The last Pleniglacial and the human settlement of central Europe: new information from the Rhineland site of Wiesbaden-Igstadt.

Street, Martin ; Terberger, Thomas

Introduction

It is generally considered that the extreme climatic deterioration of the late Weichselian Pleniglacial led to the complete desertion of northern Central Europe by humans (Bosinski 1992: 84; 1990: 131; Gamble 1986: 205) and, although this viewpoint has also been questioned (Weniger 1990: 173), it is at least certain that there was a considerable reduction of settlement intensity following the Gravettian represented in the region before the Pleniglacial (Hahn 1969; Bosinski et al. 1985; Bosinski 1995a; 1995b; 1995c; Conard et al. 1995).

The recolonization of northern Central Europe is usually interpreted as a relatively late expansion of Upper Magdalenian groups (of ultimately southwestern French origin) in a direct response to the sudden lateglacial climatic amelioration c. 13,000 BP (Bolus et al. 1988; Rensink 1993). This view has been somewhat modified by first attempts to calibrate radiocarbon ages, and to correlate them with climatic data from ice cores, deep-sea cores and varve sequences (Street et al. 1994), an adjustment which increases the age of Magdalenian samples by more than 2000 calendar years [ILLUSTRATION FOR FIGURE 1 OMITTED]. On this evidence, and supported by several consistent new series of radiocarbon dates, Upper Magdalenian groups were already established at the northern fringe of the Mittelgebirge (Upland Zone) before the lateglacial rise in temperature (Street et al. 1994; Housley et al. 1997).

Nevertheless, while it still seems certain that the occupation of northern latitudes indeed intensified in response to late glacial climatic amelioration, there is increasing evidence that regions peripheral to proposed Pleniglacial refugia were also occupied sporadically or at low intensity much earlier than hitherto supposed. This paper will suggest that there is evidence for occupation of the Rhineland before the Upper Magdalenian.

The age of Last Glacial events is currently expressed by several different (and often incompatible) relative and absolute dating methods. However, recent advances in the calibration of radiocarbon dating suggest that it will soon be possible directly to compare 14C ages and those obtained by other methods (such as thermoluminescence, counts of varves and ice cores and biostratigraphic evidence) and establish a standard time-scale for the Last Pleniglacial and the Late Glacial (Street et al. 1994; Lanting & van der Plicht 1996; Kitagawa & van der Plicht 1998; Joris & Weninger 1998). This paper will quote radiocarbon dates from archaeological contexts as uncalibrated years 14C BP. Until there is a consensus for a common calibration system, the presentation of the 'raw dates' avoids confusion and even the uncalibrated radiocarbon dates provide convincing evidence for continuity or hiatus of settlement in different regions of Europe.

Pleniglacial settlement in Europe - hiatus or continuity

The classic southwestern French Upper Palaeolithic cultural sequence is divided into the four major groups: Aurignacian, (Upper) Perigordian (= Gravettian), Solutrean and Magdalenian (with sub-divisions), complemented by further groups such as the Aurignacian V (Peyrony & Peyrony 1938), Protomagdalenian (= Perigordian VII) (Bordes & de Sonneville-Bordes 1966), ProtoSolutrean and Badegoulian (= Magdalenian 0/I). In this region Upper Palaeolithic development is characterized by continuity of settlement through the Pleniglacial [ILLUSTRATION FOR FIGURE 2 OMITTED].

In contrast to southwestern France, Central Europe is apparently characterized by a hiatus between the Gravettian occupation before the Pleniglacial and the lateglacial sites. The Gravettian in the Rhineland [ILLUSTRATION FOR FIGURE 3 OMITTED] is represented particularly at the sites of Koblenz-Metternich, Rhens, Mainz-Linsenberg and Sprendlingen (Hahn 1969; Bosinski et al. 1985; Bosinski 1995a; 1995b; 1995c; Conard et al. 1995), for which no 14C dates are available. For the absolute chronology we must turn to southern Germany. Almost without exception, the 14C dates for German Gravettian sites lie between 30,000 and 23,000 BP [ILLUSTRATION FOR FIGURE 4 OMITTED], which is in accord with results for the Central and Eastern European Gravettian generally [ILLUSTRATION FOR FIGURE 5 OMITTED]. German radiocarbon dates outside this range should be mentioned. AMS dates of 30,550[+ or -]550 BP (OxA-4601) and 31,100[+ or -]600 BP (OxA-4600) from layers III and IV at the south German Hohler Fels site are not associated with a diagnostic industry and cannot be assumed to date the Gravettian (Hahn 1995), while the large standard deviation of a conventional date of 21,160[+ or -]500 BP (H-5314-4899) from layer IIb and a second date of 23,100[+ or -]70 BP (Pta-2746) from the same layer suggest caution (Hahn 1995: 89). An industry from Bockstein-Torle layer VI dated to 20,400[+ or -]220 BP was originally described as Aurignacian (Hahn 1977: 85) but subsequently compared (Hahn 1977: 297) with the French Perigordian VI and VII (= Protomagdalenian, radiocarbon dated to 21,980[+ or -]250 BP at Laugerie Haute Est) and considered to be a late Gravettian containing 'aurignacoid' elements. A faunal assemblage from Aschenstein in northern German Lower Saxony has been radiocarbon dated to 18,820[+ or -]180 BP (Weniger 1990: 171), but the archaeological nature of the site is questionable.

