Limits...
No band color effects on male courtship rate or body mass in the zebra finch: four experiments and a meta-analysis.

Seguin A, Forstmeier W - PLoS ONE (2012)

Bottom Line: Red-banded males were found to have higher fitness than green-banded males, and most empirical evidence suggests that this effect is mediated by female mating preferences rather than by male-male competition.Combining this new experimental data with all the published evidence in a meta-analysis shows that color bands seem to affect neither male courtship rate (average effect size d=0.02) nor male body mass (d=-0.07).The present case is a reminder that replication of experiments lies at the heart of distinguishing between real effects and false positive findings.

View Article: PubMed Central - PubMed

Affiliation: Max Planck Institute for Ornithology, Seewiesen, Germany.

ABSTRACT
Replication of experiments is essential for distinguishing real effects from type 1 errors and idiosyncrasies. One of the most replicated experiments in behavioral ecology is the presumed manipulation of male attractiveness in zebra finches by adding red or green color bands. Red-banded males were found to have higher fitness than green-banded males, and most empirical evidence suggests that this effect is mediated by female mating preferences rather than by male-male competition. A recent study, however, reported that color bands affected male courtship rate and body mass independently of female behavior. If this effect was real, some earlier findings of female preferences and maternal effects on offspring traits could potentially be reinterpreted as being mediated indirectly via effects on male behavior. This new perspective seems appealing also in light of a growing interest in bi-directional feedback mechanisms between endocrinology and ornamentation. However, here we report four independent failures to replicate this effect of color bands on courtship rate and body mass. Combining this new experimental data with all the published evidence in a meta-analysis shows that color bands seem to affect neither male courtship rate (average effect size d=0.02) nor male body mass (d=-0.07). The present case is a reminder that replication of experiments lies at the heart of distinguishing between real effects and false positive findings.

Show MeSH
Effect sizes (d ±95% CI) of the color band treatments on (A) male courtship rate and (B) male body mass.A positive effect size refers to red-banded males obtaining higher courtship rate or mass compared to green-banded males. For individual experiments, the size of the square reflects sample size, and for the overall estimate the diamond marks the mean and the 95% CI. Longitudinal analyses as marked with (L); the remainder are cross-sectional analyses. The experiments in (a) refer to Pariser et al. [25], Ratcliffe and Boag [24] inexperienced birds (1) and experienced birds (2), Gleeson [39] individual housing and group housing, Burley et al. [14], wild-type population (W) of the present study, domesticated population (D) of the present study, juvenile (juv) and adult (ad) birds, with housing in uni-color and mixed color groups, and housing in duos or larger groups; (b) Zann [16] at the start (t1) and end (t2) of the breeding season, Cuthill et al. [23] with mass measured at either dawn or dusk, Schuett and Dall [7], and others as above.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC3369886&req=5

pone-0037785-g005: Effect sizes (d ±95% CI) of the color band treatments on (A) male courtship rate and (B) male body mass.A positive effect size refers to red-banded males obtaining higher courtship rate or mass compared to green-banded males. For individual experiments, the size of the square reflects sample size, and for the overall estimate the diamond marks the mean and the 95% CI. Longitudinal analyses as marked with (L); the remainder are cross-sectional analyses. The experiments in (a) refer to Pariser et al. [25], Ratcliffe and Boag [24] inexperienced birds (1) and experienced birds (2), Gleeson [39] individual housing and group housing, Burley et al. [14], wild-type population (W) of the present study, domesticated population (D) of the present study, juvenile (juv) and adult (ad) birds, with housing in uni-color and mixed color groups, and housing in duos or larger groups; (b) Zann [16] at the start (t1) and end (t2) of the breeding season, Cuthill et al. [23] with mass measured at either dawn or dusk, Schuett and Dall [7], and others as above.

Mentions: The combined effect size of color bands on male courtship rate was estimated at d = 0.016 (95% CI -0.252–0.283). Twelve effect size estimates from six different populations contributed to this overall estimate (Fig. 5a). Effect size was unrelated to male housing density (rs = −0.51, n = 11, p = 0.11, trend against the prediction). Heterogeneity variance among estimates was 0.085, which was still short of significance (p = 0.075).


