The Paleobiology of the Extinct Venomous Shrew Beremendia (Soricidae, Insectivora, Mammalia) in Relation to the Geology and Paleoenvironment of Dmanisi (Early Pleistocene, Georgia) more

by Oscar Sanisidro

Co-authored with Marc Furió, Jordi Agustí Alexander Mouskhelishvili and Andrés Santos-Cubedo

Journal of Vertebrate Paleontology 30(3):928–942, May 2010 © 2010 by the Society of Vertebrate Paleontology ARTICLE THE PALEOBIOLOGY OF THE EXTINCT VENOMOUS SHREW BEREMENDIA (SORICIDAE, INSECTIVORA, MAMMALIA) IN RELATION TO THE GEOLOGY AND PALEOENVIRONMENT OF DMANISI (EARLY PLEISTOCENE, GEORGIA) ´ ´ ´ MARC FURIO,*,1 JORDI AGUSTI,2 ALEXANDER MOUSKHELISHVILI,3 OSCAR SANISIDRO,4 ´ and ANDRES SANTOS-CUBEDO1,5 1 ` ` ` ` Institut Catala de Paleontologia (Modul ICP, Universitat Autonoma de Barcelona) Cerdanyola del Valles (08193), Barcelona, Spain, marc.furio@icp.cat; 2 ICREA-Institute of Human Paleoecology, University Rovira i Virgili, Pl. Imperial Tarraco, 1 (43005), Tarragona, Spain, jordi.agusti@icrea.es; 3 Georgian State Museum, Georgian Academy of Sciences, Purtseladze Street 3 (380007), Tbilisi, Georgia, amouskhelishvili@museum.ge; 4 Museo Nacional de Ciencias Naturales, Departamento de Paleobiolog´a, C/Pinar 25 (28006), Madrid, Spain, osamo@alumni.uv.es; ı 5 ´ Grup Guix, C/Santa Luc´a 75 Vila-real (12540), Castello, Spain, andres.santos@icp.cat ı ABSTRACT—The uses of toxic substances in the animal kingdom are usually explained as adaptations to reach bigger prey—venom, or to defend from the attack of predators—poison. This is a quite simplistic explanation of the reality, which offers other, less evident, uses for the possession of these compounds. In the present work, we analyze the characters of Beremendia Kormos, 1934, an extinct Eurasian genus of shrews, which was recently said to have been venomous. The envenomation apparatus of these shrews was correlated with its uncommonly large size, justifying a possible adaptation to hunt big prey. Examining its dental characters, we do reassess the venomous nature of the species included in this genus, but we deduce that the diet of Beremendia was highly specialized in coleopterans and gastropods instead of large animals. The use of venom in shrews feeding on non-struggling prey can be reliably explained as a mechanism to subdue the prey without killing them before the real time of consumption. The induction of victims into a comatose-state permits their hoarding for a longer time in a better state of preservation than if they were dead, thus diminishing the risk of starvation. Such strategy provides important beneﬁts to their users under irregular conditions, because the effects of environmental unpredictability are consequently reduced. This interpretation of Beremendia is supported by the ethology of some extant shrews, and correlated at local scale with the geological context of Dmanisi, and at global scale with the Plio-Pleistocene climatic trends. INTRODUCTION The study of toxic substances is not only of great interest for zoologists and ecologists, but it has also become an unexpected source of ﬁnds with practical applications in pharmacology. The total of venomous and poisonous vertebrates accounts about 2000 species (Smith and Wheeler, 2006), a non-negligible number considering a total of around 50,000 extant species of vertebrates alive. Toxic substances acquire two different names, depending on their use and localization. On the one hand, ‘poisons’ refer to toxins accumulated in subcutaneous reservoirs spread all over the body of animals that are potential prey. In this case, toxic substances have the function of inhibiting capture and their usefulness is usually reinforced by the development of shiny colors or spectacular forms in an attempt to alert possible predators of the unhealthy consequences of their ingestion. On the other hand, ‘venoms’ are toxic substances concentrated in few parts of the body and associated to delivery systems of mainly predatory animals. In contrast to reptiles and amphibians, the list of extant venomous mammals is quite short. Toxic substances in mammals are commonly related with an improvement of the hunting efﬁciency, as in Solenodon cubanus Peters, 1864, Neomys fodiens (Pennant, 1771), Blarina carolinensis (Bachman, 1837), and Blarina brevicauda (Say, 1823). In the other known extant venomous mam*Corresponding author. mal, the monotreme platypus (Ornithorhynchus anatinus Shaw, 1799), only the males are venom deliverers and the function of the extratarsal spur injection is probably a weapon used by one on another when conﬂicting for possession of the females (Grant and Temple-Smith, 1998). In extinct mammals, the role that venom could have played is a matter of debate. Hurum et al. (2006) showed that extratarsal spurs like that of the platypus were present in most Mesozoic mammal groups, possibly with a defensive function and to a lesser extent for intra-speciﬁc competition or predation. Fox and Scott (2005) found grooved structures in the upper canines of Bisonalveus browni Gazin, 1956, and in loose lower teeth of another Paleocene mammal, similar to the one that S. cubanus displays in its second lower incisors to inject venom. Those evidences led the authors to hypothesize that venomous mammals could have been more widespread after the K-T boundary than previously thought. A third example of a fossil venomous mammal came from the Spanish Early Pleistocene site of Sima del Elefante, where ´ Cuenca-Bescos and Rofes (2007) noticed the possession of ´ a venom delivery apparatus in Beremendia ﬁssidens (Petenyi, 1864). In fact, B. ﬁssidens was a rather frequent form in the European micromammal communities from the Pliocene until the middle Pleistocene, as demonstrated by the more than 100 European fossil sites from where it has been reported (see RzebikKowalska, 1998). An Asian species of the genus, B. pohaiensis (Kowalski and Li, 1963), has been found in some localities from 928 ´ FURIO ET AL.—PALEOBIOLOGY OF BEREMENDIA China (Jin and Kawamura, 1996). Surprisingly, a relationship between the distribution and the paleobiology of the genus was never established. Based on its ubiquitous presence, several authors concluded that the genus Beremendia should be considered an opportunistic element within the micromammal community (Reumer, 1984; Rzebik-Kowalska, 2000; Popov, 2003). Assuming such opportunistic strategy, Fanfani and Masini (1998) even considered the presence of Beremendia to be useless for paleoecological interpretations. Although these few remarks are based on good observations, none of these studies properly attempted to elucidate the paleobiological proﬁle of Beremendia in detail, but rather speculated on it. This blank in our knowledge has been probably enhanced by the lack of a living relative with which to ´ compare. In the contribution of Cuenca-Bescos and Rofes (2007), the possession of a venom delivery apparatus was suggested to be related to its uncommonly large size as an adaptation to the hunt of big prey. We believe that such explanation is not so evident considering some characters of B. ﬁssidens from a functional point of view. Making use of the new ﬁndings of Beremendia ﬁssidens from the Early Pleistocene hominid site of Dmanisi (Georgia) and the Late Pliocene ﬁssure inﬁlling of Almenara-Casablanca 1 (Spain), a new ecological explanation for the occurrence of this genus is herein provided. A review of the published data about Beremendia is compiled with the analysis of this new fossil material in order to assess its functionality from morphological and biomechanical points of view. The comparative morphology with some extant species of shrews and their ecological preferences provide a good approach to restore the most likely behaviour of Beremendia in life. Simultaneously, its paleobehaviour provides clues to the environments in which the poisonous nature of these shrews was favored. Methodology Localities and Storage of the Material Studied—The fossil material of Beremendia ﬁssidens analyzed comes from the Early Pleistocene hominin site of Dmanisi (Georgia, Southern Caucasus) and the Late Pliocene karstic inﬁlling of AlmenaraCasablanca 1 (Eastern Spain). The location of these localities is provided in Figure 1. The material from Dmanisi (one complete maxillar with left I1–M2 and right A1–M3 dental rows, and one left anterior side of the maxillar preserving the I1–A4 row) belongs to the collection of the Georgian National Museum. The material from Almenara-Casablanca 1 (two left series P4–M2 [IPS-5952, IPS-5953] with their corresponding bone areas for the insertion of the roots, and one complete right hemimandible [IPS-5954] with its complete dentition) is part of the ` collection of the Institut Catala de Paleontologia. The recent material of Crocidura russula (Hermann, 1780) and Sorex araneus Linnaeus, 1758, belongs to the personal collection of M.F. Equipment—The study has been carried out using a stereoscopic binocular Leica MZ6 for direct observation, and a Hitachi S-570 scanning electronic microscope (SEM) from the Servei de ` Microscopia (Universitat Autonoma de Barcelona) to take general photos of the fossil and recent material, their dental wear pattern, and for detailed analysis of the external histology of the symphyseal areas. Measurements and Nomenclature—The nomenclature employed to describe the features of the soricid mandibles follows Repenning (1967). The nomenclature of the dental elements follows Reumer (1984). The names of the craniomandibular muscles ¨ are the ones used by Dotsch (1994). MORPHOLOGICAL CHARACTERS OF BEREMENDIA Position and Relationships of Beremendia The genus Beremendia includes three well-differentiated species: B. ﬁssidens, B. minor Rzebik-Kowalska, 1976, and B. 929 FIGURE 1. Geographical situation of the localities where the material comes from. For further detail of these sites, see Lordkipanidze et al. ´ (2007) for Dmanisi, and Furio et al. (2007) for Almenara-Casablanca. Almenara-Casablanca 1 is placed 400 m to the East of the ﬁssure inﬁlling of Almenara-Casablanca 4. pohaiensis, according to Jin and Kawamura (1996) and RzebikKowalska (1998). Although the classiﬁcation of the soricids by Repenning (1967) placed the genus among the Neomyiini, an updated vision of the phylogeny considers it as a member of the Beremendiini, a tribe of pigmented-toothed shrews without any extant representative (Reumer, 1984, 1998). Size differentiates the big forms B. ﬁssidens and B. pohaiensis from the somewhat smaller B. minor. Other than that, the three species included in the genus share the features analyzed and discussed in the following paragraphs. The members of the fossil genus Beremendia stand out from all the Euroasiatic Plio-Pleistocene soricines by some peculiar traits that are easily recognizable (Fig. 2). Most of these dental and mandibular characters of Beremendia have been exhaustively described for taxonomical purposes (see Repenning, 1967, Gureev, 1971; Rzebik-Kowalska, 1976; Jammot, 1977; Reumer, 1984; Jin and Kawamura, 1996; Fanfani and Masini, 1998; Popov, 2003; Ro´ fes and Cuenca-Bescos, 2009), but no special