EXPERIMENTAL TESTS OF RAPID ECO-EVOLUTIONARY DYNAMICS IN A PLANT- HERBIVORE SYSTEM

EXPERIMENTAL TESTS OF RAPID ECO-EVOLUTIONARY

DYNAMICS IN A PLANT- HERBIVORE SYSTEM

CHAPTER ONE: A GENERAL INTRODUCTION TO THE

STUDY OF RAPID ECO-EVOLUTIONARY DYNAMICS

ABSTRACT

In this introductory chapter I first briefly review the historical

development of the ‘new’ field of eco-evolutionary dynamics and

summarize current theoretical and empirical findings. I then present

the objectives of my dissertation and how these aim to address

outstanding questions in this emerging field of study.

BACKGROUND

Interactions between Ecology and Evolution

Ecology and evolutionary biology are fields that share a long and

complex history with periods of either strong integration or

independent development (Collins 1986). It has long been recognized

that ecology and evolution influence one another. For example, Darwin

(1859) described how ecological interactions, especially competition,

shapes the selective environments in which species evolve, and how

evolution will impact extinction and species distributions. Few

biologists would argue that evolution and ecology do not interact at

least on long timescales. As the fields of population ecology and

population genetics developed in the 1920s, they also grew apart

(Collins 1986). Population ecologists often ignored evolution, since

evolution was perceived to have little

impact, and population geneticists focused on how the ecological

environment causes evolution (Ford 1964). The classic evolutionary

paradigm was established where ecology caused evolution (arrow 1 in

Fig. 1.1). The alternate causal pathway, where evolution impacts

ecology (arrow 2 in Fig. 1.1), was recognized but usually relegated to

long-term effects. For example speciation will impact species diversity

(Sax and Gaines 2003, Pelletier et al. 2009). This separation is still

enduring today as population genetics and population ecology are still

largely studied separately (Levins et al. 2003, Lewontin et al. 2003,

Hairston et al. 2005, Saccheri and Hanski 2006, Pelletier et al. 2009).

Two common assumptions in the ecological literature often

exclude any possible impact that evolution might have on short-term

population dynamics, dynamics occurring within a few dozen

generations or within 100 years (Thompson 1998, Hairston et al.

2005). First, populations are assumed to be genetically homogeneous

(Cappuccino and Price 1995, Sibly and Hone 2002, Hairston et al.

2005). Yet, within population genetic variation in ecologically

important traits has been demonstrated repeatedly since the 1960s

(Ayala 1968, Berry et al. 1978). Moreover, such variation can

significantly influence population dynamics in the laboratory (Schlager

1963, Leips et al. 2000) and in nature (Hanski and Saccheri 2006,

Hazell et al. 2006). Another, common assumption is that evolution

occurs on a much slower (and thus separate) timescale than short-

term ecological processes. Even if genetic variation is considered,

evolution is ignored because it is perceived to be too slow to have an

effect (Slobodkin 1980,

Endler 1991, Cappuccino and Price 1995, Hairston et al. 2005, Pelletier

et al. 2009). Thus most ecological studies assume that evolution is not

occurring and utilize non-evolving trait values in their models.

This assumption has now been challenged by dozens of studies

documenting rapid evolutionary changes in nature occurring on

‘ecological timescales’, sometimes within a few generations (Dyer

1968, Thompson 1998, Hendry and Kinnison 1999, Bone and Farres

2001, Reznick and Ghalambor 2001, Ashley et al. 2003). Rapid

evolution has been documented in many ecologically important traits

such as life-history traits, foraging traits, morphological traits,

phenology, and enemy resistance traits and has been documented in

many major taxonomic groups including fish, arthropods, microbes,

mammals, vascular and non-vascular plants, lizards, amphibians, and

mollusks. These findings have led some researchers to reexamine how

evolution and ecology interact given that the arbitrary distinction of

ecological and evolutionary time is no longer valid.

Rapid Eco-Evolutionary Dynamics

Although biologists have been studying the interactions between

ecology and evolution the reason this topic is currently receiving

renewed interest is that it focuses primarily on interactions occurring

on short timescales that have traditionally been overlooked (reviewed

in Hairston et al. 2005, Fussmann et al.

2007, Johnson and Stinchcombe 2007, Pelletier et al. 2009, Schoener

2011).

This current field of study is called ‘Eco-Evolutionary Dynamics’ and

emphasizes the reciprocal and concurrent interactions between

ecology and evolution and is the focus of my dissertation. Eco-

evolutionary dynamics differs from other sub- categories of

evolutionary-ecology because it focuses not on how the ecological

environment causes evolution, i.e. ecological genetics (Ford 1964), but

on how genetic variation and rapid evolution impact ecology

(population dynamics, community structure, ecosystem functioning…).

Yet the ultimate goal of eco- evolutionary dynamics is to study how

evolutionary and ecological dynamics causally influence each other at

the same time and how this might alter both ecological and

evolutionary outcomes (full cyclical causality, Fig. 1.1; Bull et al. 2006,

Kokko and Lopez-Sepulcre 2007, Ezard et al. 2009, Pelletier et al.

2009).

Previous Studies of Eco-Evolutionary Dynamics

Theoretical studies, investigating eco-evolution dynamics, date

back at least 50 years (Pimentel 1961) and since then have diversified

into different approaches, based on very different biological

assumptions (reviewed in Bergelson et al. 2001, Abrams 2005, Day

2005, Fussmann et al. 2007). Eco- evolutionary dynamic models allow

ecological and evolutionary processes to interact and assess how such

interactions influence ecological and evolutionary dynamics. Such eco-

evolutionary dynamics have been shown to influence the trajectory of

growth of single populations (Anderson and King 1970), the density

and stability of victim-exploiter systems (Pimentel 1961, Fussmann et

al. 2003,

Bull et al. 2006, Duffy and Sivars-Becker 2007), the structure of multi-

species communities (Loeuille and Leibold 2008), and even ecosystem

processes (Loeuille et al. 2002). Overall these theoretical studies, by

comparing models with and without evolution, overwhelmingly

demonstrate that eco-evolutionary dynamics can qualitatively and

quantitatively alter ecological and evolutionary outcomes (Day 2005,

Fussmann et al. 2007). Yet, empirical studies of such interactions are

still very rare and most of this body of theory remains untested

(Fussmann et al. 2007).

One sub-discipline of eco-evolutionary dynamics, called

‘Community Genetics,’ (Antonovics 1992) explores how genetic variation,

mostly in plants, influences the structure and composition of arthropod

communities they support, how it influences competition with other

plant species, and ecosystem fluxes (Agrawal 2003, Johnson and

Agrawal 2005, Whitham et al. 2006, Hughes et al. 2008, Bailey et al.

2009). These studies show the strong influence of intraspecific

genetically based variation and suggest that if evolution changed the

frequency of plant genotypes then evolutionary dynamics would

influence ecological dynamics (Johnson et al. 2009). A similar approach

consists of post-hoc comparisons of the ecological properties of

populations thought to have undergone recent evolutionary

diversification. Such studies assess how evolution has altered life

history traits (Reznick and Bryga 1996), population dynamics (Hanski

and Saccheri 2006), community structure (Post et al. 2008) and

ecosystem processes (Bassar et al. 2010). These two empirical

approaches

however have yet to quantify the impact of evolution as it occurs and

thus could miss dynamic aspects of the eco-evolutionary interactions.

Empirical studies have also studied try to explain ecological

changes using models that include evolutionary change (Anderson and

May 1982, Tuda 1998, Sinervo et al. 2000, Hairston et al. 2005, Duffy

and Sivars-Becker 2007, Ezard et al. 2009). In such studies, rapid

evolution is usually strongly correlated and ecological predictions are

usually significantly improved by including evolutionary dynamics. For

example, an ecological model correctly predicted the start date of

epidemic parasitic outbreaks in natural Daphnia populations but failed

to predict their termination (Duffy et al. 2005). An evolutionary-

ecological model where susceptibility evolved, correctly predicted the

date of termination (Duffy and Sivars-Becker 2007). Although these

observational studies establish the generality of evolutionary-

feedback, they remain correlational and only suggest causality.

The Experimental Approach in Eco-Evolutionary Dynamics

The experimental approach quantifies the causal impact of rapid

evolution as populations evolve compared to populations that cannot.

This approach addresses the limitations of the methods above since

they tract evolution as it occurs and can establish causality. Pimentel

first used this approach to show how the population dynamics of a

parasitoid wasp were changed as its housefly host evolved resistance

compared to a non-evolving control (Pimentel et al.

1963, Pimentel and Al-Hafidh 1965, Pimentel 1968). By replacing the

control housefly population every generation, he prevented the

evolution of resistance. Rapid evolution within a few years in the host

reduced the parasitoid’s population size and variance even though

host population size was held constant. This experimental approach

has only been attempted a handful of times (Bohannan and Lenski

1999, Fussmann et al. 2003, Agashe 2009, Terhorst et al. 2010).

Some experimental systems have quantified the full feedback cycle,

where both ecological and evolutionary dynamics influence each other

concurrently. In Yoshida et al.’s (2003) study of rotifers and algae in

chemostats, rapid evolution in algae caused the predator-prey

population dynamics to change from being 1/4 out-of-phase to being

perfectly out-of-phase. This was caused by a change in the frequency

of resistant algal clones that increased the density of algae while

reducing that of the rotifer predator (evolution impacting ecology).

Then because of frequency and density-dependent clonal selection the

faster growing, but less defended algal clone, increased in frequency

leading to increased predation and lower algal density resetting the

cycle (ecology impacting evolution; Shertzer et al. 2002, Yoshida et al.

2003, Yoshida et al. 2004). These empirical studies demonstrate how

evolutionary dynamics and genetic variation influence the dynamics

and outcome of short-term ecological phenomena and argue

convincingly for causality, but only under carefully controlled

laboratory conditions.

DISSERTATION OBJECTIVES

My overall objective was to develop a study system wherein I

could manipulate rapid evolution experimentally, altering its

occurrence and rate. I could then use this system to experimentally

assess the conditions under which eco-evolutionary dynamics are

occurring, dissect how they operate, quantify their impact, and

eventually determine their importance compared to other ecological

processes. Although much progress is being made in the study of eco-

evolutionary dynamics many outstanding questions remain that my

dissertation aims to address.

Plant-herbivore interactions are thought to be one of the most

common and important ecological interactions, generating much of the

species and phenotypic diversity in nature as well as having immense

economic importance (Ehrlich and Raven 1964, Futuyma and Agrawal

2009). Yet none of the experimental model systems used in eco-

evolutionary dynamics, where evolution is manipulated, to my

knowledge utilize a plant-herbivore system (except one study of

herbivorous spider mites but it was not framed in this context Agrawal

2000). All other studies use predator-prey (Fussmann et al. 2003,

Yoshida et al. 2003, Terhorst et al. 2010) or host-parasitoid systems

(Pimentel 1968, Tuda 1998, Bohannan and Lenski 2000) exclusively.

The eco-evolutionary dynamics in plant-herbivore systems could

differ greatly from those observed in predator-prey or host-parasitoid

interactions. Herbivore dynamics, especially those of pest species can

often be in non-

equilibrium states (Wallner 1987, Karley et al. 2004). Also insect

herbivores usually have a weaker impact on plant population dynamics

than do predators on predator-prey dynamics because herbivores do

not necessarily kill their host. The magnitude of impact of insect

herbivores on plant population dynamics has been debated for years

(Crawley 1989) and only in the last decade have a dozen or so studies

found support for this (Maron and Crone 2006). Finally, it is important

to study eco-evolutionary dynamics in plant-herbivore systems

because of the immense economic importance these processes might

have if they alter our ability to accurately predict pest population

dynamics and pest evolution. Thus my first objective was to develop a

plant-herbivore system to study eco- evolutionary dynamics. My first

chapter describes the selected study system, consisting of a local

aphid-mustard population. I then genetically characterized the aphid

population by identifying neutral genetic variation that could be used

to identify and track aphid clones. My next objective was to

ecologically characterize the clones and identify ecologically relevant

trait variation that could be manipulated in order to prevent or induce

rapid evolution in future experiments.

Chapter 2 addressed my second objective, which was to quantify the

impact of rapid evolution on concurrent ecological dynamics (arrow 2 in

Fig. 1.1). This aspect of eco-evolutionary dynamics has received less

attention than the impact of ecology on evolution (Bull et al. 2006,

Ezard et al. 2009, Pelletier et al. 2009). In order to establish causality I

utilized the experimental

approach where evolution itself is manipulated and its impact on

ecological dynamics quantified directly. Since this has never been

done using plant- herbivores I did so under partially controlled

laboratory conditions. This experiment will help determine whether

rapid evolution can impact short-term ecological dynamics. Previous

investigations of eco-evolutionary dynamics study periods representing

dozens of generations since they mostly use very fast reproducing

microorganisms (Bohannan and Lenski 2000, Fussmann et al. 2003,

Yoshida et al. 2003, Terhorst et al. 2010). I will explore the impact of

even faster bouts of rapid evolution. My experiments focus on a single

growing season of the host (5-6 aphid generations or less), which could

have important implications for pest management in agricultural

systems

To my knowledge all experimental test of eco-evolutionary

dynamics, where evolution is manipulated, have been conducted

under highly controlled and simplified laboratory environments

(Fussmann et al. 2007). Although laboratory experiments establish the

potential impact of rapid evolution, field experiments are crucial

because ecological context can influence both ecological and

evolutionary processes (Holt 2005). Experiments conducted in the wild

within realistic communities encompass more realistic levels of biotic

and abiotic variation as well as gene flow. These confounding factors

could impose different selective pressures, altering the rate or

direction of evolution itself, or they could interfere with the manner in

which rapid evolution impacts population dynamics,

e.g. by altering the strength of density regulation (Saccheri and Hanski

2006). All

of these problems imply that eco-evolutionary dynamics should ideally

be studied in the wild since non-intuitive results could occur that differ

significantly from predictions based only upon laboratory experiments.

Even strong laboratory results could be overwhelmed by

environmental variation in nature.

An important issue to consider is the source of the study

population. To properly study the importance of eco-evolutionary

dynamics it is important to use genotypes that actually interact in

nature. Some studies (for an example see Agashe 2009) magnify the

genetic variation in their experimental populations by using genotypes

from multiple independent populations. If more genetic variation is

used than is commonly found in wild populations, then the

experimental populations might evolve more quickly which could

overestimate the importance of eco-evolutionary dynamics. To avoid

such a bias I only collected genotypes from a single population where

the experiments were conducted. Thus my third goal, the focus of

Chapter 3, is to address whether eco-evolutionary dynamics have

significant impacts in the wild in the face of environmental variation using

a local population. If eco-evolutionary dynamics have strong impacts in

nature this has important implications for the study of population

dynamics since evolution is traditionally not considered in these

studies. For example I know of no pest population dynamic models

that incorporate pest evolution within the growing season.

Experimental studies quantifying interspecific evolutionary-

feedback have only been attempted in the laboratory (Fussmann et al.

2007) and those in nature

have been observational or focus on genetic variation and not

evolution per se. It thus remains an open issue whether rapid evolution

can have interspecific impacts in plant-herbivore systems. Chapter 3

will also address my fourth objective, which is to experimentally determine

whether aphid rapid evolution significantly impacts their host plant’s

fitness. This objective will be addressed in all experiments but

especially in Chapter 3 since a field experiment permits more accurate

quantification of host fitness.

Many ecological forces have strong influences on ecological

dynamics (e.g. interspecific competition, population density…).

Ecologists should focus their limited resources on understanding and

quantifying important drivers of ecological dynamics. Recent studies

and many reviews claim that rapid evolution should be included in this

list (Thompson 1998) yet very few studies have addressed this

question explicitly (Johnson and Stinchcombe 2007). In community

genetics only a few experiments quantify the relative importance of

genotypic variation versus other ecological forces such as habitat

variation and induced plant resistance (Johnson and Agrawal 2005,

McGuire and Johnson 2006). The only eco-evolutionary dynamic study

that compares the relative impact of evolution itself compared to

ecological processes, is a correlation approach proposed by Hairston

(2005). Thus my fifth goal is to experimentally test the relative impact of

rapid on population growth rate compared to that of intraspecific density

(Chapter 4).

Eco-evolutionary dynamics are defined as the reciprocal

interactions between short-term ecological and evolution dynamics

timescales (Kokko and Lopez-Sepulcre 2007). My objectives thus far

have focused on quantifying the less studied half of that interaction,

how rapid evolution impacts ecology. Yet much more complicate

dynamics are possible if both arrows of causality are occurring

concurrently (Fig. 1.1). Ecological changes induced by rapid evolution

could alter future bouts of evolution by changing the selective

environment experienced by the target organism. Thus my sixth and

final objective is to assess both arrows of causality in the same

experiment and determine whether both ecological and evolutionary

dynamics are influencing each other (Chapter 4). If such an interaction is

occurring this implies that much more complex dynamics are possible

in this system and would put into question models that do not couple

ecological and evolutionary changes together.

Importance of Eco-Evolutionary Dynamics

The growing interest in eco-evolutionary dynamics stems from

its potential implications for many aspects of biology. Theoretical

studies suggest that such a process could greatly alter not only

evolutionary but also ecological interactions, dynamics, and outcomes.

Acquiring a better understanding of this process should improve our

understanding and our predictive ability (Duffy and Sivars- Becker

2007, Pelletier et al. 2009). Given the ever increasing examples of

rapid evolution, considering evolutionary-ecological interactions as a

working

hypothesis (Thompson 1998) might also provide important insight into

many applied issues, e.g. disease epidemics (Real et al. 2005),

fisheries management (Law 2000), bio-control (Hufbauer and Roderick

2005), and conservation biology (Ashley et al. 2003).

Figure 1.1: Diagram of Eco-Evolutionary Dynamics

Diagram representing eco-evolutionary dynamics illustrating the cyclical

causality between evolution and ecology dynamics.

LITERATURE CITED

Abrams, P. A. 2005. 'Adaptive Dynamics' vs. 'adaptive dynamics'.

Journal of Evolutionary Biology 18:1162-65.

Agashe, D. 2009. The stabilizing effect of intraspecific genetic

variation on population dynamics in novel and ancestral habitats.

The American Naturalist 174:255-67.

Agrawal, A. 2000. Host-range evolution: Adaptation and trade-offs in

fitness of mites on alternative hosts. Ecology 81:500-08.

Agrawal, A. 2003. Community genetics: New insights into community

ecology by integrating population genetics. Ecology 84:543-44.

Anderson, R. and R. May. 1982. Coevolution of hosts and parasites.

Parasitology 85:411-26.

Anderson, W. W. and C. E. King. 1970. Age-specific selection.

Proceedings of the National Academy of Sciences 66:780-86.

Antonovics, J. 1992. Towards community genetics. P. 426-49 in R. S. Fritz

and

E. L. Simms, editors. Ecology and Evolution of Plant Resistance to

Herbivores and Pathogens: Ecology, Evolution, and Genetics. .

University of Chicago Press.

Ashley, M. V., M. F. Willson, O. R. W. Pergams, D. J. O'Dowd, S. M.

Gende, et al. 2003. Evolutionarily enlightened management.

Biological Conservation 111:115-23.

Ayala, F. 1968. Genotype, environment, and population

numbers. Science 162:1453-9.

Bailey, J. K., J. A. Schweitzer, F. Ubeda, J. Koricheva, C. J. LeRoy, et al.

2009. From genes to ecosystems: a synthesis of the effects of

plant genetic factors across levels of organization. Philos T R Soc B

364:1607-16.

Bassar, R. D., M. C. Marshall, A. López-Sepulcre, E. Zandona, S. K. Auer, et

al.

2010. Local adaptation in Trinidadian guppies alters ecosystem

processes. Proceedings of the National Academy of Sciences 107:3616-

21.

Bergelson, J., G. Dwyer, and J. Emerson. 2001. Models and data on

plant- enemy coevolution. Annual Review of Genetics 35:469-99.

Berry, R., R. Anderson, B. Turner, and L. Taylor. 1978. Genetical

factors in animal population dynamics. P. 53-80 Population

Dynamics.

Bohannan, B. and R. Lenski. 1999. Effect of prey heterogeneity on the

response of a model food chain to resource enrichment. The

American Naturalist 153:73-82.

Bohannan, B. and R. Lenski. 2000. Linking genetic change to

community evolution: insights from studies of bacteria and

bacteriophage. Ecology Letters 3:362-77.

Bone, E. and A. Farres. 2001. Trends and rates of microevolution in plants.

Genetica 112-113:165-82.

Bull, J. J., J. Millstein, J. Orcutt, and H. A. Wichman. 2006. Evolutionary

feedback mediated through population density, illustrated with viruses

in chemostats.

American Naturalist 167:E39-E51.

Cappuccino, N. and P. W. Price. 1995. Population Dynamics: New

Approaches and Synthesis. Academic Press.

Collins, J. 1986. Evolutionary ecology and the use of natural selection

in ecological theory. Journal of the History of Biology 19:257-88.

Crawley, M. 1989. Insect Herbivores and Plant Population Dynamics.

Annual Review of Entomology 34:531-62.

Darwin, C. 1859. On The Origin of Species By Means of Natural Selection.

Day, T. 2005. Modelling the ecological context of evolutionary change:

Déjà vu or something new? P. 273-309 in K. Cuddington and B. E.

Beisner, editors.

Ecological Paradigms Lost Routes of Theory Change. Elsevier

Academic Press.

Duffy, M., S. Hall, A. Tessier, and M. Huebner. 2005. Selective

predators and their parasitized prey: Are epidemics in Zooplankton

under top-down control? Limnology and Oceanography 50:412-20.

Duffy, M. A. and L. Sivars-Becker. 2007. Rapid evolution and

ecological host- parasite dynamics. Ecology Letters 10:44-53.

Dyer, K. F. 1968. Evolution observed -- Some examples of evolution

occurring in historical times. Journal of Biological Education 2:317-

38.

Ehrlich, P. R. and P. H. Raven. 1964. Butterflies and Plants: A Study in

Coevolution. Evolution 18:586-608.

Endler, J. A. 1991. Genetic heterogeneity and ecology. P. 315–34 in R. J.

Berry,

T. J. Crawford, and G. M. Hewitt, editors. Genes in Ecology: The 33rd

Symposium of the British Ecological Society. Blackwell Scientific

Publications.

Ezard, T. H. G., S. D. Cote, and F. Pelletier. 2009. Eco-evolutionary

dynamics: disentangling phenotypic, environmental and population

fluctuations.

Philosophical Transactions of the Royal Society B: Biological

Sciences 364:1491-98.

Ford, E. B. 1964. Ecological Genetics. Chapman and Hall.

Fussmann, G. F., S. P. Ellner, and N. G. J. Hairston. 2003. Evolution as a

critical component of plankton dynamics. Proceedings of the Royal

Society of London. Series B, Biological Sciences 270:1015-22.

Fussmann, G. F., M. Loreau, and P. A. Abrams. 2007. Eco-evolutionary

dynamics of communities and ecosystems. Functional Ecology

21:465-77.

Futuyma, D. J. and A. A. Agrawal. 2009. Macroevolution and the

biological diversity of plants and herbivores. Proceedings of the

National Academy of Sciences of the United States of America

106:18054-61.

Hairston, N. G. J., S. P. Ellner, M. A. Geber, T. Yoshida, and J. A. Fox.

2005. Rapid evolution and the convergence of ecological and

evolutionary time. Ecology Letters 8:1114-27.

Hanski, I. and I. Saccheri. 2006. Molecular-level variation affects

population growth in a butterfly metapopulation. PLoS Biology

4:e129.

Hazell, S. P., I. A. D. McClintock, and M. D. E. Fellowes. 2006.

Intraspecific heritable variation in life-history traits can alter the

outcome of interspecific competition among insect herbivores. Basic

and Applied Ecology 7:215-23.

Hendry, A. P. and M. T. Kinnison. 1999. Perspective: The pace of modern

life: Measuring rates of contemporary microevolution. Evolution

53:1637-53.

Holt, R. D. 2005. On the integration of community ecology and

evolutionary biology: Historical perspectives and current prospects.

P. 235-72 in K. Cuddington and B. Beisner, editors. Ecological

Paradigms Lost Routes of Theory Change. Elsevier.

Hufbauer, R. A. and G. K. Roderick. 2005. Microevolution in biological

control: Mechanisms, patterns, and processes. Biological Control

35:227-39.

Hughes, A. R., B. D. Inouye, M. T. J. Johnson, N. Underwood, and M.

Vellend. 2008. Ecological consequences of genetic diversity.

Ecology Letters 11:609- 23.

Johnson, M. T. J. and A. Agrawal. 2005. Plant genotype and

environment interact to shape a diverse arthropod community on

evening primrose (Oenothera biennis). Ecology 86:874-85.

Johnson, M. T. J. and J. R. Stinchcombe. 2007. An emerging synthesis

between community ecology and evolutionary biology. Trends in

Ecology and Evolution 22:250-57.

Johnson, M. T. J., M. Vellend, and J. R. Stinchcombe. 2009. Evolution in

plant populations as a driver of ecological changes in arthropod

communities.

Philosophical Transactions of the Royal Society B: Biological

Sciences 364:1593-605.

