Butler, H.C. 1913a Ancient Architecture, Pp. 149-213 in Syria.

Publications of the
Princeton University Archaeological Expedition to Syria (Div. II, Part 3, Umm
Idj-Djimal, Leyden) 149-213.

Publications of the Princeton University Archaeological
Expeditions to Syria in 1904-1905 and 1909
DIVISION II
ANCIENT ARCHITECTURE IN SYRIA
BY
HOWARD CROSBY BUTLER
SECTION A
SOUTHERN SYRIA
I
PART 3
UMM IDJ-DJIMAL.
72. UMM IDJ-DJIM
A
L (THANT/A?)
LATE E. ]. BRILL
PUBLISHERS AND PRINTERS
LEYDEN - I9I3.
-

Printed by E. J. BRILL, - LEYDEN (Holland).
Abbreviations of Periodicals and Publications Frequently Mentioned.
A. A. E. S. Publications of an American Archaeological
Expedition to Syria in 1899-1900, 1, II, III, IV.
A. ]. A. Atw-ican Journal of Arclaeology.
Ann. Ep. DAnnee Epigrapltique.
B. C. :. Bulletin de Correspondance Hellnique.
C. 1. G. Corpus Inscl-iptiolZum Graecaru1l.
C. 1. L. Corpus Inscriptionum Latinaru1Z.
C. 1. S. Corpus Imcription1mz Semiticarutt.
S. C. Marquis de Vogue; La Syrie Cen t1-a Ie, Architecture
,Civile et Religieuse.
G. G. A. Gottingische Geleil1'te Ameigen.
H. Hermes.
1. G. R. Inscriptiones Graecae ad Res Romanas pertinentes.
1. S. O. G. Dittenberger: Orientis Graed Inscriptiones Se-
lectae.
]. A. Journal Asiatique.
J. K. D. A. I. Jahrbuch des Kaiserliclt Deutschen Archao
logisclun Instituts.
]. K. P. K. Jalzrbuclt tier Koniglielt Preuszischen Kunst
sam 11 ltmgel.
K. A. Strzygowski; Klein-Asien, ein Neuland der Kunst
geschichte.
M. N. D. P.-V. Mittheilungen 2md Nachrichten des Deut
schen Palastina- Vel-eins.
M. S. M. Dussaud and Mader; Mission dans les regions
desertiques de la Syrie moyenne.
P. A. Brunnow; Provincia Arabia.
P. E. F. Quartedy
,
Statement of tlte Palestine Exploration
Fund.
P. M. Guy Ie Strange; Palestine under the Moslems.
P. R. G. S. Proceedings of tlt Royal Geographieal Society.
R. A. Revue Arclzologique.
R. A. O. Clermont-Ganneau; Recueil d'Arclleologie Orientale.
R. B. Revue Biblique.
S. E. P. Conder; Survey oj Eastern Palestine.
V. A. S. Dussaud; Voyage Archiologique au Safa.
Z. G. E. Zeitscltrit del' Gesellscltaft fur Erdkunde zu Bedin.
Z. D. M. G. Zeitschrift del Deutschen Morgenlandischen
Gesellsclurft.
Z. D. P.-V. Zeitscltrit des Deutschen Palastina- Vereins.
Explanation of Ground Plans.
SCALE: 0.0025 M. = 1 M. except when otherwise indicated on the plan.
Walls standing to a height of 2 M. or more.
Fallen walls, or foundations.
material in situ.
Foundations only in situ, or top of low wall.
Conjectured walls.
Column standing to height of 2 M. or more.
" less than 2 M.
Conjectured column.
Columns and architrave in situ.
.:::.' :.
Columns and arch in situ.
r B
Bases in situ, arch fallen.
1::::'--:".
.. ::.
,:::T
-

r
u
I
@
Arch in situ.
Arch fallen .
Conjectured arch.
Opening high in standing wall.
" " fallen
Pavement.
Tunnel-vault.
Cross-vaul t.
Cistern.
Explanation of Elevations and Sections.
SCALE: 0.005 M.
I M. except when otherwise indicated in the drawings.
o
Conjectured.
Limestone.
Basalt.
Brick.
SCALE OF DETAILS: 5 cm.
I M. except when some other scale is given in drawing,
NOTE. It has not been possible to carry the above scheme into effect with absolute consistency; but it has been
applied in a large majority of the drawings. Departures from the scheme are made clear by the text.
'
UMM IDJ-DJIMAL
AND
SURROUNDINGS
FROM SURVEY BV
F.A.NORRIS. C.E.
tANUARV. 1905
SCALE - le ... - 60I.
C To,b
f BIM
_ M.u_ B t.
Dout .. o R
.jQ
o
TOME Of
01ELD08
'"HOI 1111'"

t
I
(1
,
MAP NO. 1.

".0.
o
l1EAThE ..N TOME
IgJ< . Nb. 000'
UMM IDJ-DJIMAL.
72. UMM IDJ-DJIM
A
L (THANTIA?)
Far out in the desert, in the midst of a rolling plain, beside the dry bed of an
ancient stream, there is a deserted city. The plain about it is not a waste of sand ,
its surface is composed of dry and exhausted soil, overgrown with grey lichen, thinly
sprinkled with parched desert plants, and strewn with rounded bits of black basalt,
from the size of an egg to the size of a man's head, which are no l onger black, as
they were when the peasant' s plough-share turned them over from time to time; for
Ill. 130. Ruins of Umm idj-Djima.. View from the Southeast.
the desert mosses have covered them with a lace-work of white and grey, so disguising
their real nature as to have led one traveller, at least, to mistake the plain for a bed
of limestone. The walls of the ancient deserted city, its half-ruined gates, the towers
and arches of its churches, the two and three-storey walls of its mansions, all of basalt,
rise black and forbi dding from the grey of the plai n. Many of the buildings have fallen
in ruins, but many others preserve their ancient form in such wonderful completeness,
that, to the traveller approaching them from across the plain (Ill . 1 30), or viewing
them with the aid of a feld glass from the nearer crests of the Djebel Iauran, the
Publications of the Princeton University Archaeological Expeditions to Syria, Div. II, Sec. A, Pt. 3. 20
Di vision II Section A Part 3
deserted ruin appears
l
ike a
l
iving city, a
l
l of b
l
ack, rising from a grey-w
h
ite sea.
T
h
is ancient ruined city
h
as
l
ong been ca
l
led by t
h
e Arabs U mm idj-Djimal, w
h
ic
h
,
being translated, is "Mot
h
er of Came
l
s". Certain Bib
l
e sc
h
olars
h
ave tried to identify
t
h
e place wit
h
Beth Gamul of t
h
e
Ol
d Testament.1 It is not defnite
l
y known w
h
at
t
h
e city was ca
l
led in Roman and ear
l
y C
h
ristian times; but " Mot
h
er of Camels" it
is now, and no name could ft it better, especia
ll
y in t
h
e springtime w
h
en t
h
e Bedawin,
wit
h h
undreds of breeding camels, pitc
h
t
h
eir tents around t
h
e walls of t
h
e city, and
t
h
e new-born camels are s
h
e
l
tered wit
h
in t
h
e ruins against t
h
e winds t
h
at b
l
ow from
Hermon's snow-capped peak.
U mm idj-Djima
l
was t
h
e centra
l
city, and t
h
e metropo
l
is, of t
h
e once t
h
ick
l
y
settled region t
h
at
l
ies sout
h
of t
h
e Djebe
l
Iauran, and w
h
ic
h
is sty
l
ed in t
h
ese pub
l
ications, t
h
e Sout
h
ern Iauran. It was certainly one of t
h
e
l
argest, if not by a
ll
means t
h
e
l
argest, o
f
t
h
e cities of t
h
e Roman province of Arabia, sout
h
of Bostra
and east of Philadelhia. T
h
e
l
ine betwixt Syria and Arabia today is not very
defnitely drawn in t
h
is
l
oca
l
ity; but, if Syria is taken to be part of t
h
e Turkis
h
Empire,
and Arabia to be t
h
e independent country in
h
abited so
l
e
l
y by t
h
e Bedawin, and free
from garrisons, soldiers, taxes and ot
h
er features of Imperia
l
government, t
h
en t
h
is is
Arabia stil
l
, as it was in t
h
e days w
h
en it was ru
l
ed from Rome and Byzantium.
As it stands today, wit
h
all its monuments of t
h
e past still recognizable, it is t
h
e pre
Is
l
amic C
h
ristian Arabic city par excelence, practica
ll
y untouc
h
ed by Roman or
G
reek
infuence, and unc
h
anged by t
h
e stream of Is
l
am t
h
at swept by it.
T
h
is ancient site is not difcult of access for t
h
ose prepared for desert trave
l
and
wi
ll
ing to entrust t
h
eir safety to t
h
e Bedawin; but t
h
e
l
ack of water p
l
aces restrictions
upon camping among its ruins. A number of exp
l
orers
h
ave reac
h
ed Umm idj-Djimal
for a few
h
ours at a time ; but no Europeans
h
ad encamped on t
h
e spot unti
l
J
anuary
1 905 w
h
en t
h
e Princeton Expedition pitc
h
ed its tents beside t
h
e o
l
d fortress for two
weeks, and were fo
l
lowed,
l
ater
i
n t
h
e same winter, by Miss Gertrude
L
owt
h
ian Be
ll
2
w
h
o spent a nig
h
t in camp among t
h
e ruins. Neit
h
er party sufered t
h
e slig
h
test in
convenIence from t
h
e Bedawin, a
l
t
h
oug
h
trave
ll
ing wit
h
out t
h
e usual Turkis
h
guard.
T
h
e Arabs w
h
o were encamped t
h
ere w
h
en t
h
e Princeton Expedition arrived, pointed
out to our muleteers a small poo
l
of water about two mi
l
es west of t
h
e ruins, and
l
eft t
h
e p
l
ace quiet
l
y next day, saying t
h
ere was not sufcient water for t
h
em and for
us too. Indeed I may add t
h
at unusually
h
eavy falls of snow upon t
h
e mountains of
t
h
e
Iauran, and unwonted s
h
owers in t
h
e p
l
ain, made t
h
e winter of 1 904-1 905 a
singularly opportune season for exp
l
oration in t
h
is region. I was told by Bedawin
s
h
ep
h
erds t
h
at t
h
e supp
l
y of water was more plentifu
l
t
h
ere t
h
an it
h
ad been wit
h
in
t
h
eir memory, and I could we
ll
be
l
ieve it w
h
en I returned to t
h
is loca
l
ity in 1 909.
During two weeks of uninterrupted work, we were ab
l
e t
h
oroug
hl
y to exp
l
ore t
h
e ruins.
A survey of t
h
e ancient city was made by Mr. Norris, in w
h
ic
h
t
h
e
l
ines of t
h
e wa
ll
s
and al
l
t
h
e n10re important bui
l
dings wit
h
in t
h
em, and a c
h
urc
h
and many tombs out
side t
h
e wa
l
ls, were
l
ocated. As a result, two maps are presented
h
erewit
h
, Map No. I,
w
h
ic
h
s
h
ows t
h
e city and its immediate surroundings, and Map No. 2 w
h
ic
h
gives t
h
e
city itse
l
f on a muc
h l
arger sca
l
e, and many of its buildings drawn to sca
l
e. T
h
e
ground p
l
ans of a large number of buildings were carefully measured, and t
h
ese
h
ave
t Jer. XLVIII, 23. 2 The Desert and the Sown, p. 73-77.
Umm idj-Dji mal ( Tha1Ztza ?)
bee inserted in the
l
arger map. A
l
arge collect
i
on of photographs was taken, and
measurements were secured for the pub
li
cat
i
on, and presentation
i
n deta
il
, of over th
i
rty
buildings. Includ
i
ng the 2 9
i
nscript
i
ons
i
n
G
reek and
L
at
i
n, and the 3
i
nscr
i
ptions
i
n
Nabataean that had been cop
i
ed and publ
i
shed by former exp
l
orers, the Princeton
Exped
i
t
i
on gathered 5
L
at
i
n, 276
G
reek, 3 I Nabataean, I pre-Is
l
amic Arab
i
c and 1 3
Safa't
i
c
i
nscript
i
ons, wh
i
ch appear i n
Di
v. III, A, 3 and
Di
v. IV, Sects. A and c,
res pecti ve
l
y .
U mm idj-Dj
i
ma
l
was probab
l
y the great c
i
ty of the desert described by the Arabs
to some of the ear
l
iest travellers
i
n the Jauran. It was frst reached
i
n 1 857 by
Cyri
l G
raham 1 who pub
li
shed a br
i
ef descr
i
pt
i
on of the ru
i
ns in the fol
l
ow
i
ng year.
In 1 86 1 -62 W. H. Wadd
i
ngton 2 copied severa
l i
nscriptions here, and, over a decade
l
ater,
i
n 1 875-76, Char
l
es M.
D
oughty 3 passed through the ruined c
i
ty. At about
the same t
i
me Se
l
ah Merr
i
l
l
4, then American Consu
l
at Jerusa
l
em, v
i
sited the s
i
te, and,
the year after,
i
n 1 877, W
ill
iam M. Thomson 5 reached
i
t. It was not unt
il
thirteen
years later,
i
n 1 890, that He
i
nr
i
ch Frauberger 6 came to U mm
i
dj-
D
jimal ; he was
fo
ll
owed in 1 893 by
G
. Robinson
L
ees 7 who has recent
l
y pub
l
ished anew the account
of h
i
s journey in Southern Bashan. The frst p
l
ans to be pub
l
ished of bu
il
dings in
the anc
i
ent c
i
ty were those of one church and of one of the city gates made by
G
. Schu
macher 8 in 1 894, and publ
i
shed w
i
th descr
i
pt
i
ve notes the fo
ll
owing year. In 1 90 I
Rene
D
ussaud 9 and Freder
i
c Macler made a hasty excurs
i
on to the p
l
ace, and cop
i
ed
severa
l i
nscr
i
pt
i
ons ; the Princeton Expedit
i
on fo
ll
owed
i
n 1 905, and M
i
ss Be
ll
conlp
l
eted
the list of vis
i
tors
i
n the same year.
I
t will be noted at the head of this chapter, that Than tza , with an
i
nterrogat
i
on
mark, appears as the anc
i
ent name of Umm
i
dj-Dj
i
ma
l
. My reasons for this tentat
i
ve
i
dentifcation are given in an append
i
x
10,
devoted to a descr
i
pt
i
on of the sect
i
on of
Trajan's great road between Bostra and Philadelphia, wh
i
ch I prepared for
D
iv. III,
S
ect. A. , Part 2. They need not be repeated here further than to say that the
i
dent
i

fcation is based upon a new

i
nterpretation of the read
i
ng of the number of m
il
es given
from po
i
nt to po
i
nt
i
n Roman numerals on the Tabula Peutingeriana, which are not
eas
il
y reconci
l
ed w
i
th the actua
l
fgures which appear upon the mi
l
estones that are stil
l
in place on Trajan's road ; further evidence be
i
ng deduced from the or.der
i
n wh
i
ch
the name Thainatha,
i
.e. Thantza of Peutinger's map, appears
i
n the Notztia Digni
tatum. A suggest
i
on
i
s perhaps to be found in an
i
nscr
i
ption discovered at Umm
i
dj
D
j
i
mal by the Expedition of 1 909, which ment
i
ons the a(AEu 8avo1Yvwv;
i
t is the
G
reek ha
l
f of a N abataean-
G
reek b
i
lingua
l
11. Cons
i
der
i
ng the two variants of the form
of the name wh
i
ch appear on the anc
i
ent map and in the Noltia, it
i
s not surprising
that both shou
l
d difer slight
l
y from the N abataean origina
l
and
i
ts
G
reek trans
li
terat
i
on.
1 Journal of the Royal Geographical Society, ' 58 (XX-VIII) pp. 226-263.
2 Wadd. inscs. 2057a-2068, nnd de Vogie, La Syrie Centrale, Inscriptions Smitiques, Paris, 1 868-77, pp. 1 20-1 23.
3 Travels in Arabia Deserta, Cambridge Univ. Press, 1 883, Vol. I, p. 11 .
Q.S. ' 76, pp. 5 I-5 5, East of Jordan, New-York, Scribner, 1 883, (new ed.) p. 86.
5 The Land and the Book (3 vols.), London, Nelson, 1 88 1-1 886, p. 509.
6 Globus, ' 93 (LXIII) pp. 1 0-1 3 and 1 67-1 72.
7 Geog. Jour. '95, pp. 1-26, Lie and Adventures beyond Jordan, New-York, Appletons, 1909, pp. 263-267.
8 M.N.D.P.-V.
' 95 (I) pp. 33-35. Z.D.P.-V. ' 97 (xx) pp. 65-227.
9 M.S.M. pp. (608; 682), 438.
10
III, A, 2, p. XIV.
11 Published by Dr. Littmann in Florilegium Melchior de Vogue, p. 76. III, A, 3 insc. 2381.
Division II Section A Part 3
EXTENT AND CONDITION OF THE RUINS. T
h
e ruins of t
h
e city itself cover an irre
gular space over 800 m. long, and from 300 to Sao m. wide (see Map No. I) all
enclosed by a wall built at diferent periods and now in varying
s
tates of dilapidation.
O
n al
l
sides of t
h
e city, to a distance of Sao m. from t
h
e walls, are t
h
e ruins of large
built tombs. All along t
h
e west side of t
h
e ruins, about 200 m. distant, is t
h
e dry
and nameless wadi in w
h
ic
h
t
h
e ruins of a bridge lie about opposite t
h
e middle of t
h
e
west wall, and w
h
ere t
h
e ruins of a masonry dam are to be seen about 400 meters
to the nort
h
west of t
h
e city. From t
h
is dam t
h
e remains of an aqueduct can be traced
to t
h
e nort
h
west angle of t
h
e city wall, along its nort
h
wall, and by t
h
e w
h
ole lengt
h
of t
h
e east wall. T
h
e main aqueduct, w
h
ic
h lies just below t
h
e surface t
h
roug
h
out
t
h
e greater part of its lengt
h
, was tapped at several points by branc
h
aqueducts t
h
at
carried t
h
e water into reservoirs in diferent parts of t
h
e city. From t
h
e west gate t
h
e
remains of a built road are traceable, leading nort
h
of west, in t
h
e direction of Iar
il-Ba
c
i1, as is s
h
own on Map No. I.
Wit
h
in t
h
e walls t
h
ere is a broad open space, like an irregularly s
h
aped common,
w
h
ic
h
extends from t
h
e main west gate sout
h
ward, a little to t
h
e west of t
h
e main
axis of t
h
e city, as far as t
h
e sout
h
wall. This open space is interrupted only by t
h
ree
or four single buildings, of public c
h
aracter ; but it is surrounded by close set ranks
of buildings of all kinds, between w
h
ic
h
open the narrow streets, leading east and west
to t
h
e outer walls of t
h
e city. All t
h
e space between t
h
e open common and t
h
e walls
is flled wit
h
confused masses of ruins. Many buildings are in a remarkable state of
preservation, others are more or less destroyed, but t
h
e great mass of buildings, con
sisting naturally of private
h
ouses,
h
ave been completely destroyed. It was possible
to measure t
h
e ground plans of all of t
h
e public buildings, t
h
e c
h
urc
h
es, t
h
e barracks,
t
h
e government
h
ouse, etc. , and a large number of t
h
e private residences of diferent
classes, and to secure data for making elevations, sections, and restorations of buildings
of almost every one of t
h
ese varieties. In Map No. I the bui
l
dings t
h
at were mea
sured are s
h
aded grey, and t
h
e groups of private
h
ouses, are numbered by Roman
numerals, t
h
e groups of w
h
ic
h
only t
h
e outlines were measured are left w
h
ite; and it
will be seen t
h
at t
h
ere still remain large areas wit
h
in t
h
e walls t
h
at
h
ave been left
blank in t
h
e map. T
h
ese are strewn wit
h
ruins most of w
h
ic
h
are too completely
destroyed for t
h
e making of satisfactory ground plans, yet I
h
ave no doubt t
h
at later
explorers could add muc
h
to t
h
e detail of t
h
e map by tracing t
h
e walls of ruined
h
ouses among t
h
e masses of debris. Map No. 2 gives t
h
e plan of t
h
e greater part
of t
h
e city on a
l
arger scale, wit
h
all t
h
e buildings t
h
at were measured by t
h
e Princeton
Expedition, drawn to scale. Many of t
h
ese buildings will appear separately on still
larger scale, wit
h
sections and restorations ; but a large number of t
h
e private
h
ouses
s
h
own
h
ere are
h
ardly of sufcient importance to require individual presentation, and
only suc
h
as represent a particular class, or
h
ave some feature of special interest, will
receive separate attention in t
h
is Part.
T
h
e ruins preserve ample remains for t
h
e study and restoration of t
h
e city walls
and gates, t
h
e reservoirs and aqueduct, a temple and details of t
h
e Nabataean period,
c
h
urc
h
es of t
h
ree diferent centuries,
h
ouses t
h
at probably represent every century from
t
h
e frst to t
h
e sevent
h
, and tombs of every period from t
h
e time of t
h
e Nabataean settle
ment to t
h
e fall of t
h
e C
h
ristian city. T
h
us, we
h
ave a sufcient variety of buildings,
and a
l
ong enoug
h
period of building activity, to render possible not only a recon-
Umm idj-Djimal ( Thant-a ?) 1
5
3
struction of t
h
e
h
istorical development of arc
h
itecture in t
h
is ancient city, but a resto
ration of an important C
h
ristian Arabic centre of civilization at t
h
e end of t
h
e sixt
h
century.
THE BUILDINGS. T
h
e arc
h
itectural forms illustrated in t
h
e ruins of Umm idj.Djiml,
i . e. t
h
e ground plans, t
h
e construction of t
h
e superstructures of t
h
e buildings, and t
h
eir
ornament, present somet
h
ing more t
h
an t
h
e common types t
h
at are familiar to us in
t
h
e numerous towns of t
h
e Jauran, and t
h
at are described in a general way in t
h
e
introduction to II, A, 2. It
h
as been said by more t
h
an one of t
h
e earlier visitors to
t
h
is ruined city t
h
at t
h
e style of t
h
e arc
h
itecture is typically Jauranian, and t
h
is is
true of t
h
e general principles of construction employed in its buildings. T
h
e girder
arc
h
, t
h
e corbel courses, and roofng slabs, all t
h
e principal details and constructive
principles t
h
at were developed in t
h
e purely lit
h
ic arc
h
itecture in' basalt t
h
roug
h
out
Sout
h
ern Syria, are exemplifed in an
h
undred diferent buildings
h
ere ; but t
h
ere are
to be observed, at t
h
e same time, a num ber of features, important in construction and
signifcant in ornament, w
h
ic
h
are not common in t
h
e arc
h
itecture of t
h
e lauran, w
h
en
viewed as a w
h
ole, and w
h
ic
h
add interest to t
h
e style as it was expressed in t
h
is
Arabic metropolis. Taken up in c
h
ronological order t
h
ese peculiarities appear frst in
t
h
e arc
h
itecture of t
h
e N abataean period of w
h
ic
h
t
h
e remains are scanty at best, but
sufcient for comparative purposes. Not
h
ing new or unexpected is found in t
h
e matter
of plan, and t
h
e actual stonework s
h
ows t
h
e same beauty and skillful workmans
h
ip
t
h
at are to be seen in ot
h
er centres of N abataean in
f
uence ; but t
h
e ornamental details
are plain in t
h
e extreme, severe indeed in comparison wit
h
t
h
ose ric
h
decorations w
h
ic
h
are most c
h
aracteristic of Nabataean arc
h
itecture in ot
h
er parts of t
h
e lauran. T
h
e
mouldings are
f
at and simple in profle, and always wit
h
out carved enric
h
ments. T
h
e
four-pointed capital wit
h
its exaggerated abacus and depressed ec
h
inus, seen at Petra,
Bora and st, is
h
ere ; but t
h
e foliate capital is entirely wanting.
T
h
e few buildings of t
h
e Roman period t
h
at remain s
h
ow none of t
h
e elaborate
ornamental features t
h
at are most expressive of t
h
e Imperial arc
h
itecture in t
h
e nort
h
ern
parts of t
h
e Province, and t
h
e scattered details of Roman buildings t
h
at peris
h
ed during
t
h
e early C
h
ristian period s
h
ow only delicate and complex moulded surfaces, but no
carving of any kind.
T
h
e c
h
urc
h
es present a far greater variety of ground plans and superstructures
t
h
an can be found in any ot
h
er part of t
h
e ljauran. T
h
e most typical of all t
h
e
c
h
urc
h
forms in t
h
e lauran, i.e. t
h
e t
h
ree-aisled basilica wit
h
transverse arc
h
es, does
not appear at all ; t
h
e long, single,
h
all-c
h
urc
h
, wit
h
broad transverse arc
h
es, w
h
ic
h
is
per
h
aps t
h
e oldest c
h
urc
h
plan in t
h
e region, is found in several diferent edifces ; t
h
e
t
h
ree-aisled basilica, wit
h longitudinal arc
h
es,
h
ig
h
side walls and a
f
at roof of stone
over all , - t
h
e commonest form of c
h
urc
h
plan in t
h
e sout
h
ern Iauran , occurs in
several examples ; but t
h
ere are ot
h
er forms w
h
ic
h
are eit
h
er rare or w
h
ol
l
y wanting
among t
h
e c
h
urc
h
es of t
h
e neig
h
bor
h
ood. T
h
ese are, frst, a plan like t
h
e c
h
urc
h
es
of t
h
e fft
h
century in Nort
h
ern Syria,
h
aving t
h
ree aisles wit
h longitudinal arc
h
es carried
on piers, and a semi-circular apse between side c
h
ambers, and a superstructure in w
h
ic
h
t
h
e side aisles were roofed wit
h
stone slabs and t
h
e n1ain aisle was covered wit
h
a
d
ouble pitc
h
ed roof of timber raised upon a clearstorey, - t
h
e plan and superstructure
in fact of t
h
e well known c
h
urc
h
at
alb
L
auze
h
in t
h
e region east of Antioc
h
;
secondly, a plan like t
h
e above, but wit
h
columns substituted for piers, and wit
h h
igh
1
5
4
Division II Section A Part 3
side-walls, fat roof of stone, an
d
no clearstorey ; and t
h
ird, t
h
e same general arrangement,
wit
h
a double pitc
h
ed roof of timber over t
h
e main aisle, set on t
h
e level of t
h
e stone
roofs of t
h
e side aisles. Columnar nart
h
exes, rare in t
h
e ot
h
er parts of t
h
e Baurn,
are found in a number of c
h
urc
h
es
h
ere and t
h
e rectangular presbyterium exists as
well as t
h
e semi-circular. T
h
e ornament
d
oes not di
f
er from that of ot
h
er c
h
urc
h
es
in t
h
e basalt region of Sout
h
ern Syria.
T
h
e type of domestic arc
h
itecture is t
h
e one common in all t
h
e towns of t
h
e
locality ; but t
h
e proportion of large and tall h
ouses is muc
h larger. T
h
e use of
columns, in some porticos and loggias, adds interest to t
h
e general appearance of t
h
e
h
ouses.
H
ouses wit
h
two large arc
h
ed storeys in front and four storeys of narrow
c
h
ambers, wit
h
out arc
h
es, be
h
ind, are common t
h
roug
h
out t
h
e ruins.
O
rnament of all kinds is meagre in t
h
e buildings of t
h
e C
h
ristian period as it
was in t
h
e Roman and N abataean periods. A few Roman mouldings of intricate pro
fle were employed as second-
h
and decoration in C
h
ristian buildings ; but t
h
e mouldings
of t
h
e C
h
ristian period are usually of simple rig
h
t-lined profle. T
h
ey are used as
string courses and impost mouldings, and, in a few cases, as door frames. T
h
e use
of t
h
e column introduced ot
h
er ornamental features. T
h
e Doric column, wit
h
plain
s
h
aft, and wit
h
or wit
h
out a base, is t
h
e most con1mon form used ; but t
h
e typical
Doric entablature does not appear, being replaced by a perfectly plain arc
h
itrave and
a simply moulded cornice. T
h
e Ionic order appears in t
h
e columns of t
h
e " Praeto
rium" and in t
h
e pilaster caps of t
h
e C
h
urc
h
of N umerianos. No Corint
h
ian types
were found in t
h
e ruins, t
h
oug
h
very plain campaniform capitals appear in a number
of places. T
h
e remains of mosaic pavements are to be seen in several of t
h
e c
h
urc
h
es ;
t
h
ese were in sinlple geometrical patterns of four colors. T
h
e subjects of moulded
ornament in plaster, and of painted decoration, in Umm idj-Djimal are discussed on
page 69 of Part 2, describing t
h
e arc
h
itecture of t
h
e Sout
h
ern Bauran.
HISTORY OF THE MONUMENTS: Certain deductions wit
h
regard to t
h
e
h
istory of
t
h
e city are to be drawn from t
h
e monuments of arc
h
itecture, wit
h
t
h
e aid of a fe

