to Pleistocene large mammal diversity and turnover in North Mediterranean region: the Greek Peninsula with respect to the Iberian and Italian ones

Dimitris S. KOSTOPOULOS Department of Geology, University of Thessaloniki, GR-54124 Thessaloniki (Greece) [email protected]

Maria-Rita PALOMBO Dipartimento di Scienze della Terra, Università “La Sapienza”, CNR, Istituto di Geologia Ambientale e Geoingegneria, P.le A. Moro 5, I-00185 Roma (Italy) [email protected]

Maite-Teressa ALBERDI Museo Nacional de Ciencias Naturales, CSIC, J. Gutierez Abascal 2, E-28006 Madrid (Spain) [email protected]

Andrea M. F. VALLI Dipartimento di Scienze della Terra, Università “La Sapienza”, P.le A. Moro 5, I-00185 Roma (Italy) [email protected]

Kostopoulos D. S., Palombo M.-R., Alberdi M.-T. & Valli A. M. F. 2007. — Pliocene to Pleistocene large mammal diversity and turnover in North Mediterranean region: the Greek Peninsula with respect to the Iberian and Italian ones. Geodiversitas 29 (3) : 401-419.

ABSTRACT During the Plio-Pleistocene several important climatic events constrained faunal renewal phases, aff ecting the structure of mammal communities, their richness and diversity. In order to complete the analysis of the Pliocene to Pleistocene large mammal faunal renewal in North Mediterranean region, the Greek turnover and diversity patterns from Ruscinian (about 5 Ma BP) to the KEY WORDS end of middle Pleistocene (about 175-127 ka BP) are compared with those of Large mammals, Italy and Spain. Th e analysis takes into account the changes in the taxonomical Plio-Pleistocene, composition and ecological structure of the Greek, Italian and Spanish succes- diversity, turnover, sion of regional faunal complexes, which correspond to high rank biochrons Southern Europe. (Mammal Ages or Mammal Subages) or they are provided from a cluster analysis.

GEODIVERSITAS • 2007 • 29 (3) © Publications Scientifi ques du Muséum national d’Histoire naturelle, Paris. www.geodiversitas.com 401 Kostopoulos D. S. et al.

Th e results indicate a longitudinal diversifi cation from Greece to Spain and a signifi cant early Pleistocene (Galerian) faunal renewal that af- fected more or less simultaneously the entire North Mediterranean region as a response to the establishment of cooler-dryer climatic conditions around the Jaramillo magnetochron.

RÉSUMÉ Diversité et changement des faunes de grands mammifères plio-pléistocènes sur le pourtour nord de la Méditerranée : comparaison des faunes de la péninsule grecque avec celles d’Italie et d’Ibérie. Pendant le Plio-Pléistocène plusieurs événements climatiques importants ont provoqué des phases de renouvellement faunique, aff ectant la structure des communautés à mammifères, et également leur richesse et leur diversité. Dans l’objectif de compléter l’analyse des changements des faunes à grands mam- mifères dans la région nord-méditerranéenne du Pliocène au Pléistocène, le renouvellement et la diversité faunique de la Grèce ont été comparés à ceux de l’Italie et de l’Espagne, du Ruscinien (environ 5 Ma BP) à la fi n du Pléistocène moyen (175-127 ka BP). L’analyse prend en compte les changements de com- position taxonomique et de structure écologique des successions des complexes fauniques régionaux de Grèce, d’Italie et d’Espagne. Ils correspondent aux com- plexes biochronologiques majeurs (« Mammal Ages » ou « Mammal Subages ») ou dérivent de l’analyse de regroupement par comparaison (« cluster analysis »). Les résultats montrent qu’il se vérifi e, pendant le Villafranchien, une diversifi - MOTS CLÉS cation longitudinale de l’est vers l’ouest (de la Grèce jusqu’à l’Espagne), et qu’il Grands mammifères, se produit un renouvellement signifi catif concernant, plus au moins à la même Plio-Pléistocène, époque (Pléistocène inférieur, Galérien), toute la région nord-méditerranéenne. diversité, turnover, Ce renouvellement est la réponse à l’établissement des conditions climatiques Europe du Sud. froides et sèches correspondant au magnétochron de Jaramillo.

INTRODUCTION great infl uence on the biogeographic distribution of mammalian species, aff ecting the taxonomical Th e crucial role of both the physical and the biotic composition and the structure of mammal com- factors in confi guring mammal communities has munities through time. Environmental changes been extensively discussed by several authors, who were assumed to be a driving force in changing propose diff erent models to explain the regulation diversity: taxonomic groups undergo long periods of faunal diversity with respect to time (e.g., Van of quasi-static dynamic that are interrupted by Valen 1973; Turner 1995; Vrba 1995; Brett et al. intervals of especially high rates of extinction/emi- 1996; Behrensmeyer et al. 1997; Alberdi et al. 1997; gration and origination/immigration (Vrba 1985, Azanza et al. 1999, 2003, 2004; Palombo et al. 1995; Azanza et al. 2004; Palombo & Valli 2004). 2003; Fortelius et al. 2006 and literature therein). With respect to the role played by biotic factors in Physical and biotic environmental conditions exert mammal radiation and extinction the “Turnover

