Species number with confidence

Benthic sampling by J. Frederick Grassle and Nancy J. Maciolek during the 1980s contributed to a recognition of the enormous diversity of biota in the deep sea. The problem of predicting on the basis of such sampling, the number of species in a community is a classical one in statistical ecology. Here, we show how to construct a lower prediction bound on species number using the sequential broken-stick model of relative abundances. We illustrate the method using some data from Grassle and Maciolek's original sampling effort.