UDK 902.6:502.3(497.4+439) Documenta Praehistorica XXVI The ecology of Neolithic environmental impacts -re-evaluation of existing theory using case studies from Hungary & Slovenia Adam Gardner Godwin Institute for Quaternary Research. University of Cambridge, UK arga@pcmail.nerc-bas.ac.uk ABSTRACT - Traditional interpretations ofSeolitiiic environmental imfiacts are discussed in light of neir palaeoecologicai sequences from Slovenia and Hungary. Popular perceptions of targe-scale Neolithic landsca/H? clearance are not apparent and instead the environmental response to agricultural activity is represented as a shij) in the species composition of forest following observed cycles of decline in the forest dominant. An existing archaeological model, hitherto unused by palaeoecolo-gisls. is adapted to explain the forest response to known Xeotithic activity. IZMEČEK - V članku razpravljamo o tradicionalnih razlagali neolitskih vplivov na okolje v luči norih fMileoekoloikili sekt ene iz Slovenije in Madžarske. Priljubljene razlage, da so v neotitiku v velikem obsegu čistili pokrajino, nismo mogli potrdili Namesto tega je odgovor okolja iui kmetovanje predstavljala sprememba sestave trsi v gozdu, kiji' ciklično sledila nazadovanju gozdne dominante. I)a hi razložili odziv gozda na znane neolitske aktivnosti, smo prilaga/iti že obstoječ arheološki model. ki ga ptdeoekologi doslej še niso uporabili. KEY WORDS - Ljubljansko barje: Matra hills \eotithic:palaeoecologv. pollen sequence; eniironmen-tai change; transition to fanning, forest clearance INTRODUCTION Conventional environmental research informs us that the Neolithic saw a period of intense deforestation throughout Europe which has been identified in sedimentär) sequences as a decline in forest taxa and an expanse in herbaceous taxa. particularly grami-nids. Thea' exist numerous records of forest clearance attributed to Neolithic activity, but many of these were either published prior to routine 14C determinations or were subject to inadequate '-»C dating and were consequently dependent on unreliable methods of relative dating such as the Blytt-Sernander peat-based classification. Ultimately, many such records should be discounted, but the concept of a dramatic widespread environmental change is so well established in the literature that it could be considered a paradigm of archaeological research. Consequently, intellectual inertia is extensive and presents a considerable hurdle for new research to overcome. The concept of forest clearance during the Neolithic is well established and is derived largely from analyses of sediments in northern Europe. Several models have been proposed to account for the expansion of agriculture during the early Neolithic, and palaeoecological schemes such as the Landnam of Iversen (1941), the leaf-foddering' model of Troels-Smith (1954) and the forest utilisation model of Goran-sson (1986; 1987) are widely cited in both the palaeoecological and archaeological literature. The schemes of Iversen and Troels-Smith pioneered the notion of forest manipulation by humans rather than forest clearance, yet suffered problems of chronology, resolution and logistics, which became apparent with a more objective and quantitative appreciation of human activity (e.g. Edwards 1979: ¡993; Rackham 1986). The Göransson scheme has been regarded as highly controversial (Edwards 163 Adam Gardner 1993) and suggests that pollen production increases as the landscape is first opened (sensu striclo Aaby 1988: 1994) due to greater exposure of tree crowns to sunlight, thus demanding a complete reinterpretation of Holocene palaeoecological records. Although these models have moderated the conventions for interpreting human activity, they arise from a palaeoecological perspective and do not benefit from modern archaeological opinion regarding the onset of agriculture (Ammerman & Caralli-Sforza 1973: 1984: Barker 1985: Halstead 1981; 1989: Zvelebil ¡988: ¡990; ¡994; Edwards el at. ¡996). In addition, all of these models have been formulated with respect to the unique Scandinavian transition to agriculture, which cannot be applied to the central or south-east European situation. These interpretations fail somewhat when applied to the archae-ologically and ecologically more complex situation evident in south-east Europe at the time of Neolithic settlement. Much of south-eastern Europe was covered w ith deciduous forest from the Early Holocene (Bennett et at. 1991: Willis 1994: 1995a) and the earliest Neolithic activity is presumed to have exploited natural open spaces such as river terraces and forest gaps (ran Andel & Runnels 1995; Simmons & ¡lines ¡996). Such activity is difficult to detect in the palaeoenvironmental record as there is little deliberate environmental disturbance and the dominant signal is that of the pristine' landscape. However, the subsequent intensification of agriculture during the consolidation phase (sensu Zielebil & Rouley-Conuy 1984) should be detected in the palynological record is a forest clearance event, as is evident in northern Europe (Edwards & McDonald ¡991). Recent research (Willis & Bennett 1994; Willis 1995a: 1995b; Gardner 1998: 1999; Gardner & Willis 1999) has demonstrated that the timing and magnitude of such events varies spatially and that, contrary to traditional palaeoecological interpretations, the earliest discernible forest clearance arising from the transition to agriculture in south-east Europe occurs several millenia after archaeological evidence for the earliest intensive farming. Hie factors contributing to this discrepancy are complex and have been described in detail elsewhere (IIIOf c- Bennett 1994; Willis 1995b; Gardner ¡998:1999), but it is useful to introduce them here. Briefly, the contributory factors identified include spatial representation of the pollen source area, the location of palaeoecological sites in relation to archaeological sites and the temporal resolution of the palaeoecological samples. The absence of any firm palaeoecological evidence for the early Neolithic, despite abundant archaeological evidence for agricultural settlements, suggests that either established palaeoecological methods are unsuitable for the interpretation of southeast European Neolithic impacts or that the impacts of Neolithic agriculture were so small that they remain undetected. This paper presents results from two sites in south-exst Europe w hich are situated in the vicinity of settlements occupied during the Neolithic. The palaeoenvironmental records from each of the sites w ill be discussed with reference to modern ecological studies of forest dynamics and used in a comparative analysis of Neolithic impacts. The concept of Neolithic landscape clearance w ill be addressed by recourse to new models which account for the shortcomings of established theories. STl DY SITES AND METHODS Sirok Nyirjes To (47° 55' 81" N, 20° 11' 14" E) b a small oligotrophy peat bog on the fringe of the Matra hills in lieves county, north-east Hungary (Fig. 1). Situated at 200 m a.s.l., die basin is an ellipse approximately 2(H) iti long and KM) m wide, surrounded by steep slopes supporting a gallery forest of Quercus cerris, Carpinus betulus and Con/us acel-lana. Podpesko Jezero (45° 58' 58" N, 1-»° 28' 30" E. 300 m a.s.1.) (Fig. 2) is a small circular lake of 80 m diameter at the north-western end of an elliptical basin which has an infilled with organic deposits. The surrounding steep slopes are covered by a dun rendzina soil supporting a Picea abies and Fagus sylralica forest plantation. All methods used are identical to those presented in Gardner (1998). with the exception of the coring method and age-modelling procedures adopted for •kmt J \ : BuM. ~> * Budapest • • • Tama bod /.....Km „ 0- 50 Fig, 1. 