The Ecology of Heteroblasty in Acacia

The ecology of heteroblasty in Acacia

Michael A. Forster

PhD Candidate Evolution and Ecology Research Centre School of Biological, Earth and Environmental Sciences University of New South Wales

Supervisor: Dr Stephen P. Bonser

COPYRIGHT STATEMENT ‘I hereby grant the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all proprietary rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation. I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstract International (this is applicable to doctoral theses only). I have either used no substantial portions of copyright material in my thesis or I have obtained permission to use copyright material; where permission has not been granted I have applied/will apply for a partial restriction of the digital copy of my thesis or dissertation.'

Signed ……………………………………………......

Date ……………………………………………......

AUTHENTICITY STATEMENT ‘I certify that the Library deposit digital copy is a direct equivalent of the final officially approved version of my thesis. No emendation of content has occurred and if there are any minor variations in formatting, they are the result of the conversion to digital format.’

Signed ……………………………………………......

1 Date ……………………………………………...... THE UNIVERSITY OF NEW SOUTH WALES Thesis/Dissertation Sheet

Surname or Family name: FORSTER

First name: MICHAEL Other name/s: ANTHONY

Abbreviation for degree as given in the University calendar: PhD

School: Biological, Earth and Environmental Sciences Faculty: Science

Title: The ecology of heteroblasty in Acacia

Abstract 350 words maximum: (PLEASE TYPE)

Heteroblasty defines a dramatic change in leaf form and function along a shoot and is a prominent feature of the genus Acacia (Mimosaceae). Function of different leaf types in Acacia (i.e. compound leaf versus phyllode) is well established yet it is unknown whether heteroblasty is a plastic trait. A fully factorial designed experiment established the light environment, and not nutrients or water, had a significant influence on heteroblastic development. Compound leaves, which have higher specific leaf area (SLA), are retained for longer under low irradiance and, specifically, under a low Red:Far Red light environment. Plants grown in high intraspecific density environments also retained compound leaves for longer. Blue light signals and greater ultraviolet radiation had no effect on heteroblastic development. Heteroblasty is thought to aid in seedling establishment however across all experiments there was no consistent evidence of improved plant performance. Rather, there was an optimal allocation of biomass to organs where resources were most limiting and this was more influential in assisting seedling establishment. Lastly, a meta-analysis of a global dataset of leaf traits found compound leaves to be similar to simple leaves but offset towards the cheap to construct with fast returns region of the leaf economics spectrum.

Declaration relating to disposition of project thesis/dissertation

I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation.

I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstracts International (this is applicable to doctoral theses only).

…………………………………………………………… ……………………………………..……………… ……….……………………...…….… Signature Witness Date

The University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptional circumstances and require the approval of the Dean of Graduate Research.

FOR OFFICE USE ONLY Date of completion of requirements for Award:

ACKNOWLEDGEMENTS

To my family for all their on-going support: Mum and Dad; Marge and John; Luke; John and Therese; and all the others.

To my friends who helped out and kept my motivation: Clayton, Jo and Benny; Sally, Brian, Todd and Ant; Leah, Natalie, Mick and Jamie.

To my fellow colleagues for all their professional support: Brenton, Ross, Geoff, Frank, Angela, Kylie, Paul, Louise, Burhan, Craig, David, David, Derek, Sigfredo, Mel, Cate, Markus, Marion, Yan-Shi, Remko, Keyne and Kristin.

To all those students who entertained and kept me updated with the latest things, especially Leah and Rhiannon.

To University of Western Sydney for loan of invaluable scientific equipment.

And lastly to my supervisor, Stephen Bonser, for giving me the freedom to pursue what I wanted.

4 CONTENTS

Introduction (p. 11)

Chapter 1: Heteroblastic development and the optimal partitioning of traits among contrasting environments in Acacia implexa. (p. 23)

Chapter 2: Heteroblastic development and shade-avoidance in response to Blue and Red light signals in Acacia implexa. (p. 56)

Chapter 3: Ultraviolet radiation alters whole-plant and leaf trait strategies in a heteroblastic species, Acacia implexa. (p. 88)

Chapter 4: Optimal allocation of resources in response to shading and neighbours in the heteroblastic species, Acacia implexa. (p. 124)

Chapter 5: Global ecology of compound and simple leaves. (p. 160)

Discussion (p. 184)

Supplementary Materials (p. 189)

List of Tables (p. 6)

List of Figures (p. 8)

5 LIST OF TABLES

Chapter 1

Table 1.1 Summary of the traits, abbreviations and statistical transformations that were used in the Chapter 1. (p. 48)

Table 1.2 Whole-plant multivariate and univariate analysis. (p. 48)

Table 1.3 Heteroblastic multivariate and univariate trait analysis. (p. 49)

Table 1.4 ANCOVA table following analysis of CompNodes. (p. 50)

Chapter 2

Table 2.1 (A) Brief description of populations of sourced seeds. (B) Specifications of the light filters used to obtain Blue and Red light treatments. (p. 77)

Table 2.2 Summary of traits measured in Chapter 2 with abbreviations and transformations made for statistical analysis. (p. 78)

Table 2.3 ANOVA table following analysis of node of first transitional leaf. (p. 79)

Table 2.4 Multivariate and univariate analysis of variance. (p. 80)

Chapter 3

Table 3.1 Brief description of populations of sourced seeds. (p. 113)

Table 3.2 Summary of traits measured in Chapter 3 with abbreviations and transformations made for statistical analysis. (p. 114)

Table 3.3 Results from multivariate analysis of variance. (p. 115)

Table 3.4 Results from analysis of variance. (p. 116)

Chapter 4

Table 4.1 Brief description of populations of sourced seeds. (p. 149)

Table 4.2 Summary of traits measured in Chapter 4 with abbreviations and transformations made for statistical analysis. (p. 150)

6 Table 4.3 Results from multivariate analysis of variance. (p. 151)

Table 4.4 Results from analysis of variance. (p. 152)

Chapter 5

Table 5.1 Distribution of compound and simple leaf species according to growth form. (p. 175)

Table 5.2 Distribution of compound leaf and simple leaf species according to biome (based on Whittaker, 1975). (p. 176)

7 LIST OF FIGURES

Introduction

Figure 2. Phyllodes and transitional leaves may be displayed in any sequence as demonstrated by Acacia melanoxylon. (p. 20)

Figure 3. The delay in heteroblasty is known to occur along a rainfall gradient in Acacia melanoxylon. (p. 21)

Chapter 1

Figure 1.1 Estimated marginal means and standard errors of whole-plant traits. (p. 51)

Figure 1.2 Estimated marginal means and standard errors of heteroblastic traits. (p. 52)

Figure 1.3 Grouping of treatments and traits following discriminant function analysis. (p. 53)

Figure 1.4 Number of nodes displaying a compound leaf (CompNodes) with standard error bars. (p. 54)

Figure 1.5 Estimated marginal mean and standard errors of total biomass (g) at point of harvest. (p. 55)

Chapter 2

Figure 2.1 Node of first transitional leaf represents the onset of heteroblasty. (p. 81)

Figure 2.2 Means (±S.E.) of measured traits across experimental treatments. (p. 82)

Figure 2.3 Means (±S.E.) of measured traits within the population by Red light interaction treatment. (p. 84)

Figure 2.4 Grouping of population, Blue and Red light treatments following discriminant function analysis. (p. 85)

8 Figure 2.5 Mean (±S.E.) of relative growth rate (RGR) across experimental treatments. (p. 86)

Figure 2.6 Mean (±S.E.) of relative growth rate (RGR) across the full factorial combination of population by Blue by Red light treatments. (p. 87)

Chapter 3

Figure 3.1 Transitional leaves from plants grown in enhanced and reduced UV treatments. (p. 117)

Figure 3.2 The node at which the first transitional leaf developed on plants corresponds to timing of heteroblastic development. (p. 118)

Figure 3.3 Comparisons of whole-plant traits in the UV and population by UV treatments. (p. 119)

Figure 3.4 Comparisons of leaf-level traits in the UV and population by UV treatments. (p. 120)

Figure 3.5 Mean (±S.E.) dry weight of three leaf organs: phyllode, rachis and pinnae; of the first transitional leaf. (p. 121)

Figure 3.6 Grouping of population by UV treatment following discriminant function analysis. (p. 122)

Figure 3.7 Mean (±S.E.) of growth traits in the UV and population by UV treatments. (p. 123)

Chapter 4

Figure 4.2 Means (±S.E.) of whole-plant traits across experimental treatments. (p. 154)

Figure 4.3 Means (±S.E.) of leaf-level traits across experimental treatments. (p. 156)

Figure 4.4 Grouping of population, light and density treatments following discriminant function analysis for whole-plant traits. (p. 157)

Figure 4.5 Grouping of population, light and density treatments following discriminant function analysis for leaf-level traits. (p. 158)

Figure 4.6 Mean (±S.E.) of growth traits across experimental treatments. (p. 159)

Chapter 5

Figure 5.1 (A) Compound and (B) simple leaf showing equivalent structures. (p. 177)

Figure 5.2 Box plots of compound (grey fill) and simple (white fill) of traits associated with the leaf economics spectrum. (p. 178)

Figure 5.3 Box plot summary of Anderson-Rubin factor score derived from principal components analysis of four key leaf economics traits (LMA, LL, Nmass, and Amass). (p. 179)

Figure 5.4 (a) Box plots of photosynthetic nitrogen use efficiency (PNUE) for compound and simple leaf species. (b) Regression of Amass on Nmass for compound (solid circles and line) and simple (empty circles and dashed line) leaf species. (p. 180)

Figure 5.5 Box plots of compound (grey fill) and simple (white fill) of mass-based traits associated with the leaf economics spectrum within growth forms. (p. 181)

Figure 5.6 Box plots of compound (grey fill) and simple (white fill) of mass-based traits associated with the leaf economics spectrum within biomes. (p. 182)

10 INTRODUCTION

Plants display different types of leaves throughout their life-cycle. Each leaf type is thought to serve a function that enables an individual plant to establish, grow, survive and eventually reproduce. While certain types of leaves serve an obvious function, for instance the cotyledons serve the purpose of capturing energy following the depletion of nutrition provided by the seed, certain plant species produce leaves that appear incongruous with its growth and ecology. Prevalent among these species are those that undergo a developmental process called heteroblasty.

Heteroblasty has its linguistic roots in Greek as hetero- means “different” and - blasty means “germ” or “sprout”. Goebel (1898, cited in Wells and Pigliucci,

2000) defines heteroblasty as the switch between juvenile and adult foliage where adulthood is further defined as reproductive maturity. It is considered to be a non- reversible developmental switch whereby the juvenile foliage does not develop following the initiation of adult foliage. This definition however is too simplistic to account for the vast complexity that plant life presents. Instead, argued Jones

(1999), the switch from juvenile to adult foliage and the implicit assumption of reproductive competence should be referred to as “phase change”. This leaves heteroblasty as a more general term that includes “sequential changes among metamers that occur as a normal expression of whole-plant ontogeny” (Jones,

1999; p: S109). In other words, the expression of different leaf types throughout the entire life-cycle of a plant (Jones, 1999), or along a shoot (Leroy and Heuret,

2008), or even over a single growing season (Critchfield, 1960; 1970), is a highly complex thing and the modern usage of heteroblasty encompasses all of these.

The complexity of heteroblasty has led to many instances where its ecological and evolutionary significance is not so obvious. Generally, heteroblasty may be considered to be a generalist strategy where one leaf type may have advantages over an alternative leaf given an environmental context. The ability to produce either leaf type enables the plant to cope with an unpredictable environment.

However heteroblastic development may not always have such inherent plasticity and “sequential changes among metamers” may be genetically fixed (i.e. canalized; Wells and Pigliucci, 2000). For example, many divaricate shrubs in the

New Zealand flora switch from small, carbon rich leaves to large, nitrogen rich leaves at defined heights thought to be out of reach of browsing herbivorous birds

(Greenwood and Atkinson, 1977; Atkinson and Greenwood, 1989; Bond et al.,

2004). In any case there is no consensus as to the universality of heteroblasty and its significance is most likely context dependent.

A ubiquitous group of species that undergo heteroblasty belong to the genus

Acacia subgenus Phyllodineae (Mimosaceae). Acacia is a cosmopolitan genus typically found in tropical and sub-tropical forests and woodlands (New, 1984).

The group is divided into five subgenera each fluidly confined to different continents (Orchard and Maslin, 2003). Phyllodineae is the most species rich subgenus (approx. 1000 spp.) with its greatest representation in Australia although members are found in Asia and across the Pacific (Orchard and Maslin, 2003).

Phyllodineae is unique among the subgenera of Acacia in that it is the only one to exhibit heteroblasty (Kaplan, 1980).

12 The expression of heteroblasty is highly complex between and within species of

Acacia where there is an example of single species displaying eight different leaf types on seedlings (Leroy and Heuret, 2008). Leaf types can nonetheless be categorised into three general groups: juvenile, transitional and adult. Juvenile leaves are double compound – that is the leaflets of the compound leaf are further subdivided into more leaflets (Figure 1a). The leaflets are thus called pinnae and the sub-leaflets are called pinnules. There may be one or more pair of pinnae attached to a rachis and this is species and developmentally dependent.

Transitional leaves still have pinnae attached however the petiole elongates vertically to produce an organ called a phyllode (Figure 1b; Kaplan, 1980). Finally, the adult leaf is purely a phyllode without the development of any pinnae (Figure

1c). Following the onset of transitional leaves, juvenile leaves do not develop thereafter. However adult and transitional leaves may develop in any sequence and transitional leaves can persist for many years in certain species (e.g. A. melanoxylon, Figure 2). Other species (e.g. A. urophylla) may not develop a transitional leaf and may only develop one or two juvenile leaves before the permanent development of adult leaves (Leroy and Heuret, 2008). Very few species (most notably A. rubida) develop both transitional and adult leaves throughout their entire lifetime.

Compound leaves and phyllodes have advantages and disadvantages given multiple ecological challenges. Generally, compound leaves tend to be shade- adapted whereas phyllodes have adaptations to higher irradiances (Brodribb and

Hill, 1993). The amount of mass per photosynthetic area tends to be greater in phyllodes than compound leaves. The greater structural reinforcement of phyllodes

13 means they are less susceptible to damage or herbivory (Turner, 1994). Phyllodes additionally have adaptations suited to arid environments. For example, stomata have a more rapid response to water deficits on phyllodes than on compound leaves (Walters and Batholomew, 1984; Hansen, 1986). Moreover, Hansen and

Steig (1993) found similar photosynthetic rates in phyllodes and compound leaves but lower stomatal conductance in phyllodes. Phyllodes therefore tend to have superior water use efficiency over compound leaves (Brodribb and Hill, 1993).

In spite of our growing understanding of the functional significance of compound leaves and phyllodes there is still no consensus as to the significance of heteroblastic development in Acacia. Aridity is perhaps the most apparent environmental factor that may affect heteroblasty as phyllodes have superior water conservation traits (Boughton, 1986; Brodribb and Hill, 1993). In support of this,

Farrell and Ashton (1978) observed earlier phyllode development along a high to low rainfall gradient (Figure 3). The greater resistance to physical damage and herbivory of phyllodes increases the ability to maintain foliage cover and is an important adaptation in low nutrient environments where the replacement of lost leaf tissue can be extremely costly (Turner, 1994). Lastly, compound leaves have been observed to persist for longer on seedlings growing in forest understoreys

(Withers, 1979) suggesting heteroblasty may be delayed in shaded environments.

Water, nutrients and light are not found in isolation in the natural environment and a combination of these factors may also be driving heteroblastic development in

Acacia.

14 The delay or hastening of certain leaf types is not the only response available to heteroblastic plants encountering contrasting environmental conditions. Optimal partitioning theory (OPT) predicts that plants allocate biomass to above or below ground organs to maximise acquisition of the most limiting resources (Thornley,

1972; Bloom et al., 1985). Above ground parts of the plant, the shoot, captures light whereas the below ground plant, the root, captures water and nutrients. When irradiance decreases plants have a relatively high allocation to above ground biomass. Conversely, when water and/or nutrients decrease, relatively more biomass is allocated below ground. Although a simple model of how a plant functions, OPT has provided a powerful framework for further studies on plant biomass allocation (Tilman, 1988). Yet, OPT has not been applied to heteroblastic species. In particular, it is not known whether heteroblastic species can alter biomass allocation patterns within the shoot (i.e. greater leaf mass ratio of one leaf type over the other) in response to contrasting environments.

The principal aim of this thesis was to determine the environment in which heteroblasty could be delayed or hastened in a model species of Acacia. The second aim was to establish whether heteroblasty was independent of optimal allocation of biomass to organs where resources were most limiting. The third aim was to determine whether heteroblasty or optimal allocation of biomass could increase seedling performance through measures of total dry biomass and relative growth rate. And the fourth aim was to assess heteroblasty and optimal allocation across populations within a species of Acacia in order to determine genetic variability in the expression of these traits.

15 The first chapter examines the onset of heteroblasty and optimal allocation of biomass in a single population of Acacia implexa across nutrient, light and water environment gradients. The findings of this chapter set the basis for the remaining experiments in the thesis as it was found that only a low light environment delayed heteroblastic development and that there was an optimal allocation to organs where resources were most limiting. The second chapter examines Blue and Red light components of photosynthetically active radiation (PAR) and which wavelength is the primary driver of heteroblasty and optimal allocation. The second chapter also introduces three populations within A. implexa sourced from regions of contrasting rainfall (high, medium and low). The third chapter builds on spectral wavelength effects on plant development and biomass allocation by experimentally manipulating ultraviolet radiation. The fourth chapter looks at the effects of intense intra-specific competition for above and below ground resources has on seedling establishment, heteroblasty and optimal allocation. Finally, a meta-analysis is presented in the fifth chapter of a global dataset of leaf physiological traits

(GLOPNET) examining differences between compound and simple leaf species.

This compilation of data allows us to conclude whether differences seen in compound and simple leaves along the shoot in Acacia (i.e. heteroblasty) can be generalised at a global scale as a difference between compound and simple leaf species.

A Note on Thesis Format

Each chapter in this thesis was originally written as independent pieces of work under a common theme. Therefore the reader may find some of the methodology,

16 particularly species description and statistical analyses, and citation of references repetitive.

REFERENCES

Atkinson IAE, Greenwood RM. 1989. Relationships between moas and plants. New Zealand Journal of Ecology, 12S:67-98.

Bloom AJ, Chapin FS, Mooney HA. 1985. Resource limitations in plants - an economic analogy. Annual Review of Ecology and Systematics, 16:363- 392.

Bond WJ, Lee WG, Craine JM. 2004. Plant structural defenses against browsing birds: a legacy of New Zealand’s extinct moas. Oikos, 104:500-508.

Boughton VH. 1986. Phyllode structure, taxonomy and dstribution in some Australian acacias. Australian Journal of Botany, 34:663-674.

Brodribb T, Hill RS. 1993. A physiological comparison of leaves and phyllodes in Acacia melanoxylon. Australian Journal of Botany, 41:293-305.

Critchfield WB. 1960. Leaf dimorphism in Populus trichocarpa. American Journal of Botany, 47:699-711.

Critchfield WB. 1970. Shoot growth and heterophylly in Ginkgo biloba. Botanical Gazette, 131:150-162.

Farrell TP, Ashton DH. 1978. Population studies of Acacia melanoxylon. I. Variation in seed and vegetation characteristics. Australian Journal of Botany, 26:365-379.

Goebel K. 1898. Organographie der Pflanzen, Jena: Fischer.

Greenwood RM, Atkinson IAE. 1977. Evolution of divaricating plants in relation to moa browsing. Proceedings of the New Zealand Ecological Society, 24:21-29.

Hansen D, Steig E. 1993. Comparison of water-use efficiency and internal leaf carbon dioxide concentration in juvenile leaves and phyllodes of Acacia koa (Leguminosae) from Hawaii, estimated by two methods. American Journal of Botany, 80:1121-1125.

Hansen DH. 1986. Water relations of compound leaves and phyllodes in Acacia koa var. latifolia. Plant, Cell and Environment, 9:439-445.

Jones CS. 1999. An essay on juvenility, phase change, and heteroblasty in seed plants. International Journal of Plant Sciences, 160:S105-S111.

Leroy C, Heuret P. 2008. Modelling changes in leaf shape prior to phyllode acquisition in Acacia mangium Willd. seedlings. C R Biologies, 331:127- 136.

New TR. 1984. A Biology of Acacias, Melbourne: Oxford University Press.

Orchard AE, Maslin BR. 2003. Proposal to conserve the name Acacia (Leguminosae: Mimosoideae) with a conserved type. Taxon, 52:362-363.

Thornley JHM. 1972. A balanced quantitative model for root:shoot ratios in vegetative plants. Annals of Botany, 36:431-441.

Tilman D. 1988. Plant Strategies and the Dynamics and Structure of Plant Communities, Princeton: Princeton University Press.

Turner IM. 1994. Sclerophylly - primarily protective. Functional Ecology, 8:669- 675.

Walters GA, Batholomew DP. 1984. Acacia koa leaves and phyllodes: gas exchange, morphological, anatomical and biochemical characteristics. Botanical Gazette, 145:351-357.

Wells CL, Pigliucci M. 2000. Adaptive phenotypic plasticity: the case of heterophylly in aquatic plants. Perspectives in Plant Ecology, Evolution and Systematics, 3:1-18.

Withers JR. 1979. Studies on the status of unburnt Eucalyptus woodland at Ocean Grove, Victoria. IV. The effect of shading on seedling establishment Australian Journal of Botany, 27:47-66.

18 (a) (b) (c)

Figure 1. The three types of “leaf” displayed by Acacia: a) a double compound leaf; b) a transitional leaf with two pinnae attached to a flattened rachis/petiole; c) a modified leaf, the phyllode; a flattened rachis/petiole. Leaves were harvested from a non-experimental plant 4 months after sowing grown in full sunlight from the 4th, 8th and 13th node respectively. Scale bar is 10mm.

Figure 2. Phyllodes and transitional leaves may be displayed in any sequence as demonstrated by Acacia melanoxylon.

Figure 3. The delay in heteroblasty is known to occur along a rainfall gradient in Acacia melanoxylon. OB – high rainfall (Otway Ranges, 1736 mm p.a.); TC – mid rainfall (Turritable Creek, 872 mm p.a.); BR – low rainfall (Brisbane Ranges, 658 mm p.a.). Photo after Farrell and Ashton 1978.

Chapter 1

Heteroblastic development and the optimal partitioning of traits among contrasting environments in Acacia implexa

Authors: Michael A. Forster and Stephen P. Bonser

Previously published as:

Forster, M.A. and Bonser S. P. (2009). Heteroblastic development and the optimal partitioning of traits among contrasting environments in Acacia implexa. Annals of Botany 103: 95-105.

23 ABSTRACT

Optimal partitioning theory (OPT) predicts plants will allocate biomass to organs where resources are limiting. Studies of OPT focus on root mass ratio, stem mass ratio and leaf mass ratio where roots and stems are often further subdivided into organs such as fine roots/tap roots or branches/main stem. Leaves however are rarely subdivided into different organs. Heteroblastic species develop juvenile and adult foliage and provide an opportunity of subdividing leaf mass ratio into distinct organs. Acacia implexa (Mimosaceae) is a heteroblastic species that develops compound (juvenile), transitional and phyllode (adult) leaves that differ dramatically in form and function. Our aims were to grow A. implexa to examine patterns of plastic development of whole-plant and leaf traits under the OPT framework. Plants were grown in a glasshouse under contrasting nutrient, light and water environments in a full factorial design. Allocation to whole-plant and leaf- level traits was measured and analysed with multivariate statistics. Whole-plant traits strongly followed patterns of allocation predicted by OPT. Leaf-level traits showed a more complex pattern to experimental treatments. Compound leaves on low nutrient plants had significantly lower specific leaf area (SLA) and were retained for longer as quantified by a significantly greater compound leaf mass ratio after 120 days. There was no significant difference in SLA of compound leaves in the light treatment yet transitional SLA was significantly higher under the low light treatment. The timing of heteroblastic shift from compound to transitional leaves was significantly delayed only in the low light treatment.

Therefore plants in the light treatment responded at the whole-plant level by adjusting allocation to productive compound leaves and at the leaf-level by adjusting SLA. There were no significant SLA differences in the water treatment

24 despite strong trends at the whole-plant level. Explicitly subdividing leaves into different types provided greater insights into OPT than would otherwise had been the case.

25 INTRODUCTION

Optimal partitioning theory (OPT) predicts that plants allocate biomass to above or below ground organs to maximise acquisition of the most limiting resources

(Thornley, 1972; Bloom et al., 1985) and provides a framework for the study of plasticity in plants. A simple model of a basic plant divides the body into two parts: that above ground and below ground. The above ground plant primarily captures light and carbon resources whereas the below ground plant captures water and mineral resources necessary for nutrition. When light and/or carbon levels decrease plants have a relatively high allocation to above ground biomass.

Conversely, when water and/or mineral resources decrease, relatively more biomass is allocated below ground. This is clearly a basic model of how a plant functions yet it provides a powerful framework for further studies on plant biomass allocation (Tilman, 1988). For example, OPT has provided context for plant competition for limiting resources (Tilman, 1982; Wedin and Tilman, 1993), community succession (Gleeson and Tilman, 1990; 1994) community diversity

(Tilman, 1988; Tilman and Pacala, 1993; Chesson, 2000) and more context specific examples such as resource allocation in resprouting and seeder species

(Bellingham and Sparrow, 2000).

Experiments and studies of OPT often rely on broad categorisation of traits into units such as root to shoot ratio (R:S) (e.g. Li et al., 1999; Andrews et al., 2001;

Zhang et al., 2005; McCarthy and Enquist, 2007). R:S is relatively simple to measure and changes in the ratio are straightforward to interpret. However, Poorter and Nagel (2000) argued that better insights into OPT can be gained by subdividing R:S into more specific traits. Dividing R:S into root mass ratio, stem

26 mass ratio and leaf mass ratio (or fractions, rather than ratios) provides greater functional insights - particularly into growth analyses and the carbon economy

(Poorter and Nagel, 2000). Root, stem and leaf ratios are the bare minimum compartments that should be examined (see Poorter and Nagel, 2000), yet more predictions of optimal partitioning theory could be tested with further subdivisions of these organs. Previous studies have subdivided root mass ratio into fine roots and tap roots (e.g. Wullschleger et al., 2005) and stem mass ratio has been subdivided into main stem and petioles (e.g. McConnaughay and Bazzaz, 1995).

Leaves on the other hand are rarely subdivided since most plants tend to have only a single type of leaf organ or photosynthetic blade.

Leaves are complex organs and are highly variable across species, and sometimes, within species. Different types of leaves have been shown to play various biological and ecological roles (e.g. Boardman, 1977). Therefore the presence of two or more leaves undertaking different roles allows for a subdivision of leaf mass ratio into different organs. For instance, rosette and cauline (stem) leaves have been demonstrated to be favoured in contrasting light environments in

Arabidopsis thaliana (e.g. Bonser and Geber, 2005) and could be subdivided into components of leaf mass ratio. Sun leaves and shade leaves on the same plant are often significantly different (Boardman, 1977) and is another example of two different leaf types that can be easily subdivided within leaf mass ratio.

There are cases of dramatic and sudden changes in leaf form and function that occur in many plant lineages. Heterophylly, where an individual plant exhibits two or more contrasting leaf types (Winn, 1999), defines such plants. Heterophylly is

27 particularly prominent where strong and conflicting selection pressures favour the expression of different leaf types on the same plant. Well known cases of heterophylly involve amphibious or submerged plants where leaf type is plastic according to whether the stem is in an aqueous or terrestrial environment

(Sculthorpe, 1967; Bruni et al., 1996; Wells and Pigliucci, 2000; Mommer and

Visser, 2005; Mommer et al., 2006). A similar concept is heteroblasty where a plant may exhibit two or more leaf types during development that is not necessarily related to a change in environment (Jones, 1999). The shift from one leaf type to another occurs only once and is often used to mark the transition from juvenile to adult development (Goebel, 1898; cited in Wells and Pigliucci, 2000).

