sign in sign up
<<
Home , Phenotypic plasticity

Opinion TRENDS in Ecology & Vol.16 No.10 October 2001 555

Canalization Mapping : Canalization was first defined by Waddington11,12 as the ability to produce a consistent in spite of variable genetic and/or environmental features canalization, plasticity (Box 1). Later13,14, Waddington broadened this definition, focusing on phenotypes that, if not strictly invariable, are ‘to some extent resistant to and developmental modification’. Simultaneously, he developed the idea of canalizing selection, implying a genetic control of canalization. Waddington thought that the same stability process counteracted both genetic and environmental disturbances. This duality is the basis of ‘’, the process by which a response to Vincent Debat and Patrice David unusual environments can be converted by selection into a permanent, genetically determined, phenotypic change. In parallel, Schmalhausen15 developed the The relationship between genotype and phenotype is not one to one. This related concept of the ‘auto-regulatory mechanism’ statement is central to our understanding of how shapes (Box 1), a process that stabilizes the phenotypic evolution. Here, we clarify the links between canalization, plasticity against environmental influences and . He and developmental stability, the three major processes involved in the control also recognized that some labile showed of phenotypic variability. We present a short historical review, including the environmentally induced changes in development, original definitions of these concepts, and then summarize their current although he never used the term ‘plasticity’. meaning and use, highlighting possible sources of confusion. Some of the Thoday16, in defining ‘developmental flexibility’, perspectives allowed by a more synthetic conceptual framework are presented, explicitly introduced an to the in the light of the recent advances in molecular and developmental genetics. processes involved in developmental control (Box 1).

One of the most important advances achieved through Phenotypic plasticity the evolutionary synthesis is the consideration of Woltereck17 introduced the concept of the ‘reaction phenotypic variation as a quantity of interest, whereas, norm’ as early as 1909, but Johannsen was the first to according to the classic typological view, it is considered point out the general importance of environmental a nuisance. Thus, the factors affecting phenotypic influences for genotype–phenotype relationships18. variation have been given considerable attention by Schmalhausen (Box 1) used two different terms to evolutionists. It is widely accepted that variability describe such influences, depending on their adaptive (defined as the ability to vary1–3) results from two or nonadaptive nature15. The definition of plasticity antagonistic trends4: on the one hand, sources of given by Bradshaw19 encompasses both variation, including genetic mutations, environmental Schmalhausen’s concepts (Box 1). Smith-Gill20 effects and developmental errors; on the other hand, a recently distinguished two different components in set of regulatory processes, including buffering and plasticity, focusing on the discrete versus continuous enhancing mechanisms. Three such processes are variation of the traits in relation to their adaptive commonly considered: canalization, phenotypic value (Box 1). plasticity and developmental stability. Much effort has been made recently to understand the genetic Developmental stability bases, and evolutionary implications of Physiological homeostasis ensures a constant end these processes5–8. However, the literature can be product in spite of disturbances10. Lerner’s21 confusing, because the historical definitions usually extension of this concept (‘genetic homeostasis’,

Vincent Debat* referred to do not take into account later conceptual Box 1) is the basis of many later studies dealing with Laboratoire Génétique et and empirical advances. Most modern authors correlations between heterozygosity and Environnement, Institut consider each of the three processes separately, morphological variance22,23. In this context, des Sciences de 5 l’Evolution, UMR 5554 although their distinction might appear arbitrary . individual deviations from the population mean CNRS, CC064, Université were often interpreted as developmental errors, Montpellier II, 34095 Historical context: the origin of disorder revealing a lack of developmental stability or Montpellier Cedex 05, The study of developmental regulatory processes is developmental homeostasis, taken as synonyms of France. *e-mail: rooted in the old concept of physiological homeostasis, Waddington’s canalization. However, the use of the [email protected] defined by Bernard9 and Cannon10. word ‘homeostasis’ at both the individual and montp2.fr Developmentalists, such as Waddington and population levels is confusing and was rejected by 13 Patrice David Schmalhausen, suggested that developmental Waddington , and many researchers argued that Centre d’Ecologie pathways, as well as , must be strongly these concepts were unnecessary to explain the Fonctionnelle et Evolutive controlled. The concepts of canalization, plasticity observed correlations24. In addition to (CEFE) CNRS, route de Mende, 34095 Montpellier and stability were developed to describe such morphological variance, the deviation from Cedex 05, France. control systems. bilateral symmetry, or ‘fluctuating asymmetry’(FA),

http://tree.trends.com 0169-5347/01/$ – see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S0169-5347(01)02266-2 556 Opinion TRENDS in Ecology & Evolution Vol.16 No.10 October 2001