Karley, A., W. Parker, J. Pitchford, and A. Douglas. 2004. The mid-

season crash in aphid populations: why and how does it occur?

Ecological Entomology 29:383-88.

Kokko, H. and A. Lopez-Sepulcre. 2007. The ecogenetic link between

demography and evolution: can we bridge the gap between theory

and data? Ecology Letters 10:773-82.

Law, R. 2000. Fishing, selection, and phenotypic evolution. ICES

Journal of Marine Science 57:659-68.

Leips, J., J. Travis, and F. H. Rodd. 2000. Genetic influences on

experimental population dynamics of the least killifish. Ecological

Monographs 70:289-309.

Levins, R., R. S. Singh, and M. K. Uyenoyama. 2003. Towards a

population biology, still. P. 21–47 The Evolution of Population

Biology.

Lewontin, R., R. S. Singh, and M. K. Uyenoyama. 2003. Building a

science of population biology. P. 7–20 The Evolution of Population

Biology.

Loeuille, N. and M. A. Leibold. 2008. Ecological consequences of

evolution in plant defenses in a metacommunity. Theoretical

Population Biology 74:34-45.

Loeuille, N., M. Loreau, and R. Ferriere. 2002. Consequences of plant-

herbivore coevolution on the dynamics and functioning of

ecosystems. Journal of Theoretical Biology 217:369-81.

Maron, J. L. and E. Crone. 2006. Herbivory: effects on plant

abundance, distribution and population growth. Proceedings of the

Royal Society B: Biological Sciences 273:2575-84.

McGuire, R. and M. T. J. Johnson. 2006. Plant genotype and induced

responses affect resistance to herbivores on evening primrose

(Oenothera biennis).

Ecological Entomology 31:20-31.

Pelletier, F., D. Garant, and A. P. Hendry. 2009. Eco-evolutionary

dynamics.

Philosophical Transactions of the Royal Society B: Biological

Sciences 364:1483-89.

Pimentel, D. 1961. Animal population regulation by the genetic feed-

back mechanism. The American Naturalist 95:65-79.

Pimentel, D. 1968. Population regulation and genetic feedback.

Science 159:1432-37.

Pimentel, D. and R. Al-Hafidh. 1965. Ecological control of a parasite

population by genetic evolution in the parasite-host system. Annals

of the Entomological Society of America 58:1-6.

Pimentel, D., W. P. Nagel, and J. L. Madden. 1963. Space-time

structure of the environment and the survival of parasite-host

systems. The American Naturalist 97:141-67.

Post, D. M., E. P. Palkovacs, E. G. Schielke, and S. I. Dodson. 2008.

Intraspecific variation in a predator affects community structure and

cascading trophic interactions. Ecology 89:2019-32.

Real, L., J. Henderson, R. Biek, J. Snaman, T. Jack, et al. 2005. Unifying

the spatial population dynamics and molecular evolution of

epidemic rabies virus.

Proceedings of the National Academy of Sciences of the United

States of America 102:12107-11.

Reznick, D. N. and H. A. Bryga. 1996. Life-history evolution in guppies

(Poecilia reticulata: Poeciliidae): V. Genetic basis of parallelism in life

histories.

American Naturalist 147:339-59.

Reznick, D. N. and C. K. Ghalambor. 2001. The population ecology of

contemporary adaptations: what empirical studies reveal about the

conditions that promote adaptive evolution. Genetica 112-113:183-

98.

Saccheri, I. and I. Hanski. 2006. Natural selection and population dynamics.

Trends in Ecology & Evolution 21:341-47.

Sax, D. and S. Gaines. 2003. Species diversity: from global decreases to

local increases. Trends in Ecology & Evolution 18:561-66.

Schlager, G. 1963. The ecological genetics of the mutant Sooty in

populations of Tribolium castaneum. Evolution 17:254-73.

Schoener, T. W. 2011. The newest synthesis: Understanding the

interplay of evolutionary and ecological dynamics. Science

331:426-29.

Shertzer, K., S. Ellner, G. F. Fussmann, and N. G. J. Hairston. 2002.

Predator- prey cycles in an aquatic microcosm: testing hypotheses

of mechanism.

Journal of Animal Ecology 71:802-15.

Sibly, R. and J. Hone. 2002. Population growth rate and its determinants:

an overview. Philosophical Transactions of the Royal Society of London

Series B- Biological Sciences 357:1153-70.

Sinervo, B., E. Svensson, and T. Comendant. 2000. Density

cycles and an offspring quantity and quality game driven by

natural selection. Nature 406:985-88.

Slobodkin, L. B. 1980. Growth and Regulation of Animal Populations. 2nd

edition.

Dover Publications Inc.

Terhorst, C. P., T. E. Miller, and D. R. Levitan. 2010. Evolution of prey

in ecological time reduces the effect size of predators in

experimental microcosms. Ecology 91:629-36.

Thompson, J. N. 1998. Rapid evolution as an ecological process. Trends

in Ecology & Evolution 13:329-32.

Tuda, M. 1998. Evolutionary character changes and population

responses in an insect host-parasitoid experimental system.

Researches on Population Ecology 40:293-99.

Wallner, W. 1987. FACTORS AFFECTING INSECT POPULATION-DYNAMICS:

Differences between outbreak and non-outbreak species. Annual

Review of Entomology 32:317-40.

Whitham, T. G., J. K. Bailey, J. A. Schweitzer, S. M. Shuster, R. K. Bangert,

et al.

2006. A framework for community and ecosystem genetics: from genes

to ecosystems. Nature Reviews Genetics 7:510-23.

Yoshida, T., N. G. J. Hairston, and S. Ellner. 2004. Evolutionary trade-off

between defence against grazing and competitive ability in a simple

unicellular alga, Chlorelia vulgaris. P Roy Soc Lond B Bio 271:1947-

53.

Yoshida, T., L. E. Jones, S. P. Ellner, G. F. Fussmann, and N. G. J.

Hairston. 2003. Rapid evolution drives ecological dynamics in a

predator-prey system. Nature 424:303-06.

CHAPTER TWO: MOLECULAR AND ECOLOGICAL

CHARACTERIZATION OF STUDY SYSTEM

ABSTRACT

I developed a system to experimentally study eco-evolutionary

dynamics in both the laboratory and in the field. I selected the green

peach aphid, Myzus persicae (Sulzer), and a wild mustard host

Hirschfeldia incana (Lagrèze-Fossat), system because the aphid’s

reproductive biology makes it an excellent candidate for experimental

evolution. The preliminary studies presented in this chapter introduce

and characterize the study system for the experiments presented in

future chapters. I first justify why I selected an asexually reproducing

model system. I then present the study system and describe how I

sampled a local population and genetically identified multiple clonal

lineages that were maintained in the greenhouse. I then ecologically

characterized these clones by assessing differences in fitness using

two experiments. I found that aphids differed by as much as 17% in

intrinsic growth rate and selected a subset of three clones to use in

subsequent experiments that directly quantify eco- evolutionary

dynamics.

INTRODUCTION

Experimentally Testing Eco-Evolutionary Dynamics

The overall objective of my dissertation is to test experimentally

the conditions under which rapid evolution interacts with concurrent

ecological dynamics. The key to accomplish this end is to manipulate

the rate of evolution. One commonly used approach is to compare

non-evolved and evolved populations under common garden

conditions in a post-hoc manner (Hanski and Saccheri 2006, Post et al.

2008, Harmon et al. 2009, Bassar et al. 2010). A related approach

consists of repeatedly testing changes in ecological parameters as a

population is evolving (e.g. fitness on a new host Agrawal 2000, the

strength of predation Terhorst et al. 2010) compared to unselected

control populations.

A different approach consists of manipulating only the response

to selection without changing the selective environment. One can thus

compare the ecological properties (e.g., population dynamics,

interspecific interactions, ecosystem effects…) of populations as they

are evolving to non-evolving controls. One method used to prevent

evolution consists of continually replacing the control population under

selection with unselected individuals (Pimentel 1968). This is

unfeasible in most systems because it would alter population

dynamics. An alternate approach, used here, is to manipulate

available genetic variation (see also Yoshida et al. 2003, Agashe

2009). Populations with less or no genetic variation will evolve more

slowly or not at all (Fisher 1930). The

population dynamics in these treatments can be directly compared to

those of populations that have natural levels of genetic variation. Yet,

this approach also has its limitations, including the confounding effect

of inbreeding if the study organism is sexual. Inbreeding as well as

genetic variation is known to impact ecological dynamics (Haag et al.

2002, Hanski and Saccheri 2006, Hughes et al. 2008). However,

naturally asexual organisms do not suffer from these issues.

Evolution in these asexual systems occurs through changes of clonal

frequencies, which also can change mean population trait values (Via

and Shaw 1996). This form of evolution is common in many taxa

including aphids (Via and Shaw 1996, Vorburger 2006), algae

(Fussmann et al. 2003, Yoshida et al. 2003), crustaceans (Lynch 1984),

protozoans (Terhorst et al. 2010), ‘genotypic selection’ in clonally

reproducing plants (Pan and Price 2001), bacteria (Bohannan and

Lenski 2000), and has been observed in snails (Jokela et al.

2003) and fish (Vrijenhoek and Pfeiler 1997).

Study System

I developed a local aphid-mustard study system specifically for

this dissertation research. To my surprise no other eco-evolutionary

experimental system uses an insect-plant model system where the

focal species is the insect (although Agrawal (2000) used herbivorous

spider mites). Other studies use predator-prey (Fussmann et al. 2003,

Yoshida et al. 2003, Terhorst et al. 2010) or host-parasitoid systems

(Tuda and Iwasa 1998, Bohannan and Lenski 2000).

This is surprising given that a growing body of research in eco-

evolutionary dynamics is focused on how plant genetics impacts insect

communities (‘community genetics’ reviewed in Whitham et al. 2006,

Hughes et al. 2008).

I selected the green peach aphid Myzus persicae (Sulzer) (Insecta:

Hemiptera: Aphididae) for my studies. M. persicae has a global

distribution but probably originated in China (Blackman 1974). It is

considered the world’s most important crop pest due to its enormous

host-range (over 40 families of plants) and its ability to transmit over

100 plant viruses (Mackauer and Way 1976, Blackman and Eastop

2000). Given its importance, its population dynamics, basic biology

and ecology are well studied (Van Emden et al. 1969, Mackauer and

Way 1976, Ro and Long 1999, Blackman and Eastop 2000, Karley et

al.

2003). Moreover many genetic resources are available (Sloane et al.

2001, Wilson et al. 2003, Wilson et al. 2004). This host-alternating

aphid reproduces through cyclical parthenogenesis which is the most

common reproductive mode in aphids (Dixon et al. 1989). It

reproduces sexually on its primary host (Prunus species) and the

offspring migrate to secondary hosts (crops or wild plants) where they

reproduce asexually for multiple generations or remain asexual

switching between a series of secondary hosts (Van Emden et al.

1969). The asexual reproductive phase in aphids is apomictic, i.e. that

offspring are exact genetic copies of their mothers except for rare

point mutations (Wilson et al.

2003). This implies that microsatellite markers can reliably identify

clonal lineages. Clonal lines differ in their propensity to reproduce

sexually (Blackman

1974) and in certain warmer climates these clones are favored since

the sexual over-wintering egg stage is not required for survival

(Vorburger et al. 2003).

Myzus persicae is highly amenable to experimental evolution

because of its short generation time (5-8 days depending on

temperature), high fecundity, ease of culture, and facultative asexual

reproduction. Aphids in their asexual phase have telescoping

generations, i.e. grand-daughters are developing within adult females

thus shortening development time from birth to maturity (Dixon et al.

1989). Also this aphid has a large magnitude of genetic variation

between clones identified in multiple traits, including intrinsic growth

rate (Weber 1985b, a, Vorburger 2005). Given their cyclical

parthenogenetic reproduction and migration of clones over larger

distances, spring populations are replete with clonal variation (Dickson

and Laird 1967, Vorburger 2006). These populations naturally undergo

rapid evolution, through clonal selection, within a matter of weeks

leading to significant changes in clonal frequencies (Vorburger 2006).

This occurs in many aphid species (de Barro et al. 1995, Sunnucks et

al. 1997, Fuller et al. 1999). Such changes in clonal frequencies alter

mean phenotypic trait values within weeks (Via and Shaw 1996).

Multiyear temporal studies have identified highly successful and

widespread genotypes (Wilson et al. 2003, Vorburger 2005, Vorburger

2006). In agricultural settings certain clones have evolved pesticide

resistance and rapidly increase in frequency (Foster et al.

2002). Although evolution and ecological dynamics are often studied in

aphids, to my knowledge no studies have looked at how rapid

evolution within a season

might impact the aphid’s population dynamics (e.g., growth rate, peak

densities, peak density date) even in the laboratory (Roush and

McKenzie 1987).

I selected the short-pod mustard Hirschfeldia incana (Lagreze-

Fossat) (Brassicaceae), formerly Brassica geniculata, as a host species.

This mustard, probably of Mediterranean origin, has invaded Western

Europe, Australia, New Zealand, and the South Western United States

(Horovitz and Galil 1972). It was selected for its local abundance

throughout the field site and because it is easily reared. This primarily

self-incompatible (Horovitz and Galil 1972) annual plant completes its

growth from seed to seed within a few months permitting me to

quantify the impact of the aphids on its host.

The experiments focus on a single population of aphids collected

at the University of California Motte-Rimrock Reserve near Perris,

California (MRR,

http://nrs.ucop.edu/reserves/motte/motte_rimrock.htm). The UC

Natural Reserve System provides accessible yet protected natural

sites that are excellent locations to study non-agricultural species

interactions in natural settings. The MRR is particularly good location

to study this insect-plant system since it is very close to the UCR

campus, which reduces travel, facilitates experimental execution, and

permits more thorough sampling. The MRR is a mix of Riversidean

sage scrub habitat, coastal-desert grassland, and willow riparian

thickets (Minnich and Dezzani 1998). All collections of aphids and host

seeds were made within the reserve and the field experiment (Chapter

4) was conducted here. I focused on a single population of aphids as

opposed to

sampling multiple populations because I wanted to assess how local,

i.e. available, genetic variation might lead to eco-evolutionary

interactions. Artificially increasing genetic variation by combining

genotypes from multiple populations that would never interact in

nature would limit my ability to understand the importance of eco-

evolutionary interactions in natural populations (for an example see

Agashe 2009).

Population Sampling

In March 2008, I collected 22 adult female apterous (non-

winged) M. persicae feeding on Hirschfeldia incana from the Motte-

Rimrock Reserve. I sampled early in the season because this is period

when the population should contain the highest number of clonal

lineages from sexual reproduction and migration (Vorburger 2006).

These asexually reproducing females were used to create isofemale

colonies that were maintained in the greenhouse on individually caged

H. incana seedlings under conditions that maintain asexual

reproduction (16hrs light / 8hrs dark) (Blackman 1974). Given that the

clonal composition of aphid population in future experiments was to be

manipulated I genotyped the isofemale colonies to determine whether

they differed genetically.

Clone Molecular Identification

I identified clonal lineages using six published microsatellite

markers (Sloane et al. 2001, Wilson et al. 2004). Microsatellites are

short repeated

genomic sequences (usually 1-5 base pairs) with highly variable

number of repeats and can be used to identify aphid clones (Wilson et

al. 2003). I selected these microsatellite markers over other neutral

genetic markers because they are highly reliable, can be genotyped

cheaply because they do not require sequencing, are co-dominant and

highly polymorphic, and can be multiplexed, meaning that multiple loci

can be genotyped with a single PCR. Because they are used

commonly, the methods for M. persicae were readily available (Sloane

et al. 2001, Vorburger et al. 2003, Wilson and Swenson 2003, Wilson et

al.

2004).

Three individuals from each isofemale colony were genotyped.

PCR methods were modified from previous studies (Sloane et al. 2001,

Wilson et al. 2004). One microsatellite primer from each pair was 5’

labeled with fluorescent dyes. Loci with overlapping lengths were

labeled with dyes of different wavelengths (loci-dye: myz2-6FAM,

myz3-6FAM, M40-6FAM, M86-HEX [Sigma- Aldrich], M49-PET [Applied

Biosystems], and myz9-HEX [Invitrogen]). All six microsatellite were

composed of dinucleotide repeats. DNA from a single aphid was

extracted using 5% Chelex 100 resin (Bio-Rad) and incubated for

35min at 56°C and for 15min at 95°C. Samples were then centrifuged

at 12000rpm for 2min and the supernatant retained. PCR reactions

were 10µL in volume that included 0.6 units of Taq polymerase (New

England Biolabs), 1µL of 10x Mg-free standard reaction buffer, 0.2mM

dNTP, 1µM of forward and reverse primer,

1.5mM MgCl2, and 0.8µL of DNA extract. Touchdown PCR was used to

amplify

these loci. Two different PCR cycling programs were used that follow

those of Sloane et al. (2001) except that one additional amplification

cycle was added to the last annealing temperature. Loci myz2, myz9,

M40, and M86 used PCR program PMS1 and myz3 and M49 used

PMS2. The lengths of PCR products were determined using an ABI

3100 Genetic Analyzer, using the GeneScan 500 LIZ size standard, and

GeneMapper software (Applied Biosystems).

Our genotyping identified 10 genetically unique clonal lineages.

Table 1.1 presents the each clones genotype at these loci.

Clone Maintenance

Multiple clonal populations were maintained in the greenhouse

throughout the dissertation research. Populations were always kept on

H. incana grown from seeds collected at the Motte-Rimrock Reserve.

Initially, in 2008-2009, colonies were maintained on seedlings that

were contained in large plastic jars with screen windows. Every three

or four weeks approximately two dozen aphids were transfer to a fresh

seedling. This was repeated for each clonal colony. Older colonies were

kept as backups. In 2010-2011, I switched to using larger plants (at

the end of the rosette stage) within large cages. These populations

were larger and the plants survived longer. Transfers were still

conducted every three to four weeks but consisted of hundreds of

aphids. The three focal clones (813, 815, and 828) were maintained in

duplicate. All colonies were kept within a partially temperature

controlled greenhouse without humidity control under

natural lighting as well as metal halide lights that extended daylight to

16hrs a day. These conditions maintained asexual reproduction

(Blackman 1974).

Every 4-6 weeks a group of 3-5 aphids from each clonal

population was tested for clonal contamination. If this was detected

individual aphids were used to establish new isofemale colonies and

these were tested until the original clone was rescued. In the hundreds

of tests I never found new alleles/genotypes in the microsattelite loci

tested. In other words, all contaminating aphids were from a known

aphid clonal lineage and none of these had mutations at microsatellite

loci. Due to contamination and greenhouse cooling failures I did,

however, permanently lose some clones.

Population Sampling: Clone Identification and Characterization

For adaptive evolution to occur within a population composed of

different clones these clones must differ in relative fitness. Given that

these aphids grow and reproduce very quickly I was able to assess

fitness by measuring their intrinsic growth rate. The intrinsic growth

rate (rm) is a good index of fitness in this system because populations

typically grow exponentially, then crash as their host plant senesces

(Wallner 1987, Karley et al. 2003, Karley et al. 2004). Intrinsic growth

rate is known to vary greatly between aphid clones within species and

between species (Weber 1985a, Vorburger 2005). The rm of a given

clone is sensitive to the plant species or genotype on which it grows,

so it is commonly used to measure host suitability (Wyatt and White

1977). Intrinsic growth rate,

and how it relates to temperature, is also a key component in

predictive models used in integrative pest management (Guldemond

et al. 1998, Ro and Long 1999).

Intrinsic growth rate is commonly estimated in aphids using full or

truncated life table analyses (Birch 1948, Wyatt and White 1977, Le

Roux et al. 2004).

These methods however have their limitations in that they isolate

individual adults in small clip-cages. This procedure can cause stress

to the insect, harm the plant, alter the microclimate, and restricts

selection of feeding sites (Guldemond et al. 1998). I instead extracted

rm from observed population growth rates on whole caged plants using

population growth models (Vehrs et al. 1992, Guldemond et al. 1998).

These conditions mimic future experimental conditions and are more

informative than life table approaches.

EXPERIMENT 1: ECOLOGICAL CHARACTERIZATION OF M. PERSICAE

CLONES

The objective of this experiment is quantifying intrinsic growth

rates for the unique clonal lineages collected from the Motte-Rimrock

population. H. incana, seeds collected at the Motte-Rimrock Reserve

were grown in small pots (~500mL) using UC soil Mix III, a sand/peat

moss mix supplemented with micronutrients. Once the seedlings, still

in the rosette stage, reached approximately 25cm wide they were

placed in cages within the greenhouse.

Cages were constructed to individually house each plant. A cage

consisted of an

eight liter pot with a wire frame creating a 75cm high dome that held

up thin transparent mesh (Bridal Organza, #664-7242, Jo-Ann). Aphid

clones were cleared of plant-viruses by using the approach suggested

in Raybould et al. (1999). On day 0 of the experiment each plant (the

unit of replication) received 12 apterous third instar M. persicae from a

single clonal lineage. I attempted to test all lineages: however, clone

831 did not have enough aphids to initiate the experiment and clone

820 was contaminated with another clone. Thus I tested 8 different

clonal treatments and each was replicated three times. This

experiment was conducted in a partially temperature controlled

greenhouse (mean daytime temperature = 26.8°C, range = 19°C to

32°C, mean nighttime = 21°C, range = 14°C to 29°C). Additional

lighting to extend light to 16hrs / 8hrs day as a way of maintaining

asexual reproduction (Blackman 1974).

Population size was measured, by counting all aphids, on days 0,

3, 6, 9, 12, and 23. Because it took longer than a full day to count all

the aphids (there were other treatments not presented here), certain

treatments were counted on day 13 and others on days 24 and 25. I

tested for differences in growth rate by fitting an exponential growth

model to the population dynamics observed on days 0 to 13 because

after this day population growth declined. I fit a linear mixed- effect

model (LME) with a linear exponential growth equation. I set a

common intercept (mean density of aphids on day 0) across

treatments. Thus the dependent variable was LN(x) transformed

number of aphids, the fixed effect was aphid treatments and day (as

the main covariate). Given that plants were

repeatedly counted violating the assumption of independent

observations I set unique plant identity as a random effect on

population growth rate and used an autoregressive correlation error

structure (Pinheiro and Bates 2000). Increasing variance through time

was modeled by using a variance function within the LME that

increases with the power of the variance covariate (varPower). All

analyses were implement in R (v. 2.11.1; R Development Core Team

2009) using the nlme package (Pinheiro et al. 2009).

RESULTS

This first experiment revealed significant clonal variation in

intrinsic growth rate (Fig. 2.1, LME, p = 0.002). Clonal lineages differed

by as much as 17%, which causes a range in doubling time of 41 to 48

hours. Over a period of 10 days of exponential growth the fastest

clone should reach 85% higher population size than the slowest clone.

I thus identified clonal lineages that differed in fitness. However, this

experiment had limitations that I wanted to correct.

EXPERIMENT 2: FOCUSED CHARACTERIZATION

Experiment 1 identified clones that differed greatly in fitness but it

had limitations. I repeated the above experiment but improved upon it

by using more replicates (5), by counting more often, and initializing

the population with a stable age distribution. I decided to focus on a

subset of clones that had different growth rates (813, 815, and 828) to

confirm their differences. Clone 820 was selected

also since it was contaminated in the previous experiment and thus its

growth rate was not determined. I was able to re-isolate this clone

before experiment 2.

The methods were very similar to experiment 1 and I here focus

on the differences. H. incana, were grown in larger four liter pots. Once

the seedlings, still in the rosette stage, reached approximately 20cm

wide they were placed in cages within the greenhouse. On day 0 of the

experiment each plant received 10 apterous M. persicae from a single

clonal lineage. From Experiment 1 I estimated a stable age structure

after a few generations. I estimated the mean proportion of aphids in

each growth stage during the last days of exponential growth. I

replicated this stable age distribution by introducing two adults, one

4th, two 3rd, and five 1st or 2nd instars to each plant. Instar stages

were distinguished by size. These treatments were replicated five

times and the position of plants randomized within the partially

temperature controlled greenhouse (mean daytime temperature =

23°C, range = 13°C to 37°C, mean nighttime = 14°C, range = 8°C to

21°C).

Population size was measured by counting all aphids on days 0, 3, 6,

9, 12,

15, 18, 21, 25, 29, and 36. Some treatments were counted one day

later on days 19, 22, 26, and 30 because of time limitations. I fit a LME

exponential growth model to the population dynamics but excluded

day 36 since the plants started senescing. The analysis mirrored that

of experiment 1.

RESULTS

Clones grew exponentially until day 30, reaching densities

between 10 to 20 thousand aphids per plant (Fig. 2.2). Exponential

growth lasted much longer than in the first experiment because the

plants were larger and could grow faster as they had more soil. Also,

aphids grew more slowly probably due to lower temperatures during

the experiment. M. persicae grows faster at a mean fluctuating

temperature of 24°C (Exp. 1) than 18.5°C (Exp. 2) (Davis et al. 2007).

