Content uploaded by Moshe Fischer
Author content
All content in this area was uploaded by Moshe Fischer on Jan 03, 2017
Content may be subject to copyright.
THE ROMAN TEMPLE AT KEDESH,
UPPER GALILEE:
A PRELIMINARY STUDY
Moshe Fischer, Asher Ovadiah and Israel Roll
I. INTRODUCTION
Ancient Kedesh lies about 15 km. north of Safed and 2 km. northwest of the
moshav
Ramot
Naphtali near the Israel-Lebanon border (Fig. I). The site comprises a tell at the western edge
of the Kedesh Valley and a low hill overlooking the valley some 200 m. to the east. (The site
appears on maps of the Survey of Israel as Tel Qedesh.) The importance of Kedesh in
antiquity is attested both by the surface remains strewn over the site and the numerous
references to it in the historical sources. Kedesh is mentioned in one of the Zenon papyri
(PZen: 59004), as well as in the account ofthe battle between Jonathan the Hasmonean and the
generals of Demetrius according to I, Mace. XI 53:73 (Abel 1949; see also
los. Ant.
XIII,
154). Josephus
(War
IV, 104) relates that Titus' army set up camp at Kedesh during the
campaign against Yohanan of Gush Halav and that following the massacre of the Jews of
Caesarea in 66 C.E. the latter avenged themselves by attacking many Gentile settlements,
including K«iliacra ('trov) 'wpirov, "Kedesh (of) the Tyrians," i.e. Kedesh in the region of Tyre
(los. War
II, 459). Kedesh was known as one of the strongholds of Tyre in the Galilee and in
unceasing conflict and enmity with the "Galileans"
(los. War
IV, 105).
According to both the epigraphical and archaeological evidence, Kedesh flourished during
the 2nd-3rd centuries C.E. The town is mentioned also in later historical sources. Thus
Eusebius
(Eus. Onom.
116:10), notes that Kedesh lies 20 miles from Tyre. The 10th century
Arab geographer Muqaddasi refers several times to the town (Marmardji 1951:5,96,102,105,
165).
On the hill east of the tell are the remains of a well preserved monumental temple (PI. 25: 1)
and west of this are remnants of mausolea and decorated sarcophagi. Rock-hewn tombs were
discovered at the north and northwestern fringes of the hill. During the 19th century the site
was surveyed in varying detail by Renan (1871:685-686), Guerin (1880:355-362), Wilson
(1869:66-71) and especially by Conder and Kitchener of the British Survey of Palestine (SWP
I: 226-230) whose work has been the main source for our knowledge of the temple until
recently. Since these 19th century investigations the site has been explored by only a few
scholars (e.g., Fenner 1905:101-102; Dalman 1914; Watzinger 1935:98-99). During 1976-1977
the authors made an extensive survey of the site and in 1981 and 1983-1984 they conducted
three seasons of excavations on behalf of the Institute of Archaeology and the Department of
Classical Studies of Tel Aviv University.
I
The archaecological survey, in which the architect Dr. D. Chen participated, was conducted on behalf
of the Department of Classical Studies of Tel Aviv University; we would like to thank Prof. Shalom
Perlman, Head of the Department, for his encouragement. The excavations, headed by the authors,
were carried out with the participation ofM. Aviam, A, Amitai and D. Michaeli of the Society for the
Protection of Nature in Israel (Mt. Meiron Field School) and the architect G. Solar, Jerusalem.
146
Fischer, Ovadiah and Roll: Roman Temple at Kedesh
~
NN
\0
~
N
000
290
A
~
I
Qlryat Shemona
290
\
0
i
eKefar Szotd
~
i
~
i
q, ieShamir
lq
i
'v
(
i
eGonen
280
KEDESHi
280
fi
~ eMalkiyya
._._.1 EMEQ
IjULA
Tel Mal!!a
eVesud HaMctaia
0
1
km.
270
!
VO
~
NN
\0
0N
000
Fig. I. Location of Kedesh.
II. THE TEMPLE COMPLEX
The architectural complex at Kedesh, located on the upper part of the southern slope of the
hill (map ref. 20022/27984), includes a large temenos (sacred enclosure) with a monumental
temple at its centre (Fig. 2). Numerous architectural fragments are scattered in and around the
temenos.
1. Theperibolos
The peribolos enclosing the temenos is rectangular in plan and measures about 55x80 m., as
revealed by the traces of a wall visible above ground. A number of segments, varying in length
between 2.00-16.50 m., that project above the surface have enabled us to reconstruct its
ground plan. These segments are located along the western side of the peribolos, at its
southwest corner, along its southern wall (three segments) and at its southeast corner, where
the wall makes a double right angle. On the western side is an area 10 m. wide, which is
separated from the rest of the temenos by a north-south wall. About 5 m. of the southern part
of this wall is preserved. A threshhold stone in secondary use was found incorporated into this
wall. At one end it has a deep round socket for the door pivot and in its centre there is a
rectangular slot for the door bolt; the other end is broken off.
Students of the Department of Classical Studies, members of the Society for the Protection of Nature
in Israel and youth groups from all over the country also participated (see Ovadiah a.o. 1982). The
manuscript of this article was translated by E. Heineman.
147
148
I
I
I
I
,
I
I
/I
II
II
II
II
II
II
II
II
II
II
II
,I
il
Tel Aviv II (1984)
~...................
------------
<F5
I
\ J
\ I
'-
/
----
Fischer, Ovadiah and Roll: Roman Temple at Kedesh
The masonry of the peribolos is not uniform. Some segments, such as the surviving stretch
of the western wall, are built of alternate headers and stretchers, as typical of the Hellenistic
period, but in others the stones are arranged randomly_ The space between the outer and inner
faces of the wall is filled with earth and stones, a method of construction also typical of the
Hellenistic period (Pergamon V, I: Pis. XXII; XXIV:2). The cornerstone is preserved in the
southwest angle, where only one course is visible. However, farther east the wall still stands to
a height of five courses. Here headers and stretchers were used only in a few places.
Occasionally, instead of one broad stretcher there are two smaller ones. A few of the stones
have a central boss and marginal dressing. Only one course is preserved in the central part of
the southern segment. Here the stones are roughly dressed and arranged as double headers
and stretchers. Eastward along this segment three courses have survived (PI. 25:2). The stones
of the two lower courses are randomly arranged, while the upper course, which shifted slightly
inward forming a narrow ledge, is composed of headers and stretchers. Most of the stones
here are crudely hewn. Further along there is an additional segment of the south wall built of
crudely hewn headers and stretchers, of which only one course is preserved. In the southeast
corner there are four courses of headers and stretchers randomly inserted among stones of
various shapes, some of which have marginal dressing and a roughly worked projecting boss.
This corner was especially interesting because it forms a rectangular projection extending
outwards from the peribolos. Judging by its location and the solidity of its construction,
compared to the other segments of the peribolos, this segment was presumably part of a
rectangular corner tower. The eastern fa«ade of the peribolos, only partially preserved, was
probably symmetrical in design and had a monumental entrance leading into the temenos,
such as those of the sacred complexes of Heiios at Kasr Naus and Zeus at Hossn Soleiman
(Baetocaece) in Syria, where the periboloi are very similar to the one at Kedesh (Krencker and
Zschietzschmann 1938:8-12,67-78; Figs. 17,23, 104, 110; Pis. 5, 31, 37). The "baroque"
characteristics of this design correspond well with the general trend of Roman architecture in
the Orient (Lyttelton 1974:221-222).
2. The temple
In view of the parallel walls of the peribolos and the temple, it is obvious that the two
structures are an integrated architectural complex, although the temple did not stand in the
exact centre of the temenos. The approximate distances between the temple and the walls of
the peribolos are 27 m. on the west, 15 m. on the south, 19 m. (estimated) on the north and 42
m. on the east.
The temple consists of a rectangular cella and a portico (Fig. 3). Its outer dimensions
(without the portico) are 20.66x22.63 m., its inner dimensions 17.60x20.04 m. and its overall
length (including the portico) is 31.25 m. All that remains standing is the east fa«ade of the
cella, which is preserved to a height of 7.50 m., or 14 courses (PI. 26). The fa~ade has three
carved and ornamented entrances: a central opening symmetrically flanked by two smaller
ones. Of its three other walls, only the foundations or lower courses remain. The neatly
worked ashlars are laid according to the typical Roman dry-wall method, smoothly fitted to
each other. Most of the blocks are either crystallized limestone (malaka) or dolomite (the
latter from the region of Manara), while some cretaceous and lithographic limestone was also
149
150
r.:::,---
I( )I
L':::.J- __
r::;,..,
I( )I
L:=...J
I
I
I
I
~
II
II
II
II
II
II
i'
I
I
I
L.
II
II
II
II
II
II
r
I
I
I
L.
II
II
II
II
II
II
II
l'
I
I
I
L.
