Given the paucity of Bronze Age evidence in the Guadalquivir Valley itself, the previously unpublished site of Mesa Redonda, higher up on the edge of the valley, has been the subject of intensive research, since Bronze Age remains are known from there. Mesa Redonda is located in the municipality of Villaverde del Río -about 30 km north-east of the city of Seville- on the first elevations of the Sierra Morena mountain ridge (Fig. 1). The site is located on a prominent hilltop, some 200 m above the valley, with a wide panoramic view (Fig. 2). The mountain, which slopes in all directions, forms a flat conical peak at 239 m. a. s. l., with a rounded to oval base, which explains its name (Fig. 3). On the surface, around the edge of the summit plateau and on the slopes immediately below, extensive massive dry-stone walls can be seen (Fig. 4). There and on the summit, a large amount of archaeological material is exposed on the surface, testifying to an intensive occupation. The summit area is surrounded by a wall that runs down the slope along the edge of the site. In some areas, the wall is only vaguely visible, while in others, particularly on the eastern slope, it is not visible at all (Figs. 5 and 7). To the east, at the foot of the slope, just below the mountain, flows a permanent watercourse, the Siete Arroyos stream. The landscape is dominated by Mediterranean woodland, with oaks (genus Quercus) and Mediterranean fan palms (Chaemerops humile), and the typical dehesa type of landscape of the Sierra Morena. It should be noted, however, that a few decades ago there was very little vegetation due to intensive pastoral use (Fig. 6). The geology around Mesa Redonda is characterized by metamorphic rocks, granodiorites, schists and quartz-diorites.

Mesa Redonda has been known as an archaeological site for several decades and has repeatedly been the target of looting excavations, with many of the objects recovered ending up in private collections through the antiquities trade. The most prominent example of this is the Fondo Arqueológico Ricardo Marsal Monzón collection, now owned by the Junta de Andalucía (Aguilera Collado et al., 2014).

The first study of Mesa Redonda was the archaeological excavation carried out by Javier Verdugo in 1977 (Huntingford, 1983). The purpose of this intervention was to carry out three test pits and to uncover a Roman cistern that had been discovered on the summit (Huntingford, 1983). Two test pits were excavated in the central area of the summit, no deeper than 30 cm, and yielded very little archaeological material. A third test pit was excavated in the north-western part of the hill, where many structures and archaeological material dating from the Late Bronze Age to the Middle Ages were found. Following this work, Mesa Redonda became an officially registered archaeological site in the SIPHA (Sistema de Información del Patrimonio Histórico de Andalucía) database (code ID 01411010003).

In 1978 and 1979, two excavation campaigns were carried out by the University of Barcelona under the direction of Elizabeth Huntingford Antigas, with the aim of clarifying the archaeological potential of the site. On this occasion, two excavation sections of 3.5 × 4 m and 5 × 4 m were conducted on the northern and western edges of the summit plateau, where archaeological strata were traced to depths of 5.5 m, and 3.5 m, respectively. However, due to the potential instability of the trench profiles, the danger of going deeper became too great and the work was stopped. In the first trench, which began with High Medieval finds close to the surface, the team reached layers containing materials from the 7^th century BC. In contrast, the second trench yielded layers from the 8^th to 9^th centuries BC (Huntingford, 1983). However, Late Bronze Age and possibly earlier material was repeatedly identified during the excavations, both on the surface and in the stratified material of the trenches. Unfortunately, the results of the project were never published. However, a detailed cultural and historical evaluation of the excavations was produced by the excavator in 1983 as part of an extensive doctoral thesis submitted to the University of Barcelona (Huntingford, 1983).

The find material collected on the surface in the summit area and on the slopes shows an almost uninterrupted sequence from the Bronze Age to the 12^th century AD. Characteristic forms of the Bronce Final (Final Bronze Age) (Fig. 9, G-M), which correspond directly to those of comparable sites in the Lower Guadalquivir Valley (e. g., Montemolín: Ferrer Albelda et al., 2022, figs. 33-45, 47-55, 57; Setefilla: Aubet et al., 1983, figs. 24-30). Among them, there are also some that may well be older and allow dating to the Bronce Pleno (Full Bronze Age) (Fig. 9, A-F), as shown by comparisons with corresponding materials from other sites (Chala Aldana, 2022, figs. 28-38; 2024, 119-127, 215-292; Setefilla: Aubet et al., 1983, figs. 15-22; El Trastejón (Zufre, Huelva): Hurtado Pérez et al., 2011, figs. 2.68-2.69; Cerro de San Juan (Coria del Río, Seville): García Rivero and Escacena Carrasco, 2015, fig. 7). However, since there are hardly any closed settlement complexes of the Bronce Pleno in the lower Guadalquivir region (Bartelheim et al., 2021), a typological assignment of the material found on Mesa Redonda to this period is difficult. The dating possibilities into the Bronce Pleno are clearer with the metal objects from the Marsal collection (Fig. 10), which, according to the origin noted there, come from Mesa Redonda, presumably from looted grave finds. With the halberd, the dagger and the Palmela point we have finds that can be clearly assigned to this period (Brandherm, 2003, pp. 175-177; 245-249; 380-413). This is confirmed by the discovery of materials from the Bronce Pleno at the Siete Arroyos site (Díaz-Zorita Bonilla et al., 2019). In addition, the concentrations of Chalcolithic settlement pottery found at the foot of the hill on the terrace above the Siete Arroyos stream, where Bronze Age materials are absent, as well as the Chalcolithic material on the slopes and the summit, indicate the existence of an earlier site there. This could evidence that the settlement activity within the Mesa Redonda Complex was transferred to the summit at the beginning of the Bronze Age.

