit., fig. 7); other
examples ibid.
10'-1n later examples like the fresco in the Aphendiko, Mistra, A.D. 1366 (Xyngopoulos, op. cit.,
fig. IO), SS. Gregory and Basil are represented side by side with John Chrysostom, with fountains
issuing from the writings of each.
102 S. RadojCi6, "Die Entstehung der Malerei der palaologischen Renaissance," Jahrbuch der &err.
Byzant. Gesellschaft, V I I (1958), p. 116.
l o 3 A A S S , Jan. 11, p. 760.
112 OTTO D E M U S
unity of the Church, that prompted the propagators of the ecclesiastic union
with Rome in the thirteenth and fourteenth centuries to turn to his writings.
John Chrysostom was, therefore, often quoted by the unionist patriarch of
Constantinople, John Beccus, in his work on Church Union, a book that helped
to bring about an ephemeral union of the two Churches at the Council of
Lyonsin 1274. Beccus was deposed and imprisoned in 1282 by Andronicus I1
who renounced the short-lived union.lo4I t would be tempting to try to connect
the Dumbarton Oaks icon, which is most likely to have been commissioned by
or for an important ecclesiastic, with John Beccus, but the style of the icon
points to a considerably later date. However, Beccus was not the last to at-
tempt a rapprochement of the two Churches. About two generations later,
Demetrius Cydones pursued the same policy, a policy which led eventually to
the Union of Rome in 1369.1°5 AS a matter of fact, a date towards the middle
of the fourteenth century would approximately fit the style of the mosaic icon.
In any case, the figure of St. John Chrysostom had gained an especial actuality
at that time, so that his portrait must have held a deep, perhaps even a
controversial meaning.
The use of the term "portrait" may appear questionable, but the represen-
tation of the Saint is so specific that it gives the impression of being actually a
portrait. This impression is strengthened on comparing the Dumbarton Oaks
icon with what contemporaries of John Chrysostom said about his character
and his personal appearance. I t is said of him that though dignified, his per-
sonal appearance was not imposing; that his stature was diminutive and his
limbs so emaciated that he himself compared his body to that of a spider
(&paxvcj6q~). His forehead is described as very lofty and deeply furrowed with
wrinkles, expanding widely at the top, his head bald "like that of Elisha,"
his eyes deeply set but keen and piercing. His cheeks were pallid and withered
and his chin pointed and covered with a short beard. Being dyspeptic, he
was considered morose; in addition, he is said to have been choleric, irritable,
impatient, and somewhat intolerant of other peoples' weaknesses, easily
offended, suspicious, and violent in his anger. Taken all in all, there emerges
from contemporary sources the picture of a very remarkable, but equally
difficult man, a picture which accords so well with the Dumbarton Oaks icon
that it is indeed tempting to see in the latter an actual portrait.
As there are many likenesses of John Chrysostom which present the same
individual characteristics, it is quite reasonable to speculate on the chance of
there having existed an authentic portrait of the Saint which might have been
the ultimate source of all later representations; one might even suppose that
such a portrait formed part of the series which is known to have existed in the
Patriarchate of Constantinople. To this it might be objected that portraits of
bishops and patriarchs who had been subjected to a damnatio memoriae were
removed and destroyed, a fate which a portrait of Chrysostom could hardly
have escaped, not to mention the wholesale destruction of saints' portraits-
104 p. 694.
See A. A. Vasiliev, History of the Byzantine Empire (Madison, 1958)~
105 Ibid. p. 694ff.
 T W O P A L A E O L O G A N MOSAIC I C O N S 113
and John was held a saint as early as 438-during the period of Iconoclasm.1o6
Nevertheless, an authentic pictorial tradition may be thought to have lingered
on, to blossom up again after the iconoclastic controversy had passed.
However, an examination of the earliest extant representations of the Saint
yields results which, at first glance, cannot be easily reconciled with this as-
sumption. For nothing could be further from a portrait in the specific sense
of this term than the paintings that are, in all likelihood, the oldest represen-
tations of John Chrysostom to have come down to us, all of them in the church
of Sta Maria Antiqua in Rome. The first of these107 forms part of a group of
four Church Fathers, standing to the right and left of the main apse; it is
identified by the inscription I O A N N , and can be dated, together with the
other figures, in the period of Pope Martin I, that is, in the middle of the seventh
century. Unfortunately, the face of the figure is seriously damaged: its lower
part, including the beard, is missing, but there is enough left to show that the
face did not agree either with the Dumbarton Oaks icon or with the early
descriptions of the Saint's features.
