Revue internationale d’écologie méditerranéenne ... · In this study the effects of wood-pasturage on species composition and forest structure in the Quercus frainetto forest

Revue internationaled’écologie méditerranéenne

International Journalof Mediterranean Ecology



SOMMAIRE – CONTENTS

P. D. DIMOPOULOS, E. BERGMEIERWood pasture in an ancient submediterranean oak forest . . . . . . . . . . . . . . . . . . . .137

M. DUBAR, BUI-THI-MAI, S. NICOL-PICHARD & M. THINONÉtude palynologique du carottage de Pont d’Argens (Roquebrune-sur-Argens, Var) : histoire holocène de la végétation en Provence cristalline ; facteurs naturels et anthropiques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .147

K. KASSIOUMIS, K. PAPAGEORGIOU, T. GLEZAKOS & I.N. VOGIATZAKISDistribution and stand structure of Taxus baccata populations in Greece; Results of the first national inventory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .159

CARSTEN F. DORMANN, RACHEL KINGComparing the palatability of Mediterranean or non-native plants in Crete . . . . . .171

M. CONEDERA, M.C. MANETTI, F. GIUDICI & E. AMORINIDistribution and economic potential of the Sweet chestnut (Castanea sativa Mill.) in Europe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .179

J.-C. THIBAULT, R. PRODON & P. MONEGLIAEstimation de l’impact des incendies de l’été 2000 sur l’effectif d’un oiseau endémique menacé : la sitelle corse (Sitta whiteheadi) . . . . . . . . . . . . .195

C. RATHGEBER, L. BLANC, C. RIPERT & M. VENNETIERModélisation de la croissance en hauteur du pin d’Alep (Pinus halepensis Mill.) en région méditerranéenne française . . . . . . . . . . . . . . . . . .205

M. BOUAZZA, N. BENABADJI, R. LOISEL & G. METGEÉvolution de la végétation steppique dans le sud-ouest de l’Oranie (Algérie) . . . . .219

FAITS DE CONSERVATION EN MÉDITERRANÉE

MEDITERRANEAN CONSERVATION NEWS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .233

ANALYSES D’OUVRAGES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .245

ANNONCE DE COLLOQUE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .248

ec

olo

gia

me

dite

rra

nea

Tome 30 fascicule 2, 2004ISSN 0153-8756

Tome 30 Fascicule 2, 2004ISSN 0153-8756

Tom

e 3

0 f

asc

icu

le 2

, 2

00

4ecologia mediterranea

Revue indexée dans Pascal-CNRS et Biosis

ecologiamediterranea

Tome 30 • Fascicule 2 • 2004



ISSN 0153-8756

Rédacteur en chef • Managing editor Secrétariat • Secretariat FRÉDÉRIC MÉDAIL MICHELLE DOUGNY

Rédacteurs • Editors LAURENCE AFFRE PHILIP ROCHE

THIERRY DUTOIT THIERRY TATONI

JÉRÔME ORGEAS ERIC VIDAL

Fondateur • FounderPROFESSEUR PIERRE QUÉZEL

Comité de lecture • Advisory board

ARONSON J., CEFE-CNRS, Montpellier

BARBERO M., IMEP, Université Aix-Marseille III

BEAULIEU J.-L. DE, IMEP, Université Aix-Marseille III

BROCK M., University of New England, Armidale, Australie

CHEYLAN M., EPHE, Montpellier

DEBUSSCHE M., CEFE-CNRS, Montpellier

FADY B., INRA, Avignon

GRILLAS P., Station biologique Tour du Valat, Arles

GUIOT J., CEREGE-CNRS, Aix-en-Provence

HOBBS R. J., CSIRO, Midland, Australie

KREITER S., ENSA-M-INRA, Montpellier

LE FLOC’H E., CEFE-CNRS, Montpellier

MARGARIS N. S., University of the Aegean, Mytilène, Grèce

OVALLE C., CSI-Quilamapu, INIA, Chili

PEDROTTI F., Universita degli Studi, Camerino, Italie

PLEGUEZUELOS J. M., Université de Grenade, Espagne

PONEL P., IMEP, CNRS, Marseille

PRODON R., EPHE, Montpellier

RIDCHARSON D. M., University Cape Town, Afrique du Sud

SANS F. X., Université de Barcelone, Espagne

SHMIDA A., Hebrew University of Jérusalem, Israël

TROUMBIS A., University of the Aegean Mytilene, GrèceURBINATI C., Agripolis, Legnaro, Italie

Abonnements • Subscription (contacter Edisud)Un an : 2 numéros • One year : 2 issues

— France : 61 € + 9,12 € de frais de port— Europe : 61 € + 12,2 € de frais de port

— Amérique, Afrique, Asie : 61 € + 18,3 € de frais de port

ecologia mediterranea

Ecologia MediterraneaEuropôle méditerranéen de l’Arbois, Bâtiment Villemin, B.P. 80

F-13545 Aix-en-Provence CEDEX 04, FranceTél. : + 33 04 42 90 84 06 – Fax : + 33 04 91 28 80 51

Internet : http://www.imep-cnrs.com/ecologia/ – e-mail : [email protected]

Éditions Édisud / Atelier graphique Édisud

3120 route d’Avignon, F-13090 Aix-en-Provence, FranceTél. : 00 33 04 42 21 61 44 – Fax : 00 33 04 42 21 56 20

Internet : http://www.edisud.com – e-mail : [email protected]

© Édisud, 2004, tous droits réservés.

ecologiamediterranea



Tome 30 • Fascicule 2 • 2004

137

ecologia mediterranea, tome 30, fascicule 2, 2004, p. 137-146

AbstractIn this study the effects of wood-pasturage on species composition and forest structure in the Quercus frainetto forest of Folói are des-cribed. This is the most extensive broadleaved forest of Peloponnese and southern Greece and unique in that there is evidence of several thousand years of existence. The variation in plant species compo-sition among, and the differences between, grazed and ungrazed forest stands are analysed by means of ordination (correspondence analysis). Species indicative for grazing or its withdrawal are listed. Annuals and certain perennials with good regeneration capacity are indicative for grazed plots, while a dense shrub layer with Arbutus unedo and Erica arborea is related to ungrazed plots. Generally, in the absence of grazing the development of the herb and shrub layer is enhanced. Forest stands in exclosures tend to produce denser cano-pies, oak rejuvenation is more abundant, and the trees are higher and more vital than outside. In grazed woodland, litter and organic matter are less abundant and the degree of parasitism by Loranthus europaeus is higher. Our results suggest two possible conservation options for the study area, viz. (a) controlled grazing regime in the framework of a traditional but sustainable agro-silvopastoralistic system or (b) a concept towards a natural forest ecosystem.

Key-wordsGrazing, Greece, Old forest, Quercus frainetto, Silvopastoralism

RésuméDans la présente étude sont décrits les effets du sylvopastoralisme sur la composition spécifique et la structure de peuplements dans la forêt à Quercus frainetto de Folói. Cette forêt, la plus étendue du Péloponnèse et du sud de la Grèce, est unique par le fait de son existence probablement plurimillénaire. La variation de la com-position spécifique au sein de peuplements pâturés ou non, et les différences entre ces deux types de peuplements ont été analysées par ordination (analyse des correspondances). Des espèces indicatrices du pâturage et de son absence sont listées. Les espèces annuelles et certaines vivaces avec une bonne capacité de régénération sont liées aux placettes pâturées, tandis qu’une strate arbustive dense formée par Arbutus unedo et Erica arborea caractérise les placettes non pâturées. De façon générale, en l’absence de pâturage, les strates herbacées et arbustives sont mieux développées. Les forêts à l’inté-rieur des enclos ont tendance à produire des canopées plus denses, la régénération des chênes y est plus abondante, et les arbres sont plus hauts et plus vigoureux qu’à l’extérieur. Dans les peuplements pâturés, la litière et la matière organique sont moins abondantes, et le degré de parasitisme par Loranthus europaeus est plus élevé. Nos résultats suggèrent deux options possibles pour la conservation de la zone étudiée : a) un régime de pâturage contrôlé dans le cadre d’un système agropastoral traditionnel mais durable ou b) un régime visant un écosystème forestier à caractère naturel.

Mots-clésPâturage, Grèce, vieille forêt, Quercus frainetto, Sylvopastoralisme

Wood pasture in an ancient submediterranean oak forest (Peloponnese, Greece)Sylvopastoralisme dans une ancienne forêt méditerranéennede chênes (Péloponnèse, Grèce)

P. D. Dimopoulos1, E. Bergmeier2

1. Department of Environmental and Natural Resources Management, University of Ioannina, Seferi 2, GR-30100 Agrinio, Greece; Fax +30 641 39576: E-mail: [email protected]

2. Albrecht-von-Haller-Institut für Pflanzenwissenschaften, Georg-August-Universität Göttingen, Untere Karspüle 2, D-37073 Göttingen, Germany; Fax +49 551 39 2287; E-mail: [email protected] *corresponding author

138

◆ P. D. DIMOPOULOS & E. BERGMEIER

ecologia mediterranea, tome 30, fascicule 2, 2004

INTRODUCTION

Interest in wood pasture has increased a great deal lately in many European countries (Papanastasis et al., 1999; Redecker et al., 2002). In western and central Europe re-introduction of wood pasture is currently under discussion, with the principal aims of enhancing forest dynamics and increasing biodiversity (Pott, 1999; Vera, 2000; Schmidt & Heile, 2001; Spencer, 2002). Modern forestry supports tall dense forest for economic reasons and, owing to the browsing of seedlings and juve-nile trees, considers wood-pasturage as detrimental to the forest and chiefly responsible for the decline of wooded areas and the structural senescence of the tree stands. Present silvopastoralism in Europe is largely restricted to countries of the wider Mediterranean and the Balkans. Grazing by domestic animals has widely been practiced in virtually all forests except for the most remote ones. In traditional silvopastoral farming systems, submedi-terranean deciduous and mediterranean sclerophyllous woodlands are chiefly involved. Deciduous oak forest is highly esteemed due to its mast production in autumn as food for pigs. But also other domestic animals such as sheep, goats and cattle benefit from the relatively light conditions in oak woodlands which support a fairly dense and plant-rich ground vegetation.

For the present study, the forest of Folói (Kápellis; Pholóë in antiquity; Ilía, Peloponnisos, Greece) was cho-sen. It has been used for charcoal burning and pasturage for centuries, as travellers’ reports suggest (Philippson, 1892; Pritzel, 1908; Rothmaler, 1943). The major part has been used as wood pasture for sheep and pigs within the local rural economies but remote parts show little or no signs of grazing at present.

In Greece, as elsewhere in the Mediterranean, most oak woodlands have been, or still are, subject to coppi-cing at more or less regular intervals. Single-stemmed old-growth oak woodlands as in our study area, however, are exceedingly rare (Bergmeier et al., 2004). The forest of Folói is the most extensive submediterranean non-cop-piced oak forest in Peloponnisos, and certainly among the oldest existing. It was known already in Greek mythology, according to which it was frequented by centaurs, hor-ses with human body, the personifications of mountain forest wilderness. Among the many myths around the Greek hero Herakles is one with the forest of Folói as the scenery of a guest meal provided for the hero that ended up in a massacre among the centaurs. Herakles also encouraged prehistoric people to clear part of the extensive forest. The myth can be interpreted figurati-

vely as an attempt to civilize wilderness and to establish cultivation in less favourable regions. Although close to Olympia the hinterland of Ilía remained a little attended region in antiquity. To our knowledge, the first mentio-ning of an historical event in the forest area of Folói dates back to the battle between Alarichos and Stilichon in AD 397 (Christopoulos, 1978).

Grazing by domestic herbivores causes quantitative (number of plant individuals, species numbers) and qua-litative (species composition, phenological traits) effects on Mediterranean open habitats (Noy-Meir et al., 1989; Fernandez Alés et al., 1993; Bergmeier & Matthäs, 1996; Bergmeier, 1998). However, not many studies have dealt with woodland grazing in the Mediterranean, or the effects of cessation of grazing to Mediterranean deci-duous woodlands (Di Pasquale & Garfi, 1998; Debussche et al., 2001). Few studies explicitly state which species in oak woodlands increase after silvopastoralism has been given up, and which decrease (Debussche et al., 2001). Based on field studies in the Foloi forest, our paper attempts to provide such information and addresses the following questions:

Is the species composition of oak forest a suitable indicator for wood pasture?

Which species of oak forest profit from grazing and which from its withdrawal?

To which extent differs the structure of ungrazed oak forest from that of grazed stands?

Is silvopastoralism an obstacle to the rejuvenation of oak?

Ancient Mediterranean forests are in urgent need of protection but conservation priorities have hardly been discussed, owing chiefly to the lack of ecological infor-mation. Our final point of discussion is therefore: Which conclusion can be drawn from our findings for the con-servation and sustainable use of the specific study area?

STUDY AREA

The study area, the forest of Folói (Kápellis), com-prises 3100 ha of broadleaved forest. It is situated in the eastern part of Ilía (Elis in antiquity) in western Peloponnisos (fig. 1). The Folói plateau constitutes the upper and most extensive in a series of conglomerate tables between Mt Erimanthos and the river Alpheios (Philippson, 1959). The plateau is very slightly inclined with less than 700 m of altitude in the north and almost 800 m in the south. It consists of Pleistocene continental

139

WOOD PASTURE IN AN ANCIENT SUBMEDITERRANEAN OAK FOREST ◆


deposits, represented chiefly by conglomerates (IGME, 1983). The soils are commonly fairly deep cambisols, acidic, base-poor and waterpermeable. The climate is Mediterranean-type, with mild humid winters and dry warm summers, but in the western Peloponnese modi-fied towards sub-oceanic conditions, particularly so in the mountains. That is why, in spite of the southern position, the annual precipitation is above 1000 mm, and the arid period restricted to comprise about 90 days (meteorologi-cal station Andritsena, unpubl. data). Due to the infertility of the soils, the area was never densely populated. This is why a considerable extent of the forest was preserved

to our days. Philippson (1892: 36), however, mentioned forest degradation caused by extensive charcoal produc-tion. Beside of charcoal industry, grazing by domestic animals (sheep, pigs, more rarely goats) has widely been practiced in the forest (Rothmaler, 1943) and is still of considerable importance for the local rural economies. The deciduous Quercus frainetto is the predominant oak, as was already mentioned by Heldreich (1862), Pritzel (1908) and Rothmaler (1943). There are no traces of fire in the present oak forest, nor is fire mentioned as a means of forest or grazing management by the early geobotani-cal travellers cited above. Extensive nearby areas of pine

Figure 1. Position of the study area and distribution of Quercus frainetto forest drawn from an aerial photograph of 1992. Tree crowns in stands with open canopy (1) do not overlap and the cover is open or sparce (< 40 %); closed stands (2) with overlapping crowns

represent dense canopy covers (> 60 %); intermediate stands (40-60 %) are infrequent and included into (2).

140



forest, on the other hand, burnt in the 1990s. The fire did not expand to the oak forest. In spite of the serious human impact, major areas remained spacious tall forest of single-stemmed, non-coppiced trees, today an excep-tionally rare type of woodland in southern Greece and elsewhere in the Mediterranean.

METHODS

Silvopastoral activities (expansion or decrease of the grazed area; intensity; composition of livestock) have always been fluctuating in history, along with various socio-economic factors. Major parts of the forest are currently subject to grazing, but as for the rest, there is no way telling when exactly a given site was grazed last. Therefore, we distinguished between stands which are presently grazed, and others without present grazing impact. The absence of grazing was judged from the absence of browsed or grazed plants and from the lack of droppings. The species composition of grazed and non-grazed stands was studied in 30 and 12 quadrats, respectively, each of 400 m². Minimum distance between quadrats was c. 150 m, but usually more than 300 m. In order to restrict the selection to sites comparable in terms of abiotic parameters, only forests with more or less closed Quercus frainetto canopy, i.e., with more than 60 % canopy cover were included. Most stands have 65-85 % canopy cover (figure 1). Stands on steep slopes and in ravines were excluded. Species abundance in tree (t), shrub (s) and herb (h) layers were distinguished (t > 4 m; 4 m > s > 100 cm; h < 100 cm).

Four of the ungrazed quadrats were in two grazing exclosures of about two hectares in total which had been established in April 1964. In one of the exclosures (N37°46’39”, E21°44’46”), for silvicultural purposes a stand analysis had been performed in November 1964 (Panagiotidis, 1965). Among other parameters, stem diameter at breast height (BHD) and tree height (TH) had been assessed. We performed a similar analysis in 1999 using one plot of 8 × 50 m in the exclosure, the other of the same size in a grazed site further east 30 m outside the fence. Stand profiles and crown projection maps were drawn, and pH as well as visual properties of soil profiles were assessed. The following parameters were recorded per tree: BHD, TH, number of oak mistles (Loranthus europaeus) as an indicator for reduced vitality (only medium-sized to large Loranthus individuals were counted since smaller ones would easily have been

overlooked). The number of juvenile oaks in ungrazed forest was assessed in 8 plots of 1 m² each, laid out at regular intervals along a diagonal inside the exclosure, and for the grazed forest the same number of plots was arranged along an outside extension of the diagonal.

In order to explore the relevance of grazing for explaining the variation in the data set, the 42 relevés (30 in grazed sites, 12 in ungrazed sites) were subjected to indirect gradient analysis (Correspondence analysis, CA). The settings were biplot scaling with focus on inter-species distances, no downweighting of species, and no transformation. The ordinations were performed using CANOCO 4 (ter Braak & Šmilauer, 1998). For statistical calculations on species frequency, forest and tree parameters, seedling numbers, and the degree of mistle infection, Mann-Whitney U-Test was used, with the significance levels expressed by 2-tailed Monte Carlo significance.

Nomenclature of taxa follows Flora Europaea (Tutin et al., 1968-1980, 1993).

RESULTS

The ordination of the 42 quadrats of Quercus frainetto forest by means of CA revealed an arched plot structure which is expected for data sets with one predominant gradient (figure 2). Along the horizontal axis (axis 1), gra-zed plots formed the left wing of the arch, while the right wing was composed of non-grazed plots. Hence grazing regime constitutes the most important gradient explaining a great deal of variation in species composition. Species such as Cynosurus echinatus, Poa bulbosa and Trifolium campestre scored in the far left of the diagram, indicating their preferential occurrence in grazed stands. In contrast, Arbutus unedo and Erica arborea shrubs turned out to be characteristic for non-grazed plots. In the central part of the diagram taxa without clear preference to any mana-gement regime were assembled.

Species preferences for grazed and non-grazed forests are displayed in more detail in table 1. Annual species were found to be restricted largely to the grazed plots. Among the perennials, certain species with the poten-tial to resprout from basal buds or subterranean tubers (Oenanthe pimpinelloides, Asphodelus ramosus, Poa trivialis subsp. sylvicola, Poa bulbosa) occurred significantly more frequently in grazed plots. Shrub species, in particular Arbutus unedo and Erica arborea, to somewhat lesser degree also juvenile plants of these species in the herb layer, are a specific feature of non-grazed plots.

141



The mean cover values of the Quercus frainetto canopy and the field layer tend to be higher in the non-grazed plots though not significantly (table 2). Quercus frainetto in the shrub layer occurred with high constancy in both regime types but was more abundant in the non-grazed plots. The cover of the shrub layer was very variable both in grazed and non-grazed stands, chiefly due to the varia-tion in cover values of the Q. frainetto understorey, but significantly and altogether more than three times higher in non-grazed than in grazed stands (table 2).

Stand analyses of a non-grazed plot in an exclosure and a grazed one nearby outside the fence revealed significantly higher values for tree height (TH) in the non-grazed plot while mean stem diameter (BHD) was

lower than in the grazed plot (table 3). If compared with the mean values for all trees in 1964, the trees have become 5 m taller at an average within 35 years, and the BHD increment was 7 cm. The ratio TH/BHD remai-ned almost constant in the exclosure while in the grazed stand a considerable decrease was noted (table 3). The age structure of the trees is uneven, and in 1964 11 % of the trees had BHD values > 40 cm (Panagiotidis, 1964). The crown projection revealed more than 80 % canopy cover in the fenced plot, as against about 70 % in the neighbouring grazed one (figures 3 and 4). The lower non-branched part of the stems is generally longer in the exclosure, and the branches in the lower two thirds of the trees are more scattered and with less foliage. Dead

Figure 2. Ordination diagrams (correspondence analysis) displaying floristic similarities of 30 grazed and 12 non-grazed plots (dots, left diagram) and selected species scores (right diagram). Species names are abbreviated by 4 letters of the generic name and 3 of the specific epithet (full names in table 1). Their positions were slightly adjusted if necessary to avoid overlap. Eigenvalues of axis 1: 0.229, axis 2: 0.168.

142



branches are more numerous in the grazed plot. The oak mistle Loranthus europaeus occurred with a mean of 3.2 (± 2.6) individuals per tree while the ratio was 0.9 (± 1.2) Loranthus individuals per tree in the exclosure (p = 0.006). We have found 27 (± 19) juveniles of Q. frainetto per m² in the grazed and 60 (± 24) in the ungrazed plots (p = 0.012) (table 3).

The soil profiles were roughly 3-layered both inside and outside the exclosure, and the soil type was identified as cambisol (Braunerde). Base saturation is low, and the deep lime-free B horizon is markedly acidic (pH 4.6-5.5). Differences between the plots refer particularly to the humus layer. In the grazed plots, the latter is absent or, if present, thin (up to 2 cm) and largely without obvious mycelia. Litter is sparse and often absent. In the exclosu-res, there is high fungal activity in the humus layer, and litter in various stages of decomposition was 2-7 cm thick and covers most of the surface area.

