In the last decade, a number of specialist analyses were made on the material from old excavations of Mesolithic-Neolithic sites in the Danube Gorges. These new results altered significantly our understanding of the Lepenski Vir culture. The question of chronology of this regional phenomenon has been acute since the discovery of Lepenski Vir in the 1960s, and it remains of key importance for understanding the character of Mesolithic-Neolithic transformations in this and the neighbouring regions. The most heated debate was fuelled by the initial stratigraphic and chronological attribution of the type-site itself. There remained the question about the adequate dating of the most prominent phase at this site characterized by buildings with trapezoidal bases covered with limestone floors and with rectangular stone-lined hearths placed in the centre of these features. There have been suggestions that these features also contain Early Neolithic Starčevo type pottery and other similar items of material culture and should thus be dated to the Early Neolithic historical context. Moreover, the first series of conventional radiocarbon determinations (21 dates) also suggested that the absolute chronology of these features should be confined to the period from around 6400-5500 cal BC (Fig. 1). Due to the importance of defining more precisely the chronology for the start of construction of these particular features at Lepenski Vir and for establishing the life-span of these buildings and their associated material culture, we have AMS (Accelerator Mass Spectrometry) dated a number of contexts from this site. The results are presented in this paper. The project was made possible through the Oxford Radiocarbon Accelerate Dating Service (ORADS) programme funded by the Arts and Humanities Research Council (AHRC) and the Natural Environment Research Council (NERC) of the Great Britain. Apart from those dates presented in this paper, there are 29 previously published radiometric dates from Lepenski Vir 21 made by the conventional 14C dating of charcoal and 8 AMS dates made on animal (3 dates) and human bones (5 dates). There is also a series of 20 new AMS dates made on human bones from Lepenski Vir. The last group of dates has not been published with all the contextual details and are of limited use in our analyses of absolute chronology and stratigraphy of Lepenski Vir. New dates are listed in Table 1. From 32 dated contexts from Lepenski Vir, 27 contexts are stratigraphically related to trapezoidal buildings while 5 dates are connected with the area outside of buildings. From those contexts related to trapezoidal structures, 24 contexts are dated with animal and 3 with human bones. The emphasis on the dating of animal bones is connected with problems of precision when dating samples made on human and dog bones due to the reservoir effect and the deposition of 'old' carbon. Dated remains of animal and human bones originate from the following types of stratigraphic contexts a) beneath building floors, i.e. stratigraphically the oldest contexts in the settlement (2 dates) (Fig. 2); b) between two superposed floors of trapezoidal buildings as 'sealed' contexts (8 dates) (Fig. 3); c) lying directly on top of the floors of trapezoidal buildings but not overlapped by a later floor (17 dates) (Fig. 3); d) outside of trapezoidal buildings, found in contexts such as pits, domed ovens, and burials, or in contexts that can be attributed to the occupation layer only (6 dates) (Fig. 4). The new dates indicate a very long duration of the Mesolithic period, from around 9400 to around 7500 cal BC (Fig. 2, 23). These early dates are concentrated in two particular periods that may point to two separate phases within these two millennia, with settlement discontinuities. It remains possible that there were many more occupation episodes that these dates do not encompass, and more AMS dates may indicate whether these two groupings with three dates per grouping are meaningful and representative of two separate and defined phases of occupation at this locale. This early period would correspond with the phase that the excavator of Lepenski Vir defined as Proto-Lepenski Vir although his ideas about the spatial distribution of this phase, its interpretation, duration and relation to the later phase of trapezoidal buildings must be revised in the light of new AMS dates and other available data. The phase with trapezoidal buildings most likely starts only around 6200 cal BC and most of the trapezoidal buildings might have been abandoned by around 5900 cal BC. The absolute span of only two or three hundred years and likely even less, for the flourishing of building activity related to trapezoidal structures at Lepenski Vir significantly compresses Srejović's phase I. Thus, it is difficult to maintain the excavator's five subphases which, similarly to Ivana Radovanović's more recent re-phasing of Lepenski Vir into I-1-3, remain largely guess works before more extensive and systematic dating of each building is accomplished along with statistical modeling in order to narrow the magnitude of error. On the whole, new dates from these contexts better correspond with Srejović's stratigraphic logic of sequencing buildings to particular phases on the basis of their superimposing and cutting than with Radovanović's stylistic logic, i.e. her typology