ELSEVIER Quaternary Science Reviews 23 (2004) 2369-2387 QSR Trees or no trees? The environments of central and eastern Europe during the Last Glaciation Katherine J. Willisa*, Tjeerd H. van Andelb ^School of Geography and the Environment, University of Oxford, Mansfield Road, Oxford OX1 3TB, UK ^Department of Earth Sciences, Cambridge University, Cambridge CB2 3EQ, UK Received 27 January 2004; accepted 5 June 2004 Abstract The location and survival of trees in the coldest stages of the last full-glacial has long been of interest to palaeoecologists, biogeographers, archaeologists and geneticists alike. In particular, where species survived in isolated refugia and the influence that this has had upon the long-term ancestry of the populations, remain key research questions. However, the exact location of refugia during the coldest stages of the full-glacial still remains illusive for many species of fauna and flora, with different lines of evidence often being at odds. This is particularly true for Europe. Emerging evidence from various fossil proxies, palaeoclimatic modelling and genetic research is starting to suggest that the traditional paradigm that trees were restricted to southern Europe and in particular the three southern peninsulas (Balkan, Italian and Iberian) during the full-glacial is questionable. This is backed by increasing evidence, including 151 14C-dated and identified pieces of macrofossil charcoal wood from 40 localities in central and eastern Europe to indicate that during the last full-glacial populations of coniferous and some deciduous trees grew much further north and east than previously assumed. This paper reviews the fossil evidence and considers it alongside genetic and palaeoclimatic evidence in order to contribute towards a newly emerging synthesis of the full-glacial refugial localities in Europe and their influence upon the ancestry of European species. Plotted against a new high-resolution millennial time-scale for the interval ~32-~ 16 ka BP in Greenland our evidence shows that coniferous as well as some broadleaf trees were continuously present throughout those interstadial/stadial cycles for which there are adequate data. © 2004 Elsevier Ltd. All rights reserved. 1. Introduction Ever since it was first suggested that during the coldest periods of the last full-glacial (~37-16kaBP), the most probable location for cold-stage refugia of temperate tree taxa was in southern Europe (Frenzel and Troll, 1952), many researchers have looked to these locations for evidence of refugial populations (e.g. Huntley and Birks, 1983; Bennett et al., 1991; Willis, 1992; Tzedakis, 1993; Willis, 1994; Follieri et al., 1998; Carrion, 2002; Tzedakis et al., 2002). The theory that southern Europe and the Near East provided conditions suitable for refugia of temperate tree taxa is based on a number of assumptions relating to the full-glacial environments of "■Corresponding author. Tel.: +44-1865-271919 E-mail address: kathy.willis@geog.ox.ac.uk (K.J. Willis). 0277-3791/$-see front matter © 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.quascirev.2004.06.002 those regions and their ability to supply the necessary conditions for growth. Those included an ice-free and in most places permafrost-free terrain, a soil able to support woody vegetation, and a climate with extremes that did not exceed the physiological capabilities of the tree taxa. There was plenty of evidence to support such a model given little or no geomorphological evidence for permafrost below 45 °N (Washburn, 1979) and palaeoclimatic modelling (COHMAP, 1988; Wright et al., 1993) indicating that a steep climatic gradient existed across southern Europe at 21kaBP).! This modelling 'The notation kaBP stands for calendar dates (ice-core, U-Th, TL, OSL and ESR) and calibrated 14C dates in millennia (cal ka) before present (BP). Where appropriate 14C years are appended in brackets (xx + xxx 14C yr b.p.) to specific dates. Calibration with CalPal (Joris and Weninger, 2000) available from the CalPal website; http:// www.calpal.de. 2370 K.J. Willis, T.H. van Andel / Quaternary Science Reviews 23 (2004) 2369-2387 therefore predicted that in the southern peninsulas of Europe January temperatures would have been no less than 0°C and precipitation 2-4 mm/day (Wright et al., 1993), well within the physiological extremes of most European temperate tree taxa (Kienast, 1987; Prentice and Helmisaari, 1991; Nikolov and Helmisaari, 1992; Larcher, 1995). Although the original refugial model implied that 'forests' could have survived in southerly locations during the cold stages of the Quaternary, extensive populations of trees have never been detected. Instead, palaeoecological sequences from the Near East and the three southern peninsulas of Europe (Balkan, Italian and Iberian) indicate that temperate tree taxa possibly survived in small pockets at microenvironmentally favourable locations with usually only a few tree taxa detected in low concentrations in any one place. Aridity was probably a significant limiting factor for tree growth, a fact supported by evidence for an increased presence of trees in mid-to-high altitude sites during the full and Lateglacial where there would have been more orographical precipitation (e.g. Wohlfarth et al., 2001; Carrion, 2002; Bjorkman et al., 2003; Wick et al., 2003). In many of the lowland regions, however, trees also survived and a number of pollen sequences extending back through multiple glacial-interglacial cycles show a small but persistent presence of mesophilous and thermophilous tree pollen throughout the cold stages. This full-glacial vegetation has commonly been described as 'open parkland' (van Andel and Tzedakis, 1996, 1998; Tzedakis et al., 1997; Follieri et al., 1998; Tzedakis et al., 2002; Davies et al., 2003). The idea that southern Europe and the Near East provided full-glacial refugia for flora (and associated fauna) is now well established in the literature and constitutes an important biogeographical basis for a number of molecular phylogenetic studies into the ancestry of populations (e.g. Hewitt, 2000, 2004; Willis and Whittaker, 2000; Petit et al., 2003; Kadereit et al., 2004; Lascoux et al., 2004). The isolationary aspect of refugia is of particular interest because it provides a mechanism for allopatric speciation whereby the genetic differences apparent in present-day populations are attributed to cold-stage refugial isolation (e.g. Taberlet and Bouvet, 1994; Konnert and Bergmann, 1995; Demesure et al., 1996). It is this isolationary aspect of the refugia model however, that now appears to be coming under closer scrutiny (Petit et al., 2003; Lascoux et al., 2004; Willis and Niklas, 2004). A key assumption associated with the southern European refugia model is that further north and east into central Europe the palaeolandscape was tree-less. The tree populations in southern refugia were therefore 'isolated' for long intervals and it was during these periods of isolation that genetic differences became established between them. The suggestion that central and eastern Europe remained tree-less during the full-glacial is based primarily on the assumption that the greater environmental extremes in these regions, would have restricted tree growth severely. No long palaeoecological sequences from central and eastern Europe (east of 10°E) exist to support this interpretation. Instead, there is increasing evidence from macrofossil charcoal records that the interpretation of a tree-less full-glacial central and eastern European landscape may not be correct. In addition, new palaeocli-matic simulations of the Stage 3 Project suggest that full-glacial conditions in central and eastern Europe were not nearly as severe as previously anticipated (Barron and Pollard, 2002; Barron et al., 2003; Pollard and Barron, 2003). Pollen-based "tree-less tundra" models for Europe north of the transverse mountain ranges of the Pyrenees, Alps and Carpathians (e.g. Frenzel, 1992; Alfano et al., 2003; Huntley et al., 2003) have repeatedly been questioned by researchers of the Late Pleistocene Mammalian fauna such as Guthrie (1990, 2000), Guthrie and van Kolfschoten (1999), Lister and Sher (1995), Yurtsev (2001) and Stewart and Lister (2002), because the carrying capacity adequate to feed numerous large herbivores such as mammoth, woolly rhinoceros, reindeer, giant deer, bison or auerochs demands a very productive environment (named "steppe-tundra" or "mammoth steppe" by Guthrie, 1990). This is not the place to go deeply into this issue, but the arguments, generally dismissed or ignored by palynologists, have gained considerable strength from the mammalian evidence contained in the environmentally oriented Mammalian database of the Stage 3 Project (Stewart et al., 2003) and from hitherto rather neglected central and eastern European data compiled by Musil (2003). The large number of Upper Palaeolithic sites spread across much of central and eastern Europe north of the mountain barrier (Musil, 2003; van Andel et al., 2003) bears witness that animal resources were plentiful. This paper sets out to review the macrofossil charcoal evidence alongside other fossil evidence in the context of our current understanding of the climates and climate changes of the time, and to consider all in the light of the traditional 'southern European refugial model'. We shall do this in four ways. First, we review published macrofossil charcoal evidence for full-glacial trees from 40 localities (in 7 countries) that are north and east of the traditional 'southern refugia' localities. Second, we examine other lines of fossil evidence that appears to support this macrofossil charcoal evidence. Third, we review recent palaeoclimatic modelling for central and eastern Europe and consider whether tree growth was possible under those climatic conditions. Finally, we briefly consider the macrofossil charcoal evidence alongside recent genetic evidence to address the question K.J. Willis, T.H. van Andel / Quaternary Science Reviews 23 (2004) 2369-2387 2371 whether a new interpretation of the full and Lateglacial environments for central and eastern Europe is needed. 2. Trees or no trees: the macrofossil charcoal evidence Macrofossil charcoal is a term used to describe organic plant material (fragments > 2 mm diameter) preserved in the fossil record through the process of burning and charcoalification (Schweingruber, 1978, 1990). It is a particularly useful record for the identification of fossil wood because the level of preservation is often good enough to examine the transverse and longitudinal ring structures and thus identify the tree to genus (and in some cases species) level. Traditionally, macrofossil charcoal analysis has been a tool used by archaeologists for the identification of charcoal from occupation layers for 14C dating and to understand the use of local materials in subsistence. 