1 Introduction

The organism often responds to environmental changes through alterations or modifications of its structure and development.Such environment-dependent alterations are more dramatic in plants than in animals, because plants have modular architecture and indeterminate growth as well as a sessile habit.The capacity of a plant to alter its morphology and function to match changing environments is mediated by phenotypic plasticity.Phenotypic plasticity refers to different phenotypic expression of the same genotype in a variety of environmental conditions (Sultan, 1987).Phenotypic plasticity has been recognized to play an important role in plant adaptation by buffering the effect of natural selection acting on genotypes (Bradshaw, 1965; Grant, 1985; Scheiner, 1993a; Schlichting, 1986; Westcott, 1986).Where genetic variation for plasticity exists, a population with a different mean plasticity can evolve (Bruno et al., 1997).Phenotypic plasticity also applies to character correlations that may be altered by the environment, in their sign as well as in their magnitude (Schlichting, 1989; Stearns et al., 1991; van Tienderen et al., 1996).In the past decade, the study of how phenotypic diversity in different environmental conditions is generated by developmental plasticity has been a key challenge in plant evolutionary biology (Schlichting et al., 1998).Particular emphasis has been placed on the adaptive significance of phenotypic plasticity (Lortie et al., 1996; Pigliucci et al., 1996; Schmitt et al., 1996) and the conditions under which plasticity may evolve (Gillespie et al., 1989; Schlichting et al., 1995; Via et al., 1985; 1987;1995).

Two different viewpoints have been proposed to understand the evolution of phenotypic plasticity.Via(1993) suggested that the phenotypic plasticity of a trait evolves since the trait' s mean value evolves.Scheiner(1993b)and Schlichting and Pigliucci (1993)argued that phenotypic plasticity can evolve as an independent "trait" because plasticity is under the control of a genetic system different from that for the trait mean.Quantitative genetic analyses can be employed to judge these two viewpoints by estimating the genetic correlation between the mean and plasticity of a trait between two different environments (Wu, 1997).The magnitude of genetic correlation reflects the genetic dependence of two quantitative traits, due to either pleiotropy or genetic linkage, or both (Falconer, 1989).If a genetic correlation is significantly different than 1.0, this indicates that the traits studied are not under the control of the completely same genetic systems.If the genetic correlation is significantly larger than zero, this indicates that at least the partially same genetic basis is shared between the two traits.The genetic mechanisms underlying phenotypic plasticity have been further studied by locating the so-called plasticity genes on chromosomes with the aid of molecular markers (Pigliucci, 1996; Stratton, 1998; Wu, 1998a).

A field that has not been well synthesized yet in the literature is the phenotypic plasticity of tree architecture, the genetic variation of this plasticity and the way in which the plasticity evolves.The form and structure of trees present phenomenal phenotypic variation (Hall et al., 1978; Tomlinson, 1983; Wu et al., 1998), and have been studied both empirically and theoretically (Chen et al., 1994; Honda et al., 1978; 1979;Niklas, 1986; 1994;Powell et al., 1986; Tomlinson, 1982; Wu, 1994a; Wu et al., 1994; 1996).Although there is a major debate about whether tree architecture is the consequence of adaptation to particular environments (Horn, 1971), strong evidence exists for the environment-induced sensitivity of tree morphogenesis (Brunig, 1976; Hinckley et al., 1992; Isebrands et al., 1988; Kuuluvainen et al., 1991;Wu et al., 1998).Perhaps the most straightforward example comes from the formation of sylleptic branches in relation to environmental variation.Sylleptic branches are a type of branches that develop from lateral buds of the parent shoots without an intervening period of dormancy (Hall et al., 1978.Fig. 1A).Experiments based on several temperate tree species have shown that sylleptic branching is strongly conditioned by the environment in which trees are grown (Ceulemans et al., 1990; Ceulemans et al., 1996; Dunlap et al., 1995; Hinckley, 1996; Powell et al., 1986; Scarascia-Mugnozza, 1991; Tromp, 1996; Tromp et al., 1996).However, the degree to which the environment determines the number and size of sylleptic branches varies among populations (Dunlap et al., 1995) and genotypes (Wu et al., 1998).

