Skip to: Site Menu | Main content

Three Notes on Qumran

Herod had extensive building programs at Masada, Machaerus, and Hyrcania and probably used Qumran as a convenient facility for the collection and distribution of building materials and food stores suitable for his royal palaces. He thus needed to make it possible for a limited number of quartermasters to live permanently at Qumran by increasing the available water, an essential requisite for even the most limited of year-round occupation.

By David Stacey
Field Archaeologist (1975-1987), Jericho Excavations

March 2009

The Dam

While much has been written about the plastered pools and the water system within the settlement at Qumran, far less thought has been given to the dam that was built across the mouth of a basin in the Nahal Qumran to capture occasional flood water and steer it into an aqueduct that would carry it to those plastered pools. No trace of the dam has survived the torrential floods of the past 2000 years but its existence is certain because the elevation of the aqueduct is some 4 or 5 m above the bedrock of the basin.1 A description of the basin and a plan of it showing the approximate location of the dam was published by Ilan and Amit (Ilan and Amit 2002: 381-382, Fig. 1). The dam must have been “substantial” (Ilan and Amit 2002: 382, fn 6). As it must have stood four or five meters high, the walls at their base were probably at least two or three meters thick2 and based solidly on, or more likely, cut into the bedrock. Such a dam was unusual and would most certainly have “demanded both extensive hydrological knowledge and a large budget” (Hirschfeld 2004: 115). It would also have required frequent and expensive maintenance.

Although the construction of aqueducts burgeoned in the Second Temple period most were constructed to supply either Royal palaces or large urban settlements themselves under Royal or Temple patronage and served either to capture run-off rain water falling in their immediate vicinity, or to transport water from distant springs. Dams are rare but there are two, both usually dated to the time of Herod, which were integral parts of the water systems of two of the most prestigious towns, Jerusalem and Caesarea (Patrich and Amit 2002). Neither would have had to withstand the forces to which that at Qumran would have been subjected by the flash floods whose destructive power was all concentrated into a narrow canyon. How these forces were overcome we do not know but the dam had to have been an exceptional construction technologically.

A short aqueduct, some of it tunnelled through the cliff-face of the gorge, took water that overflowed the dam (Ilan and Amit Fig. 1; 3, 6-16) into an earlier run-off channel that gathered the occasional rain that fell on the slopes north west of the settlement (ibid 16-19, Stacey 2007: 228) and which had served to fill L110, L117 and L118 in the Hasmonean period. Rain-fall at Qumran is very infrequent but can be briefly heavy. When it occurs rain often also falls in the Buqe’ah causing floods in Nahal Qumran, at the same time, or soon after, the Qumran rains. However rain is more frequent in the Buqe’ah than in Qumran and there can be unexpected floods when there has been no sign of precipitation in Qumran. The Hasmoneans clearly constructed the three pools, L110, L117 and L118, so that they would usually be filled by the total water collected by the run-off section of the aqueduct. To increase the volume of water that could be gathered meant capturing some of the flood water. The construction of the dam and of the rock-cut channel required the sort of labour and financial resources, technical expertise and determination demonstrated by Herod in other of his building projects and should be associated with his expansion of the settlement in 31 BCE (Stacey 2007)i. Herod had extensive building programmes at Masada, Machaerus and Hyrcania and probably used Qumran as a convenient facility for the collection and distribution of building materials and food stores suitable for his royal palaces (Stacey 2007: 239; 2008: 22). He thus needed to make it possible for a limited number of quartermasters to live permanently at Qumran by increasing the available water, an essential requisite for even the most limited of year-round occupation.

The dam was located approximately 25 m from the back wall of the “waterfall” and the “pond” which it would have retained was roughly circular but with several extensive irregular fissures which would also have retained water. Although it is impossible to know the effect of two millennia of floods, one can deduce that the capacity of the ponded water was c. 2,500 m³ (25 x 25 x 4 m) which is twice the total capacity of all the plastered pools that existed in Qumran after the additions dateable to Herod.3 Water would only have overflowed into the aqueduct on the two or three days in the year when floods raged. However it would have been simple enough to lift water from the pond into the channel by means of a shaduf, a beam with a counter weight,4 and it seems extremely unlikely that, having exerted so much effort and expertise into building the dam, the water gathered by it would have been allowed to go to waste.