Box 1. Canalization, plasticity and developmental stability: some historical landmarks in the evolution of three concepts

We provide original definitions whenever • Bradshaw, 1965e: phenotypic plasticity: • Bradshaw, 1965e: developmental they are short and self-explanatory. ‘Plasticity is shown by a genotype when instability is plasticity. According to Otherwise we provide a concise summary its expression is able to be altered by Bradshaw, stochastic errors during of the author’s ideas. environmental influences…it does not development (referred to as have any implication concerning the ‘developmental instability’) must be Canalization adaptive value of the change occurring, viewed as a manifestation of plasticity, • Waddington, 1942a: canalization: although many types of plasticity may considering that ‘such changes ‘Adjustment of developmental reactions have important adaptive effects’. ultimately have definite environmental so as to bring about one definite end result • Smith-Gill, 1983f: plasticity, developmental causes’. regardless of minor variations in conversion and phenotypic modulation. • Zakharov, 1992h: developmental conditions during the course of the Plasticity is separated by Smith-Gill into homeostasis, homeorhesis, reaction’. two different components: developmental stability. Zakharov • Waddington, 1961b: canalization: ‘The ‘developmental conversion’ is an redefined developmental homeostasis property of a developmental process, of adaptive discrete change; whereas (or ‘homeorhesis’) as the set of being to some extent modifiable, but to ‘phenotypic modulation’ is a continuous mechanisms ensuring phenotypic some extent resistant to modification’. nonadaptive variation. Only the former consistency, including two different and • Schmalhausen, 1949c: autoregulatory is supposed to be genetically based and well-separated processes: canalization, mechanism: ‘a set of processes historically therefore moulded by selection. corresponding to Schmalhausen’s selected, monitoring developmental path’. Additionally, she proposed modulation ‘autonomous-regulatory development’; • Thoday, 1953d: developmental flexibility: to result from ‘a failure of the to and developmental stability, a set of ‘an individual organism may be said to completely buffer development against mechanisms buffering developmental possess flexibility either if its genotype is environmental perturbations’, or explicitly noise sensu Waddington, that is, such that it can develop different formulated, as a lack of canalization. developmental variation among phenotypes in different environments, replicated or symmetrical organs within each phenotype better adapted than the Developmental stability a single organism. others to the environment that evokes it, • Lerner, 1954g: genetic homeostasis, or if its genotype is so balanced that developmental homeostasis, References development is buffered against developmental stability. Lerner defined a Waddington, C.H. (1942) Canalization of environmental variables and hence genetic homeostasis as ‘the property of development and the inheritance of acquired characters. Nature 150, 563–565 apparently the same adaptive the population to equilibrate its genetic b Waddington, C.H. (1961) Genetic assimilation. phenotype results in a range of composition and to resist to sudden Adv. Genet. 10, 257–293 environmental conditions’. changes’. According to him, this property c Schmalhausen, I.I. (1949) Factors of Evolution, depends on population heterozygosity, University of Chicago Press Plasticity as heterozygous individuals buffer d Thoday, J.M. (1953) Components of . • Schmalhausen, 1949c: morphosis and developmental variation more efficiently Symp. Soc. Exp. Biol. 7, 97–113 e Bradshaw, A.D. (1965) Evolutionary dependent autoregulatory than do homozygous ones. Later significance of phenotypic plasticity in . . Schmalhausen researchers adopted this view to interpret Adv. Genet. 13, 115–155 distinguishes two types of ‘environment- the existence of negative correlations f Smith-Gill, S.J. (1983) Developmental dependent development’: ‘morphosis’, a between individual heterozygosity and plasticity: developmental conversion versus nonadaptive, continuous change in morphological variance: homozygous phenotypic modulation. Am. Zool. 23, development correlated to environmental individuals would exhibit variable 47–55 g Lerner, I.M. (1954) Genetic Homeostasis, Oliver variations; and ‘dependent phenotypes and/or high levels of and Boyd autoregulatory morphogenesis’, an fluctuating asymmetry because of a lack h Zakharov, V.M. (1992) Population phenogenetics: environmentally induced variation of of developmental stability or analysis of developmental stability in natural adaptive value, historically selected. developmental homeostasis. populations. Acta Zool. Fenn. 191, 7–30