The key difference between the concepts relies on reversibility where a heterophyllic species can adjust leaf type many times according to environment conditions whereas a heteroblastic species is constrained to have a single change in the expression of leaf form through development (Jones, 1999; Wells and

Pigliucci, 2000). Nevertheless, a heteroblastic plant can possibly delay or hasten a shift in leaf type given certain environmental conditions and this constitutes an example of plasticity (Wells and Pigliucci, 2000; Burns, 2005). By studying heterophyllic or heteroblastic species and treating their different leaf types as subdivisions of leaf mass ratio may improve our capacity to test predictions from optimal partitioning theory similarly to subdividing traits within root mass ratio and stem mass ratio.

Species in the cosmopolitan genus Acacia subgenus Phyllodineae undergo heteroblastic development from double compound leaves, transitional leaves and then to a modified petiole/rachis called a phyllode (see Figure 1, Introduction;

28 New, 1984). Transitional leaves are unique by having both pinnae of compound leaves and the flattened petiole/rachis of phyllodes. The switch from compound to transitional leaves is heteroblastic in Acacia and its timing is species dependent.

The switch from transitional leaves to phyllodes is also heteroblastic in most species however juvenile leaves can be retained in certain species meaning this is a case of heterophylly. Phyllodes are generally considered to be better adapted to hot, dry and high light conditions whereas compound leaves are more suited to cool, wet and low light (Brodribb and Hill, 1993). It has also been recognised that heteroblastic development is delayed under high rainfall (Farrell and Ashton,

1978) and in forest understoreys (Withers, 1979). Furthermore, the rate of photosynthetic return per unit biomass invested in the leaf is significantly lower in phyllodes than compound leaves (Brodribb and Hill, 1993). This strategy is known to be used by plants grown in nutrient impoverished environments and therefore phyllodes are probably better adapted to these environments. Under three contrasting environments of nutrients, light and water species of Acacia may optimally partition resources to compound leaves, transitional leaves or phyllodes.

The aim of our study was to examine shifts in biomass allocation in a heteroblastic species, Acacia implexa, under contrasting environments of nutrients, light and water in order to test predictions of biomass allocation under the OPT framework.

These three environments have been extensively study under OPT (e.g. Givnish,

1988; Tilman, 1988; Ryser and Eek, 2000; Navas and Garnier, 2002; Chan et al.,

2003; Moriuchi and Winn, 2005) however studies with heteroblastic species have been comparatively rare. Explicitly, the aims of our study were to: (i) to test for differences in the capacity for plants to respond to multiple environmental

29 variables; (ii) subdivide leaf mass ratio into categories of different leaf types (i.e. compound versus transitional leaves) to determine the capacity of the different leaf types to respond to contrasting environments; and (iii) determine whether heteroblasty is a variable trait in A. implexa and under which environment it is most pronounced.

METHODS

Acacia implexa is a tree growing between 5m and 12m (Kodela, 2002). It is a relatively fast growing species primarily inhabiting slopes on shallow, well- drained soils of woodlands and open forest (Maslin and McDonald, 2004). It occurs in a rainfall zone of approx. 400mm to 1000mm and its natural range extends from northern Queensland to southern Victoria (Maslin and McDonald,

2004). Seeds of A. implexa were obtained from the CSIRO Australian Tree Seed

Centre, Canberra, Australia (http://www.ffp.csiro.au/tigr/atscmain) seedlot number

19770. The seeds were sourced from a population of 12 parents from Bylong New

South Wales (32° 37´S, 150° 03´E, mean rainfall 675mm per annum; avg temperature 23°C).

Experimental design

Seeds were pre-treated in boiling water for 2 minutes (scarification is necessary for this species as it germinates following heat induced by fire). Six seeds were sown into a single 115mL pot filled with soil. Soil in each pot consisted of 33%

Australian Native Landscape supply of “Organic Garden Mix” (with a ratio of

50% black soil, 20% coarse sand and 30% organics), 33% washed river sand and

33% cocopeat. Our previous experience in growing Australian Acacia species

30 found good results using this complex soil mixture. Plants within each pot were randomly thinned to one per pot following the emergence of the first double compound leaf. The first double compound leaf was the second leaf to develop following a single compound leaf that developed after the cotyledons.

All pots were assigned to one of either high or low nutrient, light and water treatments. The high nutrient treatment was established by adding a general fertilizer containing 36% dolomite, 14% blood and bone, 14% ammonium nitrate,

14% gypsum, 14% lime and 7% trace elements to the soil mix. Additionally, a 4 month Osmocote low phosphorus slow release fertiliser was added with the N:P:K ratio of 17:1.5:8.5. The low nutrient treatment had no additional fertiliser added to the soil mix. Plants in the high light treatment received natural sunlight whereas the plants in the low light treatment were grown inside a 57cm cylinder (open ended) of green filter plastic (Lee Filters, Andover, UK; number 121 Lee Green).

The light filters simulated a natural canopy where photosynthetic flux density was reduced to 65% and the red : far red ratio was reduced from 1.0 to 0.2 (Bonser and

Geber, 2005). The high water treatment was kept well watered at all times. Plants in the low water treatment were watered after early signs of wilting were shown.

Low water plants were typically watered about once per week. Seeds were sowed directly into high/low nutrient soil and the light and water treatments began after the emergence of the first double compound leaf.

Plants were grown according to a complete randomised full-factorial design. The size of the experiment was: 2 x nutrients, 2 x light, 2 x water, 25 x replicates gave a total sample size of 200 plants. All plants were grown in the School of

31 Biological, Earth and Environmental Sciences glasshouse at the University of New

South Wales, Sydney, Australia. The experiment was conducted over 120 days between February and June 2006 with glasshouse temperatures maintained between 19°C and 26°C. Previous experience with growing this species showed most individuals had undergone heteroblastic development by 120 days and we were interested in determining a “snap-shot” of the plant phenotype under different treatments at a single point in time. 170 plants were harvested while 30 died.

Despite efforts to maintain sterility 9 plants were found to have symbiotic rhizobium (which fix atmospheric nitrogen and thus improve fertility and growth) and these were removed from analysis to give a total sample size of 161 plants.

Mortality was largely random across light and water treatments however the high nutrient treatment had n = 68 and the low nutrient treatment had n = 93.

We focused on measuring a core set of morphological traits that were classified as either whole-plant traits or those traits associated with heteroblasty (i.e. within module leaf traits). Whole plant traits included data from all leaves irrespective of leaf type whereas heteroblastic traits were derived from leaves explicitly divided into different leaf types. A summary of these traits is presented in Table 1.1. Stem length and diameter at soil level were taken before harvest. At harvest leaves were separated from the stem and classified as either compound, transitional or phyllode. Leaves were flattened and refrigerated at 4°C until area measurements were undertaken. This prevented pinnules on the compound leaves from folding and modifying leaf area. Each individual leaf area was measured using LeafA software (G Williamson, University of Adelaide, Australia, ‘pers. comm.’) with a flatbed scanner. Individual leaves were then oven dried at 60°C for at least 7 days

32 and then mass taken. Specific leaf area was calculated as fresh leaf area (cm2) divided by dried mass (g) for each leaf and then averages were taken. Soil was carefully removed from roots in a plastic tray to ensure all root material was collected. Stem and roots were oven dried at 60°C for at least 7 days and then mass was taken.

Data analysis

Multivariate analysis of variance (MANOVA: Wilks’ λ) was used to test for the effects of nutrient, light and water environment on the following traits: RMR,

StMR, LMR, LAR, HtoD, Internode and SLA (Table 1.1). Data were checked for normality using Shapiro-Wilk Test and transformed where appropriate (details of transformations are shown in Table 1.1). Homogeneity of variances in MANOVA were checked with Levene’s test of equality and outliers checked with

Mahalanobis distances. Multivariate sets of covarying traits were further characterised using discriminant function analysis. DFA is a classification procedure that can be applied on factorial multivariate datasets to assess whether groups can be uniquely separated (Quinn and Keough, 2002). In this study we were interested in whether our eight treatment combinations could be classified into unique groups or whether there were overlaps in groups. Means with 95% confidence intervals of experimental treatments were plotted against the first and second discriminant functions using procedures outlined in Johnson et al. (2007).

Univariate F-tests on each trait were examined for each environment and interaction and were ordered from highest to lowest to determine their contribution to the MANOVA (Quinn and Keough, 2002; Monroe and Poore, 2005). In order to examine how heteroblastic traits affect whole-plant traits we undertook the exact

33 same MANOVA procedure as described above. However, in this case the following traits were analysed: RMR, StMR, CompLMR, TransLMR, CompLAR,

TransLAR, HtoD, Internode, CompSLA and TransSLA. Phyllodes were not included as an individual trait in the multivariate analysis as very few plants produced phyllodes across the experiment. Nevertheless, phyllode mass data was included in overall biomass calculations in those plants that did develop phyllodes.

Pairwise comparisons within treatments were sequential Bonferoni corrected to control for multiple comparisons.

We also specifically examined if developmental shift from a compound leaf to a transitional leaf (i.e. heteroblasty) was a plastic trait in A. implexa. An analysis of co-variance (ANCOVA) was performed where the number of nodes which developed a compound leaf (CompNodes - Table 1.1) was the dependent variable, biomass was the co-variate, and nutrient, light and water were fixed factors. All analysis of variance and correlation data analyses were performed using SPSS

15.0.1 (SPSS Inc., Chicago, IL, USA).

RESULTS

Whole-Plant Response

Acacia implexa displayed a suite of plastic responses induced by all three experimental treatments and there were some significant interactions among treatments. MANOVA indicated a strong effect caused by the nutrient and light treatments with the water treatment having a marginally significant effect (Table

1.2a). The light x water interaction was also marginally significant however there

34 was a strong effect in the three-way nutrient x light x water interaction (Table

1.2a). In the nutrient treatment univariate F-tests indicated that the greatest plastic responses were between partitioning of biomass to roots in a low nutrient environment and leaves in the high nutrient environment (Table 1.2b, Figure 1.1).

Only StMR and Internode had no plastic response to nutrients (Table 1.2b).

Conversely, StMR and Internode were highly plastic traits in the light environment as were HtoD and SLA (Table 1.2b). These traits were significantly greater in the low light treatment whereas RMR and LAR were significantly greater in the high light treatment (Figure 1.1). There were less plastic trait responses in the water treatment (Table 1.2b) however RMR and HtoD were higher in the low water treatment and Internode and LMR were higher in the high water treatment (Figure

1.1). We also note that in spite of significant multivariate interactions in the nutrient x light and nutrient x light x water treatments these interactions were not attributable to interactions in any single trait as there were no significant univariate interactions [Supplementary materials 1.1].

Heteroblastic Trait Responses

A suite of plastic responses were also observed when leaf-related traits were subdivided into heteroblastic components. MANOVA results indicated that only the nutrient and light treatments had a significant effect whereas the water treatment had no effect on the multivariate dataset (Table 1.3a). The nutrient x light interaction was highly significant across the heteroblastic traits (Table 1.3a).

Ordering of the F-values in the nutrient treatment showed RMR had the highest rank followed by TransLMR, CompLMR, TransLAR and CompLAR (Table 1.3b).

TransLMR and TransLAR were significantly greater in the high nutrient treatment

35 yet CompLMR and CompLAR were significantly greater in the low nutrient treatment (Figure 1.2). The light treatment still produced a strong plastic response in stem traits (StMR, HtoD and Internode; Table 1.3b) however plasticity in leaf traits was complex. For example, CompLAR and CompLMR were significantly higher in a low light environment yet there was no difference in TransLAR or

TransLMR (Table 1.3b, Figure 1.2). On the other hand, there was no significant difference in CompSLA yet TransSLA was significantly greater in the low light environment (Table 1.3b, [Supplementary materials 1.2]).

Discriminant Function Analysis

The first two discriminant functions explained 93.2% and 84.9% of the total variance among whole plant and heteroblastic traits respectively. The first discriminant function accounted for the variability among the nutrient treatment whereas the second discriminant function accounted for the light treatment (Figure

1.3). Accordingly, biplots of the whole plant traits revealed LMR to be positively and RMR to be negatively associated with the first discriminant function and

StMR, SLA, HtoD and Internode were found to be positively associated with the second discriminant function (Figure 1.3c). Biplots of the heteroblastic traits revealed a more complex pattern also revealed from the MANOVA and ANOVA analysis. CompLMR and CompLAR were positively associated with the first discriminant function and TransLMR and TransLAR were negatively associated corresponding to the groupings of the low and high nutrient treatments respectively

(Figure 1.3d).

36 Heteroblastic Developmental Plasticity

After 120 days of growth not all plants had undergone heteroblastic development with the least amount occurring in the low nutrient treatment [Supplementary materials 1.3]. The heteroblastic shift between compound leaves and transitional leaves was significantly different under the light treatment yet there was no difference in the nutrient or water treatment (Figure 1.4). None of the interactions between treatments delayed the shift in leaf type. Total biomass had no significant effect on the model as a co-variate (Table 1.4) despite only the light treatment showing no difference in total biomass at point of harvest (Figure 1.5).

DISCUSSION

Optimal partitioning theory (OPT) predicts that biomass will be allocated to traits where there are limited belowground or aboveground resources. Patterns in whole- plant allocation of Acacia implexa reported here largely support this prediction.

For example, a greater allocation to roots and leaves in a low and high nutrient environment respectively has been well documented (Reynolds and Dantonio,

1996; Li et al., 1999; Poorter and Nagel, 2000; Navas and Garnier, 2002). We demonstrated that root mass ratio is higher in the low nutrient treatment and leaf mass ratio and leaf area ratio are greater in the high nutrient treatment (Figures 1.2,

1.4c). Further, OPT predicts a greater allocation to stem traits in a low light environment (Tilman, 1988; Poorter and Nagel, 2000). We found that stem mass ratio, height to diameter ratio and mean internode length all increased under the low light treatment (Figures 1.2, 1.4c) supporting the prediction of increased allocation to stem traits. OPT predicts that plants in a low water environment will

37 have proportionally greater allocation to roots and less to leaves (Tilman, 1988) and this prediction is supported by significantly greater root mass ratio in the low water treatment and greater leaf mass ratio in the high water treatment (Figure 1.1).

In spite of strong conformity of whole-plant traits to predictions made by OPT, specific leaf area (SLA) seemed to show conflicting results. Firstly, the water treatment had no effect on SLA (Figure 1.1) despite much research suggesting aridity should decrease SLA as plants increase structural leaf mass to prevent desiccation (e.g. Cunningham et al., 1999; Specht and Specht, 1999; Zhang et al.,

2005). Moreover, leaf area expansion tends to decrease with lack of water as a driving factor. Similarly, nutrient impoverishment should lower SLA as leaves increase structural defences to avoid loss of expensive photosynthetic tissue to herbivory or other damage (e.g. Cunningham et al., 1999; Li et al., 1999; Specht and Specht, 1999; Navas and Garnier, 2002). However in this study SLA was significantly greater in the low nutrient treatment (Table 1.3). From a whole-plant perspective incongruent SLA results are not necessarily conflicting. For example, leaf mass ratio was significantly lower in the low water treatment but leaf area ratio was not significantly different across water treatments (Figure 1.1), suggesting plants produced fewer leaves but maintained a relatively high photosynthetic area. In the high nutrient treatment, plants were significantly larger and developmentally older than plants in the low nutrient treatment (Figure 1.5).

SLA tends to decrease with age (Thomas and Winner, 2002) as the value of highly productive but short-lived leaves diminishes (Bonser, 2006). The response of SLA in the light treatment was much easier to interpret than the nutrient and water treatments as SLA was significantly larger in low light (Figure 1.1). This is a

38 widely accepted pattern allowing plants a greater photosynthetic surface area in a light limited environment (e.g. Thompson et al., 1988; Ellsworth and Reich, 1992;

Dong, 1993, Gunn et al., 1999; James and Bell, 2000; Evans and Poorter, 2001;

Poorter et al., 2006; Rozendaal et al., 2006).

Subdividing leaf traits into different categories of compound and transitional leaves provided a greater understanding of plant growth strategy and biomass partitioning, particularly in the nutrient treatment. The high nutrient treatment had significantly higher transitional leaf allocation (both transitional leaf mass and area) and the low nutrient treatment had higher allocation to compound leaves

(Figures 1.3, 1.4d). This result suggests a case of apparent plasticity where plants in the high nutrient treatment are larger (Figure 1.5) and more developmentally advanced than low nutrient plants and therefore would naturally have more transitional leaves. However results from subdividing SLA into compound SLA and transitional SLA suggest there was a more complex response to the contrasting nutrient treatment - particularly in the development of compound leaves.

Compound SLA is significantly larger in the high nutrient than the low nutrient treatment (Figure 1.2). Plants in the high nutrient treatment produced cheaply constructed but short-lived compound leaves that led to relatively fast growth of aboveground biomass. Plants in the low nutrient treatment on the other hand invested more resources into compound leaves to ensure they were not as cheaply constructed and therefore could obtain a longer lifespan from them. This investment of more resources increased SLA at the expense of further aboveground development. In this way more resources could be allocated to belowground biomass.

The light treatment also produced complex patterns and there was evidence of a plastic response of increasing photosynthetic capture area in the low light environment. Using compound leaves, plants respond to low light by producing more leaf material on an area basis (Figure 1.2). Using transitional leaves, plants respond to low light by producing leaves with higher SLA (Figure 1.2) Compound leaves are more productive than transitional leaves in a low light environment

(Brodribb and Hill, 1993) and therefore there appears to be an adaptive value in developing more of these before the onset of heteroblasty. Yet it is not clear why plants did not also produce compound leaves with relatively higher SLA in the low light treatment as they did with transitional leaves and is generally known to be the case in shade-avoiding species (e.g. Thompson et al., 1988; Ellsworth and Reich,

1992; Dong, 1993; Gunn et al., 1999; James and Bell, 2000; Evans and Poorter,

2001; Poorter et al., 2006; Rozendaal et al., 2006).

At the point of harvest in this experiment it was not obvious whether having more transitional leaves was adaptive as not all plants had developed phyllodes. The development of phyllodes in this species is not heteroblastic but rather is heterophyllic. Individuals at later stages of development that display only phyllodes can subsequently develop transitional leaves if placed in a shaded environment (M Forster, Personal observation). This experiment was run only over

120 days and running a similar experiment for a longer period will assist in examining patterns of allocation between transitional leaves and phyllodes.

40 The developmental timing of heteroblasty was only significantly delayed (i.e. developmentally delayed and not the actual timing) in the low light environment

(Figure 1.4). Maintaining more productive compound leaves in a light limited environment should allow plants to maintain a high growth rate (Brodribb and

Hill, 1993). For example, at the end of the experiment the low nutrient and low water plants were significantly smaller in overall plant size than their high treatment counterparts yet plants in the light environment were the same size

(Figure 1.5). There have also been previous studies documenting no difference in biomass of plants grown in high and low light (Dong, 1993; Tienderen and

Hinsberg, 1996; Hinsberg and Tienderen, 1997). This suggests that some species do not differ in growth according to light environment. Nevertheless, plasticity in heteroblastic development due to differences in the light environment has also been observed in other species (e.g. Day, 1988; James and Bell, 2000; Burns,

2005). In Acacia pycnantha shaded juveniles retained compound leaves for a longer period than juveniles grown in full sun (Withers, 1979). Farrell (1973) also suggests that compound leaves are shade leaves and phyllodes are sun leaves in

Acacia melanoxylon. This would seem an important adaptive strategy in Acacia as members of the genus are pioneer species competing intra- and interspecifically for light as juveniles while adults generally inhabit a relatively higher light environment (New, 1984).

In this experiment plastic response of traits to contrasting environments largely followed predictions made by OPT. However we found that explicitly subdividing leaves into different types produced complex plastic responses that would not have been observed if all leaves were treated as a single functional type.

Funding

UNSW Faculty Research Grant and Early Career Research grant to S.P.B.

Acknowledgements

We thank Brenton Ladd and Geoff McDonnell for assistance and Grant Williamson for leaf area software. We also thank Kevin Burns, David Causton and an anonymous reviewer for helpful comments.

REFERENCES

Andrews M, Raven JA, Sprent JI. 2001. Environmental effects on dry matter partitioning between shoot and root of crop plants: relations with growth and shoot protein concentration. Annals of Applied Biology 138: 57–68.

Bellingham PJ, Sparrow AD. 2000. Resprouting as a life history strategy in woody plant communities. Oikos 89: 409–416.

Bloom AJ, Chapin FS, Mooney HA. 1985. Resource limitations in plants - an economic analogy. Annual Review of Ecology and Systematics 16: 363– 392.

Boardman NK. 1977. Comparative photosynthesis of sun and shade plants. Annual Review of Plant Physiology 28: 355–377.

Bonser SP. 2006. Form defining function: interpreting leaf functional variability in integrated plant phenotypes. Oikos 114: 187–190.

Bonser SP, Geber MA. 2005. Growth form evolution and shifting habitat specialization in annual plants. Journal of Evolutionary Biology 18: 1009– 1018.

Brodribb T, Hill RS, 1993. A physiological comparison of leaves and phyllodes in Acacia melanoxylon. Australian Journal of Botany 41: 293–305.

Bruni NC, Young JP, Dengler NC. 1996. Leaf developmental plasticity of Ranunculus flabellaris in response to terrestrial and submerged environments. Canadian Journal of Botany 74: 823–837.

Burns KC. 2005. Plastic heteroblasty in beach groundsel (Senecio lautus). New Zealand Journal of Botany 43: 665–672.

Chan SS, Radosevich SR, Grotta AT. 2003. Effects of contrasting light and soil moisture availability on the growth and biomass allocation of Douglas-fir and red alder. Canadian Journal of Forestry Research 33: 106–117.

Chesson P. 2000. Mechanisms of maintenance of species diversity. Annual Review of Ecology and Systematics 31: 343–366.

Cunningham SA, Summerhayes B, Westoby M. 1999. Evolutionary divergences in leaf structure and chemistry, comparing rainfall and soil nutrient gradients. Ecological Monographs 69: 569–588.

Day JS. 1998. Light conditions and the evolution of heteroblasty (and the divaricate form) in New Zealand. New Zealand Journal of Ecology 22: 43– 54.

Dong M. 1992. Morphological plasticity in the clonal herb Lamiastrum galeobdolon (L.) Ehrend. & Polatschek in response to partial shading. New Phytologist 124: 291–300.

Ellsworth DS, Reich PB. 1992. Leaf mass per area, nitrogen content, and photosynthetic carbon gain in Acer saccharum seedlings in contrasting forest light environments. Functional Ecology 6: 423–435.

Evans JR, Poorter H. 2001. Photosynthetic acclimation of plants to growth irradiance: the relative importance of specific leaf area and nitrogen partitioning in maximizing carbon gain. Plant, Cell & Environment 24: 755–767.

Farrell TP. 1973. The Ecology of Acacia melanoxylon. M.Sc thesis University of Melbourne, Melbourne, Australia.

Farrell TP, Ashton DH. 1978. Population studies of Acacia melanoxylon. I. Variation in seed and vegetation characteristics. Australian Journal of Botany 26: 365–379.

Givnish TJ. 1988. Adaptation to sun and shade: a whole-plant perspective. Australian Journal of Plant Physiology 15: 63–92.

Gleeson SK, Tilman D. 1990. Allocation and the transient dynamics of succession on poor soils. Ecology 71: 1144–1155.

Gleeson SK, Tilman D. 1994. Plant allocation, growth rate and successional status. Functional Ecology 8: 543–550.

Goebel K. 1898. Organographie der Pflanzen. Jena: Fischer.

Gunn S, Farrar JS, Collis BE, Nason M. 1999. Specific leaf area in barley: individual leaves versus whole plants. New Phytologist 143: 45–51.

43 Hinsberg Av, Tienderen Pv. 1997. Variation in growth form in relation to spectral light quality (red/far-red ratio) in Plantago lanceolata L. in sun and shade populations. Oecologia 111: 452–459.

James SA, Bell DT. 2000. Influence of light availability on leaf structure and growth of two Eucalyptus globulus ssp. globulus provenances. Tree Physiology 20: 1007–1018.

Johnson HE, Llyod AJ, Mur LAJ, Smith AR, Causton DR. 2007. The application of MANOVA to analyse Arabidopsis thaliana metabolomic data from factorially designed experiments. Metabolomics 3: 517–530.

Jones CS. 1999. An essay on juvenility, phase change, and heteroblasty in seed plants. International Journal of Plant Sciences 160: S105–S111.

Kodela PG. 2002. Acacia. In: Harden GJ, ed. Flora of New South Wales. Kensington: University of New South Wales, 381–476.

Li B, Suzuki J, Hara T. 1999. Competitive ability of two Brassica varieties in relation to biomass allocation and morphological plasticity under varying nutrient availability. Ecological Research 14: 255–266.

Maslin BR, McDonald MW. 2004. AcaciaSearch: Evaluation of Acacia as a woody crop option for southern Australia. Canberra: Rural Industries Research and Development Corportation.

McCarthy MC, Enquist BJ. 2007. Consistency between an allometric approach and optimal partitioning theory in global patterns of plant biomass allocation. Functional Ecology 21: 713–720.

McConnaughay KDM, Bazzaz FA. 1992. The occupation and fragmentation of space - consequences of neighboring shoots. Functional Ecology 6: 711– 718.

Mommer L, Pons TL, Visser JW. 2006. Photosynthetic consequences of phenotypic plasticity in response to submergence: Rumex palustris as a case study. Journal of Experimental Biology 57: 283–290.

Mommer L, Visser EJW. 2005. Underwater photosynthesis in flooded terrestrial plants: a matter of leaf plasticity. Annals of Botany 96: 581–589.

Monroe K, Poore AGB. 2005. Light quantity and quality induce induce shade- avoiding plasticity in a marine macroalga. Journal of Evolutionary Biology 18: 426–435.

Moriuchi KS, Winn AA. 2005. Relationships among growth, development and plastic response to environment quality in a perennial plant. New Phytologist 166: 149–158.

44 Navas M, Garnier E. 2002. Plasticity of whole plant and leaf traits in Rubia peregrina in response to light, nutrient and water availability. Acta Oecologica 23: 375–383.

New TR. 1984. A Biology of Acacias. Melbourne: Oxford University Press.

Poorter H, Nagel O. 2000. The role of biomass allocation in the growth response of plants to different levels of light, CO2, nutrients and water: a quantitative review. Australian Journal of Plant Physiology 27: 595–607.

Poorter H, Pepin S, Rijkers T, Jong Yd, Evans JR, Korner C. 2006. Construction costs, chemical composition and payback time of high- and low-irradiance leaves. Journal of Experimental Biology 57: 355–371.

Quinn GP, Keough MJ. 2002. Experimental Design and Data Analysis for Biologists. Port Melbourne: Cambridge University Press.

Reynolds HL, Dantonio C. 1996. The ecological significance of plasticity in root weight ratio in response to nitrogen. Plant and Soil 185: 75–97.

Rozendaal DMA, Hurtado VH, Poorter L. 2006. Plasticity in leaf traits of 38 tropical tree species in response to light; relationships with light demand and adult stature. Functional Ecology 20: 207–216.

Ryser P, Eek L. 2000. Consequences of phenotypic plasticity vs. interspecific differences in leaf and root traits for acquisition of aboveground and belowground resources. American Journal of Botany 87: 402–411.

Sculthorpe CD. 1967. The Biology of Aquatic Vascular Plants. London: Edward Arnold.

Specht RL, Specht A. 1999. Australian Plant Communities: dynamics of structure, growth and biodiversity. South Melbourne: Oxford University Press.

Thomas SC, Winner WE. 2002. Photosynthetic differences between saplings and adult trees: an integration of field results by meta-analysis. Tree Physiology 22: 117–127.

Thompson WA, Stocker GC, Kriedemann PE. 1988. Growth and photosynthetic response to light and nutrients in Flindersia brayleyana F. Muell., a rainforest tree with broad tolerance to sun and shade. Australian Journal of Plant Physiology 15: 299–315.

Thornley JHM. 1972. A balanced quantitative model for root:shoot ratios in vegetative plants. Annals of Botany 36: 431–441.

Tienderen PHv, Hinsberg Av. 1996. Phenotypic plasticity in growth habit in Plantago lanceolata: how tight is a suite of correlated characters? Plant Species Biology 11: 87–96.

