a. Essa?
.nﬁnﬁhwﬁh my}. a
3...!
:23:

 

3.0.
a 5:

 

VI 1‘. I.
l! oruriilil... .
;I!

to ‘
.3. 1x532.
131......3‘34.
.x I. I .. l

.2.
1111..
I...

T . 1 .1 i. 3...;
.l 53:15.:

0“;

 

3

7'3

910%
o

This is to certify that the
dissertation entitled

RESTORATION OF PRAIRIE FEN PLANTS, INSECTS, AND
ABIOTIC PROCESSES

presented by

ANNA KATHERINE FIEDLER

has been accepted towards fulfillment
of the requirements for the

Doctoral [\ degree in Entomologl

 

 

Oamajor Pr/ofessor’s Signature

8/23/ 20/0

Date

MSU is an Afﬁnnafive Action/Equal Opportunity Employer

 

‘ LIBRARY
Michigan State
University

 

 

 

I

—v—o-I-I-o-I-I-o-c—a-o—-—-~a—,_ . _

-.-‘-I-I-.-.-I-.-C-l-

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE, DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 ICIProj/AMPres/ClRC/Dateoue.indd

 

RESTORATION OF PRAIRIE FEN PLANTS, INSECTS, AND ABIOTIC
PROCESSES

By

Anna Katherine Fiedler

A DISSERTATION
Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Entomology

2010

ABSTRACT

RESTORATION OF PRAIRIE FEN PLANTS, INSECTS, AND ABIOTIC
PROCESSES

By

Anna Katherine Fiedler

There are growing concerns about biodiversity decline and species extinctions due
to habitat fragmentation, invasive species, and climate change. Habitat restoration is
increasingly used to reverse degradation of rare ecosystems and maintain biological
diversity, species interactions, and ecosystem function. The effects of restoration
activities on these properties, however, are often not measured. The Midwestern US. has
a relatively high density of globally rare prairie fen wetlands which support high plant
diversity and support a number of rare and endangered plants, insects, and vertebrates.
The most common exotic invasive species in Michigan prairie fens is F rangula alnus
(glossy buckthorn). Frangula alnus invasion is known to alter soil and plant community
conditions in fens but changes following restoration have never been quantiﬁed. The goal
of my research was to assess changes in abiotic conditions, species diversity, and species
interactions after restoration, and to assess whether they are on a trajectory toward those
in uninvaded fen. Toward that end, I quantiﬁed differences in invaded and uninvaded
prairie fen before restoration began, as well as tracking shifts in resource availability,
plant community, pollinator diversity, plant-pollinator networks, and pollinator ﬁmction
in the two growing seasons following restoration of prairie fen. I found a range of
responses to restoration across resources and plant community metrics. Light availability,

herbaceous plant cover, and relative graminoid abundance increased in restored plots,

while soil pH, surface spatial heterogeneity and ﬂoristic quality index were initially
greater in uninvaded areas versus invaded fen and remained so in the ﬁrst two years after
restoration. These results indicate that some soil and plant community factors change
rapidly while others may remain altered for years.

A diverse community of mobile generalist pollinators rapidly re-colonized
restored areas. Bee and butterﬂy communities were nearly absent in invaded prairie fen
and responded rapidly to restoration in abundance, diversity, and composition. However,
plant species diversity and composition in restored plots remained signiﬁcantly different
than in reference plots, suggesting the plant community may take longer to recover.

I examined plant—pollinator networks and the ecosystem function of pollination
using quantitative food webs describing plant-pollinator interactions, and by assessing
pollinator function using sentinel Asclepias incarnata plants. Plant-pollinator networks in
invaded plots were depauperate, with signiﬁcantly lower plant and pollinator species
richness than restored or reference plots. Network connectance, compartment diversity,
generality, and vulnerability did not differ between restored and reference plots, with
marginally higher interaction evenness in reference than restored plots. Pollinator
function was restored in cleared areas in the ﬁrst growing season following restoration,
with no signiﬁcant differences in pollinator abundance or diversity between cleared and
reference areas. This work indicates that rapid restoration of plant community structure,
pollinator diversity, and function are possible in open-structured ecosystems. Combined
consideration of species diversity, ecosystem function, and species interactions provides a

process-based ecosystem analysis that can inform ecological theory and restoration.

ACKNOWLEDGEMENTS

I am deeply grateful to Dr. Doug Landis for his support through my change in
research focus aﬁer my Masters degree, his commitment to ﬁnding research funding for
my project, and for providing the space for me to grow into a colleague over the last four
years. Conversations with Dr. Rufus Isaacs, Dr. Patrick Doran, and Dr. Carolyn
Malmstrom have helped me to focus and develop my research ideas, as well as provided
career direction.

Dave Cuthrell and Ryan O’Conner shared their contagious enthusiasm and love
for prairie fens with me. Greg Kowalewski, Steve Woods, Rodolfo Zuniga-Villegas and
their crews provided indispensable help with plot establishment and ongoing restoration
work. I am grateful to Chris Sebolt for putting himself in very real danger of contracting
poison sumac each of the many times he helped with ﬁeld work and restoration activities.

Jane Memmott and her lab group, especially Jo Brooks, Rachel Gibson, Darren
Evans, and Mariano Devoto, made learning about the intricacies of species networks, as
well as the meaning of “toodlepip”, a joy.

I am extremely grateful for the enthusiastic help that Mitch Lettow provided on
this project, and the consistent and humorous assistance that Colin Philippo provided. I
simply could not have done this project without their help.

Rob Ahern provided much-needed support and perspective, as well as
encouragement to try the parts of this research that I have found most rewarding. I am
also grateful to Ben Werling, Mary Gardiner, and Jeff Evans for their input,
encouragement, and high ﬁves. Emily Scott and the chocolate chip cookies at Gone

Wired provided an impetus to work on weekends to ﬁnish this work.

iv

I am indebted to the community of people I have met and become ﬁiends with
during my time in Lansing. You have been the source of creativity that has kept me
interested in, and taking an inspired approach to, this research. Anna Porubcansky has
provided constant reminders to take things as they come, Laura Wies has been a safe
haven. I am grateful to my family for their continued support of me and interest in this
work.

Most of all, I appreciate the support that Jake McCarthy provided, in his near
super-human ability to tolerate me after long work days, his interest in my R-values, and

his inspiration for ways to clearly communicate my ﬁndings to others.

TABLE OF CONTENTS

 

 

 

 

 

 

LIST OF TABLES .......................................................................................................... IX
LIST OF FIGURES ......... -- XI
CHAPTER 1. LITERATURE REVIEW-- - 1
Introduction -- - _ - - -- - _ _ -- - 2
Glacial History ................................................................................................................. 3
Prairie Fen. - - ........ -- .......... -- . - - -4
Unique Wetland Type ...................................................................................................... 4
Plant Community Gradient .............................................................................................. 7
Insects in Prairie Fen ..................................................................................................... 10
Historical Disturbance Patterns ..................................................................................... 12
Threats to prairie fen systems and their management ................................................... l3
Conversion ................................................................................................................. l4
Fragmentation ............................................................................................................ 14
Hydrological changes ................................................................................................ 15
Changes in disturbance regime .................................................................................. 15
Pollution, nutrient inputs, and invasive species ......................................................... 18
Glossy Buckthom _ - _ -- - - - - - 20
Life history and arrival in Michigan .............................................................................. 20
Impacts on habitat .......................................................................................................... 21
Management techniques and implications ..................................................................... 22
Physical control techniques ....................................................................................... 23
Chemical control techniques ..................................................................................... 24
Long-term management ............................................................................................. 25
Conservation of biodiversity and ecosystem function - 27
Functions of wetlands .................................................................................................... 27
Insect conservation and function ................................................................................... 28
Pollination: a key process for ecosystem persistence .................................................... 29
Literature Cited 33

 

CHAPTER 2. BIOTIC AND ABIOTIC CONDITIONS IN MICHIGAN PRAIRIE

 

 

 

 

FEN INVADED BY GLOSSY BUCKTHORN (FRANGULA ALNUS) ..................... 40
Abstract 41
Introduction - - - - . 42
Glossy buckthom ........................................................................................................... 43
Effects on habitat ........................................................................................................... 43
Methods -- _ 45
Experimental design ...................................................................................................... 45
Sampling ........................................................................................................................ 46
Statistical analysis .......................................................................................................... 47
Results -- ........... - - -- 48
Differences between invaded and uninvaded plots ....................................................... 48
Differences with distance from water’s edge ................................................................ 49

vi

Discussion - -- -- - - - - ........ 50

 

 

Nutrient Availability ...................................................................................................... 52
Plant community ............................................................................................................ 53
Patterns in Frangula alnus invasion ............................................................................... 54
Acknowledgements - -- - - -- - - - 56
Literature cited- ---_ - - -- - - - 57

 

CHAPTER 3. DIFFERENTIAL SHIFTS IN RESOURCE AVAILABILITY AND
PLANT COMMUNITY STRUCTURE FOLLOWING PRAIRIE F EN

 

 

 

 

 

 

 

RESTORATION - - - - -- -68
Abstract - - -- -- - - - - - - - - - ......... - 69
Introduction - - -- - -- - -- - - -- 70
Methods -- - - -- -- - - 73
Experimental design ...................................................................................................... 73
Sampling ........................................................................................................................ 74
Statistical analysis .......................................................................................................... 75
Results - - - -- 77
Resource availability ..................................................................................................... 77
Plant community characteristics .................................................................................... 78
Discussion - - - -- - - ........ 79
Resource Availability .................................................................................................... 80
Plant Community Structure and Diversity ..................................................................... 82
Long-term Restoration Success ..................................................................................... 84
Acknowledgements ....... - - ..... - 86
Literature cited - ............... . - -- ........ 88

 

CHAPTER 4.1NVASIVE SPECIES REMOVAL LEADS TO RAPID RECOVERY

 

 

 

 

 

 

 

 

OF POLLINATOR COMMUNITIES IN PRAIRIE FEN ...... - -- - ----100
Abstract ..... - - - - . .......... . --101
Introduction -- -- - ........... - - 102
Methods - ...... - -- - - 105

Experimental design ....................................... ' ............................................................. 105

Sampling ...................................................................................................................... 105

Statistical analysis ........................................................................................................ 107
Results - - - -- -- - -- - 108
Discussion - - - - - -- - ....... - 113
Acknowledgements - - -- - - ....... - - - 118
Literature cited ...... -- - - - - 119

CHAPTER 5. RAPID RESTORATION OF PLANT-POLLINATOR NETWORKS

 

 

 

 

AND FUNCTION FOLLOWING HABITAT RESTORATION ......... “138
Abstract -- -- - - - ....... -- ..... 139
Introduction “ - -- - - - -- - - - 140
Methods - - - - -143
Study system and experimental design ........................................................................ 143
Sampling ...................................................................................................................... 144

vii

Pollinator networks .................................................................................................. 144

 

 

 

Pollinator function ................................................................................................... 145
Statistical analysis ........................................................................................................ 147
Pollinator networks .................................................................................................. 147
Pollinator function ................................................................................................... 149
Results - - - - - - . -- - - - 149
Plant—pollinator networks ............................................................................................ 149
Pollinator function ....................................................................................................... 151
Discussion -------- - -- ---------------------- - - - - - 152
Acknowledgements -- - - - . - - - - 156
Literature Cited -- ---------- - - 158

 

CHAPTER 6. CONCULSION. IMPROVING THE PRACTICE OF

 

 

 

RESTORATION THROUGH RESTORATION ECOLOGY"..- - - 172
Literature Cited - - -- - -- - - - -- - 178
Appendix A. Soil nutrient levels from 1-10cm cm depth soil cores. ......................... 180
Appendix B. Plant voucher data ------------------------- ----1801
Appendix C. Insect voucher data ............................................................................... 1806

viii

LIST OF TABLES

Table l. The 25 federally endangered (E), candidate (T), and special concern (SC)
insect species in Michigan prairie fen communities (MNF I 2007) ....................... 10

Table 2. Hypothesized and actual differences in abiotic and biotic factors between
F rangula alnus invaded and uninvaded reference areas of prairie fen. A result in
parentheses indicates a marginally signiﬁcant result ....................................... 60

Table 3. Plant species identiﬁed at research site and the percent of quadrats each
species occupied, 112 total ..................................................................... 61

Table 4. Differences in abiotic and biotic factors between upland and water’s edge
transects in prairie fen. P-values are bolded where less than 0.10 ........................ 64

Table 5. Differences in resource availability between invaded, removal, and

uninvaded treatments pre-restoration and year 1 to 2 post-restoration (September

2007, 2009, and 2009, respectively). Signiﬁcant p-values (a=0.05) are bolded.
Treatrnent*time interactions not shown if neither treatment nor time were

signiﬁcant ......................................................................................... 93

Table 6. Differences in plant community metrics between invaded, removal, and
uninvaded treatments pre-restoration and year 1 to 2 post-restoration (September
2007, 2009, and 2009, respectively). Signiﬁcant p-values (or=0.05) are bolded ....... 94

Table 7. ANOSIM based on similarity matrix by time. All resource factors and

plant factors presented in tables 1 and 2 included, plus % bare ground, soil Ca and

Mg. Signiﬁcant p-values are bolded. R ranges up to 1, a value closer to 1 indicates
greater between treatment differences .......................................................... 95

Table 8. Plant species identiﬁed at research site and the percent of quadrats each

species occupied in year 2 (2009), 177 total. Bolded values are species whose

frequency varied most by treatment, Inv=Invaded, Remo=Removal and

Ref=Reference .................................................................................... 124

Table 9. Number of bees observed using observational sampling per replicate and
treatment in years 1 and 2 post restoration. The percent of bees each genus

composed of the total number observed is in the ﬁrst column for years 1 and 2.

Bolded values are the most common species whose abundance varied most by

treatment. Dashes indicate no insects of that species were seen in a given year.

Values are by replicate and treatment, averaged over sample dates ...................... 128

ix

Table 10. Number of bees collected using pollinator bowl sampling per replicate
and treatment, Inv=Invaded, Remo=Removal and Ref=Reference, in years 1 and
2 post restoration. The percent of bees each species composed of the total number
trapped is in the ﬁrst column for years 1 and 2. Bolded values are the most
common species whose abundance varied most by treatment. Dashes indicate no
insects of that species were seen in a given year. Values are by replicate and
treatment, averaged over sample dates. . . .. .........................................

Table 11. Number and identity of butterﬂies seen during 5 minute observational
sampling in years 1 and 2 post restoration, % obs=percent of total observed per
year, Inv=Invaded, Remo=Removal and Ref=Reference. Bolded values are the
most abundant species whose abundance varied by treatment. Dashes indicate no
insects of that species were seen in a given year. Values are by replicate and
treatment, averaged over sample dates.. ..................................................................

Table 12. Plant species in quantitative ﬂower visitation webs included in Figure
13 ................................................................................................

Table 13. Insect species in quantitative ﬂower visitation webs included in Figure
13 ................................................................................................

Table 14. Characterization of plant-pollinator networks in invaded, restored, and

uninvaded reference treatments. Mean
iSEM ...........................................................................................

Table 15. Differences in plant—pollinator network structure metrics between
restored and uninvaded reference plots, based on an analysis of covariance with
the number of plant and pollinator species in each network accounted for, with
DF 3,9 ..........................................................................................

Table 16. Differences in pollinator function and insect visitation at A. incarnata
plants between invaded, restored, and uninvaded treatments. Insect abundances
are scaled by the number of A. incamata ﬂower clusters open during sampling.
Signiﬁcant p-values (0L=0.05) are

bolded ..........................................................................................

129

131

162

163

164

165

166

LIST OF FIGURES

Figure 1. Characterization of wetlands by water source and water level. Fen is
deﬁned by groundwater, but poor fen may have a combination of surface and
groundwater inﬂow .............................................................................. 5

Figure 2. Figure 2. Layout of research plots at MacCready Reserve, Clarklake, MI.

A: Ten 25 x 25m replicates of buckthom invaded plots are shown as outlined

rectangles and two uninvaded areas as white ﬁlled rectangles. B: Quadrats in black

lie 3m from the upland edge (upland). White quadrats lie 3m from the water’s edge
(water) and dark grey quadrats are in the center of plots (center). C: Schematic of
typical elevational differences within plots, arrows showing where elevation was
measured ......................................................................................... 65

Figure 3. Comparisons of abiotic and biotic factors between F. alnus invaded and
uninvaded plots. Error bars are +SEM, 01:005. Treatment effects from a 2-way
ANOVA with treatment and transect are shown

here .................................................................................................. 66

Figure 4. The effect of distance to the upland/water’s edge on abiotic and biotic

factors in F. alnus invaded plots. Patterns in uninvaded plots are shown for

comparison; no statistical tests were performed on these data. Error bars are +SEM.
Bars with different letters within a ﬁgure are signiﬁcantly different using

Bonferonni adjusted LS means comparisons (p<0.05). ................................... 67

Figure 5. Comparisons of soil resource availability, light and microhabitat between

F. alnus invaded, removal, and uninvaded plots. Error bars are +SEM, or=0.05.

Where treatment or treatment*time were signiﬁcantly different using LS means
separations, letters of signiﬁcance are shown ............................................... 96

Figure 6. Comparisons of plant metrics between F. alnus invaded and uninvaded

plots. Error bars are +SEM, 0t=0.05. Treatment effects from a 2-way ANOVA with
treatment and transect are shown. Additive herbaceous cover represents the sum of
individual plant species cover/m2 plot. Where treatment or treatrnent*time were
signiﬁcantly different using LS means separations, letters of signiﬁcance are

shown ............................................................................................. 97

Figure 7. Two dimensional NMDS plots based on a mean of each resource and

plant metric by treatment and replicate using the Euclidian dissimilarity index.

There were two reference plots pre-restoration; a third was added before year 1.
Pre-restoration and year 2 statistical differences by treatment are shown with

letters .............................................................................................. 99

xi

Figure 8. Comparisons of A) percent cover of forbs and B) Simpsons diversity

between F. alnus invaded, removal, and uninvaded plots in the ﬁrst and second

years following restoration. Maximum values for percent cover recorded for each
growing season are used. Error bars are +SEM, a=0.05. Treatment effects from a

2-way ANOVA with treatment and time are shown; year effects were not

signiﬁcant ........................................................................................ 132

Figure 9. Two dimensional NMDS ordinations of the plant community show no
difference between invaded and removal plots in year 1 but a shift in removal plots

away from invaded plots by year 2. The ordination is based a Bray-Curtis

dissimilarity matrix for the ﬁrst and second years following restoration. Maximum
values for percent cover recorded for each growing season are used, data were

square root transformed. Stress values were 0.08 and 0.09 in years 1 and 2 ............ 133

Figure 10. Comparisons of A) bee abundance B) bee diversity C) butterﬂy

abundance and D) butterﬂy diversity in the ﬁrst and second seasons following
restoration. Bee abundance, butterﬂy abundance, and diversity are based on the
season-long mean number of individuals observed in 5-minute observational

sampling periods in invaded, removal, on-site reference (reference) and off-site

reference (off-site). Bee diversity is based on bees collected with bowl sampling in
invaded, removal, and on-site reference plots. Treatment effects ﬁ'om 2-way

ANOVAs with treatment and time are shown. Letters in upper and lower case

within the same graph indicate that treatments differed by year or treatment*year. .. 134

Figure 11. Two dimensional NMDS ordinations of the bee community show

distinct bee Communities in invaded plots in both years 1 and 2, with bee

communities in removal and reference plots overlapping. The ordination is based a
Bray-Curtis dissimilarity matrix (using n+1), using mean values per replicate and
treatment; data were square root transformed. Stress values were 0.14 in both

years .............................................................................................. 136

Figure 12. Two dimensional NMDS ordinations of the butterﬂy community show

that the invaded and removal butterﬂy communities differ in both years 1 and 2. In

year 2, all six replicates of invaded contained 0 butterﬂies, and all points overlap.

