-
Title
-
Pollen limitation and reproductive success in five South Puget Sound prairie plants
-
Date
-
2021
-
Creator
-
Richard, Savannah
-
Identifier
-
Thesis_MES_2021_RichardS
-
extracted text
-
Pollen Limitation and Reproductive Success in Five South Puget Sound Prairie Plants
by
Savannah Richard
A Thesis
Submitted in partial fulfillment
Of the requirements for the degree
Master of Environmental Studies
The Evergreen State College
December 2021
©2021 by Savannah Richard. All rights reserved.
This Thesis for the Master of Environmental Studies Degree
by
Savannah Richard
has been approved for
The Evergreen State College
by
_______________________________
Sarah Hamman Ph.D.
Member of Faculty
December 14, 2021
_______________________________
Date
ABSTRACT
Pollen Limitation and Reproductive Success in Five South Puget Sound Prairie Plants
Savannah Richard
Disruption of pollen distribution can negatively impact pollen quality and quantity, leading to
reduced seed set. Pollen limitation is measured by comparing pollen supllementation treatments
to naturally pollinated treatments. If pollen supplemented treatments increase in reproductive
success then pollen limitation is occurring. Pollen limitation is caused by land degradation and
environmental disruptions that change pollinator networks and pollen deposition. Better
understanding the patterns of pollen limitation and reproductive success can aide in restoration
efforts for sensitive plant species. This study included three native south Puget Sound prairie
plant species (Eriophyllum lanatum, Lupinus lepidus, and Plectritis congesta) and two nonnative invasive plant species (Hypochaeris radicata and Leucanthemum vulgare). These plants
were sampled across six restoration sites varying in land use and restoration treatment history.
No evidence of pollen limitation was found but reproductive success was shown to vary among
restoration sites. L. lepidus had the highest seed quantity at a high-quality restoration site. This
result aligned with previous predictions of native plant species having higher reproductive
success at high-quality restoration sites. This differs from P. congesta which had the highest seed
quality at a low-quality restoration site. These findings highlight the complex interactions of
plant life history, habitat fragmentation, and restoration practices on plant reproductive success.
Table of Contents
List of Figures ................................................................................................................................. v
List of Tables ................................................................................................................................. vi
Acknowledgements ....................................................................................................................... vii
Introduction ................................................................................................................................... 1
Research Questions and Hypothesis ............................................................................................... 4
Literature Review ......................................................................................................................... 6
Background ................................................................................................................................. 6
Pollen Deposition and Fertilization ............................................................................................ 7
Evolution of Floral Morphology in Relationship to Pollination ................................................. 9
Theoretical Framework ............................................................................................................. 13
Pollen Supplementation Studies ............................................................................................... 19
Habitat Fragmentation .............................................................................................................. 21
Invasive Plant Species............................................................................................................... 24
Restoration Applications ........................................................................................................... 26
Methods ........................................................................................................................................ 28
Study Site .................................................................................................................................. 28
Study Species ............................................................................................................................ 31
Experimental Design ................................................................................................................. 32
Data Collection ......................................................................................................................... 32
Statistical Analysis .................................................................................................................... 33
Results .......................................................................................................................................... 36
Treatment Effect ....................................................................................................................... 36
Seed Quality – Individual Seed Weight .................................................................................... 36
Seed Quality – Seed Viability ................................................................................................... 38
Quantity – Proportion of Fertilized Seeds ................................................................................ 39
Discussion..................................................................................................................................... 42
Treatment Effect ....................................................................................................................... 42
Plectritis congesta ..................................................................................................................... 43
Lupinus lepidus ......................................................................................................................... 45
Eriophyllum lanatum ................................................................................................................ 46
Hypochaeris radicata................................................................................................................ 46
Leucanthemum vulgare ............................................................................................................. 47
Conclusion ................................................................................................................................... 48
iv
List of Figures
Fig 1: Willamette Valley Puget Trough Georgia Basin .................................................................. 2
Fig 2: Pollen Deposition Diagram .................................................................................................. 6
Fig. 3: Sequential and Simultaneous Flowering Diagram ............................................................ 12
Fig. 4: Haig and Wetoby Equilibrium Model ............................................................................... 14
Fig. 5: Stochastic Resource Allocation ......................................................................................... 16
Fig. 6: Stochastic Resource Allocation in Low and High Quality Environments ........................ 17
Fig. 7: IPU and Resource Allocation ............................................................................................ 21
Fig. 8: Pollinator Trait Responses to Habitat Fragmentation ....................................................... 22
Fig. 9: Ecological Restoration Practices ....................................................................................... 26
Fig. 10: Plant Species .................................................................................................................... 31
Fig. 11: P. congesta Seed Weight ................................................................................................. 37
Fig. 12: L. vulgare Total Individual Seed Weight ........................................................................ 37
Fig. 13: H. radicata Fertilized Seed Weight ................................................................................. 38
Fig. 14: H. radicata Proportion of Viable Seeds .......................................................................... 39
Fig. 15: L. lepidus Proportion of Fertilized Seeds ........................................................................ 40
Fig. 16: H. radicata Proportion of Fertilized Seeds...................................................................... 40
v
List of Tables
Table 1: CNLM Restoration Treatment Application ................................................................... 28
Table 2: Plant Life History .......................................................................................................... 30
vi
Acknowledgements
This project was made possible by a grant from The Evergreen State College Master’s in
Environmental Studies thesis fund. This grant assisted with purchasing laboratory supplies. I
would also like to thank the Science Support Center, Jenna Nelson, John Kirkpatrick, and Dylan
Fischer for facilitating lab access during the COVID-19 pandemic. I would especially like to
thank Susan Waters for providing me with the opportunity to intern on multiple projects and
introducing me to the world of pollination ecology. Susan Waters also graciously shared her
data, seed samples and pollination ecology knowledge with me. She also assisted me with coding
in R and statistical analysis. I would also like to thank Sanders Freed and the Center of Natural
Lands Management for providing management plans for the sites used in this study.
My success during the thesis process was highly influenced by the support and guidance
of my thesis advisor Sarah Hamman. Sarah greatly aided in the thesis writing process and
statistical analysis. Her feedback and expertise were essential for my successful completion of
this project. I would also like to thank members from my cohort: Tim Atkinson and Sarah
Larson, for their advice and support. Lastly, I would like to thank my life partner Jesse Hunnicutt
who has been a constant source of emotional support and humor.
vii
Introduction
Pollination services are essential to plant reproduction and underpin healthy functioning
ecosystems. Although important, pollination ecology receives little research especially in the
context of ecological restoration science. Animal mediated pollination is responsible for 88% of
flowering plant species sexual reproduction globally (Cariveau et al. 2020). This means most of
the plants used in restoration are impacted by changes in pollinator networks and pollen
limitation.
Plant reproduction hinges on pollinator networks, pollen deposition and pollen quality. If
there is a disruption in the pollination process this has impacts on not only plant reproduction but
plant populations and ecological functions (Harder and Barrett 1996, Ashman et al. 2004, Aizen
and Harder 2007). This is of particular concern for the restoration of rare plants or endangered
plants that are experiencing limited pollen quantity or quality. Limits in pollen deposition can
change plant community compositions and drive evolutionary changes such as an increase in
self-pollination (Knight et al. 2005).
A major contributing factor to pollen limitation is land degradation and land use
conversion which results in fragmented landscapes. Such ecological disruptions can alter
pollinator networks and restrict access to plant populations (Knight et al. 2005). Invasive plant
species colonization can also play a role in pollen limitation. Invasive species often produce
excessive amounts of pollen which can effectively clog plant stigmas and limit access to
conspecific pollen (Arceo-Gomez et al. 2016).
1
South Puget Sound Prairies
The Pacific Northwest is the unlikely
home to rare prairie ecosystems. Known as the
Willamette Valley-Puget Trough-Georgia
Basin Ecoregion, prairie habitat extended from
British Columbia to the Willamette Valley,
Oregon, now only fragmented remnants remain
(Fig.1) (Hamman et al. 2011). In Washington
remnant prairies exist in the South Puget Sound
region clustered in Joint Base Lewis-McChord
(JBLM) and near Olympia. These prairies were
formed by the retreating of the Vashon glacier
over 14,000 years ago. This geologic` event
caused glacial outwash soils rich in gravel,
Fig 1: Willamette Valley Puget Trough Georgia
Basin extending from British Columbia to
Southern Oregon (Hamman et al. 2011).
which drain quickly and are ideal for
grasslands (Kruckeburg 1991).
Climatologically, South Puget Sound
prairies differ significantly from midwestern prairies. The growing season for perennial forbs
and grasses begins in the winter and extends into spring. While many perennial native plants
experience a summer dormancy, due to drought conditions (Sinclair et al. 2006). Midwestern
prairie vegetation primarily consists of perennial grasses with less forb diversity. The Pacific
Northwest experiences increased levels of precipitation from October to May, but the Puget
Sound trough exists in a rain shadow, shielded from high levels of precipitation by the Olympic
2
Mountain range (Bowcutt and Hamman 2016). The average rainfall for Olympia, Washington for
2020 was 127 cm, while the mean temperature was 16°C (NOAA). The temperate climate of the
Pacific Northwest also means that temperatures rarely drop below freezing in lowland regions.
This contrasts with Midwestern U.S. prairies that experience freezing temperatures throughout
the fall and winter months.
Pacific Northwest prairies contain mosaics of plant communities ranging from Garry oak
woodlands, savannas, grasslands, and wetlands. The primary vegetation composition consists of
perennial grass and forbs, annual forbs and sparsely dispersed shrub and trees (Sinclair et al.
2006). Of the recorded 278 plant species on South Puget Sound prairies in the early 2000’s, 59%
were native and the remaining 41% were nonnative (Dunwiddie et al. 2006). Of the 23,000 acres
of Puget Sound lowland prairies only 3,000 acres have predominantly native plant species
(Storm 2006).
The disappearance of prairies in this region began with the arrival of European colonists.
Starting in the 1800’s prairie land was used for agriculture and grazing livestock. As the Euro
American population expanded so did the development of residences, towns, and roads. This
development heavily fragmented prairies, impacting animal populations and altering pollinator
networks. Introduction of invasive plant species and tree encroachment further degraded prairie
sites (Kruckeburg 1991).
Prior to European settler colonization, Indigenous peoples-maintained prairie lands
through frequent, low intensity, controlled burns. These burns promoted growth of plant species
used for food, medicine, and fibers. Burning prairies also eliminated tree establishment and
created an open landscape that attracted elk and deer (Storm 2002). The disappearance of fire
3
from the landscape has resulted in reduced native plant and animal diversity, proliferation of
invasive species and encroachment of trees and shrubs.
Since the cessation of Indigenous land management practices such as controlled burns,
95-99% of western Washington prairies have been lost (Hamman et al. 2011). A small portion of
prairies remain although scattered and heavily fragmented. Currently prairie sites are managed
by Washington State Department of Fish and Wildlife, Washington State Department of Natural
Resources, Joiint Base Lewis McChord (JBLM) and non-profit land trusts such as the Center for
Natural Lands Management and Capital Land Trust. Restoration efforts include prescribed burns,
herbicide treatments, Douglas fir removal, mowing and mechanical removal of invasive species
and native planting and seeding.