14C dating of charcoal is often suggested as the most reliable material for dating from archaeological sites because of its high inert carbon content and the least risk of chemical contamination (Damblon and Haesaerts, 1997). Surprisingly, however, it has rarely been used in a comprehensive way for palaeoenvironmental reconstructions although there are notable exceptions (e.g. Badal et al., 1994; Damblon, 1997; Hansen, 2001; Rudner and Siimegi, 2001; Damblon and Haesaerts, 2002; Figueiral and Terral, 2002; Ntinou, 2002). One reason for the apparent lack of use of this technique in mainstream palaeoenvironmental research is that it does not provide a continuous stratigraphic sequence (unlike, e.g. pollen analysis from lake sediments). Rather, it represents a 'snapshot' of the vegetation in and around the site during the time of occupation. In environments where continuous sequences of lake sediments are available, macrofossil charcoal analysis is therefore not the first choice of technique. There are also concerns over lack of stratigraphic control, possible downward movement through the sequence, or introduction from elsewhere. However, there are many sites now where stratigraphic control is paramount to the research and numerous pieces of charcoal have been identified and dated from clearly defined stratigraphic horizons. As to the suggestion that the material may have been brought into the archaeological site from elsewhere, it is obvious that this is often true for some of it. The question that really needs to be addressed, however, is how far woody material will be carried/dragged by humans and/or other transportation agents? Research into the transportation of woody material for burning on hearths suggests that in almost all cases it is probably from a radius of no more than 2 hours walking (<10km) from the site (Vita-Finzi and Higgs, 1970; Bailey, 1997; Hansen, 2001; Bailey, 2004). Finding macrofossil charcoal at an occupation site, therefore, may not prove that it was growing in immediate proximity to the site, but it can be assumed that it was growing somewhere in the region and that dating this charcoal gives an age for its presence. In some sites, however, the evidence for local growth is compelling since silhouettes of tree trunks plus roots marked out in charcoal have been discovered (Willis et al., 2000) that indicate in situ growth and burning. To counter some of the above concerns regarding macrofossil charcoal evidence only samples from sites with good temporal and stratigraphic control have been included in this review. These sites should have clearly identifiable and well-dated occupation horizons with 14C dates based on identified pieces of charcoal. Ideally the site should also have material cross-dated from more than one laboratory. In order to examine the coldest periods of the last glaciation but within the confines of reliable radiocarbon dating, this review includes published sites in central and eastern Europe that contained macrofossil charcoal evidence for woody plant species dated between ~42 and 19kaBP (see footnote 1). Countries included in the review are Austria, Czech Republic, Croatia, Hungary, Moldova, Poland, and Romania (Fig. 1)—all regions traditionally described as being beyond the full-glacial refugial localities for tree taxa. Therefore, they provide an interesting west-east transect in which to examine the distribution of woody vegetation during the last full-glacial. The macrofossil charcoal evidence from each region will be briefly described. 2.1. Austria Macrofossil charcoal assemblages from two Upper Palaeolithic sites in Austria have been included in this review (Table 1) namely Willendorf II and Schwallen-bach (Haesaerts et al., 1996; Damblon, 1997; Damblon and Haesaerts, 1997). Both sites are situated on the banks of the Danube (Fig. 1) and have well-defined occupation layers contained within loess sequences. Willendorf was occupied between 44 and 26kaBP (42,000-23,000 14C yr b.p.) and Schwallenbach between 43 and 35kaBP (40,000-29,500 14C yr b.p.). In Willendorf, six main occupation layers have been identified with 25 14C dates on macrofossil charcoal wood ranging between 44 and 26kaBP (42,000-23,000 14C yr b.p.: Table 1) (Haesaerts et al., 1996; Damblon and Haesaerts, 1997). Species identified are predominantly needle-leaved (Pinus cembra, Pinus, Picea-Larix) although there are also several deciduous taxa including Betula and Salix. The presence of tree species in all six occupation layers leads the authors to conclude that they were an important element of the full-glacial palaeolandscape in this region (Haesaerts et al., 1996; Damblon, 1997). The site of Schwallenbach is also located on the Danube, approximately 1 km upstream from Willendorf SITE LOCATIONS .39 v , 13 34 -.19 /2-^7'12153{ ...... 38 40 \ • 23—. 21 ™ -32—« «29 , --,.....'--'7X33?' 255.°.35 - 2'8 ' /' 36 \ -.31' 11 \ 20 — .25*- 35,000-30,000 Cal.yr B.P. 25 46-1* rj 40,000-35,000 Cal.yr B.P. 30,000 -25,000 Cal.yr B.P. KEY TO TREE SYMBOLS i Abies ^ Pinus Abies ^alba A. Pinus ' cembra A P/'nus t mugo A P/'nus T sylvestris Q Populus ^ Quercus k Rhamnus sylvatica ♦ cathartica ^ Fraxinus (Jfh Sa/« Juniperus ^ Sorf)us 9Juniperus communis () Betula ^ Corylus Q Carpinus Fagus . Sorbus * aucuparia " /.a/w /.a/w L decidua ff) Taxus I baccata 2& /.am/ T" P/'cea ^ P/'cea/ r /.am ^ P/'cea A P/cea t excelsa Fig. 1. Location of the 40 full-glacial macrofossil charcoal sites plus the tree species/genera identified at each site for the time intervals 40,000-35,000, 35,000-30.000 and 30,000-25,000 ka BP (full citation to sites is given in text and Tables 1-7). Numbers correspond to site names that are detailed in Tables 1-7. K.J. Willis, T.H. van Andel / Quaternary Science Reviews 23 (2004) 2369-2387 2373 Table 1 Identification of tree species/genera in macrofossil charcoal layers from sites in Austria and 14C dates (based on the identified macrofossil charcoal wood within these layers) Site number (with reference to Figs. 1 and 2): Site location: longitude (E)/latitude (N) Site name Site type containing macrofossil charcoal layers2 Tree species/genera identified in macrofossil charcoal layers Published 14C dates (uncalibrated yr b.p.r Lab. code Cal yr BP References to source of data (1) 15°24748°19' Willendorf II Occupation layers situated upon loess (2) 15°24748°19' Schwallenbach Occupation layers situated upon loess Pinns cembra type, Pinns 23,200 + 140 GrA-893 26,270 + 210 sp., Picea sp. 23,400 + 190 GrA-493 26,400 + 220 23,670 + 120 GrA-494 26,580 + 210 Pinns cembra type, Pinns 24,710 + 180 GrA-894 28,010 + 780 sylvestris, Picea sp. 25,230 + 320 GrN-17801 28,710 + 1000 25,660 + 350 GrN-17802 29,290 + 960 25,400 + 170 GrN-21690 28,970 + 910 25,400 + 170 GrN-20767 28,970 + 910 25,800 + 800 GrN-11191 29,230 + 1330 Pinns cembra type, 23,830 + 200 GrA-491 26,720 + 270 Picea sp. 23,990 + 130 GrA-492 26,840 + 280 Pinns cembra, Pinns 26,500 + 480 GrN-20768 30,440 + 640 sylvestris, Picea sp. 26,150 + 110 GrA-1016 30,300 + 450 Pinns sylvestris, Pinns 28,560 + 520 GrN-17804 32,550 + 700 cembra, Picea sp. Betula sp., Pinns 30,500 + 900 GrN-11193 34,460 + 1000 sylvestris, Picea sp. 32,000 + 3000 H-246/231 36,270 + 3250 Pinns sylvestris, Picea 31,210 + 260 GrA-501 35,060 + 520 sp.; Betula sp. 31,700 + 1800 H-249/1276 36,320 + 2280 32,060 + 250 GrN-1273 36,610 + 1400 Picea sp., Salix sp. 37,930 + 750 GrA-896 41,710 + 280 38,800 + 1530 GrN-17805 42,120 + 670 Picea sp., Picea-Larix 39,500 + 1500 GrN-11190 42,510 + 740 41,600 + 4000 GrN-17806 44,360 + 2900 Picea-Larix > 36,000 GrN-17807 — 41,700 + 3700 GrN-11195 44,360 + 2900 Picea sp. 29,530 + 800 GrA-5219 33,300 + 890 Picea sp. 31,680 + 290 GrA-6899 35,410 + 570 Picea sp. 39,000 + 2600 GrA-5221 41,780 + 2110 Picea sp. 37,400 + 2600 GrA-5222 39,840 + 2500 Pinns cembra, Picea sp. 39,920 + 1300 GrN-21801 42,660 + 690 Haesaerts et al. (1996) , Damblon and Haesaerts (1997) , Damblon (1997) (1996) aSpaces between dates indicate different occupation layers. II (Damblon, 1997). Here seven occupation layers have been identified in the loess sequence with dates ranging between 43 and 33kaBP (-40,000-29,000 14C yr b.p.; Table 1). Of the five 14C dates obtained on macrofossil charcoal from this site all were based on charcoal remains of Picea and P. cembra, providing further evidence for the regional presence of these coniferous trees during the coldest stages of the last full-glacial. 2.2. Czech Republic In the Czech Republic at least four Upper Palaeolithic sites that contain well-dated macrofossil charcoal 2374 K.J. Willis, T.H. van Andel / Quaternary Science Reviews 23 (2004) 2369-2387 Table 2 Identification of tree species/genera in macrofossil charcoal layers from sites in the Czech Republic and charcoal wood within these layers) C dates (based on identified macrofossil (3) 16°40749°11' Stránská skála Occupation layers situated upon loess (4) 16°35749°10' Bohunice Occupation layer situated upon loess (5) 16°38748°53' Dolní Occupation Věstonice II layers situated upon loess (6) 16°38748°52', Dolni Věstonice I Occupation layers situated upon loess (7) 16°40748°52' Pavlov I Occupation layer situated upon loess Pinus sp. cf. Pinus, Abies 41,300 + 3100 sp. cf. Abies, Picea sp. 38,500 +1400 38,200 + 1100 cf. Picea, Betula sp., Salix, Larix, Picea-Larix, Pinus cf. sylvestris, Pinus sp. Picea sp., Sorbus aucuparia, Corylus sp., Quercus sp. Abies sp., Picea sp., Larix sp., Pinus sylvestris, cf. Alnus sp., Fraxinus sp. Larix sp., Juniperus sp., Taxus baccata Abies alba, Abies sp., Larix decidua, cf. Larix sp., Picea excelsa, cf. Picea sp., Picea-Larix, Pinus sylvestris, Pinus cembra, Pinus mugo Abies alba, Larix decidua, Picea excelsa, Pinus sylvestris, Pinus cembra, Pinus mugo, Salix sp., Ulmus sp., Juniperus communis, Fagus sylvatica Abies alba, cf. Abies sp., Picea excelsa, cf. Picea sp., Picea-Larix, Pinus sylvestris, Pinus cembra, Pinus mugo, Pinus sp., cf. Pinus sp., Ulmus sp., Populus sp. 30,980 + 360 32,350 + 900 32,600 + 1700 36,000 + 1100 40,173 + 1200 41,400 + 1400 42,900 + 700 26,390 + 270 24,470: 28,220: 29,940: 27,600: 25,570: 25,740: 26,390: 26,640: 26,920: 27,070: 26,550: 26,970: 27,070: 27,080: 