Abstract
Many biological studies require the capture of individuals for sampling, for example for measurement of morphological or physiological traits, or for marking individuals for later observations. Capture methods employed often vary both within and between studies, and these differing methods could be more or less effective in capture of different individuals based on their morphology or behavior. If individuals that are prone to capture by the selected method differ with respect to traits of interest, such sampling bias could generate misleading or simply inaccurate results. The selection of capture methods could introduce two different forms of sampling bias, with the individuals that are sampled differing from the population at large or with individuals sampled via one method differing from individuals that could be sampled using a different method. We investigated this latter form of sampling bias by comparing individual birds sampled using two common capture techniques. We caught free-ranging black-capped chickadees (Poecile atricapillus) using walk-in traps baited with seed and mist nets paired with playback of an audio stimulus (conspecific mobbing calls). We measured 18 traits that we expect might vary among birds that are trappable by these differing methods-one that targets birds that are food motivated and potentially less neophobic and another that targets birds that respond readily to a perceived predation risk. We found no differences in the sex, morphology, initial and stress-induced corticosterone concentrations, behavioral response to a novel object, or behavioral response to a predator between individuals captured by these two methods. Individual variation in the behavioral response to a novel object was greater among birds caught by mist nets, suggesting this method might provide a sample that better reflects population-level individual variation. We do not know if the birds caught by these two methods provide a representative sample of the population at large, but can conclude that selection of either of these two common capture methods can similarly sample mean trait values of a population of interest. To accurately assess individual variation, particularly in behavior, mist nets might be preferable.