272 Documenta Praehistorica XLIV (2017) Introduction The starting point for some considerations concern- ing the reliability of 14 C dates are three new studies concerning the absolute chronology of the Central European early and middle Neolithic, all three of which are based mainly or entirely on dates on bone collagen. Two are from the north-western Carpathian basin (Tasi≤ et al. 2015; Jukacs et al. 2016). The first produced an absolute chronology for the stratigra- phic sequence at Vin≠a-Belo Brdo; the second dated the contact between Vin≠a A and earliest LBK (eLBK) and between Ra∫iste and the later LBK at Szederkény from bone collagen dates only and imbedded the re- sults in a wider context. The Vin≠a chronology poses no obvious problems. In contrast, the parallelisation of all four different ceramic styles at Szederkény, although pairwise spa- tially separated, contradicts the usual archaeological chronologies, postulating the (not necessarily direct) sequence Vin≠a A-Ra∫i∏te on the one hand, and the eLBK-later LBK on the other. No explanation is given, but three much later bone dates for Ra∫i∏te from Discrepancies between archaeological and 14 C-based chronologies> problems and possible solutions Hans-Christoph Strien Institut für Altertumswissenschaften, Vor- und frühgeschichtliche Archäologie, Johannes Gutenberg Universität Mainz, Mainz, DE strien@uni-mainz.de ABSTRACT – 14 C dating of bone collagen is believed to produce the most reliable absolute dates for the Central European Early Neolithic, as the selection of bones in anatomical context minimises ta- phonomic problems. In contrast, a comparison of three newly published local or regional chronolo- gical models as well as a comparison of several series of dates from bone collagen, charcoal and cereals highlights problems probably caused by diagenetic influences, especially on collagen. There- fore, at least the checking of bone collagen 14 C dates against charcoal or cereal dates from the same contexts seems to be indispensable. IZVLE∞EK – Za 14 C datiranje kostnega kolagena velja, da omogo≠a najbolj zanesljive absolutne datu- me za srednjeevropski zgodnji neolitik, saj izbor kosti iz anatomsko pravilnih kontekstov zmanj∏a te∫ave s tafonomijo. V ≠lanku opozarjamo na te∫ave, ki lahko nastanejo predvsem pri vplivu diage- neza na kolagen, kar prika∫emo s primerjavo treh novo objavljenih lokalnih ali regionalnih krono- lo∏kih modelov kot tudi s primerjavo ∏tevilnih datumov, ki so bili pridobljeni iz kostnega kolagena, oglja in ∫it. Posledi≠no opozarjamo na nujnost preverjanja 14 C datumov, ki so bili pridobljeni iz kost- nega kolagena, z datumi, pridobljenimi iz oglja ali ∫it iz istih kontekstov. KEY WORDS – 14 C; taphonomy; diagenesis; Central Europe; Early Neolithic KLJU∞NE BESEDE – 14 C; tafonomija; diageneza; Srednja Evropa; zgodnji neolitik Neskladnosti med arheolo[kimi in 14 C kronologijami> te/ave in mo/ne re[itve DOI> 10.4312\dp.44.16 Discrepancies between archaeological and 14 C-based hronologies> problems and possible solutions 273 high even when assuming the fast assimilation of the autochthonous hunter-gatherer population), there would have been 20 sites in the first year of eLBK expansion in order to arrive at the number of 400 sites 150 years later – one third of the number of known eLBK sites in the region – and this means 100 sites only in the above-mentioned areas of Germany. Even with a mean of 5 contemporaneous houses per site (too low in the case of sites like Vaihingen with up to 45 contemporaneous houses) and the lowest estimate of 6 inhabitants per house, this would mean that 3000 persons left the Balaton-Vienna area in the first year for south-western and central Germany, and we have to add the thousands of immigrants be- tween Bavaria and Volhynia, an obviously unrealis- tic number. This calculation still ignores the additio- nal population growth needed for the colonisation of the regions west of the Rhine only two genera- tions after the start of the expansion, e.g., with 36 known sites in Lower Alsace at about 5250 cal BC. Even this simplified demographic estimate shows quite clearly that the three chronologies do not fit together. A start of eLBK expansion not later than 5500 cal BC would fit much better with the date of 5300 cal BC for the start of Flomborn W of the Rhine, and the date of 5350 cal BC for the start of eLBK expansion, with a start of Flomborn in Alsace not earlier than 5150 cal BC. But which chronology is wrong (perhaps all of them?), and even more im- portantly, why? At first glance there is no evident rea- son to doubt any of them. Dendrochronology, 14 C and relative chronology Typological correlations with other absolute dates show that the dates for the Alsation chronology are entirely plausible. The