Quantifying interspecific competition effects of Herb Robert (Geranium robertianum) on three native Western Washington forbs

Item

Title
Quantifying interspecific competition effects of Herb Robert (Geranium robertianum) on three native Western Washington forbs
Date
2021
Creator
Larson, Sarah
Identifier
Thesis_MES_2021_LarsonS
extracted text
QUANTIFYING INTERSPECIFIC COMPETITION EFFECTS OF HERB ROBERT
(GERANIUM ROBERTIANUM) ON THREE NATIVE WESTERN WASHINGTON FORBS

by
Sarah L. Larson

A Thesis
Submitted in partial fulfillment
of the requirements for the degree
Master of Environmental Studies
The Evergreen State College
December 2021

© 2021 Sarah L. Larson. All rights reserved.

This Thesis for the Master of Environmental Studies Degree
by
Sarah L. Larson

has been approved for
The Evergreen State College
by

________________________________
Sarah Hamman, Ph.D.
Member of the Faculty

12/10/2021
_____________________________
Date

ABSTRACT

Quantifying interspecific competition effects of Herb Robert (Geranium robertianum) on three
native western Washington forbs
Sarah L. Larson

Invasive species pose a threat to ecosystems and biodiversity around the world. While many
hypotheses have been offered to explain why some species successfully invade, they tend to be
broad and inadequate at explaining invasiveness for any one species. Reductionist studies can
reveal patterns that are species- and habitat-specific that improve our understanding of
invasiveness at a local level. Geranium robertianum L. is an annual/biennial Eurasian herb that
has invaded western Washington and continues to spread across the Pacific Northwest coast. Its
ability to invade intact forest understory plant communities makes this invasive species
somewhat distinctive among invasive plants and concerning for native biodiversity. In order to
determine the interspecific competition effect of G. robertianum, a greenhouse study was
conducted between G. robertianum and three native forbs: Geum macrophyllum, Tellima
grandiflora, and Dicentra formosa, all herbaceous perennials. Each native species was grown
with and without G. robertianum. The plants grown with G. robertianum had two separate soil
treatments in addition to one group without soil treatments. Activated charcoal was added to
assess the potential allelopathic interference of G. robertianum and native mineral soil was added
to assess the effect the soil microbial community has on the competitive ability of the native
plants. After 12 weeks, plants were harvested, dried, separated into above- and belowground
components, and weighed. G. macrophyllum experienced the highest level of interspecific
competition by G. robertianum but the effect was ameliorated by the native soil treatment. None
of the treatments significantly affected the growth of T. grandiflora, which appears to be a robust
competitor to G. robertianum. The charcoal treatment produced a generally negative effect for
all species, indicating that allelopathic interference is not a primary invasive mechanism for G.
robertianum. However, patterns in above- and belowground biomass for T. grandiflora and two
of the G. robertianum groups showed some allelopathic interference but lacked statistical
significance. This suggests that allelopathy is not the primary invasive strategy used by G.
robertianum but potentially contributes a weak effect in conjunction with other above- and
belowground invasive mechanisms.

Table of Contents
List of Figures ............................................................................................................................... vi
List of Tables ............................................................................................................................... vii
Acknowledgements .................................................................................................................... viii
Introduction ................................................................................................................................... 1
Literature Review ......................................................................................................................... 4
Introduction ............................................................................................................................... 4
Invasion Biology ........................................................................................................................ 4
Invasive Species ...................................................................................................................... 4
Invasion Hypotheses ................................................................................................................ 7
Geranium robertianum, L. (Herb Robert) ............................................................................. 16
Distribution ............................................................................................................................ 16
Description............................................................................................................................. 17
Habitat ................................................................................................................................... 18
Phenology and Reproduction................................................................................................. 19
Physiology ............................................................................................................................. 21
Materials & Methods .................................................................................................................. 23
Species Selection ....................................................................................................................... 23
Experimental Design ................................................................................................................. 26
Data analysis ............................................................................................................................. 30
Results .......................................................................................................................................... 31
Geum macrophyllum ................................................................................................................. 31
Tellima grandiflora ................................................................................................................... 32
Dicentra formosa....................................................................................................................... 32
Geranium robertianum .............................................................................................................. 36
Discussion..................................................................................................................................... 39
Geum macrophyllum ................................................................................................................. 39
Tellima grandiflora ................................................................................................................... 41
Dicentra formosa....................................................................................................................... 43
Geranium robertianum .............................................................................................................. 45
Conclusion ................................................................................................................................... 49
References .................................................................................................................................... 52
iv

Appendices ................................................................................................................................... 60
Appendix A .............................................................................................................................. 60
Appendix B .............................................................................................................................. 62
Appendix C .............................................................................................................................. 63

v

List of Figures
Figure 1. Schematic diagram of the PAB relationship of invasion ................................................ 8
Figure 2. Distribution of Geranium robertianum in Washington state........................................ 16
Figure 3. Photograph of Geranium robertianum ......................................................................... 18
Figure 4. Ecoregions of Washington state ................................................................................... 23
Figure 5. Photographs of native species used in experimental design: Geum macrophyllum,
Tellima grandiflora, and Dicentra formosa .................................................................................. 25
Figure 6. Photographs of the full setup of all species and treatment ........................................... 29
Figure 7. Photograph of Geum macrophyllum Control group and Geum macrophyllum Carbon
group after ~10 weeks of growth .................................................................................................. 31
Figure 8. Effect of treatments on native biomass for Geum macrophyllum and Tellima
grandiflora .................................................................................................................................... 32
Figure 9. Effect of treatments on biomass for Dicentra formosa ................................................ 33
Figure 10. Effect of treatments on biomass for Geranium robertianum ..................................... 38
Appendix C, Figure 1. Effect of treatments on biomass of all three native species. .................. 63

vi

List of Tables
Table 1. Primary invasion hypotheses relevant to the success of Geranium robertianum as an
invasive species in western Washington. ........................................................................................ 9
Table 2. Summary of treatment groups by species with shorthand ............................................. 27
Table 3. Analysis of variance (ANOVA) results for biomass of native species for all treatment
groups ............................................................................................................................................ 34
Table 4. Pairwise comparison of treatment groups; post hoc for ANOVA & Tukey’s HSD and
Kruskal-Wallis & Dunn’s test ....................................................................................................... 35
Table 5. Analysis of variance (ANOVA) results for biomass of Geranium robertianum by native
species group for all treatment groups .......................................................................................... 37
Appendix A, Table 1. Data analysis summary for native species by response variable ............. 60
Appendix A, Table 2. Data analysis summary for Geranium robertianum by response variable
....................................................................................................................................................... 61
Appendix B, Table 1. Mean aboveground (AG), belowground (BG), total biomass, aboveground
to belowground ratio (AG:BG) for each treatment group by species ........................................... 62

vii

Acknowledgements

A thesis does not happen in a vacuum. This final product was only possible with the help
and support from many others. I would like to thank my thesis reader, Sarah Hamman, for her
patience, guidance, and positivity throughout the entire process. I am indebted to my friend
Adam Martin for his invaluable help in producing beautiful plots out of my data. I am also
grateful for Claire Olson who shared her love and knowledge of botany and supported me
through what were some challenging years. She went above and beyond to help me throughout
this process: scrubbing nursery pots outside in the freezing cold, scouting good locations to
collect herb Robert, and even picked up some of the native plants for me in Bellingham. This
thesis process would have been far harder and lonelier without her help and friendship. This
journey was, at times, difficult for my children and I can only hope that someday they will
understand why I chose this path and that it wasn’t just for me, but also for them. I also owe my
mother a heartfelt thank you for uprooting her life for my benefit.
Finally, I would be remiss to not acknowledge the physical and emotional labor that
nursery workers put into growing the plants I used for my thesis. The availability of native plants
is a crucial component to conservation and restoration efforts and I am grateful that there are
people who dedicate the time and energy to grow native plants. I would also like to
acknowledge, specifically, the incarcerated technicians at the Washington Corrections Center for
Women in Gig Harbor for their dedication and enthusiasm for the native plant conservation
nursery operated in partnership with the Sustainability in Prisons Project.

viii

Introduction

The global redistribution of non-native plant species by humans is an ever-increasing
dilemma that can have profound consequences for ecosystems and biodiversity (Kueffer, 2017).
The costs can be great, both in terms of ecological and economic damage. Often cited as the
second leading cause of biodiversity loss and species extinctions, invasive species ultimately
become drivers that reshape biotic communities and alter ecosystems (Wilcove et al., 1998;
Catford et al., 2009; Jose et al., 2013). There are a handful of invasive plant species that cause a
disproportionate amount of harm and therefore generate ample amounts of research. This leaves
a paucity of research on species that are perhaps not yet widespread in their new range or the
magnitude of their impacts is unknown. Invasive plants species that disrupt, damage, or
otherwise harm agriculture are often give the greatest amount of attention and funding, as the
economic impacts tend to be more quantifiable and are not always synonymous to the ecological
impacts (Pyšek & Richardson, 2006).
As invasion success is likely context- and species-dependent and due to a wide variety of
factors and mechanisms, a robust and comprehensive theoretical framework is necessary to
advance our understanding of what makes invasive species so successful (Parker et al., 1999;
Catford et al., 2009). While there is likely no “holy grail” mechanism or equivalent hypothesis
for successful invasion, small-scale studies can improve our understanding of which mechanism
and factors contribute to success for any given species in an ecosystem-specific context (Davis et
al., 2000; Catford et al., 2009). The aim of this study is just that – a reductionist examination of
Geranium robertianum (herb Robert), a successful invader in western Washington. This plant is
not problematic for agriculture and has not cost millions of dollars in attempts to eradicate. So

1

why does it matter? It could be argued that its impacts to biodiversity justify exploring this
question and it is with this lens that this thesis evaluates the effects of G. robertianum on the
growth of native species found in forest understory plant communities. Pacific Northwest
forested habitats are generally resistant to invasive plant species, but they are readily invaded by
G. robertianum, where it is able to flourish and dominate the plant community.
There are several characteristics that are attributed to the success of G. robertianum as an
invasive species: self-pollination, abundant seed production, not palatable to herbivores, and high
ecological tolerance – able to grow in both full shade and full sun (Bertin, 2001; Barndt, 2008).
Additionally, the role of allelopathy – the release of chemical compounds potentially harmful to
other plants – is still under consideration as a primary mechanism for invasion success. The
intent of this thesis is to address these mechanisms and gain a better understanding of whether
allelopathy is indeed a primary mechanism utilized by G. robertianum. Specifically, the
following research questions were posed: What are the effects of G. robertianum on the growth
of native species? Is there evidence of allelopathic interference?
These questions were addressed using several different experimental treatments,
specifically, activated charcoal and native mineral soil added to an artificial potting mix.
Activated charcoal is efficient at absorbing biochemical compounds and is often used in
experiments assessing allelopathy (Del Fabbro et al., 2014). The addition of activated charcoal to
the potting mix as a treatment group assesses the potential allelopathic effect of G. robertianum
on the growth of the native species. Similarly, the relationship between plants and the soil
microbial community is integral to growth and development. The addition of native mineral soil
to the potting mix as a treatment group assesses whether the native species were given a
competitive advantage over the invasive species. The native species grown in the presence of G.

2

robertianum without the addition of either soil amendment serve as a reference to gauge the
effect of interspecific competition by this successful invasive species.

