Journal of Vertebrate Paleontology 23(1):260±262, March 2003 q 2003 by the Society of Vertebrate Paleontology RAPID COMMUNICATION LATE PLEISTOCENE SURVIVAL OF THE SABER-TOOTHED CAT HOMOTHERIUM IN NORTHWESTERN EUROPE JELLE W. F. REUMER1, LORENZO ROOK2, KLAAS VAN DER BORG3, KLAAS POST1, DICK MOL1 and JOHN DE VOS4 1Natuurmuseum Rotterdam, P.O. Box 23452, 3001 KL Rotterdam, The Netherlands, [email protected]; 2UniversitaÁ di Firenze, Dipartimento di Scienze della Terra and Museo di Storia Naturale (Sezione Geologie e Paleontologia), Via G. la Pira 4, 50121 Firenze, Italy; 3Faculteit Natuur- en Sterrenkunde, R.J. van de Graaff Laboratorium, Universiteit Utrecht, P.O. Box 80.000, 3508 TA Utrecht, The Netherlands; 4Naturalis, National Museum of Natural History, P.O. Box 9517, 2300 RA Leiden, The Netherlands Saber-toothed cats (Machairodontinae) were wide-spread, top The age of the machairodontine dentary from the North Sea, predators of the Pleistocene terrestrial ecosystem. In Europe, about 28,000 yrBP, corresponds to the middle Weichselian Gla- they previously were known only from the early and middle cial, and, in the context of Dutch Pleistocene glacial chronol- Pleistocene (Adam, 1961; Turner and AntoÂn, 1997), and their ogy, more precisely just after the relatively warm Denekamp fossils are always extremely rare. The previous youngest record interstadial which lasted from about 32,000 to 29,000 yrBP was from Steinheim a/d Murr, Germany, where an upper canine (Dansgaard et al., 1993; Berendsen, 1998). It is thus by far the of Homotherium was found in a stratum dated to ca. 0.3 Ma youngest saber-toothed cat found in Europe. (Adam, 1961). In this paper, we describe a well-preserved den- tary of Homotherium latidens dated by 14C to be late Pleisto- DESCRIPTION cene, ca. 28,000 yrBP. The North Sea jaw (Fig. 1; Natuurmuseum Rotterdam cata- On March 16, 2000, the ®shing vessel UK33 trawled a partial logue number 02-011) is a right dentary with p3 and p4; the mandible of a large felid from a locality southeast of the Brown i1±i3, canine, and m1 are missing. Some of the missing teeth Bank in the North Sea. This speci®c area is only known to yield are reported to have been present at the time of recovery by late Pleistocene and early Holocene mammalian fossils (pers. the ®shing crew of the UK33, but they were lost during the six obs.). Therefore, the discovery struck our attention as being of weeks between this moment and the time of acquisition of the possible interest because no unambiguous late Pleistocene ma- fossil by one of us (K.P.). The original presence of these easily chairodontines have been reported from Europe. KurteÂn (1968) lost teeth indicates that the specimen was probably not re- mentioned localities in England that yielded saber-toothed cats worked. Several small, black, peaty grains found in some of of WuÈrmian age. Of these, Pin Hole is a Devensian (5Wei- the osseous trabeculae suggest that the jaw was most probably chselian or WuÈrmian) deposit, but without a direct date; Kent's deposited in peat. The bone of the ramus is a light brown color Cavern, the type locality of Homotherium latidens, is now con- with small black spots, and has a brittle appearance. Unlike sidered to be middle Pleistocene (Cromerian or pre-Hoxnian) North Sea fossils of a geologically older age, it does not pro- (Sutcliffe and Kowalski, 1976). duce a metallic sound when tapped upon with a hard object, We selected the posterior root of the p4 and some bone ma- thus indicating absence of permineralization (De Vos et al., terial from the mandibular ramus for radiocarbon analysis. The 1998). The apex of the angular process is missing, as are the analysis was performed on collagen extracted using the method medial half of the condylar process and the labial edges of the described by Longin (1971) at the Utrecht University AMS fa- alveoli of the incisiform teeth. The remainder of the jaw is well cility (Van der Borg et al., 1997). The extraction ef®ciency was preserved. 