27,900: 26,970: 26,390: 26,100: 24,560: 25,790: 25,820: 26,430: 20,270: 27,790: 29,180: 31,700: 32,850: 25,950: 26,000: 26,650: 26,170: 26,580: 30,010: 190 ;370 ;300 ;80 ;280 ;210 ;270 no ;250 ;300 160 160 ;170 ;170 ;550 ;200 190 100 640 ;320 ;170 190 ;210 ;370 460 1000 660 600 ;350 ;230 ;170 460 460 GrN-12606 GrN-12298 GrN-12297 GrN-12605 GrN-14829 GrN-16918 GrN-16920 Q-1044 GrN-6802 GrN-6165 ISGS-1744 GrN11003 GrN-11196 GrN-10525 GrN-13962 GrN-15276 GrN-15277 ISGS-1744 GrN-14831 GrN-15279 GrN-15278 GrN-15325 GrN-15326 GrN-15324 GrN-15327 GrN-15280 GrN-21122 GrN-21123 GrN-14830 GrN-20392 GrN-6857 GrN-1286 GrN-10524 GrN-11004 GrN-6859 GrN-6860 GrN-11189 GrN-6858 GrN-18189 GIN-104 GrN-19539 GrN-20391 KN-1286 KN-286 44,160 + 2150 41,940 + 570 41,810 + 410 34,910 + 570 36,810 + 1590 37,250 + 2260 39,910 + 1350 42,750 + 680 43,550 + 1010 44,290 + 800 30,510 + 450 27,710: 32,260: 33,830: 31,590: 29,200: 29,600: 30,510: 30,800: 30,980: 31,100: 30,720: 31,040: 31,110: 31,120: 32,050: 31,030: 30,560: 30,230: 28,110: 29,580: 29,780: 30,600: 23,820: 31,880: 33,030: 36,310: 37,340: 29,510: 29,880: 30,780: 30,300: 30,570: 33,880: ;730 :590 ;420 ;310 :930 ;710 ;450 ;310 ;330 ;340 ;340 :290 ;280 ;280 ;670 :300 :400 ;470 ; 1110 ;830 ;630 ;380 ;380 ;530 ;680 : 1710 : 1280 : 1130 ;760 ;360 ;480 ;570 ;550 Svoboda and Svoboda (1985) Musil (2003); Damblon et al. (1996) Svoboda and Svoboda (1985); Damblon et al. (1996); Musil (2003) Mason et al. (1994) Opravil (1994); Damblon et al. (1996) ; Damblon (1997) Opravil (1994); Damblon et al. (1996) ; Damblon (1997) ; Slaviková-Veslá (1950) Opravil (1994); Damblon et al. (1996) ; Damblon (1997) aSpaces between dates indicate different occupation layers. assemblages indicating a full-glacial presence of trees (Table 2). The most extensive record is from Dolní Věstonice I, and II (Fig. 1), a cluster of adjacent sites that were settled between ~33 and 28kaBP (-29,000-24,500 14C yr b.p.). These sites have undergone extensive investigation and recently detailed K.J. Willis, T.H. van Andel / Quaternary Science Reviews 23 (2004) 2369-2387 2375 examination and revision of the earlier dating of these sequences has resulted in a detailed Upper Palaeolithic record with excellent temporal and stratigraphic resolution (Damblon et al.,1996; Damblon, 1997). From Dolní Věstonice I, and II over 3000 pieces of macrofossil charcoal have been identified, with site descriptions of parts of trunks up to 40 cm in length and 30 cm in diameter (Opravil, 1994; Musil, 1997, 2003). Species include both needle-leaved {Picea excelsa, Pinus sylves-tris, P. cembra, Pinus mugo, Larix decidua, Juniperus communis, Taxus baccata) and broad-leaved trees {Fagus sylvatica, Ulmus, Populus, Salix and Betula) (Mason et al., 1994; Opravil, 1994) although it is noted that up to 80% of the material appears to be from conifers (in particular pine and spruce). Results from this site led several authors to conclude that former interpretations of this region as being a tree-less steppe during the full-glacial are probably incorrect and that it was a cold-forest steppe dominated by coniferous trees with deciduous trees surviving in more protected areas (Slaviková-Veslá, 1950; Mason et al., 1994; Opravil, 1994; Damblon, 1997; Musil, 2003). In further support of this interpretation are macrofossil charcoal records from three other Upper Palaeolithic sites, namely Pavlov I, Bohunice and Stránská skála (Fig. 1; Table 2) dated to ~30kaBP (-26,000 14C yr b.p.), ~43kaBP (40,000 14C yr b.p.) and -42-37 ka BP (-38,000-32,000 14C yr b.p.), respectively. Again these represent occupation layers in loess sequences where the stratigraphy is well determined and numerous 14C dates have been obtained from identified pieces of macrofossil charcoal with cross-dating carried out by several laboratories (Damblon et al., 1996). Thus even though each of these sites only represents snapshots in time, viewed as a whole they provide a convincing record of the full-glacial presence of trees in the Czech landscape, and depict an environment between 43 and 24kaBP (-40,000-20,000 14C yr b.p.) most closely described as cold-forest steppe (Damblon, 1997). With the presence of various temperate taxa including Quercus, Fraxinus, Fagus and Taxus at Bohunice, Stránská skála and Dolní Věstonice there is again the suggestion that these regions may have been refuges for some temperate deciduous taxa (Damblon, 1997). 2.3. Croatia The site of Sandalja II is a karst cave near Pula in Croatia (Fig. 1) with evidence of Upper Palaeolithic occupation from -29 to 12kaBP (27,800 + 850-10,830 + 70 14C yr b.p.). Analysis and dating of macrofossil charcoal samples from several hearths (Table 3) in the cave indicates a full-glacial presence of Pinus, Abies, Sorbus sp., F. sylvatica, Ulmus, Populus and Rhamnus cathartica between —28 and 24kaBP (-25-21,00 14C yr b.p; Culiberg and Sercelj, 1995). The authors interpret these findings as representing trees in close proximity to the caves occurring in dispersed groups or communities in ecological niches (microrefu-gia) (Culiberg and Sercelj, 1995). 2.4. Hungary A review of macrofossil charcoal sites in Hungary incorporating both data from Upper Palaeolithic archaeological sites and palaeosols in loess sequences (Table 4) has revealed over 29 sites containing macrofossil charcoal assemblages spanning the time interval 29-25 ka BP (-31,500-16,500 14C yr b.p.) (Willis et al., 2000). Macrofossil charcoal wood identified from these sites includes Pinus, Picea, Larix, Betula, Juniperus, Salix and at one site Carpinus betulus with again a predominance of coniferous wood but also the presence of broad-leaved deciduous trees at many of the sites. An advantage of this Hungarian data is that not all of the assemblages have been obtained from archaeological sites, but are records of tree growth that have been preserved within loess sequences. Since these records are Table 3 Identification of tree species/genera in macrofossil charcoal layers from sites in Croatia and 14C dates (based on identified macrofossil charcoal wood within these layers) Site number (with Site type Tree species/genera Published 14C dates Lab. code Cal yr BP Reference to reference to Figs. 1 containing identified in based on macrofossil source of data and 2): Site location: macrofossil macrofossil charcoal charcoal longitude (E)/latitude charcoal layers layers2 (uncalibrated yr (N) Site name b.p.)* (8) 13°53/44°54 Occupation layers Fagus sylvatica 21,740 + 45 Zg 25,260 + 240 Culiberg and Sandalja II in karst cave Sercelj (1995) Sorbus sp. Pinus sp. 23,540 + 180 GrN 26,490 + 220 Rhamnus cathartica 22,660 + 460 Zg 25,890 + 380 Fagus sp., Ulmus sp., 25,340 + 170 GrN 28,870 + 930 Populus sp. aSpaces between dates indicate different occupation layers. 2376 K.J. Willis, T.H. van Andel / Quaternary Science Reviews 23 (2004) 2369-2387 Table 4 Identification of tree species/genera in macrofossil charcoal layers from sites in Hungary and 14C dates (based on the identified macrofossil charcoal wood within these layers) Site number (with Site type Tree species/genera Published 14C Lab. code Cal yr BP References to reference to Figs. containing identified in dates based on source of data 1 and 2) Site macrofossil macrofossil charcoal macrofossil location: charcoal layers layers charcoal longitude (E)/ (uncalibrated yr latitude (N) b.p.) (9) 18°50747°35' Loess deposit Pinus sp., Picea sp., 32,500 ± 2,170 Hv-1776 37,070 + 2570 Geyh et al. (1969) Juniperus sp. (10) 21°05/48°05' Loess deposit Pinus sylvestris 30,174 + 1,101 Deb-4347 34,070 + 1250 Willis et al. (2000) (11) 19°10746°05' Loess deposit Pinus sylvestris, 29,828 + 554 Deb-3058 33,640 + 670 Willis et al. (2000) Betula sp. (12) 19°15747°50' Loess deposit Pinus sylvestris-Pinus 29,800 ± 600 Mo-422 33,600 + 710 Geyh et al. (1969) Occupation layer cembra, Picea sp. 27,200 ± 1400 Deb-3035 30,810 + 1830 Willis et al. (2000) (13) 21°15748°20' Loess deposit Picea-Larix, 28,700 ± 3000 GXO-195 32,480 + 3470 Krolopp (1977) Picea sp. 26,316 + 365 Deb-3051 30,330 + 590 Willis et al. (2000) (14) 21°20748°05' Occupation layer Picea sp. 28,225 + 360 Deb-3035 32,270 + 590 Willis et al. (2000) (15) 19°25747°45' Loess deposit Picea sp. 27,700 ± 300 Deb-1901 31,770 + 470 Willis et al. (2000) (16) 21°25747°55' Loess deposit Picea sp. 27,323 + 644 Deb-2657 31,430 + 730 Willis et al. (2000) (17) 21°10748°00' Loess deposit Picea sp. 27,491 ± 362 Deb-3034 31,560 + 460 Willis et al. (2000) (18) 21°10748°15' Loess deposit Picea sp. 27,251 + 288 Deb-4345 31,260 + 350 Willis et al. (2000) (19) 21°30748°10' Loess deposit Carpinus sp. 26,962 + 657 Deb-5052 30,900 + 700 Willis et al. (2000) (20) 21°15747°55' Loess deposit Picea sp. 26,851 + 398 Deb-3049 31,950 + 560 Willis et al. (2000) (21) 20°05746°55' Loess deposit Pinus sylvestris 26,736 + 629 Deb-4346 30,600 + 740 Willis et al. (2000) (22) 21°25748°10' Loess deposit Picea sp. 26,618 + 532 Deb-3042 30,420 + 810 Willis et al. (2000) (23) 18°00/47°10' Loess deposit Picea-Larix, Pinus 26,350± 3111 Hv-1777 30,030 + 3270 Geyh et al. (1969) sylvestris-Pinus cembra (24) 18°50/47°50' Loess deposit Picea-Larix, Picea sp. 21,165 + 865 Hv-5426 24,530 + 860 Pecsi (1977) 24,030 ± 317 Deb-3353 27,180 + 600 Willis et al. (2000) (25) 19°05746°10' Loess deposit Pinus sylvestris 23,749 ± 494 Deb-3064 26,990 + 670 Willis et al. (2000) (26) 21°20/47°50' Loess deposit Picea sp. 23,571 + 486 Deb-1562 26,660 + 480 Willis et al. (2000) (27) 21°15/47°45' Loess deposit Picea-Larix 20,350± 470 Hv-1775 23,850 + 620 Geyh et al. (1969) (28) 20°50746°30' Lake gytta Picea sp. 23,519± 494 Deb-4350 26,600 + 460 Willis et al. (2000) (29) 20°10746°50' Loess deposit Salix sp. 23,290 ± 285 Deb-4572 26,330 + 260 Willis et al. (2000) (30) 19°15/46°15' Loess deposit Pinus sylvestris 22,110± 300 Deb-1562 25,520 + 320 Willis et al. (2000) (31) 18°30746°00 Loess deposit Pinus sylvestris 21,937 + 252 Deb-3104 25,400 + 310 Willis et al. (2000) (32) 18°10746°55' Occupation layer Picea-Larix, Pinus 21,740± 320 Hv-4189 25,200 + 280 Krolopp (1977) sylvestris-Pinus 21,725 + 560 Hv-2590 25,130 + 550 Marosi and cembra Szilärd (1974) (33) 17°50746°45' Occupation layer Pinus sylvestris- 18,900± 100 GrN-1783 22,210 + 270 Vogel-Waterbolk Pinus cembra, (1964) Picea-Larix (34) 21°30748°30' Occupation layer Pinus sylvestris- 18,700± 1900 A-518 21,700 + 2180 Vertes (1964) Pinus cembra 17,050± 350 GrN-4038 19,980 + 660 Picea-Larix (35) 19°30746°10' Occupation layer Pinus sylvestris-Pinus 18,080± 405 Hv-1619 21,170 + 580 Dobosi (1967) cembra, Picea-Larix (36) 18°00746°40' Loess deposit Pinus sylvestris-Pinus 17,760± 150 GrN-1957 20,940 + 410 Gabori-Csänk cembra, Picea-Larix (1960) (37) 19°15747°50' Occupation layer Pinus sylvestris- 16,750± 400 — 19,560 + 680 Geyh et al. (1969) Pinus cembra, Picea-Larix, Betula sp. K.J. Willis, T.H. van Andel / Quaternary Science Reviews 23 (2004) 2369-2387 2377 not just associated with dates of human occupation, this has important implications for understanding the temporal extent of trees on the eastern European landscape during the full-glacial. Evidence from 34 14C dates clearly indicates the presence of trees in Hungary during each 1000-year time-slice between 37 and 20kaBP (35,500-16,500 14C yr b.p.) with no obvious clustering of dates during the warmer Dansgaard/ Oeschger (D/O) events. In addition, at several sites there is evidence for in situ growth of trees with charcoal silhouettes of tree trunks and extensive tree root systems in the cleaned loess face. Conclusions drawn from this study suggest that the full-glacial and lastglacial maximum (LGM) landscape was probably a mixture of forest-steppe—the nearest present-day equivalent being open boreal-type forest (Willis et al., 2000). Other data associated with this study to support this interpretation included fossil molluscan assemblages at many of the sites containing a number of species that at present occupy a woody habitat. 