A. Trevor Hodge
The qanat was a type of water system that originated in Iran, spread throughout the Mediterranean with the expansion of the Saracen Empire, and is still in widespread use throughout the Middle East. It apparently did not get as far as Morocco, and though qanats exist there, the technique seems to have been brought there by Islam, from Spain after the Islamic conquest, arriving in Morocco only around the eleventh century A.C.; in Antiquity it did nevertheless get as far along the North African shore as Algeria.
The technique was probably derived from mining, and the earliest qanats seem to go back well before the eighth century B.C. Essentially, the qanat was a tunnel driven deep inside a hillside to tap a spring or aquiferous stratum inside it. From this underground source the tunnel had to run downhill, to keep the water flowing, until it eventually surfaced at ground level to deliver the water to the consumers (see figure 1). This meant that the hillside slope had to be steeper than the tunnel, so that the further the tunnel ran downhill, the closer it got to the surface. Throughout its downhill length, the tunnel was served by a regular series of vertical ventilation shafts, 5–20 m apart.
The name qanat is derived from the Akkadian qanu, “reed”, from which one finds the Biblical city of Qanatha, named from its qanats (Num. 32.42; I. Chron. 2.23; in the King James version it is spelled “Kenath”). In North Africa the name “foggara” was often used, also “kariz” and “madjira”, which last, from the profusion of Saracen qanats, gave its name to the city of Madrid (“Place of qanats”). In the New World the qanat even got as far as Los Angeles, where the earliest water supplies were so organised by the conquistadores.
The heart of the qanat was a vertical shaft dug down to intersect
1 On qanats see Henri Goblot 1979; this is the definitive work. See also the trea-
tise on qanat-building by Mohammed al Kardji (c. 1010 A.D.), available in a French
edition: Mazaheri 1973. Further bibliography in Hodge 1992, 391–2 n. 6. A read-
ily available account is Wulff 1968.