Madīnat az-Zahrā’ was the capital of the 10th century Umayyad caliphate in al-Andalus and is today one of the most important sites of Islamic culture in the west. A new project is dedicated to the investigation of the Plaza de Armas, the main public square of the caliphal city. The aim of the first season of field work was the investigation of the Great Portico, which delimits the Plaza de Armas in the west and forms the façade of the caliphal palace. Based on archival material and observations in the field, the building history of this part of the site was studied.
The main focus of the first season was the documentation of some 1200 building elements that had been recovered during the excavation of the Great Portico in 1974-1975. Among the elements recorded this season are column bases, shafts, capitals, imposts, modillions and glazed roof tiles which derive from a pavilion that had stood above the central nave of the Great Portico. Based on these elements, a detailed reconstruction of the portico can be proposed. Several building elements can be attributed to a second, previously unknown pavilion that must have been added to the portico by al-Hakam II.
In addition, a geophysical survey was conducted in the area to the east of the Plaza de Armas, in order to study the eastern limit of the plaza. The dimension and structure of several buildings could be determined. The results of the survey suggest that an audience hall stood to the east of the plaza, with a courtyard located in front at a much lower level.
Fig. 1: View across the Plaza de Armas toward the Great Portico
Results of the First Season of work at Madīnat az-Zahrā’
Madīnat az-Zahrā’ is one of the most important archaeological sites of the Islamic culture in the western Mediterranean region. The city was founded in 936 CE by cAbd ar-Raḥmān III near Córdoba (Spain) as the capital of the Umayyad caliphate. Archaeological work at the site has so far focused mainly on the central palatial area. While the architecture of the palaces is thus fairly well known, the relationship between these palaces and the surrounding city has not been studied in detail so far. The aim of the current project is the archaeological investigation of a large public space which was located in front of the palace gate. Known from historical sources as the staging ground for military parades, public audiences and executions, the so-called Plaza de Armas has never been studied before. The investigation of this public space will not only add to the understanding of the overall design and evolution of Madīnat az-Zahrā’ as a caliphal city, but will also provide new insight into the relationship between the caliph and his subjects and the way the caliphal state was interpreted by the Umayyads by architectural means.
The first season of field work was dedicated primarily to the study of the so called Great Portico (gran portico), the only element of the Plaza de Armas known so far. The portico was excavated by Félix Hernández in 1974-1975 and partially restored by Rafael Manzano Martos in 1975-1982. The Great Portico functioned as the façade of the caliphal palace and constituted the western limit of the Plaza de Armas. It comprises a 120 m long arcade, composed of 15 arches, with a wide central arch that indicates the location of the main access to the palace. Above the central arch lay a pavilion from which the caliph could present himself to the people gathered in the public space.
Work conducted at the site in the past decades by Antonio Vallejo, a reevaluation of the existing documentation and a preliminary survey of the visible building elements suggest that the portico was not part of the original design of the city. In fact, at least five major building phases can be distinguished in this area of the site, providing valuable evidence for the reconstruction of the evolution of the city as a whole (Fig. 2).
Fig. 2: Plan of the excavated area of Madînat az-Zahrâ' with building phases
Phase I (brown): At first the city appears to have occupied only part of the site, the back wall of the portico forming its eastern limit. The Plaza de Armas was originally located outside the city walls, as an open space in front of the palace gate.
Phase II (black): The main mosque was erected in 940-941 outside the original perimeter wall, to the south of the Plaza de Armas. At the same time the palace walls appear to have been reinforced.
Phase III (red): The city was later extended to the east by about 680 m and surrounded by a fortification wall with square towers. Inside the palace the famous Salón Rico and the great three-ailed hall were erected at this time. Inscriptions indicate that these buildings were under construction between 953 and 957. The great portico appears to have been part of this refurbishment of the palace, with the aim of creating a new palace façade.
Phase IV (blue): At a later stage some of the arches were blocked and office built inside the portico. A new gate was added in the northern wall of the Plaza de Armas, with a ramp leading down to the Plaza and providing a direct access from the outside. Building elements found in the portico suggest furthermore that a second pavilion or hall was erected above the portico. An inscription on a capital suggests that these changes were undertaken in the reign of al-Hakam II, possibly between 970 and 973.
