Close Home
iForest - Biogeosciences and Forestry
vol. 3, pp. 150-152
Copyright © 2010 by the Italian Society of Silviculture and Forest Ecology
doi: 10.3832/ifor0552-003

Research Articles

A comparative fluctuating asymmetry study between two walnut (Juglans regia L.) populations may contribute as an early signal for bio-monitoring

Y.A.I. Kourmpetis (1), F.A. Aravanopoulos (2)Corresponding author


Adaptation of organisms to changing environments is becoming a very important issue, especially in the context of climate change. Developmental stability is the ability of an individual to eliminate environmental disturbances while expressing a heritable phenotypic trait. Developmental stability is an important notion for evolutionary and ecological studies, since it provides valuable information on the adaptation of organisms or populations to certain environments. There is evidence that developmental stability has a genetic basis ([1]), nevertheless the relationship between genes and developmental stability is still under debate ([21]).

The lack of data and models does not allow direct estimations of developmental stability. In practice, the pattern of bilateral variation in a random sample of the population, called fluctuating asymmetry ([27], [14]) is widely used as an indirect developmental stability estimator. Fluctuating asymmetry refers to the random deviation from perfect symmetry of bilateral traits resulting from extrinsic and intrinsic perturbations not buffered during development ([27], [14]). Fluctuating asymmetry has been suggested to result from poorly co-adapted gene complexes and therefore to be useful as a measure of developmental stability ([6]). The relation between fluctuating asymmetry and developmental stability is based on the assumption that the symmetric parts of bilateral traits are controlled by the same groups of genes and also that both sides share the same environment, during the process of development. Therefore a perfectly symmetrical phenotype can be expected. During development however, several random errors occur independently and additively in each side introducing a source of uncertainty into the phenotype development mechanism and consequently leading to random deviations from perfect symmetry. The variance of the random deviations (that is equivalent to fluctuating asymmetry) is proportional to the error rate during development. Assuming linear correlation, this variance parameter can be used as an estimator of this error rate and furthermore of developmental stability. Detailed reviews on fluctuating asymmetry are given by Palmer & Strobeck ([15]), Palmer & Strobeck ([16]), Van Dongen ([26]) and Graham et al. ([4]).

There is some underlying additive genetic variance associated with fluctuating asymmetry and relevant QTLs have been identified in poplars ([28]) and mice ([8]), but the extent of its genetic basis remains unresolved ([12], [13], [11], [7], [24]). Nevertheless, there is at least circumstantial evidence that has linked fluctuating asymmetry to environmental stress ([17], [18], [22], [5]) and even to genetic bottlenecks ([6]).

In this study, we compared the fluctuating asymmetry as an indicator of developmental stability in two adjacent walnut (Juglans regia L.) populations: an artificial and a natural. We have estimated fluctuating asymmetry by bilateral leaf morphometrics assuming that leaf asymmetry provides a reliable estimate of the developmental stability in plants and the populations that they belong to. Leaves are highly sensitive tree organs that are subject to environmental, as well as phenological growth cycles and fluctuations ([2]). Our working hypothesis is that comparative developmental stability levels estimated by fluctuating asymmetry provide a cost effective way to obtain insight with regards to population analysis and biomonitoring. Such insight may be valuable for the design of future more thorough studies.

Materials and methods 


The two populations compared in this study are located in the region of Sohos, near Thessaloniki (Greece) in the area of Mt. Vertiskos at an elevation of 650-700 m (40o 49’ N, 23o 22’ E), characterized by sand-clay soils and typical Mediterranean climate. The region is renown for its walnut and features both wild populations and walnut orchards with trees that attain a height of up to 25 m. The average age of the natural population is about 60 years, however older trees can be found. The artificial population comprises of locally selected genotypes of younger age (40 years old on the average). We chose randomly 29 individuals to sample from each population at a minimum distance of 20 m from each other, avoiding sampling filial structures. Isolated trees (no other tree present at a radius of 100 m) were excluded from sampling. Trees exhibiting natural (infection) or anthropogenic (pruning) disturbance were excluded as well. Ten leaves per tree were randomly selected from lower branches at a height of 2 m, being approximately equally distributed on the circumference of the tree canopy. Measurements were taken after careful drainage.


We measured the perimeters of the left (L) and the right (R) side of the leaf in a resolution of 1 mm and used the log-ratio: log(L/R) as Signed Asymmetry (SA) measure per leaf. We were interested to study developmental stability (DS) at the population level, thus fluctuating asymmetry (FA) was estimated using the pooled samples per population. We approached our analysis from a Bayesian perspective. We employed a model to infer SA and Directional Asymmetry (DA - i.e., when one side is systematically larger compared to the other side), similar to the one proposed by Van Dongen ([23], [25], [26] - eqn. 1):

\begin{equation} Y_{ip} \sim N \left(\mu_p,\sigma^{2}_{FA(p)} \right) \end{equation}

where Yip denotes the i-th SA observation taken from the population p (i.e., natural or artificial); μp corresponds to the expected value of the observation. Deviation from zero indicates the presence of DA in both populations. Otherwise, in case of DA absence, μp should be equal to zero. Parameter σ2FA(p) denotes the variance of SA in the population p (artificial or natural). By using a single parameter for SA per population, we assume that all the trees in each population share a common SA. We used this parameter to observe the developmental stability of the corresponding population. The prior distributions for the parameters in our model are similar to those proposed by Van Dongen ([25]): μp ~ N(0.10) and σ-2FA(p) ~ Γ(0.001, 0.001), where Γ denotes the Gamma distribution. In order to compare the difference in developmental stability between the two populations we estimated the posterior probability (eqn. 2):

\begin{equation} P \left( \sigma_{FA(nat)} > \sigma_{FA(art)} | Y \right) \end{equation}