II
II
II
II
II
II
II
i'
I
Tel Aviv
II
(1984)
E
on
o
Fischer, Ovadiah and Roll: Roman Temple at Kedesh
used (e.g., in the lower courses of the central entrance threshold).2
No remnants or indications of columns were found inside the cella to suggest that this inner
space was divided into aisles. Nor were there traces of any piers against the north and south
walls that might have supported the roof. This suggests that the roof, despite the considerable
width of the hall (17.60 m.), was not supported by columns but by other me.ans. Although
uncommon, this would not be exceptional, as may be seen from the reconstructions of some
Roman temples in Syria, such as the Great Temple at Kal'at Fakra and the temple at Der el
Kal'a (Krencker and Zschietzschman 1938:1-3,40-46; Pis. 2,20), both of which have an inner
width of 15 m. The roof could have been supported by transverse wooden beams, which, in
turn, carried a series of lighter timbers or crossrafters intersecting and slotted into the trans-
verse beams. It is also possible that in order to reduce the span of the beams they rested on
consoles fitted against the north and south walls of the temple. This was probably the method
used to support the roof structure in several Roman temples in Syria, like those at Kasr N aus,
Hossn Niha and Majdal Anjar (ibid.: 13-19, 128-129, 182-191; Pis. 7, 8, 56, 58, 59, 74d-e).
During the excavation of the portico many fragments of terracotta tiles were found, both of
solenes and kalypteres, which obviously had been used to cover the roof of the cella.
A paved section of neatly fitted rectangular slabs was found in the eastern part of the hall,
approximately on the same level as both the southern entrance threshold and the pavement of
the portico. There were also traces of the underpinnings of three steps leading down from the
central entrance threshold to the floor of the cella.
An apse, measuring 5.60 m. across, abutted onto the inner side of the western wall of the
cella, opposite its central entrance. This was evidently intended for a statue or relief of the
deity to whom the temple was dedicated, a kind of adyton, such as those of the temples at
Rahle, Burkush and Palmyra (ibid.:226-228, 240-243; Pis. 94, 96, 100; Gawlikowski and
Pietrzykowski 1980). This apse was apparently not contemporary with the temple but added
at a later date. Within the apse were stones that had fallen from its vault, which should enable
its eventual reconstruction. It is possible that further excavation of the temple will yield
architectural finds of the sort that will help to clarify its interior layout. Some of the masonry
courses of the east fal<ade are clearly shifted out of line (PI. 27: I), and a similar disturbance is
evident in the keystones above the two side entrances. This could have been caused by an
earthquake some time in the past. One likelihood is the devastating earthquake of May 19,
363 C.E. that affected the entire region, from northern Galilee to Petra and from the Mediter-
ranean coast to the Jordan Valley (Russel 1980; Hammond 1980).
The central entrance
The central entrance is not entirely preserved; its north side was demolished at some stage
and partially but carelessly restored. Its southern doorjamb is preserved in situ (PI. 28), while
several of its architectural elements were recovered during the excavation of the portico (see
below). From the excavations it emerged that the threshold of the southern entrance, which
consists of a well hewn stone slab, was some 90 cm. below the level of the central entrance
threshold.
2 For this information we are indebted to A. Livnat of the Department of Geophysics of Tel Aviv
University, who examined the various kinds of building stones of the temple during a visit to the site.
151
Tel Aviv II (1984)
It is not yet clear how the cella was entered, though at present the evidence seems to indicate
that the priests used the two side entrances, since the southern entrance threshold is more or
less level with the floors of the portico and the cella. Presumably, the central entrance
threshold, higher by one metre than the portico floor, was reached either by means of a
staircase, like that in the portico of the "Exedra" at Hossn Soleiman (Krencker and
Zschietzschmann 1938:99-101; 41-42), or by a ramp, like that of the temple of Zeus in
Olympia (Mallwitz 1972:214; Lawrence 1962: 154-155; Figs. 89-90). There is a vertical groove
1.43 m. in height in the bottom third of the doorjamb on the side facing the threshold. The
threshold, which is 1.04 m. wide and 4.64 long (including the width of the jambs), has two
step-like levels. There are a number of regularly spaced square sockets cut into the upper level
and two larger square sockets at each end of it (PI. 27:2). The doorjamb groove and threshold
sockets were used to hold a screen or lattice that barred access into the temple when its doors
were swung open but allowed worshippers to see the statue or relief of the deity inside the
apse. There are four square sockets in the lower step of the threshold, which faces the cella.
The pivots of the double doors fitten into the outer, larger sockets, while the two inner sockets
took the door bolts. The estimated width of the entrance between its two jambs is 2.70 m. The
southern jamb, which is preservred to a height of 4.50 m., is 1. 03 m. wide. Cut into the inner
face of this jamb are five parallel horizontal slots, each 14 cm. wide and set 93 cm. apart,
which were probably meant to provide clearance for the slightly protruding bars of the door
panels, thereby permitting the doors to be fully opened. Similar slots are found on the Golden
Gate of the Temple Mount in Jerusalem (de Vogue 1864:12: Fig. 10; Creswell 1969:464; Fig.
527).3
The northern entrance
The northern entrance is completely preserved. Its neatly carved lintel and jambs are each
cut from a single block (PI. 29: 1). The width of the opening is 1.03 m., and its estimated
height, based on the ascertained height of the southern entrance, is 1.69 m. (So far, we have
been unable to excavate the northern entrance due to the massive stones blocking it and the
danger of the entire fac;:ade collapsing.). The width of the jambstones and height of the lintel
are identical, i.e. 54 cm. The mouldings consist of a fillet (4 cm.), a cyma recta (7 cm.) a torus
(2 cm.) and three stepped fasciae, increasing in width from inside to outside (9, 12 and 18 cm.
respectively). Thefasciae are surmounted by a cyma reversa and a fillet. The southern end of
the lintel is broken but held in place by the adjoining stones.
The southern entrance
This entrance, also consisting of a monolithic lintel and two monolithic jambs, is similar to
the northern one but narrower (width 0.85 m.). The height of the lintel and the width of the
jambs are identical: 54 cm. The mouldings consist of a fillet (4.5 cm.). a cyma recta decorated
with a floral motif (7 cm.), a torus (2 cm.) and three fasciae, 10, 13 and 16 cm. wide respec-
tively (PI. 29:2).
3 de VogUe'sinterpretation as to the function of the slots in the Golden Gate (which was also accepted
by Creswell) differs from ours.
152
Fischer, Ovadiah and Roll: Roman Temple at Kedesh
Only three courses are preserved above this entrance, although when the site was visited by
Wilson in 1866 he found six courses (SWP I: facing p. 226). The second course above the lintel
bears a tabula ansata (0.42x 1.54 m.) on which no trace of the inscription remains except for
the letter '0'.
In the middle of the first course above the lintel (both here and above the northern entrance)
is a carefully dressed keystone with a projection on each side, which originally fit tigh,tly into
the corresponding projections on the two adjacent stones, which held it in place. Below the
keystone (before it slippped) and its adjacent stones there was a narrow space that was
deliberately left open in order to relieve and balance out any excessive pressure on the centre
of the lintel by distributing it to either side. The use of a flat arch for relieving the weight
above the lintel is a structural solution that was very common in Roman architecture and has
generally proved itself (Robertson 1969:321; PI. XXIVb; Marasovic and Marasovic 1970:
Figs. 8, 14).
The krater reliefs and the niches in the far;ade
North and south of the two side entrances, one on each side, are reliefs of volute kraters with
ribbed bodies, and above them small apsidal niches. A downward slanting groove runs from
each niche, cutting through the width of the wall and leading to a larger niche on its interior
face. The internal southern niche is rectangular (PI. 30:2) while the northern one is arched (PI.
26:2). The northern krater (PI. 30: 1), which is 1.67 m. north of the jamb of the northern
entrance, is set in the course below the one that incorporates the lintel. The niche here is 40
cm. high and 30 cm. wide. The southern krater relief is 1.58 m. south of the jamb of the
southern entrance and set into the two courses below the lintel course. The stone on which the
relief is carved is fractured and the relief is poorly preserved, although the base of the krater,
part of its body and one of its volute handles are still indentifiable. The mouths of the kraters
are rounded concavities.
The function of the grooves between the outer and inner niches has been variously inter-
preted by the scholars who have examined them in the past. Thus Wilson (SWP 1:227)
suggested that the grooves were intended for confessional purposes, while Conder and Kit-
chener (ibid.: 227) thought that worshippers could drop donations of coins for the temple
treasury into the grooves and in return get the priest's oracular message, without the latter
being required to show himself. Our renewed examination of the kraters, grooves and niches
indicates that these were elements of a system of libations. During the libation ceremony a
liquid of some sort - blood from sacrificed animals, wine or oil - was poured into the outer
niche and flowed along the groove to the inner niche where it was collected in a receptacle
placed at the bottom. This ceremony was apparently part of a rite that connected the world of
the living with the world of the dead (Oeconomus 1921). The fact that the temple at Kedesh is
closely surrounded on three sides by tombs and mausolea strengthens this interpretation. The
two eagles on the fa<;:adeof the cella (see below) were perhaps also a funerary symbol (Ronze-
valle 1911:14-15), as well as being associated with the god Baalshamin (see Section IV). By
dedicating their libation offerings to the exalted and worshipful godhead of the temple, the
relatives of the departed would seek to communicate directly with the deity in order to ease
the plight of the dead. To the best of our knowledge, such a libation system for a funerary
153
Tel Aviv II (1984)
ritual is unique in the temple architecture of the Roman world, and hence its existence at the
Kedesh temple is of the greatest interest.
