Misidentification of the Bray-Curtis similarity index

Abstract

The Bray-Curtis (BC) similarity index is misidentified in various software programs, and the index used by these programs is more appropriately credited to Czekanowski (Cz). The BC and Cz indices differ because abundant species have the greatest impact on the Cz index, while all species (and samples) 'count the same' in the BC index, due to a double standardization of the dataset. To illustrate the difference between these indices, I analyzed the example dataset in the PRIMER manual. In one instance, similarity values differ by 40 % (21 vs. 61 %). These discrepancies must be considered in comparisons between analyses conducted with software programs that use the true BC index, and programs that use the Cz index but mislabel it as 'BC'. © Inter-Research 2008.