443 Water Management in Assyria from the Ninth To

ARAM, 13-14 (2001-2002), 443-460 S. DALLEY 443

WATER MANAGEMENT IN ASSYRIA FROM THE NINTH TO THE SEVENTH CENTURIES BC1

STEPHANIE DALLEY

The Assyrians, building up from the ninth century onwards, ruled an empire that eventually stretched from Tarsus in Cilicia to the eastern border of Egypt, from Armenia to Bahrein. Enormous wealth and talent flowed into the heart- land through tax, tribute and trade, through deportation and the magnetic attraction of working for the most powerful kings in the world. From an Assyrian perspective it was the duty of a successful king not only to enrich the nation through conquest, but also to display power through fine buildings and the patronage of great art, through engineering works and the conspicuous management of water provided for great cities, parks and gardens, notably at Nineveh, Nimrud and Arbela. Provided that the technology was available, the Assyrian king had the manpower and the raw materials to achieve whatever he wanted, regardless of time, wastefulness and the health and safety of his workmen. Assyrian achievements in water management are seldom mentioned in books which claim to describe the early development of technology. Many books and articles on ancient technology, such as those of White, Landels and Oleson, begin with the Greek world, or restrict themselves to purely archaeological evidence from the Near East. The volume of World Archaeology in 1980 which was dedicated to the subject of Water Management did not cover the Assyrian evidence. The Bulletin of Sumerian Agriculture did not cover the Assyrians in its volume on irrigation and water control.2 Another cause for neglect lies in differences in the way that various branches of learning have developed. As a result of several factors including especially the Renaissance, Greek culture has an unbroken thread of tradition, whereas the civilisation of Mesopotamia has had to be reconstructed after complete loss. By comparison, the literature of the latter is specialist, wide- spread and disorganised, and the sources fragmentary. The complexities and ambiguities of the cuneiform writing system mean that one often has to wait

1 I would like to thank Dr. Shafiq Abou-Zayd for arranging such a useful conference; Dr. Andrew Wilson for bibliography and subsequent fruitful discussions on qanats; and Dr. Peter Kingsley for help with bibliography. 2 Notable exceptions include M. Drower, “Water supply, Irrigation and Agriculture”, in A History of Technology, ed. C. Singer et al., (Oxford 1954), esp. pp. 531 and 553; and F.W. Rob- ins, The Story of Water Supply, (Oxford 1946), 36-7. 444 WATER MANAGEMENT IN ASSYRIA for a duplicate with variants to turn up before one can be certain of understanding a long-known inscription. Moreover, the phenomenon of the named, national culture-hero / inventor in Greek tradition stands in sharp contrast to the anonymity of inventions and achievements in the ancient Near East. This is as true of technical inventions as of authorship: we do not know who engineered Sennacherib’s aqueduct, but we do know that an engineer named Eupolinos constructed a tunnel and water system for Polycrates on Samos. In ancient Mesopotamia, creative efforts, whether in literature or technology, were either anonymous or attributed to the royal patron. But among Greeks and Romans, famous men such as Archi- medes had inventions or principles attributed to them, and some of these attri- butions are demonstrably wrong. In the case of Pythagoras, we know that some of the work credited to him is certainly earlier. In the case of Hippo- damus of Miletus, the grid-plan for city layout, attributed to him, is attested much earlier, in the East Mediterranean at Tell Dor,3 and even earlier at the Urartian fortress of Bastam in NW Iran.4 Critias5 attributed to the Athenians the invention of the potter’s wheel and the clay oven, but few archaeologists now would give him credence.6 Terpander was credited with inventing the 7-string lyre, which is now attested from the Minoan Bronze Age.7 In some of these cases re-invention may be at issue rather than deliberate falsification or misguided chauvinism. Above all, however, there is still uncertainty over exactly what technology was available to the Assyrians from the ninth century onwards. This is partly due to difficulties involved in reading and understanding cuneiform texts, along with difficulties in finding and interpreting material remains. On this occasion two items in particular are investigated: whether or not the qanat was known from the late eighth century BC, as has been claimed in the past, and whether the water-raising screw was in use at that time.

1. THE QUESTION OF QANATS

The qanat is usually defined as a sloping underground tunnel with vertical access shafts at regular intervals, designed to collect and conduct ground water, sometimes accessing a “mother well” at its starting point. It was thought by some to derive its inspiration from shaft-mining, which is hard to date, but definitely attested by the Roman period. Qanats are likewise hard to date, but

3 E. Stern, Dor, ruler of the seas, (Jerusalem 1994), 159. 4 W. Kleiss, Bastam (Berlin 1979). 5 As quoted by Athenaeus in Deipnosophistae I 28. 6 The phenomenon is discussed by A.Kleingünther, Protos Heuretes (Leipzig 1933). 7 A. Lesky, Geschichte der griechischen Literatur, 3rd ed. (Bern and München 1971), pp.155-156. S. DALLEY 445 were certainly used in the Roman period, and are generally thought to have spread from origins in Persia. One of their great advantages is that the flow of water is protected from pollution, by comparison with an open watercourse.

THE NEGOUB TUNNEL The earliest great construction work in which the Assyrians are known to have exhibited their ability to manage water on a grand scale is the series of tunnels with its associated canal, which brought clear mountain water from the Upper Zab river into Nimrud, ancient Calah, via a tunnel cut through an out- crop of conglomerate rock at Negoub.8 The original canal named Patti-hegalli “Abundance Canal” was cut by Assurnasirpal II (884-859), who also built the great Northwest Palace at Nimrud. It extended for more than 19.5 km, requir- ing several weirs on its way. It was probably associated with 7km of tunnel, dug by joining between the bottoms of vertical shafts which had been cut through the conglomerate at regular intervals, and was intended to irrigate fields outside the city as well as for urban use, especially for the palace garden described in the famous Banquet Stela inscription of Assurnasirpal.9 Owing to the flat gradient, the canal may have silted up rather fast. Of two later tunnels in the same area, the earlier was cut by Tiglath-pileser III (744-727 BC).10 It was blocked up, and both the earlier installations were replaced under Esarhaddon (680-669 BC), who left an inscription there (now very damaged),11 recounting that the older tunnel of Assurnasirpal II had silted up. Esarhaddon’s tunnel had shafts about 15m deep with stairs cut along one side, and the height of the tunnel itself is estimated between 4 and 5m. The entrance to the tunnel consists of three narrower tunnels which join the main tunnel after 8m, and their purpose was probably to accommodate sluice gates of man- ageable size. Grooves for those gates were still observable in 1985. In order to slow the flow of water as it entered the tunnel, there was a 3 m rise, a deliberate “negative grade”, over the length of the tunnel, to dissipate the water current and prevent erosion of the banks at the entrance to the tunnel. These works demonstrate a system of waterworks with canals and tunnels, in which the total length was at least 24 km. The evidence shows how trial and error rather than any application of theoretical principles governed the scheme, and how different phases of work were spread over three centuries. Although the tunnel was constructed with vertical shafts at intervals, Davey stated clearly that it was not to be confused with a genuine qanat, which collects ground