attention has ever been paid to their biomechanical implications. One of the most peculiar characters of Beremendia is the structure of its articular mandibular condyle. The articular condyle is a major characteristic that divides the family Soricidae into its corresponding subfamilies. Based on this trait, Repenning (1967) established ﬁve main divisions, distinguishing the three subfamilies Limnoecinae, Crocidurinae, and Soricinae, the latter being subdivided into the three different tribes Blarinini, Soricini, and Neomyini. Reumer (1984) reﬁned this classiﬁcation, dividing the Soricinae into seven different tribes by the addition of the Notiosoricini, the Soriculini, the Beremendiini, the Amblycoptini, and the Allosoricini. In the last update of the taxonomy of the group, Reumer (1998) established a quite stable classiﬁcation of ﬁve different subfamilies, in which the soricines contained the tribes Soricini, Notiosoricini, Beremendiini, Anourosoricini, Blarinini, Blarinellini, and Neomyini (or Nectogalini according to Hutterer, 2005). In this latter classiﬁcation, the articular condyle is also a key feature that characterizes the tribe. It must be emphasized that the morphology of the condyle is not just a feature randomly selected along the evolution of the soricids, but has a biomechanical signiﬁcance. Because the anterior part of the skull became tube-like in the most primitive shrews, this region absorbed the stresses of the bite force. The zygomatic arches were therefore free from mechanical stresses and they were consequently reduced (Preuschoft and Witzel, 2002). This is one of the main synapomorphies that characterizes the family Soricidae. The absence of a supporting point for muscle attachment resulted in a very peculiar way of moving the mandible, which is reﬂected in the morphology of the condyles. The main muscle in the chewing action became the temporal muscle in the most archaic crocidosoricine 930 JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 30, NO. 3, 2010 FIGURE 2. Speciﬁc characters of Beremendia ﬁssidens. The anterior (AFC) and the posterior (PFC) functional complexes are separated at right and left sides of the plate by means of solid parallel lines. The AFC comprises: 1, ﬁssident upper incisor; 2, wear pattern of the talon in I1 and anterior crest of the A1; 3, anterior extreme of the i1 strongly pigmented and turned upwards; 4, well-delimited groove running along the medial face of the i1; 5, double-bulged symphysis delimiting an inner longitudinal fossa. The PFC comprises: 6, Strong pigmentation in upper and lower molars; 7, ﬂat wear of the cusps and crests of the molars; 8, tip of the coronoid process tilted anteriorly; 9, ascending ramus leaning labially; 10, articular condyle separated in two articular facets, being the lower one very anterolingually placed; 11, short angular process; 12, broad and half-twisted horizontal ramus; 13, small inner temporal fossa. shrews from the Oligocene and the Miocene, instead of the masseter (Dannelid, 1998). This primitive condition is retained by the extant white-toothed shrews, all of them included in the subfam´ ilies Crocidurinae and Crocidosoricinae sensu Furio et al. (2007). In these shrews, the mass of the masticatory muscles in relation to the total body mass is larger than in the red-toothed shrews ¨ (Dostsch, 1985, 1994). Another soricid clade, the pigmented-toothed shrews, i.e., subfamily Soricinae, developed a mandibular articulation with two separated facets, subsequently modifying the original structure displayed in crocidurines and crocidosoricines. This new model provided a greater mobility of the mandible with relatively lower energy expenditure or, at least, proportionally less masticatory muscle mass required. Among soricines, the separation of the facets acquired different degrees, ranging from low differentiation, e.g., Soricini or ‘Sorex-like,’ with moderate interarticular areas, to a high one, e.g., Soriculini, or ‘Neomys-like,’ with narrow interarticular areas, and Beremendiini and Blarinini, or ‘Bla- rina-like,’ with broad interarticular areas. The three main types of soricid articular condyles are shown in Figure 3 to summarize the main changes linked to the general construction of the mandible. The members of the tribe Beremendiini, clearly exempliﬁed by its type-genus Beremendia, led this trend of separating the articular facets to its maximum extreme. The lower articular facet of Beremendia is displaced towards an anterolingual position with regard to the upper articular facet (Fig. 3). The most immediate beneﬁts of such construction can be successfully explained as an improvement of the efﬁciency of the mandibular mechanics. The double articulation condyle guides two different movements. The ﬁrst one is the natural rotation of the mandible parallel to the sagital plane, in which the lower facet acts as a hinge. The second one is the lateral sliding in the horizontal plane, a displacement allowed by the upper facet. It is noteworthy that as a counterbalance to such mechanic improvements, one action immediately excludes the other, so the mandibular movements of ´ FURIO ET AL.—PALEOBIOLOGY OF BEREMENDIA 931 FIGURE 3. Schemes of lingual views of right hemimandibles in Crocidura, Sorex, and Beremendia (above) and posterior views of their corresponding articular condyles (SEM photographs, below) in a representative species for each genus: C. russula, S. araneus, and B. ﬁssidens (IPS-5954). The structure of the condyle and the orientation of its elements are related with the modiﬁcation of the whole mandible and the insertion areas of the main masticatory muscles. Note that the condyles of Crocidura and Sorex do not deviate signiﬁcantly from the axis deﬁned by the ascending ramus of the coronoid process, clearly differing from that of Beremendia. In the latter genus, the condyle is laterally twisted to the lingual side, and the upper and lower facets are clearly separated to guide the movements of biting and chewing. Besides, this disposal leads to a more medial insertion of the temporal muscles in Beremendia (see posterior view), which is related with the anterior tilting of the coronoid process (scheme of mandible in lateral view). The more anterior position of the lower facet of the articular condyle of Beremendia permits a greater surface for the attachment of the external pterygoid than in Sorex and Crocidura. However, the short angular process reduces the available area for the attachment of the internal pterygoid and the masseter muscles. The digastricus muscles attach at different lengths of the horizontal ramuses in each species. Angle α is drawn following Carraway and Verts (1994), subtended by a line from the apex of the coronoid process through the distalmost extension of the lower condylar facet and a line along the ventral edge of the dentary. Abbreviations: c.p., coronoid process; dig., digastricus muscle; ex. pter., external pterygoid muscle; in. pter., internal pterygoid muscle; l.f., lower facet of the articular condyle; mass., masseter muscle; temp., temporal muscle; u.f., upper facet of the articular condyle. biting and chewing cannot take place at the same time. The jaw movements are limited by the restrictions imposed by the articular condyle. All the dental, maxillar, and mandibular modiﬁcations are related with one of these two movements. Therefore, the grouping of some features deﬁnes a ‘posterior functional complex’ (PFC here after) linked to the chewing action and an ‘anterior functional complex’ (AFC here after) linked to the biting action. The PFC concerns modiﬁcations on the coronoid and angular processes of the mandible, on the internal temporal fossa, and the strong pigmentation and bulbosity (also known as amblyodonty) of the molars. The AFC includes modiﬁcations on the ﬁrst upper and lower incisors, and the use of the ﬁrst upper antemolar. Some other peculiar characters are involved in both actions, like the shape of the symphyseal area or the half-twisted horizontal mandibular ramus, and they are expected to have served in a dual way by being involved in both bite and chew. The occlusal wear pattern has a different meaning depending on the functional complex involved. Finally, the cranial stoutness is a feature probably linked to other specializations on the postcranial skeleton. All these characters and their functionality are analyzed in the following paragraphs. Analysis of the Posterior Functional Complex Coronoid Process—In most shrews, the coronoid process does not have a signiﬁcant angle with respect to the condyle in posterior view, and thus all the force employed in the longitudinal rotation of the mandible is supplied by the temporal muscle. Carraway and Verts (1994) made use of this conﬁguration to estimate the biting force in different extant species of Sorex Linneaus, 1758, approaching the mechanical action of their mandibles as simple ‘Type I levers,’ in which their lower facets were the fulcrae. Although such assumption is not so directly applicable in the case of Beremendia, it constitutes a good starting point to develop the explanation of its mandibular mechanical modiﬁcations. In Beremendia, the tip of the coronoid process is rather short and pointed, and it bends strongly forwards (Reumer, 1984). With respect to the model of Carraway and Verts (1994), the expected bite force of Beremendia would be consequently reduced, because the angle α decreases—it increases the complementary ‘θ’—and therefore the different points at the resistance moment arm—teeth—lose relative force. Furthermore, Beremendia displays its coronoid process laterally tilted outwards with respect to the condyle when the axis of the lower articular facet lies 932 JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 30, NO. 3, 2010 Beremendia as well. The combination of amblyodonty with exodaenodonty is functionally useful in the tasks of crushing, although they considerably decrease the capabilities of piercing. Reumer (1984) considered the presence of exodaenodont teeth in shrews to be indicative of a diet based on abrasive items, probably gastropods. Pigmentation—All of the species of Beremendia display strongly pigmented teeth, with a dark red—sometimes even completely black—stain that covers the upper halves of the labial faces in lower teeth, and most of the lingual faces of the main cusps of the upper teeth (Rzebik-Kowalska, 1976; Reumer, 1984). The presence of pigmentation in the teeth is one of the main traits characterizing the soricine shrews. The amount of pigmentation varies between genus and species within the subfamily, from almost unnoticeable remains at the tips of the main cusps, to those covering most of the labial and lingual faces of the teeth. The coloration is due to the presence of iron in the outermost layer of the enamel, and depending on the concentration of this element therein, the tonality ranges from an inconspicuous orange to a strong dark one. The function of the pigmentation in the teeth of the shrews is still to be demonstrated. However, it can be ascertained that the pigmented parts of the teeth are more