Twenty-four measures of species diversity, richness and equitability were c
ompared using both real species abundance data (bird censuses on a standard
study plot) and simple simulation tests. Based on 20 criteria, the most re
commendable measures for the estimation of alpha diversity of species have
been Fager's "number of moves per specimen", exponential Shannon's informat
ion (or H'), reciprocal Simpson's lambda, and species richness (number of s
pecies). However. the last index is only appropriate when sample sizes are
approximately equal otherwise it can be misleading and Hurlbert's rarefacti
on method should be applied. The other measures tested, especially all equi
tability indices, have been shown to be problematic and should not be used
regularly for the measurement of species diversity.