Antagonistic co-evolution and the evolution of genotypic randomization.

Antagonistic co-evolution, such as the pursuit and flight of host-parasite co-evolution, easily generates cyclical co-evolutionary dynamics. It is well known that a fluctuating, contrary environment may favour the evolution of recombination, but previous analyses have shown that the optimal rate of recombination declines as the period of the environmental fluctuations gets longer. It is here shown that the direction of selection for recombination (for higher or lower rates) may only be sensitive to fluctuation period if the fluctuations are generated by non-co-evolving features of the environment, such as changes in climate. In the simple model of this paper, co-evolutionary cycling provides an advantage for recombination that is independent of period length. This independence of period length necessitates a new framework for understanding the advantage of recombination in the context of co-evolution. In the model studied in this paper it is the phase difference of the host and parasite trajectories that is the relevant feature, and phase relationships are independent of the co-evolutionary cycle time. That is, the phase difference of the oscillations is independent of their frequency.