On examination, the German classic Gravettian industries date to before the Pleniglacial, and no certain Gravettian industries are absolutely dated to later than 23,000 BP [ILLUSTRATION FOR FIGURE 4 OMITTED]. The absolute dating of the re-occupation of northern Central Europe by Magdalenian groups is still subject to some discussion. Radiocarbon dates appreciably before 13,000 BP are uncommon [ILLUSTRATION FOR FIGURE 4 OMITTED]. By contrast, numerous dates are clearly associated with Magdalenian presence between 13,000 and 12,000 BP in Belgium (Charles 1993; 1996), the Rhineland (Street et al. 1994; Street 1998a), Thuringia (Street 1998b; Street & Gaudzinski 1998; Street & Hock 1998; Housley et al. 1997) and southern Germany and Switzerland (Honeisen et al. 1993; Housley et al. 1997; Pasda 1998).

Earlier radiocarbon dates c. 16,200 BP have been discussed for the Belgian site Trou des Blaireaux, but a critical review of the dated material concludes that the dates are unconnected with human activity (Charles 1996: 15ff). Radiocarbon dates of 14,520[+ or -]240 BP and 15,490[+ or -]310 BP [ILLUSTRATION FOR FIGURE 4 OMITTED] suggest that the Middle Magdalenien a navettes (Allain et al. 1985) extended at least as far east as the Polish Maszycka Cave (Kozl/owski et al. 1995: 120). Nevertheless, at the Kniegrotte in Thuringia, a site with numerous triangular geometric microliths considered typical of the Middle Magdalenian, a consistent series of AMS dates [ILLUSTRATION FOR FIGURE 4 OMITTED] is only slightly older than dates for typically Upper Magdalenian assemblages (Street & Hock 1998). In the Rhineland, the Magdalenian IV age of the double burial at Bonn-Oberkassel (Verworn et al. 1914; 1919; Rosinski 1978; Street 1995) has been called into question (Schmitz & Thissen 1996; Street & Wuller 1998; Baales & Street 1998).

Older dates from southern German Magdalenian contexts can also be discussed here [ILLUSTRATION FOR FIGURE 4 OMITTED]. Although an age of 17,100_.150 14C BP for layer IIa at the Hohler Fels site is believed to be too old for the dated context, Weniger (1990: 171) suggested that the occupation layer can probably be dated to before 15,000 BP. A more recent study argues that the material is reworked and can all be assigned to the Upper Magdalenian (Hahn 1995: 86). At Munzingen, close to Freiburg (Pasda 1994), AMS dates obtained by the Oxford laboratory are very heterogeneous, with older ages potentially difficult to interpret (Housley et al. 1997). It is, however, not impossible that they indeed represent repeated occupations of the site from the earlier Magdalenian until the late glacial (Pasda 1998).

In summary, while indications for a Pleniglacial occupation of Germany (in the form of absolute dates from archaeological contexts) are very rare and often ambiguous, it is unclear whether all indications for settlement during this period should be rejected and the possibility remains that there was very late 'Gravettian' sensu lato survival or a (Magdalenian?) presence earlier than believed. In the light of the above arguments, the discovery of a new open-air site close to Wiesbaden provides new impetus for the discussion of the possibility of settlement in northern Central Europe at a period close to the Pleniglacial.

The site of Wiesbaden-Igstadt

Research history and stratigraphy In the 1980s patinated lithics and bone and tooth fragments on the surface of a field in the community of Igstadt to the east of Wiesbaden (Hessen) were recognized by A. Kratz as originating from a new Upper Palaeolithic site disturbed by ploughing. In 1991 one of the authors (TT) was commissioned by the Landesamt fur Denkmalpflege Hessen to excavate test trenches, leading to the discovery of an archaeological horizon (Terberger 1992; 1998a). Excavations during the summer of 1992 and 1995 identified evident features. The character of the lithic artefacts suggested that the site could be referred to the Aurignacian.

The Igstadt site lies on the south-facing slope of the valley of the Waschbach stream, which, flowing from the Taunus hills to the northwest, is diverted towards the Rhine by a low hill of Tertiary material. The site is located in the bend of the valley, on a slightly inclined loess slope (166 m OD) c. 100 m from the present Waschbach. By contrast, the opposite side of the valley is a steep slope with an outcrop of Tertiary Cyrenenmergel (shell marl).

The archaeological horizon lies within a calcareous loess in the first few centimetres below the approximately 30-cm thick plough soil. Projection of the level of the finds allows recognition of only one, vertically dispersed archaeological horizon, and the unity of the assemblage is not in doubt. It is not clear to what extent reworking has affected the find horizon, but archaeological features suggest that this is largely in situ. The stratigraphic sequence of the site is difficult to interpret. Drilling showed the existence of several metres of loess underneath the archaeological stratum, but there is no adequate basis for stratigraphically dating the artefacts. Core samples suggested that there had been heavy erosion of the loessic slope and distinct accumulation of sediment in the valley. A soil horizon c. 1.5 m below the archaeological layer is dated by TL to 28,500[+ or -]2500 BP (L. Zoller (Heidelberg) written comm.) and serves as a terminus post quem for the occupation of the site.

Character of the assemblage and site structures The Igstadt assemblage comprises a loose scatter of lithics, relatively poorly preserved faunal remains and other material (Terberger 1998a). Only in a few areas of the site is the concentration of finds greater [ILLUSTRATION FOR FIGURE 6 OMITTED]. Bone and tooth fragments of horse (Equus sp.), together with teeth and antler fragments of reindeer (Rangifer tarandus), some teeth of an ovicaprid (Capra ibex?) and a few carnivore remains have been determined, but the faunal material is of no biostratigraphic value for dating the horizon more precisely and seasonal data are still unknown.