Cut into the block above the southern krater and niche is an apsidal niche,4 in which a
human figure wearing a toga is carved in sunken relief (PI. 31: I). The upper part of the figure
is unclad, while the toga covers its right leg almost to the ankle and its left leg from the knee
upward, then crosses the body diagonally until it reaches the left shoulder and folds over the
arm to hang down in a flat drape. The folds of the toga are carved in broad, shallow relief;
both the figure and robe lack plasticity and depth. The figure grasps in its right hand what
seems to be some kind of basket-handled pear-shaped vessel with a pointed base; the body of
the vessel is marked off by a small, shallow depression from its handle. The figure holds a
spear in its left hand. Apparently the libation ceremony is symbolized by this relief, the vessel
representing the pouring of the libation, and the heroic posture of the figure reflecting the true
attitude of the worshipper. There are some incised lines, mostly curving, to the left of the
figure, whose significance so far escapes us. The figure stands firmly on its right leg, while the
left leg is disposed somewhat freely (contraposto position). The feet point in opposite direc-
tions. The heroic stance of the figure is reminiscent of many of the statues of Hellenistic rulers
(Bieber 1967: Figs. 299, 685). Stylistically and technically similar reliefs dating to the 2nd-3rd
centuries C.E. were discovered in Rome (Toynbee and Ward-Perkins 1956: PI. 15) and in
Stobi, Yugoslavia.5
Slightly north of the jamb of the northern entrance is a recess measuring 38x72 em. with a
roughly cut stone slab set into it. The ashlars around the four sides of the recess have
countersunk borders that were evidently designed to take a stone slab that bore a relief or an
inscription that is now lost.
The portico
The portico in front of the temple is of the same length as the width of the far,:ade of the cella
(20.66 m.). In our excavations we uncovered the stylobate (width 1.32 m.), three pedestals
in
situ
(PI. 31:2) and the emplacements of two more. These elements allow us to reconstruct a
sixth pedestal, making the portico a
hexastylos
consisting of two groups of three columns
each. The distance between the pedestals in each group is 3.62-3.63 m. from axis to axis, while
the distance between the two pedestals opposite the central entrance of the cella is 4.10 m. The
central entrance of the portico between the two central pedestals is slightly south of that of the
cella, thereby upsetting the symmetry of the temple far,:ade.
The depth of the portico, between the stylobate and the far,:ade of the cella is 5.75 m.
Between the southeast pedestal of the portico and the southeast corner of the cella another
pedestal was found
in situ.
Presumably, another pedestal in the same position is to be found
opposite of the northern side, but this area has not yet been excavated. These pedestals
4 A similar example of a single apsidal niche to the left of the main entrance of the cella is found in the
temple inside the small temenos at Hossn.Soleiman (Krencker and Zschietzschmann 1938:98; PI. 40).
In the temple at 'Ain Hersha two niches are discernible, one on each side of the cella entrance
(ibid.:252; Pis. 106-107, 109, lower). In the temple at Hibbariye, there are four niches in the fa~ade of
the cella, two on each side and one above the other (ibid.:218; Fig. 327; PI. 89).
5 Our thanks are due to Dr. S. Diihl of the German Institute of Archaeology in Instanbul, who has
kindly shown us the photograph of this relief.
154
Fischer, Ovadiah and Roll: Roman Temple at Kedesh
reduced the north-south length of the portico, thereby making it easier to span its roof.
The pedestals, which are 56 em. high, are square (1.28x 1.28 m.), and consist of identical
elements (Fig.5; PI. 32:1): a plinth, an elaborate profile with two cymae above and below a
high torus, and a fillet. The column bases standing on the pedestals were of the Attic type, but
so iar only one base has been found (Fig. 5; PI. 33:1). The plinth measures 1.26x 1.26 m.; the
overall height of the base is 48 em.; the diameter of its upper flangs is 90 em., exactly the same
as the bottom diameter of the monolithic column that stood on it. The elements making up
the base are a plinth, a torus, ascotia with two fillets, and an upper torus. There is a round
socket (6 em. in diameter, 5 em. deep) in the centre of the base to enable the column to be
joined to it by means of a dowel. So far four monolithic columns have been found (Fig. 5).
One of them, which is 5.62 m. high, was entirely uncovered. On the basis of what is preserved
of the eastern wall of the cella and from the various decorated architectural elements that
belonged to the entablature found near the portico (see Section III below), the original height
of the temple may be estimated at approximately 12 metres.
Several sections of the courtyard and portico paving were uncovered in the excavations. It
consists of rectangular and triangular limestone slabs, carefully cut and fitted, apparently part
of a dromos or paved ceremonial way used for religious processions, probably leading from
the main entrance in the eastern wall of the peribolos to the temple.
The southern wall of the temple
At a distance of 8.81 m. from the southwest corner and 10 m. from the southeast corner of
the building there is an opening in the wall which has been blocked (PI. 32:2). The opening
was 3.82 m. wide and its corners are smoothed and right angled. The blocking stones are set
back some 10 em. from the outher face of the wall and are about 10 em. lower than the
preserved upper courses of the wall. Presumably there had once been a side opening at this
spot, which was blocked up at some stage in the past, but the function of the original opening
and the date of its blockages are still unclear.
Outerface of temple wall
Four courses of the temple wall are visible in its southwest corner. The first course, which
here appears as the lowermost course, is 45 em. high. Further east, where the ground level
drops, there is another, lower course, 42 em. high, which protrudes about 11 em. from the face
of the wall (PI. 32:2).The second course of neatly dressed stones, 36 em. high, is carefully laid
on top of the underlying one. The third and fourth courses, 50 and 60.em. high respectively,
are set back about 31 em. from the first course. This was necessitated by the carvings of the
second course of carefully fitted blocks, whose profile of pseudo-pilasters is meticulously and
skillfully worked (Fig. 4). Another delicate profile of pseudo-pilasters was discovered running
for about five metres along the western end of the north wall on its outer face (PI. 32:3)). This
profile is very similar to that at the southwestern corner of the temple. We therefore infer that
the entire outer face of the temple incorporated pseudo-pilasters at regular intervals.
A rough dating of these profiles to the 2nd-3rd centuries C.E. is permissible on the grounds
of their similarity to those on the city wall of Jerusalem below the present Nablus Gate
(Hamilton 1940:5-8; Figs. 2, 4, 5; PIs. I-II; Hennessy 1970:22-25; PI. XIV). This dating
accords well with the other architectural details of the temple at Kedesh. Similar profiles are
155
o
,
JOcm.
,
Tel Aviv II (1984)
Fig. 4. Profile at southwest outer angle of the temple.
156
Fig.
S.
Reconstructed column of portico.
Fischer, Ovadiah and Roll: Roman Temple at Kedesh
known in temples in the Syro-Phoenician region (Krencker and Zschietzschmann 1938: Figs.
19a, 40d, 397).
III. ARCHITECTURAL DECORA nON
The decorated architectural elements found in and around the Kedesh temple came from its
portico (PI. 33:I), the fa9ade of the cella and the external faces of the other walls; they include
fragments of the entablature (Pis. 33; 34), a pilaster capital (PI. 35:2), elements of the entran-
ces of the cella fa9ade (Pis. 28:2; 29; 35:1; 36:1) and the pseudo-pilaster profiles described
above (PI. 32:3).
1. The entablature
So far eight decorated items have been found that apparently belong to the cornice of the
entablature. Seven of these, which fell from the upper part of the portico, have a uniform style
of mouldings. The eighth item, however, which has different dimensions as well as differently
decorated mouldings, most likely belongs to the entablature of the central entrance of the
cella. Because of the small number and the state of preservation of the fragments, it has so far
proved impossible to reconstruct the cornice of the portico. The preserved dimensions of the
fragments are (height x length x width):
Item 1:0.68x 1.00x 1.33 m.
Item 2: 0.38-0.45x 1.04xO.63m.
Item 3: 0.3lx1.26x1.04 m. (PI. 33:2).
Item 4: O.70x 1.11x1.20 (PI.34: I).
Item 5: 0.60xO.70xI.60 m.
Item 6: 0.45xO.70x1.45 m.(PI.34:2).
Item 7: 0.70xO.95x1.50 m. (P1.34:3).
Item 8: 0.22x0.49x 1.00 m. (PI. 35: I).
On the first seven items the following mouldings are distinguishable: (a) ovolo (egg-and-
dart); (b) dentils; (c) cyma reversa (identical to moulding e); (d) modillions-and-coffers framed
by an ovolo strip; (e) cyma reversa (identical to moulding c); (f) fluted corona; (g) astragal;(h)
sima with anthem ion, containing ancanthus leaves, palmettes and lotus flowers.
Although the eighth item (pI. 35:I) has the ovolo, dentils, cyma reversa and modillions-and
coffers, which are similar to the mouldings listed above, it differs in several respects from the
entablature items of the portico:
(I) The cyma reversa above the ovolo is not decorated with the usual relief of stylized
leaves but left plain. This combination of decorated and plain bands was common in the
architecture of the East Roman Empire (Baalbek I: PI. 102, upper).
(2) There is no decoration between the modillions-and coffers and the corona.
(3) The corona was left blank. The astragal is simply a moulding without the bead-and.-reel
decoration, and the sima is an undecorated torus.
This item must have come from just above the right-hand corner of the central doorway of
the cella fa~ade (cf. Baalbek II: Fig. 39; PI. 50).
Following is a detailed description of the eight different types of mouldings of the seven
items from the entablature of the portico and an assessment of their place within the architec-
157
Tel Aviv II (1984)
tural decoration of the East Roman Empire in general and of the Syro-Phoenician region in
particular.
(a) OvoIo.
