AMQ28(2)131-143 TOTI rivisto 12_2015 edi.pub Available online http://amq.aiqua.it ISSN (print): 2279-7327, ISSN (online): 2279-7335 Alpine and Mediterranean Quaternary, 28 (2), 2015, 131 - 143 INTERGLACIAL VEGETATION PATTERNS AT THE EARLY-MIDDLE PLEISTOCENE TRANSITION: A POINT OF VIEW FROM THE MONTALBANO JONICO SECTION (SOUTHERN ITALY) Francesco Toti Department of Earth Sciences, University of Florence, Florence, Italy Corresponding author: F. Toti ABSTRACT: The Montalbano Jonico succession (southern Italy) represents a rich stratigraphic archive of the Early to Middle Pleistocene environmental and climatic changes in the central Mediterranean. Pollen analyses associated to multivariate statistical methods (principal component and cluster analyses) have been carried out in the sedimentary portion including Marine Isotope Stages (MIS) 21 to 17 (~200 kyrs and ~190 m). The pollen ratio between mesotherm and steppic taxa (Pollen Temperature Index), used as a proxy of temperature, permits to precise consecutive warmer and cooler phases associated, respectively, to interglacials/interstadials and glacials/stadials (sub- stages 21.3 to 18.3 plus stage 17). Pollen data suggest major expansions of mixed oak forests during warm phases, whereas open vegeta- tion typically marks cool to cold phases. The statistical processing puts in evidence both analogies and differences among the successive interglacials/interstadials. Mid- to high-altitude trees and Mediterranean taxa appear to expand slower than hydrophytes plus deciduous oaks and other herbs, which conversely act like pioneers. Moreover Tsuga, Cedrus plus steppic plants and Poaceae are inversely related to Mediterranean xerophytes and mesotherm deciduous taxa, possibly due to the different tolerance to temperature decreases. On this basis, three main vegetation patterns seem to occur during warm periods. The 1st pattern (including MIS 21.3, 19.3 and 17) shows a tripar- tite structure with a middle warm phase including higher abundances of slow-spreading taxa, sandwiched between cooler intervals (at the beginning and the end of the sub-stage) with the increase of faster-expanding taxa. The 2nd pattern (in MIS 21.1 and 19.1) shows a com- plex structure characterized by rapid changes in proportions between fast- and slow-spreading taxa as well as among plants with different tolerance to temperature decrease. The 3rd pattern (in two interstadials within MIS 18), is characterized by relatively high abundances of fast-spreading taxa and low abundances of warmth-demanding taxa. This work represents a first attempt for a better understanding of interglacial vegetation dynamics during the Early-Middle Pleistocene Transition; deeper investigations are in progress with the specific aim to better characterize the observed multiple patterns and possibly to define their causes. KEYWORDS: Early-Middle Pleistocene Transition, interglacials, MIS 19, Montalbano Jonico, southern Italy, principal component analysis, pollen analysis. 1. INTRODUCTION During the transition between Early and Middle Pleistocene (EMPT, 1.2-0.5 ka; e.g. Head & Gibbard, 2005; Maslin & Ridgwell, 2005), Earth’s climate experi- enced a gradual change in the Glacial-Interglacial (G-I) oscillations mode. The orbital obliquity-driven 41 kyrs cyclicity, which had dominated the earlier part of the Pleistocene, was overshadowed by a high-amplitude, lower-frequency 100 kyrs-rhythm. A profound impact on terrestrial and marine biota is recorded across the EMPT as a result of changes in the intensity and dura- tion of both glacials and interglacials (e.g. Head et al., 2008). In the Italian pollen and calcareous nannoplank- ton records a major shift seems to coincide with the Marine Isotope Stages (MIS) 25 to 20, at the end of the Jaramillo sub-chron, where several turnovers in the nannofossil communities and the progressive disap- pearance of the most thermophilous pollen taxa oc- curred (Bertini et al., 2010 and references therein). On the other hand, no remarkable events have been re- corded at the Calabrian-Ionian transition (e.g. Head & Gibbard, 2005; Bertini et al., 2015). Quaternary Euro- pean and Mediterranean pollen records display recur- rent patterns of vegetation succession paralleling G-I cyclicity. In Italy, four main patterns have been summa- rized on the basis of latitudinal, geomorphological and local factors by Bertini (2010). Pattern 1, is expressed by open vegetation-thermophilous forest alternations especially in Mediterranean littoral zones (e.g. Suc et al., 1995). Pattern 2, which is typical of several Northern Apennines sites, exhibits prevalent alternations between altitudinal coniferous and thermophilous forests (e.g. Bertini 2001 and references therein). Patterns 3 and 4 show the presence of thermophilous and coniferous forest, respectively, during glacials, and open vegetation and steppe, respectively, during interglacials (e.g. Ber- toldi et al., 1989; Capraro et al., 2005). The interglacial structure, its duration and intensity, as well the season- ality, strongly depend on the Earth orbital geometry, which is a major factor of climate forcing (e.g. Hays et al., 1976; Berger, 1981; 1988). According to many au- thors, the vegetation succession of each interglacial is “unique” because of the peculiar combination of its astro -climatic variables (e.g. Watts, 1988; Bartlein & Prentice, 1989; Huntley & Webb, 1989). Moreover, even where some interglacials exhibit virtually identical orbital con- figurations, the occurrence of stochastic factors (e.g. relationships among taxa and/or between taxa and cli- mate) can make the system unpredictable (e.g. Bennett et al., 1991; Tzedakis et al., 2012a). At the same time, the existence of interglacials characterized by very dif- ferent climate conditions but showing similar vegetation patterns cannot be excluded (Tzedakis & Bennett, 1995); in this case, internal biotic forcings should be invoked. One of the longest composite Quater- nary pollen record, including subsequent interglacials since the base of the Ge- lasian, has been reconstructed by the study of both marine and continental suc- cessions of southern Italy (Fig. 1) (e.g. Combourieu-Nebout et al., 1990; Com- bourieu Nebout and Vergnaud Grazzini, 1991; Combourieu-Nebout, 1993; Capraro et al., 2005; Klotz et al., 2006; Joannin et al., 2008; Russo Ermolli et al., 2010a, b, 2014; Suc et al., 2010; Amato et al., 2014; Petrosino et al., 2014; Robustelli et al., 2014; Bertini et al., 2015; Marino et al., 2015). The knowledge of past interglacials is particularly