inscriptions t
h
at
h
ave to do wit
h
t
h
e buildings. T
h
e remains of no less t
h
an two
Nabataean temples, and one monumental tomb, besides a large number of Nabataean
stelae t
h
at belonged to ot
h
er tombs now destroyed, indicate t
h
at t
h
e site was an im
portant centre of Nabataean civilization as early as t
h
e frst century B. C.-A. D. A
gateway bearing an inscription of t
h
e Emperors Marcus Aurelius and Commodus, and
parts of an early wall , prove t
h
at t
h
e city was defended certainly before t
h
e end of t
h
e
second century ; t
h
e great reservoir near t
h
e centre of t
h
e town probably belonged
to t
h
e same epoc
h
. T
h
ere are no monuments t
h
at can be assigned definitely to t
h
e
t
h
ird century ; but t
h
e fourt
h
century is represented by t
h
e " Praetorium", and by a
c
h
urc
h
dated in t
h
e year 345. In t
h
e next century t
h
e barracks were built ; for t
h
oug
h
t
h
e inscription referring to t
h
e building of t
h
e castellunz is not in situ, t
h
ere is very
little doubt t
h
at t
h
e Dux Pelagios Antipatros, w
h
ose name is mentioned in it, and
w
h
o built t
h
e fortress at
I
ar il-B
c
il in 41 1-1 2, was also t
h
e builder of t
h
e barracks
h
ere. A large portion of t
h
e city as it stands today was erected in t
h
e sixt
h
century;
and t
h
ere is only slig
h
t evidence to s
h
ow t
h
at any building operations were carried
on
h
ere afterwards. T
h
e frst settlers were Arabs of t
h
e nation we
h
ave called N a
bataean, and t
h
e absorption of t
h
at nation into t
h
e Roman Empire seems to
h
ave
h
ad
little e
f
ect upon t
h
e racial c
h
aracteristics of t
h
e city ; for, t
h
oug
h L
atin and Greek be-
Umm idj-Djimal ( Thantia?) 1
5
5
came the languages of a majority of the i nscripti ons, the art that fourished under
Roman rule in other parts of Syri a, and even i n other parts of the Province of Arabia,
did not reach this distant frontier city. In the following centuries, under Roman po
litical, and Christi an religious, infuence, we fnd that the inhabitants in many instances
adopted Roman and Greek names; but no corresponding i nfuence is perceptible in the
art of the city which undoubtedly remained purely Arabic in blood until the end.
Monumental inscriptions i n Safa'tic script \ and an example of a pre-Islamic Arabi c
inscription, show how strong the Arabic national life was throughout the city's history.
Nabataean Archztecture.
As has been remarked above, the architecture of the Nabataean peri od at Umm
idj-Djimal has none of the rich or elaborate character of the same style as we fnd it
at st, and other centres of Nabataean art. The building which I believe to have
been a temple of this peri od is situated in the southwestern quarter of the town within
the west gate (see Map NO. 2), in the midst of a group of houses of a later ti me.
It i s very evident that the walls were utilized i n the Christian period without alteration,
being wrought into a complex of domestic buildings. The beautiful quality of the
stonework, and its high fnish, stand out in strong contrast to the walls that surround,
and almost hide, it. The temple faces the nort h. Its plan (Ill. 13 I ) i s prostyle, distyle
in-antis, the cella is 8 m. wide inside, and 6.80 m.
deep ; there are three portals in the front wall,
- a wide doorway between two narrow ones.
The anta-walls are returned 76 cm. toward the
two columns. The columns are no l onger in
situ; but their two bases lie in the courtyard
not lor. from their original position. The stone
work of the walls is of the best quality, well
cut and closely joined. The top of the wall
is fnished in a cornice of the simplest profle
(A in Ill. 13 I ) . The bases of the columns con
sist of a heavy lower, and a thin upper, torus,
without any dividing scotia, and a slender bead
8.
316
NABATAEAN
TEMPLE
A
B
.CALI: 10CMw 1M
Ill. 131.
just below the shaft. The capitals and their details are wanting. I t will be remembered
that there are three portals in the Nabataean facade of the temple-mosque at il-Um
ta
C
iyeh (Pt. 2, Ill. 69). It i s quite probable that there was originally a small inner
sanctuary here with a passage all around it, as was found in the two temples at st.
The cornice and the column-bases are substantially like corresponding details i n the
ruins of N abataean buildings at st. In a group of houses (VI on Map 2) , about 70 m.
to the north, several N abataean inscripti ons 2 were found together with an altar with
a bili ngual inscription 3 in Greek and Nabataean upon it, giving the name of the Na
bataean deity Dusares. The altar and some of the inscriptions undoubtedly were ori
ginally set up withi n the precinct of the temple.
1 Diy. IV, Sec. c. 2 Diy. IV, Sect. A. lnscs. 39, 40, 41 .
3 Diy. III, insc. 238. Div. IV, Sect. A, insc. 38.
Division II Section A Part 3
Th
e remains of anot
h
er Nabataean building, presumab
l
y a temp
l
e, were discovered
we
ll
to t
h
e nort
h
east, about
h
a
l
f way between t
h
e "Gate of Commodus" and t
h
e east
wa
ll
of t
h
e city.
O
n
l
y a few sections of t
h
e foundations are visib
l
e, wit
h
t
h
e bases
and capita
l
s of two co
l
umns and fragments of t
h
e base moulding of t
h
e wal
l
some of
w
h
ic
h
is in situ. T
h
ese fragments are s
h
own in I
ll
. 1 32. T
h
e bases of t
h
e columns (A)
5CALE:10CM1
are of t
h
e ordinary Attic form; but t
h
e torus
mou
l
dings
h
ave t
h
e fat profle t
h
at c
h
aracterizes
t
h
e bases of t
h
e smal
l
er of t
h
e two temp
l
es at st.
T
h
e capita
l
is of t
h
at pecu
l
iar variety, known on
l
y
FRAGMENTS-
at Petra, Bora and st, in w
h
ic
h
a great
l
y exag-
t
m
Ill. 132.
gerated abacus wit
h
concave sides is t
h
e main
feature of t
h
e capital. T
h
e i
l
lustration (A in
I
ll
. 132) s
h
ows t
h
is capita
l
wit
h
a round bead
mou
l
ding be
l
ow it, and a section of t
h
e s
h
aft
attac
h
ed to t
h
e capital. T
h
e base mou
l
ding (B)
is typica
l
in profle. No p
l
an of t
h
e ter pIe could
be made out in t
h
e ruins, wit
h
out removing much
debris ; but, judging by t
h
e sca
l
e of t
h
e details, t
h
is temple was
l
arger t
h
an t
h
e ot
h
er.
Not far from t
h
e site of t
h
is temp
l
e anot
h
er a
l
tar was found, - an altar of beautifu
l
form wit
h
interesting ornament -, wit
h
a Greek inscription 1 giving t
h
e name of
t
h
e Nabataean god So
l
mos. T
h
is probab
l
y be
l
onged to t
h
e temp
l
e; and, being
l
arge
and very
h
eavy, it was not removed far from its origina
l
p
l
ace.
Gates and Walls.
T
h
ere are six gates sti
ll
visib
l
e in t
h
e wa
l
ls of Umm idj-Djima
J
, two in t
h
e west
wal
l
, two in t
h
e sout
h
and two in t
h
e east.
O
nly one of t
h
ese, t
h
e more nort
h
erly
of t
h
e two gates in t
h
e west wall, w
h
ic
h
bears t
h
e inscription of Marcus Aurelius and
Commodus and w
h
ic
h
, for t
h
e sake of brevity, I
h
ave sty
l
ed on t
h
e map, t
h
e "Gate
of Com modus",
h
as any pretensions to monumental c
h
aracter.
O
ne of t
h
e sout
h
gates
was fanked by towers, and may
h
ave been similar to t
h
e
G
ate of Commodus, but it
is muc
h
ruined. T
h
e main east gate also
h
as towers, but is smaller and simp
l
er in de
sign. All t
h
e ot
h
ers are plain arc
h
ed openings in t
h
e wall , in many cases placed be
tween buildings built against t
h
e wa
l
l.
GATE OF COMMODUS: Date, 1 7 6-1 80 A.D. T
h
is was apparently t
h
e principal gate
of t
h
e city, situated in t
h
e nort
h
west quarter, at t
h
e end of a road from Iar il-Ba
c
il$
- t
h
e nearest ruin on Trajan's great road w
h
ic
h
was t
h
e n1ain artery of COlnmerce in
t
h
e
A
rabian province. T
h
e gate is a very simple structure (Il
l
. 133), consisting of
two towers, w
h
ic
h
project outside t
h
e walls, connected by two arc
h
es w
h
ic
h
spring
from piers p
l
aced against t
h
e opposite faces of t
h
e towers. T
h
e towers probab
l
y served
as guard-
h
ouses, and eac
h h
ad a doorway between t
h
e piers of t
h
e arc
h
es. T
h
e gate
was almost devoid of ornament. T
h
e caps of t
h
e piers
h
ave a rig
h
t-lined, but sa
l
ient,
prof
l
e; t
h
e arc
h
es were less t
h
an semi-circ
l
es, and
h
ad no mouldings. Ill. 1 34 s
h
ows
t
h
e present condition of t
h
e gate: one tower is well preserved,
h
aving
l
ost only one
1 Div. III, insc. 239.
Ill. 134. "Gate of Commodus"; View from the West.
Ill. I 35. East Gate; East ace.
Umm i dj-Dji mal ( Thantiar) 1
57
or two of its upper courses and its cornice. The two piers on that side are standing
with the spring-stones of the arches sti
ll
above them. The opposite tower preserves
on
l
y a third of its origina
l
height. The restoration ( I
ll
. 133) was easi
l
y made ; for a
ll
the materia
l l
ies as it fe
ll
. Pieces of the cornice, which is of simp
l
e, right-
l
ined prof
l
e,
CATEoF'COMMODVS
DATE: 17G-180A:
SECTIONA-B

.
Ill. 133.
l
ie on a
ll
sides, and the inscription 1 of Marcus
A
ure
l
ius and Com modus
l
ies between
the two towers ; there can be
l
itt
l
e doubt as to its origina
l
position.
EAST GATE: The towers of the East Gate (I
ll
. 135) project inside the wa
ll
s:
there is but one arch, and this is narrower than the arches of the other gate. The
photograph shows that on
l
y the arch has fa
ll
en, and
restoration was a very simp
l
e matter (I
ll
. 1 36).
This gate was bui
l
t in a wa
ll
of heavy, rough
l
y
quadrated masonry ; on
l
y the piers of the arch and
the arch itse
l
f were of fnished stone. This arch,
l
ike the others, is not a comp
l
ete semi-circ
l
e, and
the archivo
l
t is quite p
l
ain. The caps of the piers
are of right-
l
ined prof
l
e,
l
ike those of the other
gate, and I do not hesitate to assign it to the
same period, i.e. , the end of the second century.
O
THER GATES: The West Gate, south of the
Gate of Commodus, is a sma
ll
afair, on
l
y 1. 80 m.
wide, in a
l
ate, and more crude
l
y bui
l
t, portion
of the city wa
ll
. It had a
l
inte
l
that was probab
l
y
re
l
ieved by an arch of discharge. It seems to
have been high enough to have permitted the
EAT GATE
- ES
0.
Ill. 1 36.
passage of came
l
s.
O
utside this gate there are some very
l
ate bui
l
dings that were
never enc
l
osed with a wa
ll
, within it is an open space, from which streets
l
ead in a
ll
directions. The more wester
l
y of the gates in the south wa
ll
which I have ca
ll
ed the
t Div. III. insc. 232. Wadd. 2057 and P.A., pp. 289 and 3 1 7.
Publications of the Princeton University Archaeological Expeditions to Syria, Div. II, Sec. A, Pt. 3.
2 1
Division II Section A Part 3
Sout
h
west Gate was only an arc
h
ed opening between buildings t
h
at were built on t
h
e
wa
ll
. T
h
e principa
l
Sout
h
Gate is completely ruined ; but t
h
e remains of a tower are
to be seen on t
h
e east side of t
h
e opening, as is s
h
own on Map 2. T
h
e wall
h
ere
seems to
h
ave been t
h
e original second-century wall, and most of t
h
e good materia
l
used in its construction was carried away at a late period, leaving only t
h
e smaller blocks
and fl
l
ing. T
h
e more nort
h
erly of t
h
e gates in t
h
e east wall, w
h
ic
h
we may cal
l
t
h
e Nort
h
east Gate, is an arc
h
, also in a part of t
h
e origina
l
wall ; it joins t
h
e nort
h
side of t
h
e East C
h
urc
h
, t
h
e east end of w
h
ic
h
was built against t
h
e city wall . T
h
ere
were probably ot
h
er gates in t
h
e northern part of t
h
e town ; but we found no traces
of t
h
em.
WALLS: As I
h
ave said above, t
h
e lines of t
h
e origina
l
wa
l
ls are well wit
h
in
t
h
ose of t
h
e later wal
l
s at a number of points ; but, w
h
erever it
h
as been preserved,
t
h
e structure of t
h
e more ancient wall is of so muc
h
better quality, as masonry, t
h
at
it may be recognized at a glance and easily distinguis
h
ed from t
h
e later wall. T
h
e
older wall was about 1 ))0 m. t
h
ick at mid-
h
eig
h
t. It was built in even courses of
blocks of basalt of good size, quadrated but roug
h
faced. Nort
h
of t
h
e Gate of
Commodus its foundations, about 2 m. t
h
ick, may be traced ; sout
h
of t
h
at gate it
h
as
been rep
l
aced by t
h
e
l
ater wal
l
. East of t
h
e "Praetorium", and from t
h
at point to
t
h
e sout
h
west angle of t
h
e city, it is lost among t
h
e buildings j for t
h
is entire section
appears to
h
ave been taken down w
h
en t
h
e present west wall was built a considerable
distance to the west of t
h
e older wall. T
h
e line of the original wall on t
h
e sout
h
side of t
h
e city was not changed in the later wall ; but the remains of t
h
is end of t
h
e
city suggest eit
h
er t
h
at t
h
e larger blocks of t
h
e" older wall were carried away at some
late period w
h
en t
h
e city walls were dispensed wit
h
, leaving only t
h
e f
ll
ing, or t
h
at
a very poor wall replaced an older and better one.
O
n t
h
e east side of t
h
e city t
h
e
later walls were built well to t
h
e east of t
h
e older ones, nort
h
and sout
h
of t
h
e East
Gate j but t
h
e East Gate itself appears t9 be on t
h
e line of t
h
e origina
l
wall. Traces
of the o
l
der wall are to be seen in sections that were utilized as t
h
e rear walls of
buildings at several points. A piece of
h
eavy wall, about 2 m. t
h
ick, is still standing
to t
h
e sout
h
west of t
h
e East C
h
urc
h
, well wit
h
in t
h
e later wall, and anot
h
er section
of t
h
ick wall, running east and west, is to be traced in t
h
e ruins well to the east of
t
h
e Gate of Commodus j this may be a part of t
h
e original nort
h
wall of t
h
e city.
T
h
e later walls were built to include a num bel of buildings t
h
at
h
ad been erected
outside t
h
e original walls j consequently t
h
ey follow most irregular lines. T
h
e con
struction is not s6 good, and t
h
e walls are not so thick as t
h
e more ancient con
structions. A few sections are sufciently wel
l
preserved to s
h
ow w
h
at kind of wall
building t
h
ey presented.
O
ne of t
h
ese (Ill. 1 37) is a curved piece of wall outside t
h
e
East Gate, and to t
h
e sout
h
of it, at a point w
h
ere t
h
e later wa
l
l was broug
h
t into
t
h
e line of t
h
e older wall. As may be seen by t
h
e p
h
otograp
h
, the
l
ower
h
alf of t
h
e
wall, to a
h
eig
h
t of about 2 m. , is laid in roughly quadrated work about 1 . 20 m.
t
h
ick ; above t
h
is i s a n1etre of
h
eight laid i n
h
erring-bone fas
h
ion, and, above this,
a lig
h
t loose wall that may or may not be original.
I
h
ave con1e to t
h
e conclusion, after carefu
l
observation of t
h
e older and t
h
e later
wall and t
h
e barracks, t
h
at t
h
e original walls were out-grown in t
h
e t
h
ird or fourt
h
century, t
h
at t
h
e later walls were constructed at t
h
at time to increase t
h
e area enclosed,
and t
h
at the old walls were destroyed w
h
erever t
h
ey no longer served their propose.
Ill. 137. Secti on of Later Wall on East Side of the City.
Ill . 138. The great Reservoir; View from the Southeast.
Umm idj-Dji mal ( Thantia ?) 1
5 9
It a
l
so seems probable t
h
at t
h
ese walls, old and new, were abandoned after t
h
e build
ing of t
h
e barracks or fortress, and t
h
e introduction of a garrison, early in t
h
e fft
h
century. T
h
e walls were once more preyed upon for building material for a century,
or a little more, unti
l
t
h
e garrison was wit
h
drawn, and t
h
e barracks were converted
into a monastery. T
h
e walls t
h
en existing were repaired w
h
erever necessary, and some
very crude walls were built to take t
h
e place of suc
h
sections as
h
ad been taken away
for building material during t
h
e period w
h
ile t
h
e walls were deemed unnecessary.
Reservozr and Aqueduct.
T
h
e water supply of Umm idj-Djiml was well provided for. Needless to say, t
h
e
dry, sandy, and pebbly bottom of t
h
e wadi west of t
h
e city was, in ancient times, t
h
e
bed of a perennial stream. T
h
e dam above t
h
e city retained t
h
e water for distribution,
t
h
roug
h
t
h
e aqueduct, into t
h
e various reservoirs. T
h
ere were no less t
h
an four reser
voirs fed by t
h
e aqueduct, beside a number of independent birke
h
s, or cisterns, t
h
at
were replenis
h
ed wit
h
rain water from t
h
e roofs of buildings. It is an open question
whet
h
er t
h
e reservoirs and cisterns were built of necessity, for t
h
e conservation of water
during an annua
l
dry season ; but it is certain t
h
at t
h
ey were made for convenience ;
for t
h
ey are so planned t
h
at every group of residences
h
ad its supply of water near
at
h
and. T
h
e great reservoir, 40 m. by 30 m. square, is situated a little east of t
h
e
centre of t
h
e town. It was partly excavated in t
h
e solid rock, and, like all t
h
e ot
h
er
reservoirs and cisterns, undoubtedly served as a quarry for building material w
h
ile in
process of construction ; for t
h
e entire city seems to be underlaid wit
h
a bed of basalt
rock from 3 to 4 metres below t
h
e surface. T
h
e p
h
otograp
h
(Ill . 138) s
h
ows t
h
at
t
h
e bottom of t
h
e resevoir, and t
h
e
l
ower parts of its side walls, are of t
h
e natural
basalt rock. T
h
e upper parts of t
h
e side walls were built in courses of draug
h
ted
masonry w
hi
c
h
is an indication t
h
at t
h
e reservoir was built in t
h
e t
h
ird century or
earlier, and probably under military infuence ; for t
h
is kind of masonry was commonly
employed in military arc
h
itecture during t
h
e period of strong Roman infuence in Syria.
T
h
e walls were covered wit
h
t
h
ick water-tig
h
t cement, i. e. opus szgnznum, and a
fig
h
t of steps led down to t
h
e bottom of t
h
e reservoir at its sout
h
west angle. T
h
e
ground level at t
h
e top of t
h
e wall falls away on all sides to prevent t
h
e wash from
t
h
e city streets coming into t
h
e reservoir, and t
h
e top of t
h
e wall
h
as a coping of
fat and well ftted slabs, two metres wide, furt
h
er to insure cleanliness. T
h
e mound
of sand outside t
h
e sout
h
west angle of t
h
e reservoir is undoubtedly t
h
e remains of
sediment taken from time to time from t
h
e bottom of t
h
e great reservoir. T
h
e branc
h
conduit leading to t
h
is reservoir from t
h
e main aqueduct was sunk just below t
h
e
surface; it was paved,
h
ad side walls, and was roofed wit
h
slabs. It was completely
h
idden w
h
en we frst visited
U
mm idj-Djimal in 1905 ; but an attempt to clean it out
and restore it
h
ad been made w
h
en I saw it in 1909. T
h
is work, as I was informed
by t
h
e Bedawin encamped
h
ere,
h
ad been undertaken, a few mont
h
s before, by Druses
from t
h
e mountain, and t
h
e Arabs believe t
h
at t
h
ese mountain people are about to
make an efort to settle in t
h
e deserted city ; t
h
is would be possible only in case water
can be broug
h
t to t
h
e ruins; and, since t
h
e wadi w
h
ic
h
gave t
h
e city its original water
supply is now perennially dry, it will be necessary to cut an aqueduct from some point
in t
h
is nameless wadi to a point on t
h
e WadI Butm a few miles to t
h
e nort
h
; for
160 Division II Section A Part 3
t
h
at wadi becomes a rus
h
ing stream once a year, for a mont
h
or
l
ess. We found it
fu
ll
of water in January 1904; but abso
l
ute
l
y dry in llarc
h
1909. T
h
e ot
h
er reser
voirs and cisterns are far sma
ll
er ; a few of t
h
em are rectangu
l
ar in form, but t
h
e
majority are ova
l
. A
ll
were cut in t
h
e natura
l
rock and were wa
ll
ed up to t
h
e sur
rounding
l
eve
l
and provided wit
h
a coping. Many of t
h
e sma
ll
er rectangu
l
ar cisterns
were arc
h
ed over and given a covering of stone s
l
abs ; in ot
h
ers, piers wit
h
bean1s
from one to anot
h
er carried t
h
e stone covering. T
h
e wa
ll
s of t
h
e cisterns t
h
at are
st[
'
l
protected by a roof sti
ll
.
preserve t
h
eir cement
l
inings in exce
ll
ent condition. It is
pr bab
l
e t
h
at many of t
h
ese covered cisterns are sti
ll
intact and t
h
at t
h
ey remain to
be discovered. Many of t
h
ose t
h
at are known are ferti
l
e sources of inscriptions ; for,
w en t
h
ey were bui
l
t, t
h
e ste
l
ae of t
h
e tombs of past generations were broug
h
t In
from cemeteries outside t
h
e wa
ll
s and used as s
l
abs for t
h
e covering.
From its source at t
h
e dam in t
h
e wadi to a point near t
h
e nort
h
east ang
l
e of
t
h
e city wa
ll
, t
h
e aqueduct is buried from I to 2 m. deep in t
h
e soi
l
. Its course na
tura
ll
y fo
ll
ows t
h
e contour
l
eve
l
s. Below t
h
e point near t
h
e nort
h
east ang
l
e of t
h
e
city, it appears as a paved and wa
ll
ed conduit wit
h
its covering of s
l
abs on a
l
eve
l
wit
h
t
h
e soi
l
. T
h
is section is s
h
own by doub
l
e
l
ines on Map NO. 2. A branc
h
of
t
h
is section is carried t
h
roug
h
t
h
e wa
ll
, but its termination was not discovered. Be
l
ow
t
h
e water-gate, t
h
at is, t
h
e opening in t
h
e wa
ll
w
h
ere t
h
e conduit to t
h
e great reservoir
is carried t
h
roug
h
, t
h
e aqueduct again disappears entire
l
y be
l
ow t
h
e
l
eve
l
of t
h
e soi
l
to reappear at a point about 80 m. from its termination. Here t
h
e ground fa
ll
s away,
and t
h
e conduit is raised upon a so
l
id substructure ( I
ll
. 139) over a metre
h
ig
h
. T
h
e
p
h
otograp
h
presented
h
erewit
h
s
h
ows t
h
e e
l
evated portion of t
h
e aqueduct in a very
ruinous state. At t
h
e end of t
h
e aqueduct t
h
ere is a deep rectangu
l
ar cistern wit
h
a
ruined bui
l
ding beside it. T
h
e fa
ll h
ere is sufcient
l
y great to
h
ave made possib
l
e t
h
e
introduction of a mi
ll
-w
h
ee
l
,
l
ike t
h
ose
l
ong in use in ot
h
er parts of Syria; but it is
difcu
l
t to know
h
ow ear
l
y t
h
is sort of mi
ll
was emp
l
oyed in t
h
e Arabian province.
T
h
ere are water mi
ll
s in t
h
e Ledja w
h
ic
h
appear to be of considerab
l
e antiquity ; many
of t
h
em are in complete ruins and seem not to
h
ave been in use since t
h
e region
h
ad a sett
l
ed Mos
l
em popu
l
ation,
l
ong before t
h
e coming of t
h
e Druses.
Civil Architecture.
"PRAETORIUM" : Date 37 1 A. D. T
h
e bui
l
ding w
h
ic
h
I
h
ave ca
ll
ed t
h
e Praetorium
( I
ll
. 140) is situated near t
h
e west wa
ll
, sout
h
of t
h
e Gate of Commodus, and beside
t
h
e broad, open, space t
h
at extends t
h
roug
h
t
h
e midd
l
e of t
h
e city. Its major axis
l
ies east and west. Sout
h
and east of it is a great courtyard, origina
ll
y enc
l
osed wit
h