2.5. Moldova The site of Cosautsi is an Upper Palaeolithic site located on the west bank of the Dniester river in northeast Moldova. It comprises an 18.5m thick stratigraphic sequence in loess with occupation layers indicating that the site was occupied for approximately 3000 years between -23 and 18kaBP (-19,500-16,000 14C yr b.p). Numerous 14C dates have been obtained for these occupation layers, with many based on macrofossil charcoal fragments. Most recently a new dating project was carried out on the Cosautsi sequence (Haesearts et al., 1998) involving detailed identification of the macro- fossil charcoal to be dated and also cross-dating of all these samples with dates on bone fragments found in the same stratigraphic layers. Fourteen dates on macrofossil charcoal were obtained (Table 5) with the majority of the fragments identified as Picea sp. although in two of the layers P. cembra and Salix sp. were also noted. The predominance of Picea in this macrofossil charcoal record, plus the lack of any temperate tree taxa and a somewhat depauperate flora in comparison to records from other Upper Palaeolithic sites further west has led to the suggestion that although there were trees in the vicinity of this site during the full-glacial, the environment was harsh and unable to provide refugia for temperate flora (Damblon, 1997). 2.6. Poland The site of Spadzista in Southern Poland is another Upper Palaeolithic occupation sequence situated in loess. Occupation at this site occurred between 28 and 26kaBP (-24,000-23,000 14C yr b.p.: Damblon et al., 1996; Musil, 2003). There is also a date from a humic soil further down in the sequence ~36kaBP (~32,000 14C yr b.p.). Macrofossil charcoal fragments identified from these two layers on which the dates were based include Pinus, Larix and Abies (Table 6), indicating a full-glacial presence of these species in Southern Poland. 2.7. Romania The Upper Palaeolithic site of Mitoc Malu Galben is situated east of the Carpathians on the river Prut (Fig. 1) and is comprised of three main occupations Table 5 Identification of tree-species/genera in macrofossil charcoal layers from Cosautsi, Moldova and 14C dates (based on the identified macrofossil charcoal wood within these layers) Site number (with reference to Figs. 1 and 2): Site location: (longitude E/latitude (N) Site name Site type containing macrofossil charcoal layers Tree species/genera identified in macrofossil charcoal layers Published 14C dates Lab. based on macrofossil charcoal (uncalibrated yr b.p.) code Cal yr BP References to source of data (38) 28°18748°30' Occupation layer Picea sp., Pinus 16,050 + 170 GrA-4209 18,520 + 240 Damblon and Cosautsi situated upon cembra, Salix 17,130 + 180 GrA-5217 20,060 + 480 Haesaerts (1997); loess 17,230 + 140 GrN-217792 20,210 + 450 Haesaerts et al. 17,620 + 210 GrN-21793 20,720 + 490 (1998) 17,910 + 80 GrN-21360 21,210 + 280 18,030 + 150 GrN-21359 21,330 + 280 17,780 + 190 GrA-7554 20,940 + 440 17,950 + 100 GrN-21794 21,260 + 270 18,260 + 210 GrA-5218 21,560 + 270 19,200 + 130 GrN-21361 22,560 + 360 19,120 + 100 GrA-7555 22,470 + 340 19,440 + 100 GrA-6746 22,800 + 380 19,070 + 100 GrA-7557 22,410 + 330 19,410 + 100 GrN-21795 22,770 + 380 2378 K.J. Willis, T.H. van Andel / Quaternary Science Reviews 23 (2004) 2369-2387 Table 6 Identification of tree species/genera in macrofossil charcoal layers from Spadzitsta Street, Southern Poland and macrofossil charcoal wood within these layers) C dates (based on the identified Site number (with reference to Figs. 1 and 2): Site location: longitude (E)/latitude (N) Site type containing macrofossil charcoal wood Tree species/genera identified in macrofossil charcoal layers Published 14C dates Lab. based on macrofossil charcoal (uncalibrated yr b.p.) code Cal yr. BP Sources of data (39) 19°55750°03' Spadzitsta Street Occupation layer situated upon loess Pinus sp., Larix sp., Abies sp. 23,040 + 170 24,380 + 180 GrN-6636 GrN-11006 26,170: 27,590: ;220 700 Damblon et al. (1996); Musil (2003) Table 7 Identification of tree species/genera in the macrofossil charcoal layers from Mitoc Malu Galben, Romania and 14C dates (based on the identified macrofossil charcoal wood within these layers) Site number (with Site type Tree species/genera Published 14C dates Lab. code Cal yr BP References to source reference to Figs. 1 containing identified in based on of data and 2): Site macrofossil macrofossil charcoal macrofossil location: longitude charcoal layersa layers charcoal (E)/latitude (N) Site (uncalibrated yr name b.p.f (40) 27°02748°07' Occupation layers Picea sp. 20,300 + 700 GrN-14031 23,720 + 850 Damblon et al. (1996); Mitoc Malu Galben situated upon Damblon (1997) loess Picea sp., Alnus sp. 23,850 + 100 GrA-1353 26,700 + 220 Picea sp. 23,390 + 280 GrN-20438 26,390 + 260 Picea sp. 23,830 + 330 GrN-14034 26,820 + 400 Picea sp., Pinus 23,990 + 250 GrN-20439 26,960 + 410 cembra Picea sp., Alnus sp. 25,610 + 500 GrN-20440 29,120 + 1120 Picea sp., Alnus sp. 26,180 + 290 GrN-18811 30,220 + 580 Picea sp. 26,500 + 460 GrN-18815 30,470 + 610 Picea sp. 26,020 + 650 GrN-18880 29,550 + 1170 Picea sp., Alnus sp. 26,380 + 600 GrN-18881 30,120 + 900 Picea sp. 26,300 + 450 GrN-18879 30,210 + 720 Picea sp. 25,080 + 500 GrN-18882 28,570 + 1080 Picea sp. 26,110 + 1050 GrN-18883 29,460 + 1510 Picea sp. 28,910 + 480 GrN-12636 32,810 + 700 Picea sp., Alnus sp. 25,380 + 120 GrA-1355 28,950 + 900 Picea sp. 31,850 + 800 GrN-12637 36,410 + 1610 Picea sp. 31,000 + 330 GrA-1648 34,920 + 550 Picea sp. 31,100 + 900 OxA-1646 35,090 + 940 Picea sp. 30,240 + 470 GrN-20443 34,180 + 570 Picea sp. 31,160 + 570 GrN-20770 35,060 + 670 Picea sp. 30,920 + 390 GrN-20442 34,870 + 580 Picea sp. 31,160 + 550 32,730 + 220 GrN-20444 GrA-1357 35,060 37,250 + 660 + 1100 aSpaces between dates are used to indicate different occupation layers. phases between 37 and 33kaBP (33,000-29,500 14C yr b.p.), 30-28 ka BP (27,500-25,000 14C yr b.p.) and 27-26 ka BP (24,000-23,000 14C yr b.p.) (Damblon et al., 1996). Similar to the majority of sites described above, this is a loess sequence containing well-determined stratigraphic horizons for the occupation layers. This site has been described as one of the most dated in Central Europe (Damblon et al., 1996) with over 46 14C determinations. Recent scrutiny of these dates plus the provision of additional 14 dates has led to a well-constrained temporal sequence containing 23 dates carried out on identified pieces of macrofossil charcoal assemblages (Damblon, 1997; Table 7). Similar to the records from Moldova the majority of the macrofossil charcoal obtained from this sequence was Picea with Alnus charcoal fragments found in five samples and P. K.J. Willis, T.H. van Andel / Quaternary Science Reviews 23 (2004) 2369-2387 2379 cembra in one sample. These results have led to the suggestion that spruce was the main tree species in the woody component of the landscape near Mitoc for the period between 37 and 26kaBP (-32,000-23,000 14C yr b.p.) with local occurrences of alder growing along the river Prut (Damblon, 1997). The lack of temperate deciduous trees in the macrofossil charcoal assemblage from this site, again similar to the record from Moldova, is taken to suggest that this full-glacial environment of this region was too harsh to provide refugia for temperate elements of the flora. In summary, there is now macrofossil charcoal evidence from at least 40 sites in central and eastern Europe (Figs. 1 and 2) to indicate the presence of trees growing either in situ or in close proximity to the sites during the full-glacial between 44 and 19kaBP (-41,000-16,000 14C yr b.p.). At all of these sites, the macrofossil charcoal assemblages have been used to identify the trees present and provide the material for 14C dating. Over 20 different tree taxa have been identified and 151 14C dates (often cross-dated with bone material from the same sequence) obtained from seven different dating laboratories (Tables 1-7). Although the majority of the wood identified in this macrofossil charcoal record is from coniferous trees there is also a notable presence of broad-leaved temperate deciduous trees including Fagus, Ulmus, Quercus, Corylus, Sorbus, Carpinus, Rhamnus and Populus at some sites (Figs. 1 and 2). The absence of these types from other sites is probably due to more severe climatic conditions although in some cases it might also be due to a sampling bias. In several site reports, e.g. it is stated that pieces of temperate deciduous taxa were identified in the assemblages but discarded from further analyses and dating since it was presumed that they must represent contamination of the sample (see, e.g. Damblon et al., 1996, 1997). 