Palaeogeography, Palaeoclimatology, Palaeoecology 449 (2016) 455–462

Contents lists available at ScienceDirect

Palaeogeography, Palaeoclimatology, Palaeoecology

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

Latitudinal gradients and indicator species in ungulate paleoassemblages during the MIS 3 in W Europe

Diego J. Álvarez-Lao a,⁎,MarcosMéndezb a Departamento de Geología, Universidad de Oviedo, C/ Jesús Arias de Velasco s/n., E-33005 Oviedo, b Área de Biodiversidad y Conservación, Universidad Rey Juan Carlos, c/ Tulipán s/n., E-28933 Móstoles, Madrid, Spain article info abstract

Article history: Studies of mammal assemblages for the European MIS 3 have focused on a regionalization using presence– Received 9 October 2015 absence data and traditional cluster analysis. Although this approach has revealed mammal assemblages with Received in revised form 6 February 2016 no modern analogs, it can easily miss two important dynamical aspects of the MIS 3 assemblages, namely, Accepted 23 February 2016 transitional regions and faunal mixtures. We revisited this regionalization using a complementary approach Available online 3 March 2016 based on the study of quantitative data for ungulate species. The main aim of this work is to evaluate whether Keywords: the MIS 3 climatic oscillations led to a mixture of or a latitudinal gradient in faunal composition, as Cold-adapted large mammals as to identify indicator species for different areas in ungulate mammal paleoassemblages of Western Europe. Indicator value Ungulate faunal compositions of 86 fossil assemblages from 63 Western European sites of MIS 3 age were gathered from published sources. The distribution area was divided into four regions: central W Europe, S Mammal palaeoecology , Cantabrian region and rest of the Iberian Peninsula. Chronologies of the assemblages were collected Western Palearctic and calibrated. Multivariate ordination and indicator value analysis were carried out. Community analysis Our study provided four main results: (i) there was a latitudinal faunal turnover during the MIS 3 period consis- tent with a temperature gradient; (ii) the Cantabrian region hosted a transitional faunal assemblage between the northern cold faunas and the southern temperate faunas; (iii) it was possible to find indicator species of large herbivores for different geographical areas along that latitudinal gradient, although not all of them were linked to climate but to orography (mainly in Cantabrian and other Iberian sites); and (iv) within those indicator spe- cies, a signal of mixed temperate and cold faunas was discernible. Together, this evidence supports a scenario in which the succession of cold and warm episodes typical from MIS 3 left signals of mixed cold and temperate faunal assemblages but without totally blurring the latitudinal gradient in climatic conditions. © 2016 Elsevier B.V. All rights reserved.

1. Introduction between mammal assemblages and climate could be blurred based on several paleoecological considerations. First, late Pleistocene faunal and The distribution patterns of large herbivore mammals changed environmental heterogeneity was greater than that of the throughout the Pleistocene in response to glaciation-related climate and, consequently, many Late Pleistocene communities do not have changes. This led to the latitudinal assemblage of mammal faunas ac- modern analogs (Graham et al., 1996; Stewart, 2005, 2008). Second, cording to climatic tolerance of different species. Cold-adapted mammal due to vagaries of climate, migrating species responding individualisti- faunas were predominant during the cold stages in middle and high lat- cally to environmental change (Stewart, 2008) were mixed so that itudes of Eurasia and North America (Kahlke, 1999; Stewart et al., 2003; distinct assemblages could be obscured (Graham et al., 1996). Third, Markova et al., 2010), reaching southern latitudes only during the species, taken separately, could have a reduced value as indicators of coldest events (Álvarez-Lao et al., 2009; Álvarez-Lao and García, 2011a, climatic conditions (Stewart, 2005; De Cáceres et al., 2012). Sommer et al., 2014). Temperate mammal faunas, which were predom- These considerations were potentially prevalent during the Marine inant in most of the Holarctic during the Pleistocene interglacial episodes Isotope Stage (MIS 3), an episode of the Late Pleistocene from 60 to (Kurtén, 1968; Stuart, 1982; Guérin and Patou-Mathis, 1996), persisted 27 ka, placed between the two cold maxima of the last glaciation (MIS in southern glacial refugia (Iberian, Italian and Balkan peninsulas) during 4 and 2) (van Andel, 2002; Roucoux et al., 2005, Clark et al., 2009). the cold stages (Hewitt, 2000; Sommer and Nadachowski, 2006; Climatically, MIS 3 was characterized by several sharp oscillations Jennings et al., 2011; Stewart and Stringer, 2011). However, a clear link between cold stadials and temperate interstadials (Dansgaard et al., 1993; Barron and Pollard, 2002; North GRIP Members, 2004; Van ⁎ Corresponding author. Tel.: +34 985103143; fax: +34 985103103. Meerbeeck et al., 2011). In most cases, rapid decadal warming episodes E-mail address: [email protected] (D.J. Álvarez-Lao). of 8–16 °C were followed by several centuries of gradual cooling,

http://dx.doi.org/10.1016/j.palaeo.2016.02.050 0031-0182/© 2016 Elsevier B.V. All rights reserved. 456 D.J. Álvarez-Lao, M. Méndez / Palaeogeography, Palaeoclimatology, Palaeoecology 449 (2016) 455–462 reaching stadial temperature values (North GRIP Members, 2004; value, which can be more powerful than subjective expert assessment Huber et al., 2006; Van Meerbeeck et al., 2011). or previous statistical tools in these scenarios of mixed assemblages. Although MIS 3 is considered a climatically mild episode by some Moreover, quantitative data are used since they provide a more accurate authors (van Andel, 2002), the fossil record of large herbivore mammals expression of the faunal differences between nearby geographical areas from North and Middle Europe shows a predominance of cold-adapted than mere presence–absence records. mammals (e.g. Rangifer tarandus, Mammuthus primigenius and Coelodonta antiquitatis) during most of this episode (Delpech, 1983, 1984; Turner, 2. Material and methods 2000; Münzel and Conard, 2004a, 2004b; David et al., 2005; Schreve, 2006). In addition, cold-adapted mammals were detected throughout 2.1. Study sites and species MIS 3 even as far South as the Iberian and Italian peninsulas (García and Arsuaga, 2003; Palombo and Ferretti, 2005; Álvarez-Lao et al., 2009; A total of four regions were considered in this study. Following Álvarez-Lao and García, 2010, 2011a, 2012; Álvarez-Lao, 2014). In fact, Stewart (2005), we considered Central