resistant to the wear than the non-pigmented ones. In the case of rodents, in which the anterior faces of their ever-growing incisors are covered by pigmented enamel, the greater resistance against the wear than the non-pigmented zones creates a frontal shearing blade useful in the tasks of gnawing. Strait and Smith (2006) demonstrated that the higher concentrations of iron in the enamel of the extant shrew Blarina brevicauda were found in the cusps related with crushing and grinding functions, instead of those involved in shearing. Consequently, the stained teeth of Blarina were rather favored to prevent the rapid erosion produced by handling abrasive items. Such assumption can be extended to other soricid dentitions that display an extremely similar pattern of coloration to that of Blarina such as the strong and widespread pigmentation of Beremendia. Further evidences on the function of the pigmentation come from the analyses of the dental wear pattern. Occlusal Wear of the Posterior Elements—The occlusal wear pattern is the result of the erosion of the enamelled surface by the interaction with the opposite teeth or food items. The pattern is strongly inﬂuenced by the occlusal morphology and the enamel microstructure (Evans and Sanson, 1998, and references therein), and by the distribution of the pigmentation on the outer layers of the enamel (Strait and Smith, 2006). In Blarina brevicauda, the best-studied example among the Recent shrews, the wear pattern of the molars and the antemolars tend to create ﬂat surfaces in which the dentine only crops out when the thick layer of enamel is completely eroded (Dannelid, 1998). The wear pattern in the teeth of Beremendia is extremely similar to the one found in Blarina. The tallest crests of the labial face in upper and lower teeth—mainly in molars—show strong ﬂat erosion, whereas depressions and minor cusps tend to remain practically untouched (Fig. 4). In some other species of shrew, the cutting action developed between the posterior crest of the upper fourth premolars and the paralophid of the ﬁrst lower molars creates ‘carnassiallike’ blades in both. Their absences in Beremendia better agree with a chopping-action of their teeth rather than a cutting one. horizontally. The resulting force of the temporal muscle is obviously reoriented with respect to the ‘Sorex-condition,’ thus getting a noticeable medial component (see Fig. 3). In this aspect, Beremendia reproduces the condition of the genus Blarina Gray, 1838. Both genera share a similar structure of the articular condyle, only differing in the superior pterygoid fossa. In Beremendia this fossa is represented by a deep pit, opposite to Blarina, in which it is ﬁlled with bone to thicken the condyloid process (Repenning, 1967). In the latter genus, the large ‘part 1’ of the temporal muscle faces laterally the attachment to the coro¨ noid process (Dotsch, 1994). Proportionally, the forces provided by the temporal muscles of Blarina and Beremendia are longitudinally reduced but medially enhanced compared with those of Sorex. Angular Process—The angular processes of the three known species of Beremendia are very short. The external faces of the angular process serve for the attachment of the masseter muscles, whereas their internal ones act in the same way for the internal pterygoid muscles. Given the orientation of these muscles in shrews, they are not only involved in the adduction of the jaws—as usual in most mammals—but also in their protraction (Gasc, 1963). Consequently, the small size of the angular process in Beremendia indicates that the masseter and the internal pterygoid muscles were certainly reduced and thus the longitudinal displacements of the jaws played a minor role. Nonetheless, the masseter was still functional in this genus, judging by the rather well-developed zygomatic process (Fig. 4). In any case, the force lost to carry out the adduction of the jaws should have been consequently supplied by the action of somewhat more developed external pterygoid muscles. Internal Temporal Fossa—The internal temporal fossa of Beremendia is relatively small to its general size. Compared with other shrews—even other soricines—there is a noticeable decrease in size of the internal temporal fossa. This pocketed opening serves for the attachment of the pinnated third part of the temporal muscle, so its importance in the mastication movements is also deduced to have been consequently reduced. The major role played by this part of the muscle is thought to be the stabilization of the lower jaw, especially in the area of the double joint ¨ (Dotsch, 1994). It is therefore expected that not only did it not constrain the lateral rotation of both hemimandibles, but it also permitted so in a considerable degree during the chewing action of Beremendia. Molar Bulbosity—All the teeth of Beremendia ﬁssidens are characterized by an enlargement of their cusps and their enamel thickness. The bulbosity of the molars—known as ‘amblyodonty’ by several authors—has been reported for some other fossil soricid genera, but was never highlighted for Beremendia. In fact, this is a quite subjective property, and it strongly depends on the author whether a dentition is considered bulbous or it is not. However, it can be ascertained that the teeth of Beremendia are really stout when they are compared with those of Sorex, Neomys Kaup, 1829, and of most Crocidura Wagler, 1832 (Fig. 4). The sharpness of the cusps and crests in the teeth of insectivores has a direct relationship with their capabilities to break down the exoskeletons of insects. The sharper the elevations are, the easier they penetrate the cuticles and the inner tissues of insects. However, this character is apparently constrained by the risk of tooth fracture when hard or brittle food is dealt with (Evans and Sanson, 1998). Therefore, the thickening of the enameled layer and the blunting of the angles on the occlusal surface provide the best ratio between the crushing properties and the resistance of the structures. In the insectivores, amblyodonty is usually related with exodaenodonty, extremely exaggerated in some dimylids, another phenomenon in which the teeth overlap the jaw bone (Ziegler, 1999). A certain degree of exodaenodonty can be appreciated in Analysis of the Anterior Functional Complex Grooved First Lower Incisor—The ﬁrst lower incisor of Beremendia ﬁssidens is acuspulate and there is a well-marked groove running along the medial side of its crown. This feature was already noticed in the ﬁrst accurate descriptions of the species, and even a couple of piped structures in the lower incisors have been reported in some cases (see Rzebik-Kowalska, 1976, 2000). ´ FURIO ET AL.—PALEOBIOLOGY OF BEREMENDIA 933 FIGURE 4. SEM photographs of different parts of the PFC in Crocidura, Sorex, and Beremendia. A, the wear pattern of the molars reveals that the erosion in Crocidura and Sorex creates cutting edges, whereas in Beremendia (IPS-5954) the general trend is to create rather ﬂat surfaces from the more elevated cusps and crests. B, the size and shape of the zigomatic processes reveal striking differences between the extant forms and the extinct genus. In the living shrews, the size ranges from the inconspicuous of Crocidura to the thin but well-deﬁned one of Sorex. In Beremendia (IPS-5952), the zigomatic process is stout and broad, but comparatively short to the general size of the skull. Abbreviations: m., lower molar; z.p., zigomatic process. Fox and Scott (2005) documented the presence of a very similar structure in the upper canines of Bisonalveus browni, a Paleocene North American pantolestid mammal. According to these authors, such specialized structures were involved in the injection of venomous saliva, and no similar ones could be found in the extant shrew Blarina brevicauda, which transmits the toxic saliva by means of rapid bites, not necessarily using the anterior teeth. Nevertheless, there is some controversy on the functionality of the canine grooves of B. browni (see Folinsbee et al., 2007; Orr et al., 2007), because these structures are present in non-venomous extant species of mammals as well. ´ Cuenca-Bescos and Rofes (2007) stated that the structure present in the lower incisor of Beremendia was analogous to the one found in the canines of Bisonalveus, thus implying its similar venomous nature. Paradoxically, in the case of Beremendia the venomous function is more reliable than in its supposed pantolestid analogue (Folinsbee et al., 2007), because the saliva of some living shrews has been demonstrated to be toxic. In fact, the existence of this groove in some extant shrews and its relationship with the venomous nature of their salivas was already known a long time ago, and the ﬁrst observations regarding this date back to the end of the 19th century (Jammot, 1983). The link between this structure and the venomous bites of the shrews was perfectly described by Jammot (1983:255), “En vue linguale, on peut observer ces grandes incisives; au tiers de la hauteur est une rainure assez profonde qui suit le trac´ inf´ rieur de la dent e e jusqu’` son extr´ mit´ . On comprend mieux l’int´ rˆ t de cette struca e e ee ture lorsque l’on sait que les glandes sousmaxillaires s’ouvrent a ` la base de ces incisives et que la salive s’´ coule par le canal cone stitu´ par les deux goutti` res lat´ rales juxtapos´ es de chacune des e e e e incisives. La salive des glandes sousmaxillaires des Soricidae est venimeuse (. . .)” The toxic properties of the bites of some recent species of soricids are well documented, and even some injuries on humans have been reported several times (Nowak, 1991). There is no doubt that the lower incisor is always the main element involved in this venomous injection, and that the groove acts as a channel to conduct the venom from the salivary glands to the anterior part of the lower incisor. Nevertheless, although the grooves do exist in most of the species (Fig. 5), they are usually difﬁcult to discern due to their shape and/or smaller dimensions. In other words, this structure is clearly visible in Beremendia, for two reasons. The ﬁrst one is that the species of this genus reached bigger dimensions than the soricids usually do, and the character was thus overdeveloped. The second one is that the furrow was better delimited and canalized than in other species, in which it is just outlined as a slight concavity of the medial surface of the lower incisor. It must be added that the grooved structure of Beremendia terminates right in the most anterior extreme of the tooth. This 934 JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 30, NO. 3, 2010 FIGURE 5. SEM photographs showing some differences found in the AFC of Crocidura, Sorex, and Beremendia. The shaded areas correspond to the magniﬁed SEM photograph immediately below. A, the grooved structure in the medial face of the lower incisor is present in the three species, although it is only ﬁnely piped in the fossil one. Moreover, the anterior extreme is curved upwards more abruptly than in the other two shrews. Such conﬁguration ensures a more reﬁned ability for the conduction and injection of the venom into the prey. B, articulation areas of the symphyses (intermediate photographs), and detail of the enamel-symphysis zone tissues magniﬁed ×100 in respect to their corresponding photos above (lower photographs), both showing signiﬁcant differences. The recent specimens show porous tissue cropping out in the junction area, indicating that both hemimandibles were somehow fused in life. In the contrary, the surfaces of the longitudinal bulges are completely covered by compact bone in Beremendia, indicating that each half-part of the mandible preserved a certain mobility degree for lateral rotation. Abbreviations: c.b.t., compact bone tissue; en., enamel; gr., groove; p.b.t., porous bone tissue; sym., symphysis. ´ FURIO ET AL.