Clonal differences in the second experiment were smaller than the first

but were still highly significantly different among clones. Analysis of

the population dynamics from days 0 to 30 showed that clone 813

grew fastest (daily rm mean ± 1 SE: 0.268 ± 0.002) which was 2.2%

faster (planned contrast, p<0.001) than

clone 815 (0.263 ± 0.004). In turn, clone 815 grew 4.6% faster (p<0.001)

than

clone 828 (0.251 ± 0.004). Thus clone 813 grew 6.9% faster

(p<0.001) than the clone 828. Only three clonal lineages were

required for future experiments and I decided not to use clone 820

(which had a growth rate of 0.251 ± 0.004) since it had a similar

value to clone 828.

An analysis restricted to the first 19 days of the experiment

revealed the same pattern of growth rates but differences between

clones were even larger. Clone 813 grew 9% faster (p<0.001) than the

815 that in grew 8.9% faster than 828 (p=0.005). This implies that 813

grew 18.9% faster (p<0.001) than the 828. These growth rates predict

population densities that differ by as much as 2.3 fold by day 19.

Comparing the analyses of days 0-30 with 0-19 suggest that fitness

differences between clones changed slightly with density.

DISCUSSION

My experiments identified significant genetically based variation

among clones in fitness quantified as exponential growth rate. I

identified up to 17-18.9% variation in fitness depending on the dataset

and analysis. Because these clones grew under controlled

environmental conditions, differences between clones are genetically

based (Via and Shaw 1996). Such intraspecific variation between

clones is not uncommon in M. persicae and even larger differences have

been observed if clones are collected from different host plant species.

Weber (1985a) found up to 8 fold variation in population size after 12

days in 1137 unreplicated isofemale lines, whereas Vorburger (2005)

found 60% variation in his measure of fitness using 19 clones. My

sampling of the Motte-Rimrock Reserve was not extensive; I began

with 22 females, representing 10 clones, all of which were feeding on

the host plant species used in the future experiments.

One surprising result from my experiments is that clone 813 in the

first experiment was one of the slower genotypes (Fig. 2.1). However,

in the second experiment it was the fastest growing clone. This could

be explained by a variety of causes such as the lack of proper temporal

sampling and replication in the first experiment and potentially human

error (I was still improving the counting technique). It is also possible

that given differences in temperature between experiments, this clone

could simply grow relatively more quickly at lower

temperatures. These results highlight that experiments need to be

self-contained i.e., having all controls needed for the analysis

concurrently being studied.

Given the variation I identified in the clones I assigned the three

focal clones a letter for easier reference. Clone 813 is henceforth clone

A, 815 is clone B, and 828 is clone C. Each of the clones was selected

because they differed in growth but also because their genotypes at

three of their microsatellite markers were unique. These markers had

PCR conditions that permitted multiplexing (using one PCR reaction to

amplify all three), alleles were very diverse, and thus one genetic

analysis provided information on 3 loci that could each identify the

clones in case some loci did not amplify in the sample (loci myz2, M40,

M86 in Table. 1.1). This streamlined genotyping in the large

experiments presented in chapters 2, 3, and 4.

CONCLUSIONS

I identified unique clonal lineages from a natural population,

characterized them genetically, identifying markers for easy

genotyping, and quantified how they differed in fitness. The

differences in intrinsic growth suggest that clonal frequency could

rapidly evolve in a mixed clone population. Because aphids undergo

multiple generations within a growing season, changes in frequency

should be measurable within the time course of short-term population

dynamics and could potentially have an impact on population growth

rate thereby linking evolution with concurrent population dynamics.

Table 2.1: Aphid Microsatellite Genotypes

The multilocus microsatellite genotypes of the 10 clonal M. persicae lineages

collected at the reserve. Numbers represent the length of each allele at each

locus. Unique clones have alleles of unique length or unique combinations of

alleles. The three bolded clonal lines will be used in future experiment. Their

alphabetical coding is listed for reference.

The bolded loci are used to quickly genotype aphids in those experiments.

The number of unique alleles and unique genotypes are listed for each locus.

Microsatellite Loci

Clonal Lineage

myz2 myz3 myz9 M40 M49 M86

813- Clone A 188/202 119 203/209 125 156/171 98/135

815 – Clone B 186/196 115/121 195/207 121/125 201/203 110/112

820 162/186 105/115 207/209 125/131 166 117/133

822 186/198 117/119 195/223 121/131138/156 123

825 186/198 117/ 119* 221/223 119/125152/166117/140*

828 – Clone C 186/188 117 195/209 121 154 135

831 186/188 117 195/209 119/121 154 117/135

834 174 117 195/223 121/133 138/ * 98/100

836 174/188 117 203 121/133136/156 110/123

207 /

M1 172/186 113/119 121/125 143/179

108/112 238*

# Alleles 8 6 7 5 11 9 or 10

# Genotypes 8 5 8 8 8 or 9 10

* alleles where microsatellite lengths were inconsistent due to bad amplification.

Figure 2.1: Aphid Clonal Variation in Intrinsic Growth Rate

Daily intrinsic growth rates, from experiment 1, of the eight clonal lineages

estimated during population exponential growth in pure (single clone)

populations.

Figure 2.2: Population Dynamics of Pure Clone Populations

Population dynamics of the clonal lineages in experiment 2. Values represent

mean number of aphids (±1 SE) and the lines are the best model fit more

predictions from the analysis. The three clones that are used in subsequent

experiment have their alphabetical code listed in the legend. Populations

crashed after day 30.

LITERRATURE CITED

Agashe, D. 2009. The stabilizing effect of intraspecific genetic

variation on population dynamics in novel and ancestral habitats.

The American Naturalist 174:255-67.

Agrawal, A. 2000. Host-range evolution: Adaptation and trade-offs in

fitness of mites on alternative hosts. Ecology 81:500-08.

Bassar, R. D., M. C. Marshall, A. López-Sepulcre, E. Zandona, S. K. Auer, et

al.

2010. Local adaptation in Trinidadian guppies alters ecosystem

processes. Proceedings of the National Academy of Sciences 107:3616-

21.

Birch, L. 1948. The intrinsic rate of natural increase of an insect

population. The Journal of Animal Ecology 17:15-26.

Blackman, R. 1974. Life-cycle variation of Myzus persicae (Sulz.)(Hom.,

Aphididae) in different parts of the world, in relation to genotype

and environment. Bulletin of Entomological Research 63:595-607.

Blackman, R. and V. Eastop. 2000. Aphids on the World’s Crops: an

Identification and Information Guide. 2nd edition. John Wiley &

Sons.

Bohannan, B. and R. Lenski. 2000. Linking genetic change to

community evolution: insights from studies of bacteria and

bacteriophage. Ecology Letters 3:362-77.

Davis, J., E. Radcliffe, and D. Ragsdale. 2007. Effects of high and

fluctuating temperatures on Myzus persicae (Hemiptera: Aphididae).

Physiological ecology 35:1461-66.

de Barro, P., T. Sherratt, C. Brookes, O. David, and N. Maclean. 1995.

Spatial and temporal genetic variation in British field populations of

the grain aphid Sitobion avenae (F.)(Hemiptera: Aphididae) studied

using RAPD-PCR. Proceedings of the Royal Society B: Biological

Sciences 262:321-27.

Dickson, R. and E. J. Laird. 1967. Fall dispersal of green peach aphids to

desert valleys. Annals of the Entomological Society of America 60:1088-

91.

Dixon, A. F. G., A. J. Minks, and P. Harrewijn. 1989. Evolution and

adaptive significance of cyclical parthenogenesis in aphids. P. 289-

97 Aphids: their Biology, Natural Enemies and Control.

Fisher, R. 1930. The Genetical Theory of Natural Selection. Clarendon

Press. Foster, S., R. Harrington, A. Dewar, I. Denholm, and A.

Devonshire. 2002.

Temporal and spatial dynamics of insecticide resistance in Myzus

persicae

(Hemiptera: Aphididae). Pest Management Science 58:895-907.

Fuller, S., P. Chavigny, L. Lapchin, and F. Vanlerberghe-Masutti. 1999.

Variation in clonal diversity in glasshouse infestations of the aphid,

Aphis gossypii Glover in southern France. Molecular Ecology 8:1867-

77.

Fussmann, G. F., S. P. Ellner, and N. G. J. Hairston. 2003. Evolution as a

critical component of plankton dynamics. Proceedings of the Royal

Society of London. Series B, Biological Sciences 270:1015-22.

Guldemond, J., W. J. Brink, and E. Belder. 1998. Methods of assessing

population increase in aphids and the effect of growth stage of the

host plant on population growth rates. Entomologia Experimentalis

et Applicata 86:163- 73.

Haag, C. R., J. W. Hottinger, M. Riek, and D. Ebert. 2002. Strong

inbreeding depression in a Daphnia metapopulation. Evolution

56:518-26.

Hanski, I. and I. Saccheri. 2006. Molecular-level variation affects

population growth in a butterfly metapopulation. PLoS Biology

4:e129.

Harmon, L. J., B. Matthews, S. Des Roches, J. M. Chase, J. B. Shurin, et al.

2009. Evolutionary diversification in stickleback affects ecosystem

functioning. Nature 458:1167-70.

Horovitz, A. and J. Galil. 1972. Gynodioecism in East mediterranean

Hirschfeldia incana Cruciferae. Botanica Gazette 133:127-31.

Hughes, A. R., B. D. Inouye, M. T. J. Johnson, N. Underwood, and M.

Vellend. 2008. Ecological consequences of genetic diversity.

Ecology Letters 11:609- 23.

Jokela, J., C. Lively, M. Dybdahl, and J. Fox. 2003. Genetic variation in

sexual and clonal lineages of a freshwater snail. Biological Journal of

the Linnean Society 79:165-81.

Karley, A., W. Parker, J. Pitchford, and A. Douglas. 2004. The mid-

season crash in aphid populations: why and how does it occur?

Ecological Entomology 29:383-88.

Karley, A. J., J. W. Pitchford, A. E. Douglas, W. E. Parker, and J. J. Howard.

2003. The causes and processes of the mid-summer population

crash of the potato aphids Macrosiphum euphorbiae and Myzus persicae

(Hemiptera: Aphididae). Bulletin of Entomological Research 93:425-

37.

Le Roux, V., J. Saguez, C. Vincent, and P. Giordanengo. 2004. Rapid

method to screen resistance of potato plants against Myzus persicae

(Homoptera : Aphididae) in the laboratory. Journal of Economic

Entomology 97:2079-82.

Lynch, M. 1984. The genetic structure of a cyclical parthenogen.

Evolution 38:186-203.

Mackauer, M. and M. J. Way. 1976. Myzus persicae Sulz. an aphid of world

importance. P. 51-119 in V. L. Delucchi, editor. Studies in biological

control. Cambridge University Press.

Minnich, R. and R. Dezzani. 1998. Historical decline of coastal sage

scrub in the Riverside-Perris Plain, California. Western Birds 29:366-

91.

Pan, J. and J. Price. 2001. Fitness and evolution in clonal plants: the

impact of clonal growth. Evolutionary Ecology 15:583-600.

Pimentel, D. 1968. Population regulation and genetic feedback.

Science 159:1432-37.

Pinheiro, J. and D. Bates. 2000. Mixed-Effects Models in S and

S-PLUS. Springer-Verlag.

Pinheiro, J. C., D. M. Bates, S. DebRoy, and D. Sarkar. 2009. nlme: Linear

and Nonlinear Mixed Effects Models.

Post, D. M., E. P. Palkovacs, E. G. Schielke, and S. I. Dodson. 2008.

Intraspecific variation in a predator affects community structure and

cascading trophic interactions. Ecology 89:2019-32.

R Development Core Team. 2009. R: A language and environment for

statistical computing. R Foundation for Statistical Computing,

Vienna, Austria.

Raybould, A., L. Maskell, M. Edwards, J. Cooper, and A. Gray. 1999. The

prevalence and spatial distribution of viruses in natural populations

of Brassica oleracea. New Phytologist 141:265-75.

Ro, T. and G. Long. 1999. GPA-Phenodynamics, a simulation model for

the population dynamics and phenology of green peach aphid in

potato: formulation, validation, and analysis. Ecological Modelling

119:197-209.

Roush, R. and J. McKenzie. 1987. Ecological genetics of insecticide

and acaricide resistance. Annual Review of Entomology 32:361-80.

Sloane, M. A., P. Sunnucks, A. C. C. Wilson, and D. F. Hales. 2001.

Microsatellite isolation, linkage group identification and determination

of recombination frequency in the peach-potato aphid, Myzus persicae

(Sulzer) (Hemiptera: Aphididae). Genetical Research 77:251-60.

Sunnucks, P., P. DeBarro, G. Lushai, N. Maclean, and D. Hales. 1997.

Genetic structure of an aphid studied using microsatellites: Cyclic

parthenogenesis, differentiated lineages and host specialization.

Molecular Ecology 6:1059-73.

Terhorst, C. P., T. E. Miller, and D. R. Levitan. 2010. Evolution of prey

in ecological time reduces the effect size of predators in

experimental microcosms. Ecology 91:629-36.

Tuda, M. and Y. Iwasa. 1998. Evolution of contest competition and its

effect on host–parasitoid dynamics. Evolutionary Ecology 12:855-

70.

Van Emden, H., V. Eastop, R. Hughes, and M. Way. 1969. The ecology of

Myzus persicae. Annual Review of Entomology 14:197-270.

Vehrs, S., G. Walker, and M. Parrela. 1992. Comparison of population

growth rate and whitin-plant distribution between Aphis gossypii

and Myzus persicae (Homoptera : Aphididae) reared on potted

chrysanthemums. Journal of Economic Entomology 85:799-807.

Via, S. and A. J. Shaw. 1996. Short-term evolution in the size and

shape of pea aphids. Evolution 50:163-73.

Vorburger, C. 2005. Positive genetic correlations among major life-

history traits related to ecological success in the aphid Myzus

persicae. Evolution 59:1006- 15.

Vorburger, C. 2006. Temporal dynamics of genotypic diversity reveal

strong clonal selection in the aphid Myzus persicae. Journal of

Evolutionary Biology 19:97-107.

Vorburger, C., M. Lancaster, and P. Sunnucks. 2003.

Environmentally related patterns of reproductive modes in the

aphid Myzus persicae and the predominance of two 'superclones'

in Victoria, Australia. Molecular Ecology 12:3493-504.

Vrijenhoek, R. and E. Pfeiler. 1997. Differential survival of sexual and

asexual

Poeciliopsis during environmental stress.

Wallner, W. 1987. FACTORS AFFECTING INSECT POPULATION-DYNAMICS:

Differences between outbreak and non-outbreak species. Annual

Review of Entomology 32:317-40.

Weber, G. 1985a. Genetic variability in host plant adaptation of the

green peach aphid, Myzus persicae. Entomologia Experimentalis et

Applicata 38:49-56.

Weber, G. 1985b. Population genetics of insecticide resistance in the

green peach aphid, Myzus persicae (Sulz.) (Homoptera, Aphididae).

Zeitschrift für Angewandte Entomologie 99:408-21.

Whitham, T. G., J. K. Bailey, J. A. Schweitzer, S. M. Shuster, R. K. Bangert,

et al.

2006. A framework for community and ecosystem genetics: from genes

to ecosystems. Nature Reviews Genetics 7:510-23.

Wilson, A., P. Sunnucks, and D. Hales. 2003. Heritable genetic

variation and potential for adaptive evolution in asexual aphids

(Aphidoidea). Biological Journal of the Linnean Society 79:115-35.

Wilson, A. C. C., B. Massonnet, J. C. Simon, N. Prunier-Leterme, L.

Dolatti, et al. 2004. Cross-species amplification of microsatellite loci

in aphids: assessment and application. Molecular Ecology Notes

4:104-09.

Wilson, D. and W. Swenson. 2003. Community genetics and

community selection. Ecology 84:586-88.

Wyatt, I. and P. White. 1977. Simple estimation of intrinsic increase

rates for aphids and tetranychid mites. The Journal of Applied

Ecology 14:757-66.

Yoshida, T., L. E. Jones, S. P. Ellner, G. F. Fussmann, and N. G. J.

Hairston. 2003. Rapid evolution drives ecological dynamics in a

predator-prey system. Nature 424:303-06.

CHAPTER THREE: EXPERIMENTAL

ASSESSMENT OF THE IMPACT OF RAPID

EVOLUTION ON POPULATION DYNAMICS IN

THE GREENHOUSE

ABSTRACT

Most short-term population dynamic models and studies assume

that evolution occurs on slower timescales and thus do not allow for

parameter values to evolve during their study. Yet multiple recent

examples of rapid evolution in many systems could invalidate this

approach if evolution impacts short-term ecological dynamics. The

objective of this study is to quantify the impact of rapid evolution on

short-term population dynamics using an aphid (Myzus persicae) and an

undomesticated host (Hirschfeldia incana). This is the first experimental

Eco-Evolutionary Dynamics study system using an insect-herbivore to

my knowledge. I manipulated the amount of genetic variation in

intrinsic growth rate within replicated aphid populations by altering the

clonal composition. Aphid populations evolved rapidly changing

significantly from their initial frequency within four weeks, well within a

growing season and approximately four or five aphid generations. As

populations were evolving I quantified their population dynamics.

Evolving populations grew significantly faster, between 28% and 34%,

and reached higher densities, compared to non-evolving control

populations. Evolving aphid populations did not cause increased

damage to their host plant. My results countermand to prevailing

approach that assumes that short-term population dynamics are too

fast for evolution to have an influence.

INTRODUCTION

Few studies or models concerning short-term population

dynamics consider the possibility that evolution could alter population

parameters (e.g. intrinsic growth rates, carrying capacity, interspecific

interactions) during the study period (Thompson 1998, Levins et al.

2003). Although studies take into account changes in population

parameters, due to changes in age structure or spatial distribution,

they rarely consider genetic changes (Cappuccino and Price 1995,

Sibly and Hone 2002) even though certain ecologists have advocated

such a consideration for decades (Pimentel 1961, Chitty 1967,

Anderson and King 1970, Berry et al. 1978). This view, however, seems

to be changing (Thompson 1998, Hairston et al. 2005, Saccheri and

Hanski 2006, Hughes et al. 2008, Pelletier et al. 2009).

Life tables are commonly used to estimate intrinsic growth rate,

a key parameter in models forecasting population dynamics (Kocourek

et al. 1994, Guldemond et al. 1998, Ro and Long 1999). Alternatively,

models are fit to historical density data to forecast future population

density (Onstad et al. 2005). These approaches however assume that

genotypic frequencies and hence

parameter values do not evolve. The justification for ignoring

evolution in short- term ecological studies is based on two

assumptions. Firstly, populations are often assumed to be genetically

homogeneous, at least in traits that impact population dynamics

(Roughgarden 1979, Cappuccino and Price 1995). Yet, population

genetic variation in ecologically important traits has been

demonstrated repeatedly since the 1960s reviewed in (Ayala 1968,

Berry et al. 1978). Moreover, such variation can significantly influence

population dynamics in the laboratory (Schlager 1963, Leips et al.

2000) and in nature (Hazell et al. 2006, Saccheri and Hanski 2006).

Secondly, evolution is assumed to occur on a much slower (and thus

separate) timescale, over hundreds or thousands of generations, than

short-term ecological processes, occurring over a dozen or fewer

generations (Slobodkin 1980, Thompson 1998, Hairston et al. 2005).

Thus even if genetic variation is considered, evolution is ignored

because it is perceived to be too slow to have an ecological effect

(Endler 1991). Studies falsifying this assumption by identifying ‘rapid

evolution’ occurring on ‘ecological time’, often within a few

generations, in both natural and human disturbed environments, have

recently become very common (see reviews in Dyer 1968, Thompson

1998, Hendry and Kinnison 1999, Bone and Farres 2001, Reznick and

Ghalambor 2001, Ashley et al. 2003) but what remains unclear is

whether such rapid evolution actually impacts short-term ecological

dynamics (Pelletier et al. 2009).

Theoretical models have shown that rapid evolution in

population parameters and genotypic frequencies can significantly

alter population growth trajectories (Pimentel 1961, Anderson and

King 1970, Fussmann et al. 2003, Duffy and Sivars-Becker 2007).

These effects could be straightforward. For example, if a genotype

with a higher growth rate becomes more common it could accelerate

the whole population’s growth rate. On the other hand, more complex

effects are possible. If this common genotype’s relative fitness

advantage decreases with increasing density and frequency it might

slow the growth of the evolving population (Agrawal 2004). Because of

such potentially non-intuitive interactions it is important to

experimentally assess the impact of rapid evolution.

Whether the effect of evolution is simple or complex does not

imply that evolution should be considered in all population dynamic

studies. It is important to quantify how strongly evolution can impact

concurrent population dynamics in different contexts (Hairston et al.

2005). The strength of this effect might depend on the ecological

context (e.g. community composition, level of disturbance) and

evolutionary context (e.g. amount of genetic variation present, rate of

evolution, mechanism of evolution). For example, Yoshida et al. (2003)

showed that rapid evolution within their rotifer-algal chemostat system

causes the predator-prey cycles to become almost perfectly out-of-

phase and not one quarter out-of-phase as predicted by ecological

theory. Yet this predator-prey cycling in chemostats is highly

dependent on nutrient flow and will not occur if the dilution rate is too

low

or too high (Shertzer et al. 2002). This implies that empirical studies

under differ conditions are required to quantify the importance of this

process.

A small but growing number of experimental studies have begun

quantifying the impact of rapid evolution on population dynamics

using different methods (Fussmann et al. 2007). Bohannan and Lenski

(2000) observed changes in mean density, and the fluctuation in

density in both bacteria and phage, as the bacteria evolved resistance

to this phage. Fussmann et al. (2003) used a combination of modeling

and an experimental verification to show that rapid evolution of

asexual reproduction qualitatively changed populations dynamics

within 30 days. Without evolution the populations has a single peak in

density and then crashes, however, with evolution a second peak in

density occurs. Other studies experimentally compare the population

dynamics of evolving populations to those that cannot evolve due to

replacement of the population with unselected individuals (Pimentel et

al. 1963, Pimentel and Al- Hafidh 1965, Pimentel 1968), a lack of

genetic variation (Yoshida et al. 2003, Fussmann et al. 2007, Agashe

2009), or because the non-evolving population is in an environment

without the key selective pressure such as predation (Agrawal 2000,

Terhorst et al. 2010). These studies demonstrate how evolutionary

dynamics influence the dynamics and outcome of short-term

ecological phenomena and argue convincingly for causality.

Plant-herbivore interactions are thought to be one of the most

common and important ecological interactions in natural populations,

generating much of

the species and phenotypic diversity in nature as well as having

immense economic importance (Ehrlich and Raven 1964). Surprisingly,

there have been no experimental studies in which evolution was

manipulated and its impact on population dynamics quantified in

plant-herbivore interactions (except spider mites in Agrawal 2000). Yet

a growing body of research in eco-evolutionary dynamics is focused on

how plant genetics impacts insect communities (‘community genetics’

reviewed in Whitham et al. 2006, Hughes et al. 2008). This is striking

given that Wallner (1987), in his highly cited review of the causes of

insect pest outbreak, strongly advocated for a consideration of the role

of evolution in such outbreaks. The ecological effects of rapid evolution

in a plant- herbivore system could differ greatly from those observed

in predator-prey or host-parasitoid interactions for two reasons. First,

the patterns of population dynamics often differ. In many plant-

herbivore systems, especially in agricultural pests, the pest dynamics

are often in a non-equilibrium state consisting of outbreaks and

crashes (reviewed in Wallner 1987, Karley et al. 2004) as opposed to

equilibrium or predator-prey cycles observed in the rapid evolution

studies listed previously. Second, herbivores might have a weaker

affect on host population dynamics than would a predator on predator-

prey dynamics because herbivores do not necessarily kill their host.

The magnitude of impact of insect herbivores on plant population

dynamics has been debated for years (Crawley 1989) and only in the

last decade have a dozen or so studies found support for this process

(Maron and Crone 2006).

My goal was thus to develop a model plant-herbivore system to

study how rapid evolution through natural selection acting on genetic

variation present within natural populations impacts concurrent

population dynamics. To do so, I developed a study system wherein

evolution can easily be manipulated. I selected the green peach aphid

(Myzus persicae) and a local wild invasive annual mustard, Hirschfeldia

incana, as a host. In this aphid species populations are replete with

clonal variation at the beginning of the season, resulting from

immigration and sexual reproduction, which permits rapid evolution at

the beginning of the growing season (Vorburger 2006). Variation

declines throughout the growing season leading to the evolutionary

changes in population mean trait values (Via and Shaw 1996). To

study the impact of such rapid evolution on population dynamics I

experimentally manipulated aphid populations’ genetic composition

and evolutionary potential by controlling which clones were present in

replicated populations. By selecting different pairings of clones I

altered the level of genetic variation in an ecologically important trait

and could thus test how the evolutionary context might alter the

impact of rapid evolution on concurrent population dynamics. I

specifically tested the following hypotheses-predictions: 1) If rapid

evolution impacts population dynamics, then the observed population dynamics

of evolving aphid populations will differ significantly from those of non- evolving

aphids, 2) If evolutionary context is important, then the impact of rapid evolution

on population dynamics will differ between the different evolution treatments,

and 3) If rapid evolution in aphids impacts their host plant, then plant

fitness will differ significantly when exposed to evolving versus non-evolving

aphids.

