THE PLACE OF ANCIENT AGRICULTURAL PRACTICES AND TECHNIQUES IN YEMEN TODAY:

PROBLEMS AND PERSPECTIVES

Sanaa, Yemen

The prospecting carried out throughout this campaign (November17th to December 4th)* had the aim of affirming whether, in effect,the modulation in size and technical sequence detected in earlierhydraulic prospecting and analyses in the zone of Zafar (the 1998campaign) were equally discernible in the valleys adjoining thearea.

The systems that had been studied earlier (see figure 1) hadturned out to be of a size that was noticeably smaller than previousarchaeological studies had supposed. (1) It rapidly became obviousthat we were, in fact, dealing with hydraulic systems that could beobserved in detail with the same resolution that we had used withknown al-Andalus systems. Their compactness, technical articulationand size, modulated around a 10 ha area, clearly indicated that thesayl water - the torrential stream that follows the monsoon rains -was concentrated and made to run in a circular fashion, as with anyother continuous-flow water. Consequently, procedures for the studyof sayl systems could be the same as those used for the study ofcontinuous-flow water. At the same time, the prospecting of Zafarwould allow for reasonable predictions to be made with reference tothe ordering and functioning of the size of other systems in thearea. It was also clear that a consistent widening of the sample ofcase studies would not only reveal the technical congruency of thisstable, agricultural construction, but would also highlight theregularity of population sizes for the peasant construction groups,as expressed by surface area.

The list of asdad (singular: sadd) drawn up byal-Hamdani (in the mid IV H/ X AD), in conjunction with the currentcommentaries of Muhammad b. Ali al-Akwa , his editor and publisher,have established the orientation of the prospecting. (2) The asdadto which al-Hamdani refers are, in his opinion, already very old,certainly pre-Islamic. Both the architecture and the samples ofceramic ware collected from the surface of the prospect area (Halidal-Ansi, 1999) indicate a Himyaritic chronology.

As a result, there was the opportunity to study the production ofextremely old agricultural areas and the technical reasons for theirstability and permanence, none of which issues had, to our knowledge,been analysed with clarity to date; it was simply assumed that Islamsignified the start of a change that archaeologically had beenneither identified nor measured.

The valley of Bayt al-Aswal lies 2 Km from Zafar. The whole valleyis organised as a hydraulic complex of converging sequences. Eachsequence originated in a sadd. There are 5 of these asdad and theyhad been inventoried by al-Akwa in his commentaries onal-Hamdani.

The work consisted of prospecting the land in detail on the basisof aerial photography (1:10,000), and of undertaking sectional andelevational planimetric representations of the water containmentwalls (see figure 3), of some of the terraces, of the irrigationditches (saqiya-s) of the maqlah-s and other elementsof irrigation. Detailed planimetry was also undertaken of all thehydraulic systems, plot by plot, following the entire course of thewater through the systems, ditch by ditch and watergate by watergate.This work is carried out on the basis of aerial photography and bytracing the water circuit, since this is not directly visible in thephotographs. In addition, the names of all the girab were taken andnumbered, and will subsequently be analysed philologically.

The results obtained are the following:

• Sadd Gubayr (N 14° 11. 850; E 44° 25,300) Water from this sadd irrigates 17 girab, with a total extension of 7.3 ha.

• Sadd al-Maliki (N 14° 11. 690; E 44° 25,405) 14 girab, 3.46 ha.

• Sadd al-Nuqaq (N 14° 11. 965; E 44° 24,889) 6 girab, 2.8 ha.

• Sadd al-Amra (N 14° 11. 646; E 44° 25,001) 9 girab, 1.71 ha.

• Sadd al-Umran (N 14° 11. 656; E 44° 25,341) 33 girab, 7.56 ha.

In the map obtained (figure 4) it can clearly be seen that theorientation of the valley's irrigation is based upon two principalsystems, both being greater than 7 ha each: sadd Gubayr andsadd al-Umran. The other systems (al-Nuqaq, al-Maliki,al-Amra) are evidently complementary with respect to the irrigationsystem as a whole, yet they are fundamental in ensuring thecompactness of the overall system and in reducing to a minimum theerosive effects of uncontrolled torrential rain.