—PALEOBIOLOGY OF BEREMENDIA 935 FIGURE 6. The wear pattern in the I1 and A1 (grey areas in the upper schemes) found in Crocidura and Sorex are different from that of Beremendia. In the former genera, the wear areas of the upper incisors and the ﬁrst upper antemolars are not continuous, just affecting the zone between the apex and the talon of the I1, and the main cusps and the posterior crests of the upper antemolars. In the latter genus, the talon of the I1 disappears, thus connecting the occlusal area of the ﬁrst upper incisor with the anterior crest of the ﬁrst upper antemolar. This pattern indicates that both teeth served as a single functional unit in Beremendia to grasp some kind of abrasive objects. apical part is strongly bent upwards, resulting in a sharp single elevation in all the species of the genus (see Reumer, 1984; Jin and Kawamura, 1996; Rzebik-Kowalska, 1976). Such conﬁguration properly denotes its function as a ‘puncturing-and-injecting’ structure. Fissidence of the First Upper Incisors—A very peculiar feature displayed in Beremendia is the shape of its ﬁrst upper incisors. The upper incisors of Beremendia have strongly developed medial tines. Such character is called ‘ﬁssidence’ or ‘biﬁdence,’ and it has occurred in different evolutionary lineages of soricids. A strong ﬁssidence is characteristic of all the species included in the genus, and it is a character clearly reﬂected in the name of its most widespread species, B. ﬁssidens. The ﬁssidence has been said to be a burrowing and/or a carcass-feeding structure (Dannelid, 1998). Carraway and Verts (1994) demonstrated that there is an inverse relationship between the degree of ﬁssidence and/or non-parallelism of the apexes in the upper incisors, and the biting force of the lower incisors in Sorex. This is at ﬁrst a quite unexpected correlation, because the lack of a sharp apex increases the biting surface reducing the pressure applied over the prey, so the weakest pressure of biting should be supplied by an increase in the power of the jaw muscles to get the same effect. It must be assumed that a softer bite than should be expected to correspond to their size beneﬁtted the ﬁssident shrews in some way. Wear Pattern of the Anterior Elements—On inspecting recent specimens of Crocidura, Sorex, or Neomys, little wear was found on the I1 and A1 and in the latter element only on the posterior crest (Fig. 6). In the specimens of Beremendia from Dmanisi, the wear of the talon of the I1 is noticeably more developed, and it is uninterrupted until the highest cusp of the ﬁrst upper antemolar, showing that the two served as a single functional unit (Fig. 6). Occlusion cannot explain this pattern, because the upper and the lower anterior dentitions do not touch completely when the oral cavity is closed. The zone between the incisors and the last premolars is even left as an empty space. Consequently, the wear of the ridges and cusps of the antemolars can only be reliably explained as a consequence from their function in grasping of food using the oral cavity. Features Involved in Both Anterior and Posterior Functional Complexes The Symphyseal Area—A conspicuous fossa is present in the symphyseal area of the hemimandibles of Beremendia ﬁssidens, just behind the medial base of the crown of the lower incisor. This structure is delimited by the presence of two longitudinal bulges ´ running back till beneath the talonid of m1. Cuenca-Bescos and Rofes (2007) considered the presence of a conspicuous fossa in the symphyseal area of the mandible of Beremendia as evidence of large amounts of connective tissue formerly housed there to reduce the mobility between both hemimandibles to improve the bite force exerted over the prey. In fact, double-articulated symphyseal areas are not uncommon in shrews. The modiﬁcations of this structure in other genera are a product of the relative size of their horizontal ramuses (see illustrations in Repenning, 1967, and Reumer, 1984, or any other work depicting mandibles of shrews in medial views), and the kind and quantity of the ligaments attached on this area. The mobility of the symphyseal articulation directly depends on this latter feature, and it can be consequently deduced from the histological study of the bony tissue in this area. In crocidurines, the rotational movements are prevented in ¨ favor of transversal ones (Dotsch, 1994), thus strongly reducing the mobility in the symphyseal area. Because of that, the bulges of each hemimandible tend to be fused in some degree with their corresponding opposites. Thus, the disarticulation of the mandible results in the cropping out of porous bone tissue. In some soricines the junction area can be even wider. Examples of both cases are shown in Figure 5. 936 JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 30, NO. 3, 2010 FIGURE 7. SEM photograph of the right hemimandible of Beremendia ﬁssidens from Almenara-Casablanca 1 (IPS-5954) in frontal view, with its specular medial image, showing three different moments. A, when temporal muscles are contracted. In this case, the lower symphyseal bulges diverge, whereas the upper ones converge. B, when muscles are relaxed. In this case, both symphyseal bulges remain at a similar distance. Note that in this resting position, the axes of the lower articular facets and most of the dental crests stay horizontal. C, when the internal pterygoid muscles are contracted. This is an inverse situation to A in which the lower bulges converge whereas the upper ones diverge. White arrows indicate the main stresses occurring in each point. Dotted-line arrows indicate the contraction of the main muscle involved in each position of the lateral rotation. The solid line indicates the connection of the axes deﬁned by the medial extension of the lower articular facets. However, Beremendia displays outer surfaces of the bulges covered by compact bone (Fig. 5), indicating that these convexities were not fused, but were rather designed to support the strong rotational movements of the hemimandibles. Even if these bulges were attached by connective tissue, it would not reduce the mobility, but would rather permit it to a certain extend. It must be emphasized that the chewing in soricids includes a great variety of movements depending on the existence of mobile symphyses and some lateral divergent tilting of both hemimandibles during ¨ the jaw opening does exist (Dotsch, 1994). This is not surprising because a fused symphysis is indicative in most mammals of a greater degree of transverse movements and is correlated with low cusped teeth with horizontally oriented occlusal wear facets (Lieberman and Crompton, 2000), which is by no means the case of Beremendia. The morphology of the mandible in shrews can be modiﬁed because of the stress-forcing of masticatory muscles during the ontogeny (Carraway et al., 1996; Badyaev and Foresman, 2004) or due to evolutionary processes along the geological time, i.e., by enhancement of this trait along a certain lineage (Jammot, 1983). In whatever way, no ossiﬁed symphysis has ever been reported in any specimen of Beremendia, although it would be a good expected solution to reinforce it if fusion was mechanically required. This can be considered as a further indication that the anteromedial junction of the mandible of Beremendia was well ﬁxed, but unavoidably mobile, for the masticatory action. Lateral rotations of the hemimandibles undoubtedly played a major role in the chew of these shrews, as shown in Figure 7, whereas transverse motions were certainly prevented and/or extremely reduced. Horizontal Ramus—The horizontal ramus of Beremendia is stout and half-twisted. When the upper and the lower teeth occlude, the anterior part acquires a rather vertical disposal, whereas its posteroventral extreme dips medially as a result of the twist. That leads to just one suitable disposition, in which the ascending ramuses of the hemimandibles are opened laterally whereas the cusps and crests of the upper and the lower teeth are squeezed in their complementary depressions. No other orientation for the ramus is permitted when the upper and the lower tooth rows occlude. Such partial twist can be seen as a mechanical solution to resist and release the pressure exerted over the ascending ramus of the coronoid process of each hemimandible, and drive it along the horizontal ramus up to the symphyseal area. Cranial Stoutness—The skull of Beremendia ﬁssidens is notably hard and robust. The cranial capsule of the shrews is usually thin, so it is rarely preserved in the fossil record. However, noticeable differences in the properties of the skull get reﬂected in the construction of the maxillar bones. It is noteworthy that a lot of authors have reported good specimens of complete palates of Beremendia, including the complete dental series (see RzebikKowalska, 1976; Reumer, 1984; Fanfani and Masini, 1998; Popov, 2003; and the material from Dmanisi in this work), a rather rare phenomenon within fossil shrews, which gives an idea about its high mechanical resistance. Among the extant shrews, the shortest and most heavily constructed skulls are typically found in the digging forms Anourosorex and Diplomesodon and the large commensal species Suncus murinus (Hutterer, 1985). The cranial stoutness of Beremendia can thus be considered indicative of a similar fossorial way of life. DISCUSSION The Beremendia Contradiction Having checked the characters involved in the AFC and the PFC, it is noteworthy that some contradictions exist if all the available explanations on the functionality of each feature are ´ taken into account. The hypothesis of Cuenca-Bescos and Rofes (2007) that Beremendia was specialized in hunting big prey could be supported by the grooved lower incisor to inject venom and by the ﬁssident upper incisor, if its functionality was to serve as a carcass-feeding tool, as stated by some authors (see Dannelid, 1998). Nonetheless, the rest of the characters do not support this idea. Although the AFC could be functionally designed for a hunter or a scavenger animal, the inter´ pretation of Cuenca-Bescos and Rofes (2007) of the immobilized symphyseal junction does not seem tenable (see above). Moreover, in living shrews the lower condylar facet does not permit to reach more than 30 degrees of amplitude (Jammot, 1983). If the ﬁrst lower incisor was used to inject the venom throughout the skin of some vertebrates, the protraction movement of the mandible should have been enhanced to provide a good thrust. The length and the shape of the angular process are not in agreement with the idea that Beremendia displayed such a capability. Judging by the attachment surfaces in the zygomatic and angular processes, the masseter was aligned almost straight and horizontal, parallel to the anteroposterior axis of the skull (Fig. 8). So it would have served for the protraction of the mandible. However, its size is rather small in relation to the general cranial dimensions, and the thrust movement would not have been long and powerful enough to penetrate the skin of big prey. Moreover, the apexes of the upper incisors should not have been ﬁssident, but rather sharp-pointed, ´ FURIO ET AL.