From the Lower Pleistocene Hominid-Bearing Level of Trinch

Gloria Biochronological implications of the Cuenca-Bescós, Arvicolidae (Rodentia, Mammalia) from César Laplana & the Lower Pleistocene hominid-bearing Jose Ignacio Canudo level of Trinchera Dolina 6 (TD6, Paleontología Ciencias, Atapuerca, Spain) Universidad de Zaragoza, 50009 Zaragoza, Spain. UA Level TD6 of the Trinchera Dolina Section in the railway cutting of Museo Nacional Ciencias the Sierre de Atapuerca (Trinchera del Ferrocarril) has yielded a rich E-mail: small mammal assemblage (26 species) in association with fossil [email protected] human remains of Homo antecessor. The arvicolids of TD6 are identified as: Mimomys savini, Microtus seseae, Stenocranius gregaloides, Received 30 April 1998 Terricola arvalidens, Iberomys huescarensis, Allophaiomys chalinei, and Revision received Pliomys episcopalis. The rodent association also includes large rodents 13 October 1998 (i.e., Castor fiber, Marmota sp., and Hystrix refossa) and the small and accepted 16 March Allocricetus sp., Eliomys helleri, Micromys minutus, and Apodemus aff. 1999 flavicollis. The small vertebrate remains also include Insectivora Keywords: hominid site, (Beremendia fissidens, Sorex sp, Neomys sp., Crocidura sp., Galemys sp., Mimomys savini, Microtus, Talpa sp., Erinaceus sp.), Chiroptera (Miniopterus schreibersii, Myotis biochronology, Lower sp., Rhinolophus sp.), and Lagomorpha (Oryctolagus sp., Lepus sp.), as Pleistocene. well as lizards, birds and amphibians. The H. antecessor remains are derived from a 15 cm thick layer at the top of TD6 (TD6–T36–43), where A. chalinei, H. refossa and Marmota sp. do not occur. The paleomagnetic Matuyama/Brunhes boundary is found in the over- lying level TD7 of the Gran Dolina Section. On the basis of the arvicolids, TD6 can be referred to the Biharian biochron. The Matuyama/Brunhes boundary is fixed in the late Biharian (Microtus– Mimomys rodent Superzone). The species M. savini (without M. pusillus), as well as the evolutionary stage of Microtus s.l., are characteristic of the Late Biharian. The evolutionary level of the species M. savini, T. arvalidens, S. gregaloides indicates that TD6 is older than West Runton (type Cromerian). In the Trinchera Dolina Section we are able to calibrate, for the first time, the evolutionary level of important biochronological markers with magnetostratigraphy. We propose that a radiation of Microtus s.l., along with the first appear- ance of primitive S. gregaloides, T. arvalidens and Iberomys, took place just before the Matuyama/Brunhes boundary. These species can be considered as characteristic elements of early Pleistocene faunas. 1999 Academic Press

Journal of Human Evolution (1999) 37, 353–373 Article No. jhev.1999.0306 Available online at http://www.idealibrary.com on

Introduction et al., 1995; Parés & Pérez-González, 1995). The unique combination of traits in these In July 1994, the Atapuerca team discovered hominids led Bermúdez de Castro et al. fossil human remains at the Trinchera (1997) to recognize a new species, Homo Dolina site. The fossils were recovered from antecessor. Since 1994, our eﬀorts have been the upper part of the TD6 level, from T36 focused on the study of the small vertebrate (Aurora stratum=TD6 T36–43), in a layer associations of the hominid-bearing layer. of brownish-red clay 15 cm thick (Carbonell Consequently, the aim of this paper is to

0047–2484/99/090353+21$30.00/0 1999 Academic Press 354 . -´  ET AL. discuss the biochronologically significant ( 4) The rodent association in upper levels rodents (i.e., the voles and cricetids) from at Atapuerca includes Terricola grega- the Aurora level, and to review their system- loides, Terricola arvalidens and Iberomys atics. The results help to establish the brecciensis. biochronological position of the fossil ( 5) The rodent association has mixed human remains from the TD6 level within elements derived from the early the broader context of the European Middle Pleistocene and the late Early Pleistocene. Pleistocene. In order to avoid any misleading interpret- ation of the current research and results, Previous work especially those related to the biostratigra- Some earlier work led to contradictory con- phy and dating, the following points of clusions about the Trinchera Dolina rodent clarification are presented: assemblages and the biochronology of the ( 1) Atapuerca is not a section. It includes TD section. The problem probably origi- several sites with different geological, nated during the early years of the research stratigraphical, paleontological and project, when several researchers took small archaeological contexts as indicated samples of sediments from one or more of above. For further details the reader is the Atapuerca sites. From these preliminary referred to Pérez González et al. studies, references to ‘‘Atapuerca 1’’, ‘‘2’’, (1995), Parés & Pérez González ‘‘3’’, ‘‘Atapuerca section’’, ‘‘Atapuerca (1995), Arsuaga et al. (1996, 1997), lower levels’’, ‘‘Atapuerca upper levels’’, and Rosas et al. (1998). ‘‘Atapuerca’’ ‘‘ATA B’’ appeared in the ( 2) If there are any upper levels at Ata- literature (Brunet-Lecomte, 1989; Brunet- puerca they should be considered Lecomte & Chaline, 1991; Agustí, 1991; either the upper levels of the Trinchera Agustí et al., 1993, 1997). The lack of Dolina Section (TD10, TD11), the stratigraphical and geological context upper levels of the Galeria complex generated confusion about the rodent sites (TG), the upper levels of the assemblages at Atapuerca. An exception was Elefante Section (TE), the upper the study by Martínez Navarro et al. (1997), levels of Sima de los Huesos Site (SH, in which the Atapuerca rodents were cor- see Cuenca-Bescós et al., 1997), or the rectly attributed. Moreover, during the early upper levels of any of the cavities filled phases of the Atapuerca project (1978– with Pleistocene and Holocene sedi- 1984), few small mammal samples were ments in Atapuerca Hill. However, collected. Since the entire faunal assemblage there are no known upper levels at the was not evaluated some results were Matuyama/Brunhes boundary in any ambiguous (Gil, 1996a,b,c, 1997a,b; Gil & of the caves or fissures. The only pub- Sesé, 1991; Sesé & Sevilla, 1996), as will be lished reference to the Matuyama/ discussed below. Brunhes boundary are those of the From the previous literature on Atapuerca TD2 level (Carracedo et al., 1987), readers would conclude that: TD7 level (Parés & Pérez González, ( 1) Atapuerca is a single site or section 1995) and Sima del Elefante (Arsuaga with several paleontological levels. et al., 1997). ( 2) The upper levels are placed around the ( 3) There is no question that the rodent Matuyama/Brunhes boundary. faunas of Trinchera Dolina are more ( 3) The rodent association is more mod- modern than those of Orce 7. The ern than Orce 7. latter site has Allophaiomys pliocaenicus,      355

a primitive and probably ancestral lower levels of Trinchera Dolina, form of the Microtus s.l. group. How- Iberomys is represented by its primitive ever, it is uninformative, since an early form, I. huescarensis. Moreover, T. sub- Microtus radiation took place in the terraneus never reached the Ibero- early Lower Pleistocene (Chaline, Lusitanian region (Brunet-Lecomte, 1972; van der Meulen, 1973; 1989; Musser & Carleton, 1993). The Rabeder, 1981; Agustí, 1991). More ‘‘subterraneus’’ form at Atapuerca has interesting is the fact that Allophaiomys been confused with T. arvalidens from chalinei is found in association with I. the lower TD levels, and with Terricola huescarensis, an early representative of atapuerquensis from the upper TD the I. brecciensis–I. cabrerae lineage in levels and TG. the lower levels TD3–TD8a of Trinchera Dolina (Cuenca-Bescós Small mammals from TD6 et al., 1998), as well as at Cueva The present list of small mammals from Victoria and Almenara 3 (Casablanca). TD6 is as follows: Moreover, M. savini has a high proportion of individuals with an enamel Rodentia: Mimomys savini, Microtus seseae, islet (Laplana & Cuenca-Bescós, Stenocranius gregaloides, Terricola arvalidens, 1996, Figure 3). This is a primitive Iberomys huescarensis, Allophaiomys chalinei, trait that indicates that M. savini from Pliomys episcopalis, Allocricetus sp., Apodemus TD is older than the M. savini assem- aff. flavicollis, Micromys minutus, Eliomys hel- blage from the type locality of West leri, Marmota sp., Hystrix refossa, Castor fiber Runton (dated at 450 ka, Rink et al., Insectivora: Beremendia fissidens, Sorex sp., 1996), as well as Voigstedt, Prezletice, Neomys sp., Crocidura sp., Galemys sp., and Huescar 1. In TD3–TD6 we Talpa sp., Erinaceus sp. found primitive representatives of the Lagomorpha: Oryctolagus sp., Lepus sp. S. gregaloides and T. arvalidens line- Chiroptera, Myotis sp., Rhinolophus sp., ages. The recently defined M. seseae is Miniopterus schreibersii a new species not previously known This list is still preliminary, since several from other localities. taxa will need revising as new material ( 4) There is no ‘‘Atapuerca level’’ with a becomes available. For example, our current rodent association comprising T. views on Microtus s.l., Pliomys, Apodemus, arvalidens, T. gregaloides (which is S. Allocricetus, Marmota, Chiroptera, Lagomor- gregaloides, not Terricola, Cuenca- pha and Insectivora could well change with Bescós et al., 1995) and I. brecciensis. additional information. Iberomys brecciensis is a derived species from the Middle Pleistocene which appears to be associated with Arvicola Material and methods in the Atapuerca sites TG, TD10 and 11, and in other European sites such Excavations at Atapuerca have been con- as Cúllar de Baza 1, Aragó, and Saint ducted by our team since 1978. The bio- Estève Janson (Chaline, 1971, 1972; stratigraphic framework has been advanced Ruiz Bustos, 1976). The references of significantly by the introduction in 1991 of a Gil (1996a,b)toI. brecciensis in the programme of screening in the River lower levels of Trinchera Dolina (as Arlanzón (Figure 1) of all sediments well as the presence of T. subterraneus removed from the excavations (Cuenca- and A. bursae) are not valid. In the Bescós et al., 1995, 1997; Carbonell et al., 356 . -´  ET AL.