DISCUSSION

The species composition of grazed Q. frainetto forests differs considerably from that of non-grazed stands. Annual species in particular qualify as grazing indicators, at least in dense stands with a more or less closed canopy. Among the perennial herbs and subshrubs, most species of deciduous oak forest seem to be not or negatively affected by grazing. Exceptions include Oenanthe pimpi-nelloides and Poa trivialis subsp. sylvicola, both supplied with subterranean tuberous or knotted swellings which enable regeneration; Asphodelus ramosus, largely avoided by herbivores and locally abundant in overgrazed pastu-res, and Poa bulbosa, a mat-forming pasture grass.

The forest structure in non-grazed stands, as exempli-fied by the exclosures, is denser, the trees are taller, and there is pronounced canopy competition. The trees in the

Treatment grazed not grazed p

Number of plots 30 12

Woody species

Arbutus unedo s . 100 ***

Arbutus unedo h 3 83 ***

Erica arborea s 20 91 ***

Rubus canescens 53 91 ***

Sorbus torminalis h 3 33 ns

Annuals

Cynosurus echinatus 63 . ***

Trifolium campestre 46 . *

Aira elegantissima 30 . ns

Cerastium brachypetalum 30 . ns

Perennial herbs and subshrubs

– more frequent in grazed plots

Oenanthe pimpinelloides 96 58 ***

Poa trivialis ssp. sylvicola 90 50 ***

Asphodelus ramosus 67 25 **

Trifolium physodes 88 83 **

Poa bulbosa 50 16 ns

– more frequent in ungrazed plots

Stipa bromoides 46 100 *

Clinopodium vulgare 36 75 *

Potentilla micrantha 76 100 *

Brachypodium sylvaticum 85 100 *

Aremonia agrimonoides 50 83 *

Symphytum bulbosum 43 75 *

Brachypodium rupestre 13 50 ns

Teucrium chamaedrys 3 41 ns

Dorycnium hirsutum 6 41 ns

Cephalanthera longifolia 6 41 ns

Achillea ligustica 6 33 ns

Lathyrus laxifl orus 76 100 ns

Luzula forsteri 82 100 ns

Table 1. Constancy values (given in %) and frequency differences of selected species in grazed

and not grazed Quercus frainetto forest. s – shrub layer (1-4 m), h – herb layer (< 1 m). Frequency differences indicated by Mann-Whitney U-test significance levels: P < 0.05: *;

P < 0.01: **; P < 0.005: ***; ns: not significant.

Grazed Not grazed p

number of plots 30 12

canopy cover 71.0 ± 11.8 75 ± 8.4 0.086 (ns)

cover shrub layer 11.1 ± 16.1 39.5 ± 22.2 0.000 (***)

cover herb layer 43.7 ± 20.7 50.5 ± 18.5 0.297 (ns)

Table 2. Oak forest parameters of the grazed and non-grazed plots. Mean cover values are in % with standard deviation. Significance levels

are * (p < 0.05); ** (p < 0.01); *** (p < 0.001); ns, not significant.

143



grazed stands tend to be less high, with more branches in the trunk area and with more investment into radial stem growth. The higher degree of infection by parasitic Loranthus suggests that oak trees are less vital in heavily grazed forest. Since geological and topographic condi-tions are largely identical for all plots, differences in the

vitality of trees and in species composition are almost certainly related to the considerable differences obser-ved in present litter and humus layers. The litter layer and the higher content of organic matter are important in maintaining rapid infiltration rates, absorbing many times its own weight of water (Pritchett & Fisher, 1987).

Year 1964 1999 1999

Treatment grazed grazed not grazed

number of trees 524 10 9

TH (m) 20.6 ± 1.6 23.9 ± 2.8 25.6 ± 1.8 p = 0.009 **

BHD (cm) 22.4 ± 3.1 34.5 ± 4.5 29.5 ± 4.5 p = 0.049 *

TH/BHD 92 69 87

Loranthus individuals per tree 3.2 ± 2.6 0.9 ± 1.2 p = 0.006 **

number of juv. oaks 27 ± 19 60 ± 24 p = 0.012 *

Table 3. Tree and site parameters obtained from stand analyses in 1964 when the exclosure was installed, and in 1999 in the exclosure and outside next to it. Mean values with standard deviation are given. TH = mean tree height, BHD = mean breast height diameter.

Significance levels refer to stand analyses in 1999; levels as in table 2.

Figure 3. Stand profile (a depth of 8 m is recognized) and crown projection of a grazed Quercus frainetto forest stand. Dots on the projection indicate the position of stems, triangles that of weathered stumps.

144



In the exclosure, litter covers the ground almost totally, and moisture is retained much more effectively. Annuals occur only on naked mineral soil, in conditions which are absent in the exclosure (or restricted to stem bases). This is the result of wind turbulence dislocating foliage and preventing accumulation of organic matter (Wilke et al., 1993). The wind effect is reduced if a shrub or subshrub layer is developed. Such a layer is represented in the oak forest by the regrowth of Q. frainetto, the most frequent species in the herb layer, and, in non-grazed stands, also by Arbutus unedo and Erica arborea. In a study on post-grazing successional oak woodland in southern France Debussche et al. (2001) found shrub species among the increasing taxa but not among the decreasing. Grazing (by sheep and pigs as in the study area) does not prevent oak rejuvenation but young oaks are less abundant. In fact, since grazing is likely to prevent a dense Arbutus and Erica understorey, moderate silvopastoralism might even favour Q. frainetto rejuvenation. In exclosures, the

dense herb layer of oak seedlings and saplings supports litter and humus accumulation, thus improving soil water and nutrient conditions which, in turn, are favourable for rejuvenation. Browsing may be of little direct effect on the juveniles but trampling is destructive to the herb layer. Soil compaction and depletion are common features in grazed woodlands (Bezkorowajnyj et al., 1993; Sibbald, 1999). Our findings suggest that they may be interpreted as an indirect effect of animals due to decreased litter accumulation rather than directly by trampling.

CONCLUSION AND FINAL REMARKS

The forest of Folói forms part of an area named ‘Oropedio Folois’ (9723 ha), chosen to become a Special Conservation Area, eligible to be included in the European ‘Natura 2000’ network of Sites of Community

Figure 4. Stand profile (a depth of 5 m is recognized) and crown projection of an

ungrazed Quercus frainetto forest exclosure plot. Dots on

the projection indicate the position of stems; triangles

that of weathered, sixangles that of resprouting stumps.

Quadrats in the profile indicate Arbutus

rejuvenation of >1 m, circle: resprouting Q.frainetto

stump.

145



Interest. Regulative arrangements and administrative measures for Special Conservation Areas are currently initiated, including the establishment of a management plan. Such a plan, however, cannot be worked out unless the conservation priorities are clearly defined. From our study two possible concepts may be suggested:

(a) Folói represents an outstanding example of an agro-silvopastoral system. Such ecosystems are vanishing in Europe and almost lost in most countries. They are considered a traditional asset worthy of protection. The history of human interference in the Folói area is long but today’s combined impacts on the forest (grazing, charcoal production, forestry, agriculture) are far from sustainable. Maintaining wood pasture in Folói requires a balanced grazing regime and a strict control of other kinds of impact.

(b) On the other hand, Folói constitutes a unique exam-ple of submediterranean tall oak forest. As our study shows, it it severely suffering in places from grazing but the conditions of regeneration towards a natural forest are better than anywhere else.

Any management plan and conservation measures depend on which alternative is given priority. It is clear, from the results of our paper, that the two conservation visions can hardly be realized simultaneously in one and the same site. It is also evident that a Special Conservation Area cannot be established without the acceptance and co-operation of the resident farmers and villagers. The management plan will have to make an attempt to accomodate both options: e.g., by installing a core zone where grazing is to be prohibited, and a buffer zone with controlled grazing regime. With the establishment of a European network of conservation sites the problem of harmonization of the options natural forest and traditio-nal silvopastoralism will be of increasing relevance. Folói may well serve as a model on how to balance the respec-tive management and conservation measures.

ACKNOWLEDGEMENTS

We thank L. Boskos, F. Galanos, G. Karetsos and K. Varelides, Research Forest Institute of Athens (NAGREF), for supplying us with stand analysis data from the late N. Panagiotidis; H. Dres, retired forester of the Folói forest, for discussions and informations on the study area; U. Bergmeier for support and assistance in the field; C. Adamidis, Ioannina, for statistical advice; P. Lampropoulos, Patras, and G. Amschlinger, Freiburg,

for skilfully preparing the figures; M. Klescewski and H. Gondard, both Montpellier, for their linguistic help with the résumé; E. Vidal and an anonymous reviewer for sug-gestions to improve the manuscript.

ReferencesBERGMEIER E., 1998. Flowering intensity of phrygana plants

after fencing. Israel. J. Plant Sci. 46: 41-46.

BERGMEIER E., DIMOPOULOS P., THEODOROPOULOS K. & ELEFTHERIADOU E., 2004. Zonale sommergrüne Laubwälder der südlichen Balkanhalbinsel. Tuexenia (in press).

BERGMEIER E. & MATTHÄS U., 1996. Quantitative studies of phenology and early effects of non-grazing in Cretan phrygana vegetation. J. Veg. Sci. 7: 229-236.

BEZKOROVAJNYJ P.J., GORDON A.M. & MCBRIDE A.A., 1993. The effect of cattle foot traffic on soil compaction in a silvo-pastoral system. Agroforestry Systems 21: 1-10.

CHRISTOPOULOS G. (ED.), 1978. Istoria tou Ellinikou Ethnous. [The history of the Greek Nation.] Vol. 7 (Vizantinos Ellinismos-Protovizantini chroni). Ekdoti Athinon, Athína.

DEBUSSCHE M., DEBUSCHEE G. & LEPART J., 2001. Changes in the vegetation of Quercus pubescens woodland after cessation of coppicing and grazing. J. Veg. Sci. 12: 81-92.

DI PASQUALE G. & GARFI G., 1998. Analyse comparée de l’évolution de la régénération de Quercus suber et Quercus pubescens après élimination du pâturage en forêt de Pisano (Sicile sud-orientale). Ecol. Medit. 24: 15-25.

FERNANDEZ ALÉS R., LAFFARGA J.M. & ORTEGA F., 1993. Strategies in Mediterranean grassland annuals in relation to stress and disturbance. J. Veg. Sci. 4: 313-322.

HELDREICH TH. VON, 1862. Die Nutzpflanzen Griechenlands. Athens.

IGME (ED.), 1983. Geological Map of Greece 1/500 000 (red. J. Bornovas & T. Rondogianni-Tsiambaou). 2nd ed. Athína.

NOY-MEIR I., GUTMAN M. & KAPLAN Y., 1989. Responses of Mediterranean grassland plants to grazing and protection. J. Ecol. 77: 290-310.

PANAGIOTIDIS N.TH., 1965. Makrochronii dhasike piramatike epithanie. [Langfristige forstliche Versuchsflächen.] Athine. 51 pp.

PAPANASTASIS V.P., FRAME J. & NASTIS A.S. (EDS.), 1999. Grasslands and woody plants in Europe. Proceedings Internat. Occ. Symp. European Grassland Federation, Thessaloniki, Greece, May 27-29, 1999. Thessaloniki.

PHILIPPSON A., 1892. Der Peloponnes. Versuch einer Landeskunde auf geologischer Grundlage. Berlin.

PHILIPPSON A., 1959. Die griechischen Landschaften 3 (2). Der Peloponnes, Teil 2: Der Westen und Süden der Halbinsel. Klostermann, Frankfurt am Main.

146



POTT R., 1999. Diversity of pasture-woodlands of north-western Germany. In: Kratochwil A. (ed.), Biodiversity in ecosystems. Kluwer Acad. Publ., Dordrecht: 107-132.

PRITCHETT W. & FISHER R., 1987. Properties and management of forest soils. Wiley, New York.

PRITZEL E., 1908. Vegetationsbilder aus dem mittleren und südlichen Griechenland. Bot. Jahrb. Syst. Pflanzengesch. Pflanzengeogr 41: 180-214.

REDECKER B., FINCK P., HÄRDTLE W., RIECKEN U. & SCHRÖDER E. (EDS.), 2002. Pasture landscapes and nature conservation. Springer, Berlin, Heidelberg.

ROTHMALER W., 1943. Die Waldverhältnisse im Peloponnes. Intersylva 3, 329-342.

SCHMIDT M. & HEILE H., 2001. Beweidung von Hutewäldern im Reinhardswald – Pro und Kontra. Jahrb. Naturschutz Hessen 6: 184-190.

SIBBALD A.R., 1999. Silvopastoral agroforestry: soil-plant-animal interactions in the establishment phase. In: Papanastasis V.P., Frame J. & Nastis A.S. (eds.), Grasslands and woody plants

in Europe, Proceedings Internat. Occ. Symp. European Grassland Federation, Thessaloniki, Greece, 27-29 May 1999: 133-144.

SPENCER J., 2002. Managing wood pasture landscapes in England: the New Forest and other more recent examples. In: Redecker B., Finck P., Härdtle W., Riecken U. & Schröder E. (eds.), Pasture landscapes and nature conservation. Springer, Berlin, Heidelberg: 123-136.

TER BRAAK C.J.F. & ŠMILAUER P., 1998. Canoco Reference Manual and User’s Guide to Canoco for Windows: Software for Canonical Community Ordination (version 4). Microcomputer Power, Ithaca, NY, USA.

TUTIN T.G. ET AL. (EDS.), 1968-1993. Flora Europaea, Vols. 2-5 and Vol. 1, 2nd ed. Cambridge University Press, Cambridge.

VERA F.W.M., 2000. Grazing ecology and forest history. Oxon Cabi Publ.

WILKE B., BOGENRIEDER A., WILMANNS O., 1993. Differenzierte Streuverteilung im Walde, ihre Ursachen und Folgen. Phytocoenologia 23: 129-155.

147


RésuméL’analyse palynologique des 28 derniers mètres du remblaiement holocène du delta de l’Argens, correspondant à l’intervalle 8 000-3 000 BP, permet une première restitution de l’histoire de la forêt de la Provence cristalline. Cette histoire est marquée, à l’origine, par le développement d’une forêt de chênes caducifoliés et de bruyères arborescentes. L’absence ou quasi-absence de taxons aujourd’hui majeurs comme le châtaignier et le chêne-liège est remarquable. Cette forêt commence à décliner à partir de 6 500 BP, sous l’effet d’un début d’anthropisation ou de causes naturelles (climat, niveau de la mer, évolution des sols et facteurs internes à l’écosystème fores-tier). Les modifications sont nettes vers 5 500 BP et deviennent plus radicales vers 3 000 BP avec une très large dominance des taxons héliophiles. Ces transformations sont incontestablement attribuables à l’intervention humaine.

Mots-clésVégétation, holocène, Provence cristalline, anthropisation néolithique

AbstractThe palynological analysis of the upper 28 meters of the Holocene infill of the Argens Delta coinciding with the 8000-3000 BP period, has allowed a preliminary reconstruction of the history of the vegetation of Siliceous Provence. Originally there was a mixed deciduous oak and arborescent healther forest. The absence or near-absence of taxa which are nowadays very common such as the chesnut or the cork-oak is to be noted. The forest cover decreased progressively from 6500 BP in response to natural (climate, sea-level, soils, and internal factors of the forest ecosystem) or anthropic causes. A first threshold of decline around 6000 BP and later a drastic modification of the forest cover around 3000 BP, may be certainly interpretated as man-made.

Key-wordsVegetation, holocene, siliceous Provence, neolithic anthropisation

Étude palynologique du carottage de Pont d’Argens(Roquebrune-sur-Argens, Var) : histoire holocène de la végétationen Provence cristalline ; facteurs naturels et anthropiquesPollen analysis of the core of Pont d’Argens(Roquebrune-sur-Argens, Var): Holocene vegetation historyin the siliceous Provence; natural and anthropic factors

Dubar Michel1, Bui-Thi-Maï1, Nicol-Pichard Sylvie2 & Thinon Michel3

1. CEPAM (UMR-CNRS 6130), bât. 1, 250 rue Albert-Einstein, 06560 Valbonne. [email protected]. Museum d’Histoire naturelle de Marseille, palais Longchamp, 13004 Marseille3. IMEP (UMR-CNRS 6116), université Paul-Cézanne / Aix-Marseille III, faculté des sciences et techniques, case 462, 13397 Marseille cedex 20

148

◆ M. DUBAR ET AL.


Abridged versionThe questions which we attempt to answer focus on the state

of the vegetation in the Cristalline Provence before Man’s impact and the history of some present major taxa such as Quercus suber L., Erica and Castanea sativa L.

A palynological study was thus carried out on a core bored in of the Holocene Delta of the Argens River. Originating in the calcareous Western Provence, the Argens River drains in its lower watercourse large areas of the Maures and Esterel cristalline blocks (fig. 1).

The Argens Delta (fig. 2) was built by continuous accretion of fine-grained sediments during the Holocene sea-level uprise (Dubar & Anthony, 1995; Dubar, 2003). The core of Pont d’Argens (fig. 3) cuts through the upper 28 meters of these sediments which date from the 8000-3000 yr BP interval (Fiches et al., 1995).

Palynological analysis has enabled us to produce a diagram with 52 superposed spectra (fig. 4). It comprises 33 arboreal taxa, 64 herbaceous taxa and 21 taxa of ferns and mosses.

The dominant taxa which are continously represented throu-ghout are the arborescent heather (30-50 %), deciduous oak (5-15 %) and pines (4-30 %). Quercus suber is very slightly represented (inf. 1 %), and Castanea is absent.

The diagram shows three zones corresponding to an evolution in three phases (fig. 4).

The first (from the basis to 21,5 m) indicates a stable compo-sition of the vegetation, particularly for the deciduous trees (oak, lime).

The second phasis (from 21,5 m to the hiatus) starts with a con-tinuous curve of Alnus and with a peak of Corylus. Some heliophil-ous plants (herbaceae, shrubs and pines) increased; correlatively the deciduous trees decreased significantely from about -19 m.

The third phasis (from the hiatus to the top) marked a strong change: pines, Erica and other heliophilous plants are at the height of representation when the deciduous oak drop at the minima.

We note the curve of Alnus which certainly corresponds with a temporary expansion of the riparian forest.

The flat but clear Abies curve also is observed on the other dia-grams of Provence (Beaulieu, 1977, Pichard, 1987, Nicol-Pichard & Dubar, 1998, Andrieu-Ponel et al., 2000). Besides, making use of these diagrams we can reconstitute the vegetation and its history in Crystalline Provence between 8000 and 3000 BP.

The original vegetation before anthropisation combined decidu-ous oak and arborescent healther in a mixed forest. This was the precise time of the “Climatic Optimum” of the Holocene which resulted from forest reconquest at the end of the glaciation. Of course after the optimum, the forest cover declined slowly however without its mesophilous characteristics was really modified.

In Eastern Provence, further East than the Tanneron Block, humidified and bounded by Alpine influences, the mesophilous state lasted until 5000 BP, whereas in Western Provence, drier due to the Mistral wind, the first signs of the thinning of the forest occurred earlier around 7500 BP.

The slow decline of the deciduous forest after the Atlantic Optimum might have been be caused by independant factors, notably:

— Morphodynamic modifications such as the uprise in sea-level (closed to 1 cm/y) which primaraily determined the construction of the alluvial and coastal plain, and finally the edification of sand bars near the mouths and along the coasts. As a conse-quence of the modifications of landscape, environment and soils, the riparian forest and later the pine forest, became widespread (Dubar, 2001).

— A global climatic tendency towards a cooling of the climate, linked to the solar radiation ratio (Berger, 1992) occurred after the Atlantic Optimum.

— The evolution of internal factors of the forest caused a slight imbalance in the geo-ecosystem (Heinrich & Hergt, 1993).

In this context of slow forest decline, Neolithic man probably took the opportunity to extend agro-pastoralism, precociously in Western Provence, as indicated by the archaeological data, then later in Eastern Provence. At Pont d’Argens, as in the whole of Crystalline Provence, this scenario appears to have taken place between the two at a date close to 6000 BP. The impact on the forest cover will be however irreversible only much later, at the Iron Age (towards 3000 BP), everywhere in Provence (Triat-Laval, 1979) and in particular in Pont d’Argens.

149

ETUDE PALYNOLOGIQUE DU CAROTTAGE DE PONT D’ARGENS (PROVENCE CRISTALLINE, VAR) ◆


INTRODUCTION

La végétation forestière actuelle de la Provence cristal-line est essentiellement constituée par des groupements à chêne-liège (Quercus suber L.), pin maritime (Pinus pinaster Aiton subsp. pinaster), pin pignon (Pinus pinea L.) et châ-taignier (Castanea sativa Miller). Le chêne vert (Quercus ilex L.), le chêne pubescent (Quercus pubescens Willd.) et le pin d’Alep (Pinus halepensis Miller) sont nettement moins fréquents. Ces taxons arborescents sont généralement associés à des formations ligneuses plus ou moins basses constituant le maquis et assez régulièrement parcourues par des incendies. Parmi les espèces les plus fréquentes du maquis, on peut citer la bruyère arborescente (Erica arborea L.), l’arbousier (Arbutus unedo L.), le cytise velu (Cytisus villosus Pourret), le ciste de Montpellier (Cistus monspeliensis L.), le ciste à feuilles de sauge (Cistus sal-vifolius L.) et la lavande stéchas (Lavandula stoechas L.). Cependant, comme l’avait noté Molinier (1954, 1973), le châtaignier paraît d’introduction récente et l’expan-sion du chêne-liège serait en grande partie liée à l’action de l’homme. Ces points ont été contestés par certains

auteurs, notamment Loisel (1976) qui envisageait un indigénat du châtaignier et considérait que les forêts de chêne-liège constituaient une large part des groupements sylvatiques potentiels de la basse Provence siliceuse.