11 nul ion map of Sirok Syirjes To and surrounding area. 164 The ecology ol Neolithic environmental impacts - re-evaluation o( enisling theory using case studies from Hungary & Slovenia Fig. 2. location map of Podpesko Jezcro and surrounding area. Sirok. The peaty sediments from Sirok were collected using a hand-held Russian corer w ith a chamber of 50 cm length and 5 cm diameter. Hie Sirok age model was constructed using the Bernshtein polynomial curve (Bennett 1997), w hich w as then extrapolated following the routines of Maher (1998) using the sequence of Willis et at. (1995) to constrain the basal dale. RESULTS AND DISCUSSION' Sirok Nyirjes To Hie pollen sequence from Sirok Nyirjes To (Fig. 3a-b) has been divided into 7 zones, which are summarised in table 2. The 4.5 m profile spans the last 10000 years and is composed primarily of Sphagnum peat, except for the lowermost 50 cm. which comprise a unit of silly clay overlain by a 6 cm unit of day. The resolution of the sequence is one sample- even 113 calendar years, except for a fine-interval section where one sample represents 30 calendar years. The base of the sequence exhibits features characteristic of a change from a predominantly coniferous to a mixed deciduous forest. This type of event is typical of the Holocene after a period of total dominance by Pinus during the lateglacial/Holocene transition. The establishment of a mixed deciduous forest dominated by Conlus is complete by ca. 8300 cal BP and persists without change for ca 000 years. This type of forest hits no modern analogue (Rack-ham 1988). but is a feature apparent in several other pollen records from the region (e.g. sites in Wilis 1994; Willis et. al 1997) and represents a distinct phase in south-east European vegetation development (Huntley & Birks 1983). From ca. 6900 cat BP a series of high frequency cycles occur during which Conlus declines and other taxa, most notably Carpinus bet id us. expand briefly. It is plain that the previously stable forest Is being disturbed in some manner and that Corytus is being selectively removed, allow ing minor flourishes of C. betidus. Furthermore, minor increases in wet-habitat field-layer taxa such is Ihelypteris patu-stris and some members of the Liliaceae, in addition to expansions in Sphagnum and Filicales, imply a change both in hydrological conditions and light penetration to forest floor. The remov al of this disturbance is evident in the stabilisation of Conlus in reduced abundance and the expansion of C. betulus to a position of dominance within the forest, a component of the ty pical post- lab sample sample uncalib. cal. years cal. years calibration 8,3Cpdb code code material years bp BP ( 2o) BC AO (la) dataset ±0.1V used AA-27177 Sir (H cm Sphagnum peat 380 ±50 516 (467) 302 1434 (14S3)1648 1 -27.9 AA-27178 Sir 160 cm Sphagnum peat 2955 ±55 3324 (3100) 2945 1374(1150)995 2 -25.14 AA27I79 Sir 2-10 cm Sphagnum peat *580 ±55 5451 (5300) 5046 3501 (3350) 3096 1 -25.12 AA-271 KO Sir 300 cm Sphagnum peat 5135±W) 5989 (5910) 5738 4039 (3960) 3788 I -296 AA-27185 Sir 394 cm wood 5805 ±55 6742 (6640) 6469 4792 (4690) 4519 1 -29.2 AA-27186 JZ 233 cm wood 365 ±45 509 (345) 301 1441 (1605) 1649 1 -27.5 AA-27187 JZ 348 cm wood 930 ±45 935 (829) 727 1015 (112111223 1 -25.14 AA-27188 J7. 477 cm wood 61IO±75 7176 (6970) 6786 5226 (5020) 48.36 1 -31 2 AA-27189 JZ 651 cm wood 9075 ±70 10 279(10 030) 9922 8329 (8080) 7972 3 -29.0 Tab. I. "CAMS determinations and calibration results. Calibration performed using method I of CA LIU 3.0 /Stuiver & Reimer 1993/ I Risulis fn>m radiocarbon analyses are presented in table 1 and are incorporated in the stratigraphie diagrams (Figs. 3 & l). 165 Adam Gardner S I o J C s £ 1 i ? 1 ç ? Ï * 0. en CL S! Ôr ........................Ill.....I I I I I..... Il 8 | S I Fig. ib 167 Adam Gardner BP as C. betuius is reduced. The removal of the dominant forest element permits expansion in several taxa, all of which arc adapted to wetter conditions. At this time the Sphagnum which forms the peat dramatically increases production of spores, indicating a change in surface wetness (Barber ¡981). Subsequently, the space created by the removal of C. betuius is filled by an expansion of Quercus from ca. l-t(M) cal BP that persists to the present day. PodpeskoJezero The pollen sequence from Podpesko Jezero (Fig. 4a-b) has been divided into six zones, which are summarised in table 3. The fo.25 m of sediment recovered from PodpeSkoJezero span 11 350 years within a sequence of (from the base upw ards) marl, silty-marl and gyttja (Figs ta-b). Throughout the sequence. each sample represents 108 calendar y ears, except for the close-interval section between 74(H) and 5600 cal BP where each sample represents 64 calendar years. The base of the sequence incorporates the la-teglarial-Holocenc transition with a coniferous forest dominated by I'in us. Although its pollen is present in abundance, the Pitius forest at this time is not complete (sensu Aaby 1988; Peterson ¡983) and the pollen spectra probably represent a coniferous parkland' ty pe of environment w ith open spaces. The shift from a coniferous to a mixed-deciduous forest begins with the expansion of Betuia in the open spaces. follow ed by expansion of a full range of temperate deciduous forest taxa. From about 9000 cal BP, an expansion of Corylus occurs w hich is followed by the dramatic expansion of Fagus. Once established. Fagus persists in the forest until the present, but is subject to changes in the vegetation structure. Initially . Fagus dominates the Zone Age (cal. BP) Dominant features Charcoal ".AP Cone (grains cm J) AP: NAP Palyn. richness Sir-P7 1750-present Increase of Quercus. herbaceous elements and Sphagnum. Sharp increase of Hetula in top of zone high >70 20 (MM) to 30 (HH) low to moderate maximum Sir-PO 34001750 Maximum of Carpinus betuius: declines as Fagus expands low; 1 peak >75 20000 to 30000 maximum declines high Sir-PS 52003700 Increase of C betuius and C orienlalis. very low >75 ca. 200(H) moderate; increases high; steady SirP4 69005200 High frequency fluctuations in Corylus. Quercus and C betuius low >75 5000 to 15000 moderate high; steady Sir-P5 83006900 Mixed assemblage of Quercus, Tilia and l lmus dominated by Corylus low >80 ca. 15000 moderate increasing Sir-P2 89508300 Maxima in Tilia and Filicales, reductions in all other taxa. low- 50 ca. 20000 low decreasing Sir PI 10(H)-8950 Dominance of Poaceae. high v alues of Picea, Qiurcus, Corylus and Cyperaceae. moderate 50 ca. 12000 low increasing Tab. 2. Summary of major patynotogkat events by zone for Sirok. AH values are approximate: for further detail see figures 3a-b. 168 The ecology ol Neolithic environmental impacts - re-evaluation o( enisling theory using case studies from Hungary & Slovenia forest until the arrival of Abies ca. 7000 cal BP, when an Abies-Fagus forest similar to those currently present in the region w as formed. Soon after the establishment of the Abies-Fagus stage, external influences manipulate the forest structure to the extent that the dominant Fagus is replaced briefly by Cory-lus. The partial removal of Fagus and Abies provided conditions suitable for a secondary expansion of Cotylus not typically seen in Slovenian Holocene sequences (Sercelj 1996). Cotylus produces little or no pollen when growing as an understorey shrub {Rackham ¡988). therefore this event represents a real dominance by mature trees. A significant thinning of the canopy occurred locally around the basin, resulting in enhanced light levels penetrating to the field layer. From about 5000 cal BP the canopy openings contracted ¡is Fagus. Abies and Quer cus increased to form a forest, w hich although dense, nonetheless allowed sufficient light penetration for Carpinus betulus to establish and flourish. From ca. 3200 cal BP the landscape surrounding the basin underwent rapid and dramatic change, during which herbaceous taxa (Poaceae. Cyperaceae, Cannabis) and Filicales increased and several arboreal taxa (Fagus. Abies. C. betulus) were reduced. Low values for the AP:NAP ratio (Fig. tb) support this assumption and reveal that non-arboreal taxa dominated the pollen rain from ca. 3000 cal BP. This event probably represents the formation of the modern landscape of mixed (predominantly Abies-Fagus) forest surrounding open land w hich was exploited for agriculture. DISCUSSION Sirok Nyirjes To In the early postglacial, the slopes around Sirok supported an open parkland composed of Picea. Querent, Tilia and Cotylus, with open spaces dominated by Poaceae (Fig. 3a). Moderate burning of the vegetation w as occurring up to ca. 8900 cal BP. A transition from lake to peat deposits occurred in the basin from 10000-8300 cal BP. Over the same interval distinct changes were also occurring in the vegetation, each of which corresponds to a different sedi- Zone Age (cal. BP) Dominant features Charcoal "»AP Cone (grains cm-*) IK NAP Palyn. richness JZP6 2100-present Expansion of Poaceae, Cyperaceae, Cannabis and Filicales maximum values ca. 50 500030 ooo minimum maximum JZ-P5 64002100 Secondary (and maximum) Cotylus peak. Abies declines. Appearance of Carpinus betulus virtually absent >80 ca. 5000 low- increases JZ-P-t 70006400 Abies maximum. Fagus declining sharply. low >90 ca. 5000 high fluctuates: generally low JZ-P3 85007000 Total dominance by Fagus. negligible >90 ca. 5000 maximum values declines JZ-P2 98508500 Transition from coniferous to mixed deciduous assemblage. Corylus rise and fall, Fagus rise begins. low 80-90 ca, 12000 increasing fluctuates JZ-PI 113009850 Pinus dominates. High v alues of Picea, Betuia. Poaceae. moderate >75 ca. 7000 low low Tab. J. Summon of major palynological events by zone for Pttdpesko Jezero. AU values are approximate; for further detail see figures ta-h. 169 Aaam Gardner ■ I ■J, c a i ¥ ■-TurT'» —^m-TYMVN ■ ,-nvrT-/11 -o frrmtTAr: :MTTTrtiM7irTnirr^ b \r A rt ii ■ ■■-t-.