STUDIES IN

LATE ANTIQUITY AND EARLY ISLAM

1

THE BYZANTINE AND EARLY

ISLAMIC NEAR EAST
VI
ELITES OLD AND NEW IN THE BYZANTINE AND EARLY
ISLAMIC NEAR EAST

EDITED BY

JOHN HALDON AND LAWRENCE I. CONRAD

(Papers of the Sixth Workshop on

Late Antiquity and Early Islam)

THE DARWIN PRESS, INC.

PRINCETON, NEW JERSEY
2004
Copyright © 2004 by THE DARWIN PRESS, INC., Princeton, NJ 08543.
All rights reserved.

No part of this publication may be reproduced, stored in a retrieval system, or

transmitted, in any form, by any means, electronic, mechanical, photocopying,
recording, or otherwise, without the prior permission of the publisher, except in
the case of brief quotations in critical articles or reviews.

Library of Congress Cataloging-in-Publication Data

Workshop on Late Antiquity and Early Islam (1st: 1989: London, England)
The Byzantine and early Islamic Near East: papers of the First Workshop on Late
Antiquity and Early Islam / edited by Averil Cameron and Lawrence I. Conrad
p. cm. - Studies in late antiquity and early Islam: 1-)
Includes bibliographical references and index.
Contents: 1. Problems in the literary source material.
ISBN 0-87850-080-4 (v. 1): $29.95
1. Middle East—History—To 622—Historiography. 2. Islamic Empire—History—622-
661—Historiography. 3. Islamic Empire—History—661-750—Historiography. I. Cameron,
Averil. II. Conrad, Lawrence I., 1949- . III. Title. IV. Series: Studies in late antiquity
and early Islam: 1, etc.
DS62.25.W67 1989
939.4'0072-dc20 92-352
CIP
ISBN 0-87850-144-4

The paper in this book is acid-free neutral pH stock and meets the guidelines for perma-
nence and durability of the Committee on Production Guidelines for Book Longevity of
the Council on Library Resources.

Printed in the United States of America

•RESS, INC., Princeton, NJ 08543.

reproduced, stored in a retrieval system, or

ELITES AND PATRONAGE IN EARLY
leans, electronic, mechanical, photocopying, BYZANTIUM: THE EVIDENCE FROM
prior permission of the publisher, except in
1 articles or reviews. HAGIOS DEMETRIOS AT
THESSALONIKE*
in-Publication Data
Islam (1st: 1989: London, England) Leslie Brubaker
;ar East: papers of the First Workshop on Late
^veril Cameron and Lawrence I. Conrad
(University of Birmingham)
y and early Islam: 1-)
nd index.
y source material.
THIS PAPER deals with material culture, and with monuments or artefacts
listoriography. 2. Islamic Empire—History—622- commissioned by private individuals or families. Its focus is exclusively on
-History—661-750—Historiography. I. Cameron, images, and on images that can be pinned to people acting primarily in
. III. Title. IV. Series: Studies in late antiquity their own interests rather than in the service of institutions, ecclesiasti-
cal or imperial. We can now determine that an image was commissioned
92-352 by a specific person only when we are told so in a surviving text or in-
CIP scription, or when the image itself provides the information by incorporat-
0-87850-144-4 ing a portrait of the donor.1 Such portraits provide our only information
on the self-presentation —though always via the mediating artisan—of the
patron, and it is accordingly on this group of images that I shall focus
here.
The most important collection of patron portraits produced during the
ral pH stock and meets the guidelines for perma-
early Byzantine centuries (ca. 500-ca. 750) are the sixth- and seventh-century
on Production Guidelines for Book Longevity of
mosaics commissioned for the church of Hagios Demetrios by the wealthy in-
habitants of Thessalonike to commemorate their patron saint Demetrios, his
holy companions, and themselves.2 The panels have been interpreted as ex
*In addition to the participants of the workshop, and particularly Mark Whitlow, I
thank Nancy Palterson Sevcenko and Chris Wickham for comments on an earlier draft of
this article.
1 On donor portraits as a genre, see Nancy Patterson Sevccnko, "The Representation of
Donors and Holy Figures on Four Byzantine Icons," Ddtion, Ser. 4, 17 (1993 94), 157-64;
eadem, "Close Encounters: Contact between Holy Figures and the Faithful as Represented
in Byzantine Works of Art," in Andre Guillou and Janinc Durand, eds., Byzance e.t les
images (Paris, 1994), 255-85.
2 Colour reproductions of most surviving panels appear in Chrysanthe Mauropoulou-
Tsioume, Byzantine Thessalonike (Thessalonikc, 1993), 76-79. The most recent study of

63
 64 Leslie Bru baker

voto images, made to thank Demetrios and the other holy figures pictured
for some previous intervention on the donor's behalf.3 In most cases, nei-
ther the saints nor the donors are named; we can, however, rest assured that
any individual or family who could afford to commission a panel recording
Demetrios' favour was sufficiently wealthy (and socially significant) to be
counted amongst the elite.4 Though the point is rarely made, these panels
are significant not only as votive images recording the special importance of
S. Demetrios in early Byzantine Thessalonike, but as portraits of the city
notables who paid to have themselves displayed as recipients of the saint's
grace on the walls of Dernetrios' church for posterity. They tell us about
evolving definitions of sanctity; but they also tell us about how patronage
was used in the construction of elite identity in early Byzantine Thessa-
lonike.
The main point of this chapter is that the mosaics at Hagios Demetrios
track changes in the way that elite identity was represented in Byzantium:
as Demetrios was transformed from a healing saint into a focus of urban
identity,5 his elite clients changed from individuals and families to repre-
sentatives of civic authority. A second point concerns a series of portraits
that have usually been interpreted as representing different stages in the de-
velopment of a single child. This interpretation is, I will argue, incorrect;
rather than identifying a single individual, the motif of the gold cross on
the foreheads of four children indicates that they have been miraculously
healed.
Modern understanding of the panels has been complicated by their tran-
sience and, to a lesser extent, by problems surrounding their dating. Re-
discovered only in 1907, when the building (a mosque since 1492-93) was

Demctrios, his church and his cult, is James C. Skedros, Saint Demetrios of Thessaloniki,
Civic. Patron and Divine Protector 4th-7th Centuries CE (Cambridge, MA, 1999); though
this appeared after the present study had been drafted, Skcdros makes some good points
about the relationship between the Thcssalonikan elite and the cult of S. Demetrios that
I have signalled briefly in the notes.
3See esp. Andre Grabar, Mnrtyrium. Recherches sur le culte des reliques et I'art
chretien antique, 2 (Paris, 1946), 87-100; Hans Belting, Likeness and Presence: a History
of the Image before the Era of Art, trans. Edmund Jephcott (Chicago, 1994), 82-88.
4See Robin Cormack, "The Making of a Patron Saint: the Powers of Art and Ritual
in Byzantine Thessaloniki," in Irving Lavin, cd., Themes of Unity and Diversity, Acts of
the XXVI International Congress of the History of Art 3 (University Park, 1989), 547-
56.
5See Skedros, Saint Demetrios, esp. 70-82, 100-102, 115-20.

V
 Leslie Brubaker Elites and Patronage in Early Byzantium 65

os and the other holy figures pictured renovated, most of the works were destroyed a decade later by the disas-
e donor's behalf.3 In most cases, nei- trous fire of 1917. Records of the monument produced before the fire are
icd; we can, however, rest assured that not exhaustive,6 and most questions about the commissions that require ar-
ifford to commission a panel recording chaeological or technical stylistic information will never be answerable. The
ealthy (and socially significant) to be excavation and consolidation of the building's remains after 1917 did, how-
the point is rarely made, these panels ever, reveal a few additional mosaic panels that had been concealed behind
j;es recording the special importance of Ottoman plaster for centuries.7 Including these, nine panels survived the fire
ssalonike, but as portraits of the city and are still in place inside of the church; another sequence, destroyed in
s displayed as recipients of the saint's 1917, is known through photographs taken shortly after 1907 and through
irch for posterity. They tell us about drawings and watercolours made by N.K. Kluge and W.S. George on the site
they also tell us about how patronage at the same time.8 A niche on the outer south wall of the apse also retains
e identity in early Byzantine Thessa- a fragmentary mosaic, with the remnants of a votive inscription.9 The rel-
atively well-preserved extant panels are, however, all inside of the building,
that the mosaics at Hagios Demetrios a basilica with a nave flanked by two aisles on either side (Fig. 1). One is
ientity was represented in Byzantium: on the short wall formed by the end of the inner entrance tribelori, facing
a healing saint into a focus of urban the inner north aisle (Fig. 6); two cover sections of the entrance (west) wall,
•om individuals and families to repre- and are visible from the side aisles (Figs. 7-8); the other six occupy faces of
nd point concerns a series of portraits the two pillars located immediately before the sanctuary at the east end of
representing different stages in the de- the church, and two of them—placed on the east-facing side of the southeast
terpretation is, I will argue, incorrect; pier—could have been visible only to the clergy for whom the eastern apse
vidual, the motif of the gold cross on was reserved (Figs. 9-14). The lost sequence, now best known from George's
ites that they have been miraculously watercolours (Figs. 2-5), occupied the inner face of the arcade that separated
the two aisles on the north side of the church. The early fresco decoration
els has been complicated by their tran- that supplemented the mosaic work has largely perished.
•oblems surrounding their dating. Re-
milding (a mosque since 1492-93) was 6 On the discovery and recording of the mosaics see Robin Cormack, "The Mosaic
Decoration of S. Demetrios, Thessaloniki: a Re-examination in the Light of the Drawings
i C. Skedros, Saint Demetrios of Thessaloniki, of W.S. George," Papers of the British School at Athens 64 (1969), 17-19 (repr. in idem,
Centuries CE (Cambridge, MA, 1999); though The Byzantine Eye: Studies in Art and Patronage [London, 1989], as Study I); and Andre
ieen drafted, Skcdros makes some good points Grabar, "Notes sur les mosaiques de Saint-Demetrios a Salonique," Byzantion 48 (1978),
onikan elite and the cult of S. Demetrios that 64-69.
7The excavation was directed and published by George and Margaret Soteriou, He.
Recherches sur le cults des reliques et I'art basilike t,ou Hagiou Demetriou, Thessalonikcs (Athens, 1952).
ans Belting, Likeness and Presence: a History 8 Published and discussed by Cormack, "Mosaic Decoration," and idem, "The Church
Mmund Jephcott (Chicago, 1994), 82-88. of Saint Demetrius: the Watercolours and Drawings of W.S. George," Catalogue of an
i Patron Saint: the Powers of Art and Ritual Exhibition Organised by the British Council (Thessalonike, 1985), 44-72 (repr. in idem,
i, ed., Themes of Unity and Diversity, Acts of Byzantine Eye, as Study II).
istory of Art 3 (University Park, 1989), 547- 9 Soteriou reconstructed the inscription as [ikep sO/^? name] KAI THE STMBIOT
ATTOT IQ[dvva<;]: Hagios Demetrios, 198-99, PI. 71b; Andreas Xyngopoulos, Ta Psephi-
82, 100402, 115-20. dota tou naou ton Hagiou Dcmctriou Thessalonikcs (Thessalonikc, 1969), 30-31, PI. 28.
66 Leslie Brubaker

