The interaction between large mammals and Acheulean tools during the Middle Pleistocene in the Manzanares valley (Madrid, Spain): new evidence for Santa Elena and Oxígeno sites

The fluvial deposits of the Manzanares and Jarama rivers present one of the largest concentrations of lithic and faunal remains of Pleistocene sites in Europe. In the Manzanares River close to the confluence of the Jarama River, the stepped terrace system disappears and gives way to the Complex Terrace of Butarque (CTB), where the sites of Santa Elena and Oxígeno are located. Different numerical dates obtained from the visible CTB’s bottom suggest that it was deposited during the MIS 6 or even MIS 7. This paper provides the first taphonomic and palaeoecological interpretation of both collections. A total of 445 fossil elements have been recorded in Oxígeno. The most represented are cranial fragments of Elephas sp. About Santa Elena, 130 fossil elements have been recorded. The best represented are fragments of Bos/Bison sp. horns and Elephas sp. tusks. The taphonomic processes identified at both sites bring to light a complex taphonomic history. The incidence of processes related to fluvial transport is higher in Oxígeno than in Santa Elena. Besides, dismemberment cut marks have been identified exclusively on the humerus of Bos primigenius in Santa Elena. Several tooth marks from carnivores have been recorded: small pits and scores at Oxígeno as well as pits and furrowing on Bos primigenius and Megaloceros sp. bones from Santa Elena. The faunal association depict an open landscape of wet meadows with herbaceous vegetation adjacent to riparian wood habitats associated with the Manzanares alluvial plain, where hominids and predators competed to some unknown extent.


Introduction
The lithic assemblages associated with the Acheulean technocomplex in the Iberian Peninsula have been recorded almost entirely in fluvial deposits where no faunal remains have been preserved, which hinders the study of an important part of the behaviour of the human groups that developed this technology in Europe.
The fluvial deposits of the rivers Manzanares and Jarama, located in the centre of the sedimentary basin of Madrid, in the Tagus Middle Basin-Province of Madrid-have sealed one of the largest concentrations of Acheulean sites in Europe, with lithic and faunal assemblages assigned to a chronological frame spanning from the Middle to the Late Pleistocene Soto 2000, 2002;Rubio-Jara et al. 2002; Rubio-Jara and Panera 2019; Yravedra et al. 2019b). The Acheulean of the Iberian Peninsula is characterised by a significant presence of macrotools, often knapped on large flakes, as well as by the preferential use of quartzite as a raw material, except in the Manzanares valley, where flint is virtually exclusively used in the lithic industry.
The Oxígeno and Santa Elena sites were discovered in the so-called Complex Terrace of Butarque (CTB), in the valley of the river Manzanares, chronologically framed between the MIS 7 and MIS 6. They were retrieved from archeopaleontological interventions carried out during the motoring and control of the exploitation of the Manzanares sand quarries during the 1950s and the 1960s (Rus and Querol 1981;Rubio-Jara et al. 2016). The lithic assemblages retrieved from these sites fit within the last evidence of the Acheulean technocomplex in Europe. Additionally, further evidence from these sites suggests that this technocomplex coexisted with Early Middle Palaeolithic lithic industries ). The coexistence from MIS 9, and perhaps from MIS 10, of Acheulean industries together with others defined as Middle Ancient Palaeolithic (European Ancient Middle Palaeolithic EAMP) has been detected in the Iberian Peninsula. These last probably evolved from a technological substratum of flake industries and cores such as those recorded in Lower Pleistocene sites in Orce (Granada), Atapuerca-Sima Elefante and TD6 in Atapuerca (Burgos) and Cueva Negra de Quípar (Murcia) (Santonja 2020). Currently, a good number of deposits with EAMP industries are known in the Iberian Peninsula with MIS 8/ MIS 9 chronologies both in the open air, for example, in Cuesta de la Bajada (Teruel), Ambrona (Soria) and Solana del Zamborino (Granada), as in cave, cases of Bolomor (Valencia), Gran Dolina-Atapuerca (Burgos) and Cueva de las Grajas (Málaga), which coexist with Acheulean assemblages clearly later than MIS 8, even up to MIS 6, such as Valdocarros (Madrid) or Torralba (Soria) Santonja et al. 2016). In the Manzanares valley, where there is evidence of Middle Palaeolithic industries during the MIS 5 and probably since the end of the Middle Pleistocene (Rubio-Jara 2011), sites such as Oxígeno and Santa Elena as well as other sites located in the interfluvium of the Manzanares and Jarama rivers (Bárez et al. 2010;Baena et al. 2015), with lithic assemblages that contain Acheulean elements such as handaxes, cleavers on flake and trihedral picks associated with a high percentage of retouched tools on flake, bifaces-tool support (Boëda 2001) and a stronger levallois presence, could be the result of the spatial-temporal relationship between the Acheulean and the EAMP in the region Santonja 2020).Together with thousands of lithic pieces, hundreds of faunal remains have been found in these sites, which enables the analysis of exploitation and processing of macromammals carried out by human groups with Acheulean technology. This type of analyses has been possible only in a limited number of open-air sites in the Iberian Peninsula (Villa et al. 2005;Yravedra 2010;Yravedra et al. 2010Yravedra et al. , 2012Yravedra et al. , 2014Yravedra et al. , 2019aYravedra et al. , b, c, 2021Baena-Escudero et al. 2014;Panera et al. 2015;Domínguez-Rodrígo et al. 2015;Rubio-Jara et al. 2016;Pineda et al. 2017;Pineda and Saladié 2019).