3. Trees or no trees: other lines of fossil plant evidence At many of the sites described above, pollen has been identified in the same loess/cave sequences as the macrofossil charcoal. These records, however, have not been included in this review. This is due to the inherent problem of obtaining a stratigraphically meaningful pollen record in such sediments (Turner, 1985; but see Carrion et al., 1999; Navarro Camacho et al., 2000). Lake or peat sediments containing pollen sequences and extending back to the full-glacial are SITE LOCATIONS .39 4si 13 , r22 20 27' \ 23-./ 21 38 40 \ -------25"E- 25,000-20,000 Cal.yr B.R 20,000- 15,000 Cal.yr B.P. KEY TO TREE SYMBOLS ^ Abies ^ Abies ^alba l^Alnus {) Betula ^ Corylus ^ Carpinus f» Fagus T sylvatica |j| Fraxinus ^ Juniperus 9Juniperus communis ^ Larix ± Larix T decidua ^ Larix/ I Picea a Picea/ i Larix Picea Quercus A Picea • excelsa ^ Pinus ▲ Pinus ™ cembra A Pinus TT mugo A Pinus " sylvestris Populus k Rhamnus * cathartica /Hfl* Salix Sorbus I*. Sorbus aucuparia f) Taxus T baccata ^ Ulmus Fig. 2. Location of the 40 full-glacial macrofossil charcoal sites plus the tree species/genera identified at each site for the time intervals 25,000-20,000 and 20,000-15,000 ka BP (full citation to sites is given in text and Tables 1-7). Numbers correspond to site names that are detailed in Tables 1-7. 2380 K.J. Willis, T.H. van Andel / Quaternary Science Reviews 23 (2004) 2369-2387 more stratigraphically secure, but are rare in central and eastern Europe (see, e.g. the coverage of sites in Huntley et al., 2003). However, there are several notable exceptions that add further support to the findings described above regarding a presence of trees in central and eastern Europe during the full-glacial. The most notable of these is from Bulhary in the Czech Republic where pollen and macrofossil analyses have been carried out on a buried peat dated to ~28kaBP (25,657 + 2750 14C yr b.p.) (Rybníčková and Rybníček, 1991). The pollen diagram indicates a coniferous forest containing P. sylvestris, Pinus cf. mugo, P. cembra, Picea, Larix, J. communis and Betula. There is also evidence for the scattered presence of temperate deciduous trees including Ulmus, Acer, Cor-ylus, Quercus and Tilia. Plant macrofossils obtained from these peat deposits, which from numerous other studies are known to be a reliable indicator of a very local presence (Birks, 2003), include leaves, seeds and wood of Betula cf. pubescens and Salix sp. Three lake sites that contain pollen diagrams extending back into the end of the LGM at ~20kaBP (-18,000 14C yr b.p.) are Svarcenberk in the Czech Republic (Pokorný and Janlovská, 2000), Sarret in Hungary (Willis et al., 2000) and Steregoiu in NW Romania (Bjorkman et al., 2002, 2003). In all three cases the pollen diagrams indicate an open-forested vegetation during the LGM (with tree pollen accounting for over 60% of the total pollen). The pollen diagrams indicate that Pinus and Betula were predominant in this environment but there was also (in Hungary) a low but persistent presence of other tree taxa including Juniperus, Salix, Alnus, and Larix. In Svarcenberk plant macrofossils have been obtained from the same sequence indicating Betula nana seeds and Salix wood during the LGM. From the Czech Republic and Poland numerous finds of Larix, Betula, P. excelsa and P. cembra seeds and needles in early Lateglacial lake sediments (Allerod: -14,000 BP) (Jankovská, 1988, 1998) led to the suggestion that the Lateglacial vegetation of this area of eastern Europe bore a close analogy to the present-day Siberian boreal (taiga) forest zone (Jankovská, 1998). In addition, pine megafossils (logs) dating to ca 17,000 ka BP have been recently found at Magherus in Northwestern Romania (B. Wohfarth pers. comm., 2004). 4. Trees or no trees: the genetic evidence Another line of evidence used in the interpretation of former distributions of vegetation is that of genetics. This involves the extraction and amplification of genetic signatures (DNA or other genetic markers) from present-day species in order to understand their ancestry, past movement of populations, and present-day patterns of genetic diversity. Much of this work is based on the assumption that in consequence of prolonged isolation during glacial periods, extant tree populations situated close to refugia should be highly divergent with a declining diversity away from the refugia indicating post-glacial colonization and successive founder events (Hewitt, 1996, 2000, 2004). Over the past 10 years, there have been a number of important publications examining the genetic legacy of European tree species including, e.g. Quercus, Fagus, and Abies alba (Konnert and Bergmann, 1995; Demesure et al., 1996; Dumolin-Lapegue et al., 1997) with the results from a number of these studies suggesting that indeed clear genetic differentiation between different populations of the same species exist. In many cases this genetic differentiation appears to have a distinct geographical coverage in southern Europe and is often used in conjunction with the fossil evidence to argue for the importance of refugia in their genetic diversification (e.g. Taberlet and Bouvet, 1994; Hewitt, 1996, 2000; Taberlet and Cheddadi, 2002; Tzedakis et al., 2002). Amongst all the evidence supporting the southern European refugial model there is, however, an increasing amount of genetic evidence that indicates a somewhat different full-glacial distribution of some tree types and subsequent post-glacial migration routes. In particular, it has been demonstrated that several taxa including Picea abies, Betula, Alnus, Populus and Salix (Lagercrantz and Ryman, 1990; Palme et al., 2002, 2003a, b; Petit et al., 2003; Lascoux et al., 2004) do not conform to a southern European refugial pattern but demonstrate a more northerly and diffuse distribution. Such results have led to the suggestion that this is a reflection of full-glacial survival of these species during the LGM in northerly refugia including central and eastern Europe (Palme and Vendramin, 2002; Stewart and Lister, 2002; Palme et al., 2003a, b; Petit et al., 2003; Lascoux et al., 2004). 5. Trees or no trees: could the trees identified in the macrofossil charcoal record have survived the full-glacial and Lateglacial environments of central and eastern Europe? The records of full-glacial tree growth in central and eastern Europe presented in this review are clearly at odds with the most recent landscape reconstructions based on long-sequence pollen-based biomes which suggest an open, essentially tree-less environment (see, e.g. Alfano et al., 2003; Huntley et al., 2003). It is therefore not surprising that palynologists have tended to discount occasional dated macrofossil charcoal records from one site or another. However, as a whole those records (Figs 1 and 2; Tables 1-7) alongside equally piecemeal evidence from macrofossils and pollen provide a convincing case for the survival of trees in K.J. Willis, T.H. van Andel / Quaternary Science Reviews 23 (2004) 2369-2387 2381 many parts of central and eastern Europe throughout the cold stages of the full and Lateglacial (~42-18kaBP). So could trees have withstood the predicted climatic extremes of the full and Lateglacial in central and eastern Europe? Or were these climates too severe to have allowed trees in all but a few small microenvironmentally favourable locations? In order to address these questions we need to examine both the climatic simulations for the full and Lateglacial in central and eastern Europe and the physiological tolerances of the trees identified in the macrofossil charcoal record. Recently, high-resolution climate simulations have been obtained for the early full-glacial and LGM climates as part of the Stage 3 Project (Barron and Pollard, 2002; van Andel, 2002; Barron et al., 2003; Pollard and Barron, 2003) to provide millennium-scale 'snapshot' simulations representing the main states of the glacial climate by means of: a "modern" state tested satisfactorily against modern data, a "warm state" typical for the 45-38 ka BP interval, and a "LGM cold state for the acme of the LGM (27-~18kaBP). The European climate then—as it does today, displayed two main continental-scale regions in addition to the Mediterranean zone: a north-south gradient from arctic to cold temperate climates north of the trans-European mountain barrier and a west-east transition from the maritime Atlantic climate to the continental climate of eastern Europe; both are clearly reflected in the snapshot simulations (Barron et al., 2003). The 'LGM cold simulation', for example predicted winter (December, January, February) surface temperatures that were significantly cooler than the modern simulation, with much of northern Europe 10-20 °C cooler, southern Europe and parts of central Europe 7-10 °C, and only Spain and the area around the Black Sea a mere 2-4 °C cooler (Barron and Pollard, 2002). Winter and summer precipitation also had a strong east-west gradient, bringing to central and eastern Europe more precipitation than to the three southern peninsulas of Europe. In general, the 'early warm' and 'LGM cold' simulations are therefore much as they had been expected from previous anecdotal information. The climate simulated for the interval between the early warm and LGM cold states (~38-27kaBP), however, predicted temperature and precipitation regimes for central and eastern Europe quite different than those permitted in the tree-less tundra concept suggested by the Stage 3 pollen-based biomes; predicted temperatures and precipitation regimes for central and eastern Europe were much warmer than anticipated with, for example winter temperatures only 4-7 °C cooler (Fig. 4: Barron and Polard, 2002; Pollard and Barron, 2003); it also suggested that central and eastern Europe had much more rainfall both during the winter and summer than southern Europe. Since this simula- tion did not lit with the pollen evidence of a tree-less tundra, the climate simulation for ~38-27kaBP was rejected (Barron and Pollard, 2002; Alfano et al., 2003; Barron et al., 2003). It is interesting to note however that the palynological data used in the Stage 3 Project and the underlying tree-less tundra concept, covered only the maritime and Mediterranean regions and not continental Europe with the exception of one site in Estonia at 39°E, 56°N (Liivrand, 1990, 1991). So could the trees identified in the macrofossil charcoal records have survived the extremes of the simulated full and Lateglacial climates? Environmental factors that potentially could have limited tree growth include low temperatures, reduced amounts of precipitation, cold podzolic soils with permafrost in some regions, shorter growing seasons and low atmospheric C02 concentrations. Dealing with the latter point first, laboratory experiments have demonstrated that the physiological effects of C02 limitation only become predominant when other nutrients, water and light are available at high levels (Tilman, 