8 C.J. Davey, “The Negub Tunnel”, Iraq 47 (1985), pp. 49-55. See also J.N. Postgate, “Negub”, in Reallexikon der Assyriologie (1999), p. 207. 9 A.K. Grayson, Assyrian Royal Inscriptions vol. 2 (Wiesbaden 1976), p. 172 gives a translation. 10 D. Oates, Studies in the Ancient History of Northern Iraq, (Oxford 1968), p. 47. 11 R. Borger, Die Inschriften Asarhaddons (Graz 1956), pp. 35-36. 446 WATER MANAGEMENT IN ASSYRIA water and is not connected to river or spring water. However, it clearly has most of the features typical of a genuine qanat.

URARTIAN WATERWORKS IN E. ANATOLIA Some inspiration for Assyrian water management of this kind may have come from other countries, for instance from Urartu in eastern Turkey, where Assyrian kings campaigned, especially during the late eighth century, but also at an earlier date. To look at this evidence involves a new investigation concerning the origins of the qanat, for which improved textual understanding goes hand-in-hand with detailed evidence from archaeological work. When the Urartian king Sarduris I c. 830 BC chose Teushpa near Lake Van as his capital, he and his successor Menua were obliged to bring drinking water over a distance of 66km. because the lake itself has undrinkable water and none of its feeder streams is perennial. Retaining walls for this canal still stand up to a height of 20m, and the aqueduct which brought the water across a valley was built using tree trunks and earth packing, a construction which is thought to have survived with regular repairs and maintenance into the 19th century AD.12 Subsequently Rusa I (contemporary with Sargon II) or Rusa II (just after Sennacherib’s reign) built dams and made an artificial lake with a sluice gate, and a number of small reservoirs. Parts of these works are securely dated by inscriptions, some of them still in place. In a passage from the Letter to the Gods reporting in cuneiform Akkadian the Urartian campaign of 714 BC, Sargon II gave a brief description of the water supply system around the city of Ulhu, which he destroyed. When Laessøe in 1951 and 1953 wrote about this passage, claiming to have eluci- dated a description of a genuine qanat, the understanding of several lexical items was more uncertain than appears from his publications.13 It seemed plau- sible to suppose that the origin of qanats might be there, because many “Ar- menian” qanats are found in that area. However, it has become clear from extensive recent work on Urartian water management that qanats are not linked to Urartian settlement patterns, and so must be of a later date.14 Laessøe quoted two passages from Sargon’s description of waterworks which he found on campaign, admired and then destroyed in the Urartian town Ulhu. The translation given here is my own, making use of the various entries in the Chicago

12 G. Garbrecht, “The Water Supply System at Tuspa (Urartu)”, World Archaeology 11/3 (1980), pp. 306-312. 13 J. Laessøe, “The Irrigation System at Ulhu”, Journal of Cuneiform Studies 5 (1951), pp. 21-32. 14 For a review of Urartian waterworks, see O. Belli, “Urartian dams and artificial lakes in E. Anatolia”, in: Anatolian Iron Ages 3, ed. D. French and A. Çilingiroglu (1994), 9-30, and M. Salvini, Geschichte und Kultur der Urartäer, (Darmstadt 1995), section Wasserwirtschaft, pp. 126-128. S. DALLEY 447

Assyrian Dictionary, and differs from that of Laessøe in several important re- spects. Lines 202-204: Ursa sarru maliksunu ina bibl[at libbisu… mu]Òê mami usaklim palgu babil mê surduti ih[rima…. mê] nuhsi kima puratti usardi adappi la mina Òurrussa useÒâ[mma…. ]-lu-u usamkira qarbati “Rusas the king their ruler discovered… water-outlets… through his own inge- nuity. He dug a canal to bring flowing water and…. he made abundant water flow like the Euphrates. He made innumerable branch-canals come out from it, and…. had meadows irri- gated.” Lines 221-222 íd hiritu nar tuklatisu arursa askirma mê du[ss]uti ana rusumdi uter adappi halili silitte Òurrisa uha††im-ma sa […]-ti isqillassina ukallim samsu “I (Sargon) blocked the outlet of the canal, his mainstay, and turned the abundant water-supply into a quagmire. I sealed up the branch-canals, the halili-courses, branches from it, and I exposed the pebbles of [….] to the Sun.” Two of the words used, silittu and isqillatu, were not understood when Laessøe wrote his paper, and they are now understood, thanks to more evidence and to thorough lexicographical work carried out for the two great Akkadian dictionaries. Laessøe thought (by implication) that isqillatu must mean a shaft, since he restored the missing word before it as sapilti and trans- lated it tendentiously as “(even) the d[ee]p(est one) of their isqillatus”, giving a plural meaning to a singular noun, and not recognising that the sign sa before the break is probably to be connected with the suffix -sina to indicate an an- ticipatory genitive. We now know that isqillatu means a pebble, and silittu means a branch-canal. He took the phrase muÒê mami (literally “water-outlets”) to mean “qanat-openings in the hillsides” (virtually the same meaning as he implied for isqillatu), which cannot be justified, since qanat-openings or shafts are not outlets for water. If he meant qanat-openings to be the very end of the qanat, when the water reaches ground level, he had found no expression specific to qanat construction. An omen text makes it clear that when a watercourse is dammed or blocked, its pebbles “see the sun”, i.e. dry out on the old river-bed.15 The expression in line 222 uses the same imagery. Above all, however, Sargon’s description of causing a mire, a muddy place, rusumtu, is not applicable to a qanat, since qanats are blocked by filling in the shafts so that no water comes through, as is clear, for example, from Polybius’ description of

