In nature, larger species assembly appears more stable, while May's random-assembly model proves the opposite is true. This study experimentally tests whether addition of a species stabilizes host-parasitoid laboratory systems. As a measure for stability of the systems, dominant Lyapunov exponents (LEs) were estimated by fitting nonmechanistic (LENNS model, incorporating neural networks) and mechanistic nonlinear models to the population dynamics data. Both approaches showed that addition of the parasitoid, Heterospilus prosopidis, increased the LEs to be slightly positive values, indicating the species addition destabilized the system to become chaotic. Sensitivity analysis further revealed that the high searching efficiency and short handling time of the parasitoid primarily contributed to the destabilization. When the other parasitoid species was added to a host-Heterospilus system, the system was stabilized, or LEs decreased.
The two contrasting results indicate that it is not the number of species that determines the stability of species assembly but the identity of the species to add to. The data analyzed in the present study were provided by Drs. S. Utida and M. Shimada.