Shrews of the Genus Crocidura from El Harhoura 2

Palaeogeography, Palaeoclimatology, Palaeoecology 436 (2015) 1–8

Contents lists available at ScienceDirect

Palaeogeography, Palaeoclimatology, Palaeoecology

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

Shrews of the genus Crocidura from El Harhoura 2 (Témara, Morocco): The contribution of broken specimens to the understanding of Late Pleistocene–Holocene palaeoenvironments in North Africa

Raphaël Cornette a,b,⁎, Emmanuelle Stoetzel c, Michel Baylac a,b, Sibyle Moulin a, Rainer Hutterer d, Roland Nespoulet c, Mohammed Abdeljalil El Hajraoui e,ChristianeDenysa, Anthony Herrel f,g a Muséum National d'Histoire Naturelle (MNHN), Institut de Systématique, Évolution, Biodiversité, (ISYEB), UMR 7205 CNRS/MNHN/UPMC/EPHE, 45 rue Buffon, 75005 Paris, France b UMS CNRS/MNHN 2700, “Outils et Méthodes de la Systématique Intégrative”, Plate-forme de morphométrie, 45 Rue Buffon, 75005 Paris, France c UMR CNRS/MNHN 7194, “Histoire Naturelle de l'Homme préhistorique”, 43 rue Buffon CP 48, 75005 Paris, France d Zoologisches Forschungsmuseum Alexander Koenig, Section of Mammals, Adenauerallee 160, 53113 Bonn, Germany e Institut National des Sciences de l'Archéologie et du Patrimoine, angle rues 5 et 7 Rabat instituts, Madinat Al Irfane, Rabat Hay Riyad, Morocco f UMR CNRS/MNHN 7179, “Mécanismes adaptatifs: des organismes aux communautés”, 55 rue Buffon, 75005 Paris, France g Ghent University, Evolutionary Morphology of Vertebrates, K.L. Ledeganckstraat 35, B-9000 Gent, Belgium article info abstract

Article history: The El Harhoura 2 cave (Morocco) presents a high density of white-toothed shrew (Crocidurinae) remains in the Received 22 January 2015 eight studied archaeological levels, ranging from around 108,000 to 5800 BP. A detailed understanding of the spe- Received in revised form 5 May 2015 cies contained in these assemblages is important for palaeoecological studies, yet often difficult to achieve be- Accepted 12 June 2015 cause the remains are systematically broken. Here we apply geometric morphometric methods that use sliding Available online 21 June 2015 landmarks allowing us to assign broken mandibles to five potential species present in Morocco today. Four species were detected, with a distinct distribution in the different levels. Surprisingly, Crocidura russula, a species Keywords: Shrews typically associated with relatively humid habitats, was found to be more abundant in periods characterized Mandible by dry conditions. Moreover, the abundance of the Saharan species Crocidura tarfayensis and Crocidura lusitania Fragmentation followed an inverse pattern. Both groups appear to affect each other's abundance throughout the studied period. Species identification We suggest that the coastal position of El Harhoura 2 as well as the inter-specific competition between C. russula Taphonomy and other species may explain this pattern. Palaeoecology © 2015 Published by Elsevier B.V.

1. Introduction El Harhoura 2 cave is situated in the region of Témara, few kilometres to the south of Rabat on the Northern-Atlantic coast of 1.1. Background Morocco (Fig. 1). Caves in this region were formed by marine erosion of the sandstone cliff, probably during the MIS 6 to MIS 5 transition North Africa constitutes a transition zone between the arid Sahara (Nespoulet et al., 2008a, 2008b). The region of Témara has an important and the more humid Mediterranean regions. The Quaternary fossil role with respect to the understanding of the origin and dispersal of an- record in this region is rather well documented and allows one to follow atomically modern humans in and out of North Africa. The caves of the last climatic ﬂuctuations and to document their effects upon the Témara offer a unique succession of different human occupations rang- different fossil communities (Stoetzel, 2013). Among these, small ing from the Middle and Upper Palaeolithic to the Middle Neolithic. At mammals are considered as good palaeoenvironmental indicators present, the El Harhoura 2 cave presents an archaeo-stratigraphy divid- (Valenzuela et al., 2009; Escudé et al., 2013), and they enable one to ed into eleven levels (Fig. 2): level 1 = Neolithic; level 2 = upper reﬁne the palaeoecological interpretations when they are studied in a Palaeolithic (Iberomaurusian); levels 3 to 11 = middle Palaeolithic well-established chronological context and with appropriate tapho- (Aterian). However, the bedrock has not yet been attained and addition- nomic and taxonomic methods. al levels will probably be discovered in future excavations. New 14C, TL, OSL and U-Th/ESR datings are currently in progress and some are already available (Jacobs et al., 2012; Janati Idrissi et al., 2012)and provide a more detailed temporal framework for this study. This cave ⁎ Corresponding author at: Muséum National d'Histoire Naturelle (MNHN), Institut de has yielded both lithic and bone industries, ceramics, several human re- Systématique, Évolution, Biodiversité, (ISYEB), UMR 7205 CNRS/MNHN/UPMC/EPHE, 45 rue Buffon, 75005 Paris, France. Tel.: +33 40 79 80 15. mains and burials, and numerous faunal remains (Campmas et al., 2008, E-mail address: [email protected] (R. Cornette). in press; Nespoulet et al., 2008a, 2008b; Michel et al., 2009, 2010; El

Fig. 1. Distribution of the ﬁve extant white-toothed shrews present in Morocco (data from www.iucnredlist.org) and the localisation of the El Harhoura 2 cave.