15 Details are given in CAD s.v. silittu, and s.v. isqillatu. 448 WATER MANAGEMENT IN ASSYRIA

Arsaces’ cavalry “choking up the shafts which went down into the underground channels”.16 Therefore the passage must refer to open canals. Thus it is clear that the passage cannot be taken to describe Urartian use of the qanat.17 Canals of various types, weirs, artificial lakes and cisterns were used to supply water for Urartian settlements, but not qanats.18

SENNACHERIB’S WATERWORKS FOR ARBELA A problem in identifying qanats similar to that described for the Negoub tunnels arises for an inscription of Sennacherib (704-681 BC) found on the opening of a tunnel built with vertical shafts at Arbela in eastern Iraq. Since there are several hundred qanats still traceable in the vicinity of that city,19 Laessøe thought that Sennacherib made a qanat there. The inscription is badly damaged, but enough remained for Fuad Safar to conclude that, since the engi- neers gathered water from three mountain rivers at Bastura, in the same way as was done at Bavian, this was not a kariz (qanat) system tapping subterranean waters, and should not be compared with the later qanat (kariz) which supplied the city. His report was in Arabic only, and so hard for most western scholars to evaluate, although it was well used by Drower.20 From the evidence of Arbela’s tunnel she deduced that “Most Assyrian cities, even those on a great river, probably relied for some of their fresh water upon this system.”

A QANAT AT BABYLON? Similar to the Negoub and Arbela tunnel systems is the one described for Babylon by Rassam, who described his discovery in the following words: “At the mound of Babel… uncovered four exquisitely-built wells of red granite in the southern center of the mound; three of which were situated in a parallel line21 within a few feet of each other, and one was some distance from them in a south- easterly direction… Each well is built of circular pieces of granite…. Each stone, which is about three feet in height, had been bored and made to fit the one below it so exactly that one would imagine that the whole well was hewn in one solid rock. On digging to the bottom of these wells, it was found to communicate with an aqueduct supplied with water from the Euphrates, or a canal which must have skirted the north-eastern corner of the mound. Even when I dug into the water-

16 Polybius, History X 28. 17 Laessøe’s conclusions were accepted by H. Goblot, Les Qanats, Une Technique d’acqui- sition de l’eau, (Paris 1979), pp. 67-69. 18 See now M. Salvini, “Pas de qanats en Urartu!" in Irrigation et drainage dans l'Antiquité, ed. P. Briant, (2001) pp. 143-155. 19 O. Mason, Iraq, Naval Intelligence Handbook, (1946), p. 93. 20 F. Safar and F. Basmachi, “Sennacherib’s Project for Supplying Erbil with Water”, Sumer II (1946) pp. 50-52 and III (1947) (Arabic only. I am grateful to Dr. Lamia Al-Gailani-Werr for translating it for me.) See also J. Laessøe, “Reflexions on modern and ancient Oriental Water- works”, Journal of Cuneiform Studies 7 (1953), p. 29. 21 Presumably he meant a straight line. S. DALLEY 449

course when the river was high, the water oozed out through the debris, though the Euphrates ran then about a mile from it. These wells, which were about one hundred and forty feet high when I uncovered them, and could not have been less than fifty or sixty feet higher originally …..”22 Rassam supposed that this underground water-conduit was supplied from a canal running northeast of the main mound at Babylon, and current knowledge of the topography suggests that the Nar-Banitu canal would fit.23 The position of the vertical shafts is not given on any plans of Babylon as far as I have been able to discover. No direct evidence for dating is available, but two main possibilities can be recommended: either the time in the mid sixth century BC when Esarhaddon and his son Assurbanipal rebuilt the city after its sack by Sennacherib; or in the mid fifth century when Nebuchadnezzar undertook major building work there. A third option would be the reign of Sargon II, whose building work in Babylon late in the eighth century is attested by inscribed bricks.24

“OVERSEER OF QANATS”? Although the Akkadian word qanû “reed; pipe, tube” (Assyrian plural qanâti) lies behind the Arabic qanat,25 and although a Neo-Assyrian profession sa eli qanâte “overseer of qanâte” might appear to justify a connection with qanats, in fact the word qanû means also a pipe or tube, and may refer generally to fittings for water supply and drainage. Interlocking terracotta pipes were used to carry water from at least the EBA in Mesopotamia, and Assyrian palaces were planned and built with sophisticated systems of water supply and drainage, as the evidence from Khorsabad shows. Therefore the profession sa eli qanâte, attested in a neo-Assyrian professions list and in two late Assyrian legal records,26 may mean overseer of qanats, but does not neces- sarily do so. Many of the great Assyrian palaces were excavated by tunnelling, or excavations were focussed mainly on sculpture and rich finds above floors, ignor- ing questions of water supply and drainage. Exceptional in this respect was Victor Place at Sargon’s new city Khorsabad.27 Still intact after 2,700 years were the fine vaulted conduits and channels underground that delivered water