We applied Markov Chain Monte Carlo (MCMC) to estimate the probability in eqn. 2 and used the JAGS ([19]) software to perform this posterior simulation. From a total of 120 000 iterations, the first 20 000 were discarded as burn-in stage, while we sampled from the Markov Chain every 100 iterations in order to reduce auto-correlation between samples. Four parallel Markov Chains were simulated and convergence diagnostics were obtained using the CODA R package ([20]). Further inferences were made by combining the samples from the four chains. We used parameter r = σFA(nat) / σFA(art) to monitor the ratio of developmental stability values in the two populations. In addition to our Bayesian analysis, we tested the hypothesis that the variances between the two populations are equal by performing a F-test with null hypothesis H0: σ2FA(nat) = σ2FA(art) and alternative H1: σ2FA(nat) > σ2FA(art) and using the statistic r2 = σ2FA(nat) /σ2FA(art).

Results and Discussion 

Summary statistics and posterior densities are shown in Tab. 1 and Fig. 1 respectively. Fluctuating asymmetry in the natural population is larger than the fluctuating asymmetry in the artificial, with a probability of 0.982. This result is also supported by the F-test (test statistic r2 = 1.27) resulting to p-value = 0.021 (with the classical statistics interpretation) with confidence intervals CI = (1.047, +inf). We regard this outcome as noteworthy. Our results show that the right side of the leaf is slightly larger than the left side, but still the difference is small and additional experimentation is needed to further elucidate this finding. The natural population can be regarded as presenting reduced levels of developmental stability. It is noted that a significant amount of genetic erosion has been indicated for European J. regia populations and while genetic diversity parameters for northern Greek populations were higher that the average, they were notably lower than typical values reported for widespread plant species ([3]). The artificial population has been long cultivated for the production of regionally renowned walnuts. Nevertheless, any positive influence of cultivation activities on the artificial population does not apply to the adjacent natural population that may be more prone to environmental perturbations and stochastic fluctuations of effective population size. In this case, our results may call for more detailed genetic studies in order to investigate the evolutionary potential of the natural population (genetic monitoring).

Tab. 1 - Summary statistics for the posterior densities of parameters. Column Mean denotes the posterior mean. SD is the standard deviation. The last four columns show quantiles for each variable.
Fig. 1 - Traces and posterior densities for the parameters of interest. Plots were produced by concatenating the four MCMC chains.

We consider our findings as a potentially early warning indication for problems that may occur in the natural population under scenarios of future environmental uncertainty. Given that expression of both sides of a leaf bilateral trait is due to the same genes, then any asymmetry between the sides can be a consequence of environmental disturbance. Hence, fluctuating asymmetry could be considered as a proxy for environmental or genetic stress ([27], [14]). Our approach is characterized by simple data collection, while a thorough analysis can be streamlined by specific software that we plan to develop. Therefore the careful application of such an approach can present an additional tool to a battery of available applications for population bio-monitoring ([10], [9]). Natural biodiversity is extremely important and it may be vital for biomonitoring and conservation policies to take into account the contribution of simple potential “signals”.


Two anonymous reviewers are thanked for their constructive comments and suggestions.


