Climate Change and Reproductive Phenology: Context-

Dependent Responses to Increases in Temperature and

Implications for Assisted Colonization

by

Susana Wadgymar

A thesis submitted in conformity with the requirements for the degree of Doctorate of Philosophy

Graduate Department of Ecology and Evolutionary Biology University of Toronto

©Copyright by Susana Wadgymar 2015

 

ii    

Climate change and reproductive phenology: context-dependent

responses to temperature and implications for assisted colonization

Susana Wadgymar

Doctorate of Philosophy

Department of Ecology and Evolutionary Biology

University of Toronto

2015

Abstract

Contemporary changes in climate have rapidly increased temperatures worldwide,

extending the length of the growing season and eliciting large shifts in reproductive and growth

traits across a diversity of plant taxa. The role of phenotypic plasticity in alleviating immediate

changes in selection pressures must be thoroughly explored in order to identify the circumstances

under which the survival of particular species may require active management. The major goals

of my thesis were to characterize the contexts in which responses to warming occur and are

adaptive, and to provide insight on the feasibility of assisted colonization (the movement of

species beyond their current range boundary to climatically favorable habitat) and assisted gene

flow (the relocation of multiple, genetically distinct populations to facilitate local adaptation).

Focusing on the annual legume, Chamaecrista fasciculata, I applied artificial warming to

simple plant communities to mimic the thermal regimes expected by the mid-21st century.

Among experiments, I manipulated aspects of the abiotic and biotic environment likely to

contribute to variation in plastic responses to warming, including plant genotype, community

 

iii    

diversity, population density, internal patterns of resource allocation, and the frequency of

rainfall.

Reproductive phenological traits varied in their degree of response to warming, and

photoperiodic constraints prevented optimal responses in populations of C. fasciculata from

lower latitudes. In all cases, temperature-induced phenotypic plasticity was adaptive or neutral,

but only sufficiently alleviated selection pressures in particular situations. Variation in

competitive dynamics, pollinator access, and rainfall frequency did not modify responses to

changes in temperature.

This work identified barriers to assisted colonization across latitudes that arise when

reproductive phenology is dependent on photoperiodic cues. Phenotypic plasticity may

ameliorate some of the negative effects of increases in temperature, but persistent, directional

selection pressures will require the evolution of life history traits for adaptation to climate

change.

 

iv    

Acknowledgements The completion of this thesis would not have been possible without the encouragement

and support of many people. My time at the University of Toronto was enriched by the

exemplary, supportive, and diverse academic community within the Department of Ecology and

Evolutionary Biology, and although I only list a few by name below, all members of the EEB

department contributed to my inspiration and success.

I am forever indebted to my advisor and mentor, Dr. Arthur E. Weis, for his continual

support, reassurance, and advice. Art consistently challenged me to think bigger and bolder, and

has always championed for my success louder than any other. I thank my committee members,

John Stinchcombe and Stephen Wright, for their valuable guidance during the development of

my thesis, and I am grateful to Shannon McCauley (University of Toronto Mississauga) and

Hugh Henry (Western University) for agreeing to serve as my internal and external examiners,

respectively. Additionally, I appreciate Benjamin Gilbert for discussions on statistical

methodology, Helen Rodd for her perpetual understanding and encouragement, particularly in

the final year of my PhD, and Jonathan Gammal for assistance with installing and programming

software at KSR.

I would like to thank my lab mate and best friend, Emily Austen, whose intellectual

contributions, mutual appreciation of the ridiculous, and unwavering confidence in me made my

thesis possible. I am grateful to Matthew Cumming, a good friend and coauthor on the three

main chapters of this thesis, for acting as a sounding board for my ideas.

I have had the pleasure of acquiring many friends who supported me with scientific

discussion, intellectual stimulation, and Friday beers, including Alison Parker, Amanda Gorton,

Anna Simonson, Arvid Ågren, Bergita Petro, Brandon Campitelli, Brechann McGoey, Cheryl

Partridge, David Punzalan, Donna Hopkins, Emily Josephs, Geoff Legault, Florain Busch,

Heather Coiner, Jane Ogilvie, Jean Mitchell, Jennifer Ison, Joanna Bundus, Kyle Turner, Lesley

Campbell, Nathaniel Sharp, Patrick Friesen, Patrick Vogan, Penelope Gorton, Robert

Williamson, Stephen DeLisle, and Young Wha Lee.

And lastly, I thank my husband, Ryan Corcoran, and my good friend, Gil Martinez, for

their emotional and mental support throughout my studies, and my parents and brother, who are

deeply proud of my accomplishments and who always encouraged me to pursue my dreams.

 

v    

Table of Contents

Abstract………………………………………………..………………………………………….ii

Acknowledgements……………………...………………………………………………………..iv

List of Contents……………………………………...…………………………………………….v

List of Tables……………………………………...……...………………………………………ix

List of Figures………………………………………..……………………………………………x

List of Appendices………………………………………………………….……………………xii

CHAPTER ONE: GENERAL INTRODUCTION……….………………………………………1

Phenology as a climatic indicator………...……………………………………………….2

Phenology and assisted colonization…………...………………………..………………..3

Independent vs. correlated responses to warming……………….………...……………...5

Competitive dynamics and phenological responses to warming…..……………………...6

Research objectives and thesis outline……...…………………………………………….7

References cited……………………………………………………………………….....10

CHAPTER TWO: THE SUCCESS OF ASSISTED COLONIZATION AND ASSISTED GENE

FLOW DEPENDS ON PHENOLOGY………………………………………………..………...17

Abstract…………………………………………………………………………………..17

Introduction………………………………………………………………………………18

Methods…………………………………………………………………………………..21

Study species……………………………………………………………………..21

Experimental design….…………………………………………………………..22

Statistical analyses……………………………………………………………….24

 

vi    

Phenotypic selection analysis……………………………………………………25

Temporal reproductive isolation…………………………………………………26

Results……………………………………………………………………………………27

Thermal environment…………………………………………………………….27

Population differences and responses to warming………………………………27

Phenotypic selection analysis……………………………………………………29

Temporal reproductive isolation…………………………………………………29

Discussion………………………………………………………………………………..30

Phenotypic selection and responses to warming………………………………...31

Considerations for assisted colonization and assisted gene flow………………..32

Tables and Figures…………………………………………………………………….....35

References Cited..………………………………………………………………………..45

CHAPTER THREE: SIMULTANEOUS PULSED FLOWERING IN A TEMPERATE

LEGUME: CAUSES AND CONSEQUENCES OF MULTIMODALITY IN THE SHAPE OF

FLORAL DISPLAY SCHEDULES……………..………….…………………………………...54

Abstract..…………………………………………………………………………………54

Introduction…………………………………..…………………………………………..55

Methods…………………………………………………………………………………..58

Study species…………………..…………………………………………………...58

Summary of experiments…………………………………………………………58

Comparisons of flowering phenologies among populations……………………..61

Relationships between flowering phenology and environmental variables……...62

Synchrony and phenological assortative mating………………………………...63

Results……………………………………………………………………………………66

 

vii    

Comparison of flowering phenologies among populations……………………...66

Relationships between flowering phenology and environmental variables……...67

Synchrony and phenological assortative mating………………………………...69

Discussion………………………………………………………………………………..70

Multimodality and display schedule shape………………………………………70

Causes of variation in display and deployment schedules……………………….71

Population and individual synchrony……………………………………………72

Phenological assortative mating, natural selection, and schedule shape……….73

Tables and Figures……………………………………………………………………….74

References cited...………………………………………………………………………..82

Appendix A………………………………………………………………………………87

CHAPTER FOUR: THE INFLUENCE OF COMPETITION ON PHENOLOGICAL

RESPONSES TO WARMING………………………….……………………………………...101

Abstract…………………………………………………………………………………101

Introduction……………………………………………………………………………..102

Methods…………………………………………………………………………………105

Study organisms………………………………………………………………...105

Experimental design…………………………………………………………….105

Statistical analyses……………………………………………………………...106

Selection analyses………………………………………………………………108

Results…………………………………………………………………………………..109

Treatment differences…………………………………………………………...109

Phenotypic responses to warming……………………………………………...109

 

viii    

Phenotypic responses to competitive dynamics………………..……………….110

Modified responses to warming and subsequent patterns of selection…………111

Discussion………………………………………………………………………………112

Variation in phenological responses to warming………………………………113

Competition and phenology…………………………………………………….113

Summary………………………………………………………………………..115

Tables and Figures……………………………………………………………………...116

References cited...………………………………………………………………………123

CHAPTER FIVE: CONCLUDING DISCUSSION…….……………………………………..128

Phenotypic plasticity and evolution in response to warming…………………………..128

Are all species advancing their phenologies?.....................................................128

Are all advances in phenology adaptive?............................................................130

Do individual traits respond to warming independently or in a correlated

manner?...............................................................................................................131

Future directions and implications for assisted colonization…………………………...132

References cited……….…..……………………………………………………………134

 

ix    

List of Tables

Table 2.1 Linear mixed effects analyses for differences in focal traits across populations and

thermal treatments………………………………………………………………..35

Table 2.2 Linear mixed effects analyses for independent differences in focal traits across

populations and thermal treatments……………………………………………...36

Table 2.3 Hurdle model showing the effects of focal traits on survival and seed

production………………………………………………………………………..37

Table 2.4 Estimates of direct and total phenotypic linear selection coefficients…………...38

Table 3.1 Summary of experiments contributing flowering phenology data………………74

Table A1 Estimates of individual synchrony, population synchrony, and the strength of

phenological assortative mating for populations of C. fasciculata across

experiments and for species native to KSR……………………………………...95

Table 4.1 Linear mixed effects analyses for differences in focal traits across species and

thermal, culture, and density treatments………………………………………..116

Table 4.2 Generalized linear mixed effects analyses for the influence of focal traits and

experimental treatments on reproductive biomass……………………………...117

Table 4.3 Estimates of direct phenotypic linear selection coefficients……………………118

 

x    

List of Figures

Figure 2.1 Map showing the distribution of C. fasciculata and the source populations and

experimental site at KSR………………………………………………………...39

Figure 2.2 Reaction norms showing plasticity in reproductive phenological traits across

populations and thermal treatments……………………………………………...40

Figure 2.3 Differences in plant size and seed production across populations and thermal

treatments………………………………………………………………………...41

Figure 2.4 Logistic regressions of survival as a function of flowering onset and plant size in

each thermal treatment…………………………………………………………...42

Figure 2.5 Relationships between seed number and focal traits in each thermal treatment…43

Figure 2.6 Estimates of the degree of temporal isolation between populations and treatments

and the average total flower production and flowering duration………………...44

Figure 3.1 Individual- and population-level flower display schedules from the warming

experiment………………………………………………………………………..75

Figure 3.2 Population-level display curves for open pollinated and pollinator excluded

treatments………………………………………………………………………...76

Figure 3.3 Heatmaps summarizing cross-correlations between pulsed phenologies and

unimodal simulations…………………………………………………………….77

Figure 3.4 Logistic regression relating floral longevity to average daily temperatures……..78

Figure 3.5 Display and deployment schedules of populations in the water manipulation

experiment………………………………………………………………………..79

Figure 3.6 Heatmaps summarizing cross-correlation coefficients between both display and

deployment schedules and average daily temperatures………………………….80

Figure 3.7 Histograms showing the distribution of estimates of individual synchrony,

population synchrony, and the strength of phenological assortative mating for

both C. fasciculata populations and native species to KSR……………………...81

Figure A1 Additional display schedules from the warming experiment……………………87

 

xi    

Figure A2 Display and deployment schedules for the warming experiment…………….….88

Figure A3 Display and deployment schedules for the pollination experiment………….…..89

Figure A4 Heatmaps relating display and deployment schedules to humidity……………...90

Figure A5 Heatmaps relating display and deployment schedules to precipitation………….91

Figure A6 Display schedules for each population in experiment 3…………………..……..92

Figure A7 Heatmaps relating display and deployment schedules with volumetric water

content in the watering manipulation experiment……………………………….93

Figure A8 Population-level display schedules for native species to KSR…………………..94

Figure A9 Individual synchrony estimates as a function of flowering duration…………….97

Figure A10 Hypothetical flowering schedules and corresponding estimates of synchrony.....98

Figure A11 Individual synchrony estimates as a function of sampling interval for a set of

hypothetical flowering schedules……………………………………………….100

Figure 4.1 Differences in flowering onset date and final plant size among species in thermal,

density, and culture treatments…………………………………………………119

Figure 4.2 Differences in reproductive biomass among species in thermal, density, and

culture treatments……………………………………………………………….120

Figure 4.3 Scaled differences in flowering onset date and corresponding phenotypic linear

selection coefficients among species in thermal, density, and culture

treatments……………………………………………………………………….121

Figure 4.4 Scaled differences in final plant size and corresponding phenotypic linear

selection coefficients among species in thermal, density, and culture

treatments……………………………………………………………………….122

 

xii    

List of Appendices

Appendix 1 Supplementary materials for chapter 3

Additional figures of phenotypic data……………………………………………………87

Estimates of individual synchrony, population synchrony, and the strength of phenotypic

assortative mating………………………………………………………………..95

Details on the individual synchrony metric……………………………………………...97

 

1    

Chapter 1

General Introduction

Rapid increases in global temperatures are inciting ecological and evolutionary responses

in species across a wide variety of taxa (Parmesan 2006). Some species may be able to track

their favorable climatic envelope and shift their distributions to higher latitudes and elevations

(Perry et al. 2005; Lenoir et al. 2008). Others must evolve at the same rate, or faster, than the

changing environment to avoid extinction (Bürger & Lynch 1995; Etterson & Shaw 2001).

Phenotypic plasticity, or the ability of organisms to alter their phenotypes in response to

changing conditions, can buffer against the ill effects of climate change for a time, depending on

the fitness consequences of the plastic response (Nicotra et al. 2010). Ultimately, the strong

directional selection pressures imposed by warming may require species to employ both

adaptation and migration strategies in order to ensure survival (Davis & Shaw 2001).

Conservationists have developed several management plans for species vulnerable to the

effects of climate change (McLachlan et al. 2007; Galatowitsch et al. 2009; Mawdsley et al.

2009). However, variability in species’ responses to warming, and uncertainty in their fitness

effects, has challenged our ability to foresee where conservation measures can be most

successfully applied (Lepetz et al. 2009). Identifying the additional factors promoting or

impeding responses to environmental change, and their adaptive value, may help to inform these

conservation policies.

For my dissertation, I examined the context-dependent nature of warming-induced

phenotypic plasticity to illuminate the potential causes of variation in responses across traits and

plant species. With a focus on plasticity in the timing of reproductive traits (i.e. phenological

traits), I explored the cumulative effects of increases in temperature on subsequently expressed

traits to ask whether the responses of individual traits are independent or correlated. A central

goal of this work was to identify potential constraints in the feasibility of the conservation

management practice of assisted colonization, or the facilitated movement of vulnerable species

to climatically favorable habitat beyond their current range boundary (Hunter Jr. 2007). I later

examined plasticity in patterns of flower deployment to gauge the effects of increased

 

2    

temperatures, fluctuations in internal resource allocation, and the frequency of rainfall on

patterns of synchrony and phenological assortative mating. Lastly, I investigated the potential

for competitive dynamics within a community of flowering plants to modify individual species’

responses to warming. In most studies, I used phenotypic selection analyses to determine the

circumstances in which plasticity is adaptive, and to verify whether evolutionary responses will

be required to maintain fitness in warmer climates.

Below, I describe the usefulness of plant phenological traits as biological markers of

environmental change. I then review the conservation management plan of assisted colonization

and discuss how an improved awareness of the factors influencing phenotypic responses to

warming can inform the successful application of this program. Lastly, I briefly discuss whether

limitations in trait-specific responses to warming or variation in the competitive community have

the potential to confound our understanding of phenological responses to climate change.

Phenology as a climatic indicator

Longitudinal studies of natural plant communities have demonstrated that phenological

traits can serve as indicators of environmental change (Walther et al. 2002). In general,

increases in temperature are prompting earlier transitions between life history stages across a

wide variety of taxa (Parmesan 2007). In many plant species, the dates of first flowering are

advancing with the earlier onset of spring, particularly for early-flowering species (Fitter & Fitter

2002; Menzel et al. 2006; Bertin 2008). Though informative, observational studies cannot

distinguish between shifts due to plasticity or those due to genetic changes, and often cannot

determine whether shifts were adaptive (Gienapp et al. 2008; Merilä & Hendry 2014). By

default, plasticity is frequently evoked as the mechanism of change, leading to speculation on the

limitations of plasticity and evolution in mitigating extinction risks and conjecture on the specific

strategies employed by plant lineages or morph types. While most phenological traits are

generally advancing as temperatures warm, the processes mediating species- or population-

specific responses to warming remain elusive.

The adaptive nature of phenological traits makes them ideal characters for studies of

climate change. Local adaptation in phenological traits can enable a population to respond to

local abiotic or biotic cues, ensuring the appropriate timing of various life history stages (Elzinga

 

3    

et al. 2007). Indeed, populations along climatic gradients are often differentiated for

phenological traits (Etterson & Shaw 2001; Hall & Willis 2006; Montague et al. 2007).

Phenological traits frequently have high levels of genetic variation within and among populations

(Etterson 2004; Weis & Kossler 2004; Burgess et al. 2007), providing the raw material necessary

for adaptive evolution (Franks et al. 2007). The timing of flowering in particular can be a

primary determinant of fitness (Milla et al. 2009), affecting rates of pollination and frugivory

(Augspurger 1981; Elzinga et al. 2007; Pais & Guitian 2007) and ensuring the favorable timing

of fruit maturation and seed dispersal (Rathcke & Lacey 1985). Furthermore, differences in

flowering times in sympatric subspecies can result in temporal reproductive isolation and

contribute to speciation (Hall & Willis 2006; Ellis et al. 2006; Levin 2006, 2009), while limited

phenological variation for flowering time in marginal populations can stabilize the range edge of

a species’ distribution (Griffith & Watson 2005, 2006; Chuine 2010). Aside from their far-

reaching effects, many phenological traits are conspicuous and can be easily measured and

quantified, making them ideal gauges of changes in climate.

The onset of flowering is one of the most studied life history stages in plants and the

processes that control it are complex and intertwined (Simpson et al. 1999). Several

developmental pathways for flowering have been discovered and are well explored in the model

organism Arabidopsis thaliana (Simpson & Dean 2002). The transition to flowering involves a

suite of interacting components from the photoperiodic, vernalization/autonomous, sucrose, and

gibberellins pathways that all regulate the expression of a few key genes (Blazquez 2000; Ausín

et al. 2005). Increased temperatures may only have an effect on flowering onset, and perhaps

other phenological traits, if the other internal and external conditions experienced by a plant are

conducive to progressions in development. For example, warmer temperatures may not advance

the onset of flowering in species with obligate photoperiodic requirements (Samach & Coupland

2000), but they may further accelerate flowering for plants experiencing reductions in the red:far

red light ratio imposed by neighboring plants (Halliday et al. 2003). Our understanding of the

molecular pathways governing floral transitions can help us predict the circumstances in which

phenotypic change may be facilitated or constrained by variation in the abiotic or biotic

environments.

Phenology and assisted colonization

 

4    

Assisted colonization, also known as assisted range expansion, is the facilitated

movement of species that are highly threatened or economically valuable to climatically

favorable habitat beyond their current range boundary (Hunter Jr. 2007, McLachlan et al. 2007,

Hoegh-Guldberg et al. 2008, Wilson et al. 2009). Here, I distinguish between assisted

colonization, as described above, and assisted migration, which refers to relocations occurring

between habitats within a species’ current distribution.

Assisted colonization has been suggested as a conservation measure against extinction for

species vulnerable to changes in climate. This proposal has stirred passionate debates in the

scientific community, clearly indicating a lack of consensus on the availability, reliability, and

interpretation of science to inform policy and management decisions. Proponents argue that,

lacking the luxury of time, action must be taken immediately to ensure the persistence of species

that are economically valuable or have limited distributions, dispersal abilities, or adaptive

potential (Sax et al. 2009; Schlaepfer et al. 2009; Vitt et al. 2009; Wilson et al. 2009).

Opponents feel that the potential risks associated with relocating species to new territories have

been historically disastrous, and have included the creation of invasive species, the spread of

diseases and pests, disturbances to food webs and other ecosystem dynamics, and a loss of

genetic diversity due to potential hybridization with native species (Fazey & Fischer 2009;

Ricciardi & Simberloff 2009; Wilson et al. 2009). Regardless, the ethical uncertainties

surrounding this conservation strategy must not prevent it from being scientifically examined

(Schwartz et al. 2009).

The evolution of phenological traits will likely play a large role in the success of assisted

colonization programs. For instance, the evolution of earlier reproduction in northern marginal

populations may be necessary for successful establishment and range expansion for the annual

cocklebur Xanthium strumarium (Griffith & Watson 2006) and the pitcher-plant mosquito

Wyeomyia smithii (Bradshaw & Holzapfel 2001). However, reductions in plant size associated

with the evolution of earlier flowering may limit population growth and northward spread, as

seen in the wetland plant Lythrum salicaria (Colautti et al. 2010). Plasticity in phenological

traits, if adaptive, may relieve selection pressures and provide more time for evolution to

generate favorable phenotypes (Nicotra et al. 2010). Alternatively, the relocation of seeds or

individuals from genetically, and phenologically, distinct populations may facilitate rapid

 

5    

evolution, a conservation strategy known as assisted gene flow (Aitken & Whitlock 2013).

Many of these predicaments have been explored independently and under various motivations,

but have yet to be explicitly linked to the success or failure of assisted colonization.

Independent vs. correlated responses to warming

In annual plants, the onset dates of reproductive phenological traits (e.g. flower bud

production, flowering onset, fruiting onset) are staggered sequentially after seedling emergence.

Plasticity in the timing of a specific trait may ensure its expression during favorable conditions

(Sultan 2000). However, genetic correlations among sequentially expressed phenological stages

(e.g. dates of emergence, flower bud production, flowering onset, and fruit maturation) could

constrain the individual stages from reaching their optimal values (Schlichting & Levin 1986).

Despite their inherent developmental association, phenological traits are often studied

independently and without consideration of other life history stages (Schlichting 1986). The

adoption of a cumulative life cycle perspective may illuminate how growth and development are

affected by increases in temperature and whether characters are environmentally or genetically

correlated.

While correlations between phenological traits have been frequently observed in nature,

less is known about how these correlations can be influenced by environmental change

(Antonovics 1976). Sherry et al. (2007) artificially warmed natural plant communities and

observed delayed flowering onset for later flowering species in heated conditions relative to

ambient. However, flower bud formation occurred at the same time or earlier in warmer plots,

suggesting that plasticity in flowering onset date is independent of flower bud formation or

earlier expressed traits. Conversely, plasticity in the onset of fruiting was largely attributable to

prior shifts in flowering onset date. Haggerty and Galloway (2011) found that increased

temperatures accelerated the onset of later expressed reproductive phenological stages relative to

earlier expressed traits in several population of Campanulastrum americanum. These studies

demonstrate that correlations between characters can be environmentally determined, and that

warming-induced phenotypic plasticity is both trait and species specific.

Negative genetic correlations between traits can impede adaptive evolution if both traits

are selected for in the same direction (Etterson & Shaw 2001). While phenotypic correlations

 

6    

can only generally be examined as proxies for genetic correlations, variation in the magnitude or

direction of phenotypic plasticity among correlated traits can identify characters that may be

capable of independent evolution (Falconer & Mackay 1996). Conversely, correlated responses

among traits may serve as indications that genetic correlations, through either pleiotropy or

linkage disequilibrium, may retard evolution and warrant further investigation (Lynch & Walsh

1998). Identifying disparities in the responses of distinct phenological phases to warming may

reveal which traits are sensitive to changes in climate and which are capable of responding to

patterns of direct selection.

Competitive dynamics and phenological responses to warming

At temperate latitudes, plants stagger their growth and development throughout the

growing season (Rabinowitz et al. 1981; Herrera 1986; Weis et al. 2014). Abiotic conditions

vary seasonally, and often in a predictable way, allowing plants to rely on dependable cues for

the appropriate timing of life history traits. Competitive dynamics may also vary temporally,

according to variation in the presence and abundance of species through time (Wiens 1977).

Accordingly, species occupying distinct temporal niches may experience contrasting competitive

regimes, with phenotypic plasticity in reproductive timing influencing the degree of temporal

overlap with conspecifics (Price & Waser 1998; Sherry et al. 2007; Aldridge et al. 2011). If

competition intensifies as the extent of overlap between species increases, then the adaptive

value of plasticity in reproductive traits may depend on the warming-induced responses of

competing species.

Many have observed that the responses of early-flowering species to warming may differ

in magnitude or direction from those flowering later in the season (Fitter & Fitter 2002; Menzel

et al. 2006; Sherry et al. 2007; Bertin 2008). Resource competition among developmentally

distinct species could produce these patterns in phenology if plants flowering earlier in the year

are able to accelerate reproduction without the same competitive repercussions experienced by

reproductive shifts in later flowering individuals. The disproportionate division of resources

among competing species (e.g. asymmetric competition) typically increases in intensity as

species similarity decreases (Keddy & Shipley 1989; Weiner 1990; Johansson & Keddy 1991).

That is, species that vary in growth and development tend to experience disparate degrees of

 

7    

competition. Competitive dynamics can influence plasticity in reproductive phenological traits

(Weiner 1988), and if competition is asymmetric, it may differentially influence flowering onset

dates in developmentally distinct species. Asymmetric competition is commonly found in

herbaceous plant communities (Keddy & Shipley 1989), and could be a contributor to variation

in phenological responses to warming among species.

The effects of competitive interactions are the net result of competition for multiple

resources. Competition asymmetries need not exist equally across required resources, and

limitations for various resources do not always impact phenotypes in the same way (Weiner

1990). For instance, competition for pollination services may be greatest between species that

are similar in phenology and are pollinated by the same suite of biotic vectors (Levin &

Anderson 1970). In this case, traits like flowering onset date would be unaffected by this

limiting resource, whereas pollen limitation can strongly reduce fruit set and increase final plant

size (Burd 1994). Conversely, competition for access to light can be magnified if some species

are faster to develop than others (Weiner 1986). In many cases, plants that are shaded have

accelerated flowering onset dates and achieve a smaller final plant size and lower fitness relative

to those with access to light (Schmitt & Wulff 1993; Schwinning & Weiner 1998; Franklin

2008). Competition for nutrients and soil moisture may influence life history traits in a similar

manner as competition for light (Chapin III 1980; Weiner 1988). The mechanisms of

competition can be teased apart experimentally (Weiner 1986; Wilson 1988), and the effects of

competition on phenotypic responses to warming may depend on the limiting resource as well as

the developmental variation and competitive abilities of the species present.

Research objectives and thesis outline

The objective of my thesis was to explore the adaptive role of reproductive phenotypic

plasticity in responses to increases in temperature. To achieve this, I conducted a series of field

experiments informed by previous work on my model species, Chamaecrista fasciculata, in

combination with several common garden experiments I ran in the greenhouse. I examined

variation in traits expressed within individuals, among individuals in a population, among

populations, and among competing species, in order to elucidate some of the factors contributing

to variation in responses to warming and to reveal the contexts under which plasticity alone may

 

8    

maintain fitness. Below, I briefly describe the objectives of each my thesis chapters, all of which

were written as independent manuscripts for publication.

Chapter 2 – Assisted colonization and assisted gene flow depend on phenology

Increases in temperatures are threatening the persistence of species in their current

geographic locations. For species of ecologic or economic importance, assisted colonization

may provide a reprieve from the negative effects of a changing climate, providing time for

adaptive evolution, and the relocations of genetically distinct populations, or assisted gene flow,

may further facilitate evolutionary responses. To better inform these policies, we planted seeds

from latitudinally distinct populations of Chamaecrista fasciculata in a potential future

colonization site north of its current range boundary. We exposed plants to either ambient

temperatures or those expected by mid century, and monitored a suite of life history traits to

determine the adaptive value of plastic responses. Population success was dependent on latitude

of origin, with southern populations performing the most poorly, even under elevated

temperatures. Differences in flowering phenology limit the potential for genetic exchange

among latitudinally disparate populations. Our results demonstrate that assisted colonization and

assisted gene flow may be feasible options for preservation provided that photoperiodic

constraints do not limit plasticity or evolution in reproductive phenological traits.

This chapter has been accepted for publication in Global Change Biology and is currently

in press. The work was completed in collaboration with Matthew N. Cumming (previously of

University of Toronto) and Arthur E Weis.

Chapter 3 - Simultaneous pulsed flowering in a temperate legume: causes and consequences of

multimodality in the shape of floral display schedules

The opportunities for pollen exchange among plants are dependent on temporal patterns

of floral displays, or display schedules. The shape of these schedules can influence the degree of

synchrony among individuals or populations and can influence the strength of phenological

assortative mating. However, studies of plasticity in flowering phenology are often limited to

variation in flowering onset date and assume correlated responses in subsequent patterns of

flower deployment. I monitored daily flower production for individual plants in several

 

9    

populations of Chamaecrista fasciculata exposed to treatments that differed in temperature,

pollinator availability, and watering schedules. Display schedule shape was plastic and

independent of shifts in flowering onset date in all populations and across experiments. Our

results indicate that this plasticity is likely due to the effects of seasonal changes in temperature

on patterns of flower deployment and floral longevity. We show that plasticity in schedule shape

resulted in a reduction of the average strength of phenological assortative mating for flowering

onset date and we discuss the potential for consequences on the efficacy of selection on

flowering time and correlated traits.

This chapter was published in the Journal of Ecology and is reprinted here with copyright

permission:

Wadgymar, Susana M., Austen, Emily J., Cumming, Matthew N., and Weis, Arthur E. (2015)

Simultaneous pulsed flowering in a temperate legume: causes and consequences of

multimodality in the shape of floral display schedules. Journal of Ecology, 103, 316-327.

Chapter 4 – The influence of competition on phenological responses to warming

Variation in species’ responses to increases in temperature has challenged researchers to

identify the additional factors promoting changes in growth and development. In plants, the

degree of shifts in reproductive phenological traits has been observed to vary according to a

species’ developmental position within a community of plants, with early flowering species

advancing more often, and to a larger degree, than those flowering later. Variation in

competitive dynamics has the potential to differentially affect species occupying distinct yet

overlapping temporal niches, making species’ responses to warming dependent on the

composition and density of the surrounding community. To investigate the influence of

competition on phenological responses to warming, we manipulated thermal regime, community

composition, and planting density in a field experiment using three species that varied in growth

and reproduction. In general, competitive dynamics did not modify responses to warming across

traits or species, nor did they alter the patterns of selection imposed by warming. This work

suggests that variation in the competitive environment may not act to constrain potential

responses to increases in temperature in most cases, and that other ecological or evolutionary

processes may be contributing to species-level differences in responses to warming.

 

10    

This work was completed in collaboration with Benjamin Gilbert (University of

Toronto), Matthew N. Cumming (previously of University of Toronto), Caroline M. Tucker

(University of Colorado Boulder), Marc W. Cadotte (University of Toronto Scarborough), and

Arthur E. Weis.

Chapter 5 - Concluding discussion

In my final chapter, I discuss the contributions of my thesis to studies of phenology and

climate change. I then outline how future trials of assisted colonization and assisted gene flow

may reveal when conservation attempts are likely to be successful. Lastly, I summarize areas of

future research that can further our understanding of how species may adapt to rapid changes in

climate.

References Cited

Aitken SN, Whitlock MC (2013) Assisted Gene Flow to Facilitate Local Adaptation to Climate

Change. The Annual Review of Ecology, Evolution, and Systematics, 44, 367-388.

Aldridge G, Inouye DW, Forrest JRK, Barr WA, Miller-Rushing AJ (2011) Emergence of a mid-

season period of low floral resources in a montane meadow ecosystem associated with

climate change. Journal of Ecology, 99, 905-913.

Antonovics J (1976) The nature of limits to natural selection. Annals of Missouri Botanical

Garden, 224-247.

Augspurger, C. K. (1981) Reproductive Synchrony of a Tropical Shrub: Experimental Studies on

Effects of Pollinators and Seed Predators in Hybanthus prunifolius (Violaceae). Ecology 62,

775-788.

Ausín, I., Alonso-Blanco, C. & Martínez-Zapater, J. (2005) Environmental regulation of

flowering. International Journal of Developmental Biology 49, 689-705.

Bertin RI (2008) Plant phenology and distribution in relation to recent climate change. Journal of

the Torrey Botanical Society, 135, 126-146.

 

11    

Blázquez MA (2000) Flower development pathways. Journal of Cell Science, 113, 3547-3548.

Bradshaw WE and Holzapfel CM (2001) Genetic shift in photoperiodic response correlated with

global warming. PNAS, 98, 14509-14511.

Burd MB (1994) Bateman’s principle and plant reproduction: the role of pollen limitation in fruit

and seed set. Botanical Review, 60, 83-129.

Bürger Rand Lynch M (1995) Evolution and extinction in a changing environment: a

quantitative-genetic analysis. Evolution, 151-163.

Burgess, K. S., Etterson, J. R. & Galloway, L. F. (2007) Artificial selection shifts flowering

phenology and other correlated traits in an autotetraploid herb. Heredity 99, 641-648.

Chapin III FS (1980) The mineral nutrition of wild plants. Annual Review of Ecology and

Systematics, 11, 233-260.

Chuine I (2010) Why does phenology drive species distribution? Philosophical Transactions of

the Royal Society B, 365, 3149-3160.

Colautti RI, Eckert CG, Barrett SCH (2010) Evolutionary constraints on adaptive evolution

during range expansion in an invasive plant. Proceedings of the Royal Society B, 277, 1799-

1806.

Davis M, Shaw RG (2001) Range Shifts and Adaptive Responses to Quaternary Climate Change.

Science, 292, 673-679.

Ellis, A. G., Weis, A. E. & Gaut, B. S. (2006) Evolutionary radiation of “Stone Plants” in the

genus Argyroderma (Aizoaceae): Unraveling the effects of landscape, habitat and flowering

time. Evolution 60, 39-55.

Elzinga JA, Atlan A, Biere A, Gigord L, Weis AE, Bernasconi G (2007) Time after time:

flowering phenology and biotic interactions. TRENDS in Ecology and Evolution 22, 432-

439.

Etterson JR, Shaw RG (2001) Constraint to Adaptive Evolution in Response to Global Warming.

Science, 294, 151-154.

 

12    

Etterson JR (2004) Evolutionary potential of Chamaecrista fasciculata in relation to climate

change. II Genetic architecture of three populations reciprocally planted along an

environmental gradient in the great plains. Evolution, 58, 1459-1471.

Falconer DS, Mackay TFC (1996) Introduction to quantitative genetics. Longman, 19(8), 1.

