Eight food crusts from Seedorf LA 245 and 23 from Bebensee LA 76 were radiocarbon-dated in the 1990s at the Leibniz-Labor, Kiel, following Nadeau et al. (1998), by the excavators, Klaus Bokelmann and Jürgen Hoika. Any remaining material from these food crusts (held by the laboratory or the Schleswig-Holstein State Archaeological Museum, Schloss Gottorf) was sampled for stable isotope analysis in 2011-12. Samples from Schlamersdorf and Kayhude were analysed as part of B. Philippsen's PhD project. A variety of clearly terrestrial and clearly aquatic samples, in addition to several food crusts, was dated from each site.
Note 1: There are other 14C ages from Bebensee LA 76, including six food crust results from the Utrecht laboratory (Hartz 2011, table 2), which are generally under 5000 BP. Several bones date to the Funnel-Beaker period, while others appear to date to the early (pre-ceramic) Ertebølle period. Nevertheless, some diagnostic Ertebølle pottery was recorded.
Most of the potsherds with food crusts at these sites are not typologically diagnostic, and their attribution to Ertebølle or Funnel Beaker cultures on the basis of radiocarbon dates is hazardous, given the potential for large 14C reservoir effects. Indeed, all four food crusts dated from diagnostic Funnel-Beaker sherds at Bebensee gave 14C ages greater than 5300 BP, when the Funnel-Beaker tradition would normally be dated to 5000 BP or later (Andersen 1986; 1990, Fischer 2002a; Hartz and Lübke 2006; Glykou 2011). Radiocarbon ages under 5000 BP can only be associated with Funnel-Beaker (or later) pottery, but older 14C ages from undiagnostic sherds in mixed deposits might represent either Funnel-Beaker or Ertebølle pottery (and given the predominance of late Funnel-Beaker pottery types among the diagnostic sherds at Bebensee, it is unlikely that many of the dated food crusts are from Ertebølle sherds). Eight samples were available for stable isotope analysis, including one of those from a diagnostic Funnel-Beaker sherd and three that gave 14C ages under 5000 BP; the other four are potentially from Ertebølle sherds [Note 1].
Note 2: It is unclear whether the four Funnel-Beaker samples from the same square metre and depth represent four different vessels. Three 14C ages of food crusts from unidentified potsherds from an adjacent site, Seedorf LA 296, were reported by the excavator (Bokelmann 1994: 5690±65 BP (OxA-4481, δ13C=-27.7‰), 5835±70 BP (OxA-4483, δ13C= 26.6‰) and 5935±65 BP (OxA-4482, δ13C= 26.1‰)). Another food crust from Seedorf was dated previously: (OxA-4016, 5160±70BP, δ13C=-31.3: Hedges et al. 1993). It is unclear whether this sample was from LA 245, LA 296, or another site at Seedorf.
The Seedorf LA 245 samples were clustered in two areas of the site – five samples with very similar 14C ages, of late Funnel-Beaker date, were from one small area, and three potentially Ertebølle dates are from samples found 11–12m away at a lower depth [Note 2]. Three of the younger food crusts and all three of the older samples were available for analysis. Most other 14C samples from Seedorf LA 245 were from bone or antler artefacts found in other parts of the site, and these are typically much older or younger than the food crusts, but a bone awl was dated to the late Ertebølle period and could be contemporary with the older food crusts.
At Schlamersdorf, evidence including 14C ages of terrestrial samples (charcoal and bones) confirms that the Funnel-Beaker and Ertebølle occupations are spatially distinct, and at the LA 5 location it is likely that we are only dealing with Ertebølle pottery. Only Ertebølle pottery was recorded at Kayhude, and 14C dates of worked wood and antler samples (as well as the typology of these artefacts) all belong to the later Mesolithic period (Clausen 2007).
Note 3: Where both untreated sample material and Kiel's 14C dating extract were analysed, we report only the extract results. Unpublished EA-IRMS data on dating extracts from coastal sites in Schleswig-Holstein are generally consistent with those reported by Craig et al. (2011), who analysed untreated food crusts from the same assemblages. Post-depositional carbonates and organics removed by acid-alkali-acid pretreatment should be carbon-rich and nitrogen-poor, but we found no systematic difference in C/N values between extracts and untreated samples of the same food crusts (n=15).
For 14C dating, food crust samples were pre-treated by the normal acid-alkali-acid method (e.g. Olsson 1976); 1M HCl at 80°C for one hour, 1M NaOH at 80°C for at least three hours and 1M HCl at 20°C overnight, although the NaOH concentration and temperature was reduced to 0.5 or 0.2M at room temperature for archaeological food crusts. In a few cases, food crusts were treated using a sequence of solvents (Bruhn et al. 2001) before acid-alkali-acid extraction, although there is no evidence that this affected either the 14C ages or stable isotope results. For EA-IRMS (Elemental Analysis-Isotope Ratio Mass Spectrometry), the extract used for 14C dating was analysed where possible. Stable isotope measurements of extracts and untreated food crust samples show no systematic differences, and results are usually close (differences <1–2‰). Discrepancies between repeat measurements from a single sherd appear to be related more to lack of homogeneity in the food crusts themselves than to differences in sample preparation [Note 3].
The food crusts from Schlamersdorf and Kayhude were analysed at the AMS 14C Dating Centre at Aarhus University, by combustion in a EuroVector elemental analyser coupled to an IsoPrime stable isotope ratio mass spectrometer. Most samples yielded enough material for doublet measurements. Food crust samples from Bebensee and Seedorf were analysed by Dr Ulrich Struck at the Berlin Natural History Museum, using a THERMO/Finnigan MAT V isotope ratio mass spectrometer, coupled to a THERMO Flash EA 1112 elemental analyser. δ13C values are reported as ‰ VPDB, δ15N values as ‰ AIR. Atomic C/N ratios were derived from %C and %N measurements.