https://www.questiaschool.com/read/1G1-20202121/the-site-of-saar-d...
The Site ofSaar:D ilm un Reconsidered
By Craw ford,H arriet Antiquity,Septem ber 1997
Crawford, Harriet, Antiquity
Geoffrey Bibby went to the Arabian Gulf 'looking for Dilmun' in the 1950s and '60s as part of a Danish team led initially
by Professor Glob. For 20 years after this little archaeological research took place. In the 1980s the picture changed
again; much of the work undertaken was driven either by the need for rescue work, or by the accident of discovery. A
major conference in 1983, followed in 1986 by the publication of an important book Bahrain through the ages (al Khalifa
& Rice 1986), gave fresh impetus to work in the area and, for almost the first time, some problem-oriented research
was undertaken to answer specific questions.
In summary, the state of knowledge prior to 1986 was as follows: thanks to the cuneiform texts, the earliest examples
of which date back to the Uruk period of south Mesopotamia in the late 4th millennium, Dilmun's identification with
Bahrain and its position as a vital entrepot in the trade route linking south Mesopotamia with the copper mines of Oman
was already well known (Potts 1990: 85ff). The texts and the archaeology together showed that by the early 2nd
millennium Dilmun was centred on the island of Bahrain, but the archaeological evidence for this period was still limited.
The first extensive modern archaeological investigations of Bahrain were undertaken by Danish teams under Professor
Glob, Peder Mortensen and Geoffrey Bibby at the Qala'at al Bahrain, a large stratified tell site on the north coast of the
main Bahrain island. The site apparently covered almost 50 ha in the so-called Early Dilmun period, usually dated to the
first quarter of the 2nd millennium. Much of the early material at this site is deeply stratified under later material; the
Danish expedition was only able to establish a stratified sequence of pottery from the pre Early Dilmun, conventionally
dated to about 2400 BC, to the Hellenistic periods, and to identify, among other buildings, a short length of the early
2nd-millennium city wall and the fragmentary plans of half a dozen houses (Hojlund 1994).
Mortensen also uncovered a temple at the village of Barbar 3 km to the southwest, which dated to the Early Dilmun
period and gave further indications of its prosperity and the technical skills of its builders (Mortensen 1986). Finds from
the temple reinforced the image of Dilmun as an important trading entrepot. There was lapis from Afghanistan, copper
from Oman and artefacts such as the cylindrical vessels, mirror handle and copper bull's head which pointed to
contacts with eastern Iran or even with Central Asia (Lombard & Kervran 1989: 30). The wide-ranging contacts of the
people of Dilmun were already well known because of the finds of their seals in a distinctive local style in excavations at
Ur in south Mesopotamia (Gadd 1932), at Susa (Harper 1992) and, more recently, from the island of Failaka (Kjaerum
1983). Rare examples are even known from the Indus valley (Rao 1986).
Finds from the burial mounds which were for centuries the most notable archaeological feature of Bahrain, and which as
a result have been extensively robbed, painted the same picture. The so-called Royal Graves at Aali (Reade & Burleigh
1978) in particular demonstrated engineering and building skills of a high order as well as yielding the remains of a
1 of 7
28/01/2015 03:43
https://www.questiaschool.com/read/1G1-20202121/the-site-of-saar-d...
sophisticated material culture. These included fragments of ivory figurines and scraps from furniture as well as imported
Iranian pottery and a quadruple - spiral gold bead which can be parallelled as far away as Troy (Lombard & Kervran
1989: 32). Rescue work on many other less imposing and less well-furnished types of graves provided additional useful
information on the artefacts of the period and underlined the special nature of the Aali burials. Unfortunately, the skeletal
remains were frequently too badly preserved to allow demographic studies to take place.
What was missing from the picture of Dilmun was detailed information on the domestic economy and social structure of
Bahrain in the Early Dilmun period, information on how local people actually lived and on how their society was
organized. The London-Bahrain Archaeological Expedition (LBAE) was set up in 1989 to identify and explore a
settlement site of the period using modern methods of data collection and analysis in order to fill this gap.(1)
Excavations at Saar: the setting
Some (unpublished) work had been undertaken by a joint Bahraini-Jordanian expedition at the settlement site of Saar in
the northwest of the island in the 1980s [ILLUSTRATION FOR FIGURE 1 OMITTED], but had been suspended.
