lable at ScienceDirect

Quaternary International 366 (2015) 25e36

Contents lists avai

Quaternary International

journal homepage: www.elsevier .com/locate/quaint

A charcoal record of Holocene woodland succession from sandstonerock shelters of North Bohemia (Czech Republic)

Jan Nov�ak a, *, Ji�rí Svoboda b, Petr �Sída c, Jan Prost�redník d, Petr Pokorný e

a Department of Botany, Faculty of Science, University of South Bohemia, Brani�sovsk�a 31, CZ 370 05 �Cesk�e Bud�ejovice, Czech Republicb Institute of Archaeology, AC CR, 691 29 Dolní V�estonice, Czech Republicc Department of Archaeology, West Bohemian University in Plze�n, 30100 Plze�n, Czech Republicd Museum of the Bohemian Paradise, Sk�alova 71, 511 01 Turnov, Czech Republice Center for Theoretical Studies, Charles University in Prague and the Academy of Sciences of the Czech Republic, Jilsk�a 1, 110 00 Praha 1, Czech Republic

a r t i c l e i n f o

Article history:Available online 2 October 2014

Keywords:CharcoalRock sheltersSandstone areaVegetation historyPollen analysis

* Corresponding author.E-mail address: prourou@gmail.com (J. Nov�ak).

http://dx.doi.org/10.1016/j.quaint.2014.08.0421040-6182/© 2014 Elsevier Ltd and INQUA. All rights

a b s t r a c t

Archaeological excavations at North-Bohemian sandstone rock shelters have uncovered complex evi-dence of intermittent human presence since the Late Palaeolithic to recent times. In this paper, weinvestigate the history of Holocene woodlands based on charcoal assemblages from stratified sandyaccumulations under archaeologically investigated rock shelters. In total, we carried out anthracologicalanalyses of eight profiles excavated under rock shelters.

Our study shows that the number of charcoal taxa correlated with local environmental diversityaround the rock shelters. Charcoals of Pinus sylvestris were abundant in all the profiles analysed. Profilesin bottom parts of valleys recorded a higher abundance of broadleaf tree charcoal. The abundance of oakgradually decreased from the Bronze Age onwards, as oak got replaced by pine, and locally beech, fir andspruce. Today, several of the recorded woody species are rare or even absent in the sandstone regionunder study.

We focused on comparing anthracological results with results of pollen analyses carried out in thesame areas. Our results demonstrate that when results from different types of archives are compared, it ispossible to gain a more differentiated insight into local vegetation composition. This may carry widemethodological implications.

© 2014 Elsevier Ltd and INQUA. All rights reserved.

1. Introduction

Late Quaternary vegetation and palaeoenvironmental studieshave traditionally been based on interpreting pollen data. This can,however, lead to distorted results due to the complexity of thepollen-vegetation relationship. Pollen profiles are usually situatedin wetlands, although the matrix of surrounding vegetation iscomposed of terrestrial habitats. A more realistic view of pastvegetation changes can be obtained by combining data resultingfrom analyses of pollen, macrofossils (seeds and fruits) and char-coal (e.g. Ritchie, 1995; Birks and Birks, 2004; Nelle et al., 2010).Macrofossil and charcoal analyses provide information about localvegetation, which may not necessarily correspond to pollen re-cords, as already pointed out by Couvert (1968). Past vegetation andtree species composition at terrestrial sites over centuries and

reserved.

millennia canmore reliably be reconstructed based on soil charcoal,that is, by performing pedoanthracological analyses (e.g. Clarket al., 1998; Blackford, 2000; Robin et al., 2012; Nelle et al., 2013).

Vegetation reconstructions from charcoal data obtained fromarchaeological layers are usually based on the assumption thatthese data are not seriously biased by intentional selection duringwood collection or by post-depositional transformation processes.The first serious attempts to perform such reconstructions weremade in the 1940s (Salisbury and Jane, 1940). Criticisms of the basicassumptions of such reconstructions followed soon (Godwin andTransley, 1941), and discussions have been ongoing ever since.Whereas collecting firewood is generally considered to be mostlynon-selective, collecting and felling wood for specialized produc-tion (e.g. metallurgy or carpentry), funerary customs or housebuilding is obviously highly selective (Marston, 2009; Th�ery-Parisotet al., 2010; Rubiales et al., 2011 and others). Samples containingselectively collected wood are generally assumed to be character-ized by reduced species composition (e.g. Nov�ak et al., 2012a). Onthe other hand, people probably obtained firewood from woody

J. Nov�ak et al. / Quaternary International 366 (2015) 25e3626

species growing near archaeological sites (Shackleton and Prins,1992; Jansen et al., 2013).

Continual or long-term charcoal records from Europeanarchaeological sites are still infrequent, and each new localityprovides new perspectives. Some of the best examples of sites witha potential for such investigations are caves and rock shelters withlong-term or repeated human presence (e.g. Badal et al., 1994;Heinz and Barbaza, 1998; Henry et al., 2012). In this regard,North-Bohemian Cretaceous sandstone areas seem to be quiteunique in the Central European context. Rock shelters are a char-acteristic geomorphological formation of sandstone regions (Fig. 7).In many cases, the stratigraphy of deep sandy sediments has beenpreserved intact under rock shelters. Sandy accumulations underrock shelters are generally formed in a specific way. Most of thesandy sediment comes from the sandstone rocks above. Biological(e.g., earth worm bioturbation; e.g. J�egou et al., 1998; Preston andSchmidt, 2006) and physical factors (e.g., uprooting; Samonilet al., 2010) are frequently discussed as influences which signifi-cantly affect the deposition of charcoals in soil. However, the dryand acidic sandy sediments under rock shelters are usually lessstrongly affected by biological factors than common soil horizons.Nevertheless, charcoal records in accumulations of sedimentprobably result from a very complex process of charcoal formation,recording and preservation (Forbes et al., 2006; Fr�ejaville et al.,2013).

Archaeological investigations show that rock shelters of thestudy region (North Bohemia) were widely used as temporary re-treats by hunters and gatherers since the Late Palaeolithic untilrecently (e.g. Svoboda, 2003; Svoboda et al., 2007; �Sída andProst�redník, 2007; Svoboda, 2014). Repeated visits of rock shel-ters over decades, centuries and even millennia are an importantaspect of hunter-gatherer land use (e.g. Smith and McNees, 2011).Former fireplaces are one of the visible traces of their repeatedvisits. Plant macroremains, bones, chipped tools and ceramics havebeen found besides charcoal pieces in sediment layers (e.g. Svobodaet al., 2013).

The first systematic archaeological investigation of Bohemiansandstone rock shelters was carried out approximately 20 years ago(Svoboda, 2003). Some anthracological and macrofossil research ofsandstone rock shelters has been carried out (e.g. Opravil, 2003;Svoboda et al., 2007), but few samples of charcoal or plant mac-roremains have been systematically collected and examined.

Fig. 1. Location of the study area and sites of anthracological profiles: 1 e Janovaz�atoka, 2 e Údolí samoty, 3 e Dvojit�a br�ana, 4 e Konejlova cave, 5 e Velbloud, 6 e

Kristova cave, 7 e V�e�z�ak, 8 e Hlavat�a sk�ala; and pollen profiles: JL e Jelení lou�ze, SH e

Sk�rítk�uv hrnec.

Palaeoecological studies from sandstone areas in Central Europehave so far been mainly based on pollen analyses (e.g. Jankovsk�a,1992; Kune�s and Jankovsk�a, 2000; Pokorný and Kune�s, 2005;Kune�s et al., 2007). Here we present the results of the first sys-tematic anthracological survey of eight rock shelter profilescoupled with pollen-analytical investigations. The main goals ofthis paper are to: (1) determine the history of local forests, (2) drawattention to the relationship between the dynamics of local vege-tation and adjacent environmental conditions, (3) compareanthracological results with results of pollen analyses carried out inthe same areas, and (4) investigate the human impact on localvegetation.

