Confidence limits of species proportions in microfossil assemblages

We analyzed the confidence limits of benthic foraminiferal counts obtained from fossil deep-sea assemblages to figure out the optimum number of specimens needed for a reliable interpretation of species proportions. Precision and reliability are analyzed by calculating the binomial confidence limits. We demonstrate that in studies based on the interpretation of the proportional distribution of only the dominant taxa, counts of as low as 100 specimens are sufficient. Studies, however, that aim to investigate species diversity patterns usually require higher counts. Generally, we suggest that percent abundance given in micropaleontological studies should include the binomial error estimate.