The nature of the deposits where faunal remains and lithic industries have been preserved has had a direct influence on the taphonomic processes that have affected the origin, preservation and condition of both assemblages. Thus, the taphonomic analysis contributes to the understanding of the biological, physical and anthropic agents involved in their formation (Yravedra et al. 2019b). The study of the faunal remains of Oxígeno and Santa Elena presented here constitutes the first taphonomic research for both collections on the basis of an exhaustive analysis of the cortical surfaces of the bone remains. Its aim was to identify any agents that may have left traces on them and to establish whether the remains were exploited, and to what extent, by human groups with Acheulean technology (Behrensmeyer 1978;Behrensmeyer and Hill 1980;Behrensmeyer et al. 1986;Lyman 1994a, b). This study also brings to light a complex taphonomic history when trying to understand these faunal accumulations associated with Acheulean industries. Preservation of cut marks on faunal remains linked to Acheulean assemblages is not frequent. Santa Elena provides new evidence of human intervention on megafauna, as dismemberment cut marks have been identified on a humerus of Bos primigenius. Together with other sites with similar evidence, this suggests that exploitation of megafauna by hominids in these valleys during the Middle and Late Pleistocene was a recurrent strategy for obtaining meat resources (animal protein). Additionally, several carnivore tooth marks have been recorded: small pits and scores at Oxígeno, as well as pits and furrowing on Bos primigenius and Megaloceros sp. bones at Santa Elena, suggesting that carnivores also interacted with the same fauna as the hominids. Both predators, carnivores and hominids, therefore, shared the same spaces in competition for animal resources. Evidence indicates that humans had primary access to the consumption of megafauna ) and that their environment exploitation patterns could have been defined by itinerant routes in a familiar landscape and by the recurrent use of specific places, for example, for the consumption of megafauna meat (Yravedra et al. 2019b). In this regard, the species identified in Oxígeno and Santa Elena reflect a mixed environment of open spaces, of wet meadows and of wooded riverside areas, associated with a fluvial environment, as is deduced from the associations of pollens, small mammals and herpetofauna identified in different sites of the river Manzanares area (Álvarez Catalán et al. 2009;Sesé et al. 2011a, b;Blain et al. 2013Blain et al. , 2019Laplana et al. 2015;Gil-García et al. 2019). At certain times of the year, mainly during drier seasons, those conditions could have favoured the concentration of fauna in this environment Yravedra et al. 2019b).

Geomorphological and chronological framework
The sites of Santa Elena and Oxígeno were located in SE Madrid (Villaverde), approximately 1.5 km from each other (Rus and Querol 1981;Gamazo 1982Gamazo , 2002Baena et al. 2002). The river Manzanares valley is located in the socalled Tagus Basin, in the South Sub-plateau of the Iberian Peninsula (Pérez-González 1994). The Manzanares, tributary of the river Jarama, is a low sinuosity river embedded in an asymmetrical valley in which two terrace systems can be distinguished (Pérez-González and Uribelarrea 2002). Up to 13 stepped terraces can be identified in the high and middle course (between + 4-5 and + 95 m) (Pérez-González 1994), no thicker than 6-7 m. The terraces were constructed by isostatic rising, together with blocks adjustment and lithological structured controls (Alia 1960;Pérez-González 1971, 1980, 1994Silva et al. 1988a, b), and date from the last third of the Lower Pleistocene to the Final Late Pleistocene (Rubio-Jara and Panera 2019). The chronostratigraphy of the Manzanares valley has been established based on the sequence of terraces with geomorphological criteria that allow establishing equivalences with the terraces of the rivers of the nearby valleys (Jarama and Tagus), complemented with numerical dating and biostratigraphic criteria from the known faunal associations (Moreno et al. 2019). Thus, the highest terrace (+ 60 m) may date to the end of the Lower Pleistocene, as the Matuyama Chron (1st) limit was detected in equivalent terraces of the Jarama and Tagus rivers, or in other rivers in the interior of the peninsula such as the Arlanzón, in the Duero basin, where this limit has been set by ESR at least 0.780 Ma. However, there are no numerical dates or biostratigraphic data for the higher terraces of + 52-54 m, + 44-46 m and + 35-40 m, although they must be between MIS 19 and MIS 12, since the terrace average of + 25-30 m can be attributed from the faunal associations to the MIS 11/9. On the other hand, we have AAR and ESR numerical dates on the equivalent terrace of the Jarama that refer to this chronology and IRSL numerical dates on the equivalent terrace of the Tagus River that dates it at MIS 9 or early MIS 8. We also do not have numerical dating or biostratigraphic data for the + 18-20-m terrace, although dating has been obtained on the equivalent terrace of the Jarama River by AAR with dates between MIS 7 and MIS 5 and by ESR that date it in MIS 6. Finally, on the lower terrace of the Manzanares (+ 11-12 m), two dates have been obtained by TL (40 + 2.7 ka and 40.2 + 4.6 ka) and another two by OSL on the + 8 m terrace, which date these deposits between 23.8 and 29.6 ka (Rubio-Jara and Panera 2019). In the low stretch of the river, downstream from Villaverde, and until it meets the river Jarama, the terrace system is complex (Goy et al. 1989) due to the underlying karst of evaporitic rocks that create synsedimentary overlapped terraces and affects the terraces + 25-30 m, + 18 + 20 m and + 12-15 m. This phenomenon gives rise to a thickening of the alluvial deposits (Uribelarrea 2008) and is known as the Complex Terrace of Butarque (Goy et al. 1989). The geographic identification of the sites Santa Elena and Oxígeno (Baena et al. 2002;Gamazo 2002), together with the georeferencing of published maps (Gamazo 1982), and the overlaying of orthophotos and the MAGNA (National Geological Maps) scale 1:50,000 from Madrid and Getafe have enabled an accurate location of their geomorphological position at the CTB (Rubio-Jara and Panera 2019) (Fig. 1).  