-r-. i ■ i r >-r— I.......... -r I .rjTl fi itIi— \ \ 1 -—____/fMTTN.-. ^N Jc ^MDlri^J VrilV,JL - ——r^fTyr^fv uM M i \ oacnti , —- . fPVTi Tlvl K-, . . . _j-1---l i JlimLlniL.iimnJi.i.lliiiii 1111111111 .11 ■ 111 . 11........111, . , i ■ 11111, J—l—l-1-i_L I ■ I . , ■ ■ I -1 I ' ■ I i I I I ■ I I I I I I I : I .. I . i . ■ . i I.........Ill f§ § § § § § I #Xgs. 4 a-h. Pollen percentage diagrams for Podpesko Jezero plotted against core Hthology. depth and interpolated '*C timescaie; 4a Percentage diagram of major taxa; 4b Summary diagram showing total p*dlen percentages and supplementary spectra. See text for explanation of •Harris zones \ 170 The ecology ol Neolithic environmental impacts - re-evaluation o( enisling theory using case studies from Hungary & Slovenia mentary unit: 1. a declinc of Picea. Pin us and Bo tula and open-ground herbaceous elements (lake sediment); 2. a dramatic increase of TUia and Filicales (silt); and 3. the establishment of closed deciduous forest (Sphagnum peat). Tilia and Filicales were present at the base of the sequence, but the expansion to 25-30% of the pollen sum between 8900 and 8300 cal BP is unusual and not apparent in any other published diagram, although a sharp increase in Tilia at ca. 9500 BP (ca. 10 600 cal BP) has been recorded at Batorliget, eastern Hungary (IViUisetal. 1995), and a peak in Filicales dated to ca. 9500 cal BP is apparent at Kis Mohos To (Willis et al. 199 7), 50 km north of Sirok. Given the change in the hydrological regime and the poor dispersal characteristics of Tilia pollen, an expansion of this sort is difficult to reconcile in terms of ecology . Tilia typically thrives on well-drained soils such as those of the English lowland (Tutin et al. 1989: Rackham 1988; Pack/mm & Harding 1992) and is unlikely to have flourished on the waterlogged basin surface. However, dispersal of Tilia pollen is so poor that individual trees must have If ¡? ? n a. JZ-P2 a. N 9 N -3 N -a V vAyV-yv \ .............. ? 8 8 o '8 - t o o ° i o _ o 5 3 11111 ii 11 1111 3 1 I • 1 I 1 1 5000 6000 7000 8000 9000 10000 11000 8 CM § 8 1 § § 8 Fig. 4b 171 Adam Gardner been within a few metres of the basin (Nilsson c-Praglowski 1992) and the possibility exists that a dense gallery forest (which is composed of Corylus avellaiM. Carpinus betulus and Quercus cerris at present), of which Tilia was a dominant component, developed on the well-drained slopes at the periphery of the basin (cf. EUenberg 1988). The order Filicales includes taxa which characteristically thrive on burnt or nutrient-poor soils, but specifically certain taxa (such as Thefypteris palustris and Athyrium sp.) which grow on wet ground (7m-lin etal 1989). The charcoal concentration for this section of the core decreases prior to the expansion of Filicales (Fig. 3b), therefore fire would appear to be an unlikely factor to account for dtis event. However. the gradual infilling of the lake rev ealed newly exposed land which could have been colonised by advantageous taxa such as Filicales. Therefore, the abundance of Filicales spores at this time can probably be explained in terms of colonisation of newly exposed land. From 8300 to 5200 cal BP the forest around the site became dominated by Corylus with Quercus. Tilia and litmus. The expansion of Corylus to v alues >40% is a common feature of European Holocene pollen diagrams (lluulley & Birks 1983) and is conv entionally taken to represent a shrub layer in a forest. However, Corylus produces little pollen w hen growing in the understorey (Rackham 1988) and the population must therefore have either been a dominant canopy component or grow ing on the fringe of the basin. The latter is a distinct possibility in small basins such as Sirok, yet the ubiquity of a pronounced Corylus phase in European Holocene pollen diagrams (Huntley & Birks 1983) is such that there may once have existed a forest community dominated by Corylus which has no modern analogue. By extension, this implies that Corylus may have existed as a canopy tree. Three palynologically indistinguishable species of Corylus exist in Europe, of which two (C. avellana and C. maxima) are shrubs and one (C. colurna) is a canopy tree reaching 22 m (Huntley & Birks 1983: Tutin et at. 1989). All are currently distributed throughout the Balkans (Tutin el at. 1989) and are presumed to have been present throughout the Holocene (Huntley & Birks 1983). Given the possibility of two alternative species for the Balkan region, it is impossible to say whether the Corylus spectrum represented in this sequence is a shrub layer on the fringe of the basin or a now-extinct forest assemblage with Contus in the canopy. From 6900 to 5200 cal BP. there were several forest cycles in which the dominance of Corylus in the sequence was periodically reduced and in which Carpinus betulus rapidly expanded. These cycles occurred over v arious timescales, the shortest being ca. 36 years and the longest ca. 288 years. This selective removal of Corylus and the subsequent expansion of Carpinus betulus (Fig. 3a) may be due to anthropogenic intervention on the landscape, given the respectiv e ecology of the two taxa. However, the climate of this period was significantly different to that of today (Kutzbach & Guetter 1986: Huntley & Prentice 1988:1993: Kutzbach et at. 1993: Chedda-di et at. 1997) and underlying climatic factors which could account for these events must also be examined. Between approximately 7800 and 5700 cal BP the climate of Europe experienced conditions which have traditionally been described as optimal' and which were designated as Atlantic under the Blytt-Sernander classification (Roberts 19%). These optimal climatic conditions were considered to be uniformly warmer and moister than present across Europe. yet recent research has shown this not to be the case. Cheddadi et at. (1997) demonstrated that although the idea of a climatic optimum' is acceptable for northern Europe, conditions in south-east Europe at 6ka BP (6800 cal BP) were up to 4°C cooler than present and precipitation was up to 200 mm year-' greater. Given that Corylus is more tolerant of cool summers and of wetter conditions than Carpinus betulus (Huntley & Prentice 19931 there is no apparent climatic reason for the onset of the periodic declines observed in the Corylus curve. ALP and early copper-using cultures flourished on the sparsely wooded plains between 7300 and 5800 cal BP (Sherratt 1982a, ll)82b.- Willis et al. 1998) and the forests of the Matra and Bukk mountains represented the only reliable source of wood for raw materials and fuel, in particular for the nearby Tar-nabod (Fig. 1) settlements which were occupied from 72-iO cal BP. The possibility exists that forest grazing and browsing by livestock from the ALP settlements gradually introduced a change in the forest composition, but this is unlikely, as the At/old provided vast expanses of grassland suitable for both arable and pastoral use (Kosse 1979). Therefore, the coincident timing of these events in the pollen record and of the occupation of the Tarnabod tells suggests that the changes to the forest could be attributed to human activity. 172 Tha ecology oI Neolithic environmental impact» - re-«valuation ol existing Iheory using case stuOes from Hungary & Slovenia Coppicing of Corylus is an effective land-use strategy which would have ensured a continuous supply of raw material (Evans 1992) from the Sirok area. Carpinus betulus was present as a minor element of the stable Corylus dominated forest for ca. 600 y ears, and selective removal of Corylus through coppicing created the conditions into which C. betulus had the opportunity to expand. C betulus seedlings are light-demanding and severe disturbance or coppicing allows the establishment of new specimens (Rackham 1980) which, if left undisturbed, will grow and cast dense shade. Therefore, there must have been an external influence between 7000 and 52(H) cal BP which prevented the expansion of C. betulus into the forest canopy, C. betulus also coppices extremely well (Rackham 1980, Evans & Barkham 1992) and although it has virtually no uses as timber, burns at an extremely high temperature (Evans 1992; Mabberley 1997) and is hence valuable as fuel. It is therefore reasonable to suggest that two coppice cy cles were operative; Corylus coppice with a short rotation of 6-10 years and C betulus coppice with a longer rotation of 15-35 years (Evans 1992). Clearly these cy cles are too brief to register w ithin the limits of the available temporal resolution (I sample = ca. 36 years), but the varied timescales of these cycles (36-288 years) suggests that there may be a threshold level at which the signal of coppicing activity