Two features characterized only by darkening of the sediment and a small number of burned artefacts and other stones were interpreted as hearths. Numerous small burned bone fragments in the hearth fills are the only evidence of fuel. The absence of formal construction elements, such as a stone setting typical of the younger Upper Palaeolithic (Terberger 1998b), and the use of bones as fuel are reminiscent of the features at the Rhineland Aurignacian open site of Lommersum (Hahn 1989: 79), but are also known from sites of different age, e.g. Grubgraben (Brandtner 1996: 123). The distribution of bone charcoal recovered by wet-sieving confirms the observed evident structures. The denser distribution of finds around both features is interpreted as showing activities carried out around the fires. Two concentrations c. 4 m in diameter with more or less disturbed peripheral areas are reconstructed. The presence of a third, ephemerally used hearth at the centre of the site [ILLUSTRATION FOR FIGURE 6 OMITTED] is suggested by a greater concentration of burnt material here (Serangeli 1996).

The area immediately adjacent to the eastern hearth is characterized by mainly small finds with a few larger artefacts. To the south, the area 0.5-2.0 m from the hearth also contains other elements such as a larger cobble, part of a massive quartzite slab, various large bone fragments and larger lithic artefacts. Another aspect of activity at this hearth is shown by the presence of haematite and Tertiary molluscs used as raw material for jewellery. Bone fragments, thin stone slabs, hammerstones and lithic artefacts extend some 2 m to the east and up to 3 m to the south-southeast of the western hearth, which has evidence for more varied activities than the eastern hearth. Faunal remains show the processing of game, and haematite was distributed densely around the fireplace. Several small fragments of ivory are possibly waste from the production of jewellery.

The assemblage size and the distribution of the finds suggest a short occupation of small groups of hunter-gatherers. The central hearth represents a brief episode (a few hours?) with little evidence for activities. The eastern and western hearths represent a longer period of time (several days?) with evidence for more activities. Their regular placement and evidence that the different features 'respect' each other's space, suggest that they were mutually recognizable and at least quasi contemporary, if not actually in use at the same time.

A possible reason for the choice of site location may be an outcrop of Tertiary Cyrenenmergel on the opposite bank of the Waschbach from which fossil mollusc shells could be collected. Similar molluscs were found on the Gravettian sites of Mainz-Linsenberg and Sprendlingen [ILLUSTRATION FOR FIGURE 3 OMITTED], only a few kilometres away (Hahn 1969; Bosinski et al. 1985; Bosinski 1995a; 1995b). A further fragment of mollusc shell can probably be identified as Spisula solida (W. Rahle (Tubingen) written comm.), a species found in the Atlantic and the North Sea, and possibly indicates long-range exchange systems to the west.

The lithic assemblage

Lithic artefacts of Upper Palaeolithic type form the largest part of the Igstadt material [ILLUSTRATION FOR FIGURE 7 OMITTED]. The lack of chronologically significant organic artefacts means that typological dating of the Igstadt assemblage must be based on the lithic assemblage. The following discussion is based upon the observations of one of the authors (TT) and an analysis of the assemblage carried out by J. Serangeli (1996).

The assemblage (including both excavated finds and material recovered as surface finds) contains c. 2700 artefacts, which together weigh some 6.7 kg. If only the artefacts [greater than]1 cm are considered, their number falls to 1585. Chips and fragments [less than]1 cm make up 57.6% of the excavated material, a first indication that it represents a 'normally' balanced assemblage.

The lithic raw material is dominated by chalcedony, which, although its exact source has not yet been located, was probably obtained in the Mainz basin, not far (up to 20 km) from the site (cf. Floss 1994: 49 & 53). Pieces with cortex, numerous flakes and chips and hammerstones imply that primary manufacture of artefacts took place at the site. The origin of one artefact of opal is suggested to be the Siebengebirge close to Bonn (Serangeli 1996), which implies a raw material source more than 100 km to the northwest.

Controlled blade technology is demonstrated by a number of blade cores, mainly with single platforms, but there is no evidence for bipolar production of blades. Bladelet cores and bladelets are both very rare. More homogeneous specimens of raw material show that working [TABULAR DATA FOR TABLE 1 OMITTED] was skilled, several blades measuring more than 10 cm in length. Comparative analysis shows similarities in the technology of blade production with southern German Aurignacian assemblages (Serangeli 1996: 36).

The retouched lithic assemblage comprises c. 88 formal tools (e.g. scrapers, burins, pieces esquilles) with a total of 103 working ends. The proportion of combination tools is quite high (11.4%). End scrapers on blades are the commonest tool type and often show marked edge retouch. Among the scrapers are three carinated end scrapers manufactured on thick blanks and seven nosed scrapers. Burins are the second most common tool type in Igstadt and are dominated by burins on truncations. The multiple burins include two carinated burins. Among the other, far less common tool types are borers and splintered pieces, represented by isolated finds only.

The absence of backed retouched forms (backed bladelets and points) is a particular feature of the complex and clearly distinguishes the Igstadt assemblage from both Gravettian and Magdalenian industries. Specifically, there is no resemblance to the assemblages of the nearby Gravettian sites Sprendlingen and Mainz-Linsenberg which are characterized by very high proportions of backed tools of 44% and 34% respectively (Bosinski et al. 1985: 69). The carinated and nosed scrapers and carinated burins at Igstadt are typical elements of Aurignacian assemblages, and Serangeli (1996) points to similarities between the Igstadt assemblage and the industry at the Lower Rhineland Aurignacian open site of Lommersum (Hahn 1989). Nevertheless, differences can also be found, e.g. the high proportion of combination tools (11.4%) at Igstadt contrasted with Lommersum (2-6%), a feature more characteristic of later, Magdalenian industries.

Radiocarbon dating

Bone samples were dated by the Oxford Radiocarbon Accelerator Unit, the University of Zurich/Eidgenossische Technische Hochschule and by the Heidelberg radiocarbon laboratory (Pettitt et al. 1998).