In year 2, 1 butterﬂy was seen in invaded plots and 5 of 6 points overlap. The

ordination is based a Bray-Curtis dissimilarity matrix (using n+1), with mean

values per replicate and treatment; data were square root transformed. Stress values

were 0.04 for year 1 and 0.11 for year 2 ..................................................... 137

Figure 13. Quantitative ﬂower visitation webs for representative replicates of the

A) invaded, B) restored, and C) reference prairie fen plots. Plant species are

represented by rectangles on the bottom of the ﬁgure, insects are shown on the top

level. The width of a bar represents the relative abundance of pollinators or plants

and the frequency of the interaction is represented by the width of the line from

pollinators to plants. Flowers that were observed but no pollinators were observed

at are present in these webs, but with no line connecting them to pollinators .......... 167
x11

Figure 14. Two dimensional NMDS ordinations of A) the ﬂowering plant and B)
ﬂower—visiting insect communities in plant—pollinator interactions. Both A) and

B) indicate that there are distinct plant and pollinator communities at the

uncleared and cleared plots. A) shows no distinction between the plant

communities in the on— and off—site references, while B) reveals signiﬁcantly

different plant-visiting insect communities between on— and off-site references.

The ordination is based on a Bray—Curtis dissimilarity matrix (using n+1) with

mean values per replicate and treatment; data were square root transformed.

Stress values were 0.14 and 0.11 for ﬂowering plants and insects, respectively. . .. 169

Figure 15. Patterns of A) pollinator function and at A. incarnata ﬂower—visiting

B) insect richness C) insect abundance and D) bee abundance in the ﬁrst and

second seasons following restoration. Insect richness and abundances are based

on means by treatment and replicate over three sampling dates. Error bars are

+SEM, or=0.05. Treatment effects from a 2—way ANOVA with treatment and

time are shown where signiﬁcant. Letters in upper and lower case within the

same graph indicate that treatments differed by year or treatrnent*year .............. 170

xiii

Chapter 1

LITERATURE REVIEW

Introduction

The retreat of the most recent glaciation from what is now the Northern U.S.
approximately 12,000 years ago created a set of abiotic conditions that paved the way for
the development of unique communities (Albert 1995). One of these is prairie fen, a
unique groundwater fed wetland habitat type in the Midwestern U.S. Prairie fen is
exceptional in that it supports very high biodiversity within communities that compose
less than 0.1% of the glaciated landscape in the Midwestern U.S. (N ekola 1994). Prairie
fen communities in Michigan are home to a suite of rare and endangered plants and
animals, including the federally listed Mitchell’s Satyr butterﬂy (Neonympha mitchellii
mitchellii). The plant community in prairie fen contains many rare species and is
increasingly imperiled.

Prairie fen is a ﬁre—dependent community that is typically embedded in a matrix
of prairie and savanna habitats. Together, the upland and wetland elements create a
landscape of high biodiversity. Since European settlement in the Midwestern U.S., prairie
fen and the surrounding matrix have undergone a variety of changes. These changes have
resulted in a number of threats to prairie fen, including conversion, ﬁagmentation,
hydrological changes, altered disturbance regimes, introduction of non—native species,
pollution, and increased nutrient inputs. The result of these combined threats is that a
group of plant and animal species endemic to prairie fen are are risk of extinction.
Perhaps more importantly, ecosystem functions such as water and nutrient cycling that
were once supported by this habitat mosaic may be disrupted. Insects are key to a variety
of ecosystem processes: decomposition, insect predation, herbivory, and pollination.
Pollinator species loss may represent a direct threat to plant persistence in fen due to

decreased seed set, and therefore persistence of already rare plant species. To prevent

2

further species, habitat, and ecosystem function loss and to protect watershed health, an
understanding of the importance and restoration of pollinator function in prairie fen

communities is vital.

Glacial History

During the Wisconsonian glaciation in the Pleistocene, lobes of glacial ice ﬂowed
down from the North, eventually covering all of Michigan and much of Wisconsin and
Minnesota (Albert 1995). The advancing glaciers passed over bedrock composed of
Paleozoic limestone, dolomite, sandstone, and shale, scouring uplands and carrying soil
and rock in and on the ice (Albert 1995). When the glaciers melted about 13,000-16,000
years Before Present (BP), the resulting glacial till was redeposited to form the current
landscape. Where glacial lobes meet along their edges, unique processes occur resulting
in the formation of interlobate landscapes. These areas are typically of higher relief than
the surrounding landscape and contain many ice-contact features, including end-moraine
ridges composed of elongate mounds of glacial till at the edge of glaciers, eskers formed
of glacial outwash carried by melting water under and between glaciers, and kettlehole
lake depressions left by melting ice chunks which often have surrounding wetlands
(Albert 1995). The moraine and esker soils are particularly well-drained, but poorly

drained soils are typical of the kettlehole depressions (Albert 1995).

Prairie Fen

Unique Wetland Type

Prairie fen habitats are one of the rarest wetland types in the US, and occur most
commonly in the glaciated Midwest (Spieles et al. 1999, Amon et a1. 2002). The
difference between current day fen status and that before European settlement is not well
known because fen communities were not speciﬁcally delineated in early land surveys
(Bedford and Godwin 2003). However, prairie fens are undoubtedly much more rare
today than they were historically; Pearson and Loeschke (1992) compared the amount of
prairie fen in Iowa and determined that nearly 40% had been destroyed by cultivation or
drainage. In Michigan, there are currently approximately 130 fens, distributed primarily
in the southeastern and southwestern interlobate regions (Spieles et al. 1999). Prairie fens
are typically small, those in New York are primarily smaller than 5 hectares (Bedford and
Godwin 2003), but contain high plant species diversity. Amon et al. (2002) identiﬁed
1169 vascular plant species in Midwestern fens, many of them rare. This combination of
small size and high species richness makes fens biodiversity hotspots. In Iowa, New
York, and New Jersey, fens compose less than 0.1% of total land, yet contain 7-18% of
the rare taxa that occur in those states (N ekola 1994).

The term fen is used to describe a variety of wetland habitats throughout North
America and Europe (Amon et al. 2002, Bedford and Godwin 2003). Although they
occur across a range of soils, climate, and water availability, fens are distinct from other
wetland communities in that all are groundwater fed (Bedford and Godwin 2003) (Fig.
1). Because the primary water input is groundwater, water level variation within

4

undisturbed fens is quite low (Amon et al. 2002, Bedford and Godwin 2003). The
resulting system is one where plants grow in constantly saturated soils but are rarely
inundated with standing water. The size of the catchment area surrounding a fen is one
determinant of how much this water level changes; if the recharge area is small and local,
water levels in a fen will drop during extended dry periods (Amon et al. 2002). Water
ﬂows through the prairie fen in either diffuse surface ﬂow or in more concentrated stream
ﬂow (Spieles et al. 1999). The amount of water that ﬂows from fens into surface lakes
and rivers varies greatly, and depends on the amount of water input into the system
(Amon et al. 2002).

Fen habitats have a high degree of mineral inﬂow through groundwater, and the
characteristics of the surrounding upland determine the amount and type of nutrients in

prairie fens. An alkaline fen results in locations where the surrounding substrate is

water source

rainwater

stable water . bog
level,

inundated

yes —+ marsh
next to open
surface water I water body <
high water level "0 —* swamp

 

ﬂuctuation
gneiss, granite, iron- poor fen (pH 4.5-5.5).
groundwater TICh bedrock, /' wl sphagnum moss
stable water levels surface+groundwater
no long-term
inundation alkaline J 1 f 1
bedrock fen (pH >5.5), calcareous fen, marl fen,
Peatland with rich fen, (pH 6-7.5), extreme rich fen
graminoids, brown CaC03 deposits. (pH >7.5), marl at
mosses, shrubs. brown mosses, char- surface, vegetation
"668- acteristic dicot forbs. sparse and stunted.

Figure 1. Characterization of wetlands by water source and water level. Fen is deﬁned by
groundwater, but poor fen may have a combination of surface and groundwater inﬂow.

primarily composed of limestone or dolomite bedrock or calcareous rich glacial deposits

(Bedford and Godwin 2003). Water ﬂowing through all of these substrates becomes rich

in calcium and magnesium. As groundwater ﬂows through surrounding uplands, nutrient
availability often decreases because formation of C8C03, CaPO4 and the presence of iron
promote phosphorous adsorption or precipitation into forms of phosphorous that are
unavailable to plants (Bedford and Godwin 2003, Grootjans et a1. 2006). These
communities are often referred to as rich fens due to high species diversity that occurs on
soils depauperate in nitrogen and phosphorous. A drop in the partial pressure of C02
causes carbonates to form when groundwater rich in calcium bicarbonate reaches the
surface (Bedford and Godwin 2003). This causes calcium to precipitate out into CaCO3
deposits, creating a marl substrate that occurs only in alkaline fens. Most commonly this
deposit is clay-like marl, however, on occasion CaC03 precipitates into hard deposits
known as tufa (Amon et al. 2002).

In Europe, prairie fens are considered one type of peatland, with peat substrate
and occasional mineral soils from CaCO3 deposits (Rydin et a1. 2006). In the US,
however, fens can develop on carbonate or peat substrates, or a combination of the two
(Amon et al. 2002). Peat in fens is composed of organic matter provided by brown moss,
sedges, and sphagnum moss which break down slowly in the anoxic conditions of the
water saturated substrate in prairie fen (Amon et al. 2002, Grootjans et al. 2006). If the
water inﬂow to a fen is small, there may not be sufﬁcient water saturation to lead to peat
formation, leading to a fen with mineral substrate only (Amon et al. 2002). Prairie fens,

therefore, contain a mosaic with mineral deposits, as well as peat histosols in a

progression of decompositional states, including ﬁbric, hemic, and sapric deposits (Buol
et al. 1997). F ibric peat is primarily undecomposed brown and sphagnum moss with a tan
to light brown color. When water levels drop, sapric peat formed from decomposed
brown moss, sedges, and sphagnum is likely to form (Buol et al. 1997, Amon et al. 2002).
Sapric peat based soils from 0.5 to >12 m deep are black and high in organic matter, and
are the typical prairie fen substrate (Spieles et al. 1999, Amon et a1. 2002). Miner and
Ketterling (2003) examined the process of marl and peat deposition in a prairie fen in
Illinois. They found that basal peat deposits bordering marl ﬂats ranged in age from
14,700 years old to less than 500 years old. This age range and the patterns of peat and
marl they found indicate that the formation of peat and marl ﬂats is a cyclic process, with
peat erosion forming a marl ﬂat that is subsequently covered by accumulating peat
(Miner and Ketterling 2003). Through the process of marl deposition, peat formation, and
erosion, a continuum of soils are present within prairie fen that form the substrate for a

variety of plant communities.

Plant Community Gradient

The plant community in prairie fen has been characterized as a gradient that
includes four (Spieles et al. 1999) to ﬁve (Bowles et al. 1996, Bowles et al. 2005)
vegetation zones. Here, I use the deﬁnitions outlined by Spieles et a1. (1999), which
include sedge meadow, inundated ﬂats, wooded fen, and marl ﬂat. The occurrence and
amount of each vegetation zone varies from fen to fen.

Sedge meadow plant communities primarily occur on sapric peat and have the
greatest herbaceous cover and diversity of the plant communities, including shrubs,

composites, and grasses (Spieles et al. 1999). The sedge-shrub association contains

7

Carex stricta (meadow sedge) and Carex aquatilis, in addition to low-growing shrubs,
especially Potentillaﬁwicosa (shrubby cinquefoil) (Spieles et a1. 1999). The sedge-
composite association contains Carex stricta, C. saquatilis, and C. sterilis, along with
Aster spp. (asters), Eupatorium perfoliatum (boneset), and Eupatorium maculatum (J oe-
pye weed) (Spieles et al. 1999). The sedge-grass association is commonly dominated by
the same Carex species as sedge-composite, as well as Andropogon gerardii (big
bluestem), A. scoparious (little bluestem), and Sorghastrum nutans (Indian grass)
(Spieles et al. 1999). Pycnanthemum virginianum (Virginia mountain mint),
Muhlenbergia glomerata (marsh wild-timothy), Thelypteris palustris (marsh fern),
Solidago ohioensis (Ohio goldenrod), and Rudbeckia hirta (black-eyed Susan) are
common species in all of the sedge meadow associations.

The inundated ﬂats zone occurs in depressions near rivers or lakes within the fen,
and can be localized around the water body or more expansive. Inundated ﬂat is the
wettest area in the fen, with up to 0.3 m of standing water in spring and summer.
Dominant plant species include a group of rushes: Cladium mariscoides (twig-rush),
Scirpus acutus (hardstem bulrush), Eliocharus rostellata (spike-rush), E. elliptica
(golden-seeded spike rush), and Juncus brachycephalus (Spieles et al. 1999).

Prairie fen may include a wooded zone dominated by shrubs and trees and
grading into uplands. In addition, wooded fen zones may also occur in wetter areas, with
Larix Iaricz'na (tamarack) dominating the tree composition. In some cases Ulmus
americana (American elm) and Acer rubrum (red maple) are present in this zone.

Common shrub species include Cornusfoemina (gray dogwood), C. stolomfera (red-osier

dogwood), T oxicodendron vernix (poison sumac), and Spiraea alba (meadowsweet)
(Spieles et al. 1999).

Areas where calcareous groundwater seepage enters the prairie fen contain marl
ﬂat vegetation. Sparse vegetation dominated by calciﬁles characterizes these areas, which
may be small and broken or broad and ﬂat. Species common in this area include Labelia
kahlmii (bog lobelia), Carexﬂava (sedge), Rhynchospora alba (beak-rush), and
Parnassia glauca (grass-of-parnassus). Carnivorous plants, including Sarracem’a
purpurea (pitcher plant) and Drosera rotundifolia (round-leaved sundew) may also be
found in this plant community.

The result of this soil and vegetation gradient is that the habitat frequently known
as prairie fen contains a variety of sub-communities. This continuum is part of the reason
for the unusually high species diversity in prairie fens, although all of these plant
communities are not present in each prairie fen. In summary, habitats discussed as prairie
fen here are groundwater fed, water saturated, alkaline habitats with soils that range from
mineral marl to ﬁbric and sapric peat.

In addition to a plant community gradient, tussocks formed by C. stricta provide
unique microhabitat for a range of plant species occurring from wet to dry. Peach and
Zedler (2006) found high species richness on tussocks due to the increased surface area,
the seasonal change in composition, and the presence of multiple microhabitats on each
tussock. The result was greater species richness per area on tussocks than the fen

substrate.

Insects in Prairie Fen

Several studies have examined the insect community in prairie fen. Panzer et al. (1995)
reviewed literature and surveyed habitat remnants in the Chicago region to determine a
set of remnant-dependant insects associated with the savanna-prairie habitat continuum.
They found that Lepidoptera were one of the most remnant-dependant insect groups.
Michigan prairie fens alone contain 25 insect species of conservation concern, including
6 moth species in the genus Papaipema, ranked as special concern, threatened, or

endangered, in large part due to their endemism (Table 1). Of the arthropods in prairie

Table 1. The 25 federally endangered (E), candidate (T), and special concern (SC) insect
species in Michigan prairie fen communities (MNF12007).

 

 

. State
Order Genus and specres Common Name Status
Odonata Cordulegaster erronea Tiger Spiketail SC
Williamsonia ﬂetcheri Ebony Boghaunter SC
Orthoptera Neoconocephalus lyristes Bog Conehead SC
Oecanthus laricus Tamarack Tree Cricket SC
Orchelimum concinnum Red-faced Meadow Katydid SC
Paroxya hoosieri Hoosier Locust SC
Homoptera Flexamia huroni Huron River Leafhopper SC
Flexamia reﬂexus Leafhopper SC
Lepyronia angulifera Angular Spittlebug SC
Prosapia ignipectus Red-legged Spittlebug SC
Coleoptera Liodessus cantralli Cantrall's Bog Beetle SC
Stenelmis douglasensis Douglas Stenelmis Rifﬂe Beetle SC
Lepidoptera Calephe/is mutica Swamp Metalmark SC
Euphyes dukesi Dukes' Skipper T
Hemileuca maia Barrens Buckmoth SC
Meropleon ambifusca Newman's Brocade SC
Neon ympha mifchelii mitchellii Mitchell's Satyr E
Oarisma poweshiek Poweshiek Skipperling T
Papaipema beeriana Blazing Star Borer SC
Papaipema cerina Golden Borer SC
Papaipema maritime Maritime Sunﬂower Borer SC
Papaipema sciata Culvers Root Borer SC
Papaipema silphii Silphium Borer Moth T
Papaipema speciosissima Regal Fem Borer SC
Spartinipﬁga lnops Spartina Moth SC

 

10

fen, butterﬂy species are likely the most well-recognized and documented group in
Southern Lower Michigan. The greater Washtenaw County butterﬂy survey has
contributed to this knowledge, with documented Lepidopteran sightings from 1994
onward (Kuhlman 2007). Bultman (1992) studied the cursorial spiders associated with
prairie fens and oak hickory forest. He found that the spider community in the prairie and
seep regions of prairie fen was not the same as that in oak-hickory forest, while the
community in forested fen was similar to that in oak hickory forest. Similarly, Lammers-
Campbell (1998) examined the community of Chironomidae in a fen and surrounding
vegetation. She found that chironomid species and plant community corresponded Well

with each other, but did not ﬁnd any chironomid species that were speciﬁc to fen only.

Arthropods may not only be indicators of prairie fen community, but also play a
key role in formation of fen structure and vegetation. Lesica and Kannowski (1998)
examined the role that ants play in forming structure and altering the plant community of
fens. They found that while there was more magnesium, phosphate, sodium, and
potassium in vegetation hummocks, there were higher nutrient levels in ant mounds,
indicating that hummocks may be abandoned ant mounds and providing evidence that
ants are key to the formation of soil microtopography. They also found that several
grasses, including Muhlenbergia sp., grow on active ant mounds. Although each of the
experimental studies focuses on a distinct group of insects, it is clear that insects in
prairie fen are frequently associated with vegetation structure and type, regardless of

whether they are the cause of (or response to) that vegetation.

ll

Historical Disturbance Patterns

Although perennially moist, the large volume of sedge and grass species in prairie
fen historically provided enough fuel to support frequent ﬁre (Curtis 1971). Fire
maintained the open character of fens, so that while shrubs and trees are part of the
system, they did not historically dominate large areas of prairie fens. Fires occurred in the
past as a result of drought, lightning strikes and burning by native Americans (Albert
1995, Anderson et al. 2000).

Since about 12,000 BP, with the ﬁrst evidence of humans in North America,
humans have modiﬁed the landscape with ﬁre for a variety of purposes (Buckner 2000,
Dey and Guyette 2000, Williams 2000a). Paleo-lndian hunter-gatherers used ﬁre
indirectly and directly for hunting. Fire was used directly to trap large mammals for
hunting, and indirectly to increase the production of forbs and grasses that provide habitat
for game (Dey and Guyette 2000). In addition, ﬁre was used to increase the yield of nuts,
berries, and acorns, all of which provided food directly to humans and also increased
browse for game species (Dey and Guyette 2000). European settlers used ﬁre to clear
agricultural ﬁelds, as well as to improve the quality of forage for range livestock
(Anderson et al. 2000, Dey and Guyette 2000). In a site in Southern Ontario, Dey and
Guyette (2000) found that the mean ﬁre return interval in red oak forest was 15 years
from 1600-1900, but that varied from 5 to 76 years, illustrating that even with
anthropogenic ﬁre regimes, ﬁre frequency was variable.

The General Land Ofﬁce surveyor notes for the Jackson interlobate region of
Michigan in the early 1800’s contained historic references to American Indian ﬁres in the
oak savanna and barrens (Albert 1995). Even though European settlers used ﬁre for

management, ﬁre frequency decreased with European settlement. The average ﬁre return

12

interval in a forest, savanna and fen mosaic in the Missouri Ozarks from 1710 to 1830
ranged from every 3.0 to 4.8 years, while during European settlement that increased to
every 6.9-12.5 years (Dey et a1. 2004). In the early 20th century, complete ﬁre
suppression became the primary mode of land management. The result of lengthening
ﬁre return intervals in the 1900’s was that the prairie savanna landscape, including prairie
fen communities, was colonized by trees and shrubs over time (Buckner 2000, Dey and
Guyette 2000). One factor preventing prairie fens from becoming forested since the
increase in ﬁre interval is the high water table that prevents most trees, shrubs, and
invasive plant species from establishing.

