In northern Mesopotamia, it is during the Late Uruk period that the first urbanization process46 is thought to have started. This ended rather abruptly for still unknown reasons and seems to have been followed by a short period of dark age. few excavated large sites, except Brak, give an uninterrupted sequence from the fourth to the third millennium, and up to now there has not been enough information to understand what happened. It is only several centuries later, slightly before the middle of the millennium, that a second phase of urbanization is detected, contemporary with the Early Dynastic III of Southern Mesopotamia (Akkermans and Schwartz 2003: 233-87).
Curiously though, it is between these two urban phases, somewhere during the Early Jezirah I and/or II period47 that a whole set of new circular settlements made their first appearance. These are usually known as Kranzhugel because most of them consist of a circular mound surrounded by a ring. Because they are being excavated, the best known are Tell Chuera, Tell Beydar, and Mari,48 but there are about a dozen of them (see fig. 10.2). Their size is mostly much larger than the usual size of settlements in this area.49 Other smaller
5 Sites like Al-Rawda, Sheirat, Umm el-Marra.
6 In Syria, no major city other than Habuba Kabira is known for this period. But it now seems that Brak and Hamoukar, like Gawra, extended to “city” sizes on local grounds already at the beginning of the Uruk Period or Late Chalcolithic 2-3 (beginning of the fourth millennium). Unfortunately, there is no comparable information in southern Mesopotamia at that time. How different the organization of those northern
BERTILLE LYONNET
Contemporary sites,50 51 such as Khazne, Rad Shaqrah, or Hassek, are also circular in shape though they do not present this ring.
Besides their architecture, another odd feature of all these circular sites is that they are exclusively situated in the semi-arid zone of northern Mesopotamia, where previous sedentary occupation was extremely rare (hole 2002; hole and Kouchoukos in press). Finally, it should be stressed that the cultural zone to which they belong is distinct from the area where “Ninevite 5” pottery dominates, since only a few sherds of this type of pottery have been found there (Lyonnet 1996; ur and Wilkinson 2008). Though pots and peoples rarely match, this distinction, added to all the other differences mentioned above, could also be a sign of a different ethnic group from that of the Ninevite area. We consider that question in the conclusion, but let us first consider three of these round cities more precisely.
The case of mari
Mari is the only round city that has been excavated enough to give some information on the period of its foundation (Ville I) (Margueron 2004: 49-123), even though it is known mainly through small soundings. It is also the only circular site known along the Euphrates11 and this position in the river valley has led some researchers to doubt that it was one of the Kranzhugel because all the others are in the steppe, even if along wadis or depressions favorable for the collection of temporary water. however, the site is not close to the river but at some distance on the holocene terrace; we return to the implications of this position below. The site has been extremely damaged and only part of it still stands, but the undisputable circularity of the inner city,52 as well as the existence of the ring outside are, I think, sufficient features to correlate it with the other round cities (fig. 10.3).
According to J. Margueron Mari was founded as a new town on virgin territory in the beginning of the third millennium.53 The outer ring was established at the same time and consists mainly of a huge heap of earth over a core made of a stone (gypsum) wall surmounted with mudbricks. No repair of this exterior ring has been done during the entire period I, said to have lasted about 250 years.14 Though considered as a protection, it is not seen as a real defensive rampart at that time, and Margueron thinks that it mainly played a role against flooding from the Euphrates. The space between the ring and the town itself (300 m) was free of any structure.
The inner town (diameter 1,200 m) is enclosed by a real rampart of mudbricks on a stone base, with rectangular towers and doors. No repair was made on the inner rampart during the course of period I. From one excavated door, one street goes north probably to the center of the town, and several other streets are supposed to also converge there from the other doors.
13 The foundation of Mari is dated between 2950 and 2860 B. C. (Early Dynastic I) by thermoluminescence and radiocarbon data, but ceramics are rather considered as belonging to the Early Dynastic II period (Margueron 2004: 8-9, 57-59). The end of period I is not yet precisely dated but it should be either late within the Early Dynastic II or at the beginning of Early Dynastic III (see nn. 14 and 15). Ville II starts during the Early Dynastaic III period (thermoluminescence date around 2550-2500 B. C.).
