Embed Size (px)
Citation preview
The effects of biotic interactions on Ambrosia artemisiifolia L.
by
Arthur Andrew Meahan MacDonald
A Thesis submitted in conformity with the requirements
for the degree of Master of Science,
Graduate Department of Ecology and Evolutionary Biology,
in the University of Toronto
© Copyright by Andrew MacDonald 2009
ii
The effects of biotic interactions on Ambrosia artemisiifolia L.
Arthur Andrew Meahan MacDonald
Master’s of Science
Ecology and Evolutionary Biology
University of Toronto
2009
Abstract
Plant invasions can ensue when plants are introduced to regions without their specialist enemies (the
Enemy Release Hypothesis). This assumes natural enemies limit survival and fecundity in an invader's
native range. I tested this assumption for a native invasive species, Common Ragweed (Ambrosia
artemisiifolia), by excluding natural enemies from seeds, seedlings and adults. At the adult stage, I
added disturbance and conspecific density treatments. Protection from herbivores slightly improved
performance at the seedling stage only, while disturbance greatly increased survival and fecundity.
Increasing conspecific density reduced performance only in disturbed plots. I also tested herbivore
tolerance using simulated damage. Heavy (75%) damage did not reduce fecundity; light damage even
increased seed production. These results suggest enemies do not limit Ambrosia in its native range,
especially compared to the effects of habitat disturbance. While enemy release may have occurred
during Ambrosia’s invasions, it is not likely to be their principal cause.
iii
Acknowledgements
“What admirable training is science for the more active warfare of life. Indeed, the unchallenged
bravery, which these studies imply, is far more impressive than the trumpeted valor of the warrior.
Science is always brave, for to know, is to know good; doubt and danger quail before her eye. But
cowardice is unscientific; for there cannot be a science of ignorance.” --- Henry David Thoreau
This project has in one sense been about testing the limits of ragweed’s performance in various
ecological contexts. Besides testing the limits of this plant, I have also explored my own limits, and
perhaps sometimes the limits of the patience of those in my environment! The following people have
my gratitude, as I have learned a great deal from each of them.
First of all I must thank my supervisor Dr. Peter Kotanen. I was terrifically fortunate to end up in his
lab (for that, thanks also go to Sean Blaney and Dave McCorquodale). Peter has been precisely the
mentor I needed to transition from a naturalist trained at a small university to an empirical ecologist.
Also, I am very grateful to my supervisory committee, Dr. John Stinchcombe and Dr. Spencer Barrett
for their frank and candid advice. Among the wider faculty, I am indebted to Ivana Stehlik, Helene
Wagner and especially Monika Havelka, an inspiration.
I had great luck with labmates as well– James MacKay and Steve Hill are good friends and good
teachers. Steve in particular has been a statistics mentor of the first rank. Every error bar on the
following thesis would have been larger were it not for the efficient and competent help of Kateryna
Kostyukova. Our (hon.) labmate Megan Saunders helped keep the ecology culture alive in our lab.
Outside our lab there have been too many grad students to list easily. In particular I’d like to thank the
Gwynne Lab: Laura Robson, Kyla Ercit, Murray McConnell, Edyta Piascik and especially Kevin
Judge. These intelligent and friendly people kept up my morale and tolerated my frequent presence in
their Lab. Kevin in particular is a better friend and housemate than I could wish for; no opinion of
mine as been uninfluenced by his thoughtful lifestyle. I would also like to thank the on campus grad
community: Yessica Rico, Shekar Biswas, George Lamont, I-San Chan, April Wong, Mike Adorjan and
most especially Megan S., Steve Carradona (and Lumi and Chelcea). These wonderful folks have fed
me and kept me company throughout the cold winter months. Off-campus, Marion Andrew has
warmed my opinion of Toronto dramatically with her hospitality.
iv
Probably the best part of my education here were the summers spent at KSR. I am forever grateful to
the Kofflers for the opportunity. The staff (Ann Zimmerman, John Jensen, Paul, and (in 2008) Art
Weis) were always helpful and made working there a pleasure. Everyone who lived at Willow Ridge
can attest to the sense of community there and the fun and learning that is guaranteed when smart
people spend time in Nature. In particular: Anna Simonsen, Russel Dinnage, Brechann McGoy, Rob
Colautti, Emily Austen, Viet Pham and most of all Zuzana Burivalova and Ron McKenzie.
I want to briefly acknowledge some literary sources of support: Steven Covey, David Allen, Yamuna
Devi, H.D. Thoreau, E.O. Wilson, G. M. Hopkins.
I would not even be a scientist if it were not for a very supportive and loving family back in Cape
Breton. Thanks Mom, Dad, Brian and Marybeth for being there, all the way from giving me science
books in elementary school to helping with field work in 2009! Thanks to my entire extended family,
in particular Sr. Rita Clare. Also, thanks for all the friendship and love from all of Angela Petersen’s
extended family. Finally, to Angela Petersen herself, my constant source of love, encouragement and
support – its been twice as great and half as difficult. Thanks so much.
v
Table of Contents
Abstract ...................................................................................................................................ii
Acknowledgements ...............................................................................................................iii
List of Figures .......................................................................................................................vii
List of Tables........................................................................................................................viii
List of Tables........................................................................................................................viii
1 General Introduction .......................................................................................................1
1.1 Invasions.....................................................................................................................................................................1 1.1.1 Type of herbivore damage. .....................................................................................................................................3 1.1.2 Life history stage ....................................................................................................................................................4 1.1.3 Plant habitat ............................................................................................................................................................7
1.2 Native range studies are important..........................................................................................................................8
1.3 Study species ..............................................................................................................................................................8
1.4 Questions ..................................................................................................................................................................11
2 The effects of disturbance and enemy exclusion on performance of Ragweed (Ambrosia artemisiifolia L.) at three life-history stages. ...................................................23
2.1 Introduction .............................................................................................................................................................23
2.2 Methods....................................................................................................................................................................26 2.2.1 Site........................................................................................................................................................................26 2.2.2 Species..................................................................................................................................................................27 2.2.3 Experimental Design ............................................................................................................................................27 2.2.4 Statistical Analyses ...............................................................................................................................................31
2.3 Results ......................................................................................................................................................................32 2.3.1 Seed survival ........................................................................................................................................................32 2.3.2 Seedling survival ..................................................................................................................................................32 2.3.3 Adult: 2007...........................................................................................................................................................33 2.3.4 Adult: 2008...........................................................................................................................................................34
2.4 Discussion.................................................................................................................................................................35 2.4.1 Seed survival ........................................................................................................................................................35 2.4.2 Seedling survival ..................................................................................................................................................37 2.4.3 Adult survival and reproduction ...........................................................................................................................38
2.5 Figures ......................................................................................................................................................................42
vi
2.6 Tables........................................................................................................................................................................50
3 Leaf damage has weak effects on fecundity in Ambrosia artemisiifolia L. ..............59
3.1 Introduction .............................................................................................................................................................59
3.2 Methods....................................................................................................................................................................61 3.2.1 Site........................................................................................................................................................................61 3.2.2 Species..................................................................................................................................................................62 3.2.3 Experimental design .............................................................................................................................................62 3.2.4 Data collection......................................................................................................................................................63 3.2.5 Analyses ...............................................................................................................................................................64
3.3 Results ......................................................................................................................................................................66 3.3.1 i) Direct effects of damage ...................................................................................................................................66 3.3.2 ii) Size-dependent effects .....................................................................................................................................66
3.4 Discussion.................................................................................................................................................................67 3.4.1 Weak effects of damage on plant performance .....................................................................................................68 3.4.2 Leaf clipping changes the pattern of allocation to biomass ..................................................................................70 3.4.3 Tolerance effects on invasions..............................................................................................................................71 3.4.4 Tolerance and the ERH.........................................................................................................................................72
3.5 Conclusions ..............................................................................................................................................................72
3.6 Acknowledgments....................................................................................................................................................73
3.7 Figures ......................................................................................................................................................................74
3.8 Tables........................................................................................................................................................................82
4 General Discussion .......................................................................................................91
4.1 Summary of results .................................................................................................................................................91
4.2 Ambrosia and invasions ..........................................................................................................................................92 4.2.1 Ragweed in North America ..................................................................................................................................92 4.2.2 Ragweed’s Invasions ............................................................................................................................................93 4.2.3 Invasions in general ..............................................................................................................................................93
4.3 Suggestions for future work....................................................................................................................................94 4.3.1 Seeds.....................................................................................................................................................................94 4.3.2 Seedlings ..............................................................................................................................................................95 4.3.3 Mechanisms of tolerance and allocation...............................................................................................................95 4.3.4 Indirect effects of tolerance ..................................................................................................................................98
vii
List of Figures
Figure 2-1) Relationship between seed biomass and fecundity .............................................................. 42
Figure 2-2) Germinating seeds from samples of 10 seeds ...................................................................... 43
Figure 2-3) Seedling survivorship in response to herbivore exclusion treatments................................. 44
Figure 2-4) Damage to adult plants during the summer of 2007 ........................................................... 45
Figure 2-5) Damage to individual adult plants during the summer of 2008.. ........................................ 46
Figure 2-6) The incidence of stem-galling insects in 2008..................................................................... 47
Figure 2-7) Measures of plant performance in response to factorial field experiment in 2007. ............ 48
Figure 2-8) Measures of plant performance in response to factorial field experiment in 2008. ............ 49
Figure 3-1) Stem biomass in response to removal of leaves ................................................................... 74
Figure 3-2) Seed production in response to leaf removal ....................................................................... 75
Figure 3-3) The relationship between seed and plant biomass decreases with clipping damage ......... 76
Figure 3-4) Back-transformed elevations of the lines in Fig. 3-3 .......................................................... 77
Figure 3-5) The comparison of allometric reproduction between control and 75% damage................. 78
Figure 3-6) Seed mass in response to apical meristem clipping............................................................. 79
Figure 3-7) The relationship between seed and plant biomass not affected by apical removal. ........... 80
Figure 3-8) 95% confidence intervals for the regressions described in Fig 3-7 .................................... 81
Figure 4-1) A rough estimate of male flower mass ................................................................................ 99
viii
List of Tables
Table 2-1) A Table of the common enemies which occur on Ambrosia at our field site. ........................ 50
Table 2-2) ANOVA of enemy exclusion on seed germination.................................................................. 50
Table 2-3) ANOVA of enemy exclusion on seedling survival. ................................................................. 51
Table 2-4) Factorial ANOVA on the adult experiment in 2007.............................................................. 51
Table 2-5) Factorial ANOVA on the adult experiment in 2008............................................................... 52
Table 2-6) Proportion of plants surviving to reproduction in summer 2007 .......................................... 52
Table 2-7) Proportion of plants attacked by stem-galling insects in 2008. ............................................ 52
Table 3-1) Comparison of natural and simulated damage types ............................................................ 82
Table 3-2) ANOVA table for the linear model including plant size (stem biomass) as a continuous
variable and clipping as an ordered factor.. ........................................................................................... 83
Table 3-3)Parameter estimates (SE) for slopes and intercepts from the linear model of log(seed mass)
on log(stem mass). .................................................................................................................................. 83
1
1 General Introduction
1.1 Invasions
Human activities, particularly habitat alteration and the intentional and unintentional transport
of organisms, have caused rapid, unprecedented mixing of the Earth’s biota in a breakdown of what
Charles Elton (1958) called “Wallace’s Realms”: great biogeographic regions of distinctive, coadapted
organisms. While invasions do occur in nature, humans have accelerated that rate tremendously
(Vitousek et al. 1997) and are creating management crises that cost billions to control in the United
States (Pimentel et al. 2000, Mack et al. 2000) and in Canada (Colautti 2006). In particular, the
introduction of alien plant species has caused profound change in many now-threatened ecosystems
around the world with a concomitant loss of conservation, economic and aesthetic value. For example,
Bromus tectorum has altered the fire regime in North American grasslands and shrublands (Mack
1981), while Melaleuca quinquenervia has replaced entire areas of the Florida Everglades (Serbesoff-
King 2003). Invaded landscapes can be less economically productive: Leafy Spurge (Euphorbia esula)
prevents cow grazing in pastures (Leistritz et al. 1992) and Water Hyacinth (Eichhornia crassipes)
clogs fish habitat in the African Great Lakes making fishing difficult (Ogutu-Ohwayo et al. 1997).
Wind-pollinated introduced plants can cause allergies and pose a public health concern, as Common
Ragweed (Ambrosia artemisiifolia) has done in Europe (Igrc 1995). Despite decades of active research
(Callaway and Maron 2006), these invasions are likely to increase still more as the developing world
becomes more industrialized; for example, in China, biological invasions are rapidly increasing in
frequency (Ding et al. 2008, Weber and Li 2008).
Despite their negative consequences for the biosphere – indeed, often because of them –
invasions can be tools in our exploration of ecology and evolution (Kolar and Lodge 2001, Lodge
1993, reviewed in Callaway and Maron 2006). To this end they have been studied for a century and a
2
half (Darwin 1859, Elton 2000, Callaway and Maron 2006) and have furthered our understanding of
rapid evolution, community assembly and top-down control of plant populations. Invasion biology has
increased our understanding of population biology (Sakai et al. 2001), the phenomena of rapid
evolution (Blossey and Notzold 1995, Sakai et al. 2001) and the importance of biotic interactions in
determining plant distribution and abundance (Klironomos 2002, Mitchell et al. 2006). Theoretical
approaches to understanding invasions and the communities they invade has progressed from verbal
ideas (Elton 1958) to sophisticated modeling (e.g. Eppstein and Molofsky 2007) taking into account the
multiple interacting factors which together influence invasion.
Much research in Invasion biology is organized around the Enemy Release Hypothesis. The
Enemy Release Hypothesis (ERH) suggests that invaders benefit from a reduced consumer and/or
pathogen load in their new range; it is often invoked to explain the success of non-natives (Elton 1958,
Mitchell et al. 2006, Agrawal and Kotanen 2003, Keane and Crawley 2002, Mitchell and Power 2003,
Maron and Vila 2001, Wolfe 2002). This popular theory has generated a tremendous amount of
discussion and research since it was suggested by Charles Elton 50 years ago and continues to be
refined and explored by modern invasion ecologists (Richardson and Pysek 2008).
Liu and Stiling (2006) identify several assumptions of the ERH, most importantly that top-down
control is important for many plant species and that plants are likely to be introduced without their
oligophagous consumers. Because generalists are present in most habitats and are likely to be able to
feed on invaders (Liu and Stiling 2006, but see Jogesh et al. 2008), invaders continue to experience
some damage in their new habitat. However, because oligophagous consumers in the introduced range
are unlikely to recognize the invader as food, individual plants incur less tissue damage and so enjoy
enhanced probabilities of survival and reproduction. Additionally, invasive plants may be those able to
repel generalist attack as well: for example, in several recent studies (e.g. Jogesh et al. 2008,
3
Cappuccino and Carpenter 2005) exotics were found to be less preferred by native generalists.
Nevertheless the importance of Enemy Release as a cause of invasion is not totally clear;
studies continue to return equivocal results. Colautti et al (2004) found some of this variable evidence
to be linked to the type of study: comparisons of invasives in both their native and exotic range
supported the ERH, while comparisons of exotics and natives within the same community tended to
reject the ERH. In a quantitative review of ERH studies, Liu and Stiling (2006) found support for the
basic assumptions of ERH: introduced populations are attacked by fewer species of insects (in
particular, fewer specialists), and within a community natives receive more damage than invasives. Liu
and Stiling also conclude that studies across the life cycle are necessary to understand how all these
factors together determine plant performance and therefore, invasion success. While whole-lifecycle
studies are increasingly performed, we still have much to learn about the population consequences of
herbivory and its relevance to invasions (Maron and Crone 2006).
The importance of enemy release on the invasiveness of plant populations will depend on many
factors, including the degree of escape from herbivores, the plants’ defensive adaptations and the
habitat in which the herbivore-plant interaction occurs (Mitchell et al. 2006). Below is discussed
variation in the importance of enemies with reference to herbivore natural history, plant life stage when
attacked, the adaptations of plants to herbivory and finally, the abiotic environment wherein these
organisms are interacting.
1.1.1 Type of herbivore damage.
Kotanen and Rosenthal (2000) point out that there are many qualitatively different types of
damage inflicted by invertebrate herbivores, which may be tolerated by different mechanisms and to
different degrees. For example, damage to an apical meristem (i.e. as by some stem galling insects)
can result in the activation of auxiliary meristems, while leaf damage may stimulate an accelerated
4
growth rate (Tiffin 2000) . Long-term demographic studies sometimes reveal counterintuitive
consequences of the effects discussed above: e.g., no effects of conspicuous vegetative damage (Rose
et al. 2005) and large effects of light meristem damage (Doak 1992).
1.1.2 Life history stage
It is necessary to place plant-consumer interactions in the context of the life cycle in order to
understand and predict how plant populations will respond to enemy release (Halpern and Underwood
2006, Liu and Stiling 2006). Survival probability (for vegetative stages) and growth and reproduction
(for mature individuals) together determine the fitness of an individual and the growth rate of the
population (Caswell 2000, Morris and Doak 2002). If survival, growth or reproduction are depressed
by enemies, then there is potential for enemy release to result in a faster-growing – and perhaps more
invasive – population. While many plants have complex life cycles, most terrestrial plants can be
considered in three fundamental life history stages: seed (dispersing and dormancy), seedling (growth)
and adult (reproduction). The enemies which are likely to attack, and the consequences of the
interaction, are different among these three stages.
Seeds, which protect a plant embryo during dispersal and dormancy, provide the food needed
for germination and growth before photosynthesis begins (Baskin and Baskin 1980). As such, they are
readily eaten by other organisms, in particular mammals, birds, insects and fungi (Mitschunas et al.
2006). For example, in old fields in North America, small mammals (Hulme 1998) and ground beetles
(Coleoptera: Carabidae) (Honek et al. 2003) are common post-dispersal predators (Wall et al. 2005).
Importantly, seeds are frequently the life stage most easily introduced to new areas – for example, in
contaminated crop seed (Dessaint 2005). If the predictions of the ERH hold for seeds, then invasions
may be more likely to begin from such introductions.
Seedling mortality caused by pathogens and herbivores can be very high, a phenomenon noted
5
by Darwin in The Origin of Species (1859):
Seedlings, also, are destroyed in vast numbers by various enemies; for instance, on a
piece of ground three feet long and two wide, dug and cleared, and where there could
be no choking from other plants, I marked all the seedlings of our native weeds as
they came up, and out of the 357 no less than 295 were destroyed, chiefly by slugs
and insects.
Germinating plants are especially vulnerable while they acquire the resources for survival and
reproduction into the adult stage, making seedling survival an important demographic transition with
the potential to control population growth. Seedling survival is rarely studied on its own, and is usually
included in studies which examine the entire life cycle of a plant.
For example, Erhlén (2003) demonstrated that excluding slugs augmented population growth
across several years in Lathyrus vernus; these experimental treatments had consequences for all life-
history stages but the change in seedling survival was among the largest responses. When enemy
exclusion treatments which enhance seedling survival result in increased population growth, there is
the potential for the ERH mechanism to increase invasiveness following introduction.
While damage to seeds and seedlings is usually fatal, adult plants have more developed
defensive strategies which prevent or mitigate the effect of herbivores (Boege 2005). The defensive
adaptations of adult plants are of two kinds: resistance reduces the probability of damage, while
tolerance is the ability to incur that damage without a fitness cost (Kotanen and Rosenthal 2000). The
degree of resistance and tolerance that plants possess determines how strongly affected they will be by
herbivory – and therefore, how they will respond to release from herbivory. Both of these concepts are
central to our understanding of plant-herbivore relationships, and therefore to our understanding of the
ERH; both may be advantageous to invaders as their population adapts to a new set of biotic
interactions (Stastny et al. 2005).