402 GEODIVERSITAS • 2007 • 29 (3) Plio-Pleistocene mammal diversity and turnover in North Mediterranean

pulse hypothesis” (Vrba 1993, 1995) suggests that a great opportunity to study the evolution of the most of lineage turnovers occurred when groups of ecological features of regional faunas with respect organism spread in pulses correlated with changes to environmental changes. Alberdi et al. (1998), in the physical environment. Furthermore, renewal Azanza et al. (1999, 2002, 2004), Palombo et al. patterns might be constrained by climatic-driven (2003) and Palombo & Valli (2004) already stud- eff ect at intermediate to large temporal scales. ied the infl uences of global climatic changes on Hence, changes in mammal faunal diversity have the structural evolution of southwestern European often been considered as the result of bioevents, mammal faunas. In this paper we extend this study correlated with major climatic changes, which, in eastwards, by comparing the diversity and structural their turn, can determine concurrent bioevents in dynamics of the reconstructed Greek palaeocom- multiple lineages. munities with the SW European ones. Apparently, the composition of regional faunal complexes is the result of both ecological and his- torical factors that control the structure of commu- METHODOLOGY nities even inside a particular biogeographic area. Indeed, the possibility of dispersion, diff usion and In order to evaluate the possibility of correlating survival of taxa – especially when forced by climatic trans-South European faunal renovation with changes events – were greatly aff ected by the physiography, in environmental conditions that assumed to have the microclimate conditions, and the availability occured at the transition between faunal complexes, of ecological niches. As far as the North Mediter- we compare the turnover patterns of Greece with ranean is concerned, past and present large mam- those of Spain and Italy from Ruscinian (about 5 mal distribution is not homogeneous. Although Ma BP) to the end of middle Pleistocene (about the South European realm is supposed to outline a 175-127 ka BP). uniform zoogeographic province, its environment A change in the taxonomical composition between and topography are highly heterogeneous. Despite two successive biochrons can be considered as a faunal general similarities, the three well separated penin- turnover, when alterations in species composition sulas maintained diff erent confi guration through is provided by the concurrent extinction of exist- time and had diff erent biogeographic signifi cance ing species (exits) and their replacement through because of geographical and ecological barriers, immigration by new ones (entries). An increase represented by local mountain chains, relations or decrease in richness depending on just one of between their latitudinal and longitudinal exten- these factors should be regarded as a dispersal or sion (e.g., Spain versus Italy), and infl uence of the extinction event (see Alberdi et al. 1998; Torre et Mediterranean Sea on their climatic conditions. al. 1999; Azanza et al. 2003; Palombo 2004, 2005 From the end of , the “Greek arm” of and literature therein). Actually, even if constrained the Balkan Peninsula constituted a reception and by paleoenvironmental changes, the extinction/ diff usion area of several migration waves directed emigration and the origination/immigration events mainly from the East to the West (Kostopoulos et were scattered in time. In order to evaluate average al. 2002; Spassov 2003). On the other hand, it is trend in ecological structure of extinct faunas, we generally claimed that at the same time-span the have to establish a theoretic faunal complex that Iberian and Italian peninsulas formed a kind of “cul can be used as a “block of coordinated stasis” dur- de sac”, where ended most of the Eurasian dispersal ing which no turnovers are expected (Brett et al. events which aff ected Western Europe (Alberdi et al. 1996). As local faunal assemblages (LFAs) arranged 1997; Palombo et al. 2003). In addition, Southern in biochron units could be regarded as represent- Europe is considered as a refuge area for northern ing not overlapping and “ecologically adjusted taxa during the Pleistocene climate worsening. groups of animals with specifi c geographic limits Th e quite rich Late Cenozoic mammal record and chronological range” (Tedford 1970), we as- from the North Mediterranean region provides sume biochrons as “blocks of coordinated stasis”

GEODIVERSITAS • 2007 • 29 (3) 403 Kostopoulos D. S. et al.

TABLE 1. — Relative age distribution of Greek, Italian and Spanish local faunal assemblages (LFAs) used in the analyses. Data from Palombo et al. (2006) and literature therein.

Mammal ages/ Greek LFAs Italian LFAs Spanish LFAs subages Aurelian Petralona, Megalopolis Montignoso, Casal de La Carihuella, Los Casares, Pazzi,Torre in Pietra C. Negra, Valdegoba, (lower and upper), Villacastin, Pinarillo-1, Sedia Diavolo Torrejones Congosto, Pinilla del Valle Torralba, Ambrona, Atapuerca 10, La Solana del Zamporino Galerian Marathoussa-Megalopolis, Fontana Ranuccio, Visogliano, Cullar de Baza, Atapuerca 4, Kaiafas, Ravin Voulgarakis, Isernia la Pinetta, Valdemino, Atapuerca 6, Huescar Apollonia Slivia, P. Galeria, Redicicoli, Colle Curti Late Villafranchian Alikes, Pyrgos, Libakos Pirro Nord, Pietrafi tta, M. Ric- Venta Micena, Quibas, Ponton Gerakarou, Vassiloudi cio, Casafrata, Tasso, Olivola, de la Oliva, Fonelas Valdarno superiore, P. Rosso Middle Villafranchian Volakas, Dafnero, C. S. Giacomo, Montopoli La Puebla de Valverde, Sesklo, Vatera Huelago, El Rincón Early Villafranchian – Colle Pardo, Triversa, G. S. Villaroya Barbara, Barga, Pieve Fosciana Ruscinian Apolakkia, Megalo Emvolon Val di Pugna La Calera, La Gloria 2, Layna, Kessani, Maramena Alcoy and we estimate turnover rates between successive following Harper’s (1975) method: Nrs = Nbda + biochrons. Consequently, faunal renewals can be Nrt + ½ (Nf + Nl - No) (Nrs, number of taxa that extracted from the number of extinctions (last potentially occur at a given time interval; Nbda, appearances, LA) at the end of a biochron and number of species present before, during and after the new appearances (fi rst appearances, FA) at the the faunal unit; Nrt, number of species present beginning of the successive one. Turnover indices before and after but not in the faunal unit; Nf, (TI = % FA + % LA/2) are calculated using fi rst number of fi rst appearances [FA]; Nl, number of and last appearance percentages after normalising last appearances [LA]; No, number of taxa present the LA and FA values (% FA = FA/RM × 100; only in the faunal unit). % LA = LA/RM × 100; RM = N - (FA/2 + LA/2); A basic problem remains, however, on how we can N = number of taxa) as in Torre et al. (1999). ascertain a kind of successive faunal complexes that Richness, used as an estimate of faunal diversity, enable us to correlate among diff erent geographic is currently measured by the total number of taxa regions (e.g., Pickford 1990). Actually, correlation actually (Nt) or potentially occurring in every bio- among faunal complexes belonging to diff erent chronological interval. Nevertheless, the richness of geographic areas is not an easy matter because of a single time interval might be overestimated when diff erences among local fossil records (both by means fi rst local appearances are considered to occur at the of taxonomical denomination and completeness), beginning of the interval, and the disappeared taxa taphonomic biases, endemic phenomena, and thought to persist until its end, whereas they might diff erences in the local biochronological/biostrati- not actually overlap in time. In order to reduce this graphical schemes. Hence, faunal renewals have bias, it is possible to standardise the number of taxa been calculated at the transition between biochrons that potentially occur at a given time interval by (Mammal Ages, MAs, and Mammal subAges, considering species richness at its midpoint. Con- MsAs) according to the biochronological schemes sequently, we calculated standing richness values already established for European mammal faunas

