Neolithic\Eneolithic settlement patterns and Holocene environmental changes in Bela Krajina (south-eastern Slovenia)

Bela krajina is a lowland karst region in south-eastern Slovenia. It is bounded on the east and south by the river Kolpa and on the west by the high dinaric uplands of Ko≠evski rog. It is separated from central Slovenia by the Gorjanci hills. The lowland karst plateau has limited surface water, particularly rivers and streams in deeply incised valleys or gorges. The karst itself is characterised by a range of karst features, such as uvalas and swallow holes, but there are extensive areas of Pleistocene deposits in the Kolpa valley and to a lesser extent in the catchments of the river Lahinja (Plut 1985.13–15; Radov≠i≤, Galovi≤ 2002.10–31).


Introduction
Bela krajina is a lowland karst region in south-eastern Slovenia.It is bounded on the east and south by the river Kolpa and on the west by the high dinaric uplands of Ko≠evski rog.It is separated from central Slovenia by the Gorjanci hills.The lowland karst plateau has limited surface water, particularly rivers and streams in deeply incised valleys or gorges.The karst itself is characterised by a range of karst features, such as uvalas and swallow holes, but there are extensive areas of Pleistocene deposits in the Kolpa valley and to a lesser extent in the catchments of the river Lahinja (Plut 1985.13-15;Radov-≠i≤, Galovi≤ 2002.10-31).
The region has a relatively long tradition as an area of research into the Neolithic and Eneolithic, due to ABSTRACT -This paper examines the archaeological settlement pattern and vegetation history of Bela krajina region of Slovenia in order to better understand the interaction of human activities and environmental processes in the landscape.Pollen record of two small palaeoecological sites (Mlaka and Griblje) indicates that human impact on the vegetation at circa 4150 calBC was intensive (forest cutting/burning, beech decline and formation of fields, pastures, meadows) and can be associated with numerous Neolithic/Eneolithic sites, located in the Lahinja river basin and the Kolpa lowlands.Human pressure on the (lowland/riverine) environment slightly decreased between c. 3750-2850 calBC.This coincides with the appearance of a more dispersed settlement pattern, including the formation of short-term settlement/activity areas on the karst plateau.This change to a more extensive Eneolithic settlement pattern can be presumably associated with change in economy (more intensive pastoralism and transhumance, possibly also soil erosion) and is partially borne out by evidence from excavated sites in the area.

Vegetation history and human impact on environment
To date palaeoecological research in the area focused on studies of vegetation development in lowland Bela krajina.The vegetation history of the karst plateau was not investigated, therefore only lowland vegetation can be presented and compared with the archaeological settlement pattern (Pl.1).
Palynological research at Mlaka and Griblje wetlands (Figs. 1, 2) showed that, on the local scale, the Holocene vegetation development was very dynamic, with significant human impact on the environment (Andri≠ 2007).This became apparent due to natural characteristics of selected study sites, which are small (with small relevant source area of pollen, Sugita 1994) and therefore sensitive to local vegetation changes and human impact on the environment.Both study sites are located in the vicinity of Neolithic/Eneolithic settlements (Figs. 3, 4; R∫i∏≠a, Pusti Gradac, Griblje).
The results of pollen analysis at Mlaka (Fig. 1, Tab. 2) suggest that at c. 6900 calBC the early Holocene open woodland of oak (Quercus), hazel (Corylus), lime (Tilia), birch (Betula) and pine (Pinus) was suddenly replaced by beech (Fagus) forest.This is presumably associated with an increase in precipitation; a similar spread of beech also occurred at Griblje (Fig. 2, Tab.3), probably simultaneously with Mlaka (Andri≠ 2007).Between c. 6900 and 5500 calBC beech canopy was occasionally opened by small-scale landscape burning at both study sites.However it is not clear from the present state of research, whether this fire disturbance regime was natural or anthropogenic (e.g.Mesolithic people were using fire to open the landscape).
The initial spread of beech forest was followed by beech decline and a change in forest composition, with an increase of monolete fern spores and initially lime at Mlaka and pine and trilete fern spores at Griblje.Landscape also became more open.This unusual beech decline is specific for Bela krajina region and it seems unlikely to be triggered by a global cold climatic fluctuation, although it is possible that vegetation composition was affected by local climate (e.g.drier and hotter summers in Bela krajina region of Slovenia).It is also possible that this vegetation change was caused by human impact, but the main problem of this explanation is that no archaeological sites, dated to 5500-5000 calBC were discovered in the area, so further archaeological and palaeoclimatological research is needed to better understand this unusual vegetation change.
After 5000 calBC open areas were regrown by hazel and oak and, between 4700 and 4100 calBC, hornbeam forest at Mlaka.An increase of 'anthropogenic indicator' herb taxa, e.g.ribwort plantain (Plantago lanceolata), plants of the Centaurea family (e.g.