L’état originel de la végétation de cette Provence sili-ceuse est donc encore mal connu. C’est pour cette raison que l’étude pollinique d’une « archive sédimentaire » ayant conservé les pollens d’une période antérieure au processus d’anthropisation nous a semblé intéressante à réaliser. Le carottage de Pont d’Argens est, en fait, le premier enregistrement sédimentaire holocène provenant du cœur même de la Provence cristalline entre les Maures et l’Esterel (fig. 1).

La carotte de Pont d’Argenset son contexte géomorphologique

L’Argens est un petit fleuve, long de moins de 100 km qui prend sa source près de Saint-Maximin, sur le revers septentrional de la chaîne de la Sainte-Baume, en Provence occidentale (massif de Mourre d’Agnis). Il rejoint la « dépression permienne » un peu avant Vidauban

Fig. 1. La Provence cristalline : socle magmatique et cristallophyllien, roches volcaniques des massifs des Maures, de l’Esterel et du Tanneron et leur auréole sédimentaire silicatée.

Fig. 1. Siliceous Provence: basal complex of magmatic, metamorphic and volcanic rocks ot the Maures, Esterel and Tanneron blocks and their belt of silicated sediments.

150

◆ M. DUBAR ET AL.


après avoir traversé une région de plateaux calcaires. Il ne quitte plus alors la zone des terrains silicatés, d’abord les pélites du Permien, puis la partie orientale du massif cristallin des Maures. Son cours est donc situé, pour parts sensiblement égales, en zone calcaire et en zone siliceuse. Le delta lui-même a plus de 12 km de longueur et atteint 5,5 km dans sa plus grande largeur.

Réalisé dans la partie amont du delta (fig. 2), le son-dage a atteint, à 28 m de profondeur, le substrat rocheux permien, après avoir traversé l’intégralité du remblaie-ment holocène présent en ce point. Ce remblaiement s’est constitué au cours de la remontée postglaciaire du niveau de la mer entre 12 000 ans BP, époque à laquelle la mer était vers -100 m, et 3 000 ans BP, date à laquelle ce niveau a atteint pratiquement le zéro actuel (Dubar & Anthony, 1995 ; Dubar, 2003). Les basses vallées sont alors ennoyées et transformées en rias (fig. 2). Piégés

dans la ria, les sédiments apportés par l’Argens ont ainsi provoqué son colmatage. Le colmatage est complet lors-que le niveau de la mer s’est approché du zéro actuel et l’émersion a eu lieu peu après.

Les sédiments déposés présentent toujours une granu-lométrie fine et sont bien classés : ce sont des vases, des limons et des sables fins souvent chargés en débris orga-niques (fig. 3, A). Au sommet, dans les derniers mètres, ces dépôts qui deviennent plus grossiers, graveleux et rougeâtres, résultent de phénomènes de ruissellement ou d’alluvionnement postérieurs à l’émersion.

Trois dates 14C ont été obtenues sur des lits tourbeux, dans le tiers inférieur de la carotte. Nous les utilisons en âge BP non calibré. La date de 7440 +/-90 BP (Ly 5868) obtenue à 26,60 m permet de situer la base du remblaie-ment vers 7 900 BP. Les dates intermédiaires sont de 7 080 +/-70 BP (Ly 5867) à 23,50 m, de 6 000+/-60

Fig. 2. Le delta de l’Argens et la position des deux carottages de Pont d’Argens et du Verteil.

Fig. 2. The Argens delta: location of the two cores of Pont d’Argens and Verteil.

151



BP (Ly 5866) à 19,25 m et de 5790 +/-60 (Ly 5865) à 18,60 m. La partie supérieure n’a pas été datée directe-ment, cependant on peut évaluer son âge en se référant à un autre carottage réalisé plus en aval, au Verteil (fig. 1). Dans cette carotte, un petit lit tourbeux situé à -7 m (fig. 3, B) a été daté de 3 050 BP (Ly 5889). Sans qu’on puisse établir une parfaite correspondance altimétrique entre les deux carottages, il semble toutefois que le som-met du remblaiement deltaïque de Pont d’Argens puisse être situé vers 3 000 BP.

Compte tenu de la genèse du remblaiement et de son caractère d’accrétion continue (Fiches et al., 1995), nous pouvons admettre que la carotte de Pont d’Argens couvre

Fig. 3. Les carottages de Pont d’Argens (A) et du Verteil (B) et leur étalonnage C 14 (dates BP non calibrées).

Fig. 3. Pont d’Argens and Verteil cores with 14 C datings (non calibrated BP).

l’intervalle 7 900-3 000 BP. Il manque malheureusement 5 m de sédiments, car la récupération des dépôts a été mauvaise entre -10 et -15 m.

L’analyse palynologique

L’analyse a porté sur l’ensemble de la carotte, à l’ex-ception de la tranche 10 à 15 m. Ce hiatus ne semble toutefois pas avoir gravement affecté la continuité du diagramme pollinique.

Normalement les analyses ont été réalisées avec un pas de 20 cm, mais parfois, afin de suivre les variations verticales de la lithologie, cet espacement a été modifié, ce qui explique une certaine irrégularité des prélèvements, en particulier dans les dix derniers mètres. Cette séquence supérieure est pauvre en pollens et les grains sont mal conservés. En revanche, dans la série sédimentaire infé-rieure, qui se développe en-dessous de 15 m, les spectres sont très riches en pollens et spores. Le diagramme com-prend 52 spectres (fig. 4) qui ont permis d’identifier 33 taxons arborescents, 64 taxons herbacés, suffrutescents et cryptogamiques. Les pourcentages de ces derniers taxons ont été calculés séparément de façon à ne pas amoindrir la représentation globale du couvert forestier. La première colonne située à gauche du diagramme, montre les cour-bes respectives des pollens d’arbres (AP) et d’herbacées (NAP).

Les familles des herbacées non représentées gra-phiquement (tableau 1, p. 22) sont les Orchidaceae, Papaveraceae, Rutaceae, Saxifragaceae et Valerianaceae. Les genres et espèces non représentés sont : Aphyllanthes monspeliensis L., Crocus sp., Potentilla sp., Thalictrum sp. ainsi que Alisma sp., Hippuris vulgaris L., Lythrum sp., qui sont des plantes d’eau douce et Ruppia sp. qui indique la présence d’eau saumâtre.

Les trois essences dominantes et représentées de manière continue sont des Ericaceae (Erica arborea, 30 à 40 %), des chênes caducifoliés (5 à 15 %) et des pins (4 à 30 %). On remarque, en revanche, l’extrême discré-tion (moins de 1 %) de Quercus suber qui est aujourd’hui caractéristique des massifs siliceux environnants. Erica arborea, qui domine constamment, a une fréquence rela-tivement homogène de la base au sommet (le hiatus de 5 m entre les cotes 15 et 10 m ne produit pas de distorsion importante de sa représentation).

Nous avons subdivisé le diagramme en trois zones correspondant à une évolution triphasée (fig. 4) :

— La première s’étend de la base jusqu’aux environs du niveau 21,5 m. Les différentes représentations sont

152

◆ M. DUBAR ET AL.


relativement stables. Les chênes caducifoliés se main-tiennent à des valeurs élevées, immédiatement après la bruyère arborescente. On peut noter la présence pratiquement constante du tilleul (Tilia), espèce carac-téristique des milieux forestiers caducifoliés évolués.

— La seconde zone lui succède jusqu’au hiatus. Elle est assez hétérogène mais nous définissons sa base par la concomitance de l’apparition d’une courbe continue de l’aulne, d’un pic remarquable de Corylus et de l’accroissement des Cichorioideae. Cette zone peut être subdivisée elle-même en deux sous-zones 2a et 2b : la première de 21,5 m jusqu’à 17,5 m, la seconde de 17,5 m jusqu’au hiatus. Ces deux sous-zones se différencient au niveau du fonds arboréen : la pre-mière voit la persistance d’une bonne représentation de taxons arborescents forestiers comme les chênes à feuillage caduc et le sapin (Abies), la seconde partie est caractérisée par la nette diminution de ces arbres, tandis que les héliophiles comme les Cichorioideae, les Chenopodiaceae et Pinus prennent de l’importance. Le rapport AP/NAP diminue corrélativement.

— La troisième couvre la partie supérieure du diagramme. Les taux de Pinus, d’Erica arborea et des héliophiles sont à leur apogée, tandis que ceux des chênes sont à leur minimum.

Bien développée entre 21 et 17,40 m avec un pic de 40 % à 17,80 m, la courbe de l’aulne (Alnus) est certaine-ment représentative de l’essor momentané de la ripisylve.

La discrète courbe du sapin (Abies) entre 21 et 18 m est bien conforme à ce qui est connu par ailleurs en Provence pour cette période (Beaulieu, 1977 ; Triat-Laval, 1978, Pichard, 1987 ; Nicol-Pichard & Dubar, 1998 ; Andrieu-Ponel et al., 2000) et dénote sans doute l’existence de sapinières dans des massifs voisins du bas-sin de l’Argens. En effet, la présence de grains de pollens de sapin au faible pouvoir dispersif (Triat-Laval, 1971) implique l’existence régionale d’arbres producteurs.

Les herbacées ne présentent pas de variations significa-tives, on note cependant une légère progression des NAP à partir de la cote -17,50 m. Dès cet instant, la courbe de l’aulne s’arrête et celle des fougères progresse consi-dérablement. Ce changement voit aussi l’accroissement de la courbe des pins et la diminution de celle du chêne pubescent. Ces tendances se confirment après le hiatus (cote -10 m) et semblent donc bien valider les résultats de cette deuxième partie du diagramme. Les variations qui y sont observées, en particulier la progression des pins et celle de Erica arborea, s’inscrivent bien dans une

tendance significative d’une modification de la végétation. De même, la progression de Cistus et des chicorées peut être considérée comme étant liée à l’ouverture du milieu et à l’érosion probable des sols.

La restitution de la végétation de la zone cristalline à partir de l’étude palynologique de la carotte de Pont d’Ar-gens pose le problème de l’apport probable de pollens des zones lointaines du bassin versant situées en Provence calcaire. Il semble cependant que ces apports soient rela-tivement peu importants. En effet, la forte dominance, en nombre de grains, d’Erica arborea, qui est une espèce calcifuge réputée pour ne diffuser ses pollens qu’à très faible distance, tend à montrer que la composition polli-nique des spectres est principalement d’origine locale. Ce cortège paraît donc relativement bien représentatif de la flore de la zone cristalline proche du carottage.

DISCUSSION

Couvrant une durée de près de 5 millénaires, le diagramme de Pont d’Argens montre l’évolution de la végétation holocène régionale. Il peut être comparé aux autres diagrammes obtenus en Provence comme ceux de Tourves (Nicol-Pichard, 1987) et de Biot (Nicol-Pichard & Dubar, 1998). Le premier est situé en Provence calcaire à l’ouest de Pont d’Argens, tandis que le second, localisé en région niçoise également calcaire, est plus à l’est. Ces deux diagrammes montrent le développement considé-rable de la forêt mésophile, tout particulièrement de la chênaie caducifoliée dès l’Holocène ancien. Ce type de végétation paraît d’ailleurs constituer un état d’équilibre durable dans toute la Provence, Provence occidentale et rhodanienne comprises (Triat-Laval, 1979) ainsi que dans le Sud des Alpes (Beaulieu, 1977) : la reconquête de la forêt depuis la fin du glaciaire est rapide et continue et conduit à l’optimum atlantique entre 8 000 et 7 500 BP. En Provence cristalline, sur les roches mères compactes (granites, rhyolithes, gneiss), assez peu altérables sous climat méditerranéen, les facteurs édaphiques sont forte-ment exprimés, certainement en raison de la chimie des sols, mais aussi vraisemblablement dans le faible dévelop-pement des profils. De ce fait, les espèces de la chênaie caducifoliée sont localement peu favorisées par rapport à d’autres taxons qui, comme Erica arborea, sont moins exigeants au point de vue édaphique. Dans ce complexe du chêne caducifolié et de la bruyère arborescente, il est évidemment difficile de savoir si la répartition végétale se faisait selon une mosaïque de zones (par exemple des

153



Fig. 4. Diagramme pollinique de Pont d’Argens (les taxons trop rares n’ont pas été portés sur le diagramme ; voir texte).

Fig. 4. Pollinic diagram of Pont d’Argens (taxa very rare were not related to the diagram ; see text).

154

◆ M. DUBAR ET AL.


taxons altitude Croc Aphymons MAL ORC PAP PRI Thal Filip RUT SAX Viola Lythr Alism Hipp Rupp VAL

542 1,50

769 0,80

860 1,70

901 0,80 0,80 0,80

925 0,80 0,80 0,80

980 1,00 1,00 1,00

1610

1650 0,20

1705 0,50

1720 0,40

1740 0,20 0,40

1760

1780 0,50 1,20 0,90

1820 0,30 0,60

1900 0,20

1923 0,30

1960 0,50

2020 0,20

2060 0,30 0,30

2080 0,20 0,20

2170 0,30 0,30

2310 0,30

2330 0,70

2390 0,30

2410 0,30

2470 0,30

2490 0,20 0,20

2530 0,30

2550 0,30

2650 0,30

2670 0,20 0,20

2690 0,40 0,40

2707

altitude taxons Croc Aphymons MAL ORC PAP PRI Thal Filip RUT SAX Viola Lythr Alism Hipp Rupp VAL

Croc = Crocus ; Aphymons = Aphyllanthes monspeliensis ; MAL = MALVACEAE ; ORC = ORCHIDACEAE ; PAP = PAPAVERACEAE ; PRI = PRIMULACEAE ; Thal = Thalictrum ; Filip = Filipendula ; RUT = RUTACEAE ; SAX = SAXIFRAGACEAE ; Lythr = Lythrum ; Alism = Alisma ; Hipp = Hippuris ; Rupp = Ruppia ; VAL = VALERIANACEAE

Tableau 1. Contenu du varia du diagramme pollinique (figure 4).

Table 1. Detailed content of the varia in the polllinic diagram (figure 4).

155



groupements à bruyère dominante sur les adrets et les sols superficiels et des secteurs plus forestiers à chênes pubescents sur les ubacs et les sols profonds) ou bien si une véritable association chêne-bruyère occupait l’ensem-ble des massifs. Un faciès de ce type existe aujourd’hui dans certaines zones protégées, non perturbées par les incendies depuis plus de cinquante ans, comme la haute vallée du Reyran dans le massif de l’Esterel. Mais toutes les observations écologiques montrent que la bruyère régresse lorsqu’elle est dominée par des ligneux de taille supérieure. La coexistence de ces deux taxons avait déjà été observée par Reille (1984), à basse altitude en Corse, avec des taux élevés de pollen d’Erica arborea dès le début de l’Atlantique, ce qui avait conduit cet auteur à s’interroger sur la place de la bruyère arborescente dans les écosystèmes naturels sur terrains siliceux. On peut remarquer que, en Corse, des taux relativement élevés de cette espèce ont été relevés à des altitudes approchant 1 800 m (Reille, 1975), ce qui conduit à penser que son pollen est relativement diffusable par le vent. Quoi qu’il en soit, cette végétation à l’architecture mal connue, cons-tituée globalement de chênes caducifoliés et de bruyères arborescentes semble représenter un certain état de sta-bilité (climax), qui peut être dit « originel », c’est-à-dire antérieur à l’anthropisation.

Sur le diagramme de Pont d’Argens, comme d’ailleurs sur les autres séries polliniques de Provence orientale, l’intervention de l’homme néolithique semble plus faible et plus tardive qu’en Provence occidentale, peut-être discrètement vers 6 500 BP, au début de la phase 2, où le pic de Corylus peut être, lui aussi, interprété comme une ouverture ménagée du milieu qui profite momen-tanément à cette héliophile mésophile. L’ouverture du paysage est nettement plus sensible dans la seconde partie de cette phase où le pic des Chenopodiaceae peut représenter l’extension des rudérales ou bien l’installation d’une végétation lagunaire halophile. Cependant, le pic des Cichorioideae et la régression des taxons forestiers, notamment avec la disparition de Tilia, fait pencher pour la première hypothèse. L’anthropisation s’accentue forte-ment dans la dernière phase, qui semble correspondre au Bronze final et à l’âge du Fer, d’après l’âge de 3 000 BP obtenu sur le carottage du Verteil. Les pollens de pins atteignent des taux inégalés, ainsi que la bruyère dans le niveau supérieur, alors que les cistes, comme tous les taxons héliophiles (Artemisia, Helianthemum, Plantago, Apiaceae, Cichorioideae, Chenopodiaceae, etc.) sont en forte augmentation ; Calluna, Asphodelus et Juniperus sont des indicateurs de pastoralisme.

La lente régression de la forêt après l’optimum atlanti-

que pourrait être aussi le résultat de changements morpho-dynamiques locaux. En effet, tout particulièrement dans les basses vallées ou à proximité du rivage, il a été montré que sous l’effet de la remontée du niveau marin (Dubar, 2001), vers la fin de la transgression holocène, à partir de 6 000 BP, la formation des plaines alluviales côtières et des basses vallées conduit à l’extension des ripisylves puis à celle des pinèdes au détriment de la chênaie. L’extension tardive des pins sur la frange côtière et près des estuaires est essentiellement d’ordre édaphique et liée à la mise en place, alors que le niveau marin s’est presque stabilisé, des cordons sableux et graveleux (Dubar & Anthony, 1995). Ce phénomène est très perceptible sur le diagramme de Biot (Nicol-Pichard & Dubar, 1998), mais il l’est moins sur celui du Pont d’Argens.

Des variations climatiques au cours de l’Holocène, après l’optimum atlantique, ont également été invoquées pour expliquer les changements de végétation. Depuis la fin du glaciaire, la tendance linéaire a été au réchauffe-ment jusqu’à l’optimum climatique atlantique. Un lent rafraîchissement aurait suivi, provoqué par des change-ments de l’insolation terrestre (Berger, 1992). Certains auteurs font intervenir des « crises » climatiques. Ainsi, la détérioration du Subboréal correspondrait à l’accomplis-sement d’une série modulée de phases sèches à caractère « méditerranéen » accusé (Jalut et al., 2000). Cependant l’origine de ces crises climatiques et même leur réalité physique restent hypothétiques.

On a également fait appel à la dynamique interne des écosystèmes, sans qu’il n’y ait aucune intervention externe. Après la reconquête rapide post-glaciaire et l’acmé de l’optimum atlantique (Kremer & Petit, 2001), la chênaie caducifoliée décline lentement comme cela est fréquent dans la dynamique des populations. Les causes peuvent être multiples, mais il semble que pour les communautés végétales à extension très rapide, des rétroactions positives, cumulatives et durables pour-raient s’établir, menant à un certain déséquilibre du géo-écosystème (Heinrich & Hergt, 1993).

Il n’est pas possible de faire la part respective de ces divers éléments. Seul le constat d’une évolution forestière est certain et cette évolution se fait, à l’Atlantique, dans le sens d’une diminution sensible du couvert mésophile, le changement étant diachrone :

— En Provence occidentale et rhodanienne, au climat plus sec et soumis à l’action du mistral, le changement enregistré par les archives polliniques est plus précoce. On observe notamment les premières manifestations et fluctuations des chênaies sclérophylles à partir de 7 500 BP, date qui coïncide avec l’installation des pre-

156

◆ M. DUBAR ET AL.


mières communautés néolithiques du Cardial ancien (Triat-Laval, 1978).

— En Provence la plus orientale (au-delà du Tanneron), plus humide car subissant l’influence du relief des Alpes, l’équilibre mésophile de la forêt de la période atlantique paraît se maintenir pratiquement jusque vers 5 000 BP (Dubar et al., 1986).

On peut émettre l’hypothèse que la Provence orientale bien que occupée sur son littoral dès 7 000 BP (Binder & Maggi, 2001) a été plus faiblement et plus tardivement mise en exploitation que la Provence occidentale (Dubar & Roscian, 2001). En retour, l’impact sur le couvert végétal est aggravé également de manière retardée dans le temps. Il ne faut cependant pas oublier que les diffé-rences climatiques peuvent aussi jouer sur l’expression pollinique du degré d’anthropisation. À Pont d’Argens, la situation paraît être intermédiaire avec l’enregistrement d’une première régression de la chênaie caducifoliée vers 6 500 BP suivie d’une accentuation à partir de 5 500 BP. L’impact reste cependant modéré et la dégradation ne deviendra irréversible que beaucoup plus tard, vers 3 000 BP.

On doit remarquer que ces données polliniques mon-trent la grande rareté du chêne-liège (Quercus suber) et l’absence du châtaignier (Castanea sativa), tous deux forts pollinisateurs, qui sont aujourd’hui les arbres caractéristi-ques des massifs des Maures et de l’Esterel. Par contre, la discrétion de l’arbousier (Arbutus) dans le diagramme est peut être à mettre au compte de sa faible dispersion polli-nique. Comme en zones calcaires, la chênaie caducifoliée constituait l’essentiel de la forêt originelle et, de la même façon que pour ces régions, les données phytohistoriques s’inscrivent en faux vis-à-vis des spéculations phytoso-ciologiques. Ces dernières accordaient, pour les massifs siliceux, une place prépondérante soit au chêne vert (Molinier, 1973), soit au chêne-liège (Lavagne & Moutte, 1974 ; Loisel, 1976). L’absence du châtaignier rend égale-ment peu probable un indigénat suggéré par l’individuali-sation actuelle d’une association végétale particulière dans les massifs des Maures et de l’Esterel (Loisel, 1976).