Hajraoui et al., 2012; Stoetzel et al., 2014). Moreover, a high concentra- This site is unique as it involves studies focusing on different verte- tion of small vertebrate remains has been discovered, which has brates including small mammals, amphibians, and reptiles. Indeed, a allowed for rather complete faunistic and palaeoecological studies of systematic, taphonomic, and palaeoecological approach was undertak- the region (Stoetzel et al., 2007, 2008, 2010, 2011, 2012, 2013; en for this North-African site. The study of the micro vertebrates from Stoetzel, 2009). El Harhoura 2 provided important information on the evolution of

Fig. 2. Stratigraphy of the El Harhoura 2 cave with average 14C(Nespoulet and El Hajraoui, 2002–2003)andOSLages(Jacobs et al., 2012). Levels 1 to 8 are the levels included in the present study. R. Cornette et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 436 (2015) 1–8 3 biodiversity and palaeoenvironments of this region during the Late suggest a North African origin with a recent expansion into Spain, Quaternary, particularly in the absence of palynological data. However, France and nearby countries of the Maghreb during the Late Pleistocene some North-African small vertebrates remain poorly understood with (Cosson et al., 2005; Nicolas et al., in press). Moreover, it has been sug- respect to their systematic position, ecology or bone morphology. gested that the dispersal across the Strait of Gibraltar may have been Moreover, as only limited reference collections exist, this makes identi- human-mediated (Cosson et al., 2005; Nicolas et al., in press). fications difficult for a considerable part of the fossil material, which is, In general, the identification to species level is difficult in Crocidura. in addition, often highly fragmented. Thus, it is necessary to develop Indeed, the genus Crocidura presents a conserved phenotype and for new methods for taxonomic identification and discrimination to allow several species, both the skull and mandible or the teeth need to be an improved identification of the fossil remains (e.g. Cornette et al., carefully studied to correctly discriminate species or populations 2015a). (Leroy et al., 2004; Cornette et al., 2012; Cornette et al., 2015a, 2015b). Consequently, misidentifications are common (Vogel et al., 2006). Cur- 1.2. Model rently, genetic or cytogenetic studies are often used to correctly assign specimens. But palaeogenetic analyses on fossil small mammals are Small mammals (Escudé et al., 2013), and especially shrews very difficult, mainly because of their small size and high degree of frag- (Stoetzel et al., 2007; Valenzuela et al., 2009), are considered very useful mentation. Although the russula and the whitakeri morphotypes have in the context of palaeoclimatic variation interpretation. Indeed, some been identified in the El Harhoura 2 material, some “intermediate” species such as Crocidura russula are ecologically specialized and, in forms and some other morphotypes were detected. But the similarity Morocco, are tied to Mediterranean climates, contrary to Crocidura in morphology between these taxa, as well as the high fragmentation lusitania or Crocidura tarfayensis that are more related to drier environ- of the material, renders precise identification difficult. Small mammals ments. Moreover, they allow the investigation of ancient human pres- represent the bulk of the small vertebrate material from El Harhoura 2 ence and displacements. Indeed, despite the fact that crocidurine (87.5% of the total number of remains), and Crocidura represents 5.8% shrews are not considered to be strictly commensal, some species, like (number of remains) and 17% (minimum number of individuals) of C. russula, are anthropophilic (Churchfield, 1990). Finally, shrews are the small mammal remains (Stoetzel, 2009). Nevertheless, recent known to interact with each other and with other small mammals studies have suggested that small fragments may offer a relevant way (Kirkland, 1991; Liesenjohann et al., 2011; McDevitt et al., 2014; to assign archaeological specimens to extant species using geometric Cornette et al., 2015b). As a consequence, precisely discriminating the morphometrics (Cornette et al., 2015a). Crocidura species at the site, their relative frequency, and their relations with other vertebrates is a primordial step for further studies on 1.3. Aims of the study palaeobiodiversity, palaeoecology, or archaeology in this area. Five species of Crocidura are currently recognized in Morocco The identification of fossil small mammal material of El Harhoura (Fig. 1): Greater White-toothed shrew, C. russula; Whitaker's shrew, 2 is often limited by the conservation state of the remains, especially Crocidura whitakeri; Saharan shrew, C. tarfayensis; Mauritanian shrew, for shrews that are difficult to identify to species level. However, C. lusitania and Savannah path shrew, Crocidura viaria (Hutterer, 1986; this genus with its different species is a key to understand Aulagnier et al., 2008). Only the first two species (C. russula and palaeoenvironments. The present study proposes to take into account C. whitakeri) are currently present in the region of Témara (Aulagnier broken mandibles and assign them to extant species using fragmented and Thévenot, 1986; Aulagnier, 1987; Aulagnier et al., 2008). The material from all levels using recently described geometric morphomet- first occurrence of Crocidura in Morocco is identified from the Early ric methods (Cornette et al., 2015a). Observations of the presence and Pleistocene sites of Irhoud Ocre (Crocidura jaegeri; Jaeger, 1975; relative abundance of different species in the context of the eight stud- Rzebik-Kowalska, 1988), Thomas 1—Unit L (Crocidura darelbeidae; ied levels of the site are discussed and provide new information on Raynal et al., 2001; Geraads, 2010; Geraads et al., 2010), and Sidi paleoenvironmental evolution (climate and landscape) in North Africa. Abdallah (Crocidura abdallahi; note that the validity of C. jaegeri and C. darelbeidae is debated, (M. Vergilino, pers. com.)). The first represen- 2. Material and methods tatives close to modern C. cf. russula also date from this period (Jebel Ressas, Tunisia; Mein and Pickford, 1992). During the Middle 2.1. Archaeological material Pleistocene, the North African Crocidurinae were highly diversified (Stoetzel, 2013), with several fossil taxa (Crocidura thomasi, Crocidura The archaeo-stratigraphy of the cave is composed of eleven levels. maghrebiana, Crocidura marocana, Crocidura tadjerensis), as well as For this study levels 1 to 8, for which the fossil material was available modern ones occurring for the first time during this period for study were considered. Remains of Crocidurinae were retrieved (C. whitakeri and C. tarfayensis, while C. russula was already present) from every level. Fragments were previously identified to genus level, and probably originating from the local C. marocana or C. maghrebiana Crocidura spp., and at least two morphotypes were immediately ob- (Hutterer, 1991; Kowalski and Rzebik-Kowalska, 1991; Aulagnier, served (C. russula and C. whitakeri). Although a few skull fragments 1992; Vogel et al., 2006; Stoetzel, 2013). The Middle to Late Pleistocene have been found, they will not be used here given the low number boundary is characterized by the extinction of numerous small mammal and bad state of preservation. Consequently, only mandibles are used taxa, including all the fossil shrew species (Stoetzel, 2013). During the in the current paper. The typology (Fig. 3) of fragmentation established Late Pleistocene, most of the present Moroccan species were recorded by Cornette et al. (2015a) based on the study of 609 broken specimens from several Late Pleistocene and Holocene sites, such as Irhoud Nean- from El Harhoura 2 was followed here. Four hundred and seventy-three derthal (Jaeger, 1970, 1975; Amani and Geraads, 1998), Oulad Hamida fragments were classi fied into six groups (Table 1) and concern 78 % of 2 (Felids cave; Raynal et al., 2008; Geraads et al., 2010), Chrafate the total number of Crocidura remains. The first fragment (COMP)isthe (Ouabhi et al., 2003), Ez Zarka (Ouabhi et al., 2003), Guenfouda most complete one including the ascending branch and the main part of (Aouraghe et al., 2010; Lopez-García et al., 2013), and El Harhoura 2 the body of the mandible except the region situated near the incisor. The (Stoetzel, 2009; Stoetzel et al., 2010, 2011). In Morocco, the presence latter part is systematically virtually absent. The second fragment of crocidurines concerns mostly C. russula, C. whitakeri and to a lesser (CMM) comprises the articular condyle (but not the coronoid process) extent also C. tarfayensis. Moreover, a few fossils have been attributed and the body of the mandible up to the first molar. The third fragment to C. viaria in Morocco and Tunisia (Jebel Ressas 4; Mein and Pickford, (MM) is the body of the mandible from the third molar up to the first 1992). To date, C. lusitania has never been described from Maghrebian one. The fourth one (ABC) is the ascending branch with the articular fossil sites. Recent genetic studies of the widespread species C. russula, condyle. The fifth (AB) is the ascending branch without the articular 4 R. Cornette et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 436 (2015) 1–8