Plastic variation in sylleptic branchiness results from the altered development of a tree.The underlying developmental components of this plasticity are associated with the coherent change of other integrated traits.For example, in two clonally replicated trials of Populus, Wu and Stettler(1998)found that genotypes displaying strong plasticity in sylleptic branches were also very responsive to the environment in stem size and form.Thus, an individual tree' s developmental program for expressing sylleptic branches can explain and constrain the tree' s capacity to respond phenotypically to environmental variation through its life.Yet, the genetic mechanism underlying sylleptic branching is mediated by the environment and shared by other functionally related physiological processes.In this article, we describe evidence for the environmental and developmental plasticity of sylleptic branches and the adaptive significance of such plasticity in changing environments.We also discuss results from a molecular genetic mapping experiment that reveal the nature of genetic control over sylleptic formation and strength.

2 Tree structure:syllepsis vs.prolepsis

For tropical trees, branches can develop from axillary buds by two opposed branching processes, syllepsis and prolepsis (Hall et al., 1978).Syllepsis is a continuous branching process in which an axillary bud can grow without previous rest and form sylleptic branches, whereas prolepsis is a rhythmic branching process in which an axillary bud can grow after a period of dormancy and form proleptic branches.Although these two processes can occur in individual trees, they normally display marked discrepancies in morphogenesis (Tomlinson, 1983).In syllepsis, the bud develops immediately into a branch that grows simultaneously with its parent trunk, producing a hypopodium without basal bud-scales; the first leaves are of nearly normal size.By contrast, in prolepsis the branch develops after a period of rest as a lateral bud.A series of bud-scale leaves are present, together with a scar from the subtending leaf, and there is a gradual transition to normally developed foliage leaves.

Although branching in most temperate trees is entirely by prolepsis, syllepsis can be detected for species in such genera as Acer, Alnus, Betula, Cedar, Fagus, Larix, Malus, Populus, Prunus, and Tsuga.In these species, alternate occurrences of sylleptic and proleptic branches constitutes the framework structure of a tree (Fig. 1B and 1C).However, the tree' s overall form is determined by the interaction of these two branch types and the leader shoot, which displays dramatic diversity among different species(Zimmermann et al., 1971).Trees with no sylleptic branches are considered to have strong apical dominance, whereas those with many sylleptic branches, weak apical dominance.The relative vigor of the leader and proleptic branches represents the degree of apical control (Cline, 1991).Strong apical control occurs when the leader outgrows proleptic branches.For a tree whose branching is purely proleptic, the lateral buds for the current year remain unextended, that is, the leader shoot remains unbranched until next year.In this case, strong apical dominance and weak apical control lead to the formation of a decurrent crown.On the other hand, if a tree has strong sylleptic branching, then lateral buds extend contemporaneously with their parental axis, that is, the leader shoot can be branched.Vigorous sylleptic branching frequently inhibit the growth of proleptic branches due to the competition for growth resources (Wu et al., 1996).Thus, an excurrent crown is formed under weak apical dominance and strong apical control (Zimmermann et al., 1971).

Because the relative strength of sylleptic and proleptic branches determines the form and structure of a crown, great attention has been paid to examining their differences in structure.Comparative studies using Populus show that these two branch types are very different in many morphological features and physiological processes.Sylleptic branches are generally smaller, more tapered, and less curved than proleptic branches formed in the same year (Ceulemans et al., 1990; Wu, 1994a; Wu et al., 1994; 1996).In spring, buds on sylleptic branches tend to flush earlier and, therefore, be faster to produce mature leaves for photosynthesis, than buds on proleptic branches.At three different growth stages, Scarascia-Mugnozza (1991)found that sylleptic branches exceed proleptic branches in the proportion of carbon translocated to the stem.Hydraulic architecture was also compared between the two branch types.In young sylleptic branches, conductivity is higher than that in young proleptic branches, whereas this trend become inverse when branches thicken (Tomlinson, 1983).

Mature leaves located on sylleptic and proleptic branches display different sizes, shapes, orientations, arrange patterns and total areas (Wu, 1994a; Wu et al., 1994; 1996;1997).Proleptic branches tend to produce leaves with a larger size, more wider shape, larger number and larger leaf area than do sylleptic branches.Leaves on the leader shoot have a strikingly larger size than leaves on both branch types.Yet, leaf shape on the leader shoot is more similar to leaf shape on proleptic than sylleptic branches.This may be due to the same developmental origin for the leader shoot and proleptic branches both of which experience a period of rest (Wu et al., 1996).

