Recovery techniques for waterlogged archaeological sediments : a comparison of different treatment methods for samples from Neolithic lake shore settlements

This paper presents the first comparable overview of different recovery techniques used for waterlogged Neolithic sediments in the surroundings of the Alps in the last decades. Such an investigation became necessary because it was not known which parts of plant taxa and remain types were lacking or completely underrepresented due to the application of inappropriate recovery techniques in Slovenian archaeobotany until recently. During the 2007 excavation of the approx. 5200 years old Neolithic pile dwelling site Stare gmajne, Ljubljansko barje, Slovenia, we compared three methods for the investigation of botanical macroremains: Method 1 (M1) included rough wet-sieving and subsequent drying of the fractions, method 2 (M2) rough wet sieving and keeping the fractions wet, and method 3 (M3) washing over (flotation) and keeping the fractions wet. M3 with gentle washing, systematic subsampling, examination, and sorting of macroremains in wet condition, as well as using 0.355 mm as smallest mesh size gave the best results. When using the rougher methods M2 or M1, waterlogged (uncarbonized) seeds of taxa like Linum usitatissimum, Papaver somniferum and Brassica rapa, waterlogged chaff of Cerealia, pericarps of Maloideae and Quercus sp., which are all fragile were underrepresented or even completely lacking, and therefore the plant spectra strongly biased. On the contrary, taxa like with lignified seed/fruit walls like Cornus mas, Corylus avellana or Rubus sp. were overrepresented when using the M2 and particularly the M1 method. The application of the M3 instead of the M1 method, which has been traditionally used in Slovenian archaeobotany until now, helped us to identify uncarbonized remains of Linum usitatissimum and Triticum div. spec. for the first time in a Neolithic wet site in Slovenia. Our study is to be seen as a contribution to the desperately needed standardisation of methods in archaeobotany. It shows clearly how strongly the plant spectra can be biased by inappropriate handling techniques. The conclusions hold for all kinds of waterlogged sediments of different periods.