—PALEOBIOLOGY OF BEREMENDIA 937 FIGURE 8. Reconstruction of Beremendia based on the partial remains from Dmanisi (maxillar) and Almenara-Casablanca 1 (dentary bone). Mandible and maxillar in left lateral view (A), and in ventral view (B). Reconstruction of the muscles involved in mastication in Beremendia in left lateral view (C), and in ventral view (D). The digastrics muscles are involved in the opening of the mandible. They have a posterior double attachment, anchoring in the anterior of the paraoccipital process and the lateral border of the stylohyoid muscle. Anteriorly, the digastric muscles are anchored in the ventral border of the dentary, but the exact point of attachment is difﬁcult to discern in Beremendia, due to the absence of a similar digastric tubercle to that of Blarina. The anterior attachment of the digastric muscles is deduced to be intermediate between that of Blarina and the mandibular symphysis. The internal pterygoid connects the basal zone of the skull posterior to the maxillar with the medial surface of the angular process. The cranial fossil fragments of Beremendia available do not preserve the zone of anterodorsal attachment, but the mandibles do show the area of posteroventral insertion, which is remarkably excavated and delimited by a central ridge running along the medial face of the angular process ´ in B. pohaiensis, according to Jin and Kawamura (1996). (Artwork by O. Sanisidro.) so they could serve to get a better hold on the victim. The short protraction displacement allowed in Beremendia rather indicates more utility for the mastication. Apart from this, there are also other non-adaptive features in the PFC dentition of an expected hunter. The amblyodonty and the incipient exodaenodonty, the absence of ‘carnassial-like’ blades, the distribution of the pigmentation, and the wear pattern of the molars do not at all support the idea of Beremendia as a ﬂesh eater. Nevertheless, rejecting the idea of Beremendia as a big-prey hunter revives the questions as to the degree in which the possession of a venom-delivery system beneﬁts a small insectivore micromammal versus a nonvenomous sister species (Folinsbee et al., 2007). New light can be shed on these questions if some of the above mentioned characters are correlated with current knowledge of physiology, ecology, and ethology of shrews. Not Hunting Big, But Hoarding Small The analysis of the muscular attachments and their implications for the mobility of the mandible provide extremely valuable data in bringing Beremendia back to life. As already stated, the condyle structure of this genus—and soricines in general—does not allow biting and chewing movements at the same time. Nevertheless, the temporal, the masseter, the internal ptergoid, and the external pterygoid muscles are involved in the closure of the ¨ mandible (Dotsch, 1985), and by analogy with the case of Blarina ¨ remarked by Dotsch (1994), a great variety of different movements were expected to take part in the chewing action. The parts ‘1’ and ‘2’ of the temporal muscle are the main ones ex- ¨ erting pressure on the food in all shrews (Dotsch, 1994). Given that the coronoid processes of Beremendia are laterally bent outwards, the resulting rotational movements are reinforced. The non-fused symphysis (Fig. 5) clearly permits this mobility (Fig. 7). Thus, the posterior teeth would act as ‘nutcrackers’ at each side, holding the hard items between the upper and the lower molar rows and progressively increasing the pressure by the contraction of the internal parts—1 and 2—of the temporal muscle. To summarize, transverse motions were extremely constrained in the adduction, whereas the lateral rotation of the hemimandibles was enhanced. Horizontal movements, protraction ¨ and retraction, were also reduced and limited. Dotsch (1994) noticed that the mastication of the food in shrews was carried out on just one side of the mandible, and that it changed regularly following a certain pattern depending on the kind of food dealt with and the stage of the reduction sequence. A heterogeneous food—cat food mixed with resistant particles—was chewed slightly more irregularly than mealworms, and much more so than homogeneous food. The equalization of adduction, protraction, retraction, rotation, and transverse movements is a clear advantage in changing the food in the oral cavity from one ipsilateral active side to the other. The term ‘intractability’ was recently coined by Evans and Sanson (2005) to make reference in a broad sense to this difﬁculty to deal with invertebrate exoskeletons before their ingestion, instead of the sometimes oversimpliﬁed and imprecise one of ‘hardness.’ Thus, the muscular condition displayed by Beremendia is well described as indicative of a diet mainly based on ‘intractable’ items. 938 JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 30, NO. 3, 2010 is because most of the shrews are clear examples of opportunistic feeders, following a diet strongly dependent on the relative abundance of their potential victims (Churchﬁeld, 1994). That explains why, for instance, Blarina combines molars that can easily break gastropod shells (Dannelid, 1998) with strong venomous saliva that can paralyze or even kill small vertebrates (Kita et al., 2004). The case of Beremendia is somewhat peculiar in this aspect. The members of this genus show the most extreme modiﬁcations in dental and mandibular characters among all the PlioPleistocene shrews, clearly indicating that Beremendia followed a highly specialized diet. At ﬁrst, extreme specializations represent a considerable handicap for shrews, because trophic specialists are much more frequently confronted with scarcity of their particular kind of food than generalists (Rychlik and Jancewicz, 2002). However, the handicap derived from the high specialization can be minimized if the potential prey is abundant in most environments, highly diversiﬁed within the group, or geographically widely spread. This is exactly the case of the coleopterans and the terrestrial gastropods. On a large scale, a diet specializing in these increases the chances of surviving an ecological crisis compared with other species specialized in feeding on less abundant or more restricted prey. Nevertheless, on a local scale, any kind of specialized diet in a high-metabolic-rate shrew increases the possibility of dying due to starvation if unstable environmental conditions affect the populations of the preferred prey. In those cases, the capability to hoard food in good state of preservation allows the predator to overcome periods of scarcity. In extant species of shrews, the killing and the consuming of preys are completely separate actions, and different situations have been documented (Rychlik, 1999; Rychlik and Jancewicz, 2002). Of all the hunted preys, the immediately consumed usually represent a small percentage in most of the studied forms. Other victims are killed but left in the same place, where they can be consumed later, or be eaten by other conspeciﬁcs. A third option consists in carrying and hoarding the prey in nests or burrows exclusively for later consumption by the hunter. The last case is thought to have been of considerable interest to face changeable conditions. Nonetheless, the storage of dead bodies is limited, because although it reasonably diminishes the risk of starvation, the risk of infections from putrefaction in the nest or tunnel rapidly increases. It must be emphasized that the toxicity of the saliva is not the same in all the shrews (Jammot, 1983). Killing a prey is not a required goal for hoarding. Paralyzation is enough to ensure that the victim will remain in the nest for hours or even days, whilst the hunter is searching for more prey. An essential direct beneﬁt extracted is that the victim is thus preserved for a longer time in better conditions. In fact, some shrews have been documented to store animals alive but in comatose state (Jammot, 1983; Nowak, 1991; Dannelid, 1998, and related references therein). Using its ﬁssident upper incisors, Beremendia could have provided a soft bite and careful handling of its victims, thus avoiding unnecessary deaths before the real time of consumption. The delicate touch of these ‘pincers’ could have been also useful to grasp food and take it into the burrows. The Pliocene-Pleistocene Climatic Pulses and the Success of Beremendia and Homo The Late Pliocene to Early Pleistocene climatic pulses had deep effects in the evolution of the terrestrial fauna of that time interval, starting with the onset of the ﬁrst Arctic glaciation MIS 110 at 2.7 Ma and the glacial-interglacial dynamics in the Northern Hemisphere (Mudelsee and Raymo, 2005). After the 2.7 Ma event a major change is observed at 1.8 Ma, close to the PliocenePleistocene boundary, when strong glacial pulses are clearly visible in the oxygen isotope record of the oceans (Burckle, The idea can be reinforced by the consideration of the strong pigmentation of the teeth, not only in molars, but also in upper and lower antemolars, an adaptation to protect those teeth against the abrasive action of the exoskeletons of the most frequent ‘intractable’ animals (usually invertebrates such as coleopterans [Arthropoda] or gastropods [Mollusca]). The wear patterns of the teeth observed in the material from Dmanisi and Almenara-Casablanca 1 are in full agreement with such an hypothesis. Assuming a diet for Beremendia mainly based on those ‘intractable’ invertebrates, i.e., gastropods and beetles, the ﬁssidence of the upper incisor acquires new functional beneﬁts. The presence of medial tines increases the contact points of the ﬁrst upper incisors with the prey from barely four to a good six. The combined action of both upper incisors—left and right— by means of their corresponding spatulated apexes and talons provides a more reﬁned ability for the subjection, i.e., holding and handling, of the rounded hard-bodied exoskeletons of their victims. The pattern of wear found in the talon of the upper incisors and the anterior crests of the ﬁrst upper antemolars would even improve the subjection. Moreover, the consequent softer biting force (experimentally associated by Carraway and Verts, 1994, with the ﬁssidence in some American