MATERIALS AND METHODS

I- Experimental Design

Evolving and non-evolving M. persicae populations on H. incana

plants were studied in a partly cooled greenhouse (mean daytime

temperature = 31°C, range = 16°C to 47°C, mean nighttime = 19°C,

range = 13°C to 30°C). Asexual reproduction was maintained by using

addition lighting providing 16hrs light/ 8hrs dark (Blackman 1974). To

minimize variation in this primarily outcrossing plant, plants used in

the experiment were grown from the seeds of a single H. incana plant

collected in 2008 at the Motte-Rimrock Reserve. These seeds were

planted in four liter pots, using UCR soil mix III, a sand/peat moss mix

supplemented with micronutrients, and watered every three days.

Cages were constructed to individually house each plant. Cages

consisted of an 8 liter pot with a wire frame creating a 75cm high

dome that held up thin transparent mesh (Bridal Organza, #664-7242,

Jo-Ann). On day 0 of the experiment, the six week old seedlings in the

rosette stage, approximately 10-15cm wide, were inoculated with

seven different aphid treatments by placing 20 third instar aphids onto

each plant.

These aphids came from stock greenhouse clonal populations,

regularly tested for contamination. Replicates were initiated on three

consecutive days starting October 1st 2009. On day 1 missing aphids

were replaced with fourth instar

aphids. Each treatment combination was assigned in a randomized

block design and replicated 10 times.

I used three aphid clonal lineages, identified as ‘A’, ‘B’, and ‘C’,

that were collected in 2008 from the University of California Motte-

Rimrock Reserve near Perris, California (see Chapter 1). These clones

differ in microsatellite markers (Appendix 3.1). A preliminary

greenhouse experiment revealed that they also differ in exponential

growth rates when grown in pure cultures (Chapter 2, daily growth

rate of clone A = 0.268 ± 0.002, B = 0.263 ± 0.004, and C = 0.251 ±

0.004). I used these three clones to establish seven aphid treatments.

Three evolution treatments consisted of aphid populations (on a

single plant) that have two different clones. I established evolution

treatments that consisted of all three two-way combinations (A-B, B-C,

and A-C). Based on the ranking of growth rate in the preliminary

experiments (Chapter 2), I initiated these evolving populations with 5

individuals of the faster growing clones and 15 individuals of the

slower growing clones. The A clone in the A-B and A-C evolution

treatments as well as the B clone in the B-C evolution treatment

represented 25% of the initial clonal frequency. These populations

have genetic variation in fitness (e.g. A clone’s rm is greater than that

of the C clone) and thus could evolve by changing in clonal frequency

(away from the 25% : 75% initial ratio). Three different non-evolution

(pure clone) treatments received 20 aphids of only one of the three

clones.

Because all individuals were of the same genotype within these pure

populations,

gene frequencies could not change, thus preventing evolution. Finally,

a ‘no- aphid control’ treatment did not receive any aphids.

II- Rates of Evolution

On day 28 I collected 100 aphids from every population to track

changes in clonal frequencies (evolution). Between 16 and 24 aphids

from each sample were genotyped (for a total of 497 aphids) at three

microsatellite loci using a multiplex approach (see Appendix 3.1 for

detailed genetic methods). I calculated the frequency of the faster

clone for each treatment and replicate separately. For each evolution

treatment I then determined whether the mean frequency of the

faster clone differed significantly from the initial clonal frequency of

25% using one-sample t-tests.

III- Aphid Population Dynamics

Aphid population dynamics were quantified by counting all

aphids on days 3, 7, 10, 14, 17, 21, 24, 28, and 33. When populations

rose above 2000 aphids per plant I sub-sampled by counting one half

of every leaf. Plant senescence caused the aphid populations to crash

after day 28 so I excluded the census taken on day 33. I fit an

exponential population growth model and I found, by looking at the

residuals of every treatment, that exponential growth lasted until day

14. Afterwards populations grew linearly, as determined by fitting a

separate linear model on this portion of the time-series, days 14 to 28.

The exponential growth phase, days 0-14, was analyzed with a

linear mixed-effect model (LME) where the dependent variable was

LN(x) transformed number of aphids, the fixed effect was aphid

treatments and day (as the main covariate). Because the repeated

aphid counts on the same plant violated the assumption of

independent observations, I set unique plant identity as a random

effect on population growth rate and intercept and used an

autoregressive correlation error structure (Pinheiro and Bates 2000). I

modeled increasing variance through time by using a variance function

within the LME that increases with the power of the variance covariate

(varPower). Block (day of initiation of the replicate and spatial position

in the greenhouse) and initial plant size did not improve model fit and

were not included in the final model. For the linear growth phase, days

14-28, the same LME model was applied except that the number of

aphids was not LN(x) transformed. All analyses were implement in R

(v. 2.11.1; R Development Core Team 2009) using the nlme package

(Pinheiro et al. 2009).

IV- The Impact of Evolution on Population Dynamics

My objective is not to predict which clone will out-compete the

other but to statistically test the impact of changes in clonal frequency

on concurrent population dynamics. Ideally, one would compare the

observed aphid population dynamics in the evolution treatment to

those of a non-evolving mixed population containing the same two

clones that remain at a frequency of 25% : 75%. This is impossible

since clones will change in frequency because of fitness differences. I

thus generated the expected population growth parameters of such a

non- evolving population by using the pure aphid treatments. I tested

three a priori null hypotheses that the population growth rate (slope)

and density (intercept) do not differ between each evolution treatment

and their corresponding pure treatments,

e.g. A-C evolution treatment vs pure A and pure C treatments. I did so

with the use of planned contrasts that are orthogonal comparisons

between a subset of the aphid treatment levels within the LME

analysis. Different hypotheses are tested by assigning weights to

treatments levels. I set the planned contrast coefficients of the no-

evolution expectation to match those of the initial clonal frequency

(e.g. pure A= -0.25 and pure C= -0.75 and these are compared to the

AC evolution treatment= 1). My three hypotheses were tested using

the following simplified contrast matrix:

Aphid

Treatments

Hypothesis 1:

AB vs Pure A and

Pure B

Hypothesis 2:

BC vs Pure B and

Pure C

Hypothesis 3:

AC vs Pure A and

Pure C

Pure clone A -0.25 0 -0.25

Pure clone B -0.75 -0.25 0

Pure clone C 0 -0.75 -0.75

Evolution AB +1 0 0

Evolution BC 0 +1 0

Evolution AC 0 0 +1

Thus differences in growth rate or density between the evolution

treatment and the no-evolution expectation represent the impact of

changes in the frequency of clones (rapid evolution) on population

dynamics.

V - Host Plant Fitness

Finally, to quantify the impact of aphid rapid evolution on its

host’s fitness I measured the above ground dry biomass of the plants

at the end of the experiment as a proxy for host fitness (Mitchell-Olds

and Bradley 1996). I fit a general linear model on LN(x) transformed

plant weight measurements. The factors were aphid treatment, block

and initial plant size (width of rosette on day 3). The interactions

between these factors were non-significant and thus removed from the

final model. I again used planned contrasts to determine whether

plants with evolving aphid populations were smaller than expected

from no-evolution treatments.

RESULTS

I - Pure Clone Treatments

Pure clone treatments differed in their population dynamics as

the rank order of growth rates changed throughout the experiment

(Fig. 3.1). In the exponential phase (day 0- to 14) the B clone grew

fastest (5.6% faster than A), the A clone was second fastest (4.8%

faster than the C clone; Fig. 3.1.a). In the linear growth phase (day 14

to 28) the B clone grew significantly slower than the other two clones.

The A clone grew 80% and the C clone grew 37% faster than the B

clone respectively and the A clone grew 12% faster than clone C (Fig.

3.1.b).

II - Evolution Treatments - Genetic Analyses

On day 28 I tested for changes in the frequency of clones in the

evolution treatments away from the initial frequency of 25% : 75%. In

the A-B evolution treatment clonal frequency did not change

(frequency of A clone = 23%, one sample t-test, p = 0.63; Fig. 3.2).

The B-C and A-C evolution treatments did significantly evolve as the

frequency of the C clone decreased. The B clone reached 44%

(p=0.031) and the A clone reached 47% (p=0.001) almost doubling

their initial frequency of 25% (Fig. 3.2).

III - Impact of Aphid Evolution on Aphid Population Dynamics

To test the impact of rapid evolution on concurrent population

dynamics I compared the observed population dynamics in evolving

populations to those observed in both corresponding pure treatments

by using planned contrasts proportional to the initial frequency of

clones (i.e., population dynamics without evolution). For example in

the exponential growth phase the A-C evolution treatment grew with

an exponential rate of 0.321 which is 8.5% slower than the expected

growth rate of a population at a constant (non-evolving) frequency of

25% (for clone A with a growth rate of 0.0363) and 75% (for clone C

with a growth rate of 0.347) which has an expected growth rate of

0.351.

In the exponential phase (days 0 to 14), the only evolution

treatment that differed in daily growth rate from its no-evolution

expectation was the A-C treatment (Fig. 3.3 a-b-e, Table 3.1). Oddly,

although the faster growing A clone

increased in frequency in this treatment, the evolution treatment grew

8.5% slower (LME, p = 0.0015) than the no-evolution expectation (Fig.

3.3.e). In the second growth phase, days 14 to 28, the A-B evolution

treatment did not differ from the no-evolution expectation in either

intercept (density on day 14) or growth rate (Fig. 3.3.b, Table 3.1).

Evolution in the B-C treatment did not alter density on day 14 but

significantly accelerated population growth rate compared to the no-

evolution expectation afterwards (+28.2%, p = 0.008, Fig. 3.3.d, Table

3.1).

Finally, although the A-C evolution treatment grew slower in the

exponential stage leading to a significant decrease in density at day

14 (-28.5%, p= 0.009), evolution significantly accelerated population

growth rate in the second growth phase (+33.8%, p < 0.001, Fig. 3.3.f,

Table 3.1).

IV - Impact of Aphid Evolution on Host Plant Fitness

Aphid feeding severely reduced plant size. The no-aphid control

plants were five times heavier than plants with aphids (p < 0.001).

Although in certain treatments evolution led to higher aphid densities,

this did not magnify the impact of aphids on the host plants’ above

ground biomass. Final plant weight did not differ significantly between

the three evolution treatments and their corresponding no-evolution

expectations (ANOVA, all p-values > 0.1).

DISCUSSION

I experimentally assessed the impact of rapid evolution on

concurrent ecological dynamics in a plant-herbivore system. I found

that rapid evolution, occurring within weeks, significantly accelerated

population growth rates and density as clonal frequencies changed.

Yet, this rapid evolution of aphids did not have a detectable effect on

the plant host. The presence of aphids had a large and significant

impact on plant growth, but faster growing evolving aphids did not

damage their host more than non-evolving aphids. These results have

important implications for the study of population dynamics and pest

management.

Over the course of only 28 days, approximately 4-5 aphid

generations, natural selection significantly altered aphid clonal

frequencies in two of the three evolution treatments. Similar changes

have been observed in non-experimental aphid infections in

greenhouses (Fuller et al. 1999) as well as in the wild populations (de

Barro et al. 1995, Vorburger 2006). In the A-C evolution treatment,

the A clone almost doubled its initial frequency, which is expected

given that the A clone grew faster than the C clone in pure

treatments in both growth phases (Fig 3.1). Yet, the evolutionary

outcome was not always predictable from the difference in the growth

rates of single clone cultures. The A-B treatment did not evolve (Fig.

3.2), even though on day 28 the pure A clone was 41% more dense

than the pure B clone (Fig. 3.1). This might be explained by the fact

that the pure B clone grew faster than the A during the first growth

phase but slower during the second growth phase, reversing the

evolution that

occurred during the first phase. This reversal may in turn suggest that

the fitness difference between the clones is density dependent. Also,

the B clone in the B-C treatment reached a frequency of 44% on day

28, even though its density on that day was 30% lower than the pure C

clone. This is surprising since in the early phase pure B grew faster

than C but the opposite occurs in the later growth phase. One possible

explanation for these two unpredictable results is that one clone

reduces the others’ growth (Rochat et al. 1999). In my experiment it is

possible that the B clone interferences with the C and A clones’

feeding, thus lowering their relative fitness when mixed, resulting in

higher clone B frequencies than expected on day 28. Also, it is possible

that the early growth phase determines the evolutionary outcome

since populations expand approximately 170 fold during the early

phase, as opposed to only 4.5 fold in the later phase.

The novelty of my study is not to determine the exact process

leading to these evolutionary changes but to assess the impact of

these changes on concurrent population dynamics. I observed strong

impacts of rapid evolution on population dynamics in both treatments

that evolved. Rapid evolution accelerated population growth rate by

28% and 33% in the two mixed treatments where evolution occurred,

supporting my first hypothesis (Fig. 3.3, Table 3.1). Evolution increased

population density by as much as 17% for the B-C and 19% for the A- C

treatments compared to no-evolution expectations (best linear

approximations on day 24 where maximum differences are seen in the

raw data). These effects are similar in magnitude to other ecological

forces usually deemed as important.

For example increasing temperature from 20-25ºC and 25-30ºC

causes M. persicae’s intrinsic growth rate to increase by 14% and 3.6%

respectively (Davis et al. 2007). Gurevitch et al (2000)’s factorial

meta-analysis of dozens of experiments manipulating predation and

competition revealed impact sizes similar to those reported here. The

strong effects I observed suggest that population density and growth

rate might not be predictable by simply averaging the demographic

parameters of a mixed genotype population (Wallner 1987, Endler

1991).

Rapid evolution could have even stronger effects if I had used

clones with larger fitness differences. Such variation would be likely if I

had 1) sampled more than a dozen clones from a natural population,

2) collected aphids from different host species, or 3) conducted the

experiment in an agricultural setting with pesticide application that

would select for M. persicae clones that differ in resistance. Other

studies in M. persicae have reported higher levels of genetic variation

between clones. Weber (1985a) found up to 8 fold variation in

population size after 12 days in 1137 unreplicated isofemale lines,

whereas Vorburger (2005) found 60% variation in his measure of

fitness using 19 clones. In another study, Weber (1985b) found 3000

fold variation in resistance to parathion. Also, the acceleration in

growth rate due to evolution that I observed only occurred in the latter

half of the experiment (Table 3.1). This delay is likely a consequence

of a lag of at least a few generations before evolution can change the

population’s growth parameters. The fact that such a change occurred

within

only 30 days directly challenges the assumption that evolutionary

change happens on a much longer time scale than the change

wrought by ecological interactions. These results suggest that if the

experiment had lasted longer, either by having a longer-lived host or

by having new host plants, the effect of evolution could have been

magnified. The large acceleration of growth rate within only a few

weeks suggests that rapid evolution on naturally occurring genetic

variation can be a strong driver of population dynamics on ‘ecological

timescales’.

It is difficult to compare the effect of rapid evolution on

population dynamics between very different study systems. Rapid

evolution in certain predator-prey or host-parasite systems has been

shown to alter the mean density and the pattern and magnitude of

density cycles (Pimentel 1968, Bohannan and Lenski 2000, Fussmann

et al. 2003, Yoshida et al. 2003, Terhorst et al. 2010).

These systems might inherently have more opportunity for qualitative

changes in population dynamics since both species undergo multiple

generations within the experiment. Thus the population dynamics and

potentially the evolutionary dynamics of both species might be

altered. This is not what occurs in many plant- herbivore systems.

Many insect populations grow rapidly then crash because of plant

senescence, predation, parasitism or climate (Wallner 1987, Ro and

Long 1999, Karley et al. 2004). Such dynamics often occur within one

generation of the plant. Given differences in these types of

interspecific interactions in nature it

is important to investigate the impact of rapid evolution in different

model systems.

My second hypothesis tests whether the evolutionary context

(identity of genotypes and rate of evolution) influences the impact of

rapid evolution on population dynamics. Given that one evolution

treatment did not evolve and the other two did so at similar rates

nullifies my ability to quantify the relationship between rate of

evolution and impact size since I have dichotomous treatments (rapid

and no evolution). These results suggest that, in these experimental

conditions, genotypic identity of clones, as long as they evolve at

similar rates, does not change the magnitude of the impact of rapid

evolution on population growth rate. Yet a closer examination of my

results suggests that clonal identity might have an important impact.

In the A-C evolution treatment it is clear why the growth rate

accelerates; the faster growing A clone becomes more common. For

the B-C treatment however, the B clone increases in frequency but the

B clone grows more slowly in the second growth phase (by 37%) than

the C clone did in pure treatments (Fig. 3.1). Why exactly this

evolutionary change accelerates the growth of the evolving population

remains unresolved. One possibility is that the B clone experiences

more severe density-dependent growth when at high frequency which

only occurs in the pure B treatment. Reduced growth rate does not

occur in the B-C evolution treatment because the B clone only reaches

44% frequency. Experiments at different initial clone frequencies could

help resolve this issue. Such intraspecific variation in the strength of

density dependence has

been document in other aphid species (Agrawal et al. 2004) and is

obvious from the pure clone treatments (Fig. 3.1). I have also

observed that the B aphids become smaller in body size and are more

likely to produced winged aphids but only at very high densities and

only when it represents a dominant frequency of the population (M. M.

Turcotte pers. observation in multiple experiments). Aphid clones are

known to differ in their propensity to create winged individuals during

crowding (Muller et al. 2001, Hazell et al. 2005).

An obstacle I faced was how to statistically compare evolving

and non- evolving populations. If one compares an evolving mixed

population (e.g. clones A and B) to either pure treatment (pure A or

pure B) then genetic variation, clonal identity, and evolution are

confounded. My aim was to assess the impact of evolution itself and

not the former factors. Hence I created planned contrasts that

compare the observed population parameters in the evolution

treatment to those of both corresponding pure treatments in the ratio

of the initial frequency of clones (25:75). This is akin to having a

population composed of a constant (non- evolving) ratio of clones. The

limitation of this approach is that it assumes that interclonal

interactions are equivalent to intraclonal interactions. This caveat

might explain why the A-C evolution treatment is initially grows 8.4%

slower (Fig. 3.3.E, Table 3.1). This result suggests that one clone,

probably A, is interfering with the other. This decreases total

population growth rate and only once the faster growing A clone has

increased significantly in frequency, days 14-28, can evolution

compensate for this effect and accelerate population growth rate.

Alternatively, interclonal interference might be density- and/or

frequency- dependent and thus change throughout the experiment.

Clones with aphid populations are known to vary genetically in

competitive ability and this variation can also change with ecological

context (Hazell et al. 2006).

This study quantified one half of the eco-evolutionary feedback

cycle and showed that rapid evolution can significantly alter

population density. Whether this occurs under different ecological

conditions remains to be tested. Another opened question is whether

changes density reciprocally influences future bouts of evolution. If

this is the case it would complete the eco-evolutionary feedback cycle

(Fussmann et al. 2007, Kokko and Lopez-Sepulcre 2007). Although

selection within aphids is known to be density-dependent (Agrawal et

al. 2004) direct tests of this hypothesis in this system should be

undertaken. Accurate predictions of population dynamics are crucial

in many applied fields, such as fisheries, pest management,

conservation biology, invasion biology, and epidemiology. Rapid

evolution, in many ecologically relevant traits has been documented

repeatedly in these systems (Ashley et al. 2003) yet evolution is

usually not considered in population dynamic studies. My

experimental results strongly countermands this approach and

suggest that rapid evolution can have a large effect on growth rate

and density. It follows that investigating the impact of rapid evolution

can improve predictions and population management (Hufbauer and

Roderick 2005, Duffy and Sivars-Becker 2007).

Table 3.1: Analysis of Population Dynamics Comparing Evolving and Non-

Evolving Populations

Planned contrasts from a linear mixed-effect model, comparing each type of

evolving population to its corresponding no-evolution expectation, generated

from the pure aphid treatments following the initial frequency of clones (see

Methods for details). The percent change represents the change in slope or

intercept from the non-evolving expectation to that of the observed evolution

treatment. Thus positive changes represent increases due to evolution. Slope

represents the rate of growth of aphid populations and intercept represents

density at the start of each time period. All p-values are for 2-tailed tests.

Significant results were bolded for easier identification.

Evolutio

Treatme

Days 0-14 Days 14-28

(Clones) d.f. t p % Change d.f. t p % Change

Intercept

A-B

50 -0.51 0.613 -2.6 50 0.16 0.875 +1.7

Slope 216 -0.38 0.703 -0.9 215 -0.77 0.440 -9.4

Intercept

B-C

50 -0.77 0.448 -3.7 50 -1.33 0.187 -14.4

Slope 216 -0.89 0.375 -2.2 215 2.67 0.008 +28.2

Intercept 50 -0.35 0.723 -1.8 50 -2.77 0.009 -28.5

A-C

Slope 216 -3.22 0.002 -8.5 215 3.41 0.001 +33.8

Figure 3.1: Population Dynamics of Pure Clone Treatments

Population dynamics of pure clonal treatments. Values represent mean

number of aphids (±1 SE) through time separated into two time periods for

easier visualization. a) early growth phase during days 0 to 14 and b) the late

growth phase during days 14 to

28. The y-axes differ between panels.

Figure 3.2: Rapid Experimental Evolution of Clonal Frequencies

Rapid clonal evolution as shown by the mean frequency of the faster

growing aphid clones in each evolution treatment (±1 SE). X-axis shows

which clone’s frequency is being tested in each evolution treatment. Dashed

horizontal bar indicates initial clonal frequency of 25% and (*) indicate

significant divergence from initial frequency. A, B, and C are the aphid clones.

Figure 3.3: Populations Dynamics of Evolving and Non-Evolving Aphids

Population dynamics of the three observed evolution treatments (black

diamonds) with the best fit model from LME analysis (black line). Left panels

are for the early growth phase and panels on the right for the late growth

phase for easy visualization. The dashed gray line represents the best fit

model that combines both pure treatments using the constant (non-evolving)

frequency of clones (25:75). For each treatment I added the corresponding

pure clone treatments (grey symbols) used to generate the no-evolution

expectation. Values represent mean number of aphids (±1 SE) and the y-axes

differ between left and right panels. Evolution treatments have two letters.

Figure 3.3

APPENDIX 3.1: DETAILS OF MULTIPLEX GENETIC ANALYSIS

Methods for Multiplex Genotyping used during Experiment

During the experiment I genotyped aphids using three microsatellite loci

for which the three clones had unique genotypes (Appendix Table 3.1). This

reduced cost and workload while providing robust genotyping. A single

multiplex PCR reaction amplified three microsatellite loci at once (myz2, M40,

and M86) using the PMS1 program. The PCR reactions were identical to those

used to genetically characterize all clones in Chapter 1 except for the use of

a mixed primer solution that contained three forward and three reverse-

labeled primers. To normalize signal strength the concentrations of the primer

pairs were: 1, 3, and 5µM for loci myz2, M40, and M86 respectively.

Appendix Table 3.1: Microsatellite Genotype of Focal Aphid Clones

Microsatellite genotypes of the three clonal lineages of green peach

aphid (Myzus persicae) used in this experiment. These were collected from the

Motte-Rimrock Reserve in spring 2008 from the mustard plant Hirschfeldia

incana. Numbers represent the length of the each allele at each locus. The

three bold loci were used to identify clones during the experiment.

Microsatellite Loci

Clonal Lineage

myz2 myz3 myz9 M40 M49 M86

A186 202 119 119 203 209 125 125 156 171 98 135

B186 196 115 121 195 207 121 125 201 203 110 112

C186 188 117 117 195 209 121 121 154 154 135 135

LITERATURE CITED

Agashe, D. 2009. The stabilizing effect of intraspecific genetic

variation on population dynamics in novel and ancestral habitats.

The American Naturalist 174:255-67.

Agrawal, A. 2000. Host-range evolution: Adaptation and trade-offs in

fitness of mites on alternative hosts. Ecology 81:500-08.

Agrawal, A. 2004. Plant defense and density dependence in the

population growth of herbivores. American Naturalist 164:113-20.

Agrawal, A. A., N. Underwood, and J. R. Stinchcombe. 2004.

Intraspecific variation in the strength of density dependence in

aphid populations.

Ecological Entomology 29:521-26.

Anderson, W. W. and C. E. King. 1970. Age-specific selection.

Proceedings of the National Academy of Sciences 66:780-86.

Ashley, M. V., M. F. Willson, O. R. W. Pergams, D. J. O'Dowd, S. M.

Gende, et al. 2003. Evolutionarily enlightened management.

Biological Conservation 111:115-23.

Ayala, F. 1968. Genotype, environment, and population

numbers. Science 162:1453-9.

Berry, R., R. Anderson, B. Turner, and L. Taylor. 1978. Genetical

factors in animal population dynamics. P. 53-80 Population

Dynamics.

Blackman, R. 1974. Life-cycle variation of Myzus persicae (Sulz.)(Hom.,

Aphididae) in different parts of the world, in relation to genotype

and environment. Bulletin of Entomological Research 63:595-607.

Bohannan, B. and R. Lenski. 2000. Linking genetic change to

community evolution: insights from studies of bacteria and

bacteriophage. Ecology Letters 3:362-77.

Bone, E. and A. Farres. 2001. Trends and rates of microevolution in plants.

Genetica 112-113:165-82.

Cappuccino, N. and P. W. Price. 1995. Population Dynamics: New

Approaches and Synthesis. Academic Press.

Chitty, D. 1967. The Natural Selection of self-regulatory behaviour in

animal populations. Proceedings of the Ecological Society of

Australia:51-78.

