SEARCH

SEARCH BY CITATION

Keywords:

  • developmental instability;
  • fluctuating asymmetry;
  • habitat disturbance;
  • inbreeding;
  • microsatellite DNA marker

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Bibliography

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.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Bibliography

Duplicate copies of morphological structures such as arms, legs, wings, etc. show minor, yet significant, nondirectional differences ( Palmer & Strobeck, 1986). This so-called fluctuating asymmetry (FA) in bilateral expression ( Ludwig, 1932) cannot be explained by either genotypic or environmental differences, as the development of bilateral trait sides is ensured by the same genotype under identical environmental conditions ( Reeve, 1960). Rather, the observed levels of FA are believed to reflect the inability of an individual to buffer its development against random disturbances (developmental instability, DI) and, hence, to accurately express its expected phenotype ( Palmer & Strobeck, 1986). Although the magnitude of the effect varies, individual- and population levels of FA often correlate positively with various estimates of ‘environmental stress’, both in controlled laboratory experiments and under natural conditions ( Leung & Forbes, 1996; but see Bjorksten et al., 2000 ). Furthermore, many studies have reported a positive association between FA and estimates of inbreeding, suggesting that higher levels of inbreeding reflect higher levels of ‘genetic stress’ ( Soulé, 1979; Vrijenhoek & Lerman, 1982; Palmer & Strobeck, 1986). Inconsistency in the pattern ( Britten, 1996; Völlestad et al., 1999 ) and uncertainty about its causality ( Mitton, 1993), however, have led to a controversy over the generality of the latter relationship ( Leary et al., 1984 ; Clarke, 1993).

Discordance in the observed association between FA and inbreeding might have different, nonexclusive causes. First, the association might be obscured due to the typically weak association between single-trait individual FA and the underlying DI ( Whitlock, 1996). Such downward bias can be corrected for by dividing the observed correlation by the hypothetical repeatability R ( Whitlock, 1998). Secondly, the magnitude of the inbreeding effect might be small relative to the effect of other genetic processes ( Clarke, 1993), or may vary between ontogenetic stages and, thus, between traits that develop at different ages (see Clarke, 1998 for a comparable rationale for the overall weak between-trait correlation in individual FA). Thirdly, the magnitude of the inbreeding effect might depend on the level of environmental stress that organisms experience during their development. Energetic stress has previously been shown to affect developmental precision ( Nilsson, 1994; Palmer, 1996, and references therein). If the relationship between stress and developmental instability is affected by the genotype, the magnitude of the association between FA and inbreeding may, itself, depend on how much stress is experienced during development ( Palmer, 1996). In particular, the association has been hypothesized, yet not tested, to be pronounced under stressful conditions, but weak or nonexistent under more relaxed ones ( Palmer, 1996).

We studied the relationship between individual FA estimates and individual inbreeding coefficients (defined as the probability that both alleles at any locus are identical by descent; Falconer & Mackay, 1996) in 237 free-living Taita thrushes (Turdus helleri). This critically endangered, Kenyan forest endemic currently survives in three spatially ( Lens et al., 1999 ) and genetically ( Galbusera et al., 2000 ) isolated populations that are differentially exposed to habitat disturbance following severe forest fragmentation since the 1960s ( Brooks et al., 1998 ). Elsewhere we showed that population level FA was positively correlated with the level of habitat disturbance in the Taita thrush and six sympatric bird species ( Lens et al., 1999 ). In this paper we use genetic information from microsatellite-DNA markers to estimate individual levels of recent inbreeding ( Ritland, 1996) and a measure of genomic diversity (internal distance) assumed to reflect inbreeding events deeper in the pedigree ( Goldstein et al., 1995 ; Slatkin, 1995). As human impact caused between-population variation in the level of habitat disturbance, but not in mean inbreeding, we can test the interaction between both factors.