The ovolo is commonly decorated with an alternating pattern of eggs-and-tongues, eggs-
and-darts, eggs-and-Ieaves, etc. The egg-and-dart pattern first appeLred in the 1st century
B.C.E. During the Flavian period its constituent elements became more pro-
nounced. This design, which produces an attractive interplay of light and shadow, continued
with minor variations throughout the Roman period (Weigand 1914:51; Weigand
1924-1925:169-170; von Blanckenhagen 1940:110). During the Hadrianic period, the egg-
and-leaf design (known in Classical and Hellenistic times) had a brief revival (Strong
1953:148; Fig.1), but during the Antonine period the egg-and-dart regained its popularity
(Kiihler 1939:70-71).
On our ·egg-and-dart pattern the shell enclosing the egg is emphasized by the cleavage
intervening between the two and by the distinct carving of the darts. The plasticity of the egg
within its broad shell and the strongly accented darts are typical of the Antonine period (von
Blanckenhagen 1940:110). The strict use of the dart pattern indicates a chronological range
within the above periods, but more likely the latter, since already during the Severan period
the dart was often replaced by a broader curved line, while later also leaves and flowers
alternated with the egg form (BaaIbek II: Figs. 143-144).
(b) The dentils
On Roman entab1atures the dentils are often combined with the egg-and-dart pattern in one
moulding. The occurance of ovolo, dentils and cyma reversa in a decorative sequence between
the frieze and the cornice allows us to fix both the typology and chronology of these elements
on the basis of their dimensional ratios. Thus, during the Classical period the dentils were
large and broad and the ratio between them and the ovolo was 2:1 (Weigand 1924-1925:174).
Over time a more moderate ratio prevailed and the sizes of these elements became almost the
same. During the Antonine and Severan periods the balance turned in favour of the ovolo, as
its elements were more strongly emphasized, both by increasing the space between the egg and
its shell and by making the separating device (dart, leaf, riband, etc.) more conspicuous. On
the other hand, the dentils changed very little over this period. The dentils on the architectural
items at Kedesh, as usual in the Syro-Phoencian region in the 2nd-3rd centuries C.E., lack
connecting links: this contrasts with the practice in Asia Minor where the dentils are always
linked by bars or rings (Strong 1953:148-150; Leon 1971:269).
(c) The cyma reversa
The cyma reversa appears twice on the cornice, once on the third moulding from below and
once on the fifth. The cyma reversa, which is identical on both mouldings and similar to that
on the southern jamb of the central doorway of the cella (see below), is of the so-called
"Syrian" type (Weickert 1913:105). The flower of the cyma reversa is fitted to the contour of
the central leaf and divided into two parts by a deep groove (Weigand 1924-1925: Figs. A5,
A6). A thin link is discernible between the two sides of the groove on almost every flower
(Weigand 1914: Figs. 33A, 34B). The separating flower has outspread wing-like petals, like
those decorating the Temple of Bacchus at Baalbek (Weigand 1924-1925: Fig. A7). The petals
gradually became more outspread until the flower was depicted as completely open (ibid.: Fig.
158
Fischer, Ovadiah and Roll: Roman Temple at Kedesh
AIO). It is notable that the fifth moulding forms the soffit of the corona; this is quite a rare
phenomenon and a good indication of the artists' ability to vary their ornamental patterns.
6
(d) M odillions-and-coffers
Modillions-and-coffers below the corona are integral elements of the traditional Roman
Corinthian style as known from the Augustan period onwards. This decoration originated in
the ornamental stone and stucco work of the Hellenistic period (Strong 1963:78-79).
The modillions-and-coffers of the Kedesh entablature are framed by an ovolo., as was
common in the Eastern Mediterranean from the 2nd century B.C.E. on, whereas in the
Roman West a cyma was used for this frame (Lyttelton 1974:92, 246-247, 264, 266). At
Baalbek a special type of modillion with two parallel grooves was developed (Weigand
1914:86), which was employed at other Syro-Phoenian sites as well (Baalbek II: Fig. 15c;
Krencker and Zschietzschmann 1938:Figs. 141,285). The modillions at Kedesh are decorated
with reliefs of acanthus leaves and palmettes , as was usual in Asia Minor in the 2nd century
C.E. (Lyttelton 1974:269). On the entablature fragments that have been preserved, the decora-
tion of the soffit of the modillions includes stylized flowers, branches, acanthus leaves and
S-form spirals (normal and reversed). Within the coffers appear stylized floral motifs in low
relief, including four-and five-petalled rosettes. Parallels are found in the Syro-Phoenician
region in general and at Baalbek in particular (Baalbek I: PIs. 90, 102, 103; Baalbek II: Figs.
29; PIs. 67-68; Weigand 1914:PI. 5:1). There are also thr-eeunique motifs that were previously
unknown on these mouldings: a cithara, a basin standing on a tripod and a crescent with a star
(PI. 33:2). Modillions with floral decoration, which are of Hellenistic origin, are well known in
the architecture of the Augustan period and common on many buildings of the 2nd-3rd
centuries C.E. in the Syro-Phoenician region (Weigand 1914:54;Pergamon II:8l).
(e) The cyma reversa
See moulding c, above.
(/) The corona
The transition from the soffit of the corona decorated with a cyma reversa to the corona
itself consists of a torus. The corona is carved with flutes, a common ornamentation of the
2nd-3rd centuries C.E. in the Eastern Mediterranean (Weigand 1914:84).The fluted corona is
also known in Rome during the Augustan period, where it continues throughout the Flavian
and Antonine periods, but afterwards a plain corona was preferred in Rome and in Asia
Minor (Strong 1953:Figs. 4-5), although on certain 2nd century C.E. buildings in Asia Minor
the fluted corona still appeared Pergamon IV: Pis. 34-35). This may have been due to Syrian
influence. At any rate, in Syria during the 2nd-3rd centuries C.E. the fluted corona was used
almost exclusively (e.g., at Baalbek, Gerasa, Bostra, Palmyra; Baalbek I: PI. 25; Butler 1903:
ill. 226; Weigand 1914: Fig. 9; Lyttelton 1974: Fig. 165), although sometimes the meander
pattern was used instead (Baalbek I: PI. 22). Fluting is connected with a development in the
architectural decorative style that came to prefer less elaborate elements, such as the "fluting
and acanthus capitals" (Ward-Perkins 1948:68-70). The concavity of the flutes produced an
6 A similar instance of the location of a cyma reversa moulding is found at the small temple in ancient
Philadelphia, modern Amman (Butler 1903:43).
159
Tel Aviv II (1984)
attractive three-dimensional effect allowing an interplay of light and shadow like that already
noted with the egg-and-dart moulding.
(g) The astragal
The corona is separated from the sima by an astragal, as characteristic of 2nd-3rd centuries
C.E. entablatures, especially in Syria. The astragal pattern of the entablature closely resem-
bles the astragals on the doorjamb and lintel of the fa9ade of the cella (see below).
(h) The sima
The sima, the uppermost element of the Roman cornice, is decorated with reliefs of acan-
thus leaves and palmettes, sometimes with intertwined lotus flowers in a recurring rhythmic
pattern. The fragmentary items that have survived at Kedesh indicate that the sima was
decorated with curled palmettes alternating with stylized split acanthus leaves, natural acan-
thus and lotus flowers or lotus calyces. Unfortunately, the fragmentary nature of the architec-
tural items makes it difficult to reconstruct a continuous sequence of the sima, although it
probably consisted of these four motifs (Riegl 1893:252-257; Weigand 1914:87). This is a
more varied range of motifs than the "intermitierende Akanthusranke" customary in Roman
architecture and resembles those found at Baalbek ( Baalbek I: PI. 25; Baalbek II: Figs. 14,
19). Another similarity with Baalbek is the anthemion of the sima, which continues the
tradition of the acanthus-scroll motif (Riegl 1893:248-252; Weigand 1914:55; Weigand
1924-1925:87).
2. The pilaster capital
Among the capitals that decorated the temple at Kedesh is a Corinthian pilaster capital (PI.
35:2), found in front of the portico (PI. 33:1). This capital has several characteristics in
common with elements of the entablature. A number of criteria permit us to establish its
typological and chronological classification. The decorative elements clearly assign it to the
extensive series of Corinthian capitals of the Syro-Palestinian type made of local limestone
that can be roughly dated to the 2nd-3rd centuries C.E. (Watzinger 1935:99; Fischer 1979:
No. 260). The capital is carved on three of its four sides, and the pilaster that it crowned
probably stood at the northeast corner of the cella (see Fig. 3). The arrangement of the motifs
displays a certain rigidity of conception, as expressed by the placement of the acanthus leaves
of the first row as single and discrete units, without any sort of link between them. The leaves
of the second row, also unconnected units, project somewhat above those of the first; their
central vein merges with the mass of the calathus, which enhances the impression of their
growing naturally out of the latter. In the centre of the capital above the second row of leaves
there is an additional acanthus leaf. All the leaves are of the Syrian acanthus type, which is
common in the architectural decoration of the Roman East (Heilmeyer 1970:105; PIs. 25-28,
31-33; Fischer 1979:51-52). The lobes of these leaves are divided by numerous deep fissures,
the central lobe having five tips instead of the usual four. Apparently the five-tipped lobe
originated in the workshop of the Trajaneum at Pergamon and from there spread throughout
the Roman world in the second century C.E. (Heilmeyer 1970:90-97). The calyces growing
behind the second row of leaves lack cauliculi and are divided into two by two triangles, set
one above the other; the helices and volutes (the latter are not preserved) stem from the
calyces. The helices are narrow with a wide central channel and terminate in an emphasized
160
Fischer, Ovadiah and Roll: Roman Temple at Kedesh
knob. A large stemless flower occupies the centre of the moulded abacus.
