Despite the need to rid the land of water, which was arguably as important as
bringing water to it, especially to safeguard against salinisation, there is little reference
to drains (Postgate 1994 : 181 ). Nevertheless, it is likely that some linear soil marks evi-
dent on the ground and satellite images (and perhaps of later date than the Sumerians)
are probably drains rather than canals.
Features along canalsAfter water was conducted along canals it was necessary to get this water into sec-
ondary, tertiary and smaller channels, and ultimately into the fields. This was probably
effected by using either regulators or weirs. Probably the best-known archaeological
example of a canal regulator was that excavated by Parrot at Tello, ancient Girsu (Parrot
1948 : 213 ; Potts 1997 : 20 – 21 ). Constructed of baked bricks and bitumen, this feature
appears to have taken a large volume of water and funnelled it through a narrow brick-
built channel which would have constrained flow to a given discharge after which the
water then flowed into a widening funnel, presumably back into another channel
(Postgate 1994 : fig. 9. 2 ; Pemberton et al. 1988 : fig. 9 ).
In modern irrigation systems it is necessary to employ distributors, with or without
gates, to subdivide flow into smaller canals. In Iraq, this role is traditionally played by
dams or weirs constructed of reeds, palm trunks and mud, which operate to both raise
and divert flow into subordinate channels (Rost and Hamdani in press). Because the
water level is usually low when it is required for irrigating cereals, a regulator is required
to raise the water to a sufficient level for it to flow into smaller channels or fields.
Accordingly ‘the cultivators build a series of dams each of which holds up the water
till the fields in the neighbourhood have been flooded, and it is then broken to let the
water pass on to the next dam’ (GB Admiralty: 1918 : 434 ).
Here it has been necessary to digress into traditional irrigation methods because
there is no universal agreement regarding the meaning of all the Sumerian terminology
for features referred to in the texts. For example, the meaning of terms such as a-ga-
am(perhaps an artificial pond to slow down the flow of water (Civil 1994 : 130 )) and
nag-ku 5 (lateral ponds or reservoirs into which excess flood waters were diverted for
later use (Civil 1994 : 133 ; Zettler 2003 : 32 )) continue to be debated. Particularly vexing
is that because these Sumerian features do not seem to occur in the modern landscape,
it appears that the Sumerian landscape may have been rather different from that of
today. Adding to the enigmatic nature of these features is that if the nag-ku 5 were in fact
reservoirs, they did not appear to have held sufficient water for the irrigation of fields
of grain (Hunt 1988 : 194 ), nor for holding large amounts of flood water.
Finally, the water reached the fields, which were frequently long and narrow with
the upper (short) end facing a water course (Civil 1994 : 125 ; Liverani 1996 ). However,
occasionally there is reference to the irrigation of palm orchards with an under-storey
of other plants such as pomegranates (Civil 1994 : 131 ).
Overall, what would have originally been a natural fluvial system that evacuated
excess water and sediments from the Tigris–Euphrates basin into the Gulf, under
human management eventually became a massive sediment trap. This is because most
irrigation practices either imposed impediments to the flow, slowed the flow, or by
allowing water to infiltrate the soil, precipitated excess silts and clays. Therefore not
only was it a massive labour-intensive task to remove the accumulated silt and clay, as
–– Hydraulic landscapes and irrigation systems ––