Supplemental material for Lister et al., 2 November 2001 issue Stratigraphy and dating of deposits yielding mammoth remains

AFRICA M. subplanifrons Measurements of the earliest Mammuthus are taken from Maglio (1973) and Kalb and Mebrate (1993). Maglio’s sample is a composite, from a variety of early Pliocene localities in southern and eastern Africa, including Virginia and the Vaal river terraces, S. Africa; the Kaiso Formation, Uganda; the Vogel River Series, Tanzania; and the Chemeron Formation, Kenya. Kalb and Mebrate’s material was collected from the Awash Group, Ethiopia, specifically from the Aramis Member of the Sagantole Formation and the Kuseralee Member of the Adu-Asa Formation. The fossil-bearing horizons alternate with a series of tuffs which are correlated lithostratigraphically and dated radiometrically (Kalb and Mebrate 1993). M. subplanifrons from the Aramis member is between the VT-1 and CT tuffs, dated to ”0DDQG0D respectively. From the Kuseralee Member the fossils are from below the LASS tuff dated to ” Ma.

EUROPE Red Crag, Montopoli and Cernatesti The earliest European mammoth sample in this study is a composite of four teeth from the Red Crag, England, three from Montopoli, Italy, and one from Cernatesti, Romania. The Red Crag Formation comprises a series of inshore marine shelly sands. As indicated by Gibbard et al. (1998) and Head (1998a&b), the Red Crag covers the Waltonian, Pre-Ludhamian and Ludhamian Stages of the British Late Pliocene. However, the majority of surface exposures, except those around Walton, are of Pre-Ludhamian age (Zalasiewicz, pers. comm.), so it is likely that the bulk of the collected mammalian fauna is of this age. Head (1998a & b) and Funnell (1998), on the basis of dinoflagellates and foraminifera, respectively, concur in placing this part of the Red Crag in the interval 2.6 - 2.4 Ma. Head (1998b) correlates the Pre-Ludhamian with either the cold Praetiglian Stage of The Netherlands (c. 2.5 Ma), or the preceding Reuverian C Substage (c. 2.6 Ma), with the latter slightly favoured. In the Italian sequence, the earliest occurrence of elephantid fossils is in the Montopoli Faunal Unit, of Early Villafranchian age. The Montopoli deposits are littoral sands (Benvenuti et al., 1995) immediately above a palaeomagnetic reversal regarded as the Gauss/Matuyama, and are therefore dated to c. 2.6-2.5 Ma (Lindsay et al. 1980, Sardella et al. 1998). The Romanian fossil is from the Daçiç Basin. In their revision of Romanian Villafranchian faunas, Feru et al (1983) and Radulesco and Samson (1995) confirm the presence of very primitive mammoth at two localities: Tuluçesti and Cernatesti. A molar from the latter locality is

included in the present study, using measurements from Radulesco and Samson (1995) and kindly provided by Dr C. Radulesco (pers. comm.1999). Associated mammalian fauna includes Mammut borsoni, Anancus arvernensis, Stephanorhinus elatus and S. etruscus, Cervus pardinensis, C. perrieri, Plesippus sp. and Paracamelus sp. The authors correlate the Cernatesti and Tuluçesti faunas with Montopoli, and regard them as early Villafranchian in age, mammal zone MN16a (c. 3.0-2.9 Ma). The Red Crag, Montopoli and Cernatesti sample (referred to as ‘Red Crag +’ in the Figures) is here given a median age of 2.6 Ma. Khapry The sample comes from the Khapry Sands - an extensive fluvial body of near-mouth sediments of the palaeo-Don River (Azov Sea near Rostov-on-Don, Russia). Since 1903, fossils have been collected in a series of sand quarries, such as Khapry, Morskaya, Liventsovka and others, and this collection was chosen as the type for the Khapry Faunal Complex, of Late Pliocene age (Gromov, 1948). The elephant material from these sites served as the basis for the primitive mammoth named “Archidiskodon” gromovi (Alexeeva & Garutt, 1965), a concept supported by some scientists (Baigusheva, 1971; Azzaroli, 1977) and rejected by others (Dubrovo, 1989), and now most commonly treated as a primitive subspecies of M. meridionalis (see Lister, 1996a for review). The Khapry Sands have reversed palaeomagnetic polarity, and on the basis of molluscan fauna are correlated with the middle Akchagylian of Caspian stratigraphy. This places them in the earlier part of the Matuyama Chron just above the Gauss-Matuyama reversal at 2.5 Ma, and before the Akchygalian/Apsheronian boundary at c. 2.0 Ma. This is supported by the mammalian fauna, by which the Khapry Complex is correlated with the Middle Villafranchian (St. Vallier) and is believed to correspond to Zone MN 17 of European mammal biostratigraphy (c. 2.5-1.5 Ma). The most recent revision of the Khapry Complex (including previously unpublished material) and its stratigraphic position confirms that this fauna represents a single chronological unit (Titov, 1999). Upper Valdarno The type material of M . meridionalis is from the Upper Valdarno, Italy. The Upper Valarno Basin, SE of Florence, is filled with 550m of fluvio-lacustrine deposits. The sedimentary infilling of the basin is divided into three main stages: Middle Pliocene, Early Pleistocene and Middle to Late Pleistocene (Abbate, 1983). The second sequence, comprising the Montevarchi lacustrine and fan-delta deposits, has yielded the bulk of the classic ‘Upper Valdarno’ fauna. Azzaroli (1977) and Azzaroli et al. (1988) show that while the precise horizon of many of the older finds is uncertain, almost all came from the ‘Montevarchi Group’, covering the Olivola and Tasso Faunal Units. This has been confirmed in recent years by more systematic collecting, with the addition of a possibly new Faunal Unit between them, represented at Matassino and Poggio Rosso. The large mammal faunas from all of these Units are clearly younger than the ‘St Vallier’ stage of Seneze and Chilhac, France, both below the base of the positive Olduvai palaeomagnetic chron (c. 1.95 Ma) (Azzaroli et al., 1998). New palaeomagnetic work on the Montevarchi sequence (Napoleone et al., in press) places all of the classic ‘Upper Valdarno’ faunas in positively-magnetised deposits in the later part of the Olduvai chron (i.e. c. 1.8 Ma). This is consistent with rodent biostratigraphy (Sardella et al., 1998). Palynology reveals an alternation of steppic and forest environments through the Montevarchi sequence, presumed to correspond to at

least one glacial-interglacial cycle (Bertini, 1994), but the precise relation of these to the mammalian faunas is unclear. Pietrafitta Commercial excavations in a lignite mine at Pietrafitta, west of Perugia in Umbria, Italy, have yielded a rich mammalian fossil assemblage. The lignite fills a basin formed during Middle Pliocene times (Ambrosetti et al., 1989). Plant remains, including Pterocarya and Tsuga, suggest an Early Pleistocene age for the deposit (Ricciardi, 1961). The mammalian assemblage, in addition to Mammuthus meridionalis (Ferretti, 1999), includes Mimomys pusillus, Microtus (Allophaiomys) chalinei, M. (A.) cf. ruffoi, Leptobos aff. vallisarni, Stephanorhinus cf. hundsheimensis, Pseudodama ‘farnetensis’ , and an early megacerine deer related to Megaloceros verticornis (Abbazzi, 1995; Ficcarelli et al., 1996; Gentili et al., 1996). This combination of species suggests correlation to Early Biharian sites of Eastern Europe, and in Italian stratigraphy indicates a ‘Middle Villafranchian’ age, more precisely correlated to the Farneta Faunal Unit of Italian stratigraphy (Gentili et al., 1996; Sardella et al., 1998). This Unit is midway in stratigraphic sequence between the top of the Olduvai palaeomagnetic event at c. 1.7 Ma (Olivola F.U.), and the Jaramillo event at c. 1.0 Ma (Colle Curti F.U.), and is regarded as c. 1.4 Ma in age (Sardella et al., 1998). Untermassfeld The fossil site of Untermassfeld was discovered in 1978 on the right slope of the River Werra, southern Thuringia (central Germany). The Pleistocene sequence begins with 7-8m of coarse gravels which, according to the terrace sequence of the Werra Valley, were deposited during the Eburonian (Early Pleistocene) (Ellenberg & Kahlke, 1997). These are superceded by 21m of floodplain deposits, comprising the Lower and Upper Fluviatile Sands. The latter are preserved as the infill of a channel cutting into the former. The basal 3m of the Upper Sands forms the main fossiliferous horizon of the site (Kahlke, 2000). Wiegank (1997) showed that the Lower Fluviatile Sands incorporate a magnetic reversal, from reversed to normal. This normal polarity persists in the find horizon. The rich mammalian assemblage includes several taxa indicative of the late ‘Villfranchian’, such as Mimomys pusillus and Alces carnutorum (Kahlke, 1997), so that it seems very likely that the magnetic reversal is the base of the Jaramillo event rather than the Brunhes/Matuyam boundary. This indicates an age of approximately 1.0 Ma for the fauna (Shackleton, 1995). Oriolo The ‘Sabbie gialle’ Formation of the north-eastern Apennines (Italy) stratigraphically overlays the Early Pleistocene ‘Argille Azzurre’ and has been divided into two distinct sedimentary cycles by Marabini et al (1987). The first represents a regressive phase, the second, unconformably overlying the first, a minor transgression characterised by shoreline depositional structures. The quarry ‘La Salita di Oriolo’ at Faenza, Emilia Romagna, exposes the upper part of the latter unit, which makes up the bulk of the ‘Sabbie gialle’. The mammalian fossil assemblage from this unit has been described by Masini et al. (1995), and indicates an age in the late Early Pleistocene, corresponding approximately to the Colle Curti faunal unit of Italian biostratigraphy. A palaeomagnetic study of a 20m outcrop at the Oriolo quarry indicates a transition from reversed to normal polarity (Torre et al. 1996). Part of the fauna pertains to the lower (reversed) interval, while the remainder, inclusing the elephant remains described by Ferretti (1999) were retrieved