is recorded. The recovery of the woodland from ca. 5200 cal BP coincides w ith the abandonment of the settlements at Tarnabod (Kalicz & Makkay 1977) and a gradual eastward shift in settlement pattern during the Copper Age (Sherratt 1982a). It is highly probable, therefore, tiiat woodland recovery is linked to abandonment of coppicing. This w ould have lead to the inevitable closure of the canopy and the gradual accession of C. betulus to a position of dominance within the forest. Figures 3a-b shows that this process occurred over more than a millennium. More significant, however, is the apparent lack of any sedi-mentological influx through erosive disturbance during the entire period of coppicing and C. betulus expansion, indicating the validity of coppicing as a sustainable land-use strategy. The removal of human influence allowed natural forest processes to resume at Sirok and the expansion of Eagus occurred at the expense of the C. betulus canopy. Recent ecological studies (Peters 1997) have demonstrated that the deep shade cast by Eagus restricts growth in all the forest taxa present, and the decline in C. betulus appears to he direcdy associated with the expansion of Eagus. However, Eagus does not achieve total dominance in the forest and merely restricts C. betulus to a position within a mixed deciduous assemblage containing Quer-cus. Corylus. Tilia, t lmusand C. orientalis. Thus, the expansion of Eagus from ca. 3500 cal BP appears to have coincided with a reduction in total forest cover. From ca. 1700 cal BP to the present day, the v egetation around the basin changed once again and became dominated by Quercus with a greater proportion of non-arboreal taxa Taxa tolerant of wetter conditions (e.g. Alnus and Sphagnum) expanded and charcoal concentrations increased (Fig. 3b). suggesting burning of the landscape. The most striking feature of these events is the expansion of Quercus to a position of dominance in the pollen assemblage. Clearly, the combination of increased burning, the removal of the deciduous tree cover and the apparent shift in the water table are related to increased impacts from anthropogenic activity, perhaps as a result of increased populations, lack of effective controls (e.g. coppicing) to mitigate against landscape degradation, or a combination of both. The period from 1700 BP to the present represents a turbulent period in die cultural history of Hungary. The early Middle Ages (2000-1625 BP; AD 0375) saw the colonisation of the Matra region by Barbarian groups of Celts. Dacs. Vandals and Sarma-tians, followed by migration groups of Huns, Avars and Slavons between 1625 and 1105 BP (AD 375895) and die Hungarian conquest of 1105-1045 BP (AD 895-955) (Trogmayer 1980; Willis et al. ll)97). Not until the evolution of the Hungarian Kingdom from 995 BP (AD 955) did the population resettle into village communities and return to managed use of woodland (Eiigedi 1986). by w hich time the forest composition had totally changed. The response evident between 7000 and 5200 cal BP is therefore not apparent over the past millennium. In addition, Quercus. the dominant taxon present from ca 995 BP (ca. 920 cal BP), does not coppice well and is a highly valued timber tree (Rackham 1980; 1986; Mabberley 1997). Furthermore, Quercus mast pro vides good fodder for liv estock (Aeivbohl 1983) and the forests may therefore have been subject to an alternative form of managed use. Podpesko Jezcro The Holocene pollen stratigraphy of Podpesko Je/ero is similar to the regional pollen record estab- 173 Adam Gardner lished for Slovenia (Cultberg 1991; Sercelj 1996). The characteristic lateglacial assemblage of predominantly coniferous types with Poaceae and Cypera-ceae is evident up to ca. 9800 cal BP. From 9800 cal BP the vegetation changed rapidly and became dominated by a mixed deciduous forest composed of Querais, Tilia, Ulmus and Corylus. and the open areas decreased in size. This persisted for approximately 600 years until the expansion of Corylus at ca. 9000 cal BP. The basin at this time was occupied by a diverse, dominantly deciduous forest into which Corylus expanded rapidly. The forest was dominated by Corylus for ca. 300 years until the expansion of Fagus at ca. 8500 cal BP. An almost pure Fagus forest existed in the basin until the expansion of Abies at ca. 7000 cal BP when a mixed Abies-Fugus occupied the slopes surrounding the lake. From ca. 7000 cal BP a series of changes to the forest occurred which does not follow the sequence of Sercelj (1996) or the general trends suggested for south-east Europe (Huntley & Birks 1983; Huntley & Prentice 1993). The decline of Fagus and Abies and the subsequent expansion of Corylus, Quercus and Carpinus betulus from ca. 6400 cal BP (Fig. 4a) is an unusual event in forest development The characteristic internal dynamics of Fagus forest are driven by its canopy dominance and tolerance of extremely low light levels, particularly during the juvenile stage of growth (Newbold 1983; Peters 1997). Shade tolerance w hilst juvenile is especially important to maintain Fagus dominance. Furthermore, Huntley and Prentice (¡993) show that elsewhere in south-east Europe. Fagus and Abies were expanding at this time and the eastern range of Corylus was in decline. Thus, there is no readily apparent reason for a Fagus-Abies population to become reduced by Corylus through competition alone. A possible explanation for the retraction of Fagus is a change to an unfavourable climate more suitable for other deciduous forest taxa. However, this does not account for the initial expansion of Fagus at 8500 cal BP during conditions w hich were less favourable for growth than those of ca 6400 cal BP (Huntley & Prentice 1993). The microclimate of the Ljubljana Moor region is modulated by the constant temperature of inflowing karstic streams front the southern highlands. Consequently, local climatic conditions throughout the Holocene have been warmer and moister than elsewhere (Sercelj 1996). favouring the grow th of temperate deciduous forest Coupled with the close proximity of refugial populations (Willis 1994), this explains the early expansion of Fagus at 8500 cal BP during climatic conditions described by Huntley and Prentice (1993) as unfavourable. However, given this constant microclimate, there is no apparent climatic explanation for the decline of Fagus at 6400 cal BP, Recent research (M. Biulja. unpublished data, personal communication. 1999) has produced the earliest reliable radiocarbon date for human settlement on Ljubljana Moor. New excavations at Breg have revealed Mesolithic artefacts and occupation layers dated to 9180 cal BP, w hilst at Babna Gorica early-Neolithic occupation layers containing monochrome pottery have been dated to 6290 cal BP. The early Neolithic has hitherto been considered absent from the region and these excavations provide a cultural framework in which to view the forest changes at PodpeSko Jezero. Forest grazing by livestock has been shown to reduce deciduous forest by limiting the regeneration of seedlings (Pigott 1983) w ith the effect that forest stands gradually decline. The immature shoots of deciduous taxa especially Fagus, are exceptionally palatable to grazing animals (Rackham 1980) and consumption of seedlings drastically reduces the success rate of regeneration (Neu bold 1983). Furthermore. consumption of beechnuts by herbivores (New-bold 1983) places an additional strain 011 the marginal success of Fagus replacement. Grazing and browsing by animals could therefore account for the decline in Fagus apparent from 64OO cal BP, hut this would not have operated in isolation and additional agencies must have contributed to the selective removal of forest elements. The selective remov al of Fagus and Abies could be regarded as the result of repeated small-scale clearances within the catchment with the intention of creating useful pockets of land for cultivation or grazing. The opening of the forest canopy would have promoted the establishment of a rich field-layer and advantageous fast-growth arboreal taxa such as Cory lus respond rapidly and soon dominate the canopy gap. In addition, present day ecological studies have demonstrated that Fagus will not regenerate if gaps are large (Peters ¡997). However, these findings have not been incorporated into any of the popular models proposed to characterise the environmental response to Neolithic agriculture. Carugati et al (1996) present a model for small-scale agricultural activity in the forest of Neolithic Sammardenchia on the Friuli Plain of north-east 174 Thg BCQlog» 0> NcoMhiç envitonmefit.1l impacts - re-cvaluatioo ol existing theory using cas« studies Irpm Hungary & Sloven a Italy. In this example, a small gap is created in Quer-cus-Fraxinus forest by felling or