On the basis of lithic technology and typology an Aurignacian context had been expected (Serangeli 1996; Terberger & Serangeli 1996). The majority of European Aurignacian assemblages date to between 32,000 and 26,000 BP. In Germany this range can be narrowed to between 32,000 and 29,000 BP (Dombek & Hahn 1989: 58ff), although there is increasing evidence of older material (Hahn 1995; Richter 1996; Uthmeier 1996). While all the radiocarbon dates obtained on the Igstadt assemblage (TABLE 1) are compatible with (younger than) the TL date of 28,500 BP for the palaeosol underlying the archaeological horizon, they are far too young to support an interpretation of an Aurignacian occupation of the site.

A surface find of postcranial bone gave a conventional 14C age of 13,940[+ or -]690 (Hd-15742/15440). An excavated tooth dated at Zurich gave a lateglacial age of 12,000[+ or -]90 (UZ-3767/ETH-13'379). The six Oxford AMS results on excavated bone and the AMS date of 17,210[+ or -]135 (UZ-3768/ETH-13'380) on bone from the University of Zurich are very similar, falling close to the Pleniglacial. Since the absolute dates are crucial for the interpretation of the site their reliability must be considered here. Various explanations for the discrepancy between typological and radiocarbon dating can be suggested:

1 All the radiocarbon dates are acceptable and identify several phases of occupation at the site, none of which are Aurignacian.

2 The site is Aurignacian, but all the radio-carbon dates are too young due to inaccuracy resulting from unknown causes (contamination?).

3 Some of the radiocarbon dates are reliable (dating a non-Aurignacian occupation), but others are inaccurate (contamination?).

The first possibility outlined above seems unlikely at Igstadt, where there are arguments for close chronological proximity of the dated material. The second and third possibilities (contamination) must be considered. The lateglacial age of 13,940[+ or -]690 on a surface find of postcranial bone can be rejected, since the collagen content of this conventional sample was too low for reliable dating (B. Kromer written comm.). The remaining seven Pleniglacial dates and one lateglacial date are not overtly linked with any methodological problem, but it might be asked if a combination of long-term use of agricultural fertilizers and shallow depth of sediment covering the archaeological site might have led to contamination of samples (with consequently younger results?). The date for the tooth fragment is appreciably younger than that for the seven samples of postcranial bone, and may indeed have been differently influenced. It seems, nevertheless, unlikely that uptake of contaminant would permit all seven bone samples to give such consistent results if they were indeed much older (Aurignacian). At the Central Rhineland Magdalenian sites Gonnersdorf and Andernach, mammoth ivory has also produced anomalously young results (R. Housley written comm.), which might suggest that teeth generally can be unsuitable for dating. At present the authors believe that the most parsimonious explanation for the Igstadt radiocarbon dates is the third possibility, that the seven AMS results for excavated postcranial bone represent the true age of the site and that the conventional date for a stray find (low collagan) and the AMS date on tooth can be regarded as unreliable and anomalous respectively. A Pleniglacial date for Wiesbaden-Igstadt?

To summarize the problem, whereas technological and typological analyses of the Igstadt lithic assemblage show that it is very different from Gravettian and Magdalenian industries and had suggested it could best be compared with early Upper Palaeolithic Aurignacian material (Serangeli 1996), absolute dating places the occupation of the site close to the last Pleniglacial. Because no lithic assemblages dated to this period are known from the Rhineland (Bosinski 1990: 65; 1992: 84) it is necessary to widen the search for potentially contemporary similar assemblages of 'Aurignacian' type.

At the site of Breitenbach in Thuringia, a lithic assemblage with carinated scrapers but lacking backed pieces is generally accepted as Aurignacian (Pohl 1958; Hahn 1977: 101ff; 1989: 57; Richter 1987). Nevertheless, mammoth remains at this site were dated by radiocarbon to the Pleniglacial (18,100[+ or -]200 BP) and Lateglacial (12,320[+ or -]200 BP) (Richter 1987: 92). The latter date was on a specimen of ivory (see above), but the former date on bone closely resembles the Igstadt AMS results and Breitenbach is now being redated by an extended series of samples. Looking further afield, the period between 20,000 and 16,000 BP is characterized in western Europe by industries of the Solutrean and the Badegoulian/initial Magdalenian. Although the Igstadt AMS dates resemble dates for the Upper Solutrean at Combe Sauniere (Gowlett et al. 1986) and Abri Fritsch (Chollet 1989) and the Middle Solutrean at the Grotte de la Salpetriere (Combier 1989), the assemblages are totally different, there being nothing similar to the Solutrean leaf points or surface retouch at Igstadt. Several French lithic assemblages dated to this period do, however, broadly resemble the Igstadt industry. They are characterized by the absence or low frequency of backed laminar pieces typical of the preceding Upper Perigordian/Gravettian and by the presence of flake tools, burins and carinated and nosed scrapers resembling those of the Aurignacian sensu stricto. They are referred variously to the Aurignacien terminal/aurignacoide and Magdalenien 0/1 = Badegoulien (Djindjian 1996; Bazile 1996). Such industries are, for example, the Aurignacien V assemblage at Laugerie-Haute (Peyrony & Peyrony 1938) and the 'couche inferieure' (Badegoulian) of Beauregard at the south of the Paris Basin (Schmider 1971).

The 'Aurignacien V/final' in western Europe, first recognized at Laugerie Haute (Peyrony & Peyrony 1938), lies stratigraphically between the Gravettian and the early Solutrean (Djindjian 1996: 45). It is dated by radiocarbon to between 21,500 and 20,500 BP in layer 30A at the Grotte de Salpetriere (Bazile 1996: 56f) and does not therefore appear to have a particularly close chronological relationship with the Igstadt assemblage.