Individual levels of FA are calculated for three traits that develop at different ages during ontogeny, and observed relationships with inbreeding are translated into patterns of the presumed underlying DI by estimating the hypothetical repeatability of FA ( Whitlock, 1996, 1998). Based on these data, we test the prediction that the association between DI and inbreeding, both recent and ancestral, depends on the level of habitat disturbance to which individuals are exposed during ontogeny. Hence, in statistical terms, we test the interaction between habitat disturbance and inbreeding, rather than the effect of inbreeding per se, on FA.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Bibliography

Level of habitat disturbance

Data were collected in the only three indigenous forest fragments of the Taita Hills (south-east Kenya, 03°20′S, 38°15′E) that are inhabited by the endemic Taita thrush. In fragment CH (50 ha), NG (92 ha) and MB (220 ha), we established transects of 40 equally spaced points and measured the composition and structure of trees, shrubs and leaf litter as indicators of forest quality ( Table 1). Fragment CH (most disturbed) showed lower biomass, lower stem density, less tree species, lower tree diversity, lower tree equitability, less closed canopy cover, less open shrub layer, smaller leaf litter cover, and less open herbaceous layer than fragment MB (least disturbed), with intermediate values in fragment NG. Likewise, survival rates of Taita thrushes were the lowest in CH, intermediate in NG and highest in MB ( Table 1; Fisher’s Exact Test, one-tailed: P=0.03). Thus, while the three fragments might have differed in aspects other than habitat disturbance, population survival rates were inversely related to the level of disturbance. A recently developed assignment test based on individual genotypes ( Davies et al., 1999 ) confirmed that all Taita thrushes were captured in their natal fragment ( Galbusera et al., 2000 ) and, thus, were correctly assigned to the levels of habitat disturbance experienced during development.

Table 1.  Vegetation parameters and survival rates in three indigenous forest fragments. Thumbnail image of

Individual estimates of FA and DI

Between July 1996 and September 1999, each of four ringers (unaware of the hypothesis tested) measured Taita thrushes in fragments CH, NG and MB. Two measurements (left-right-left-right or right-left-right-left, with slide calipers reset to zero after each measurement) were made of left and right tarsus length (TL), outer rectrix length (RL) and postocular patch width (PW), per individual (details in Lens & Van Dongen, 1999). Although tarsus length was measured for all 237 individuals, traits RL and PW were measured in a subsample of catches only, due to the ruffled plumage and moulting stage of some individuals.

Bilateral trait asymmetry was analysed with the use of mixed regression analysis (with restricted maximum likelihood parameter estimation; REML), by the following procedure. First, we examined whether variance due to measurement error (ME) was heterogeneously distributed between the three populations (e.g. due to consistent differences in measurement accuracy between ringers), for each trait separately. As this was not the case (likelihood ratio tests: all P > 0.05), we estimated a single ME component for each trait. Secondly, we separated ME from ‘real’ FA (i.e. variance components of the random side effect), and tested for the presence of directional asymmetry by F-statistics (adjusting the denominator degrees of freedom by Satterthwaite’s formula; Verbeke & Molenberghs, 1997). Thirdly, we tested the significance of FA by comparing the likelihood of models with and without random side effect. Fourthly, we calculated unbiased FA values per individual as the variance components of the slopes of the individual regression lines in the mixed regression model (for methodological details see Palmer & Strobeck, 1986; Van Dongen et al., 1999 ). Individual FA estimates were log transformed when applied in hypothesis testing.

Patterns of individual FA were translated into patterns of DI through division by the hypothetical repeatability R, an estimate of the proportion of variation in FA that reflects between-individual variation in DI (sensuWhitlock, 1996); R was calculated as 1 – (((VFA + VME) × (π − 2)/π)/V|FA|), where VFA, V|FA| and VME are the variances of signed and unsigned FA and measurement error, respectively ( Van Dongen, 1998).

Individual inbreeding coefficients

Taita thrushes were blood sampled by puncture of the brachial vein and genotyped with six polymorphic microsatellite-DNA markers ( Table 2). Details on DNA extraction and PCR amplification conditions are described in Galbusera et al. (2000) . Genotypes were scored on a 6% acrylamide gel in an automated sequencer (ALF express, Pharmacia Biotech). No allelic disequilibrium was detected between any two loci, and all locus-population combinations were in Hardy–Weinberg equilibrium ( Table 2). Given the absence of a heterozygote deficit, null alleles were assumed absent.