The same is true of the second representation of St. John in Sta Maria
Antiqua,lOs (fig. 24) one of a row of standing figures in the north aisle which
can be dated in the third quarter of the eighth century: Chrysostom is here
represented as a comparatively young man with an oval face fringed with a
sparse beard, without the slightest trace of the ascetic features of our mosaic.
The two frescoes, it is true, are the work of provincial Greek painters; but
these painters were firmly enough grounded in "official" iconography to
represent other saints "correctly." Furthermore, these painters were not alone
in thus delineating the features of St. John Chrysostom: a Vienna manuscript
of the Saint's Homilies on the Gospel of St. Matthew, written and illuminated
in Salzburg in the early ninth century, after an eastern model, shows the Saint
in a similar guise, as a young man with little hair and less beard.lOg Curiously
enough, these very words, vios, mavbs, 6Atyoyivqs, are used to describe St. John
Chrysostom in the Painters' Guide of Mount Athos.lloAs a matter of fact, John
was only fifty-three years of age when he died, and he rose to eminence -long
before he was fifty.
From these early representations and from the description of the Herwzerteia
it might be deduced that there existed, in pre-iconoclastic times, a recognized
iconographic type of St. John Chrysostom which had nothing to do with the
106 On this entire question cf. A. Grabar, L'iconoclasme byzantin, Dossier arche'ologique (Paris, 1957),
p.213f.
1°7 J. Wilpert, Die romischen Mosaiken, op. cit., p. 662, pl. 142b; E. Kitzinger, Romische Malerei
vom Beginn des 7. b i s zur Mitte des 8. Jahrhunderts (Munich, 1934), pp. 8, 41.
los J. Wilpert, op. cit., p. 708, pl. 192; E. Kitzinger, op. cit., p. 33. The row of saints t o which the
figure of St. John Chrysostom belongs is dated in the period of Pope Paul I (757-67) or Hadrian I
(772-95)-
log Vienna Xat. Bibl., Cod. 1332, fol. I": H. J. Hermann, "Die friihmittelalterlichen Handschriften
des Abendlandes," Beschreibendes Verzeichnis der illuminierten Handschriften in ~sterreich, VIII
(Leipzig, 1923)~p. 143; color reproduction in F. Unterkircher, La miniatura Austriaca (Milan-Florence,
19531, PI. I.
11° Dionysius of Fourna, The Painters' Guide ( ' E p y q v ~ i aTGVZwyphqwv), ed. A. Papadopoulos-
Kerameus (St. Petersburg, ~ g o g )pp.
, 128, 154, 222, 267.
8
114 OTTO D E M U S
highly individualized character portrait that appeared at a later period, perhaps
not too long after the anasteZosis of icons. This latter type seems to have been
developed by the beginning of the eleventh century at the latest. The Saint's
face, as it is depicted in the mosaics of Kievl1l (fig. 2 5 ) or the frescoes of
Ohrid112 (both of the second quarter of the century), is full of wrinkles and
furrows, pointed and angular in its features, with sharply defined ascetic
features-a type that could never be adequately described in the words used
by the Hermeneia, but that comes close to the Dumbarton Oaks icon. True,
the eyes are somewhat larger in the two eleventh-century portraits, and the
forehead is a little less bulbous, but all the other features are already there:
the emaciated cheeks outlined with a V-shaped design, the heavy dark circles
under the eyes, the drooping moustache, the sparse, two-pronged beard, the
characteristic distribution of the thin, brown hair, and the curiously pointed,
triangular shape of the face.
Between these works, the two monumental representations on the one hand,
and the small icon on the other, there is a difference in date that amounts to
about three centuries, within which lies the entire development from the late
Macedonian to the full Palaeologan style-yet, the basic concept of the portrait
underwent little change. Of course, this iconographic formula was subjected to
different interpretations in the various phases of its development. The
forceful characterization of the Saint in the works of the eleventh century113
was replaced, a hundred years later,l14 by an extremely delicate and psycho-
logically refined representation, in which Chrysostom is shown, not as the
fighting Patriarch, the pillar of orthodoxy, but as an ascetic introvert (fig. 26).
His ascetism becomes, from that period on, one of the chief motifs of epigram-
matic poetry. The Palaeologan poet Manuel Philes, for instance, has left us
several epigrams on icons of the Saint. One of these expresses astonishment
that the Saint's portrait could be so much alive in view of the fact that he had
starved himself so rigorously in his real life that there was scarcely any life in
him. In another, the painter is admired for having painted the shadow of a
shadow; in yet another, for having succeeded in painting Chrysostom with
material colours (fiha~s)whereas there was little or no earthly matter in the
Saint's body during his lifetime.l15 The point in this and other epigrams is
driven home mercilessly by the repetition of words like &aap~os(fleshless),
dtol-ria (lack of food), ~ E T I T ~ T(thinness),
~S etc.