That it was not is indicated by the ‘overflow’ channel from the pool L117. This does not appear to be merely for draining unwanted, excess water at the height of a flood. Such an overflow would more easily be conducted directly into the wadi through the short channel that exited the cistern L110 i (Magen and Peleg 2007: Fig. 4). The channel from L117 relates to the late stage of the pool after its sides had been raised, together with that of the round cistern L110 (de Vaux 1973: 9), when the ‘main’ aqueduct was built (Stacey 2007). The channel was well built and covered with capstones and post-dated an earlier refuse dump (Magen and Peleg 2007: Fig. 11). The material culture recovered from this dump is said to date from the ‘first half of the first century BCE’ (Magen and Peleg 2007: 8). It is to be hoped that it will soon be published as it could help confirm my dating of the ‘main’ aqueduct to no earlier than 31 BCE (Stacey 2007). The channel gives every impression that it was built, not as an occasional drain operating only for a day or two each year, but to transport water to an unknown destination with some regularity. Elsewhere I have suggested that this and other channels (from L60, L71 and L91) may have been utilised to irrigate small garden areas where ‘an enterprising gardener could have encouraged limited crops of quick-growing plants such as herbs, onions, pulses, gourds or mallow with which to vary the diet, or of woad for dyeing’ (Stacey 2008: 17). It seems likely that this ‘dump’ was, at least in part, deliberately filled to provide passage for the ‘drain’ and to create an irrigated garden. At some point the sub-surface channel from L117 would have had, either to flow into a settling pool from which distribution channels could lead to different terraces or, perhaps more likely, merely come to the surface and irrigate lower lying gardens via sluices within the channel. Although theoretically it would have been possible to feed these channels from pools in the built-up area it makes far more sense that, so long as water survived in the dam pond, it would have been used to replenish all, or most, of the pools, covered to reduce evaporation, and allow extra water to flow in these channels to irrigate much needed seasonal gardens.

Although for many Jews the use of a shaduf to replenish pools would not necessarily negate the use of those pools as mikva’ot, providing 40 se’ah (c. 750 litres) of ritually pure water remained in them, it would appear that for the authors of the Damascus Document, any water that had been contained in a vessel (which would include a shaduf) was not acceptable for purification (CD 10.11-14).5 The regular drawing of water from pools for industrial or other purposes and its replacement by the use of a shaduf would probably have made them increasingly unacceptable for purification purposes for any but the least observant. The pool L138, whose method of connection to the water system is uncertain (Stacey 2007: 230), could probably have been easily isolated from it to prevent the entry of “impure” water and is thus the only pool likely to have been a mikva at all times.6


There are many references in the writings of Josephus to the slaves, even eunuchs, of the Hasmonean kings and of Herod.7 Many who are actually mentioned were in positions of some importance within the royal households (inter alia, Ant. 15. 226: 16. 230-3; 17. 55-6; Bell 2. 57). The slaves were presumably of Jewish origin, or were gentiles who had undergone circumcision and ritual immersion so that there was no danger that by their very touch they would have rendered food or wine impure. Although “Roman agricultural writers advised estate owners to employ day laborers rather than slaves for the most difficult and excruciating tasks to avoid damaging their slaves’ bodies’”(Hezser 2005: 249) such considerations would scarcely have concerned the Hasmonean Kings or Herod. The rapid expansion of the royal estate in Jericho “could only have been achieved with a large influx of laborers” (Stacey 2007: 238) and some of these may well have been slaves who “were available throughout the year” and “could be organized in work gangs which were easy to supervise” (Hezser 2005: 250). It should be noted that following Herod’s death in Jericho, his funeral procession was led by troops followed by “five hundred of Herod’s slaves and freemen, carrying spices” (Bell. 1. 673), many of whom must have been employed on the Jericho estate. However, since the day-to-day life of slaves was of no more concern to Josephus than were the lives and conditions of domestic animals, we have little idea as to how many slaves were owned by the royal estate, or to what work they all were put. Elsewhere I have suggested that Qumran was a seasonal industrial suburb of Jericho and that most of the work carried out there was of an unpleasant, smelly, and even ritually impure nature (Stacey 2008). As there was no nearby settlement from which day laborers could be drawn, slaves may have been drafted in from Jericho. It seems probable that not all slaves of non-Jewish origins (inter alia, Ant. 13. 319, 397; Bell. 1. 376), especially those who had been captured in some numbers during Herod’s military campaigns, would have been subjected to circumcision, but would have been impressed to undertake hazardous jobs (such as building the northern palace at Masada, or the dam at Qumran) or those activities, such as the preparation of skins, which would have been ritually polluting for Jews.

