The seasonal cycle of the planktonic diatom Melosira italica subsp. subarctica has been studied in a small lake in the English Lake District for the last 24 years. In 1967 there were four periods when the population increased. The vernal population was the largest ever recorded. The early summer and early autumn increases were small and the late summer increase large. The late summer increase was caused by the artificial destratification of the lake and favourable nutrient conditions. The size of the population was greater than that in spring for most of the years since 1945. The effect of a previous artificial destratification in 1961 was small because the experiment did not start early enough or last long enough, and the weather conditions were unfavourable. In 1967, the population began to die by the end of August. Though the available data point to nutrient depletion as the cause of the death of the cells, it does not seem that there was any lack of available silicon, nitrogen or phosphorus. Reasons are given for the view that, if the lake was isothermal and vertically mixed each summer, large populations of Melosira would not always develop. It is suggested that the normal, seasonal alternation of periods of stratified and unstratified water are of biological advantage to Melosira, enabling it to compete with other diatoms which can also produce large populations but cannot live for long on or in the deposits in the dark or in the absence of oxygen.