CONCLUSION

Le diagramme de Pont d’Argens offre une première approche phytohistorique de la Provence cristalline anté-rieure à l’intervention de l’homme néolithique. Il nous renseigne sur la composition de la végétation depuis

environ 7 500 BP, la dominance (attendue) de certains taxons comme la bruyère arborescente, le rôle impor-tant de la chênaie caducifoliée et l’absence de taxons, aujourd’hui prépondérants, comme le chêne-liège ou le châtaignier. Sur le plan de l’évolution de la couverture végétale de la Provence cristalline au cours de l’Holocène, ce diagramme est conforme à ce qui est déjà connu en Provence calcaire. La tendance forestière mésophile est exprimée précocement (c’est la suite normale de l’évo-lution climatique postglaciaire), modulée par les facteurs édaphiques locaux. Des signes de modifications d’origine anthropique ou à déterminisme naturel apparaissent à partir de 6 500 BP, avec un décalage d’un millier d’an-nées par rapport aux données obtenues en Provence occidentale, et donc, chronologiquement intermédiaires avec ceux observés en zone calcaire orientale. À partir de 5 500 BP, les manifestations d’une exploitation par l’homme producteur deviennent évidentes pour aboutir à une anthropisation généralisée vers 3 000 BP.

REMERCIEMENTS

L’interprétation palynologique et la rédaction de l’article ont bénéficié des nombreux conseils de Michel Girard. Qu’il en soit vivement remercié ici. Nous remer-cions également Claudine Dauphin pour la traduction en anglais de la version abrégée ainsi que les deux rap-porteurs anonymes qui ont contribué à éclairer certains points importants de la discussion.

BibliographieANDRIEU-PONEL V., PONEL P., JULL A.J.T, BEAULIEU J.-L.,

BRUNETON H. & LEVEAU P., 2000. Towards the reconstruc-tion of the Holocene vegetation history of Lower Provence: two new pollen profiles from Marais des Baux. Veget. Hist. Archeobot., 9: 71-84.

BEAULIEU J.-L. DE, 1977. Contribution pollenalytique à l’histoire tardi- et post-glaciaire de la végétation des Alpes méridiona-les françaises. Thèse ès-sciences Aix-Marseille III, 358 p.

BERGER A., 1992. Le climat de la Terre. Un passé pour quel avenir ? Editions De Boeck-Wesmael. 479 p.

BINDER D. & MAGGI R., 2001. Le Néolithique ancien de l’arc liguro-provençal. Bull. Soc. Préhist. Fr., 98 : 411-422.

DUBAR M. & ROSCIAN S., 2000. Scénario climatique et dévelop-

157



pement de l’agro-pastoralisme néolithique en Provence et en Ligurie. Bull. Soc. Préhist. Fr., 98 : 391-398.

DUBAR M., 2001. Évolution de la végétation littorale en relation avec les changements physiographiques et édaphiques à la fin de la transgression holocène en Provence orientale (France). Actes du XVIIe Congrès de l’APLF, Arles, sept. 2001, 38.

DUBAR M., 2003. The Holocene deltas of Eastern Provence and the French Riviera : geomorphological inheritance, genesis and vulnerability. Géomorphologie: relief, processus, environne-ment, 2003, 4: 263-270.

DUBAR M. & ANTHONY E.J., 1995. Holocene Environmental Change and River-Mouth Sedimentation in the Baie des Anges, French Riviera. Quatern. Res. 43, 3: 329-343.

DUBAR M., DAMBLON F., NICOL-PICHARD S., VERNET J.-L., CHAIX L., IRR F. & BABINOT J.-F., 1986. L’environnement côtier des Alpes-Maritimes à la fin de la transgression versi-lienne d’après l’étude biostratigraphique du site de l’Etoile à Nice (France). Revue de Paléobiologie, 5 : 289-310.

FICHES J.-L, BÉRATO J., BRENTCHALOV D., CHOUQUER G., DUBAR M., GAZEENBEEK M., LATOUR J. & ROGERS B., 1995. Habitats de l’Age du fer et structures agraires d’épo-que romaine aux Escaravatiers (Puget-sur-Argens, France). Gallia, 52 : 205-262.

HEINRICH D. & HERGT M., 1993. Atlas de l’Ecologie. Encyclopédie d’aujourd’hui. Le Livre de poche, Paris, 286 p.

JALUT G., AMAT A.E., BONNET L., GAUQUELIN T. & FONTUGNE M., 2000. Holocene climatic changes in the Western Mediterranea, from South-east Spain. Palaeogeogr., Palaeoclim., Palaeoecol., 160 : 255-290.

KREMER A. & PETIT R., 2001. L’épopée des chênes européens. La Recherche, 342 : 40-43.

LAVAGNE A. & MOUTTE P., 1974. Feuille de Saint Tropez au 1/100 000. Bull. Carte Végét. de la Provence et des Alpes du Sud (I) : 3-43.

LOISEL R., 1976. La végétation de l’étage méditerranéen dans le sud-est continental français. Thèse ès-sciences Aix-Marseille III, 384 p.

MOLINIER R., 1954. Les climax côtiers de la Méditerranée occi-dentale, Vegetatio, 4 : 284-308.

MOLINIER R., 1973. Les études phytosociologiques en Provence cristalline, Bull. Hist. Nat. Marseille, 33 : 7-46.

NICOL-PICHARD S. & DUBAR M., 1998. Reconstruction of Late-Glacial and Holocene environments in Southeast France based on the study of a 66 m long core from Biot, Alpes-Maritimes. Veget. Hist. and Archeobot., 7 :11-15.

REILLE M., 1975. Contribution pollenanalytique à l’histoire tar-diglaciaire et holocène de la végétation de la montagne corse. Thèse ès-sciences Aix-Marseille III, 206 p.

REILLE M., 1984. Origine de la végétation de la Corse sud-orientale ; analyse pollinique de cinq marais côtiers. Pollen et spores, 26 (1) : 43-60.

TRIAT-LAVAL H., 1971. Contribution à l’étude de la dissémina-tion pollinique du pin, du hêtre, du sapin et de l’épicéa. Ann. Univ. de Provence, Sciences, 46, p. 155-160.

TRIAT-LAVAL H., 1978. Contribution pollenalytique à l’histoire tardi- et post-glaciaire de la végétation de la basse vallée du Rhône. Thèse ès-sciences Aix-Marseille III, 343 p.

TRIAT-LAVAL H., 1979. Histoire de la forêt provençale depuis 15 000 ans d’après l’analyse pollinique. Forêt médit., 1 : 19-24.

159


AbstractYew, Taxus baccata L. is a declining species that occupies a limited range throughout the Mediterranean basin. The community struc-ture and spatial distribution of yew trees were investigated in Greece by means of a questionnaire survey administered to Forest District Offices nationwide. The results of the survey show that the species population is confined mainly to mountainous areas extending from south Peloponnese to Evros prefecture. Findings suggest yew as a rare and potentially endangered species that occurs at different degrees of population fragmentation ranging from individual trees to more rarely clumps of trees. Yew distribution is formed by isolated populations, mostly in mountain ravines at an altitudinal range between 500 and 1 500 metres; it increasingly becomes an isolated storey tree within fir pure forests (Abies cephalonica) and various kinds of mixed woodlands of beech and oak, as well as mixed forests of fir and beech. Populations are small, most in shrubby groups of 5 to 50 individuals with height ranging between 3 and 6 metres and low proportion of saplings and seedlings. The rarity of yew populations and biochemical interest implicate the adoption of appropriate management actions to facilitate broader expansion, aiming at restoring existing Taxus woodland areas and increasing the share of Taxus forests under statutory protection within the Natura 2000 network of protected areas.

Key-wordsYew tree, Taxus, yew forests, questionnaire survey, Greece

RésuméL’if, Taxus baccata L., est une espèce en déclin qui occupe un espace limité dans le bassin méditerranéen. La structure des peuplements et la répartition spatiale des ifs ont été étudiées en Grèce au moyen d’une enquête sous forme de questionnaire adressé aux bureaux forestiers nationaux. Les résultats de l’enquête montrent que l’espèce est confinée principalement dans les zones montagneuses s’étendant du Péloponnèse sud à la préfecture d’Evros en Thrace. Les résultats suggèrent de considérer l’if comme une espèce rare et potentielle-ment en danger dont les populations se situent à différents degrés de fragmentation, allant de spécimens isolés à, plus rarement, des petits boisements. La répartition de l’if en Grèce est donc constituée de populations isolées, principalement dans les ravins de montagne, à une altitude comprise entre 500 et 1 500 m ; il se rencontre de plus en plus comme une espèce de sous-étage, isolée dans les forêts de sapin (Abies cephalonica) et dans différents types de forêts mélangées de hêtre et de chêne, ainsi que dans des forêts mélangées de sapin et de hêtre. Les populations sont petites, la plupart en groupes de 5 à 50 individus d’une hauteur de 3 à 6 m avec une faible proportion d’arbres et de jeunes plants. La grande rareté des populations d’if et l’intérêt biochimique impliquent l’adoption d’actions appropriées de gestion visant à faciliter une expansion de l’espèce, afin de reconstituer les zones sylvestres existantes à Taxus et d’augmenter la part de forêts de Taxus sous la protection statutaire du réseau Natura 2000.

Mots-clésIf, Taxus, forêts d’if, enquête, Grèce

Distribution and stand structure of Taxus baccata populationsin Greece; Results of the first national inventoryDistribution et structure des peuplements de Taxus baccataen Grèce ; résultats du premier inventaire national

K. Kassioumis1, K. Papageorgiou1, T. Glezakos2 & I.N. Vogiatzakis3

1. Correspondence author: NAGREF – Agricultural Research Station of Ioannina (ARSI), E. Antistasis 1, Katsikas, 45500, Ioannina, Greece. E-mail: [email protected]. NAGREF – Information and Documentation Section, Egialias 19 & Chalepa, 15125 Athens, Greece3. Landscape and Landform Research Group Department of Geography, University of Reading Whiteknights RG6 6AB, Reading Berks, UK

160

◆ K. KASSIOUMIS, K. PAPAGEORGIOU, T. GLEZAKOS & I. N. VOGIATZAKIS


INTRODUCTION

The contribution of the Holarctic or Eurasiatic element in the Mediterranean flora and its importance to the post glacial flora of Southern Europe and North Africa has often been highlighted (see review in Quézel, 1985). Yew, Taxus baccata L. is one of the species belonging to this element of pre-Miocene origin that now occupies a limited range not only in the Mediterranean basin but all over Europe (Ellenberg, 1988; Thomas & Polwart, 2003). The distribution of the species in the Mediterranean Basin includes the North Mediterranean countries (Euro-Mediterranean) (Jalas & Suominen, 1973; Di Benedetto et al., 1983; Barbero & Quézel, 1994; Quézel & Médail, 2003), Morocco, Algeria and Turkey (Hulten & Fries, 1986). Yew is at present confined to mountainous areas of the basin, following the climatic regression after the last ice age (Garcia et al., 2000). Palaeoecological evidence from various sites in the Mediterranean suggest that yew contributed a significant amount of tree pollen during the Holocene indicating that it was a co-dominant element in the vegetation formations at the time (Grove & Rackhan, 2001; Goni & Hannon, 1999; Penalba, 1994). This has made yew a reliable indicator for ancient temperate forests (Barbero & Quézel, 1994). Today yew grows sparsely along the Mediterranean basin and does not correspond to the places it used to occupy millions years ago (Charles, 1982). Large population of Taxus can still be found in northern Spain and Italy but the majority of natural yew woods are fragmented and isolated (Quézel & Médail, 2003).

Despite its decline in Europe with few exceptions (see Svenning & Magard, 1999) it is currently acknowl-edged there is a lack of information on the species distribution and abundance all over Europe including the Mediterranean area (IPGRI, 2003). This is also the case of Greece, where the information about the species distribution is fragmented despite the fact that numerous accounts of the species can be found in a number of floristic surveys (Dimadis, 1916; Voliotis & Athanasiadis, 1971; Boratynski et al., 1990)

Taxus baccata is a native species in Greece; woodlands grow dispersed in its native habitat in deciduous or mixed forests on mountain slopes and in ravines (Christensen, 1997). They are notable on limestone and grow from an elevation of 800 m up to a maximum limit of 2 200 m, reaching a height of 10-20m in maturity (Athanasiadis, 1986; Tutin et al., 1993; Arampatzis, 1998). As a result of climate change and human disturbance over centuries, yew distribution is more frequently formed by individual

trees within larger forests and to a lesser extent in small groups (Voliotis, 1986). Taxus baccata has a stress-tole-rant life strategy sensu Grime, being slow growing, slow to reach maturity (70 years), long lived (> 1 000 years), shade tolerant but can withstand full sun (Thomas & Polwart, 2003). It is also of high ornamental value and sometimes in Greece is planted in parks and gardens (Christopoulos & Bastias, 1990).

The seeds and foliage of Taxus baccata contain the alkaloid taxin (Vidakovic, 1991), a compound that exhibits significant anticancer properties. The biochemical interest of taxanes and fear of extinction in Greece has been a major influence for increasing managerial interest for the woodland species of Taxus. However, until now there has never been an attempt to carry out a detailed inventory about the spatial distribution and ecological characteristics of the species in the country. For example Voliotis (1986), Strid (1980; 1986) as well as Boratynski et al. (1992) refer to the distribution and limited ecological information of the species. The plant population of the Taxus species is not precisely known. Population estimates are rare and usually not accurate. Moreover the conservation status of Taxus baccata is inadequate in Greece. Only two habitat types that include Taxus baccata have been identified as priority habitat types, according to habitats’ directive 92/43/EEC (Dafis et al., 1997), and limited information can be derived on how well the network of protected areas conserves yew forests in Greece.

The above realisations set out the demand to undertake a nation-wide survey aiming at firstly, acquiring a better knowledge of the geographical distribution and topographical conditions of the existing populations of Taxus baccata in Greece and secondly analysing major community structure parameters of native forests. Furthermore, based on the knowledge of natural distri-bution and community structure, the research sets out a framework of management actions to aid conservation and improve its management.

RESEARCH METHODOLOGY

The need for building up a national inventory in conjunction with the sporadic occurrence of yew in Greece, prohibits the use of random survey plot sampling along transects, on account of the increased cost, workload and time required. The nature of the research implies the use of a specially designed questionnaire targeting forest experts employed in Forest District Offices, as the most

161

DISTRIBUTION AND STAND STRUCTURE OF TAXUS BACCATA POPULATIONS IN GREECE ◆


appropriate survey instrument to collect nationwide information for a dispersed species regarding its major biophysical parameters. Forest District Office (FDO) is the statutory forest authority for each of the 104 forest districts divided in Greece, responsible for managing and conserving forest resources. The questionnaire layout aimed to facilitate filling in time, to help the respondents answer all questions with the highest consistency level and to assist with coding and subsequent statistical analysis. It was deemed imperative from the onset of the research, to reduce the unit of analysis from a surface area to a single point location due to the scarcity and relative rarity of yew trees. Each competent Forest District Office completed a separate questionnaire for each yew population, either in the form of a single tree or clump of trees, found within

its own district. Four major subject areas were covered to acquire information about (i) the geographical distribution of yew populations; (ii) major climatic and topographical conditions; (iii) ownership status and habitat structure of native forests; and (iv) the stand structure characteristics of Taxus baccata. Following the questionnaire analysis, a thorough procedure of personal communication between members of the research team and forest experts was initiated, in order to clarify information and identify possible cause of yew decline. In addition, on-site visits were carried out to illuminate all non-clear situations. The nomenclature and classification of the taxa mentioned are according to Flora Europaea (Tutin et al., 1993).

Data were collected from a systematic postal survey carried out between July and December 1995. Forest

Figure 1. Geographical distribution of Taxus baccata in Greece: 1) (●) found by this study and 2) (--) reported by Voliotis (1986).

162



Figure 2. Distribution of Taxus baccata according to major climatic and topographical parameters.

163



Figure 3. Ownership status and structure of forests including Taxus baccata.

164



experts were asked to complete the questionnaire and return it in the attached pre-paid envelope. In addition, telephone communication with forest experts, helped explaining the research scope and served as a persuasive measure to stimulate participation. Overall, 104 question-naires were distributed to all Forest District Offices and a usable data set of 103 questionnaires was eventually collected and used in the analysis representing a response rate of 99 %, which ensured a nationwide data coverage. The high response rate obtained is similar to that repor-ted by Pantera & Papanastasis (2003) concerning the inventory of Valonia Oak in Greece and can therefore be regarded as a normal average for a field study of this kind. Of the 103 Forest District Offices, 42 (40.78 %) provided information for the occurrence of yew forests within their administrative region; no yew woodlands were identified for the remaining 61 units (59.22 %). In total, the research gathered information on 117 forests containing 173 locations of Taxus baccata. Response rate in each question varies from a low rate of 79.2 % to as high as 100 % indicating that overall, forest respondents had a satisfying knowledge of the biophysical and phy-siographic parameters of yew populations.

Based on the information obtained from questionnai-res, an inventory form was created for every Taxus baccata location identified, as shown in the appendix. A special information system was developed to facilitate analysis and interpretation of the results using the inventory forms as input information source. The information system is essentially an electronic data bank to allow immediate access and management of the information as well as to update the results with new records. The programme dBase IV ver. 2.0 for DOS was used as the platform for developing the data bank. It provided advanced options for data management and the possibility to transferred data to Microsoft Access and edit results in various kinds of format, whether in print or electronically.

RESULTS

Geographical distribution of Taxus baccata

This survey provided information of the geographical occurrence of yew locations as well as of the distribution of various stand sizes across the country. The occurrences mapped in figure 1 give the best information available. Large yew populations are confined in mountainous

areas of central and northern Greece, especially along the Pindos mountain range, the mount Olympus, the Rodopi mountain range and the mount Cholomontas in the peninsula of Halkidiki. Further, small natural stands of yew are found in the Peloponnese and in the island of Evia. Individual yew trees are widespread in several locations throughout the country. Aggregate data suggest that yew populations are found in 28 out of the 51 pre-fectures of Greece.

Climatic and topographical conditions

Research findings indicate that Taxus baccata in Greece occurs in a broad elevational range from sea level to over 1 500 m high (fig. 2a), but it is primarily a montane tree with most populations growing in 501-1 000 m (49.09 %) and 1 001-1 500 m altitude range (41.82 %). Eighteen sites (10.4 %) are found in low alti-tudes (< 500 m) and only five sites (2.89 %) are found at an altitude above 1 500 m. Within this altitudinal range, yew tends to grow on the north-eastern (60.13 %) and north-western (20.89 %) slopes where there is high humi-dity and high insolation. Stands of yew thrive on slopes of almost any inclination but they were primarily found on moderate slopes (57.5 %), less on even surfaces and slopes less than 40 % (26.25 %) and only 16.25 % on moderate to steep slopes (> 71 %). Furthermore, ravines are the most common habitat for yew populations. Of the 127 locations reported, 78.74 % were found to grow on ravines, 16.54 % on smooth mountainsides and only 4.72 % on steep cliffs. According to the bioclimatic types identified for Greece by Mavromatis (1980), the vast majority of yew sites (74.75 %) generally occur in the sub-Mediterranean zone (40>x>0; x= number of biolo-gical dry days during dry period, according to the method of Bagnouls-Gaussen) and 20.23 % in the low-mid-Mediterranean (75>x>40). In addition, only 3.47 % of the identified sites belong in intense mid-Mediterranean (100>x>75) and a tiny 1.73 % of the sites occur in non-arid type (x=0). Taxus baccata populations are typically associated with limestone substrate (46.67 %). They also tend to grow well upon schist (33.94 %) and with lower frequency occurrence on flysch (21.82 %). Shallow soil depth was rarely limiting; most yew stands grow on shallow (38.16 %) and medium soils (36.64 %) and only occasionally on medium-deep (13.74 %) and deep soil depth (11.45 %). This is expected as yew is renowned for having an extensive horizontal root system (Rodwell, 1991).

165



Ownership status and structure of native habitat

The research findings indicate that 173 sites contain Taxus baccata populations, which are distributed along 117 different forests nationwide. Most yew populations occur in state owned forests (65.0 %), followed by municipal forests (19.13 %) as shown in figure 3a. This is not of surprise if one bears in mind that two thirds of the forest area in Greece is public land (65 %) and the remaining 35 % is shared by all other non state own-ers (Papageorgiou et al., 2004). There are only 3 yew populations (2.61 %) under private ownership while in 12.17 % of the reported yew sites, the ownership status remains unclear.

Yew sites are almost equally encountered in pure (45.30 %) and mixed (52.00 %) forests and rarely on partially forested areas (1.71 %). This reflects that yew saplings are favoured by more sheltered, shady and moist locations. In pure stands, yew forms part of an understory of conifers being especially prominent in fir forests (Abies cephalonica) (71.7 %) and very occasionally in Black pine (Pinus nigra) (11.32 %), Beech (13.2 %) or Oak (3.77 %, largely Quercus frainetto) stands. In the Fagion alliance, yew occurs in the Luzulo-Fagetum woodlands especially alongside the Pindus range, in the Asperulo-Fagetum in central and northern Macedonia and in the Cephalanthero-Fagion in Mount Olympus (Dafis et al., 1997). In mixed forests occurrence is less selective. Yew

Figure 4. Community structure of yew populations.