Research Questions and Hypothesis
This thesis project aims to address three over-arching questions. (1) Is pollen limitation
occurring for any plant species? (2) Does reproductice success vary by restoration site? (3) How
do native and non-native plant species differ in reproductive success among sites? The first
question will address if pollen limitation is occurring by comparing hand pollinated treatments to
open pollination. If seed quantity or quality is significantly higher in hand pollinated treatments,
then we can presume that pollen limitation is occurring. If there is no difference between
treatments, then pollen limitation is not occurring. The second question investigates the impact
of restoration sites on seed quantity and quality. The third question is a qualitative comparison of
non-native invasive species and native species reproductive success among restoration sites.
Six prairie restoration sites were chosen in the south Puget Sound region with varying land
use histories and restoration treatments. Some sites have decades of invasive species
management and native plant reintroduction while others have low-quality prairie with a limited
4
history of restoration. If a restoration site is particularly associated with pollen limitation for
multiple plant species, then the restoration history can be examined to improve future practices.
For this study I hypothesized that plant species experiencing pollen limitation will be
associated with prairie sites with limited restoration history. I also surmise that native plants will
have greater reproductive success at comparatively more restored sites while invasive species
will be reproductively successful no matter the restoration history of the site. The null hypothesis
for this study is that seed quantity and quality are not affected by restoration site or pollination
treatment.
5
Literature Review
Background
Flowering plants rely on abiotic and biotic vectors as a means of sexual reproduction, due
to their immobile nature. This process is known as pollination and involves a multitude of
interactions that influence reproduction and floral evolution (Harder and Barrett 1996, Knight et
al. 2006). For successful animal-facilitated pollination to take place, a pollinator must deposit
pollen from one member of a species to another (Fig.2). Intraspecific plant species pollen, also
known as conspecific pollen, must land on the receptive area of the plant, the stigma. Chemical
reactions then occur triggering the growth of a pollen tube to the ovum. Once the pollen tube
reaches the ovum fertilization occurs and seeds are produced (Cheung 1996).
Pollen limitation is a potential outcome that may occur when there is insufficient pollen
quantity being produced or distributed. The quality of pollen can also negatively influence plant
reproduction, resulting in aborted ovules or infertile seeds (Harder and Barrett 1996, Ashman et
al. 2004). Pollen limitation can result from a multitude of external factors such as habitat
Fig 2: Pollen Deposition Diagram. Pollen deposition diagram displaying the complex process of insect
mediated pollination among hermaphroditic plants of the same species. Anthers are shown exporting pollen
while stigmas are shown receiving pollen. Ovules are colored to match the corresponding fertilizing individual
(Anderson and Minnaar 2020).
6
fragmentation, decreases in pollinator populations or environmental disruptions (Alvarez 2002,
Knight et al. 2005, Aguilar et al. 2006, Newman et al. 2013). The evolutionary life history and
reproductive strategy of plants can also play an integral role in pollen limitation.
Plant populations experiencing pollen limitation can be negatively impacted through
reduced reproductive success and population declines. Chronic pollen limitation can lead to
plants selecting for self-pollination as a preferred reproductive strategy. This switch to selfpollination greatly reduces genetic complexity in plant populations and could potentially lead to
plant communities being less evolutionarily adaptable. Although, plants that have evolved to be
hermaphroditic or partially self-compatible have a much lower rate of pollen limitation. It is
thought that these are reproductive strategies are used to avoid pollen limitation (Knight et al.
2005).
Pollen limitation studies are often used to assess the availability of pollen for a plant
species in a specific environment. To test for pollen limitation, researchers implement pollen
supplementation experiments. This type of natural experiment involves supplementing a portion
of flowers with conspecific pollen and allowing a control group to be openly pollinated.
Reproductive success is then quantified by seed set or other metrics to account for the
supplemental pollen impact. If the hand pollinated treatment has a higher rate of reproductive
success when compared to the control group, then pollen limitation is determined to be occurring
(Ashman et al. 2004).
Pollen Deposition and Fertilization
Pollination can be a major driving force in floral evolution. Pollen donors compete to
fertilize ovules by locating females, excluding other pollen donors, and being accepted by female
7
receptive floral anatomy (Burd 1994, Snow 1994). Locating female flowers involves strategies
such as having a similar phenology, sharing pollinators, or increasing patch density. Research
has shown that reproductive success is not limited by the male flower population but by the
ability of males to locate females (Burd 1994). This is especially true for dioecious plants that
have male and female flowers on separate plants. There is less possibility of selfing but more
energy spent locating intraspecific plants of the opposite sex.
Hermaphroditic flowers differentiate reproductice structures on temporal or physical
scales to avoid self pollination (Harder and Barrett 1996). When self pollination occurs, it can
cause pollen discounting, which reduces the amount of pollen grains for cross-pollination. To
limit the effects of pollen discounting, strategies such as having separate timing for male and
female reproductive parts within an individual inflorescence are used. Herkogamy is another
common strategy which involves a physical separation of anthers and stigmas. Heterostyly is a
form of herkogamy where the stigma grows at a different length than the anthers. This can
reduce pollen discounting costs across other floral species morpho-types rather than intraspecific
individual flowers (Harder and Barrett 1996).
These evolutionary competitive dynamics also play out on the microscopic scale during
post-pollination on the stigma. Plants can exert little sexual selection control over the pollen
being deposited on the stigma but post pollination process such as the rate of pollen tube growth
or fertilization rates are a few ways in which selection can take place (Burd 1994, Snow 1994).
After male pollen granules germinate on the stigma a pollen tube grows towards the ovary.
Pollen tube growth rates vary depending on genetic makeup. Biochemical reactions can occur on
the surface of the stigma inhibiting pollen tube growth. This can occur for self-incompatible
plants when self-pollen is deposited on the stigma (Cheung 1996). Post pollination selection is
8
referred to as “female choice” due to the selective pressures exerted on the pollen from the
stigma (Snow 1994).
Evolution of Floral Morphology in Relationship to Pollination
Plant life history and physiology, impact the ways in which pollen is produced, received, and
distributed (Harder and Barrett 1996, Fenster et al. 2004, Knight et al. 2005). Plants associated
with abiotic pollen distribution such as wind, tend to produce larger quantities of pollen and have
smaller, less colorful floral displays. This contrasts with animal pollinated plants that often
produce larger and more colorful floral display to attract pollinators (Harder and Barrett 1996,
Fenster et al. 2004). Elaborate floral displays and nectar rewards require allocation of more
resources. This resource allocation causes less pollen to be produced than wind pollinated
flowers but there is higher efficiency at conspecific pollination (Knight et al. 2005).
Flowering plants associated with specialist pollinators tend to be more impacted by pollen
limitation than generalist pollinator associated plants (Knight et al. 2005, Fernández et al. 2012,
Hagen et al. 2012, Newman et al. 2013). Plants associated with specific insect functional groups
co-evolve morphological structures that aide in pollination (Harder and Barrett 1996, Fenster et
al. 2004). This leads to a reduction in diversity of overall pollinator visitations but improves
conspecific pollination rates and limits heterospecific pollen transfer. The limited diversity of
pollinator visits means specialist associated plants are more susceptible to pollen limitation due
to changes in pollinator networks or patch density (Knight et al. 2005, Fernández et al. 2012,
Newman et al. 2013). Pollinator preference is often expressed through morphological patterns
such as radially symmetrical flowers attracting generalist pollinators and bilaterally symmetrical
flowers attracting specialist pollinators (Harder and Barrett 1996). These morphological
9
expressions correspond with low pollen limitation for radially symmetrical flowers and a higher
risk of pollen limitation among bilaterally symmetrical flowers (Knight et al. 2005).
It is commonly observed that hermaphroditic flowers are more prevalent than mono-sexual
(dioecious) flower types (Knight et al. 2005). A potential advantage of hermaphroditic flowers is
the ability for outcrossing and self-pollination (Lloyd 1992). By retaining both modes of
reproduction hermaphroditic flowers can ensure reproduction when stochastic environmental
variations occur. Hermaphroditic flowers are also known to over produce floral structures
compared to seed capsule production (Ehrlen 1991, Burd 1998). This could be due to pollen
limitation where not all inflorescences are receiving adequate pollen deposition. Alternatively,
Cohen and Durkas (1990) conclude that hermaphroditic floral production increases in relation to
increased pollination variance. Hermaphroditic flower structures also require less resource
investment than complete female flowers (Wesselingh 2007). This means the overproduction of
flowers in hermaphroditic flowers is explained by the low cost of flower production and an
increased variance in pollination (Ehrlen 1991, Burd 1998).
Pollen deposition can vary among multiple inflorescences on one plant depending on
arrangement and bloom time. Sequential flowering or when flowers bloom on the same plant at
varying times, allows the opportunity for a response to low pollen deposition. If pollen
deposition is low a sequential blooming plant can reallocate resources to produce more
subsequent flowers (Wesselingh 2007). Whereas a plant that has a synchronous bloom of flowers
all at the same time does not have the opportunity to reallocate resources to produce more
flowers later (Fig.3). Instead, synchronous blooming plants invest in more flowers than
sequential blooming plants and have higher rates of fruit abortion (Casper and Neissenbaum
1993).
10
The spatial arrangement of inflorescence on a plant can also cause varying amounts of pollen
deposition among flowers. Accounting for how flowers are arranged in relation to one another
and how many flowers per inflorescence are all important factors when considering variations in
pollen deposition. (Wesselingh 2007). Basal flowers on plants with sequential blooms, tend to
bloom earlier and produce high quantity and quality seeds, while distal flowers bloom later and
produce less seed set (Diggle 1995, Ashman and Hitchens 2000). The spatial arrangement of
flowers on a stalk can influence pollinator interaction and ultimately seed set.
Phenology of flowers on a plant is also an important factor in pollen deposition. Plants with
extended floral bloom times tend to not experience as much variability in seed set as plants with
shorter bloom cycles (Galen and Stanton 1991). This is due to variation in environmental
conditions and pollinator networks and visitations. Galen and Stanton (1991) found that late
season blooming alpine Ranunculus experienced reduced seed weight compared to earlier season
flowers. This is a general pattern described for sequentially blooming flowers. Early season basal
flowers tend to serve as resource sinks and produce higher quantity and quality seeds. Later
season flowers suffer from limited resources and tend to produce seeds of less quantity and
quality (Casper and Neissenbaum 1993).
11
Fig. 3: Sequential and Simultaneous Flowering Diagram
(a) Shows a sequential flowering plant where (pd) represents pollen deposition. The
thickness of lines shows the strength of pollen deposition. Diagram (b) represents
simultaneous flowering where a set number of resources are predetermined no matter the
pollen deposition (pd) (Wesselingh 2007).
12
Theoretical Framework
Pollen limitation research is often carried out using pollen supplementation studies based
on multiple theories foundational to evolutionary biology. The optimality theory is based on
Darwin’s theory of natural selection and is used in evolutionary biology modeling to determine
fitness. Natural selection presumes that organisms are in constant competition to evolve adaptive
traits or behaviors to maximize fitness. This maximization of fitness translates to the assumption
that in an ideal environment, organisms will reproduce and function at an optimum rate (Parker
and Smith 1990, Knight et al. 2005). If this is not the case, external pressures such as
competition or environmental changes may be to blame. For pollen limitation studies, the
optimality theory assumes that a plant existing in an optimal abiotic and biotic environment will
not be any more reproductively successful when more pollen is introduced. If the plant has
higher reproductive success when pollen is introduced this would indicate that an ecological
disturbance is taking place which is limiting pollen (Knight et al. 2005).
Rosenheim et al. (2010) outlines Liebig’s law of the minimum which states that plant
growth is limited by a single nutrient or external factor. When Liebig’s law is applied in
ecological experiments, typically nutrients are added and then plant fitness is quantified to
determine which is the limiting growth factor. This law also applies to pollen limitation
experiments. If more pollen is added through hand pollination and this positively impacts plant
reproduction, then we can assume pollen availability is a key limiting factor. Depending on the
plant and environment there may be other simultaneous limiting factors negatively impacting
plant reproduction.