Fig. 3. Examples and abbreviations for the six mandible fragments of white-toothed shrews representing the six patterns of fragmentation. Anatomical location of landmarks (in red) and sliding semi-landmarks (in blue) used in the study of these six fragments. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) condyle. The sixth (M) is the part of the body situated under the second or C. whitakeri are represented by only a few specimens in the reference and third molar. The other pieces were too small and irregularly data set because they are rare and often even absent in natural history fragmented to be used in this analysis. The teeth were not included in collections. our analysis because they are not uniformly present on the mandible fragments, and also present other limitations related to wear. With re- 2.3. Morphometrics analyses spect to the eight archaeological levels, all fragments were manually classified in six groups following the established pattern of fragmenta- To assign the fossil remains of mandibles to different species, geo- tion and then, each archaeological group was added to its corresponding metric morphometric methods based on sliding semi-landmarks were modern divided referential for analyses (Cornette et al., 2015a). used (Zelditch et al., 2012; Gunz and Mitteroecker, 2013). Combining Landmarks and sliding semi-landmarks were digitized using TPSdig2 anatomical landmarks and sliding semi-landmarks has been successful- (Rohlf, 2013) as illustrated in Fig. 3. ly performed for other archaeological studies (Evin et al., 2013; Cucchi et al., 2014). This approach allowed the description of the form of the 2.2. Reference sample broken mandibles where few anatomical landmarks are present, thus increasing the level of specific discrimination as described previously As mentioned higher, five species of the genus Crocidura occur today (Cornette et al., 2015a). in Morocco. For this study, 136 mandibles were used, 64 belonging to In order to classify the archaeological specimens, Discriminant C. russula,9toC. tarfayensis,5toC. whitakeri,26toC. lusitania and 32 Analyses were performed using modern species found in Morocco to C. viaria (see Cornette et al., 2015a for all details regarding this refer- today as a reference sample, and a posteriori specific assignations for ence sample). This ‘actualistic approach’ is particularly relevant in the fossil specimens were performed (Baylac and Friess, 2005; Evin et al., case of El Harhoura 2, as all the Late Pleistocene and Holocene North 2013; Stoetzel et al., 2013). Size (logarithm of the centroid size), playing African shrew species still exist today, despite changes in their geo- an important role to discriminate white-toothed shrews was included graphic distribution (Stoetzel, 2013). Nowadays, C. russula is the most in the analyses in order to optimize the specific classification abundant and C. whitakeri is rare but also present in the Témara region. (Cornette et al., 2015a). To reduce dimensionality, especially in case of In the archaeological site, the potential presence of other species, absent sliding semi-landmarks, 90% of the total shape variation was kept for today in this region, could be linked to climate changes during the last analyses after a PCA had been performed (Baylac and Friess, 2005). 130,000 years. Indeed, two taxa (the Common Toad Bufo bufo,andthe Results of the specific a posteriori assignation were accepted for Jerboa Jaculus cf. orientalis) were found in the cave, yet are completely each fragment considering the highest probability. Subsequently, in absent from the Témara region today (Stoetzel et al., 2010). In addition, order to compare the specific composition of different levels, specific a sub-Saharan rodent (Arvicanthis sp.) is also represented in the nearby reassignments for all fragments were pooled by species for each level. cave El Mnasra (Stoetzel et al., 2014). Thus, even if rare today in the con- In order to summarize and visualize the relationships between different sidered area, all extant species of the genus Crocidura were used in the levels and their specific composition a Correspondence Analysis reference sample (Cornette et al., 2015a). Some taxa like C. tarfayensis (Benzécri and Coll, 1973; Greenacre, 1984) was performed on the percentage of these pooled results for each level. To relate these faunistic results with palaeobiodiversity a regression Table 1 with the same pooled results and the Shannon's diversity index Number of fragments studied for each level. (Spellerberg and Fedor, 2003) of this region for this period following COMP CMM MM M ABC AB Total Stoetzel et al. (2013) was performed. Shannon's diversity index permits Level 1 2 2 9 7 0 1 21 to evaluate the palaeobiodiversity of small vertebrate (small mammals, Level 2 2 2 18 14 1 4 41 reptiles and amphibians). The calculated Shannon index is 1.14, 0.78, Level 3 3 1 15 17 1 1 38 0.88, 0.91, 0.82, 0.73, 0.69, and 0.61 from level 1 to level 8 respectively Level 4 2 1 14 13 0 1 31 Level 5 10 3 21 22 2 0 58 (Stoetzel, 2009). Its relationships with data concerning the diversity of Level 6 10 9 34 29 4 2 88 Crocidura in the different levels could highlight a potential interspecific Level 7 9 10 51 10 13 7 100 effect. Level 8 18 11 43 8 10 6 96 Statistical analyses were performed using R (R Development Core Total 56 39 205 120 31 22 473 Team, 2013), with the specialized morphometrics library R-morph, R. Cornette et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 436 (2015) 1–8 5

