An empiricists’ guide to sexual conflict over parental investment: a comment on Paquet and Smiseth

Research on sexual conflict over parental investment is oddly out of step with analyses of other forms of this evolutionary conflict. Whereas other researchers focus on the never-ending evolutionary instability that ensues when the sexes have conflicting evolutionary interests (Arnqvist and Rowe 2005), we persist in believing that conflict over parental investment has an evolutionarily stable outcome (Lessells 2012). Although others realize that proper analyses of evolutionary conflict must involve quantifying fitness (e.g., Mock and Forbes 1992; Holland and Rice 1998), we are often content to measure only behavior and to cross our fingers that it is a plausible proxy for fitness (even though it is commonly not: Sheldon 2002). And, although research on other aspects of sexual conflict has been led by incisive and elegant experimentation (Arnqvist and Rowe 2005), theoretical analyses dominate research on sexual conflict over parental investment (Lessells 2012). Paquet and Smiseth’s (2016) timely and stimulating review shows how empiricists can now lead the way in developing new thinking about sexual conflict over parental investment that brings it more into line with other research on conflict between the sexes (although, modestly, they stop short of saying this themselves). By calling our attention to the role that maternal effects might have on the outcome of sexual conflict over parental investment, they show how there might not be a stable outcome to sexual conflict over parental investment after all. The outcome might vary from pair to pair, according to the powers of manipulation shown by each individual breeding female. The outcome might also fluctuate in response to microecological conditions that influence the female’s capacity to manipulate her partner via a maternal effect. In fact, the capacity for individual variation in maternal effects to influence the outcome of sexual conflict is even greater than that described in Paquet and Smiseth (2016). Recent empirical work with burying beetles shows that parental effects can also influence sexual conflict not just within a generation, but by acting across generations too, to influence the provision of care when offspring mature into parents themselves (Kilner et al. 2015). The outcome of sexual conflict then varies according to the extent of mismatch in parental quality between breeding partners (Kilner et al. 2015). A second empirical lesson to be learned from Paquet and Smiseth’s review is that sexual conflict cannot properly be analyzed without measuring the fitness consequences of parental behavior, separately for each parent and their young. This is essential because it is the only way to evaluate which partner has been manipulated into behaving in a way that is suboptimal for them. Measuring parental behavior alone will not suffice nor, in fact, is it even necessary for understanding the evolutionary conflict of interest here (De Gasperin and Kilner 2015). With more studies measuring the fitness consequences of sexual conflict over provisioning, the theoretical assumption that the outcome is evolutionarily stable will quickly become testable. These 2 lessons, then, should form the benchmarks of quality for any future empirical work on sexual conflict over parental investment. It should start with the assumption that the outcome will vary from family to family and across generations and might proceed to determine why any such variation arises. It should also focus on measuring the fitness consequences of parental care for each member of the family, rather than quantifying parental behavior alone. Studies that set out to test the specific hypotheses outlined in Paquet and Smiseth’s review must additionally be designed to experimentally eliminate closely related competing explanations for outcomes that might be observed. For example, they must show unambiguously that males are responding to the maternal effect alone, and not some correlate in the mother, the offspring, or the local environment. They must further demonstrate that changes in maternal effects function to be manipulative and are not simply a component of prenatal maternal investment that precipitates changes in the extent of paternal investment, in a division of labor that is to the mutual advantage of each sex (Lessells 2012). Finally, females must be shown to cause changes in male investment. There should be no scope for reversing the causal arrow, with variation in male quality instead driving strategic changes in prenatal maternal investment (Sheldon 2000). Surprisingly, few of the existing studies that are claimed to show maternal effects manipulating fathers actually meet these exacting standards. Nevertheless, with carefully designed experiments, measuring effects on (correlates of) fitness, empiricists will be excitingly placed to lead the next generation of research on sexual conflict over parental investment.