Description of the Mosaics

The mosaic panels from Hagios Demetrios have been described many times;10
here I will note only the features pertinent to subsequent discussion.
TRIBELON WALL, FACING INNER NORTH AISLE (Fig. 6): The badly dam-
aged panel shows a nimbed male—presumably S. Demetrios flanked by four
ecclesiastics, two on each side, standing before a curtain. Demetrios has his
arms around the shoulders of the two figures nearest to him, both of whom
are dressed as bishops; the lateral figures may be identified as a priest (right)
and a deacon (left), the latter of whom reappears on the east face of the
southeast sanctuary pier and in a medallion inserted into the inner north ar-
cade sequence. The faces of the central three figures have been removed, but
enough remains to determine that Demetrios had a circular nimbus; the oth-
ers have square haloes.11 Demetrios' chlarnys is decorated with a star-filled
medallion and trilobes with basal ribbons, both variants on the fabric design
worn by S. Sergios on the southeast sanctuary pier. Another motif shared
by the tribelon panel and the pier mosaics is the framing pattern: all but
the panel of Demetrios and the builders (on the north face of the southeast
pier) are framed with an inner row of grey-blue tesserae followed by a row of
black, surrounded by a double row of white cubes.12
WEST WALL, INNER NORTH AISLE (Fig. 7): The upper left corner of a
panel, framed in red and white triangular shapes, shows the upper torso of
a youthful nimbed saint in a chlamys with tablion (presumably Demetrios)
standing against a cloud-filled sky; a half-figure of an angel blowing a golden
trumpet swoops down above the saint, arid the wings of a second angel are
visible at the right edge of the fragment.13
WEST WALL, INNER SOUTH AISLE (Fig. 8): On the south side of the
west wall, the right portion of a panel has been preserved. This shows an
orant nirnbed saint with golden hands- again presumably S. Demetrios
1()The most complete description of the extant panels in English appears in R..F. Hod-
dinott, Early Byzantine. Churches in Macedonia and Southern Serbia (London, 1963),
141-55; for the lost mosaics of the north arcade, see Cormack, "Mosaic Decoration".
u Hoddinott, Early Byzantine Churches, 155, PI. 33; Xyngopoulos, Hagios Demetrios,
Pis. 8-10.
12The pier panels supplement this with a thick band of red followed by another white
border, but it is no longer clear whether or not this detail appeared on the damaged
tribelon mosaic.
13Hoddinott, Early Byzantine, Churches, 142-43, PL IV; Xyngopoulos, Hayios
Demetrios, Pis. 5-7; Robin Cormack, Writing in Gold: Byzantine Society and its Icons
(Oxford, 1985), 79-80, 82-83, Fig. 22.
 Leslie Drubaker Elites and Patronage in Early Byzantium 67

standing in front of a structure perhaps meant to represent the main focus

of Demetrios' cult in the church, his silver kiborion (see below).14 To the
,rios have been described many times;10
viewer's right, an adult male propels a youth toward the saint; both ex-
iinent to subsequent discussion.
tend covered hands toward him, and this gesture is repeated by a fragmen-
NORTH AISLE (Fig. 6): The badly dam-
tary figure on the left.15 The frame is composed of gold lozenges against a
sumably S. Demetrios -flanked by four
greenish-blue background; the inner rim is gold, the outer one white.
ng before a curtain. Demetrios has his
NORTHEAST SANCTUARY PIER, WEST FACE (Fig. 9): Beardless and
) figures nearest to him, both of whom
nimbed, a standing saint in a tablion chlamys decorated with trefoils set
ires may be identified as a priest (right)
within lozenges lifts his right hand (palm outwards) and rests his left on
lorn reappears on the east face of the
the shoulder of a small child with raised and covered hands; an even shorter
iallion inserted into the inner north ar-
child, also with raised and covered hands, stands beneath the saint's blessing
al three figures have been removed, but
hand.16 The figures stand on a light brown undulating ground line topped
metrics had a circular nimbus; the oth-
with two rows of yellow-green tesserae. The backdrop consists of a green
chlamys is decorated with a star-filled
"wall" with coloured bands along the top, one a thick red strip decorated
jons, both variants on the fabric design
with alternating ovals and vertical lines, and with a tent-like piece of drap-
sanctuary pier. Another motif shared
ery (?) at the top that appears to mimic the roof shape of the structure
aosaics is the framing pattern: all but
tentatively identified as the kiborion in the west wall, south aisle mosaic
,ers (on the north face of the southeast
described above. The saint is usually identified either as Demetrios or as
' grey-blue tesserae followed by a row of
Bakchos, the latter because his cohort Sergios occupies the corresponding
F white cubes.12
west face of the southeast sanctuary pier.
LE (Fig. 7): The upper left corner of a
NORTHEAST SANCTUARY PIER, EAST SIDE (Fig. 10): The badly damaged
gular shapes, shows the upper torso of
panel shows an orant saint wearing the same fabric, set against the same
s with tablion (presumably Demetrios)
backdrop, and standing on the same type of groundline as the saint on the
half-figure of an angel blowing a golden
west face. Below the panel is the inscription "a prayer for one whose name
it, and the wings of a second angel are
God knows", a formula found also in the mosaics of the north arcade (see
ent.13
below).17
;LE (Fig. 8): On the south side of the
lei has been preserved. This shows an
14For another saint with golden hands—this time S. Stephen, but also part of a votive
ids—again presumably S. Demetrios composition- usually dated to ca. 525, see the references to Dyrrachium in n. 64 below.
ixtant panels in English appears in R..F. Hod- 15 Hoddinott, Early Byzantine Churches, 142-43, PL IV; Ernst Kitzingcr, "Byzantine

edonia and Southern Serbia (London, 1963), Art in the Period between Justinian and Iconoclasm," Berichte zum XI. internatlonalen
•cade, see Cormack, "Mosaic Decoration". Byzantinisten-Kongress 4.1 (Munich, 1958), Fig. 22; repr. in idem, The Art of Byzantium
155, PI. 33; Xyngopoulos, Hagios Demetrios, and the Medieval West: Selected Studies, ed. W.E. Kleinbauer (Bloomington, 1976), Essay
VI; Xyngopoulos, Hagios Demetrios, Pis. 1-4; Corrnack, Writing in Gold, 80-83, Fig. 23.
a thick band of red followed by another white ""Excellent colour reproductions in Nano Chatzidakis, Byzantine Mosaics (Athens,
or not this detail appeared on the damaged 1994), Pis. 16-17; Hoddinott, Early Byzantine Churches, 153, PL 33a; Kitzinger, "Byzan-
tine Art in the Period between Justinian and Iconoclasm," 25, Fig. 23; Xyngopoulos,
,es, 142-43, PI. IV; Xyngopoulos, Hagios Hagios Demetrios, Pis. 15-18.
ting in Gold: Byzantine Society and its Icons lV TnEP ETXHS OT [o]IAEN 0 0[eo]S TO ONOMA: Xyngopoulos, Hagios De.me.-
trios, PL 27.
68 Leslie Bruhaker

NORTHEAST SANCTUARY PIER, SOUTH SIDE (Fig. 11): On the adjacent,

but broader, south side of the same pier, the Virgin and S. Theodore stand
on an identical undulating light brown ground topped with rows of yellow-
green cubes, against a continuation of the same "wall", now blue but with
the same red band decorated with alternating ovals and vertical lines that
are here sufficiently well preserved to identify as a simple fieur-de-lys motif.18
The position of the Virgin appears to have been altered: the "shadow" of her
former right shoulder is clearly visible. She holds an unfurled scroll on which
is written: "Supplication. Lord God, hearken to the voice of my prayer,
for I pray for the world."19 The orant Theodore, identified by his beard and
facial type,20 stands frontally, in a tablion chlamys decorated with a pattern
identical to that on Demetrios' chlamys on the east face of the southeast
pier. Above their heads a half-figure of Christ reaches down from an arc of
light and gestures toward the Virgin, as if in response to her supplication.21
A second inscription, partially destroyed, runs along the base of the panel; it
reads "... discouraged by men, made alive by your strength, in thankfulness
I dedicate [this]."22
SOUTHEAST SANCTUARY PIER, NORTH FACE (Fig. 12): Across the nave,
the broad panel facing the Virgin and S. Theodore shows a nimbed S. Deme-
trios, in a tablion chlamys, with his arms around the shoulders of a bearded
bishop and a secular official dressed in gold.23 The figures stand on the un-
dulating light brown ground topped with a few rows of yellow-green tesserae

18 Chatzidakis, Byzantine Mosaics, PL 18; Hoddinott, Early Byzantine Churches, 154-

55, PI. 34; Xyngopoulos, Hagios Demetrios, Pis. 21-23.
19 AEHSI£. K[upi]E O 8[e6]S, EISAKOYEON THE $ONHE THE AAIHEEOS MOT,
OTI THE? TOT KOSMOT AEOME; English trans, from Hoddinott, Early Byzantine.
Churches, 155.
20See Henry Maguire, The Icons of their Bodies: Saints and Their Images in Byzantium
(Princeton, 1996), 88-89. As Cormack, "Mosaic Decoration," 29, has already noted, an
undated invocation to the Virgin, Theodore and Demetrios was found on the church walls:
Soteriou, Hagios Df.metrios, Fig. 93.
21This motif recurred in the lost mosaics of the north arcade: see below.
2 2 . . . E N AN0PnnOI£ AIIEAniEeEIE IIAPA AE THS EHE ATNAMEQS
ZnOHOIHGEIE ETXAPIETQN ANE6EMHN. The inscription is visible in Maguire, The.
Icons of Their Bodies, Fig. 37, and is easily legible in George's drawing of the panel: Cor-
mack, "Church of Saint Demetrius," no. 42. Pace Hoddinott, Early Byzantine Churches,
154, no name appears.
23Chatzidakis, Byzantine Mosaics, PI. 15; Hoddinott, Early Byzantine Churches, 150-
52, PI. 32a; Xyngopoulos, Hagios Demetrios, Pis. 11-14; Cormack, Writing in Gold,
51-53, Fig. 14.
 Leslie Brubaker Elites and Patronage in Early Byzantium 69

