The importance of population heterogeneities in detecting social learning as the foundation of animal cultural transmission

Abstract

High levels of within-population behavioural variation can have drastic demographic consequences, thus changing the evolutionary fate of populations. A major source of within-population heterogeneity is personality. Nonetheless, it is still relatively rarely accounted for in social learning studies that constitute the most basic process of cultural transmission. Here, we performed in female mosquitofish (Gambusia holbrooki) a social learning experiment in the context of mate choice, a situation called mate copying (MC), and for which there is strong evidence that it can lead to the emergence of persistent traditions of preferring a given male phenotype. When accounting for the global tendency of females to prefer larger males but ignoring differences in personality, we detected no evidence for MC. However, when accounting for the bold-shy dichotomy, we found that bold females did not show any evidence for MC, while shy females showed significant amounts of MC. This illustrates how the presence of variation in personality can hamper our capacity to detect MC. We conclude that MC may be more widespread than we thought because many studies ignored the presence of within-population heterogeneities.

Keywords: Gambusia holbrooki; mate copying; mosquitofish; personality; social learning.

Figure 1.

Top view on the experimental set-up and design of the mate-copying experiment. (a) First mate-choice test: the grey fields mark the two mate-choice zones. The observer female (black) is in the large tank and two males (grey), a large (lower left) and a small (upper right), are placed diagonally in one of the small stimulus tanks at each end of the large tank. (b) Demonstration phase for 10 min. A stimulus female (black) is placed in a separate tank next to the small male (lower right tank) so that it is visible to the observer female. A pseudo-stimulus female (black) is also placed in a separate tank next to the large male (upper left) but behind a screen (thick black bar) and thus not visible to the observer female. (c) Second mate-choice test (like the first test).

Figure 2.

Top view on the experimental test maze to determine boldness as exploratory behaviour. The tank was separated into five chambers by four dark grey plastic boards, each with an opening in their middle for females to explore all chambers. The female was gently placed into chamber 1 at the beginning of the test and given a maximum of 10 min to reach chamber 5.

Figure 3.

Residual plots for (a) mate-copying trials, (b) control 1 for consistency in mate choice trials in the absence of social information, and (c) control 2 for group size effects trials. Null or negative residuals indicate no copying while positive values indicate copying. Numbers next to the points indicate the sample size for each specific size-ratio category.

Figure 4.

Predicted SLSs in shy females of the mate-copying treatment according to our selected statistical model, including the effects of male size-ratio (bottom left axis), its square and the SB of the observer female (bottom right axis). Here, the SLI is social learning values predicted by the model accounting for these effects. Those predicted values are bell shaped as revealed by the quadratic term, and the level of the surface globally goes up as the SB during the first preference test (see Methods). Note that here we use the SB during the first mate-choice test in shy individuals of the mate-copying treatment, but that using the SB during the second mate-choice test gave similar results. (Online version in colour.)