Short- and long-term repeatability and pseudo-repeatability of bird song: sensitivity of signals to varying environments

Estimating heritability of song considering within-individual variance in a wild songbird: The collared flycatcher

Heritable genetic variation is a prerequisite for adaptive evolution; however, our knowledge about the heritability of plastic traits, such as behaviors, is scarce, especially in wild populations. In this study, we investigated the heritability of song traits in the collared flycatcher (Ficedula albicollis), a small oscine passerine with complex songs involved in sexual selection. We recorded the songs of 81 males in a natural population and obtained various measures describing the frequency, temporal organization, and complexity of each song. As we had multiple songs from each individual, we were able to statistically account for the first time for the effect of within-individual variance on the heritability of song. Heritability was calculated from the variance estimates of animal models relying on a genetic similarity matrix based on Single Nucleotide Polymorphism screening. Overall, we found small additive genetic variance and heritability values in all song traits, highlighting the role of environmental factors in shaping bird song.

Intrinsic individual variation in daily activity onset and plastic responses on temporal but not spatial scales in female great tits

In a variety of species, individuals appear to be consistent in the daily timing of their activity onset. Such consistent among-individual differences can result from both intrinsic factors, as individuals may e.g. differ genetically, and extrinsic factors, as the environment may vary on spatial and temporal scales. However, previous studies typically did not differentiate between their respective contributions on individual variation in the timing of activities. Here, we repeatedly measured the onset of activity in female great tits (Parus major) on consecutive days during the egg laying phase of the breeding season in four consecutive years. Subsequently, we used a variance partitioning analysis in order to determine which part of the total variation could be attributed to intrinsic (female identity) and extrinsic (nest box identity) factors. Overall, 27% of the total variation could be attributed to female identity. In addition, we found temporal variation in the activity onset, indicating that individuals can plastically adjust their timing. Yet despite their general ability to change the timing of activities over time, spatial environmental factors did not contribute significantly to the observed variation. Individuals may choose a habitat that matches the preferred timing of activities, or might not benefit from adjusting their timing to environmental factors that might vary on spatial scales.

Beyond illness: Variation in haemosporidian load explains differences in vocal performance in a songbird

In animal communication, signals are expected to evolve to be honest, so that receivers avoid being manipulated by signalers. One way that signals can evolve to be honest is for them to be costly, with only high-quality individuals being able to bear the costs of signal expression. It has been proposed that parasites can introduce costs that affect the expression of sexually selected traits, and there is evidence to support the role of parasitism in modulating animal behavior. If host infection status or intensity is found to relate to differences in signal expression, it may indicate a fitness cost that mediates honesty of signals. Birdsong is a good model for testing this, and physically challenging songs representing complex motor patterns provide a good example of sexually selected traits indicating individual condition. We performed a field study to evaluate the relationship between song performance and avian malaria infection in a common songbird. Previous work on this subject has almost always evaluated avian malaria in terms of binary infection status; however, parasitemia-infection intensity-is rarely assessed, even though differences in parasite load may have profound physiological consequences. We estimated parasitemia levels by using real-time PCR. We found that birds with higher parasitemia displayed lower vocal performance, providing evidence that this song trait is an honest signal of parasitic load of haemosporidian parasites. To our knowledge, this study links parasite load and the expression of a sexually selected trait in a way that has not been addressed in the past. Studies using song performance traits and parasitemia offer an important perspective for understanding evolution of characters via sexual selection.

Low repeatability of aversive learning in zebrafish (Danio rerio)

Aversive learning - avoiding certain situations based on negative experiences - can profoundly increase fitness in animal species, yet no studies have systematically quantified its repeatability. Therefore, we assessed the repeatability of aversive learning by conditioning approximately 100 zebrafish (Danio rerio) to avoid a colour cue associated with a mild electric shock. Across eight different colour conditions, zebrafish did not show consistent individual differences in aversive learning (R=0.04). Within conditions, when zebrafish were conditioned to the same colour, blue conditioning was more repeatable than green conditioning (R=0.15 and R=0.02). Overall, aversive learning responses of zebrafish were weak and variable. We speculate that the effect of aversive learning might have been too weak to quantify consistent individual differences, or directional selection might have eroded additive genetic variance. We also discuss how confounded repeatability assays and publication bias could have inflated estimates of repeatability in the literature.