166



trees can be found more frequently in mixed Oak-Beech stands (23.08 %), Fir-Beech (17,95 %), Fir-Black pine (15.38 %) and occasionally in Beech-Fir (7.69 %) and Oak-Fir/Black pine (5.98 %) stands. Owing to its extreme tolerance of shade, yew grows up to form a shrub layer, when scattered and occasionally a part of the canopy in denser populations. Figure 3e shows that most yew populations are found in the understory (56.49 %), less usually in the middle layer (24.43 %), and only 9 sites (6.87 %) have been recorded to contain yew trees as an associate in the canopy. Yew seldom forms a true shrub layer and is usually intermingled with various Juniperus species (29.8 %) and Ilex aquifolium (29.8 %). Other frequently encountered species are Ostrya carpinifolia (14.9 %), Carpinus orientalis (10.6 %) and Quercus ilex (10.6 %) with occasional Acer pseudoplatanus, Corylus avellana and various Ulmus species (4.2 %).

Community structure

Figure 4 depicts a variety of parameters to illustrate the stand structure of yew trees in Greece and give an indication of the scarcity of the species. The study revea-led that populations are rather small, most between 5-50 individuals (48.56 %), or exist in scattered patches of 2-4 trees (13.29 %) and more frequently in isolated specimens (23.7 %). Large yew stands are rare, only 12.13 % of the identified sites contain populations between 51 and 500 individuals while dense stands (over 500 trees) are for-med in only 2.31 % sites. The area each yew community covers was found to vary in proportion to stand size. Thus with the exception of single-tree sites, most populations are found to occupy a land area of 0,2-1 ha (56.88 %) and less frequently extending to 1,1-5 ha (27.52 %). Following the rarity of large yew stands, only 13 popula-tions (11.93 %) extent to an area greater than 5 ha. The largely scattered and small isolated populations impede yew to develop progressively into pure yew forests. It can be inferred from the above, that population fragmentation constitutes the major cause for the degeneration of the species in Greece. This is verified by communication with the forest experts who also suggested indiscriminate fel-ling and to a lesser extent grazing, as other less significant causes of yew decline.

Growth of yew is slow compared to most other trees even under optimum conditions. Consequently, even the oldest individuals do not attain considerable height. The survey provided a measure of height and girth for each recorded individual. For groups we recorded the number of trees in each height and girth class. In Greek forests,

yew occurs most frequently in the undergrowth reaching a height between 3 and 5 m (73.02 %). Only 17.06 % of the trees have a low canopy of 6-10 m in height and in 8.31 % of the identified sites, individuals have exceeded 11 m hei-ght. The bell-shaped form of yew distribution, according to girth, indicates that most of the yew populations recorded are represented by individuals in medium diameter classes (49.1 % in 11-20 cm and 42.23 % in 21-50 cm). There are only a few specimens of yew with a girth below 11cm and above 50 cm. The paucity of individuals in the lowest girth class is a survey weakness; forest experts could not provide accurately information on saplings and seedlings and emphasized particularly on the largest individuals. Although limited numeric data is available to judge yew recolonization, the questionnaire sought to describe the general state of regeneration in yew populations. Findings suggest poor yew regeneration in the identified populations with replacement of individual trees occurring only in 21 out of the 108 yew locations (19.44 %).

Based on the data for all juvenile and mature Taxus trees recorded, the information system estimated the total number of Taxus baccata trees throughout the country to be about 9,000 individuals (recorded data for 9,070 trees). However, the figure is likely to be an underestimate given that infor-mation of new recruitments and seedlings could not be adequately documented by the questionnaire survey.

DISCUSSION

The present study has made explicit that the use of a questionnaire survey to build a Taxus inventory allows only a macroscopic evaluation of major structural parameters pertaining to the natural distribution of yew forests in Greece. Despite the methodological inadequa-cies, the questionnaire survey remains a useful instrument for providing baseline information on the geographical and topographic distribution, population dynamics and the community structure of yew trees at a country level. Such information could provide a significant insight into the conservation potential of the species and set a broad framework of actions to improve its management and conservation. This study provides a comprehensive account of the current state of the species in Greece and considers how forestry practices might be adjusted.

The state of Taxus baccata in GreeceForemost, Taxus baccata is fairly rare and never found

in a large quantity; it occurs at different degrees of

167



population fragmentation ranging from individual trees to more rarely clumps of trees. Indeed, consistent with other researchers (Voliotis, 1986), this study has confirmed that yew grows naturally in high mountains of central and northern Greece but some disparities are also apparent. Voliotis (1986) reported Taxus baccata growing naturally in the islands of Thassos, Samothraki, south of Evia and in forests in central and southeastern Peloponnese that are not verified by the present study. However, small natu-ral stands of yew that extend further east, in the Evros prefecture and in the Sithonia peninsula were reported by this study.

Based on research findings, most yew populations in Greece are restricted to north-eastern, ravines in medium steep slopes at an altitudinal range between 500 and 1 500 m. Taxus baccata increasingly becomes an isolated understory tree within fir pure forests (Abies cephalonica) and various kinds of mixed woodlands, most notably in deciduous mixed stands of beech and oak, mixed conifer stands of fir (Abies cephalonica) and black pine (Pinus nigra) as well as mixed forests of fir and beech. Perhaps a major finding of the research is the marked structural homogeneity of yew stands. The adult population structure is skewed towards small isolated shrubby groups of 5 to 50 individuals having a height range between 3 and 6 m. Large and dense stands of yew trees are rare with most trees forming small groups or being scattered individually within forests throughout the country. The occurrence of such small isolated populations has been a common status along the Mediterranean basin. For instance, Garcia et al. (2000) noted that in southern Spain, T. baccata is restricted to a small number of isolated patches, most with fewer than 10-20 individuals dominated by senescent individuals with a low proportion of saplings and seedlings. Thomas and Polwart (2003) report a similar situation in Portugal and the islands of Corsica and Sardinia.

It has been often pointed out by many authors (e.g. Thomas & Polwart, 2003) that accounting for the distribution of the species is a difficult task. Apart from climate change considered to be responsible for the species decline, particularly in the Mediterranean area, there are also other factors that have been put forward. These include its poor competitive ability with respect to light if compared to other species that grow together with yew such as hornbeam and beech and its extermination by shepherds/farmers since it was poisonous to animals (Ellenberg 1988). Although in exceptional circumstances high regeneration has been reported (e.g. Garcia et al., 2000), in general, the opposite is the case with regeneration

been further impeded by herbivory. The research implies an apparent regeneration failure of most yew populations in Greece. Communication with forest experts indicated grazing as a potential factor in lack of recruitment but factual evidence is lacking to substantiate the above assertion. Instead, the weakness of the questionnaire survey to provide an accurate record of seedlings and saplings, may be a more plausible explanation.

Implications for conservationand management

Currently, there is minimal forestry importance for the woodland species of Taxus baccata in Greece, apart from the spiritual value of particularly old and large individuals in churchyards (Strid, 1980). Forest management today aims explicitly at the foresighted sustainable timber yield of the species with the most economic importance and management plan is the main planning tool. Hence, no separate management plans have been conducted for yew populations and no concrete actions are prescribed in the plans drawn for forests with Taxus in order to preserve the integrity of yew populations. However, the importance of T. baccata as an alternative source of taxoid production (Appendino et al., 1992), has brought forward the scientific interest for the conservation of the species. There is, at present, increasing research on the anticancer properties of taxanes isolated from yew trees (Guéritte, 2001). Originally extracted from the bark of the Pacific yew (T. brevifolia), taxol (paclitaxel) can now be synthesized from a taxol congener 10-deacetylbaccatin III) which can be extracted at approximately 10 times the quantity per unit weight from the leaves of Taxus baccata and other Taxus species than from Taxus brevifolia bark (Dennis, 1988). In Greece, yew belongs to the group of rare and potentially endangered species on account of a number of causes, most notably its highly fragmented geographical distribution followed by indiscriminate felling of mature trees. The largely small and scattered populations restrict yew to evolve gradually into denser stands. Anzalone et al. (1997) noted that the rarity of Taxus in the Mediterranean flora of southern Europe may be due to the yew being left as a declining relict as the climate has become less oceanic. Consequently, in the face of predicted climate change, it seems likely that yew populations in Greece will further deplete.

The fringe occurrence and biochemical interest of Taxus baccata yield certain managerial implications for the forest authorities. Most importantly, the restoration of Taxus woodland areas could be alleviated through applied

168



management and possibilities of increasing the area of habitat types. Population viability management methods should be restricted to the few dense stands identified by the survey. Appropriate silvicultural manipulation needs to be applied to include a management by thinning to let in more light and help regeneration. In the multiplicity of remaining yew locations that include a few number of individuals, usually grown in the understory of pure or mixed forests, management should attempt to facilitate yew recolonization by creating more gaps. This approach is more likely to increase the possibility of survival of yew populations and expand tree number. However, individual trees are found more difficult to evolve into a forest community and success is a function of favourable biotic and abiotic conditions (Carvalho et al., 1999).

Applied management interventions should be followed up by policy reforms such as changes in the forest legislation to prohibit yew cutting, in order to stimulate protection of yew individuals against indiscriminate felling in managed forests. Also the underpinning of state forest ownership prevalence for the management and restoration of Taxus baccata populations should not be underestimated. It can be a supportive factor in the sense that integrated, comprehensive and explicit forest management actions can be evenly applied to two-thirds of the country’s forestland.

Another implication concerns the conservation of the species, especially in relation to national parks and other protected areas. Granting to yew woodlands a sta-tutory protection could improve the population viability primarily as a virtue of cessation of grazing and human disturbance. Currently, the conservation status of Taxus baccata is deficient in Greece. Despite the classification of the habitats “Mixed beech forests with Taxus and Ilex” and “Mountainous coniferous woods with Taxus baccata” (92/43/EEC codes 9120 and 9580 respectively) as priority habitat types (Dafis et al., 1997), only twelve representative sites (4,05 %) accounting for an area of 3,717 ha or 0,16 % of the total proposed protected land, contain Taxus woodlands. Moreover, with the exception of a national park (Mt. Olympus) and a protected natural monument (Mt Voras peaks), which provide strict statu-tory protection, all remaining sites are hunting reserves with a rather ambiguous conservation potential. Thus, increasing the share of forests including Taxus under a statutory protection regime is believed to aid broader expansion of yew in Greece.

ACKNOWLEGEMENTS

The General Secretariat for Research and Technology of the Greek Ministry of Development for supporting this research within the framework of a wider project to study the pharmaceutical biosynthesis of Taxol. We thank D. Trakolis from the Forest Research Institute of Thessaloniki and F. Galanos from the Institute of Mediterranean Forest Ecosystems and FPT in Athens for their valuable help in communicating with Forest District Offices and verifying the data on the inventory form for several sites with Taxus baccata trees.

ReferencesANZALONE, B., LATTANZI, E., LUCCHESE, F. & PADULA, M. 1997.

The vascular flora of the Circeo national park. Webbia, 51: 251-341.

APPENDINO, G., 1993. Taxol (paclitaxel): historical and ecological aspects. Fitoterapia, 1: 5-24.

APPENDINO, G., GARIBOLDI, P., PISETTA, A., BOMBARDELI, E. & GABETTA, B., 1992. Taxanes from Taxus baccata. Phytochemistry, 31: 4253-4257.

ARAMPATZIS, TH., 1998. Shrubs and trees of Greece, vol. I. Technological and Educational Institute of Drama. Drama. 322 p. (in Greek).

ATHANASIADIS, N., 1986. Forest Phytosociology. Aristotle University of Thessaloniki, Thessaloniki. 309 p. (in Greek).

BARBERO, M. & QUÉZEL, P., 1994. Place, rôle et valeur historique des éléments laurifoliés dans les végétations préforestières et forestières ouest-méditerranéennes. Annali di Botanica, 52: 81-133.

BORATYNSKI, BRAWICZ, K. & ZIELINSKI, J., 1992. Chorology of Trees and Shrubs in Greece. Polish Academy of Sciences, Kormik. 286 p.

CARVALHO, A., REBELO, A. & DIAS, J. 1999. Distribution and natural regeneration of yew trees in the national parks of Peneda-Geres (Portugal) and Baixa Limia Serra-Xures (Spain). Revista de Biologia, 17: 43-49.

CHARLES, J.-P., 1982. Etude climatique, floristique et statistique des peuplements d’ifs (Taxus baccata L.) du sud de la France (Provence, Languedoc, Corse). Thèse de doctorat de 3e cycle, université d’Aix-Marseille III, Marseille, 226 p.

CHRISTENSEN, K.I., 1997. Taxus L. In: STRID, A., & TAN, K. (ed), Flora hellenica vol 1. Koeltz Scientific Books, Königstein: p. 15.

CHRISTOPOULOS, G. & BASTIAS, I., 1990. Phytologia. Ekdotiki Athinon, Athens. 435 p. (in Greek).

169



DAFIS, S., PAPASTERGIADOU, E., GEORGIOU, K., BABALONAS, D., GEORGIADIS, TH., PAPAGEORGIOU, M., LAZARIDOU, TH. & TSIAOUSI, B., 1997. Habitat Directive 92/43/EEC. Ecotypes in Greece: Natura 2000 network. Gen. Directorate XI EC, Goulandri Natural History Museum, EKBY, Thessaloniki. 923 p.

DENNIS, J.N., 1988. A highly efficient, practical approach to natural taxol. J. Am. Ch. Soc., August 1988: 5917.

DI BENEDETTO, L., LEONARDI, S. & POLI, E., 1983. Taxus baccata L. in Sicilia. Not. Fitosoc., 18: 1-18.

DIMADIS. D., 1916. Forest vegetation of Greece. National Printing Office of Greece, Athens. 159 p.

ELLENBERG, H., 1988. Vegetation ecology of central Europe. Cambridge University Press, Cambridge, UK.

GARCIA, D., ZAMORA, R., HODAR, J.A., GOMEZ, J.M. & CASTRO, J., 2000. Yew (Taxus baccata L.) regeneration is facilitated by fleshy-fruited shurbs in Mediterranean environments. Biol. Conserv., 95: 31-38.

GONI, M.F.S. & HANNON, G.E., 1999. High altitude vegetational pattern on the Iberian Mountain chain (north-central Spain) during the Holocene. The Holocene, 9: 39-57.

GROVE, A.T. & RACKHAM, O., 2001. The Nature of the Mediterranean Europe. Yale University Press, New Haven.

GUÉRITTE, F., 2001. General and recent aspects of the chemisttry and strucutre-activity relationships of taxoids. Cur. Pharm. Des., 7: 1229-1249.

HULTEN, E. & FRIES, M., 1986. Atlas of north European vascu-lar plants: north of the tropic of cancer I-III. Koeltz Scientific Books, Königstein.

IPGRI, 2003. EUFORGEN www.ipgri.cgiar.org/networks/euforgen/.

JALAS, J. & SUOMINEN, J., 1973. Atlas Florae Europaeae: distribution of vascular plants in Europe. Vol. II Gymnospermae (Pinaceae to Ephedraceae).

MAVROMATIS, G., 1980. Bioclimate of Greece; correlating climate and natural vegetation, bioclimatic maps. For. Res., 1: 54-63. (in Greek).

PANTERA, A. & PAPANASTASIS, V.P., 2003. Inventory of Valonia oak (Quercus ithaburensis Decaisne subsp. macrolepis (Kotschy) Hedge & Yalt. in Greece. Geot. Sc. Iss., 14: 34-44.

PAPAGEORGIOU K., VAKROU, A. TRAKOLIS, D. & MALAMIDES, G., 2004. Sustainable Forest Management: The challenge ahead for the Greek State Forestry. In: Humphreys, D. (Ed) Forests for the future: findings on National Forest Programmes from COST Action E19, European Science Foundation (in press).

PENALBA, M.C. 1994. The history of the Holocene vegetation in northern Spain from pollen analysis. J. Ecol., 82: 815-832.

QUÉZEL, P., 1985. Definition of the Mediterranean region and the origins of the flora. In: Gomez-Campo, C. (Ed) Plant Conservation in the Mediterranean Area. (ed. C.), Dordrecht, pp. 9-24.

QUÉZEL, P. & MÉDAIL, F., 2003. Ecologie et biogéographie des forêts du bassin méditerranéen. Elsevier, Paris, 573 p.

RODWELL, J.S., 1991. British plant communities, vol. 1. Woodlands and scrub. Cambridge University Press, UK.

STRID, A., 1980. Plants of Mt. Olympus. Goulandri Natural History Museum, Athens. 420 p.

STRID, A., 1986. Mountain Flora of Greece. Vol. 1. Cambridge University Press, Cambridge. 822 p.

SVENNING, J.C. & MAGARD, E., 1999. Population ecology and conservation status of the last natural population of English yew Taxus baccata. Denmark. Biol. Conserv., 88: 173-182.

THOMAS, P.A., & POLWART, A., 2003. Taxus baccata L. J. Ecol., 91: 489-524.

TUTIN, T.G., BURGES, N.A., CHATER, A.O., EDMONDSON, J.R., HEYWOOD, V.H., MOORE, D.M., VALENTINE, D.H. WALTERS, S.M. & WEBB, D.A., 1993. Flora Europaea, 5 vol. Cambridge University Press, London.

VIDAKOVIC, M., 1991. Conifers: morphology and variation. Graficki Zavod Hrvatske, Zagreb.

VOLIOTIS, D., 1986. Historical and environmental significance of the yew (Taxus baccata L.). Isr. J. Bot., 35: 47-52.

VOLIOTIS, D. & ATHANASIADIS, N., 1971. Trees and shrubs. Thessaloniki. 294 p.

170



APPENDIX

Inventory form of the Taxus baccata populations in GreeceLocation code number:

Prefecture’s name, land area (Ha) and forest area (Ha)Forest District Office’s name and category

Characteristics of forests with Taxus trees — name of the forest— municipality (where the forest belongs to)— ownership status (state, private, municipal, monasteries, other)— forest structure (pure forest – one main species –, mixed, partially forested area) — silvicultural type (high forests, coppice, high and coppice)— dominant tree species (in pure forests)— dominant tree species (in mixed forests)— other tree species present in the forest

Characteristics of sites with Taxus trees— name of the site— altitude above sea level (m) (< 500 m, 501-1 000 m, 1 001-2 000 m, > 2 000 m)— bioclimatic character (based on Mavromatis, G. 1980)— site exposure (orientation) (east-southeast, south-southwest, north-northeast, west-northwest)— soil slopes (%) (< 40 %, 41-70 %, 71-100 %, > 100 %)— bedrock (limestone, flysch, schist, other):— soil depth (non-deep < 50 cm, mean depth 51-100 cm, deep > 1m)

Growth conditions of Taxus populations— population density (only 1 tree, 2-4 trees, 5-20, 21-50, 51-100, over 100 trees)— spread of Taxus trees in the forest (in single trees, in groups, in thickets, in clumps)— covered area (Ha)— participation in the structure of the forest (in the upper layer, in the middle layer, in the understory)— distribution of Taxus trees (along streams, on smooth mountainsides, on steep sides)— general appearance (excellent, good, fair, poor)— existence of regeneration (yes, no)

Characteristics of Taxus trees— breast diameter (cm) (< 5 cm, 6-10 cm, 11-20 cm, 21-50 cm, > 50 cm)— height (m) (< 2 m, 3-5 m, 6-10 m, 11-20 m, over 20 m)

Geographical location of the site (on a map scale 1/50 000) and also information on how to approach the site and contact persons.

171


AbstractHerbivory is thought to be an important factor in the ecology of introduced species. A lower palatability to the herbivores may con-tribute to the success of invasive species in their new habitat. Here we investigate the palatability (to the generalist herbivore snail Cepaea hortensis) of 11 non-native plant species found on Crete and compare it to that of 13 native species. These were collected from three different habitats (dunes, olive groves and shrublands), so as to be able to reconstruct a community background palatability. Our results indicate that non-natives fall into the range of palatabi-lities found among the natives, with no significant overall difference between these groups. In all three tested habitats, non-natives were more palatable than the native community background. Only in dunes was one non-native species, Acacia saligna, markedly less palatable than the community average. Palatability of the species was not related to their commonness on Crete, independent of being native or not.

Key-wordsAlien species, Cepaea hortensis, community palatability, Crete, dune, herbivory, Mediterranean island, olive-grove, shrubland

RésuméL’herbivorie est connue comme étant un facteur important interve-nant dans l’écologie des espèces introduites. Une faible palatabilité vis-à-vis des herbivores est susceptible de contribuer au succès des espèces invasives dans leur nouvel habitat. Dans ce travail, nous avons étudié la palatabilité (vis-à-vis d’un gastéropode herbivore et généraliste Cepaea hortensis) de 11 espèces végétales introduites en Crète et nous l’avons comparée à celle de 13 espèces végétales indigè-nes. Ces espèces sont issues de trois catégories d’habitats (dunes, oli-veraies et matorrals), de façon à permettre une reconstruction de la palatabilité à l’échelle de la communauté. Nos résultats indiquent que la palatabilité des espèces introduites s’insère dans la gamme de palatabilité rencontrée chez les espèces indigènes, sans différence significative globale entre les deux groupes. Toutefois, pour chacun des trois habitats testés, la palatabilité des espèces introduites s’est avérée plus élevée que la palatabilité moyenne à l’échelle de la communauté. Une seule espèce introduite, Acacia saligna, présente dans les formations dunaires, est apparue comme nettement moins appétante que la moyenne de la communauté. La palatabilité des espèces n’est pas reliée à leur abondance en Crète, que l’espèce soit indigène ou introduite.