pottery of the earliest Phase IIB is quite similar to the pottery from the Flomborn cemetery. The same holds for Vedrovice; so the start of this cemetery can be dated to immediately after the end of eLBK, where it too is placed by a new cor- respondence analysis of eLBK ceramic inventories (Strien 2017). The 14 C dates from Vedrovice suggest a start around 5300 cal BC (Pettitt, Hedges 2008), perhaps slightly earlier. So the 14 C age accords well with the Alsatian model. Moreover, this date is back- ed by dendrochronology. A well was constructed at Leipzig-Plaußig, a site with no traces of eLBK, in 5259/58 den BC (Friederich [no year].3; Herbig et al. 2013.268). Since the well has not been fixed in any relative chronology, this date can serve only as terminus ante quem for the start of the early LBK in Western Saxony. This date confirms that the early Ivandvor are cited, without the consequences being debated. This problem will be discussed later. The site was fixed within the chronology of the Central European eLBK by using dates from varying materi- als, partially from dubious contexts (such as Eilsle- ben), some others dating from later contexts (early, but not the earliest LBK (Vedrovice, Kleinhaders- dorf)). The beginning of the eLBK and its expansion is dated to about 5350 cal BC. The resulting relative date – slightly earlier than Vin≠a A in Vin≠a-Belo Br- do – is not completely contrary to archaeological ex- pectations. The ‘formative phase’, dated to 5500– 5350 cal BC exclusively by charcoal, may be ignor- ed at this point, even more so as its existence as a separate chronological unit is questionable (Strien 2017). The last study comprised bone collagen dates for the whole of the lower Alsatian Early and Mid- dle Neolithic sequence (Denaire et al. 2017). The be- ginning of LBK settlement in this region was prob- ably around 5300 cal BC or the first half of 53 rd cen- tury. If we accept these three 14 C-chronologies, we have to face the problem that the start of the earliest LBK in trans-Danubian Hungary dates only 50 years ear- lier than the start of the early LBK and the settle- ment of Alsace and the Rhine-Meuse region, and only 150 years earlier than the end of the early LBK (Flomborn) in the West (even taking the upper end of the 95% probability range for the end of phase IIC in Alsace). This is not so much a chronological as a demographical problem. It means that, starting from the relatively restricted area between Lake Ba- laton and Vienna, the whole area between the Rhine and western Ukraine was densely occupied within only 50 years. Even this time span is probably an overestimate, as on the one hand it starts with the Bíňa phase, which until now has not been attested outside NW Hungary and SW Slovakia, while on the other hand the earliest pits in Alsace in the corres- pondence analysis are not 14 C dated (Denaire et al. 2017.Fig. 5), so the beginning of the expansion of the eLBK may date later than 5350 cal BC, and the start of the Flomborn earlier than 5300 cal BC. Demographic considerations In south-western Germany (southern Hessen, Lower Franconia, the Neckar Valley) 200 eLBK sites are known, and 100 in Central Germany. In the Neckar Valley 194 Flomborn sites are known, but as only half of all known LBK sites are dated, we have to double this number. Even calculating with an unrea- listic population growth rate of 2% per year (too Hans-Christoph Strien 274 LBK (Flomborn) settlement of Lower Alsace as well as the Vedrovice cemetery started not later than in the first half of the 53 rd century cal BC. The pit alignment at Herxheim can be dated with 95.4% probability somewhere in the time span 5208– 5053 cal BC (mean of 8 dates, calculated with Ox- cal: 6161±10 BP, Haack 2015.61; this and all fur- ther calibrations done with OxCal 4.3 online). This fits quite well with the Alsatian chronology; the pits produced ceramics of late Palatinate style – phase V (Jeunesse et al. 2009) and respectively V/VI (Lindig 2002) – which is most probablly contemporaneous with the Lower Alsatian phases IVa2 and IVb (Le- franc 2007.Tab. 14), starting after 5185 cal BC and ending before 5020 cal BC (95% probability, De- naire et al. 2017). And Herxheim enables a further (indirect) dendrochronological control for the 14 C age with the rich πarka finds. πarka itself is contem- poraneous with the house generations (HG) XII and XIII of the Rhine-Meuse chronology, shown by im- ports at Elsloo (Strien 2010.502; in a new seriation (Strien, work in progress) pit 655 can be dated to HG XIII, and pit complex 653 to HG XII and XIII) and Cologne-Mengenich (Schröter 1970.Abb. 6, 10; dated to HG XII). The second well at Kückhoven, built about 5057 den BC, is dated by decorated ce- ramics from the destruction horizon of