Western Europe, a transition during MIS 3, the woolly mammoth reached its southernmost geograph- zone including Southern France and the Cantabrian region of the Iberian ical distribution both in Europe (Álvarez-Lao et al., 2009) and Asia Peninsula, and a Southern region including the rest of the Iberian Penin- (Takahashi et al., 2007). Therefore, mixed faunal assemblages seem to sula. For our analysis, we treated the Cantabrian region and Southern have occurred during MIS 3, potentially obscuring a paleoecological France separately, to take into account the differences in topography reconstruction of this period. among these two regions. Using a literature search, we compiled 86 Studies of mammal assemblages for the European MIS 3 (Stewart faunal assemblages of MIS 3 age from 63 sites: 19 assemblages from et al., 2003; Markova et al., 2010) have focused on a regionalization 14 sites in Central Western Europe, 17 assemblages from nine sites in using traditional presence–absence based cluster analysis (Kreft and Southern France, 34 assemblages from 25 sites in the Cantabrian region, Jetz, 2010). This approach offers valuable insights into the composition and 16 assemblages from 14 sites in the rest of the Iberian Peninsula of the mammal assemblages, including those with no modern analogs. (see Appendix A in Supporting Information). These sites covered a However, it can easily miss two important dynamical aspects of the wide latitudinal gradient from central Western Europe to southern MIS 3 assemblages, namely, transitional regions and faunal mixtures. Spain (Fig. 1). Dynamics of faunal assemblages has been shown for North America The studied species are associated with different environments (Graham et al., 1996), or for single species in Europe (Sommer et al., based on paleontological and current ecological evidences: R. tarandus, 2011, 2014). Current regionalization of W Europe during MIS 3 Saiga tatarica, Ovibos moschatus, C. antiquitatis and M. primigenius are considers that the Cantabrian region in northwestern Spain belongs to considered indicative of cold conditions and open landscapes; Capra the same assemblage than the rest of the Iberian Peninsula (Markova ibex/C. pyrenaica and Rupicapra rupicapra indicate rocky environments; et al., 2010; see, however, Stewart, 2005). This assessment ignores Capreolus capreolus, Dama dama and Sus scrofa are normally associated the biogeographical peculiarity of this area during this period. with forested landscapes while Megaloceros giganteus, Equus caballus, The Cantabrian region is a belt of land of nearby 300 km long and Equus hydruntinus and Stephanorhinus hemitoechus indicate open and 30–50 km wide where short river valleys run perpendicular to the herbaceous areas; Cervus elaphus and the bovines (Bison priscus and coast. Topography is highly rough and, consequently, large fluvial plains Bos primigenius)areflexible species in their palaeoclimatic parameters are lacking. While the Iberian Peninsula is relatively isolated from and not specifically indicative of any particular landscape (Kurtén, continental Europe by the , which acted as a barrier limiting 1968; Corbet and Ovenden, 1980; Guthrie, 1982; Vereschagin and faunal dispersals, a corridor west of the Pyrenees, that opened during Baryshnikov, 1982; Guérin and Patou-Mathis, 1996; Kahlke, 1999, the Late Pleistocene's coldest stages, allowed the entrance of large 2014). mammal faunas from continental Europe into the Cantabrian region In previous studies on MIS 3 faunal assemblages, presence–absence (Álvarez-Lao and García, 2011a, 2011b). In addition, this region has rel- analyses have been mainly utilised (Stewart et al., 2003; Markova et al., ative isolation from the Iberian areas further south by the Cantabrian 2010). However, ignoring differences in abundance between assem- Mountain Range, which also acted as a barrier for large mammal faunas blages leads to loss of valuable information. Therefore, we took a quan- (Álvarez-Lao and García, 2011a). Consequently large mammal faunas titative approach using percentages of species, based on the number of from this area are highly influenced, not only by latitude, but also by identified specimens (NISP). Controversy exists on the suitability of topography. Moreover, pollen and marine isotope data indicate that NISP compared to MNI (minimum number of individuals) as quantita- environmental conditions recorded in at NW Iberia were significantly tive estimates in paleoassemblages (Lyman, 2008). Both units show ad- cold and dry during the MIS 3 coldest events (Lézine and Denèfle, vantages and drawbacks. First, NISP is an observed or direct measure 2003; Roucoux et al., 2005). The recent improvement in the Cantabrian while MNI is a derived measure that can be calculated by different MIS 3 archaeo-paleontological record (Altuna, 1996, 2004; Altuna and (more or less comprehensive) methods and, therefore, more subjective Mariezkurrena, 2000, 2010, 2011; Castaños, 1996; Castaños and and less comparable across studies (see a detailed discussion in Lyman, Castaños, 2007; Castaños et al., 2009; Álvarez-Lao and García, 2010, 2008). Second, MNI data are not always provided in the literature. Final- 2011a, 2011b, 2012; Álvarez-Lao, 2014) currently allows a reassess- ly, Lyman (2008) showed that there is a strong correlation between ment of the Cantabrian region faunal assemblages. NISP and MNI. In our case, MNI data were available for only 33 assem- The main goal of this work is to compare the ungulate assemblages blages (6 from W Europe, 1 from S France, 21 Cantabrian and 5 from along the latitudinal gradient from Central Western Europe and South other Iberian sites). A high correlation was found among total NISP France to the Cantabrian region and the rest of the Iberian Peninsula and total MNI across sites (r = 0.70, p b 0.001). From 13 species, for during the MIS 3. Our specific questions are: (1) Did the climatic oscilla- which both NISP and MNI were available, a significant correlation was tions during MIS 3 lead to a mixture of faunas or was a latitudinal found for nine species present at nine or more sites (r ≥ 0.65, (temperature) gradient still recognizable in faunal composition? We p ≤ 0.003). The remaining four species had nine or less data and there- predict that this gradient should be detected, with Cantabrian sites fore correlations were nonsignificant due to low sample sizes. Overall, showing an intermediate faunal assembly compared to northern and we preferred to use NISP, which was available for a broader set of southern sites. (2) Can some species be considered as indicators of assemblages. some faunal assemblages along this latitudinal gradient? We predict Within each site, we selected faunal assemblages from layers with that faunal mixture would decrease the value of single species as indica- NISP higher than 100. When sites had several layers containing faunal tors of particular areas along the latitudinal gradient. We take advantage assemblages meeting this criterium, two or three layers were consid- of new statistical tools from community ecology to assess indicator ered, in order to capture temporal variation in faunal assemblages. In D.J. Álvarez-Lao, M. Méndez / Palaeogeography, Palaeoclimatology, Palaeoecology 449 (2016) 455–462 457