(Baylac, 2012) and the statistical libraries MASS (Venables and Ripley, Table 3 2002) and ade4 (Dray and Dufour, 2007). Pooled results by species for each level. C. lusitania C. russula C. tarfayensis C. viaria C. whitakeri 3. Results Level 1 4 10 2 0 5 Level 2 4 23 9 0 5 Results of specific assignations (Table 2) were pooled by level Level 3 6 14 11 0 7 (Table 3) and the proportions of the presence of species in each level Level 4 7 10 8 0 6 were plotted (Fig. 4). Specimens attributed to C. russula are the most Level 5 4 28 13 0 13 Level 6 8 46 20 0 14 abundant in each level. They represent more than 60% of the sample Level 7 4 63 16 0 17 in level 7 and nearly 30% for level 4, which is lowest number recorded Level 8 8 51 22 0 15 for C. russula. In contrast, no specimens were attributed to C. viaria. Specimens attributed to C. whitakeri are most abundant in level 1 and least abundant in level 2, yet generally present a relatively stable 4. Discussion proportion f the total community. It must be noticed that level 1 is the most recent and specimens attributed to C. whitakeri represent the 4.1. Methodological limitations second most abundant species. Proportionally rare in level 1, specimens attributed to C. tarfayensis are more common in level 4 and very abun- The results have to be considered as a global faunistic list for dant in level 3, being the second most common species after specimens Crocidurinae at the site based on the phenotypic proximity of fossil re- attributed to C. russula in levels 2, 3, 4, 6 and 8. Specimens attributed to mains relative to the extant reference sample. Discriminant analyses C. lusitania present their peaks of abundance in levels 1 and 4 and their allowed us to assign broken mandibles from the site to extant species peak of scarcity in level 7. Overall, specimens attributed to C. lusitania assumed to be potentially present in the different archaeological levels. are proportionally less abundant from level 5 to 8. This approach necessarily imposes a current-day context comparing Considering reciprocal abundance or scarcity of species, specimens fragments to modern fauna (Stoetzel et al., 2010, 2013; Cornette et al., attributed to C. tarfayensis are the most abundant when specimens at- 2015a). However, like most studies based on inferences, the results tributed to C. russula are the least abundant (levels 3 and 4). Another are significantly dependent on the reference sample used to allow as- clear opposite pattern of proportional abundance concerns specimens signments. An unknown species, or a species not included in the refer- attributed to C. russula that follow the inverse pattern of specimens at- ence sample (e.g. currently not occurring in the country of reference), tributed to C. lusitania. This opposition is systematically observed from will not be detected and by necessity will be attributed to one of the levels 8 to 1. It is particularly clear in level 7 where specimens attributed species of the reference sample. Although based on a small sample for to C. russula reach their maximum of abundance and specimens attrib- some rare species like C. whitakeri, the reference sample was established uted to C. lusitania are proportionally least abundant. Moreover in to take into account the presence of taxa currently not present near the level 4, where specimens attributed to C. russula show their lowest site. Indeed, currently only two species are known in this area, C. russula abundance, specimens attributed to C. lusitania obtain their peak of and C. whitakeri (Aulagnier, 1987). Moreover, all fragments were tested abundance. The same kind of opposite pattern of abundance is also ob- for a possible attribution to Suncus etruscus, another Crocidurinae shrew served in level 2. from Morocco. None of the fragments were, however, attributed to this The overall results of the correspondence analysis (Fig. 5)highlight species (data not shown). Nevertheless, the assignment of specimens to two main specific patterns of opposition and show their affinities with C. whitakeri and C. tarfayensis has to be interpreted with care, as sample the respective levels. The first one, between C. russula and C. lusitania, sizes of the reference sample are small and as both are morphologically especially in levels 7 and 4. The second one concerns C. tarfayensis and very similar (Vogel et al., 2006). However, the reference sample and the C. whitakeri, especially in levels 1 and 3. The regression performed on use of a geometric morphometric approach correctly discriminated species composition by level and the Shannon index is significant these two species and previously allowed the detection of them at the (adjusted R2 = 0.74; P =0.03). El Harhoura 2 site (Cornette et al., 2015a). Although ideally additional

Table 2 Results of the speciﬁc assignments for each level and for each fragment. Abbreviations are “lus” for Crocidura Lusitania, “rus” for C. russula, “tar” for C. tarfayensis, “via” for C. viaria,and“whi” for C. whitakeri.

COMP CMM MM

lus rus tar via whi lus rus tar via whi lus rus tar via whi

Level 1 0 1 1 0 0 0 0 0 0 2 1 6 0 0 2 Level 2 0 0 1 0 1 0 0 1 0 1 3 12 1 0 2 Level 3 0 0 0 0 3 0 0 1 0 0 3 7 4 0 1 Level 4 0 0 1 0 1 0 0 0 0 1 4 6 3 0 1 Level 5 0 3 1 0 6 0 0 2 0 1 1 13 4 0 3 Level 6 0 2 4 0 4 0 1 8 0 0 2 20 4 0 8 Level 7 0 5 1 0 3 0 5 2 0 3 4 35 5 0 7 Level 8 0 7 6 0 5 0 6 0 0 5 7 23 8 0 5

M AB ABC

lus rus tar via whi lus rus tar via whi lus rus tar via whi

Level 1 3 3 0 0 1 0 0 1 0 0 0 0 0 0 0 Level 2 1 9 3 0 1 0 2 2 0 0 0 0 1 0 0 Level 3 3 7 4 0 3 0 0 1 0 0 0 0 1 0 0 Level 4 3 4 3 0 3 0 0 1 0 0 0 0 0 0 0 Level 5 3 12 5 0 2 0 0 0 0 0 0 0 1 0 1 Level 6 6 18 3 0 2 0 1 1 0 0 0 4 0 0 0 Level 7 0 7 2 0 1 0 5 2 0 0 0 6 4 0 3 Level 8 1 5 2 0 0 0 5 1 0 0 0 5 5 0 0 6 R. Cornette et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 436 (2015) 1–8

of category 2 to 3 of the Andrew's classification (owl, such as Bubo ascalaphus), by a predator of category 3 to 4 for levels 2, 3 and 6 (diurnal raptor, such as Falco tinnunculus, or small mammalian carnivore), while levels 4 and 5 present an intermediate pattern (unknown predator or mix of predators). All of the identified potential predators are supposed to be opportunistic and thus inducing few biases of faunal representa- tion. The most common prey (i.e. the most abundant species in the ac- cumulations) are small ones in levels 1 (Mus spretus and white- toothed shrews), while they are larger in levels 3 to 8 (Meriones shawii). As there is no relationship between the type of predator, the most hunted species and the palaeoecological signal of each level, differences in species abundances are likely due to environmental changes rather than to biases of predation (Stoetzel, 2009; Stoetzel et al., 2011). Never- theless, the present study cannot respond directly to these origins of accumulation. But in every case, independently from the predator responsible for the accumulation of remains, the species identified were present in the area. Here again, the results highlight the availabil- ity of prey in their relative abundance in the environment, especially in Fig. 4. Relative proportions of species present in the eight different stratigraphic levels. Arid levels (2, 5 and 7) are indicated in white, more humid ones (1, 3, 4, 6 and 8) are shad- case of generalist predators consuming broad prey spectra (Stoetzel ed in grey. et al., 2011).