OUTH SIDE (Fig. 11): On the adjacent, familiar from the northeast pier, and here too they stand against a wall. The
)ier, the Virgin and S. Theodore stand wall depicted here is, however, white and distinguished by its crenellations;
ra ground topped with rows of yellow- these, draped with cloth, rise behind the heads of the two men accompa-
}f the same "wall", now blue but with nying Demetrios and thus do double-duty as square haloes, presumably to
Iternating ovals and vertical lines that indicate that the portraits represented living men. A metrical inscription
identify as a simple fieur-de-lys motif.18 along the base of the panel—that is, in the same position as the votive
have been altered: the "shadow" of her inscription that accompanies the panel opposite—explains: "The founders
i. She holds an unfurled scroll on which of this glorious house you behold at both sides of the martyr Demetrios,
d, hearken to the voice of my prayer, he who diverted a barbarous flood of barbarian ships and ransomed the
t Theodore, identified by his beard and city."24
blion chlamys decorated with a pattern SOUTHEAST SANCTUARY PIER, WEST FACE (Fig. 13): Adjacent to the
,mys on the east face of the southeast donor portrait, facing the nave as a pendant to the unidentified saint on the
of Christ reaches down from an arc of west face of the northeast sanctuary pier, is a portrait of S. Sergios, identified
, as if in response to her supplication.21 by inscription.25 The undulating ground, backdrop "wall" and top curtain
yed, runs along the base of the panel; it are virtually identical to the pendant panel on the northeast pier, while the
alive by your strength, in thankfulness fabric pattern on Sergios' chlamys resembles that worn by Demetrios in the
tribelon panel.
«)RTH FACE (Fig. 12): Across the nave, SOUTHEAST SANCTUARY PIER, EAST FACE (Fig. 14): Not visible from
i S. Theodore shows a nimbed S. Deme- the public areas of the church, the panel facing the apse shows S. Demetrios
arms around the shoulders of a bearded with his arm around the shoulder of the white-bearded deacon familiar from
in gold.2'5 The figures stand on the un- the tribelon mosaic. The background details are identical to those on the
with a few rows of yellow-green tesserae adjacent north face.26 Here too an inscription runs beneath the panel; it
reads: "blessed martyr of Christ, friend of the city, protect both citizens and
8; Hoddinott, Early Byzantine Churches, 154- outsiders".27 Robin Cormack has suggested that this invocation responds to
s, Pis. 21-23.
TSON THS *ONHS THS AAIHSEQS MOT,
nglish trans, from Hoddinott, Early Byzantine 24 KTisTAS OEHHEIS TON HANENAOSON AOMON EKEIOEN ENOEN MAP-
TTPOS AHMHTPION TOT BAPBAPON KATAfiNA BAPBAPfiN STOAH[v] META-
Bodies: Saints and Their Images m Byzantium TPEHONTOS K[al] HOAIN ATTPOTMENON: transcription and translation: Nancy
Mosaic Decoration," 29, has already noted, an Patterson Sevcenko (with Jeffrey Featherstone) in Kurt Weitzmann, cd., Aye of Spiritual-
e and Dcmctrios was found on the church walls: ity: Late Antique and Early Christian Art, Third to Seventh Century (New York, 1979),
554-55. See also Jean-Marie Spieser, "Les inscriptions de Thessalonique," Travaux et
s of the north arcade: see below. memoires 5 (1973), no. 7, 155-56, PI. 1,4; Denis Feissel and Jean-Marie Spicscr, "Invcn-
SIS IIAPA AE THS SHS ATNAMEHS taires en vue d'un recueil des inscriptions historiques de Byzance II: Les inscriptions de
MHN. The inscription is visible in Maguire, The Thessalonique. Supplement," Travaux et memoires 7 (1979), 334.
y legible in George's drawing of the panel: Cor- 25 Hoddinott, Early Byzantine Churches, 153, PI. 33b; Xyngopoulos, Hagios Demetrios,

:2. Pace Hoddinott, Early Byzantine Churches, Pis. 19-20.

26 Chatzidakis, Byzantine Mosaics, PI. 14; Hoddinott, Early Byzantine Churches, 152-

L5; Hoddinott, Early Byzantine Churches, 150- 53, PI. 32b; Xyngopoulos, Hagios Demetrios, Pis. 24-26.
trios, Pis. 11-14; Cormack, Writing in Gold, 27 nANOABIE TOT XPISTOT MAPTTS, $IAOHOAIS, SPONTIAA TI0H K[ai]
ITOAITON K[al] HENQN.
 70 Leslie Brubaker

the deacon's role as liaison between ecclesiastical and civic authority, and to
his pastoral duties.28
Amongst the extant mosaics, the pier and tribclon panels are closely
related technically and stylistically. Excepting the portrait of Demetrios with
the donors, the borders are identical in all compositions; fabric patterns are
shared (Demetrios' chlamys in the tribelon panel compares with that worn by
Sergios, that worn by Demetrios on the east face of the southeast pier with
that worn by Theodore, and that worn by the unidentified saint on the east
face of the northeast pier with that worn by the saint sometimes identified as
Bakchos), and so are background motifs: the undulating light brown ground
topped by rows of yellow-green tesserae appears in all pier mosaics, as does
the backing wall. Though both the tribelon mosaic and the panel portraying
the Virgin, Christ and S. Theodore are sometimes dated later than the others
in this group,29 these points of comparison (to which numerous others could
be added30) suggest instead that the pier panels and, probably, the tribelon
mosaic as well were designed to complement one another and belong to the
same campaign of decoration. The two panels on the west wall do not belong
with this group, and indeed show different kinds of scenes: one, an enigmatic
episode from the life of S. Demetrios, the other a votive image not from an
ecclesiastical or civic official but from what appears to be a family. This latter
theme recurred in the now-lost mosaics from the north arcade, to which we
shall now turn.
INNER NORTH AISLE, NORTH ARCADE: The mosaics covered the arcade
above the columns that separated the inner and outer aisles on the north
side of the church; applied to the south face of that arcade, they were visible
from the nave and covered the westernmost eight spandrels. The remaining
three spandrels, toward the apse, were covered with late medieval painting
over at least two earlier layers of fresco; when this section of the church

28 Cormack, Writing in Gold, 91.

29E.g. Hoddinott, Early Byzantine Churches, 155, thinks that the tribelon mosaic should
be dated well after the panel of Demetrios and the two donors (which he thinks it copies),
though he disagrees with those who have proposed a late dating for the Virgin and
Theodore panel (ibid., 154).
'i0E.g. the use of contrasting colours to form ray-like edges (compare the area below
the raised forearm of the saint on the west face of the northeast pier, with that below
Theodore's left arm and along his right forearm), the doubled and nearly horizontal drap-
ery folds at the neckline of Theodore, Demetrios with the donors, and Demetrios with the
deacon; and the way tesserae are laid to form haloes.

L
 Leslie Brubaker Elites and Patronage in Early Byzantium 71

2cclcsiastical and civic authority, and to was destroyed in 1917 these had not yet been uncovered.31 Before their
destruction, the arcade mosaics were recorded photographically arid, more
le pier and tribclon panels are closely clearly, in watercolours and drawings.32 The watercolours were published
Excepting the portrait of Demetrios with by Cormack in 1969, who for ease of reference identified the spandrels with
in all compositions; fabric patterns are letters (from A to K, running from west to east and ignoring the half-spandrel
.belon panel compares with that worn by where the arcade abuts the west wall, where nothing remained).33 I will follow
the east face of the southeast pier with that convention here.
>rn by the unidentified saint on the cast SPANDREL A (Fig. 2): Only partially preserved, the orant S. Demetrios
rorn by the saint sometimes identified as stands before a niche, with the head of a diminutive figure to the left. To the
)tifs: the undulating light brown ground right, an inscription reads: "As a prayer for one whose name God knows",34
;rae appears in all pier mosaics, as does the formula also found on the east face of the northeast pier (and further-
ribclon mosaic and the panel portraying east along the north arcade). The apex of the arch between Spandrels A and
re sometimes dated later than the others B did not survive, but the two spandrel compositions seem to be enclosed in
larison (to which numerous others could the same frame (green rectangles against a gold ground, with dark borders
3 pier panels and, probably, the tribclon dotted with white) and may thus have been conceived as a single unit.35
iplement one another and belong to the SPANDREL B (Fig. 2): The orant S. Demetrios again stands before a
ivo panels on the west wall do not belong niche, this time with a more extensive architectural setting that originally
ferent kinds of scenes: one, an enigmatic incorporated two medallion portraits, one of which remained and was duly
s, the other a votive image not from an recorded by George. On the right, a small bearded figure moves toward
L what appears to be a family. This latter Demetrios with draped and extended hands.30 No inscription was recorded.
aics from the north arcade, to which we SPANDREL C (Fig. 3): The composition extends from the apex of the arch
between Spandrels B and C to the apex of the arch between Spandrels C and
The mosaics covered the arcade D, and is clearly delineated by a red frame on which simulated gemstones
the inner and outer aisles on the north alternate with a pattern formed of four white dots. In the centre, the Virgin,
ith face of that arcade, they were visible seated on a jewelled lyre-back throne, holds the Christehild in her lap and
ernmost eight spandrels. The remaining is flanked by two angels. To the left, a small saint (presumably Demetrios)
rere covered with late medieval painting presents an even smaller figure, who moves toward the Virgin with covered
fresco; when this section of the church
31See Cormack, "Mosaic Decoration," 41-42.
32 Photographs by N.K. Kluge were published by T. Uspensky in hvestija Russkogo
Archeologicheskogo Instituta v Konstantinopolje 14 (1909), and those by M. le Tourneau
ches, 155, thinks that the tribclon mosaic should by C. Diehl in Monuments et memoires (Fondation Eugene Piot) 18 (1910); they have
and the two donors (which he thinks it copies), since been republished often, most accessibly in Hoddinott, Early Byzantine Churches.
re proposed a late dating for the Virgin and For the drawings and watcrcolours, see n. 8 above.
•"Cormack, "Mosaic Decoration," 23 (with diagram).
) form ray-like edges (compare the area below 34TIIEP ETXHS OT OIAEN 0 0[e6]E TO ONOMA; Cormack, "Mosaic Decoration,"
jst face of the northeast pier, with that below 24-25; Cormack, "Church of Saint Demetrius," no. 30 (watercolour).
rearm), the doubled and nearly horizontal drap- 35So Cormack, "Mosaic Decoration," 25.
letrios with the donors, and Dcinetrios with the 36 Hoddinott, Early Byzantine Churches, 145, Fig. 31a (photograph); Cormack, "Mosaic
orrri haloes. Decoration," 25-26; Cormack, "Church of Saint Demetrius," no. 30 (watercolour).
72 Leslie Brubaker