The effect of social environment on bird song: listener-specific expression of a sexual signal

Animal signals should consistently differ among individuals to convey distinguishable information about the signalers. However, behavioral display signals, such as bird song are also loaded with considerable within-individual variance with mostly unknown function. We hypothesized that the immediate social environment may play a role in mediating such variance component, and investigated in the collared flycatcher (Ficedula albicollis) if the identity and quality of listeners could affect song production in signalers. After presenting territorial males with either a female or male social stimulus, we found in the subsequent song recordings that the among-stimulus effects corresponded to non-zero variance components in several acoustic traits indicating that singing males are able to plastically adjust their songs according to stimulus identity. Male and female stimuli elicited different responses as the identity of the female stimuli affected song complexity only, while the identity of male stimuli altered also song length, maximum frequency, and song rate. The stimulus-specific effect on song in some cases decreased with time, being particularly detectable right after the removal of the stimulus and ceasing later, but this pattern varied across the sex of the stimulus and the song traits. We were able to identify factors that can explain the among-stimulus effects (e.g., size and quality of the stimuli) with roles that also varied among song traits. Our results confirm that the variable social environment can raise considerable variation in song performance, highlighting that within-individual plasticity of bird song can play important roles in sexual signaling.

Sequential organization of birdsong: relationships with individual quality and fitness

Many vocalizing animals produce the discrete elements of their acoustic signals in a specific sequential order, but we know little about the biological relevance of this ordering. For that, we must characterize the degree by which individuals differ in how they organize their signals sequentially and relate these differences to variation in quality and fitness. In this study, we fulfilled these tasks in male collared flycatchers (Ficedula albicollis). We characterized the sequential order of syllables with a network analysis approach and studied the consistency of network variables on distinct time scales (within day, between days, and between years), and assessed their relationship with such quality indicators like age, body condition, arrival date, and fitness related proxies like survival to the next year and pairing success. We found that the syllables were associated nonrandomly with one another and both the frequency differences of consecutive syllables and the number of motif types were higher in the original than in randomized syllable sequences. Average degree and small-worldness showed considerable among-individual differences and decreasing repeatability with increasing time scale. Furthermore, we found relationships between male age and average degree among and within individuals. Accordingly, older males produce syllable sequences by using common syllables less often than younger individuals. However, the network variables showed no relationship with fitness-related variables. In conclusion, the sequential organization of birdsong has the potential to encode individual-specific characteristics, which thus could be used as signal in social interactions and thus potentially could be subject to sexual selection.

Zebra finches bi-directionally selected for personality differ in repeatability of corticosterone and testosterone

Consistent between-individual differences in behaviour have been documented across the animal kingdom. Such variation between individuals has been shown to be the basis for selection and to act as a pacemaker for evolutionary change. Recently, equivocal evidence suggests that such consistent between-individual variation is also present in hormones. This observation has sparked interest in understanding the mechanisms shaping individual differences, temporal consistency and heritability of hormonal phenotypes and to understand, if and to what extent hormonal mechanisms are involved in mediating consistent variation in behaviour between individuals. Here, we used zebra finches of the fourth generation of bi-directionally selected lines for three independent behaviours: aggression, exploration and fearlessness. We investigated how these behaviours responded to artificial selection and tested their repeatability. We further tested for repeatability of corticosterone and testosterone across and within lines. Moreover, we are presenting the decomposed variance components for within-individual variance (i.e. flexibility) and between-individual variance (i.e. more or less pronounced differences between individuals) and investigate their contribution to repeatability estimates. Both hormones as well as the exploration and fearlessness but not aggressiveness, were repeatable. However, variance components and hence repeatability differed between lines and were often lower than in unselected control animals, mainly because of a reduction in between-individual variance. Our data show that artificial selection (including active selection and genetic drift) can affect the mean and variance of traits. We stress the importance for understanding how variable a trait is both between and within individuals to assess the selective value of a trait.

Temporal autocorrelation: a neglected factor in the study of behavioral repeatability and plasticity

Quantifying individual variation in labile physiological or behavioral traits often involves repeated measures through time, so as to test for consistency of individual differences (often using repeatability, “R”) and/or individual differences in trendlines over time. Another form of temporal change in behavior is temporal autocorrelation, which predicts observations taken closely together in time to be correlated, leading to nonrandom residuals about individual temporal trendlines. Temporal autocorrelation may result from slowly changing internal states (e.g., hormone or energy levels), leading to slowly changing behavior. Autocorrelation is a well-known phenomenon, but has been largely neglected by those studying individual variation in behavior. Here, we provide two worked examples which show substantial temporal autocorrelation (r > 0.4) is present in spontaneous activity rates of guppies (Poecilia reticulata) and house mice (Mus domesticus) in stable laboratory conditions, even after accounting for temporal plasticity of individuals. Second, we show that ignoring autocorrelation does bias estimates of R and temporal reaction norm variances upwards, both in our worked examples and in separate simulations. This bias occurs due to the misestimation of individual-specific means and slopes. Given the increasing use of technologies that generate behavioral and physiological data at high sampling rates, we can now study among- and within-individual changes in behavior in more detailed ways, including autocorrelation, which we discuss from biological and methodological perspectives and provide recommendations and annotated R code to help researchers implement these models on their data.