3

Literature Review
Introduction
Throughout this literature review, I will synthesize the relevant academic literature and
prevailing themes in invasion biology. The first section of this literature review will provide an
overview of invasive species and the discipline of invasion biology. Specifically, this section will
describe what an invasive species is and their impacts on ecosystems, as well as a brief overview
of the history of invasion biology. The second section will provide a holistic framework for
invasion hypotheses that attempts to capture all the factors and mechanisms that contribute to
successful invasions. The invasion hypotheses that are most relevant to the success of Geranium
robertianum as an invasive species will be discussed more in depth. Finally, the last section will
provide an overview of Geranium robertianum – its distribution, biology, and ecology. The
literature on G. robertianum is limited and this review is based on a handful of comprehensive
sources. Research on the role of this species as an invader is limited as well, particularly in
western North America where it continues to spread. The aim of my research is to contribute
quantitative data to a relatively small body of knowledge on this invasive species.
Invasion Biology
Invasive Species
Most nonnative species are introduced by humans either intentionally or unintentionally
(Simberloff, 2013; Jose et al., 2013). Intentional introductions tend to include species used for
agriculture or horticulture; unintentional introductions ironically can result from the intentional
ones – garden escapees, contaminated soil, shipping. However, not all nonnative species are
considered invasive – under optimum conditions some may develop traits that enable them to
outcompete and eventually dominate recipient communities, thus becoming an invasive species
4

(Jose et al., 2013). The definition of invasive also differs depending on who is referring to the
species in question. Policymakers generally consider a species invasive if it causes negative
impacts to the environment, human health, or the economy. In his extensive treatise on invasive
species, Simberloff (2013) argues this is in stark contrast to biologists who consider a species’
evolutionary history in determining invasiveness: where did it evolve and was its transportation
aided by humans? Once established, an invasive species can be conveyed and spread by human
activities and disturbances, especially as our modern globalized economy facilitates the largescale redistribution of plant species (Kueffer, 2017).
There are a few invasive species that have caused a disproportionate amount of
ecological alteration and extinctions: feral hogs, rats, and predatory snakes are common
examples, especially in isolated island ecosystems (Gurevitch & Padilla, 2004). Some
researchers argue that most invasive plant species are likely to alter plant community
composition or cause displacement, rather than extinctions (Sax & Gaines, 2003; Gurevitch &
Padilla, 2004). Several widely known and studied invasive plant species impact North American
ecosystems, costing millions of dollars annually in control efforts: English ivy (Hedera helix),
kudzu (Pueraria spp.), garlic mustard (Alliaria petiolata), cheatgrass (Bromus tectorum), and
knapweeds (Centaurea spp.). Knapweed alone occupies more than 7 million acres, negatively
impacting grazing for livestock and wildlife (Callaway & Ridenour, 2004). The impacts to
recipient ecosystems by these invasive plant species, and others, include alterations to nutrient
cycling, natural succession patterns, soil chemistry, water availability, herbivory patterns, and
natural fire regimes (Jose et al., 2013; Simberloff, 2013).
Despite Charles Eton’s 1958 seminal work on invasive species, little attention was paid to
invasives species by the scientific community in the mid-1900s. However, by the 1980s, with
5

increasing awareness of problems associated with invasive species, invasion biology emerged as
an independent discipline (Simberloff, 2013). Researchers continually build and modify
frameworks for understanding the many aspects of invasives: what allows invasives to be so
successful? What are their impacts? How do we manage them? Researchers continue to discuss
the impacts of invasive species, often without a solid or specific definition of “impact” (Parker et
al., 1999). Authors tend to use the term equally for different plants with varying degrees of
effects on recipient ecosystems. Parker et al. (1999) suggests quantifying impact by looking at
range, abundance, and the per-biomass effect of the invader. However, the issue of quantifying
the impacts of non-natives/invasives is further confounded by the presence of multiple invasives,
as well as other environmental stressors, such as anthropogenic disturbances and climate change.
Additionally, the interactions of multiple invasives – invasional meltdown – may impact
ecosystems on a larger scale than what would be expected had their impacts been measured
individually (Simberloff & Von Holle, 1999; Von Holle, 2011). Biotic and abiotic factors –
pollinators, seed-dispersers, herbivory pressures, soil types, and fire regimes – in conjunction
with anthropogenic stressors can influence the degree of impact (Catford et al., 2009).
As Parker et al. (1999) point out in their attempt at formalizing a framework for invasive
species, the problem with too many generalizations around impacts is more than an academic
problem: it is ultimately a problem for management efforts. Distinguishing the effects of
invasives as minor or major is necessary in order to prioritize management efforts. A common
metric used by managers for assessing the impact of invasives species is to assess their economic
cost in terms of damage or eradication efforts (Parker et al., 1999). This method can be limited
and myopic in nature, focusing primarily on human needs, favoring short-term outcomes over
long-term consequences and ignoring ecosystem function as a factor. Quantifying the value of

6

ecosystem function remains problematic, as it also requires placing a human-based measure of
value on a non-human system. Barney et al. (2013) argue that the focus should come from a
conservation and ecosystem protection perspective. While the economic consequences of
invasives should not be ignored or underrated, neither should the ecological consequences.
Invasion Hypotheses
Dozens of hypotheses have been presented to explain, at least partially, why some species
are so successful in recipient ecosystems. Many of these hypotheses focus on either the attributes
of the invader or attributes of the recipient community, while struggling to incorporate the two
(Pyšek & Richardson, 2006; Crystal-Ornelas & Lockwood, 2020). The simplest and broadest of
them hypothesizes that successful invaders possess attributes that give them an edge over native
communities, that they are overall competitively superior (Lewis & Jerde, 2012). This broad
hypothesis performs more as a definition of an invasive species versus a theory to describe why.
Because successful invasion is so context-dependent, the vast majority of invasive species are
likely successful due to different combinations of biotic and abiotic factors and thus, are not
easily quantifiable under one broad invasion hypothesis (Catford et al., 2009). Indeed, it likely
requires multiple small-scale hypotheses to begin to describe a successful invasion versus the
“holy grail” approach to explaining and predicting invasiveness which has resulted in the dozens
of disparate hypotheses (Catford et al., 2009; Crystal-Ornelas & Lockwood, 2020).
Catford et al. (2009) describe a broad, overarching framework for successful invasion
that is structured around plant reproduction, or propagule pressure, and abiotic and biotic
characteristics of the recipient plant community and invading species (PAB). This framework
cleverly unifies all of the disparate hypotheses into a top-down approach that enables
reductionist, or small-scale, experiments to hone in on the number of potential mechanisms

7

contributing to successful invasion (Pyšek & Richardson, 2006; Catford et al., 2009). The three
components represent the most fundamental characteristics necessary for the success of invaders
in novel ecosystems: (P) successful and sufficient propagation across temporal and spatial scales;
(A) hospitable abiotic environmental characteristics of an invaded site; and (B) species-specific
and ecology-evolutionary interactions across biotic communities (Figure 1). An important
component of this framework is the inclusion of anthropogenic interference as a key driver of
invasions.

H

P
B
H

I

Figure 1. Schematic diagram
illustrating the relationship
between propagule pressure
(P), abiotic characteristics
(A), and biotic characteristics
(B). Degree of interference
by humans (H), solid line
greater interference than
dashed line. Larger circles
represent greater extent of
influence. This interaction
drives invasion (I). Adapted
from Catford et al., 2009.

A
H

Of the dozens of invasion hypotheses situated in the PAB framework, 12 standout as
plausible mechanisms contributing to the success of Geranium robertianum as an invasive
species. Of these theories, many of them focus primarily on biotic factors of successful invasion.

8

Due to the complexity of biological and ecological systems, many of these theories possess
similarities and overlap. The 6 most relevant to this thesis are described below (Table 1).
Table 1. Primary invasion hypotheses relevant to the success of Geranium robertianum as
an invasive species in western Washington.
Hypothesis
Propagule pressure*
Sampling
Ideal weed*
Enemy Release
Enemy of my enemy
Biotic indirect effects*
Invasional meltdown*
Adaptation*
Novel weapons*
Habitat filtering
Disturbance
Naturalization

Human
Interference
++
+
+

Propagule
Pressure
++
+
+

Abiotic
Factors

+
+
++

+
+

++
++
++

Biotic
Factors
+
++
++
++
++
++
++
+
++
+
+
++

Symbology: ++ major driver of invasion; + secondary driver influencing invasion; * theories discussed in
more depth in this thesis. Adapted from Catford et al. 2009.

Propagule pressure, specifically the number of propagules reaching a new location,
serves as a key factor in the success of invasive species. Propagule pressure acts like a filter for
invasion – without sufficient propagation at sufficient frequency, establishment cannot occur. So
important to successful invasion, many invasion hypotheses automatically consider it to be a
requirement of invasion versus simply a driver (Catford et al., 2009). In order to serve as a
driver, the plant must already have a large amount of its resources allocated to reproductive
output (Leishman & Harris, 2011). Reproductive output can effectively be increased when a
plant is released into a novel ecosystem and freed from natural enemies. This concept overlaps
with the Enemy Release theory, which refers to the reallocation of resources that were previously
used for defense to increased growth and reproduction (Catford et al, 2009; Leishman & Harris,
9

2011). A study by Mason et al. (2008) found that, on average, invasive species produced more
seeds than native species – upwards of seven times more seeds per individual per year. While the
authors did not determine a mechanism responsible for the difference in seed production, it does
point to the possible reallocation of resources as proposed by the Enemy Release theory. An
additional concept related to propagule pressure pertains to the role of seed banks for successful,
long-term invasion. Seed banks enable plants to spread the risk of germination and establishment
over time periods greater than a single growing season, assisting in the stabilization of
population dynamics and aiding in recovery after disturbances. Many successful North American
invaders, including purple loosestrife (Lythrum salicaria) and garlic mustard (Alliaria petiolata),
have long-lived seeds that can dominate soil seed banks (Leishman & Harris, 2011; Jose &
Holzmueller, 2013). This serves both as a mechanism to outcompete native species, but also
makes long-term control of invasives incredibly challenging.
Many of the characteristics that contribute to successful invasion overlap under the Ideal
Weed hypothesis. This early hypothesis focuses on the traits of plants only and has limited
success in predicting invasiveness (Pyšek & Richardson, 2006). The plant traits associated with
invasiveness – high and early seed production, high phenotypic plasticity, rapid growth, and
small seed size – are all biotic drivers of successful invasion. However, this hypothesis ignores
the abiotic factors of successful invasion, making Ideal Weed an incomplete, albeit important,
hypothesis. Similarly focusing on biotic factors, the Biotic Indirect Effects hypothesis looks
specifically at interactions that cascade through plant communities – the effects one species has
on a second, which in turn effects a third (Callaway et al., 2004; Catford et al., 2009). These
interactions are more holistic than the Ideal Weed hypothesis, expanding the biotic factors

10

beyond a single species. If expanded to include related interactions with soil biota, these two
hypotheses cover a wide breadth of biotic factors that contribute to successful invasion.
Interactions between the invader and recipient community can influence overall invasive
success, where strong competitive interactions, such as beneficial mutualistic relationships and
novel pathogens can reduce the likelihood of successful invasion (Lewis & Jerde, 2012). Where
strong competitive interactions are lacking, an overall reduction in species richness can be a
consequence of invasion. This in turn increases the likelihood of further invasions. This
mutualistic and facilitative interaction between successive invaders is the basis for the Invasional
Meltdown hypothesis (Von Holle, 2011). These interactions can occur over a range of trophic
levels with the potential to lead to synergistic impacts on the recipient community. The
distinction between this hypothesis and the similar Biotic indirect effects hypothesis is that the
impacts of invasive species can be greater than the sum of their individual impacts (Von Holle,
2011; Braga et al., 2018). A somewhat recent and alarming example of invasional meltdown is
the introduction and subsequent spread of the spotted lanternfly (Lycorma delicatula), a
planthopper native to northern China, whose preferred host is the highly invasive Ailanthus
altissima (‘tree of heaven’). First introduced to the United States in the mid-1800s as an
ornamental specimen, A. altissima has since spread to nearly all 50 US states and is considered
one of the worst invasive plant species in North America (USDA NRCS Plant Database, 2021).
The establishment of this invasive plant species is facilitating the spread of L. delicatula, which
was first detected in Pennsylvania in 2014 and has the potential to cause substantial damage to
agriculture, particularly grapes (Vitis spp.) and apples (Malus spp.) (Dara et al, 2015; WA
NWCB, 2021).

11

It was previously proposed by Darwin (1859) that introduced plant species would be less
likely to establish themselves in places with pre-existing congeneric native species due to similar
resource requirements (Duncan & Williams, 2002). However, a study focusing on invasive
species in New Zealand found the opposite to be true – the naturalization rate was much higher
among introduced genera with pre-existing native species. Modeling based on plant families with
the highest number of non-native species revealed that genus was a highly significant predictor
of invasive probability for an introduced species (Duncan & Williams, 2002). The Adaptation
hypothesis is based on this concept and suggests that invasive species may be successful in a new
range given shared competitive abilities and traits to native congeneric species (Catford et al.,
2009). While this study is based on concepts that can be related to island endemism and thus
places at especially high risk of invasion, it makes an important case for considering congeneric
species as pre-adapted invaders.
Del Fabbro et al. (2013) and others argue that allelopathy is one of several mechanisms
that contribute to the success of invasive species (Callaway & Ridenour, 2004; Thorpe et al.,
2009; Callaway, 2011). Allelopathic chemical compounds are secondary metabolites and,
therefore, not required by the plant for metabolism – growth, development, or reproduction
(Inderjit, 2011). First described in 1937, allelopathy is not a new concept. However, it was
Callaway and Ridenour (2004) who first made the connection between allelopathy and invasive
species, developing what they termed the ‘Novel Weapons Hypothesis’ in an attempt to explain
why some invasive species are so successful in recipient ecosystems. It is theorized that
allelopathic compound-producing plants can quickly dominate plant communities that possess no
adaptation or tolerance to the compounds, effectively suppressing the germination and growth of
native plants (Callaway & Ridenour, 2004; Del Fabbro et al., 2013).