Permission was granted by the Minister of Information(2) for the LBAE to re-start work there in the spring of 1990. The
remains of the essentially single-period Early Dilmun site, covering about 2-5 ha, lay immediately under the surface with
no overburden and appeared relatively undisturbed. The site lies in the lee of a north/south limestone ridge on which
stood an extensive mound field composed of thousands of burials, now largely destroyed. To the east is an area rich in
springs, today used for date gardens. The sea now lies about 3 km to the west and about 7 km to the east; geological
maps of the area indicate that the sabkha, or compacted sand, on the east comes within 3 km of the site and a tongue
of sea reached much closer to the ridge within living memory. It seems likely that this was the case in the early 2nd
millennium too, and there may well have have been a sheltered anchorage below the site on this side.
Subsistence
The importance of the sea to the domestic economy of Saar was immediately obvious. Fish seem to have been the
major source of protein and the remains include many of both the inland and deep-water species found in the waters
around Bahrain today. Copper fish-hooks of various sizes were present, and analogy with modern practice suggests
that fixed nets may also have been used. Large numbers of shells are also found, and include the pearl oyster. Contrary
to modern perceptions which see the pearl oyster as inedible, the quantities present suggest it was probably eaten, or
used for bait, rather than sought for the pearls. A number of tiny seed pearls found in the course of the excavations are
unpierced and were probably too small to have been of any value as ornaments. The absence of Terebralia from the
shell collection is interesting, as this shell is found in large numbers further south on the coast of the UAE. It lives in the
roots of mangroves and its absence suggests that the mangrove swamps did not occur around this coast in the early
2nd millennium (Glover 1995). Some fragments of mangrove charcoal have, however, been identified from Saar, but the
author is inclined to suggest that the wood may have been imported (Gale 1994: 223).
Evidence for agriculture was also recovered. Scholars have in the past assumed that most of the grain eaten in Dilmun
was imported from Mesopotamia (e.g. Edens 1992), but Potts (1993) has challenged this view. In addition, recent
history demonstrates that, with irrigation, cereals can be successfully grown on the island.
Preservation of carbonized plant remains at Saar is poor; only a few grains of cereals were retrieved together with an
even smaller amount of chaff so the question remains unresolved. However, there are large numbers of quernstones
and rubbers used in some form of (food?) processing, though flint sickle blades are very rare. Date palms provided the
most common plant remains and dates were eaten in quantity, the leaves, fibres and wood also providing important raw
materials (Nesbitt 1993).
2 of 7
28/01/2015 03:43
https://www.questiaschool.com/read/1G1-20202121/the-site-of-saar-d...
In summary, the environmental evidence points to a largely self-sufficient community, dependent on the sea and
date-gardens for its subsistence. There is evidence for the presence of a small amount of cereals, but it is not at the
moment possible to tell if these are indigenous or imported.
External contacts
The site was not entirely self-sufficient as the artefacts recovered show: many of the larger stone tools such as mortars
and door sockets are made from locally available limestone, but finer-grained hard and softstones were imported,
apparently from the Arabian mainland. Copper, present in small quantities, has yet to be analysed; the presence of a
limited amount of 2nd-millennium Wadi Suq pottery from Oman suggests that the copper also came from here. The
bitumen from the site is of Iranian origin (Conan pers. comm.), by contrast with that from the Qala'at which comes from
Mesopotamia (Hojlund 1994). External contacts seem to have varied between large centres such as the Qala'at and the
small local communities, perhaps reflecting individual enterprise by local fishermen in the latter as opposed to state-run
trade in a major centre.
There is also evidence for a flourishing trade within Dilmun itself. Hundreds of seal impressions have now been retrieved
from the houses at Saar originating from bales, packages and jars. This adds another dimension to the study of the
local economy. The fact that all the sealings are in the native Early Dilmun style shows plainly that, in spite of Dilmun's
status as an international trading entrepot, most of the commerce at this level was within the Dilmun polity itself, or even
within the Saar township. Their wide distribution also indicates that the majority of the inhabitants of Saar were
economically active.
The architecture
The settlement, of which almost 70% has now been excavated, offers further evidence on the social organization of the
community. It was dominated by the temple which stood on the highest point at an elevation of around 14 m at the
junction of two major roads, now called Main Street and Temple Road ([ILLUSTRATION FOR FIGURE 2 OMITTED].