Clarke GM (1998). The genetic basis of developmental stability. IV. Individual and population asymmetry parameters. Heredity 80: 553-561.
::CrossRef::Google Scholar::
Fair JM, Breshears DD (2005). Drought stress and fluctuating asymmetry in Quercus undulata leaves: confounding effects of absolute and relative amounts of stress? Journal of Arid Environments 62: 235-249.
::CrossRef::Google Scholar::
Fornari B, Cannata F, Spada M, Malvolti ME (1999). Allozyme analysis of genetic diversity and differentiation in European and Asiatic walnut (Juglans regia L.) populations. Forest Genetics 6: 115-127.
::Google Scholar::
Graham JH, Raz S, Hel-Or H and Eviatar Nevo (2010). Fluctuating asymmetry: methods, theory, and applications. Symmetry 2: 466-540.
::CrossRef::Google Scholar::
Hodar JA (2002). Leaf FA of Holm oak in response to drought under contrasting climatic conditions. Journal of Arid Environments 52: 233-243.
::CrossRef::Google Scholar::
Hoelzel R, Fleischer RC, Campagna C, Le Boeuf BJ, Alvord G (2002). Impact of a population bottleneck on symmetry and genetic diversity in the northern elephant seal. Journal of Evolutionary Biology 15: 567-575.
::CrossRef::Google Scholar::
Houle D (1997). Comment on “A meta-analysis of the heritability of developmental stability” by Moller and Thornhill. Journal of Evolutionary Biology 10: 17-20.
::CrossRef::Google Scholar::
Leamy LJ, Routman EJ, Cheverud JM (2002). An epistatic genetic basis for fluctuating asymmetry of mandible size in mice. Evolution 56: 642-653.
::CrossRef::Google Scholar::
Lens L, Van Dongen S (2002). Fluctuating asymmetry as a bio-indicator in isolated populations of the Taita thrush: a Bayesian perspective. Journal of Biogeography 29: 809-819.
::CrossRef::Google Scholar::
Leung B, Forbes MR, Houle D (2000). Fluctuating asymmetry as bioindicator of stress: comparing efficacy of analyses involving multiple traits. American Naturalist 155: 101-115.
::CrossRef::Google Scholar::
Markow TA, Clarke GM (1997). Meta-analysis of the heritability of developmental stability: a giant step backward. Journal of Evolutionary Biology 10: 31-37.
::CrossRef::Google Scholar::
Moller AP, Thornhill R (1997a). A meta-analysis of the heritability of developmental stability. Journal of Evolutionary Biology 10: 1-16.
::CrossRef::Google Scholar::
Moller AP, Thornhill R (1997b). Developmental instability is heritable. Journal of Evolutionary Biology 10: 69-76.
::CrossRef::Google Scholar::
Palmer AR (1994). Fluctuating asymmetry analyses: a primer. In: “Developmental instability: its origins and evolutionary implications” (Markow T ed). Kluwer Academic Publishers, Dordrecht, Germany, pp. 335-364.
::Google Scholar::
Palmer AR, Strobeck C (1986). Fluctuating asymmetry: measurement, analysis, patterns. Annual Review of Ecology and Systematics 17: 391- 421.
::CrossRef::Google Scholar::
Palmer AR, Strobeck C (2003). Fluctuating asymmetry analyses revisited. In “Developmental instability (DI): causes and consequences” (Polak M ed). Oxford University Press, Oxford, UK, pp. 279- 319.
::Google Scholar::
Parsons PA (1990). Fluctuating asymmetry: an epigenic measure of stress. Biological Reviews 65: 131-154.
::CrossRef::Google Scholar::
Parsons PA (1992). Fluctuating asymmetry: a biological monitor of environmental and genomic stress. Heredity 68: 361-364.
::CrossRef::Google Scholar::
Plummer M (2003). JAGS: A program for analysis of Bayesian graphical models using Gibbs sampling. In: Proceedings 3rd International Workshop on “Distributed Statistical Computing”.
::Google Scholar::
Plummer M, Best N, Cowles K, Vines K (2006). CODA: Convergence diagnosis and output analysis for MCMC. R News 6: 7-11.
::Google Scholar::
Pertoldi C, Kristensen TN, Andersen DH, Loeschcke V (2006). Developmental instability as an estimator of genetic stress. Heredity 96: 122-127.
::CrossRef::Google Scholar::
Silva MC, Lomonaco C, Augusto SC, Kerr WE (2009). Climatic and anthropic influence on size and fluctuating asymmetry of Euglossine bees (Hymenoptera, Apidae) in a semideciduous seasonal forest reserve. Genetics and Molecular Research 8: 730-737.
::CrossRef::Google Scholar::
Van Dongen S (2000a). Unbiased estimation of individual asymmetry. Journal of Evolutionary Biology 13: 107-112.
::CrossRef::Google Scholar::
Van Dongen S (2000b). The heritability of fluctuating asymmetry: a Bayesian hierarchical model. Annales Zoologici Fennici 37: 15-24.
::Google Scholar::
Van Dongen S (2001). Modelling developmental instability in relation to individual fitness: a fully Bayesian latent variable model approach. Journal of Evolutionary Biology 14: 552-563.
::CrossRef::Google Scholar::
Van Dongen S (2006). Fluctuating asymmetry and developmental instability in evolutionary biology: past, present and future. Journal of Evolutionary Biology 19: 1727-1743.
::CrossRef::Google Scholar::
Van Valen L (1962). A study of fluctuating asymmetry. Evolution 16: 125-142.
::CrossRef::Google Scholar::
Wu J, Zhang B, Cui Y, Zhao W, Xu L, Huang M, Zeng Y, Zhu J, Wu R (2007). Genetic mapping of developmental instability: design, model and algorithm. Genetics 176: 1187-1196.
::CrossRef::Google Scholar::


Kourmpetis YAI, Aravanopoulos FA (2010).
A comparative fluctuating asymmetry study between two walnut (Juglans regia L.) populations may contribute as an early signal for bio-monitoring
iForest - Biogeosciences and Forestry 3: 150-152. - doi: 10.3832/ifor0552-003
First Previous Next Last
© iForest

Download Reference

Paper ID# ifor0552-003
Title A comparative fluctuating asymmetry study between two walnut (Juglans regia L.) populations may contribute as an early signal for bio-monitoring
Authors Kourmpetis YAI, Aravanopoulos FA
Close Download