12

Overall, the evidence for this hypothesis is largely based on greenhouse and laboratory
studies. How applicable or relevant greenhouse allelopathic studies are to real-world scenarios
has been debated. An allelopathic study on knotweed (Fallopia x bohemica) found that
allelopathic potential and impact was directly related to the physical and chemical properties of
the soil; essentially the discrepancy was between artificial soil mixes that are often used in lab
and greenhouse experiments and native soil (Parepa & Bossdorf, 2016). The higher pore space
and permeability of artificial soil mixes may contribute to higher levels of allelopathic effect
compared to native soils. Furthermore, unless native soils that are used for lab or greenhouse
studies is sterilized, the presence of soil biota could contribute to erroneous study results.
However, field evidence of allelopathy does exist. Thorpe et al. (2009) produced strong
evidence for the hypothesis in their field study of Centaurea stoebe (spotted knapweed), an
aggressive Eurasian invader of North American rangelands. Allelopathic effects were assessed in
experimental in situ plots in both native and invaded ranges. Interestingly, they demonstrated
that secondary metabolites produced by C. stoebe reduced growth of native plant species in
Montana, but had no effect on plants in its native range in Romania (Thorpe et al., 2009).
However, they do point out that soil chemistry likely plays an important role in the retention and
activity of the secondary metabolites produced by C. stoebe, which requires further study. In a
laboratory study, Inderjit et al. (2009) demonstrated similar results using Nicotiana attenuata
(wild tobacco) seedlings. A portion of the N. attenuata seedlings were genetically modified to be
‘silent’ – unable to synthesize or release specific secondary metabolites. It was demonstrated in
this experiment that the release of secondary metabolites in unsilenced seedlings did have an
influence on neighboring seedlings, whereas the silenced seedlings did not.

13

The Novel Weapons hypothesis is not without criticism and skepticism. It is important to
point out that allelopathy is a conditional and species-specific effect and not a mechanism used
by all successful invasive plant species (Inderjit et al., 2009; Thorpe et al., 2009). Additionally,
soil chemistry is an increasingly an important component for this mechanism that is still understudied. How secondary metabolites respond to specific soil chemistries and regional climate is
still being debated. Researchers have pointed out that one of the secondary metabolites of
interest, (±)-catechin, degrades more quickly in wet soils compared to dry soils, putting into
doubt its role as an allelopathic compound that explains the success of Centaurea stoebe (Blair et
al., 2006). However, other research showed the (±)-catechin may be benefiting C. stoebe in terms
of competition for nutrients, specifically phosphorous: C. stoebe contained twice the
phosphorous levels in comparison to neighboring native species (Thorpe et al., 2006). This
difference was six times greater in areas with very low soil phosphorus availability. A
greenhouse study on the same species demonstrated that it could also obtain up to 15% of its
carbon from a native grass (Festuca idahoensis) through mycorrhizal networks (Carey et al.,
2004).
Despite somewhat conflicting conclusions in these studies, they demonstrate the
complexity of the direct and indirect effects plants have on soil chemistry and microbial
communities (Weidenhamer & Callaway, 2010). It is estimated that mycorrhizae are responsible
for up to 75% of all phosphorus taken up annually by terrestrial plants worldwide (van der
Heijden et al., 2008). Because of the reliance plants have on mycorrhizal fungi symbiosis to
collect and deliver phosphorous and other nutrients, disruption to these networks by invasive
plants and allelopathic compounds can significantly transform ecosystem structure and function
(Reinhart & Callaway, 2006; Hagan & Shibu, 2013). Over time these changes can create positive

14

feedback loops that impart greater benefit to the invader. Alliaria petiolata (Garlic mustard), an
aggressive invader of eastern North American forests, was found to reduce mycorrhizal fungal
densities and impede the growth of native tree seedlings despite not forming mycorrhizal
associations itself (Hagan & Shibu, 2013). Interestingly, a study on forest soils invaded by A.
petiolata found that nutrient levels were consistently and significantly higher than non-invaded
soils. The difference appears not to be due to allelopathic compounds, but instead to increased
rates of decomposition of native trees, fundamentally altering nutrient cycles (Rodgers et al.,
2008).
Overall, this presents a complicated duality where invasive species can alter soil
microbial communities both directly and indirectly, becoming drivers of coevolutionary
trajectories that permanently alter plant community composition (Callaway et al., 2004). Given
the complex interactions between abiotic and biotic factors, along with anthropogenic
interference, it is very likely that no single hypothesis could explain the invasion process. An
integrated approach to test multiple hypotheses and factors to identify the primary causes of
invasion should be considered on a species by species basis when considering management
decisions (Catford et al., 2009; Batish et al., 2013).

15

Geranium robertianum, L. (Herb Robert)
Distribution
Occurring widely in its native Eurasia, Geranium robertianum L. (common name Herb
Robert), can be found in northern Africa northward to Great Britain and Scandinavia and
eastward to Russia and central Asia (Tofts, 2004). Introduced populations are found in Japan,
Chile, New Zealand, Australia, South Africa, and North America. Whether or not it is native to
North America is unclear. Washington and Oregon states list it as a Class B noxious weed,
whereas East Coast and Midwest populations are considered naturalized, with Indiana,
Maryland, and Rhode Island listing G. robertianum as Threatened, Endangered, and of Special
Concern, respectively (USDA NRCS; WA NWCB).

Figure 2. Distribution of Geranium robertianum (Herb Robert) in Washington State. 2018 data
from Washington State Noxious Weed Control Board, map by author.

16

First recorded in Washington state in 1911 in Klickitat county, it was listed as an invasive
species in 1998 and is now present in at least 23 counties (Figure 2; WA NWCB). Of the 17
species from the genus Geranium found in Washington state, 12 are introduced, including G.
robertianum, leaving only 5 species considered native to the region (Hitchcock & Cronquist,
2018; Giblin & Legler, 2021).
Description
Plants range from 10 – 50 cm in size, stems branching from the base, further pairs of
branches forking from the previous base. Forking of branches continues, creating a dense
branching system that eventually ends at a dormant bud and an inflorescence develops (Tofts,
2004). Leaves are bright green with red margins, rarely entirely red, 3.5 – 10 cm wide, ternately
to palmately divided, with somewhat hairy pinnately-lobed leaflets with rounded margins (Tofts
2004; Hitchcock & Cronquist, 2018; Giblin & Legler, 2021). Leaves are fragile and produce a
pungent odor when crushed, described by Tofts (2004) as a “strong, disagreeable smell.” Flowers
are pale to bright pink, sometimes with white stripes, and generally paired in axillary cymules
that arise from the axils of the uppermost leaves. Flowers are radially symmetric, perfect
(containing male and female parts) and complete with five sepals, five petals, ten stamens in two
whorls, and five carpels; seeds are brown, smooth, and 1-2 mm long with a sticky string attached
(Tofts, 2004; Hitchcock & Cronquist, 2018; Giblin & Legler, 2021). The majority of its biomass
is distributed within the stems and leaves, with lesser amounts in its shallow, fibrous root system
(Boerner, 1990).

17

Figure 3. Photograph of Geranium robertianum. Note the deeply dissected leaves and hairy
stems and flowers. Photo credit: King County Noxious Weed Control Program.

Habitat
Geranium robertianum possesses a wide ecological amplitude, occurring in biomes that
range from Mediterranean to boreal and continental conditions (Tofts, 2004). While Herb Robert
is tolerant of varying amounts of light, found in environments that range from full light to 1/370 th
ambient light, it most often occurs in moist, shaded woodland habitats (Tofts, 2004; Bertin,
2001). Light conditions were found to impact biomass production and flowering, where
moderate ambient light conditions produced more biomass and had a higher percentage of
blooms than plants grown in full light (Tofts, 2004). G. robertianum is equally tolerant of a
range of soil conditions, even growing as an epiphyte on trees and cracks in rocks and walls.
North American habitat has been variably described as anything from moist, rocky, or deciduous
18

woods to ravines, gravelly shores, and clearings along roads and trails (Gleason & Cronquist,
1991; Hitchcock & Cronquist, 2018). In the British Isles, most G. robertianum populations are
encountered on moderately nitrogen-rich soils (Tofts, 2001). Poorer soils tend to produce plants
with red leaves, which indicates phosphate or nitrogen deficiency. Soils in Pacific Northwest
forests tend to be nitrogen-poor, suggesting that G. robertianum is able to outcompete native
species that are adapted to low-nitrogen conditions.
Research on populations in Massachusetts by Bertin (2001) found a correlation between
climatic conditions and second-year survivorship of biennial individuals. In Maryland and Rhode
Island, where populations are less secure, hot, dry summers and exceptionally cold winters
without snow cover reduced overall winter survival. In comparison, West Coast climatic
conditions, where the summers are cooler and the winters milder, may lead to greater secondyear survivorship. This difference is supported by similar observations in Britain when compared
to seedling survival in Poland, which was significantly lower in the colder climate (Tofts, 2001).
Phenology and Reproduction
Within its native habitat of Eurasia, and particularly in Great Britain, Geranium
robertianum exhibits striking differences between populations and subspecies (Tofts, 2004;
Baker, 1956). The subspecies robertianum, present in Washington state, generally presents as a
spring and fall annual, occasionally as a biennial. Plants that germinate in fall overwinter as
dense rosettes of leaves, elongating (bolting) in early spring and fruiting in early to mid-summer
(WA NWCB, 1997). Morphological differences between populations continue with second-year
growth, including growth patterns – erect or prostrate – and degree of hairiness (Tofts, 2004).
Similar to the variations in morphology, G. robertianum also exhibits variations in recruitment
patterns. Researchers in Poland identified three seasonal peaks of seedling emergence (late

19

spring, mid-summer, and late summer to early fall), which was similar to the recruitment patterns
in Massachusetts populations (Falińska & Pirožnikow, 1983; Bertin, 2001). They noted in both
Polish and Northeastern U.S. populations that late season seedlings had the poorest rates of
overwinter survival with only 5-6% and 0-23% seedlings surviving into spring, respectively.
Observations of overwinter survival in Great Britain contrasted with the findings in Poland and
Massachusetts: dense aggregates of plants with retained cotyledons persisted through winter
(Tofts, 2001). Again, this suggests that milder climates result in better overwinter survival and
more robust population sizes.
Populations in Great Britain flower from April to September and set fruit from June to
November; populations in the Northeastern U.S. have a similar phenology (Bertin, 2001; Tofts,
2004). Flowering can occur throughout the year in sufficiently mild climates. Pollinator data for
populations in Great Britain include visits from flies, butterflies, and bees (Tofts, 2004). In
Washington state populations pollination is apparently unspecialized (WA NWCB, 1997).
However, most authors consider the plant to be primarily self-pollinated due to its relatively
short flowering period of approximately 48 hours and ability to set seed in greenhouse conditions
lacking pollinators (Bertin, 2001; Tofts, 2004). After approximately three weeks of ripening, five
seeds develop within the capsule, which are then ballistically ejected up to seven yards away
from the plant in response to the drying of the capsule (WA NWCB, 1997; Bertin, 2001; Tofts,
2004). Minute threads attached to the seeds further aid dispersal via wind, insects, and
herbivores.
Seed set quantity in natural conditions varies by location, with plants in Poland producing
100 – 200 seeds per plant and plants in the British Isles producing 10 – 310 seeds (Tofts, 2004).
Research by the Washington State Noxious Weed Control Board found seed production to be
20

relatively low under full canopy with low light conditions, however under 50-60% canopy cover
with higher light levels, plants produced approximately 3100 seeds/m 2 (WA NWCB, 1997).
Some disagreement exists on germination requirements. Scarification requirements have been
reported by some authors, whereas others have reported the seeds germinating freely soon after
dispersal (Tofts, 2004, WA NWCB, 1997). Germination rates varied by both age of seed and by
seasonal cohort. In general, peak germination rates occurred with 1-year old seeds from the
spring cohort, with summer and fall cohort seeds having much lower germination rates (Tofts,
2004; Falińska & Pirožnikow, 1983). The study by Bertin (2001) generally agreed with the 1983
study, reporting that most seeds germinated in the first and second years and none by the seventh
year (Bertin, 2001). Based on these two studies, it would appear as though the seed bank for G.
robertianum is not particularly long lived in the soil.
Physiology
Geranium robertianum is reported to generally be infected with arbuscular mycorrhiza
fungi. Depending on the species of mycorrhiza and soil nutrient levels, mycorrhizal associations
in G. robertianum may result in increased biomass or nutrient-uptake compared to uninfected
plants (Boerner, 1990; Tofts, 2004). Inoculation with the arbuscular mycorrhiza Paraglomus
occultum (syn. Glomus occultum) was found to increase phosphorous uptake efficiency even in
low phosphorous conditions (Boerner, 1990).
G. robertianum is reported as containing no alkaloids. However, it possesses three types
of secretory trichomes: Types I and II secrete terpenoids and phenols, and Type III accumulate
anthocyanins in the apical cells – cells capable of dividing and forming new cells – and secrete
flavonoids (Tofts, 2004). Studies on flavonoids have identified several types excreted by the
plant, but contain no discussion as to their function (Ivancheva & Petrova, 2000). Barndt (2008)

21

examined the allelopathic potential of G. robertianum through leaf leachate and decomposing
leaves. Difficulty with germination of test species proved problematic for the study and results
were not clear, leaving the allelopathic potential of G. robertianum still in question.