Figure 1. Small mammal washing–sieving device and part of the ‘‘river’’ team during the Atapuerca ﬁeld campaign of July 1995 in the Arlanzon River (Burgos, Spain).

1995; Laplana & Cuenca-Bescós, 1996, between TD6 T36–43 and the lower part of 1997). TD6 (see Figure 7). The levels being excavated are carefully labelled with pertinent stratigraphic infor- Systematic descriptions of rodents mation. With the help of students from from TD 6 T36–43 Castilla y León, who assist every season, we have been able to process several tons of Since the samples are large, we use only the sediment. Sorting of the fossils is carried out first lower molar (M1) from each of the in our laboratory, and the specimens are analyzed vole species. Descriptions and labelled and stored in the Zaragoza Museum taxonomic discussions follow the terminol- of Paleontology. Each site is designated with ogy of van der Meulen (1973), Rabeder a letter, such as TD for Trinchera Dolina, (1981) and Rekovets & Nadachowski followed by the stratigraphical level and the (1995). Measurements in the text and tables archaeological sublevel (T), and the letter are minimum (Min), mean (M) and maxi- and number of the 1 m2 grid which divides mum (Max) values of the parameters L the excavation. This is followed by the let- (length of the anteroconid complex ACC+ ters ATA (for Atapuerca) and a number length of the trigonid–talonid complex (indicating the year of collection). An TTC) and W (width of the posterior part of example of a label for the Aurora stratum the ACC) in millimetres, followed by the small mammal sample would read as fol- ratios A/L (=a/L*100), B/W (=b/W*100), lows: TD6T40G17ATA95. In the case of C/W (=c/W*100) and D/E (=d/e*100). N Aurora, the level is stratigraphically uniform is the total number of observations for each and lithologically it represents a single geo- parameter or ratio. As in van der Meulen logical unit. The small mammal assemblages (1973) measurements do not include the of the Aurora level are also homogeneous, enamel band. Arvicolid molars are distin- and represent a single time horizon. guished by their flat occlusal surface, con- Although the taphonomy of the small mam- sisting of dentine bordered by a strip of mal fossils from TD6 is homogeneous enamel. The enamel ribbon shows a lingual (Fernández Jalvo & Andrews, 1992), there and buccal series of folds more or less are differences in species composition pointed inward and outward (reentrant, RA      357 and salient angles, SA). The RA depth and describe here only the main peculiarities of angle are variable and generally show the vole species identified at TD6 T36–43. diagnostic differences between species. The Family Arvicolidae GRAY, 1821 surface can be divided into different ‘‘fields’’ Species Mimomys savini Hinton, 1910; of dentine with slight or no communication Figures 2 and 3; Table 1. with the dentine of neighbouring areas of the tooth (posterior lobe, PL, triangles, T, and anterior cap, AC). The TTC is formed by Description the PL and T1–T3. The ACC is formed by Mimomys savini is the largest vole found in the mesial T in front of T3 and the AC. The TD6. The occlusal surface of M1 shows RA may have crown cementum. The SA features typical of this species. The M1 is constitutes the outline of T. broad. Adult molars have two roots which Following van der Meulen (1973): the form late in development. Crown cementum parameter a is the length of the anteroconid is abundant in adults, but may be absent in complex and A/L expresses the relative young individuals. The enamel thickness length of the ACC in comparison to L. The is well differentiated, and thicker on the parameter b measures the shortest distance distal sides of SA (T1–T5). In young between the buccal reentrant angle 3 specimens (without roots) the enamel islet is (BRA3) and the lingual reentrant angle 4 present in 14% of the assemblage, and the (LRA4). It is the expression of the constric- typical Mimomys enamel differentiation tion of the neck between T4T5 and AC, and is not evident. There are four lingual and three buccal SA. RA are 31 lingual and B/W gives the degree of separation between 2 21 buccal. these fields. The parameter c measures the 2 shortest distance between BRA3 and the lingual reentrant angle 3 (LRA3), and C/W Discussion is the expression of the separation (or con- As for the presence of roots, the enamel islet fluency) of T4 and T5. Primitive Microtus is less frequent in more modern populations s.l. species molars yield high values of B/W of M. savini (i.e., West Runton 7%, Huescar and C/W and low A/L. High C/W values are 14%, Voigtstedt and Prezletice 0–1%) than characteristic of species with a ‘‘pitymyan in older populations (i.e., Chiscau1 and rhombus,’’ such as Terricola and Pitymys. TD6 14%). This seems not to be a gradual High C/W also characterizes primitive trend, since there are assemblages such as Microtus s.l. species and Allophaiomys s.l. those at Hohen Sülzen and Kozi Grzbiet Parameter d measures the shortest distance with fewer individuals with enamel islets. between BRA4 and LRA5, and e represents Hohen Sülzen is supposed to be older than the greatest width of the anterior part of the TD6 (Storch et al., 1973; Von Koenigswald ACC measured in T6T7 (or the distance & Tobien, 1990). However, Fejfar & LSA5–BSA4). The ratio D/E shows how Heinrich (1990) remark that the disappear- T6T7 is developed in relation to the width ance of M. pusillus (or small-sized Mimomys) of ACC (Rekovets & Nadachowski, 1995). provides an insufficient time marker since it Parameters e and d can be measured only is found in localities above the Matuyama/ when T6 and T7 are slightly to well- Brunhes boundary, such as at Kärlich E developed. The arvalid–nivalid type is (van Kolfschoten & Turner, 1996). expected to yield greater values of e than the The Polish locality Kozi Grzbiet poses gregalid–hintonid type (see D/E ratios of another problem: it has a derived M. savini Tables 2, 3 and 4). A more detailed study of assemblage with no enamel islet on the the small mammals is in progress, so we will occlusal surface of M1. However, its age of 358 . -´  ET AL.

Figure 2. Occlusal surface of left molars of Mimomys savini from Atapuerca TD6 hominid-bearing level.

(a) M1 with enamel islet (TD6T38H-18ATA94A-1). (b) M1 (TD6T39G-17ATA94A-1). (c) M1 with prism fold (TD6T37I-16ATA94A-75). (d) M1 (TD6T39H-18ATA94A-1). (e) Left mandible with M1, M2,M3 (TD6T40G-18ATA95A-1). Scale bar represents 1 mm.

