Benthic diatom investigations of some Hungarian streams have been carried out in order to find reference sites in assistance to the ecological quality analysis of the different river types according to the Hungarian river typology. On the basis of the coevaluation of biological and chemical parameters, among the investigated streams, some part of Kemence stream seems to be the appropriate reference site for the classification of the highland, siliceous rivers of Hungary. However, further invastigations are needed for the selection of reference sites for the other types of Hungarian river typology. This is the most difficult in the case of lowland rivers, which are exposed to more severe anthropogenic impacts. Important is the question which biological parameters should be investigated and considered when assessing the ecological condition of our waters. In the United States, e.g. the assessment of the relative abundance of Achnanthidium minutissimum is quite widespread (Stevenson and Bahls 1999), the extent of which indicates the level of disturbance. The use of this method is hampered, on one hand, by the fact that A. minutissimum is an early coloniser species which means that if we do not take samples from mature periphyton, we might find it dominant simply because of that, and, on the other hand, that distinguishing between the species varieties is not always unambiguous under light microscope because of its small size, and these can have different ecological needs. It is also recommended to calculate species richness, diversity and evenness (Stevenson and Bahls 1999), but based on our investigations these parameters are not in correlation with water quality. Investigation of other diatom features such as tolerance and sensitivity, some autecological parameters (as it was applied and recommended by Fore and Grafe 2002 for investigating the reaction of diatom associations to anthropogenic disturbances) might also be useful in completing the water quality analysis of rivers, however, these methods need further investigations. Some of the indices (EPI-D, IPS, IBD) calculated with the software Omnidia are suitable for the classification of our rivers, but the investigation of more rivers, flowing through different types of stone are necessary to choose the appropriate indices. The reason for this is that the different indices work with a different register of species and obviously those indices will prove to be suitable the species register of which shows the highest similarity to the diatom associations of the certain water type.