Besides the routine updating of the dataset, one of the HISTALP-subsets – the LSS (long series subset) of 32 temperature series starting in the 1840s and earlier – was subject to a major re-analysis. The probability of a systematic bias of early instrumental warm season temperatures in the GAR was first shown and discussed at the basis of a detected decoupling of tree-ring reconstructed and directly measured temperatures (Frank and Esper, 2005, Büntgen et al., 2006b, Frank et al., 2006 ). Hiebl, 2006 generally discussed the case of the possibly too warm measured early summer temperatures comparatively with similar findings from Skandinavia and in historic climate model runs. The physical reason of the EI-bias (Early Instrumental - bias) was insufficient shading of the sites in the times before the new regulations in respect to standardised screens were introduced in the “weather-services period”. Böhm et al, 2008 proposed and applied a method to correct the temperature LSS based on longterm parallel measurements and on intensive additional metadata studying. The EI bias correction cooled the mean GAR-temperatures of the warm season (April-September) by about 0.4°C, leaving winter temperatures unchanged. The version 2008 of station-mode as well as grid-mode GAR temperatures consists of EIB-corrected series henceforth.