in a
h
ig
h
wa
ll
. In t
h
e sout
h
west ang
l
e of t
h
is yard, and extending for s
o
me distance
a
l
ong its west wa
ll
, is a row of residences of t
h
e same period, and in t
h
e same genera
l
sty
l
e, as t
h
e main bui
l
ding.
Plan: T
h
e p
l
an of t
h
e Praetorium ( I
ll
. 14 I) is a para
ll
e
l
ogram divided
l
ongi
tudina
ll
y into two main sections ; on on t
h
e sout
h
subdivided into t
h
ree main apart
ments (G.P. and R.), and one on t
h
e nort
h h
aving fve sma
ll
er divisions ( S.T.V. W.X. ) .
T
h
e midd
l
e division (G) of t
h
e t
h
ree main apartments, consists of a square atrium
Doricum, according to t
h
e description of
V
itruvius, i.e. , a square court wa
ll
ed on a
ll
sides and
h
aving four co
l
umns, at t
h
e ang
l
es of a square comp
l
uvium, carrying a
Il l . 1 39. Elevated Section of the Aqueduct, Outside East \Vall ; Vi ew from the South.
Ill . 140. The " Praetori um"; View from the Southeast.
Umm i dj-Dji mal ( Thantia ?) 1 6r
slanting roof between t
h
e columns and t
h
e walls. T
h
e sout
h
wall of t
h
e atrium con
tains t
h
e principal entrance to t
h
e building and two large windows above two rect
angular nic
h
es on t
h
e inside. In t
h
e east wall of t
h
e atrium, to t
h
e rig
h
t as one
enters, is a single doorway between two pilasters w
h
ic
h
correspond to t
h
e columns of
t
h
e impluvium. T
h
is doorway opens into ( p) one of t
h
e t
h
ree main divisions of t
h
e
building. T
h
is division consists of cruciform c
h
amber, t
h
e four arms of t
h
e cross being
1
'I
I
15. 9

I.OS
98
I
!
1$:0
: 90
1
537

I
1,01
j

b\80
:: : :

:4o
578
1 !u
PRAETORlv
DATE: 371 AD (?)
!
Ill. 1 41 .
o
equal, and covered by tunnel vaults w
h
ic
h
meet in a square now open to t
h
e sky.
T
h
e two square spaces between t
h
e arms of t
h
e cross and t
h
e east wall of t
h
e build
ing are flled by two very small square c
h
ambers (p and x) ; t
h
e spaces at t
h
e oppo
site side are occupied by a passage ( Y. Y. ) w
h
ic
h
extends across t
h
e western arm of
t
h
e cruciform c
h
anlber, opening out of doors on t
h
e sout
h
, turning t
h
e angle of t
h
e
nort
h
ern arm, and leading into a c
h
amber ( X) in t
h
e nort
h
east angle of t
h
e building.
T
h
is passage was a private entrance of some sort.
O
n t
h
e west side of t
h
e atrium
is a large
h
all ( R) wit
h
two broad transverse arc
h
es, opening upon t
h
e atrium by
means of t
h
ree portals, - a large portal between two narrower ones -, separated by
pilasters. T
h
e frst bay of t
h
is
h
all
h
as a small doorway and a window in its sout
h
162 Division II Secti on A Part 3
end ; t
h
e second and t
h
ird bays
h
ave windows to t
h
e sout
h
, t
h
e west end of t
h
e
h
a
ll
h
as t
h
ree
l
arge windows, and t
h
e t
h
ird bay
h
as a
l
so a doorway in t
h
e nort
h
. T
h
is
section of t
h
e bui
l
ding
h
ad two storeys. In t
h
e ot
h
er
l
ongitudina
l
division of t
h
e bui
l
d
ing, (S) is a square c
h
arber wit
h
a transverse arc
h
, connecting wit
h
t
h
e westernmost
bay of t
h
e great
h
a
ll
, (T) is a simi
l
ar c
h
amber adjoining it on t
h
e east, and opening
upon t
h
e atrium by a doorway in t
h
e nort
h
west ang
l
e, (V) is a sma
ll
er c
h
amber, arc
h
ed
l
ike t
h
e ot
h
ers, wit
h
its doorway in t
h
e midd
l
e of t
h
e nort
h
wa
ll
of t
h
e atrium, and
(W) is a c
h
amber corresponding to (T) , but
h
aving t
h
ree windows in its nort
h
wa
ll
w
h
ere t
h
e ot
h
er c
h
ambers
h
ave on
l
y one or two. T
h
is c
h
amber is not so deep as
t
h
e ot
h
ers ; for its front wa
ll
is set back to give room for t
h
e passage ( Y) mentioned
above. (X) is an ob
l
ong c
h
amber
l
ying nort
h
and sout
h
, not arc
h
ed ; but
h
aving two
storeys of corbe
l
-and-s
l
ab construction.
In its
l
ower storey t
h
is c
h
amber
h
as a row of six recesses,
l
ike cupboards, i n t
h
e
east wa
ll
, and was apparent
l
y a c
l
oak room, or a room for depositing arms ; it opens
into a minute c
h
amber in t
h
e nort
h
east ang
l
e of t
h
e cruciform c
h
amber, and
h
ad a
staircase on t
h
e west
l
eading up to a sma
ll
doorway over t
h
e passage ( Y), w
h
ere an
ot
h
er stair, above t
h
e passage,
l
ed up to t
h
e roof. An outer stair, corbe
ll
ed out from
t
h
e sout
h
wa
ll
,
l
eads up from t
h
e sout
h
east ang
l
e of t
h
e bui
l
ding to t
h
e roof above
t
h
e cruciform c
h
amber ; both stairs conducted to about t
h
e same point on t
h
e roof.
Superstructure. Upon t
h
e ground-p
l
an described above, wit
h
t
h
e aid of t
h
e parts
of t
h
e bui
l
ding t
h
at are sti
ll
in situ, and by observation of t
h
e fa
ll
en parts, t
h
e aCCOlll
panying e
l
evations and sections (I
ll
. 142) were drawn. T
h
e p
h
otograp
h
given in I
ll
.
140 s
h
ows just
h
ow muc
h
of t
h
e sout
h
wa
ll
is standing, and t
h
e sout
h
e
l
evation, pre
sented in I
ll
. 142 , gives t
h
ese parts s
l
ig
h
t
l
y s
h
aded, w
h
i)e t
h
e restore
d
portions are
l
eft w
h
ite. It wi
ll
be seen t
h
at t
h
e c
h
ief entrance to t
h
e atrium, wit
h
a window on
eit
h
er side of it, and t
h
e
l
itt
l
e side entrance of t
h
e great
h
a
ll
on one side, and t
h
e
doorway of t
h
e passage (Y) on t
h
e ot
h
er, give symmetry to t
h
e midd
l
e of t
h
e faade.
East of t
h
is are t
h
e
l
arge window in t
h
e end of t
h
e sout
h
arm of t
h
e cruciform c
h
amber
and t
h
e outside stair.
O
n t
h
e west of t
h
e midd
l
e of t
h
e faade, stands t
h
e two-storey
portion of t
h
e bui
l
ding, t
h
e entire
h
eig
h
t of w
h
ic
h
is preserved at its sout
h
east ang
l
e,
w
h
ere a
l
ofty fragment, embracing sections of wa
ll
on bot
h
sides of t
h
e ang
l
e, and
carrying a smal
l
piece of t
h
e raking cornice, supp
l
ies data for t
h
e restoration of t
h
e
sout
h
faade and t
h
e transverse section (E-F). T
h
e p
h
otograp
h
( I
ll
. 140) s
h
ows one
comp
l
ete window, wit
h
its
h
ood, in t
h
is section of t
h
e sout
h
wa
ll
, and t
h
e upturned
si
ll
of t
h
e window above it. T
h
ere can be no doubt t
h
at t
h
ere were t
h
ree windows
in t
h
e
l
ower storey on t
h
is side, - two beside t
h
e one s
h
own in p
h
otograp
h
- and it
is equa
ll
y beyond question t
h
at t
h
ere were t
h
ree simi
l
ar windows in t
h
e upper storey.
T
h
e t
h
ree windows in t
h
e west end of t
h
e
h
a
ll
are a
l
so to be seen in t
h
e p
h
otograp
h
.
T
h
e
l
ongitudina
l
section (A-B) cuts t
h
roug
h
t
h
e great
h
a
ll
(R), t
h
e atrium (0) ,
and t
h
e tunne
l
-vau
l
ted cruciform c
h
amber (P) . T
h
e sections t
h
roug
h
t
h
e great
h
a
ll
and t
h
e atrium are se
l
f-exp
l
anatory in view of t
h
e description of t
h
e p
l
an a
l
ready given.
T
h
e cut t
h
roug
h
t
h
e cruciform c
h
amber s
h
ows t
h
e
h
eig
h
t and arrangement of t
h
e
tunne
l
vau
l
t springing from t
h
e crowns of t
h
e ot
h
ers, over t
h
e square intersection w
h
ic
h
is now open to t
h
e sky. T
h
e intersection cou
l
d not
h
ave been covered by a dome,
for t
h
e entire square is provided wit
h
an over
h
anging cornice,
l
eaving no p
l
ace for
pendentives, so t
h
at t
h
e c
h
oice must
l
ie between a tunne
l
vau
l
t and a c
l
oistered vau
l
t.
, -
.
U mm i dj - Dj i maJ ( Thanta ?)
_rlIDJ-DJlL

(
,

, - ]_][|\
)
>-
' . . . . . .

____|/'~ `
. r.-_. " ,
_ '
.'

. __
_
_
E
L
E
VA
TIO
N
LLJ l /
.L'i Cl
LLJ
Ill. 1 42,
.