Table 2.  Characteristics of six polymorphic microsatellite loci from which individual inbreeding coefficients were estimated. Thumbnail image of

The probability that both alleles at any locus are identical by descent (i.e. inbreeding coefficient f;Falconer & Mackay, 1996) is traditionally estimated by the proportion of homozygous loci in an individual. However, estimates of the probability of identity-by-descent, calculated from mean homozygosity at loci with different expected levels of identity-by-state, are biased downward ( Ritland, 1996). Recently, Ritland (1996) developed an unbiased estimator for individual inbreeding coefficients using information from several polymorphic marker loci. Ritland estimates, f^, are calculated for each allele at each locus as: f^=(Σi,l((Sil – Pil2)/Pil))/(Σl(nl – 1)) (where i and l denote alleles and loci, respectively; Sil equals 0 if allele i is not contained, 0.5 if i is contained once, or 1 if i is contained twice; Pi is the frequency of allele i at locus l; and nl is the number of alleles at locus l), and combined into a single inbreeding coefficient as a weighted sum (see Ritland, 1996 for procedure and assumptions). Given the relatively high mutation rates of microsatellite loci ( Ellegren et al., 1995) , f^ reflects recent inbreeding. Because CH, NG and MB were genetically differentiated and gene flow, as estimated from assignment tests, was virtually absent ( Galbusera et al., 2000 ), allele frequencies were estimated per population. Average f^ did not differ between the three populations (mean ± SD: CH: –0.066 ± 0.19, NG: –0.019 ± 0.14, MB: +0.003 ± 0.17; F2,234=1.61, P=0.20; allele frequencies estimated from pooled data as population mean values have zero expectation; Ritland, 1996).

Under the assumption that the stepwise mutation model adequately describes the evolution of microsatellite alleles, internal distance estimates provide information about the time since microsatellite coalescence ( Goldstein et al., 1995 ; Slatkin, 1995) and, thus, reflect probabilities of common ancestry of alleles over a larger time-span (i.e. when mutations become frequent). Internal distance estimates, combining information from multiple loci, are calculated as: mean d2= ((1/L) ((ΣLi=1)(a1,i – a2,i)2)) (where a1,i and a2,i denote lengths of alleles at locus i for allelic position 1 and 2, respectively; and L is the number of loci examined) ( Coulson et al., 1998 ; see Table 2 for locus-specific d2-values). Average mean d2 did not differ between the most disturbed (CH: 61.2 ± 18.8) and least disturbed (MB: 72.5 ± 6.5) population (F1,234=0.01, P=0.94), but was actually higher in the intermediately disturbed one (NG: 121.5 ± 9.7) (NG – MB: F1,234=10.0, P=0.002; NG – CH: F1,234= 2.82, P=0.10).

Relationship between habitat disturbance, FA, and inbreeding

Individual inbreeding coefficients were treated as linear covariates in the analyses. As all analyses involved repeated measurements at the level of the individual (i.e. three traits measured per thrush), we performed mixed regression analysis with repeated-measure structure and factor trait as random factor, using Proc Mixed in SAS ( Littell et al., 1996 ). The initial model included FA as dependent variable, and Population, Inbreeding and the Population × Inbreeding interaction as fixed factors (interaction could be tested as levels of habitat disturbance were not confounded by mean inbreeding, see above). To test whether the two-way fixed effects interaction was trait-dependent, we added the three-way random effects interaction Trait × Population × Inbreeding (and all two-way interactions with factor Trait) to the model. Applying backward selection, significance of variance components was tested by likelihood ratio tests, while fixed effects were tested by F-tests, adjusting the denominator d.f. by Satterthwaite’s formula ( Verbeke & Molenberghs, 1997). To test whether inbreeding effects resulted from recent or ancestral events, factor Inbreeding was inferred from Ritland estimates and internal distance estimates, respectively.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Bibliography

Signed FA estimates showed no directional component (mean values of the signed FA did not differ from zero; F-tests: all P > 0.3) and were highly significant, i.e. VFA was much larger than VME for all traits (likelihood ratio tests: all P < 0.0001) ( Table 3). The leptokurtic distribution of the signed FA (high kurtosis K) indicated substantial between-individual variation in DI, which was reflected in high hypothetical repeatabilities (mean R-values: CH=0.42, NG=0.49, MB=0.50; Table 3).

