This paper describes an investigation into the function of the breakdown of algae in the cycle of elements in a lake. The liberation and mineralisation of phosphorus, nitrogen, silicium and iron during sterile autolysis of Scenedesmus quadricauda and some other algae were determined. Autolysis was induced by chloroform treatment, U. V. irradiation or heating to 60° C. 70-80 % of the P-compounds are liberated in a few days, only the phosphate of the nucleic acids and proteins not being liberated. The forms in which phosphorus is liberated, and the mechanisms involved, were followed up. Some compounds are liberated by enzymatic activity. Only 20-30 % of the N-compounds are liberated. The rest of the N-compounds belongs to the proteins and the nucleic acids, and remains as a slag. The liberation of P- and N-compounds does not depend on size or nature of the killed algae investigated. The liberation of Si-compounds is a non-enzymatic slow process. In 5 weeks 20-30 % of the Si-compounds were liberated from Stephanodiscus Hantzschii. This amount could be increased by adding NaHCOj to the suspensions of autolysing diatoms. The mineralisation of iron, which is also a slow process, depends on the chelate-forming capacity of the milieu. Preliminary experiments were carried out to digest the slag of sterile autolysis by means of bacteria. When the slag was suspended in lake water one half of the nitrogen was converted to ammonia in 5 days, while the other half was divided up into bacterial and Scenedesmus nitrogen, the mutual ratio of which could not yet be determined. It appeared not yet to be possible to digest the slag by cultures of Pseudomonas fluorescens, Pseudomonas aeruginosa, Proteus vulgaris and Bacillus subtilis. The fact, however, that Ps. fluorescens and Ps. aeruginosa showed an increased respiration, when the slag was given as a substrate, and could also be cultured on a lye extract of the slag, is a stimulus for further investigations in this direction. The combination of high light intensity and high temperature as a cause of the dying off of algae in nature is suggested to be a more frequently occurring phenomenon than is generally believed. A preliminary experiment is mentioned, which gave probably a positive answer to the question whether autolysis can play a part in a lake, where bacterial putrefaction always occurs.