from levels above the reversal. Vai (1984) and Marabini et al. (1987) proposed a ‘Milazzian’ (Middle Pleistocene) age for the upper sequence of the ‘Sabbie gialle’, which woud suggest that the palaeomagnetic reversal was te Brunhes/Matuyama boundary at c. 780 Ka. However, this is at variance with the mammalian fauna which suggests that it represents the base of the Jaramillo event, giving the elephant remains an age of c. 1.0 Ma (Ferretti 1999). Saint-Prest This site, in Eure-et-Loire, northern France, has been known as a rich fossiliferous locality since the middle of the 19th century (Laugel, 1862). Dewolf & Lautridou (1973) and Dewolf et al. (in press) show, on the basis of recent exposures, that the fossiliferous site is located in a karstic depression in the upper part of a slope of the Eure valley. The depression is filled by a sequence of periglacial deposits, within which is a layer of clayey sands which are the source of the mammalian fossils. This layer is at the top of the periglacial alluvial sands and gravels of the river Eure. The whole is capped by Saalian and Weichselian loesses. The mammalian assemblage, which indicates a temperate, forested episode, includes on the one hand Early Pleistocene taxa such as Equus stenonis, on the other hand species of the late Early Pleistocene and early Middle Pleistocene, such as Bison schoetensacki (Guerin et al., n.d.; Dewolf et al. in press). A cervid antler previously identified as Megaloceros verticornis can be reidentified as Eucladoceros cf. giulii (Lister, pers. obs.), a species recently discovered at Untermassfeld (Kahlke, 1997; see above). Taken together, this assemblage indicates an age in mammalian biozone zone MNQ 20, in the region 1.2 - 0.8 Ma (Guérin, 1982; Sardella et al., 1998). This is corroborated by negative palaeomagnetism of the St.-Prest deposits (Dewolf et al., in press), indicating the Matuyama Chron and a minimum age of 0.78 Ma, which however excludes the Jaramillo positive subcron (c. 0.9-1.0 Ma). An extensive sample of mammoth molars from the site (Coppens & Beden 1972) has been restudied in the present investigation, and plotted at an age of 1 Ma. Taman’ The Siniaya Balka locality in the Taman’ peninsula, Azov Sea, is the type locality of the Tamanian faunal complex. Hundreds of bones, including a large series of elephant molars, have been collected here since 1912 in a stone/bone breccia of mud flow origin, filling an ancient gully. The gully is cut into strongly tilted clay and sand sediments of Late Pliocene age (Upper Kuyalnik or Akchagylian), which on the basis of small mammals and reversed magnetic polarity are referred to the lower part of the Matuyama Chron, shortly before 1.96 Ma (Pevzner et al., 1998). Since the gully fill has no upper contact with any sediment, closer limits to its age are provided mainly from biostratigraphic considerations in comparison with other, dated localities of the local Tamanian Complex and the broader European succession. No rodents are known from the Siniaya Balka breccia, but the presence of Bison s.l. and more advanced horses than Equus stenonis makes it younger than most Villafranchian faunas but not older than the latest faunas of that stage (1.6-1.2 Ma). On the other hand, the bison and horses from Siniaya Balka are clearly less advanced than those from the subsequent faunal complex, the Tiraspolian, broadly correlated with the Cromerian (c. 800-500 Ka). Sediments capping the area of Siniaya Balka (though not the gully itself) are correlated to the Bakinian marine terrace, regarded on the basis of marine fauna as broadly early Middle Pleistocene (c. 800-500 Ka) in age; the Bakinian sediments have normal polarity, and their continental equivalents include mammal faunas of Tiraspolian type. Since the earliest faunas of the Tiraspolian complex appear at the very top of

the Matuyama Chron, immediately before the Matuyama/Brunhes transition at 780 Ka, and some Tamanian s.l. faunas also lie in the upper part of Matuyama, the upper limit of the Tamanian complex is designated at about 0.8-0.9 Ma (Vangengeim et al., 1991). The position of Siniaya Balka within the Tamanian time-range (1.1-0.8 Ma) cannot be precisely determined, so we provisionally assign it an average age of c. 1.0 Ma. West Runton and Beeston The West Runton Freshwater Bed (WRFB), the stratotype of the Cromerian interglacial, is a 1.52 m thick deposit of sand and silt exposed at the base of the cliff at West Runton, Norfolk, UK. Pollen from the bed was interpreted by West (1980) as representing the first half of the interglacial - substages Ia to IIb. The deposit, part of the Cromer Forest-bed Formation, is rich in organic detritus, macroscopic plant remains, beetles, molluscs and vertebrates. Over 40 species of mammal have been recorded (Stuart, 1996). Foreshore deposits in the vicinity also yield mammalian remains but these are of much older, Pre-Pastonian to Pastonian (Early Pleistocene) age (Lister, 1996b, 1998). Fossils from West Runton have been collected for over 150 years. Those from the WRFB can be identified with relative ease: they are often labelled as having been collected in situ or from the base of the cliff, and most have a characteristic dark brown shiny preservation, although some are redder and appear to have come from oxidised, sandy facies which can also be observed today. Early Pleistocene fossils, of heavier preservation and usually indicated as having been found on the foreshore, are excluded from the sample used in this study. Excavations in 1992 and 1995 resulted in the recovery of an almost complete skeleton of Mammuthus trogontherii from the WRFB (Ashwin & Stuart 1996). This specimen is incorporated in the sample. At Beeston, in Norfolk, less than two km west of West Runton, a complete lower jaw of M. trogontherii was discovered in the cliff in 1978, in situ within gravels of the Beestonian stage. The jaw lay at the bottom of West’s (1980) Bed o, overlying temperate Pastonian sediments, and forming part of the type Beestonian with a pollen spectrum indicating a cold environment and largely treeless, herbaceous vegetation. The bed was succeeded by a series of further cold climate deposits, capped eventually by bed u, incorporating a temperate, tree-dominated pollen assemblage which West (1980) tentatively correlated with the upper part of the interglacial sequence (‘Cromerian IVa’) of West Runton. Recent work has greatly refined our understanding of early Middle Pleistocene stratigraphy in Britain (Turner, 1996; Preece & Parfitt, 2000; Stuart & Lister, 2000 and in press), with five or six distinct temperate episodes suggested for the interval between the Brunhes/Matuyama inversion (c. 780 Ka) and the Anglian/Elsterian glaciation (c. 450 Ka). Both molluscan and mammalian evidence concur in placing the West Runton Freshwater Bed as the earliest of these episodes. First, the presence of the vole Mimomys savini places the WRFB in the lower half of this interval, compared to later sites with Arvicola terrestris. Second, the presence of the mollusc Valvata goldfussiana, extinct mid-way through the Mimomys biozone, limits the WRFB, with Pakefield/Kessingland (Suffolk), to the earlier two of three Mimomys episodes identified. Finally, the absence from the large WRFB mammal collections of Palaeoloxodon, Hippopotamus and Megaloceros dawkinsi, all present at Pakefield/Kessingland, suggests that the WRFB is the earlier of the two (Preece & Parfitt, 2000; Stuart & Lister, 2000 and in press). The