630–560 ka, based on complex ﬂuoro- over 10% are Early Pleistocene and those chloro-apatite and collagen dating methods below 10% are Middle Pleistocene. Further using bones (Nadachowski, 1985, 1990), studies of M. savini from chronostrati- would suggest that Kozi Grzbiet should have graphically well calibrated Pleistocene sec- M. savini assemblages more primitive than tions can be used to support or to refute this those from West Runton, which have been proposition. dated to 450–500 ka using ESR dating of tooth enamel (Rink et al., 1996). Genus Microtus SCHRANK, 1798 M. savini evolutionary stages seem to be Microtus seseae, Terricola arvalidens, Iberomys good biostratigraphical markers when the huescarensis, and Stenocranius gregaloides are percentages of enamel islets and root devel- grouped together as Microtus s.l. In our opment are considered. We propose that M. opinion it is more convenient for the pur- savini assemblages in which more than 10% poses of biostratigraphy to consider Microtus of M1 have an enamel islet are older than the s.s., Terricola, Iberomys and Stenocranius sep- assemblages of M. savini with fewer than arately (Fejfar & Heinrich, 1990; Musser & 10%. Using the Trinchera Dolina Section as Carleton, 1993; García-Perea & Gisbert, a standard, we propose that the assemblages 1997).      359

SAINT ESTEVE-JANSON, CULLAR DE BAZA (–%), I (I) ISERNIA Mimomys Arvicola II Arvicola III II Mimomys IV PREZLETICE (0%), WEST RUNTON (7%), VOIGTSTEDT (<1%), HUESCAR 1 (4%), III KOZI GRZBIET (0%), HOHENSULZEN (2%), IV Middle Pleistocene

Early Pleistocene rooted rootless TD6 (14%), CHISCAU 1 (14%)

a b

Figure 3. (a) Model of evolutionary transition from Mimomys to Arvicola, modiﬁed from von Koenigswald (1982) and von Koenigswald & Kolfschoten (1996), and (b) succession of localities with Mimomys savini

and Arvicola cantianus showing the general trend towards the reduction of the enamel islet in M1 from the Early to the Middle Pleistocene. In (a) the horizontal arrow indicates the direction of evolution, the vertical arrow indicates the ontogenetic pathway. During the transition, the ﬁrst ontogenetic stage is accelerated, the two last stages are retarded or reduced, and the second one is prolonged. As shown by this model, the proportion of individuals with enamel islets (a character associated with ontogenetic stage I), and with closed linea sinuosa or presence of roots (characters associated with stages III and IV), are

gradually reduced with time. In (b), populations from the Early Pleistocene (with more than 10% of M1 with enamel islets) precede those from the Middle Pleistocene (which have lower proportions). In Arvicola cantianus this character is vestigial (Cúllar de Baza, Ruiz Bustos, 1976). Figure (b) after data from Chaline (1972), Pasquier (1972), Storch et al. (1973), Ruiz Bustos (1976), Sala (1983), Nadachowski (1985), Terzea (1989), Maul (1990a), von Koenigswald & van Kolfschoten (1996) and this work.

Table 1 Measurements of the M1 of Mimomys savini from the Atapuerca TD6 hominid-bearing level

Mean S.D. S.E. n Minimum Maximum

L (mm) 3·508 0·245 0·061 16 3·053 3·895 W (mm) 1·251 0·070 0·017 18 1·128 1·383

Species Microtus seseae (GIL, 1997); Figure Emended diagnosis 4(a–e); Table 2 The most striking feature of M. seseae is its pars. 1995 Terricola arvalidens—Cuenca- M1 provergent occlusal surface and its rela- Bescós et al., 1995 tively elongated provergent T6. M. seseae 1997 Microtus lignani—Gil, 1997 has a uniform morphology, except in the 360 . -´  ET AL.

Figure 4. Occlusal surface of Microtus seseae and Terricola arvalidens from the Atapuerca TD6 hominid-

bearing level. All specimens are from the right side except (a) and (e). Microtus seseae: (a) M1 (TD6T36H-16ATA94B). (b) M1 (TD6T37I-16ATA94A-2). (c) M1 (TD6T40/41H-16ATA95A-28). (d) M1 (TD6T40/41H-16ATA95A-31). (e) M1 (TD6T40/41H-16ATA95A-12). Terricola arvalidens: (f) M1 (TD6T40/41G-17ATA95A-4). Scale bar represents 1 mm.

Table 2 Measurements of the M1 of Microtus seseae from the Atapuerca TD6 hominid-bearing level

Mean S.D. S.E. n Minimum Maximum

L (mm) 2·711 0·153 0·025 36 2·339 3·110 W (mm) 0·895 0·055 0·009 38 0·753 1·027 A/L 53·501 1·877 0·313 36 46·866 56·936 B/W 6·722 3·722 0·604 38 2·388 22·306 C/W 17·440 4·520 0·733 38 9·814 28·060 D/E 58·999 12·396 2·095 35 39·228 79·747

variable confluence of T4 and T5, and in the type in most specimens, but there are some development of T6. BRA4 is absent or specimens (less than 1%) with no enamel incipient; when present, it lacks or has little differentiation. BRA3 and LRA4 are alter- crown cementum. LRA5 is reduced and nating, as are BSA4 and LSA5. T6 can be may or may not have crown cementum. The smaller than T7, but is always present. In enamel differentiation is of the microtus some specimens the T6 is provergent,      361 strongly elongated in the bucco-distal plane. Microtus lineages (i.e., M. seseae), the closed The BRA are mesially more elongated than T4T5 are typical of M. nivaloides, M. malei the LRA in some specimens (provergent, and M. nivalinus. The elongated T6 is typi- sensu Rabeder, 1981). The AC2 is simple, cal of M. seseae, T. henseli and M. nivalinus, short and confluent with T6 and T7, both and separates these species from the others of which are broadly confluent. T6T7 is considered here. The enamel-free areas are separated from T4T5 by an asymmetrical, extensive in I. huescarensis. Stenocranius narrow neck (M–B/W 6·72). In the ACC, gregaloides, T. henseli and M. nivalinus the T4 and T5 may be confluent or nearly show an intermediate condition, whereas M. closed (M–C/W 17·44). The anterior seseae, T. arvalidens, M. nivaloides and M. enamel-free area is extended to the buccal malei have the primitive condition of a less side of the crown; in some individuals it developed enamel-free area on the mesial extends to the BSA 4. side. The most mesial LRA and BRA are parallel in I. huescarensis, T. arvalidens, S. Differential diagnosis and discussion gregaloides and M. nivalinus, whereas in M. The aim of the emended diagnosis is to seseae, T. henseli, M. nivaloides and M. malei provide a better understanding of the species they are alternating. T4 in M2 is open in M. described by Gil (1997b). Gil failed to com- seseae and T. arvalidens and closed in I. pare M. seseae with Microtus species of the huescarensis and S. gregaloides. In the other same age and/or morphologically related species we cannot study this character. forms, and his diagnosis and description Our preliminary work leads us to propose could lead to the conclusion that M. seseae is that M. seseae is closely related to M. nival- a variant of T. subterraneus. This discussion oides, M. malei, M. nivalinus and Tyrrheni- compares the morphology of M. seseae to cola. It represents one of the earliest species other Microtus s.l. species of small size (i.e., of a second radiation of the Microtus s.s. I. huescarensis, T. arvalidens, S. gregaloides, lineage that occurs at the end of the Early Tyrrhenicola henseli, Microtus nivaloides, Pleistocene. Microtus malei and Microtus nivalinus). Up to Gil (1997b) failed to compare his T. 10 characters have been studied on the seseae/M. lignani to the species M. nivaloides occlusal surface of M1, and one character and M. nivalinus. Moreover, we have to has been studied in M2. remark that M. nivalinus as used here is We observe that M. seseae and modern based on the holotype drawn by A. P. populations of T. henseli share a provergent Currant from the Natural History Museum occlusal surface and deep RA. The elon- in London (personal communication). The gated T6 is shared also by M. nivalinus. The holotype of M. nivalinus shows considerable general M. nivalinus morphotype seems to affinities with M. oeconomus, and some point to a close relationship between the authors (e.g., Rekovets & Nadachowski, three species. The BRA4 is deep only in 1995, based on Ukrainian material) place T. arvalidens (as in other Terricola species) both taxa in M. oeconomus nivalinus. The and this separates Terricola from Microtus s.l. Ukrainian species is somewhat different The deep LRA is shared by I. huescarensis from the type material from West Runton and M. nivaloides. The LRA5 is deep only in and, after examining the holotypes, it seems T. henseli. Except in S. gregaloides and M. improbable that M. nivalinus and M. nivalinus, where it is reduced or absent, this oeconomus are synonyms. Sutcliffe& feature is developed (but shallow) in other Kowalski (1976) mention the occurrence of species. The T4T5 confluency (‘‘pitymyan intermediate specimens between M. nivali- rhombus’’) is characteristic of primitive nus and M. nivaloides, but the proposed 362 . -´  ET AL.