22

Materials & Methods

Species Selection
Intact forest understory, specifically the Alnus rubra/Polystichum munitum (Red
alder/Sword fern) plant association, influenced species selection, as it is both a common plant
association throughout the Puget Trough ecoregion and the invasion by Geranium robertianum is
of increasing concern there (Figure 4; Chappell, 2004).

Figure 4. Ecoregions of Washington State. Plant association chosen for this study based on the Alnus
rubra/Polystichum munitum plant association commonly found in the Puget Trough ecoregion
(purple). Map by Washington State Dept. of Natural Resources.

Plugs were chosen over seed due to time constraints involved in stratifying seed and the wide use
of plant plugs over seed in restoration projects. Three different native species of herbaceous
angiosperms, all from different plant families, were chosen to represent different growth forms

23

and phenologies: Geum macrophyllum Willd., Rosaceae; Tellima grandiflora (Pursh) Douglas ex
Lindl., Saxifragaceae; and Dicentra formosa (Haw.) Walp, Papaveraceae (Figure 5). All three
species were sourced from the Puget Sound region: 40 G. macrophyllum (10 cc plugs) from
South Sound Native Plants, Thurston County; 40 T. grandiflora (8 cc plugs), seed sourced from
the SE Olympic lowlands and grown at Washington Corrections Center for Women, Pierce
County; and 35 D. formosa (4 inch square pots) of unknown plant stock from Plantas Nativas
nursery, Whatcom County. G. robertianum plants were collected between January and February
2021 from three different locations in Thurston County. Plants varied in size, but all were
established rosettes. G. robertianum plants were placed in trays of artificial soil mix for up to a
week prior to replanting with the native species.
Geum macrophyllum Willd. is a widely distributed herbaceous perennial in the Rosaceae
family found in western North America from Alaska to Baja California and east across Canada
to the northeast Atlantic coast (Figure 5; Hitchcock & Cronquist, 2018; Giblin & Legler, 2021).
In Washington state it is found on both sides of the Cascade crest. Classified as a facultative
wetland species, it is common in moist woodlands, frequently found along wet meadows and
streambanks from sea level to subalpine zones (Hitchcock & Cronquist, 2018; Fertig, 2021;
Giblin & Legler, 2021). Plants grow from short rhizomes with a fibrous root system and are
tolerant of a variety of soil types, including moderately calcareous soils. Flowering occurs
between April and August (Hitchcock & Cronquist, 2018; Giblin & Legler, 2021). Cymose
inflorescences of 3 – 16 solid yellow flowers have a strong UV signature and attract a range of
pollinators, particularly small flies (USDA NRCS, 2021). Fruits are aggregated achenes with
hooks which may aid in seed dispersal by animals. G. macrophyllum is not known to be toxic or
allelopathic (USDA NRCS, 2021).

24

Tellima grandiflora (Pursh) Douglas ex Lindl. is a monotypic herbaceous perennial in the
Saxifragaceae family primarily found west of the Cascade crest in Washington state, its range
extending along the west coast of North America from Alaska to California and east to western
Montana (Figure 5; Hitchcock & Cronquist, 2018; Giblin & Legler, 2021). A common forb in
Alnus rubra/Polystichum munitum forest associations in the Puget Trough, it grows in rich,
organic soil from sea level to mid-elevations in moist woodlands, forest edges, and along
streambanks (Chappell, 2004). Semi-evergreen during mild winters, plants grow from short
rhizomes with fibrous roots, blooming from April to July. It produces long, spike-like racemose
inflorescences of 10 – 35 greenish-white to sometimes pink flowers. The plant self-seeds readily,
producing 100 – 150 seeds per plant (Hitchcock & Cronquist, 2018; Giblin & Legler, 2021). T.
grandiflora is noted as species useful for restoration by the Washington Native Plant Society, as
it can create thick patches and outcompete invasive species in disturbed and shady areas (WNPS,
2021).

Figure 5. Photographs of native species used in experimental design, left to right: Geum macrophyllum,
Tellima grandiflora, and Dicentra formosa. Photos by author.

25

Dicentra formosa (Haw.) Walp. is an herbaceous perennial in the Papaveraceae family
found primarily in low to mid-elevation moist woodlands west of the Cascades, from British
Columbia south to California (Figure 5; Hitchcock & Cronquist, 2018; Giblin & Legler, 2021). It
grows from relatively shallow, brittle rhizomes. Tolerant of varied soil textures and full shade, it
does poorly in calcareous soils and drought conditions (USDA NRCS, 2021). Dormant in winter,
it emerges in mid-spring and flowers between March and July, producing panicles of 2 – 30
distinct pink-colored, heart-shaped flowers. It can go dormant during the dry summer season,
emerging and flowering again in late summer to early fall. Seeds are borne in elongated capsules
with elaiosomes, fleshy structures containing fats. Seed dispersal is aided by ants, which are
attracted to the elaiosomes, and seeds require cold stratification for germination. The plant is
considered moderately toxic to herbivores but is not known to be allelopathic (USDA NRCS,
2021).
Experimental Design
Growing media for all treatments was an artificial soil mix composed of a 2:1:1 mixture
of peat moss, perlite, and coarse sand sourced from a local hardware store. Components were
measured by volume and mixed in 40 L batches. Ten plants of each species were randomly
selected for each of four treatment groups (Table 2). The first treatment group served as the
control group and consisted of a single native species grown in the absence of G. robertianum in
the artificial soil mix with no other soil amendments (n = 30). The second treatment group
consisted of a single native species grown in the presence of G. robertianum with no soil
amendments (n = 30). The third treatment group consisted of a single native species grown in the
presence of G. robertianum with the addition of 20 g granulated activated horticultural charcoal
(20 g/pot, n = 30; Rio Hamza Trading Co., Corbin, KY). The final treatment group consisted of a

26

single native species grown in the presence of G. robertianum with the addition of native mineral
soil (200 g/pot, n = 30). The native mineral soil was collected from two locations adjacent to
Schneider Creek, Thurston County, in undisturbed Alnus rubra/Polystichum munitum plant
association forest, then mixed for uniformity before being added to each treatment pot.
Altogether there were 4 different treatments for 3 different species group with each combination
replicated 10 times, for a total of 4 x 3 x 10 = 120 pots.
G. macrophyllum and T. grandiflora plants were uniform in size and development within
their species group. D. formosa varied the most in terms of rhizome size and development per
pot. Plants varied from singular rhizomes with minimal root development to extensive rhizome
and root development that filled out the entire nursery pot. Plants with exceptionally large
rhizomes were divided to accommodate the need for an additional five plants to bring the count
up to 40.
Table 2. Summary of treatment groups by species with shorthand.
Shorthand

Treatment 1
Control

Treatment 2
Invaded

Treatment 3
Carbon

Treatment 4
Soil

Geum macrophyllum

GEMA

GM

GMGR*

GMGRC

GMGRS

Tellima grandiflora

TEGR

TG

TGGR

TGGRC

TGGRS

Dicentra formosa

DIFO

DF

DFGR

DFGRC

DFGRS

Geranium robertianum
w/GEMA

GERO

-

GRGM

GRGMC

GRGMS

Geranium robertianum
w/TEGR

GERO

-

GRTG

GRTGC

GRTGS

Geranium robertianum
w/DIFO

GERO

-

GRDF

GRDFC

GRDFS

Species

*Geum Macrophyllum grown with Geranium Robertianum; C = activated charcoal treatment (Carbon
group); S = native soil treatment (Soil group)

27

For each treatment group, plant plugs were transferred to 1-gallon plastic nursery pots
filled with the artificial soil mix and any additional soil treatments (native mineral soil or
charcoal). Plugs were either planted in the center of the pot or equidistant to G. robertianum,
depending on the treatment group. Pots were placed by treatment group on nursery racks in a
utility room adjacent to a large, SE facing window with full natural light. Due to low-light levels
in winter, light was supplemented with 2-bulb, 4-foot light fixtures fitted with 6500K fluorescent
light bulbs (Philips F40T12/Daylight) positioned at equal heights above the pots. Lights were
controlled by a timer and left on for 12 hours per day from 8 AM to 8 PM. Pots were rotated
weekly to minimize any possible effects of shelf level or proximity to the window. The
temperature of the room was kept between 17 – 20° C (62 – 68° F) for the duration of the
experiment. Water was added as needed to maintain consistent moisture levels across all species
and treatment groups. Water was sourced from a domestic community well with chlorine levels
kept at or below 1 ppm and left in open containers at room temperature for 24 hours before
using.
At approximately three weeks into the experiment, aphids infested all of the G.
robertianum, but did not infest the native species. In order to limit the impact of the infestation,
all G. robertianum was treated with insecticidal soap (Natria© Insecticidal Soap, 1% potassium
salts of fatty acids), regardless of level of infestation, and any overspray onto the native species
was immediately rinsed with water. This was repeated 10 days after the first treatment. Aphids
did return, but at significantly lower numbers. At eight weeks (56 days), all pots were fertilized
with 200 mL of diluted Alaska fish fertilizer (5-1-1 NPK, 15 mL per 1-gallon of water). At 12
weeks (84 days), all plants were harvested and aboveground and belowground parts separated.
Due to the nature of commercially available plant plugs often consisting of more than one

28

seedling, many of the plugs for T. grandiflora and G. macrophyllum contained multiple plants.
The number of plants present and their relative size in each pot for these species was recorded
during harvesting. For D. formosa, all aboveground biomass was collected, but only a subset of 5
replicates per treatment group were selected for belowground biomass collection. Each of these 5
replicates were chosen using a random number generator. All above- and belowground samples
were then oven-dried in separate paper bags at 80° C (176° F) for 24 hours and weighed with an
analytical scale (RADWAG AS 82/820.R2). Aboveground and belowground biomass data were
also collected for G. robertianum.

Figure 6. Photographs of the full setup of all species and treatment groups.

29

Data analysis
A one-way analysis of variance (ANOVA) was used to analyze the effects of each
treatment on the biomass of the native species and G. robertianum using R Statistical Software
(RStudio version 2021.9.0+351). Aboveground, belowground, and total biomass of the native
species and G. robertianum were all analyzed individually (see Tables 1 & 2 in Appendix A).
For G. macrophyllum and T. grandiflora, each response variable was also analyzed on a perplant basis where there was more than one plant per plot. One-way ANOVAs were used to
evaluation treatment effects on the following response variables: aboveground biomass,
belowground biomass, ratio of aboveground to belowground biomass for each species, and ratio
of native total biomass to G. robertianum total biomass. Due to the high variability in existing
rhizome development across the 40 plants D. formosa plants, a ratio for aboveground to
belowground growth was not calculated for this species.
Individual treatment groups within a dataset not meeting the assumptions of normality
required the entire dataset to be transformed with Log10 transformation. Datasets that could not
be transformed to meet the assumptions of normality and homogeneity of variance were
analyzed instead with the Kruskal-Wallis test for non-parametric data. Differences among
treatment groups were examined with the Dunn’s (Benjamini-Hochberg) post hoc test. For data
meeting ANOVA assumptions, differences among treatment groups were examined using
Tukey’s HSD post hoc test.

30

Results
Geum macrophyllum
There was a statistically significant difference between aboveground and belowground
biomass by treatment group (F7,72 = 61.04, p < 0.001; Figure 7; Figure 8; Table 4). Additionally,
there was a statistically significant treatment effect on all response variables except for the
belowground per plant analysis (Table 3). The Invaded and Carbon treatment groups exhibited
the greatest decrease in both aboveground and total biomasses compared to the Control group (p
< 0.001 for both) and Soil treatment group (p < 0.001 for both) (Table 4). There was no
significant treatment effect for the Soil treatment group compared to the Control group (p =
0.164). When comparing the ratio of aboveground to belowground biomass across treatment
groups, three out of six pairwise comparisons were statistically significant (Table 4). The mean
aboveground-belowground ratio (AG:BG) for the entire species group was 2.03 (Table 1,
Appendix B). None of the plants died, nor produced flowers.

