A Circular Depletion Sampler (the CDS) was constructed in order to collect removal data for estimating benthic sample populations and species catchability. The removal sampling method was compared to Modified Hess, Surber, and Kicknet sampling methods, and gave consistently higher and less variable estimates than other methods. Taxa showing the lowest catch abilities included the case-building Trichoptera, the Mollusca, and the Annelida. The highest catch abilities were shown by the smaller Chironomidae and the nonburrowing Ephemeroptera. Taxa with high catchability generally showed high emigrability for Surber and Kicknet methods, particularly the aquatic Insecta. The determination of species richness and evenness diversity was significantly biased by low catchabilities or emigration or both with Modified Hess, Surber, and Kicknet sampling methods. Removal samples have a higher probability of representing rare species, and as expected, pooled removal collections accumulated species at a higher rate than the pooled collections of other methods. The dominance diversity of pooled CDS collections became stable more rapidly than those of other methods, and gave higher resolution between benthic communities of different structure. DPLETE, a computer program for making removal population estimates by the method of Carle and Strub (1976) and making dominance diversity estimates based on information theory is presented, and is suitable for use in any study where removal data has been collected.