Crawley, M. 1989. Insect Herbivores and Plant Population Dynamics.

Annual Review of Entomology 34:531-62.

Davis, J., E. Radcliffe, and D. Ragsdale. 2007. Effects of high and

fluctuating temperatures on Myzus persicae (Hemiptera: Aphididae).

Physiological ecology 35:1461-66.

de Barro, P., T. Sherratt, C. Brookes, O. David, and N. Maclean. 1995.

Spatial and temporal genetic variation in British field populations of

the grain aphid Sitobion avenae (F.)(Hemiptera: Aphididae) studied

using RAPD-PCR. Proceedings of the Royal Society B: Biological

Sciences 262:321-27.

Duffy, M. A. and L. Sivars-Becker. 2007. Rapid evolution and

ecological host- parasite dynamics. Ecology Letters 10:44-53.

Dyer, K. F. 1968. Evolution observed -- Some examples of evolution

occurring in historical times. Journal of Biological Education 2:317-

38.

Ehrlich, P. R. and P. H. Raven. 1964. Butterflies and Plants: A Study in

Coevolution. Evolution 18:586-608.

Endler, J. A. 1991. Genetic heterogeneity and ecology. P. 315–34 in R. J.

Berry,

T. J. Crawford, and G. M. Hewitt, editors. Genes in Ecology: The 33rd

Symposium of the British Ecological Society. Blackwell Scientific

Publications.

Fuller, S., P. Chavigny, L. Lapchin, and F. Vanlerberghe-Masutti. 1999.

Variation in clonal diversity in glasshouse infestations of the aphid,

Aphis gossypii Glover in southern France. Molecular Ecology 8:1867-

77.

Fussmann, G. F., S. P. Ellner, and N. G. J. Hairston. 2003. Evolution as a

critical component of plankton dynamics. Proceedings of the Royal

Society of London. Series B, Biological Sciences 270:1015-22.

Fussmann, G. F., M. Loreau, and P. A. Abrams. 2007. Eco-evolutionary

dynamics of communities and ecosystems. Functional Ecology

21:465-77.

Guldemond, J., W. J. Brink, and E. Belder. 1998. Methods of assessing

population increase in aphids and the effect of growth stage of the

host plant on population growth rates. Entomologia Experimentalis

et Applicata 86:163- 73.

Gurevitch, J., J. Morrison, and L. Hedges. 2000. The interaction

between competition and predation: A meta-analysis of field

experiments. American Naturalist 155:435-53.

Hairston, N. G. J., S. P. Ellner, M. A. Geber, T. Yoshida, and J. A. Fox.

2005. Rapid evolution and the convergence of ecological and

evolutionary time. Ecology Letters 8:1114-27.

Hazell, S., D. Gwynn, S. Ceccarelli, and M. Fellowes. 2005. Competition

and dispersal in the pea aphid: clonal variation and correlations across

traits. Ecological Entomology 30:293-98.

Hazell, S. P., I. A. D. McClintock, and M. D. E. Fellowes. 2006.

Intraspecific heritable variation in life-history traits can alter the

outcome of interspecific competition among insect herbivores. Basic

and Applied Ecology 7:215-23.

Hendry, A. P. and M. T. Kinnison. 1999. Perspective: The pace of

modern life: Measuring rates of contemporary microevolution.

Evolution 53:1637-53.

Hufbauer, R. A. and G. K. Roderick. 2005. Microevolution in

biological control: Mechanisms, patterns, and processes.

Biological Control 35:227-39.

Hughes, A. R., B. D. Inouye, M. T. J. Johnson, N. Underwood, and M.

Vellend. 2008. Ecological consequences of genetic diversity.

Ecology Letters 11:609- 23.

Karley, A., W. Parker, J. Pitchford, and A. Douglas. 2004. The mid-

season crash in aphid populations: why and how does it occur?

Ecological Entomology 29:383-88.

Kocourek, F., J. Havelka, J. Berankova, and V. Jarosik. 1994. Effect of

temperature on development rate and intrinsic rate of increase of

Aphis gossypii reared on greenhouse cucumbers. Entomologia

Experimentalis et Applicata 71:59-64.

Kokko, H. and A. Lopez-Sepulcre. 2007. The ecogenetic link between

demography and evolution: can we bridge the gap between theory

and data? Ecology Letters 10:773-82.

Leips, J., J. Travis, and F. H. Rodd. 2000. Genetic influences on

experimental population dynamics of the least killifish. Ecological

Monographs 70:289-309.

Levins, R., R. S. Singh, and M. K. Uyenoyama. 2003. Towards a

population biology, still. P. 21–47 The Evolution of Population

Biology.

Maron, J. L. and E. Crone. 2006. Herbivory: effects on plant

abundance, distribution and population growth. Proceedings of the

Royal Society B: Biological Sciences 273:2575-84.

Mitchell-Olds, T. and D. Bradley. 1996. Genetics of Brassica rapa .3.

Costs of disease resistance to three fungal pathogens. Evolution

50:1859-65.

Muller, C. B., I. S. Williams, and J. Hardie. 2001. The role of nutrition,

crowding and interspecific interactions in the development of winged

aphids. 26:330-40.

Onstad, D., S. Fang, and D. Voegtlin. 2005. Forecasting seasonal

population growth of Aphis glycines (Hemiptera: Aphididae) in

soybean in Illinois. Journal of Economic Entomology 98:1157-62.

Pelletier, F., D. Garant, and A. P. Hendry. 2009. Eco-evolutionary

dynamics.

Philosophical Transactions of the Royal Society B: Biological

Sciences 364:1483-89.

Pimentel, D. 1961. Animal population regulation by the genetic feed-

back mechanism. The American Naturalist 95:65-79.

Pimentel, D. 1968. Population regulation and genetic feedback.

Science 159:1432-37.

Pimentel, D. and R. Al-Hafidh. 1965. Ecological control of a parasite

population by genetic evolution in the parasite-host system. Annals

of the Entomological Society of America 58:1-6.

Pimentel, D., W. P. Nagel, and J. L. Madden. 1963. Space-time

structure of the environment and the survival of parasite-host

systems. The American Naturalist 97:141-67.

Pinheiro, J. and D. Bates. 2000. Mixed-Effects Models in S and

S-PLUS. Springer-Verlag.

Pinheiro, J. C., D. M. Bates, S. DebRoy, and D. Sarkar. 2009. nlme:

Linear and Nonlinear Mixed Effects Models.

R Development Core Team. 2009. R: A language and environment for

statistical computing. R Foundation for Statistical Computing,

Vienna, Austria.

Reznick, D. N. and C. K. Ghalambor. 2001. The population ecology of

contemporary adaptations: what empirical studies reveal about the

conditions that promote adaptive evolution. Genetica 112-113:183-

98.

Ro, T. and G. Long. 1999. GPA-Phenodynamics, a simulation model for

the population dynamics and phenology of green peach aphid in

potato: formulation, validation, and analysis. Ecological Modelling

119:197-209.

Rochat, J., F. Vanlerberghe-Masutti, P. Chavigny, R. Boll, and L.

Lapchin. 1999. Inter-strain competition and dispersal in aphids:

evidence from a greenhouse study. Ecological Entomology 24:450-

64.

Roughgarden, J. 1979. Theory of Population Genetics and Evolutionary

Ecology: An introduction.

Saccheri, I. and I. Hanski. 2006. Natural selection and population

dynamics.

Trends in Ecology & Evolution 21:341-47.

Schlager, G. 1963. The ecological genetics of the mutant Sooty in

populations of Tribolium castaneum. Evolution 17:254-73.

Shertzer, K., S. Ellner, G. F. Fussmann, and N. G. J. Hairston. 2002.

Predator- prey cycles in an aquatic microcosm: testing hypotheses

of mechanism.

Journal of Animal Ecology 71:802-15.

Sibly, R. and J. Hone. 2002. Population growth rate and its determinants:

an overview. Philosophical Transactions of the Royal Society of London

Series B- Biological Sciences 357:1153-70.

Slobodkin, L. B. 1980. Growth and Regulation of Animal Populations. 2nd

edition.

Dover Publications Inc.

Terhorst, C. P., T. E. Miller, and D. R. Levitan. 2010. Evolution of prey

in ecological time reduces the effect size of predators in

experimental microcosms. Ecology 91:629-36.

Thompson, J. N. 1998. Rapid evolution as an ecological process. Trends

in Ecology & Evolution 13:329-32.

Via, S. and A. J. Shaw. 1996. Short-term evolution in the size and

shape of pea aphids. Evolution 50:163-73.

Vorburger, C. 2005. Positive genetic correlations among major life-

history traits related to ecological success in the aphid Myzus

persicae. Evolution 59:1006- 15.

Vorburger, C. 2006. Temporal dynamics of genotypic diversity reveal

strong clonal selection in the aphid Myzus persicae. Journal of

Evolutionary Biology 19:97-107.

Wallner, W. 1987. FACTORS AFFECTING INSECT POPULATION-DYNAMICS:

Differences between outbreak and non-outbreak species. Annual

Review of Entomology 32:317-40.

Weber, G. 1985a. Genetic variability in host plant adaptation of the

green peach aphid, Myzus persicae. Entomologia Experimentalis et

Applicata 38:49-56.

Weber, G. 1985b. Population genetics of insecticide resistance in the

green peach aphid, Myzus persicae (Sulz.) (Homoptera, Aphididae).

Zeitschrift für Angewandte Entomologie 99:408-21.

Whitham, T. G., J. K. Bailey, J. A. Schweitzer, S. M. Shuster, R. K. Bangert,

et al.

2006. A framework for community and ecosystem genetics: from genes

to ecosystems. Nature Reviews Genetics 7:510-23.

Yoshida, T., L. E. Jones, S. P. Ellner, G. F. Fussmann, and N. G. J.

Hairston. 2003. Rapid evolution drives ecological dynamics in a

predator-prey system. Nature 424:303-06.

CHAPTER FOUR: THE IMPACT OF RAPID EVOLUTION ON

POPULATION DYNAMICS IN THE WILD: EXPERIMENTAL

TEST OF ECO-EVOLUTIONARY DYNAMICS

ABSTRACT

Rapid evolution challenges the assumption that evolution is too

slow to impact short-term ecological dynamics. This has led to a push

to empirically study how evolution and ecological processes

reciprocally impact each other on short time scales termed ‘Eco-

Evolutionary Dynamics’. In this study I tested how rapid evolution

impacts concurrent population dynamics using an aphid (Myzus

persicae) and an undomesticated host (Hirschfeldia incana) in replicated

wild populations. I manipulated the amount of genetic variation in

intrinsic growth rate within aphid populations, which altered rates of

evolution (changing clonal, or gene, frequencies) in both caged and

uncaged populations. Evolving populations grew significantly faster,

up to 42%, and reached higher densities, up to 67% higher, compared

to non-evolving control populations. Moreover, the magnitude of the

impact of evolution on population growth rate increased with observed

rates of evolution. Yet this effect only occurred in uncaged treatments

that were open to a natural spectrum of herbivores and predators.

Also, the relative fitness of competing clones changed with density.

This suggests that as evolution

changes density, density feeds back changes the selective

environment leading to reciprocal eco-evolutionary dynamics at the

same timescale. Finally, aphid evolution did not significantly influence

their host’s fitness.

INTRODUCTION

Ecological and evolutionary forces are usually thought to

influence each other asymmetrically, i.e. ecology shapes evolution

(Levins and Lewontin 1980, Hairston et al. 2005, Kokko and Lopez-

Sepulcre 2007). Ecological changes are often assumed to occur

independently of evolution, e.g. an organism’s population size is

reduced due to a drought. Evolutionary changes, however, are usually

a consequence of the ecological environment, e.g. desiccation

resistance evolves in response to droughts. Most ecological models

and studies make the simplifying assumption that evolution does not

impact short-term ecological processes because evolution is perceived

to act on a much slower time scale relative to ecological interactions

(Slobodkin 1980, Endler 1991, Thompson 1998, Hairston et al. 2005,

Pelletier et al. 2009). This assumption has now been challenged by

dozens of studies documenting rapid evolutionary changes in nature

occurring on ‘ecological time scales’, sometimes within a few

generations (Thompson 1998, Hendry and Kinnison 1999, Reznick and

Ghalambor 2001, Ashley et al. 2003). Given the convergence of time

scales, the next step is to assess whether rapid evolution and

concurrent ecological dynamics influence each other reciprocally

(Pelletier et al. 2009). This reciprocal interaction defines

eco-evolutionary dynamics (Hairston et al. 2005, Bull et al. 2006,

Kokko and Lopez-Sepulcre 2007). This process could have

important implications for fisheries, pest, and infectious diseases

management where accurate evolutionary and population

dynamic predictions are required (Hufbauer and Roderick 2005).

A growing body of eco-evolutionary dynamic theory, based on

very different biological assumptions, generally concludes that when

rapid evolution occurs during the course of an ecological interaction, it

can significantly alter quantitative and qualitative ecological

predictions (reviewed in Day 2005, Fussmann et al. 2007). Theoretical

models suggest that such eco-evolutionary dynamics can influence the

trajectory of growth of single populations (Anderson and King 1970),

the density and stability of victim-exploiter systems (Pimentel 1961,

Fussmann et al. 2003, Bull et al. 2006, Duffy and Sivars-Becker 2007),

the structure of multi-species communities (Loeuille and Leibold

2008), and even ecosystem processes (Loeuille et al. 2002).

I focus on quantifying the impact of rapid evolution on

concurrent ecological dynamics. This aspect of eco-evolutionary

dynamics has received less attention than the impact of ecology on

evolution (Bull et al. 2006, Ezard et al.

2009, Pelletier et al. 2009). One fruitful empirical approach consists of

using models to assess the influence of ecological and evolutionary

processes on observed ecological dynamics (Hairston et al. 2005,

Duffy and Sivars-Becker 2007, Ezard et al. 2009). In such studies,

rapid evolution is usually strongly

correlated with ecological dynamics and ecological predictions are

usually significantly improved by including evolutionary dynamics. A

second approach consists of post-hoc comparisons of the ecological

properties of ancestral populations versus populations that have

undergone evolution. Such studies assess how evolution has altered

life history traits (Reznick and Bryga 1996), population dynamics

(Hanski and Saccheri 2006), community structure (Post et al. 2008)

and ecosystem processes (Bassar et al. 2010). These field studies

demonstrate the strength and generality of rapid evolution’s

ecological effects.

Other empirical studies can quantify the impact of rapid

evolution as populations evolve. These powerful studies test the

causal impact of rapid evolution on population dynamics by

experimentally manipulating the occurrence of evolution itself thus

directly comparing evolving and non-evolving populations. Pimentel

first used this approach to show how the population dynamics of a

parasitoid wasp were changed as its housefly host evolved resistance

compared to a non-evolving control (Pimentel et al. 1963, Pimentel

and Al-Hafidh 1965, Pimentel 1968). By replacing the control housefly

population every generation, he prevented the evolution of resistance.

Rapid evolution within a few years in the host reduced the parasitoid’s

population size and variance even though host population size was

held constant. This experimental approach has only been attempted a

handful of times (Tuda 1998, Bohannan and Lenski 1999, Yoshida et

al. 2003, Fussmann et al. 2007, Agashe 2009, Terhorst et al. 2010). In

my previous study (Chapter 3) I experimentally quantified the impact

of aphid rapid

evolution on population growth rate compared to non-evolving

control populations in the greenhouse. I found that certain rapid

evolution treatments significantly accelerated population growth rate

by as much as 34%. These studies demonstrate how evolutionary

dynamics influence the dynamics and outcome of short-term

ecological phenomena and argue convincingly for causality, but only

under carefully controlled laboratory conditions.

Here I assess whether and how strongly rapid evolution impacts

concurrent ecological dynamics in natural populations. While

recognizing the value of lab studies, field experiments are crucial

because ecological context can influence both ecological and

evolutionary processes (Holt 2005). Experiments conducted in the wild

within realistic communities encompass more realistic levels of biotic

and abiotic variation as well as gene flow. These confounding factors

could impose different selective pressures, altering the rate or

direction of evolution itself, or they could interfere with the manner in

which rapid evolution impacts population dynamics, e.g. by altering

the strength of density regulation. All of these problems imply that

eco-evolutionary dynamics should be studied in the wild since non-

intuitive results could occur that differ significantly from predictions

based only upon laboratory experiments.

Most eco-evolutionary dynamics studies focus on interspecific

interactions and all experimental systems to my knowledge utilize

predator-prey or host- parasitoid model systems. Whether short-term

interspecific eco-evolutionary dynamics are important in plant-

herbivore systems remains an open question.

Although many studies show that plant genotypes can influence the

composition of arthropod communities (reviewed in Whitham et al.

2006, Hughes et al. 2008) they have yet to document these effects of

rapid evolution itself. In my previous greenhouse experiment I found

that aphids significantly harm their host plant, reducing above ground

biomass by a factor of five (Chapter 3). Yet aphid evolution, although

accelerating population growth rate, did not alter damage to the host.

Field populations permit better quantification of host plant fitness

differences because pollination can occur. Thus in this experiment I

more thoroughly quantify host fitness by investigating host

characteristics that are more tightly linked to fitness such as flower

and seed production in order to explore interspecific eco-evolutionary

dynamic effects.

My specific objective was to experimentally assess the impact of

rapid evolution on concurrent population dynamics in the wild. To do

so, I used my study system wherein evolution can easily be

manipulated. I selected the green peach aphid (Myzus persicae) and a

local wild invasive annual mustard, Hirschfeldia incana, as a host. I

manipulated aphid populations’ genetic composition and evolutionary

potential by controlling which clones were present within each

treatment. In order to explore the importance of ecological context I

conducted this study in caged and uncaged populations in the wild; the

difference is that uncaged populations admit a natural spectrum of

herbivores and predators. I specifically tested the following

hypotheses-predictions: 1) If rapid evolution impacts population dynamics,

then the observed population dynamics

of evolving aphid populations will differ significantly from those of non-evolving

aphids, 2) If ecological context is important, then the impact of rapid evolution on

population dynamics will differ between caged and uncaged treatments, and 3) If

rapid evolution in aphids impacts their host plant, then plant fitness will differ

significantly when exposed to evolving versus non-evolving aphids.

MATERIALS AND METHODS

I- Study System

Myzus persicae is considered the world’s most important crop

pest and thus its life-history and ecology are well-studied (Mackauer

and Way 1976). It is highly amenable to experimental evolution

because of its short generation time (6-10 days), ease of culture, and

the large magnitude of genetic variation identified in multiple traits

(Chapter 1; Vorburger 2005). This cyclically parthenogenetic aphid

undergoes sexual reproduction to survive cold winters. In the spring

populations are replete with multiple clonal lineages that reproduce

asexually until the fall (Mackauer and Way 1976). Aphid populations

rapidly

evolve through natural clonal selection within months, changing gene

frequencies and mean trait values (Via and Shaw 1996, Vorburger

2006). This parthenogenetic lifestyle permitted the experimental

manipulation of the level of genetic variation within a population by

controlling the initial frequency of clones therein.

In 2008, I collected multiple clonal lineages from a single wild

population from the Motte-Rimrock Reserve in Perris California. I

identified clones using 6 microsatellite markers and characterized their

intrinsic per capita growth rates experimentally (Chapter 2; Sloane et

al. 2001, Wilson et al. 2004). I selected three of these clones for this

and previous experiment (Chapters 2 & 3) that differ in intrinsic growth

rate (detailed below).

II- Field Experiment Design

The focal experiment was conducted in a wash area (20m by

12m), which was cleared of vegetation, at the Reserve where the

aphids were collected. A wire fence was erected to keep out large

vertebrate herbivores. Clonal aphid reproduction was maintained by

the long days and high temperature during the month of July

(Blackman 1974). To minimize variation in this primarily outcrossing

plant, plants used in the experiment were grown from the seeds of a

single H. incana plant collected in 2008 at the Reserve and germinated

in the greenhouse. Two week old seedlings, were planted 1.4 m apart

in the field site. Plants were watered three times a week because there

was no measurable precipitation during the experiment. Plants were

caged in a thin transparent mesh (Bridal Organza, #664-7242, Jo-Ann)

to prevent insect damage and permit aphid populations to become

established. On day 0 of the experiment, seven different aphid

treatments were initiated by placing 20 third instar aphids onto each

plant. These aphids came from stock greenhouse clonal populations,

regularly tested

for contamination, that were maintained on H. incana. Replicates were

initiated on three consecutive days starting June 30th 2009. On day 1

missing aphids were replaced with fourth instar aphids. On day 13, for

half of the plants, the cage mesh was lifted but tied to the top of the

wire frame. This maintained consistent shade between treatments but

allowed full access to the arthropod community. Thus aphid

treatments were fully crossed with caging treatments. In uncaged

plants, competitors, predators, pollinators, and other herbivores were

seen interacting with the aphids and their host.

I used three aphid clonal lineages, identified as ‘A’, ‘B’, and ‘C’,

that differ in microsatellite markers (Appendix 3.1). A preliminary

greenhouse experiment revealed that they also differ in exponential

growth rates when grown in pure cultures (Chapter 2, daily per capita

growth rate of clone A = 0.268 ± 0.002, B = 0.263 ± 0.004, and C =

0.251 ± 0.004). Using these clones, seven aphid treatments were

established as follows: three evolution treatments consisted of aphid

populations (on a single plant) that were initially composed of two

different clones (10 individuals of each clone for a total of 20 aphids).

These populations have genetic variation in fitness (e.g. A clone’s rm is

greater than that of the C clone) and thus could evolve by changing in

clonal frequency (away from the initial ratio). To explore how the rate

of evolution might alter the magnitude of evolution’s impact on

population dynamics, I established evolution treatments that consisted

of all three two-way combinations (A-B, B-C, and A-C). Three different

non-evolution (pure clone) treatments received aphids of only one of

the

three clones. Because all individuals were of the same genotype within

these pure populations, gene frequencies could not change, thus

preventing evolution. Finally, ‘no- aphid control’ treatments did not

receive any aphids. Each treatment combination was assigned in a

randomized block design and replicated 8 times for a total of 112

plants.

III - Rates of Evolution

On days 13, 20, and 31 I collected 20, 50, and 70 aphids

respectively from every population in order to track changes in clonal

frequencies (evolution).

Between 16 and 32 aphids from each sample were genotyped (for a

total of 2213 aphids) at three microsatellite loci (for detailed genetic

methods see Appendix 3.1). I calculated the frequency of the faster

clone for each treatment, sampling day and replicate separately. For

each treatment I then determined whether the mean frequency of the

faster clone differed significantly from the initial clonal frequency of

50% using one-sample t-tests. I also tested how caging, evolution

treatment, and their interaction impacts the final frequency of clones

using a general linear model. All analyses were implement in R (v.

2.11.1; R Development Core Team 2009) using the nlme package

(Pinheiro et al. 2009).

IV - Aphid Population Dynamics

Aphid population dynamics were quantified by counting all

aphids on days 3, 6, 10, 14, 17, 20, 24, 27, 31, and 36. When

populations rose above 2000

aphids per plant I sub-sampled by counting one half of every leaf. The

removal of cages on day 13 for half the treatments qualitatively

altered population dynamics. Because of this I analyzed caged and

uncaged treatments separately.

Treatments that remained caged for the entire experiment grew

exponentially until day 27 (Fig. 4.1.a). On day 31, population growth

started to slow and on day 36 populations crashed due to plant

senescence. The uncaged treatments were initially caged for 13 days

during which the aphids grew exponentially (Fig. 4.1.b). Once cages

were removed many lower leaves were damaged or consumed by

vertebrate herbivores which reduced growth rate. However, after this

reduction populations once again grew exponentially until day 31

before crashing on day 36 (Fig. 4.1.c). Thus I analyzed these time

periods separately (days 0 to 13 and then starting day 13 to 31) for

the uncaged treatments, this greatly improved residuals normality for

all treatments.

I thus had three separate analyses; caged from days 0 to 27,

uncaged days 0 to 13, and uncaged 13 to 31. For each of these I

tested for differences in population growth rate between treatments

using the following exponential

growth model:

Nt = N0 * e(rm *day)

where Nt is the number of aphids on day t, N0 is the number of aphids

on day 0, and rm is the intrinsic growth rate. Models were analyzed

using linear mixed- effect models (LME) where the dependent variable

was LN(x) transformed number of aphids and the fixed effects were

aphid treatment, day (as the main

covariate), and their interaction. Because the repeated aphid counts on

the same plant violated the assumption of independent observations, I

set unique plant identity, nested within block, as a random effect on

population growth rate and intercept and used an autoregressive

correlation error structure (Pinheiro and Bates 2000). I modeled

increasing variance through time by using a variance function that

increases with the power of the variance covariate (varPower).

Because plants differed slightly in size and stage of development at

the start of the experiment I explore whether the inclusion of these

covariates improved model fit. Initial plant size was measured as the

number of true leaves and rosette width and combined into a principal

component score. Only the first principal component was used since it

explained 73% of the variation. Stage of development was quantified

into an ordinal scales (1= rosette, 2 = low bolt, 3= bolt). Only in the

caged treatments did the covariates improve fit and were kept in the

final model. All analyses were implement in R (v. 2.11.1 R

Development Core Team 2009) using the nlme package (Pinheiro et al.

