Figure 1.

(A and B) The relationship between $P_2$ and $P_1$ either assuming no genetic correlation ($r_{A_{v,1},A_{m,2}}=0$) between environmental variance in $P_1$ and the mean of $P_2$ using Monte Carlo simulation (MC_lin) and multiple regression (MR_lin) or assuming a negative genetic correlation ($r_{A_{v,1},A_{m,2}}=-0.5$) between them (MC_nonlin, MR_nonlin) when $r_{A_{m,1},A_{m,2}}=0.1$ (A) or when $r_{A_{m,1},A_{m,2}}=0.5$ (B) (${\sigma }_{A_{m,1}}^2={\sigma }_{A_{m,2}}^2=0.3,{\sigma }_{A_{v,1}}^2=0.05,{\sigma }_{e_1}^2\text{exp}\left(\left(1/2\right){\sigma }_{A_{v,1}}^2\right)=0.7,$ ${\sigma }_{e_2}^2=0.7,$ $r_{A_{m,1},A_{m,2}}=0.1,$ $r_{A_{m,1},A_{v,1}}=0,$ $r_e=0$).