Tilman D. 1988. Plant Strategies and the Dynamics and Structure of Plant Communities. Princeton: Princeton University Press.

Tilman D, Pacala S. 1993. The maintenance or species richness in plant communities. In: Ricklefs RE, Schluter D, eds. Species Diversity in Ecological Communities. Chicago: University of Chicago Press, 13–25.

Wedin DA, Tilman D. 1993. Competition among grasses along a nitrogen gradient: initial conditions and mechanisms of competition. Ecological Monographs 63: 199–229.

Wells CL, Pigliucci M. 2000. Adaptive phenotypic plasticity: the case of heterophylly in aquatic plants. Perspectives in Plant Ecology, Evolution and Systematics 3: 1–18.

Winn AA. 1999. The functional significance and fitness consequences of heterophylly. International Journal of Plant Sciences 160: S113–S121.

Withers JR. 1979. Studies on the status of unburnt Eucalyptus woodland at Ocean Grove, Victoria. IV. The effect of shading on seedling establishment Australian Journal of Botany 27: 47–66.

Wullschleger SD, Yin TM, Fazio SPD, Tschaplinski TJ, Gunter LE, Davis MF, Tuskan GA. 2005. Phenotypic variation in growth and biomass distribution for two advanced-generation pedigrees of hybrid poplar. Canadian Journal of Forestry Research 35: 1779–1789.

Zhang X, Wu N, Li C. 2005. Physiological and growth responses of Populus davidiana ecotypes to different soil water contents. Journal of Arid Environments 60: 567–579.

46 Table 1.1 Summary of the traits, abbreviations and statistical transformations that were used in the Chapter 1.

Trait Abbrev. Definition Trans. a) Whole-Plant Traits

Root Mass Ratio RMR Ratio of total root mass to log (g g-1) sum of stem and leaf mass Stem Mass Ratio StMR Ratio of total stem mass to log (g g-1) sum of root and leaf mass Leaf Mass Ratio LMR Ratio of total leaf mass to log (g g-1) sum of stem and root mass Leaf Area Ratio LAR Ratio of total leaf area to log (cm2 g-1) total biomass Height to Diameter HtoD Ratio of stem length to log+1 (mm mm-1) basal stem diameter Internode Length Internode Stem length divided by the log+1 (mm) total number of nodes Specific Leaf Area SLA Ratio of leaf area to leaf log (cm2 g-1) mass b) Heteroblastic Traits

Compound Leaf Mass CompLMR Ratio of total compound sqrt Ratio leaf mass to sum of root, (g g-1) stem and other leaf mass Transitional Leaf Mass TransLMR Ratio of total transitional log Ratio leaf mass to sum of root, (g g-1) stem and other leaf mass Compound Leaf Area CompLAR Ratio of total compound sqrt Ratio leaf area to total biomass (cm2 g-1) Transitional Leaf Area TransLAR Ratio of total transitional log+1 Ratio leaf area to total biomass (cm2 g-1) Compound Specific Leaf CompSLA Ratio of compound leaf ln Area area to compound leaf mass (cm2 g-1) Transitional Specific TransSLA Ratio of transitional leaf log+1 Leaf Area area to transitional leaf (cm2 g-1) mass Number of Compound CompNodes Number of nodes which log Nodes developed a compound leaf before the first transitional leaf.

47 Table 1.2 Whole-plant multivariate and univariate analysis. The univariate ANOVAs have been ranked from highest to lowest F-value as an indication of the influence of traits over the MANOVA model. MANOVA degrees of freedom are 7, 147. ANOVA degrees of freedom are 1, 153. Full ANOVA tables are presented in [Supplementary materials 1.1]. a) MANOVA

Wilks’ λ F-value P-value Nutrient 0.326 43.489 <0.0001 Light 0.379 34.362 <0.0001 Water 0.906 2.170 0.040 N x L 0.957 0.936 0.480 N x W 0.942 1.304 0.252 L x W 0.908 2.125 0.044 N x L x W 0.881 2.829 0.009

b) ANOVA

Rank Nutrient F-value P-value Light F-value P-value Water F-value P-value 1 RMR 191.396 <0.0001 Internode 159.305 <0.0001 RMR 12.543 0.001 2 LMR 71.291 <0.0001 HtoD 153.654 <0.0001 HtoD 7.399 0.007 3 HtoD 54.197 <0.0001 StMR 39.505 <0.0001 Internode 5.159 0.025 4 LAR 21.035 <0.0001 SLA 39.035 <0.0001 LMR 4.704 0.032 5 SLA 11.076 0.001 RMR 10.035 0.002 LAR 2.958 0.087 6 StMR 2.433 0.121 LAR 9.975 0.002 StMR 0.467 0.496 7 Internode 0.498 0.481 LMR 1.881 0.172 SLA 0.079 0.779

48 Table 1.3 Heteroblastic multivariate and univariate trait analysis. The univariate ANOVAs have been ranked from highest to lowest F- value as an indication of the influence of traits over the MANOVA model. MANOVA degrees of freedom are 10, 64. ANOVA degrees of freedom are 1, 73. Full ANOVA tables are presented in [Supplementary materials 1.2]. a) MANOVA

Wilks’ λ F-value P-value Nutrient 0.292 15.500 <0.0001 Light 0.421 8.808 <0.0001 Water 0.810 1.503 0.159 N x L 0.707 2.659 0.009 N x W 0.898 0.724 0.699 L x W 0.936 0.437 0.923 N x L x W 0.873 0.928 0.514

b) ANOVA

Rank Nutrient F-value P-value Light F-value P-value Water F-value P-value 1 TransLMR 64.611 <0.0001 Internode 64.435 <0.0001 RMR 7.573 0.007 2 RMR 59.678 <0.0001 HtoD 41.758 <0.0001 Internode 4.371 0.040 3 CompLMR 44.805 <0.0001 TransSLA 24.545 <0.0001 HtoD 3.651 0.060 4 TransLAR 38.174 <0.0001 StMR 11.607 0.001 CompSLA 1.572 0.214 5 CompLAR 34.531 <0.0001 CompLAR 5.014 0.028 CompLAR 0.903 0.345 6 HtoD 8.731 0.004 TransLAR 3.238 0.076 TransLAR 0.126 0.724 7 TransSLA 5.369 0.023 RMR 3.026 0.086 TransLMR 0.098 0.755 8 CompSLA 4.218 0.044 CompLMR 1.671 0.200 CompLMR 0.075 0.785 9 StMR 1.348 0.249 CompSLA 1.216 0.274 StMR 0.062 0.804 10 Internode 0.246 0.622 TransLMR 4.08x10-7 0.999 TransSLA 0.041 0.839

Table 1.4 ANCOVA table following analysis of CompNodes. Within treatment trait means are presented in Figure 1.4. Total biomass at point of harvest was chosen as the covariate in order to control for possible differences in plant size on CompNodes.

Treatment df MSS F-value P-value

Total Biomass 1 0.010 1.450 0.230 Nutrient 1 0.010 1.392 0.240 Light 1 0.092 12.987 <0.0001 Water 1 0.002 0.224 0.636 Nutrient * Light 1 0.003 0.441 0.508 Nutrient * Water 1 0.019 2.608 0.108 Light * Water 1 0.017 2.436 0.121 Nutrient * Light * Water 1 0.011 1.590 0.209 Error 152 0.007 Total 161

-0.2 **** ** ***

-0.4

log RMR log -0.6

-0.6 ****

-0.7

-0.8 log StMR

-0.9 -0.2 **** * -0.3

-0.4

log LMR log -0.5

-0.6 1.5 **** ** 1.4

1.3

log LAR 1.2

1.1 2.0 **** **** **

1.8

1.6 log+1 HtoD

1.4 **** *

1.2

log+1 Internode 1.0 1.8 *** **** 1.7

1.6 log SLA

1.5 High Low High Low High Low Nutrient Light Water Figure 1.1 Estimated marginal means and standard errors of whole-plant traits. Traits and their units of measurement are outlined in Table 1.1. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.

0.6 ****

0.4

sqrt CompLMR 0.2 0.0 **** -0.3

-0.6

-0.9 log TransLMR log 4 **** *

sqrt CompLAR 1 1.5 ****

1.0

0.5 log+1 TransLAR log+1 4.5 *

4.2 ln CompSLA ln 3.9 1.9 *****

1.7

log+1 TransSLA log+1 1.5 High Low High Low High Low

Nutrient Light Water Figure 1.2 Estimated marginal means and standard errors of heteroblastic traits. Traits and their units of measurement are outlined in Table 1.1. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001

a) b) 4 4

3 3 t t n L ie L n o r o e w t w 2 i Li 2 Lig r 8 gh u 34 h t 7 t N t u h 1 N 1 g 7 3 i w 4 t H o n t 0 L e 0 8 i n 5 r 12 e 6 t i u 5 tr -1 -1 u N DF2 (38.6%) DF2 1 (25.5%) DF2 N H h H 6 i g i -2 gh 2 i -2 g w L h L o ig H i ht ght L -3 -3

-4 -4 -4 -3 -2 -1 0 1 2 3 4 -4-3-2-101234 DF1 (54.6%) DF1 (59.4%) c) d) 1.0 1.0 HtoD Internode Internode t n t SLA L ie L n ow HtoD r ow e L 0.5 t StMR L 0.5 i ig u TransLAR ig r h TransLMR h TransSLA t StMR t N t u h RMR N CompSLA ig w LAR t H o n t 0.0 L e 0.0 CompLAR i n r ie t r u t CompLMR N u

DF2 (25.5%) DF2 (38.6%) DF2 LMR N H h H i g i -0.5 gh i -0.5 gh w Li Li o RMR ght H ght L

-1.0 -1.0 -1.0 -0.5 0.0 0.5 1.0 -1.0 -0.5 0.0 0.5 1.0 DF1 (54.6%) DF1 (59.4%)

Figure 1.3 Grouping of treatments and traits following discriminant function analysis. Variance explained by first and second discriminant functions displayed parenthetically. a,b: Treatments are mean populations with circles indicating 95% confidence intervals for (a) whole plant and (b) heteroblastic traits. Numbers correspond to treatment group (1, high nutrient, high light, high water; 2, high nutrient, high light, low water; 3, high nutrient, low light, high water; 4, high nutrient, low light, low water; 5 low nutrient, high light, high water; 6, low nutrient, high light, low water; 7, low nutrient, low light, high water; 8, low nutrient, low light, low water). The nutrient and light treatments are highlighted due to significant outcomes following MANOVA. c,d: Factor loadings of (c) whole plant and (d) heteroblastic traits across the first and second discriminant functions.

10 **** 9

Nodes with Compound Leaf Nodes with Compound 4 HighLow High Low High Low Nutrient Light Water

Figure 1.4 Number of nodes displaying a compound leaf (CompNodes) with standard error bars. Analysis was performed on log data yet untransformed data is presented. Only the light treatment showed a significant difference (**** p < 0.0001). Full ANCOVA table is presented in Table 1.4.

1.4 **** ****

1.2

1.0

0.8 Total Biomass (g) 0.6

0.4 HighLow High Low High Low Nutrient Light Water

Figure 1.5 Estimated marginal mean and standard errors of total biomass (g) at point of harvest. Analysis was performed on log data yet untransformed data is presented. The nutrient and water treatments showed a significant difference (**** p < 0.0001). Full ANOVA table is presented in [Supplementary materials 1.4].

Chapter 2

Heteroblastic development and shade-avoidance in response to Blue and Red light signals in Acacia implexa

Authors: Michael A. Forster and Stephen P. Bonser

ABSTRACT

Information from Blue (400-500nm; B) and Red (660-730nm; R) wavelengths is used by plants to determine proximity of neighbours or actual shading. Plants undergo trait changes in order to out-compete neighbours or accommodate shading. Heteroblasty, the dramatic shift from one leaf type to another during juvenility, can be influenced by the light environment although it is unknown whether cues from B or R (or both) are driving the developmental process.

Seedlings of three populations of Acacia implexa (Mimosaceae) sourced from low, medium and high rainfall habitats were grown in a factorial design of high/low B and R light to determine how light signals affect heteroblasty and patterns of biomass allocation. Low B light significantly delayed heteroblasty in the low rainfall population and low R light significantly delayed in the low and high rainfall populations. Low B light increased stem elongation and decreased root biomass whereas low R light had a more wide-ranging impact on plant phenotypes.

These results were consistent across populations although the low rainfall population showed greater trait variability in response to R light signals. We conclude that R light conveys a greater information signal that affects heteroblasty and seedling development in A. implexa.

INTRODUCTION

Plant’s perception of the light environment is fundamental to growth and reproduction. Light not only determines the total amount of photosynthesis but acts as an information source in plant-plant signalling (Ballaré, 1999; Franklin, 2008).

Light signalling occurs across two major components of the spectral wavelength: blue (400-500nm; B) and red (660-730nm; R) light bandwidths. Leaves preferentially absorb B and R light (Holmes and Smith, 1977) and the decrease of these wavelengths relative to other wavelengths acts as an indicator of neighbour plants or actual shading (Ballaré and Scopel, 1997; Ballaré, 1999; Franklin and

Whitelam, 2007; Franklin, 2008). Low levels of B and R light induce a series of trait changes that accommodates the light environment and directs development towards higher quality light.

Photomorphogenesis defines changes in plant traits caused by the light environment (Weller and Kendrick, 2007). Under shaded environments typical changes in traits include a decrease in root allocation, an increase in stem allocation, elongation of the stem and internodes, an increase in leaf area but a decrease in leaf mass (e.g. Smith, 1982; Tilman, 1988; Kitajima, 1994; Poorter and

Nagel, 2000; Sleeman et al., 2002). These trait changes direct the growing individual above or away from the source of shade and into a full sunlight environment. Not all plants show such trait changes and many species can tolerate extremely low levels of light (Franklin, 2008). Photomorphogenesis is more common in shade-avoiding species, a term for those species needing relatively high light levels in order to maintain optimal plant functioning.

The overwhelming number of studies on photomorphogenesis and shade- avoidance has occurred by reducing red light (660-670nm; R) relative to far red light (720-730nm; FR) while less attention has been directed towards B light

(Galen et al., 2004). R:FR light is detected by phytochrome molecules and is known to have a significant influence on plant development and shade-avoidance

(Holmes and Smith, 1975; Whitelam and Smith, 1981; Casal and Smith, 1989;

Ballaré, 1999). B light is detected by cryptochrome and phototropin molecules and is largely thought to influence redox reactions in the photosynthetic system and stem orientation during seedling development (Briggs and Huala, 1999; Casal,

2000). Recently, there has been an increasing appreciation that phototropins act as a more general photoreceptor that can induce responses such as reorientation of root growth (Galen et al., 2004; Galen et al., 2007) and chloroplast and stomatal development in the leaf (Kinoshita et al., 2001; Sakai et al., 2001). Therefore B light may have a more general role in photomorphogenesis and the shade- avoidance syndrome.

Heteroblastic development is a trait in certain plant species that can be influenced by the light environment. Heteroblasty is a life-history event where plants switch from highly distinct juvenile to adult foliage (Goebel, 1898; cited in Wells and

Pigliucci, 2000). It occurs across a number of taxonomic groups and is a particularly distinct trait in the genus Acacia subgenus Phyllodineae (New, 1984).

The progression from juvenile to adult leaves in Acacia typically commences with a double compound leaf, followed by a transitional leaf consisting of a flattened petiole/rachis (called a phyllode) with double compound leaf attached, to the adult leaf which lacks a compound leaf component and solely consists of the phyllode

(Leroy and Heuret, 2008; Forster and Bonser, 2009). Compound leaves are known to be retained for longer in plants growing under closed canopies (Withers, 1979) and heteroblasty was experimentally delayed under low irradiance in the species A. implexa (Forster and Bonser, 2009). Compound leaves have larger photosynthetic capture area than phyllodes and have greater rates of photosynthesis under lower irradiance (Brodribb and Hill, 1993). Therefore the delay in heteroblastic development and retention of compound leaves conveys an amount of growth advantage to a developing Acacia seedling in an adverse light environment.

Despite the interest in the light effects on Acacia morphology and physiology little is known about the specific role that B or R light may play as an information signal.

Populations of Acacia species sourced from regions of different mean annual rainfall develop different phenotypes when grown in a common environment

(Farrell and Ashton, 1978; Cody, 1989; 1991; Daehler et al., 1999). It is thought that these different phenotypes are adapted to the moisture regimes experienced by the source populations however we believe the light environment is a key selective force driving the evolution of growth forms in these trees (Chapter 4). Vegetation cover and cloud cover decline with decreasing rainfall thus increasing irradiance levels. Moreover, Acacia species tend to dominate the canopy in low rainfall regions and the sub-canopy or understorey in higher rainfall regions (Specht and

Specht, 1999). Therefore populations within a species subjected to a strong rainfall gradient should expect to have adaptations to the concurrent light gradient.

Whether there is a genetic capacity to respond to contrasting B and R light signals

(or both) in Acacia has yet to be explored.

Our first objective in this study was to perform a fully factorial designed experiment of high and low B and R light to determine the timing of heteroblastic development across populations of the species A. implexa. We predict that heteroblastic development will be delayed under low B and/or R light and this pattern will be consistent across populations. The second objective was to explore patterns of photomorphogenesis under varying light environments. Strong photomorphogenic responses in the shade-avoiding A. implexa are expected to be observed under low R light and the same developmental pattern may also occur under low B light. Our final objective was to assess whether photomorphogenesis in response to B and R light environments altered relative growth rates in A. implexa.

METHODS

Acacia implexa is a tree growing to approximately 12m that inhabits woodlands and forests of eastern Australia (Kodela, 2002; Maslin and McDonald, 2004).

Seeds were sourced from three populations representing a rainfall gradient from

830mm p.a. to 450mm p.a. (Table 2.1a). We chose these populations because a previous study found heteroblastic development to be hastened along a rainfall gradient in the closely related A. melanoxylon (Farrell and Ashton, 1978). Seeds were sourced from the CSIRO Australian Tree Seed Centre, Canberra, Australia

(www.csiro.au/content/pt6.html).

We applied a full-factorial design with three factors: 3 x population (sourced from high, medium and low rainfall regions), 2 x B light (high and low) and 2 x R light

(high and low). B and R light were manipulated using filter plastic placed around growing plants. Table 2.1b outlines the percentage transmission of B, R and FR light allowed by the various filter plastics. All filter plastics were sourced from Lee

Filters, Andover, UK. Plants were grown at the University of New South Wales,

School of Biological, Earth and Environmental Sciences glasshouse in Sydney

Australia. Humidity in the glasshouse was kept constant and temperatures ranged between 19°C and 26°C. Plants were grown in a blocked design and were kept well spaced on glasshouse benches.

Plants were grown in 115mL pots consisting of uniform soils. Soil consisted of

33% Australian Native Landscape supply of “Organic Garden Mix”, 33% washed river sand and 33% cocopeat. A four month Osmocote low phosphorus slow release fertiliser was added with the N:P:K ratio of 17:1.5:8.5. No additional fertiliser was added throughout the experiment due to this species sensitivity to excess nutrients and the fact the experiment ran within the release time of the fertiliser. All soil, pots and any other equipment used in the experiment were completely sterilised to prevent infection of roots of symbiotic rhizobium (which fix atmospheric nitrogen thus improving fertility and growth of plants). Plants were kept well watered at all times.

This species germinates following heat from fire therefore we placed seeds in boiling water for two minutes to simulate this heat. Six seeds were then sowed directly into potted soil and were later thinned to one plant per pot. Previous attempts at growing this species found high amounts and uniform germination using this technique.

At the start of the experiment initial measurements were taken on stem length and diameter. A subset of 20 plants per population was harvested at this point with these same measurements taken. This subset of plants was then oven dried at 60°C for at least seven days and biomass was measured. Regression of biomass on stem length and diameter gave initial biomass necessary for growth rate calculations.

Plants were checked after 20 days of growth for signs of transitional leaf development. Plants that had at least one fully developed transitional leaf were harvested. Thereafter, plants were checked regularly and after 75 days all plants were harvested. At the time of harvest the following traits were measured: stem length, diameter at base of plant, total number of nodes, node of first transitional leaf, number of leaves, and number of transitional leaves. At time of harvest no plants had developed phyllodes. Leaves and petioles were removed from the stem and flattened (to prevent pinnules from folding inward) for area measurements.

Roots were removed from the stem and soil was carefully removed over a plastic container in order to capture all root material. All plant material was oven dried at

60°C for at least seven days before mass measurements were taken. The area of compound and transitional leaves were taken separately on a flatbed scanner

(Epson Stylus, CX4900) and using LeafA software (G Williamson, University of

Adelaide, Australia, ‘pers. comm.’). The mass of the different leaf types was also made separately.

Plant traits that were derived from the measurements just described are summarised in Table 2.2. Relative growth rate (RGR) was determined by:

RGR = (Ln harvest biomass) - (Ln initial biomass) / Number of days grown (1)

Specific leaf area (cm2 g-1) and leaf mass ratio (g g-1) were determined from measurements made at harvest (described above). Net assimilation rate (g cm-2 day-1) was determined by reorganising the common relative growth rate equation

(e.g. Portsmuth and Niinemets, 2006):

NAR = RGR / (SLA x LMR) (2)

Statistical Analyses

To test the hypothesis that the shift from compound leaves to transitional leaves will be hastened under low quantity and/or quality light environment a Type III univariate ANOVA was performed with population, light quantity and light quality as fixed factors and the node number of the first transitional leaf as the dependent variable. This same ANOVA model was performed on growth traits (total biomass, relative growth rate, relative height growth).

Photomorphogenesis was assessed using multivariate analysis of variance on the same ANOVA design. Data were checked for normality using Shapiro-Wilk Test and Log transformed where appropriate (details of transformations are shown in

Table 2.1). Following transformations all data were standardized to a mean of zero and standard deviation of one as traits were measured on different scales and standardization overcomes bias that may arise from this (Quinn and Keough,

2002). Homogeneity of variance in MANOVA was checked with Levene’s test of equality and outliers checked with Mahalanobis distances. Additionally we chose

to use Pillai’s Trace as a test statistic as this measure is relatively robust to the assumptions of MANOVA (Quinn and Keough, 2002). Multivariate sets of covarying traits were further characterised using discriminant function analysis.

DFA is a classification procedure that can be applied on factorial multivariate datasets to assess whether groups can be uniquely separated (Quinn and Keough,

2002). In this study we were interested in whether the factorial treatment combinations of B and R light could be classified into unique groups or whether there were overlaps in groups. We were separately interested in population effects therefore an additional DFA was run on the three populations. Means with 95% confidence intervals of experimental treatments were plotted against the first and second discriminant functions using procedures outlined in Johnson et al. (2007).

Lastly, univariate ANOVAs were performed on all plant traits to individually assess mean differences and Student-Newman-Keuls (S-N-K) post-hoc analysis was performed where appropriate. Sequential Bonferroni correction was applied in order to overcome bias associated with running a large number of statistical tests

(Rice 1989). All statistical analyses were performed using SPSS 15.0.1.

RESULTS

Heteroblastic development was significantly delayed under the R light treatment and there was no difference in the B light treatment (Table 2.3, Figure 2.1a). There was also a significant population effect (Table 2.3, Figure 2.1a) and a significant population by blue light and population by R light effect (Table 2.3). Low B light and low R light significantly delayed heteroblastic development in the low rainfall

population and low R light delayed development in the high rainfall population however light had no effect on the medium rainfall population (Figures 2.1b,c).

Acacia implexa displayed patterns of shade-avoidance under both low B and R light with a more pronounced phenotypic response to R light (MANOVA, Table

2.4a). There was also a significant population and a population by R light effect however there was no significant effect between population and B light (Table

2.4a). Additionally, a significant three-way effect between population, B and R light was observed (Table 2.4a).

A shade-avoidance syndrome was most apparent in the low R light treatment By contrast, only stem elongation and lower root biomass allocation were evident in the low B light treatment (Table 2.4b, Figure 2.2). Low R light elongated the stem and increased biomass allocation to the stem and away from roots (Table 2.4b,

Figure 2.2). There was no significant difference in leaf mass allocation under low

R however leaf area was significantly increased as indicated by an increase in LAR and SLA (Table 2.4b, Figure 2.2). NAR was significantly lower in the low R light treatment (Table 2.4b, Figure 2.2). Across populations there was an elongation of the stem and an increase in biomass allocation to the stem from the low rainfall to the high rainfall population (Table 2.4b, Figure 2.2). There was also a significant increase in LAR and SLA however LMR significantly decreased across the rainfall gradient (Table 2.4b, Figure 2.2). In the population by R light treatment interaction, the low rainfall population was most sensitive to decreased R light

(Figure 2.3). This pattern was particularly evident in the traits HtoD, StMR, LAR and SLA.

Discriminant function analysis of the B and R light treatments explained 93.6% of the variance across the first two discriminant functions. The high R light treatment was positively associated with the first discriminant function and the low R light treatment was negatively associated (Figure 2.4a). The B light treatment could be reliably separated into groups within the low R light treatment however 95% confidence intervals overlapped in the high R light treatment (Figure 2.4a). A biplot of the traits revealed RMR and NAR to be associated with the high R light treatment and the stem and leaf traits to be associated with the low R light treatment (Figure 2.4b). We also performed a discriminant function analysis on the population by R light treatment due to these factors having a significant effect following MANOVA (Table 2.4a). The first two discriminant functions explained

76.3% of the variance with the high R light treatment being positively associated with the first discriminant function (Figure 2.4c). The R light treatment could reliably be separated within populations as there was no overlap in 95% confidence intervals (Figure 2.4c). However, low R light by medium rainfall population treatment could not be separated from the high R light by high rainfall population (Figure 2.4c) meaning that these two groups had similar multivariate distribution of traits. RMR and NAR were positively associated with the first discriminant function and all other traits were negatively associated (Figure 2.4d).

Therefore the low rainfall population and high R light treatment could be discriminated by larger root mass and higher assimilation rates. Conversely, the high and medium rainfall populations and low red R treatment could be discriminated by greater biomass allocation to stem traits (StMR, HtoD, Internode) and leaf area (SLA, LAR).

There was a highly significant difference in RGR in the B and R light treatments with RGR being greater under high R light and the low B light treatments (B: F1,152

= 15.361, p = 0.0001; R: F1,152 = 14.989, p = 0.0002, Figure 2.5). There was also a significant B light by R light effect on RGR with the high B light by low R light treatment having a much slower growth rate than the other treatment groups (F1,152

= 22.356, p < 0.0001, Figure 2.6). The slow growth rate in this treatment could be attributed to the medium rainfall population which showed the slowest growth rate of all populations and treatments in the high B/low R treatment (Figure 2.6). We also note that there was a significant three way population by B by R light treatment effect (F2,152 = 3.453, p = 0.0344, Figure 2.6).

DISCUSSION

Species response to contrasting light environments

Photomorphogenesis in Acacia implexa is largely driven by information obtained from the R:FR wavelength with the blue (B) wavelength having a minor effect.

Under the low R:FR treatment A. implexa displayed a classical shade-avoidance syndrome of thinner and longer stems (i.e. increased HtoD and longer internodes), greater biomass allocation to stems and less to roots, larger photosynthetic capture area (i.e. increased LAR and SLA) and lower assimilation rates (Figures 2.3,2.4).

The low B treatment only had a significant effect on stem elongation and root biomass allocation. However the most novel finding of this experiment is heteroblastic development can also be classified as a shade-avoiding syndrome with low R:FR significantly delaying the onset of transitional leaves (Figure 2.2).

The pattern of trait responses observed in this experiment strongly suggests photoreceptors in the red light region are driving juvenile development in A. implexa. Phytochromes are the principal photoreceptors for light in the R wavelength and are widely known to drive shade-avoidance phenotypes (Holmes and Smith, 1975; Whitelam and Smith, 1981; Casal and Smith, 1989; Ballaré,

1999; Franklin, 2008). It was therefore highly predictable that A. implexa, a shade- intolerant, pioneer-colonising species of high irradiance habitats, should develop all the classic symptoms of shade-avoidance (Figures 2.3,2.4). On the other hand, phototropins detect changes in B light and low B light signals to the plant to undergo stem elongation (or stem bending) towards better quality light

(phototropism; Casal, 2000; Franklin, 2008). Under high B light, on the other hand, there is increasing evidence phototropins additionally act to increase root growth (Galen et al., 2004; Galen et al., 2007). In this experiment we observed a significant increase in stem elongation (higher HtoD and Internode, Figure 2.3) under low B light and a significant increase in root biomass allocation under high

B light. Overall, though, plant traits were much more responsive to alterations of the red light wavelength.