Division I I Section A Part 3
T
h
e former was used in a tomb of simi
l
ar p
l
an at il-
A
ndedn 1 wit
h
its axis at rig
h
t
ang
l
es to t
h
e major axis of t
h
e cruciform room be
l
ow it ; t
h
e
l
atter appears in t
h
e
bat
h
s at Bo-ra, and was apparently used in t
h
e Praetorium at Mismiye
h
. 2
T
h
e cross .section (C-D) i s taken t
h
roug
h
t
h
e great
h
a
l
l (R) and t
h
e arc
h
ed c
h
amber (T)
beside it. T
h
e
h
eig
h
t of t
h
e piers of t
h
e arc
h
es of t
h
e
h
a
ll
is known. T
h
e arc
h
was, of course, a semi-circ
l
e, probab
l
y a
l
itt
l
e stilted : t
h
is, in execution, gives t
h
e
approximate
h
eig
h
t of t
h
e foor of t
h
e upper storey, w
h
ic
h
, judging from t
h
e ruins, was
composed of stone slabs. T
h
is
h
eig
h
t suits t
h
e
h
eig
h
t of t
h
e windows of t
h
e upper
storey w
h
ic
h
is a known quantity.
T
h
e ot
h
er cross section (E-F) is taken t
h
roug
h
t
h
e' atrium and t
h
e middle room (V)
on t
h
e nort
h
of it. It s
h
ows, at t
h
e
l
eft, t
h
e two-storey, gab
l
ed wa
ll
at t
h
e end of
t
h
e great
h
a
ll
, wit
h
t
h
e parts in situ and restored diferent
l
y represented, and, at t
h
e
rig
h
t, a section t
h
roug
h
t
h
e arc
h
ed room (V) , w
h
ic
h
is but one storey
h
ig
h
. T
h
e co
l
umns w
h
ic
h l
ie in t
h
e atrium are p
l
aced in position, and t
h
e roof above t
h
em
h
as
been restored. T
h
e entrance to t
h
e storey above t
h
e arc
h
ed
h
a
ll
must
h
ave been in
t
h
e restored wa
ll
j for t
h
e stairs t
h
at
l
ed to it are part
l
y preserved and are s
h
own in
a p
h
otograp
h
( I
l
l. 5 2 pub
l
is
h
ed in Part 2). T
h
ese stairs rose from t
h
e top of t
h
e sout
h
wa
ll
of t
h
e atrium; t
h
ey were approac
h
ed by a wa
l
k, corbe
l
led out like a balcony,
from t
h
e top of t
h
e wa
ll
, and
h
e
l
d in place by a parapet. T
h
e doorway could not
h
ave been in t
h
e centre of t
h
e wa
ll
un
l
ess it opened upon a
l
ong ba
l
cony, as I
h
ave
drawn it. T
h
e circu
l
ar window in t
h
e gable
h
as been restored from fragments of its
framing stones found just wit
h
in t
h
e middle doorway of t
h
e great
h
all. T
h
ese stones
of irregu
l
ar s
h
ape are quite similar to t
h
e framing stones of a gab
l
e window sti
ll
in
place at Djemardn, just nort
h
of Bo-ra.
Ornament. T
h
e on
l
y ornament in t
h
is bui
l
ding is t
h
at of t
h
e co
l
umns and episty
l
e
of t
h
e atrium (Il
l
. 1 4 I ) . T
h
e bases are we
ll
turned, and of good proportion, t
h
e capita
l
s
are of t
h
e Ionic order,
h
aving a plain square abacus and a very
h
eavy ec
h
inus wit
h
a narrow torus below it. T
h
e interesting feature of t
h
e capita
l
is t
h
at t
h
e volutes
were not carved, but were executed in stucco upon
f
lat faces ; t
h
e egg-and-dart in t
h
e
middle of t
h
e ec
h
inus was a
l
so app
l
ied in stucco to a smoot
h
ovo
l
o surface. T
h
e
drawing of t
h
e capita
l
(I
ll
. 1 4 I ) s
h
ows a bit of t
h
e stucco volute stil
l
ad
h
ering to t
h
e
capita
l
. T
h
e arc
h
itrave
h
as a broad band be
l
ow a
h
eavy cymatium. T
h
e upper
members of t
h
e entablature were omitted, as was common in arc
h
itecture in Syria in
w
h
ic
h
c
l
assica
l
forms appear,
l
ater t
h
an t
h
e t
h
ird century. It is very probable t
h
at
ot
h
er decorative features in stucco were emp
l
oyed in t
h
is bui
l
ding, and t
h
e remains of
t
h
e
h
ardest kind of p
l
aster upon
h
ig
hl
y fnis
h
ed fat surfaces indicate t
h
at painted
decoration was applied to t
h
e wal
l
s of t
h
e atrium ; t
h
oug
h
no remnants of co
l
or were found.
Purpose and Date. T
h
ere is no bui
l
ding t
h
us far known in Syria, and none e
l
se
w
h
ere, so far as I can discover, t
h
at reproduces t
h
e plan of t
h
is bui
l
ding. It is a
curious and interesting mixture of native and foreign e
l
ements, some of its features
being pecu
l
iar to t
h
e Iauran, and ot
h
ers distinctive
l
y of Greek or Roman origin.
T
h
e vaulted cruciform c
h
amber wou
l
d seem to be t
h
e most signifcant part of t
h
e edi
fce, and t
h
is is p
l
ain
l
y not re
l
igious or funera
l
in c
h
aracter j for no c
h
urc
h
in Syria
1 II. B. 2. Ills. 54, 55.
2 See fragment of vault, S.C. pI. 7. Note : This building i s now destroyed, but records of i t were made. by M.M. Rey
and de Vogue.
Umm idj -Dj imal ( Thantia ?)
is known to have had this form, and it is too large, and too well lighted, to have
been a tomb. The plan of its vaults, however, i s an almost complete reproduction of
those of the so-called Praetorium 1 formerly existing at Mismiyeh (ancient Phaena) i n
the northern part of the Ledja, the only detail lacking being a large niche, or apsis,
i n the eastern branch of the cross, ad columnar, i nstead of solid, supports. It was
this similarity that suggested the probability of the building before us having served
a similar purpose. The Praetorium of P haena was an earlier building, and its classi
cal . elements are far more in evidence ; but, as we have seen in the general discussion
of the architecture of this regi on 2, classical architecture never penetrated these southern
parts of the Province of Arabia to the same extent tha t it did the northern parts, and
the sl ight diferences existing between the vaults of the two buildings, as enumerated
above, are hardly sufcient to deny a similarity of purpose. The building at Mismiyeh
stood by itself, a rectangular structure, cruciform withi n, and having a portico of col
umns i n front. The present structure, as we have seen, was part of a larger build
ing, of a plan very highly articulated for the locality, and absolutely unique in its
form. It appeared, at frst sight, that the edifce might have been a large public bath ;
but a more minute examinati on proved this conjecture to be wholly erroneous ; for the
most signifcant details of the baths in southern, and i n northern Syria (the conduits
and water pipes) are wanting. As I have said above, it is not possible to reconcile
the building, as a whole, with any religious or sepul chral purpose ; but if we assume
that the vaulted chamber was of the same character as the Praetorium of Phaena,
it is not difcult to conceive of the whole structure as a government building, the seat
of the chief ofcial , or ofcials, of this portion of the Arabian province. One may
tentatively thi nk of the vaulted chamber - the Praetorium proper - as the bureau of
the chief civil and mil itary authority, and of the great hall s a basilica, . or court
room, in which cases were tried ; the atrium aforded a suitably digni fed and secluded
entrance to both, and a passage from one to the other, as well as a waiting-room
for persons having business i n either of the other rooms. The square, arched cham
bers along the north wall may be regarded as various important dependencies, (S) as
a private ofce, or retiri ng-room, opening of from a dais at the west end of the
court-room, (T) , (V) , and (W) as bureaus of vari ous ofcers, clerks or secretaries,
and (X) as a general cloak room, or, perhaps, a library. (p) might be regarded as
a store-room for archives, and (x) , I believe, was a l atri na. It would be possible, in
the same manner, to assign uses for the upper foor ; but I have proceeded far enough
in this tentative allotment of uses to the various divisi ons of the building to convey
my idea of the probable purpose of the structure.
The dating of the building involves a curious crux ; for I propose to date it by
an inscription 3 found not in the building itse1 f, but upon a stone now serving as a
li ntel of the southernmost of the three west portals of the " Cathedral " , about 70 m.
northeast of the edifce in questi on. This lintel i s not the main lintel of the portal
of which it is a part, but spans the doorway in a second wall, directly within the
present front wall, built, apparently after the front wall , to strengthen it. The i n
scription i s the famous Lati n one, published by Waddi ngton, giving the names of the
Emperors Valentinian, Valens, and Gratian, and recording the erecti on of a bur
g
us.
I s. c. pI. 7. 2 I I . A. 2. p. 66. 3 III, Insc. 233.
Publications of the Princeton Lniversity Archaeological Expeditions t o Syria, Div. II, Sec. A, Pt. 3 22
1 66 Division I I Section A Part 3
T
h
e inSCrIptIOn, of course, is not in situ ; but it fts t
h
e
l
inte
l
-
l
ess doorway of t
h
e
entrance to t
h
e courtyard on t
h
e sout
h
side of t
h
e Praetorium, from w
h
ic
h
p
l
ace I be
l
ieve it was taken at a
l
ate period, after t
h
e government bui
l
ding
h
ad
l
ost muc
h
of
its origina
l
importance, and after t
h
e origina
l
bui
l
ding of t
h
e " Cat
h
edra
l
" w
h
ic
h
con
tains a dated inscription of t
h
e year 5 5 9, A. D. , probab
l
y t
h
e date of t
h
e erection of
t
h
e bui
l
ding. T
h
e word burgus is rare ; according to Waddington 2 it was adopted
from t
h
e German burg, and was not derived direct
l
y from t
h
e
l
ate Greek rup,o of
w
h
ic
h
t
h
ere are so many examp
l
es in Syrian inscriptions ; it passed into Arabic as
bordj, and is of frequent occurrence. In t
h
e Greek inscriptions of Syria t
h
e word .
rGp,o is usua
ll
y app
l
ied to towers, numerous examples of w
h
ic
h
appear in t
h
e sma
ll
er
rums. Many of t
h
ese towers were undoubted
l
y government bui
l
dings ; t
h
eir exact pur
pose IS not known, but t
h
ey must
h
ave been symbo
l
s of Imperia
l
power, and were
per
h
aps seats of Imperia
l
aut
h
ority in t
h
ese sma
ll
communities. T
h
e Arabic form is '
app
l
ied not on
l
y to towers, but to many ot
h
er kinds of bui
l
dings ; Burdj Balrla
is a ruined Roman temp
l
e, Burd id-DenlnI 4 is a c
h
urc
h
in a wonderful state of pre
servation ; bot
h
of t
h
ese examples are from N
o
rt
h
ern Syria. T
h
e burg, or bourg, of
nort
h
ern Europe, in Roman times, is be
l
ieved to
h
ave been a military structure ; and
it was a
l
so t
h
e seat of t
h
e Roman government. If t
h
is word found its origin in t
h
e
R
h
ine country, it
h
ad travel
l
ed far before it reac
h
ed Umm idj-Djima
l
, and it is not at
a
ll
improbab
l
e t
h
at its meaning
h
ad expanded : ru
p,
o is a very simi
l
ar, and per
h
aps
t
h
e same, word, and t
h
is we fnd used as t
h
e name for towers, origina
l
ly per
h
aps t
h
e
on
l
y government bui
l
dings in sma
l
l towns. U mm idj-Djima
l
was a
l
arge city, its
government bui
l
ding, i.e. its 7Up'o, was not a tower, but an extensive edifc, and, its
inscription being in
L
atin, burgus seems to
h
ave been t
h
e term app
l
ied to it. T
h
e
ot
h
er government bui
l
ding - t
h
e barracks - is ca
ll
ed Xra!cAAO; , or, at
l
east, t
h
ere
is anot
h
er inscription referring to a mi
l
itary structure
h
ere by t
h
is name, and we
know t
h
at xia!poy and Xra!eAA'; were t
h
e Greek words, an
d
castellu7n t
h
e Latin
word, for t
h
e fortresses, barracks, and similar mi
l
itary structures in Syria. T
h
e date
of t
h
e inscription of Va
l
entinian, Va
l
ens, and Gratian, in t
h
e second consu
l
s
h
ip of
Gratian
'
and Probus, wou
l
d fa
ll
about t
h
e year 37 1 of our era. T
h
e construction,
sty
l
e, and genera
l
appearance of t
h
e building, are w
h
o
ll
y in keeping wit
h
t
h
is date.
T
h
e
h
ig
h
exterior fnish of t
h
e wa
l
ls, and t
h
e interior fnis
h
of t
h
e wal
l
s of t
h
e atrium,
suggest an ear
l
y date ; t
h
ey are t
h
e best examp
l
es of wa
ll
bui
l
ding in t
h
e town, wit
h
t
h
e so
l
e exception of t
h
e Nabataean temp
l
e. T
h
e form and proportions of t
h
e Ionic
co
l
umns are muc
h
nearer to c
l
assical mode
l
s t
h
an t
h
e co
l
umns of t
h
e same order 7
found in t
h
e c
h
urc
h
of SSe Sergius and Bacc
h
us at Umm is-Surab, dated 489 A. D. ,
and t
h
is fact, I be
l
ieve, postulates an ear
l
ier date.
BARRACKS : Date, about 41 2 A. D. T
h
e largest sing
l
e bui
l
ding in t
h
e ruined city,
situated near t
h
e midd
l
e of t
h
e sout
h
wa
ll
, and standing quite by itse
l
f, free from sur
rounding bui
l
dings ( I
ll
. 143) is ca
ll
ed by t
h
e Bedawin "id-Der", and t
h
is word, - in
trans
l
ation t
h
e Convent -,
h
as been app
l
ied to it by some of t
h
e few trave
ll
ers w
h
o
h
ave reac
h
ed
U
mm idj-Djima
l
. A convent it may
h
ave been before t
h
e end of its
career as a
h
abitab
l
e bui
l
ding ; but t
h
ere can be no doubt t
h
at it was origina
ll
y bui
l
t
to serve as a fort or barracks, preferab
l
y t
h
e
l
atter ; because it was built wit
h
in t
h
e
I II. A. 3. Inscr. 260.
II. B. 4. p. 201 .
^ cf. Waddington's commentary under Insc. 2058.
III. insc. 233. " III . insc. 237.
A .A.E. S. II. p. 66.
, See Ill. 78. II. A. 2.
U mm i dj-Dji mal ( Thantza ?)
city walls, and was i ntended for the housing of soldiers rather than as a stronghold
for defence. The similarity, i n plan and in constructi on, which i t bears to the fortresses
of or il-lallabat, ar il-Baci and Dr il-Kahf, leaves no doubt of its military
character. The last named building, it will be noticed, is also called Der by the Arabs
of the desert, though, in its i nscripti on, it i s designated by the word castellum.
Plan. The building i s a rectangle ( 3 3 . 7 5 m. X 5 5 m. ) with i ts longer axi s lying
north and south (Ill. 1 44) ; just south of the mi ddle of its east side, a rectangular
chapel j uts out from the main building. The present entrance, - a low and narrow
doorway with a stone door -, in the east wall, j ust north of the chapel , is not the
original ; for this part of the wall shows evidence of having been rebuilt at a period
consi derably later than that of the origi nal structure. Rooms of varying depth are
disposed on all sides within the heavy outer wall, leaving a yard of somewhat i rregular
Ill . 1 43. The Barracks ; View from the Southwest.
shape. A high tower, 4. 40 m. square, and about 1 5 m. high, stands at the southeast
angle ; this does not break out from the straight li nes of the two walls, like the towers
at oer il- ljallabat, but simply occupi es the angle made by them, in the same manner
as was seen i n the fortress of ar il-Baci. 2 The other three angles of the building
were not planned to receive towers. The south side of the rectangle i s occupied by
a row of shallow rooms of the same depth as the tower ; the southern half of the
west side, by a deep set of apartments of various shapes and sizes, and the northern
half by a double row of barrack-rooms, deeper than the others, and having thei r
front wall set a metre forward of the wall of the southern half The north side is com
posed of chambers only 4 m. deep, but varying i n width from 6 to 7 m. ; three of
these rooms have girder-arches . On the east side of the yard there are long shallow
rooms on either side of the entrance, and then a double row of six rooms, three in each
row, between which and the tower i s a large square apartment with two girder-arches.
II, /. 2, Ill. 5 5 . ^ II. A. 2 , Ill. 6 I .
1 68
`

Di vision !! Secti on A Part

"
|
.
.r

`
,lL L /
I-- 7 Gc o{ -
9 g . . . .

0
1
n
' BARRCKS
j T -1
/
l r `-.!. :. ./L
J T

\ l . 5^
^8u s i 2 8I
h *;`0
2o

g . O ,

, 4
_ ' ' -- L
T1 C L
1 ; 0
g , , .
.
2.ao
~" ! / <. 8 ! __ .9
1

:
_
:
_
l

g .

. %0
'

;;

1 4
f-
G --
1

:

. .' .
.
- _ " -

I
JJ/

\ 4. q o

..
.LLl Cl
Ill. 1 44.
Il l . 1 45 . The Barracks
; Vi ew from the Sou theas t, showing Tower and South Wall of t he Chapel.
Ill. 146. The Barracks ; Part of Wal l on West Side of Court yard.
Umm i dj-Dj lmal ( Thantia ?)
Superstructure. On the ground-plan described above, the superstructure is of
height on all sides, except where the two towers stand, though some parts have
storeys and others only one, the former having an i ntermediate foor and a roof
slabs carried on corbels, the latter having high girder-arches also carrying a roof of
. The angle tower rises three storeys above the roof of the apartments adj oin
(Ill. 1 45), and there were two high rooms, one above the other, i n the mi ddle of
west side. making an oblong tower-like structure at that point. The outer wall
solid and thick (from 0. 80 m. to I . l o r. )
"
; i t i s broken only by l oop-hole windows,
at regular i ntervals on the south side, and irregularly pl aced on the other
On the outside the wall was carried up 3 m. above the roofs. The wall itself
constructed i n the same manner as the outer wall of the fortress at ar i l-BSl}.
tower, above the ground storey, is of totally diferent workmanship, being of the
highly fnished quality of stonework found in the Christian architecture of Southern
(Ill . 1 45) . The northern half of the west wall of the court-yard of the barracks
the same high quality of workmanship (Ill . 1 46) ; the courses are comparatively
but the blocks are long and accurately joi ned ; the edges are sharp, and the
is so smooth as to give the efect of havi ng been polished. The upper foor
this part was reached by a stair which led up in the angle of the proj ecti on, and
balcony, composed of long and highly fnished slabs corbelled out from the wall ,
across the front ; traces of this are to be seen in the photograph (Il l . 1 46),
h most of the slabs have been broken of. The remainder of the structure was
i n the more crude style of Dr il- Kahf and ar i l-B,fil, where rough quadrated
nry was used. The entire northern half of the west side of the building was
to barrack-rooms, like those at Der il-Kahf \ the southernmost of them was
.
fnished on the i nterior, and has a nl0lIlded corbel course below the ceiling ; it
probably an ofce of some sort ; for it i s larger and better lighted than the other
The southern half of the same side has one large arched apartment, one
ler room, two passages, and a long narrow unlighted room in the rear ; the group,
it belonged , to the original military structure, may have constituted the residential
quarters of the commandant, or, if it belongs to a later reconstruction, which from the
diference in construction appears probable, i t ' may have no relati on to the building
a military structure. The small rooms on the south wall appear to have been
. (
The large arched room in the southeast angle of the yard, which had another
arched room above i t, was perhaps another ofcers' quarters. There are two piers
within it, which show that the room was origi na
l
Iy spanned by a single girder-arch set
t right angles to the present pair which support its ceiling of stone slabs. Adj oi ni ng
this large room on the north is a long passage, also i n two storeys,_ and, beyond this,
the second series of double barrack-rooms, two storeys high. The purpose of the
narrow arched rooms on the north side i s somewhat problematical ; they were probably
barrack-rooms of another sort, like the large arched rooms on the south side of the
fortress at or il-Jallabat. The entrance, as I have said above, does not seem to
me to belong to the original structure ; yet, from my observati on of the wall at this
I am convinced that the original portal , a much larger one, occupied the same
t,",. . n .. ; for an entrance in the mi ddle of the east side, like those at oer i l-Jallabat
1 II. A. . Ill. 127.
1 70
Division II Section A Part 3
and Dr il- Kahf, would have opened into one of the barrack-rooms which I believe
belong to the early building. It seems as if the frame of the original d
o
orway had
been removed, and the wall on either side of it and above it had been rebuilt, and
the present narrow entrance substituted for the old one. At thi s ti me the lintel bear
ing the building i nscription was carried away, perhaps with other material , and used
i n the building of a private house about 2 00 m. to the west, near the west wall of
the town. Thi s house, by every sign, is one of the very latest buildings i n the city.
Tower. The high tower in the southeast angle, which is almost compl etely pre
served, and is perhaps the most striking feature of the city as viewed from a distance,
rises practi cally unbroken in six storeys. No architectural distinction is made on the
outside between the ground foor and the one above i t ; but each of the other storeys
is marked by a narrow course which projects slightly. The storey below the upper
most foor has a small square window in each side. The top foor has four l arge
openings, one in each face of the tower, those on the north and south faces being
set two courses higher than those on the other two sides. Each opening is four courses
- about 1 . 50 m. - high. At the bottom, on either side of each opening, a corbel projects
from the wall , and upon these is built a structure, like a high-walled and bottoml ess
balcony, three courses high (see Sect. C-D in Ill . 1 44
)
. On the face of each of these
structures is a large incised cross and the naine of an archangel ; Michael , Gabriel ,
Raphael or Uriel. The use or purpose of these proj ecti ons is not apparent ; similar
constructions, with stone foors having one or two holes in them, are not rare in
_
towers
and other structures i n northern and southern Syria ; some of these, when placed
above doorways, were of the nature of machicolations for the discharge of missiles
upon the heads of enemies forcing an entrance ; others were unquesti onably latri nae.
I n the present instance, however, the placing of this feature on all four si des precludes
either of the uses menti oned above ; while the absence of a foor and the presence of
i nscriptions fail to suggest any particular purpose. I am inclined to l ook upon this
tower as an early belfry, in which some sort of resounding instrument, the se1nanter,u17z
perhaps, was placed. The outer structures may have been for the protection of this.
There are several other inscriptions 1 upon the tower, one in beautiful large letters
upon a belt-course, and others which are partial belt-courses. All of the inscripti ons
are of a religious character : but I beli eve they are coeval with the original building
of the barracks.
It is plain that the tower, like all the rest of the building, was covered with
stucco. A band of this material is still plainly visible near the top of the tower, and
patches of it are to be seen in many parts of the building. Here, as in numerous
other examples, one fnds that two kinds of stucco were employed ; one, of ordinary
quality, which was used to cover most of the surfaces, and which has almost entirely
disappeared ; and another, of particularly durab
l
e quality, whi ch was laid on in bands
and discs, usually upon smoothly fnished walls, and was probably intended to receive
a painted design. This quality of stucco, in many cases, is well preserved. If there
were ornamental features in this building they must have been executed i n stucco or colour.
Chapel. The chapel is a basilica of medi um size. Like many of the churches
in this regi on, it has a nave of three bays. Its presbyterium is almost square and
1 III, inscs. 245-25 5 .
Umm i dj -Dj imal ( Than#a ?)
opens into the prothesis, on the south side, by a broad arch, almost as wide, though
not so high, as the chancel arch ; an arrangement unique in the churches of southern
Syri a, as far as I have observed. The prothesis opens upon the south aisle by an
other arch. All three of these arches are in situ. The diaconicum i s an oblong
cham ber on the north of the presbyteri um; it has a narrow doorway and a small
window. The superstructure i s exceedingly simple. The outer ' walls were built of rough
quadrated masonry, the i nterior piers and arches of large blocks of stone highly
fnished ; the walls were built up to the height of the crowns of the arches ; a corbel
course projected from the walls, double corbels were laid upon the arch walls, and the
whole i nteri or was covered by a fat roof of stone slabs. A few small windows were
placed high in the walls, and a narrow doorway i n the north wall, with another i n
the south wall , provided the only entrances.
Character and Date. As I have said above, the resemblance which this build
i ng bears, i n its construction and i n parts of its plan, to the fortresses of Dr il-Kahf
and ar il-Baci , gives sufcient evi dence of its military character. The double bar
rack-rooms on the east and west sides are the only rooms of this particular plan and
arrangelnent that I have seen in southern Syria, outside the fortress of Der il-Kahf.
The outer walls have the thickness common to walls of many castles and forts i n
Syria ; while there i s nothi ng i n the plan or arrangement of the building that recalls
features in any of the monastic edifces of the neighbourhood. The religious tenor of
the i nscripti ons upon the tower need not ofer a barrier to the belief that the original
building was a barracks, or that the tower belongs to the military period of the history
of the building ; for religious i nscripti ons abound in the military structures of the
Christi an period in Syria. 1 There i s no other building in U mm idj-Dj i mal, or ' remai ns
of one, to whi ch the i nscription of Pelagios Antipatros might have belonged. This
dux Arabiae built the castle of ar il-Baci i n 41 1-1 2, and there can be little doubt
that the i nscription found i n Umm i dj-Dj i mal belongs to about the same year. It is
not possible to say at what time the mai n entrance to the barracks gave place to the
small doorway which we see today, or to venture more than a conjecture as to the
peri od when the barracks became a monastery. One cannot assume that the erecti on
of the chapel was coeval with the conversion of the purpose of the building ; for
chapels are found in fortresses and barracks from one end of Syria to the other,
though, i n all known examples, the chapel is withi n the enclosure of the military structure.
There is now no part of the i nterior of this barracks that might have been the chapel
of the guard, and I am i nclined to believe that the present church edifce was added
to the barracks while it sti ll served its original military purpose. \hen the present
entrance was made, the crude overhanging structure above i t was added. This i s a
single machicolation corbelled out from the wall and provided with thin walls. It i s
of late and poor construction.
EcclesiasHcal Arckitecture.
The ffteen churches of Umm idj-Dji mal may be classifed under two general
types : those with undivided naves spanned by transverse arches, i . e. the hall-churches,
and the basilicas. Both types present numerous varieties from a fxed form, i n ground-
1 cf. III, inscs. 865, 9 1 5, 947.
1 72 Division II Section A Part 3
plan as well as in superstructure. The former class - the hall-churches - may be
said to embrace all the churches without si de aisles, from the extremely elongated
form with many transverse arches, to those which are nearly square in plan and have
but one transverse arch. In this class we fnd the presbyterill m sometimes rectangular
and sometimes semi-circular, i n one case concealed on the exterior by fanking cham
bers, and i n another showing a complete exterior curve ; while a narthex is present in
one example and absent i n another, or i s replaced by a lateral portico. The ground
plans of the basilicas also present several varieties of arrangement ; while the variati ons
i n their superstructures are even more diverse, and consequently very interesti ng. We
fnd curved protrudi ng apses, and apses concealed between side chambers, naves divided
i nto aisles by broad arches carried on piers and narrower arches supported by columns.
In some of the basilicas the si de walls are of equal height with the i nteri or arches,
the whole nave being covered either by fat slabs on one level, if the middle aisle be
narrow, or the side aisles covered with slabs and the main nave with a simple doubl e
pitched roof of wood. I n other i nstances the superstructure has the typical Latin basi
lical form with a cl earstorey raised above low side aisles, with this diference from the
Lati n type, that the side aisles are roofed with fat slabs of stone. A si ngle example
has clearstorey windows directly above the nave arches, opening i nto a gallery, as if
the clearstorey and a trifori um gallery had been combined. Many of the basilicas have
a colonnaded porch, or narthex, others have a porch composed of arches and pi ers ;
while an isolated church had a broad single-arched narthex between proj ecting towers.
The diversity of plan and arrangement is i ndeed great, and especially signifcant i n a
city so far removed from the great art centres.
The churches represent various grades of excellence in the matter of construction;
though none of them can boast of having been built in the best manner known i n
the I:auran, neither does any fnd its place among the poorest examples of masonry
in this region of extremes i n the matter of constructi on. Three churches were built
in part of highly fnished quadrated masonry ; each one of them contains fragments of
classical ornanlent built i nto the walls, suggesting that the highly fnished quadrated
blocks also were taken from some abandoned building of the Roman peri od, or earlier,
perhaps a temple. The maj ority of the church buildi ngs were constructed of smooth
quadrated masonry, and the remainder of rough quadrated work, the best of its class.
One-half of these church edifces seem to have been connected with monasteries,
or, at least, are i ntegral parts of larger or smaller groups of resi dential buildings ; the
other half stand quite free, or have no more than very small residential quarters at
tached to them; one is the chapel of the barracks. Only two are defnitely dated by
i nscriptions ; one in the frst half of the fourth century, the other in the second half
of the sixth ; and these two dates are hardly sufcient to be useful i n determining the
dates of the other churches. I do not believe that the types represented by the two
dated churches were confi ned to the fourth and sixth centuries respectively, although
they are wi dely diferent types. The earlier of the two dated churches, which i s the
earliest of the dated churches i n all Syria, and the earliest of churches anywhere i n
the world that i s given a defi nite date by an i nscripti on, i s of the hall-church type of
whi ch there are other examples that are certainly very early, like the church at cAnz 1
1 II. A, 2. p. 1 33.
Umm idj-Dj i mal ( Thanta ?) 1
73
the ornament of which can be hardly later than the frst half of the fourth century.
The type was undoubtedly derived from the long halls in later Roman buildings i n
the lauran, like that i n the Iaiarlyeh at Sha11a which i s a second or third century
building, and was i tself converted i nto a church. But i t is quite possible, even pro
bable, that the type persisted throughout the Christian peri od in the lauran, if not i n
its original very elongated form, at least i n the shorter form with one or two trans
verse arches ; for the reason that it was the simplest kind of structure to build when
a small church, or chapel , was desired.
The later church, built over two hundred years after the other, is a basilica.
This type too is early in Northern Syria where there are examples dating from the
fourth century. But ' in the South the type i s rare at best, except i n the form, un
known i n Northern Syri a, in which the clearstorey i s omitted and the entire roof i s
fat and . on one level . However, I shall begi n the description of the churches wi th
the earliest dated examples, and let the other hall-churches follow i mmediatel y, leaving
the basilicas until the last.
One cannot fai l to observe, in looking over the larger map of the town, that the
churches, though havi ng their apses i n a general way toward the east, show great
irregularity in their orientation, the diferences between them varying at the maxi mum
as much as 20 degrees. Recent attempts, - successful , I believe, i n the mai n -, have
been made, as i s generally known, to date Greek temples accordi ng to the degrees
of their divergence from an ori entati on based on the true north ; but I am unable to
say whether a si milar rule would be applicable here. Nor have I been able thus far
to discover whether the churches that have the same, or nearly the same, beari ng are
to be assigned to approximately the same dates. It is true that the Church of Julianos,
built i n the fourth century, i s so placed that the bearing of i ts front wall is almost
exactly due north, and that the Cathedral , which i s a sixth-century structure is so
oriented that the beari ng of its front wall i s north 2 5