Goy et al. 1989;Vegas et al. 1975) The deposits of the CTB, near the mouth of the Manzanares into the Jarama River, seem to be related to successive fining cycle characteristic of meandering rivers. This pattern has been observed in different outcrops in the lower course of the river Manzanares, at PRERESA, some 9 km away from the mouth of the Manzanares River into the Jarama (Carrillo et al. 1978;Arche 1983), in Perales del Río (14 km away) (Silva 2003) and Los Estragales, also 14 km away , where fluvial levels including channel gravel and sand deposits at the base, sealed with a layer of mud that represents flood plain facies, have been described. Most of these contain mammal fauna and lithic industry (Gaibar Puertas 1974). The CTB includes at least the T + 12-15 m and T + 18-20 m, and probably also T + 25-30 m of the river Manzanares, and in some places can reach up to 70 m in thickness (Uribelarrea 2008), although virtually only the top 15 m are accessible for visual inspection. In this stretch a timeframe ranging from MIS 7/6 to MIS 4 has been obtained with various techniques and procedures along the river Manzanares (Table 1). The series of numerical dates obtained, together with the association of micromammals identified in different sites, particularly the evolutionary stage of Microtus brecciensis, as well as the presence of Microtus arvalis (Laplana et al. 2015), suggests that the base of the CTB dates to the final MIS 8 or early MIS 7. Although this species turns up in the Middle Pleistocene in the northern area of the Iberian Peninsula, it only starts to be identified in the Manzanares and Jarama terraces records to the end of the Middle Pleistocene, as it happens in Arriaga (Sesé and López Martínez 2013) and H-02 (Laplana et al. 2015). This occurs simultaneously with a cold spell that would explain spreading to southernmost areas. This species lives currently in the northern half of the Peninsula, where it would have sheltered after the Holocene climatic improvement (Laplana et al. 2015). The TL dates are considered therefore to be a minimum age, when taking into account the AAR (Panera et al. 2011), the ESR dates (Moreno et al. 2019) and the Mediterranean warm and semihumid climatic conditions with dry summers suggested by the study of the herpetofauna (Blain et al. 2012).

Material and methods
Although the lithic series from Oxígeno (9432 pieces) is barely published (Rus and Querol 1981) and Santa Elena (2496 pieces) is unpublished, their industries have been described as Large Flake Acheulean on flint, and possibly Middle Palaeolithic industries . In this paper, a first approach to both lithic assemblages, based on the inventories preserved in the Museo Arqueológico Nacional (hereinafter MAN) and the data collected in published papers, is offered (Santonja and Querol 1980;Rus and Querol 1981).
Part of the findings retrieved from the archaeological works undertaken by different researchers in the Manzanares valley from the end of the nineteenth century and throughout the twentieth century (Quero 2002;Vega 2002, 2008;Rubio-Jara et al. 2002) was incorporated into the MAN collections (Cacho and Martos 2002). Thus, the fauna and lithic industries of the Oxígeno and Santa Elena sites, located in fluvial deposits of the Manzanares, entered the MAN in 1973 from the archaeological works carried out at the beginning of the 1950s by Julio Martínez Santa Olalla (Rus and Querol 1981;Quero 2002) (Goy et al. 1989;Santonja and Vega 2002). The urban growth of the city of Madrid since the middle of the twentieth century has eliminated or buried a large part of the deposits from the river Manzanares, which is why the collections deposited in museums such as the MAN are a fundamental element for studying human behaviour during the Pleistocene in the Manzanares valley. Recent works have enabled the discovery of the Oxígeno and Santa Elena sites, with fauna and lithic industries that remain largely unpublished, within the sequence of Pleistocene terraces of the river Manzanares (Rubio-Jara and Panera 2019). Its georeferencing, supported by biostratigraphic contextualization based on the faunal series, has provided a chronological context that allows its integration into current models of interpretation of the Acheulean technocomplex, both from a regional perspective, in the middle Tagus Basin, and from a larger scale within the framework of the Iberian Peninsula (Santonja et al. 2001(Santonja et al. , 2011Rubio-Jara et al. 2002; Rubio-Jara 2011; Santonja and Vega 2008; Rubio-Jara and Panera 2019).
In this paper, we have analysed a sample of 575 bone remains belonging to the Middle Pleistocene deposits of the Santa Elena (n = 130) and Oxígeno (n = 445) sites that are currently deposited in the MAN. The bone remains have been recorded according to the NISP (number of identified specimens); age patterns have also been analysed based on the tooth-wear patterns and the degree of epiphyseal fusion and tooth replacement, according to Barone (1976). After determining the species, the taphonomic analysis of the remains has been carried out using 10-20 × hand magnifiers, following the criteria of Blumenschine (1986). From the analysis of the bone surfaces, cut marks have been observed and identified according to Binford (1981), Shipman (1981) and Yravedra (2006). The identification of tooth marks has been carried out according to Binford (1981), Shipman (1981) and Blumenschine (1988). With regard to other alterations, weathering has been analysed according to Behrensmeyer (1978), taking into consideration hydric alterations, rounding and polishing of the bones, as well as chemical alterations, according to Fernández-Jalvo and Andrews (2017).

Lithic industry
In both series, the presence of flint is dominant (over 90%). The presence of quartzite, in which large cutting tools (LCTs) have generally been configured, except for the trihedrals, which are all made from flint, is residual. The most represented technological group in both series is that of flakes and cores (representing between 80 and 85% in each series), with a token presence of flakes and levallois cores (Table 2). Among the retouched tools, the best represented are the retouched flakes and scrapers and, to a lesser extent, denticulates. Simple scrapers are better represented than transverse, convergent, bifacial and double types.