Table 3 Measurements of the M1 of Terricola arvalidens from the Atapuerca TD6 hominid-bearing level

Mean S.D. S.E. n Minimum Maximum

L (mm) 2·714 0·214 0·062 12 2·236 3·083 W (mm) 0·883 0·047 0·013 12 0·809 0·984 A/L 53·256 1·714 0·495 12 50·843 56·361 B/W 5·062 1·054 0·304 12 3·500 7·031 C/W 19·717 4·193 1·210 12 13·056 26·042 D/E 48·306 11·427 3·299 12 26·706 63·816

synonymy seems inappropriate. More con- and LRA5 and shorter AC3) to that of the vincing is the synonymy proposed by van early Tiraspolian faunas of the Ukraine than der Meulen (1973) and Rekovets & it is to the late Tiraspolian and younger Nadachowski (1995) between M. nivaloides faunas. The neck between T4T5 and the and M. arvalinus, with M. nivaloides being anterior field is very long and straight in the species name used in our work. The some specimens. This species differs in size TD6 levels with M. seseae and Le Vallonnet and in its more primitive aspect from typical which contains M. nivaloides, are younger Middle Pleistocene T. arvalidens, and it than Fuente Nueva 3. This latter locality could possibly be distinguished as a separate contains Allophaiomys burgondiae, a species chronospecies when more material becomes more primitive than the M. nivaloides– available. M. seseae group (Martínez Navarro et al., The separation of T. arvalidens and M. 1997). seseae is problematic. Both species have similar length, width and A/L ratios (B/W and Terricola arvalidens Kretzoi, 1957; Figure C/W are not significantly different). How- 4(f); Table 3 ever, they are morphologically dissimilar (for synonyms see Rekovets & in T6, T7, and ACC3. The main differences Nadachowski, 1995) are the lack of provergent RA and an elongated T6, the lower d (and D/E) par- Description ameters, and the more or less parallel mesial The species is well represented in levels RA in T. arvalidens (see discussion above). TD3, TD4 and TD4BW (Cuenca-Bescós The ratio D/E is highly variable (see Tables et al., 1995). The species from level TD6 is 2 and 3), but it differs between the two similar to that in levels TD3–TD4. species. We believe that M. seseae and T. arvalidens Discussion are part of a cladogenetic event that leads to Terricola arvalidens from Trinchera Dolina is the Terricola species on one branch and to somewhat smaller and more primitive than M. nivalinus, M. nivaloides, Tyrrhenicola and the species described from the Ukraine M. seseae on the other. (Rekovets & Nadachowski, 1995) and from the type locality of West Runton (Cuenca- Iberomys huescarensis Ruiz Bustos, 1988 Bescós et al., 1995). The T6 is somewhat Iberomys brecciensis—Mazo et al., 1985 less developed than in the Ukrainian species. Terricola gregaloides—Agustí et al., 1993 The anteroconid complex is more similar pars. 1997 Iberomys brecciensis—Gil, 1997a; in its simplicity (less development of BRA4 Figure 5; Table 4      363

Figure 5. Occlusal surface of Iberomys huescarensis from the Atapuerca TD6 hominid-bearing level. All

specimens are from the right side except (a), (b), (c) and (i). (a) M1 (TD6T37I-16ATA94A-97). (b) M1 (TD6T40/41G-17ATA95A-12). (c) M1 (TD6T39H-17ATA94A-2). (d) M1 (TD6T40/41H- 16ATA95A-38). (e) M1 (TD6T40/41G-17ATA95A-1b). (f) M1 (TD6T40/41H-16ATA95A-26). (g) M1 (TD6T40/41G-17ATA95A-6). (h) M1 (TD6T40/41G-17ATA95A-7). (i) Left mandible with M1,M2, M3 (TD6T39H-17ATA94A-1). Scale bar represents 1 mm.

Table 4 Measurements of the M1 of Iberomys huescarensis from the Atapuerca TD6 hominid-bearing level

Mean S.D. S.E. n Minimum Maximum

L (mm) 2·696 0·109 0·030 13 2·539 2·906 W (mm) 0·954 0·038 0·010 14 0·894 1·018 A/L 50·011 1·980 0·549 13 46·814 53·043 B/W 3·416 1·532 0·409 14 0·750 7·109 C/W 19·612 4·506 1·204 14 12·360 26·265 D/E 30·603 4·129 2·920 2 27·684 33·523

Description Discussion Iberomys huescarensis is a small-sized vole The small size and ‘‘plesiomorphic Iberomys’’ with molars characterized by bucco-lingual morphology of I. huescarensis from TD3–6 asymmetry. M1 has a short and triangular- point to a primitive form of the Iberomys shaped AC. A more detailed description of lineage: I. brecciensis (Middle Pleistocene)–I. the species from the Trinchera Dolina lower cabrerae (extant Iberian vole of Cabrera). The levels (TD3–4) is given in Cuenca-Bescós primitive features of Iberomys can be found et al. (1995). The morphology and biometry in the gregaloid morphotype of Allophaiomys of I. huescarensis from TD6 are similar to nutiensis from Les Valerots and Monte Peglia those from TD3–4. (Chaline, 1972; van der Meulen, 1973). 364 . -´  ET AL.

We have also found primitive morpho- which separated from the Allophaiomys com- types of Iberomys in M. hintoni described plex group by the early Lower Pleistocene. from several localities by Rabeder (1981), Iberomys huescarensis is the first represen- and M. thenii from Untermassfeld and tative of the lineage, showing evident bucco- Neuleiningen 5, 15 (Maul, 1990a, 1996). lingual asymmetry, a triangular shape of the The study of the relationships between these anterior part and large B/W and C/W. This species is still in progress. The specimens species, as well as early I. brecciensis, shows described as ‘‘S. gregaloides’’ ATA B by Iberomys morphotypes that have been Agustí et al. (1993, Figures 3 and 4) are I. described as Pitymys species. The assign- huescarensis, judging from the asymmetry ment of these taxa to Terricola (=Pitymys)is and the triangular shape of AC. The based on the unclosed ‘‘pitymyan rhom- species A. chalinei from TD3-TD6T51 is bus’’, a plesiomorphic feature in several clearly distinguished from I. huescarensis by lineages of the Microtus group (i.e., Steno- its larger size and the ample ‘‘c’’ (see cranius). Iberomys brecciensis appears later in Cuenca-Bescós et al., 1995). Middle Pleistocene localities, such as Cúllar Previous studies consider I. brecciensis to Baza 1, Aragó, and Saint Estève Janson. co-occur with T. aguirrei in levels TD3–TD6 Stenocranius gregaloides Hinton, 1923; of the Trinchera Dolina section (Gil & Sesé, Figure 6(e) 1991). Gil (1997a) notes the synonymy of T. aguirrei with T. huescarensis ‘‘sensu Ruiz Bustos, 1988’’. We verify the presence of I. Description heuscarensis in levels TD3, 4, 5, 6, 8a and Only two specimens can be assigned to S. confirm the synonymy of I. huescarensis with gregaloides. The measurements are N: 2; L: P. aguirrei. We also include in this synonymy 2·76; W: 0·912–0·921; A/L: 51; B/W: the forms identified as I. brecciensis of 24–37; C/W: 19–23; D/E: 64. It is a small- Huescar 1 (Mazo et al., 1985; Ruiz Bustos, sized vole similar in morphology and size to 1988; Sesé, 1989) and levels TD3–TD6 T. arvalidens and M. seseae. The main differences are observed in the AC of the M and (Gil, 1997a). The typical I. brecciensis M1 is 1 larger, broader, and more complex and it the development of T4 in M2, which is does not co-occur with Mimomys. It is found closed in I. huescarensis and S. gregaloides in the Middle Pleistocene with Arvicola and open in M. seseae and T. arvalidens. The species (i.e. Saint Estève Janson, Aragó, M1 of S. gregaloides either has no T6 devel- Cúllar de Baza 1). In the Middle Pleistocene opment or it is reduced. The T4 and T5 are populations I. brecciensis has a relatively con- confluent and BRA3 is generally parallel fluent T4 and T5, which is a plesiomorphic to the LRA4. The BRA is shallow in S. retention from the Microtus ancestral form. gregaloides,asinI. huescarensis, and unlike This led Ruiz Bustos (1988) to recognize M. seseae. The lack of provergent BRA Pitymys cullarensis in Cullar de Baza 1 based is also a difference between T. seseae and S. gregaloides. on an M1 with confluent T4T5, but this is undoubtedly an I. brecciensis specimen. In conclusion, Iberomys evolved as a line- Discussion age with an asymmetrical and triangular AC This species is more frequent in the lower shape that could have its origin in M. nutien- levels, TD3–TD5 (Cuenca-Bescós et al., sis, M. thenii or M. hintoni. Nevertheless, the 1995). As in the preceding discussion of T. narrow neck between the AC and posterior arvalidens, the differences between S. grega- part of the ACC places I. huescarensis in a loides of Trinchera Dolina and S. gregaloides clade with the Microtus complex group from Middle Pleistocene localities could be      365

Figure 6. Other rodent taxa from the Atapuerca TD6 hominid-bearing level. Pliomys episcopalis left M1: (a) (TD6T40/41H-16ATA95A-17). (b) (TD6T40/41H-16ATA95A-18). Allophaiomys chalinei right M1: (c) (TD6T50G-16ATA96A-16). (d) (TD6T51G-17ATA96A-41). Stenocranius gregaloides right M1: (e) (TD6T33I-16ATA94A-162). Castor fiber: (f) Left lower cheek teeth with P4,M1,M2,M3 (TD6T38H- 18ATA94Ca31). Hystrix refossa: (g) Right mandible fragment (reversed) with P4 and M1 (TD6T42H- 18ATA95Hy86). Allocricetus sp.: (h) Right mandible with M1M2M3 (TD6T37H-16Ca). In a–e, h scale bar represents 1 mm. In f, g scale bar represents 1 cm. sufficient justification for the recognition of Allocricetus sp. [Figure 6(h)]—for the this form as a new chronospecies. The holo- moment, we cannot asign the Allocricetus type of S. gregaloides from West Runton is assemblages of TD6 to any of the known somewhat different from S. gregaloides species. The hamster from TD6 differs mor- from TD3–TD6, in its deeper LRA5. phologically from Allocricetus bursae in its