14 It probably lasted longer.
Several excavations and soundings into different parts of the site show, for Ville I, a succession of different levels. Each phase presents a totally new arrangement even within one small area, and it is not yet possible to establish connections between the different areas:
• in the west (ishtar temple area), from bottom to top, one large building has been discovered, made of bricks (only a few left) on large stone (gypsum) foundations (of the same kind as the wall inside the outer ring); it contains several small square rooms and one larger rectangular room but its function is unknown. it is followed by what seems to be a stone platform, then by small mudbrick multi-room houses and, finally, at the end of ville i, by three large stone-vaulted tombs that apparently contained rich material but have been disturbed.
• in the same zone (area L), five levels were excavated with craft installations (metallurgy and pottery kilns) associated with modest architecture. in one of the levels, impressions of wooden wheels from a cart and two equid skeletons have been discovered. The upper level gave tombs with a very rich material and cylinder seals, one of which is dated to the late early Dynastic II period.15
• in the north (area B), eight levels have been noticed, each showing thin straight or curvilinear walls and tombs. unfortunately, the sounding was not large enough to understand the architecture.
• under the palace, the first level encountered on virgin soil is that of six tombs16 deprived of funerary material, built in bricks and compared to those of Kheit Qasim in the Hamrin (iraq): they are deposited over a layer of earth said to have been laid artificially on virgin soil.54 Over this level of tombs, circular hearths and layers of ash have been discovered. in the following phases, the area served for metallurgy, evidenced by the presence of a large hearth and installations associated with poor architectural remains. several layers of decay follow until the appearance of another craft installation with a pottery kiln, hearths, and tombs. All this is finally covered by 1 m of both decay and leveling done before the reconstruction of ville ii; because of the presence of tombs, Margueron thinks that it is houses that had probably been leveled.55
• To the east of the palace (area H), nine different floor levels were found, eight of which are associated with small mudbrick architecture and two hearths, probably houses.
• under the Ninhursag temple, a wide wall in mudbricks could be part of a platform 2.2 m high.
To sum up, the city of Mari ville i does not present any proof of real urban organization (margueron 2004: 103), except for the general shape and street plan. No system of collecting water (sewer or drain) is visible. Only one important building (said to be probably an administrative building) and platforms (one of which is mentioned as possibly a religious building) have been found, but the rest of the architecture is said to be modest, and the houses are small
15 This information appears only on p. 296 under the legend of fig. 285 (margueron 2004). The author considers that it means that ville i lasted longer than said earlier.
16 The tombs are said to be probably part of a cemetery.
BERTILLE LYONNET
And often rebuilt and/or restored. Furthermore, the architecture seems very varied. It is difficult to have complete plans of the houses but one is said to have a very small central space, while several others have only one standardized rectangular room surrounded by a courtyard with small utility rooms around it. The author compares this architecture to houses at Tell Arad in southern Levant or to others at Tell Mohammed Arab in northern iraq (Margueron 2004: 105). Tombs are also of various kinds and present diverse degrees of wealth. The richest and largest are apparently all in the last level of ville i. finally, ceramics also combine different traditions: some, like “Grey Polished Ware” and “scarlet Ware,” are imported from mesopotamia, others, like “metallic” and “Ninevite Ware” from northern mesopotamia, while the local ceramics seem to mainly consist of hole-mouth cooking ware with crescent-shape lugs and of rather coarse ware (margueron 2004: 105, legend of fig. 76).
All these data do not fit well with the capital city that margueron claims it to be. Rather than an administrative place, the large building could just as well be a collective storage bin, since only two doorways have been found. The very modest architecture of the houses, their constant repair and rebuilding, and the comparison proposed by margueron himself with Arad, could rather make them temporary dwellings for mobile groups, living there only in winter time. This does not mean that during the course of ville i some of them would not have settled longer. This could well have been the case at the end of this period around what can be considered as the tombs of ancestors (for such tombs, see Peltenburg 1999; Porter 2002).
Nevertheless, margueron has vigorously rejected this idea of mobile groups arguing that the main building presents a degree of skill too high to have been built by semi-mobile groups even in the course of sedentarization (margueron 2004: 103) and that metallurgical activities could not have been managed by such people, though there are several examples of such connections (Levy, this volume; mouton 1999; Kohl 2007: 149-50).
Because he defends the idea of a city on a mesopotamian model, he also considers that, being huge in size, mari was inhabited by a large sedentary population. This statement implies a series of other postulates dealing with regional water management, and especially the construction of three major canals (margueron 2004: 68-82):
• Water was needed inside the city, both for everyday life and for trade (harbor) and, since there is no evidence of wells and the Euphrates is too far away, therefore a bychannel, or derivation canal (canal de derivation) was built that passed through the city itself.