The Badegoulien = Magdalenien 0/1 is attributed to the Lascaux intersradial c. 17,500 BP (Djindjian 1996: 45). At the Abri Fritsch, in the southern Paris Basin, an older phase (layer 6) and a younger phase (layer 3) of the Badegoulian are distinguished (Schmider 1990), the Magdalenien 0 dating to 17,980[+ or -]150 BP (Chollet 1989). The Magdalenien 0 is dated to 18,260[+ or -]360 BP at Laugerie-Haute Est (Delibrias et al. 1976) and to 18,300[+ or -]200/18,400[+ or -]200 BP at Cuzoul de Vers (Chollet 1989), while the Magdalenien 1 is dated to 17,490[+ or -]520,17,420[+ or -]390 and 17,320[+ or -]460 BP at Grotte de Pegourie and 16,800[+ or -]170 and 15,980[+ or -]150 BP at Cuzoul de Vers (Lorblanchet 1989). Radiocarbon dating [ILLUSTRATION FOR FIGURE 2 OMITTED] thus places the French Badegoulien slightly younger than the Igstadt results. A possibly contemporary lithic assemblage from the Kastelhohle-Nord in Switzerland has also been tentatively referred to the Badegoulian (Leesch 1993). The general similarities of the Badegoulian to the Igstadt assemblage can be noted. The Badegoulian contains 'Aurignacian' elements such as carinated and nosed scrapers, but backed forms are rare or absent. The raclette, a typical form ([greater than]25%) in the younger Badegoulian inventories but less common in older asssemblages (Schmider 1990: 49; Djindjian 1996: 48), is absent at Igstadt.

In eastern Central Europe [ILLUSTRATION FOR FIGURE 5 OMITTED] a hiatus appears between the Gravettian and Epigravettian occupations similar to that between the Gravettian and Magdalenian in Germany. Reliably dated Gravettian sites are before 22,000 BP and Epigravettian sites are younger than 20,000 BP. Chronologically close to WiesbadenIgstadt, and with apparent typo-/technological similarities, is the KS3 lithic industry from the Austrian site Grubgraben (Brandtner 1996), known for its Epigravettian industries with characteristic backed elements (Montet-White 1990). Thermoluminescence measurements of loess deposits date the occupation of Grubgraben to between 22,000 and 19,000 BP (Brandtner 1996: 123), while the archaeological horizons are dated by radiocarbon to between 16,800[+ or -]280 and 19,270[+ or -]80 BP, layer 3 being dated to 18,030[+ or -]270 BP (Damblon et al. 1996).

Several dated eastern Central European assemblages from the period close to the Pleniglacial are described as 'Epi-Aurignacien' (Oliva 1996: 70) or 'latest Aurignacian' (Kozlowski 1996). At the Moravian site of Stranska skala (Czech Republic), the Stranska skala IV assemblage contains almost no backed forms but does have carinated scrapers and burins, and is dated to 18,820[+ or -]120 and 17,740[+ or -]90 BP (Oliva 1996). An assemblage of some 500 artefacts from Langmannersdorf in Austria has been attributed to the Aurignacian but is dated on bone charcoal to 20,580[+ or -]170 BP (GrN-6659) and 20,260[+ or -]200 BP (GrN-6660). As was the case for the Bockstein Torle VI assemblage (see above), Hahn (1977:297) suggests that the Langmannersdorf A & B assemblage might not be Aurignacian, but late Gravettian or even something completely different.

It is thus apparent that several eastern Central European sites with absolute dates similar to those obtained at Wiesbaden-Igstadt are characterized by lithic assemblages which have 'Aurignacian' elements and low frequencies of backed elements compared to the typical Epigravettian industries. An explanation for the character of these sites might be functional. An alternative explanation would be that such 'aurignacoid' assemblages have a limited chronological distribution around the Pleniglacial, and are an eastern equivalent of the French Aurignacien V and Badegoulian industries.

In this search for chronological and technological/typological parallels to Igstadt, it must be pointed out that Badegoulian sites have not been found further to the northeast than the Paris Basin (Leroi-Gourhan et al. 1976: 1329ff; Schmider 1990: 50; Djindjian 1996: 47) and that the closest Badegoulian site to the Mainz Basin is thus over 450 km distant. The sites in Lower Austria and Moravia are even further away, at least 600 km from the Rhineland.

Summary

Whereas the Wiesbaden-Igstadt lithic assemblage was initially believed on typological grounds to represent an early phase of the Upper Palaeolithic (Aurignacian), a series of radiocarbon dates suggests that the occupation of the site actually lies close to the Pleniglacial (c. 19,000-17,000 BP), a period for which evidence of human presence has hitherto been lacking in the Rhineland. Comparison with other sites to the west (e.g. Abri Fritsch and Beauregard in the Paris Basin, sites of the Badegoulian generally), and the southeast (Grubgraben and, possibly, Langmannersdorf in Austria, Stranska skala in the Czech Republic) shows that several assemblages with dates similar to the Igstadt assemblage commonly contain typical 'Aurignacian' elements such as carinated scrapers and burins and are characterized by low frequencies of backed pieces. It therefore appears that the radiocarbon age of Igstadt can indeed be reconciled with the typological features of the assemblage, and that the Igstadt site was occupied close to the last Pleniglacial, a time when the Rhineland was, until now, believed to have been deserted by humans.