Fig. 1. Map of Western Europe with the location of the sites included in the study. Cantabrian sites: 1. Liñares; 2. Cueva del Hueso; 3. Cueva del Conde; 4. Balmori; 5. La Riera; 6. Jou Puerta; 7. Esquilleu; 8. El Castillo; 9. Covalejos; 10. Morín; 11. El Ruso; 12. Covacho de Arenillas; 13. El Cuco; 14. Arrillor; 15. Lumentxa; 16. Santimamiñe; 17. Bolinkoba; 18. Urtiagako Leizea; 19. Ekain; 20. Amalda; 21. Aitzbitarte III; 22. Aitzbitarte IV; 23. ; 24. Labeko Koba; 25. Lezetxiki. Other Iberian sites: 26. Prado Vargas; 27. Peña Miel; 28. Cueva Millán; 29. Ermita; 30. Gabasa; 31. Arbreda; 32. Casares; 33. Cova Beneito; 34. Les Cendres; 35. Zafarraya; 36. Nerja; 37. Columbeira; 38. Pego do Diablo; 39. Figueira Brava. South France sites: 40. Isturitz; 41. Tournal; 42. Abri des Battuts; 43. Le Piage; 44. Roc de Combe; 45. Grotte XVI; 46. ; 47. ; 48. Le Flageolet 1. Central Western Europe sites: 49. Saint-Césaire; 50. Theillat; 51. Arcy-Sur-Cure, gr. Renne; 52. Arcy-Sur-Cure, gr. Bison; 53. Maisières-Canal; 54. Trou Magritte; 55. Spy; 56. Goyet; 57. Huccorgne-Hermitage; 58. ; 59. Geissenklösterle; 60. Brillenhöhle; 61. Vogelherd; 62. Paviland, Goat's Hole ; 63. Lynford. ten cases we included assemblages with NISP lower than 100, for or electron spin resonance) have also been considered. Results are completeness. Exclusion of these assemblages did not qualitatively expressed in cal. ka BP (calibrated kiloanne before present). For assem- modify our conclusions, so we kept them in our data set. blages dated by association with archaeological contexts, we followed All the analyzed assemblages but three correspond to archaeological the time spans inferred for each “technocomplex” by Rasilla (1994), contexts. We also performed the analyses excluding these three assem- Arrizabalaga (2000, 2005), Cabrera Valdés et al. (2002) and Maroto blages corresponding to paleontological contexts (carnivore activity or et al. (2005) for N Iberian Peninsula, and Delporte (1991), Bosselin and natural traps) and the pattern did not change. Consequently, we includ- Djindjian (1994) and Djindjian and Bosselin (1994),forFrance.Although ed them in the data set. the MIS 3 covers a wide time span, from approximately 27 to 60 ka, the A total of 16 species of ungulates were included in our data set age of most of the assemblages included in our study is younger than (Appendix A). Bos–Bison were pooled due to the impossibility in 45 ka (see Appendix B). Older assemblages were unreliable to date by distinguishing them in most cases. Our analyses focused on ungulates the radiocarbon method, so normally they are not absolute dated with because they are the most environmentally indicative large mammals the exception of a few of them dated by U-Th or ESR methods. (Guthrie, 1982) and they are usually well represented into Pleistocene fossil assemblages. 2.2. Statistical analysis Chronologically, all the assemblages are either absolutely dated or in association with a well-defined archaeological context. All radiocarbon A correspondence analysis (CA) was performed with CANOCO 4.5 dates (except those previously calibrated in the literature) have been (Lepš and Šmilauer, 2003). CA is a multivariate method to reduce calibrated by means of the CalPal software (Weninger et al., 2008)to dimensionality in community (species per sites) data based on 1σ (68% range) in order to obtain calibrated ages that could be correlated Euclidean dissimilarity in species composition and abundance between with the palaeoclimatic scale GRIP. Absolute dates by methods other sites. It is suitable when the response of species to the environmental than 14C (i.e. U-Th series, amino acid racemization, thermoluminescence gradient is unimodal rather than linear, i.e. a high species turnover 458 D.J. Álvarez-Lao, M. Méndez / Palaeogeography, Palaeoclimatology, Palaeoecology 449 (2016) 455–462 between sites is present (Lepš and Šmilauer, 2003). CA output is an or- to the presence of D. dama, S. hemitoechus and R. rupicapra (Table 1). dination of sites in a reduced (optimally two or three) dimensional Axis 2 was positively related to M. primigenius and negatively related space of dissimilarity that allows to detect not only groups of sites to O. moschatus and R. tarandus (Table 1). Axis 3 was positively related with similar faunal composition but also continuous gradients in faunal to C. ibex/C. pyrenaica (Table 1). Sites were arranged in the ordination change across sites (Lepš and Šmilauer, 2003). This method was thus space in a smooth gradient (Fig. 2). 1 and 2 showed a continuous suitable to reveal the potential continuous gradient between sites. transition from S France and other W European sites to Cantabrian Coding D. dama and O. moschatus, present in a single site each, as passive and the rest of Iberian sites (Fig. 2). Axes 2 and 3 further separated, al- species did not change the results qualitatively, so the results of the CA though less clearly, Cantabrian from the rest of Iberian sites (Fig. 2). including all species are presented. Only the ordination axes explaining Four Cantabrian assemblages fell outside this pattern: Esquilleu III, more than 10% of the variance in species data were retained. For the in- Covacho de Arenillas II and Bolinkova VI clustered with the other Iberian terpretation of the ordination axes, only those species with scores ≥0.7 assemblages due to high percentages of C. ibex/C. pyrenaica remains; were considered. Urtiagako Leizea clustered with S France and W Europe assemblages An analysis of indicator species was carried out following Dufrêne due to a high percentage of R. tarandus. and Legendre (1997) and De Cáceres et al. (2010, 2012). Species are The analysis identified indicator species for each of the four areas considered as good indicators of a region if they are restricted to that considered (Fig. 3). Nevertheless, consideration of the possibility region and are frequent within it (Dufrêne and Legendre, 1997). The that some species could be indicating groups of areas, rather than single extent to which a species is restricted to a region is calculated as (De areas, increased the indicator value of some species, such as R. tarandus, Cáceres et al., 2012) C. elaphus, R. rupicapra, C. ibex/C. pyrenaica, C. capreolus