In the group of LCTs, handaxes dominate, with a significant presence of cleavers on flake, while trihedral picks are practically non-existent from Santa Elena (Fig. 2).
In Oxígeno we have the study of a sample of LCTs from the collection deposited in the MAN (219 handaxes, 38 trihedrons and 21 cleavers) (Rus and Querol 1981). The handaxes, cleavers on flake and trihedral picks analysed do not show signs of intense fluvial alteration (except for one trihedral, one cleaver on flake and approximately 10% of the handaxes, all flint, which show abraded edges). Most of the handaxes were made from flint (95%) and rarely from quartzite. The most used support is flake on flint (n = 127; 58%) and occasionally quartzite (n = 7; 5.5% of the handaxes on flake).
Thick handaxes dominate (n = 195; 89%) with lanceolate-Micoquian, amygdaloid and protolimande silhouettes (according to Bordes 1961), while among the flat handaxes cordiform, discoid and oval types stand out. In general, the series consists of handaxes with bilateral and bifacial configuration, which occasionally show retouch and edges (9 handaxes were categorized as handaxe-scraper) conformed with soft hammer (use of soft hammer was identified on one third of the sample). The trihedrals, all flint, were also knapped mainly on flake (68%) and had thick and cortical butts. Soft hammer is rarely used for knapping trihedrals (this is obvious only in two instances). The use of soft hammer on the cleavers studied is not frequent either, although they tend to show very regular and symmetrical silhouettes, with type V being the dominant one over II, 0 and I (classification Tixier 1956;Inizan et al. 1999: pp. 55-57).

Faunal remains
The analysed faunal assemblages consist of 445 remains from Oxígeno (Fig. 3) and 130 remains from Santa Elena (Fig. 4) (Table 3). Although the species representation is similar to that of Oxígeno, the aurochs slightly outnumber the proboscideans, and the horses are less represented. As in Oxígeno, all individuals were adults.
The faunal association of both sites corresponds to the Middle Pleistocene, defined, above all, by the species Bos primigenius, Megaloceros matritensis, Palaeoloxodon antiquus and Stephanorhinus hemitoechus. The fauna is the characteristic of temperate climates, where open steppe-type environments predominated and developed in intercalations of closed and open areas, as suggested by the presence of rhinoceros, horse, Equus hydruntinus and proboscideans. Skeletal profiles are highly skewed due to cranial elements predominating in all taxa, which present worse preservation conditions than the appendicular elements, since some elements such as tusks are highly fragmented. Among the postcranial elements, the lower appendicular elements stand out, such as the metapodials, which appear complete in both sites, although other long bones such as humeri are also present.

Taphonomic processes
Despite the small size of the bone samples from Oxígeno (Fig. 5) and Santa Elena (Fig. 6), the incidence of various taphonomic processes has been observed at both sites. The state of bone surfaces shows regular conditions, with moderate or severe weathering, defined by the presence of cracking on the cortical surface. In Santa Elena, transverse and longitudinal marks overlapping the bone axis, with an open and superficial morphology, have been observed. This has been interpreted as weathering-the bones seem to show more severe weathering than those of Oxígeno, which would suggest a slower sedimentation process in Santa Elena. Along with weathering, other alterations have been recorded, such as trampling, which can easily be confused with cut marks. Among the different taphonomic processes that can be recorded on the bones, trampling marks are easily mistaken with cut marks because both have similar characteristics. Both are elongated marks with variable orientations and lengths, both can appear anywhere on the bone and both can be found in groups or isolated. Traditionally, cut marks were said to be V-shaped, and deeper than trampling marks can sometimes be V-shaped and can be deep. Although this is not usual, sometimes uncertain marks appear that can only  be differentiated by analysing the marks in detail, paying attention to elements such as the continuity/brokenness of the line or the presence/absence of micro-striations. This topic has generated an extensive bibliography for 40 years.
There are the works of Potts and Shipman (1981), Fiorillo (1984), Oliver (1984, Andrews and Cook (1985); Behrensmeyer et al. (1986) and Fisher (1995) who already raise the problem of trampling marks and how they affect the interpretation of cut marks and sites. In some cases, important debates have been generated about the correct identification of cut marks, since sometimes the authors discuss whether the marks are from cuts or trampling (Domínguez Rodrigo et al. 2011;Alcalá 2016 vs McPherron et al. 2010). Therefore, given the difficulty of correctly differentiating among cut or trampling marks, some authors have tried to generate methodological protocols that allow the two types of alterations to be correctly differentiated (Domínguez Rodrigo et al. 2009;Pineda et al. 2014;Courtenay et al. 2019Courtenay et al. , 2020. Hydric alterations have been recorded on both sites; however, the incidence is less pronounced in Santa Elena, where episodes of rounding have not occurred. On the contrary, in Oxígeno, the water action is more pronounced, giving rise to rounding and alterations of the bone surfaces. As for alterations of biological origin, the marks of biochemical origin produced by bacteria, fungi or insects, as well as those caused by vermiculation by roots, have been observed in both sites.