Fazey, I. & Fischer, J. (2009) Assisted colonization is a techno-fix. Trends in Ecology and

Evolution 24, 475.

Fitter AH, Fitter RSR (2002) Rapid Changes in Flowering Time in British Plants. Science, 296,

1689-1691.

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

Franks SJ, Sim S, Weis AE (2007) Rapid evolution of flowering time by an annual plant in

response to a climate fluctuation. PNAS, 104, 1278-1282.

Galatowitsch S, Frelich L, Phillips-Mao L (2009) Regional climate change adaptation strategies

for biodiversity conservation in a midcontinental region of North America. Biological

Conservation, 142, 2012-2022.

Gienapp, P., Teplitsky, C., Alho, J. S., Mills, J. A. & Merilä, J. (2008) Climate change and

evolution: disentangling environmental and genetic responses. Molecular Ecology 17, 167-

178.

Griffith, T. M. & Watson, M. A. (2005) Stress avoidance in a common annual: reproductive

timing is important for local adaptation and geographic distribution. Journal of Evolutionary

Biology 18, 1601-1612.

Griffith, T. M. & Watson, M. A. (2006) Is evolution necessary for range expansion?

Manipulating reproductive timing of a weedy annual transplanted beyond its range. The

American Naturalist 167, 153-164.

Haggerty BP, Galloway LF (2011) Response of individual components of reproductive

phenology to growing season length in a monocarpic herb. Journal of Ecology, 99, 242-253.

Hall, M. C. & Willis, J. H. (2006) Divergent selection on flowering time contributes to local

adaptation in Mimulus guttatus populations. Evolution 60, 2466-2477.

 

13    

Halliday KJ, Salter MG, Thingnaes E, Whitelam GC (2003) Phytochrome control of flowering is

temperature sensitive and correlates with expression of the floral integrator FT. The Plant

Journal, 33, 876-886.

Herrera J (1986) Flowering and fruiting phenology in the coastal shrublands of Doñana, south

Spain. Vegetatio, 68, 91-98.

Hoegh-Guldberg, O., Hughes, L., McIntyre, S., Lindenmayer, D. B., Parmesan, C., Possingham,

H. P., & Thomas, C. D. (2008) Assisted Colonization and Rapid Climate Change. Science

321, 345-346.

Hunter Jr., M. L. (2007) Climate Change and Moving Species: Furthering the Debate on Assisted

Colonization. Conservation Biology 21, 1356-1358.

Johansson ME and Keddy PA (1991) Intensity and symmetry of competition between plant pairs

of different degrees of similarity: an experimental study on two guilds of wetland plants.

Oikos, 60, 27-34.

Keddy PA and Shipley B (1989) Competitive hierarchies in herbaceous plant communities.

OIKOS, 54, 234-241.

Lenoir J, Gégout JC, Marquet PA, de Ruffray P, Brisse H (2008) A significant upward shift in

plant species optimum elevation during the 20th century. Science, 320, 1768-1771.

Levin, D. A. (2006) Flowering phenology in relation to adaptive radiation. Systematic Botany 31,

239-246.

Levin, D. A. (2009) Flowering-time plasticity facilitates niche shifts in adjacent populations.

New Phytologist 183, 661-666.

Levin DA and Anderson WW (1970) Competition for pollinators between simultaneously

flowering species. The American Naturalist, 104, 455-467.

Lepetz V, Massot M, Schmeller DS, Clobert J (2009) Biodiversity monitoring: some proposals to

adequately study species’ responses to climate change. Biodiversity and Conservation, 18,

3185-3203.

Lynch M, Walsh B (1998) Genetics and analysis of quantitative traits. Sunderland: Sinauer.

 

14    

Mawdsley JR, O’Malley R, Ojima, DS (2009) A review of climate-change adaptation strategies

for wildlife management and biodiversity conservation. Conservation Biology, 23, 1080-

1089.

Menzel A, Sparks TH, Estrella N, Koch E, AAsa§ A, Ahas§ R, Alm-Kübler K, Bissolli P, et al.

(2006) European phenological response to climate change matches the warming pattern.

Global Change Biology, 12, 1969-1976.

Merilä J, Hendry AP (2014) Climate change, adaptation, and phenotypic plasticity: the problem

and the evidence. Evolutionary applications, 7, 1-14.

McLachlan, J. S., Hellmann, J. J. & Schwartz, M. W. (2007) A Framework for Debate of

Assisted Migration in an Era of Climate Change. Conservation Biology 21, 297-302.

Milla, R., Escudero, A. & Iriondo, J. M. (2009) Inherited variability in multiple traits determines

fitness in populations of an annual legume from contrasting latitudinal origins. Annals of

Botany 103, 1279-1289.

Montague, J. L., Barrett, S. C. H. & Eckert, C. G. (2007) Re-establishment of clinal variation in

flowering time among introduced populations of purple loosestrife (Lythrum salicaria,

Lythraceae). Journal of Evolutionary Biology 21, 234-245.

Nicotra AB, Atkin OK, Bonser SP, et al. (2010) Plant phenotypic plasticity in a changing

climate. Trends in plant science, 15, 684-692.

Pías, B. & Guitián, P. (2007) Flowering phenology and female reproductive success in the mass-

flowering tree Sorbus aucuparia (Rosaceae). Ecoscience 14, 205-213.

Parmesan C (2006) Ecological and Evolutionary Responses to Recent Climate Change. Annual

Review of Ecology, Evolution, and Systematics, 37, 637-669.

Parmesan C (2007) Influences of species, latitudes, and methodologies on estimates of

phenological response to global warming. Global Change Biology, 13, 1860-1872.

Perry AL, Low PJ, Ellis JR, Reynolds JD (2005) Climate Change and Distribution Shifts in

Marine Fishes. Science, 308, 1912-1915.

Price MV, Waser NM (1998) Effects of experimental warming on plant reproductive phenology

in a subalpine meadow. Ecology, 79, 1261-1271.

 

15    

Rabinowitz D, Rapp JK, Sork VL, Rathcke BJ, Reese G, Weaver JC (1981) Phenological

properties of wind- and insect-pollinated prairie plants. Ecology, 62, 49-56.

Rathke, B. & Lacey, E. P. (1985) Phenological Patterns of Terrestrial Plants. Annual Review of

Ecology and Systematics 16, 179-214.

Ricciardi A, Simberloff D (2009) Assisted colonization is not a viable conservation strategy.

Trends in Ecology and Evolution, 24, 248-253.

Samach A, Coupland G (2000) Time measurement and the control of flowering in plants.

Bioessays, 22, 38-47.

Sax, D. F., Smith, K. F. & Thompson, A. R. (2009) Managed relocation: a nuanced evaluation is

needed. Trends in Ecology and Evolution 24, 472-473.

Schlaepfer, M. A., Helenbrook, W. D., Searing, K. B. & Shoemaker, K. T. (2009) Assisted

colonization: evaluating contrasting management actions (and values) in the face of

uncertainty. Trends in Ecology and Evolution 24, 471-472).

Schlichting CD, Levin DA (1986) Phenotypic plasticity: an evolving plant character. Biological

Journal of the Linnean Society, 29, 37-47.

Schmitt J, Wulff RD (1993) Light spectral quality, phytochrome and plant competition. Trends

in Ecology and Evolution, 8, 47-51.

Schwartz MW, Hellmann JJ, McLachlan JS (2009) The precautionary principle in managed

relocation is misguided advice. Trends in Ecology & Evolution, 24, 474.

Schwimming S, Weiner J (1998) Mechanisms determining the degree of size asymmetry in

competition among plants. Oecologia, 447-455.

Sherry RA, Zhou X, Gu S, Arnone III JA, Schimel DS, Verburg PS, Wallace LL, Luo Y (2007)

Divergence of reproductive phenology under climate warming. PNAS, 104, 198-202.

Simpson GG, Dean C (2002) Arabidopsis, the Rosetta stone of flowering time? Science, 296,

285-289.

Simpson GG, Gendall AR, Dean C (1999) When to switch to flowering. Annual Review of Cell

and Developmental Biology, 99, 519-50.

 

16    

Sultan SE (2000) Phenotypic plasticity for plant development, function and life history. Trends

in Plant Science, 5, 537-542.

Vitt, P., Havens, K. & Hoegh-Guldberg, O. (2009) Assisted migration: part of an integrated

conservation strategy. Trends in Ecology and Evolution 24, 473-474.

Walther, G., Post, E., Convey, P., Menzel, A., Parmesan, C., Beebee, T. J. C., Fromentin, J.,

Hoegh-Guldberg, O., & Bairlein, F. (2002) Ecological responses to recent climate change.

Nature 416, 389-395.

Weiner J (1986) How competition for light and nutrients affects size variability in Ipomea

tricolor populations. Ecology, 67, 1425-1427.

Weiner J (1988) The influence of competition on plant reproduction. Plant reproductive ecology:

patterns and strategies, 228-245.

Weiner J (1990) Asymmetric competition in plant populations. Trends in Ecology and Evolution,

5, 360-364.

Weis AE, Kossler TM (2004) Genetic variation in flowering time induces phenological

assortative mating: quantitative genetic methods applied to Brassica rapa. American Journal

of Botany, 91, 825-836.

Weis AE, Nardone E, Fox GA (2014) The strength of assortative mating for population date and

its basis in individual variation in flowering schedule. Journal of Evolutionary Biology, 27:

2138:2151.

Wiens JA (1977) On competition and variable environments: populations may experience

‘ecological crunches’ in variable climates, nullifying the assumptions of competition theory

and limiting the usefulness of short-tern studies of population patterns. American Scientist,

590-597.

Wilson JB (1988) Shoot competition and root competition. Journal of Applied Ecology, 279-296.

Wilson, J. R. U., Dormontt, E. E., Prentis, P. J., Lowe, A. J. & Richardson, D. M. (2009)

Something in the way you move: dispersal pathways affect invasion success. Trends in

Ecology and Evolution 24, 136-144.

 

17    

Chapter 2

The success of assisted colonization and assisted gene flow depends on phenology

This chapter resulted from collaboration with Matthew N. Cumming and Arthur E. Weis.

Susana M. Wadgymar carried out the experiment, performed the analyses, and wrote the

manuscript. MNC assisted with fieldwork, while AEW contributed to ideas and manuscript

editing. This manuscript has been accepted with minor revisions in Global Change Biology.

Abstract

Global warming will jeopardize the stability and genetic diversity of many species.

Assisted colonization, or the movement of species beyond their current range boundary, is a

conservation strategy proposed for species with limited dispersal abilities or adaptive potential.

However, species that rely on photoperiodic and thermal cues for development may experience

conflicting cues if transported across latitudes. Relocating multiple, distinct populations may

remedy this quandary by expanding genetic variation and promoting evolutionary responses in

the receiving habitat - a strategy known as assisted gene flow.

In order to better inform these policies, we planted seeds from latitudinally distinct

populations of the annual legume, Chamaecrista fasciculata, in a potential future colonization

site north of its current range boundary. Plants were exposed to ambient or elevated

temperatures via infrared heating. We monitored several life history traits and estimated patterns

of natural selection in order to determine the adaptive value of plastic responses. To assess the

feasibility of assisted gene flow between phenologically distinct populations, we counted flowers

each day and estimated the degree of temporal isolation between populations.

Increased temperatures advanced each successive phenological trait more than the last,

resulting in a compressed life cycle for all but the southern-most population. Warming altered

patterns of selection on flowering onset and vegetative biomass. Population performance was

dependent on latitude of origin, with the northern-most population performing best under

 

18    

ambient conditions and the southern-most performing most poorly, even under elevated

temperatures. Among-population differences in flowering phenology limited the potential for

genetic exchange among the northern and southern-most populations.

All plastic responses to warming were neutral or adaptive, however photoperiodic

constraints will likely necessitate evolutionary responses for long-term persistence, especially

when involving populations from disparate latitudes. With strategic planning, our results suggest

that assisted colonization and assisted gene flow may be feasible options for preservation.

Introduction

Environmental change wrought by increasing global temperatures can compromise the

persistence of species in their current geographic locations. Although some species may possess

sufficient developmental and physiological plasticity to tolerate novel conditions, others may be

prone toward extinction (Thomas et al. 2004; Carpenter et al. 2008). Migration to newly suitable

locations may alleviate some of the pressures imposed by a changing climate, and indeed

numerous taxa have shifted their ranges to higher latitudes and elevations over the past few

decades (Perry et al. 2005; Parmesan 2006; Lenoir et al. 2008). However, thermal conditions

define only part of a species’ niche, and migrating species may still experience unfamiliar

conditions in the receiving habitat, including novel community assemblages (Hellmann et al.

2012; Nooten et al. 2014) and photoperiodic regimes (Griffith &Watson 2006). For some,

survival will depend on adaptation to novel conditions at current localities, or on a combination

of migration and evolution, as has been argued for the pole-ward migration of species following

glaciation (Davis & Shaw 2001). Limitations in dispersal abilities or adaptive capacities may

necessitate the preemptive management of vulnerable or economically valuable species (Aitken

et al. 2008; Hobbs et al. 2009).

Several conservation measures have been proposed to mitigate extinction risks in the face

of climate change. Assisted colonization, or assisted range expansion, refers to the movement of

a species beyond its current range boundary and has been recently recommended for species

unable to adapt or migrate in response to global warming or for species of economic or ecologic

importance (McLachlan et al. 2007; Kreyling et al. 2011; Loss et al. 2011). Despite being hotly

debated as a management strategy (Mueller et al. 2008; Ricciardi & Simberloff 2009; Vitt et al.

 

19    

2010), formal trials to assess the prospects and limitations of assisted colonization have been few

(Hewitt et al. 2011). Additionally, academic, conservation, government, and industry sectors

often differ in their management goals, and the focuses of existing studies of assisted

colonization are not always applicable across disciplines (Pedlar et al. 2012). For example, with

regards to plants, the interests of forestry professionals (e.g. maximizing woody growth, Lu et al.

2014) do not align with those considering assisted colonization as a species conservation strategy

(e.g. maximizing population size, Willis et al. 2009) or for ecosystem maintenance or restoration

(e.g. maximizing primary productivity, Grady et al. 2011; Lunt et al. 2013). Whatever the

motivation may be, long-term success will depend on the ecological and evolutionary responses

of newly established populations to the selective pressures imposed by a continuously changing

climate.

The success of assisted colonization across latitudes is contingent upon the choice of

relocation sites and source populations (Kreyling et al. 2011; Leech et al. 2011). Potential

relocation sites should lie within areas projected to have similar climatic conditions found within

the historic range (e.g. north of the current range boundary, Kreyling et al. 2011). The choice of

source populations, however, can present a conundrum. Those residing along the leading edge

of a species’ historic distribution may be best matched to the current thermal conditions just

outside of it (Hill et al. 2011), improving the chances of short-term success. However, rapid

evolutionary responses may be necessary for long-term persistence as the climate continues to

warm. Populations from lower latitudes may be better suited to the higher temperatures expected

of the future conditions in the new site (Grady et al. 2011), but may be unable to establish under

current conditions. One challenge in implementing assisted colonization is in identifying a

source population, or mixture of source populations, that ensures sufficient genetic variation in

key traits to generate genotypes suited to future climatic conditions.

Many factors can limit the establishment of species in more pole-ward sites, and these

restrictions will be particular to the species involved. However, there is one factor that changes

with latitude in an absolutely predictable fashion – the annual photoperiodic cycle. Species that

rely on photoperiodic and thermal cues for growth or development may experience conflicting

signals if relocated further north (Bradshaw & Holzapfel 2010). While evolutionary responses to

changes in environmental conditions can proceed rapidly (Bradshaw & Holzapfel 2001; Franks

 

20    

et al. 2007), in the case of assisted colonization, immediate, plastic responses to novel

temperatures or photoperiods may determine whether relocations are successful. While difficult

to assess in a field setting, it may be important to consider the implications of photoperiodic

mismatches when examining the responses of populations to increases in temperature (Bradshaw

& Holzapfel 2008) and when selecting candidate species or populations for assisted colonization.

Phenological traits can be influenced by both thermal (Blázquez et al. 2003) and

photoperiodic (Kobayashi & Weigel 2007) regimes, and the timing of these traits may serve as

indicators of environmental change (Fitter & Fitter 2002; Parmesan 2007). Despite their

potential adaptive value (Stevenson & Bryant 2000; Chuine 2010), the fitness impacts of

climate-induced shifts in development are seldom measured (Merilä & Hendry 2013).

Additionally, phenological traits are expressed sequentially (e.g. in plants, flowering onset

precedes fruiting onset), the plastic responses of individual traits are rarely measured

independently of previously expressed traits (but see Haggerty & Galloway 2011; Kim &

Donohue 2011). The monitoring of phenological traits may reveal whether relocated individuals

are well suited to conditions in the receiving habitat, yet failing to examine the cumulative

influence of environmental change across the life cycle may result in misidentifying the true

targets or agents of selection or their capacity for evolutionary response (Ehrlén 2015).

Relocating individuals from multiple, distinct populations, or their hybrids, may expand

genetic variation in the newly founded population, increasing the chance that some individuals

respond favorably to relocation and have reproductive rates high enough to sustain the

population in the short term. Recombination among genetic variants would enable evolutionary

adaptation to novel combinations of habitat and climate (Rice & Emery 2003; Tallmon et al.

2004; Loss et al. 2011; Aitken et al. 2008). However, population differences in reproductive

phenology could limit, and bias, the potential for genetic exchanges among migrants (Weis

2015). Implementing a combination of assisted colonization with assisted gene flow would

require knowledge of the potential for natural genetic exchange among the populations relocated

together under current and future conditions (Aitken & Whitlock 2013). Experiments founded in

proposed relocation sites can explore these patterns, and can also help to reveal which traits

contribute to fitness and how selection regimes will change as temperatures continue to warm

(Aitken et al. 2008; Lawler & Olden 2011).

 

21    

In this study, we examined phenotypic responses to warming and patterns of natural

selection beyond the range and discuss our results as they relate to the feasibility of assisted

colonization in combination with assisted gene flow. We planted seeds from latitudinally

distinct populations of the annual legume Chamaecrista fasciculata north of its current range

boundary and exposed them to present-day and future climatic conditions using artificial climate

warming arrays. We assessed the responses of phenological traits in units of calendar days as

well as growing degree-days, and we also monitored growth, survival, and seed production in

order to estimate patterns of natural selection. With an emphasis on comparing the feasibility of

these management plans under current and future climatic conditions, we ask: (1) In an assisted

colonization program, which source population(s) are likely to succeed? and (2) In an assisted

gene flow program, do phenological differences between populations impede genetic

introgression?

Methods

Study species

Chamaecrista fasciculata Michx. (Fabaceae, subfamily Caesalpinoideae) is a frost-

intolerant annual plant of tropical descent (de Souza Conceição et al. 2009) distributed in North

America across the Great Plains and eastward towards the Atlantic (Irwin & Barneby 1982). Its

upper range boundary includes all of the northern U.S. states from Minnesota to New York but it

has not yet been known to occur in Canada (Fig 2.1). It is frequently found in sandy soils and

occupies prairie habitat or sites that have been recently disturbed (Foote & Jackobs 1966).

Growth and flowering are indeterminate, with an individual plant producing anywhere from 1 to

several hundred flowers in its lifetime.

During the summer of 2009 seeds were collected from populations located along two

latitudinal transects in the US (Fig. 2.1): one through the Midwest from Minnesota (MN,

44.8011°N, 92.9647°W) south to Missouri (MO, 38.4979°N, 90.5610°W) and the other along the

East Coast from Pennsylvania (PA, 40.1790°N, 76.7248°W) south to North Carolina (NC,

35.8900°N, 79.0092°W). When possible, three fruits were collected from 50-100 individuals per

population from plants spaced approximately 5 m apart (an estimated genetic neighborhood size

for this species, Fenster 1991). Others have shown that populations of C. fasciculata are locally

 

22    

adapted across large geographic distances (Galloway & Fenster 2000; Etterson & Shaw 2001)

and that the spatial scale of gene flow is limited via both pollen movement and seed dispersal

(Fenster 1991; Fenster et al. 2003).

Experimental design

We used the Experimental Climate Warming Arrays at the University of Toronto’s field

station in southern Ontario, the Koffler Scientific Reserve at Joker’s Hill (KSR, 44.0300°N,

79.5275°W), to expose plants to either present day thermal regimes or those predicted of the area

by mid-century (OMNR 2007). Each warming array consisted of a steel triangular structure

anchored 1.25 meters above the ground with six infrared heaters mounted in a hexagonal

configuration along the sides (design per Kimball et al. 2008). Heating elements were angled

inward and down from horizontal, producing a uniform heat shadow of 3 meters in diameter. Six

plots were heated by 1.5°C during the day and 3°C at night (Easterling et al. 1997), while six

remained unheated. Temperatures were monitored at the plot level in three arrays per treatment

using infrared radiometers (SI-111 infrared radiometer, Campbell Scientific, Edmonton,

Canada). Measurements were taken every 15 minutes, and a comparison of the average

temperatures within a treatment were used to determine the degree of heat output necessary to

maintain the target level of warming. Due to technical issues, data from one of the heated plots

were dropped from all analyses.

In May of 2011 seeds were scarified, stratified for 3 days, and planted within the

warming arrays in a hexagonal design with 20 cm spacing between plants. Individuals from each

population were randomized within each plot. Due to record levels of precipitation in the area,

all seedlings drowned and seeds had to be replanted in pots in the greenhouse adjacent to the

heating arrays while the experimental plots drained of water. To expose seedlings to different

temperatures from emergence onward, half of the plants were moved just outside of the

greenhouse to experience ambient thermal conditions while the remaining plants inside the

insulated greenhouse were exposed to elevated temperatures. Temperature measurements from

several iButton Temperature Loggers (1992L, Maxim Integrated, San Jose, California, USA)

placed at soil level in random pots indicated an average temperature difference of 2.7°C between

locations inside and outside of the greenhouse. Plants were given 0.5 oz of fertilizer (20-20-20,

 

23    

1g/L) at 21 days after planting, and at 25 days seedlings were transplanted into the warming

arrays as outlined previously, with 5 plants per plot from each population. A ring of non-focal

plants was planted around the focal individuals in each plot to absorb any edge effects. Plots

were watered every few days for the first two weeks after transplanting, after which they

received natural levels of precipitation. Plots were weeded periodically throughout the

experiment to minimize interspecific competition and plants were harvested upon first frost, 178

days after planting.

Plants were measured each day for several sequentially expressed phenological traits, or

phenophases; the date of emergence, first flower bud, first open flower, and first mature fruit.

The date of first bud was recorded when a flower bud first reached a length of 0.5 cm while the

date of first mature fruit was noted as the date when the first fruit pod browned and seeds rattled

within. All fruit were collected at this stage, before the pods explosively dehisced their seed.

Flowers were counted daily on all individuals for 86 out of the 93 days where flowers were

present, allowing us to examine the effects of temperature on the total number of flowers

produced and the duration of flowering. We measured aboveground vegetative biomass after

harvesting, and female fecundity was defined as the total number of seeds produced by an

individual.

Our ability to assess responses to climate change can be dependant on the way that we

measure progression through life history stages. The growth and development of plants is

frequently influenced by temperature, and in the absence of other limiting factors, a minimum

accumulation of heat can be required before proceeding to the next developmental stage (Wang

1960). Due to variable conditions among years, the number of days that elapse before this

minimum heat sum is reached can vary. Accordingly, studies of the effects of warming on the

timing of life history traits (in units of days) can be complimented by also examining the

accumulation of heat across days, or growing degree-days (GDD), upon the expression of those

traits (in units of °C·day, Neuheimer & Taggart 2007). For example, a phenological trait that

requires a fixed heat sum before developing will exhibit a constant value of accumulated GDD

even if the timing of that trait shifts in response to changes in climate. In this way, temporal

changes in phenological traits do not necessarily reflect plasticity to increasing temperatures;

rather, plant development is contingent on specific patterns of heat accumulation. While this

 

24    

methodology is often overlooked outside of agriculture or entomology, phenological traits

measured in units of GDD can yield less variable results and can increase predictive power

(Neuheimer & Taggart 2007).

We calculated the accumulated GDD upon the onset of budding, flowering, and fruiting

for each individual. Growing degree-days are calculated by comparing the average daily

temperature to a base temperature, Tbase, below which growth does not occur: GDD = (Tmax –

Tmin) / 2 – Tbase, where Tmax and Tmin are the maximum and minimum daily temperatures,

respectively (Miller et al. 2001). Tbase ranges from 5°C to 10°C in most commercial species,

and Tmax can be capped at 30°C because growth often does not continue to accelerate at higher

temperatures (Wang 1960). However, many tropical species may require temperatures in excess

of 30°C for the development of certain traits (Trudgill et al. 2005 and references therein). Due to

C. fasciculata’s tropical origin, we calculated daily GDD in a variety of scenarios, with 0.5°C

increments of Tmax capped from 30°C to 35°C and of Tbase ranging from 5°C to 10°C (121

combinations in total). If average daily temperatures dropped below Tbase, GDD was set to 0.

In order to obtain weather data for the period of time prior to planting within the warming

arrays, we used temperature data collected by Environment Canada at the nearby Buttonville

Airport (43.8608°N, 79.3686°W) to calculate GDD for plants in ambient treatments. For heated

treatments, we added the average of the daytime and nighttime increases in temperature due to

artificial warming (2.25°C) to temperature data. For a given trait, accumulated GDD is

calculated as the sum of daily GDD from the date of planting to the date of trait onset.

Statistical analyses

We confirmed that heated plots were maintained at a warmer temperature throughout the

experiment by comparing temperature differences between treatments with a repeated measures

linear model with thermal treatment, day, and their interaction as fixed effects. We incorporated

an auto-regressive error structure of order 1 to account for any autocorrelation in observations

among days, nested within plot (Zuur et al. 2009). This analysis was performed using the nlme

package (Pinheiro et al. 2014) in R (R Development Core Team 2014).

 

25    

Differences in the timing of budding, flower, and fruiting onset, as well as flowering

duration, the total number of flowers produced, and above ground vegetative biomass were

analyzed via linear mixed models, again using the nlme package in R. In order to meet the

assumption of residual normality, we analyzed the log of vegetative biomass +1. Population,

thermal treatment, and their interaction were included as fixed effects while plot was included as

a random effect if it improved model fit. A significant effect of temperature is indicative of

phenotypic plasticity in the trait of interest, a population effect reflects genetic differentiation

among populations, and a significant interaction reveals genetic differences among populations

in their plastic responses to warming temperatures. Variance heterogeneity among populations

or treatments was corrected using error variance covariates, if necessary (Zuur et al. 2009). The

random term and error covariate components of the model structure were selected by minimizing

AIC values, after which models fit via maximum likelihood were used to optimize the fixed

effects. Here, we present final, optimized models selected via log likelihood ratio tests. We

further assessed the influence of thermal treatment on development by repeating these analyses

and substituting the accumulated GDD at the onset of budding, flowering, and fruiting in lieu of

calendar days.

If a series of sequentially-expressed phenological traits are accelerated or delayed by

warmer temperatures, two factors may be involved. An early initiation of the first trait in the

sequence will contribute to an early initiation of the remainder, while the intervals between these

transitions, or phenophases, may show no change. However, trait-specific plastic responses can

independently diminish or extend the phenophase intervals. We estimated the degree of

independence for budding, flowering, and fruiting responses by repeating the analyses described

above with the onset date of the previous phenophases as a covariate. In annuals, plant size is

often inherently correlated to flowering onset (Bolmgren & Cowan 2008; Weis et al. 2014), and

we included flowering onset in as a covariate in the reanalysis of vegetative biomass. Similarly,

we included flowering onset as a covariate in the examination of the independent effects of

thermal regime on total flower number and flowering duration.

Phenotypic selection analysis

 

26    

We estimated whether any temperature-induced changes in phenology or vegetative

biomass were adaptive by assessing the magnitude of phenotypic selection. We used a hurdle

model to separately examine the effects of phenotype on two components of fitness; survival to

fruiting and total seed production. We first analyzed survival with a binomial generalized linear

model with log link (the zero component) using the glm function in R, after which we modeled

seed production (excluding zeros) with a zero-truncated negative binomial generalized linear

model (the count component) using the VGAM package (Yee 2010). Analyses estimating the

strength of direct selection on individual traits include the date of flowering onset and vegetative

biomass as fixed effects in both components, while fruiting onset was also included in the count

component. Total selection was calculated with separate univariate analyses for each of the

aforementioned traits. Trait values were rescaled to a mean of 0 and a standard deviation of 1

across populations and treatments, and we included temperature as an interacting fixed effect

with all traits. A significant interaction between a trait and temperature would indicate that

patterns of selection on that trait differ between thermal regimes. We also included population as

a fixed effect to account for any unmeasured differences among populations that may affect

survival or seed production. We report the significance of fixed effects from the final, optimized

models via chi-squared values from analyses of deviance.

The coefficients from the saturated hurdle model will be reported as estimates of direct

and total linear selection, however we caution that they are not comparable to selection gradients

and differentials as calculated by multiple regression (Lande & Arnold 1983). We chose

statistically sound methodology to estimate relationships between phenotype and fitness,

whereas linear multiple regression would have violated a number of assumptions (Mitchell-Olds

& Shaw 1987). For comparison, we report selection gradients and differentials as well, but our

figures and discussion of selection will focus on the results derived from the hurdle model. To

obtain gradients and differentials, we standardized traits and calculated relative fitness within

each treatment, with relative fitness defined as the total number of seeds produced by an

individual divided by the average number of seeds produced by all individuals in a given

treatment.

Temporal reproductive isolation

 

27    

We estimated the potential for gene flow between populations, as would occur under a an

assisted gene flow scenario, by analyzing the overlap in flowering schedules in order to

determine the proportion of opportunities for pollen exchange between populations. For the few

days with missing flower count data, we used a linear function running from the day before to

the day after the missed count to interpolate the expected number of flowers. Within a treatment,

and for each population pairing, we used daily flower counts to construct an n x n matrix of pair-

wise mating opportunities between all individuals. We assigned a score of 0 for one population

and 1 to the other, and we calculated a correlation between the population of origin for pollen

recipients and that of their potential pollen donors, weighted by the mating probabilities

produced in the mating matrix (see Weis & Kossler 2004). Values of the resulting correlation

coefficient, ρ, range from 0 to 1, and reflect completely random mating to complete reproductive

isolation, respectively. We obtained 95% confidence intervals on ρ by bootstrapping 1000x with

replacement (Weis & Kossler 2004).

Results

Thermal environment

Average temperatures were consistently higher in artificially warmed plots than in

ambient plots throughout the season (average daily temperature difference = 2.17°C;

Temperature, F1, 631=11.75, p<0.001; Day, F1, 631=346.71, p<0.001; Temperature*Day, F1,

631=0.46, p=0.5). Consequently, plants in heated plots had the opportunity to accumulate more

GDD from the timing of planting to first frost than those in ambient plots (2036 °C·day ± 22.8

SE vs. 1707 °C·day ± 21.4 SE, respectively, across all combinations of Tbase and Tmax).

Population differences and responses to warming

Populations differed genetically in the average onset dates of budding, flowering, and

fruiting according to their latitude of origin, with northern populations progressing through

developmental phases earlier and more rapidly than southern populations (Table 2.1, Population

term, Fig. 2.2a). Warming advanced these traits in the MN, PA, and MO populations, with each

phenophase advancing more than the last (Table 2.1, Temperature term). Ultimately, warming

compressed the life cycle of plants in these populations. In the southern-most NC population,

 

28    

increased temperatures advanced budding and flowering onset, but not fruiting. However, more

NC individuals survived to produce fruit in the heated treatment than in the unheated (27% vs.

68%, respectively). Variation among populations in plasticity also increased in later-expressed

traits, with the onset of fruiting responding to warming most variably among populations (Table

2.1, Population*Temperature term).

To determine if the time intervals between the phenophases responded to warming, we

amended the analyses by including the onset date of the previous phenophase as a covariate

(Table 2.2). Budding, flowering, and fruiting onset were influenced by temperature

independently of shifts in the previous phenophases (emergence, budding, and flowering,

respectively). However, the degree of independent response for flowering onset averaged just

one day beyond shifts due to previous traits. Variation among populations in plasticity for

budding and flowering onset disappear once the preceding responses to warming are accounted

for. Only fruiting onset displayed significant independent variation in plasticity among

populations.

Plants in the heated treatment had accumulated more growing degree-days at the onset of

budding, flowering, and fruiting than plants in ambient conditions, despite the temporal

acceleration of most traits in all populations (Table 2.1, Fig. 2.2b). Additionally, we detected

much more variation in plasticity when phenological responses were assessed via heat-sums than

when analyzed using calendar days. These patterns remain when including the accumulated

GDD of the previous phenophases in analyses (Table 2.2), and are a direct indication of thermal

plasticity in phenological traits.

Increased temperatures had no effect on the final, aboveground vegetative biomass of any

population, although plants from the NC population were significantly larger those from the

other populations (Table 2.1, Fig. 2.3a). Unlike most annual plants, there seems to be no

association between flowering onset date and plant size in C. fasciculata (Table 2.2), suggesting

that the evolutionary potential of phenological traits may not be strongly affected by patterns of

selection on growth.

In ambient conditions, the average number of seeds produced by a plant varied by

population latitude of origin and was highest in the MN population and lowest in the NC

 

29    

population (Fig. 2.3b). All populations produced more seed in the warmer environment, with

proportionately larger increases in the southern populations. For the PA and MO populations,

this warming-induced increase resulted in seed production levels equivalent to that of the MN

population.

Phenotypic selection analysis

We examined the effects of temperature on the relationship between focal traits and two

different fitness components; survival and the number of seeds produced. Despite responding to

warming, the influence of flowering onset on survival was marginal and similar in both thermal

environments (Table 2.3, Fig. 2.4a), with direct selection favoring early flowering (Table 2.4).

In contrast, early flowering only increased seed production under ambient conditions and was

negligible in the warmer environment (Table 2.3, Fig. 2.5a). Thus, with regards to seed

production, the advancement of flowering onset when heated was adaptive and may have

ameliorated the maladaptive timing of flowering in the more southern populations. These results

illustrate that increasing temperatures can influence patterns of selection through shifts in

phenological traits, and that warming can differentially affect selection through separate

components of fitness.

Early fruiting onset increased seed production in both thermal environments, and direct

selection on this trait was slightly stronger in the ambient treatment, again suggesting that

plasticity may have relieved selection on this trait (Fig. 2.5b). Final vegetative biomass strongly

influenced survival when warmed (Table 2.3, Fig. 2.4b) resulting in selection for larger plant size

only in the heated conditions (Table 2.4), whereas larger plant size increased seed production

similarly in both thermal environments (Fig. 2.5c). In this case, warming altered patterns of

selection on a trait unaffected by temperature, but this was only apparent when examining the

effects of trait variation on survival.