WRFB is normally magnetised (West, 1980), so as the earliest of 5 or 6 known temperate episodes in the interval 780-450 Ka, the West Runton Freshwater Bed is here assigned an approximate age of 700 Ka. The age of the Beeston jaw is more difficult to determine. The Beestonian was defined as the interval between the Pastonian and Cromerian interglacials (West, 1980), but since that time it has been realised that the Pastonian is much older than originally thought, and that several temperate episodes occur within the ‘Cromerian Complex’. In the absence of significant biostratigraphic information from the Beestonian, it cannot be ascertained with certainty whether these cold-stage deposits pre- or post-date the type Cromerian of West Runton. The specimen is provisionally incorporated with the West Runton sample, with which it is morphologically conformable. Voigtstedt The main mammalian assemblage from Voigtstedt, Thüringia, Germany (including all the remains studied here) was collected from a single limnic horizon, the Lehmschicht, sandwiched between fluviatile sands (Kahlke, 1965; Steinmüller, 1972), the whole overlain by Elsterian glacial deposits. Pollen and vertebrate evidence indicate an early Middle Pleistocene interglacial. On mammalian evidence, this has been correlated with the type Cromerian of West Runton, UK (Stuart, 1981), and more recent consideration of the mammals (including the presence of Mimomys savini, and absence of taxa such as Hippopotamus and Palaeoloxodon) places the deposit in the first half of the ‘Cromerian Complex’ (v. Koenigswald & Heinrich, 1999; Stuart & Lister, 2000 and in press). Since the ‘Cromerian Complex’ extends from MIS 19 - 13, a date in the region of MIS 19-17 seems likely. Approximately 2 km north of Voigtstedt, at Edersleben, a Mammuthus skeleton was excavated (Garutt & Nikolskaya, 1988). Steinmüller (1972) shows that it came from the lower sands, the unit underlying the Lehmschicht at Voigtstedt, and pertaining to the Helme cold stage immediately preceding the Voigtstedt interglacial. This specimen is incorporated with the main Voigtstedt sample in the present study. It was formerly referred to Mammuthus trogontherii on the basis of foot morphology, but this feature is of debatable value (Lister 1996a), and the specimen is dentally indistinguishable from the advanced M. meridionalis of the main Voigtsedt assemblage. Süssenborn The Süssenborn deposits are an extensive sequence of fluvial sands in Thüringia, Germany. On the basis of lithostratigraphy and mammalian fauna, a date within a cool phase of the early Middle Pleistocene ‘Cromerian Complex’ is indicated for Süssenborn (Steinmuller, 1972; Maul, 1990). The presence of Mimomys savini places the assemblage in the first half of the ‘Cromerian Complex’ (v. Koenigswald & Kolfschoten, 1996; v. Koenigswald & Heinrich, 1999). It has been debated whether the Süssenborn deposits are younger or older than those of Voigtstedt (Steinmuller, 1972; Maul, 1990), but von Koenigswald and Heinrich (1999) suggest a later date because the assemblage includes the first entry of several cold-adapted species into Europe.

Mosbach A thick sequence of fluviatile sands of the Rhine at Mosbach has yielded a succession of mammalian assemblages. Lithostratigraphy and mammalian fauna indicate a date within the Cromerian Complex, probably a cool phase of an interglacial, for the main assemblage (Mosbach II) (Brüning, 1978; Igel, 1985; Koenigswald & Tobien, 1987). On the basis of the mammals, such as the presence of Arvicola, Palaeoloxodon and Hippopotamus, an interglacial in the second half of the Cromerian Complex’ (in the region MIS 15-13) is indicated (v. Koenigswald & Heinrich, 1999; Stuart & Lister, in press). Since Voigtstedt, Süssenborn and Mosbach II are all within the interval of the ‘Cromerian Complex’, i.e. c. 780-450 Ka, and biostratigraphy indicates that they are chronologically in the order indicated above, estimated dates of 700 Ka, 600 Ka and 500 Ka are allocated, respectively, to the three samples in this study. Steinheim The gravel deposits at Steinheim are up to 22m thick and have yielded a rich mammalian assemblage from several pits over the course of the past century. At the Sammet pit, the lower levels, with an interglacial fauna including Palaeoloxodon antiquus and the Steinheim hominid, have been referred to the Holsteinian Interglacial (Adam et al., 1995). Above this is a great thickness of cold-stage gravels with a mammalian assemblage rich in mammoth remains. These have been regarded as ‘early Rissian’ in age, and are surmounted by a further interglacial horizon (Adam et al. 1995). A few of the mammoths in the sample came from other pits, particularly the Sigrist pit, which has mammoth-bearing gravels both above and below the P. antiquus horizon. The Palaeoloxodon and Mammuthus horizons here have also been referred to the Holsteinian and early Rissian (Adam et al., 1995), though Ziegler (pers. comm.) suggests that they might relate to the next interglacial/glacial cycle. However, Schreve & Bridgland (in press) find that the mammalian fauna of the Palaeoloxodon horizons at Steinheim as a whole resembles that of Swanscombe, referred to MIS 11, rather than that of the lower terrace deposits at Grays, which are referred to MIS 9. They concur that the overlying Mammuthus gravels relate to the succeeding cold stage. On the assumption that the Holsteinian is the first interglacial after the Elsterian and is equivalent to MIS 11 (c. 400 Ka), an approximate age of 350 Ka (MIS 10) is used for the mammoth sample in the present study. Von Koenigswald & Heinrich (1999) caution that an unequivocal palynological correlation of Steinheim with the Holsteinian is not available. If the Holsteinian and/or one or both of the Steinheim interglacial deposits represent MIS 9, then a date of c. 250Ka would be more likely for at least part of the mammoth assemblage. This does not affect its position in the sampling sequence of this paper. Ariendorf A series of Late Middle Pleistocene deposits is represented at the Schneider Quarry near Koblenz, Germany (Turner, 1997). At the base, gravels of the Leubsdorf (Middle) terrace of the Rhine are dated by the Ar40/Ar39 method, on an intercalated tephra, to c. 490 Ka BP. Channelled into this, the Ariendorf Interglacial is represented by a series of tephra and palaeosol deposits dated by Ar40/Ar39 to between 410 - 419 Ka BP. These are succeeded by a series of loess deposits within which four faunal and archaeological levels have been investigated; from the bottom up: the channel, and Ariendorf 1, 2 and 3. A series of TL dates on loess place the channel and Ariendorf 1 in the region 300-230 Ka, Ariendorf 2 in the region 140-200 Ka, and loess

above Ariendorf 3 in the region 160-30 Ka. An Ar40/Ar39 date on tephra immediately below the Ariendorf 3 horizon gave an age of 215 Ka BP (Turner, 1997). The mammoth remains examined in this study are from the channel and Ariendorf 2. They therefore span the interval c. 300-150 Ka, but appear morphologically homogeneous. Tourville At Tourville-la-Rivière, NW France, a 35m sequence of deposits in a low terrace of the river Seine is exposed (Carpentier & Lautridou, 1986). The key horizons are two interglacial estuarine silty beds, B below and D above, and an intervening unit of periglacial sands and gravels (Bed C). Mammalian assemblages have been obtained from all three layers; the mammoth sample studied here is from Bed C, and was found in association with other ‘cold’ elements such as Rangifer and Coelodonta. Shells from Bed D have been dated by ESR to c. 200 Ka (Stremme, 1985), and feldspar grains to 196 +/- 23 Ka by thermoluminescence (Balescu et al., 1997). Bed B yielded a TL date of 314 Ka. Beds B, C & D are therefore referred to MIS 9, 8 & 7, respectively. An age in the region 250 +/- 40 Ka is therefore very likely for the mammoth sample from Bed C. Ilford Interglacial ‘brickearth’ (silt and clay) deposits of the river Thames have been exposed in several pits at Ilford, Essex, UK, the most important being Uphall Pit. These deposits are formed within the Taplow/Mucking Formation of Thames stratigraphy. According to the model of Sutcliffe & Kowalski (1976) and Bridgland (1994), this is the third post-Anglian formation and is correlated to MIS 7, c. 200 Ka BP. Biostratigraphy, such as the presence of mammoth, horse, and the mollusc Corbicula fluminalis, and absence of hippopotamus, strongly supports the allocation of this deposit, and correlates such as Aveley, to a pre-Ipswichian interglacial (Sutcliffe & Kowalski, 1976; Meijer & Preece, 2000; Schreve, in press). The Mucking Terrace is immediately above the Ipswichian (Eemian MIS 5e) Trafalgar Square terrace, supporting the allocation to MIS 7 (Bridgland, 1994). Stanton Harcourt At Stanton Harcourt, 7km west of Oxford (UK), Pleistocene river gravels overlie Oxford Clay bedrock. The main body of gravel forms the Stanton Harcourt Gravel Member of the Summertown-Radley Terrace Formation, and was deposited in cold, periglacial conditions. On the basis of Thames terrace stratigraphy, Bridgland (1994) attributes this period to MIS 6. In a restricted area, the gravels overlie the Stanton Harcourt channel deposits, within which large numbers of mammoth and other fossils have recently been excavated (Buckingham et al., 1996). Fauna and flora clearly indicate a temperate, wooded environment. Both its stratigraphic position and fossil complement, particularly in comparison with the better-known lower Thames sequence, strongly suggest an age in the pre-Ipswichian interglacial (cf. MIS 7). This evidence includes the presence of the beetle Anotylus gibbulus, the mollusc Corbicula fluminalis, and the abundance of mammoth living under temperate conditions. An MIS 7 age (c. 200 Ka) is supported by amino-acid racemisation data (Bowen et al., 1979). Marsworth A sequence of deposits exposed at College Farm, Marsworth, UK, comprises two interglacial channels separated by periglacial ‘coombe rock’ (consolidated chalk and flint debris with frost polygons) (Green et al., 1984; Murton et al., in press). The upper channel contains a mammalian