hands; to the right stands an orant figure, bearded and nimbed, who may
represent S. Theodore.37 Over the apex of the left arch are two medallion
portraits, one of an unnamed female and one of an unnamed male saint; over
the apex of the right arch are three more medallions, with an unidentified
male saint joined by S. Pclagia and the local Thessalonikan S. Matrona. A
small kneeling figure, apparently female, occupies the narrow space between
Matrona and the edge of the arch. No votive inscription was recorded.38
SPANDREL D (Fig. 3): The composition extends from Spandrel D to just
before Spandrel H, and all is enclosed in a frame similar to that of Spandrel
A, but with alternating red and green rather than gold and green rectangles.
Spandrel D itself shows Demetrios seated in front of his kiborion, gesturing
toward a defaced medallion from which a hand emerges in response; the
medallion figure holds a book and has therefore been identified as Christ,
who appears in similar form (but bookless) in the pier panel of Theodore
and the Virgin. A woman holds a small child in covered hands and presents
her to Demetrios, who touches the child—marked with a gold cross on its
forehead—with his right hand. An uii-narned nimbed woman- -sometimes,
but without compelling evidence, identified as the Virgin—stands to the
right of the medallion portrait and gestures toward the woman and child
with one hand while raising the other. To her right is an architrave enclosing
what appears to be a sarcophagus.39 Further right still, at the apex of the
fifth arch, is another medallion portrait of an unidentified nimbed male that
may represent a later insertion into the composition.40 No inscription was
recorded.
SPANDREL E (Fig. 4): The Virgin, standing this time, is again flanked by
two angels; here, however, the angels rest their outer hands on the shoulders
of two figures, one a mother holding a small child with a gold cross on its
forehead in her draped arms and the other apparently also a female with
draped arms but without a child, whom they seem to be presenting to the
Virgin. The Virgin gestures toward the right and upward, and probably

37On Theodore, see above and n. 20.

38 Hoddinott, Early Byzantine Churches, 145, Figs. 29a, 30c (photographs); Cormack,
"Mosaic Decoration," 26-31; Cormack, "Church of Saint Demetrius," no. 32 (watercolour).
39 Hoddinott, Early Byzantine Churches, 145, Figs. 29d, 30c (photographs); Cormack,
"Mosaic Decoration," 31-33; Cormack, "Church of Saint Demetrius," nos. 32, 34 (water-
colours).
40So Hoddinott, Early Byzantine Churches, 145, and, with caution, George, in a pencil
notation visible on the watercolour: see Cormack, "Mosaic Decoration," 33.
 Leslie Brubaker Elites and Patronage iu Early Byzantium 75

he centre of the composition, Dernetrios on children, as, indeed, was the focus of many saintly miracles recorded in
ht, in a four-columned structure the top the early Byzantine period.48
.ead, had disappeared before 1907. To the There is, in short, no compelling reason to assume that the gold cross
•up of three women, a beardless man, and was the prerogative of a single individual, and in fact two early seventh-
n her forehead incline themselves toward century texts suggest a different interpretation. The first is a collection of
the saints, and the couple behind her may miracle stories compiled in honour of S. Demetrios, probably by John, bishop
icr parents. Behind the figures is a shrine of Thessalonike, around 610. Here, in the third miracle, the saint appears
' the seventh arch, a medallion portrait of during the night to those who have fled to the church for protection against
ios. The right side of the composition had the plague, and marks with a cross those who will survive.49 A context
e may have been placed next to Demetrios, for this detail is supplied by the History of Theophylact Simocatta, written
i above the eighth arch, and a fountain in probably in the late 620s. Theophylact describes the emperor Maurice's
t to the concluding frame. The inscription close encounter with a wild boar; the boar did not attack, and "after the
\nd you, my Lord Saint Demetrios, aid us disappearance of the beast, the emperor traced on his forehead the sign of
aria, whom you gave to us."4' the cross, as it is customary for Christians to do at miracles" .50 Although,
is entire composition concerns one family, as we shall see, the mosaic sequence of the north arcade apparently predates
in the last spandrel, who, it is argued, ap- both of these texts, their combined evidence nonetheless suggests that the
nt, once as a small girl, and once as a young sign of the cross on one's forehead was not a personalising attribute but a
1 cross on her forehead.47 This assumption sign of miraculous favour. Rather than identifying a single girl, the crosses on
ge of the four relevant compositions within the children's foreheads mark them as recipients of miracles performed by S.
Ision to assign names to anonymous figures, Demetrios. The sequence may be seen as a record of successful supplications
i cross marks an individual rather than the on behalf of four children: perhaps all had recently been healed by Demetrios
tus that might be shared by several people, through the sign of the cross,51 or—in the case of Maria had been conceived
on, but it also raises considerable difficul- after his intervention, or—as in a dream-vision recorded in the early seventh
'gical narrative development that must be century- -had been marked by Demetrios with a cross as a sign that they
A of the four children portrayed as Maria would survive an illness or other catastrophe.52
uence. Nor is there any particular reason to SPANDREL H (Fig. 5): In the final mosaic composition recorded along
.ted to a single major protagonist indeed, I the north arcade, framed in red with simulated gemstones and pearls in a
f this long sequence would seem to contra- pattern that generally recalls the border of Spandrel A, the orant Demetrios
all four compositions in this sequence focus stands in front of a niche with a scalloped conch, closed in the back by a
lat we must assume that the same child is
48See below, 83.
)f other panels at Hagios Demetrios is also 49 On the collection, Paul Lemerle, Les plus anciens recueils des miracles de Saint
Demetrius (Paris, 1979-81), csp. II, 27-34; for the miracle itself, ibid., 75-82.
E AIMHT[p]I [p]OH0I HMIN TOIE AOTAOIS 50 Mary and Michael Whitby, The History of Theophylact Simocatta, an English Trans-
nglish trans, from Cormack, "Mosaic Decoration," lation with Introduction and Notes (Oxford, 1986), 157.
Churches, 146, Figs. 30b, 31b (photographs); Cor-1 51 On healing through the sign of the cross, sec further below, 83. For an alternative

irmack, "Church of Saint Demetrius," no. 36 (wa- suggestion that the children may be commended to Dcmetrios and the Virgin because
they have been recently baptised, see Andre Grabar, L 'iconoclasme byzantin. Dossier
ne Churches, 145-46; Cormack, "Mosaic Decora- archeologique (Paris, 1957), 86-88.
',d, 88-89. r'2 Reference in n. 78 below.
76 Leslie Brubaker

curtain. On either side of the conch are small medallion portraits, both of
youthful males; the fragmentary identifying inscriptions are uncertain, but
one may have read Alexandros. To Demetrios' right, two small figures raise
covered hands toward him; to his left, a more richly dressed beardless male
follows suit. The inscription at the far right of the composition reads: "As
a prayer for one whose name God knows", which is the same formula found
in Spandrel A and on the east face of the northeast pier.53
Whether or not the arcade mosaics were set in a single campaign or
individually was not determined by the early twentieth-century observers,
and without the archaeological evidence provided by suture lines it is unlikely
ever to be known. The twisted ribbon pattern above the figural panels, which
seems to unite them, is, however, entirely above the horizontal suture line
recorded by George; it was therefore not part of the original mosaic work but
part of the later restorative campaign that also inserted the three medallion
portraits over the apex of the sixth arch. The ribbon encloses simple fleur-
de-lys- like motifs, those we have noted already on the pier panel portraying
S. Theodore and the Virgin. This has obvious implications for the relative
dating of the mosaics.

Date of the Mosaics and the Significance of Their Location in the

Church
The dating of the mosaics relies in part on the chronology of the building.
The church, built upon the foundations of a Roman bath complex that had
apparently been modified to accommodate a cult site before the basilica was
itself constructed,54 is attributed to the second half of the fifth, or possibly
even the early sixth, century.55 While the mosaics in the church obviously
5:iTII[ep e]TX% o]T OIA[ev] 0 0[£o]£ [10] ONO[na]: Hoddinott, Early Byzantine
Churches, 146, Fig. 31b (photograph); Cormack, "Mosaic Decoration," 37-39; Cormack,
"Church of Saint Demetrius," no. 36 (watercolour).
54So, most recently, Aristotle Mentzos, To proskynema tou Hayiou D&netriou Thessa-
lonikes sta Byzantina chronia (Athens, 1994), 42-45.
55The dating is primarily based on brickstarnps—see esp. Michael Vickers, "Fifth-
Century Brickstamps from Thessaloniki," Papers of the British School at Athens 68 (1973),
285-94- -and capital types. A date of ca. 460-470 is proposed by the Soterious (as in n.
7 above); ca. 450-75 by W. Eugene Kleinbauer, "Some Observations on the Dating of
S. Demctrios in Thessaloniki," Byzantion 40 (1970), 36-44; the third quarter of the fifth
century by Richard Krautheimer, Early Christian and Byzantine Architecture, 4th cd.,
rev. (London, 1986), 125-28; the second half of the century, and possibly as late as ca. 500
by Cormack, "Church of Saint Demetrius," 55-57. The late date of 510-20 is argued by
 Leslie Brubaker Elites and Patronage in Early Byzantium 77