In Haig and Westoby’s (1988) seminal paper, they created a graphical model describing
plant resource allocation and reproductive success. The authors reasoned that ovule fertilization
13
would increase when increased resources were allocated to pollinator attraction; this is known as
the fitness gain curve. The ability of ovules to mature and form seeds would decrease because of
the resources diverted to pollinator attraction. On the graphical model this is known as the
resource cost curve (Fig. 4). Haig and Westoby (1988) refer to the intersection of the fitness gain
curve and the resource costs curve as equilibrium. They argued that plants would evolve traits to
maximize attractiveness to pollinators while maintaining enough resources for adequate seed
production. If a plant population were to reach equilibrium, then supplementation of pollen
would have no effect on seed set.
Fig. 4: Haig and Wetoby Equilibrium Model. The pollen-limitation function increases as resources are
spent in pollen attraction effort. The provisioning-limitation function decreases as resources are
allocated to pollen attraction. This reduces effectively pollinated ovules and seed set. The intersection
of the provisioning-limitation function and the pollen-limitation function (a*) represents the
equilibrium (Haig and Westoby1988).
Fig. 5: Haig and Wetoby Equilibrium Model. The pollen-limitation function increases as resources are14
spent in pollen attraction effort. The provisioning-limitation function decreases as resources are
allocated to pollen attraction. This reduces effectively pollinated ovules and seed set. The intersection
of the provisioning-limitation function and the pollen-limitation function (a*) represents the
equilibrium (Haig and Westoby1988).
Burd (2008) revisited the Haig and Westoby equilibrium model and revised it by adding
stochastic variation. Stochastic variation includes variability in pollen supply, resource allocation
and environmental factors from year to year. The first model (Fig. 5) represents a plant species
response to stochastic resource allocation. The line labeled ‘R’ represents mean resource
constraint, while the dotted lines on either side represent high and low variations of resource
allocation. ‘F’ represents number of ovules fertilized as resources are allocated to pollinator
attractiveness. There are also two dotted lines on either side of F to represent the upper and lower
stochastic variations. This model shows an equilibrium (a*) at the same point as Haig and
Westoby’s model, where lines F and R intersect. The dashed line ‘St+R’ represents average seed
set, which would be impacted by resource variability.
The second model created by Burd (2008) (Fig. 6) represents a plant species existing in
two habitats, where one is richer in resources than the other. ‘R1’ represents a level of resource
allocation while R0 represents a low level of resource allocation. ‘S0’ represents the range of seed
fertilization in a low-quality environment while S1 represents the range of seed fertilization in a
high-quality environment. The low-quality environment with limited resource allocation has a
lower mean seed set (a0) than the high-quality environment with increased resource allocation
(a1). This model differs from the Haigh and Westoby equilibrium model, showing how stochastic
variations in environmental resources cause variations in seed set. Burd’s second model can be
applied in pollen limitation studies to better understand how environmental conditions can
impact resource allocation in plants. In poor quality habitats with habitat fragmentation or low
species richness, the plant may be operating at a lower than optimal equilibrium.
15
Fig.5: Stochastic Resource Allocation
The Burd (2008) model shows a plant species that has stochastic resource allocation, where F
represents mean ovule production and the dotted lines on either side represent low and high
Fig. 12: Stochastic
Resource
Allocation
variations
in response
to resource
allocation that is driven by resource constraint (R).
Fig. 8: Stochastic Resource AllocationThe Burd (2008) model shows a plant species that has
Fig. 13: Stochastic
Resource Allocation
stochastic
resource allocation,
where F represents mean ovule production and the dotted lines on
either side represent low and high variations in response to resource allocation that is driven by
resource constraint (R).
Fig. 14: Stochastic Resource Allocation
Fig. 9: Stochastic Resource Allocation
Fig. 10: Stochastic Resource AllocationThe Burd (2008) model shows a plant species that has
stochastic resource allocation, where F represents mean ovule production and the dotted lines on
either side represent low and high variations in response to resource allocation that is driven by
resource constraint (R).
Fig. 11: Stochastic Resource AllocationThe Burd (2008) model shows a plant species that has
stochastic resource allocation, where F represents mean ovule production and the dotted lines on
either side represent low and high variations in response to resource allocation that is driven by
resource constraint (R).
16
Fig.6: Stochastic Resource Allocation in Low and High Quality Environments
Burd’s (2008) second revised model represents a single plant species’ ovule production response
(F) to its allocation to pollinator attraction, driven by a stochastic environment, where S0
Fig.
20: IPU
and quality
Resource
AllocationFig.
Stochastica high
Resource
Allocation
represents
a low
environment
and S21:
quality
environment.
1 represents
in Low and High Quality Environments
Fig. 15:
Stochastic
Resource
Allocationanalysis
in Low and
Quality EnvironmentsBurd’s
(2008)
Recent
theoretical
framework
andHigh
meta-analysis
has suggested that
pollen
second revised model represents a single plant species’ ovule production response (F) to its
Fig.
22: Stochastic
Resource
Allocation
and Highenvironment,
Quality
allocation
to pollinator
attraction,
driven in
byLow
a stochastic
where S0 represents a low
limitation
may
not
only
be
a
consequence
of
ecological
disturbance
but also a response to
Environments
quality environment and S1 represents a high quality environment.
stochastic
variation in pollen delivery, plant life history and ovule fertilization (Ashman 2004,
Fig. 16: Stochastic Resource Allocation in Low and High Quality Environments
Fig. 23: IPU and Resource AllocationFig. 24: Stochastic Resource Allocation
in Low2006,
and High
Knight
BurdQuality
2008).Environments
Pollen limitation may vary among plant species even if co-flowering
Fig. 17: IPU and Resource AllocationFig. 18: Stochastic Resource Allocation in Low and High
inQuality
the same
vicinity. This could
be due
to differing
life histories
reproductive
strategies such
EnvironmentsBurd’s
(2008)
second
revised model
representsand
a single
plant species’
ovule production response (F) to its allocation to pollinator attraction, driven by a stochastic
where S0 represents
a low quality
environment
and S1 represents
a high
quality
asenvironment,
rates of self-compatibility
or specialist
pollinator
associations.
Plants with
higher
rates of selfenvironment.
compatibility tend to be less impacted by pollen limitation (Knight 2006).
Fig. 19: Stochastic Resource Allocation in Low and High Quality EnvironmentsBurd’s (2008)
secondThe
revised
model represents
single plantdetected
species’ in
ovule
production
response
(F) tomay
its be
prevalence
of pollena limitation
studies
suggests
that plants
allocation to pollinator attraction, driven by a stochastic environment, where S0 represents a low
quality environment and S1 represents a high quality environment.
adapted to stochastic environments that experience varied amounts of pollen distribution (Burd
1994). Knight et al. (2006) suggests that the phenomenon of frequent pollen limitation may be
17
due to a hedge-betting strategy of over-producing flowers in case an influx of pollen distribution
occurs. Such an adaptation would allow a plant to take advantage of pollen fluctuations to create
a higher seed set. This hedge-betting strategy theory conflicts with Haig and Westoby’s model
which is predicated on the concept that resource allocation has costs for plants, hence plants
would tend to not over-produce flowers. Knight et al.’s (2006) meta-analysis reveals that a high
percentage of pollen limitation studies have found a high proportion of flowers to set seed,
indicating an overproduction of flowers. If plants are existing at equilibrium then we would
expect that flower to seed set ratios would be equal.
Cohen and Dukas’ (1990) findings point to both resource allocation and hedge-betting
strategies. Cohen and Dukas (1990) conclude that over production of hermaphroditic flowers in
relationship to seed set, is due to increased variance in pollination and resource allocation.
Hermaphroditic floral structures require less resources for production then a complete female
flower (Wesselingh 2007). Over production of flowers is also an adaptation to stochastic
pollination, allowing plants to take advantage of influxes of conspecific pollen (Ehrlen 1991,
Burd 1993).
An opposing theory to pollen limitation is genetic load theory. This theory suggests that
self-incompatible plants dependent on outcrossing pollination have higher genetic loads than
self-compatible plants. This theory presumes that low levels of sexual reproduction may be due
to genetic load rather than a lack of availability of pollen (Charlesworth et al. 1990). If genetic
load is the underlying cause of reduced plant reproduction, then pollen supplementation through
hand pollination would produce negligible impacts compared to naturally pollinated plants (Burd
1994). To entirely resolve the issue of genetic load being an underlying cause, selfing rates of
plants would need to be determined.
18
Pollen Supplementation Studies
In recent years there has been considerable reexamination of pollen limitation studies
methodologies. This has been due in part to reviews of pollen supplementation experiments
revealing that the majority of research resulted in pollen limitation (Burd 1994, Ashman et al.
2004, Knight et al. 2005). This can be interpreted as most plants studied are experiencing pollen
limitation or there are faults in the experimental design. Seed set is often counted as the only
measurement of pollen limitation. This leads to the appearance of pollen limitation without
taking into consideration pollen deposition, phenology, and plant life history. Due to the ongoing
debate surrounding quantifying pollen limitation, many studies have cited the need for more
measurements that quantify pollen limitation such as pollen quantity and quality (Burd 1994,
Ashman et al. 2004, Knight et al. 2006, Aizen and Harder 2007).
A common proxy measurement for pollen quantity is seed set, due to the direct effects of
pollen deposition and seed production. Pollen quality can impact post-pollination process of
ovule fertilization and viable seed development. Low quality pollen can result in unsuccessful
pollen tube growth, unfertilized or aborted seeds (Ashman et al. 2004, Aizen and Harder 2007).
Poor pollen quality can occur for self-incompatible plants when self-pollen is deposited on the
stigma and for self-compatible plants experiencing an inbreeding depression that receive selfpollen (Aizen and Harder 2007). Pollen quality can be accounted for by measuring pollen tube
growth, seed viability rates, seed weight, counting aborted seeds or determining pollen
deposition ratios on the stigma (Ashman et al. 2004). Seed weight has been found to correspond
to seedling survivability and plant establishment (Galen and Stanton 1991, Aizen and Harder
2007). Accounting for pollen quality in pollen limitation experiments can provide a more
accurate picture of pollen distribution and plant reproduction.
19
A meta-analysis of pollen limitation studies done by Knight et al. (2006), found several
ways in which to improve pollen limitation studies moving forward. Their research found that
previous hand pollination experiments did not measure amount of pollen being applied and
applied pollen as a singular treatment. Over pollinating flowers could reduce the production of
more flowers in the future due to resource allocation amongst flowers. A single pollen
supplementation treatment does not capture flowers produced later in the season. Unequal
resource allocation could also take place if fractions of plants are supplemented with pollen
(Zimmerman and Pyke 1988). More research still needs to be done on best practices when
conducting pollen supplementation experiments but treating all flowers multiple times
throughout the phenology period can account for variables such as resource allocation.
Determining an appropriate morphological unit of measurement for plants sampled in
pollen limitation studies is an important factor to consider when designing pollen
supplementation experiments. Wesselingh (2007) suggests using an integrated physiological unit
(IPU) which is used to group morphological arrays into functional units or subunits. An IPU can
be an individual flower, inflorescence, or an entire plant (Fig. 7). Plant life history is an
important factor when considering an appropriate IPU. For instance, perennial plant species are
often able to re-allocate unused resources to be used in preceding seasons, while annual plants
cannot (Diggle 1995, Casper and Neissenbaum 1993, Wesselingh 2007). In some cases it has
been shown that inflorescences comprised of many florets act as a single morphological unit, for
instance the ray and disc florets that compose an Asteraceae inflorescence. If singular florets are
removed there is not a significant difference in pollen or seed production (Herrera 1991).