fauna including hippopotamus, and is regarded as Ipswichian (MIS 5e) in age. The lower channel comprises three layers: a fossiliferous gravelly sand at the base (Layer 3), organic muds (Layer 2), and chalk muds with waterlain sand and fine gravel (Layer 1). Mammalian remains were recovered in situ from all three layers, but the large majority of the fossils (including all the mammoth remains studied in the present paper) were from Layer 2 (Murton et al., in press; D. Parrish, pers. comm.). The mammalian, molluscan and beetle assemblages from Layer 2 suggest temperate but predominantely grassland habitats. Uranium series dating of travertine blocks reworked into the Lower Channel gave a range of age-estimates between 140Ka and 170 Ka (Green et al., 1984). Schreve (in press), on the basis of the mammalian and coleopteran assemblage, correlates the Lower Channel with MIS7 (c. 200 Ka). A dating in late MIS 7 is favoured, although an interstadial early in MIS 6 cannot be ruled out (Murton et al., in press). Therefore, the mammoth sample is plotted at a median age of 175 Ka in this study. Brundon Jordan’s Pit, near Brundon, Suffolk (UK) exposed gravels and ‘brickearths’ of the River Stour. Mammalian and molluscan remains indicate a temperate but rather open, grassy environment. Uranium-series dating of bone gave age estimates of 230 Ka and 174 Ka (Szabo & Collins, 1975). On the basis of the mammalian and molluscan assemblage, Schreve (in press) correlates the fauna to the upper part of the sequence at Aveley, and therefore to the later part of MIS 7. Ehringsdorf Weimar-Ehringsdorf is one of several localities exposing a considerable depth of travertine within terrace gravels of the River Ilm in Thüringia, Germany. At Ehringsdorf, the Lower Travertine yields rich floral and faunal remains as well as evidence of Middle Paleolithic human activity. This is superceded by a loess-rich ‘Pariser’ horizon, above which is the Upper Travertine, also richly fossiliferous. The sequence was originally attributed to the Eemian (Kahlke, 1974), but more recent work on small mammals has inclined to a pre-Eemian interglacial (von Koenigswald & Heinrich, 1999). Uranium-series analysis of the Lower Travertine gave an age-estimate of 225 +/- 28 Ka (Blackwell & Schwarcz, 1985). Schreve and Bridgland (in press) show how the large mammal assemblage includes several markers (such as Stephanorhinus kirchbergensis and Homo sp.) which in comparison with the British sequence indicate an MIS 7 rather than MIS 5e (Ipswichian/Eemian) age. They further suggest that the Lower and Upper travertines represent successive parts of the MIS 7 interglacial (c. 200 kyr). Mammoth material from the site comes mainly from the Upper Travertine. Balderton Four suites of sands and gravels were mapped in the area of the River Trent, Lincolnshire (UK) by Brandon & Sumbler (1991). These are interpreted as terrace structures, and are in order of descending height (and by inference, decreasing age), the Eagle Moor, Balderton, Fulbeck, and Floodplain terraces. The flint-rich Eagle Moor Sand and Gravel is interpreted as outwash from the Anglian glaciation, while the Fulbeck Sand and Gravel has yielded a mammalian fauna including hippopotamus, which is interpreted as Ipswichian in age. The Balderton Sand and Gravel, intermediate in position and therefore age, is typically 7-8 m deep and includes icewedge casts. Above, the Whisby Sand includes traces of rubification indicating a palaeosol, and the whole has been cryoturbated by subsequent cold conditions. This corroborates the position of the Balderton Sand and Gravel as pre-dating the last glacial-interglacial cycle. ESR dates on

mammoth teeth from the deposit (Grün, in Lister & Brandon, 1991) give ages in the range 130190 Ka. Subsequently, an interglacial deposit has been exposed at the base of the Balderton Sand and Gravel, containing biostratigraphic indicators of MIS 7 age (Brandon, pers. comm.). A median age of 160 Ka is used for the mammoth sample in this study. Tattershall Thorpe At Tatterhall Thorpe, Lincolnshire (UK), a depression in glacial till (the Wragby Till) contains interglacial clays and detritus muds. This is succeeded by several metres of sands and gravels of the River Witham, with silt lenses and ice-wedge casts. A large quantity of mammalian bones (including mammoth molars) was recovered from the sands and gravels. Holyoak and Preece (1980) studied the interglacial biota and indicated that they might correlate to the Ipswichian or to the preceding interglacial. Radiocarbon dates on plant matter from silt horizons in the gravels gave ages in the range 28 Ka - 46 Ka BP, which might indicate a Middle Devensian range but are close to the limit of radiocarbon dating. Subsequent evidence points to an older age. The Tattershall Thorpe deposits are higher than those at nearby Tattershall Castle, for which there is good evidence of an Ipswichian-Devensian sequence. The mammalian assemblage from Tattershall Thorpe is very different from that of Tattershall Castle (Rackham, 1984), but very similar to that of the Balderton Sands and Gravels (see above). The mollusc Corbicula fluminalis has been found in the Tattershall Thorpe interglacial (Holyoak and Preece 1980); this is unknown in the Ipswichian but is characteristic of the ‘MIS 7’ interglacial (Meijer & Preece, 2000). Finally, ESR datings on three mammoth teeth from the gravels have given very similar dates, all in the range c. 130 Ka BP (Grün, pers. comm.). Current geological surveying thus interprets the Tatterhall Thorpe deposits as a higher terrace of the Witham than that of Tattershall Castle (Brandon, pers. comm.). La Cotte The site of La Cotte de St Brelade, on the southern side of Jersey (UK Channel Islands), has yielded rich mammalian faunas in association with Palaeolithic industries (Scott, 1986). The most important faunal layers are broadly Saalian in age. The mammoth remains examined here come from (oldest first) Layers C, B, A, 3 and 6. The underlying Layer D is dated by TL to 238 +/-35 Ka BP (Ox-TL 222); the sequence is capped by a raised beach deposit referred to the Eemian (Callow & Cornford 1986). Since the faunal layers were deposited under a cold climate, an age in the range c. 190-130 Ka is indicated for the mammoth remains. Zemst The sample comes from fluvial and aeolian sands at Zemst-Bos van Aa, in the eastern branch of the Flemish Valley, Belgium. Two Formations have been recognised: the Gent Formation (largely aeolian) in the lower part of the sequence, and the Zemst Formation (fluvial in origin) above. The mammoth sample studied here comes from the blue-grey sands of the lower part of the Bos van Aa Member of the Zemst Formation. The mammal assemblage from this horizon is known as Zemst IIb (Germonpré et al., 1993). It directly overlies the Grimbergen Member, which on the basis of plant and molluscan remains, especially the presence of the mollusc Corbicula fluminalis, was correlated with the Eemian (Bogemans 1993). Resting on the Bos van Aa Member are a fine clastic topstratum and point bar deposits correlated on sedimentological and palynological grounds with the Hombeek Member, known from boreholes in the area and regarded as representing the Amersfoort and Brorup interstadials. The Bos van Aa Member has

therefore been regarded as early Weichselian in age. However, recent recognition of additional Late Middle Pleistocene interglacials throws into question the attribution of the Grimbergen deposits to the Eemian. Meijer and Preece (2000) show that in Britain and The Netherlands, the mollusc Corbicula fluminalis does not occur in the Eemian (MIS 5e, c. 120 Ka), but is highly characteristic of the preceding interglacial, correlated to MIS 7 (c. 200 Ka). ESR dating of a mammoth molar from the Zemst IIb fauna produced dates of 126, 200 +/-9,300 BP (early Uuptake model) or 131,900 +/- 7,800 BP (late U-uptake model) (Grün, in Germonpré et al. 1993). Although this might, within the range of error, correspond to the earliest Weichselian, it could also correspond to the late Saalian, MIS 6, corresponding to a ‘Stage 7’ age for the Grimbergen interglacial. Taken together with the biostratigraphic information, an MIS 6 age for the mammoth sample is considered likely. 3 HGPRVWLDQG/HD9DOOH\*UDYHOV

A celebrated early Upper Palaeolithic site in the Czech Republic, representative of the Eastern Gravettian complex (Soffer 1993). Recent AMS radiocarbon dates on charred bone (van der Plicht et al. 1997) have yielded the following results: 26870 +/- 250 BP (GrN-6801) and 26320 +/- 240 BP (GrN-6852). Low terrace gravels of the River Lea, a tributary of the Thames NE of London, have yielded a cold-adapted mammalian fauna including M. primigenius, Coelodonta antiquitatis, Rangifer tarandus and Dicrostonyx torquatus. Fossils have been collected from several localities, especially in the vicinity of Broxbourne, Edmonton and Nazeing. Peat rafts within the gravels, known as the ‘Lea Valley Arctic Bed’, contain a ‘steppe tundra’ flora, and radiocarbon dates indicate an age of c. 28 Ka (Allison et al. 1952; Stuart, 1982). In view of the near identity of these two European sites in terms of both age and mammoth morphology, they are plotted together in this paper, as ‘Predmosti +’.