The temple is isolated from the other buildings by two smaller alleys which run north and south of it, while a third runs
further to the south and parallel to the other two. east of Main Street two more small roads have been identified leading
eastwards, parallel to Temple Road, towards a well. The smaller roads divide up the houses in the centre of the
settlement into blocks of not more than four or five. (The situation is less regular on the outskirts of the settlement.)
Main Road itself runs south from the temple for about 200 m before running into a garden area where it becomes
impossible to trace. The impression is of a surprisingly regular layout which appears to result from some sort of overall
plan.
This impression of planning is reinforced by a study of the houses [ILLUSTRATION FOR FIGURE 3 OMITTED]. About
60 units have now been identified, though many have not been fully investigated. All are built of roughly finished local
stone; the majority conform to a single plan with minor variations. In its simplest form this plan consists of two rooms,
an outer L-shaped area and a smaller inner room, giving a total area of about 80 sq. m; some are smaller. The inner
room was normally roofed, but it is not entirely clear if the L-shaped one was. In House 207 it appears to have been, as
ceiling plaster was recovered from both areas. In others, we may suggest, a light palm-leaf roof was supported on the
shallow buttresses which are still a feature of traditional houses on the island today. Some of the inner rooms are
subdivided, as in House 50 for example, forming such small areas, sometimes cluttered with tools and other artefacts,
that we have to ask if they were storage rather than living areas.
It is not merely the plans of these houses which are standardized; a range of fixtures and fittings are common to many
of them. A double basin [ILLUSTRATION FOR FIGURE 4 OMITTED] is often found in the entrance passage with two
bowls, a higher and a lower one, usually well-plastered in grey plaster. A well-furnished house had as many as three
3 of 7
28/01/2015 03:43
https://www.questiaschool.com/read/1G1-20202121/the-site-of-saar-d...
cooking installations, each apparently for a different purpose. There are small open hearths with a clay rim, more
elaborate double ones, sometimes with a clay tripod on one side on which to balance a round-bottomed pot, and
well-built tannurs for the making of fiat bread. In addition plastered pits of uncertain function occur, and storage jars
have also been found sunk into the floors of the outer L-shaped areas.
The remarkable uniformity of the layout of the settlement and of the domestic architecture at Saar gives rise to
important questions. Is this a normal small town, or is it a special-purpose site, such as the Workmen's village at
Amarna, or a military barracks? Who was overseeing the planning of the settlement? We cannot answer these
questions. We can only say that there is no evidence from the small finds for anything other than normal domestic
activities taking place. There are no recognizable tool-kits occurring on a regular basis; there is no evidence for
large-scale manufacturing; nor is there any indication that it was a military barracks. Few weapons have been found,
and there is neither observable hierarchy in the domestic architecture, nor any evidence for fortification. In sum, there is
no evidence for Saar being a special-purpose site. If we turn to the second question, the temple is the only
non-domestic public building uncovered on the site. We have to consider the possibility that the temple was the
organizing force behind the planning.
The temple is built of the same local stone as the houses, held together with gypsum mortar; the walls were originally
heavily plastered both inside and out. It is trapezoidal in shape, 17-5 m long, with a curious bulge in the exterior wall in
the northern corner. In its second and last phase [ILLUSTRATION FOR FIGURE 5 OMITTED] the rear of the building is
subdivided into three separate areas and the roof supported on three centrally placed pillars: one square one at the
western end; one originally square one in the centre, which had an additional skirt added to it to make it look round at a
later date; and a third round one at the eastern end, where a single rather narrow door gives access to the building.
There are two altars, one on the south wall and one on the north side of the central pillar. Each is decorated with a
semi-circular plastered feature at the back whose interpretation has aroused much interest: it may be a schematic
rendering of bulls' horns, or it may represent the crescent moon. There is evidence for burnt offerings of fish and
vegetable matter having been made on each altar.