of hearth forms, ash-places, entrance platforms, and presence/absence of -supports around rectangular hearths used as reliable chronological indicators. The short chronological span for phase I also suggests that phase Lepenski Vir II is not realistic. This has already been shown by overlapping plans of the phase I buildings and stone outlines that the excavator of the site attributed to Lepenski Vir II phase. According to Srejović, Lepenski Vir phase II was characterized by buildings with stone walls made in the shape of trapezes, repeating the outline of supposedly earlier limestone floors of his phase I. However, the trapezoidal buildings must be envisioned as dug-in features with their rear, narrow side dug deep into the slope since these features were dug into the sloping terrace where the site is situated. It is more likely that these stone constructions assigned to a separate phase were part of the same trapezoidal buildings with limestone floors assigned by the excavator to phase I. Thus, vertical stone walls existed on the level above limestone floors, built in dry wall technique around buildings' floors and cuts. The visual overlap of phases I and II clearly shows the match between these stone constructions and trapezoidal limestone floors (Fig. 27). Even on the published section of the western part of the settlement of Lepenski Vir which runs through the backs of Houses 43, 34, 27, 20, 33 and 32, phase II is not marked (see Fig. 28), which might lend further support to our conclusion about its elusive character. Furthermore, no activity areas were reported with regard to the 'floor' level of these structures, with the exception of the largest building at the site, XLIV. Therefore, trapezoidal stone walls previously attributed to phase II were part of the same phase I buildings. Henceforth we suggest treating Srejović's phases I and II as a single phase and we refer to this building horizon as Lepenski Vir I-II (see Table 2). The new dating programme also suggests no temporal break between phases Lepenski Vir I-II and phase III. The dates indicate that Srejović was right to separate the latter as it seems that most of the trapezoidal buildings were abandoned by 5900 cal BC and that new and different occupation pattern commenced at the site in the period following 5900 cal BC. Yet, some of the dates indicate that, at the current resolution of the chronological scale there could have been some overlapping between the use of particular trapezoidal buildings, perhaps primarily for the interment of human burials (e.g. House 21 and Burials 7/I and II, see Fig. 11), and the new types of contexts that appear around 5900 cal BC. These new contexts included a number of pits, dug primarily in the rear area of the site, outside of the zone with trapezoidal buildings. There are also several domed ovens the function of which remains unclear. Also, crouched inhumations became the dominant burial rite (of possibly migrant individuals) during this phase. Some of these crouched burials were found lying on the floors of trapezoidal buildings. This seems to have been the time of significant changes in patterns of habitation of the Lepenski Vir community. Bones of domestic animals were also found in those features assigned to phase III. We have directly dated four samples that come from domestic animals (sheep/goat, cattle and pig) found in these contexts at Lepenski Vir. In this way, it was possible to directly date the introduction of domestic animals to the site. The results suggest that the these domestic animals must have been introduced to Lepenski Vir in the post-5900 cal BC period. The upper limit of this Middle Neolithic phase Lepenski Vir III remains to be elucidated further. While previous charcoal dates indicated that the site was used up until around 5400 cal BC, the upper limit of new dates is around 5700 cal BC. Existing dates from the neighbouring site of Padina suggest that some trapezoidal buildings at that site (Houses 15 and 18) might have been used up until 5500 cal BC. It is possible that future dates will move this limit to the end of the Middle Neolithic, i.e. c. 5500 cal BC. At this time, previously occupied sites on the Danube, such as Lepenski Vir, Padina, Vlasac, etc., were abandoned for more than a millennium. Lepenski Vir is used again during the Eneolithic period, when a burial of the Salkuţa culture, AMS dated to around 4300 cal BC, was interred here (see Burial 2, Fig. 25). New radiometric dates from Lepenski Vir, together with all other newly available data, demand a revision of conclusions previously made with regard to the absolute chronology of particular phases as well as stratigraphic attribution of certain contexts. Such a revision inevitably leads us to suggest a new stratigraphic division and phasing of Lepenski Vir (see Table 2). This revised phasing largely keeps the old nomenclature of the excavator of Lepenski Vir. We would like to avoid confusions and complications in suggesting completely new labels for particular phases when there is no need for such a radical break from the original understanding of the site's stratigraphy. We are aware that the future dating of Lepenski Vir may affect certain elements of our conclusions and that the suggested changes thus remain tentative. Yet, at present, our conclusions are firmly grounded in the available data.