Regarding alterations of anthropogenic origin, only Santa Elena has preserved cut marks on several remains specifically on a humerus of Bos primigenius. This shows the connection between the lithic industry recorded on the site to the fauna. However, the sample is small, and the scarcity of postcranial elements has prevented more cut and percussion marks from being recorded. In the case of Oxígeno, on the one hand, the poor preservation of the bone surfaces and the rounding of the bones may have masked the anthropic traces. On the other

Lithic technology
The sites of Santa Elena and Oxígeno are located on fluvial deposits of the Complex Terrace of Butarque (CTB) of the river Manzanares, whose base has been chronologically framed between MIS 7 and MIS 6 (Rubio-Jara and Panera 2019). The collections, retrieved in the early 1950s (Rus and Querol 1981), have provided lithic assemblages with a very high number of elements associated with fauna, an exceptional circumstance in the whole of the Middle Pleistocene archaeo-paleontological record of the Iberian Peninsula. The materials from the Santa Elena and Oxígeno sites have remained practically unpublished and lacked a precise geomorphological context until now. Its location, based on the georeferentiation and the biostratigraphic contextualization of its faunal series, in the so-called CTB, within the sequence of the fluvial terraces of the river Manzanares, has allowed the integration of both sites into the current interpretation model of the Acheulean in the Middle Tagus Basin, in the interior of the Iberian Peninsula. In the current geochronological frame, the earliest Acheulean evidence in the region could date from sometime between MIS 15 and MIS 13. Although no excavated site has been dated earlier than MIS 13, infrequent finds of small assemblages of lithic industry have been located in some middle and high terraces of the river Tagus (Santonja and Pérez-González 1997, 2000-2001Rodríguez de Tembleque et al. 2010). Additionally, lithic industry within a stratigraphic context and showing clear Acheulean features has been recorded in different locations in terrace levels of the river Tagus. They probably date from MIS 15/MIS 12, for example, Hornaguera and Paridera (Santonja and Pérez-González 1997; Rodríguez de Tembleque 2008), Buenavista inferior and Salchicha inferior (Alférez 1977;Santonja 1981) or Gébalo (Rodríguez de Tembleque 2008). In all cases, quartzite is the predominant material used where, together with flakes and cores, some LCTs appear, such as choppers, cleavers, trihedrals and handaxes. Puente Pino, located in an intermediate position between two middle terraces of the river Tagus, could actually be the oldest Acheulean site excavated in the Tagus Basin (Rodríguez de Tembleque 2004, 2008. It can be dated to just before the expansion of the Acheulean technocomplex in the area, around MIS 11/MIS 9, in accordance with the numerous lithic assemblages found in stratigraphic contexts and the sites excavated throughout the Tagus Basin. Sites such as Pinedo (Querol and Santonja 1979), Cañete Bajo (Santonja and Pérez-González 1997, 2000-2001Rodríguez de Tembleque 2008), Cañete Alto (Rodríguez de Tembleque 2008), Cien Fanegas (Baena et al. 2010) in the Tagus valley, Áridos 1 and 2  in the Jarama valley, San Isidro (Santonja 1977) and Transfesa/Tafesa (Baena and Baquedano 2010) in the Manzanares valley have provided industries with significant percentages of macrotools, among which the use of large flakes for shaping bifaces is common, together with retouched tools barely elaborated and scant levallois presence. Handaxes, often made on flake, are the most common bifaces and tend to be poorly elaborated, with nonregularized or non-retouched edges, while trihedral picks are scarcely represented. Cleavers on flake (0 and I Tixier's types mainly) are also present. The latest evidence of the Acheulean technocomplex in the Tagus Basin dates as late as the end of the Middle Pleistocene MIS 8 to MIS 6 in sites such as Las Acacias (Santonja and Querol 1977;Santonja 1981) and Valdocarros (Blain et al. 2012) in the Jarama valley, Santa Elena and Oxígeno (Rus and Querol 1981) in the Manzanares valley or Los Ahijones and El Cañaveral, found on the dividing platform along the Manzanares-Jarama river valleys (Bárez et al. 2010;Baena et al. 2015). A significant presence of more elaborately made handaxes than those of previous dates has been detected among the macrotools of these deposits, often with edges regularized by retouching, as well as complex retouched tools on flake. In Santa Elena and Oxígeno also, levallois cores and débitage have been described. These assemblages may correspond to the interrelation between Acheulean and Early Middle Palaeolithic .
Therefore, the sites Santa Elena and Oxígeno fit into the last beats of the Acheulean in the Iberian Peninsula. Fluvial loads from the rivers Manzanares and Jarama have been influential in the choice of raw materials for manufacturing these industries, but not decisive. For example, in the Tagus and Jarama valleys, where there is plenty of quartzite, this raw material prevails for knapping macrotools, whereas in the Manzanares valley, flint predominates and quartzite is absent. Nevertheless, in this last valley, manufacturing of bifaces on flake is frequent, and in the Jarama valley, flint pieces are more frequent as retouched tools, despite being a rare raw material. The lithic industries of Santa Elena and Oxígeno do not show signs of intense fluvial alterations, due to the presence of flint hegemonic. This last issue is an anomaly in the Acheulean technocomplex in the Iberian Peninsula, where the macrotooling is represented mainly in quartzite. The almost exclusive use of flint in the Manzanares valley to produce macrotools, mainly from the configuration of large flakes, could initially be related to availability of the raw material in the environment, but other factors such as functionality could have played a part. The fluvial loads of the river Manzanares in the area where the sites Santa Elena and Oxígeno were located consist mainly of quartz, which has sporadic representation in the lithic assemblages of both sites, apart from flint in gravels and pebbles bars, whose diameters are seldom over 7-8 cm in length (Goy et al. 1989). The flint used in the elaboration of the macrotools of these sites could come from the Miocene outcrops located in the interfluve of the Manzanares and Jarama rivers (Rus and Enamorado 1991;Baena et al. 2015;Bárez et al. 2016;Yravedra et al. 2019b), situated some 10 km to the northeast, where several Acheulean sites associated with flint exploitation, production of large flake blanks and preconfiguration of bifacial products in contexts related to MIS 7/MIS 6 have been excavated (Baena et al. 2015;Bárez et al. 2016). In the same way, the quartzite, although residual, since it was preferably used for macrotools, would have been introduced from outside the Manzanares valley, perhaps from the Jarama River valley, to which the Manzanares is a tributary, since at the fluvial loads, quartzite is abundant. Additionally, in the sites along the river Jarama, there is a greater use of flint in retouched tools, which is actually absent in the fluvial loads .