The capital bears a number of drill marks. It should be noted that the use of the drill in
sculptural and architectural carvings is common from the Flavian period onwards, becoming
more frequent in the Antonine period and virtually standard during the Severan period.
Pilaster capitals similar to ours are found at Pergamon and Baalbek in architectural assem-
blages dated to the middle of the 2nd century C. E. (Heilmeyer 1970:92-93; PI. 36:1; Baa/bek
II: PI. 54). In the light of the above data, our pilaster capital can be assigned to the repertoire
of the capitals typical. of the Antonine period. It constitutes an important link in the
prolonged building stages of the temple complex at Kedesh.
3. The central entrance
Of the central doorway of the cella the following decorative architectural items are pre-
served: the southern doorjamb, the northern doorjamb, half the lintel and one of the two
consoles. To these should be added the two entablature fragments, one of which was described
in detail above (Item 8; PI. 35:I).
The southern doorjamb
The southern doorjamb (PI. 28:2) found in situ, is a monolith measuring 4.50 m. (height),
0.60 m. (width across front) and 1.03m. (depth). The outer face includes three fasciae, the first
separated from the second by an astragal and the second from the third by a cyma reversa; the
third fascia is framed by an anthem ion of stylized leaves, which is bordered by a cavetto with a
running pattern of stylized tendrils. The ratio of the widths of the fasciae is I:2:2, and the
outer moulding with the anthem ion comprises one quarter of the overall width of the jamb.
The general arrangement of the profiles and the ratio between them, along with the stepping
of the fasciae, the types of dividing mouldings and the nature of the outer moulding are
similar to those of Corinthian architraves in general and the architraves of second century
C.E. Syria architecture in particular (Weigand 1924-1925:83).
The astraga/ here is similar to the one on the cornice. It consists of a "long angular bead
pointed at the ends and connected by two short reels," a type "typical of the Antonine period"
(Strong 1953:148;see also Weigand 1924-1925:172-174; Fig. B4; Baa/bek II: Fig. 15a).
The cyma reversa is also similar to that of the cornice of the portico, including the "ball" at
the top of the central leaf. The "ball" motif occurs on the cyma reversa of architectural
elements at Caesarea and Beth-shan (unpublished) and various sites in Asia Minor (Weigand
1924-1925: Fig. AI-2; Lyttelton 1974:Figs. 184-185), whence it reached the Syro-Phoenician
region during the Roman period.
The anthem ion has series of stylized acanthus leaves alternatedly reversed (i.e. one stem up
and one stem down) and linked by a leafed tendril (Riegl 1893:252-257; Weigand 1914:87).
This style of depicting the linking leaves of the anthemion originated in Asia Minor in the
Hellenistic period (Lyttelton 1974:264) and became common from the Flavian period on-
wards. It should be noted that this type of anthemion is more commonly found as the
crowning moulding of the architrave along with the ovolo and as the decoration of the sima
(Lyttelton 1974: Pis. 212,220-222,227; Applebaum 1978:100-101; Figs. D-E).
The cavetto is decorated with a wavy, tendrilled stalk, partly naturalistic in depiction. This
type of cavetto, combined with an astraga/ and/ or ovolo frequently crowns the architraves of
161
Tel Aviv II (1984)
Asia Minor buildings dating to the 2nd century C.E. In Syria this motif appears as the
crowning moulding of the architrave instead of the anthem ion (Lyttelton 1974:Figs. 141, 143,
148;Baalbek II: Fig. 73). This cavetto design originated in the Greek painted pottery of the
Classical period (Rieg 1893: 168-169; Fig. 76). It is a common ornamentation on the sima of
several synagogues in the Galilee (Kohl and Watzinger 1916:116;Fig. 228), but it also occurs
as the crowning moulding on the architrave on some of the pagan buildings in Palestine, such
as the theatre in Caesarea (unpublished).
Our comparative analysis of the four mouldings of the southern doorjamb permits us to fix
its date between the second half of the 2nd century to the beginning of the 3rd century C.E.
The northern doorjamb
The northern jamb (which was found among the architectural items recovered in the exca-
vations in the cella) is basically similar to the southern one, except for its cyma reversa design,
which is identical to that of the lintel of the central doorway.
The lintel
The lintel of the central doorway is only partially preserved (PI. 35:3). It was discovered
where it had fallen into the portico with its upper part facing downward. This upper part, still
partly buried in the earth, was examined in the 19th century by Wilson.7The preserved length
of the fragment is 2.18m., the height of its far;:ade0.80 m., and the width of the soffit 0.41 m.
(this width corresponds to the inner width of the doorjamb). The far;:adeof the lintel includes
three fasciae, separated from the bottom upwards by an astragal and a cyma reversa. The
ratio of the widths of the three fasciae is I:2:2, although admittedly the exact width of the
damaged lower one is hard to determine. The dimensions of the mouldings are identical to
those of the doorjamb. The upper fascia is crowned by an anthemion of alternately arranged
stylized leaves similar to those of the anthem ion of the southern jamb, and the cavetto above
the anthem ion is identical to that of the jamb. The mouldings on the right-hand side of the
lintel turn downwards at a ninety degree angle to match up with those of the northern
doorjamb. This, it may be assumed, was repeated on the left-hand side of the lintel, thus
enabling us to reconstruct the entire decorative frame of the central doorway, although
apparently the designs of all three sides were not completly uniform. For example the cyma
reversa of the lintel is identical to that of the northern jamb but differs from that of the
southern jamb in that its central leaf, instead of being a double arch, is divided into three
unconnected elements, while the separating flower is V-shaped. This type of cyma reversa is
rare in Roman architectural decoration in general and particularly rare on architraves, lintels
or doorjambs.
As the architectural item bearing the frieze of the lintel is still partly buried in the ground,
we shall confine ourselves here to a brief description of its main decorative elements. The
frieze, which is set back in relation to the anthem ion, has three naturalistic acanthus leaves on
its right-hand side, along with scrolls of intertwining vine tendrils creating oval medallions
which enclose bunches of grapes and vine leaves, stalks with tendrils, individual bunches of
grapes, as well as figurative motifs such as a gazelle and a male head, perhaps a mask. All
7 Wilson
(SWP I:
228-229) made a small-scale probe at this spot, which prior to our work, was the only
excavation conducted in the temple complex.
162
Fischer, Ovadiah and Roll: Roman Temple at Kedesh
these are characteristic of the "peopled scrolls," a common decorative style in Roman art
(Toynbee and Ward-Perkins 1950). A cornice (PI. 35:1), parts of which were found in the
excavations.(see Item 8 of the entablature), ran above the frieze. Similar cornices are to be
found in other sites also (Baalbek II: Fig. 39; Pis. 25-26, 51-52; Krencker and Zschietzsch-
mann 1938: Figs. 103, 131,202).
The soffit of the lintel fragment bears two decorative panels. The outer panel is a square
framing a relief of a wreath of leaves intertwined with flowers and buds. The wreath is tied
with a "Hercules" knot, from which issue a short stalk and two leaves joined to an eight-
petalled rosette. The central panel, which is only partially preserved, bears an eagle with
outspread wings carved in low relief, although only the lower part of the eagle's body, its two
legs and its outspread left wing are visible. The wing feathers, in two layers, are schematically
depicted. A palm branch and a curling riband appear below the wing. This schematized
depiction is typical of the eagles carved on lintels in temples in Syria and Lebanon (Krencker
and Zschietzschmann 1938: Figs. 100-101, 200, 346-347; Baalbek II: Fig. 38). The eagle,
along with wreaths and ribands, is a common decorative motif on the lintels of synagogues in
the Galilee and the Golan, apparently due'to the "baroque" influence of the gentile architec-
ture in the region8(Kohl and Watzinger 1916:158; Ma'oz 1981).
Although the left part of the lintel is not preserved, its overall length can be calculated by
taking the distance from its middle (i.e. the centre of the eagle's body) to its right-hand end
(1.94 m.) and multiplying it by two, since the ruling principle of symmetry in Roman architec-
ture presumably applied here as well. This would make the overall length of the lintel 3.88 m.
The console
A decorated console (PI. 36:I), still partially buried in the ground, was discovered in the
portico area opposite the central doorway. The console has a massive base, partially worked,
that anchored it into the wall of the fa~ade and a decorated part that projected outward from
the wall. It measures 1.49 m. (length), 0.50 m. (width) and 0.60 m. (height). The decoration
includes spiral volutes, the upper volute larger than the lower one. A lace-like floral design of
rosettes and vine leaves is worked in the side area between the two volutes, evidently with the
aid of a drill. The plasticity of the acanthus leaves that mantle the front of the volutes testifies
to a high degree of carving skill. The console was originally fitted to one side of the lintel of
the central doorway. In its style it recalls the "baroque" character of decorative elements at
sites throughout the Roman Empire in general and particularly in the East (Krencker and
Zschietzschmann 1938: Fig. 144; Baalbek II: Fig. 39; PI. 50; see also Lyttelton 1974:229-239),
and is a good example of the "baroque" tastes of the temple builders at Kedesh.
The above-described doorjamb, lintel and console of the central entrance are typical of
monumental doorways of the Ionic type, which evolved during the Classical and Hellenistic
periods. This type was described in detail by Vitruvius (IV, 6,3), who emphasized the continu-
ous banding and matching up of the horizontal and vertical mouldings, and also referred to
the proportions between them (Busing-Kolbe 1978:82-83, 130-131; Gruben 1976: Figs.