Figure 7. Photograph of
Geum macrophyllum
Control group (left) and
Geum macrophyllum
Carbon Group (right)
after ~10 weeks of
growth. The differences
between these two
treatment groups was
statistically significant
(p < 0.001).

31

Tellima grandiflora
There were no significant differences between treatment groups for this species. Mean
aboveground biomass did not significantly differ from the corresponding belowground biomass
for each treatment group (Figure 8). The mean aboveground-belowground ratio (AG:BG) for the
entire species group was 0.99 (Table 1, Appendix B). Out of the 40 T. grandiflora plants, none
died and only three produced flowers, all of which were in the Soil treatment group.

Figure 8. Effect of treatments on native biomass for Geum macrophyllum (left) and Tellima grandiflora
(right); green and brown portions represent aboveground and belowground biomass, respectively (note:
numbers on y-axis are not negative). Values are means and standard errors for each treatment group.

Dicentra formosa
The aboveground biomass was the only response variable which showed a significant
treatment effect for D. formosa (F3, 33 = 10.84, p < 0.001; Table 3). This effect was limited to the
Soil treatment group, with significantly less aboveground biomass than all other treatment
groups (Figure 9; Table 4). All four treatment groups produced flowers, with the Soil group
32

representing the lowest number of blooms (DF: 9 flowers; DFGR: 8; DFGRC: 8; and DFGRS:
3). Three plants (DFGR: 1 and DFGRC: 2) either did not produce aboveground growth or they
died back, though the rhizomes were intact and alive. New rhizome growth was conspicuous and
noted during harvesting. Not all plants produced new rhizome growth during the course of the
experiment. Of the five subsamples taken from each treatment group for belowground biomass,
the Control group had two plants with new rhizome growth, all five in the Invaded group had
new growth, none of the Carbon group had new growth, and two in the Soil group had new
growth.

Figure 9. Effect of treatments on biomass of Dicentra formosa; green and brown portions represent
aboveground and belowground biomass, respectively (note: numbers on y-axis are not negative). Values
are means and standard errors for each treatment group.

33

Table 3. Analysis of variance (ANOVA) results for biomass of native species Geum
macrophyllum, Tellima grandiflora, and Dicentra formosa grown with and without
Geranium robertianum for all treatment groups. Significant p-values (p < 0.05) are bolded.
Species & Response Variable
G. macrophyllum
Aboveground Biomass*
Aboveground Biomass per plant*
Belowground Biomass
Belowground Biomass per plant**
Total Biomass
Total Biomass per plant**
AG:BG Ratio**
AG:BG by Group
T. grandiflora
Aboveground Biomass*
Aboveground Biomass per plant*
Belowground Biomass
Belowground Biomass per plant
Total Biomass
Total Biomass per plant
AG:BG Ratio
AG:BG by Group
D. formosa
Aboveground Biomass*
Belowground Biomass**
Total Biomass**

df

SS

F

p

3
3
3
3
3
3
3
7

1.908
0.486
0.251
6.645
7.614

67.34
9.690
6.471
48.7
61.04

< 0.001
< 0.001
0.001
0.052
< 0.001
< 0.001
< 0.001
< 0.001

3
3
3
3
3
3
3
7

0.176
0.054
0.299
0.338
0.902
0.649
0.703
0.621

0.059
0.224
1.583
1.267
1.344
0.684
2.867
1.336

0.196
0.879
0.210
0.300
0.275
0.568
0.050
0.246

3
3
3

3.595
-

10.84
-

< 0.001
0.268
0.300

*Transformed data; **Nonparametric data analyzed with Kruskal-Wallis

34

Table 4. P-value results from pairwise comparisons of treatment groups; post hoc for ANOVA &
Tukey’s HSD and Kruskal-Wallis & Dunn’s Test. Significant p-values (p < 0.05) are bolded.
Green highlight indicates nonparametric analysis; blue indicates parametric analysis after
transformation.
Corresponding p
Values for:

AG
Biomass

AG
Biomass
per plant

BG
Biomass

BG
Biomass
per plant

Geum macrophyllum (GM)
GMGR – GM
0.175
< 0.001
0.017
0.006
GMGRC – GM
0.062
< 0.001
< 0.001
0.002
GMGRS – GM
0.164
0.867
0.121
0.493
GMGRC – GMGR
0.351
0.969
0.503
0.017
GMGRS – GMGR
0.104
0.582
0.426
< 0.001
GMGRS – GMGRC
0.324
0.164
< 0.001
0.002
Geranium robertianum grown with G. macrophyllum (GRGM)
GRGMC – GRGM
0.067
0.660
GRGMS – GRGM
0.945
0.478
GRGMS – GRGMC
0.113
0.041
Tellima grandiflora (TG)
TGGR – TG
0.207
0.890
0.392
0.916
TGGRC – TG
0.997
0.931
0.825
0.801
TGGRS – TG
0.957
0.908
0.197
0.240
TGGR – TGGRC
0.287
0.999
0.881
0.994
TGGR – TGGRS
0.452
0.999
0.973
0.588
TGGRS – TGGRC
0.989
0.999
0.651
0.746
Geranium robertianum grown with T. grandiflora (GRTG)
GRTGC – GRTG
0.740
0.992
GRTGS – GRTG
0.995
0.576
GRTGS – GRTGC
0.688
0.651
Dicentra formosa (DF)
DFGR – DF
0.660
0.589
DFGRC – DF
0.969
0.368
DFGRS – DF
0.447
< 0.001
DFGRC – DFGR
0.431
0.620
DFGRS – DFGR
0.668
< 0.001
DFGRS – DFGRC
0.669
0.006
Geranium robertianum grown with D. formosa (GRDF)
GRDFC – GRDF
0.200
0.928
GRDFS – GRDF
0.433
0.987
GRDFS – GRDFC
0.846
0.980
-

Total
Biomass

Total
Biomass
per plant

AG:BG
Ratio

< 0.001
< 0.001
0.008
0.402
< 0.001
< 0.001

0.034
0.001
0.516
0.232
0.123
0.007

0.099
0.003
0.619
0.188
0.039
0.001

0.083
0.900
0.040

-

0.046
0.707
0.270

0.283
0.983
0.627
0.477
0.929
0.834

0.805
0.910
0.502
0.995
0.957
0.878

0.999
0.537
0.090
0.488
0.076
0.707

0.788
0.993
0.848

-

0.300
0.536
0.039

0.727
0.466
0.403
0.589
0.668
0.789

-

-

0.802
0.957
0.945

-

0.585
0.801
0.940

Notes: Native plants grown without G. robertianum are the Control Group (GM, TG, DF); plants grown
with G. robertianum but without additional treatments are the Invaded group (GMGR, TGGR, DFGR);
Addition of C or S indicates Carbon treatment and Soil treatment groups, respectively (e.g., TGGRC and
TGGRS).

35

Geranium robertianum
All three groups of G. robertianum had greater mean aboveground biomass than
corresponding belowground biomass (Figure 10; Table 1, Appendix B). Many of the plants
produced flowers and seeds (GRGM: 18/30; GRTG: 29/30; and GRDF: 21/30). After the
completion of the experiment, the 20 leftover D. formosa pots with intact soil were placed
outside. Within several weeks, several of the pots had newly sprouted G. robertianum seedlings.
For G. robertianum grown with G. macrophyllum, the Soil and Invaded treatment groups
were nearly equal in terms of mean aboveground biomass. There was a statistically significant
treatment effect for the Carbon group when compared to the Soil group (p = 0.041; Table 4),
with both the aboveground and belowground biomass significantly lower for the Soil group.
Similarly, the aboveground and belowground biomass for the Invaded treatment group was
smaller relative to the Carbon treatment group, however the difference was not statistically
significant (p = 0.067; Table 4). The ratio of aboveground to belowground biomass for the
Carbon treatment group was significantly greater than the Invaded treatment group (p = 0.046),
indicating aboveground growth was greater than belowground growth for the Carbon treatment
group. Four plants died during the experiment, three in the Invaded treatment group (GMGR)
and one in the Soil treatment group (GMGRS).
There were no significant treatment effects for the G. robertianum grown with T.
grandiflora, with the exception of the aboveground-belowground ratio for the Soil treatment
group compared to the Carbon treatment group, with larger ratios for the Soil treatment group (p
= 0.039). None of the plants died in this species group. Similarly, there were no significant
treatment effects for the G. robertianum grown with D. formosa (Table 5). A total of six plants
died in this group: four in the Invaded group, three after flowering, and two in the Soil group.

36

Table 5. Analysis of variance (ANOVA) results for biomass of Geranium robertianum by native
species group for all treatment groups (Invaded, Carbon, and Soil). Significant p-values (p <
0.05) are bolded.
Species & Response Variable
G. robertianum w/G. macrophyllum
Aboveground Biomass*
Belowground Biomass*
Total Biomass*
AG:BG Ratio*
G. robertianum w/T. grandiflora
Aboveground Biomass
Belowground Biomass
Total Biomass
AG:BG Ratio
G. robertianum w/D. formosa
Aboveground Biomass
Belowground Biomass
Total Biomass
AG:BG Ratio

df

SS

F

p

2
2
2
2

0.679
0.178
0.577
0.242

4.528
2.226
4.396
3.519

0.024
0.134
0.026
0.049

2
2
2
2

0.259
0.017
0.211
22.33

0.416
0.611
0.250
3.388

0.664
0.550
0.781
0.049

2
2
2
2

1.007
0.002
0.256
2.346

1.716
0.068
0.207
0.561

0.204
0.935
0.817
0.592

*Transformed data

37

G. robertianum
Figure 10. Effect of treatments on Geranium robertianum biomass; green and brown portions represent aboveground and belowground biomass,
respectively. Values are means and standard errors for each treatment group. Groups are as follows: DICFOR = G. robertianum grown with
Dicentra formosa; GEUMAC = G. robertianum grown with Geum macrophyllum; TELGRA = G. robertianum grown with Tellima grandiflora.

38

Discussion

Examining the direct effects of the invasive Geranium robertianum on native species was
the intent of this research. Specifically, what are the effects on the growth of the native species
and is there any evidence of allelopathic interference from G. robertianum. While this plant is
somewhat wide spread in western Washington, little is known about its direct effects on native
plants or whether its presence has long-term consequences for native plant community
composition. This is somewhat concerning as it is capable of invading intact, uninvaded forest
understory. Overall, this study revealed interesting patterns when comparing results across
species groups. In terms of competitive ability and effects of treatments, Geum macrophyllum
and Tellima grandiflora stood out as highly contrasting species. The results for Dicentra formosa
are less informative, which is likely due to the high variability in preexisting rhizome size from
one replicate to the other. Additionally, due to the rhizomatous growth habit of D. formosa, it is
possible the 12-week duration of the experiment was insufficient to effectively capture the
competitive interactions between these species. The presence of multiple plants per pot for G.
macrophyllum and T. grandiflora had the potential to confound the results, but this was not the
case with T. grandiflora and was of little impact for G. macrophyllum, which will be discussed
below.
Geum macrophyllum
This species group exhibited the most dramatic changes in biomass relative to treatments,
with all but one response variable indicating statistically significant differences among treatment
groups (Table 3). The fact that both aboveground and belowground biomass were reduced in the
Invaded and Carbon treatments relative to the Control group suggest that G. robertianum was

39

directly reducing the growth of this species. In order to demonstrate an allelopathic effect, the
response of the native species (above- and/or belowground biomass) should decrease with
exposure to a potential allelopathic invader but should recover with the addition of an activated
charcoal treatment (Rúa et al., 2008; Del Fabbro et al., 2014). A lack of this specific response by
G. macrophyllum suggests that the reduction in total biomass for both the Invaded and Carbon
groups is likely due to general interspecific competition for resources versus allelopathic
interference from G. robertianum. In contrast to the Invaded and Carbon treatment groups, the
effect of G. robertianum on G. macrophyllum in the Soil treatment group appears to have been
ameliorated by the addition of native soil, as the total biomass G. macrophyllum in the Soil
treatment group was nearly 80% of that of the Control group. This result contrasts with the
majority of studies on the effects of invasives on soil microbial communities, which are
generally characterized as negative. For example, invasive species are able to exploit the benefits
of symbioses while avoiding the mutual cost of the network, or alter microbial communities by
increasing the abundance of generalist pathogens (Reinhart & Callaway, 2006; Hagan & Jose,
2013; Lankau, 2013). However, arbuscular mycorrhizas are important mediators of competitive
interactions between invasive and native plants and, despite generally lacking host-specificity,
specificity in growth responses for individual infected plants does exist (Reinhart & Callaway,
2006). Given the complex interactions of plant-mycorrhizal associations, the results for the Soil
treatment group suggests that the soil microbial community is important for G. macrophyllum in
terms of both growth and competitive ability.
Overall, this species had greater mean aboveground biomass than belowground biomass.
The mean AG:BG ratio across all treatment groups was 2.03, revealing an emphasis on
aboveground growth for this species (Table 1, Appendix B). However, there was a notable shift