Historic disturbance in prairie fen also included grazing by ungulates, including
white-tailed deer, red deer, elk, moose, and auroch (Middleton et al. 2006a, Rydin et a1.
2006). In addition, beaver historically created ponds by damming waterways, which after
abandonment succeeded to marsh meadows or fens (Rydin et al. 2006). Fire and grazing
thus combined to play a role in shaping the Midwestern landscape, with a matrix of open

communities and forested habitats in upland areas (Anderson et al. 2000).

Threats to prairie fen systems and their management

There are a number of threats to prairie fen habitats, all of which are either a
direct or indirect result of human impacts on the landscape. These include habitat
fragmentation, changes in hydrology, invasive species, changes in nutrient inputs and
increased pollution, and changes in the fen disturbance regime. In many cases these
factors interact to impact prairie fen communities. The effects of these threats both
individually and together mean that prairie fen systems will need to be managed for their

persistence.

l3

Conversion

Worldwide, the primary threat to prairie fen habitat worldwide is the direct
conversion of prairie fens to agricultural and other land uses (Grootjans et al. 2006, van
Diggelen et al. 2006). In Europe and Central Asia, wetlands have been drained since the
Middle Ages for hay production and cattle grazing. In the 20th century, drainage
technology developed further, so that crop production was possible on drained wetlands,
leading to a large increase in wetland drainage (van Diggelen et al. 2006). Wetlands are
also drained for peat harvest for garden soil enrichment and as a fuel, though bogs are
more commonly mined for peat than fens, in areas where peat is rare prairie fen may be
used to provide peat. After peat extraction, an area of open water remains (Grootjans et
al. 2006), which no longer provides substrate for fen plants. Marl in prairie fens has also
been extracted to provide lime on agricultural soils, as well as for cement production
(Morrison 1945).
Fragmentation

In addition to direct conversion of prairie fens, the landscape surrounding these
habitats in the Midwestern U.S. historically contained oak savanna and upland prairie
communities. Both of these community types have been converted to agricultural land
since the 1830’s. More recently these lands are being intensively developed, so that
residential development adjoins prairie fen habitats (Amon et al. 2002) and disturbance
has increased or changed in the area between prairie fens (Bedford and Godwin 2003). In
many cases, plant and animal species in prairie fen communities, which by their nature
were already isolated from each other, have less suitable habitat for dispersal to nearby

prairie fen communities. The result of this decreased dispersal is that plant establishment

14

depends on the seed bank exclusively in many cases, rather than on a combination of a
seed bank and seed dispersal (Middleton et al. 2006a). Additional negative genetic effects
of small populations, such as genetic drift, could be detrimental to fen biodiversity in the
long-term.
Hydrological changes

As groundwater-fed communities, prairie fen habitats are extremely sensitive to
changes in hydrology. Alteration of groundwater recharge areas surrounding the prairie
fen can change the volume of water and nutrient that enter the fen system (Boeye and
Verheyen 1992). Quarrying, ﬁeld tiling, ditching, irrigation, ﬁlling, addition of
impermeable surfaces (such as roads), and residential development in the uplands
bordering prairie fen all disrupt groundwater ﬂow into fens (Amon et al. 2002, Bedford
and Godwin 2003). Water withdrawal from aquifers caused by irrigation can alter the
depth to the water table, reducing moisture in the plant rooting zone in fens (Amon et al.
2002). In prairie fens with reduced groundwater input, rainwater can ﬂush away minerals
key to plant community and soil development (Spieles et al. 1999). Oxygen also becomes
more common in soil with depleted water levels, disrupting the anoxic environment
characteristic of prairie fen (Amon et al. 2002). Both of these changes lead to succession
that is not characteristic of fens.
Changes in disturbance regime

Decreasing frequency of ﬁre, grazing, and mowing have led to a variety of
changes in prairie fen communities. All three disturbance types historically created open
niches for new plant establishment, and likely played a key role in maintaining high plant

diversity in prairie fen. With no grazing, mowing, or ﬁre management, Jensen and

15

Schrautzer (1999) found that 23 species were likely to become extinct in a German fen
system. This group is most likely to include small-seeded species, which depend on bare
soil provided by disturbance for establishment and germination.

Several studies indicate that prairie fen plant diversity decreases over time
without ﬁre. Kost and De Steven (2000) found increased perennial forb cover in one
Wisconsin sedge meadow following a prescribed burn, and their ﬁndings indicate that
ﬁre provides 1-2 years for forb recruitment due to the decreased litter layer, leading to an
increase in plant diversity. Weltzin et al. (2005) examined the effect of litter removal and
plant community composition in a fen and found that litter removal increased light
availability, soil temperature, and phosphorous in aboveground plant material, along with
altering plant communities. Fire burns leaf litter and provides openings in plant
communities that likely lead to similar effects on plant communities. Bowles and Jones
(2006) examined a group of wetlands in the Chicago region, and found that fens burned
every 5 years over a 30 year period maintained species richness, while those burned less
frequently did not. In addition, Middleton (2002) performed controlled burns in sedge
meadow communities in Wisconsin, and found an increase in forb species richness after
burning, including the presence of species that had not been documented in the habitat
over the prior 20 years.

While ﬁre may negatively affect insect populations in the short term (Swengel
1998), it is clear that the plant community that provides insect habitat does not persist in
the absence of ﬁre. Low-intensity patchy ﬁre, such as that historically present in this
landscape, would provide refuges for insect re-population. Panzer and Schwartz (2000)

found increased species richness and population densities of most remnant-dependant

16

insect species with rotational burning and ﬁre return intervals of 2-5 years. Williams
(2000b) recommends reintroducing ﬁre into the landscape mosaic that includes prairie
fen, with the caveat that it be “Indian-type” ﬁre. This type of burn involved a mosaic of
habitats that were exposed to low-intensity, patchy ﬁre. Additionally, burns can be timed
to avoid periods when insects of concern are in immobile stages, can be carried out in
rotating burn units, and burns can be conducted at varying times in the growing season to
minimize the effects on one group of insects or plants (Reed 1997, Panzer 2002).

Mowing of prairie fen has been practiced in Europe for hundreds of years for hay
production. Beginning in the 1970’s many of those fens were no longer mowed, which
has led to increased shrub cover and decreased plant diversity (Middleton et al. 2006b).
In addition to controlling shrub growth and maintaining high species diversity, mowing
has been used effectively as a way to decrease the effects of high nitrogen inputs
(V erhoeven et al. 1996).

Prairie fen communities, with their grass and sedge rich ﬂora, presented a habitat
for cattle grazing, and fen meadows and partially drained fens were used for this purpose
through much of the 20th century in both Europe and North America (Middleton et al.
2006b). Grazing has both positive and negative impacts on prairie fen. Positive impacts
include maintenance of open habitat in the absence of ﬁre. On the other hand, cattle can
break down the characteristic hummock structure in prairie fens, decreasing the
microhabitat variability available to plant species. In addition, when cattle are removed
ﬁom the habitat, shrubs that cattle grazing had kept small often grow rapidly and become
dominant in the fen (Middleton 2002). Reintroducing grazing in prairie fen, however, is

not likely to lead to former plant diversity, as species with seeds that do not persist in the

17

seed bank will not return with reintroduction of grazing after long periods (Hald and
Vinther 2000).
Pollution, nutrient inputs, and invasive species

A variety of pollution types increasingly enter wetland systems, including
chlorine and sodium from road salt, nutrients from septic tanks, herbicide runoff, and
fertilizer runoff. Panno et al. (1999) found increased abundance of invasive narrow-
leaved cattail and decreasing diversity of native vegetation near a site of nutrient addition
and pollution from a nearby septic system and roadway, which contributed Cl- and Na+
to the groundwater. Evidence indicates that pollution leads to changes in the fen plant
community.

Changes in nutrient availability may lead to changes in plant dominance within
prairie fens. Increased nitrogen deposition in Europe has changed nutrient inputs in
wetlands so that some plant communities that were formerly nitrogen limited become
phosphorous limited (Verhoeven et al. 1996). Without management such as mowing,
plant community succession changes and decreased species diversity often result
(Verhoeven et al. 1996).

Prairie fens are increasingly impacted by invasive plant species. Bowles and Jones
(2006) compared changes in 31 wetland communities in the Chicago region between
1976 and 2002. They found increased abundance of invasive plant species purple
loosestrife (Lythrum salicaria), glossy (F rangula alnus) and common buckthom
(Rhamnus cathartica), and narrow-leaved cattail (Typha angustifolia). Additionally, they

determined that increased abundance of both invasive narrow-leaved cattail and native

18

broad-leaved cattail (Typha latifolia) were associated with signiﬁcant decreases in native
species richness within study plots (Bowles and Jones 2006).

The effects of invasive species are often compounded by increased nutrient
availability, which has potential to increase growth of invasive species but not native
species, leading to a competitive advantage for the invasive. Woo and Zedler (2002)
determined that nitrogen and phosphorous addition to Typha x. glauca led to increased
height, biomass, and rarnet density, but had no signiﬁcant affect on native graminoids.
Rickey and Anderson (2004) found a similar pattern with Phragmites austrailis, with
increased Phragmites biomass with nitrogen addition but no signiﬁcant change in native
Spartina pectinata biomass. Lindig-Cisneros and Zedler (2002) found that reed canary
grass (Phalaris arundinaceae) was more likely to establish in areas with ﬂuctuating water
levels than static water levels. In addition, they found that P. arundinaceae required light
availability for germination, which would be caused by a disturbance in prairie fen.

F our changes in environmental conditions are linked to the spread of aquatic
invasives: changes in wetland hydrology, increased nutrient levels, increased salinity, and
removal of vegetation (Galatowitsch et al. 1999). All of these changes occur with
increasing frequency in prairie fens. The result is that a veritable maelstrom of factors is
interacting to affect the existence of prairie fen communities. Although the factors
threatening prairie fens all stem from human activities, ironically, prairie fens are
unlikely to persist without human intervention to maintain them. In Europe, many areas
considered natural fens area actually fen meadows that have been slightly drained and
used for grazing or mowing, and in North America many prairie fens have been grazed

historically (van Diggelen et a1. 2006).

19

Glossy Buckthorn

Life history and arrival in Michigan

Glossy buckthom (Rhamnaceae: F rangula alnus P. Mill. (previously Rhamnus
frangula L.)) is a shrub of Eurasian origin that is especially aggressive in bogs and fens
(Voss 1985). Frangula alnus was likely introduced into North America by 1800,
cultivated as a hedge plant and for wildlife habitat, but did not naturalize until the 1900’s.
F rangula alnus establishes well in acid to alkaline peat soils with some moisture,
although it is does not survive permanent waterlogging (Godwin 1943). In peat soils, the
root system grows very near the surface, above the water table (Godwin 1943). Frangula
alnus forms even-aged stands and individuals grow to 7 m tall; the species is sometimes
considered a small tree (Converse 1984). Glossy, smooth, oval leaves with parallel
venation open on this shrub earlier than many native woody species in the US. Midwest,
and stay green late into the fall (Godwin 1943). '

Reproduction is primarily sexual, although asexual reproduction is reportedly
possible in the absence of insect pollination. Frangula alnus blooms from the end of May
— September, and ﬂowers are visited by bees, wasps, ichneumonids, and beetles (Godwin
1943). Ovate drupes containing 2-3 pyrenes each are produced in large numbers annually
(Godwin 1943). Drupes are green when immature, then red changing to black (Converse
1984). A large variety of bird species eat F. alnus berries and disperse seeds (Catling and
Porebski 1994). Seed dispersal is also carried out by mice and elk, and dispersal in water
may also be possible; dry F. alnus seeds ﬂoat for one week, and ﬁ'uits ﬂoat up to 19 days
(Converse 1984). Seedlings establish most easily in areas with exposed soil and available

‘20

light (Godwin 1943), and F. alnus has greatest seedling density near the source of seed
(Converse 1984). Seeds are viable for at least 6 years (Walsh et al. 2003), although if in a
dry location for several months the germination rate is very low (Converse 1984).

F rangula alnus is currently widespread in North America, from Nova Scotia
west to Manitoba in Canada and in the Eastern US. as far south as Kentucky (Sinclair
and Catling 1999). F rangula alnus was ﬁrst collected in Michigan in Delta County in the
upper peninsula in 1934, although it was likely present in the state before this time (Voss
1985). F rangula alnus is most common in Michigan in fens, bogs, disturbed clamp to wet
areas, lake shores, and thickets along rivers (Voss 1985), although it will survive in

forests, as well.

Impacts on habitat

There is evidence that F. alnus displaces native plant species in its invasive range.
In an Ontario wetland, Sinclair and Catling (1999) compared plant growth in plots with
F. alnus removed versus beneath F. alnus cover. They found both greater herbaceous
percent cover and species richness in removal plots. In contrast, Houlahan and Findlay
(2004) examined wetlands in Ontario and found that when considering the entire wetland,
exotic plant species, including F. alnus, were not more likely to be dominant plants than
native species. Their results indicate that exotic plant species do not competitively
exclude native plant species across wetlands. This pattern is focused on species
dominance and not species identiﬁcation; species composition in the presence of F. alnus
may change in wetlands. There may also be effects of scale on this pattern, so that at

smaller scales native plant diversity is low where exotic plant cover is high.

21

F rangula alnus may also impact a variety of abiotic factors in prairie fen habitats.
Transpiration of F. alnus in midsummer has been shown to lower the water table
(Godwin 1943), making prairie fen systems increasingly suitable for F. alnus and less
suitable for native prairie fen species. The organic matter hurnmocks characteristic of
prairie fen communities may provide a wet but not inundated microhabitat conducive to
F. alnus seedling establishment. In the presence of F. alnus, those organic matter
hummocks are broken down, reducing microhabitat differences within the fen. This
process may be assisted by an overall decrease in the water table. In addition, F. alnus
may alter nitrogen cycling in prairie fen systems; related common buckthom (Rhamnus
cathartica L.) is associated with elevated nitrogen levels in Illinois woodlands (Heneghan
et al. 2006). Concurrent with increased nitrogen levels, leaf litter decomposition rates
may be altered in the presence of F. alnus, resulting in changes in nitrogen mineralization
rates. Weltzin et al. (2005) found that the plant community in a northern prairie fen
changed with litter removal. This indicates the potential that changes in leaf litter
thickness below F. alnus lead to changes in plant community composition in prairie fen
communities. The potential for F. alnus to alter communities that it invades is very real,
and exploration of further potential factors affecting prairie fen communities may yield

evidence of impacts on a variety of abiotic as well as biotic factors.

Management techniques and implications

A variety of management techniques have been tested for control of adult and
seedling F. alnus; some have proven more successful than others. Biological control of F.
alnus and Rhamnus cathartica has been investigated but candidate insects did not feed on

F. alnus. Finding and testing herbivores of F. alnus is a current priority (CABI

22

bioscience, www.cabi.org). Oat rust (Puccina coronata) does infect F. alnus, but it is a
secondary host and P. coronata does not appear to cause F. alnus mortality (Peturson
1949). Cultural control methods of F. alnus in wetlands include excavation, mowing,
stump cutting, burning, and stem girdling (Converse 1984). F rangula alnus resprouts
vigorously after top kill or top removal, so that a single cutting results in thick regrowth
with more stems than the initial growth. Techniques for herbicide application, however,
have been extensively tested to maximize effective control (Converse 1984, Reinartz
1997, Pergams and Norton 2006).
Physical control techniques

Excavation may be useful for small plants and seedlings. With small plants, hand-
pulling or removal with a grubbing hoe may be successful. Larger F. alnus plants may be
pulled out using heavy equipment, but this method results in soil disturbance, not only
leaving an open area to be colonized, but also bringing seeds of F. alnus or other invasive
plants to the surface where they germinate readily (Converse 1984). Repeated mowing or
cutting reduces F. alnus plant vigor. Sinclair and Catling performed one cutting of F.
alnus in an invaded Ontario wetland and found greater native plant species cover and
diversity after one growing season in the cut area (Sinclair and Catling 1999). Mowing
may be a temporary way to increase native plant growth in areas where there aren’t
resources for herbicide. In addition, annual mowing has been used in European prairie
fens to decrease the effect of large nutrient inputs and maintain plant species richness
(Gusewell and Le Nedic 2004, Middleton et al. 2006b). Burning is another technique that
has been used with relatively low success in an attempt to control F. alnus. A burn in late

April or early May in the Midwest, when F. alnus has leafed out but native plant species

23

have not, has potential to top kill it, eliminating the main growing point, and reduce
resprouting. Stem girdling involves destroying the surface of a 2-5 cm band to interrupt
carnbium ﬂow (Converse 1984). Although girdling doesn’t disrupt the soil or require
herbicide, Reinartz (1997) found that girdling alone did not cause F. alnus mortality in a
Midwestern wetland. This method is yet another that suppresses F. alnus growth but does
not effectively cause mortality.
Chemical control techniques

While physical control techniques may lead to suppression of buckthom, physical
techniques combined with chemical control have proven extremely effective in killing F.
alnus adults. Herbicide may be applied to uncut stems (basal bark treatment), girdled
stems, cut stumps, ﬁilled stems, or foliage, or injected into the trunk. Foliage spray of F.
alnus requires the largest volume of chemical, and has been carried out with fosamine
(ammonium salt) in September with high F. alnus mortality rates (Converse 1984). Stem
cutting or girdling plus herbicide application lead to high buckthom mortality rates. The
most commonly used herbicides include glyphosate (Rodeo, among others), a non-
selective herbicide, and 2-4 D and triclopyr (Garlon and Pathﬁnder), which are speciﬁc to
broadleaf plants. Pergams and Norton (2006) performed a comparison of physical and
chemical control methods of R. cathartica (common buckthom) and found that girdling
or cutting plus stump treatment with herbicide of one stem on the plant usually led to
death. Reinartz (1997) compared mortality rates of F. alnus using stem girdling and
cutting alone versus stern cutting plus glyphosate herbicide treatment. He found that
neither girdling or cutting alone led to plant mortality, while cutting plus treatment of

25% concentration glyphosate led to 92 - 100% mortality of individual plants. Herbicide

24

safety in wetlands is of concern, as any surfactant used to increase herbicide penetration
into plants will also lead to ready herbicide spread in water. A formulation of glyphosate
that is safe for use in wetlands must be used to prevent water contamination. Glyphosate
should not be harmful to the surrounding watershed or non-target vegetation without a
surfactant, but will degrade more slowly in anaerobic conditions presented by most
prairie fen communities (Converse 1984).

Time of year also plays a role in control efﬁcacy of F. alnus. Stump cutting plus
herbicide treatment of F. alnus is known to be highly effective at any time of year except
when sap is rising from March — May. In addition, invasive management in prairie fen
habitats in winter minimizes trampling to plant communities, making December —
February an ideal time to cut and stump treat F. alnus. While no references mention
disposal of woody material after it is cut, The Nature Conservancy in Michigan uses a
method that involves stacking material, allowing it to dry, and burning bmsh piles (R.
Zuniga-Villegas, pers. comm). This reduces the area where the soil is heat sterilized,
which can decrease native plant re-establishment ﬁom the seed bank.

Long-term management

Even with herbicide treatment, some stumps resprout and followup treatment is
necessary to kill all buckthom plants. In addition, the proliﬁc production of F. alnus seed
leads to a large seed bank near adult plants. As seedling density is highest near adult F.
alnus plants, the target treatment area will be near previous adult populations (Converse
1984). Hand-pulling of seedlings is possible but time-consuming, and also has potential
to increase germination of additional F. alnus seeds due to disturbance. Use of a propane

ﬂame torch in areas with little other plant growth in the ﬁrst season after treatment of

25

adults is an effective treatment method (R. Zunigo-Villegas, pers. comm). Although the
largest ﬂush of seedlings will likely occur in the ﬁrst year following clearing (Frappier et
al. 2004), seedling management will need to continue for several years after adult F.
alnus management.