22 H. Rassam, Asshur and the Land of Nimrod (Cincinnati 1897), pp. 352-54. I am indebted to Ian Cheyne for this reference. 23 See R. Zadok, Repertoire Géographique des textes cunéiformes 8 (Wiesbaden 1985), p. 366. 24 F. Wetzel, Die Stadtmauern von Babylon (Leipzig 1930) p. 64. 25 Goblot, Qanats, p. 19. 26 ADD 618 r.9 (=NALK 334), dated 657 BC, and ADD 112:9' (= NALK 356), dated 700 BC. The lexical list containing the profession, K.4395 (MSL XII 238-241) is very probably 7th century BC in date. 27 V. Place, Ninive et l’Assyrie, (Paris 1867), vol. 1, pp. 273-79; vol. III, plates 38 and 39. 450 WATER MANAGEMENT IN ASSYRIA and drained away sewage. A sturdily-built privy consisted of a large paving stone with a 10cm hole, beneath which a long vertical drop ended at the head of a sloping channel which in turn led into a splendidly vaulted main drain. This was constructed in baked brick and bitumen with a stone base, and included an arrangement whereby a few bricks could easily be removed from the vault, as with a manhole cover, to give access for repairs and maintenance. The excavator acutely remarked that a separate system of great complexity appears to have worked as a siphon to bring fresh water up into the palace, a conclusion which is linked to the total absence of wells or cisterns, according to their observation, within the palace. This is in contrast to the ninth century North-West Palace of AssurnaÒirpal II at Nimrud, in which wells were found. The contrast suggests a major technical advance linked to extensive planning, and makes possible a date as early as the reign of Sargon for the works at Babylon observed by Rassam. On present evidence one cannot rule out the possibility that the true qanat was invented late in the Assyrian period. At Negoub and at Arbela the technique of digging a tunnel for water by joining the bases of a line of vertical shafts was used, and when one such tunnel was blocked in order to re-route the water, it may have been observed that ground water continued to accumulate and flow in the abandoned tunnel. The conditions were ripe for the discovery of the true qanat. If so, it is possible that sa eli qanâte was a profession newly created in the late Assyrian period. The adaptation of the system, from carry- ing river water to tapping underground water in riverless piedmont, may have been made either at this period or later. If it is true that a genuine qanat in the Bahariya oasis of Egypt is cut by a 26th dynasty tomb,28 the Assyrian adminis- tration in Egypt could have introduced the system into the Bahariya oasis. It would therefore be possible that the Achaemenid Persians inherited the system from the Assyrians. Evidence for use of the qanat in Roman times, and so pre-Islamic, seems certain for Libya,29 for southern Palestine,30 for Roman forts in Syria31 and for

28 A. Fakhry, “Bahriya and Farfara Oases, Third Preliminary Report”, in Recent Explora- tions in the Oases of the Western Desert (Cairo 1942), p. 83. See also Fakhry, Bahria Oasis vol. 1 (Cairo 1942) fig. 4 on p. 37, but cf. fig. 52 on p. 95. Supporting evidence is still required. The many springs in the oasis and canals built from them may account for prosperity there under Ramesses III (1196-1080 BC) attributed uncritically to a supposed New Kingdom invention of qanats by S.Lee Gosline, Bahariya Oasis Expedition Season Report for 1988 (Texas 1990), Varia Aegyptiaca Supplement 3 (1990), pp. 27-30. 29 D. Mattingly et al., “The Fezzan Project 1998” Libyan Studies 29 (1998), pp. 137 and 141; “The Fezzan Project 1999”, Libya Studies 30 (1999), p. 139. 30 Zvi Ron, “Qanats and Spring Flow tunnels in the Holy Land”, in Qanat, Kariz and Khattara, ed. P.Beaumont, M.Bonine, K.McLachlan, (Cambridge 1989), pp. 211-236. 31 R. Mouterde and A. Poidebard, Le Limes de Chalcis (Paris 1945), chapter by Soubhi Mazloum, “L’organisation hydraulique de deux oasis antiques: Qdeym et ¨Amsareddi”, pp. 117- 126, dated late second or early 3rd century AD, p. 119. S. DALLEY 451

Oman.32 Whether it genuinely spread from Iran must now be open to question; Polybius’ statement, that the Persians who ruled Asia had built the qanats in Iran long before the reign of Antiochus III is unquestionable,33 but the technique could have come from elsewhere. It remains to be proved whether the earliest of the several qanats visible at Persepolis and Pasargadae can be dated to the Achaemenid period.34

2. WATER-RAISING SCREWS

The issue of invention prior to the lifetime of a traditionally credited inventor is keenly felt in the case of the water-raising screw. Archimedes certainly wrote a mathematical treatise on spirals, but scholars are divided over whether he invented the mechanical device too. Some scholars are inclined to dismiss the statement of Plutarch, who described how mechanics were entirely distinct from geometry, and how Hieron of Syracuse had trouble pursuading Archime- des to turn from abstract concepts to material things.35 Evidence is presented here to support the hypothesis that the water-raising screw was in use by 700 BC, long before the life-time of Archimedes in the third century BC, and belongs to a period of innovation and royal patronage for water engineering in general.

JERWAN, BAVIAN AND WATERWORKS FOR NINEVEH By far the most impressive work of water management is the one recorded in detail by Jacobsen and Lloyd in 1935. In a stupendous work of engineering Sennacherib channelled water from several mountain streams East of Nineveh across varied terrain, making 18 different channels, beginning at Bavian, some 50km away. The water crossed a wide valley on an aqueduct more than 280m long, built up on free-standing stone arches, 22 m wide with a surface incorpo- rating 40cm of waterproof lime cement, with a fall of 80 cm over 64 m, a fall rate of 1/80. It flowed into Nineveh at just the right height to water the garden built on the high citadel mound outside his palace. More than two million dressed stones were used. Some of the water was diverted to irrigate orchards and fields to the North of Nineveh.36 To accommodate surges in the water flow, Sennacherib made an artificial marshland in which he proudly an- nounced:

32 P.M. Costa and T.J. Wilkinson, “Traditional Methods of Water Supply”, ch. 3 in “The Hinterland of Sohar”, Journal of Oman Studies 9 (1987), p. 54. 33 Polybius, History X, 28. 34 E.F. Schmidt, Flights Over Ancient Cities of Iran, (Chicago 1940), esp. plates 2 and 14. 35 Plutarch, Life of Marcellus XIV 4-6. 36 J. Reade, “Studies in Assyrian Geography. Sennacherib and the Waters of Nineveh”, Re- vue d’Assyriologie 72 (1978), pp. 47-72 and 157-180. 452 WATER MANAGEMENT IN ASSYRIA

“herons from far away built their nests, and wild boar of the canebrake and forest cattle had large litters”. He was particularly pleased to have designed an automatic sluice which “opened by itself, without using a spade or a shovel, and allows the waters of prosperity to flow. Its gate is not opened by any action of men’s hands”.37 These works, in addition to those already discussed, show how completely the Assyrians had mastered certain principles of water engineering, although there is no evidence for the extrapolation of general principles rather than gaining experience through trial and error. The building of permanent bridges had a long history in Mesopotamia,38 and the Assyrians of the eighth and seventh centuries inherited a long tradition of expertise which was put to use in the Jerwan aqueduct. An example of Mesopotamian bridge building from the sixth century BC comes from excavation in an abandoned section of river bed at Babylon.39