significant not only because it allows the documentation of ancient pa- leoenvironments but also for its key role in modeling the natural course of the present interglacial (Holocene), permitting predic- tions about the future climate changes. For this study a selected portion of the Montal- bano Jonico (MJ) section, spanning the MIS 21-16, has been submitted to detailed palynological analyses which have strongly improved those recently produced in Ber- tini et al. (2015), and extend the previous pollen documentation concerning MIS 37- 23 (Joannin et al., 2008). The main purpose of this paper is to recognize similarities and differences in the vegetation dynamics between subsequent warm periods (either interstadials or inter- glacials), assuming that each one is marked by a peculiar astronomical and vegetation signature. To achieve this aim, warmer phases have been singled out and compared using a multivariate statistical approach. The significance of the MJ pol- len record is enhanced by the possibility to exclude the presence of consistent strati- graphic gaps thanks to the strong strati- graphical frame established for this marine section (e.g. Marino et al., 2015 and refer- ences therein). 2. CASE STUDY: THE MONTALBANO JONICO SEC- TION 2.1. Geological and stratigraphical setting MJ composite succession is located in the Basili- cata region (southern Italy), near the eponymous village, at 40°17’N 16°34’E (292 m above the sea level). Its deposition occurred within the Bradanic foredeep, a perisutural basin bounded westward by the Apennine orogenic front and eastward by the Murge highlands. The MJ succession covers more than 630 kyrs (Lower to Middle Pleistocene; Ciaranfi et al., 2010) and has been reconstructed sewing up 10 sub-sections (Ciaranfi et al., 1997; 2001). The succession is divided in two distinct portions by a stratigraphic gap (Fig. 2). The lower portion (“Interval A”) consists of ~180 m thick muddy turbidites including a volcaniclastic layer (V1). Benthic paleocommunities assign “Interval A” to an up- per-slope environment involved in a prevalent shallow- ing upward trend (D’Alessandro et al., 2003; Stefanelli, 2004). The upper portion (“Interval B”) is more than 300 m thick and includes silty sands and silty clays in which eight volcaniclastic layers (V2-V9) are embedded; re- peated deepening-shallowing cycles, from outer to inner shelf environment, have been recognised (D’Alessandro et al., 2003; Stefanelli, 2004) (Fig. 2). Bathymetric changes have been firstly related to the north-eastward thrusting of the southern Apennines units (Ciaranfi et al., 1996; D’Alessandro et al., 2003), and secondly to the orbital-scale G-I ciclicity and climate-driven eustatic variations, in the frame of the progressive Plio- Pleistocene cooling (e.g. Joannin et al., 2008; Ciaranfi et al., 2010). Based on the most recent radiometric datings 132 Toti F. Fig. 1 - (a) Location of Montalbano Jonico and the main sites of southern Italy cited in the text. (b) View of the Montalbano Jonico succession in the “Calanchi” badlands area, with the Agri Valley in the background. (39Ar/40Ar) on V1-V5 tephra layers (Petrosino et al., 2015) and the constrains provided by the calcareous plankton (Maiorano et al., 2004), sapropel (Stefanelli et al., 2005) and isotopic (Brilli et al., 2000) stratigraphy, Marino et al. (2015) have proposed the age-model shown in Fig. 2, closely related to the astronomical tun- ing of the section by Ciaranfi et al. (2010). Correlations with LR04 benthic stack (Lisiecki and Raymo, 2005), Pacific (Mix et al., 1995a,b; Shackleton et al., 1995) and Atlantic (Bickert et al., 1997) δ18O benthic records allow the isotopic shifts related to MIS 22-16 to be outlined (Ciaranfi et al., 2010; Marino et al., 2015). According to the calcareous plankton-inferred relative chronology (Maiorano et al., 2004), “Interval B” extends from the top of Pseudoemiliania lacunosa Zone to the base of Emil- iania huxleyi Zone of Rio et al. (1990) (Fig. 2). Two nannofossil events are recorded: the beginning (826.89 ka) and the end (771.04 ka) of the second temporary disappearance of Gephyrocapsa omega (Maiorano et al., 2004). Although no sapropels occur within the 133 Interglacials at the EMPT in southern Italy Fig. 2 - Lithostratigraphy and benthic δ18O record of the Montalbano Jonico composite section (Brilli et al., 2000; Ciaranfi et al., 2010; Maio- rano et al., 2010; Marino et al., 2015); the age of volcaniclastic layer V4 is after Petrosino et al. (2015); the studied portion is indicated on the right. (Modified from Bertini et al., 2015). “Interval B” (Stefanelli et al., 2005), three bands with higher organic carbon content are visible (Fig. 2): the oldest one encloses V4 tephra (773.9 ±1.3 ka, Petrosino et al., 2015), the second and last one are located below and above the end of the second temporary disappear- ance of Gephyrocapsa omega. The present work is based on the study of eighty- three palynological samples from the 67.7 m (858 ka, MIS 21)-257.5 m (656.05 ka, MIS 16) portion of “Interval B” (Fig. 2), this generating a mean time resolution of ~2463 years. However this result is based on highly uneven values throughout the succession, ranging from 60 years to more than 14 kyrs. 2.2. Present climate and vegetation The MJ section lies between the Metapontine Jonian coast, to the south-east, and the Agri valley, to the north-west. The latter is part of a wider hydrographic basin including all the Apennines rivers with drainage towards the Taranto gulf (Fig. 1). The Agri valley is bor- dered northward by reliefs ranging from ~1700 to ~2000 m. 50 km southwest of MJ, the Pollino massif, with sev- eral peaks over 2000 m, separates Basilicata from Calabria region. The Agri valley approximates the inter- section between three climatic zones: the coastal Meta- pontine area, characterized by a meso-Mediterranean regime, the eastern hinterlands, with semi-continental climatic features, and the western mountain area, with rainfalls well distributed throughout the year and strong thermal and humidity gradients depending on orogra- phy. The MJ section is embedded in the evocative bad- lands landscape, molded by the erosion of Plio- Pleistocene marine clays. The vegetation cover is firstly influenced by edaphic factors, being the substrate the result of sediment weathering (Di Pietro et al., 2004). Halo-xerophytic components, such as Camphorosma monspeliaca, Lygeum spartum and Mantisalca duriaei, colonize the steepest slopes, enriched by Sulla coronata when slopes are weaker. Terophytes