The presence of all technological categories in the Santa Elena and Oxígeno assemblages points to a complete set. The chaînes opératoires of the débitage dominate the lithic assemblages, with flakes and cores better represented than retouched tools and LCTs. Handaxes with bilateral and bifacial configuration, which occasionally show retouch and edges shaped with soft hammer, and among which amygdaloidal silhouettes predominate, are outstanding in both Santa Elena and Oxígeno. Cleavers on flake, also present in both series, are mostly shaped with bifacial retouch (type V), whereas trihedrals are sporadic in the Santa Elena series. Retouched elements are more common in Oxígeno than in Santa Elena, of which scrapers are the best represented. The existence of levallois elements-flakes and cores-and the presence of handaxes with silhouettes regularized by retouch or retouched edges, containing biface-tool support (according to Boëda 2001), suggest a scenario of coexistence of Acheulean industries.
This lasted up to the end of the Middle Pleistocene, together with Early Middle Palaeolithic type industries in the middle Tagus Basin, which has already been suggested by other researchers in different regions of the Iberian Peninsula Santonja 2020;Ferreira et al. 2021).

Human-carnivore interactions
The finds from Santa Elena and Oxígeno have not been retrieved from a systematic excavation, and therefore, the available information is limited. Nevertheless, both sites are part of the short list of Middle Pleistocene open air sites in the Iberian Peninsula which have fauna associated with lithic industries, among which sites in the Manzanares Rubio-Jara and Panera 2019;Yravedra et al. 2019c) and Jarama valleys are included , the Orce region Yravedra et al. 2021), and to a lesser extent the Pleistocene terraces of the river Guadalquivir (Caro et al. 2004;Baena-Escudero and Guerrero 2011;Baena-Escudero et al. 2013 in the south of the Iberian Peninsula, as well as some sites scattered in the peninsular interior, such as Torralba and Ambrona (Villa et al. 2005;Domínguez-Rodrígo et al. 2015;Rubio-Jara et al. 2016;Santonja et al. 2014b;Pineda et al. 2017;Santonja et al. 2018;Pineda and Saladié 2019). It is even rarer to record a relationship between lithic industry and fauna associated with anthropogenic alterations, in the form of green fractures, percussion or cut marks on bones, as in the case of Santa Elena. Without these signs of anthropic alteration, it is difficult to discriminate whether there is a direct relationship and what degree of association is present or if they belong to completely independent events (Bunn et al. 2010;Yravedra et al. 2014;Haynes 2022;Haynes et al. 2020). In the Iberian Peninsula, cut and percussion marks or green fractures have been recorded in different open air sites, such as Torralba and Ambrona (Shipman and Rose 1983;Villa 1990;Villa et al. 2005; Pineda and Saladié 2019), Solana del Zamborino (Díez 1992) or Pinedo (Yravedra personal observation) and also in the Middle Tagus Basin in Áridos Villa 1990;Yravedra et al. 2010), Tafesa (Yravedra 2010) and Valdocarros (Yravedra in preparation) in the Jarama valley and ETB-H02 (Yravedra et al. 2019b), PRERESA (Yravedra et al. 2012) and Arriaga (Yravedra et al. 2019b) in the Manzanares valley. Cuts and percussion marks and green fractures have also been recorded in cave sites such as Galería (Diez et al. 1999) and TD10 (Rodríguez Hidalgo et al. 2017) at the karst complex of Atapuerca, and in southern areas of the Iberian Peninsula, as in Cueva del Ángel (García Solano 2014).
A great deal of these anthropogenic alterations on bones is related to megafauna, mainly proboscideans. At the Barranc de la Boella site, located to the northeast of the Iberian Peninsula, and dated between 0.78 and 0.99 Ma, possible cut marks on two rib diaphyses of Mammuthus meridionalis have been identified (Pineda et al. 2017). In Ambrona, located in the Northern Meseta of the Iberian Peninsula, and dated to MIS 11/MIS 9, as well as in Torralba, located near Ambrona and dated to MIS 7, butchering marks produced by hominids on elephant bones (Palaeoloxodon antiquus) have been recorded, although they are rare in the whole assemblage of the recorded fauna (Shipman and Rose 1983;Villa 1990;Pineda and Saladié 2019). In the Middle Basin of the river Tagus, in Áridos 2, at the Jarama valley, MIS 11/ MIS 9, cut marks (defleshing and evisceration) were identified on a scapular blade and another on the ventral side of one rib of an old Palaeoloxodon antiquus bull. They were probably made with a handaxe ). In the Tafesa site, also in the Jarama valley, dated to MIS 11-MIS 12, cut marks were recorded on long bone shafts and axial fragments of Palaeoloxodon antiquus (Yravedra 2010). In the river Manzanares valley, at the CTB, dating from MIS 6 to MIS 5, cut marks and green fractures (for extraction of marrow) on Palaeoloxodon antiquus bones were identified in Arriaga IIa (Rus and Vega 1984;Panera et al. 2015), PRERESA (Yravedra et al. 2012) and also on a mammoth (Mammuthus intermedius or Mammuthus primigenius) in EDAR Culebro 1 (Yravedra et al. 2014).