conch are small medallion portraits, both of postdate the building's construction, there is no way of knowing whether or
y identifying inscriptions are uncertain, but, not this basilica was decorated with mosaics from the start.56 It is clear,
To Demetrios' right, two small figures raise however, that the mosaics known to us were never the defining images of the
iis left, a more richly dressed beardless male cult site. The decoration of the apse is unknown, but we may be certain that
; the far right of the composition reads: "As none of the mosaic work in the church ever appeared in the main ceremonial
od knows", which is the same formula found space of the nave, which is completely sheathed in finely worked marble inlay
face of the northeast pier.53 contemporary with the original architecture.57 If the early miracle stories
5 mosaics were set in a single campaign or connected with the building are to be trusted, the main focus of Demetrios'
3d by the early twentieth-century observers, cult was not in fact representational imagery at all, but a large hexagonal
evidence provided by suture lines it is unlikely silver kiborion located in the nave, where Demetrios' presence was somehow
•ibbon pattern above the figural panels, which tangible.58 This was not centrally located: it nearly touched the north arcade
rei, entirely above the horizontal suture line and sat slightly closer to the western entrance end than to the eastern apse
efore not part of the original mosaic work but (Fig. I).59 Whether or not the mosaics were part of the original plan of the
npaign that also inserted the three medallion building, their placement appears to have been affected by the location of
sixth arch. The ribbon encloses simple fleur- the kiborion, which was aligned with the fifth arch, between Spandrels D and
/e noted already on the pier panel portraying E. This is where the longest sequence of mosaics on the inner north arcade
Phis has obvious implications for the relative started: beginning with the image of S. Demetrios sitting in his kiborion
receiving the small child with a gold cross upon its forehead, the children with
gold crosses on their foreheads were sited directly in line with the kiborion
le Significance of Their Location in the itself (Figs. 3-5). This lateral alignment of cult site and particularised images
is unlikely to be coincidental, and presumably was meant to visualise the
es in part on the chronology of the building,
Jean-Marie Spieser, Thcssalonique et ses monuments du 4e au 6e siecle. Contribution a
indations of a Roman bath complex that had I'etude d'une villa paleochretienne (Paris, 1984), 165-214, csp. 212-13. For a good survey
:commodate a cult site before the basilica was of the literature, see Skedros, Saint Demetrios, 29-39, who argues in favour of a date in
ed to the second half of the fifth, or possibly the last quarter of the fifth century.
'' While the mosaics in the church obviously 56As argued by, e.g., Kleinbauer, "Some Observations".
57 See Paul Lemerle, "Notes sur les plus anciennes representations de saint-Demetrius,"

1 ©[EO]E [TO] ONOfoaa]: Hoddinott, Early Byzantine Deltion, Ser. 4, 10 (1980-81), 9; and also, on the opus sectile itself, Spiescr, Thessalonique,
h); Cormack, "Mosaic Decoration," 37-39; Cormack, 174-82.
3 (watercolour). 58This was not a martyhon in the strict sense of the term, for there was no tomb. The

tltzos, To proskynema ton Hagiou Deme.triou Thessa- church apparently had no relics of the saint either, other than a piece of blood-drenched
sns, 1994), 42-45. cloth, kept in a small crypt under the altar that was originally accessible only to the
on brickstamps—see esp. Michael Vickers, "Fifth- priest and that was anyway sealed over early in the history of the church; this relic is
niki," Papers of the British School at Athens 68 (1973), never mentioned in early documents. Sec Grabar, Martyrium, I, 450-57; idem, "Quelqucs
of ca. 460-470 is proposed by the Soterious (as in n. reliquaires dc Saint Demetrios et le martyrion du saint a Salonique," Dumbarton Oaks
ic Kleinbaucr, "Some Observations on the Dating of Papers 5 (1950), 8-11; Paul Lemerle, "Saint Demetrius de Thessalonique et les problemcs
mtion 40 (1970), 36-44; the third quarter of the fifth du martyrium et du transept," Bulletin de correspondance hellenique 77 (1953), 660-94,
Early Christian and Byzantine. Architecture, 4th ed., esp. 661-73; Demetrios Pallas, "Le ciborium hexagonal dc St-Demetrios de Thessalonique,"
:ond half of the century, and possibly as late as ca. 500 Zograf 10 (1979), 44-58.
letrius," 55-57. The late date of 510-20 is argued by 59Soteriou, Hagios Demetrios, 10-20, 100-101, 179 82.
78 Leslie Brubaker

association between S. Demetrios, with his locus in the kiborion, and the
miracles recorded by the crosses.
The secondary importance of the mosaics in the hierarchy of Hagios
Demetrios' decoration may suggest that they were not a significant part of
the original plan, but such speculation brings us no nearer to a date for the
panels. For this, we are reliant upon style and archaeology, and both aspects
are problematic in the case of a building now largely destroyed. It is clear
from the inscription over the sixth arch of the north arcade that a fire dam-
aged the church at some point, and the George drawings demonstrate that
most of the arcade mosaics actually preceded that fire: the post-fire repairs
there are, as already noted, limited to the inserted portrait medallions and
the work above the horizontal suture line that George carefully indicated on
all of the relevant drawings. Connections between the inserted portraits and
the decorative motifs of the post-fire work in the arcade and on the sanctu-
ary piers strongly suggest that the piers belong to this post-fire campaign as
well. In other words, whether or not the pre-fire sequence along the north
arcade all dated from the same period, it seems that the post-fire work hangs
together in a fairly coherent programme.
Much hinges on the date of the fire, and for this we are reliant on accounts
of the miracles that were associated with S. Demetrios. Paul Lernerle has
demonstrated that the oldest version of the Miracula is associated with John,
bishop of Thessalonikc soon after Eusebios (mentioned in letters of Pope
Gregory between 597 and 603); John's version of the Miracula is therefore
usually dated to ca. 610.60 Miracles six and twelve of this collection mention
a fire during the time of bishop Eusebios, but it seems only to have affected
the kiborion of the saint.61 The great fire that left the church without a
roof is described only in a subsequent miracle collection, usually dated to
the later seventh century; here the fire is said to have occurred soon after
the death of bishop John, and Lemerle concluded that a date of ca. 620 was
most likely.62 Accepting these dates as at least general guidelines, the pre-fire
mosaics could have been installed at any time between ca. 500 and ca. 620,
while the post-fire mosaics might be dated sometime in the second quarter
of the seventh century.

GO Lemerle. Miracles, II, 27-34; for John's dates see also Louis Petit, "Les eveques de
Thcssalonique," Echos d'Orient 4 (1900-1901), 212. A good survey of the literature on
the date of the Miracula appears in Skedros, Saint Demetrios, 107-15.
61 Lernerle, Miracles, I, 90-95, 120-29.
6 2 Ibid., I, 190-97; II, 107-10.
 Leslie Brubakn- Elites and Patronage in Early Byzantium

•s, with his locus in the kiborion, and the The only surviving mosaics that might belong to the first phase of dec-
oration are those of the west wall, showing Demetrios and two angels and
f the mosaics in the hierarchy of Hagios Demetrios with supplicants (Figs. 7 8 ) . These have been dated to the late
sst that they were not a significant part of sixth or early seventh century by Ernst Kitzinger,63 but arguments for a date
latiori brings us no nearer to a date for the as early as ca. 500 have also been advanced.64 The formal parallels with the
»on style and archaeology, and both aspects votive mosaic at Dyrrachium in Albania (ca. 525) and with the mosaics at the
building now largely destroyed. It is clear church of the Panagia Kanakaria on Cyprus (ca. 530) suggested by Cormack
h arch of the north arcade that a fire dam- are general,65 but do argue for a date in the first rather than the second half
nid the George drawings demonstrate that of the sixth century.60 We are likely, in other words, to have two groups of
Jly preceded that fire: the post-fire repairs votive mosaics separated in date by close to a century, and the length of this
ed to the inserted portrait medallions and timespan suggests that different types of evidence about sanctity and about
,ure line that George carefully indicated on elite patronage may be expected from each group.
lections between the inserted portraits and
fire work in the arcade and on the sanctu- The Evidence of the Miracula
e piers belong to this post-fire campaign as The first collection of miracles—the Miracula associated by Lemerle with
not the pre-firc sequence along the north bishop John, and dated to ca. 610—was thus probably compiled after the first
riod, it seems that the post-fire work hangs sequence of mosaics had been installed, but before the second. Significantly,
ramme. however, it makes no mention whatsoever of any mosaic decoration inside the
i fire, and for this we are reliant on accounts church, and indeed scarcely refers to representational imagery of any sort.67
ted with S. Demetrios. Paul Lemerle has Instead, as noted earlier, the Miracula make it clear that the focus of the S.
.on of the Miracula is associated with John, Demetrios cult was the silver kiborion. This apparently included a portrait
r Euscbios (mentioned in letters of Pope of the saint the first miracle in John's collection mentions "the divine effigy
tohn's version of the Miracula is therefore of the holy victorious [one]" in the kiborion68—but that portrait is never
es six and twelve of this collection mention itself credited with any activity. The author of the Miracula assumes that
usebios, but it seems only to have affected portraits of the saint exist, but their only role is to allow people to identify
great fire that left the church without a
quent miracle collection, usually dated to 63 Kitzingcr, "Byzantine Art between Justinian and Iconoclasm," 20-24.
64Sce Cormack, "Mosaic Decoration," 49-50; A.H.S. Mogaw and Ernst J.W. Hawkins,
le fire is said to have occurred soon after
The Church of the Panagia Kanakaria at Lythrankomi in Cyprus, its Mosaics and Frescoes
nerle concluded that a date of ca. 620 was
(Washington, DC, 1977), 62 n. 178.
3s as at least general guidelines, the pre-fire G5Rcferences in previous note.
at any time between ca. 500 and ca. 620, ™Guglielmo Matthiae, "La cultura figurativa di Salonicco nei secoli V e VI," Rivista di
be dated sometime in the second quarter archeologia cristiana 38 (1962), 180 84, argued that the panel showing Demctrios and the
angels should be dated to ca. 500, that showing Demetrios with golden hands and patrons
to the mid-sixth century; however, the two panels show sufficient formal similarities to
)hn's dates see also Louis Petit, "Les evequcs do argue against a half-century difference in date.
( ' 7 Nor do the passions associated with Demetrios mention portraits of the saint: Lemerle,
D-1901), 212. A good survey of the literature on
dros, Samt Demetrios, 107-15. "Representations," 1-2.
° 8 ... TO OeoeiSe? Kpooconov TOU ocOiou nava^TCTOU d6Xotp6pou: Lemcrle, Miracles I, 66:27
28 (I, 22).
80 Leslie Bmbaker Elites .

