The aim was to build a prototype concrete tank which was suitable for the seasonal storage of solar heat. The building material was watertight concrete (``Densit``) of the type used for constructing floors and the tank was to be integrated in a district heating plant. Problems arose concerning leaks and cracks in the concrete construction. It was thought that these were caused by an accumulation of water resulting from the use of two kinds of membrane, one of which was more open to diffusion than the other. Measurements were taken during 4 periods from September 20th 1990 until January 3rd 1991. The current of energy and the temperatures in 5 layers of the storage facility were measured. Heat losses were registered as 328 kWh per 24 hours. The fact that this was more than expected was thought to be due to leakage and surface evaporation. The testing confirmed the general suspicion that no usable cheap membranes are currently available. It is suggested that clay-sealing could help to solve the problem of leakage. (AB).