Evidence of anthropic intervention on other species is scarcer. In Vallparadís, to the northeast of the Iberian Peninsula and dated to over 0.8 Ma, remains of rhinoceros and hippopotamus with cut marks have been described (Martínez et al. 2010); and in Cuesta de la Bajada, also in the northeast and dated to MIS 9-MIS 8, the same marks were recorded on equid and cervid bones (Santonja et al. 2014a;Domínguez-Rodrigo et al. 2015). In the Middle Basin of the river Tagus, cut and/or percussion marks have been recorded on the remains of rhino, large bovine, deer, horse and even beaver (for example, in Áridos 1, MIS 11-MIS 9) (Yravedra personal comment). In Tafesa, MIS 11-MIS 9, some cut and percussion marks were identified on cervid and bovine bones (Yravedra 2010). Cut marks have also been observed on horse and rhino bones in Pedro Jaro (Yravedra et al. 2019c), located in a middle terrace of the river Manzanares, dated to MIS 11-MIS 9. Some cut marks and green fractures were identified on the bones of a large bovine, probably auroch, in PRERESA, in the CTB of the Manzanares (Yravedra et al. 2012), and rare cut marks in ETH-H02 (also in the CTB) on equids, bovines and cervids associated with defleshing and disarticulation processes; also skinning marks were identified on a skull (Yravedra et al. 2019a). Now we have to add to this collection the marks from Santa Elena, where dismemberment cut marks on a humerus of Bos primigenius were identified. These are the only cut marks identified in the fauna of Santa Elena, and although no cut marks were identified in Oxígeno, human activity cannot be ruled out, as the scarcity of postcranial elements and the poor preservation of the bone surfaces may have made anthropic alterations invisible on the bones of both sites. In this sense, the large presence of remains of proboscideans in Santa Elena and Oxígeno and the well-documented exploitation of these animals in the area through green fractures and cut and percussion marks on proboscidean bones in the sites of Áridos 2 (Villa 1990;Yravedra et al. 2010), PRERESA (Yravedra et al. 2012), Arriaga Yravedra et al. 2019b), EDAR Culebro 1 (Yravedra et al. 2014) and Tafesa (Yravedra 2010) would imply that similar exploitation was carried out in Santa Elena and Oxígeno. Cut marks recorded on the bones of other taxa in different sites of the Jarama and Manzanares valleys Rubio-Jara et al. 2016;Yravedra et al. 2019b), such as Valdocarros (Yravedra personal observation), Tafesa (Yravedra 2010), ETB-H02 (Yravedra et al. 2019b), Pedro Jaro and also Santa Elena, where rhinoceros, aurochs, bison, deer and horse show evidence of anthropic exploitation, together with other small animals like the beaver from Áridos 1 , suggest a broader exploitation of taxa in Santa Elena and Oxígeno, such as that already recorded on the large bovine in Santa Elena. Additionally, Santa Elena and Oxígeno also provide evidence of carnivore intervention on the fauna. Examples of sites where the alternating activity of carnivores and humans on the same assemblages can be recorded are not frequent in the fossil record of the Iberian Peninsula. Scenarios of competition for carcasses have been identified in Fuente Nueva-3, a site located to the southeast of the Iberian Peninsula dated to 1.07-0.98 Ma, where hyenas (Pachycrocuta brevirostris) scavenged on the remains of the carcass after hominids had dealt with it Yravedra et al. 2021), and also, in Barranc de la Boella, in the northeast of the Iberian Peninsula, where cut and carnivore marks have been recorded on the same fauna (Pineda et al. 2017). In the Pleistocene sites of the Jarama and Manzanares valleys, in ETB-H02, some cut marks have also been recorded on the bones of equids, bovines and cervids associated with defleshing and disarticulation processes, along with others produced by small carnivores. This has been interpreted that humans and carnivores succeeded each other in different events in the occupation of the site space (Yravedra et al. 2019a). In EDAR Culebro 1, where anthropic marks have been recorded on mammoth bones, tooth marks on bones of rhinoceros were also recorded (Yravedra et al. 2019b). Similarly, in Tafesa, bones modified by small carnivores were identified within a basically anthropic fossil accumulation (Yravedra 2010). In Áridos 2, on the same elephant remains in which cut marks associated with defleshing and evisceration were recorded, tooth marks, intense furrowing on a distal epiphysis of the humerus and chewing on one rib end were caused by hyena . The sites of Santa Elena and Oxígeno provide new evidence of carnivore intervention on fauna accumulations in which there has also been human intervention. From the presence of human and carnivorous activity on the same type of prey detected in these sites, relevant conclusions are derived, since they may have conditioned the behaviour of human populations of the Middle Pleistocene. First, carnivores and humans shared the same spaces, succeeding each other in their wanderings through the landscape without interacting with each other. Second, according to the evidence recorded in some sites such as ETB-H02 (Yravedra et al. 2019a), Tafesa (Yravedra 2010), Áridos 2 , PRERESA (Yravedra et al. 2019c) and Valdocarros (Yravedra et al. in preparation), humans gained access to meat resources prior to carnivores. Third, the alternation between humans and carnivores in the same spaces indicates that the former did not remain for long periods of time in the places where they took advantage of these resources, although they probably visited them on a recurring basis as can be deduced from the variety of species represented in Oxígeno, Santa Elena (this paper), in ETB-H02 (Yravedra et al. 2019a), in Tafesa (Yravedra 2010) and in Valdocarros (Yravedra et al. in preparation).