grow in more sandy and detritic soils (Fascetti et al., 2001). Amaranthaceae (i.e. Camphorosma, Atriplex, Suaeda) occupy flat and more humid zones at the base of the slopes. In weakly steep areas, poorly affected by erosion processes, terophytic Fabaceae (including He- dysarium and Scorpiurus) dominate the floristic assem- blage, with sporadic ingressions of Plantago afra. On flat to sub-flat surfaces, some patchy Mediterranean maquis occurs; here Pistacia lentiscus dominates in association with Rubia peregrina, Phillyrea latifolia, Juni- perus oxycedrus, Asparagus acutifolius and Helicto- trichon convolutum (Biondi et al., 1992; Fascetti et al., 2001). The basal horizon of the surrounding relief, which extends from a few meters to 300-400 m, belongs to the olive tree-carob tree alliance (Mediterranean climax), also including holm oak; it is worth noting that the latter is an important component of higher (submontane) belts, forming mid-altitude (up to 800-1000 m) wood- lands (Corbetta, 1974). Cypresses and Aleppo pines occur as reforestation elements. SubMediterranean to submontane horizons are marked by the presence of Quercus pubescens, commonly with Fraxinus ornus (plus Ostrya carpinifolia, Carpinus betulus and C. orien- talis in moisture-rich stations), Pyrus communis, Sorbus domestica, Crataegus oxyacantha, Ligustrum vulgare, Spartium junceum, Osyris alba, and sclerophyll taxa such as Asparagus acutifolius and Pistacia terebinthus. At higher elevations Quercus pubescens tends to be replaced by Q. cerris, which, together with chestnut woods, characterizes the submontane horizon. Beech- woods and Abies alba relict formations punctuate moun- tain belts, together with Pteridium aquilinum brackens. The undergrowth composition varies with the altitude and includes noteworthy taxa such as Ilex aquifolium, Geranium striatum, and Stellaria nemonum (Corbetta, 1974). 3. METHODS 3.1. Palynological analyses Palynological analyses have been performed on samples from the MJ section’s “Interval B” (“Ideale”, “JS” and “VCT” sub-sections: see Fig. 2) during the first year of PhD of the author; the pollen study is still in progress in order to increase the resolution in selected intervals. At present, eighty-three sediment samples have been processed using a standard palynological technique, at the Laboratory of Palynology of the Department of Earth Science of the Florence University. Samples have been first weighted before starting the physical-chemical treat- ment. Lycopodium tablets have been added to each sample to estimate palynomorph concentrations. Attacks with acid (HCl and HF), KOH and hexametaphosphate solutions have been followed by enrichments proce- dures (ZnCl2 and 10µm sieving in ultrasonic bath). Resi- dues have been mounted in slides using glycerol and finally analyzed by optical microscope (using x750 and 1250 magnifications) for quantitative pollen analyses. A mean of 284 pollen grains per sample have been counted, Pinus and Cupressaceae excluded. Pinus, which is generally over-represented in the marine sedi- ments (Heusser, 1988; Beaudouin et al., 2005), at MJ reaches very high percentage values and for this reason has been excluded from the total pollen sum (see pollen diagram in Fig. 3); for the taphonomic and palaeo- environmental significance of Pinus pollen grain, see Bertini et al. (2015) and Marino et al. (2015). Cupres- saceae pollen grains have been also removed from the total pollen sum as they exhibit in some levels both very high frequencies and morphologic features which do not support their unequivocal determination. Pollen data are expressed as percentages normal- ized to the total pollen sum, excluding Pinus, Cupres- saceae, indeterminate and indeterminable grains; the summary pollen diagram (Fig. 3) shows eleven selected informal pollen groups established on the basis of cli- mate and ecological requirements of correlative modern taxa, which are: (1) Cedrus and Tsuga; (2) High-altitude taxa (Abies and Picea plus Betula and Fagus); (3) de- ciduous Quercus; (4) Broad-leaved deciduous taxa mi- nus Quercus, e.g. Carpinus, Ostrya, Alnus, Ulmus, Zelk- ova, Alnus, Carya, Pterocarya, Ericaceae and Hedera; (5) Mediterranean sclerophyll taxa, e.g. Quercus ilex type, Olea, Pistacia and Cistaceae; (6) Asteraceae ex- cept Artemisia; (7) Poaceae; (8) Hydrophytes, principally Cyperaceae and Sparganium/Typha angustifolia type; 134 Toti F. (9) Halophytes, e.g. Amaranthaceae, Caryophyllaceae and Plumbaginaceae; (10) Steppic taxa, e.g. Artemisia and Ephedra; (11) Other non-arboreal plants, e.g. Bras- sicaceae, Dipsacaceae, Plantago, Rosaceae and Saxi- fragaceae. Arboreal Pollen (AP) relative abundances have also been plotted (Fig. 3) as they are expected to reflect the spreading/shrinkage of the woodland in response, respectively, to the increase/decline of the atmospheric humidity. In order to discriminate warm (temperate) from cooler to cold phases, a pollen-derived palaeo- temperature index (Pollen Temperature Index: PTI, Fig. 3; e.g. Joannin et al., 2008, 2011; Bertini et al., 2015) has been used. PTI, which is given by the ratio between mesothermic and steppic taxa, puts in evidence the glacial/stadial (lower values)-interglacial/interstadial (higher values) alternation in the pollen record. We have included deciduous Quercus, Corylus, Carpinus, Ostrya, Alnus, Ulmus, Zelkova, Carya, Pterocarya, Tilia, Hedera and Ericaceae (just to mention the most abundant) among mesothermic taxa and Artemiasia, Ephedra, Lygeum and Hippophae among steppic taxa. A thresh- old-value of the PTI has been proposed at ~2 to divide warmer by cooler phases, so that they have distinct separations (Fig. 3). The so defined glacial/stadial (PTI < ~2) and interglacial/interstadial (PTI > ~2) phases are in good agreement - in terms of both chronologic posi- tion and time extension - with the major shifts in the δ8O curve (e.g. sub-stages 21.3, 21.1, 19.3 and 19.1 after: Marino et al.; 2015) of the same succession. 3.2. Multivariate statistical analyses Pollen data relative to the samples with a PTI > ~2 (assumed to belong to interglacials/interstadials, as described in Methods), in which the variables are repre- sented by the eleven (1-11) selected informal groups, have been examined through Principal Component Analysis (PCA) using variance-covariance matrix as data input. PCA has been adopted as a tool for trans- posing the original dataset variability in a lower dimen- sion representation. In this simplified panorama, few principal components are assumed to summarize the associations among correlated variables. Centered log- ratio transformation as proposed by Aitchison (1982) have been carried out to avoid biases deriving from the compositional (constrained) nature of the data. Variabil- ity similarities between selected pollen groups have been also evaluated through hierarchical cluster analy- sis, that has been applied on log-centered variables by using the Euclidean distance as similarity measure and- Ward's method as the subsequent linkage algorithm. 