4.3. Palaeoecological results specimens of these species are needed, the data nonetheless allow us to detect the presence of others species of white toothed shrews where The small vertebrate associations of El Harhoura 2 indicate that the only two taxa were previously identified or expected. Late Pleistocene was characterized by a succession of relatively humid (levels 3, 4, 6, and 8) and arid (levels 2, 5, and especially 7) periods 4.2. Source of accumulation (Stoetzel, 2009; Stoetzel et al., 2011). Landscapes were always of mosaic type, mostly open (steppes/meadows), but with more bushy/woody The different levels clearly present a different species composition. areas and fresh water places during more humid phases. The Middle These differences could represent changes in the palaeofauna for each Holocene (level 1) was characterized by a Mediterranean climate level, but have to be interpreted given potential sources of bias. The ac- close to, but a little more humid and warmer than the present one cumulations in caves may originate from different processes (Andrews, (Stoetzel et al., 2007, 2008, 2011; Stoetzel, 2009). 1990; Denys, 2002; Stoetzel, 2009). In the context of the present study, Putting these results in this palaeoclimatic context highlights a sin- the small vertebrate assemblages are suspected to be the result of accu- gular particularity compared to similar studies on other taxa (Stoetzel mulations of pellets/scats of different predators (Stoetzel et al., 2011, et al., 2008, 2011). Indeed, in contrast to what was expected, specimens 2012). Thus, differences in prey preferences of predators could explain attributed to C. russula reach their maximal relative abundance in levels some of the results. According to the taphonomic analysis, small verte- 2 and 7. According to the faunal associations and palaeoecological indi- brates of levels 1, 7 and 8 seem to have been accumulated by a predator ces, these two periods correspond to more arid palaeoclimate compared to other levels (Stoetzel, 2009; Stoetzel et al., 2011). C. russula is a small mammal typically associated to Mediterranean climatic zones as suggested by its current distribution in Africa (Fig. 1). Moreover, it seems to prefer more humid and more vegetated areas than C. whitakeri, although the ecology of the latter remains poorly known (Aulagnier and Thévenot, 1986; Aulagnier et al., 2008). Although the relative abundance in some levels may be the result of a predation bias as discussed above, it follows exactly the inverse pattern of the a priori prediction. Moreover, a scarcity or abundance of specimens attributed to C. russula is always associated with an inverse pattern of abundance or scarcity of C. lusitania and C. tarfayensis, two dry climate adapted species. These two species are expected to be found during more arid periods given their desert habits, as suggested by their current distribution (Fig. 1). Note, however, that C. lusitania originated in Sub-Saharan Africa and may have arrived in the area of study during more humid periods or by following the coast, allowing it to traverse the Sahara. This unexpected result may potentially be explained by the geo- graphical position of El Harhoura 2 cave. Indeed, situated on the coast (Fig. 1), the oceanic influence could maintain a micro-climate that acted to buffer global climatic tendencies. This situation could have created particular inter-specific competition between different shrew species. C. russula is always more abundant, even in arid periods, and as such appears dominant in the context of potential interspecific competition. Indeed, C. russula has completely eliminated the native C. suaveolens from numerous Atlantic islands suggesting its competitive superiority (Cosson et al., 1996). A second example of its competitive Fig. 5. Correspondence analysis performed on the pooled specific assignments for each superiority is found in Switzerland where it outcompetes C. leucodon studied level showing relationships between species and archaeological levels. (Vogel et al., 2002). Moreover, C. russula seems to be the only species R. Cornette et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 436 (2015) 1–8 7 currently present in the area of El Harhoura 2 as the other species Morocco and the Institut Scientifique, Rabat, Morocco for their help in C. whitakeri has neither been captured in the Témara region, nor in the providing modern and fossil shrew specimens. We also thank the editor rest of Morocco despite numerous trapping sessions. This is of interest and two anonymous reviewers for constructive and helpful comments given that specimens attributed to C. whitakeri were still present in on a previous version of the manuscript. This study was supported by the most recent level (Middle Holocene). Moreover, C. russula shows ANR-09-PEXT-004 MOMHIE and by ATM-MNHN Formes possibles, no, or only a weak current overlap with the three others species of formes réalisées. We are very grateful to Peter Vogel for lending modern white-toothed shrews (Fig. 1) which may suggest climatic constraints specimens of C. viaria and C. tarfayensis, and to Silvia Valenzuela-Lamas but also inter-specific competition. Indeed, the northern distribution for lending specimens of C. russula from Spain. We also thank Céline of these three species finishes where the southern one of C. russula be- Houssin for her help during different steps of this work, Julio Pedraza gins. Previous studies (Cornette et al., 2015b) have demonstrated that for computing help, Maxime Cammas for the distribution map and ahighbiteforceinC. russula could explain a part of this competitive su- Allowen Evin, Jan Decher and Silvia Valenzuela-Lamas for improving periority. Moreover, the results clearly show that the proportion of dif- drafts of this paper. ferent species of white-toothed shrews at El Harhoura 2 is correlated to the biodiversity of vertebrates found in the site as indicated by the cal- References culated Shannon index further suggestion a role for competition in the observed patterns of abundance. Amani, F., Geraads, D., 1998. Le gisement moustérien du Djebel Irhoud, Maroc: précisions sur la faune et la paléoécologie. Bull. Archeol. Maroc. 