Temporal reproductive isolation

We calculated the strength of temporal reproductive isolation to gauge the potential for

gene flow between pairs of populations. Mating opportunities were greatest between populations

most similar in latitude of origin. In ambient conditions, ρ was lowest between the MN and PA

 

30    

populations and the PA and MO populations (Fig. 2.6a, below diagonal). In contrast, ρ was

highest between the southern-most NC population and the northern MN and PA populations,

indicating that these populations are almost entirely reproductively isolated.

The strength of ρ was significantly weaker for four out of six cases when in heated (Fig.

2.6a, above diagonal) versus ambient conditions, reflecting a greater potential for gene flow

among populations as temperatures warm, and was significantly stronger in another comparison.

Flowering duration was slightly variable among populations and thermal treatment, with the NC

population flowering longer in ambient conditions than in heated and vice versa for the MO and

MN populations (Table 2.1, Fig. 2.6b). Temperature had no effect on the total number of

flowers produced in any population (Table 2.1, Fig. 2.6c). Although small, the differences in the

degree of temporal isolation due to warming could be due, in part, to shifts in flowering onset

date and the duration of flowering.

Discussion

We have presented an experiment to mimic the assisted colonization of a species to a

pole-ward site beyond its historic geographic range under both current and anticipated future

thermal regimes. We show that colonists from northern populations are the most fit under

ambient temperatures, and that mean fitness steadily declines for colonists from lower latitudes.

Warmer temperatures alleviate the fitness decline for all but the southern-most (NC) colonists.

Nearly one in three plants from the NC locality failed to produce any seed in the heated

treatment, and total seed production was only half of that of the northern-most population. Thus,

the successful pole-ward colonization by a population that is expected to be pre-adapted to future

thermal regimes may be limited by other environmental factors.

Our experiment also assessed the potential for assisted gene flow; that is, the potential for

colonists from the south to interbreed with northern populations, and thereby introduce genes

that may be adaptive under warmer temperatures. Differences in flowering time severely

restricted mating opportunities between northern- and southern-most populations, regardless of

thermal regime. Here we discuss the contributions of phenological and growth traits to fitness

under ambient and warmed conditions and expand upon the implications for assisted

colonization and assisted gene flow.

 

31    

Phenotypic selection and responses to warming

Increased temperatures accelerated the onset of reproductive traits in all populations of

Chamaecrista fasciculata, resulting in a compression of life cycle length for all but the southern-

most population. Flexibility in life cycle length can have profound consequences, including the

potential to alter demographic processes (Galloway & Burgess 2009; Kai Zhu et al. 2013),

community structure or composition (Sherry et al. 2007), interactions with pollinators (Elzinga et

al. 2007), or traits expressed in the offspring generation (Galloway & Etterson 2007). Although

seldom examined in plants, warming has elicited abbreviated periods of growth and reproduction

in two arctic shrubs (Post et al. 2008), a monocarpic herb (Haggerty & Galloway 2011), and

three perennial grassland species (Frei et al. 2014), suggesting that this phenomenon may be

more common than currently appreciated.

Plasticity in phenological traits can modify the timing of subsequently expressed

phenophases (Donohue 2002), with the potential for individual traits to shift in opposing

directions (Sherry et al. 2007). In C. fasciculata, we found that each phenological trait advanced

more than the last, with the onset of fruiting displaying the greatest degree of plasticity in all

populations. In herbaceous species, flowering onset date is the most commonly examined trait in

studies of warming (Fitter & Fitter 2002; Parmesan 2006). However, our results suggest that

fruiting onset can display higher sensitivity to temperature, is under stronger selection, and

displays greater variation in plasticity among populations. These trends were amplified when

examining phenological shifts in units of GDD. The temperature-driven responses revealed by

the GDD analyses imply that the timing of life history traits is not entirely dependent on the

accumulation of heat sums and that other factors, like photoperiod, may be influential in the

expression of phenological traits.

Once flowering begins, plants must allocate resources between growth and reproductive

functions, often producing a relationship between plant size and flowering onset date (Bolmgren

& Cowan 2008). Warming-induced shifts in flowering time may have consequences for

competitive ability and reproductive capacity indirectly through their influence on size. In C.

fasciculata, reproductive phenological traits responded to warming while final vegetative

biomass and biweekly stem diameter measurements (data not shown) did not, suggesting that

 

32    

growth and development are not linked in this species. However, it has been demonstrated that

accelerated growth in warmer conditions can compensate for the earlier onset of maturity

(Neuheimer & Grønkjær 2012; Zhang et al. 2012). Additionally, increased levels of carbon

dioxide can differentially influence reproductive and vegetative traits (Reekie & Bazzaz 1991),

as demonstrated in C. fasciculata (Farnsworth & Bazzaz 1995), and the effects of elevated CO2

and temperature may interact synergistically to affect the expression of life history traits or

allocation patterns between growth and reproduction (Morison & Lawlor 1999).

Selection on flowering onset date and final plant size differed between thermal

environments, demonstrating that patterns of selection may change as temperatures warm, and

that such changes are not always a result of plastic responses to warming. Additionally, the

magnitude of direct selection imposed by warming differed among fitness components. The

plastic responses observed for all traits were either neutral or adaptive, and climate-related

genetic variation in plasticity among populations could facilitate evolution and further bolster

population performance. However, others have shown that the evolutionary responses of C.

fasciculata to warming may be constrained by genetic correlations antagonistic to the direction

of selection (Etterson & Shaw 2001) or by low heritabilities in fitness related traits (Etterson

2004). These previous findings should be interpreted with some caution (Bradshaw & Holzapfel

2008); the elevated temperature regime was achieved by transplanting northern population to

southern latitudes, thus confounding thermal and photoperiodic effects on the expression of loci

contributing to phenology. Nevertheless, evolutionary restrictions like these may limit a species’

ability to adapt in pace with a rapidly changing climate, and should be considered in the

decision-making criteria for assisted colonization.

Considerations for assisted colonization and assisted gene flow

Thermal and photoperiodic regimes vary by latitude. Adaptation to the local thermal and

photoperiodic cycle results in a latitudinal cline in genes associated with circadian rhythms and

development (Hut & Beersma 2011). For species that flower after the summer solstice, like C.

fasciculata, any photoperiods experienced at a particular latitude will occur later in the year in

locations further north. Thus, relocating populations across latitudes could expose them to novel

temperature and photoperiod combinations, which may elicit opposing developmental responses.

 

33    

Relocating populations across large spatial scales may impair attempts at assisted

colonization across latitudes, although less so as temperatures warm. The northern-most

population performed best, even when experiencing the thermal regimes typical of locations

further south. The populations originating from intermediate latitudes displayed the highest

reproductive efficiency (seeds per unit biomass) in the heated treatment. Although warmer

temperatures increased seed production in the southern population by over 800%, reproductive

output was still lower in comparison to all other populations. Warming-induced shifts in

phenological traits were adaptive and most likely contributed to the increased fecundity seen in

all populations. However, selection strongly favored early fruiting in the heated treatment

despite phenological shifts, and simultaneously favored larger plant size. Photoperiodic

constraints could limit further plasticity in these traits.

In C. fasciculata, long day lengths may promote vegetative growth over reproductive

development (Lee & Hartgerink 1986). In this experiment and in others, flowering onset for the

NC population planted at KSR always occurred after August 28th, when the photoperiod in

southern Ontario was 13.75 hours, which is similar to the that experienced upon flowering at the

NC home site on August 7th (data not shown). This occurred despite warming-induced advances

in phenology (this experiment) and even when planting seeds two months ahead of the MN

population (Wadgymar et al. 2015). This supports that flowering onset date is under at least

partial photoperiodic control in C. fasciculata, and that the evolution of genes associated with

photoperiodic responses would be necessary for successful long-term establishment of

populations relocated to northern latitudes.

Scattering seed or planting individuals from multiple populations may inflate genetic

variation and enhance responses to selection of an endangered local population, but only if there

is sufficient overlap in flowering periods of local and colonizing populations. We found almost

no overlap in the flowering schedules of the northern and southern-most population of C.

fasciculata, with little potential for increases in mating opportunities as temperatures warm. For

the northern populations, only the flowers produced during the end of the flowering season

overlapped with open flowers from the NC population. As with many plants (Austen et al.

2015), the probability of fruit set in C. fasciculata declines with later-produced flowers (Lee &

Bazzaz 1982), further reducing the likelihood that any mating opportunities will be realized

 

34    

between these populations. The introgression of genetic material from the southern to northern

population is thus unlikely to occur naturally in the field. Efforts for assisted gene flow

involving species with phenologically distinct populations may require captive breeding

programs to create F2 (or later) generation hybrids in order to produce genotypes with

combinations of thermal and photoperiodic responses that ensure successful establishment.

Many factors can influence the success of assisted colonization beyond those discussed

here, including the availability of hosts (Moir et al. 2012) or mutualists (Keel et al. 2011;

Kranabetter et al. 2012), novel species interactions (Hellmann et al. 2012), competitive

interactions (Stanton-Geddes et al. 2012), or genetic constraints (Etterson & Shaw 2001; Sheldon

et al. 2003; Both & Visser 2005). Plants are at their most vulnerable when in the seedling stage,

as we encountered with our attempt to plant this experiment from seed. As relocations will

likely be carried out with seeds, emergence rates and seedling survival may increase if seeds are

pretreated (e.g. stratified, scarified, inoculated, etc.) prior to planting (McLane & Aitken 2012).

A lack of compatible rhizobia prevented individuals of C. fasciculata from establishing beyond

its northwestern range edge, and inoculations with known strains improved emergence and

growth (Stanton-Geddes & Anderson 2011). While we did not find such limitations in this

experiment, examinations of factors influencing the ecology and evolution of range limits may

further reveal the circumstances under which relocations are likely to succeed.

As the climate continues to warm, assisted colonization may prove to be a viable

adaptation strategy to alleviate risks of extinction or decreased productivity. However, we have

revealed several underappreciated complications in its implementation. For immediate

relocations, populations originating from near the current range boundary may fare best, while

those from too far within the range may do poorly, even in the future thermal regimes of the

newly established site. Plasticity will likely make a strong contribution to initial survival and

establishment, but the long-term success of relocations may ultimately depend on the capacity

for adaptive evolution in the newly founded population, particularly when reproductive

phenologies do not match the photoperiodic conditions of the receiving habitat. Additionally,

among-population differences in flowering phenology may limit the potential for assisted gene

flow in the field. With strategic planning, our results suggest that assisted colonization and

assisted gene flow may be feasible options for preservation.

 

35    

Table 2.1 Linear mixed effects analyses of phenological traits, aboveground biomass, and

growing degree-day (GDD) accumulations for populations planted in both ambient and heated

conditions. F-values are reported for fixed effects in the final, optimized model.

Trait Temperature Population Temperature:Population

Budding onset 74.88*** 322.70*** 2.11+

Flowering onset 112.82*** 388.24*** 2.98*

Fruiting onset 32.95*** 281.54*** 6.25***

Biomass NS 51.39*** NS

Flowering duration 0.48 23.89*** NS

Flower number NS 12.53*** NS

GDD at Budding onset 10.88** 338.92*** 3.25*

GDD at Flowering onset 25.25*** 367.19*** 4.88**

GDD at Fruiting onset 506.66*** 355.40*** 24.45***

Significance: NS Not Significant, p<0.1+, p<0.05*, p<0.01**, p<0.001 Num. df: Treatment 1; Population 3; Treatment:Population 3 Denom. df.: 158-204

 

36    

Table 2.2 Linear mixed effects analyses of independent responses of phenological traits,

aboveground biomass, and growing degree-day (GDD) accumulations for populations planted in

both heated and ambient conditions. Independent responses to warming are indicated by a

significant temperature effect when the previous phenophase (included in parentheses) is

included as a covariate in the model. F-values are reported for fixed effects in the final,

optimized model.

Trait

(covariate) Temperature Population Temperature*Population Covariate

Budding onset

(Emergence date) 75.90*** 329.53*** 2.30† 6.70*

Flowering onset

(Budding onset) 455.28*** 1198.63*** NS 565.93***

Fruiting onset

(Flowering onset) 27.00*** 318.42*** 7.36*** 11.53***

Biomass

(Flowering onset) NS 51.05*** NS 0.23

Flowering duration (Flowering onset) 6.74*** 40.45*** 2.47† 64.41***

Flower number (Flowering onset) NS 12.85*** NS 12.27***

GDD Budding onset

(GDD Emergence date) 10.66** 339.46*** 3.42* 0.01

GDD Flowering onset

(GDD Budding onset) 95.38*** 1212.76*** NS 568.85***

GDD Fruiting onset

(GDD Flowering onset) 689.36*** 387.09*** 22.89*** 17.58***

Significance: NS Not Significant, p<0.1†, p<0.05*, p<0.01**, p<0.001*** Num. df: Treatment 1; Population 3; Treatment:Population 3; Covariate 1 Denom. df: 157-201

 

37    

Table 2.3 A hurdle model demonstrating the effects of flowering onset, fruiting onset, and final

plant size on survival and seed production in populations of Chamaecrista fasciculata planted in

both ambient and artificially warmed conditions. This two-part analysis first models the

probability of surviving to produce seed using a generalized linear model with a binomial

distribution and logit link (zero component). Seed production, excluding zeros, is then modeled

by a negative binomial generalized linear model with log link (count component). Chi-squared

values are reported for fixed effects in the final, optimized model.

df Survival Seed number

Flowering onset 1 NS 3.90*

Fruiting onset 1 -- 21.61***

Biomass 1 7.96** 100.94***

Temperature 1 13.98*** 0.48

Population 3 58.42*** 22.48***

Flowering onset*Temperature 1 NS 6.99**

Fruiting onset*Temperature 1 -- NS

Biomass*Temperature 1 6.77** NS Den. df: Zero 198-203, Count 318-321 Significance: NS: Not Significant, --: not included in model, p<0.1†, p<0.05*, p<0.01**, p<0.001***

 

38    

Table 2.4 Estimates of direct and total phenotypic linear selection coefficients +/- SE for

Chamaecrista fasciculata planted in ambient and artificially warmed conditions. Coefficients

were derived from a hurdle model that examined relationships between phenotypes and seed

number (negative binomial distribution with log link) separately from those of phenotypes and

survival (binomial distribution with logit link). Fruiting onset was not included in the survival

analysis, as survival was scored as the production of at least one fruit. Significant differences in

the strength of selection between treatments are indicated in Table 3. For reference, selection

gradients and differentials derived by multiple regression per Lande and Arnold (1983) are also

reported.

Hurdle model Multiple Regression Survival Seed Number Seed Number

Direct Selection Ambient Heated Ambient Heated Ambient Heated

Flowering onset -0.80 (0.65)

-1.10 (0.84)

-0.45* (0.23)

-0.19 (0.21)

-0.89*** (0.19)

-0.13 (0.21)

Fruiting onset -- -- -0.69*** (0.14)

-0.51** (0.17)

-0.45** (0.15)

-0.58** (0.18)

Biomass 0.25 (033)

1.72** (0.64)

1.00*** (0.14)

0.90*** (0.10)

1.21*** (0.18)

0.81*** (0.10)

Survival Seed number Seed number

Total Selection Ambient Heated Ambient Heated Ambient Heated

Flowering onset -1.19 (0.63)

-0.79 (0.64)

-0.68*** (0.25)

-0.35 (0.25)

-0.70*** (0.10)

-0.32** (0.11)

Fruiting onset -- -- -0.69*** (0.17)

-0.45** (0.16)

-0.50*** (0.13)

-0.33** (0.12)

Biomass 0.37 (0.31)

1.61** (0.51)

0.77*** (0.13)

0.97*** (0.09)

-0.04 (0.12)

0.52*** (0.52)

Significance: NS: not significant, †P<0.1, *P<0.05, **P<0.01, ***P<0.001

 

39    

Figure 2.1 A map of the eastern U.S. showing the northern range limit of Chamaecrista

fasciculata (dashed line), as well as the seed collection sites in Minnesota (MN), Pennsylvania

(PA), Missouri (MO) and North Carolina (NC). The experimental relocation took place in

southern Ontario at the Koffler Scientific Reserve at Joker’s Hill (KSR). The distribution of C.

fasciculata was estimated from herbarium specimens, field observations, communications with

other researchers, and the PLANTS database maintained by the United States Department of

Agriculture.

MN

MO

NC

PA

KSR

N

0 200 400 600km

 

40    

Figure 2.2 Reaction norms showing the mean ± 2 standard errors of (a) reproductive

phenological traits and (b) growing degree day accumulations upon the expression of those traits

in the Minnesota (MN), Pennsylvania (PA), Missouri (MO), and North Carolina (NC)

populations in ambient (A) and artificially heated (H) conditions.

��

��

�

�

�

�

A H A H A H A HMN PA MO NC

40

110

180Da

ys si

nce

plant

ing

��

��

�

�

�

�

�

Fruiting onsetFlowering onsetBudding onset

� �

�

� � �

��

A H A H A H A HMN PA MO NC

500

900

2100

� �

�

� � �

��

�

Fruiting onsetFlowering onsetBudding onset

Grow

ing d

egre

e da

ys °C

·day

(a) (b)

 

41    

Figure 2.3 The mean ± standard error of (a) the log of above ground biomass and (b) the number

of seeds produced in the Minnesota (MN), Pennsylvania (PA), Missouri (MO), and North

Carolina (NC) populations in ambient and artificially warmed conditions.

0

2

4

log V

eget

ative

biom

ass (

g)

AmbientHeated

(a)

MN PA MO NC0

350

700

Numb

er o

f see

ds

AmbientHeated

(b)MN PA MO NC

 

42    

Figure 2.4 Logistic regressions portraying the probability of surviving to produce seed in heated

and ambient conditions as a function of (a) flowering onset or (b) aboveground vegetative

biomass, scaled to a mean of 0 and a standard deviation of 1, per the zero component of the

hurdle model (Table 4). Histograms depict the trait values of individuals that survived (upper

panels) or died (lower panels) from each population.

MN0

20 MN

PA0

20 PA

MO0

20 MO

NC0

20 NC

� �

�� �

�

� �

�

� � �

��

�� �� � � �

�

� �

� �

� �

�

�

�

� �� �

�

� �� �

�

� � �� � ��

�

�

�

�� ��

�

� ��

�

� �� �

�

��

� � �

� �� � �� ���� �

�

� �

�

� �

� � �

�� ���

�

� ���

��

�

�

� �� �

�

�� �

Flowering onset (scaled)!2 0 2 4

� �

�� �

�

� �

�

� � �

��

� �� �� � �

�

��

��

��

�

�

�

��� �

�

� � ��

�

� ��� � ��

�

�

�

����

�

� ��

�

��� �

�

��

� ��

� � ��� �� ����

�

��

�

� �

� ��

�� ���

�

� ���

� �

�

�

�� ��

�

�� �

!2.5 0.5 3.5log Vegetative biomass (scaled, g)

MN0

20 MN

PA0

20 PA

MO0

20 MO

NC0

20 NC

0

20

0

20

0

20

0

20

0

20

0

20

0

20

0

20

0

0.5

1

0

0.5

1

Prob

abilit

y of s

urviv

alSu

rviva

l F

requ

ency

Morta

lityFr

eque

ncy

Prob

abilit

y of s

urviv

al

AmbientHeated AmbientHeated

Survi

val

Fre

quen

cyMo

rtality

Freq

uenc

y

(a) (b)

 

43    

Figure 2.5 Relationships between seed number and (a) flowering onset, (b) fruiting onset, and

(c) the log of vegetative biomass per the partial regression coefficients obtained from the count

component of the hurdle model in Table 4. Note that these relationships are linear on a log scale,

and the response variable was log transformed for ease of viewing and comparison.

���

�

�

�

�

�

�

�

�

�

�

� �

�

� �

�

��

�

�

�

�

�

�

��

��

�

� �

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

��

�

�

�

��

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

�

�

Flowering onset (scaled)log

Num

ber o

f see

ds!2 0 2

0

4

8

� �

�

�

�

�

�

�

�

�

�

�

� �

�

��

�

��

�

�

�

�

�

�

��

��

�

��

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

��

�

�

�

��

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

�

�

log N

umbe

r of s

eeds

!2 0 20

4

8

Fruiting onset (scaled)

��

�

�

�

�

�

�

�

�

�

�

� �

�

��

�

��

�

�

�

�

�

�

��

��

�

� �

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

�

��

�

�

�

��

�

�

�

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

�

�

log N

umbe

r of s

eeds

!2.5 0.5 3.50

4

8

log Vegetative biomass (scaled, g)

� AmbientHeated

(a)

(b)

(c)

� AmbientHeated

� AmbientHeated

 

44    

Figure 2.6 (a) Estimates of the degree of temporal isolation, ρ, between populations of C.

fasciculata in ambient (below diagonal) and artificially heated (above diagonal) conditions, and

population and treatment differences in (b) average flowering duration and (c) total flower

production ± standard error. Estimates of ρ span from 0 (random mating between populations)

to 1 (populations are reproductively isolated). We constructed 95% confidence intervals via

bootstrapping 1000x with replacement. Estimates marked with an asterisk lie outside of the

interval range of the corresponding population comparison in the opposing thermal regime.

MN PA MO NC

NC

MO

PA

MN

0.978 0.950* 0.754

0.675* 0.379* 0.675*

0.341* 0.213* 0.915*

0.469* 0.603* 0.969

0 0.2 0.4 0.6 0.8 1

�Flowe

ring d

urati

on

20

35

50

�

��

�

�

�

�

�

Ambient Heated

100

300

500

Total

numb

er of

flowe

rs

��

�

�

�

�

�

(a)

(b)

�

(c)

Ambient Heated

NCMOPAMN

NCMOPAMN

MN PA MO NC

NC

MO

PA

MN

0.978 0.950* 0.754

0.675* 0.379* 0.675*

0.341* 0.213* 0.915*

0.469* 0.603* 0.969

0 0.2 0.4 0.6 0.8 1

�Flowe

ring d

urati

on

20

35

50

�

��

�

�

�

�

�

Ambient Heated

100

300

500

Total

numb

er of

flowe

rs

��

�

�

�

�

�

(a)

(b)

�

(c)

Ambient Heated

NCMOPAMN

NCMOPAMN

 

45    

References cited

Aitken SN, Yeaman S, Holliday JA, Wang T, Curtis-McLane S (2008) Adaptation, migration or

extirpation: climate change outcomes for tree populations. Evolutionary Applications, 1, 95-

111.

Aitken SN, Whitlock MC (2013) Assisted Gene Flow to Facilitate Local Adaptation to Climate

Change. The Annual Review of Ecology, Evolution, and Systematics, 44, 367-388.

Anderson JT, Inouye DW, McKinney AM, Colautti RI, Mitchell-Olds T (2012) Phenotypic

plasticity and adaptive evolution contribute to advancing flowering phenology in response to

climate change. Proceedings of the Royal Society B: Biological Sciences, 279,

rspb20121051.

Austen EJ, Forrest JRK, Weis AE (2015) Within-plant variation in reproductive investment:

consequences for selection on flowering time. Journal of Evolutionary Biology, 28, 65-79.

Blázquez MA, Ahn JH, Weigel D (2003) A thermosensory pathway controlling flowering time

in Arabidopsis thaliana. Nature Genetics, 32, 168-171.

Bolmgren K, Cowan PD (2008) Time – size tradeoffs: a phylogenetic comparative study of

flowering time, plant height and seed mass in a north-temperate flora. Oikos, 117, 424-429.

Both C, Visser ME (2005) The effect of climate change on the correlation between avian life-

history traits. Global Change Biology, 11, 1606-1613.

Bradshaw WE, Holzapfel CM (2001) Genetic shift in photoperiodic response correlated with

global warming. Proceedings of the National Academy of Sciences of the United States, 98,

14187-14744.

Bradshaw WE, Holzapfel CM (2008) Genetic responses to rapid climate change: it’s seasonal

timing that matters. Molecular Ecology, 17, 157-166.

Bradshaw WE, Holzapfel CM (2010) Physiology of Biotic Response to Rapid Climate Change in

Animals. Annual Review of Physiology, 72, 147-166.

Carpenter KE, Abrar M, Aeby G, et al. (2008) One-Third of Reef-Building Corals Face Elevated

Extinction Risk from Climate Change and Local Impacts. Science, 321, 560-563.

 

46    

Chuine I (2010) Why does phenology drive species distribution? Philosophical Transactions of

the Royal Society B, 365, 3149-3160.

Davis M, Shaw RG (2001) Range Shifts and Adaptive Responses to Quaternary Climate Change.

Science, 292, 673-679.

de Souza Conceição A, de Queiroz LP, Lewis GP, de Andrade MJG, de Almeida PRM,

Schnadelbach AS, van den Berg C (2009) Phylogeny of Chamaecrista Moench

(Leguminosae-Caesalpinioideae) based on nuclear and chloroplast DNA regions. Taxon, 58,

1168-1180.

Donohue K (2002) Germination timing influences natural selection on life-history characters in

Arabidopsis thaliana. Ecology, 83, 1006-1016.

Easterling DR, Horton B, Jones PD, et al., (1997) Maximum and minimum temperature trends

for the globe. Science 277, 364-367.

Ehrlén J (2015) Selection on flowering time in a life-cycle context. Oikos, 124, 92-101.

Elzinga JA, Atlan A, Biere A, Gigord L, Weis AE, Bernasconi G (2007) Time after time:

flowering phenology and biotic interactions. TRENDS in Ecology and Evolution 22, 432-

439.

Etterson JR (2004) Evolutionary potential of Chamaecrista fasciculata in relation to climate

change. II Genetic architecture of three populations reciprocally planted along an

environmental gradient in the great plains. Evolution, 58, 1459-1471.

Etterson JR, Shaw RG (2001) Constraint to Adaptive Evolution in Response to Global Warming.

Science, 294, 151-154.

Farnsworth EJ, Bazzaz FA (1995) Inter- and intra-generic differences in growth, reproduction,

and fitness of nine herbaceous annual species grown in elevated CO2 environments.

Oecologia, 104, 454-466.

Fenster CB (1991) Gene flow in Chamaecrista fasciculata (Leguminosae) I. Gene dispersal.

Evolution, 45, 398-409.

 

47    

Fenster CB, Vekemans X, Hardy OJ (2003) Quantifying gene flow from spatial genetic structure

data in a metapopulation of Chamaecrista fasciculata (Leguminosae). Evolution, 57, 995-

1007.

Fitter AH, Fitter RSR (2002) Rapid Changes in Flowering Time in British Plants. Science, 296,

1689-1691.

Foote LE, Jackobs JA (1966) Soil factors and the cocurence of partridge pea (Cassia fasciculata

Michx.) in Illinois. Ecology, 47, 968-975.

Franks SJ, Sim S, Weis AE (2007) Rapid evolution of flowering time by an annual plant in

response to a climate fluctuation. PNAS, 104, 1278-1282.

Frei ER, Ghazoul J, Matter P, Heggli M, Pluess AR (2014) Plant population differentiation and

climate change: responses of grassland species along an elevational gradient. Global Change

Biology, 20, 441-455.

Foote LE, Jackobs JA (1996) Soil factors and the occurrence of partridge pea (Cassia fasciculata

Michx.) in Illinois. Ecology, 6, 968-975.

Galloway LF, Burgess KS (2009) Manipulation of Flowering Time: Phenological Integration and

Maternal Effects. Ecology, 90, 2139-2148.

Galloway LF, Etterson JR (2007) Transgenerational plasticity is adaptive in the wild. Science,

318, 1134-1136.

Galloway LF, Fenster CB (2000) Population differentiation in an annual legume: local

adaptation. Evolution, 54, 1173-1181.

Grady KC, Ferrier SM, Kolb TE, Hart SC, Allan GJ, Whitham TG (2011) Genetic variation in

productivity of foundation riparian species at the edge of their distribution: implications for

restoration and assisted migration in a warming climate. Global Change Biology, 17, 3724-

3735.

Griffith TM, Watson MA (2006) Is Evolution Necessary for Range Expansion? Manipulating

Reproductive Timing of a Weedy Annual Transplanted beyond Its Range. The American

Naturalist, 167, 153-164.

 

48    

Haggerty BP, Galloway LF (2011) Response of individual components of reproductive

phenology to growing season length in a monocarpic herb. Journal of Ecology, 99, 242-253.

Hellmann JJ, Prior KM, Pelini S (2012) The influence of species interactions on geographic

range change under climate change. Annals of the New York Academy of Sciences, 1249, 18-

28.

Hewitt N, Klenk N, Smith AL, Bazely DR, Yan N, Wood S, MacLellan JI, Lipsig-Mumme C,

Henriques I (2011) Taking stock of the assisted migration debate. Biological Conservation,

144, 2560-2572.

Hill JK, Griffiths HM, Thomas CD (2011) Climate Change and Evolutionary Adaptations at

Species’ Range Margins. Annual Review of Entomology, 56, 143-158.

Hobbs RJ, Higgs E, Harris JA (2009) Novel ecosystems: implications for conservation and

restoration. Trends in Ecology and Evolution, 24, 599-605.

Hut RA, Beersma DGM (2011) Evolution of time-keeping mechanisms: early emergence and

adaptation to photoperiod. Philosophical Transactions of the Royal Society B, 366, 2141-

2154.

Irwin HS, Barneby RC (1982) The American Cassiinae: A synoptical revision of Leguminosae

tribe Cassieae subtribe Cassiinae in the New World. Memoirs of the New York Botanical

Garden, 35, The New York Botanical Garden, New York.

Keel BG, Zettler LW, Kaplin BA (2011) Seed Germination of Habenaria repens (Orchidaceae)

in situ Beyond its Range, and its Potential for Assisted Migration Imposed by Climate

Change. Castanea, 76, 43-54.

Kim E, Donohue K (2011) Population differentiation and plasticity in vegetative ontogeny:

effects on life-history expression in Erysimum capitatum (Brassicaceae). American Journal

of Botany, 98, 1752-1761.

Kimball BA, Conley MM, Wang S, Xingwu L, Luo C, Morgan J, Smith D (2008) Infrared heater

arrays for warming ecosystem field plots. Global Change Biology, 14, 309-320.

Kobayashi Y, Weigel D (2007) Move on up, it’s time for change – mobile signals controlling

photoperiod-dependent flowering. Genes & Development, 21, 2371-2384.

 

49    

Kranabetter JM, Stoehr MU, O’Neill GA (2012) Divergence in ectomycorrhizal communities

with foreign Douglas-fir populations and implications for assisted migration. Ecological

Applications, 22, 550-560.

Kreyling J, Bittner T, Jaeschke A, Jentsch A, Steinbauer MJ, Thiel D, Beierkuhnlein C (2011)

Assisted Colonization: A Question of Focal Units and Recipient Localities. Restoration

Ecology, 19, 433-440.

Lande R, Arnold SJ (1983) The Measurement of Selection on Correlated Characters. Evolution,

37, 1210-1226.

Lawler JJ, Olden JD (2011) Reframing the debate over assisted colonization. Frontiers in

Ecology and the Environment, 9, 569-574.

Lee TD, Bazzaz FA (1982) Regulation of Fruit Maturation Pattern in an Annual Legume, Cassia

fasciculata. Evolution, 63, 1374-1388.

Lee TD, Hartgerink AP (1986) Pollination intensity, fruit maturation pattern, and offspring

quality in Cassia fasciculata (Leguminosae). In: Biotechnology and ecology of pollen (ed.

Mulcahy DL, Mulcahy GB, Ottaviano E). pp. 417-422, Springer, New York, 1986,

Leech SM, Aluedo PL, O’Neill G (2011) Assisted Migration: Adapting forest management to a

changing climate. BC Journal of Ecosystems and Management, 12, 18-34.

Lenoir J, Gégout JC, Marquet PA, de Ruffray P, Brisse H (2008) A significant upward shift in

plant species optimum elevation during the 20th century. Science, 320, 1768-1771.

Loss SR, Terwilliger LA, Peterson AC (2011) Assisted colonization: Integrating conservation

strategies in the face of climate change. Biological Conservation, 144, 92-100.

Lu P, Parker WH, Cherry M, Colombo S, Parker WC, Man R, Roubal N (2014) Survival and

growth patterns of white spruce (Picea glauca [Moench] Voss) rangewide provenances and

their implications for climate change adaptation. Ecology and Evolution, 4, 2360-2374.

Lunt ID, Byrne M, Hellmann JJ, et al., (2013) Using assisted colonisation to conserve

biodiversity and restore ecosystem function under climate change. Biological Conservation,

157, 172-177.

 

50    

McLane SC, Aitken SN (2012) Whitebark pine (Pinus albicaulis) assisted migration potential:

testing establishment north of the species range. Ecological Applications, 22, 142-153.

McLachlan JS, Hellmann JJ, Schwartz MW (2007) A Framework for Debate of Assisted

Migration in an Era of Climate Change. Conservation Biology, 21, 297-302.

Merilä J, Hendry AP (2014) Climate change, adaptation, and phenotypic plasticity: the problem

and the evidence. Evolutionary applications, 7, 1-14.

Miller P, Lanier W, Brandt S (2001) Using growing degree days to predict plant stages. Montana

State Unviersity, USA, Extension Service.

Mitchell-Olds T, Shaw RG (1987) Regression analysis of natural selection: statistical inference

and biological interpretation. Evolution, 41, 1149-1161.

Moir ML, Vesk PA, Brennan KEC, Poulin R, Hughes L, Keith DA, McCarthy MA, Coates DJ.

(2012) Considering Extinction of Dependent Species during Translocation, Ex Situ

Conservation, and Assisted Migration of Threatened Hosts. Conservation Biology, 26, 199-

207.

Morison JIL, Lawlor DW (1999) Interactions between increasing CO2 concentration and

temperature on plant growth. Plant, Cell and Environment, 22, 659-692.

Mueller JM, Hellmann JJ (2008) An Assessment of Invation Risk from Assisted Migration.

Conservation Biology, 22, 562-567.

Neuheimer AB, Taggart CT (2007) The growing degree-day and fish size-at-age: the overlooked

metric. Canadian Journal of Fisheries and Aquatic Science, 64, 375-385.

Neuheimer AB, Grønkjær P (2012) Climate effects on size-at age: growth in warming waters

compensates for earlier maturity in an exploited marine fish. Global Change Biology, 18,

1812-1822.

Nooten SS, Andrew NR, Hughes L (2014) Potential Impacts of Climate Change on Insect

Communities: A Transplant Experiment. PLoS ONE, 9, e85987.

doi:10.1371/journal.pone.0085987

 

51    

OMNR: Colombo SJ, McKenney DW, Lawrence KM, & Gray PA (2007) Climate change

projections for Ontario: practical information for policymakers and planners. Ontario

Ministry of Natural Resources, Queen’s Printer for Ontario.