It should be pointed out that the morphological differences between sylleptic and proleptic branches are not unchanged depending on genotypes, populations, environments and years.Results from Populus strongly suggest that the environmental plasticity of tree architecture is mainly attributable to the plasticity of sylleptic branches.This is because sylleptic branches are more plastic to environmental changes than proleptic branches and because the plasticity of sylleptic branches can results in the plasticity of other elements in tree construction.

3 Plasticity of sylleptic branches

The mechanisms underlying tree architecture are based on deterministic plans or models of development that may subsequently be modified by reiteration and other opportunistic changes in structure (Fisher, 1992; Hall et al., 1978; Tomlinson, 1983).Whereas the deterministic, or genetic, factors predispose a tree towards a specific architecture (prototype), the opportunistic components modify it in response to the unique environment the individual experiences.The opportunistic components may include predictable (e.g., different climate or soil types) and unpredictable (e.g., stochastic errors internal or external to the organism)environmental variation (Wu, 1997).The relative importance of deterministic and opportunistic factors in designing tree architecture has been unclear.Using a clonal trial of a segregated F₂ pedigree derived from two morphologically diverged poplar species, Populus trichocarpa and P.deltoides, Wu and Stettler(1994;1996)were able to estimate the relative contribution of deterministic (genetic) and unpredictable opportunistic factors to tree morphology and development.The estimates of broad-sense heritability, i.e., the proportion of the total phenotypic variance accounted for by genetic factors, ranged from 0.40 to 0.80 for crown structural traits including branch number, length, leaf size, leaf number and leaf area.However, the heritability values for these traits were larger on sylleptic than proleptic branches, which suggests that as compared to proleptic branches, sylleptic branches have genetically a greater potential to defend against unpredictable environmental errors.

This insight into the control of crown architecture was further enhanced by analyzing the predictable opportunistic influence on branch structure and leaf arrangement in two different clonal trials of the same poplar pedigree including the original parents, F₁ parents and F₂ progeny (Wu et al., 1997).In two contrasting environments, one east of the Cascade Range in Boardman, northeastern Oregon (continental climate) and the other west of the Cascades near Clatskanie, northwestern Oregon (maritime climate), the same genotypes from the pedigree displayed different developmental patterns and growth forms.Sylleptic branches produced in the preceding year were compared to proleptic branches produced in the current year, because the latter actually were preformed in the same year and, thus, were responding to the same environmental cues as the former.It was found that sylleptic branches are much more plastic to the growth environment than are proleptic branches.For example, total leaf area on sylleptic branches was higher, by a factor of 1.2 ~ 36.5, in three generations of the pedigree from a higher-light, warmer and well-irrigated site (Boardman)than from a lower-light, cooler and non-irrigated site (Clatskanie)(Fig. 2A).By contrast, total leaf area on proleptic branches was not significantly increased despite an improvement of growth environment (Fig. 2B). For the majority of genotypes in the F₂ family, heavier(more and larger)sylleptic branches were produced in the optimal growth environment of Boardman than in the sub-optimal environment of Clatskanie (Fig. 3A).By contrast, numbers and lengths of proleptic branches showed only a minor upwards shift at Boardman (Fig. 3B).The strong plasticity of sylleptic branches in number and size may reflect a poplar' s capacity for architectural alterations by which the tree can respond opportunistically to changes in its environment.However, it will be interesting to further observe whether sylleptic branches display larger environment-dependent differences in leaf thickness, leaf orientation relative to the sun, and leaf structure than proleptic branches, because these morphological and anatomical properties of leaves are often associated with plant adaptation to the amount of sunlight and stress (Smith et al., 1997).Moreover, terrestrial plants respond to the environmental regimes of a given habitat by evolving leaf structural characteristics in concert with leaf orientational capabilities.

1) Wu R.1999.Manuscript in preparation.North Carolina State University, Raleigh, NC.

The environmental influences on sylleptic branching were also observed in other temperate species.In Malus, Tromp(1992a;b)found that sylleptic shoot growth increase with increasing temperature.By exposing young apple trees to two temperature conditions during three different periods, Tromp and Boertjes(1996)demonstrated that continuous high temperature in all periods was critical for vigorous sylleptic growth.Although high humidity may result in enhanced sylleptic growth, its effect is largely dependent on the interaction with soil temperature (Tromp, 1996). A gradually decreasing photoperiod can decrease the ability of Malus to form sylleptic shoots later in the season (Génard et al., 1994).However, genetic variation occurs for such photoperiod sensitivity; for example, certain genotypes produce sylleptic branches early in the growing season while others, much later (Ceulemans et al., 1990; 1996).Because of a close relationship between the rate of sylleptic-branch growth and environmental conditions, sylleptic branchiness has been used as an indicator of environmental adaptation (Powell et al., 1986), although the underlying ecological mechanisms of this relationship are unknown (Tomlinson, 1983).