We know of one of Herod’s slaves by name. Following Herod’s death, Simon, a tall, handsome slave, who it seems had served Herod in Jericho where, because, he was “far superior to others of his order” … “had had great things committed to his care,” burned down and looted the royal palace in Jericho, set a diadem on his head, and attempted to set himself up as king. To help finance his ambition, he and his followers attempted to plunder other of the royal palaces, including, possibly, that in Jerusalem, but eventually retreated to Perea,8 apparently his place of origin, where he was pursued by Roman soldiers who cut off his head (Ant. 17. 273-277; Bell. 2. 57; and Tacitus Histories 5. 9). The details are sketchy but the following scenario is not impossible. Simon may have been familiar with Qumran, perhaps even had certain aspects of its daily life “committed to his care.” Part of the loot from the palace in Jericho included silver coins, which he chose to bury for safe keeping in Qumran while he went off looking for more treasure. His unfortunate demise meant that he never returned to retrieve his loot, represented by the two pots of Tyrian silver coins found in L 120. The unlikelihood that these represented a collection for payment of the temple tax as has been suggested (Magness 2002:191-192) is pointed out by the Lonnqvists. They note that in the limited amount of the hoard that they have been able to study at least four of the coins came from the same die indicating that the coins represent a currency hoard, possibly belonging to “a single wealthy family” (Lonnqvist 2006: 139-140). That wealthy family may have been that of Herod himself.

We know little about the daily life of slaves and even less about what happened to them once they were dead. It has been claimed that the cemetery at Qumran is exclusively male and therefore proof that the site was occupied by a sect of celibate Essenes. Even if the cemetery was indeed predominantly male (and too few graves have been excavated to be certain of this), it is possible that some of the peculiarities of the graves (when compared to rock cut tombs known elsewhere which clearly belonged to wealthy families) are because they belonged at least in part to slaves. If the five hundred slaves that were assembled in Jericho for Herod’s funeral procession is any indication of the size of the slave population of the Jericho estate, there would have been a considerable number of slaves requiring burial over the years. Although a few of the most favored slaves may have been buried in their master’s family tomb many would not and certainly had no family tomb in which to be interred. A small area of poor Hellenistic burials, dug simply into the soil, was excavated near to Herod’s hippodrome, Tell es-Samarat (Stacey 2004: 226). They were covered by a mound of soil that was the spoil from leveling operations for the nearby hippodrome. If these were burials of the poor and of slaves, then the encroachment of buildings and the introduction of irrigation agriculture on to previously uncultivated land would seriously have restricted the area available to them, and it is possible that the bodies of some slaves were transported to Qumran for burial instead. Slaves were not allowed to marry Israelite women (Hezser 2005: 197), and it may be that for some male slaves, particularly non-Jewish prisoners of war, the prospect of finding an acceptable partner, even if they were encouraged to do so, would be unlikely. Thus slaves buried in Qumran would have been predominantly men. And if they had the misfortune to be eunuchs, they would have been exclusively so!

Many shaft graves similar to those in Qumran have been found in a large cemetery at Qazone in Nabataea. They are dated a few centuries later and definitely contain burials of men, women, and children. Could it be that the type of grave originated in Nabataea and that those in Qumran belonged primarily to slaves who had been captured during the various conflicts between the Nabataeans and both the Hasmoneans and Herod?