Fig. 2. Griblje G3: Percentage pollen diagram (prepared by Tamara Koro∏ec).
cornflower), Cereal type pollen grains and other herb taxa, characteristic for agricultural fields and pastures, can be associated with Neolitihic settlements, located in the of both study sites.It is possible that people living near the Mlaka site were using local hornbeam forest (coppicing and wood pasture, which prevented beech regeneration).At about 3800 calBC human pressure on the environment increased, hornbeam forest was probably burnt and the landscape became open again, with meadows and fields located in vicinity of the Mlaka site.Similar vegetation development, but without a hornbeam phase and with less intensive forest clearance, was detected also at Griblje.Later, with decreased human impact on the environment, beech forest returned between 3700 and 2800 calBC.Forest composition changed again at 2800 calBC, when fir (Abies) replaced beech at the Mlaka site, whereas fir increase is not that pronounced at Griblje.This vegetation change could be associated with wetter climate and/or beech cutting (if beech wood was needed for metallurgy).Human impact on the environment slightly increased by 2500 calBC and after c. 1200-1000 calBC.The landscape was gradually becoming more open at both study sites, which could be associated with numerous Bronze Age and Iron Age sites in the area (Andri≠ 2007.763-776).

The early Holocene archaeological evidence
An examination of the archaeological evidence for the Neolithic/Eneolithic settlement pattern in Bela krajina provides a more episodic view of human activity in the early Holocene within the broad trends, shown by the palaeoecological research.
There is only minimal archaeological evidence for human activity in the region prior to the 5 th millennium BC.This is confined to a single site of Epigravettian occupation in a cave site, Judovska hi∏a, in the Krupa gorge (Pohar 1985.7-15).Thus the palaeoecological evidence gives us the strongest for potential Mesolithic activity outside the area of the single known site (Tabs.2, 3).

Models of Neolithic/Eneolithic settlement
The existing model of spatial exploitation in the Neolithic and Eneolithic in Bela krajina was developed in the 1980's.It was based on the results of the Moverna vas field survey project in the Krupsko polje, which was supplemented by small scale excavation on the central site of Moverna vas in 1988 (Budja op. cit.;Toma∫ 1997.113-142).The model formulated from this work also includes field survey work around Pusti Gradac and Zorenci on the upper reaches of the river Lahinja (Budja op. cit.).The small-scale excavation on the Gradac settlement from 1993-1995 showed a broadly similar chronological situation to that at Moverna vas (Mason 1995.183-199).The field survey data and the data from the two limited stratigraphic excavations were employed in conjunction with data from previously known sites that had not been subject to either modern excavation or extensive field survey to extend the model to whole of the river Lahinja river system and by inference to the whole of Bela krajina.
The model can now be supplemented by the results of developer-funded field survey and excavation, as well as research oriented field survey over the last ten years.This has led to the discovery of a range of settlement sites and off-site data, as well as a range of associated radiocarbon dates and pedological data.It also means that there are now other excavated sites in the upper reaches of the Lahinja system (R∫i∏≠a, Gradinje) and two large sites (Griblje, Podklanec) in the Kolpa valley (Mason 2001.10;2008.20-21;Mason, Bricelj 2006.41-42;Mason, Pintér 2001.141-142;Mason, Toma∫i≠, Nov∏ak 2006.95-96;Pintér 1998).These were defined by field survey that was supplemented by small-scale excavation and monitoring of infrastructure projects.This has permitted the extension of the model with a greater degree of veracity to the rest of Bela krajina.
Neolithic settlement (Fig. 3) The model posits the first appearance of agricultural settlement in the Middle Neolithic in the 5 th millennium BC.The initial colonisation of the interior of Bela krajina was centred on the drainage system of the river Lahinja and its tributaries.The primary settlements complexes were located on fertile soils, close to the rivers.Two typical locations can be identified -river meanders, e.g.Pusti Gradac and Gradac, and canyon or terrace edge sites, e.g.Griblje, Moverna vas, Podklanec, R∫i∏≠a and Zorenci (Mason 1995.192;2008.20).The cave site of Judovska hi∏a also continued in use (Pohar op.cit).
The earliest dates for this initial Neolithic phase are 4900 calBC at Moverna vas and 4840 calBC at R∫i∏-≠a, which would coincide with the hornbeam phase of strongest human impact (Budja 1994.20, Fig. 5; Mason 2008.20)(Tabs.1, 2, 3).The other sites lack radiocarbon dates, but it can be asserted with some degree of certainty that those that are only known from small-scale excavation (Gradac) and field survey (Pusti Gradac, Gradac, Zorenci and Griblje) have produced material from secure contexts, which indicates that they probably have a similar early origin.