8 It seems that the eagles on the lintels of the synagogues in the Galilee and the Golan are decorative
rather than symbolic. Moreover, it should be noted that there are no references in the Talmud to any
symbolic significance of the eagle.
163
Tel Aviv 11 (1984)
164-166; Strong 1963:82-83). The architectural material from Syria in general and Kedesh in
particular adheres to Vitruvian principles, but there are also divergencies. Although in its
overall design the central entrance of the Kedesh temple is similar to the doorways of other
Roman temples in the Syro-Phoenician region, its decorative elements have their own indi-
viduality, and we have been unable to trace any exact parallels.
4. The northern doorway
The front of the lintel is symmetrically decorated by an eagle with outspread triangularly
tipped wings (PIs. 29:I; 37:2). Although the feathers of the wings are blurred by weathering,
their stylized depiction is still apparent. The eagle's head is turned to the right, while its body
faces front. Its eye is indicated by a small, drilled hole, and in its well-preserved beak it holds
what may be a wreath, like that depicted at Dabburah (Ma'oz 1981:109-110). The legs are
partly obliterated, yet their feathers, as well as the stylized tail feathers, are still visible. The
carving is flat and lacking in plasticity. The bird stands on a concave object, probably
Jupiter's thunderbolt (fulmen) (Cumont 1927: Fig. I; XXXVIII:2; Avi-Yonah 1946:89; PI.
XXV:6). There is a rectangular base, perhaps an altar, under the fulmen. A wall painting in
the Casa dei Vettii in Pompeii shows the baby Hercules strangling two snakes in front of an
altar, on which stands an eagle (Curtius 1929: Fig. 97). The base andfulmen on the Kedesh
lintel occupy the lowerfascia, the eagle's legs and part of its tail the middlefascia and its body,
part of its tail and its wings the upperfascia, while the wing tips and the head protrude beyond
it into the torus and the cyma recta.
If a Jupiter's thunderbolt is indeed depicted here, it provides us with an invaluable clue as to
the identification of the deity to whom the temple was dedicated-namely Baalshamin, who is
identified in this case with Jupiter (See Section IV below).
Plant motifs are symmetrically arranged on either side of the eagle: on its left is a stylized
six-petalled rosette and on its its right a festoon of flowers and leaves (PI. 36:2). The festoon
with its tendency toward plasticity, is richer and more elaborate than the six-petalled rosette.
Modelled in the Hellenistic-Roman tradition, this type of festoon is a common architectural
decoration (Riegl 1893:248-251; Figs. 129-130; Weigand 1914:79-80; Toynbee 1953: PI.
XXVII). A similar combination of flowers and leaves may be seen in the festoons of the Ara
Pacis in Rome, which represents the classicist trend of the beginning of Imperial art (Toynbee
1953:90-91; PIs. XXV-XXVI; Kraus 1953:26). In the course of time this type of festoon
became more "baroque" in character, and this is the style in which it was carved at Kedesh.
The rosette to the eagle's left, with sharp-pointed leaves emerging from its calyx, is a
common type known in various types of Roman art; in architecture it appears as a decorative
motif in the coffers of the ceiling or geison.
5. The southern doorway
The entire southern doorway is preserved in situ (PI. 30:2). Its architectural elements are
identical to those of the northern doorway but some of them are decorated differently. The
upper and lower fasciae of the lintel are plain (PI. 29:2). The middle fascia is decorated with a
double garland of stylized leaves, with a six-petalled stylized rosette at its centre. (Drill holes
are visible in the spaces between the leaves and inside the rosette.) The leaves of the garland,
which are elongated and fleshy with a raised midrib (spina), point towards the centre of the
164
Fischer, Ovadiah and Roll: Roman Temple at Kedesh
lintel and are alternately arranged in twos and threes. At each end of the garland is a
plastically rendered double riband that "ties" it together and then undulates down freely.
These tied ends are carved to appear as if they are "suspended" from two projecting L-shaped
"hooks," each of which faces the opposite direction. Despite his efforts to produce an effect of
plasticity in the "suspension" of the garland, the artist did not succeed in breaking free from a
certain stiffness and immobility of rendition. This was probably due to the narrow width and
flat surface of the fascia, which inhibited any free plastic expression.
Garlands of this type are especially common on the soffiti of architraves; they also appear
occasionally on the mouldings that separate the various decorative elements of the Corinthian
order (Gruben 1976:371; Fig. 310; Baalbek I: PIs. 82,109-110). Such garlands with ribands
descending toward the doorjamb belong to a category of illusionist architectural decoration
that is found also on certain monumental tombs in Jerusalem dating to the end of the Second
Temple period (Kon 1947: PI. VII).
The cyma recta that crowns the lintel is carved in low relief with a spiralic scroll design
incorporating six-petalled rosettes. This decoration survives also on the left-hand upright
cyma recta of the lintel but was not preserved on the right-hand upright side of the lintel. This
scroll design is one of the most common decorative motifs in Roman architecture. It recalls
one of the patterns of ancient mosaic pavements (Avi-Yonah 1933:138, pattern BI; see also
AIEMA 1973: Nos. 221, 223; Ovadiah 1980: 109-110). The scroll is also found on doorways in
Syria and Transjordan (Butler 1903: ill. 431).
6. Conclusions
Our comparative analysis of the decorated architectural items from the Roman temple at
Kedesh points to a continuation of the Hellenistic tradition, as reflected by the use of the same
motifs and by the variety and richness of the combinations of decorative frames and mould-
ings (Lyttelton 1974:246-247; Weigand 1914:86-87; Kohl and Watzinger 1916: 170-171).
However, some divergencies from the Hellenistic legacy are observable in the Kedesh temple,
for example, the additional cyma reversa moulding between the corona and geison of the
cornice, which is lacking in the entablatures of most Syro-Phoenician sites, although in
various places in Asia Minor and North Africa there may be an additional moulding in the
form of an ovolo (Lyttelton 1974: Figs. 170, 184-185, 227; Caputo 1959: PI. 66; Pergamon
V,2: PI. V:2). The additional moulding of the Kedesh temple entablature may, of course, have
been due to the influence of the architecture of these regions.
Actually, the deviation at Kedesh from the norm of architectural decoration is not signifi-
cant. If we ignore the additional cyma reversa moulding, what we have, in the final analysis, is
the standard entablature of the Syro-Phoenician region. The entablature mouldings of the
portico, especially those of its cornice, are designed according to a canon developed in the
region during the middle of the 2nd century C.E. that prevailed during the second half of this
century and throughout the 3rd century C.E. (Weigand 1914:86-87; Weigand 1924-1925:86;
Kohl and Watzinger 1916:170-171; Lyttelton 1974:246-247). On the other hand, in Asia
Minor a greater freedom of design in the architectural items as well as a greater richness in the
repertoire of decorative elements is observable. These characteristics are equally rooted in the
Hellenistic tradition, whose influence began to be felt throughout the Roman empire from the
165
Tel Aviv II (1984)
2nd century C.E. onward (Lyttelton 1974:201-202; Strong 1953:134).The architectural items
from the Roman temple at Kedesh are typical of the Corinthian style prevalent during the 2nd
century C.E. The same decorative elements are consistently present at the sacred enclosure of
Baalbek, where this style found its most brilliant expression. Undoubtedly, Baalbek was the
centre from which these decorative motifs spread elsewhere and the artistic inspiration for
various other sites in the Syro-Phoenician region, including Kedesh in the Upper Galilee.
However, it would appear that although the builders and artists at Kedesh remained faithful
to the artistic concepts and principles prevailing in the Roman period, th.;y allowed them-
selves a certain latitude in the choice and execution of the decorative motifs. In this respect,
we are reminded of the comments of Lyttelton (1974:271): "We miss in Asia Minor the
touches of eccentricity and individuality which often characterize the architecture of Syria in
this period."
IV. IDENTIFICATION OF THE TEMPLE DEITY
In the past it was suggested that the temple at Kedesh was dedicated to the god Helios
(Wilson as cited in SWP /:228-229; Dalman 1914; Goodenough 1958:114). However, the
epigraphical and iconographical data, in our opinion, point towards a different identification.
These data include: (1) two Greek inscriptions, one of which was found close to the temple
and the other in secondary use in the wall of the nearby village police station at Qadis; (2) two
eagles with outspread wings, one carved on the soffit of the lintel of the central doorway (PI.
37:1) and the other on the front of the lintel of the northern doorway (PI. 37:2).
1. The
inscriptions
Wilson (as cited in SWP /:229) referred to the first inscription mentioned above in a brief
sentence, which is important to the discussion here because he found it himself and could
therefore be reasonably sure of its relevance to the temple. He wrote: "Close to the temple,
and evidently belonging to it, an altar with a Greek inscription was found, which I cannot
make out, but have copied and taken a squeeze of."
Subsequently the altar disappeared and when ClermontcGanneau (1903) wanted to trans-
cribe and translate the inscription, he had to rely on an outline sketch made on the site by
Wilson, which he found in the archives of the Palestine Exploration Fund. According to this
sketch, on one side of the altar was depicted the head of a bearded man covered by a veil and
surrounded by a radiated halo, along with an undecipherable Greek word. On the other side
of the altar, between two stylized trees, was incised a 13-line Greek inscription, with an
average of four to six letters in each line. Clermont-Ganneau found decipherment of this
inscription exceedingly difficult. As he writes: "After a few palaeographical corrections -
some obvious, others perhaps doubtful - have been made, the following is the reading which
I believe may be obtained:
(lEi!>uyiCll..... ~E1ttilltO~Zi]vro(v)ZrouillOU(a)vsElTlK(E)V,lYE'? IlTl(vo~)'A(p)tElltuiou TIt'.