Phase V (green): Additional structures were added later, before the site was finally abandoned and sacked in 1013.
Documentation of architectural elements
During the excavation of the Great Portico in 1974-1975 a large number of objects were found in the debris, including building elements, pottery and other small finds. Many hundred pieces from this excavation are now stored in the magazine of the museum of Madīnat az-Zahrā’. Additional pieces remain in boxes in a storage room at the site. Most of the building elements were registered in 1986 and given a number indicating their provenance (65 for the portico, 64 for the ramp leading up from the portico to the palace plateau in the west). The pieces now in the museum were given an additional inventory number, based on a new numbering system (34 for the area of the portico).
The main focus of the first campaign of the present project was the comprehensive documentation of these building elements. More than 1200 pieces were drawn, photographed, described and entered into a database. As far as possible existing inventories were integrated, providing additional information on the precise provenance of the pieces. The objects documented so far include 483 fragments of floor slabs, 27 fragments of thresholds, 14 column bases, 114 column shaft fragments, 44 capitals, 69 imposts, 87 voussoirs, 11 decorated wall slabs, 31 decorated frieze fragments, 6 modillions, 8 merlons, a door frame, a small arch, 74 window grills, 14 water basins, a fragment of a Cufic inscription, four fragments of Roman sculptures, several game boards and a large quantity of glazed roof tiles.
The building elements studied so far provide detailed information on the original appearance of the great portico. Many pieces can be attributed to the pavilion which stood above the central arch of the portico. These include fragments of bases, shafts and capitals of columns as well as imposts and modillions. At least seven column shafts of three varieties of pink marble are preserved, four of them nearly complete. The shafts have a diameter of 24.5 cm at the base and 21.5 cm at the top and are 187-191 cm long. Six almost complete capitals of white marble are preserved, two of Corinthian type, four of a composite type (Fig. 3). Including bases, shafts and capitals the columns were 2.32 m high. The diameter of the columns suggests that the columns were placed about 1.65 m apart. If the pavilion was as wide as the bay of the portico below its façade could have comprised a 5.2 m wide arcade divided into three bays (Fig. 4). The number of columns preserved suggests that at least two sides of the pavilion were furnished with such an arcade, possibly three or four sides.
Fig. 3: Capital 34.24089, deriving from the pavilion above the Great Portico
Fig. 4: Reconstruction of a column (left) and the pavilion above the central nave of the portico
The fragments of floor slabs found in the debris of the portico (find complex 65) are predominantly made of white marble, those found in the adjoining ramp leading to the palace (find complex 64) predominantly of white limestone. Since the floor of the portico was paved with yellow limestone and the ramp with pebbles, the slabs must derive from the upper floor. Quite possibly the pavilion above the portico was paved with white marble, the space above the ramp with white limestone. This space above the ramp could have served as an antechamber or a forecourt to the pavilion.
During the excavation of the portico a large quantity of glazed roof tiles was discovered. They are the earliest such tiles to be found in Spain. They probably derive from the roof of the pavilion. The tiles are of the traditional barrel type and were glazed on the exterior, on the interior, or on both. Sherds of a minimum of 52 tiles were documented, 44 glazed in green color, 8 in yellow color. The intention may have been to create a yellow pattern on a green background, possibly a Cufic inscription or a geometric pattern.
Several modillions found among the building elements suggest that the roof of the pavilion was supported by a cornice. The façade of the central bay of the portico would thus have resembled the design of the gates of the Great Mosque of Córdoba, with a wide arch below, an arcade above and a row of modillions at the top (Fig. 4). The prototype for both may have been the façade of the Alcázar in Córdoba. Historic sources indicate that an audience hall existed above some of the palace gates of the Alcázar. The overall design of the portico (Fig. 5) is reflected by some later building façades, including the façade of the palace of Pedro I at Seville.
Fig 5: Reconstruction of the Great Portico. Original design (below) and later transformation (top)
A number of building elements cannot be attributed to the pavilion known so far. These include at least three column shafts of gray limestone as well as bases and two capitals of white marble, all deriving from columns of a size significantly smaller than those of the pavilion. All pieces were found to the south of the central arch of the portico, either in the 11th or the 12th bay of the portico counting from the north. One fragment of a capital was inscribed with the name of al-Hakam II, with caliphal titles known only from the years 970-973. The pieces may originate from a second pavilion, which was added by al-Hakam II to the existing portico, possibly above the 11th or 12th bay (Fig. 5 top).
