The Statistics of Rarity1
Ecologists deal with rarity in many guises--species can be rare, particular interactions may
be uncommon, and catastrophic events that reshape landscapes are, by definition, infrequent.
Although ecologists often seek out abundant species or events for their investigations out of
convenience, rare species are often of special central concern to conservation biologists, reserve
managers, and legislators, and historical legacies of rare events are pervasive in ecosystems.
Statistical analysis and modeling of rare events is often necessary, but as MacKenzie et al. point
out in their contribution to this Special Feature, ``rare species [or events] are simultaneously the
species for which strong inferences about state variables and vital rates are most needed and the
species for which such information is most difficult to obtain.'' The standard set of statistical
tools used by the majority of ecologists are difficult or inappropriate to use when analyzing rare
species or events, either because assumptions such as normality or homoscedasticity do not hold,
or because the required sample sizes are impossibly large. The overall goal of this Special Feature
is to present a cross-section of techniques for sampling, quantifying, and modeling rarity.
Methods for analyzing data on rare species and events come from a variety of disciplines and
were originally designed for addressing specific questions. Most of these questions, including
problems related to quality control in manufacturing, frequency of flooding, and econometrics,
are unrelated to ecological questions, but the mathematical and statistical tools transcend disci-