Mots-clésEspèces introduites, Cepaea hortensis, palatabilité des communautés, Crète, dune, herbivorie, île de Méditerranée, oliveraie, matorral

Comparing the palatability of Mediterranean or non-nativeplants in CreteÉtude comparative de la palatabilité de végétaux méditerranéens ou exotiques en Crète

Carsten F. Dormann1, Rachel King2

1. Applied Landscape Ecology, UFZ Centre for Environmental Research Leipzig-Halle, Permoserstr. 15, 04318 Leipzig, Germany. Email: [email protected]. Tel: ++44-(0)341-235 2953 – Fax: ++44-(0)341-235 2511

2. Scott Wilson, Scott House, Basing View, Basingstoke, RG21 4JG, UK

172

◆ C. F. DORMANN & R. KING


INTRODUCTION

Biological invasions of natural communities by non-native plant species is one of the most serious threats to biodiversity (Heywood, 1995), especially in Mediterranean-type ecosystems (Huenneke, 1988; Vitousek, 1988). The extend to which introduced spe-cies become established, and later become pests, differs widely (Williamson & Fitter, 1996; Williamson, 1999), and seems to be a function of both species traits (“inva-siveness”) and community susceptibility (“invasibility”: Alpert et al., 2000).

One of the interfaces between invasiveness and invasibility is the interaction between non-native plants and their herbivores. It has been argued that palatability differences allow non-native plant species to become invaders (Blossey & Nötzold, 1995). This would imply that grazers, browsers and plant parasites are important in controlling the invasion process (Noble, 1989).

Anti-herbivore chemistry of the non-native may be very different from that of the natives (often helped by the fact that the non-natives are recruited from families new to the community, Rejmánek, 1996). Thus, they present a novel suite of feeding deterrents to the local herbivores, which make them more effective in their defence against the consumers (it has even been argued that co-evolu-tion between plants and their consumers is the cause of the diverse secondary metabolism found in plants, see Harborne, 1997; Hartley & Jones, 1997, for review). To a generalist herbivore, which is confronted with different defence cocktails in each plant species, such differences are probably less important, as it is far less engaged in co-evolution with its forage than a monophagous specialist.

To date, very little information on the palatability of non-native plant species compared to that of the local natives is available (Crawley et al., 1996; Williamson & Fitter, 1996). More work has concentrated on related topics, such as comparing the performance of a non-native invader in its native and its new range (Callaway and Aschehoug, 2000; Bossdorf et al., 2004) or the effect of herbivory on exotic and indigenous congeneric species (e.g. Schierenbeck et al., 1994, e.g. Radford & Cousens, 2000). These studies however did not take account of the background palatability of native species when addressing the effect of herbivory on the establishment of a non-native.

In this study, we report on a bioassay-palatability trial comparing 11 non-native and 13 native plant species common to three habitats on the Greek island of Crete. We hypothesis that invasive introduced plant species on

Crete have a lower palatability to generalists than natives. The richness of the local flora, its many island endemics and the numerous non-native plant species present on the island make Crete an island of high invasion risk (Vitousek, 1988). Two of the selected habitats (dunes and shrublands) are important communities for rare species, while the third (recently abandoned olive groves) repre-sents the anthropogenic habitat most commonly invaded by non-native species (Lonsdale, 1999).

METHODS

The study was performed on Crete, Greece (35.5° N 25° E) from 10 to 20 May 2002. Ten sites were selected for each of three habitat types: olive grove, coastal dune and phrygana (the arid dwarf-shrubland frequent in the eastern Mediterranean). In each site, cover of the most common plant species and all those used in this study were recorded. Non-native species occurred in none of the plots chosen for the recordings, because they usually produce monodominant stands within the investigated communities.

Sampling procedureand bioassay-palatability trial

Ca. 10 g leaf material of 13 typical common native and 11 common non-native plant species (table 1) was collected near the above sites. The five replicates for each plant species were several km apart. Within eight hours extracts were prepared following Grime et al. (1993): 1 g fresh leaf material was ground in 10 ml H2Odist., filtered through Whatman #1 filter paper and then frozen until further use.

Palatability trials were run according to (Grime et al., 1993): the test was a comparison between 0.18 ml extract on a 1.5 cm × 1.5 cm piece of Whatman #1 filter paper (added in three doses) and filter paper without extract. These filter papers were dried at 30 °C in a drying oven and then weighed to the nearest mg. Both filter papers were re-wetted with equal amounts of water just before offering them to the snails. Pre-trials have shown that snails eat the moist filter paper. After the trial, the filter paper was left to dry again and then re-weighed. Each sample extract was fed to two different snails (i.e. two subsamples per replicate).

As a bioassay agent for palatability, some 50 Cepaea hortensis garden snails were collected. Grime et al.

173

COMPARING THE PALATABILITY OF NATIVE AND NON-NATIVE MEDITERRANEAN PLANTS ◆


(1993) used the species Cepaea nemoralis, which is rare in Scotland (Kerney, 1976). We selected C. hortensis not only because it is an accepted generalist herbivore (Grime et al., 1993), but also because all plant species tested are unknown to the snails. Thus, there was no ‘native’ plant species in the palatability trial from the snails’ perspective. Hence, no bias with respect to coevolution was introdu-ced. This could have led to a false representation of the general palatability of that species. Moreover, most larger snails in the cultivated areas of Crete are in fact non-native (e.g. Helix aspera), being imported from the mainland (Francisco Welter-Schultes, personal communication). By focussing solely on the palatability of water-soluble leaf content, we disregard the importance of leaf tough-ness, hairiness, etc. As dozens of different herbivores are consuming the plants (sheep, goat, different species of snails, beetles and bugs), it is impossible to test for specific palatability to all these herbivores. We therefore resorted to only investigate water-extractable leaf content, which will be consumed by all these leaf herbivores.

The individually marked snails were kept in a cage under near-natural conditions until used for the trial. For this, they were starved for 24 h, then put together with the extract and the control filter paper in a moist plastic bowl and left for 16 h overnight. After the trial snails were fed on lettuce for two days before the starving for the next trial. Each snail was used approx. five times for different plant species extracts. No snail received the same com-bination of plant species, and sequences of plant species were randomised. Feeding trials took place between 11th

June and 7th July 2002. Air temperature during the fee-ding period was recorded at half-hourly intervals.

A palatability index was calculated as the preference for extract over control per total amount eaten:

PI ranges from Ð1 to +1, with negative values indica-ting rejection of extract compared to water. The nonlinear index used by Grime et al. (1993) = extract eaten/control eaten) is not sensible when very little of the control or extract sample has been eaten, as it produces a bias in favour of high index values (which can be cured by the taking the logarithm of the ratio Elston et al., 1996). Using the same index (but calling it “acceptability index”), Dirzo (1980) had to discard those trials where the test disk has been rejected or where control disks were con-sumed less than half (although stating that these tests are a valid measure of acceptability). This was the case in some of our trials. However, their index is related to ours by the formula:

Species distribution data

The data for the species’ distribution on Crete were taken from Turland (1992) and Turland et al. (1993). They mapped all species in a 8.25 km × 8.25 km grid. We used the number of occupied cells as an index of distribution.

Statistical analysis

Arcsin (0.1* square-root (x+1)) transformed data from the feeding trials were analysed with a mixed effect

Native Non-native

Ammophila arenaria, Poaceae (Aa) Acacia saligna, Fabaceae (As)

Calicotome villosa, Fabaceae (Cv) Agave americana, Agavaceae, (Ag)

Ceratonia siliqua, Fabaceae (Cs) Ailanthus altissima, Simaroubaceae (Aia)

Cupressus sempervirens,Cupressaceae(Cu) Arundo donax, Poaceae (Ad)

Forb mixture from the undergrowth (F) Carpobrotus acinaciformis, Aizoaceae (Ca)

Medicago marina, Fabaceae (Mm) Nicotiana glauca, Solanaceae (Ng)

Olea europaea, Oleaceae (Oe) Opuntia fi cus-indica, Cactaceae (Of)

Otanthus maritimus, Asteraceae (Om) Oxalis pes-caprae, Oxalidaceae (Op)

Pancratium maritimum, Amaryllidaceae (Pm) Phytolacca americana, Phytolaccaceae (Pa)

Quercus coccifera, Fagaceae (Qc) Ricinus communis, Euphorbiaceae (Rc)

Sarcopoterium spinosum, Rosaceae (Ss) Robinia pseudoacacia, Fabaceae (Rp)

Thymus capitatus, Lamiaceae (Tc)

Urginea maritima, Liliaceae (Um)

Table 1. Native and non-nativespecies sampled (abbreviations).

174



model (function “lme” in R: Pinheiro and Bates; 2000), as subsamples were nested within replicates. ‘Status’ (native or non-native) was used as the fixed effect and ‘species’ as a random effect, since we were not interested in the specific identity, but rather in the difference between native and non-native. ‘Temperature at time of feeding’, ‘woodiness of the species’ and ‘snail weight’ (as well as all interactions) were used as additional explanatory varia-bles, but were excluded from the final model as their contributions were far from significant (P > 0.4, model simplification following suggestions of Crawley 2002).

RESULTS

Species differed widely in their palatability to snails (fig. 1). However, our hypothesis, that native and intro-duced species differ in their palatability to a generalist herbivore was not confirmed. They did not differ signifi-cantly in their palatability (fig. 2; F1, 22 = 2.12, P= 0.160; log-ratios produced the same results).

Within habitats, we compared the palatability of all non-native species to that of the natives occurring in this habitat, weighted by their abundance according to our vegetation recordings. This background palatability

had values of PIolive grove = Ð0.407, PIphrygana = Ð0.499 and PIdune = Ð0.265. Apart from the dunes, non-native palatability was always higher than this background level. Only in the dunes did the PI-value of the introdu-ced Acacia saligna indicate lower palatability than that of the native community (see fig. 1). Palatability and cover for native species was unrelated in all three habitats (Pearson’s correlation: P > 0.3 for all three habitats).

The species also differ widely in their distribution on Crete as measured by the number of occupied grid cells. This was, however, not related to their palatability (fig. 3). While non-natives were overall less common than natives (F1, 21 = 11.32, P < 0.01), palatability was unrelated to distribution for both types (F1, 21 = 0.01, P = 0.93).

DISCUSSION

In this comparison of the palatability of 13 native and 11 non-native Cretan plant species we could not detect a difference using a generalist herbivore as a bioassay. In fact, non-natives were even slightly more palatable than natives (although not significantly so: fig. 2). We therefore have to conclude that there is no a priori reason to believe that the ‘nativeness’ status of a species has any relevance

Fig. 1. Palatability sequence of the tested 24 species. Positive values indicate

preference of extract over water. Abbreviations

as in table 1.

175



for the probability of generalist herbivores limiting its success at establishment in a given habitat. Consistent with this finding, there was no relationship between pala-tability and distribution of the species on Crete.

Our study does not address palatability to specialist herbivores, therefore an extrapolation to the effect of her-bivory per se on the establishment of non-native species is not possible. Possibly species that suffer heavily from a native specialist herbivore might evolve higher compe-titive performance in its absence (the EICA hypothesis: Blossey & Nötzold, 1995; Keane & Crawley, 2002), which is based on the ideas of Herms and Mattson (1992). A test of this coevolutionary hypothesis is beyond the scope of this study.

Palatability values measured by our experiments showed a greater variability than those given by Dirzo (1980) and Grime et al. (1993). In these two studies, (converted) PI values for water extracts of grasses and dicotyledons are usually in the range of 0 to Ð0.15 (Grime et al., 1993). For plants where cell sap was known to be distasteful, PI values went down to Ð0.33, but clearly neither above 0 nor below Ð0.5, as was the case in our experiment. This is probably a consequence of their rejection of trials where only control material was consumed, biasing against low palatability. In another palatability experiment, when simultaneously offering 43 species to a snail and a cric-ket, consumption data show clear rejections (i.e. PI values of Ð1) for 24 and 10 species, respectively (Grime et al., 1996). Moreover, since filter paper ranking 12th in the list, snails consumed less of 73 % of all species than of the

Fig. 2. Palatability of natives compared to non-natives.

Fig. 3. Distribution of the 24 species on Crete was unrelated to palatability. Grey dots refer to non-natives, black dots to natives.

filter paper control. However, these trials were performed on fresh material, not extracts, and are thus not directly comparable to our situation.

Comparing the palatability of non-natives not only to that of natives, but more specifically to common native species within a given habitat has not been attempted before. Given that non-natives are slightly more pala-table than natives, it is not surprising to find that the background PI-values are also generally more negative, i.e. natives are less palatable. The exception of Acacia saligna (syn. Acacia cyanophylla) in the dunes is remar-kable, as this species is a pest in South Africa (Roux & Middlemiss, 1963) and became invasive more recently in western Mediterranean coastal dunes (Cronk &Fuller, 1995). Investigations into this coincidence of high com-munity background palatability and low palatability of Acacia saligna may be fruitful. As for the other species, it is remarkable that the species with the highest palatability, Ricinus communis or Castor Oil Plant, ironically has seeds highly toxic to mammals (ricin leads to the agglutination of red blood cells). This high toxicity does not hold for leaves and snails, apparently, as none of them died in the weeks following the experiments.

Palatability is related to a plant’s growth rate (Herms & Mattson, 1992). Faster growing species allocate less assi-milates to anti-herbivore defense, thus being more pala-table (Hartley & Jones, 1997, Jones & Hartley, 1999).

176



177



However we have no data for the species tested in this study, and a recent review found no supportive evidence for the hypothesis that non-natives have higher growth rates than natives (Daehler, 2003). Nevertheless, it may be that our specific selection of non-natives is indeed more palatable because the have a higher growth rate.

Another trait of the foliage that reduces consumption is leaf and tissue toughness. As we produced extracts, we did not test this trait, but it may play an important role in the field. Sclerophylly is very common in woody Mediterranean plant species, due to the ecophysiologi-cal constraints of the climate (Larcher, 1995) and the deterrent effect of tough leaves on herbivores (Davidson, 1993). This holds true as much for the native (Ammophila arenaria, Ceratonia siliqua, Cupressus sempervirens, Olea europaea, Quercus coccifera and Thymus capitatus) as for the non-native species in this experiment (Acacia sali-gna, Agave americana, Arundo donax, Nicotiana glauca and Opuntia ficus-indica).

The Cretan landscape has been subject to intense grazing by livestock for centuries (Rackham & Moody, 1996). Those plants now present must therefore have adapted to this situation. As additional winter feeding keeps livestock densities usually well above the popula-tion density supported by the vegetation alone, one could assume the plants are accustomed to very high levels of grazing. Hence new plant species are more likely to come from habitats with lower grazing intensity. For a generalist herbivore this could mean a higher palatability of non-natives compared to the native Cretan species. This is in fact what we found, although the slightly higher mean palatability was statistically not significant.

ACKNOWLEDGEMENTS

The work was carried out as part of the EU-funded 5th framework project EPIDEMIE (EVK2-CT-2000-00074). We gratefully acknowledge comments of Phil Lambdon, Phil Hulme and two anonymous referees on an earlier version.

APPENDIX

The table in the appendix gives data on PI, nativeness, woodiness, distribution and cover in the three vegetation types for all 24 species.

ReferencesALPERT, P., E. BONE & C. HOLZAPFEL, 2000. Invasiveness, inva-

sibility and the role of environmental stress in the spread of non-native plants. Persp. Plant Ecol. Evol. Syst. 3, 52-66.

BLOSSEY, B. & R. NÖTZOLD, 1995. Evolution of increased com-petitive ability in invasive nonindigenous plants: a hypothesis. J. Ecol. 83, 887-889.

BOSSDORF, O., S. SCHRÖDER, D. PRATI & H. AUGE, 2004. Palatability and tolerance to simulated herbivory in native and introduced populations of garlic mustard (Alliaria petio-lata). Am. J. Bot. 91, 856-862.

CALLAWAY, R. M. & E. T. ASCHEHOUG, 2000. Invasive plants versus their new and old neighbors: a mechanim for exotic invasion. Science 290, 521-523.

CRAWLEY, M. J., 2002. Statistical Computing: An Introduction to Data Analysis using S-Plus. Wiley, New York.

CRAWLEY, M. J., P. H. HARVEY & A. PURVIS, 1996. Comparative ecology of the native and alien floras of the British Isles. Philos. Trans. R. Soc. Lond. Ser. B 351, 1251-1259.

CRONK, Q. C. B. & J. L. FULLER, 1995. Plant Invaders. Chapman & Hall, London.

DAEHLER, C. C., 2003. Performance comparisons of co-occurring native and alien invasive plants: implications for conservation and restoration. Annu. Rev. Ecol. Evol. Syst. 34, 183-211.

DAVIDSON, D. W., 1993. The effects of herbivory and granivory on terrestrial plant succession. Oikos 68, 23-35.

DIRZO, R., 1980. Experimental studies on slug-plant interac-tions. I. The acceptability of thirty plant species to the slug Agriolimax caruanae. J. Ecol. 68, 981-998.

ELSTON, D. A., A. W. ILLIUS & I. J. GORDON, 1996. Assessment of preferences among a range of options using log ratio analysis. Ecology 77, 2538-2548.

GRIME, J. P., J. H. C. CORNELISSEN, K. THOMPSON & J. G. HODGSON, 1996. Evidence of a causal connection between anti-herbivore defence and the decomposition rate of leaves. Oikos 77, 489-494.

GRIME, J. P., S. F. MACPHERSON-STEWART & R. S. DEARMAN, 1993. Palatability. In G. A. F. Hendry and J. P. Grime, edi-tors. Methods in Comparative Plant Ecology. Chapman & Hall, London, p. 117-121.

HARBORNE, J. B., 1997. Plant secondary metabolism. In M.J. Crawley, editor. Plant Ecology. Blackwell, Oxford, p. 132-155.

HARTLEY, S. E. & C. G. JONES, 1997. Plant chemistry and her-bivory, or why the world is green. In M.J. Crawley, editor. Plant Ecology. Blackwell, Oxford, p. 284-324.

HERMS, D. A. & W. J. MATTSON, 1992. The dilemma of plants: to grow or defend. Quart. Rev. Biol. 67, 283-335.

178



HEYWOOD, V. H., 1995. The Mediterranean flora in the context of world biodiversity. Ecol. Medit. 20, 11-18.

HUENNEKE, L., 1988. SCOPE program in biological invasions: a status report. Conserv. Biol. 2, 8-10.

JONES, C. G. & S. E. HARTLEY, 1999. A protein competition model of phenolic allocation. Oikos 86, 27-44.

KEANE, R. M. & M. J. CRAWLEY, 2002. Exotic plant invasions and the enemy release hypothesis. Trends Ecol. Evol. 17, 164-170.

KERNEY, M. P., 1976. Atlas of the Non-Marine Mollusca of the British Isles. Pemberley Books, Bucks, UK.

LARCHER, W., 1995. Physiological Plant Ecology, 3rd edition edi-tion. Springer, Berlin.

LONSDALE, W. M., 1999. Global patterns of plant invasions and the concept of invasibility. Ecology 80, 1522-1536.

NOBLE, I. R., 1989. Attributes of invaders and the invading process: terrestrial and vascular plants. In J. A. Drake, H.A. Mooney, F. di Castri, R. H. Groves, F. J. Kruger, M. Rejmanek, and M. Williamson, editors. Biological Invasions: a Global Perspective. John Wiley & Sons, Chichester, p. 301-313.

PINHEIRO, J. C. & D. M. BATES, 2000. Mixed-Effect Models in S and S-plus. Springer, New York.

RACKHAM, O. & J. MOODY, 1996. The Making of the Cretan Landscape. Manchester University Press, Manchester.

RADFORD, I. J. & R. D. COUSENS, 2000. Invasiveness and com-parative life-history traits of exotic and indigenous Senecio species in Australia. Oecologia 125, 531-542.

REJMÁNEK, M., 1996. A theory of seed plant invasiveness: the first sketch. Biol. Conserv. 78, 171-181.

ROUX, E. R. & E. MIDDLEMISS, 1963. Studies in the autecology of Australian acacias in South Africa. The occurrence and distribution of Acacia cyanophylla and A. cyclops in the Cape Province. S. Afric. J. Sci. 59, 286-294.

SCHIERENBECK, K. A., R. N. MACK & R. A. SHARITZ, 1994. Effects of herbivory on growth and biomass allocation in native and introduced species of Lonicera. Ecology 75, 1661-1672.

TURLAND, N. J., 1992. Floristic notes from Crete. Bot. J. Linn. Soc. 108, 345-357.

TURLAND, N. J., L. CHILTON & J. R. PRESS, 1993. Flora of the Cretan Area. Annotated Checklist and Atlas. Natural History Museum, H.M.S.O., London.

VITOUSEK, P. M., 1988. Diversity and biological invasions of oceanic islands. In E. O. Wilson, editor. Biodiversity. National Academy of Science, USA, Washington, p. 181-189.

WILLIAMSON, M., 1999. Invasions. Ecography 22, 5-12.

WILLIAMSON, M. H. & A. FITTER, 1996. The characters of suc-cessful invaders. Biol. Conserv. 78, 163-170.