40

for the Invaded and Carbon treatment groups – the mean AG:BG ratios were lower than the
Control and Soil treatment groups but the effects were not proportional. For example, the
aboveground biomass of the Invaded group was ~60% lower than the Control group, whereas
belowground biomass was only ~40% lower. It appears as though G. macrophyllum altered
resource allocation and conservation in response to invasion. The changes in AG:BG ratios and
statistical significance between treatment effects is only moderately reduced when the number of
plants per pot are taken into consideration. While growing several individuals of the same
species with a different species will produce both interspecific and intraspecific competition
effects, it is theorized that individuals from the same species will have higher levels of
competition due to sharing similar resource needs (Mangla et al., 2011). The results for the
Control and Invaded groups contradict this theory, however. Since all four treatment groups
contained more than one individual of G. macrophyllum, the significant difference between the
Control and Invaded groups points to higher levels of interspecific competition, not intraspecific,
as the driver behind decreased biomass for G. macrophyllum.
Tellima grandiflora
Based on the hypothesis that the native species would exhibit some effect from the
presence of G. robertianum, the strongest effect exhibited by T. grandiflora was by the
belowground biomass for the Soil treatment group relative to the Control group, but this was still
a rather large p-value of 0.197. All treatment groups had belowground biomass that very closely
mirrored its corresponding aboveground biomass, with mean AG:BG ratios ranging from 0.82 to
1.12 (species group mean of 0.99). This suggests that whatever effect a treatment or the presence
of G. robertianum had on the plants, it more or less affected aboveground and belowground
growth equally. It is difficult to determine if or how much allelopathy impacted growth versus

41

other modes of interspecific competition, such as resource availability. The total biomass of the
Invaded group did slightly decrease compared to the Control, which appears to have been
improved by the activated charcoal treatment in the Carbon group. This is the pattern expected
for allelopathic interference, however there was no statistically significant differences among the
treatment groups. While also lacking statistical significance, T. grandiflora responded positively
to the native soil treatment compared to the Invaded group in terms of aboveground growth. This
is somewhat similar to G. macrophyllum where the positive effect was highest for aboveground
growth, implicating soil biota as an important factor in resource acquisition for these native
species. An additional qualitative observation for the Soil treatment group that highlights the
importance of the soil microbial community was that the only plants that flowered (out of all 40
plants in the species group) were in this treatment group, despite all plants being comparable in
size and health at the start of the experiment.
Focusing on two different modes of allelopathy, Brandt (2008) used both G. robertianum
leaf litter and leaf leachate to investigate whether either had a negative impact on the growth of
T. grandiflora. Decomposing leaf litter had no significant impact on aboveground or
belowground biomass, whereas leaf leachate reduced both. However, it was noted that the leaf
leachate was very acidic (pH = 4.0), so it is difficult to determine whether it was allelopathy or
acidity that impacted growth. Brandt also commented that other modes of allelopathy should be
explored for G. robertianum, including the possibility of root exudates being the primary mode
of its allelopathic potential. While the study design for this thesis did not extract or measure root
exudates, plants would have been exposed to any exudates released by G. robertianum.
Although the growth of T. grandiflora appeared to be overall robust, the presence of G.
robertianum may affect this species in other ways not examined in this study, such as flowering
42

and seed set. T. grandiflora was specifically noted by the Washington Native Plant Society as an
ideal species for restoration, owing to its ability to grow well in disturbed and shady areas and
outcompete invasive weeds (WNPS, 2021). Even when accounting for the discrepancies in the
number of seedlings per pot (AG, BG, and Total Biomass per plant) and intraspecific
competition, the effects of G. robertianum and the treatments were still insignificant, suggesting
that T. grandiflora is an effective competitor against invasion.
While there is a paucity of studies examining the competitive ability of native plants
found in Puget lowland forests, many studies on shrub-steppe ecosystems of western North
America have examined various methods to out-compete highly invasive Bromus tectorum
(cheatgrass). A 2014 study explored whether native “weedy” forbs could effectively compete
with B. tectorum and enhance the growth of native perennial grasses, finding three of the seven
test species were highly effective at suppressing the growth and seed production of B. tectorum
(Leger et al., 2014). Applying the invasion hypothesis of biotic resistance – competitive
resistance of a recipient community that impedes invaders – this example, along with the
resilience of T. grandiflora, highlights the need for more studies on native species that are
potentially resistant to invasion and able to enhance restoration efforts.
Dicentra formosa
The most prominent finding for this species was the significant decrease in aboveground
biomass for the Soil treatment group when compared to the other treatment groups (Figure 9).
Given the high variability in rhizome size across all groups during planting, it would be easy to
conflate this variability with the treatments and the presence of G. robertianum. However, the
belowground biomass for the Soil treatment group was not proportionately small and did not
significantly differ from the other groups. An additional outcome of the Soil treatment group was

43

the reduced number of flowering individuals compared to the other groups – only 3 out of 10
plants flowered. There are a few possible explanations for the reduced aboveground biomass for
this group: 1) the lack of inflorescences and associated foliage reduced the overall aboveground
biomass; 2) the particular plants selected for this group contained some of the extra-large plants
that were divided to reach the required 40, with the plants potentially focusing energy on
belowground growth in response to division; or 3) the presence of G. robertianum impacted
flowering and growth. Considering the other two groups with G. robertianum had similar
flowering rates to the Control group, it is less likely that the presence of G. robertianum would
have impacted flowering. However, due to the complexity of belowground interactions with the
soil microbial community, including promoting antagonists, the possibility that G. robertianum
negatively impacted growth and flowering cannot be completely ruled out (Reinhart & Callaway,
2006; Stinson et al., 2006).
Another notable outcome for this species group was the lack of new rhizome growth for
the Carbon treatment group. During harvesting, new rhizome growth was conspicuous – pale
white in color and easily broken – whereas existing rhizomes were darker in color and more
resistant to breakage. Activated carbon is frequently used in allelopathic studies due to its ability
to adsorb biochemical compounds (Rúa et al., 2008; Del Fabbro et al., 2014), but it can also
interfere with soil nutrient availability, water retention, soil pH, and mycorrhizal development,
negatively impacting plant growth (Wurst et al., 2010). Weißhuhn and Prati (2008) found
activated charcoal substantially altered substrate chemistry by increasing available phosphate and
decreasing the ratio of organic carbon to total nitrogen. The results of these studies indicate that
even if activated charcoal is found to reduce allelopathic chemicals, its ability to alter soil
chemistry has the potential to confound results and erroneously attribute greater negative effect

44

to allelopathy than is warranted (Weißhuhn & Prati, 2008; Wurst et al., 2010). Considering the
Carbon treatment group was the only group that failed to develop new rhizomes, it suggests that
the activated charcoal had a negative effect on the soil chemistry.
Geranium robertianum
The native soil treatment had an interesting effect on the G. robertianum grown with
Geum macrophyllum and Tellima grandiflora. Aboveground biomass of the Soil treatment group
for G. robertianum from both species groups was nearly equal to the aboveground biomass of the
Invaded groups, but belowground biomass decreased by ~25% for both. In contrast, both native
species had positive responses for aboveground biomass in the Soil treatment group, increasing
by ~100% and ~36% relative to the Invaded groups, respectively. The plants grown with
Dicentra formosa had somewhat different results – instead of aboveground biomass of the Soil
treatment group being nearly equal to the Invaded group, it decreased by ~30%. Belowground
biomass also decreased but only by ~6%. These results both agree and conflict with studies on
plant-soil biota feedbacks, where the invader is generally theorized to be the one that receives the
most benefits. One study on Centaurea stoebe (spotted knapweed) found positive plant-soil biota
feedbacks (increased biomass) when grown in North American soil with a native grass, but not
when grown in soil from its native region of Europe (Callaway et al., 2004). When grown in
sterilized soil, plant-soil biota feedbacks were eliminated and the biomass of the native grass
increased. These results show that C. stoebe benefits from being removed from co-evolutionary
pathogens in its native range, potentially gaining mutualistic associations and outcompeting the
native grass. The conflicting results for G. robertianum across native species demonstrate how
competitive relationships and soil microbial community interactions are species-specific and

45

difficult to predict even within the same plant community (Weißhuhn & Prati, 2008;
Weidenhamer & Callaway, 2010).
In a mycorrhizal study on G. robertianum, it was noted that the plant is capable of
forming mycorrhizal associations and that these associations were particularly helpful in low
nutrient soils (Boerner, 1990). Plants inoculated with arbuscular mycorrhizal fungi had higher
uptake of phosphorous and nitrogen compared to plants that did not receive inoculation. This
increase in nutrient uptake may be due to either the mycorrhiza facilitating greater root mass and
length, thereby allowing for greater uptake rates, or the mycorrhiza directly assisting in the
increased nutrient uptake, or possibly both (Boerner, 1990, Reinhart & Callaway, 2006). Nonnative species that are capable of making novel mycorrhizal associations can enhance their
success, tipping the scales from being non-native to invasive. This interaction is two-way, with
soil microbial communities affecting and responding to changes in plant community composition
(Lankau, 2013). New mycorrhizal associations can facilitate a feedback system that alters
existing plant-soil biota interactions, such as increasing the abundance of pathogens or changing
the abundance of mutualists, which can be beneficial or inhibitory to native species (Reinhart &
Callaway, 2006; Lankau, 2013). The lack of a significant negative effect of the native soil
treatment on G. robertianum suggests that novel mycorrhizal associations and an absence of coevolutionary inhibitory soil pathogens may be mechanisms contributing to its success as an
invader.
The plants grown with G. macrophyllum had a significant increase in aboveground
biomass for the Carbon treatment group, whereas aboveground biomass for G. macrophyllum
was lowest for this is treatment group. This is contrary to what would be expected for a
competitive effect driven by allelopathic interference, where you would anticipate an increase in

46

biomass for the native, not the invader. Unexpectedly, the results were opposite for the plants
grown with T. grandiflora and D. formosa, both of which had a reduction in aboveground
biomass relative to the Invaded groups. The patterns for plants grown with T. grandiflora and D.
formosa are what you would expect to see if allelopathic chemicals were being inhibited by an
activated charcoal treatment, however, none of the differences were statistically significant.
Overall, these patterns suggest that the Carbon treatment was not particularly detrimental to
growth for G. robertianum, nor did they indicate any strong signs of allelopathic interference.
One further notable pattern for G. robertianum that may offer an explanation of
invasiveness was the amount of its aboveground biomass relative to belowground biomass,
which stands out compared to both T. grandiflora and G. macrophyllum. When considering the
phenology and life cycle of each of these species, it does not seem unusual that the native
species, which are perennials, dedicate more energy and resources towards belowground biomass
than the annual-biennial G. robertianum. The mean AG:BG ratio for G. robertianum across all
species and treatment groups was 4.66, whereas G. macrophyllum and T. grandiflora only have
mean ratios of 2.03 and 0.99, respectively (Table 1, Appendix B). A comparative study of
perennials and annuals in Argentina found that root traits of annuals were associated with
enhanced resource acquisition via low-density roots with high nitrogen concentration, whereas
the perennials demonstrated enhanced root persistence with the presence of thick, dense root
systems (Roumet et al., 2006). This study demonstrates resource acquisition and distribution are
crucial differences between these plant life cycles. The high AG:BG ratio for G. robertianum
suggests that an annual life cycle with energy and resources directed towards abundant
aboveground growth and proportionally abundant seed quantity are significant contributors to the
invasiveness of this species (Table 1, Appendix B; Figure 1, Appendix C). The AG:BG ratios for

47

G. macrophyllum and T. grandiflora are consistent with this notion, as both species directed far
more resources to belowground growth compared to G. robertianum. Likewise, due to its
rhizomatous growth habit, D. formosa has a similar pattern, directing more resources to
belowground growth compared to G. robertianum.
Limitations and Changes
Given the opportunity to reflect on this thesis process, there are several things that I
would have done differently. This list is not exhaustive, but merely points out the more obvious
aspects that could have been improved upon. Given the interesting plant interactions I was
looking at it would have been incredibly useful to have a full factorial study design that included
not just the native grown in monoculture but also with the same treatments as the other groups
(activated carbon and native mineral soil). Growing G. robertianum by itself with the same
treatments would also have been informative. To avoid confounding the results with aspects of
intraspecific competition, I should have thinned the plugs down to just a few seedlings at the
start of the experiment and finally down to a single, dominant seedling once the plugs were well
established in the pots. In a similar vein, the rhizomatous growth habit of Dicentra formosa made
for a challenging set up and interpretation of the results. A standardized rhizome length would
have been better to start with and make for more informative results at the end. Using a native
species with a more fibrous root structure more similar to T. grandiflora and G. macrophyllum
for the third native in the experimental design may have been more informative. Additionally, I
should have been more mindful of phenology, as many of the G. robertianum plants had gone to
seed and were entering senescence and beginning to die back before harvesting began.

