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

(tabulated similarly as given in Table 4 for 1990) were processed by a group linkage method using average linkage to gain dendrograms (Figs 1-2). The pattern of the occurrence of species at the four sites is briefly exposed through values of the Jaccard index in Table 5. Since the number of the samples taken and of the specimens collected is very low at Visegrád, only the other three sites are comparable (to "test" the effect of this insufficiency, we "complemented" the Visegrád list with the three synanthropic species, which actually were not collected there). Aggtelek, the least disturbed area, is more similar to Magyarkút, the most disturbed area of the four, than to the Garadna valley (B), for the total and for the rare species, respectively. The representation of the rare species at the four sites is illustrated in Table 6. The dominant species and the possible relation of the dominants to the number of the rare species are shown in Table 7. To enable a rough estimate of the sample size needed for a higher number of species representation, we made a list for the highest number of species per sample with their actual number of specimens collected (Table 8). As we. can see, 400 or more specimens in a sample seem to be necessary to stand for a better chance that a fair number of rare species is included. For three sites the summarized frequencies of the species were ranked from the highest to the lowest on a logarithmic scale (Figs 3-5). The curves show a logarithmic tendency rather than a lognormal distribution, the size of these accumulated samples and the number of the species involved is too low for a fitness test. It is less precise but more easily perceiveable to order the species in logarithmic frequency categories, like this: dominance Aggtelek Bükk N.P. Magyarkút % 101.00.10.01X spp. 3 13 10 5 31 % 10 42 32 16 spp. 1 11 6 11 29 % 3 38 21 38 spp. 3 6 15 10 34 % 9 18 44 29 It seems probable from the table that the least steep section of the curves is likely in the second or third frequency category. DISCUSSION Though the size of the 35 samples taken is rather uneven, we can suppose that the drosophilid assemblages in the low mountain creek valleys of Hungary are rich in species. The number of specimens of several species is smaller than in Western Europe (cf. e.g. Burla and Bächli (1991). The ratio of the rare species (those of unknown catchability) is rather high: in total five species were represented by a sing-