The sensitivity of sylleptic branching to environmental variation is genetically variable, which thus provides fun-damental raw material for the evolution of plant structure and form.In a common-garden study of Populus trichocarpa, clones from mesic river valleys were found to display more and larger sylleptic branches than clones from xeric river valleys in Washington (Dunlap et al., 1995).Populus deltoides displays many sylleptic branches in a continental environment but forms no sylleptic branches in a maritime environment, whereas P.trichocarpa only has slight changes in the number and length of sylleptic branches between the two environments (Wu et al., 1998).In the F₂ population derived from these two species, the difference in the number and length of sylleptic branches between the two environments varied, with some genotypes displaying consistency between the two sites and others pronounced differences (e.g., Fig. 3A).A few F₂ recombinants showed stronger plasticity in sylleptic branches than the original parents.The degree to which genotypes respond differently to environment can be quantified by a quantitative genetic method (Wu, 1997).The estimated broad-sense heritabilities for the phenotypic plasticity of the number and length of sylleptic branches were around 0.50, similar to the heritabilities for those traits within individual environments (Wu et al., 1998).However, the genetic correlation between the plasticity and trait mean of sylleptic branches was significantly different from 1.0 (Fig. 4A), indicating that the plastic response of sylleptic branches could be viewed as an independent character on their merit and under the control of unique genetic systems (Schlichting, 1986; 1989;Scheiner, 1993).However, since this genetic correlation is significantly greater than zero, genotypes with a larger capacity to express syllepsis tend to be more plastic to variable environments than genotypes with a lower capacity.As compared to sylleptic branches, the phenotypic plasticity of proleptic branches seems to be more independent of their trait mean across environments (Fig. 4B), suggesting that proleptic branches and their plasticity do not co-evolve to respond to a change in environment (Wu, 1997).

Some developmental factors may also be critically important for the formation and growth of sylleptic branches. The vigor of the main stem shows a strong positive association with syllepsis in a number of woody perennials, such as Larix (Powell, 1987; 1988;Powell et al., 1986), Malus(Tromp, 1996), Myrsine(Wheat, 1980), Populus (Wu, 1994b; Wu et al., 1994; 1996), Prunus(Kervella et al., 1995) and Tsuga (Powell, 1991).In contrast, main-stem vigor is not useful in predicting the mumber of preformed leaves on proleptic branches (Powell, 1991).At equal stem growth rates, sylleptic branching occurs more frequently in early than late season (Génard et al., 1994), which suggests that the emergence date of the meristems interacts with growth characteristics to determine the probability of sylleptic branching (Kervella et al., 1995).

In summary, sylleptic branches, as an adaptive strategy of many temperate woody plants to environment, are affected by genetic, developmental and environmental factors.The phenotypic plasticity of sylleptic branches is the consequence of interactions among these factors.At this point, however, little is known about how these factors interact to determine the morphological and developmental alteration of sylleptic branches in response to environmental changes.

4 The function of sylleptic branches :a plasticity analysis 4.1 Environmental plasticity of functioning

Biological function is implied by the close association between a particular structure and a known physiological or ecological process (Farnsworth et al., 1995).In forest trees, the rate of stem growth is the physiological process that is targeted for successful natural competition and breeding practice.The close relationships between stem growth and crown structural traits, such as branch number, branch length, branch orientation, leaf size, total leaf number and total leaf area, have been observed in many studies (Dunlap et al., 1995; Ford E D, 1985;1992;Ford E D et al., 1990; Ford R et al., 1990; Harrington et al., 1997; Hinckley et al., 1989; 1992;McCurdy et al., 1987; Nelson et al., 1981; Ridge et al., 1986; Roden et al., 1990).These relationships frequently display a significant genetic basis that can be quantified by genetic correlations (Wu, 1994b; Wu et al., 1996).

