Branch wood from the lake shore settlements of Horgen Scheller, Switzerland: Evidence for economic specialization in the late Neolithic period

596 waterlogged, uncarbonized branch wood specimens from four cultural layers of the late Neolithic lake shore site of Horgen Scheller (between 3080 and 3030 B.C.) on Lake Zilrich, Switzerland, were examined. The following parameters were analysed: taxonomy, diameter and age of the twigs, cutting season, and distribution of the branches over the area as well as their distribution in the different cultural layers. The twigs were divided into the following three groups, based on differences observed in these analyses: conifers, deciduous trees with catkins, and deciduous trees without catkins. Branch diameter examinations revealed that branches of deciduous trees were thicker on average than those of conifers. While branches from deciduous trees without catkins were gathered at the age of seven years on average, those of conifers and deciduous trees with catkins were older. Waney edge analyses showed that the majority of the twigs were cut within the winter months. Among the deciduous trees without catkins there is, however, a markedly higher proportion of branches cut in summer than in the other two groups. The taxonomic examinations demonstrated that 70% of the specimens consisted of twigs of eitherCorylus avellana L. (hazel),Taxus baccata L. (yew), orAbies alba Mill. (silver fir), thus implying the great economic importance of these species in the settlements of Horgen Scheller. To our knowledge this is the first report that a remarkably high proportion of yew twigs (26.8%) has been found in an analysis of subfossil twigs. The possibility is discussed that the inhabitants had specialized in obtaining yew from the region. Concentrations of wood finds, possible remains of buildings, became evident from the analyses of the distribution of the branch wood over the excavated area. A very similar spatial distribution pattern of the taxa was found through the different cultural layers. Therefore it seems likely that existing structures were reused when a new settlement was built. The data also show that the building outlines slowly moved towards land over time. A rising lake water level could have led to gradual dislocation of the settlement