404 GEODIVERSITAS • 2007 • 29 (3) Plio-Pleistocene mammal diversity and turnover in North Mediterranean

(e.g., Palombo et al. 2006 and literature therein). grassland, steppe or savanna) and mixed dwellers Th e following biochronological faunal complexes (ubiquitous taxa, which can equally live in forests (BFCs) have been considered: Ruscinian (MN14, and open landscape, or at the edge of both). MN15), early Villafranchian (MN16a or Italian Triversa faunal unit, FU), middle Villafranchian (MN16b, MN17 or Italian FUs from Montopoli TURNOVER ANALYSIS to Costa San Giacomo and MNQ18 partim), late Villafranchian (Italian FUs from Olivola to Pirro, GREECE MNQ18 [partim]-MNQ19), Galerian (Italian FUs Th e BFCs and CFCs analyses results are almost identi- from Colle Curti to Fontana Ranuccio, MNQ20- cal (Fig. 1). Th e main problem concerns the paucity of MNQ23), and Aurelian (early Aurelian = Italian Greek early Villafranchian LFAs and the consequent Torre in Pietra FU, MNQ24; late Aurelian includ- inability to establish a distinct faunal complex for ing late Pleistocene LFAs) (sensu Gliozzi et al. 1997; this mammal stage, based on the selected data. Fol- Guerin 1990; Mein 1990). lowing the results of the cluster analysis (Palombo et During the last years multivariate techniques al. 2006), we decide to join the Ruscinian and the have been extensively and successfully applied in early Villafranchian faunal complexes, admitting an the recognition of successive, not overlapping and a-priori distortion, which causes an overestimation ecologically adjusted assemblages of taxa, living of the turnover between the Ruscinian and the Vil- together in a given space and time (Azanza et al. lafranchian, characterised by high rates of extinction 1997, 1999; Palombo et al. 2003; Palombo & Valli and much more fi rst appearances (Fig. 1). Although 2004 and literature therein). Th erefore, we have also fi ctitious, these results broadly refl ect reality, especially analysed the faunal renewals for each geographic regarding the extinction phase at the Ruscinian-Villa- area on the basis of large mammal complexes (Ta- franchian boundary and the richness increase after ble 1) resulting from a similarity analysis based on Ruscinan. Both in the BFCs and CFCs analyses, the the common presence of taxa (see Palombo et al. middle-late Villafranchian transition is character- 2006). Nineteen Greek faunas ranging from the ized by the predominance of fi rst appearances over latest Miocene to the middle Pleistocene, 29 Ital- extinctions, even with a low turnover index, leading ian faunas ranging from the earliest Villafranchian to the increase of richness, which reaches its maxi- to the end of middle Pleistocene and 29 Spanish mum during the late Villafranchian. An important faunas ranging from the earliest Pliocene to the renewal event occurs at the boundary between the late Pleistocene have been involved (Palombo et late Villafranchian and the Galerian, in the BFCs al. 2006). According to the results of this analysis, analysis, or at the transition between the early and successive groups of LFAs are considered as repre- middle Galerian, in CFCs analysis (Fig. 1). In both senting succession of local “biochrons”, here named cases, the turnover index reaches its maximum values, cluster faunal complexes (CFCs). whereas the phase is characterised by an important For ecological purposes, we estimate the percent- increase of extinctions and a rather decline of fi rst age of taxa according to the preferred habitat and appearances (CFCs analysis), providing a richness dietary guild. Carnivores were regarded all together, reduction, especially during the middle Galerian while the herbivores were separated into grazers (Gr), (CFCs analysis). At the Galerian-Aurelian transi- browsers (Br), and mixed feeders (B-G), including tion, last appearances are minimized and rates of taxa which, on a seasonal, regional or occasional new appearances slightly rise, leading together to a basis, indiff erently eat grass, leaves or both. Th e signifi cant richness increase (Fig. 1). categories of frugivores (Fr) and omnivores (Om) were also considered. Th ree major ecological habitat ITALY groups were retained: forest dwellers (taxa inhabit- During the early and the middle Villafranchian, ing closed woodland, clear woodland, miscellaneous diff usion phases have occurred, characterised by woodland), open land dwellers (taxa inhabiting the predominance of fi rst appearances (Fig. 2). Th is

GEODIVERSITAS • 2007 • 29 (3) 405 Kostopoulos D. S. et al.

BFC CFC 70.00 70.00

60.00 60.00

50.00 50.00

40.00 40.00

30.00 30.00

20.00 20v00

10.00 10.00

0.00 0.00 R-eV/mVmV/IVIV/GG/A R-eV/mV mV/IV-eG IV-eG/mG mG/IG-A RM Total number of FA Total number of LA Total number of species

120.00 120.00

100.00 100.00

80.00 80.00

60.00 60.00

40.00 40.00

20.00 20.00

0.00 0.00 R-eV/mV mV/IV IV/G G/A R-eV/mV mV/IV-eG IV-eG/mG mG/IG-A

% FA % LA RM TI

40.00 40.00

30.00 30.00

20.00 20.00

10.00 10.00

0.00 0.00 R eV mV IV G A R eV mV IV-eG mG IG-A

Nrs Nt

FIG. 1. — Turnover pattern analysis of Greek biochronological (BFC) and cluster (CFC) faunal complexes. Abbreviations: FA, fi rst appearances; LA, last appearances; Nrs, standing species richness; Nt, species richness; RM, running mean; TI, turnover index; A, Aurelian; G, Galerian; R, Ruscinian; V, Villafranchian; e, early; l, late; m, middle. trend allows richness to reach its maximum at the documented either at the boundary between the beginning of the late Villafranchian, while according late Villafranchian and the Galerian in the BFCs to CFCs analysis (Fig. 2), richness decreases dur- analysis or during the transition from the early to ing the latest Villafranchian and at the transition the middle Galerian in the CFCs analysis (Fig. 2). between Villafranchian and Galerian. Th e end of In the fi rst case, the passage is characterised by an early Pleistocene marks the major faunal renewal, increase of extinctions whereas rates of appear-

406 GEODIVERSITAS • 2007 • 29 (3) Plio-Pleistocene mammal diversity and turnover in North Mediterranean

BFC CFC 80.00 80.00 70.00 70.00 60.00 60.00 50.00 50.00 40.00 40.00 30.00 30.00 20.00 20.00 10.00 10.00 0.00 0.00 R/eV eV/mV mV/IV IV/G G/eAR-eV/m-IV m-IV/IV-eG IV-eG/m-IG m-IG/eA RM Total number of FA Total number of LA Total number of species