It is increasingly clear that at a time close to the last Pleniglacial both western and eastern Europe were occupied by groups with lithic assemblages lacking the characteristic backed elements common to the earlier Gravettian and the succeeding Magdalenian and Epigravettian industries. In view of the presence of the Badegoulian as far north as the Paris Basin and the contemporary occupation of sites in Moravia and Lower Austria [ILLUSTRATION FOR FIGURE 8 OMITTED], the existence of contemporary groups in the intermediate Rhineland region no longer appears problematic. Whether the Igstadt assemblage represents penetration of the Rhineland from one or other of these regions, or even shows that a corridor of contact existed between the two, cannot be answered, although an allochthonous mollusc shell identified as Spisula solida might suggest contacts to the west.

The hypothesis presented here should be tested by new series of dates for other potentially Pleniglacial sites (Breitenbach, Langmannersdoff), by comparison of the Igstadt assemblage with those we have suggested are contemporary [ILLUSTRATION FOR FIGURE 8 OMITTED] and, in the near future, by the comparison of the calibrated radiocarbon time-scale of occupation with data for short-term climatic change obtained from other archives. Even without this environmental information, the new evidence from Igstadt for a Pleniglacial human presence in the Rhineland means that it is no longer possible to regard northern Central Europe as an abandoned 'arctic desert' and it will be necessary to discuss the implications of this for models of the lateglacial expansion (most recently Housley et al. 1997).

Acknowledgements. The authors would like to thank the Oxford Radiocarbon Accelerator Unit (ORAU) and, in particular, Paul Pettitt for their continuing help in radiocarbon dating the Rhineland Palaeolithic. Bernd Kromar and Ludwig Zoller (Heidelberg), and the Zurich AMS laboratory also provided dates and useful comments, while Olaf Joris (Neuwied) and Bernhard Weninger (Cologne Radiocarbon Laboratory) generously allowed us to use their calibration data. Thanks are particularly due to Jordi Serangeli for allowing the authors access to his analysis of the Wiesbaden-Igstadt lithic assemblage.

References

ALLAIN, J., R. DESBROSSE, J.K. KOZLOWSKI & A. RIGAUD. 1985. Le Magdalenien a navettes, Gallia Prehistoire 28: 87-124.

BAALES, M. & M. STREET. 1998. Late Palaeolithic backed-point assemblages in the northern Rhineland: current research and changing views, Notae Praehistoricae 18: 77-92.

BAZILE, F. 1996. La question de 'L'Aurignacien Terminal' en Languedoc, in Palma di Cesnola & Montet-White (ed.): 55-68.

BOLUS, M. G. BOSINSKI, H. FLOSS, H. HUSMANN, U. STODIEK, M. STREET, T. TERBERGER & D. WINTER. 1988. Le sequence Bolling-Dryas III en Rhenanie, in M. Otto (ed.), De la Loire a l'Oder. Les civilisations Paleolithique final dans le nord-ouest europeen 2: 475-510. Oxford: British Archaological Reports. International series 444(2).

BORDES, F. & D. DE SONNEVILLE-BORDES. 1966. Protomagdalenien ou Perigordien VII? L'Anthropologie 70: 113-22.

BOSINSKI, G. 1978. Oberkassel, krfr. St. Bonn, in S. Veil (ed.), Alt- und mittelsteinzeitliche Fundplatze des Rheinlandes: 118, 125. Cologne and Bonn: Rheinland Verlag. Kunst und Altertum am Rhein 81.

1990. Homo sapiens. L'histoire des chasseurs du Paleolithique superieur en Europe (40,000-10,000 av. J.-C.). Paris: Editions Errance.

1992. Eiszeitjager im Neuwieder Beckon. Archaologie des Eiszeitalters am Mittelrhein. Koblenz: Archaologische Denkmalpflege Amt. Archaologie an Mittelrehin und Mosel 1.

1995a. Sprendlingen, in Bosinski et al. (ed.): 872-3.

1995b. Mainz-Linsenberg, in Bosinski et al. (ed.): 873-5.

1995c. Rhens, in Bosinski et al. (ed.): 882.

BOSINSKI, G., H. BOSINSKI, K. BRUNNACKER, E. CZIESLA, K.P. LANSER, F.O. NEUFFER, J. PREUSS, H. SPOERER, W. TILLMANNS & B. URBAN. 1985. Sprendlingen. Ein Fundplatz des mittleren Jungpalaolithikums in Rheinhessen, Jahrbuch des Romisch-Germanischen Zentralmuseums 32: 5-91.

BOSINSKI, G., M. STREET & M. BAALES (ed.). 1995. The Palaeolithic and Mesolithic of the Rhineland, Quaternary Field trips in Central Europe 15(2): 829-99. Munich: Dr. Friedrich Pfeil. 14th INQUA-Kongress Berlin.

BRANDTNER, F. 1996. Zur geostratigraphischen und kulturellen Zuordnung der Palaolithstation Grubgraben bei Kammern, NO, in J. Svoboda (ed.), Paleolithic in the Middle Danube region. Anniversary volume to Bohuslav Klima: 121-46. (Svasek 5). Brno: Institute of Archaeology.

CHARLES, R. 1993. Towards a new chronology for the Belgian Lateglacial: Recent radiocarbon dates from the Oxford AMS system, Notae Praehistoricae 12: 59-62.

1996. Back into the North: the radiocarbon evidence for the human recolonisation of the northwestern Ardennos after the Last Glacial Maximum, Proceedings of the Prehistoric Society 62: 1-18.

CHOLLET, A. 1989. Le Magdalenien zero, in Mohen (ed.): 296-7.

COMBIER, J. 1989. Gravettien et Solutreen dans la vallee du Rhone, in Mohen (ed.): 286-89.

CONARD, N.J., G. BOSINSKI & D. ADLER. 1995. Koblenz-Metternich, in Bosinski et al. (ed.): 882-5. DAMBLON, F., P. HAESAERTS & J. VAN DER PLICHT. 1996. New datings and considerations on the chronology of Upper Palaeolithic sites in the Great Eurasiatic plain, Prehistoire Europeenne 9: 177-231.