In addition to the altars there are three platforms, two against the east wall and one along the north, at right angles to
the other two. All these features are very finely plastered and on the top of the largest, which survived intact from the
earlier building, the imprint of a rectangular base was still preserved. There were also traces of purple paint on
fragments of plaster found in this area. In the open area outside the front of the temple are five circular bases, possibly
the remains of further offering tables or altars; as it is impossible to reconstruct their original height we cannot entirely
exclude the possibility that they were pillars. Similarly constructed bases can also be seen at the site of Diraz, 4 km to
the north, where they seem to have been pillars supporting a portico or loggia.(3)
There is relatively little space within the building and, with the exception of the storage area at the west end, all of it
seems to have used for the same, apparently ritual purpose. The small finds provided no evidence for any
administrative activity either. Only one seal was found, plus 77 fragments of sealing impressions. There are no
impressions which match the single seal and very few duplicate impressions. This suggests that the sealings were
coming into the temple from a wide range of different sources. The goods from which the sealings came included jars
and bundles; no definite door sealings were identified. As there is no evidence that the temple had a redistributive
function, the incoming goods were probably used for the support of the temple itself. In short, as there is no
indisputable evidence for administrative tasks being carried out in the temple, it is difficult to see it as the organizing
force within the community with control of planning.
Nor, as we have seen, is there any marked hierarchy within the houses, so there is no support for the presence of a
'secular' administration either. This impression of egalitarianism is reinforced by the distribution of the small finds. There
is some concentration of imported goods in the houses down Main Street and at the top of Temple Road, but a
4 of 7
28/01/2015 03:43
https://www.questiaschool.com/read/1G1-20202121/the-site-of-saar-d...
remarkable feature is the wide distribution of seals and sealings throughout the settlement. Many houses have not yet
been fully excavated; perhaps half the total have already yielded seals or sealings.
Seals and sealings
More than 80 seals have now been found. This unexpected quantity of glyptic material from a small town or large village
raises again the question of function. Their wide distribution throughout the settlement suggests that many, perhaps
most, members of the community owned a seal. Should we perhaps be looking again at the seal as an amulet, a
protective device, as well as an administrative tool? Four or five seals are frequently found in a single house; as each
house is apparently only large enough to have housed a nuclear family, this suggests that each member of the family
may have owned one. It is hard to believe that each was also involved in administrative or economic activity. (Or did
one or two individuals in each domestic unit have several seals each?) No common devices or motifs link the seals
found in a single house so designs do not seem to have been specific to families. Nor do duplicate designs occur,
suggesting that the design of seals may also have been controlled in some way. There is no correlation between the
designs on seals and sealings from a single location; this is to be expected as sealings were certainly attached to
outgoing goods, not to ones remaining within the house.
Conclusions
Seven seasons of excavations at Saar and recent work at the Qala'at al Bahrain by a French team, so far unpublished,
are beginning to clarify the internal organization of Dilmun society. The economic structure is certainly becoming clearer.
Dilmun's internal economy was based on the sea and on date-garden agriculture, with the possibility that some cereals
were being grown as well. There was a thriving internal market in which a broad sector of the population participated,
with some limited evidence for local participation in external trade too. It is not yet clear how essential supplies of
softstone and copper reached Saar; it is possible that they were redistributed from a larger centre like that at the
Qala'at whose prosperity certainly rested on the well-known international trade.
If Saar is taken as representative of Dilmun settlement - and there are no grounds for seeing it as a special-purpose
site - then the floor areas of the houses suggest the nuclear family was the basic unit within society; the regularity of
layout demonstrated at Saar suggests the presence of an overall plan rather than organic growth. As Saar itself had
little in the way of administrative structures, its regular layout seems the product of administration from another centre.
The nature of this administration, and its location, is unclear; references in the Mari letters to a king of Dilmun
(Groneberg 1992) and the unusually large and well-supplied graves at Aali suggest a possible answer. The
administrative centre might have been at the Qala'at al Bahrain, the only fortified site of this date yet identified, and the
largest.
There has been a tendency among scholars to see countries geographically peripheral to the great ancient civilizations
of Mesopotamia and Egypt largely in terms of their relationship to the centre: their indigenous culture has been
presumed derivative, inferior, of little interest. Today it can be shown that the countries of the Gulf developed in
distinctive ways and evolved in very different directions. Their relationship with Mesopotamia was undoubtedly important
in economic terms; in the case of Dilmun at least, the indigenous civilization adapted foreign influences, of which
Mesopotamia was only one, to form a distinctive whole. No mere shadow of its larger neighbour, it deserves to be
studied on its own terms.
1 The directors of the expedition were Robert Killick, Jane Moon and the author who retired as a director after five
years.
2 Without the support and encouragement of the Ministry of Information the work of the Expedition would not have been
5 of 7
28/01/2015 03:43
https://www.questiaschool.com/read/1G1-20202121/the-site-of-saar-d...
possible. Preliminary reports on the work have appeared as follows: Killick et al. 1991; Crawford 1993; Woodburn &
Crawford 1994; Moon et al. 1995.