Tab. 3. Radiocarbon dating of vegetational phases at Griblje and the associated archaeological settlement pattern.
It has been suggested that arable farming was an important part of the economy, but the site locations suggest that they were placed to exploit riverine environments for summer grazing, fodder, fishing and wildfowling, as well as the arable potential of the first terrace and the grazing potential of the drier karst hinterland (Mason 1995.185-187).
It is possible that the Middle/ Late Neolithic core settlements may have seen a seasonal element in their occupation throughout their use.The presence of midden deposits might be indicative of seasonal gatherings on the sites, which were otherwise occupied by smaller populations during the rest of the year.The presence of fine wares with burnt food deposits in the midden deposits are perhaps related to seasonal symbolic feasting, reintegrating an otherwise scattered population at a centre at certain times of the year.
Eneolithic settlement expansion (Fig. 4) The settlement pattern changed in the 4 th millennium BC.There is a visible expansion out from the Late Neolithic settlement centres into the drier karst hinterland through a process of secondary colonisation (Budja 1989.93-98;Mason 1995.193-195).This can best be seen in the original Moverna vas/ Krupsko polje field survey and in more recent work.Similar expansion can be seen around the Pusti Gradac site (meander and first terrace) and Eneolithic activity is clearly present on core settlements at Gradac, Zorenci and R∫i∏≠a (Budja 1992.102-109;Dular 1985.65;Mason 1995.191;2008.21).The most complete new evidence for a Neolithic/Eneolithic settlement pattern comes from the Griblje area and mirrors that of the Krupsko polje that is a Neolithic core settlement with later Eneolithic expansion into the drier hinterland (Mason 2001.10).However, it should be noted that seasonal ponds are present within the hinterland and probably formed foci for Eneolithic and possibly earlier Neolithic activity.Expansion was more likely to take place into the karst hinterlands and not laterally along the terrace -acti-vity here was present in the form of satellite activity areas throughout the Neolithic and Eneolithic.
It has been hypothesised that this expansion was linked to increased population and was made possible by the increasing importance of the stock-raising element in the economy, which led to a more mobile lifestyle (Budja 1989.93-98;Mason 1995.193-194).
This may be seen in the expansion of occupation of cave sites, such as Veliki zjot in the Kolpa valley (Leben 1991.169-191;Turk 1991.189).Enclosed upland sites, such as those at Spaha, Stra∫a, Ωidovec, Topli vrh and potentially Metlika, also appear in the Lengyel and Lasinja phases (Bre∏≠ak 1992.255-256;Dular 2001.89-106) and may be connected to increased competition in the area.They have also been tentatively connected to the appearance of transhumance (Mason op. cit. 194).However, it must be admitted that this hypothesis can only be tested against a large stratified faunal assemblage, which are unfortunately completely absent due to poor conditions for the preservation of animal bone in the area.The decline and disappearance of this settlement pattern in the second half of the 3 rd millennium BC has been seen as a result of overexploitation, leading to environmental degradation (Mason op. cit. 195).
Recent excavations in some swallow holes have provided vital insights into the nature of Eneolithic occupation and/or environmental degradation in the karst hinterland.The site of ∞ardak II in the lowland on the edge of the valley of the river Do-bli≠ica has no direct evidence of Eneolithic settlement, but the charcoal in primary erosion deposit in the is dated to the mid to late 4 th millennium BC (Beta-229151: calBC 3640-3420 (calBP 5460-5380), calBC 3380-3320 (calBP 5330-5270) calBC 3230-3110 (calBP 5180-5060) (Mason 2008.21;Mason, Predan, Murko, Pintér 2006a.23-24;2006b.24-26)(Tab.1).Eneolithic activity here would then fall within the Eneolithic expansion into the karst hinterland, but also within the phase of beech regeneration, noted at Mlaka.