2. Study areas

The North-Bohemian Cretaceous sandstone area is one of thelargest sandstone landscapes in Europe (H€artel et al., 2007). Itcomprises several neighbouring sandstone sub-regions. For ourstudy, we selected three (Bohemian Switzerland, Lu�zick�e hory andthe Bohemian Paradise; Fig. 1) because these are the only oneswhere data-bearing systematic archaeological investigations tookplace in the last decade or so. Our set of study areas is characterizedby the presence of rocky sandstone hummocks and rock shelters,flat or gently sloped tablelands, sunken valleys, and sporadic vol-canic hills (Demek, 1987). Valleys with steep rocky slopes and hu-mid gorges represent only a small part of the landscape. There areonly a few watercourses within the sandstone areas, so the lack ofwater always significantly influenced human presence and activ-ities. Humans often visited rock shelters located near water sources,though settlements have also been discovered under rock shelterssituated in upper parts of valleys (Svoboda et al., 2013). Most of theinvestigated territory is situated 200e400 m above sea level, butseveral volcanic hills rise above 500 m.a.s.l.

According to soil maps of the Czech Republic (N�eme�cek et al.,2008; soil taxonomical units according to; Mich�eli et al., 2007),arenosols, lithosols, podzols, cambisols, luvisols and fluvisols arethe most common soils on the sandstone bedrock in the area,which is now predominantly forested. Bottoms of slopes and val-leys are characterized by the presence of luvisols, cambisols andfluvisols. Arenosols, lithosols and podzols are typical for upperparts of slopes and places near rocks.

The recent forests in the area are mostly composed of Pinussylvestris and Picea abies. Also common are Fagus sylvatica, Betulapendula, Quercus petraea and Quercus robur. Xerophilous pine for-ests (un. Dicrano-Pinion, with e.g. Avenella flexuosa, Calluna vulgaris,Festuca ovina, Rumex acetosella and Cladonia sp.) are typical habitatsin upper parts of slopes. Today's presence of Fagus is often con-nectedwith bottoms of slopes or valleys. Picea is also more frequentat valley bottoms, but it mostly occurs there due to coniferousplantations. Quercus sp. and Corylus avellana is admixed in all typesof habitats, but their abundance is slightly higher at bottoms ofvalleys and slopes. Carpinus betulus and demanding broadleaf trees(Tilia sp., Acer sp., Fraxinus excelsior and Ulmus sp.) occur infre-quently at bottoms of slopes and valleys. The rare presence of Alnussp. and Salix sp. is restricted to watercourses and springs. Micro-climatically inverse stands are characterized by oreophytic in-dicators (e.g. Blechnum spicant, Homogyne alpina, Huperzia selago,Petasites albus, Viola biflora).

According to the Archaeological Database of Bohemia (Kuna,2002), the study area borders with an agricultural landscapedotted with sites of ancient settlement from the Late Palaeolithic(14,000e11,500 cal BP), Mesolithic (11,500e7500 cal BP), Neolithic(7500e6000 cal BP), Late Neolithic (6000e4000 cal BP), BronzeAge (4000e2800 cal BP), Hallstatt (2800e2400 cal BP), La-T�ene(2400e2000 cal BP) and Roman periods (2000e1650 cal BP).

Fig. 2. Percentage proportion of charcoal taxa in the profiles under study. US e Údolí samoty, KOC e Konejlova cave, VB e Velbloud, VZ e V�e�z�ak, KRC e Kristova cave, HS e Hlavat�a sk�ala, DB e Dvojit�a br�ana, JZ e Janova z�atoka. LatePalaeolithic: 14,000e11,500 cal BP; Mesolithic: 11,500e7500 cal BP; Neolithic: 7500e6000 cal BP; Late Neolithic:6000e4000 cal BP; Bronze Age: 4000e2800 cal BP; Hallstatt: 2800e2400 cal BP; La-T�ene: 2400e2000 cal BP; MiddleAges: 900e500 cal BP.

J.Nov

�aket

al./Quaternary

International366(2015)

25e36

27

0 10 20 30 40 50 60 70 80 90 100

110

120

130

140

150

160

170

180

190

200

210

220

230

240

250

260

270

280

290depth (cm)

1790

±30

BP

2940

±30

BP

4500

±35

BP

5330

±35

BP

5650

±40

BPdates

20

Abies a

lba Acer Alnus g

lutino

sa t.

2040

60

Betula

alba t

.

Callun

a vulg

aris

Carpinu

s betu

lus

Castan

easa

tiva

2040

Corylus

avell

ana

2040

Fagus

sylva

tica Fran

gula

alnus

Fraxinu

s exc

elsior

Junip

erus 20

Picea a

bies

2040

Pinus s

ylves

tris t. Pop

ulus

2040

Quercu

s

Rubus Salix Sambu

cus n

igra

Sorbus

t. 20

Tilia co

rdata

t.Ulm

us Vaccin

ium Viscum

2000

4000

micros

copic

charc

oal /

1000

LPA

Z

JL3

JL2

JL1

Jele

ní lo

uže

(per

cent

age

polle

n di

agra

m -

woo

dy ta

xa)

Fig. 3. Jelení lou�ze, Percentage pollen diagram of woody taxa and numbers of microscopic charcoal pieces/1000.

J. Nov�ak et al. / Quaternary International 366 (2015) 25e3628

Fig. 4. Sk�rítk�uv hrnec, Percentage pollen diagram of woody taxa and numbers of microscopic charcoal pieces/1000.

J. Nov�ak et al. / Quaternary International 366 (2015) 25e36 29

Fig. 5. Percentage pollen diagram from the LaTene layer (20e30 cm) in the Kristova cave.

J. Nov�ak et al. / Quaternary International 366 (2015) 25e3630

Attempts to colonize the forested sandstone areas were rather localand temporary in all periods, but the evidence of their ancientcolonization is poor and restricted to rock shelters. Colonizationhad been most intensive in the High Middle Ages (13th century;Meduna and S�adlo, 2009). From the Middle Ages to the present, thepredominantly forested areas have been affected, for example, byhunting, woodcutting, glass and charcoal production, and localmining of iron ore (Prost�redník and �Sída, 2009).

Table 1AMS-14C dates of the single charcoal pieces from the three investigated profiles.UGAMS: UGAMS: Centre of Applied Isotope Studies, The University of Georgia, USA;Poz: Pozna�n Radiocarbon Laboratory, Poznan, Poland; Gd: Gadam Centre, Gliwice,

3. Methods

In total, eight rock shelter profiles were sampled for anthraco-logical analysis (Table 3). In the Bohemian Paradise, rock sheltersare approximately 5e15 km apart. The Údolí samoty rock shelter

Fig. 6. Ordination diagram showing results of a PCA analysis of charcoal taxa and theprofile/period relationship. The first axis explained 30.8% of the variability, and the firsttwo axes together explained 51.3% of the variability. The first part of each abbreviationidentifies the profile (US e Údolí samoty, KOC e Konejlova cave, VB e Velbloud, VZ e

V�e�z�ak, KRC e Kristova cave, HS e Hlavat�a sk�ala, DB e Dvojit�a br�ana, JZ e Janovaz�atoka), and the second part identifies the period (Palaeo e Late Palaeolithic, Meso e

Mesolithic, Bronze e Bronze Age, Iron e Iron Age, Mid e Middle Ages), e.g. KOC Iron e

Konejlova cave/Iron Age.

and the Janova z�atoka rock shelter are located approximately 60 kmand 80 km from the Bohemian Paradise, respectively (Fig. 1). Theanthracological research was carried out in the context of archae-ological investigations. Charcoal samples were collected fromsandy sediment profiles sectioned into 10 cm thick layers down tothe bedrock; only part of the profile Údolí samoty was divided into20 cm thick layers and Janova z�atoka was sampled in 5 cm thicklayers. Twenty litres of sediment samples were taken from everylayer of each profile.

Poland; OxA: University of Oxford laboratory; Oxford, England, UK.

