Variability in predator-prey experiments: simulation using a stochastic model

Abstract

Simulation of several actual predator-prey experiments was carried out, using a stochastic version of the Lotka-Volterra equations. Great variability was found from run to run, with occasional extinction of either predator or prey with the same parameters. It is suggested that single experimental runs are very unreliable for parameter estimation, multiple replicates of experiments being necessary. It was not found necessary to postulate such mechanisms as hiding places for most experimental results, the discrete stochastic equations being adequate for this purpose.