In some cases resistance may increase the probability of invasion. For example, while
6
generalists are common in most habitats, invaders more resistant than native fauna may nevertheless
experience less herbivory from generalists (Jogesh et al. 2008, Cappuccino and Carpenter 2005).
However, resistance is not always advantageous: if costly resistance traits have evolved in response to
specialist damage (e.g. highly toxic secondary chemistry), introduced plants expend energy needlessly
on their production. In such cases, the invasiveness of a population will increase if plants evolve the
ability to divert resources away from costly (and now-useless) defenses, towards traits that are
advantageous in the introduced range . This hypothesized mechanism is called the Evolution of
Increased Competitive Ability Hypothesis, or EICAH (Blossey and Notzold 1995).
In contrast to resistance traits, tolerance adaptations maintain fitness after damage (Kotanen and
Rosenthal 2000, Weis and Franks 2006, Tiffin 2000). Qualitatively different damage types are
tolerated by different mechanisms and to different degrees; these responses are not necessarily
correlated (Kotanen and Rosenthal 2000). Interestingly, mechanisms of tolerance may also differ
between male and female fitness, with the effect that these likewise are uncorrelated (Strauss et al.
2003). The phenomenon of tolerance can very rarely take the form of overcompensation, wherein
herbivore-attacked plants seem to outproduce their undamaged neighbours. Agrawal (2000) suggests
that this is an evolved response to anticipated environmental conditions.
If introduced plants are more tolerant than members of the invaded community, they may incur
equal amounts of herbivory but suffer a lower relative fitness cost, giving them an advantage.
Additionally, they may possess an indirect advantage by augmenting herbivore populations, thereby
increasing the strength of apparent competition for the native community.
Tolerance adaptations may be frequent in invaders not because of direct or indirect benefits of
tolerance itself, but because tolerance adaptations correlate with other advantageous traits. Blumenthal
(2005, 2006) suggested that ERH may have an even greater effect when the introduced species is able
7
to grow quickly in response to nutrient-rich conditions. In such cases, plants are able to quickly take
advantage of invasion ‘windows’ of disturbed sites. Fast-growing, disturbance-adapted plants are often
highly tolerant of leaf damage (Coley et al. 1985), as such traits are often correlated with tolerance.
Indeed, rapid regrowth ability is a commonly-cited mechanism of tolerance (Tiffin 2000), and high
resource sites are more likely to contain tolerant plants (Hawkes and Sullivan 2001)
1.1.3 Plant habitat
In addition to changing across a plant’s life history, the strength of enemy interactions varies
among habitats. For example, DeWalt et al (2004) found that enemy-release in Clidemia hirta varied
between habitats in the introduced and native zones: the plant was excluded from forests in Costa Rica
by fungi and insects, but was able to invade Hawaiian forests in their absence. Chemical exclusion of
both fungi and insects increased survival of this invasive shrub by 41% in Costa Rican forests. Such
manipulative experiments are necessary to understand how enemy attack limits the size and distribution
of populations in the native range of a species, and how release from these enemies causes invasiveness
(Mitchell et al 2006).
Invasion success is predicted to vary across many such environmental variables (Moles et al.
2008), including disturbance and density. Frequent disturbance events free space in which invasive
species can establish; anthropogenic disturbance in particular can decrease a community’s resistance to
invasion (Mack 2000). In a review of herbivore effects on plant population size, Maron and Crone
(2006) found a positive relationship between disturbance frequency and the strength of herbivore-plant
interactions: in disturbed sites, populations are not limited by the distribution of suitable germination
sites and therefore any herbivory that lowers seed production is likely to have an effect. However,
although the absolute effect of herbivory is less in competitive environments, the combined effect of
herbivory and strong suppression by competitors can reduce population growth considerably, as
8
McEvoy and Coombes (1999) found for Senecio jacobaea. Frequent disturbance may also reduce the
density of conspecifics. In order for population processes to have limiting effects, they must be
density-dependent (Halpern and Underwood 2006). Demographic studies which manipulate
conspecific density (e.g. Gustaffson and Ehrlen 2003) can be used to understand how rates of
herbivory, growth and reproduction change as populations become more dense.
1.2 Native range studies are important
A fundamental assumption of the enemy release hypothesis is that enemies have a negative
effect on plant populations in the native range, and therefore introduction without these enemies causes
plants to become invasive (Liu and Stiling 2006). However, for many species – in particular those with
long-lived life history stages such as seed banks – enemies may not limit population size in the native
range (Maron and Vila 2001). The limiting effects of biotic interactions on an invader can be revealed
by factorial, manipulative studies in the native range of both species, where these interactions evolved
(Hierro et al. 2005, Guo 2006, Mitchell et al. 2006). Such studies allow the verification of a basic
assumption of the ERH: that enemies limit plant populations (Liu and Stiling 2006). As in the DeWalt
(2004) study outlined above, factorial manipulations of enemy density, habitat variables (e.g.
disturbance) and conspecific density can demonstrate that plant populations are limited in size and/or
restricted in distribution by enemy interactions in their native range (Mitchell et al. 2006). My
experiments focus on such native-range, factorial manipulations of a single species – Ambrosia
artemisiifolia L. – which has become invasive throughout Eurasia, where it has apparently escaped
herbivory (Genton et al. 2005b).
1.3 Study species
Common Ragweed, Ambrosia artemisiifolia L. (Asteraceae), is a weedy, annual species native
to North America. Ragweed is a common early-successional old-field species that can grow
9
aggressively in newly disturbed ground, forming up to 44% of the biomass (Teshler et al. 1981, Stevens
and Carson 1999). Its natural history is consistent with the classic ‘weedy traits’ described by Baker
(1965): it is a short-lived and rapidly-growing annual plant that produces abundant wind-dispersed
pollen; however, it differs from these stereotypical traits in being highly self-incompatible (Freidman
and Barrett 2008). Its copious pollen production is a major cause of hay fever; the pollen coat contains
potent allergens which are a principal cause of hay fever in humans (Bagarozzi and Travis 1998).
The huge quantities of pollen produced have created usable pollen records (Bassett and
Crompton 1975, Bassett and Terasmae 1962) which place the genus in North America for more than 60
000 years. Pollen records reveal its dominance in the Great Plains in the post glacial period (Grimm et
al 1993). Pollen records also have revealed it to be common in Native North American agriculture;
after European contact, sharp increases in ragweed pollen counts indicate the larger European
agricultural areas (Grimm et al. 2000). Ragweed has therefore been a crop weed for thousands of
years.
The weedy capacity of this species is further enhanced by its long-lived, tough achenes, borne
singly in female flowers. Seeds of Ambrosia can survive for decades in the seed bank: Toole and
Brown (1946) reported 22% germination of seeds buried at 42cm for 39 years, with much lower
germination at shallower depths. Ambrosia seeds germinate in late spring; cool temperature cycles
induce a light requirement which prevents earlier germination (Baskin and Baskin 1980). During
germination, Ambrosia forms strong AM mycorrhizal relationships with soil fungi (Fumanal et al.
2006, Koide and Li 1991, MacKay and Kotanen 2008) which dramatically increase growth.
The strong positive response of Ambrosia to disturbance extends beyond germination to growth
and reproduction. Plants grow quickly, often dominating newly-cleared land (Bazzaz and Mezga.
1973, Bazzaz 1974, Maryushkina 1991, Stevens and Carson 1999) and attaining densities of about 30
10
plants/m2 (Foster et al. 1980). Ragweed plants are able to moderate the relative proportions of male and
female flowers in response to heavy crowding (Paquin and Aarssen 2004). However, growth is
strongly suppressed in shaded conditions (Raynal and Bazzaz 1975) or root competition (Kosola and
Gross 1999). In fact, the competitive effect on ragweed is so dramatic that it has been suggested that
competitors be included in biological control efforts (Teshler 2002)
Invasion history
Ragweed is native to Canada (Bassett and Terasmae 1962, Bassett and Crompton 1975) but is
introduced to many other areas of the world, including Hawaii, Australia, China (Ding 2008, Weber
and Li 2008) and most problematically Eurasia (Igrc 1995). Throughout its introduced range, it can be
an aggressive crop weed in addition to an important source of hayfever-causing pollen.
By the early half of the 20th
Century Ragweed was widely introduced throughout France, but
was not abundant until after the second World War, perhaps as a result of the disturbance of large areas
of land (Allard 1943, Chauvel 2006). Most likely it is introduced in agricultural seed and animal
bedding, as was likely the case in France (Chauvel 2006) . Because the pathway between North
America and Europe was easily traversed for so long, the ragweed population in Europe is the result of
multiple introductions (Genton et al. 2005a).
Herbivory and defense
Herbivores attacking Ambrosia in North America include the specialists Zygogramma suturalis
(Coleoptera: Chrysomelidae, Reznik 1991) and several polyphagous consumers including beetles
(Systena blanda and Ophraella communa, Coleoptera: Chrysomelidae, see Gassman et al. 2006), a
sucking bug (Corythucha sp., Hemiptera: Tingidae) and caterpillars (Tarachidia spp. Lepidoptera:
Noctuidae), plus numerous generalists. While all of these enemies do damage to leaves, Ambrosia is
also attacked by a stem-galling insect, moths in the genus Epiblema. The chemistry of its defense
11
against herbivores is not well known, although sesquiterpene lactones have been characterized from
ragweed tissue (David et al. 1999). Ragweed has lost most of these enemies in colonizing Europe,
however reciprocal transplant experiments show no loss of defense: genotypes from Europe experience
damage equal to native plants (Genton et al 2005b).
This plant is a seemingly ideal candidate for 'classical biological control' because it has several
specialist herbivores and has reached high densities in disturbed sites in Europe (Sheppard et al. 2006).
However, biocontrol efforts have not been successful in Russia, Croatia and China (Igrc 1995). There
may be several reasons for the apparent failure of biocontrol, most likely a failure of pest populations
to establish large, stable populations. This may have been the case with European introductions of
biocontrol agents. Alternatively, biocontrol efforts may fail because plants are not sensitive to damage
caused by introduced enemies. In fact, leaf herbivory can stimulate Ambrosia’s growth (Throop 2005);
such compensatory growth is sometimes a mechanism of tolerance (Tiffin 2000).
1.4 Questions
The enemy release hypothesis makes the assumption that enemy attack in the native range is
severe enough to reduce plant populations in their native range. Testing this assumption is essential to
understanding whether the ERH is the mechanism for an invasion. In the two following chapters I ask
two related questions about the relationship between Ambrosia performance and enemy attack:
1) Is performance of Ambrosia reduced by insect damage? (Chapter 2)
This question is tested with a manipulative experiment in the native environment of Ambrosia,
crossing insect exclusion with a range of environmental conditions, particularly disturbance (important
for many weedy species) and conspecific density (essential for population limitation). In addition I
examine several life stages, as a major limitation of previous studies is their sole reliance on data from
adult plants.
12
2) Could performance of Ambrosia be reduced by insect damage? (Chapter 3)
I test this hypothesis with an artificial clipping experiment. Such manipulations are interesting
for two reasons: first, because natural levels of damage can be quite variable, and second, because
insects introduced for biocontrol often inflict more damage than they did in the native range. Clipping
experiments supply information about how plants may respond, in principle, to different intensities of
damage, including very high levels. The results of this experiment are described in Chapter Three.
Finally in Chapter 4 I present a summary of my results, synthesizing them with the invasion
literature and suggesting future experiments.
13
LITERATURE CITED
Agrawal A. A. 2000. Overcompensation of plants in response to herbivory and the by-product benefits
of mutualism. Trends in Plant Science 5:309-313.
Agrawal A. A., P. M. Kotanen. 2003. Herbivores and the success of exotic plants: a phylogenetically
controlled experiment. Ecology Letters 6:712-715.
Allard. 1943. The North American ragweeds and their occurance in other parts of the world. Science
98:292-294.
Bagarozzi D. A., J. Travis. 1998. Ragweed pollen proteolytic enzymes: Possible roles in allergies and
asthma. Phytochemistry 47:593-598.
Baker H. G. 1965. Chacteristics and modes of origin of weeds. Pages 147-172 In H.G Baker and G.L.
Stebbins, editors. Academic Press, New York, NY, .
Baskin J. M., C. C. Baskin. 1980. Ecophysiology of secondary dormancy in seeds of Ambrosia
artemisiifolia. Ecology 61:475-480.
Bassett I. J., C. W. Crompton. 1975. Biology of Canadian weeds 11. Ambrosia artemisiifolia L. and A.
psilostachya Dc. Canadian Journal Of Plant Science 55:463-476.
Bassett I. J., J. Terasmae. 1962. Ragweed, Ambrosia spp, in Canada and their history in post-glacial
time. Canadian Journal of Botany 40:141-150.
Bazzaz F. A. 1974. Ecophysiology of Ambrosia artemisiifolia - successional dominant. Ecology
55:112-119.
Bazzaz F. A., D. M. Mezga. 1973. Primary productivity and microenvironment in an Ambrosia
dominated old-field. American Midland Naturalist 90:70-78.
14
Blossey B., R. Notzold. 1995. Evolution of increased competitive ability in invasive nonindegenous
plants - a hypothesis. Journal of Ecology 83:887-889.
Blumenthal D. 2005. Interrelated causes of plant invasion. Science 310:243-244.
Blumenthal D. M. 2006. Interactions between resource availability and enemy release in plant invasion.
Ecology Letters 9:887-895.
Boege K., R. J. Marquis. 2005. Facing herbivory as you grow up: the ontogeny of resistance in plants.
Trends In Ecology and Evolution 20:441-448.
Callaway R. M., J. L. Maron. 2006. What have exotic plant invasions taught us over the past 20 years?
Trends In Ecology and Evolution 21:369-374.
Cappuccino N., D. Carpenter. 2005. Invasive exotic plants suffer less herbivory than non-invasive
exotic plants. Biology Letters 1:435-438.
Caswell, H. 2000. Matrix Population Models: Construction, Analysis, and Interpretation. Sinauer
Associates.
Chauvel, B., F. Dessaint, C. Cardinal-Legrand and F. Bretagnolle. 2006. The historical spread of
Ambrosia artemisiifolia L. in France from herbarium records. Journal of Biogeography 33:665-673
Colautti R. I., A. Ricciardi, I. A. Grigorovich, and H. J. MacIsaac. 2004. Is invasion success explained
by the enemy release hypothesis? Ecology Letters 7:721-733.
Colautti R. I., S. A. Bailey, C. D. A. van Overdijk, K. Amundsen, and H. J. MacIsaac. 2006.
Characterised and projected costs of nonindigenous species in Canada. Biological Invasions 8:45-59.
Coley, P. D., J.P. Bryant and F. S. Chapin. 1985. Resource availability and plant antiherbivore defense
Science 230: 895-899
15
David J. P., A. J. D. Santos, M. L. D. Guedes, J. M. David, H. B. Chai, J. M. Pezzuto, C. K.
Angerhofer, and G. A. Cordell. 1999. Sesquiterpene lactones from Ambrosia artemisiifolia
(Asteraceae). Pharmaceutical Biology 37:165-168.
Darwin, C. R. 1859. On the Origin of Species by Means of Natural Selection, or the Preservation of
Favoured Races in the Struggle for Life. London: John Murray. 1st edition, 1st issue.
Dessaint F., B. Chauvel, and F. Bretagnolle. 2005. Ragweed (Ambrosia artemisiifolia L.): expansion
history of a biological pollutant in France. Medecine Sciences 21:207-209.
DeWalt S. J., J. S. Denslow, and K. Ickes. 2004. Natural-enemy release facilitates habitat expansion of
the invasive tropical shrub Clidemia hirta. Ecology 85:471-483.
Ding J. Q., R. N. Mack, P. Lu, M. X. Ren, and H. W. Huang. 2008. China's booming economy is
sparking and accelerating biological invasions. Bioscience 58:317-324.
Doak D. F. 1992. Lifetime impacts of herbivory for a perennial plant. Ecology 73:2086-2099.
Ehrlen J. 2003. Fitness components versus total demographic effects: Evaluating herbivore impacts on
a perennial herb. American Naturalist 162:796-810.
Elton C. S. 1958. The Ecology of Invasions by Animals and Plants. University Of Chicago Press.
Eppstein M. J., J. Molofsky. 2007. Invasiveness in plant communities with feedbacks. Ecology Letters
10:253-263.
Foster M. M., P. M. Vitousek, and P. A. Randolph. 1980. The effects of ragweed (Ambrosia
artemisiifolia L.) on nutrient cycling in a 1st-year old-field. American Midland Naturalist 103:106-
113.
Friedman J., S. C. H. Barrett. 2008. High outcrossing in the annual colonizing species Ambrosia
16
artemisiifolia (Asteraceae). Annals of botany 101:1303-1309.
Fumanal B., C. Plenchette, B. Chauvel, and F. Bretagnolle. 2006. Which role can arbuscular
mycorrhizal fungi play in the facilitation of Ambrosia artemisiifolia L. invasion in France?
Mycorrhiza 17:25-35.
Genton B. J., J. A. Shykoff, and T. Giraud. 2005a. High genetic diversity in French invasive
populations of common ragweed, Ambrosia artemisiifolia, as a result of multiple sources of
introduction. Molecular ecology 14:4275-4285.
Genton B. J., P. M. Kotanen, P. O. Cheptou, C. Adolphe, and J. A. Shykoff. 2005b. Enemy release but
no evolutionary loss of defence in a plant invasion: an inter-continental reciprocal transplant
experiment. Oecologia 146:404-414.
Grimm E. C., G. L. Jacobson, W. A. Watts, and B. C. S. H. a. K. A. Maasch. 1993. A 50,000-Year
record of climate oscillations from Florida and its temporal correlation with the Heinrich Events.
Science 261:198-200.
Grimm, E.C. 2001. Trends and palaeoecological problems in the vegetation and climate history of the
northern Great Plains, USA. Biology and Environment: Proceedings of the Royal Irish Academy
101:47-64
Guo Q. 2006. Intercontinental biotic invasions: what can we learn from native populations and
habitats? Biological Invasions 8:1451-1459.
Gustafsson, C. and J. Ehrlen. 2003. Effects of intraspecific and interspecific density on the demography
of a perennial herb, Sanicula europaea. Oikos 100: 317-324
Halpern S. L., N. Underwood. 2006. Approaches for testing herbivore effects on plant population
dynamics. Journal of Applied Ecology 43:922-929.
17
Hawkes C. V., J. J. Sullivan. 2001. The impact of herbivory on plants in different resource conditions:
A meta-analysis. Ecology 82:2045-2058.
Hellström K., P. Rautio, A. Huhta, and J. Tuomi. 2004. Tolerance of an annual hemiparasite, Euphrasia
stricta agg., to simulated grazing in relation to the host environment. Flora - Morphology,
Distribution, Functional Ecology of Plants 199:247-255.
Hierro J. L., J. L. Maron, and R. M. Callaway. 2005. A biogeographical approach to plant invasions: the
importance of studying exotics in their introduced and native range. Journal of Ecology 93:5-15.
Honek A., Z. Martinkova, P. Saska, and S. Pekar. 2007. Size and taxonomic constraints determine the
seed preferences of Carabidae (Coleoptera). Basic and Applied Ecology 8:343-353.
Hulme P. E. 1998. Post-dispersal seed predation: consequences for plant demography and evolution.
Perspectives in Plant Ecology, Evolution and Systematics 1:32-46.
Igrc J., C. J. Deloach, and V. Zlof. 1995. Release and establishment of Zygogramma suturalis F
(Coleoptera, Chrysomelidae) in Croatia for control of common ragweed (Ambrosia artemisiifolia L).
Biological Control 5:203-208.
Jogesh T., D. Carpenter, and N. Cappuccino. 2008. Herbivory on invasive exotic plants and their non-
invasive relatives. Biological Invasions 10:797-804.
Keane R. M., M. J. Crawley. 2002. Exotic plant invasions and the enemy release hypothesis. Trends in
Ecology and Evolution 17:164-170.