east. Most of the churches in
the town conform to one of these angles or the other. In one or two cases the dia
gonal of the nave lies due east and west, throwing the axis of the apse to the south
of east. It is an i nteresti ng feld for further i nvestigation.
CHURCH OF JULIANOS : Date 345 , A. D. This most ancient of the churches of Dmin
idj-Djimal is hidden i n a large group of buildi ngs to the north of the centre of the
town (See plan of city). Some of the buildings which surround i t on three sides be
longed to it, as residential buildings for the clergy who served withi n i ts walls, others
certainly were not dependencies of the church, and were probably ordinary resi dences,
either owned by private i ndividuals or rented from the church. They are not directly
connected by doorways or passages with the church enclosure ; though they make use
of the walls of the church buildings as party walls. The building is unfortunately i n
a sad state of dilapi dati on. So great i s the mass of debris wi thi n i t that it i s not
easy to trace i ts outlines wi thout time and care ; but when these are applied, a com
plete plan of the church i tself can be made out owing to the very simplicity of the
design. The apse i s preserved to the springing of its half dome, and the residential
buildings are exceptionally well preserved, presenting no difculties of restoration. The
inscription which gives the name by which I have designated this church, and the
date 345 A. D. , is not in situ. It was i nscribed upon the lintel of the middle portal
in the south wall of the church, and now lies i n the court, not four metres from i ts
Publications of the Princeton Cniversity Archaeological Expedition to Syria, Div. II, Sec. A, Pt. 3.
I 7
4
Di vision II Section A Part 3
original posItI on. I have no doubt that this inscription, set up by J ulianos, and re
cording the erecti on of a memorial , was the ori ginal dedicatory i nscri ption set up at
the time of the completi on of the church. One or two epigraphical scholars, with
whom I have discussed the questi on, appeared to be somewhat in doubt as to whether
the Greek word {'Y[EOV used i n this i nscription, was apllied to churches at so early
a peri od, and raised the question as to whether this li ntel might not have been brought
here from a tomb while the church was i n process of construction. The word, how
ever, probably occurs in the i nscripti on 1 on the li ntel of the main west portal of the
Church of S. S. Sergius and Bacchus at Un1m is-Surab, 2 where i t is applied to the
church ; the i nscription i s dated 489 A. D. The i nscription of Julianos, moreover, i s
placed upon the lintel of the principal portal of the church - the middle portal on the
south si de -, and is a stone far l arger than the lintels of tombs i n Southern Syria,
and shows no signs of havi ng been re-cut. I may add that there is no example i n
Syria, within my knowledge, i n which there i s evidence of a Christi an tomb having
been despoiled to build a church. The beari ng of the axis of this church i s almost
due east.
Plan. The nave (Ill . 1 47) has ten bays divided each from those adj oi ning i t by
arches of seven metres span which are carried on piers that proj ect . 7 5 m. from the
walls. The bays are all of equal depth except the westernmost which is nearly twice
as deep as the others. The presbyteri um i s a semi-circular apse, almost as wide as
the nave, separated from the nave by an arch, only fve metres wide, carri ed by two
deeply proj ecting piers. The apse wall is preserved up to the level of the half dome
(Il l. 1 48) . Along the north si de of the nave extends a l ow, narrow building divided
i nto rooms by three walls which, with its east wal l , abut against the frst, third, ffth
and eighth arches of the nave. Three of the rooms are connected with the church by
doorways, and a window in every other bay of the nave opened out over the roof of
the side building. A colonnade of eight columns, three of which are i n situ, was car
ried along the south side of the church, from i ts western end to the li ne of the seventh
arch ; beyond this is a vestibule opening upon the end of the colonnade by an arch,
and out of doors by a broad doorway. The colonnade gives upon an oblong court,
at the east and west end of which there i s a small residence, comprising each a large
room, i n whi ch the well-built transverse arch i s still preserved, and one or two smaller
apartments i n two storeys. The south side of the court is occupied by shllower build
ings, in two storeys, with rooms of diferent sizes and shapes, one of whi ch i s a l ong
stable ftted with stone mangers. These buildings preserve the walls of their two storeys
in completeness.
Superstructure. The restorati on of the east end was not diffcult for the reason
that the apse wall and both of the apse piers are preserved. The wall was built of well
fnished quadrated blocks, as is shown i n Ill. 1 48, with narrow bonding courses at i nter
vals. One or two piers were preserved on the south side of the nave adj oi ni ng the
apse, giving the diference i n height between them and the apse pi ers. The arches
which sprang from both had their crowns on the same level , as is shown in Section
A-B of Ill. 1 47. The restoration of the nave as shown in Section C-D, Ill. 1 47, is
based upon the height of the pi ers, whi ch is a known quanti ty, and semi-circular arches
t Ill. A. 2. insc. 5 I . 2 II. A. 2. pp. 95-99.
Umm idj-Dji mal ( Thantia ?) 1
75
3. CHVRCH
O F
I VL I ANOS
1 3. 3

c
b. 0
DATE: 345AD
DICTION E-F-
:ESTORED
VlIDJ-TJI/
Di visi on II Section A Part 3
the width, and consequently the height, of which is known. There can be little doubt
as to the accuracy of the restored Section E-F, which shows the south side of the church
and a cut through the residence at the east end of the court ; for the residence is stil1
standing, and two columns of the colonnade on the south side of the church are stil1
in place, while the bases of the others are also in situ. The presence of architrave
blocks and of Doric columns of a smaller scale than those still in place proves conclus
ively that the col onnade had two storeys.
CHURCH of MASECHOS : This church forms the northern boundary of a large grou
of buildings i n the southeast quarter of the town (See Map N. 2) , just south of the
East Gate. The ancient wall of the city 'as apparently removed when this church and
the buildings south of it were erected, and a new wall was built out to the eastward to
i nclude them ; the north wall of the church being made to serve duty as a part of this
new wall. A short inscri ption 1 above a doorway at the east end of the narrow chamber
on the south side of the apse contains the proper name that I have chosen for the
title of this church ; for one Masechos was probably the builder of it. The photograph
(Ill . 1 49) gives a picture of the street upon which the Masechos church fronts. Looking
up the street one faces north ; at the extreme left of the picture, but on the right side
of the street, one may see the front wall of this church with the corbels of the porch
roof protruding from it, above a mass of debris. At the right of the picture stands the
west front of the Southeast Church, which will be taken up next. A view of a part
of the north wall of the church is given in Ill. 1 3 5 ; it shows the character of its
construction.
Plan. An undivided nave of six bays separated by arches of over six metres
F AC AD rSTOED-
CHVUCH
MSrCHOS ,
__ NM IDJ-DJlM
D
Ill. 1 50.
span, and carried on sali ent wall piers, a deep-set semi-circular apse between long
,
narrow side chambers
, and a colonnaded porch of four bays, closed at one end and
1 III. A. 3. , inscr. 264. .
Ill. 1 48. Apse of JuJianos Church ; Vi ew from the Sout heas t .
Ill . 1 49. Street with Faades of Masechos Church and Southeast Church on the Right .
Umm i dj-Dj imal ( Thantia r) 1
77
open at the other, make up the simple plan of this church. This church, with one
other i n the city, has the unusual feature of a double west portal , and the columns of
the porch are so spaced, with reference to the wal l and pi er at the south end, that
one column stands i n the middle to accommodate the two doorways. The porch is a
paved platform raised about 60 cm. above the street. There is a doorway in the south
wall of the church leading into a room which forms a part of the group of residential
buildings about the court, or cloister, on the south side of the church. The prothesis
and diaconicum are long rooms of extreme narrowness, being hardly more than a metre
wide. The room on the south side of the apse, which was probably the prothesis,
has a doorway i n its eastern end opening out into the space between the church and
'the town wal l .
Superstructure. It seemed hardly necessary to give a l ongitudinal secti on of this
building ; for its arrangenlent is similar to that of the Julianos Church. The three
westernmost arches of the nave, and the apse arch, are still standing. The nave arches
having outlived those of the J ulianos Church owing, I believe, to the fact that their
piers are more salient and that they are consequently more frmly buttressed. The
drawing (Section A-B i n Ill . I 5 0) shows the proportions of the nave arches and their
relati on to the apse arch which is fanked by the doorways leading i nto the side chambers.
A few of the roofng slabs are still i n place. The drawing of the fa<ade explains itself ;
a disti nction i s shown between the parts of the wall that are standing and those which
are restored. The height of the columns was easily discovered from the fallen shafts
and capital s. The moulding of the edge of the platform, probably taken from some
other building, is shown i n Ill . I So. The capitals and bases of the columns are
illustrated by (D) and (E) .
SOUTHEAST CHURCH. At the extreme right i n Ill. 1 49, the dilapidated front of this
church may be seen. It stands in the midst of houses, and connects, by means of a
doorway in its north wall , with the court yard on the south side of the Masechos church.
The building is less well constructed than the former, and appears to be of l ater
date. Its plan and arrangement are totally diferent from those of the two hall-churches
described above, and I have prepared a full set of drawings to illustrate its structure.
The plan (Ill . I S I ) is oblong, with a semi-circular apse, showing 010St of its exterior
curve, opening into the nave by a comparatively narrow arch, and having a minute
chatnber on its north side. The nave, i nstead of being provided with several trans
verse arches for the support of slabs, has but one arch which spans the middle of the
nave, and was carried up to gable-form to support a double pitched roof of timbers.
This church, like the one just described, has two front portals with a porch of three
col Ulnns between returned end walls. The columns are not in situ, but lie i n the ruins
of the porch. The two secti ons and the drawi ng of the fa<ade, all given in Ill. I S I ,
illustrate the peculiar features of the church, and show, by shading the preserved
portions, which of the walls are sti ll in place.
EAST CHCRCH. This is a complete l ittle convent situated near the middle of the
east wall of the town, and utilizing a part of the ancient town wall as its own east wal l .
The church itself abuts upon the south si de of a small arched gate i n the wall, and
the resi denti al buildings with the cloister court, lie to the south of the church. The
si ngle nave has six bays, a smal l rectangular presbyterium fanked by narrow chanlbers,
and a colonnaded narthex of three bays (Ill. 1 5 2). Near the middle of the south wall
178 Di vision II Secti on A Part 3
there is a doorway opening into one of three apartments that fank the church on that
side, and occupy the north side of the court. Two of these apartments are square
and are spanned each by a transverse arch which brings their fat stone roofs to the
height of the similar roof of the church : the other is divided by a transverse wall i nto
two rooms on the ground foor, and is two storeys high. To the westward of this
apartment are two small square rooms in two storeys, which proj ect beyond the front
wall of the church. pn the west side of the courtyard are three rooms, two of them
small and in two storeys, the other somewhat l arger, and spanned by a high trans
verse arch. N ext to this, on the south side of the court, is a small vestibule with a
`LL L1l'
`L L'
Ill. 1 5 I .
A /DL
``lL/O
'l\ll
t1
AND
-.U/'b
doorway on the interior and an arch opening i nto the street. The bearing of the
maj or axis of the church it almost the same as that of the Julianos Church, and, if
the same angle of orientati on i ndicates coincidence of date, we have here another
fourth-century church. Consi derable portions of this buildi ng are in situ : the front
presents quite a striking appearance with its steps partly visible and with one of its
columns still standing full high and others half high. F rom the ruins, that is from
the spaci ng of the two mi ddle columns and from the voussoirs that lie i n the portico,
it seems quite certain that the middle i ntercolumniation was arched. Two of the trans
verse arches at the west end of the nave are still i n place with the slabs of the stone
roof above them, and the narrow arch of the presbyterium and parts of its roof are still
preserved. The buildings about the court present a well preserved and interesting group.
Umm i dj -Dj i mal ( Thantia ?) 1
79
CHAPEL OUTSIDE EAST 'T ALL. On the top of a knoll , about 60 m. from the east
wall, and half way between the two gates, is a small church, or chapel , which belongs
J.3o
o
.--10. ---
ZIO
Ill. 1 5 2.
EAST CHVRCH
`ll lJ-DJIMAL
H
to the class of aisleless buildings with transverse arches. The nave (Il l . 1 5 3), only
1 2 m. long and 6 m. wide, is spanned by two
transverse arches, and has a semi-circular pro
j ecti ng apse. Along the south wall is a narrow
low structure (Sect. A-B) divided i nto two cham
bers ; east of this is a small chamber adj oi ni ng
the apse . but not connected wi th i t. The two
smal ler of these rooms probably served in place
of the side chambers that often fl ank the apse,
providi ng a prothesis and a diaconi cum. The
western half of the nave of thi s chapel , and the
east wal l with the arch of the apse and its curved
wall, up to the springing of the half dome, are
lLJJ:ESORED-
'lL
well preserved though the building belongs to
V
the poorer class of structures being of roughly
quadrated masonry throughout, except i ts arches A.--- ... _ ___
and doorways. I presume it is a rather late
church and may have been either a memorial
chapel or a cemetery chapel ; for i t is i n the
ancient burial grounds on the east side of the city.
DOUBLE CHURCH. This building may be
Ill. 1 53.
taken as the l ast of the hall-church group and the frst of the basilicas ; for both classes
are represented in this pair of twin churches which were constructed as one. The double
church is situated in the southeastern quarter of the city, well within the walls, and
quite surrounded by private houses, except on the north side where an open place
stretches away toward the Roman Reservoi r. It seems probable that none of t he resi
dences in the i mmediate neighborhood belonged to the church which appears to have
1 80 Division II Section A Part 3
been an ordinary parish church and not part of a monastery. The ground-plan (Ill . 1 5 4)
shows the two naves side by side, the basilica on the north, the hall-church on the
south ; both having deeply curved apses which protruded conspicuously from the east
wall. It is quite plain, after a careful examination of the structure of the wal1 s at the
points where the two churches join, that the hall-church was added to the other. The
nave of this half of the church has three bays, divided by transverse arches of broad
span, and an apse which embraces the entire width of the nave, although the apse
arch was somewhat narrower than the nave-arches. The west portal i s unusually high
and broad : the only i ndication that there was a porch here is a row of corbels pro
jecting from the wal l . The west and north walls are preserved to their ful l height ;
I '

A-J

l.'J
CAPITAL-
J`LL'l`l.l.
SLCI OI C)cm

`ll L1L! l l/ L H
111. 1 54.
the apse wall is standing to the height of about three metres ; but the south wall has
collapsed, carryi ng the interi or arches and the stone roof with it. An i nteresting
feature here i s a stoup, or basin for holy water, called colmbon i n the early Greek
church, which protrudes, i n bracket-from, from the wall of the church j ust at the right
of the mai n entrance (Ill. 1 5 5) . There are similar details i n other churches of this city,
and at least one example from another ruin i n the Southern Jaurn, viz. , that i n the
church of St. John at il-Umta
C
1yeh 1 . The photograph (Ill . 1 5 6) gives a view of the west
front of this southern half of the church. The street, as may be seen, is flled with
the debris of a house i mmediately west of the other half of the church, which has
collapsed. The broad and high main portal , with the colymbi on beside it, show well
cf. II, A, 2. , Ill. 74.
Umm i dj - Dj i mal ( Thantia?)
1 8 1
I n the picture ; while the fallen arches of the i nterior and part of the apse wall may be
seen through the doorway. The quality of the roughly quadrated masonry of the walls
also may be seen in the photograph.
The northern half of the double church is one of the most i nteresting examples of
ecclesi astical architecture in all Southern Syria, a unique speci men of the basilica type ;
yet one which would appear to have i nfuenced early Moslem architectural forms throughout
Syria. The ground- plan (Ill. 1 5 4
)
is that of a simple basili ca, with an unusually broad
middle aisle and extremely narrow side aisles which, however, are not of equal width ;
the north aisle being half as wide as the other, and only a metre wide. The deep
protruding apse is of equal width with the middle aisl e. There are two doorways i n
the west end, - a peculiarity which we have seen also i n the Church of Masechos and
i n the Southeast Church ; these opened i nto a closed narthex with only one entrance,
that in its south end. The
church has three other door
ways, one opening upon the
space to the north, one open
ing into the westernmost bay
of the other church, and a very
narrow doorway also leading
i nto the other church just west
of the chancel arch. The most
. _ " o_ _
: I \ I
'
I,
1
e
',

. f h I
I I >/ ,:' , I
InterestI ng eature 0 t e p an '
/ " ', /, ! , ` 1
however ; one which is more 1 l . . 1, ,. , : ' .
fully illustrated in the Sections
I

_ _ . " ' V '

' '
"

A-B and C-D, is the use of

" " o
. ':".'

_ '_ " I'

* .

. ` r
s +

col umns for the nave arcades.
Ill. 1 5 5. Waterbasin in wall of Double Church.
These columns are widely spaced and carry arches of broad span ; they are employed
exactly as piers are employed in all the large class of broad-arched churches i n Southern
and Northern Syria. The columns are quite sl ender and well proportioned, and the
arches which they carry sprang from small cubical plinth bl ocks which gave the arches
a very slender impost. The arches carried roofs con1posed of stone slabs over the side
aisles, and a timber roof over the middle aisle. This roof of wood had the gable form,
and was set directly upon a low wall above the arches, corresponding to the parapet
of the outer edge of te roof ; there was no provision made for a clearstorey, unless
there were open spaces in this low wall, or parapet, which seems improbable. The
north wall is devoid of window openings, so that all the light must have been admitted
through windows in the west end. The piers of the chancel arch are much higher
than the columns of the nave arcade, and the bema was el evated upon three steps.
The capitals of the col umns (See drawing i n Ill. 1 5 4
)
are i nteresting examples of late
Doric, with a perfectly straight echinus and a banded abacus ; all are well executed.
It i s a curious and i nteresting fact that among the numerous and varied examples
of arch building furnished by the hundreds of known churches i n Southern and N orth
ern Syria, there is no other example of columnar supports for an arcade of arches of
broad span ; the universal arrangement being piers for broad arches and col umns for
narrow ones. It is quite as if the great majority of builders did not realize the strength
of a comparatively slender column to resist a simple downward pressure, nor under-
Publications of the Princeton University Archaeological Expeditions to Syria, Div. II, Sec. A, Pt. 3. 24
1 82 Di vision II Secti on A Part 3
stand that a row of arches of almost any width could be supported by very slender
uprights provided the row was suffciently buttressed at the ends. The drawi ngs show
how deep were the pi ers which buttressed these two arcades, and the structure, though
so much lighter and more open than was common, proved the techni cal skill of its
builders ; for, by every sign, these arches have fallen only very recently, i ndeed the
fractures i n the fallen arch stones looked very fresh in 1 905 . But however uncommon
thi s method of arch constructi on may have been among the Christian architects of
Syria before the seventh century, i t became the generally accepted method for the con
struction of the arches of mosques of the early and late Moslem builders i n Southern
Syri a, even in buildings where no other technical skill was shown. I n a great number
of ancient mosques throughout the laurn and the N u1rah, many of which have been
long deserted and are i n a parti al state of rui n, a slender column shaft, often not a
whole column, taken from some Christian or Roman building, will be found acting as
the support of two i nterior arches, if the mosque be small as i s usual ; or two columns
i n arcades of three arches, or more, according to the si ze of the mosque.
This church was built throughout of rough quadrated work, with door-frames,
columns, piers and arches of excepti onally high fnish. The walls were treated with
three coats of plaster, frst a coarse coat to fll up the uneven interstices, then a thi nner
coat which was roughened to receive the third l ayer - a thi n and extremely fi ne hard
coat which was laid over the other coats and over the details which were executed in
highly fnished stone. The photograph (Ill. 1 5 7) shows some of this plaster still ad
hering to the walls and the more highly fnished details such as the pier and part of
the arch near the middle of the picture. and the fallen capital in the i mmediate fore
ground. These bits of plaster have been so long preserved for the reason that they
were protected until quite recently by the roofs of the side aisles, which, as I have said
above, collapsed not so very long ago. A fragment of this roof may be seen i n the
photograph, consisting of a few slabs over the west end of the north aisle, with the
fne mortar, the pebbles, the volcanic scoriae and the beaten clay, which were laid in
successive layers upon the slabs i nsuring a perfectly water-tight roof. Two i nscriptions
were found i n this church, and both of them i n situ ; one 1 a painted i nscripti on i n
Greek over the north doorway, outside, now very faint, and one i n Arabic of the ear
l iest form and therefore very important as an example of pre-Islamic Arabic. Thi s is
on the south side of the pier at the west end of the northern arcade of the nave,
di rectly behi nd the arch-stones shown i n the photograph (Ill. 1 5 7) . That this i nscripti on
bel ongs to the Christian period cannot be gainsai d, for it was covered with plaster at
some ti me when the i nterior of the church was undergoing repairs and decorati on, and
there i s no evidence that the building was ever used by the Moslems. The plaster
which covered the i nscripti on is of the same quality as that which covered the highly
fnished details, which is to say the best quality employed by the Christian builders
of Syri a.
" CATHEDRAL" : Date 5 5 7 A. D. The ruined basilica which we have called the
" cathedral " is by far the largest of the churches of Umm idj-Dj imal and stands in the
most commandi ng positi on, free from other buildings, i n the great open space, or common,
which occupies the middle part of the city, and upon the mai n street which extended
t Ill, insc. 263.
Il l . 1 5 6 . Sout her Hal f o f Doubl e Church ; Vi ew of West Front from t he Street.
Il l. 1 5 7. Interi or of }orthern Half of Doubl e Church ; West End and North Wal l.
U mm i dj -Dj imal ( Thantia ?)
southward from the Gate of Commodus, passing the Praetori um, to the Barracks. It
is a pity that this largest of the churches, and one of the two that i s dated by an
inscripti on, shoul d be i n so rui ned a conditi on. Its plan (Ill . I S 8) may be easily traced ;
but the form of the superstructure can only be conjectured from fragments i n the fallen
masses of stone. One storey of the west faade (Ill. I S 9), the wall of the north aisle
and a part of the apse wall, are standi ng ; but the entire south wall and all the i nterior
arches have fall en. The date \ in large l etters, readi ng only E TS t Y N A , was found
upon the fallen cap of one of the pi ers of the apse arch.
The plan (Il l . I S 8) is that of a nave of four broad bays, the arches being carri ed
on high oblong piers, and a deep semi-circular apse between side chambers, making a
straight east wall , and a shallow narthex closed at both ends and having fve arches
carri ed on columns. Adj oi ning the north wall of the church, at its east end, i s a low .
structure comn1unicati ng with the south aisle ; thi s may have been a sacristy or an
outside chapel of some sort. The whole plan is such as one would expect to fnd in
Northern Syria ; but whether the s