was used as a canalization index25,26. However, Current uses of the three concepts Waddington12,13 again argued that FA had Several definitions of canalization, plasticity and nothing to do with his canalization concept, because developmental stability are currently in use (Box 2). the FA of a character does not diminish following The latest review on canalization is by Gibson and canalizing selection. Waddington proposed the Wagner6, who define it as ‘the reduction in terms ‘developmental noise’, to refer to variability of a trait’. However, unlike Waddington, differences among homologous replicated parts they highlight the difference between genetic and within a single individual, and ‘developmental environmental canalization, possibly based on stability’, to refer to the lack of noisiness. different (although potentially overlapping) Waddington’s position has been recently re-advocated genetic and developmental systems. by Zakharov1 (Box 1). Environmental canalization, defined as the

http://tree.trends.com Opinion TRENDS in Ecology & Evolution Vol.16 No.10 October 2001 557

Box 2. Current definitions of the three concepts

Canalization • Eshel and Matessi, 1998g: ‘Plasticity References • Zakharov, 1992a: ‘Similarity of the occurs in the many cases in which a Zakharov, V.M. (1992) Population phenogenetics: expression of the phenotypic character organisms are able to respond to analysis of developmental stability in natural populations. Acta Zool. Fenn. 191, 7–30 under different conditions of changing conditions of the environment b Stearns, S. et al. (1995) The differential genetic development’. with specific modifications of certain and environmental canalization of fitness • Stearns et al., 1995b: ‘The process by traits, in a way that is appropriate to components in melanogaster. which phenotypic variation is reduced preserve the quality of its vital activities’. J. Evol. Biol. 8, 539–557 c Wilkins, A.S. (1996) Canalization: a molecular h by developmental mechanisms.’ • Behera and Nanjundiah, 1995 : genetic perspective. BioEssays 19, 257–262 • Wilkins, 1996c: ‘Genetic capacity to buffer ‘Plasticity is roughly analogous to a d Wagner, G.P. et al. (1997) A population genetic developmental pathways against mixed strategy in the Evolutionary theory of canalization. Evolution 51, 329–347 mutational or environmental Stable Strategy (ESS) language’. e Via, S. et al. (1995) Adaptive phenotypic plasticity: consensus and controversy. Trends perturbations’. Ecol. Evol. 10, 212–217 • Wagner et al., 1997d: ‘The suppression of Developmental stability f Callahan, H.S. et al. (1997) Developmental phenotypic variation.’ • Zakharov, 1992a: ‘The capability for phenotypic plasticity: where ecology and evolution stable development under given meet molecular biology. BioEssays 19, 519–525 g Eshel, I. and Matessi, C. (1998) Canalization, Plasticity conditions’. genetic assimilation and preadaptation: a • Via et al., 1995e: ‘A change in the • Palmer, 1994i: ‘The developmental quantitative genetic model. Genetics 149, phenotype that depends on the stability of an organism is reflected in its 2119–2133 environment…Historically, two main ability to produce an “ideal” form under h Behera, N. and Nanjundiah, V. (1995) An investigation into the role of phenotypic plasticity types of phenotypic plasticity have been a particular set of conditions’. in evolution. J. Theor. Biol. 172, 225–234 recognized: graded responses and • Clarke, 1998j:‘Developmental stability i Palmer, P.A. (1994) Fluctuating asymmetry discrete or switched responses’. refers to buffering processes that reduce analysis: a primer. In Developmental Instability: • Callahan et al., 1997f: ‘The ability of an the variation resulting from Its Origins and Evolutionary Implications (Markow, T.A., ed.), pp. 335–364, Kluwer organism to alter its physiology, developmental accidents’. Academic Publishers morphology or development in • Auffray et al., 1999k: ‘The ability of j Clarke, G. (1998) The genetic basis of response to changes in its organisms to withstand genetic or developmental stability. V. Inter- and intra- environment’. environmental disturbances during individual character variation. Heredity 80, 562–567 d • Wagner et al., 1997 : ‘Macroenvironmental development, so as to produce a k Auffray, J-C. et al. (1999) Shape asymmetry and sensitivity’. predetermined phenotype’. developmental stability. In On Growth and Form: • Stearns et al., 1995b: ‘The sensitivity of a • Van Dongen and Lens, 2000l: ‘An Spatiotemporal Patterning in Biology (Chaplain, trait to change in an environmental individual’s ability to buffer its M.A.J., ed.) pp. 310–324, John Wiley & Sons l Van Dongen, S. and Lens, L. (2000) The factor measures its phenotypic development against random evolutionary potential of developmental plasticity’. perturbations’. instability. J. Evol. Biol. 13, 326–335