The other karst hinterland site is that of Vinji vrh near Semi≠ was subject to limited excavation in 2004 (Mason, Britov∏ek, Pintér 2006.182-183).It proved to be an late Eneolithic or Early Bronze Age settlement site, which at least partially lay within a swallow hole.The settlement was of low intensity without deep occupation layers and was later subject to intensive erosion.It has produced a range of 14 C dates, but the final date lies within in the second half of the 3 rd millennium or at the beginning of the 2 nd millennium BC (Beta-229158: calBC 2200-1880 (calBP 4150-3830) (Mason 2008.22)(Tab.1).This is just outside the phase of more intensive human impact detected at Mlaka around 2500 calBC.It is clear that the intensity of human impact in the Eneolithic phase varied in space and time and was not a unitary phenomenon.The impact at Mlaka may be locally related to the Gradinje settlement site, which is without radiocarbon dates (Mason, Toma∫i≠, Nov∏ak 2006. 95-96).It should be also noted the final phase of occupation at Gradac terminated in a major erosion event, which is contemporary with Moverna vas phase 8 (Mason 1995. 195).It is now becoming increasingly clear that the main feature of the transition from the Eneolithic to the Early Bronze Age was the abandonment of the Neolithic/Eneolithic core settlements, but the continuation of the small-scale ephemeral settlements in their hinterland -indicative of an increasingly mobile way of life.
In conclusion, the evidence suggests that the existing model for Neolithic and Eneolithic settlement patterns in Bela krajina still offers the best means of interpreting the growing body of data for these periods.The model can now be extended with some confidence to the Kolpa valley and to other parts of the Lahinja catchment.What is now needed is the expansion of field survey work in these areas, particularly on sites that are incompletely defined, e.g.Podklanec.It is only then that the veracity of this model can be tested.The true nature of the Neolithic core settlements remains equally enigmatic, but as this can only be tested by large-scale excavation, it must remain so for the moment.On-going work in the spheres of palaeobotanical studies and off-site landscape studies may also be expected to provide vital results on the development of the Neolithic and Eneolithic landscape, particularly with regard to possible environmental degradation during and at the end of the Eneolithic, something which at the moment remains on the level of an intrasite phenomenon on the few excavated sites.It is also to be hoped that the tantalising palaeobotanical evidence for Mesolithic activity will be matched by archaeological evidence.

∴ ∴
The authors would like to thank prof.M.

Radiocarbon dating of vegetational phases at Mlaka and the associated archaeological settlement pattern.
Budja for the opportunity to present this paper at the 15 th Neolithic seminar in Ljubljana.Ildikó Pintér provided valuable help in the preparation of the powerpoint presentation.We would like to thank Tamara Koro∏ec for preparing the text figures, which place the text in context and make it understandable The palynological research was financed by the Slovenian Research Agency, the Dulverton Trust, ORS scholarship St. Hugh's College Oxford and Selwyn College, Cambridge.The Ministry of Culture of Slovenia funded the archaeological fieldwork at Griblje from 1998 to 2001 and at Gradac from 1993 to 1995, which was also partly funded by the Municipality of Metlika.The excavations at R∫i∏≠a in 2004 were funded by Elektro Ljubljana d.d., whilst the Municipality of ∞rnomelj funded the excavations at Vinji vrh in 2005 in the context of the Bela krajina watermain project.The major excavations at ∞ardak from 2004 to 2006 on the ∞rnomelj Western bypass were financed by the Slovenian Roads Agency.