The establishment of F. alnus monocultures depends on disturbance (Godwin
1943). The reality of prairie fen systems, however, is that they are a community with
frequent anthropogenic disturbance and changes in water table. These factors mean that
continued management and scouting for F. alnus and other invasive species will be
necessary in many prairie fen systems. Although there are limited resources to control F.
alnus and other invasive plant species in natural habitats, the use of several techniques
will maximize the effectiveness of those efforts. First, the use of best practices to control
invasion will minimize time spent with control followup (Webster et al. 2006). For F.
alnus, this means cut stump treatment with herbicide rates and formulations that are
known to be effective. In addition, prioritizing control areas allows maximum use of
resources. Control directed at small satellite populations prevents satellite populations
ﬁ'om becoming new populations centers, reducing future control efforts (Webster et al.
2006). During the ﬁrst and second year of F. alnus management, graminoid vegetation is
recovering, so that there is often little dry biomass. In subsequent years, ﬁre can be
carried through prairie fen systems and prescribed burns every 2-5 growing seasons are a
key long-term element of fen management to prevent further invasion of non-natives and

maintain plant species diversity and richness.

26

Conservation of biodiversity and ecosystem function

As human impacts on the biosphere expand, conservation of natural habitats is
increasingly vital for the persistence of biodiversity on Earth. The perception of
conservation is changing, as well, as we recognize that merely preserving land does not
ensure the continued persistence of plant and animal communities and active
management is often required to maintain ecosystem services.

Services provided as a result of ecosystem function are known as ecosystem
services (Costanza et al. 1997). Ecosystem services fall into four broad categories:
supporting, provisioning, regulating, and cultural (Reid et al. 2005). Supporting services
include soil formation and nutrient cycling, provisioning include supplying fuel, food,
and fresh water, regulating services include regulation of climate, ﬂooding, and disease,
and cultural services include educational, spiritual, and aesthetic beneﬁts to humans.
Restoration of prairie fen habitats has potential to support and provide many of these

services.

Functions of wetlands

Freshwater wetlands provide a variety of ecosystem services which are invaluable
to humans. Wetlands speciﬁcally provide ﬂood control, nutrient decomposition, waste
puriﬁcation, water supply, and the ability to stabilize and moderate natural and urban
microclimates and to retain nutrients and sediments (Moser et al. 1996, Braga 1999,
Bedford and Godwin 2003). In addition, wetlands support biodiversity, which humans
value for the intrinsic and spiritual worth placed on them and also due to the perception

that biodiversity can provide beneﬁts in the future (Thompson and Starzomski 2007).
27

Healthy prairie fens may be deﬁned by their ability to provide these resources, all of
which are dependent on the ecosystem processes of water and mineral cycling,
community dynamics, and solar energy ﬂow. Therefore, one measurement of successﬁrl
prairie fen restoration is that it is a system with key ecosystem processes intact.
Reintroduction of the historical disturbance regime and elimination of invasive species

are two key steps to restoring that function.

Insect conservation and function

Insects are the most diverse group of taxa on earth, comprising 80% of all
metazoan species that have been described (Sarnways 2005). From 90 to 95% of insect
species remain undescribed; estimates of insect diversity on Earth range from 1.8 to 8.8
million species (Sarnways 1994), with recent estimates placing the total at 3.7 to 5.9
million species (N ovotny et al. 2002). Because so many insect species remain
undescribed, we do not know current insect extinction rates, but an estimated 11,200
insect species have become extinct since 1600, and 100,000-500,000 more are predicted
to become extinct by 2300 (Sarnways 2005). In North America, an estimated 44% of
insects have not been discovered or adequately described (Redak 2000). In 2000, 37
arthropod species were listed as federally endangered in the US, and estimates of the
actual number of species at risk range from 2000 to 29,000 species (Redak 2000). The
same processes that are putting rare habitats and the plant species in them at risk:
fragmentation, development, agriculture, pollution, altered disturbance regimes, and
nutrient inputs, are also putting insect species endemic to those systems at risk. Insect

species extinction may have far-reaching impacts on ecosystem structure and function.

28

Insects and their activities support a variety of key ecosystem functions and
processes. The majority of ﬂowering plants are insect-pollinated (Klein et al. 2006).
Detritivores, omnivores, and carrion feeders contribute to decomposition and nutrient
cycling (New 2005). Predator and parasitoid insects control herbivore populations, which
may be crop pests (Van Driesche and Bellows 1996). Herbivores control plants that may
otherwise become weedy (Goeden and Andres 1999).

In addition to these functions that humans value directly, insects also support a
multitude of key ecosystem processes. Insects are the primary energy link between plants
and other animals, and therefore are the basis for animal food webs. Insects also are
habitat architects. Herbivorous insects affect plant community structure, including
potential plant species presence or absence, through herbivory. Insects such as termites
and ants act as ecosystem engineers by moving and processing large volumes of soil and
detritus (Sarnways 2005). A variety of ecosystem processes, therefore, are dependent on

the presence of insects.

Pollination: a key process for ecosystem persistence

Pollination is a key insect-supported ecosystem function (Kevan 1999). Animal
pollination is carried out by a diversity of taxa, including bees (which are all obligate
ﬂower visitors), ﬂies, butterﬂies and moths, beetles, birds, bats, and small mammals
(Keams et al. 1998). Insect pollinators, like other arthropods, are threatened by habitat
loss, effects of pesticides, increased atmospheric carbon dioxide, and introduction of non-
. native plants and pollinators (Keams and Inouye 1997, Memmott and Waser 2002).
Pollination has signiﬁcant direct effects for humans: pollination services are valued at

$117/ha/yr (Costanza et a1. 1997) and 35% of all crops beneﬁt ﬁom insect pollination
29

(Klein et al. 2006). Beyond direct monetary effects that beneﬁt humans, pollination
provides additional indirect effects on gene ﬂow, metapopulation dynamics, and genetic
diversity (Neal 1998). The potential negative impacts of decreasing pollinator abundance
and diversity are very real. Over 90% of angiosperrns are animal pollinated (Keams et al.
1998). Biesmeijer et al. (2006) found decreased local bee diversity in Britain and the
Netherlands after 1980 versus pre-1980. They saw a concomitant decrease in the species
diversity of outcrossing plants, a pattern which could result in decreased populations of
outcrossing plant species and subsequent reductions in gene ﬂow within a species. It is
unclear whether plants or pollinators were lost ﬁrst; regardless, both groups are key
components in the persistence of natural ecosystems.

Declining species diversity and abundance may affect speciﬁc organisms in
ecosystems differently depending on their origin and habitat specialization. Memmott and
Waser (2002) examined the pollinator-plant interaction web of a community in Illinois.
They found that, on average, non-native plant species attracted fewer pollinator species
than native plants. In addition, generalist pollinators were more likely to visit non-native
plants than pollinator specialists. In their study of pollinators in Britain and the
Netherlands, Biesmeijer et al. (2006) found a greater decline in species abundance of
specialist than generalist bee species. The disappearance of specialist pollinators may or
may not have a broad impact on plant community diversity and persistence. Memmott
(1999) examined the plant-pollinator food web of a meadow, and found a complex web
of interactions dominated by general plant-insect and insect-plant relationships, rather
than the speciﬁc relationships that are often focused on in pollination biology research.

Similarly, Memmott et al. (2004) used a modeling approach to examine the effect of

30

pollinator extinction on plant species diversity. They found that when generalist
pollinators were removed from the model system, a more rapid decline in plant species
diversity resulted than with specialist pollinators removed. In addition, this decrease was
linear, and was not as low as was predicted.

Restoration of prairie fen, as a plant community that commonly occurs in small
remnants and contains many rare plant species, may be severely limited by the seed bank
and seed dispersal of native plants (Middleton et al. 2006a). Restoration is more likely to
be successful with functional interactions in place (Neal 1998). In addition, restoration of
prairie fen and other native habitats may help alleviate the pollinator crisis beyond habitat
edges, although recent ﬁndings indicate that effect of remnant high-quality habitats on
agricultural ﬁelds extends <150 m from them in an agricultural matrix (Kohler et al.
2008). Many of the rare habitats in the prairie fen continuum contain a diversity of
ﬂowering forbs (Spieles et al. 1999), and there has been a call to conserve diverse plant
patches to prevent and reverse pollinator decline (Kearns and Inouye 1997, Kearns et al.
1998). In contrast to high quality prairie fen, buckthom-invaded prairie fen contains very
little ﬂowering herbaceous vegetation (pers. obs.).

An additional concern is that while restoration of pollinator richness and
abundance may be enough to restore pollinator function, it may not be considered
sufﬁcient to restore a resilient habitat. In some cases, systems reach full function at a
portion of the total biodiversity (Schwartz et a1. 2000). This is related to the concept that
species within communities are functionally redundant (N aeem 1998, Wohl et al. 2004).
F orup and Memmott (2005) found no signiﬁcant difference in plant and insect richness or

abundance between restored and old hay meadows. Although the old and restored

31

meadows were similar in function, they found that there was a greater proportion of
potential species links in the old meadows. This ﬁnding suggests that in the case of
pollinators, increased species diversity may lead to increased ecosystem stability across
functional groups, a relationship that has been found in a number of studies (Peterson et
al. 1998, Lhomme and Winkel 2002). The increased species diversity could buffer
communities from disturbance and human impacts, indicating that although current
ecosystem function is of concern, ﬁrture resiliency of that function will likely need to be

considered.

32

Literature Cited

Albert, D. A. 1995. Regional landscape ecosystems of Michigan, Minnesota, and
Wisconsin: a working map and classiﬁcation. Gen. Tech. Rep. NC-178, St. Paul, MN.

Amon, J. P., C. A. Thompson, Q. J. Carpenter, and J. Miner. 2002. Temperate zone fens
of the glaciated Midwestern USA. Wetlands 22(2):301-317.

Anderson, R. C., J. E. Schwegrnan, and M. R. Anderson. 2000. Micro-scale restoration:
A 25-year history of a southern Illinois barrens. Restoration Ecology 8(3):296-306.

Bedford, B. L., and K. S. Godwin. 2003. F ens of the United States: distribution,
characteristics, and scientiﬁc connection versus legal isolation. Wetlands 23(3):608-629.

Biesmeijer, J. C., S. P. M. Roberts, M. Reemer, R. Ohlemuller, M. Edwards, T. Peeters,
A. P. Schaffers, S. G. Potts, R. Kleukers, C. D. Thomas, J. Settele, and W. E. Kunin.
2006. Parallel declines in pollinators and insect-pollinated plants in Britain and the
Netherlands. Science 313(5785):351-354.

Boeye, D., and R. F. Verheyen. 1992. The Hydrological Balance of a Groundwater
Discharge Fen. Journal of Hydrology 137(1-4):149-163.

Bowles, M., and M. Jones. 2006. Trends of change in composition and structure of
Chicago region wetland vegetation. Chicago Wilderness Journal 4(3):25-34.

Bowles, M., J. McBride, N. Stoynoff, and K. Johnson. 1996. Temporal changes in
vegetation composition and structure in a ﬁre-managed prairie fen. Natural Areas Journal
16(4):275-288.

Bowles, M. L., P. D. Kelsey, and J. L. McBride. 2005. Relationships among
environmental factors, vegetation zones, and species richness in a North American
calcareous prairie fen. Wetlands 25(3):685-696.

Braga, M. I. J. 1999. Integrating Freshwater Ecosystem Function and Services with
Water Development Projects. Inter-American Development Bank, Washington DC.

Buckner, E. 2000. In summary: ﬁre in the evolution of the Eastern landscape- a timeline.
Page 120 in Workshop on ﬁre, people, and the central hardwoods lanscape, Newtown
Square, PA: US. Department of Agriculture, Forest Service, Northeastern Research
Station.

Bultman, T. L. 1992. Abundance and Association of Cursorial Spiders from Calcareous
Fens in Southern Missouri. Journal of Arachnology 20(3):]65-172.

Buol, S. W., F. D. Hole, R. J. McCracken, and R. J. Southard. 1997. Histosols: Organic
soils. Page 527 Soil genesis and classiﬁcation. Iowa State University Press, Ames.

33

Catling, P. M., and Z. S. Porebski. 1994. The history of invasion and current status of
Glossy Buckthom, Rhamnusfrangula, in Southern Ontario. Canadian Field-Naturalist
108(3):305-310.

Converse, C. K. 1984. Element stewardship abstract for Rhamnus cathartica, Rhamnus
frangula (syn. F rangula alnus). The Nature Conservancy, Arlington, VA.

Costanza, R., R. dArge, R. deGroot, S. Farber, M. Grasso, B. Harmon, K. Limburg, S.
Naeem, R. V. Oneill, J. Paruelo, R. G. Raskin, P. Sutton, and M. vandenBelt. 1997. The
value of the world's ecosystem services and natural capital. Nature 387(6630):253-260.

Curtis, J. T. 1971. The vegetation of Wisconsin; an ordination of plant communities.
University of Wisconsin Press, Madison, WI. 657.

Dey, D. C., and R. P. Guyette. 2000. Anthropogenic ﬁre history and red oak forests in
south-central Ontario. Forestry Chronicle 76(2):339-347.

Dey, D. C., R. P. Guyette, and M. C. Starnbaugh. 2004. Fire history of a forest, savanna,
and fen mosaic at white ranch state forest. 132-137.

Forup, M. L., and J. Memmott. 2005. The restoration of plant-pollinator interactions in
hay meadows. Restoration Ecology 13(2):265-274.

Frappier, B., R. T. Eckert, and T. D. Lee. 2004. Experimental removal of the non-
indigenous shrub Rhamnusﬁ—angula (glossy buckthom): Effects on native herbs and
woody seedlings. Northeastern Naturalist 11(3):333-342.

Galatowitsch, S. M., N. 0. Anderson, and P. D. Ascher. 1999. Invasiveness in wetland
plants in temperate North America. Wetlands 19(4):733-755.

Godwin, H. 1943. F rangula alnus Miller (Rhamnusfrangula L.). Journal of Ecology
3 1 :77-92.

Goeden, R. D., and L. A. Andres. 1999. Biological control of weeds in terrestrial and
aquatic environments. Pages 871-889 in R. D. Goeden, L. A. Andres, T. W. Fisher, T. S.
Bellows, and L. E. Caltagirone, editors. Handbook of biological control : principles and
applications of biological control. Academic, San Diego ; London.

Grootjans, A. P., E. B. Adema, W. Bleuten, H. Joostn, M. Madaras, and M. Janakova.
2006. Hydrological landscape settings of base-rich fen mires and fen meadows: an
overview. Applied Vegetation Science 9(2): 1 75-1 84.

Gusewell, S., and C. Le Nedic. 2004. Effects of winter mowing on vegetation succession
in a lakeshore fen. Applied Vegetation Science 7(1):41-48.

Hald, A. B., and E. Vinther. 2000. Restoration of a species-rich fen-meadow after

abandonment:
response of 64 plant species to management. Applied Vegetation Science 3:15-24.

34

Heneghan, L., F. Faterni, L. Umek, K. Grady, K. Fagen, and M. Workman. 2006. The
invasive shrub European buckthom (Rhamnus cathartica L.) alters soil properties in
Midwestern US woodlands. Applied Soil Ecology 32(1):142-148.

Houlahan, J. E., and C. S. Findlay. 2004. Effect of invasive plant species on temperate
wetland plant diversity. Conservation Biology 18(4):]132-1138.

Jensen, K., and J. Schrautzer. 1999. Consequences of abandonment for a regional fen
ﬂora and mechanisms of successional change. Applied Vegetation Science 2(1):79-88.

Kearns, C. A., and D. S. Inouye. 1997. Pollinators, ﬂowering plants, and conservation
biology - Much remains to be learned about pollinators and plants. Bioscience 47(5):297-
307.

Keams, C. A., D. W. Inouye, and N. M. Waser. 1998. Endangered mutualisms: The
conservation of plant-pollinator interactions. Annual Review of Ecology and Systematics
29:83-112.

Kevan, P. G. 1999. Pollinators as bioindicators of the state of the environment: species,
activity and diversity. Agriculture Ecosystems & Environment 74(1-3):373-393.

Klein, A. M., I. Steffan-Dewenter, and T. Tschamtke. 2006. Rain forest promotes trophic
interactions and diversity of trap-nesting hyrnenoptera in adjacent agroforestry. Journal of
Animal Ecology 75(2):315-323.

Kohler, F ., J. Verhulst, R. van Klink, and D. Kleijn. 2008. At what spatial scale do high-
quality habitats enhance the diversity of forbs and pollinators in intensively farmed
landscapes? Journal of Applied Ecology 45(3):753-762.

Kost, M. A., and D. De Steven. 2000. Plant community responses to prescribed burning
in Wisconsin sedge meadows. Natural Areas Journal 20(1):36-45.

Kuhlman, R. 2007. Greater Washtenaw County Butterﬂy Survey Homepage. Ann Arbor,
MI.

Lammers-Campbell, R. 1998. Ordination of chironomid (Diptera : Chironomidae)
communities characterizing habitats in a Minnesota peatland. Journal of the Kansas
Entomological Society 7 l (4):414-425.

Lesica, P., and P. B. Kannowski. 1998. Ants create hummocks and alter structure and
vegetation of a Montana Fen. American Midland Naturalist 139(1):58-68.

Lhomme, J. P., and T. Winkel. 2002. Diversity-stability relationships in community
ecology: Re-examination of the portfolio effect. Theoretical Population Biology
62(3):271-279.

35

Lindig-Cisneros, R., and J. B. Zedler. 2002. Phalaris arundinacea seedling establishment:
effects of canopy complexity in fen, mesocosm, and restoration experiments. Canadian
Journal of Botany-Revue Canadienne De Botanique 80(6):617-624.

Memmott, J. 1999. The structure of a plant-pollinator food web. Ecology Letters
2(5):276-280.

Memmott, J ., and N. M. Waser. 2002. Integration of alien plants into a native ﬂower—
pollinator visitation web. Proceedings of the Royal Society B: Biological Sciences 269(
1508 ):2395 - 2399.

Memmott, J ., N. M. Waser, and M. V. Price. 2004. Tolerance of pollination networks to
species extinctions. Proceedings of the Royal Society of London Series B-Biological
Sciences 271(1557):2605-2611.

Middleton, B. 2002. Winter burning and the reduction of Camus sericea in sedge
meadows in southern Wisconsin. Restoration Ecology 10(4):723-730.

Middleton, B., R. van Diggelen, and K. Jensen. 2006a. Seed dispersal in fens. Applied
Vegetation Science 9(2):279-284.

Middleton, B. A., B. Holsten, and R. van Di ggelen. 2006b. Biodiversity management of
fens and fen meadows by grazing, cutting and burning. Applied Vegetation Science
9(2):307-3 16.

Miner, J. J ., and D. B. Ketterling. 2003. Dynamics of peat accumulation and marl ﬂat
formation in a calcareous fen, Midwestern United States. Wetlands 23(4):950-960.

MNFI. 2007. Rare Species Explorer.in M. N. F. Inventory, editor., Lansing, MI.

Morrison, P. C. 1945. Cement plant migration in Michigan. Economic Geography
21(1):1-16.

Moser, M., C. Prentice, and S. Frazier. 1996. A global overview of wetland loss and
degradation.in 6th Meeting of the Conference of the Contracting Parties. Wetlands
International, Brisbane, Australia.

Naeem, S. 1998. Species redundancy and ecosystem reliability. Conservation Biology
12(1):39-45.

Neal, P. R. 1998. Pollinator restoration. Trends in Ecology & Evolution 13(4):132-133.

Nekola, J. C. 1994. The environment and vascular ﬂora of northeastern Iowa fen
communities. Rhodora 96(886):121-169.

New, T. R. 2005. Invertebrate conservation and agricultural ecosystems. Cambridge
University Press, Cambridge, UK; New York. xii, 354 p.

36

Novotny, V., Y. Basset, S. E. Miller, G. D. Weiblen, B. Bremer, L. Cizek, and P. Drozd.
2002. Low host speciﬁcity of herbivorous insects in a tropical forest. Nature
416(6883):841-844.

Panno, S. V., V. A. Nuzzo, K. Cartwright, B. R. Hensel, and I. G. Krapac. 1999. Impact
of urban development on the chemical composition of ground water in a fen-wetland
complex. Wetlands 1 9(1 ):236—245.

Panzer, R. 2002. Compatibility of prescribed burning with the conservation of insects in
small, isolated prairie reserves. Conservation Biology 16(5):1296-1307.

Panzer, R., and M. Schwartz. 2000. Effects of management burning on prairie insect
species richness within a system of small, highly fragmented reserves. Biological
Conservation 96(3):363-369.