Klironomos J. N. 2002. Feedback with soil biota contributes to plant rarity and invasiveness in
communities. Nature 417:67-70.
Kolar C. S., D. M. Lodge. 2001. Progress in invasion biology: predicting invaders. Trends in Ecology
and Evolution 16:199-204.
18
Koide, R. T. and M. Li. 1991. Mycorrhizal fungi and the nutrient ecology of three oldfield annual plant
species. Oecologia 85: 403-412
Kosola K. R., K. L. Gross. 1999. Resource competition and suppression of plants colonizing early
successional old fields. Oecologia 118:69-75.
Kotanen P. M., J. P. Rosenthal. 2000. Tolerating herbivory: does the plant care if the herbivore has a
backbone? Evolutionary Ecology 14:537-549.
Leistritz F. L., F. Thompson, and J. A. Leitch. 1992. Economic impact of leafy spurge (Euphorbia
esula) in North Dakota. Weed Science 40:275-280.
Liu H., P. Stiling. 2006. Testing the enemy release hypothesis: a review and meta-analysis. Biological
Invasions 8:1535-1545.
Lodge D. M. 1993. Biological invasions: Lessons for ecology. Trends in Ecology and Evolution 8:133-
137.
Mack R. N. 1981. Invasion of Bromus tectorum L. into western North America: an ecological
chronicle. Agro-Ecosystems 7:145-165.
Mack R. N., D. Simberloff, W. M. Lonsdale, Evans, H. and Clout,M., and F. A. Bazzaz. 2000. Biotic
invasions: Causes, epidemiology, global consequences, and control. Ecological Applications 10:689-
710.
MacKay J., P. M. Kotanen. 2008. Local escape of an invasive plant, common ragweed (Ambrosia
artemisiifolia L.), from above-ground and below-ground enemies in its native area. Journal of
Ecology 96:1152-1161.
Maron J. L., E. Crone. 2006. Herbivory: effects on plant abundance, distribution and population
growth. Proceedings Of The Royal Society B-Biological Sciences 273:2575-2584.
19
Maron J. L., M. Vila. 2001. When do herbivores affect plant invasion? Evidence for the natural
enemies and biotic resistance hypotheses. Oikos 95:361-373.
Maryushkina V. Y. 1991. Peculiarities of Common Ragweed (Ambrosia artemisiifolia L.) strategy.
Agriculture Ecosystems and Environment 36:207-216.
McEvoy, P. B. & Coombs, E. M. 1999 Biological control of plant invaders: regional patterns, field
experiments, and structured population models. Ecological Applications 9, 387–401.
Mitchell C. E., A. A. Agrawal, J. D. Bever, G. S. Gilbert, R. A. Hufbauer, J. N. Klironomos, Maron, J.
L. and Morris,W.F., I. M. Parker, A. G. Power, E. W. Seabloom, M. E. Torchin, and D. P. Vazquez.
2006. Biotic interactions and plant invasions. Ecology Letters 9:726-740.
Mitschunas N., M. Wagner, and J. Filser. 2006. Evidence for a positive influence of fungivorous soil
invertebrates on the seed bank persistence of grassland species. Journal of Ecology 94:791-800.
Moles, A. T., M. A. M. Gruber and S.P. Bonser. 2008. A new framework for predicting invasive plant
species. Journal of Ecology 96:13-17
Morris, W. F. and Doak, D. F. 2002. Quantitative Conservation Biology: Theory and Practice of
Population Viability Analysis. Sinauer Associates.
Ogutu-Ohwayo R., R. E. Hecky, A. S. Cohen, and L. Kaufman. 1997. Human impacts on the African
Great Lakes. Environmental Biology of Fishes 50:117-131.
Paquin V., L. W. Aarssen. 2004. Allometric gender allocation in Ambrosia artemisiifolia (Asteraceae)
has adaptive plasticity. American Journal of Botany 91:430-438.
Pimentel D., L. Lach, R. Zuniga, and D. Morrison. 2000. Environmental and economic costs of
nonindigenous species in the United States. Bioscience 50:53-65.
20
Raynal, D. J. and F.A. Bazzaz. 1975. Interference of winter annuals with Ambrosia artemisiifolia in
early successional fields. Ecology 56: 35-49
Rejmanek M., D. M. Richardson. 1996. What attributes make some plant species more invasive?
Ecology 77:1655-1661.
Richardson, D. M. and Pysek, P. 2008. Fifty years of invasion ecology - the legacy of Charles Elton.
Diversity And Distributions 14: 161-168
Rose K. E., S. M. Louda, and M. Rees. 2005. Demographic and evolutionary impacts of native and
invasive insect herbivores on Cirsium canescens. Ecology 86:453-465.
Sakai A. K., F. W. Allendorf, J. S. Holt, Lodge, D. M. and Molofsky,J., K. A. With, S. Baughman,
Cabin, R. J. and Cohen,J.E., N. C. Ellstrand, D. E. McCauley, P. O'Neil, I. M. Parker, J. N.
Thompson, and S. G. Weller. 2001. The population biology of invasive species. Annual Review of
Ecology and Systematics 32:305-332.
Serbesoff-King K. 2003. Melaleuca in Florida: a literature review on the taxonomy, distribution,
biology, ecology, economic importance and control measures. Journal of Aquatic Plant Management
41:98–112.
Sheppard, A., R. Shaw, and R.Sforza, R. 2006. Top 20 environmental weeds for classical biological
control in Europe: a review of opportunities, regulations and other barriers to adoption. Weed
Research 46: 93-117
Stastny M., U. Schaffner, and E. Elle. 2005. Do vigour of introduced populations and escape from
specialist herbivores contribute to invasiveness? Journal of Ecology 93:27-37.
Stevens M. H. H., W. P. Carson. 1999. Plant density determines species richness along an experimental
fertility gradient. Ecology 80:455-465.
21
Strauss S. Y., W. Watson, and M. T. Allen. 2003. Predictors of male and female tolerance to insect
herbivory in Raphanus raphanistrum. Ecology 84:2074-2082.
Teshler M. P., A. DiTommaso, J. A. Gagnon, and A. K. Watson. 1981. Biological control programmes
against insects and weeds in Canada 1981-2000. Pages 290-294 In P. G. Mason and J. T. Huber,
editors. CABI Publishing, .
Throop, H. L. 2005. Nitrogen deposition and herbivory affect biomass production and allocation in an
annual plant. Oikos 111: 91-100
Tiffin P. 2000. Are tolerance, avoidance, and antibiosis evolutionarily and ecologically equivalent
responses of plants to herbivores? The American Naturalist 155:128-138.
Toole, E.H., and E. Brown. 1946. Final results of the Duvel buried seed experiment. Journal of
Agricultural Research 72:201-210
Vitousek P. M., H. A. Mooney, J. Lubchenco, and J. M. Melillo. 1997. Human domination of Earth's
ecosystems. Science 277:494-499.
Wall S. B. V., K. M. Kuhn, and M. J. Beck. 2005. Seed removal, seed predation, and secondary
dispersal. Ecology 86:801-806.
Weber E., B. Li. 2008. Plant invasions in China: what is to be expected in the wake of economic
development. Bioscience 58:437-444.
Weis A. E., S. J. Franks. 2006. Herbivory tolerance and coevolution: an alternative to the arms race?
New Phytologist 170:423-425.
Wise M. J., W. G. Abrahamson, and K. Landis. 2006. Edaphic environment, gall midges, and goldenrod
clonal expansion in a mid-successional old-field. Acta Oecologica 30:365–373.
22
Wolfe L. M. 2002. Why alien invaders succeed: support for the escape-from-enemy hypothesis. The
American Naturalist 160:705-711.
23
2 The effects of disturbance and enemy exclusion on performance of Ragweed (Ambrosia artemisiifolia L.) at three life-history stages.
2.1 Introduction
Plant invasions have occurred throughout the biosphere (Elton 1958, Mack et al. 2000) and are
major contributors to a contemporary biotic homogenization of unprecedented speed and scale
(Vitousek et al. 1997). Management procedures to control plant invasions cost billions in the US
(Pimentel et al. 2000, Mack et al. 2000) and Canada (Colautti et al. 2006). Despite such efforts, the
frequency of plant invasions is likely to increase globally as the developing world becomes more
industrialized, increasing the incidence of habitat disturbance and import of propagules from elsewhere
(e.g. China, Ding et al. 2008).
A principal explanation for biotic invasion is the Enemy Release Hypothesis (herein ERH, see
Keane and Crawley 2002, Elton 1958), which suggests that introduced plants are able to become
invasive because their herbivores and pathogens are absent in their new range. Freed from the costs of
tissue loss to enemies and in some cases having reallocated resources from expensive defence traits,
invasive plants may enjoy increased reproduction and vigour. This augmented performance may give
introduced plants the advantage over other members of their new community.
ERH remains controversial: in particular, it is unclear to what extent invaders are limited by
enemies within their home range, and therefore whether escaping enemies during invasion provides a
significant advantage (Maron and Vila 2001). The limiting effects of biotic interactions on an invader
can be revealed by a study of their natural history in their native range, where these interactions
evolved (Hierro et al. 2005, Guo 2006). A particularly effective approach includes the direct
manipulation of enemies and other relevant habitat variables in factorial field experiments (Mitchell et
al. 2006). The utility and generality of such manipulative approaches can be improved by including
24
three variables. First, density should be considered, as only density-dependent processes are capable of
limiting population size (Halpern and Underwood 2006). Second, studies of enemy limitation should
study as many life-history stages as possible, since interactions which limit a population’s growth can
occur at several stages (Liu and Stiling 2006, Maron and Crone 2006). Finally, disturbance often
increases the probability of invasion, and so it is important to understand a species’ adaptation to
disturbance, and the interactions between damage and herbivore damage (Mack et al. 2000, Blumenthal
2006, Cadotte et al. 2006).
Studying multiple life stages of plants is an underemployed, yet very powerful, means of
understanding the relationship between plants and their enemies. Studies of invasions rarely observe
the effects of enemies across the life cycle, focussing instead on a single life stages or interactions (Liu
and Stiling 2006). For example, studies have examined rates of rodent predation on seeds (e.g, Hulme
1998, Manson and Stiles 1998), fungal mortality on seeds (e.g. Blaney and Kotanen 2001) and most
frequently leaf damage in adult plants (see reviews by Liu and Stiling 2006). Observations of enemy
effects at all life stages is necessary to understand the long-term consequences for a population
(Halpern and Underwood 2006, Maron and Crone 2006, Mitchell et al. 2006); these results are not
always obvious from an observation of individuals in one year. For example, using an 11-year dataset
(Rose et al. 2005) found that vegetative damage to Cirsium canescens Nutt. had no consequences for
population growth, while damage to the flower heads had a sufficiently strong effect to act as
population control.
Interactions with enemies are not the only factors relevant to plant invasions: species which are
adapted to frequent disturbance within their native range also may possess an advantage over members
of the introduced community. The potential for competitive interactions within the community to
exclude invaders is termed ‘biotic resistance’ and can be a strong impediment against an invader's
25
population growth (Parker et al. 2006). Disturbance often permits plant species to invade habitats,
allowing them to avoid competition with plants in the resident community (Blumenthal 2006, Cadotte
et al. 2006, Blumenthal 2005). Herbivory and disturbance can interact: a review by Maron and Crone
(2006) found that the effect size of herbivory was greatest in disturbed, competition-free habitats –
however, the strongest population limitation occurred when competition and herbivory were both
present. For example, a factorial experiment by McEvoy et al. (1993) found that Senecio jacobaea
populations were reduced by both heterospecific competition and herbivory by the ragwort flea beetle.
Blumenthal (2005, 2006) suggests that the converse is true in biological invasions: escape from
enemies coupled with resource-rich, disturbed habitat could together stimulate an even greater effect
than either alone. Thus the effect of disturbance on plant performance, and especially plant responses to
herbivore attack, is particularly important in understanding biological invasions.
Common (or Annual) ragweed (Ambrosia artemisiifolia L.) is well-suited to address these
questions. This plant is native to North America and has been widely introduced throughout Eurasia
(Kiss 2007) and is abundant in parts of both eastern and western Europe. Within this introduced range,
Ambrosia is much less attacked by enemies; however, trans-Atlantic transplant experiments showed no
evidence of an evolutionary loss of defence (Genton et al. 2005b). Importantly, this suggests that
native-range plants remain relevant models for assessing herbivore impacts on European populations as
well. In both its native and invaded range, ragweed is also highly disturbance-dependent (Bazzaz and
Mezga 1973, Allard 1943), suggesting high susceptibility to biotic resistance by competitors.
Populations of ragweed are able to grow to high density when sites are cleared. As well, enemy attack
is known to increase as in ragweed plants growing closer to established populations (MacKay and
Kotanen 2008). However, it is not known how populations of ragweed respond to natural enemies;
such knowledge would help us understand whether Ambrosia invasion in Europe is in part caused by
26
enemy release, and inform control efforts on both sides of the Atlantic.
To understand the importance of biotic interactions to Ambrosia within its native range, we
excluded enemies from seeds, seedlings and adult plants in a series of field experiments. We also
manipulated disturbance, and intraspecific density and insect presence/absence to look for effects on
reproduction, growth and survival. Specifically, we hypothesize that enemies lower survival and/or
reproduction a three major life history stages, each with a different group of important enemies:
1. Seeds: fungal pathogens and mammals.
2. Seedlings: slugs and insects.
3. Adult plants: disturbance, herbivores, and conspecific density
At each life history stage we predicted that enemy exclusion would increase survivorship. At
the adult stage (3) we included disturbance and density as additional factors, and included growth and
reproduction as response variables, in recognition of the importance of competition for this species.
Increases in disturbance, decreases in density, and the absence of insects were all predicted to increase
the performance of adult plants.
2.2 Methods
2.2.1 Site
These experiments were conducted at the Koffler Scientific Reserve (KSR) at Joker's Hill, a
350-ha field station owned by the University of Toronto and situated 50km north of Toronto, Ontario
(44 03'N, 79 29'W). A complete plant species list for KSR is provided on the reserve's webpage
(http://www.ksr.utoronto.ca/). Ragweed is common at the site in general and at our experimental sites
in particular, emerging from the seed bank once our experimental site was cleared.
27
2.2.2 Species
Common Ragweed, Ambrosia artemisiifolia L. (Asteraceae), is a weedy, annual species native
to North America. Ambrosia can grow aggressively in newly disturbed ground, forming up to 44% of
the biomass in the first year after clearing (Teshler 1981, Stevens and Carson 1999). Its natural history
is consistent with the classic ‘weedy traits’ described by Baker (1965): it is a short-lived and rapidly-
growing annual plant that produces abundant wind-dispersed pollen. However, it departs from these
stereotypical traits in being highly self-incompatible (Freidman and Barrett 2008). Its copious pollen
production is a major cause of hay fever; the pollen coat contains potent allergens which are a principal
cause of hay fever in humans (Bagarozzi and Travis 1998). Ambrosia is highly disturbance-dependent;
its growth is strongly suppressed in shaded conditions (Raynal and Bazzaz 1975) or root competition
(Kosola and Gross 1999). In fact, the competitive effect on ragweed is so dramatic that it has been
suggested that competitors be included in biological control efforts (Teshler 2002)
Ambrosia is a problematic plant on both sides of the Atlantic. In its native North America, it
was a problematic crop weed in Native agriculture (Grimm et al. 1993) and continues to be a serious
crop weed (Bassett and Crompton 1975). In Eurasia, particularly France, ragweed is introduced to
agricultural settings and has become invasive (Kiss 2007, Igrc et al. 1995). Data from France indicate
that this plant has escaped enemies during this invasion (Genton et al. 2005b), most of the enemies
which eat it are North American (Table 2-1). It appears that enemy release may have accompanied
invasion, yet enemy effects on plant populations have never been demonstrated for this species. This
knowledge is required for understanding invasion by Ambrosia.
2.2.3 Experimental Design
1) Seed survival
The effects of vertebrate (bird and mammal) granivory and fungal pathogens on seed mortality
28
were assessed in a seed survival experiment. We established this experiment in early November 2007,
just after natural ragweed plants had dispersed seeds, using a mixture of seeds collected in 2005 from 4
mother plants from each of 10 populations in Southern Ontario. We ploughed nine blocks (approx. 4m
x 20m), and within each we marked nine small plots (approx. 10cm x 10cm, in a 3x3 grid) separated
from one another and the block’s edge by a 1m buffer. Within each plot we scattered 10 seeds on the
ground surface and randomly assigned one of three nested treatments: control (no treatment), cage
protection, and cage protection plus fungicide addition. The cages had a mesh size of approx. 1cm2 and
were meant to exclude large (vertebrate) granivores. Fungal seed pathogens were controlled by soaking
seeds and surrounding soil with a 10g/L solution of Captan (as recommended by product directions; see
MacKay and Kotanen 2009). Captan is a heterocyclic phthalimide fungicide (Zeneca Agro, Stoney
Creek, ON, Canada) and is very effective against fungi in the Oomycetes, Ascomycetes, and
Basidiomycetes (Sharvelle 1961, Torgeson 1969, Neergaard 1977), previous experiments at KSR have
successfully used Captan to exclude fungal pathogens from seeds (e.g. Blaney and Kotanen 2001,
MacKay and Kotanen 2008, Schafer and Kotanen 2003). In May and June 2008, emerging seedlings
were marked and censused weekly; we counted seedlings until the end of June, when new seedlings
had stopped appearing. Our response variable is germinated seeds, making inference about the fate of
ungerminated seeds difficult (Maron and Crone 2006); such seeds could have been moved by
earthworm activity, washed away, died or have re-entered secondary dormancy (Baskin and Baskin
1980).
2) Seedling survival
The effects of attack by slugs and other herbivores on seedling survival were tested in a
seedling survival experiment in spring 2008. In early June we transplanted wild ragweed seedlings
(approx 2-4cm high, with at least 2 pairs of adult leaves) from a population at KSR to five blocks
29
around the site; the same locations used for the seed survival experiment (described above). In each
block we planted 10 rows of 3 plots (approx. 10cm diameter) with ten ragweed seedlings per plot (300
plants/block). Three nested treatments were then randomly applied within rows: control (no enemy
exclusion), aluminum barrier plus slug pellets (slug exclusion), and aluminum barrier, slug pellets, plus
a covering of 1mm2 nylon netting (herbivore exclusion). We used about 60ml of slug pellets,
containing 2.75% metaldehyde, which kills slugs and snails when ingested (Spectrum brands IP inc,
Brandfort, On). Aluminum barriers were made of a sheet of aluminum flashing and are effective
barriers to slugs with added protection from the poison in the slug pellets (Schüder et al 2003). Netting
excluded all but the smallest herbivores, including most insects. We counted the seedlings throughout
the seedling stage to determine the rate of seedling mortality from these enemies.
3) Adult survival and reproduction
This experiment was conducted twice at KSR, once in 2007 and again in 2008. In both years we
used seeds from the same populations as the seed survival experiment. The seeds were stratified in
moist soil at approx. 5°C for one week, before being transferred to cell packs filled with standard, non-
sterilized potting soil and germinated in the greenhouse on the campus of the University of Toronto at
Mississauga. When experimental plants had developed two pairs of adult leaves and were 6-10 cm
high, (approximately the same size as wild plants) they were brought to the field site and allowed to
acclimatize in a sheltered location for at least 3 nights.
In both years, 5 experimental blocks were marked within old field vegetation, each containing
12 1m2 plots separated by about 2m. Blocks were at least 30m apart. Each plot within a block was
randomly subjected to one of 8 factorial combinations of three treatments: a conspecific density
treatment (2 levels: 9 and 25 plants/m2), an insect exclusion treatment (2 levels: pesticide or water
sprayed), and a disturbance treatment (2 levels: plots cleared of existing vegetation or not cleared).
