Call Duration as an Indicator of Genetic Quality in Male Gray Tree Frogs

See allHide authors and affiliations

Science  19 Jun 1998:
Vol. 280, Issue 5371, pp. 1928-1930
DOI: 10.1126/science.280.5371.1928


The “good genes” hypothesis predicts that mating preferences enable females to select mates of superior genetic quality. The genetic consequences of the preference shown by female gray tree frogs for long-duration calls were evaluated by comparing the performance of maternal half-siblings sired by males with different call durations. Offspring of male gray tree frogs that produced long calls showed better performance during larval and juvenile stages than did offspring of males that produced short calls. These data suggest that call duration can function as a reliable indicator of heritable genetic quality.

The “good genes” model of sexual selection predicts that some attributes of male courtship displays advertise genetic quality. Preferences for such attributes should allow females to mate with high-quality males and thereby benefit indirectly through enhanced quality of offspring (1). Although the good genes hypothesis has been tested several times (2), few studies have provided direct genetic evidence supporting this hypothesis (3). Only one such study involved a species in which females cannot benefit directly from their choice of mates (4). Because selection for direct benefits such as courtship feeding or parental care should overwhelm any selection for indirect (genetic) benefits (5), the role of good genes selection in the evolution and maintenance of female preferences is best tested in species in which females do not benefit directly from mate choice.

Female gray tree frogs (Hyla versicolor) strongly prefer male advertisement calls of long duration in laboratory experiments (6, 7). In the field, females freely initiate matings with calling males and do not always choose the first male encountered (7). Because males do not defend oviposition sites, offer nuptial gifts, or contribute parental care (8,9), and no difference has been found in fertilization success as a function of call duration (10), there are no apparent direct benefits of a female's mate choice. We therefore predicted that females selecting mates with long calls should benefit indirectly through increased fitness of offspring. This prediction can be tested by evaluating the relation between paternal call duration and the genetic quality of offspring.

Male gray tree frog advertisement calls consist of rapidly repeated pulses. In dense choruses and in response to playbacks, males tend to increase call duration by increasing the number of pulses per call (11, 12). Nonetheless, some males consistently produce longer calls than others in the same acoustic environment (7,12–14). Although long calls are usually produced at slow rates, thereby keeping aerobic metabolic costs relatively constant (11,14), males that produce long calls spend less time calling per night (11) and attend fewer choruses per season (8) than males that produce short calls. Long calls thus appear to impose higher nonaerobic costs than short calls. Call duration may, therefore, be an honest indicator of male genetic quality.

We tested whether call duration indicates heritable genetic quality by using maternal half-siblingships (half-sibships) to compare the performance of different males' offspring while experimentally controlling for all maternal effects. Maternal half-sibships were generated by artificially crossing each female with two males that had been giving calls of distinctly different durations in the same social environment (Table 1). Thus, within each maternal half-sibship, one sibship was sired by a male with calls of longer duration than the male siring the other sibship. Because call duration varies with chorus density, males' calls must be assessed in the same social context in order to be validly compared. Thus, in 1995 we selected nine sets of two males that had been calling within 2 m of each other, and in 1996 we selected six sets of two field-caught males that had been calling simultaneously in a small captive chorus. The mean difference in call duration between the long- and the short-caller in each set was 10.1 pulses per call in 1995 and 15.8 pulses per call in 1996 (Table 1); in laboratory experiments, female H. versicolor routinely base preferences on differences of as few as 2 pulses per call (15). Furthermore, the average call durations of individual males classified as long-callers did not overlap with the average call durations of individuals classified as short-callers (16). Long- and short-callers did not differ in body mass. External artificial fertilization allowed unambiguous assignment of paternity, and rotation of the egg-stripping of each female between the pair of males eliminated the possibility of effects of fertilization order (17). In 1995, each of nine gravid females was artificially crossed with a different set of males to generate nine maternal half-sibships. In 1996, each of 11 gravid females was artificially crossed with at least one set of males to generate 16 maternal half-sibships. All frogs were collected near a pond in Boone County, Missouri.

Table 1

Average calling performance of sires exhibiting long versus short calls. For each 1995 male, approximately 25 consecutive calls were analyzed from field recordings. For each 1996 male, at least 20 min of consecutive calls were analyzed from digitally collected data.

View this table:

Because the relative performance of different genotypes can vary significantly with environmental conditions (18), we reared the resulting tadpoles at two food levels, thereby creating an unfavorable and a favorable growth environment in which to compare the performance of offspring (19). Comparison of our results with those from field studies indicates that our high food treatment was a realistic approximation of conditions encountered in nature (20). Tadpoles from the crosses (1995, n = 538; 1996, n = 384) were raised individually in containers filled with 1.0 liter of water in the laboratory at the two food levels; 15 tadpoles per family were reared at each food level in 1995 and six tadpoles per family in 1996 (21). To assess offspring performance, we used several variables (22) that are important determinants of fitness in anurans, predicting future survival and age and size at maturity, which influence lifetime reproductive success (23).

Offspring of males with long calls always performed significantly better than or not significantly differently from offspring of males with short calls (24) (Table 2). In multivariate analyses where responses were combined to account for correlations among response variables (25), the main effect of call duration was significant at both food levels in 1996 and showed the same trend at both food levels in 1995 (Table 2), with offspring of males with long calls showing a general performance advantage over offspring of males with short calls. The probability of obtaining these four multivariate results that independently support the same directional hypothesis was calculated as P = 0.0008 (Table 2) with the use of a combined probability test (26). The multivariate tests therefore support the hypothesis that offspring performance is predicted by paternal call duration.

Table 2

Relative performance of offspring of males exhibiting long versus short calls. A shorter larval period is interpreted as better performance. For all other variables, larger values indicate better performance. NS, not significant. Dashes indicate data not collected in 1995.

View this table:

The specific benefits realized by offspring of long-callers differed among our experimental environments (Table 2). Because variation in the quality of the growth environment is predicted to influence the relation between larval growth and development (27), this difference in responses among environments is not unexpected. The consistency of the general benefit realized by offspring of long-callers in our experimental environments suggests that a general performance advantage may be applicable in other environments as well.

Overall, these results provide strong evidence that males with long calls relative to those of other males in the same social environment sired offspring of significantly higher phenotypic quality than males with short relative call durations. We can attribute these observed phenotypic differences to differences in paternal genetic contribution, because our comparison of maternal half-sibships controls for maternal genetic contributions and maternal effects. Thus, our results demonstrate that relative call duration reliably reflects genetic quality in H. versicolor. Our data suggest a genetic correlation between sire call duration and offspring performance, which implies that each trait has a heritable basis. The preference for long calls should, therefore, enable females to select high-quality mates and benefit indirectly through increased fitness of offspring. Because female H. versicolor do not gain direct benefits from their choice of mate, the indirect genetic benefits we have documented suggest good genes selection as a probable explanation for the evolution and maintenance of the female preference in this species.

  • * To whom correspondence should be addressed. E-mail: awelch{at}


View Abstract

Navigate This Article