and E. hydruntinus (Fig. 3). In this analysis, only two regions, central West a =N A ¼ X p p Europe and the rest of the Iberian Peninsula, had indicator species of K = their own ( Fig. 3). Otherwise, three groups of areas were found ¼ ak Nk k 1 to have indicator species: S France + central West Europe, South France + Cantabrian region + the rest of the Iberian Peninsula, and S where ap is the sum of abundances (NISP values in our case) of the France + the rest of the Iberian Peninsula (Fig. 3). Finally, consideration species in the target region, Np is the number of sites belonging to the of pairs of species as indicators of single areas revealed that some target region, K is the number of regions (four in our case), ak is the species pairs had higher indicator values than single species for South sum of abundances of the species at region k, and Nk is the number of sites belonging to region k. France and for the rest of the Iberian Peninsula (Fig. 3). The frequency of a species within a given region is calculated as (De Cáceres et al., 2012) 4. Discussion

n Our study provided four main results: (i) there was a latitudinal B ¼ p Np faunal turnover during the MIS 3 period consistent with a temperature gradient; (ii) the Cantabrian region hosted a transitional faunal assem- where np is the number of occurrences of the species within the region. blage between the northern cold faunas and the southern temperate The indicator value, IV, is calculated as (Dufrêne and Legendre, faunas, due to the presence of species associated with cold faunas; 1997) (iii) it was possible to find indicator species of large herbivores for different geographical areas along that latitudinal gradient; and (iv) in IV ¼ A B 100 many cases, however, indicator species were stronger indicators of combinations of areas, rather than of single areas. Together, this and ranges from 0 to 100 (maximum indicator value). evidence supports a scenario in which the succession of cold and The analysis of indicator species proceeded in three steps. First, indi- warm episodes typical from MIS 3 left signals of mixed cold and temper- cator species for each region considered in this study (Central W ate faunal assemblages but without totally blurring the latitudinal gradi- Europe, S France, Cantabrian region and rest of Iberian Peninsula) ent in climatic conditions. were identified. Second, we considered the possibility that species A continuous gradient in faunal assemblages was found between could be related to the conditions present in more than one region West European and Iberian sites. This was clearly indicated by the (De Cáceres et al., 2010), and calculated a modification of IV which al- latitudinal arrangement of sites in the ordination diagram, which lows to assess whether species are indicator of combinations of regions was defined by the presence of four species typical from cold faunas. (De Cáceres et al., 2010). This analysis can also identify indicator species for single regions (i.e. groups of regions consisting of a single region) Table 1 (De Cáceres et al., 2010). Third, we tested whether a combination of Loading of the ungulate species in each of the three axes extracted by the correspondence species, rather than a single species, could be a better indicator of a analysis. given region (De Cáceres et al., 2012). For this last analysis, only pairs Species Axis 1 Axis 2 Axis 3 of species were considered and the values for At, the lower threshold Bos–Bison −0.276 −0.082 −0.499 for the minimum predictive value, and Bt, the minimum value for sensi- Capra ibex/C. pyrenaica −0.640 0.181 1.672 tivity, were set to 0.6 and 0.4, respectively (De Cáceres et al., 2012). All Capreolus capreolus −0.555 0.083 −0.342 three indicator value analyses were performed with the module Cervus elaphus −0.692 0.156 −0.422 Coelodonta antiquitatis 0.189 0.097 −0.491 indicspecies from the statistical program R 2.15.2 (R Development Dama dama −0.833 0.236 −0.354 fi Core Team, 2012). This module provides IV as well as a signi cance Equus caballus 0.207 −0.046 −0.121 value for this index, based on randomization. Equus hydruntinus −0.333 0.017 −0.125 Mammuthus primigenius 2.217 3.185 −0.022 − − 3. Results Megaloceros giganteus 0.301 0.051 0.357 Ovibos moschatus 1.036 −0.949 0.014 Rangifer tarandus 1.155 −0.859 0.085 The CA identified three axes absorbing 60.0% of the variance in the Saiga tatarica 1.104 −0.054 0.273 species data. The percentage absorbed by axes 1–3 were 25.3, 19.7 Rupicapra rupicapra −0.713 0.155 −0.257 and 15.0, respectively. Axis 1 was positively related to the presence of Stephanorhinus hemitoechus −0.724 0.200 0.271 Sus scrofa −0.307 −0.054 −0.064 M. primigenius, O. moschatus, R. tarandus and S. tatarica,andnegatively D.J. Álvarez-Lao, M. Méndez / Palaeogeography, Palaeoclimatology, Palaeoecology 449 (2016) 455–462 459

Fig. 2. Ordination diagram of the sites along the three first axes of the correspondence analysis. Crosses: S France; open circles: central W European sites; black circles: Cantabrian sites; downward triangles: other Iberian sites.

Latitude and longitude were positively correlated in our data set and supported by several of our results. First, the CA showed a position of provided the same picture in a CCA (results not shown), thus ruling Cantabrian sites that smoothly joined the central W European sites, out major effects of a longitudinal gradient. Cantabrian sites were clear- the S France sites and the sites from the rest of the Iberian Peninsula ly transitional. This transitional nature of the Cantabrian region was (Fig. 2). Second, our indicator analysis of combined regions failed to

Fig. 3. Indicator species and their indicator value for the four areas included in this study. The list under each area includes the indicator species identified by the simple indicator analysis, as well as the pairs of species identified by the additional indicator analysis using combinations of species. Species in bold are the indicator species for groups of areas (joined by lines) identified in an additional indicator analysis of groups of areas. Note that some of those groups consisted of a single area. 460 D.J. Álvarez-Lao, M. Méndez / Palaeogeography, Palaeoclimatology, Palaeoecology 449 (2016) 455–462 isolate the Cantabrian region while still retrieved indicator species for were present in most parts of the latitudinal gradient studied during some the two regions at the ends of the latitudinal gradient. Instead, it joined times of the MIS3. Indication of groups of areas is related to broad the Cantabrian to the S France sites and to the rest of Iberian sites. The ecological tolerance (De Cáceres et al., 2010). transitional nature of the Cantabrian region can be explained by the