Parmesan C (2006) Ecological and Evolutionary Responses to Recent Climate Change. Annual

Review of Ecology, Evolution, and Systematics, 37, 637-669.

Parmesan C (2007) Influences of species, latitudes, and methodologies on estimates of

phenological response to global warming. Global Change Biology, 13, 1860-1872.

Pedlar JH, McKenney DW, Aubin I, et al., (2012) Placing Forestry in the Assisted Migration

Debate. BioScience, 62, 835-842.

Perry AL, Low PJ, Ellis JR, Reynolds JD (2005) Climate Change and Distribution Shifts in

Marine Fishes. Science, 308, 1912-1915.

Pinheiro J, Bates D, DebRoy S, Sarkar D (2014) nlme: Linear and nonlinear mixed effects

models. (R package, version 3.1-117)

Post ES, Pedersen C, Wilmers CC, Forchhammer MC (2008) Phenological sequences reveal

aggregate life history response to climatic warming. Ecology, 89, 363-370.

R Development Core Team (2014) R: A language and environment for statistical computing. R

Foundation for Statistical Computing. Vienna, Austria, ISBN 3-900051-07-0. Available at

http://www.R-project.org.

Reekie EG, Bazzaz FA (1991) Phenology and Growth in four annual species grown in ambient

and elevated CO2. Canadian Journal of Botany, 69, 2475-2481.

Ricciardi A, Simberloff D (2009) Assisted colonization is not a viable conservation strategy.

Trends in Ecology and Evolution, 24, 248-253.

Rice KJ, Emery NC (2003) Managing microevolution: restoration in the face of global change.

Frontiers in Ecology and the Environment, 1, 469-478.

Sheldon BC, Kruuk LEB, Merilä J (2003) Natural selection and inheritance of breeding time and

clutch size in the collard flycatcher. Evolution, 57, 406-420.

Sherry RA, Zhou X, Gu S, Arnone III JA, Schimel DS, Verburg PS, Wallace LL, Luo Y (2007)

Divergence of reproductive phenology under climate warming. PNAS, 104, 198-202.

 

52    

Stanton-Geddes J, Anderson CG (2011) Does a facultative mutualism limit species range

expansion? Oecologia, 167, 149-155.

Stanton-Geddes J, Tiffin P, Shaw RG (2012) Role of climate and competitors in limiting fitness

across range edges of an annual plant. Ecology, 93, 1604-1613.

Stevenson IR, Bryan DM (2000) Avian Phenology: Climate change and constraints on breeding.

Nature, 406, 366-367.

Tallmon DA, Luikart G, Waples RS (2004) The alluring simplicity and complex reality of

genetic rescue. TRENDS in Ecology and Evolution, 19, 489-496.

Thomas CD, Cameron A, Green RE, et al. (2004) Extinction risk from climate change. Nature,

427, 145-148.

Trudgill DL, Honek A, Li D, Van Straalen NM (2005) Thermal time – concepts and utility.

Journals of Applied Biology, 146, 1-14.

Vitt P, Havens K, Kramer AT, Sollenberger D, Yates E (2010) Assisted migration of plants:

Changes in latitudes, changes in attitudes. Biological Conservation, 143, 18-27.

Wadgymar SM, Austen EJ, Cumming MN, Weis AE (2015) Simultaneous pulsed flowering in a

temperate legume: causes and consequences of multimodality in the shape of floral display

schedules. Journal of Ecology, 103, 316-327.

Wang JY (1960) A Critique of the Heat Unit Approach to Plant Response Studies. Ecology, 41,

785-790.

Weis AE, Kossler TM (2004) Genetic variation in flowering time induces phenological

assortative mating: quantitative genetic methods applied to Brassica rapa. American Journal

of Botany, 91, 825-836.

Weis AE, Nardone E, Fox GA (2014) The strength of assortative mating for population date and

its basis in individual variation in flowering schedule. Journal of Evolutionary Biology, 27:

2138:2151.

Weis AE (2015) On the potential strength and consequences for non-random gene flow caused

by local adaptation in flowering time. Journal of Evolutionary Biology, 28, 699-714.

 

53    

Willis SG, Hill JK, Thomas CD, Roy DB, Fox R, Blakeley DS, Huntley B (2009) Assisted

colonization in a changing climate: a test-study using two U.K. butterflies. Conservation

Letters, 2, 45-51.

Yee TW (2010) VGAM: Vector Generalized Linear and Additive Models. (R package, version

0.7-10)

Zhang R, Post E, Shea K (2012) Warming leads to divergent responses but similarly improved

performance of two invasive thistles. Population Ecology, 54, 583-589.

Zhu K, Woodall CW, Ghosh S, Gelfand AE, Clark JS (2013) Dual impacts of climate change:

forest migration and turnover through life history. Global Change Biology, 20, 251-264.

Zuur AF, Ieno EN, Walker NJ, Saveliev AA, Smith GM (2009) Mixed effects models and

extensions in Ecology with R. Springer New York, New York.

 

54    

Chapter 3

Simultaneous pulsed flowering in a temperate legume: causes and consequences of multimodality in the shape of floral display schedules

This chapter resulted from collaboration with Emily J. Austen, Matthew N. Cumming, and

Arthur E. Weis. Susana M. Wadgymar carried out the experiments, performed the analyses, and

wrote the manuscript. MNC assisted with fieldwork while EJA and AEW contributed to ideas

and manuscript editing. This manuscript has been accepted for publication in the Journal of

Ecology.

Abstract In plants, the temporal pattern of floral displays, or display schedules, delimits an

individual’s mating opportunities. Thus, variation in the shape of display schedules can affect

the degree of population synchrony and the strength of phenological assortative mating by

flowering onset date. A good understanding of the mechanisms regulating the timing of

flowering onset has been developed, but we know less about factors influencing subsequent

patterns of floral display.

We observed unusual multimodal display schedules in temperate populations of the

annual legume Chamaecrista fasciculata. Here we ask if ‘flowering pulses’ are simultaneous

among individuals and populations and explore potential underlying mechanisms and

consequences of pulsed flowering.

We monitored daily flower production for individual plants from genetically divergent

populations during a series of field experiments that manipulated three potential influencers of

display schedule shape: average daily temperature, pollinator availability, and watering

schedules. We measured floral longevity to isolate the contributions of flower retention and

flower deployment to display schedules. We assessed relationships between flowering and

environmental variables and compared estimates of population synchrony, individual synchrony,

 

55    

and the strength of assortative mating with those of 29 unimodally-flowering species from the

area.

We observed simultaneous flowering pulses in all experiments, with peaks aligned

among individuals and populations despite variation in flowering onset and/or duration. Pulses

were not the result of increases in average temperature, pollinator availability, or variation in

watering schedules. Seasonal fluctuations in temperature correlated with floral longevity and

flower deployment, suggesting that the shape of display schedules may be plastic in response to

temperature. Average population and individual synchrony differed only slightly from those of

the species with unimodal schedules, while the average strength of assortative mating for

flowering onset date was strongly reduced (0.21 in C. fasciculata vs. 0.35 for the 29 other

species).

Researchers should take caution in assuming that components of display schedules are

genetically or developmentally correlated with flowering onset. Variation in the shape of display

schedules can influence patterns of gene-flow within or between populations, with potential

effects on the strength of phenological assortative mating and subsequent responses to selection.

Introduction

The opportunities for pollen exchange among plants are dependent on temporal patterns

of flower production, or flowering phenology. Coupled with other factors, including the

composition of the pollinator community, the spatial layout of members of the population, or the

mating system of a species, synchrony in flower production among individuals can affect

outcrossing rates within populations (Loveless & Hamrick 1984; Young 1988; Ims 1990; Ison et

al. 2014). Individual variation in patterns of flowering determines the potential for nonrandom

mating among plants with distinct flowering onset dates (phenological assortative mating),

potentially influencing the efficacy of selection on flowering onset and correlated traits (Weis &

Kossler 2004; Weis 2005). Despite these far-reaching effects, we have a limited understanding

of the factors affecting the schedule of flowers on display across the season, or how variation in

the shape of these display schedules influences the temporal structure of a plant’s mating pool.

 

56    

Species in temperate regions tend to flower in a unimodal fashion over the span of

several weeks, with flower production increasing at a rapid rate to a maximum, or peak, display

size, followed by a steady decline in flower number as all internal resources are diverted away

from flower production and towards maturing fruit (Rabinowitz 1981; Herrera, 1986; Weis et al.

2014). Display schedules of this shape have been described by algebraic functions that estimate

peak flowering dates and the dispersion of flower production over the season (Malo 2002; Clark

and Thompson 2011). However, in perennials, the shape of individual- or population-level

display schedules can be variable across years (Picó & Retana, 2001). This suggests that

plasticity in the symmetry of display schedules (skew), the magnitude of peak flowering

(kurtosis), the number of days where flowers are produced (duration), or the number of flowering

peaks (modality) may be more common than currently appreciated. In fact, a close examination

of cases where individual display schedules have been tallied typically reveals short-term

fluctuations in flower number about a smoother, underlying seasonal pattern (Malo 2002; Clark

& Thompson 2010; Austen et al. 2014; Weis et al. 2014). This phenological ‘noise’ may be the

result of immediate developmental responses to factors influencing flower deployment (the

opening of new flowers) or floral longevity (the retention of previously open flowers), which

together comprise the flowers on display each day. Fluctuations in display and deployment

schedules may occur simultaneously across individuals, indicating a shared plastic response to

the same external stimuli. But what causes these day-to-day fluctuations?

The display schedule can be sensitive to many factors, including environmental

conditions or resource availability (Bustamante & Búrquez 2008), resource partitioning among

vegetative, defensive, and reproductive functions (Bazzaz et al. 1987), or meristem availability

and allocation (Bonser & Aarssen 1996). The contribution of these factors can be teased apart

experimentally (Diggle 1995); however understanding the proximal mechanisms underlying

schedule shape presents several challenges. Subtle variation in schedule shape may only be

detected with fine-scale monitoring of flowering at the individual level (Miller-Rushing et al.

2008; Morellato et al. 2010). Display schedules may be influenced by environmental factors that

present both seasonal trends and daily fluctuation (e.g. precipitation, temperature, etc.). These

time-series data are often messy and autocorrelated, requiring special methods for analyzing

relationships between variables (Hudson 2010; Brown et al. 2011). Additionally, the daily

environment can influence both flower deployment and floral longevity (Augspurger 1983;

 

57    

Primack 1985), requiring fine-scale data to distinguish between the contributions of each to the

display schedule.

Temporal shifts in internal resource allocation can also affect the shape of display and

deployment schedules (Stephenson 1981; Stanton et al. 1987). Plants face a trade-off between

current reproduction (resource investment in seed maturation) and future reproduction (resource

investment in the production of new flowers). The potential number of flowers deployed in a

given day can be inversely related to the number of fruit being matured (Primack 1978; Lloyd

1980), where a decrease in flower number in the deployment schedule reflects a temporary

resource shift to fruit maturation after a period of effective pollination (Stephenson 1981). This

phenomenon has been observed in the tropics, where a lack of seasonality permits flowering

year-round and display schedules are often multimodal (Newstrom et al. 1994). The dependence

of flower deployment on internal resource availability can be detected when monitoring flower

number where pollinator services are limited or absent and few to no fruit are being matured.

Fluctuations in display schedules influence the probability of pollen exchange between

any two individuals: plants that fluctuate in synchrony will share more mating opportunities than

those fluctuating independently. The shape of display schedules dictates the degree of

phenological synchrony within and among individuals, which in turn can influence rates of

outcrossing and selfing via pollinator movements, or geitonogamy. Synchrony in display

schedules is an essential requirement for random mating in plants. Thus, variation in schedule

shape may also alter the strength of phenological assortative mating by flowering onset date in a

population (Fox 2003), although this has yet to be formally tested.

In this paper we examine mechanisms contributing to, and consequences of, multimodal

display schedules in temperate populations of the annual legume Chamaecrista fasciculata

(Michx.). The occurrence of multiple flowering peaks, or pulsed flowering, is unusual at these

latitudes and offers an opportunity to examine the factors that shape display schedules and how

variation in shape ultimately influences synchrony and phenological assortative mating. We aim

to (1) verify pulsed flowering in Chamaecrista fasciculata, (2) determine whether flowering

pulses occur simultaneously across individuals and genetically differentiated populations, (3)

assess whether flowering pulses, and the intervals between them, are the result of intermittent

 

58    

shifts in internal resource allocation away from flower production and toward fruit maturation,

(4) establish whether flowering pulses correlate with fluctuations in abiotic variables, and (5)

evaluate effects of flowering pulses on synchrony and assortative mating for flowering onset

date. To accomplish this, we account for variation in floral longevity to distinguish between

display schedules, which include flowers of any age available for pollination, and deployment

schedules, which involve the opening of new flowers each day.

Methods

Study species

Chamaecrista fasciculata (Fabaceae, subfamily Caesalpiniodeae), or the partridge pea, is

a self-compatible annual legume that prefers sandy soils in prairie and disturbed habitats (Foote

& Jackobs 1966). Its distribution spans the eastern half of the United States and Mexico, with

the northern range limit running along the Canadian border from Minnesota to New York (Irwin

& Barneby 1982).

Plants consist of a central stalk with several branches that each develop multiple

compound racemes (Garrish & Lee 1989). The flower buds produced on each raceme can be

held in stasis for 4 to 10 days until blooming, resulting in multiple buds awaiting anthesis at the

same time. Flowering and growth continue until first frost, and plants typically generate 100 to

800 flowers over the course of 30 to 60 days. Flowers produce no nectar, are exclusively buzz-

pollinated, and are reported to remain open for just one day (Thorp & Estes 1975).

We collected seeds from five populations of C. fasciculata in the fall of 2009. Two were

from the U.S. midwestern states of Minnesota (MN, 44.8011°N, 92.9647°W) and Missouri (MO,

38.4979°N, 90.5610°W) while three were from the eastern states of Pennsylvania (PA,

40.1790°N, 76.7248°W), Virginia (VA, 37.5061°N, 77.7342°W), and North Carolina (NC,

35.8900°N, 79.0092°W). Where possible, three fruit were collected from each of 50-100 plants

located at least 5 m apart along a transect (the approximate genetic neighborhood size for this

species, Fenster 1991).

Summary of experiments

 

59    

We collected daily flower counts for individuals from all populations of Chamaecrista

over the course of three common garden experiments conducted over two years in a field setting

(Table 3.1). This paper explores the patterns of flower deployment and floral longevity in these

experiments; data addressing other questions will be reported elsewhere.

All experiments took place at the University of Toronto’s field station, the Koffler

Scientific Reserve at Joker’s Hill (KSR, 44.0300°N, 79.5275°W). This site is just north of

Chamaecrista’s current distribution limit in eastern North America, but just within the latitudinal

limit west of the Great Lakes. Each experiment included treatments that ultimately manipulated

aspects of flowering phenology (Table 3.1). In each study, seedlings that had been planted on

the same day were transplanted into the field 20 cm apart in a hexagonal array with a ring of

equally spaced plants around focal individuals to absorb edge affects. All other competitors

within the plots were cleared. Unless otherwise noted, all experiments began in May. With few

exceptions (detailed below), flowers were counted on each individual every day until a killing

frost occurred. Daily precipitation data were collected from a rain gauge monitored by

Environment Canada at the nearby Buttonville Airport (43.8608°N, 79.3686°W) while

temperature and humidity measurements were recorded by a weather station at KSR (HC-S3

probe, Campbell Scientific, Edmonton, Canada).

In experiment 1 we manipulated thermal regimes in order to extend the growing season to

that of a latitude approximately 5 degrees further south. Temperature has been shown to strongly

advance flowering onset dates in many species (Parmesan 2007), however no studies have

monitored subsequent patterns of flowering to see if display schedules are similarly affected.

We used infrared heaters to warm 3-metre diameter plots by a desired amount above ambient

(design per Kimball et al. 2008). Temperatures were monitored at the plot level with infrared

radiation scans (SI-111 infrared radiometer, Campbell Scientific, Edmonton, Canada). Six of

these plots were heated by 1.5˚C during the day and 3˚C at night, in accordance with diurnal

warming projections (Easterling et al. 1997) and local warming predictions (OMNR 2007), while

six identical plots were unheated. Heated and ambient plots were otherwise exposed to natural

conditions. Each plot contained five randomly selected individuals from each of four

populations.

 

60    

In experiment 2 we manipulated pollinator access to plants, and thus resource allocation

to fruit/seed maturation. This study took place in a field that had been undisturbed for two

decades. For each population, seven individuals were planted within each of six plots that either

allowed or excluded pollinators. Pollinator excluded plots were covered with a tent made of

fine-mesh bridal tulle to prevent pollination. We hung a sheet of this netting on the south (sun-

facing) side of plots open to pollinators to account for any shading affects.

In experiment 3 we manipulated watering schedules to determine whether variation in

water availability influences the shape of display schedules. We randomly assigned 12 plots to 1

of 3 watering regimes; ~91 liters of water applied every 2 weeks, ~45.5 liters of water applied

every week, or ~13 liters of water applied every 2 days. Thus, all plots received the same total

volume of water, but at different schedules. We staggered planting dates to obtain concurrently

flowering cohorts from select populations that otherwise have minimal flowering overlap (Table

1). We planted seven seedlings from each population-cohort combination (hereafter simply

referred to as population) per plot, each in their own quadrant to avoid asymmetric competition

among planting groups. We excluded natural precipitation by covering plots with 2.7x3.7 meter

roofs made of clear plastic, slanted southward towards the direction of summer rain events. Rain

gutters along the southern edges directed precipitation away from the plots. We measured the

percent volumetric water content (VWC) within each quadrant of each plot for a portion of the

days where flowers were counted (TDR 100 Soil Moisture Meter, Spectrum Technologies, Inc.,

Illinois, USA).

In each experiment, we counted flowers on all individuals every day. In all, we tallied

over 148,000 flower observations. Experiments 1 and 2 had days of missing data (7 of 93 and 8

of 86 days, respectively). We interpolated expected flower counts from a linear function running

from the day before to the day after the missed counts. There were never more than two

consecutive days without data collection, so our interpolations were unlikely to distort the true

pattern of flowering. Interpolated data were used in all graphs and analyses.

In experiment 3, we measured floral longevity for a subset of consecutive days. On a

given day, every flower on each individual was marked with a felt tip marker on the inside of the

rigid upper petal. On subsequent days, the number of flowers remaining open and unwithered

 

61    

with colors from previous days was recorded and new flowers were counted and marked with a

different color. We used these data to measure floral longevity and to determine the proportion

of newly deployed flowers contributing to display schedules. In total, the longevities of 7011

flowers were monitored. Markings did not seem to deter pollinators from visiting flowers or

produce any adverse reactions in the flowers themselves (personal observation).

Comparison of flowering phenologies among populations

For data from experiment 2, we used linear mixed models to determine whether

pollination treatment influenced the total number of flowers produced or the flowering duration

in all populations. We included pollination treatment, population, and their interaction as fixed

effects and plot as a random effect. In these analyses, and in subsequent models, we account for

any variance heterogeneity among groups with error variance covariates per Zuur et al. (2009)

using the nlme package (Pinheiro et al. 2014) in R (R Development Team, 2005).

We formally assessed whether flowering pulses occurred simultaneously across

populations by examining the cross-correlation functions between pairs of populations, which

produces correlation coefficients between time series that are aligned or shifted (lagged) by a

certain number of days. This analysis is not a complete estimate of synchrony among

populations; rather, it correlates patterns of flowering between populations only for the period of

time where both were in flower. All correlation analyses were calculated using the proportion of

total flowers in bloom each day, which standardizes flowering output across populations with

different display sizes. We chose to compare populations that vary in flowering onset date,

duration, genetic origin, treatments within experiments, and across experiments conducted in the

same year in order to capture the extent of phenological concordance between distinct groups.

Cross-correlations between time series that are themselves autocorrelated can result in

inflated variances that produce erroneously large cross-correlation coefficients (Zuur et al. 2009;

Brown et al. 2011). To account for autocorrelation in any of the phenological data, we applied

auto-regressive integrated moving average (ARIMA) models to each time series prior to

calculating cross-correlation coefficients (Box & Jenkins 1970). The order of auto-regressive

and moving average terms were chosen by examining the extended sample autocorrelation

functions for each time series and minimizing the Akaike information criterion (AIC). Only

 

62    

significant coefficients were included in the final models. All ARIMA models were analyzed

using the TSA package (Chan & Ripley 2012) in R (R Development Core Team, 2014).

To test whether flowering pulses were augmenting the phenological correlations between

populations, we compared the cross-correlation functions of our observed data to the cross-

correlation functions of simulated unimodal data. For each population, a simulated, unimodal

display schedule was created by rearranging daily flower counts so that the maximum flowers on

display occurred at the midpoint of the flowering duration, and the remaining flower counts were

arrayed in descending order on either side of the new peak date of flowering. In this way we

preserved the total number of flowers produced, the variation in daily flowering display size

(flower counts), the onset date, and the duration of flowering for each population, and can

compare cross-correlation functions when only the modality of the display schedule had been

altered. As before, we converted data to proportions and employed ARIMA models to remove

autocorrelation prior to each analysis.

If flowering pulses occur simultaneously across populations, we expect to see a strong,

positive correlation coefficient at a lag of 0 days despite differences in flowering onset or

duration between populations. With our multimodal data, we predict that correlations will

decrease rapidly at larger lags, eventually becoming negative, because the flowering peaks of the

display schedules being compared would be misaligned if shifted by more than a day or two. In

contrast, comparisons of unimodal display schedules would produce correlation coefficients that

varied in size and direction at larger lags, depending on the difference in flowering onset and

duration between the populations being compared. Lastly, if the display schedules of different

populations were completely independent, correlation coefficients would be small and less

consistent in sign at all lags and in all comparisons, regardless of schedule modality.

Relationships between flowering phenology and environmental variables

We calculated cross-correlation functions lagged up to 5 days to determine whether

display schedules correlate with average daily temperature, total daily precipitation, and average

daily humidity. Again, we fit phenological and environmental times series with ARIMA models

to remove any autocorrelation from the data. In experiment 3, where natural precipitation was

excluded, we analyzed the relationship between display schedules and the volumetric water

 

63    

content of the soil within each plot. Due to data availability for soil moisture readings, we were

only able to do this for the MN population planted early.

The relationship between temperature and floral longevity in experiment 3 was examined

using logistic regressions. For each population, the mean proportion of flowers open for two

days was regressed on the average temperature for the 24 hours proceeding flower deployment

(calculated here as the average of temperature readings taken every 15 minutes from 8 AM the

day of flower deployment to 8 AM on the subsequent day). We used the average slope and

intercept from these logistic regression equations to calculate the number of newly deployed

flowers each day based on that day’s average temperature. This allowed us to estimate floral

deployment schedules for each population. To examine effects of environmental variables on

patterns of flower deployment, we repeated the cross-correlation analyses between deployment

schedules and abiotic variables.

Where sufficient data were available, we analyzed the effects of watering treatment on

display schedules (for all but the NC late population) and soil moisture content (for the MN early

population) in experiment 3 via generalized least squares fitted models, with watering treatment,

day, and their interaction included as fixed terms (again using the nlme package in R). A

significant interaction term would indicate that the display schedules or soil moisture levels were

variable among watering treatments throughout the growing season. We accounted for temporal

auto-correlation in observations among days (nested within plot) by incorporating an auto-

regressive error structure of order 1 (display schedule analysis) or an exponential correlation

error structure that can account for irregularly spaced observations through time (soil moisture

analysis, Zuur et al. 2009). Models with the lowest AIC values were selected for these analyses.

Synchrony and phenological assortative mating

We estimated population synchrony as per Weis et al. (2014), where synchrony

information is extracted from an n × n matrix of pair-wise mating opportunities, Φ, among all of

the display schedules of studied individuals (n). Each matrix element of Φmf is calculated as the

product of a father f’s proportional contribution to the pollen pool in the population each day of

the flowering season and the number of opportunities for pollen receipt presented by mother, m

(each estimated by their number of open flowers, Weis & Kossler 2004). In a perfectly

 

64    

asynchronous population (i.e. no plant flowers at the same time as any other), the diagonal

elements of Φ are 1/n and the non-diagonal elements are 0. In contrast, in a perfectly

synchronous population (i.e. all individuals display the same number of flowers each day), all

elements equal (1/n)2. The degree of synchrony among plants in a population, Sp, is calculated as

the ratio of the first eigenvalue of the mating matrix, λ1, to the sum of all n eigenvalues:

Sp = λ1 / Σ λ Equation 1

In the case of complete synchrony, all elements of Φ are equal; the first eigenvalue will

be 1/n and the remaining ones will be zero. Thus, Sp = 1 when all plants have identical display

schedules. With a completely asynchronous population, all eigenvalues are equal to 1/n, and per

equation 1, Sp = 1/n. Thus, this measure scales between 1/n and 1. When n is large, 1/n

approaches 0.

We developed a measure of individual synchrony, Si, to quantify the opportunity for

geitonogamous pollen transfer (see Appendix 1, for details). This measure incorporates the

effects of uniformity of display schedules (scaled to total flower production) and flowering

duration, such that Si = CVi / √Di, where CVi is the coefficient of variation of the individual i’s

schedule, and Di is the schedule duration. This equation can be easily rearranged to:

Equation 2

where schedule synchrony is estimated by the sum of the squared deviation of the observed daily

flower production (SSi) relative to the total flowers counted (Ti) and corrected for schedule

duration. A value of Si = 1 indicates that all flowers within a plant could exchange pollen with

every other flower on that plant, while Si = 0 when flowers are distributed evenly across days (SSi

= 0), thus minimizing the opportunities for geitonogamy. Synchrony is technically undefined at

the limit where Di = 1, but we assign this maximum possible synchrony a value of 1. We make

the assumption that any open flower, regardless of age, can receive and contribute pollen

equally, and so conduct calculations on display schedules rather than deployment schedules.

 

65    

The strength of phenological assortative mating for a given trait can be quantified by the

phenotypic correlation between potential mates, ρ (Weis & Kossler 2004). Here we estimate the

potential for assortative mating by flowering onset date; however variation in any component of

schedule shape can influence an individual’s mating pool (Fox 2003). We can characterize ρ by

extracting the proportion of all mating opportunities in a population that occur between two

individuals from the mating matrix, Φ. For hermaphrodites, like C. fasciculata, ρ is:

Equation 3

where z is the date of flowering onset, m and f represent the mother and father of a potential

mating pair, φmf is the element of Φ corresponding to the proportion of mating opportunities

between mother, m, and father, f, and Xm is the proportion of flowers in the population produced

by m. When ρ = 0, the population is mating randomly with respect to flowering onset date. We

make the assumptions that all flowers on display by an individual are equally likely to set seed

and that all flowers open on the same day have the same potential to exchange or receive pollen.

To place our estimates of Sp, Si, and ρ into a broader context, we compare them with

estimates for 29 other species naturally occurring at KSR. Most of these old-field species

exhibited unimodal display schedules that are more typical of temperate regions (Weis et al.

2014; see Fig. A8). The data were collected in 2008 from approximately 50 individuals per

species. Flowers were counted every 3 days, so individual synchrony was estimated with a

modified version of equation 2:

Equation 4

where Ii represents the sampling interval (e.g. Ii = 3 when counts are made every 3 days) and Ci is

the number of days where flowers were counted. We analyzed differences in mean Sp, ρ, and Si

between Chamaecrista populations and the KSR species using two-sample t-tests (if variances

were equal) or Welch’s two-sample t-tests (if variances were unequal).

 

66    

Results

Comparison of flowering phenologies among populations

Population-level flowering pulses were simultaneous across populations and were the

result of simultaneous pulsing at the individual level. Individuals produce one or more flowering

pulses in alignment with their neighbors despite differences in flowering onset and flowering

duration within and among populations or treatments. We first present the observed display

schedules, and below present the cross-correlation analyses that formally test for simultaneity of

flowering pulses.

Consider the example of the MN and MO populations in experiment 1 (Fig. 3.1). Under

the ambient temperature regime, the former began flowering 17 days after the latter, on average,

yet both share a flowering peak on day 234 and another near day 241. This is also true of plants

that were artificially warmed in this experiment, where flowering pulses between heated and

ambient treatments overlap despite the advancement of flowering onset in heated plots.

Flowering pulses were produced simultaneously across the experiment regardless of thermal

treatment or genetic origin. Similar patterns were seen between ambient and heated treatments

within the PA and NC populations (Fig. A1), although the scant temporal overlap in display

schedules precluded simultaneous pulsing between populations.

Overlap among populations in display schedules was enhanced in experiment 2, which

included a pollinator exclusion treatment. Exclusion led to increased resource investment in

flower production (Fig. 3.2), including an extension of flowering duration until the end of the

season for the early-flowering populations (Pollination F(1, 20) = 129.30, P<0.001; Population F(4,

20) = 35.00, P < 0.001; Pollination*Population F(4, 20) = 10.53, P < 0.001), and an increase in the

total number of flowers produced in all populations (Pollination F(1, 20) = 14.51, P < 0.01;

Population F(4, 20) = 1.51, P > 0.05; Pollination*Population F(4, 20) = 1.40, P > 0.05). Multiple,

simultaneous flowering pulses were still produced when resources were not diverted towards

fruit maturation. Thus, flower pulses in display schedules are not a consequence of periodic

diversions of internal resources away from flower production and toward fruit maturation.

 

67    

The artificially extended flowering durations in experiment 2 (Fig. 3.2) enabled us to

observe phenological overlap between populations that were otherwise completely or partially

temporally isolated. Flowering pulses in display schedules appear to be aligned between

pollinated and unpollinated groups, both within and between populations during periods of

overlap. The most striking example is the simultaneous pulsing of the early-flowering MN

population when unpollinated in experiment 2 and the later-flowering NC population from

experiment 1, which was planted 0.5 km away (shown in Fig. 3.2a).

Statistical support for simultaneous pulsing in display schedules is presented in Figure 3a.

Cross-correlation coefficients are shown for the 10 population and treatment comparisons with

sufficient temporal overlap to allow meaningful tests. The display schedules of the various

populations and experimental treatments of Chamaecrista are significantly positively correlated

at lag 0, with a mean correlation coefficient ofr0 = 0.50 (Fig. 3.3a). As predicted, this positive

correlation disappears when time series are misaligned by one day or more. Almost all pairs are

weakly negatively correlated at a lag of 3 days (r3 = -0.15 ), suggesting that 6 days may be the

most common length of time between the pulses of these display schedules. When repeating this

analysis using simulated, unimodal data, all consistent associations among the display schedules

of these groups disappeared (r0 = 0.04, Fig. 3.3b ). Together, these results imply that display

schedules in Chamaecrista are highly synchronized among these genetically differentiated

populations grown under varied thermal and pollination environments, and that this can be

attributed to the simultaneous pulsing of floral displays.

Relationships between flowering phenology and environmental variables

Flower pulses can be the result of temporary increases in either flower deployment or

floral longevity. Logistic regressions revealed a negative relationship between average daily

temperature and floral longevity in experiment 3 (Fig. 3.4, MN early odds ratio = 0.64, Z = -

15.98, P < 0.001; NC early odds ratio = 0.52, Z = -32.22, P < 0.001; MN late odds ratio = 0.62, Z

= -14.49, P < 0.001), with floral longevity increasing sharply from 1 day to 2 days in all

populations when temperatures declined below 16-19ºC. Monitoring floral longevity allowed us

to distinguish newly deployed flowers from all that were on display, and with this distinction we

constructed deployment schedules for each population. Deployment schedules are multimodal,

 

68    

with pulses of deployed flowers occurring simultaneously among populations (Fig. 3.5 and see

Figs A2 and A3). Furthermore, flowers retained from previous days also appear to occur in

pulses. These data suggest that floral longevity is mediated by temperature in this species. Thus,

pulses in display schedules are the result of pulses of deployed flowers, but can be amplified by

the retention of day-old flowers when temperatures are low.

To examine the direct influence of temperature on flower deployment and retention, we

calculated cross-correlation coefficients between average daily temperatures and population-

level display schedules, and repeated analyses with population-level deployment schedules.

When examining all flowers displayed, we observed a negative correlation (higher temperatures,

fewer flowers) at a lag of 1 (r1 = -0.24 ) and a positive correlation at a lag of 4 (r4 = 0.25 ) in

all populations and treatments across experiments (Fig. 3.6a). When accounting for temperature-

mediated floral longevity, we find deployment schedules to be less consistently correlated to

temperatures at a lag of 1 and 4 (r1 = -0.12,r4 = 0.15, respectively, Fig. 3.6b). The fluctuation

in the sign of correlation coefficients as lags increase may reflect the fluctuations in average

daily temperatures found in both years (Fig. 3.5d and see Fig A3k for temperature data). In

many populations, average daily humidity negatively correlated to both display and deployment

schedules at lag 0 (r0 = -0.14 and -0.12, respectively) and lag 4 (r4 = -0.17 and -0.16,

respectively), while there were no consistent relationships between precipitation and display or

deployment schedules (see Figs A4 and A5).

Altering the watering schedule in experiment 3 did not affect the display schedules of any

population (Fig. A6, Day*Treatment MN early F(2, 500) = 0.67, P > 0.05; NC early F(2, 483) = 1.30,

P > 0.05; MN late F(2, 369) = 0.73, P > 0.05). However, there was a strong negative correlation in

all three treatments between VWC and display or deployment schedules at a lag of 0 (r0 = -0.40

and -0.24, respectively), as well as a positive correlation at a lag of 5 (r5 = 0.36 and 0.25,

respectively, Fig. A7). Soil moisture content differed slightly among treatments throughout the

season (Watering treatment F(2, 287) = 3.89, P < 0.05; Day F(1, 287) = 4.42, P < 0.05;

Treatment*Day F(2, 287) = 2.58, P < 0.10), with greater levels of VWC in the two-week treatment

than in the one-week or control treatments. Display schedules would have differed among

treatments if VWC directly influenced flower deployment or retention. It is possible that

 

69    

correlations between display or deployment schedules and VWC are driven by unmeasured

factors that correlate with VWC (e.g. soil porosity).

Synchrony and phenological assortative mating

Average synchrony among populations, Sp, in C. fasciculata was comparable to that of

natural populations of species located at KSR (Sp = 0.63 vs 0.66, respectively; t43 = -0.77, P >

0.05; Fig. 3.7a, b and see Table A1) while the average synchrony within individuals, Si, was

significantly higher (Si = 0.17 vs. 0.14, respectively; t42.5 = 3.41, P < 0.01; Fig. 3.7c, d and see

Table A1), indicating a greater opportunity for geitonogamy.