A/L increases from older to younger more simple M1 occlusal surface and a nar- assemblages, as shown by Rekovets & rower anterior portion of M1. Nevertheless, Nadachowski (1995). The species S. grega- it is similar in size to the late Middle Pleis- loides and T. arvalidens (sensu the West tocene A. bursae from Sima de los Huesos, Runton descriptions of Hinton, 1923) occur TD10 and TD11, and larger than A. bursae together at numerous localities in central from Cúllar de Baza (early Middle Pleis- and southern Europe dated to Cromerian tocene). We can infer that there is an s.l. (Sutcliffe & Kowalski, 1976; Rekovets & increase in the size of A. bursae during the Nadachowski, 1995). The primitive appear- Middle Pleistocene, since the early Middle ance of S. gregaloides of TD3–TD6 could Pleistocene hamsters are smaller than those indicate an early radiation of this species at from the late Early Pleistocene (Chaline, the end of the Early Pleistocene. 1972; Cuenca Bescós et al., 1997). Our work in progress shows that there are simi- Other rodent species in TD6 larities in size between Allocricetus from Pliomys episcopalis [Figure 6(a), (b)] Early Pleistocene sites such as TD6, and Allophaiomys chalinei [Figure 6(c), (d)], Hys- Allocricetus from late Middle Pleistocene trix refossa [Figure 6(g)], Marmota sp., sites, although there are slight morphologi- Micromys minutus—these rodents are found cal differences and a size gap between the only in the lower part of TD6. Early Pleistocene and early Middle Pleis- Castor fiber [Figure 6(f)] tocene assemblages. This size discontinuity 366 . -´  ET AL. can be interpreted either as the difference The species composition of the fauna between two separate species, or as the from TD6 (and the levels TD3–TD8a in disapperance of A. bursae at the end of the general) has analogues at other European Early Pleistocene and its replacement by a localities (Figure 8), and there seems to be new, small form of A. bursae from Asia at the no mixing of faunas (e.g., Jánossy, 1986; beginning of the Middle Pleistocene. For Agustí, 1990, 1991; Fejfar & Heinrich, these reasons, at the moment, we do not 1990; Markova, 1990; Maul, 1990a,b; assign Allocricetus from TD6 to any of the Repenning et al., 1990; von Koenigswald known species. & Tobien, 1990; Agustí & Moyá-Solá, 1991; Rekovets & Nadachowsky, 1995; van Kolfschoten & Turner, 1996). On the basis of the arvicolids, TD3–TD8a Biochronological implications of the can be referred to the Biharian. A remark- voleo and cricetids of TD6 able feature of the Biharian is the evolutionary radiation of Microtus which occurs very The distribution of the rodents in the quickly at the end of the Early Pleistocene, Trinchera Dolina Section is shown in Figure in the Late Biharian of Europe and Irving- 7. TD6–T36–43 has no A. chalinei, H. tonian II of North America (Dispersal Event refossa, M. minutus or Marmota. Also of 8ofRepenning et al., 1990)(Chaline, 1972; interest are the faunal differences between van der Meulen, 1973; Rabeder, 1981; TD8 lower and TD8 upper. There is a Markova, 1990; Repenning et al., 1990; major faunal change during the time of the Rekovets & Nadachowski, 1995). In Spain, formation of TD8 and an important faunis- the Early Pleistocene ends with the Late tic hiatus can be related with it. The Arvicola Biharian localities of the M. savini Zone of cantianus zone, which characterizes the Agustí & Moyá-Solá (1992). Middle Pleistocene of Europe, is not found The Biharian comprises the Microtus– in the Trinchera Dolina Section. Mimomys rodent superzone, which is The species M. savini, M. seseae, S. grega- divided into the Rodent Zone M. savini/M. loides, I. huescarensis and T. gregaloides from pusillus and the Rodent Zone M. savini TD6 have the potential to make an import- (Fejfar & Heinrich, 1990). The Biharian ant contribution to our knowledge of the mammal age is characterized by the concur- early radiation of the Microtus group. Some rent ranges of Mimomys and Microtus, with of these taxa are typical elements of the Mimomys savini the characteristic species of Biharian. The Late Biharian M. savini the Biharian, extending from the Early Pleis- Rodent Zone includes the paleomagnetic tocene to early Middle Pleistocene (between boundary Matuyama/Brunhes (780 ka, 1·65 and 0·45 Ma according to Fejfar & Cande & Kent, 1995) elsewhere in Europe, Heinrich, 1990). The beginning of the such as at Stranska Skala, Grace, Shamin, Biharian is the FAD of Microtus s.l. (1·4– Mahlis, Kärlich, Atapuerca TD7 (Biquand, 1·7 m.y.a.), which is also the FAD of M. 1974; Chaline, 1974; Koci & Sibrava, 1976; savini, while the end of the Biharian corre- Markova, 1990, 1998; Parés & Pérez sponds to the LAD of M. savini and FAD of González, 1995; van Kolfschoten & Turner, Arvicola (0·5–0·45 m.y.a.). The LAD of M. 1996). TD6 is below the TD7 level, which pusillus is supposed to be concurrent with records the paleomagnetic Matuyama/ Dispersal Event 8, and the radiation of Brunhes boundary in the Trinchera Microtus s.l. (S. gregaloides, M. hintoni, T. Dolina Section (Parés & Pérez González, arvalidens, M. nivalinus, M. nivaloides, M. 1995). malei, M. oeconomus, M. ratticepoides, and I.      

Figure 7. Stratigraphic distribution of rodents from the Trinchera Dolina Homo antecessor site. Shaded areas indicate levels without small mammals. Pointer in TD6 level marks the position of the Aurora Stratum (T36–43). Stratigraphic column modiﬁed from Parés & Pérez González (1998). 367 JAR. MATUYAMA BRUNHES POLARITY antecessor iue8 isrtgahcadcrnsrtgahcpsto ftemi oaiisctdi h etadthe and text the in cited localities main the of position chronostratigraphic and Biostratigraphic 8. Figure

21 20 19 18 17 13 12 Marine oxygen isotope stages CHRONOSTRA- 990 865 810 787 780 760 712 500 485 Ky TIGRAPHY rnhr oialvlTD6. level Dolina Trinchera

PLEISTOCENE

LOWER MIDDLE EMILIAN SICILIAN MILAZZIAN Mammal BIHARIAN Ages Continental Biochronology inEurope Northern BAVELIAN "Cromerian Complex" ELSTERIAN Europe British CROMERIAN ANGLIAN Isles

GALERIAN Italy Eastern TAMANIEN TIRASPOLIAN Europe North IRVINGTONIAN America Rodent Zones M. savini Arvicola cantiana M. pusillus Mimomys savini Biohorizons Arvicola cantianus

M. pusillus Mimomys savini TD 3–4 TD6 TD5 TD7 Cúllar deBaza1 West Runton Le Vallonnet Huescar 1 Western Europe Grâce Stranska Skala Hohensülzen Voigtstedt Kozi Grzbiet Kozi Kärlich G Eastern Europe Sites inEurope Middle & Prezletice Kärlich B

Isernia Homo

Shamin Slivia

368  .   ´  - TAL. ET      369 huescarensis) at the end of the Early Pleis- Localities and the rodent zone Mimomys tocene. The Microtus s.l. species of this age savini are characterized by great variability of the Localities with M. savini, but without small-

