Studies of mate choice typically assume that individuals prefer high quality mates and select them based on condition-dependent indicator traits. In species with bi-parental care, mutual mate choice is expected to result in assortative mating for quality. When assortment is not perfect, the lower quality pair members are expected to compensate by increased parental investment to secure their partner (positive differential allocation). This framework has been assumed to hold for monogamous species like the zebra finch (Taeniopygia guttata), but progress has been hampered by the difficulty to define individual quality. By combining multiple measures of causes (inbreeding, early nutrition) and consequences (ornaments, displays, fitness components) of variation in quality into a single principal component, we here show that quality variation can be quantified successfully. We further show that variation in quality indeed predicts individual pairing success, presumably because it reflects an individual's vigor or ability to invest in reproduction. However, despite high statistical power, we found no evidence for either assortative mating or for positive differential allocation. We suggest that zebra finch ornaments and displays are not sufficiently reliable for the benefits of choosiness to exceed the costs of competition for the putative best partner. To assess the generality of these findings unbiased quantification of signal honesty and preference strength is required, rather than selective reporting of significant results.
PCA of female zebra finches_table2
PCA of 219 female zebra finches based on six traits: fecundity, mass day 8, mass day 100, inbreeding coefficient (Fshort), brood size and beak color (BC).
PCA of male zebra finches_table2
PCA of 217 male zebra finches based on 10 traits: siring success, Natal brood size, Mass day 8, Mass day 100, Inbreeding_coefficient (Fshort), Beak color(BC), Cheek patch size, Courtship rate(meanroot11), Repertoire size, Amplitude.
Pearson r of 7 traits_population wide_table3
Test for assortative mating for various quality traits of males and females (seven traits) calculated at the population-wide. At the population level, sample size (n) indicates the number of pairs for which a Pearson correlation coefficient was calculated. ‘Fitness’ refers to the number of eggs laid (female fecundity) or fertilized (male siring success). PCA score refers to the estimate of overall quality (Table 2) based on the other breeding round, i.e. when not paired to the focal partner.
Pearson r of 7 traits_within aviary_table3
Test for assortative mating for various quality traits of males and females calculated at the within-aviary level. At the aviary level, the sample size indicates the number of aviaries for which a Pearson correlation coefficient was calculated; shown are the mean correlation coefficients for each trait. ‘Fitness’ refers to the number of eggs laid (female fecundity) or fertilized (male siring success). PCA score refers to the estimate of overall quality (Table 2) based on the other breeding round, i.e. when not paired to the focal partner.
Pearson r of Pairing status_table 4
Test for assortative mating for male versus female estimates of quality (PCA scores, Table 2) across pairs of various status categories (all pair bonds observed, all exclusively monogamous relationships, the subset of monogamous relationships that lasted until the end of a breeding round, and all polygamous relationships).
model testing the differential-allocation hypothesis_table5
Data for the linear mixed-effect model testing the differential-allocation hypothesis based on observations of the proportion of female incubation in 283 out of 292 lasting monogamous pairs that initiated breeding. The dependent variable is the relative count of female versus male nest visit (FIDInc and MIDinc) records (using the ‘cbind’ function in R) within each pair. The random effects male and female identity (ID) reflect the individual repeatability across different partners (variance component V), while the random effect pair ID controls for overdispersion in the binomial counts. The predictor of interest (quality difference) is the difference in estimated quality between the partners (female PCA score minus male PCA score:diff_FM).
Pairing success ~ quality_figure1
Data for the Figure1: Pairing success of individuals as a function of their quality (scored as the first principal component reflecting the quality indicators given in Table 2).
data for figure2
Data for figure2: Principal component scores reflecting individual quality of males and females that formed a pair bond during the first and second breeding round. Note that PC scores are calculated from data of the other breeding round.
pair status & quality_ FigS1
Data for figureS1: PC scores of monogamous, polygamous and unpaired individuals of both sexes. Pairing status during the first breeding round in relation to the PC score from the second breeding round (PC1_other_round); Pairing status from the second round in relation to the PC score from the first round (PC1_other_round).
pair status and quality_ FigS2
Data for FigS2: Principal component scores of mated pairs of different pairing status (as in Fig. 2, but both breeding rounds combined).
dynamic assortment in aviary_FigS3
Data for FigS3: Male and female quality (PC scores from the other breeding round) and level of assortative mating for PC scores (Pearson correlation coefficient r) in relation to the order of pairing within aviaries (AV). All pair bonds that were formed within an aviary were ranked by the order of pair formation (Pairing order within aviary).