The question that now arises is whether the specific potrait-type of the Saint
111 0. Powstenko, op. cit., pl. 75.
112 0. Bihalji-Merin, Fresken und Ikonen, Mittelalterliche Kunst i n Serbien und Makedonien (Munich,
1958), PI. 8.
11s Here are a few portraits of Chrysostom, chosen a t random from works of the eleventh century:
Harbaville triptych; London, Victoria and Albert Museum, bronze triptych; St. Mark's, Pala d'oro;
Oxford, Bodleian, Roe 6, fol. I ; Oxford, Merton 28, fol. 3 ; Sinai, Cod. 364, fol. I ; Paris, Bibl. Nat.,
gr. 224, fol. 7 ; Coislin 79, fol. zv; Geneva, Bibl. MS 24.
114 Twelfth century: Palermo, Cappella Palatina; Cefalh; Monreale; Spas Nereditsa; Paris, Bibl.
Nat., gr. 1208.
116 Manuelis Philae carmina, I, p. 33 f. (nos. LXIX, LXX, LXXI, LXXII, LXXIII), p. 58 f.
(CXXXV), p. 319 (CCXXIX). The conceit of the shadow's shadow is, of course, derived from Plotinus;
cf. A. Drews, Plotin (Jena, 1go7), p. 41 f.
 T W O PALAEOLOGAN MOSAIC ICONS 115
was actually created in the post-iconoclastic period or whether it was a much
older one that was rediscovered and re-elaborated after the iconoclastic em-
bargo had been withdrawn. There is much to be said for the second alternative
which has been championed by Andr6 Grabar who has come to the conclusion
that the late portraits of St. John Chrysostom "gardent. . . les traces kvidentes
du chef-d'oeuvre disparu du Ve sikcle auquel elles remontent toutes."l16 To
harmonize this view with the objections stated above, it must be assumed that
the realistic portrait type which may have gone back to the time of Chrysostom
himself, was temporarily superseded by a more generalized and abstract manner
of representing the Saint's features. I t has been shown by E. Kitzinger that
there was indeed a strong abstract current in late sixth- and seventh-century
art and that it is this style, and not that of contemporary Hellenism, that is to
be "found most frequently associated with the portrayal of holy persons or
Saints."l17 To put it more concisely, the late sixth century was the time when
a saint's image lost its portrait character in order to acquire that of an icon.118
I t was the increased veneration accorded to images that led to this change119
which can be described as the almost exclusive use of "realistically non-
descript" types and abstract forms in Byzantine icon painting of the century
and a half that preceded the outbreak of iconoclasm. The victory of orthodoxy
in the middle of the ninth century brought about, not so much a return to the
abstract iconic images that were in use immediately before the controversy,
as a re-establishment of the realistic types that had preceded them. This
re-establishment, however, was only partial, and the process of re-creating and
"mediaevalizing" each particular portrait-typedidnot take place independently
for, as was almost the rule in the development of Byzantine iconography,
the crystallization of one "type" was assisted and influenced by the develop-
ment of related types.
Thus, an ascetic figure represented in the drum of the Palatine Chapel, of
about the middle of the twelfth century120 (fig. 27), shows exactly the same
features as contemporary representations of St. John Chryso~tom~~~-hollow
cheeks, aquiline nose, bulbous forehead, short two-pronged beard, and sparse
hair-the only difference being one of costume (figs. 26, 28). However, this
mosaic of the Palatina does not represent John Chrysostom; it depicts the
Prophet Jonah. One might, at first, be tempted to suggest that the mosaicist
of Palermo, thinking perhaps of the intrinsic similarity between the two
embittered censors of a great city's vanities, simply used the likeness of
118A. Grabar, Martyrium, I1 (Paris, 1946)~p. 25, note 2.
117E. Kitzinger, "Art in the Period between Justinian and Iconoclasm," Berichte zum XI. Internat.
Byzant. Kongress (Munich, 1958), IV/I, passim, esp. p. 45.
11s E. Kitzlnger, "On some Icons of the Seventh Century," Late Classical and Mediaeval Studies in
Honor of A . M. Friend, Jr. (Princeton, 1955)~p. 13zff., esp. p. 145.
119 E. Kitzinger, "The Cult of Images in the Age before Iconoclasm," Dumbarton Oaks Papers, 8
(1954), P. 85ff.
lZ0 0. Demus, The Mosaics of Norman Sicily, pp. 39, 64, pl. 12. I am indebted to the Dumbarton
Oaks Research Library and Collection for permission to reproduce this photograph as well as that
for figure 29 (St. Luke).
lZ1 Cf., for example, the portrait of the Saint likewise in the Palatina (0.Demus, op. cit., pl. 23 B)
or, better, a three-quarter profile representation like that of the Paris Coislin 79 (fig. 28).
8*
116 O T T O DEMUS
Chrysostom to depict Jonah or vice versa. However, the case is more compli-
cated, because the same features are found in the usual type of one of the
saints most frequently portrayed, the Evangelist St. Luke. This type122(fig. 29)
is distinguished from the portrait of John Chrysostom only by the different
hairdress : St. Luke is represented with ample curls, arranged in two tiers, a
"coiffure" which covers a large part of the forehead and thus changes the entire
