‘Good genes’ is the default explanation for the evolution of elaborate ornaments, despite abundant evidence that the most attractive mates are seldom those that produce the most viable offspring

It’s Not about Him: Mismeasuring ‘Good Genes’ in Sexual Selection. Angela M. Achorn, Gil G. Rosenthal. Trends in Ecology & Evolution, December 16 2019, https://doi.org/10.1016/j.tree.2019.11.007

Highlights
. ‘Good genes’ remains the default explanation for the evolution of elaborate ornaments, despite abundant evidence that the most attractive mates are seldom those that produce the most viable offspring.
. ‘Good genes’, in which preferred traits predict offspring viability, is often conflated with other indirect benefits, including genetic compatibility, heterozygosity, and offspring attractiveness.
. Few studies in fact test the key predictions of ‘good genes’ models and, as predicted by theory, they show scant evidence for additive effects of mating decisions on offspring viability.
. Direct tests of indirect genetic benefits should measure the attractiveness and viability of offspring from a large number of matings, distinguish between additive and nonadditive benefits, and control for differential investment in offspring.

Abstract: What explains preferences for elaborate ornamentation in animals? The default answer remains that the prettiest males have the best genes. If mating signals predict good genes, mating preferences evolve because attractive mates yield additive genetic benefits through offspring viability, thereby maximizing chooser fitness. Across disciplines, studies claim ‘good genes’ without measuring mating preferences, measuring offspring viability, distinguishing between additive and nonadditive benefits, or controlling for manipulation of chooser investment. Crucially, studies continue to assert benefits to choosers purely based on signal costs to signalers. A focus on fitness outcomes for choosers suggests that ‘good genes’ are insufficient to explain the evolution of mate choice or of sexual ornamentation.

Keywords: genetic qualitymate choiceindirect benefitsgenetic benefits


Reevaluating the Evidence: ‘Good Genes’ in Context

The question of whether sexual selection is good or bad for choosers and populations is a central one in evolutionary biology, animal communication, and conservation biology. By focusing on signals and signal costs, studies often fail to test the basic premise that ornaments are the target of mate choice [46], let alone that they confer any benefits on choosers.

When studies do test for ‘good genes’, evidence suggests that this process accounts for a modest fraction of variance in sexual fitness [2,6,13]. A recent meta-analysis [6] showed that attractiveness was highly heritable, consistent with FLK models, but good genes received mixed support. Attractiveness did not correlate with traits directly associated with fitness (life-history traits). However, attractiveness did positively correlate with physiological traits, such as immunocompetence and condition.

Similarly, recent studies provide at best mixed support for the intuition that the prettiest males have the best genes, although perhaps the most tenacious males do. The clearest evidence for good genes (Table 1) has found them for traits where choosers have limited agency to make mating decisions. Persistent courtship or mating is likely to increase courter success no matter how choosers behave before mating [47], and frequently impose direct costs on females [48]. The one study that controlled for differential allocation [38] yielded equivocal results. There is widespread evidence that choosers invest more in the offspring of attractive males, but more-ornamented courters may manipulate choosers into investing in a mating beyond the lifetime fitness optima of the choosers.

For conspicuous display traits, weak signals of good genes should be the rule. A seeming paradox of good genes models is that preferences for good genes are most likely to be maintained if genetic effects on viability are weak, since this slows the depletion of genetic variation by selection [49]. ‘Good genes’ are likely to be less important to preference evolution than self-reinforcing coevolution channeled by mating biases [13] and direct selection on mating decisions [13,14] (see Outstanding Questions).

More rigorous measures of ‘good genes’ speak to another central question, namely whether sexual selection is more likely to confer a positive or negative effect on population mean fitness [50]. On the one hand, populations can benefit from accelerated purifying selection through sexual selection on the courting sex, meaning that sexual selection can increase population fitness if there is a positive correlation between preference and fitness. On the other hand, sexual selection can decrease population fitness through reduced viability as a consequence of sexual conflict.

