Sperm quality and quantity evolve through different selective processes in the Phasianidae

Abstract

Sperm competition is often considered the primary selective force underlying the rapid and diversifying evolution of ejaculate traits. Yet, several recent studies have drawn attention to other forms of selection with the potential of exceeding the effects of sperm competition. Since ejaculates are complex, multivariate traits, it seems plausible that different ejaculate components vary in their responses to different selective pressures. Such information, however, is generally lacking as individual ejaculate traits tend to be studied in isolation. Here, we studied the macroevolutionary patterns of ejaculate volume, sperm number, sperm length and the proportion of viable normal sperm in response to varying levels of sperm competition, body size and the duration of female sperm storage in pheasants and allies (Phasianidae). Ejaculate volume, sperm number and sperm viability were all relatively higher in polygamous than in monogamous mating systems. However, whereas ejaculate volume additionally covaried with body size, sperm number instead increased with the female sperm-storage duration, in conjunction with a decrease in sperm length. Overall, our results revealed important details on how different forms of selection can jointly shape ejaculates as complex, composite traits.

Figure 1

Graphical representation of the results of Table 1, showing the response of four ejaculate traits to (left) the social mating system as a proxy of sperm competition and (right) either the female egg-laying period or body mass. The left-hand panels depict the least-squares means with 95% confidence intervals, after controlling for body mass and the female egg-laying period (numbers at the bottom of each panel indicate sample sizes). The right-hand panels show the partial regressions derived from the same models, controlling for body mass and the social mating system on both axes. The labelled data point in panel B indicates Dendragapus obscurus, the removal of which considerably strengthened the negative relationship (see text and Table 1). The proportion of viable normal sperm was arcsine-square-root transformed and then converted to percentages by multiplying by 180/π. Total sperm length was measured in μm, ejaculate volume in μl, and total sperm number in millions.

Figure 2

Graphical representation of the response of the relative sperm investment, ln(sperm length/sperm number), to (A) the social mating system as a proxy of sperm competition and (B) the female egg-laying period. (A) depicts the least-squares means with 95% confidence intervals and sample sizes, after controlling for the female egg-laying period and body mass. (B) shows the partial regression derived from the same model, controlling for body mass and the social mating system on both axes.

Figure 3

Pairwise correlations between three ejaculate traits. The proportion of viable normal sperm was arcsine-square-root transformed and then converted to percentages by multiplying by 180/π. Total sperm length and total sperm number were quantified in μm and millions, respectively.