Search for a command to run...
Evenness is an important property of all ecological communities. It may be defined as degree to which the abundances are equal among the species present in sample or (Molinari 1989). The most accepted measures of evenness are the J index (H'/H'max) (Pielou 1969), and Alatalo's (1981) simple but important modification of Hill's (1973) E21 ratio. Molinari (1989) studied the behavior of these indices over gradient of increasing evenness and found that both J and F21 have two important shortcomings: a) they tend to strongly overestimate evenness specially when its real value is low, and b) their behavior is decidedly nonlinear, the indices grow very fast at low equitabilities and then level off and change very slowly. Consequently, he developed G21 a calibrated version of F21, to overcome these limitations. G21 is an empirical mathematical transformation of F21 that effectively linearizes it and greatly improves its accuracy, at least in collections of only two species. As practical tool it is the best index already available, but it has both conceptual and practical drawbacks. The most important are: 1) It is purely empirical measure, without any clear ecological meaning. As pointed out by Molinari (1989), the usual interpretation of F21 as ratio of very abundant to abundant species is not correct and should be avoided. For G21 the problem is even greater, because of the complexity of the mathematical transformations involved. 2) The statistical properties of G21 are unknown. Consequently, it is impossible to make formal statistical comparisons among samples or communities. This is problem shared by many other indices. 3) G21 (as F21) measures evenness within those species with intermediate and high proportional abundances. Its value is little influenced by the species with low abundances. Consequently, the index fails to discriminate between communities whose main difference rests in the relative proportion of rare species. This last point is of great importance. In many ecological communities the proportion of rare species is an important component of evenness, perhaps the one that most clearly distingishes one community from another (Medina and Huber 1992; and see Fig. 4 for real example). Besides, the rare species are very sensitive to perturbations or environmental changes. Their sudden decrease or even disappearance produce changes in evenness, that good index should be able to detect. In the past, the lack of sensitivity to rare species was considered as desirable property in an evenness index, because it makes it robust against sampling bias. However, the occurrence of rare species in sample is not an effect of sampling bias, but consequence of important characteristics of the community that should not be ignored. This article proposes new evenness index that tries to overcome these problems. It is much simpler than the already discussed indices. It shows linear behaviour over gradients of increasing equitability, and has an immediate and realistic ecological interpretation. It also gives equal weight to all species independently of their abundances. It allows the derivation of simple diversity index, and its statistical distribution is known (Appendix 1).