Table 3.  Variance components and the distribution of signed and unsigned FA (three traits) for three Taita thrush populations. Thumbnail image of

The magnitude of the relationship between f^ and FA differed significantly between populations (Population × Inbreeding interaction: P=0.038, Table 4). As predicted, the relationship was significantly positive in the most disturbed population (CH, Fig. 1) and decreased under decreasing levels of habitat disturbance (estimated regression slopes: CH: 1.65 ± 0.81, P=0.022; NG: 0.99 ± 0.50, P=0.026; MB: –0.14 ± 0.31, P=0.32; one-tailed), explaining zero variation in FA in the least disturbed population ( Fig. 1). This pattern was consistent across traits (Trait × Population × Inbreeding interaction not significant; Table 4). After backward exclusion of all nonsignificant terms, the final regression model contained the two-way fixed Population × Inbreeding interaction (and, thus, its respective main effects), the random intercept (Trait) and the random Trait × Population interaction (the latter two modelling between-trait variation in average and population-level FA and therefore not of prime interest to our hypothesis). This model explained 67% of the total variation in FA. When translating patterns of FA into patterns of DI by the hypothetical repeatability R, inbreeding explained 86% (TL), 52% (RL) and 25% (PW) of variation in DI in the most disturbed population.

Table 4.  Test of fixed and random effects on individual estimates of fluctuating asymmetry. Thumbnail image of
image

Figure . 1. Relationship between individual levels of inbreeding and of FA, in three populations exposed to different levels of habitat disturbance. Pairwise regression slopes were significantly different between the least (MB) and most (CH) disturbed population (F1,242=4.21, P=0.041) and nearly significant between the least and intermediately (NG) disturbed one (F1,223=3.65, P=0.057). All subgraphs contain hidden values.

Download figure to PowerPoint

When recalculating f^ for all possible combinations of five of six loci, the above effect on the relationship between f^ and FA persisted (Population × Inbreeding interaction: 0.019 < P < 0.058). When replacing f^ by mean d2, in contrast, the Population × Inbreeding interaction (and main inbreeding effect) no longer explained any variation in FA ( Table 4), also when re-calculating mean d2 for all combinations of five loci (significance range: 0.50 < P < 0.96).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Bibliography

Results presented in this paper provide the first evidence that the magnitude of the relationship between developmental instability and inbreeding in natural populations may depend on environmental factors, such as the level of habitat disturbance experienced during ontogeny. This pattern proved consistent across traits that are expressed during different stages of the development. As it is virtually impossible to control for among-population differences in genetic and/or environmental factors, results of this paper (i) provide a possible explanation for the discordance in observed relationship between DI and inbreeding stemming from nonexperimental studies, and (ii) highlight the difficulties in defining relationships between both factors and DI in natural populations.

Apart from inconsistency in the pattern, uncertainty about its causality has fostered the controversy over the association between DI and inbreeding ( Palmer, 1996). First, the relationship between the heterozygosity of individual loci and that throughout the genome is assumed to be low in randomly mating populations (e.g. Mitton, 1978; Nevo, 1978). As the effective population size in fragment CH was much smaller than in the other fragments (Ne-estimates under the assumption of mutation-drift equilibrium (μ=5 × 10–4): CH=30, NG=150, MB=750; Galbusera et al., 2000 ), the likelihood of mating between closely related individuals can be expected to be higher in CH as compared with the larger populations. This might substantially have increased variation in the level of inbreeding in CH, in which case inbreeding coefficients calculated over a small number of microsatellite loci may better reflect genome-wide inbreeding than in populations NG and MB ( Mitton, 1997). Consequently, as closer association between the estimated individual inbreeding coefficients and the true underlying level of inbreeding can be expected to increase statistical power to detect associations with other variables (such as FA), the observed heterogeneity in association between DI and inbreeding among populations could be suspected to be due to heterogeneity in statistical power. Yet, we consider this alternative explanation unlikely, because mean inbreeding did not differ between the three populations, and because the slope of the regression line was significant in the intermediately sized population (NG) too. Secondly, Mitton (1993) argued that the observed correlations might arise as by-products of direct effects of particular heterozygous loci on physiological efficiency. This view was supported by perceptive studies of enzymatically nonfunctional null-alleles in salmonid fishes ( Leary et al., 1993 ). Although correlative studies, such as ours, do not allow to distinguish between putative mechanisms underlying the relationship between DI and inbreeding, the persistent inbreeding effect following the removal of any single locus from the calculation of the Ritland estimates, suggested that the observed pattern was not due to linkage of single microsatellite-DNA markers to a particular FA-affecting locus.