Lab. code Profile Depth(cm)

14C age(BP)

Range Calibrated 14Cage (BC)range 95.4%

Datedcharcoltaxa

UGAMS 9516 Dvojit�a br�ana 60e70 6730 30 5711e5571 PinusUGAMS 11223 Dvojit�a br�ana 80e90 7900 30 7023e6648 PinusUGAMS 11224 Hlavat�a sk�ala 85 8010 30 7058e6824 PinusGdA-531 Hlavat�a sk�ala 125 8480 50 7593e7482 PinusUGAMS 11225 Hlavat�a sk�ala 147 8630 30 7716e7586 PinusPoz-23176 Janova z�atoka 35 9250 60 8630e8300 PinusUGAMS 11226 Konejlova cave 45 2160 25 357e112 PinusUGAMS 9511 Konejlova cave 65 3430 25 1875e1665 PinusUGAMS 9512 Konejlova cave 147 8880 30 8225e7945 PinusGdA-975 Kristova cave 107 6900 40 5882e5716 PinusGdA-977 Kristova cave 123 7130 40 6071e5916 PinusGdA-976 Kristova cave 140 7440 40 6404e6232 PinusPoz-48373 Údolí samoty 140 7960 50 7043e6697 CorylusPoz-43850 Údolí samoty 165 8730 50 7942e7604 CorylusPoz-43849 Údolí samoty 180 9360 50 8766e8531 CorylusOxA-25772 Údolí samoty 270 11,750 50 11,808e11,482 PinusUGAMS 12207 Velbloud 70 2950 20 1263e1057 PinusUGAMS 12208 Velbloud 115 7410 30 6375e6228 PinusGdA-1207 V�e�z�ak 140 8060 40 7141e6825 Pinus

The extraction of charcoals and plant macrofossils from thesediments followed standard wet-sieving procedures (Carcailletand Thinon, 1996) using a 0.4 mm sieve. Charcoal pieces weremanually selected from the sieved fractions under a stereoscopicmicroscope. Charcoal analysis was performed only on fragmentsfrom the largest fraction (>1 mm). Smaller charcoal pieces thatwere not taxonomically identifiable were excluded because theyhardly provide any information on past forest composition (Robinet al., 2014).

Charcoals were identified with the help of a reference collectionand standard identification keys (Greguss, 1972; Schweingruber,1990) using an episcopic interference microscope with200e500 � magnification. Anthracomass (mg of charcoal per kg ofsoil; Talon et al., 1998) was determined from charcoals larger than

J. Nov�ak et al. / Quaternary International 366 (2015) 25e36 31

1mm. Percentagemacrocharcoal diagrams were created using Tiliav. 1.7.16 (Grimm, 2011).

Pollen analysis was performed on continuous peat sectionssampled using piston (Jelení lou�ze site) and auger (Sk�rítk�uv hrnecsite) corers. Sites for pollen analysis were selected following thenecessity of finding another type of proxy record for habitats of thesame type as that of the rock shelters under study (i.e. interiors of

Table 3Anthracomass and numbers of species in study profiles.

Site/profile Period LaPa Mes LaNe BrAg LaT/Hal MidAg LaPa Mes LaNe BrAg LaT/Hal MidAg

Anthracomass (mg/kg) Number of species

Position on a slope

Dvojit�a br�ana N slope 181.2 253.1 286.2 192.7 6 8 6 2Hlavat�a sk�ala N slope 573.9 69.7 188.1 147.9 12 12 10 5Janova z�atoka Bottom 1.1 195.6 397.1 1277.8 1 7 13 11Konejlova cave Bottom 124.4 612.9 429.7 355.8 8 11 16 15Kristova cave Upper 406.2 134.7 326.7 175.4 11 5 10 5Údolí samoty Upper 34.9 434.8 300.5 881.1 3750.1 4 8 6 7 1Velboud Upper 8.1 206.8 33.4 515.7 5 8 4 8V�e�z�ak Bottom 78.4 100.5 58.8 9 13 15

sandstone valleys or gorges). Although no pollen profile has so farbeen discovered in the direct vicinity of the investigated rockshelters, we have discovered two sites that exactly meet the re-quirements specified above. The Jelení lou�ze site is locatedapproximately 8 km from the Janova z�atoka rock shelter, and theSk�rítkuv hrnec site is located 15 km from the Údolí samoty rockshelter. Both are identical in character, being peat bogs situated atthe bottom of a sandstone gorge. Samples of peat were prepared bythe standard acetolysis methods (Moore et al., 1991), and at least700 pollen grains were counted in each sample. Percentage pollendiagrams (Figs. 3 and 4) were plotted based on the total terrestrialpollen sum. Only woody taxa are presented because only these arerelevant for our comparison with anthracological results. The dia-grams are dated using radiocarbon (AMS) measurements per-formed on macrofossils of terrestrial plants (seeds, fruits and/orneedles e Table 2).

Table 2AMS-14C dates from pollen profiles. Dated material: terrestrial macrofossils. Poz:Pozna�n Radiocarbon Laboratory, Pozna�n, Poland.

Lab code Profile Depth (cm) C14 age (BP) Range

Poz-9132 Jelení lou�ze 48 1790 30Poz-9133 Jelení lou�ze 122 2940 30Poz-9134 Jelení lou�ze 192 4500 35Poz-9135 Jelení lou�ze 258 5330 35Poz-8121 Jelení lou�ze 280 5650 40Poz-59611 Sk�rítk�uv hrnec 180e185 2240 30Poz-59612 Sk�rítk�uv hrnec 221e225 2665 30Poz-59613 Sk�rítk�uv hrnec 253e257 3820 30Poz-59615 Sk�rítk�uv hrnec 288e291 4070 35

Some charcoal samples were also subjected to 14C AMS radio-carbon dating. The age of others was determined based on anarchaeological chronology. The age of each layer was determinedbased on typical archaeological artefacts (chipped industry in theLate Palaeolithic andMesolithic period, pottery fragments of typicalvessels in younger periods). The radiocarbon dating was performedusing the 14C AMS method at the Pozna�n Radiocarbon Laboratory,Poland; CAIS Athens, Georgia, USA; Gadam Centre, Gliwice, Poland;and University of Oxford laboratory, Oxford, England. All theradiocarbon dates were calibrated using OxCal 4.1. (Bronk-Ramsey,

2001). Charcoal samples for AMS 14C dating (Table 1) were takenfrom archaeologically important layers of each profile. Onerandomly chosen charcoal piece per base layer was dated.Considering the archaeological context of individual samples, it waspossible to differentiate among periods, including the Late Palae-olithic, the Mesolithic, the Late Neolithic, the Bronze Age, the IronAge and the Middle Ages.

Statistical analyses in CANOCO v. 4.5 (ter Braak and �Smilauer,2002) allowed us to compare charcoal taxa and examine the pro-file/period relationship. A detrended correspondence analysis(DCA) was carried out to check the length of the gradient. Themaximum length of the gradient was 1.18, so a principal compo-nents analysis (PCA) was performed. A logarithmic transformationof the percentage data centred by species was used.

4. Results

4.1. Anthracological research results

From the eight rock shelter profiles, a total of 12,181 charcoalfragments from 365 samples were identified. The anthracologicalanalysis revealed 22 taxa (Fig. 2). The amount of charcoal fragmentsand the number of charcoal species differed significantly among thesamples or profiles (see Table 3). The spectrum is generally domi-nated by Pinus (n ¼ 7487) and Quercus (n ¼ 1780). Other commontaxa are Picea (n ¼ 639), Fagus (n ¼ 540), Corylus (n ¼ 433), Abies(n ¼ 261), Fraxinus (n ¼ 140), Populus/Salix (n ¼ 144) and Tilia(n ¼ 127). Further taxa include Taxus (n ¼ 120), Ulmus (n ¼ 63),Betula (n ¼ 71), Acer (n ¼ 48), Alnus (n ¼ 42), Salix (n ¼ 16), Sorbus(n¼ 10), Lonicera (n¼ 9), Carpinus (n¼ 5), Frangula (n¼ 5), Juniperus(n ¼ 4), Prunus cf. Padus (n ¼ 1) and Vaccinium (n ¼ 1). Charcoal ofCorylus shells is very common (n ¼ 572), especially in Mesolithiclayers. The highest species diversity (17 species) was found in theprofile Konejlova Cave, the lowest in the profile Velbloud (9 species).

4.2. Woodland history reconstructed from charcoal assemblages

The charcoal assemblages differed according to the profile andtime period (Table 3, Fig. 2). Late Palaeolithic layers(14,000e11,500 cal BP) are remarkable for their low species di-versity and anthracomass. Pinus was a dominant species, andspruce/larch and birch were common. A very small charcoal frag-ment of a deciduous tree (cf. Frangula sp.) was also found.