New recognition about the genetic correlations between stem growth and crown structural traits has been obtained from the study of poplar hybrids by Wu and Stettler(1998), who found that these genetic correlations are subjected to a strong environmental impact.Changes in the correlation structure of morphological traits among different environments have been found in a number of annual species (Chapin et al., 1993; Schlichting, 1986; van Tienderen et al., 1996).The change of trait correlations with environments occurs when a difference exists between two traits in the relative degree and rank with which various genotypes respond to an environmental stimulus.For example, in poplars, total leaf area on sylleptic branches responds to the environment similarly among the original parents, F₁ parents and F₂ progeny of the pedigree (Fig. 2A), whereas total leaf area on proleptic branches does not (Fig. 2B).It is thus expected that a change occurs in the across-generation relationships of total leaf areas on sylleptic and proleptic branches between the two environments.

The changes of genetic correlations suggest that plastic responses may involve major reorganization of the rela tionships among traits, thus raising questions about how morphological integration is maintained in the face of environmental change.Thomas et al.(1971)and Primack and Antonovics (1981)suggested that morphological integration increases as the environment becomes more stressful.Indeed, in forest trees, leaf and branch components of a canopy are more closely associated with stem growth and production under sub-optimal than optimal conditions (Hinckley et al., 1992; Wu et al., 1998).The response of plants to low-resource environments often involves an integrated physiological process, such as a decline in both growth rate and the rate of acquisition of all resources (Chapin, 1991).

Sylleptic branches may play a greater role in regulating the plastic response of physiological integration than proleptic branches.To test this hypothesis, we compare the results from three different field trials of hybrid poplars. The first trial was established with unrooted cuttings near Puyallup, Washington.In this trial, resource acquisition should be limited at early stages due to less developed rooting systems.We found that sylleptic branches, their numbers, sizes and leaf area they carry, have closer genetic relationships with stem growth than do their proleptic branch counterparts (Wu et al., 1994; 1996).The other two trials were established with rooted cuttings in two contrasting environments (Boardman and Clatskanie, Oregon).In these two trials, trees should be less resource-limited, because they have well-developed rooting systems.No significant difference was detected in the genetic correlations of sylleptic and proleptic branches with stem growth traits at both sites, regardless of their remarkable differences in environments (Wu et al., 1998).

What physiological mechanisms are associated with the process of physiological integration conferred by sylleptic branches?Here, we offer an explanation.The relationship between morphology and growth can be biochemically ascribed to the production and distribution of photosynthate within trees, which are affected by the gradient in assimilate concentration and the distance between source and sink (Gifford et al., 1981; Watson et al., 1984).In a resource-limited environment, the formation and elongation of roots are essential for a tree' s survival at its early stages of growth.These processes can stimulate root meristems to become strong sinks for photosynthate.In order to meet growing requirements for photosynthate, plants tend to increase photosynthesizing leaf areas through the production of sylleptic branches in the current year.Thus, leaves on sylleptic branches provide immediate sources to supply photosynthate for root growth.By the uptake of more water and nutrients from soils, well-developed roots lead in turn to a favorable feedback for stem growth.It should be noted, however, that the final number and size of sylleptic branches produced under resource-limited condition depend on the balance between the strength to which roots induce these branches and the level of resources.The relationship of sylleptic growth and root development was also proposed in Malus by Tromp and Boertjes (1996), who speculated that cytokinin production in the growing root tips plays a signaling role in the outgrowth of lateral buds.Recent hormone studies in annual plants have suggested that roots use hormones, or their precursors, to provide shoots with early warning of deteriorating soil conditions in ways that increase resilience to stress (Jackson, 1993; 1997).

The other evidence for strong integration between sylleptic branches and stem growth has been obtained from a radioactive tracer experiment in two-year-old poplars from unrooted cuttings (Scarascia-Mugnozza, 1991).In this experiment, sylleptic branches was observed to translocate a greater proportion of carbon to the stem than proleptic branches, which tended to allocate carbon to their own apical region.This difference in carbon export behavior seemed to increase during a growing season.In mid-May, sylleptic branches on the first-year stem height increment (SYL1)exported 45 % of the fixed carbon to the stem, whereas proleptic branches on the same increment (PRO1) exported only 38 %(Fig. 1C).In mid-September, SYL1 were exporting 70 %while PRO1 were still exporting 38 %. Molecular genetic evidence from Bradshaw and Stettler' s (1995)quantitative trait locus (QTL)mapping strongly supports these physiological findings in that several QTLs governing stem radial growth were found to be coincident with QTLs for leaf traits on sylleptic branches on the same genetic linkage group.