Our study indicates a particular combination of environmental connection between this region and West Europe (Álvarez-Lao and features in the Cantabrian area during the MIS 3 period. While taxa García, 2011a) and had already been suggested by Stewart (2005). of interglacial character (C. elaphus, C. capreolus and R. rupicapra) The transitional nature of S France in Stewart (2005) was supported were predominant, the presence of R. tarandus, M. primigenius and by our CA analysis and by our indicator species analysis. Nevertheless, C. antiquitatis in several assemblages (see Appendix A) shows that differences in orography between S France and Cantabrian sites separat- cold-adapted species entered northern Iberia during the MIS 3 cold ed these two transitional areas. events. R. tarandus occurred at eleven of the studied Cantabrian assem- It is interesting to note that the abundance of R. tarandus dramatical- blages (Appendix A) but did not reach statistical significance in the ly decreased in presence and abundance between South France (always indicator value analysis. Probably due to their very low abundance in present and normally over 50% of the remains) and the Cantabrian this area, only a weak signal was produced in our analysis. The presence Region (presence at one-third of the studied assemblages and with of other cold-adapted taxa in the rest of Iberia, although much more percentages normally below 1%). In addition, this species was never no- sporadically, has also been recorded during MIS 3. M. primigenius has ticed southwards of the Cantabrian region (Álvarez-Lao and García, been found in northern Iberian assemblages and, interestingly, further 2011a), which reflects the transitional nature of this area. By contrast, south in central and in Granada (Padul peat bog, not included C. elaphus showed the opposite pattern: it was the dominant species in the present study due to lack of detailed information on faunal in almost all the Cantabrian (and the rest of Iberian) sites, though it composition from this site; Álvarez-Lao et al., 2009). was very scarce at most southern French localities (and almost absent A potential indication of faunal mixture would have been strong northwards). These contrasting abundance patterns between two differences in ungulate assemblages at different levels of a given site. adjacent geographical areas located at similar latitudes suggest that In our data set, we had 23 sites represented by two or three levels. the Pyrenees constituted an important ecological barrier to these spe- However, in general, no significant variations in faunal compositions cies. Since both taxa occur in both regions, mere presence–absence were observed at each area throughout the varying climatic conditions analysis would overlook such abundance differences, thus underlying of MIS 3 (results not shown), so the climatic gradient could be recog- the value of quantitative analyses. nized during the entire episode. The current resolution of dating may Looking at single species, each region had indicator species that fitted not be sufficient for an accurate correlation of the within-site faunal expectations based on a latitudinal gradient in environmental condi- assemblages with the sharp climatic oscillations of MIS 3. tions, especially central W Europe. M. primigenius and C. antiquitatis, Assemblages studied here were mainly due to human, or mixed two typical cold-adapted taxa (Kurtén, 1968; Guthrie, 1982; Kahlke, human and carnivore, activity. This raises the question of whether the 1999), were indicator species from the central Western European region, faunal assemblages analyzed are representative of the original biocoe- and the cold-adapted R. tarandus was an indicator species from S France, noses because the selectivity of humans and carnivores as bone accu- while taxa of interglacial character such as C. elaphus and C. capreolus mulators can bias the composition and abundance of the assemblages. (Kurtén, 1968; Guérin and Patou-Mathis, 1996) and, to a lesser extent, This could be the case for many assemblages with over 75%–80% of S. hemitoechus were indicator species from the southernmost studied remains from a single ungulate species (Appendix A). For some of the areas (Cantabrian and the rest of Iberian regions). The Iberian Peninsula assemblages in our data set, the dominance of C. elaphus (Cabrera, seems to have acted during MIS 3 as a glacial refuge for some temperate 1984; Cabrera et al., 2001; Dari, 1999; Castaños, 2005)orR. tarandus species such as C. elaphus, S. hemitoechus and D. dama, the last two of (Delpech, 1984; Grayson and Delpech, 2005) has been attributed to which were not noticed in mainland Europe during this time. Persistence human selectivity, while the dominance of C. ibex/C. pyrenaica and of D. dama and S. hemitoechus in the late Pleistocene was also noticed at R. rupicapra has been attributed to either human (Castaños, 1986; the Italian Peninsula (Sardella et al., 2005, Iurino et al., 2015; Pandolfi Jordá Pardo et al., 2003) or carnivore (Yravedra, 2007, 2010) selectivity. and Tagliacozzo, 2015). Nevertheless, a high percentage of C. elpahus is also found in assem- Two of the indicator species from Cantabrian and other Iberian as- blages with no human or carnivore influence in our data set (López semblages were not linked to climate but to orography. The Cantabrian González, 2001). The literature discussing the taxonomical composition area has a rugged orography which could explain both the absence of of bone accumulations produced by humans or carnivores (Cruz-Uribe, suitable lowland landscape for species like E. hydruntinus (Burke et al., 1991; Pickering, 2002; Fosse et al., 2010; Mallye et al., 2012) has found 2003), and the abundance of a rocky species like R. rupicapra (Kurtén, that carnivore remains are generally more abundant in carnivore accu- 1968; Guérin and Patou-Mathis, 1996). The ibex (C. ibex/ C. pyrenaica) mulations than in those made by humans. However, no clear patterns is also a typical species of mountain areas, which could explain its abun- are found for the relative frequency of ungulate species. Thus, despite dance in a number of Iberian sites located near mountain areas (Iberian, the potential bias caused by different accumulator agents, it is reason- Pyrenean and Baetic mountain ranges; Appendix A, Fig. 1), and especial- able to accept that the most abundant species within an assemblage ly in Zafarraya and Nerja, both located in the Penibaetic Mountain was also the most abundant one in the corresponding biocoenosis Range. In general, the indicator value was higher for the indicator and, therefore, the more available for the predator (human or carni- species in the two northern areas, compared to the two southern