THE KING’S INTEREST IN TECHNOLOGY One of the social factors which must have enabled water technology to progress to such an extent is the interest which two particular kings, Sargon II and his son Sennacherib, showed in engineering work. It was quite a short- lived fashion, and is particularly evident in inscriptions which relate to mining, smelting and casting bronze. One of the ways in which we can highlight this fashion in a more general way is to pick out a reference to a little known god who was the patron god of blueprints: engineering, design and architecture, and to pick out references to the profession he patronised. His name is Musda. Previously known only from inscriptions of the late third millennium BC, he has emerged only recently in the first millennium in a new edition of a cylinder of Sargon. Musda is the chief engineer, designer, ar- chitect, sitimgallu, of Enlil, the god who headed the traditional Mesopotamian pantheon. He is named by Sargon II in the context of planning and executing the new capital city at Khorsabad.40 The profession of which he is the divine patron is mentioned by Sennacherib in connection with the design and execu- tion of his palace and its gardens at Nineveh,41 and then by his successor

37 Bavian inscription, edited in T. Jacobsen and S. Lloyd, Sennacherib’s Aqueduct at Jerwan, Oriental Institute Publications 24 (Chicago 1935). 38 The Akkadian words are gisru (Assyrian) and titurru (Babylonian). For an early bridge see M-T. Barrelet, “Une ‘construction énigmatique’ à Tello”, Iraq 27 (1965), pp. 100-118. 39 Wetzel, Stadtmauern, pl. 51. 40 A. Fuchs, Die Inschriften Sargon IIs aus Khorsabad, (Cuvillier Verlag, Göttingen 1994), Cyl. line 60. 41 D.D. Luckenbill, The Annals of Sennacherib, Oriental Institute Publications 2 (Chicago 1924), 129 vi 57. S. DALLEY 453

Esarhaddon in connection with the rebuilding of Babylon.42 At other periods that type of person is simply not mentioned. It is probably significant for the Siloam inscription in Hezekiah’s tunnel in Jerusalem that this is a period in which it was acceptable for a king to recount engineering work in royal inscriptions, and in the case of Sennacherib, he even speaks of the toils and dan- gers which his workmen suffered in trying to achieve work of heroic propor- tions.

HANGING GARDENS AND THE WATER-RAISING SCREW Sennacherib’s South-West palace, and the gardens which were raised up beside the palace were described together in a text written on a clay prism, and called the palace with its gardens “a wonder for all peoples”. They were built expressly for the queen: “For Tashmetum-sharrat the queen my beloved wife whose form the Mother of the Gods has made more beautiful than that of all other women, I had a palace of love, joy and happiness built…. At the command of Ashur, Father of the Gods, and of Queen Ishtar, may we both live long together in health and happiness in these royal lodges and have our fill of contentment there.”43 Sennacherib wrote the long inscription in which he described his main achievements in improving Nineveh with a palace, a palace garden, and various measures for water supply. In places we can show that the text has incor- porated a shortened version of another text which dealt with one aspect in much greater detail; in other places the description makes abrupt transitions for which we suspect a similar editorial process is responsible.44 To show this feature we include here a long quotation from the clay prism inscription known now as the Palace Without a Rival. It contains both of the topics with which we are concerned: bronze-casting and water-control. In particular, it includes a passage concerning large bronzes of various shapes for various pur- poses, cast using a new method, and the installation of some of them for the purpose of raising water continuously or regularly. “Whereas in former times the kings my forefathers had created copper statues imitating real forms, to put on display inside temples, and in their method of work they had exhausted all the craftsmen for lack of skill and failure to understand pinciples(?); they needed so much oil, wax and tallow (or, lanolin) for the work that they caused a shortage in their own lands – I, Sennacherib,… knowledgeable in all kinds of work, took much advice and deep thought over making that work:

42 R. Borger, Die Inschriften Asarhaddons (Graz 1956), Ep. 24 iii 36. 43 R. Borger, “König Sanheribs Eheglück”, Annual Report of the Royal Inscriptions of Meso- potamia Project, 6 (1988), 5-7. 44 Examples in this instance are the version in which dedication to the queen was included, Borger, Eheglück, (it is omitted in the two main exemplars now extant) and the Bavian inscription which gives much greater detail about the engineering outside Nineveh. 454 WATER MANAGEMENT IN ASSYRIA

great pillars of copper, colossal striding lions, such as no previous king had ever constructed before me, with the technical skill that (the god) Ninshiku brought to perfection in me, and… I invented a technique for copper and did it skilfully. I created clay moulds as if by divine intelligence for great cylinders and alamittu- palms, tree of riches; 12 fierce lion-colossi together with 12 mighty bull-colossi which were perfect castings; 22 cow-colossi invested with joyous allure, plenti- fully endowed with sexual attraction; and I poured copper into them over and over again; I made the castings of them as perfectly as if they had only weighed half a shekel each. Two of the copper bull-colossi were then coated with electrum. I installed the alabaster bull-colossi alongside the white limestone bull- and cow- colossi at the bolts of my royal lodges. I bound tall pillars of copper alongside pillars of mighty cedar, the gift of the Amanus mountains, with bands of copper and tin, and stood them on the lion bases,45 and then positioned door-leaves to crown their gateways….. (a description of elaborate doors and thresholds follows)… In order to draw water up all day long I had ropes, bronze wires and bronze chains made, and instead of a shaduf I set up the great cylinders and alamittu-palms over cisterns. I made those royal lodges look just right. I raised the height of the sur- roundings of the palace to be a Wonder for all Peoples.46 I gave it the name: “In- comparable Palace”.47 A park imitating the Amanus mountains I laid out next to it, with all kinds of aromatic plants, orchard fruit trees, trees that sustain the mountains and Chaldaea, as well as trees that bear wool, planted within it.” It has long been held that these are the gardens shown on a broken stone relief sculpture from the palace of Sennacherib’s grandson, Assurbanipal, at Nineveh. Ever since they were uncovered, some of the details on that sculpture have been compared with the description of the Hanging Gardens by Greek writers: the aqueduct which brings water into the garden half way up the slope; the artificial terraces built up on stone vaults, which meant that trees could not reach the water-table, and must be watered entirely artificially. To this we can add the dedication to the Queen (only recently found), and the screw for watering, if our identification of gismahhu and alamittu in the prism inscription are correct.48 The water from the aqueduct would also have been led into the palace, from the garden raised up beside the building. Some of the bronze-castings were architectural items in the form of animals, probably used as column-bases. They weighed up to 43 tons, and the new methods allowed them to be more than twice as heavy as those of Senna-