aspect of a face that is otherwise almost identical with that of Jonah in the
Palatine Chapel and with the most common type of St. John Chrysostom. If
it could be proved that this bearded type of st: Luke were considerably older
than the post-iconoclastic "portrait" of John Chrysostom, one might be able
to assume that the latter followed the model of the dominant, because more
frequently represented, type of the Evangelist .123 However, the origins of this
latter can be traced back hardly further than the second quarter of the ninth
century;12*the fully developed type is found only in the eleventh ~ e n t u r y . 1 ~ ~
Thus, the crystallization of the portrait type of St. Luke happened at the same
time as that of the Chrysostom portrait; the two developments were parallel
and, it seems, interrelated. There is even a strong probability that the portrait
of John Chrysostom was originally "in the lead" since it may actually have
been derived from a genuine portrait tradition, while the representation of
St. Luke had no such roots. The more specific and better founded "author's
portrait" might, therefore, have become, to a certain extent, the model of the
"holier" Evangelist's portrait. The fact, on the other hand, that the represen-
tation of the latter was much more widely disseminated than that of the former,
may have speeded up the development of a highly differentiated, "ascetic"
type which, in its turn, may again have influenced the evolution of the
"character-portraitJ' of John Chrysostom.
The two processes seem to have been practically completed by the beginning
of the eleventh century and seem to have been more or less universal; more
or less, but not compl~telyso. For, among the representations of both saints
there can be found an exceptional type which does not conform to those I
have described. In the case of the Evangelist, it is the image of a beardless
youth with short, dark hair, low forehead, straight, thick nose, and strong chin,
a "Hellenistic" type which is found in its purest form in the Gospelbook no. 43
of S t a ~ r o n i k i t a and
, ~ ~ ~which survives into the thirteenth and fourteenth
centuries in illuminated books inspired by tenth-century prototypes.127If this
survival of a Hellenistic type is hardly more than an interesting sideline of the
main iconographic development of St. Luke, the matter stands differently as
1z2 A. M. Friend, Jr., "The Portraits of the Evangelists in Greek and Latin Manuscripts," I, A r t
Studies (1927), p. I 15 ff.
123 This was t h e opinion-wrong, as I now believe-which I expressed in my Symposium lecture
at Dumbarton Oaks in 1958.
1z4 Rome, Vat. gr. 699, Cosmas Indicopleustes ( A. M. Friend, Jr., op. cit., fig. 72) ; Paris, Bibl. Nat.
gr. 70, Gospels, (ibid., fig. 3.)
lZ5 Cf. Nicaea, Koimesis church, mosaic (ibid., fig. 123); further examples (very numerous),
e.g. ibid., figs. 127, 130, 134, 146.
lZ6 A. &I. Friend, Jr., op. cit., fig. 97.
lZ7 E.g. Paris gr. 54 (ill. in V. N. Lazarev, Istoviia, 09. cit., 11, pl. 2 5 2 ) ; on the use of this type by
Cimabue, see 0. Demus, "Die Entstehung," op. cit., p. 40, fig. 32.
 T W O P A L A E O L O G A N MOSAIC I C O N S 117
regards St. John Chrysostom, since there are several representations of the
Saint which are quite different from the common type, and which must be
accorded a certain importance because of their high artistic merit and their
central position, being, as they are, works produced in the capital. As a matter
of fact, the earliest representation of St. John Chrysostom that has been
preserved in Constantinople shows this different, "secondary" type. I t is the
impressive mosaic portrait of the Saint in Hagia Sophia, in the series of saints
and Church Fathers on the north tympanum128 (fig. 30). The Patriarch is
represented full length, clad in the ample garments of his high office. The
elongated oval of his face is surrounded with a short beard and dark hair; he
is not bald, his features are neither pointed nor emaciated, he has a strong
chin and nose; in short, there is hardly any resemblance between this and the
familiar ascetic type. Since the grandiose mosaic of Hagia Sophia was probably
made in the late ninth century and is, therefore, earlier than the oldest examples
of the ascetic portrait, it might be thought to antedate the formation of the
latter type, in which case it could be discounted for the purposes of this enquiry.
However, the Hagia Sophia image is not isolated. A related figure appears on
the lid of the reliquary of theTrue Cross in the treasure of the Sancta Sanctorum
in Rome, a Constantinopolitan work of the tenth century129(fig. 31). The head
and the body are more elongated, so that the Saint appears almost like a giant,
but the features are so similar to those of the Hagia Sophia mosaic that the
two portraits must be assumed to belong to a common tradition. And, should
the Roman icon also be thought too early, there is another Constantinopoli-
tan representation of the same type which belongs to a much later period: the
