Figure 2.

The even-moment effect. (A): Case in which individuals with phenotype 0 have a higher expected fitness than do those with phenotype 1 (${\widehat{w}}_0$ = 1.04, ${\widehat{w}}_1$ = 1) as well as a higher variance in fitness (var(w₀) = 0.9984, var(w₁) = 0). (B): The mean value of Δ$\overline{\varphi }$ from monte-carlo simulations, given the fitness distributions in (A) and δ = 0. (C): Case in which both phenotypes have the same mean fitness, and the expected direction of evolution is towards the phenotype with the higher variance in fitness (but the lower fourth moment). The variance, fourth moment, and $\widehat{\Omega }$ for each phenotype are shown below the distributions (D): Example of a case in which, for N = 2, the expected direction of evolution is towards the phenotype with lower geometric mean fitness.