Fig. 2.

Predator-prey dynamics for two species X and Y lead to a scatterplot (relating sampled species abundances) that is interpretable when successive time-points are connected (a). If, however, the information about time were not included (b), these dynamics would appear uncorrelated because when X is high, Y can be either high or low, and vice versa. Thus, even in a completely deterministic system, it is impossible to tell whether two species interact with each another simply by examining multiple samples in which both are present. However, this technique is widely used in practice despite its limitations. Figure adapted from (Holling 1973)