2009).

IV - The Impact of Evolution on Population Dynamics

My objective is not to predict which clone will out compete the

others but to statistically test the impact of changes in clonal

frequency on concurrent population dynamics (Hypothesis 1). Ideally

one would compare the observed aphid population dynamics in the

evolution treatment to those of a mixed population, containing the

same clones, but cannot evolve (remains at a

frequency of 50:50 even though clones differ in fitness). This is not

possible in my field experiment. I thus generated the expected

population growth parameters of a non-evolving population (that

remains at constant clonal frequency) by using the pure aphid

treatments. I tested three a priori null hypotheses that the population

growth rate (slope) and density (intercept) do not differ between each

evolution treatment and their corresponding pure treatments, e.g. A-C

evolution treatment vs pure A and pure C treatments. I did so with the

use of planned contrasts that are orthogonal comparisons between a

subset of the aphid treatment levels within the LME analysis. I set the

planned contrast coefficients of the no-evolution expectation to match

those of the initial clonal frequency (e.g. pure A= -0.5 and pure C= -0.5

and these are compared to the AC evolution treatment= 1). The

contrast matrix is found in Chapter 3. Thus differences in growth rate

or density between the evolution treatment and the no-evolution

expectation represent the impact of changes in the frequency of clones

(rapid evolution) on population dynamics. I present the results of

planned contrasts for the slope and intercept estimates of these

comparisons.

V - The Effect of Density on Natural Selection

I also investigated whether competing clones reacted differently

to changes in density in order to explore whether more complex eco-

evolutionary dynamics are occurring such as density-dependent

selection. To test this possibility I compared how the relative fitness of

competing clones, in each

evolution treatment, depends on density. For each population I

calculated the quantity of each clone on days 0, 13, 20, and 31 by

multiplying their genotypic frequencies by total population size. Next I

calculated per capita growth rate in each time period (days 0-13, 13-

20, and 20-31) using ln(N2)- ln(N1) / (day2-day1), where N= number of

aphids of this clone (Agrawal et al. 2004). This was done for each clone

in each evolution treatment. I then fitted a linear mixed-effect model

on these growth rates. I fit separate models for caged and uncaged

populations. Fixed effects were the treatment (combination of identity

of the focal clone, identity of the competitor; e.g. clone A competing

with clone B) and initial density in that time period and their

interaction. I included plant identity and block as a random effect as

well as autocorrelation error. With the use of planned contrasts I

determined if density differentially reduced growth rate between

competing clones.

VI - Host Plant Fitness

I first determine whether aphid treatments impact their host’s

growth by measuring final above ground dry biomass as done in

previous experiments. I fit a general linear model with plant weight as

the response variable, the fixed effect was aphid treatment, including

the 7th treatment which did not receive any aphids, the covariates

were initial plant stage of development and the first principal

component of plant size. Planned contrasts were used to determine

whether aphid herbivory in general influence plant final weight. Other

contrasts,

mimicking those used in the population dynamic models, were used to

determine whether plants with evolving aphid populations differed in

weight compared to plants with non-evolving populations.

I also quantified plant fitness traits. I first calculating flower-days

by summing my counts of the number of flowers present on every

sampling day multiplied by the length of that counting period. When

harvesting the plants I estimated the total number of seeds and mean

seed dry mass by sub-sampling. I performed a similar analysis as

described above but given correlation among traits I did so in a

MANCOVA framework.

Finally I determined whether final plant dry biomass predicted

plant fitness by fitting three separate linear models one focusing on

flower days, one on seed number, and one on mean seed weight. The

dependent variable was final plant weight, and the covariates were

stage and size of the plant at the start of the experiment. In all three

analyses these covariates were non-significant and did not improve

model fit and were thus removed. I thus present results from Pearson

correlations without the covariates.

RESULTS:

I - Pure Clone Treatments

Pure clone treatments differed greatly in their population

dynamics in the field (Fig. 4.1) as expected from greenhouse studies.

In the caged treatments, the exponential population growth rate

(slope) of the A clone was fastest. It grew

9.4% faster than the B clone and 14.5% faster than the C clone (Fig.

4.1.a). The B clone in turn grew 4.7% faster than the C clone. The

caged treatments reached very high densities approximately ten times

higher than the uncaged treatments (see y-axis scales in Figs. 4.1.a &

c). This difference occurred because of strong vertebrate herbivory I

observed when the cages were removed from the uncaged treatments

on day 13. This herbivory reduced plant biomass and in addition to

predation on the aphids slow the uncaged aphid’s population growth.

The pattern of growth rate between the pure clone populations was

similar in the uncaged treatments. Within the first 13 days, clones with

higher expected rm grew at faster rate. The A clone grew 9.8% faster

than the B clone and 19.3% faster than the C clone (Fig. 4.1.b). The B

clone in turn grew 8.6% faster than the C clone. After cages were

removed, the A clone still grew fastest, 12.1% faster than the B clone

and 19.3% faster than the C clone and finally the B clone grew 6.4%

faster than the C clone (Fig. 4.1.c).

II - Evolution Treatments - Genetic Analyses

Within only 20 days, in caged and uncaged populations, the

faster growing clones in the A-C and in the B-C evolution populations

significantly increased in frequency compared to their initial frequency

of 50% (one sample t-tests, all p- values < 0.03; Fig. 4.2). By day 31 all

evolution treatments showed significant evolution in the expected

direction (all p-values <0.01, except for caged A-B p= 0.1). The faster

growing clones in the caged populations reached on average

71% frequency, which is significantly less than the uncaged population

reaching 79% (p = 0.04, Fig. 4.2). This implies that caged treatments

evolved 38% slower than in uncaged (29% increase in frequency

divided by 21%). Aphid treatments differed significantly in their rate of

evolution (p < 0.001) the A-C and B-C treatments differed significantly

from A-B but not between each other (Fig. 4.2). Caging did not interact

with evolution treatment (p = 0.46, Fig. 4.2). Thus the A clone in the A-

C evolution treatments reached on average 85%, the B clone in the B-

C treatments reached 80%, and the A clone in the A-B treatments

reached 62% (Fig. 4.2). Thus evolution treatments that differ in their

rate of evolution were successfully created.

III - Impact of Aphid Evolution on Aphid Population Dynamics

To test the impact of rapid evolution on concurrent population

dynamics I compared the observed population dynamics in evolving

populations to those observed in both corresponding pure treatments

by using planned contrasts proportional to the initial frequency of

clones (i.e., population dynamics without evolution). In the caged

treatments, although rapid evolution occurred leading to an increase

in the frequency of the faster growing clone this did not alter the

evolving populations growth rate compared to their no-evolution

expectations (Table 4.1, Fig. 4.3). The B-C evolution treatment, even

though it grew 4.2% faster than its no-evolution expectation this

differences was not significant (p = 0.09). As for the uncaged

treatments, in the early growth phase, none of the

evolution treatments differed in growth rate or density from their no-

evolution expectations (Table 4.1, Figs. 4.4.a, b, c). After day 13, in

the uncaged treatments all evolution treatments grew at a

significantly faster rate than their corresponding no-evolution

expectations (between 33% and 42% faster, Table 4.1, Figs. 4.4.d, e,

f). I also created predicted population densities for each treatment

based on the model fitted parameters. This approach is more reliable

than simply comparing densities on day 31 because the fitted values

incorporate all the time series data. Thus predicted population sizes on

day 31 in the evolution treatments reached much higher densities

than expected without evolution (A-B treatment +13.7%, B-C +67%,

and A-C +17.5%, Table 4.1).

Although evolution in the A-B and B-C treatments have similarly

accelerated population growth rate (by 33% and 35%) their

predicted densities on day 31 differed greatly because the fitted

models incorporate differences in intercept. Moreover, the evolution

treatment with the slowest rate of evolution (A-B) showed the

smallest effect of rapid evolution on population growth rate. The

faster evolving treatments (A-C and B-C), which do not differ in their

rate of evolution, showed a stronger positive impact on population

dynamics.

IV -The Effect of Density on Natural Selection

I tested for density-dependent natural selection by fitting a

linear mixed- effect models that compared how strongly density

reduces the growth rate of each clone in the evolution treatments.

Increases in density more severely

reduced the growth of clone C than clone A (caged p<0.001; uncaged

p < 0.001, Table 4.2) and also compared to clone B (caged p= 0.065;

uncaged p < 0.001, Table 4.2). Thus the relative fitness of clone C

decreases in response to increased population density when

competing with another clone.

V - Impact of Aphid Evolution on Host Fitness

Because rapid evolution did not impact aphid dynamics in the

caged treatments I only present results for the uncaged treatments.

Firstly, my analysis revealed that aphid herbivory significantly reduced

final host biomass (mean ± SE weight with aphids = 15.3 g. ± 2.6,

weight without aphids = 20.8 g. ±2.2; GLM, p = 0.014). However,

although rapid aphid evolution significantly increased aphid growth

rate and density this did not cause more damage to the host plant than

the corresponding no-evolving aphid populations (GLM, all three p-

values > 0.11). My multivariate analysis of fitness for flower-days, seed

number, and seed weight found no difference between any aphid

treatments including the no-aphid treatment (MANCOVA, Wilks = 0.57,

F= 1.1, overall treatment p-value = 0.37, planned contrasts p-values

all > 0.12). Although no-aphid treatments had higher plant fitness

means there was a lot of variation that overwhelmed any trends (e.g.

mean ± SE for flower-days for no-aphid treatment = 1683 ± 373,

mean for aphid treatments = 1542 ± 135).

I also documented significant correlations between the above

ground dry biomass of the plants and the number of seeds produced

and total flower-days. Mean seed weight, however, was marginally

non-significant (p=0.084, Table 4.3).

DISCUSSION

I experimentally assessed the impact of rapid evolution on

concurrent ecological dynamics in the wild for the first time. I found

that rapid evolution, occurring within weeks, significantly accelerated

population growth rates and density as the frequency of faster

growing clones increased (Hypothesis 1), yet this change in aphid

density and growth rate did not impact host plant fitness (Hypothesis

3). I also found that the magnitude of the impact of rapid evolution on

population dynamics depended positively on the rapidity of evolution

and on ecological context (Hypothesis 2). Unexpectedly, in caged

populations, although rapid evolution occurred this had no impact on

population dynamics. This highlights the important affect that

ecological context can have the strength of eco-evolutionary

dynamics. My results have important implications for basic as well as

applied ecological and evolutionary biology.

In my field experiment, pure aphid treatments grew at

significantly different rates (Fig. 4.1). As expected, when populations

contained two clones their genetic composition quickly changed as the

faster clones increased in frequency (Fig. 4.2). For example the

frequency of the C clone in the uncaged A-C evolution populations was

reduced by more than four fold within only 31 days.

Similar large changes in clonal frequencies have been observed in

other wild aphid populations (Vorburger 2006). This change in clonal

frequency is not surprising nor is it novel. The novelty of my study lies

in quantifying the effect of this evolutionary change on the

population’s growth rate within that short time period. Hypothesis 1 was

supported but only in the natural uncaged treatments. Focusing on

these results, I saw that evolving populations grew at a significantly

faster rate than the expected rate if evolution is not taken into

account (Fig. 4.4, Table 4.1). Endler (1991) foreshadowed these results

while discussing early ecological-evolutionary models: “The time course

of total population size is not predictable from the average of the demographic

parameters of all genotypes. … genotypes contribute unequally and differently to

population size as they change in frequency during the course of natural

selection.” Moreover, the magnitude of this effect was dependent on the

observed rate of evolution (Hypothesis 2). The more quickly a

population evolves the strong impact of that rapid evolution on

population growth rate (Figs. 4.2 and 4.4, Table 4.1).

Even more complex and unexpected qualitative changes in

population dynamics have been observed in some laboratory studies

of rapid evolution. In Yoshida et al.’s (2003) study of rotifer and algae

in chemostats, rapid evolution in algae caused the predator-prey

population dynamics to change from being 1/4 out-of-phase to being

perfectly out-of-phase. In another rotifer study, Fussmann et al. (2003)

found that rapid evolution caused two peaks in population size but

only a single peak when populations could not evolve. These systems

might

inherently have more opportunity for qualitative changes in population

dynamics since both species undergo multiple generations within the

experiment. Thus the population dynamics and potentially the

evolutionary dynamics of both species might be altered. This is not

what occurs in many plant-herbivore systems. Many insect populations

grow rapidly then crash because of plant senescence, predation,

parasitism or climate (Wallner 1987, Ro and Long 1999, Karley et al.

2004). Such dynamics often occur within one generation of the plant.

Given differences in these types of interspecific interactions in nature

it is important to investigate the impact of rapid evolution in different

model systems.

All three uncaged evolution treatments accelerated population

growth rate. However evolution seems to have had a more pronounced

impact on population dynamics in the B-C evolution treatment (Fig

4.e). This occurred because the A- B and A-C evolution treatments had

lower day 13 densities than their corresponding no-evolution

expectations (Fig. 4.d & f), although these differences were not

significant (Table 4.1). Evolution in these two treatments compensated

for this effect and eventually overtook the no-evolution expectations.

On the other hand the B-C evolution treatment has similar intercept

values thus a 35% increase in growth has a very large effect on final

density.

Given these differences it is clear that the increase in growth rate in the

evolution treatments is not simply due to some evolution treatments

having lower day 13 densities. Overall my results, in addition to my

previous greenhouse experiment (Chapter 3 and those of Pimentel 1968,

Fussmann et al. 2003, Yoshida et al.

2003) argue that evolution should no longer be assumed to be too

slow to impact short-term ecological dynamics (Slobodkin 1980, Endler

1991, Thompson 1998, Hairston et al. 2005). Rapid evolution can be a

strong driver of population dynamics.

My experiment, however, demonstrated that ecological context

can have an important impact on rapid evolution’s ecological effects.

When I designed the experiment I created caged and uncaged

treatments because I hypothesized that environmental biotic variation

might overwhelm any effects of evolution. The impact of rapid

evolution in the uncaged treatments was evident even in the face of

potentially confounding factors. First, when cages were opened on day

13, vertebrate herbivores chewed off many large leaves causing a

reduction in aphid growth. Second, a competing aphid species

(Brevicoryne brassicae) colonized some plants (although remained at low

density) as did aphid predators (e.g. spiders, coccinellid larvae and

adults) that seemingly influenced the treatments randomly, adding to

variation between replicates. Finally, on average, 10% of the aphids

genotyped were immigrants. I did not remove the effect of these

ecological processes because they occur in natural populations. These

confounding factors and environmental variation strengthen the

importance of my results because the impact of rapid evolution on

population dynamics was still measurable and large.

I expected a stronger impact of rapid evolution on population

dynamics in the caged treatments because migration, interspecific

competition, predation,

and non-aphid herbivory were prevented. Surprisingly, even though

populations rapidly evolved in the caged treatments this did not

significantly impact population dynamics (Fig. 4.3, Table 4.1). I posit

that my ability to detect an effect of evolution might have been

reduced by the large population sizes reached in the caged

treatments (up to 50 000 aphids) because counts become less

accurate at such densities. The coefficient of variation for triplicate

counts of the same plant at a density of 6000 aphids was 0.02

whereas it was 0.13 for a plant with 26000 aphids. Moreover caging

significantly reduced the rate of evolution by 38% (Fig. 4.2) thus

potentially reducing the effect of evolution.

Aphids clones might differ in susceptibility to predation (Muller 1983,

Pilson 1992) a selective force not present in cages thus potentially

slowing the rate of evolution cages. Another possibility is that the rate

of evolution was reduced because populations were less density-

regulated than the uncaged populations. Stronger competition

between clones could magnify selective advantages and change the

selection environment (Table 4.2, and see text below). Aphid clones

are known to differ in how their growth rates respond to increased

population density (Agrawal et al. 2004). I argue that caged aphid

populations were under less severe population regulation because per

capita growth rates were much higher throughout the experiment

(mean daily rm= 0.25) compared to uncaged populations in the second

growth phase (rm = 0.08). This difference can be attributed to the fact

that the uncaged host plants were damaged by herbivores, which

reduced plant size. Final mean dry weight for caged plants was 52%

higher than uncaged (t-test, p < 0.001). Uncaged plants also had

smaller leaves most of which were lost by the end of the experiment in

contrast to caged plants. Stronger competition among aphids might

have caused more rapid evolution, in caged treatments, and hence a

stronger impact of evolution on population growth rate. A greenhouse

experiment using the same aphid clones supports these findings in

that aphid density can alter not only the clonal selection environment

but also the impact of rapid evolution on population dynamics (Chapter

5).

Saccheri & Hanski (2006) proposed that population density is less

likely to be influences by natural selection or evolution if that

population is under strong density regulation, which should overwhelm

for eco-evolutionary effects. The population dynamics in my system do

not behave as though they are tightly regulated yet my results still

suggest that density affects the selective environment leading to non-

intuitive outcomes.

Conducting studies in this emerging field in realistic ecological

conditions is crucial. The choice of genotypes are studied can also

alter the rate of evolution, which can change the impact strength of

rapid evolution on population dynamics. Together my results

emphasize that to properly assess the relative importance of evolution

on population dynamics requires experiments under natural

conditions. These should manipulate ecological variables in a realistic

manner, e.g. mimicking natural levels of variation in competition, and

also use genetic variation present within populations (i.e. not

artificially magnifying variation by using non-local genotypes).

The importance of genetic variation and rapid evolution on other

members of the community is currently receiving much attention

(Hughes et al. 2008, Johnson et al. 2009). I investigated whether this

is occurring in my system by assessing host plant fitness. Aphid rapid

evolution could impact its host’s fitness if one clone is more damaging

per capita, and thus an increase in its frequency would

correspondingly reduce plant fitness. More directly an increase in

aphid density through rapid evolution could also increase damage.

Although aphid evolution increased aphid population growth rate and

density, by up to 42%, this did not impact plant fitness or plant final

biomass (Hypothesis 3). One possibility is that non-aphid herbivore

damage determined plant fitness, overwhelming the aphid’s impact.

This suggestion is supported by the result that plants being consumed

by aphids, although having lower biomass, did not differ in fitness

from plants without aphids. Finally, micro-environmental variation

might have increased variation between replicates in plant fitness

(coefficients of variation among replicates: range of 0.1-1.9 with a

mean of 0.6). To properly quantify the interspecific effects of rapid

evolution in natural populations one must not experimentally exclude

such confounding ecological processes.

My experiment focused on the less-studied half of the eco-

evolutionary dynamics cycle. Rapid evolution impacts concurrent

population dynamics and its host fitness but is the reciprocal causal

process also occurring? I found that the competitive advantage of the

A and the B clones over C clone significantly increases at higher

density, in both caged and uncaged treatments. This

suggests the possibility of a full eco-evolutionary feedback cycle

where both rapid evolution and ecological dynamics influence each

other on similar timescales (Kokko and Lopez-Sepulcre 2007). As rapid

evolution leads to higher densities this alters natural selection thus

potentially modifying future bouts of evolution. Comparing my caged

and uncaged genetic results the strength of density regulation might

be more important than the absolute density in influencing the

selective environment. I limit my interpretation of these results since

my experimental design cannot disentangle the effect of density and

time. The C clone might simply have a fitness disadvantage on mature

plants, which is confound with density in this experiment.

In many applied fields such as fisheries, pest management,

conservation biology, invasion biology, and epidemiology, accurate

predictions of population dynamics are crucial. Rapid evolution in

many traits has been observed in these systems (Ashley et al. 2003).

My results suggest that considering rapid evolution as a force

impacting concurrent population dynamics might provide important

insight into many of these applied issues (Hufbauer and Roderick

2005, Duffy and Sivars-Becker 2007). For example in pest

management, models are used to establish guidelines for the

economic use of pesticides yet these almost always assume non-

evolving population growth parameters. My results argue that this

oversight might be problematic. Even in a small population, where

drift might be stronger, within only 31 days, predictions made without

considering rapid evolution could underestimate actual population

growth rate by up to 42% and

population size by 67%. This could significantly alter optimal control

strategies. My novel study suggests predictions of population

dynamics and rapid evolution will be more accurate if they explicitly

consider these processes and their interactions at all timescales.

Table 4.1: Analysis of Population Dynamics Comparing Evolving Aphids to Non-Evolving Controls for

Caged and Uncaged Treatments

Planned contrasts from linear mixed-effect models, comparing each type of evolving population to its

corresponding predicted non-evolving expectation for the caged treatments (days 0 to 27), for the early growth

phase (days 0 to 13) and the late phase (days 13 to 31) of the uncaged treatments. Evolution treatments are

identified by their clonal composition. Intercept represents density at the start of each time period, slope

represents aphid population growth rate, and predicted density represent expected density on the last day of

the time period based on the best fit model parameter estimates. The percent change represents the change

in intercept, slope, or final density, from the non-evolving expectation to that of the observed evolution

treatment. Positive values represent increase due to rapid evolution. All p-values are for 2-tailed tests.

Significant results were bolded for easier identification.

Table 4.1

Evolution

Treatment Caged (days 0 to 27 Uncaged (days 0 to 13) Uncaged (days 13 to 31)

(Clones) d.f. F p %

Change d.f. F p %

Change

Intercept 29 1.3 0.212 14.7% 30 -1.3 0.208 -13.2% 30 -1.9 0.071 -25.7%

A-B Slope 345 -0.8 0.443 -1.7% 166 -0.9 0.361 -4.9% 200 2.5 0.012 33.3%

Predicted 0.19% -29.0% 13.7%

Density

Intercept 29 -0.7 0.506 -7.9% 30 1.0 0.307 12.4% 30 0.6 0.568 9.1%

B-C Slope 345 1.7 0.088 4.2% 166 -0.7 0.487 -4.2% 200 2.4 0.016 35.2%

Predicted 12.6% -2.8% 67.0%

Density

Intercept 29 2.3 0.026 27.8%30 -1.3 0.188 -15.1% 30 -1.8 0.079 -29.0%

A-C Slope 345 -0.6 0.581 -1.3% 166 -0.1 0.905 -0.7% 200 2.8 0.006 41.7%

Predicted

Density 0.1% -21.5% 17.5%

Table 4.2: Analysis of Density-Dependent Clonal Selection

Summary of density-dependent clonal selection analysis. Results of planned contrasts comparing

how

strongly density reduces the growth rate of each clone in the evolution treatments. Slopes

represent daily growth rate as a function of initial density in each time period (see methods for

details). All slope values (means and SE) should be multiplied by 10-4. All p-values are for 2-tailed

tests.

Treatments

Slope of

Faster

Slope of % Difference

Slower in Slope d.f. t-value p-value

Clone (SE) Clone (SE)

A-B -0.35 (0.14) -0.27 (0.12) 30.1% 85 -0.8 0.4284

Caged B-C -0.22 (0.11) -0.59 (0.2) -62.7% 85 1.9 0.0653

A-C -0.29 (0.13) -0.66 (0.14) -56.6% 85 3.5 0.0008

A-B -4.2 (0.7) -3.2 (0.7) 32.7% 78 -1.5 0.1497

Uncaged B-C -2.8 (0.5) -7.2 (1.1) -61.3% 78 4.1 0.0001

A-C -2.7 (0.7) -7.8 (1.2) -65.0% 78 4.3 0.0001

116

Table 4.3: Correlation Between Final Plant Weight and Plant Fitness Traits

Summary of Pearson correlation analyses between final plant dry

above ground biomass and three fitness traits. All correlations were

positive. All p-values are for 2-tailed tests.

Number

Weight

Fitness Trait d.f. t-value p-value Pearson Correlation

Total Seed 49 5.19 0.0000 0.60

Mean Seed 40 1.77 0.0838 0.27

Flower-Days 49 3.62 0.0007 0.46

117

Figure 4.1: Population Dynamics of Pure Clone Populations for Caged and

Uncaged Treatments

Partial residual plots of population dynamics, over the time periods used in

the analyses, of pure clonal treatments in field experiment. Values represent

mean number of aphids (±1 SE) through time once all explained variation in

the model has been removed. Panel

(a) represents the caged treatments during days 0 to 27, (b) early growth

phase of the uncaged plants before cages were removed on day 13, and (c)

late growth phase of uncaged plants once cages were removed. Notice that

panels differ in y-axis scales. Functions represent the best fit exponential

model for each treatment (clone A is gray, clone B is the full black line, and

clone C is the dashed black line).

118

119

Figure 4.2: Temporal Changes in Clonal Frequency During the Experiment

Rapid clonal evolution as shown by the mean frequency of the faster growing

aphid clones in each evolution treatment (±1 SE). Panels separate caged (a)

and uncaged (b) treatments Horizontal bar indicates initial clonal frequencies

of 0.5.

120

Figure 4.3: Population Dynamics Comparing Evolving Aphids to Non-

Evolving Controls in Caged Treatments

Partial residual plot of population dynamics of the three observed evolution

treatments (black diamonds) with the best fit model from LME analysis (black

line) for caged treatments. The dashed gray line represents the best fit model

that combines both pure treatments using the constant (non-evolving)

frequency of clones (50:50). For each treatment I added the corresponding

pure clone treatments (grey symbols) used to generate the no-evolution

expectation. Evolution treatments have two letters. Values represent mean

number of aphids (±1 SE) once all explained variation in the model has been

removed.