In fact, the observation of this compactness is fundamental inunderstanding the sequence of the design and the construction of thesequence. We observe in figure 5 how the al-Gubayr sadd systemand its irrigation space constitute the central part both because oftheir total extension and also because of the singular extension oftheir terraces. We can also perceive the decisive function of thearchitecture in underpinning all the valley's linked systems. On theother hand, it is less obvious to perceive that this centralitycannot exist without the sadd al-Nuqaq- (Nuqaq means'toad'), which diverts, collects and administers one of the strongestsayl water channels in the valley. Al-Nuqaq protects the al-Gubayrterracing. Strictly speaking, it is al-Nuqaq that provides a stableexistence for al-Gubayr. To limit ourselves to the observation ofal-Nuqaq's size (6 girab, 2.8 ha) or to its apparentlyperipheral area would be a grave distortion of reality.

In such a way, the observation of the compactness and conditionedstability of al-Gubayr allows us to appreciate that the juxtapositionof these areas cannot have been produced in long chronologicalintervals. The stable al-Gubayr area requires the contention of thecurrent channels that is carried out by al-Nuqaq. This suggests thatthe hydraulic organisation within the valley as a whole was conceivedas an ensemble, and that its realisation was successive butintegrated. We should therefore abandon the commodity of postulatingevolutional causality. For example, it is not empirically consistentto propose that the extension of hydraulic areas in the valley wasthe result of gradual demographic pressure. A better explanation isthat of considering the ensemble as a whole, congruent with the sizeof the initial population. This is not to suggest that all areascreated within the valley based on asadd have the same degree ofnecessity with respect to one another. And in fact the 14girab (3.46 ha) of al-Maliki o the 9 girab (1.71 ha) ofal-Amra are not necessary to the existence of Gubayr or al-Nuqaq.Nor, on the other hand, is it at all certain that the 33 girab(7.56 ha) of al-Umran, as we shall see later in this discussion, areessential to the design, construction and maintenance of al-Gubayrand al-Nuqaq.

This kind of particularly interwoven articulation was foreseeablefrom the studies of the hydraulic areas of al-Andalus. The noveltylies in the fact that the organisation of the complementarity andcompactness of the spaces was undertaken not to control &endash;concentrate and distribute &endash; continuous water currents, butinstead was for temporary channels for the monsoon. It isspecifically the fact of the monsoon that introduces variations ofrange in hydraulic circuits, variations that had not been detected inthe modest continuous-flow waters of al-Andalus.

In the eastern sector of figure 4, bordering the saddGubayr and the first terraces of the Gubayr and Maliki, there are20 girab the irrigation of which depends on a sadd, Harara,located at a point 2,760 m in the mountains that overlook the valley,which is the true hydrographic nucleus of the basin. The Hararagirab are clearly additional to the valley's hydraulicorganisational ensemble. Part of the descent water that suppliesthe girab is collected by the sadd Gubayr itself. Thesuperposed nature of the water collection undertaken by the enormoussadd Harara was evident from the first moment. The sadd had ananomalous appearance, given the unnecessary connection with thevalley's hydraulic sequences, the logic of which was clearlydiscernible. And the monumental nature of this sadd is reinforced bythe architectural remains of a small, paved walled site, clearlyvisible and associated with the sadd, at about 100m distant. It wouldseem that the sadd Harara was known to the writers of aHimyaritic inscription of the IV H/ X AD, whose context wasundetermined, and the geographical origin of which was also unknown.Its final identification, which we were able to carry out, hasallowed us to understand the meaning of the 'palatial' inscription,related to the construction of the sadd, in Harara.

Consequently, the sadd Harara, so clearly superposed above thevalley's hydraulic ensemble - in agreement with epigraphical evidence- dates at least from the IV H/ X AD. Therefore, the peasants'regulation of the valley is at least as old as this.

Current peasant memory attributes malignity to the waters thatfall from the sadd Harara, situated at such a height.According to the peasants now living in the area, this water would betoo cold and would damage the sorghum harvest. They also claim thatthe reflection of the waters in the sadd, so high above, was of suchstrength that it dazzled the peasants in the valley below, stoppingthem from working. Whether this is true or not, the opinion -manifestly very old - reveals the perception of this sadd assomething external, and that is certainly true. Perhaps we ought tothink along the lines of sadd Harara, although clearlyexternal, representing an interference in the hydraulicadministration of the valley, since it allowed for the control of thewater supply to two of the sadd from afar. One of these, saddGubayr, was the most important in the entire valley sequence. Ithas also been observed that sadd Harara was able to intervene indetermining the hydraulic circuits in neighbouring valleys. Far moredetailed fieldwork will have to be undertaken in future in order toassess the ways in which the existence of this enormous saddregulated change in all these valley circuits.