20
Fig.7: IPU and Resource Allocation
This diagram shows integrated physiological unit (IPU) of an entire plant (dashed line). Dotted
lines indicate individual flower. Arrows denote allocation of resources (Wesselingh 2007).
Fig. 29: IPU and Resource Allocation
Fig.Habitat
25: Pollinator
Trait Responses to Habitat FragmentationThis diagram shows integrated
Fragmentation
physiological unit (IPU) of an entire plant (dashed line). Dotted lines indicate individual flower.
Fig. 30:denote
Pollinator
Trait Responses
to Habitat
Fragmentation
Arrows
allocation
of resources
(Wesselingh
2007).
Anthropogenic environmental disturbances through changes in land use have caused
Fig. 26: IPU and Resource Allocation
Fig.
31: Pollinator
Traitdeclines.
ResponsesHabitat
to Habitat
32:are
IPUdriving
and Resource
global
biodiversity
lossFragmentationFig.
and fragmentation
forces Allocation
in changing
pollinator
networks
andAllocationThis
plant communities
et al. 2012,
Newman etunit
al. (IPU)
2013,of
Cariveau
et
Fig.
27: IPU and
Resource
diagram(Hagen
shows integrated
physiological
an
Fig. 33: IPU and Resource Allocation
entire plant (dashed line). Dotted lines indicate individual flower. Arrows denote allocation of
resources
al. 2020).
(Wesselingh
This occurs
2007).
through large expanses of habitat with intricate plant – animal interactions
Fig.being
28: Pollinator
Traitinto
Responses
Habitat
FragmentationThis
diagram
integrated
transformed
severaltosmall,
isolated
patches. This
spatialshows
separation
disrupts pollinator
physiological unit (IPU) of an entire plant (dashed line). Dotted lines indicate individual flower.
Arrows
denotealters
allocation
resources (Wesselingh
2007).
networks,
plantofcommunity
composition,
and contributes to pollen limitation and
21
ultimately reduces plant reproduction. (Knight et al. 2005, Aguilar et al. 2006, Hagen et al. 2012,
Kaiser-Bunbury 2017, Bennett et al. 2018). These cascading effects of alterations in pollinator –
plant relationships are important to understand when attempting to restore ecosystem functions
and conserve sensitive plant species.
Fig.8: Pollinator Trait Responses to Habitat Fragmentation
A framework displaying differing pollinator traits and responses to habitat fragmentation (HF)
(Hagen et al. 2012).
Fig. 34: Pollinator Trait Responses to Habitat Fragmentation
A framework
displaying
differing
pollinatorplant
traitslife
andhistories
responsesand
to habitat
fragmentation
(HF)react
Research
has shown
that varying
reproductive
strategies
(Hagen et al. 2012).
Fig. 35: Pollinator Trait Responses to Habitat Fragmentation
differently to fragmented habitats (Aguilar et al. 2006, Hagen et al. 2012, Bennett et al. 2018).
A framework displaying differing pollinator traits and responses to habitat fragmentation (HF)
(Hagen et al. 2012).
Fig. 36:
Pollinator
Trait
Responsespollination
to Habitat Fragmentation
Plants
with
generalist
associated
syndromes have shown a greater capability to cope
A framework displaying differing pollinator traits and responses to habitat fragmentation (HF)
with
fragmented
habitats on a localized scale (Fig.8) (Newman et al., 2013). Alternatively,
(Hagen
et al. 2012).
Aguilar et al. (2006) found that generalist and specialist pollinator associated plants both
responded negatively to fragmented habitats with reduced seed set. In addition, self-incompatible
plant reproduction were negatively impacted by fragmentation, while self-compatible plants
were not. This outcome illustrates the adaptive advantages of self-compatibility and the
22
vulnerability to ecological changes self-incompatible plants.
For rare plant species habitat fragmentation can be especially dangerous, driving
population declines. Ashman et al. (2004) suggests the Allee effect plays a role in pollen
limitation and reduced reproduction. The Allee effect is a biological concept that refers to the
positive association between population size and fitness. In pollination ecology low density floral
patches attract fewer pollinators and thus suffer from pollen limitation. The Allee effect can
exacerbate this situation by negatively impacting seed quantity or quality. In self-compatible
plant communities, selfing rates increase as patch density declines (Fernández et al. 2012). This
can lead to an inbreeding depression due to lack of genetic out crossing.
Overall plant population size can also cause Allee effects regardless of floral patch
density. If plant species are self-incompatible then the reduction in available compatible mates
will produce pollen limitation symptoms and lower seed set (Ashman et al. 2004). A reduced
species population could also suffer from increases in heterospecific pollen transfer, which can
block receptive areas on the stigma for conspecific pollen (Arceo-Gomez et al. 2016). These
scenarios would happen regardless of pollinator visitations or reduced patch density due to lack
of outcrossing.
Declines in plant populations can also lead to pollen limitation and potential Allee effects
(Agren 1996). Small population size can lead to reduction in self-compatible alleles leading to
inbreeding (Hagen et al. 2012). This can lead to the Allee effect and threaten small populations
with reproductive incompatibility and inability to self-pollinate (Knight 2004). The Allee effect
is known to impact small populations more severely and can lead to risk of extinction. Better
understanding minimum population sizes for plant species can aide in reintroduction efforts of
threatened or endangered plant species.
23
Changes in floral patch densities is another form of habitat fragmentation that can lead to
pollen limitation. Reductions in floral patch densities has been associated with increases in
pollen limitation (Knight 2004). Newman et al. (2013) found that fragmentation on a localized
spatial scale drastically impacted plant reproduction. Bare percent ground cover of over 40% was
found to be associated with total reproductive failure of plants sampled. These findings
demonstrate that fragmentation on a local level can drastically impact pollen distribution and
subsequently depress reproduction.
Invasive Plant Species
Introduced invasive plant species are one of the largest ecological threats to native plant
species causing habitat fragmentation and disruptions of pollinator networks resulting in pollen
limitation and reduced reproductive capacity (Flory and Claye 2009, Morales and Traveset 2009,
Orrock and Witter 2010). Invasive plant species often have adaptive characteristics that allow
them to alter their environment and transform conditions for co-flowering native plant
communities. This can include changing plant community composition, floral patch densities and
altering pollinator networks (Lopezaraiza et al. 2007, Orrock and Witter 2010). Some invasive
plants can produce large quantities of pollen which can effectively clog native plants stigmas and
block conspecific pollen reception (Kanchan and Chandra 1980, Arceo-Gomez and Ashman
2016). These environmental disruptions can ultimately negatively impact native plant pollen
distribution and plant reproduction (Flory and Claye 2009, Morales and Traveset 2009, Orrock
and Witter 2010).
Invasive plant species often share pollinators with native plant species, causing either
facilitation or disruptions in visitations or pollen quality (Orrock and Witter 2010). In some
24
cases, invasive plant species can act as a pollinator magnet, drawing more pollinator visitations
to floral patches (Lopezaraiza et al. 2007, Ramula and Pihlaja 2012, Waters et al. 2014). This
pollinator magnet effect can impact native plants differently depending on pollinator preferences.
pollination network analysis by Lopezaraiza et al. (2007) found that invasive plant species I.
glandulifera attracted an increase of pollinator richness and abundance to the co-flowering patch.
Although I. glandulifera received the most pollinator visitations and native flowers received
increases in heterospecific pollen transfer. This study reveals the pollinator magnet effects of
invasive species on native plant communities and the disproportionate rewards reaped by the
invading species.
Introduced invasive species have been shown to cause both facilitative and competitive
effects on native plant communities (Morales and Traveset 2009, Waters et al. 2014). Research
by Waters et al. (2009) showed that two native plants have differing responses to the introduction
of an invasive plant species. One native (M. laciniata) had reduced seed set in floral patches with
the invasive species (H. radicata). While the other native (E. lanatum) experienced a higher seed
set in the presence of the invasive non-native. This research highlights the complex nature of
invasive plant species colonization impacts on native plant communities.
Additionally, Morales and Traveset’s (2009) research shows that morphological
similarity to the invading plant species is related to facilitative impacts on native plants. Their
research revealed native plant seed set was reduced in the presence of invasive species. However,
native plants that had morphological similarities to the invasive species did not have reduced
seed set or pollinator visitation.
25
Restoration Applications
Common ecological restoration practices used to improve prairie habitat include
prescribed burns, invasive species removal, mowing, seeding and out-planting native species.
Restoration treatments applied on the landcape scale can increase habitat quality and address
habitat connectivity. Treatmetns focused on the local scale can more directly impact biological
communities and pollination functions (Fig.9). These treatments can be tailored to include plant-
Fig.9: Ecological Restoration Practices
Restoration treatments on the landscape level have impacts on habitat. Restoration treatments on the
local level have direct impacts on biotic communities and pollination function (Cariveau 2020).
Table 2: CNLM Restoration Treatment Application for Each Site (Unpublished data, Susan
Waters).Fig.
41: Ecological
Practices
Fig.
37: Ecological
Restoration
treatments
on the
landscape
level haveregimes,
impacts
pollinator
relationships
suchRestoration
asPracticesRestoration
increasing
floral patch
densities,
altering
disturbance
on habitat. Restoration treatments on the local level have direct impacts on biotic communities and
pollination
function
(Cariveau
2020).Restoration treatments such as invasive species removal have
and
improving
pollinator
diversity.
Table 3: CNLM Restoration Treatment Application for Each Site (Unpublished data, Susan
Waters).
Fig.
38: Ecological Restoration Practices
been shown to improve pollinator network diversity, increase pollinator visitation rates and
Table 1:
4: Plant
history (Andersson
Adderlyfor
et Each
al. 2015).Table
5: CNLM
Restoration
Table
CNLMlife
Restoration
Treatment2008,
Application
Site (Unpublished
data,
Susan
Treatment
Application
for
Each
Site
(Unpublished
data,
Susan
Waters).Fig.
42:
Ecological
Waters).Fig. 39: Ecological Restoration PracticesRestoration treatments on the landscape level have 26
Restoration
Practices
impacts
on habitat.
Restoration treatments on the local level have direct impacts on biotic
communities and pollination function (Cariveau 2020).
Table 6: CNLM Restoration Treatment Application for Each Site (Unpublished data, Susan
increase seed set (Flory and Claye 2009, Morales and Traveset 2009, Orrock and Witter 2010,
Kaiser-Bunbury et al. 2017).
Restoring pollinator diversity has been associated with improving native plant seed set
and establishment. Research done by Albrecht et al. (2012) found a positive association between
pollinator diversity and native plant seed set, although this effect began to wane with the highest
level of pollinator diversity causing declines in seed set. This research shows that pollinator
diversity alone will not uniformly improve native plant seed set. It is likely that plant association
with specific pollinators and pollinator functional group abundance are important components to
include with diversity (Sabatino et al. 2021). Using increased biodiversity as a restoration metric
may not be appropriate for every plant community. Better understanding site specific pollinator
network – plant relationships and measuring reproductive success could prove to be a more
useful tool in restoring pollination functions.
