Fig. 2.

Each isofemale line’s intercept as a random effect in a null Cox regression (“Extinction Risk”) plotted against (a) Antagonism and (b) Concordance (see Methods and Fig. 1), demonstrating, respectively, that male-benefit/female-detriment and generally low-fitness (i.e. low-male/low-female) isofemale lines were more sensitive to inbreeding relative to their respective opposite extremes. Linear relationships (for visual purposes only) provided by ordinary least squares regression.