vore) responsible of the bone accumulation. Even assuming these areas (Fig. 3). This indicates a higher reliability of cold-adapted species biases, the diverse agents and processes that have influenced the forma- as indicators, and a wider ecological tolerance of the indicator species tion of the analyzed assemblages have not produced modifications so in the southern part of the latitudinal gradient. deep as to blur the clear latitudinal signal detected in our analyses and Within this general picture, it was possible to detect the signal of coherent with previous studies. mixed faunal assemblages due to the alternation of cold and temperate Previous work about MIS 3 faunal assemblages has focused on the episodes. In particular, this was indicated by the higher indicator value characterization of biogeographical faunal provinces (Stewart et al., of some species when they were allowed to indicate groups of areas, rath- 2003; Markova et al., 2010). This approach, based on presence–absence er than single areas. The main example is the presence of C. elaphus as in- data and statistical classification techniques, necessarily yields groups of dicator of all areas except central West Europe. Although the mere sites, i.e. faunal provinces. Our approach, based on quantitative data and presence of C. elaphus is not indicative of specific climatic conditions ordination techniques, allowed to identify transitional zones between (Kurtén, 1968; Guérin and Patou-Mathis, 1996), high proportions of this faunal provinces and thus provided additional details to previous anal- species are usually associated to temperate climates. Temperate climates yses. Both qualitative and quantitative approaches are useful and D.J. Álvarez-Lao, M. Méndez / Palaeogeography, Palaeoclimatology, Palaeoecology 449 (2016) 455–462 461 complementary, rather than alternative. The use of presence–absence Arrizabalaga, A., Altuna, J. (eds.), Labeko Koba (País Vasco). Hienas y humanos en los albores del Paleolítico Superior. Munibe (Antropologia-Arkeologia) 52, pp. 15–72. data has been a traditional logical choice due to the suspicions raised Arrizabalaga, A., 2005. LasprimerasocupacioneshumanasenelPirineoOccidentalyMontes against paleoecological quantitative data (Damuth, 1982). However, Vascos. Un estado de la cuestión en 2005. Munibe (Antropol. Arkeol.) 57 (2), 53–70. our analysis showed that quantitative data provided congruent results Barron, E., Pollard, D., 2002. High-resolution climate simulations of Oxygen Isotope Stage – with the qualitative approach and, at the same time, yielded additional 3 in Europe. Quat. Res. 58, 296 309. Bosselin, B., Djindjian, F., 1994. La chronologie du Gravettien Français. Préhistoire insights. We make a call for further integration of qualitative and Européenne 6, 77–115. quantitative data in future studies of faunal paleoassemblages. Burke, A., Eisenmann, V., Ambler, G.K., 2003. The systematic position of Equus hydruntinus, an extinct species of Pleistocene equid. Quat. Res. 59, 459–469. Cabrera, V., 1984. El Yacimiento de La Cueva de “El Castillo” (Puente Viesgo, Santander). 5. Conclusions Biblioteca Praehistorica Hispana, XXII. Instituto Español de Prehistoria, Madrid, C.S.I.C. Cabrera Valdés, V., Bernardo de Quirós, F., Maíllo Fernández, J.M., Valladas, H., Lloret Martínez de la Riva, A., 2002. El Auriñaciense arcaico de El Castillo (Cantabria): Our analysis, combining quantification and indicator value of large descripción tecnológica y objetivos de la producción. Espacio Tiempo y Forma (Ser. I) herbivores, provides a useful methodology to understand faunal assem- 15, 67–86. blages during MIS 3. Succession of cold episodes left a signal that can be Cabrera, V., Maíllo, J.M., Lloret, M., Bernardo de Quirós, F., 2001. La transition vers le Paléolothique supérieur dans la grotte du Castillo (Cantabrie, Espagne): la couche interpreted as waves of intrusions of cold elements into southern lati- 18. l'Anthropologie 105, 505–532. tudes. The presence of cold-adapted species at very southern latitudes Castaños, P., 1986. Los macromamíferos del Pleistoceno y Holoceno de Vizcaya. and of temperate-adapted species in South France, deduced from a sim- Unpublished Ph.D., Universidad del País Vasco. Castaños, P., 1996. Hallazgos de rinoceronte lanudo en Legintxiki (Etxauri, Navarra). ple qualitative look at fossil evidence, is thus compatible with a gradual Principe de Viana. Supl. Cienc. 14 (15), 77–80. gradient in faunal composition derived from the decreasing influence of Castaños, P., 2005. Revisión actualizada de las faunas de macromamíferos del Würm – cold episodes at lower latitudes. During the MIS 3, almost all species antiguo en la Región Cantábrica. Museo de Altamira. Monografías 20, 201 207. Castaños, P., Castaños, J., 2007. Estudio de la del Abrigo del Cuco. In: were at almost all the regions, but with differential incidence and Muñoz, E., Montes, R. (eds.), Intervenciones arqueológicas en Castro Urdiales 3, persistence along the latitudinal gradient. pp. 161–170. Supplementary data to this article can be found online at http://dx. Castaños, P., Murelaga, X., Bailon, S., Castaños, J., Saez de Lafuente, X., Suárez, O., 2009. Estudio de los vertebrados del yacimiento de Lezikako Koba (Kortezubi, Bizcaia). doi.org/10.1016/j.palaeo.2016.02.050. Kobie (Ser. Paleontol.) 28, 25–50. Clark, P.U., Dyke, A.S., Shakun, J.D., Carlson, A.E., Clark, J., Wohlfarth, B., Mitrovica, J.X., Hostetler, S.W., McCabe, A.M., 2009. The Last Glacial Maximum. Science 325, Acknowledgements 710–714. Corbet, G., Ovenden, D., 1980. The Mammals of Britain and Europe. Collins. This research has been supported by the Universidad de Oviedo Cruz-Uribe, K., 1991. Distinguishing hyena from hominid bone accumulations. J. Field Archaeol. 18, 467–486. project UNOV-13 EMERG-12. We thank Miquel de Cáceres for his help Damuth, H., 1982. Analysis of the preservation of community structure in assemblages of with the technical details in the use of the module indicspecies of R fossil mammals. Paleobiology 8, 434–446. and Marcelino de la Cruz for advice in the code to draw Fig. 2. We are Dansgaard, W., Johnsen, S.J., Clausen, H.B., Dahl-Jensen, D., Gundestrup, N.S., Hammer, C.U., Hvidberg, C.S., Steffensen, J.P., Sveinbjörnsdottir, A.E., Jouzel, J., Bond, G., 1993. very grateful to Florent Rivals and an anonymous reviewer for their Evidence for general instability of past climate from a 250 kyr ice-core record. Nature constructive comments that greatly helped to improve the manuscript. 364, 218–220. Dari, A., 1999. LesgrandsmammifèresdusitePleistoceneSupérieurdelagrottedu Castillo: Étude archéozoologique: donnes paleontologiques, taphonomiques et References palethnographiques. Espacio Tiempo y Forma (Ser. I) 12, 103–127. David, F., Connet, N., Girard, M., Miskovsky, J.