a figure seen in a vision as S. Dcmetrios himself: in the eighth miracle,

Demetrios appears to the boat captain Stephanos "dressed [or posed] as one
sees him in images";69 in the tenth, the saint appears to a well-born man,
who sees Demetrios seated in his kiborion "in the dress [or pose] in which
one sees him in images";70 and in miracle fifteen, another "man of good
birth" sees Demetrios open the door of his kiborion "looking like he did in
the old images".71 In the second, late seventh-century collection of miracles,
a variant on the same formula appears: the African bishop Cyprian is able
to identify the man who rescued him from captivity as Demetrios when he
sees his portrait. 72 These are the only specific references to portraits of the
saint in the Miracula.
There are, however, also two references to images associated with mir-
acles. One appears in the later edition of the miracles, which claims that
Demetrios' salvation of the city is represented at his sanctuary "in wood" (i\. While

civic gratitude, my main interest here is in a mosaic described as on the

outside wall of Hagios Demetrios itself. This appears in the first miracle of
John's collection, which concerns the eparch Marianos. Marianos was struck
with paralysis by the devil; after refusing the help of a magic amulet, he
dreamt that his friend Demetrios an important man at court—told him to
come to his house to be healed. One of the eparch's servants, inspired by
God, realised that the church of Hagios Demetrios was meant. Mariaiios
was carried there, and in a dream-vision saw S. Demetrios, who said to him
"Christ our Lord returns your strength, he who heals those in need". On
awakening, Marianos repeated this phrase and was instantly cured. He gave
thanks to the saint, and many gifts to the church. If anyone doubts this

°9. . . i\ xai 0xr)|_ioru £v TCXU; etxooi: Lemcrle, Miracles I, 102:9 (VIII, 70).
70. . . OUTCOC 6 a^rporti xcrra ia<; elxovoe<; iyYpdtpETai.: Lemerlc, Miracles, I, 115:16-17
(X, 89).
7 ' . . . t f ) I5ea xaia ity YP a 9^ v T^lv ^ v tort? cxp)(ai.OT£pai!; autou Eixoaiv £'YY E: YP a M^ v ' r ) v:
Lemerle, Miracles, I, 162:17 (XV, 167). All of these examples are rioted and discussed by
Lernerle, "Representations," 2-3.
72Lemerle, Miracles, I, 239:5-7 (II. VI, 311).
73Lemerle interprets this to mean a second sanctuary dedicated to Demctrios, in the
forest: Miracles, I, 179:18-19; see also 174 n. 19. It could equally indicate a wooden
image, possibly three-dimensional.
 Leslie Brubaker Elites and Patronage in Early Byzantium 81

Demetrios himself: in the eighth miracle, story, they should examine the mosaic on the outside of the church, on the
iptain Stephanos "dressed [or posed] as one wall facing the stadium (?), and they will be convinced.74
nth, the saint appears to a well-born man, The Miracula description raises three issues. The first is the correlation
s kiborion "in the dress [or pose] in which of the text account and the visual record: the Miracula mention a mosaic on
in miracle fifteen, another "man of good the outer wall of the building, and such an image— though not the one noted
door of his kiborion "looking like he did in in the Miracula- -still exists. To judge by the fragmentary inscription, the
late seventh-century collection of miracles, existing mosaic is an ex voto, a visual prayer in the name of a man and his
ppears: the African bishop Cyprian is able wife.75 Presumably, the Marianos mosaic was also votive, though whether
him from captivity as Demetrios when he or not it recorded the narrative described in the Miracula or simply showed
3 only specific references to portraits of the Marianos giving thanks to Demctrios is not clear; the latter is perhaps more
likely, given the other evidence preserved at Hagios Demetrios. But what is
references to images associated with mir- more important is that written and visual evidence together indicate that
edition of the miracles, which claims that the exterior as well as the interior of the church acted as a frame for votive
; represented at his sanctuary "in wood" (ev imagery. This raises a second issue, the question of why the Marianos mosaic
as an example of a votive image recording was installed on the exterior wall of the church, when we know there were
it here is in a mosaic described as on the already votive mosaics inside. Was the interior wall space along the aisles
; itself. This appears in the first miracle of the only space where mosaics and frescoes were possible76 -already filled? Or
the eparch Marianos. Marianos was struck was an exterior site considered more important, so that the mosaic's location
•i refusing the help of a magic amulet, he there indicated Marianos' high status? This latter is plausible since, as we
--an important man at court—told him to have already seen, there was no room for votive panels in the main ceremonial
One of the eparch's servants, inspired by space of the nave, and the early votive panels inside the church were relegated
F Hagios Demetrios was meant. Marianos to the aisles. Whether or not there were mosaics on the sanctuary piers — or
i-vision saw S. Demetrios, who said to him anywhere; else in the eastern end of the church- before the fire of ca. 620 is
:rength, he who heals those in need". On not known. But there was never much wall space for decoration in this high-
is phrase and was instantly cured. He gave status area, and- -though further from the kiborion where Marianos' healing
rifts to the church. If anyone doubts this took place— the exterior of the church was certainly a more visible and more
prominent location than the side aisles. It is at least possible that Marianos,
whose gifts to the church were evidently extensive, was allotted space on an
exterior wall in preference to an interior wall for reasons of status.

!i<;elx6ai: Lemerlc, Miracles I, 102:9 (VIII, 70). 74 Lemcrle, Miracles, I, 57-67. For the conclusion — Et TI? 4>eu8r) ^ Xeyeiv 6noTOK«Coi,
EYYpatpEToa: Lemerlc, Miracles, I, 115:16-17 xr]v EX ^ouaEiou auvie6£ijaevr)v EXEWE ypacpr)v w TOU vaoO itpo? TOV dcpopuvia
TO TTJ? noXewt; 0Ta5iov, xai nXr)potpopT]8ei(; ToTc; KpoeiKr)(jevoic; -ibid.,
v TCUC; otpxaioiEpaii; autoO etx6oiv 67:14-17.
All of these examples are noted and discussed by 75See above, 64, and references in n. 9.
70 Fragments of wall paintings were discovered in the south aisle, at the end of the north

1,311). arcade, and along the north wall by Soteriou; they are, however, too badly damaged to
second sanctuary dedicated to Dcmetrios, in the date: see Soteriou, Hagios Demetrios, 103,204. Spieser, Thessalonique, 170-71, attempted
o 174 n. 19. It could equally indicate a wooden to date some of these to the post-iconoclast period on the basis of costume details, but
the argument is not compelling.
82 Leslie Brubaker

A final issue raised by the Miracula account of the healing of Marianos

concerns the use and role of images in the text. Bishop John cited the mo-
saic as a witness to his words: he had not made the story up, for he had
an independent testimony of Marianos' healing in the mosaic. The image
here confirmed the truth of John's words, just as the portraits of the saint
confirmed that the man who appeared to the ship captain Stephanos, the
well-born man, and later the African bishop Cyprian was Demetrios him-
self. It is important for our understanding of the development of the theory
of images in Byzantium to note that the images in the Miracula validate,
they do not act. Throughout the Miracula, in both the early and the later
collections, images are passive agents of commemoration and authorisation.
Never is an image the object of prayers or requests, and never is an image
presented as a conduit to the saint himself.

Votive Mosaics and Sanctity

The Miracula texts and the mosaics of Hagios Dcrnetrios alike demonstrate
that images of Demetrios existed, but it is evident that those images were
not locations of saintly power. The saint did not work through his image,
but through his physical appearance, usually in a dream or vision.
Nine of the fifteen miracles in John's collection describe dream-visions
of the saint, often on several occasions. In the first miracle, the paral-
ysed Marianos—who has rejected the imagery of an amulet—has two dream-
visions of the saint; the first directs him to the church, the second heals him
when there.77 As we have seen, a mosaic is then installed to commemorate
the healing, but the image itself plays no part in the event itself. In the third
miracle, discussed earlier, Demetrios appears in the church and marks with
a cross those who will survive the plague.78 In miracle six, Demetrios gives
instructions to the priest at his church three times, all in dreams.79 Miracle
seven recounts the story of Onesiphoros the sacristan, who substituted small
for the large candles left by worshippers before Demetrios' kiborion: the saint
appeared to him three times in dreams to ask him riot to do this, and finally,
when he continued, struck him almost dead.80 The ninth miracle concerns
the arrival of food at Thessalonike during a famine, an event clarified by

77Lcmerle, Miracles, I, 57-67.

78 Ibid., I, 75-82.
7albid., I, 93-95.
80 Ibid., I, 97 99.
 Leslie Brubaker Elites and Patronage in Early Byzantium

(iracula account of the healing of Marianos Demetrios' appearance to a pious man in a dream.81 In miracle ten, a man
,ges in the text. Bishop John cited the mo- has a dream-vision of Demetrios interceding with God for the salvation of
he had not made the story up, for he had Thessalonike.82 In the final trio of miracles included in the first collection,
arianos' healing in the mosaic. The image Demetrios appears at the walls of Thessalonike as a soldier helping to defend
n's words, just as the portraits of the saint the city; he is seen by the Thessalonikans and by the enemy, who flee at the
ipeared to the ship captain Stephanos, the sight.83 Demetrios is also seen in a dream refusing to leave the city to its
frican bishop Cyprian was Demetrios him- fate.84
irstanding of the development of the theory This same pattern continues in the collection attributed to the later sev-
that the images in the Miracula validate, enth century. Here Demetrios is seen at the battle line by friend and foe
e Miracula, in both the early and the later alike,85 directs the miraculous appearance of supplies during various sieges
rents of commemoration and authorisation, by diverting ships from their anticipated routes,86 and leads the African
prayers or requests, and never is an image bishop Cyprian to safety.87 He also continues to appear in dreams,88 but
nt himself. saintly presence is never associated with saintly imagery.
The Miracula texts make it clear that Demetrios was an efficacious inter-
cessor with God on behalf of those who came to his church and of Thessa-
lonike in general. But neither the portrait of the saint that we are told was in
aics of Hagios Demctrios alike demonstrate the silver kiborion at Hagios Demetrios, nor the sixth- and seventh-century
i, but it is evident that those images were mosaics there, participated in the construction of Demetrios' sanctity. As
The saint did not work through his image, with other miracle collections compiled before ca. 680, it is the physical
ince, usually in a dream or vision, presence of the saint rather an active image that effects cures, and it is the
in John's collection describe dream-visions "tomb" of the saint, his kiborion, that is honoured with votive candles, not
iccasions. In the first miracle, the paral- his icon.89 Why, then, were the mosaics at Hagios Demetrios made, and what
l the imagery of an amulet—has two dream- do they record?
cts him to the church, the second heals him
a mosaic is then installed to commemorate The Elite Patron
)lays no part in the event itself. In the third
Nothing in the Miracula texts nor in the panels themselves suggests that the
;rios appears in the church and marks with
mosaics at Hagios Demetrios were made as objects of devotion; there is no
le plague.78 In miracle six, Demetrios gives
indication that they are "icons" in the later Byzantine sense of the term;
:hurch three times, all in dreams.79 Miracle
phoros the sacristan, who substituted small 81 Ibid., I, 106-108.
.ippers before Demetrios' kiborion: the saint 92 Ibid., I, 112-16.
83 Ibid., I, 133-38, esp. 135 (XIII, 120); 146-58, csp. 157 (XIV, 161).
reams to ask him not to do this, and finally,
MIbid., I, 161 65.
ilmost dead.80 The ninth miracle concerns 85E.g. ibid., I, 177-78 (I, 188-89), 189 (II, 214), 195 (III, 220), 216 (IV, 260-61).
like during a famine, an event clarified by 86E.g. ibid., I, 188 (II, 209), 231-32 (V, 298-300).
S7Ibid., I, 237-41.
88E.g. ibid., I, 196-97 (III, 227), 231-32 (V, 298-300).
89On this point, see my "Icons before Iconoclasm?," Se.ttim.ane. di studio del Centra
italio.no di studi sull'alto medioevo 45 (1998), 1215-54, esp. 1231-37, on the miracles of S.
Artcmios.
84 Leslie Brubaker

