The Benninghoff method of estimating the total number of fossil pollen grains of a particular type in a sample of sediment involves adding an estimated number of exotic pollen grains to the sediment sample and then counting the numbers of fossil and exotic grains in a small portion of the sediment sample. Several methods for forming confidence intervals from such data have been suggested in the literature. When the exotic concentration is known, exact conditional intervals can be formed. When variation in the estimate of the exotic concentration needs to be considered, confidence intervals based on a normal approximation to the logarithm of the maximum likelihood estimate perform well except in extreme cases. For a given expected total count, the number of fossil grains of a particular type will be estimated with smallest variance when the fossil and exotic concentrations are equal.