30
Plots were disturbed by completely hand-removing existing sod. In the conspecific density
treatments, plants were laid out in a grid with approximately equal distances between each other and
from the edge of the plot. Finally, in 2007, insects were excluded by spraying the upper and lower
surfaces of all leaves with a weekly spray of the insecticidal plant extract Rotenone (5% Rotenone,
30ml in 3L water, Wilson Laboratories, Dundas, ON). In 2008 we followed the same procedure, but
used the organophosphate pesticide Malathion (50% Malathion, 25ml in 3L water, Wilson
Laboratories, ON). We changed the pesticide treatment because Rotenone proved insufficiently
effective in 2007 (see Results). In both years control plots were sprayed with an equivalent amount of
water.
2007
The amount of insect damage (proportion of all leaves with visible damage, proportion of stems
attacked by stem borers) and two measures of plant size (plant height and leaf number), were measured
at the beginning of the experiment (18 June) and throughout the growing season (July and August) on
all plants. At the end of the summer (September), stem biomass and fecundity were measured on a
random sample of three plants. We estimated fecundity by counting all seeds produced by an
individual.
2008
The amount of insect damage (proportion of all leaves with visible damage, proportion of stems
attacked by stem borers) and two measures of plant size (plant height and leaf number), were measured
at the beginning of the experiment (18 June) and throughout the growing season (July and August) on
six randomly chosen focal individuals per plot. Damage was most likely caused by the insects in Table
2-1, we did not identify different damage types. At the end of the summer (September), we measured
stem biomass and fecundity of focal plants. We estimated fecundity by weighing all seeds produced by
31
an individual, based on a regression of 38 plants from one of our experimental blocks. Weight is an
excellent predictor of seed number (linear regression: F1,36=947.94, p<<0.001, r2=0.96, both variables
log transformed, Figure 2-1).
2.2.4 Statistical Analyses
Seed and seedling survival (experiments 1 and 2)
We tested for the effects of enemy exclusion on the number of germinated seeds (1) and
surviving seedlings (2) with randomized-block ANOVAs. Because both experiments began with 10
seeds or seedlings, respectively, analyzing these data as proportions is unnecessary. We first calculated
the means of all within-block replicates (3 for the seed experiment, 10 for the seedling) and performed
our analysis on these. Block means were log-transformed to normalize residual variance, and block
was treated as a random effect in our analyses. In each experiment, we tested the a priori hypothesis
that a more complete exclusion of enemies would increase survival. Specifically, we hypothesized the
following ranks of magnitude in our response variables for the seed and seedling experiments,
respectively: Control<cage<cage & fungicide (seeds germinating), Control<barrier<barrier & net
(seedlings surviving). The analysis of ordered factors was conducted with orthogonal polynomial
contrasts (Sokal and Rohlf 1994, Quinn and Keough 2002).
Adult plants (experiment 3)
We tested for treatment differences in our measured variables with randomized block factorial
ANOVAs. All statistical models were mixed, including the three treatments (disturbance, density and
pesticide spray) as fixed effects and block as random.
In both years, our measurements of plant size (stem weight and leaf number), and fecundity
(seed number in 2007, seed mass in 2008) were measured for multiple plants within each plot (see
32
experimental design, above). We then calculated plot means and performed the analysis on these.
In contrast, survivorship and stem gall incidence were measured as proportions of plants within
each plot. We analyzed these data with Generalized Linear Mixed Models (GLMM, Bolker 2008,
Crawley 2007), using the same structure of fixed and random effects as described above for the GLMM
linear predictor. To accommodate proportion data, the GLMM uses a logit link and assumes binomial
errors. The significance of terms was determined by simplifying the model and checking for a loss of
explanatory power with a Likelihood-ratio test. For example, the 3-way interaction was removed and
this simplified model was compared to the full; this term was left out if the change in deviance was not
significant at the p<0.05 level. Note that 2-way interactions were not deleted sequentially but were
replaced after testing, so that each interaction was tested against a model containing all other 2-way
interactions. This procedure is described in Crawley (2007). All analyses were performed in R 2.8.1
(R Development Core Team 2008).
2.3 Results
2.3.1 Seed survival
Of the nine experimental blocks that were set up in November 2007, three were damaged over
the winter, making it impossible to distinguish treatments; analyses are limited to the six remaining
blocks wherein treatments could be identified clearly. About 50% of seeds germinated in situ (out of
10 seeds, 4.78±0.5, mean ± SE). Protection from both herbivores and fungi had no effect: by the end of
June there were no significant differences in the number of germinated seeds among treatments (F2,10 =
0.354, p=0.71, Figure 2-2,Table 2-2).
2.3.2 Seedling survival
Seedling survivorship was very high; between early and late June an average of 7.8 out of 10
33
seedlings survived in each plot (Figure 2-3). Our a priori hypothesis of increasing survivorship with
treatment rank was supported (Linear contrast in randomized block ANOVA: F1,8=6.25, p=0.04,).
Increasing levels of enemy exclusion (control, slug, slug and insect) had increasingly greater seedling
survival: 7.14±0.95, 7.76±0.56 and 8.6±0.44, respectively. Excluding slugs therefore creates a 8.7%
increase in survivorship; excluding insects and slugs a further 10.8%.
2.3.3 Adult: 2007
In 2007, the proportion of leaves damaged was high: on average 81.9% of leaves in the control
treatment were damaged. Our pesticide treatment was weakly effective, reducing damage by 16.7%
(F1,28 = 6.35, p=0.02,Table 2-4, Figure 2-4). There was no clear pattern of increasing damage with
density or disturbance (p=0.28, F1,28=1.25; p=0.97, F1,28<0.01 respectively), although the data suggest
that high-density, undisturbed plots receive more damage than high-density, disturbed plots.
Survivorship over the summer was clearly enhanced by disturbance, with 93.5% of plants
surviving in disturbed sites and 42.6% in undisturbed. Model simplification on a GLMM demonstrated
that density (p=0.45) and pesticide (p=0.72) had no effect (Table 2-6). However, there was a slight
trend for survivorship to decrease with increasing density in disturbed plots (Figure 2-7).
Plant growth and fecundity were also sensitive to disturbance and additionally showed a strong
interaction with density, but no effect of pesticide. Our measures of plant growth (number of leaves
and stem biomass) were strongly increased by disturbance (F1,28=1109.8, p<<0.01 and F1,28=62.18,
p<0.01, respectively). Increasing density reduced average leaf number by 38.9% and stem biomass by
67%, but only in the disturbed treatment (F1,28=5.72, p=0.03 and F1,28=16.61, p<<0.01, respectively).
There was no significant effect of pesticide on either variable (Leaf: F1,28=2.95, p=0.10; Stem:
F1,28<<0.01, p=0.99).
Pesticide, while effectively reducing leaf damage, had no effect on fecundity (F1,28=0.72,
34
p=0.40). We did see strong differences in response to disturbance (F1,28=64.87, p<<0.01), while
increasing density lowered seed production by 74% in disturbed plots only.
2.3.4 Adult: 2008
The following summer damage rates were much lower: on average, 33.4% of leaves per plant
were damaged in control plots. Malathion was very effective, reducing this damage level by 72%
(F1,28=84.62, p<<0.01, Table 2-5, Figure 2-8). However, in contrast to 2007, we found significant
interactions of pesticide application with both disturbance and density. Disturbance lowered damage in
both pesticide and control plots, but had a proportionally greater effect on pesticide-sprayed plots.
Denser plots experienced 31% less damage per plant in control plots only; density had no effect on
herbivory when pesticide was applied.
Survivorship over the summer of 2008 was very high: only 10 focal plants died in the entire
year (4.2% mortality), making it impossible to detect treatment effects on survivorship (Figure 2-8C).
Plant growth and fecundity followed a similar pattern to 2007 data: disturbance had the largest
effect on plants, density had an effect only where plots were disturbed, and insect exclusion, while
effective, had no effect on either growth or fecundity. Number of leaves and stem biomass were
strongly increased by disturbance (F1,28=152.05, p<<0.01 and F1,28=235.96, p<<0.01, respectively).
Increasing density reduced average leaf number by 41.4% and stem biomass by 63.8%, but only in the
disturbed treatment (F1,28=6.95, p=0.01 and F1,28=19.77, p<<0.01, respectively). There was no
significant effect of pesticide spray on either variable (F1,28=1.98, p=0.17 and F1,28=0.12, p=0.73,
respectively). Our measure of reproduction (total seed mass) also showed a strong response to
disturbance (F1,28=56.34, p<<0.01), while increasing density lowered seed production by 65.6% in
disturbed plots only. Again, there was no effect of pesticide application.
Effects of density were comparable in both years: leaf number, stem biomass and fecundity
35
were lower at high density by approximately 40%, 65% and 70%, respectively. However, while these
effect sizes were similar between years, the absolute value of the plant responses varied widely: for
example, among plants from low density, disturbed plots only (i.e. the largest plants), those in 2008 had
fewer leaves (103.8 vs 114.9) and heavier stems (8.4g vs 2.8g) than their conspecifics in the previous
year.
2.4 Discussion
The major findings of our study across three life stages in ragweed are that enemies have i) no
effect on seed mortality, ii) slight effect on seedling survivorship and iii) No effect on adult
survivorship, growth or fecundity. We demonstrate a lack of evidence for strong enemy effects at all
life stages in this plant. Additionally, we document a large degree of tolerance to damage in ragweed,
based on its consistent growth and reproduction across our herbivore manipulation treatments and
across two years which varied widely in background levels of herbivory. Our results suggest that
enemy release is not likely to have encouraged invasion in this species. These results do, on the other
hand, suggest that tolerance may be important in the invasion process. Ragweed can serve as an
interesting model system for investigating the relationship between tolerance and invasions, as well as
investigating the mechanism of tolerance in this plant.
2.4.1 Seed survival
Ragweed seeds have high survivorship between growing seasons (Baskin and Baskin 1980),
and our results suggest that enemy release did not cause an increase in this already high rate of
survivorship. We were unable to detect any evidence for the importance of vertebrate granivores or
fungal pathogens, though both are known to be important in other systems (e.g. Maron and Kauffman
2008). We did not attempt to manipulate seed losses from insects, however insect post-dispersal
predation may be low because of the hard seed coat of ragweed seeds: once hardened, the seed coat is
36
very difficult to crack and may resist breaking by even strong-jawed organisms such as crickets and
ground beetles (pers. obs.). Because of these hard achenes, it is possible that the important time for
seed predation in this species is pre-dispersal, from granivores such as Harpalus rufipes (Coleoptera:
Carabidae) or Eurosta bellis (MacKay and Kotanen 2008). Carabids, and the Harpalini in particular, are
broadly generalist in their diet habits and consume seeds based on size (Honek et al. 2003, Honek et al.
2007); ragweed seeds are well within their preferred size range. These beetles are common in
agricultural fields (Menalled et al. 2007) and therefore a more careful study of the relationships
between their abundance and ragweed seed survivorship would be interesting.
Our study contrasts in several important ways with work in a related species, A. trifida
(Harrison and Regnier 2003). In this annual congeneric, rodent and invertebrate seed predation was as
high as 88%. Rodent predation was strongest in winter, and invertebrate predation in May-November.
We found little evidence of rodent predation; this is potentially because seed mortality is likely to be
low in disturbed areas where small mammals risk predation (Manson and Stiles 1998). In A. trifida,
another Harpalus species, (H. pensylvanica) was found to consume giant ragweed seeds much less than
other, smaller-seeded species (Harrison and Regnier 2003). However, A. trifida possesses seeds much
larger than those of A. ambrosiifolia, therefore a study on seed preferences of invertebrates A.
ambrosiifolia in this species is potentially interesting. A. trifida differs in two other important respects
from our study species: it produces few seeds which live for only a few years, whereas ragweed
produces a great number of long-lived seeds. Interestingly, Harrison et al. (2003) point out that even
with this high rate of seed predation, population limitation by this means alone is unlikely; our results
suggest that this is even more true for A. artemisiifolia.
An unknown environmental variable may be significantly influencing ragweed survival, as
suggested by the large variance in our data accounted for by experimental block (Table 2-2). While
37
sites were outwardly similar in terms of surrounding vegetation, aspect and soil type, they may differ in
some other but less obvious quality that differentially favours the survivorship and germination rate of
ragweed seed.
2.4.2 Seedling survival
Excluding enemies increased survivorship at the seedling stage; however, in general , seedling
survival was high and the effect of enemies was slight. As noted by Darwin (see quote above) such
high seedling survivorship is unusual: seedlings are quite vulnerable to predation and fungal infection.
Snails can have a very large effect on seedling survival (Ehrlen 2003, Hill and Silvertown 1997), while
many leaf-eating insects are present in spring, including the overwintered adults and newly-laid eggs of
Zygogramma suturalis. A future experiment on seedling survivorship in Ambrosia could examine
fungal morality, as our largest single mortality observation (5 of 10 seedlings) was a potential fungal
infection.
Our data suggest that Ambrosia potentially has very high seedling survivorship. Since plants,
once grown, are highly tolerant to leaf damage (Chapter 3), high seedling survivorship will likely result
in dense adult populations and a corresponding high level of seed production. Our results from this
experiment are consistent from observations of the germinated seeds in the seed experiment: of 257
observed germinated seeds, only 8 resulting seedlings died before the end of June. However, our
results should be interpreted with caution as such physical exclosures are well-known to have
experimental effects. In our case these may have been both positive and negative: while netting may
have increased moisture content and hence seedling survival, it may also have restricted light and
caused spurious mortality. A future study of seedling survival in this species should compare our cage
results with chemical enemy-exclusion treatments.
38
2.4.3 Adult survival and reproduction
No effect of enemies
Ragweed plants are not sensitive to natural levels of damage. Neither when plants were heavily
damaged in 2007, nor when we were able to better control damage levels in 2008, did we find
treatment effects on growth or reproduction.
The absence of strong enemy effects at the adult stage was robust to manipulations of density
and the presence of competitors (Figure 2-8). This was somewhat unexpected, because it is possible
that the stress caused by competition could interact with herbivory to cause a fitness loss in damaged
plants, as found by Maron and Crone (2006). Our inability to detect a fitness cost suggests that
herbivory often has weak effects on performance. However, stronger asymmetries between ragweed
and a competitor – for example, in competitive ability or herbivore load – could, at least potentially,
lower ragweed’s performance.
Disturbance effects on attack and tolerance
Ragweed is a disturbance-loving plant: every measurement of performance was higher when
plots were cleared of sod. Interestingly, clearing plants also augmented the amount of damage plants
incurred, perhaps because plants in disturbed areas are more apparent to foraging consumers and are
therefore more likely to be attacked. However, plants of disturbed areas were also more fecund. In a
separate study (Chapter 3) we show that ragweed is highly tolerant to vegetative damage when grown
in a disturbed field. This high level of response to disturbance, coupled with Ambrosia’s rapid growth
rate and highly tolerant life history, likely have played an important role in its invasion of Europe.
As reviewed by Mitchell et al (Mitchell et al. 2006), an invader’s interactions with herbivores
will vary with abiotic conditions. One such interaction was suggested by Blumenthal (2005, 2006) in
his ‘resource-enemy release hypothesis’ (R-ERH), i.e. that enemy release should have even greater
39
effects when the invading species is adapted to respond rapidly to high resource levels. Ragweed is
one such disturbance-dependent, high-responding species, which has undergone enemy release. It may
be that disturbance increases the amount of available nutrients or light and so augments the level of
tolerance in Ambrosia. Nevertheless our experiment did not demonstrate any interaction of enemy
exclusion and disturbance. However, it may be that we did not reduce damage to the low levels
observed by Genton et al (2005a,b), or that our disturbed sites are not as nutrient-rich as the agricultural
sites in France where Ambrosia is invasive. If so, it may be possible for an R-ERH effect to occur and
generate a biologically relevant performance increase.
Individual plant performance is density-dependent
Ragweed plants performed more poorly at higher densities. Density decreased all measures of
plant performance, including reproduction, but only when plots were disturbed. In such disturbed
locations, individuals were able to grow to sufficiently large sizes and directly interact with one
another, whereas heavy competition kept individual plants too small to interact in undisturbed sites.
However, while the presence of density-dependence in this species makes population regulation
possible in principle, in reality the highly-disturbed nature of ragweed’s natural habitat makes density-
dependent population regulation unlikely.
The highest density we used (25plants/m2) represents approximately the mean density of a
natural ragweed population (Foster et al. 1980),and therefore we can consider 9 plants/m2 treatment to
represent reduced intraspecific competition, as might be found in new ragweed populations. Results
from a previous study (MacKay and Kotanen 2008) also have shown a negative relationship between
the distance of ragweed from established populations and enemy attack (10 to 100m scale). The
present study reports that damage to individual plants increases at low density (1m2 scale). Newly-
established ragweed populations may experience similar conditions of isolation and low density.
40
Conclusions
Ragweed
This is the largest and most complete study of effects of biotic interactions on Ambrosia
artemisiifolia. Throughout its present range, there is interest in controlling the spread of this
agricultural weed and source of allergenic pollen. The work presented here suggests that natural levels
of consumer damage do not reduce Ambrosia survival or seed production, even while competitors are
present. These experiments suggest that top-down control is not strong in this species; consequently,
biological control programs are unlikely to provide an effective means of population control unless
they can produce damage well in excess of native-range levels. The level of damage required will
depend on the level of tolerance in Ambrosia (Chapter 3).
Biocontrol efforts involving the introduction of many ragweed-feeding invertebrates, including
the specialist Zygogramma suturalis (Coleoptera: Chrysomelidae) have not been successful (Kiss
2007). Our results suggest that this lack of success may be due to a weak effect of natural enemies of
plant performance. In particular, since ragweed invasions are worst in disturbed sites (e.g. agricultural
fields), then they may be especially tolerant to any damage. However, because ragweed pollen is an
important component of its impact, measures that reduce pollen output or broadcast may still be useful,
even if they do not result in an actual decrease in population size. We did not collect data on pollen
production directly, and therefore our results do not let us directly infer the consequences of enemy
damage for pollen production. However, collecting data on pollen production following herbivory in
this species would be interesting: a biological control agent which reduced pollen output but not
population size would nevertheless be desirable in parts of invaded Eurasia.
41
Invasions in general
While intercontinental studies of invasions are very powerful, native range studies on their own
can be important means of determining the population consequences for invaders of negative biotic
interactions. Escape from enemies has occurred during Ambrosia’s invasion in Europe, and
consequently has been suggested to contribute to its invasiveness (Genton et al. 2005a). However,
when we employed factorial manipulations to separate the effects of disturbance and herbivory
(Mitchell et al. 2006), we found a far greater effect of disturbance treatment compared to herbivore
exclusion. Enemy release, therefore, may have occurred but is unlikely to be the sole cause of its
successful invasion. The effect of conspecific crowding was likewise small: while ragweed plants had
higher performance when grown at low density in disturbed plots, the improvement in their seed
production and size was significant but smaller than the effect of disturbance itself. Ambrosia is so
strongly disturbance-dependent that the presence of suitable open habitat may be far more important
than escape from herbivory in determining its range and population growth rate. Control measures that
reduce the availability of open ground for ragweed colonization may therefore be a more effective
means of control than leaf-eating biocontrol insects.
42
2.5 Figures
0.05 0.10 0.20 0.50 1.00 2.00 5.00
510
20
50
100
200
500
1000
Mass (g) of seeds produced by one plant
Num
ber
of seeds p
roduced b
y o
ne p
lant
Figure 2-1) Relationship between seed biomass and fecundity (Linear Regression:
F1,36=947.94, p<<0.001, r2=0.96). Dotted lines represent 95% confidence limits
associated with predictions of seed number. Regression was performed on log-
transformed variables; note logarithmic axis scales.
