Functional response: an overview and experimental guide

Abstract

The functional response of predators is central to any description of the process of predation, and is probably the best example of the influence of the individual behaviour on the population dynamics of predator- prey interactions. The estimation of the functional response is useful because of its implications on stability, but probably more importantly because it allows appropriate estimates for predation parameters and the comparison between potential agents of biological control. There are, however, several limitations in its estimation that affect the quality of the results obtained from empirical studies and, consequently, set a limit to the biological conclusions derived from them. These involve all stages of the estimation process: the design and implementation of the experiments, the analysis of data and the model used to estimate functionalresponse parameters. Our aim in this paper is to establish a base line for the empirical study of the functional response. Throughout the paper, we set our focus on parasitoids because these insects are a classical group in studies of functional response, and also because of their importance as agents of biological control of pests. We first describe the functional response and present an overview of the main problems arising in its estimation. Then, we summarise and exemplify the procedure that allows the best estimation of the functional response. Our report differs from past attempts in that it involves all stages of the estimation process. The best protocol involves variable time experiments, because they allow the studied animals to display behaviours under a more natural scenario. Along these lines, prey patch densities should mimic natural densities and individual behaviour should be observed throughout the assay. Also, other individual traits such as the abilities to switch prey and learn must be standardised. Functional response is, after all, a behavioural study and thus careful observations allow detecting inconsistencies and flaws of the experiment and data collection. With the data, a Logistic Regression Analysis is appropriate, because it allows distinguishing readily between type II and type III functional responses. Once the shape of the response is determined, appropriate model parameters may be estimated by a non-linear least squares procedure. We finally illustrate this protocol, providing analysis details in an appendix, with a study of the functional response of the parasitoid Ibalia leucospoides.