Significant differences in charcoal records between rock shelterslocated in the (i) upper parts (Velbloud, Údolí samoty, Kristovacave) and (ii) at the bottoms of valleys or on the northern slopes(Dvojit�a br�ana, Hlavat�a sk�ala, Janova z�atoka, Konejlova cave, V�e�z�ak)were apparent in layers from the Mesolithic period (Fig. 6).

Fig. 7. Velbloud rock shelter situated in the upper part of a slope during excavation in2012 (Photo e J. Prost�redník).

J. Nov�ak et al. / Quaternary International 366 (2015) 25e3632

(i) Profiles in upper parts of slopes documented only reducedspecies composition with a strong dominance of Pinus, andthe rare presence of broadleaf species (Quercus, Corylus, Tilia,Populus/Salix, Betula) and spruce throughout the Holocenehistory. The most abundant presence of Corylus shells wasfound in the Mesolithic period (11,500e7500 cal BP). Highamounts of oak charcoal were recorded in the upper Meso-lithic and the beginning of the Late Neolithic period(6000e4000 cal BP). Charcoals of Fagus were recorded fromthe Late Neolithic or Bronze Age onwards, but their occur-rence was scarce and episodic. Abies charcoal was also veryrare at these sites.

(ii) In theMesolithic period,Quercus and Pinusdominated,Corylus,Picea and Tiliawere common in most of the profiles, and onlyscarce occurrences of Acer, Alnus, Betula, Fraxinus, Frangula,Lonicera, Sorbus, Salix/Populus, Taxus and Ulmuswere recorded.The charcoal records show that the vegetation at localities atbottoms of valleys and on northern slopes was rich in specieswith a high abundance of oak and demanding broadleaf spe-cies. During the Bronze Age (4000e2800 cal BP), Quercuscharcoal started to decline, and Picea and Fagus charcoalincreased significantly in spite of the common presence ofdemandingbroadleaf species (Corylus,Tilia, Fraxinus,UlmusandAcer); Alnus and Taxus, Betula Carpinus, Salix/Populus charcoalswere recorded rarely in this period. Hazelnut shells were rare.Abies charcoal started to increase at the end of the Bronze Age.The increase in Fagus, Abies and Picea charcoals continuedduring the Iron Age (Hallstatt period 2800e2400 cal BP; La-T�ene period 2400e2000 cal BP). The highest species diversitywas recorded in samples from the Hallstatt/La-T�ene period,althoughhazelnut shellsdisappeared. Broadleaf trees andAbiesalba were still abundant in the High Medieval period(900e500 cal BP); however, the dominance of Pinus waselevated. Thepresence of light-demanding species and Corylluscharcoals also increased in the Middle Ages.

4.3. Anthracomass

Most profiles showed differences in anthracomass (mg/kg)among charcoal samples/periods (Table 3). Low anthracomass wasrecorded in the profile V�e�z�ak, and the highest values of anthraco-mass were found the profile Údolí samoty. Late Palaeolithic

samples contained the lowest anthracomass. In the Mesolithicperiod, the highest amounts of anthracomass were recorded in theprofiles Hlavat�a sk�ala, Údolí samoty and Kristova cave. Sampleswith high anthracomass levels (profile Hlavat�a sk�ala, Kristova cave)in the Mesolithic period often contained noticeably lower amountsin the following period. These Mesolithic samples were alsonoticeable for their high numbers of hazelnut shells. Late Neolithicsamples often contained lower anthracomass than Mesolithicsamples. Samples from the profiles Konejlova cave and Údolísamoty contained high anthracomass in the Bronze Age and IronAge.

4.4. Analysis of pollen from peat accumulations

The two investigated pollen profiles (see diagrams in Figs. 3 and4) come from the same type of habitat (a small peat bog at thebottom of a sandstone gorge), but from two different sub-regions(see Fig. 1). Both records show many similarities, but also severalimportant differences in vegetation development. The commonand most striking feature is succession from broadleaf assemblageswith Quercus, Corylus, Tilia, Fraxinus and Ulmus (Stage 1: pollenzones JL1 and SH1) to assemblages with Fagus, Abies and Picea(Stage 2: pollen zones JL 2 to JL3 and SH2 to SH3). This succession isprogressive, and its most dynamic phase is in both cases datedbetween 3000 and 4000 BP.

Despite obvious common features, there are also certain dif-ferences between the two investigated pollen records. These dif-ferences lie in the relative proportions of individual forest trees andin the succession sequence of dominant tree species during theearly phase of Stage 2. In the case of the Sk�rítk�uv hrnec site, Tiliaplays a muchmore important role in Stage 1 compared to the Jelenílou�ze site. The opposite is the case of Quercus. The transitionalphases from Stage 1 to Stage 2 in the Jelení lou�ze and Sk�rítk�uvhrnec profiles (JL1/JL2 transition and SH1/SH2 transition, respec-tively) are characterized by an expansion of Fagus in the formerprofile and an expansion of Picea in the latter. In the late phases ofStage 2 (JL3 and SH3 pollen zones), the greatest difference betweenthe two sites lies in the relative importance of Betula and Pinus invegetation composition.

4.5. Analysis of pollen from rock shelters

Even though samples from sandy accumulations under the rockshelters studied were systematically collected also for pollenanalysis, a large majority of the samples did not contain any pollengrains. Only one sample contained well preserved pollen grains,andwe attribute this exception to the uniqueness in local chemistrye high content of humic acids leached from plant biomass formingan occupation layer. This only positive sample was extracted from aLa-T�ene horizon (layer 20e30 cm) in the profile Kristova cave(Fig. 5). Its pollen spectrum consists mainly of arboreal taxa. Abiesand Pinus pollen grains were the most frequent. Pollen of Betula,Corylus, Picea and Quercus was common, while pollen grains ofAlnus, Carpinus, Tilia, Ulmus and Fagus were among rare ones. Non-arboreal pollen and spores belonged to forest undergrowth acid-ophilous herbs (such as, most of all, Calluna, Polypodium vulgare andPteridium aquilinum).

5. Discussion

5.1. Relationship between woodland history and localenvironmental conditions

The Late Palaeolithic period is characterized by the lowestnumber of charcoal fragments and species. The surrounding

J. Nov�ak et al. / Quaternary International 366 (2015) 25e36 33

vegetation can be reconstructed as an open pine forest where thespecies composition indicates unfavourable environmental condi-tions. Our record of Late Palaeolithic charcoal assemblages is one ofonly a handful from sandstone rock shelters in Central Europe. Asimilar composition of the charcoal and pollen taxa has beendocumented in the peat profile Voron�e�z in the northern part of thePolomen�e hory sandstone region (Nov�ak et al., 2012b).

A remarkable increase of species diversity is recorded since theMesolithic period (11,500e7500 cal BP). Our anthracological studyclearly highlights that the noticeable differences in species di-versity depend on the location of each rock shelter and on envi-ronmental diversity in the vicinity (e.g. bottom or upper part of thevalley, distance from a watercourse; Fig. 6). Relatively smallchanges in species composition during the Holocene (charcoal as-semblages dominated by Pinus) are documented in profiles locatedin dry upper parts of valleys (Údolí samoty, Kristova cave, Vel-bloud). The surroundings of these rockshelters are recently char-acterized by the presence of arenosols, lithosols and podzols. Thepresence of these shallow, dry and acid soils is probably related notonly with topographical position, but also with long-term presenceof pine forest (e.g. Bendel et al., 2006). This confirms what is pre-served in the charcoal records, which speak of a high dominance ofPinus and a rare presence of broadleaf species during the Holocenehistory. Vegetation succession in these habitats was probably morefrequently affected by fire events, as shown pedoanthracologicalstudies (e.g. Nov�ak et al., 2012b; Bobek, 2013).