SIBERIA The Early Olyorian The type region of the Olyorian Land Mammal Age is in the Yana-Kolyma Lowland of northeastern Yakutia. In this region, many river bluffs expose mammal-rich deposits of Olyorian age. The combined sample of Early Olyorian mammoths comes from numerous localities where the lower member of the Olyor Suite, referred to the Chukochyan Horizon of the regional stratigraphic chart (Lower Pleistocene), is exposed. Fine-grained sands, silts and organic layers of the Olyor Suite with numerous ice-wedge casts have been studied in most detail in the type area of the Olyorian, the Bolshaya Chukochya River (Sher, 1971; Virina et al., 1984) and the key section at the Krestovka River (Loc. 6, lower course of the Kolyma) (Sher et al., 1979). Palaeomagnetic studies at selected sites of the Olyor Suite consistently indicate that its lower part (exposed up to 20 m height in some sections and traced by boreholes to a depth of 30 m below the river level in others) is of reversed magnetic polarity, while its upper part (as well as all the overlying younger formations) are of normal polarity. A positive excursion is found some depth below the reversal. This sequence is interpreted as the late Matuyama and early Brunhes Chrons, the former including the Jaramillo Subchron. The Chukochyan (Early Olyorian) lies completely

within the Matuyama zone, with its upper boundary almost reaching the Matuyama-Brunhes inversion (c. 780 Ka BP). Important Early Olyorian mammoth fossils come from Locs. NN 23, 25, 27, where the exposed part of the Chukochyan runs down well below the suggested Jaramillo Subchron, so the lower limit of this sample is estimated as at least 1200 Ka. In addition to mammoth, Early Olyorian fauna includes Predicrostonyx compitalis, Lemmus sheri, Allophaiomys reservatus (Zazhigin, 1997,1998), Equus verae, Praeovibos beringiensis, Soergelia sp. and others (Sher et al., 1987). This broadly Early Pleistocene assemblage strongly supports the interpretation of the palaeomagnetic sequence: none of these taxa was in existence in the Gauss reversed Chron, and several (e.g. Allophaiomys) were extinct by the early Middle Pleistocene. Other Early Olyorian mammoth fossils come from similar sites in the Alazeya and Bolshoy Khomus-Yuryakh River valleys, where the large thicknesses of exposed Chukochyan have again been proven with both paleomagnetism and mammals (Sher, 1981; Grinenko and Minyuk, 1985; Virina, 1997). Finally, a few teeth in this sample were found at the Adycha River (Verkhoyansk region), where Early Olyorian sediments (Adychan Suite) lie mostly below the common river level, but many mammalian fossils of that age are delivered to the river bank by erosion (Sher et al, 1987). The age of the Early Olyorian assemblage can therefore be defined as the interval c. 1200 - 800 Ka. Most specimens in the Early Olyorian sample with more certain stratigraphic position are related to the middle part of this interval - either closely above the Jaramillo Subchron (where it is recognized) or immediately below it. The sample as a whole is therefore plotted at a median age of 1 Ma. The Late Olyorian The upper member of the Olyor Suite, referred to the Akanian Horizon, is known in various parts of north-eastern Siberia, but is best studied in the same areas as the Early Olyorian (see above). Its lower boundary is defined by the transition in the lineage of collared lemmings from the Early Olyorian Predicrostonyx compitalis to Dicrostonyx renidens, and by the replacement of Allophaiomys reservatus by true Microtus spp. (Sher, 1984; Sher et al., 1987; Zazhigin, 1976, 1997). These important faunal changes occur in the sections immediately below the Matuyama/Brunhes inversion (c. 780 Ka), so the main part of the Akanian fauna is from sediments of normal polarity, and occupies the early part in the Brunhes Chron. A key mammoth finding is from Loc.35 at the Bolshaya Chukochya R., where a partial skeleton of mammoth with four molars was found in situ with Late Olyorian rodents just above the Matuyama/Brunhes inversion (inferred from a borehole and near-by bluff) (Sher, 1971; Virina, 1997). Many additional teeth of similar morphology and characteristic Olyorian preservation were collected at the same and neighbouring localities (Locs. 34 and 36), where mostly Akanian and later sediments are exposed. This includes an associated upper dp4 and M1 found in situ in the upper part of the Akanian. Other specimens in the sample were collected at sites of the Alazeya R. with similar stratigraphy (Sher, 1981; Grinenko and Minyuk, 1985). Since the top of the Akanian is everywhere marked with an erosional unconformity, the upper time-limit of the Late Olyorian sample cannot be defined precisely. The earliest known postOlyorian site, Loc. 83 at the Bol. Khomus-Yuryakh R., is likely to be as old as the Late Cromerian Complex in Europe. This correlation is based on the occurrence of an early form of Dicrostonyx simplicior, a collared lemming with a broad North Eurasian range (V. Zazhigin, pers. comm.). This species probably evolved in Northern Siberia from the Late Olyorian D. renidens, prior to its appearance in Europe in the late Cromerian Complex c. 500 kyr, thus giving

an indirect indication of the upper age limit for the Late Olyorian. The likely time range for the Late Olyorian can be roughly estimated as 800-600 Ka, and we take 700 Ka as the median age of the Late Olyyorian mammoth sample. Indeed, most specimens included in this sample occurred either immediately above the Matuyama/Brunhes inversion, or seem to come from the lowermost part of the Akanian (the main upper part of this member at most sites having been eroded before younger sediments started to deposit). The Late Middle Pleistocene This compound sample comes from various sedimentary bodies in the Yana-Kolyma Lowland, generally referred to the Russian “Middle Pleistocene” (the Late Middle Pleistocene of European stratigraphy, MIS 11-6, c. 400 - 125 Ka). Their common feature is that they always underlie the Late Pleistocene “Ice Complex” (see below), have normal polarity, and in places overlie the Olyor Suite sediments. Many of them have yielded D. simplicior - the guide fossil for this time span in northeastern Siberia Having a variety of local stratigraphic names, they were joined into the Keremesitian Horizon (Sher et al., 1987) after the type formation at the Keremesit River (Lower Indigirka). Some of these units are lithologically similar to the Olyorian and contain icewedge casts, but no (or very rare) ice wedges. However, they have no typical Olyorian mammals; most remarkable is the replacement of the huge and archaic Olyorian horses by more advanced large caballine forms of Equus scotti type. Although the latter could have appeared in the region at the end of the Akanian (Eisenmann, 1992), the full replacement (extinction of E. verae group) had taken place by the Keremesitian. Another feature of this stage is the dominance of bison among the artiodactyls, not observed in the Olyorian which is characterised instead by various extinct musk-oxen. Sites of this type, where the mammoth fossils came from, include the Khroma Suite (Kaplina et al., 1983) and Allaikha Suite (Kaplina et al., 1980), named after the rivers of their type locations on the Yana-Indigirka Lowland; and the Utkin Beds on the Maliy Anyuy River (the Lower Kolyma basin) (Sher, 1971). The lower part of the Khomus Suite (Loc. 83, see above) can be referred to the same group (Virina, 1997). Another group of sedimentary units referred to the Keremesitian are built mostly by fluvial sands and include numerous though not very large ice wedges along with ice wedge casts. Among them is the Keremesit Suite itself (Bashlavin et al., 1987), the Maastakh Suite on the Bol. Chukochya River (Sher, 1971, 1984) and its equivalents in the Krestovka R. sections (Sher et al., 1979), the thick sequence of sands on top of the Akanian on the Adycha R., and sediments underlying the Ice Complex in the Khroma R. estuary (Khaptashinskiy Yar). The chronological limits for this mammoth sample can be approximately estimated as between 600 Ka (cf comments on Loc. 83 in the Late Olyorian section) and 200-150 Ka. We therefore plot the sample at a median age of 400 Ka. The distribution of the specimens over the sites, however, together with the geological information from these localities, suggests that the majority of specimens in the sample come from the earlier rather than the later part of this chronological interval. The Late Pleistocene (“Ice Complex” sample) The upper member of the Pleistocene succession in Northeast Siberia is very remarkable. The most famous sedimentary bodies with tons of mammoth ivory, millions of bones, and complete frozen mammoth carcasses belong to this unique formation, where huge ice wedges plus

structure-forming ice make up to 80-85% of its volume (the rest is silt or fine sand). They normally build the high Pleistocene plain, or terrace, or the hills remaining from its dissection, called ‘Yedoma‘. The most common current name for this climatically driven, syncryogenic formation of polygenetic origin is ‘Ice Complex', and in this sense, it has no particular chronological or stratigraphic implications. Since most of it, however, was deposited during the Late Pleistocene, many local stratigraphic names have been suggested for it, and the most official is the Yedoma Suite (Sher, 1971) or the Yedoma Superhorizon (see Sher et al., 1987). The upper part of the Ice Complex is accessible to radiocarbon dating. Its upper limit corresponds approximately to the terminal Pleistocene at about 12 Ka. Its lower limit is more difficult to ascertain. At two sites at least, a Late Middle Pleistocene age is suggested by the presence of advanced Dicrostonyx simplicior (which is thought to be replaced by D. gulielmi in the Late Pleistocene) (Kaplina et al., 1980; Sher, 1997a). Mammalian fossils from the Ice Complex usually have perfect preservation with soft tissues, typical for unthawed permafrost. Our latest mammoth sample, named ‘Ice Complex’, includes some specimens that have finite radiocarbon ages, or came from sites radiocarbon dated to the Late Pleistocene, such as the famous Duvannyy Yar in the lower course of the Kolyma R. In addition, many specimens, lacking strict chronological control, but demonstrating fresh preservation typical of the ‘Ice Complex’ but not of underlying units, have been included. While the majority of the material is undoubtedly of Late Pleistocene age, the Ice Complex extends into the latest Middle Pleistocene, so some of the specimens may be of this age. Critically for the present study, however, all of it is definitely younger than any material in the ‘Late Middle Pleistocene’ sample since the latter comes from sediments underlying the main Ice Complex. To minimize confusion, we call our sample ‘Ice Complex’, its chronological framework falling within the range of MIS 2-6, i.e. between c. 150200 and 12 Ka. We have excluded ’terminal’ mammoth populations in the region 15-10 Ka, such as the famous Berelyokh sample (12.5 Ka) since they demonstrate a sharp decrease in body size, believed to be related to the extinction, with consequences for dental morphometrics (Lister & Joysey 1992). Considering the Ice Complex sample as a whole, we note that a large collection of radiocarbon dates on mammoth from northeast Siberia (Sulerzhitzky, 1995), including specimens incorporated in the present study, indicates that the majority fall within the age range 50-30 Ka. Stratigraphic information available for most sites contributing to our sample, confirms that only a few of them can be referred to a younger (30-12 Ka) or older (>50 Ka) age. For that reason, we take 40 Ka as a median age for the Ice Complex sample.