no evidence suggests that they are meant to be understood as mediating

between the saint in heaven and his worshippers on earth. Instead, as the in-
scriptions that still accompany some of the panels attest, they are "prayers"
to the saint. Apart from the two post-fire donor or dedication inscriptions,
which commemorate both the restorers of the building and the saint,90 all
of the other preserved legends from Hagios Demetrios are prayers of thanks
for Demetrios' previous intercession or prayers in hope of future blessings.
As we have seen, two of the four inscriptions associated with the pre-fire
panels simply read "a prayer for one whose name God knows", one fragment
addressed the Virgin, and one invoked Demetrios.91 Those attached to the
later panels are more extensive, but repeat the same sentiments. The in-
vocation "a prayer for one whose name God knows" recurs, as do the more
civically orientated prayers "for the world" and for "citizens and outsiders";
one inscription specifically gives thanks.92
But while most of the inscriptions insist on the anonymity of the donor,
the images do not. Amongst the eighteen compositions described above,
only two certainly never included an image of the donor(s), the northeast
sanctuary pier panel of the Virgin with S. Theodore, and the southeast pier
panel of S. Sergios (Figs. 12-13).93 Though the damaged state of the mosaics
makes categorical distinctions impossible, we can at least conclude that most
of the mosaics depicting Demetrios included portraits of supplicant-donors.
The pre-fire mosaics along the north arcade, and that on the south side
of the west wall, portray a range of supplicants, many of whom are young
(Figs. 2-5, 8). This is not true of the post-fire mosaics on the sanctuary piers
and tribelon wall, where only one panel—that closest to the north arcade se-
quence, on the west face of the northeast pier (Fig. 9) - includes children,
and where the compositions record not supplications but rather Demetrios'
protection. I do not think that this can be simply an accident of survival.
The difference between the pre- and post-fire mosaics suggests a changed
role for mosaic decoration: the Hagios Demetrios panels appear to docu-
ment the transformation of ideas about votive imagery between roughly the

90See nn. 24 and 43 above.

91Sce nn. 34, 45, 46, and 53 above.
92Sec nn. 17, 19, 22, and 27 above; cf. n. 9.
93Two other mosaics now omit donor portraits, but both are fragmentary. The north
arcade mosaics show that the portraits of supplicants were sometimes diminutive, and the
damaged east face of the northeast sanctuary pier could just possibly have included one
in the area no longer preserved; so too the west wall mosaic of Demctrios and the angels.
 Leslie Brubaker Elites and Patronage in Early Byzantium sr>

eant to be understood as mediating mid-sixth century and the mid-seventh. Though other distinctions between
•shippers on earth. Instead, as the in- the pre- and post-fire mosaics can be made, one little-noticed but major
' the panels attest, they are "prayers" shift at Hagios Demetrios involves the identity of the "appropriate" donor,
•fire donor or dedication inscriptions, and another involves a change in the way those donors interact with the
s of the building and the saint,90 all saint.
igios Demetrios are prayers of thanks
: prayers in hope of future blessings. Changing Patterns of Elite Patronage
;riptions associated with the pre-fire The pre-fire compositions are more complex than the later panels, with more
tiose name God knows", one fragment elaborate settings that include architectural features, landscape elements and
Demetrios.91 Those attached to the medallion portraits of a battalion of saints to assist the main protagonists,
•epeat the same sentiments. The in- the Virgin and, especially, S. Demetrios. The nine votive compositions—
e God knows" recurs, as do the more two centred on the Virgin, seven on Demetrios—have the following array of
>rld" and for "citizens and outsiders"; supplicants:
s.92
insist on the anonymity of the donor, Location Votive Supplicants
;hteen compositions described above, Figure
image of the donor (s), the northeast West wall, Demetrios adult male presents child;
ti S. Theodore, and the southeast pier south aisle child (fragmentary)
3Ugh the damaged state of the mosaics North arcade, Demetrios adult male (?) (fragmentary)
)le, we can at least conclude that most spandrel A
eluded portraits of supplicant-donors, North arcade, Demetrios adult male
th arcade, and that on the south side spandrel B
North arcade, Virgin Demetrios presents an adult male;
applicants, many of whom are young
spandrel C adult female (?)
iost-fire mosaics on the sanctuary piers
North arcade, Demetrios mother presents child
;1—that closest to the north arcade se-
spandrel D
least pier (Fig. 9)- includes children, North arcade, Virgin mother presents child; adult female
)t supplications but rather Demetrios' spandrel E
:an be simply an accident of survival. North arcade, Demetrios mother presents child; adult male (?)
post-fire mosaics suggests a changed spandrel F
.os Demetrios panels appear to docu- North arcade, Demetrios parents present child; two adult
at votive imagery between roughly the spandrel G females; adult (fragmentary)
North arcade, Demetrios three adult males
spandrel H

j. As is evident from the table, just over half of the compositions involve a
•traits, but both are fragmentary. The north
child, presented to Demetrios or the Virgin by an adult (or adults), usually
pplicants were sometimes diminutive, and the
ry pier could just possibly have included one the mother. This pattern finds no match in the miracle collections associ-
vest wall mosaic of Demetrios and the angels. ated with S. Demetrios, which include only four healings, one of the eparch
86 Leslie Brubakcr

Marianos, one of an official in the prefecture, one of a possessed soldier, and

one of unspecified victims of the plague.94 It does, however, find parallels
in other collections of miracles, notably the Miracles of S. Artemios, the
core of which was written in Constantinople between 658 and 668.95 Here,
nine of the 45 miracles involve mothers and their children (normally sons).96
The healings follow a set pattern: Artemios appears to the mother in a
dream-vision, touches the ailing child or makes the sign of the cross over
it, and explains that the child is healed through Christ.97 This formula is
of particular interest in regard to the Hagios Demetrios mosaics because of
its reference to the healing sign of the cross,98 as visualised by the children
marked with crosses in the mosaics of the north arcade (Spandrels D---G; Figs.
3 5), arid its insistence that the saint heals through Christ, a concept appar-
ently represented in Spandrel D, where Demetrios gestures toward Christ,
who in turn extends his arm out from his medallion and down toward the
child held in its mother's arms (Fig. 3)." The comparison suggests that
the Hagios Demetrios votive panels from the period before the fire fit into a
larger context, and record cures or supplications of a type familiar elsewhere
in the early Byzantine world, where the "appropriate" supplicants were of-
ten families, and especially mothers on behalf of their children, rather than
representatives of officialdom. At Hagios Demetrios these are elite families
a supposition confirmed by the high-status clothing worn by most of the
supplicants—who were able to record their gratitude to the patron saint of
the city, and to memorialise their special relationship with him; but they

94 Miracles 1 4 : Lemerle, Miracles, I, 57-67, 69-71, 75-82, 84-86.

n5 Ed. A. Papadopoulos-Kerameus, Varia graeca sacra (S. Petersburg, 1909), 1 75; repr.
with English trans, and commentary in Virgil S. Crisafulli and John W. Nesbitt, The
Miracles of S. Artemios (Leiden, 1997); on the date see John F. Haldon, "Supplementary
Essay: the Miracles of Artemios and Contemporary Attitudes, Context and Significance,"
in Crisafulli and Nesbitt, csp. 33-35.
96Miracles 10-12, 28, 31, 36, 42-43, 45: Crisafulli and Nesbitt, Miracles, 94-101, 154-57,
162-65, 188-93, 216-19, 222-25.
97Discussion in "Icons," 1236-37.
98In miracle 10, Artemios makes the sign of the cross "all over [the child's] body"; in
miracle 31, he makes the sign of the cross over the child's diseased testicles: Crisafulli and
Ncsbitt, Miracles, 96, 164.
"A variant on this formula, with the saint (Peter, in this case) below a medallion
portrait of Christ flanked by portraits of the mother and her child, recurs on a Sinai icon
(B.5) usually dated to the sixth or seventh century: see my "Icons," 1236-37.
 Leslie Brubaker Elites and Patronage in Early Byzantium S7

