Reconstruction of one stretch of the Hittite city walls

A project in experimental archaeology


Since 2005, the Hittite capital Hattusha in Central Anatolia boasts a newly added attraction. In autumn the reconstruction of a 65-m long stretch of the mudbrick city walls reached completion. Three segments of curtain-wall seven-to-eight meters high and two defense towers reaching a height of 12 to 13 m now convey the impression of how well fortified the city really was-and how impressive a sight it must have been to visitors of the period. This project represents the first major reconstruction in unbaked mudbricks; indeed, worldwide there are few projects comparable in dimension.

The reconstructed wall as seen from the north with the cliffs of Sarıkale appearing in the background



A stretch of the so-called Abschnittsmauer in the Lower City was selected for reconstruction. This internal wall screened off the area of the Great Temple and the habitation immediately surrounding it from the remainder of the Lower City to the northwest. [Ill. Plan] Some 200 m to the west runs the Postern Wall of the outer fortifications, constructed in the 16th century BC. According to our present understanding, the Abschnittsmauer was in use during the period of the Great Hittite Empire (14th-13th cent. BC).
The two factors most critical in deciding which section of the fortifications to reconstruct were accessibility and visibility. Visitors to the site generally follow an itinerary which guides them in a circle within the city. The line of the outer city wall lies beyond the pathway, often out of sight on high natural ridges with a steep drop-off to the exterior. The wall most easily accessible-both for the construction work itself and later for the visitors-was the Abschnittsmauer in the Lower City. At this point, too, is located the present entryway to Hattusa-so that the visitors' circular itinerary both begins and ends at this stretch of reconstructed wall.


Plan of the Lower City of Hattusa with the reconstructed stretch of wall marked in red

Hittite clay model of a defense tower, a find from Hattusa



The reconstructed wall seen from the west

General considerations/concept of the reconstruction

The project, begun in 2003 with the sponsorship of the JTI firm, has two principal objects: (1) to rebuild a monumental structure in mudbrick, and (2) to benefit from this reconstruction as a controlled and documented experiment.

1) The idea of restoring an entire Hittite structure here in the capital of Hattusa is by no means a new idea; it has long been a dream. Not simply to complete and stabilize the foundations exposed by excavation as had become the practice at the site, but to raise the walls and rebuild the roof. A complete reconstruction there before one's eyes would allow the visitor to grasp at one glance much of what we know about Hittite building techniques and to realize that Hittite architecture was essentially a mudbrick architecture, yes, that once mudbrick had been effectively employed for monumental structures. Thus an important task of archaeological research would be fulfilled: the handing down of what we have learned.

The decision of which type of structure to restore fell out in favor of the defense system. For it is only of the fortifications that we have pictorial images in the form of clay models from Hittite times [Ill. Clay model]. An attempted restoration of a house or a palace or a temple would be beset with question marks, for we have hardly any information on the appearance of the facades; we know only what we have excavated to socle-height. Neither would such large structures with flat roofs be aesthetically impressive from the exterior. Were the interiors opened to the public, we would be responsible for providing furnishings-furniture, wall decoration, flooring, and cultic objects, of which we know too little; once again our attempt would be highly speculative. A stretch of fortification wall, on the other hand, everyone finds to a great extent self-explanatory. Each and every visitor need not be invited inside, a great advantage when one considers the probable wear and tear from the increasing throngs of tourists.

2) The second objective of the construction is a contribution to experimental archaeology. The construction procedure, the work force, and the time allotted to each step in the process as well as to the amount of material involved have been documented, and later the observation of the reconstructed structure will be as well: the type and frequency of upkeep and repairs necessary. Both of the above are integral components of the project. Because-insofar as possible-we have employed the same materials available 3500 years ago, we now have a fair idea of how the work progressed in Hittite times. The bricks were prepared from a hand-mixed mass of loam, straw, and water. Quite intentionally, any additional stabilizer was avoided to provide a more accurate observation of the original material's resistance to erosion by wind and precipitation.

The enormity of the task faced by the Hittite master builders becomes obvious when one considers that our reconstruction represents only one percent of the ancient 6.6-km-long outer fortifications. (The sum of all elements in the city walls of Hattusa totals more than nine kilometers.) For the transport of heavy materials we benefited from the use of machines; some 2,400 tons of loam, 100 tons of straw, and 1,500 tons of water were needed for the mudbrick mixture alone. Around 1, 750 tons of earth then had to be moved to provide access ramps for construction, and a great number of logs brought in for the construction of the upper rooms in the towers. Today we have bulldozers, trucks and tank-trucks; the Hittites had only oxcarts and manpower. Thus the frequent mention in the cuneiform tablets of thousands of prisoners of war brought to the city as laborers is hardly puzzling, most particularly when you consider that in Central Anatolia mudbrick production is only possible in the summer months. Only from mid-June to mid-September, when the weather is hot and dry, is it possible to build mudbrick constructions-just at the time when military campaigns are carried out and the local labor force is kept busy in the fields!