species of Sorex) provides the opportunity to control the force of the incisors to manage the prey by ﬁxing the hard and convex dorsal part of the exoskeletons in touch with the upper incisor without breaking it. That position leaves the most vulnerable parts of the victim in contact with the anterior part of the lower incisor, as shown in Figure 8, irrespective of whether it is a coleopteran or a gastropod. This way, the toxic substances can be easily injected through the weaker ventral protection of the prey, thus penetrating their open circulatory systems. A good canalization in the lower incisors as found in the species of Beremendia could be expected to be valuable for a good, precise injection of the venom. It is noteworthy that among all the Palearctic forms of Sorex dealt in Dannelid (1994), the one possessing the most ﬁssident upper incisors was S. roboratus Hollister, 1913, which had been previously reported to feed almost exclusively on coleopterans. This species was also the biggest, and the one with the most heavily pigmented teeth among all the species considered in there, thus indicating that these three characters are more than probably linked from the functional point of view. Not by chance, these three characters are overdone in Beremendia (see previous points). The capability of a big form like Beremendia to inject venom into non-struggling prey like gastropods or coleopterans raises another question: if such a reﬁned mechanism to hold the victims carefully was actually needed. Why not simply smash them? Recent discoveries on the behavior of some extant shrews can provide an explanation. It is a well-known fact that soricids are in general terms the group with the highest basal metabolic rate, i.e., consumption of oxygen per gram per hour, among all the terrestrial mammals. This trait is especially developed in the redtoothed shrews. Higher metabolic rates are advantageous to face colder environmental conditions, but as a counterbalance, they require a longer time of activity during the day to ensure the needed feeding supply. In some species of shrews, starvation periods lasting more than a few hours can lead to death. In order to avoid risky prolonged time of starvation, most soricines have restricted their distribution to rather warm and moist environments in which invertebrates easily proliferate (Reumer, 1984; Churchﬁeld, 1994). More widely distributed genera and species have developed the capability to store food. The hoarding behaviour has been reported in species of the extant genera Blarina by Robinson and Brodie (1982), in Neomys by Rychlik (1999), Rychlik and Jancewicz (2002), and Haberl (2002), and in Cryptotis Pomel, 1848, by Formanowicz et al. (1989). These shrews have been documented as storing different kinds of prey. This ´ FURIO ET AL.—PALEOBIOLOGY OF BEREMENDIA 1995). At the same time, the palynological analysis from Northern Europe indicates a new extension of cold steppe, which ended the warm-temperate conditions of the last Pliocene interglacial (Tiglian) and led to a new glacial phase known as the Eburonian. This cold event was global and has been recognized also in other parts of the world such as Eastern Africa, where a trend towards increasing aridity is reﬂected by a marked increase in grasses and other C4 plants (Cerling, 1992). Among terrestrial fauna, these climate pulses led to environmental changes and a number of dispersal events, best exempliﬁed by the ﬁrst settlement of Eurasia by early Homo (Vekua et al., 2002). The dispersal out of Africa of the early Homo Linnaeus, 1758, species and the paleoenvironmental causes that favored their development are of great interest in paleoanthropology. Due to its age and its fossil assemblage, Dmanisi has become an indispensable place for the elucidation of some of these hitherto open questions. This is because this locality has provided a large collection of fossil human remains, including ´ teeth (Martinon-Torres et al., 2008), mandibles (Gabunia et al., 2002), complete skulls (Rightmire et al., 2006), and postcranial bones (Lordkipanidze et al., 2007), but especially because it can be chronologically placed at the end of the Olduvai chron and the base of the upper Matuyama chron (Lordkipanidze et al., op. cit.), thus being the oldest occurrence of the genus Homo out of Africa. But further beyond than its impressive hominid representation, Dmanisi has delivered a rich vertebrate fauna composed of typical late Pliocene, ‘Villafranchian’ elements, such as Mammuthus meridionalis (Nesti, 1825), Stephanorhinus etruscus (Falconer, 1868), Equus stenonis Cocchi, 1867, Ursus etruscus Cuvier, 1823, Megantereon cultridens (Cuvier, 1824), and Pachycrocuta perrieri (Croizet and Jobert, 1828). Within the microvertebrate assemblage, the rodent fauna is mainly based on warm steppic elements, such as gerbils (Parameriones aff. obeidiensis, Haas, 1966; 40%) and migratory hamsters (Cricetulus n. sp., 28%; Lordkipanidze et al., 2007). The large mammal association is consistent with this environmental reconstruction, according to the presence of dry-adapted bovids such as diverse Antilopini, Gallogoral sickenbergii Kostopoulos, 1997, ` ` Soergelia cf. minor Moya-Sola, 1987, and Capra sp. However, the abundance of deer (Pseudodama cf. nestii [Major, 1879], Cervus abesalomi Kahlke, 2001, Eucladoceros aff. ctenoides [Nesti, 1841]) suggests the presence of some kind of woodland close to the rivers. The study of fossil fruits from the site conﬁrms a local open habitat dominated by xerophytic plants (Messager et al., 2008). In contrast to other coeval early Pleistocene sites, insectivores are very rare in Dmanisi. Within the scarce material found, only Beremendia ﬁssidens is clearly represented. Its presence in the site is not to be surprising, because the genus Beremendia, and especially its most frequent species B. ﬁssidens, reached its climax in terms of distribution and relative abundance during this time interval. From the beginnings of the Late Pliocene until the end of the Early Pleistocene, Beremendia spread over a wide geographic area, from southwestern Europe to eastern Asia (Jin and Kawamura, 1996). It is evident that the capabilities of Beremendia to store its prey provided extra beneﬁts under the climatic deterioration that affected most of the Central European Pliocene soricids during the interval between the RuscinianVillanyian boundary and the Early Pleistocene, in the shrews’ crisis noticed by Reumer (1985). It has been reported that in extant shrews, those capable of hoarding food, e.g., Blarina brevicauda, can minimize with great success the effects of lower temperatures, shorter days and winter food restrictions by simply modifying their daily activity pattern (Brandt and McCay, 2005, and references therein). But more than the short/medium-term climatic oscillations that affected Dmanisi in the past, the geological context of this local- 939 ity provides extra clues to understand the success of Beremendia. The fossil bones are notably found in a succession of volcanicash beds, which resulted from the accumulation of several eruptive episodes (Lordkipanidze et al., 2007). These volcanic events were determined to have been particularly destructive (Lumley et al., 2008), and thus the unpredictability of the environment would have generated regular local expirations. Furthermore, the upper part of the stratigraphy dominated by xerophilous vegetation of Boraginaceae is in line with this interpretation, given that these herbs are known to colonise open spaces (Messager et al., 2008). Apart from that, there are geological evidences of vertebrate burrowing and tunnelling in the site of Dmanisi. Animal burrows in the deposits of Dmanisi range in size from less than 1 cm of diameter, attributed to invertebrates—principally insects—to one burrow about 30 cm diameter, probably constructed by an animal the size of Hystrix Linnaeus, 1758, or perhaps a small canid. However, the most distinctive and most common burrows at Dmanisi are called ‘krotovina’—from the Russian ‘mole hole’—because their morphology and context make it highly certain that they were constructed by fossorial micromammals. Krotovina are not only restricted to Dmanisi, but they also occur in similar localities in the surrounding area (see Tappen et al., 2002b). The Dmanisi krotovina appear to have quite similar shape throughout the stratigraphic sequence, yet their relative density (n/m2 in proﬁle) and distinctiveness vary considerably among the different sedimentary-soil facies thus far documented. Overall, krotovina at Dmanisi have relatively low density. Low krotovina densities can be the result either of minimal micromammal burrowing activity, or prolonged bioturbation—including that of invertebrates and plant roots—that can homogenize sediments, rendering individual burrows less conspicuous. These narrow tunnels are found in all the layers, sometimes penetrating the immediately inferior ones, but never the upper ones. Two conclusions can be extracted from these observations. The ﬁrst one is that the tunnels of each layer were dug before the deposition of the immediately upper sediments/ashes. The second one is that the animals that carried out the construction of the tunnels did not burrow very deeply. Although the diameter of most krotovina ﬁts quite well with the expected body size of Beremendia, it is difﬁcult to establish a direct relationship between both. At any rate, it is a well-known fact that soricids tend to make use of tunnels made by other micromammals like moles or voles. In Dmanisi, Parameriones and Cricetulus are good candidates to have dug most krotovina, judging by the ethology of their extant relatives. So, independently of whether Beremendia was the digger of these tunnels or not, they were probably used by individuals as a refuge. A lot of terrestrial gastropod shells of the genera Helicopsis Fitzinger, 1833, Pseudochondrula Hesse, 1933, Helix Linnaeus, 1758, and Jaminia Risso, 1826 (Taktakishvili, pers. comm. in Gabunia et al., 2000), are not only found scattered in the burrows of Dmanisi, but even sometimes concentrated in small areas that could correspond to ‘paleolarders’ of Beremendia. Considering all these aspects, Beremendia ﬁssidens would have been one of the few species of insectivores capable of facing these conditions. Its capacity to hoard large amounts of comatose, immobile prey would have been of great help to survival after volcanic ash-rains like those that repeatedly affected the surrounds of Dmanisi in the past. The behavior deduced for Beremendia is not at all unlikely considering that other species present at the site, such as Homo and Pachycrocuta, displayed the capability to face unpredictable environments by means of hoarding more food than the intended for immediate consumption. Preliminary observations on the taphonomy of Dmanisi point to the involvement of both hominins and carnivores in the accumulation of the large-mammal fossil assemblage (Tappen et al., 2002a). 