REFERENCES Abbate, E. 1983. Schema stratigrafico della successione neoautoctona del Valdarno Superiore e del bacino do Arezzo. In: Abbate, E. & Sagri, M. (eds.) I Depositi Fluvio-lacustri del Valdarno Superiore, 1-6 (Florence: Ist. Geol. Paleont. Univ. Firenze). Abbazzi, L. 1995. Megaceroides obscurus from the Val di Chiana (Cava Liberatori, Tuscany, Central Italy, Farneta f.u., Early Pleistocene). remarks on the early evolution and systematics of Megaceroides. Bollettino della Società Geologica Italiana 34, 223-234.

Adam, K.-D., Bloos, G. & Ziegler, R. 1995. Stop 11: Steinheim/Murr, N of Stuttgart - locality of Homo steinheimensis. In: Schirmer, W. (ed.) Quaternary field trips in Central Europe, 13, 727728. (Munich: Verlag Pfeil) Alekseeva, L.I. & Garutt, V.E. 1965. New data on the evolution of elephants of the genus Archidiskodon. Byull. Kom. Izuch. Chetv. Per., 30, 161-166 (In Russian) Allison, J., Godwin, H. & Warren, S.H. 1952. Late-glacial deposits at Nazeing in the Lea Valley, North London. Phil. Trans. R. Soc. Lond. B 236, 169-230. Ambrosetti, P., Cattuto, C. & Gregori, L. 1989. Geomorfologia e neotettonica nel bacino di Tavernelle/Pietrafitta (Umbria). Il Quaternario 2, 57-64. Ashwin, T. & Stuart, A.J. 1996. The West Runton elephant. Current Archaeology 149, 164-168. Azzaroli, A. 1977. Evolutionary patterns of Villafranchian elephants in central Italy. Atti Accad. Naz. Lincei, Mem., Cl. Sc. Fiz., mat. nat. Ser. 8, 14, Section 2a, 149-168. Azzaroli, A., de Giuli, C., Ficcarelli, G. & Torre, D. 1988. Late Pliocene to early mid-Pleistocene mammals in Eurasia: faunal succession and dispersal events. Palaeogeog., Palaeoclimatol., Palaeoecol. 66, 77-100. Baigusheva, V.S. 1971. Fossil theriofauna of the Liventsovka quarry (north-east Azov sea region). In: Matepialy po faunam antpopogena SSSR. Trudy Zoologicheskogo Instituta AN SSSR, 49, 5-29 (In Russian) Balescu, S., Lamothe, M., Lautridou, J.P. 1997. Luminescence evidence for two Middle Pleistocene interglacial events at Tourville, northwestern France. Boreas 26, (1) 61-72 Bashlavin, D.K., Zhigulyovtseva, S.N., Ovander, M.G. 1987. Keremesit River section as a key section for the Middle Pleistocene of the Yana-Indigirka Lowland. In: V.P.Pokhialaynen (Ed.), Chetvertichniy period v severo-vostochnoy Azii. Magadan, SV KNII DVO AN SSSR, 110-118. Benvenuti, M., Dominici, S. & Rook, L. 1995. Inquadramento stratigrafico-deposizionale delle faune a mammiferi villafranchiane (Unita Faunistiche Triversa e Montopoli) del Valdarno Inferiore nella zona a sud dell’Arno (Toscana). Il Quaternario 8, 457-464. Bertini, A. 1994. Palynological investigations on Upper Neogene and Lower Pleistocene sections in central and northern Italy. Mem. Soc. Geol. It. 48: 431-443. Blackwell, B. & Schwarcz, H.P. 1985. U-series analysis of the Lower Travertine at Ehringsdorf, DDR. Quaternary Research 25, 215-222. Bogemans, F. 1993. Quaternary geological mapping on the basis of sedimentary properties in the eastern branch of the Flemish Valley (sheets Boom-Mechelen & Vilvoorde-Zemst). Mem. Cartes. geol. min. Belg. 35, 1-49.

Bowen, D.Q., Hughes, S., Sykes, G.A. & Miller, G.H. 1989. Land-sea correlations in the Pleistocene, based on isoleucine epimerisation in non-marine molluscs. Nature340, 49-51. Brandon, A. & Sumbler, M.G. 1991. The Balderton Sand and Gravel: pre-Ipswichian cold stage fluvial deposits near Lincoln, England. J. Quat. Sci. 6, 117-138. Bridgland, D. 1994. Quaternary of the Thames. Geological Conservation Review Series 7. Chapman & Hall, London, 441 pp. Brüning, H. 1978. Zur Untergliederung der Mosbacher Terassenabfolge und zum klimatischen Stellwert der Mosbacher Tierwelt im Rahmen des Cromer-Komplexes. Mainzer Wissenschaftl. Arch. 16, 143-190. Buckingham, C.M., Roe, D.A. & Scott, K. 1996. A preliminary report on the Stanton Harcourt Channel Deposits (Oxfordshire, England): geological context, vertebrate remains and palaeolithic stone artefacts. J. Quat. Sc. 11, 397-415. Callow, P. & Cornford, J.M. (eds.) 1986. La Cotte de St Brelade 1961-1978: Excavations of CBM McBurney. Geo Books, Norwich. Carpentier, G. & Lautridou, J.-P. 1986. The low terrace of the Seine: the alluvium, periglacial deposits, interglacial fluviomarine deposits, slope deposits and palaeosols, fauna. In: The loess and other Pleistocene periglacial deposits of N.W. Europe, including their relationships with marine formations and features, J.P. Lautridou, D. Keen & J.L. Mornier, (Eds.), Caen, 64-68. Coppens, Y. & Beden, M. 1982. Mammuthus meridionalis depereti n. subsp. (Mammalia, Elephantidae), nouveau mammouth du Pléistocène inférieur d’Europe occidentale. C R. Acad. Sc. Paris Ser. 2, 294, 291-294. Dewolf, Y. & Lautridou, J.-P. 1973. Révision de la coupe de Saint-Prest (Eure-et-Loire) (feuille de Chartres au 1/50000). C. R. somm. Soc. Géol. Fr. 1973, 2, 55-57. Dewolf, Y., Guérin, C. & Lautridou, J.-P. in press. Révision d’un site célèbre - Saint-Prest (Chartres, France). Quaternaire. Dubrovo, I.A. 1989. Systematic position of the Khapry elephant. Paleontologicheskiy zhurnal 1, 78-87 (In Russian). Eisenmann, V. 1992. Origin, dispersals, and migrations of Equus (Mammalia, Perissodactyla). Courier Forsch.-Inst. Senckenberg 153, 161-170. Ellenberg, J. & Kahlke, R.-D. 1997. Die quartärgeologische Entwicklung des mittleren Werratals und der Bau der unterpleistozänen Komplexfundstelle Untermassfeld. In (Kahlke, R.-D., ed.) Das Pleistozan von Untermassfeld bei Meiningen (Thuringen), Teil 1. Monogr. Röm-Germ. Zentralmus. Mainz 40, 29-62.