At the species level the observed RGR under the R light treatment was expected: a low R:FR generally leads to reduced rates of growth (e.g. Huante and Rincon,

1998). On the other hand, the effect of B light on RGR has not been widely studied. Intuitively, RGR would be expected to be greater under high B light due to the better quality light (e.g. Takemiya et al., 2005). Counter-intuitively this experiment found RGR to be significantly greater under low B light. In examining

the components of RGR (i.e. NAR, LMR and SLA; Hunt and Cornelissen, 1997), we found SLA and NAR to be similar across the B light treatments yet LMR was greater under low B light, albeit marginally non-significant (p ~ 0.05). In this experiment low B light induced a greater elongation of the stem which allowed the development of more leaf material for a given amount of plant biomass which in turn increased RGR. Therefore light signals conveyed by different wavelengths can have a complex effect on growth rates of developing seedlings of A. implexa.

The control of heteroblastic development under low R:FR, and the subsequent implied involvement of phytochrome photoreceptors, offers insights into hormonal signalling and the shade-avoidance syndrome. The study of hormonal effects on

R:FR signalling and shade-avoidance has not been extensively studied and it has been suggested that either auxin, ethylene or gibberellic acid may be involved

(Vandenbussche et al., 2005; Franklin and Whitelam, 2007). The application of gibberellic acid to lateral branches of Acacia melanoxylon suppressed heteroblastic development and applying gibberellins following phyllode development caused a reversion to compound leaves (Borchet, 1965). Application of gibberellic acid to other heteroblastic species has also induced a reversion to juvenile leaves (Horrell et al., 1990). Furthermore, gibberellins are known to cause a number of traits to undergo changes associated with shade-avoidance, particularly stem elongation

(Vandenbussche et al., 2005). Therefore shade-avoidance and heteroblastic development in Acacia may have a hormonal basis in gibberellic acid however the role of other hormones, particularly auxin, cannot be discounted.

Population response to contrasting light environments

Acacia species are known to have high genetic variability between populations

(Daehler et al., 1999) and the patterns of morphological development observed for

A. implexa were no exception. Irrespective of light treatment there were highly significant differences between low, medium and high rainfall populations across a range of traits. Stem elongation, in particular, appears to be highly variable across populations with the high rainfall population having the longest stems. The pattern of trait variability is consistent with adaptation to varying light environments from where seeds were sourced. For example, canopy cover in the region of the high rainfall population is denser than canopy cover in the low rainfall region (McRae and Cooper, 1985; Benson, 1986; Scott, 1992). Additionally, the number of cloud free days is greater in the low rainfall region than the high rainfall region (BOM,

2009). Greater canopy cover and less total solar radiation in the high rainfall region suggests light quality and quantity in that region is lower. Therefore we observed greater carbon allocation to the shoot in the high rainfall population when compared to the low rainfall population.

In this experiment there were strong population by R light treatment effects yet there were no population by B light effects further suggesting that seedling development in A. implexa is being driven by changes in the R:FR ratio. This species is an early successional type that typically germinates following community wide disturbance such as fire. Selective pressure to respond to a reduction in the R:FR ratio would be strong as this signals competing plants are potentially growing an overhanging canopy. Experiments have shown that B photoreceptors (cryptochromes) do not respond to crowding whereas R

photoreceptors (specifically, phytochrome B) played the most important role in photodetection of neighbours and potential competition (Ballaré and Scopel,

1997).

The low rainfall population appeared to be more sensitive to a reduction in R light than the medium or high rainfall populations. In this experiment stem elongation

(HtoD), allocation of biomass to stem tissue (StMR) and leaf area (LAR, SLA) were particularly sensitive traits to a reduction in R light in the low rainfall population (Figure 2.3). Stem elongation is a ubiquitous response across a number of populations sourced from open habitats when these populations are subjected to shade (Dudley and Schmitt, 1995; Schmitt et al., 2003) and is implicated as the most common shade-avoidance response in plants (Dudley and Schmitt, 1996).

Evidence of greater response to shading in populations from more open habitats confirms the adaptive value of shade-avoidance (Dudley and Schmitt, 1995;

Weinig, 2000; Schmitt et al., 2003). RGR was used as a measure of plant performance in this experiment and RGR was not significantly different in the R light by low rainfall population treatment. This result suggests a possible adaptive value for shade-avoidance in the low rainfall population.

Different parts of the spectral wavelength convey information and plants have evolved numerous mechanisms to detect plant-plant signals. Acacia implexa is highly responsive to a reduction in the R:FR ratio which signals to developing seedlings to undergo photomorphogenesis and shade-avoidance. There was some shade-avoidance response to B light but the overwhelming response was to R light.

Novel to this experiment was the finding that heteroblastic development is also

affected by the R light component whereby plants retain juvenile leaves for longer under low R light.

Acknowledgements

We thank Brenton Ladd, Geoff McDonnell and Natalie Wynyard for assistance and Grant Williamson for leaf area software. Project was funded by UNSW Faculty Research Grant and Early Career Research grant to S.P.B.

REFERENCES

Ballaré CL. 1999. Keeping up with the neighbours: phytochrome sensing and other signaling mechanisms. Trends in Plant Science, 4:97-102.

Ballaré CL, Scopel AL. 1997. Phytochrome signalling in plant canopies: testing its population-level implications with photoreceptor mutants of Arabidopsis. Functional Ecology, 11:441-450.

Benson DH. 1986. The vegetation of the Gosford and Lake Macquarie 1:100 000 vegetation map sheet. Cunninghamia, 1:467-490.

Borchett R. 1965. Gibberellic acid and rejuvenation of apical meristems in Acacia melanoxylon. Die Naturwissenschaften, 3:65-66.

Briggs WR, Huala E. 1999. Blue-light photoreceptors in higher plants. Annual Review of Cell Development Biology, 15:33-62.

Brodribb T, Hill RS. 1993. A physiological comparison of leaves and phyllodes in Acacia melanoxylon. Australian Journal of Botany, 41:293-305.

Casal JJ. 2000. Phytochromes, cryptochromes, phototropin: Photoreceptor interactions in plants. Photochemistry and Photobiology, 71:1-11.

Casal JJ, Smith H. 1989. The function, action and adaptive significance of phytochrome in light-grown plants. Plant Cell and Environment, 12:855- 862.

Cody ML. 1989. Morphological variation in mulga. I. Variation and covariation within and among Acacia aneura populations. Israeli Journal of Botany, 38:241-257.

Cody ML. 1991. Morphological variation in mulga. II. Covariation among morphology, phenology, and spatial patterns in Acacia -dominated vegetation. Israeli Journal of Botany, 40:41-59.

Daehler CC, Yorkston M, Sun WG, Dudley N. 1999. Genetic variation in morphology and growth characters of Acacia koa in the Hawaiian Islands. International Journal of Plant Sciences, 160:567-573.

Dudley SA, Schmitt J. 1995. Genetic differentiation in morphological responses to simulated foliage shade between populations of Impatiens capensis from open and woodland sites. Functional Ecology, 9:655-666.

Dudley SA, Schmitt J. 1996. Testing the adaptive plasticity hypothesis: Density- dependent selection on manipulated stem length in Impatiens capensis. The American Naturalist, 147:445-465.

Farrell TP, Ashton DH. 1978. Population studies of Acacia melanoxylon. I. Variation in seed and vegetation characteristics. Australian Journal of Botany, 26:365-379.

Forster MA, Bonser SP. 2009. Heteroblastic development and the optimal partitioning of traits among contrasting environments in Acacia implexa. Annals of Botany, 103:95-105.

Franklin KA. 2008. Shade avoidance. New Phytologist, 179:930-944.

Franklin KA, Whitelam GC. 2007. Red:far-red ratio perception and shade avoidance. In Light and Plant Development, pp. 211-234.

Galen C, Huddle J, Liscum E. 2004. An experimental test of the adaptive evolution of phototropins: Blue-light photoreceptors controlling phototropism in Arabidopsis thaliana. Evolution, 58:515-523.

Galen C, Rabenold JJ, Liscum E. 2007. Functional ecology of a blue light photoreceptor: effects of phototropin-1 on root growth enhance drought tolerance in Arabidopsis thaliana. New Phytologist, 173:91-99.

Goebel K. 1898. Organographie der Pflanzen, Jena: Fischer.

Holmes MG, Smith H. 1975. The function of phytochrome in plants growing in the natural environment. Nature, 254:512-514.

Holmes MG, Smith H. 1977. The function of phytochrome in the natural environment. II. The influence of vegetation canopies on the spectral energy distribution of natural daylight. Photochemistry and Photobiology, 25:539-545.

Horrell BA, Jameson PE, Bannister P. 1990. Plant growth regulation and phase change in some New Zealand heteroblastic species. New Zealand Journal of Botany, 28:187-193.

Huante P, Rincon E. 1998. Responses to light changes in tropical deciduous woody seedlings with contrasting growth rates. Oecologia, 113:53-66.

Hunt R, Cornelissen JHC. 1997. Components of relative growth rate and their interrelationships in 59 temperate plant species. New Phytologist, 135:395- 417.

Johnson HE, Llyod AJ, Mur LAJ, Smith AR, Causton DR. 2007. The application of MANOVA to analyse Arabidopsis thaliana metabolomic data from factorially designed experiments. Metabolomics, 3:517-530.

Kinoshita T, Doi M, Suetsugu N, Kagawa T, Wada M, Shimazaki K. 2001. phot1 and phot2 mediate blue light regulation of stomatal opening. Nature, 414.

Kitajima K. 1994. Relative importance of photosynthetic traits and allocation patterns as correlates of seedling shade tolerance of 13 tropical trees. Oecologia, 98:419-428.

Kodela PG. 2002. Acacia. In Flora of New South Wales, pp. 381-476.

Leroy C, Heuret P. 2008. Modelling changes in leaf shape prior to phyllode acquisition in Acacia mangium Willd. seedlings. C. R. Biologies, 331:127- 136. Maslin BR, McDonald MW. 2004. AcaciaSearch: Evaluation of Acacia as a woody crop option for southern Australia, Canberra: Rural Industries Research and Development Corportation.

McRae RHD, Cooper MG. 1985. Vegetation of the Merriwa area, New South Wales. Cunninghamia, 1:351-370.

New TR. 1984. A Biology of Acacias, Melbourne: Oxford University Press.

Portsmuth A, Niinemets Ü. 2006. Structural and physiological plasticity in response to light and nutrients in five temperate deciduous woody species of contrasting shade tolerance. Functional Ecology, 21:61-77.

Quinn GP, Keough MJ. 2002. Experimental Design and Data Analysis for Biologists, Port Melbourne: Cambridge University Press.

Rice WR. 1989. Analyzing tables of statistical tests. Evolution, 43:223-225.

Sakai T, Kagawa T, Kasahara M, Swartz TE, Christie JM, Briggs WR, Wada M, Okada K. 2001. Arabidopsis nph1 and npl1: Blue light receptors that mediate both phototropism and chloroplast relocation. Proceedings of the National Academy of Sciences USA, 98:6969-6974.

Schmitt J, Stinchcombe JR, Heschel MS, Huber H. 2003. The adaptive evolution of plasticity: Phytochrome-mediated shade-avoidance responses. Integrative and Comparative Biology, 43:459-469.

Scott JA. 1992. The natural vegetation of Balranald - Swan Hill area. Cunninghamia, 2:597-652.

Sleeman JD, Dudley SA, Pannell JR, Barrett SCH. 2002. Responses of carbon acquisition traits to irradiance and light quality in Mercurialis annua (Euphorbiaceae): Evidence for weak integration of plastic responses. American Journal of Botany, 89:1388-1400.

Smith H. 1982. Light quality, photoperception and plant strategy. Annual Review of Plant Physiology, 33:481-518.

Specht RL, Specht A. 1999. Australian Plant Communities: dynamics of structure, growth and biodiversity, South Melbourne: Oxford University Press.

Takemiya A, Inoue S, Doi M, Kinoshita T, Shimazaki K. 2005. Phototropins promote plant growth in response to blue light in low light environments. The Plant Cell, 17:1120-1127.

Tilman D. 1988. Plant Strategies and the Dynamics and Structure of Plant Communities, Princeton: Princeton University Press.

Vandenbussche F, Pierik R, Millenaar FF, Voesenek LACJ, Straeten DVD. 2005. Reaching out of the shade. Current Opinion in Plant Biology, 8:462- 468.

Weinig C. 2000. Plasticity versus canalization: Population differences in the timing of shade-avoidance responses. Evolution, 54:441-451.

Weller JL, Kendrick RE. 2007. Photomorphogenesis and photoperiodism in plants. In Photobiology: The Science of Life and Light, pp. 417-463.

Wells CL, Pigliucci M. 2000. Adaptive phenotypic plasticity: the case of heterophylly in aquatic plants. Perspectives in Plant Ecology, Evolution and Systematics, 3:1-18.

Whitelam GC, Smith H. 1981. Retention of phytochrome-mediated shade avoidance responses in phytochrome deficient mutants of Arabidopsis, cucumber and tomato. Journal of Plant Physiology, 139:119-125.

Withers JR. 1979. Studies on the status of unburnt Eucalyptus woodland at Ocean Grove, Victoria. IV. The effect of shading on seedling establishment Australian Journal of Botany, 27:47-66.

Table 2.1 (A) Brief description of populations of sourced seeds. Climate data were collated from Bureau of Meteorology, Australia climate data of nearest weather station. (B) Specifications of the light filters used to obtain Blue and Red light treatments. Values are percent transmission of light through filters. Filters were sourced from Lee Filters (Andover, UK).

(A) Rainfall Avg Temp. Population Seedlot Location (mm p.a.) (°C)

Low Rainfall 19780 35°57′S 450 23 145°45′E Medium Rainfall 19770 32°37′S 650 23 150°03′E High Rainfall 18859 31°56′S 830 24 151°11′E

(B) Filter (product code) 400-500nm 665nm 730nm B Light R Light

0.3 ND (209) 50 45 85 High High Lime Green (088) 0 45 85 Low High 0.6 ND (210) 25 20 75 High Low 0.3 ND + Lime (209 + 088) 0 20 75 Low Low

Table 2.2 Summary of traits measured in Chapter 2 with abbreviations and transformations made for statistical analysis. Trait Abbrev. Definition Trans. a) Whole-Plant Traits

Height to Diameter HtoD Ratio of stem length to basal (mm cm-2) stem diameter Internode Length Internode Stem length divided by the (mm) total number of nodes Root Mass Ratio RMR Ratio of total root mass to sum log (g g-1) of stem and leaf mass Stem Mass Ratio StMR Ratio of total stem mass to sum log (g g-1) of leaf and root mass Leaf Mass Ratio LMR Ratio of total leaf mass to sum log (g g-1) of stem and root mass Leaf Area Ratio LAR Ratio of total leaf area to total (cm2 g-1) biomass Net Assimilation Rate NAR The amount of carbon fixed log (g cm-2 day-1) per day for a given amount of photosynthetic tissue Specific Leaf Area SLA Ratio of leaf area to leaf mass (cm2 g-1) Relative Growth Rate RGR Amount of biomass fixed per (g g-1 day-1) day; summarised by Eq. 3

Table 2.3 ANOVA table following analysis of node of first transitional leaf. Within treatment trait means are summarised in Figure 2.1. d.f. MSS F-value P-value

Population 2 14.551 8.269 0.0004 B Light 1 1.629 0.926 0.3376 R Light 1 38.848 22.077 < 0.0001 Population x B Light 2 7.561 4.297 0.0154 Population x R Light 2 11.749 6.677 0.0017 B x R 1 4.063 2.309 0.1309 Population x B x R 2 1.854 1.054 0.3514 Error 140 1.760 Total 152

Table 2.4 Multivariate and univariate analysis of variance. F-values are displayed for the univariate ANOVAs and significant values are in bold. Degrees of freedom are given parenthetically. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001. a) MANOVA

Pillai’s F-value P-value Trace Population 0.690 8.299 < 0.0001 Blue Light 0.204 3.999 0.0003 Red Light 0.380 9.593 < 0.0001 Pop x Blue Light 0.092 0.762 0.7281 Pop x Red Light 0.226 2.007 0.0133 Blue x Red 0.246 5.106 < 0.0001 Pop x Blue x Red 0.219 1.934 0.0181

b) ANOVA

Trait Population Blue Light Red Light Pop x Blue Pop x Red Blue x Red Pop x Blue x Red (2) (1) (1) (2) (2) (1) (2) HtoD 33.030**** 15.276*** 13.594*** 0.913 4.703** 4.123* 3.465* Internode 26.437**** 4.026* 11.195** 0.516 2.913 1.583 1.434 RMR 1.906 11.965*** 21.793**** 1.476 3.730* 2.213 0.776 StMR 24.847**** 1.643 12.971*** 1.123 3.662* 0.683 1.069 LAR 3.464* 0.460 35.342**** 1.048 3.603* 0.016 0.453 LMR 9.980**** 3.874 2.403 0.058 2.546 0.001 0.403 SLA 11.651**** 0.107 33.576**** 1.374 4.646* 0.130 0.289 NAR 1.415 0.861 51.802**** 1.311 0.277 11.547*** 3.318*

(A) 10 *** ****

5 Node of of FirstNode Transitional Leaf 4 L MH HLH L Population Blue Red (B) 10 *

5 Node of First Transitional Leaf First Transitional of Node 4 HL HLH L Low Medium High (C) 10 **** **

5 Node of First Transitional Leaf First Transitional of Node 4 HL HLH L Low Medium High

Figure 2.1 Node of first transitional leaf represents the onset of heteroblasty. Displayed are means (±S.E.). H: High; M: Medium; L: Low. *** p < 0.001; **** p < 0.0001. Complete statistical results are presented in Table 2.3. (A) Population, Blue and Red light treatments; (B) interaction of the population by Blue light treatment; and (C) the interaction of the population by Red light treatment.

1 **** *** ***

0 HtoD

-1 1 **** * **

0 Internode

-1 1 *** ****

0 RMR

-1 1 **** ***

0 StMR

-1 L M H H L H L Population Blue Red

Figure 2.2 Means (±S.E.) of measured traits across experimental treatments. H: High; M: Medium; L: Low. * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001. All traits have had units of measurement standardised to a mean of zero and standard deviation of one.

1 *****

0 LAR

-1 1 ****

0 LMR

-1 1 **** ****

0 SLA

-1 1 ****

0 NAR

-1 L M H H L H L Population Blue Red

Figure 2.2 (continued) Means (±S.E.) of measured traits across experimental treatments. H: High; M: Medium; L: Low. * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001. All traits have had units of measurement standardised to a mean of zero and standard deviation of one.

** 1 1

0 0 HtoD

-1 Internode -1

* * 1 1

0 0 RMR StMR

-1 -1

* 1 1

0 0 LAR LMR

-1 -1

* 1 1

0 0 SLA NAR

-1 -1

High Low High Low

Figure 2.3 Means (±S.E.) of measured traits within the population by Red light interaction treatment. * p < 0.05; ** p < 0.01. High and Low correspond to Red light treatment. ● High rainfall populations; ○ medium rainfall population; and ▼ low rainfall population. All traits have had units of measurement standardised to a mean of zero and standard deviation of one.

(A) (B) 3 3

2 2

1 1 MH ML LH HL HH 0 0 HH LH HL

DF2 (31.1%) DF2 (18.3%) DF2 LL -1 LL -1

-2 -2

-3 -3 -3 -2 -1 0 1 2 3 -3 -2 -1 0 1 2 3 DF1 (62.5%) DF1 (58.0%)

RMR

0.5 0.5 RMR

NAR NAR

0 0 HtoD

StMR LMR DF2 (18.3%) DF2 DF2 (31.1%) DF2 LMR SLA SLA InternodeLAR -0.5 StMR -0.5 LAR Internode

HtoD

-1 -1 -1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1 DF1 (62.5%) DF1 (58.0%)

Figure 2.4 Grouping of population, Blue and Red light treatments following discriminant function analysis. Displayed are group centroids with circles indicating 95% confidence intervals (overlapping circles indicate no difference between groups). Variance explained by first and second discriminant functions displayed parenthetically. (A) HH, HL, LH, LL: high Red, high Blue; high Red, low Blue; low Red, high Blue; low Red, low Blue treatments. (B) HH, MH, LH: high rainfall, medium rainfall, low rainfall population by high Red light treatments; HL, ML, LL: high rainfall, medium rainfall, low rainfall population by low Red light treatments. (C and D) Factor loadings of measured traits corresponding to DFA of treatments across first and second discriminant functions.

* *** *** 0.16 ) -1

day 0.12 -1

0.08 RGR (g g (g RGR

0.04 L MH H LH L Population Blue Red

Figure 2.5 Mean (±S.E.) of relative growth rate (RGR) across experimental treatments. H: High; M: Medium; L: Low. * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001.

0.20

0.16 a a ) ab abc ab -1 abc abc abc abc 0.12

day bc c -1

0.08 d RGR (g g (g RGR 0.04

0.00 Blue Light: H L H L H L H L H L H L Red Light: High Low High Low High Low Population: Low Medium High

Figure 2.6 Mean (±S.E.) of relative growth rate (RGR) across the full factorial combination of population by Blue by Red light treatments. H: High; L: Low. * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001. Letters indicate groupings following S-N-K post-hoc analysis.

Chapter 3

Ultraviolet radiation alters whole-plant and leaf trait strategies in a heteroblastic species, Acacia implexa

Authors: Michael A. Forster and Stephen P. Bonser

ABSTRACT

Ultraviolet radiation is inevitably absorbed by leaves yet plants rarely show damage to ambient levels. Nevertheless different types of leaves may better cope with varying levels of UV. Heteroblastic species display functionally different leaf types and may show difference in their capacity to cope with variations in UV.

Three populations of Acacia implexa were grown in a glasshouse under no UV and enhanced UV to test the hypothesis that heteroblastic development to sun-adapted leaves will be hastened under enhanced UV. Further, we tested the hypothesis that populations sourced from a high UV environment will better tolerate enhanced UV as expressed by whole-plant and leaf-level traits. We found no significant difference in the timing of heteroblastic development. Relative growth rate (RGR) was not significantly different between UV treatments due to plants in enhanced

UV treatment having larger specific leaf area (SLA) and plants in the reduced UV treatment having a higher net assimilation rate (NAR). Leaves in enhanced UV were similar in area yet significantly smaller in mass due to less structural support and therefore had a higher SLA. These results were not consistent across populations where whole-plant and leaf-level trait responses were highly complex.

Generally, two populations from higher UV environment showed no difference in

RGR yet the third population from the lowest UV light environment increased

RGR under enhanced UV. We conclude that differential levels of UV can have complex effects on whole-plant and leaf trait strategies which in turn can affect

RGR and plant performance.

INTRODUCTION

Ultraviolet (UV) radiation is an important component of the light spectrum as it falls outside the region of photosynthetically active radiation (PAR) but is still absorbed by leaves (Caldwell et al., 2003; Caldwell et al., 2007). UV has been divided into three zones across the light spectrum. UV-A falls closest to PAR at

315-400nm and is generally considered benign to biological organisms. UV-B

(280-315nm) and UV-C (200-280nm) can be extremely dangerous to biological organisms however the earth’s atmosphere absorbs all UV-C and the majority of

UV-B (McKenzie et al., 2007). Depletion of ozone through atmospheric pollutants increases the risk of greater UV-B reaching the earth’s surface however significant ozone depletion is needed for an increase in UV-C (McKenzie et al., 2007). The degree of response to enhanced levels of UV across all wavelengths is not completely understood yet is important in an overall understanding of how plants function under variable light environments.

The capacity to display different leaf forms will be associated with an increased ability to function across environments. Heterophylly is defined as the capacity to produce at least two different types of leaves grown on a single individual and can be drastically different in form and function (Winn, 1999). Heteroblastic development is often used in conjunction with heterophylly with the subtle difference being a single developmental shift between different leaf types versus reversible shifts in leaf types (Jones, 1999). Heterophylly has been shown to be adaptive under varying light environments where different leaf types are developed given certain light conditions (e.g. Bodkin et al., 1980; Cronin and Lodge, 2003).

Heteroblastic development has also been shown to be delayed or hastened

depending on the light environment (Burns, 2005; Forster and Bonser, 2009).

However it is not known whether heterophylly or heteroblastic development is associated with plant function under variable UV environments.

Heteroblastic development is known to occur in the cosmopolitan genus, Acacia

(Mimosaceae), most commonly in juveniles. Within Acacia heteroblastic development is constrained to the subgenus Phyllodineae (New, 1984) and involves shifting from true leaves, to transitional leaves and then modified leaves as adults. True leaves, or the juvenile leaves, are double compound whereas the modified leaves, or the adult leaves, are flattened petioles/rachis known as phyllodes (New, 1984). Transitional leaves consist of a phyllode with at least two pinnae attached distally (Figure 3.1). In a test of the adaptive value of compound leaves and phyllodes to a high light environment, Brodribb and Hill (1993) found phyllodes to perform far more optimally in greater intensity light. In Acacia the development of juvenile or adult leaves may have important adaptations to the UV environment. For example, Forster and Bonser (2009) experimentally demonstrated that juvenile leaves were retained for longer in a low light environment and transitional leaves were developed earlier in a high light environment. The optimisation of leaf types to a high and low light environment consequently led to similar relative growth rates in the different light environments. It is not known whether different UV light environments can also induce changes in the developmental timing of leaf types in Acacia and whether or not such changes can lead to differences in relative growth rates.

The timing of heteroblastic development in Acacia has been shown to vary according to habitat and across populations. For example, the switch from compound leaves to phyllodes was observed to be delayed under a forest canopy in

A. pycnantha Benth. (Withers, 1979). Populations of A. melanoxylon R. Br. delayed heteroblastic development along a rainfall gradient where juvenile leaves were retained for longer under higher rainfall (Farrell and Ashton, 1978).

Understoreys of forests and higher rainfall environments (with greater cloud cover) generally receive lower doses of UV radiation (Brown et al., 1994; Paul and

Gwynn-Jones, 2003) and consequently heteroblastic development may also affect different populations. Populations from higher rainfall regions with a greater canopy cover may be expected to tolerate less any increased level of UV relative to other populations.

UV radiation may additionally alter whole-plant and leaf-level traits. Whole-plant traits are defined as the allocation of biomass to entire organs such as roots, leaves or stem. Leaf-level traits are defined as those traits associated within the organ of the leaf such as width, length or mass. Stressful levels of UV light (particularly

UV-B) are known to have adverse affects on whole-plant and leaf-level traits including decreased relative growth rates, stunted height growth, increased branching, increased root mass, smaller leaves, and an increase in the amount of leaf mass per photosynthetic area (specific leaf area) (Tevini and Teramura, 1989;

Rozema et al., 1997a; Frohnmeyer and Staiger, 2003; Jenkins and Brown, 2007).

Conversely, ambient levels of UV radiation have been demonstrated to benefit certain species. For example, enhanced UV-B can increase reproduction in certain species (Björn et al., 1997; Stephanou and Manetas, 1998) and can also improve

drought tolerance by decreasing stomatal density (Gitz et al., 2005) and increasing leaf cuticle thickness (Björn et al., 1997). There are also cases of positive effects on stem growth, total leaf number, leaf mass and leaf area (Barnes et al., 2005). As a result, the affects of UV radiation on whole-plant and leaf-level traits is not unidirectional but rather can be highly complex.

The major aim of this study was to test the effects of enhanced versus reduced levels of UV radiation on the growth performance of seedlings of a heteroblastic

Acacia species, A. implexa Benth., from three populations along a light gradient.

Our first hypothesis was increased UV will induce earlier heteroblastic leaf development. Our second hypothesis was enhanced UV should alter whole-plant traits such that plants would be stunted with decreased growth rates. Thirdly, we hypothesised that leaf-level traits would accommodate the more intense light environment by being smaller and shorter. Lastly, we hypothesised that plants from a high light population would tolerate enhanced levels of UV radiation more than plants from the low light population.