Panzer, R., D. Stillwaugh, R. Gnaedinger, and G. Derkovitz. 1995. Prevalence of remnant
dependence among the prairie-inhabiting and savanna-inhabiting insects of the Chicago
region. Natural Areas Journal 15(2): 101-1 16.

Peach, M., and J. B. Zedler. 2006. How tussocks structure sedge meadow vegetation.
Wetlands 26(2):322-335.

Pearson, J. A., and M. J. Loeschke. 1992. Floristic composition and conservation status
of fens in Iowa. Journal of the Iowa Academy of Sciences 99:41-52.

Pergams, O. R. W., and J. E. Norton. 2006. Treating a single stem can kill the whole
shrub: a scientiﬁc assessment of buckthom control methods. Natural Areas Journal
26(3):300-309.

Peterson, G., C. R. Allen, and C. S. Holling. 1998. Ecological resilience, biodiversity,
and scale. Ecosystems l (1)16-1 8.

Peturson, B. 1949. The occurrence of Puccinia coronata on Rhamnusfrangula in
Canada. Phytopathology 39(2): 142-148.

Redak, R. A. 2000. Arthropods and multispecies habitat conservation plans: Are we
missing something? Environmental Management 26:897-8107.

Reed, C. C. 1997. Responses of prairie insects and other arthropods to prescription burns.
Natural Areas Journal 17(4):380-385.

Reid, W. V., H. A. Mooney, A. Cropper, D. Capistrano, S. R. Carpenter, K. Chopra, P.
Dasgupta, T. Dietz, A. K. Duraiappah, R. Hassan, R. Kasperson, R. Leemans, R. M.
May, T. McMichael, P. Pingall, C. Samper, R. Scholes, R. Watson, A. Zakri, Z. Shidong,
N. Ash, E. Bennett, P. Kumar, M. Lee, C. Raudsepp-Heame, H. Sirnons, J. Thonell, and
M. B. Zurek. 2005. Ecosystems and human well-being: synthesis. Island Press,
Washington, DC.

37

Reinartz, J. A. 1997. Controlling glossy buckthom (Rhamnus frangula L) with winter
herbicide treatments of cut stumps. Natural Areas Journal 17(1):38-41.

Rickey, M. A., and R. C. Anderson. 2004. Effects of nitrogen addition on the invasive
grass Phragrnites australis and a native competitor Spartina pectinata. Journal of Applied
Ecology 41 (5):888-896.

Rydin, H., J. K. J eglurn, and A. Hooijer. 2006. The biology of peatlands. Oxford
University Press, Oxford ; New York. ix, 343 p.

Samways, M. J. 1994. Insect conservation biology. lst edition. Chapman & Hall, London
; New York. xiii, 358 p.

Samways, M. J. 2005. Insect diversity conservation. Cambridge University Press,
Cambridge, UK ; New York, NY. xi, 342 p.

Schwartz, M. W., C. A. Brigham, J. D. Hoeksema, K. G. Lyons, M. H. Mills, and P. J.
van Mantgem. 2000. Linking biodiversity to ecosystem function: implications for
conservation ecology. Oecologia 122(3):297-305.

Sinclair, A., and P. M. Catling. 1999. The value of cutting in the management of glossy
buckthom (Rhamnusfrangula L.). Wetland Journal 11(2):25-27.

Spieles, J. B., P. J. Comer, D. A. Albert, and M. A. Kost. 1999. Natural community
abstract for prairie fen. Michigan Natural Features Inventory, Lansing, MI.

Swengel, A. B. 1998. Effects of management on butterﬂy abundance in tallgrass prairie
and pine barrens. Biological Conservation 83(1):77-89.

Thompson, R., and B. Starzomski. 2007. What does biodiversity actually do? A review
for managers and policy makers. Biodiversity and Conservation 16:1359-1378.

van Diggelen, R., B. Middleton, J. Bakker, A. Grootjans, and M. Wassen. 2006. Fens and
ﬂoodplains of the temperate zone: Present status, threats, conservation and restoration.
Applied Vegetation Science 9(2): 157-162.

Van Driesche, R. G., and T. S. J. Bellows. 1996. Biological Control. Chapman and Hall,
New York

Verhoeven, J. T. A., W. Koerselrnan, and A. F. M. Meuleman. 1996. Nitrogen- or
phosphorus-limited growth in herbaceous, wet vegetation: Relations with atmospheric
inputs and management regimes. Trends in Ecology & Evolution 11(12):494-497.

Voss, E. G. 1985. Michigan ﬂora; a guide to the identiﬁcation and occurrence of the
native and naturalized seed-plants of the State, Part II: Dicots (Saururaceae-Comaceae).
Cranbrook Institute of Science. University of Michigan Herbarium, Bloomﬁeld Hills, MI.

3v.

38

Walsh, D. G. F., S. Waldren, and J. R. Martin. 2003. Monitoring seed viability of ﬁfteen
species after storage in the Irish threatened plant genebank. Biology and Environment:
Proceedings of the Royal Irish Academy lO3B(2):59-67.

Webster, C. R., M. A. Jenkins, and 8. Jose. 2006. Woody invaders and the challenges
they pose to forest ecosystems in the eastern United States. Journal of Forestry
104(7):366-374.

Weltzin, J. F., J. K. Keller, S. D. Bridgham, J. Pastor, P. B. Allen, and J. Chen. 2005.
Litter controls plant community composition in a northern fen. Oikos 110(3):537-546.

Williams, G. W. 2000a. Introduction to Aboriginal ﬁre use in North America. Fire
Management Today 60(3):8-12.

Williams, G. W. 2000b. Reintroducing Indian-type ﬁre: implications for land managers.
Fire Management Today 60(3):40-48.

Wohl, D. L., S. Arora, and J. R. Gladstone. 2004. Functional redundancy supports
biodiversity and ecosystem function in a closed and constant environment. Ecology
85(6): 1 534-1540.

Woo, L, and J. B. Zedler. 2002. Can nutrients alone shift a sedge meadow towards
dominance by the invasive T ypha x glauca? Wetlands 22(3):509-521.

.39

Chapter 2

For Natural Areas Journal

BIOTIC AND ABIOTIC CONDITIONS IN MICHIGAN PRAIRIE FEN INVADED BY
GLOSSY BUCKTHORN (FRANGULA ALNUS)

Anna Katherine Fiedler

40

Abstract

Michigan prairie fens are unique in their high plant and animal diversity, but
many are currently threatened by invasion of exotic glossy buckthom (F rangula alnus).
Land managers and conservationists have observed a variety of qualitative changes in
prairie fen after F. alnus invasion, but the relationship between its presence and biotic
and abiotic factors has not been quantiﬁed. Here, we compare a set of abiotic factors and
the plant community in areas of a Michigan prairie fen invaded by buckthom with that of
uninvaded sites. Many of the conditions typical of fens were different in buckthom-
invaded plots. We found signiﬁcantly lower soil pH, fewer vegetative hmnmocks, less
light availability, less total plant cover, and lower graminoid relative abundance in
invaded versus uninvaded areas. We also examined the buckthom invasion process by
looking at differences from the upland edge to the water’s edge of invaded plots. In F.
alnus-invaded areas, we found signiﬁcantly greater percent soil organic matter, fewer
vegetative hummocks/ m2 and lower mean coefﬁcient of plant conservatism ﬁom upland
edge toward the water. Our ﬁndings indicate a variety of signiﬁcant differences in F.
alnus-invaded areas of prairie fen, which will be useful to wetland managers and
conservationists. These differences provide a baseline against which restoration success

can be gauged.

Keywords: prairie fen, glossy buckthom, F rangula alnus, restoration, invasive species.

41

Introduction

The retreat of the most recent glaciation from what is now the northern U.S.
approximately 12,000 years ago created a set of abiotic conditions that paved the way for
the development of distinctive plant communities (Albert 1995). One of these is prairie
fen, a unique groundwater fed wetland habitat found in the Midwestern U.S. Prairie fens,
typically less than 5 ha in size (Bedford and Godwin 2003), support very high
biodiversity within areas that comprise less than 0.1% of the glaciated landscape in the
Midwestern U.S. (Nekola 1994). Prairie fen is one of the rarest wetland types in the US,
occurring most commonly in the glaciated Midwest (Spieles et al. 1999, Amon et al.
2002). This combination of small size and high species richness makes prairie fens
biodiversity hotspots within these landscapes; however, this wetland community and its
associated biodiversity are increasingly imperiled.

There are a number of threats to prairie fen habitats, all of which are either a
direct or indirect result of human impacts on the landscape. These include habitat
fragmentation and conversion (Grootjans et al. 2006, van Diggelen et al. 2006), changes
in hydrology (Amon et al. 2002), invasive species (Bowles and Jones 2006), changes in
nutrient inputs (V erhoeven et al. 1996), increased pollution (Panno et al. 1999), and
changes in the fen disturbance regime (Middleton et al. 2006). In many cases these
factors interact to impact prairie fen communities. To minimize the effects of these
threats and ensure their persistence, many prairie fens will need to be actively managed.
Here, we provide information that will help land managers considering glossy buckthom

removal in'prairie fens.

42

Glossy buckthom

Glossy buckthom (Rhamnaceae: F rangula alnus P. Mill. (previously Rhamnus
frangula L.)) is a shrub of Eurasian origin that is especially aggressive in bogs and fens
(Galatowitsch et a1. 1999). It was introduced in North America, likely before 1800, and
was cultivated for forestry uses, wildlife habitat, and as an ornamental hedge plant. It did
not become naturalized and exhibit landscape spread until the early 1900’s (Converse
1984). F rangula alnus is currently widespread in North America, occuning from Nova
Scotia west to Manitoba in Canada and in the Eastern US. as far south as Kentucky
(Sinclair and Catling 1999, USDA 2008).

F rangula alnus tolerates a range of soil and moisture conditions. It establishes
well in acid to alkaline peat soils with some moisture, although it is does not survive
permanent waterlogging (Voss 1985). This species blooms proliﬁcally and produces
abundant ﬁ'uits, each containing 2-3 seeds, annually (Godwin 1943). A variety of bird
species eat F. alnus berries and disperse seeds, and seedlings establish best in areas with
exposed soil and available light (Converse 1984). Frangula alnus is most common in
fens, bogs, disturbed damp to wet areas, lake shores, and thickets along rivers (V 033

1985), although it will survive in forests, as well.

Effects on habitat

F rangula alnus may be altering its environment, making it better able to persist in
areas where it is established. First, there is evidence that F. alnus displaces native plant
species in its invasive range. In an Ontario wetland, Sinclair and Catling (1999)
compared plant growth in plots with F. alnus removed versus beneath F. alnus cover.
They found both greater herbaceous percent cover and species richness in removal plots.

43

In addition, Houlahan and Findlay (2004) examined wetlands in Ontario and found that,
along with certain natives such as Salix petiolaris, exotic plant species, including F.
alnus, became dominant species in these wetland systems. F rangula alnus may also alter
a variety of abiotic factors in prairie fen habitats. The organic matter hummocks
characteristic of prairie fen communities provide a wet but not inundated microhabitat
(Peach and Zedler 2006) that may be conducive to initial F. alnus seedling establishment,
leading to subsequent breakdown of those hummocks. In addition, F. alnus may alter
nitrogen cycling in prairie fen systems. Many invasive species are known to alter nitrogen
ﬁxation rates and increase nitrogen availability (Ehrenfeld 2003) and F. alnus has been
associated with elevated soil nitrogen in upland forests (Huebner et al. 2009). In sum, a
suite of characteristics may differ between areas with and without F. alnus invasion, and
characterizing these differences is the ﬁrst step to understanding invasion.

Land managers concerned with the conservation of prairie fen have observed a
variety of changes in prairie fen upon invasion of F. alnus. However, the relationship
between F. alnus presence and these differences in prairie fen has not been quantitatively
measured. In addition, baseline data are rarely gathered before restoration, but are crucial
to gauging its effectiveness. Therefore, the goal of this study was to quantify the
differences in biotic and abiotic factors between invaded and uninvaded prairie fen and
examine the importance of glossy buckthom invasion. Our hypothesis was that biotic and
abiotic characteristics would be different in invaded prairie fen, with lower pH and light
availability, greater nutrient availability, fewer hummocks, lower herbaceous plant
diversity, lower percent cover of herbaceous species, decreased relative abundance of

grarninoids, and an increased relative abundance of exotic species. A common

44

observation is that F. alnus invasions often progress from the drier upland edges of fens
towards wetter lowland areas. Therefore, we hypothesized that fens undergoing
buckthom invasion would be the least similar to uninvaded conditions near the upland
edge. We anticipated our results would provide information on the potential impacts of
glossy buckthom invasion on wetlands. This information can guide management and

restoration efforts to maximize their effects.

Methods

Experimental design

The study took place in Jackson County, Michigan at the Michigan State
University MacCready Reserve in Clarklake, Liberty Township. In a single fen, we
delineated ten 25 x 25 m replicate plots containing mature glossy buckthom plants
(invaded), as well as two 25 x 25 m Lurinvaded plots located within 300 m of
experimental plots (Figure 2A). Nine 1 m2 sampling quadrats were established in each
plot in Fall 2007. Due to possible effects of slope and position on plant communities and
abiotic factors, three transects were established in each plot: 3 m from the upland edge
(upland), 3 m from the water edge (water), and in the center of the treatment (center),
equidistant to both edges (Figure 28). All replicates had a distinct upland edge. However,
replicates on the East side of the pond lacked a distinct water edge. In these plots the
transects began at the upland edge (southwest) and ran down slope (northwest). Each
transect contained three replicate quadrats. To further characterize this gradient, we
measured the difference in elevation between the center of the “upland” quadrats and
center of “water” quadrats (Figure 2C).

45

Sampling

To assess differences in soil pH and nutrients between buckthom invaded and
uninvaded fen, soil nutrients, organic matter, and pH were measured from soil cores
taken at 0-10cm depth. In September—October 2007, two soil samples were taken within
40 cm of each sampling quadrat. Samples ﬁ'om each transect within a plot replicate were
homogenized. Samples were sent to the Michigan State University soils lab for
measurement of pH (water pH method), percent organic matter (loss on ignition at
500°C), nitrate and ammonia (KCl extraction), phosphorous (Bray P-l or Olsen in
samples with free calcium carbonate) and potassium (neutral 1 normal ammonium acetate
procedure).

To assess differences in light availability, microhabitat, and plant community
between buckthom invaded and uninvaded fen, a set of measurements was taken in all
nine 1m2 quadrats in each replicate and treatment in September 2007. Light availability
was measured on a clear, sunny day with a l? m long photosynthetically active radiation
(PAR) ceptometer (Sunﬂeck, Decagon Devices Inc., Pullman, WA) at 100 cm height
within each invaded and uninvaded plot. Soil rnicrotopography was measured at the same
location by counting the number of vegetation-based hurnmocks within the 1 m2 quadrat.
In the same 1 m2 quadrat, we estimated total and species percent cover of forbs, grass,
and shrubs less than 1.5 m tall (estimating percent cover including the vertical projection
of all vegetation rooted within the quadrat). Mature F rangula alnus shrubs were not
included in the vegetation analysis. To measure the gradient in soil elevation between
upland and water transects within replicates, we used an electronic self-leveling single-

beam rotary laser (LMH-C series, CST/Berger, Watseka, IL) on June 15 and 16, 2009.

46

Statistical analysis

To determine the relative abundance of different plant growth forms per plot, we
divided the proportion of graminoid and exotic plant cover, respectively, by total plant
cover. These proportion data were arcsin square root transformed for analyses; light
availability was log (x+1) transformed to address variance heterogeneity. We calculated
diversity indices for within-plot species cover with two metrics: Simpsons diversity
index, D (using percent cover) (Magurran 2004) and the Floristic Quality Index (F OD. To
make the Simpsons index more intuitive, it is expressed here as l-D, so that the value of
D increases with increasing diversity. Floristic Quality Assessment is performed on a
plant species list using the coefﬁcient of conservatism (C) assigned to each plant species
found in Michigan (Herman et al. 2001). As C increases, the probability that a species is
likely to occur in intact, uninvaded habitat increases.

To characterize differences between buckthom invaded and uninvaded areas and
examine differences between transects, we performed a 2-way ANOVA on all measured
characteristics with treatment (buckthom-invaded or uninvaded), transect (upland, center,
and water), and the interaction between them, including subsarnples within transect (SAS
Institute 2003, PROC MIXED). Number of live hummocks was non-normally distributed
so a poisson distribution with a log link was used (SAS Institute 2003, PROC
GLIMMIX). Because transects represent blocks within each replicate, we used a split-
plot design with subsampling, with replicate and replicate by treatment as random factors.
Due the uneven number of replicates in each treatment, we used Satterthwaite adjusted
degrees of ﬁ'eedom (Satterthwaite 1946). Bonferroni adjusted p-values for multiple

comparisons are used for transect comparisons.

47

Results

Differences between invaded and uninvaded plots

Buckthom-invaded plots differed from uninvaded plots with several key abiotic
features altered in buckthom-invaded habitats. Soil pH, number of vegetative hummocks
and light availability were signiﬁcantly lower in invaded areas (Table 2, Figure 3 a, c, d).
There was a trend (p=0.090) toward higher organic matter content in invaded than
uninvaded areas (Table 2, Figure 3b). There were no signiﬁcant differences in soil nitrate,
ammonia, phosphorous, or potassium concentrations, although all were numerically
greater in buckthom invaded than uninvaded plots. The transect by treatment interaction
was not signiﬁcant in any of the statistical comparisons (p>0.15).

Buckthom—invaded plots supported fewer plants characteristic of fens. The mean
plant coefﬁcient of conservatism, total plant cover, and graminoid relative abundance
were all signiﬁcantly lower in buckthom-invaded than uninvaded plots (Table 2, Figure 3
e, f). Although there were no signiﬁcant differences in plant diversity (Simpsons D) or
exotic plant relative abundance between invaded and uninvaded areas, there again were
numeric differences in the direction we predicted, with lower diversity, greater
dominance and greater exotic cover in buckthom invaded plots (Table 2).

There were marked differences in plant community between buckthom invaded
and uninvaded areas. Shade-tolerant woodland species, including Lindera benzoin (L.)
Blume, Rubus strigosus Michx., Solanum dulcamara L., Rubus pubescens Raﬁ,
Equisetum arvense (L.), Viola nephrophylla Greene, and Circaea lutetiana L., were
absent from uninvaded plots but present in at least 10% of buckthom invaded plots
(Table 3). In contrast, light-loving Pedicularis canadensis L., Muhlenbergia glomerata

48

(Willd.) Trin., Scirpus acutus Muhl. ex Bigelow, Gentianopsis crinita (Froel.) Ma,
Rudbeckia hirta L., Solidago ohioensis Frank ex Riddell, Cladium mariscoides (Muhl.)
Torn, Sorghastrum nutans (L.) Nash, and Potentillaﬁ'uticosa L. were present in 17-50%
of uninvaded plots but absent from all buckthom-invaded plots (Table 3). Several plant
species characteristic of prairie fen: Cirsium muticum Michx., Aster lanceolatus Willd.,
Eupatorium macalatum L., T helypteris palustris Schott, and Carex stricta Lam. occurred

in 44-100% of uninvaded plots but only 2-24% of buckthom-invaded plots.

Differences with distance from water’s edge

We observed a gradient of increasing graminoid cover and light availability from
the upland edge to the water’s edge of invaded plots. This plant community gradient
coincided with a soil elevational gradient of 16-82 cm from upland plots to those near the
water’s edge for the replicates on the western side of the lake, and differences of 0-28 cm
on the eastern side of the lake (Figure 2c).

