MEASURING INDIVIDUAL-LEVEL RESOURCE SPECIALIZATION

TLDR

Many species that appear generalized actually consist of individual specialists using only a subset of resources, yet niche variation is often inferred by testing a null hypothesis of common resource use, a practice that biases against negative results and obscures true specialization, and has traditionally been measured by partitioning total niche width into within‑ and between‑individual components. The paper aims to provide four quantitative indices of intrapopulation resource‑use variation and proposes two alternative measures that quantify mean resource overlap between individuals and their population, while discussing each measure’s strengths and weaknesses. The authors employ four quantitative indices and two alternative overlap measures, and recommend supplementing cross‑sectional data with isotope ratios, parasite loads, or diet–morphology correlations to confirm temporal consistency of specialization. The indices are only useful when data are high quality; coarse‑grained resource measurement can make specialists look generalized, while cross‑sectional sampling in patchy environments can inflate apparent specialization.

Abstract

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.