S. Van Dongen

The statistical analysis of fluctuating asymmetry: REML estimation of a mixed regression model

The unbiased estimation of fluctuating asymmetry (FA) requires independent repeated measurements on both sides. The statistical analysis of such data is currently performed by a two‐way mixed ANOVA analysis. Although this approach produces unbiased estimates of FA, many studies do not utilize this method. This may be attributed in part to the fact that the complete analysis of FA is very cumbersome and cannot be performed automatically with standard statistical software. Therefore, further elaboration of the statistical tools to analyse FA should focus on the usefulness of the method, in order for the correct statistical approaches to be applied more regularly.

Fluctuating asymmetry increases with habitat disturbance in seven bird species of a fragmented afrotropical forest

We studied fluctuating asymmetry (FA) in the tarsus length of seven forest-restricted bird species, two of which are globally critically endangered, in three indigenous forest remnants of a recently fragmented, afrotropical biodiversity hot spot. Based on mixed regression analysis and an extension of Levenes test, individuals from the most degraded fragment showed four- to sevenfold higher asymmetry levels compared to those from the least degraded one, with intermediate levels in the moderately disturbed fragment. When comparing contemporary FA levels with measurements of museum specimens collected 50 years ago, we found highly significant increases in asymmetry in the most degraded fragment but no differences in the least degraded one. These strikingly parallel spatial and temporal patterns across species confirm that repeated measurements of FA can provide a sensitive early warning system for monitoring stress effects in highly threatened ecosystems.

Fluctuating asymmetry as an animal welfare indicator : A review of methodology and validity

It has been suggested that fluctuating asymmetry (FA) reflects an animals ability to cope with the sum of challenges during its growing period and, thus, is a potential welfare indicator. In this review we investigate the evidence of associations between FA and other welfare indicators measured at the level of the individual and of effects of welfare-relevant environmental conditions on FA in populations of captive birds and mammals including humans. As the question of validity cannot be treated independently from the quality of the available data, first a checklist for the proper measurement and analysis of FA is drafted and used to evaluate the methodological quality of the various studies. We recommend this checklist to be used as a standard for future FA studies. We found 17 relevant studies on associations between FA and other welfare indicators, and 36 studies on effects of welfare-relevant factors on FA. Frequent methodological shortcomings or insufficient methodological information allow for only cautious conclusions. The proportion of significant results supporting the link between higher FA and poorer welfare is only moderately high. Independent from statistical significance, almost all studies found the relationship between FA and welfare to be prevailingly in the expected direction. FA is a promising measure of animal welfare, despite a great number of open questions, e.g. relating to the ontogeny of FA or its sensitivity to various stressors. The considerable potential of FA as a welfare indicator makes it worthwhile to pursue more intensely validation studies as well as applied studies. These studies should pay particular attention to an appropriate methodological approach.

Developmental instability and inbreeding in natural bird populations exposed to different levels of habitat disturbance

As habitat disturbance and inbreeding increasingly stress natural populations, ecologists are in urgent need of simple estimators to measure their impact. It has been argued that developmental instability (DI) could be such a measure. Observed associations between DI and environmental or genetic stress, however, are largely inconsistent. We here test whether an interaction between habitat disturbance and inbreeding could, at least partly, explain these discordant patterns. We therefore studied individual estimates of fluctuating asymmetry (FA) and of inbreeding in three populations of the critically endangered Taita thrush that are differentially exposed to habitat disturbance following severe forest fragmentation. As predicted, the relationship between DI and inbreeding was pronounced under high levels of disturbance, but weak or nonexistent under less disturbed conditions. Examining this relationship with mean d2, an allelic distance estimator assumed to reflect ancestral inbreeding, did not reveal any significant trend, hence suggesting that inbreeding effects in the Taita thrush are fairly recent.

GENETIC POPULATION STRUCTURE OF THE WINTER MOTH (OPEROPHTERA BRUMATA L.) (LEPIDOPTERA, GEOMETRIDAE) IN A FRAGMENTED LANDSCAPE

As a result of human activities natural environments have been altered in many different ways. One important effect of human disturbance is the fragmentation of natural habitats. As a consequence, genetic differentiation among habitat islands is expected to increase, whereas within-area genetic diversity is expected to decrease. Indirect estimates from allozyme polymorphisms are used to investigate the effects of habitat fragmentation in the winter moth on a very small geographical scale. We demonstrated that genetic differentiation increased whereas genetic diversity decreased with fragmentation, with habitat patches isolated by only a few hundred metres up to 3 km. These results were even more striking considering that no genetic differentiation was detected at a larger scale (10–40 km). This pattern of distribution of genetic variation is in agreement with temporarily variable densities and gene flow levels which prevent an equilibrium being reached between genetic drift and gene flow. Consequently the effects of fragmentation probably remain limited.

The evolutionary potential of developmental instability.

