Salinity
Soil mapping in advance of major irrigation projects has demonstrated that water-
logged so-called hydromorphic soils are particularly common within the floodbasins
and lower levee slopes of southern Mesopotamia. These soils tend to become saline
when the water table remains close to the ground surface so that osmotic processes
cause salts to rise and accumulate near the surface (Buringh 1960 ). Since Jacobsen and
Adams ( 1958 ) suggested that a shift towards the more salt tolerant barley was associated
with increased salinisation, there has been a vigorous debate concerning the role of
salinisation in the demise of Sumerian civilisation. Although this model has been
disputed (Powell 1985 ), there can be little doubt that if irrigation spread down levee
towards flood basins, there would have been greater loss of crops due to salinisation.
However, as discussed below, because cereal cultivation may have been mainly confined
to the levee slopes during the third and even the early second millennium BC, the worst
affects of salinisation may have been avoided.
BROAD PATTERNS OF WATER DISTRIBUTION
Whereas the conventional model suggests that the lion’s share of irrigation water
during the third millennium BCderived from the more readily controlled and less
incised Euphrates River (Jacobsen 1960 ), it has been argued that the Tigris also played
a role in the supply of water to the lower plains (Potts 1997 : 7 – 10 , 26 ). This debate must
be contextualised within the dynamic and branching nature of the Mesopotamian
rivers which shifted significantly throughout the Holocene.
During the earlier Holocene a joint Tigris–Euphrates channel is thought to have
flowed through the centre of the plain (Paepe 1971 ), with the two rivers subsequently
separating so that the Tigris migrated to the east and the Euphrates to the west (Adams
1981 ; Algaze 2008 , fig. 5 ; Hritz 2010 : 187 ). After these early and rather obscure phases,
the record becomes clearer and ancient levees in the region of Sippar demonstrate that
during much of the third and second millennium BCthe Mesopotamian plains received
a significant amount of water from branches of the Euphrates that radiated out from
the region of Sippar and contributed multiple channels that ultimately flowed to the
Sumerian cities of the lower plains (Jacobsen 1960 ; Cole and Gasche 1998 ). On the
other hand, Heimpel ( 1990 : 213 ) has shown that the watercourse that flowed past
Girsu, Umma and Adab was the Tigris or a branch of it. Unfortunately, this recon-
struction runs into the problem that east of Sippar in the northern plains levees of
branches of the Euphrates extended east as far as the Tigris River and appear to have
watered the southern plains (Cole and Gasche 1998 : map 5 ). Alternatively, Hritz shows
that a massive levee within the Diyala region may have been an earlier course of the
Tigris, which would then have supplied southern Sumer with irrigation water from a
northerly direction (Hritz 2010 : figs. 13 and 14 ). This reconstruction also places the
Tigris at a level which would have enabled it to supply irrigation water by gravity flow
to a much larger area without the need for labour intensive water-lifting devices.
Illustrative of the complexities of these dynamic river systems is that of the water
supply of Umma which had two principal watercourses: the Idigna and the Iturungul,
both of which were its major lifelines (Adams 2008 : 6 ). Whereas the former was a
branch of the ancient Tigris, its clearly canalised offshoot to the south, the Iturungul,
ultimately joined a major branch of the ancient Euphrates not far downstream from
–– Tony J. Wilkinson ––