4. RESULTS Palynological analyses point out a rich flora con- sisting of more than 120 taxa. They are largely herba- ceous; among them Asteraceae, Poaceae, Amarantha- ceae and Ephedra are dominant. Artemisia and Ephedra show subsequent phases of significant increase, rea- sonably linked to steppe expansion during (cold/dry) glacial periods at the expenses of forest taxa, Amaran- thaceae usually paralleling the steppe taxa trends (Fig. 3). Among arboreal taxa, Pinus pollen grains reach the highest percentage values. Quercus is also well repre- sented and its increase has been associated to the ex- pansion of the forest during warm-temperate conditions. Among the other deciduous broad-leaved taxa, Carpinus 135 135 Fig. 3 - Pollen Temperature Index (PTI) and percentage values of selected pollen groups in the MIS 21-17 interval of the Montalbano Jonico section. On the left, the age-model (Marino et al., 2015) and the benthic δ18O curve (Brilli et al., 2000; Ciaranfi et al., 2010). The horizontal dashed lines define informal pollen zones corresponding to the successive warm periods (i.e. Zones I-VI plus T), as specified in the column on the right. AP: Arboreal Plants. Interglacials at the EMPT in southern Italy betulus, C. orientalis/Ostrya type, Ulmus, Zelkova, etc. follow. Along with Zelkova, other taxa which progressively disappears in the course of the Pleistocene, such as Carya, Pterocarya and Liquidambar, are still found in this portion of the MJ section, where Taxodium type shows scattered occurrences. Cedrus and Abies are the most abundant taxa among mid- and high- altitude conifers, respectively. As a whole the floristic composition at MJ is quite uniform as pointed out and discussed for the MIS 21 to MIS 18 interval by Bertini et al. (2015); in fact neither dis- appearance nor appearance events have been recorded. 4.1. Vegetation and climatic signature of interglacials and interstadials The subsequent changes in the arbo- real and non-arboreal pollen taxa through- out the MJ succession (858 to 656.05 ka) permit to recognize a clear alternation be- tween forest and open landscapes. They are well expressed especially by the con- traposition between steppe plus other non- arboreal taxa and Quercus plus other mesophilous taxa (Fig. 3). Based on these vegetation data, it is possible to discrimi- nate between warmer and cooler phases. It is further possible to determine the exten- sion of such zones through the PTI. In fact, as specified in Methods, PTI values higher than ~2 have been assumed to depict warm-temperate conditions associated to interglacials/interstadials whereas values lower than ~2 approximate glacials/stadials (Fig. 3). Forty-three sam- ples have been thus included into seven informal “warm” zones: those labeled with roman numbers (I-VI) include samples with PTI stably above 2, whereas that labeled with T (=Transition) includes samples with PTI quickly oscillating between 1.3 and 3 (Fig. 3 and Tab. I). Each zone can be generally correlated with a stage or sub-stage of the marine isotope stratigraphy by consid- ering the astronomical tuning after Marino et al. (2015) and the oxygen isotope stack after Bassinot et al. (1994) (Fig. 3). A short description of the informal zones is pro- vided below. Zone I - It is described by three pollen samples covering ~12 kyrs. AP percentages are constantly high, even if changes in proportions between mid- to high-altitude trees and mesophilous taxa pollen are recorded. Zone I is consistent with MIS 21.3. Zone T - AP are in low percentages, with a main fall at 827.5 ka. Both steppic and halophytic taxa reach a peak at 824.5 ka. Short-term variations of Cedrus and Tsuga diffusely occur. This zone correlates with sub-stage 21.1. Zone II - Mesophilous arboreal taxa show quite high percentages, decreasing only in correspondence with two short-term events (at 783.5 and 774.8 ka, Bertini et al., 2015) marked by the expansion of Asteraceae due to humidity drops. Mediterranean taxa show a significant abundance at 780.56 ka, predating the acme of Quercus and a subsequent increase of high-altitude taxa. This zone correlates with isotopic sub-stage 19.3. Zone III - The first portion of the interval is marked by a positive excursion of oak and Mediterranean taxa. After- wards, the other mesophilous taxa plus the mid- to high- elevations trees slightly increase. A correlation of this zone with MIS 19.1 is consistent. Zone IV - As a whole AP percentage values are similar to those of the previous Zone III (Tab. I).The AP curve shows a peak centered at 741.41 ka. Such an event coincides with an increase in mid- to high-altitude taxa and predates the start of Quercus increase at 740.12 ka. The zone extension ranges between 18.4 and 18.3 iso- tope sub-stages. Zone V - AP percentages stay quite similar to those of the previous zones. Quercus relative abundances rap- idly increase between 728.28 and 726.9 ka, with secon- dary superimposed oscillations well appreciable due to the high time-resolution. A correlation with the 18.3 iso- tope sub-stage can be proposed. Zone VI - It corresponds to MIS 17 and represents the longest interglacial. It is also marked by the highest AP values (a huge shift is located at 675.42 ka). Between 697.31 and 677.96 ka, high-altitude taxa are considera- bly abundant, whereas Cedrus and Tsuga show sparse occurrences. Mediterranean taxa peak at 695 ka, while deciduous Quercus is involved in a gradual increase up 136 Toti F. Tab. I - Montalbano Jonico section: pollen-based zonation for the studied interval, associated with summary vegetation indexes. Correlation with isotopic stages is also shown. Tab. II - Montalbano Jonico section: principal component loadings for the first three axes. to 680.69 ka, followed by a dramatic rise 5 kyrs after, when high-altitude elements collapse. 