120.00 120.00

100.00 100.00

80.00 80.00

60.00 60.00

40.00 40.00

20.00 20.00

0.00 0.00 R/eV eV/mV mV/IV IV/G G/eA R-eV/m-IV m-IV/IV-eG IV-eG/m-IG m-IG/eA

% FA % LA RM TI

50.00 50.00

40.00 40.00

30.00 30.00

20.00 20.00

10.00 10.00

0.00 0.00 ReVmVIVGeA R-eV m-IV IV-eG m-IG eA

Nrs Nt

FIG. 2. — Turnover pattern analysis of Italian biochronological (BFC) and cluster (CFC) faunal complexes. Abbreviations as in Fig. 1. ances show just a slight increase compared to the Afterwards, a progressive renewal takes place, reach- preceding phases, while in the CFCs analysis, it ing its apex at the late Villafranchian-early Galerian is characterised by the predominance of fi rst ap- transition (at c. 1 Ma). Th is renewal, already evident pearance data. at the middle-late Villafranchian boundary (Fig. 3), is fi rstly characterised by a predominance of new SPAIN appearances over extinctions and allows richness Th e BFCs analysis shows a turnover at the Ruscin- to increase. Th en, extinctions exceed new appear- ian-early Villafranchian boundary, followed by an ances (BFCs analysis) causing a progressive drop extinction phase and a richness decline (Fig. 3). in the richness. At the Galerian-Aurelian transition

GEODIVERSITAS • 2007 • 29 (3) 407 Kostopoulos D. S. et al.

BFC CFC 70.00 70.00

60.00 60.00

50.00 50.00

40.00 40.00

30.00 30.00

20.00 20.00

10.00 10.00

0.00 0.00 R/eV eV/mV mV/IV IV/GG/A R/e-mV e-mV/IV IV/G G/A RM Total number of FA Total number of LA Total number of species

120.00 120.00

100.00 100.00

80.00 80.00

60.00 60.00

40.00 40.00

20.00 20.00

0.00 0.00 R/eV eV/mV mV/IV IV/GG/A R/e-mV e-mV/IV IV/G G/A

% FA % LA RM TI

40.00 40.00

30.00 30.00

20.00 20.00

10.00 10.00

0.00 0.00 R eV mV IV G eA R e-mV IV G A

Nrs Nt

FIG. 3. — Turnover pattern analysis of Spanish biochronological (BFC) and cluster (CFC) faunal complexes. Abbreviations as in Fig. 1.

extinctions still prevail, but the progressive uplift of PALAEOECOLOGICAL ANALYSIS new appearances leads to a bend in richness. Th e CFCs analysis (Fig. 3) indicates both the dispersion In the absence of data, early Villafranchian complex phase at the middle-late Villafranchian transition and is excluded from Greek BFCs calculations. Greek the extinction increase from the late Villafranchian carnivores number is rather stable during Villa- to the Galerian-Aurelian. franchian but it decreases in Galerian, recovering

408 GEODIVERSITAS • 2007 • 29 (3) Plio-Pleistocene mammal diversity and turnover in North Mediterranean

during Aurelian (BFCs analysis; Fig. 4). Omnivo- decrease. In Spain the highest herbivore percent- rous and frugivorous species show a gradual decline age corresponds to the Ruscinian, followed by a from the Ruscinian to the late Villafranchian/early decrease of grazers in the following Pliocene, an Galerian, but they increase afterwards (Fig. 4). increase during the late Villafranchian and the From the Pliocene to the early middle Pleistocene Aurelian and by a drop during the Galerian (also herbivores remain stable with a signifi cant increase similarly to the Greek trend; Fig. 9). Browsers are during Galerian (Fig. 4). Nevertheless, browsers diff erently represented in the three peninsulas, gradually decrease from the Ruscinian to the late being much more abundant in Spain than in Italy Villafranchian/early Galerian and then they slightly and very low in Greece (Figs 5; 7; 9). However, recover, whereas grazers show a general increase they show a relatively uniform trend in the three throughout this period, interrupted by a reversal in peninsulas, decreasing towards the late Villafran- Galerian, when mixed feeders predominated over chian-early Galerian and then recovering towards them (Fig. 5). During Aurelian, grazers reach their the middle Pleistocene, in association with a reduc- maximum, while mixed feeders and browsers are tion of mixed feeders. in balance (Fig. 5). Habitat analysis shows that mixed dwellers were Comparison of the Greek data with those taken always poorly represented in Spain in comparison from the analysis of the Italian and Spanish faunal with Greece and Italy, whereas forest dwellers were complexes (Figs 6-9) shows that Ruscinian faunal much less frequent in Greece that in Italy and complex is characterized by a clearly distinct faunal Spain (Figs 10; 11). From the Ruscinian to the structure (even on taxonomic level), whereas dif- late Villafranchian open inhabitants predominate ferences are less pronounced during the successive in Spain, interrupted perhaps by a more forestrial Plio-Pleistocene complexes. phase during the early Villafranchian (Fig. 11). As far as the carnivores are concerned, it seems On the contrary, at the same time in Greece and that they are always more abundant in Spain than Italy forest and mixed dwellers are rather gradu- in Italy and even more in Greece (Figs 4; 6; 8). Th e ally displaced by open land dwellers, reaching a Spanish complexes show that carnivores are almost maximum during the Galerian (Figs 10; 11). From as much numerous as herbivores (early and late Vil- Galerian to Aurelian the entire area shows the same lafranchian) or exceed them (Aurelian), reaching decreasing tendency of open landscapes associated (but not dramatically) their minimum percentage with an increase of mixed dwellers, as a result of during Galerian (Fig. 6). In Italy, carnivores were the climatic mildness and the fragmentation of the more abundant than herbivores only during the environment (Figs 10; 11). early Villafranchian and then their number decreases towards the Galerian, recovering slightly in the Aurelian (Fig. 8). Apart from the general reduction DISCUSSION of carnivore percentage during Galerian, the three regions were also characterised by a less evident Evidently, major faunal renewals and reconstruc- decline during the middle Villafranchian that has tions do not always correspond to “true turnover been followed by a more or less marked increase phases” (i.e. the complete or almost complete re- during the late Villafranchian, corresponding grosso placement of a mammal fauna by a new one) but modo to the “wolf event” of Azzaroli (1983). as the renewal often allows to a change in diversity, Herbivore guild presents quite diff erent trends in we have to deal with an extinction phase or a dis- the Greek, Italian and Spanish succession of faunal persion event. Moreover, a faunal renewal usually complexes (both in the CFCs and BFCs analyses; marks the beginning of a more or less long period Figs 5; 7; 9). Similarly to Greece, a progressive in- of richness increase, which is followed by a period crease of grazers, emphasized in the CFCs analysis of decreasing richness. Accordingly, it seems that (Fig. 7), characterizes all the Villafranchian faunal in some cases an environmental change encourages complexes of Italy, interrupted by a weak Galerian a more effi cient exploitation of resources, which

GEODIVERSITAS • 2007 • 29 (3) 409 Kostopoulos D. S. et al.

