[en] Following an introductory chapter, Chapter II gives a short summary of the characterisitics of the Hard X-ray Imaging Spectrometer. Chapters III and IV are devoted to the origin of the impulsive hard X-ray bursts. In chapter III it is investigated whether, from a theoretical point of view, the creation of a hot thermal region is feasible. Observations of hard X-ray spikes require that the heating to temperatures of the order of 5 X 10₈ K takes place in several seconds. It is shown that this is indeed possible, provided the resistivity of the plasma in the dissipation regions is considerably enhanced over its classical value, determined by Coulomb collisions. The observational tests of the thermal and non-thermal models are discussed in chapter IV. Three spatially well resolved flares with strong hard X-ray bursts have been analysed, using X-ray, UV and optical data. It is concluded that for these flares the bulk of the impulsive hard X-rays originate from the footpoints of loops. An important discovery with HXIS is the presence of a long-lived hot component after the occurrence of the impulsive spikes. For the example treated in chapter V its temperature is about 4 x 10₇ K. Though slowly varying, the source can be considered quasi-stationary. Finally in chapter VI the response of transition region and chromosphere to the injection of fast electrons is investigated. The results of a 1D-numerical simulation indicate that the observed large increase in the density of a flaring loop after the impulsive phase cannot be fully explained by invoking only the injection of non-thermal electrons. (Auth.)