No band color effects on male courtship rate or body mass in the zebra finch: four experiments and a meta-analysis.

Seguin A, Forstmeier W - PLoS ONE (2012)

Effect sizes (d ±95% CI) of the color band treatments on (A) male courtship rate and (B) male body mass.A positive effect size refers to red-banded males obtaining higher courtship rate or mass compared to green-banded males. For individual experiments, the size of the square reflects sample size, and for the overall estimate the diamond marks the mean and the 95% CI. Longitudinal analyses as marked with (L); the remainder are cross-sectional analyses. The experiments in (a) refer to Pariser et al. [25], Ratcliffe and Boag [24] inexperienced birds (1) and experienced birds (2), Gleeson [39] individual housing and group housing, Burley et al. [14], wild-type population (W) of the present study, domesticated population (D) of the present study, juvenile (juv) and adult (ad) birds, with housing in uni-color and mixed color groups, and housing in duos or larger groups; (b) Zann [16] at the start (t1) and end (t2) of the breeding season, Cuthill et al. [23] with mass measured at either dawn or dusk, Schuett and Dall [7], and others as above.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC3369886&req=5

pone-0037785-g005: Effect sizes (d ±95% CI) of the color band treatments on (A) male courtship rate and (B) male body mass.A positive effect size refers to red-banded males obtaining higher courtship rate or mass compared to green-banded males. For individual experiments, the size of the square reflects sample size, and for the overall estimate the diamond marks the mean and the 95% CI. Longitudinal analyses as marked with (L); the remainder are cross-sectional analyses. The experiments in (a) refer to Pariser et al. [25], Ratcliffe and Boag [24] inexperienced birds (1) and experienced birds (2), Gleeson [39] individual housing and group housing, Burley et al. [14], wild-type population (W) of the present study, domesticated population (D) of the present study, juvenile (juv) and adult (ad) birds, with housing in uni-color and mixed color groups, and housing in duos or larger groups; (b) Zann [16] at the start (t1) and end (t2) of the breeding season, Cuthill et al. [23] with mass measured at either dawn or dusk, Schuett and Dall [7], and others as above.
Mentions: The combined effect size of color bands on male courtship rate was estimated at d = 0.016 (95% CI -0.252–0.283). Twelve effect size estimates from six different populations contributed to this overall estimate (Fig. 5a). Effect size was unrelated to male housing density (rs = −0.51, n = 11, p = 0.11, trend against the prediction). Heterogeneity variance among estimates was 0.085, which was still short of significance (p = 0.075).

Bottom Line: Red-banded males were found to have higher fitness than green-banded males, and most empirical evidence suggests that this effect is mediated by female mating preferences rather than by male-male competition.Combining this new experimental data with all the published evidence in a meta-analysis shows that color bands seem to affect neither male courtship rate (average effect size d=0.02) nor male body mass (d=-0.07).The present case is a reminder that replication of experiments lies at the heart of distinguishing between real effects and false positive findings.

View Article: PubMed Central - PubMed

Affiliation: Max Planck Institute for Ornithology, Seewiesen, Germany.

ABSTRACT
Replication of experiments is essential for distinguishing real effects from type 1 errors and idiosyncrasies. One of the most replicated experiments in behavioral ecology is the presumed manipulation of male attractiveness in zebra finches by adding red or green color bands. Red-banded males were found to have higher fitness than green-banded males, and most empirical evidence suggests that this effect is mediated by female mating preferences rather than by male-male competition. A recent study, however, reported that color bands affected male courtship rate and body mass independently of female behavior. If this effect was real, some earlier findings of female preferences and maternal effects on offspring traits could potentially be reinterpreted as being mediated indirectly via effects on male behavior. This new perspective seems appealing also in light of a growing interest in bi-directional feedback mechanisms between endocrinology and ornamentation. However, here we report four independent failures to replicate this effect of color bands on courtship rate and body mass. Combining this new experimental data with all the published evidence in a meta-analysis shows that color bands seem to affect neither male courtship rate (average effect size d=0.02) nor male body mass (d=-0.07). The present case is a reminder that replication of experiments lies at the heart of distinguishing between real effects and false positive findings.

Show MeSH