Mud, Sand and Sod's Law in Dusty Old Zabid

by Edward J. Keall
[West Asian Department, Royal Ontario Museum]

Yemen Update 30/31 (1992):24-26

There's not much green sod in Zabid. In the Yemeni citidel where the Canadian Archaeological Mission has its headquarters, the soldiers who occupied the abandoned barracks during the Gulf Crisis dug up the only bit of savannah grass growing there and planted a sorghum patch. One of the soldiers was enterprising enough to plant papaya where the overflow from our water tank made the ground constantly moist. But, for the most part, we found our compound looking about the same as it ever had, without a blade of grass to keep the dust down. Without irrigation, we must cultivate thorny acacia instead of green sod.

There is, however, a Sod's Law in archaeology: Tantalizing Discoveries Will Always be Made When it is Time to Close Down the Operations. The last days of a dig bring mayhem as new features suddenly show up in the baulks, beckoning with their enticements of "dig me, dig me". At a time when funds are hard to raise and when for other reasons one cannot guarantee that one will return, it is tempting to seize the opportunity to dig more, even if in haste and with the risk of poor quality documentation.

This season's work at Zabid between December '91 and February '92 was sponsored by the Royal Ontario Museum as usual, with significant funding from the National Geographic Society for the excavations and a supporting grant from the Royal Netherlands Embassy in Yemen for restoration work. The excavations involved investigation of what the region's environment was during the heyday of the city (9th-16th centuries). Many questions had been raised during the 1988 season, when we had probed various parts of the city's environs.

The first site investigated was a potter's workshop (area ZBE) just outside of the present town, in an area increasingly used for the tipping of garbage. Our aim was to find out before it was too late more about the workshop and the field where the kilm had been located. The work promised to provide signficant information about Zabid and the transition from rural to industrial land use on the outskirts of the city in the 9th century. To assist in interpretation of the field and its crops the Project had enlisted the help of Irene Good, an archaeo-botanist with experience in Turkey, and Ingrid Hehmeyer, an agricultural specialist who had worked with the German team that surveyed the Ma'rib dam and its irrigation networks.

On opening up the old trench we learned a useful lesson when we realized that a hard band of fine sand exposed in cross-section represented a deposit blown by the wind into the trench which we ourselves had back-filled in 1988. Distinguishing between wind-blown and water-borne sediments became an important part of the excavations. Seeing alternating layers of silt and sand had made me think they represented controlled irrigation with sand blowing into the field during the fallow season. Our irrigation specialist, however, declared, "No, mudir (Arabic for "director"), it's not a field." There was nothing in the mud to suggest it had been farmed or cropped. Upon excavation the seemingly regular pattern of depressions that we had taken to represent furrows or crop marks turned out to be the products of flood water strong enough to create eddying currents that left ripples in the mud like those on a beach.

Although we reluctantly accepted that we had not found an irrigated field of the 9th century, the discovery did help explain why the pottery workshop was there: clearly the potter was exploiting a clay source deposited there by flood water from Wadi Zabid. The various aggregates visible in cross-section could readily have been used for the variety of pots made on the site, ranging from small pitchers of finely levigated clay to coarser sand-tempered pots for cooking. As the potter continued to divert wadi water into the area to create more sediments, the clay beds rose to an alarming height. One disastrous flood brought about the collapse of the edge of the "settling pond"; to remedy this, the potter dragged a whole lot of debris to the edge, to form a new bank. The debris included kiln wasters and other workshop remnants.

The loss of the expected information on farming was disappointing, but the archaeo-botanist found herself concentrating on how to judge seasonality, how to discover when the activity in the workshop occurred. Was it year round or, for instance, closely tied to the flood seasons? To answer this question, we extracted soil samples that will be analyzed for pollens. Although the pollens found in this soil may have been airborne from quite some distance away, they may help reflect the season of the year when these surfaces were exposed to the air. The rapid accumulation of the sediments and the sudden tipping of refuse to form a bank hold the promise of good seasonal separation for the pollens, because around Zabid the flood waters come only during the two limited rainy seasons of the spring and late summer monsoons.

Strong flood waters reached the area during its last phase of use, some time around the 11th century. We know this from the presence of flood-borne gravel and boulders in the topmost layer of the site. In fact Sod's Law came into play in this discovery, because most of the digging crew had already been transferred to another site when the significance of the gravels at such a high level became apparent. The trench received attention only because of some last minute "trouble shooting" regarding the definition of a mud berm, which we then found to have been cut by the flood and which may represent the last time a potter deliberately tried to control water in the area.