Whether or not developmental instability (DI) has evolutionary potential is subject to much debate. Generally, studies fail to detect significant heritability for fluctuating asymmetry (FA), a trait assumed to reflect DI. In addition, between‐trait correlations in FA are low, suggesting that DI is trait‐ rather than individual‐specific. Among the various attempts to explain these patterns, the overall weak correlation between FA and DI at the individual level has received most attention. Presently, the concept of hypothetical repeatability (R) of individual FA allows us to correct for this weak relationship, transforming patterns of FA into unbiased patterns of DI. By applying R to data presented in the literature, we show that heritability of DI remains lower than predicted but between‐trait correlations in DI substantially increase after transformation. We further provide evidence that DI changes from a trait‐ to an individual‐specific property with higher values of R. As increasing hypothetical repeatability might co‐occur with increased environmental or genetic stress, we discuss the potential implications of our results for the study of evolution of stress resistance. From this we conclude that there is an urgent need for studies that compare the evolutionary potential of developmental instability under a variety of stress conditions.

Evidence for organism-wide asymmetry in five bird species of a fragmented afrotropical forest

Empirical evidence for between-trait correlation in fluctuating asymmetry (FA) at the individual level is generally lacking or contradictory. Yet the assumption of organism-wide asymmetry, estimated by the asymmetry of any given trait, is inherent to most studies that use FA as a measure of developmental stability (DS). A commonly assumed reason for this weak between–trait correlation is the low repeatability of individual, single–trait asymmetry. In this paper we describe high repeatability and significant between–trait correlation in population– and individual–level FA in five afrotropical bird species inhabiting a fragmented cloud forest. Absence of anti–symmetry and of between–trait correlation in signed FA levels permits us to translate the observed patterns into the presumed underlying DS, using the concept of hypothetical repeatability. This correction, which has not been applied before in this context, proved adequate as it yielded correlations comparable to those found at the population level.

Stabilizing selection on blue tit fledgling mass in the presence of sparrowhawks

Like British great tits, Belgian blue tits have a lower winter body mass when sparrowhawks are present. Since body mass affects manoeuvrability in small birds, tits may balance the risks of starvation and the risk of hawk predation by varying the amount of extra fat carried during winter. Predation pressure by sparrowhawks on young and inexperienced fledglings is at least as intense as that on the adults during winter. We therefore expected that tit fledgling body mass could also be reduced in the presence of sparrowhawks. In the years after one pair of sparrowhawks settled in a study plot, the mean body mass of blue tit fledglings was lower compared with that in years when there were no sparrowhawks. Furthermore, the shape of the curve relating juvenile survival to fledging mass changed, because the survival of the heaviest fledglings was reduced, which altered the selection differential of juvenile survival as a function of body mass from directional to stabilizing. Of seven published studies on the fledgling body mass–survival relation in tits, all three of the studies conducted in the absence of sparrowhawks showed the highest survival rates for the heaviest young, whereas in all four studies with sparrowhawks present this was no longer the case.

MATE SELECTION BY MALE WINTER MOTHS OPEROPHTERA BRUMATA (LEPIDOPTERA, GEOMETRIDAE): ADAPTIVE MALE CHOICE OR FEMALE CONTROL?

Mate choice is an important component of sexual selection. It is expected to evolve if the benefits of choice outweigh the costs. Yet, the relative importance of costs and benefits in the evolution of mate choice remain poorly understood. In this study we present experimental evidence for adaptive mate choice by males, but not females, in the winter moth. In a series of experiments we show that: (a) males have a higher probability of attempting to copulate, and consequently also a higher probability to copulate with a larger, more fecund female; (b) if males are given the choice between two females they are more likely to copulate with the larger female; and (c) females do not seem to show any mate discrimination. A sample of winter moths collected in copula in the field did not show any assortative mating for body size. This is the first demonstration of male choice in a moth species with chemical communication. This choice is possibly based on variation in female pheromone quality and/or quantity. We argue that the relatively higher variation in female quality and the limited number of male matings probably have led to male choosiness despite a strongly male biased operational sex ratio. This is consistent with recent studies indicating that choice and competition may occur more frequently in the same sex than previously thought.

How should we bootstrap allozyme data

The use of the bootstrap to estimate the distribution of statistics from allozyme data is examined. The different loci are often used as the unit of resampling. Since the interpretation and validity of the bootstrap is affected by the unit of resampling, and since resampling over loci, in most practical cases, does not conform to the three basic assumptions of the bootstrap, this method should be avoided. Resampling over individual genotypes may provide a valid alternative approach.The use of the bootstrap to estimate the distribution of statistics from allozyme data is examined. The different loci are often used as the unit of resampling. Since the interpretation and validity of the bootstrap is affected by the unit of resampling, and since resampling over loci, in most practical cases, does not conform to the three basic assumptions of the bootstrap, this method should be avoided. Resampling over individual genotypes may provide a valid alternative approach.