the first well (built in 5089 den BC) and from the construction pit, which date both very close together between HG XII and XIII. The well at Altscherbitz, built in 5099 den BC and probably filled in after a relatively short time span – along with other ceramic of the region- al late LBK style – contained two pots with a pasted decoration known from πarka contexts (Tegel et al. 2012). Together, the two wells show an age of about 5080–5030 den BC for πarka. The 14 C date for Herx- heim and consequently for the Alsatian phase IVb may be slightly too old, but the 95% range of both at least overlaps with the date interpolated from dendrochronology and typochronology, so we should accept the Alsatian 14 C chronology at least as a good approximation for the both start and the end of the dated part of the Lower Alsace LBK sequence, with only phase V remaining undated. The same cannot be said for the start of the Middle Neolithic sequence; a date in the 48 th century for Hinkelstein is clearly too late, probably by about 200 years. Most chronologies based on typology publish- ed in the last 30 years postulate at least an overlap between Hinkelstein and late western LBK, often even a similar end date for both (e.g., Jeunesse, Strien 2009; Spatz 1996; Lefranc 2007; Stehli 1994). Sometimes a gradual succession LBK/Hinkel- stein can be mapped (Strien 2013). This is confirm- ed beyond doubt by the sequence of πarka/early Stroked Pottery Culture (SBK) in the East, demon- strated at the household level at Dresden-Prohlis (Link 2014); πarka being contemporaneous with late, but not the latest western LBK, and SBK II/III with Hinkelstein, no time is left for a gap between the LBK and Hinkelstein in the West. This may be the case on a regional scale, but not generally, and not to the extent of up to two centuries. A closer look at some local series may shed light on the question. Differences between dates from char- coal, cereals and bone collagen can regularly be seen, with dates on bone usually being the youngest group. Even a simple comparison of unweighted means of uncalibrated dates shows this effect quite clearly: ● Rottenburg: cereals (2 dates) 6155 BP, bone (12 dates) 5965 BP (Bofinger 2005.112–120, Tab. 15) – difference 190 years (cereals/bone); ● Bruchenbrücken (only dates earlier than 5900 BP in order to exclude late LBK intrusions; Stäuble 2005.Anhang D): cereals (4 dates) 6215 BP, char- coal (3 dates) 6377 BP, bone (4 dates) 6144 BP – difference 233 years (charcoal/bone) resp. 71 years (cereals/bone). Only if all dates later than 6200 BP were excluded, the differences would for the most part disappear, but in this case, the whole bone collagen series from Strögen has to be rejected; ● Strögen pit 5 (Lenneis, Lüning 2001): cereals (1 date) 6506 BP, charcoal (2 dates) 6313 BP, bone (4 dates; VERA-896 excluded as an outlier) 6152 BP – 161 years (charcoal/bone) resp. 354 years (cereals/bone); ● Schwanfeld (Stäuble 2005.Anhang D; only dates earlier than 5900 BP in order to exclude late LBK intrusions and only dates with a standard devia- tion of ≤100 years): cereals (1 date) 6380 BP, char- coal (3 dates) 6493 BP, bone (3 dates) 6321 BP– 172 years (charcoal/bone) resp. 59 years (cereals/ bone). A systematic deviation of bone collagen dates to a younger age is clearly visible; even if there were an overlap between the 95.4% ranges of floral and fau- nal dates (which is often but not always the case, Fig. 1) for each single site, any kind of group calibra- tion would produce a later age for the bone collagen series. Discrepancies between archaeological and 14 C-based hronologies> problems and possible solutions 275 below, are all eLBK (or La Hoguette), as far as can be recognised. So there can be no reasonable doubt that the dated bones belong to the eLBK sherds. The last example is the Hinkelstein and Großgartach cemetery at Trebur (Spatz 1999.213–217). Here we see a difference even between conventional and AMS dates on bone collagen. One outlier excluded nine graves are dated with both methods. The mean dif- ference between the uncalibrated dates is 134 years, with the AMS dates in all nine cases being earlier. Nevertheless, even the AMS dates are considered to be too late (Spatz 1999.216–217). Observations like these are far from new, and only confirm earlier results (Denaire 2009). A difference of 160–230 years (uncalibrated) cannot be explain- ed only by the old-wood effect – even the trunk of a 300-year-old oak consists mainly of wood younger than 200 years, not to speak of the branches. This is shown, too, by the dates on cereals from the above- mentioned eLBK sites, which are clearly older than the dates on bone from the same sites. And finally, the old wiggle matching calibration of charcoal dates from Merzbach