121

Figure 4.4: Population Dynamics Comparing Evolving Aphids to Non-

Evolving Controls in Caged Treatments

Partial residual plot of population dynamics of the three observed evolution

treatments (black diamonds) with the best fit model from LME analysis (black

line) for uncaged treatments (caged results can be seen in Appendix 3.1).

The dashed gray line represents the best fit model that combines both pure

treatments using the constant (non-evolving) frequency of clones (50:50). For

each treatment I added the corresponding pure clone treatments (grey

symbols) used to generate the no-evolution expectation. Evolution treatments

have two letters. Values represent mean number of aphids (±1 SE) once all

explained variation in the model has been removed. The y-axes differ

between panels. Left panels are for the early growth phase (days 0 to 13) and

right panels are for the late growth analysis (days 13 to 31).

122

Figure 4.4

123

LITERATURE CITED

Agashe, D. 2009. The stabilizing effect of intraspecific genetic

variation on population dynamics in novel and ancestral habitats.

The American Naturalist 174:255-67.

Agrawal, A. A., N. Underwood, and J. R. Stinchcombe. 2004.

Intraspecific variation in the strength of density dependence in aphid

populations.

Ecological Entomology 29:521-26.

Anderson, W. W. and C. E. King. 1970. Age-specific selection.

Proceedings of the National Academy of Sciences 66:780-86.

Ashley, M. V., M. F. Willson, O. R. W. Pergams, D. J. O'Dowd, S. M.

Gende, et al. 2003. Evolutionarily enlightened management.

Biological Conservation 111:115-23.

Bassar, R. D., M. C. Marshall, A. López-Sepulcre, E. Zandona, S. K. Auer, et

al.

2010. Local adaptation in Trinidadian guppies alters ecosystem

processes. Proceedings of the National Academy of Sciences

107:3616-21.

Blackman, R. 1974. Life-cycle variation of Myzus persicae (Sulz.)(Hom.,

Aphididae) in different parts of the world, in relation to genotype and

environment. Bulletin of Entomological Research 63:595-607.

Bohannan, B. and R. Lenski. 1999. Effect of prey heterogeneity on the

response of a model food chain to resource enrichment. The

American Naturalist 153:73-82.

Bull, J. J., J. Millstein, J. Orcutt, and H. A. Wichman. 2006. Evolutionary

feedback mediated through population density, illustrated with viruses

in chemostats.

124

American Naturalist 167:E39-E51.

125

Day, T. 2005. Modelling the ecological context of evolutionary change:

Déjà vu or something new? P. 273-309 in K. Cuddington and B. E.

Beisner, editors.

Ecological Paradigms Lost Routes of Theory Change. Elsevier

Academic Press.

Duffy, M. A. and L. Sivars-Becker. 2007. Rapid evolution and ecological

host- parasite dynamics. Ecology Letters 10:44-53.

Endler, J. A. 1991. Genetic heterogeneity and ecology. P. 315–34 in R. J.

Berry,

T. J. Crawford, and G. M. Hewitt, editors. Genes in Ecology: The 33rd

Symposium of the British Ecological Society. Blackwell Scientific

Publications.

Ezard, T. H. G., S. D. Cote, and F. Pelletier. 2009. Eco-evolutionary

dynamics: disentangling phenotypic, environmental and population

fluctuations.

Philosophical Transactions of the Royal Society B: Biological

Sciences 364:1491-98.

Fussmann, G. F., S. P. Ellner, and N. G. J. Hairston. 2003. Evolution as a

critical component of plankton dynamics. Proceedings of the Royal

Society of London. Series B, Biological Sciences 270:1015-22.

Fussmann, G. F., M. Loreau, and P. A. Abrams. 2007. Eco-evolutionary

dynamics of communities and ecosystems. Functional Ecology

21:465-77.

Hairston, N. G. J., S. P. Ellner, M. A. Geber, T. Yoshida, and J. A. Fox.

2005. Rapid evolution and the convergence of ecological and

evolutionary time. Ecology Letters 8:1114-27.

Hanski, I. and I. Saccheri. 2006. Molecular-level variation affects

population growth in a butterfly metapopulation. PLoS Biology

4:e129.

Hendry, A. P. and M. T. Kinnison. 1999. Perspective: The pace of

modern life: Measuring rates of contemporary microevolution.

126

Evolution 53:1637-53.

Holt, R. D. 2005. On the integration of community ecology and

evolutionary biology: Historical perspectives and current prospects.

P. 235-72 in K.

127

Cuddington and B. Beisner, editors. Ecological Paradigms Lost

Routes of Theory Change. Elsevier.

Hufbauer, R. A. and G. K. Roderick. 2005. Microevolution in

biological control: Mechanisms, patterns, and processes.

Biological Control 35:227-39.

Hughes, A. R., B. D. Inouye, M. T. J. Johnson, N. Underwood, and M.

Vellend. 2008. Ecological consequences of genetic diversity.

Ecology Letters 11:609- 23.

Johnson, M. T. J., M. Vellend, and J. R. Stinchcombe. 2009. Evolution

in plant populations as a driver of ecological changes in arthropod

communities.

Philosophical Transactions of the Royal Society B: Biological

Sciences 364:1593-605.

Karley, A., W. Parker, J. Pitchford, and A. Douglas. 2004. The mid-

season crash in aphid populations: why and how does it occur?

Ecological Entomology 29:383-88.

Kokko, H. and A. Lopez-Sepulcre. 2007. The ecogenetic link between

demography and evolution: can we bridge the gap between theory

and data? Ecology Letters 10:773-82.

Levins, R. and R. Lewontin. 1980. Dialectics and reductionism in ecology.

Synthese 43:47-78.

Loeuille, N. and M. A. Leibold. 2008. Ecological consequences of

evolution in plant defenses in a metacommunity. Theoretical

Population Biology 74:34-45.

Loeuille, N., M. Loreau, and R. Ferriere. 2002. Consequences of plant-

herbivore coevolution on the dynamics and functioning of

ecosystems. Journal of Theoretical Biology 217:369-81.

Mackauer, M. and M. J. Way. 1976. Myzus persicae Sulz. an aphid of

world importance. P. 51-119 in V. L. Delucchi, editor. Studies in

128

biological control. Cambridge University Press.

129

Muller, F. P. 1983. Differential alarm pheromone responses between

strains of the aphid Acyrthosiphon pisum. . Entomologia

Experimentalis et Applicata 34:347–48.

Pelletier, F., D. Garant, and A. P. Hendry. 2009. Eco-evolutionary

dynamics.

Philosophical Transactions of the Royal Society B: Biological

Sciences 364:1483-89.

Pilson, D. 1992. Insect distribution patterns and the evolution of host

use. P. 120- 39 in R. S. Fritz and E. L. Simms, editors. Plant resistance

to herbivores and pathogens: ecology, evolution, and genetics

Unviersity of Chicago Press.

Pimentel, D. 1961. Animal population regulation by the genetic feed-

back mechanism. The American Naturalist 95:65-79.

Pimentel, D. 1968. Population regulation and genetic feedback.

Science 159:1432-37.

Pimentel, D. and R. Al-Hafidh. 1965. Ecological control of a parasite

population by genetic evolution in the parasite-host system. Annals

of the Entomological Society of America 58:1-6.

Pimentel, D., W. P. Nagel, and J. L. Madden. 1963. Space-time

structure of the environment and the survival of parasite-host

systems. The American Naturalist 97:141-67.

Pinheiro, J. and D. Bates. 2000. Mixed-Effects Models in S and

S-PLUS. Springer-Verlag.

Pinheiro, J. C., D. M. Bates, S. DebRoy, and D. Sarkar. 2009. nlme:

Linear and Nonlinear Mixed Effects Models.

Post, D. M., E. P. Palkovacs, E. G. Schielke, and S. I. Dodson. 2008.

Intraspecific variation in a predator affects community structure and

cascading trophic interactions. Ecology 89:2019-32.

130

R Development Core Team. 2009. R: A language and environment for

statistical computing. R Foundation for Statistical Computing,

Vienna, Austria.

Reznick, D. N. and H. A. Bryga. 1996. Life-history evolution in guppies

(Poecilia reticulata: Poeciliidae): V. Genetic basis of parallelism in life

histories.

American Naturalist 147:339-59.

Reznick, D. N. and C. K. Ghalambor. 2001. The population ecology of

contemporary adaptations: what empirical studies reveal about the

conditions that promote adaptive evolution. Genetica 112-113:183-

98.

Ro, T. and G. Long. 1999. GPA-Phenodynamics, a simulation model for

the population dynamics and phenology of green peach aphid in

potato: formulation, validation, and analysis. Ecological Modelling

119:197-209.

Saccheri, I. and I. Hanski. 2006. Natural selection and population

dynamics.

Trends in Ecology & Evolution 21:341-47.

Sloane, M. A., P. Sunnucks, A. C. C. Wilson, and D. F. Hales. 2001.

Microsatellite isolation, linkage group identification and

determination of recombination frequency in the peach-potato

aphid, Myzus persicae (Sulzer) (Hemiptera: Aphididae). Genetical

Research 77:251-60.

Slobodkin, L. B. 1980. Growth and Regulation of Animal Populations. 2nd

edition.

Dover Publications Inc.

Terhorst, C. P., T. E. Miller, and D. R. Levitan. 2010. Evolution of prey in

ecological time reduces the effect size of predators in experimental

microcosms. Ecology 91:629-36.

Thompson, J. N. 1998. Rapid evolution as an ecological process. Trends

in Ecology & Evolution 13:329-32.

131

Tuda, M. 1998. Evolutionary character changes and population

responses in an insect host-parasitoid experimental system.

Researches on Population Ecology 40:293-99.

132

Via, S. and A. J. Shaw. 1996. Short-term evolution in the size and shape

of pea aphids. Evolution 50:163-73.

Vorburger, C. 2005. Positive genetic correlations among major life-

history traits related to ecological success in the aphid Myzus

persicae. Evolution 59:1006- 15.

Vorburger, C. 2006. Temporal dynamics of genotypic diversity reveal

strong clonal selection in the aphid Myzus persicae. Journal of

Evolutionary Biology 19:97-107.

Wallner, W. 1987. FACTORS AFFECTING INSECT POPULATION-DYNAMICS:

Differences between outbreak and non-outbreak species. Annual

Review of Entomology 32:317-40.

Whitham, T. G., J. K. Bailey, J. A. Schweitzer, S. M. Shuster, R. K. Bangert,

et al.

2006. A framework for community and ecosystem genetics: from

genes to ecosystems. Nature Reviews Genetics 7:510-23.

Wilson, A. C. C., B. Massonnet, J. C. Simon, N. Prunier-Leterme, L.

Dolatti, et al. 2004. Cross-species amplification of microsatellite loci

in aphids: assessment and application. Molecular Ecology Notes

4:104-09.

Yoshida, T., L. E. Jones, S. P. Ellner, G. F. Fussmann, and N. G. J.

Hairston. 2003. Rapid evolution drives ecological dynamics in a

predator-prey system. Nature 424:303-06.

133

CHAPTER FIVE: AN EXPERIMENTAL ASSESSMENT OF THE

FULL ECO-EVOLUTIONARY DYNAMICS CYCLE BETWEEN

RAPID EVOLUTION AND POPULATION DENSITY AND THEIR

RELATIVE IMPACTS

ABSTRACT

This study builds on previous findings that aphid rapid evolution

impacts concurrent population dynamics by experimentally testing

whether such changes in density alter the course of evolution. This

represents the first explicit experimental test of the full eco-

evolutionary dynamic cycle (two-way causality between ecological and

rapid evolutionary dynamics) in a plant-herbivore system. Using

aphids and mustard, I manipulated aphid initial density and aphid

rapid evolution. I found strong evidence for density-dependent

selection. Initial aphid density altered the rate and outcome of

evolution, as measured by changes in clonal frequency. Density also

quantitatively and qualitatively altered how rapid evolution impacts

concurrent population growth rate. Aphid evolution, within only 16

days, significantly accelerated population growth rate (by up to

10.3%) compared to non-evolving controls in some combinations of

134

clones. Yet in one treatment combination, rapid evolution reduced

population growth by 6.4%. These effects were similar in magnitude to

the reduction in population

135

growth rate caused by a three-fold increase in initial density. These

results suggest that the full eco-evolutionary dynamic cycle is

occurring in this system and can have relatively strong effects on both

population dynamics and rapid evolution. I also identified one

treatment where aphid evolution significantly augmented the damage

caused to the host plant. These results have important implications for

the study of population dynamics and pest management.

INTRODUCTION

The reciprocal causal influences between evolution, genotypic

variation, and ecological dynamics termed ‘Eco-Evolutionary Dynamics’

(Fig. 1.1) is currently receiving much attention because such

dynamics can alter the predicted outcome of ecological interactions

(reviewed in Fussmann et al. 2007, Johnson and Stinchcombe 2007,

Pelletier et al. 2009, Schoener 2011). This renewed interested

emanates from recent empirical studies of eco-evolutionary dynamics

occurring on short-time scales, i.e. within a few dozen generations

(Fussmann et al. 2003, Yoshida et al. 2003, Hairston et al. 2005). Eco-

evolutionary dynamics, occurring on these timescales, have

traditionally been ignored because of a widely held assumption that

evolution is too slow to influence ecological dynamics (Hairston et al.

2005, Fussmann et al. 2007).

136

Thus much of the work in evolutionary-ecology and ecological genetics

focused on how ecological conditions (abiotic, competitive

environment) and ecological dynamics (density fluctuations) cause

selection and evolution (focused on arrow

137

1 in Fig. 1.1; Levins et al. 2003, Johnson and Stinchcombe 2007). The

reciprocal arrow of causality was rarely studied but important

exceptions do exist (Pimentel 1961, Chitty 1967, Pimentel 1968).

Many reviews listing examples of rapid evolution have recently

challenged the assumption that evolution is too slow to have an

impact on contemporary ecological interactions (Thompson 1998,

Hendry and Kinnison 1999, Bone and Farres 2001, Reznick and

Ghalambor 2001). Such examples of rapid evolution inspired a few but

growing number of empirical studies that quantify how rapid evolution

alters ecological dynamics (Bohannan and Lenski 2000, Fussmann et

al. 2003, Yoshida et al. 2003, Terhorst et al. 2010). For example, in my

previous studies using the green peach aphid, Myzus persicae, I

experimentally quantified that aphid rapid evolution significantly

accelerates population growth rate within a few weeks compared to

non-evolving control populations in the greenhouse (Chapter 3) and in

the field (Chapter 4). Yet, my previous experiments focused on one

arrow of causality in the eco-evolutionary dynamics cycle, which is

how rapid evolution might alter ecological dynamics (arrow 2 in Fig.

1.1). Eco- evolutionary dynamics can potentially be more complex in

this system if both processes are linked through cyclical causality (Fig.

1.1). Ecological changes induced by rapid evolution could alter future

bouts of evolution by changing the selection experienced by the target

138

organism. Full eco-evolutionary dynamics seem to be occurring in a

side-blotched lizards field system (Sinervo et al. 2000, Svensson and

Sinervo 2000) and in rotifer-algae chemostat experiments

139

(Yoshida et al. 2003). In both cases density cycles lead to, and are

caused by, cycles in phenotypic composition thus linking ecological

and evolutionary dynamics. In most studies in this emerging field

however, only parts of the cyclical pathway are investigated.

My previous studies have shown that rapid evolution can

increase population growth rate and density (arrow 2 in Fig. 1.1;

Chapters 3 & 4) but it remains unknown whether the full eco-

evolutionary dynamic cycle is occurring. The ecological change that I

observed in the first bout of eco-evolutionary dynamics (altered

population growth rate) could alter the affect of evolution on

population dynamics in the second bout in at least two ways. First,

changes in density might alter the selective environment directly

through density-dependent aphid clonal selection (thus modifying

arrow 1 in Fig. 1.1). Aphid clones are known to differ in how their

growth rates respond to increased population density (Agrawal et al.

2004). My field experiment (Chapter 4) found correlerative evidence

for density-depdendent clonal selection but more robust experiments

are required that do not confound host plant age with density. If the

higher population densities, caused by aphid evolution, feed back and

alters the selective environment then the aphids might not evolve at

the rate or even direction as predicted within a constant selective

environment. Second, increased density could also affect the strength

140

of population regulation, which might interfere with the ability of rapid

evolution to impact population growth rate

141

or density (altering arrow 2 in Fig. 1.1). Strong population regulation

could overwhelm any effect of rapid evolution (Saccheri and Hanski

2006).

Even if rapid evolution occurs within a system and has an

ecological impact, this does not guarantee a full eco-evolutionary

dynamic cycle. This could be the case if the ecological change does not

change the selective environment and does not feedback to influence

future evolution. Theoretical studies suggest that a break in the

feedback cycle could lead to very different ecological and evolutionary

outcomes (Anderson 1971, Abrams and Matsuda 1997, Shertzer et al.

2002, Day 2005). Thus the objective of this study is to explicitly assess

whether the full eco-evolutionary dynamic feedback cycle is occurring

in my study system. I will extend previous studies by testing 1) whether

aphid clonal selection is density-dependent by comparing the final aphid clonal

frequency in replicated populations initiated at different densities. I am thus

testing whether ecological context impacts the selective environment

(if arrow 2 impacts arrow 1 in Figure 1.1). Given these density

treatments I will also test 2) whether the impact of rapid evolution on

concurrent population dynamics depends on initial density (if arrow 1

influences arrow 2 in Fig. 1.1) by comparing whether treatments initiated at

different densities differ in how strongly evolution alters population growth rate.

Also by conducting these experiments with different aphid clones I will

142

test 3) whether the evolutionary context (genotypic composition of the

population) affects the results. Finally, simply showing that eco-

evolutionary dynamics occur fails in telling us how important this

process is. It is thus

143

imperative to compare its effect size with other commonly studied

ecological forces (Hairston et al. 2005, Johnson and Stinchcombe 2007,

Pelletier et al. 2009). Thus I will also 4) tests the relative impact of rapid

evolution on population growth rate compared to the effect of manipulating

density.

METHODS

I- Study System

The system used in this study is the green peach aphid, Myzus

persicae, and one of its many hosts the annual mustard, Hirschfeldia

incana. M. persicae is a globally distributed species best know for its

tremendous impact as an agricultural pest (Blackman 1974, Mackauer

and Way 1976). I developed this system to study eco-evolutionary

dynamics because the aphid naturally evolves rapidly, through

changes in clonal frequency, during its asexual growth phase in the

spring-summer months (Foster et al. 2002, Vorburger 2006). In 2008, I

collected multiple clonal lineages from a single wild population from

the Motte- Rimrock Reserve in Perris California. I identified clones using

6 microsatellite markers (Chapter 2; Sloane et al. 2001, Wilson et al.

2004) and characterized their intrinsic per capita growth rates

experimentally (Chapter 2). I selected three of these clones for this and

previous experiment (Chapters 2, 3, & 4) that differ in intrinsic growth

144

rate (detailed below).

145

II- Experimental Design

Evolving and non-evolving M. persicae populations on H. incana

plants were studied in a partly cooled greenhouse (mean daytime

temperature = 33°C, range = 21°C to 47°C, mean nighttime = 21.5°C,

range = 16°C to 28.5°C).

Asexual reproduction was maintained by using addition lighting

providing 16hrs light/ 8hrs dark (Blackman 1974). To minimize

variation in this primarily outcrossing plant, plants used in the

experiment were grown from the seeds of a single H. incana plant

collected in 2008 at the Motte-Rimrock Reserve. These seeds were

planted in four liter pots, using UCR soil mix III, a sand/peat moss mix

supplemented with micronutrients, and watered every three days.

Cages were constructed to individually house each plant. Cages

consisted of an 8 liter pot with a wire frame creating a 75cm high

dome that held up thin transparent mesh (Bridal Organza, #664-

7242, Jo-Ann). On day 0 of the experiment, the eight week old plants,

were inoculated with 6 different aphid treatments by placing a given

number of third instar aphids onto each plant. These aphids came

from stock greenhouse clonal populations, regularly tested for

contamination, that were maintained on H. incana (Chapter 2).

Replicates were initiated on three consecutive days starting July 25th

2010. On day 1 missing aphids were replaced with fourth instar

146

aphids.

I used three aphid clonal lineages, identified as ‘A’, ‘B’, and ‘C’,

that differ in microsatellite markers (Appendix 3.1). A preliminary

greenhouse experiment revealed that they also differ in exponential

growth rates when grown in pure

147

cultures (Chapter 2, daily per capita growth rate of clone A = 0.268 ±

0.002, B = 0.263 ± 0.004, and C = 0.251 ± 0.004). I used these three

clones to establish 12 aphid treatments in a factorial design that

manipulated density (low and high ~3x) and aphid evolution

(possibility of evolution or not). Three evolution treatments consisted

of aphid populations (on a single plant) that have two different clones

at an initial frequency of 50% : 50%. I established evolution treatments

that consisted of all three two-way combinations (A-B, B-C, and A-C).

These populations have genetic variation in fitness (e.g. A clone’s rm is

greater than that of the C clone) and thus could evolve by changing in

clonal frequency (away from the initial ratio). Three different non-

evolution (pure clone) treatments received aphids of only one of the

three clones. Because all individuals were of the same genotype within

these pure populations, gene frequencies could not change, thus

preventing evolution. I also manipulated initial density by initiating

these six treatments with either 20 or 60 third instar aphids on each

plant. I tripled the initial density because in my field collections I often

observed such a range of density on single plants (M. M. Turcotte

personal observation). Each treatment combination was assigned in a

randomized block design and replicated 7 times for a total of 84 plants.

III- Rates of Evolution

148

On day 22 I collected 50 aphids from every population in order to

track changes in clonal frequencies (evolution). Between 16 and 40

aphids from each

149

sample were genotyped (for a total of 892 aphids) at three

microsatellite loci using a multiplex approach (see Appendix 3.1 for

detailed genetic methods). I calculated the frequency of the faster

clone for each treatment and replicate separately. For each treatment

I then determined whether the mean frequency of the faster clone

differed significantly from the initial clonal frequency of 50% using

one-sample t-tests. I also tested whether the frequency of the faster

clone in each evolution treatment differed between initial density

using 2-sample t- tests.

IV- Aphid Population Dynamics

Aphid population dynamics were quantified by counting all

aphids on days 3, 6, 10, 13, 16, 22, 25, and 29. When populations rose

above 2000 aphids per plant I sub-sampled by counting one half of

every leaf. By fitting different population growth models I found that

the results were best described by fitting an exponential model from

days 0 to 16. On day 22 plants were starting to senesce, having

multiple yellow and/or drying leaves. When the logistic growth model

was extended to include day 22 many treatments had severely

overestimated population sizes on that day. On day 26, population size

declined as the populations crashed. Given the crash of aphid density

due to plant senescence, logistic growth models no longer yielded

150

accurate estimates of carrying capacity.

151

Differences in population growth rate between treatments were

tested using the following exponential growth model:

Nt = N0 * e(rm *day)

where Nt is the number of aphids on day t, N0 is the number of aphids

on day 0, and rm is the intrinsic growth rate. This model was analyzed

with a non-linear mixed-effect model (NLME). The fixed effect for N0

was initial density (low or high) and the fixed effects for rm were initial

density, aphid treatment, and their interaction. Because the repeated

aphid counts on the same plant violated the assumption of

independent observations, I set unique plant identity as a random

effect on population growth rate and intercept and used an

autoregressive correlation error structure (Pinheiro and Bates 2000). I

modeled increasing variance through time by using a variance

function that increases exponentially with the variance covariate

(varExp) but that differs between both density treatments. Block (day

of initiation of the replicate and spatial position in the greenhouse)

did not improve model fit and was not included in the final model.

Because plants differed slightly in size and stage of development on

day 0 these data were included as covariates for population growth

rate. Initial plant size was measured as the number of true leaves and

rosette width and combined into a principal component score. Only the

first principal component was used since it explained 70% of the

152

variation. Stage of development was quantified into an ordinal scales

(1= low bolt, 2 = bolt, 3= flowers). All analyses were implement in

153

R (v. 2.11.1 R Development Core Team 2009) using the nlme package

(Pinheiro et al. 2009).

V- The Impact of Evolution on Population Dynamics

My objective is not to predict which clone will out-compete the

other but to assess the impact of changes in clonal frequency on

concurrent population dynamics. Ideally one would compare the

observed aphid population dynamics in the evolution treatment to

those of a non-evolving mixed population containing the same two

clones (remains at a frequency of 50% : 50%). This is impossible since

clones will change in frequency because of fitness differences. This is

not possible in my field experiment. I thus generated the expected

population growth parameters of a non-evolving population (that

remains at constant clonal frequency) by using the pure aphid

treatments. I tested three a priori null hypotheses that the population

growth rate (slope) do not differ between each evolution treatment

and their corresponding pure treatments, e.g. A-C evolution treatment

vs pure A and pure C treatments. I did so with the use of planned

contrasts that are orthogonal comparisons between a subset of the

aphid treatment levels within the NLME analysis. I set the planned

contrast coefficients of the no-evolution expectation to match those of

the initial clonal frequency (e.g. pure A= -0.5 and pure C= -0.5 and

154

these are compared to the AC evolution treatment= 1). The contrast

matrix is found in Chapter 3. Thus differences in population growth

rate between the evolution treatment and the no-evolution

155

expectation represent the impact of changes in the frequency of clones

(rapid evolution) on population dynamics. To enhance interpretation of

the magnitude of differences in population growth rate I calculated

expected densities on day 16 from the model parameters and set

common covariate values (mean values).