4.2 Developmental plasticity of functioning

Plastic responses in sylleptic branches represent developmental events-alterations of the trajectory of ontogeny that can have profound effects on later stages of a tree' s life cycle.Alterations in trajectory are referred to as developmental plasticity (Pigliucci, 1998).For one- or two-year-old poplars, tree architecture is simple and affected only by sylleptic and proleptic branches formed in the first two years (Fig. 1B and C).However, for older poplars, these two branch types develop alternatively in many years and are located at multiple crown positions (Wu, 1994a).The architecture of these older trees is the consequence of the interaction between branch types and crown positions at which sylleptic or proleptic branches develop.In a trial established with unrooted cuttings of poplars, stem growth in year 2 is largely a function of branch types, with growth traits being correlated more strongly with sylleptic than proleptic branches (Wu et al., 1994; 1996).However, by year 3, sylleptic branches become less important to growth, which is in turn determined by branches at upper crown positions.Again, this developmental plasticity of the function of sylleptic branches may be related to rooting systems.Owing to the rapid development of roots, clones with heavier (more and larger)sylleptic branches have a great opportunity to outperform clones without, or with lighter (fewer and smaller), sylleptic branches at early stages of growth.However, this advantage disappears after stable rooting systems are formed (Scarascia-Mugnozza, 1991).

The developmental plasticity of sylleptic branches in function may be associated with the environment.Under optimal environmental conditions, sylleptic branches display larger genetic correlations with stem height than stem basal area during the establishment year (Wu et al., 1998).It is not surprising that more resources tend to be allocated into height growth than radial growth during the establishment phase in high-resource environments, since a taller tree can better compete against its neighbors that also fully express their growth potential.When trees develop into more complex structures in subsequent years, however, more resources should be allocated to radial growth in order to maintain the taller tree' s spatial advantage.In sub-optimal conditions, a different pattern was observed for the developmental plasticity of the integration between sylleptic branches and growth traits (Wu et al., 1998).

In general, the developmental and environmental control of syllepsis and its functioning can provide an explanation for a biologically important phenomenon, that is, why plants can well match changing environments throughout their lives (Watson et al., 1995).As Cheverud(1984)has suggested, if the structure of genetic correlations among traits has arisen under certain environmental conditions, then the flexibility of that correlation structure in response to heterogeneous environmental and developmental conditions should evolve as well.

5 Developmental and environmental impacts on the genetic structure of sylleptic branches

The fact that different leaf forms or branch types can be produced on the same tree shows that the plant develop mental system can accommodate major morphological switches without the disruption of development.Plant develop ment mayinvolve a particular hierarchical design in which plants activate complex developmental pathways as a response to either environmental or internal signals.As sessile organisms, plants need to respond to local environmental signals and thereby alter their phenotype in a way that best adjusts them to local conditions.

The genetic bases for these morphological changes in trees have been examined using quantitative genetic analysis.In an F₁ hybrid population of Populus deltoides ×P.simonii, Wu(1994a;b)detected different degrees of genetic controlover the same traits, e.g., leaf size and leaf orientation, at different crown positions, the current terminal and sylleptic and proleptic branches.A further structural analysis on the F₂ trees of P.trichocarpa ×P.deltoides suggested that broad-sense heritabilities for leaf and branch morphologies are a function of branch type and crown position (Wu et al., 1996).In year 2, the heritability levels were higher on sylleptic branches or upper crown positions than on proleptic branches or lower positions.This trend was maintained into year 3, but the differences in heritability were mainly a function of position.Significant genetic correlations were not detected for the same morphologi cal traits between different branch types or crown positions, suggesting that different genetic bases exist for different developmental components of a tree.

Nevertheless, environment is very important in affecting the difference in the degree of genetic control over syllepsis and prolepsis.For example, heritabilities for traits on sylleptic branches were greater than, or similar to, those for traits on proleptic branches in optimal conditions, whereas in many cases heritabilities were less for traits on sylleptic than proleptic branches under sub-optimal conditions (Wu et al., 1998).The changes of heritability values for the same trait with environments have been reported in many species (reviewed in Hoffmann et al., 1997a). There are ongoing debates as to whether specific environmental conditions will tend to increase or decrease heritabilities in a consistent way (Hoffmann et al., 1997b; 1998).