A promising approach in restoring ecological pollination function, is to improve habitat
connectivity (Kaiser-Bunbury 2017, Betts et al. 2019). Focusing on connecting restoration site
locations can bridge pollinator communities and support plants with self-incompatible life
histories in reproduction (Aguilar et al. 2006). This approach along with site specific pollen
limitation, reproductive success, and plant-pollinator relationships, are all important components
in restoring pollination functions to degraded ecosystems.
27
Methods
Study Site
Sampling sites were established from spring to summer of 2020, on six Puget Trough prairie
restoration sites. These sites vary in restoration history and land management techniques applied
(Table 1). Glacial Heritage Preserve is owned by Thurston County and was managed by the
Nature Conservancy and the Center for Natural Lands Management since 1994. Glacial Heritage
Preserve is renowned for having some of the highest quality prairie ecosystems in south Puget
Sound. In recent years Glacial Heritage was the site of Taylor’s checkerspot butterfly and
Castilleja levisecta reintroduction (personal corrrespondance, Sanders Freed).
Table 1: CNLM Restoration Treatment Application for Each Site (Unpublished data, Susan Waters).
Table 7: Plant life history (Andersson 2008, Adderly et al. 2015).Table 8: CNLM Restoration Treatment
Application for Each Site (Unpublished data, Susan Waters).
Table 9: Plant life history (Andersson 2008, Adderly et al. 2015).
Fig. 44: Plant SpeciesTable 10: Plant life history (Andersson 2008, Adderly et al. 2015).Table 11: CNLM
Restoration Treatment Application for Each Site (Unpublished data, Susan Waters).
Table 12: Plant life history (Andersson 2008, Adderly et al. 2015).Table 13: CNLM Restoration
Treatment
Application
for Each Site
(Unpublished
data, Susan
Waters).
This
site was originally
dominated
by Scotch
broom,
requiring years of intensive mowing
and invasive removal. Prescribed burns have taken place regularly since 2001 on 956 acres of the
1020 total acres (unpublished data, Susan Waters). The Scotch broom population has been
successfully reduced through regular herbicide treatments and mechanical and hand removal.
Other nonnative forbs targeted for herbicide treatments include Hypocheris radicata,
Leucanthemum vulgare and Jacobaea vulgaris.
28
Wolf Haven Preserve is in Tenino, Washington and is owned by Wolf Haven
International. The 36 acres prairie is managed by CNLM and contains rare plant species such as
Castilleja levisecta. This restoration site is small in comparison to other sites included in this
study but boasts high quality prairie. Regular prescribed burns have occurred at this site since
2009 with 22 acres burned and seeded (Unpublished data, Susan Waters).
Tenalquot Prairie Preserve is a 126 acres site established in 2006 and located near the
town of Rainier and JBLM. This property is predominantly grassland bordered by an oak and
Douglas fir forest. Prescribed burning has been conducted since 2007 and herbicide treatments
and native plant seeding began in 2008 (Unpublished data, Susan Waters). Established
populations of Castilleja levisecta currently exist at Tenalquot. This was achieved through plug
plantings and seedings beginning in 2007. The success of native plant community establishment
at this site has made it eligible for introduction of the federally listed Taylor’s Checkerspot
butterfly (personal correspondence, Sanders Freed).
Johnson Prairie is 194 acres of high-quality prairie located in a designated military
training area on Joint-Base Lewis McChord. This restoration site has a history of military related
ecological disturbances, which has benefited native prairie communities that are adapted to
disturbance regimes. CNLM began prescribed burn and herbicide treatmetns in 2009 and started
native seeding in 2014, however large section of this site are considered very high qulaity, with
diverse and abundant native species (unpublished data, Susan Waters).
Cavness Ranch is managed by CNLM through a conservation easement established in
2005. Cavness Ranch is 613 acres near Tenino, Washington. This site has a long history of
varied land use such as logging, agriculture, and cattle ranching. Currently, land within the
conservation easement is being used for Christmas tree farming and cattle grazing. This property
29
also has a diversity of habitats including wetlands, riparian forest, Oregon white oak - Douglas
fir forest, mixed forest, and grassland (personal correspondence, Sanders Freed).
Restoration at Cavness Ranch began in 2009 with prescribed burning prescribed burning.
Native seeding and plug planting followed starting in 2011, then herbicide treatments started in
2014 (unpublished data, Susan Waters). This site is heavily fragmented due to varying
ecosystems and historical land use activities throughout the property.
The 140 acres Deschutes River Preserve was acquired by CNLM in 2014. This property
was previously used as an equestrian center and cattle ranch. Of the sites sampled in this study
Deschutes River Preserve has had the least amount of restoration treatments. The majority of this
property (100 acres) is degraded prairie grasslands with little native plant cover. The remaining
portion of this property is dominated by Oregon white oak and Douglas fir forests. A high
priority restoration goal for this site is to improve habitat for the federally listed Mazama pocket
gophers, which has been observed at this site (personal correspondence, Sanders Freed).
Table 2: Plant life history (Andersson 2008, Adderly et al. 2015).
Fig. 45: Plant SpeciesTable 14: Plant life history (Andersson 2008, Adderly et al. 2015).
Table 15: Plant life history (Andersson 2008, Adderly et al. 2015).
Fig. 46: Plant SpeciesTable 16: Plant life history (Andersson 2008, Adderly et al. 2015).
30
Restoration treatments such as prescribed burns and herbicide treatments began in 2015 followed
by native seeding in 2016 (unpublished data, Susan Waters). Deschutes River Preserve has the
shortest restoration history of all sites sampled in this study and has the highest proportion of
degraded low-quality prairie.
Study Species
Five common prairie plant species were chosen to sample: Eriophyllum lanatum,
Plectritis congesta, Lupinus Lepidus, Leucanthemum vulgare and Hypochaeris radicata.
(Fig.10). The plants chosen in this study have varying floral morphologicies, life histories and
pollination syndromes (Table 2). Of these five plant species three are native species: Eriophyllum
lanatum, Plectritis congesta and Lupinus lepidus. Two plants are nonnative: Leucanthemum
vulgare and Hypochaeris radicata. Among the plants chosen, four are perennial (Eriophyllum
lanatum, Lupinus lepidus, Hypochaeris radicata and Leucanthemum vulgare) and one is an
Fig. 1049: Plant Species chosen for this study.
Photo credits: 1,2,3: Legler, Ben (2004), 4: Houck, Douglas (2012), 5: Skotland, Bruce (2004) Burke
Herbarium Image Collection.
Fig. 50: Plant Species
Plant species chosen for this study. Photo credits: 1,2,3: Legler, Ben (2004), 4: Houck, Douglas (2012),
5: Skotland, Bruce (2004) Burke Herbarium Image Collection.
31
Fig. 51: Plant Species chosen for this study.
Fig. 47: Plant Species chosen for this study.
Fig. 52: Plant Species
annual (Plectritis congesta). These plant species cannot be compared amongst each other due to
varying life histories, pollinator syndromes and floral morphologies.
Experimental Design
At each site selected plant species were located, although not all plant species occurred at
each site. Plants were randomly selected at each site then inflorescence were randomly selected
on each plant for sampling. Inflorescence were then subjected to one of two treatments: open
pollination or hand pollination. These paired treatments were conducted ten times for each
species at each site. Selected hand pollinated plants had their stigmas dusted with a paintbrush
containing conspecific pollen from that site. Sampled flowers were hand pollinated once then
were snooded with fine mesh nylon bags, as to not allow for pollinator access. Flowers selected
for open pollination were not hand pollinated nor snooded. This treatment group was allowed
access to pollinators.
Data Collection
Physical characteristics of each sampled plant were recorded such as plant height, flowering
stalk, and number of flowers. These characteristics did not apply evenly to each plant species due
to varying floral morphologies. Seeds were cleaned and counted by hand or using a dissecting
microscope. Due to the quantity of seed produced by L. vulgare, a count estimation technique
was used. Fertile seeds were visibly distinguished form infertile or immature seeds. Fertile and
infertile seeds were both counted, providing a total seed count. Fertile seeds were then collected
and weighed using an analytical scale, to determine individual fertile seed weight.
32
Two plant species were chosen for Tetrazolium staining process to determine seed viability
P. congesta and H. radicata. These plants were chosen to examine native and non-native and
annual and perennial differences. Seeds were stained using Tetrazolium: An important test for
physiological seed quality evaluation (Franca-Neto et al. 2019). A stain was created using 1%
2,3,5-triphenyltetrazolium chloride in a solution of deionized water. pH was adjusted to a 6.8-7.4
range. All fertilized seed samples were soaked in the staining solution for 48 hours at room
temperature. Seeds were then dissected to determine embryo staining. Red staining indicated
viability while no staining or a white embryo indicates non-viability.
To determine pollen limitation seed quantity and quality were compared to open
pollination and hand pollination treatments. Seed quantity was measured calculating proportion
of fertilized seeds out of total seed set. Seed quality was determined by individual seed weight
for each sample. Proportion of viable seeds out of total seed set was used as an additiona seed
quality measurement for P. congesta and H. radicata.
Statistical Analysis
Individual fertilized seed weight was calculated by dividing total fertilized seed set
weight by total fertilized seed set. For L. vulgare, fertilized, immature or aborted seeds were all
included in seed set count. Therefore, seed weight is a measurement of total individual seed
weight for L. vulgare.
Fertilized seed set was divided by total seed set to get a percentage of fertilized seed. For
L. vulgare, fertilized seed was indistinguishable from non-fertile seeds, so total seed count was
measured. Number of viable seeds was divided by total seeds to get the percentage of viable
seeds. Where needed the percentage values were arcsine square root transformed to better fit a
normal distribution.
33
Distribution normality for all other response variables were tested by graphing them out
in a histogram then running a Shapiro-Wilkes test. For non-normally distributed data, Akaike
Information Criterion (AIC) tables were used to compare negative binomial regression, gamma,
and Poisson distributions to select the best fit. Gamma distributions were used for nonparametric
continuous data and negative binomial regression, or Poisson were used for count or proportional
data.
Two factor mixed effect models were selected for each dependent variable. Covariates
such as height and number of flowering stalks were included in model selection as to account for
plant robustness. Pair number was included in the models as a random effect. Several nested
models were compared using the (AIC). Eight models were used in the selection process: (1) site
and treatment interactions, pair number as a random effect and number of flower stalks and
height (2) site and treatment interactions and flower stalks (3) site and treatment interactions and
height (4) site and treatment interactions (5) site and treatment (6) site (7) treatment (8) a null
model.
For all normally distributed parametric data a general linear model (GLM) was used. A
one-way analysis of variance (ANOVA) test was run on models selected with three or less
independent factors that had a normal distribution. An alpha of 0.05 was set for all tests. If
ANOVA tests had significant variables, then a Tukey post hoc test was run to compare each
variable against each other for significance.
For non-normally distributed continuous variables a gamma distributed general linear
mixed model (GLMM) was fit an estimated marginal means post hoc test was run to evaluate
pairwise comparisons. A pairwise p-value matrix was created to compare restoration sites
amongst each other. Non-normally distributed data that did not fit general linear models or
34
general linear mix model distributions, were tested using the Kruskal Wallace test and then a
Wilcoxon rank sum test was used to make pairwise comparisons.
35
Results
Treatment Effect
Pollen supplementation treatments and open pollination treatments were not significant
factors in any mixed models for any of the plant species in this study.This results means that
neither open or hand pollination treatment had any significant impact on seed quality or quantity
among restoration sites.