uper-
:
tr
::rn
w
:si:
i
::d
o
:
t
as
o
oe i:e:
imiIr
in wood, one cannot determi ne defni tely ;
though the comparatively smal l amount of
debris lying i n the nave would i ndi cate
that nothing more than the nave arches
and a clearstorey wall had fallen here.
At some period in the church' s his
tory the west wall was reinforced by the
construction of an i nner wal l directly withi n
3.70
: o:::
the origi nal one : new lintels were placed
i nside the older ones of the three door
ways on a slightly lower l evel . One of
these was not a newly cut stone ; but was
a lintel from another and older building.
tATHEDOeo DAL
Ill. 1 58.
This i s the lintel with the famous burgus inscripti on 2 upon it, already referred to as
having been perhaps the li ntel of the gateway of the Praetori um. It was set up In
the reign of the emperors Valens, Valentinian and Grati an, about 3 7 1 A. D.
SOUTHWEST CHURCH. This church stands on the northwest l imits of a compact
group of buildings in the southwest angle of the city. The other buildi ngs of the
group are of a domesti c residential character, and though they are well preserved, and
the plan of the whole group i s easily traced, yet I am unable to say, with any degree
of certai nty, whether the houses were dependencies of the church, and formed the
residential part of a monastic i nstituti on, or whether they were the homes of i ndividual
citizens which grew up i ndependently in the space between the church building and the
angle of the city wall . A third sol uti on of the problem, dependent upon questi ons of
land . tenure, and of church proprietorship, in Umm i dj-Dj i mal between the fourth and
the seventh centuries - questions about which practically nothing i s known -, would
be to assume that the land belonged to the church, and the householders either rented
1 III. insc. 260. 2 TIL insc. 233. ladd. insc. 2058.
Division II Secti on A Part 3
the lana and built their houses upon it, or rented houses built by the church. The
houses are of the better class of domestic architecture found in the city, superior i n
construction and fnish to the architecture displayed i n the church edifce itself. The
church building is quite well preserved (Ill . 1 60) ; the high outer walls are almost i ntact ;
but the i nterior arches and the half-dome of the apse have collapsed, though the piers
of the apse and part of its wall are standing.
The ground plan of the church (Ill . 1 6 I
)
is that of a simple basilica, its nave
divided into three aisles by two arcades of three arches each, its semi-circular apse
being set deeply back and showing its whole proj ection ; the only singular feature being
the presence of two doorways i n the east wall, one on either side of the apse, and in
the end of each side aisle, exactly where the doorways ' to side chambers would be
placed ; but openi
t
g out of doors. There are two other entrances, one i n the north
side and one in the south. Upon this ordinary plan was erected a type of basilica
peculiar to the I:auran ; the outer walls were built up to the level of the crowns of
the arches of the nave, and the whole bui ldi ng was covered with a fat roof all on
one level . The side aisles were certainly roofed with slabs of stone ; and I am i ncl ined
to believe that the middle aisle was covered i n a similar way ; but al l the roofng
materials of this church, even most of the arch stones, were carried of for the con
structi on of a bridge on the line of the Hedjaz railway. The width of the mi ddle aisle,
5 . 20 m. was not too great a span to be roofed with slabs on proj ecti ng corbels i n the
fashi on of the country. The corbel course of the side aisles and the parapet above
them are to be seen i n the photograph (Ill . 1 60) at the extreme right. The picture
also shows a r
\
w of brackets i n the south wall , which carried the stone beams of a
portico, and the quality of the masonry, with remnants of stucco still adhering to it.
Attention should be called parti cularly to the fragment of ornament i n stucco which
appears on the exterior, over the doorway beside the apse ; it is a l ittle mouldi ng
describing a trefoil, whi ch gives the appearance of a cusped arch over the entrance.
The windows are mere slits placed high i n the side walls.
NORTH CHURCH. There are two churches of the basilica class, both of medi um
dimensions, i n the northern part of the city. One, the North Church, i s shown on
Map No. I ; but it i s too for north to be i ncluded i n the l arge map. This church
(Ill . 1 62) is of diferent proportions from the church described immediately above ; its
nave is more nearly square and its middle aisle is wider in proporti on to the width of
its side aisless ; but otherwise the plan i s quite the same, but for the doorways in the
east end of the side aisles, which are peculiar to the other building. The north aisle
i s considerabl y narrower than the south, - a peculiarity found i n two or three other
churches i n these ruins. The chief entrance was a rather small doorway i n the west
end, and another entrance opened southward from the easternmost bay of the south
aisle. West of this entrance a long low building (See Sect. A-B) extended along the
south wall of the church as far as the west wall. Thi s low structure was divided i nto
two chambers, the westernmost openi ng i nto the church by three spaces divided by
monolithi c piers, the other, much smaller, bei ng entered directly from the south aisl e.
The apse wall and the chancel arch of the church are both standing, but the half
dome has fallen i n. The i nterior arches of the nave have collapsed, carrying the roof
with them. Thi s roof certainly consisted of fat slabs laid on corbel courses over the
aisle. It seems probable that the middle aisle was provided wi th a fat roof of wood,
Ill. 1 5 9. Facadl: o f the " Cathedra1 . "
Ill. 160. Southwest Church at Right, - Buildi ngs on Court at Left ; Vi ew from t he Southeast.
(
`
/_
l |
Umm idj-Djimal ( Tlzantia?)
z :
. ' .
I/ Lll 'LC `llIt
A
I ' I,, ' ,
l Mo:
oC
| =
gO i . ?

m. lD
I
.
I
|
I :
1 ---__ 1B
' 76 :
i

:
I
r
41 |MC+
f -Q.x 68 9C=>1. 274
SOVTHWEST
CHVRCH

. .
'
,
~
( I
.Ll U l|
Ill. I 6 I .
^
J
\
.
.
_ J
1 86 Di vision II Section A Part _
on the same level as the aisle roofs, and covered with the same protective layers of
cinders and beaten clay.
1 C
SECTI ONA-B ETOIED
NOTH CHVRCH
Il l . 1 62.
NORTHEAST CHURCH. Almost precisely si milar to the building just described, . n
ground plan (Ill. 1 63) and superstructure, this church also preserves its apse wall and
apse arch, but not its half dome. Three of the four i nterior piers, made of well
dressed stone i n small blocks, are preserved i ntact, and the outer wall stands to i ts
full height on two sides. The west entrances are two ; one at the end of the mi ddle
aisle, the other at the end of the south aisle. The long, low building along the
south wall i s entered by a single doorway and i s not divided as i n the N orth Chuch.
The entire building, with the excepti on of i ts piers and arches, and i ts door and
A -
window frames, i s built of
roughly quadrated work ; but
the walls were coated with
stucco on both si des, as one
may still see i n the lower parts
of the apse. An i nteresti ng
detail i n thi s part of the build
i ng is a stone bench, - 50 cm.

wide and apparently of about

5 EC TI ON A-B the same height -, which ex-
LL
tended all around the apse,
NORTHEAST
CHRCH
from pier t o pier. The bench
was built of roughly quadrated
E

blocks, and was originally

-
Il
-
l
. I
6
-.

cov

ed w
h
h ucro. So fr
as I could see, the bench
was not broken i n the middle for a seat of special dignity, like the bishop' s throne which
was ofen given this place i n early apses. This bench was a feature in all the apses of
Southern Syria ; though i n the great majority of cases they have been destroyed, or at least
concealed by the collapse of the masonry half-domes, almost all of which have fallen. A well
built bench, i t will be remembered, is still to be seen in the Church of St. George at Zorca
in the Jauran, one of the few churches in which the half-dome of the apse is preserved.
Ill. 1 64. West Church ; Facade from the West.
Ill. 1 65. \Nest Church ; Vi ew from the South.
Umm i dj -Dj i mal ( Thantia?)
WEST CHURCH. One of the largest, and quite the best preserved, of all the ffteen
churches of U mm i dj-Dji mal , and one which was apparently the most i mposing and the
most beautiful , is the church near the gate of Commodus, j ust outside the west wall
of the city. Though outside the city wall, it i s connected with the city by one wall
that connects its east end with the Roman gate and another that extends southward
from i ts southwest angle and turns to j oi n the city wall north of the Praetori um. A
ground plan and a photograph of the church were published by Schumacher 1 i n 1 897 .
I have made only one, slight correcti on i n his plan by deepening the apse, adding a
cross secti on and a l ongitudinal secti on and two new photographs to what has already
been published. The west end preserves two storeys . quite completely (Ill. 1 64) . The
south arcade of the nave is standi ng with the clearstory wall above i t (Ill . 1 65) , the
east end is almost complete, the crown of the half dome only having fall en, and a
large secti on of the south aisle wall still stands with a portal i n it.
In taking up the study of this bui ldi ng, we fnd ourselves i mmediately transported
from Southern Syri a, from the lauran, to the regi ons of Northern Syri a, for in ground
pl an and i n superstructure, thi s church i s foreign to the regi on i n whi ch i t is located ;
but conforms to plans and methods of building which are common i n the North. Even
the black basalt of which i t is made does not stamp i t as a product of the South ; for
this material is found i n many northern churches i n the basalt country between the
mountains and the Euphrates, i n the cAla, at Kerratl n, i l- Anderln, and the Djebel il-la.
The plan (Ill. 1 66), i t will be observed, consists of a well proporti oned nave divided
i nto aisles by two arcades of four arches each, carried on square pi ers. The
d
eep-set
apse
i
s fanked by spacious side chambers, giving a straight east wall to the building,
and the west end has two square towers with an arched narthex between them. The
broad arches of the nave were built of very shallow voussoirs, i. e. , voussoirs which
were l ong in the sofft but very short from extrados to introdos, and these were brought
together upon small i mpost blocks, not nearly so large as the piers upon which they
rest (See Sect. A-B and C-D). The tall clearstorey wall above the arches was pierced
with square-topped windows, one above each of the arches of the nave. That the
roof of the nave was gabled may be seen from the fragment of a gable still i n place
and shown in Ill. 1 64. The west wall was pierced on this l evel by a large round
arched window. The half dome of the apse was built of concrete with volcanic scoriae
in it. The side aisles were roofed with fat slabs of stone resting upon the aisle walls
and upon a corbel course above the nave arches which may be seen i n Ill . 1 65 .
This whole system o f roofng i s exactly like that empl oyed i n the church at Ialb Lauzeh
in Northern Syri a.
A great part of the church edifce, i ncluding the aisle walls and the l ower storey
of the west facade, in additi on to the piers and arches, was built of highly fni shed
quad rated work, and the presence of a Classical moulding dividi ng the storeys of the
facade suggests that all this beautifully fnished material was taken from some building
of the Roman peri od. The whole of the clearstorey, the upper part of the facade, and
the walls of the chambers beside the apse were built of smooth quadrated masonry
of good quality. The towers which fank the west facade, and which originally had an
arch between them, like the western towers of the church at I:alb Lauzeh in Northern
1 Z. D. P. -V
'
97 pp. 1 5 7, 1 58.
I bb Division Section A art
Syria, are built of roughly quadrated masonry, and are certai nly l ater than the rest of
the church ; because the highly fnished wall of the facade and its Classical moulding
are visible behind the walls of the towers at the points where they join, as may be
seen in Ill. 1 64. Schumacher' s plan shows buttresses in the west walls, outside, a
t
the p oints opposite the ends of the interior arcades.
:
; I I
' :
I I
I I

I :

, .

: :

. ,
: J

I
: :
, '
, .
, I
-
., '
I '
: :
, .
. ,
:
WLST
CHVRCH
3E CTI ON A-B
Ill. 1 66.
The i nterior ornament of this church is more profuse and more interesting than
that of any other of the churches in this place. The caps of the piers of the nave
have mouldings of excellent profle, well executed ; though the piers have no base
mouldings. The soffts of some of the arches are ornamented at the springing with
carved discs contai ning the cross and other Christian emblems ; one of these may be
seen in Ill. 1 64, by looking through the western doorway to the frst arch on the south
of the apse. The entire foor of the nave was paved with mosaic work of geometrical
Umm i dj -Dj i mal ( TJantia?)
patterns i n various bright colours. Wherever thi s pavement was uncovered it showed patches
In cement, showing that the pavement was of good age before the church was abandoned.
LJl JJ
. / RRSTOBD
4. 1 0
. _
_

KLAVDIAOS
CHVRCH'
/ '
(
" .
SECTION C-D RESTORED
Il l . 1 67.
.
_ .
.
CHURCH OF KLAUDIANOS. The na.me which I have gIven to this church was taken
Publications of the Princeton University Archaeological Expeditions to Syria, Div. II, Sec. A, Pte 3
Di vision II Secti on A Part 3
from a Greek i nscri ption 1 on the lintel of the mai n west portal, which names a citizen
Klaudianos, who was the buil der, or at least a benefactor, of the church. The building
forms a part of a group of houses almost directly east of the Gate of Commodus. The
church faces upon the open space which extends north and south through the mi ddle'
of the city, and stands on the north side of the group of houses whi ch were i n all
probability the residential buildings of the clergy connected with the church ; they are
hardly extensive enough to have forn1ed a monastery proportional to the size of the c
h
urch.
Again we have a three-aisled basilica ( I ll . 1 67) with apse and si de chambers
making a fat east wal l , three west portals, and a narthex of four col umns closed at
the ends. In the nave the supports of the four longitudi nal arches on either side of
the mi ddle ai sle are columns, and the responds at the ends of the arcades are columns
placed against the walls. The entrance to the prothesis, which i n this case i s on the
north of the apse, i s an arch. There seems 'o have been a porch of two columns
outsi de of the north portal .
The structure is in a sadly ruined condi ti on ; the apse wall and the chambers beside
the apse are preserved, the west wall i s standing (Ill. 1 68) well above t
h
e lintels of
the portals, and two of the col umns of the i nteri or are i n place ; but the side walls of
the church and the i nteri or arches have fallen, fi lling the nave wi th the ruins of arches
and roof.
It seems i mprobable that there were galleries above the side aisles. Si nce
the three aisles are of about the same wi dth, the roof must have been of stone, fat,
and continuous over the entire building.
There i s a large court, or cloister, at the southeast angle of the church ; this i s
bounded by buildings on all sides. Those at the west and along the south wall of the church
form an i nteresting group built of well fnished stone. The ph

tograph (Il l . 1 69) shows

the rui ns of the faade of the church at the left and the arch of one of the large buildings
south of the church. One of these large arched apartn1ents i s fnely fni shed in the
i nterior, and is provided with a latrz'na. The entrance to the cloister was a double
arched passage on the east side, fanked by buildi ngs one storey high, and havi ng a
lJ orter' s lodge i n its upper storey. These buildings certai nly belonged to the church,
and, with i
t
, may have constituted a small monastery ; but it is quite as probable that
they were the residential quarters of the ordinary clergy. I am of the opinion that
the monasteries of Southern Syria were isolated, l i ke that at i d- Dr, 2 and that these groups
of ecclesiastical resi dences about the city churches were
CHPE L or B.ARACKS
the homes, not of monks; but of the regular priesthood.
.]LJ |l L
Ill. 1 70;
I
I III, insc. 261 .
CHAPEL OF THE BARRACKS. The small church,
or chapel , attached to the barracks has been discussed
i n connecti on with the greater building of which it
was a part, on page 1 70. I wish, i n this place,
only to call attention to some of its more i mportant
features so that they may not be omi tted from the
general di scussion of the religious architecture of the
city. The plan of the church and a longitudi nal
secti on on a small scale are given in Ill . 1 44, to these
I S now added a cross section on larger scale (
Ill . 1 70) .
2 H. A, 2, p. 1 02.
Il l . 1 68. Farade of Klandi anos Church ; Vi t; w from the West.
Ill. 1 69. View of Klaudianos Church and Bui ldi ngs South of it, from the West.
Umm i dj - Dj i mal ( Thantia?)
The outer walls are quite completely preserved, the arches of the nave have fallen,
but their piers are standing. The unusual part of the church i s at the east end, where
there is a chamber of the ordi nary kind on the north of the rectangular sanctuary, but
no chamber on the south side, the space for the prothesis being di vi ded from the sanc
tuary and from the end of the aisle by arches. The arch of the sanctuary and these
two arches of the prothesis are still standing, all three being carried by a si ngle pier.
The remains show conclusively !u! the whole church wa covered by a fat roof at one
level , the si de aisles certai nly with slabs of stone, and the middle aisl e i n the same
manner, in .ll probability. The roofs of the sanctuary and side spaces were also of
stone and on the same level . The i nteri or still preserves bi ts of plaster i n three coats
upon its walls, and shows other signs of having been well fnished. The exterior was
also covered with stucco, but few traces of it are now to be found.
CHURCH OF NUMERIANOS. North of the barracks, and i n the middle of the open
common, I S this i nteresting church, built, accordi ng to the i nscripti ons, 1 as a vow by
Numeri anos, Johannes and Mari a, and which I have cal led the Church of Numerianos
for the sake of brevity. A compact group of residential buildings surrounds a small
cloister court on the north. The ground plan (Ill . 1 7 I ) presents nothing that is un
usual i n church plans, and it i s onl y when the superstructure is exami ned for resto
rati on that the unique features of the building become evident. The plan i s that of
the ordi nary three-aisled basilica with apse and side chamber givi ng a fat east wall,
and wi th a colonnaded porch at the west. The supports of the broad arches of the
nave are square piers, two on a side, to carry the six arches. The frst unusual feat
ure to be noticed is a wall-arch above the central west portal , springing from one to
the other of the piers at the west end. The opening of the apse has been closed
by a wall built probably by Arabs after the rise of Islam ; the front wall of the cham
ber on the south side of the apse was apparently i n part rebuilt at the same time.
The west wall i s completely preserved i n its lower storey, and, at its north end,
a section of the upper storey wal l , crowned with an angle pilaster cap of the I oni c
order, gi ves the height of the wall at that poi nt. The enti re south wall and the south
arcade of the nave have collapsed, and the south chamber besi de the apse is in ruins ;
but the apse wall still remains up to the l evel from which the semi-dome sprang ; the
chamber on the north side of the apse i s completel y preserved. The northern half of
the church preserves much of its original form i ntact (Il l . 1 7 2) , enough, i n fact, to give
ample material for a complete restoration of the nave. The three arches on this side
of the nave are not only standing, but they still carry a large part of the clearstorey
wall i ntact to its full height, and the stone roof of the north aisle, which is composed
of well ftted slabs. This, of course, i nvolves the preservation of the north aisle wall.
Secti on C-D i n Ill. 1 7 I , shows, by shading, j ust how much of the clearstorey is i n place :
at . the east end its full height is given, the remai nder shows that there was a rect
angular wi ndow centered over each arch, and one over each pier. There were corbe
l
s
i n the wall above the windows, which had to do with the support of a wooden roof
above the middle aisl e. The curi ous and i nteresti ng feature of this ,church i s that the
apparent clearstorey windows did not open to the l ight of day, but i nto a gallery above
the side aisl e. The wall . of the side aisle was carried up in an upper storey equal i n
1 III, inscs. 257-259.
i visi on !! Scction A Part
height to the wall above the mai n arches. This wall was pierced with wi ndows opposite
those i n the clearstorey. The opemng which may be seen 1 0 the photograph (Ill. 1 7 2)
J .
! ..
l8
\-1 l. 80
.8
:
:




6.
1\JL'
,
" - " )
')
/
.ECTION C-D -
'CHCH UNtR IOS ATVM I DJDJlM
Ill. 1 7 I .
partly concealed by the ruins of the front wall of the nave is one of these windows
and i s not i n the same wall with the other two windows as would appear at frst sight.
This outer wall terminates toward the west i n a sort of pilaster as may be seen in
Ill. 1 7 2 and in the drawing of the fa<ade (Ill . 1 7 I ) ; the lower courses of another pi
laster, or pier stand at the end of the clearstore wal l . I have drawn a column i n the
Il l . 1 7 2. Numerianos Church ; Vi ew from the Southwest.
Ill. r . Numerianos Church ; Part of Fa\ade, Vi ew from the \Vest .
U mm i dj - Dj i mal ( Thantia ?) i
>1
space between the two pilasters, havi ng found a small col umn i n the rmns of the porch.
The gallery thus fornled was roofed with slabs of stone, the corbels for which are still
visi ble. It i s most probable that the south aisle i s to be restored exactly like the
north aisle, as I have done in Ill. 1 7 I , Section A-B. It was possible to pass from
one gallery to the other by a narrow platform corbelled out above the transverse wall
arch at the west end of the main aisle, as is shown i n Secti on C-D, Ill. 1 7 I .
I have restored the remai nder of the facade from small columns and other broken
details lying in the porch, on the basis of certain facades i n Northern Syria i of the
a
rchitrave and cornice there can be little doubt. The porch was a free-standi ng col
onnade carrying a roof of slabs : the walls, which are now found i n ruins at the ends
of the porch, are late additi ons. The I onic angle-pilaster cap from the south end of
facade is now lyi ng near the southeast angle of the chapel of the Barracks. The entire
lower storey of this church, i ncluding the outer walls and the pi ers and arches of the
i nteri or, was built of the most highly fnished stone, admi rably laid and fi tted (Ill. 1 7 3) ;
the upper parts were built of smooth quadrated masonry i n blocks of unusual size.
The mouldings of the main west portal , which are very simple and the only mouldings
visibl e in the church, have the appearance of not having been made for the place
which they now occupy i for the moulding of the lintel i s not returned and brought
down at the ends of the li ntel i but the i nscripti on of N umerianos i s complete and i s
undoubtedly coeval with the building. The bases of the columns of the porch were
well moulded Attic basis, but the capitals were i n plain Dori c style.
The ecclesiastical buildi ngs on the north side of the church ( Il l . 1 7 I ) are grouped
about a nearly symmetrical court with a circul ar cistern i n the middle i the cistern was
.
origi nally covered with slabs level with the pavement of the court. The buildings are
all small , only those adj oining the church and those on the west side being two storeys
high ; a tower of three storeys occupied the northeast corner. The two rooms adjoi n
ing the church, with a passage, or sl ype, between them, are still preserved i n part
and have a stone ceiling a little l ower than the ceiling of the side aisle withi n. Upon
the outer wall of these rooms there apparently stood a colonnade of columns about
_ In. hi gh carryi ng a roof of wood. None of these columns i s i n situ ; but there are
several i n the ruins, and I cannot fnd any other place where they might have stood
than that I have shown i n Secti on A-B, Ill. 1 7 I . This open l oggia would have been
reached by a long fight of steps leading up from the court along the east wall, and
it would have been necessary to have steps to reach the gallery of the church from
the open loggia. Most of the rooms in the court were for resi dence, one was a l ong
stahl e with fve mangers. Beside the doorway i n the southeast angle of the court, at
the left hand, was a bracket, i n the form of a di nlinutive altar, with a slight depression
in i ts top and an i nscription 1 upon its si de. Similar altar-like brackets 2 were found be
side other doorways i n Umm i dj- Dj i mal ; I believe that . they belonged originally to
.
Pagan buildings, but I am unable to suggest what their use may have been. This is
one of the few buildings i n Umm i dj -Dji mal , if not the only one, which shows un
mistakable evidence of Moslem occupati on. No demonstration is required to prove
that Moslem hands built the wall which closes the apse, and there i s little doubt, if
the south wall of the church were not in ruins, that we should fnd it pierced to receive
1 III, insc. 240 2 . p. 21 I, Ill. 1 93.
1
9
4
Di vision' - lr ' Secti on A Part 3
a mi brab, or its portal ftted with the Mohammedan prayer niche. It may well have
been that the Moslem occupants of the city were nomads who lived in tents and did
not i nhabit the houses of their Christian predecessors , and that this church was the
only building restored or altered for Moslem uses.
DOltestic Archztecture.
The greater part of the ruins of Umm idj-Dji mal are, naturally, those of domes
tic architecture. Al most the entire area within the wal l s, excl usive of that devoted
to the open common i n the mi ddl e of the city, and of that devoted to the churches
and their courts , is occupied by private residences i n a better or a worse state of
preservation. These houses represent such various classes of domesti c architecture,
as may be indi cated by diferences of si ze, qual ity of constructi on, and mi nor details
of luxury. They are, for the most part, grouped i n irregular blocks divided from
each other by narrow, crooked streets and lanes. The houses i n a block . usual ly
open upon one large court-yard, though, in a number of cases, two or three courts
are found in a block of unusual size separated by rather narrow buildings. The
outer walls present plain unbroken surfaces to the streets, with no windows except in
the upper storeys . The entrances to the court-yards, usually two i n each, are occasi on
ally arched, and several of the houses have towers which rise one or two storeys above
the surrounding roofs. The doors, and most of the windows, the staircases and bal
coni es, i n fact all of the more attractive features of the houses, are to be seen only
from the court-yards. The exteri ors were severe and forbi dding i n the extreme, j ust
as one fnds the case to have been i n those resi dential quarters of Pompei i i n which
the private insulae were not surrounded by shops ; and for plainness they were not unl ike
the street fronts of Engli sh and Ameri can city houses of the 1 8th and early 1 9th
centuries. All the di ferent qual i ties of stone work and wall bui lding that have been
descri bed as belonging to the architecture of the Chri stian period i n the JIauran 1 appear
in these houses most of which are two storeys hi gh, though there are bl ocks of one
storey hOllses, and many resi dences of three and even four storeys. The n10st common
form of hOllse i s that in whi ch two hi gh arched storeys composed of l arge rooms are
combi ned wi th four storeys of smal ler rooms, all under one fl at roof; like the maj ority
of houses in other parts of the Southern Hauran, 2 except that the smaller rooms are
placed at the si de of the large arched rooms instead of behind them. The transverse
arch, the corbel courses and long stone slabs for roofs and i ntermedi ate foors, described
in a former Part,3 are the i mportant features of constructi on in all these houses, and
determi ne their form and styl e. Stucco on the outsi de, and several coats of plaster
withi n, gave a fnish to these buil di ngs which the present appearance of their rui ns
woul d hardly suggest. It i s qui te unnecessary to draw an odi ous comparison between
the appearance of these ruins and that of wrecked houses in any great modern city.
The l arger of the two maps of the city presented herewith shows twenty groups
of private residences, wi th detailed ground pl ans measured and drawn to scale. Al most
as many more houses were in a state 6f preservation to have been added to the map
i n the same manner, i f there had been ti me to devote to the l ong and son1 ewhat
tedi ous process ; most of these are given .on the map as blocks with shaded sides, the
I II. A, 2
,
p. 67. 2 Ibi d. pp. 1 20
,
1 22. 3 II, A. 2, p. 68.
Umm i dj - Dj i mfd ( Thantia ?) 1
9
5
outlines having been measured and drawn to scal e. These, taken together with the
blocks drawn i n detail , would have represented about two-ffths of the resi dences of the
town ; the other three-ffths, situated chiefy i n the north and northeast quarters, have
been so completely demolished that it would be very di fcult to secure plans of the
houses, or even trustworthy outlines of the blocks. I esti mated, on the occasi on of a
visit to the ruins i n . the spring of 1 909, after this map had been drawn, that there
were origi nally from 1 30 to I S O blocks of houses si milar to those given in detail on
the map. The blocks presented i n measured drawi ngs on the map were chosen to
represent as many types of houses as possible, or by reason of some particularly
interesti ng feature, and not entirely because they were the best preserved ; for there
are a number of blocks given here only in outlines that contain houses i n a remarkable
state of preservation. It will be noticed that the houses are nUlnbered, not by blocks,
but by court-yards, each court beari ng a Roman numeral : the houses grouped about
the churches are not numbered. It seemed to n1e that, having so complete a pl an of
about one-ffth of the resi dential porti ons of the town, it should be possible by some
means to arrive at an approximate estimate of the population of the . city i n the ffth
and sixth centuries. In attempting to make such an esti mate it would be necessary to
know how many separate residences were included in a block of houses. A careful
examination of many diferent blocks suggests that, ordi narily, each fami ly had one
large room, usually arched, with several smaller rooms, square or obl ong, disposed
about it and openi ng i nto it ; this seems parti cul arly probable i n many groups in which
we fnd one latrina to every large rOOln. In cases where there were two or more
storeys I esti mated both, or all , the storeys as belonging to one fami ly, and, where
two large arced roms had a connecting doorway, I gave both to a fami ly which
might be supposed to be . of somewhat larger means than a family which had only one
such room. According to thi s scheme, there woul d have been from eight, to ten fami li es
i n each block, and, taking seven as the average n umber i n a family there would have
been from ffty to seventy persons tO
I
a block, which seems a conservative esti mate,
taki ng i nto consi derati on the fact that a Inajority of the blocks are composed of houses
of more than one storey. The whole city, then, with its one hundred and ffty blocks,
would have contained a popul ation of from seven to ten thousand persons within the
wal l s. The houses of most of the smaller towns of the Southern Hauran, were usually
provided wi th stables which were a part of them ; i n U mm i dj - Dj i mal , on the contrary,
only a few houses have stables, and there is no evi dence to show that any considerable
number of domestic ani mals were housed within the city gates. For this reason the
above esti mate of the probable populati on i s rather conservati ve, i t allows, moreover,
for a degree of l uxury which is to be thought of as a modern development, providi ng,
as i t does, at least one room to each member of a fami ly which is probably too much.
I n anci ent ti mes, the lower and mi ddle classes, the tradesmen and merchants, every
body, i n fact, except royalty and the great families, lived in far more cramped quarters
than the same classes do today, and I believe it is not i mprobabl e that, at the height
of i ts i mportance as a city, Umm i dj -Dj i mal could boast of a populati on of perhaps
ffteen thousand souls.
These extensive rui ns of domestic architecture undoubtedly represent houses of
several centuries ; but, so far as one may j udge by comparing them with bui ldi ngs of
known date, there are probably few houses standing whieh are earlier than the latter
Dlvlslou II Sectlou A Part 3
part of the frst century A. D. There are probably many which were built during the
second, third, and fourth centuries ; but the great mass of private architecture here is
undoubtedly to be assigned to the ffth and sixth centuries. It will be seen that many
buildings of one class or kind of masonry have been repaired, or added to, in masonry
of another sort. The repairs and additions are, of course, later than the original
building, and, as it h
a
ppens that they are constructed in a poorer manner, i t is safe
to assume that whole buildings built in the poorer style are later than those which are
better built. Furthermore, those buildings which I believe to be of the more remote
dates, besides being constructed in a better manner, do not contain i nscribed stones
placed at random in their walls ; while the buildi ngs which represent the poorer class
of work often contain Pagan inscriptions and old grave- stones worked in in hap-hazard
fashion.
Owing to the absence of dated inscripti ons upon the houses, and the consequent
diffculty of determi ning their dates, in spite of the diferences of construction, it is not
possible to take up the description of the residences i n chronological order. I shall
begin with an example of the domestic architecture of the town, i n whi ch the plan is
vey simple, and the superstructure typical of a large class, and. then take up the houses
of more compl icated ground-plan.
HOUSE No. XIX. Thi s house is probably to be considered as a single residence
of the better class. It occupies the east side of a small court, the south and west sides
of which were once occupied by other houses which are now fallen in ruins. It had
fourteen rooms, yet it is so planned that it could have been divided for the use of two
8.30
2. 83

HOVstNX
5LCTI ONA-B ESTOED
Ill. 1 74.
families, only on the plan of a modern fat, allowing a storey and a half to each family.
The ground-plan (1
11
. 1 74) speaks for itself; it shows a l arge arched room, facing west,
with three small rooms in a row on the north of it, one of which protrudes beyond
the west wall of the large room, and with a long narrow room on the south. The
transverse arch of the l arge room raises the ceiling to such a height that there is room
ll l . 1 7 5. HOLlse No. XIX, Vi ew from the SOLl t hwest.
I l l . 1 7 6. House No. III, View from t he Suuth.
Umm i dj - Dj i mal ( Thantia?) 1
97
for two storeys of corbel constructi on in the three small rooms, within the same space.
The upper rooms were reached by two stairs, one i n an angle of the l arge room and
. one i n the court outside. The pl an of the upper part of the house is a repetiti on of
that of the l ower ; a second arched chamber being placed above the frst and a third
pair of small rooms in two storeys on the north of it ; but the projecti on was complete
in two stori es, its fat roof serving as a terrace for the higher rooms beside it. A con
siderable part of this house i s still standing as the photograph (Ill. 1 7 5 ) shows ; that is,
the l ower girder arch, and the pi ers and springers of the upper, with all four storeys
of the rooms in the northeast angle, and three storys of the i ntermediate rooms. The
projection preserves its two storeys i ntact, with stone foors and roof. A few of the long
slabs of the roof of the upper arched chamber are to be seen i n the photograph. The
greater part of the front wall of the house i s standing ; but most of the steps of the
stairs which where corbelled i nto the walls, have been broken of; yet it i s still possible
to see that the outside stair led up to the upper doorway on the north si de of the
court, and then mounted by two steps to a platform in the angle, from which it led
on, by another fight, to the upper storey of the mai n part of the house. The i nterior
stairs are likewise traceable in stumps of corbels proj ecting from the walls. It is worth
while to call attention to a bracket adorned with good Roman mouldings, which i s one
of two that supported a shadestone above the doorway of the upper arched room.
This house i s the best example i n the ruins of a type quite common i n Dmm i dj-Dji mal,
viz. , the four-storey and two-storey house i n one.
HOUSE No. I I I .
.
An extension of the plan illustrated i n the house j ust described
is found in this residence (Il l . 1 7 6) which occupies the north side of a large courtyard,
in the southwest quarter of the town. The plan (Ill. 1 7 7) shows parts of buildings on
the east side of the court, but these, with the residences on the south and west sides,
are i n rui ns. The part I shall describe i s i n a remarkable state of preservati on, and
represents the best quality of domestic architecture i n the city. The arrangement of
the house on the north side of the court gives a double set of arched rooms i n two
storeys with four storeys of narrow roonlS between them, and three storeys at the wes't
end. The larger pair of arched roonlS, with the three-storey section of the house on
one side of i t and the four-storey secti on on the other, originally constituted a si ngle
residence ; the smaller pair were added to this at a later period and in poorer masonry,
as the photograph (Ill . 1 7 6) pl ainly shows. This earlier residence consisted of two l arge
rooms and seven smaller ones. The foor of the l owest storey of the four-storey part
was sunk below the l evel of the arched room in order to give suffcient height for
three foors above i t. Thi s sunken room (Ill . 1 7 7 , Sect. C-D) has a ni che i n its north
end with an opening i n the wall above it, like a rudimentary chimney, and may have
been a kitchen. It is not directly connected with the large room. The foors above
this room are of about equal height ; that directly above it was reached by an exterior
stair, the third opened upon the upper arched room, and the fourth was probably reached
by a stair i n the northeast angle of the large upper room. The lowest room on the
other side has a very low ceiling, i t has an outer doorway and one connecting with the
l arge room. Above i t i s a chamber reached from a landing in the main outer stair.
The third, or uppermost, chamber opens upon the upper arched room, and i ts ceiling
i s as high as that of the arched room. The photograph (Ill . 1 76) and the elevation
(Il l . I 7 7) show the mai n outer stair corbelled out from the walls. It begins on the
Publications of the Princeton University Archaeological Expeditions to Syria, Div. II, Sec. A 1t. 3.
Di vision I I Section A Part 3
west side of the court, rIsmg to the plqtform just menti oned, and then, ascending to U
"
' ' )
'
.
` (.
'
/
-
--

` \l !L`
$
+

.. r
( )
)
SLCTI ON C-D .
HOVSL Ngr
Ill. I ] ].
-
1.
- -.
|

. _l
c
;

. `
., - ' \
(
'. ~
platform before the door of the upper arched chamber, continued to the roof. The
great arch of the
l
ower room i s stil
l
i n place ; the upper arch is represented only by
Umm i dj -Dj i mal ( Thantia ?) 199
its piers. The main room on the ground-foor was l ighted by three windows in its front
wall, a circular openi ng over the door, square within, to accommodate a shutter, and
a large rectangular window on either si de. The newer part of the house, whi ch perhaps
constituted a separate residence, consists of two arched rooms, one above the other, the
foor and roof of the upper one being a little lower than those of the adj oi ni ng resi
dence. The l ower room was lighted in a manner typical in the Southern Jauran,
having a hooded window directly over the l intel of its doorway. The room adj oi ni ng
this, on the east side of the court, probably belonged to the same resi dence. The
upper fl oor of the later resi dence was reached by a stair partly corbelled out from the
wall and partly bui lt up as a soli d structure. The hoods represented over the doorways
in the upper story in the drawing are based upon hood-stones which were found i n
the debris i nsi de the house.
HOUSES XVII and XVI I I . The most i nteresti ng block of houses on the east side
of the city (Ill . 1 7 8) i s that which compri ses the two court-yards marked XVI I and
XVII I on the maps of the city. Number XVII is bounded on two sides by narrow
'
I

-
1 . 1 3
'1.20
9.40 XVI I
/
1. 1 1
10 I / : _ ~
/ R
I I
, I
' `
1I!. 1 78.
7

\
B
streets, and on the third side by a small enclosed garden belonging to number XVI I I .
Number XVI I I itself gives upon a street t o the south, and upon a small open square
to the west ; but its east wall faces a broad space just inside the east wall of the city
a li ttle to the north of the mai n East Gate, occupied in part by an open cistern.
House XVI I is uni que among the houses of the city ; it is not planned like an ordinary
residence or group of resi dences. Its courtyard would be symmetrical but for its west
2OO , Di vision II Section A Part 3
wall, l. e. the wall separati ng it from the street, which is brought out at a sharp angle
to accommodate a vestibule i n the southwest corner. Withi n the vestibule, i mmediately
to the left, and facing the court on its south side, is a building about 8 m. by 1 0 m.
square ; its lower storey, entered by a narrow doorway, i s divided transversely by walls
into three oblong chambers. Its upper foor, which I have chosen to represent i n the
plan (Ill . 1 7 8, R) had an open l oggia of four columns above tLe entrance, and two
broad arches above the transverse walls below, making one large room, larger by half
than the l argest of the arched living rooms in ordinary houses. Across the court, in
the angl e, stands a tower of fve storeys that must have been fully 1 5 m. high. Beside
the tower i s a second entrance to the court-yard i n the form of a large vestibule.
On the remaining si de of the court is an irregular group of seven rooms of various
sizes opening into each other, and two arched rooms, of the ordinary type, not con
nected with the group of seven, but opening upon the court of number XVI I I . Of
the seven rooms, not one resembles the ordinary living room with its transverse arch ;
three of them are rather small and gi ve upon the court ; three others are large, but
long and narrow ; one of them i s spanned by an arch across its shorter axis and
another has a square pier i n the middle, whi ch supports a cross wall i n the foor above.
This room is also provided with a latrzna. This house, with its spacious arched apart
Inent and loggia on one side, its tower on the other, and its carefully planned group
of rooms i n two storeys, seems more like a public, or semi-publi c, building of some
ki nd, than li ke the residences wi th whi ch we are already familiar in the towns of the
Southern Ijauran.
House XVI I I i s more the typical large group of residences of the better class.
On the north si de of i ts court is a residence, or two residences, composed of two
arched rooms of medium ' size above a low groundfoor without arches. Its west side
is occupied by one side of the double-arched room of House XVI I , and by a fne large
arched room of one story, fanked by rooms of i rregular shape, the group constituting
probably two residences. The bui ldings of the south side of the court are shallow, only
3 . 80 m. wide i nside. In the mi ddle i s the vestibule, having a doorway outside and an
arch withi n, which was the chief entrance to the compound. I t is the residences of
the east side of the court that are of particular i nterest, and to which I shall devote
three ill ustrations. The plan is perfectly well seen in Ill. 1 7 8 where two suites are
shown, probably two large residences, each consisti ng of an arched living room with a
narrow room on either side of it, making six rooms in all on the ground foor. The
presen't state of the building . is shown in two photographs (Ills. 1 7 9 and 1 80), both of
whi ch give the northern half of the building, 1 79 being the front on the court and
1 80 the exteri or wall. It will be seen that the arches, lower and upper, have fallen ;
but the three- storey part at the north end, and the east wall with the large windows
in its upper storey, are standi ng. By means of these photographs, and with the aid
of the Secti ons given i n Ill . 1 8 I , a clear idea of the arrangement of the house may
easily be had. The drawi ng, also i n Ill. 1 8 I, showi ng the actual state of part of the
building i s the same view as Ill. 1 80, but drawn to scale ; the other drawing i s an
enlarged view of the , great window which gi ves thi s house i ts special distincti on : it i s
al so drawn to scal e. The two secti ons should be exami ned together. The two resi
dences are almost precisely simi lar although arranged reversely ; the four-storey parts
coming together, and the ' three-storey parts at the ends. I n plan and superstructure,
IlI. i g. Part of House o. XVI I I, Northeast Angle of Courtyard.
Ill. 1 80. Part of House No. XVIII, East Wall, Exteri or.
Umm i dj-Dj i mal ( Thantia ?)
.
I
SECTION C-D O L
, .. . . _ I
H OVSE N XI II
,
IAST\ALL' NORTH END
ACIVAL 5TATE
Ill. 1 81 .
WINDOVATX
SCAE : .1

.5eM-1_
20 1
202 Di vision I I SeCtion A Part 3
and in number and arrangement of rooms, these residences resemble House I I I ; but
they are built on a l arger scale. There can be no questi on as to the correctness of
the general li nes of the restoration ; for what one residence lacks another supplies.
Only the double arched wi ndow i n the residence on the south and the placing of one
or two small windows could be questioned. Section A-B shows just how much of the
front wall i s standing ; the full height of the walls is preserved in one place in this
wall , and i n three other places in other walls. The corbelled stairs are restored only
i n part, and show how most of the rooms were reached ; the others were reached by
short stairs inside the upper arched rooms. The whole structure is interesting i n point
of plan and construction ; but the most important details are to be found i n the pas
sage on the ground . foor at the north end, and i n the coupled arches of the upper
storey wi ndows. The passage is a l ong vestibule entered from without by a doorway
framed in draughted stone-work of excellent quality, and closed by stone doors in two
leaves, two metres high, beautifully fnished on both surfaces and swung upon ball-and
socket hinges ; one of these doors is still i n place. The corbel course which supports
the ceiling of the passage is a moulding of salient cyma-recta profle highly fnished,
and the walls are covered with plaster i n three coats, the uppermost being carried
over the surface of the moul
ding and over the ceiling, and painted in designs i n red
and yellow, the pattern of which is no longer to be distinguished. The arched window
i s an i nteresti ng feature as an architectural detail , and because it commands a superb
view, over the city wall, of the rolling country that lies to the east of U mm idj-[)ji mal .
The capital of the colonette which divides the two arched openings is important i n
that i t reproduces the form of Nabataean capitals i n the Dj ebel lauran. The second
window with coupled arches, which I have shown i n the other resi dence, i s not now i n
existence ; but i ts si l l and the lower courses of one of i ts piers were found i n place,
and l eave no doubt that i t once existed. The date of this house, or rather of these
two residences, like that of most of the domestic architecture of the city, i s very much
i n doubt. Two or three features, however, are suggestive, not so much by themselves
as when studied i n connecti on with one of the tombs on the east side of the city.
The tomb i n questi on is one in which the Nabataean i nscriptions are in situ (Ill . 1 8 5 ,
see p. 206) ; these i nscriptions quite defnitely fx the date of the tomb within the frst
century or the early part of the second. It i s only a matter of stone-cutting and sur
face fnish that seem to link the house described above with this tomb ; but thi s, taken
together with the capital of the colonnette which i s strikingly like Nabataean capitals
in the mountai n, one of which bears a N abataean i nscripti on, and with the draughted
masonry about the doorway, is sufcient basis for an hypothesis i n favor of an early
date for the house, a date not later than the middle of the second century after Christ.
HOUSES XII and XI I I . In quite another quarter of the city, almost in the middle
of the town, east of the " Cathedral" , and west of the Roman reservoir, is a block com
prising four courts, the southern half of w
h
ich, with two courtyards, was measured, and
is herewith presented as houses XII and XIII (Ill. 1 8 2) . There is nothing particularly
signifcant about the former ; its rather wide and shallow arched rooms are but one
storey high ; i ts smaller rooms have two foors. In i ts northwest angle are two l arge
apartments, separate residences in all probability, which open upon another court and
belong to another house. Al l the larger rooms here are provided with latri nae. But
house XIII i s peculiar ; it has i ts own l ittle court, though it has al so an entrance upon
Umm l dj -Dj | ml ( Thantia ?) 203
the court of No. XII. It has but one large room, and that without the usual trans
verse arch. Its roof, as is shown in Sections A-B and C-D in Ill . 1 8 2 , is carried on
three tiers of corbels which project 1 . 20 m. from the side walls. A photograph of these
corbel courses is given i n Ill. _I , I I , A. 2 ; the slabs of the roof were 2 . 60 m. in
length. At the west end of this room is a screen composed of thin slabs of basalt
pierced with holes, which divides it from the northern half of the long room adjoining
it. This screen is one of the most interesting details of domestic architecture to be
seen in Umm idj- Dji mal , but it is not easy to suggest a plausibl e use for it. Its position
`L'l Ll

.

J
' `'
.`

, .,, _
.'
'.
J
.
` =
.


c,:",
.

. `` `

.