Table 1. The interpretation of canalization, phenotypic plasticity and developmental stability in a quantitative genetic context Partition of phenotypic variancea Genetic mechanisms Effect on variance

Genetic canalization VP = VA + VNA + VE Epistatic modifiers (genetic-canalizing genes) Negative on VA

Environmental canalization VP = VA + VNA + VE Environment-canalizing genes Negative on VE

Phenotypic plasticity VP = VA + VNA + VµE + VME Gene expression dependent on macroenvironment Positive on VME

Developmental stability VP = Vind + Vwithin Genes controlling homogeneous development of Negative on Vwithin homologous body parts

a Vp, phenotypic variance; VA, additive genetic variance; VNA, nonadditive genetic variance; VE, environmental variance; VµE, microenvironmental variance;

VME, macroenvironmental variance; Vind, among-individual variance; Vwithin, within-individual variance.

‘insensitivity of a character to environmental factors’3 Canalization, plasticity and developmental stability as seems to be the opposite of phenotypic plasticity27,28. patterns: a quantitative genetic interpretation By contrast, genetic canalization and phenotypic Each of the concepts of canalization, plasticity and plasticity are not mutually exclusive: they can developmental stability is associated with a ‘combine to form canalized reaction norms’27. Gibson component of phenotypic variance that can be and Wagner6 do not distinguish developmental measured using the appropriate quantitative genetic stability from canalization, because they list FA design (Table 1). Genetic canalization has studies29,30 as examples of canalization. Therefore, traditionally been associated with directional they assume implicitly that the regulation of selection experiments: if selection proves relatively symmetry is just a special case of the canalizing inefficient in moving the trait value outside a certain mechanisms controlling phenotypic variance, still a interval, this trait is said to be canalized within this controversial position31–33. interval. The quantitative genetic interpretation of

http://tree.trends.com 558 Opinion TRENDS in Ecology & Evolution Vol.16 No.10 October 2001

this resistance to selection is a decrease in additive mutual adjustment of growing left and right parts

genetic variance (VA) owing to a network of epistatic during development. Therefore, developmental interactions. Two populations of genes are usually stability, as measured by symmetry, might involve considered3: additive genes that increase or decrease mechanisms other than environmental canalization. the deviation from the mean by a certain number of phenotypic units; and genetic-canalizing genes that Canalization, plasticity and developmental stability as multiply these deviations by a constant, pushing the processes: the mark of adaptationism phenotype away from, or close to, the mean. Similarly, Since they were first established, the concepts of environmental canalization can reflect the action of canalization, plasticity and developmental stability

environment-canalizing genes that decrease VE, the have been associated with adaptive interpretations, environmental variance. whether in the form of open statements or hidden Phenotypic plasticity is usually measured using an assumptions. experimental design involving several replicates of a Canalization is considered to be adaptive in a genotype (such as clones or full siblings) distributed context of stabilizing selection. Both Waddington’s among a few treatments (e.g. temperatures). The ‘canalization’ and Schmalhausen’s ‘auto-regulatory environmental variance is decomposed into a mechanism’ are assumed to evolve as adaptive fine- microenvironmental component (within treatment, tunings following a major environmental and/or

VµE ), and a macroenvironmental component (among genetic change that initially increased phenotypic treatment, VME). Only VME is used to quantify plasticity. variance. For example, after a major has This suggests that environmental canalization and been fixed, many minor genes will be selected to plasticity involve different kinds of phenotypic canalize the new developmental pathway. Genetic