Our analysis of habitat factors from the upland to water’s edge in invaded areas
yielded several interesting patterns in nutrient availability and soil characteristics. There
were greater nitrate levels in the center (20.3 i 4.0 ppm) than water (9.8 :l: 3.7 ppm) and
upland (8.6 i 1.1 ppm) transects, while ammonia levels were signiﬁcantly greater in the
water (17.6 d: 1.3 ppm) and center (14.4 :t 2.0 ppm) transects than the upland (9.7 i 1.3
ppm). Potassium levels, on the other hand, were greater in the water (1 12.8 d: 11.4 ppm)
than center (72.0 :t 8.5 ppm) and upland (70.9 i 9.6 ppm) transects. There was a
signiﬁcant increase in percent soil organic matter between upland (50.3 i 7.9) versus
center (67.5 i 6.0) and water (71.5 i 4.4) in buckthom-invaded areas, and a similar
pattern in uninvaded plots (Table 4, Figure 4a). The number of vegetative hummocks/ m2

49

was signiﬁcantly greater near the water (3.4 d: 0.4) than in the center (0.97 :l: 0.27) or
upland transects (0.37 i 0.18) (Figure 4b). There were no signiﬁcant differences in
phosphorous levels, soil pH, or light availability between transect (Table 4). There was
the same general trend in number of hammocks in uninvaded areas, but there was a
smaller magnitude of difference in hummock density between transects in the uninvaded
plots.

Plant community measures also differed by transect within invaded areas, with
greater similarity to fen closer to the water’s edge. The mean coefﬁcient of conservatism
in buckthom-invaded areas was signiﬁcantly greater near the water (4.8 d: 0.2) and center
(4.5 :t 0.2) than in the upland transect (3.0 :l: 0.2) (Table 4, Figure 4c). Again, the
differences in plant conservatism for uninvaded plots between blocks were smaller. The
relative abundance of graminoids in buckthom-invaded areas was signiﬁcantly greater in
quadrats near the water (0.25 i 0.04) than the center (0.05 a: 0.02) and upland (0.02 i
0.01) areas of the plots (Table 4, Figure 4d), while abundance in uninvaded blocks was
much greater overall with less distinct differences between blocks. There were no

signiﬁcant differences in Simpsons D or relative exotic cover by transect (Table 4).

Discussion
In support of our hypotheses (Table 2), we found a number of signiﬁcant
differences in both abiotic and biotic factors between prairie fen invaded by F. alnus and
nearby uninvaded areas. The differences in uninvaded fen versus buckthom-invaded
areas included signiﬁcantly greater light availability, higher soil pH, more hummocks,
greater plant conservatism, greater percent plant cover, and greater relative graminoid

abundance. The primary exception to our predictions was in organic matter content,

50

which we predicted would be lower in invaded areas due to increased decomposition
rates. We found a trend toward greater organic matter content in the 0-10 cm soil layer in
invaded than uninvaded areas. This either indicates that invasion inﬂuences organic
matter or that F. alnus is better able to invade areas of fen with high organic matter rather
than marl-based soils that were more prevalent in the uninvaded areas.

Because we did not follow the entire invasion process in this study we do not
know whether F. alnus was able to invade areas with different characteristics or if it
altered them after its arrival. The study was performed at one buckthom-invaded site, so
we cannot make inferences about patterns between sites. The history of F. alnus invasion
is not well-documented for this site but we do have access to color photos, likely dating
to the 1930’s, which show the currently F. alnus invaded area with sedge-dominated
vegetation and very little shrub growth. We also have a general report from 2002
indicating “minor cover” of F. alnus in the fen area. Because shrub-dominated growth in
our site was not the historic state and we see buckthom in areas with peat soils near
uplands but not near the water, here we use plot proximity to water as a proxy for
understanding the invasion process. While these are critical factors to keep in mind, this
study documents a number of striking contrasts between buckthom-invaded and
uninvaded fen which haven not been quantiﬁed before. In addition, some patterns, such
as differences in herbaceous vegetation and microhabitat, were unlikely to exist pre-
invasion.

In addition to differences between buckthom-invaded and uninvaded areas of a
prairie fen, we also found differences in a number of factors between plots nearest the

upland edge and water’s edge. These differences allow us to examine the possible

51

changes in fen with buckthom invasion. Within buckthom invaded replicates, there were
areas nearest the water’s edge where buckthom had not established. These areas are
likely too wet to allow buckthom establishment. In the transect nearest the water’s edge
we found signiﬁcantly greater potassium levels, hummock abundance, total herbaceous
cover, and graminoid relative abundance. A second group of factors, the mean coefﬁcient
of conservatism, percent organic matter, and soil ammonia, were all greater near the
water and in the center of replicates than near the upland edge. These differences likely
represent, in part, a transition to upland soil characteristics and an upland plant

community along the upland fen edge.

Nutrient Availability

Nutrient availability in prairie fen is low, a factor that contributes to the high plant
species diversity (Boeye et al. 1997, Nekola 2004, Bowles et al. 2005). In our study, we
did not ﬁnd signiﬁcant differences in soil N, P, or K between invaded and uninvaded fen,
although N03, NH4, P, and K levels were all numerically higher in invaded areas. There
is evidence that F. alnus alters nutrient properties in white-pine dominated uplands
(Fagan and Peart 2004) and has been associated with higher nitrogen levels in hardwood
dominated uplands (Huebner et al. 2009), but we found no published data on its effect on
nutrients in wetlands. Increasing eutrophication of wetlands in North America may lead
to decreased species richness within wetlands (Bedford et al. 1999). Increased nutrient
availability as a result of F. alnus invasion is more likely in this case than larger scale
eutrophication leading to decreased species richness, since both uninvaded areas are
within 25 m of invaded areas. In our on-going project we are studying differences and
changes in soil nutrients after removal of F. alnus to examine legacy effects of its

52

invasion on soil properties. This will allow us to examine whether F. alnus was able to
invade areas with greater soil nutrient availability, lower pH and greater organic matter

content, or whether it created these changes after invasion and establishment.

Plant community

We found marked differences in the herbaceous plant community of invaded and
uninvaded plots. Speciﬁcally, there was signiﬁcantly more total cover, a signiﬁcantly
greater mean coefﬁcient of conservatism, and higher proportion of graminoid cover in
uninvaded than buckthom-invaded areas. These patterns ﬁt with the sedge and grass-
dominated prairie fen, and the higher C indicates a higher quality plant community in
uninvaded areas. Common fen species C. stricta, II palustris and E. maculatum (Amon et
al. 2002) were markedly more abundant in uninvaded areas. Surprisingly, there was no
difference in the proportion of exotic herbaceous cover between invaded and uninvaded
areas. This indicates that when F. alnus is present, it is either preventing establishment or
suppressing the presence of smaller non-natives, including F. alnus seedlings.

F rangula alnus has been found to have signiﬁcant negative effects on native
herbaceous cover and diversity in pine forests (Frappier et al. 2003). However, we found
no signiﬁcant differences in plant diversity or evenness between buckthom-invaded and
uninvaded herbaceous communities. This result is similar to that of Houlahan and
Findlay (2004), who found that invasive species were not more likely to be dominant
plants than native wetland species. This suggests that F. alnus is likely outcompeting
other exotic species, whereas in uninvaded fen, sedges are primarily ﬁlling this niche.
Similarly, Mills et al. (2009) found that at larger spatial scales F. alnus was not
associated with decreased herbaceous diversity in a bog community over 15 years, and

53

they found no effect on herbaceous cover of invasion. This may be due to differences in
scale and a study site with high initial high shrub cover, whereas in our study the

uninvaded plots had low shrub cover.

Patterns in Frangula alnus invasion

Previous reports have observed that F. alnus is abundant around the perimeter of
prairie fen (McCormac and Schneider 1994) and land managers in Michigan frequently
report that their fens are ‘ringed’ by F. alnus. In contrast, areas within fens that are
nearest to groundwater inﬂows appear more resistant to invasion by F. alnus. Our
examination of biotic and abiotic gradients from the upland toward the water’s edge
suggests a potential mechanism for such invasion.

F rangula alnus is known to require water during the entire growing season but
does not tolerate long periods of inundation. In water-saturated peat soils the root system
of F. alnus grows very near the surface, above the water table (Godwin 1943). Most fens
have a slight to medium slope that permits water drainage, resulting in habitat that is wet
but not inundated (Amon et al. 2005). Frequently the drier areas of fens occur near
upland edges. For example, the lake-edge fen in our study slopes gradually from a
forested upland edge towards the lake (Figure 2). Overhanging tree limbs provide
perches from which birds deposit F. alnus seeds directly into the fen edge, and the
generally drier conditions near the perimeter favor F. alnus seedling establishment. We
have also observed that F. alnus seedlings frequently take root on the top of Carex spp.
hummocks. Such hummocks have been shown to provide a range of moisture availability
to native plants (Peach and Zedler 2006) and may provide a moist but not inundated

microhabitat conducive to F. alnus seedling establishment.

54

Once established, F. alnus likely facilitates its own further invasion. Transpiration
of F. alnus in midsummer has been shown to lower the water table in wetlands (Godwin
1943). This would make the portion of fen near F. alnus infestations less suitable for
native species that require wetter conditions, but increasingly suitable for its own
invasion. Finally, with a lowered water table under the mature F. alnus, decomposition
may be accelerated, leading to breakdown of hummocks and reducing microhabitat
differences in the fen. Our data support this pattern of hummock degradation, with
signiﬁcantly fewer hummocks found in F. alnus invaded areas.

The patterns we found in this Michigan fen are similar to those that land managers
have commonly observed. We have now quantitatively established this set of both abiotic
and biotic factors in F. alnus invaded and uninvaded fen before restoration activities
occur at the site. Several key differences between invaded and uninvaded prairie fen,
notably the lack of hummocks, lower plant conservatism, and lower graminoid cover in
invaded fen, were unlikely to be present pre-invasion. These ﬁndings coupled with those
from our on-going restoration efforts at this site will be useful to land managers to
determine whether the effects of F. alnus are mitigated during restoration. F rangula
alnus may not only be effectively invading certain areas of fen, it also appears to be
facilitating its persistence via microhabitat degradation. While relatively rapid recovery is
often seen in prairie fens (e. g. seeding is not a common practice in fen restoration as the
existing seed bank appears to be sufﬁcient), F. alnus may be altering some conditions

over longer periods, leaving a legacy that may increase future fen invasibility.

55

Acknowledgements

We would like to thank G. Kowalewski, J. Curtis, M. Trimmer and J. Vigneron
from MSU forestry, R. Zuniga-Villegas and D. Tison of The Nature Conservancy and the
Landis Lab, especially C. Sebolt, for help with plot establishment. D. Cuthrell and R.
O’Connor provided invaluable input on research direction. A. Reznicek and R. O’Connor
assisted with plant specimen identiﬁcation. R. Ahem and N. Bello provided signiﬁcant
input on data analysis. B. Werling, P. Doran and two anonymous reviewers helped
improve this manuscript. Research funding was provided by the MSU MacCready
Endowment and National Fish and Wildlife Foundation. Funding support for A.K.F. was
provided through a sustainable agriculture predoctoral fellowship by the CS. Mott

foundation and a MSU Barnett Rosenberg predoctoral fellowship in biological sciences.

56

Literature cited

Albert, D. A. 1995. Regional landscape ecosystems of Michigan, Minnesota, and
Wisconsin: a working map and classiﬁcation. Gen. Tech. Rep. NC-178, St. Paul, MN.

Amon, J. P., C. S. Jacobson, and M. L. Shelley. 2005. Construction of fens with and
without hydric soils. Ecological Engineering 24(4):341-357.

Amon, J. P., C. A. Thompson, Q. J. Carpenter, and J. Miner. 2002. Temperate zone fens
of the glaciated Midwestern USA. Wetlands 22(2):301-317.

Bedford, B. L., and K. S. Godwin. 2003. Fens of the United States: distribution,
characteristics, and scientiﬁc connection versus legal isolation. Wetlands 23(3):608-629.

Bedford, B. L., M. R. Walbridge, and A. Aldous. 1999. Patterns in nutrient availability
and plant diversity of temperate North American wetlands. Ecology 80(7):2151-2169.

Boeye, D., B. Verhagen, V. VanHaesebroeck, and R. F. Verheyen. 1997. Nutrient
limitation in species-rich lowland fens. Journal of Vegetation Science 8(3):415-424.

Bowles, M., and M. Jones. 2006. Trends of change in composition and structure of
Chicago region wetland vegetation. Chicago Wilderness J oumal 4(3):25-34.

Bowles, M. L., P. D. Kelsey, and J. L. McBride. 2005. Relationships among
environmental factors, vegetation zones, and species richness in a North American
calcareous prairie fen. Wetlands 25(3):685-696.

Converse, C. K. 1984. Element stewardship abstract for Rhamnus cathartica, Rhamnus
ﬁangula (syn. F rangula alnus). The Nature Conservancy, Arlington, VA.

Ehrenfeld, J. G. 2003. Effects of exotic plant invasions on soil nutrient cycling processes.
Ecosystems 6(6):503-523.

F agan, M. E., and D. R. Peart. 2004. hnpact of the invasive shrub glossy buckthom
(Rhamnusfrangula L.) on juvenile recruitment by canopy trees. Forest Ecology and
Management 194(1-3):95-107.

Frappier, B., R. T. Eckert, and T. D. Lee. 2003. Potential impacts of the invasive exotic
shrub Rhamnus frangula L. (glossy buckthom) on forests of southern New Hampshire.
Northeastern Naturalist 10(3):277-296.

Galatowitsch, S. M., N. 0. Anderson, and P. D. Ascher. 1999. Invasiveness in wetland
plants in temperate North America. Wetlands 19(4):733-755.

Godwin, H. 1943. F rangula alnus Miller (Rhamnusfrangula L.). Journal of Ecology
3 1 :77-92. -

57

 

 

 

 

Grootjans, A. P., E. B. Adema, W. Bleuten, H. Joostn, M. Madaras, and M. Janakova.
2006. Hydrological landscape settings of base-rich fen mires and fen meadows: an
overview. Applied Vegetation Science 9(2):]75-184.

Herman, K. D., L. A. Masters, M. R. Penskar, A. A. Reznicek, G. S. Wilhelm, W. W.
Brodovich, and K. P. Gardiner. 2001. Floristic quality assessment with wetland
categories and exapmles of computer applications for the state of Michigan. Michigan
DNR, Wildlife Division, Natural Heritage Program, Lansing, MI.

Houlahan, J. E., and C. S. Findlay. 2004. Effect of invasive plant species on temperate
wetland plant diversity. Conservation Biology 18(4):]132-1138.

Huebner, C. D., R. S. Morin, A. Zurbriggen, R. L. White, A. Moore, and D. Twardus.
2009. Patterns of exotic plant invasions in Pennsylvania's Allegheny National Forest
using intensive Forest Inventory and Analysis plots. Forest Ecology and Management
257(1):258-270.

Magurran, A. E. 2004. Measuring biological diversity. Blackwell Publishers, Malden,
MA. viii, 256 p.

McCormac, J. S., and G. J. Schneider. 1994. Floristic diversity of a disturbed western
Ohio fen. Rhodora 96(888):327-353.

Middleton, B. A., B. Holsten, and R. van Diggelen. 2006. Biodiversity management of
fens and fen meadows by grazing, cutting and burning. Applied Vegetation Science
9(2):307-316.

Mills, J. E., J. A. Reinartz, G. A. Meyer, and E. B. Young. 2009. Exotic shrub invasion in
an undisturbed wetland has little community-level effect over a 15-year period.
Biological Invasions 1 1(8): 1 803-1 820.

Nekola, J. C. 1994. The environment and vascular ﬂora of northeastern Iowa fen
communities. Rhodora 96(886): 121-169.

Nekola, J. C. 2004. Vascular plant compositional gradients within and between Iowa
fens. Journal of Vegetation Science 15(6):771-780.

Panno, S. V., V. A. Nuzzo, K. Cartwright, B. R. Hensel, and I. G. Krapac. 1999. hnpact
of urban development on the chemical composition of ground water in a fen-wetland
complex. Wetlands 19(1):236-245.

Peach, M., and J. B. Zedler. 2006. How tussocks structure sedge meadow vegetation.
Wetlands 26(2):322-335.

SAS Institute. 2003. SAS/STAT User's guide for personal computers. SAS Institute,
Cary, NC.

58

Satterthwaite, F. E. 1946. An approximate distribution of estimates of variance
components. Biometrics Bulletin 2:1 10-114.

Sinclair, A., and P. M. Catling. 1999. The value of cutting in the management of glossy
buckthom (Rhamnusfrangula L.). Wetland Journal 11(2):25-27.

Spieles, J. B., P. J. Comer, D. A. Albert, and M. A. Kost. 1999. Natural community
abstract for prairie fen. Michigan Natural Features Inventory, Lansing, MI.

van Diggelen, R., B. Middleton, J. Bakker, A. Grootjans, and M. Wassen. 2006. Fens and
ﬂoodplains of the temperate zone: Present status, threats, conservation and restoration.
Applied Vegetation Science 9(2): 157-162.

Verhoeven, J. T. A., W. Koerselman, and A. F. M. Meuleman. 1996. Nitrogen- or
phosphorus-limited growth in herbaceous, wet vegetation: Relations with atmospheric
inputs and management regimes. Trends in Ecology & Evolution 11(12):494-497.

Voss, E. G. 1985. Michigan ﬂora; a guide to the identiﬁcation and occurrence of the
native and naturalized seed-plants of the State, Part II: Dicots (Saururaceae-Comaceae).
Cranbrook Institute of Science. University of Michigan Herbarium, Bloomﬁeld Hills, MI.
3 v.

59

F rangula alnus invaded and uninvaded reference areas of prairie fen. A result in

Table 2. Hypothesized and actual differences in abiotic and biotic factors between
parentheses indicates a marginally signiﬁcant result.

032050 <>Oz< no“. 000: $00 008.82g 05020 goo. 90:3
.9 2 F Co 0.00» 0 co Emzmtomcoo ho E06580 .
0620.5 <>Oz< no. now: 900 005.805.: Capone...”
2300.. E00536 2.05905 0 0.0065 0000£c0._0n +

.ﬂoa 00095:: .m> 00002: :_ C0595" + .026." - v 53026 006605 0.: 02006:. a

 

 

vod H Ed Nod H Que No: Nd Sﬁo m2 + .050 0:98 030.01

mod a $6 mod n 3.0 o: mdm Sodv - - L0>mo EoEEmb 025.01

no.0 « 5.8 8.. a ~08 2; 3.2 Sodv - - . 98.233 98.85

etc a he Ed a «S. 2.- mm 35.0 - - o :82 £8350

«no a $4 9.0 a 2mm OE 0N onto 02 - a Lease $202350

8.0 « No.5 84 a 5m o: 4.3 50.9. - - Is In 52.2.4. a___nm__m>mzm_._

cod H and omd a ooé o: v.3 wood - - «at 3858:: 200582925.

8.: a 9.8 can a 3.8 9.. on 82. 3 - same 2:85 .x.

8.9 a 2.8 Be a 8.8 o: co Etc 02 + Easy.

84 a te 86 a one one am 83 02 + ESE

man a N3 03 a 3.9 o: 3 83 02 + Acne-:2

m: a 8.0 8d a SS or; mam 03.0 02 + Eaemoz

one a me.» mod « ta 2; 3m Sod - - In new
00002.2: 00002: E9325 “.0 u. n. +£3.00". #530605 .580".

 

60

Table 3. Plant species identiﬁed at research site and the percent of quadrats each species
occupied, 112 total.