A significantly large number of window grills are found among the building elements (74 fragments recorded so far). Each grill appears to have been of a different type, much like those at the Great Mosque of Córdoba and in the Salón Rico. Many of the grills could derive from windows placed at the back wall of the portico, or in adjacent buildings. Some of the grills could originate from the pavilion, however, and could have served to close the bays of the arcades. The caliph may thus not have been visible to the public gathered in the square, while he himself would have been able to look onto the public space through the grills.
Interesting are several fragments of sculptures of white marble dating to the Roman period. Several pieces might derive from a sarcophagus that had been reused like other examples at Madīnat az-Zahrā’ as a water basin. The presence of a Roman portrait bust is more difficult to explain. The sculpture was possibly placed above the central arch of the Great Portico. Examples for this are known from the city gates of Toledo. Historic sources mention the sculpture of a female figure placed above the main city gate of Madīnat az-Zahrā’.
Overall the documentation of about half of the total material stored in the magazines was completed this season. The aim of the coming season will be to record the remaining half, thus providing a comprehensive documentation of all known building elements deriving from this part of the site.
Fig. 6: Fragments of a decorated window grill from the area of the Great Portico
The 120 m wide Great Portico constitutes the western limit of the Plaza de Armas. Aerial images and surveys indicate that the plaza itself was about 155 m long from west to east and was delimited on the side opposite the portico by a large, so far unknown building. In order to gain further insight into the structure of this building David Jordan of the University of Liverpool conducted a geophysical survey, using Electrical Resistance Imaging (ERI). ERI is a widely-used geophysical survey method in which an electric current is injected into the ground between pairs of electrodes and the potential difference it produces is sampled using further electrode pairs. By varying the location and distance between these electrodes it is possible to sample potentials produced in different volumes of the ground. Computerized numerical methods can then be used to reconstruct the 3-dimensional distribution of resistivity in the ground, within certain limits of depth, spatial resolution and confidence.
ERI surveys of complex, stony archaeological sites have proved able to distinguish volumes of intact masonry, rubble, soil and bedrock in three-dimensions and under a wide range of conditions. Thus, since two-dimensional ER survey has proved very effective elsewhere at Madīnat az-Zahrā’, there was good reason to favor ERI survey and to reject alternative methods to map the three-dimensional structure of remains around the Plaza de Armas. Difficulties were nonetheless evident. The complex surface topography across the area requires that ERI results are corrected for the changes in apparent electrical resistivity due to geometry alone. The abundance of stone within the soil, moreover, makes it difficult to insert electrodes between which measurements could be made – especially under the dry conditions at the time of this survey. Finally ERI is a much slower method than magnetometry and GPR. Thus the area which could be covered by a test-survey would naturally be smaller – or require more widely spaced samples.
In the first season David Jordan was able to study 12 profiles using ERI, up to 100 m in length each. The profiles provide information on the location, size and preserved height of several large structures located to the east of the Plaza de Armas. As expected, the buildings appear to be located on individual terraces. Directly to the east of the plaza lie two terraces, one in the north and one in the south (Fig. 7). The floor level of the northern terrace (about 194 m a.s.l.) appears to be not much higher than that of the Plaza de Armas (about 193 m a.s.l.), while the floor level of the southern terrace is much lower, about 5 m lower than the terrace to the north (about 188 m a.s.l.). The northern terrace is densely built over by thick masonry walls preserved up to a height of 2-4 m (red and orange in the profile), while the southern terrace appears to be occupied by a relatively open space. The size and layout of the terraces would suggest that the northern, upper terrace was occupied by a large audience hall, resembling in layout the great three-ailed hall to the west, while the southern, lower terrace by the corresponding courtyard. The difference in floor level between hall and courtyard is familiar from other audience halls, though a difference of the present magnitude – about 6 m – is known so far only from the Dar al-Jund, the earliest hall at Madīnat az-Zahrā’ known so far.
Fig. 7: North-south profile east of the Plaza de Armas. Left, remains of buildings on a higher terraces, right, an open courtyard on a lower terrace