gives rise to an increase in mammal diversity and areas indicates highly arid conditions, though it richness. is possible that, due to latitudinal and altitudinal Our results show that at the transition from the diff erences, the development of a mosaic vegeta- Ruscinian to the Villafranchian the marginal parts tion-cover furnishes new available niches. of the studied area (Spain, Greece) are rather af- Around 1.9-1.7 Ma the so-called “wolf event” fected by a true turnover. At the West, this event is (Azzaroli 1983) diff erently aff ected Greek, Italian followed by a distinction phase while in Italy and and Spanish faunal complexes. In the Italian and Greece a diff usion phase occurs, associated with a Greek peninsulas, most Pliocene species disap- predominance of new appearances. Although based peared, particularly among herbivores, and the on insuffi cient data, the obtained results confi rm percentage of ruminant taxa reached lower levels. the relevance of the so-called Equus-Elephant dis- Consequently, the dispersal phase was essentially persal event (Aguirre et al. 1976; Azzaroli 1983; related to the progressive appearance of carnivores. Steininger et al. 1985; Azzaroli et al. 1988; Aguirre Th e hypothesis that the increase in carnivore diver- & Morales 1990) that signifi ed a marked increase sity and their predatory pressure might have con- in diversity due to the high number of fi rst appear- tributed to herbivore decrease cannot be ruled out, ances. Th is event should be related to the double even if an important role might have been played seasonality caused by the 3.5 Ma establishment of by changes in vegetation structure and the conse- cool winters that generated a thermic seasonality, quent presence of more monotonous environments superimposed on the pre-existing North hydric (Torre et al. 2001). Although herbivore standing seasonality (Shackleton 1995; White 1995; Suc et richness dwindled during the latest Pliocene, the al. 1995 and literature therein). increase in large-sized species suggests that the From the early to the late Villafranchian, car- environment could sustain a quite large biomass. nivores decrease in Greece and Italy but remain Moreover, the dominance of mixed feeders might balanced with herbivores and quite abundant in also imply minimal specialisation in restricted Spain, where browsers and grazers decrease with ecological niches. a simultaneous increase of mixed feeders. At the During the early Pleistocene the climatic trend same time in Italy and Greece, grazer’s participa- was characterised by a progressive decrease in mean tion gradually raises together with an elimination of temperatures, even if climatic oscillations were browsers. While open land dwellers predominate in not accentuated. Indeed, the establishment of Spain during this period, forest and mixed dwellers 40 ka climate cyclicity permitted the alternation of are progressively displaced by open land dwellers “warm” steppes under xeric cool temperate climates taxa toward the East. and deciduous forests under humid and warm- Th e apparent – less dramatic – reorganization of temperate climates (Suc et al. 1995; Bertini 2000 the faunal structure from 3.5-2.6 Ma seems to be and references therein). Climate changes are fully related to the occurred onset of bi-polar glaciations, refl ected on the mammal assemblage as a broadly followed by glacial-interglacial cycles of moderate simultaneous early-middle Galerian renewal phase amplitude, sustained at the orbital periodicity of documented in the entire region. During or little 41 ka (Williams et al. 1988; Suc et al. 1995; Shack- after this period, carnivores and browsers reach leton 1995). During this time the Mediterranean their minimum values, whereas open landscapes region experienced a modern climate, which allowed predominate and grazers prevail over other her- the establishment of the current fl oral assemblage. bivore categories. Th is faunal renewal represents an Furthermore, the glacial-interglacial cycles starting essential community reorganization that shows a at 2.6 Ma permitted the institution of alternations total rejuvenation of the fauna, and a general trend between open vegetation of steppe type and decidu- towards larger body size, reduction in the number ous broad-leaved forests (plus Taxodium/Glyptostrobus of medium-sized herbivores, and increase of large type during the Pliocene) (Bertini 1994; Pontini & herbivores and megaherbivores. At that moment, a Bertini 2000). Th e increasing extension of grassy great part of the lineages that constitute the living

410 GEODIVERSITAS • 2007 • 29 (3) Plio-Pleistocene mammal diversity and turnover in North Mediterranean

mammal communities appears, including humans of herbivore standing richness due to the extinction (Th ouveney & Bonifay 1984; Lister 1992; Foley of the Galerian taxa, diversity slightly increased 1994; Turner 1995; Eronen & Rook 2004). because of the appearance of new carnivores and Th e faunal renewal was speeded up by signifi - herbivores, including small or medium sized forms. cant climatic variations with cooler temperatures Furthermore, some new medium-sized taxa ap- and lower humidity, starting around the Jaramillo peared possibly at the beginning of MIS 6, when submagnetozone, and possibly corresponding to more severe, arid climatic oscillations facilitated the the Menapian of the Duct palynilozones (Zagwijn spread of some “mountain” bovids along with the 1992a, b). Th e peaks of last occurrences around dispersion of the “steppe horse” Equus hydruntinus, 1 Ma coincide with low sea levels and cooler-drier and the appearance of a primitive representative of conditions, representing the so-called “Glacial Pleis- Mammuthus primigenius (Palombo 2005). tocene”. Th is event is fairly well documented in the isotope record that changes from low amplitude (40-50 ka quasiperiodic cycles, orbital obliquity) CONCLUSION to lower frequency (100 ka cycles, orbital eccen- tricity) and the higher amplitude nature of the Our inferences are in good agreement with those signal (Williams et al. 1988). Th is cyclic change of previous works in the same line (Alberdi et al. resulted in exceptional shifts of seasonal pattern 1998; Azanza et al. 1999, 2003, 2004; Torre et al. in precipitation and temperature across a range 2001; Palombo 2004, 2005) and complement the of latitudes. Th e change in vegetation prolonged study of the Plio-Pleistocene North Mediterranean global drought and increased seasonal variation; the faunal diversity with respect to time and space. Th e Italian faunal record suggests a consistent decrease succession of local faunal complexes groups, pro- in temperature and humidity (Torre et al. 2001; vided by cluster analysis (CFCs), is characterised Palombo 2005). by chronological ranges that slightly diff er between Th e early-middle Pleistocene faunal renewal takes geographic areas (Palombo et al. 2006). Th is could its origin in two distinct trends: 1) the progres- suggest that time and pattern of renewal phases sive reduction in richness characterising late early also diff er for each region. However, diff erences Pleistocene faunas (during which last appearances in the LFA richness, diversity of localities in local prevailed); and 2) the subsequent dispersal related, scale and disparity of localities per cluster might above all, to the progressive diff usion of taxa from aff ect the analysis, consenting to overestimation of Eastern and Central Europe that led to the im- diff erences in time and overall pattern. portant and relatively rapid increase in diversity Since the particular composition of a regional (due fi rstly to immigration and secondly to local faunal complex is the result of multiple ecological origination bioevents) occurring at the beginning parameters, as well as, of historical factors, some of the middle Pleistocene. minor diff erences can be discerned between the From the middle Galerian to the Aurelian an West and the East of the North Mediterranean increase in extinctions and richness occurs, associ- region. From this point of view, the Greek Rus- ated with an increase in carnivore taxa. Browsers cinian and Villafranchian turnover and diversity recover in Greece and Italy while in Spain grazers patterns and the large mammal guilds are more enhance again. In a general way, open landscape closely related to the Italian that to the Spanish dwellers decline with a contemporaneous increase ones. During this time-interval, open country taxa of ubiquitous large mammals. Italian, Greek and gradually substitute forest dwellers in the Greek Spanish faunal complexes, belonging to the early and Italian faunal complexes, while in Spain, open Aurelian show that the most fl exible species be- land mammals prevailed since the beginning of the came the predominant groups, browser percent- Villafranchian, at least. Th e present work stresses age increased and the fauna progressively acquired out, however, that signifi cant faunal renewals and modern characteristics. Moreover, after the decrease reorganizations, already recognized in Spain and