Multimodal display schedules have the potential to drastically reduce Sp if flowering

pulses are misaligned, and the high levels of synchrony we observed can only be maintained if

individuals pulse concurrently. To confirm this, we shuffled the flowering onset dates of all

individuals in each population in order to randomize the occurrence of flowering pulses among

individuals. We repeated this randomization 1000 times, recalculating Sp for each population,

and compare the average of these estimates to those of the KSR species. When the alignment of

flowering pulses is randomized, average Sp in Chamaecrista significantly decreases to 0.52 (t(36.2)

= -8.24, P < 0.001). Variation in the onset and end dates of individual display schedules can also

influence Sp in C. fasciculata by affecting the frequency with which early-flowering plants

produce flowering pulses outside of the display schedules of late bloomers, and vice versa.

However, the average standard deviation in onset and end dates for populations of Chamaecrista

were not significantly different than those of the KSR species (t(24.3) = 1.47, P > 0.05 and t(48) = -

1.21, P > 0.05, respectively). Together, these results suggest that the levels of Sp observed in

Chamaecrista are equivalent to those from the KSR species because flowering pulses were

produced simultaneously across individuals.

The average strength of phenological assortative mating by flowering onset date was

significantly lower in Chamaecrista than in other species (ρ = 0.21 vs 0.35, respectively; t43 = -

2.63, P < 0.01; Fig. 3.7e, f and see Table A1). When flowering peaks are randomized,ρ

becomes indistinguishable from that of the other species (ρ = 0.39 vs. 0.35, respectively, t(43) =

0.91, P > 0.05). Simultaneous flowering pulses in C. fasciculata may offer more mating

opportunities between early- and late-flowering individuals than typically seen in the

 

70    

phenologies of temperate species, reducing the strength of assortative mating by flowering onset

date.

Discussion

Multimodality and display schedule shape

Display schedules in Chamaecrista fasciculata are multimodal, with flowering pulses

produced simultaneously among individuals and populations. In temperate regions,

simultaneous pulsing has only been demonstrated in several wind-pollinated Juncus species

(Michalski & Durka 2007). Chamaecrista is of tropical descent, perhaps evolving from a rain

forest tree to a savannah shrub prior to its colonization of temperate zones (de Souza Conceição

et al. 2009). Multimodality in display schedules is more common in the tropics (Newstrom et al.

1994), and the unique pattern of flowering found in C. fasciculata may be explained by its

tropical origin. However, in the tropics, simultaneous flowering pulses have only been formally

documented in several Brazilian Myrtaceae species (Proença & Gibbs 1994). It is likely that

examples of simultaneous flowering pulses, and multimodality in general, are rare simply

because few studies have monitored phenology at the level of individuals (Augspurger 1983)

with high enough frequency to capture fluctuations in display schedules (Miller-Rushing et al.

2008; Morellato et al. 2010).

Statistical methods have been developed for describing unimodal display schedules

through fitting flexible regression functions (Malo 2002; Clark & Thompson 2011). However,

adequate regression models for display schedules of other shapes may prove elusive, particularly

if the number of modes is variable and they occur at irregular intervals. Several approaches have

been taken to identify modality in display schedules, including the use of cumulative flowering

density curves to examine bimodality (Aldridge et al. 2011), coefficients of variation to quantify

temporal variability in flower production (Picó & Retana 2001; Michalski & Durka 2007) and

principal coordinates analyses to distinguish between unimodal and bimodal phenologies

(Austen et al. 2014). Each method has its own merits and limitations, and like the time series

analyses used here, many of these approaches cannot yield concrete estimates for the number of

modes or the dates at which they occur (but see Aldridge et al. 2011). The development of

methodology capable of describing variation in modality, and detecting significant departures

 

71    

from unimodality, may be necessary for characterizing many fine- and broad-scale phenological

patterns.

Causes of variation in display and deployment schedules

In C. fasciculata, display schedule shape is likely dictated by environmental conditions.

Temperature affected the shape of display schedules by influencing the life span of individual

flowers (Fig. 4), as is seen in other species (Vesprini & Pacini 2005). This may occur because

cooler temperatures preserve floral tissues (Primack 1985) or because bee activity is also

temperature dependent (Corbet et al. 1993) and flowers are left unpollinated (Blair & Wolfe

2007; Elzinga et al. 2007; Castro et al. 2008). Additional work may distinguish between

mechanisms contributing to variation in floral longevity (Yasaka et al. 1998) and may reveal

whether newly deployed and retained flowers have the potential to contribute equally to fruit

production (i.e. comparable stigma receptivity, pollinator attraction, etc.).

Temperature may also influence display schedules by triggering the opening of flowers

(Fig. 3.6b). Correlations between display schedules and temperature have been found in other

temperate species where the display schedules of individuals and populations can be multimodal

(Picó & Retana 2001; Michalski & Durka 2007). However, this is the first attempt to identify

associations between environmental variables and flower deployment independent of display

schedules. In these studies, and in ours, inter-annual variation in temperature profiles through

the growing season may partially explain variation in modality within and across years. We

observed temperatures at KSR to fluctuate while generally decreasing throughout the growing

season (Fig. 3.5d and see Fig A3k), and our results suggest that the degree of modality in display

schedules of C. fasciculata may be a plastic response to temperature or a correlated variable at

the time of flower bud maturation and flower opening.

Flowering pulses in C. fasciculata are not caused by the intermittent diversion of internal

resources away from flower production and towards fruit maturation, although we found that the

total number of flowers and flowering duration were resource limited (Fig. 3.2). Patterns of

flower and fruit production are inherently linked because both are constrained by a shared

resource pool, and adjustments to resource allocation can be made via changes in flower

production or through seed and fruit abortion (Stephenson 1981). Flowering and fruiting

 

72    

schedules can also be correlated if fruit development times are constant and flowering peaks

produce subsequent pulses of maturing fruit (Rojas-Sandoval & Meléndez-Ackerman 2011).

Such tight correlation does not seem to occur in C. fasciculata, which can hold initiated fruit in

stasis for weeks until an unknown stimulus prompts the selective maturation of some fruit and

abortion of others (Lee & Bazzaz 1982a, b). While patterns of flower and fruit production may

be independent in C. fasciculata, variation in display schedule shape may influence the rate and

timing of fruit and seed dispersal or predation in other species (Mahoro 2002).

Population and individual synchrony

Population synchrony was, on average, indistinguishable from that of the 29 species

studied at KSR (Fig. 3.7a, b), while estimates of individual synchrony for C. fasciculata were

only slightly higher than the KSR species (Fig. 3.7c, d). If display schedule shape were the only

determinant of rates of outcrossing or the occurrence of geitonogamy, this result suggests that

multimodal display schedules at population and individual levels may not alter the potential for

geitonogamy from that of unimodal schedules. However, we might expect large display sizes

(‘pulses’) at the individual level to increase opportunities for geitonogamous selfing if pollinators

move less frequently among individuals (Harder & Barrett 1995). On the other hand, pulsed

flowering across neighboring plants may promote outcrossing if pollinators forage among

individuals with large display sizes. In C. fasciculata, levels of individual and population

synchrony may interact with several aspects of floral morphology shown to reduce geitonogamy,

including herkogamy (Webb & Lloyd 1986), enantiostyly (Fenster 1995; Jesson & Barrett 2002),

and the presence of a stiff hooded petal that acts as a flight guide (Wolfe & Estes 1992). The net

effect of these floral traits and of the shape of population-level display schedules may contribute

to the high outcrossing rates observed in several populations of C. fasciculata (~80%, Fenster

1991).

The display schedules observed here may also partially explain reports of population

structure in Chamaecrista. In this species, pollen movement is localized and populations are

subdivided into small patches of related individuals (Fenster 1991). Flowering pulses in

neighboring plants produce large floral displays that may encourage pollinators to forage

primarily within small groups of individuals. Coupled with short seed dispersal distances

 

73    

(Fenster 1991), the effects of population-level flowering pulses on pollinator movements may

generate this fine-scale population structure.

Phenological assortative mating, natural selection, and schedule shape

Patterns of selection and assortative mating guide the evolution of phenological traits

(Fox 2003). If plant mating is assortative by flowering onset, the genetic variance for flowering

onset (and correlated traits) will be inflated, accelerating responses to natural selection. The

shape of display schedules dictates the potential for genetic exchanges among individuals, and

plasticity in schedule shape can influence the degree of phenological assortative mating within

population (Weis 2005). This, in turn, can alter the effectiveness of selection on phenological (or

correlated) traits (Fox 2003).

We found the average strength of phenological assortative mating by flowering onset date

to be lower in Chamaecrista than in other species found at KSR. Simultaneous flowering pulses

offer greater mating opportunities among individuals with distinct flowering onset dates, i.e.

mating is closer to random. If display schedule shape is partially plastic, as our results suggest,

and mating opportunities are environmentally dictated, the potential evolutionary responses of

flowering onset date to selection may be reduced and may vary among years or among

populations experiencing contrasting environmental conditions.

While the genetic and environmental influences on flowering onset are well known in

several model systems (Mouradov et al. 2002; Putterill et al. 2004), it is often assumed that

components of display schedules are genetically or developmentally correlated with flowering

onset. Here, we have shown evidence that flower display and deployment may be plastic in

response to temperature or a correlated variable in Chamaecrista, producing distinct, multimodal

display schedules in alignment with the thermal regimes typically experienced at KSR. Detailed

phenological data from other systems may reveal that responses to daily fluctuations in the

environment are more widespread than currently appreciated. Chamaecrista fasciculata may

prove to be an excellent candidate for understanding the ecological and evolutionary causes and

consequences of variation in display schedule shape.

 

74    

Table 3.1 A summary of several experiments conducted at the Koffler Scientific Reserve at

Joker’s Hill; the treatments applied, the year of study, the populations involved [Minnesota

(MN), Pennsylvania (PA), Illinois (IL), Missouri (MO), Virginia (VA), and North Carolina

(NC)], and the approximate number of individuals per population and treatment combination

Experiment Treatment Year MN PA MO VA NC n

1 +/- Heat 2011 X X X X 30

2 +/- Pollination 2011 X X X X X 75

3 Watering schedule, +/- Planting date 2012 X X 84

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

75    

Fig. 3.1 (a) Individual-level display schedules from experiment 1. Individuals from the MN and

MO populations from both heated and ambient treatments are staggered along the y-axis in order

of flowering onset date. The size of each circle reflects the proportion of total flowers on display

by an individual on a given day. (b) Population-level display curves for each of the same

population and treatment combinations. The height of these lines reflects the proportion of total

flowers on display by that group on a given day. We show data from a subset of populations for

visual clarity; see Fig. A1 for the remaining data.

Julian date212 232 252 272 292

0.1

0

0

20

40

60

80

100 MN HeatedMN AmbientMO HeatedMO Ambient

Indiv

idual

(a)

(b)

Prop

ortio

n of

flowe

rs pr

oduc

ed

 

76    

Fig. 3.2 The population-level display schedules for populations in experiment 2 from the open

pollination (solid) and pollinator excluded (dotted) treatments. Data is shown for the (a) MN, (b)

PA, (c) MO, (d) VA, and (e) NC populations. In addition, panel (a) includes the display

schedule for the NC population in ambient conditions (dashed) from experiment 1 located ~0.5

km away.

0.0

0.1 (a)PollinatedUnpollinatedNC Ambient

0.0

0.1 (b)

0.0

0.1

Prop

ortio

n of

flowe

rs pr

oduc

ed

(c)

0.0

0.1 (d)

212 232 252 2720.0

0.1

Julian date

(e)

 

77    

Fig. 3.3 Heatmaps summarizing cross-correlation coefficients between the display schedules of

select populations. (a) Cross-correlations for observed, pulsed display schedules. (b) Cross-

correlations for unimodal display schedule simulations. Correlation coefficients were calculated

between time series lagged up to five days. The color and shade of a specific box indicates the

sign and magnitude of the correlation coefficient. An S signifies that the correlation was

significant.

Lag0

Lag1

Lag2

Lag3

Lag4

Lag5

MNLate!NCEarly

VAPollinated!NCUnpollinated

VAPollinated!NCPollinated

PAUnpollinated!MOUnpollinated

MNUnpollinated!PAUnpollinated

MNUnpollinated!NCPollinated

MNUnpollinated!NCAmbient

MNPollinated!MNUnpollinated

MOHeated!MOAmbient

MNHeated!MNAmbient

S S

S

S

S

S

S S

S

S

S

S

Lag0

Lag1

Lag2

Lag3

Lag4

Lag5

S

S S

S

S S

S S S

S

!0.5 !0.25 0 0.25 0.5(a) (b)

 

78    

Fig. 3.4 Logistic regressions relating floral longevity to average daily temperatures in

experiment 3. Data from all watering treatments are combined within a population and planting

time because the treatments did not significantly affect flower production. Each point represents

the proportion of two-day old flowers on a given day, averaged across all individuals in a

population, regressed on the average daily temperatures of the 24 hours preceding flower

deployment. The adjusted r-squared values for MN early, NC early, and MN late are 0.86, 0.98,

and 0.62, respectively (per Naglekerke 1991).

��

�

�

�

�

�

�

�

�

�

�

�

�

�

10 12 14 16 18 20 22 24

0.0

0.5

1.0 MN earlyNC earlyMN late

Average daily temperature (°C)

Prop

ortio

n of

flowe

rs op

en fo

r two

day

s

�

 

79    

Fig. 3.5 The display schedules (a) MN planted early, (b) NC planted early, and (c) MN planted

late populations from experiment 3, and (d) average daily temperatures. Display schedules are

shown with deployment schedules highlighted in dark grey and retained flowers highlighted in

light grey. Deployment schedules were estimated from average daily temperatures using the

logistic regression models shown in Fig. 3.4.

0.00

0.05

0.10

0.00

0.04

0.08

0.00

0.04

0.08

10

15

20

25

RetainedDeployed

(a)

(b)

(c)

(d)

209 219 229 239 249 259 269Julian date

Prop

ortio

n of

flowe

rs pr

oduc

ed d

aily

Temp

erat

ure

(°C)

 

80    

Fig. 3.6 Heatmaps summarizing the cross-correlation coefficients between population-level (a)

display schedules or (b) deployment schedules and average daily temperatures. As in figure 3,

the color and shade of a box reflect the sign and magnitude of the correlation coefficient, and

those with an S indicate significant correlations. For each population, the number of new

flowers each day was estimated from that day’s average temperature using the average

coefficients from the three logistic regression models in Fig. 3.4.

Lag0

Lag1

Lag2

Lag3

Lag4

Lag5

MN LateNC EarlyMN Early

NC UnpollinatedNC Pollinated

VA UnpollinatedVA Pollinated

MO UnpollinatedMO Pollinated

PA UnpollinatedPA Pollinated

MN UnpollinatedMN PollinatedNC AmbientNC Heated

MO AmbientMO HeatedPA AmbientPA Heated

MN AmbientMN Heated

SS

S SS S

S S S SSS S SS S SS SSS S

SS SS

SS

S

Lag0

Lag1

Lag2

Lag3

Lag4

Lag5

SS

S S

S SS

SS

S SS SS

S

!0.5 !0.25 0 0.25 0.5(a) (b)

 

81    

Fig. 3.7 Histograms showing the distribution of estimates of population synchrony (Sp),

individual synchrony (Si), and the strength of phenological assortative mating by flowering onset

date (ρ) for all populations and treatments of C. fasciculata from experiments 1-3 (panels a, c,

and e, respectively) and for the 29 species located at the Koffler Scientific Reserve (panels b, d,

and f, respectively). Calculations were not performed for treatments where pollinators were

excluded. Vertical, dotted lines represent the average estimate in each panel.

 

 

 

 

 

 

 

 

0

7

Freq

uenc

y

(a)

0

11 (c)

0

7 (e)

0.0 0.2 0.4 0.6 0.8 1.00

3

6

Sp

Freq

uenc

y

(b)

0.0 0.2 0.4 0.6 0.8 1.0

0

13

Si

(d)

0.0 0.2 0.4 0.6 0.8 1.0

0

5

!

(f)

 

82    

References  Cited  

Aldridge G, Inouye DW, Forrest JRK, et al. (2011) Emergence of a mid-season period of low

floral resources in a montane meadow ecosystem associated with climate change. Journal of

Ecology, 99, 905-913.

Augspurger CK (1983) Phenology, flowering synchrony, and fruit set of six neotropical shrubs.

Biotropica, 15, 257-267.

Austen EJ, Jackson DA, Weis AE (2014) Describing flowering schedule shape through

multivariate ordination. International Journal of Plant Sciences, 175, 70-79.

Bazzaz FA, Chiariello NR, Coley PD, et al. (1987) Allocating resources to reproduction and

defense. BioScience, 37, 58-67.

Blair AC, Wolfe LM (2007) The association between floral longevity and pollen removal, pollen

receipt, and fruit production in the flame azalea (Rhododendron calendulaceum). Canadian

Journal of Botany, 85, 414-419.

Bonser SP, Aarssen LW (1996) Meristem allocation: a new classification theory for adaptive

strategies in herbaceous plants. Oikos, 77, 347-352.

Box G, Jenkins G (1970) Time series analysis: forecasting and control. Holden-Day, San

Francisco.

Brown CJ, Schoeman DS, Sydeman WJ, et al. (2011) Quantitative approaches in climate change

ecology. Global Change Biology, 17, 3697-3713.

Bustamante E, Búrquez A (2008) Effects of plant size and weather on the flowering phenology

of the organ pipe cactus (Stenocereus thurberi). Annals of Botany, 102, 1019-1030.

Castro S, Silveira P, Navarro L (2008) Effect of pollination on floral longevity and costs of

delayed fertilization in the out-crossing Polygala vayredae Costa (Polygalaceae). Annals of

Botany, 102, 1043-1048.

Chan K, Ripley B (2012) Time series analysis (R package, version 1.01)

Clark RM, Thompson R (2011) Estimation and comparison of flowering curves. Plant Ecology

& Diversity, 4, 189-200.

 

83    

Corbet SA, Fussell M, Ake R, et al. (1993) Temperature and the pollinating activity of social

bees. Ecological Entomology, 18, 17-30.

de Souza Conceição A, de Queiroz LP, Lewis GP, et al. (2009) Phylogeny of Chamaecrista

Moench (Leguminosae-Caesalpinioideae) based on nuclear and chloroplast DNA regions.

Taxon, 58, 1168-1180.

Diggle PK (1995) Architectural effects and the interpretation of patterns of fruit and seed

development. Annual Review of Ecology and Systematics, 26, 531-552.

Easterling DR, Horton B, Jones PD, et al. (1997) Maximum and minimum temperature trends for

the globe. Science 277, 364-367.

Elzinga JA, Atlan A, Biere A, et al. (2007) Time after time: flowering phenology and biotic

interactions. TRENDS in Ecology and Evolution 22, 432-439.

Fenster CB (1991) Gene flow in Chamaecrista fasciculata (Leguminosae) I. Gene dispersal.

Evolution, 45, 398-409.

Fenster CB (1995) Mirror image flowers and their effect on outcrossing rate in Chamaecrista

fasciculata (Leguminosae). American Journal of Botany, 82, 46-50.

Foote LE, Jackobs JA (1996) Soil factors and the occurrence of partridge pea (Cassia fasciculata

Michx.) in Illinois. Ecology, 6, 968-975.

Fox GA (2003) Assortative mating and plant phenology; evolutionary and practical

consequences. Evolutionary Ecology Research, 5, 1-18.

Garrish RS, Lee TD (1989) Physiological integration in Cassia fasciculata Michx.: inflorescence

removal and defoliation experiments. Oecologia, 81, 279-284.

Harder LD, Barrett SCH (1995) Mating costs of large floral displays in hermaphrodite plants.

Nature, 373, 512-515.

Herrera J (1986) Flowering and fruiting phenology in the coastal shrublands of Doñana, south

Spain. Vegetatio, 68, 91-98.

Hudson IL (2010) Interdisciplinary approaches: towards new statistical methods for phenological

studies. Climate Change, 100, 143-171.

 

84    

Ims RA (1990) The ecology and evolution of reproductive synchrony. Trends in Ecology &

Evolution, 5, 135-139.

Irwin HS, Barneby RC (1982) The American Cassiinae: A synoptical revision of Leguminosae

tribe Cassieae subtribe Cassiinae in the New World. Memoirs of the New York Botanical

Garden, 35, The New York Botanical Garden, New York.

Ison JL, Wagenius S, Reitz D, et al. (2014) Mating between Echinacea angustifolia (Asteraceae)

individuals increases with their flowering synchrony and spatial proximity. American

Journal of Botany, 101, 180-189.

Jesson LK, Barrett SCH (2002) Solving the puzzle of mirror-image flowers. Nature, 417, 707.

Kimball BA, Conley MM, Wang S, et al. (2008) Infrared heater arrays for warming ecosystem

field plots. Global Change Biology, 14, 309-320.

Lee TD, Bazzaz FA (1982a) Regulation of fruit and seed production in an annual legume, Cassia

fasciculata. Ecology, 63, 1363-1373.

Lee TD, Bazzaz FA (1982b) Regulation of fruit maturation pattern in an annual legume, Cassia

fasciculata. Ecology, 63, 1374-1388.

Lloyd DG (1980) Sexual strategies in plants I. An hypothesis of serial adjustment of maternal

investment during one reproductive session. New Phytologist, 86, 69-79.

Loveless MD, Hamrick JL (1984) Ecological determinants of genetic structure in plant

populations. Annual Review of Ecology and Systematics, 15, 65-95.

Malo JE (2002) Modelling unimodal flowering phenology with exponential sine equations.

Functional Ecology, 16, 413-418.

Mahoro S (2002) Individual flowering schedule, fruit set, and flower and seed predation in

Vaccinium hirtum Thunb. (Ericaceae). Canadian Journal of Botany, 80, 82-92.

Michalski SG, Durka W (2007) Synchronous pulsed flowering: analysis of the flowering

phenology in Juncus (Juncaceae). Annals of Botany, 100, 1271-1285.

Miller-Rushing AJ, Inouye DW, Primack RB (2008) How well do first flowering dates measure

plant responses to climate change? The effects of population size and sampling frequency.

Journal of Ecology, 96, 1289-1296.

 

85    

Morellato LPC, Camargo MGG, Neves FFD, et al. (2010) The influence of sampling method,

sample size, and frequency of observations on plant phenological patterns and interpretation

in tropical forest trees. Phenological Research, Springer, New York.

Mouradov A, Cremer F, Coupland G (2002) Control of flowering time: interacting pathways as a

basis for diversity. The Plant Cell, S111-S130.

Newstrom LE, Frankie GW, Baker HG (1994) A new classification for plant phenology based on

flowering patterns in lowland tropical rain forest trees at La Selva, Costa Rica. Biotropica,

26, 141-159.

OMNR: Colombo SJ, McKenney DW, Lawrence KM, et al. (2007) Climate change projections

for Ontario: practical information for policymakers and planners. Ontario Ministry of

Natural Resources, Queen’s Printer for Ontario.

Picó FX, Retana J (2001) The flowering pattern of the perennial herb Lobularia maritima: an

unusual case in the Mediterranean basin. Acta Oecologica, 22, 209-217.

Pinheiro J, Bates D, DebRoy S, et al. (2014) nlme: Linear and nonlinear mixed effects models.

(R package, version 3.1-117)

Primack RB (1978) Regulation of seed yield in Plantago. Journal of Ecology, 66, 835-847.

Primack RB (1985) Longevity of individual flowers. Annual Review of Ecology and Systematics,

16, 15-37.

Proença CEB, Gibbs PE (1994) Reproductive biology of eight sympatric Myrtaceae from central

Brazil. New Phytologist, 126, 343-354.

Putterill J, Laurie R, Macknight R (2004) It’s time to flower: the genetic control of flowering

time. BioEssays, 26, 363-373.

Rabinowitz D, Rapp JK, Sork VL, et al. (1981) Phenological properties of wind- and insect-

pollinated prairie plants. Ecology, 62, 49-56.

Rojas-Sandoval J, Meléndez-Ackerman E (2011) Reproductive phenology of the Caribbean

cactus Harrisia portoricensis: rainfall and temperature associations. Botany, 89, 861-871.

Stanton ML, Bereczky JK, Hasbrouck HD (1987) Pollination thoroughness and meternal yield

regulation in wild radish, Raphanus raphanistrum (Brassicaceae). Oecologia, 74, 68-76.

 

86    

Stephenson AG (1981) Flower and fruit abortion: proximate causes and ultimate functions.

Annual Review of Ecology and Systematics, 12, 253-279.

Thorp RW, Estes JR (1975) Intrafloral behavior of bees and flowers of Cassia fasciculata.

Journal of the Kansas Entomological Society, 48, 175-184.

Vesprini JL, Pacini E (2005) Temperature-dependent floral longevity in two Helleborus species.

Plant Systematics and Evolution, 252, 63-70.

Webb CJ, Lloyd DG (1986) The avoidance of interference between the presentation of pollen

and stigmas in angiosperms. II. Herkogamy. New Zealand Journal of Botany, 24, 163-178.

Weis AE, Kossler TM (2004) Genetic variation in flowering time induces phenological

assortative mating: quantitative genetic methods applied to Brassica rapa. American

Journal of Botany, 91, 825-836.

Weis AE (2005) Direct and indirect assortative mating: a multivariate approach to plant

flowering schedules. Journal of Evolutionary Biology, 18, 536-546.

Weis AE, Nardone E, Fox GA (2014) The strength of assortative mating for population date and

its basis in individual variation in flowering schedule. Journal of Evolutionary Biology, 27:

2138:2151.

Wolfe AD, Estes JR (1992) Pollination and the function of floral parts in Chamaecrista

fasciculata (Fabaceae). American Journal of Botany, 79, 314-317.

Yasaka M, Nishiwaki Y, Konno Y (1998) Plasticity in flower longevity in Corydalis Ambigua.

Ecological Research, 13, 211-216.

Young HJ (1988) Neighborhood size in a beetle pollinated tropical aroid: effects of low density

and asynchronous flowering. Oecologia, 76, 461-466.

Zuur AF, Ieno EN, Walker NJ, et al. (2009) Mixed effects models and extensions in Ecology

with R. Springer New York, New York.

 

 

 

 

 

87    

Appendix  A  

Supplemental  Information  for  Chapter  3  

 

Figure  A1    (a)  Individual-­‐level  display  schedules  from  experiment  1  not  included  in  the  main  text.    Individuals  from  the  Pennsylvania  (PA)  and  North  Carolina  (NC)  populations  in  both  heated  and  ambient  treatments  are  staggered  along  the  y-­‐axis  in  order  of  flowering  onset  date.    The  size  of  the  circles  corresponds  to  the  proportion  of  total  flowers  produced  by  an  individual  on  a  given  day.    (b)  Population-­‐level  display  schedules  for  each  of  the  same  population  and  treatment  combinations.    The  height  of  these  lines  reflects  the  proportion  of  total  flowers  produced  by  that  group  on  a  given  day.    

0

20

40

60

80

100

212 232 252 272 2920

0.1

Julian date

Indiv

idual

(a)

(b)

Prop

ortio

n of

flowe

rs pr

oduc

ed

 

88    

 

Figure  A2    The  display  schedules  of  (a)  MN  Heated,  (b)  MN  Ambient,  (c)  PA  Heated,  (d)  PA  Ambient,  (e)  MO  Heated,  (f)  MO  Ambient,  (g)  NC  Heated,  and  (h)  NC  Ambient  populations  and  treatments  of  experiment  1,  with  deployment  schedules  highlighted  in  dark  grey  and  retained  flowers  highlighted  in  light  grey.    Deployment  schedules  were  estimated  by  accounting  for  temperature-­‐mediated  variation  in  floral  longevity  (see  main  text).    For  reference,  average  daily  temperatures  are  shown  in  panel  (i).      

 

 

 

0.0

0.1(a)

0.0

0.1(b)

0.0

0.1(c)

0.00

0.14 (d)

0.0

0.1(e)

0.00

0.12(f)

0.00

0.08 (g)

0.00

0.06(h)

212 232 252 272 2925

30 (i)

Julian date

Pro

porti

on o

f flo

wers

pro

duce

d da

ilyTe

mpe

ratu

re (°

C)

RetainedDeployed

 

89    

 

Figure  A3  The  display  schedules  of  (a)  MN  Pollinated,  (b)  MN  Unpollinated,  (c)  PA  Pollinated,  (d)  PA  Unpollinated,  (e)  MO  Pollinated,  (f)  MO  Unpollinated,  (g)  VA  Pollinated,  (h)  VA  Unpollinated,  (i)  NC  Pollinated,  and  (j)  NC  Unpollinated  populations  and  treatments  of  experiment  2,  with  deployment  schedules  highlighted  in  dark  grey  and  retained  flowers  highlighted  in  light  grey.    Deployment  schedules  were  estimated  by  accounting  for  temperature-­‐mediated  variation  in  floral  longevity  (see  main  text).    For  reference,  average  daily  temperatures  are  shown  in  panel  (k).    

 

 

 

0.00

0.12 (a)

0.00

0.04 (b)

0.00

0.12 (c)

0.00

0.06 (d)

0.0

0.1 (e)

0.00

0.06 (f)

0.00

0.06 (g)

0.00

0.08 (h)

0.00

0.06 (i)

0.00

0.06 (j)

212 222 232 242 252 262 272 28210

25 (k)

Prop

ortio

n of

flow

ers

prod

uced

dai

lyTe

mpe

ratu

re (°

C)

Julian date

RetainedDeployed

 

90    

 

Figure  A4    Heatmaps  summarizing  the  cross-­‐correlation  coefficients  between  population-­‐level  (a)  display  schedules  or  (b)  deployment  schedules  and  average  daily  humidity  for  populations  and  treatment  combinations  in  experiments  1,  2,  and  3.    The  color  of  a  specific  box  reflects  the  sign  and  magnitude  of  the  correlation  coefficient,  with  significant  correlations  marked  with  an  S.      

 

Lag0

Lag1

Lag2

Lag3

Lag4

Lag5

S SS S

S S

SS

S SS

Lag0

Lag1

Lag2

Lag3

Lag4

Lag5

SSS

S SS

S S

SS S

SS

MN LateNC EarlyMN Early

NC UnpollinatedNC Pollinated

VA UnpollinatedVA Pollinated

MO UnpollinatedMO Pollinated

PA UnpollinatedPA Pollinated

MN UnpollinatedMN PollinatedNC AmbientNC Heated

MO AmbientMO HeatedPA AmbientPA Heated

MN AmbientMN Heated

!0.5 !0.25 0 0.25 0.5(a) (b)

 

91    

 

Figure  A5    Heatmaps  summarizing  the  cross-­‐correlation  coefficients  between  population-­‐level  (a)  display  schedules  or  (b)  deployment  schedules  and  total  daily  precipitation  for  population  and  treatment  combinations  in  experiments  1,  2,  and  3.    The  color  of  a  specific  box  reflects  the  sign  and  magnitude  of  the  correlation  coefficient,  with  significant  correlations  marked  with  an  S.      

Lag0

Lag1

Lag2

Lag3

Lag4

Lag5

S

S

SS

SS

Lag0

Lag1

Lag2

Lag3

Lag4

Lag5

S SS

SS

SS

NC UnpollinatedNC Pollinated

VA UnpollinatedVA Pollinated

MO UnpollinatedMO Pollinated

PA UnpollinatedPA Pollinated

MN UnpollinatedMN PollinatedNC AmbientNC Heated

MO AmbientMO HeatedPA AmbientPA Heated

MN AmbientMN Heated

!0.5 !0.25 0 0.25 0.5(a) (b)

 

92    

 

Figure  A6    Population-­‐level  display  schedules  for  the  constant  watering  treatment  (solid),  one-­‐week  watering  treatment  (dashed),  and  two-­‐week  watering  treatment  (dotted)  for  the  (a)  MN  early,  (b)  NC  early,  and  (c)  MN  late  populations  in  experiment  3.  

0.0

0.1 (a) ConstantOne weekTwo week

0.0

0.1

Prop

ortio

n of

flowe

rs pr

oduc

ed

(b)

209 219 229 239 249 259 2690.0

0.1

Julian date

(c)

 

93    

 

Figure  A7  Heatmaps  summarizing  the  cross-­‐correlation  coefficients  between  population-­‐level  (a)  display  schedules  or  (b)  deployment  schedules  and  the  volumetric  water  content  measured  within  different  replicates  of  the  watering  treatments  for  the  MN  early  cohort  of  experiment  3.    The  color  of  a  specific  box  reflects  the  sign  and  magnitude  of  the  correlation  coefficient,  with  significant  correlations  marked  with  an  S.  

 

 

 

Lag0

Lag1

Lag2

Lag3

Lag4

Lag5

Two Week 4

Two Week 3

Two Week 2

Two Week 1

One Week 4

One Week 3

One Week 2

One Week 1

Constant 4

Constant 3

Constant 2

Constant 1

S

S

S

S

S

S

S

Lag0

Lag1

Lag2

Lag3

Lag4

Lag5

S

S

!0.5 !0.25 0 0.25 0.5(a) (b)

 

94    

 

Figure  A8    The  population-­‐level  display  schedules  for  the  29  flowering  species  naturally  occurring  at  the  Koffler  Scientific  Reserve  at  Joker’s  Hill.    See  Table  S1  for  full  species  names.    

 

 

 

 

 

 

95    

Table A1 Estimates of population synchrony (Sp), average within-individual synchrony (Si ± standard error of the mean), and the strength of phenological assortative mating (ρ) for each population and treatment combination of experiments 1-3 and for the 29 species naturally occurring at the Koffler Scientific Reserve at Joker’s Hill. Population synchrony was calculated per Weis et al. (2014), individual synchrony per the metric presented in the main text, and the strength of assortative mating per Weis and Kossler (2004).