By contrast, rock shelters located at bottoms of valleys or on theirnorthern slopes (Konejlova cave, V�e�z�ak, Janova z�atoka, Hlavat�ask�ala, Dvojit�a br�ana) demonstrate more pronounced changes of thesurrounding vegetation. The highest species diversity (17 taxa) wasfound in the profile Konejlova cave, which is located at the bottom ofa humid gorge. The humid microclimatic conditions surroundingthe Konejlova cave are corroborated by findings of burnt frog bonesin archaeological layers (Kova�cíkov�a et al., 2012). The high speciesdiversity at the site is caused by a diverse range of habitats in itsvicinity. The lowest species diversity (9 taxa) was recorded in theprofiles Velbloud, which is situated in a dry upper part of a valley.Although the regional climatic conditions of the localities understudy differ only slightly, the vegetation around the rock shelters isprimarily influenced by local environmental conditions. Upper partsof rocks and valleys are dry and get overheated on sunny days.Bottoms of sandstone gorges, by contrast, are notably humid andhave an inversion microclimate (H€artel et al., 2007).

Profiles in bottom parts of valleys recorded a higher abundanceof Quercus charcoal in the Mesolithic period. The decline of oak anddemanding broadleaf species coupled with the increase in pine,beech, spruce and fir is often associated with environmentalchanges in the Late Bronze Age (Lo�zek, 1998). In northern Germany,for example, the constant or rising curve of Fagus is often connectedwith increased anthropogenic activities (Küster, 1997; Wolters,1999; Brande, 2003; Jahns, 2007). It is difficult to assess whetherthese vegetation changes were related more to human influence,the Holocene spread of trees from glacial refugia, natural processes(e.g. the Iversen cycle or wildfire events) or some combination ofthese factors. During the Iron Age, the abundance of Fagus and Abiescharcoals was increased; on the other hand, the charcoal recordsindicate a decrease of Quercus and Pinus. The medieval increase inthe abundance of Pinus and Abies accompanied by the decrease inspecies diversity can be interpreted as a consequence of medievalcolonization of sandstone regions.

5.2. The charcoal record and human impact

Our rock shelters (Údolí samoty, Kristova cave) with highamounts of anthracomass and hazelnut shells from the Mesolithic

period are situated in dry upper parts of valleys, but relatively closeto natural water sources. The amounts of anthracomass suggestthat there were many fireplaces and that humans frequently stayedat the rock shelters. The question remains whether these sitesrepresent longer-term or wintering shelters. Because the dry upperparts of the valleys around the rock shelters have for a long timebeen species-poor habitats (pine forests), the number of speciesremained constant despite substantial increases in anthracomass.Nevertheless, hazelnut shells have also been found at small orshort-term Mesolithic sites (e.g. Opravil, 2003). The oldest exam-ined hazelnut shell was dated to the Preboreal period(9360 ± 50 BP, profile Údolí samoty). This is one of the oldest datesfor a hazelnut shell found in Central Europe (compare e.g. Pokornýet al., 2010; �Sída et al., 2011; Regnell, 2012). The Mesolithic was anera of hunter-gatherers, so wild plants are integrated in virtually allmodels and concepts dealing with this period (e.g. Clarke, 1976;Rowley-Conwy, 1986; Price, 1989; Zvelebil, 1994). Several studiesdescribe the importance of hazelnuts as a food source during theMesolithic period (e.g. Kune�s et al., 2008; Regnell, 2012). Remark-ably, the number of hazelnut shells was not directly linked with theabundance of Corylus charcoal. Corylus charcoal was present in allprofiles, but hazelnut shells were found only in some of the rockshelters. The high abundance of Corylus shells is neverthelesscorrelated with high amounts of anthracomass in Mesolithic layers.During the transition from the Upper Mesolithic period to theagricultural period, the presence of hazelnut shells diminishedsignificantly. Corylus charcoal and Corylus pollen, however,remained abundant until the Bronze Age. The mass of Coryluscharcoals subsequently increased in the Middle Ages.

The presence of Alnus and Salix charcoal was recorded only atsites roughly within 50 m of a watercourse. The abundance ofspecies characteristic of wetlands was very low in these samples. Asimilar result has been obtained at the Dombas rock shelter(Opravil, 2003), which is situated on the bank of the river Plou�cnice.This is probably related to the narrow floodplain and limited spacefor riverside vegetation in the area. Differences between charcoalspectra and pollen data are sometimes explained by selectivecollection or avoidance due to burning properties (e.g. Kreutz,1990;Out, 2010). The absence of Alnus or Tilia charcoals is sometimesexplained as above. In our charcoal assemblages, however, Tilia isconstantly present, indicating that it was commonly used as fire-wood. The presence of Tilia or Alnus as common firewood has beenrevealed by many anthracological studies (e.g. Jansen et al., 2013;Kuna et al., 2013). The availability and proximity of firewood inthe environment is one of the most important factors determiningthe ease with which prehistoric communities collected wood asfuel (Asouti and Austin, 2005). It is nevertheless possible that thedominance of the pine charcoal may have been influenced by thecommon presence of (dry) pine wood in the vicinity of rock shel-ters, regardless of the fact that Pinus is best adapted to the condi-tions of sandstone rocks. Moreover, it is good to bear in mind thatcharcoal records in accumulations of sediment are probably shapedby a very complex process of charcoal formation, recording andpreservation (Forbes et al., 2006; Preston and Schmidt, 2006;Fr�ejaville et al., 2013).

Our records indicate that Taxus charcoals were rarely presentedin the Mesolithic layers, became abundant in the Bronze Age andstarted to disappear during the Middle Ages. Rich findings of Taxuscharcoal are unusual not only for Bohemian sandstone areas butalso for neighbouring regions. The unusual abundance of Taxuscharcoal in certain fireplaces hints at the hypothesis that peopleused some of the rock shelters they visited to process yew wood.Taxuswas the wood of choice for making bows, other weapons andalso tools (Coles et al., 1978; Oeggl and Schoch, 2000). The BronzeAge increase in Taxus and Fagus charcoals is sometimes interpreted

J. Nov�ak et al. / Quaternary International 366 (2015) 25e3634

as a consequence of lower temperatures and higher precipitation.Pernaud (2001), for example, inferred that the climate in theBronze Age was cooler and more humid from Taxus charcoal inlayers at the la Karelsl�e cave in Luxembourg. The presence of Taxuscharcoal was recorded mainly under rock shelters situated at thebottoms of gorges or on northern slopes. At present, Taxus does notoccur in the Bohemian sandstone regions under study.

The archaeological evidence of human activities in the forestedstudy areas is poor and mainly restricted to rock shelters. Meso-lithic settlements were recorded in all of the profiles, but theirinfluence on the surrounding woodland compositionwas relativelyweak. During agricultural prehistory (7500e650 BP), a noticeabledecline of trees accompanied by an increase of Pinus, Fagus, Abiesand Picea was recorded. Still, it is difficult to estimate the extent towhich this succession of woodland vegetation is attributable tohumans. Human presence in the region was relatively local andtemporary. On sandstone bedrock, however, even a low humanimpact can set off dramatic environmental changes. The influenceof human activities was more apparent in the Middle Ages. Some ofthe profiles document amedieval increase in the abundance of pineand fir coupled with a decrease in species diversity. However, thecomposition of vegetation was more affected by human activityduring the modern period (Meduna and S�adlo, 2009). From theMiddle Ages to the present, the predominantly forested areas havebeen affected, for example, by hunting, woodcutting, glass andcharcoal production, and local mining of iron ore (Prost�redník and�Sída, 2009).