DE LUMLEY, H. (ed.). 1976. La Prehistoire Francaise I(2): Lee civilisations paleolithiques et mesolithiques de la France. Paris: CNRS.

DELIBRIAS, G., G. DELIBRIAS, M.-TH. GUILLIER, J. EVIN, J. THOMMERET & Y. THOMMERET. 1976. Datations absolues des depots quaternaires et des sites prehistoriques par la methode Carbone 14, in de Lumley (ed.): 1499-1514.

DJINDJIAN, F. 1996. Los industries aurignacoides en Aquitaine entre 25,000 BP et 15,990 BP, in Palma di Cesnola & Montet-White (ed.): 41-54.

DOMBEK, G. & J. HAHN. 1989. Radiocarbondatierungen - Die chronologische Stellung von Lommersum im Aurignacien nach 14C-Daten, in Hahn (1989): 54-62.

FLOSS, H. 1994. Rohmaterialversorgung im Palaolithikum des Mittelrheingebietes. Mainz & Bonn: Romisch-Germanisches Zentralmuseum. Monograph 21.

GAMBLE, C. 1986. The Palaeolithic settlement of Europe. Cambridge: Cambridge University Press.

GOWLETT, J.A.J., R.E.M. HEDGES, I.A. LAW & C. PERRY. 1986. Radiocarbon dates from the Oxford AMS system. Archaeometry Datelist 4, Archaeometry 28: 206-21.

HAHN, J. 1969. Gravettien-Freilandstationen im Rheinland: Mainz-Linsenberg, Koblenz-Metternich und Khans, Bonner Jahrbucher 169: 44-87.

1977. Aurignacien. Das altere Jungpalaolithikum in Mittelund Osteuropa. Cologne & Vienna: B. Fundamanta A9.

1989. Genese und Funktion der jungpalaolithischen Freilandstation: Lommersum im Rheinland. Cologne: Rheinland Verlag. Rheinische Ausgrabungen 29.

1995. Nane Beschleuniger-14C-Daten zum Jungpalaolithikum in Sudwestdeutschland, Eiszeitalter und Gegenwart 45: 86-92.

HEDGES, R.E.M., P.B. PETTITT, C, BRONK RAMSEY & G.J. VAN KLINKEN (ed.). 1998. Radiocarbon dates from the Oxford AMS system: Archaeometry Datelist 25, Archaeometry 40: 227-39.

HONEISEN, M., D. LEESCH & J.-M. LE TENSORER (ed.). 1993. Das spate Jungpalaolithikum/Le Paleolithique superieur recent, in SPM 1, Die Schweiz vom Palaolithikum bis zum fruhen Mittelalter/La Suisse du Paleolithique a l'aube du Moyen-Age: 153-202. Basle: Schweizerische Gesellschaft fur Urund Fruhgeschichte/Societe suisse de prehistoire et d'archeologie.

HOUSLEY, R.A., C.S. GAMBLE, M. STREET, & P. PETTITT. 1997. Radiocarbon evidence for the Lateglacial human recolonisation of Northern Europe, Proceedings of the Prehistoric Society 63: 25-54,

JORIS, O. & B. WENINGER. 1998. Extension of the 14C-calibration curve to c. 40,000 cal BC by synchronizing Greenland 18O/16O ice-care records and North Atlantic foraminifera profiles. A comparison with U/Th-coral data, Radiocarbon 40: 495-504.

KITAGAWA, H. & J. VAN DER PLICHT. 1998. Atmospheric radiocarbon calibration to 45,000 yr BP: Late Glacial fluctuations and cosmogenic isotope production, Science 279: 1187-90.

KOZLOWSKI, J.K. 1996. The latest Aurignacian and 'Aurignacoid' elements in the Epigravettian of the Carpathian Basin, in Palma di Cesnola & Montet-White (ed.): 83-98.

KOZLOWSKI, S.K., E. SACHSE-KOZLOWSKA, A. MARSHACK, T. MADEYSKA, H. KIERDORE, A, LASOTA-MOSKALEWSKA, G. JAKUBOWSKI, M. WINIARSKA-KABACINSKA, Z. KAPICA & A. WIERCINSKI. 1995. Maszycka Cave, a Magdalenian Site in Southern Poland, Jahrbuch des Romisch-Germanischen Zentralmuseums 40 1993(1995): 115-252.

LANTING. J.N. & J. VAN DER PLICHT. 1996. De 14C-Chronologie van de Nederlandse Pre- en Protohistorie, I: Laat-Paleolithicum, Paloeohistoria 37/38: 71-125.

LEESCH, D. 1993. Cadre chronologique et facies industriels, in Honeisen et al. (ed.): 153-64.

LEROI-GOURHAN, A., M. BREZILLON & B. SCHMIDER. 1976. Lee civilisations du Paleolithique superieur dans le centre et le sudest du Bassin-Parisien, in de Lumley (ed.): 1315-38.

LORBLANCHET, M. 1989. Caracteres originaux du Magdalenien du Quercy, in J.P. Rigaud (ed.). Le Magdalenien en Europe. Acres du Colloque de Mayence 1987: 239-52. Liege: Universite de Liege. ERAUL 38.

MOHEN, J.-P. (ed.). 1989. Le Temps de la Prehistoire 1. Paris: Edition Archeologia.

MONTET-WHITE, A. (ed.) 1990. The Epigravettian site of Gruhgraben, Lower Austria: the 1986 and 1987 excavations. Liege: Universite de Liege. ERAUL 40.