Seed Quality – Individual Seed Weight
Individual seed weight was found to significantly vary among sites for L. vulgare, P.
congesta and H. radicata. L. vulgare total individual seed weight varied significantly by
restoration site (Appendix A Table 1). Seed weight was significantly higher at Johnson (p=0.046
and Wolf Haven (p=0.014) than at Tenalquot (Fig.11). P. congesta fertilized seed weight had the
highest mean at Cavness when compared to all other restoration sites (Fig.12) (Appendix A
Table 2). Site was also significant for H. radicata individual fertilized seed weight (Appendix A
Table 1). A pairwise comparison for H. radicata showed that Glacial Heritage fertilized seed
weight was significantly lower than Deschutes (p=0.048) and Tenalquot (p<0.001) (Fig.13).
Seed weight did not significantly vary among restoration sites for E. lanatum (Appendix A Table
3) or L. lepidus.
36
Fig. 11: L. vulgare Total Individual Seed Weight
Significant differences in means are depicted by letters. Error bars denote one standard deviation
from the mean.
Fig. 11: L. vulgare Total Individual Seed Weight
Significant differences in means are depicted by letters. Error bars denote one standard deviation
from the mean.
Fig. 53: H. radicata Fertilized Seed Weight
Fig. 11: L. vulgare Total Individual Seed Weight
Fig. 54: H. radicata Proportion of Viable SeedsFig. 55: H. radicata Fertilized Seed
WeightFig.
11: L. vulgare
Total Individual
Seed Weight differences in means are depicted by
Fig.
11: L. vulgare
Total Individual
Seed WeightSignificant
letters. Error bars denote one standard deviation from the mean.
Fig. 11: L. vulgare Total Individual Seed Weight
Significant differences in means are depicted by letters. Error bars denote one standard deviation
from the mean.
Fig. 12: P. congesta Fertilized Seed Weight
P. congesta individual fertilized seed weight by restoration site. Error bars depict one standard deviation
from the mean. Significant differences in means are depicted using differing letters.
Fig. 11: L. vulgare Total Individual Seed WeightFig. 12: P. congesta Fertilized Seed Weight
37
Fig. 12: P. congesta Fertilized Seed Weight
Fig. 12: P. congesta Fertilized Seed Weight
Fig. 13: H. radicata Fertilized Seed Weight
Individual fertilized seed weight for H. radicata. Significant differences in means denoted by differing
letters. Error bars represent one standard error from the mean.
Fig. 56: H. radicata Proportion of Viable SeedsFig. 57: H. radicata Fertilized Seed Weight
Fig. 61:
H. radicata
Fertilized
Seed WeightIndividual fertilized seed weight for H. radicata. Significant
Seed
Quality
– Seed
Viability
differences in means denoted by differing letters. Error bars represent one standard error from the mean.
Fig. 58: H. radicata Fertilized Seed Weight
Proportion of fertilized seed viability varied among sites significantly for H. radicata
Fig. 62: H. radicata Proportion of Viable Seeds
Fig. 59: H.Aradicata
Proportion
of Viable SeedsFig.
60: H. were
radicata
Fertilized
Weight
(Appendix
Table 4),
but no significant
site variations
found
for P. Seed
congesta
(Appendix A
Fig. 63: H. radicata Proportion of Viable SeedsIndividual fertilized seed weight for H. radicata.
Table
5). Sites
Cavness,inWolf
and
had significantly
lower mean
seed viability
Significant
differences
meansHaven
denoted
byDeschutes
differing letters.
Error bars represent
one standard
error from
the mean.
than Glacial Heritage, Tenalquot and Johnson Prairie (Fig.14).
Fig. 64: H. radicata Fertilized Seed WeightIndividual fertilized seed weight for H. radicata. Significant
differences in means denoted by differing letters. Error bars represent one standard error from the mean.
38
Fig.14: H. radicata Proportion of Viable Seeds
Proportion of viable seeds for H. radicata. Significantly different means are represented by differing
letters. Error bars represent one standard deviation from the mean.
Fig. 65: H. radicata Proportion of Viable Seeds
Fig. 68: H. radicata Proportion of Fertilized SeedsProportion of viable seeds for H. radicata.
Significantly different means are represented by differing letters. Error bars represent one standard
Quantity
–radicata
Proportion
of Fertilized
Seeds
Fig.
66: H.from
Proportion
of Viable Seeds
deviation
the mean.
Theofproportion
offor
fertilized
seed Significantly
counts for L.different
Lepidusmeans
varied
bydiffering
restoration
Proportion
viable seeds
H. radicata.
aresignificantly
represented by
Fig. 67: H. radicata Proportion of Viable Seeds
letters. Error bars represent one standard deviation from the mean.
site (Appendix A Table 6). A pairwise comparison of each site showed that Glacial Heritage had
Fig. 69: H. radicata Proportion of Fertilized SeedsProportion of viable seeds for H. radicata.
Significantly
different
means are
by differing
letters.
Error bars(p<0.001)
represent one
standard
the
highest mean
proportion
of represented
fertilized seeds
compared
to Cavness
Tenalquot
deviation from the mean.
(p<0.001) and Wolf Haven (p<0.001) (Fig.15). H. radicata also had restoration site significance
for the proportion of fertilized seed counts (Appendix A Table 7). A pairwise comparison of each
site revealed that Deschutes has significantly lower proportion of fertilized seeds when compared
to Johnson Prairie (p=0.037), Tenalquot (p=0.007), and Cavness (p=0.02). Tenalquot has a
significantly higher proportion of fertilized seeds than Deschutes (p=0.007), and Wolf Haven
(p=0.02) (Fig.16). Proportion of fertilized seeds did not significantly vary among sites for
39
Fig.15: L. lepidus Proportion of Fertilized Seeds
L. lepidus proportion of fertilized seeds by site. Means were calculated using the Wilcoxon sum rank
test (Appendix A Table 6). Significant differences in means are represented by differing letters. Error
bars represent one standard error from the mean.
Fig. 78: L. lepidus Proportion of Fertilized SeedsL. lepidus proportion of fertilized seeds by site.
Means were calculated using the Wilcoxon sum rank test (Appendix A Table 6). Significant
differences in means are represented by differing letters. Error bars represent one standard error from
the mean.
Fig.16: H. radicata Proportion of Fertilized Seeds
H. radicata proportion of fertilized seeds by restoration site. Differing letters signify significant
differences in means. Error bars represent one standard deviation from the mean.
Fig. 73: L. lepidus Proportion of Fertilized SeedsFig. 74: H. radicata Proportion of Fertilized Seeds
Fig. 70: H. radicata Proportion of Fertilized Seeds
Fig. 75: H. radicata Proportion of Fertilized Seeds
Fig. 71: L. lepidus Proportion of Fertilized SeedsFig. 72: H. radicata Proportion of Fertilized SeedsH.
radicata proportion of fertilized seeds by restoration site. Differing letters signify significant
40
L. vulgare (Appendix A Table 8), P. congesta (Appendix A Table 9), or E. lanatum (Appendix A
Table 10).
41
Discussion
All plants sampled displayed varying responses in reproductive success across all
restoration sites. Pollen limitation was not found to be occurring for any plant species. Variation
in reproductive success is most likely due to the different life histories represented by the five
plants chosen. These results highlight the ecological differences among restoration sites resulting
from fragmentation and land use history.
Treatment Effect
In this study, there was no significant difference between pollen supplementation
treatments and open pollination treatments. For some mixed models, treatment was included but
for all measurements of seed quality and quantity treatment was not a significant factor. This
could be an indicator of the lack of pollen limitation among plant species and restoration sites. If
this is the case then this would indicate that the plants sampled have adequate pollen distribution
and in tact pollinator networks.
The overall lack of treatment significance could also indicate a field sampling error when
conducting pollen supplementation treatments. Hand pollination treatments were conducted on
subsets of an entire plant including a flower head or inflorescence. Number of flowering stalks
per plant sampled were accounted for and in some cases multiple flowering stalks from the same
plant were treated.
Selecting an appropriate integrated physiological unit (IPU) of measurement is an
important component of creating an accurate pollen limitation experiment. An IPU is described
by Wesselingh (2007) as a unit of measurement used to group morphological arrays or functional
subunits. In pollination studies this often means sampling subunits such as individual flowers or
treating an entire plant as one unit. This is done to account for variability of resource allocation
42
depending across plant life histories and reproductive strategies. Perennial plant species can store
unused resources in corms, rhizomes, and bulbs, for future use (Vico et al. 2016). Annual plants
are unable to store unused resources, which allows them to put extra energy into seed production.
These resource allocation trade offs vary depending on plant life history.
The complex nature of assigning an appropriate IPU for pollen limitation studies has
been researched by many pollination ecologists. In analysesis of pollen limitation methodologies
researchers suggest categorizing the entire plant as an IPU and conducting multiple hand
pollination treatments throughout the plant’s life cycle (Zimmerman and Pyke 1988, Ashman
2004). Multiple hand pollination treatments could account for varying flowering times and
variations in seasonality. While measuring the entire plant as an IPU may be appropriate for
some plants this would depend on plant life history and taxonomy.
In future pollen limitation experiments selecting an appropriate IPU for each plant
species should be carefully considered. This decision should be based on plant life history, plant
morphology and phenology. The addition of multiple hand pollination treatments throughout the
flowering season to plants such as Lupine that bloom sequentially could improve the accuracy of
pollen limitation detection.
Plectritis congesta
P. congesta was found to have higher seed quality at the low-quality prairie restoration
site Cavness. This outcome did not fit the original hypothesis predicting native plants would
have higher seed quality and quantity at high quality prairie sites. A study done by Ramula and
Pihlaja (2012) found that meadow plant species experienced declines in species richness after
non-native invasions. Although, invaded native meadow species also showed an increase in
43
reproductive output. This could be due to increased competition for resources that a non-native
invasion creates or the invasive species altering pollinator networks. Invasive species can change
pollinator networks and act as a magnet species by attracting pollinators to floral patches which
can temporarily positively impact native plant reproduction.
Research conducted by Trowbridge et al. (2016) found that P. congesta had increased in
reproductive success at sites where it was seeded after one burn treatment. This finding was
temporary, lasting only eight growing seasons before drastic declines in establishment were seen.
This contrasts with native perennials used in this study (E. lanatum, A. millefolium and R.
occidentalis) that had long-term persistence and increased abundance associated with a single
burn and seeding treatment. Differing plant life histories may explain these differing responses to
restoration treatments.
This aligns with Dunwiddie et al.’s (2014) research regarding the loss of native annuals
throughout Northwest prairie systems. Annuals such as P.congesta could require repeated burn
treatments to stimulate reproductive outputs. At Cavness, where the increase in P. congesta seed
quality was observed, regular burning and seeding is part of the restoration plan (personal
communication, Sanders Freed). This could explain the increase in seed quality seen at this lowquality site. Long term research is required to observe establishment rates and overall
reproductive success of this species at Cavness.
Research done by Waters et al. (2014) on two native Coast Salish prairie plants (E.
lanataum and M. laciniata) and one non-native (H. radicata), compared pollinator visitation
rates and plant seed set at native and invasive dominated neighborhoods. Native E. lanatum had
more pollinator visits and higher seeds set at the plot with invasive plants while M. laciniata had
more pollinator visits and higher seed set at native dominated floral communities. This research
44
reveals the facilitative and competitive interactions invasive plants have on native plant
communities.
Lupinus lepidus
In contrast with P. congesta, L. lepidus had significantly higher seed set at the highquality prairie site Glacial Heritage. This aligns with the original hypothesis of higher seed
quality occurring for native plants and higher-quality prairie sites. Glacial Heritage has
undergone over twenty years of invasive species removal and seasonal burning greatly reducing
non-native plant species richness and abundance. Reduction of invasive plant populations has
been shown to increase native plant diversity, species richness and biomass (Flory and Claye
2009).