ring-barking, and the cleared land is used for the growth of cereals. The stumps of felled trees remain in position and after removal of any undergrow th, cereals are sown in the sheltered glade surrounded by mature forest. After sev eral years of use for crops, the perimeter of the clearance is encroached by shrubs such as Cory-Ins which subsequently close the abandoned plot as the forest recolonises. Although relatively simple, this model serves as a useful example for the nature of early human impacts upon dense forest and demonstrates the manner in which small-scale activity may produce a localised shift in forest composition. Repeated openings of this type within a small forest are palynologically inv isible (Sugita ct til. 1999). yet over an extended period of time may result in a shift in the forest composition. The expansion of Corylits from 6400 cal BP reached a peak ca. S8(H) cal BP, but was rapidly reduced by the expansion of Carpinus behdus. C. orienlalis and by the partial recovery of Fagus (Fig. 4a). Small increases in Poaceae and Cyperaceae are apparent, indicating greater availability' of light in an increasingly open landscape. In addition, the summary diagram (Fig. 3b) shows a reduction in arboreal pollen to ca. 80"o, which suggests the existence of openings in the canopy (settsu Aaby 1988; 1994). Large-scale landscape clearance is apparent from 2300 cal BP and herbaceous types increase. Accumulation of organic deposits began at 2300 cal BP with the onset of eutrophication and the transition to the gyttja phase of sedimentation. At this time, the forest in the catchment w as composed of Abies-Fagus forest with Picea. Pimis and Quercus and a greater (ca. 40°i.) non-arboreal component dominated by Poaceae and Cyperaceae. The forest clearance at 2300 cal BP represents the first large-scale landscape disturbance and the first reliable appearance of anthropogenic indicators' (Behre 1981; 1986). Increases in Poaceae, Cyperaceae and Filicales palynomorphs reflect the expansion of non-arboreal vegetation within the catchment w hich, nonetheless, was floristically poor. The expansion of Cannabis -type pollen in the sequence includes Hiiniultis I it/ml us and Cannabis saliva. which are palynologically indistinguishable. Both types are used for fibre production (Polunin 1980; Mabberley 1997). although the primary use of Hamulus is in brewing. However, Godwin (1967) suggests that as only the female infructescence of Hiiniiilus is used for brewing. female plants are selectively cultivated and the pollen is scarce. Therefore pollen of Cannabis-type is more likely to represent Cannabis saliva than Hu-mulus liipulus. An expansion of this magnitude can be taken to represent either pollen deposition from a large stand grow ing nearby or retting of fibres in the lake (Bradshaw el at. 1981; Willis el al. 1998). Cannabis will not grow in large stands unless cultivated (Polunin 1980; Clapham el al. 1987) so it is reasonable to assume that both processes were activ e. The range of economically important taxa present in the pollen assemblage from 2300 cal BP is surprisingly poor in consideration of the number and density of archaeological settlements in the region (Bre-gant elal 1980). Other than those taxa mentioned above, the only crops present in more than trace abundance are represented by pollen of cereal-type. Fagopyrum. Apiaceae and Brassicaceae, and of those, none reach 5% of the total terrestrial pollen. In addition, arboreal pollen remains at ca SO4'» from 2300 cal BP to the present day. Surface samples taken from the sediment-water interface (Gardner 1998) display a similar pollen assemblage, with low v alues of crop pollen and a total arboreal pollen value of 48%. The only marked difference is a higher proportion of Fagopyrum and the total absence of Cannabis iype pollen. Thus the present day landscape was formed ca. 2300 cal BP. DYNAMICS OF HOLOCENE ENVIRONMENTS IN SOUTH EAST EUROPE Theoretical Framework In the Holocene palaeoecological record of southeast Europe there appear three dominant phases of env ironmental change (Gardner 1999): 1. Early Holocene 'primary' forest development; recolonisation of the landscape by forest upon climatic improvement; rapid change characterised by high species turnover. 2. Mid Holocene secondary' forest development: maturation of die forest soils and canopy structure; expansion of the dominant forest (axon. 175 Adam Gardner 3. Late Holocene large-scale forest clearance by human populations; high species diversity , essentially static-equilibrium achieved. Each of these three phases is characterised by regional differences in floral composition and timing of events, but the broad explanation for the mechanisms of change is simple. In the early Holocene, the magnitude of climatic events exerted a greater influence upon the environment than did any aspect of human activity . Thus, the rapid environmental change recorded for the early Holocene in southeast Europe was driven by climate. Conversely, the late Holocene was characterised by a comparatively stable climate and exploitation of the environment by human populations which exceeded climatic forces in driving environmental change. However, in the mid Holocene, a combination of climatic factors and steadily increasing human impact produced a palaeoenvironmental signal which is complex and extremely difficult to define (Birks & Line 1993). Explaining the mechanisms of change The dramatic environmental change apparent in die early Holocene phase was driven entirely by an increase in summer temperatures of I- 1()°C betw een 12 and 9ka BP (Kutzbach & Guetter I'M: Kutz-bach et al. 1993) All other factors relating to environmental change during this period are linked to increased temperature, in particular moisture availability (Willis 1994; Bennett & Willis 1995) and soil development (Penning/on 1986: Willis et al. 1997). Expansion of primary forest at the onset of the Holocene in southern Europe occurred in response to climate change and w as characterised by a rapid succession from raw mineral soils supporting coniferous parkland to a mixed deciduous forest on organic soils, characterised by high v ariability and high species turnover (Bennett & Willis 1995). The mid Holocene phase of south-east Europe is characterised by (loristic stability, with the development of a secondary deciduous forest of usually one dominant taxon (e.g. Carpinus sp. in Greece, C. bet ul us in Bulgaria. Quercus ilex in Croatia and Fagus in Slovenia - Willis 1994 and references therein) which persists to the present day, albeit in reduced importance. This phase is eloquently summarised by-Bennett & Willis (1995) as pattern 2 of their scheme for Holocene vegetational development. In climatic terms, the mid Holocene forest developed during an optimum growth period when temperatures were 2°C higher than present (Huntley & Prentice 1993) and humidity more suitable for dense deciduous woodland (sensa Magri 1996). In human terms, the mid Holocene forest matured at a crucial point in the establishment of sedentary societies in southeast Europe who. in addition to growing arable crops, exploited the forest for raw materials, fuel, pasture, fodder and wild food resources. The late Holocene phase is characterised by almost complete domination by human activity as an environmental driving mechanism. Climatic forces remained very much in evidence and periodic oscillations such as the Little Ice Age from ca. Al) 1590-1N50 (Lamb 1977) were of sufficient magnitude to cause local re-advance of Alpine glaciers (Grove 1988). However, increasingly intensive land-use from the late Bronze Age/early Iron Age (ca. 3ka BP) onwards initiated irreversible soil erosion (van Andel et al. 1990: Hahtead 1996). a reduction of 50% of global forest cover (Birks & Line 1993) and the expansion of arable field and meadow plant communities. Synthesis of Mid Holocene Human Impacts The dominant driving forces behind early and late Holocene environmental change are undoubtedly climatic and human agencies respectively, but the situation for mid Holocene environmental change is not so clearly defined. Climatic change alone cannot account for the subtle changes in forest composition evident in this study, yet other than the proximity of known archaeological sites, there is no evidence from the sedimentary or charcoal records to suggest a dominantly human origin for these changes. Comparison of the characteristics of Holocene forest species decline with long pollen sequences from previous w arm stages serves as a guide to unrav el human impacts from natural processes. Several such sequences exist in southern Europe (e.g. loannina (Tzedakis 1993: ¡994) and Tenaghi Phillipon (Wijmstra 1969; van der Wiel & Wijmstra 1987a; 1987b) in Greece; Yalle di Castiglione (Foltieri et al. 1988) in Italy; and Les Echets (de Ben id ten & Reilie 1984) and La Grande Pile (Woillard 1978; de Beau-lieu & Reilie 1992) in France), each of which extend back to at least the oxy gen isotope stage 5e (Eemian interglacial) and reveal the development of a forest assemblage composed of ecological groups which bear striking similarities to those seen in the Holocene of south-east Europe. The distinction should he made that species may or may not adopt the same 176 The ecology ol Neolitnic envirwimental impacts - re-evaluation ot existing theory mmg case studies from Hungary K Slovenia positions in previous warm-stages as they do in the Holocene or may be totally absent (Bennett & W illis 1995) (cf. Fagus at Ioannina and Carpimis beta-Iiis at Les Echets and La Grande Pilej, but the overall ecological classification (e.g. temperate deciduous. boreal etc.) Ls similar. Therefore, at the onset of each warm-stage a coniferous woodland assemblage existed which changed rapidly to a mixed temperate deciduous forest. This was followed by a mature secondary forest with a dominant broad-leaved taxon w hich developed by the middle of each stage. All of the mid warm stage forest phases from the long terrestrial sequences outlined above are terminated by the expansion of cold tolerant forest elements (e.g. Picea or Pinus) over secondary deciduous taxa. thus completing the interglacial cycle' (Birks 1986). In no instance is there a shift in forest composition as seen in the PodpeSko Jezero sequence (Figs, -ta-b) and no secondary expansion of advantageous taxa such as Corylus. Similarly , the nature of the transition to secondary forest at Sirok Nytrjes To w hereby a Corylus dominated forest (which has no modern analogue) is punctuated by cycles of Carpimis betulus (the secondary forest dominant) expansion is not apparent in any of the long sequences. The implication of underlying climatic trends producing distinct ecological phases in w armstage forest development serves as a guide in evaluating the additional external factors experienced during the mid Holocene. Thus, although a climatic cause could possibly produce the change in forest composition at these sites, it is not apparent in previous warm stages and, given the presence of agricultural communities, is more likely to have resulted from human activity. Comparing the Evidence The on-going excavations at Breg have rev ealed lith-ic and ceramic evidence for Mesolithic and early Neolithic occupation of Ljubljana Moor and biological analyses in progress are expected to produce data on subsistence strategies (M Budja, personaI communication, 1999). Similarly, excavations at Maharski Prekop (Bregant 1974a; 1974b; 1975) have revealed the full range of crops and animals used by the inhabitants of the region and have demonstrated a community reliant on arable agriculture. but which maintained important livestock herds and continued to exploit wild resources. In contrast, there is no firm palaeoecological evidence for Neolithic arable agriculture in any of the available sequences for the area (e.g. Ibis study; Cidiberg & $er- celj. 1978a. 1978b; Amine, 1997). Similarly, archaeological excavations of the nearby Tarnahod tells have revealed a typical ALP assemblage comprising a full range of domesticated plant and animal remains (Kalicz & Makkay 1977) which is not recorded in the Tarnabod palaeoecological sequence (Gardner 1999). The attendant off-site sequences, Podpesko Jezero and Sirok Nyirjes To, do not display any evidence of arable farming activity from the pollen record (sensu Behre 1981; 1986) during these settlement phases, nor do they show any evidence of landscape instability from the sedimentological record. In contrast, both sites display a subtle suite of changes to the forest composition which are interpreted from this study to be a result of managed forest-use. The forest-farming model of Carugati et al. (1996) has already been presented and serves as a useful scheme for visualising the changes apparent in a south-east European landscape during the early Neolithic, yet offers no cultural basis for the sequence of human activities suggested. Harris (1996a; 1996b) advances a more detailed conceptual model for plant and animal exploitation in which he proposes a multi-stage transition from dependence on wild resources to dependence on domesticated resources. For the plants he suggests two phases of cultivation in which initially small clearings are used for morphologic-ally wild plants before intensive agriculture begins with fully domesticated crops. Similarly, an intermediate protection phase is suggested by Harris (1996a; 1996b) to demarcate exploitation of wild animals and full domestication. Harris' (1996a) model serves as a useful conceptual basis on which to explain the forest changes apparent in this study. Table 4 shows a development of die Harris model adapted to clarify proposed impacts on mature south-east European forest during the preliminary stages of Neolithic agriculture. Initial agriculture. both arable and pastoral, impinged on relatively untouched forest and influenced forest development to the extent that the composition changed and there was a shift in canopy dominance. Subsequent agricultural expansions introduced greater influence oil the env ironment by a change to large-scale clearance practices, which did not introduce any clearance-abandonment cy cles and allowed the partial recovery of certain forest elements (sensu Iversen 1941). Ihe scheme in table 4 can be fitted readily to die sequences from Sirok Nyirjes To (Figs. 3a-b) and Pod- 177 Adam Gardner Tab. 4. Conceptual basis for evolution of agricultural systems and attendant impacts on forested environments ^modified from Harris. I «)%a; 1996c/ Sole that individual phases are not rigid and the entire scheme is temporally flexible. Wild food procurement Food production wild plants and animals important as food domesticated food production dominant Gathering and collecting \ Cultivation Agriculture • small clearances • large clearances • minimal tillage • increased tillage \ \ Hunting Scavenging Fishing \ Livestock keeping Livestock raising • pasture • transhumance • forest grazing . nomadic and browsing pastoralism \ Decreasing dependence on wild resources ........-...... Increasing dependence on domesticated resources forest virtually unmodified forest thinned; compositional change partial recovery of forest large scale landscape clearance peSko Jezero (Figs, 4a-b) and serves to illustrate the extent of activity occurring within the catchments. In both cases the basal Harris zone is relatively pristine forest, more or less unaffected by human activities. Previous work (Mellars 1975. Simmons & Innes 1996) has proposed forest firing by Mesoli-thic populations as a strategy for hunting success, although Rackham (1980) has discredited such theories on account of the incombustibility of temperate deciduous forest. Despite the excav ations at Jaszbe-reny (Kertesz el al. 1994) in Hungarv and Breg (Chapman & Midler 1990. Budja 1997) in Slovenia demonstrating Mesolithic occupation of the two study regions, there is no evidence from charcoal records or pollen data to suggest Mesolithic manipulation of forest. Therefore, the early llolocene forest at both sites can confidently be termed pristine'. The second Harris zone' reveals two different manifestations of forest thinning. The Sirok sequence (Fig. 3b) shows a pollen assemblage interpreted as the forest response to a coppice regime, demonstrated by a high rate of change and fluctuations in Corylus and Carpinus belulus. Plant food produc- tion occurred on the fertile terr aces of the A (fold adjacent to the ALP settlements, as did pasture for livestock. At Podpesko Jezero. (Fig. -ib) repeated cycles of clearance for small cultivation or pasture plots are suggested in this study to have caused a shift in the composition of the forest canopy as a result of competitive interactions between tree species. Suitable land for cultiv ation was available on Ljubljana Moor, but this was probably under extreme pressure as a result of loss of low-lying land to unpredictable floods. At Sirok and Podpesko Jezero the forest recovered partially as the smaller scale disturbance became more restricted. At Sirok. abandonment of coppicing initiated further development of secondary forest, leading to the expansion of Fagus. At Jezero. the Abies-Fagus forest recovered slowly and was interrupted by a phase of Carpinus belulus as changing hydrological conditions on Ljubljana Moor permitted greater exploitation of the land surface there. The final phase of the Harris scheme occurs ca. 2000 BP at Sirok and PodpeSko Jezero, during large- 178 The scotogy o» Neolilh«c environmental impacts - re-evaluation ol Quisling