Inbreeding per se explained as much as 36% of the variation in FA (TL) in the most disturbed fragment, despite the fact that Ritland estimates are typically hampered by large errors of inference ( Ritland, 1996; Lynch & Ritland, 1999) and individual FA is only weakly correlated with DI as it attempts to estimate a variance with two data points ( Palmer, 1994). Given the high levels of hypothetical repeatability of FA in all populations, the decrease in association between inbreeding and DI with diminishing habitat disturbance was not due to a parallel decrease in statistical power. Rather, the expression of the genetic variance of a trait is believed to depend on environmental conditions ( Hoffmann & Parsons, 1991; Sgró & Hoffmann, 1998), and stressful conditions may cause nonadditive genetic variance to be converted into additive genetic variance ( Van Dongen & Lens, 2000 and references therein). Hence, as the evolutionary potential of a trait can be expected to increase under increasing stress because more of the underlying genetic variability is being revealed, strong association between inbreeding and DI under severe habitat disturbance, but not under more relaxed conditions, is in line with the hypothesized association between the evolutionary potential of DI and stress ( Van Dongen & Lens, 2000). In view of the present lack of a theoretical framework on the genetic basis of developmental instability, results from this study therefore strongly support the current plea for further research on the genetic architecture of DI in stressed populations ( Jenkins et al., 1997 ).

Empirical studies recently provided evidence that mean d2 can be a more powerful indicator of the fitness consequences of inbreeding and outbreeding than heterozygosity per se ( Coltman et al., 1998 ; Coulson et al., 1998 ). Unless the presumed underlying mutation model was incorrect ( Estoup & Angers, 1998), the absence of an effect of mean d2 in our study suggests that inbreeding effects, as revealed by f^, are fairly recent. Earlier, comparison of contemporary FA levels with those in museum specimens collected before the main forest fragmentation, revealed strong, parallel increases in FA in the most disturbed fragment but no change in the least disturbed one, in five Taita forest bird species ( Lens et al., 1999 ). Results presented in this paper provide additional evidence that the effects of human impact on the Taita bird community are, indeed, of recent origin.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Bibliography

We are grateful to E. Waiyaki, D. Gitau, S. Karimi and T. Imboma for measuring and blood sampling Taita thrushes, and S. Griffith, K. Otter, O. Hanotte, H. Bickle and T. Burke for sharing unpublished sequences of the microsatellite-DNA primer sets. Two anonymous referees provided the most valuable comments that improved our paper considerably. Fieldwork was funded by a project of the Flemish Inter University Council (to W.N. Verheyen and EM), a research grant (to EM) and two postdoctoral fellowships (to LL and SVD) of FWO-Flanders, and an IWT specialization grant (to TVC).