Developmental genetics has shown that single genes can activate complex developmental pathways, and that mutations in these genes can dramatically alter the course of morphogenesis.In most cases these genes encode either known or putative transcription regulators.The developmental switch between syllepsis and prolepsis may be controlled by a single dominant gene, but contingent upon the environment (Wu et al., 1998).For example, in the clonally replicated trial of Clatskanie, 375 F₂ genotypes derived from P.trichocarpa and P.deltoides were classified into two groups based on the presence or absence of sylleptic branches in year 1²⁾.A χ² test indicated that these two groups did not deviate significantly from the expected 3 :1 segregation ratio of a dominant gene within each of the three replicates studied (Tab. 1).Given that the shoots of P.trichocarpa are prone to syllepsis whereas those of P.deltoides are largely proleptic, it is possible that the favorable allele for sylleptic branches is derived from the P.trichocarpa parent and dominant over the other negative allele from the P.deltoides parent.However, as shown by a molecular genetic experiment (Wu, 1998), it seems also possible that the P.deltoides parent contributes favorable sylleptic alleles to its F₂ progeny.

2) Wu R and Stettler R F.1999.M anuscript in preparation.University of Washington, Seattle, WA.

In this trial, we further found that some F₂ genotypes did not produce sylleptic branches consistently in all replicates.For example, a total of 42 genotypes generated sylleptic branches in replicate 1 but not in replicate 3.This finding suggests that the direction of additive effect and the dominant recessive relationship of the gene affecting the formation of sylleptic branches are affected by other minor genes, mediated by environmental signals²⁾.In the other trial at Boardman, a 3 :1 segregation pattern was not observed for the same F₂ progeny, suggesting that multiple genes exist to simultaneously determine the developmental switch between syllepsis and prolepsis.Thus, the genetic control of syllepsis is environmentally sensitive, potentially with more genes to affect sylleptic branches under favorable than non-favorable conditions.

The presence or absence of sylleptic branches was used as a morphological marker to associate with the number and length of sylleptic branches in the second year from the same or different trials.We found that the sylleptic marker could account for over 30 % of the total genetic variance for these traits in either trial.Such results suggest that variation in sylleptic branchiness is controlled by two genetic components, a major dominant gene, as identified by a χ²-test(see above), that determines the presence or absence of sylleptic branches by specifying the fate of axillary meristems, and many other epistatic genes that modify the strength of sylleptic branches.With the aid of DNA based molecular markers, Wu(1998)was able to map individual quantitative trait loci (QTLs)that govern the numbers, increments, and mass of sylleptic branches formed in different years in Populus.In most cases, these QTLs showed gene action from partial dominance to overdominance.Both major gene (called syl1) and epistatic genes are regulated by environmental signals.For example, P.deltoides is able to respond to unfavorable environmental conditions (such as shading, low temperature and restricted moisture)by forming no or few slender sylleptic branches; when environmental conditions are favorable, it forms robust sylleptic branches (Dickmann et al., 1990; Wu et al., 1998).It is possible that syl1 is involved in regulating this response by altering plant architecture.We propose the following model for the function of syl1 in P.deltoides, which likely contributes a negative recessive for this major gene.Under favorable environmental conditions, the allele of syl1 for this species is inactive, allowing axillary meristems to develop into lateral sylleptic branches.Yet, further elongation of these branches in mediated through other quantitative genes.Under unfavorable conditions, the allele of syl1 for this species is activated so that the plant produces no or few sylleptics.

The syl1 gene may have pleiotropic effects on stem growth or be linked with genes that condition growth traits, because it explains large proportions (20 %~ 30 %)of the genetic variance for height and basal area in both Populus trials²⁾.In the corresponding QTL experiment, the QTLs for both the number of sylleptic branches and stem basal area growth in year 2 mapped to the same linkage groups E and O (Bradshaw et al., 1995).This coincidence was confirmed from the third-year data, although only a single common linkage group, O, remained that carries these two kinds of QTLs (Wu, 1998).A similar conclusion can be inferred for Larix laricina in that greater stem cross-sectional area and dry mass were observed for trees with heavy than light sylleptic branches (McCurdy et al., 1987; Powell 1987).In fact, many "qualitative" genes with large effects on the phenotype have been identified in annual plants, such as maize, using DNA-based genetic linkage maps (Doebley et al., 1995a; 1995b;Dorweiler et al., 1993; 1997).These genes, for example, Sos1, tb1, and tga1, were found to determine several key steps in the morphological evolution of maize from its progenitor, teosinte.The cloning of tb1 enabled Doebley et al.(1997)to observe different expression of tb1 in maize and teosinte.We anticipate that tree architecture adapted to a particular environment could be selected through the genetic manipulation of syl1-like genes.