theory using case studies from Hungary 8 Slovenia scale landscape clearance. However, Quercus expands during this stage at Sirok and the total forest cover at both sites remains at 50%. Results from this study suggest that extensive settlement of both regions by Iron Age agricultural communities has resulted in widespread disruption of the natural environment and the establishment of die modern landscape. Thus, the adoption of a modified Harris' model for Neolithic activity can be used to account for the environmental changes recorded in the palaeoecolo-gical sequences from Sirok Nyirjes To and Podpesko Jezero. No model is infallible, yet the Harris model serves to place the small scale palaeoecologjcal changes observed at the two sites into a tangible archaeological context. CONCLUSION Palaeoecological models previously proposed to explain the earliest human activity are considered unsuitable for characterising the environmental re- sponse to Neolithic agriculture in Hungary and Slovenia. Such models over-emphasise the ability of Neolithic communities to change their environment and suffer problems concerning sampling resolution and chronology. Archaeological models fare better in terms of temporal resolution, but include few considerations of forest dy namics or broader environmental concepts. A new scheme, based on the existing Harris archaeological model, is presented in an attempt to classify changes apparent in this study according to the human activity apparent from the archaeological record. By application of this model, it is evident that the environmental response to Neolithic activity is complex and can not be represented by simple models of forest clearance and arable field expansion. Consideration of contemporary ecological research is crucial to deciphering pollen signals of forest use by prehistoric societies, and by adopting appropriate methods and models suitable for unique geographical situations, a more realistic impression of human-environment interaction may be gleaned. REFERENCES AABY B. 1988. The cultural landscape as reflected in percentage and influx diagrams from two Danish ombrotrophic mires. In Birks H. H.. Birks II.J. B.. Ka-land P. E. & Moe D. (eds.), Die Cultural Landscape, Past. Present and Future. 209-228. 1994. NAP percentages as an expression of cleared areas. In I'rcnzel B. (ed.). Evaluation of the land Surfaces Cleared from Forests in the Roman Iron Age and the Time of Migrating Germanic Tribes Based on Regional Pollen Diagrams: 1327. Palaeoclimate Research X. AMMERMAN A.J. and CAVALLI-SFORZA L L 197.3. A population model for the diffusion of early fanning in Europe. In Renfrew C. (ed). Die Explanation of Culture Change. 1973 343-357. Models in Prehistory. 198-1. 7lie Neolithic Transition and the Genetics of Population in Europe. Princeton. ANDRIČ M. 1997. Paleoetnobotanika v neoHtskih študijah M. A. Thesis, University of Ljubljana. BARBER K. E. 1981. Peat Stratigraphy and Climate Change. Balkema, Rotterdam. BARKER G. 1985. Prehistoric Farming in Europe. Cambridge. BEIIRE K. E. 1981. The interpretation of anthropogenic indicators in pollen diagrams. Pollen el Spores 23 225-245. ed. 1986. Anthropogenic Indicators in Pollen Dui-grams. Balkema, Rotterdam. BENNETT K. D. 1997. PSIMPOLL 3.00 & PSCOMB 1.03: 'C' programs for analysing and plotting pollen data. Available from the Department of Quaternary Geology, University of Uppsala at URL:http://www. kv.geo. uu. se/software .html BENNETT K. D.. TZEDAK1S P. C. and WILLIS K. J. 1991. Quaternary refugia of north European trees. Journal of Biogeograph v 18. 103-115. 179 Adam Gardner BENNETT K. D. and WILLIS K, J. 1995 The role of ecological factors in controlling vegetation dynamics on long temporal scales. 6'tórnale Botánico Italiano 129: 243-254. BIRKS H.J. B. and LINE J. M. 1993. Glacial refugia of European trees - a matter of chance? Disserlationes Botanicae 196. 283-291 BIRKS H.J. B. 1986. Late Quaternary bioticchanges in terrestrial and lacustrine environments, with particular reference to north-west Europe. In Berglund B. E. (ed.), Handbook of Holocene Palaeoecology and Palaeohydrotogy, 1986:3-65. BRADSIIAW R. II. W„ COXON P.. GRIEG J. R. A. and HALL A. R. 198-1. New fossil evidence for the past cultivation and processing of hemp (Cannabis sativa L.) in Eastern England Netv Pliyologist 89. 503510. BREGANTT. 1974a. Kolišče ob Maharskem Prekopu pri Igu - raziskovanja leta 1970. Poročilo o raziskovanju paleolita. neolita in eneolita v Sloveniji lil: 7-35. 1974b. Kolišče ob Maharskem Prekopu pri Igu -raziskovanja leta 1972. Poročilo o raziskovanju paleolita. neolita in eneolita v Sloveniji IV: 3967. 1975. Kolišče ob Maharskem Prekopu pri Igu -raziskovanja 1973 in 1974 leta. Poročilo o raziskovanju paleolita, neolita in eneolita v Sloveniji IV: 7-115. BREGANT T., JERMAN M.. SLABE M. and YUGA I). 1980. Varovanje arheoloških terenskih spomenikov Ljubljanskega Barja. Arheološka zaščitna raziskovanja na ljubljanskem Barju r letu 1979: 11-18. BCDJA M. 1997. Landscape changes in the Neolithic and Copper Ages in Slovenia. Case Study: Ljubljana Basin microrcgiou In Chapman J. and Doiukhanov P. M. (eds.), landscapes in FIilv. Colloquia Pontica 3: 75-78. CARUGAH m. C., CASTELLETTI L. and ROTTOLI m. 1996. L'agricoltora nel primo neolítico del Friuli. Le ricerche a Sammardenchia Fagninola a Valer. In Ferrari A. & Pessina A. (eds.), Sammardenchia e i primi agricoltori del Friuli Banca di Crédito Cooperativo di Basiliano: 103-112. CHAPMAN J. C. and MÍJLLERJ. 1990. Early farmers in the Mediterranean basin: the Dalmatian evidence. Antiquity 64: 127-134. CHEDDADI R., YU G„ GUIOTJ., HARRISON S. P. and PRENTICE I. C. 1997. The climate of Europe 6000 years ago. Climate Dynamics 13 1-9. CLAPHAM A. R.. TUTIN T. G. and MOORE D. M. 1987. Flora of the British Isles (3rd edition). Cambridge University Press. CU1.1BF.RG M. 1991. Late Glacial Vegetation in Slovenia Razprave SA7.U IV XXIX10. 1-52. CUUBERG M. and SERCELJ A. 1978a Ksilotomske in palinološke analize rastlinskih ostankov s kolišča na Partih pri Igu - Izkopavanja leta 1977. Poročilo o raziskovanju paleolita. neolita in eneolita v Sloveniji VI: 95-99 1978b. Ksilotomske analize lesa iz s kolišča ob Maharskem prekopu - Izkopavanja 1976 in 1977. Poročilo o raziskovanju paleolita, neolita in eneolita v Sloveniji VI 103-107. DE BEAULIEU J. L. and RF.ILLE M. 1984. A long Upper Plesitocene pollen record from Les Echets. near Lyon. France. Boreas 13 111-132. 1992. Tlie last climatic cycle at La Grande Pile (Vosges. France): a new pollen profile. Quaternary Science Reviews 11. 431-438. EDWARDS K.J. 1979. Palynological and temporal inference in the context of prehistory, with special reference to the evidence from peat and lake deposits Journal of Archaeological Science 6 255-270. 1993. Models of mid-Holocene forest farming for north west Europe. In Chambers F. M. (ed.). Climate Change and Human Impact on the landscape. 132-145. EDWARDS K. J„ HALSTEAD P. and ZVELEBIL M. 1996. The Neolithic transition in the Balkans - archaeological perspectives and palaeoecological evidence: a comment on Willis & Bennett. The Holocene 6.1. 120-122. ELLENBERG H. 1988. Vegetation Ecology of Central Europe (4th edition). Cambridge University Press. 180 The ecology ol Neolithic environmental impacts - re-evaluation o( enisling theory using case studies from Hungary & Slovenia EVANS M N. and BARKHAM J. P. 1992. Coppicing and natural disturbance in temperate woodlands - a review. In Buckley G. P. (ed.), Ecology and Management of Coppice Woodlands. 79-98. EVAN'S J. 1992. Coppice forestry - an overview. In Buckley G. P. (ed.), Ecology and Management of Coppice Woodlands: 18-28. FOLLIF.R1 M„ MAGRI 1). and SADORI L. 198«. 250000-year pollen record from Valle di Castiglione (Roma). Pollen et Spores 30.329-356. FUGEDI E. 198b. Castle and Societies in Medieval Hungary. Akademiai Kiado. Budapest. GARDNER A. R. 1998. Biotic response to early Določene human activity: results from palaeoenviron-mental analyses of sediments from Podpeško Jezero. Poročilo o raziskovanju paleolita. neolita in eneo-lita v Sloveniji XXIV: 63-77. 1999. Tite Impact of Neolithic Agriculture in the Environments of South-East Europe. Ph.D. Thesis. University of Cambridge. GARDNER A. R and WILLIS K. J. 1998. Prehistoric farming and the postglacial expansion of beech and hornbeam: a comment on Kiister. Ihe Hdocene 9/1: 143-145. GODW IN H. 1967. The ancient cultivation of hemp. Antiquity 41: 42-49. 137-138. GORANSSON H. 1986. Man and the forest of nemoral broad-leaved trees during the Stone Age. Striae 24: 143-152. 1987. On arguing in a circle, on common sense, on the smashing of paradigms, on thistles among flowers, and on other things. Norwegian Archaeological Review 18. 43-45. GROVE J. M. 1988. 