At the new site of Qasr Bani Najah (area QBN) north of the city, the two U.K.-based archaeologists, Christopher Evans and Prince Chitwood, blossomed. Although they had claimed to be very much at home on the kiln site because the pits filled with rubbish and washed out by floods reminded them of Anglo-Saxon village archaeology, Qasr Bani Najah offered them a challenge in the sand, something that Cambridgeshire could never give them. The purpose of their work on this site was to open up and expand another of our 1988 probes. The original work had been structured to kill two birds with one stone, to learn what we could about the legend of the 11th century Najahid Palace and to develop our knowledge of the 13th-15th century potsherds that littered the ground. As it turned out we found out little about either, except that enormous amounts of bricks had been dug up from the site in years past to build the houses in Zabid, and hence there was little left of the "palaces" and not much pottery below ground.

Disappointing though this discovery was, the 1988 work had raised other questions that the 1992 campaign set out to answer. The fact that there were massive amounts of wind-blown sand around the site when the structures were built tells us that the environment of medieval Zabid must have been little different from that of today. Three separate building phases were encountered in the excavations, and we can tell that all were erected in sandy conditions. Although the walls were gone, their alignment could be traced by following the robber's trenches. The original building was set down directly into pure sand. Even now we cannot be absolutely sure that there was not an earlier occupation of the site beneath this dune, but it was too dangerous to dig deeper to find out.

Members of our team battled horrendous conditions as they tried to draw sections in the late afternoon when the sand swirled around, but it was apparent to me that conditions in the area had improved considerably since our 1988 season. A dune stabilization program had been initiated just north of the site, with acacia and other shrubs planted and sustained initially with irrigation pipes to anchor the sand. In 1988 we had found it almost impossible to work on site after mid-day when the winds picked up. The judgment as to whether conditions really were as bad as that 500 years ago still hangs in the balance. Again, the pollen analysis will address the question of seasonality; over what period of time did the sand accumulate? Was the dune outside the building formed in a day, a month, or a year?

Excavations on the palace mound were still in full swing when news came that construction work had exposed a solid brick feature in a field on the other side of the main road. Our investigations on site CZB revealed that it was a water-supply system, something we had been meaning to investigate anyway at the third of our proposed work sites. This was an opportunity not to be missed; the underground water conduits of Zabid have high repute in the folklore of the city, but no one had ever been able to show me where they were. Our attention to the emergency rewarded us with the discovery that the system consisted of three pipelines of glazed tile set into a bed of bricks laid with lime mortar. Two sections were exposed, over one hundred meters apart. The pipes continue underground in both directions. There was virtually no soil sediment in the pipes, and clearly a great deal of effort had been taken to ensure that the quality of the water was good. Soil samples taken from the original building trench can be dated to at least earlier than the 16th century on the basis of the small scraps of pottery we found in the trench fill.

The discovery of the pipes was a bonus, but it brought Sod's Law into effect once again. It meant that we started work on the third site rather late in the season. I had been taken to see the feature in 1988, when I was shown how a flood had exposed a length of brick masonry in the bank of a field. As a reminder of why in archaeology one should always act sooner rather than later, I was shocked to find that since 1988 the enterprising landowner had planted fruit trees along the stream bed. When it came time to do work that earlier might have taken only a couple of days, we were hampered by this new plantation. In fact, a deep probe inside what may have been a catchment basin had to be aborted when our trench filled up with water seeping down from the irrigation ditch higher up the bank.

As chance would have it, however, only the day before we started our work a bulldozer exposed more brickwork just upstream. Our last-minute probe suddenly became a last-ditch effort to extract as much information as we could from the dirt before the information disappeared forever. Two parallel pieces of brick masonry could be seen snaking through the field where the bulldozer had scraped along their top. Envisaging on the last day of scheduled excavation an easy exposure of a conduit some 35 cm wide and perhaps of the same depth, I returned to the "palace" site (QBN) for, as usual, last-minute recording. Imagine my surprise on returning to the "water" site (ZHB) to find that the sides of the conduit were going on down, beyond a meter in depth. After a meter and a half it became impossible to dig anymore, because the space was less than shoulder width. From the deep layers of sediment within the narrow space we could tell that the conduit had carried flood water and that it had been necessary to build it up higher on one occasion. Perhaps this is a reflection of the sediments rising, in the same way they rose rapidly on the kiln site.

Since the conduit appears to run directly towards the first feature, now surounded by fruit trees, it seemed logical to connect the two. Excavation revealed this other feature to be a circular distribution chamber with a flat entry trough and outlets of glazed pipes. One of the pipes was badly worn and had been repaired with a sleeve of lead inserted into the mouth. Two of the pipes had been deliberately blocked with a pack of lime plaster. We hope that this sealed dirt will provide us with some interesting environmental information. From the pottery found inside the abandoned drum chamber, we know that the system had already fallen into decay well before the 16th century. These conduits hold the promise of being remnants of Zabid's most famous water systems of the 12th-15th centuries. Clearly, a future season has to be directed towards tracing the course of these conduits underground, both towards their planned destinations and from their respective sources. I hope there will be enough time to avoid some of the worst pitfalls of Sod's Law and that we will be able to figure out how they really worked.

[This article is reprinted from ROM Archaeological Newsletter, series II, No. 48, March, 1992]

Search Site

Search Library Collection