The protohistoric phase is represented by a wide range of materials from the entire first millennium BC. Some of these materials can be associated with the Early Iron Age (8^th to 6^th century BC), while others, such as several hand-made pottery pieces found during profile cleaning, could belong to the Late Bronze Age. It is important to note that this pottery can sometimes be confused with Late Antiquity and Early Medieval handmade pottery. The absence of clear stratigraphic sequences along the profile makes it impossible to be more precise. Similarly, during the first two centuries of the Roman occupation of the Iberian Peninsula (2^nd and 1^st centuries BC), some locally produced pottery from the Late Iron Age (5^th to 3^rd centuries BC) is still in use. Imports of amphorae and tableware from other parts of the Mediterranean complemented the local production of traditional pottery. During the surface survey it was noted that the Early and Late Iron Age materials were more concentrated on the southern and south-western edges of the Mesa Redonda, next to the wall line and the terraces south of the hilltop (Fig. 11). In the Early Iron Age assemblage, handmade, local, traditional pottery represents 90 % of the material, while wheel-manufactured pottery constitutes 10 %. The greyish ‘Orientalizing style’ pottery (bowls, plates and unspecified vessels), the common use pottery (tableware and storage jars) and the typical western Phoenician amphorae (G-10 according to J. Ramon, 1995) make up this 10 % (Fig. 11). In terms of handmade pottery, the well-treated or smoothed surfaces represent 75 % of the material collected, while 17 % is ‘rough-surfaced’ pottery. The smoothed pottery included cups, plates, carinated vessels, hemispherical bowls, straight rimmed cups and biconical cups. In the case of the rough-textured pottery (Kuass ware), its sole presence at the top of the Mesa Redonda is a significant finding. It is the first time that this type of pottery has been identified in settlements in the Sierra Morena region of Seville.

The materials collected in the survey are similar to other materials documented in several settlements along the Middle and Lower Guadalquivir Valley during the first millennium BC, especially with regard to their technological and typological aspects. As already mentioned, the handmade pottery can be associated with the transition between the Late Bronze Age and the Early Iron Age, especially considering the double truncated conical vessels and the burnished reticulated decorations on several bowls. By comparing the pottery found at Mesa Redonda with other sites with protohistoric levels, it is possible to date the earliest protohistoric phase at the site to the first half of the 8^th century BC. This phase is contemporaneous with the phase II of Montemolín (Marchena, Seville) (Ferrer et al., 2022, pp. 89-102); the stratum I of Cerro Macareno (La Rinconada, Seville) (Pellicer et al., 1983, figs. 74-75); the strata I and II of Cerro de la Cabeza (Santiponce, Seville) (Domínguez de la Concha et al., 1988, pp. 156-157, figs. LVII-LIX) and the strata X and IV from the test pits CA80/A and CA80/B respectively, in Carmona (Pellicer and Amores, 1985, figs. 13-16 and 148-149). However, as mentioned above, Early Iron Age wheel-manufactured pottery is scarce. The least represented phase corresponds to the 5th and 4th centuries BC. During this period there was a general contraction and restructuring, or even disappearance, of several settlements in the Middle and Lower Guadalquivir Valleys. This does not mean that there were no settlements at all. Although there are some surface materials from this period in Mesa Redonda, their small number could be a sign that these materials were brought to a part of the site that may not have been inhabited at the time. This phenomenon can also be observed in other settlements, such as Cerro de la Cabeza (Santiponce, Seville) or Mesa de Setefilla (Lora del Río, Seville) (Aubet, 1989, p. 308). This represents a socio-political and economic reorganisation of this part of the Guadalquivir Valley (Ferrer and García, 2019). In this regard, the Late Iron Age pottery documented during the 2019 survey campaign belongs to the typical late production style, based on its morphological characteristics and sparse decoration (Fig. 11). In fact, some of these materials could be contemporary with the beginning of the Roman period (2^nd and early 1^st century BC), and therefore with some of the Italic imports identified at the site and the first ‘Romanised’ productions (Figs. 11- 12).