GEODIVERSITAS • 2007 • 29 (3) 411 Kostopoulos D. S. et al.

Italy at the transition from the Villafranchian to ropean Plio-Pleistocene: some remarks. Th ird French the Galerian (according to BFCs analysis) or be- Symposium on Stratigraphy, Lyon: 25-26. AZANZA B., PALOMBO M. R. & ALBERDI M. T. 2003. — tween the early and the middle Galerian (accord- Similarity relationships between large mammal faunas ing to CFCs analysis) (Azanza et al. 1999, 2004; of Spanish and Italian peninsulas from the latest Mi- Palombo 2004), are well refl ected on the Greek ocene to the middle Pleistocene. Neues Jahrbuch für faunal complex either. Th is early Galerian faunal Geologie und Paläontologie, Abh. 229 (1): 95-127. renovation seems, therefore, to aff ect more or less AZANZA B., PALOMBO M. R. & ALBERDI M. T. 2004. — simultaneously the entire North Mediterranean Large mammal turnover and diversity from the Pliocene to the Pleistocene in Italian peninsula. Rivista Italiana region, which responds rather uniformly to the di Paleontologia e Stratigrafi a 110 (2): 531-545. establishment of cooler-dryer climatic conditions AZZAROLI A. 1983. — Quaternary mammals and the around Jaramilo magnetochron (1 Ma). “End-Villafranchian” dispersal event – A turning point in the history of Eurasia. Palaeogeography Pal- aeoclimatology Palaeoecology 44: 117-139. AZZAROLI A., DE GIULI C., FICCARELLI G. & TORRE D. Acknowledgements 1988. — Late Pliocene to early Mid-Pleistocene We are grateful to Dr. B. Azanza (Madrid) and an mammals in Eurasia: faunal succession and dispersal anonymous referee for their valuable criticism and events. Palaeogeography Palaeoclimatology Palaeoecol- comments on the manuscript. ogy 66: 77-100. BEHRENSMEYER A. K., TODD N. E., POTTS R. & MC- BRINN G. E. 1997. — Late Pliocene faunal turnover in the Turkana Basin, Kenya and Ethiopia. Science REFERENCES 278: 1589-1594. BERTINI A. 1994. — Palynological investigations on up- AGUIRRE E. & MORALES J. 1990. — Villafranchian faunal per and lower Pleistocene sections in central record of Spain. Quartärpaläontologie 8: 7-11. and northern Italy. Memorie della Società Geologica AGUIRRE E., LOPEZ N. & MORALES J. 1976. — Con- Italiana 48: 431-443. tinental faunas in Southeast Spain related to the BERTINI A. 2000. — Pollen record from Colle Curti and , in Messinian Seminar, 2: Il signifi cato Cesi: early and middle Pleistocene mammal sites in Geodinamico della crisi di Salinità del Miocene termi- the Umbro-Marchean Apennine Mountains (central nale nel Mediterraneo. Messinian Seminar, Gargnano, Italy). Journal of Quaternary Science 15: 825-840. September 5-12, 1976: 62-63. BRETT C. E., IVANY L. C. & SCHOPF K. M. 1996. — Co- ALBERDI M. T., AZANZA B., CERDEÑO E. & PRADO J. L. ordinated stasis: an overview. Palaeogeography Palaeo- 1997. — Similarity relationship between Mammal climatology Palaeoecology 127: 1-20. faunas and biochronology from latest Miocene to ERONEN J. & ROOK L. 2004. — Th e Mio-Pliocene Eu- Pleistocene in the Western Mediterranean area. Ec- ropean primate fossil record: dynamics and habitat logiae Geolica Helvetiae 90 (1): 115-132. tracking. Journal of Human Evolution 47: 323-341. ALBERDI M. T., CALOI L., DUBROVO I., PALOMBO M. R. FOLEY R. A. 1994. — Speciation, extinction and climatic & TSOUKALA E. 1998. — Large mammal faunal com- change in hominid evolution. Journal of Human Evo- plexes and palaeoenvironmental changes in the late lution 26: 275-289. middle and late Pleistocene: a preliminary comparison FORTELIUS M., ERONEN J., LIU L., PUSKINA D., TESAKOV between the Eastern European and the Mediterranean A.,VISLOBOKOVA I. & ZHANG Z. 2006. — Late Mi- areas. Geologija 25: 8-19. ocene and Pliocene large land mammals and climatic AZANZA B., ALBERDI M. T., CERDEÑO E. & PRADO changes in Eurasia. Palaeogeography Palaeoclimatology J. L. 1997. — Biochronology from latest Miocene Palaeoecology 238: 219-227. to middle Pleistocene in the Western Mediterranean GLIOZZI E., ABBAZZI L., AMBROSETTI P. , A RGENTI P. , A Z- area. A multivariate approach. Mémoires et Travaux ZAROLI A., CALOI L., CAPASSO BARBATO L., DI STEFANO de l’EPHE, Montpellier 21: 567-574. G., ESU D., FICCARELLI G., GIROTTI O., KOTSAKIS AZANZA B., ALBERDI M. T. & PRADO J. L. 1999. — Lar- T., MASINI F. , M AZZA P., MEZZABOTTA C., PALOMBO ge mammal diversity and turnover patterns during M. R., PETRONIO C., ROOK L., SALA B., SARDELLA the Plio-Pleistocene in Western Mediterranean Area. R., ZANALDA E. & TORRE D. 1997. — Biochronol- Revista de la Sociedad Geológica de España 12 (1): ogy of selected mammals, molluscs, ostracods from 113-122. the middle Pliocene to the late Pleistocene in Italy. AZANZA B., PALOMBO M. R. & ALBERDI M. T. 2002. — Th e state of the art. Rivista Italiana di Paleontologia Mammalian biochronological framework of the Eu- e Stratigrafi a 103 (3): 369-388.