Basketry and Coracles

I have suggested that Qumran was primarily a seasonal industrial site and that some of the industrial activities, particularly the curing of skins, the scouring of wool, and the retting of flax were carried out in readily made, and ultimately disposable, containers made of local reeds and/or palm fronds made watertight with skins or, more likely, bitumen collected off the Dead Sea (Stacey 2008: 15, 17). Several different processes were undertaken in which raw materials were soaked in various liquids which required regular dilution and/or a complete change of liquid over several stages. There are few permanent pools at Qumran suitable for these activities, but it would have been more convenient, particularly where, as at Qumran, there was no running water, to have many portable containers which could be placed near to available water. Skins were processed initially by being soaked in excrement, which, together with the necessary contact with dead animals, would have made the activity particularly polluting. Thus it would have been advantageous to have “disposable” containers. I also suggest that both wool and flax were processed at Qumran, and, because it was forbidden for Jews to mix wool and linen in the same garment (Deuteronomy 22:11), it was simpler to process them in entirely separate containers.

We shall probably never know if remains of such containers were found at the site. The Donceels mention objects of “vegetable substances, such as palm-leaf-work or wicker work, of which most have no inventory number; some were photographed, few were described and many have disappeared, as in the case of.... large mats found in some rooms” (Donceel 1994: 14). The simplest and most obvious container would have been a spherical basket similar to the Mesopotamian “quffah9 or coracle, depictions of which appear on Assyrian reliefs at Nineveh and Nimrud10; a description of which was written by Herodotus (Book 1: 194); and which survived well into the 20th century. The raw materials, or close equivalents, utilized in Mesopotamia would have been freely available near the Dead Sea.

In construction, a quffah is just a huge lidless basket, strengthened within by innumerable ribs radiating from around the centre of the floor” (Hornell 1938: 153). In India, where similar shaped coracles are still made, the builders use split bamboo with which to build a strong interwoven framework which can then be covered, originally with hides, although today with plastic fertiliser sacks (McGrail et al 2003). Bamboo was not available in Mesopotamia (or by the Dead Sea), nor were there any other plants growing nearby from which strong ribbing could be made. Consequently, the first part of the Mesopotamian guffah to be built was a coiled “mat” made of “a continuous and flattened spiral, formed of a stout bundle of parallel lengths of some fibrous material – straw, reeds or palm leaflets – bound by a parceling or whipping into a rope-like cylinder. By concentric coiling of this ‘filled rope,’ the shape required is gradually built up.... The gunwale consists of a bundle of numerous withies, usually of willow, forming a stout cylindrical hoop bound round and attached to the uppermost and last-formed coil by closely set series of coir lashings. The inner framework giving strength and rigidity to the coiled walls of the quffah is formed of a multitude of curved ribs, closely set; usually split branches of willow, poplar, tamarisk, juniper or pomegranate are employed; when these are not available the midribs of date-palm leaves are used, but these are less esteemed.... As each of these ribs and frames is placed in position, it is sewn with coir cord to the basketry walls” (Hornell 1938:153-155 Pls III & IV).11 The outside of the finished “basket” would then be waterproofed with asphalt. They were generally about 2 or 3 m in diameter but some as large as 5 m were reported.12 Very similar baskets, except waterproofed internally, would have been ideal containers for the various industrial processes at Qumran. In India, a coracle can be built, once all the raw materials are to hand, by a team of three in ten hours (McGrail 2003: 246). Due to the inferiority of the materials available in Mesopotamia, which meant that a mat had to be made before a framework could be added, the construction must have required more man-hours overall.

There is no doubt that basket and mat making date back to very early times in Palestine, nor that quffat flourished in Mesopotamia at the time that the Assyrians invaded Palestine. As very similar raw materials used for making quffat on the Tigris-Euphrates were freely available on the shores of Lake Tiberias and the Dead Sea and there was a similar scarcity of timber for planked boats, it seems extremely likely that similar craft were made there too. One could go further and say that it is extremely unlikely that such easily and cheaply constructed craft that were stable (in experienced hands) and able to take quite heavy loads, did not exist on these inland seas. Although I know of no historical or ethnographic evidence for their existence, both Strabo and Diodorus (possibly using the same source) mention that asphalt was collected on the Dead Sea using reed boats which appear to be simply bundles of reeds tied together (similar vessels existed together with quffat in Mesopotamia and were also known in Egypt).