18, 11–18. These data also show that white-toothed shrews respond in a differ- Andrews, P., 1990. Owls, Caves and Fossils. Natural History Museum Publications, London, ent way to known climatic variations in comparison to other verte- p. 231. brates. The El Harhoura 2 palaeoenvironmental reconstructions were Aouraghe, H., Agusti, J., Ouchaou, B., Bailon, S., Lopez-Garcia, J.M., Haddoumi, H., El Hammouti, K., Oujaa, A., Bougariane, B., 2010. The Holocene vertebrate fauna from based on the macro- and micro-faunal composition of the different Guenfouda site, Eastern Morocco. Hist. Biol. 22 (1), 320–326. levels, and the calculation of palaeoecological indices (Stoetzel et al., Aulagnier, S., 1987. Les Crocidures (Insectivora) du Maroc septentrional: biométrie, 2007, 2011, 2014). Among the most noteworthy evidences, the pres- systématique. Bull. Inst. Sci. Rabat 11, 193–204. Aulagnier, S., 1992. Zoogéographie des Mammifères du Maroc: de l'analyse spécifiqueàla ence of some desert-adapted rodents (Jaculus cf. orientalis) during arid typologie de peuplement à l'échelle régionale (State Thesis). University Montpellier periods (Stoetzel et al., 2011), and more aquatic and humid-adapted 2, France. amphibians and reptiles during more humid periods (Stoetzel et al., Aulagnier, S., Thévenot, M., 1986. Catalogue des mammifères sauvages du Maroc. Travaux fi fi 2008, 2012) can be mentioned. de l'Institut Scienti que, Série Zoologie 41. Institut Scienti ue de Rabat, Rabat, p. 163. Aulagnier, S., Haffner, A.J., Mitchell-Jones, A.J., Moutou, F., Zima, J., 2008. Guide des A quite similar unexpected pattern was, however, found in a geo- Mammifères d'Europe, d'Afrique du Nord et du Moyen-Orient. Delachaux et Niestlé, metric morphometric study of molar shape of the mice of El Harhoura Paris, p. 271. 2(Stoetzel et al., 2013). This study showed an overall morphological Baylac, M., 2012. Rmorph: a R geometric and multivariate morphometrics library Available from the author:. [email protected]. stability of M. spretus throughout the Late Pleistocene despite several Baylac, M., Friess, M., 2005. Fourier descriptors, procrustes superimposition, and data environmental changes. However, a slight shift in molar shape was ob- dimensionality: an example of cranial shape analysis in modern human populations. served during Middle Holocene, and the size of the fossil M. spretus mo- In: Slice, D.E. (Ed.), Modern Morphometrics in Physical Anthropology. Kluwer – fi Academic/Plenum Publishers, New York, pp. 145 165. lars of El Harhoura 2 is signi cantly larger than that in the modern Benzécri, J.P., Coll, 1973. L'analyse des données. II. L'analyse des correspondances. Bordas, populations of the same species. The impact of human activities and Paris, p. 620. inter-specific competition were likely involved (other murid species ar- Campmas, E., Michel, P., Amani, F., Cochard, D., Costamagno, S., Nespoulet, R., El Hajraoui, M.A., 2008. Comportements de subsitance à l'Atérien et au Néolithique au Maroc riving in Morocco during Holocene, especially the house mouse around Atlantique: premiers résultats de l'étude taphonomique et archéozoologique des 3000 BP) and seem to have had a stronger influence than climatic faunes d'El Harhoura 2 (Région de Témara, Maroc). Actes du colloque RQM4, Le changes on the morphological evolution of Mus. It is interesting to Quaternaire marocain dans son contexte méditerranéen, Oujda, 15–17 Novembre 2007pp. 236–254. note here that despite a clear morphological stability throughout the Campmas, E., Michel, P., Costamagno, S., Amani, F., Stoetzel, E., Nespoulet, R., El Hajraoui, El Harhoura 2 stratigraphy (no statistical difference between the 8 stud- M.A., 2015. Were Upper Pleistocene human/carnivore occupations at the Témara ied levels), Stoetzel et al. (2013) demonstrated a significant relationship caves (El Harhoura 2 and El Mnasra, Morocco) influenced by climate change? – between tooth shape (m1 and M1 first DFA axis) and the diversity indi- J. Hum. Evol. 78, 122 143. http://dx.doi.org/10.1016/j.jhevol.2014.08.008. Churchfield, S., 1990. The Natural History of Shrews. Comstock Publishing, Cornell ces (Shannon and Simpson). This may reflect changes in morphology University, Ithaca, New York, p. 178. related to biotic or abiotic environmental changes (e.g. global changes Cornette, R., Herrel, A., Cosson, J.F., Poitevin, F., Baylac, M., 2012. Rapid morpho-functional in primary productivity and temperature through time), or this might changes among insular populations of the greater white-toothed shrew. Biol. J. Linn. Soc. 107, 322–331. also be the consequence of concomitant trends through time of tooth Cornette, R., Herrel, A., Stoetzel, E., Moulin, S., Hutterer, R., Denys, C., Baylac, M., 2015a. shape and faunal diversity. Specific information levels in relation to fragmentation patterns of shrew mandibles: do fragments tell the same story? J. Archaeol. Sci. 53, 323–330. http://dx.doi.org/10. 1016/j.jas.2014.10.020. 5. Conclusion Cornette, R., Tresset, A., Houssin, C., Pascal, M., Herrel, A., 2015b. Does bite force provide a competitive advantage in shrews? The case of the greater white-toothed shrew. Biol. Four species of white-toothed shrews were recovered at El Harhoura J. Linn. Soc. 114, 795–807. http://dx.doi.org/10.1111/bij.12423. Cosson, J.F., Pascal, M., Bioret, F., 1996. Origine et répartition des musaraignes du genre 2 based on an analysis of the phenotypic proximity of the form of man- Crocidura dans les îles bretonnes. Vie et milieu 16 (3– 4), 233–244. dible fragments. Their relative abundance in different stratigraphic Cosson, J.F., Hutterer, R., Libois, R., Sarà, M., Taberlet, P., Vogel, P., 2005. Phylogeographical levels reveals an original pattern inverse to the predicted pattern footprints of the Strait of Gibraltar and Quaternary climatic fluctuations in the west- ern Mediterranean: a case study with the greater white-toothed shrew, Crocidura of abundance for humid and arid periods. White-toothed shrews russula (Mammalia: Soricidae). Mol. Ecol. 14 (4), 1151–1162. present an important contribution to the palaeobiodiversity and Cucchi, T., Barnett, R., Martínkov_a, N., Renaud, S., Renvoisé, E., Evin, A., Sheridan, A., palaeoclimatology of the studied area and the observed patterns may Mainland, I., Wickham-Jones, C., Tougard, C., Quéré, J.P., Pascal, M., Pascal, M., be caused by the coastal position of the El Harhoura 2 cave, or rather Heckel, G., O'Higgins, P., Searle, J.B., Dobney, K.M., 2014. The changing pace of insular life 5000 years of microevolution in the Orkney vole (Microtus arvalis orcadensis). the ancient inter-specific competitions among shrews. Evolution 68 (10), 2804–2820. Denys, C., 2002. Micromammifères. In: Miskovsky, J.C. (Ed.), Géologie de la Préhistoire. – Acknowledgements Géopré, Presses Universitaires de Perpignan, Paris, pp. 841 858. Dray, S., Dufour, A.B., 2007. The ade4 package: implementing the duality diagram for ecologists. J. Stat. Softw. 22 (4), 1–20. We thank the Mission Archéologique El Harhoura-Témara, Ministère El Hajraoui, M.A., Nespoulet, R., Debénath, A., Dibble, H., 2012. Préhistoire de la Région de des Affaires Etrangères et Européennes, France, Ministère de la Culture, Rabat-Témara, Villes et Sites Archéologiques du Maroc. vol. III, p. 300. Escudé, E., Renvoisé, E., Lhomme, V., Montuire, S., 2013. Why all vole molars (Arvicolinae, Morocco (dir. R. Nespoulet and M.A. El Hajraoui), the Institut National Rodentia) are informative to be considered as proxy for Quaternary paleoenvironmental des Sciences de l'Archéologie et du Patrimoine (INSAP), Rabat, reconstructions. J. Archaeol. Sci. 40, 11–23. 8 R. Cornette et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 436 (2015) 1–8