Bibliography

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Bibliography
  • 1
    Bjorksten, T.A., Fowler, K., Pomiankowsky, A. 2000. What does sexual trait FA tell us about stress? Trends Ecol. Evol. 15: 163 166.DOI: 10.1016/s0169-5347(99)01788-7
  • 2
    Britten, H.B. 1996. Meta-analysis of the association between multilocus heterozygosity and fitness. Evolution 50: 2158 2164.
  • 3
    Brooks, T., Lens, L., Barnes, J., Barnes, R., Kihuria, J.K., Wilder, C. 1998. The conservation status of the forest birds of the Taita Hills, Kenya. Bird Conserv. Intern. 8: 119 139.
  • 4
    Clarke, G.M. 1993. The genetic basis of developmental stability: I. Relationships between stability, heterozygosity and genomic coadaptation. Genetica 89: 15 23.
  • 5
    Clarke, G.M. 1998. The genetic basis of developmental stability: IV. Individual and population asymmetry parameters. Heredity 80: 553 561.
  • 6
    Coltman, D.W., Don Bowen, W., Wright, J.M. 1998. Birth weight and neonatal survival of harbour seal pups are positively correlated with genetic variation measured by microsatellites. Proc. Royal Soc. London Series B 265: 803 809.
  • 7
    Coulson, T.N., Pemberton, J.M., Albon, S.D., Beaumont, M., Marshall, T.C., Slate, J., Guinness, F.E., Clutton-Brock, T.H. 1998. Microsatellites reveal heterosis in red deer. Proc. Royal Soc. London Series B 265: 489 495.
  • 8
    Davies, N., Villablanca, F.X., Roderick, G.K. 1999. Determining the source of individuals: multilocus genotyping in nonequilibrium population genetics. Trends Ecol. Evol. 14: 17 21.DOI: 10.1016/s0169-5347(98)01530-4
  • 9
    Ellegren, H., Lifjeld, J.T., Slagsvold, T., Primmer, C.R. 1995. Handicapped males and extrapair paternity in pied flycatchers: a study using microsatellite markers. Mol. Ecol. 4: 739 744.
  • 10
    Estoup, A. & Angers, B. 1998. Microsatellites and minisatellites for molecular ecology: theoretical and empirical considerations. In: Advances in Molecular Ecology (G. R. Carvalho, ed.), pp. 55–86. IOS Press, Ohmsa.
  • 11
    Falconer, D.S. & Mackay, T.F.C. 1996. Introduction to Quantitative Genetics, 4th edn. Longman, New York.
  • 12
    Galbusera, P., Lens, L., Schenck, T., Waiyaki, E., Matthysen, E. 2000. Genetic variability and gene flow in the globally, critically-endangered Taita thrush. Conserv. Gen. 1: 45 55.
  • 13
    Goldstein, D.B., Linares, A.R., Cavalli-Sforza, L.L., Feldman, M.W. 1995. An evaluation of genetic distances for use with microsatellite loci. Genetics 139: 463 471.
  • 14
    Hoffmann, A.A. & Parsons, P.A. 1991. Evolutionary Genetics and Environmental Stress. Oxford University Press. Oxford.
  • 15
    Jenkins, N.L., Sgró, C.M., Hoffmann, A.A. 1997. Environmental stress and the expression of genetic variation. In: Environmental Stress, Adaptation and Evolution (R. Bijlsma & V. Loeschcke, eds), pp. 79–96. Birkäuser-Verlag, Basel.
  • 16
    Leary, R.F., Allendorf, F.W., Knudsen, K.L. 1984. Superior developmental stability of heterozygosity in enzyme loci in salmonid fishes. Am. Nat. 124: 540 551.
  • 17
    Leary, R.F., Allendorf, F.W., Knudsen, K.L. 1993. Null alleles at two lactate dehydrogenase loci in rainbow trout are associated with decreased developmental stability. Genetica 89: 3 13.
  • 18
    Lens, L. & Van Dongen, S. 1999. Evidence for organism-wide asymmetry in five bird species of a fragmented afrotropical forest. Proc. Royal Soc. London Series B 266: 1055 1060.
  • 19
    Lens, L., Van Dongen, S., Wilder, C.M., Brooks, T.M., Matthysen, E. 1999. Fluctuating asymmetry increases with habitat disturbance in seven bird species of a fragmented afrotropical forest. Proc. Royal. Soc. London Series B 266: 1241 1246.
  • 20
    Leung, B. & Forbes, M.R. 1996. Fluctuating asymmetry in relation to stress and fitness: effect of trait type as revealed by meta-analysis. Ecoscience 3: 400 413.
  • 21