OLIVA, M. 1996. Epiaurignacien en Moravie: Je changement economique pendant le deuxieme interpleniglaciare wurmien, in Palma di Cesnola & Montet-White (ed.): 6982.

PALMA DI CESNOLA, A. & A. MONTET-WHITE (ed.). 1998. The late Aurignacian, in A. Palma di Cesnola (ed.), The Upper Palaeolithic: 11-150. Forli: ABACO Edizioni. Colloquia of the XIII International Congress of the UISPP (Forli) 6.

PASDA, C. 1994. Das Magdalenien in der Freiburger Bucht. Stuttgart: Konrad Theiss. Materialhefte zur Archaologie in Baden-Wurttemberg 25.

1998. Der Beginn des Magdaleniens in Mitteleuropa, Archaologisches Korrespondenzblatt 28: 175-90.

PETTITT, P.B., M. STREET & T. TERBERGER. 1998. Comments on the dating of Wiesbaden-Igstadt. Archaeometry Datelist 26, Archaeometry 40: 443-4.

PEYRONY, D. & E. PEYRONY. 1938. Laugerie-Hautepres des Eyzies (Dordogne). Paris: Masson. Archives de l'Institut de Paleontologie Humaine, Memoire 19.

POHL, G. 1958. Die jungpalaolithische Siedlung Breitenbach, Kr. Zeitz, und ihre bisherige Beurteilung, Jahresschrift for mitteldeutsche Vorgeschichte 41/42: 178-90.

RENSINK, E. 1993. Moving into the north: Magdalenian occupation and exploitation of the Loess landscapes of northwestern Europe. Ph.D thesis, University of Leiden.

RICHTER, J. 1987. Jungpalaolithische Funde aus Breitenbach/Kr. Zeitz im Germanischen Nationalmuseum Nurnberg, Quartar 37/37: 63-96.

1996. 'Out of Africa II'. Die Theorie uber die Einwanderung des modernen Menschen nach Europa auf dem archologischen Prufstand, Archaologischen Informationen 19(1 & 2): 67-73.

SCHMIDER, B. 1971. Lee industries lithiques du Paleolithique superieur en Ile-de-France, Suppl. Gallia Prehistoire 6.

1989. Le Magdalenien dans le centre du Bassin Parisien, in Mohen (ed.): 307-9.

1990. The last Pleniglacial in the Paris Basin (22,500-17,000 BC), in Soffer & Gamble (ed.): 41-53,

SCHMITZ, R.-W. & J. THISSEN. 1996. Aktuelle Untersuchungen zum endpleistozanen/fruhholozanen Fundplatz Bonn-Oberkassel. Ein Vorbericht, Archaologische Informationen 19: 197-203.

SERANGELI, J. 1996. Die Steinartefakte der Freilandstation Wiesbaden-Igstadt und ihre Verteilung. Magisterarbeit, Tubingen.

SOFFER, O. & C. GAMBLE (ed.). 1990. The world at 18,000 BP 1: High latitudes. London: Unwin Hyman.

STREET, M. 1995. Bonn-Oberkassel, in Basinski et al. (ed.): 940-41.

1998a. Comments on the dating of Gonnersdorf, in Hedges et al. (ed.): 231.

1998b. Comments on the dating of the Teufelsbrucke, in Hedges et al. (ed.): 233.

STREET, M., M. BAALES & B. WENINGER. 1994. Absolute Chronologie des spaten Palaolithikums und Fruhmesolithikums im nordlichen Rheinland, Archaologisches Korrespondenzblatt 24: 1-28.

STREET, M. & S. GAUDZINSKI. 1998. Comments on the dating of Oelknitz, in Hedges et al. (ed.): 233-4.

STREET, M. & C. HOCK. 1998. Comments on the dating of the Kniegrotte, in Hedges et al. (ed.): 231-2.

STREET, M. & B. WULLER. 1998. Comments on the dating of Oberkassel, in Hedges et al. (ed.): 229-31.

TERBERGER, T. 1992. Ein neuentdeckter jungpalaolithischer Fundplatz bei Wiesbaden-Igstadt, Archaologisches Korrespondenzblatt 22: 161-75.

1998a. Siedlungsspuren zwischen 20,000 und 16,000 BP am Mittelrhein? Vorbericht zu den Ausgrabungen an der jungpalaolithischen Freilandstation Wiesbaden-Igstadt, Hessen. Germanio 76: 403-37.

1998b. Die Siedlungsbefunde des Magdalenien-Fundplatzes Gonnersdorf, Konzentrationen III und IV. Stuttgart: Franz Steiner. Der Magdalenien-Fundplatz Gonnersdorf 6.

TERBERGER, T. & J. SERANGELI. 1996. Wiesbaden-Igstadt, Germany: Site structure and economy of an Aurignacian open air site, in UISPP Congres de Forli 1996, Commission Paleolithique superieur. Abstracts 181-2.

UTHMEIER, T. 1996. Ein bemerkenswertes fruhes Inventar des Aurignacien van der Freilandfundstelle 'Keilberg-Kirche' bei Regensburg, Archaologisehes Korrespondenzblatt 26: 233-48.

VERWORN, M., R. BONNET & G. STEINMANN. 1914. Diluviale Menschenfunde in Obercassel bei Bonn, Die Naturwissenschaften 2, Heft 27: 645-50.

1919. Der diluviale Menschenfunde in Obercassel bei Bonn. Wiesbaden: J.F. Bergmann.

VIALET, P. 1989. Le Magdalenien dans lee Bassins de l'Allier et de la Loire, in Mohen (ed.): 320-22.

WENIGER, G.-C. 1990. Germany at 18,000 BP, in Softer & Gamble (ed.): 171-92.