7he Little Ice Age. Routledge, London. HALSTEAD P. 1984. Stategies Eor Survival: An Ecological Approach to Social and Economic Change in Early Farming Communities ofThessaty, N. Greece. Pli.D Thesis, University of Cambridge. 1989. Like rising damp? An ecological aproach to the spread of farming in south-east and central Europe. In Milles A., Williams D. & Gardner N. (eds.). The Beginnings of Agriculture. BAR International series 496. 23-53 1996. The development of agriculture and pasto-ralism in Greece: when, how, who and what? In Harris I). R. (ed). The Origins and Spread of Agriculture and Pastoralism in Eurasia. 296-309. HARRIS 1). R. 1996a. Introduction: themes and concepts in the study of early agriculture. In Harris 1). R. (ed). The Origins and Spread of Agriculture and Pastoralism in Eurasia: 1-11. 1996b. The origins and spread of agriculture and pastoralism in Eurasia: an overview. In Harris D. R. (ed.). Hie Origins and Spread of Agriculture and Pastoralism in Eurasia: 552-574. 1996c. Domesticatorv relationships of people, plants and animals. In Ellen R. & Fukui K. (eds.), Redefining Nature■ Ecology, Culture and Domestication Berg: 437-463■ HUNTLEY B. and BIRKS H. J. B. 1983. -4« Atlas of Past and Present Pollen Maps for Europe: 0-13.000 years ago. Cambridge University Press. HUNTLEY B. and PRENTICE I. C. 1988. July tempera tures in Europe from pollen data, 6000 years before present. Science 241: 687-690. 1993- Ilolocene vegetation and climates of Europe. In Wright H. F... Kutzbach J. E„ Webb T. Ill, Ruddiman W. F„ Strect-Pcrrott F. A. and Bartlein P. J. (eds ), Global Climates Since the Last Glacial Maximum. 136-118. IVERSEN J. 1941. Land occupation in Denmark's stone age. Geological Survey of Denmark II series 66:5-68. KALICZ N. and MAKKAYJ. 1977. Die Linienbandke-ramik in der Grossen Ungarischen Tiefebene. Akademiai Kiado, Budapest. KKRTESZ R.. SUMEGI P.. KOZAK M„ BRAUN M„ FE-LEGYHAZI E. and HERTELENDI E. 1994. Archaeological and palaeoecological study of an early Holocene settlement in the Jaszag area. Acta Geographica Ac Geologica el Meteorologica Debrecina 32. 5-47. KOSSE K. 1979. Settlement ecology of the Koros and Linear Pottery Cultures in Hungary. BAR International Series 64. 181 Adam Gardner Kl'TZBACHJ. E. and GUETTER P.J. 1986. The influence of changing orbital parameters and surface boundary conditions on climate simulations for the past 18,000 years. Journal of Atmospheric Sciences 43:1726-1759. KUTZBACH J. E.. GUETTER P. J., BEHL1NG P. J. and SELIN R. 1993. Simulated climate changes: results of the COHMAP climate-model experiments. In Wright H. E.. Kutzbach J. E.. Webb T. III. Ruddiman W. F.. Street-Perrott F. A. and Bartlein P. J. (eds.). Global Climates Since the Last Glacial Maximum: 24-93- LAMB H. H. 1977. Climate: Present. Past and Future (2 vois.). Methuen. London. MABBERLEY D.J. 1997. The Plant Book: a Portable Dictionary of the Higher Plants. Cambridge University Press. MAGRI D. 1996. The Neolithic transition and pa-laeoecology in the Balkans: a comment on Willis & Bennett Tlie /lolocene 6.1. 119-123. MAHER L. J. 1998. SLOTDEEP v. 1.8 adds DC profiles to its DC map. INQUA Commission for the I lolocene. Sub-Commission on Data-Handling Methods Newsletter 18 (available from URL http://www.kv.geo.uu. se/inqua/newsl8/nl8-ljm2.htm) MELL\RS P. 1975. Fire ecology, animal populations and man: a study of some ecological relationships in prehistory. Proceedings of the Prehistoric Society 42: 14-45. NEWBOLD A. J. 1983. Some aspects of the regeneration of oak and beech. Ph.D. Thesis. University of London. NILSSON S. and PRAGLOWSKI J. (eds.) 1992. F.rdt-man 's Handbook of Palynology (2nd Edition). Munksgard. Copenhagen. PACKHAM J. R. and HARDING D.J. L. 1982. Ecology of Woodland Processes. Edward Arnold, London PENNINGTON W. 1986. Lags in adjustment of vegetation to climate caused by the pace of soil development: evidence from Britain. Vegetatio 67:105-118. PETERS R. 1997. Beech Forests. Kluwer Academic, Dordrecht. PETERSON G. M. 1983. Recent pollen spectra and zonal vegetation in the western U SSR. Quaternary Science Renews 2. 281-321. PIGOTT C. D. 1983. Regeneration of oak-birch woodland following exclusion of sheep. Journal of Ecology 71: 629-646. POLUNIN 0. 1980. Flowers of Greece and the Balkans. Oxford University Press. RACKHAM 0. 1980. Ancient Woodland: Its History, Vegetation and Uses in England. Edward Arnold. London. 1986. the History of the Countryside. Dent, London. 1988. Trees and woodland in a crowded landscape - the cultural landscape of the British Isles. In Birks H. H„ Birks H. J. B. Kaland P. E. & Moe D. (eds.). The Cultural Landscape, Past. Present and Future: 53-78. ROBERTS N. 1998. The Holocene (2nd edition). Blackwell. Oxford. SERCELJ A. 1996. The Origins and Development of Forests in Slovenia. Razprare SAZVIV: 35. SIIERRATT A. 1982a. The development of Neolithic and Copper Age settlement in the Great Hungarian Plain part 1: the regional setting. Oxford Journal of Archaeology 1.3 287-316. 1982b. The development of Neolithic and Copper Age settlement in the Great Hungarian Plain part 2: site survey and settlement dynamics. Oxford Journal of Archaeology 2,1: 13-41. SIMMONS I. G. and INNES J. B. 1996. Disturbance phases in the Mid-Holocene vegetation at North Gill. North York Moors: form and process. Journal of Archaeological Science 23: 183-191. SUGITA S„ GAILLARD M.-J. and BR0STR0M A. 1999-Landscape openness and pollen records: a simulation approach. The Holocene 9,4: 409-421. TROELS-SMITH J. 1954. Ertebolle culture - farmer culture. Results of the past ten years' excavations in Aamosen Bog, West Zealand. Aarhogerfor Sor disk Oldkyndighed og Historic. 182 The ecology o< Neolithic environmental impacts - re-evaluation ol enisling theory using case studies Irom Hungary & Slovenia TROGMAYER 0. 1980. Pannonia before die Roman conquest. In Lengyel A. C. & Radan G. T. (eds.). The Archaeology of Roman Pannonia: 65-82. TUTLN T. G.. HEYWOOD V. H., BURGESS \. A.. MOORE D. M.. VALENTINE D. H.. WALTERS S. M. and WEBB D. A. (eds ). Flora Europaea (2nd Edition, 5 volumes), 1989. Cambridge University Press. TZEDAKIS P. C. 1993. Long-term tree populations in northwest Greece through multiple Quaternary climatic cycles. Nature 364: 437-440. 199-». Vegetation change through glacial-interglacial cycles: a long pollen sequence perspective. Philosophical Transactions of the Royal Society of London B345. 403-432. VAN ANDEL T. H. and RUNNELS C. N. 1995. The earliest farmers in Europe. Antiquity 69: 481-500. VAN ANDEL T. H., ZANGGER E. and DEM1TRACK A. 1990. Land use and soil erosion in prehistoric and historical Greece. Journal of Field Archaeology 17: 379-396. VAN DER WILL A. M. and WIJMSTRA T. A. 1987a. Pa lynology of the low er part (78-120 in) of the core Tenaghi Phillipon II. Middle Pleistocene of Macedonia. Greece. Review of Paleobotany and Palyno-logy 52: 73-88. 1987b. Palynology of the 112.8-197.8 m interv al of the core Tenaghi Phillipon III, Middle Pleistocene of Macedonia. Greece. Review of Paleobotany and Palynology 52: 89-117. WIJMSTRA T. A. 1969. Palynology of the first 30 metres of a 120 m deep section in northern Greece. Acta Botanica Seerlandica 18. 511-527. WILLIS K. J. 1994. The vegetational history of the Balkans. Quaternary Science Reviews 13 769-788. 1995a. Land degradation in the Balkans: variations in time and space. In van der Leeuw S. (ed.), L Homme el la degradation de I. Environment. Editions APDCA. 1995b. The pollen-sedimentological evidence for the beginning of agriculture in southeast Europe and Anatolia. In Poročilo o raziskovanju paleoli-ta. neolita in eneolita v Sloveniji XXII: 9-25. WILLIS K. J. and BENNETT K. D. 1994. The Neolithic transition - Fact or Fiction? Palaeoecological evidence from the Balkans. The Holocene 4/3: 326330. WILLIS K. J.. BRAUN M., SUMEGI P. and TOTII A. 1995. The Late Quaternary Environmental History of Batorliget. N.E. Hungary . Palaeogeograpliy. Pa-laeoclimatology. Palaeoecology 118. 25-47. 1997. Does soil development cause vegetation change or vice versa? A temporal perspective from Hungary. Ecology 73(2): 730-740. 1998. Prehistoric land degradation in Hungary: Who. How & Why? Antiquity 72: 101-113- W01LLARD G. M. 1978. Grande Pile peat hog: a continuous pollen record for the last 140.000 years. Quaternary Research 9: 1-21. ZVELEB1L M. 1990. Mesolithic prelude and neolithic revolution. In Zvelebil M. (ed ), Hunters in Transition. 5-15. 1994. Plant use in the Mesolithic and its role in the transition to farming. Proceedings of the Pre historic Society 60:35-74. ZVELEBIL M. and ROWLEY-CONWY P. 1984. Transition to farming in northern Europe: a hunter - gatherer perspective. Norwegian Archaeological Relieu' 17,2: 104-128. ZVELEBIL M. and ZVELEBIL K. V. 1988. Agricultural transition and Indo-European dispersals. Antiquity 62: 574-583. 183