412 GEODIVERSITAS • 2007 • 29 (3) Plio-Pleistocene mammal diversity and turnover in North Mediterranean

GUERIN C. 1990. — Biozones or mammal units? Methods (Tiberino basin). Geobios 33 (5): 519-526. and limits in biochronology, in LINDSAY E. H., FAHL- SHACKLETON N. J. 1995. — New data on the evolu- BUSCH V. & M EIN P. (eds), European Neogene Mammal tion of Pliocene climatic variability, in VRBA E. S., Chronology. Plenum Press, New York: 119-130. DENTON G. H., PARTRIDGE T. C. & BURCKLE L. H. HARPER C.W. JR. 1975. — Standing diversity of fossil (eds), Paleoclimate and Evolution, with Emphasis on group in successive intervals of geologic time: a new Human Origins. Yale University Press, New Haven: measure. Journal of Paleobiology 4: 181-207. 242-248. KOSTOPOULOS S. D., VASSILIADOU K. & KOUFOS G. D. SPASSOV N. 2003. — Th e Plio-Pleistocene vertebrate 2002. — Th e beginning of Pleistocene in the Balkan fauna in South-Eastern Europe and the megafaunal area according to the mammal record; palaeozooge- migratory waves from the east to Europe. Revue ographic approach. Annales géologiques des Pays hél- paléobiologique 22 (1): 197-229. leniques 39 A: 253-278. STEININGER F. F. , R ABETER G. & ROEGL F. 1985. — Land LISTER A. 1992. — Mammalian fossils and Quater- mammal distribution in the Mediterranean Neogene: nary biostratigraphy. Quaternary Science Reviews 11: a consequence of geokinematic and climatic events, in 329-344. STANLEY D. J. & WEZEL F. C. (eds), Evolution of the MEIN P. 1990. — Updating of MN zones, in LINDSAY Mediterranean Basin. Raimondo Selli Commemoratio, E. H., FAHLBUSCH V. & M EIN P. (eds), European Springer Verlag: 559-571. Neogene Mammal Chronology. Plenum Press, New SUC J.-P., BERTINI A., COMBOURIEU-NEBOUT N., DI- York: 73-90. NIZ F. , L EROY S., RUSSO-ERMOLLI E., ZHENG Z., PALOMBO M. R. 2004. — Guilds of large mammals BESSAIS E. & FERRIER J. 1995. — Structure of West from the Pliocene to the late Pleistocene in the Ital- Mediterranean vegetation and climate since 5.3 Ma. ian peninsula, in BAQUEDANO E. & RUBIO S. (eds), Acta Zoologica Cracoviensia 38: 3-16. Homenaje a Emiliano Aguirre. Museo Archeologico TEDFORD R. H. 1970. — Principles and practices of Regional, Madrid, Zona Archaeologica 4 (2 Palae- mammalian geochronology in North America. Ameri- ontologia): 20-29. can Paleontological Convention, Chicago 1969, Proc- PALOMBO M. R. 2005. — Middle Pliocene-late Pleis- ceedings F: 666-703. tocene herbivore guilds of Italy, in CRÉGUT-BONNOURE TORRE D., ROOK L. & FICCARELLI G. 1999. — Th e E. (ed.), Les ongulés holarctiques du Pliocène et du mammal fauna turnover in Italy at the early to mid- Pléistocène. Quaternaire, hors série 2: 123-136. dle Pleistocene transition, in GIBERT G., SANCHEZ F. , PALOMBO M. R. & VALLI A. M. F. 2004. — Similarities GIBERT L. & RIBOT F. (eds), Th e Hominids and their between large mammal faunas of the Italian Peninsula Environment During the Lower and Middle Pleistocene and France from the Pliocene to the middle Pleistocene. of Eurasia. Museo de Prehistoria y Paleontologia Neues Jahrbuch für Geologie und Paläontologie, Abh. “J. Gibert”, Orce: 541-548. 233 (1): 69-102. TORRE D., ABBAZZI L., BERTINI A., FANFANI F. , PALOMBO M. R., AZANZA B. & ALBERDI M. T. 2003. — FICCARELLI G., MASINI F. , M AZZA P. & R OOK L. Italian mammal biochronology from latest Miocene to 2001. — Structural changes in Italian late Pliocene- middle Pleistocene: a multivariate approach. Geologica Pleistocene large mammal assemblages, in ROOK Romana 36: 335-368. L. & TORRE D. (eds), Neogene and Quaternary PALOMBO M. R., VALLI A. F. M., KOSTOPOULOS D. S., continental stratigraphy and mammal evolution. ALBERDI M.T., SPASSOV N. & VISLOBOKOVA I. 2006. — Bollettino della Società Paleontologica Italiana 40 Similarity relationships between the Pliocene to middle (2): 303-306. Pleistocene large mammal faunas of Southern Europe TURNER A. 1995. — Plio-Pleistocene correlations be- from Spain to the Balkans and the North Pontic re- tween climatic change and evolution in terrestrial gion, in KAHLKE R.-D., MAUL L. C. & MAZZA P. P. A . mammals. Th e 2.5 Ma event in Africa and Europe. (eds), Late Neogene and Quaternary biodiversity and Acta Zoologica Cracoviensia 38: 45-58. evolution: regional developments and interregional THOUVENEY N. & BONIFAY E. 1984. — New chrono- correlations. Courier Forschungsinstitut Senckenberg logical data on European Plio-Pleistocene faunas and 256: 329-347. hominid occupation sites. Nature 308: 355-358. PICKFORD M. 1990. — Dynamics of Old World biogeo- VAN VALEN L. 1973. — A new evolutionary law. Evo- graphic realms during the Neogene: implications for lutionary Th eory 1: 1-30. biostratigraphy, in LINDSAY E. H., FAHLBUSCH V. & VRBA E. S. 1985. — Environment and evolution: al- MEIN P. (eds), European Neogene Mammal Chronol- ternative causes of the temporal distribution of evo- ogy. Plenum Press, New York: 413-442. lutionary events. South African Journal of Sciences PONTINI M. R. & BERTINI A. 2000. — Late Pliocene 81: 229-236. vegetation and climate in central Italy: high resolu- VRBA E. S. 1993. — Turnover pulse, the Red Queen, tion pollen analysis from the Fosso Bianco succession and related topics. American Journal of Science 293:

GEODIVERSITAS • 2007 • 29 (3) 413 Kostopoulos D. S. et al.

