Evaluation of community indices in seasonal assemblages of butterflies (Lepidoptera) at different frequency of transect count

Community indices were evaluated at different frequency of transect counts of butterflies by randomization. Randomized samples were obtained by re-arranging 1, 3, 6, 12 and 18 counts from 18 counts at a study site (Hiraoka), and 1, 2, 3, 5 and 15 from 15 at another site (Narukawa), in 1994. Morisita's new estimate of species richness, QM gave a good estimate when the number of counts was five or more. QM estimated from the sum of both assemblages (58 species) was 70.7, while 73 species were recorded at least once from 1990 to 1994. The Octave method always resulted in an underestimation. Diversity indices, Shannon's H, Simpson's D and , Hurlbert's E(Sn), Itow's b, and Morisita's H gave good estimates when the number of counts was five or more. Temporariness of univoltine species in flight period made a bias in the estimates. We concluded that the count is needed at least six times (once a month) and the estimates were improved by separating the multivoltine species from the univoltine species. Intensive counting during the flight period of the univoltine species was recommended.