M1 occlusal surface. sized Mimomys, include the Trinchera Dolina Section (TD3–TD8a), West Run- The Mimomys savini/Mimomys pusillus ton, Huéscar 1, Casablanca-3, Chlum 6, Rodent Zone Westbury-sub-Mendip 1, Bave A, Kärlich This Zone is purportedly older than the M. E/F, Maasvlakte I, Zuurland 5, Kozi savini Rodent Zone. Localities with both M. Grzbiet, Tarko, Slivia, West Runton, pusillus and M. savini are Betfia V, Hohen Voigstedt, Beftia VII, Puerto Lobo1y4, Sülzen near Worms, Les Valerots 4, Orce 3, Rosia, Tasso, Grâce, and Stranska Skala Untermassfeld, Zalesiaki 1A/1, 8, 9, 10, (Maul, 1990b; Fejfar et al., 1998; Sardella et Razvodje, Tatinja draga, Podumci 1, al., 1998). The Matuyama/Brunhes bound- Neuleiningen 5, 15 (Maul, 1990a,b, 1996; ary is above TD6, Stranska Skala and Terzea, 1995; Paunovic & Rabeder, 1996; Grâce. At Stranska Skala, S. gregaloides is Agustí et al., 1997). The fauna from Unter- present, as in the case of TD3–6. Another massfeld has a paleomagnetic age determi- rodent association found immediately below nation that includes the Jaramillo event the Matuyama/Brunhes boundary is that (Kahlke, 1995). There are some localities from Mahlis, which has M. nivaloides (=M. where there is only M. pusillus, such as arvalinus) and M. ratticepoides, but with no Breitenberghohle, Deutsch Altenburg 2c1, occurrence of M. savini (Fuhrmann et al., 4b, 10, 16, 30a, Holstejn, Kamyk, Les 1977). Valerots 6, Orce 2, Sackdilling, Shamin, In Hungary, transitional faunas are repre- Zabia (Markova, 1990; Maul, 1990b; sented at Kovesvarad (Janossy, 1986), Paunovic & Rabeder, 1996). Among these where M. savini is found with a diversity of localities, only Shamin in the Don basin has Microtus s.l. species (7), some very modern a paleomagnetic geochronology, and is (such as M. gregalis) and others very primi- placed immediately below the Matuyama/ tive (such as M. hintoni). In the upper layers Brunhes boundary (Markova, 1990). Other of Tarko, A. bursae appears with A. can- localities have small-sized Mimomys, such as tianus at the Tarkoian (Middle Pleistocene), Beremend 16, 17 (either M. pusillus or M. where the Microtus diversity is low. The reidi; Janossy, 1995), or M. blanci without insectivore Beremendia fissidens is also M. savini, at Barranco de los Conejos present at Tarko in the lower layers where (Agustí et al., 1997). Other localities have M. savini occurs, but does not appear in the both M. savini and M. blanci: Maasvlakte I, upper layers with A. cantianus. The Paleo- Zuurland 5, Monte Peglia 11, 63, and lithic sites of Vertesszölös are similar to the Courterolles (Chaline, 1972; van der upper levels of Tarko, where A. cantianus Meulen, 1973; van Kolfschoten, 1990). The and A. bursae are both represented (Jánossy, small Mimomys oswaldoreigi is found to 1986). co-occur with M. savini at Fuente Nueva 3 In Poland, the succession of mammal (Martínez Navarro et al., 1997), in which assemblages of the Early Pleistocene ends the evolutionary stage of Allophaiomys bur- with Kozi Grzbiet (Nadachowski, 1985, gondiae and A. chalinei indicate an earlier age 1990), which is correlated with the than TD3–8a. The localities where small- Templomhegyian phase of the Biharian or sized Mimomys are found together with M. Mindel I/Mindel II interglacial. Samples of savini can be correlated with the M. bones from unit 2, analysed by the complex savini/M. pusillus Zone. fluoro-chloro-apatite and collagen method, 370 . -´  ET AL. give dates of 630–560 ka, and paleomag- entire distribution of the species (in Western netic samples are normally magnetized. Kozi Europe) because assemblages such as those Grzbiet shares with TD6 M. savini, M. of Hohen Sülzen and Kozi Grzbiet have arvalidens and S. gregaloides. Allocricetus bur- fewer individuals with enamel islets. Hohen sae is also found in Kozi Grzbiet, but we Sülzen is thought to be older than TD6 have not compared it with Allocricetus sp. based on its M. savini–M. pusillus associ- from TD6. On the basis of the evolutionary ation. However, the disappearance of M. stage of M. savini, S. gregaloides and T. pusillus (or small-sized Mimomys) provides arvalidens we propose that TD6 is older than an insufficient time marker, since small Kozi Grzbiet. Mimomys is found in localities above the The M. savini from the type locality of Matuyama/Brunhes boundary, such as at West Runton (450–500 ka) shows derived Kärlich E. features on the occlusal surface of M1 (i.e., The species M. seseae, found only in TD5 low number of specimens with enamel islets, and TD6, is morphologically close to M. reduction of roots, Figure 3). TD6 repre- nivaloides, M. malei, M. nivalinus and Tyr- sents an earlier stage in the M. savini rhenicola. M. seseae represents one of the lineage based on its high proportion of first species of a second radiation of the plesiomorphic characters. Microtus s.s. lineage at the end of the Early Pleistocene. The primitiveness of the assemblages of Conclusions T. arvalidens and S. gregaloides in TD3– The TD6 level in the Trinchera Dolina TD6 levels could also indicate an early Section at Atapuerca, Burgos, Spain, is a radiation of those species at the end of the hominid-bearing level associated with a rich Early Pleistocene. The primitive stage small mammal assemblage. Our study shows of M. savini, Iberomys, S. gregaloides, T. that the rodent assemblage differs from that arvalidens and Allocricetus indicate an reported by previous authors. Early Pleistocene age. On the basis of The arvicolids are represented by M. sa- Mimomys and the arvicolid assemblage in vini, I. huescarensis, M. seseae, T. arvalidens, general, TD6 is more modern than the S. gregaloides, P. episcopalis and A. chalinei. Jaramillo faunas from Untermassfeld and The rodent assemblage is typical of the Les Valerots and could be equivalent to end of the Early Pleistocene and beginning Shamin, early Tiraspolian, and early of the Middle Pleistocene. A more precise Cromerian faunas. age for this assemblage has been gained On the basis of the vole species from TD6 through the study of the evolutionary stages we propose that the hominids of the of some of the rodent species. M. savini Trinchera Dolina level 6 are of late Early evolutionary stages seem to be good chrono- Pleistocene age. This makes them younger logical markers when percentages of enamel than the humans that made the artefacts at islets and root development are considered. Les Valerots, Le Vallonnet, Fuente Nueva We propose that M. savini assemblages with 3, but older than the Middle Pleistocene more than 10% of M1 with enamel islets are humans of Cúllar de Baza 1, Boxgrove, older (i.e., Early Pleistocene) than assem- Mauer and Arago. In conclusion, micro- blages of M. savini with less than 10% (i.e., vertebrates (this work); macrovertebrates Middle Pleistocene). Further studies of M. [herbivores, (van der Made, 1999) and savini in well calibrated sections should pro- carnivores (Garcıá & Arsuaga, 199)] and vide a test of this hypothesis. This conclu- radiometric data (Falgue`res et al., 1999) sion may not apply equally throughout the provide confirmation of a late Early      371

Pleistocene age (780–857 ka) for the Homo lia) en el Pleistoceno Inferior del sur de Espanã. Acta antecessor bearing level TD6. Geol. Hisp. 16, 203–205. Arsuaga, J. L., Bermúdez de Castro, J. M., Carbonell, E., Cuenca-Bescós, G., Lorenzo, C., Pérez González, Acknowledgements A. & Rodríguez, X. P. (1996). Los yacimientos pleistocenos de la Sierra de Atapuerca. In (M. Financial support for research in Atapuerca Segura, I. Bustamante & T. Bardají, Eds) Itinerarios has been funded by Consejeria de Educación geológicos desde Alcalá de Henares, pp. 153–174. Alcalá de Henares: Sociedad Geológica de Espanã. y Cultura of Junta de Castilla y León, Direc- Arsuaga, J. L., Martínez, I., Gracia, A., Carretero, ción General de Investigación Científica y J. M., Lorenzo, C., García, N. & Ortega, A. I. Técnica (PB93-0066-C03, PB96-00-C01) (1997). Sima de los Huesos (Sierra de Atapuerca, Spain). The site. J. hum. Evol. 33, 109–127. and Principe de Asturias Foundation. The Bermúdez de Castro, J. M., Arsuaga, J. L., Carbonell, excavation team of ‘‘cata’’ TD recovered the E., Rosas, A., Martínez, I. & Mosquera, M. (1997). small-mammal bearing sediments, and the A hominid from the Lower Pleistocene of Atapuerca, Spain: possible ancestor to Neanderals and modern samples were processed by our team, humans. Science 276, 1392–1395. students and laboratory assistants. A. P. Biquand, D. (1974). Position chronologique de la très Currant provided drawings of the holo- haute nappe alluviale de Grâce (vallée de la types held in the Natural History Museum Somme) par rapport à la limite paléomagnétique Brunhes-Matuyama. Bull. Assoc. Fr. Quat. 41–42, (London). Lutz Maul, Nuria Garcia and 157–159. Leonid Rekovets provided useful comments. Brunet-Lecomte, P. (1989). La répartition géo- Lutz Maul showed the M. thenii collection graphique des campagnols souterrains (Arvicolidae, Rodentia) au Quaternaire en Europe occidentale. from Weimar to one of us (G.C.-B.) during Mammalia 53, 605–620. a stay in Germany (MEC, HA-97-22). Brunet-Lecomete, P. & Chaline, J. (1991). Morpho- Thanks are due to the three anonymous re- logical evolution and phylogenetic relationships of the European ground voles (Arvicolidae, Rodentia). viewers, the JHE staff and editor, Dr Terry Lethaia 24, 45–53. Harrison, who provided constructive and Cande, S. C. & Kent, D. V. (1995). Revised calibration useful editorial comments. of the geomagnetic polarity time scale for the late Cretaceous and Cenozoic. J. geophys. Res. 100, 6093–6095. References Carbonell, E., Bermúdez de Castro, J. M., Arsuaga, J. L., Díez, J. C., Rosas, A., Cuenca-Bescós, G., Sala, Agustí, J. (1990). The Plio-Pleistocene arvicolid suc- R., Mosquera, M. & Rodríguez, X. P. (1995). Lower cession in southern and eastern Spain. In (O. Fejfar Pleistocene hominids and artifacts from Atapuerca- & W.-D. Heinrich, Eds) International Symposium TD6 (Spain). Science 269, 826–829. Evolution, Phylogeny and Biostratigraphy of Arvicolids Carracedo, J. C., Heller, F., Soler, V. & Aguirre, E. (Rodentia, Mammalia), pp. 21–32. Munich: Pfeil. (1987). Estratigrafía magnética del complejo de Agustí, J. (1991). The Allophaiomys complex in Atapuerca. Determinación del límite Matuyama- Southern Europe. Geobios 25, 133–144. Bruhnes. In (E. Aguirre, E. Carbonell & J. M. Agustí, J., Castillo, C. & Galobart, A. (1993). Hetero- Bermúdez de Castro, Eds) El hombre fósil de Ibeas y el chronic evolution in the Late Pliocene–Early Pleis- Pleistoceno de la Sierra de Atapuerca, I, pp. 193–200. tocene arvicolids of the Mediterranean area. Quater- Soria: Junta de Castilla y León, Consejería de nary International 19, 51–56. Cultura y Bienestar Social. Agustí, J., Oms, O., Garcés, M. & Parés, J. M. (1997). Calibration of the Late Pliocene–Early Pleistocene Chaline, J. (1971). L’âge des Hominiens de la Caune transition in the continental beds of the Guadix-Baza de l’Arago à Tautavel (Pyrénées-Orientales), d’après Basin (Southern Spain). Quaternary International 40, l’étude des Rongeurs. C. r. Acad. Sci. Paris 272, 93–100. 1743–1746. Agustí, J. & Moyà–Solà, S. (1991). Les faunes de Chaline, J. (1972). Les rongeurs du Pléistocène Moyen mammifères du Pléistocène inférieur et moyen de et Supérieur de France. Cahiers de Paléontologie, l’Espagne: implications biostratigraphiques. 1–410. L’Anthropologie (Paris) 95, 753–764. Chaline, J. (1974). Les Rongeurs, l’âge et l’environne- Agustí, J. & Moyà-Solà, S. (1992). Mammalian disper- ment de la très haute terrasse de Grâce à Montières sal events in the Spanish Pleistocene. Cour. Forsch. (Somme). Bull. Assoc. Fr. Quat. 41–42, 151–157. Inst. Senck. 153, 69–77. Cuenca-Bescós, G., Canudo, J. I. & Laplana, C. Alcalde, G., Agustí, J. & Villalta, J. F. (1981). Un (1995). Los arvicólidos (Rodentia, Mammalia) de nuevo Allophaiomys (Arvicolidae, Rodentia, Mamma- los niveles inferiores de Gran Dolina (Pleistoceno 372 . -´  ET AL.