179


AbstractNo official and coherent data on the distribution of the European chestnut exist despite its wide range of distribution and the impor-tant economic role it has played in many countries. In 1997, in the framework of the COST action G4 “Multidisciplinary Chestnut Research”, quantitative and qualitative data on chestnut forests were collected, mostly from the National Forest Inventories, in order to provide as sound a picture as possible of this important European resource. A total of 2.25 million hectares of forest dominated by ches-tnut were recorded, with 1.78 million hectares (79.0 %) cultivated for wood and 0.43 million hectares (19.3 %) for fruit production. The remaining 0.04 million hectares (1.7 %) were classified as irre-gular structures or without any indication. A further 0.31 million hectares are thought to be mixed forest with chestnut.Three types of chestnut countries can be distinguished: (i) coun-tries with a strong chestnut tradition (e.g. Italy, France, southern Switzerland, Spain, Portugal and Greece), where the chestnut stands are cultivated with intensive and characteristic silvicultural systems (coppices and orchards); (ii) countries with only a partially developed chestnut tradition due to the country’s particular geogra-phy (e.g. England) or history (e.g. Croatia, Turkey, Georgia); (iii) countries where the chestnut only sporadically occurs (e.g. Hungary, Bulgaria, Belgium) or has been recently introduced (e.g. Slovakia, Netherlands).A comparison of the present distribution of traditional silvicultural systems and historical data on chestnut distribution supports the hypothesis that the large-scale chestnut forest plantations are of post-Roman origin. Chestnut cultivation is now at a turning point as the changed needs of society have changed as it has moved away from a rural-based to an industrial and urban-oriented organization. The evolution of the chestnut market in recent decades confirms the potential of this resource for both traditional products and new services and goods related to organic-food and environmentally friendly products.

Key-wordsChestnut coppice, chestnut orchard, silvicultural systems, chestnut resources, Europe

RésuméIl n’existe aucune donnée officielle et cohérente sur la distribution du châtaignier en Europe, en dépit de sa vaste aire de répartition et de l’important rôle économique que cette espèce a joué dans de nombreux pays. En 1997, dans le cadre de l’action COST G4 “Multidisciplinary Chestnut Research”, des données quantitatives et qualitatives ont été collectées au sujet des châtaigneraies européennes, principalement à partir des inventaires forestiers nationaux. L’objectif était de fournir un état des lieux aussi précis que possible pour cette importance ressource économique. Un total de 2,25 millions d’hec-tares de boisements dominés par le châtaignier a été inventorié, avec 1,78 millions d’hectares (79,0 %) cultivé pour la production de bois et 0,43 million d’hectares (19,3 %) pour la production fruitière. Les 0,04 million d’hectares restants (1,7 %) ont été classés en structures irrégulières ou n’ont pas pu être classifiés. Plus de 0,31 million hec-tares sont classés en tant que boisements mixtes à châtiagnier. Trois types de pays peuvent être distingués du point de vue de la castanéiculture : (i) les pays dotés d’une forte tradition castanéicole (ex. Italie, France, sud de la Suisse, Espagne, Portugal et Grèce), où les châtaigneraies sont conduites en taillis ou futaies, grâce à des sylvicultures intensives et caractéristiques ; (ii) les pays où la tradi-tion castanéicole n’est que partiellement développée en raison de contraintes écologiques (ex. Angleterre) ou historique (ex. Croatie, Turquie, Géorgie) ; (iii) les pays où la castanéiculture n’est que sporadique (ex. Hongrie, Bulgarie, Belgique) ou a été récemment introduite (ex. Slovaquie, Pays-Bas).La confrontation de la distribution actuelle des pratiques tradition-nelles et des données historiques sur la répartition du châtaignier supporte l’hypothèse d’une origine post-romaine pour les peuplements de châtaignier présents sur de vastes étendus. La castanéiculure a été profondément bouleversée par les changements socio-économiques liés au passage d’une économie essentiellement rurale à une économique de type industriel. Toutefois, l’évolution du marché durant ces dernières décennies montre que le potentiel économique de cette ressource reste bien présent, à la fois pour des produits traditionnels mais aussi pour de nouveaux services et biens.

Mots-clésChâtaigneraies, systèmes sylvoculturaux, castanéiculture, Europe

Distribution and economic potential of the Sweet chestnut(Castanea sativa Mill.) in EuropeDistribution et potentiel économique du châtaignier(Castanea sativa Mill.) en Europe

M. Conedera1, M.C. Manetti2, F. Giudici1 & E. Amorini2

1. WSL Swiss Federal Institute for Forest, Snow and Landscape Research, Sottostazione Sud delle Alpi, CH 6504 Bellinzona, Switzerland.2. C.R.A. Forest Research Institute – Viale S. Margherita 80, I-52100 Arezzo, Italy

180

◆ M. CONEDERA, M. C. MANETTI, F. GIUDICI & E. AMORINI


INTRODUCTION

The Sweet chestnut (Castanea sativa Mill.) is the only native species of the genus in Europe. The main chest-nut refugia are though to be the Transcaucasian region, north-western Anatolia, the hinterland of the Tyrrhenian coast from Liguria to Lazio along the Apennine range, the region around Lago di Monticchio (Monte Vulture) in southern Italy, the Cantabrian coast on the Iberian penin-sula, and probably also the Greek peninsula (Peloponnese and Thessaly) and north-eastern Italy (Colli Euganei, Monti Berici, Emilia-Romagna) (Krebs et al., 2004).

Chestnut cultivation started early and has a long tra-dition in Europe, as it is a tree species that is suitable for both timber and fruit production. The first written evi-dence of chestnut management is found in Theophrastus’ “Inquiry into plants”, 3rd Century B.C. The Romans then introduced in most parts of Europe the idea of systemati-cally cultivating and using the chestnut tree and, in certain cases, the tree itself (Conedera et al., 2004). In Medieval Times, people in several parts of Europe became greatly interested in cultivating chestnut, mostly for fruit produc-tion. Chestnut cultivation was extended to the ecological limits of the species (Pitte, 1986).

Despite the historical and economic importance of the European chestnut, no official and coherent data exist on the distribution of the species. The only attempt to survey chestnut resources in Europe was made in the fifties within the framework of the Chestnut International Commission. The International Commission consisted of a group of experts from different chestnut countries (Spain, Portugal, France, Switzerland, Italy, Yugoslavia, Turkey, Greece) who were charged with the analysis of the effects on chestnut cultivation of the increasing abandonment of rural areas and the onset of pathologies such as Cryphonectria para-sitica and Phythophtora spp. Creating a chestnut distribu-tion map was listed as a priority (Groupe des Experts du Châtaignier 1951; Commission Internationale du Châtaignier 1953, 1955, 1958).

One of the most interesting products of the International Commission was the chestnut distribu-tion map of Europe and the related quantitative data provided by the country reports presented during the meetings of the Commission (Commission Internationale du Châtaignier 1958). Unfortunately, the data are quite rough, due to different silvicultural approaches and clas-sifications in chestnut stands throughout Europe (simple coppices, coppices with standards, high forests for timber production, high forests for fruit production, orchards, mixed stands). Moreover, several countries in the north-

ern and eastern part of Europe were not considered in the survey.

Starting in the 1960s, new data on chestnut distri-bution became available as national forest inventories were carried out in several countries. But there were still problems with the classification of the different chestnut forest types as the surveying methods among the coun-tries were only partially comparable. The search for accu-rate data on chestnut resources in Europe has intensified in recent decades, since the chestnut tree was recognised as a multipurpose species for landscape conservation in marginal areas. Responding to this need, some authors have tried to synthesise the existing data, providing an overview of chestnut resources in the main European countries (Bourgeois et al., 1991).

In 1997, within the framework of the COST action G4 “Multidisciplinary Chestnut Research” experts from all European countries where chestnut is present, were regrouped for the first time. During this action, quantita-tive and qualitative data on chestnut forests were collected in order to provide as sound a picture as possible of this important European resource. In this paper we describe the project and discuss the collected data.

MATERIAL AND METHODS

Data collection

The experts (data contributors, see table 1) of the countries involved in the COST G4 action were asked to provide data about the area covered by chestnut forests according to the typologies reported in table 2. The data were made available as a national report for each country in which basic statistics were linked to other additional information about silvicultural management, products and research activities. The information so obtained was then compiled together with local and international lite-rature on chestnut cultivation in Europe.

Defining chestnut silvicultural typologies

There are many different silvicultural systems applied in chestnut stands in European countries, which makes it difficult to define a strict typology silvicultural practices. In this work, we adopted a hierarchical classification (table 2), first separating the chestnut forests (stands consisting of more than 50 % of chestnut) from the mixed stands (ches-tnut less than 50 %). We then distinguished between stands

181

DISTRIBUTION AND ECONOMIC POTENTIAL OF THE SWEET CHESTNUT (CASTANEA SATIVA MILL.) IN EUROPE ◆


Country Data contributor Data sources

Albania Maxhun Dida and Caush Elezi Source not given

Andorra Sebastià Semene GuiltartNational Forest Inventory 1997IEA Centre de Biodiversitat 2003

Austria Eva Wilhelm Source not given

Azerbaijan Vagid Gadjiev Source not given

Belgium Hugues Lecomte and Klaartje van LoyFlemish Forest Inventory (1996-99)Walloons: source not given

Bosnia-Herzegovina Ahmet Lojo Source not given

Bulgaria Svetla Doncheva Forest National Service

Czech Republic Haltofová and Jankovsky (2004)

Croatia Sanja Novak Agbaba Croatian Forest Service

France Eric Sevrin National Forest Inventory (1999)

Germany Volker André Bouffi er Seeman et al. (2001)

Georgia Yuri Michailov Source not given

Greece Gregor Chatziphilippidis and Stephanos Diamandis National Forest Inventory (1992), Diamandis (2002)

Hungary Lazslo Radocz and Norbert Frank National inventories and others

Italy Maria Chiara ManettiNational Forest Inventory (1985)National Statistics Institute (1993)Regional Forest Inventories, Ad hoc questionnaires

Macedonia Sotirovski Kiril and Sumarski Fakultet Source not given

Netherlands Anne Oosterbaan Estimated by the referent

Portugal Afonso MartinsNational Forest Inventory (1998)National Statistics Institute (1987-1999)

Romania Valentin Bolea and Danut Chira Source not given

Russian Federation Michhail Pridnya, Gennadyi Solntsev and Yuri Michailov Source not given

Serbia-Monte Negro Glisic (1975)

Slovakia Milan Bolvansky and Ferdinand TokarLesprojekt (General Directorate of State Forest)Institute of Forest Ecology, Nitra

Slovenia Anita Solar, Dusan Jurc Jurc (2002)

Spain Juan Gaillardo Lancho and Santiago Lorenzo Pereira National Forest Inventory (1996)

Switzerland Fulvio Giudici National Forest Inventory (1985)

Turkey Necdet Guler, Ümit Serdar Agricultural statistics

United Kingdom Nigel Braden and Karen RussellNational Forest Inventory (1995-99)Forestry Commission Census (1979-82)

Chestnut forests (chestnut area) Stands with more than 50 % chestnut

Timber production Stands where wood production is prevalent

Coppices Simple coppices. Coppices with standards

High forests« Natural » stands. PlantationsStands converted into high forestsAbandoned stands (coppices, orchards) with the structure of high forests

Fruit production Stands where fruit production is prevalent

Orchards Stands with grafted trees (groves), including row plantations for fruit production

High forests « Natural » stands. Plantations

Irregular structure Stands without a codifi ed management

Mixed forests with chestnut Stands with less than 50 % chestnut

Table 2. Definition of the chestnut forest types.

Table 1. Participating countries and data contributors.

182



with defined productive purposes (timber or fruit) and those without a codified management (irregular structure). The stands for timber production were subsequently divi-ded into high forests and coppices and the stands for fruit production into orchards and high forests. Coppices are defined as pure chestnut forests regenerated from dormant or adventitious buds of the stump for wood production (mostly poles and firewood). Orchards (or groves, as some authors call them) are traditionally open stands composed of grafted chestnut trees (selected varieties) for fruit pro-duction with intercropping of cereals (silvo-arable system, usually wheat, oat or rye), hay or pasture (silvo-pastoral system). We classified as orchard also new plantations of chestnut trees of selected varieties for fruit production or row plantations of grafted chestnut trees. In contrast, high forest was defined as chestnut stands that originated directly from seedlings (i.e. without coppicing or grafting) and that were used for timber or fruit production.

Chestnut mapA GIS-based chestnut map was constructed using the

original data from the International Chestnut Commission (Commission Internationale du Châtaignier 1958). The experts were asked to update the map for their own coun-try or, for the countries with missing data in 1958, to pro-vide a distribution map indicating the geographic location (polygons) of existing chestnut forest types. Where no forest type was indicated, the default classification of high forest was assumed. After scanning, the original chestnut maps were georeferenced and the polygons digitalised by hand with the best possible accuracy.

RESULTS AND DISCUSSION

Chestnut areaThe reported chestnut area covers in total 2,53 million

hectares in Europe, of which 2,25 million hectares are chestnut forests, i.e. forests where chestnut is the domi-nant tree species and the remaining 0,31 million hectares are mixed forests with chestnut (table 3). The distribu-tion area ranges from southern Europe (e.g. Crete) to the North (southern England, Belgium) (fig. 1). The European chestnut forests are concentrated in just a few countries with a long tradition of chestnut cultivation (fig. 1). France and Italy together account for 79,3 % of the whole chestnut forest area, with the other traditional chestnut countries, Spain, Portugal and Switzerland, accounting for a further 9,7 %. The remaining 11,0 %

are dispersed in the other countries (table 3). For certain countries, such as Albania, Austria, the Czech Republic and Serbia-Montenegro, little information is available. We know roughly where chestnut forests may be found (e.g. Glisic, 1975; Haltofová & Jankovsky, 2004), but no further details about stand type, distribution and man-agement are provided. For the countries not included in table 3, we assume that chestnut does not occur. A comparison of the distribution map of current chestnut forests (fig. 1) with both the chestnut pollen map at the end of the Roman Period (ca. 570 AD, fig. 2a) and with that at the end of the Middle Ages (ca. 1460 AD, fig. 2b), shows that, especially in western Europe, the main chest-nut areas coincide with the chestnut stands created since the Middle Age, as already reported by several authors (Pitte, 1986; Fernandez de Ana Magán, 2002).

The proportion of chestnut stands with respect to the total forest area is usually very low (< 1-3 %). Exceptions to this general picture are Georgia (16,1 %) and the two main West-European chestnut countries, Italy (7,7 %) and France (6,6 %). In some cases, the irregular distribu-tion pattern of chestnut stands implies a contrast between the low percentages at the national level and the high concentration of the species at the regional level. This is, for instance, the case for Switzerland where the chestnut percentage is 1,2 % nationally and 12,9 % in the region south of the Alps, the main chestnut-growing area (EAFV 1988). In most countries, chestnut forests are cultivated along the mountain ranges, highlighting the orophilous (mountainous) character of the species.

Unfortunately, it is not possible to analyze reliably the evolution of the total chestnut area for the countries included in the FAO survey of the fifties as the methods of data collection are not comparable. The FAO-survey considered only the managed and productive chestnut areas (Commission internationale du châtaignier, 1958), a category not considered in the present study.

Chestnut forests for timber production

In Europe, chestnut-growing area devoted to timber production is 1,78 million hectares (table 3), correspond-ing to 79,0 % of the total chestnut-growing area. The fur-ther division of these stands into coppices and high forests is in some cases not clear, because of the problematic classification of the coppices with standards. They tend to be classified as coppices or high forests, depending on the different weight given to the reserve-trees. In this study, we accepted the classifications proposed by the data contributors.

183



Fig. 1. Present distribution of the chestnut in Europe.

The traditional silvicultural method of timber pro-duction, the “coppice system”, is still widely applied in countries where chestnut cultivation was widespread in the past to satisfy the needs of rural populations located in marginal or mountainous areas and where the chestnut found suitable climate and soil conditions in Italy, France, Spain, Greece and Southern Switzerland (fig. 3 and 4; table 3).

There are several reasons for the large number of coppice stands. They used to be popular because small and medium traditional chestnut products (e.g. poles) were important in the national and local economies. In addition, chestnut stools have a high resprouting capac-ity, and the shoots grows remarkably quickly (Manetti et al., 2001). Moreover, many coppices in France, Italy, Spain and Switzerland were originally ancient orchards,

that were abandoned at the beginning of the last century and then coppiced because of the high incidence of the chestnut blight and of the demand for chestnut wood for tannin, mining and – especially in Spain – for barrels (Pitte, 1986; Conedera et al., 1997; Fernandez de Ana Magán, 2002). In these areas, there are now a number of atypical coppice stands, derived from orchards and characterized by very low stool density and poor quality shoots. According to Tani et al. (2003), two or three gen-erations of coppicing are needed in these cases, to regain a satisfactory stool density in the stands.

Since the late fifties, the silvicultural rules applied to coppice stands required a rotation period from 5 to 12 years (simple coppicing), or 25-30 years in cases with 2-3 spacings and thinnings, depending on the desired prod-ucts and the type of ownership. Changes in the socio-

184



coun

try

Tota

l co

untr

y ar

ea 1)

Tota

lfo

rest

area

1)

CH

ESTN

UT

FOR

ESTS

(che

stnu

t > 5

0%)

MIX

ED

FOR

ESTS

(c

hest

nut <

50

%)

rem

arks

TIM

BER

PRO

DU

CT

ION

FRU

IT P

ROD

UC

TIO

NIr

regu

lar

struc

ture

CH

ESTN

UT

AR

EA

ches

tnut

fo

rest

s vs

tota

l for

est

area

Cop

pice

sH

igh

fore

stsTO

TAL

Orc

hard

sH

igh

fore

stsTO

TAL

km2

km2

haha

haha

haha

haha

%%

ha

Alb

ania

2740

991

8600

8600

8600

0.4

0.9

1800

And

orra

4539

00.

00.

0sp

orad

ic p

rese

nce

of c

hestn

ut in

fore

st sta

nds

Aus

tria

8273

9402

miss

ing

valu

e

Aze

rbai

jan83

5910

9415

0015

0015

000.

10.

170

0hi

gh fo

rests

mos

tly fo

r fru

it pr

oduc

tion

Belg

ium

3025

620

750

300

1050

1050

0.0

0.2

4450

in m

ixed

fore

sts c

hestn

ut c

over

s abo

ut 3

0% o

f the

ba

sal a

rea.

Bosn

ia-

Her

zego

vina

5100

2273

3057

3057

3057

0.1

0.1

4585

Bulg

aria

1105

536

9020

0010

021

0072

072

014

029

600.

10.

148

0di

strib

utio

n be

twee

n co

ppic

e an

d hi

gh fo

rest

estim

ated

.

Cro

atia

5592

1783

1458

042

015

000

1500

00.

70.

822

362

Cze

ch

Rep

ublic

7728

2632

ches

tnut

pre

sent

as s

ingl

e tre

e or

tree

coh

orts

in m

ore

than

300

loca

lities

. No

quan

titat

ive

data

avai

lable

Fran

ce55

010

1534

186

2500

5800

092

0500

1000

0010

0000

1020

500

45.3

6.7

copp

ice

inclu

des c

oppi

ce w

ith st

anda

rds

Geo

rgia

6831

298

4800

048

000

4800

02.

116

.1da

ta so

urce

unc

erta

in

Ger

man

y34

927

1074

044

0044

0044

000.

20.

016

00on

ly to

tal a

rea

avai

lable,

inclu

ding

aven

ue a

nd so

litar

y tre

es; c

oppi

ce st

ands

are

ass

umed

to b

e pr

evale

nt

Gre

ece

1289

035

9933

051

3305

160

060

033

651

1.5

0.9

Hun

gary

9234

1840

300

800

1100

900

900

2000

0.1

0.1

650

Italy

2940

610

003

4827

5115

119

4978

7023

5620

2356

2032

347

7658

3734

.07.

7irr

egul

ar st

ruct

ure

inte

nded

as f

ores

t with

out a

co

difi e

d m

anag

emen

t

Mac

edon

ia25

4390

60

5058

5058

0.2

0.6

irreg

ular

stru

ctur

e in

tend

ed a

s for

est w

ithou

t a

codi

fi ed

man

agem

ent

Net

herla

nds

3392

375

5050

500.

00.

025

0

Portu

gal

9150

3666

3390

033

900

1960

919

609

5350

92.

41.

521

400

copp

ice

are

inclu

ded

in th

e hi

gh fo

rest

area

Rom

ania

2303

464

4828

9028

9010

010

029

900.

10.

020

0

Rus

sian

Fede

ratio

n16

8885

185

1392

4000

040

000

040

000

1.8

0.0

7000

Serb

ia-M

onte

N

egro

1020

028

87ch

estn

ut p

rese

nts i

n fe

w sp

ots.

No

quan

titat

ive

data

av

ailab

le

Slov

akia

4808

2177

1613

0213

1892

9295

1505

0.1

0.1

45

Slov

enia

2012

1107

3000

030

000

185

185

3018

51.

32.

720

2308

no d

etai

led in

form

atio

n on

the

spec

ial d

istrib

utio

n of

th

e di

ffere

nt si

lvic

ultu

ral s

yste

ms

Spai

n49

945

1437

049

909

5003

999

948

3767

937

679

1376

276.

11.

0co

nfus

ion

betw

een

high

fore

sts a

nd o

rcha

rds m

ay

exist

; sin

gle

or d

isper

sed

orch

ard

trees

are

not

in

clude

d

Switz

erlan

d39

5511

9919

000

4700

2370

034

0034

0027

100

1.2

2.3

6800

mix

ed fo

rests

inte

nded

as c

hestn

ut p

rese

nce

less t

hen

25%

of t

he b

asal

area

T urk

ey76

963

1022

536

1436

1425

278

2527

828

892

1.3

0.3

2800

0co

nfus

ion

betw

een

high

fore

sts a

nd o

rcha

rds m

ay

exist

Uni

ted

Kin

gdom

2416

027

9479

1310

875

1878

818

788

0.8

0.7

1087

1iso

lated

woo

ds <

2 h

a ar

e no

t inc

lude

d

TOTA

L20

9922

896

1891

1483

891

2964

4517

8033

639

7585

3669

843

4283

3764

022

5225

910

0.0

3135

01

% o

f che

stnut

area

65.9

13.2

79.0

17.7

1.6

19.3

1.7

100.

