S. Mahunka szerk.: Folia Entomologica Hungarica 53. (Budapest, 1992)

le specimen each, two species by two ex., four species by three ex., two species by six ex. and one species by seven ex. (Table 2). The collecting data suggest that the effect of human impact on the drosophilid assemblages is detectable also at the least disturbed sites of Hungary: domestic species are found at the Ménes-forrás, Aggtelek. On the other hand, there are some of the rarest species in the habitats strongly deformed by human activity (Magyarkút). The frequencies in the assemblages changed extremely from one year to another. Both dendrograms of the Renkonen index and the Czekanowski index, respectively, for 1988-89 define two groups: on the upper part of the dendrograms one can find those samples where the black "fruit breeding" species (D. obscura, D. subobscura) prevailed and the mushroom feeders were present in low numbers and in low frequencies (on the Czekanowski dendrogram 9 samples of the 13 are from 1989). On the lower part we find those samples where the numbers and frequencies of D. phalerata, testacea, etc. were high (on the Czekanowski dendrogram 8 samples of the 12 are from 1988). As for the partition of the samples at the lowest level, the only remarkable difference in the two dendrograms is with the sample of M30079, where all the three species of D. phalerata, testacea and D. obscura were subdominant (altogether nearly 87%). On the Renkonen dendrogram one sample (A12108) is found in a separated position: this is where D. kuntzei was strongly dominant by 66 %. Both dendrograms show a group of 4 samples where the "domestic" species were present with high dominance values. This analysis does not prove any distinct separation of the samples according to collecting places or seasons. It seems likely that though human activity also deforms the drosophilid assemblages in the mountain creek valleys. However, the small populations of the rare species can survive the acting disturbance of that size even in the "worst" of the four sites. Below a certain level of frequency we lose power for judging the species richness: similarly to the situation found in other communities (dung heap flies, coprophagous flies on pastures, agromyzids in cereal crops, etc.): a good part of the drosophilid species is so rare (their populations are so small or their catchability is so low) that only an indefinite part of the extant rare species can be detected at all in a given area. This level of frequency depends mainly on the sample size. We are tending to believe that a majority of the dipterous species is rare (i.e. belongs to this category) at any site. However, ratios of the populations of the dominant-subdominant, constant-subconstant species are hypothesized, behind which more non-detected species are to be expected than behind population frequencies deformed by human impact. Although we are convinced that true ecological research must be focussed on the dominant-subdominant species for various reasons. We must also continue to develop more effective methods for a better understanding of the rare species as for their survival, population dynamics, density, life strategies, breeding substrates and so on.