45 Cast bronze lion-bases beneath pillars are probably shown on a sculpture of Assurbanipal. See J. Reade, Assyrian sculpture, (London 1983) pl. 56. 46 This presumably refers to the artificial stone-built hill of terraces created for the gardens, and to the open, pillared walkway with trees planted on its roof, described by some of the Greek sources. See S. Dalley, “Nineveh, Babylon and the Hanging Gardens”, Iraq 56 (1994), pp. 45- 58. 47 The sequence of sentences is particularly disjointed here, and results from the way in which royal inscriptions reused and recombined passages from other, more extensive inscriptions. 48 The confusion between Nineveh and Babylon is the subject of S. Dalley, “Why did Herodotus not mention the Hanging Gardens of Babylon?”, in Herodotus and his World: Essays in Memory of George Forrest, ed. P. Derow and R. Parker, (OUP) forthcoming. S. DALLEY 455 cherib’s father Sargon II, which weighed up to 17 tons or more.49 The items used to raise water are described as a “great tree trunk”, gismahhu, which is known to be the word used for a cylinder (e.g. in mathematical problem texts) and an alamittu-palm-tree. In fact, it is the moulds which are described as being these two, separate items, not the casting. The weight of one bronze screw cannot be calculated with any accuracy, since length, diameter and thickness of metal are all unknown – between two and three tons has been suggested. For comparison, one of the famous pair of cast bronze “barriers” of the 12th century BC, found at Susa and securely dated by the inscription on them of the Elamite king Silhak-Insusinak II (c.1150-1120 BC), has known meas- urements: 4.34m long, 0.18m diameter, thickness irregular, around 1.5cm.50 Since Strabo said the hanging gardens were watered with a screw, and since the cylinder is a component part of the screw, the hypothesis was put forward that alamittu was a word used in Akkadian for the spiral.51 In Mesopotamian architectural design facades were constructed in mud brick showing two kinds of palm-trunk decoration: the scalloped and the spiral;52 and some stone and terracotta sculptures show palms with a spiral-patterned trunk. An alternative explanation, that Sennacherib’s innovation describes an orna- mental façade, can be dismissed because of the vocabulary used, which includes none of the well-known terms for decoration.53 Another suggestion, that Sennacherib “rebuilt the pulling system of the wells in his palace”54 does not explain the terms used (the cylinder and the alamittu); and since no wells were found in either of the great palaces excavated at Nineveh, we may deduce that Sennacherib’s water-installations were at least as sophisticated as those of his father, as found by Place at Khorsabad.

DID ARCHIMEDES INVENT THE SCREW? Posidonius55 of Apamea in inland Syria, (c.135-50 BC) wrote that Archime- des “found” (heuren) screws for raising water, when he was in Egypt. A com-

49 S. Dalley, “Neo-Assyrian textual evidence for bronzeworking centres”, in Bronzeworking centres of Western Asia c. 1000-539 B.C. ed. J. Curtis, (London 1988), pp. 104-105. A rough calculation showed that a lion casting weighing 32 tons could be about 2m high and 3.5m long, roughly the same dimensions as the great stone statues found in the gateways of Assyrian palaces. 50 F.W. König, Die elamischen Königsinschriften, (Graz 1965), p. 17, no. 45; personal in- spection of the one on exhibition in the Louvre. 51 This is part of an argument that Sennacherib’s gardens at Nineveh are the original “Hang- ing Gardens of Babylon”. See Dalley, “Hanging Gardens”, pp. 45-58. 52 For example, at Larsa, Ur, Tell al Rimah, Tell Haddad, Tell Basmusian. 53 Unpublished suggestion of Michael Roaf. 54 M. van de Mieroop, The Ancient Mesopotamian City (Oxford 1997), p. 161. His contention that privies did not exist in ancient Mesopotamia is contradicted by the evidence found at Khorsabad by Place. A.R. George, BANEA conference paper delivered in Cambridge 1999, identified the Akkadian word asurru as lavatory, and adduced further material and textual evidence. 55 Quoted by Strabo III 2, 9. 456 WATER MANAGEMENT IN ASSYRIA mon translation for this context is “invented”, and it rests on analogy with the famous cry heureka with which Archimedes in his bath is said to have marked his sudden understanding of a new physical principle.56 Archimedes cannot be said to have “invented” displacement in this context: the principle as a law of physics was there before Archimedes realised how to understand and define it. Thus the verb heuren here means “to come across, discover” rather than “to devise first, to originate” (in the sense known to patent law).57 Therefore heuren may mean “found, discovered” in the passage of Posidonius, and that meaning may be preferable since Posidonius calls them “Egyptian screws”. Some scholars have interpreted this to mean that the screw existed in Egypt before Archimedes found it there.58 One might suppose that Archimedes, having discovered the screw in use in the Nile Delta, was inspired to write his treatise on spirals, his interest being mathematical and theoretical, as Plutarch maintained. A similar passage is quoted by Diodorus Siculus, writing in the first century BC, in describing the events and customs of semi-mythical times in the first six books of his world history. In I 34.259 in the context of the Nile Delta, he wrote of screws which Archimedes epenoese.60 This verb could imply that Archimedes invented them but it could also mean that he observed and thought about them,61 for example with a view to putting them to a new use, as he did for Hieron’s bilge-pumps, or to formulating mathematical principles involved in spirals, as he did for his treatise on spirals. Again in book V 37.3 Diodorus Siculus (probably quoting Posidonius of Apamea) mentioned Egyptian screws which Archimedes found or invented in Egypt. He gave an account of very deep shaft mines in Spain, such as are known archaeologically for the Roman period, although it is not known whether they go back earlier.62

