A computer program has been developed which performs cluster analysis of microorganisms using methods of numerical taxonomy. The program is designed to group related strains, identify the groups by reference to known strains, and calculate a hypothetical median organism (HMO) for each group. The HMO serves to condense taxonomic information and provides a tag for each strain cluster. Every strain in a group is compared with the HMO established for that group. A representative strain for the group is obtained by selection of the strain showing highest similarity to the HMO. New data sets can be compared with data sets of previous analyses. Hence, the occurrence of the same taxonomic groups within separate data sets can be determined. Quantitative or qualitative differences in distribution of taxonomic groups within or between data sets can be measured. The output from the computer is a graphical display, using an on-line plotter; thus, the investigator is provided with visual comparison of data sets. Results obtained from a study applying the computer program in an analysis of taxonomic data obtained for 43 bacterial strains isolated from Chesapeake Bay indicate the usefulness of this method of taxonomic analysis in microbial ecology.