valley fits quite well with the Alsa- tian chronology, with a start of Flomborn at around 5300 cal BC (Stehli 1989). These dates from the 1970s are not very reliable, but at least they do not show any important old wood effect, compared to the Alsatian dates on bone collagen. When bone dates accord with dendrochronologically controlled age, there seems to be no serious difference to char- coal dates. Too young, but sometimes too old: the case of Ra∫i∏te/Sopot IB-II A look at the Carpathian basin shows that there are rare cases when bone collagen dates are too old. Starting with Ivandvor, a Sopot IB-II/Ra∫iste-site with the usual pattern (dates from Jakucs et al. 2016): Fig. 1. Comparison of calibrated ages of charcoal (two dates), cereals (one date), and bone collagen (four dates) from Strögen pit 5. The Middle Neolithic series from Western Germany show the same effect, resulting in a systematic diffe- rence between the Lower Rhine area (only charcoal dates) and SW Germany (mainly bone dates; Friede- rich 2011.Tab. 89–90), resulting in a strange chro- nology: the Middle Neolithic in SW Germany starts with Hinkelstein and in Rhenania only slightly later with late (!) Großgartach. The complete Großgartach- Planig-Friedberg-Rössen-Bischheim sequence, present in both regions, should be delayed in SW Germany by at least 150 years, but follow the same stylistic development in both regions (Friederich 2011.397– 406, Fig. 347). Here the 14 C-based chronology evi- dently contradicts the archaeological facts and each plausible reconstruction of cultural history, and it does not become much better by the third, Alsatian chronology for the same time span (Fig. 2). A special case is the eLBK site at Rottenburg (Bofin- ger 2005). The two dates on cereals and the two earliest dates on bone – both identified as outliers from the rest – are in a range one can argue as per- haps correct (although more probably too young, the lower limit of the 95.4% probability range being 5290 cal BC for the cereals, and 5305 cal BC for the bones), the 10 remaining bone dates are five to six centuries too young, 4846–4722 cal BC (95.4% range) – almost exactly the assumed time span for Hinkelstein in Alsace, which is thought to be too late. The context is quite clear: the bones came from the long pits of the houses, below the preserved low- er part of the original soil horizon. On top was a layer mixed from the disturbed upper part of the soil and material that went down the hill slope, dated to the late pre-Roman period, and a Roman road (Kind 2005.255–262, Fig. 1–5: units 4 and 3). In this layer, only three small sherds with eroded surfaces of late LBK/Hinkelstein date were identified (along with much more eLBK material and some Ro- man pieces; data: own analysis), the other, Neolithic, sherds, almost 18 000 of them, most from the layers Hans-Christoph Strien 276 charcoal (3 dates) 5947 BP, bone (3 dates) 5717 BP – 230 years older date for charcoal. These dates are not only contra- dictory in themselves, but do not fit with other 14 C series for the Ra∫i∏te/Sopot IB-II context: the site was dated by János Ja- kucs et al. (2016.35) to “the se- cond quarter of the fifth millen- nium”. The Sopot graveyard at Alsónyék is dated somewhere between 5200–5020 cal BC (95% probability for the beginning) and 4850–4680 cal BC (95% probability for the end) (Oross et al. 2016.166), so it dates main- ly to the first quarter of the fifth millennium. At Szederkenyi, Ra- ∫i∏te is dated “between the 54 th and 52 nd centuries cal BC” (Ja- kucs et al. 2016.32). If all these series are accepted, Sopot IB-II – usually taken as transitional between LBK and Lengyel (Oross et al. 2016. 151– 152) and synchronised with late LBK (πarka) and early SBK in Bohemia (overview in Gleser 2012) – would span more than 600 years, ranging from the start of eLBK expansion, about 5350– 5300 cal BC, at least until 4700 cal BC or even later, far beyond the latest estimate for the end of late LBK. On the one hand Sopot IB-II is connected typo- chronologically to Vin≠a C1 (phase 6 at Vin≠a-Belo Brdo; Oross et al. 2016.158–159) – dated somewhere between 5040 and 4855 cal BC (Tasi≤ et al. 2016. Tab. 8), almost exactly the minimum time span given for Alsónyék – and on the other hand at Szederkény by 14 C to Vin≠a A1 resp. A1-A3, although without much contact between the two groups except for some isolated Ra∫i∏te sherds in Vin≠a context, but not vice versa (Jakucs et al. 2016.30), which makes later intrusions more probable than contact. In ad- dition, Ra∫i∏te is associated with later LBK, and Vin- ≠a A with eLBK. So there are archaeological argu- ments that clearly contradict the very early dates from Szederkény, and no other 14 C dates support them, as the authors admit (Jakucs et al. 2016. 