Ferretti, M. 1999. Mammuthus meridionalis (Mammalia, Probocidea, Elephantidae) from the ‘Sabbie Gialle’ of Oriolo (Cava La Salita, Faenza, Northern Italy) and other European late populations of southern mammoth. Eclogae geol. Helv. 92, 503-515. Feru, M.U., Radulesco, C. & Samson, P. 1983. Succession des mammifères plio-pléistocènes dans le Basin dacique (Roumanie). Ann. Inst. Geol. Geophys. 59, 161-167. Ficcarelli, G., Masini, F. & Torre, D. 1996. The mammals of the latest Villafranchian in Italy. In (C. Turner, Ed.) The Early Middle Pleistocene in Europe, 263-272. Rotterdam: Balkema. Forsten, A. 1999. A review of Equus stenonis Cocchi (Perissodactyla, Equidae) and related forms. Quaternary Science Reviews 18, 1373-1408. Funnell, B.M. 1998. Climatic evolution of the North Sea and North Atlantic during the PliocenePleistocene (2.7-1.7 Ma). Meded. Nederlands Inst. Toegepaste Geowet. TNO 60, 227-238. Garutt, V.E. & Nikolskaya, V.N. 1988. Uber das Skelett vom Steppenelefant aus Edersleben. Beitr. Heimatf. 9, 3-13. Gentili, S. Abbazzi, L., Masini, F., Ambrosetti, P., Argenti, P. & Torre, D. 1996. Voles from the Early Pleistocene of Pietrafitta (central Italy, Perugia). Acta zool. Cracov. 39, 185-199. Germonpre, M., Bogemans, F., van Neer, W. & Grün, R. 1993. The dating of two Pleistocene mammal assemblages from the Flemish Valley, Belgium. Contr. Tert. Quatern. Geol. 30, 147153. Gibbard, P.L., Zalasiewicz, J.A. & Mathers, S.J. 1998. Stratigraphy of the marine PlioPleistocene crag deposits of East Anglia. Meded. Nederlands Inst. Toegepaste Geowet. TNO 60, 239-262. Green, C.P., Coope. G.R., Currant, A.P., Holyoak, D.T., Ivanovich, M., Jones, R.L., Keen, D.H., McGregor, D.F.M. & Robinson, J.E. 1984. Evidence of two temperate episodes in Late Pleistocene deposits at Marsworth, UK. Nature 309, 778-781. Grinenko, O.V., Minyuk, P.S. 1985. Paleomagnetism and stratification of the Upper Cenozoic deposits of the Alazeya River basin (Northern Yakutia). In: V.I.Goncharov, Ed., Paleomagnetism i problemy stratigraficheskoy korrelyatsii. Magadan, SV KNII DVNTs, 124-132. Gromov, V.I. 1948. Paleontological and archeological background for the stratigraphy of the continental Quaternary deposits in the USSR (mammals, Paleolithic). Trudy Instituta geologicheskikh nauk AN SSSR, seriya geologicheskaya, 64, 1-521 (In Russian). Guérin, C. 1982. Première biozonation du Pléistocène européen, principal résultat biostratigraphique de l’étude des Rhinocerotidae (Mammalian, Perissodactyla) du Miocène terminal au Pléistocene supérieur d’Europe occidentale. Géobios 15, 593-598.

Guérin, C., Dewolf, Y. & Lautridou, J.-P. n.d. Le site paléontologique de Saint-Prest. Chartres: Muséum des Sciences Naturelles et de Préhistoire. Head, M.J. 1998a. Marine environmental change in the Pliocene and early Pleistocene of eastern England: the dinoflagellate evidence reviewed. Meded. Nederlands Inst. Toegepaste Geowet. TNO 60, 199-225. Holyoak, D.T. & Preece, R.C. 1980. Late Pleistocene interglacial deposits at Tattershall, Lincolnshire. Phil. Trans. R. Soc. B 311, 193-236. Igel, W. 1985. Paläogeographie durch sediment-Morphoskopie. Grundlagen, Methodik u. Anwendung. Unpubl. PhD thesis, Univ. Mainz. Kahlke, H.-D. 1965. Die stratigraphische Stellung der Faunen von Voigtstedt. Paläont. Abh. A, II, 2/3, 691-692. Kahlke, H.-D. (Ed.) 1974. Das Pleistozän von Weimar-Ehringsdorf. Teil I. Paläontologische Abhandlungen 21, 1-351. Kahlke, R.-D. (Ed.) 1997. Das Pleistozan von Untermassfeld bei Meiningen (Thuringen), Teil 1. Monogr. Röm-Germ. Zentralmus. Mainz. Kahlke, R.-D. 2000. The Early Pleistocene (Epivillafranchian) faunal site of Untermassfeld (Thuringia, Central Germany) - synthesis of new results. In (Lordkipanidze, D., Bar-Yosef, O. & Otte, M. Eds.) Early Humans at the Gates of Europe. Etudes et Recherches Archéologiques de l’Université de Liège 92, 128-138. Kalb, J.E. & Mebrate, A. 1993. Fossil elephantoids from the hominid-bearing Awash Group, Middle Awash Valley, Afar Depression, Ethiopia. Trans. Am. Phil. Soc. 83, 1-114. Kaplina, T.N., Sher, A.V., Giterman, R.E., Zazhigin, V.S., Kiselyov, S.V., Lozhkin, A.V. & Nikitin, V.P. 1980. Key section of Pleistocene deposits on the Allaikha river (lower reaches of the Indigirka). Byull. Kom. Izuch. Chetv. Per., 50, 73-95 Kaplina, T.N., Ovander, M.G., Lozhkin, A.V., Zhigulyovtseva, S.N., Rybakova, N.O., Nikitin, V.P., Kiselyov, S.V., Bakhtin, A.V., Pirumova, L.G. & Virina, E.I. 1983. Quaternary deposits in the middle current of the Khroma River (Yana-Indigirka Lowland). In: Shilo N.A., Ed., Stratigrafiya i paleogeografiya pozdnego kaynozoya Vostoka SSSR, Magadan, SV KNII DVNTs AN SSSR, 80-95. Koenigswald, W. v. & Heinrich, W.-D. 1999. Mittelpleistozäne Säugetierfaunen aus Mitteleuropa - der Versuch einer biostratigraphischen Zuordnung. Kaupia 9, 53-112.

Koenigswald, W.v. & Kolfschoten, T. v. 1996. The Mimomys-Arvicola boundary and the enamel thickness quotient (SDQ) of Arvicola as stratigraphic markers in the Middle Pleistocene. In: The Early Middle Pleistocene of Europe (C. Turner, Ed.). Rotterdam: Balkema, 211-226. Koenigswald, W. v. & Tobien, H. 1987. Bemerkungen zur Altersstellung der pleistozänen Mosbach-Sande bei Wiesbaden. Geol. Jb. Hessen 115, 227-237. Kolfschoten, T. van. 1990. The Early Biharian mammal faunas from Bave and Dorst-Surae. Quartärpaläontologie 8, 265-272. Laugel, A. 1862. La faune de Saint-Prest, près de Chartres (Eure-et-Loire). Bull. Soc. Géol. Fr. (2), 19, 709-718. Lindsay, E.H., Opdyke, N.D. & Johnson, N.M. 1980. Pliocene dispersal of the horse Equus and late Cenozoic mammal dispersal events. Nature 287, 135-8. Lister, A.M. 1996a. Evolution and taxonomy of Eurasian mammoths. In: Shoshani, J. & Tassy, P. (Eds.). The Proboscidea. Oxford: OUP, 203-213. Lister, A.M. 1996b. The stratigraphical interpretation of large mammal remains from the Cromer Forest-bed Formation. In The Early Middle Pleistocene in Europe(ed. C. Turner), 25-44. Balkema, Rotterdam. Lister, A.M. 1998. The age of Early Pleistocene mammal faunas from the ‘Weybourne Crag’ and Cromer Forest-bed Formation (Norfolk, England). Meded. Nederl. Inst. Toegepaste Geowetensch. TNO 60, 271-280. Lister, A.M. & Brandon, A. 1991. A pre-Ipswichian cold-stage mammalian fauna from the Balderton Sand and Gravel, Lincolnshire, England. J. Quat. Sci. 6, 139-157. Lister, A.M. & Joysey, K.A. 1992. Scaling effects in elephantid dental evolution - the example of Eurasian Mammuthus. In Structure, Function and Evolution of Teeth (eds. Smith, P. & Tchernov, E.), 185-213. Jerusalem: Freund. Maglio, V.J. 1973. Origin and evolution of the Elephantidae. Trans. Am. Phil. Soc. 63, 1-149. Marabini, S., Lenaz, R. & Vai, G.B. 1987. Pleistocene superiore e Olocene del margine pedeappenninico romagnolo: rapporto preliminare. Rend. Soc. Geol. It. 10, 33-37. Masini, F. Abbazzi, L., Ferretti, M. & Mazza, P. 1995. Fauna a mammiferi terrestri dale ‘Sabbie gialle’ Quaternarie di Imola. In (Pacciarelli, M. & Vai, G.B., Eds) Musei Civici di Imola. La collezione Scarabelli. I. Geologia, 314-326. Bo: Casalecchio. Maul, L. 1990. Überblick über die unterpleistozänen Kleinsäugerfaunen Europas. Quartärpaläontologie 8, 153-191.

