Scots pine (Pinus sylvestris L.) trees growing on shingle fields offer a unique possibility to reconstruct precipitation and study climate variability in the fairly humid eastern part of central Sweden. Tree-ring characteristics were compared with monthly (1890-2001) and daily (1961-2001) climate data from an adjacent meteorological station. Chronologies for latewood (LW), earlywood (EW), and tree-ring widths (RW) were constructed from 73 living and dead trees. Correlation analyses show that tree growth is most sensitive to early summer precipitation. EW shows the strongest correlation with precipitation in May and June while LW is best correlated with June and July precipitation. A reconstruction model for May-June precipitation was calculated using principal component analysis (PCA) regression (regular regression) including EW, LW, and RW for present and previous years. The model explained 46% of the variation in May-June precipitation and allowed a reconstruction back to 1560. Information about wet and dry years was collected from historical documents and was used to validate the result. Periods with precipitation above and below the mean show agreement with previous reconstructions of spring precipitation from tree rings in Finland and of spring floods from estuary sediments in the region. Analyses of correlations between meteorological stations and reconstructed precipitation show that the model is valid for the coastal part of central Sweden. The authors conclude that Scots pine trees on shingle fields are well suited for precipitation reconstruction, and the separate analyses of LW and EW improve the reconstruction.