35). The dates on charcoal from Ivandvor seem to match with archaeology, the two dates from feature SU 407 dating to 5021–4842 cal BC (95.4%) – the minimum range for Alsónyék – and the single date from fea- ture SU 194 to 4766–4499 cal BC, slightly too late. A date about or shortly after 5000 BC is most plau- sible, comparing relative chronology (πarka/early SBK) with dendrochronology (πarka dating around 5050 den BC, as shown above). So we have two sets of dates (Alsónyék, Ivandvor charcoal) that fit with a dendrochronological age derived from relative chronology, as well as with a 14 C series from a site with the same relative age (Vin≠a-Belo Brdo), one set much too old (Szederkény) and one too late (Ivandvor bone collagen). Why bone collagen? In the three series with very large differences from expected date – Rottenburg, the Aisne Valley and Trebur (conventional dates) – the common element is geology: all these sites are located on fluviatile se- diments, with either thin or no loess coverage. We can even add two more examples. At Ulm-Eggingen Fig. 2. Chronological table 5500–4500 cal BC, based on 14 C. Discrepancies between archaeological and 14 C-based hronologies> problems and possible solutions 277 (sited on tertiary sands with a thin loess coverage), the charcoal dates (Dombek 1989) are clearly too young: house 12, archaeologically dated shortly after the end of Flomborn, and so some generations earli- er than the Kückhoven wells, was dated in the range 4946–4743 cal BC (95.4% probability, one outlier ex- cluded), and the waterlogged wood from the Mohel- nice well, with a questionable dendrochronological date around 5600 den BC, but in any case not later than 5400 den BC (Schmidt, Gruhle 2003.56), dated to 5364–5007 cal BC (99.7% range; 3 dates from Neustupný, Vesely 1977). The possibility of hydro- geological influences on the diagenesis of charred and uncharred organic matter, especially bone colla- gen, should be taken into consideration. The diffe- rence between conventional and AMS dates from Tre- bur strengthens this idea, as later contaminations had probably been removed much better from ma- terial for the AMS dates than from the bigger pieces for conventional dates (Spatz 1999.217). It is further reinforced by a fourth date from the Mohelnice well (Schmidt, Gruhle 2003.56; KN-4339: 6580±75 BP, not included in the data list by Jakucs et al. 2016), which has an uncalibrated age 300 14 C-years older than the mean of the other dates. It is the only one that fits with the dendrochronological age (87.7% probability: 5641–5461 cal BC, 7.7%: 5451–5377 cal BC), so we may assume that for this sample the re- moval of contaminations was much more successful than for the samples dated in the 1970s, and that contamination by water may play an important role even in an undisturbed context. Szederkény is an interesting case. The dates for the eastern and central part of the site with Vin≠a A and eLBK ceramics are probably too young. The dates for the western part with Ra∫i∏te and later LBK finds, on the other hand, are clearly too old as shown above. So far, this is the only case in which the bone dates are too old. Here we possibly see different dia- genetic influences on bone collagen in parts of the same site separated by a distance of little more than 100m. A closer look at geological and hydrological properties of the different parts of the site might teach us something about the nature of the diagene- tic processes involved. Apart from hydrogeological factors, no other reasons are clearly visible. But as there seem to be regional as well as local differen- ces, it is not very probable that only one factor lies behind the problem. If it is really hydrology, we may even assume the possibili- ty that the lowest levels of a tell – near to groundwater level – are af- fected, whereas the upper layers yield correct dates. The start of Vin≠a-Belo Brdo might be dated too late, and the later phases from Vin≠a B or C onwards might not (at least the date for C1 corresponds to the expected date; Fig. 3). We may conclude that dates on bone collagen are often too young, but not in every case; the best ex- amples of series that accord with other dates are the Alsatian LBK and Rössen to Epirössen sequen- ces. Compared with the Alsatian chronology, the same holds for the Vedrovice cemetery and the pit alignment at Herxheim. Most other bone collagen series from the Central European Neolithic in the second half of the 6 th and the 5 th millennia BC are too young, in individual cases even up to 500 years. Fig. 3. Chronological table 5500–4900 BC, based on relative chrono- logies, archaeological correlations and dendrochronologically dated wells. In bold: date in accordance with 14 C age. Hans-Christoph Strien 