Production of the mudbricks


Large pits in which the mud for the bricks was mixed by hand



Although unbaked mudbricks are a very stable building material, exposed to the rain and frost they crumble over time if not protected. This explains why so few Hittite superstructures have survived; there are fortunately a great number of accidentally burnt walls at Hattusa, some preserved to a good height. From these "fired" walls we know that the dimensions of the bricks averaged ca 45 x 45 x 10 cm. Such a brick would weigh about 34 kg.

The first step in the production of the bricks was the preparation of a mixture of loam, straw, and water in large pits. The straw was added as temper to prevent cracks from forming in the mass as it dried. It functioned as a "reinforcement" that worked against the traction and tension liable to create fissures during the drying process. The paste was mixed by hand with shovels and stomped with the feet. The next day it was mixed again and then left to settle for two or three days, which increased the eventual strength of the bricks. Simple wooden frames were used to form the bricks. They were filled with paste; the upper surface was flattened, and the frames lifted off. With continuous effort, one frame would shape as many as 120 bricks per day. A total of 64,000 bricks were produced in the course of the project.

Wooden frames used to shape the bricks

The bricks were then laid out in the open fields to dry. In sunny weather they were ready to be turned over in five days, and on the eighth or ninth day they were stood upright on one lateral edge to provide more air circulation. On the tenth or eleventh day the bricks were stable enough-if still quite damp at the core-to be built into the wall. The remaining dampness would dry very slowly, with a positive result: the lime in the mud would bond more gradually, thus increasing the strength of the wall. Furthermore, the remaining moisture provided a certain plasticity in the brick that enabled one to adjust the form to attain a regular height in each course-important for the solidity and stability of the construction.



Like all the other Hittite buildings, the mudbrick city walls were built upon socles of stone, and like the Hittite city walls, the Abschnittsmauer was also of casemate construction, that is to say two parallel walls connected by transverse walls at fairly regular intervals. The towers were constructed upon separate foundations (see plan). Within the area of reconstruction, a new socle of dry fieldstone masonry was raised upon the old foundations. The socles were laid out in steps reflecting the slope. Originally, a grid of timbers might have been set into or upon the socle to provide a base for the mudbrick superstructure. Such a technique was employed at other sites to absorb the effects of consequent settling and disturbances in the socle as well as to compensate for any horizontal stresses, thus averting a possible collapse. Because no traces whatsoever of such wooden grids were found at Bogazköy, however, we did not feel their use was justified in the reconstruction. For fear, however, of jeopardizing our reconstruction efforts through a possible misjudgment on our part here, we took the liberty of stabilizing the stone socle with cement mortar invisible from the exterior-a "non-Hittite" concession.

The widest walls, four bricks across (ca 1.9 m)

The upper story of the North Tower during construction


In accordance with the dimensions of the socle beneath, the curtain walls were begun at a width of three or four bricks; the thickness of the walls then decreases as they rise. The cists within the "casemate" walls were actually too small to have been used as rooms, and were therefore filled with earth. The walls of the lower stories of the towers were constructed vertically, with no reduction in width, while the walls of the accessible chambers above were constructed two bricks in thickness-still a good 90 cm across. The mudbricks were set in a mud mortar mixed in like proportions. Water was poured over the bricks before application of the mortar so that the dry surface of the bricks would not absorb water from the mortar and prevent stable bonding.

The upper floors have two doors that open onto the ramparts, and vertical rectangular windows styled after those of the clay model illustrated here were installed not only in the exterior and interior walls, but also in the narrower lateral walls looking out over the ramparts.

As the walls rose, building became more and more difficult. Because there is no evidence for the use of any crane-like apparatus in Hittite times, during the reconstruction we used no machines to lift the materials; instead, earthen ramps were built up against the back side of the wall, and the level of this sloping access raised whenever necessary until we reached the crown. After completion of the wall, the ramps were dismantled and the earth hauled away.

Timber elements/roofing

Large larch logs were used in the construction of the tower rooms. The lower wooden construction consists of a peripheral anchorage and a rough internal grid, the latter serving both to prevent distortion and to bind the central vertical support of the tower firmly in place; the upper construction serves the same purpose-as well as supporting the roof.