48

Conclusion

Overall, in evaluating whether any of the treatments were effective at improving
competitiveness of the native species, the results were quite mixed. The only native species to
respond positively to the activated charcoal treatment was T. grandiflora, though the difference
was not statistically significant. The carbon treatment for G. macrophyllum had the exact
opposite outcome expected when looking for allelopathic interference from an invasive species:
the invasive biomass increased significantly while the native biomass decreased. Either G.
robertianum benefited from the presence of the activated charcoal or benefited from the
inhibition of G. macrophyllum by the activated charcoal. Given the decreased biomass of G.
robertianum in the Carbon treatment groups grown with D. formosa and T. grandiflora, it is
more likely that G. robertianum benefited not from the presence of activated charcoal, but from
the inhibitory effect on G. macrophyllum. Unlike the activated charcoal treatment, the native soil
treatment appeared to be generally neutral to beneficial for all species, with the exception of D.
formosa. The robust response of G. macrophyllum to the soil treatment suggests that it may be a
more resilient to invasion than was demonstrated by its performance in the Invaded and Carbon
treatment groups.
While there was a clear pattern of interspecific competition effects on the native species
by G. robertianum, there is no strong evidence of allelopathic interference. The patterns you
would expect to see with allelopathic interference are present for several of the groups, however
they all lacked statistical significance. Considering the patterns are present, it is possible that a
study with more than 10 replicates per treatment group could reveal greater significance. Also, it
is possible that allelopathy is only partially responsible for the invasive success of G.

49

robertianum and that its effects are weak, but not entirely insignificant when considered as a part
of a complement of invasive mechanisms. In criticism of the Novel Weapons hypothesis, it has
been pointed out that allelopathy does not neatly explain the invasiveness of many invasive
species (Del Fabbro et al., 2014). The presence of patterns demonstrating allelopathic
interference for a very successful invasive species but lacking statistical significance
substantiates this claim. Moreover, allelopathic chemicals are not necessary for growth or
reproduction and are therefore costly to produce, possibly only produced when sufficient
resources are available, suggesting that use of allelopathic chemicals by G. robertianum is
possibly dependent on life cycle and habitat quality (Parepa & Bossdorf, 2016).
In terms of invasion hypotheses relevant to G. robertianum, there does not appear to by a
“holy grail” hypothesis or mechanism that adequately explains its invasive success. The
abundant aboveground growth and proportionally abundant seed quantity support both the
Propagule pressure and Ideal Weed hypotheses as explanations for its success as an invader.
Propagule pressure is considered by some as a prerequisite for invasiveness and has been
demonstrated to be the primary determinant of habitat invisibility (Van Holle & Simberloff,
2005). High propagule pressure may enable invaders to become established through saturation of
the seed bank, which quickly overwhelms any biotic resistance to invasion (Van Holle &
Simberloff, 2005; Catford et al., 2009). The broad ecological tolerances exhibited by G.
robertianum suggest it possesses high phenotypic plasticity, which also falls under the Ideal
Weed hypothesis. Evidence for allelopathy and the Novel Weapons hypothesis is weak and does
not appear to be the most relevant hypothesis to explain its invasiveness, but is potentially a
minor contributing factor. It is worth pointing out, however, that allelopathic interactions can be
species-specific and that long-term changes to soil chemistry facilitated by invasives is not well

50

understood. Given that it is widely recognized that plants can influence the structure and
composition of the soil microbial community, there is the potential that G. robertianum is
influencing native soil biota, triggering feedback loops that are either beneficial to itself or
detrimental to natives (Del Fabbro & Prati, 2015). As this invasive species shows no signs of
slowing its spread in western Washington, it is clear that more research is needed, particularly in
terms of impacts to the soil microbial community. Similarly, as invasive species are an
unfortunate byproduct of the modern world and rates of introductions show no signs of slowing
down, more studies of native plants and their interactions with established invasive species can
increase our knowledge of which natives are the most resistant to invasion, which may help
inform management and restoration decisions in the future.

51

References

Barndt, K. J. (2008). An Assessment of the Allelopathic Potential of Herb Robert (Geranium
robertianum) [M.S.]. University of Washington.
Barney, J. N., Tekiela, D. R., Dollete, E. S., & Tomasek, B. J. (2013). What is the “real” impact
of invasive plant species? Frontiers in Ecology and the Environment, 11(6), 322–329.
https://doi.org/10.1890/120120
Batish, D. R., Singh, H. P., Kaur, S., Priyanka, & Kohli, R. K. (2013). Novel weapons hypothesis
for the successful establishment of invasive plants in alien environments: A critical
appraisal. In D. R. Batish, R. K. Kohli, S. Jose, & H. P. Singh (Eds.), Invasive Plant
Ecology (pp. 19–28). CRC Press.
Bertin, R. I. (2001). Life cycle, demography, and reproductive biology of herb Robert (Geranium
robertianum). Rhodora, 103(913), 96–116.
Blair, A. C., Nissen, S. J., Brunk, G. R., & Hufbauer, R. A. (2006). A Lack of Evidence for an
Ecological Role of the Putative Allelochemical (±)-Catechin in Spotted Knapweed
Invasion Success. Journal of Chemical Ecology, 32(10), 2327–2331.
https://doi.org/10.1007/s10886-006-9168-y
Boerner, R. E. J. (1990). Role of mycorrhizal fungus origin in growth and nutrient uptake by
Geranium robertianum. American Journal of Botany, 77(4), 483–489.
https://doi.org/10.1002/j.1537-2197.1990.tb13579.x

52

Braga, R. R., Jeschke, J. M., Heger, T., Aparicio, L. G., & Simoes Vitule, J. R. (2018).
Invasional Meltdown Hypothesis. In J. M. Jeschke & T. Heger (Eds.), Invasion Biology:
Hypotheses and Evidence (Vol. 9, pp. 79–91). CABI.
Callaway, R. M. (2011). Novel Weapons Hypothesis. In D. Simberloff & M. Rejmánek (Eds.),
Encyclopedia of Biological Invasions (pp. 779–781). University of California Press.
Callaway, R. M., & Ridenour, W. M. (2004). Novel Weapons: Invasive Success and the
Evolution of Increased Competitive Ability. Frontiers in Ecology and the Environment,
2(8), 436–443. https://doi.org/10.1890/1540-9295(2004)002[0436:NWISAT]2.0.CO;2
Catford, J. A., Jansson, R., & Nilsson, C. (2009). Reducing redundancy in invasion ecology by
integrating hypotheses into a single theoretical framework. Diversity and Distributions,
15(1), 22–40. https://doi.org/10.1111/j.1472-4642.2008.00521.x
Chappell, C. B. (2004). Upland Plant Associations of the Puget Trough Ecoregion, Washington.
Washington Department of Natural Resources (Washington Natural Heritage Program).
https://www.dnr.wa.gov/publications/amp_nh_upland_puget.pdf
Crystal-Ornelas, R., & Lockwood, J. L. (2020). The ‘known unknowns’ of invasive species
impact measurement. Biological Invasions, 22(4), 1513–1525.
https://doi.org/10.1007/s10530-020-02200-0
Dara, S. K., Barringer, L., & Arthurs, S. P. (2015). Lycorma delicatula (Hemiptera: Fulgoridae):
A New Invasive Pest in the United States. Journal of Integrated Pest Management, 6(1),
20. https://doi.org/10.1093/jipm/pmv021

53

Davis, M.A., Grime, J.P. & Thompson, K. (2000). Fluctuating resources in plant communities: a
general theory of invasibility. Journal of Ecology, 88, 528–534.
Del Fabbro, C., Güsewell, S., & Prati, D. (2014). Allelopathic effects of three plant invaders on
germination of native species: A field study. Biological Invasions, 16(5), 1035–1042.
https://doi.org/10.1007/s10530-013-0555-3
Duncan, R. P., & Williams, P. A. (2002). Darwin’s naturalization hypothesis challenged. Nature,
417(6889), 608–609. https://doi.org/10.1038/417608a
Falińska, K. and Pirožnikow, E. (1983). Ecological structure of Geranium robertianum L.
populations under natural conditions and in the garden in Poland. Polish Journal of
Ecology. 31(1), 93–121.
Fertig, W. (2021). Plant of the Week: Large-leaf Avens (Geum macrophyllum). U.S. Forest
Service. Retrieved October 31, 2021, from https://www.fs.fed.us/wildflowers/plant-of-theweek/geum_macrophyllum.shtml
Giblin, D.E. & B.S. Legler (eds.). (2003+). WTU Image Collection Web Site: Vascular Plants,
Macrofungi, & Lichenized Fungi of Washington State. University of Washington
Herbarium. Accessed 10 Oct 2021.
http://biology.burke.washington.edu/herbarium/imagecollection.php
Gleason, H. A. & Cronquist, A. (1991). Manual of Vascular Plants of Northeastern United
States and Adjacent Canada. The New York Botanical Garden, Bronx, NY.
Gurevitch, J., & Padilla, D. (2004). Are invasive species a major cause of extinctions? Trends in
Ecology & Evolution, 19(9), 470–474. https://doi.org/10.1016/j.tree.2004.07.005

54

Hagan, D. L. & Jose, S. (2013). Mycorrhizae and alien plants. In D. R. Batish, R. K. Kohli, S.
Jose, & H. P. Singh (Eds.), Invasive Plant Ecology (pp. 79–92). CRC Press.
Hitchcock, C. L., & Cronquist, A. (2018). Flora of the Pacific Northwest: An illustrated manual
(D. Giblin, B. Legler, P. F. Zika, & R. G. Olmstead, Eds.; Second edition). University of
Washington Press.
Inderjit. (2011). Allelopathy. In D. Simberloff & M. Rejmánek (Eds.), Encyclopedia of
biological invasions (pp. 61–63). University of California Press.
Inderjit, von Dahl, C. C., & Baldwin, I. T. (2009). Use of silenced plants in allelopathy
bioassays: A novel approach. Planta, 229(3), 569–575. https://doi.org/10.1007/s00425008-0856-4
Ivancheva, S. & Petrova, A. (2000). A chemosystematic study of eleven Geranium species.
Biochemical Systematics and Ecology, 28, 255-260.
Jose, S., Batish, D. R., Kohli, R. K., Bardhan, S., & Singh, H. P. (2013). Invasive plant ecology:
The horse behind the cart? In S. Jose, H. P. Singh, D. R. Batish, & R. K. Kohli (Eds.),
Invasive Plant Ecology (pp. 1–6). CRC Press.
Jose, S., & Holzmueller, E. J. (2013). What makes alien plants so successful? Exploration of the
ecological basis. In H. P. Singh, D. R. Batish, R. K. Kohli, & S. Jose (Eds.), Invasive Plant
Ecology (pp. 7–18). CRC Press.
Kueffer, C. (2017). Plant invasions in the Anthropocene. Science, 358(6364), 724–725.
https://doi.org/10.1126/science.aao6371

55

Lankau, R. A. (2013). Species invasion alters local adaptation to soil communities in a native
plant. Ecology, 94(1), 32–40. https://doi.org/10.1890/12-0675.1
Leger, E. A., Goergen, E. M., & Forbis de Queiroz, T. (2014). Can native annual forbs reduce
Bromus tectorum biomass and indirectly facilitate establishment of a native perennial
grass? Journal of Arid Environments, 102, 9–16.
https://doi.org/10.1016/j.jaridenv.2013.10.015
Leishman, M. R., & Harris, C. J. (2011). Seed Ecology. In D. Simberloff & M. Rejmánek (Eds.),
Encyclopedia of Biological Invasions (pp. 985–991). University of California Press.
Lewis, M. A., & Jerde, C. L. (2012). Invasion Biology. In A. Hastings & L. J. Gross (Eds.),
Encyclopedia of Theoretical Ecology (pp. 384–391). University of California Press.
Mangla, S., Sheley, R. L., James, J. J., & Radosevich, S. R. (2011). Intra and interspecific
competition among invasive and native species during early stages of plant growth. Plant
Ecology, 212(4), 531–542. https://doi.org/10.1007/s11258-011-9909-z
Mason, R. A. B., Cooke, J., Moles, A. T., & Leishman, M. R. (2008). Reproductive output of
invasive versus native plants. Global Ecology and Biogeography, 17(5), 633–640.
https://doi.org/10.1111/j.1466-8238.2008.00402.x
Parepa, M., & Bossdorf, O. (2016). Testing for allelopathy in invasive plants: It all depends on
the substrate! Biological Invasions, 18(10), 2975–2982. https://doi.org/10.1007/s10530016-1189-z
Parker, I. M., Simberloff, D., Lonsdale, W. M., Goodell, K., Wonham, M., Kareiva, P. M.,
Williamson, M. H., Holle, B. V., Moyle, P. B., Byers, J. E., & Goldwasser, L. (1999).