278 How to solve the problem Of course, a large-scale study on causes and possi- ble remedies is not only desirable, but even neces- sary. As long as this study is lacking, we have to look for ways to minimise the risks. The first step is that dates on collagen must always be checked against charcoal or cereal dates from the same context; fi- nancial constraints concerning the number of dates are not an excuse for methodological deficits. In the case of differences of more than 50 BP-years be- tween dates from faunal and floral carbon, the bone dates at least should be questioned, if not rejected. From a purely taphonomic point of view, dates from bones in anatomical contact are the best choice, but in the case of settlement pits, it may be preferable to use botanic material for 14 C dating, making tapho- nomic problems improbable by taking three or more samples from different depths of a column. The high- er cost of dating more samples is the price for the time saved collecting the samples compared to the time-consuming identification of bones in anatomical contexts and make misleading results caused by dia- genetic processes less probable, without excluding them, as shown by the examples of Ulm-Eggingen and, probably, Rottenburg. Archaeology first Finally, this paper shows that even the most refined, technically advanced methods used by the natural sciences can provide only an additional source of in- formation. Our subject being archaeology, we always have to check the external data to see whether they fit with our archaeological data, and if not, why not. In the last few years there has been a tendency to- wards an unfounded reliance on 14 C dating alone, culminating in the conclusion concerning the four late and ten extremely late dates from Rottenburg: “presumably there is a so-far unresolved problem with the detection of later activity” (Jakucs et al. 2016.53), thus a priori excluding any problem with the dates themselves. As Rottenburg is probably the site with the best preservation of all known eLBK sites, with an unbroken layer of eLBK sherds above the 14 C-dated pits (Kind 2005.Abb. 3–5), thorough- ly excavated and documented by an experienced ex- cavator, this shows an impressingly uncritical atti- tude towards 14 C and distrust in archaeology. Scien- tific measurements and statistical analyses of them are not necessarily more reliable or accurate than our own methods, even if they are sometimes assumed to achieve a level of obvious pseudo-accuracy, as in the conclusions on Hinkelstein in Alsace: “the appear- ance of Hinkelstein ceramics in Alsace was swift: taking 1–115 years (95% probability…)”; “Hinkel- stein ceramics were probably used in this area for one or two generations” (Denaire et al. 2017.43); and “the re-establishment of ways of doing things may have been quite rapid (1–40 years at 68% probability)” (Denaire et al. 2017.71) – conclusions derived from two graves and one pit included in a seriation, only the two graves being dated by 14 C. How can we get any useful information about time spans and rapid or slow developments from a data base such as this; from a data base which was not checked to ascertain whether the suspected prob- lems with collagen dates (Denaire 2009) were real, but enriched by two dates from a site almost 150km away, in another region with a longer duration of Hinkelstein occupation? Calculations like these may be statistically satisfying, but nothing more. The re- sults cannot be the reliable approach to prehistoric reality that should be the aim of our research. Of course, we may assume that the dates from Vin≠a and the Hungarian sites (and consequently also those from the early middle Neolithic in Alsace and SW Germany) are correct, shifting the whole sequence by 150–200 years; however, in this case, the LBK and Rössen series from Alsace would be too old, and even worse, the dendrochronological dates from the wells would not fit with the 14 C chronology. And how should we explain a series like Rottenburg? A model that assumes they are all correct (Fig. 2) is completely incompatible with relative chronology (Fig. 3). A serious problem at least with some of the 14 C dates on bone collagen has to be admitted – or a completely new chronological system has to be constructed, with, for example, eLBK lasting for six centuries (5350–4750 cal BC), ending at about the same time as Hinkelstein, as we have no evident rea- son to discard the dates from Rottenburg while ac- cepting the Ra∫i∏te dates from Szederkény. In the light of these results, the ‘formative phase’ of eLBK in Transdanubia must be reconsidered: the dif- ference of about 150 years between its beginning and the start of the rest of the eLBK might be ex- plained by the different material used for 14 C