MEASURING INDIVIDUAL-LEVEL RESOURCE SPECIALIZATION

Many apparently generalized species are in fact composed of individual specialists that use a small subset of the population's resource distribution. Niche variation is usually established by testing the null hypothesis that individuals draw from a common resource distribution. This approach encourages a publication bias in which negative results are rarely reported, and obscures variation in the degree of individual specialization, limiting our ability to carry out comparative studies of the causes or consequences of niche variation. To facilitate studies of the degree of individual specialization, this paper outlines four quantitative indices of intrapopulation variation in resource use. Traditionally, such variation has been measured by partitioning the population's total niche width into within- and between-individual, sex, or phenotype components. We suggest two alternative measures that quantify the mean resource overlap between an individual and its population, and we discuss the advantages and disadvantages of all four measures. The utility of all indices depends on the quality of the empirical data. If resources are measured in a coarse-grained manner, individuals may falsely appear generalized. Alternatively, specialization may be overestimated by cross-sectional sampling schemes where diet variation can reflect a patchy environment. Isotope ratios, parasites, or diet–morphology correlations can complement cross-sectional data to establish temporal consistency of individual specialization.