The Roman pottery shows a very clear trend when considering the evidence from across the site. In general, over 68.5 % of the datable sherds belong to an occupation period between 100 BC and 100 AD. Roman pottery outside of this precise time span is barely present, accounting for no more than 5.13 % of the total period considered (Fig. 12). Although we do not find significant local trends in the different sectors of the site, we would like to highlight some facts that need to be considered when interpreting the nature of the Roman settlement at Mesa Redonda. The only place where evidence for continuous occupation until the Late Roman and Islamic periods has been identified is within Trench 1. The abundance of this late material is less substantial in Trenches 2 and 4, and completely absent in Trench 5, where the Roman pottery disappears by the end of the 1^st century AD. Overall, Trench 2 exhibits the strongest presence of Roman pottery, but within a specific time range spanning from the late Republican to the early Empire period. It should also be said that the Late Roman pottery is rather insignificant.

Nevertheless, the ceramic evidence suggests that Mesa Redonda and its surroundings were occupied until the beginning of the 2^nd century AD. There is no other evidence to confirm the existence of a Roman settlement on the hilltop after this date, except for the possibility that the plateau was frequented or used for other purposes, such as defense and other special purposes, as evidenced by the cistern dating from the 2^nd century AD. In Late Antiquity, the use of Mesa Redonda has not yet been established for all phases, but in general a continuous occupation seems feasible (Fig. 12).

The medieval archaeological material recovered from Mesa Redonda can be roughly divided into four relatively well-defined chrono-cultural moments within the Middle Ages on the Iberian Peninsula from the 7^th to the 12^th century AD, corresponding to a period of transition towards the Islamic world. From the pre-Islamic period (7^th to early 8^th century AD) and the Emirate period (early 8^th to early 10^th century AD), most of the pieces belong to the functional groups of storage and transport, domestic use and kitchenware. The ceramic fragments attributed to the subsequent Caliphate period (early 10^th to early 11^th century AD) all belong to the functional groups of storage and transport, domestic use, fire vessels, kitchenware and tableware (Fig. 13). Although a significant proportion of the items in the inventory are exclusively attributed to the Caliphal period, a considerable number of pieces also exhibit characteristics that align with those of the subsequent century. These pieces display few technical or morphological differences when compared to the productions of the Taifa period, suggesting that they may also belong to that period. For this reason, in the present analysis of the materials, they have been treated as a single ceramic group of the 10^th and 11^th centuries. It should be noted, however, that in the individual and overall analysis of the productions and pieces, their relative chronology has been indicated whenever possible.

The largest group of pottery analyzed in this study is that of the Taifa period (11^th century AD), comprising kitchen pottery, domestic pottery and, most notably, tableware. The latter category includes a significant number of pieces from Sevillian production of the Abbadí period. In fact, there is compelling evidence that the Taifa pottery found on the Mesa Redonda was produced in the workshops that were recently identified in the vicinity of the Alcazar and the Cathedral of Seville. This is because the pieces from this period exhibit technical and morphological characteristics that are very similar to those found in these two Sevillian monuments (Huarte and Lafuente, 2000). The Almoravid period (late 11^th to mid-12^th century AD) is the least represented in the documentation. Nevertheless, the recognizable pottery fragments also include a substantial collection of the earliest known productions from the North African/Berber period. The total absence of one of the so-called ‘diagnostic artifacts’ of the Almohad period, the ‘rib casserole’ (cazuela de costillas), from the ceramic record could indicate that the upper part of the hill was abandoned in the mid-12^th century. However, until further work is carried out at the top of the Mesa Redonda, it is not possible to rule out the possibility that the site was occupied during the period of Unified Rule.

The presence of early Islamic wares within the medieval pottery assemblage indicates that the hilltop was occupied during a period spanning the medieval era. The earliest pottery analyzed is datable to the pre-Emiral period (7^th to the first third of the 8^th century AD) and the ‘paleo-Andalusian’ period (8^th to 9^th century AD). In contrast, the most recent pottery has a chronology that does not extend the first half of the 12^th century AD. This can be attributed to the North African phase, specifically the Almoravid period (Rosselló-Bordoy, 1991). The documented multilayered occupational sequence is of great importance for the knowledge of the medieval period in southwestern Iberia, as it is one of the few sites where material from the first two centuries of the Islamic presence in Al-Andalus has been found (Alba Calzado and Gutiérrez Lloret, 2008). Furthermore, it contributes to the documentation of the process of Islamisation of the Guadalquivir Valley. This is markedly distinct from the processes observed in Sharq Al-Andalus or in other areas of the Garb Al-Andalus, such as Mérida (Alba Calzado and Gutiérrez Lloret, 2008), or in the Seville area.