1993). Therefore, although lower atmospheric C02 during the LGM may have had an important influence on low-latitude environments (e.g. Street-Perrott et al., 1997; Cowling et al., 2001; Harrison and Prentice, 2003), its impact on mid-to-high latitude vegetation would have been much less pronounced (Bennett and Willis, 2000). When considering the impact of the other environmental factors (reduced temperature, precipitation and cold podzolic soils), it is interesting to note the present-day tolerances in central and northern Europe and the Eurasian boreal forest of many of the species identified in the macrofossil charcoal record (Table 8; Kienast, 1987; Ellenberg, 1988; Prentice and Helmisaari, 1991; Nikolov and Helmissari, 1992; Larcher, 1995). All could have coped with the conditions predicted by the rejected 38-27 ka BP simulations but more importantly, also the cold conditions predicted for the LGM (27-18 ka BP) (Table 8). A reduced growing season and permafrost might have caused more limitations to temperate tree growth but it is probable that such conditions would not have had a significant impact on the physiological capabilities of the hardened boreal species. Neither climatic scenario, the LGM cold nor the rejected 38-27 ka BP snapshots can therefore be used to assert that taiga-forests could not have existed in central and eastern Europe between ~40 and ~20kaBP. In fact, the coupled climate vegetation model (Haxeltine and Prentice, 1996) for the 38-27 ka BP interval shows as a dominant vegetation type in central and northern Europe evergreen taiga/montane forest (Alfano et al., 2003: Fig. 6) as our data also suggest. Nonetheless, the simulation for 37-28 ka BP remains contentious and the pollen data continue to be seen as demanding a predominantly stepic herbaceous vegetation of low productivity for central and eastern Europe (Huntley 2382 K.J. Willis, T.H. van Andel / Quaternary Science Reviews 23 (2004) 2369-2387 Table 8 Present day tolerances of minimum growing temperature, drought and nutrient availability for tree species/genera recognised in the macrofossil charcoal record (Tables 1-7) Tree species/ Tcold Drought Ability to grow on genus (°C)» resistance (1: very poor soils0 recognised in low to 5: very (1 — tolerant, macrofossil high) 2 — intermediate, charcoal record2 3 — intolerant) Needle-leaved Abies alba -5 3 — Abies sibirica -35 2 2 Juniperus — 3 1 communis Larix decidua -10 2 — Larix gmelinii -45 4 1 Larix sibirica -33 5 1 Larix sukaczewii -22 3 1 Picea abies -17 1 2 Picea ajanensis -38 2 2 Picea obovata -40 1 2 Pinus cembra -10 5 — Pinus mugo — 5 1 Pinus pumila -45 2 1 Pinus sibirica -35 2 2 Pinus sylvestris -40 5 1 Taxus baccata -4 4 1 Broad-leaved Acer platanoides -15 3 3 Alnus incana — 1 1 Alnus glutinosa -15 1 1 Betula pendula -40 5 1 Betula pubescens -40 1 1 Betula verrucosa — 2 2 Carpinus betulus -8 3 2 Corylus avellana -15 — 2 Fagus sylvatica -3 2 1 Fraxinus -16 2 3 excelsior Populus tremula -40 4 1 Populus nigra — 1 1 Quercus robur -16 5 1 Quercus petrea -4 3 2 Salix alba — 1 1 Sorbus aria — 4 1 Sorbus — 4 1 aucuparia Tilia cordata -18 3 3 Ulmus glabra -15 — 3 Ulmus scabra — 3 1 — Data not available. aWhere taxa are only recognised to genera in the macrofossil charcoal record, species of this genera presently growing in central and northern Europe and the Eurasian boreal forest (for which there are data available), (Ellenberg, 1988; Keinast, 1987; Prentice and Helmisaari, 1991; Nikolov and Helmisaari, 1992) are included for comparison. bNumbers in bold represent lowest possible average January temperature for species (from Keinast, 1987); those in italics represent mean temperature of the coldest month at the border of the species' geographical range (from Prentice and Helmisaari, 1991; Nikolov and Helmisaari, 1992). cEstimate based on tolerance to low nitrogen availability. et al., 2003; Huntley and Allen, 2003), in direct conflict with the other evidence presented in this paper as well as the well-documented mammalian fauna of the period (van Andel, 2003b). 6. The temporal and spatial distribution of trees and the high-frequency oscillations of the Lateglacial climate We regard the evidence presented in Tables 1-8 as conclusive for the existence during the last glaciation of a wide range of tree species in central and eastern Europe, so compelling a substantial revision of current views on the glacial landscape, its resources and its capacity to support human exploitation in continental Europe. This raises a number of important questions. Was the presence of trees limited to warm phases of the D/O oscillations, retreating at each stadial and returning at the next interstadial? Did trees occur only in small enclaves in favoured locations or gallery woodland along rivers or did extensive open woodland (taiga) exist? What were the glacial climate variations indicated by the mix of some nine coniferous and 20 broadleaf tree species recorded for the interval ~40-16kaBP at various time and locations (Figs. 1 and 2)? And what are the consequences of the presence of trees for the carrying capacity in human terms of the full-glacial landscape. Those questions and many others are beyond the scope of our paper or even beyond currently available evidence, except for the question as to whether the presence of trees was limited to short, warm D/O events or was continuous. A manuscript on high-resolution calibration of the Greenland time-scale generously made available to us by Nick Shackleton (Shackleton et al., 2004) has enabled us to distinguish with some confidence those macrofossil charcoal dates that are from warm interstadials from those present in cold stadials for Greenland Interstadials 2-7 (GIS: Shackleton et al., 2004; Fig. 2). Beyond GIS 7, the sparse data do not permit this determination. Table 9 shows all sites with large sets of charcoal dates and a few others for comparison, plotted against a millennial time-scale for the late Oxygen Isotope Stage 3 and the LGM (32—16 uncal. kaBP). Because of differences between the 14C calibration method used in this paper and that of Shackleton et al. (2004), all charcoal and D/O cycle dates are uncalibrated. The sites listed in Table 9 are all south of the 49°N parallel and east of 15°E and therefore well within the continental climate region. Because the transitions from interstadial to stadial and back are steep (see Shackleton et al., 2004: Fig. 2), spanning at most a few centuries, the numbers in each category can be reasonably estimated except very near boundaries. K.J. Willis, T.H. van Andel / Quaternary Science Reviews 23 (2004) 2369-2387 2383 Table 9 Chronological distribution across Dansgaard/Oeschger oscillations of14C-dated charcoal samples on 14C time scale; GIS (Greenland Interstadial) interstadial bases are also in uncalibrated 14C time Age uncal. ka BP Interstadial _Base_ <19. 9 20.0 21.0 22.0 23.0 24.0 25.4 26.0 27.0 28.0 29.6 30.0 31.0 32.0 GIS 2 19.9 GIS 7 31.9 14 : GIS 5 28 5 1 1 2 1 GIS 6 29 5 Site codes: A: Willendorf; B: Dolní Věstonice I+ 11; C: Pavlov; D: Šandalja; E: Hungarian region; F: Cosautsi; G: Mitoc Malu Galben. Dated boundaries labelled GIS are the interstadial bases 2-7 (after Shackleton et al., 2004: Table 1). Interstadials shaded. From our best data set (Table 9) the overwhelming majority of dates are found in cold events. Among the 121 dates for the seven sites displayed, only 16 are securely located within interstadials—as expected, given the difference in duration of stadials and interstadials. It is obvious that the presence of coniferous as well as broadleaf trees was continuous throughout the interval for which there are adequate data (~3 5-20 uncal. kaBP). The trees did not migrate northward during warm D/O events, only to perish during the subsequent cold event; they were always there. A decline in numbers towards the LGM would be expected and can indeed be seen in most sites together with a northward increase in the dominance of coniferous over broadleaf trees (Figs. 1 and 2; Table 9). The Hungarian region and the Cosautsi site in Moldavia, however, are striking exceptions that indicate a milder LGM there. 7. Conclusions A review of the published macrofossil charcoal evidence from 40 localities in central and eastern Europe has revealed evidence for the presence of at least 20 different tree types growing in this landscape during the last full-glacial interval. In total, 151 14C dates have been obtained from the macrofossil charcoal assemblages that date these samples between ~42 and 19 kaBP. Identified tree types are predominantly needle-leaved, including Picea, Pinus, Larix, Juniperus, although broad-leaved types including Salix, Betula, Fagus, Ulmus, Quercus, Corylus, Sorbus, Carpinus, Rhamnus and Populus have been found at some localities. Additional data from plant macrofossils and pollen in lake sequences, although much less extensive in temporal and spatial coverage, adds further support to suggestion of the full-glacial presence of trees in central and Eastern Europe. In addition, genetic evidence from extant European tree populations suggests that many boreal tree types do not have a genetic distribution pattern that conforms to a model predicting full-glacial isolation in southern European refugia. The evidence presented in this paper makes it difficult to support a European full-glacial refugial model that excludes trees from central and Eastern Europe. It is even more difficult, however, to establish from the fossil evidence alone whether the trees grew in isolated pockets on an otherwise open tundra landscape or in an open-taiga forest. Recent palaeoclimatic and biome model simulations for the last full-glacial, however, indicate that, contrary to previous interpretations, the central and eastern European landscape could have supported taiga forest, called "a speculative model" by Huntley and Allen, 2003: Fig. 6.12). The evidence presented in this paper fully supports these model interpretations and leads to the undoubtedly somewhat controversial suggestion that during the last full-glacial interval the central and eastern European landscape was partly K.J. Willis, T.H. van Andel / Quaternary Science Reviews 23 (2004) 2369-2387 2384 covered by taiga/montane woodland, which in some regions also contained isolated pockets of temperate trees. Acknowledgements We are grateful to the many members of the Stage 3 Project of the Department of Earth Sciences, Cambridge University, who generously helped us in compiling the data on which this paper rests. This paper has also benefited from interesting and helpful comments from the following: Keith Bennett, Geoff Bailey, Freddie Damblon, Nick Shackleton, Ian Woodward and Barbara Wohlfarth. We would also like to thank Ailsa Allen and the cartography department in the School of Geography and the Environment, University of Oxford, for drawing of the figures. References Alfano, M.J., Barron, E.J., Pollard, D., Huntley, B., Allen, J.R.M., 2003. Comparison of climate model results with European vegetation and permafrost during Oxygen Isotope Stage 3. Quaternary Research 59, 97-107. Badal, E., Bernabeu, J.Y., Vernet, J.L., 1994. Vegetation changes and human action from the Neolithic to the Bronze Age (7000-4000 BP) in Alicante, Spain, based on charcoal analysis. Vegetation History and Archaeobotany 3, 155-166. Bailey, G.N., 1997. Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece. McDonald Institute for Archaeological Research, Cambridge. Bailey, G.N., 2004. Site catchment analysis. In: Renfrew, C, Bahn, P. (Eds.), Key Concepts in Archaeology. Routledge, London, in press. Barron, E.J., Pollard, D., 2002. High-resolution climate simulations of Oxygen Isotope Stage 3 in Europe. Quaternary Research 58, 296-309. Barron, E., van Andel, T.H., Pollard, D., 2003. Glacial environments II. Reconstructing the climate of Europe in the Last Glaciation. In: van Andel, T.H., Davies, S.W. (Eds.), Neanderthals and Modern Humans in the European Landscape during the Last Glaciation. McDonald Institute for Archaeological Research, Cambridge, pp. 57-78. Bennett, K.D., Willis, K.J., 2000. Effect of global atmospheric carbon dioxide on glacial-interglacial vegetation change. Global Ecology and Biogeography 9, 355-361. Bennett, K.D., Tzedakis, P.C., Willis, K.J., 1991. Quaternary refugia of north European trees. Journal of Biogeography 18, 103-115. Birks, H.H., 2003. The important of plant macrofossils in the reconstruction of Lateglacial vegetation and climate: examples from Scotland, western Norway and Minnesota, USA. Quaternary Science Reviews 22, 453-473. Bjorkman, L., Feurdean, A., Cinthio, K, Wohlfart, B., Possnert, G., 2002. Late Glacial and early Holocene woodland development in the Gutaiului Mountains, NW Romania. Quaternary Science Reviews 21, 1039-1059. Bjorkman, L., Feurdean, A., Wohlfarth, B., 2003. Late-Glacial and Holocene forest dynamics at Steregoiu in the Gutaiului Mountains, Northwest Romania. Review of Palaeobotany and Palynology 124, 79-111. Carrion, J.S., 2002. Patterns and processes of Late Quaternary environmental changes in a montane region of southwestern Europe. Quaternary Science Reviews 21, 2047-2066. Carrion, J.S., Munuera, M., Navarro, C, Burjachs, F., Dupre, M., Walker, M.J., 1999. The palaeoecological potential of pollen records in caves: a case study of Mediterranean Spain. Quaternary Science Reviews 18, 1061-1075. COHMAP Members, 1988. Climatic changes of the last 18,000 years: observations and model simulations. Science 241, 1043-1052. Cowling, S.A., Maslin, M.A., Sykes, M.T., 2001. Paleovegetation simulations of lowland Amazonia and implications for neotropical allopatry and speciation. Quaternary Research 55, 140-149. Culiberg, M., Sercelj, A., 1995. Anthracotomical and palynological research in the palaeolithic site Sandalja II (Istria, Croatia). Razprave IV, Razreda Sazu 3, 49-57. Damblon, F., 1997. Palaeobotanical study of representative Upper Palaeolithic sites in the central European Plain: a contribution to the SC-004 project. Prehistoire Europeenne 11, 245-253. Damblon, F., Haesaerts, P., 1997. Radiocarbon chronology of representative Upper Palaeolithic sites in the central European Plain: a contribution to the SC-004 project. Prehistoire Europeenne 11, 255-276. Damblon, F., Haesaerts, P., 2002. Anthracology and radiochronology of the Upper Pleistocene in the loessic areas of Eurasia. In: Thiebault, S. (Ed.), Charcoal Analysis: Methodological Approaches, Palaeoecological Results and Wood Uses, pp. 65-71. Damblon, F., Haesaerts, P., Van der Plicht, J., 1996. New datings and considerations on the chronology of Upper Palaeolithic sites in the Great Eurasiatic Plain. Prehistoire Europeenne 9, 177-231. Davies, S.W., Valdes, P.V., Ross, C, van Andel, T.H., 2003. The human presence in Europe during the last glacial period: III Site clusters, regional climates and resources. In: van Andel, T.H., Davies, S.W. (Eds.), Neanderthals and Modern Humans in the European Landscape during the Last Glaciation. McDonald Institute for Archaeological Research, Cambridge, pp. 191-220 Chapter 11. Demesure, B., Comps, B., Petit, R.J., 1996. Chloroplast DNA phylogeography of the common beech (Fagus sylvatica L.) in Europe. Evolution 50, 2515-2520. Dobosi, V., 1967. U'j felsö-paleolit telep zs Alföldön Contribution. Archeolögiai Ertesitö 94, 184-193. Dumolin-Lapegue, S., Demesure, B., Fineschi, S., Le Corre, V., Petit, R.J., 1997. Phylogeographic structure of white oaks throughout the European continent. Genetics 146, 1475-1487. Ellenberg, H., 1988. Vegetation Ecology of Central Europe, fourth ed. Cambridge University Press, Cambridge. Figueiral, I., Terral, J.-F., 2002. Late Quaternary refugia of Mediterranean taxa in the Portuguese Estremadura: charcoal based paleovegetation and climatic reconstruction. Quaternary Science Reviews 21, 549-558. Follieri, M., Giardini, M., Magri, D., Sadori, L., 1998. Palynostrati- graphy of the last glacial period in the volcanic region in central Italy. Quaternary International 47.48, 3-20. Frenzel, B., 1992. European Climate Reconstructed from Documentary Data, Methods and Results 1. Gustav Fischer Vela, Stuttgart. Frenzel, B., Troll, C, 1952. Die Vegetationszonen des nördlichen Eurasiens während der letzen Eiszeit. Eiszeitalter und Gegenwart 2, 154-167. Gabori-Csänk, V., 1960. A sägväri telep abszolüt kormeghatärozäsa. Archeolögiai Ertesitö 18, 5-18. Geyh, M.A., Schweitzer, F., Vertes, F., Vogel, I.C., 1969. Neue chronologische Angaben der Würm-Vereisung in Ungarn. Föl- drajzi Ertesitö 18, 5-18. Guthrie, R.D., 1990. Frozen Fauna of the Mammoth Steppe: The Story of Blue Babe. University of Chicago Press, London. K.J. Willis, T.H. van Andel / Quaternary Science Reviews 23 (2004) 2369-2387 2385 Guthrie, R.D., 2000. Origin and cause of the mammoth steppe: a story of cloud cover, woolly mammal tooth pits, buckles, and inside-out Beringia. Quaternary Science Reviews 20, 549-574. Guthrie, D., van Kolfschoten, T., 1999. Neither warm and moist nor cold and arid: the ecology of the Mid Upper Palaeolithic. In: Roebroeks, W., Mussi, W., Svoboda, J., Fennema, K. (Eds.), Hunters of the Golden Age, 31. Analecta Prehistorica, Leiden, pp. 13-20. Haesaerts, P., Damblon, F., Bachner, M., Trnka, G., 1996. Revised stratigraphy and chronology of the Willendorf II sequence, Lower Austria. Archaeologia Austriaca 80, 25-42. Haesaerts, P., Borziak, I., Van der Plicht, J., Damblon, F., 1998. Climatic events and Upper Paleolithic chronology in the Dniester Basin: new 14C results from Cosautsi. Radiocarbon 40, 649-657. Hansen, J., 2001. Macroscopic plant remains from Mediterranean caves and rock shelters: avenues of interpretation. Geoarchaeology 16, 401-432. Harrison, S.P., Prentice, C.I., 2003. Climate and C02 controls on global vegetation distribution at the Last Glacial Maximum: analysis based on palaeovegetation data, biome modeling and palaeoclimate simulations. Global Change Biology 9, 983-1004. Haxeltine, A., Prentice, I.C., 1996. An equilibrium terrestrial biosphere model based on ecophysical constraints, resource availability and competition among plant functional types. Global Geochemical Cycles 10, 693-709. Hewitt, G.M., 1996. Some genetic consequences of ice ages and their role in divergence and speciation. Botanical Journal of the Linnean Society 58, 247-276. Hewitt, G.M., 2000. The genetic legacy of the Quaternary ice ages. Nature 405, 907-913. Hewitt, G.M., 2004. Genetic consequences of climatic oscillations in the Quaternary. Philosophical Transactions of the Royal Society, Series B 359, 183-197. Huntley, B., Allen, J.R.M., 2003. Glacial environments III. Palaeovegetation patterns in late glacial Europe. In: van Anděl, T.H., Davies, S.W. (Eds.), Neanderthals and Modern Humans in the European Landscape During the Last Glaciation. McDonald Institute for Archaeological Research, Cambridge, pp. 79-102 Chapter 6. Huntley, B., Birks, H.J.B., 1983. An Atlas of Past and Present Pollen Maps for Europe: 0-13,000 Years Ago. Cambridge University Press, Cambridge. Huntley, B., Alfano, M.J., Allen, J.R.M., Pollard, D., Tzedakis, P.C., de Beaulieu, J.-L., Grůger, E., Watts, W., 2003. European vegetation during Marine Oxygen Isotope Stage 3. Quaternary Research 59, 195-212. Jankovská, V., 1988. A reconstruction of the Late-Glacial and Early Holocene evolution of forest vegetation in the Proprad Bain, Czechoslovakia. Folia Geobotanica 23, 303-319. Jankovská, V., 1998. Late Glacial and Early Holocene of Tatras' foreground basins—an analogy of Siberian boreal and subboreal zone?. In: Bencatova, B., Hrivnak, R. (Eds.), Rastliny a clovek. Technická univerzita vo Zvolené, Brno, pp. 89-95. Joris, O., Weninger, B., 2000. Radiocarbon calibration and the absolute chronology of the late Glacial. In: L'Europe Centrále et Septentrionale au Tardiglaciaire, Table Ronde de Nemours 13-16 mai 1997. Mémoires de la Musée de Préhistoire de Pile de France 7, pp. 19-54. Kadereit, J.W., Griebeker, E.M., Comes, H.P., 2004. Quaternary diversification in European alpine plants: pattern and process. Philosophical Transactions of the Royal Society, Series B 359, 265-275. Kienast, F., 1987. FORECE—A Forest Succession Model for Southern Central Europe. ORNL/TM-10575. Oak Ridge National Laboratory, Tennessee pp. 1-78. Konnert, M., Bergmann, F., 1995. The geographical distribution of genetic variation of silver fir (Abies-alba, Pinaceae) in relation to its migration history. Plant Systematics and Evolution 196, 19-30. Krolopp, E., 1977. Absolute chronological data of the Quaternary sediments in Hungary. Földrajzi Kozlemények 26, 228-232. Lagercrantz, U., Ryman, N., 1990. Genetic structure of Norway spruce (Picea abies): concordance of morphological and allozymic variation. Evolution 44, 38-53. Larcher, W., 1995. Physiological Plant Ecology, third ed. Springer, Berlin. Lascoux, M., Palme, A., Cheddadi, R., Latta, R.G., 2004. Impact of the ice ages on the genetic structure of trees and shrubs. Philosophical Transactions of the Royal Society of London, Series B 359, 197-209. Liivrand, E., 1990. Type section of the lower and middle-Valdaian interstadial