dating. Looking from the West, a date about 5500 cal BC for the beginning of eLBK expansion seems to be more reliable, while the date of 5350 cal BC based on bone collagen dates is about 150 years too young. Interpreting the ‘formative phase’ as a regional va- riant of the eLBK and not as its predecessor – as sup- posed by typological maps and the results of a cor- respondence analysis of some 600 features of eLBK Discrepancies between archaeological and 14 C-based hronologies> problems and possible solutions 279 and the intial Flomborn/Notenkopf phases (Strien 2017) – both starting at 5500 cal BC, would avoid the demographic problems connected with the late date for the expansion horizon. If this model were accepted, the start date for Vin≠a-Belo at least Brdo would also be too young. What becomes evident in this context is the lack in SE Central Europe of any modern typochronology based not only on a clearly defined typology, but also on appropriate statistical instruments. So far, for the whole area between Bylany and Vin≠a-Belo Brdo no modern chronology has been published that covers at least a major part of the time span between the start of the LBK and the transition to its succes- sors. The establishment of fine-grained regional typo- logical chronologies and their interregional correla- tion using classical analyses of foreign forms and importations as well as modern statistical methods (e.g., Procustes analysis of seriations) and additional information like the demographic considerations dis- cussed above, correlated with precise absolute dates from the wells, may and should be supported, but can in no way be replaced by 14 C dating. Radiocar- bon dating is not the Holy Grail of prehistoric archa- eology, especially as long as environmental influen- ces on its results are neither fully recognised nor un- derstood. Bofinger J. 2005. Untersuchungen zur neolithischen Be- siedlungsgeschichte des Oberen Gäus. Materialhefte zur Archäologie in Baden-Württemberg 68. Konrad Theiss Verlag. Stuttgart. Denaire A. 2009. Radiocarbon dating of the western Eu- ropean Neolithic: comparison of the dates on bones and dates on charcoals. Radiocarbon 51(2): 657–674. Denaire A., Lefranc Ph., Wahl J., Bronk Ramsey C., Dun- bar E., Goslar T., Bayliss A., Beavan N., Bickle P. and Whit- tle A. 2017. The cultural project: formal chronological mo- delling of the early and middle Neolithic sequence in Low- er Alsace. Journal of Archaelogical Method and Theory 24 (4): 1072–1149. doi 10.1007/s10816-016-9307-x Dombek G. 1989. Die 14 C-Daten von Ulm-Eggingen. In C.- J. Kind, Ulm-Eggingen. Bandkeramische Siedlung und mittelalterliche Wüstung. Konrad Theiss Verlag. Stuttgart: 401–412. Friederich S. 2011. Bad Friedrichshall-Kochendorf und Heilbronn-Neckargartach. Forschungen und Berichte zur Vor- und Frühgeschichte in Baden-Württemberg 123. Kon- rad Theiss Verlag. Stuttgart. (no year). Archäologische Ausgrabungen. Bandkera- mischer Brunnen, Leipzig-Plaußig. www.archaeologie. sachsen.de/download/II_22_plauss_02.pdf Gleser R. 2012. Zeitskalen, stilistische Tendenzen und Re- gionalität des 5. Jahrtausends in den Altsiedellandschaften zwischen Mosel und Morava. In R. Gleser, V. Becker (eds.), Mitteleuropa im 5. Jahrtausend vor Christus. Beiträge zur internationalen Konferenz in Münster 2010. Neolithi- References ∴ ∴ kum und ältere Metallzeiten. Studien und Materialien Bd. 1. Lit Verlag Dr. W. Hopf. Berlin: 35–103. Haack F. 2016. Die frühneolithische Grabenanlage von Herxheim bei Landau: Architektur, Verfüllungsprozesse und Nutzungsdauer. www.diss.fu-berlin.de/diss/receive/ FUDISS_thesis_000000102455 Herbig C., Maier U., Stäuble H. and Elburg R. 2013. ‘Neo- lithische Füllhörner’. Archäobotanische Untersuchungen in fünf linienbandkeramischen Brunnen in Westsachsen. Offa 69/70: 265–293. Jeunesse C., Lefranc Ph. and van Willigen S. 2009. Die pfälzische Bandkeramik: Definition und Periodisierung einer neuen Regionalgruppe der Linearbandkeramik. In A. Zeeb-Lanz (ed.), Krisen-Kulturwandel-Kontinuitäten. Zum Ende der Bandkeramik in Mitteleuropa. Beiträge der Internationalen Tagung in Herxheim bei Landau (Pfalz) vom 14.–17.06.2007. Internationale Archäologie. Arbeitsgemeinschaft, Symposium, Tagung, Kongress 10. Verlag Marie Leidorf GmbH. Rahden/Westf.: 61–77. Jeunesse C., Strien H.-C. 2009. Bemerkungen zu den stich- bandkeramischen Elementen in Hinkelstein. In A. Zeeb- Lanz (ed.), Krisen-Kulturwandel-Kontinuitäten. Zum En- de der Bandkeramik in Mitteleuropa. Beiträge der Inter- nationalen Tagung in Herxheim bei Landau (Pfalz) vom 14.