    Littell, R.C., Milliken, G.A., Stroup, W.W., Wolfinger, R.D. 1996. SAS System for Mixed Models. SAS Institute Inc, New York.
  • 22
    Ludwig, W. 1932. Das Rechts-Links Problem Im Tierreich und Beim Menschen. Springer. Berlin.
  • 23
    Lynch, M. & Ritland, K. 1999. Estimation of pairwise relatedness with molecular markers. Genetics 152: 1753 1766.
  • 24
    Mitton, J.B. 1978. Relationship between heterozygosity for enzyme loci and variation of morphological characters in natural populations. Nature 273: 661 662.
  • 25
    Mitton, J.B. 1993. Enzyme heterozygosity, metabolism, and developmental stability. Genetica 89: 47 65.
  • 26
    Mitton, J.B. 1997. Selection in Natural Populations. Oxford University Press, Oxford.
  • 27
    Nevo, E. 1978. Genetic variation in natural populations: patterns and theory. Theor. Pop. Biol. 13: 121 177.
  • 28
    Nilsson, J.A. 1994. Energetic stress and the degree of fluctuating asymmetry: implications for a long-lasting, honest signal. Evol. Ecol. 8: 248 255.
  • 29
    Palmer, A.R. 1994. Fluctuating asymmetry analyses: A primer. In: Developmental Stability: It’s Origins and Evolutionary Implications (T. A. Markow, ed.), pp. 335–364. Kluwer Academic Publishers, Dordrecht.
  • 30
    Palmer, A.R. 1996. Waltzing with asymmetry. Is fluctuating asymmetry a powerful new tool for biologists or just an alluring new dance step? Bioscience 46: 518 532.
  • 31
    Palmer, A.R. & Strobeck, C. 1986. Fluctuating asymmetry: measurement, analysis and patterns. Ann. Rev. Ecol. Syst. 17: 391 421.
  • 32
    Raymond, M. & Rousset, F. 1995. An exact test for population differentiation. Evolution 49: 1280 1283.
  • 33
    Reeve, E.C.R. 1960. Some genetic tests on asymmetry of sternopleural chaetae number in Drosophila. Gen. Res. 1: 151 172.
  • 34
    Ritland, K. 1996. Estimators for pairwise relatedness and individual inbreeding coefficients. Gen. Res. 67: 175 185.
  • 35
    Sgró, C.M. & Hoffmann, A.A. 1998. Effects of temperature extremes on genetic variances for life-history traits in Drosophila melanogaster as determined by parent-offspring comparisons. J. Evol. Biol. 11: 1 20.
  • 36
    Slatkin, M. 1995. A measure of population subdivision based on microsatellite allele frequencies. Genetics 139: 457 462.
  • 37
    Soulé, M.E. 1979. Heterozygosity and developmental stability: another look. Evolution 33: 396 401.
  • 38
    Van Dongen, S. 1998. How repeatable is the estimation of developmental stability by fluctuating asymmetry. Proc. Royal Soc. London Series B 265: 1423 1427.
  • 39
    Van Dongen, S. & Lens, L. 2000. The evolutionary potential of developmental instability. J. Evol. Biol. 13: 326 335.DOI: 10.1046/j.1420-9101.2000.00169.x
  • 40
    Van Dongen, S., Molenberghs, G., Matthysen, E. 1999. The statistical analysis of fluctuating asymmetry: REML estimation of a mixed regression model. J. Evol. Biol. 12: 94 102.DOI: 10.1046/j.1420-9101.1999.00012.x
  • 41
    Verbeke, G. & Molenberghs, G. 1997. Linear Mixed Models in Practice: a SAS-Oriented Approach. Lecture notes in statistics, p. 126. Springer-Verlag, New York.
  • 42
    Völlestad, L.A., Hindar, K., Möller, A.P. 1999. A meta-analysis of fluctuating asymmetry in relation to heterozygosity. Heredity 83: 206 218.DOI: 10.1046/j.1365-2540.1999.00555.x
  • 43
    Vrijenhoek, R.C. & Lerman, S. 1982. Heterozygosity and developmental stability under asexual breeding systems. Evolution 36: 768 776.
  • 44
    White, G.C. & Burnham, K.P. 1999. Program MARK: survival estimation from populations of marked animals. Bird Study 46: S120 S138.
  • 45
    Whitlock, M. 1996. The heritability of fluctuating asymmetry and the genetic control of developmental stability. Proc. Royal Soc. London Series B 263: 849 854.
  • 46
    Whitlock, M. 1998. The repeatability of fluctuating asymmetry: a revision and extension. Proc. Royal Soc. London Series B 265: 1428 1430.