• irrigated agriculture was necessary to feed the people, and therefore an irrigation canal was built on the Holocene terrace.
• Trade was the vital concern of this capital city, and to favor communications both with the north and with mesopotamia, therefore a navigation canal 120 km long was built on the opposite (left) bank of the Euphrates, with its head about 20 km up the mouth of the Khabur and its end in the Baghouz area.
Mentions of boats and of canals do exist in the mari texts of the eighteenth century B. C., but for the moment nothing in them allows one to say that a canal and a harbor were located within the city itself or to correlate the mentioned irrigation channels — the dimensions of which are unknown — with the ruins still visible on the surface of the Holocene terrace.
Each of the canals proposed by margueron is a huge enterprise and the first two would have dangerously threatened the city. The depth of the derivation canal needed for boats to enter the city would have made it a gateway both for the annual floods of the Euphrates and for human invaders, as mentioned by the author himself (Margueron 2004: 69). Dikes would have been necessary but none has ever been found. As for the irrigation canal, it is over 100 m wide and can be followed at least 17 km along the Holocene terrace but would have been much longer depending on where its head was. Not only does it present the same problem as the derivation canal (gateway for the annual floods), but also, as stressed by B. Geyer (Geyer and Monchambert 2003: 70-74), irrigation on this terrace without draining causes an accumulation of salt which ultimately puts any cultivation in danger. Furthermore, the necessity of such a huge irrigation canal has to be questioned since it has been established by the Mari texts that only winter crops were produced in the early second millennium,56 so that water would have been needed only during late winter and early spring (Geyer and Monchambert 2003: 111, 194-95; Margueron 2004: 37, 81), when it rains the most and when the level of the Euphrates is at its highest. Simple small channels with elevating machines, such as shadufs,57 would have been able then to provide enough water for the fields.
Finally, all attempts to date these canals have failed to show any trace going back to the third millennium. No section in what is considered to be a trace of the derivation canal has been excavated, and no sites dating to the third or early second millennium have been found along the irrigation canal (Margueron 2004: 37, 81). Only three sherds collected on the surface of a canal or of a secondary branch are dated to the Middle Bronze Age (Geyer and Monchambert 2003: 194).
The navigation canal, Nahr Dawrin, is 120 km in length and is sometimes dug within the rock of the above plateau. If really made during the course of Ville 1, it would have been a colossal work, unique in the ancient Middle East for that period. Its position on the other side of the river, correlated to the fact that neither its head nor its end are directly across from Mari, makes it difficult to believe that it was done under and for Mari’s control. Only three sites dating to a phase earlier than 1slamic have been identified in a survey made along the canal and, as said by S. Berthier, they could be the ruins of a seasonal occupation (Berthier 2001: 163).58 Later dates have a much more secure basis. Its construction could be due to the Assyrians in relation either to Tell Shekh Hamad/Dur-Katlimmu (Ergenzinger and Kuhne 1991), because of the existence close to it of a canal on the right bank of the Khabur joining the Nahr Dawrin, or to Tell Masaikh/Kar Assurnasirpal because it stands right above the city. But it could also date as late as the 1slamic period since only 1slamic sherds have been found in the four sections made at different places along the canal (Berthier 2001: 36, 52, 63). Furthermore, the height of the water flowing in the canal is said not to have exceeded 1 m, making the Nahr Dawrin not navigable for boats.
To sum up, there is no proof for any of the assumptions that Mari 1, when founded as a Kranzhugel, was a capital city on the same model as those of southern Mesopotamia. This statement rests only on the a priori idea advanced by the excavator, “if we exclude the irrigation and navigation canals, we are condemned not to understand the foundation of Mari”
21 1n Geyer and Monchambert (2003: 212, n. 73), only one site, Diban 4 (site number 84) is said to be Early or Middle Bronze Age, but the data concerning this site (page 91) mention that ceramic is rare and the date Early Bronze Age or Middle Bronze is followed by a question mark.
BERTILLE LYONNET
(translated from Margueron 2004: 76). But, as I try to show, there are other possibilities for understanding the position of this settlement. if we see it as a central place for the gathering of mobile groups living there only part of the year, or as a place where only part of the tribe lived all year around while the rest wandered in the steppe around, there is no need for such huge water management; simple irrigation with elevating machines would have been sufficient to grow enough barley for the members of the tribe and their sheep. Mari would then have fitted naturally in its environment.
World History