Repeatability of signalling traits in the avian dawn chorus

Background

Birdsong, a key model in animal communication studies, has been the focus of intensive research. Song traits are commonly considered to reflect differences in individual or territory quality. Yet, few studies have quantified the variability of song traits between versus within individuals (i.e. repeatability), and thus whether certain song traits indeed provide reliable individual-specific information. Here, we studied the dawn chorus of male great tits (Parus major) to determine if key song traits are repeatable over multiple days and during different breeding stages. Additionally, we examined whether repeatability was associated with exploration behaviour, a relevant personality trait. Finally, we tested if variation in song traits could be explained by breeding stage, lowest night temperature, and exploration behaviour.

Results

We show that the start time of an individual’s dawn song was indeed repeatable within and across breeding stages, and was more repeatable before, than during, their mate’s egg laying stage. Males started singing later when the preceding night was colder. Daily repertoire size was repeatable, though to a lesser extent than song start time, and no differences were observed between breeding stages. We did not find evidence for an association between exploration behaviour and variation in dawn song traits. Repertoire composition, and specifically the start song type, varied across days, but tended to differ less than expected by chance.

Conclusions

Our findings that individuals consistently differ in key song traits provides a better understanding of the information receivers can obtain when sampling songs of different males. Surprisingly, start time, despite being influenced by a highly variable environmental factor, appeared to be a more reliable signal of individual differences than repertoire size. Against expectation, singers were more repeatable before than during their mate’s egg laying stage, possibly because before egg laying, females are less constrained to move around unguarded and thus may then already sample (and compare) different singers. Combining repeated dawn song recordings with spatial tracking could reveal if the sampling strategies of receivers are indeed important drivers of repeatability of song traits. Such a complementary approach will further advance our insights into the dynamics and evolution of animal signalling systems.

MHC-mediated sexual selection on birdsong: Generic polymorphism, particular alleles and acoustic signals

Several hypotheses predict that the major histocompatibility complex (MHC) drives mating preference in females. Olfactory, colour or morphological traits are often found as reliable signals of the MHC profile, but the role of avian song mediating MHC-based female choice remains largely unexplored. We investigated the relationship between several MHC and acoustic features in the collared flycatcher (Ficedula albicollis), a European passerine with complex songs. We screened a fragment of the class IIB second exon of the MHC molecule, of which individuals harbour 4-15 alleles, while considerable sequence diversity is maintained at the population level. To make statistical inferences from a large number of comparisons, we adopted both null-hypothesis testing and effect size framework in combination with randomization procedures. After controlling for potential confounding factors, neither MHC allelic diversity nor the presence of particular alleles was associated remarkably with the investigated qualitative and quantitative song traits. Furthermore, genetic similarity among males based on MHC sequences was not reflected by the similarity in their song based on syllable content. Overall, these results suggest that the relationship between features of song and the allelic composition and diversity of MHC is not strong in the studied species. However, a biologically motivated analysis revealed that individuals that harbour an MHC allele that impairs survival perform songs with broader frequency range. This finding suggests that certain aspects of the song may bear reliable information concerning the MHC profile of the individuals, which can be used by females to optimize mate choice.

How should we interpret estimates of individual repeatability?

Individual repeatability (R), defined as the proportion of observed variance attributable to among-individual differences, is a widely used summary statistic in evolutionarily motivated studies of morphology, life history, physiology and, especially, behaviour. Although statistical methods to estimate R are well known and widely available, there is a growing tendency for researchers to interpret R in ways that are subtly, but importantly, different. Some view R as a property of a dataset and a statistic to be interpreted agnostically with respect to mechanism. Others wish to isolate the contributions of 'intrinsic' and/or 'permanent' individual differences, and draw a distinction between true (intrinsic) and pseudo-repeatability arising from uncontrolled extrinsic effects. This latter view proposes a narrower, more mechanistic interpretation, than the traditional concept of repeatability, but perhaps one that allows stronger evolutionary inference as a consequence (provided analytical pitfalls are successfully avoided). Neither perspective is incorrect, but if we are to avoid confusion and fruitless debate, there is a need for researchers to recognise this dichotomy, and to ensure clarity in relation to how, and why, a particular estimate of R is appropriate in any case.