4.2. Principal component analysis Statistical calculations on selected pollen data allow three principal components to be extracted. They are able to take into account for about 69% of the total variance, thus providing a reasonable summary of the information contained in the dataset (Tab. II). The first principal component (PC 1) encloses 30.62% of the total variance and expresses a balance in which hydro- phytes, Quercus, Poaceae and steppic taxa (with posi- tive loadings) are opposed to Tsuga and Cedrus, high- altitude taxa and Mediterranean taxa (with negative loadings). PC 1 could reflect the different timing of ex- pansion exhibited by taxa with positive loadings (faster- expanding taxa) versus negative loadings (slower- expanding taxa). The analysis of vegetation successions in many sites of southern Europe with sufficient moisture availability confirms that non-synchronous taxa expan- sions occur at the onset of warm-humid periods. Decidu- ous Quercus often expands early, followed by other mesophilous trees, such as Carpinus and Ostrya, and then by Abies, Fagus and sometimes Picea (Tzedakis, 2007; Brauer et al., 2007). In fact, some water- demanding arboreal taxa can show delays in migration 137 Interglacials at the EMPT in southern Italy Fig. 4 - Montalbano Jonico section: plot of sample scores. (a) First vs second principal component, (b) first vs third principal component and (c) second vs third principal component. Samples colours indicate the correspondent pollen zone. because they require well developed soils, which do not form until several millennia from the beginning of an interglacial; con- versely, deciduous oaks can grow on less organic soils (Sadori et al., 2011). Further, we may expect that hydrophytes and other herbaceous taxa react more rapidly to an increase of humidity with respect to arbo- real taxa. The second principal component (PC 2) accounts for the 27.73% of the total variance, and expresses a balance be- tween mid-altitude taxa, halophytes, step- pic taxa (with positive loadings) and Medi- terranean elements, Quercus, plus other broad-leaved thermophilous trees (with negative loadings) (Tab. II). Conceivably, this balance can be considered expression of the differential vegetation’s response to cooler (with mid-altitude taxa and herbs increase) and warmer (with broad-leaved deciduous taxa and Mediterranean xero- phytes increase) climate contexts. The third principal component (PC 3), contribut- ing for the 10.89% to the total data variabil- ity, appears to be almost exclusively con- trolled by high-altitude taxa (with positive loadings) and Mediterranean taxa (with negative loadings). This component can reflect the different moisture requirements of Mediterranean sclerophylls (adapted to seasonal draught) and high-altitude trees (demanding more constant humidity). Samples from the seven statistical units associated with Zones I-VI and T, plotted as scores (the coordinates of the samples in the space of the components) on the three main components (Fig. 4) show scattered distributions and overlap- ping domain areas. In other words, “warm” zones, regarded as groups of samples defined a priori, cannot be discriminated if the variance-covariance structure of taxa groups is applied. 4.3. Cluster analysis Cluster analysis on the centered-log data has therefore been run in order to discover the presence of natural associa- tions in the dataset. The obtained dendro- gram (Fig. 5a) permits to point out the presence of two well distinct groups of samples, marked with different colors (blue cluster vs green cluster) which further help us to sight them across the seven “warm” zones (Fig. 5b). The main clusters ob- tained have been separated by considering that theiroccurred only for a high rescaled distance representing anof high dissimilar- ity. The two clusters have been then plot- ted as scores on the three main compo- nents revealed by multivariate analysis (Fig. 6). In Fig. 6a we can see that PC 1 and PC 2 are both involved in marking the separation between the two clusters. The blue cluster shows higher values for both axes with re- spect to the green cluster. In Fig. 6b the two clusters are plotted in a PC 1 vs PC 3 diagram, and discrimination is 138 Toti F. Fig. 5 - (a) Cluster analysis on the selected warm periods from the Montalbano Jonico record; dendrogram has been obtained using Ward method. (b) Clustered samples are sorted in function of the age. Fig. 6 - Montalbano Jonico section: plot of sample scores on the component axes already used in Fig. 4. Samples are marked with colours of natural associations re- vealed by the cluster analysis. operated by the former. Again, the blue cluster assumes higher values than green cluster. No significant discrimi- nation is obtained by plotting the samples in a PC 2 vs PC 3 diagram. 5. DISCUSSION Palynological analyses point out, in agreement with the isotopic record (Marino et al., 2015), the succession of climatic oscillations between MIS 21 and MIS 16. In particular interglacials and interstadials are well ex- pressed by major increases in deciduous Quercus and other mesophilous taxa and by PTI values higher than ~2. On the other hand, glacials and stadials are charac- terized by the large expansion of steppe and halophytic taxa as well as by PTI values lower than ~2. Cluster analysis on the pollen samples from the “warm” zones (I-VI plus T) indicates that the variability in taxa percentages is expressed by two main associations of samples (i.e. blue and green clusters), showing well sorted distributions on the PCA plots (Fig. 6a and b). Each studied interval (Zones I-VI and T) includes sam- ples associated to both blue and green cluster, except for the Zone V, wich is exclusively represented by blue cluster’s samples (Fig. 7). A first observation is that a similar “blue-green-blue” arrangement unites Zone II and Zone VI (MIS 19.3 and 17, respectively) (Fig. 7). (Zone I represents the last ~12 kyrs of MIS 21.3 and shows a pattern very close to those better expressed in both Zone II and Zone VI). Analogies can also be recog- nised between Zone IV and Zone V (MIS 18.3-18.4 in- terval and MIS 18.3, respectively), since they both show the predominance (up to totality) of “blue” samples (Fig. 7). Instead Zone T and Zone III (MIS 21.1 and 19.1, respectively) show a thickly alternate blue-green pat- tern. Key remarks arise when samples are plotted into the PC 1 vs PC 2 diagram (Fig. 6a). Increasing values of the PC 1 highlight major differences on the rapidity of plant community changes: hydrophytes and deciduous Quercus as well as other herbaceous taxa (with positive loadings), are expected to react more quickly to palaen- vironmental changes than mid- to high-altitude forest taxa and Mediterranean sclerophylls (with negative loadings) (Tab. II). On