recently uncovered fresco of St. John Chrysostom in the apse of the Par-
ecclesion of the Kariye Camii130 (fig. 32). But for the difference of garments,
the figure is very similar to that of the Sancta Sanctorum reliquary and
the face is essentially the same, except that it is rendered in a different
style.
Thus, the abstract iconic type as it is preserved in the early frescoes of
Sta Maria Antiqua, and the realistic, ascetic portrait type which was the
standard one from the eleventh century onwards were not the only icono-
graphic possibilities of rendering the features of St. John Chrysostom ; there
was yet a third type which depicted the Patriarch in an "idealized" manner,
equally far removed from the abstract as from the realistic modes. I t does
not depict the saint, as did the iconic type, nor does it portray the ascetic, as
did the realistic type: instead, it characterizes the patriarch and the scholar.
I t might fittingly be called the humanistic type. There is nothing surprising
in the fact that the humanistic portrait of the Patriarch should have survived
in Constantinople, since it is very likely that this type was created in the
128 Color illustration in A. Grabar, L a fieinture byzantine, p. 96.
128 Ph. Lauer, "Le Trksor du Sancta Sanctorum," Monuments et Me'moires de l a Fondation Piot,
XV (1906), p. 92, pl. x ~ v b .
130 P. A. Underwood, "Second preliminary Report on the Restoration of the Frescoes in the Kariye
Camii at Istanbul by the Byzantine Institute, 1955," Dumbarton Oaks Papers, 11 (1g57), p. 212ff.,
fig. 47.
 OTTO D E M U S
capital. There exist, it is true, Constantinopolitan examples of the other types;
but the opposite, the existence of the humanistic type in the provinces, would
be difficult to prove.
The three types did not therefore, exist on equal terms, either locally or
chronologically. The iconic type, having perhaps replaced a lost fifth-century
portrait, held its sway in the time between Justinian and the iconoclastic
controversy, while the ascetic type seems to have dominated later Macedonian
and Comnenian art. As to the third type, it may be of some significance that
the specimens I have adduced belong to the ninth, tenth, and fourteenth
centuries, the Macedonian and Palaeologan Renaissances respectively. I t was
in these Renaissance periods, in the capital itself, that the humanistic type
came to the fore.
The Dumbarton Oaks portrait -to return to our point of departure -al-
though generally conforming to the ascetic type, also contains some features
of the humanistic portrait; to mention only one, it shows a somewhat fuller
growth of hair and beard. hforeover, the features are rendered not by a system
of lines but by a modelling technique that is specifically painterly, a technique
that is more akin to the humanistic than it is to the ascetic type. Soft gra-
dations, transparent shadows, and impasto highlights combine to produce an
effect of brush work, as in water-color or fresco painting. This technique does
not employ any of the optical or calligraphic abbreviations that are so charac-
teristic of the mosaics of the Kariye Camii and, mutatis mutandis, of the
frescoes in the Parecclesion of the same church. However, very close parallels
to the soft, detailed "tvash" technique of the miniature mosaic can be found
in another part of the Kariye, namely in the decoration of an arcosolium tomb
in the fifth bay of the outer narthex.131 In particular, the bust of St. John
Damascene invites comparison as regards the modelling of the face. Although
the tomb is not dated exactly, it is likely that its paintings belong to the second
quarter or the middle of the fourteenth century, a date which also appears to
be appropriate for the Dumbarton Oaks icon. Thus this icon seems to belong
to a period considerably later -by at least one, possibly two generations -than
that of the icon of the Forty Martyrs. If the latter exemplifies the art of the
early Palaeologan era with its experimental character, then the portrait of
St. John Chrysostom is typical of the completely developed style with its
absolute sureness of touch, a style which already shows the first, faint signs
of becoming dry, even tired. True, the difference in subject matter may be held
partly to account for the difference in "key." The icon intended to evoke the
somewhat arid personality of the great churchman and ascetic, is executed in
a severe, abstract key. I t was not only the formal tendency prevalent in his
time but also the requirements of his subject that led the mosaicist towards
the realm of iconic geometry, expressed in the upright rectangle of the book,
the regular pattern of the fiolystavrion, the plastic halo decorated with a
geometric all-over pattern of tiny crosses-an element of the frame invading
131 P. A. Underwood, "Kotes on t h e Work of t h e Byzantine Institute in Istanbul, 1955-1956,''
Dunzbarton Oaks Papers, 12 (1958), p. 269 ff ., esp. 279f., figs. I 1-13.
1. Dumbarton Oaks Collection. Mosaic Icon, The Forty Martyrs
2. Dumbarton Oaks Collection. Mosaic Icon, The Forty Martyrs, detail
 . r._