Experiment Population*Treatment Sp Si ρ 1 MN Heated 0.682 0.161 ± 0.007 0.082 MN Ambient 0.583 0.165 ± 0.006 0.126 PA Heated 0.704 0.197 ± 0.008 0.088 PA Ambient 0.638 0.215 ± 0.007 0.183 MO Heated 0.556 0.148 ± 0.008 0.199 MO Ambient 0.671 0.180 ± 0.008 0.175 NC Heated 0.758 0.162 ± 0.007 0.191 NC Ambient 0.718 0.151 ± 0.012 0.173

2 MN Pollinated 0.558 0.212 ± 0.009 0.164 MN Unpollinated 0.705 0.132 ± 0.006 0.216 PA Pollinated 0.586 0.210 ± 0.009 0.224 PA Unpollinated 0.668 0.143 ± 0.006 0.090 MO Pollinated 0.599 0.173 ± 0.006 0.133 MO Unpollinated 0.752 0.134 ± 0.005 0.053 VA Pollinated 0.449 0.181 ± 0.007 0.542 VA Unpollinated 0.685 0.161 ± 0.005 0.117 NC Pollinated 0.686 0.174 ± 0.007 0.503 NC Unpollinated 0.705 0.159 ± 0.005 0.238

3 MN Early 0.598 0.154 ± 0.002 0.121 NC Early 0.750 0.134 ± 0.004 0.176

MN Late 0.572 0.174 ± 0.006 0.202

KSR species Alliaria petiolata 0.822 0.110 ± 0.003 0.283 Aquilegia canadensis 0.603 0.120 ± 0.008 0.465

Arctium minus 0.709 0.126 ± 0.004 0.202 Asclepias syriaca 0.561 0.255 ± 0.044 0.571 Chelidonium majus 0.746 0.140 ± 0.004 0.088 Cirsium vulgare 0.586 0.100 ± 0.004 0.651 Daucus carota 0.637 0.074 ± 0.005 0.213 Erigeron philadelphicus 0.750 0.076 ± 0.003 0.135 Erigeron pulchellus 0.576 0.135 ± 0.019 0.258 Eupatorium maculatum 0.761 0.106 ± 0.002 0.227 Galium aparine 0.831 0.145 ± 0.008 0.388 Geranium robertianum 0.788 0.163 ± 0.031 0.612 Geum canadense 0.564 0.182 ± 0.034 0.479 Hesperis matronalis 0.615 0.170 ± 0.005 0.178 Inula helenium 0.697 0.075 ± 0.005 0.232 Lactuca serriola 0.517 0.106 ± 0.005 0.352 Leonurus cardiaca 0.905 0.129 ± 0.002 0.053

 

96    

Leucanthemum vulgare 0.692 0.099 ± 0.005 0.335 Monarda fistulosa 0.677 0.130 ± 0.005 0.190 Phryma leptostachya 0.746 0.112 ± 0.004 0.217 Plantago major 0.645 0.180 ± 0.009 0.512 Prunus serotina 0.722 0.240 ± 0.009 0.341 Ranunculus acris 0.656 0.198 ± 0.050 0.446 Rudbeckia hirta 0.718 0.075 ± 0.004 0.191 Solidago altissima 0.435 0.146 ± 0.004 0.721 Sonchus arvensis 0.419 0.089 ± 0.004 0.436 Verbascum thapus 0.527 0.155 ± 0.019 0.601 Verbena urticifolia 0.867 0.103 ± 0.002 0.052 Vicia cracca 0.364 0.232 ± 0.040 0.698

 

97    

Individual  synchrony,  Si  

 

Mating  opportunities  between  flowers  on  a  given  individual,  i,  are  minimized  when  open  flowers  are  distributed  evenly  across  days  and  maximized  when  all  flowers  are  open  at  once.    Thus,  uniform  display  schedules  (where  the  variance  in  flowers  produced  per  day  =  0)  should  have  the  lowest  levels  of  synchrony  regardless  of  flowering  duration  or  the  total  number  of  flowers  produced  (Si  =0).    This  property  is  captured  by  the  coefficient  of  variation,  CVi,  measured  as  the  standard  deviation  of  flower  number  across  days  (SDi)  divided  by  the  mean  number  of  flowers  across  days  ( i);  however,  this  measure  scales  with  the  duration  of  flowering,  Di.    Standardizing  the  coefficient  of  variation  by  the  square  root  of  the  flowering  duration  ensures  that  Si  is  largely  insensitive  to  flowering  duration  (Fig.  A9).    

Display  schedules  with  short  flowering  durations  are  sensitive  to  the  distribution  of  flowers  among  days  (Fig.  A9).    Specifically,  as  Di  approaches  0,  Si  becomes  increasingly  lower  as  display  schedules  become  more  uniform  (i.e.  as  SDi  approaches  0).    We  suggest  caution  when  applying  this  metric  to  display  schedules  that  are  2-­‐5  days  in  length,  especially  when  the  distribution  of  flowers  among  days  is  nearly  uniform.    

 Figure  A9    Si  as  a  function  of  the  duration  of  flowering  for  four  example  display  schedules.    In  each  schedule,  all  flowers  are  produced  on  the  first  and  last  date  of  flowering  with  zero  flower  counts  on  all  other  days.    The  

display  schedules  differ  in  the  proportion  of  total  flowers  produced  on  the  first  versus  the  last  day,  represented  by  the  different  line  types  above.    

0 5 10 15 20 25 30

0.0

0.2

0.4

0.6

0.8

1.0

Flowering duration

Si

95-575-2566-3350-50

 

98    

Si  is  technically  undefined  when  Di=1.    Individuals  that  produce  all  flowers  in  a  single  day  are  assigned  a  synchrony  of  1.    Individual  synchrony  is  always  0  for  uniform  display  schedules,  regardless  of  flowering  duration  (Fig.  A10A  vs.  A10B)  or  the  total  number  of  flowers  produced  (Fig.  A10A  vs.  A10C).    Si  will  be  equivalent  for  display  schedules  that  allocate  identical  proportions  of  flowers  among  days  (Fig.  A10D  vs.  A10F  and  A10E  vs.  A10G)  irrespective  of  schedule  shape  (Fig.  A10F  vs.  A10H).      

 Figure  A10    Eight  different  hypothetical  flowering  schedules  for  individuals  A-­‐H,  along  with  the  corresponding  values  of  Si,  Di,  and  Ti.    

 

1 2 3 4 50

10

20

Num

ber o

f flo

wer

s

1 2 3 4 5 6 7 8 9 100

10

20

1 2 3 4 50

50

100

Num

ber o

f flo

wer

s

1 2 3 4 5 6 7 8 9 100

5

10

1 2 3 4 50

50

100

Num

ber o

f flo

wer

s

1 2 3 4 5 6 7 8 9 100

5

10

1 2 3 4 50

15

30

Day

Num

ber o

f flo

wer

s

1 2 3 4 5 6 7 8 9 100

5

10

Day

10 10 10 10 10

100 100 100 100 100

490 0 0 1

24

0 0 0 1

5 5 5 5 5 5 5 5 5 5

12

34

5 54

32

1

2

4

6

8

10 10

8

6

4

2

2

6

10

8

4 4

8

10

6

2

125

SA = 0DA= 5TA = 50

SC = 0DC= 5TC = 500

SE = 0.975DE= 5TE = 50

SG = 0.951DG= 5TG = 25

SB = 0DB= 10TB = 50

SD = 0.157DD= 10TD = 30

SF = 0.157DF= 10TF = 60

SH = 0.157DH= 10TH = 60

 

99    

Si  can  be  rewritten  as  follows:  

 

where  SSi  is  the  sum  of  squares  for  individual  i  and  Ti  is  the  total  number  of  flowers  produced.    This  form  of  the  equation  can  be  easily  adjusted  to  accommodate  flower  counts  that  occur  at  equal  intervals  throughout  the  growing  season:  

 

where  Ii  equals  the  census  interval  (e.g.  Ii  =  3  if  flowers  are  counted  every  3  days)  and  Ci  equals  the  total  number  of  census  days.    Here  we  assume  that  the  patterns  observed  during  census  days  are  representative  of  the  data  that  were  not  sampled,  so  that  the  duration  of  flowering  is  equal  to  Ii*Ci,  the  total  number  of  flowers  produced  is  equal  to  Ii*Ti,  and  the  sample  sum  of  squares  is  equal  to  Ii*SSi.      

Figure  A11  shows  how  well  this  metric  captures  true  levels  of  synchrony  as  interval  lengths  increase.    We  used  four  sample  individual  display  schedules  from  experiment  3,  and  we  estimated  Si  when  Ii=1  (the  true  value)  up  to  an  interval  length  of  6.    Estimates  start  to  deviate  from  the  true  value  when  I>4,  but  adequately  descries  Si  when  Ii  is  three  or  less.  

 

 

 

100    

 Figure  A11      Estimates  of  Si  for  different  display  schedules  (A-­‐D)  as  the  census  interval  increases.    The  true  Si  is  shown  when  I=1.  

1 3 5 7 9 12 15 18 21 1 4 7 11 15 19 23 27 31 35

1 4 7 11 15 19 23 27 31 35 1 4 7 10 14 18 22 26

0.20

0.15

0.10

1 2 3 4 5 6

Interval I

Si

Num

ber o

f !ow

ers

Num

ber o

f !ow

ers

40

20

0

50

25

0

80

40

0

20

10

0

A B

C D

A

C

B

D

Day Day

 

101    

Chapter Four

The Influence of Competition on Phenotypic Responses to Warming

This chapter resulted from collaboration with Benjamin Gilbert, Matthew N. Cumming, Marc W.

Cadotte, Caroline M. Tucker, and Arthur E. Weis. Susana M. Wadgymar carried out the

experiment, performed the analyses, and wrote the manuscript. MNC assisted with fieldwork,

BG lent advice on statistical analyses, BG and MWC assisted with funding, CMT helped develop

the motivation for the study, all coauthors contributed to experimental design, and AEW

contributed to manuscript editing.

Abstract

Global warming has influenced the timing of life history traits in many plant species.

The extent of shifts in reproductive phenological traits has been observed to vary according to a

species’ developmental position within a community of plants, with early flowering species

advancing more often, and by a larger degree, than those flowering later. Species may also

experience temporal variation in competition as the surrounding community changes in density

and composition throughout the growing season. Warming-induced plasticity in reproductive

phenology may vary among species in magnitude, direction, or adaptive value if phenotypic

shifts alter the degree of overlap with competing species. In this way, phenological responses to

warming may vary among species occupying distinct yet overlapping temporal niches, and may

depend on the presence, abundance, and species-specific responses of the heterospecific

competitors.

To investigate the influence of competition on phonological responses to warming among

phonologically distinct species, we manipulated competitive and thermal regimes for 3 weedy

plant species that differ in growth and development: Sinapis arvensis (early flowering),

Chamaecrista fasciculata (intermediate flowering), and Ambrosia artemisiifolia (late

flowering). We constructed communities that varied in the form and strength of competition,

with each species planted individually in monoculture and together in polyculture at both low

and high densities. These communities were then exposed to either ambient or elevated

 

102    

temperatures via artificial warming in a field setting. We monitored plasticity in flowering onset

date (a temperature-sensitive trait) and plant size (a trait representative of competitive ability) for

each species in each environment, and we used total reproductive biomass to estimate patterns of

selection across treatments.

For all traits and species, differences in community composition and plant density did not

interact with thermal treatment to influence flowering onset dates or final plant sizes. Planned

contrasts revealed that increased temperatures only significantly influenced flowering onset date

in two of the four competitive environments in the early flowering S. arvensis, indicating that

competitive regimes can sometimes constrain potential phenotypic responses to warming. In all

cases, plasticity in flowering onset date was adaptive and selection regimes did not differ

significantly between treatments. The patterns of selection imposed by warming on final plant

size were dependent on culture type for C. fasciculata, but were otherwise similar across

treatments for S. arvensis and A. artemisiifolia.

Our results demonstrate that phenotypic responses to warming and subsequent patterns of

selection are species and trait-specific. In general, variation in the competitive environment may

not act to constrain potential responses to increases in temperature in cases where competing

species are phenologically distinct, and other ecological or evolutionary processes may be

contributing to species-level differences in responses to warming.

Introduction

The widespread advances in plant reproductive phenology observed over the past few

decades are viewed as indicators of global climate change (Fitter & Fitter 2002; Parmesan 2007;

Menzel et al. 2006). However, variation in the responses of species remains largely unexplained,

even when accounting for phylogenetic non-independence (Willis et al. 2008). Some have

observed that earlier-flowering species are advancing more than those flowering later in the

growing season (Fitter & Fitter 2002; Menzel et al. 2006; Bertin 2008), suggesting that species

occupying distinct temporal niches are experiencing contrasting abiotic and biotic conditions that

may act in concert with increases in temperature to influence flowering onset dates. Biotic

factors, including competition for resources, can also vary seasonally and have the potential to

independently influence life history traits (Dyer & Rice 1999; Elzinga et al. 2007; McGoey &

 

103    

Stinchcombe 2009; Wright et al. 2015), or to act synergistically or antagonistically with the

effects of climate change (Kareiva et al. 1993). Observational and experimental work aimed at

documenting phenotypic responses to increasing temperatures often cannot distinguish between

the cumulative effects of the abiotic and biotic environment, leading to reduced predictive power

and misidentifications of the factors promoting phenological change (Gilman et al. 2010; Van

der Putten et al. 2010).

Long-term monitoring studies are invaluable because they relay phenological responses

in natural plant assemblies and under natural settings (Fitter & Fitter 2002; Forrest et al. 2010;

Willis et al. 2008). Such studies, however, are unable to differentiate between plastic and

genetic changes in flowering time. Generally, it is difficult to partition the effects of warming

from those of other uncontrolled factors, including biotic interactions (Gienapp et al. 2008;

Merilä & Hendry 2014). For example, species may vary in competitive abilities, and it is

possible that the response of a species to warming observed in one community is constrained in

another because of the presence of a superior competitor (Goldberg & Barton 1991; Weiner

1988). Additionally, we frequently do not have fitness data to accompany these observations,

resulting in speculation on the fitness consequences of any phenological shifts (Merilä & Hendry

2014).

In contrast, phenological data collected from manipulative warming experiments control

for many other factors that may also influence the timing of flowering, and by design distinguish

between plastic shifts in flowering onset within a growing season and evolutionary shifts

between seasons (Dunne et al. 2004; Anderson et al. 2012). However, these manipulations are

typically applied to natural communities where phenotypes and fitness are not followed on

individual plants (Price & Waser 1998; de Valpine & Harte 2001; Sherry et al. 2007), or where

competition is not quantified (Post et al. 2008). Other experiments have manipulated

temperatures for plant populations constructed to resemble monocultures with constant densities

(Wadgymar et al., in press). As such, we have little idea of how plasticity in growth and

development may be restricted by aspects of the biotic environment, including spatial or

temporal overlap with competitors. On account of this, many of the effects seen in these

experiments might only be roughly indicative of potential outcomes in more natural settings

(Wolkovich et al. 2012).

 

104    

Although relatively unexplored, it is plausible that warming-induced shifts in life history

traits, whether plastic or genetic, can be influenced by the competitive dynamics within

communities. The timing of flowering onset can influence patterns of resource allocation

between growth and reproductive functions (Bazzaz et al. 1987). Plant size is often indicative of

competitive ability (Gaudet & Keddy 1988), and species-specific shifts in reproductive timing

may influence the degree of temporal overlap with conspecifics (Price & Waser 1998; Sherry et

al. 2007; Aldridge et al. 2011), potentially altering competitive dynamics between species for

pollinator access, resources for fruit maturation, or future seedling establishment (Ågren &

Fagerström 1980; Forrest et al. 2010). Consequently, fitness may be influenced directly by shifts

in phenology or indirectly through associated changes in plant size or growth rate (Weiner 1988).

Ultimately, the magnitude of warming-induced phenological shifts, and their adaptive value, may

depend on the presence and responses of competing species within the same community.

At the community level, the growth and reproductive development of species are

staggered throughout the growing season at temperate latitudes (Rabinowitz et al. 1981; Herrera

1986; Weis et al. 2014). Species that differ in phenological traits may occupy distinct, yet

overlapping, temporal niches, where early- and late- flowering species experience competition

from those developing later or earlier, respectively, and intermediate-flowering species

experience competition from both groups (Kareiva et al. 1993; Pau et al. 2011). While empirical

evidence and support for temporal niche occupation is rare (Dante et al. 2013; Zhang et al.

2014), the potential for temporally asymmetric competition to differentially constrain the

potential responses of species to warming or alter patterns of selection on growth or phenological

traits remains unexplored.

To investigate the potential for competitive regimes to enhance or repress phenological

responses to climate change differentially among phonologically distinct species, we exposed

plant communities of different composition to either ambient or elevated temperatures via

artificial warming. We monitored the growth and flowering phenology of three annual species

planted from seed: early-flowering Sinapis arvensis, intermediate flowering Chamaecrista

fasciculata, and late-flowering Ambrosia artemisiifolia. We examined the effects of intra- verses

inter-specific competition by planting these species with conspecifics in monocultures or with

each other in polycultures at both low and high densities. We applied a factorial combination of

 

105    

three treatments: two thermal environments, two culture types, and two densities. We ask (1)

Does the competitive environment alter phenotypic responses to warming? and (2) Does the

competitive environment alter the selection pressures imposed by warming?

Methods

Study organisms

We selected three weedy, summer annual species that differed in patterns of growth and

phenology to create simple communities with varied competitive dynamics. Seeds of early

flowering Sinapis arvensis (Brassicaceae), or wild mustard, were collected in 2009 from a large

population in the margin of an agricultural field near Honfleur, Québec (46.6560°N,

70.8788°W). Growth is determinate in this species, with plants growing vegetatively as rosettes

until bolting and the formation of an indeterminate inflorescence (Mulligan & Bailey 1975).

Flowers are hermaphroditic and are pollinated by a wide variety of species in the orders

Hymenoptera and Diptera (Mulligan & Kevan 1973).

Seeds of the intermediate flowering Chamaecrista fasciculata (Fabaceae), or partridge

pea, were collected in 2009 from a naturalized population in the Grey Cloud Dunes south of

Minneapolis, Minnesota (44.8011°N, 92.9647°W). This species has indeterminate growth and

flowering, with a branching, semi-woody morphology (Garish & Lee 1989). Flowers are

hermaphroditic and are exclusively buzz pollinated (Thorp & Estes 1975).

Seeds of the late flowering Ambrosia artemisiifolia (Asteraceae), or common ragweed,

were collected in 2005 from various established populations around Mississauga, Ontario

(43.5890°N, 79.6441°W). This wind-pollinated species also has indeterminate growth, growing

and maturing seeds until first frost (Bassett & Crompton 1975). Ambrosia artemisiifolia is

monoecious, with distinct male and female flowers that differ in average onset date (Deen et al.

1998).

Experimental design

This study was conducted in the Experimental Climate Warming Arrays at the Koffler

Scientific Reserve at Joker’s Hill (44.0300°N, 79.5275°W), where plants were exposed to either

 

106    

present-day or projected future thermal regimes (per OMNR 2007). Each warming array

consisted of six infrared heaters mounted on a steel triangular structure 1.25 meters above soil

level (design per Kimball et al. 2008). Heaters were angled inward and down from horizontal,

creating a uniform circular heat shadow of 3 meters in diameter. Six arrays, or plots,

experienced ambient temperatures while another six were heated to 1.5˚C above ambient during

the day and 3˚C above ambient at night (per Easterling et al. 1997). Plants were otherwise

exposed to natural conditions. Plot-level temperatures were monitored in three plots per

treatment using infrared radiometers (SI-111 infrared radiometer, Campbell Scientific,

Edmonton, Canada).

Competition treatments were applied at the subplot level, with each plot divided into

eight equally sized, wedge-shaped subplots (~0.69 m2) using 6-inch edging buried at soil level to

minimize belowground plant contact between subplots. Each species was planted in

monoculture and polyculture communities to distinguish between the effects of intra- versus

interspecific competition. These mono- and polycultures were then replicated at low and high

densities (18 vs. 150 total seeds per subplot, respectively) to manipulate the strength as well as

the type of competition. Plots were cleared of all natural vegetation prior to planting and were

weeded periodically throughout the growing season.

All seeds were stratified for 8 days prior to planting, and C. fasciculata seeds were also

scarified. Seeds were scattered at random in their appropriate subplots on June 8th, 2012. After

approximately two weeks, we measured the distance to, and identity of, the first and second

nearest neighbors for a subset of focal plants to confirm desired differences in density and

community composition. We periodically surveyed focal individuals for survival and the date of

first flower. For A. artemisiifolia, we monitored the onset of male flowering when pollen was

presented and female flowering when a stigma was first observed to protrude from the flower.

We collected all fruit when matured, and fecundity was estimated as total mass of seeds and

fruit. Upon first frost, plants were harvested at soil level to measure aboveground vegetative

biomass.

Statistical Analyses

 

107    

We confirmed that temperature differences between heated and ambient plots were

maintained throughout the experiment using a repeated measures linear mixed model with

thermal treatment, day, and their interaction as fixed effects and plot as a random effect. To

account for any autocorrelation in temperature measurements among days, we incorporated an

auto-regressive error structure of order 1, nested within plot (Zuur et al. 2009), using the nlme

package (Pinheiro et al. 2014) in R (R Development Core Team, 2014). We verified density

differences between low- and high-density treatments using the average distance between our

focal plants and their first and second nearest neighbors as the response variable in a linear

mixed model with density, culture, species, and their interaction as fixed effects and subplot

nested within plot as a random effect, again using the nlme package in R. With these models,

and with those subsequently described, variance heterogeneity among treatments or species was

corrected using error variance covariates, if necessary (Zuur et al. 2009). We selected the

random terms and error covariates by minimizing AIC values, and optimized fixed effects via

maximum likelihood estimations.

We examined treatment effects on the average date of flowering onset and final

aboveground vegetative biomass for each species using linear mixed models, and also analyzed

differences in reproductive biomass using a generalized linear mixed model with a gamma

distribution and log link (via the lme4 package in R, Bates et al. 2014). All three treatments, and

their interactions, were included as fixed effects, while subplot nested within plot was included

as a random effect. Treatment effects on the onset of male and female flowering in A.

artemisiifolia were assessed separately. We analyzed the log of vegetative biomass +1 in order

to meet assumptions of residual normality, and we define reproductive biomass as the mass of

seeds and fruit. Gamma distributions exclude zero; accordingly, we added 0.01 to reproductive

biomass estimates. A significant interaction between the warming treatment and either or both

of the density or culture treatments indicates that the competitive environment has modified a

species’ response to warming. We present the results from the final, optimized models selected

via log likelihood ratio tests.

The analyses described above will reveal whether treatment combinations yielded

differences in average phenotype and fitness. We used planned contrasts from least-squares

means to identify the specific competitive environments in which phenotypic responses to

 

108    

warming were statistically significant. We standardized flowering onset date and vegetative

biomass within a species to a mean of zero and a standard deviation of 1. This allows for

comparisons among species, treatments, and traits in the degree and direction of plasticity.

Within each competitive environment, we performed a pairwise contrast between the

standardized trait values in heated and ambient conditions using the lsmeans package in R (Lenth

& Hervé 2015). Error rates were Tukey HSD adjusted and contrasts yielded standard errors of

the differences in means between thermal treatments.

Selection analyses

To examine whether responses to thermal or competitive environments were adaptive, we

performed phenotypic selection analyses to estimate patterns of selection on the onset date of

flowering and on final plant size. We analyzed the covariance between each trait and fitness

using a generalized linear mixed model with a gamma distribution and log link. While this

methodology yields statistically sound estimates of direct selection on traits, they are not on a

linear scale, and are thus not directly comparable to selection gradients calculated via multiple

regression (Lande & Arnold 1983).

There were no clear differences among treatments in the proportion of individuals

surviving to produce seed, with an average of 88% survival across species and treatments (data

not shown). We thus focus our selection analyses on the mass of fruit and seeds produced.

Within a species, traits were mean standardized, and we calculated relative fitness for each

individual as the total reproductive biomass divided by the average reproductive biomass

produced by all individuals of the same species. Gamma distributions exclude zero, and again

we added 0.01 to all individual fitness values prior to calculating relative fitness. For A.

artemisiifolia, we could not include the male and female flowering onset dates due to a high

degree of collinearity, so we assessed selection on each trait separately, with vegetative biomass

included in both models. Due to a limited sample size, we were unable to calculate patterns of

selection for A. artemisiifolia planted in low-density monocultures.

To determine whether selection regimes differed among thermal or competitive

environments, we included all three treatments in interactions with each trait in a separate

analysis. An interaction between trait and treatment(s) indicates that the magnitude or direction

 

109    

of selection on that trait is dependent on variation in the treatment. We report chi-squared values

from analyses of deviance for the final, optimized model.

Results

Treatment differences

Artificially warmed plots were maintained at a higher temperature than ambient plots

throughout the duration of the growing season (1.97°C average difference; Treatment F1,

692=15.30, p=0.001; Day F1, 692=156.89, p<0.001; Treatment*Day F1, 692=0.48, p>0.5). Density

treatments were also effectively constructed, with the average distance to the first and second

nearest neighbors significantly lower in the low-density treatment than in the high, regardless of

culture type (Density, F1, 77=45.55, p<0.001). However, within the low-density treatment, S.

arvensis had closer neighbors, on average, than C. fasciculata or A. artemisiifolia (15.01 vs.

17.22 vs. 20.29 cm, respectively, Density*Species, F2, 852=3.25, p<0.05), reflecting a potential

difference in the intensity of competition among species planted at low-density.

Phenotypic responses to warming

Species varied in their phenological responses to thermal and competitive environments.

Increased temperatures marginally advanced flowering onset in the early-flowering S. arvensis,

strongly accelerated flowering onset in the intermediate-flowering C. fasciculata, and did not

affect the onset of male or female flowering in the late-flowering A. artemisiifolia (Thermal

term, Table 4.1, Figs. 4.1a-c). The responses of flowering onset to warming were not influenced

by culture treatment for any of the three species (Thermal*Culture term, Table 4.1), and shifts in

phenology due to thermal treatment were only marginally modified by density for the onset of

female flowering in A. artemisiifolia (Thermal*Density term, Table 4.1). These results suggest

that species vary in their phenological sensitivity to changes in temperature, and that any plastic

responses elicited by increasing temperatures may be largely unaffected by competitive

dynamics.

Increased temperatures slightly decreased aboveground vegetative biomass in A.

artemisiifolia, but had no influence on size in the earlier-developing S. arvensis or C. fasciculata

 

110    

(Table 4.1, Figs. 4.1d-f). Furthermore, temperature did not interact with either culture or density

treatments for any species, again suggesting that competition may not play a large role in

managing phenotypic responses to warming.

Phenotypic responses to competitive dynamics

For all three species, the onset of flowering responded to variation in the competitive

environment. For S. arvensis, both density and culture treatments had interacting effects, with

the greatest degree of plasticity in low-density, monocultures and in high-density, polycultures

(Table 4.1, Fig. 4.1a). This effect was predominantly driven by the delay of flowering in low-

density polycultures, where this early-flowering species experienced the least competition

amongst the slower-growing species at low density. For C. fasciculata, culture had a weaker and

additive effect with temperature, with flowering onset delayed in polyculture conditions relative

to monoculture regardless of thermal treatment (Fig. 4.1b). The onset of male and female

flowering was delayed in polyculture communities in A. artemisiifolia (Fig. 4.1c), and the onset

of female flowering was further delayed in ambient, low-density treatments. Both C. fasciculata

and A. artemisiifolia experience competition from earlier-developing species, and our results

imply that the presence of interspecific competitors, and not planting densities, is the most

influential component of the competitive environment for these species.

Final plant size was also affected by competitive regimes in all species. For the early-

developing species, S. arvensis, high-density conditions reduced plant size only in monoculture

communities, where competition was expected to be strongest (Table 4.1, Fig. 4.1d). In contrast,

for C. fasciculata, high-density conditions reduced plant size in polyculture communities at high

densities (Fig. 4.1e). For the later developing species, A. artemisiifolia, high densities

consistently reduced plant size regardless of culture type, although increased temperatures

seemed to ameliorate the degree of decline (Fig. 4.1f).

Differences in reproductive biomass largely reflected treatment effects on plant size

(Table 4.1, Figure 4.2), with fecundity reduced at high densities in monoculture conditions in S.

arvensis and in polyculture conditions in C. fasciculata. Ambrosia artemisiifolia performed the

most poorly at high densities, although this decline was alleviated in elevated thermal conditions.

Clearly, larger plant size results in more-fit individuals, and competitive regimes govern the most

 

111    

variation in this trait. However, phenotypic selection analyses can account for variation in size

and reveal whether plasticity in flowering onset confers additional fitness effects.

Modified responses to warming and subsequent patterns of selection

The results of the mixed models reported previously suggest that phenotypic responses to

warming are not influenced by variation in the strength and form of competition; that is, the

slopes of the reaction norms are not distinct from one another. We applied planned contrasts

within competition and density treatment combinations to reveal whether flowering onset date

and final plant size differed between thermal environments. A significant difference would

indicate that the slope of an individual reaction norm is different than zero, even if it was not

distinct from those of the other competitive environments.

Planned contrasts revealed that competitive regimes modified the response of flowering

onset to warming in one of the three species. For S. arvensis, flowering onset was significantly

advanced in only two of the four competitive environments, demonstrating that competitive

dynamics may moderate phenological responses to warming (Fig. 4.3a). Selection gradients did

not differ between thermal regimes (Table 4.2), however early flowering was favored in the

polyculture communities, but was only weakly favored or neutral in monoculture conditions

where competition was likely strongest (Fig. 4.3d, Table 4.4).

For C. fasciculata, flowering onset advanced in response to warming in the same manner

across all competitive environments (Fig. 4.3b). Selection on flowering onset date was more

varied. In the monoculture communities, selection only favored early flowering when plants

were exposed to ambient temperatures (Table 4.3). In high-density conditions, the shift to earlier

flowering alleviated the strength of selection on flowering onset date (Fig. 4.3e). This suggests

that warming-induced shifts to earlier flowering were adaptive in C. fasciculata.

In contrast to the two earlier-developing species, the onset of male flowering was not

affected by thermal or competitive conditions in A. artemisiifolia (Fig. 4.3c), and selection on

this trait was consistently neutral (Fig. 4.3f, Table 4.3). Selection on the onset of female

flowering was also neutral in all cases but one; later flowering was strongly favored in low-

density, polyculture communities when warmed (Table 4.3).

 

112    

Aboveground vegetative biomass was not significantly affected by either temperature or

competitive environment in any of the three species (Fig. 4.4a-c). However, in S. arvensis, plant

sizes tended to be smaller in heated conditions when compared to ambient for all but the

monoculture communities planted at high density (Fig. 4.4a). This may be due, in part, to the

associated shifts to earlier flowering in the elevated temperature treatment (Fig. 4.3a). This

warming induced trend towards smaller size resulted in an increase in the strength of selection

for larger plant size (Table 4.3, Fig. 4.4d). For C. fasciculata, selection gradients did not differ

between treatments, although selection was strongest in the polyculture communities when low

densities are heated and when high densities experience ambient temperatures (Table 4.3, Fig.

4.4e). Selection for larger size was consistent in all treatments for A. artemisiifolia (Table 4.3,

Fig. 4.4f).

Discussion

Global warming has prompted shifts in life history traits across a wide array of taxa

(Parmesan 2007), yet our ability to predict a given species’ response to warming is limited by

unaccounted for evolutionary or ecological processes contributing to phenotypic variation. Here,

we explored the potential for the competitive environment to modify the phenotypic responses of

developmentally distinct species to increases in temperature. For all three focal species,

differences in community composition and plant density did not interact with temperature to

influence flowering onset dates or final plant sizes. Planned contrasts revealed that increased

temperatures only significantly influenced flowering onset date in two of the four competitive

environments in the early flowering S. arvensis, indicating that competitive regimes can

sometimes constrain potential phenotypic responses to warming. However, we saw no evidence

of this for plasticity in final plant size or in the other species examined.

In all cases, plasticity in flowering onset date was adaptive and selection regimes did not

differ significantly between treatments. Patterns of selection imposed by warming on final plant

size were dependent on culture type for C. fasciculata, but were otherwise similar across

treatments for S. arvensis and A. artemisiifolia. Cumulatively, our results demonstrate that

phenotypic responses to warming may generally be insensitive to variation in competitive

dynamics, and that the degree of plasticity induced by warming, and subsequent patterns of

 

113    

selection, are species and trait-specific. Below we briefly review among-species variability in

the response of flowering onset to warming, and then discuss whether competition may be a

contributor to unexplained phenological variation.

Variation in phenological responses to warming

Proposed explanations for variation in responses to warming include differences in

mating system, pollination mechanism, geographical distribution, phylogenetic history, life form,

and flowering time relative to other members of the community (Fitter & Fitter 2002; Peñuelas et

al. 2002; Menzel et al. 2006; Sherry et al. 2007; Bertin 2008; Willis et al. 2008). For example,

in a survey of changes in flowering onset for 385 British plants, it was demonstrated that annuals

were more likely to have shifted their flowering onset dates than perennials, and that larger shifts

are found in insect-pollinated species than in wind-pollinated species, particularly if they were

already relatively early flowering (Fitter & Fitter 2002). The latter suggestion is supported here,

as we observed an advancement of flowering when warmed for the insect-pollinated S. arvensis

and C. fasciculata, but not for the wind-pollinated A. artemisiifolia. The magnitude and

direction of phenological responses to climate change can be phylogenetically conserved, with

the most responsive species belonging to a nonrandom assortment of plant families (Willis et al.

2008). Some have proposed that earlier-flowering species are more sensitive to change in

temperature than those flowering later in the season, with larger shifts in the onset of

reproduction seen in early-developing species relative to late (Fitter & Fitter 2002, Menzel et al.

2006; Sherry et al. 2007; Bertin 2008; but see Peñuelas et al. 2002).

Regardless of the potential causes of variation in responses to warming, it has been

demonstrated that non-responsive species are suffering negative demographic consequences

(Willis et al. 2008). Thus far, most examinations of the mechanisms proposed to explain

variation in responses to climate change have been correlative and have focused on the outcomes

of evolutionary processes (e.g. pollination mechanism). Below we discuss how short-term

ecological processes, like competition, may be more influential than currently appreciated.

Competition and phenology

 

114    

Competitive dynamics can elicit variable physiological, morphological, and

developmental responses among species, with the effects of intra- and interspecific competition

potentially acting in opposing directions (Connell 1983; Linhart 1988; Stoll & Prati 2001). The

occurrence and strength of competition may depend on the life history stages of species at the

time of interaction (Callaway & Walker 1997) or on habitat quality (Aerts 1999). Community

composition changes through space and time as species emerge and senesce, producing transient

competitive regimes that are unique to each species present (Connell 1983). If developmentally

distinct species experience competition in a systematic way, the effects of competitive

interactions should be predictable based on the presence, abundance, and development times of

heterospecifics.

In our experiment, we competed species that varied drastically in growth and

development times, creating the potential for asymmetric competitive pressures across species.

While we found little evidence that competitive regimes constrain phenotypic responses to

increases in temperatures, we did find that competitive dynamics on their own strongly

influenced plasticity in flowering onset date and final plant size, and that each of our focal

species was affected by different aspects of the competitive environment.

If we consider effects on final plant size as proxies for the intensity of competition, we

see that final plant size was reduced in S. arvensis when intraspecific competition was strong

(high density monocultures), whereas strong interspecific competition (high density

polycultures) was more detrimental for the intermediately flowering C. fasciculata. The last

species to flower, A. artemisiifolia, had suppressed growth at high densities irrespective of the

composition of the surrounding community (Fig. 4.1d-f). With the relative importance of intra

and interspecific competition across species in mind, a reexamination of phenological responses

reveals that the strongest competitive dynamics accelerated flowering onset date in S. arvensis,

delayed onset in C. fasciculata, and had little to no influence on onset dates in A. artemisiifolia

(Table 4.1, Fig. 4.1a-c). While there are differences in treatment effects on plasticity among

species, we see general trends in phenotypic selection among species. For all three species, the

competitive environment influenced patterns of selection on plant size (Fig. 4.4), but not on

flowering onset date (Fig. 4.3). Future work should focus on whether the species-specific

competitive effects observed here are a consequence of each species’ temporal sequence within

 

115    

the community, and whether this may add predictive power to explanations of phenological

variation in nature.