56 J.P. Oleson, Greek and Roman Mechanical Water-lifting Devices (Toronto 1984), pp. 92- 93. 57 For the verb heurisko, the Greek Lexicon of Liddell and Scott gives the meanings “find, find out, discover, devise, invent, get, gain, fetch (a price), earn” 58 E.g. E.J. Dijksterhuis, Archimedes (Copenhagen 1956), 21-22; B. Gille, “Machines” in: A History of Technology II (Oxford 1956), ed. C. Singer, E.J. Holmyard, A.R. Hall and T.I. Williams, p. 631; B. Cotterell and J. Kamminga, Mechanics of pre-industrial technology, (CUP 1990), 94; D.L. Simms, “Archimedes the Engineer”, History of Technology 17 (1995), p. 58. Contrary arguments are collected by Oleson, Waterlifting Devices, pp. 291-92. 59 Probably quoting Agatharchides of Knidos, who was writing ahout 25 years after the death of Archimedes. 60 Meanings given in Liddell and Scott for epinoeo include “think on, think of, contrive, purpose, note, observe, invent”. 61 I am grateful to Prof. George Huxley for helping with this point. Gille, “Machines” p. 677 takes this view of Diodorus’ work, that he intentionally mentioned the screw as used for a time before Archimedes. 62 See e.g. B. Rothenberg and A. Blanco-Freijeiro, Studies in Ancient Mining and Metallurgy in SW Spain, (London 1981), p. 40; and the remarks of R.J. Hopper, “The Laurion Mines: a re- consideration”, Annual of the British School of Archaeology at Athens 63 (1968), pp. 293-326, pointing out that evolutionary theories about developments in mining are not based on proper evidence. S. DALLEY 457

“And what is the most surprising thing of all, they drain out the waters of the streams they meet by means of what are called Egyptian screws, which Archimedes the Syracusan found when he was going round in Egypt. And by the use of such devices they carry the water in successive lifts as far as the entrance… Since this machine is an exceptionally ingenious device, an enormous amount of water is thrown out, to one’s astonishment, with a trifling amount of labour… And one may well marvel at the inventiveness of the craftsman63 in connection not only with this invention but with many other, greater ones… we shall give a precise and detailed account when we come to the period of Archimedes”. This account was presumably in a later book and is not now extant. This passage in Diodorus could be understood to imply that an anonymous craftsman invented the screw before the time of Archimedes, during the semi-mythical age of world history. Alternatively, Archimedes may be referred to; although the word technites “craftsman” might be reckoned to be insufficiently elevated for such a great scholar, whom Plutarch names and calls demiourgos, a more complimentary term.64 Diodorus does not mention the means of pro- pulsion, nor the angle of installation, nor the size of the screws, but says the Phoenicians developed the mining industry in Spain. The reference to “successive lifts” is interesting because a similar series would have been required to raise water up an artificially terraced garden such as Sennacherib built. The screw was presumably known in wood before it was cast in bronze, for one would not experiment in such an expensive way in a material which could not be easily modified. But there is no evidence at present discovered which might show whether the Assyrians invented the device shortly before casting it in bronze, or whether they discovered it abroad during a campaign or as a result of a merchant’s travels. For raising water over the bank of a ditch into a field or garden for irrigating plants, the screw is often installed at a low angle, and can be rotated with the feet, as shown on Roman-period terracottas and mosaics, and as described by Philo of Alexandria.65 Presumably this is the practice to which Deuteronomy 11:10-11 also refers in saying that Egypt is a place “where you sowed your seed and watered it with your feet”,66 a text datable to the seventh century or to the Persian period, sixth to fourth centuries BC. This method is not suitable, however, for a steeper angle of installation such as would be needed in deep mines and in a steeply terraced garden, circumstances for which no illustrations are extant to show how rotation was achieved. Sennacherib’s inscription may imply that ropes and chains were used, perhaps by pulling them under

63 Oleson, Waterlifting Devices, p. 92, following e.g. the note in the Loeb translation, assumes Archimedes is meant, but the word for craftsman here, technites, is odd if applied to Archimedes. 64 Plutarch, Life of Marcellus XIV.9. 65 On the Confusion of Tongues, 38; translation by Oleson, in Greek and Roman Technology: a sourcebook (London 1998), by J.W. Humphrey, J.P. Oleson and A.N. Sherwood, p. 318. 66 A date before the time of Archimedes is almost certain for this bit of text, although its relevance to the screw for irrigation has been questioned by Oleson, in O. Wikander, Handbook of Ancient Water Technology, (Leiden 2000), p. 234. 458 WATER MANAGEMENT IN ASSYRIA tension from below the cylinder, but this is uncertain. The crank handle is not attested until rotary querns were invented in the second century BC, although one cannot preclude the possibility that it was invented earlier for a different machine. The passage in Deuteronomy suggests that the screw was character- istically used to water vegetable gardens, and this restricted use may account for the fact that Herodotus, describing irrigation of field crops in Egypt, men- tions only the shaduf. Three different texts of pre-Archimedean times may thus attest to the use of the water-raising screw: Sennacherib’s Palace Inscription, Deuteronomy 11:10, and Strabo’s description of the Hanging Gardens. The principles of making a bearing upon which rotation could take place would have been adapted from the door pivots and sockets, made for enormous, monumental doors, in which the sockets were often made of a hard stone coated with bronze. Doorkeepers were issued with an allocation of oil or lard, for lubricating the socket as well as locks and bolts.67 Strabo (XVI 1.5) said that the hanging gardens of Babylon were watered with screws set alongside stairs leading to the upper terraces of the garden, so presumably out in the open rather than concealed inside a building. Diodorus Siculus (II.10) said the water-raising devices worked without anyone outside seeing the water being raised. This description excludes the shaduf, the water- wheel, and the cerd,68 all of which are patently visible. Strabo knew that the gardens were earlier than Archimedes by many centuries, so by implication he seems to acknowledge that the screw was in use before the lifetime of Archimedes. One might claim that the screw was fitted retrospectively,69 but this argument suffers from the weakness that no such modification is mentioned in texts, and that the previous watering system would have worked perfectly well during the previous four centuries, until the modification was made. If it worked, why fix it? In any case, Sennacherib’s inscription states clearly that the new system was “instead of a shaduf” and included a cylinder. Since Oleson has suggested that the cerd might have been used to raise water for the gardens until a screw was fitted retrospectively,70 a few observations may be made in an attempt to evaluate its merit. The device requires a pole and an ox or two, with a flat piece of land sloping down, away from a well or cistern. It was used for a Persian garden at Susa in Hellenistic times. But it is not shown on Assyrian relief sculptures nor on dynastic Egyptian paintings, despite the fact that it would be a highly visible feature (unlike the screw). The new watering method described in Sennacherib’s palace inscription cannot refer to a cerd, because it does not have the parts – leather bags and animals – required, and because the cerd could not be constructed with a bronze cylinder.