6 Conclusions

Tree architecture is a unique predictor of growth rate and life-history traits for woody plant species and their diversity represents one important form of morphological evolution.For tropical species and some temperate species, the relationships between tree architecture and key physiological or ecological processes can be modified by sylleptic branches, which are subject to strong natural selection and are amenable to genetic and developmental study.Extensive analyses indicate that dramatic diversity occurs in sylleptic branching among species and populations, as well as among the environments in which trees are grown.It is recognized that the plastic response of sylleptic branches can explain and constrain a tree' s capacity to respond to changes in environment.Under favorable conditions, considerable production of sylleptic branches indicates that these branches can effectively use excess resources.On the other hand, strong integration with growth traits under non-favorable conditions indicates that sylleptic branches have great capacity to buffer environmental stress.

The plasticity of sylleptic branches is controlled by genetic factors.Recently, a debate has arisen about how the phenotypic plasticity of a quantitative trait is genetically generated in heterogeneous environments.Whereas some genetic analyses suggest that the phenotypic plasticity of a trait is the consequence of natural selection on this trait in individual environments, other studies suggest that plasticity itself is an independent trait that is subject to natural selection.We have used quantitative genetic approaches to estimate the genetic correlation between the plasticity of sylleptic branches and their mean value across two environments, which provides a powerful means for testing these two hypotheses.Our results suggest that the plasticity of sylleptic branches is under the control of its unique genetic system that is different from that for the expression of sylleptic branches in individual environments.Thus, the alteration of sylleptic branches in response to environmental changes is achieved by directly activating genes for plasticity.

The development and plasticity of sylleptic branches cannot be isolated from other elements of tree architecture. Interactions of sylleptic branches with growth characteristics, proleptic branches and below-ground carbon allocation determine the adaptive response of tree architecture.However, as indicated by results from field trials established with rooted and unrooted cuttings in poplars, sylleptic branches may play a particular role in regulating physiological and morphological integration in the circumstances where resources are limited.In the trial of unrooted cuttings, sylleptic branches were more strongly correlated with growth than proleptic branches, whereas such a difference did not happen in the trial of rooted cuttings.We propose that such functional mechanisms of sylleptic branches may be associated with the development of root systems, regulated by hormones, and can be explained by a source-sink relationship.

The formation and growth of sylleptic branches are also related to the resource status of a tree' s shoot leader and their functioning changes with stand development.Classic studies have shown that the suppression of lateral bud growth by the plant apex is controlled by auxin.More recent work has demonstrated the importance of auxin-cytokinin interactions in regulating such apical dominance (Coenen et al., 1997).Therefore, further studies that attempt to explain the relationship between plant adaptation and the phenotypic plasticity of sylleptic branches should be directed toward understanding how a tree' s developmental "blueprints" and hormonal signaling interact with the environment to affect the occurrences of sylleptic branches and identifying the genetic mechanisms that underlie this interaction.Although these are not an easy task, new techniques and materials have now been available.By a means of asexual reproduction, such as cutting propagation, tissue culture or embryogenesis, the same genotype can be repeated both temporarily and spatially, which thus provides a robust way to estimate phenotypic plasticity.For some species, such as Populus, the crossing of trees with contrasting features with regard to syllepsis and prolepsis can be used to study the pattern of segregation of branching habits in the progeny.Molecular markers and gene mapping allow an assessment of how many, and which quantitative trait loci control the environmental variation in sylleptic branches.Recent developments of molecular techniques, such as cDNA microarrays (Service, 1998), lead to faster and more efficient discoveries of genes that are differentially expressed in sylleptic branches across a variety of environments.With these techniques, we may be close to answering some of our oldest and most exciting questions regarding tree adaptation.

Acknowledgments

The senior author is grateful to Prof.R.R.Sederoff and other members of the Forest Biotechnology Group at North Carolina State University for encouragement and support on this and other studies.He expands his thanks to R.F.Stettler, T.M.Hinckley, R.Ceulemans, D.M.O' Malley, S.E.McKeand, J.M. Dunlap, J.G.Isebrands, M.Topa, B.Reztlaff, and J.E.Grissom for helpful discussions regarding syllepsis and tree development.This study is partially supported by the NCSU Biotechnology Industrial Associates and a grant from the Ministry of Forestry, China.