the other hand, the PC 2 can be thought to reflect the duality among plants with different tolerance to cold: Tsuga, Cedrus, halophytes and step- pic taxa - with positive loadings -, though diffused in very different environments, can tolerate decrease in temperature; conversely, Mediterranean taxa and de- ciduous broad-leaved elements - with negative loadings - have a lower tolerance to temperature decreases. Blue cluster’s samples are marked by higher values for both principal component axes than green cluster’s samples (Fig. 6a). Hence, the former reflect periods marked by (a) not very high temperatures (higher values on PC 2) and (b) the increase of fast-spreading taxa (higher val- ues on PC 1); vice versa, the latter suggest contexts with (c) higher temperatures (lower values on PC 2) and (d) lower abundances of pioneer taxa (lower values on PC 1). This allow some considerations to be made, in the light of the three pattern previously described: - Vegetation succession during MIS 19.3 and 17 (and possibly 21.3) is characterized by a warm mid-phase with a relatively high percentage of slow-spreading taxa, rimmed by cooler intervals with higher abun- dance of fast-expanding plants (sandwich pattern). - The vegetation landscape within MIS 21.1 and 19.1 constantly changes from warmer states, with the in- crease of slow-spreading taxa, to cooler states, in which higher frequencies of fast-spreading taxa are recorded. - Interstadial between MIS 18.4 and MIS 18.3, as well as MIS 18.3, show a rather monotonous vegetation pattern, marked by the constant relatively low pres- ence of both slow-spreading and warmth-demanding taxa. This is probably due to the shortness of the warm phase that prevent those plants to settle. Selected warm periods at MJ section, “Interval B” are different under many points of view, being their dura- tion, time-scale position, intensity and orbital configura- tions widely varying. Nevertheless, cluster analysis shows that the most of the variability is described by only two main classes of samples, which generally coex- ist within each interval. At one level, warm periods glob- ally show more similarities than initially predicted. As hinted by Tzedakis & Bennet (1995), internal biotic fac- tor such as inter- and intra-specific competition can be invoked to justify these analogies. At the same time, it is worth nothing that each period is characterized by a specific vegetation dynamics (see the presence of three patterns), which induce to stress the role of the forcing by external factors. 6. CONCLUSIONS AND PERSPECTIVES MJ section represents an important stratigraphic archive for the reconstruction of the Early-Middle Pleis- tocene G-I cycles in the central Mediterranean basin. In- depth studies on flora and vegetation changes have been previously carried out by Joannin et al. (2008) and Bertini et al. (2015), for the MIS 37-23 and MIS 21-18 intervals, respectively. Here, high resolution palynological analyses and statistical processing of pollen data have been per- formed for the interval MIS 21 to 16. Such an approach has permitted to characterize and compare successive warm periods in terms of duration, intensity and vegeta- tion dynamics. The PCA carried out on PTI-based statis- tical groups has not permitted to discriminate the differ- ent warm periods. On the other hand, cluster analyses underline the presence of two natural association of samples, which bear vegetational signatures when they are plotted on principal component axes. Such a double type of samples is not randomly arranged across the succession and points out three main pattern in the warm periods. Since the age-model of the MJ section allows cor- relations with the orbital geometry predicted for each period (e.g. after Laskar et al., 2004), it would be possi- ble to verify if astronomical forcing can be (at least par- tially) responsible for the different structure of warm periods. The chronologic model suggests a major role especially of the 100 kyrs-period oscillations. In particu- lar, (absolute or relative) eccentricity maxima correlate 139 Interglacials at the EMPT in southern Italy well with the first (sandwich) structure. An in-depth sta- tistical approach would permit to test relationships be- tween orbital and vegetation phenomena but also to better understand how astronomical factors modulate vegetation changes during interglacials and warm phases (Tzedakis, 2012a, b). The present study also fits the issue concerning the extent of the impact of the EMPT on terrestrial ecosys- tems. Finally the high-resolution record of one of the investigated interglacials, i.e. MIS 19, represents a par- 140 Toti F. Fig. 7 - Selected samples of the warm periods of the Montalbano Jonico pollen record - marked with colours from the cluster analysis -, “warm” pollen zones and oxygen isotopic curve (Brilli et al., 2000; Ciaranfi et al., 2010; Marino et al., 2015). ticularly appreciable datum as it is now considered one of the “best analogues” of the present interglacial (Berger et al., 2012; Tzedakis at al., 2012a). ACKNOWLEDGEMENTS This research has been financially supported by the Università degli Studi di Firenze (Fondi di Ateneo, A. Bertini 2011-2013) and a PhD fellowship. The author is deeply grateful to Adele Bertini and Antonella Buccianti (Università degli Studi di Firenze) for the many construc- tive advices. The author wish to thank Luigi Forte, Neri Ciaranfi, Angela Girone, Patrizia Maiorano and Maria Marino (Università degli Studi di Bari, Aldo Moro). A thanks goes to Elda Russo Ermolli and an anonymous reviewer for the thoughtful comments on the previous version of the manuscript. REFERENCES Aitchison J. (1982) - The statistical analysis of composi- tional data (with discussion). Journal of the Royal Statistical Society, Series B (Statistical Methodol- ogy), 44(2), 139-177. Amato V., Aucelli P., Cesarano M., Jicha B., Lebreton V., Orain R., Pappone G., Petrosino P., Russo Ermolli E. (2014) - Quaternary evolution of the largest intermontane basin of the Molise Apennine (central-southern Italy). Rendiconti Lincei, http:// dx.doi.org/10.1007/s12210-014-0324-y Bartlein P.J., Prentice I.C. (1989) - Orbital variations, climate and paleoecology. Trends in Ecology and Evolution, 4, 195-199. Bassinot F.C., Labeyrie L.D., Vincent F., Quidelleur X., Shackleton N.J., Lancelot Y. (1994) - The astro- nomical theory of climate and the age of the Brun- hes-Matuyama magnetic reversal. Earth and Planetary Science Letters, 126, 91-108. Beaudouin C., Suc J.P., Cambon F., Touzani A., Gir- esse P., Pont D., Aloïsi, J.C. Marsset T., Cocho- nat P., Duzer D., Ferrier J. (2005) - Present-day rhythmic deposition in the Grand Rhône prodelta (NW Mediterranean) according to high-resolution pollen analyses. Journal of Coastal Research, 21, 292-306. Bennett K.D., Tzedakis P.C., Willis K.J. (1991) - Quater- nary refugia of north European trees. Journal of Biogeography, 18(1), 103-115. Berger A.L. (1981) - The astronomical theory of paleocli- mates. In: Berger A.L. (eds), Climatic Variations and Variability: Facts and Theories. Reidel, Dordrecht, 501-525. Berger A. (1988) - Milankovitch theory and climate. Re- views of Geophysics, 26, 624-657. Berger A., Yin Q.Z., Herold N. (2012) - MIS-11 and MIS- 19 analogues of our Interglacial. 