Figure 2.

Between-female agreement in mate choices in a two-way choice chamber. Each female had eight trials and proportion of agreements in dichotomized preferences (identity of the male that a female spent the larger fraction of time with) was calculated between pairs of genetic sisters, pairs of foster sisters and pairs of unrelated females. The three plots show all trials (A), only trials with time allocation to the preferred male of >70% (B), and only trials with time allocation to the preferred male of >85% (C). Because limiting the comparisons to only clear choices means excluding trials with less clear choices, sample sizes vary among plots. Differences between the agreement among genetic sisters and the agreement among foster sisters relative to the agreement among unrelated females were tested in a generalized linear model (GLM) with binomial error structure and logit link and a single categorical predictor (type of female pair) with three levels. P-values refer to the contrasts between the two types of females and the reference category (unrelated females).