Inferior, Atapuerca, Burgos, Espanã). Rev. Esp. Pale- water Bed of West Runton, Norfolk. Ann. Mag. Nat. ont. 10, 202–218. Hist. 12, 541–542. Cuenca-Bescós, G., Laplana Conesa, C., Canudo, J. I. Jánossy, D. (1986). Pleistocene Vertebrate Faunas of & Arsuaga, J. L. (1997). Small mammals from Sima Hungary. New York: Elsevier. de los Huesos. J. hum. Evol. 33, 175–190. Jánossy, D. (1995). Lower Pleistocene vertebrate fau- Falgue`res, Ch., Bahain, J.-J., Yokoyama, Y., Arsuaga, nas from the Loc. 16 and 17 from Beremend (South- J. L., Bermu´dez de Castro, J. M.. Carbonell, E., ern Hungary). Fragmenta Min. Pal. 18, 91–102. Bischoff, J. & Dolo, J. M. (1999). Earliest Humans in Kahlke, R.-D. (1995). Untermassfeld near Europe: the age of TD6 Gran Dolina, Atapuerca, Meiningen—Early Pleistocene fauna. In (W. Spain. J. hum. Evol. 37, 343–342. Schirmer, Ed.) Quaternary Field Trips in Central Fejfar, O. & Heinrich, W.-D. (1990). Muroid rodent Europe, 2, pp. 730–732. Munich: INQUA. biochronology of the Neogene and Quaternary in Koci, A. & Sibrava, V. (1976). The Brunhes– Europe. In (E. H. Lindsay, V. Fahlbusch & P. Mein, Matuyama boundary at Central European localities. Eds) European Neogene Mammal Chronology, pp. 91– In (J. Easterbrook & V. Sibrava, Eds) Quaternary 118. NATO ASI A, 180. New York: Plenum Press. Glaciation in the Northern Hemisphere, Inter. Geol. Fejfar, O., Heinrich, W.-D. & Lindsay, E. H. (1988). Corr. Project 73/1/24, Report 3, pp. 135–160. Updating the Neogene Rodent biochronology in Washington D.C.: IUGS-UNESCO. Europe. In (Th. van Kolfschoten & P. L. Gibbard, Laplana, C. & Cuenca-Bescós, G. (1996). Mimomys Eds) The Dawn of the Quaternary, pp. 533–554. (Arvicolidae, Rodentia) del nivel TD6 de Trinchera Haarlem: Mededelingen Nederlands Instituut voor Dolina (Atapuerca, Burgos, Espanã; Pleistoceno Toegepaste Geowetenschappen. Inferior). In (T. Palacios & R. Gozalo, Eds) Comuni- Fernández Jalvo, Y. & Andrews, P. (1992). Small caciones XII Jornadas de Paleontología, pp. 75–76. mammal taphonomy of Gran Dolina, Atapuerca Badajoz: Sociedad Espanõla de Paleontología. (Burgos), Spain. J. Archaeol. Sci. 19, 407–428. Laplana, C. & Cuenca-Bescós, G. (1997). Los arvicó- Fuhrmann, R., Heinrich, W.-D., Mai, D. H. & lidos (Arvicolidae, Rodentia) del límite Pleistoceno Wiegank, F. (1977). Untersuchungen am präelster- Inferior-Medio en el relleno cárstico Trinchera zietlichen Löss von Malis (Bezirk Leipzig). Z. geol. Dolina (Sierra de Atapuerca, Burgos, Espanã). In (A. Wiss. 5, 717–743. Grandal-D’Anglade, Ed.) Libro de resúmenes y comu- nicaciones XII Jornadas de Paleontología, pp. 192–195. Garcıá, N. & Arsuaga, J. L. (1999). Carnivores from A. Corunã: Sociedad Espanõla de Paleontología. the Early Pleistocene hominid-bearing Trinchera Malez, M. & Rabeder, G. (1984). Fundmaterial von Dolina 6 (Sierra de Atapuerca, Spain). J. hum. Evol. kleinsäugern aus der Altpleistozäne spaltenfüllung 37, 415–430. Podumci I in NordDalmatien (Kroatien, Jugoslav- García-Perea, R. & Gisbert, J. (1997). Lista patrón de ien). Beitr. Paläont. Österr. 2, 439–510. los mamíferos de la Península Ibérica, Islas Baleares y Markova, A. K. (1990). Pleistocene microtheriofauna Canarias. Galemys 9, 1–38. of the European part of the USSR. In (O. Fejfar & Gil, E. (1996a). Precisiones bioestratigráficas y evolu- W.-D. Heinrich, Eds) International Symposium Evolu- tivas sobre Terricola subterraneus (Arvicolidae, Roden- tion, Phylogeny and Biostratigraphy of Arvicolids tia) en el Pleistoceno de la Sierra de Atapuerca (Rodentia, Mammalia), pp. 313–338. Munich: Pfeil. (Burgos, Espanã). In (G. Meléndez, M. F. Blasco & Markova, A. K. (1998). Early Pleistocene small mammal I. Pérez-Urresti, Eds) II Reunión de Tafonomía y faunas of Eastern Europe. In (T. van Kolfschoten & P. Fosilización, pp. 129–134. Zaragoza: Institución L. Gibbard, Eds) The Dawn of the Quaternary, pp. Fernando el Católico (CSIC). 313–326. Haarlem: Mededelingen Nederlands In- Gil, E. (1996b). Consideraciones bioestratigráficas stituut voor Toegepaste Geowetenschappen. sobre Allocricetus bursae (Cricetidae, Rodentia) en el Martínez Navarro, B., Turq, A., Agustí Ballester, J. & Pleistoceno inferior y medio de Atapuerca (Burgos, Oms, O. (1997). Fuente Nueva—3 (Orce, Granada, Espanã). Bol. Geol. y Min. 107, 45–52. Spain) and the first human occupation of Europe. Gil, E. (1996c). Terricola atapuerquensis (Arvicolidae, J. hum. Evol. 33, 611–620. Rodentia) nueva especie en el Pleistocene Medio de Maul, L. (1990a). Biharische Kleinäugerfunde von Atapuerca (Burgos, Espanã). Geogaceta 20, 250–252. Untermassfeld, Voigtstedt und Süssenborn und ihre Gil, E. (1997a). Bioestratigrafía y evolución de Microtus chronologische Stellung im Rahmen der biharischen brecciensis en los yacimientos del Pleistoceno de Ata- Micromammalia-Faunen Europas. Ph.D. Disser- puerca (Burgos, Espanã). Geogaceta 21, 113–116. tation, Humboldt-Universität Berlin. Gil, E. (1997b). Los micromamíferos del tránsito Maul, L. (1990b). Ûberblick über dei unterpleis- Pleistoceno inferior-medio de Atapuerca (Burgos, tozänen Kleinsäugerfaunen Europas. Quatärpaläon- Espanã). Implicaciones bioestratigráficas y evoluti- tologie 8, 153–191. vas. Geogaceta 22, 73–76. Maul, L. (1996). Biochronological implications of the arvi- Gil, E. & Sesé, C. (1991). Middle Pleistocene small colids (Mammalia: Rodentia) from the Pliocene and mammals from Atapuerca. Cahiers du Quaternaire 16, Pleistocene faunas of Neuleiningen (Rheinland-Pfalz, 337–347. southwest Germany). Acta Zool. Cracov. 39, 349–356. Hinton, M. A. C. (1923). Diagnoses of species of Mazo, A. V., Sesé, C., Ruiz-Bustos, A., Penã,J.A. Pitymys and Microtus occurring in the Upper Fresh- (1985). Geología y paleontología de los yacimientos      373