One of the most impactful environmental disturbances facing native plant populations is
the invasion of non-native plant species. Mechanisms by which invasive plants colonize new
environments include high seed set production, heterospecific pollen transfer and changes in
pollination networks (Flory and Claye 2009). Invasive plant species not only alter environmental
conditions, but also increase competition for resources among native plants communities. This is
achieved through changes in floral patch density, abundance, and distribution, resulting in
reductions in reproductive success (Flory and Claye 2009, Orrock and Witter 2010). Meadow
dwelling forbs adapted to high levels of light exposure have been found to be particularly
negatively impacted by non-native plant invasions when compared to other habitat types
(Alvarez 2002).
45
Eriophyllum lanatum
E.lanatum was the only plant species sampled that had no significant difference of seed
quantity or quality among restoration sites. This is most likely due to E. lanatum selfcompatibility or asexual reproduction (Appendix B Figure 1). This adaptation allows plants to be
less impacted by environmental changes such as habitat fragmentation and pollinator networks
(Aguilar et al. 2006). Evolving asexual reproduction while retaining the ability to sexually
reproduce is an adaptive strategy when experiencing environmental disruptions that would
impact pollination (Knight et al. 2005). This could explain why E. lanatum showed no
significant variations among restoration sites for reproduction success.
Hypochaeris radicata
The results for H. radicata reveal complex relationships between seed quality, quantity,
and restoration sites. Cavness Ranch had significantly higher fertilized seed quantity when
compared to Deschutes, Glacial Heritage and Johnson Prairie. For restoration sites such as
Cavness, seed quality and quantity do not align. Seed quality measured by viability varied among
sites with Cavness, Deschutes and Wolf Haven significantly lower seed viability. Glacial
Heritage, Johnson Prairie and Tenalquot had significantly higher seed viability. Fertilized seed
weight also presented nuanced results with Glacial Heritage, having significantly lower seed
weight when compared to Tenalquot and Deschutes.
These findings show that measurements of seed quality and quantity do not vary
uniformly among restoration sites. The restoration site Cavness had high seed quantity with low
seed viability. This result shows that H. radicata produced high seed quantity at Cavness but
46
those seeds had comparatively low viability. This result could reveal environmental stressors
such as increased co-flowering competition that influences resources allocation.
Measurements of seed quality, such as seed weight and seed viability, are typically
expected to reveal similar results (Burd 1994, Ashman et al. 2005, Knight et al. 2006). This was
not the case for H. radicata at Glacial Heritage where seed viability was relatively high and seed
weight was significantly low when compared to other restoration sites. These conflicting results
illustrate the complexities of environmental differences at heavily fragmented sites.
Leucanthemum vulgare
The results of L. vulgare revealed that seed weight was significantly lower at Tenalqout
when compared to Wolf Haven and Johnson Prairie. Seed weight is correlated to seedling
survivorship and plant establishment (Galen and Stanton 1991, Aizen and Harder 2007). This
could indicate that the L.vulgare population at Tenalquot is being impacted by restoration
treatments causing low seed weight. L. vulgare is specifically listed as an invasive species of
restoration focus in the CNLM Tenalquot management plan (personal correspondence, Sanders
Freed).
47
Conclusion
This study produced surprising results that drastically differed from earlier hypotheses
and predictions. Pollen limitation was not detected for any plant species sampled at any
restoration site. The absence of treatment significance may be due to field sampling methods.
Alterations to pollen limitation field sampling methodologies have been outlined in this study
and could be applied to future pollen limitation studies.
All species sampled had varying responses in reproductive success among restoration
sites. L. lepidus seed quality followed the hypothesis that native plant species would display
higher reproductive success at the more restored sites. In contrast, P. congesta did not respond as
expected, producing the highest seed quality at a low-quality prairie site. Invasive plant species
reproductive success among sites showed more nuanced site specific results. Including other
variables such as soil chemistry, soil moisture and plant species richness in future studies on
plant reproductice success will elucidate these complex results (Gornish 2016).
The findings in this study can be utilized by restoration managers to better understand
variations in reproductive success among plant populatons for restoraiton sites. In the future
measurements of pollen limitation and reproductive success can be added to site specific
monitoring efforts. This can be especially important for assessing plant establishment for rare
plants of conservation concern. Better understanding complicated plant-pollinator interactions
can aide in restoring ecological functions to degraded landscapes and add to the greater body of
knowledge within pollination ecology.
48
Bibliography
Adderley, Lorraine & Vamosi, Jana. (2015). Species and Phylogenetic Heterogeneity in Visitation
Affects Reproductive Success in an Island System. International Journal of Plant Sciences. 176.
186-196. 10.1086/679617.
Aguilar, R., Ashworth, L., Galetto, L., & Aizen, M. A. (2006). Plant reproductive susceptibility to
habitat fragmentation: Review and synthesis through a meta-analysis. Ecology Letters, 9(8),
968–980. https://doi.org/10.1111/j.1461-0248.2006.00927.x
Aizen, M. A., & Harder, L. D. (2007). EXPANDING THE LIMITS OF THE POLLENLIMITATION CONCEPT: EFFECTS OF POLLEN QUANTITY AND QUALITY. Ecology,
88(2), 271–281. https://doi.org/10.1890/06-1017
Albrecht, M., Schmid, B., Hautier, Y., & Müller, C. B. (2012). Diverse pollinator communities
enhance plant reproductive success. Proceedings of the Royal Society B: Biological Sciences,
279(1748), 4845–4852. https://doi.org/10.1098/rspb.2012.1621
Alvarez, M. E., & Cushman, J. H. (2002). COMMUNITY-LEVEL CONSEQUENCES OF A PLANT
INVASION: EFFECTS ON THREE HABITATS IN COASTAL CALIFORNIA. Ecological
Applications, 12(5), 1434–1444. https://doi.org/10.1890/10510761(2002)012[1434:CLCOAP]2.0.CO;2
Anderson, B., & Minnaar, C. (2020). Illuminating the incredible journey of pollen. American Journal
of Botany, 107(10), 1323–1326. https://doi.org/10.1002/ajb2.1539
Andersson, S. (2008). Pollinator and nonpollinator selection on ray morphology in Leucanthemum
vulgare (oxeye daisy, Asteraceae). American Journal of Botany, 95(9), 1072–1078.
http://www.jstor.org/stable/41922355
49
Arceo‐Gómez, G., Abdala‐Roberts, L., Jankowiak, A., Kohler, C., Meindl, G. A., Navarro‐Fernández,
C. M., Parra‐Tabla, V., Ashman, T.-L., & Alonso, C. (2016). Patterns of among- and withinspecies variation in heterospecific pollen receipt: The importance of ecological generalization.
American Journal of Botany, 103(3), 396–407. https://doi.org/10.3732/ajb.1500155
Arceo-Gómez, G., Schroeder, A., Albor, C., Ashman, T.-L., Knight, T. M., Bennett, J. M., Suarez, B.,
& Parra-Tabla, V. (2019). Global geographic patterns of heterospecific pollen receipt help
uncover potential ecological and evolutionary impacts across plant communities worldwide.
Scientific Reports, 9(1), 1–9. https://doi.org/10.1038/s41598-019-44626-0
Ashman, T., & Hitchens, M. S. (2000). Dissecting the causes of variation in intra‐inflorescence
allocation in a sexually polymorphic species, Fragaria virginiana (Rosaceae). American Journal
of Botany, 87(2), 197–204. https://doi.org/10.2307/2656906
Ashman, T.-L., Knight, T. M., Steets, J. A., Amarasekare, P., Burd, M., Campbell, D. R., Dudash, M.
R., Johnston, M. O., Mazer, S. J., Mitchell, R. J., Morgan, M. T., & Wilson, W. G. (2004).
POLLEN LIMITATION OF PLANT REPRODUCTION: ECOLOGICAL AND
EVOLUTIONARY CAUSES AND CONSEQUENCES. Ecology, 85(9), 2408–2421.
https://doi.org/10.1890/03-8024
Bennett, Joanne. M., Steets, Janette. A., Burns, Jean. H., Durka, W., Vamosi, Jana. C., Arceo-Gómez,
G., Burd, M., Burkle, Laura. A., Ellis, Allan. G., Freitas, L., Li, J., Rodger, James. G., Wolowski,
M., Xia, J., Ashman, T.-L., & Knight, Tiffany. M. (2018). Data from: GloPL, a global data base
on pollen limitation of plant reproduction (Version 1, p. 2274543 bytes) [Data set]. Dryad.
https://doi.org/10.5061/DRYAD.DT437
Betts, M. G., Hadley, A. S., & Kormann, U. (2019). The landscape ecology of pollination. Landscape
Ecology, 34(5), 961–966. https://doi.org/10.1007/s10980-019-00845-4
50
Burd, M. (1994). Bateman’s Principle and Plant Reproduction: The Role of Pollen Limitation in Fruit
and Seed Set. Botanical Review, 60(1), 83–139.
Burd, M. (2008). The Haig‐Westoby Model Revisited. The American Naturalist, 171(3), 400–404.
https://doi.org/10.1086/527499
Bowcutt, F.,& Hamman, S. (2016). Vascular Plants of the South Sound Prairies. The Evergreen State
College.
Cariveau, D. P., Bruninga-Socolar, B., & Pardee, G. L. (2020). A review of the challenges and
opportunities for restoring animal-mediated pollination of native plants. Emerging Topics in Life
Sciences, 4(1), 99–109. https://doi.org/10.1042/ETLS20190073
Casper, B. B., & Niesenbaum, R. A. (1993). Pollen versus resource limitation of seed production: a
reconsideration. Current Science, 210-214.
Charlesworth, D., Morgan, M. T., & Charlesworth, B. (1990). INBREEDING DEPRESSION,
GENETIC LOAD, AND THE EVOLUTION OF OUTCROSSING RATES IN A
MULTILOCUS SYSTEM WITH NO LINKAGE. Evolution, 44(6), 1469–1489.
https://doi.org/10.1111/j.1558-5646.1990.tb03839.x
Cheung, A. Y. (1996). The pollen tupe growth pathway: its molecular and biochemical contributions
and responses to pollination. Sexual Plant Reproduction, 9(6), 330-336.
Cohen, D., & Dukas, R. (1990). The Optimal Number of Female Flowers and the Fruits-to-Flowers
Ratio in Plants Under Pollination and Resources Limitation. The American Naturalist, 135(2),
218–241. https://doi.org/10.1086/285040
Cruzan, M. B., & Barrett, S. C. (1996). Postpollination mechanisms influencing mating patterns and
fecundity: an example from Eichhornia paniculata. The American Naturalist, 147(4), 576-598.
51
Diggle, P. K. (1995). Architectural effects and the interpretation of patterns of fruit and seed
development. Annual Review of Ecology and Systematics, 26(1), 531-552.
Dunwiddie, P. W., Alverson, E. R., Martin, R. A., & Gilbert, R. (2014). Annual Species in Native
Prairies of South Puget Sound, Washington. Northwest Science, 88(2), 94–105.
https://doi.org/10.3955/046.088.0205
Ehrlen, J. (1991). Why do plants produce surplus flowers? A reserve-ovary model. The American
Naturalist, 138(4), 918-933
Fenster, C. B., Armbruster, W. S., Wilson, P., Dudash, M. R., & Thomson, J. D. (2004). Pollination
Syndromes and Floral Specialization. Annual Review of Ecology, Evolution, and Systematics,
35(1), 375–403. https://doi.org/10.1146/annurev.ecolsys.34.011802.132347
Fernández, J. D., Bosch, J., Nieto-Ariza, B., & Gómez, J. M. (2012). Pollen limitation in a narrow
endemic plant: Geographical variation and driving factors. Oecologia, 170(2), 421–431.
https://doi.org/10.1007/s00442-012-2312-1
Ferrero‐Serrano, Á., Hild, A. L., & Mealor, B. A. (2011). Can invasive species enhance competitive
ability and restoration potential in native grass populations?. Restoration Ecology, 19(4), 545551.