67 A. Salonen, Die Türen des alten Mesopotamien, (Helsinki 1961), pp. 120-121. 68 For a description of this device, which requires a pole and an ox on a flat piece of land, see Oleson, in Wikander, Handbook, pp. 222-225. 69 Oleson, Waterlifting devices, p. 293 70 Oleson in Wikander, Handbook, p. 242. S. DALLEY 459

Nor do any of the Greek descriptions of the Hanging Gardens mention such a device. The terms used for the cerd in modern Iraqi Arabic, charad and bakra, do not correspond to words in Sennacherib’s Palace inscription, and, as Laessøe has concluded, there is no evidence at all for the use of the cerd by Babylonians and Assyrians.71 The suggestion, therefore, does not stand up to scrutiny. Above all, a garden watered in such a way would not have qualified as a World Wonder, nor could it have been converted into one simply by the retrospective introduction of the screw, at a time when screws were common.

WAS THE SCREW INTRODUCED INTO THE DELTA OF THE NILE BY THE ASSYRIANS? At first sight it might seem to be a valid objection that no clear evidence from Egypt is available for the use of the screw before the third century BC. This is less serious when one considers that genre scenes, of a kind likely to show watering devices, are seldom found after the New Kingdom period in Egypt, for they became unfashionable; there is a very long gap in time until terracotta figurines, mosaics and papyri reveal the existence of screws in the Ptolemaic period. The late Assyrian palace sculptures which show gardens at Nineveh are far too damaged to be used as evidence for absence,72 and they come from the very end of the period when scenes were shown in sculpture. We have no comparable scenes from the Neo-Babylonian period nor from the whole of the Achaemenid period. As far as contemporary written and pictorial sources are concerned, our knowledge of ancient Mesopotamia is much better informed for the late Assyrian period than for the subsequent three centuries. If the screw was invented or adopted by the Assyrians in the seventh century BC, it would be likely that they introduced it into Egypt, when they took control of the Nile valley in the mid seventh century, to replace the less effi- cient shaduf which had served the purpose for the previous millennia. Unfortu- nately the period of Assyrian occupation and its aftermath is a time to which very few archaeological discoveries bear witness, particularly in the Delta.

WAS HEZEKIAH’S TUNNEL + INSCRIPTION MADE IN IMITATION OF SENNACHERIB? Several commentators have remarked that Hezekiah’s tunnel in Jerusalem would have taken so long to dig that it could not have been made quickly enough when a siege was threatened. This means that we should consider the

71 J. Laessøe, “Reflexions on modern and ancient Oriental water works”, Journal of Cunei- form Studies 7 (1953), p. 26. 72 WA 124939, of Assurbanipal, is broken off both where the topmost part of the gardens should have been and for a very large part of the right-hand side. The Original Drawing IV 77, which Reade, “Assyrian Illustrations of Nineveh”, Iranica Antiqua 33 (1998), pp. 81-94, has identified as gardens at Nineveh, shows only small fragments of the whole scene, but includes the vital pillared walkway with trees on top, corresponding to Philo’s description of columns with beams set on top supporting soil and trees. Also reproduced by Reade, “Nineveh”, Real- lexicon der Assyriologie, Band 9 (2000), p. 404. 460 WATER MANAGEMENT IN ASSYRIA possibility that it was a peace-time achievement. It was the blocking of the Gihon spring that constituted an act of defiance against Sennacherib, and the biblical text clearly separates that act from the act of digging the tunnel. The tunnel brought the water out at the new pool of Siloam, and it was only a slight improvement on the old channel which brought the water out at the old Siloam pool, a little lower down the slope of the hill. David Ussishkin has suggested that its purpose was to water the royal garden, and by bringing the water out higher up, a considerably larger area could be watered by gravity, through irrigation channels.73 Tsvi Tsuk during discussion at this conference mentioned the likelihood that a water-tunnel at Amman dates to the same period, and so could also have imitated the waterworks at Nineveh.74 The new evidence from royal tombs at Nimrud has shown that Hezekiah was closely related to Sennacherib’s mother Atalya.75 This relationship suggests the possibility that Hezekiah was imitating Sennacherib’s great engineering work at Nineveh in making his tunnel and garden. What is left of his inscription, describing how the workmen tunnelling from each end heard each other and broke through to join up the two sections, belongs to the genre fash- ionable in the time of Sennacherib, in which the labours of the workforce could be reported within a royal inscription, as given in the translation above. In conclusion, there is good evidence both textual and archaeological for re- markable developments in water management in Assyria over a period of more than two centuries, culminating in works which are still to be seen. They reveal a heroic level of technical achievement which deserves to be better known among historians of technology. Whether the Assyrians promoted the use of the true qanat, they certainly had all the elements needed to do so, and used a very similar system to bring water underground into their cities. As for the water-raising screw, they knew it well enough to cast one in bronze, for watering a terraced garden at Nineveh, recording the achievement in cuneiform, and verifying the account of the Hanging Gardens given by Strabo.

POSTSCRIPT After this paper went to press, Assyrische Wasserbauten by Ariel M. Bagg was pub- lished by Philipp von Zabern, Mainz am Rhein 2000. It is reviewed in detail by the present writer in Archiv für Orientforschung, forthcoming 2001. More detail about the irrigation of Sennacherib's palace garden at Nineveh is given by the writer in “More about the Hanging Gardens", in a forthcoming volume in hon- our of David Oates, ed. A. McMahon (Cambridge).

73 D. Ussishkin, “The Water Systems of Jerusalem during Hezekiah's Reign" in Meilenstein, Festgabe für Herbert Donner, ed. M. Görg (Wiesbaden 1995), 289-307. 74 See M. Burdajewicz, in New Encylopaedia for Archaeological Excavations in Eretz-Israel, ed. E. Stern, vol. 4, s.v. Rabbath-Ammon, p. 1247. 75 This issue is discussed in more detail by S. Dalley, “Yabâ, Atalya and the Foreign Policy of Late Assyrian kings”, State Archives of Assyria Bulletin 1998, (Helsinki), in press.