5.3. Comparison between the charcoal and pollen records

The two pollen records presented in this article come from smallpeat bogs located at bottoms of sandstone gorges, so they pre-sumably reflect local vegetation composition of the interior of thosesandstone gorges (i.e. the same environment where the investi-gated rock shelters are located). The significant influence of thelocal environment on pollen sequences recorded in such smallwetlands, in relevant literature often called “forest hollows”, hasbeen documented by several studies (e.g. Janssen, 1966; Sugita,1994; Bj€orkman and Bradshaw, 1996; Brostr€om et al., 2008;Overballe-Petersen and Bradshaw, 2011). Besides the small peatbogs Jelení lou�ze and Sk�rítk�uv hrnec, several pollen profiles havepreviously been analysed in the Bohemian sandstone regions, butmostly outside areas of our interest or outside the target environ-ment (Jankovsk�a, 1992; Kune�s and Jankovsk�a, 2000; Kune�s et al.,2007; Nov�ak et al., 2012b). However, the local trends in vegeta-tion history observed in our two pollen profiles appear to begenerally comparable with ones known from other Bohemiansandstone regions: In the Early Holocene (11,500e10,000 BP), Pinuswas absolutely dominant in sandstone regions. Picea might haveoccurred locally, but its importance in the vegetation was negli-gible. At around 9500 BP, broadleaf trees expanded, and thisexpansion reached its maximum at around 6500 BP. In thisparticular developmental stage (i.e. the Middle Holocene), Coryluswas dominant at first, but then Quercus gained absolute dominanceover other broadleaf trees. After 6500 BP, broadleaf trees starteddeclining in favour of Picea and, in some places (like Jelení lou�zesite), also in favour of Fagus. This process (i.e. a decline of QuercetumMixtum associated with an expansion of Picea) reached itsmaximum around 3500 BP, when Abies started to occur. Such avegetation history may be attributed to general acidification of thelocal environment that, according to Pokorný and Kune�s (2005),has predominantly natural causes and can be ascribed to the mostdynamic mid-interglacial phase of the Iversen cycle (Iversen, 1958;Ralska-Jasiewiczowa et al., 2003; Birks and Birks, 2004). AlthoughFagus-Abies forests had been locally dominant since that time until

almost recently, Picea retained its importance. The dominance ofspruce in the pollen spectra is nevertheless probably caused bytaphonomic factors, as pollen sites are always situated withininversion valleys, whose wet substrates and inverse microclimateare favourable for spruce.

If we directly compare the pollen record with the charcoal recordfrom the bottom of valley, we can see many analogous trends invegetation history, even though humid valley bottoms take up onlysmall parts of sandstone areas. The profile Jelenní lou�ze documentedthree distinct peaks of microcharcoals throughout the pollen recordwhereas the profile Sk�rítkuv hrnec recorded higher amounts ofmicrocharcoal only before the late Bronze Age period. The subse-quent expansion of beech is probably connected with a decrease infire disturbances. Pollen analyses of the profile Jelení lou�ze haverevealed repeated declines in beech pollen after the reappearance offire events. Several studies document similar consequences of pastfire events (e.g. Nelle et al., 2010; Robin et al., 2012; Bradley et al.,2013). Fire-sensitive species such as Fagus are disfavoured by fireevents, while pine, spruce and pioneer taxa (i.e. Betula, Sorbus) takeadvantage of them. These considerations may apply to the inter-pretation of pedoanthracological results, pointing to possible biasesin the representation of the above-mentioned taxa.

Our pollen and charcoal analyses from the same layers in theKristova cave profile document different records of tree compo-sition. The pollen analysis recorded the presence of Abies, Alnusand broadleaf trees (Carpinus, Tilia), while the charcoal analysisdid not. The rock shelter is located in the upper part of a slope, andthe nearest wetland is situated 500 m from the site. These resultssuggest that not only the pollen of Alnus, but also of other, moremesophilous, taxa, might have come from more distant anddiverse habitats than wood intercepted by our anthracologicalanalysis.

Our analyses of charcoal from rock shelters seem to generallydocument higher vegetation diversity compared to pedoanthraco-logical results from nearby plateaus (Nov�ak et al., 2012b; Bobek,2013). The pedoanthracological profiles in Bohemian sandstoneregions were usually situated on flat sandstone plateaus are char-acterized by long-term dominance of Pinus, a common occurrenceof Picea, and only a limited presence of deciduous trees and Abies.These charcoal records showed a similar composition of vegetationas profiles under rock shelters from upper parts of slopes andconfirmed the linkage between charcoal records and the characterof local vegetation.

The recent character of the vegetation exhibits analogous fea-turese humid gorges, bottoms of valleys and banks of watercoursesharbour a large part of the species diversity found in the sandstonearea (S�adlo et al., 2011). Without doubt, both charcoal and pollenrecords can be considered related to past vegetation conditions.Both the pollen record from small peat bogs, and the anthracologicalrecord reconstruct similar trends in the history of local vegetation.However, the results of our pollen analysis may suggest a lessdifferentiated mixture of vegetation in the vicinity due to tapho-nomic factors. Pollen spectra are in such situations predominantlycollected from a single type of environment that is anomalous in thespatial sense e wet bottom parts of sandstone gorges. Due to thisfact, charcoal spectra, despite being not so well stratified anddatable, are probably more appropriate for reconstructing the treelayer composition of regional forests. However, integrating the re-sults of both investigation techniques is likely to paint the bestpicture.

6. Conclusion

Excavations at North-Bohemian rock shelters have uncoveredcomplex evidence of intermittent human presence since the Late

J. Nov�ak et al. / Quaternary International 366 (2015) 25e36 35

Palaeolithic to recent times. The eight profiles examined in thisstudy contained abundant charcoal assemblages.

The variability in charcoal taxa reflects changing environmentalconditions in the vicinity of the study sites. Significant differencesbetween charcoal records from rock shelters located in the upperparts of slopes and bottoms of valleys are apparent in the Meso-lithic period. Charcoals of Pinus were abundant, and Quercus char-coals were common, in all profiles analysed. At some of the sites,charcoals of oak and demanding broadleaf trees (Fraxinus, Ulmus,Tilia, Acer) gradually decreased since the Bronze Age, while pinecharcoals increased accordingly. Beech charcoals, albeit generallyrare, increased since the Bronze Age and decreased again duringthe Middle Ages. Several recorded woody species are now rare oreven absent in the sandstone regions under consideration (e.g.Taxus).

The two investigated pollen diagrams come from a small peatbog at the bottom of a sandstone gorge. They show some similar-ities, but also certain differences, in vegetation development. Thecombination of charcoal and pollen data provides a more detailedreconstruction of the vegetation history on a local scale than any ofthese methods would if used alone. The results of our charcoalanalyses closely correspond with the past arboreal composition inthe vicinity of the rock shelters under study. Our pollen analyses,however, present a less distinguishable mixture of surroundingvegetation, being biased by the massive influence of the localwetland environment.

Acknowledgements

We would like to thank anonymous referees for helpful com-ments on the previous version of this paper. This study was sup-ported by two grants from the Czech Science Foundation (GA�CR 13-11193S and GA�CR 13- 08169S).

References

Asouti, E., Austin, P., 2005. Reconstructing woodland vegetation and its exploitationby past societies, based on the analysis and interpretation of archaeologicalwood charcoal macro-remoins. Environmental Archaeology 10, 1e18.

Badal, E., Bernabeu, J., Vernet, J.L., 1994. Vegetation changes and human action fromthe Neolithic to the Bronze Age (7000e4000 B.P.) in Alicante (Spain) based oncharcoal analysis. Vegetation History and Archaeobotany 3, 155e166.

Bendel, M., Tinner, W., Ammann, B., 2006. Forest dynamics in the Pfyn forest inrecent centuries (Valais, Switzerland, Central Alps): interaction of pine (Pinussylvestris) and oak (Quercus sp.) under changing land use and fire frequency. TheHolocene 16, 81e89.

Birks, H.J.B., Birks, H.H., 2004. The rise and fall of forests. Science 305, 484e485.Bj€orkman, L., Bradshaw, R., 1996. The immigration of Fagus sylvatica L. and Picea

abies (L.) Karst. into a natural forest stand in southern Sweden during the last2000 years. Journal of Biogeography 23, 235e244.

Bobek, P., 2013. Dlouhodobý vliv po�z�ar�u na slo�zení vegetace. In: Seiler, U., Wild, J.(Eds.), Historische Waldentwicklung in der S€achsisch-B€ohmischen SchweizHistorický vývoj lesa v �Ceskosask�em �Svýcarsku. Rhombos-Verlag, Berlin,pp. 225e245.

Blackford, J.J., 2000. Charcoal fragments in surface samples following a fire and theimplications for interpretation of subfossil charcoal data. PalaeogeographyPalaeoclimatology Palaeoecology 164, 33e42.

Bradley, L.R., Giesecke, T., Halsall, K., Bradshaw, R.H.W., 2013. Exploring therequirement for anthropogenic disturbance to assist the stand-scale expansionof Fagus sylvatica outside southern Scandinavia. The Holocene 23, 579e586.

Brande, A., 2003. Late Pleistocene and holocene pollen stratigraphy of Lake Stechlin.Advances in Limnology 58, 281e311.