Flory, S. L., & Clay, K. (2009). Invasive plant removal method determines native plant community
responses. Journal of Applied Ecology, 46(2), 434–442. https://doi.org/10.1111/j.13652664.2009.01610.x
França-Neto, J. de B., & Krzyzanowski, F. C. (2019). Tetrazolium: An important test for
physiological seed quality evaluation. Journal of Seed Science, 41(3), 359–366.
https://doi.org/10.1590/2317-1545v41n3223104
52
Galen, C., & Stanton, M. L. (1991). Consequences of Emergence Phenology for Reproductive
Success in Ranunculus adoneus (Ranunculaceae). American Journal of Botany, 78(7), 978–988.
https://doi.org/10.2307/2445177
Gornish, E. S., & Ambrozio dos Santos, P. (2016). Invasive species cover, soil type, and grazing
interact to predict long‐term grassland restoration success. Restoration Ecology, 24(2), 222-229.
Haig, D., & Westoby, M. (1988). On Limits to Seed Production. The American Naturalist, 131(5),
757–759. https://doi.org/10.1086/284817
Hagen, M., Kissling, W. D., Rasmussen, C., De Aguiar, M. A. M., Brown, L. E., Carstensen, D. W.,
Alves-Dos-Santos, I., Dupont, Y. L., Edwards, F. K., Genini, J., Guimarães, P. R., Jenkins, G.
B., Jordano, P., Kaiser-Bunbury, C. N., Ledger, M. E., Maia, K. P., Marquitti, F. M. D.,
Mclaughlin, Ó., Morellato, L. P. C., … Olesen, J. M. (2012). Biodiversity, Species Interactions
and Ecological Networks in a Fragmented World. In Advances in Ecological Research (Vol. 46,
pp. 89–210). Elsevier. https://doi.org/10.1016/B978-0-12-396992-7.00002-2
Hamman, S. T., Dunwiddie, P. W., Nuckols, J. L., & McKinley, M. (2011). Fire as a Restoration Tool
in Pacific Northwest Prairies and Oak Woodlands: Challenges, Successes, and Future Directions.
Northwest Science, 85(2), 317–328. https://doi.org/10.3955/046.085.0218
Harder, L. D., & Barrett, S. C. H. (1996). Pollen Dispersal and Mating Patterns in Animal-Pollinated
Plants. In D. G. Lloyd & S. C. H. Barrett (Eds.), Floral Biology: Studies on Floral Evolution in
Animal-Pollinated Plants (pp. 140–190). Springer US. https://doi.org/10.1007/978-1-4613-11652_6
Herrera, J. (1991). ALLOCATION OF REPRODUCTIVE RESOURCES WITHIN AND AMONG
INFLORESCENCES OF LAVANDULA STOECHAS (LAMIACEAE). American Journal of
Botany, 78(6), 789–794. https://doi.org/10.1002/j.1537-2197.1991.tb14480.x
53
Kaiser-Bunbury, C. N., Mougal, J., Whittington, A. E., Valentin, T., Gabriel, R., Olesen, J. M., &
Blüthgen, N. (2017). Ecosystem restoration strengthens pollination network resilience and
function. Nature, 542(7640), 223–227. https://doi.org/10.1038/nature21071
Kanchan, S., & Chandra, J. (1980). POLLEN ALLELOPATHY—A NEW PHENOMENON. New
Phytologist, 84(4), 739–746. https://doi.org/10.1111/j.1469-8137.1980.tb04786.x
Knight, T. (2004). Floral density, pollen limitation, and reproductive success in Trillium
grandiflorum. Oecologia, 137, 557–563. https://doi.org/10.1007/s00442-003-1371-8
Knight, T. M., Steets, J. A., Vamosi, J. C., Mazer, S. J., Burd, M., Campbell, D. R., Dudash, M. R.,
Johnston, M. O., Mitchell, R. J., & Ashman, T.-L. (2005). Pollen Limitation of Plant
Reproduction: Pattern and Process. Annual Review of Ecology, Evolution, and Systematics,
36(1), 467–497. https://doi.org/10.1146/annurev.ecolsys.36.102403.115320
Knight, T., Steets, J. A., & Ashman, T.-L. (2006). A QUANTITATIVE SYNTHESIS OF POLLEN
SUPPLEMENTATION EXPERIMENTS HIGHLIGHTS THE CONTRIBUTION OF
RESOURCE REALLOCATION TO ESTIMATES OF POLLEN LIMITATION. American
Journal of Botany, 93, 271–277. https://doi.org/10.3732/ajb.93.2.271
Krukeberg, A. 1991. The Natural History of Puget Sound Country. University of Washington Press,
Seattle, Washington
Lloyd, D. G. (1992). Self- and Cross-Fertilization in Plants. II. The Selection of Self- Fertilization.
International Journal of Plant Sciences, 153(3, Part 1), 370–380. https://doi.org/10.1086/297041
Lopezaraiza–Mikel, M. E., Hayes, R. B., Whalley, M. R., & Memmott, J. (2007). The impact of an
alien plant on a native plant–pollinator network: An experimental approach. Ecology Letters,
10(7), 539–550. https://doi.org/10.1111/j.1461-0248.2007.01055.x
54
Morales, C. L., & Traveset, A. (2009). A meta-analysis of impacts of alien vs. native plants on
pollinator visitation and reproductive success of co-flowering native plants. Ecology Letters,
12(7), 716–728. https://doi.org/10.1111/j.1461-0248.2009.01319.x
National Oceanic and Atmospheric Administration (NOAA). National Weather Service.
https://www.weather.gov/ accessed 11/2/2021.
Newman, B. J., Ladd, P., Brundrett, M., & Dixon, K. W. (2013). Effects of habitat fragmentation on
plant reproductive success and population viability at the landscape and habitat scale. Biological
Conservation, 159, 16–23. https://doi.org/10.1016/j.biocon.2012.10.009
Noland, S., Carver, L. (2011). Prairie Landowner Guide for Western Washington. The Nature
Conservancy, Seattle, WA.
Orrock, J. L., & Witter, M. S. (2010). Multiple drivers of apparent competition reduce reestablishment of a native plant in invaded habitats. Oikos, 119(1), 101–108.
https://doi.org/10.1111/j.1600-0706.2009.17831.x
Parker, G. A., & Smith, J. M. (1990). Optimality theory in evolutionary biology. Nature, 348(6296),
27–33. https://doi.org/10.1038/348027a0
Ramula, S., & Pihlaja, K. (2012). Plant communities and the reproductive success of native plants
after the invasion of an ornamental herb. Biological Invasions, 14(10), 2079–2090.
https://doi.org/10.1007/s10530-012-0215-z
Rosenheim, J. A., Alon, U., & Shinar, G. (2010). Evolutionary Balancing of Fitness‐Limiting Factors.
The American Naturalist, 175(6), 662–674. https://doi.org/10.1086/652468
Sabatino, M., Rovere, A., & Meli, P. (2021). Restoring pollination is not only about pollinators:
Combining ecological and practical information to identify priority plant species for restoration
55
of the Pampa grasslands of Argentina. Journal for Nature Conservation, 61, 126002.
https://doi.org/10.1016/j.jnc.2021.126002
Snow, A. A. (1994). Postpollination selection and male fitness in plants. The American
Naturalist, 144, S69-S83.
Storm, L. (2002). Patterns and Processes of Indigenous Burning How to Read Landscape Signatures
of Past Human Practices. International Society of Ethnbiology, University of Georgia Press pp.
496-508 ref. 49.
Trowbridge, C. C., Stanley, A., Kaye, T. N., Dunwiddie, P. W., & Williams, J. L. (2017). Long-term
effects of prairie restoration on plant community structure and native population dynamics.
Restoration Ecology, 25(4), 559–568. https://doi.org/10.1111/rec.12468
Vico, G., Manzoni, S., Nkurunziza, L., Murphy, K., & Weih, M. (2016). Trade‐offs between seed
output and life span–a quantitative comparison of traits between annual and perennial congeneric
species. New Phytologist, 209(1), 104-114.
Waters, S. M., Fisher, S. E., & Hille Ris Lambers, J. (2014). Neighborhood-contingent indirect
interactions between native and exotic plants: Multiple shared pollinators mediate reproductive
success during invasions. Oikos, 123(4), 433–440. https://doi.org/10.1111/j.16000706.2013.00643.x
Wesselingh, R. A. (2007). Pollen limitation meets resource allocation: Towards a comprehensive
methodology. New Phytologist, 174(!), 26-37.
https://nph.onlinelibrary.wiley.com/doi/full/10.1111/j.1469-8137.2007.01997.x
Zimmerman, M., & Pyke, G. H. (1988). Reproduction in Polemonium: Assessing the Factors Limiting
Seed Set. The American Naturalist, 131(5), 723–738.
56
Appendices
Appendix A
Table 1
ANOVA Table
Individual Seed Weight by Site
Plant Species
Model
DF
SS
MS
F
P
Lupinus lepidus
ANOVA
3
0.003
<0.001
1.546 0.215
Hypochaeris
radicata
ANOVA
5
<0.001
<0.001
3.636 0.004
Leucanthemum
vulgare
ANOVA
4
<0.001
<0.001
3.132 0.018
Table 2
Plectritis congesta
Individual Seed Weight by Site
P-values from General Linear Mixed Model using Gamma Loglink Model
Estimated Marginal Means Pairwise Comparison Table
Site
CV
DS
GH
JP
TQ
CV
DS
GH
JP
TQ
WH
0.004
0.001
0.016
<0.001
<0.001
0.934
1.00
0.092
0.090
0.928
0.708
0.07
0.095
0.080
1.00
WH
57
Table 3
Eriophyllum lanatum
Fertilized Seed Weight
P-values from Pairwise Comparison of Means
Site
CV
GH
JP
CV
GH
JP
TQ
WH
1.00
1.00
0.977
0.890
1.00
0.988
0.881
0.988
0.890
TQ
0.538
WH
Table 4
Hypochaeris radicata
Proportion of Stain Viable
P-values from Wilcoxon Rank Sum Test Pairwise Comaparison Table
58
Table 6
Lupinus lepidus
Proportion of Fertilized Seed
P-values from Wilcoxon Rank Sum Test Pairwise Comparison Table
Site
CV
GH
TQ
GH
<0.001
-
-
TQ
0.23
<0.001
-
WH
0.23
<0.001
0.08
59
Table 7
Hypochaeris radicata
Proportion of Stain Viable
P-values form Wilcoxon Rank Sum Test Pairwise Comparison Table
60
Table 9
Plectritis congesta
Proportion Fertilized Seed
P-values from Estimated Marginal Means Pairwise Comparison Table
Table 10
Eriophyllum lanatum
Proportion Fertilized Seeds
P-values from the Wilcoxon Sum Tank Test Pairwirse Comparison Table
61
Appendix B
Self-compatibility data for south
Puget Sound prairie plant species.
Mean seed set for open and closed
pollination treatments
(unpublished data, Susan Waters).
62