Diodorus, however, describes a sea battle on the Dead Sea in which a Nabataean force of 6000 men defeated the naval force of Antigonus by killing many of them with arrows fired from their reed boats (Diodorus Book 19: 98-100). Since each boat was said to have two rowers and an archer, we can deduce that the Nabataean “navy” consisted of 2000 boats. Even if we discount this number as exaggerated, the implication is that many boats were available to the Nabataeans, and where wood was in very short supply, we can suppose that reeds were used. Quffat would have been more stable and more readily maneuverable than crude rafts, so it may be that part of the Nabataean’s success was due to their having superior vessels with which they were familiar from regular use on the Dead Sea.

Herodotus says that quffat were made in Armenia and used to float goods down to Baghdad where the hide was stripped off the framework, which was sold off, and taken back to Armenia for re-use, “for it is not by any means possible to go up stream by water” (Herodotus I.194). Similarly in 19th century, Tamil Nadu hide-boats, each carrying 4 tons of rice, were floated down river in time of flood. Once at their destination, the “bamboo framing was sold off and the hide retained for further use” (McGrail 2003: 244). Quffat could have been made near Lake Tiberias (waterproofed with asphalt brought up from the Dead Sea) and used to carry agricultural produce from the Galilee down the Jordan when in flood. The craft would then remain on the Dead Sea where they could be used, particularly in the Second Temple period, for carrying building and food supplies in short “hops” along the shore, inter alia for transporting balsam from Ein Gedi to Qumran for onward movement to the royal estate in Jericho for final processing and sale. They would have been smaller, though more numerous than, wooden vessels that undoubtedly did exist on the Dead Sea over several centuries (Hadas 2008, Nissenbaum 1993, Oren 2007/8) although “until now anchors have been the only archaeological evidence of sailing on the lake” (Hadas 2008: 31). A ship-dock has been identified at Qasr el-Yahud and a possible jetty at Rujm el-Bahr but no docking facilities have been discerned at sites where they would be expected, such as Ein Gedi and Masada (or Qumran). Quffat would have been ideal for lightering from larger vessels anchored offshore. The great advantage of these buoyant vessels is that they would have been cheap, readily available and required only the most rudimentary of landing stages, if any at all. A clearance of rocks may have been all that was necessary. It is to be hoped that with the rapid shrinkage in the shore line of the Dead Sea sharp eyes will look out, not only for anchors, but also for surviving relics of more humble craft not built of wood.

1 This could be seen on a visit to the site in June 2007 soon after a particularly strong flood which had cleaned out the basin to bedrock making it no longer possible for a person, like myself, no longer in the first flush of youth, to climb up from the basin into the channel of the aqueduct. Less aggressive torrents can partially fill the basin with debris making access to the aqueduct from the basin simple (as it was on a previous visit in May 2005).

2 There are two long dams associated with the low level aqueduct at Caesarea, both dated, probably, to the Byzantine period. One is 3.5 m high and has a width of between 2.1 and 2.5 m, the other is 7 m high and is 5.5 m wide at its base and 4.5 m wide at its crown (Peleg 2002:141-148).

3 Although a total capacity of 11,000 m³ is claimed (Ilan and Amit 2002: 381), this would appear to be a typographical error with the generally accepted figure being around 1,100 m³ (Hirschfeld 2004: 111, Galor 2003: 293).

5 For a discussion on the halakhah of Mikva’ot, see Magness 2002: 134-162.

6 cf, although with different reasoning, “We believe that the only pool that may have served as a ritual bath is pool L138” (Magen and Peleg 2007: 37).

7 All of the references can be found discussed in a recent comprehensive book by Catherine Hezser, Jewish Slavery in Antiquity (2005) OUP.

8 Peraea was a previously gentile area gradually taken over during Hasmonean and Herodian times. Although much of the population had been converted or replaced by Jews, resentment may have lingered among those forcibly converted.