43
Control Cage Cage+Fungicide
Germ
inate
d
02
46
810
Control Cage Cage+Fungicide
Figure 2-2) Germinating seeds from samples of 10 seeds sown on the soil surface
in November 2007 (One-way ANOVA: F2,10 = 0.231, p=0.8,). Seeds were
protected either by a cage (mammal exclusion) or a cage and fungicide (mammal
and fungal exclusion). Control plots were left untouched. Three within-block
replicates were averaged before analysis. Points show means ± 1 SEM (n=6).
44
Control BP BPN
Seedlin
gs s
urv
ivin
g to the e
nd o
f June
02
46
810
Figure 2-3) Seedling survivorship in response to herbivore exclusion treatments.
Points represent the mean, bars the SEM across five blocks. There is an increase
in seedling survivorship with enemy exclusion (Linear contrast in Randomized
block ANOVA: F1,8=6.25, p=0.04,). BP = Barrier and pellets (excluding slugs),
BPN = Barrier, pellets and net (excluding slugs and most insects). Data were log-
transformed in analysis; raw data are shown for clarity.
45
Pro
po
rtio
n o
f le
ave
s d
am
ag
ed
by A
ug
ust
0.0
0.2
0.4
0.6
0.8
1.0
9 plants/m2
25 plants/m2
9 plants/m2
25 plants/m2
Rotenone Water
Disturbed
Undisturbed
Figure 2-4) Damage to adult plants during the summer of 2007. Treatments are
disturbed (cleared plots, filled circles) and undisturbed (uncleared plots, open
circles); high and low density; and insecticide application (Rotenone) and control
(water). Points are means±SEM (n=5). See Table 2-4 for ANOVA results.
46
Pro
po
rtio
n o
f le
ave
s d
am
ag
ed
by A
ug
ust
0.0
0.2
0.4
0.6
0.8
1.0
9 plants/m2
25 plants/m2
9 plants/m2
25 plants/m2
Malathion Water
Disturbed
Undisturbed
Figure 2-5) Damage to individual adult plants during the summer of 2008.
Treatments are disturbed (cleared plots, filled circles) and undisturbed (uncleared
plots, open circles); high and low density; and insecticide application (Malathion)
and control (water). Points are means±SEM (n=5). See Table 2-5 for ANOVA
results.
47
Pro
po
rtio
n o
f p
lan
ts w
ith
ste
m g
alls
0.0
0.2
0.4
0.6
0.8
1.0
9 plants/m2
25 plants/m2
9 plants/m2
25 plants/m2
Malathion Water
Disturbed
Undisturbed
Figure 2-6) The incidence of stem-galling insects on our plants during the summer
of 2008. Treatments are disturbed (cleared plots, filled circles) and undisturbed
(uncleared plots, open circles); high and low density; and pesticide application
(Malathion) and control (water). Points are means±SEM (n=5). See Table 2-7 for
GLMM results.
48
To
tal n
um
be
r o
f le
ave
s
020
40
60
80
100
9 plants/m2
25 plants/m2
9 plants/m2
25 plants/m2
Rotenone Water
A
Ste
m b
iom
ass (
g)
01
23
4
9 plants/m2
25 plants/m2
9 plants/m2
25 plants/m2
Rotenone Water
Disturbed
Undisturbed
BS
urv
ivin
g p
lan
ts
0.0
0.2
0.4
0.6
0.8
1.0
9 plants/m2
25 plants/m2
9 plants/m2
25 plants/m2
Rotenone Water
CT
ota
l n
um
be
r o
f se
ed
s
0200
400
600
800
1200
9 plants/m2
25 plants/m2
9 plants/m2
25 plants/m2
Rotenone Water
D
Figure 2-7) Measures of plant performance in response to a factorial field
experiment in 2007. Treatments are disturbed (cleared plots, filled circles) and
undisturbed (uncleared plots, open circles); high and low density; and pesticide
application (Rotenone) and control (water). Points are means±SEM (n=5). See
Table 2-4 for ANOVA results (Figs A,B and D), and Table 2-6 for GLMM results
(Fig C).
49
To
tal n
um
be
r o
f le
ave
s
20
40
60
80
100
120
9 plants/m2
25 plants/m2
9 plants/m2
25 plants/m2
Malathion Water
A
Ste
m b
iom
ass (
g)
02
46
810
9 plants/m2
25 plants/m2
9 plants/m2
25 plants/m2
Malathion Water
Disturbed
Undisturbed
BS
urv
ivin
g p
lan
ts
0.0
0.2
0.4
0.6
0.8
1.0
9 plants/m2
25 plants/m2
9 plants/m2
25 plants/m2
Malathion Water
CS
ee
d b
iom
ass (
g)
02
46
9 plants/m2
25 plants/m2
9 plants/m2
25 plants/m2
Malathion Water
D
Figure 2-8) Measures of plant performance in response to a factorial field
experiment in 2008. Treatments are disturbed (cleared plots, filled circles) and
undisturbed (uncleared plots, open circles); high and low density; and pesticide
application (Malathion) and control (water). Points are means±SEM (n=5). See
Table 2-5 for ANOVA results (Figs A,B and D), data in Fig C was not analyzed
because almost all plants survived.
50
2.6 Tables
Table 2-1) A table of the common enemies which occur on Ambrosia at KSR. This list was compiled from personal observations and data collected by MacKay and Kotanen (2008)
Taxon Species Damage type Life stage
attacked
Tarachidia
candefacta Hbn.
T. erastrioides Hbn.
(Noctuidae)
Chewing Adult
Zygogramma
suturalis F.
(Chrysomelidae)
Chewing Seedling and
Adult
Beetles
(Coleoptera)
Systena blanda
Melsheimer
(Chrysomelidae)
Chewing Seedling and
Adult
Snails
Trichia striolata
Pfeiffer
(Hygromiidae)
Chewing Seedling
Moth
(Lepidoptera)
Epiblema sp Stem galling Adult
Small
mammals
Seed eating Seed
Table 2-2) ANOVA of enemy exclusion on seed germination. The variation has been partitioned into orthogonal polynomial contrasts to test for an increase in seed survivorship with more complete enemy exclusion. In order, the treatments are: no protection (control), cage (mammal exclusion) or a cage and fungicide (mammal and fungal exclusion). The response variable is mean count of emerged seedlings, natural-log transformed to normalize variance. All terms are tested against the residual with 10 degrees of freedom.
Source df MS F p
Treatment 2 0.025 0.354 0.71 Linear 1 0.031 0.442 0.52
Quadratic 1 0.019 0.265 0.62
Residual 10 0.071
51
Table 2-3) ANOVA of enemy exclusion on seedling survival. The variation has been partitioned into orthogonal polynomial contrasts to test for an increase in seed survivorship with more complete enemy exclusion. In order, the treatments are: no protection (control), BP (aluminum barrier and slug pellets = slug exclusion) or BPN (aluminum barrier, slug pellets and net cover = slug and insect exclusion). Data are natural-log transformed counts of surviving plants. All terms are tested against the residual variance with 8 degrees of freedom.
Source df MS F p
Treatment 2 0.06 3.13 0.1 Linear 1 0.12 6.25 0.04
Quadratic 1 <0.01 <0.01 0.97
Residual 8 0.16
Table 2-4) Factorial ANOVA on the adult experiment in 2007. Leaf number and seed number were log(x+1) transformed before analysis to normalize residuals. F-statistics and p-values are shown in bold typeface when significant.
Leaf Number Stem Biomass Seed Number Damage
Source p F1,28 p F1,28 p F1,28 p F1,28
Density <<0.01 77.19 <0.01 62.51 0.11 2.73 0.28 1.25
Dist. <<0.01 1109.8 <<0.01 62.18 <<0.01 64.87 0.97 <0.01
Spray 0.10 2.95 0.99 <<0.01 0.40 0.72 0.02 6.35
Density x Dist. 0.03 5.72 <<0.01 16.61 0.03 5.23 0.27 1.26
Density x Spray 0.31 1.09 0.86 0.03 0.53 0.40 0.66 0.20
Dist. x Spray 0.65 0.21 0.67 0.18 0.76 0.09 0.69 0.16
Density x Dist. x Spray 0.14 2.31 0.74 0.11 0.30 1.11 0.79 0.07
52
Table 2-5) Factorial ANOVA on the adult experiment in 2008. All values were log-transformed before analysis to normalize residuals, with the exception of damage, which was untransformed. F-statistics and p-values are shown in bold typeface when significant.
Leaf Number Stem Biomass Seed Biomass Damage
Source p F1,28 p F1,28 p F1,28 p F1,28
Density 0.16 2.12 <0.01 12.97 0.15 2.21 0.03 5.57
Dist. <<0.01 152.05 <<0.01 235.96 <<0.01 56.34 <<0.01 45.12
Spray 0.17 1.98 0.73 0.12 0.15 2.16 <<0.01 84.62
Density x Dist. 0.01 6.95 <<0.01 19.77 <0.01 10.39 0.31 1.05
Density x Spray 0.83 0.04 0.55 0.37 0.59 0.29 0.02 5.83
Dist. x Spray 0.05 4.29 0.96 0.00 0.47 0.54 0.03 5.53
Density x Dist. x Spray 0.37 0.84 0.32 1.03 0.47 0.53 0.25 1.36
Table 2-6) Proportion of plants surviving to reproduction in summer 2007, as predicted by all three treatment variables. Likelihood ratio tests were used to simplify a maximal model, which
contained the six parameters shown and additional β for two- and three-way interactions.
Parameter Estimate S.E. z-value P Odds ratio
Fixed effects β0 -0.90 0.18 -4.89 <0.001
βDisturbed 0.87 0.13 6.70 <0.001 2.39 Random effects
Φblock 0.047 0.217
Table 2-7) Proportion of plants attacked by stem-galling insects in 2008, as predicted by all three treatment variables. Likelihood ratio tests were used to simplify a maximal model, only the minimal adequate model is shown.
Parameter Estimate S.E. z-value P Odds ratio
Fixed effects β0 -4.19 1.01 -4.16 <<0.001 0.02
βUndisturbed -2.15 0.53 -4.06 <<0.001 0.12
βwater 3.92 1.03 3.80 <<0.001 50.4
53
LITERATURE CITED
Allard. 1943. The North American ragweeds and their occurance in other parts of the world. Science
98:292-294.
Bagarozzi D. A., J. Travis. 1998. Ragweed pollen proteolytic enzymes: Possible roles in allergies and
asthma. Phytochemistry 47:593-598.
Baker H. G. 1965. Chacteristics and modes of origin of weeds. Pages 147-172 In H.G Baker and G.L.
Stebbins, editors. Academic Press, New York, NY, .
Baskin J. M., C. C. Baskin. 1980. Ecophysiology of secondary dormancy in seeds of Ambrosia
artemisiifolia. Ecology 61:475-480.
Bassett I. J., C. W. Crompton. 1975. Biology of Canadian weeds 11. Ambrosia artemisiifolia L. and A.
psilostachya Dc. Canadian Journal Of Plant Science 55:463-476.
Bates, D., M. Maechler and B. Dai. 2008. lme4: Linear mixed-effects models using S4 classes. R
package version 0.999375-28. http://lme4.r-forge.r-project.org/
Bazzaz F. A., D. M. Mezga. 1973. Primary productivity and microenvironment in an Ambrosia
dominated old-field. American Midland Naturalist 90:70-78.
Blaney C. S., P. M. Kotanen. 2001. Post-dispersal losses to seed predators: an experimental comparison
of native and exotic old field plants. Canadian Journal Of Botany-Revue Canadienne De Botanique
79:284-292.
Blumenthal D. 2005. Interrelated causes of plant invasion. Science 310:243-244.
Blumenthal D. M. 2006. Interactions between resource availability and enemy release in plant invasion.
Ecology Letters 9:887-895.
Bolker, B. M. 2008. Ecological Models and Data in R. Princeton University Press
54
Cadotte M. W., B. R. Murray, and J. Lovett-Doust. 2006. Ecological patterns and biological invasions:
Using regional species inventories in macroecology. Biological Invasions 8:809-821.
Colautti R. I., S. A. Bailey, C. D. A. van Overdijk, K. Amundsen, and H. J. MacIsaac. 2006.
Characterised and projected costs of nonindigenous species in Canada. Biological Invasions 8:45-59.
Crawley, M. J. 2007. The R Book. Wiley
Darwin, C. R. 1859. On the Origin of Species by Means of Natural Selection, or the Preservation of
Favoured Races in the Struggle for Life. London: John Murray. 1st edition, 1st issue.
Ding J. Q., R. N. Mack, P. Lu, M. X. Ren, and H. W. Huang. 2008. China's booming economy is
sparking and accelerating biological invasions. Bioscience 58:317-324.
Ehrlen J. 2003. Fitness components versus total demographic effects: Evaluating herbivore impacts on
a perennial herb. American Naturalist 162:796-810.
Elton C. S. 1958. The Ecology of Invasions by Animals and Plants. University Of Chicago Press.
Foster M. M., P. M. Vitousek, and P. A. Randolph. 1980. The Effects of Ragweed (Ambrosia
artemisiifolia L.) on Nutrient Cycling in a 1st-year Old-field. American Midland Naturalist 103:106-
113.
Friedman J., S. C. H. Barrett. 2008. High outcrossing in the annual colonizing species Ambrosia
artemisiifolia (Asteraceae). Annals of botany 101:1303-1309.
Genton B. J., J. A. Shykoff, and T. Giraud. 2005a. High genetic diversity in French invasive
populations of common ragweed, Ambrosia artemisiifolia, as a result of multiple sources of
introduction. Molecular ecology 14:4275-4285.
Genton B. J., P. M. Kotanen, P. O. Cheptou, C. Adolphe, and J. A. Shykoff. 2005b. Enemy release but
55
no evolutionary loss of defence in a plant invasion: an inter-continental reciprocal transplant
experiment. Oecologia 146:404-414.
Grimm E. C., G. L. Jacobson, W. A. Watts, and B. C. S. H. a. K. A. Maasch. 1993. A 50,000-Year
record of climate oscillations from Florida and its temporal correlation with the Heinrich Events.
Science 261:198-200.
Guo Q. 2006. Intercontinental biotic invasions: what can we learn from native populations and
habitats? Biological Invasions 8:1451-1459.
Halpern S. L., N. Underwood. 2006. Approaches for testing herbivore effects on plant population
dynamics. Journal of Applied Ecology 43:922-929.
Harrison S. K., E. E. Regnier. 2003. Postdispersal predation of giant ragweed (Ambrosia trifida) seed in
no-tillage corn. Weed Science 51:955-964.
Hierro J. L., J. L. Maron, and R. M. Callaway. 2005. A biogeographical approach to plant invasions: the
importance of studying exotics in their introduced and native range. Journal of Ecology 93:5-15.
Hill B. H. C., J. Silvertown. 1997. Higher-order interaction between molluscs and sheep affecting
seedling numbers in grassland. Acta Oecologica-International Journal Of Ecology 18:587-596.
Honek A., Z. Martinkova, and V. Jarosik. 2003. Ground beetles (Carabidae) as seed predators.
European Journal of Entomology 100:531-544.
Honek A., Z. Martinkova, P. Saska, and S. Pekar. 2007. Size and taxonomic constraints determine the
seed preferences of Carabidae (Coleoptera). Basic and Applied Ecology 8:343-353.
Hulme P. E. 1998. Post-dispersal seed predation: consequences for plant demography and evolution.
Perspectives in Plant Ecology, Evolution and Systematics 1:32-46.
56
Igrc J., C. J. Deloach, and V. Zlof. 1995. Release and establishment of Zygogramma suturalis F
(Coleoptera, Chrysomelidae) in Croatia for control of common ragweed (Ambrosia artemisiifolia L).
Biological Control 5:203-208.
Keane R. M., M. J. Crawley. 2002. Exotic plant invasions and the enemy release hypothesis. Trends in
Ecology and Evolution 17:164-170.
Kiss L. 2007. Is Puccinia xanthii a suitable biological control agent of Ambrosia artemisiifolia?
Biocontrol Science and Technology 17:535-539.
Kosola K. R., K. L. Gross. 1999. Resource Competition and Suppression of Plants Colonizing Early
Successional Old Fields. Oecologia 118:69-75.
Liu H., P. Stiling. 2006. Testing the enemy release hypothesis: a review and meta-analysis. Biological
Invasions 8:1535-1545.
Mack R. N., D. Simberloff, W. M. Lonsdale, Evans, H. and Clout,M., and F. A. Bazzaz. 2000. Biotic
invasions: Causes, epidemiology, global consequences, and control. Ecological Applications 10:689-
710.
MacKay J., P. M. Kotanen. 2008. Local escape of an invasive plant, common ragweed (Ambrosia
artemisiifolia L.), from above-ground and below-ground enemies in its native area. Journal of
Ecology 96:1152-1161.
Manson R. H., E. W. Stiles. 1998. Links between microhabitat preferences and seed predation by small
mammals in old fields. Oikos 82:37-50.
Maron J. L., E. Crone. 2006. Herbivory: effects on plant abundance, distribution and population
growth. Proceedings Of The Royal Society B-Biological Sciences 273:2575-2584.
Maron, J. L. and M. J. Kauffman. 2008. Habitat-specific impacts of multiple consumers on plant
57
population dynamics. 87: 113-124
Maron J. L., M. Vila. 2001. When do herbivores affect plant invasion? Evidence for the natural
enemies and biotic resistance hypotheses. Oikos 95:361-373.
McEvoy P. B., N. T. Rudd, C. S. Cox, and M. Huso. 1993. Disturbance, competition, and herbivory
effects on ragwort Senecio jacobaea populations. Ecological Monographs 63:55-75.
Menalled F. D., R. G. Smith, J. T. Dauer, and T. B. Fox. 2007. Impact of agricultural management on
carabid communities and weed seed predation. Agriculture Ecosystems and Environment 118:49-54.
Mitchell C. E., A. A. Agrawal, J. D. Bever, G. S. Gilbert, R. A. Hufbauer, J. N. Klironomos, Maron, J.
L. and Morris,W.F., I. M. Parker, A. G. Power, E. W. Seabloom, M. E. Torchin, and D. P. Vazquez.
2006. Biotic interactions and plant invasions. Ecology Letters 9:726-740.
Neergaard, P. 1977. Seed Pathology. Volume I. Halsted Press, New York. 839 pp.
Parker J. D., D. E. Burkepile, and M. E. Hay. 2006. Opposing effects of native and exotic herbivores on
plant invasions. Science 311:1459-1461.
Pimentel D., L. Lach, R. Zuniga, and D. Morrison. 2000. Environmental and economic costs of
nonindigenous species in the United States. Bioscience 50:53-65.
Quinn, G. P. and M.J. Keough. 2002. Experimental Design and Data Analysis for Biologists.
Cambridge University Press.
R Development Core Team. 2008. R: A language and environment for statistical computing. R
Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-
project.org.
Raynal, D. J. and F.A. Bazzaz. 1975. Interference of winter annuals with Ambrosia artemisiifolia in
58
early successional fields. Ecology 56: 35-49
Rose K. E., S. M. Louda, and M. Rees. 2005. Demographic and evolutionary impacts of native and
invasive insect herbivores on Cirsium canescens. Ecology 86:453-465.
Schafer M., P. M. Kotanen. 2003. The influence of soil moisture on losses of buried seeds to fungi.
Acta Oecologica-International Journal Of Ecology 24:255-263.
Schüder, I., G. Port, and J. Bennison. 2003. Barriers, repellents and antifeedants for slug and snail
control. Crop Protection 22: 1033-1038
Sharvelle, E.G. 1961. The Nature and Uses of Modern Fungicides. Burgess Publishing, Minneapolis,
MN.
Sokal, R. R. and F. J. Rohlf. 1994. Biometry: The Principles and Practices of Statistics in Biological
Research. W. H. Freeman
Stevens M. H. H., W. P. Carson. 1999. Plant density determines species richness along an experimental
fertility gradient. Ecology 80:455-465.
Teshler M. P., A. DiTommaso, J. A. Gagnon, and A. K. Watson. 1981. Biological control programmes
against insects and weeds in Canada 1981-2000. Pages 290-294 In P. G. Mason and J. T. Huber,
editors. CABI Publishing, .