In principle, then, We believe that future fieldwork should bedirected at determining how the sadd Harara mediated in allthe nearby hydraulic circuits, situated in the valley below.Construction by the Himyarite state of Zafar, or by their clientgroups, seems initially evident. If this is the case, the orderestablished by the peasants at the bottom of the valley would beprior to the intromission of Zafar in the peasants' hydrauliccircuits. Were these circuits not to have existed first, it wouldhave been very difficult to subsequently establish interventionthrough control of the sporadic sayl channels from placesoutwith the peasants' ordering. This seems to us, all in all, to bean issue of greater complexity than the usual description ofhydraulic areas.

The sequence of hydraulic organisation in the Bayt al-Aswal valleyis possible from the systematic use of specific hydraulic knowledge.The empirical recognition of the geography of the channels determinedby the monsoon is accompanied by the knowledge of how to constructcultivation terraces. To simplify, we could say that the buildersresolve two integrated questions. How to construct an area ofcultivation, how to concentrate and distribute the necessary water,and how to avoid the constant erosion of the soil that is involved.The technical expression of this issue is the maqlah, a depression,in the form of a deep channel, transversal to the natural slope, inwhich the rainwater and channel filtration collected. Its functionwas to moderate the current flow that is inevitably beyond thecontrol of the channels, and to concentrate the filtration of theirrigation water. The maqlah therefore contributed to thedraining of the terraces and constituted a much-needed element ofprotection for these areas. But they must also have had a primordialfunction in the process of terrace construction. The maqlahchannel is dug into the ground, but on its upper section closest tothe direction of the slope, there was usually an upper-terracecontention wall of considerable magnitude, in some cases reachingseveral metres' width. Their function in the process of constructioncould be, therefore, that of halting or slowing and accumulatingsediment brought along with the channels in order to create a'terraceable' surface just above the height at which they weresituated. The direction of construction in the cultivating area is,then, from bottom to top, causing the deposit of sediments broughtalong by the channels. For this reason, the siting of themaqlah-s is spaced, at the key points of concentration for theearlier natural current, and may be used in establishing phases ofconstruction. Clearly, they are a device, and may be confused withthat of a deep irrigation channel. They seem to establish intersticesin the irrigated area, yet do not constitute any type of break withthe final circulation of the water (figure 4). In order to betterappreciate their importance in the construction and maintenance ofthe sequence as a whole, we represent only the circuit of connections(figure 6), their role in the construction and maintenance of theterraces and of their drainage being manifestly clear. Themaqlah, therefore, is a constructive device combining a greatdeal of peasant knowledge that was essential both at the moment ofits construction and when it began to carry out its function.

There is the opportunity at Bayt al-Aswal to verify the specificuse of this device, by excavation. In girab 13 and 14 insadd al-Umran, three metres underground, there are the ruinsof channels dug into the rock that must have constituted an essentialconstruction phase in the process of organising the terrace siltingagainst the direction of the water current. One of the aims of thenext campaign would be to verify, by systematic excavation, what are,up to the moment, casual peasant findings.

Similarly, one of the future objectives would be that of verifyingup to which point an excavation of a "palace" spatially associatedwith sad Harara, whose epigraphic name deserves to be kept, mightobtain results that would justify the cost of such an excavation.

* Maite Márquez, a graduated student of the UniversitatAutònoma of Barcelona, assisted helpfully in the fieldwork.

(1) In reference to this zone, see GIBSON, M.; WILKINSON, T.J.,"The Dhamar Plain, Yemen: A preliminary study of the archaeologicallandscape". Proceedings of the Seminar for Arabian Studies,25. 1995

(2) Hamdani, Abu Muhammad al-Hasan b. Ahmad al-/Muhammad b. 'Alial-Akwa (ed. 1986) Al-Iklil. Vol. VIII, Beirut.