-C., Mourer-Chauviré, C., Roblin-Jouve, A., Altuna, J., 1996. Faunas de Clima frío en La Península Ibérica Durante El Pleistoceno 2005. Les niveaux du paléolithique supérieur à la grotte du Bison (Arcy-sur-Cure, Superior. In: Ramil-Rego, P., Fernández Rodríguez, C., Rodríguez Gutián, M. (Eds.), Yonne): Couches A à D. Rev. Archéol. l'Est 54, 5–50. Biogeografía Pleistocena-Holocena de La Península Ibérica. Xunta de Galicia, Santiago De Cáceres, M., Legendre, P., Moretti, M., 2010. Improving indicator species analysis by de Compostela, pp. 13–39. combining groups of sites. Oikos 119, 1674–1684. Altuna, J., 2004. Estudio biométrico de Vulpes vulpes LyAlopex lagopus L. Contribución a su De Cáceres, M., Legendre, P., Wiser, S.K., Brotons, L., 2012. Using species combinations in diferenciación en los yacimientos paleolíticos cantábricos. Munibe (Antropol. Arkeol.) indicator value analyses. Methods Ecol. Evol. 3, 973–982. 56, 45–59. Delpech, F., 1983. Les Faunes Du Paléolithique Supérieur Dans le Sud-Ouest de La France. Altuna, J., Mariezkurrena, K., 2000. Macromamíferos del Yacimiento de Labeko Koba Cahiers Du Quaternaire 6. Editions du Centre National de la Recherche Scientifi que. (Arrasate, País Vasco). In: Arrizabalaga, A., Altuna, J. (Eds.), Labeko Koba (País Delpech, F., 1984. Les ongulés en périgord et Nord-Ouest du Quercy durant le Würm III. Vasco): Hienas Y Humanos en Los Albores del Paleolítico Superior. Munibe Chronoclimatologie, paléobiogéographie, palethnologie. Geobios 17, 531–548. (Antropologia-Arkeologia) vol. 52, pp. 107–181. Delporte, H., 1991. La séquence aurignacienne et périgordienne sur la base des travaux Altuna, J., Mariezkurrena, K., 2010. Tafocenosis en yacimientos del País Vasco con récents réalisés en Périgord. Bull. Soc. Préhistorique Fr. 88 (8), 243–256. predominio de grandes carnívoros. Consideraciones sobre el yacimiento de Amalda. Djindjian, F., Bosselin, B., 1994. Perigordien et Gravettien: L'épilogue d'une contradiction? Actas de la 1ª Reunión de Científicos sobre cubiles de hiena (y otros grandes carnívoros) Préhistoire Eur. 6, 117–131. en los yacimientos arqueológicos de la Península IbéricaZona Arqueológica 13. Alcalá de Dufrêne, M., Legendre, P., 1997. Species assemblages and indicator species: the need for a Henares, pp. 214–228. flexible asymmetrical approach. Ecol. Monogr. 67, 345–366. Altuna, J., Mariezkurrena, K., 2011. Estudio de Los macromamíferos del Yacimiento de Fosse, P., Avery, G., Fourvel, J.-B. Lesur-Gebremariam, J., Monchot, H.,·Brugal, J.P., Kolska Aitzbitarte III (excavación de La Entrada). In: Altuna, J., Mariezkurrena, K., Ríos, J. Horwitz, L., Tournepoche, J.-F., 2010. Los cubiles actuales de hiena: síntesis critica de (Eds.), Ocupaciones Humanas en Aitzbitarte III (País Vasco) 33.600–18.400 BP sus características tafonómicas a partir de la excavación de nuevos yacimientos (Zona de Entrada a La Cueva)Ekob vol. 5. Servicio Central de Publicaciones del (República de Djibuti, África del Sur) y la información publicada. Actas de la 1ª Gobierno Vasco, Vitoria, pp. 395–480. Reunión de Científicos sobre cubiles de hiena (y otros grandes carnívoros) en los Álvarez-Lao, D.J., 2014. The Jou Puerta Cave (Asturias, NW Spain): a MIS 3 large mammal yacimientos arqueológicos de la Península Ibérica. Zona Arqueológica 13, Alcalá de assemblage with mixture of cold and temperate elements. Palaeogeogr. Henares, pp. 97–105. Palaeoclimatol. Palaeoecol. 393, 1–19. García, N., Arsuaga, J.L., De Vos, J., Mol, D., 2003. Last glaciation cold-adapted faunas in the Álvarez-Lao, D.J., García, N., 2010. Chronological distribution of Pleistocene cold-adapted Iberian Peninsula. In: Reumer, J.W.F. (Ed.), Advances in Mammoth Research large mammal faunas in the Iberian Peninsula. Quat. Int. 212, 120–128. (Proceedings of the Second International Mammoth Conference, Rotterdam, May Álvarez-Lao, D.J., García, N., 2011a. Geographical distribution of Pleistocene cold adapted 16–20 1999). Deinsea vol. 9, pp. 159–169. large mammal faunas in the Iberian Peninsula. Quat. Int. 233, 159–170. Graham, R.W., Lundelius Jr., E.L., Graham, M.A., Schroeder, E.K., Toomey III, R.S., Anderson, Álvarez-Lao, D.J., García, N., 2011b. Southern dispersal and palaeoecological implications E., Barnosky, A.D., Burns, J.A., Churcher, C.S., Grayson, D.K., Guthrie, R.D., Harington, of woolly rhinoceros (Coelodonta antiquitatis): review of the Iberian occurrences. C.R., Jefferson, G.T., Martin, L.D., McDonald, H.G., Morlan, R.E., Semken Jr., H.A., Quat. Sci. Rev. 30, 2002–2017. Webb, S.D., Werdelin, L., Wilson, M.C., 1996. Spatial response of mammals to Late Álvarez-Lao, D.J., García, N., 2012. Comparative revision of the Iberian woolly mammoth Quaternary environmental fluctuations. Science 272, 1601–1606. (Mammuthus primigenius) record into a European context. Quat. Sci. Rev. 32, 64–74. Grayson, D.K., Delpech, F., 2005. Pleistocene reindeer and global warming. Conserv. Biol. Álvarez-Lao, D.J., Kahlke, R.-D., García, N., Mol, D., 2009. The Padul mammoth finds—on 19, 557–562. the southernmost record of Mammuthus primigenius in Europe and its southern Guérin, C., Patou-Mathis, M., 1996. Les Grands Mammifères Plio-Pléistocènes d'Europe. spread during the Late Pleistocene. Palaeogeogr. Palaeoclimatol. Palaeoecol. 278, Masson. 57–70. Guthrie, R.D., 1982. Mammals of the Mammoth as Paleoenvironmental Indicators. Arrizabalaga, A., 2000. El yacimiento de Labeko Koba (Arrasate, País Vasco): entorno, In: Hopkins, D.M., Matthews Jr., J.V., Schweger, C.E., Young, S.B. (Eds.), Paleoecology of crónica de las investigaciones, estratigrafía y estructuras: cronología absoluta. In: . Academic Press, New York, pp. 307–326. 462 D.J. Álvarez-Lao, M. Méndez / Palaeogeography, Palaeoclimatology, Palaeoecology 449 (2016) 455–462

