Factorial enrichment experiments with nitrate-N, phosphate-P and selenite-Se, were conducted in situ, with natural phytoplankton assemblages, during two consecutive spring-bloom periods (1983 and 1984) in Lake Erken, Sweden. Nutrient conditions, biomass indicators and algal physiological status, in experiments and in lake water, were compared. Characteristic peaks of phytoplankton, dominated by diatoms, developed in the lake in both years. In 1984, chlorophyll alpha biomass (30 mug l-1) reached almost twice that of 1983 (17 mug l-1). Inorganic P resources were depleted in both years, but inorganic N was exhausted only in 1984. In 1983, Stephanodiscus hantzschii GRUN v. pusillus GRUN. KRIEGER completely dominated the biomass peak. In 1984, Melosira islandica v. helvetica O. MULL occurred in great numbers along with S. hantzschii v. pusillus. The growth rate of S. hantzschii v. pusillus, surplus phosphorus dynamics and potential alkaline phosphatase activity were different during the two spring periods. In the experiments, eight alternative nutrient situations were tested during four phases of the natural phytoplankton development. The algae responded to enrichments during the entire phytoplankton bloom in 1983 and 1984. Nutrient interactions were important, especially in 1983, and occurred most frequently during the late exponential phase, biomass maximum and declining phase in both years. Interactions between N and P were most common in 1983, occurring during the late exponential phase and biomass maximum. N and Se interaction occurred occasionally. Interactions between P and Se were more frequent in 1983 than in 1984. When nutrients interacted, the combined effect was generally smaller than the sum of their individual effects. P deficiency indicators and the experiments showed clearly that the bloom of 1983 was P limited. In 1984 no such clear pattern was found.