deposits at Töravere in south-east Estonia. Proceedings of the Estonian Academy of Sciences 39, 12-17. Liivrand, E., 1991. Biostratigraphy of the Pleistocene deposits in Estonia and Correlations in the Baltic Region. Stockholm University Department of Quaternary Research, Stockholm, report 19. Lister, A.M., Sher, A.V., 1995. Ice cores and mammoth extinction. Nature 378, 23-24. Marosi, S., Szilárd, J., 1974. Neuere Angaben über das Alter das Balatons. Földrajzi Ertesitö 23, 333-346. Mason, S.L., Hather, J.G., Hillman, G.C., 1994. Preliminary investigation of the plant macro-remains from Dolní Věstonice II and its implications for the role of plant foods in Palaeolithic and Mesolithic Europe. Antiquity 68, 48-57. Musil, R., 1997. Klimatická konfrontace terestrických a marinních pleistocenních sedimentů (Climatic comparison of terrestrial and marine pleistocene sediments). Dynamika vztahů marinního a kontinentálního prostředí 93-167. Přírod. Brno, Fakulta Masarykovy Univerzity. Musil, R., 2003. The Middle and Upper Palaeolithic game suite in central and south-eastern Europe. In: van Andel, T.H., Davies, S.W. (Eds.), Neanderthals and Modern Humans in the European Landscape during the Last Glaciation. McDonald Institute for Archaeological Research, Cambridge, pp. 167-190. Navarro Camacho, C.N., Carrion, J.S., Navarro, J., Prieto, A.R., 2000. An experimental approach to the palynology of cave deposits. Journal of Quaternary Science 15, 603-619. Nikolov, N., Helmisaari, H., 1992. Silvics of the circumpolar boreal forest tree species. In: Shugart, H.H., Leemans, R., Bonan, G.B. (Eds.), A Systems Analysis of the Global Boreal Forest. Cambridge University Press, Cambridge, pp. 13-84. Ntinou, M., 2002. Vegetation and human communities in prehistoric Greece. In: Badal, E., Bernabeu, J., Marti, B. (Eds.), Neolithic Landscapes of the Mediterranean. University of Valencia, Valencia, pp. 91-104. Opravil, E., 1994. Vegetation. In: Svoboda, J. (Ed.), Pavlov I— Excavation 1952-1983, Ch. V., ERAUL 66. The Dolní Věstonice Studies 2, Liege, pp. 177-180 Liege. Palmě, A., Vendramin, G.G., 2002. Chloroplast DNA variation, postglacial recolonization and hybridization in hazel, Corylus avellana. Molecular Ecology 11, 1769-1780. Palmě, A., Semerikov, V., Lascoux, M., 2003a. Absence of geographical structure of chloroplast DNA in sallow Salix capreas. Heredity 91, 465^174. Palmě, A., Su, Q., Rautenberg, A., Manni, F., Lascoux, M., 2003b. Postglacial recolonisation and DNA variation of silver birch, Betula pendula. Molecular Ecology 12, 201-212. Pécsi, M., 1977. A hazai éurópai és az európai loszképzodmények paleogrográiai kutatása és osszehasinlítása. Geonómia és Bányá 10, 183-221. 2386 K.J. Willis, T.H. van Andel / Quaternary Science Reviews 23 (2004) 2369-2387 Petit, R.J., Aguinagalde, I., de Beaulieu, J.-L., Bittkau, C, Brewer, S., Cheddadi, R., Ennos, R., Fineschi, S., Grivet, D., Lascoux, M., Mohnty, A., Müller-Starck, G, Demesure-Musch, B., Palme, A., Pedro Martin, J., Rendell, S., Vendramin, G.G., 2003. Glacial refugia: hotspots but not melting pots of genetic diversity. Science 300, 1563-1565. Pokorný, P., Janlovská, V., 2000. Long-term vegetation dynamics and the infilling process of a former lake (Svarcenberk, Czech Republic). Folia Geobotanica 35, 433^157. Pollard, D., Barron, E.J., 2003. Causes of model-data discrepancies in European climate during Oxygen Isotope Stage 3 with insights from the last glacial maximum. Quaternary Research 59, 108-113. Prentice, I.C., Helmisaari, H., 1991. Silvics of north European trees: compilation, comparisons and implications for forest succession modelling. Forest Ecology and Management 42, 79-93. Rudner, E., Sumegi, P., 2001. Recurring Taiga forest-steppe habitats in the Carpathian Basin in the Upper Weichselian. Quaternary International 76/77, 177-189. Rybníčková, E., Rybníček, K., 1991. The environment of the Pavlovian: palaeoecological results from Bulhary, South Moravia. In: Palaeovegetational Development in Europe, Proceedings of the Pan-European Palaebotanical Conference 1991. Vienna, Museum of Natural History, pp. 73-79. Schweingruber, F.H., 1978. Microscopic wood anatomy: structural variability of stems and twigs in recent and subfossil woods from Central Europe. Zug, Zürcher. Schweingruber, F.H., 1990. Anatomie europäischer Hölzer: ein Atlas zur Bestimmung europäischer Baum, Strauch, und Zwergstrauchhölzer. Haupt, Bern. Shackleton, N.J., Fairbanks, R.G, Tzu-chien, C, Parrenin, F., 2004. Absolute calibration of the Greenland time scale: implications for Antarctic time A scales and for A14C. Quaternary Science Reviews. Slavíková-Vešlá, L, 1950. Reconstruction of the succession of forest trees in Czechoslovakia on the basis of an analysis of charcoals from prehistoric settlements. Studia Botanica Cechoslovaca 11, 198-225. Stewart, J.R., Lister, A.M., 2002. Cryptic northern refugia and the origins of modern biota. Trends in Ecology and Evolution 16, 608-613. Stewart, J.S., van Kolfschoten, T., Markova, A., Musil, R., 2003. The mammalian faunas of Europe during Oxygen Isotope Stage Three. In: van Andel, T.H., Davies, S.W. (Eds.), Neanderthals and Modern Humans in the European Landscape during the Last Glaciation. McDonald Institute for Archaeological Research, Cambridge, pp. 103-130. Street-Perrott, FA., Huang, Y., Perrott, RA., Eglinton, G, Barker, P., Ben Khelifa, L., Harkness, D.D., Olago, D.O., 1997. Impact of lower atmospheric carbon dioxide on tropical mountain ecosystems. Science 278, 1422-1426. Svoboda, J., Svoboda, H., 1985. Les industries de type Bohunice dans leur cadre stratigraphique et écologique. LAnthropologie 89, 505-514. Taberlet, P., Bouvet, J., 1994. Mitochondrial DNA polymorphism, phylogeography and conversation genetics of the Brown Bear Ursus arctos in Europe. Proceedings of the Royal Society of London, Series B 225, 195-200. Taberlet, P., Cheddadi, R., 2002. Quaternary refugia and persistence of biodiversity. Science 297, 2009-2010. Tilman, D., 1993. Carbon dioxide limitation and potential direct effects of its accumulation on plant communities. In: Kareiva, P.M., Kingsolver, J.G, Huey, R.B. (Eds.), Biotic Interactions and Global Change. Sinauer Associates, Sunderland, MA, pp. 333-346. Turner, C, 1985. Problems and pitfalls in the application of palynology to Pleistocene archaeological sites in Western Europe. In: Renault-Miskovsky, L, Bui-Thi, M., Girard, M. (Eds.), Palynologie archeologique. Centre de Recherches Archeologiques, Antibes, pp. 347-373 C.N.R.S. Tzedakis, P.C., 1993. Long-term tree populations in northwest Greece through multiple Quaternary climatic cycles. Nature 364, 437-440. Tzedakis, P.C., Andrieu, V., de Beaulieu, J.-L., Crowhurst, S.J., Follieri, M., Hooghiemstra, H., Magri, D., Reihe, M., Sadori, L., Shackleton, N.J., Wijmstra, T.A., 1997. Comparison of terrestrial and marine records of changing climate of the last 500,000 years. Earth and Planetary Science Letters 150, 171-176. Tzedakis, P.C., Lawson, LT., Frogley, M.R., Hewitt, G.M., Preece, R.C., 2002. Buffered tree population changes in Quaternary refugium: evolutionary implications. Science 297, 2044-2047. van Andel, T.H., 2002. Reconstructing climate and landscape of the middle part of the last glaciation in Europe—The Stage 3 Project. Quaternary Research 57, 2-8. van Andel, T.H., 2003. Epilogue—humans in an ice age: the Stage 3 Project. In: van Andel, T.H., Davies, W. (Eds.), Neanderthals and Modern Humans in the European Landscape of the Last Glaciation. McDonald Institute for Archaeological Research, Cambridge, pp. 257-262. van Andel, T.H., Tzedakis, P.C., 1996. Palaeolithic landscapes of Europe and Environs, 150,000-25,000 years ago. Quaternary Science Reviews 15, 481-500. van Andel, T.H., Tzedakis, P.C., 1998. Priority and opportunity: Reconstructing the European Middle Palaeolithic climate and landscape. In: Bayley, J. (Ed.), Science in Archaeology: An Agenda for the Future. English Heritage, London, pp. 37-46. van Andel, T.H., Davies, S.W., Weninger, B., 2003. The human presence in Europe during the last glacial period I. Human migrations and the changing climate. In: van Andel, T.H., Davies, S.W. (Eds.), Neanderthals and Modern Humans in the European Landscape during the Last Glaciation. McDonald Institute for Archaeological Research, Cambridge, pp. 31-56 Chapter 4. Vertes, L., 1964. Das jung Paläolithikum von Arka in Nord-Ungarn. Quartär 15/16, 132-139. Vita-Finzi, C, Higgs, E.S., 1970. Prehistoric economy in the Mount Carmel area of Palestine: site catchment analysis. Proceedings of the Prehistoric Society 36, 1-37. Vogel, I.C., Waterbolk, H.I., 1964. Groningen radiocarbon dates V. Radiocarbon 6, 349-369. Washburn, A.L., 1979. Geocryology. Wiley, New York. Wick, L., Lemke, G., Sturm, M., 2003. Evidence of late glacial and Holocene climatic change and human impact in eastern Anatolia: high resolution pollen, charcoal, isotopic and geochemical records from the laminated sediments of Lake Van, Turkey. The Holocene 13, 665-677. Willis, K.J., 1992. The late Quaternary vegetational history of northwest Greece: I. Lake Gramousti; II. Rezina Marsh; III. A comparative study of two contrasting sites. New Phytologist 121, pp. 101-117, 119-138, 139-155. Willis, K.J., 1994. The vegetational history of the Balkans. Quaternary Science Reviews 13, 769-788. Willis, K.J., Niklas, K., 2004. The role of Quaternary environmental change in plant macroevolution: the exception or the rule? Philosophical Transactions of the Royal Society of London. Series B 359, 159-173. Willis, K.J., Whittaker, R.J., 2000. The refugial debate. Science 287, 1406-1407. Willis, K.J., Rudner, E., Sümegi, P., 2000. The full-glacial forests of central and south-eastern Europe. Quaternary Research 53, 203-213. K.J. Willis, T.H. van Andel / Quaternary Science Reviews 23 (2004) 2369-2387 2387 Wohlfarth, B., Hannon, G., Feurdean, A., Ghergari, L., Onac, B.P., Possnert, G, 2001. Reconstruction of climatic and environmental changes in NW Romania during the early part of the last deglaciation (~15,000-13,600 cal yr BP). Quaternary Science Reviews 20, 1897-1914. Wright, Jr., H.E., Kutzbach, J.E., Webb III, T., Ruddiman, W.F., Street-Perrott, F.R., Bartlein, P.J. (Eds.), 1993, Global Climates since the Last Glacial Maximum. University of Minnesota Press, Minneapolis. Yurtsev, B.A., 2001. The Pleistocene "Tundra-Steppe" and the productivity paradox: the landscape approach. In: Elias, S.A., Brigham-Grette, J. (Eds.), Beringean Paleoenvironments: Festschrift in Honour of D.M Hopkins. Quaternary Science Reviews 20, 165-174.