The trends observed here might be contingent on whether our focal species are generally

representative of early, intermediate, and late flowering species and whether they naturally co-

occur and compete. Including additional phenologically distinct species in this experiment

would have come at the expense of replication, and instead we chose three annual species whose

growth and development times made predictions about responses to warming and the symmetry

of competition possible. These species are typically found in similar, disturbed habitats, making

them plausible natural competitors. We observed that C. fasciculata and A. artemisiifolia coexist

in plant communities along the eastern United States (data not shown). While S. arvensis’

distribution overlaps substantially with those of the other two species, it is most often found in

the margins of agricultural fields and may be less likely to naturally coexist with either C.

fasciculata or A. artemisiifolia. While we feel that our species selection was representative of

annual plants common to temperate regions, we caution that the outcome of this experiment may

not be broadly observed across other combinations of phenologically distinct species.

Summary

Overall, our results suggest that the effects of competition on phenotypic responses to

warming are largely additive and may be predictable when taking in to account the

developmental sequence of species within a community. The lack of evidence for an interaction

between thermal and competitive treatments is encouraging for studies where competition is not

quantified. Further explorations of the relative importance of increasing temperatures and

competition should investigate how the degree of phenological separation among species affects

phenotypic plasticity and selection. Additionally, the overwintering of seeds may introduce

important variation in emergence dates, both within and between species, which can influence

competitive dynamics later in life. Variation in reproductive phenology has the potential to

influence population demographics, community composition, and evolutionary trajectories, and

rapidly accelerated increases in temperature mandate the continued exploration of potential

contributors to species’ responses to climate change.

 Table  4.1  Analyses  of  the  responses  of  flowering  onset  date,  aboveground  vegetative  biomass,  and  reproductive  biomass  to  thermal  (ambient  vs.  heated),  density  (low  vs.  high),  and  culture  (mono-­‐  vs.  poly-­‐)  treatments.    For  A.  artemisiifolia,  we  separately  analyzed  the  onset  of  male  and  female  flowering.    Flowering  onset  and  vegetative  biomass  were  analyzed  using  linear  mixed  effects  (lme)  models,  while  reproductive  biomass  was  analyzed  using  a  generalized  linear  mixed  (glm)  model  with  a  Gamma  distribution  and  log  link.    We  report  F-­‐values  and  (for  lme  models)  or  Chi-­‐squared  values  from  analyses  of  deviance  (for  glm  models)  and  associated  p-­‐values  for  the  fixed  effects  in  the  final,  optimized  models.    F-­‐values  where  p<0.05  are  in  bold.    

S. arvensis C. fasciculata A. artemisiifolia

Flowering Onset

Vegetative Biomass

Reproductive Biomass

Flowering Onset

Vegetative Biomass

Reproductive Biomass

Male: Flowering

Onset

Female: Flowering

Onset

Vegetative Biomass

Reproductive Biomass

Thermal 3.93

p=0.08 NS NS 22.13

p<0.001 NS NS NS 0.59

p=0.44 4.87

p=0.05 NS

Density 0.53

p=0.47 19.94

p<0.001 18.45

p<0.001 NS 35.37

p<0.001 20.41

p<0.001 NS 0.34

p=0.56 17.30

p<0.001 10.33

p=0.001

Culture 4.80

p=0.04 20.22

p<0.001 10.14

p=0.001 6.66

p=0.01 32.42

p<0.001 17.56

p<0.001 5.40

p=0.02 6.84

p=0.01 NS NS

Thermal:Density NS

NS NS NS NS NS NS 3.54

p=0.06 NS NS

Thermal:Culture NS

NS NS NS NS NS NS NS NS NS

Density:Culture 7.40

p=0.01 NS NS NS 12.28

p=0.002 6.92

p=0.009 NS NS NS NS

Thermal:Density:Culture NS

NS NS NS NS NS NS NS NS NS

116  

 

117    

Table  4.2    Generalized  linear  mixed  effects  analyses  of  the  influence  of  flowering  onset  date,  aboveground  vegetative  biomass,  and  experimental  treatments  on  reproductive  biomass.    Significant  interactions  between  a  trait  and  a  treatment  signify  that  patterns  of  selection  on  that  trait  are  dependent  on  treatment  level.    All  higher  order  interactions  were  not  significant  and  we  report  chi-­‐squared  and  p-­‐values  from  analyses  of  deviance  for  the  final,  optimized  models.    For  A.  artemisiifolia,  we  separately  analyzed  selection  regimes  when  considering  the  onset  date  of  male  and  female  flowering.      

   

  S.  arvensis   C.  fasciculata   A.  artemisiifolia  Male  

A.  artemisiifolia  Female  

Flowering  onset   36.59  p<0.001  

10.36  p=0.001  

NS   NS  

Vegetative  biomass   470.04  p<0.001  

168.59  p<0.001  

263.64  p<0.001  

263.64  p<0.001  

Temperature   NS    

NS   NS   NS  

Density   NS    

3.48  p=0.06  

NS   NS  

Culture   NS    

0.16  p=0.69  

NS   NS  

Flowering  onset*Temperature   NS    

NS   NS   NS  

Flowering  onset*Density   NS    

NS   NS   NS  

Flowering  onset*Culture   NS    

NS   NS   NS  

                             Biomass*Temperature   NS    

NS   NS   NS  

                             Biomass*Density   NS    

NS   NS   NS  

                             Biomass*Culture   NS    

4.30  p=0.04  

NS   NS  

 

 

 

 

   

135  

Table  4.3    Estimates  of  direct  phenotypic  linear  selection  coefficients  (±  S.E.)  and  p-­‐values  for  species  planted  in  thermal  (ambient  vs.  heated),  density  (low  vs.  high),  and  culture  (mono-­‐  vs.  poly-­‐)  treatments.  For  A.  artemisiifolia,  we  examined  selection  on  the  onset  date  of  male  and  female  flowering  separately  to  avoid  issues  of  multicolinearity.    Coefficients  where  p<0.05  are  shown  in  bold.      

S. arvensis C. fasciculata A. artemisiifolia

Thermal Treatment

Competition Treatments Flowering onset Biomass Flowering onset Biomass Flowering onset

Male Biomass

Male Flowering onset

Female Biomass Female

Low, Mono

-0.06 (0.05) p=0.26

0.71 (0.05)

p<0.001

-0.19 (0.07)

p=0.008

0.60 (0.07)

p<0.001 NA NA NA NA

High, Mono

-0.12 (0.05) p=0.01

0.93 (0.06)

p<0.001

-0.20 (0.07)

p=0.004

0.63 (0.09)

p<0.001

0.02 (0.08) p=0.78

1.03 (0.07)

p<0.001

0.04 (0.07) p=0.59

1.03 (0.07)

p<0.001

Low, Poly

-0.38 (0.11)

p<0.001

0.65 (0.09)

p<0.001

-0.15 (0.12) p=0.20

0.60 (0.12)

p<0.001

0.07 (0.10) p=0.51

1.08 (0.05)

p<0.001

0.09 (0.08) p=0.27

1.15 (0.07)

p<0.001

Ambient

High, Poly

-0.19 (0.06)

p=0.003

0.67 (0.07)

p<0.001

-0.24 (0.14) p=0.09

1.11 (0.21)

p<0.001

0.01 (0.08) p=0.93

0.90 (0.08)

p<0.001

0.04 (0.06) p=0.52

0.91 (0.08)

p<0.001

Low, Mono

-0.21 (0.09) p=0.02

0.90 (0.08)

p<0.001

-0.14 (0.10) p=0.19

0.54 (0.10)

p<0.001

0.44 (0.31) p=0.16

1.27 (0.36)

p<0.001

0.43 (0.36) p=0.23

1.21 (0.36)

p<0.001

High, Mono

-0.11 (0.06) p=0.06

0.84 (0.08)

p<0.001

-0.01 (0.11) p=0.95

0.65 (0.11)

p<0.001

0.00 (0.05) p=0.98

0.92 (0.07)

p<0.001

-0.03 (0.07) p=0.65

0.90 (0.07)

p<0.001

Low, Poly

-0.26 (0.09)

p=0.005

0.89 (0.11)

p<0.001

-0.16 (0.24) p=0.52

1.04 (0.30)

p<0.001

0.17 (0.12) p=0.16

0.82 (0.09)

p<0.001

0.76 (0.17)

p<0.001

0.84 (0.05)

p<0.001

Heated

High, Poly

-0.16 (0.06)

p=0.003

0.86 (0.06)

p<0.001

-0.09 (0.12) p=0.44

0.70 (0.13)

p<0.001

0.01 (0.08) p=0.94

0.87 (0.11)

p<0.001

0.02 (0.09) p=0.86

0.84 (0.11)

p<0.001

118  

   

119  

 

Figure  4.1    Average  shifts  in  (A-­‐C)  flowering  onset  date  and  (D-­‐F)  above  ground  vegetative  biomass  ±  S.E.  for  (A,  D)  S.  arvensis,  (B,  E)  C.  fasciculata,  and  (C,  F)  A.  artemisiifolia  in  response  to  thermal  (ambient  vs.  heated),  density  (low  vs.  high),  and  culture  (mono  vs.  poly)  treatments.    Responses  are  shown  for  treatments  of  significant  effects  or  to  facilitate  comparisons  among  species  (see  Table  1).    For  A.  artemisiifolia,  treatment  effects  are  similar  between  the  onset  of  male  and  female  flowering,  and  panel  (C)  only  portrays  variation  in  male  flowering  onset  date.  

 

�

24

25

26Ju

lian

date

A High, PolyHigh, MonoLow, PolyLow, Mono

0

1

2

3

log V

eget

ative

biom

ass (

g)

Mono Poly

D

�

Julia

n da

te

50

54

58 B PolyMono

0

1

2

3

log V

eget

ative

biom

ass (

g) Mono Poly

E

Julia

n da

te

60

64

68 C

Ambient Heated Low density High Density0

1

2

3

4

5

log V

eget

ative

biom

ass (

g) AmbientHeated

FPolyMono

   

120  

 

Figure  4.2    Average  reproductive  biomass  ±  S.E.  for  (A)  S.  arvensis,  (B)  C.  fasciculata,  and  (C)  A.  artemisiifolia  in  thermal  (ambient  vs.  heated),  density  (low  vs.  high),  and  culture  (mono-­‐  vs.  poly-­‐)  treatments.  

 

 

 

0

11

22

Repr

oduc

tive

bioma

ss (g

) AmbientHeated

A

0

11

22

Repr

oduc

tive

bioma

ss (g

) B

Repr

oduc

tive

bioma

ss (g

)

Mono Poly Mono PolyLow Density High Density

0

17

34C

   

121  

 Figure  4.3  (A)  Average  flowering  onset  (in  units  of  sd)  in  heated  conditions  relative  to  ambient  for  various  competitive  regimes.    Error  bars  were  obtained  by  planned  contrasts  of  responses  between  thermal  treatments  within  a  particular  competitive  regime,  with  stars  indicating  a  significant  response  to  warming.    Negative  and  positive  shifts  reflect  the  acceleration  or  delay  of  flowering  onset  when  heated,  respectively.    (B)  Selection  gradients  ±  S.E.  reflecting  the  strength  of  direct  selection  on  flowering  onset  date  in  heated  and  ambient  conditions  for  each  competitive  environment  (see  Table  4.2).    Negative  and  positive  gradients  reflect  direct  selection  for  earlier  and  later  flowering,  respectively.  Due  to  a  small  sample  size,  we  were  unable  to  calculate  the  strength  of  selection  for  A.  artemisiifolia  in  low  density,  monoculture  conditions.    In  panels  (C)  and  (F),  we  convey  the  results  concerning  variation  in  male  flowering  onset  date  in  A.  artemisiifolia.  

!2 !1 0 1 2

Poly, High

Poly, Low

Mono, Low

�

�

�

�

!0.6 !0.3 0.0 0.3 0.6

AmbientHeated

!2 !1 0 1 2

�

�

�

�

!0.6 !0.3 0.0 0.3 0.6

!2 !1 0 1 2

Poly, High

Mono, Low

Flowering onset (scaled)

�

�

�

!0.6 !0.3 0.0 0.3 0.6Selection gradient

A

C

B

D

F

E

*

*

*

*

*

*

Mono, High

Poly, High

Poly, Low

Mono, Low

Mono, High

Poly, Low

Mono, High

   

122  

 Figure  4.4  (A)  Average  above  ground  vegetative  biomass  in  heated  conditions  relative  to  ambient  for  various  competitive  regimes.    Error  bars  were  obtained  by  planned  contrasts  of  responses  between  thermal  treatments  within  a  particular  competitive  regime.    Negative  and  positive  shifts  reflect  increases  or  decreases  in  plant  size  when  heated,  respectively.    (B)  Selection  gradients  ±  S.E.  reflecting  the  strength  of  direct  selection  on  vegetative  biomass  in  heated  and  ambient  conditions  for  each  competitive  environment  (see  Table  4.2).    Positive  gradients  reflect  direct  selection  for  later  size.    Due  to  a  small  sample  size,  we  were  unable  to  calculate  the  strength  of  selection  for  A.  artemisiifolia  in  low  density,  monoculture  conditions.    In  panel  (F),  we  portray  the  selection  coefficients  from  the  model  including  male  flowering  onset  date  in  A.  artemisiifolia  (see  Table  4.2).

!2 !1 0 1 2

Poly, High

Mono, Low

�

�

�

�

0.40 0.65 0.90 1.15 1.40

!2 !1 0 1 2

Poly, High

Mono, Low

�

�

�

�

0.40 0.65 0.90 1.15 1.40

!2 !1 0 1 2

Poly, High

Mono, Low

Vegetative biomass (scaled)

�

�

�

0.40 0.65 0.90 1.15 1.40Selection gradient

AmbientHeated

A

C

B

D

F

E

Poly, Low

Mono, High

Poly, Low

Mono, High

Poly, Low

Mono, High

   

122  

REFERENCES  CITED  

Aerts R (1999) Interspecific competition in natural plant communities: mechanisms, trade-offs

and plant-soil feedbacks. Journal of Experimental Botany, 50, 29-37.

Ågren GI, Fagerström T (1980) Increased or decreased separation of flowering times? The joint

effect of competition for space and pollination in plants. Oikos, 35, 161-164.

Aldridge G, Inouye DW, Forrest JRK, et al. (2011) Emergence of a mid-season period of low

floral resources in a montane meadow ecosystem associated with climate change. Journal of

Ecology, 99, 905-913.

Anderson JT, Inouye DW, McKinney AM, et al. (2012) Phenotypic plasticity and adaptive

evolution contribute to advancing flowering phenology in response to climate change.

Proceedings of the Royal Society B, 279, 3843-3852.

Bassett IJ, Crompton CW (1975) The biology of Canadian weeds. 11. Ambrosia artenmisiifolia

L. and A. psilostachya DC. Canadian Journal of Plant Science, 55, 463-476.

Bates D, Maechler M, Bolker B, et al. (2014) lme4: Linear mixed-effects models using Eigen and

S4. R package version 1.1-7, http://CRAN.R-project.org/package=lme4

Bazzaz FA, Chiariello NR, Coley PD, et al. (1987) Allocating resources to reproduction and

defense. BioScience, 37, 58-67.

Bertin RI (2008) Plant phenology and distribution in relation to recent climate change. Journal of

the Torrey Botanical Society, 135, 126-146.

Callaway RM, Walker LR (1997) Competition and facilitation: a synthetic approach to

interactions in plant communities. Ecology, 78, 1958-1965.

Connell JH (1983) On the prevalence and relative importance of interspecific competition:

evidence from field experiments. The American Naturalist, 122, 661-696.

Dante SK, Schamp BS, Aarssen LW (2013) Evidence of deterministic assembly according to

flowering time in an old-field plant community. Functional Ecology, 27, 555-564.

De Valpine P, Harte H (2001) Plant responses to experimental warming in a montane meadow.

Ecology, 82, 637-648.

   

123  

 Deen W, Hunt T, Swanton CJ (1998) Influence of temperature, photoperiod, and irradiance on

the phenological development of common ragweed (Ambrosia artemisiifolia). Weed Science

Dunne et al. (2004), 46, 555-560.

Dyer AR, Rice KJ (1999) Effects of competition on resource availability and growth on a

California bunchgrass. Ecology, 80, 2697-2710.

Easterling DR, Horton B, Jones PD, et al., (1997) Maximum and minimum temperature trends

for the globe. Science 277, 364-367.

Elzinga JA, Atlan A, Biere A, et al. (2007) Time after time: flowering phenology and biotic

interactions. TRENDS in Ecology and Evolution 22, 432-439.

Fitter AH, Fitter RSR (2002) Rapid Changes in Flowering Time in British Plants. Science, 296,

1689-1691.

Forrest J, Inouye DW, Thomson JD (2010) Flowering phenology in subalpine meadows: does

climate variation influence community co-flowering patterns? Ecology, 91, 431-440.

Garish RS, Lee TD (1989) Physiological integration in Cassia fasciculata Michx.: inflorescence

removal and defoliation experiments. Oecologia,81, 279-284.

Gaudet CL, Keddy PA (1988) A comparative approach to predicting competitive ability from

plant traits. Nature, 334, 242-243.

Gienapp P, Teplitsky C, Alho JS, et al. (2008) Climate change and evolution: disentangling

environmental and genetic responses. Molecular Ecology 17, 167-178.

Gilman SE, Urban MC, Tewksbury J, et al. (2010) A framework for community interactions

under climate change. Trends in Ecology and Evolution, 25, 325-331.

Goldberg DE, Barton AM (1991) Patterns and consequences of interspecific competition in

natural communities: a review of field experiments with plants. The American Naturalist,

139, 771-801.

Herrera J (1986) Flowering and fruiting phenology in the coastal shrublands of Doñana, south

Spain. Vegetatio, 68, 91-98.

   

124  

Kareiva PM, Kingsolver JG, Huey RB (1993) Biotic Interactions and global change. Sinauer

Associates, Sunderland.

Kimball BA, Conley MM, Wang S, et al. (2008) Infrared heater arrays for warming ecosystem

field plots. Global Change Biology, 14, 309-320.

Lande R, Arnold SJ (1983) The Measurement of Selection on Correlated Characters. Evolution,

37, 1210-1226.

Legth RV, Hervé H (2014) lsmeans: Least-squares means. http://cran.r-

project.org/web/packages/lsmeans/

Linhart YB (1988) Intrapopultion differentiation in annual plants. III. The contrasting effects of

intra- and interspecific competition. Evolution, 42, 1047-1064.

McGoey BV, Stinchombe JR (2009) Interspecific competition alters natural selection on shade

avoidance phenotypes in Impatiens capensis. New Phytologist, 183, 880-891.

Merilä J, Hendry AP (2014) Climate change, adaptation, and phenotypic plasticity: the problem

and the evidence. Evolutionary applications, 7, 1-14.

Menzel A, Sparks TH, Estrella N, et al. (2006) European phenological response to climate

change matches the warming pattern. Global Change Biology, 12, 1969-1976.

Mulligan GA, Bailey LG (1975) The biology of Canadian weeds. 8. Sinapis arvensis L.

Canadian Journal of Plant Science, 55, 171-183.

Mulligan GA, Kevan PG (1973) Color, brightness, and other floral characteristics attracting

insects to the blossoms of some Canadian weeds. Canadian Journal of Botany, 51, 1939-

1952.

OMNR: Colombo SJ, McKenney DW, Lawrence KM, et al. (2007) Climate change projections

for Ontario: practical information for policymakers and planners. Ontario Ministry of

Natural Resources, Queen’s Printer for Ontario.

Parmesan C (2007) Influences of species, latitudes, and methodologies on estimates of

phenological response to global warming. Global Change Biology, 13, 1860-1872.

Pau S, Wolkovich EM, Cook BI, et al. (2011) Predicting phenology by integrating ecology,

evolution and climate science. Global Change Biology, 17, 3633-3643.

   

125  

Peñuelas J, Filella I, Comas P (2002) Changed plant and animal life cycles from 1952 to 2000 in

the Mediterranean region. Global Change Biology, 8, 531-544.

Pinheiro J, Bates D, DebRoy S, et al. (2014) nlme: Linear and nonlinear mixed effects models.

(R package, version 3.1-117)

Post ES, Pedersen C, Wilmers CC, et al. (2008) Phenological sequences reveal aggregate life

history response to climatic warming. Ecology, 89, 363-370.

Price MV, Waser NM (1998) Effects of experimental warming on plant reproductive phenology

in a subalpine meadow. Ecology, 79, 1261-1271.

Rabinowitz D, Rapp JK, Sork VL, et al. (1981) Phenological properties of wind- and insect-

pollinated prairie plants. Ecology, 62, 49-56.

Sherry RA, Zhou X, Gu S, et al. (2007) Divergence of reproductive phenology under climate

warming. PNAS, 104, 198-202.

Stoll P, Prati D (2001) Intraspecific aggregation alters competitive interactions in experimental

plant communities. Ecology, 82, 319-327.

Thorp RW, Estes JR (1975) Intrafloral behavior of bees and flowers of Cassia fasciculata.

Journal of the Kansas Entomological Society, 48, 175-184.

Weiner J (1988) The influence of competition on plant reproduction. Plant reproductive ecology:

patterns and strategies, 228-245.

Weis AE, Nardone E, Fox GA (2014) The strength of assortative mating for population date and

its basis in individual variation in flowering schedule. Journal of Evolutionary Biology, 27:

2138:2151.

Willis CG, Ruhfel B, Primack RB, et al. (2008) Phylogenetic patterns of species loss in

Thoreau’s woods are driven by climate change. Proceedings of the National Academy of

Sciences, 105, 17029-17033.

Van der Putten WH, Macel M, Visser ME (2010) Predicting species distribution and abundance

responses to climate change: why it is essential to include biotic interactions across trophic

levels. Philisophical Transactions of the Royal Society B, 365, 2025-2034.

   

126  

Wolkovich EM, Cook BI, Allen JM, et al. (2012) Warming experiments underpredict plant

phenological responses to climate change. Nature, 485, 494-497.

Wright A, Schnitzer SA, Reich PB (2015) Living close to your neighbors – the importance of

both competition and facilitation in plant communities. Ecology, DOI: 10.1890/13-1855.1

Zhang C, Willis CG, Burghardt LT, et al. (2014) The community-level effect of light on

germination timing in relation to seed mass: a source of regeneration niche differentiation.

New Phytologist, 204, 496-506.

Zuur AF, Ieno EN, Walker NJ, et al. (2009) Mixed effects models and extensions in Ecology

with R. Springer New York, New York.

   

127  

Chapter 5

Concluding discussion

Global temperatures are increasing at unprecedented rates. Uncertainty in how species

will fare in warmer environments warrants research on the limitations of phenotypic plasticity in

traits under selection and of the feasibility of active management programs for vulnerable

species. In my thesis, I investigated the contexts under which plasticity in reproductive

phenological traits are adaptive and whether plasticity adequately relieves the selection pressures

imposed by warmer temperatures. In this final chapter, I review current evidence for how

climate change is eliciting plastic and evolutionary responses in the life history traits of plants

and how the results reported here expand this understanding. I then synthesize the findings of

each chapter to discuss the successful application of assisted colonization and assisted gene flow.

Lastly, I suggest potential areas for future research.

Phenotypic plasticity and evolution in response to warming

The sessile nature of plants suggests a crucial role for phenotypic plasticity in

ameliorating the immediate selection pressures imposed by climate change. While only the

underlying heritable component of phenotypic variation will directly determine responses to

selection, selection itself is acting on an organism’s phenotype as a whole, making any other

contributions to phenotypic variation important in a population’s evolutionary trajectory

(Scheiner 1993). The ecological and evolutionary consequences of plasticity have been

extensively explored in well-understood traits (Stearns 1989, Pigliucci 2001), yet we know little

about the relative roles of plasticity and evolution in governing the responses of reproductive

phenological traits to climate change.

Are all species advancing their phenologies?

The majority of plant species are accelerating their reproductive phenologies as

temperatures warm (Parmesan 2007). When looking at the average shifts in first flowering dates

across the growing season, some have observed that earlier flowering species are more

responsive to increases in temperature than those flowering later in the season (Fitter and Fitter

2002, Menzel et al. 2006, but see Peñuelas et al. 2002). In chapter 4, we hypothesized that

   

128  

asymmetric competition among species may be driving these patterns. While the few species

included in our experiment preclude any generalizations about the responses of phenologically

distinct species to warming, our results do fall in line with this trend, with the early- and

intermediate-flowering Sinapis arvensis and Chamaecrista fasciculata responding to warming

while the later flowering Ambrosia artemisiifolia did not. For S. arvensis, the degree of plastic

response of flowering onset to warming depended on the competitive environment. However, in

no case did the effects of competition significantly interact with thermal treatment to influence

flowering onset date or final plant size.

Previous efforts to characterize species-specific responses to warming have focused on

differences between species that arise over evolutionary timescales, including differences in

pollination mechanism, phylogenetic history, mating system, and life cycle (Fitter and Fitter

2002, Peñuelas et al. 2002, Menzel et al. 2006, Bertin 2008, Willis et al. 2008). We are among

the first to explore the potential for a short-term ecological process (e.g. competition) to modify

phenological responses to warming. The lack of evidence for variation in the competitive

environment to differentially affect species’ responses to warming is encouraging for those who

study single species in isolation, who rely on observational data collected from citizen science

programs, or who are otherwise unable to account for spatial or temporal differences in

community composition or density.

Population-level differences in phenotypic responses to warming may contribute to

patterns of species-level variation. Populations can vary drastically in genetic variation for life

history traits (Loveless & Hamrick 1984; Geber & Griffen 2003) and may also differ in their

genetic variation for plasticity in those traits (Stearns 1989). Site-specific differences in

historical abiotic or biotic environments among populations could generate variation in the

capacity of populations to respond to increases in temperature. Data that are collected from

single populations or sites may not be representative of a species’ average response to climate

change. In chapter 2, we exposed geographically distinct populations of C. fasciculata to the

same ambient and heated thermal treatments and monitored differences in reproductive

phenological traits. We detected little to no significant variation in the population-level

responses of budding and flowering onset date to warming, while plasticity in fruiting onset was

highly variable among populations. These data serve to illustrate how our assessments of

   

129  

species-level differences in responses to environmental change may depend on the focal trait

under consideration as well as the population sampled.

Are all advances in phenology adaptive?

A recent meta-analysis concluded that selection generally favors early flowering,

particularly in temperate latitudes (Munguía-Rosas et al. 2011). The majority of observational

studies have found that flowering plants are accelerating their flowering onset dates as

temperatures and atmospheric CO2 concentrations increase (Fitter and Fitter 2002, Menzel et al.

2006, Parmesan 2007). Together, these findings imply that shifts to earlier flowering onset dates

may typically be adaptive. Manipulative experiments have also commonly found shifts towards

early flowering to be adaptive, but caution that plasticity is insufficient to assuage strong

negative directional selection pressures (Etterson and Shaw 2001, Haggerty and Galloway 2011,

Anderson et al. 2011). Only some of the results reported in this thesis align with these findings,

with exceptions explained by differences in responses and selection pressures among species.

Species-specific differences in the selection pressures imposed by warming could

contribute to variable phenotypic shifts among species. The meta-analysis mentioned above was

largely composed of estimates of the strength of selection on flowering onset date in perennials

(Munguía-Rosas et al. 2011). The few annual species included in the analyses experienced

stronger negative directional selection, and more variable selection pressures among-species,

than did perennials. Annual species have one season in which to mature seeds, and variation in

the strength of selection on flowering time may result from the potential trade off between

flowering time and plant size, where earlier flowering guarantees reproductive success before

season’s end while later flowering allows more time for growth and the accumulation of

resources for the production of offspring (Dorn & Mitchell-Olds 1991; Weis et al. 2014). The

results from chapter 4 indicate that the selection regimes imposed by increased temperatures vary

by species, with early flowering strongly favored in the early-flowering species and with

flowering onset selectively neutral in the last species to flower. Our results, combined with the

scarcity of studies that measure the fitness consequences of plasticity in flowering onset date,

demonstrate the need for caution when making general assumptions about the adaptive nature of

phenological responses to warming.

   

130  

Our results also illustrate that the adaptive value of plasticity may depend on the

ecological context in which it was assessed. Chapters 2 and 4 of this thesis together demonstrate

that selection only favors early flowering in C. fasciculata when low-density monoculture

communities experience ambient temperatures. Under these conditions, warming-induced

advances in flowering onset are adaptive. However, when competitive dynamics are stronger,

through either increases in density or the presence of heterospecifics, selection on flowering

onset date is neutral regardless of thermal environment. Furthermore, the results from chapter 2

reveal that selection analyses are dependent on the component of fitness under consideration,

with differences in selection on flowering onset date between thermal environments only

detected when seed production was used as a proxy for fitness. Our results serve to illustrate

how interpretations of the adaptive nature of plasticity can be contingent on the ecological

circumstances or fitness measures from which we evaluate the relationship between phenotype

and fitness.

Do individual traits respond to warming independently or in a correlated manner?

Individuals are made up of a suite of traits with varying degrees of plasticity and adaptive

value, and these characters may respond to environmental conditions independently or in an

integrated manner. The detection of individual plastic responses among traits indicates that

selection may act independently and effectively on each trait. With many species encountering

novel selection pressures as the climate changes, we must determine whether the ‘mosaic nature

of plasticity’ (Ghalambor et al 2007) fosters adaptive developmental responses across the life

cycle.

The independent responses of sequential reproductive life history traits to warming have

been found in several systems (Post et al. 2008; Haggerty & Galloway 2011), suggesting that

flexibility in associations between phenological traits may be common. In chapter 2, we found

that the onset dates of budding and flowering did not respond to increases in temperature

independently of previously expressed traits (emergence and budding onset dates, respectively).

In contrast, the greatest and most variable degree of independent plasticity was for the onset date

of fruiting, a trait that is largely ignored in studies of climate change (but see Peñuelas et al.

2002). In chapters 2 and 4, we saw that advances in first flowering date did not correspond with

   

131  

shifts towards smaller plant size, a constraint commonly found in many plant species (Dorn &

Mitchell-Olds 1991). These studies reveal adaptive trait combinations in C. fasciculata that

could be achieved through adaptive evolution, and also identify scenarios where correlations

between traits might be genetic and challenge evolutionary processes.

Plasticity in the onset date of flowering is often regarded as a representation of how plant

reproduction as a whole is expected to respond to changes in the environment. The consideration

of flowering onset as a proxy for patterns of reproductive phenology explicitly, and perhaps

naively, assumes that the schedule of flower deployment will follow in suite after the first flower

blooms. In chapter 3, we showed that the shape of display schedules is plastic and independent

of plasticity in flowering onset date. Furthermore, plasticity in flower deployment and floral

longevity seems to be governed by seasonal variation in temperature, which in itself is projected

to be modified by climate change (Stocker et al. 2013). The novel experiments presented in this

thesis demonstrate the dynamic nature of plasticity, and the value of insights gained by adopting

a cumulative life cycle view of phenotypic responses to warming.

Future directions and implications for assisted colonization

In this thesis, I provide evidence that interpretations of warming-induced phenotypic

plasticity and subsequent fitness effects are dependant on the species or populations under

investigation, the traits or trait combinations being examined, the fitness components being

considered, and the competitive environment in which data are being acquired. The context-

dependent nature of phenotypic responses to warming warrants further investigation in order to

improve predictive abilities and maximize the effectiveness of conservation strategies.

This work is among the first to experimentally test hypotheses concerning general factors

that may limit the success of assisted colonization programs. The very limited previous

empirical work on this topic has varied in scope, objective, and motivation (Hewitt et al. 2011;

Pedlar et al. 2012). We need a standardized framework for assessing the successful

establishment of relocated populations and for integrating general inferences among separate

experimental trials. The ecological and evolutionary factors limiting range expansion have been

well explored in a wide array of species (Sexton et al. 2009; Hargreaves et al. 2014), and a

comprehensive review of this body of literature in the context of assisted colonization may reveal

   

132  

specific circumstances that may jeopardize or augment the success of future relocations. We

would benefit from studies that explore the feasibility of relocations within natural communities,

that monitor population growth rates over long time frames, and that assess the necessity and

feasibility of simultaneous relocations of mutualist species pairs. Experiments further assessing

the repercussions of relocating species that rely on photoperiodic and thermal cues for

development may be of particular interest for any considering relocations across latitudes.

Assisted gene flow, often termed genetic rescue, is implemented with the goal of

combining genotypes to expand genetic variation and facilitate evolutionary adaptation in

response to changing environmental conditions (Aitken & Whitlock 2013). In plants, the success

of this program is dependent upon the degree of phenological overlap and subsequent

opportunities for pollen exchange between populations. Currently, we can employ measures of

phenological synchrony to gauge the potential for mating opportunities between groups, as we

did in chapter 2 of this thesis. However, many factors contribute to differences between

predicted and realized gene flow, including pollinator behavior, plasticity in expected patterns of

flower deployment or longevity, and variation in fruit initiation and abortion within individuals.

In chapter 3, we demonstrated the potential for plasticity in floral display schedules to reduce the

strength of phenological assortative mating within populations. That is, mating opportunities

between early- and late-flowering individuals are closer to random than what we observe in other

species. The success of assisted gene flow may be greater in species with populations that

express similar plastic responses in patterns of flowering phenology, as seen in C. fasciculata.

Many plant species exhibit a decline in fruit set probability as plants age (Austen et al.

2015), which indicates that flowering overlap alone is not a guarantee the successful genetic

exchange between populations. Gene flow for loci contributing to flowering phenology may be

non-random (Weis 2015), and this bias can work with or against selection, depending on

phenological differences between populations and the local optimum flowering time. Efforts are

underway to gauge the success of assisted gene flow by refining estimates of realized mating

opportunities between the C. fasciculata populations studied here. This work will incorporate

daily flowering schedules and declines in fruit-set probability to estimate the degree of symmetry

in pollen exchange between populations (Wadgymar & Weis, in preparation). This work will

   

133  

provide a framework for those wishing to assess realized gene flow rates when attempting to

genetically rescue a local population with phenologically divergent migrants.

