During the spring seasons of 1983, 1986 and 1987 the development of phytoplankton in Lake Zurich was investigated (from February to May) using samples taken at short term intervals. The aim was to describe the effects of the short term dynamics of environmental factors on the algal growth. The results could then be used to discuss the existing theories to assess the start of phytoplankton growth pulses in spring. Only 7 to 10 days without wind driven vertical mixing were required in spring to start the first growth pulse, despite of a still very unstable water column (sometimes inverse thermal stratification). Mainly flagellates and Stephanodiscus hantzschii increased their biomass and achieved net growth rates of 0.1 and up to 0.65 d(-1) respectively During such a phase the mixing depth was always smaller than the euphotic depth. Later on, at the start of the spring bloom (= last growth pulse in spring before the clear water stage), the intensity of vertical mixing as well as the mixing depth were markedly reduced due to an increase in heat input and low wind. Then flagellates dominated (contribution up to 75.5 % of the areal biomass reaching 60 g fresh weight m(-2)) and the growth rate rose to a maximum of 0.65 d(-1). Standard models of critical depth considers that there is only a biomass increase if the mixing depth is smaller than the depth of a water layer positive balanced between production and respiration. This model for determining the beginning of a phytoplankton growth pulse in spring takes no account of the favorable light conditions for phytoplankton cells at calm and sunny days in February and March. The newly developed threshold value model takes these situations into account: It assumes that the phytoplankton biomass increases when the calculated effective light climate is equal or greater than a previously fixed threshold. The calculations are based on the mean light intensity within the mixed layer at windy days or within the euphotic depth (z(eu)) at calm days. In Lake Zurich a minimum of 0.2 10(6) J m(-2) d(-1) (= 0.9 mol quanta m(-2) d(-1)) has to be reached or surpassed in at least 3 days before an exponential increase of algal biomass can occur. The value does not depend on short term fluctuations in neither radiation nor mixing depth. It seems that this value is rather low comparing with those of investigations in other water bodies (up to 0.8 10(6) J m(-2) d(-1)) but high related to values from algal cultures (0.02 10(6) J m(-2) d(-1)). As the weather can only be forecasted a few days ahead with any certainty the period for a more or less accurate prediction of an algal bloom is restricted to about 1 to 5 days.