Roof grid of larch trunks in the North Tower

Floor-anchorage of larch trunks in the North Tower

The roofs have been built in the traditional Anatolian manner, which one can assume reflects the Hittite roofs as well. Across the ceiling beams are placed tightly-spaced logs of poplars. Upon this comes a layer of straw or slender branches followed by a coating of mud which is packed by treading and stamping upon it as it dries. The uppermost layer consists of a 10- to 20-cm coat of çorak, a watertight serpentinite that occurs widely in the surroundings of Hattusa and has been found time and again in various Hittite building levels at the site.

The surface of the roofs was divided into planes sloping slightly downward and directing the falling rain to gutters with water spouts. The latter, formed of branches cut in half and hollowed out, penetrated the mudbrick breastwork to spew the water forth on the city-side of the wall.

This type of roofing requires constant upkeep, such as one sees even today in the villages of southeastern Anatolia. After heavy rains, the surfaces of the roofs are compressed through a rolling process, and the slopes renewed; snowfall must be promptly removed so that the trapped and melting moisture cannot seep into the roof.

In addition, wood (oak, ash and spruce) was employed for the construction of a staircase in the North Tower and for the timber steps on the ramparts, as well as for doors, window frames, and shutters. Inside the tower rooms -over three meters high- there are also ladders leading to wooden trapdoors in the roofs.


Applying the mud coating over a layer of straw strewn upon the ceiling of poplar


Packing the watertight layer of çorak that seals the roof

Interior of the upper room in the North Tower

This type of roofing requires constant upkeep, such as one sees even today in the villages of southeastern Anatolia. After heavy rains, the surfaces of the roofs are compressed through a rolling process, and the slopes renewed; snowfall must be promptly removed so that the trapped and melting moisture cannot seep into the roof.

In addition, wood (oak, ash and spruce) was employed for the construction of a staircase in the North Tower and for the timber steps on the ramparts, as well as for doors, window frames, and shutters. Inside the tower rooms-over three meters high-there are also ladders leading to wooden trapdoors in the roofs.

Mud plaster-the armor of the mudbrick structure

Buildings of mudbrick are by no means restricted to warm and dry climates-even in Germany there exist even today multistory houses of mudbrick or packed loam and stucco, some of which have been standing for centuries. The secret of their survival is not only their solid construction but a particularly well applied plastering that is carefully looked after and repaired or renewed when need be. This "skin" protects the building material by shedding precipitation and insulating the substance from the effects of frost. Indeed the mud walls do absorb some moisture below a certain temperature, but this they can "exhale" just as easily through the minute pores in the plaster so that there is no danger of shattering through contraction and expansion.

There has been much experimentation, and a variety of methods to increase the resistance of mud-plaster suggested (e.g. Minke 2001, 75ff); these range from the addition of lime to that of ox blood, cow dung and dung-water, as well as gelatin obtained from bones. Because we have no hints that the Hittites used any such additives, we simply employed the same mixture used for the bricks and the mortar. In the preparation of plaster, however, the material remained in the pits for at least ten days. This settling process-also referred to as mauken -insures a better coagulation of the minerals in the clays, thus resulting in a more weatherproof substance.

The application of the plaster to the dampened surface of the bricks was carried out exclusively by hand; application with a trowel or wooden pallet would not insure sufficient binding. The plaster coating is relatively thin and must therefore be carefully applied. Cracks which appear as it dries are covered with a second layer of plaster, and finally any fine crevasses remaining brushed over with a fine paste. The resulting plaster does not create a smooth surface; as thin as it is, it rather enhances the structure of the wall beneath. Indeed the plaster is bonded to the masonry; it does not represent a thick separate layer that might flake or spring from the wall. Observations have thus far shown that the plaster is affected most of all by the mechanical beating of falling rain-most particularly when damp-which sets off a process of erosion that in time leads to a gradual dissipation of the plaster surfaces, most particularly those which are horizontal (the upper surfaces of the breastwork and the crenellations).

Mud-plastering following completion of the raw masonry


Application of the mud plaster by hand

The thin paste applied in the third phase of plastering


The thin "skin" of the finished plaster, still revealing the structure of the masonry beneath


The main part of the construction ended in October 2005. In 2006 some changes had to be made on various parts of the building, and at the same time the landscaping of the surroundings and construction of a path for visitors were carried out; with these the major phase of the project has reached completion and the second phase begun: the observation of the structure and the documentation of the attention and repairs it necessitates. Answering the question of what a large structure of unbaked mudbrick needs to survive in the harsh climate of Central Anatolia-how the plaster, the roofing, and the masonry itself respond to the burning sun, the rain, the snow, and the chill of the frost-constitutes an essential aspect of the project. In the coming years the observation of this structure will bring us further understanding of this type of architecture. Not only the planning and building of such a structure in Hittite times, but the problems of upkeep faced then as well will thus have been duplicated.