56

Impact: Toward a framework for understanding the ecological effects of invaders.
Biological Invasions, 1(1), 3–19. https://doi.org/10.1023/A:1010034312781
Pyšek, P., & Richardson, D. M. (2006). The biogeography of naturalization in alien plants.
Journal of Biogeography, 33(12), 2040–2050. https://doi.org/10.1111/j.13652699.2006.01578.x
Reinhart, K. O., & Callaway, R. M. (2006). Soil biota and invasive plants. New Phytologist,
170(3), 445–457. https://doi.org/10.1111/j.1469-8137.2006.01715.x
Roumet, C., Urcelay, C., & Díaz, S. (2006). Suites of root traits differ between annual and
perennial species growing in the field. New Phytologist, 170(2), 357–368.
https://doi.org/10.1111/j.1469-8137.2006.01667.x
Rúa, M., Somereet, N., Johnson, A., Rogers, W., & Siemann, E. (2008). Experimental
approaches to test allelopathy: A case study using the invader Sapium sebiferum.
Allelopathy Journal, 22(1), 1–14.
Sax, D. F., & Gaines, S. D. (2003). Species diversity: From global decreases to local increases.
Trends in Ecology & Evolution, 18(11), 561–566. https://doi.org/10.1016/S01695347(03)00224-6
Simberloff, D. (2013). Invasive species: What everyone needs to know. Oxford University Press.
Stinson, K. A., Campbell, S. A., Powell, J. R., Wolfe, B. E., Callaway, R. M., Thelen, G. C.,
Hallett, S. G., Prati, D., & Klironomos, J. N. (2006). Invasive Plant Suppresses the Growth
of Native Tree Seedlings by Disrupting Belowground Mutualisms. PLoS Biology, 4(5),
e140. https://doi.org/10.1371/journal.pbio.0040140

57

Thorpe, A. S., Archer, V., & DeLuca, T. H. (2006). The invasive forb, Centaurea maculosa,
increases phosphorus availability in Montana grasslands. Applied Soil Ecology, 32(1), 118–
122. https://doi.org/10.1016/j.apsoil.2005.02.018
Thorpe, A. S., Thelen, G. C., Diaconu, A., & Callaway, R. M. (2009). Root exudate is
allelopathic in invaded community but not in native community: Field evidence for the
novel weapons hypothesis. Journal of Ecology, 97(4), 641–645.
https://doi.org/10.1111/j.1365-2745.2009.01520.x
Tofts, R. J. (2004). Geranium robertianum L. Journal of Ecology, 92(3), 537–555.
USDA NRCS National Plant Data Team. (2021). USDA NRCS Plants Database. USDA NRCS
Plants Database. https://plants.sc.egov.usda.gov/home
Van Der Heijden, M. G. A., Bardgett, R. D., & Van Straalen, N. M. (2008). The unseen majority:
Soil microbes as drivers of plant diversity and productivity in terrestrial ecosystems.
Ecology Letters, 11(3), 296–310. https://doi.org/10.1111/j.1461-0248.2007.01139.x
Von Holle, B. (2011). Invasional Meltdown. In D. Simberloff & M. Rejmánek (Eds.),
Encyclopedia of biological invasions. University of California Press.
Von Holle, B., & Simberloff, D. (2005). Ecological resistance to biological invasions
overwhelmed by propagule pressure. Ecology, 86(12), 3212–3218.
https://doi.org/10.1890/05-0427
WA NWCB. (1997). Written findings of the Washington State Noxious Weed Control Board –
Geranium robertianum L. https://www.nwcb.wa.gov/images/weeds/Geraniumrobertianum-1997.pdf

58

WA NWCB. (2021). Washington State Noxious Weed Control Board. Tree of Heaven Ailanthus Altissima. https://www.nwcb.wa.gov/weeds/tree-of-heaven
Weidenhamer, J. D., & Callaway, R. M. (2010). Direct and Indirect Effects of Invasive Plants on
Soil Chemistry and Ecosystem Function. Journal of Chemical Ecology, 36(1), 59–69.
https://doi.org/10.1007/s10886-009-9735-0
Weißhuhn, K., & Prati, D. (2009). Activated carbon may have undesired side effects for testing
allelopathy in invasive plants. Basic and Applied Ecology, 10(6), 500–507.
https://doi.org/10.1016/j.baae.2008.10.009
Wilcove, D.S., Rothstein, D., Dubow, J., Phillips, A., & Losos, E. (1998). Quantifying threats to
imperiled species in the United States. BioScience, 48, 607-615.
WNPS. (2021). Tellima grandiflora. Washington Native Plant Society. Stanley, G. B., &
Verlinde-Azofeifa, S. (Eds.). https://www.wnps.org/native-plant-directory/27-tellimagrandiflora
Wurst, S., Vender, V., & Rillig, M. C. (2010). Testing for allelopathic effects in plant
competition: Does activated carbon disrupt plant symbioses? Plant Ecology, 211(1), 19–26.
https://doi.org/10.1007/s11258-010-9767-0
Zouhar, K. (2001). Centaurea stoebe subsp. micranthos. Fire Effects Information System (FEIS).
USDA, Forest Service, Rocky Mountain Research Station, Fire Sciences Laboratory.
https://www.fs.fed.us/database/feis/plants/forb/censtom/all.html

59

Appendices
Appendix A
Table 1. Data analysis summary for native species by response variable.
Data Analysis Summary for Geum macrophyllum
Response Variable
Aboveground
Biomass
Aboveground
Biomass per plant
Belowground
Biomass
Belowground
Biomass per plant
Total
Biomass
Total Biomass per
plant
AG:BG Ratio

Meets
assumptions
of normality?

Homogeneity
of Variance?

Transformation
Used

Test Used

Post Hoc Used

Yes

Yes*

Log10

ANOVA

Tukey’s HSD

Yes*

Yes*

Log10

ANOVA

Tukey’s HSD

Yes

Yes

-

ANOVA

Tukey’s HSD

No

Yes

-

KruskalWallis

Dunn’s Test

Yes

Yes

-

ANOVA

Tukey’s HSD

No

Yes

-

No

Yes

-

KruskalWallis
KruskalWallis

Dunn’s Test
Dunn’s Test

Data Analysis Summary for Tellima grandiflora
Aboveground
Biomass
Aboveground
Biomass per plant
Belowground
Biomass
Belowground
Biomass per plant
Total
Biomass
Total Biomass per
plant
AG:BG Ratio

Yes*

Yes

Log10

ANOVA

Tukey’s HSD

Yes*

Yes

Log10

ANOVA

Tukey’s HSD

Yes

Yes

-

ANOVA

Tukey’s HSD

Yes

Yes

-

ANOVA

Tukey’s HSD

Yes

Yes

-

ANOVA

Tukey’s HSD

Yes

Yes

-

ANOVA

Tukey’s HSD

Yes

Yes

-

ANOVA

Tukey’s HSD

ANOVA

Tukey’s HSD

Data Analysis Summary for Dicentra formosa
Aboveground
Biomass
Belowground
Biomass
Total Biomass

Yes*

Yes

Log10

No

Yes

-

No

Yes

-

KruskalWallis
KruskalWallis

Dunn’s Test
Dunn’s Test

*After transformation

60

Table 2. Data analysis summary for Geranium robertianum by response variable.
Data Analysis Summary for Geranium robertianum grown with G. macrophyllum
Response Variable
Aboveground
Biomass
Belowground
Biomass
Total
Biomass
AG:BG Ratio

Meets
assumptions
of normality?

Homogeneity
of Variance?

Transformation
Used

Test Used

Post Hoc Used

Yes*

Yes*

Log10

ANOVA

Tukey’s HSD

Yes

Yes*

Log10

ANOVA

Tukey’s HSD

Yes*

Yes*

Log10

ANOVA

Tukey’s HSD

Yes*

Yes*

Log10

ANOVA

Tukey’s HSD

Data Analysis Summary for Geranium robertianum grown with T. grandiflora
Aboveground
Biomass
Belowground
Biomass
Total
Biomass
AG:BG Ratio

Yes

Yes

-

ANOVA

Tukey’s HSD

Yes

Yes

-

ANOVA

Tukey’s HSD

Yes

Yes

-

ANOVA

Tukey’s HSD

Yes

Yes

-

ANOVA

Tukey’s HSD

Data Analysis Summary for Geranium robertianum grown with D. formosa
Aboveground
Biomass
Belowground
Biomass
Total
Biomass
AG:BG Ratio
*After transformation

Yes

Yes

-

ANOVA

Tukey’s HSD

Yes

Yes

-

ANOVA

Tukey’s HSD

Yes

Yes

-

ANOVA

Tukey’s HSD

Yes

Yes

-

ANOVA

Tukey’s HSD

61

Appendix B
Table 1. Mean aboveground (AG), belowground (BG), total biomass, and aboveground to
belowground ratio (AG:BG) for each treatment group by species; includes percent of total
biomass for aboveground and belowground biomass. GERO = Geranium robertianum; GEMA =
Geum macrophyllum; TEGR = Tellima grandiflora; DIFO = Dicentra formosa.
Species & Treatment
Groups
G. macrophyllum
Control
Invaded
Carbon
Soil
T. grandiflora
Control
Invaded
Carbon
Soil
D. formosa
Control
Invaded
Carbon
Soil
GERO w/GEMA
Invaded
Carbon
Soil
GERO w/TEGR
Invaded
Carbon
Soil
GERO w/DIFO
Invaded
Carbon
Soil

Mean AG
Biomass

Mean BG
Biomass

Mean
Total
Biomass

AG:BG
Ratio*

AG %
of
Total

BG %
of
Total

1.150 g
0.468 g
0.340 g
0.937 g

0.464 g
0.283 g
0.260 g
0.348 g

1.614 g
0.751 g
0.600 g
1.285 g

2.48
1.65
1.31
2.69

71%
62%
57%
73%

29%
38%
43%
27%

0.702 g
0.498 g
0.722 g
0.675 g

0.791 g
0.611 g
0.694 g
0.563 g

1.493 g
1.110 g
1.416 g
1.238 g

0.89
0.82
1.04
1.20

47%
45%
51%
55%

53%
55%
49%
45%

0.465 g
0.498 g
0.284 g
0.193 g

2.842 g
1.649 g
1.097 g
1.521 g

3.306 g
2.147 g
1.381 g
1.714 g

0.16
0.30
0.26
0.13

-

-

0.485 g
1.170 g
0.451 g

0.110 g
0.152 g
0.082 g

0.595 g
1.322 g
0.534 g

4.41
7.70
5.50

82%
89%
84%

18%
11%
16%

0.914 g
0.730 g
0.935 g

0.220 g
0.213 g
0.166 g

1.134 g
0.942 g
1.101 g

4.15
3.43
5.63

81%
77%
85%

19%
23%
15%

1.136 g
0.650 g
0.805 g

0.247 g
0.214 g
0.233 g

1.383 g
0.864 g
1.038 g

4.60
3.04
3.45

82%
75%
78%

18%
25%
22%

*Mean AG:BG ratio by group: GEMA = 2.03; TEGR = 0.99; DIFO = 0.21; GERO w/GEMA = 5.87;
GERO w/TEGR = 4.40; GERO w/DIFO = 3.70; all GERO groups = 4.66.

62

Appendix C

Figure 1. Effect of treatments on biomass of all three native species; green and brown portions represent aboveground and belowground
biomass, respectively (note: numbers on y-axis are not negative). Values are means and standard errors for each treatment group. DICFOR
= Dicentra formosa; GEUMAC = Geum macrophyllum; TELGRA = Tellima grandiflora
Figure 1

63