To the holy God ..... Septimius Zenon, son of Zosimos, dedicated in the year ... the 18th
of the month Artemisios.
As noted above, the decipherment and restoration of this inscription is a most difficult
undertaking. However, as an examination of Wilson's sketch reveals, there can be no doubt
with regard to the reading of its first two words: (lEi!>UyiCll.
166
Fischer,Ovadiahand Roll:Roman Templeat Kedesh
The second inscription, which was incised on a stone discovered in secondary use in the
outer wall of the nearby police station, was published in the beginning of the 1920s by
McCown (1923). The inscription has seven lines, with an average of 15 letters per line. Its
reading, as proposed by McCown, was incomplete as far as the date and the name of the deity
are concerned. The inscription was subsequently transferred to the Museum of Beirut,
reexamined and republished by Mouterde (1923; 1925:355, n. 4; cf.
SEG VIII
1937:No.2; Alt
1933:100,n. 4; Applebaum 1975:127).The inscription reads:
[E'tou~] YJ-lo'. 0EOUayiou oupaviou cmvyev(E)ta
(sid)
E7toiTJoEVlitelE7ttJ-lEATJ'troV'Avviou NayM
Kat AtcreJ-lOEO~.1aJ-la'to~EuxfJv.
[In the year] 243, the syngeneia of the Holy Sky God made (it) under a vow,
under the care of Annios (son) of Nagda and Lisemseos (son) of Damas.
As pointed out by Mouterde (1925:355, n. 4), the year 243 in the iscription is according to
the era of Tyre (126/5 B.C.E.), which is 117/8 according to the Common Era. As regards the
"Holy Sky God", this obviously refers to the famous Baalshamin, who was an important god
in the Syro-Phoenician region during the Roman period (Cumont 1929:118-124; Teixidor
1977:26-34). Baalshamin is known from quite a number of inscriptions, discovered variously
in Palmyra (Seyrig 1933:246-252; 1949), Damascus (Mouterde 1925:354-356), Hossn Solei-
man (Dittenberger 1903:No. 262) and the Hauran (SourdeI1952:19-31, 98, n. 2).
Three of the temples erected in his honour are well known to us: one in Palmyra (Collart
and Vicari 1969), one in Dura-Europos on the Euphrates (Rostovtzeff a.o. 1939:284-309) and
one in Si'a (Secia) in southern Syria (Butler 1916:373-380).
The full title of the god in the second inscription helps us considerably in identifying the
deity of the first. There can be no doubt that the first two words - SEI!>uyiC!>- of Wilson's
inscription, as published by Clermont-Ganneau (1903:133), the reading of which is quite clear,
refer likewise to the Holy Sky God, namely Baalshamin. Moreover, the bearded head on the
reverse side of the altar may also be identified as Baalshamin on the strength of certain
Palmyran coins that depict this god in the same manner (Du Mesnil du Buisson 1962:320,
740, Nos. 24-26).
McCown (1923:113) already surmised that the original place ofthe inscription was "possibly
in the temple of which the fa~ade is still standing." This surmise was made quite independently
of any knowledge of Wilson's altar and its inscription, of which he was apparently unaware. It
is therefore reasonable to assume that both inscriptions originally came from the Kedesh
temple, which means that it was closely associated with the cult of Baalshamin - and
probably even dedicated to this deity.
2. The iconographic evidence
The two eagles that decorate the central and northern doorways strengthen the evidence of
the inscriptions. However, since their relative positioning on their respective doorways is
significantly different, we have accorded them separate treatment and evaluation.
The soffit of the lintel over the central doorway was decorated by an eagle with outspread
wings (PI. 37:1). An eagle in such a position is known at a number of temples in the Syro-
167
Tel Aviv II (1984)
Phoenician region. Thus, the eagle with outspread wings is found on the soffiti of the lintels of
the Temple of Bel in Palmyra (Seyrig, Amy and Will 1975:84, 182; PI. 45), Temple A at Ni1,J.a
(Krencker and Zschietzschmann 1938:108), the Temple of Apollo at el-Hadet (ibid.: Figs.
199-200), the Temple of Bacchus at Baalbek (Baalbek II: 20, 22; Fig. 38), on one of the
temples at Rahle (Krencker and Zschietzschmann 1938: Fig. 347; Collart and Vicari 1969: PI.
CIV: 1-2), over the eastern gate of th~ temenos at Hossn Soleiman (Krencker and Zschietzch-
mann 1938: Figs. 100-101) and in the Hauran (Dunand 1934: Nos. 37-38). Although the eagle
does not directly represent Baalshamin, its association with the sky and celestial symbolism is
indisputable, and there would seem to be great justification in Seyrig's claim (1971:371, n. 4)
that the eagle should be viewed as "simplement Ie symbole de la Providence celeste" (cf.
Ronzevalle 1911:14-15; Cumont 1927:163-167; SourdeI1952:29-30; Toynbee 1973:240-243).
Of far greater relevance is the eagle depicted on the front of the lintel of the northern
entrance to the cella (PI. 37:2). The fact that this eagle appears on the front of the lintel
increases the likelihood of its being directly linked to the god Baalshamin. As evidence, we
may cite the monumental lintel in the Temple of Baalshamin in Palmyra (Collart and Vicari
1969:162-164,209; PIs. XCVII, CI, CV, CVII), whose front bears a large eagle with outspread
wings flanked by additional motifs.
In
this case there can be no doubt that the eagle represents
the god Baalshamin, while the flanking motifs symbolize his attributes and two companions.
As an eagle with outspread wings and representing one of a triad, Baalshamin is known from
additional monuments in the Syro-Phoenician region./ These include an altar found in the
vicinity of Tyre (Cumont 1927:163-167), another altar found at 'Atil (Dunand 1934: No. 19), a
monolithic basalt model of a naos that originated in Si'a (ibid.: No. 30) and a restored relief
from 'Ire in the Hauran (Sourdel1952: PI. II: 1).
We are therefore of the opinion that the eagle with outspread wings on the front of the lintel
of the northern doorway at Kedesh likewise symbolizes Baalshamin, the Holy Sky God
worshipped by the people of Syria and Phoenicia.
V.SUMMARY
The archaeological research so far carried out at Kedesh has revealed a rich and elaborate
architectural complex consisting of a temenos with a temple of the prostylos hexastylos type
at its centre. The architectural decoration is in the Corinthian style with "baroque" character-
istics. By virtue of its layout and architectural decoration, this temple is an integral part of the
architectural and decorative context of the Roman East, especially of the Syro-Phoenician
region. The temple is the southernmost link in the chain of-Roman temples in this region.
The kraters carved on the far;:ade of the cella at either side of the side entrances and the
narrow channels leading into the internal niches were apparently connected with a libation
rite that sought to establish contact between the world of the living and the world of the dead
- an assumption that is strengthened by the existence of mausolea to the west of the complex
and the tombs to its north. The two eagles carved on the cella far;:ade also support this theory_
In addition to the types of architectural decoration, there are also dated inscriptions found
in the past as well as in the recent excavations that provide evidence for dating the complex.
Three Greek inscriptions, dated respectively to 117/8, 189/190 and 214/5 C.E. (Fischer,
Ovadiah and Roll 1982) show that the temple was in use during the 2nd and 3rd centuries C.E.
168
Fischer,Ovadiahand Roll: RomanTempleat Kedesh
It may therefore be assumed that construction of the temple complex began around the
beginning of the 2nd century and continued in stages over a century or so.
Both the epigraphical evidence and the two eagles with outspread wings carved on the
falfade permit us to assume that the temple was dedicated to the god Baalshamin. Although it
is still difficult to determine when the temple was abandoned, there are indications that it was
destroyed by an earthquake, possibly the one that struck the region on May 19, 363 C.E.
What distinguishes the temple at Kedesh and makes it unique in comparison to the remains
of other temples in Israel is its excellent state of preservation, the richness of its architectural
decoration and, especially, the distinctive nature of the cult that was practised in it.
REFERENCES
Abel, F.M. 1949. Les livres des Maccabees. Paris.
AIEMA 1973.Repertoire graphique du decor geometrique dans la mosai"queantique. Bulletin
de l'Association Internationale pour I'Etude de la Mosai'que Antique 4. Paris.
Alt, A. 1933. Anhang. PJb 29:98-103.
Applebaum, S. 1975. The Struggle for the Soil and the Revolt of 66-73 C.E. EII2:125-128.
(Hebrew).
Applebaum, S. 1978.The Roman Theatre of Scythopolis. Scripta Classica Israelica 4:77-105.
Avi-Yonah, M. 1933.Mosaic Pavements in Palestine. QDAP2:136-18!.
Avi-Yonah, M. 1946.Newly Discovered Latin and Greek Inscriptions. QDAP 12:84-102.
Baalbek I-II. Wiegand, T. ed. 1921-1923. Baalbek: Ergebnisse der Ausgrabungen und Unter-
suchungen in den Jahren 1898 bis 1905. Berlin.