METHODS

Study Species

Acacia implexa is a tree growing between 5m and 12m inhabiting woodlands and open forests of eastern Australia (Kodela, 2002; Maslin and McDonald, 2004). A. implexa undergoes heteroblastic development as a seedling. Early shoot development is characterised by double compound leaves that develop from approximately the first ten nodes before a shift to transitional leaves and phyllodes.

The shift from compound leaves to transitional leaves is heteroblastic in this species whereby compound leaves do not develop after the developmental shift.

However transitional leaves can develop following the onset of phyllodes and this is an example of heterophylly. As mature adults A. implexa only develops phyllodes.

Experimental Design

We ran a split-block full-factorial design with two factors: 2 x UV level and 3 x population. The UV treatment corresponded to a reduced and enhanced radiation level. The reduced treatment was achieved by the glasshouse where the experiment was conducted which excluded all UV radiation. The enhanced treatment was achieved by supplementing plants in the glasshouse with UV radiation emitted from two lamps spaced 0.3m apart (type G15T8.E, 436mm long and 25.5mm diameter; Sankyo Denki Co. Ltd, Japan). Lamps had a peak spectral output at

306nm and 14.7W lamp wattage. Lamps were kept above growing plants at approximately 50cm. During a trial run lamps were hung less than 50cm above growing plants and plants exhibited clear signs of stress particularly with “burnt” leaves. The purpose of this experiment was not to induce extreme stress but to moderately enhance UV radiation and a level of about 50cm was deemed most appropriate. Using an automatic timer lamps were switched on at 10am and turned off at 4pm each day.

Three populations from a rainfall gradient were used in this experiment. A rainfall gradient was chosen as this is correlated with canopy density and light intensity. In higher rainfall regions A. implexa tends to be a sub-canopy species whereas in

lower rainfall regions A. implexa is the main canopy (pers. ob.). UV radiation is known to be absorbed by the canopy (e.g. Brown et al., 1994; Gies et al., 2007).

Additionally, the low rainfall population has more clear days and less cloudy days when compared to the medium and high rainfall populations. Therefore the three populations chosen in this experiment were chosen to represent populations with different histories and tolerances to UV radiation (i.e. plants from low rainfall will have greater tolerance than plants from high rainfall). Site descriptions for the three populations are provided in Table 3.1. Seeds were sourced from the CSIRO

Australian Tree Seed Centre, Canberra, Australia

(http://www.ffp.csiro.au/tigr/atscmain).

Plants were grown in 115mL pots in a potting mix that consisted of 33%

Australian Native Landscape supply of “Organic Garden Mix” (with a ratio of

50% black soil, 20% coarse sand and 30% organics), 33% washed river sand and

33% cocopeat. Our previous experience in growing Australian Acacia species found good results using this complex soil mixture. Additionally, a 4 month

Osmocote low phosphorus slow release fertiliser was added with the N:P:K ratio of 17:1.5:8.5. No additional fertiliser was added throughout the experiment due to this species sensitivity to excess nutrients and the fact the experiment ran for less than 4 months. All soil, pots and any other equipment used in the experiment were completely sterilised to prevent infection of roots of symbiotic rhizobium (which fix atmospheric nitrogen thus improving fertility and growth rates of plants). At time of final harvest no plants were found to have rhizobium. Plants were kept well watered at all times.

The factorial design consisted of 15 individuals assigned to three populations and two UV light levels to give a total sample size of 90. 15 plants died during the experiment (mortality was random across treatment types) giving a total harvest size of 75. Pots were located on two glasshouse benches oriented north-south with the eastern and western benches separated by opaque plastic. One bench contained all reduced UV treatment pots whereas the other bench contained all enhanced UV pots. Each day pots were moved about randomly within benches to control for positional effects from UV lamps. Glasshouse positional effects were minimised by switching the UV treatments from the east to west bench and vice versa once per week. In early-December 2007 seeds were pre-treated in boiling water for approximately 2 minutes to promote germination. Six seeds were sown directly into potted soil as previous trials had found this led to more or less uniform germination. Following the emergence of the first double compound leaf pots were thinned so as there was only one plant per pot. At this point UV lamps were switched on and the treatments commenced. Plants were checked after 28 days of growth for signs of transitional leaf development. Plants that had at least one fully developed transitional leaf were harvested. Thereafter, plants were checked on a weekly basis for fully developed transitional leaves and after 66 days all plants were harvested.

At the start of the experiment initial measurements were taken on stem length and diameter. A set of 20 non-experimental plants per population were harvested at this point with these same measurements taken. These plants were then oven dried at

60°C for at least seven days and biomass was measured. Regression of biomass on stem length and diameter gave initial biomass necessary for growth rate

calculations. At the time of harvest the following traits were measured on all experimental plants: stem length, diameter at base of plant, total number of nodes, node of first transitional leaf, number of leaves, and number of transitional leaves.

At time of harvest no plants had developed phyllodes. Leaves and petioles were removed from the stem and flattened (to prevent pinnules from folding inward) for area measurements. The first transitional leaf was separately removed and phyllode length, pinnae length and pinnae width, leaf area and leaf mass were measured.

This leaf was chosen as a representative of leaf trait variability among treatments and populations as it corresponded to a common developmental stage. Roots were removed from the stem and soil was carefully removed over a plastic container in order to capture all root material. All plant material was oven dried at 60°C for at least seven days before mass measurements were taken. The area of compound and transitional leaves were taken separately on a flatbed scanner (Epson Stylus,

CX4900) and using LeafA software (G Williamson, University of Adelaide,

Australia, ‘pers. comm.’). The mass of the different leaf types was also made separately.

Whole-plant and leaf-level traits that were derived from the measurements just described are summarised in Table 3.2. Relative growth rate was determined by:

(Ln Harvest Biomass) - (Ln Initial Biomass) / Number of days grown (1)

RGR = NAR x SLA x LMR (2)

Statistical Analyses

To test the hypothesis that the shift from compound leaves to transitional leaves will be hastened under enhanced UV a Type III univariate ANOVA was performed with population and UV level as fixed factors and the node number of the first transitional leaf as the dependent variable.

Multivariate Analysis of variance (MANOVA) was used to test the hypothesis of lower whole-plant performance under enhanced UV radiation. MANOVA considered UV level and population as fixed factors and dependent variables included: HtoD, Internode, StMR, RMR, LMR, LAR, NAR and SLA (see Table

3.2 for definitions of abbreviations). Homogeneity of variances in MANOVA were checked with Levene’s test of equality, outliers checked with Mahalanobis distances and Pillai’s Trace was used as the test statistic. Multivariate sets of covarying traits were further characterised using discriminant function analysis.

DFA is a classification procedure that can be applied on factorial multivariate datasets to assess whether groups can be uniquely separated (Quinn and Keough,

2002). In this study we were interested in whether the UV by population interaction could be classified into unique groups or whether there were overlaps in groups. Group centroids with 95% confidence intervals of experimental treatments were plotted against the first and second discriminant functions using procedures outlined in Johnson et al. (2007). A series of univariate ANOVAs were then performed on each dependent variable to assess for treatment differences. All

p-values were corrected for Type I error using sequential Bonferroni technique

(SPSS ver. 15.0.1). The same multivariate model was performed on leaf-level traits however a co-variate was included. Leaf traits are known to vary with plant size

(Rice and Bazzaz, 1989) and total dry biomass at time of harvest was used as a covariate to correct for possible bias arising out of different plant sizes imposed by experimental treatments. Leaf-level variables included: area, mass, phyllode length, pinnae length, and pinnae width (see Table 3.2 for definitions of abbreviations).

Two measures of growth were analysed: relative growth rate (RGR) and number of nodes developed per day (Nodes Day). RGR is a common trait that measures how well plants grow in a particular environment and Nodes Day is a measure of the potential to develop leaves. The two-way ANOVA model described above was also used on these traits.

All traits were assessed for normality using the Shapiro-Wilk test and were transformed where appropriate. Traits and their transformations have been summarised in Table 3.2. All statistical analyses were performed using SPSS

15.0.1.

RESULTS

Heteroblastic Development

Enhanced UV radiation did not alter the timing of heteroblastic development in

Acacia implexa (F1,69 = 0.279, p = 0.599; Figure 3.2) and the timing between populations was marginally non-significant (F2,69 = 2.485, p = 0.091). Additionally

there was no significant interaction between UV level and source population on heteroblastic development (F2,69 = 2.325, p = 0.105; Figure 3.2).

Whole-Plant Traits

MANOVA indicated that whole-plant traits differed significantly under the UV treatment, across populations, and there was a significant interaction between population and UV-B (Table 3.3). Univariate ANOVAs indicated that HtoD,

StMR, RMR, LAR, NAR and SLA were significantly different under UV treatment (Table 3.4). Pairwise comparisons showed that RMR and NAR were significantly greater in reduced UV whereas HtoD, StMR, LAR and SLA were significantly greater in enhanced UV (Figure 3.3). Univariate ANOVAs indicated that all whole-plant traits differed significantly across populations (Table 3.4) and pairwise comparisons showed complex patterns (Figure 3.3). Across the UV by population interaction there was a significant difference in HtoD, Internode, StMR,

LAR and SLA (Table 3.4). Student-Newman-Kuels (S-N-K) post-hoc analysis revealed that the low and medium light populations had higher HtoD and longer

Internode than the high light population irrespective of UV treatment and that

StMR was higher in the UV treatment irrespective of population source (Figure

3.3). Patterns within the LAR and SLA traits were complex however both traits had significantly lower values in the reduced UV by low light population treatment.

100

Leaf-Level Traits

Plant size did have an effect on leaf-level traits as indicated by a significant biomass effect following MANCOVA (Table 3.3). With biomass as a covariate, univariate ANCOVAs indicated that only phyllode length and leaf mass were significantly different under the UV treatment (Table 3.4) and that these were larger under reduced UV (Figure 3.4). Only pinnae length differed among populations (Table 3.4) with the medium light level population have longer pinnae

(Figure 3.4). Leaf area was the only significantly different trait under the population x UV interaction (Table 3.4) with area decreasing under enhanced UV for the medium and high light populations yet increasing for the low light population (Figure 3.4). Removing biomass as a covariate and running a

MANOVA on leaf-level traits revealed significant UV, population, and UV by population treatment effects. Univariate ANOVAs found transitional leaves to be much larger under reduced UV as indicated by longer phyllodes, longer and wider pinnae, and larger area and mass (Table 3.4, Figure 3.4). This same pattern held across all populations with the exception of the low light population which had similar pinnae width, area and mass.

Due to the significant difference in leaf mass revealed by the MANCOVA on leaf- level traits we separated the leaf into three components (phyllodes, rachis and pinnules, Figure 3.1) to determine if any single component was contributing to the significant leaf mass difference. Following MANCOVA (with total plant biomass as the covariate) there was no significant UV treatment effect (Table 3.3).

However univariate ANCOVA showed phyllode mass to be significantly greater in the reduced UV treatment and there was no difference in rachis mass and pinnule

101

mass (Table 3.4, Figure 3.5). MANCOVA of the leaf-level traits also revealed phyllode length to be longer in the reduced UV treatment therefore we ran an

ANCOVA with phyllode mass as the dependent variable and phyllode length as the covariate to determine if phyllode mass was significantly greater in reduced

UV because the phyllode was longer. Results from the ANCOVA showed that the covariate (phyllode length) had a significant effect on the model and there was no significant difference in phyllode mass within the UV treatment. Hence the significant difference in leaf mass detected in the MANCOVA of leaf-level traits was a consequence of phyllodes being longer.

Discriminant Function Analysis

DFA of whole-plant traits successfully separated treatments into groups (Wilk’s λ

= 0.079, F40,269 = 5.298, p < 0.0001) with the first discriminant function explaining

62.7% of the variance and the second discriminant function explaining 21.2% of the variance. Populations could be separated along the first discriminant function whereas the UV treatment could be separated along the second discriminant function (Figure 3.6). Stem elongation traits (HtoD and Internode) were positively associated with the first discriminant function and RMR and NAR were positively associated with the second discriminant function. StMR, LMR, LAR and SLA were negatively associated with the second discriminant function.

DFA of leaf-level traits also successfully separated treatments into groups (Wilk’s

λ = 0.411, F25,239 = 2.589, p = 0.0001) with the first discriminant function explaining 44.8% of the variance and the second discriminant function explaining

34.2% of the variance. Visual inspection of the group centroids with 95%

102

confidence circles shows populations could not be separated within the enhanced

UV treatment and that the high light by reduced UV treatment could not be separated from that group (Figure 3.6).

Growth Traits

There were no significant differences in RGR between reduced and enhanced UV treatments (F1,69 = 0.010, p = 0.922; Figure 3.7A) however there were significant differences across populations (F2,69 = 42.353, p < 0.001). The medium level light population had the fastest growth rate followed by the low light then high light populations (Figure 3.7A). There was also a significant interaction between population x UV treatment (F2,69 = 9.181, p < 0.001) whereby RGR was not significantly different yet increased for the low light population (S-N-K = 3.698, p

= 0.001, Figure 3.7A).

There were highly significant differences in the number of nodes developed per day in the UV treatment (F1,69 = 15.153, p < 0.0001) as well as across populations

(F2,69 = 45.173, p < 0.0001) and their interaction (F2,69 = 7.484, p < 0.0001). The rate of node development was significantly greater under enhanced UV (Figure

3.7B). Across populations nodes developed at a greater rate from the medium to low to high light level populations (Figure 3.7B).

103

DISCUSSION

Heteroblastic Development

Increased levels of UV did not alter heteroblastic development in Acacia implexa in this experiment. We predicted a shift from compound to transitional leaves would be hastened under enhanced UV levels as a previous study has found A. implexa to develop transitional leaves earlier in high light environments (Forster and Bonser, 2009). The fact that this did not occur in this experiment suggests: (i) some other component of the light spectrum principally controls the expression of leaf type (e.g. the red to far red ratio); (ii) compound leaves can perform relatively well under high light conditions (Brodribb and Hill, 1993) and therefore the light environment experienced in this experiment was not perceived as stressful; or (iii) plants lacked the ability to detect the experimental levels of UV radiation.

UV Treatment Effects

Enhanced and reduced levels of UV had complex effects on plant phenotypes in this experiment. Our hypothesis that the whole-plant would be smaller generally held (i.e. less total biomass at harvest, albeit marginally non-significant [Figure

3.7C]) however there was no difference in relative growth rates (RGR). Leaf-level traits were shorter and smaller and this supported our hypothesis. The partitioning of biomass within leaves ultimately explained how plants under contrasting UV environments had similar RGR and also explained the complexity of phenotypes developed in this experiment.

Firstly, enhanced UV radiation is known to decrease overall leaf size (Rozema et al., 1997a; Jansen et al., 1998; Barnes et al., 2005; Jenkins and Brown, 2007) and

104

the results from the leaf-level trait analysis in this study did find smaller leaves in the enhanced UV treatment. More precisely, phyllode length and leaf mass were significantly lower in the enhanced UV treatment and there was no difference in pinnae length, width or leaf area (Table 3.4, Figure 3.4). A closer examination of the components of the leaf (i.e. phyllode, rachis and pinnules; Figure 3.5) showed that only phyllode mass was significantly heavier in the reduced UV treatment and this was due to the fact that phyllodes were also significantly longer.

As a result enhanced UV principally stunted the structural support of the main photosynthetic component (i.e. pinnae) of juvenile and transitional leaves. The structural support (i.e. petiole in juvenile leaves, phyllode in transitional leaves) contributes significantly to the overall mass of the leaf yet negligibly to overall leaf area. Consequently specific leaf area (SLA) is greater in plants grown under enhanced UV. This explains the difference in our study versus previous studies which have found SLA to decrease in enhanced UV environments (Teramura and

Sullivan, 1994).

A second discrepancy between our study and previous studies was the marginally non-significant difference in overall plant size and no difference in RGR. Stressful levels of UV radiation (particularly UV-B radiation) decrease total plant size

(Tevini and Teramura, 1989; Smith et al., 2000; Searles et al., 2001; Jenkins and

Brown, 2007) and leads to slower RGR (Ballaré et al., 1996). Ambient levels of

UV-B, that is levels in the contemporary environment, does not necessarily stunt growth and may actually improve performance in certain species (Björn et al.,

1997; Stephanou and Manetas, 1998; Zaller et al., 2004; Barnes et al., 2005).

105

However smaller total plant size and smaller leaves suggests enhanced UV did have a slight adverse affect on our experimental species. Dividing RGR into its three components (SLA, net assimilation rate [NAR], and leaf mass ratio [LMR]) revealed how contrasting UV light environments led to similar RGR. LMR did not differ between treatments yet SLA was higher and NAR lower in the enhanced UV treatment and this led to similar RGR. As explained above, the lower leaf mass and similar area increased SLA in the enhanced UV treatment. Consequently, the partitioning of biomass within the leaf ultimately explained how plants under contrasting UV environments had similar RGR.

The pattern of biomass investment into various components of the leaf (i.e. phyllode, rachis and pinnules) had further consequences for other whole-plant traits. Smaller leaves with high SLA allowed plants in enhanced UV to increase height growth relative to diameter. Although internodes were not shorter under enhanced UV the greater height allowed for the production of more nodes. The increased development of nodes corresponds to more leaf meristem which in turn allows for more development of leaves. In a positive feedback loop, plants under enhanced UV allocated less biomass to individual leaves yet maintained a proportional leaf mass ratio by growing a higher number of leaves. Furthermore, leaf area of individual leaves was similar for both UV treatments however total leaf area (i.e. leaf area ratio [LAR]) was significantly greater in the enhanced UV treatment due to a higher number of leaves. There was a trade-off in this strategy as resources were diverted from belowground growth (i.e. root mass ratio) under enhanced UV. Day and Demchik (1996) also observed a decrease in RMR and an

106

increase in shoot biomass under increased UV-B radiation and attributed the pattern as a photomorphogenic response.

UV Effects on Populations

Within populations of A. implexa UV radiation had a highly complex effect on whole-plant and leaf-level traits. In regards to RGR, plants from the high and medium level light populations were not affected by enhanced UV yet the low light level population benefited from it. Plants from the low light level population significantly increased RGR under enhanced UV radiation. Across all populations

NAR was higher under reduced UV however this was statistically non-significant.

SLA, on the other hand, was higher in the enhanced UV treatment for the medium and low light level populations (Figure 3.3). The higher SLA in the low light population was the main driver behind a more rapid RGR.

The high light population showed little response to the UV treatments whereas the low light population showed more significant responses, particularly across whole- plant traits (Table 3.4, Figure 3.3). Additionally, discriminant function analysis showed close overlap of 95% confidence intervals for the whole-plant traits of the high light population and large overlap for the leaf-level traits (Figure 3.6). DFA in fact grouped the reduced UV by high light population with all the populations by enhanced UV treatments indicating that leaf-level traits were similar.

There have been previous cases where species or populations originating from a high light, or high UV, environment can better tolerate experimentally enhanced

UV radiation (Sullivan et al., 1992; Ziska et al., 1992; Rozema et al., 1997b;

107

Hofmann et al., 2000). A difference between those experiments and the populations of A. implexa in this experiment is the apparent beneficial effects of

UV radiation on the population originating from the low UV environment. Perhaps the increase in UV acted as a signal indicating that the light environment was favourable for rapid growth and height extension. In particular, UV-B can act as a damage/stress signal yet it has also been demonstrated that at lower fluence rates it can act as a growth signal (Frohnmeyer and Staiger, 2003; Jenkins and Brown,

2007). Typically the signal leads to protection mechanisms and the suppression of growth (Jenkins and Brown, 2007) rather than increased growth as demonstrated by the low light population of A. implexa.

In large doses UV radiation can be detrimental to plants however responses to ambient UV levels are mixed and not always unfavourable. Response to varying levels of UV radiation is clearly species dependent and the results presented here suggest further population dependency within species. Different trait strategies across species and within species can alter relative growth rates such that there is a mixed effect from UV radiation. The generality of this trend though needs further investigation.

Acknowledgements We thank Brenton Ladd, Geoff McDonnell and Natalie Wynyard for glasshouse assistance and Grant Williamson for leaf area software. Funded by UNSW Faculty Research Grant and Early Career Research grant to S.P.B.

108

REFERENCES

Ballaré C, Scopel A, Stapleton AE, Yanovsky MJ. 1996. Solar ultraviolet-B radiation affects seedling emergence, DNA integrity, plant morphology, growth rate, and attractiveness to herbivore insects in Datura ferox. Plant Physiology 112, 161–170.

Barnes PW, Shinkle JR, Flint SD, Ryel RJ. 2005. UV-B radiation, photomorphogenesis and plant-plant interactions. Progress in Botany 66, 313–340.

Bjorn LO, Callaghan TV, Johnsen I, Lee JA, Manetas Y, Paul ND, Sonesson M, Wellburn AR, Coop D, Heide-Jørgensen HS, Gehrke C, Gwynn- Jones D, Johanson U, Kyparissis A, Levizou E, Nikolopoulos D, Petropoulou Y, Stephanou M. 1997. The effects of UV-B radiation on European heathland species. Plant Ecology 128, 253–264.

Bodkin PC, Spence DHN, Weeks DC. 1980. Photoreversible control of heterophylly in Hippuris vulgaris L. New Phytologist 84, 533–542.

Brodribb T, Hill RS. 1993. A physiological comparison of leaves and phyllodes in Acacia melanoxylon. Australian Journal of Botany 41, 293–305.

Brown MJ, Parker GG, Posner NE. 1994. A survey of ultraviolet-B radiation in forests. Journal of Ecology 82, 843–854.

Burns KC. 2005. Plastic heteroblasty in beach groundsel (Senecio lautus). New Zealand Journal of Botany 43, 665–672.

Caldwell MM, Ballaré CL, Bornman JF, Flint SD, Björn LO, Teramura AH, Kulandaivelu G, Tevini M. 2003. Terrestrial ecosystems, increased solar ultraviolet radiation and interactions with other climatic change factors. Photochemical & Photobiological Sciences 2, 29–38.

Caldwell MM, Bornman JF, Ballaré CL, Flint SD, Kulandaivelue G. 2007. Terrestrial ecosystems, increased solar ultraviolet radiation, and interactions with other climate change factors. Photochemical & Photobiological Sciences 6, 252–266.

Cronin G, Lodge DM. 2003. Effects of light and nutrient availability on the growth, allocation, carbon/nitrogen balance, phenolic chemistry, and resistance to herbivory of two freshwater macrophytes. Oecologia 137, 32– 41.

Day TA, Demchik SM. 1996. Influence of enhanced UV-B radiation on biomass allocation and pigment concentrations in leaves and reproductive structures of greenhouse-grown Brassica rapa. Vegetatio 127, 109–116.

109

Forster MA, Bonser SP. 2009. Heteroblastic development and the optimal partitioning of traits among contrasting environments in Acacia implexa. Annals of Botany, 103:95-105.

Frohnmeyer H, Staiger D. 2003. Ultraviolet-B radiation-mediated responses in plants. Balancing damage and protection. Plant Physiology 133, 1420– 1428.

Gies P, Elix R, Lawry D, Gardner J, Hancock T, Cockerell S, Roy C, Javorniczky J, Henderson S. 2007. Assessment of the UVR protection provided by different tree species. Photochemistry and Photobiology 83, 1465–1470.

Gitz DC, Liu-Gitz L, Britz SJ, Sullivan JH. 2005. Ultraviolet-B effects on stomatal density, water-use efficiency, and stable carbon isotope discrimination in four glasshouse-grown soybean (Glyicine max) cultivars. Environmental and Experimental Botany 53, 343–355.

Hofmann R, Swinny EE, Bloor SJ, Markham KR, Ryan KG, Campbell BD, Jordan BR, Fountain DW. 2000. Responses of nine Trifolium repens L. populations to ultraviolet-B radiation: differential flavonol glycoside accumulation and biomass production. Annals of Botany 86, 527–537.

Jansen MAK, Gaba V, Greenberg BM. 1998. Higher plants and UV-B radiation: balancing damage, repair and acclimation. Trends in Plant Sciences 3, 131– 135.

Jenkins GI, Brown BA. 2007. UV-B perception and signal transduction. In: Whitelam GC, Halliday KJ, eds. Light and Plant Development, Oxford, UK: Blackwell Publishing, 155–182.

Johnson HE, Llyod AJ, Mur LAJ, Smith AR, Causton DR. 2007. The application of MANOVA to analyse Arabidopsis thaliana metabolomic data from factorially designed experiments. Metabolomics 3, 517–530.

Jones CS. 1999. An essay on juvenility, phase change, and heteroblasty in seed plants. International Journal of Plant Sciences 160, S105–S111.

Kodela PG. 2002. Acacia. In: Harden GJ, ed. Flora of New South Wales, Kensington, NSW: UNSW Press, 381–476.

Krummel P, Fraser P. 2006. Ozone depletion over Australia and the Antarctic. In: Attiwill P, Wilson B, eds. Ecology: an Australian perspective, South Melbourne, VIC: Oxford University Press, 123–140.

Maslin BR, McDonald MW. 2004. AcaciaSearch: Evaluation of Acacia as a woody crop option for southern Australia, Canberra, ACT: Rural Industries Research and Development Corportation.

110

McKenzie RL, Aucamp PJ, Bais AF, Bjorn LO, Ilyase M. 2007. Changes in biologically-active ultraviolet radiation reaching the Earth’s surface. Photochemical & Photobiological Sciences 6, 218–231.

New TR. 1984. A Biology of Acacias, Melbourne, VIC: Oxford University Press.

Paul ND, Gwynn-Jones D. 2003. Ecological roles of solar UV radiation: towards an integrated approach. Trends in Ecology and Evolution 18, 48–55.

Quinn GP, Keough MJ. 2002. Experimental Design and Data Analysis for Biologists, Port Melbourne, VIC: Cambridge University Press.

Rice SA, Bazzaz FA. 1989. Quantification of plasticity of plant traits in response to light intensity: comparing phenotypes at a common weight. Oecologia 78, 502–507.

Rozema J, Chardonnens A, Tosserams M, Hafkenscheid R, Bruijnzeel S. 1997b. Leaf thickness and UV-B absorbing pigments of plants in relation to an elevational gradient along the Blue Mountains, Jamaica. Annals of Botany 128, 150–159.

Rozema J, Staaij Jvd, Bjorn LO, Caldwell M. 1997a. UV-B as an environmental factor in plant life:stress and regulation. Trends in Ecology and Evolution 12, 22–28.

Searles PS, Flint SD, Caldwell MM. 2001. A meta-analysis of plant field studies simulating stratospheric ozone depletion. Oecologia 127, 1–10.

Smith JL, Burritt DJ, Bannister P. 2000. Shoot dry weight, chlorophyll and UV- B-absorbing compounds as indicators of a plant's sensitivity to UV-B radiation. Annals of Botany 96, 1057–1063.

Stephanou M, Manetas Y. 1998. Enhanced UV-B radiation increases the reproductive effort in the Mediterranean shrub Cistus creticus under field conditions. Plant Ecology 134, 91–96.

Sullivan JH, Teramura AH, Ziska LH. 1992. Variation in UV-B sensitivity in plants from a 3,000-m elevational gradient in Hawaii. American Journal of Botany 79, 737–743.

Teramura AH, Sullivan JH. 1994. Effects of UV-B radiation on photosynthesis and growth of terrestrial plants. Photosynthesis Research 39, 463–473.

Tevini M, Teramura AH. 1989. UV-B effects on terrestrial plants. Photochemical & Photobiological Sciences 50, 479–487.

Winn AA. 1999. The functional significance and fitness consequences of heterophylly. International Journal of Plant Sciences 160, S113–S121.

111

Withers JR. 1979. Studies on the status of unburnt Eucalyptus woodland at Ocean Grove, Victoria. IV. The effect of shading on seedling establishment. Australian Journal of Botany 27, 47–66.

Zaller JG, Searles PS, Caldwell MM, Flint SD, Scopel AL, Sala OE. 2004. Growth responses to ultraviolet-B radiation of two Carex species dominating an Argentinian fen ecosystem. Basic and Applied Ecology, 5, 153–162.

Ziska LH, Teramura AH, Sullivan JH. 1992. Physiological sensitivity of plants along an elevational gradient to UV-B radiation. American Journal of Botany 79, 863–871.