9 Arabic for basket. Plural quffat.

10 Both on display in the British Museum.

11 For a photograph of Baghdad quffat c. 1935, see Hornell 1970 Pl. IVB. Note the weight of grain they are carrying.

12 Coracles as large as 6 m in diameter are recorded in 17th cent India (McGrail 2003: 243), but they had the advantage of having split bamboo to make a more solid framework.

i A pipe that took excess water from L138 towards the wadi would also have helped relieve pressure although the comparatively narrow bore of the pipe would not have been ideal to withstand the force of the flood in full spate.


De Vaux, R., (1973). Archaeology and the Dead Sea Scrolls (Oxford).

Donceel, R., and Donceel-Voute, P., (1994). “The Archaeology of Khirbet Qumran.” Pp.1–38 in Wise,

Golb, N., Collins, J.J., and Pardee, D. (eds.), Methods of Investigation of the Dead Sea Scrolls and the Khirbet Qumran Site (New York).

Galor, K., (2003). “Plastered Pools. A New Perspective.” Pp. 291-320 in J. B. Humbert and J. Gunneweg (eds) Khirbet Qumran et ‘Ain Feshka. II Studies of Anthropology, Physics and Chemistry. Fribourg.

Hadas, G., (2008). “Dead Sea Sailing Routes during the Herodian Period” BA-I AS: Vol 26: 31-36.

Hezser, C., (2005). Jewish Slavery in Antiquity OUP.

Hirschfeld, Y., (2004). Qumran in Context (Peabody, Massachusetts).

Hornell, J., (1938). “The Coracles of the Tigris and Euphrates” Pp. 153-159 in Hornell J., British Coracles and Irish Curraghs; with a Note on the Quffah of Iraq. London.

Hornell, J., (1970). Water Transport: Origins and Early Evolution. London (orig. pub. 1946)

Ilan Z., and Amit D., (2002). “The Aqueduct of Qumran,” Pp. 380-386 in D. Amit, J. Patrich, and Y. Hirschfeld (eds.), The Aqueducts of Israel. Journal of Roman Archaeology Supplement 46; Portsmouth, RI:

Lonnqvist, K., and M., (2006). The Numismatic Chronology of Qumran. Fact and Fiction.The Numismatic Chronicle Vol 166: 121-165.

Magen, Y., and Peleg Y., (2007). The Qumran Excavations 1993–2004, Judea and Samaria

Publications 6, (Jerusalem) (also at

Magness, J., (2002). The Archaeology of Qumran and the Dead Sea Scrolls (Grand Rapids).

McGrail, S., Blue, L., and Palmer, C., (2003). “Hide Boats of the River Kaveri, Tamil Nadu” Pp. 241-257 in S. McGrail, L. Blue, E. Kentley, and C. Palmer, Boats of South Asia. London.

Nissenbaum, A., “The Dead Sea – an economic resource for 10,000 years.” Hydrobiologia 267: 127-141.

Oren, A., (2007/8) Marine Trade on the Dead Sea in the 12th Century, accessible electronically (18/2/2009) at

Peleg, Y., (2002). “The dams of Caesarea’s Low-level Aqueduct,” Pp 141-148 in D. Amit, J. Patrich, and Y. Hirschfeld (eds.), The Aqueducts of Israel. Journal of Roman Archaeology Supplement 46; Portsmouth, RI.

Patrich, J., and Amit, D., (2002). “The Aqueducts of Israel: an introduction,” Pp. 9-20 in D. Amit, J. Patrich J., and Hirschfeld Y. (eds.), The Aqueducts of Israel. Journal of Roman Archaeology Supplement 46; Portsmouth, RI.

Porath, Y., (2002). “Hydraulic Plaster in aqueducts as a chronological indicator.” Pp. 25-36 in D. Amit, Patrich J., and Hirschfeld Y. (eds.), The Aqueducts of Israel. Journal of Roman Archaeology Supplement 46; Portsmouth, RI.

Stacey, D., (2004). “Tombs in the Vicinity of the Hippodrome at Jericho.” Pp. 226–228 in E. Netzer, Hasmonean and Herodian Palaces at Jericho, Vol. II (Jerusalem).

Stacey, D., (2007). “Some archaeological observations on the Aqueducts of Qumran,” DSD

14: 222–243.

Stacey, D., (2008). “Seasonal Industries at Qumran,” B A-IAS Vol 26: 7-29.