 

 

All Unin-
Family Gsnus species Common Name C~ quadrats vaded lnvadsd
Aceraceae Acer rubrum L. red maple 7 2.78 0.00 3.33
Scrophulariaceae Agalinis purpurea (L.) Pennell purple false foxglove 6 5.56 33.33 0.00
Rosaceae Agrimonia gryposepala Wallr. tall hairy agrlmony 2 3.70 11.11 2.22
Amphlcarpaea bracteata (L.) .
Fabaceae Femald Amema" ””3"“ 4 0.93 0.00 1.11
Poaceae Andropogon gerardii Vitman big bluestem 5 0.93 5.56 0.00
Fabaceae Apios americana Medik. groundnut 7 1.85 0.00 2.22
Asclepladaceae Asclepias incamala L. swamp mllkweed 4 2.78 5.56 2.22
Asteraceae Aster lanoeolatus Willd. white panicle aster 3 11.11 50.00 3.33
Asteraceae Aster puniceus L. purplestem aster 8 11.1 1 38.89 5.56
Betulaceae Batu/a pumila (L.) bog birch 10 0.93 0.00 1.11
Asteraceae Bidens frondosa (L.) devil's beggartick 1 4.63 0.00 5.56
Asteraceae Cacalia plantaginea Raf. groovestern Indian plantain 10 0.93 5.56 0.00
Campanulaceae Campanula apannordes Pursh marsh bellﬂower 8 7.41 27.78 3.33
Cyperaceae Carex leplalea Wahlenb. bristlystalked sedge 10 1.85 0.00 2.22
Cyperaceae Carex stipata Muhl. ex Willd. awlfruit sedge 3 13.89 0.00 16.67
Cyperaceae Carex stricta Lam. upright sedge 5 37.04 100.00 24.44
Juglandaceae Carya sp. hickory 5 1.85 0.00 2.22
Celastraceae Celastrus orbiculatus Thumb oriental bittersweet 0 0.93 0.00 1.11
Apiaceae Cicuta maculata L. spotted water hemlock 6 0.93 5.56 0.00
Poaceae Cinna arundinacea L. sweet woodreed 5 1.85 0.00 2.22
Onagraceae Circaea Iutetiana L. enchanter's nightshade 1 24.07 0.00 28.89
Asteraceae Cirsium muticum Michx. swamp thistle 10 11.11 44.44 4.44
Cyperaceae asylum marrscordes (Muhl.) smooth sawgrass 10 6.48 38.89 0.00
Ranunculaceae Clematis virginiana L. western blue virginsbower 4 1.85 0.00 2.22
Comaceae Cornus foemina Mill. stiff dogwood 1 7.41 16.67 5.56
Comaceae Comus stolonifera Michx. western dogwood 6 2.78 16.67 0.00
Desmodium glutinosum (Muhl. ex .
Fabaceae Will d.) Alph. Wood porntedleaf ticktrefoil 5 0.93 0.00 1.11
Dryopteridaceae Exam“ carthus:ana (WI') H'P' spinulose woodfern 8 9.26 0.00 11.11
Onagraceae Epiloblum coloratum Biehler purpleleaf willownerb 3 3.70 11.11 2.22
Equisetaceae Equisetum arvense (L.) ﬁeld horsetail 0 12.96 0.00 15.56
Asteraceae Eupatorium macalatum L. joe-pye weed 4 10.19 50.00 2.22
Asteraceae Eupatorium perfoliatum L. common boneset 4 12.96 16.67 12.22
Asteraceae Euthamia graminifolia (L.) Nutt. ﬂat-top goldentop 4 1.85 11.11 0.00
Rhamnaceae Frangula alnus Mill. glossy buckﬂrom 0 64.81 61.11 65.56
Oleaceae Fraxinus pennsylvanica Marsh. green ash 5 2.78 0.00 3.33
Rubiaceae Gallum boreale L. northern bedstraw 7 7.41 11.11 6.67
Rubiaceae Galium circaezans Michx. licorice bedstraw 10 1.85 0.00 2.22
Rubiaceae Galium triﬂorum Michx. fragrant bedstraw 5 12.96 11.11 13.33
Gentianaceae Gentianopsis crinita (Frost) Ma greater fringed gentian 10 5.56 33.33 0.00
Rosaceae Geum canadense Jacq. white avens 1 23.15 5.56 26.67
Boraginaceae f::;”a wrgmrana (L') W" beggarslice O 1.85 0.00 2.22
Aquifoliaeeae llex verticillata (L.) A. Gray common winterberry 9 0.93 0.00 1.11

 

61

Table 3 cont’d.

 

 

All Unin-
Famlly Genus species Common Name C~ quadrats veded Invaded
Balsaminaceae Impatiens capensis Meerb. jewelweed 3 12.96 33.33 8.89
Iridaceae Iris versicolor L. harlequin blueflag 5 0.93 0.00 1.11
Juncaceae J""°”s mympha’us smallhead web 9 0.93 5.56 0.00
(Engelm.) Buchenau
Fabaceae Lathyrus palustrls L. marsh pea 8 16.67 27.78 14.44
Poaceae Leersia oryzoldes (L.) Sw. rice cut grass 4 6.48 0.00 7.78
Asteraceae Liam's spicata (L.) Willd. dense blazing star 6 0.93 5.56 0.00
Lauraceae Lindera benzoin (L.) Blume northem spicebush 7 7.41 0.00 8.89
Campanulaceae Lobe/la siphilitica L. great blue lobelia 6 3.70 0.00 4.44
Lamiaceae 162:2: amencanus MUhl' ex W. American water horehound 5 3.70 22.22 0.00
Lamiaceae Lycopus uniﬂorus Michx. northem bugleweed 7 19.44 44.44 14.44
Lythraceae Lythrum salicaria (L. ) purple Ioosestrife 0 1.85 11.1 1 0.00
Lamiaceae Mentha arvensis L. wild mint 5 1.85 11.11 0.00
Poaceae ¥::’°”b°’9’a gbmera'a (W'"d') spiked muhly 10 3.70 2222 0.00
Poaceae Muhlenbergia mexicana (L) Trin. Mexican muhly 5 0.93 5.56 0.00
Poaceae Muhlenbergia sylvatica (Torr.) woodland muhly 10 5.56 0.00 6.67
Torr. ex A. Gray
Dryopteridaceae Onoclea sensibilis L. sensitive fern 8 14.81 16.67 14.44
. Osmorhiza claytonii (Michx.) ,
Apiaceae C.B. Clarke Clayton s sweetroot 3 1.85 0.00 2.22
Caprifoliaceae Oxypolis n’gidor (L.) Raf. stiff cowbane 7 0.93 0.00 1.11
Poaceae Panicum dichotomiﬂorum Michx. fall panicgrass 0 0.93 5.56 0.00
Saxifragaceae Pamassla glauca Raf. fen grass of pamassus 10 6.48 38.89 0.00
Waceae ngznoc’ssus q“’"‘7“ef°”a (L) Virginia creeper 2 24.07 16.67 25.56
Scrophulariaceae Pedicularis lanceolata Michx. swamp lousewort 9 2.78 16.67 0.00
Crassuiaceae Penthorum sedoides L. ditch stonecrop 5 0.93 0.00 1.11
Poaceae Phalaris arundinacea L. reed canarygrass 0 12.96 16.67 12.22
Phytoiaccaceae Phytolacca americana L. American Pokeweed 1 0.93 0.00 1.11
Urticaceae Pilea pumila (L) A. Gray cleanlveed 5 68.52 44.44 73.33
Plantaginaceae Plantago major (L.) common plantain 0 0.93 5.56 0.00
Polygonaceae Polygonum amphibium L. water knotweed 4 4.63 16.67 2.22
Polygonaceae Polygonum punctatum Elliot dotted smartwew 6 3.70 0.00 4.44
Polygonaceae Polygonum sagittatum L. arrowleaf tearthumb 8 8.33 11.11 7.78
Polygonaceae Polygonum virginianum L. jumpseed 2 6.48 0.00 7.78
Salicaceae Populus lremuloides Michx. quaking aspen 4 3.70 11.11 2.22
Rosaceae Potentilla frulicosa auct. non L. shrubby cinquefoil 10 11.11 61.11 1.11
Lamiaceae Prune/Ia vulgaris L. selfheal 0 0.93 0.00 1.11
Rosaceae Prunus serotina Ehm. black cherry 1 6.48 0.00 7.78
Fagaceae Quercus sp. oak 5 4.63 0.00 5.56
Rhamnaceae Rhamnus alnifolia L'Hér. alderieaf buckthom 10 0.93 5.56 0.00
Cyperaceae Rhynchospora alba (L.) Vahl white beaksedge 10 0.93 5.56 0.00
Grossulariaceae Ribes cynosbati L. eastern prickly gooseberry 5 4.63 0.00 5.56
Rosaceae Rosa multiﬂora Thunb. multiﬁora rose 0 14.81 5.56 16.67
Rosaceae Rosa palustris Marsh. swamp rose 7 0.93 0.00 1.11
Rosaceae Rubus pubescens Raf. dwarf red blackberry 10 12.04 0.00 14.44
Rosaceae Rubus stn'gosus Michx. grayieaf red raspberry 3 6.48 0.00 7.78
Asteraceae Rudbeckia hirta L. blackeyed Susan 1 5.56 33.33 0.00

 

62

Table 3 cont’d.

 

 

All Unin-
Family Genus species Common Name C~ quadrats veded Invaded
Salicaceae Salix discolor Muhl. pussy willow 2 5.56 33.33 0.00
Caprifoliaceae Sambucus canadensis L. American black elderberry 1 2.78 0.00 3.33
Poaceae (Michx.) Nash little bluestem 5 4.63 27.78 0.00
Cyperaceae Sclrpus acutus Muhl. ex Bigelow hardstem bullrush 6 3.70 22.22 0.00
Lamiaceae Scutellaria galen'culata L.(17) marsh skullcap 5 3.70 0.00 4.44
Asteraceae Senecio aureus L. golden ragwort 7 27.78 11.11 31.11
Asteraceae Senecio pauperculus Michx. balsam groundsel 6 5.56 22.22 2.22
Smilacaceae Smilax tamnoides L. bristly greenbrler 5 6.48 0.00 7.78
Solanaceae Solanum americanum Mill. American black nightshade O 6.48 0.00 7.78
Asteraceae Solidago altissima L. Canada goldenrod 1 2.78 5.56 2.22
Asteraceae 3%:ng °’"°°"S’S Frank 9" Ohio goldenrod 9 5.56 33.33 0.00
Asteraceae Solidago patula Muhl. ex Willd. roundleaf goldenrod 9 31.48 38.89 30.00
Asteraceae 3%?ng nddel,” Frank ex Riddell's goldenrod 7 6.48 38.89 0.00
Asteraceae Solidago rugosa Mill. wrinkleleaf goldenrod 6 25.93 22.22 26.67
Asteraceae Solidago uliginosa Nutt. bog goldenrod 10 0.93 0.00 1.11
Poaceae Sorghast‘rum nutans (L.) Nash indiangrass 5 8.33 50.00 0.00
Araceae sy’l’pmca’pus mm” (L) American skunk cabbage 8 3.70 0.00 4.44
Sallsb. ex Nutt.
Ranunculaceae 1325,5251"? dasy carp um F'SCh' 8‘ purple meadow-rue 5 1.85 0.00 2.22
Thelypten'daceae Thelypleris palustris Schott eastern marsh fern 6 17.59 66.67 7.78
Anacardiaceae Toxrcodendron radicans 0“) eastern poison ivy 2 6.48 11.11 5.56
Kuntze
Liliaceae Trianfha glutinosa (Michx.) Baker sticky toﬂeldia 10 0.93 5.56 0.00
Typhaceae Typha L. cattail 1 0.93 0.00 1.11
Ulmaceae Ulmus sp. elm 3 5.56 0.00 6.67
Poaceae unknown grass 7.41 0.00 8.89
Urticaceae Urtica diolca L. sﬂnging nettle O 7.41 0.00 8.89
Violaceae Viola nephrophylla Greene northem bog violet 9 21.30 0.00 25.56
2
Mean richness I m 14-310-7 7.81:0.3

 

~Coefﬁcient of conservatism from Herman et al. (2001)

63

 

e

W... 520...... 2:005:90 0.0 0:90 v .0 A .0 000.0 2.00000 :0 0508.00... 0
.m 0.020% <>Oz< .2 000: 0.00 008.0.ch 0500.0 .00. 0.033 .
m. .9 o. F .0 0.00m 0 co 50.030200 .0 30.05000
Aw. 0.0205 <>Oz< .0. 000: 0.00 00:53:00 2.330..”
.m 0.

m. M. 09? md. omdA Nd- 8.? ad. ad wz ..050 0:98 02.0.3.

u m 8.? m6- 3.? Nd. 8.? 0.0. RN 03%002, ..0>00 0.00.505 050.0”.

W m omdn «.0 03.: EN- «86 md. omﬁ azﬂﬁag .050 o\a .000... .0>00.c0.n.
.m m «3.0 EN 03.0 tn. 8.? ed. cud a:$0u.0>> <0 c005. 3:033:05
m .m 8.? me 83. 8.0- 8.? N... 8.0 02 o 2.83.0 5.902385
m m 00? N6 3.? od- 9...? md. oNN m2 1... :0 :30 :2 axe 5:50:05 Em...
m .m cm?“ «4. ~86 Qm- 806 0.0. cud asnboag ~E\mv_00EE2. a. 502800.203
m m <8... .0 :5... 0.... 8.? mo- 8.0 898.0; .08... 0.890 as

m m 8.? to Bed B- So... o8- 8.0 nausea; Ea... v.

.w 0.. 82. ca. 8.? 3 8.? No- 8.0 02 Ea... a

m m 3...... Z «8.: 8.... 80¢ a..- 80 806003 Enefz

m Um «8... ma 8.? no. «8... 8.0 om... 80936 88.002

ﬂ m. omdn v.7 omdA n.0,. w..m.o N..- cud m2 In 0:8
m m n. . n. s n. s

m m. 20...: 80...: .280 “E 0:80”. .98”.

T .0> .250 .0> .203 .m> .0.0>>

 

64

    

Upland/lfen edge

 
 

I
I
I

“Upland” “Center” "Water" ater

Figure 2. Layout of research plots at MacCready Reserve, Clarklake, MI. A: Ten 25 x
25m replicates of buckthom invaded plots are shown as outlined rectanng and two
uninvaded areas as white ﬁlled rectangles. B: Quadrats in black lie 3m ﬁ'om the upland
edge (upland). White quadrats lie 3m from the water’s edge (water) and dark grey
quadrats are in the center of plots (center). C: Schematic of typical elevational differences
within plots, arrows showing where elevation was measured.

65

 

 

 

 

 

 

 

 

 

 

 

 

 

 

J F21.9
=o.oo1
7..
:I:
o.
‘3; 6‘
5-1
A
F3.8
60d P=0.077
8
(U
E I
.2 40.
C
(U
9
O
o\°
g 20-
F14.4
6. T P=0.004
N
E
E .
E 4
3
.C
“6 3‘
2*
E 2‘
3
z

 

 

Figure 3. Comparisons of abiotic and biotic factors between F. alnus invaded and

 

 

 

 

 

lnvaded Reference
Treatment

 

 

 

 

 

 

 

 

 

 

 

 

 

100
D F 42.7
P<0.001
so -
C
i 60«
a
E
D.
20 -
9 E F 26.4
at; e . P=0.001
5
8
*5 3.
E
'3
‘3 2‘
c 1 ‘
(U
0
2 0 _
0.6
F F 59.3
P<0.001

.0
(’1
l

P
A
l

Relative graminoid abundance
O 0
i0 0:

.0
—l

 

.0
O
I

'J

 

 

 

 

lnvaded Reference

Treatment

uninvaded plots. Error bars are +SEM, a=0.05. Treatment effects from a 2-way ANOVA
with treatment and transect are shown here.

66

 

Amodvﬁ 80.08.009.000 0:008 m4 000800.80 000000.055 9008: 00000005. 300005an 000

0am 0 5503 800000. 0000000... 5.? 800m .EMm+ 000 80.3 000m .0000 0805 .00 000.000.0000 0003
80800 3000800008 000 ”008.000.0000 00.. 00300.8 000 8020 00005.3: 5 8000000“. .8003 000020. 0330
....~ 5 80800.0 00005 0000 000050 00 030 8.00.03.05.00 05 00 00:88:. .00 000$... 00.... .v 000w...“

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.02m x005
.20; .0200 20...: .202. .980 20...:
i 0.0 - o
l 8 w
m f to . 0 m
d
0 w
New m
w. . N m
0 9.
< . 8 o m w
1..
a a
w. m S
o. v m
. 8.0 m e
A < "-9.
m < m
J! . 8... F0: J1 .8 u
9
[H1 O [01 m 0
0.8 8

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o .o
I
m
m N
.wm .ow%
m m.
o I
.VI} nJu
. .0 .
w m.
w o
w
0082...: H. . 8W m
Jl #l Bagel w m .80
Z
.8
0 < < <

 

 

 

 

 

 

67

Chapter 3

For Ecological Restoration

DIFFERENTIAL SHIFTS IN RESOURCE AVAILABILITY AND PLANT
COMMUNITY STRUCTURE FOLLOWING PRAIRIE FEN RESTORATION

Anna Katherine Fiedler

68

Abstract

Restoration activities frequently involve removal of exotic invasive species, but
subsequent changes in resource availability and the plant community are rarely
quantiﬁed. The exotic invader glossy buckthom (F rangula alnus) is known to alter soil
and plant community conditions in rare prairie fen wetlands; here, we ask whether its
removal leads to shifts in resource availability and plant communities. We examined
availability of light, soil nutrients, and spatial heterogeneity, and the resulting changes in
herbaceous cover and diversity in F. alnus invaded, removal, and uninvaded areas of
prairie fen over two years. We found rapid increases in light availability, herbaceous
plant cover, and relative graminoid abundance in restored plots. There were no signiﬁcant
differences in soil organic matter, nitrate, ammonia, potassium, and phosphorous, plant
diversity or plant species dominance between treatments before or after restoration. Other
factors, such as soil pH, surface spatial heterogeneity, and ﬂoristic quality index, were
initially greater in uninvaded areas versus invaded fen, and remained so in the ﬁrst two
years after restoration. An NMDS analysis considering soil and plant factors
simultaneously showed F. alnus removal plots were on a trajectory toward uninvaded
reference areas by end of year one, with continued shifts towards reference conditions in
year two. Our results show that in prairie fen some soil and plant community factors
change rapidly while others may remain altered for years, potentially leading to
alternative stable states. We conclude that while short-term success can be achieved,

long-term efforts are likely required for successful restoration and invasive management.

Keywords: prairie fen, restoration, F rangula alnus, glossy buckthom, invasive species.

69

Introduction

Non-native invasive species are one of the greatest threats to native communities
in the US. (W ilcove et a1. 1998). Once established, many exotic plant species can alter
ecosystem processes. For example, systems invaded by non-native plants tend to have
greater standing crop biomass, more rapidly decaying litter, shifts in the timing of
nutrient cycling, greater inorganic nitrogen levels, and increased nitrogen mineralization
rates (Ehrenfeld 2003). Wetlands are particularly invasible by non-native plants (Zedler
and Kercher 2005). In the less than 9% of earth’s land area that wetlands occupy (Zedler
and Kercher 2005), they support a number of ecosystem ﬁmctions, including water
ﬁltration and puriﬁcation, ﬂood control, and function as carbon sinks (Baron et al. 2002).
These functions, however, as well as biodiversity, are degraded in many wetland
ecosystems (Zedler and Kercher 2005).

Prairie fens are a globally rare wetland community that is most abundant in the
glaciated Midwestern US. These groundwater-fed systems were always historically
small and isolated (Bedford and Godwin 2003), yet contain very high plant diversity per
unit area (Nekola 1994). Both rare and endangered insects and plants occur in this habitat
type, making its protection a primary concern (Spieles et al. 1999). Like many natural
areas globally, prairie fen habitats are under pressure from a number of threats. These
include changes in nutrient inputs (Verhoeven et al. 1996), increased pollution (Panno et
al. 1999), changes in hydrology (Amon et al. 2002), habitat fragmentation and conversion
(Grootjans et a1. 2006, van Diggelen et al. 2006), alteration of fen disturbance regimes
(Middleton et al. 2006), and invasive species (Bowles and Jones 2006). These factors

often interact to alter the plant and animal communities within prairie fen.

70

The most common invasive species in Midwestern U.S. wetlands include reed
canarygrass (Phalaris arundinaceae L.) (Galatowitsch et al. 1999), purple Ioosestrife
(Lythrum salicaria L.), cattail (T ypha sp.), glossy (F rangula alnus Mill.), and common
buckthom (Rhamnus cathartica L.) (Bowles and Jones 2006). Reed canarygrass and
native cattail (Typha latifolia L.) have both been shown to tolerate a wider range of
moisture conditions than non invasive wetland species (Kercher and Zedler 2004a) and
reed canarygrass tolerates multiple disturbances more readily than native wet prairie
species (Kercher and Zedler 2004b). Additionally, there is evidence that the native-non-
native hybrid cattail Typha x glauca acts as an ecosystem engineer, with greater soil
organic matter, nitrogen, and litter mass and lower plant species richness in densely
invaded areas than in recently invaded areas (Tuchman et al. 2009). Common buckthom
has been associated with a number of changes in soils in upland habitats, including
elevated pH, higher percent nitrogen, and altered nitrogen mineralization rates (Heneghan
et al. 2006). Glossy buckthom (Rhamnaceae: F rangula alnus (previously Rhamnus

frangula L.)) has also been associated with elevated soil nitrogen levels in uplands
(Huebner et al. 2009). Such changes in resource availability may create a legacy of
invasion which needs to be addressed in ecosystem restoration.