112

Table 3.1 Brief description of populations of sourced seeds. Climate data were collated from Bureau of Meteorology, Australia climate data of nearest weather station. A clear day is defined as cloud observation at 9am and 3pm with ≤ 2 oktas. A cloudy day is defined as cloud observation at 9am and 3pm with ≥ 6 oktas. All values are annual averages. Rainfall Clear Cloudy Avg Temp. Population Seedlot Location (mm p.a.) Days Days (°C)

High Light 19780 35°57′S 450 131 100 23 145°45′E Medium 19770 32°37′S 650 113 108 23 Light 150°03′E Low Light 18859 31°56′S 830 92 114 24 151°11′E

113

Table 3.2 Summary of traits measured in Chapter 3 with abbreviations and transformations made for statistical analysis. An example of pinnae, phyllode, rachis and pinnules is given in Figure 3.1. Trait Abbrev. Definition Trans. a) Whole-Plant Traits

Height to Diameter HtoD Ratio of stem length to basal log (mm cm-2) stem diameter Internode Length Internode Stem length divided by the log (mm) total number of nodes Stem Mass Ratio StMR Ratio of total stem mass to sum (g g-1) of leaf and root mass Root Mass Ratio RMR Ratio of total root mass to sum log (g g-1) of stem and leaf mass Leaf Mass Ratio LMR Ratio of total leaf mass to sum (g g-1) of stem and root mass Leaf Area Ratio LAR Ratio of total leaf area to total (cm2 g-1) biomass Net Assimilation Rate NAR The amount of carbon fixed sqrt (g cm-2 day-1) per day for a given amount of photosynthetic tissue Specific Leaf Area SLA Ratio of leaf area to leaf mass (cm2 g-1) b) Leaf-Level Traits

Phyllode Length Length of phyllode of (mm) transitional leaf Pinnae Length Length of largest pinnae from (mm) phyllode to tip Pinnae Width Distance at widest point of log+1 (mm) largest pinnae Area Area of one-side of leaf sqrt (cm2) including phyllode Mass Mass of leaf including (g) phyllode Phyllode Mass Mass of phyllode taken from sqrt (g) first transitional leaf Rachis Mass Mass of rachis taken from first sqrt (g) transitional leaf Pinnule Mass Mass of pinnules taken from sqrt (g) first transitional leaf c) Growth Traits

Relative Growth Rate RGR Amount of biomass fixed per (g day-1) day; summarised by Eq. 3 Nodes per Day log

114

Table 3.3 Results from multivariate analysis of variance. * Denotes traits were analysed using multivariate analysis of co-variance with biomass as the covariate. Leaf organ mass refers to phyllode mass, rachis mass and pinnule mass (see Table 3.2). Full results are found in [Supplementary materials 3.1-4]. Pillai’s d.f. F-value P-value Trace

Whole-Plant Population 1.034 16, 124 8.294 < 0.0001 UV 0.387 8, 61 4.814 0.0001 Population x UV 0.458 16, 124 2.301 0.0054

Leaf-Level* Biomass 0.437 5, 63 9.796 < 0.0001 Population 0.327 10, 128 2.497 0.0090 UV 0.161 5, 63 2.422 0.0452 Population x UV 0.280 10, 128 2.081 0.0305

Leaf-Level Population 0.361 10, 130 2.859 0.0030 UV 0.210 5, 64 3.399 0.0087 Population x UV 0.278 10, 130 2.100 0.0288

Leaf Organ Mass* Biomass 0.436 3, 65 16.719 < 0.0001 Population 0.332 6, 132 4.377 0.0005 UV 0.091 3, 65 2.181 0.0987 Population x UV 0.083 6, 132 0.949 0.4625

Leaf Organ Mass Population 0.398 6, 134 5.544 < 0.0001 UV 0.144 3, 66 3.714 0.0157 Population x UV 0.093 6, 134 1.093 0.3701

115

Table 3.4 Results from analysis of variance. * Denotes traits were analysed using analysis of co-variance with biomass as the covariate. Full results are found in [Supplementary materials 3.1-4]. Population UV Population x UV

F-value P-value F-value P-value F-value P-value Whole-Plant HtoD 35.160 < 0.0001 9.441 0.0031 3.619 0.0321 Internode 26.048 < 0.0001 0.168 0.6829 3.678 0.0304 StMR 3.081 0.0524 19.677 < 0.0001 3.209 0.0466 RMR 3.573 0.0335 9.238 0.0034 0.337 0.7150 LMR 3.297 0.0430 1.829 0.1807 0.200 0.8193 LAR 12.373 < 0.0001 10.506 0.0018 3.267 0.0442 NAR 15.311 < 0.0001 6.704 0.0118 0.262 0.7702 SLA 10.614 0.0001 8.363 0.0051 4.534 0.0142

Leaf-Level* Phyllode Length 2.079 0.1331 6.428 0.0136 0.138 0.8713 Pinnae Length 3.371 0.0403 2.438 0.1231 0.028 0.9724 Pinnae Width 1.535 0.2230 1.839 0.1796 1.804 0.1726 Leaf Area 1.550 0.2198 1.924 0.1700 3.121 0.0506 Leaf Mass 0.002 0.9982 4.766 0.0325 2.872 0.0636

Leaf-Level Phyllode Length 1.669 0.1961 9.795 0.0026 0.339 0.7134 Pinnae Length 6.411 0.0028 4.068 0.0477 0.130 0.8786 Pinnae Width 3.805 0.0271 3.753 0.0569 2.402 0.0982 Leaf Area 4.834 0.0109 5.456 0.0225 4.112 0.0206 Leaf Mass 2.792 0.0684 9.320 0.0032 3.151 0.0491

Leaf Organ Mass* Phyllode Mass 1.029 0.3628 2.425 0.1241 1.456 0.2406 Rachis Mass 5.418 0.0066 6.646 0.0121 1.650 0.1997 Pinnule Mass 0.936 0.3972 1.979 0.1641 2.145 0.1250

Leaf Organ Mass Phyllode Mass 5.609 0.0056 6.378 0.0139 2.099 0.1304 Rachis Mass 3.654 0.0311 11.343 0.0012 2.501 0.0895 Pinnule Mass 2.366 0.1015 5.801 0.0187 2.459 0.0931

116

Enhanced Reduced

Pinnae

Rachis Pinnules Phyllode

Figure 3.1 Transitional leaves from plants grown in enhanced and reduced UV treatments. Leaves were both taken from node 7 of the medium light population. Various organs of the leaf that were measured are labelled. Scale is 10mm.

117

4 Node of First Node of Transitional Leaf Transitional 2

0 + _ + _ + _ + _ High Medium Low UV Population x UV

Figure 3.2 The node at which the first transitional leaf developed on plants corresponds to timing of heteroblastic development. + enhanced UV; – reduced UV. Means (±S.E.) are shown and there were no significant differences across treatments.

118

2.5 1.4 c c a b b b b b a a a a

) ** -2 2.0 1.1

1.5

0.8 1.0 log Internodelog (mm) log HtoD (mm cm (mm HtoD log

0.5 0.5

0.5 0.6 **** bc bc a c ab c ** 0.4 ) ) -1 -1 0.4 0.3

0.2 0.2 RMR (g g StMR (g g 0.1

0.0 0.0 2.0 180 ** b b a ab ab c

1.5 ) ) -1

-1 120 g 2

1.0

60 LMR (g g

0.5 LAR (cm

0.0 0

0.04 300

) ***bc bc a ab ab c -1 250

0.03 ) day -1 -2 g

2 200 0.02 150

0.01 (cmSLA 100 sqrt NAR (g cm (g sqrt NAR 0.00 ++_ _ + _ + _ 50 + _ + _ + _ + _ High Med Low High Med Low UV Population x UV UV Population x UV

Figure 3.3 Comparisons of whole-plant traits in the UV and population by UV treatments. + enhanced UV; – reduced UV. * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001. Letters above the figure correspond to groupings following post-hoc SNK comparisons. Units of measurement are given in Table 3.2. Complete ANOVA results can be found in [Supplementary materials 3.1].

119

160 160 * †† † 120 120

80 80

40 40 Pinnae Length (mm) Pinnae Length Phyllode Length (mm) Length Phyllode

0 0

2.0 5 † b ab b a b b 4 ) 1.6 2

1.2 2 sqrt Area (cm 1 0.8 log+1 Pinnae Width (mm) Width Pinnae log+1 0 + _ + _ + _ + _ 0.10 * High Med Low †† b b b a b b 0.08 UV Population x UV

0.06

0.04 Mass (g)

0.02

0.00 + _ + _ + _ + _ High Med Low UV Population x UV

Figure 3.4 Comparisons of leaf-level traits in the UV and population by UV treatments. + enhanced UV; – reduced UV. Significant differences in the UV treatment following ANCOVA is given by: * p < 0.05. Significant differences in the UV treatment following ANOVA are given by: † p < 0.05; †† p < 0.01. Letters above the figure correspond to groupings following post-hoc SNK comparisons. Units of measurement are given in Table 3.2. Complete ANOVA results can be found in [Supplementary materials 3.2-3].

120

0.3 *

0.2

0.1 sqrt Organ Mass (g) Organ sqrt

0.0 + _ + _ + _ Phyllode Rachis Pinnule

Figure 3.5 Mean (±S.E.) dry weight of three leaf organs: phyllode, rachis and pinnae; of the first transitional leaf. + enhanced UV; – reduced UV. * p < 0.05; n.s. = non- significant different. Units of measurement are given in Table 3.2. Complete ANOVA results can be found in [Supplementary materials 3.4].

121

Whole-Plant Leaf-Level 4 2

RM 2 1 RL

RH RM 0 0 RH EM EL EL EH EM RL EH DF2 (21.2%) DF2 (34.2%) -2 -1

-4 -2 -4 -2 0 2 4 -2 -1 0 1 2

DF1 (62.7%) DF1 (44.8%)

Figure 3.6 Grouping of population by UV treatment following discriminant function analysis. Displayed are group centroids with circles indicating 95% confidence intervals for whole plant and leaf-level traits (overlapping circles indicate no difference between groups). Variance explained by first and second discriminant functions displayed parenthetically. EH: enhanced UV by high light population; EM: enhanced UV by medium light population; EL: enhanced UV by low light population; RH: reduced UV by high light population; RM: reduced UV by medium light population; RL: reduced UV by low light population.

122

(A) 0.2 c bc a a a b ) -1 day -1 0.1 RGR (gRGR g

0.0

(B) 0.6 *** b b a a a b

0.4

0.2 Nodes per Day

0.0

1.2

0.9

0.6 Total Biomass Total 0.3

0.0 + _ + _ + _ + _ High Medium Low UV Population x UV

Figure 3.7 Mean (±S.E.) of growth traits in the UV and population by UV treatments. + enhanced UV; – reduced UV. *** p < 0.001. Letters above the figure correspond to groupings following post-hoc SNK comparisons. Units of measurement are given in Table 3.2. Complete ANOVA results can be found in [Supplementary materials 3.5- 7].

123

Chapter 4

Optimal allocation of resources in response to shading and neighbours in the heteroblastic species, Acacia implexa

Authors: Michael A. Forster, Brenton Ladd and Stephen P. Bonser

124

ABSTRACT

Heteroblasty involves the development of distinct juvenile and adult foliage and is thought to aid in seedling establishment. Acacia implexa Benth. develops juvenile compound leaves that can perform at a more optimal level under low irradiance than transitional or adult (phyllode) leaves. The lower construction costs of compound leaves should additionally assist seedling establishment in high density growth environments. We tested this hypothesis by running a full factorial design of light quality (high and low) and density levels (high, medium and low) on three populations sourced from different rainfall regions (high, medium and low). We also hypothesised that there would be an optimal allocation of biomass at the whole-plant and leaf-level according to the shade-avoidance syndrome in the shade-intolerant A. implexa.

Finally, overall performance was assessed with relative growth rates (RGR).

Heteroblastic development was significantly delayed under low quality light and in high density treatments however there were no significant interactions across treatments. Plants under low quality light developed phenotypes strongly associated with the shade-avoidance syndrome including greater stem mass, leaf area, and stem elongation; however this was not seen in the high density treatment and there were no consistent patterns across populations. Leaves were significantly longer and wider under low quality light yet there was no difference across density treatments. When the effects of plant size were removed leaves in the low density treatment were far larger than the high density treatment. There was no difference in RGR between light treatments but RGR were significantly higher when plants were sown at low density.

We conclude that heteroblasty does aid in seedling establishment, and that light is

125

clearly the most limiting factor to growth in a shaded environment however in a crowded environment there are additional limiting resources to growth.

126

INTRODUCTION

A great amount of interest has been directed towards understanding patterns of allocation under various light environments as plants not only respond to a direct decrease in light energy but can anticipate possible future decreases (Ballaré, 1999).

Photoreceptors within the plant have evolved to detect different spectral wavelengths that act as a signal to the surrounding light environment (Aphalo et al., 1999; Ballaré,

1999). The ratio of red (R) light to far red (FR) light is a particularly important signal for plants (Franklin and Whitelam, 2005). A lower R:FR ratio perceived by photoreceptors in the stem acts as a signal to detect neighbouring vegetation (Ballaré et al., 1990; Smith et al., 1990). A lower R:FR perceived by photoreceptors in the leaves on the other hand serves as a signal of actual shading by an overhanging canopy (Franklin and Whitelam, 2007). In either case plants alter allocation patterns in order to overcome the limiting light environment.

A change in allocation pattern when light energy becomes limiting to growth is known as the shade-avoidance syndrome and can be expressed at the whole-plant and leaf- level. Whole-plant traits are defined as the allocation of resources between organs such as stem, roots and leaves, whereas leaf-level traits involve characters associated within the leaf organ (Navas and Garnier, 2002). At the whole-plant level the shade- avoidance syndrome is expressed as an increase in stem mass and stem elongation at the expense of root and leaf mass (e.g. Smith, 1982; Schmitt and Wulff, 1993; Dudley and Schmitt, 1996; Franklin and Whitelam, 2005). Total leaf area however tends to increase and the ratio of leaf area to mass (specific leaf area) also increases (Morgan and Smith, 1981; Mitchell and Woodward, 1988; Yu and Ong, 2003; Poorter et al.,

127

2006). At the leaf level biomass is allocated so leaves and petioles become longer and wider (Barišić et al., 2006; Bell and Galloway, 2007). These whole-plant and leaf- level responses are typical of plants growing under low irradiance and low density or uncrowded environments. In crowded environments above-ground space and belowground resources are additionally limiting and therefore there may be a different expression of traits. For example, competition for below-ground resources may force the allocation of biomass to roots at the expense of the shoot (Weiner, 1990; Weiner et al., 1997). Therefore, in spite of light being a limiting factor to growth we may expect different phenotypes to develop under low quality light produced by shading or competitive effects produced by neighbours.

Shading and competition from neighbours not only affects patterns of biomass allocation but may also affect sequential patterns of development. For example, many plant species have dramatically different juvenile and adult leaves and the transition from one leaf type to another is known as heteroblastic development (Jones, 1999;

Wells and Pigliucci, 2000). Goebel (1898; cited in Wells and Pigliucci, 2000) noted that low light can prolong the expression of juvenile leaves. Although light is not the only factor affecting heteroblastic development, increasing experimental evidence supports the idea that heteroblasty can be hastened, or delayed, given certain light conditions (Burns, 2005; Forster and Bonser, 2009). Heteroblasty however has only been examined under actual shading and to our knowledge it has yet to be examined in a competitive context.

128

Heteroblastic development is a particularly prominent phenomenon in the subgenus

Phyllodineae and of the genus Acacia (New, 1984). Heteroblasty in Acacia involves a shift from juvenile leaves, to transitional leaves and then to adult leaves. Juvenile leaves are bipinnate compound whereas the adult leaves are modified petioles/rachis known as phyllodes (New, 1984). Transitional leaves consist of a phyllode with bipinnate leaflets attached distally. The shift from juvenile to transitional leaves is heteroblastic and the timing is species dependent. Generally, species from more arid environments switch to phyllodes at an earlier developmental stage than species from more mesic environments (Atkins et al., 1998). Additionally, there is a strong genetic component within species as different populations are also known to vary in the timing of heteroblasty (Farrell and Ashton, 1978; Daehler et al., 1999). Species and population differences in the timing of heteroblasty have been attributed to varying moisture regimes (Farrell and Ashton, 1978) however the light environment may play an important role. For example, heteroblasty has been experimentally demonstrated to be a plastic trait under contrasting light but not moisture environments (Forster and

Bonser, 2009). That is, under a low irradiance juvenile leaves are retained for longer.

Physiological and morphological traits of compound leaves are better adapted to lower irradiance whereas the traits of phyllodes are better adapted to more intense light environments (Brodribb and Hill, 1993; Yu and Li, 2007). Moreover compound leaves are less expensive to construct than phyllodes in terms of the amount of leaf biomass per photosynthetic tissue, or specific leaf area (Yu and Li, 2007; Forster and Bonser,

2009), a condition consistent with various other species grown at low irradiance

(Poorter et al., 2006). Leaves with lower construction costs and higher returns under low quality light may also be beneficial in high density or competitive growing

129

conditions and therefore heteroblastic species may retain more of these leaves in crowded or low light conditions. The timing of the developmental shift of juvenile to adult leaves should be environmentally dependent and under strong selection across environments. Thus, we should see differences in the expression of heteroblasty across populations, and variations in the expression of heteroblasty within populations.

Here we ran a full factorial designed experiment, with population, light quality, and density level as factors, to examine patterns of heteroblastic development and the expression of whole-plant and leaf-level traits on the species Acacia implexa. We hypothesised that heteroblasty will be delayed under a low quality light and a high density treatment, and that it should differ across populations. We also hypothesised that low quality light and high density treatments should induce a strong shade- avoidance syndrome in the shade-intolerant, A. implexa. Lastly, we examined how heteroblasty, whole-plant and leaf-level traits affected growth rates with the expected outcome that there should be no difference given the plasticity of these traits.

METHODS

Acacia implexa Benth. is a tree growing to approximately 12m inhabiting woodland and forest habitats across south-east and eastern Australia (Kodela, 2002). Acacia implexa is a pioneer species with seeds typically needing intense heat from fire to germinate and saplings are fast-growing (Maslin and McDonald, 2004). Seeds from three populations were obtained from the CSIRO Australian Tree Seed Centre,

Canberra, Australia. The three populations were chosen because they naturally inhabit

130

different points along a rainfall gradient (Table 4.1). Farrell and Ashton (1978) found heteroblastic development was delayed along a rainfall gradient in the closely related

Acacia melanoxylon and this motivated our choice of populations.

In mid-December 2006 seeds were pre-treated in boiling water for approximately 2 minutes to promote germination. Seeds were sown directly into potted soil as previous trials had found this led to uniform germination. Pots were 115mL in size and consisted of soil with 33% Australian Native Landscape supply of “Organic Garden

Mix”, 33% washed river sand and 33% cocopeat. A four month Osmocote low phosphorus slow release fertiliser was added with the N:P:K ratio of 17:1.5:8.5. No additional fertiliser was added throughout the experiment due to this species sensitivity to excess nutrients and the fact the experiment ran for less than four months. All soil, pots and any other equipment used in the experiment were completely sterilised to prevent infection of roots by symbiotic rhizobium (which fix atmospheric nitrogen and thus improve growth). At time of final harvest no plants were found to have rhizobium.

We ran a blocked full-factorial design with three factors: 3 populations (from high, medium and low rainfall sites), 2 levels of light quality (high and low), and 3 density levels (high, medium and low) with 15 replicates per treatment combination giving a total sample size of 270 pots. Eighty-eight pots experiences mortality and final replicates per treatment is summarised in [Supplementary materials 4.1]. Plants in the high light treatment received natural sunlight whereas the plants in the low light treatment were grown inside a 57cm cylinder (open ended) of green filter plastic (Lee

131

Filters, Andover, UK; number 121 Lee Green). The light filters simulate a natural canopy; the photosynthetic flux density was reduced by 35% and the ratio of red to far red light was reduced from 1.0 to 0.2 (Bonser and Geber, 2005). Cylinders were extended during the experiment if plants grew to within 10cm of the top of the 57cm light filtering cylinder. The density treatment was established at the time of seed sowing. 50, 15 and 6 seeds were sown for the high, medium and low density treatments. Following the emergence of the first double compound leaf, seedlings in the low density treatment were thinned to one plant per pot. Pots were spaced evenly across a glasshouse bench in the School of Biological, Earth and Environmental

Sciences glasshouse at the University of New South Wales.

At the emergence of the first double compound leaf five target plants were randomly selected in the medium and high density treatments for measurement. Five plants were chosen as a precaution against mortality during the course of the experiment. A marked toothpick was carefully placed next to each target plant as an identifier and a toothpick was also placed next to stems in the low density treatment as a control measure. Stem height and diameter were made as initial measurements on each target plant. The number of germinated stems was also counted to give initial levels of density. Plants were allowed to grow until at least three mature transitional leaves had developed. Thereafter, plants were checked weekly for at least three mature transitional leaves. Harvesting began 80 days following initial measurements and after

170 days of growth any remaining plants were harvested.

132

At time of harvest the following traits were measured: number of nodes with a compound leaf and total number of nodes, stem height and diameter. This data was used to calculate timing of heteroblasty, height to diameter ratio, and average internode length. Leaves were separated from the stem, flattened and kept refrigerated at 4°C until area measurements were taken. This prevented pinnules on the compound leaves from folding and modifying leaf area. Individual leaves were measured on a flatbed scanner and the LeafA software (G. Williamson ‘pers comm.’) was used to determine area. Individual leaves were oven dried at 60°C for at least 72 hours before determination of dry weight. Specific leaf area (SLA) was calculated as fresh leaf area divided by dried mass. Soil was carefully removed from roots in a plastic box to ensure all root material was collected. Stem and roots were oven dried at 60°C for at least 72 hours and mass taken. The addition of leaf, stem and root mass gave total plant dry weight at time of harvest. The determination of stem mass ratio (StMR), root mass ratio (RMR), leaf mass ratio (LMR) and leaf area ratio (LAR) was derived from these mass and area measurements and are summarised in Table 4.2. To derive leaf- level traits, ImageJ software (National Institute of Health, USA, http://rsb.info.nih.gov/ij) was used to measure petiole length, pinnae length and pinnae width of the last compound leaf before the onset of heteroblasty. Phyllode length, pinnae length and pinnae width was also measured on the first and third transitional leaves.

All measured traits were classified into whole-plant, leaf-level and growth traits for further statistical analysis and are described in Table 4.2. Leaf-level traits were taken

133

as the average measured from the last compound leaf, first and third transitional leaves. Relative growth rate and relative height growth rate were determined by:

RGR = (Ln x) - (Ln y) / Number of days grown (1)

where x is harvest biomass or height and y is initial biomass or height. Net assimilation rate (NAR) was determined by (e.g. Portsmuth and Niinemets, 2006):

NAR = RGR / (SLA x LMR) (2)

Statistical Analyses

To test whether heteroblastic development was delayed under low light and high density, we conducted a three-way univariate ANOVA with population, light quality, and density level as fixed factors and number of nodes that developed compound leaves as the dependent variable. The same model was used to test for the effects of population origin, light and density on relative growth rate, relative height growth and total plant dry weight at time of harvest.

A multivariate approach was then adopted to determine how experimental treatments shaped whole-plant and leaf-level phenotypes and the relation of traits relative to each other. Type III multivariate analysis of variance (MANOVA) considered population, light quality and density level as fixed factors and whole-plant dependent variables included: HtoD, Internode, StMR, RMR, LMR, LAR, NAR and SLA (see Table 4.2

134

for definitions of abbreviations). A separate MANOVA was performed on leaf-level dependent variables (petiole length, pinnae length, pinnae width, leaf area, leaf mass).

Further, leaf traits are known to vary with plant size (Rice and Bazzaz, 1989) therefore a multivariate analysis of co-variance was also performed on leaf-level traits with total plant dry weight at time of harvest as the covariate. Homogeneity of variances were checked with Levene’s test of equality [Supplementary materials 4.2]. The significance of model terms and interactions was assessed using Pillai’s trace test statistics. Multivariate sets of covarying traits were further characterised using discriminant function analysis. DFA is a classification procedure that can be applied on factorial multivariate datasets to assess whether groups can be uniquely separated

(Quinn and Keough, 2002). We performed four sets of DFAs for both whole-plant and leaf-level traits: (i) population x light x density; (ii) population x light; (iii) population x density; and (iv) light x density. Group centroids with 95% confidence intervals of experimental treatments were plotted against the first and second discriminant functions using procedures outlined in Johnson et al. (2007). A series of univariate

ANOVAs were then performed on each dependent variable to assess for treatment differences. All p-values were corrected for Type I error using sequential Bonferroni technique.

All traits were assessed for normality using the Shapiro-Wilk test and were transformed where appropriate. Traits and their transformations have been summarised in Table 4.2. Following transformations traits were then standardised with a mean of zero and standard deviation of one. All statistical analyses were performed using SPSS 15.0.1.

135

RESULTS

Heteroblastic Development

Compound leaves were retained for significantly longer in the low light over the high light treatment (F1,163 = 34.719, p < 0.0001) and in the high density over the medium and low density treatments (F2,163 = 8.660, p = 0.0003; Figure 4.1). The low rainfall population developed significantly more compound leaves over the medium and high rainfall populations (F2,163 = 43.466, p < 0.0001; Figure 4.1). There were no significant interactions between any of the experimental treatments [Supplementary materials 4.3].

Whole-Plant Traits

Whole-plant traits differed significantly under the population, light quality and density treatments however there were no significant differences across treatment interactions

(MANOVA, Table 4.3). A shade-avoidance syndrome was expressed in the high rainfall population relative to the medium and low rainfall populations and in the low light relative to the high light treatment. HtoD and Internode were longer and there was a greater allocation to stem mass and less to root mass in the high rainfall and low light treatments (Table 4.4, Figure 4.2). Plants in the low light treatment allocated greater percentage of biomass to leaf area (i.e. increased LAR and SLA, Table 4.4 and

Figure 4.2) however this was not seen across the population treatment (Figure 4.2).

NAR was also lower under low light however there were no significant differences across populations (Table 4.4, Figure 4.2). A greater allocation towards stem mass in the high density treatment was the only indication of shade-avoidance (Figure 4.2).

136

Leaf-Level Traits

Plant size did have an effect on leaf-level traits as indicated by a significant biomass effect following MANCOVA (Table 4.3). There was a significant difference between all leaf-level traits within the population, light and density treatments as well as the population by light treatment (MANCOVA, Table 4.3). Taking plant size into account there were significant differences across all leaf-level traits following pairwise comparisons in the population and light quality treatments however there were no significant differences in the crowding treatment (ANCOVA, Table 4.4). The low quality light treatment induced a shade-avoidance syndrome of longer petiole, longer and wider pinnae, larger leaf area yet lower leaf mass (Figure 4.3). There were no consistent patterns across the population treatment (Table 4.4, Figure 4.3).

Removing the effects of plant size on leaf traits shifted patterns of statistical significance away from the light quality treatment and towards the density treatment

(Table 4.4). Leaves were generally smaller in the high density treatment relative to the low density treatment. Petioles were shorter, pinnae slender, and overall leaf area was smaller in the high density treatment (Figure 4.3).

Discriminant Function Analysis

DFA of the population by light treatments separated the light treatment across the first discriminant function for whole-plant traits (Figure 4.4A). Traits positively associated with the first discriminant function were RMR and NAR whereas stem elongation traits were negatively associated which corresponded to the high light and low light treatments respectively (Figure 4.4D). This same pattern was observed following DFA

137

of the light by density treatments (Figures 4.4C,F). DFA of the population by density treatments produced considerable overlap of confidence intervals however populations were discriminated along the first axis and density treatments across the second axis

(Figure 4.4B).

At the leaf-level, DFA of the population by light treatments separated the light treatment across the second discriminant function (Figure 4.5A). The trait positively associated with the second discriminant function was leaf mass whereas longer and wider leaf traits were negatively associated which corresponded to the high light and low light treatments respectively (Figure 4.5D). The three populations marginally overlapped in confidence intervals for leaf-level traits (Figure 4.5A) and leaf elongation was associated with the high rainfall population whereas leaf widening was associated with the low rainfall population (Figure 4.5D). There was considerable overlap in confidence intervals following DFA for the population by density treatments with no clear association with either the first or second discriminant functions (Figure 4.5B). There was also considerable overlap in confidence intervals for the density treatments following DFA for the light by density treatments (Figure

4.5C). Yet the light treatments were discriminated along the first discriminant function and this correlated with longer and wider leaves for the low light treatment and heavier leaves for the high light treatment (Figure 4.5F).