3rd International Conference on Earth System Modelling, Septem- ber 17-21, 2012, Hamburg, Germany. Bertini A. (2001) - Pliocene climatic cycles and altitud- inal forest development from 2.7 Ma in the north- ern Apennines (Italy): evidences from the pollen record of the Stirone section (~5.1 to ~2.2 Ma). Geobios, 34, 253-265. Bertini A. (2010) - Pliocene to Pleistocene palynoflora and vegetation in Italy: state of the art. Quaternary International, 225(1), 5-24. Bertini A., Ciaranfi N., Marino M., Palombo M.R. (2010) - Proposal for Pliocene and Pleistocene land-sea correlation in the Italian area. Quaternary Interna- tional, 219, 95-108. Bertini A., Toti F., Marino M., Ciaranfi N. (2015) - Vege- tation and climate across the Early-Middle Pleisto- cene transition at the Montalbano Jonico section (southern Italy). Quaternary International, http:// dx.doi.org/10.1016/j.quaint.2015.01.003. Bertoldi R., Rio D., Thunell R. (1989) - Pliocene- Pleistocene vegetational and climatic evolution of the south-central Mediterranean. Palaeogeogra- phy, Palaeoclimatology, Palaeoecology, 72, 263- 275. Bickert T., Curry W.B., Wefer G. (1997) - Late Pliocene to Holocene (2.6-0 Ma) western equatorial Atlantic deep-water circulation: inferences from benthic stable isotope. In: Shackleton N.J., Curry,W.B., Richter C., Bralower T.J. (eds), Proceedings of the Ocean Drilling Program. Scientific Results 154, College Station (TX), 239-253 Biondi E., Ballelli S., Taffetani F. (1992) - La vegetazi- one di alcuni territori calanchivi in Basilicata (Italia meridionale). Documents Phytosociologiques, N.S., 14, 489-498. Brauer A., Allen J.R.M., Mingram J., Dulski P., Wulff S., Huntley B. (2007) - Evidence for Last Interglacial chronology and environmental change from south- ern Europe. Proceedings of the National Academy of Sciences, 104, 450-454. Brilli M., Lerche J., Ciaranfi N., Turi B. (2000) - Evidence of precession and obliquity orbital forcing in Oxy- gen 18 isotope composition of Montalbano Jonico section (Basilicata, southern Italy). Applied Radia- tion and Isotope, 52, 957-964. Capraro L., Asioli A., Backman J., Bertoldi R., Channell J.E.T., Massari F., Rio D. (2005) - Climatic pat- terns revealed by pollen and oxygen isotope re- cords across the Matuyama-Brunhes Boundary in the central Mediterranean (southern Italy). In: Head M.J., Gibbard P.L. (eds), Early-Middle Pleis- tocene Transitions: the Land-Ocean Evidence, Geological Society London, Special Publications, 247, 159-182. Ciaranfi N., Marino M., Sabato L., D’Alessandro A., De Rosa R. (1996) - Studio geologico-stratigrafico di una successione infra e mesopleistocenica nella parte sud-occidentale della Fossa bradanica (Montalbano Jonico, Basilicata). Bolletino Società Geologica Italiana, 115, 379-391. Ciaranfi N., D’Alessandro A., Marino M. (1997) - A can- didate section for the lower-middle Pleistocene boundary (Apennine foredeep, South Italy). In: Naiwen, W., Remane, J. (eds), Proceedings 30th International Geological Congress, 11, 201-211. Ciaranfi N., D’Alessandro A., Girone A., Maiorano P., Marino M., Soldani D., Stefanelli S. (2001) - Pleis- tocene sections in the Montalbano Jonico area and the potential GSSP for Early-Middle Pleisto- cene in the Lucania Basin (southern Italy). Memo- rie di Scienze Geologiche, 53, 67-83. 141 Interglacials at the EMPT in southern Italy Ciaranfi N., Lirer F., Lirer L., Louren, L.J., Maioran, P., Marino M., Petrosino P., Sprovieri M., Stefanelli S., Brilli M., Girone A., Joannin S., Pelosi N., Vallefuoco M. (2010) - Integrated stratigraphy and astronomical tuning of the Lower-Middle Pleisto- cene Montalbano Jonico land section (southern Italy). Quaternary International, 210, 109-120. Combourieu-Nebout N. (1993) - Vegetation response to Upper Pliocene glacial/Interglacial cyclicity in the central Mediterranean. Quaternary Research 40, 228-236. Combourieu-Nebout N., Semah F., Djubiantono T. (1990) - La limite Pliocène-Pléistocène: précisions magnétostratigraphiques par l’étude sériée de la coupe-type de Vrica (Crotone, Italie). Comptes Rendus de l’Académie de Sciences, Paris (ser. II), 311, 851-857. Combourieu Nebout N., Vergnaud Grazzini C. (1991) - Late Pliocene northern hemisphere glaciations: the continental and marine responses in the cen- tral Mediterranean. Quaternary Science Reviews, 10, 319-334. Corbetta F. (1974) - Lineamenti della vegetazione lu- cana. Giornale Botanico Italiano 108(5), 211-234. D'Alessandro A., La Perna R., Ciaranfi N. (2003) - Re- sponse of macrobenthos to changes in paleoenvi- ronment in the Lower-Middle Pleistocene (Lucania Basin, Southern Italy). Il Quaternario, 16, 167-182. Di Pietro R., Fascetti S., Pompili M. (2004) - Vegetation soil relationship in Basilicata badlands (southern Italy). Symposium of the Federation Internationale de Phytosociologie, Tenerife. Fascetti S., Di Pietro R., Pompili M. (2001) - Aspetti sinecologici della vegetazione dei rilievi argillosi della Basilicata. Atti Congresso Società Italiana Fitosociologia. La Vegetazione Sinantropica, Lipari, 2001. Hays J.D., Imbrie J., Shackleton N.J. (1976) - Variations in the Earth's orbit: pacemaker of the ice ages. Science, 194, 1121-1132. Head M.J., Gibbard, P.L. (2005) - Early-Middle Pleisto- cene transitions: an overview and recommenda- tion for the defining boundary. In: Head M.J., Gibbard P.L. (eds), Early-Middle Pleistocene tran- sitions: the land-ocean evidence: Geological Soci- ety of London, Special Publication, 247, 1-18. Head M.J., Pillans B., Farquhar A. (2008) - The Early Middle Pleistocene transition: characterization and proposed guide for the defining boundary. Epi- sodes, 31, 255-259. Huntley B., Webb T. III (1989) - Migration: species’ re- sponse to climatic variations caused by changes in the earth’s orbit. Journal of Biogeography, 16, 5 -19. Heusser