In this thesis, we present the results of a preliminary investigation of how early and late-

flowering species differ in their plastic responses to warming, and whether distinct populations

of a single species exhibit varying degrees of plasticity. We could expand our awareness of the

effects of climate change on phenological traits if more studies monitored traits and fitness at the

individual level, which would allow for phenotypic selection analyses and assessments of

adaptive plasticity. Future work could further investigate variations in plasticity and selection

regimes among populations and species. We advocate for the continued exploration of the

potential for competition to mediate the effects of increases in temperature. For instance, in

chapter 4, we found that competition did not modify phenotypic responses to warming in

communities where species markedly varied in patterns growth and development. Competition

theory predicts that the intensity of competition increases with species similarity, and

experiments that focused on competition intensity (instead of competitive asymmetry) may yield

different results.

References cited

Aitken SN, Whitlock MC (2013) Assisted Gene Flow to Facilitate Local Adaptation to Climate

Change. The Annual Review of Ecology, Evolution, and Systematics, 44, 367-388.

Aldridge G, Inouye DW, Forrest JRK, Barr WA, Miller-Rushing AJ (2011) Emergence of a mid-

season period of low floral resources in a montane meadow ecosystem associated with

climate change. Journal of Ecology, 99, 905-913.

Austen EJ, Forrest JRK, Weis AE (2015) Within-plant variation in reproductive investment:

consequences for selection on flowering time. Journal of Evolutionary Biology, 28, 65-79.

Bertin RI (2008) Plant phenology and distribution in relation to recent climate change. Journal of

the Torrey Botanical Society, 135, 126-146.

Dorn LA, Mitchell-Olds T (1991) Genetics of Brassica campestris. 1. Genetic constraints on

evolution of life-history characters. Evolution, 371-379.

   

134  

Etterson JR, Shaw RG (2001) Constraint to Adaptive Evolution in Response to Global Warming.

Science, 294, 151-154.

Fitter AH, Fitter RSR (2002) Rapid Changes in Flowering Time in British Plants. Science, 296,

1689-1691.

Geber MA, Griffen LR (2003) Inheritance and natural selection on functional traits.

International Journal of Plant Sciences, 164, S21-S42.

Ghalambor CK, McKay JK, Carroll SP, Reznick DN (2007) Adaptive versus non-adaptive

phenotypic plasticity and the potential for contemporary adaptation in new environments.

Functional Ecology, 21, 394-407.

Haggerty BP, Galloway LF (2011) Response of individual components of reproductive

phenology to growing season length in a monocarpic herb. Journal of Ecology, 99, 242-253.

Hargreaves AL, Samis KE, Eckert CG (2014) Are species’ range limits simply niche limits writ

large? A review of transplant experiments beyond the range. The American Naturalist, 183,

157-173..

Hewitt N, Klenk N, Smith AL, Bazely DR, Yan N, Wood S, MacLellan JI, Lipsig-Mumme C,

Henriques I (2011) Taking stock of the assisted migration debate. Biological Conservation,

144, 2560-2572.

Loveless MD, Hamrick JL (1984) Ecological determinants of genetic structure in plant

populations. Annual Review of Ecology and Systematics, 15, 65-95.

Menzel A, Sparks TH, Estrella N, et al. (2006) European phenological response to climate

change matches the warming pattern. Global Change Biology, 12, 1969-1976.

Munguía-Rosas MA, Ollerton J, Parra-Tabla V, De-Nova JA (2011) Meta-analysis of phenotypic

selection on flowering phenology suggests that early flowering plants are favored. Ecology

Letters, 14, 511-521.

Parmesan C (2007) Influences of species, latitudes and methodologies on estimates of

phenological response to global warming. Global Change Biology, 13, 1860-1972.

Pedlar JH, McKenney DW, Aubin I, et al., (2012) Placing Forestry in the Assisted Migration

Debate. BioScience, 62, 835-842.

   

135  

Peñuelas J, Filella I, Comas P (2002) Changed plant and animal life cycles from 1952 to 2000 in

the Mediterranean region. Global Change Biology, 8, 531-544.

Pigliucci M (2001) Phenotypic plasticity: beyond nature and nurture. JHU Press.

Post ES, Pedersen C, Wilmers CC, Forchhammer MC (2008) Phenological sequences reveal

aggregate life history response to climatic warming. Ecology, 89, 363-370.

Scheiner SM (1993) Genetics and evolution of phenotypic plasticity. Annual Review of Ecology

and Systematics, 24, 35-68.

Sexton JP, McIntyre PJ, Angert AL, Rice KJ (2009) Evolution and ecology of species range

limits. The Annual Review of Ecology, Evolution, and Systematics, 40, 415-436.

Stearns SC (1989) The evolutionary significance of phenotypic plasticity. BioScience, 39, 436-

445.

Stocker TF, Qin D, Plattner GK, et al. (2013) IPCC, 2013: climate change 2013: the physical

science basis. Contribution of working group I to the fifth assessment report of the

intergovernmental panel on climate change.

Weis AE (2015) On the potential strength and consequences for non-random gene flow caused

by local adaptation in flowering time. Journal of Evolutionary Biology (in press)

Weis AE, Nardone E, Fox GA (2014) The strength of assortative mating for population date and

its basis in individual variation in flowering schedule. Journal of Evolutionary Biology, 27:

2138:2151.

Willis CG, Ruhfel B, Primack RB, et al. (2008) Phylogenetic patterns of species loss in

Thoreau’s woods are driven by climate change. Proceedings of the National Academy of

Sciences, 105, 17029-17033.

JOHN WILEY AND SONS LICENSETERMS AND CONDITIONS

Jun 27, 2015

This Agreement between Susana M Wadgymar ("You") and John Wiley and Sons ("JohnWiley and Sons") consists of your license details and the terms and conditions provided byJohn Wiley and Sons and Copyright Clearance Center.

License Number 3657310464152

License date Jun 27, 2015

Licensed Content Publisher John Wiley and Sons

Licensed Content Publication Global Change Biology

Licensed Content Title The success of assisted colonization and assisted gene flow dependson phenology

Licensed Content Author Susana M. Wadgymar,Matthew N. Cumming,Arthur E. Weis

Licensed Content Date May 28, 2015

Pages 1

Type of use Dissertation/Thesis

Requestor type Author of this Wiley article

Format Electronic

Portion Full article

Will you be translating? No

Title of your thesis /dissertation

Climate change and reproductive phenology: context-dependentresponses to increases in temperature and implications for assistedcolonization

Expected completion date Jul 2015

Expected size (number ofpages)

150

Requestor Location Susana M Wadgymar25 Willcocks St.University of TorontoDept. of Ecology & Evolutionary BiologyToronto, ON M4P1L6CanadaAttn: Susana M Wadgymar

Billing Type Invoice

Billing Address Susana M Wadgymar25 Willcocks St.University of TorontoDept. of Ecology & Evolutionary Biology

Rightslink Printable License https://s100.copyright.com/App/PrintableLicenseFrame.jsp?pub...

1 of 7 6/27/15 5:09 PM

Toronto, ON M4P1L6CanadaAttn: Susana M Wadgymar

Total 0.00 USD

Terms and Conditions

TERMS AND CONDITIONS

This copyrighted material is owned by or exclusively licensed to John Wiley & Sons, Inc. orone of its group companies (each a"Wiley Company") or handled on behalf of a society withwhich a Wiley Company has exclusive publishing rights in relation to a particular work(collectively "WILEY"). By clicking �accept� in connection with completing thislicensing transaction, you agree that the following terms and conditions apply to thistransaction (along with the billing and payment terms and conditions established by theCopyright Clearance Center Inc., ("CCC's Billing and Payment terms and conditions"), atthe time that you opened your Rightslink account (these are available at any time athttp://myaccount.copyright.com).

Terms and Conditions

The materials you have requested permission to reproduce or reuse (the "WileyMaterials") are protected by copyright.

You are hereby granted a personal, non-exclusive, non-sub licensable (on astand-alone basis), non-transferable, worldwide, limited license to reproduce theWiley Materials for the purpose specified in the licensing process. This license is for aone-time use only and limited to any maximum distribution number specified in thelicense. The first instance of republication or reuse granted by this licence must becompleted within two years of the date of the grant of this licence (although copiesprepared before the end date may be distributed thereafter). The Wiley Materials shallnot be used in any other manner or for any other purpose, beyond what is granted inthe license. Permission is granted subject to an appropriate acknowledgement given tothe author, title of the material/book/journal and the publisher. You shall also duplicatethe copyright notice that appears in the Wiley publication in your use of the WileyMaterial. Permission is also granted on the understanding that nowhere in the text is apreviously published source acknowledged for all or part of this Wiley Material. Anythird party content is expressly excluded from this permission.

With respect to the Wiley Materials, all rights are reserved. Except as expresslygranted by the terms of the license, no part of the Wiley Materials may be copied,modified, adapted (except for minor reformatting required by the new Publication),translated, reproduced, transferred or distributed, in any form or by any means, and noderivative works may be made based on the Wiley Materials without the priorpermission of the respective copyright owner. You may not alter, remove or suppress

Rightslink Printable License https://s100.copyright.com/App/PrintableLicenseFrame.jsp?pub...

2 of 7 6/27/15 5:09 PM

in any manner any copyright, trademark or other notices displayed by the WileyMaterials. You may not license, rent, sell, loan, lease, pledge, offer as security, transferor assign the Wiley Materials on a stand-alone basis, or any of the rights granted toyou hereunder to any other person.

The Wiley Materials and all of the intellectual property rights therein shall at all timesremain the exclusive property of John Wiley & Sons Inc, the Wiley Companies, ortheir respective licensors, and your interest therein is only that of having possession ofand the right to reproduce the Wiley Materials pursuant to Section 2 herein during thecontinuance of this Agreement. You agree that you own no right, title or interest in orto the Wiley Materials or any of the intellectual property rights therein. You shall haveno rights hereunder other than the license as provided for above in Section 2. No right,license or interest to any trademark, trade name, service mark or other branding("Marks") of WILEY or its licensors is granted hereunder, and you agree that youshall not assert any such right, license or interest with respect thereto.

NEITHER WILEY NOR ITS LICENSORS MAKES ANY WARRANTY ORREPRESENTATION OF ANY KIND TO YOU OR ANY THIRD PARTY,EXPRESS, IMPLIED OR STATUTORY, WITH RESPECT TO THE MATERIALSOR THE ACCURACY OF ANY INFORMATION CONTAINED IN THEMATERIALS, INCLUDING, WITHOUT LIMITATION, ANY IMPLIEDWARRANTY OF MERCHANTABILITY, ACCURACY, SATISFACTORYQUALITY, FITNESS FOR A PARTICULAR PURPOSE, USABILITY,INTEGRATION OR NON-INFRINGEMENT AND ALL SUCH WARRANTIESARE HEREBY EXCLUDED BY WILEY AND ITS LICENSORS AND WAIVEDBY YOU

WILEY shall have the right to terminate this Agreement immediately upon breach ofthis Agreement by you.

You shall indemnify, defend and hold harmless WILEY, its Licensors and theirrespective directors, officers, agents and employees, from and against any actual orthreatened claims, demands, causes of action or proceedings arising from any breachof this Agreement by you.

IN NO EVENT SHALL WILEY OR ITS LICENSORS BE LIABLE TO YOU ORANY OTHER PARTY OR ANY OTHER PERSON OR ENTITY FOR ANYSPECIAL, CONSEQUENTIAL, INCIDENTAL, INDIRECT, EXEMPLARY ORPUNITIVE DAMAGES, HOWEVER CAUSED, ARISING OUT OF OR INCONNECTION WITH THE DOWNLOADING, PROVISIONING, VIEWING ORUSE OF THE MATERIALS REGARDLESS OF THE FORM OF ACTION,WHETHER FOR BREACH OF CONTRACT, BREACH OF WARRANTY, TORT,NEGLIGENCE, INFRINGEMENT OR OTHERWISE (INCLUDING, WITHOUTLIMITATION, DAMAGES BASED ON LOSS OF PROFITS, DATA, FILES, USE,BUSINESS OPPORTUNITY OR CLAIMS OF THIRD PARTIES), AND WHETHER

Rightslink Printable License https://s100.copyright.com/App/PrintableLicenseFrame.jsp?pub...

3 of 7 6/27/15 5:09 PM

OR NOT THE PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCHDAMAGES. THIS LIMITATION SHALL APPLY NOTWITHSTANDING ANYFAILURE OF ESSENTIAL PURPOSE OF ANY LIMITED REMEDY PROVIDEDHEREIN.

Should any provision of this Agreement be held by a court of competent jurisdiction tobe illegal, invalid, or unenforceable, that provision shall be deemed amended toachieve as nearly as possible the same economic effect as the original provision, andthe legality, validity and enforceability of the remaining provisions of this Agreementshall not be affected or impaired thereby.

The failure of either party to enforce any term or condition of this Agreement shall notconstitute a waiver of either party's right to enforce each and every term and conditionof this Agreement. No breach under this agreement shall be deemed waived orexcused by either party unless such waiver or consent is in writing signed by the partygranting such waiver or consent. The waiver by or consent of a party to a breach ofany provision of this Agreement shall not operate or be construed as a waiver of orconsent to any other or subsequent breach by such other party.

This Agreement may not be assigned (including by operation of law or otherwise) byyou without WILEY's prior written consent.

Any fee required for this permission shall be non-refundable after thirty (30) daysfrom receipt by the CCC.

These terms and conditions together with CCC�s Billing and Payment terms andconditions (which are incorporated herein) form the entire agreement between you andWILEY concerning this licensing transaction and (in the absence of fraud) supersedesall prior agreements and representations of the parties, oral or written. This Agreementmay not be amended except in writing signed by both parties. This Agreement shall bebinding upon and inure to the benefit of the parties' successors, legal representatives,and authorized assigns.

In the event of any conflict between your obligations established by these terms andconditions and those established by CCC�s Billing and Payment terms andconditions, these terms and conditions shall prevail.

WILEY expressly reserves all rights not specifically granted in the combination of (i)the license details provided by you and accepted in the course of this licensingtransaction, (ii) these terms and conditions and (iii) CCC�s Billing and Paymentterms and conditions.

This Agreement will be void if the Type of Use, Format, Circulation, or RequestorType was misrepresented during the licensing process.

Rightslink Printable License https://s100.copyright.com/App/PrintableLicenseFrame.jsp?pub...

4 of 7 6/27/15 5:09 PM

This Agreement shall be governed by and construed in accordance with the laws of theState of New York, USA, without regards to such state�s conflict of law rules. Anylegal action, suit or proceeding arising out of or relating to these Terms and Conditionsor the breach thereof shall be instituted in a court of competent jurisdiction in NewYork County in the State of New York in the United States of America and each partyhereby consents and submits to the personal jurisdiction of such court, waives anyobjection to venue in such court and consents to service of process by registered orcertified mail, return receipt requested, at the last known address of such party.

WILEY OPEN ACCESS TERMS AND CONDITIONS

Wiley Publishes Open Access Articles in fully Open Access Journals and in Subscriptionjournals offering Online Open. Although most of the fully Open Access journals publishopen access articles under the terms of the Creative Commons Attribution (CC BY) Licenseonly, the subscription journals and a few of the Open Access Journals offer a choice ofCreative Commons Licenses:: Creative Commons Attribution (CC-BY) license CreativeCommons Attribution Non-Commercial (CC-BY-NC) license and Creative CommonsAttribution Non-Commercial-NoDerivs (CC-BY-NC-ND) License. The license type isclearly identified on the article.

Copyright in any research article in a journal published as Open Access under a CreativeCommons License is retained by the author(s). Authors grant Wiley a license to publish thearticle and identify itself as the original publisher. Authors also grant any third party the rightto use the article freely as long as its integrity is maintained and its original authors, citationdetails and publisher are identified as follows: [Title of Article/Author/Journal Title andVolume/Issue. Copyright (c) [year] [copyright owner as specified in the Journal]. Links tothe final article on Wiley�s website are encouraged where applicable.

The Creative Commons Attribution License

The Creative Commons Attribution License (CC-BY) allows users to copy, distribute andtransmit an article, adapt the article and make commercial use of the article. The CC-BYlicense permits commercial and non-commercial re-use of an open access article, as long asthe author is properly attributed.

The Creative Commons Attribution License does not affect the moral rights of authors,including without limitation the right not to have their work subjected to derogatorytreatment. It also does not affect any other rights held by authors or third parties in thearticle, including without limitation the rights of privacy and publicity. Use of the articlemust not assert or imply, whether implicitly or explicitly, any connection with, endorsementor sponsorship of such use by the author, publisher or any other party associated with thearticle.

For any reuse or distribution, users must include the copyright notice and make clear toothers that the article is made available under a Creative Commons Attribution license,

Rightslink Printable License https://s100.copyright.com/App/PrintableLicenseFrame.jsp?pub...

5 of 7 6/27/15 5:09 PM

linking to the relevant Creative Commons web page.

To the fullest extent permitted by applicable law, the article is made available as is andwithout representation or warranties of any kind whether express, implied, statutory orotherwise and including, without limitation, warranties of title, merchantability, fitness for aparticular purpose, non-infringement, absence of defects, accuracy, or the presence orabsence of errors.

Creative Commons Attribution Non-Commercial License

The Creative Commons Attribution Non-Commercial (CC-BY-NC) License permits use,distribution and reproduction in any medium, provided the original work is properly citedand is not used for commercial purposes.(see below)

Creative Commons Attribution-Non-Commercial-NoDerivs License

The Creative Commons Attribution Non-Commercial-NoDerivs License (CC-BY-NC-ND)permits use, distribution and reproduction in any medium, provided the original work isproperly cited, is not used for commercial purposes and no modifications or adaptations aremade. (see below)

Use by non-commercial users

For non-commercial and non-promotional purposes, individual users may access, download,copy, display and redistribute to colleagues Wiley Open Access articles, as well as adapt,translate, text- and data-mine the content subject to the following conditions:

The authors' moral rights are not compromised. These rights include the right of"paternity" (also known as "attribution" - the right for the author to be identified assuch) and "integrity" (the right for the author not to have the work altered in such away that the author's reputation or integrity may be impugned).

Where content in the article is identified as belonging to a third party, it is theobligation of the user to ensure that any reuse complies with the copyright policies ofthe owner of that content.

If article content is copied, downloaded or otherwise reused for non-commercialresearch and education purposes, a link to the appropriate bibliographic citation(authors, journal, article title, volume, issue, page numbers, DOI and the link to thedefinitive published version on Wiley Online Library) should be maintained.Copyright notices and disclaimers must not be deleted.

Any translations, for which a prior translation agreement with Wiley has not beenagreed, must prominently display the statement: "This is an unofficial translation of anarticle that appeared in a Wiley publication. The publisher has not endorsed thistranslation."

Rightslink Printable License https://s100.copyright.com/App/PrintableLicenseFrame.jsp?pub...

6 of 7 6/27/15 5:09 PM

Use by commercial "for-profit" organisations

Use of Wiley Open Access articles for commercial, promotional, or marketing purposesrequires further explicit permission from Wiley and will be subject to a fee. Commercialpurposes include:

Copying or downloading of articles, or linking to such articles for furtherredistribution, sale or licensing;

Copying, downloading or posting by a site or service that incorporates advertisingwith such content;

The inclusion or incorporation of article content in other works or services (other thannormal quotations with an appropriate citation) that is then available for sale orlicensing, for a fee (for example, a compilation produced for marketing purposes,inclusion in a sales pack)

Use of article content (other than normal quotations with appropriate citation) byfor-profit organisations for promotional purposes

Linking to article content in e-mails redistributed for promotional, marketing oreducational purposes;

Use for the purposes of monetary reward by means of sale, resale, licence, loan,transfer or other form of commercial exploitation such as marketing products

Print reprints of Wiley Open Access articles can be purchased from:corporatesales@wiley.com

Further details can be found on Wiley Online Library http://olabout.wiley.com/WileyCDA/Section/id-410895.html

Other Terms and Conditions:

v1.9

Questions? customercare@copyright.com or +1-855-239-3415 (toll free in the US) or+1-978-646-2777.

Rightslink Printable License https://s100.copyright.com/App/PrintableLicenseFrame.jsp?pub...

7 of 7 6/27/15 5:09 PM

JOHN WILEY AND SONS LICENSETERMS AND CONDITIONS

Jun 27, 2015

This Agreement between Susana M Wadgymar ("You") and John Wiley and Sons ("JohnWiley and Sons") consists of your license details and the terms and conditions provided byJohn Wiley and Sons and Copyright Clearance Center.

License Number 3657310341897

License date Jun 27, 2015

Licensed Content Publisher John Wiley and Sons

Licensed Content Publication Journal of Ecology

Licensed Content Title Simultaneous pulsed flowering in a temperate legume: causes andconsequences of multimodality in the shape of floral displayschedules

Licensed Content Author Susana M. Wadgymar,Emily J. Austen,Matthew N. Cumming,ArthurE. Weis

Licensed Content Date Jan 9, 2015

Pages 12

Type of use Dissertation/Thesis

Requestor type Author of this Wiley article

Format Electronic

Portion Full article

Will you be translating? No

Title of your thesis /dissertation

Climate change and reproductive phenology: context-dependentresponses to increases in temperature and implications for assistedcolonization

Expected completion date Jul 2015

Expected size (number ofpages)

150

Requestor Location Susana M Wadgymar25 Willcocks St.University of TorontoDept. of Ecology & Evolutionary BiologyToronto, ON M4P1L6CanadaAttn: Susana M Wadgymar

Billing Type Invoice

Rightslink Printable License https://s100.copyright.com/App/PrintableLicenseFrame.jsp?pub...

1 of 8 6/27/15 5:07 PM

Billing Address Susana M Wadgymar25 Willcocks St.University of TorontoDept. of Ecology & Evolutionary BiologyToronto, ON M4P1L6CanadaAttn: Susana M Wadgymar

Total 0.00 USD

Terms and Conditions

TERMS AND CONDITIONS

This copyrighted material is owned by or exclusively licensed to John Wiley & Sons, Inc. orone of its group companies (each a"Wiley Company") or handled on behalf of a society withwhich a Wiley Company has exclusive publishing rights in relation to a particular work(collectively "WILEY"). By clicking �accept� in connection with completing thislicensing transaction, you agree that the following terms and conditions apply to thistransaction (along with the billing and payment terms and conditions established by theCopyright Clearance Center Inc., ("CCC's Billing and Payment terms and conditions"), atthe time that you opened your Rightslink account (these are available at any time athttp://myaccount.copyright.com).

Terms and Conditions

The materials you have requested permission to reproduce or reuse (the "WileyMaterials") are protected by copyright.

You are hereby granted a personal, non-exclusive, non-sub licensable (on astand-alone basis), non-transferable, worldwide, limited license to reproduce theWiley Materials for the purpose specified in the licensing process. This license is for aone-time use only and limited to any maximum distribution number specified in thelicense. The first instance of republication or reuse granted by this licence must becompleted within two years of the date of the grant of this licence (although copiesprepared before the end date may be distributed thereafter). The Wiley Materials shallnot be used in any other manner or for any other purpose, beyond what is granted inthe license. Permission is granted subject to an appropriate acknowledgement given tothe author, title of the material/book/journal and the publisher. You shall also duplicatethe copyright notice that appears in the Wiley publication in your use of the WileyMaterial. Permission is also granted on the understanding that nowhere in the text is apreviously published source acknowledged for all or part of this Wiley Material. Anythird party content is expressly excluded from this permission.

With respect to the Wiley Materials, all rights are reserved. Except as expresslygranted by the terms of the license, no part of the Wiley Materials may be copied,modified, adapted (except for minor reformatting required by the new Publication),

Rightslink Printable License https://s100.copyright.com/App/PrintableLicenseFrame.jsp?pub...

2 of 8 6/27/15 5:07 PM

translated, reproduced, transferred or distributed, in any form or by any means, and noderivative works may be made based on the Wiley Materials without the priorpermission of the respective copyright owner. You may not alter, remove or suppressin any manner any copyright, trademark or other notices displayed by the WileyMaterials. You may not license, rent, sell, loan, lease, pledge, offer as security, transferor assign the Wiley Materials on a stand-alone basis, or any of the rights granted toyou hereunder to any other person.

The Wiley Materials and all of the intellectual property rights therein shall at all timesremain the exclusive property of John Wiley & Sons Inc, the Wiley Companies, ortheir respective licensors, and your interest therein is only that of having possession ofand the right to reproduce the Wiley Materials pursuant to Section 2 herein during thecontinuance of this Agreement. You agree that you own no right, title or interest in orto the Wiley Materials or any of the intellectual property rights therein. You shall haveno rights hereunder other than the license as provided for above in Section 2. No right,license or interest to any trademark, trade name, service mark or other branding("Marks") of WILEY or its licensors is granted hereunder, and you agree that youshall not assert any such right, license or interest with respect thereto.

NEITHER WILEY NOR ITS LICENSORS MAKES ANY WARRANTY ORREPRESENTATION OF ANY KIND TO YOU OR ANY THIRD PARTY,EXPRESS, IMPLIED OR STATUTORY, WITH RESPECT TO THE MATERIALSOR THE ACCURACY OF ANY INFORMATION CONTAINED IN THEMATERIALS, INCLUDING, WITHOUT LIMITATION, ANY IMPLIEDWARRANTY OF MERCHANTABILITY, ACCURACY, SATISFACTORYQUALITY, FITNESS FOR A PARTICULAR PURPOSE, USABILITY,INTEGRATION OR NON-INFRINGEMENT AND ALL SUCH WARRANTIESARE HEREBY EXCLUDED BY WILEY AND ITS LICENSORS AND WAIVEDBY YOU

WILEY shall have the right to terminate this Agreement immediately upon breach ofthis Agreement by you.

You shall indemnify, defend and hold harmless WILEY, its Licensors and theirrespective directors, officers, agents and employees, from and against any actual orthreatened claims, demands, causes of action or proceedings arising from any breachof this Agreement by you.

IN NO EVENT SHALL WILEY OR ITS LICENSORS BE LIABLE TO YOU ORANY OTHER PARTY OR ANY OTHER PERSON OR ENTITY FOR ANYSPECIAL, CONSEQUENTIAL, INCIDENTAL, INDIRECT, EXEMPLARY ORPUNITIVE DAMAGES, HOWEVER CAUSED, ARISING OUT OF OR INCONNECTION WITH THE DOWNLOADING, PROVISIONING, VIEWING ORUSE OF THE MATERIALS REGARDLESS OF THE FORM OF ACTION,WHETHER FOR BREACH OF CONTRACT, BREACH OF WARRANTY, TORT,

Rightslink Printable License https://s100.copyright.com/App/PrintableLicenseFrame.jsp?pub...

3 of 8 6/27/15 5:07 PM

NEGLIGENCE, INFRINGEMENT OR OTHERWISE (INCLUDING, WITHOUTLIMITATION, DAMAGES BASED ON LOSS OF PROFITS, DATA, FILES, USE,BUSINESS OPPORTUNITY OR CLAIMS OF THIRD PARTIES), AND WHETHEROR NOT THE PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCHDAMAGES. THIS LIMITATION SHALL APPLY NOTWITHSTANDING ANYFAILURE OF ESSENTIAL PURPOSE OF ANY LIMITED REMEDY PROVIDEDHEREIN.

Should any provision of this Agreement be held by a court of competent jurisdiction tobe illegal, invalid, or unenforceable, that provision shall be deemed amended toachieve as nearly as possible the same economic effect as the original provision, andthe legality, validity and enforceability of the remaining provisions of this Agreementshall not be affected or impaired thereby.

The failure of either party to enforce any term or condition of this Agreement shall notconstitute a waiver of either party's right to enforce each and every term and conditionof this Agreement. No breach under this agreement shall be deemed waived orexcused by either party unless such waiver or consent is in writing signed by the partygranting such waiver or consent. The waiver by or consent of a party to a breach ofany provision of this Agreement shall not operate or be construed as a waiver of orconsent to any other or subsequent breach by such other party.

This Agreement may not be assigned (including by operation of law or otherwise) byyou without WILEY's prior written consent.

Any fee required for this permission shall be non-refundable after thirty (30) daysfrom receipt by the CCC.

These terms and conditions together with CCC�s Billing and Payment terms andconditions (which are incorporated herein) form the entire agreement between you andWILEY concerning this licensing transaction and (in the absence of fraud) supersedesall prior agreements and representations of the parties, oral or written. This Agreementmay not be amended except in writing signed by both parties. This Agreement shall bebinding upon and inure to the benefit of the parties' successors, legal representatives,and authorized assigns.

In the event of any conflict between your obligations established by these terms andconditions and those established by CCC�s Billing and Payment terms andconditions, these terms and conditions shall prevail.

WILEY expressly reserves all rights not specifically granted in the combination of (i)the license details provided by you and accepted in the course of this licensingtransaction, (ii) these terms and conditions and (iii) CCC�s Billing and Paymentterms and conditions.

Rightslink Printable License https://s100.copyright.com/App/PrintableLicenseFrame.jsp?pub...

4 of 8 6/27/15 5:07 PM

This Agreement will be void if the Type of Use, Format, Circulation, or RequestorType was misrepresented during the licensing process.

This Agreement shall be governed by and construed in accordance with the laws of theState of New York, USA, without regards to such state�s conflict of law rules. Anylegal action, suit or proceeding arising out of or relating to these Terms and Conditionsor the breach thereof shall be instituted in a court of competent jurisdiction in NewYork County in the State of New York in the United States of America and each partyhereby consents and submits to the personal jurisdiction of such court, waives anyobjection to venue in such court and consents to service of process by registered orcertified mail, return receipt requested, at the last known address of such party.

WILEY OPEN ACCESS TERMS AND CONDITIONS

Wiley Publishes Open Access Articles in fully Open Access Journals and in Subscriptionjournals offering Online Open. Although most of the fully Open Access journals publishopen access articles under the terms of the Creative Commons Attribution (CC BY) Licenseonly, the subscription journals and a few of the Open Access Journals offer a choice ofCreative Commons Licenses:: Creative Commons Attribution (CC-BY) license CreativeCommons Attribution Non-Commercial (CC-BY-NC) license and Creative CommonsAttribution Non-Commercial-NoDerivs (CC-BY-NC-ND) License. The license type isclearly identified on the article.

Copyright in any research article in a journal published as Open Access under a CreativeCommons License is retained by the author(s). Authors grant Wiley a license to publish thearticle and identify itself as the original publisher. Authors also grant any third party the rightto use the article freely as long as its integrity is maintained and its original authors, citationdetails and publisher are identified as follows: [Title of Article/Author/Journal Title andVolume/Issue. Copyright (c) [year] [copyright owner as specified in the Journal]. Links tothe final article on Wiley�s website are encouraged where applicable.

The Creative Commons Attribution License

The Creative Commons Attribution License (CC-BY) allows users to copy, distribute andtransmit an article, adapt the article and make commercial use of the article. The CC-BYlicense permits commercial and non-commercial re-use of an open access article, as long asthe author is properly attributed.

The Creative Commons Attribution License does not affect the moral rights of authors,including without limitation the right not to have their work subjected to derogatorytreatment. It also does not affect any other rights held by authors or third parties in thearticle, including without limitation the rights of privacy and publicity. Use of the articlemust not assert or imply, whether implicitly or explicitly, any connection with, endorsementor sponsorship of such use by the author, publisher or any other party associated with thearticle.

Rightslink Printable License https://s100.copyright.com/App/PrintableLicenseFrame.jsp?pub...

5 of 8 6/27/15 5:07 PM

For any reuse or distribution, users must include the copyright notice and make clear toothers that the article is made available under a Creative Commons Attribution license,linking to the relevant Creative Commons web page.

To the fullest extent permitted by applicable law, the article is made available as is andwithout representation or warranties of any kind whether express, implied, statutory orotherwise and including, without limitation, warranties of title, merchantability, fitness for aparticular purpose, non-infringement, absence of defects, accuracy, or the presence orabsence of errors.

Creative Commons Attribution Non-Commercial License

The Creative Commons Attribution Non-Commercial (CC-BY-NC) License permits use,distribution and reproduction in any medium, provided the original work is properly citedand is not used for commercial purposes.(see below)

Creative Commons Attribution-Non-Commercial-NoDerivs License

The Creative Commons Attribution Non-Commercial-NoDerivs License (CC-BY-NC-ND)permits use, distribution and reproduction in any medium, provided the original work isproperly cited, is not used for commercial purposes and no modifications or adaptations aremade. (see below)

Use by non-commercial users

For non-commercial and non-promotional purposes, individual users may access, download,copy, display and redistribute to colleagues Wiley Open Access articles, as well as adapt,translate, text- and data-mine the content subject to the following conditions:

The authors' moral rights are not compromised. These rights include the right of"paternity" (also known as "attribution" - the right for the author to be identified assuch) and "integrity" (the right for the author not to have the work altered in such away that the author's reputation or integrity may be impugned).

Where content in the article is identified as belonging to a third party, it is theobligation of the user to ensure that any reuse complies with the copyright policies ofthe owner of that content.

If article content is copied, downloaded or otherwise reused for non-commercialresearch and education purposes, a link to the appropriate bibliographic citation(authors, journal, article title, volume, issue, page numbers, DOI and the link to thedefinitive published version on Wiley Online Library) should be maintained.Copyright notices and disclaimers must not be deleted.

Any translations, for which a prior translation agreement with Wiley has not beenagreed, must prominently display the statement: "This is an unofficial translation of an

Rightslink Printable License https://s100.copyright.com/App/PrintableLicenseFrame.jsp?pub...

6 of 8 6/27/15 5:07 PM

article that appeared in a Wiley publication. The publisher has not endorsed thistranslation."

Use by commercial "for-profit" organisations

Use of Wiley Open Access articles for commercial, promotional, or marketing purposesrequires further explicit permission from Wiley and will be subject to a fee. Commercialpurposes include:

Copying or downloading of articles, or linking to such articles for furtherredistribution, sale or licensing;

Copying, downloading or posting by a site or service that incorporates advertisingwith such content;

The inclusion or incorporation of article content in other works or services (other thannormal quotations with an appropriate citation) that is then available for sale orlicensing, for a fee (for example, a compilation produced for marketing purposes,inclusion in a sales pack)

Use of article content (other than normal quotations with appropriate citation) byfor-profit organisations for promotional purposes

Linking to article content in e-mails redistributed for promotional, marketing oreducational purposes;

Use for the purposes of monetary reward by means of sale, resale, licence, loan,transfer or other form of commercial exploitation such as marketing products

Print reprints of Wiley Open Access articles can be purchased from:corporatesales@wiley.com

Further details can be found on Wiley Online Library http://olabout.wiley.com/WileyCDA/Section/id-410895.html

Other Terms and Conditions:

v1.9

Questions? customercare@copyright.com or +1-855-239-3415 (toll free in the US) or+1-978-646-2777.

Rightslink Printable License https://s100.copyright.com/App/PrintableLicenseFrame.jsp?pub...

7 of 8 6/27/15 5:07 PM

Rightslink Printable License https://s100.copyright.com/App/PrintableLicenseFrame.jsp?pub...

8 of 8 6/27/15 5:07 PM