variation (VµE and VME, respectively). However, the models incorporate stabilizing selection explicitly as the distinction between macro- and microenvironment is motor of the evolution of canalization3. They assume artificial: macroenvironment (e.g. temperature) that the canalized pathway is optimal in all genetic merely represents the environmental parameters one and environmental conditions prevailing at that stage chooses to manipulate, whereas microenvironmental in the population. Recently, fluctuating selection was variation represents the uncontrolled parameters shown to promote canalization, in apparent (e.g. small uncontrolled variation in developmental contradiction with previous studies34. However, temperature). In natural populations, the fluctuating selection only works when the best long- environments experienced by different individuals term strategy is to remain phenotypically invariant: will differ in many ways, and there is no distinction this kind of fluctuating selection can, therefore, be between macro- and microenvironment. Therefore, considered as a special case of stabilizing selection. plasticity and environmental canalization must act Developmental stability is considered adaptive in on the same kind of phenotypic variation. much the same way as is canalization. First, Developmental stability is not measured using the developmental homeostasis26,35 was taken by Lerner21 classic genetic/environmental partition of phenotypic to be a property directly related to fitness. Second, in variance. Instead, within-individual variance all FA studies the optimum phenotype is always (represented by FA) is separated from among- considered to be perfect symmetry (FA = 0). When there individual variance. The underlying assumption is are reasons to reject this assumption (e.g. directional that symmetrical parts of the organism represent the asymmetry or antisymmetry), the data are excluded replicated expression of a single genotype in a single or statistically corrected to obtain a mean FA of zero8. environment: the differences between replicates Several studies have questioned seriously the validity therefore reflect the intrinsic variance of the of the relationship between FA and fitness36. However, developmental process, when environmental variation direct or indirect correlations between FA and fitness is reduced to a minimum. However, this minimum have been documented. For example, wing FA might variance must originate in minute, undetected, decrease flight precision directly37, and FA in male environmental events (including statistical fluctuations sexual ornaments might decrease mating success in cellular processes) that differentially affect the directly37. FA might also reflect indirectly a fitness- development of replicated parts within an individual. related quality, such as individual condition37. Finally, It is implicitly assumed that, even in an ideal world, the existence of developmental stability, a process the environment would vary among individuals more reducing random departure from an ideal phenotype, than it would among replicated parts of the same cannot be justified without assuming that this ideal individual. According to this view, developmental (e.g. perfect symmetry) is an adaptive optimum. The noise is considered as the lowest attainable value of occasional lack of empirical evidence for a relationship

VE, and developmental stability could be interpreted between phenotypic deviance (e.g. asymmetry) and as a special component of environmental canalization, fitness might be a consequence of the efficiency of dealing with these small environmental differences developmental stability processes, keeping the among different parts of the body. However, this view departures from the optimum at an extremely low ignores potential developmental correlations among level in natural populations. Unfortunately, this these parts: for example, symmetry might depend on hypothesis seems hardly falsifiable.

http://tree.trends.com Opinion TRENDS in Ecology & Evolution Vol.16 No.10 October 2001 559

Phenotypic plasticity is used in a different context, concepts, considered as manifestations of , in which the selective optimum differs among the assumptions necessary to refer to them, and some environments. Although most quantitative of the patterns one might use as evidence for their geneticists do not refer to adaptation in the definition action in empirical contexts. We are aware of the of plasticity, experimental studies almost always difficulties involved in defining adaptation itself present adaptive arguments, by defining the precisely40. However, this problem is beyond the scope macroenvironment in such a way that different of this paper. phenotypic optima are expected under different macroenvironmental conditions. Indeed, plasticity Questions that remain plays a prominent role in adaptation to Both canalization and plasticity are often heritable environmental conditions in many organisms19. Some and can therefore evolve; but are there genes that are authors even include adaptation explicitly in the specialized in plasticity or canalization? Are the definition of plasticity38,39. mechanisms of canalization and plasticity organism- wide or character-specific? Is there any relationship The need for synthetic definitions between environmental and genetic canalization, or Although definitions are always conventional, those are they independent? What is the link, if any, of canalization, plasticity and developmental between developmental stability and canalization? stability remain unclear because evolutionists Studies of developmental stability or canalization disagree on which convention to use. First, should have often highlighted heterogeneous responses adaptive arguments be included in definitions? On among traits37,41,42. A long-standing assumption was the one hand, purely phenomenological definitions that traits highly constrained by natural selection are simpler and acknowledge the fact that were more buffered against both developmental noise phenotypic variation is often a nonadaptive matter of and environmental/genetic influences (i.e. both stable fact. On the other hand, most people referring to and canalized). For example, it has been argued that canalization, plasticity or developmental stability characters for which asymmetry largely affects are, in fact, interested in the adaptive aspects of these fitness might exhibit less FA than other traits31,33,34. processes. Second, the adaptive view itself is However it is hard to recognize such characters subdivided into two options, strongly dependent on a priori, and no consensus has been achieved, the trait being studied. People studying plasticity inconsistency being the most consistent result37. consider the ability to vary as the adaptive strategy. For those studying developmental stability or New methodologies for new perspectives canalization, the adaptive strategy is phenotypic In the past two decades, new approaches based on consistency. As a consequence, a given environment- mathematical models of shape have been developed. dependent phenotypic change will be called plasticity Geometric morphometrics is a powerful way to if it is believed to be adaptive, and lack of canalization quantify complex shape variation and has been used if it is not. increasingly in ecology, palaeontology and Synthetic definitions of canalization, plasticity and systematics43,44. Landmark-based studies45 and developmental stability must emphasize the tight outline analyses46 are now common. However, the relationship between living organisms and their determinants of phenotypic variability have been variable external environment. Organisms must keep addressed only recently. The relationship between their internal milieu constant enough to preserve developmental stability and canalization has been physiological reactions and adjust themselves to investigated32,33 using landmark data. In these changing external conditions, simultaneously. Any studies, shape variation and asymmetry of complex organism will stand within the theoretical limits of structures (fly wing and mouse skull) were quantified total environmental dependence and total autonomy, to evaluate the congruence between canalizing and all intermediary positions being influenced by stabilizing effects. Although these studies led to developmental constraints and historically inherited different results, this approach is powerful and could regulating mechanisms. Because of the history of the be extended fruitfully to a wide range of models and concepts of canalization, plasticity and developmental characters. The same methodology could also be used stability, it seems impossible to remove adaptive to investigate the relationship between genetic and considerations from the definitions. Moreover, purely environmental canalization. phenomenological definitions are not free of Recently, molecular developmental genetics has complications: for example, according to Bradshaw16, provided new perspectives on monitoring the microenvironmental variance is plasticity, mechanisms. Indeed, Rutherford and Lindquist47 because it must represent developmental changes reported an increase in morphological variation in related to unidentified and/or small variations in the Hsp90 mutants in Drosophila. Following artificial environment. However the existence of selection, these variants increased in frequency and environmental variance is not usually considered as a their expression became independent of Hsp90, proof of plasticity. Rather, some sort of suggesting that they relied on cryptic mutations is needed. In Box 3, we propose definitions of the three initially buffered by Hsp90. This escape from http://tree.trends.com 560 Opinion TRENDS in Ecology & Evolution Vol.16 No.10 October 2001