Hewitt, G.M., 2000. The genetic legacy of the Quaternary ice ages. Nature 405, 907–913. Sardella, R., Bedetti, C., Belluci, L., Conti, N., Coppola, D., Di Canzio, E., Pavia, M., Petronio, C., Huber, C., Leuenberger, M., Spahni, R., Flückiger, J., Schwander, J., Stocker, T.F., Johnsen, S., Petrucci, M., Salari, L., 2005. The Late Pleistocene vertebrate fauna from Avetrana Landais, A., Jouzel, J., 2006. Isotope calibrated Greenland temperature record over (Taranto, Apulia, Southern ): preliminary report. Geol. Alps 2, 25–29. Marine Isotope Stage 3 and its relation to CH4. Earth Planet. Sci. Lett. 243, 504–519. Schreve, D., 2006. The taphonomy of a Middle Devensian (MIS 3) vertebrate assemblage Iurino, D.A., Fico, R., Sardella, R., 2015. A pathological Late Pleistocene badger from San from Lynford, Norfolk, UK, and its implications for Middle Palaeolithic subsistence Sidero (Apulia, Southern Italy): implications for developmental pathology and strategies. J. Quat. Sci. 21, 543–556. feeding behavior. Quat. Int. 366, 96–101. Sommer, R., Nadachowski, A., 2006. Glacial refugia of mammals in Europe: evidence from Jennings, R., Finlayson, C., Fa, D., Finlayson, G., 2011. Southern Iberia as a refuge for the last fossil records. Mammal Rev. 36, 251–266. populations. J. Biogeogr. 38, 1873–1885. Sommer, R.S., Benecke, N., Lougas, L., Nelle, O., Schmölcke, U., 2011. Holocene survival of Jordá Pardo, J.F., Aura Tortosa, J.E., Rodrigo García, M.J., Pérez Ripio, M., Badal García, E., the wild horse in Europe: a matter of landscape openness? J. Quat. Sci. 26, 805–812. 2003. El registro paleobiológico cuaternario del yacimiento de la Cueva de Nerja Sommer, R.S., Kalbe, J., Ekström, J., Benecke, N., Liljegren, R., 2014. Range dynamics of the (Málaga, España). Bol. Real Soc. Esp. Hist. Nat. (Sección Geol.) 98 (1–4), 73–89. reindeer in Europe during the last 25,000 years. J. Biogeogr. 41, 298–306. Kahlke, R.-D., 1999. The History of the Origin, Evolution and Dispersal of the Late Pleisto- Stewart, J.R., 2005. The ecology and adaptation of during the non-analogue cene Mammuthus–Coelodonta Faunal Complex in Eurasia (Large Mammals). Fenske environment of Oxygen Isotope Stage 3. Quat. Int. 137, 35–46. Companies, Rapid City. Stewart, J.R., 2008. The progressive effect of the individualistic response of species to Kahlke, R.-D., 2014. The origin of Eurasian Mammoth Faunas (Mammuthus–Coelodonta Quaternary climate change: an analysis of British mammalian faunas. Quat. Sci. Rev. Faunal Complex). Quat. Sci. Rev. 96, 32–49. 27, 2499–2508. Kreft, H., Jetz, W., 2010. A framework for delineating biogeographic regions based on Stewart, J.R., Stringer, C.B., 2011. Human evolution out of Africa: the role of refugia and species distributions. J. Biogeogr. 37, 2029–2053. climate change. Science 335, 1317–1321. Kurtén, B., 1968. Pleistocene Mammals of Europe. Weidenfeld & Nicolson, London. Stewart, J.R., van Kolfschoten, M., Markova, A., Musil, R., 2003. The Mammalian Faunas of Lepš,J.,Šmilauer, P., 2003. Multivariate Analysis of Ecological Data Using CANOCO. Europe during Oxygen Isotope Stage Three. In: van Andel, T.H., Davies, W. (Eds.), Cambridge University Press. Neanderthals and Modern Humans in the European Landscape during the Last López González, F., 2001. Biogeografía Y dinámica de La Fauna De macromamíferos Glaciation, 60,000 to 20,000 years ago: Archaeological Results of the Stage 3 Project. Pleistocenos De Galicia: El Yacimiento de Liñares. Unpublished Ph.D., Universidad McDonald Institute Monograph Series, pp. 103–130. de La Coruña. Stuart, A.J., 1982. Pleistocene Vertebrates in the British Isles. Longman. Lyman, R.L., 2008. Quantitative Paleozoology. Cambridge University Press. Takahashi, K., Wei, G., Uno, H., Yoneda, M., Jin, C., Sun, C., Zhang, S., Zhong, B., 2007. AMS Mallye, J.-B., Costamagno, S., Boudadi-Maligne, M., Prucca, A., Lauroulandie, V., Thiébaut, 14C chronology of the world's southernmost woolly mammoth (Mammuthus C., Mourre, V., 2012. Dhole (Cuon alpinus) as a bone accumulator and new taphonom- primigenius Blum.). Quat. Sci. Rev. 26, 954–957. ic agent? The case of (French Pyrenees). J. Taphonomy 10, 317–347. Turner, A., 2000. The Paviland Mammalian Fauna. In: Aldhouse-Green, S. (Ed.), Paviland Markova, A.K., Puzachenko, A.Y., van Kolfschoten, T., 2010. The North Eurasian mammal Cave and the “Red Lady”:ADefinitive Report. Western Academic & Specialist Press assemblages during the end of MIS 3 (Brianskian–Late Karginian–Denekamp inter- Limited, Bristol, pp. 133–140. stadial). Quat. Int. 212, 149–158. Turon, J.-L., Lézine, A.-M., Denèfle, M., 2003. Land-sea correlations for the last glaciation Maroto, J., Vaquero, M., Arrizabalaga, A., Baena, J., Carrión, E., Jordá, J.F., Martinón, M., inferred from a pollen and dinocyst record from the Portuguese margin. Quat. Res. Menéndez, M., Montes, R., Rosell, J., 2005. Problemática del final del Paleolítico 59, 88–96. Medio en el Norte Peninsular. Museo de Altamira. Monografías 20, 101–114. van Andel, T.H., 2002. The climate and landscape of the middle part of the Weichselian Münzel, S.C., Conard, N.J., 2004a. hunting in the Hohle Fels, a cave site in the glaciation in Europe: The Stage 3 Project. Quat. Res. 57, 2–8. Ach Valley, Swabian Jura. Rev. Paléobiol. 23 (2), 877–885. Van Meerbeeck, C.J., Renssen, H., Roche, D.M., Wohlfarth, B., Bohncke, S.J.P., Bos, J.A.A., Münzel, S.C., Conard, N.J., 2004b. Change and continuity in subsistence during the Middle Engels, S., Helmens, K.F., Sánchez-Goñi, M.F., Svensson, A., Vandenberghe, J., 2011. and Upper Palaeolithic in the Ach Valley of Swabia (South-West ). Int. The nature of MIS 3 stadial–interstadial transitions in Europe: new insights from J. Osteoarchaeol. 14, 225–243. model–data comparisons. Quat. Sci. Rev. 30, 3618–3637. North GRIP Members, 2004. High resolution climate record of the Northern Hemisphere Vereschagin, N.K., Baryshnikov, G.F., 1982. Paleoecology of the Mammoth Fauna in the reaching into the last glacial interglacial period. Nature 431, 147–151. Eurasian Arctic. In: Hopkins, D.M., Matthews Jr., J.V., Schweger, C.E., Young, S.B. Palombo, M.R., Ferretti, M.P., 2005. Elephant fossil record from Italy: knowledge, prob- (Eds.), Paleoecology of Beringia. Academic Press, New York, pp. 267–279. lems and perspectives. Quat. Int. 126-128, 107–136. Weninger, B., Jöris, O., Danzeglocke, U., 2008. CalPal 2007. Cologne Radiocarbon Pandolfi, L., Tagliacozzo, A., 2015. Stephanorhinus hemitoechus (Mammalia, Rhinocerotidae) Calibration & Palaeoclimate Research Package (Available at: http://www.calpal.de/). from the Late Pleistocene of Valle Radice (Sora, Central Italy) and re-evaluation of the Yravedra, J., 2007. Nuevas contribuciones en el comportamiento cinegético de la Cueva de morphometric variability of the species in Europe. Geobios 48, 169–171. Amalda. Munibe (Antropol. Arkeol.) 58, 43–88. Pickering, T.R., 2002. Reconsideration of criteria for differentiating faunal assemblages Yravedra, J., 2010. Tafonomía en la cueva de Amalda: la intervención de carnívoros. Actas accumulated by hyenas and hominids. Int. J. Osteoarchaeol. 12, 127–141. de la 1ª Reunión de Científicos sobre cubiles de hiena (y otros grandes carnívoros) en R Development Core Team, 2012. R: A Language and Environment for Statistical los yacimientos arqueológicos de la Península Ibérica. Zona Arqueológica 13. Alcalá Computing. R Foundation for Statistical Computing, Vienna. de Henares, pp. 174–184. Rasilla, M. de la, 1994. El Solutrense en la cornisa Cantábrica. Férvedes 1, 69–87. Roucoux, K.H., Abreu, L. de, Shackleton, N.J., Tzedakis, P.C., 2005. The response of NW Iberian vegetation to North Atlantic climate oscillations during the last 65 kyr. Quat. Sci. Rev. 24, 1637–1653.