3. Dumbarton Oaks Collection. Mosaic Icon, The Forty Martyrs, detail

 5. Vatican, Cod. syr. 559, Part 1, fol. 93"

6. Rome, Sta Maria Antiqua, Chapel, Apse. Wall Painting

7. Moscow, Synodal Library. Menologium no. 183

1 x

A h -
-&
-.

8. Lesnovo. Wall Painting

 12. Berlin-Dahlem, Ehem. Staatliche Museen. Ivory

11. DePani. Wall Painting

13. Leningrad, Hermitage. Ivory Triptych,

center Panel
22. Dumbarton Oaks Collection. Mosaic Icon, St. John Chrysostom
23. Dumbarton Oaks Collection. Mosaic Icon, St. John Chrysostom, detail
26. Palermo, Cappella Palatina. Mosaic, detail
27. Palermo, Cappella Palatina. Mosaic, detail

28. Paris, Biblioth6que Nationale. Coislin 79,

fol. 2", detail 29. Palermo, Cappella Palatina. Mosaic, detail
 T W O PALAEOLOGAN MOSAIC I C O N S 119
the icon-the formal calligraphy of the inscription and, finally, the inner frame
of the mosaic: all these elements help to create a hieratic distance.
No such distance separates the Forty Martyrs of the other Dumbarton Oaks
mosaic from the beholder; there is hardly any framing motif, there is no
geometric, not even a symmetric, composition: the suffering and the heroism
of the Martyrs are presented without being formalized by decorative geometry.
There is an immediacy about this icon which is far removed from the liturgical
atmosphere of St. John Chrysostom's portrait.
But above and beyond the difference in "mode,"132 imposed on the two
artists by their subject matter, there is, undoubtedly, a strong discrepancy of
style. To recognize this discrepancy it is enough to confront the face of St. John
Chrysostom with a face of a similar ascetic type in the icon of the Forty
Martyrs, e.g. the fifth face from the right, in the middle row. The modeling of
the latter has an optical, an illusionistic, even a dramatic quality that is totally
lacking in the portrait icon. I t is, in the last resort, the antique, the renaissance
quality that is missing in the mosaic of St. John Chrysostom: one feels that
the time of the artistic rediscovery of Hellenism was over when this portrait
was made. Both mosaics are masterpieces of Palaeologan art; but only one
of them is a work of the Palaeologan Renaissance.
13%For the concept of "mode" as distinct from "style," see E. Kitzinger, Byzantine Art, passim,
esp. P. 39.