VI - Host Plant Fitness

Finally, to quantify the impact of aphid rapid evolution on its

host’s fitness I measured the above ground dry biomass of the plants

at the end of the experiment as a proxy for host fitness (Chapter 4,

Mitchell-Olds and Bradley 1996). I fit a general linear model on LN(x)

transformed plant weight measurements. The factors were treatment

(combination of aphid treatment and density), initial plant size PCA

score, and stage of development. The interactions between these

factors were non-significant and thus removed from the final model. I

again used planned contrasts to determine whether plants with

evolving aphid populations were smaller than expected from no-

evolution treatments.

RESULTS

I - Pure Clone Treatments

Pure clone treatments differed in their intrinsic growth rate and

156

these differences changed in magnitude and rank with density (Fig.

5.1). At low initial density, differences between pure treatments were

smaller between clones.

Clone A grew fastest, its rm value was 2.4% higher than clone B and 5.3%

higher

157

than clone C (Fig. 5.1.a). Clone B grew faster than clone C by 2.9%. At

high density however there was a change in rank order (Fig. 5.1.b).

Clone A grew 21% faster than clone B and 14 % faster than clone C,

and clone B grew 5.7% slower than clone C. Expected density on day

16 based on the best fit growth estimates predict that pure

population could differ by as much as 27% in low density and 117% at

high density.

II - Genetic Analyses of Density-Dependent Selection

On day 22 I tested for changes in the deviation of the frequency

of clones in the evolution treatments from the clonal initial frequency

of 50% : 50%. All treatments rapidly evolved except for one within this

period and clear evidence for density-dependent clonal evolution was

observed (Fig. 5.2). The A-B evolution treatments significantly evolved

but only at high density (frequency of A clone = 57%, one sample t-

test, p = 0.02). In the B-C evolution treatments density altered the

direction of evolution; at low density the B clone became more

frequent reaching 69% (p=0.017) whereas at high density clone C

reached 70% (p=0.004). Finally in the A-C evolution treatments the A

clone became more common reaching 71% (p=0.002) and 65%

(p=0.044) in the low and high density treatments respectively (Fig.

5.2).

158

III - Impact of Initial Aphid Density and Evolution on Aphid Population Dynamics

Although high initial density treatments reached higher densities

on day 16 (see axis scale in Figs. 5.1 and 5.3) these treatments had

significantly reduced population growth rates. Overall the high density

treatments had 9.5% lower population growth rates (NLME, F= 25, p <

0.0001). To test the impact of rapid evolution on concurrent population

dynamics I compared the observed population dynamics in evolving

populations to those observed in both corresponding pure treatments

by using planned contrasts proportional to the initial frequency of

clones (i.e., population dynamics without evolution). Given that the

interaction between initial density and aphid treatment was significant

(NLME, F= 4.8, p = 0.0003) I conducted planned contrasts for each

density separately (see Table 5.1). Evolution in the A-B evolution

treatments did not significantly alter population dynamics at either

density (Fig. 5.3.a & b, Table 5.1, both p > 0.2). The impact of

evolution in the B-C treatments was highly influenced by initial

density. At low density, rapid evolution (i.e., an increase in the

frequency of clone B) slowed population growth by 6.4% (p=0.0004,

Fig.

5.3.c) whereas at high density rapid evolution (i.e., an increase in

clone C) accelerated population growth by 10.3% (p=0.003, Fig. 5.3.d,

Table 5.1). Expected density on day 16 based on the best fit growth

159

estimates predict that rapid evolution causes differences in population

size of –25% and +48% respectively. Finally, in the A-C evolution

treatments initial density magnified the effect of rapid evolution. At

low density rapid evolution marginally accelerated

160

population growth rate by 1.1% (p=0.022, Fig. 5.2.e) yet at high

density evolution accelerated population growth by 9.4% (p= 0.0001,

Fig. 5.2.f, Table 5.1). These differences in population growth rate due

to evolution are predicted to cause differences on day 16 of +4.3% and

+43% in population size respectively.

IV - Impact of Aphid Evolution on Host Plant Fitness

My analysis of final plant weight did not reveal an overall effect

of initial aphid density (p=0.25). Of the six comparisons between

evolving populations and their corresponding no-evolution treatments

only one revealed a significant difference. The evolution treatment

with the largest effect on population growth, high density B-C,

significantly reduced its host’s biomass by 27% (GLM, p= 0.030).

DISCUSSION:

I experimentally assessed eco-evolutionary dynamics by

manipulating both density and rapid evolution in a plant-herbivore

system. I found that aphid populations rapidly evolved through

changes in clonal frequencies but the magnitude and direction of

evolution depended on initial density. Aphid evolution significantly

altered population growth rate but the direction and magnitude

depended upon initial density and genotypic composition. These

161

results suggest that the full eco-evolutionary dynamics cycles are

occurring in this system (Fig. 1.1) and can have relatively strong

effects on both population dynamics and rapid

162

evolution. Rapid aphid evolution changed population growth rate and

density influenced the course of evolution. I also identified one

treatment where aphid rapid evolution significantly augmented the

damage cause to the host plant.

These results have important implications for the study of population

dynamics and pest management because they suggest that neither

short-term ecological or evolutionary dynamics should be studied

separately.

My prior experiments in this system experimentally tested the

less studied half of eco-evolutionary dynamics. I showed that over a

period of 28 days rapid evolution significantly accelerated population

growth rate in the greenhouse by 28% to 34% (Chapter 3). In my field

experiment (Chapter 4), even in the presence of environmental

variation, immigration, the presence of competitors, herbivores,

predators, and parasitoids, rapid evolution still significantly

accelerated population growth rate by 33% to 42% within 31 days.

These experiments have thus revealed that evolution can affect

concurrent ecological dynamics (arrow 2 in Fig. 1.1). The experiment

presented here builds on the previous ones experimentally quantifying

the full eco-evolutionary cycle by assessing changes in population

density, which is the product of population growth and rapid evolution,

can in turn feed back on and influence the outcome of interclonal

163

competition.

My experiment revealed clear evidence of density-dependent

selection. Initial density had multifarious effects on the pattern of

evolution depending on genotypic composition (Fig. 5.2).

Differences in density altered whether rapid

164

evolution occurred or not (A-B evolution treatments), it affected the

direction/outcome of evolution (B-C evolution treatments, Fig. 5.2). My

goal is not to predict the outcome of clonal selection but to document

that changes in density alter the outcome. However, it remains true

that differences in population growth rate in single clone cultures most

often predicted the direction of evolution correctly (e.g. the change in

outcome of B-C treatments) but not always the rate of evolution (e.g.

when A and B were grown together at high density, A reached only

57% frequency even though the pure A clone grows 21% faster than

pure B clone). Overall my experiment revealed density-dependent

clonal evolution, which suggests that as density changes due to rapid

evolution, the selective environment is also changing in a way that

alters the evolutionary outcome, thus linking rapid evolution and

population dynamics in a cyclical causal pathway.

Multiple traits might explain why one clone out-competes

another. Aphid clones vary in many traits including life-history traits,

competitive ability, predator and pathogen resistance, adult weight,

body size, morphology, feeding rate, probing behavior, plant choice,

alate production, tendency to drop from plants, response to alarm

pheromones, photoperiod response, ability to transmit viruses, and

temperature tolerance (Muller 1983, Pilson 1992, Via and Shaw 1996,

Rochat et al. 1999, Hazell et al. 2005, Vorburger 2005, Hazell and

165

Fellowes 2009), although some of these differences might be

attributable to secondary endosymbionts (Oliver et al. 2010). I

conducted a life-table experiment on the aphids clones used in this

experiment and found that clones A, B, and C differ in

166

lifetime fecundity and survival rates (M.M. Turcotte, unpublished data).

Changes in aphid density could interact with a number of these

differences between clones thus leading to density-dependent

selection. Aphid clonal variation in the strength of density-regulation as

also been observed (Agrawal et al. 2004).

To further test the eco-evolutionary dynamics cycle, I also tested

whether density would alter how rapid evolution affects population

dynamics (how arrow 1 affects arrow 2 in Fig. 1.1). Initial density

quantitatively and qualitatively altered how rapid evolution impacts

population growth rate (Fig. 5.3, Table 5.1). In the A- C evolution

treatments an increase in initial density led to a much stronger effect

of evolution on population dynamics compared to non-evolving

controls (+1.1% versus +9.4%). This magnification of the impact of

evolution might be explained by the fact that the fitness differences

between clones A and C are exaggerated at higher densities (Fig. 5.1).

Thus an increase in frequency of clone A would more strongly

accelerate population dynamics in the high density treatment than in

the low density treatment. Overall, increased initial density led to

stronger effects of rapid evolution on population dynamics (Table 5.1).

This is consisted with my field experiment where higher levels of

competition also led to stronger impacts (Chapter 4). The mechanism

seems to be stronger clonal selection at higher densities. I still posit

167

that this magnification of the effect of rapid evolution on population

growth rate might be diminished if density regulation is strong enough

to maintain the populations near a carrying capacity (Saccheri and

168

Hanski 2006). These conditions do not seem to apply to my aphid

system where the population grows rapidly and crashes.

Initial density also qualitatively changed the affect of evolution

in the B-C treatments. At high density the C clone becomes more

frequent and since clone C’s growth rate is higher than clone B it

accelerates the evolution treatment’s population growth rate by 10.3%

above the non-evolving control (Table 5.1, Fig. 5.3.d). At low density

the B clone becomes more dominant, reaching a frequency of 69%, but

this evolutionary change slows down population growth rate by 6.4%,

even though the B clone grows more quickly than the C clone at low

initial density. I posit that the B clone interferes with its own growth

and that of the C clone when it reaches a high frequency and density.

Interference could occur through resource competition or competition

for feeding site. Alternatively, if clones vary in their propensity to

produce or respond to alarm pheromones emitted at high density

(Muller 1983) this might alter the clones feeding and reproduction.

Thus the evolving population’s growth slows as it becomes dominated

by the B clone. A similar hypothesis was proposed in Chapter 3, where

I documented evidence that clone B has reduced growth at high

density and frequency. Such intraspecific variation in the strength of

density dependence has been document in other aphid species

(Agrawal et al. 2004) as well as in this experiment (Fig. 5.1). Overall

169

my results demonstrate the dual causal interplay between rapid

evolution and concurrent ecological dynamics.

170

Study systems where the full eco-evolutionary dynamics have

been shown are very rare (Svensson and Sinervo 2000, Yoshida et al.

2003). In some systems only a tentative hypothesis is suggested. For

example in Palkovacs and Post’s (2008) aquatic system, populations of

alewifes (Alosa pseudoharengus) have become landlocked by dams and

have evolved within 300 years. The ancestral anadromous type is only

present in certain freshwater lakes in the spring and summer months.

Because of year round foraging, the landlocked fish have altered the

zooplankton community by favoring smaller species and smaller

individuals within species. It is suggested that these ecological

changes have fed back and caused the evolution of more efficient

foraging morphology (smaller gill raker spacing and gape width) and

higher prey selectivity in the landlocked alewifes compared to

anadromous forms (Palkovacs and Post 2008). Thus eco- evolutionary

dynamic cycles are inferred as a plausible explanation for existing

patterns. Given that we are just looking at patterns without a clear

establishment of cause and effect relationships, we must accept that

alternative explanations may account for them. Thus clear advantage

of my study system is that because of rapid evolution occurring within

a few weeks I can dissect the eco-evolutionary dynamics as they occur

and establish causality.

My experiment revealed that evolution altered population growth

171

by 1.1% to 10.3%, which is smaller than the effect sizes I found in

previous experiments (Chapters 3 & 4). I posit that is because I used

more mature plants in the current study. This led to an earlier

termination of the experiment as the plants

172

senesced, which in turn reduced the number of days (and hence

generations) of usable data in the experiment from 28 and 30 days in

Chapters 3 & 4 to 16 days in the current study. This reduced time

period limits the extent of evolution and the time evolution can take to

impact population dynamics. In the previous experiments I found that

evolution had no significant effects in the early time periods. Moreover

the reduced time implies smaller population sizes, which could

minimize the potential impact of an acceleration of population growth

rate. Yet, my current results are surprising in that they show a

significant effect of evolution on population growth rate over only 3-4

generations. Most other studies of the ecological effects of rapid

evolution study time periods representing many dozen generations

(Tuda and Iwasa 1998, Fussmann et al. 2003, Yoshida et al. 2003,

Terhorst et al. 2010).

Recent reviews of eco-evolutionary dynamics stress that it is of

vital importance to test the relative impact of rapid evolution

compared to other ecological forces on population dynamics compared

to more commonly studied ecological processes (Hairston et al. 2005,

Johnson and Stinchcombe 2007, Pelletier et al. 2009). My experimental

design lends itself well to address this issue. I found that a three-fold

increase in initial density reduced population growth rate by 9.5%.

Rapid evolution in two different evolution treatments had similar effect

173

sizes (Table 5.1) suggesting that rapid evolution and eco- evolutionary

dynamics can be important forces in population dynamics on short-

timescales.

174

Unlike any other experimental study system to my knowledge, I

have shown eco-evolutionary dynamics within a single species. This

suggests that short-term eco-evolutionary dynamics could be very

common, even occurring in systems that do not show strong

population dynamic coupling with other species (as opposed to

predator-prey systems). I also observed that aphid rapid evolution, in

the treatment with largest impact on aphid population dynamics, had

an interspecific impact on its host. The evolving aphid population

reduced plant biomass by 27% compared to non-evolving controls.

This suggests that eco- evolutionary dynamics could extend beyond

the aphids in this system. More experiments will be required to

explore this issue.

My results have important implication for the study of both

population dynamics and evolutionary dynamics. It suggests that

even on very short timescales of a few generations rapid evolution

can occur and can significantly alter population dynamics in complex

ways. Moreover the rate and direction of such rapid evolution can be

influenced by changes in population dynamics thus completing the

eco-evolutionary dynamics cycle. These results suggest that in

certain systems ecological and evolutionary predictions would be

improved if eco-evolutionary dynamics were explicitly considered.

Improving predicted population growth rates should enhance our

175

ability to manage agricultural pests and exploited populations such

as in fisheries.

176

Table 5.1: Analysis of Population Dynamics Comparing Evolving and Non-

Evolving Populations in Low and High Density Treatments

The effect of rapid evolution on population growth rate compared to non-

evolving controls. Results of planned contrasts from a non-linear mixed-effect

model, comparing growth rates of each type of evolving population to its

corresponding no-evolution expectation, generated from the pure aphid

treatments following the initial frequency of clones (see Methods for details).

The percent change represents the change in slope from the non-evolving

expectation to that of the observed evolution treatment. Thus positive

changes represent increases due to evolution. Slope represents the rate of

growth of aphid populations. All p-values are for 2-tailed tests. Significant

results were bolded for easier identification.

Evolutio

Treatme

Low Density High Density

(Clones) d.f. F p % Change d.f. F p % Change

A-B 397 0.3 0.577 +1.0% 397 1.6 0.205 +3.6%

B-C 397 12.8 0.0004 -6.4% 397 9.1 0.0028 +10.3%

A-C 397 5.3 0.022 +1.1% 397 15.2 0.0001 +9.4%

177

Figure 5.1: Population Dynamics of Pure Clone Treatments at High and Low

Densities

Partial residual plots of population dynamics of pure clonal treatments.

Values represent mean number of aphids (±1 SE) through time, once all

explained variation in the model has been removed, separated by a) low

initial density treatments and b) high initial density treatment. Functions

represent the best fit exponential model for each treatment (clone A is gray,

clone B is the full black line, and clone C is the dashed black line). The y-axes

differ between panels.

178

Figure 5.2: Rapid Evolution Through Changes in Clonal Frequency at Different

Initial Aphid Densities

Rapid clonal evolution as shown by the mean frequency (±1 SE) of the faster

growing aphid clones on day 22 of the experiment. The horizontal bar

indicates the initial clonal frequency of 50% on day 0. The X-axis shows which

clone’s frequency is being tested in each evolution treatment. Dashed vertical

lines separate the evolution treatments for which both initial density

treatments are shown. The (*) indicate significant divergence from initial

frequency and (**) indicates that clone frequency significantly differs between

low and high initial density within an evolution treatment.

179

Figure 5.3: Population Dynamics of Evolving and Non-Evolving Aphids at Different

Initial Densitites

Partial residual plot of population dynamics of the three observed evolution

treatments (black diamonds) with the best fit model from LME analysis (black

line). Left panels are for the low initial density treatments whereas the right

panels are for the high initial density treatments. The dashed gray line

represents the best fit model that combines both pure treatments using the

constant (non-evolving) frequency of clones (50:50).

Corresponding pure clone treatments (grey symbols) used to generate the no-

evolution expectation. Values represent mean number of aphids (±1 SE)

once explained variation in the model was removed. The y-axes differ

between panels.

180

LITERATURE CITED

Abrams, P. A. and H. Matsuda. 1997. Prey adaptation as a cause of

predator- prey cycles. Evolution 51:1742-50.

Agrawal, A. A., N. Underwood, and J. R. Stinchcombe. 2004.

Intraspecific variation in the strength of density dependence in aphid

populations.

Ecological Entomology 29:521-26.

Anderson, W. 1971. Genetic equilibrium and population growth under

density- regulated selection. The American Naturalist 105:489-98.

Blackman, R. 1974. Life-cycle variation of Myzus persicae (Sulz.)(Hom.,

Aphididae) in different parts of the world, in relation to genotype and

environment. Bulletin of Entomological Research 63:595-607.

Bohannan, B. and R. Lenski. 2000. Linking genetic change to

community evolution: insights from studies of bacteria and

bacteriophage. Ecology Letters 3:362-77.

Bone, E. and A. Farres. 2001. Trends and rates of microevolution in plants.

Genetica 112-113:165-82.

Chitty, D. 1967. The Natural Selection of self-regulatory behaviour in

animal populations. Proceedings of the Ecological Society of

Australia:51-78.

Day, T. 2005. Modelling the ecological context of evolutionary change:

Déjà vu or something new? P. 273-309 in K. Cuddington and B. E.

Beisner, editors.

Ecological Paradigms Lost Routes of Theory Change. Elsevier

Academic Press.

181

Foster, S., R. Harrington, A. Dewar, I. Denholm, and A. Devonshire. 2002.

Temporal and spatial dynamics of insecticide resistance in Myzus

persicae

(Hemiptera: Aphididae). Pest Management Science 58:895-907.

182

Fussmann, G. F., S. P. Ellner, and N. G. J. Hairston. 2003. Evolution as a

critical component of plankton dynamics. Proceedings of the Royal

Society of London. Series B, Biological Sciences 270:1015-22.

Fussmann, G. F., M. Loreau, and P. A. Abrams. 2007. Eco-evolutionary

dynamics of communities and ecosystems. Functional Ecology

21:465-77.

Hairston, N. G. J., S. P. Ellner, M. A. Geber, T. Yoshida, and J. A. Fox.

2005. Rapid evolution and the convergence of ecological and

evolutionary time. Ecology Letters 8:1114-27.

Hazell, S., D. Gwynn, S. Ceccarelli, and M. Fellowes. 2005. Competition

and dispersal in the pea aphid: clonal variation and correlations

across traits. Ecological Entomology 30:293-98.

Hazell, S. P. and M. D. E. Fellowes. 2009. Intra-specific variation

affects the structure of the natural enemy assemblage

attacking pea aphid colonies. Ecological Entomology 34:34-42.

Hendry, A. P. and M. T. Kinnison. 1999. Perspective: The pace of

modern life: Measuring rates of contemporary microevolution.

Evolution 53:1637-53.

Johnson, M. T. J. and J. R. Stinchcombe. 2007. An emerging synthesis

between community ecology and evolutionary biology. Trends in

Ecology and Evolution 22:250-57.

Levins, R., R. S. Singh, and M. K. Uyenoyama. 2003. Towards a

population biology, still. P. 21–47 The Evolution of Population

Biology.

Mackauer, M. and M. J. Way. 1976. Myzus persicae Sulz. an aphid of

world importance. P. 51-119 in V. L. Delucchi, editor. Studies in

biological control. Cambridge University Press.

Mitchell-Olds, T. and D. Bradley. 1996. Genetics of Brassica rapa .3.

Costs of disease resistance to three fungal pathogens. Evolution

183

50:1859-65.

184

Muller, F. P. 1983. Differential alarm pheromone responses between

strains of the aphid Acyrthosiphon pisum. . Entomologia

Experimentalis et Applicata 34:347–48.

Oliver, K. M., P. H. Degnan, G. R. Burke, and N. A. Moran. 2010.

Facultative symbionts in aphids and the horizontal transfer of

ecologically important traits. Annual Review of Entomology 55:247-

66.

Palkovacs, E. P. and D. M. Post. 2008. Eco-evolutionary interactions

between predators and prey: can predator-induced changes to

prey communities feed back to shape predator foraging traits?

Evolutionary Ecology Research 10:699-720.

Pelletier, F., D. Garant, and A. P. Hendry. 2009. Eco-evolutionary

dynamics.

Philosophical Transactions of the Royal Society B: Biological

Sciences 364:1483-89.

Pilson, D. 1992. Insect distribution patterns and the evolution of host

use. P. 120- 39 in R. S. Fritz and E. L. Simms, editors. Plant resistance

to herbivores and pathogens: ecology, evolution, and genetics

Unviersity of Chicago Press.

Pimentel, D. 1961. Animal population regulation by the genetic feed-

back mechanism. The American Naturalist 95:65-79.

Pimentel, D. 1968. Population regulation and genetic feedback.

Science 159:1432-37.

Pinheiro, J. and D. Bates. 2000. Mixed-Effects Models in S and

S-PLUS. Springer-Verlag.

Pinheiro, J. C., D. M. Bates, S. DebRoy, and D. Sarkar. 2009. nlme:

Linear and Nonlinear Mixed Effects Models.

R Development Core Team. 2009. R: A language and environment for

statistical computing. R Foundation for Statistical Computing,

185

Vienna, Austria.

186

Reznick, D. N. and C. K. Ghalambor. 2001. The population ecology of

contemporary adaptations: what empirical studies reveal about the

conditions that promote adaptive evolution. Genetica 112-113:183-

98.

Rochat, J., F. Vanlerberghe-Masutti, P. Chavigny, R. Boll, and L.

Lapchin. 1999. Inter-strain competition and dispersal in aphids:

evidence from a greenhouse study. Ecological Entomology 24:450-

64.

Saccheri, I. and I. Hanski. 2006. Natural selection and population

dynamics.

Trends in Ecology & Evolution 21:341-47.

Schoener, T. W. 2011. The newest synthesis: Understanding the

interplay of evolutionary and ecological dynamics. Science 331:426-

29.

Shertzer, K., S. Ellner, G. F. Fussmann, and N. G. J. Hairston. 2002.

Predator- prey cycles in an aquatic microcosm: testing hypotheses of

mechanism.

Journal of Animal Ecology 71:802-15.

Sinervo, B., E. Svensson, and T. Comendant. 2000. Density

cycles and an offspring quantity and quality game driven by

natural selection. Nature 406:985-88.

Sloane, M. A., P. Sunnucks, A. C. C. Wilson, and D. F. Hales. 2001.

Microsatellite isolation, linkage group identification and

determination of recombination frequency in the peach-potato

aphid, Myzus persicae (Sulzer) (Hemiptera: Aphididae). Genetical

Research 77:251-60.

Svensson, E. and B. Sinervo. 2000. Experimental excursions on

adaptive landscapes: density-dependent selection on egg size.

Evolution 54:1396-403.

Terhorst, C. P., T. E. Miller, and D. R. Levitan. 2010. Evolution of prey in

ecological time reduces the effect size of predators in experimental

187

microcosms. Ecology 91:629-36.

Thompson, J. N. 1998. Rapid evolution as an ecological process. Trends

in Ecology & Evolution 13:329-32.

188

Tuda, M. and Y. Iwasa. 1998. Evolution of contest competition and its

effect on host–parasitoid dynamics. Evolutionary Ecology 12:855-70.

Via, S. and A. J. Shaw. 1996. Short-term evolution in the size and shape

of pea aphids. Evolution 50:163-73.

Vorburger, C. 2005. Positive genetic correlations among major life-

history traits related to ecological success in the aphid Myzus

persicae. Evolution 59:1006- 15.

Vorburger, C. 2006. Temporal dynamics of genotypic diversity reveal

strong clonal selection in the aphid Myzus persicae. Journal of

Evolutionary Biology 19:97-107.

Wilson, A. C. C., B. Massonnet, J. C. Simon, N. Prunier-Leterme, L.

Dolatti, et al. 2004. Cross-species amplification of microsatellite loci

in aphids: assessment and application. Molecular Ecology Notes

4:104-09.

Yoshida, T., L. E. Jones, S. P. Ellner, G. F. Fussmann, and N. G. J.

Hairston. 2003. Rapid evolution drives ecological dynamics in a

predator-prey system. Nature 424:303-06.

Students also viewed