Meijer, T. & Preece, R.C. 2000. A review of the occurrence of Corbicula in the Pleistocene of North-West Europe. Geologie en Mijnbouw/Netherlands Journal of Geosciences 79, 241-255. Murton, J.B., A. Baker, D.Q. Bowen, C.J. Caseldine, G.R. Coope, A. Currant, J.G. Evans, M.H. Field, C.P. Greene, J. Hatton, M. Ito, R.L. Jones, D.H. Keen, M.P. Kerney, R. McEwan, D.F.M. McGregor, D. Parish, D.C. Schreve & P.L. Smart. A late Middle Pleistocene interglacialperiglacial-interglacial sequence (oxygen isotope stages 7 to 5e) near Marsworth, Buckinghamshire, UK. Quaternary Science Reviews, in press. Napoleone, G., Albianelli, A., Azzaroli, A. & Mazzini, M. in press. Dating the Upper Valdarno fossil vertebrate collections in the Natural History Museum of Florence from its magnetochronology in the framework of the Northern Apennine continental sequences, Italy. Boll. Soc. Paleont. It. Pevzner, M., Tesakov, A., Vangengeim, E. 1998. The position of the Tizdar locality (Taman Peninsula, Russia) in the magnetochronological scale. Paludicola, 2, 95-97. Preece, R.C. & Parfitt, S.A. 2000. The Cromer Forest-bed Formation: new thoughts on an old problem. In The Quaternary of Norfolk and Suffolk: Field Guide(eds. Lewis, S.G., Whiteman, C.A. & Preece, R.C), 1-27. Quaternary Research Association, London. Rackham, D.J. 1978. Evidence for changing vertebrate communities in the Middle Devensian. Quaternary Newsletter 25, 1-3. Radulesco, C. & Samson, P.M. 1995. Review of the Villafranchian s.s. faunas of Romania. Il Quaternario 8, 377-382. Ricciardi, E. 1961. Analisi pollinologica dei sedimenti lacustri lignitiferi di Pietrafitta (Perugia, Italia). Pollen et Spores 3 (2). Sardella, R., & 17 others. 1998. Mammal faunal turnover in Italy from the Middle Pliocene to the Holocene. Meded. Nederlands Inst. Toegepaste Geowet. TNO 60, 499-511. Schreve, D.C. in press. Differentiation of the British late Middle Pleistocene interglacials: the evidence from mammalian biostratigraphy. Quaternary Science Reviews. Schreve, D.C. & Bridgland, D.R. in press. Correlation of English and German Middle Pleistocene fluvial sequences based on mammalian biostratigraphy. Netherlands Journal of Geosciences. Scott, K. 1986. The large mammal fauna. In La Cotte de St Brelade 1961-1978: Excavations of CBM McBurney (eds. Callow, P. & Cornford, J.M.), 159-183. Geo Books, Norwich. Shackleton, N.J. 1995. New data on the evolution of climatic variability. IN (Vrba, E.S., Denton, G.H., Partridge, T.C. & Burckle, L.H. (Eds.) Plaeoclimate and Evolution with Emphasis on Human Origins, 242-248. London: Yale University Press.

Sher, A.V. 1971. Pleistocene mammals and stratigraphy of the Far Northeast USSR and North America. Moscow, Nauka, 1-310. [In English: Sher, A.V. 1974. Intern. Geology Review, 16, 1282. Sher, A.V. 1997. A brief overview of the Late Cenozoic history of the Western Beringian lowlands. In: M.E. Edwards, A.V.Sher, R.D.Guthrie, Eds., Terrestrial Paleoenvironmental Studies in Beringia. Fairbanks: University of Alaska Museum, 3-6. Sher, A.V. 1981. To the age substantiation of unconsolidated deposits in the middle course of the Alazeya R. (Kolyma Lowland). Doklady AN SSSR, 258, 179-182. Sher, A.V. 1984. Age of Quaternary deposits of the Yana-Kolyma Lowland and its mountainous surroundings. Doklady AN SSSR, 278, 708-713. Sher, A.V., Kaplina, T.N., Giterman, R.E., Lozhkin, A.V., Arkhangelov, A.A., Kiselyov, S.V., Kouznetsov, Yu.V., Virina, E.I. & Zazhigin, V.S. 1979. Late Cenozoic of the Kolyma Lowland: XIV Pacific Science Congress, Tour Guide XI, Khabarovsk August 1979. Moscow, Academy of Sciences of the USSR, 1-114. Sher, A.V., T.N.Kaplina & M.G.Ovander. 1987. Unified Regional Stratigraphic Chart for the Quaternary deposits in the Yana-Kolyma Lowland and its mountainous surroundings. Explanatory Note. In: Shilo N.A., Ed., Resheniya Mezhvedomstvennogo stratigraficheskogo soveshchaniya po chetvertichnoy sisteme Vostoka SSSR (Magadan, 1982). Ob"yasnitel’nye zapiski k regional’nym stratigraficheskim skhemam., Magadan, SV KNII DVO AN SSSR, 2969. Soffer, O. 1993. Upper Paleolithic adaptations in central and eastern Europe and man/mammoth interactions. In From Kostenki to Clovis: Upper Paleolithic paleo-Indian adaptations. (eds. O. Soffer & N. Praslov), 31-49. Plenum, New York. Steinmüller, A. 1972. Die Schichtenfolgen von Süssenborn und Voigtstedt und die Gliederung des Mittelpleistozäns. Geologie 21, 149-165. Stremme, H.E. 1985. Altersbestimmung und Paläoböden in der Quartärstratigraphie. Bull. de l’AFEQ 2-3, 159-166. Stuart, A.J. 1981. A comparison of the middle Pleistocene mammal faunas of Voigtstedt (Thuringia, GDR) and West Runton (Norfolk, England). Quartärpaläontologie 4, 155-163. Stuart, A.J. 1982. Pleistocene Vertebrates in the British Isles. London: Longman. Stuart, A.J. 1996. Vertebrate faunas from the Early Middle Pleistocene of East Anglia. In The Early Middle Pleistocene in Europe(ed. C. Turner), 9-24. Balkema, Rotterdam.

Stuart, A.J. & Lister, A.M. 2000. Pakefield, Suffolk. In The Quaternary of Norfolk and Suffolk: Field Guide(eds. Lewis, S.G., Whiteman, C.A. & Preece, R.C), 219-223. Quaternary Research Association, London. Stuart, A.J. & Lister, A.M. in press. The mammalian faunas of Pakefield/Kessingland and Corton, Suffolk, U.K.: evidence for a new temperate episode in the British early Middle Pleistocene. Quaternary Science Reviews. Sutcliffe, A.J. & Kowalski, K. 1976. Pleistocene rodents of the British Isles. Bull. BMNH (Geol) 27, 33-147. Szabo, B.J. & Collins, D. 1975. Ages of fossil bones from British interglacial sites. Nature 25, 680-682. Titov, V.V. 1999. The Late Pliocene fauna of large mammals in the northeastern Azov region. Dissertation for the degree of Candidate in Biology. Paleontological Institute RAS, 1-387 (In Russian) Torre, D., Albianelli, A., Bertini, A., Ficcarelli, G., Masini, F. & Napoleone, G. 1996. Paleomagnetic calibration of Plio-Pleistocene mammal localities in central Italy. Acta zool. cracov. 39, 559-570. Turner, C. 1996. The Early Middle Pleistocene in Europe. Balkema, Rotterdam. Turner, E. (ed.) 1997. Ariendorf: Quaternary deposits and Palaeolithic excavations in the Karl Schneider gravel pit. Jahrb. Romisch-Germanischen Zentralmuseums Mainz 44, 1-191. van der Plicht, J., Damblon, F., Haesaerts, P. & Svoboda, J. 1997. http://www.cio.phys.rug.nl/HTML-docs/Verslag/97/report_95-97.htm Vai, G.B. 1984. Quando barrivano sul Castellaccio gli ultimi elefanti. Pagine di vita e storia imolesi 2, 195-219. Vangengeim, E.A., Vekua, M.L., Zhegallo, V.I., Pevzner, M.A., Taktakishvili, I.A. & Tesakov, A.S. 1991. Position of the Tamanian faunal complex in stratigraphic and magnetochronological scales. Byull. Komm. Izuch. Chetv. Per. 60, 41-52. Virina, E.I., V.S.Zazhigin, and A.V.Sher 1984. Paleomagnetic characteristic of the type sites of the Olyorian Faunal complex (Kolyma Lowland). Izvestiya Academii Nauk SSSR. Ser. Geologicheskaya, 11, 61-72 Virina, E.I. 1997. Paleomagnetic stratigraphy of the Pliocene/Pleistocene sediments of the Kolyma Lowland and some problems of correlation with sites in Alaska and the Yukon. In: M.E. Edwards, A.V.Sher, R.D.Guthrie, Eds., Terrestrial paleoenvironmental studies in Beringia, Fairbanks: University of Alaska Museum, 19-24.

West, R.G. 1980. The Preglacial Pleistocene of the Norfolk and Suffolk Coasts. Cambridge University Press, Cambridge. Wiegank, F. 1997. Palaomagnetische Charakteristik des Unterpleistozans von Untermassfeld. In (Kahlke, R.-D., ed.) Das Pleistozan von Untermassfeld bei Meiningen (Thuringen), Teil 1. Monogr. Röm-Germ. Zentralmus. Mainz 40, 63-69. Zagwijn, W.H. & De Jong, J., 1984. Die Interglaziale von Bavel und Leerdam und inre stratigraphische Stellung im niederländischen Fruhpleistozän. Meded. Rijks. Geol. Dienst 37, 155-169. Zazhigin, V.S. 1976. Early evolutionary stages of collared lemmings (Dicrostonychini, Microtinae, Rodentia) as characteristic representatives of Beringian subarctic fauna. In: Kontpimavichus V.L., Ed., Beringiya v Kaynozoe. Vladivostok, DVNTs AN SSSR, 280-288 Zazhigin, V.S. 1997. Late Pliocene and Pleistocene rodent faunas in the Kolyma Lowland: possible correlations with North America In: M.E. Edwards, A.V.Sher, R.D.Guthrie Eds.Terrestrial paleoenvironmental studies in Beringia. University of Alaska Museum, Fairbanks, 25-29. Zazhigin, V.S. 1998. Taxonomy and evolution of the genus Allophaiomys (Microtinae: Rodentia: Mammalia) in Siberia. Paludicola, 2, 116-125