Box 3. The concepts: definitions, assumptions and evidence

Genetic canalization changes in selection regime (e.g. photoperiod allows plants to A set of processes historically selected to keep the phenotype prepare themselves to low temperatures). Although many authors constant in spite of . Assumption: the canalized trait associate discontinuous changes with adaptive plasticity and must have the same selective optimum in the range of genetic continuous responses with nonadaptive physiological contexts considered. Possible evidence: resistance to selection when constraints, we do not think that discontinuity is a strong the mean phenotype is close to an optimum; resistance to mutation argument for adaptiveness. Many traits (binary or meristic) have pressure; and existence of uncanalized variants with increased intrinsically discontinuous responses, and these responses are not response to selection and/or to mutation. necessarily optimized by natural selection. Conversely, continuous responses might often be adaptive. Environmental canalization Some reaction norms (i.e. the set of phenotypes produced by one A set of processes historically selected to keep the phenotype genotype in a range of environmental parameters) cannot be said constant in spite of environmental variation. Assumption: the a priori to be either adaptive or counteradaptive. In this case, one canalized trait must have the same selective optimum in the range of should stick to a phenomenological description such as environments considered. Possible evidence: resistance to ‘developmental sensitivity to parameter x’ (corresponding to nonspecific environmental variation, such as among-individual Bradshaw’s original definition of phenotypic plasticity). The use of environmental variation in a population. Sudden increase in the word ‘plasticity’ alone is ambiguous, given its controversial link phenotypic variance (uncanalization) when extreme with adaptation. environments are used, to which the has not been able to adapt during its history. Existence of different variants with Developmental stability

different values of VµE in the same environment. A set of mechanisms historically selected to keep the phenotype constant in spite of small, random developmental irregularities Adaptive phenotypic plasticity potentially inducing slight differences among homologous parts A set of processes historically selected to produce different within individuals. Assumption: symmetry (or identical phenotypes in relation to some environmental parameter. development of replicated body parts) represents the fitness Assumption: the plastic trait has different optima for different values optimum. Possible evidence: within-individual variation is smaller

of the environmental parameter. Possible evidence: each form than is microenvironmental variation among individuals (VµE), even has higher fitness in the environment in which it developed, than when all individuals are genetically identical (clones) and placed in that seen in other forms developed in different environments. The the best-controlled environment. Different genetic variants have case for adaptive phenotypic plasticity is strengthened when: the different levels of fluctuating asymmetry. sensitivity to environmental variation is restricted to one particular environmental parameter or set of parameters; and/or the proximal N.B. Some of the developmental and genetic mechanisms ensuring signal eliciting the developmental switch between two forms does environmental canalization might also ensure developmental stability, but this is not influence fitness by itself, but is a correlated indicator of not necessarily true.

canalization is a ‘plausible mechanism for promoting To investigate the fundamental nature of the evolutionary change in otherwise entrenched monitoring processes, the advances achieved through developmental processes’36. This is, to our knowledge, shape variation quantification, developmental the first direct evidence of a molecule involved in modelling49 and developmental molecular genetics50 canalization. Although it is not formally proven that must be integrated. The example of Hsp90 will boost Hsp90 has been historically selected for its buffering such complementary approaches that might properties, it remains the best candidate so far for canalization, plasticity and developmental stability such a scenario48. firmly in a developmental molecular framework49–52.