Torgeson, D.C. 1969. Fungicides: An Advanced Treatise, Volume II: Chemistry and Physiology.
Academic Press, New York.
Vitousek P. M., H. A. Mooney, J. Lubchenco, and J. M. Melillo. 1997. Human domination of Earth's
ecosystems. Science 277:494-499.
59
3 Leaf damage has weak effects on fecundity in Ambrosia artemisiifolia L.
3.1 Introduction
Throughout the biosphere, plant invasions are becoming increasingly common, presenting
ecologists with both challenges and opportunities. Human activities, particularly habitat alteration and
the intentional and unintentional transport of organisms, have caused rapid, unprecedented mixing of
the Earth’s biota (Elton 1958). Invasive plant species threaten the integrity of both natural and
agricultural systems, and reduce the economic and aesthetic value of land (Vitousek et al. 1997, Mack
et al. 2000). These losses are estimated to total in the billions for Canada’s economy alone (Colautti et
al. 2006); there is great applied interest in understanding invasions so that they can be managed or
prevented (Sakai et al. 2001, Halpern and Underwood 2006). In addition to presenting management
challenges, plant invasions offer an opportunity to learn more about population and community
ecology in general (Kolar and Lodge 2001, Lodge 1993). Studies in invasion biology have contributed
to population biology (Sakai et al. 2001), demonstrated when and to what degree biotic interactions
affect plant distribution and abundance (Klironomos 2002, Mitchell and Power 2003), and informed
management decisions (Colautti et al. 2006).
The Enemy Release Hypothesis (ERH, Elton 1958, reviewed in Keane and Crawley 2002 and
Liu and Stiling 2006), is an important hypothesis within plant invasion biology. The ERH posits that
when plants are introduced to new areas without their ‘enemies’ (i.e. all organisms having direct
negative effects on a plant) the result is increased reproduction and vigour of plants in their introduced
range (Mitchell and Power 2003, Keane and Crawley 2002, Mitchell et al. 2006, Agrawal and Kotanen
2003, Maron and Vila 2001, Wolfe 2002). The importance of invertebrate herbivores in particular has
been emphasized by studies of the ERH (Callaway and Maron 2006).
60
However, population limitation by enemies is not universal: some species possess tolerance
traits that maintain fitness after damage (Kotanen and Rosenthal 2000, Weis and Franks 2006),
buffering the population against top-down control (Tiffin 2000a). Different types of damage are
tolerated in different ways (Tiffin 2000b) which may not necessarily be correlated with one another
(Kotanen and Rosenthal 2000). For example, leaf damage may cause the photosynthetic rates of
remaining tissues to increase (Strauss et al. 2003, Strauss et al. 2002); however, such increases may not
always occur (Caldwell et al 1981) nor may necessarily be a tolerance response (Nowak and Caldwell
1984). On the other hand, damage to the apical meristem, another common mode of plant damage,
often results in rapid regrowth of secondary meristems (reviewed in Tiffin 2000b). Rarely, tolerance
can even take the form of overcompensation, wherein herbivore-attacked plants seem to reproduce
more than their undamaged neighbours (Hawkes and Sullivan 2001). The strength of a species’
tolerance response determines the degree to which it is limited by natural enemies, which in turn
influences how strongly it will respond to enemy release.
To understand how damage relates to fitness, manipulative studies are required; such studies are
most effective in a plant’s native range. Clipping experiments are often useful, because insect attack
rates can vary widely between years (Agrawal and Kotanen 2003), and because estimating tolerance
from natural levels of damage can give a biased estimate (Tiffin and Inouye 2000). Performing such
studies in the native range is important because they allow an assessment of plant responses in a habitat
which contains the biotic and abiotic environment in which it evolved. Because most studies of
invasives are carried out in the exotic range, where they are conspicuous problem species (Colautti et
al. 2004), they may be biased in their measurement of plant response to treatment (Hierro et al. 2005,
Guo 2006). Understanding how – and if – damage to plants causes a reduction of fecundity in their
native range is important to determine how release from damage will contribute to invasion.
61
Ambrosia artemisiifolia (L.) is an ideal system for exploring the relationship between damage
and fitness in an invasive plant. This North American annual is both an agricultural and human health
pest throughout Eurasia, where it spread during the 20th
century (Allard 1943, Chauvel et al. 2006).
Previous studies have confirmed that these plants have indeed experienced escape from enemies during
this invasion: at least in France, individuals are much less attacked than Canadian populations (Genton
et al. 2005). In an attempt to manage ragweed’s invasion, most oligophagous insects which feed on
ragweed have been introduced to Eurasia; however, these attempts have been unsuccessful (Igrc et al.
1995).
We use artificial damage treatments to investigate the relationship between damage and fitness
within a native-range population of ragweed (see Table 3-1). We simulated two damage types,
representing two different groups of enemies: stem borers (“meristem removal”) and leaf chewers
(“leaf clipping”). For each damage type, we asked the following questions: i) does damage decrease
stem biomass (an estimate of plant size) and fecundity (seed mass)? ii) Does damage affect the
relationship between size and fecundity (relative allocation of biomass)?
3.2 Methods
3.2.1 Site
These experiments were conducted at the Koffler Scientific Reserve (KSR) at Joker's Hill, a
350-ha field station owned by the University of Toronto and situated 50km north of Toronto, Ontario
(44 03'N, 79 29'W). A complete plant species list for KSR is provided on the reserve's webpage
(http://www.ksr.utoronto.ca/). Ragweed is common at the site in general and at the experimental site in
particular, emerging readily from the seed bank once our experimental site was cleared. This site
description is also included in Chapter 2, and is included here for completeness.
62
3.2.2 Species
Ragweed is a wind-pollinated spring annual (Bazzaz 1974), germinating late in May and
growing rapidly in open sites. Ragweed is highly disturbance-dependent and is excluded by dense
vegetation (Bazzaz and Mezga 1973, Foster et al. 1980, Stevens and Carson 1999) but grows readily in
weedy, open habitats such as cleared fields (Bazzaz 1974, Kosola and Gross 1999, Maryushkina 1991).
It is difficult to eradicate from a site because of its very long-lived seed bank (Baskin and Baskin
1980). It is widespread throughout its native North America (Bassett and Crompton 1975, Bassett and
Terasmae 1962, Teshler et al. 1981), and its abundant pollen is used by palynologists as an indicator of
human disturbance and agriculture (Grimm et al. 1993). This pollen is highly allergenic and a principle
cause of hayfever (Bassett and Crompton 1975, Bagarozzi and Travis 1998). Ragweed is an important
crop weed and disturbed sites in North America (Bazzaz 1974, Bassett and Crompton 1975) and has
been widely introduced throughout Eurasia, where it has formed successful invasive populations (Kiss
2007).
During invasion, Ambrosia appears to have escaped from its natural enemies: leaf damage is
common in North America but uncommon on populations in France (Genton et al. 2005). Nonetheless,
biocontrol efforts have not been successful, despite the introduction of ragweed-specialized herbivores
such as Zygogramma suturalis (Coleoptera: Chrysomelidae) (Igrc et al. 1995). The insect consumers
from which Ambrosia escaped include both leaf chewing and stem galling herbivores (Table 3-1); we
simulated these damage types in our experimental treatments.
3.2.3 Experimental design
In June 2008, we cleared an experimental plot approximately 30x30m, clearing away old-field
vegetation including a mix of grasses (Bromus, Festuca) and dicots (Cirsium, Asclepias). 150 ragweed
seedlings were gathered from a wild population at KSR and planted in a large cleared field in ten rows
63
(2m apart) of 15 plants, each with 1m between neighbouring plants. Seedlings had 1-2 pairs of mature
leaves and were less than 6cm high at the time of transplanting. Plants were watered for two days
following transplant to protect against transplant shock; plants dying within this period were replaced
(9 plants).
Each row received one of five treatments; two rows were assigned to each treatment
(n=30/treatment). Our five treatments were three levels of leaf removal (75%, 25% and 5%), apical
meristem removal (AM), and an undamaged control (Control). Damage was dispersed evenly over the
entire plant in the 75% and 25% treatments, while in the 5% treatment only fully expanded leaves near
the top of plants were removed in order to make this treatment as repeatable as possible. Leaves were
removed by clipping the petiole close to the stem. Because we removed whole leaves, we delayed
application of the 5% treatment until most plants in this treatment had at least 20 leaves: treatments
were applied to the 75%, 25%, and AM treatments on 16 July 2008 and to the 5% treatment on 28 July
2008. On 26 Aug 2008, all treatments were applied again.
3.2.4 Data collection
In September we collected all surviving plants; at this time growth had ceased and seeds had
matured. After collection, plants were allowed to dry at room temperature before weighing. For each
plant we recorded the mass of all aboveground material (total biomass) and then separated and weighed
the stem and the seeds. We separated seeds from leaf tissue with a 1.4 mm sieve. We used the total
seed biomass as a measure of fecundity, based on the strong relationship between this variable and seed
number (both variables log-transformed, F1,36 = 941.9, r2
= 0.96, p<<0.001).
64
3.2.5 Analyses
i) Are size and reproduction decreased by damage?
We tested for a significant decline in stem biomass or seed production as leaf damage treatment
increases with two separate linear regressions. In this analysis, damage is treated as a continuous
independent variable. We compared the meristem removal treatment to the control in a one-tailed t-
test; meristem removal was hypothesized to reduce growth and reproduction.
ii) Is relative allocation to reproduction decreased by damage?
In contrast to the linear regression described above, our second analysis tests two related
questions: first, does the relationship between plant size and seed biomass change with increasing leaf
damage (i.e. a significant interaction term in the model), and second, if this interaction is significant,
does clipping have a qualitatively different effect for plants of different size? We asked these questions
using the following linear model:
( 3-1) Seed weight = damage + stem weight + damage*stem weight + error
In this analysis, damage is a fixed, ordered factor, stem a continuous variable (stem biomass in
grams) and damage*stem their interaction. In the analysis of the “leaf clipping” experiment, the
damage term included four treatments: Control, 5%, 25% and 75%. Because treatment is an ordered
factor, our analysis tests the a priori hypothesis that increasing damage has increasingly strong effects
on the allocation to reproduction using orthogonal polynomial contrasts (Crawley 2007, Quinn and
Keough 2002). We analyzed the “meristem removal” experiment with the same model, this time
including only two levels of damage: control and meristem removal.
In both analyses, stem and seed biomass were loge-transformed. This was important for two
reasons. First, such a transformation normalizes the residual variance and is particularly useful when
65
looking for relationships among mass variables, which in natural populations can be log-normally
distributed (Bolker 2008). This makes our statistical model robust and allows easy comparison among
treatments using a traditional ANCOVA approach. Secondly, fitting this model is equivalent to
estimating the parameters of a power-law relationship, a very common phenomenological model that is
frequently fit to growth-fecundity data (e.g. Dodd and Silvertown 2000). Fitting the model of equation
( 3-1 is equivalent to fitting ln(seed)=β0i+ β1iln(stem), for each treatment level, where β0i and β1i
represent the intercept and slope of the relationship for the ith (0% to 75%) damage treatment. To
obtain parameter estimates for the power law equation, we back-transform ln(seed)=β0+ β1ln(stem) to:
(3-2) seed=eβ0stemβ1
Therefore, when β1 is not significantly different from unity, X (plant size) and Y (reproduction)
are in direct proportion, as determined by the constant eβ0
. When β1 is either higher or lower than one
we have evidence for differential allocation among seeds and stem.
We used stem biomass rather than total biomass because our tissue removal treatments directly
lowered leaf biomass (and therefore total biomass) while any effects on stem biomass would have been
indirect. Stem biomass is interesting for two reasons: first, it is a direct estimate of the amount of
structural biomass in the aboveground part of the plant, with potential consequences for seed dispersal
and pollen broadcast and reception. Secondly, stem biomass is an unbiased estimate of plant size and
therefore and indicator of the ability of plants to grow in a particular part of our study plot. We were
therefore able to compare fecundity among levels of damage, controlling for plant size, where the
damage*stem parameters were not significantly different. To do so, we examined the 95% confidence
intervals on the damage parameters (i.e., the intercepts of regression lines on log-transformed
variables). When a significant interaction occurred, we tested for differences in reproduction among
66
plants of different sizes by inspecting 95% confidence intervals for regression lines. Data were
analyzed in R (R Development Core Team 2008).
3.3 Results
3.3.1 i) Direct effects of damage
Survivorship was high during the summer: only two plants died after replanting, one in July in
the 75% treatment, and another in August in the 25% treatment. A third sample was damaged, leaving
117 plants in the clipping experiment and 60 in the meristem experiment. Stem biomass decreases with
increasing damage intensity (linear regression: F1,115=10.17, p<0.002), however the amount of variation
explained (r2 = 0.073) is slight (Figure 3-1). Seed mass, on the other hand, showed no difference in
response to treatment (linear regression: F1,115=0.236, p=0.628, Figure 3-2). Plants with damaged
apical meristems produced the same seed mass and had the same stem mass as plants that were
undamaged (Control: 1.62±0.38, Apical: 1.28±0.25 (mean±SE), one-tailed t-test, t=0.261, p=0.60,
Figure 3-6).
3.3.2 ii) Size-dependent effects
The relationship between seed and stem biomass does vary across treatments (Figure 3-3,
F7,109=17.59, r2=0.50, p<<0.001). Specifically, plants produce fewer seeds per unit mass under heavy
clipping (significant interaction term: F3,109=3.83, p=0.012). We found support for a linear decline in
the value of the slope as clipping intensity increased (linear contrast ± SE = -0.59±0.23, t = -2.6, p =
0.01). This decline is largely driven by the effect of removing 75% of leaf area which reduced the
slope coefficient to 0.43 ± 0.17 (SE), significantly different from the other three slopes (Figure 3-3,
inset).
Linear contrasts also reveal a significant linear increase in the intercepts of this relationship as
67
damage increases (Linear contrast ± s.e = 1.15±0.31, F1,109=12.00, p<0.001). Because the slopes of
regression lines in the 25% and 5% treatments are not different from the control, we are able to directly
compare the elevations of lines in low-damage treatments (Figure 3-4). 25% leaf removal creates a
151% increase in elevation relative to the control (Figure 3-4). The effect of removing 75% of leaves
is different for plants with different stem biomass: small, heavily damaged plants produced more seeds
than lightly-damaged plants of the same size, while among plants that grew to a large final size,
damage does not affect reproduction at all (Figure 3-5).
Removing the apical meristem had no effect on plant allocation to reproduction. The full linear
model (including treatment, plant size and their interaction) was no better at predicting the total mass of
seeds produced than a reduced model including plant size alone (RSSfull = 53.151 with 54 df, RSSreduced
= 56.696 with 56 df, F2,54=1.80, p=0.17). Although the slope of the regression line for damaged plants
is lower than undamaged (Figure 3-7, inset), this difference is not significant. In addition, the
confidence intervals for these two lines broadly overlap (Figure 3-8).
3.4 Discussion
The major findings of this experiment are i) a very high level of tolerance to both leaf and
apical meristem damage in Ambrosia and ii) evidence for plastic allocation of biomass to reproduction
in response to leaf damage. The former conclusion is supported by the slight effect of damage on stem
mass and no discernable effect on absolute seed production. The latter result is supported by an
analysis including both stem and seed mass, revealing that the relative allocation to each is not
sensitive to low amounts of damage, but at very high levels allocation to seeds is lowered relative to
stem. Our findings indicate that even heavy damage is not sufficient to reduce reproduction in ragweed
plants, and suggest that plastic allocation strategies in Ambrosia act to minimize the effect of herbivory
on fitness. This high level of tolerance is a potential cause of its success as a global invader, and an
68
obstacle to proposed biocontrol efforts.
3.4.1 Weak effects of damage on plant performance
For damage to act to limit plant populations, there must be some negative relationship between
rates of damage and plant performance. We were unable to find any relationship between our damage
treatments and seed production, though there was a slight effect on stem biomass. Because ragweed
plants are annual, the production of seeds is the key step in its life cycle, whereas stem biomass is
unlikely to have a direct effect. Therefore, because damaged plants do not contribute fewer propagules
to the seed bank, the growth rate of ragweed populations may be robust to the presence of herbivores.
Stem biomass, however, may have important indirect effects on invasion in Ambrosia. Ragweed
invasions are problematic in part because plants reach nuisance levels of population density (Igrc et al.
1995), but also because they produce copious amounts of wind-dispersed pollen. This pollen is highly
allergenic and a major cause of hayfever. If the amount of stem biomass a plant produces influences its
ability to broadcast pollen, then heavy damage may reduce this important aspect of Ambrosia invasion.
However, any reduction in pollen density is unlikely to result in fewer seeds: as in many wind-
pollinated annual plants, female flowers are not pollen-limited (Friedman and Barrett 2008).
Nevertheless, ragweed is self-incompatible (Friedman and Barrett 2008) and so there is, at least
theoretically, potential for plants to be pollen-limited.
Our study found a high level of tolerance to apical meristem damage. We were unable to find
any effects of experimentally removing apical meristems either on seed production (Figure 3-6) or on
the size-fecundity relationship (Figure 3-7). Previous studies have also used artificial meristem
damage to investigate the potential fitness effects of this damage. For example, Doak (1992) found
that experimental damage to the tops of the perennial Epilobium latifolium caused a decrease in the
population growth rate over time. Interestingly, this effect was only noticed when populations were
69
followed for three years. Our results corroborate our findings from a separate field study where plants
attacked by galls were not less fecund, even when we controlled for plant size (see Chapter 2).
Simulated damage is not always equivalent to damage by invertebrates and so herbivore
damage may have had a greater effect than our treatments. For example, Strauss and Agrawal (Strauss
et al. 2003) compared simulated damage with herbivory by caging herbivores on plants and found that
consumer damage induced a 10- to 20-fold increase in levels of defensive chemicals in Raphanus
raphanistrum. It is not known if such induced defense occurs in Ambrosia, however, the possibility
exists: putative defence chemicals have been isolated from plant tissue (David et al. 1999). This is
relevant because qualitative differences between simulated and herbivore damage sometimes prevent
broad generalizations from studies of clipping damage. This is true for leaf damage but perhaps
especially so for meristem damage, which is likely to be qualitatively different from the presence of a
stem-galling insect, which may have many additional effects on the plant.
However, artificial clipping experiments are useful because enemy attack can be highly
spatially and temporally variable, limiting the utility of observational studies of enemy damage. When
environmental variables influence both the degree of damage and plant fitness, clipping experiments
provide a more accurate estimate of tolerance than studies which rely on observed natural herbivory
(Tiffin and Inouye 2000). A high amount of year-to-year variation in enemy attack has been observed
in the few studies quantifying enemy attack over several years. For example, Agrawal et al. (2005)
found high variation in herbivory levels between two years (2002 and 2003) at the same field site
where our experiments were conducted. Similarly, we found much higher levels of attack in control
plots in 2007 compared to 2008 (See Chapter 2). One of the advantages, therefore, of this artificial
clipping experiment is its exposure of the relationship between damage and fitness in ragweed without
being constrained to available levels of natural herbivory. The biologically relevant effects of sporadic
70
but high leaf damage may only be detected via long-term monitoring studies.
3.4.2 Leaf clipping changes the pattern of allocation to biomass
We found three interesting results in our study of the relative allocation to stem and seed
biomass in ragweed. First, we found that leaf damage up to 25% does not change the relative
allocation between seed and stem biomass. Second, light damage to Ambrosia may actually increase
the number of seeds produced. Third, very heavy damage does change the relative allocation to seed
and stem biomass, with the result that heavily damaged plants with light stems produce more seeds
than undamaged plants, while among large plants there is no response to damage. The first two results
further establish ragweed as a very tolerant species to which leaf damage presents no easy means of
control. The third suggests that ragweed may possess plastic allocation strategies which contribute to
its tolerance to leaf herbivory. In fact, we have found little evidence that leaf damage can reduce
fecundity in this species – indeed, light levels of damage may enhance reproduction.
