large areas of land, albeit at a relatively low yield and return, which necessarily resulted
in the deployment of large areas of fallow also available for grazing, which resulted in
significant ‘wealth on the hoof ’. This exceeded that of the owner occupiers, who had
less land at their disposal for grazing. In other words, tenants of larger holdings, despite
their requirement to pay a significant proportion of their production to the estate
owner, would generate more wealth. As Fernea ( 1970 : 48 ) reminds us, the role of fallow
is therefore not only essential to allow for leaching of salts, it is also advantageous for
building up large herds.
Because share cropping on large estates was perhaps more profitable than income
from dependent farmers (Fernea 1970 : 48 ), it is therefore possible to imagine (but
difficult to prove) that in early Sumer, estates which taxed dependent farmers could
grow at the expense of smaller household plots. Furthermore, large land holdings such
as estates would also benefit from economies of scale because many of the tasks
performed were more efficient when they were conducted on a large scale (Postgate
1994 : 188 ). Because smaller holdings were probably associated with palm gardens along
the levee crest, the estates probably grew relative to them and extended down slope to
form the cereal zone described above.CONCLUSIONS
Water clearly played a fundamental role in the development of Sumerian civilisation,
although the Sumerian farmer often had to devote as much time in getting rid of it as
collecting it to nurture crops. Whereas Ubaid landscapes probably formed a mosaic of
riverine gardens, wetlands with ‘turtle back’ islands and desert steppe, by the third
millennium BCsettlements appear to have gravitated towards levees so that both towns
and irrigation systems became aligned. Although marshes would have remained
important, some were evidently drained whereas others would have grown as a result
of the discharge of excess water from canals.
When cuneiform and field evidence are combined, it can be inferred that areas such
as that around Umma consisted of relatively narrow cultivated zones between which
would have extended areas of desert-steppe and marsh, which was itself a valuable
resource. Although evidence exists for the construction of larger canals, especially in
the later third millennium BC, most appear to have been less than 5 km long and only
in the Ur III-Isin-Larsa period did the landscape include longer canals (Adams 1981 :
fig. 31 ), a pattern that became particularly evident by the later second millennium BC
(Ur this volume fig. 7. 8 ).
Sumerian irrigation, rather than being a monolithic system dominated by a
theocratic state, therefore was probably more heterogeneous and less centrally managed
than previously thought (Rost 2010 : 3 ). Moreover, there is now little support for the
notion that the need to irrigate large irrigation systems actually created the state as
argued by Wittfogel ( 1957 ). Instead, shorter canals leading from the main channels
could have been readily organised by Sumerian communities in the same way as recent
kin-based social systems in southern Iraq (Fernea 1970 ). There is also no reason why
the state should have controlled everything because in modern systems both state-
controlled administration (upstream) and independent or privately owned systems
(downstream) co-exist (Ertsen 2010 ). In other words, in Sumer, the state or the king
could have sponsored and built the main infrastructure and controlled estates, while
–– Tony J. Wilkinson ––