0

185



economical structure of the rural areas and the crisis in the market for poles meant that different silvicultural treatments were needed to satisfy the new demand for larger size and higher quality products.

As a results of these developments, the silvicultural systems for chestnut timber production applied are now very diverse, both within and among chestnut-growing countries. In certain cases, the rotation period has been extended without any planned active silvicultural inter-vention and no well-defined long- or medium-term objec-tives. In some countries such as Italy, Spain and the United Kingdom, coppices have been converted into high forests or, in extreme cases, have just not been cut and have thus developed into high forest-like stands. Nevertheless, in some European countries, such as France, Greece, Italy, Portugal, Spain, Southern Switzerland and the United Kingdom, private chestnut coppices are still cultivated in a short rotation period (10-20 years) without thin-nings (rarely 1, early and from below) to produce poles and small products, mainly in the private ownership. The application of long-rotation periods (30-60 years) and selective thinnings, however, are reported for Croatia, Bulgaria and Germany. Finally, there have been recent experiments in France, Italy and Spain in the face of the growing demand for chestnut wood products. Here the management policy has been to have rotation periods of 25-35 years and 1-2 thinnings from below), in order to produce timber of higher quality, especially with a low incidence of ring shake (table 4).

High forest stands, representing only 16,7 % (= ap-prox. 296 500 ha) of the chestnut-growing area devoted to timber production, are prevalent in those countries that have recently introduced chestnut cultivation (Slovakia, Hungary), and in those where there is no specific chest-nut silviculture (Romania, Russia), or where the adopted silvicultural practices early modified the forest structures, giving them the appearance of high forest-like stands (Portugal, Romania).

In mixed forests, the silvicultural treatment is generally uniform and the chestnut is managed in the same way as other species. In such stands, management generally fol-lows a rotation period of 40-60 years, or, more rarely, of 80-100 years or more (Slovakia, Bulgaria and Romania), and thinnings are carried out every 5-10 or 10-20 years.

The thinning type varies from selective to low thinning, probably according to the silvicultural tradition of the country concerned (table 4).

In summary, there appears to be variety in the silvi-cultural treatments (rotation period, type, frequency and intensity of thinnings) applied to chestnut coppice and high forest stands in Europe. Above all, it seems silvicul-ture interventions are often applied without an exhaustive analysis of their ecological impact on the forest ecosys-tem, and without a preventive analysis of the relevant ecological parameters (which include site index, climatic characteristics, and soil type) or of the main structural parameters (e.g. silvicultural system, stand density, domi-nant height, degree of mixture).

Chestnut forests for fruit production

The chestnut-growing area devoted to fruit produc-tion covers 0,43 million hectares (table 3), corresponding to 19,3 % of the total chestnut-growing area. According to our survey, Italy and France are nowadays the countries with the largest orchard areas (together they account for 84,5 % of the total reported orchard area). Thus, simi-lar to the coppice stands, orchards grown according to traditional silvicultural methods for fruit production are mostly found in countries with a long tradition of chest-nut cultivation (fig. 5; table 3). During the Middle Ages, the chestnut was an essential source of food for many mountain regions of those countries and this resulted in a wide range of products and of cultivated varieties with different ripening periods (early, mid-season, late), types of use (fresh consumption, long-term storage, drying, flour, animal feed) and ranges of distribution (higher altitudes, lower slopes, ubiquitous, etc.). This gave rise to a very complex and highly structured chestnut culture with a considerable number of different chestnut varieties cultivated for different purposes (Conedera et al., 1993).

The few regions where commercial transport routes already existed during the Middle Ages (e.g. Piedmont or Tuscany), where the only places where chestnut plantations with just one or few high-quality chestnut or marron-cultivars were started in this period. The fruits were then sold on the regional or even international

Table 3. Chestnut forest area according to defined forest types in Europe..

1 FAO, 2003: State of the World's Forests (SOFO), Rome, XIV, 151 p. Land area refers to the total area, excluding areas under inland water bodies. The source of these data is FAO (2001); they may differ slightly from those in the State of the World's Forests 2001, which used a different source. The forest cover fi gure for each country has been calibrated to the country's land area.

186



Fig. 2a. Chestnut pollen map 570 AD (source Conedera et al., 2004).

markets (Pitte, 1986). The marron varieties, then defined as elliptic-shaped fruits of medium to large size, with marked dark strips on the tegument, which are light to peel (no intrusion of the epispermatic pellicle in the cotyledons) and sweet in taste (Bassi, 1993), have traditionally been cultivated only in Italy and a few areas in France. In all other countries marron-cultivars do not exist or are of recent import (e.g. southern Switzerland, Conedera et al., 1997). In France, the definition of marrons has been recently revised in order to better fit the market and com-mercial needs. According to Bergougnoux et al. (1978), marron varieties should not display more than 12 % fruit with end-to-end epispermatic intrusions.

In some regions of the Iberian Peninsula (northern Portugal and Galicia), double-purpose varieties (fruit and timber production) are quite common. Here the chestnut trees are topped above the grafting point, usually 2 meters

above ground level (Bourgeois et al., 2004). This complex historical background makes it impossible to reliably esti-mate the number of chestnut cultivars or ecotypes exist-ing in Europe, even if the inventories so far performed at the national level suggest that there are thousands more varieties (Pitte, 1986).

In chestnut-growing areas cultivated as orchards, the quantity of chestnut produced in the main chestnut-growing countries has dramatically decreased since the beginnig of the 20ies Century although it begin to pick up again towards the end of the heigties (fig. 6). This development is mainly due to the progressive depopula-tion of the countryside, the abandonment of chestnut as a staple food, the introduction and spread of ink disease and chestnut blight, and the increased demand for wood for tannin extracts (Pitte, 1986; Bounous & De Guarda, 2002). Chestnut fruit cultivation has survived or quickly

187



Fig. 2b. Chestnut pollen map 1460 AD (source Conedera et al., 2004).

recovered best where the high-quality marron and chest-nut cultivars are cultivated (Italy and, to a lesser extent, France) or where the large-size chestnut varieties are grown (Spain, Portugal, Turkey and France) (Alvisi, 1994). In the Iberian Peninsula, intercropping with cereal is becoming rare, but the practice of soil tillage is used to increase nut production (Berrocal del Brio et al., 1998; Portela et al., 1999).

Since 1990s, people have become more aware of the value of chestnut orchards as a multifunctional landscape element. In many countries they have begun to revitalize chestnut orchards as they see them as having aesthetic and ecological value, acting as tourist attractions, and serv-ing as fire-breaks (Bounous et al., 1992, Conedera et al., 1997). Besides the revitalisation of traditional orchards in marginal chestnut-growing areas, new plantations (or even re-grafted old orchards) with high-quality varieties (marrons and similar) or large-size Euro-Japanese varie-

ties have been cultivated in several countries (mostly in France, Spain and Italy).

High forests for fruit production are rare (covering approx. 36 700 ha in Europe, which corresponds to 8,5 % of the fruit production area). Most of them are located in Turkey. According to Soylu et al. (2002), the grafting of chestnut trees is not very common in the Black Sea region. In some unclear cases, such as the attribution to the “soto”-type (= orchard) of some fruit-producing stands in Spain, the corresponding area was classified as “orchard”.

Chestnut products

The increasing demand for natural and environmentally friendly products in Europe has led to more interest in the chestnut. Moreover, some recent technological improvements (laminated veneer boards, finger-jointed

188



beams and boards, thick sliced veneer, better use of industrial wood through joint production of tannin and panels, fruit conservation and processing techniques) have also positive driving forces influencing the chestnut market (Pettenella, 2001). As a consequence, traditional chestnut products have now more opportunities on the market (poles for land consolidation work or playgrounds, logs for flooring, chestnut flour for pasta production, certification of local cultivars, etc.) and new products (chestnut parquet, chestnut-laminated veneer boards, chestnut pasta, chestnut beer, etc.) have been launched. Some of these new products and new applications, such as finger-jointed beams, and shingles from the wood, and pasta, biscuits, beer from the fruit are particularly interesting because it is possible to produce them from small-sized chestnut timber or fruit. In addition, the aesthetic, cultural and ecological value of managed chestnut ecosystems is now much more recognized. Restoring chestnut growing areas is also valued for its role in preserving landscape and a country’s traditional heritage (Conedera et al., 1997, Bounous et al., 2001). In conclusion, chestnut cultivation today provides an

exemplary model of multifunctional forestry (figure 7), playing an important social and economic role in rural areas.

The recent development of the chestnut market has not been homogeneous. It is frequently affected by agri-cultural, economic, and social and cultural local factors (Hennion & Vernin, 2000). Pettenella (2001) provides examples of niche chestnut markets, but the evidence of these is largely anecdotal and there are no figures avail-able. For example, chestnut sawnwood for solid wood furniture production is in great demand in Tuscany (Italy), logs for floorings are highly desiderable in France, poles for land consolidation, torrent and avalanche con-trol works are used in Switzerland, the production of chestnut-laminated beams and panels is increasing in north-east Italy, chestnut flour has an expanding market in Bosnia, and so on.

A precise quantification of the total amount of chestnut timber and fruit production is difficult because of both missing information in the official statistics and the existence of an unregistered parallel market (direct sales, self-consumption, fruits left on the ground by

Timber production

Coppices High Forests

rotation (years) thinnings (number) type of thinning rotation (years) thinnings (number) type of thinnings

Belgium No information Generally in mixed stands

Bosnia-Herzegovina Clear cutting when blighted No high forests

Bulgaria No information

Croatia 40-60 yes selective Generally in mixed stands

France 25-35 1 (9-12 yrs) low 40-60 2-3 (every 10-15 yrs) selective

Germany 20-30 No information 60-80 No information

Greece 20-25 2 (7, 14 yrs) low No high forests

Hungary No coppice stands No information

Italy 12-18 / 25-30 no / 1-2 low 50-60 2-3 (every 10-15 yrs) low

Netherlands Generally in mixed forests No high forests

Portugal 15 1 (at 2 yrs) low 40-45 5 (every 8-10 yrs) low

Romania No coppice stands 120 every 5-10 yrs low

Russian Federation No coppice stands Generally in mixed forests

Slovakia No coppice stands 100-120 every 10-15 yrs selective

Slovenia No information

Spain 5-8 // 18-25 no // 2 (5, 12) low 40-60 2-3 (every 10-15 yrs) selective

Switzerland 12-18 None 30-60 ?? selective

Turkey No information No high forests

United Kingdom 12-20 None Generally in mixed forests

Table 4. Main characteristics of the silvicultural management applied in the different countries for chestnut timber production.

189



Fig. 3. Present distribution of chestnut coppices in Europe.

Fruit Timber

Country chestnuts (tons) Sawnwood (m3) Poles (m3) Tannin and industry (m3)

Croatia 2,000 1,000 20,000

France 13,000 115,000 310,000 500,000

Greece 12,000

Germany 1,000 5,000

Italy 52,000 54,000 300,000 150,000

Portugal 33,000

Slovenia 16,000

Spain 10,000 8,000 12,000

Turkey 60,000 140,000

Switzerland 10,000 20,000

Other European countries 154,000

Europe 351,000

Tab. 5. Chestnut products in Europe (reference year, 2000). Source: FAO (FAOSTAT Agriculture Data - Agricultural Production - Crops Primary; http://apps.fao.org/) and Data contributors of the Cost Action G4 (see Table 1).

190



Fig. 4. Percentage distribution of the two different timber-

producing silvicultural systems coppices (grey) and high forests

(black) in the surveyed countries.

Fig. 5. Present distribution of chestnut orchards in Europe.

191



livestock, etc.). For fruit, the official statistics provided by the FAO indicate that around 350 000 tons of chestnuts are produced a year all over Europe (table 5). No official statistics exist for chestnut timber production, and only a few countries were able to provide data within the framework of the G4 Cost Action (table 5). Although incomplete, the data clearly show the poor amount of chestnut timber products of high value (i.e. sawnwood). This is probably due to the general lack of tradition in applying silvicultural treatments to improve the wood quality of the wood in the chestnut coppices.

CONCLUSIONS

Chestnut cultivation has a long tradition and deep roots in many European countries. The European countries with chestnut histories and tradition, can be divided into three main categories: (i) countries with a strong chestnut tradition (e.g. Italy, France, southern Switzerland, Spain, Portugal and Greece), where chestnut stands have been cultivated since centuries with intensive and character-istic silvicultural methods (coppice and orchards); (ii) countries with a only partially developed chestnut tra-

dition due to their particular geography (e.g. England) or history (e.g. Slovenia, Croatia, Turkey, Georgia); (iii) countries where the chestnut only occurs sporadically (e.g. Hungary, Bulgaria, Belgium) or has been recently introduced (e.g. Slovakia, Netherlands).

Despite this historical background, chestnut cultivation is now at a turning point and is being confronted with changing needs of a society that has moved from being rural to becoming industrial and urban-oriented. The development of the chestnut market in recent decades confirms the potential of this resource for both traditional products and new services and goods related to organic food and environmentally friendly products. This is particularly important for the chestnut as it is widespread in the territory particularly cut off from the main industrial developments. Silvicultural management trends are already reacting to these developments. The cultivation of high-value chestnut products (fruit or timber) is being intensified in the best chestnut-growing areas, new plantations are being created in potentially good sites for the chestnut, and traditional old orchards are being restored as part of a multipurpose landscape. At the same time, chestnut cultivation is being abandoned in the very marginal areas, where the old chestnut stands will evolve into mixed stands.

Fig. 6. Evolution of chestnut fruit production (tons) in different European countries since 1960. Source: FAO (FAOSTAT Agriculture Data - Agricultural Production - Crops Primary; http://apps.fao.org/) and Commission Internationale du châtaignier (1958).

192



Given this dynamic situation, new marketing instruments will have to be developed, such as the certifying and registering the place of origin and the system of cultivation (Pettenella, 2001). More research is also needed to develop sound silvicultural techniques to solve problems related to the correct management of chestnut stands to best take into account not only pro-ductive aspects, but also the historical value of the for-ests, the establishment of multifunctional stands, and the improvement of the ecological and environmental value of the landscape.

Acknowledgements

Our heartfelt thanks go to the data contributors for providing us with the national information, to our collea-gues François Romane and Patrick Fonti for the critical reading of the manuscript, to Florian Boller, Damiano Torriani and Daniela Furrer for the digitalisation of the chestnut maps, and to Silvia Dingwall for the English revision of the text.

ReferencesALVISI F., 1994. Aspetti economici e commerciali della casta-

nicoltura italiana. Rivista di frutticoltura e di ortofloricoltura, 56, 2: 41-47.

BASSI D., 1993. Castagno da frutto: valorizziamo la qualità. Rivista di frutticoltura, 55, 12: 39-41.

BERGOUGNOUX F., VERLHAC A., BREISCH H. & CHAPA J. 1998. Le châtaignier. Production et culture. Invuflec Malemort, 192 p.

BERROCAL DEL BRIO M., GALLARDO LANCHO J.F. & CARDEÑOSO HERRERO J.M., 1998. El castaño. Ediciones Mundi-Prensa, Madrid, 288 p.

BOUNOUS G., BOUCHET M. & GOURDON L., 1992. Ricostituzione del castagneto a frutto tradizionale: interventi in Piemonte e nel Sud della Francia. L’Informatore Agrario, 9: 155-160.

BOUNOUS G., BAGNARESI U., BELLINI E. & BECCARO G., 2001. Aspetti paesaggistici e culturali del castagno: problemati-che di tutela e valorizzazione. In: Bellini E. (ed.), Atti del “Convegno Nazionale Castagno 2001”, Marradi (Firenze) 25-27 ottobre 2001. 365-372.

BOUNOUS G. & DE GUARDA A., 2002. Origine e cenni storici. In: Bounous G. (ed.) Il castagno. Coltura, ambiente ed utilizzazioni in Italia e nel mondo. Bologna, Edagricole. 3-17.

BOURGEOIS C., LOPEZ J. & GAROLERA E., 1991. Le châtaignier en Europe. Forêt-Entreprise 76, 25-42.

BOURGEOIS C., SEVRIN E. & LEMAIRE J., 2004. Le châtaignier: un arbre, un bois. 2nd revised edition, Institut pour le développe-ment forestier, Paris, 347 p.

COMMISSION INTERNATIONALE DU CHÂTAIGNIER, 1953. Rapport de la 2e session. FAO 54/7/4069, Rome, 62 p.

— 1955. Rapport de la 3e session. FAO CH/19, Rome, 62 p.

— 1958. Rapport de la 4e session. FAO CH/26, Rome, 57 p.

CONEDERA M., MÜLLER-STARCK G. & FINESCHI S., 1993. Genetic characterization of cultivated varieties of European Chestnut (Castanea sativa Mill.) in Southern Switzerland. (I) Inventory of chestnut varieties: history and perspectives. In: Comunità Montana Monti Martani e Serano of Spoleto; Istituto di Coltivazioni Arboree University of Perugia (eds.) Proceedings of the International Congress on Chestnut Spoleto, Italy, October 20-23, 1993: 299-302.

CONEDERA M., JERMINI M. & SASSELLA A., 1997. Nouvelles perspectives pour la culture du châtaignier au sud des Alpes. Rev. suisse vitic. arboric. hortic., 29, 6: 337-344.

Fig. 7. Products and services provided by chestnut forests classified according

to the management system (modified from Pettenella 2001).

193



CONEDERA M., KREBS P., TINNER W., PRADELLA M. & TORRIANI D., 2004. The cultivation of Castanea sativa (Mill.) in Europe: from its origin to its diffusion on a continental scale. Vegetation History and Archaeobotany, 13: 161-179.

DIAMANDIS S., 2002. Grecia. In: Bounous G. (ed.) Il castagno. Coltura, ambiente ed utilizzazioni in Italia e nel mondo. Bologna, Edagricole. 253-260.

EAFV (EIDGENÖSSISCHE FORSCHUNGSANSTALT FÜR DAS FORSTWESEN), 1988. Schweizerische Landesforstinventar. Ergebnisse der Erstaufnahme 1982-1986. EAFV Bericht 305, 1-375.

FERNANDEZ DE ANA MAGÁN F.J., 2002. Spagna. In: Bounous G. (ed.) Il castagno. Coltura, ambiente ed utilizzazioni in Italia e nel mondo. Bologna, Edagricole. 265-273.

GLISIC M., 1975. Le châtaignier (Castanea sativa Mill.) en Serbie et sa variabilité biologique et écologique. PhD Thesis, Institut de sylviculture et d’Insdustrie du bois, Belgrade, 211 p. (in Serbian).

GROUPE DES EXPERTS DU CHÂTAIGNIER, 1951. Rapport de la première session. FAO 53/2/1423, Rome, 50 p.

HALTOFOVÁ P. & JANKOVSKY L., 2004. Distribution and healt condition of sweet chestnut (Castanea sativa Mill.) in the Czech Republic. In: IIIe International Symposium on Chestnut, Chaves, 20-23 October 2004, Abstracts, 60.

HENNION B. & VERNIN X., 2000. Châtaigne. Forces et faiblesses de la filière. Infos Ctifl, 166, 14-17.

JURC D., 2002. An overview of the history of the chestnut blight epidemic in Slovenia. Research reports forestry and wood science and technology, 68: 33-59.

KREBS P., CONEDERA M., PRADELLA M., TORRIANI D., FELBER M. & TINNER W., 2004. Quaternary refugia of the sweet chestnut (Castanea sativa Mill.): an extended palynological approach. Vegetation History and Archaeobotany, 13, 145-160.

MANETTI M.C., AMORINI E., BECAGLI C., CONEDERA M. & GIUDICI F., 2001. Productive potential of chestnut (Castanea sativa Mill.) stands in Europe. For. Snow Landsc. Res. 76, 3: 471-476.

PETTENELLA D., 2001. Marketing perspectives and instruments for chestnut products and services. For. Snow Landsc. Res. 76, 3: 511-517.

PITTE J.R., 1986. Terres de castanide. Hommes et paysages du châ-taignier de l’Antiquité à nos jours. Librairie Arthème Fayard, Paris, 479 p.

PORTELA E., ARAHNA J., MARTINS A. & PIRES A.L., 1999. Soil factors, farmer’s practices and chestnut ink disease: some interactions. In: Proceedings of the Second International Symposium on Chestnut (Ed. Salesses G.), Acta Horticulturae 494, 433-441.

SEEMANN P., BOUFFIER V., KEHR R., WULF A., SCHRÖDER T. & UNGER J., 2001. Die Esskastanie (Castanea sativa Mill.) in Deutschland und ihre Gefährdung durch den Kastanienrindenkrebs (Cryphonectria parasitica [Murr.] Barr.). Nachrichtenbl. Deut. Pflanzenschutzd., 53: 49-60.

SOYLU A., SERDAR Ü., ERTÜRK Ü., 2002. Chestnut growing in Turkey and researches on its problems. FAO-CIHEAM Nucis-Newsletter, 11: 34-38.

TANI A., MALTONI A. & MARIOTTI B., 2003. La produzione legnosa di castagno in Italia. Situazione attuale e prospettive. Sherwood, 92, 5-9.

Documents

Revue internationale d’écologie méditerranéenne ... · In this study the effects of wood-pasturage on species composition and forest structure in the Quercus frainetto forest