In sexually reproducing species with internal fertilization, breeding success of males depends on two stages. First, males have to mate with a female; second, sperm have to fertilize the eggs. During both stages, intense competition can occur: among-male to access a female, and among-sperm to access the egg.
The outcome of both stages of sexual selection (mating and fertilizing the egg) has attracted considerable attention and it has been established that male behaviors, ornaments and weapons are important attributes improving male chances to access mates. Similarly, many sperm traits (e.g., motility, velocity) and ejaculate attributes (e.g., number of sperm) have been identified as predictors of fertilization success during sperm competition.
However, when females mate with several males during the same reproductive bout, other parameters unrelated with the intrinsic quality of the ejaculate and/or the sperm can also shape fertilization success. For instance, given that matings always occur in sequence, the order and the timing of mating with respect to egg laying might have major effects on fertilization success, with matings occurring too early having reduced chances to fertilize the eggs. This might be due to the fact that once stored in the female reproductive tract, sperm age and are lost. Therefore, sperm transferred during matings occurring closer to egg laying might outcompete in numbers sperm transferred during previous matings. This process has been called last male sperm precedence.
But what is the most important parameter accounting for among-male differences in fertilization success during sperm competition? Is it the intrinsic quality of sperm/ejaculate, or parameters potentially unrelated to quality (such as mating order) do play a more important role?
Addressing these questions has so far been difficult for several reasons. First, experimental approaches usually tend to explore the effect of one parameter while keeping the others constant, which does not allow assessing the relative importance of different predictors of fertilization success. Second, exploring the factors predicting fertilization success (in the lab or in the field) can only be done when each of the possible predictor can be assessed, which is generally challenging. For instance, during natural matings of species with internal fertilization, it is usually impossible to assess how many sperm have been transferred by each competing male.
Ex-situ captive breeding is a conservation tool that have been deployed to preserve and restore populations of several endangered species. While captive breeding is a valuable instrument for conservation, it may also provide unparalleled opportunities to address fundamental research questions. We took advantage of the ex-situ captive breeding of the endangered North African houbara bustard (Chlamydotis undulata) to investigate the relative importance of several parameters likely to affect fertilization success during sperm competition. This was possible because this captive breeding is entirely based on artificial inseminations, and for each insemination, information on sperm/ejaculate attributes, male and female attributes is available. In addition, the houbara represents an ideal model species as, in the wild, birds are promiscuous, with males lekking to attract females and choosy females potentially having several matings before laying. We were able to explore, retrospectively, whether parameters related or not to male quality are the most important predictors of fertilization success during sperm competition.
A male North African houbara bustard performing a sexual display (Photo credit: Yves Hingrat). During the display, the male erects the white feathers of the neck and the chest and performs a circular running. The display is also accompanied by a booming.
Female houbara can store sperm in dedicated structures and use them well after insemination. Based on this rationale, we included eggs from females successively inseminated with different males and with each competing male contributing a unique insemination. A total of 2,226 inseminations and 901 eggs met the inclusion criteria. These eggs were laid by 599 females while 879 different males contributed semen for the inseminations. Thanks to this large sample size, we could use a machine learning approach to rank the importance of potential predictors of fertilization success.
We found that insemination order and the delay between the day the insemination occurred and the day of egg laying were by far the most important predictors of fertilization success (Figure 1). Inseminations occurring at the last position in the sequence and close to the day of egg laying had the highest probability to fertilize the egg. In agreement with the traditional view of sperm competition, we also found that traits usually thought to reflect ejaculate quality, such as number of sperm and motility predicted fertilization success, with males having higher number of sperm in the ejaculate and highly motile sperm being more successful to fertilize the egg. However, these traits appeared to have a much smaller weight as predictors of fertilization success compared to insemination order.
Figure 1. We used SHapley Additive exPlanations (SHAP) to interpret the output of the gradient boosted trees model. The variables on the vertical axis are ranked according to the mean of the absolute SHAP values, indicating their relative importance as predictors of siring success (higher mean absolute SHAP values corresponding to a larger contribution to the siring success prediction). Dots represent the SHAP values (positive values being indicative of higher siring success, lower values being indicative of higher siring failure) for each insemination (N = 2,226) as a function of the value of the predictor (the redder the color the higher the value of the predictor). Please note that the insemination order is inversed (1 referring to the last male in the sequence). Percentage of days displaying was used as a proxy of investment into secondary sexual traits, and number of male breeding events in captivity as a proxy of adaptation to captivity.
Although using captive individuals and artificial insemination allow exploring traits that would have been impossible to study in the field, we reckon the outcome of sperm competition under natural conditions might depend on different parameters than those included in our work.
Nevertheless, if these results reflect a pattern that applies to natural conditions, they raise the question of who controls the order of matings and its possible consequences. For instance, if order of mating is under female control, this might shift the force of selection from the post-copulatory to the pre-copulatory stage.