plio-pleistocenos de Huéscar (Depresión de Guadix- Sardella, R., Caloi, L., Di Stefano, G., Palombo, Baza, Granada). Estudios geol. 41, 647–693. M. R., Petronio, C., Abbazzi, L., Azzaroli, A., Musser, G. G. & Carleton, M. D. (1993). Family Ficcarelli, G., Mazza, P., Mezzabotta, C., Rook, L., Muridae. In (D. E. Wilson & D. M. Reeder, Eds) Torre, D., Argenti, P., Capasso Barbato, L., Mammal Species of the World (2nd edition), pp. 501– Kotsakis, T., Gliozzi, E., Masini, F. & Sala, B. (1998). 756. Washington: Smithsonian Institution Press. Mammal faunal turnover in Italy from the Middle Nadachowski, A. (1985). Biharian voles (Arvicolidae, Pliocene to the Holocene. In (T. van Kolfschoten Rodentia, Mammalia) from Kozi Grzbiet (Central &P.L.Gibbard,Eds)The Dawn of the Quaternary, Poland). Acta Zool. Cracov. 29, 13–28. pp. 499–512. Haarlem: Mededelingen Nederlands Nadachowski, A. (1990). Lower Pleistocene rodents of Instituut voor Toegepaste Geowetenschappen. Poland: faunal succession and biostratigraphy. Sesé, C. (1989). Micromamíferos de Mioceno, Plio- Quartärpaläontologie 8, 215–223. ceno y Pleistoceno de la cuenca de Guadix-Baza Parés, J. M. & Pérez-González, A. (1995). Paleomag- (Granada). Trabajos N-Q 11, 185–214. netic age for hominid fossils at Atapuerca Archaeo- Sesé, C. & Sevilla, P. (1996). Los micromamíferos del logical Site, Spain. Science 269, 830–832. Cuaternario peninsular espanõl: cronoestratigrafía e Parés, J. M. & Pérez-González, A. (1998). Contexto implicaciones bioestratigráficas. Rev. Esp. Paleont. estratigráfico y cronológico de Gran Dolina Mem. esp., 278–287. (Yacimiento de Atapuerca). In (E. Aguirre, Ed.) Storch, G., Franzen, J. L. & Malec, F. (1973). Die alt- Atapuerca y la evolución humana, pp. 51–60. Madrid: pleistozäne Säugerfauna (Mammalia) von Hohensülzen Fundación Ramón Areces. bei Worms. Senckenbergiana lethaea 54, 311–343. Pasquier, L. (1972). Etude d’une population de Mimo- Sutcliffe, A. J. & Kowalski, K. (1976). Pleistocene mys savini Hinton, 1910 (Arvicolidae, Rodentia) Rodents of the British Isles. Bull. Brit. Mus. (nat. provènant de l’Upper Freshwater bed (Quaternaire Hist.) 27, 1–147. ancien d’Anglaterre). Mammalia 36, 214–225. Terzea, E. (1989). Les arvicolidés (Rodentia, Mamma- Paunovic, M. & Rabeder, G. (1996). Die altpleistozänen lia) du Pléistocène Moyen de Chiscau-1 (Départ. de Kleinsäugerfaunen Razvodje und Tatinja draga in Bihor, Roumanie). Trav. Inst. Spéol. Emile Racovitza Süd-Kroatien. Beitr. Paläont. Österr. 21, 69–84. 28, 57–72. Rabeder, G. (1981). Die arvicoliden (Rodentia, Mam- Terzea, E. (1995). Mammalian events in the Quater- malia) aus dem älteren Pleistozän von Niederösterre- nary of Romania and correlations with the climatic ich. Beitr. Paläont. Österr. 8, 1–373. chronology of Western Europe. Acta Zool. Cracov. Rekovets, L. & Nadachowski, A. (1995). Pleistocene 38, 109–120. voles (Arvicolidae) of the Ukraine. Paleont. Evol. van der Made, J. (1999). Ungulates from Atapuerca 28–29, 145–245. TD6. J. hum. Evol. 37, 389–413. Repenning, C. A., Fejfar, O. & Heinrich, W.-D. van der Meulen, A. (1973). Middle Pleistocene smaller (1990). Arvicolid rodent biochronology of the North- mammals from the Monte Peglia (Orvieto, Italy), ern Hemisphere. In (O. Fejfar & W.-D. Heinrich, with special reference to the phylogeny of Microtus Eds) International Symposium Evolution, Phylogeny (Arvicolidae, Rodentia). Quaternaria 17, 1–144. and Biostratigraphy of Arvicolids (Rodentia, Mamma- van Kolfschoten, T. (1990). The evolution of the lia), pp. 385–418. Munich: Pfeil.. mammal fauna in the Netherlands and the Middle Rink, W. J., Schwarcz, H. P., Stuart, A. J., Lister, Rhine area (Western Germany) during the late Mid- A. M., Marseglia, E. & Brennan, B. J. (1996). ESR dle Pleistocene. Meded. Rijks. Geol. Dienst. 43, 1–69. dating of the type Cromerian Freshwater Bed at West van Kolfschoten, T. & Turner, E. (1996). Early Middle Runton, U.K. Quat. Sci. Rev. 15, 727–738. Pleistocene mammalian faunas from Kärlich and Rosas, A., Carbonell, E., Cuenca, G., García, N., Miesenheim I and their biostratigraphical implica- Fernández Jalvo, Y., van der Made, J., Ollé, A., tions. In (C. Turner, Ed.) The Early Middle Pleistocene Parés, J. M., Pérez González, A., Sánchez Marco, A., in Europe, pp. 227–253. Rotterdam: Balkema. Sánchez Chillón, B. & Valverdú, J. (1998). von Koenigswald, W. (1982). Zum Verständnis der Cronología, bioestratigrafía y paleoecología del Pleis- Morphologie der Wühlamusmolaren (Arvicolidae, toceno Medio de Galeria (Sierra de Atapuerca, Rodentia, Mammalia). Z. geol. Wiss. 10, 951–962. Espanã). Rev. Esp. Paleont. 13, 71–80. von Koenigswald, W. & van Kolfschoten, T. (1996). Ruiz Bustos, A. (1976). Estudio sistemático y ecológico The Mimomys-Arvicola boundary and the enamel sobre la fauna del Pleistoceno Medio en las Depre- thickness quotient (SDQ) of Arvicola as stratigraphic siones granadinas. El yacimiento de Cúllar de Baza I. markers in the Middle Pleistocene. In (C. Turner, Trabajos y Monografías del Departamento de Zoología de Ed.) The Early Middle Pleistocene in Europe, pp. 211– la Universidad de Granada 1. 226. Rotterdam: Balkema. Ruiz Bustos, A. (1988). Estudio sobre los arvicólidos von Koenigswald, W. & Tobien, H. (1990). Important cuaternarios. Paleomammalia 2, 1–89. arvicolid faunas from the Late Pliocene to Early Sala, B. (1983). La fauna del giacimiento di Isernia La Holocene in Western Germany (FRG). In (O. Fejfar Pineta. In (C. Pereton, C. Terzanic & M. Cremaschi, & W.-D. Heinrich, Eds) International Symposium Eds) Isernia La Pineta, un accampamento piú antico di Evolution, Phylogeny and Biostratigraphy of Arvicolids 700.000 ani, pp. 71–79. Bologna: Calderini. (Rodentia, Mammalia), pp. 231–254. Munich: Pfeil.