References 7 Via, S. et al. (1995) Adaptive phenotypic plasticity: 13 Waddington, C.H. (1961) Genetic assimilation. 1 Zakharov, V.M. (1989) Future prospects for consensus and controversy. Trends Ecol. Evol. 10, Adv. Genet. 10, 257–293 population phenogenetics. Sov. Sci. Rev. F. 212–217 14 Waddington, C.H. (1957) The Strategy of the Physiol. Gen. Biol. 4, 1–79 8 Palmer, A.R. (1994) Fluctuating asymmetry Genes, Macmillan 2 Zakharov, V.M. (1992) Population phenogenetics: analysis: a primer. In Developmental Instability: 15 Schmalhausen, I.I. (1949) Factors of Evolution, analysis of developmental stability in natural Its Origins and Evolutionary Implications University of Chicago Press populations. Acta Zool. Fenn. 191, 7–30 (Markow, T.A., ed.), pp. 335–364, Kluwer 16 Thoday, J.M. (1953) Components of fitness. Symp. 3 Wagner, G.P. et al. (1997) A population genetic Academic Publishers Soc. Exp. Biol. 7, 97–113 theory of canalization. Evolution 51, 329–347 9 Bernard, C. (1865) Introduction à l’Etude de la 17 Woltereck, R. (1909) Weiterer experimentelle 4 Palmer, A.R. (1996) Waltzing with asymmetry. Médecine Expérimentale, Baillière Untersuchungen uber Artveranderung, Speziell BioScience 46, 518–532 10 Cannon, W.B. (1932) The Wisdom of the Body, uber das Wessen Quantitativer Artunterschiede 5 Schlichting, C.D. and Pigliucci, M. (1998) Norton bei Daphniden. Versuch. Deutsch Zool. Phenotypic Evolution, a Reaction Norm 11 Waddington, C.H. (1942) Canalization of Geselschaft 19, 110–172 Perspective, Sinauer development and the inheritance of acquired 18 Johannsen, W. (1911) The genotype conception of 6 Gibson, G. and Wagner, G.P. (2000) Canalization characters. Nature 150, 563–565 heredity. Am. Nat. 45, 129–159 in evolutionary genetics: a stabilizing theory? 12 Waddington, C.H. (1953) and 19 Bradshaw, A.D. (1965) Evolutionary significance of BioEssays 22, 372–380 evolution. J. Exp. Zool. 7, 187–199 phenotypic plasticity in plants. Adv. Genet. 13, 115–155

http://tree.trends.com Opinion TRENDS in Ecology & Evolution Vol.16 No.10 October 2001 561