Damage up to 25% of leaves removed does not reduce the size-dependent fecundity of
Ambrosia. This degree of damage does not result in a significant difference in biomass allocation to
seeds or stems (i.e., the slopes of the lines in Figure 3-3), and therefore we can compare the elevations
of these lines directly. The effect of light herbivory is to slightly increase the value eβ0
(Eq. (3-2),
indicating that light damage causes Ambrosia to increase seed production slightly irrespective of
growth. Examination of the elevations of the lines (Figure 3-3) and the confidence intervals for power-
law parameters (eβ0
, Figure 3-4) shows that plants with 25% of leaves removed produce more seed than
control plants. These results suggest that damage stimulates seed production (i.e. ‘overcompensation’).
Heavy damage, in contrast, alters the relationship between seed production and stem growth.
Across the two lightest damage treatments and the control, we found that the relationship between
ln(plant size) and ln(fecundity) was neither different from unity nor altered by damage. At 75% of
71
leaves removed, the exponent in the power law equation (e.g. β1) drops below one, indicating that the
production of seeds is no longer directly proportional to plant size (Figure 3-3, inset). Inspection of
95% confidence intervals for regression lines indicate that plants in this treatment reproduced more
than undamaged plants with small final stem mass, but completely tolerated damage when plants in
both treatments grew to large final stem mass (Figure 3-3, Figure 3-5). Several mechanisms could
have caused this pattern. The observed increase in small, highly reproducing individuals could be
caused either by overcompensation only in small individuals, by a decrease in stem mass (but not seed
production) of larger individuals or a combination of these factors. While further experiments would
be necessary to elucidate the precise mechanism of tolerance in this plant, our results nevertheless
suggest that the allocation strategies of Ambrosia are adaptable to damage, and that this plasticity may
have contributed to its success as an invader.
3.4.3 Tolerance effects on invasions
Tolerance to herbivory may encourage invasions. Plants with high levels of tolerance may be
less sensitive to herbivory than other members of the invaded community, releasing them from
apparent competition with these community members. Tolerance traits may have important indirect
effects as well: if introduced tolerant plants cause an increase in enemy populations, this herbivore load
may spill over onto native flora, further skewing the relative strength of apparent competition in favour
of the exotic (Ghazoul 2002).
Tolerant plants are less sensitive to herbivory, and therefore more difficult targets for biocontrol
and artificial removal. Successful biocontrol of ragweed by an aboveground biocontrol agent is not
likely: ragweed is too tolerant of leaf damage, and any successful biocontrol agent would have to
produce more than the 75% removal of leaf biomass that we produced, since our manipulated damage
had no effect (Figure 3-2). Given the high rates of seed and seedling survivorship (see Chapter 2) and
72
the longevity of the seed bank (Baskin and Baskin 1980), population reduction via leaf damage or
meristem galling is unlikely to be effective in this species.
3.4.4 Tolerance and the ERH
While the enemy release hypothesis continues to inspire comparative studies of damage
(Colautti et al. 2004, Hierro et al. 2005, Liu and Stiling 2006,), the relationship between damage and
fitness is often overlooked (Maron and Vila 2001). In the presence of strong tolerance, loss of enemies
may occur during invasion without providing a mechanism for invasion. Because tolerance traits
prevent plants from experiencing top-down limitation, they also reduce the importance of natural
enemies in controlling population size. As well, tolerant plants are often fast-growing, and are therefore
likely to quickly capitalize on the disturbances which in some habitats cause invasion ‘windows’ (Mack
et al 2000). This is supported by our results from Chapter 2: Ambrosia responds much more strongly to
disturbance than to enemy release. By extending our damage treatments well beyond the mean levels of
leaf damage observed in 2008 (Table 3-1), we conclude that even in years where insect attack is
unusually high, ragweed plants can still present a very tolerant phenotype.
3.5 Conclusions
The slight effect of extreme damage on plant performance indicates an extremely high level of
tolerance to leaf damage in Ambrosia, a trait that may have contributed to its success as an invader. In
Eurasia, the demographic success of ragweed may be due to its disturbance-dependent ecology (see
Chapter 2), rather than to enemy release. Because leaf damage has little effect on fitness, enemy
release alone is unlikely to result in a large fitness increase in this plant. Additionally, any attempts at
biocontrol of this species, particularly using invertebrate folivores, is unlikely to meet with success
unless levels of damage exceed the 75% leaf removal that we simulated.
73
3.6 Acknowledgments
M. Saunders, K.A. Judge, L.J. Robson for useful comments. A Simonsen, R Dinnage, Z
Burivalova, R. MacKenzie, A.M. Petersen, V. Pham and especially K. Kostyukova for helpful field
work. J. Stinchcombe and S.C.H. Barrett for advice, Art Weis and KSR for logistical support. This
work would have been impossible without the donation of KSR to the U of T by the Koffler family.
This work was supported by and NSERC PGS to A.A.M.M.
74
3.7 Figures
0.0 0.2 0.4 0.6 0.8 1.0
0.5
1.0
2.0
5.0
10.0
20.0
Proportion removed
Ste
m b
iom
ass (
g)
Figure 3-1) Stem biomass in response to removal of leaves. Total mass declines
slightly as damage intensity increases (linear regression on ln-transformed stem
biomass: F1,115=6.52, p=0.012), however the amount of variation explained is
slight (r2 = 0.045).
75
0.0 0.2 0.4 0.6 0.8 1.0
0.0
10.0
50.5
05.0
0
Proportion removed
Seed p
roduction (
g)
Figure 3-2) Seed production in response to leaf removal. There is no response to
treatment (linear regression: F1,115=0.236, p=0.628).
76
0.5 1.0 2.0 5.0 10.0 20.0
0.0
10.0
50.5
05.0
0
Stem biomass (g)
Seed m
ass (
g)
Control
5%
25%
75%
Control 1/20 1/4 3/4
slo
pe
0.0
0.5
1.0
1.5
2.0
Figure 3-3) The relationship between seed biomass and plant size (Stem biomass)
changes with response to clipping (F7,109 = 17.59, p<<0.001, r2 = 0.50,). The inset
shows the slopes ± 95% confidence intervals for each line. There is a significant
decline in slope as damage increases (linear contrast = -0.59±0.23, p=0.01); the
75% treatment departs significantly from 1 (grey line).
77
Control 1/20 1/4
Treatments
Ele
vations
0.0
0.2
0.4
0.6
0.8
Figure 3-4) Back-transformed elevations of the lines in Figure 3-3, for each of the
three lowest damage treatments, with 95% confidence intervals. The slopes of
these lines are not different (Figure 3-3), and therefore comparing intercepts
allows a comparison of the different elevations of these lines. Plotting symbols
are the same as in Figure 3-3.
78
1 2 5 10 20
0.0
50.1
00.2
00.5
01.0
02.0
05.0
0
Vegetative biomass (g)
Seed m
ass (
g)
Control
apical
Figure 3-5) The effect of clipping on the relationship between plant size (stem
biomass) and seed mass. Dark lines represent regression lines (±95% CI in lighter
lines). Differences in seed production relative to size are significant (p<0.05)
where the confidence intervals of one line do not include the other. For clarity,
individual data points are omitted.
79
Apical meristem removed Control
−3
−2
−1
01
2
Seed m
ass (
ln(g
))
Figure 3-6) Box-and-whisker plots showing the response of seed mass to apical
meristem clipping. Apical meristem removal does not lower the mass of produced
seeds (one-tailed t-test, t=0.261, p=0.60, df=53.3). Lower and upper edges of
boxes represent the first and third quartile of the data, respectively, while the
upper and lower ‘whiskers’ represent the length of the box multiplied by 1.5.
Notches present an estimate of the value of the median.
80
1 2 5 10 20
0.0
50.1
00.2
00.5
01.0
02.0
05.0
0
Vegetative biomass (g)
Seed m
ass (
g)
apical
Control
control apical
0.0
1.0
2.0
Figure 3-7) Seed production as a function of aboveground biomass for control
and apical-removal treatments. The inset shows slopes ± 95% confidence
intervals. Although the slope is lower for damaged plants, this difference is not
significant (F-ratio: RSSfull = 53.151 with 54 df, RSSreduced = 56.696 with 56 df,
F2,54=1.80, p=0.17)
81
1 2 5 10 20
0.0
50.1
00.2
00.5
01.0
02.0
05.0
0
Vegetative biomass (g)
Seed m
ass (
g)
Control
apical
Figure 3-8) The effect of removing the apical meristem on the relationship
between plant size (vegetative biomass) and seed mass. Dark lines represent
regression lines (±95% CI in lighter lines). Differences in seed production relative
to size are significant (p<0.05) when there is no overlap between CIs. For clarity,
individual data points are omitted.
82
3.8 Tables
Table 3-1) Our damage treatments are intended to replicate the effect of different kinds of damage caused by natural enemies. Here, each experimental treatment is compared to the natural damage it simulates and the insect species which cause that damage. The natural levels of damage are reported from our observations of experimental plants in a separate study, using plants from low-density, cleared, water-sprayed plots (see Chapter 2).
Treatment Damage Type Insect species Natural levels (from
Chapter 2, 2008 data)
Zygogramma suturalis F.
(Coleoptera: Chrysomelidae)
Systena blanda Melshimer
(Coleoptera: Chrysomelidae)
Corythucha spp.
(Heteroptera: Tingidae)
Leaf clipping
Chewing
Tarachidia spp.
(Lepidoptera: Noctuidae)
Leaves damaged:
52.4%
Meristem
removal
Stem galling Epiblema sp. Stems galled:
0.62 (±0.32) %
83
Table 3-2) ANOVA table for the linear model including plant size (stem biomass) as a continuous variable and clipping as an ordered factor. Sums of squares are partitioned into orthogonal polynomials to test the a priori hypothesis that clipping lowers reproduction relative to plant size. All three contrasts (Linear, quadratic and cubic) are shown; higher order terms are not significant. All terms are tested against the residual.
Source df MS F P
Ln(stem biomass) 1 84.45 98.3 <0.001
Damage 3 3.84 4.47 0.005
Linear 1 10.32 12.00 <0.001
Quadratic 1 0.02 0.02 0.89
Cubic 1 1.19 1.38 0.24
Ln(stem biomass)*Damage 3 3.29 3.83 0.012
Linear 1 8.08 9.39 0.003
Quadratic 1 1.72 2.01 0.16
Cubic 1 0.07 0.08 0.78
Residual 109 0.86
Table 3-3)Parameter estimates (SE) for slopes and intercepts from the linear model of log(seed mass) on log(stem mass).
Level of Intercept Slope
Control -1.93 (0.38) 1.24 (0.26) 5% -1.77 (0.26) 1.23 (0.16)
25% -1.00 (0.32) 1.04 (0.20)
75% -0.45 (0.22) 0.43 (0.19)
84
LITERATURE CITED
Agrawal A. A., P. M. Kotanen. 2003. Herbivores and the success of exotic plants: a phylogenetically
controlled experiment. Ecology Letters 6:712-715.
Agrawal A. A., P. M. Kotanen, C. E. Mitchell, A. G. Power, W. Godsoe, and J. Klironomos. 2005
Enemy release? An experiment with congeneric plant pairs and diverse above- and belowground
enemies Ecology 86: 2979-2989
Allard. 1943. The North American ragweeds and their occurance in other parts of the world. Science
98:292-294.
Bagarozzi D. A., J. Travis. 1998. Ragweed pollen proteolytic enzymes: Possible roles in allergies and
asthma. Phytochemistry 47:593-598.
Baskin J. M., C. C. Baskin. 1980. Ecophysiology of secondary dormancy in seeds of Ambrosia
artemisiifolia. Ecology 61:475-480.
Bassett I. J., C. W. Crompton. 1975. Biology of Canadian weeds 11. Ambrosia artemisiifolia L. and A.
psilostachya Dc. Canadian Journal Of Plant Science 55:463-476.
Bassett I. J., J. Terasmae. 1962. Ragweed, Ambrosia spp, in Canada and their history in post-glacial
time. Canadian Journal of Botany 40:141-150.
Bazzaz F. A. 1974. Ecophysiology Of Ambrosia artemisiifolia - Successional Dominant. Ecology
55:112-119.
Bazzaz F. A., D. M. Mezga. 1973. Primary productivity and microenvironment in an Ambrosia
dominated old-field. American Midland Naturalist 90:70-78.
Bolker, B. M. 2008. Ecological Models and Data in R. Princeton University Press
Caldwell, M. M. , J. H. Richards, D. A. Johnson, R. S. Nowak and R. S. Dzurec. 1981. Coping with
85
herbivory: Photosynthetic capacity and resource allocation in two semiarid Agropyron bunchgrasses.
Oecologia 50: 14-24
Callaway R. M., J. L. Maron. 2006. What have exotic plant invasions taught us over the past 20 years?
Trends In Ecology and Evolution 21:369-374.
Chauvel, B., F. Dessaint, C. Cardinal-Legrand and F. Bretagnolle. 2006. The historical spread of
Ambrosia artemisiifolia L. in France from herbarium records. Journal of Biogeography 33:665-673
Colautti R. I., A. Ricciardi, I. A. Grigorovich, and H. J. MacIsaac. 2004. Is invasion success explained
by the enemy release hypothesis? Ecology Letters 7:721-733.
Colautti R. I., S. A. Bailey, C. D. A. van Overdijk, K. Amundsen, and H. J. MacIsaac. 2006.
Characterised and projected costs of nonindigenous species in Canada. Biological Invasions 8:45-59.
Crawley, M. J. 2007. The R Book. Wiley
Darwin, C. 1859. On the Origin of Species by Means of Natural Selection, or, the Preservation of
Favoured Races in the Struggle for Life. 6th
edition.
David J. P., A. J. D. Santos, M. L. D. Guedes, J. M. David, H. B. Chai, J. M. Pezzuto, C. K.
Angerhofer, and G. A. Cordell. 1999. Sesquiterpene lactones from Ambrosia artemisiifolia
(Asteraceae). Pharmaceutical Biology 37:165-168.
Doak D. F. 1992. Lifetime impacts of herbivory for a perennial plant. Ecology 73:2086-2099.
Dodd, M. E. and J. Silvertown. 2000 Size-specific fecundity and the influence of lifetime size variation
upon effective population size in Abies balsamea. Heredity 85: 604-609
Elton C. S. 1958. The Ecology of Invasions by Animals and Plants. University Of Chicago Press.
Foster M. M., P. M. Vitousek, and P. A. Randolph. 1980. The effects of Ragweed (Ambrosia
86
artemisiifolia L.) on nutrient cycling in a 1st-year old-field. American Midland Naturalist 103:106-
113.
Friedman J., S. C. H. Barrett. 2008. High outcrossing in the annual colonizing species Ambrosia
artemisiifolia (Asteraceae). Annals of Botany 101:1303-1309.
Genton B. J., P. M. Kotanen, P. O. Cheptou, C. Adolphe, and J. A. Shykoff. 2005. Enemy release but no
evolutionary loss of defence in a plant invasion: an inter-continental reciprocal transplant
experiment. Oecologia 146:404-414.
Ghazoul J. 2002. Seed predators and the enemy release hypothesis. Trends In Ecology and Evolution
17:308-308.
Grimm E. C., G. L. Jacobson, W. A. Watts, and B. C. S. H. a. K. A. Maasch. 1993. A 50,000-Year
record of climate oscillations from Florida and its temporal correlation with the Heinrich Events.
Science 261:198-200.
Guo Q. 2006. Intercontinental biotic invasions: what can we learn from native populations and
habitats? Biological Invasions 8:1451-1459.
Halpern S. L., N. Underwood. 2006. Approaches for testing herbivore effects on plant population
dynamics. Journal of Applied Ecology 43:922-929.
Hawkes C. V., J. J. Sullivan. 2001. The impact of herbivory on plants in different resource conditions:
A meta-analysis. Ecology 82:2045-2058.
Hierro J. L., J. L. Maron, and R. M. Callaway. 2005. A biogeographical approach to plant invasions: the
importance of studying exotics in their introduced and native range. Journal of Ecology 93:5-15.
Igrc J., C. J. Deloach, and V. Zlof. 1995. Release and establishment of Zygogramma suturalis F
(Coleoptera, Chrysomelidae) in Croatia for control of common ragweed (Ambrosia artemisiifolia L).
87
Biological Control 5:203-208.
Keane R. M., M. J. Crawley. 2002. Exotic plant invasions and the enemy release hypothesis. Trends in
Ecology and Evolution 17:164-170.
Kiss L. 2007. Is Puccinia xanthii a suitable biological control agent of Ambrosia artemisiifolia?
Biocontrol Science and Technology 17:535-539.
Klironomos J. N. 2002. Feedback with soil biota contributes to plant rarity and invasiveness in
communities. Nature 417:67-70.
Kolar C. S., D. M. Lodge. 2001. Progress in invasion biology: predicting invaders. Trends in Ecology
and Evolution 16:199-204.
Kosola K. R., K. L. Gross. 1999. Resource competition and suppression of plants colonizing early
successional old fields. Oecologia 118:69-75.
Kotanen P. M., J. P. Rosenthal. 2000. Tolerating herbivory: does the plant care if the herbivore has a
backbone? Evolutionary Ecology 14:537-549.
Liu H., P. Stiling. 2006. Testing the enemy release hypothesis: a review and meta-analysis. Biological
Invasions 8:1535-1545.
Lodge D. M. 1993. Biological invasions: Lessons for ecology. Trends in Ecology and Evolution 8:133-
137.
Mack R. N., D. Simberloff, W. M. Lonsdale, Evans, H. and Clout, ., and F. A. Bazzaz. 2000. Biotic
invasions: Causes, epidemiology, global consequences, and control. Ecological Applications 10:689-
710.
Maron J. L., M. Vila. 2001. When do herbivores affect plant invasion? Evidence for the natural
88
enemies and biotic resistance hypotheses. Oikos 95:361-373.
Maryushkina V. Y. 1991. Peculiarities of Common Ragweed (Ambrosia artemisiifolia L.) strategy.
Agriculture Ecosystems and Environment 36:207-216.
Mitchell C. E., A. G. Power. 2003. Release of invasive plants from fungal and viral pathogens. Nature
421:625-627.
Mitchell C. E., A. A. Agrawal, J. D. Bever, G. S. Gilbert, R. A. Hufbauer, J. N. Klironomos, Maron, J.
L. and Morris,W.F., I. M. Parker, A. G. Power, E. W. Seabloom, M. E. Torchin, and D. P. Vazquez.
2006. Biotic interactions and plant invasions. Ecology Letters 9:726-740.
Nowak, R. S. and M. M. Caldwell. 1984. A test of compensatory photosynthesis in the field:
Implications for herbivory tolerance. Oecologia 61: 311-318
Quinn, G. P. and M.J. Keough. 2002. Experimental Design and Data Analysis for Biologists.
Cambridge University Press.
R Development Core Team. 2008. R: A language and environment for statistical computing. R
Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-
project.org.
Richardson, D. M. and Pysek, P. 2008. Fifty years of invasion ecology - the legacy of Charles Elton.
Diversity And Distributions 14: 161-168
Sakai A. K., F. W. Allendorf, J. S. Holt, Lodge, D. M. and Molofsky,J., K. A. With, S. Baughman,
Cabin, R. J. and Cohen, J.E., N. C. Ellstrand, D. E. McCauley, P. O'Neil, I. M. Parker, J. N.
Thompson, and S. G. Weller. 2001. The population biology of invasive species. Annual Review of
Ecology and Systematics 32:305-332.
Sokal, R. R. and F. J. Rohlf. 1994. Biometry: The Principles and Practices of Statistics in Biological
89
Research. W. H. Freeman
Stastny M., U. Schaffner, and E. Elle. 2005. Do vigour of introduced populations and escape from
specialist herbivores contribute to invasiveness? Journal of Ecology 93:27-37.