3 A full report on the temple can be found in Crawford et al. 1997.
References
AL KHALIFA, SHAIKHA HAYA ALI & M. RICE (ed.). 1986. Bahrain through the ages: the archaeology. London: Kegan
Paul.
BIBBY, G.1972. Looking for Dilmun. Harmondsworth: Penguin.
CRAWFORD, H. 1993. London-Bahrain Archaeological Expedition: excavations at Saar 1991, Arabian Archaeology &
Epigraphy 4: 1-19.
CRAWFORD, H., R. KILLICK & J. MOON (ed.). 1997. The Dilmun temple at Saar. London: Kegan Paul.
EDENS, C. 1992. Dynamics of trade in the ancient Mesopotamian 'world system', American Anthropology 94: 118-39.
GADD, C.J. 1932. Seals of ancient Indian style found at Ur, Proceedings of the British Academy 18: 1-22.
GALE, R. 1994. Charcoal from an early Dilmun settlement at Saar, Bahrain, Arabian Archaeology & Epigraphy 5:
229-35.
Glover, E. 1995. Molluscan evidence for diet and environment at Saar in the early second millennium BC, Arabian
Archaeology & Epigraphy 6: 157-79.
GRONEBERG, B. 1992. Le Golfe Arabo-Persique vu depuis Mari, in J.-M. Durand (ed.), Florilegium marianum: receuil
d'etudes en l'honneur de Michel Fleury: 69-79. Paris.
HARPER, P. et al. 1992. The royal city of Susa. New York (NY): Metropolitan Museum of Art.
HOJLUND, F. & H.H. ANDERSON. 1994. Qala'at al-Bahrain 1: The northern city wall and the Islamic fortress. Aarhus:
Jutland Archaeological Society. Publication XXX.1.
KJAERUM, P. 1983. Failaka/Dilmun: the second millennium settlement 1(i): The stamp and cylinder seals. Aarhus:
Jutland Archaeological Society. Publication XVII.i.
KILLICK, R.G. et al. 1991. London-Bahrain Archaeological Expedition: 1990 excavations at Saar, Bahrain, Arabian
Archaeology & Epigraphy 2(2): 107-37.
LOMBARD, P. & M. KERVRAN (ed.). 1989. Bahrain National Museum: archaeological collections 1. Bahrain: Ministry of
Information.
MOON, J. et al. 1995. London-Bahrain Archaeological Expedition excavations at Saar: 1993 season, Arabian
Archaeology & Epigraphy 6: 139-56.
MORTENSEN, P. 1986. The Barbar temple: its chronology and foreign relations reconsidered, in al Khalifa & Rice (ed.):
178-85.
6 of 7
28/01/2015 03:43
https://www.questiaschool.com/read/1G1-20202121/the-site-of-saar-d...
NESBITT, M. 1993. Archaeobotanical evidence for early, Dilmun diet at Saar, Bahrain, Arabian Archaeology &
Epigraphy 4: 20-47.
POTTS, D.T. 1990. The Arabian Gulf in antiquity 1. Oxford: Clarendon Press.
1993. Rethinking some aspects of trade in the Arabian Gulf, World Archaeology 24(3): 423-40.
RAO, S.R. 1986. Trade and cultural contacts between Bahrain and India in the third and second millennia BC, in al
Khalifa & Rice (ed.): 376-82.
READE, J. & R. BURLEIGH. 1978. The 'Ali cemetery: old excavations and radiocarbon dating, Journal of Oman Studies
4: 75-83.
WOODBURN, M & H.E.W. CRAWFORD. 1994. London-Bahrain Archaeological Expedition: 1991-2 excavations at
Saar, Arabian Archaeology & Epigraphy 5: 89-105.
Q uestia School,a partofG ale,Cengage Learning.w w w .questiaschool.com
Publication inform ation:Article title:The Site ofSaar:D ilm un Reconsidered.Contributors:Craw ford,H arriet-Author.Journaltitle:
Antiquity.Volum e:71.Issue:273 Publication date:Septem ber 1997.Page num ber:701+. 2008 Antiquity Publications,Ltd.
CO PYRIG H T 1997 G ale G roup.
This m aterialis protected by copyrightand,w ith the exception offair use,m ay notbe further copied,distributed or transm itted in any
form or by any m eans.
7 of 7
28/01/2015 03:43
