Limitations of methods to test density‐dependent fecundity hypothesis

Summary1 Two main hypotheses are usually invoked to explain density dependence in fecundity: the habitat heterogeneity hypothesis (HHH) and the individual adjustment hypothesis (IAH). Although simple methods have been proposed to discriminate between the two hypotheses, their adequacy was tested for only a limited set of real and model populations. 2 In a computer simulation study based on a stochastic territory‐based approach, Ferrer, Newton & Casado (2006, Journal of Animal Ecology , 75 , 111–117) argued that a strong negative relationship between mean fecundity and its skewness in stable or increasing populations provides critical support for HHH, as this relationship should be lacking under IAH. A negative relationship between mean fecundity and its coefficient of variation (CV) was predicted under both hypotheses, although with a lower slope under IAH. 3 We used a comparable simulation approach, with model populations parameterized from an increasing Bonelli's eagle Hieraaetus fasciatus population (1992–2006), to show that both HHH and IAH can produce indistinguishable relationships between mean fecundity and both its CV and its skewness. 4 Strong negative correlations between the mean and both its CV and its skewness can emerge as statistical artifacts under biologically plausible assumptions, and so they may be largely inadequate to infer mechanisms underlying density dependence in demographic parameters.