940 JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 30, NO. 3, 2010 CONCLUSIONS and the SEM photographs taken. We are indebted to C. R. Ferring (University of North Texas) for his useful and precise observations on the geology of the site. All the people of the Dmanisi Team are greatly acknowledged for their enthusiastic work and warm treatment, especially D. Lordkipanidze and A. Vekua ´ (Georgian National Museum), M. Martinon-Torres, E. Lacasa ´ ´ and P. Fernandez Colon (National Research Center on Human ´ Evolution), and J. M. Rodr´guez-Ponga (Fundacion Duques de ı Soria) for their ever-valuable help and support during the last 5 years. We do appreciate the helpful suggestions from Editor T. Martin and two anonymous reviewers to improve the original manuscript, and those of R. Wing for doing so on the English style. LITERATURE CITED Bachman, J. 1837. Some remarks on the genus Sorex, with a monograph of the North American species. Journal of the Academy of Natural Sciences of Philadelphia 7:362–402. Badyaev, A. V., and K. R. Foresman. 2004. Evolution of morphological integration. I. Functional units channel stress-induced variation in shrew mandibles. The American Naturalist 163:868–879. Brandt, A. J., and T. S. McCay. 2005. Temperature and photoperiod effects on activity of the northern short-tailed shrew (Blarina brevicauda). BIOS Journal 76:9–14. Burckle, L. H. 1995. Current issues in Pliocene paleoclimatology; pp. 3–7 in E. S. Vrba, G. H. Denton, T. C. Partridge, and L. H. Burckle (eds.), Paleoclimate and Evolution, with Emphasis on Human Origins. Yale University Press, New Haven, Connecticut, and London. Carraway, L. N., and B. J. Verts. 1994. Relationship of mandibular morphology to relative bite force in some Sorex from western North America; pp. 201–210 in J. F. Merritt, G. L. Kirkland Jr., and R. K. Rose (eds.), Advances in the Biology of Shrews. Special Publication of Carnegie Museum of Natural History 18, Pittsburgh. Carraway, L. N., B. J. Verts, M. L. Jones, and J. O. Whitaker. 1996. A search for age-related changes in bite force and diet in shrews. American Midland Naturalist 135:231–240. Cerling, T. E. 1992. Development of grasslands and savannas in East Africa during the Neogene. Global and Planetary change 5:241–247. Churchﬁeld, S. 1994. Foraging strategies of shrews, and the evidence from ﬁeld studies; pp. 77–88 in J. F. Merritt, G. L. Kirkland Jr., and R. K. Rose (eds.), Advances in the Biology of Shrews. Special Publication of Carnegie Museum of Natural History 18, Pittsburgh. Cocchi, L. 1867. L’uomo fossile nell’Italia Centrale. Memoire della So` cieta Italiana di Scienze Naturali 2:1–287. Croizet, J. B., and A. Jobert. 1828. Recherches sur les ossements fossiles ´ ˆ du departement du Puy-de-Dome. Clermont-Ferrand, Paris, 226 pp. ´ Cuenca-Bescos, G., and J. Rofes. 2007. First evidence of poisonous shrews with an envenomation apparatus. Naturwissenschaften 94:113–116. ` ´ Cuvier, G. 1823. Recherches sur les ossemens fossiles, ou l’on retablit les ` ´ caracteres de plusieurs animaux dont les revolutions du globe ont ´ ` ´ ` detruit les especes. Nouvelle edition entierement refondue, et con´ ´ ` siderablement augmentee. Iere Partie, G. Dufour et E. D’Ocagne Libraries, France, 405 pp. ´ Cuvier, G. 1824. Recherches sur les ossemens fossiles. Nouvelle edition. Paris, G. Dufour et E. D’Ocagne 5:1–547. Dannelid, E. 1994. Comparison of pigment and other dental characters of Eastern Paleartic Sorex (Mammalia: Soricidae); pp. 217–231 in J. F. Merritt, G. L. Kirkland Jr., and R. K. Rose (eds.), Advances in the Biology of Shrews. Special Publication of Carnegie Museum of Natural History 18, Pittsburgh. Dannelid, E. 1998. Dental Adaptations in Shrews; pp. 157–174 in J. M. Wojik and M. Wolsan (eds.), Evolution of Shrews. Mammal Research Institute, Polish Academy of Science, Bialowieza. ¨ Dotsch, C. 1985. Masticatory function in shrews (Soricidae). Acta Zoologica Fennica 173:231–235. ¨ Dotsch, C. 1994. Function of the feeding apparatus in red-toothed and white-toothed shrews (Soricidae) using electromyography and cineradiography; pp. 233–239 in J. F. Merritt, G. L. Kirkland Jr., and R. K. Rose (eds.), Advances in the Biology of Shrews. Special Publication of Carnegie Museum of Natural History 18, Pittsburgh. Evans, A. R., and G. D. Sanson. 1998. The effect of tooth shape on the breakdown of insects. Journal of Zoology 246:391–400. The species of the genus Beremendia display some peculiar characteristics in their dentitions and their mandibles. Most of these peculiarities can be interpreted as a mechanical adaptation for chewing and releasing the bite force in a way different from those of other shrews. For most of the dental and mandibular features analyzed, the mechanical solutions adopted by Beremendia do not differ substantially from those observed in the genus Blarina. In fact, a re-evaluation of the characters indicating that the Beremendiini are a separate group of the Blarinini is greatly encouraged. However, the extant species of this American genus of shrews are considered to have a more opportunistic way of feeding than Beremendia because of its more diverse abilities (see Dannelid, 1998). The general structure of the mandible in Beremendia argues in favor of a greater variety of permitted movements than in the rest of shrews. Because of that, its peculiarities are better explained as feeding adaptations to deal with ‘intractable’ food items. Therefore, it is deduced that Beremendia fed on coleopterans and gastropods. Under this deduction, its ﬁssident upper incisors ensured a good subjection for rounded exoskeletons, whereas the softer ventral sides of such prey remained exposed for the puncture by the lower ones. The effects that the toxic saliva exerted over the prey are difﬁcult to evaluate, but they deﬁnitively played an important role in the feeding strategy of Beremendia. Because its preferred prey was not struggling, deadly properties would have been absolutely unnecessary. A less aggressive nature of the saliva in Beremendia is consequently proposed as a wiser option. In recent species, the injection of venomous saliva in some victims induces a coma-shocking state instead of death, thus permitting their storage in good conditions for a longer time. This is valuable in reducing the uncertainty of provision against starvation, which in some shrews may be lethal even over periods of just a few hours. The enhancement of some morphological traits, particularly in the incisors, are clear improvements that provide a more reﬁned ability to induce a comatose state in some animals so as to store them alive, thus permitting Beremendia to survive under unstable ecological conditions. Biotopes like those inﬂuenced by cold winters, or those others dominated by volcanic episodes, were therefore places where Beremendia could survive when other soricids could not. This specialization of Beremendia differs from the previous interpretation of the utility of its venom delivery mechanism proposed ´ by Cuenca-Bescos and Rofes (2007). It provides a plausible explanation to the apparent opportunistic ecological proﬁle found by Reumer (1984), Rzebik-Kowalska (2000), and Popov (2003). The concluded trophic placement for Beremendia gives an explanation to its wide—and apparently randomized—geographic distribution in the European fossil record, because beetles and/or snails can be found almost everywhere. Its specialization allowed Beremendia to be one of the scarce Eurasian genera of shrews that survived to the great extinction of soricids that took place between the Late Pliocene and the Early Pleistocene. ACKNOWLEDGMENTS ´ This study has been supported by the Fundacion Duques de Soria, which covered the travel expenses of M. F., J. A., and A. M. between Georgia and Spain by means of the project ‘Las ´ dos Iberias. Crisis climaticas e intercambios faun´sticos en el ı ´ ´ Neogeno superior del Mediterraneo occidental y el Paratethys oriental,’ in collaboration with the Georgian National Museum. The expenses of some necessary extra travels of J. A. were covered by the projects CGL2006-4548 and CGL2009-07896 (Spanish Ministry of Education and Science) and 2009 SGR 324 (Generalitat de Catalunya). This paper has been also supported by the National Science Foundation through the project, ‘Revealing Hominid Origin Initiative’ (RHOI-Hominid-NSF-BCS0321893), which covered the expenses derived from the lab work ´ FURIO ET AL.—PALEOBIOLOGY OF BEREMENDIA Evans, A. R., and G. D. Sanson. 2005. Biomechanical properties of insects in relation to insectivory: cuticle thickness as an indicator of insect “hardness” and “intractability.” Australian Journal of Zoology 53:9–19. Falconer, H. 1868. On the European Pliocene and Post-Pliocene species of the genus Rhinoceros; pp. 309–403 in Charles Murchison (ed.), Palaeontological Memoirs and Notes of the Late Hugh Falconer, Volume 2: Mastodon, Elephant, Rhinoceros, Ossiferous Caves, Primeval Man and his Contemporaries. Robert Hardwicke, 192 Picadilly, London. Fanfani, F., and F. Masini. 1998. Insectivores (Mammalia) from the Rivoli Veronese, a Late Villanyian site in Northeastern Italy. Bolletino ` della Societa Paleontologica Italiana 36:367–380. Fitzinger, L. I. 1833. Systematisches Verzeichniß der im Erzherzogthume Oesterreich vorkommenden Weichthiere, als Prodrom einer Fauna ¨ derselben. Beitrage zur Landeskunde Oesterreich’s unter der Enns 3:88–122. ¨ Folinsbee, K. E., J. Muller, and R. R. Reisz. 2007. Canines grooves: morphology, function, and relevance to venom. Journal of Vertebrate Paleontology 27:547–551. Formanowicz, D. R. J., P. J. Bradley, and E. D. J. Brodie. 1989. Food hoarding by the least shrew (Cryptotis parva): intersexual and prey type effects. American Midland Naturalist 122:26–33. Fox, R. C., and C. S. Scott. 2005. First evidence of a venom delivery apparatus in extinct mammals. Nature 435: 1091–1093. ´ Furio, M., A. Santos-Cubedo, R. Minwer-Barakat, and J. Agust´. 2007. ı Evolutionary history of the African soricid Myosorex (Insectivora, Mammalia) out of Africa. Journal of Vertebrate Paleontology 27:1018–1032. Gabunia, L., A. Vekua, and D. Lordkipanidze. 2000. The environmental contexts of early human occupation of Georgia (Transcaucasia). Journal of Human Evolution 38:785–802. Gabunia, L., M.-A. De Lumley, A. Vekua, D. Lordkipanidze, and H. De ´ ´ ` Lumley. 2002. Decouverte d’un nouvel hominide a Dmanissi (Tran´ scaucasie, Georgie). Comptes Rendus Palevol 1:243–253. ´ Gasc, J. P. 1963. La musculature cephalique chez Suncus Ehr., Crocidura Wag., Sylvisorex Thom., Myosorex Gr. (Insectivores). Mammalia 27:582–601. Gazin, C. L. 1956. Palaeocene mammalian faunas of the Bison Basin in south-central Wyoming. Smithsonian Miscellaneous Collections 131:1–57. Grant, T. R., and P. D. Temple-Smith. 