Stevens M. H. H., W. P. Carson. 1999. Plant density determines species richness along an experimental
fertility gradient. Ecology 80:455-465.
Strauss S. Y., W. Watson, and M. T. Allen. 2003. Predictors of male and female tolerance to insect
herbivory in Raphanus raphanistrum. Ecology 84:2074-2082.
Strauss S. Y., J. A. Rudgers, J. A. Lau, and R. E. Irwin. 2002. Direct and ecological costs of resistance
to herbivory. Trends in Ecology and Evolution 17:278-285.
Teshler M. P., A. DiTommaso, J. A. Gagnon, and A. K. Watson. 1981. Biological control programmes
against insects and weeds in Canada 1981-2000. Pages 290-294 In P. G. Mason and J. T. Huber,
editors. CABI Publishing, .
Tiffin P. 2000a. Are tolerance, avoidance, and antibiosis evolutionarily and ecologically equivalent
responses of plants to herbivores? The American Naturalist 155:128-138.
Tiffin P. 2000b. Mechanisms of tolerance to herbivore damage: what do we know? Evolutionary
Ecology 14:523-536.
Tiffin, P. and B.D. Inouye. 2000. Measuring tolerance to herbivory: accuracy and precision of estimates
made using natural versus imposed damage. Evolution 54: 1024-1029
Vitousek P. M., H. A. Mooney, J. Lubchenco, and J. M. Melillo. 1997. Human domination of Earth's
ecosystems. Science 277:494-499.
Weis A. E., S. J. Franks. 2006. Herbivory tolerance and coevolution: an alternative to the arms race?
90
New Phytologist 170:423-425.
Wolfe L. M. 2002. Why alien invaders succeed: support for the escape-from-enemy hypothesis. The
American Naturalist 160:705-711.
91
4 General Discussion
A reduced herbivore load following introduction is predicted to increase demographic success
of introduced plants, as predicted by the Enemy Release Hypothesis (ERH). However, studies of the
Enemy Release Hypothesis are predicated on the assumption that strong consumer effects are present at
some stage in the life cycle. In this thesis I present the results of experiments investigating the effects
of enemies on all life stages in Ambrosia artemisiifolia. This work asks two complementary questions:
first, how does enemy exclusion affect plant performance across three life history categories: seed,
seedling and adult? Second, how does leaf damage change the relationship between plant growth and
seed production in this species?
This general discussion has three parts: first, a summary of my experimental results. Second, a
discussion of their relevance to the natural history of Ambrosia in Canada, its invasion of Eurasia, and
biological invasions in general. Finally, a suggestion of interesting future research questions, including
the presentation of some suggestive ancillary results from the present study.
4.1 Summary of results The effects of enemy exclusion were in general weak in this species, as reported in Chapter 2. I
was not able to find evidence of strong enemy effects at the seed stage, while at the seedling stage such
effects were present but weak. Our pesticide spray successfully prevented insect damage and was more
effective in 2008;however in neither year did plants experience an increase in performance relative to
water controls.
Disturbance, in contrast, had a very large impact on ragweed performance. Adult plants in
disturbed areas survived better during the dry summer of 2007. The following year, though nearly all
92
plants survived, plants in disturbed plots showed greatly increased growth and reproduction relative to
plants growing among old-field vegetation.
In the tolerance experiment (Chapter 3), I found that leaf damage had no direct effects on seed
production and depressed growth very little. However, allocation to reproduction relative to stem
biomass was increased by leaf damage. This same analysis produced evidence of a small effect of
overcompensation in plants which experience light (25%) damage.
4.2 Ambrosia and invasions
4.2.1 Ragweed in North America
Disturbance is the primary influence on the population size and density of ragweed plants
(Teshler 1981, Stevens and Carson 1999). A native weed of disturbed ground in contemporary Canada,
ragweed is considered a crop weed (Bassett and Crompton 1975), a nuisance plant, and an important
source of allergenic pollen (Bagarozzi and Travis 1998). Clues to the pre-historical distribution of
ragweed are found in pollen records, which show that ragweed plants have historically lived on
disturbed sites (Grimm 1993). Lake pollen records show great fluctuation in the amount of “Ambrosia-
type” pollen throughout the Holocene, possibly in relation to disturbance caused by drought (Grimm
2001).
As European agriculture was transferred to North America, the rapid and heavy disturbance
created ideal ragweed habitat. Disturbed sites also favour the survival of seeds, as rodents do not
frequently go into open areas where they risk predation from birds (Manson and Stiles 1998). In
addition, ragweed plants have long-lived seeds (Baskin and Baskin 1980) – a common trait among
disturbance-adapted plants (Baker 1965). Seeds are not rapidly consumed by mammals and fungi,
which likely helps their seed banks build up to large sizes in cropland.
93
4.2.2 Ragweed’s Invasions
Because Ambrosia is so dependent on disturbance, it is not surprising that it is most invasive in
disturbed and agricultural land in the introduced range (Chauvel et al. 2006). However, Ambrosia is
also extremely herbivore-tolerant, which may make biological control challenging. Biocontrol agents,
such as Z. suturalis, have not been successful in Eurasia. This could be because populations have not
always established well there (Igrc 1995). However, even if large, stable populations do establish,
control efforts may prove difficult because ragweed is very tolerant to leaf damage. A successful
biocontrol agent would have to remove a great deal of leaf tissue, more than the 75% of tissue we
removed in our experiment, to reduce fecundity in ragweed.
This high level of tolerance also explains the absence of any evolutionary change in ragweed
following introduction (Genton et al. 2005): if enemies do not exert selection in the native range, then
their absence will not create a change in selection pressures in the introduced range (Franks et al.
2008).
4.2.3 Invasions in general
Disturbed sites have more invaders (Mack et al. 2000), and plants specifically adapted to
disturbance in the native range may be particularly likely to become invaders (Cadotte 2006). As well,
because disturbance creates open ground not dominated by the resident community, it can provide a
release from the ‘biotic resistance’ that introduced plants may face in their new range (Maron and Vila
2001).
The effects of disturbance can exaggerate the positive effects of enemy release (Blumenthal
2005, 2006). Blumenthal suggests that enemy release has a far greater effect when the introduced
species are adapted to rapidly use available resources. Ragweed, like many weeds of disturbed ground,
is adapted to disturbed ground with abundant resources. Our experiment could have detected this
94
synergistic effect as an interaction between disturbance and pesticide application, but no such
interaction was observed. However, it is possible that in Eurasia, ragweed is even less damaged, or that
European sites are richer in nutrients, thereby increasing the effect of enemy release to a biologically
meaningful level.
Generally, the ERH is predicated on the assumption that enemies matter. Indeed, studies of the
ERH often test hypothesis about the distribution of leaf damage, without demonstrating a fitness cost
(Maron and Vila 2001). As studies of invaders in the native and introduced ranges accumulate,
ecologists will develop a better idea of the relationship between enemy attack and population dynamics
for invasive plants. I suggest that the importance of the enemy release hypothesis has been over-
estimated.
4.3 Suggestions for future work I consider Annual Ragweed a excellent and tractable study system, which could be used in
future to continue the pattern in invasion biology of simultaneously investigating biological invasions
per se and basic ecology. There are many interesting questions remaining both within this system and
in invasion biology in general. Because each experiment I conducted raised distinct questions, this
section is again structured by experiment.
4.3.1 Seeds
Ragweed achenes have very hard coats that can not be broken easily once the seeds are mature.
Therefore, important seed predation may not occur after dispersal – rather the predispersal predation
stage may be more important. In particular, (MacKay and Kotanen 2008) found Harpalus rufipes
eating seeds before they are dropped from the plant -- interestingly, this is a European species. Eurosta
bellis was also observed consuming seeds pre-dispersal, leaving characteristic holes punched in the
seed coat. I performed some small preliminary “no-choice” trials, both with Carabids and Field
95
Crickets (Orthoptera: Gryllidae), by enclosing a sample of seeds with an insect and checking regularly.
I found it very difficult to observe Carabid predation in these artificial conditions: most of these beetles
starved, unable to break the seed coat. Crickets, on the other hand, are able to eat at least some seeds in
captivity, perhaps due to their larger size. It would be interesting to extend these trials by sampling
common omnivorous invertebrates from KSR over a summer and testing their ability to feed on the
seeds of several old-field species, including ragweed. This could establish whether ragweed is an
exceptionally tough-seeded plant, or if such durable seeds are common. Even more interesting would
be a repetition of the same study in France as well, testing the hypotheses that tougher seeds have
allowed ragweed to invade.
4.3.2 Seedlings
Ragweed plants are quite late-germinating; they may be among the last annual plants to emerge
in the spring. However, they are nevertheless often very conspicuous by the end of the summer and
have been considered ‘successional dominants’ (Bazzaz 1974, Bassett and Crompton 1975) in the early
stages of old field succession. Why is seedling germination so late? One possibility is that seedling
predation is much greater earlier in the season – this could be tested by planting out young ragweed
seedlings each week from early May to late June, and observing survival. Perhaps the phenology of
Ambrosia relative to the Eurasian flora further increases its enemy release there, avoiding important
seedling predators that might reduce populations of competitors earlier in the season.
4.3.3 Mechanisms of tolerance and allocation
The high level of tolerance in Ambrosia makes it a good system for asking questions about
mechanisms and evolution of tolerance. More detailed studies which examine the plant responses in
detail (measuring the branching pattern, photosynthetic rate, etc) are needed to unambiguously
investigate the mechanism of tolerance in a species. For example, Huhta et al. (2000), also investigated
96
several levels of clipping damage, crossing this treatment with fertilizer and competition treatments. In
addition to measuring reproduction they quantified plant architecture to test the hypothesis that release
from apical dominance was the mechanism of tolerance in this species.
I quantified seed production (female function) in my tolerance study (Chapter 3), however,
pollen dispersal (male function) is another important means of reproduction. The degrees of tolerance
to herbivory may be different between female and male function. For example, Strauss et al. (2003)
found that male and female tolerance in Raphanus raphanistrum were not correlated, although male
fitness was less variable. While measuring seed production alone is useful for predicting population
growth, measuring pollen production is also important. In particular, an important aspect of ragweed’s
impact as a weed (in Canada) and an invader (in Europe) is its production of large amounts of
windborne, highly allergenic pollen (Bagarozzi and Travis 1998). Therefore, a study which examines
herbivory on plants and pollen production would have both basic and applied significance.
In addition, I would suggest a future study to examine another possible sex-allocation response
to herbivory: plants may be using herbivory as an indicator of conspecific density. Such phenotypic
‘adjustment’ of sex ratio on the part of a growing plant has been recently demonstrated: many plant
species are capable of modifying their growth pattern when exposed to red/far red ratios indicative of
crowded conditions (Schmitt et al. 2003). Herbivores could act in a similar way as indicators, not
merely of the presence of nearby plants (as the red/far red ratio does), but of conspecific density in
particular. This would be the case if two conditions held: if the majority of damage is caused by
oligophagous or specialist consumers (as opposed to broad generalists), and if these attack plants in a
density-dependent fashion (i.e. as suggested by the Janzen-Connell mechanism, Janzen 1970, Connell
1971). This pattern of decline in attack with increasing distance from conspecifics has been
demonstrated for many tropical species (Hyatt et al. 2003) and recently for Ambrosia itself (MacKay
97
and Kotanen 2008).
Wind pollinated plants are sometimes pollen-limited, as has been demonstrated in another
anemophilous invasive, Spartina alterniflora (Davis et al. 2004). In fact, isolated Ambrosia plants set
less seed than plants in dense populations; ragweed plants are self-incompatible (Friedman and
Barrett). If pollen limitation were a problem for ragweed, then isolated individuals may be able to
increase fitness by producing more male flowers, as female flowers are likely to remain un-pollinated.
Conversely, plants near to conspecifics would not be pollen-limited. If the amount of herbivore damage
is a reliable signal of conspecific density, this may provide useful ‘cues’ to plants. Such plasticity in
within-plant sex ratios is well-documented in this species: the within-plant sex ratio of Ambrosia
becomes increasingly female under light limitation, and increasingly male under nutrient limitation
(Paquin and Aarssen 2004).
While my experiments were not designed to test these hypotheses, a rough estimate is still
possible. By weighing the clipped leaves from all plants (see Chapter 3) we were able to estimate the
relationship between leaf number and leaf mass. Then, the mass of all male flowers can be determined
by:
Male flowers = total biomass – stem(g) – seed(g) – predicted leaves (g)
where leaf mass is predicted using a count of leaves at the end of August. Changes in the
relative allocation of biomass to male vs. female function can then be investigated with an ANCOVA-
style plot (as was used in Chapter 3). The results are shown in Figure 1.
Examining the slope values in this relationship (inset) shows that ragweed produces more seed
biomass than flower biomass, and that this relationship increases with clipping damage. This pattern is
consistent with the ‘cheap pollen’ but not with my suggested mechanism and is consistent with the
98
results of (Strauss et al. 2003). However, this approach is only a rough estimate and, in particular, the
comparison is more appropriate between the mass of female flowers and male flowers. Nevertheless,
this data highlights the need to consider both male and female function when assessing reproductive
output.
4.3.4 Indirect effects of tolerance
It is possible that introduced species, when they reach high population sizes in their new range,
are able to exert indirect interactions on the community and so increase their invasiveness. It has been
suggested (Ghazoul 2002) that species which have very large floral displays are over-attractive to
pollinators, and so reduce the reproduction of native plants. Similar interactions are possible for very
tolerant species. If a plant is very tolerant, it suffers less limitations of its population growth as a result
of consumer pressure. However, herbivores of tolerant plants may spill over to neighbours. Thus,
tolerance adaptations may confer an advantage on an invading plant not only by magnifying the effect
of enemy release, but also by increasing the amount of top-down control on neighbouring plants.
99
0.01 0.05 0.50 5.00
5e−
03
5e−
02
5e−
01
5e+
00
Mass of seeds produced (g)
Estim
ate
d m
ale
flo
wer
mass (
g)
Control
5%
25%
75%
Control 1/20 1/4 3/4
Slo
pe
0.0
0.5
1.0
1.5
2.0
Figure 4-1) A rough estimate of male flower mass compared to the mass of seeds
produced for each plant in the Tolerance experiment (details in Chapter 3). The
allocation to male function over female function increases with increasing damage
(inset).
100
LITERATURE CITED
Bagarozzi D. A., J. Travis. 1998. Ragweed pollen proteolytic enzymes: Possible roles in allergies and
asthma. Phytochemistry 47:593-598.
Baker H. G. 1965. Chacteristics and modes of origin of weeds. Pages 147-172 In H.G Baker and G.L.
Stebbins, editors. Academic Press, New York, NY, .
Baskin J. M., C. C. Baskin. 1980. Ecophysiology of secondary dormancy in seeds of Ambrosia
artemisiifolia. Ecology 61:475-480.
Bassett I. J., C. W. Crompton. 1975. Biology of Canadian weeds 11. Ambrosia artemisiifolia L. and A.
psilostachya Dc. Canadian Journal Of Plant Science 55:463-476.
Bazzaz F. A. 1974. Ecophysiology Of Ambrosia artemisiifolia - Successional Dominant. Ecology
55:112-119.
Blumenthal D. 2005. Interrelated causes of plant invasion. Science 310:243-244.
Blumenthal D. M. 2006. Interactions between resource availability and enemy release in plant invasion.
Ecology Letters 9:887-895.
Cadotte M. W., B. R. Murray, and J. Lovett-Doust. 2006. Ecological patterns and biological invasions:
Using regional species inventories in macroecology. Biological Invasions 8:809-821.
Chauvel, B., F. Dessaint, C. Cardinal-Legrand and F. Bretagnolle. 2006. The historical spread of
Ambrosia artemisiifolia L. in France from herbarium records. Journal of Biogeography 33:665-673
Connell, J.H. 1971. On the role of natural enemies in preventing competitive exclusion in some marine
animals and in rain forest trees. Dynamics of populations: Proceedings of the advanced study institute
on dynamics of numbers of populations (eds P.J. Den Boer & G. Gradwell), pp. 298–312. Center for
Agricultural Publishing and Documentation, Wageningen.
101
Franks S. J., P. D. Pratt, F. A. Dray, and E. L. Simms. 2008. Selection on herbivory resistance and
growth rate in an invasive plant. American Naturalist 171:678-691.
Genton B. J., P. M. Kotanen, P. O. Cheptou, C. Adolphe, and J. A. Shykoff. 2005. Enemy release but no
evolutionary loss of defence in a plant invasion: an inter-continental reciprocal transplant
experiment. Oecologia 146:404-414.
Ghazoul J. 2002. Seed predators and the enemy release hypothesis. Trends In Ecology and Evolution
17:308-308.
Grimm E. C., G. L. Jacobson, W. A. Watts, and B. C. S. H. a. K. A. Maasch. 1993. A 50,000-Year
record of climate oscillations from Florida and its temporal correlation with the Heinrich Events.
Science 261:198-200.
Grimm, E.C. 2001. Trends and palaeoecological problems in the vegetation and cliamate history of the
northern Great Plains, USA. Biology and Environment: Proceedings of the Royal Irish Academy
101:47-64
Huhta, A-P, K. Hellstrom, P. Rautio and J. Tuomi. 2000. A test of the compensatory continuum:
fertilization increases and below-ground competition decreases the grazing tolerance of tall
wormseed mustard (Erysimum strictum)
Hyatt, L.A., Rosenberg, M.S., Howard, T.G., Bole, G., Fang, W. & Anastasia, J. 2003. The distance
dependence prediction of the Janzen-Connell hypothesis: a meta-analysis. Oikos, 103, 590–602.
Igrc J., C. J. Deloach, and V. Zlof. 1995. Release and establishment of Zygogramma suturalis F.
(Coleoptera, Chrysomelidae) in Croatia for control of common ragweed (Ambrosia artemisiifolia L).
Biological Control 5:203-208.
Janzen, D.H. 1970. Herbivores and the number of tree species in tropical forests. The American
102
Naturalist, 104, 501–528.
Mack R. N., D. Simberloff, W. M. Lonsdale, Evans, H. and Clout,M., and F. A. Bazzaz. 2000. Biotic
invasions: Causes, epidemiology, global consequences, and control. Ecological Applications 10:689-
710.
MacKay J., P. M. Kotanen. 2008. Local escape of an invasive plant, common ragweed (Ambrosia
artemisiifolia L.), from above-ground and below-ground enemies in its native area. Journal of
Ecology 96:1152-1161.
Manson R. H., E. W. Stiles. 1998. Links between microhabitat preferences and seed predation by small
mammals in old fields. Oikos 82:37-50.
Maron J. L., M. Vila. 2001. When do herbivores affect plant invasion? Evidence for the natural
enemies and biotic resistance hypotheses. Oikos 95:361-373.
Paquin V., L. W. Aarssen. 2004. Allometric gender allocation in Ambrosia artemisiifolia (Asteraceae)
has adaptive plasticity. American Journal of Botany 91:430-438.
Schmitt, J., J. R. Stinchcombe, M. S. Heschel and H. Huber. 2003. The adaptive evolution of plasticity:
phytochrome-mediated shade avoidance responses. Integr. Comp. Biol. 43: 459-469
Stevens M. H. H., W. P. Carson. 1999. Plant density determines species richness along an experimental
fertility gradient. Ecology 80:455-465.
Strauss S. Y., W. Watson, and M. T. Allen. 2003. Predictors of male and female tolerance to insect
herbivory in Raphanus raphanistrum. Ecology 84:2074-2082.
Teshler M. P., A. DiTommaso, J. A. Gagnon, and A. K. Watson. 1981. Biological control programmes
against insects and weeds in Canada 1981-2000. Pages 290-294 In P. G. Mason and J. T. Huber,
editors. CABI Publishing.