cture, one of a possessed soldier, and nonetheless fit into a pattern that, if we may generalise from the Miracles of
ic.94 It does, however, find parallels Artemios, seems to have crossed class and status boundaries.
ly the Miracles of S. Artemios, the The mosaics commissioned after the fire do not, on the whole, continue
inoplc between 658 and 668.95 Here, this model. Three of the panels on the sanctuary piers, tribelon wall and
and their children (normally sons).96 north arcade no longer include donors, though two of these the damaged
rtemios appears to the mother in a panel showing an orant saint on the east face of the northeast pier, and the
or makes the sign of the cross over panel that portrays the Virgin and S. Theodore (Figs. 10-11)- incorporate
d through Christ.97 This formula is prayer inscriptions. Only the portrait of S. Sergios (Fig. 13) on the west
iagios Demetrios mosaics because of face of the southeast pier lacks both supplicant-donors and a supplicant-
cross,98 as visualised by the children inscription, and his pendant on the west face of the northeast pier (Fig.
.e north arcade (Spandrels D G; Figs, 9) is the only saint in this later group to be associated with children. The
sals through Christ, a concept appar- remaining four compositions all show S. Demetrios with officials of the church
e Demetrios gestures toward Christ, or of the city. On the piers and tribelon wall, the saint is shown in direct
his medallion and down toward the physical contact with the donors: he rests his arms on the shoulders of two
3)." The comparison suggests that (now faceless) bishops in the tribelon panel (Fig. 6), of a bishop and a secular
n the period before the fire fit into a official on the north face of the southeast pier (Fig. 12), and of a deacon on
jlications of a type familiar elsewhere the east face of that pier (Fig. 14). As we have seen, this same deacon,
,e "appropriate" supplicants were of- and the bishop from the adjacent panel, recur in medallion portraits on
behalf of their children, rather than either side of Demetrios in the inserted panel of the north arcade; the deacon
is Demetrios these are elite families - appears again with the bishops and a priest in the tribelon mosaic. By the
tatus clothing worn by most of the middle of the seventh century, the votive mosaics have resolutely shifted
;heir gratitude to the patron saint of their focus from families to officials;100 and, as we have already seen, the
cial relationship with him; but they accompanying inscriptions respond to this shift of emphasis by being more
civically inclined than the earlier prayers. In their emphasis on the well-born
and illustrious, the post-fire mosaics at Hagios Demetrios conform much more
closely with the seventh-century Miracula texts—both versions—than do the
, 69-71, 75-82, 84-86.
lec.a sacra (S. Petersburg, 1909), 1 75; repr. earlier panels.
;il S. Crisafulli and John W. Nesbitt, The The post-fire mosaics also record a changed attitude toward the saint.
s date see John F. Haldon, "Supplementary Demetrios is no longer the object of supplication, nor is he any more pre-
orary Attitudes, Context and Significance," sented as larger than life: he is instead shown offering protection to figures
depicted in the same scale as himself. This change in attitude is accompanied
ifulli and Nesbitt, Miracles, 94-101, 154-57,
by a change in accoutrements. In the post-fire mosaics, the silver kiborion
is no longer represented, countryside and architectural fantasies have given
f the cross "all over [the child's] body"; in way to solid walls (representative of the city walls that Demetrios protects so
the child's diseased testicles: Crisafulli and

it (Peter, in this case) below a medallion 10()Skedros, Saint Demetrios, 100-102, thinks that Denietrios is being "universalised",
nothe-r and her child, recurs on a Sinai icon that his power is here "more pervasive and civic" (ibid., 101); I am not, however, convinced
itury: see my "Icons," 1236-37. that the shift to civic officials expands the saint's powers.
88 Leslie Brubaker

well?), and the saint has been abandoned by his companions in medallions.
Dernetrios has been transformed from a healing figure — a member of a large
confraternity of saints, many represented in the earlier votive panels— to
a protective figure, of unique importance to Thessalonike as its urban de-
fender; 101 his clientele has changed from individuals and families to represen-
tatives of the city; their attitude toward him has changed from supplication
to affiliation.
This shift finds parallels elsewhere in the empire. It is precisely in the
period between the two mosaic campaigns at Hagios Demetrios that a be-
lief in holy figures as urban protectors surfaces. Sometime between 550 and
590, the image of Christ not-made-by-human-hands (acheiropoieton) was
first credited with the protection of Edessa;102 shortly after 626, the Virgin
was recognised as the protectress of Constantinople, responsible for the re-
treat of the Avars.103 The Miracula texts claimed this role for Demetrios in
Thessalonike, and the second programme of mosaics visually confirmed it.
The shift of emphasis apparent in the later votive mosaics at Hagios
Demetrios prompts one final observation, on the importance of recording the
building commission. Whether or not the builders of the earlier church were
commemorated somewhere in the church, the men responsible for the post-
fire restoration ensured that their contribution was recorded in at least three
locations: the sanctuary piers, the tribelon wall, and the panel inserted into
the apex of the sixth arch of the north arcade. One wonders why they were
so insistent:104 perhaps each of these three sites had particular significance to
the rebuilding process or to the cult practices of the church? The north arcade
panel, for example, may have been inserted in a particularly badly damaged
area of the mosaic sequence — the drawings and notes of George and Kluge do
not help us here but it is also possible that the old mosaic was deliberately

101For the subsequent development of the iconography of Demetrios as a military saint,

see e.g. Andreas Xyngopoulos, O eikonographikos kyklos tes zoes ton Hagiou Demetriou
(Thessalonike, 1970); Christopher Walter, "St Demetrius: the Myroblytos of Thessa-
lonika," Eastern Churches Review 5 (1973), 157—78.
102See Averil Cameron, "The History of the Image of Edcssa: the Telling of a Story,"
in Okeanos, Essays presented to Ihor Sevcenko (Cambridge, MA, 1983), 80 94; repr. in
eadcm, Changing Cultures in Early Byzantium (Aldershot, 1996), Essay XL
103See Averil Cameron, "Images of Authority: Elites and Icons in Late Sixth-Century
Byzantium," Past & Present 84 (1979), 3-35; repr. in eadern, Continuity and Change in
Sixth-Century Byzantium (London, 1981), Essay XVIII.
104 Aiid, given this insistence, why the donors were not all named—though perhaps they
were in the tribelon panel, which has lost its lower half.
 Leslie BrubaJcer Elites and Patronage in Early Byzantium 89

d by his companions in medallions, destroyed in order to allow the portraits to be placed in the centre of the
healing figure—a member of a large arcade, and in a location close to and visible from the silver kiborion in the
ed in the earlier votive panels- to nave. But whatever the reasons for the placement of the panels, the patrons
2e to Thessalonike as its urban de- of the rebuilding wanted their work, and their association with S. Demetrios,
individuals and families to reprcsen- to be remembered by the inhabitants of and visitors to Thessalonike. In this,
hiin has changed from supplication they were not so different from the patrons of the earlier mosaics of the north
arcade and west wall of the church.
n the empire. It is precisely in the
jns at Hagios Dernetrios that a be- Elites and Their Objects: a Conclusion
urfaces. Sometime between 550 and
•human-hands (acheiropoieton) was The church of Hagios Demetrios is usually lauded as a landmark in the de-
jssa;102 shortly after 626, the Virgin velopment of sanctified space, and it has been seen as representative of the
nstantinople, responsible for the re- pre-iconoclast imagery of saints.105 This should not, however, blind us to its
ts claimed this role for Demetrios in significance as an index of the changing nature of elite patronage in the late
te of mosaics visually confirmed it. antique and early Byzantine world.
the later votive mosaics at Hagios The mid-sixth-century panels record the desire of the Thessalonikari elite
i, on the importance of recording the to be recognised as clients of Demetrios, who as individuals and as families
le builders of the earlier church were enjoyed the special relationship with the saint that was commemorated in
:h, the men responsible for the post- the public mosaics of the church. Demetrios is depicted in these panels as
.bution was recorded in at least three interacting with individuals and groups either in the church itself or in the
ilon wall, and the panel inserted into countryside, as interceding on their behalf with the Virgin and with Christ,
arcade. One wonders why they were and as able to channel access to a spectrum of other saints, his companions,
•ee sites had particular significance to represented in medallions.
:tices of the church? The north arcade By the second quarter of the seventh century, Demetrios' power has be-
rted in a particularly badly damaged come more focussed, and so has his elite clientele. Inscriptions and images
igs and notes of George and Kluge do document his transformation from a healing saint, the object of elite sup-
: that the old mosaic was deliberately plication, to a figure of urban protection, his arms around those members
of the elite who themselves represent the city—the eparch and the clergy.
ariography of Demetrios as a military saint, This transformation responds to changing attitudes toward sanctity in the
kikos kyklos tes zoes tou Hagiou Dcmc:triou early Byzantine world, but it also respects changes in the self-presentation
St Demetrius: the Myroblytos of Thessa- and self-identity of the elite who commissioned these mosaics. The appropri-
57-78.
5 Image of Edessa: the Telling of a Story," ate client is no longer an individual or a family, but a representative of the
ko (Cambridge, MA, 1983), 80-94; repr. in community; Demetrios' intercession is no longer personal, but civic.106
n (Aldershot, 1996), Essay XI.
.ty: Elites and Icons in Late Sixth-Century 105See, e.g., Belting, Likeness and Presence, 82-88; Maguire, The Icons of their Bodies,
; repr. in eadem, Continuity and Change in 101-102.
.say XVIII. 106See further the classic remarks of Peter Brown, "A Dark-Age Crisis: Aspects of the
•s were not all named—though perhaps they Iconoclastic Controversy," English Historical Review 88 (1973), 1-34; repr. in idem, Society
lower half. and the. Holy in Late Antiquity (Berkeley, 1982), 251-301. On the Miracula texts as
90 Leslie Brubaker

In this respect, the evidence from Thessalonike again seems to fit into a
larger pattern. It has been argued that the sixth and seventh centuries saw
the disintegration of the Roman model of the aristocratic urban elite; that
social structures based on extended kinship networks were supplemented and
gradually replaced by structures forged from within the official hierarchies of
the church and the state.107 This is precisely the situation visualised by the
mosaics at Hagios Demetrios, where the elite families portrayed as favoured
clients of S. Demetrios in the sixth century have disappeared by the seventh,
their places taken by the civic and ecclesiastical hierarchy that now represents
the population of Thessalonike.
ADDENDUM: In an article that appeared after this study was completed
Jeffrey Anderson argued—as I have above, but on different grounds—that
the mosaic of the Virgin and S. Theodore is contemporary with the other
pier panels at Hagios Demetrios: J.C. Anderson, "A Note on the Sanctuary
Mosaics of S. Demetrius, Thessalonike," Cahiers archeologiques 47 (1999),
55-65.

expressions of the independence of Tllessalonikc from Constantinople, see Ruth Macrides, »

"Subversion and Loyalty in the Cult of St Demetrios," Byzantinoslavica 51 (1990), 189-97.
1(17See, e.g., John F. Haldon, Byzantium in the Seventh Century: the Transformation of
a Culture (Cambridge, 1990, rev. edri. 1997), 387-402. J