1998. Field biology of the platypus (Ornithorhynchus anatinus): historical and current perspectives. Philosophical Transactions of the Royal Society B 353:1081–1091. Gray, J. E. 1838. Revision of the genus Sorex, Linn. Proceedings of the Zoological Society of London 1837:123–126. Gureev, A. A. 1971. [Shrews (Soricidae) of the Fauna of the World]. Nauka, St. Petersburg, Russia, 254 pp. [Russian] Haberl, W. 2002. Food storage, prey remains and notes on occasional vertebrates in the diet of the Eurasian water shrew, Neomys fodiens. Folia Zoologica 51:93–102. Haas, G. 1966. On the vertebrate fauna of the Lower Pleistocene site Ubeidiya. Israel Academy of Sciences and Humanities, Jerusalem, 68 pp. Hermann, J. 1780. In W. E. A. Zimmermann, Geographische Geschichte ¨ des Menschen, und der vierfussigen Thiere, Vol. 2. Weygandsche Buchhandlung, Leipzig. Hesse, P. 1933. Zur Anatomie und Systematik der Familie Enidae. Archiv ¨ fur Naturgeschichte 2:145–224. Hollister, N. 1913. Two new mammals from the Siberian Altai. Smithsonian Miscellaneous Collections 60:1–3. Hurum, J. H., Z.-X. Luo, and Z. Kielan-Jaworowska. 2006. Were mammals originally venomous? Acta Palaeontologica Polonica 51:1–11. Hutterer, R. 1985. Anatomical adaptations of shrews. Mammal Review 15:43–55. Hutterer, R. 2005. Order Soricomorpha; pp. 220–311 in D. E. Wilson and D. M. Reeder (eds.), Mammal Species of the World. A Taxonomic and Geographic Reference, third edition, Volume 1. The Johns Hopkins University Press, Baltimore, Maryland. Jammot, D. 1977. Les Musaraignes (Soricidae, Insectivora) du PlioPleistocene d’Europe. Ph.D. dissertation, University of Dijon, Dijon, France, 341 pp. ´ Jammot, D. 1983. Evolution des Soricidae (Insectivora, Mammalia). Symbioses 15:253–273. 941 Jin, C.-Z., and Y. Kawamura. 1996. The ﬁrst reliable record of Beremendia (Insectivora, Mammalia) in East Asia and a revision of Peisorex Kowalski and Li, 1963. Transactions and Proceedings of the Palaeontological Society of Japan. New Series 182: 432–447. Kahlke, H.-D. 2001. Neufunde von Cerviden-Resten aus dem Un¨ terpleistozan von Untermassfeld; pp. 461–482 in R.-D. Kahlke ¨ ¨ (ed.), Das Pleistozan von Untermassfeld bei Meiningen (Thurin¨ gen). Romisch-Germanisches Zentralmuseum Monographien 40, Bonn. ¨ Kaup, J. J. 1829. Skizzirte Entwicklungs-Geschichte und Naturliches ¨ System der Europaischen Thierwelt. Darmstadt and Leipzig, 203 pp. Kita, M., Y. Nakamura, Y. Okumura, S. D. Ohdachi, Y. Oba, M. Yoshikuni, H. Kido, and D. Uemura. 2004. Blarina toxin, a mammalian lethal venom from the short-tailed shrew Blarina brevicauda: isolation and characterization. Proceedings of the National Academy of Sciences of the United States of America 101:7542–7547. ¨ Kormos, T. 1934. Neue Insektenfresser, Fledemause und Nager aus ¨ ´ ´ dem Oberpliozan der Villanyer Gegend. Foeldrajzi Koezlemenyek 64:296–321. Kostopoulos, D. S. 1997. The Plio-Pleistocene artiodactyls (Vertebrata, Mammalia) of Macedonia. 1. The fossiliferous site “Apollonia-1,” Mygdonia basin of Greece. Geodiversitas 19:845– 875. Kowalski, K., and C. K. Li. 1963. A new form of the Soricidae (Insectivora) from the Pleistocene of north China. Vertebrata PalAsiatica 7:138–143. Lieberman, D. E., and W. Crompton. 2000. Why fuse the mandibular symphysis? A comparative analysis. American Journal of Physical Anthropology 112:517–540. Linnaeus, C. 1758. Systema naturae per regna tria naturae, secundum classes, ordines, genera, species, cum characteribus, differentiis, synonymis, locis. Editio decima, reformata. Laurentii Salvii, Stockholm, 824 pp. ´ Lordkipanidze, D., T. Jashashvili, A. Vekua, M. S. Ponce de Leon, C. P. E. Zollikofer, G. P. Rightmire, H. Pontzer, R. Ferring, O. Oms, M. Tappen, M. Bukhsianidze, J. Agust´, R. Kahlke, G. Kiladze, ı B. Mart´nez-Navarro, A. Mouskhelishvili, M. Nioradze, and L. ı Rook. 2007. Postcranial evidence from early Homo from Dmanisi, Georgia. Nature 449:305–310. Lumley, M.-A. de, J.-M. Bardintzeff, P. Bienvenu, J.-B. Bilcot, G. Flamenbaum, C. Guy, M. Jullien, H. de Lumley, J.-P. Nabot, C. Perrenoud, O. Provitina, and M. Tourasse. 2008. Impact probable du ´ ` ´ volcanisme sur le deces des Hominides de Dmanisi. Comptes Rendus Palevol 7:61–79. Major, C. J. F. 1879. Cervi pliocenici del Val D’Arno Superiore. Atti ` della Societa Toscana di Scienze Naturali, Processi Verbali 1: 100. ´ ´ ´ Martinon-Torres, M., J. M. Bermudez de Castro, A. Gomez-Robles, A. Margvelashvili, L. Prado, D. Lordkipanidze, and A. Vekua. 2008. Dental remains from Dmanisi (Republic of Georgia): morphological analysis and comparative study. Journal of Human Evolution 55:249–273. Messager, E., D. Lordkipanidze, C. R. Ferring, and B. Deniaux. 2008. Fossil fruit identiﬁcation by SEM investigations, a tool for palaeoenvironmental reconstruction of Dmanisi site, Georgia. Journal of Archaeological Science 35:2715–2725. ` ` ´ Moya-Sola, S. 1987. Los bovidos (Artiodactyla, Mammalia) del yacimiento del Pleistoceno inferior de Venta Micena (Orce, ˜ ´ Granada, Espana). Paleontologia i Evolucio, Special Issue 1:181–236. Mudelsee, M., and M. E. Raymo. 2005. Slow dynamics of the Northern Hemisphere glaciation. Paleoceanography, Vol. 20, PA4022, doi:10.1029/2005PA001153. Nesti, F. 1825. Sulla nuova specie di elefante fossile del Valdarno all’ Illustrissimo Sig. Dott. Prof. Ottaviano Targioni Tozzetti (Lettere sopra alcune ossa fossili del Valdarno non per anco descritte). Nuovo Giornale di Lettere 11;195–216. Nesti, F. 1841. In P. Savi and A. Sismonda (reps.), Adunanza del d´ 25 Setı tembre 1841. Atti della terza riunione degli scienziati italiani tenuta in Firenze nel settembre del 1841:159. Nowak, R. M. (ed.) 1991. Walker’s Mammals of the World, 5th edition. Johns Hopkins University Press, Baltimore, Maryland, 642 pp. 942 JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 30, NO. 3, 2010 Rychlik, L., and E. Jancewicz. 2002. Prey size, prey nutrition, and food handling by shrews of different body sizes. Behavioral Ecology 13:216–223. Rzebik-Kowalska, B. 1976. The Neogene and Pleistocene insectivores (Mammalia) of Poland. III. Soricidae: Beremendia and Blarinoides. Acta Zoologica Cracoviensia 21:359–386. Rzebik-Kowalska, B. 1998. Fossil history of shrews in Europe; pp. 23–92 in J. M. Wojik and M. Wolsan (eds.), Evolution of Shrews. Mammal Research Institute, Polish Academy of Science, Bialowieza. Rzebik-Kowalska, B. 2000. Insectivora (Mammalia) from the Early and early Middle Pleistocene of Betﬁa in Romania. I. Soricidae Fischer von Waldheim, 1817. Acta zoologica cracoviensia 43:1–53. Say, T. 1823. Winter Cantonment near Council Bluff—councils with the Otoes, Missouries, Ioways, Pawnees, etc.; pp. 146–166 in E. James (comp.), Account of an Expedition from Pittsburgh to the Rocky Mountains, Performed in the years 1819 and ‘20, by order of the Hon. J. C. Calhoun, Secretary of War: Under the Command of Major Stephen H. Long. From the Notes of Major Long, Mr. T. Say, and Other Gentlemen of the Exploring Party. H. C. Carey and I. Lea, Philadelphia. Shaw, G. 1799. General Zoology or Systematic Natural History. Thomas Davison. White-Friars, London, 330 pp. Smith, W. L., and W. C. Wheeler. 2006. Venom evolution widespread in ﬁshes—a phylogenetic road map for the bioprospecting of piscine venoms. Journal of Heredity 97:206–217. Strait, S. G., and S. C. Smith. 2006. Elemental analysis of Soricine enamel: pigmentation variation and distribution in molars of Blarina brevicauda. Journal of Mammalogy 87:700–705. Tappen, M., C. R. Ferring, D. Lordkipanidze, A. Vekua, and G. Kiladze. 2002a. Preliminary observations on the vertebrate taphonomy of the Dmanisi locality in The Republic of Georgia; pp. 161–170 in M. de ´ Renzi, M. V. Pardo Alonso, and M. Belinchon (eds.), Current Top´ ics on Taphonomy and Fossilization. Colleccio “Encontres,” Ayuntamiento de Valencia, Valencia. Tappen, M., D. S. Adler, C. R. Ferring, M. Gabunia, A. Vekua, and C. C. Swisher. 2002b. Akhalkalaki: the taphonomy of an Early Pleistocene locality in the Republic of Georgia. Journal of Archaeological Science 29:1367–1391. Vekua, A., D. Lordkipanidze, G. P. Rightmire, J. Agust´, R. Ferring, G. ı ´ Maisuradze, A. Mouskhelishvili, M. Nioradze, M. Ponce de Leon, M. Tappen, M. Tvalchreridze, and C. Zollikofer. 2002. A new skull of early Homo from Dmanisi, Georgia. Science 297:85–89. ¨ ¨ Wagler, J. G. 1832. Mittheilungen uber einige merkwurdige Thiere. Isis von Oken 25:275–282. ¨ Ziegler, R. 1999. Order Insectivora; pp. 53–74 in G. E. Rossner and K. Heissig (eds.), The Miocene Land Mammals of Europe. Verlag Dr. ¨ Friedrich Pfeil, Munchen. Submitted March 25, 2009; accepted September 29, 2009. Orr, C. M., L. K. Delezene, J. E. Scott, and M. W. Tocheri. 2007. The comparative method and the inference of venom-delivery systems in fossil mammals. Journal of Vertebrate Paleontology 27:541–546. Pennant, T. 1771. Synopsis of Quadrupeds. J. Monk, Chester, U.K., 382 pp. ´ ´ ˝ ´ ´ ´ Petenyi, S. J. 1864. A beremendi meszkobanya termeszetrajz- es ˝ ´ ´ ´ ´ oslenytanilag le´rva. Hatrahagyott munkai, Magyar Tudomanyos ı ´ Akademia Kiadasa 1:35–81. ¨ ¨ Peters, W. C. H. 1864. Uber die Saugethiergattung Solenodon. Abhand¨ lungen der Koniglichen Akademie der Wissenshaften zu Berlin, 1863–1864:1–22. Pomel, N. A. 1848. Etudes sur les carnassiers insectivores. 1. Insectivores fossiles. 2. Classiﬁcations des insectivores. Archives des Sciences Physiques et Naturelles 9:159–165, 244–251. Popov, V. V. 2003. Late Pliocene Soricidae (Insectivora, Mammalia) from Varshets (North Bulgaria). Acta Zoologica Cracoviensia 46: 43–72. Preuschoft, H., and U. Witzel. 2002. Biomechanical investigations on the skulls of reptiles and mammals. Senckenbergiana Lethaea 82:207–222. Repenning, C. A. 1967. Subfamilies and genera of the Soricidae. Geological Survey Professional Paper 565:1–74. Reumer, J. W. F. 1984. Ruscinian and early Pleistocene Soricidae (Insectivora, Mammalia) from Tegelen (The Netherlands) and Hungary. Scripta Geologica 73:1–173. Reumer, J. W. F. 1985. The paleoecology of Soricidae (Insectivora, Mammalia) and its application to the debate on the Plio-Pleistocene ´ boundary. Revue de Paleobiologie 4:211–214. Reumer, J. W. F. 1998. A classiﬁcation of the fossil and recent shrews; pp. 5–22 in J. M. Wojik and M. Wolsan (eds.), Evolution of Shrews. Mammal Research Institute, Polish Academy of Science, Bialowieza. Rightmire, G. P., D. Lordkipanidze, and A. Vekua. 2006. Anatomical descriptions, comparative studies and evolutionary signiﬁcance of the hominin skulls from Dmanisi, Republic of Georgia. Journal of Human Evolution 50:115–141. Risso, A. 1826. Histoire naturelle des principales productions de l’Europe ´ meridionale, et principalement de celles des environs de Nice et des Alpes maritimes. Mollusques. Paris Levrault 4, 439 pp. Robinson, D. E., and E. D. Brodie, Jr. 1982. Food hoarding behavior in the short-tailed shrew Blarina brevicauda. American Midland Naturalist 108:369–375. ´ Rofes, J., and G. Cuenca-Bescos. 2009. First record of Beremendia ﬁssidens (Mammalia, Soricidae) in the Pleistocene of the Iberian Peninsula, with a review of the biostratigraphy, biogeography and palaeoecology of the species. Comptes Rendus Palevol 8:21–37. Rychlik, L. 1999. Changes in prey size preferences during successive stages of foraging in the Mediterranean water shrew Neomys anomalus. Behaviour 136:345–365.

The Paleobiology of the Extinct Venomous Shrew Beremendia (Soricidae, Insectivora, Mammalia) in Relation to the Geology and Paleoenvironment of Dmanisi (Early Pleistocene, Georgia) more

Log In

Reset Password