418-452. New Haven: 369-384. VRBA E. S. 1995. — On the connections between paleo- WILLIAMS D. F., THUNELL R. C., TAPPA E., RIO D. & climate and evolution, in VRBA E. S., DENTON G. H., RAFFI I. 1988. — Chronology of the Pleistocene oxy- PARTRIDGE T. C. & BURCKLE L. H. (eds), Paleoclimate gen isotope record: 0-1.88 m.y. B.P. Palaeogeography and Evolution, with Emphasis on Human Origins. Yale Palaeoclimatology Palaeoecology 64: 221-240. University Press, New Haven: 24-48. ZAGWIJN W. H. 1992a. — Th e beginning of the Ice Age WHITE T. D. 1995. — African omnivores: global climatic in Europe and its major subdivisions. Quaternary change and Plio-Pleistocene hominids and suids, in Science Review 2: 538-591. VRBA E. S., DENTON G. H., PARTRIDGE T. C. & ZAGWIJN W. H. 1992b. — Migration of vegetation during BURCKLE L. H. (eds), Paleoclimate and Evolution with the Quaternary in Europe. Courier Forschungsinstitut Emphasis on Human Origins. Yale University Press, Senckenberg 153: 9-20.

Submitted on 16 December 2005; accepted on 11 April 2007.

414 GEODIVERSITAS • 2007 • 29 (3) Plio-Pleistocene mammal diversity and turnover in North Mediterranean

CFC BFC 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% Total number of species

Total number of species 20% 10% 10% 0% 0% R-eV mV IV-eG mG IG-A ReVmVIV G A

100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% Standing richness 20% Standing richness 20% 10% 10% 0% 0% R-eV mV IV-eG mG IG-A ReVmVIV G A

% Frugivorous/Omnivorous % Carnivores % Herbivores

FIG. 4. — Distribution of all large mammal diet categories in Greek cluster (CFC) and biochronological (BFC) faunal complexes. Ab- breviations as in Fig. 1.

CFC BFC 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% Total number of species 10% Total number of species 10% 0% 0% R-eV mV IV-eG mG IG-A ReVmVIV G A

100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% Standing richness Standing richness 20% 20% 10% 10% 0% 0% R-eV mV IV-eG mG IG-A ReVmVIV G A

% Browsers % Mixed feeders % Grazers

FIG. 5. — Distribution of large mammal herbivore categories in Greek cluster (CFC) and biochronological (BFC) faunal complexes. Abbreviations as in Fig. 1.

GEODIVERSITAS • 2007 • 29 (3) 415 Kostopoulos D. S. et al.

CFC BFC 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% Total number of species Total number of species 0% 0% R-eV m-IV IV-eG m-IG eA ReVmVIV GeA

100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% Standing richness Standing richness 10% 10% 0% 0% R-eV m-IV IV-eG m-IG eA ReVmVIV GeA

% Frugivorous/Omnivorous % Carnivores % Herbivores

FIG. 6. — Distribution of all large mammal diet categories in Italian cluster (CFC) and biochronological (BFC) faunal complexes. Ab- breviations as in Fig. 1.

CFC BFC 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% Total number of species 10% Total number of species 10% 0% 0% R-eV m-IV IV-eG m-IG eA ReVmVIV GeA

100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% Standing richness Standing richness 20% 20% 10% 10% 0% 0% R-eV m-IV IV-eG m-IG eA ReVmVIV GeA

% Browsers % Mixed feeders % Grazers

FIG. 7. — Distribution of large mammal herbivore categories in Italian cluster (CFC) and biochronological (BFC) faunal complexes. Abbreviations as in Fig. 1.

416 GEODIVERSITAS • 2007 • 29 (3) Plio-Pleistocene mammal diversity and turnover in North Mediterranean

CFC BFC 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20%

Total number of species 10% 10% Total number of species 0% 0% R e-mV IV G A ReVmVIV G A

100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% Standing richness Standing richness 20% 20% 10% 10% 0% 0% R e-mV IV G A ReVmVIV G A

% Frugivorous/Omnivorous % Carnivores % Herbivores

FIG. 8. — Distribution of all large mammal diet categories in Spanish cluster (CFC) and biochronological (BFC) faunal complexes. Ab- breviations as in Fig. 1.

CFC BFC 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% Total number of species Total number of species 0% 0% R e-mV IV G A ReVmVIV GA

100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% Standing richness Standing richness 10% 10% 0% 0% R e-mV IV G A ReVmVIV G A

% Browsers % Mixed feeders % Grazers

FIG. 9. — Distribution of large mammal herbivore categories in Spanish cluster (CFC) and biochronological (BFC) faunal complexes. Abbreviations as in Fig. 1.

GEODIVERSITAS • 2007 • 29 (3) 417 Kostopoulos D. S. et al.

Greece

100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% Standing richness 20% 20% Total number of species 10% 10% 0% 0% ReVmVIV GA ReVmVIV GA Italy

100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% Standing richness 20% 20% Total number of species 10% 10% 0% 0% ReVmVIV GeAReVmV IV GeA Spain 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% Standing richness 20% Total number of species 10% 10% 0% 0% ReVmVIV GA ReVmVIV GA

Forest dwellers Mixed dwellers Open land dwellers

FIG. 10. — Distribution of large mammal categories according to their habitat type in Greek, Italian and Spanish biochronological faunal complexes. Abbreviations as in Fig. 1.

418 GEODIVERSITAS • 2007 • 29 (3) Plio-Pleistocene mammal diversity and turnover in North Mediterranean

Greece

100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% Standing richness 20% 20% Total number of species 10% 10% 0% 0% R-eV m-IV IV-eGm-IG eA R-eV m-IV IV-eGm-IG eA Italy

100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40%

30% Standing richness 30%

Total number of species 20% 20% 10% 10% 0% 0% R-eV m-IV IV-eGm-IG eA R-eV m-IV IV-eGm-IG eA Spain

100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30%

30% Standing richness 20%

Total number of species 20% 10% 10% 0% 0% R e-mV IVG A R e-mV IVG A

Forest dwellwers Ubiquitous Open landscape dwellers

FIG. 11. — Distribution of large mammal categories according to their habitat type in Greek, Italian and Spanish cluster faunal com- plexes. Abbreviations as in Fig. 1.

GEODIVERSITAS • 2007 • 29 (3) 419