Growth Traits

Within the population treatment there was a significant difference in RGR (F2,163 =

3.654, p = 0.0280; Figure 4.6A), height growth (F2,163 = 17.114, p < 0.0001; Figure

138

4.6B) but no significant difference in total dry weight at time of harvest (F2,163 =

2.738, p = 0.0677; Figure 4.6C). RGR and total dry weight did not differ between high and low quality light treatments (F1,163 = 3.793, p = 0.0532; F1,163 = 2.100, p = 0.1492;

Figures 4.6A,C) however height growth was significantly faster in low quality light

(F1,163 = 25.875, p < 0.0001; Figure 4.6B). RGR and height growth were significantly faster in the low density treatment (F2,163 = 7.177, p = 0.0010; F2,163 = 4.865, p =

0.0089; Figures 4.6A,B) and total dry weight at time of harvest was also significantly greater (F2,163 = 15.574, p < 0.0001; Figure 4.6C).

DISCUSSION

The onset of heteroblastic development is a trait that has previously been related to the light environment (Burns, 2005; Forster and Bonser, 2009). However heteroblasty has rarely considered a symptom within the shade-avoidance syndrome. The current experiment demonstrated that heteroblastic development is not only delayed in a low quality light environment but it can also be delayed in a crowded environment (Figure

4.1). There are advantages of retaining compound leaves in both situations: compound leaves have greater photosynthetic capture area (per unit of biomass) and perform better under low irradiance than phyllodes (Brodribb and Hill, 1993; Yu and Li,

2007); and compound leaves are cheaper to construct than transitional leaves or phyllodes (in terms of SLA: Yu and Li, 2007; Forster and Bonser, 2009) and are more beneficial in a highly competitive environment. Therefore, in Acacia implexa heteroblastic development may be considered to be part of an overall shade-avoidance

139

strategy that aims to direct growing parts of the plant to a better quality light environment and assist in seedling establishment (Hansen 1996).

However these apparent evolutionary benefits of compound leaves may be purely coincidental, as discussed in Chapter 2. It was demonstrated that compound leaves form on Acacia plants following the application of gibberellic acid (GA) independent of the growth environment. GA is widely regarded as a phytohormone that is produced by a large number of plants when shade is encountered. Therefore, A. implexa may encounter a shade environment and release GA hormones that induce a shade-avoidance phenotype (i.e. stem elongation) and, coincidentally, produces compound leaves. This possibility needs further investigation and the role that other phytohormones may play cannot be discounted.

The reduction of light by competing neighbours or an overhanging canopy is perceived by photoreceptors in different parts of the plant yet both crowding and shading should induce a shade-avoidance syndrome (Schmitt and Wulff, 1993). Here we found that a shade-avoidance syndrome was far more evident in a low quality light environment rather than a crowded environment. Plants in the high density treatment not only competed for light but faced competition for other resources, such as moisture and nutrients (Casper and Jackson, 1997). In spite of plants in the high density treatment allocating a greater proportion of biomass to the stem, the rate of height growth was significantly faster in the low density treatment. Relative growth rates and total plant dry weight at time of harvest were also significantly lower in the higher density treatments suggesting that these plants were limited across a range of

140

critical resources. Plants grown at high and low light levels, on the other hand, had similar growth rates and total plant biomass. Relative height growth was greatly enhanced in the low quality light environment and, coupled with the shade-avoidance response, suggests that light was the most limiting resource to these plants.

Leaf-level traits in the shaded light treatment were longer and wider than the high light treatment consistent with the findings of Barišić et al. (2006). Longer and wider leaves, as well as elongation of the petiole/phyllode, enable increased capture of photosynthetic radiation. This pattern was evident only when the effects of plant size were included in the analysis and suggests that leaf-level trait responses to shade in A. implexa are adaptive (Moriuchi and Winn, 2005). In contrast, there were no differences in any leaf-level traits across the density treatments when plant size was accounted for, a finding at odds with previous experience (Barišić et al., 2006; Bell and Galloway, 2007). Rather, differences only emerged when the effects of plant size were removed. Leaf-level traits observed in this experiment are therefore not adaptive to the light environment induced by the crowding treatments imposed on them.

Above-ground space is limiting in crowded environments that can lead to narrower and longer leaves (Whitelam and Johnson, 1982).

In the genus Acacia the expression of different phenotypes across populations has previously been related to different rainfall regimes (e.g. Farrell and Ashton, 1978;

Cody, 1989; 1991). However this hypothesis has not been tested to any large extent.

Although the current experiment only examined the expression of traits from three populations derived from a single species there is a suggestion that the development of

141

different phenotypes may not be directly related to rainfall (and moisture availability) per se but other environmental variables may be important. Acacia species have excellent water use efficiency when compared to other species (Hansen, 1986; 1996) and Acacia frequently dominates vegetation communities in Australian arid and semi- arid regions (Specht and Specht, 1999). These facts suggest good tolerance to aridity and moisture deficits. Populations within species may therefore be differentiated along an alternate environmental axis and results from this study suggest light availability could be an important variable. For example, the high and medium rainfall populations were discriminated along axes related to traits associated with the low light treatment whereas the low rainfall population was grouped with traits associated with the high light treatment (Figures 4.4, 4.5). Plants from the high and medium rainfall populations expressed traits in a manner comparable to the shade-avoidance syndrome evident from the low light treatment (Figure 4.2). Data of the exact light regime of the three populations used in this experiment was not available and therefore a direct comparison is not possible. However it is reasonable to expect that as rainfall increases canopy cover also increases and light availability decreases.

Dominant vegetation communities from the high and medium rainfall populations were forests and woodlands (McRae and Cooper, 1985; Benson, 1986) whereas open woodlands dominated the region from where the low rainfall population was sourced

(Scott, 1992). Nevertheless, this is anecdotal evidence and a more extensive study is needed to support the hypothesis that populations are differentiated along a light gradient or an alternate gradient as has been suggested for A. koa (Daehler et al.,

1999).

142

Overall this study found general similarities of photomorphogenesis between low light and crowded treatments yet there were many differences. These differences may have arisen because plants in a crowded environment not only have to adjust phenotype to altered light conditions, but are most likely being limited in other critical resources.

The most significant finding of this study however was that heteroblastic development is delayed by low light as well as increased density. The fact that heteroblastic development is regulated by density dependence has not previously been demonstrated. Lastly, there were similarities in the expression of phenotypes between plants from a high rainfall population and low light treatment that suggests populations within species Acacia are differentiated along a light gradient.

Acknowledgements

We thank Geoff McDonnell, Kylie Mallit and Natalie Wynyard for glasshouse assistance and Grant Williamson for leaf area software. Funded by UNSW Faculty Research Grant and Early Career Research grant to S.P.B.

REFERENCES

Atkin OK, Schortemeyer M, McFarlane N, Evans JR. 1998. Variation in the components of relative growth rate in ten Acacia species from contrasting environments. Plant, Cell and Environment, 21:1007-1017.

Ballaré CL. 1999. Keeping up with the neighbours: phytochrome sensing and other signalling mechanisms. Trends in Plant Sciences, 4:97-102.

Ballaré CL, Scopel AL, Sanchez RA. 1990. Far-red radiation reflected from adjacent leaves: an early signal of competition in plant canopies. Science, 247:329-332.

143

Barišić N, Stojković B, Tarasjev A. 2006. Plastic responses to light intensity and planting density in three Lamium species. Plant Systematics and Evolution, 262:25-36.

Bell DL, Galloway LF. 2007. Plasticity to neighbour shade: fitness consequences and allometry. Functional Ecology, 21:1146-1153.

Benson DH. 1986. The vegetation of the Gosford and Lake Macquarie 1:100 000 vegetation map sheet. Cunninghamia, 1:467-490.

Bonser SP, Geber MA. 2005. Growth form evolution and shifting habitat specialization in annual plants. Journal of Evolutionary Biology, 18:1009- 1018.

Brodribb T, Hill RS. 1993. A physiological comparison of leaves and phyllodes in Acacia melanoxylon. Australian Journal of Botany, 41:293-305.

Burns KC. 2005. Plastic heteroblasty in beach groundsel (Senecio lautus). New Zealand Journal of Botany, 43:665-672.

Casper BB, Jackson BB. 1997. Plant competition underground. Annual Review of Ecology and Systematics, 28:545-570.

Cody ML. 1989. Morphological variation in mulga. I. Variation and covariation within and among Acacia aneura populations. Israeli Journal of Botany, 38:241-257.

Cody ML. 1991. Morphological variation in mulga. II. Covariation among morphology, phenology, and spatial patterns in Acacia -dominated vegetation. Israeli Journal of Botany, 40:41-59.

Farrell TP, Ashton DH. 1978. Population studies of Acacia melanoxylon. I. Variation in seed and vegetation characteristics. Australian Journal of Botany, 26:365-379.

Forster MA, Bonser SP. 2009. Heteroblastic development and the optimal partitioning of traits among contrasting environments in Acacia implexa. Annals of Botany, 103:95-105.

Franklin KA, Whitelam GC. 2005. Phytochromes and shade-avoidance responses in plants. Annals of Botany, 96:169-175.

Franklin KA, Whitelam GC. 2007. Red:far-red ratio perception and shade avoidance. In Light and Plant Development, pp. 211-234.

Goebel K. 1898. Organographie der Pflanzen, Jena: Fischer.

144

Hansen DH. 1986. Water relations of compound leaves and phyllodes in Acacia koa var. latifolia. Plant, Cell and Environment, 9:439-445.

Hansen DH. 1996. Establishment and Persistence Characteristics in Juvenile Leaves and Phyllodes of Acacia koa (Leguminosae) in Hawaii. International Journal of Plant Sciences, 157:123-128.

Jones CS. 1999. An essay on juvenility, phase change, and heteroblasty in seed plants. International Journal of Plant Sciences, 160:S105-S111.

Kodela PG. 2002. Acacia. In Flora of New South Wales, pp. 381-476.

Maslin BR, McDonald MW. 2004. AcaciaSearch: Evaluation of Acacia as a woody crop option for southern Australia, Canberra: Rural Industries Research and Development Corportation.

McRae RHD, Cooper MG. 1985. Vegetation of the Merriwa area, New South Wales. Cunninghamia, 1:351-370.

Mitchell PL, Woodward FI. 1988. Responses of three woodland herbs to reduced photosynthetically active radiation and low Red to Far-Red ratio in shade. Journal of Ecology, 76:807-825.

Morgan DC, Smith H. 1981. Control of development in Chenopodium album L. by shadelight: the effect of light quantity (total fluence rate) and light quality (Red:Far-Red ratio). New Phytologist, 88:239-248.

Moriuchi KS, Winn AA. 2005. Relationships among growth, development and plastic response to environment quality in a perennial plant. New Phytologist, 166:149-158.

Navas M, Garnier E. 2002. Plasticity of whole plant and leaf traits in Rubia peregrina in response to light, nutrient and water availability. Acta Oecologica, 23:375-383.

New TR. 1984. A Biology of Acacias, Melbourne: Oxford University Press.

Quinn GP, Keough MJ. 2002. Experimental Design and Data Analysis for Biologists, Port Melbourne: Cambridge University Press.

145

Rice SA, Bazzaz FA. 1989. Quantification of plasticity of plant traits in response to light intensity: comparing phenotypes at a common weight. Oecologia, 78:502-507.

Schmitt J, Wulff RD. 1993. Light spectral quality, phytochrome and plant competition. Trends in Ecology and Evolution, 8:47-51.

Scott JA. 1992. The natural vegetation of Balranald - Swan Hill area. Cunninghamia, 2:597-652.

Smith H. 1982. Light quality, photoreception, and plant strategy. Annual Review of Plant Physiology, 33:481-518.

Smith H, Casal JJ, Jackson GM. 1990. Reflectance signals and the perception by phytochrome of the proximity of neighbouring vegetation. Plant Cell Environment, 13:73-78.

Specht RL, Specht A. 1999. Australian Plant Communitites: Dynamics of structure, growth and biodiversity, South Melbourne: Oxford University Press.

Weiner J. 1990. Asymmetric competition in plant populations. Trends in Ecology and Evolution, 5:360-364.

Weiner J, Wright DB, Castro S. 1997. Symmetry of belowground competition between Holcus lanatua and Dactylis glomera. Oikos, 79:85-91.

Wells CL, Pigliucci M. 2000. Adaptive phenotypic plasticity: the case of heterophylly in aquatic plants. Perspectives in Plant Ecology, Evolution and Systematics, 3:1-18.

Whitelam GC, Johnson CB. 1982. Photomorphogenesis in Impatiens parviflora and other plant species under simulated natural canopy radiation. New Phytologist, 90:611-618.

Withers JR. 1979. Studies on the status of unburnt Eucalyptus woodland at Ocean Grove, Victoria. IV. The effect of shading on seedling establishment Australian Journal of Botany, 27:47-66.

Yu H, Li JT. 2007. Physiological comparisons of true leaves and phyllodes in Acacia mangium seedlings. Photosynthetica, 45:312-316.

Yu H, Ong BL. 2003. Effect of radiation quality on growth and photosynthesis of Acacia mangium seedlings. Photosynthetica, 41:349-355.

146

Table 4.1 Brief description of populations of sourced seeds. Rainfall and temperature data were collated from Bureau of Meteorology, Australia climate data of nearest weather station. Population Seedlot Location Rainfall (mm p.a.) Avg Temp. (°C)

Low Rainfall 19780 35°57′S, 450 23 145°45′E Medium Rainfall 19770 32°37′S, 650 23 150°03′E High Rainfall 18859 31°56′S, 830 24 151°11′E

147

Table 4.2 Summary of traits measured in Chapter 4 with abbreviations and transformations made for statistical analysis. Trait Abbrev. Definition Trans. a) Whole-Plant Traits Height to Diameter HtoD Ratio of stem length to basal stem Sqrt (mm cm-2) diameter Internode Length Internode Stem length divided by the total number Sqrt (mm) of nodes Stem Mass Ratio StMR Ratio of total stem mass to sum of root log (g g-1) and leaf mass Root Mass Ratio RMR Ratio of total root mass to sum of stem log (g g-1) and leaf mass Leaf Mass Ratio LMR Ratio of total leaf mass to sum of stem log (g g-1) and root mass Leaf Area Ratio LAR Ratio of total leaf area to total biomass Sqrt (cm2 g-1) Net Assimilation Rate NAR The amount of carbon fixed per day for a log (g cm-2 day-1) given amount of photosynthetic tissue Specific Leaf Area SLA Ratio of leaf area to leaf mass log (cm2 g-1) b) Leaf-Level Traits Petiole Length Average petiole/phyllode length of last log (mm) compound leaf,1st and 3rd transitional leaves Pinnae Length Average length of largest pinnae from last (mm) compound leaf, 1st and 3rd transitional leaves Pinnae Width Average distance at widest point of largest (mm) pinnae on last compound leaf, 1st and 3rd transitional leaves Area Average area of one-side of leaf including Sqrt (cm2) petiole/phyllode of last compound leaf, 1st and 3rd transitional leaves Mass Average mass of leaf including log (g) petiole/phyllode of last compound leaf, 1st and 3rd transitional leaves c) Growth Traits Relative Growth Rate RGR Amount of biomass fixed per day and (g g-1 day-1) determined by equation 1 Height Growth Amount of height accrued each day and (mm day-1) determined by equation 1 Total Dry Weight Amount of dry weight of all plant material log (g) at time of harvest

148

Table 4.3 Results from multivariate analysis of variance. # Denotes traits were analysed using multivariate analysis of co-variance with biomass as the covariate. Pillai’s F-value d.f. P-value Trace

Whole-Plant Population 0.497 6.249 16, 302 < 0.0001 Light 0.665 37.193 8, 150 < 0.0001 Density 0.308 3.437 16, 302 < 0.0001 P x L 0.140 1.422 16, 302 0.1294 P x D 0.238 1.210 32, 612 0.2003 P x L x D 0.088 0.871 16, 302 0.6034

Leaf-Level # Biomass 0.611 48.297 5, 154 <0.0001 Population 0.488 9.998 10, 310 <0.0001 Light 0.274 11.607 5, 154 <0.0001 Density 0.128 2.124 10, 310 0.0224 P x L 0.130 2.151 10, 310 0.0206 P x D 0.142 1.155 20, 628 0.2888 P x L x D 0.028 0.442 10, 310 0.9251

Leaf-Level Population 0.520 11.320 10, 322 <0.0001 Light 0.215 8.745 5, 160 <0.0001 Density 0.214 3.868 10, 322 0.0001 P x L 0.123 2.116 10, 322 0.0229 P x D 0.111 0.929 20, 652 0.5496 P x L x D 0.037 0.602 10, 322 0.8118

149

Table 4.4 Results from analysis of variance. # Denotes traits were analysed using analysis of co-variance with biomass as the covariate. Significant differences are given by: * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001. Full results are found in [Supplementary materials 4.4-6]. Population Light Density P x L P x D L x D P x L x D

Whole-Plant HtoD 44.280**** 166.570**** 1.528 0.066 2.049 1.154 1.968 Internode 44.134**** 100.767**** 0.010 0.099 2.658* 1.926 0.536 StMR 16.824**** 25.482**** 15.262**** 2.022 0.494 1.461 0.223 RMR 2.808 12.512*** 0.331 0.297 0.819 0.886 0.892 LMR 1.541 0.084 8.131*** 1.295 1.110 1.114 1.262 LAR 6.072** 45.830**** 6.592** 4.457* 0.110 0.744 1.297 NAR 0.982 23.751**** 5.234** 2.364 0.822 1.219 1.393 SLA 5.194** 83.117**** 1.303 3.784* 0.599 0.064 0.701

Leaf-Level # Petiole Length 5.368** 5.224* 2.485 2.068 1.240 0.103 0.750 Pinnae Length 7.164** 4.990* 0.387 2.104 0.120 0.600 0.920 Pinnae Width 10.698**** 5.119* 2.183 4.318* 0.306 0.157 0.661 Leaf Area 10.438*** 5.212* 1.909 2.283 0.349 0.895 0.610 Leaf Mass 7.221** 6.566* 2.220 2.669 1.798 0.489 3.129*

Leaf-Level Petiole Length 5.442** 1.295 6.954** 2.315 0.913 0.037 0.807 Pinnae Length 3.289* 1.207 4.538* 0.996 0.250 0.142 0.619 Pinnae Width 10.689**** 0.625 12.709**** 1.643 1.277 0.599 0.182 Leaf Area 10.915**** 0.483 12.563**** 2.028 0.344 0.870 0.651 Leaf Mass 8.025*** 5.342* 9.536*** 2.454 0.794 0.544 1.514

150

1.1 **** **** ***

1.0

0.9

0.8

log Compound Leaf Nodes Compound log Leaf 0.7 LMH H L LMH Population Light Density

Figure 4.1 Compound leaf nodes represents the total number of nodes that developed a compound leaf before the onset of transitional leaves and corresponds to timing of heteroblastic development. Displayed are means (±S.E.). *** p < 0.001; **** p < 0.0001. Complete statistical results are presented in [Supplementary materials 4.3].

151

13 **** **** 12

10 sqrt HtoD

5 **** ****

3 sqrt Internode sqrt 2

-0.5 **** **** ****

-0.6

-0.7 log StMR

-0.8

-0.32 ***

-0.36

-0.40 log RMR

-0.44 L M H H L L M H

Population Light Density

Figure 4.2 Means (±S.E.) of whole-plant traits across experimental treatments. * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001. Units of measurement are given in Table 4.2. Complete ANOVA results can be found in [Supplementary materials 4.4].

152

0.1 ***

0.0 log LMR log

-0.1

12 ** **** **

11 log LAR

-2.5 **** **

-2.8 log NAR

-3.1

2.5 ** ****

2.4 log SLA

2.3 L M H H L L M H Density Population Light

Figure 4.2 (continued) Means (±S.E.) of whole-plant traits across experimental treatments. * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001. Units of measurement are given in Table 4.2. Complete ANOVA results can be found in [Supplementary materials 4.4].

153

1.7 ** * †† ††

1.6 log Petiole log 1.5 ** * 45 † †

39 Pinn. Length Pinn. 18 **** * †††† ††††

16 Pinn. Width Pinn. 14 *** * 3.8 †††† ††††

3.5

sqrt Area 3.2

-1.2 ** * ††† † ††† -1.3

-1.4 log Mass log

-1.5 L M H H L L M H

Population Light Density

Figure 4.3 Means (±S.E.) of leaf-level traits across experimental treatments. Significant differences following ANCOVA is given by: * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001. Significant differences following ANOVA are given by: † p < 0.05; †† p < 0.01; ††† p < 0.001; †††† p < 0.0001. Units of measurement are given in Table 4.2. Complete ANOVA results can be found in [Supplementary materials 4.5-6].

154

3.0 1.0 (A) (D)

2.0

0.5 Internode MH NAR 1.0 StMR HH HtoD HL 0.0 0.0 ML RMR LH LL DF2 (16.1%) DF2 -1.0 (16.1%) DF2 LMR -0.5 -2.0 SLA LAR

-3.0 -1.0 -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 -1.0 -0.5 0.0 0.5 1.0 DF1 (81.8%) DF1 (81.8%) 2.0 1.0 (B) (E)

0.5 NAR 1.0 StMR LH ) RMR HH MM MH LM 0.0 0.0 HM HtoD SLA LL DF2 (21.1%) DF2 DF2 (21.1%DF2 Internode ML LAR HL -1.0 -0.5 LMR

-2.0 -1.0 -2.0 -1.0 0.0 1.0 2.0 -1.0 -0.5 0.0 0.5 1.0 DF1 (63.2%) DF1 (63.2%) 2.0 1.0 (C) (F) StMR

1.0 0.5 NAR HH HtoD LH LM HM Internode RMR 0.0 0.0 SLA LL DF2 (14.8%) DF2 (14.8%) DF2 LAR -1.0 -0.5 HL LMR

-2.0 -1.0 -2.0 -1.0 0.0 1.0 2.0 -1.0 -0.5 0.0 0.5 1.0 DF1 (81.0%) DF1 (81.0%) Figure 4.4 Grouping of population, light and density treatments following discriminant function analysis for whole-plant traits. Displayed are group centroids with circles indicating 95% confidence intervals for whole plant and leaf-level traits (overlapping circles indicate no difference between groups). Variance explained by first and second discriminant functions displayed parenthetically. (A) HL, ML, LL: high rainfall, medium rainfall, low rainfall population by low quality light treatments; and HH, MH, LH: high rainfall, medium rainfall, low rainfall population by high quality light treatments. (B) HH, MH, LH: high rainfall, medium rainfall, low rainfall population by high density treatments; HM, MM, LM: high rainfall, medium rainfall, low rainfall population by medium density treatments; and HL, ML, LL: high rainfall, medium rainfall, low rainfall population by low density treatments. (C) LH, LM, LL: low quality light by high, medium and low density treatments; and HH, HM, HL: high quality light by high, medium and low density treatments. (D to F) Factor loadings of whole-plant traits corresponding to DFA of treatments across first and second discriminant functions.

155

2.0 1.0 (A) (D)

1.0 0.5 HH Mass MH LH

0.0 0.0 Area HL Pinnae Length LL Pinnae Width DF2 (30.0%) DF2 DF2 (30.0%) DF2 Petiole Length ML -1.0 -0.5

-2.0 -1.0 -2.0 -1.0 0.0 1.0 2.0 -1.0 -0.5 0.0 0.5 1.0 DF1 (51.2%) DF1 (51.2%) 2.0 1.0 (B) (E)

Area 1.0 0.5 Mass

LM HL Pinnae Length LL HH HM Pinnae Width 0.0 0.0 LH MH MM DF2 (17.5%) DF2 DF2 (17.5%) DF2 Petiole Length ML -1.0 -0.5

-2.0 -1.0 -2.0 -1.0 0.0 1.0 2.0 -1.0 -0.5 0.0 0.5 1.0 DF1 (67.4%) DF1 (67.4%) 2.0 1.0 (C) (F) Area Mass Pinnae Width Petiole Length Pinnae Length 1.0 0.5

LL HL HM 0.0 0.0 LM LH HH DF2 (27.7%) DF2 DF2 (27.7%) DF2

-1.0 -0.5

-2.0 -1.0 -2.0 -1.0 0.0 1.0 2.0 -1.0 -0.5 0.0 0.5 1.0 DF1 (62.3%) DF1 (62.3%) Figure 4.5 Grouping of population, light and density treatments following discriminant function analysis for leaf-level traits. Displayed are group centroids with circles indicating 95% confidence intervals for whole plant and leaf-level traits (overlapping circles indicate no difference between groups). Variance explained by first and second discriminant functions displayed parenthetically. (A) HL, ML, LL: high rainfall, medium rainfall, low rainfall population by low quality light treatments; and HH, MH, LH: high rainfall, medium rainfall, low rainfall population by high quality light treatments. (B) HH, MH, LH: high rainfall, medium rainfall, low rainfall population by high density treatments; HM, MM, LM: high rainfall, medium rainfall, low rainfall population by medium density treatments; and HL, ML, LL: high rainfall, medium rainfall, low rainfall population by low density treatments. (C) LH, LM, LL: low quality light by high, medium and low density treatments; and HH, HM, HL: high quality light by high, medium and low density treatments. (D to F) Factor loadings of whole-plant traits corresponding to DFA of treatments across first and second discriminant functions.

156

(A) * **

) 0.040 -1 day -1 0.030 RGR (g g RGR (g 0.020

LMH H L LMH Population Light Density (B) 0.024 **** **** ** ) -1

0.020

0.016

0.012 Height Growth (cm day (cm Growth Height LMH H L LMH Population Light Density (C) 1.200 ****

1.000

0.800

0.600

0.400 Total Plant Dry Weight (g) Weight Dry Plant Total 0.200 LMH H L LMH Population Light Density

Figure 4.6 Mean (±S.E.) of growth traits across experimental treatments. * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001. Complete ANOVA results can be found in [Supplementary materials 4.7].

157

Chapter 5

Global ecology of compound and simple leaves

Authors: Michael A. Forster, Stephen P. Bonser and Ian J. Wright

158

ABSTRACT

Leaves can be categorised into two basic types, compound and simple, where compound leaves differ principally from simple leaves by being divided into a number of smaller parts. The evolutionary and ecological significance of compound versus simple leaves is unclear and under debate. Here we use the world-wide dataset, GLPONET, to compare fundamental leaf traits in compound versus simple leaves. We compared leaf mass per area, leaf life-span, leaf nitrogen, leaf phosphorus, photosynthetic rate, respiration rate and stomatal conductance and found that compound leaves differed on a leaf mass but not area basis. Mean values and multivariate analysis suggested compound leaves occupy the cheap to construct with fast returns region of the leaf economic spectrum however with overlapping confidence intervals with simple leaves. We also analysed whether compound leaves are proportionately better represented in certain biomes and growth forms. We found that deserts, grasslands and tropical forests had more compound leaf species and woodland less species than expected by chance. Herb, tree and vine functional groups had significantly more compound species and shrubs had significantly less. Generally, this study supports previous hypotheses suggesting compound leaf species have faster growth rates and are advantageous in drier environments.

159

INTRODUCTION

Leaves function to capture energy from light and have evolved independently at least six times (Niklas, 1997). The size and shape of leaves varies tremendously among species, and the search for general patterns that link all leaves has motivated a large amount of research. Recent advances strongly suggest that arrayed along a spectrum of “leaf economics” (Wright et al., 2004) running from those capable only of slow physiological rates, with relatively small investments of nutrients per unit mass or area, to leaves that are very physiologically active, with relatively large investments of nutrients per unit mass or area, but with a short return-time on those investments. Leaf trait strategies vary both among co-occurring species, and among species from different environments (Reich et al., 1992; Reich et al., 1997). These leaf-level strategies are thought to influence the rate of uptake and distribution of nutrients within plants, and in turn, help to determine various community and ecosystem properties (Reich et al., 1997)

A key trait underlying the leaf economics spectrum is leaf mass per unit area (LMA) – the ratio of leaf dry mass to projected (one-sided) surface area of the leaf lamina. On average, low LMA leaves have short leaf lifespans, high nitrogen (N) and phosphorus (P) concentrations, and rapid maximum photosynthetic (A) and respiration rates (R), while species with high LMA leaves generally have the opposite set of traits (Reich et al., 1997;

1998; Wright et al., 2004). On average, the lower the fertility or water availability in a site, the further the species occur towards the high LMA end of the spectrum. The efficiency with which leaves fix energy relative to costs has also been related to LMA.