L.E. (1988) - Pollen distribution in marine sedi- ments on the continental margin off northern Cali- fornia. Marine Geology, 80, 131-147. Klotz S., Fauquette S., Combourieu Nebout N., Uhl D., Suc J.P., Mosbrugger V. (2006) - Seasonality intensification and long-term winter cooling as a part of the Late Pliocene climate development. Earth and Planetary Science Letters, 241, 174- 187. Joannin S., Ciaranfi N., Stefanelli S. (2008) - Vegetation changes during the late Early Pleistocene at Montalbano Jonico (Province of Matera, Southern Italy) based on pollen analysis. Palaeogeography, Palaeoclimatology, Palaeoecology, 270, 92-101. Joannin S., Bassinot F., Combourieu Nebout N., Peyron O., Beaudouin C. (2011) - Vegetation response to obliquity and precession forcing during the Mid- Pleistocene Transition in Western Mediterranean region (ODP site 976). Quaternary Science Re- views, 30, 280-297. Laskar J., Robutel P., Joutel F., Gastineau M., Correia A.C.M., Levrard B. (2004) - A long term numerical solution for the insolation quantities of the Earth, Astronomy and Astrophysics, 428, 261-286. Lisieki L.E., Raymo M.E. (2005) - A Pliocene- Pleistocene stack of 57 globally distributed benthic δ18O records. Paleoceanography, 20, PA1003. Maiorano P., Marino M., Di Stefano E., Ciaranfi N. (2004) - Calcareous nannofossil events in the lower-middle Pleistocene transition at the Montal- bano Jonico section and ODP Site 964: calibration with isotope and sapropel stratigraphy. Rivista Italiana di Paleontologia e Stratigrafia, 110, 547- 557. Marino M., Bertini A., Ciaranfi N., Aiello G., Barra D., Gallicchio S., Girone A., La Perna R., Lirer F., Maiorano P., Petrosino P., Toti F. (2015) - Pa- leoenvironmental and climatostratigraphic insights for Marine Isotope Stage 19 (Pleistocene) at the Montalbano Jonico section, South Italy. Quater- nary International, http://dx.doi.org/10.1016/ j.quaint. 2015.01.043 Maslin M.A., Ridgwell A.J. (2005) - Mid-Pleistocene revolution and the ‘eccentricity myth’. In: M.J. Head and P.L. Gibbard (eds), Early-Middle Pleis- tocene transitions: the land-ocean evidence: Geo- logical Society of London, Special Publication, 247, 19-34. Mix A.C., Le J., Shackleton N.J. (1995a) - Benthic fo- raminifer stable isotope stratigraphy of Site 846: 0- 1.8 Ma. In: Pisias N.G., Mayer L., Janecek T., Palmer-Julson A., van Andel T.H. (eds), Proceed- ings of the Ocean Drilling Program. Scientific Re- sults, 138, College Station (TX), 839-856. Mix A.C., Pisias N.G., Rugh W., Wilson J., Morey A., Hagelberg T. (1995b) - Benthic foraminiferal stable isotope record from Site 849: 0-5 Ma. In: Pisias N.G., Mayer L., Janecek T., Palmer-Julson A., van Andel T.H. (eds), Proceedings of the Ocean Drill- ing Program. Scientific Results 138, College Sta- tion (TX), 371-412. Petrosino P., Jicha B., Mazzeo F.C., Russo Ermolli E. (2014) - A high resolution tephrochronological record of MIS 14-12 in the Southern Apennines (Acerno basin, Italy). Journal of Volcanology and Geothermal Research, 274, 34-50. Petrosino P., Jicha B.R., Mazzeo F.C., Ciaranfi N., Gi- rone A., Maiorano P., Marino M. (2015) - The Montalbano Jonico marine succession: An archive for distal tephra layers at the Early-Middle Pleisto- cene boundary in southern Italy. Quaternary Inter- national, http://dx.doi.org/10.1016/j.quaint.2014. 142 Toti F. Ms. received: May 18, 2015 Final text received: July 10, 2015 10.049 Robustelli G., Russo Ermolli E., Petrosino P., Jicha B., Sardella R., Donato P. (2014) - Tectonic and cli- matic controls on geomorphological and sedimen- tary evolution of the Mercure basin, Southern Ap- ennines, Italy. Geomorphology, 214, 423-435. Russo Ermolli E., Aucelli P., Di Rollo A., Mattei M., Pet- rosino P., Porreca M., Rosskopf C. (2010a) - An integrated stratigraphical approach to the late Middle Pleistocene succession of the Sessano lacustrine basin (Molise, Italy). Quaternary Inter- national, 225, 114-127. Russo Ermolli E., Sardella R., Di Maio G., Petronio C., Santangelo N. (2010b) - The late Early Pleisto- cene record of Saticula (Sant’Agata de’ Goti, Benevento, Italy). Quaternary International, 225, 128-137. Russo Ermolli E., Di Donato V., Martin-Fernandez J., Orain R., Lebreton V., Piovesan G. (2014) - Vege- tation patterns in the Southern Apennines (Italy) during MIS 13: deciphering pollen variability along a NW-SE transect. Review of Palaeobotany and Palynology, http://dx.doi.org/10.1016/j.revpalbo. 2014.11.004 Sadori L., Jahns S., Peyron O. (2011) - Mid-Holocene vegetation history of the central Mediterranean. The Holocene, 21(1) 117-129. Shackleton N.J., Hall M.A., Pate D. (1995) - Pliocene stable isotope stratigraphy of Site 846. In: Pisias N.G., Mayer L.A., Janecek T.R., Palmer-Julson A., van Andel T.H. (eds), Leg 138. Proceedings of the Ocean Drilling Program, Scientific Results, 138, 337-355. Stefanelli S. (2004) - Cyclic stages in oxygenation based on foraminiferal microhabitats:early-middle Pleistocene, Lucania basin, southern Italy. Journal of Micropaleontology, 23, 81-95. Stefanelli S., Capotondi L., Ciaranfi N. (2005) - Fo- raminiferal record and environmental changes during the deposition of early-middle Pleistocene sapropels in southern Italy. Palaeogeography, Palaeoclimatology, Palaeoecology, 216, 27-52. Suc J.-P., Bertini A., Combourieu-Nebout N., Diniz F., Leroy S., Russo-Ermolli E., Zheng Z., Bessais E., Ferrier J. (1995) - Structure of West Mediterra- nean vegetation and climate since 5.3 Ma. Acta Zoological Cracoviense, 38(1), 3-16. Suc J.P., Combourieu Nebout N., Seret G., Popescu S.A., Klotz S., Gautier F., Clauzon G., Westgate J., Insinga D., Sandhu A.S. (2010) - The Crotone series: a synthesis and new data. Quaternary In- ternational, 219, 121-133. Tzedakis P.C., Bennett K.D. (1995) - Interglacial vegeta- tion succession: a view from southern Europe. Quaternary Science Reviews, 14, 967-982. Tzedakis P.C. (2007) - Seven ambiguities in the Medi- terranean palaeoenvironmental narrative. Quater- nary Science Reviews, 26, 2042-2066 Tzedakis P.C., Wolff E.W., Skinner L.C., Brovkin V., Hodell D.A., McManus J.F., Raynaud D. (2012a) - Can we predict the duration of an Interglacial? Climate of the Past, 8, 1473-1485. Tzedakis P.C., Channell J.E.T., Hodell D.A., Kleiven H.F., Skinner L.C. (2012b) - Determining the natu- ral length of the current interglacial. Nature Geo- science, 5, 138-141. Watts, W.A. (1988) - Europe. In: Huntley B., Webb III T. (eds), Handbook of Vegetation Science 7. Vegeta- tion History. Kluwer, Dordrecht, pp. 155-192. 143 Interglacials at the EMPT in southern Italy