20 Smith-Gill, S.J. (1983) Developmental plasticity: 31 Clarke, G.M. (1998) The genetic basis of 42 Auffray, J-C. et al. (1999) Shape asymmetry and developmental conversion versus phenotypic developmental stability. V. Inter- and intra- developmental stability. In On Growth and Form: modulation. Am. Zool. 23, 47–55 individual character variation. Heredity 80, Spatiotemporal Patterning in Biology (Chaplain, 21 Lerner, I.M. (1954) Genetic Homeostasis, Oliver 562–567 M.A.J., ed.), pp. 310–324, John Wiley & Sons and Boyd 32 Klingenberg, C.P. and McIntyre, G.S. (1998) 43 Rohlf, F.J. and Marcus L.F. (1993) A revolution 22 Mitton, J.B. (1978) Relationship between Geometric morphometrics of developmental in morphometrics. Trends Ecol. Evol. 8, heterozygosity for loci and variation of instability: analyzing patterns of fluctuating 129–132 morphological characters in natural populations. asymmetry with procrustes methods. Evolution 44 Reyment, R.A. (1991) Multidimensional Nature 273, 661–662 52, 1363–1375 Paleobiology, Pergamon 23 Mitton, J.B. and Grant, M.C. (1984) Associations 33 Debat, V. et al. (2000) Independence between 45 Bookstein, F.L. (1991) Morphometric Tools for among protein heterozygosity, growth rate, and developmental stability and canalization in the Landmark Data. Geometry and Biology, developmental homeostasis. Annu. Rev. Ecol. skull of the house mouse. Proc. R. Soc. London B Cambridge University Press Syst. 15, 479–499 Biol. Sci. 267, 423–430 46 Rohlf, F.J. (1990) Fitting curves to outlines, In 24 David, P. (1999) A quantitative model of the 34 Kawecki, T.J. (2000) The evolution of genetic Proceedings of the Michigan Morphometrics relationship between phenotypic variance and canalization under fluctuating selection. Workshop (Rohlf, F.J. and Bookstein, F.L., eds), heterozygosity at marker loci under partial Evolution 54, 1–12 pp. 197–177, University of the Michigan Museum selfing. Genetics 153, 1463–1474 35 Tebb, G. and Thoday, J.M. (1954) Stability in of Zoology 25 Mather, K. (1953) Genetical control of stability in development and relational balance of 47 Rutherford, S.L. and Lindquist, S. (1998) HSP90 development. Heredity 7, 297–336 X- in . as a capacitor for morphological evolution. Nature 26 Reeve, E.C.R. (1960) Some genetic tests on Nature 174, 1109–1110 396, 336–342 asymmetry of sternopleural chaetae number in 36 Clarke, G.M. (1998) Developmental stability and 48 Rutherford, S.L. (2000) From genotype to Drosophila. Genet. Res. 1, 151–172 fitness: the evidence is not quite so clear. Am. Nat. phenotype: buffering mechanisms and the storage 27 Stearns, S.C. et al. (1995) The differential genetic 152, 762–766 of genetic information. BioEssays 22, 1095–1105 and environmental canalization of fitness 37 Moller, A.P. and Swaddle, J.P. (1997) Asymmetry, 49 Klingenberg, C.P. and Nijhout, H.F. (1999) components in Drosophila melanogaster. J. Evol. Developmental Stability, and Evolution, Oxford Genetics of fluctuating asymmetry: a Biol. 8, 539–557 University Press developmental model of developmental 28 Wilkins, A.S. (1997) Canalization: a 38 Eshel, I. and Matessi, C. (1998) Canalization, instability. Evolution 53, 358–375 molecular genetic perspective. BioEssays 19, genetic assimilation and preadaptation: a 50 Pigliucci, M. (1996) How organisms respond to 257–262 quantitative genetic model. Genetics 149, 2119–2133 environmental changes: from phenotypes to 29 Clarke, G.M. and McKenzie, J.A. (1987) 39 Behera, N. and Nanjundiah, V. (1995) An molecules (and vice versa). Trends Ecol. Evol. 11, Developmental stability of insecticide investigation into the role of phenotypic plasticity 168–173 resistant phenotypes in blowfly; a result of in evolution. J. Theor. Biol. 172, 225–234 51 Wake, D.B. et al. (1991) Development and canalizing natural selection. Nature 325, 40 Reeve, H.K. and Sherman, P.W. (1993) Adaptation evolution: the emergence of a new field. In The 345–346 and the goals of evolutionary research. Q. Rev. Unity of Vol. 1. Proceedings 30 Davies, A.G. et al. (1996) Scalloped wings: is Biol. 68, 1–32 of the Fourth International Congress of lucilia cuprina notch homologue and 41 Clarke, G.M. (1998) The genetic basis of Systematic and Evolutionary Biology candidate for the modifier of fitness and developmental stability. IV. Individual and (Dudley, E.C., ed.), pp. 582–588, College Park asymmetry of diazinon resistance. Genetics 143, population asymmetry parameters. Heredity 80, 52 Cossins, A. (1998) Cryptic clues revealed. Nature 1321–1337 553–561 396, 309–310

Newsletters A new service from BioMedNet, Current Opinion and Trends

Now available, direct to your email box: free e-mail newsletters highlighting the latest developments in rapidly moving fields of research. Teams of editors from the Current Opinion and Trends journals bring you news from a broad perspective:

Evolution of Infectious Disease Newsletter from emerging infectious diseases, host–pathogen interactions and the impact of genomics to the evolution of drug resistance

Comparative Genomics Newsletter from the evolution of genomes by gene transfer and duplication to polymorphisms and the discovery of genes involved in human disease

Each newsletter features news articles from the BioMedNet newsdesk, as well as highlights from the review content of the Current Opinion and Trends journals. Access to full-text journal articles is available through your institution. Newsletters are sent out six times a year. To sign up for Newsletters and other alerts via e-mail, visit http://news.bmn.com/alerts

http://tree.trends.com