Bronk-Ramsey, C., 2001. Development of the radiocarbon calibration programOxCal. Radiocarbon 43, 355e363.

Brostr€om, A., Nielsen, A.B., Gaillard, M.-J., Hjelle, K., Mazier, F., Binney, H., Bunting, J.,Fyfe, R., Meltsov, V., Poska, A., R€as€anen, S., Soepboer, W., Stedingk, H.,Suutari, H., Sugita, S., 2008. Pollen productivity estimates of key European planttaxa for quantitative reconstruction of past vegetation: a review. VegetationHistory and Archaeobotany 17 (5), 461e478.

Carcaillet, C., Thinon, M., 1996. Pedoanthracological contribution to the evolution ofthe upper treeline in the Maurienne Valley (North French Alps): methodologyand preliminary data. Review of Palaeobotany and Palynology 91, 399e416.

Clarke, D., 1976. Mesolithic Europe: the economic basis. In: Sieveking, G.G.,Longeworth, I.W., Wilson, K.E. (Eds.), Problems in Economic and SocialArchaeology, London, pp. 449e481.

Clark, J.S., Lynch, J., Stocks, B.J., Goldammer, J.G., 1998. Relationships betweencharcoal particles in air and sediments in west-central Siberia. The Holocene 8,19e29.

Coles, J.M., Heal, S.V.E., Orme, B.J., 1978. The use and character of wood in prehis-toric Britain and Ireland. Proceedings of the Prehistoric Society 44, 1e45.

Couvert, M., 1968. �Etude des charbons pr�ehistoriques. M�ethodes de pr�eparation etd'identification. Libyca 16, 249e256.

Demek, J., 1987. Hory a ní�ziny. In: Zem�episný lexikon �CSR. Academia, Praha (inCzech), Praha.

Forbes, M.S., Raison, R.J., Skjemstad, J.O., 2006. Formation, transformation andtransport of black carbon (charcoal) in terrestrial and aquatic ecosystems. Sci-ence of the Total Environment 370, 190e206.

Fr�ejaville, T., Carcaillet, C., Curt, T., 2013. Calibration of charcoal production fromtrees biomass for soil charcoal analyses in subalpine ecosystems. QuaternaryInternational 289, 16e23.

Godwin, H., Transley, A.G., 1941. Prehistoric charcoals as evidence of former vege-tation, soil and climate. Journal of Ecology 29, 117e126.

Greguss, P., 1972. Xylotomie of the Living Conifers. Akademiai Kiado, Budapest.Grimm, E.C., 2011. Tilia Software v.1.7.16. Illinois State Museum, Springfield IL.H€artel, H., Cílek, V., Herben, T., Jackson, A., Williams, R., 2007. Sandstone Land-

scapes. Academia, Praha.Heinz, C., Barbaza, M., 1998. Environmental changes during the Late Glacial and

Post-Glacial in the central Pyrenees (France): new charcoal analysis andarchaeological data. Review of Palaeobotany and Palynology 104, 1e17.

Henry, A., Valdeyron, N., Bouby, L., Thery, I., 2012. History and evolution of Meso-lithic landscapes in the Haut-Quercy (Lot, France): new charcoal data fromarchaeological contexts. The Holocene 23, 127e136.

Iversen, J., 1958. The Bearing of Glacial and Interglacial Epochs on the Formationand Extinction of Plant Taxa, vol. 6. Upsala Universiteit Arssk, pp. 210e215.

Jahns, S., 2007. Palynological investigations into the Late Pleistocene and Holocenehistory of vegetation and settlement at the L€oddigsee, Mecklenburg, Germany.Vegetation History and Archaeobotany 16, 157e169.

Jankovsk�a, V., 1992. Vegetationsverhaltnisse und Naturumwelt des BeckensJest�rebska kotlina am Ende des Spatglazials und im Holozan (Doksy-Gebiet).Folia Geobotanica et Phytotaxonomica 27, 137e148.

Janssen, C.R., 1966. Recent pollen spectra from the deciduous and coniferous-deciduous forests of Northeastern Minnesota: a study in pollen dispersal.Ecology 47, 804e825.

Jansen, D., Mischka, D., Nelle, O., 2013. Wood usage and its influence on the envi-ronment from the Neolithic until the Iron Age: a case study of the graves atFlintbek (SchleswigeHolstein, Northern Germany). Vegetation History andArchaeobotany 22, 335e349.

J�egou, D., Cluzeau, D., Balesdent, J., Tr�ehen, P., 1998. Effects of four ecological cat-egories of earthworms on carbon transfer in soil. Applied Soil Ecology 9,249e255.

Kova�cíkov�a, L., Nov�ak, J., Prost�redník, J., 2012. Z�achranný archeologický výzkumKonejlovy jeskyn�e v Kloko�cských skal�ach. Fontes Nissae 8, 56e68 (in Czech).

Kuna, M., 2002. The Bohemian archaeological record e an attempt at an analyticalinformation system. In: Sanjuan, L.G., Wheatley, D. (Eds.), Mappingthe Future ofthe Past. New Information Technologies for Managing the European Archaeo-logical Heritage. Universidad de Sevilla, Sevilla, pp. 45e52.

Kuna, M., Hajnalov�a, M., Kova�ríkov�a, L., Lis�a, L., Nov�ak, J., Bure�s, M., Cílek, V.,Ho�sek, J., Ko�c�ar, P., Majer, A., Makowiecki, D., Van Nieuland, J., Scott-Cummings, L., Suvov�a, Z., Sv�etlík, I., Vandenberghe, D., Yost, C.H.L., Zabilska-Kunek, M., 2013. Ran�e st�redov�eký are�al v Roztok�ach z pohledu ekofakt�u. TheEarly Medieval site at Roztoky. The evidence of ecofacts. Pam�atky Archeologick�eCIV 59e147.

Kune�s, P., Jankovsk�a, V., 2000. Outline of late glacial and holocene vegetation in alandscape with Strong geomorphological gradients. Geolines 11, 112e114.

Kune�s, P., Pokorný, P., Jankovsk�a, V., 2007. Post-glacial vegetation development insandstone areas of the Czech Republic. In: H€artel, H., Cílek, V., Herben, T.,Jackson, A., Williams, R. (Eds.), Sandstone Landscapes. Academia, Praha,pp. 244e257.

Kune�s, P., Pokorný, P., �Sída, P., 2008. Detection of the impact of early Holocenehunter-gatherers on vegetation in the Czech Republic, using multivariateanalysis of pollen data. Vegetation history and Archaeobotany 17,269e287.

Küster, H., 1997. The role of farming in the post-glacial expansion of beech andhornbeam in the oak woodlands of central Europe. The Holocene 7,239e242.

Lo�zek, V., 1998. Late Bronze Age environmental collapse in the sandstone areas ofNorthern Bohemia. In: H€ansel, B. (Ed.), Mensch und Umwelt in der BronzezeitEuropas, Kiel, pp. 57e60.

Marston, J.M., 2009. Modeling wood acquisition strategies from archaeologicalcharcoal remains. Journal of Archaeological Science 36, 2192e2200.

Meduna, P., S�adlo, J., 2009. Bezd�ezsko e Dokesko. Krajina mezi odolnosti a stagnaci.Historick�a geografie 35, 147e160 (in Czech).

Mich�eli, E., Schad, P., Spaargaren, O. (Eds.), 2007. World Reference Base for SoilResources 2006, First Update 2007. FAO, Rome. World Soil Resources ReportsNo. 103.

Moore, P.D., Webb, J.A., Collinson, M.E., 1991. Pollen Analysis, second ed. BlackwellScientific Publications, Oxford.

J. Nov�ak et al. / Quaternary International 366 (2015) 25e3636

N�eme�cek, J., Kozak, J., Lerova, Z., 2008. Soil Maps of the Czech Republic 1:50 000.Czech Geological Survey, Praha.

Nelle, O., Dreibrodt, S., Dannath, Y., 2010. Combining pollen and charcoal: evalu-ating Holocene vegetation composition and dynamics. Journal of ArchaeologicalScience 37, 2126e2135.

Nelle, O., Robin, V., Talon, B., 2013. Pedoanthracology: analysing soil charcoal tostudy Holocene palaeoenvironments. Quaternary International 289, 1e4.