8 percent, which indicated fair preservation. The ®rst ages we The p3 is small and peg-like, with a crown length of 7.4 mm obtained were 31,300 6 400 yrBP from the root and 26,900 6 and width of 5.5 mm. The extreme wear on the p4 is indicative 400 yrBP from the mandible (Table 1). The difference between of an old-aged individual. A small, round opening (a ®stula) the two dates was too large. We suspected that it was caused below the m1 alveolus indicates the presence of chronic peri- by contamination from polymers (polyvinylacetate) used for su- apical periodontitis (Hillson, 1986) that affected the anterior per®cial conservation applied to the tooth in order to avoid root. Mandibular height directly posterior to m1 is 41.9 mm; ¯aking. We next reanalyzed the original material and obtained directly anterior to p4 it is 37.4 mm. The length from the pos- similar results (Table 1). Then we collected new samples from terior end of the mandibular condyle to the lower anterior edge both the tooth and the mandibular ramus and possible contam- of the symphysis is 165 mm. The coronoid process is short and ination was removed both by visual inspection and by using robust, it does not lean posteriorly, and it has a rounded, semi- strong solvents. The new samples then gave unambiguous re- circular, upper edge. Its height is 62.2 mm. The masseteric fossa sults: 28,100 6 220 yrBP for the tooth and 27,650 6 280 yrBP is deep and pocketed to such an extent that its bottom is hidden for the bone sample (Table 1). We therefore assume that the in lateral view (Fig. 1B). Its anterior edge reaches just below original root samples had indeed not been properly decontam- m1, its ventral edge is smooth. The upper sigmoid notch be- inated. All corresponding delta13C values (between 215.3 and tween the coronoid and condylar processes is relatively long 221.2 per mil) fall within the range for collagen of land animals and has a circular contour; the smallest height of the mandible (Beavon-Ath®eld and Sparks, 2001). at the place of the upper sigmoid notch is 41.5 mm. The lower 260 REUMER ET AL.ÐLATE PLEISTOCENE HOMOTHERIUM IN EUROPE 261 TABLE 1. Results of the six performed 14C dates from the North Sea Homotherium latidens dentary. The 14C ages from the second preparation are considered the most reliable. UtC Analyzed number Sample fraction Mass [mg] d13C [p´mil] 14C age [yr BP] 10456 tooth collagen 1st preparation 2.07 218.1 31,300 6 400 10999 tooth collagen 1st preparation 2.33 217.6 31,300 6 400 10908 mandible collagen 1st preparation 2.17 218.9 26,900 6 400 11064 mandible collagen 1st preparation 1.42 215.3 26,700 6 240 11000 tooth collagen 2nd preparation 0.60 221.2 28,100 6 220 11065 mandible collagen 2nd preparation 1.01 217.7 27,650 6 280 sigmoid notch between the condylar and angular processes is formula together exclude other felid genera such as Panthera; positioned in a nearly vertical direction. The angular process we furthermore follow Ficarelli (1979) and Turner and AntoÂn shows a rather ventral extension. (1997) in synonymizing (at least) all Middle Pleistocene Eu- Two large mental foramina are present on the labial side of ropean Homotherium under Homotherium latidens. Homother- the ramus (Fig. 1B), a posterior one located below p3 and an ium is known from Africa, Eurasia, and North America. It has anterior one below the diastema between the lower canine and not previously been described from Eurasian localities younger p3. This anterior mental foramen is located at the posterior end than middle Pleistocene (Adam, 1961; Turner and AntoÂn, of a deep circular pit about 10 mm in diameter. Anterior to this 1997), but in North America Homotherium serum is known to large foramen is a minuscule third opening. The anterior or have existed through the late Pleistocene (Rawn-Schatzinger, symphyseal face of the jaw is about 18 mm wide, and it bears 1992). With a 14C date of ca. 28,000 yrBP, the North Sea den- several small foramina. The lingual side of the ramus (Fig. 1A) tary is by far the youngest specimen of Homotherium reported is rather smooth, except for the angular process which has a from Europe and proves the survival of machairodontine cats strong scar at the place of attachment of the pterygoid muscle. into the late Pleistocene on this continent, as they did in North A wide, funnel-shaped mandibular foramen is situated ventrally America. It also indicates that H. latidens must have been able below the coronoid. The mandibular symphysis is high and to live in rather harsh climatic circumstances. rough, and stands nearly vertically. Turner and AntoÂn (1997) gave the chronologic range of Hom- otherium as 3.0 to 0.5 Ma. Yet Homotherium is present in lo- CONCLUSIONS AND DISCUSSION calities slightly younger than 0.5 Ma, e.g., Artenac (France), The North Sea jaw is identi®ed as belonging to Homotherium levels IV and V, dated at about 0.4 Ma (Beden et al., 1984; latidens Owen, 1846. The low coronoid process, the contact- Tournepiche, 1996). Homotherium sp. was also identi®ed from ¯ange for the upper canine, the high and ¯attened anterior face Steinheim a/d Murr by Adam (1961). Even though later pub- of the symphysis, the morphology of the teeth, and the dental lications on the Steinheim fauna (Adam et al., 1995; Von Ko- FIGURE 1.