For example, photosynthetic N-use efficiency (PNUE, or Amass/Nmass) is a measure of the

160

amount of gain from photosynthesis per cost of unit N invested in the leaf (Wright et al.,

2005; Niinemets and Sacks, 2006). Increasing LMA lowers PNUE due to differences in structural allocation of N in the leaf, and longer CO2 diffusion paths and lower transmission of light through the leaf leading to lower Amass (Wright et al., 2005).

Overall, LMA is an easily quantifiable trait that summarises a large amount of related leaf physiological characters.

Although our general understanding of leaf trait strategies has improved immensely over recent years, there is still relatively little known about the extent to which compound- and simple-leaved species differ in ecologically-important leaf traits. Simple leaves are not dissected smaller parts (leaflets). A compound leaf, on the other hand, is divided into two or more distinct leaflets with each leaflet being roughly homologous to a simple leaf

(Figure 5.1; Givnish, 1984). Allocation of biomass into many small leaflets, rather than a single large leaf blade, is thought to be advantageous by isolating herbivory (particularly insect herbivory) or damage to leaflets rather than the entire leaf blade (Niinemets, 2006).

Compound leaves are also thought to be analogous to branches found on plants with simple leaves albeit metabolically cheaper to construct (Givnish, 1978). This advantage allows for less overall branching and slender trunks for compound leaf species (White,

1983) which in turn allows for rapid height growth both in early successional pioneer environments (Givnish, 1978; 1984) and in reaching the overstorey when gaps form in mature forests (Niinemets, 1998). Although little is known about possible differences between compound and simple leaves in leaf traits we can nevertheless hypothesise that

161

compound leaves may occupy the cheap to construct end of the leaf economics spectrum based on the idea of rapid height growth.

The highly dissected nature of compound leaves is also thought to be an advantage in the water economy of leaves. Heat dissipation is greater for highly lobed or dissected leaves versus entire leaves even at low wind speeds (Vogel, 1970). Greater heat loss leads to lower leaf temperatures when compared to simple leaves and therefore overall transpiration losses are lower (Niinemets, 1998). Consequently, it has been predicted that compound leaf species should be proportionally better represented in hotter, drier habitats

(MacArthur 1972; Givnish and Vermeij, 1976; Stowe and Brown, 1981). Furthermore, the cheap construction of compound leaves as “throw-away branches” allows for frequent discarding of leaf tissue (i.e. deciduousness) and further control of transpiration loss in arid and semi-arid environments (Givnish, 1978).

Here we use the GLOPNET dataset to assess the physiology of compound and simple leaves. GLOPNET contains leaf-trait data from over 2,500 species representing every vegetated continent, biome and plant functional type (Wright et al., 2004). Specifically we test the following hypotheses: 1) compound leaves will occupy the cheap to construct with fast returns end of the leaf economics spectrum; 2) compound leaf species will have greater representation in functional groups with slender trunks (i.e. trees) than complex trunks (i.e. shrubs) in accordance with the idea of rapid height growth; and 3) compound leaves are better adapted to hotter and drier environments.

162

METHODS

Species Traits

Species traits were obtained from the GLOPNET dataset. This dataset was compiled by

Wright et al. (2004) from 175 sites and 2548 species-site combinations. Full details of how the data were collected and traits measured can be obtained from Wright et al.

(2004). From the original dataset we chose the following traits: leaf mass per area

(LMA), leaf lifespan (LL), leaf nitrogen content (N), leaf phosphorus content (P), photosynthetic capacity (A), dark respiration rate (R) and stomatal conductance (gs). N,

P, A and R were expressed on both a leaf mass and leaf area basis. For compound leaf species these traits were measured on leaflets. We also took the ratio of N to A on a leaf mass basis in order to obtain PNUE. The published dataset did not include whether a species had compound or simple leaves therefore we present an extended version of

GLOPNET that includes species categorised as those with compound or simple leaves.

We limit our analysis to dicotyledons only. That is, monocots, conifers and lower taxonomic plants (e.g. ferns) were excluded from the dataset. The modified GLPONET dataset had 1874 samples for analysis of which 257 (13.7%) were compound (C) and

1617 (86.3%) were simple (S) leaf samples. GLOPNET also contains information on global biomes (based on Whittaker, 1975) and growth forms (herb, shrub, tree and vine) which were used to assess whether compound leaves differ in physiology in certain biomes or growth forms. Additionally, species richness of compound and simple leaf species in these biomes and growth forms was also examined.

163

Data Analysis

The proportion (i.e. species richness) of C and S species within individual biomes and growth forms was compared using contingency tables via χ2 testing.

To test whether compound and simple leaves occupy different parts of the leaf economics spectrum two approaches were used. Firstly, a series of one-way ANOVAs were performed with the following leaf traits as dependent variables: LMA, LL, Nmass, Narea,

Pmass, Parea, Amass, Aarea, Rmass, Rarea, gs and PNUE. These tests were run for the entire dataset, as well as for C and S species groups bested within biome, or within growth form. The Welch statistic, followed by Games-Howell post-hoc test, was used in order to manage statistical bias produced by unequal sample sizes (Sokal and Roff, 1995).

Secondly, a multivariate approached was used examining the following key traits associated with the leaf economics spectrum: LMA, LL, Nmass, and Amass. The first axis extracted from principal components analysis is the leaf economics spectrum and compound and simple leaves were compared along this axis in terms of their Anderson-

Rubin factor scores. Non-overlapping confidence intervals indicate that compound and simple leaves are biologically different whereas overlapping confidence intervals indicate a sharing of the leaf economics spectrum.

All statistical analyses were performed using SPSS 15.0.1.

164

RESULTS

Compound leaflets adopt a lower portion of the leaf economics spectrum with considerable overlap with simple leaves. Mean values of LMA were significantly lower in compound over simple leaves however there was no significant difference in LL

(Figure 5.2). Other traits associated with the leaf economics spectrum only showed a significant difference on a leaf mass-basis and there were no significant differences on a leaf area basis. Mean values of Nmass, Pmass and Amass were all significantly greater for compound rather than simple leaves however there was no difference in Rmass (Figure

5.2). Stomatal conductance to water (gs) was also significantly greater for compound leaves (Figure 5.2). LMA and LL were negatively separated and Nmass and Amass were positively separated along the first principal component following PCA, which explained

79.2% of the variance. A-R Factor Score for the first principal component indicated a higher mean value for compound leaf species (Figure 5.3). Compound leaflets had higher

A-R factor scores along the principal component axis than simple leaves (F1,36.8 = 10.814, p = 0.002); that is, they tended to occur further towards the “quick-return” end of the leaf economics spectrum. Still, just as for comparisons of individual traits, the two groups overlapped considerably along this multi-trait axis. There was no significant difference in mean values of PNUE between compound and simple leaves (FWelch1,193.2 = 0.004, p =

0.947) nor in the slope of Amass on Nmass (T87.4 = 0.316, p = 0.110; Figure 5.4).

Within growth forms, there was not a universal difference between compound and simple leaves along the leaf economics spectrum. Compound leaf shrubs and trees did have lower LMA yet there were no significant differences in herbs and vines (Figure 5.5).

165

There were no significant differences in LL, Amass and Rmass, and only compound leaf trees had significantly greater Pmass (Figure 5.5). Nevertheless, herbs, shrubs and trees with compound leaves had significantly greater Nmass (Figure 5.5). All statistical results can be found in [Supplementary materials 5.1].

Sample size in GLOPNET only allowed analysis for four traits across biomes: LMA, LL,

Nmass and Amass. Temperate forest biome showed the most consistent trend with the leaf economics spectrum by having compound leaves with lower LMA and LL, and higher

Nmass and Amass (albeit marginally non-significant: Figure 5.6). Tundra and woodland biomes were also consistent with the leaf economics spectrum by having lower LMA and higher Nmass (Figure 6). Grassland also had significantly greater mean values of Nmass however all other biomes showed no significant differences. All statistical results can be found in [Supplementary materials 5.2].

There was a significant difference among growth forms in the representation of compound leaf species versus simple leaf species. Among herbs, trees and vines there were significantly more compound leaf species than expected by chance alone, whereas among shrubby species there were significantly less compound species (Table 5.1).

Across the majority of biomes compound leaf species were not significantly over- or under- represented. Desert, grassland and tropical rainforest did have a significantly higher representation of compound leaf species than expected by chance (Table 5.2). The

166

woodland biome was the only habitat that had significantly less compound leaf species than expected by chance (Table 5.2).

DISCUSSION

The leaflets of compound leaves are not fundamentally physiologically different from simple leaves, supporting the early observation of Parkhurst and Loucks (1972). Rather compound leaves are offset from simple leaves towards the cheap to construct with high returns end of the leaf economics spectrum. Simple leaves occupy a wide breadth of the leaf economics spectrum and compound leaves fall within this breadth. Although on average compound leaves are cheap to construct there does exist simple leaves that are even cheaper to construct. Therefore compound leaves do not have a clear evolutionary and ecological advantage over simple leaves in terms of leaf physiology.

The integration of leaf-level and whole-plant trait strategies may assist explaining the occurrence of compound leaves towards the lower end of the leaf economics spectrum.

There is a general trade-off between support costs and the allocation of leaf biomass where leaf biomass located closer to the main stem or axis has less support costs

(Niinemets et al., 2006; Niinemets and Sack, 2006). For compound leaf species the addition of leaflets to the rachis increases the proportional investment of biomass allocated to support costs as leaf biomass is held at a greater distance from the main axis

(Niinemets et al., 2006; Niinemets and Sack, 2006). Leaflets with lower LMA may facilitate this process by minimising investment in leaf tissue that can then be redirected

167

towards greater support costs. In turn, LMA is a nexus for functionally coordinated traits

(e.g. Nmass and Narea are both related to each other through LMA; Wright et al., 2004;

Niinemets and Sack, 2006) and therefore decreasing LMA shifts these coordinated traits towards the lower end of the leaf economics spectrum.

That there are no clear advantages or disadvantages in the physiology of leaflets over leaves is additionally supported by no significant difference in physiological efficiencies.

Mean values of PNUE was the same for both compound and simple leaves and there were no differences within growth forms or across biomes. A higher PNUE may be expected for fast-growing species (Poorter et al., 1990), such as the rapid height growth alleged for compound leaf species. Yet similar PNUE observed in this study indicates no fundamental difference in leaf architecture between leaflets and leaves (Hikosaka et al.,

1998). Moreover, differences in the slope of the linear regression between Amass and Nmass would indicate inherent differences in the structural partitioning of N versus the efficiency of leaf internal diffusion of CO2 (Wright and Westoby, 2002). The slope of

Amass versus Nmass in compound and simple leaves are similar (Figure 5.4) further supporting comparable physiology for these two types of leaves.

Compound leaf species were relatively less represented within the shrub growth form but over represented in the tree, vine and herb growth forms (Table 5.1). Compound leaves tend to be found on species with a monoaxial growth form (King, 1998). That is, species with a straight trunk displayed compound leaves where the petiole/rachis acts equivalently to a branch (Aiba and Nakashizuka, 2007). Shrub species tend to be highly

168

branched and there is possibly less of a need to further branching by displaying compound leaves. The cause that determines these patterns is not to be found within leaf physiology as there was no consistent pattern between richness and leaf functional traits.

For example, LMA was lower and Nmass higher for compound leaf shrubs however this did not translate into a greater proportion of species within this growth form. Rather some other trait that is not related directly to the leaf economics spectrum, for instance allocation to stem support costs, may be a more important cause determining compound leafiness across different growth forms.

There was some support for better representation of compound leaf species in hotter, drier climates (i.e. desert) as contended by previous authors (Givnish, 1978; Stowe and

Brown, 1981) but we also found greater representation in tropical rainforests and grasslands and less in woodlands. The increased dissection of compound leaves leads to an increase dissipation of heat loads and lower levels of transpiration (Vogel, 1970;

Stowe and Brown, 1981; Niinemets and Sack, 2006) and therefore there is an advantage over simple leaves in hot and dry habitats. In tropical rainforests and grasslands the notion of rapid height growth may have increased the proportion of compound leaf species relative to other biomes. In spite of these patterns, there is no consistent relationship between leaf physiological traits and richness of compound leaf species within particular biomes. For example, LMA is much lower for compound leaves in the woodland biome however compound leaf species are significantly under-represented within this biome.

169

Conclusion

The primary finding of this study was the leaflets of compound leaves occupy the cheap to construct with fast returns region of the leaf economics spectrum but there was still substantial overlap with simple leaves. Based on leaf physiology there was no clear evolutionary or ecological advantage having compound over simple leaves but there may be advantages in certain functional groups, particularly with the tree growth form over shrubs. There was some indication that compound leaves are more prominent in hotter and drier habitats but this finding was not equivocal. The integration of leaf physiological traits with whole-plant traits may provide a greater insight into the significance of compound leaves.

Acknowledgements

We thank Peter Reich for use of the GLOPNET dataset and Marion Liberloo for assistance with figures.

REFERENCES

Aiba M, Nakashizuka T. 2007. Differences in the dry-mass cost of sapling vertical growth among 56 woody species co-occurring in a Bornean tropical rain forest. Functional Ecology, 21:41-49.

Balderrama SIV, Chazdon RL. 2005. Light-dependent seedling survival and growth of four tree species in Costa Rican second-growth rain forests. Journal of Tropical Ecology, 21:383-395.

Givnish TJ. 1978. On the adaptive significance of compound leaves, with particular reference to tropical trees. In Tropical Trees as Living Systems, pp. 351-380.

170

Givnish TJ. 1984. Leaf and canopy adaptations in tropical forests. In Physiological ecology of plants of the wet tropics. Proceedings of an International Symposium held in Oxatepec and Los Tuxtlas, Mexico, June 29 to July 6, 1983, pp. 51-84.

Givnish TJ. 1988. Adaptation to sun and shade: a whole-plant perspective. Australian Journal of Plant Physiology, 15:63-92.

Gutschick VP. 1999. Biotic and abiotic consequences of differences in leaf structure. New Phytologist, 143:3-18.

Hikosaka K, Hanba YT, Hirose T, Terashima I. 1998. Photosynthetic nitrogen-use efficiency in leaves of woody and herbaceous species. Functional Ecology, 12:896-905.

King DA. 1998. Influence of leaf size on tree architecture: first branch height and crown dimensions in tropical rain forest trees. Trees - Structure and Function, 12:438- 445.

LoGullo MA, Nardini A, Trifilo P, Salleo S. 2003. Changes in leaf hydraulics and stomatal conductance following drought stress and irrigation in Ceratonia siliqua (Carob tree). Physiologia Plantarum, 117:186-194.

MacArthur, R. 1972. Geographical Ecology: patterns in the distribution of species. Harper & Row, New York.

Niinemets U, Portsmuth A, Tobias M. 2006. Leaf size modifies support biomass distribution among stems, petioles and mid-ribs in temperate plants. New Phytologist, 171:91-104.

Niinemets U, Sack L. 2006. Structural determinants of leaf light-harvesting capacity and photosynthetic potentials. Progress in Botany, 67:385-419.

Parkhurst DF, Loucks OL. 1972. Optimal leaf size in relation to the environment. Journal of Ecology, 60:505-537.

Poorter H, Remkes C, Lambers H. 1990. Carbon and nitrogen economy of 24 wild species differing in relative growth rate. Plant Physiology, 94:621-627.

Reich PB, Walters MB, Ellsworth DS. 1997. From tropics to tundra: global convergence in plant functioning. Proceedings of the National Academy of Sciences, USA, 94:13730-13734.

Shipley B. 1995. Structered interspecific determinates of specific leaf area in 34 species of herbaceous angiosperms. Functional Ecology, 9:312-319.

171

Stowe LG, Brown JL. 1981. A geographic perspective on the ecology of compound leaves. Evolution, 35:818-821.

Vogel S. 1970. Convective cooling at low airspeeds and the shapes of broad leaves. Journal of Experimental Botany, 21:91-101.

Whittaker R. 1975. Communities and Ecosystems, New York: MacMillan.

Witkowski ETF, Lamont BB. 1991. Leaf specific mass confounds leaf density and thickness. Oecologia, 88:486-493.

Wright IJ, Reich PB, Cornelissen JHC, Falster DS, Garnier E, Hikosaka K, Lamont BB, Lee W, Oleksyn J, Osada N, Poorter H, Villar R, Warton DI, Westoby M. 2005. Assessing the generality of global leaf trait relationships. New Phytologist, 166:485-496.

Wright IJ, Reich PB, Westoby M, Ackerly DD, Baruch Z, Bongers F, Cavender- Bares J, Chapin T, Cornelissen JHC, Diemer M, Flexas J, Garnier E, Groom PK, Gulias J, Hikosaka K, Lamont BB, Lee T, Lee W, Lusk C, Midgley JJ, Navas ML, Niinemets U, Oleksyn J, Osada N, Poorter H, Poot P, Prior L, Pyankov VI, Roumet C, Thomas SC, Tjoelker MG, Veneklaas EJ, Villar R. 2004. The worldwide leaf economics spectrum. Nature, 428:821-827.

172

Table 5.1 Distribution of compound and simple leaf species according to growth form.

Numbers indicate observed frequency and values in brackets indicate expected frequency.

* 0.1 < p < 0.05; ** 0.05 < p < 0.01; *** p < 0.01.

Compound Simple

Herb 37 (22.9) *** 130 (144.1) *** Shrub 53 (94.8) *** 638 (596.2) *** Tree 162 (137.4) *** 840 (864.6) *** Vine 5 (1.9) ** 9 (12.1) **

173

Table 5.2 Distribution of compound leaf and simple leaf species according to biome

(based on Whittaker, 1975). Numbers indicate observed frequency and values in brackets indicate expected frequency. * 0.1 < p < 0.05; ** 0.05 < p < 0.01; *** p < 0.01.

Compound Simple

Boreal 7 (3.9) 21 (24.1) Desert 8 (3.2) *** 15 (19.8) *** Grassland 21 (11.7) *** 64 (73.3) *** Temperate Forest 69 (73.7) 465 (460.3) Temperate Rainforest 2 (4.7) 32 (29.3) Tropical Forest 16 (19.7) 127 (123.3) Tropical Rainforest 52 (36.4) *** 212 (227.6) *** Tundra 10 (15.0) 99 (94.0) Woodland 72 (88.7) ** 571 (554.3) **

174

Figure 5.1 (A) Compound and (B) simple leaf showing equivalent structures. A leaflet of a compound leaf is analogous to a simple leaf.

175

0 3 2 2 **** **** -1 2 1 1 area -2 1 mass LL N N LMA 0 0 -3 0

Welch1,342 = 111.470; Welch1,89 = 0.312; Welch1,314 = 224.965; Welch1,225 = 1.044; -23 -38 -4 p = 9.63x10 -1 p = 0.578 -1 p = 1.01x10 -1 p = 0.308 1 1 4 3 **** ** 0 0 3 2 area

-1 area -1 2 1 mass mass P A P A -2 -2 1 0

Welch1,110 = 72.100; Welch1,88 = 2.747; Welch1,55 = 4.176; Welch1,73 = 0.802; -13 -3 p = 1.09x10 -3 p = 0.101 0 p = 0.046 -1 p = 0.373 3 2 4 ** 2 3 1 s

1 area 2 mass g R R 0 0 1

Welch1,19 = 0.320; Welch1,18 = 0.331; Welch1,49 = 5.440; p = 0.578 p = 0.572 p = 0.024 -1 -1 0

Figure 5.2 Box plots of compound (grey fill) and simple (white fill) of traits associated with the leaf economics spectrum. All values are presented on a log-scale. LMA: leaf mass per area; LL: leaf lifespan; N: nitrogen; P: phosphorus; A: photosynthetic rate; R: dark respiration rate; gs: stomatal conductance. ** p < 0.05; **** p < 0.001.

176

A-R Factor ScoreA-R -3 -2 -1 0 1 2 3 4

Compound Simple

Figure 5.3 Box plot summary of Anderson-Rubin factor score derived from principal components analysis of four key leaf economics traits (LMA, LL, Nmass, and Amass).

177

40 3.0 (a) (b) 2.5 20

2.0 0 mass 1.5 A PNUE -20 1.0

-40 0.5 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 N Compound Simple mass

178

1 3 **** **** 0 2 -1 1

-2 LL LMA 0 -3

-4 -1 2 1 **** **** **** **** 0 1

-1 mass mass P N 0 -2

-1 -3 4 3

3 2

2 mass mass R A 1 1

0 Herb Shrub Tree Vine0 Herb Shrub Tree Vine

Figure 5.5 Box plots of compound (grey fill) and simple (white fill) of mass-based traits associated with the leaf economics spectrum within growth forms. All values are presented on a log-scale. LMA: leaf mass per area; LL: leaf lifespan; N: nitrogen; P: phosphorus; A: photosynthetic rate; R: dark respiration rate. **** p < 0.0001. Complete ANOVA tables can be found in [Supplementary materials 5.1].

179

-0.5 *** **** **** *** 2.5

-1.5 1.5 LL

LMA -2.5 0.5

-3.5 -0.5 l st a d a t d e est st st dr n d st st st d re r an or r re re n la an e re re an o ese F fo o fo u d r o o l B D assl in F n T o ssl Fo F nf od r p a p ai o ra op ai o G em R ro R W G p r R W T T p em T p mp ro T ro Te T T

1.5 3.5 *** **** ** ****

2.5 0.5 mass mass 1.5 N A

-0.5 0.5

al d st st st st ra d d st d re n re re e e d an n e est n o sla o or or n l la or or la B s Fo nf F nf u od ss F f od ra p ai p ai T ra p in o G m R ro R Wo G m a W e p T p e R T m ro T op Te T Tr

Figure 5.6 Box plots of compound (grey fill) and simple (white fill) of mass-based traits associated with the leaf economics spectrum within biomes. All values are presented on a log-scale. LMA: leaf mass per area; LL: leaf lifespan; N: nitrogen; A: photosynthetic rate. ** p < 0.05; *** p < 0.001; **** p < 0.001. Complete ANOVA tables can be found in [Supplementary materials 5.2].

180

181

DISCUSSION

The first aim of this thesis was to quantify under which environmental condition heteroblasty could be hastened or delayed in Acacia implexa. There was a significant delay in timing of heteroblastic development under a low irradiance environment and there was no significant effect under nutrient or water treatments (Chapter 1). More specifically, heteroblasty was being driven by a lower Red:Far-red (R:FR) light ratio and not by process in the Blue light (Chapter 2) or ultraviolet light region (Chapter 3).

Heteroblasty can also be delayed in a highly competitive environment (Chapter 4) where the alteration of the light environment by neighbours is delaying the onset of transitional leaves. Previously it has been suggested that the light environment does influence the retention of compound leaves on Acacia (Withers, 1979) but to our knowledge this is the first instance this hypothesis has been supported quantitatively.

The second aim was to establish whether heteroblasty occurred in isolation to optimal allocation of biomass to organs where external resources were most limiting. As it expected from previous studies (e.g. Tilman, 1988) there was a greater allocation of biomass to above-ground organs under a low irradiance environment (Chapter 1). This pattern of carbon allocation within the plant was most apparent when plants were grown under a low R:FR light environment (Chapter 2). In addition, despite the nutrient and water treatments having no significant effect on heteroblasty plants still developed phenotypes according to optimal allocation theory. That is, under low nutrients and low water there was a greater allocation of biomass to below-ground organs (Chapter 1). It

182

was also hypothesised that plants grown under a highly competitive environment would allocate biomass to above-ground organs in order to grow taller than the neighbours

(Chapter 4). This was not the case though as plants diverted resources below-ground in order to compete for mineral and water nutrition. Overall, A. implexa showed strong patterns of optimal allocation irrespective of timing of heteroblastic development.

The delay in heteroblasty was typically in the order of one or two nodes. That is, under high irradiance transitional leaves would occur at approximately node 6 whereas under low irradiance they would occur at node 7 or 8. In a perennial species that can grow to a height of 12m, is such a small delay in the shift from one leaf type to another inconsequential to the overall life-history of the plant? Hansen (1996) suggested that juvenile leaves in Acacia assist in the establishment of the plant. Higher photosynthetic rate of juvenile leaves enables rapid height growth with minimum investment in biomass

(Brodribb and Hill, 1993). In this thesis it was found that total dry weight at time of harvest was similar in shade- and sun-grown plants (Chapter 1) suggesting juvenile leaves may aid in seedling establishment. However when assessed on the basis of relative growth rate (RGR), a measure of plant performance, there were mixed results. In spite of the delay in heteroblastic development under low R:FR this did not translate into similar plant performance. RGR was found to be significantly greater under high R:FR light

(Chapter 2). Additionally, plants grown under low density treatment had far superior total dry weight at time of harvest and RGR than plants grown under a high density treatment

(Chapter 4). This was despite a significant delay in heteroblasty in the high density

183

treatment. Therefore there is no clear suggestion that a delay in heteroblasty in A. implexa leads to direct assistance in the establishment of the juvenile plant.

The fourth aim of this thesis was to determine genetic variability of heteroblasty and optimal allocation across populations of A. implexa. Three populations were chosen for examination sourced from regions differing in mean annual rainfall. The choice of these populations was motivated by Farrell and Ashton (1978) who found a delay in heteroblasty from regions of low to high rainfall in the species A. melanoxylon. There was not a similar clear distinction in the expression of leaf type in the closely related A. implexa. Rather, the population sourced from the low rainfall region tended to delay the onset of transitional leaves at a later developmental stage than the medium or high rainfall populations (Chapters 2,4). The interaction of population and experimental treatments also had a complex outcome. For example, the low rainfall population significantly delayed heteroblasty in response to low Blue light signals but the medium and high rainfall populations showed no response. Similarly, in response to low Red light signals the low and high rainfall populations significantly delayed heteroblasty but the medium rainfall population showed no response (Chapter 2). Therefore there was a highly variable expression of heteroblasty across populations of A. implexa.

On the other hand, there was a more easily interpretable optimal allocation of biomass across populations. It was hypothesised that plants sourced from the high rainfall region would have an inherent developmental pattern of greater allocation to above-ground organs when compared to the low rainfall population. This hypothesis was based on the

184

fact that in the high rainfall region below-ground resources would be less limiting than above-ground resource (i.e. light). In contrast, plants from the low rainfall region would be more limited by below-ground resources (particularly moisture) as there were abundant above-ground resources in that region. This premise was supported by the fact that vegetation communities in the high rainfall region are predominately forests (i.e. canopy cover between 30-70%; Specht and Specht, 1999) whereas the low rainfall region are dominated by open woodland (canopy cover <10%; Specht and Specht, 1999). A. implexa would inhabit the understorey in the high rainfall region whereas it would potentially form a canopy in the low rainfall region. In this thesis it was observed that the high rainfall population did allocate a greater proportion of biomass to above-ground organs (Chapters 2-4), supporting the optimal allocation of resources across populations.

Finally, the fifth chapter examined key leaf physiological traits from a global dataset

(GLOPNET) in order to assess differences between compound leaflets and simple leaves.

Compound leaflets inhabit the low construction with fast returns portion of the leaf economics spectrum but with considerable overlap with simple leaves. That is, mean values of leaf mass per area (LMA) were significantly lower for compound leaflets however there was an overlap in confidence intervals. This global finding is similar to within shoot expression of leaf type in Acacia. That is, compound leaflets of Acacia have lower LMA, higher nitrogen and phosphorus allocation per unit leaf mass, faster photosynthetic and respiration rates, and higher stomatal conductance than phyllodes (the simple leaf analogue; Walters and Batholomew, 1984; Hansen, 1986; Hansen and Steig,

1993).

185