Palaeoecology
All this evidence, considered in its regional context, suggests that Middle Pleistocene hominids carried out recurrent activity in the fluvial environments of the Manzanares and Jarama valleys. In this sense, the palaeoecological implications derived from the fossil faunas of Santa Elena and Oxígeno are also relevant.  (Garutt 1972, Garutt 1986aGarutt and Vangengeim 1982;Pushkina 2007). Lastly, Cervus elaphus is commonly associated with denser forest type and more humid environments, although several authors have discussed whether this species could also be found in open forests or forests associated with steppes (Murie 1951;Flerov 1960). Generally, the association of bovines, equids, rhinos and proboscideans in the same group is usually related to cold events, although macroinvertebrates are a generalist group with a great capacity to adapt to different environments (Sala 1987). The presence of Mammuthus in Santa Elena and Oxígeno, and of Mammuthus primigenius in EDAR Culebro (Álvarez-Lao and García 2011), also in the CTB, would suggest cold conditions, although this probably relates to the last phase of this deposit, dated to MIS 6. The environmental variability inferred from the Santa Elena and Oxígeno faunas suggests a mixed and/or alternating environment, with a predominance of open areas of herbaceous vegetation, together with more closed and isolated areas. In the case of Oxígeno and Santa Elena, this mixed landscape may respond to the proximity of the river Manzanares, with areas of humid meadow next to a riparian forest that is nourished by the direct action of the river. This hypothesis is supported by the presence of different species that predominate in different environments or that can alternate between open and closed environments, and vice versa. This is the type of landscape suggested by the pollen analyses carried out in different points of the CTB, pine forest and local forest with Mediterranean taxa ( (Sesé et al. 2011a, b) and the herpetofauna, with toads (Pelobates cultripes, Bufo bufo, Bufo calamita), frogs (Pelophylax perezi), turtles, lacertid (Timon lepidus, Psamamodromus cf. algirus), scincid (Chalcides striatus, Chalcides sp.), lizards and snakes (Natrix maura, Coronella girondica) (Blain et al. 2012(Blain et al. , 2013(Blain et al. , 2019. A mostly open landscape was dominated by wet meadows with herbaceous or shrubby vegetation, interspersed with drier and shrubby areas, and also some forest; a mixed habitat, mainly open and partially wooded (riverside forests), was favoured by humid conditions due to its proximity to a river but also typical of a Mediterranean-type climate with alternating periods of aridity during the summer until the beginning of the autumn, identical to or even more temperate than the current one, and with abundant rains in winter and spring, characterised by less continentality and lower winter temperatures than nowadays. In this scenario, the riverside ecosystems, such as that of the Manzanares valley, could have played an important role, not only for the development of plants that could provide food for the Pleistocene hominids, but also because during the dry summer months, they could have acted as an ecological corridor in a landscape where the saline substrate would have limited the development of plant species , although we have pointed out that Megaloceros matritensis present in Santa Elena and Oxígeno shows a dentition adapted to xerophytic vegetation. The existence of these corridors would favour the concentration of fauna along the riverbank of the Manzanares and would facilitate hominids access to hunt or scavenge the fauna recorded from the sites.

Conclusions
This study reveals the great potential of the finds preserved in museums such as the MAN, which come from excavations not carried out with current scientific parameters. Despite these limitations, they contribute to the knowledge of the behavioural dynamics of hominids during the Pleistocene, especially if, as in the case of Santa Elena and Oxígeno, they come from deposits that have already disappeared or are inaccessible due to the urban development of the city of Madrid. Santa Elena and Oxígeno add to the rare group of sites with lithic industries associated with Middle Pleistocene fauna. The presence of some cut and other marks related to the activity of carnivores on the fauna of Santa Elena and Oxígeno allows both collections to be related to the activities of hominids and carnivores, with some degree of competition, and in the same landscape. The existence of this type of alternation carried out by hominids and carnivores in different sites of the valleys of the Manzanares and Jarama reinforces the theory that the consumption of megafauna (proboscideans, rhinoceros, large bovines) was more than a marginal strategy of protein obtainment by hominids, and that it was a frequent strategy over time, maintained throughout the Middle and Late Middle Pleistocene, with carnivores playing a secondary role in accessing carcasses previously exploited by hominids. The association of taxa defined in Santa Elena and Oxígeno, together with the pollen data and the palaeoecological requirements of small mammals and herpetofauna collected in different sites of the CTB, suggests a mixed landscape of open spaces of more or less humid meadows and wooded areas of riparian forest within a Mediterranean climate similar to the current one with variation of temperatures and a seasonal rainfall regime. These conditions and the geological characteristics of the Manzanares valley in the surroundings of the Santa Elena and Oxígeno sites could have conditioned the alluvial plain to become an ecological corridor that favoured the concentration of fauna and then exploited by the hominids. Then again, the synsedimentary process that affected the formation of the CTB, and gave rise to the thickening of the terrace, could have preserved fauna and industries from different times (from the Late Middle Pleistocene to the Early Late Pleistocene) together in the same deposits. It is in this sense that the presence in both sites of remains of Mammuthus sp. together with Palaeoloxodon antiquus and lithic industries attributed to the Acheulean and Early Middle Palaeolithic technocomplexes should be assessed.
Funding Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. This work has been carried out in the framework of Grants PGC2018-093612-B-100, funded by MICIN/ AEI/10.13039/501100011033 and ERDF; CEN154P20 and co-financed by the ERDF 728 (European Regional Development Fund) and the Junta de Castilla y León.

Declarations
We would like to thank the National Archaeological Museum (MAN) for providing us with the facilities and technical equipment necessary for our study. All authors agreed to participate in the Archaeological and Anthropological Sciences Journal and present a new study that has not been published before. All authors read the journal policies of the Archaeological and Anthropological Sciences Journal about publishing and agreed to publish our article. All authors have worked together to reproduce a new-fangled work.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.