climate change and reproductive phenology: context ... · species beyond their current range...
TRANSCRIPT
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:[email protected]
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? [email protected] 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:[email protected]
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? [email protected] 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