Nov�ak, J., Lis�a, L., Pokorný, P., Kuna, M., 2012a. Charcoal analyses as an envi-ronmental tool for the study of Early Medieval sunken houses infills inRoztoky near Prague, Czech Republic. Journal of Archaeological Science 39,808e817.

Nov�ak, J., S�adlo, J., Svobodov�a, H., 2012b. Unusual vegetation stability in a lowlandpine forest area (Doksy region, Czech Republic). The Holocene 22, 947e955.

Oeggl, K., Schoch, W., 2000. Dendrological analyses of artefacts and other remains.In: Bortenschlager, S., Oeggl, K. (Eds.), The Iceman and his Natural Environment,The Man in the Ice, vol. 4, pp. 29e61. Innsbruck.

Opravil, E., 2003. Rostlinn�e makrozbytky. In: Svoboda, J. (Ed.), Mezolit Severních�Cech. Archeological Institute ASCR, Brno, pp. 38e42 (in Czech).

Overballe-Petersen, M.V., Bradshaw, R.H.W., 2011. The selection of small hollows forpollen analysis in boreal and temperate forest regions. Palynology 35, 146e153.

Pernaud, M., 2001. Postglacial vegetation history in Luxembourg: new charcoal datafrom the cave of la Karelsl�e (Waldbillig, eastern Gutland). Vegetation Historyand Archaeobotany 10, 219e225.

Pokorný, P., Kune�s, P., 2005. Holocene acidification process recorded in three pollenprofiles from Czech sandstone and river terrace environments. Ferrantia 44,101e107.

Pokorný, P., �Sída, P., Chvojka, O., �Z�a�ckov�a, P., Kune�s, P., Sv�etlík, I., Veselý, J., 2010.Palaeoenvironmental research of the Schwarzenberg Lake, southern Bohemia,and exploratory excavations of this key Mesolithic archaeological area. Pam�atkyarcheologick�e 101, 5e38.

Preston, C.M., Schmidt, M.W.I., 2006. Black (pyrogenic) carbon: a synthesis of cur-rent knowledge and uncertainties with special consideration of boreal regions.Biogeoscience 3, 397e420.

Price, D., 1989. The reconstruction of Mesolithic diets. In: Bonsall, C. (Ed.), TheMesolithic in Europe. John Donald Publishing, Edinburgh, pp. 48e59.

Prost�redník, J., �Sída, P., 2009. Z�achranný archeologický výzkum pozdn�e st�redov�ek�eusedlosti s vodním dílem ve V�seni u Turnova. In: Pam�atky Libereck�eho kraje.Sborník NPÚ e ÚOP, Liberec, pp. 147e158 (in Czech).

Ralska-Jasiewiczowa, M., Nalepka, D., Goslar, T., 2003. Some problems of foresttransformation at the transition to the oligocratic/Homo sapiens phase of theHolocene interglacial in northern lowlands of central Europe. Vegetation His-tory and Archaeobotany 12, 233e247.

Regnell, M., 2012. Plant subsistence and environment at the Mesolithic site Tagerup,southern Sweden: new insights on the ‘‘Nut Age’’. Vegetation History andArchaeobotany 21, 1e16.

Ritchie, J.C., 1995. Current trends in studies of long-term plant community dy-namics. New Phytologist 130, 469e494.

Robin, V., Rickert, B.-H., Nadeau, M.J., 2012. Assessing Holocene vegetation and firehistory by a multiproxy approach: the case of Stodthagen Forest (NorthernGermany). The Holocene 22, 337e346.

Robin, V., Bork, H.R., Nadeau, M.J., Nelle, O., 2014. Fire and forest history of centralEuropean low mountain forest sites based on soil charcoal analysis: the case ofthe eastern Harz. The Holocene 24, 35e47.

Rowley-Conwy, P., 1986. Between cave painters and crop planters: aspects oftemperate European Mesolithic. In: Zvelebil, M. (Ed.), Hunters in Transition.Mesolithic Societies of Temperate Eurasia and Their Transition to Farming.Cambridge University Press, Cambridge, pp. 17e32.

Rubiales, J.M., Hern�andez, L., Romero, F., Sanz, C., 2011. The use of forest resources incentral Iberia during the Late iron Age. Insights from the wood charcoal analysisof Pintia, a Vaccaean oppidum. Journal of Archaeological Science 38, 1e10.

S�adlo, J., Pet�rik, P., Boublik, K., 2011. Diverzita rostlinstva Hrad�canskych st�en(Dokesko) a jeji p�ri�ciny. Zpr�avy �Cesk�e botanick�e spole�cnosti 46, 17e38 (inCzech).

Salisbury, K.J., Jane, F.W., 1940. Charcoals from Maiden Castle and their significancein relation to the vegetation and climatic conditions in prehistoric times.Journal of Ecology 28, 310e325.

Schweingruber, F.H., 1990. Microscopic Wood Anatomy. Swiss Federal Institute ofForestry Research, Birmensdorf.

Shackleton, C.M., Prins, F., 1992. Charcoal analysis and the “principle of least effort”-a conceptual model. Journal of Archaeological Science 19, 631e637.

Smith, C.S., McNees, L.M., 2011. Persistent land use patterns and the mid-Holocenehousepits of Wyoming. Journal of Field Archaeology 36, 298e311.

Sugita, S., 1994. Pollen representation of vegetation in Quaternary sediments:theory and method in patchy vegetation. Journal of Ecology 82, 881e897.

Svoboda, J.A. (Ed.), 2003. Mezolit severních �Cech. Komplexní výzkum p�revis�u na�Ceskolipsku a D�e�cínsku, 1978e2003 (Mesolithic period in Northern Bohemia.Complex excavation of rockshelters in the �Cesk�a Lípa and D�e�cín areas, 1978 e2003). The DolníV�estonice Studies. Brno.

Svoboda, J., Hajnalov�a, M., Hor�a�cek, I., Nov�ak, M., P�richystal, A., �Sajnerov�a, A.,Yaroshevish, A., 2007. Mesolithic settlement and activities in rockshelters of theKamenice river canyon, Czech Republic. Eurasian Prehistory 5 (1), 95e127.

Svoboda, J., Nov�ak, J., Nov�ak, M., S�azelov�a, S., Demek, J., Hladilov�a, �S., Pe�sa, V., 2013.Paleolithic/Mesolithic stratigraphic sequences at Údolí samoty and Janovaz�atoka rockshelters, North Bohemia: preliminary report. Arch€aologisches Kor-respondenzblatt 43, 469e488.

Svoboda, J., 2014. Post-Glacial Transformations: Danubian Europe. Oxford Hand-books Online. http://dx.doi.org/10.1093/oxfordhb/9780199551224,013,042.

�Sída, P., Chvojka, O., Pokorný, P., Nov�ak, J., 2011. Výsledky první etapy výzkumumezolitick�e lokality �Svarcenberk 11 u Pon�edr�a�zky (okr. Jind�rich�uv Hradec). AVJ�C24, 5e18 (in Czech).

�Sída, P., Prost�redník, J., 2007. Pozdní paleolit a mezolit �Cesk�eho r�aje: perspektivypozn�aní regionu. In: Archeologick�e rozhledy, vol. LIX, pp. 443e460 (in Czech).

Talon, B., Carcaillet, Ch., Thinon, M., 1998. Etudes pedoanthracologiques des varia-tions de la limite superieure des arbres au cours de l'Holocene dans les AlpesFrancaises. G�eographie physique et Quarternaire 52, 1e14.

ter Braak, C.J.F., �Smilauer, P., 2002. CANOCO Reference Manual and CanoDraw forWindows User's Guide: Software for Canonical Community Ordination (Version4.5). Microcomputer Power, Ithaca.

Th�ery-Parisot, I., Chabal, L., Chrzavzez, J., 2010. Anthracology and taphonomy, fromwood gathering to charcoal analysis. A review of the taphonomic processesmodifying charcoal assemblages, in archaeological contexts. PalaeogeographyPalaeoclimatology Palaeoecology 291, 1e142.

Zvelebil, M., 1994. Plant use in the Mesolithic and its role in the transition tofarming. Proceedings, Prehistory Society 60, 35e74.