Evin, A., Cucchi, T., Cardini, A., Strand Vidarsdottir, U., Larson, G., Dobney, K., 2013. The occupations in the Temara Region (Rabat, Morocco): recent data on hominin con- long and winding road: identifying pig domestication through molar size and texts and behavior. Afr. Archaeol. Rev. 25 (1–2), 21–39. shape. J. Archaeol. Sci. 40, 735–743. Nicolas, V., Jacquet, F., Stoetzel, E., Lalis, A., Delapré, A., El Brini, H., Mataame, A., Rihane, A., Geraads, D., 2010. Biochronologie mammalienne du Quaternaire du Maroc atlantique, Liefried, S., Ouarour, A., Denys, C., 2015. Identification Moléculaire et Histoire dans son cadre régional. L' Anthropologie 114, 324–340. Phylogéographique et Démographique des Musaraignes du Maroc Septentrional. Geraads, D., Raynal, J.P., Sbihi-Alaoui, F.Z., 2010. Mammalian faunas from the Pliocene and Travaux de l'Institut Scientifique, Rabat (in press). Pleistocene of Casablanca (Morocco). Hist. Biol. 22 (1–3), 275–285. Ouabhi, Y., Aberkan, M., Serre, F., 2003. Recent Quaternary fossil mammals of Chrafate and Greenacre, M.J., 1984. Theory and Applications of Correspondence Analysis. Academic Ez Zarka. The origin of modern fauna in the Northern Rif (NW Morocco, Northern Press, London, p. 364. Africa). Geol. Acta 1 (3), 277–288. Gunz, P., Mitteroecker, P., 2013. Semilandmarks: a method for quantifying curves and R Development Core Team, 2013. R: A Language and Environment for Statistical Comput- surfaces. Hystrix 24 (1), 103–109. ing. R Foundation for Statistical Computing, Vienna, Austria (URL http://www.R- Hutterer, R., 1986. The species of Crocidura (Soricidae) in Morocco. Mammalia 50 (4), project.org). 521–534. Raynal, J.P., Sbihi-Alaoui, F.Z., Geraads, D., Magoga, L., MohiB, A., 2001. The earliest occu- Hutterer, R., 1991. Variation and evolution of the Sicilian shrew: taxonomic conclusions pation of North-Africa: the Moroccan perspective. Quat. Int. 75 (1), 65–75. and description of a possibly related species from the Pleistocene of Morocco Raynal, J.P., Amani, F., Geraads, D., El Graoui, M., Magoga, L., Texier, J.P., Sbihi-Alaoui, F.Z., (Mammalia: Soricidae). Bonner Zool. Beitr. 42 (3–4), 241–251. 2008. La Grotte des Félins, site paléolithique du Pléistocène supérieur à Dar Bouazza Jacobs, Z., Roberts, R.G., Nespoulet, R., EL Hajraoui, M.A., Debenath, A., 2012. Single-grain (Maroc). L' Anthropologie 112, 182–200. OSL chronologies for Middle Palaeolithic deposits at El Mnasra and El Harhoura 2, Rohlf, F.J., 2013. TpsDig2: Digitize Coordinates of Landmarks and Capture Outlines. Stony Morocco: implications for Late Pleistocene human–environment interactions along Brook, Department of Ecology and Evolution, State University of New York. the Atlantic coast of northwest Africa. J. Hum. Evol. 62 (3), 377–394. Rzebik-Kowalska, B., 1988. Soricidae (mammalia, insectivora) from the plio-pleistocene Jaeger, J.J., 1970. Découverte au Jebel Irhoud des premières faunes de rongeurs du and middle quaternary of Morocco and Algeria. Folia Quat. 57, 51–90. Pléistocène inférieur et moyen du Maroc. C. R. Acad. Sci. Paris 270, 920–923. Spellerberg, I.F., Fedor, P.J., 2003. A tribute to Claude Shannon (1916–2001) and a plea for Jaeger, J.J., 1975. Evolution des Rongeurs du Miocène à l'Actuel en Afrique nord- more rigorous use of species richness, species diversity and the ‘Shannon–Wiener’ occidentale (PhD Thesis). vol. 3. University Montpellier 2. Index. Glob. Ecol. Biogeogr. 12 (3), 177–179. Janati Idrissi, N., Falgueres, C., Haddad, M., Nespoulet, R., EL Hajraoui, M.A., Debenath, A., Stoetzel, E., 2009. Les microvertébrés du site d'occupation humaine d'El Harhoura, Bejjit, L., Bahain, J.J., Michel, P., Garcia, T., Boudad, L., EL Hammoutu, K., Oujaa, A., (Pléistocène supérieur-Holocène, Maroc): systématique, évolution, taphonomie et 2012. Datation par ESR-U/Th combinées de dents fossiles des grottes d'El Mnasra et paléoécologie (PhD Thesis). Muséum National d'Histoire Naturelle, Paris. d'El Harhoura 2, région de Rabat-Témara. Implications chronologiques sur le Stoetzel, E., 2013. Late Cenozoic micromammal biochronology of northwestern Africa. peuplement du Maroc atlantique au Pléistocène supérieur et son environnement. Palaeogeogr. Palaeoclimatol. Palaeoecol. 