Bieber, M. 1967. The Sculpture of the Hellenistic Age. (revised ed.). New York.
von Blanckenhagen, P.H. 1940. Flavische Architektur und ihre Dekoration. Berlin.
Busing-Kolbe, A. 1978. Fruhe griechische Turen. Jahrbuch des Deutschen Archiiologischen
Instituts 93:66-174.
Butler, H.C. 1903.Architecture and Other Arts. Publications of an American Archaeological
Expedition to Syria in 1899-1900, Part 2. New York.
Butler, H.C. 1916. Syria. Publications of the Princeton University Archaeological Expedition
to Syria in 1904-1905 and 1909. II.A,6. Leyden.
Caputo, G. 1959. II teatro di Sabratha e l'architectura teatrale ajricana. Roma.
Clermont-Ganneau, C. 1903. Archaeological and Epigraphic Notes on Palestine. PEFQST:
131-135.
Collart, P. and Vicari,
J.
1969. Le sanctuaire de Baalshamin
a
Palmyre, I. Rome.
Creswell, K.A.C. 1969. Early Muslim Architecture I, Part II. (2nd ed.). Oxford.
Cumont, F. 1927. Deux autels de Phenicie. Syria 8:163-168.
Cumont, F. 1929. Les religions orientales dans Ie paganisme romain. (4th ed.). Paris.
Curtius, L. 1929. Die Wandmalerei Pompejis. Leipzig.
Dalman, G. 1914. Kedes. PJb 10:46-47.
Dittenberger, W. 1903. Orientis graeci inscriptiones selectae. Leipzig.
169
Tel Aviv II (1984)
Du Mesnil du Buisson, R. 1962. Les tesseres et les monnaies de Palmyre. Paris.
Dunand, M. 1934. Le Musee de Soueida. Paris.
Eus. Onom. Klosterman, E. ed. 1904. Eusebius. Das Onomastikon der biblischen Ortsnamen.
Leipzig.
Fenner, F. 1905.In Obergalilaa. PJb 1:99-IIO.
Fischer, M. 1979. The Development of the Corinthian Capital in Palestinefrom its Beginning
until the Constantinian Period. (PhD. Thesis). Tel Aviv. (Hebrew).
Fischer, M., Ovadiah, A. and Roll, I. 1982. An Inscribed Altar from the Roman Temple at
Kadesh (Dpper Galilee). Zeitschrift fiir Papyrologie und Epigraphik 49: I55- I58.
Gawlikowski, M. and Pietrzyk owski, M. 1980. Les sculptures du temple de Baalshamin
it.
Palmyre. Syria 57:421-452.
Goodenough, E.R. 1958. Jewish Symbols in the Greco-Roman Period, VII. New York.
Guerin, V. 1880.Description geographique, historique et archeologique de la Palestine, Gali-
lee, II. Paris.
Gruben, G. 1976. Die Tempel der Griechen. Miinchen.
Hamilton, R.W. 1940. Excavations against the North Wall of Jerusalem, 1973-8. QDAP
10:1-54.
Hammond, P.C. 1980.New Evidence for the 4th Century A.C. Destruction of Petra. BASOR
238:65-67.
Heilmeyer, W.D. 1970. Korinthische Normalkapitelle. Heidelberg.
Hennessy, J.B. 1970. Preliminary Report on Excavations at the Damascus Gate Jerusalem,
1964-6. Levant 2:22-27.
Kahler, H. 1939. Die romischen Kapitelle des Rheingebietes. Berlin.
Kohl, H. and·Watzinger, C. 1916.Antike Synagogen in Galillia. Leipzig.
Kon, M. 1947. The Tombs of the Kings. Tel Aviv. (Hebrew).
Kraus, T. 1953. Die Ranken der Ara Pacis. Berlin.
Krencker, D. and Zschietzschmann, W. 1938. Romische Tempel in Syrien,
I-II.
Berlin.
Lawrence, A.W. 1962. Greek Architecture. Harmondsworth.
Leon, C.F. 1971. Die Bauornamentik des Trajansforums und ihre Stellung in der friih-und
mittelkaiserzeitlichen Architekturdekoration Roms. Wien.
Lyttelton, M. 1974. Baroque Architecture in Classical Antiquity. London.
Ma'oz A. 1981.The Art and Architecture of the Synagogues of the Golan. Apud Levine, L.I.
ed. Ancient Synagogues Revealed. Jerusalem:98-1 15.
Marasovic, J. and Marasovic, T. 1970. Diocletian Palace. Zagreb.
Marmardji, A.S. 1951. Textes geographiques arabes sur la Palestine. Paris.
Mallwitz, A. 1972. Olympia und seine Bauten. Darmstadt.
McCown, C.C. 1923. Epigraphic Gleanings, III: Inscriptions from Kades. AASOR
2-3:II3-II5.
Mouterde, R. 1923.L' inscription grecque de Kades Naphtali. Al-Machricq 21:623-625.
Mouterde, R. 1925. Inscriptions grecques relevees par l'Institut Fram;:ais de Damas. Syria
6:351-364.
Oeconomus, G.P. 1921. De Profusionum receptaculis sepulchralibus inde ab antiquitatis
temporibus usque ad nostrum fere aetatem usitatis. Athens.
170
Fischer, Ovadiah and Roll: Roman Temple at Kedesh
Ovadiah, A. 1980. Geometric and Floral Patterns in Ancient Mosaics. Rome.
Ovadiah, A., Fischer, M., Roll, I. and Solar, G. 1982.The Roman Temple at Qedesh in Upper
Galilee. Qdmoniot 15:121-125. (Hebrew).
Pergamon II. Bohn, R. 1885.Altertumer von Pergamon, II: Das Heiligtum der Athena Polias
Nikephoros. Berlin.
Pergamon IV. Bohn, R. 1896. Altertumer von Pergamon, IV: Die Theater-Terrasse. Berlin.
Pergamon V,1. Kawerau, G. and Wiegand, T. 1930. Altertumer von Pergamon, V,I: Die
Paliiste der Hochburg. Berlin.
Pergamon V,2. Stiller, H. 1895. Altertumer von Pergamon, V,2: Das Traianeum. Berlin.
PZen. Edgar, C.C. 1925.Zenon Papri. Catalogue general des antiquites egyptiennes du Musee
du Caire, I. Le Caire.
Renan, E. 1871. Mission en Phenicie. Paris.
Riegl, A. 1893. Stiljragen: Grundlegungen zu einer Geschichte der Ornamentik. Berlin.
Robertson, D.S. 1969. Greek and Roman Architecture. (2nd ed.). Cambridge.
Ronzevalle, S. 1911. Notes et etudes d'archeo1ogie orientale, No. 10: L'aigle funeraire en
Syrie, etude iconographique. Melanges de la Faculte Orientale, Beirut 5:1-62.
Russell, K.W. 1980. The Earthquake of May 19, A.D. 363. BASOR 238:47-64.
Rostovtzeff, M.1. a.o. eds. 1939. The Excavations at Dura-Europos: Preliminary Report of
the VIIth and VlIIth Seasons of Work, 1933-1934 and 1934-1935. New Haven.
SEG. Supplementum Epigraphicum Graecum.
Seyrig, H. 1933.Antiquites syriennes 13. Leculte de Bel et de Baalshamin. Syria 14:238-252.
Seyrig, H. 1949. Antiquites syriennes 41. Nouveaux monuments palmyreniens de Baalshamin.
Syria 26:29-35.
Seyrig, H. 1971. Le douteux aig1esolaire: Syria 48:371-373.
Seyrig, H., Amy, R. and Will, E. 1975. Le temple de Bel
a
Palmyre. Paris.
Sourdel, D. 1952. Les cultes du Hauran
a
lepoque romaine. Paris.
Strong, D. 1953. Late Hadrianic Architectural Ornament in Rome. Papers of the British
School at Rome 21:118-151.
Strong, D.E. 1963. Some Observations on the Early Roman Corinthian. Journal of Roman
Studies 53:73-84.
Teixidor, J. 1977. The Pagan God. Princeton.
Toynbee, J.M.C. 1953. The Ara Pacis Reconsidered and Historical Art in Roman Italy.
Proceedings of the British Academy 39:67-95.
Toynbee, J.M.C. 1973.Animals in Roman Life and Art. London.
Toynbee, J.M.C. and Ward-Perkins, J.B. 1950. Peopled Scrolls: A Hellenistic Motif in Impe-
rial Art. Papers of the British School at Rome 18:1-43.
Toynbee, J.M.C. and Ward-Perkins, J.B. 1956. The Shrine of St. Peter and the Vatican
Excavations. London.
de Vogue, M. 1864. Le temple de Jerusalem. Paris.
Ward-Perkins, J.B. 1948. Severan Art and Architecture at Leptis Magna. Journal of Roman
Studies 38:59-80.
Watzinger, C. 1935. Denkmiiler Paliistinas, II. Leipzig.
Weickert, A. 1913.Das lesbische Kymation. Munchen.
171
Tel Aviv II (1984)
Weigand, E. 1914. Baalbek und Rom, die romische Reichskunst in ihrer Entwickelung und
Differenzierung. Jahrbuch des Deutschen Archiiologischen lnstituts 29:37-91.
Weigand, E. 1924-1925. Baalbek, Datie);J.Ulgund kunstgeschichtliche Stellung seiner Bauten.
Jahrbuchfiir Kunstwissenschaft 2:77-99, 165-196.
Wilson, C.W. 1869. Remains of Tombs in Palestine. PEFQST :66-71.
172