. `
?: -
t

3i .
HOVSIS

|z1 CL` IJC! !
JI. I bz.
is indicated in the plan and sections by the letter R ; a drawing on large scale is also
given in I ll. 1 8 2 , and a photograph will be found under Il l . 1 8 3 . The drawing and
photograph render further description unnecessary ; the beautiful stone-work of the screen
is now i n strong contrast with the rough walls which were originally covered with
plaster . At frst a womans ' apartment suggested itself as the probable room on the
other side of the screen, and this n1ight have been a possible explanation of the screen
itself : but an examination of the room beyond the screen dismisses any such theory.
And, on the basis of much evidence gathered in all parts of Syria, I think it highly
improbabl e that the harem, in its modern sense, was an institution among the Christian
and pre- Christian inhabitants of the lauran, or that the women of the early centuries
of our era were screened of from masculine eyes. The room beyond the screen is
an unusually long one giving partly on the court, divided in halves by a narrow arch
204 Di vision II Section A Part 3
and provided with cupboards. The bottom of the screen rests directly upon four basins
cut from single blocks of stone and separated each from i ts neighbour by thi n walls
of highly fnished basalt as high as the screen. The basins bear a slight resemblance
to the mangers i n the better class of ancient stables in the lauran ; but I can think
of no good reason for screening the sight of cattle from the living room in this i nstance,
when it was common in all quarters i n the country for the heads of horses or cattle
to protrude i nto the lower living room. It is not i mpossible that the basi ns were
connected with a fullers shop or a dyeing establishment. It would be necessary to give
l ight and air to the operators of the one or the other ; and perhaps i mportant to protect
the bleaching or dyei ng stufs from those who commonly made use of the large room.
One of the latrinae, so common i n this town, is to be seen in the northeast corner of
the large room. It is set partly in the thickness of the two walls of the room, its own
side wall and the wall where its doorway is, being made of very thi n slabs of basalt,
highly fnished and carefully j oi ned. The edge of the doorway is carefully countersunk
to receive a wooden door, and the holes for the hinges are small and well cut. Just
wi thi n the door are a small basi n corbelled out of the wall and a shelf to hold a j ar
of water, all very much like the wash-hand-basins in the better class of city houses i n
Syria of the present day. The upper part of this l ittle structure is visible i n a photo
g raph (Ill. 5 I I I , A, 2) .
OTHER HOUSES. It is hardly worth while to multiply examples of domestic archi
tecture i n a case like this, i n which we have a whole city with hundreds of houses
built on the same plan, but, for the sake of record, I shall run rapidly over the houses
of which I have made measured plans, and which are shown only on small scale i n
Map NO. 2 . House I i s composed of two and three storey residences, rather poorly
constructed and having no arched rooms. No. I I i s shown, i n part, with the plan of
the Southwest Church (Ill . 1 6 I ) . I doubt if the house was a part of a group of reli
gious buildings of which the church was the nucleus, although it has a small rear
entrance from the courtyar
d
adj oi ning the church. It, of course, may have been in
habited by a priest or by some dignitary of the Church ; but i t i s an ordinary example
of a residence of the better class. It has one large arched room i n one storey, and
seven smaller rooms two storeys high, ffteen rooms i n all . From the standpoint of
constructi on and quality of fnish, it i s far better bui lt than the church, Interesting
features of detail are found i n the arched room whi ch, until recently, preserved i ts roof
i ntact ; the arch i s still i n pl ace. These features are two wreaths of bay ' leaves executed
in stucco, i n relief, upon the sofft of the arch at i ts springing, and a painted design of
conventi onal fowers, like tulips, which appears upon both faces of the arch. The colours
on the wreaths and in the fat design are much weathered, but still show green, red and blue.
HOUSE No. IV i s situated on the city wal l , protruding beyond i t. It is a crudely
built structure of three storeys, without arches, and its walls are full of more ancient
grave stelae brought i n from deserted cemeteries near the city. No. V i s a very wel l
built house, and very well preserved. Some holes for pipes set vertically in its walls
suggested the possibility of this having been a small bath. House No. VI i s later
than No. V, and is not so well built ; but it represents a good class of domestic archi
tecture. Its walls contain no less than fve N abataean i nscripti ons, and its court i s the
present site o( a N abataean altar. The ancient Nabataean temple was not far away.
One of the largest groups in the city is that which i ncludes VI I , VIn and IX ;
I ll . 1 83. Screen Wal l in House No. XII I .
111. 1 84. Southeast Angle of Courtyard No. VI I I .
Umm i dj-Dj i mal ( Thantia r)
205
these represent a good class of houses, in one and two storeys, built of smooth quad
rated masonry. The front of a house in the southeast angle of courtyard No. VII is
shown in a photograph (Ill . 1 84) . In these groups the arched rooms are one storey
high, the other rooms are in two storeys. I estimated that these houses accommodated
about twenty- fve famili es. Houses X and XI are of no particular i nterest, XI I and
XI I I have been described, XIV i s i rregularly planned and poorly built, XV is interesting
because it consists solely of a vestibule fanked by two small rooms, at one end of the
court, and a large room with two transverse arches at the other ; being more like a
small public buildi ng than a private residence, it should be compared with XVII described
above. No. XVI attracts attenti on only by its l ack of symmetry and the manner in
which its rooms are ftted i nto a very irregular plan, XVII, XVIII, and XI X have been
described in detail. XX is a well built house on the city wall, but has no striking features .
TOllbs .
It IS particularly fortunate that the funeral architecture of Umm idj -Dji maJ ,
though for the most part i n rui ns, preserves enough examples of tombs i n a suffci ently
well preserved state to enable us to reconstruct them with a considerable degree of
accuracy ; for few tombs of any architectural signifi cance have been di scovered i n the
ljauran, and l ittle is known about the funerary architecture of Southern Syri a, as com
pared with that of the North, and wi th that of Arabia as we know it in the tombs of
Petra and those of Medain $aleQ, recently published with ample illustrations by the
Rev. Fathers Jaussen and Savignac 1. We may beli eve that the speci mens of tombs
at U mm idj-Dj i mal are typical, and, by them, some light may be thrown upon a more
or less dark subj ect. Quite the reverse of conditions in Northern Syria and in Arabia,
none of the tombs at Umm idj- Dj imal is rock-hewn . All were wholly or i n part exca
vated i n the soil and were paved, walled, and roofed with stone. There are three kinds
illustrated i n the tombs which were found i n a conditi on to be measured and restored.
One wholly excavated in a fat surface, one excavated in a hill-side and showi ng a
front wall, and a third only partly excavated and having a building of some height
constructed in plain view above the ground level . The earli est of the tombs . has
N abataean i nscripti ons in si tu, and therefore can not well be later than the mi ddle of
the second century after Christ. In all the others the inscripti ons are in Greek, the
names are often Greek or Roman, and there is no direct evidence that they are Chris
tian ; we may therefore assume that they are earl ier than the fourth century. How
the Christians buried their dead, if di ferently from their predecessors, we do not know.
The tombs of U mm idj -Djimal were partly destroyed at an early date for building
purposes, and partly i n comparatively ancient times by marauders who rifed them ruth
l essly, and not in search of antiquities ; for the tombs are strewn with broken glass of
the highest irridescence, fragments of which show that the vases and bottles were of
beautiful and intricate forms, often highly ornamented with threads of glass. In one
tomb we found a few simple bronze obj ects, such as buttons, buckles and little bells
which had escaped the eye of the plunderers ; some of the buckles were inlaid with
glass or enamel ; the bells are of two forms, one the ordinary campaniform, the other
like two cockle shel ls holding a ti ny ball . In another tomb two small bottles, highly
irridescent, had been miraculously spared.
1 Mission A,-chiologique en Arabie, Paris, 1 909
,
pp. 307-404.
Publications of the Princeton University Archaeological Expeditions to Syria, Div. II, Sec. A, Pt. 3.
.
206 Division II Section A Part 3
NABATAEAN TOMB. This is the earliest and the most important of the tombs dis
covered by us at Umm i dj-Dji mal . It lies far to the southeast of the ruined city, where
the ground is fat and level, and is al most entirely sunk below the surface. It i s given
the number 10 in Map NO. 1 . A fight of steps on the east side descends to a door
way with moulded j ambs and l i ntel . Withi n the doorway are two more steps, of loose
stones, by which one descends to the present level of the foor of sand which i s probably
not the origi nal foor level . The plan of the i nterior (Il l . I SS ) consists of a chamber,
5CALE: .50CM l M

NABATAEAN
TOMB
Il l. 1 85 .
4. 7 7 m. wide and S . 30 m. l ong, divided by three transverse arches i nto four unequal
bays. The frst two bays withi n the entrance are equal, being 1 . 30 m. deep, and their
sides are closed with highly fnished walls i they constitute a sort of chapel in front of
the tomb proper. The other bays measure I . So m. and 2 . 50 m. deep, respectively i
their side walls are composed of the ends of loculi (Ill . I S S , Sect. A-B) arranged i n
three tiers, two on each tier i n the frst bay, and three i n each tier i n the second.
A fourth tier may be buried i n the sand below the present foor level . The loculi are
separated laterally by thin walls of basalt, and the ti ers are divided by thin slabs like
the di \' i i cn walls. The bodies of the dead were sli d into the locul i , probably upon a
board, the funeral oferings, objects of ornament, vases of pottery and glass were placed
around the body, and then the opening was closed by a single thin slab which ftted
it nicely. When the loculi were all closed, the i nteri or of the tomb presented much
the efect of an oblong church spanned by transverse arches. The stonework of the
i nteri or i s most highly fnished. It is not i mpossible that a building of some ki nd was
constructed over the whole, or a part, of this underground tomb. Two long steps are
still to be seen extending on either side of the entrance. The chief i nterest of this
Il l . 186. Nabataean Tomb wi t h Stelae. Vi ew from the Northeas t.
Il . 1 87. Tomb wi t h Many Stelae. Vi ew from the East.
Umm i dj - Dj i mal ( Thantia ?) 207
tomb lay i n the fnding of a row of eight stelae, with Nabataean i nscriptions, I fanking
the south side of the stai rcase at the entrance. There was undoubtedly a second row
on the opposite side, making a lane of stelae before the door of the tomb. Hundreds
of such stelae, with Nabataean and Greek i nscripti ons, have been found by travellers
in the Southern Iaurn. It has heretofore been supposed that these were stones
placed at the heads of simple graves like modern tombstones ; but the discovery of
these stelae in their original positi on (Ill. 1 86) and the fnding of a large collection of
Greek stelae in place in front of another tomb here at Umm i dj-Dji mal (Ill. 1 8 7) , presents
the stelae i n a new light. 'They are not to be thought of as grave-stones in the
ordinary sense, but as tablets set up outside a built tomb, giving the names of those
interred withi n. Dr. Littmann fnds that, in the maj ority of cases, the stelae with
square tops were those for the names of men, while the rounded ones were for women.
TOMB OF SAREIDOS. I n the opposite directi on from the city, i. e. to the northwest,
across the wadi , i s a good example of the second type of tomb, i. e. that which is
excavated in a hill-si de. This tomb faces the east, and is marked 1 9 on Map No. I .
An i nscription gives us the name of Sareidos. I n prepari ng the ground for this tomb
no tunnel was made ; the hill i s low, and the tomb was placed near the top. A cutti ng
was made i n the hill-side wide enough for the tomb and carried back to the requi red
depth ; then a heavy front wall of highly fnished masonry was made, the interior was
lined with l oculi of stone, and a roof
of stone sl abs was laid over the whole
excavati on. The plan, (Il l . 1 8 8) is
a sl ight variation of that descri bed
above. The tomb i s altogether on
a smaller scale and space was eco
nomized. vVithin the doorway is a
chamber, 2 . 7 5 by 3 m. square,
spanned by a single transverse arch.
S ECTI ON AB SCALE l<M, = l M'
Three walls of the chanl bel are made
up of the ends of loculi i n four
tiers arranged somewhat diferently
from those i n the Nabataean tomb.
I n Il l . 1 85 I have shown all the loculi
open, i n Ill . 1 88 I have shown all
but three of the loculi closed. In
the Nabataean tomb all the l oculi
could be opened, each havi ng its
own door-slab moveable ; but i n the
tom b of Sareidos only the loculi of
Ill. 1 88.
the second and third tiers from the bottom coul d be opened at the end and, originally,
only one i ntermediate foor was built, that i n the middle. When a burial took place, at
frst, a slab i n the second or third tier from the bottom was removed, and the body was
deposited in the tier below, as if in a sarcophagus, this receptacle was then covered
with slabs j ust below the level of the opening and an upper loculus was thus created
I IV. A. Nab. inscrs. 60-67.
208 Di vision II Section A Part 3
to receive another body which would be slid in as in the Nabataean tomb, and the
slab would be put i n place. Thus six openings in the end wall served for twelve
loculi, and eight openings in the two side walls accommodated sixteen locul i .
No stelae were found i n or near the tomb of Sareidos. There were perhaps never
any stelae here, for the names of the dead are found in i nscriptions 1 that were painted
in red upon the slabs which closed the loculi . Some of these inscriptions are still
legible. It may be that these painted i nscriptions were placed here for the convenience
of the family and of those who were admitted to the tomb chamber, and that a
conventional row of stelae outside the tomb set forth the names of the dead within
to the casual passer-by. Stelae were so well adapted to the making of corbelled stai rs,
that the Christian builders of Umm i dj-Dji mal allowed very few stelae of Pagan tombs
to remain in their places, and the mere absence of stelae fron1 the front of a tomb
does not necessarily mean that there were never any stelae there. But if the use of
these name tablets was an Arabic, that is to say a N abataean, custom, it would be
i nteresting to know for how long, and to what extent, the custom was proj ected i nto
Roman and Christian times in Southern Syri a. The square compartments i n the angles
between the loculi opening upon the sides and end of the chamber were also used for
i nterment. They were divided i nto two ti ers, instead of four, and the i ntermediate
foor and the roof were supported by a pier in the middle as shown in the plan.
These receptacles for the dead could be reached only from above, by making an
opening in the stone roof. In all probability the i ntermedi ate foor was not put i n
r- -, ! ' : I r
1; : : 1 :
r-- - - . . 76'c - - - - - 1
'___ _ _ ___ __ J
[

_
-
_
-
_
1. 0 - _J
A
PLAN 3ECTI ON A RESTORED -
TOM oF MAECHOS
until the bottom fl oor had
received its complement of
bodies.
TOMB OF MASECHOS.
A large number of the ruin
ed tombs on the west side
of the city, particularly those
near the high-road, were of
the type illustrated in the
tomb of Masechos (Il l . 1 89) .
This is the third form men
tioned i n the introductory
words on the tombs. Here
we have a little building, a
diminutive square chapel ,
with its doorway on the
south or east. Withi n the
doorway, which is on the
ground level , the foor is found to be sunk I. 1 0 m. , and the walls of this sunken
part, on three sides, are made up of the ends of loculi, as in the l arger tombs, but
in only one tier. The outer ends of these loculi proj ect far out beyond the walls of
the little tomb building. The loculi were closed in with thin slabs ; but the tombs
were further protected by a foor of stone slabs inserted at the l evel of the door-sill .
I III, inscs. 279, 280.
Umm i dj - Dj i mal ( Tlantia?) 209
Alternate posts and brackets for the support of these slabs are found i n each wall .
Wi thi n the chapel , upon this fl oor, one or two sarcophagi may have been placed ; for
I found several sarcophagi near the tombs of this type near the West Church. The
superstructure of the tomb of Masechos has been destroyed ; but pi eces of its cornice
and the li ntel of its doorway were found in the rui ns. A Greek i nscription 1 on the
lintel gives the name of Masechos, who I believe was an earlier resident of the city
than that Masechos whose name occurs in an i nscription upon the church near the
East Gate. If he was the same man, he devoted greater pai ns and skill to the build
i ng of his tomb than to the construction of the church.
STELAE TOMB. A combi nati on, one might say, of the types represented on the
Nabatean tomb and in the tomb of Masechos, is found i n this tomb of many stelae ;
SECTl ON A-B' SCALE: I C- 1 M"
D ESTOED
A- B
=== =c
STtLAE
TOMB
Ill. 1 90.
for it has the burial system of the former of these, but is built, for the most part,
above ground like the latter. Its number in Map No. I is 1 6, as I remember ; i t
stands at a good distance to the north of the city. Its plan (Ill . 1 90) i s only a slight
modi fcation of the tomb of Sarei dos. Only the two lower tiers of loculi are in existence ;
but the springers of the transverse arch of the tomb chamber are in situ, and these,
1 III, insc. 276.
2 1 0 Division II Section A Part 3
gIvmg the height of the arch, suggest the full height of the tomb, which would accommo
date two more tiers of loculi. In the Secti ons A-B and C-D, I have shown some of
the loculi open and others closed with slabs. Some of the slabs which used to close the
ends of the loculi were actually in place, the rifing of a closed loculus having been
accomplished by breaki ng through the bottom of the loculus above. In this tomb where
so many stelae were found in situ that I called it the " Tomb of Many Stel ae" , the
name tablets were not arranged to fank the approach to the entrance to the tomb
chamber, but were set up i n a l i ne (Ill . 1 8 7) agai nst the front (east) wall of the tomb,
with their l ower third well sunk i n the ground. As the number of interments i ncreased
a second li ne of stelae was set in front of the frst, and in two places a third row
was begun. The outer rows, of course,
made it i mpossible to read the i nscripti ons
on the i nside rows of stelae. There are
nearly twenty stelae visible in the photo
graph, but these do not represent the full
number of burial s, if all the loculi were
full ; for there were places for ffty-two
seperate i nterments. The square spaces i n
the angles between the loculi which open
upon the sides, and those which open upon
the end of the tomb chamber, were treated
i n a diferent manner from those in the
tomb of Sareidos, where they were reached
from above. Here, these spaces, divided
i nto four tiers, have openings 74 cm. wide
upon the tomb chamber. Each tier i n these
spaces had room for three bodies ; the frst
was sl id through the narrow opening, and
then moved sideways over to the west wall , .
and a partiti on was built beside the body ;
the second was deposited in a similar fash
ion and a second partiti on was built, leaving
a loculus of the ordinary type j ust i nside the
Ill. 1 9I . Cast of Doorwa
y
and Door of a Tomb, from the
opening. The third was slid in the ordinary
Princeton Collection.
manner, and the loculus was closed.
DOORWAY
'
AND DOOR OF A TOMB. Among the debris of a ruined tomb near the
West Church I found the li ntel and jambs and the door of a tomb of the same type
as the tomb of Masechos (Ill . 1 89) . Lintel , jambs and door were ornamented with
carving i n fat relief, quite unlike anythi ng to be seen elsewhere in , Umm idj-Dji maJ .
I had squeezes made of all theSE details, had them cast i n plaster, and set up i n
Princeton as i n thei r original positi on, wi th the door swi nging on its ball-and-socket
hi nges. A photograph of the cast (Ill. 1 9 I ) is given herewith. The lintel measures
1 . 5 0 m. long, and 45 m. high. The door of basalt was chosen by Schumacher for one
of the illustrations i n hi s article 1 on Umm idj-DjimaJ . It was 92 cm. high, and 69 cm.
1 Z.D.P.-V. xx, ' 97. p. 1 59.
U mm i dj -Dj i mal ( Thantia ?) 2 I I
wide. It is now broken in two pieces. The ornament is executed i n exceedingly fat relief.
The design of the l intel consists of a disc of i nterlacing ropes in the middle, fanked
by two pairs of columns with twisted futings which appear to support an entablature
divided into discs and squares like a l ate Doric frieze. Below the disc and the columns
is a border of square nail-head pattern above a rinceau of grape-vi ne. The ends of the
lintel are decorated with pilaster-like designs bordered with reeds and rows of square
nail-heads, which are carried down the jambs. The faces of the panels are adorned
with triangles and rings. The door has a counter-sunk panel bordered on three sides
with a runni ng grape-vine, withi n which, over the panel, i s the small fgure of an eagle
with raised wings. The panel itself, surrounded with a fat reed, contains an arch
supported on two twisted columns, and having rosettes in its spandrils. Beneath the arch
is a palm tree which flls the space. There is a small wafer on either side of the
trunk of the pal n1 tree. The grape-vi ne i s a common motive in the Pagan and
Christian ornament of Syria. The eagle was a favorite subj ect with the N abataeans and
the later i nhabitans of Southern Syria and Arabi a, and occurs frequently in the tomb
sculptures of Petra and Medin $leb, the palm tree, often associated with emblems of
Christianity, was certai nly used by the pre-Christian peoples of Syria. Neither the cross,
nor other purely Christi an symbols, occurs among these designs.
FRAGMENTS. Carved architectural fragments are not common among the ruins of
U mm idj- Dji mal ; yet a number of small details were found which are
worth noting, and a few bits of sculpture were discovered which are
enough to show that, in the Nabataean and Roman peri ods, the city
was not barren of statues. A capital (Ill . 1 9 2) was found lying i n the
courtyard of House No. V, which has no counterpart, so far as my
SCAE: 1 -10 '
Ill. 1 93.
knowledge goes, i n all Syri a. It i s the
cap of a square pier ; i ts lines follow, .CAE: l=1 0
i n general, those of the Nabataean tombs Ill. 1 92.
of Medain $leb.
The dimi nutive altars which protrude from the
j ambs of several doorways i n the city are i nteresting
bits of detail . One of them is shown here in a dra w
i ng (Ill . 1 9 3) . They are placed at mi d-height of the
doorways, and on the right-hand as one enters. They
are cut
.
with one of the j amb-stones, and stand out
like brackets. The top has a slight de
pression, and this suggests that t hey may
ha ve been actual altars of libati on to
protect the entrances from evi l i nfuences.
Another detail presented i n a drawing 1 by Schumacher i s the upper
part of a small square shaft. This also has the form of the upper
part of an altar (Ill. 1 94), but the cross represented i n relief on one
face marks it as Christian. It i s probably the upper part of the post
Ill. 1 94
of a chancel rail from one of the churches. It was found between the Double Church
and the Roman reservoir. Two real altars, both already menti oned in these pages,
1 Z.D.P.- V. xx, ' 97. p. 1 6 1 .
2 1 2 Division I I Section A Part 3
are presented in drawings in Div. I II, under i nscripti ons 2 3 8 and 2 39. The frst of
these is an exceedi ngly plain pedestal with a tall and slender die having a Nabataean
scription 1 on one side and a
Greek i nscription on another,
and with bevelled faces i n base
and cap. The second also has
the pedestal form, but is more
of the Classical type. Its base
moulding, an i nverted cyma
recta, is elevated on a high
plinth ; one face of its die i s
ornamented wi th a wreath of
l eaves i n relief, and i ts cap
is a delicate cyma recta above
a reed. The over-cap, or attic,
i s a plain block, one face of
which bears a dove-tailed plate
in relief contai ni ng a Greek
i nscripti on, dedi cati ng the altar
to Salmos, a N abataean deity.
Ill . 1 95. Ionic Angle-Pi laster Cap.
Another fragment worthy
of menti on is an Ionic pi laster
cap (Il l . 1 95 ) , which was found
near the east wall of the barracks chapel, but which must have come from the south
west angle of the N umerianos Church, since a similar cap is still in place at the
northwest angle. It is probable
that both origi nally bel onged
to a building of the late Roman
period ; for they bear no re
sem blance to the Christian
rendering of the Ionic order.
A corner acroteriun1 (Ill .
1 96) was found lying i n the
courtyard of House XV. It
is well executed in the hard
black basalt, and i s al most per
fectly preserved. The acro
terium i s 3 2 cm. high, exclusive
of its base, which is 1 9 cm.
high, and measures 37 cm. along
its base. This may have be
l onged to the roof of a building
Il l . 1 96. Acroterium found in House No. XV.
of the N abataean period ; for acroteria almost exactly si milar were empl oyed on the
Nabataean gate, and on the Temple of Dushara at s1', or it may be Roman, and
n1ight have belonged to the cover of a large monumental sarcophagus.
I IV. Nab insc. 38.
Umm i dj-Dj i mal ( Thantia ?) 2 1 3
The fragments of sculpture consist, almost entirely, of pi eces of drapery, some of
them well executed, others stif and crudely wrought. One fragment, found outside
the walls, near the chapel outside the East Gate, is the lower part of a torso (Ill . 1 97) ,
Ill. 1 97. Torso i n Basalt.
from knees to waist, 40 cm. high. It belonged to a statue a little less than l ife size,
and compares favorably with the remai ns of sculpture from about the ruined temples
of the Djebel laurfn.