392, 359–381. Quaternaire 23 (1), 25–35. Stoetzel, E., Michel, P., Nespoulet, R., El Hajraoui, M.A., 2007. Les environnements Kirkland Jr., G.L., 1991. Competition and coexistence in shrews (Insectivora: Soricidae). Holocènes du littoral atlantique du Maroc: exemple des petits et grands vertébrés The biology of the Soricidae. Special Publication of the Museum of Southwestern en contexte archéologique provenant de la grotte d'El Harhoura 2, région de Témara. Biology, University of New Mexico, Albuquerque, pp. 15–22. Quaternaire 18 (4), 299–307. Kowalski, K., Rzebik-Kowalska, B., 1991. Mammals of Algeria. Zaklad Narodowy im. Stoetzel, E., Bailon, S., El Hajraoui, M.A., Nespoulet, R., 2008. Apport sur les connaissances Ossolinskich—Wydawnictwo, Wroclaw, Poland, p. 370. des paléoenvironnements néolithiques du Maroc à partir des amphibiens-reptiles de Leroy, A., Montuire, S., Marchand, D., 2004. Analysis of teeth outlines: a morphometric la couche 1 d'El Harhoura 2, Rabat-Témara, Maroc. L' Anthropologie 112 (4–5), tool for distinguishing three species of Crocidurinae Soricidae, Insectivora, Mammalia 731–756. in France. J. Nat. Hist. 38, 259–267. Stoetzel, E., Bailon, S., Nespoulet, R., El Hajraoui, M.A., Denys, C., 2010. Pleistocene and Liesenjohann, M., Liesenjohann, T., Trebaticka, L., Haapakoski, M., Sundell, J., Ylönen, H., Holocene small vertebrates of El Harhoura 2 Cave (Rabat-Temara, Morocco): an Eccard, J.A., 2011. From interference to predation: type and effects of direct interspe- annotated preliminary taxonomic list. Hist. Biol. 22 (1–3), 303–319. cific interactions of small mammals. Behav. Ecol. Sociobiol. 65 (11), 2079–2089. Stoetzel, E., Marion, L., Nespoulet, R., El Hajraoui, M.A., Denys, C., 2011. Taphonomy and Lopez-García, J.M., Agustí, J., Aouraghe, H., 2013. The small mammals from the Holocene Palaeoecology of the Late Pleistocene to Middle Holocene Small Mammals succession site of Guenfouda (Jerada, Eastern Morocco): chronological and paleoecological of El Harhoura 2 cave (Rabat-Temara, Morocco). J. Hum. Evol. 60, 1–33. implications. Hist. Biol. 25 (1), 51–57. Stoetzel, E., Denys, C., Bailon, S., EL Hajraoui, M.A., Nespoulet, R., 2012. Taphonomic anal- McDevitt, A.D., Montgomery, W.I., Tosh, D.G., Lusby, J., Reid, N., White, T.A., Yearsley, J.M., ysis of amphibian and squamate remains from El Harhoura 2 (Rabat-Témara, 2014. Invading and expanding: range dynamics and ecological consequences of the Morocco): contributions to palaeoecological and archaeological interpretations. Int. greater white-toothed shrew (Crocidura russula) invasion in Ireland. PLoS One 9 J. Osteoarchaeol. 22 (5), 616–635. (6), e100403. Stoetzel, E., Denys, C., Michaux, J., Renaud, S., 2013. Mus in Morocco: a Quaternary Mein, P., Pickford, M., 1992. Gisements karstiques pléistocènes au Djebel Ressas, Tunisie. sequence of intra-specific evolution. Biol. J. Linn. Soc. 109, 599–621. C. R. Acad. Sci. Paris Ser. II 315, 247–253. Stoetzel, E., Campmas, E., Michel, P., Bougariane, B., Ouchaou, B., Amani, F., El Hajraoui, Michel, P., Campmas, E., Stoetzel, E., Nespoulet, R., El Hajraoui, M.A., 2009. La Macrofaune M.A., Nespoulet, R., 2014. Context of modern human occupations in North Africa: du Pléistocène supérieur d'El Harhoura 2 (Témara, Maroc): Données préliminaires. L' contribution of the Témara caves data. Quat. Int. 320, 143–161 (Special issue: Anthropologie 113 (2), 283–312. Northwest African Prehistory). Michel, P., Campmas, E., Stoetzel, E., Nespoulet, R., El Hajraoui, M.A., Amani, F., 2010. La Valenzuela, S., Poitevin, F., Cornette, R., Bournery, R., Nadal, J., Vigne, J.D., 2009. Evolving grande faune du Paléolithique supérieur (niveau 2) et du Paléolithique moyen ecosystems: ecological data from an Iron Age small mammal accumulation at Alorda (niveau 3) de la grotte d'El Harhoura 2 (Témara, Maroc): etude paléontologique, Park (Catalonia, Spain). J. Archaeol. Sci. 36, 1248–1255. reconstitutions paléoécologiques et paléoclimatiques. Hist. Biol. 22 (1–3), 327–340. Venables, W.N., Ripley, B.D., 2002. Modern Applied Statistics With S. Fourth edition. Nespoulet, R., EL Hajraoui, M.A., 2002–2003. Rapport d'activités de la Mission Springer-Verlag, New York, p. 495. Archéologique El Harhoura II. Ministère des Affaires Etrangères et Européennes Vogel, P., Jutzeler, S., Rulence, B., Reutter, B.A., 2002. Range expansion of the (France). Ministère de la Culture, Maroc, p. 39. greater white-toothed shrew Crocidura russula in Switzerland results in local extinc- Nespoulet, R., Debénath, A., El Hajraoui, M.A., Michel, P., Campmas, E., Oujaa, A., Ben Ncer, tion of the bicoloured white-toothed shrew Crocidura leucodon.ActaTheriol.47(1), A., Amani, F., Stoetzel, E., Boudad, L., 2008a. Le contexte archéologique des restes 15–24. humains atériens de la région de Témara (Maroc). Apport des fouilles des grottes Vogel, P., Mehmeti, A.M., Dubey, S., Vogel-Gerber, C., Koyasu, K., Ribi, M., 2006. Habitat, d'El Mnasra et d'El Harhoura 2. Actes du colloque RQM4, Le Quaternaire marocain morphology and karyotype of the Saharan shrew Crocidura tarfayaensis (Mammalia: dans son contexte méditerranéen, Oujda, 15–17 novembre 2007pp. 356–375. Soricidae). Acta Theriol. 51 (4), 353–361. Nespoulet, R., El Hajraoui, M.A., Amani, F., Ben Ncer, A., Debénath, A., El Idrissi, A., Zelditch, M.L., Swiderski, D.L., Sheets, H.D., 2012. Geometric Morphometrics for Biologists: Lacombe, J.P., Michel, P., Oujaa, A., Stoetzel, E., 2008b. Palaeolithic and neolithic A Primer. Academic Press, p. 478.