–17.06.2007. Internationale Archäologie. Arbeitsge- meinschaft, Symposium, Tagung, Kongress 10 10. Verlag Marie Leidorf GmbH. Rahden/Westf.: 241–248 Jakucs J., Bánffy E., Oross K., Voicsek V., Bronk Ramsey C., Dunbar E., Kromer B., Bayliss A. Hofmann D., Marshall P. and Whittle A. 2016. Between the Vin≠a and Linearband- Hans-Christoph Strien 280 keramik Worlds: The Diversity of Practices and Identities in the 54 th –53 rd Centuries cal BC in Southwest Hungary and beyond. Journal of World Prehistory 29(3): 267– 336. doi 10.1007/s10963-016-9096-x Kind C.-J. 2005. Stratigraphie und Steinartefakte der Sied- lung der Ältesten Bandkeramik von Rottenburg ‚Fröbel- weg’. In J. Bofinger, Untersuchungen zur neolithischen Besiedlungsgeschichte des Oberen Gäus. Materialhefte zur Archäologie in Baden-Württemberg 68. Konrad Theiss Verlag. Stuttgart: 255–322. Lefranc Ph. 2007. La céramique du Rubané en Alsace: contribution à l’étude des groups régionaux du Néoli- thique ancient dans la plaine du Rhin supérieur. Mono- graphies d’Archéologie du Grand-Est – Rhin, Meuse, Mo- selle 2. Université Marc Bloch. Strasbourg. Lenneis E., Lüning J. 2001. Die altbandkeramischen Sied- lungen von Neckenmarkt und Strögen. Universitätsfor- schungen zur Prähistorischen Archäologie 82. Verlag Dr. Rudolf Habelt GmbH. Bonn. Lindig S. 2002. Das Früh- und Mittelneolithikum im Ne- ckarmündungsgebiet. Universitätsforschungen zur Prähi- storischen Archäologie 85. Verlag Dr. Rudolf Habelt GmbH. Bonn. Link T. 2014. Die linien- und stichbandkeramische Sied- lung von Dresden-Prohlis. Eine Fallstudie zum Kultur- wandel in der Region der oberen Elbe um 5000 v.Chr. Veröffentlichungen des Landesamtes für Archäologie Sach- sen Bd. 60. Landesamt für Archäologie. Dresden. Neustupný E. F., Vesely K. 1977. Most radiocarbon dates I. Radiocarbon 19(2): 183–187. Oross K., Osztás A., Marton T., Köhler K., Ódor J. G., Szé- czényi-Nagy A., Bánffy E., Alt K. W., Bronk Ramsey C., Kro- mer B., Bayliss A., Hamilton D. and Whittle A. 2016. Mid- life changes: the Sopot burial ground at Alsónyék. Bericht der Römisch-Germanischen Kommission 94: 151–178. Pettitt P., Hedges R. 2008. The age of the Vedrovice ceme- tery: the AMS radiocarbon dating programme. Anthropo- logie XLVI(2–3): 125–134. Schmidt B., Gruhle W. 2003. Wuchshomogenität als ein neues Analyseverfahren zur Verbesserung der dendro- chronologischen Datierungsmethode. Die Hölzer der neo- lithischen Brunnen von Erkelenz-Kückhoven, Zwenkau und Mohelnice sowie vom Fundplatz Kaster. In J. Eckert, U. Eisenhauer and A. Zimmermann (eds.), Archäologische Perspektiven. Analysen und Interpretationen im Wan- del. Festschrift für Jens Lüning zum 65. Geburtstag. In- ternationale Archäologie Studia Honoraria 2. Verlag Ma- rie Leidorf GmbH. Rahden/Westf.: 49–60. Schröter I. 1970. Die bandkeramische Siedlung von Köln- Mengenich. Kölner Jahrbuch 11: 37–47. Spatz H. 1996. Beiträge zum Kulturenkomplex Hinkel- stein – Großgartach – Rössen. Materialhefte zur Archäo- logie in Baden-Württemberg 37. Konrad Theiss Verlag. Stuttgart. 1999. Das mittelneolithische Gräberfeld von Trebur, Kreis Groß-Gerau. Materialien zur Vor- und Frühgeschi- chte von Hessen 19. Landesamt für Denkmalpflege. Wiesbaden. Stäuble H. 2005. Häuser und absolute Datierung der Ältesten Bandkeramik. Universitätsforschungen zur Prä- historischen Archäologie 117. Verlag Dr. Rudolf Habelt GmbH. Bonn. Stehli P. 1989. Zur relativen und absoluten Chronologie der Band-keramik in Mitteleuropa. In J. Rulf (ed.), Bylany Seminar 1987. Collected Papers. Arheologický ústav ∞es- koslovenske akademie věd. Prague: 69–78. 1994. Chronologie der Bandkeramik im Merzbachtal. In J. Lüning, P. Stehli (eds.), Die Bandkeramik im Merz- bachtal auf der Aldenhovener Platte. Beiträge zur neo- lithischen Besiedlung auf der Aldenhovener Platte 5. Rheinische Ausgrabungen 36. Rheinland Verlag GmbH: 79–192. Strien H.-C. 2010. Mobilität in bandkeramischer Zeit. In D. Gronenborn, J. Petrasch (eds.), Die Neolithisierung Mittel- europas. Internationale Tagung, Mainz 24. bis 26. Juni 2005. Römisch-Germanische Zentralmuseum Tagungen Bd. 4. Mainz: 497–508. 2013. Besiedlungsgeschichte des Zabergäus 5500–5000 v. Chr. In C. Schenk, P. Wanner (eds.), Heilbronnica 5. Beiträge zur Stadt- und Regionalgeschichte. Stadtar- chiv Heilbronn. Heilbronn: 35–50. 2017. Westexpansion und Regionalisierung der ältes- ten Bandkeramik. Welt und Erde. Kerpen-Loogh. Tasi≤ N., Mari≤ M., Bronk Ramsey C., Kromer B., Barclay A., Bayliss A., Beavan N., Gaydarska B. and Whittle A. 2015. Vin≠a-Belo Brdo, Serbia: The times of a tell. Germania 93: 1–75. Tegel W., Elburg R., Hakelberg D., Stäuble H. and Büntgen U. 2012. Early Neolithic Water Wells Reveal the World’s Oldest Wood Architecture. PLOS ONE 7(12): e51374. doi: 10.1371/journal.pone.0051374 Wild E. M., Stadler P., Häußer A., Kutschera W., Steier P., Teschler-Nicola M., Wahl J. and Windl H. J. 2004. Neoli- thic massacres: local skirmishes or general warfare in Eu- rope? Radiocarbon 46(1): 377–385.