The answer likely depends on the nature of selection experienced by populations. When competing males are parasitized, sexually successful male fruit flies (Drosophila melanogaster) sire more parasite-resistant offspring, while the opposite holds true for winners of contests between unparasitized males [24] (Table 1). Along these lines, a recent meta-analysis [50] used 459 effect sizes from 65 experimental studies in which researchers manipulated the presence or strength of sexual selection, encompassing both intrasexual selection and mate choice, and then measured some aspect of fitness. The results indicate that sexual selection tends to increase population fitness, particularly when populations are exposed to novel environmental conditions.

However, in contrast to Prokop and colleagues’ meta-analysis [6], fitness traits related to immunocompetence were an exception: sexual selection covaried with weaker immunity.

Concluding Remarks

Few studies evaluate the critical predictions of benefit models of mate choice. Those that do suggest that good genes have an important role in adapting to novel environments, but that they may be more important in terms of glimpses we have suggests that good genes provide an important, if circumscribed, contribution to chooser fitness.

We can begin to tease apart the selective forces shaping mating preferences, but tests of good genes hypotheses must assess meaningful measures of offspring viability. Assessing offspring viability is conceptually straightforward, if not always easy in wild populations. This can be done by using molecular markers to reconstruct pedigrees and correlating preferred trait expression with survivorship to maturity [36], or by examining proxy measures of viability, such as juvenile growth rate or size [39]. This approach does not rule out the possibility of differential allocation. If choosers invest more in the offspring of attractive partners even at the expense of their own lifetime reproductive success, then mates may not be providing a net increase in average offspring viability [51,52]. Artificial insemination or, in externally fertilizing species, in vitro assays [8] can control for differential allocation.

A counterintuitive point about ‘good genes’ and ‘genetic quality’ is that, by definition, they are difficult, if not impossible, to infer from courter genotypic data alone. Notably, a gene favored by selection (e.g., a gene that buffers oxidative stress and helps produce attractive offspring) can carry a higher genetic load with respect to other components of viability [53,54]. An allele that is ‘good’ with respect to courter function may be in linkage disequilibrium with alleles that reduce (or increase) offspring viability. Again, direct measures of fitness are required to measure ‘good genes’. A major challenge to testing for good genes is that these effects are predicted to be weak [6,55] and, therefore, require large sample sizes. For large, long-lived animals with small populations (e.g., nonhuman primates), longitudinal samples across generations provide a feasible, if slow, approach to detecting viability consequences of mate choice. Simply measuring correlations between ornament elaboration and other courter phenotypes does not distinguish among models of signal evolution through mate choice.

‘Good genes’ is appealing because it assigns utilitarian explanations to seemingly extravagant traits and the desires that shape them. There is allure to the idea that mating preferences can increase population mean fitness and local adaptation. A loose construction of ‘good genes’ and ‘genetic quality’ remains the default explanation for mating preferences and sexual dimorphisms outside the immediate field of sexual selection and in the popular literature. We suggest that the persistence of this default view comes from a sloppy conception of these terms that leads to insufficient empirical tests of adaptive hypotheses.

Unfortunately, ‘good genes’ especially lends itself to what Bateson [56], writing about the term ‘mate selection,’ termed ‘unconscious punning’. The term conjures up so much more than ‘breeding value for viability.’ There is a precise technical term, coined by Galton in 1883 [57], which means ‘good genes’ or ‘true genes’ in Greek. Eugenics is tainted forever by the policies it incited, but we remain entranced by the intuition that Beauty marches in lockstep with Truth. Thus, evidence for ‘good genes’ and ‘genetic quality’ in the vernacular sense, is conflated with support for precise evolutionary models. We would hesitate to study the foraging ecology of koalas exclusively by grinding up eucalyptus leaves, but this is all too often the logic we invoke to study mate choice [46]. A perspective centered on choosers, rather than on the signatures that their choices leave on courters, is essential for understanding mate choice and its consequences.