Glossy buckthom is a Eurasian shrub that was introduced in North America,
becoming naturalized by the early 1900’s (Converse 1984). It is currently widespread,
occurring in the Northeastern and western plains states of the US, as well as Eastern
Canada (Sinclair and Catling 1999, USDA 2010). Glossy buckthom establishes well in
soils with a wide pH range and is aggressive in fens and bogs (Galatowitsch et al. 1999).

It is most commonly found in wet areas as it requires moist sites, but does not survive

71

permanent waterlogging (V oss 1985). Mature individuals can survive in forested areas
given adequate moisture. F rangula alnus blooms proliﬁcally and produces abundant
ﬁ'uits annually (Godwin 1943). A number of bird species eat the berries, effectively
dispersing seeds, which germinate most successfully on exposed soils with available light
(Converse 1984).

F rangula alnus is considered a key threat to prairie fens due to its ability to
invade and alter resource availability. Once established, it ﬁ'equently becomes an
additional dominant species in prairie fen (Houlahan and Findlay 2004) and alters the
open sedge-dominated structure to one dominated by up to 7m tall shrubs, with shade-
tolerant species in the herbaceous layer. In areas with F. alnus removed, herbaceous plant
growth and species richness are greater than in areas where it remains (Sinclair and
Catling 1999). This pattern is well understood by land managers, and the typical means of
restoration include removal of adult glossy buckthom and subsequent control of reprouts
and seedlings (Converse 1984). In comparison to turinvaded reference areas, invaded
areas of prairie fen differ in plant, soil, and abiotic factors (Chapter 2). However, it is
currently unknown if such features pre-exist, are caused by glossy buckthom invasion, or
a combination of both. Because plant invasions may result in long-term shifts in resource
availability that lead to alternative stable states (Suding et al. 2004), understanding if F.
alnus removal from ecosystems allows a return towards reference conditions is of
signiﬁcant concern.

The goal of this study was to assess the success of habitat restoration by
determining whether restored areas of prairie fen approach uninvaded reference areas in

resource availability and resultant plant community indices. Speciﬁcally, we examined

72

changes in soil resources, light availability, microhabitat, and plant diversity, functional
group, and quality over two growing seasons in buckthom invaded, removal, and
uninvaded areas. Our hypotheses were that 1) soil nutrients would be greatest in invaded
plots, would decrease in removal plots, and be lowest in uninvaded plots, 2) neither soil
pH nor organic matter would shift over time, 3) light availability would increase in
removal plots to a level near that of the uninvaded plots, 4) microhabitat availability
would remain high in uninvaded plots, and low in both removal and invaded plots, 5)
native herbaceous cover, grass and sedge cover, plant species diversity, and plant quality
would be greatest in uninvaded plots, would increase in removal plots, and remain low in
buckthom invaded plots, and 6) exotic invasive plant species cover would be greatest in

invaded plots, followed by removal plots, followed by the uninvaded reference.

Methods

Experimental design

The study was conducted in a F. alnus invaded prairie fen on the Michigan State
University MacCready Reserve, in Clarklake, Liberty Township, Jackson County,
Michigan. In September 2007, we assessed the site for potential restoration of prairie fen
habitats, delineating twelve 25 x 25 m replicate plots containing mature F. alnus plants,
on hydric soils and with plant species characteristic of prairie fen at the invasion margins.
In 2008, we began restoration of six of these plots, following current best management
V practices (The Nature Conservancy, R. Villegas, pers. comm). In February, we cut F.
alnus and treated cut stumps with herbicide, using glyphosate 25% AI (Rodeo, Monsanto,
St. Louis, MO) plus Cide Kick 11 surfactant, adjuvant, and penetrator (Brewer

73

International, Vero Beach, FL). Cut stumps were treated within 15 minutes of cutting,
brush was piled and burned in April-June 2008. Brush piles were located at plot margins,
outside of future sampling areas. In May 2008 and June 2009, we managed invasive
plants in the restored plots in the following ways. Using glyphosate (25% active
ingredient) and Cide Kick 11, we sprayed glossy buckthom re-sprouts and patches of P.
arundinacea L., Rosa muliﬂora Thunb., and Cirsium arvense (L.) Scop. We cut and
treated cut stems of Typha sp. and Populus tremuloides Michx. within restored plots, and
used a propane torch to ﬂame F. alnus seedlings. None of these activities was required
more than once in a growing season within a particular area.

These restoration activities created two treatments among areas that were initially
invaded by F. alnus: 1) invaded, a control where glossy buckthom remained unmanaged,
and 2) restored, where glossy buckthom was removed and invasives were managed
during 2008 and 2009. We also delineated two 25 x 25 m uninvaded reference plots
located within 300 m of experimental plots in Fall 2007, and a third uninvaded plot in
May 2008. We established nine 1 m2 sampling quadrats in each plot, laid out on a grid

across each replicate (Chapter 2).

Sampling

Soil characteristics were measured in each replicate from soil cores (0-10 cm
depth) collected 6-7 Oct. 2007, 29 Sept. 2008, and 14 Sept. 2009. We collected two soil
samples near each sampling quadrat, and sent them to the Michigan State University soils
lab for measurement of pH (water pH method) percent organic matter (loss on ignition at

500°C) phosphorous (Bray P-l or Olsen in samples with free calcium carbonate)

74

potassium (neutral 1 normal ammonium acetate procedure), nitrate, and ammonia (KCl
extraction) (Brown 1998).

To assess changes in plant community, light availability, and microhabitat
between buckthom invaded, restored, and uninvaded fen, we measured the following
parameters in all quadrats in September 2007, 2008, and 2009. We estimated species
percent and total cover of graminoids (including Poaceae, Cyperaceae and J uncaceae),
forbs, and shrubs less than 1.5 m tall including all vegetation rooted within the quadrat.
We use summed cover by species for total herbaceous cover, not including mature F.
alnus shrubs. Light availability at 100cm height was measured on a clear, sunny day with
a 1 m long photosynthetically active radiation (PAR) ceptometer (Sunﬂeck, Decagon
Devices Inc., Pullman, WA), using the average of three measurements from each quadrat.
Soil microtopography was measured in Sept. 2007, June 2008 and June 2009 by counting

the number of vegetative hummocks of 5 cm or more in height within each quadrat.

Statistical analysis

To account for shifts in total cover, we standardized the grass and sedge cover and
exotic cover as proportions of total cover. We calculated Simpsons Diversity Index
D=1/Z(p,.)2 where p, is the proportion of cover of the ith plant species per quadrat (using
VB so that the value increases with diversity) to examine the interaction of species
abundance and evenness, and the Berger Parker index of dominance to examine patterns
of plant species dominance (Magurran 2004). We used coefﬁcients of conservatism (C)
for each plant species (Herman et al. 2001) and species richness (S) within a quadrat to
calculate the ﬂoristic quality index (F QI) (Swink and Wilhelm 1994) per m2 quadrat
using: FQI= Mean C * ‘1 S. Higher values indicate a plant community that is more likely

75

to be found in high quality, undisturbed habitats. To address heterogeneity of variances
the following variables were transformed for ANOVA analyses: light availability, nitrate
levels, and number of hummocks/m2 were log(x+1) transformed, the proportion of
graminoid and exotic cover were arcsin(square root) transformed, and phosphorous and
Simpsons diversity were square root transformed.

To characterize differences in removal versus buckthom invaded and uninvaded
plots over time, we performed a 2-way ANOVA on all of the measured characteristics
with treatment (invaded, removal, and uninvaded), time: pre restoration, year 1, year 2
(Sept. 2007, 2008, and 2009, except for hummocks, which were measured in Sept. 2007,
June 2008, and June 2009), and the interaction between them (SAS Institute 2003, PROC
MIXED). The model included subsamples within treatment and replicate, with each
replicate of a treatment as the experimental unit. We speciﬁed a compound symmetry
covariance structure and used Satterthwaite adjusted degrees of freedom to account for
the uneven number of replicates between reference (n=2-3) and other treatments (n=6)
(Satterthwaite 1946).

To examine shifts in suites of characteristics over time, we used nonmetric
multidimensional scaling (N MDS). NMDS creates an ordination of data quadrats by
ranking each variable relative to the others, adjusting quadrats to minimize the stress
between them on a tWo-dimensional surface. Using untransformed data for all factors, we
calculated a mean by transect, then by row, for each variable, with one value per
treatment and replicate for each date. The analysis included all resource and plant
variables mentioned above, as well as percent bare ground lm2 quadrat, and exchangeable

soil calcium and magnesium (ppm) (neutral 1 normal ammonium acetate procedure).

76

Because factors were on different scales, data for each time were normalized before
performing the NMDS; for each value, the mean is subtracted and divided by the
standard deviation for that variable (Clarke and Gorley 2006a). We used a Euclidian
distance resemblance matrix, which is appropriate for data without zeroes and on similar
scales (Quinn and Keough 2002). The NMDS was performed using 2 and 3 dimensions,
with 25 random starting conﬁgurations and a minimum stress of 0.01 (Clarke and Gorley
2006a). To compare differences by time we performed an analysis of similarity
(ANOSIM) using the resemblance matrix for each time period with 999 random

permutations (Clarke and Gorley 2006b).

Results

Resource availability

We observed shifting patterns in soil characteristics, but never a signiﬁcant
treatment by time interaction (Fig. 5, Table 5). Soil pH was signiﬁcantly greater in
uninvaded plots than invaded plots at the beginning of the study, and did not change
following F. alnus removal (Fig. 5A). There were no signiﬁcant differences by treatment
or time in percent organic matter (Fig. 5B), although levels of both were numerically
lower in the uninvaded reference plots than in removal or invaded replicates on all dates.
Nitrate, ammonia, and potassium levels varied signiﬁcantly by time but not by treatment
(Table 5, Appendix A). There were no signiﬁcant differences by treatment or over time in
phosphorous (Table 5, Appendix A).

Initially, light availability at 1m height was signiﬁcantly greater in uninvaded
replicates than invaded replicates (Fig. 5C, Table 5), but increased after buckthom

77

removal with no signiﬁcant differences between removal and uninvaded references in
year 1 and year 2 after removal. The number of vegetative hurnmocks was signiﬁcantly
greater on all dates in the uninvaded replicates than either the invaded or removal, and
there were signiﬁcant differences by year on this metric, with more in both the invaded

and removal treatments in year 1 and year 2 than pre-restoration (Fig. 5D, Table 5).

Plant community characteristics

We observed varying responses of the plant community to fen restoration. Percent
herbaceous cover was signiﬁcantly greater in uninvaded reference plots than invaded
plots on all dates (Fi g 6A, Table 6). Herbaceous cover increased rapidly in removal plots,
becoming equivalent to reference plots by year 2. The ﬂoristic quality index per m2 was
signiﬁcantly greater on all dates in the uninvaded plots, with no signiﬁcant treatment
differences between invaded and removal treatments (Fig. 63, Table 6). The index was
signiﬁcantly greater in year 1 and 2 than pre-restoration. There were no signiﬁcant
differences in Simpsons diversity index by treatment, but there were signiﬁcant
differences in D by year, with the greatest values in year 2, followed by year 1 and pre-
restoration (Fig. 6C, Table 6). The Berger Parker index of dominance was greatest pre-
restoration, followed by year 1, with signiﬁcantly lower plant dominance in year 2 than
pre-restoration (Fig. 6D, Table 6). The index was numerically lower in uninvaded
reference areas, with the value in removal areas decreasing over time. The proportion of
graminoid (Poaceae, Cyperaceae, and J uncaceae) cover was signiﬁcantly greater in
uninvaded reference than invaded plots before treatment establishment and remained
signiﬁcantly lower in invaded plots (Fig. 6E, Table 6). Graminoid cover increased in

removal plots in years 1 and 2 (Fig. 6E). Finally, proportion of exotic cover was

78

signiﬁcantly greater in invaded and removal plots versus reference areas pre-restoration
and remained signiﬁcantly higher in invaded and removal plots through year 2 (Fig. 6F,
Table 6), although there was a numeric decrease in cover in restored plots.

An NMDS analysis that included all of the resources and plant factors above, as
well as percent bare ground and soil Ca and Mg, revealed a shift in treatments over time.
Before restoration, F. alnus invaded and removal plots were similar to each other and
signiﬁcantly different from reference plots (Fig. 7A, Table 7). In year 1 the removal plots
began to shift toward reference plots, with signiﬁcant differences between all treatments
(Fig. 78, Table 7). In year 2 post-restoration, the level of similarity continued to decrease
between the invaded and removal treatments (Fig 7C). Stress values for the 2-
dimensional plot were under 0.1 in all years, indicating that these ﬁgures accurately

represent the relationship between points.

Discussion

Shifts in resource availability that occurred with habitat degradation may alter the
course of habitat restoration, but exactly how resources shifted with degradation is rarely
known. Incorporating measurements of resource availability with plant community
metrics post-restoration allows for better assessment of potential limiting factors for
restoration, and has potential to alter the course of restoration efforts. In this study, we
found that plant community structure recovered rapidly with removal of invasive F.
alnus, while plant diversity and resource availability did not. Speciﬁcally, total
herbaceous cover and proportion graminoid cover rapidly rebounded to near or above
reference levels in restored plots. In contrast, there was no evidence for change in soil

microhabitat or nutrient availability, with most factors not different initially between

79

invaded and uninvaded plots. We also found no signiﬁcant change in plant diversity;
however, we did observe rapid increases in light availability and plant structural
response. Examination of all factors combined (N MDS) showed that after the ﬁrst
growing season, restored areas were distinct from invaded areas and continued to shift
towards reference conditions in the second growing season, suggesting that restored areas
are on an initial trajectory toward conditions found in the uninvaded reference plots.

Individual metrics, however, may take longer to recover, if at all.

Resource Availability

We measured soil nutrients and microhabitat availability, as well as light
availability, as a result of restoration, but found signiﬁcant differences only in light
availability. This manifested as an immediate shift in light availability at 1m height after
restoration activities. This is not surprising, given that F. alnus provided nearly full shade
in densely invaded areas. This ﬁnding indicates that current prairie fen restoration
practices are effective at restoring this key resource for the herbaceous plant community.
However, many wetlands in North America are also limited by nitrogen or phosphorous
availability (Bedford et al. 1999). There were no initial differences in soil nitrate,
ammonia, potassium, or phosphorous levels between F. alnus invaded and uninvaded
areas, and no changes by treatment emerged in the ﬁrst two years after restoration. This
ﬁnding was contrary to our hypothesis that soil nutrients would differ by treatment over
time. It may be that the groundwater inputs continue to supply these nutrients or that they
simply shift more slowly than our 2-year timeframe. Wetland eutrophication is a
persistent concern in Europe (Bakker and Berendse 1999) and a growing one in the US,
especially near urban areas where greater nitrogen levels are associated with greater non-

80

native plant cover (Matthews et al. 2009b). Agriculture is another source of nutrient
input, either via overland ﬂow or groundwater inﬁltration (Drexler and Bedford 2002).
These potential patterns point to the need for long-term monitoring of soil resources after
wetland restoration.

Results of soil pH and microhabitat support our hypothesis that neither would
change over the timeframe of this study. Soil pH was markedly lower in F. alnus invaded
areas before restoration and remained lower in removal areas in the ﬁrst two years after
F. alnus removal. Either F. alnus is more readily able to invade areas with pH near
neutral or it is causing shifts in alkaline soil pH toward neutral over time. Related species
are known to shift soil pH, although not necessarily in the same direction: invasion of
related R. cathartica has been associated with increases in soil pH in uplands (Heneghan
et al. 2006). We also found signiﬁcantly fewer vegetative hummocks in F. alnus invaded
areas, both before and in the two years following restoration. These hummocks of
undecomposed sedge litter are known to provide suitable microhabitat for plants with a
variety of moisture and light tolerances, and are one mechanism for supporting high plant
diversity in prairie fens (Peach and Zedler 2006). Degraded hummocks may take tens or
even hundreds of years to re-form (Zedler 2000) so this shift is likely to remain in the
long term in areas of prairie fen where mature F. alnus was established. Invasive species
may be drivers leading to change or may enter a system subsequent shifts in it
(MacDougall and Turkington 2005). While we do not have causal evidence for shifts in
pH and microhabitat resulting from F. alnus invasion, we do not have any reason to

believe there are systematic differences between the reference and invaded areas in either

of these factors.

81

Plant Community Structure and Diversity

We found that herbaceous plant cover in restored areas rapidly recovered to
levels similar to those in uninvaded references. The proportion of graminoid cover in
restored areas also appeared to recover in the ﬁrst year following restoration and
decreased in the second year, likely due to a greater increase in perennial forb cover. The
proportion of herbaceous exotic cover was initially signiﬁcantly greater in invaded than
in reference plots and remained greater in removal plots, but numerically decreased over
time. Contrary to our hypothesis, there were no signiﬁcant shifts by treatment in plant
diversity metrics during the ﬁrst two years following restoration. We did observe an
increase in the ﬂoristic quality and Simpsons diversity indices over time and a decrease in
plant dominance; however, these patterns are likely due to repeated sampling, increasing
the likelihood of ﬁnding easily overlooked species. While we did not ﬁnd signiﬁcant
differences by treatment in Simpsons diversity and plant dominance, the ﬂoristic quality
index was signiﬁcantly greater in uninvaded reference plots than either the F. alnus
removal or invaded plots.

Similar patterns in plant cover have been found in other wetland restorations. In
76 constructed and restored wetlands, the goal of minimum percent plant cover was more
frequently met than goals regarding overall plant structure or nativeness (Matthews and
Endress 2008). In contrast to our ﬁndings, the ﬂoristic quality index also increased
rapidly in these restorations, which included seed additions and was calculated on a site-
wide basis rather than in m2 plots (Matthews and Endress 2008). Even mined peatlands
with the surface peat removed have been shown to revegetate rapidly with seed additions
(Cobbaert et al. 2004). Prairie fen restorations focused on invasive species control do not

typically include seed addition. Our study provides evidence that seed addition is
82

unnecessary to regain herbaceous plant cover in this type of habitat. If, however, the
ﬂoristic quality index in F. alnus removal areas doesn’t begin to move on a trajectory
toward that in the uninvaded areas within 5 years of clearing, seeding or planting of key
species might be required. The native species pool in wetlands has been found to be
landscape-limited (Matthews et al. 2009a), pointing to the low likelihood of propagules
entering isolated prairie fen systems without assistance if they do not remain in the
seedbank at a site.

We found evidence that in the methods typically used to restore prairie fens
invaded by non-native F. alnus were successful in the short-term. The long-term success
of prairie fen restorations, however, will depend on a number of factors. Further plant
succession will be determined in part by the longevity and species present in the soil seed
bank, which may be limited at the landscape, rather than local, scale. Additionally, plant
community metrics of wetland restoration success may improve over the short term, then
decrease over the longer term (Matthews et al. 20090). There is evidence for greater seed
production in more disturbed fen communities, with few plant species producing the
majority of seeds (Klimkowska et al. 2009), thereby limiting the number of plant species
that can arise from the seed bank itself.

Through the use of uninvaded references at the same site, we have evidence that a
number of factors are on a trajectory toward that in the reference areas. The use of
references in restoration provides metrics against which restoration success may be
measured (Brinson and Rheinhardt 1996, SER Working Group 2004) but there are
difﬁculties inherent in ﬁnding sites that match the exact conditions expected in a restored

area (White and Walker 1997). This is of particular relevance in prairie fens, which

83

contain regionally variable plant species (Amon et al. 2002) several plant
subcommunities, and the potential for shifts in the entire plant community across a single
fen (Bowles et al. 2005). One of the three references in our site, in particular, differs in
soil characteristics and is characterized as a marl ﬂat subcommunity, while the remaining
two references and the areas invaded by F. alnus are characterized as sedge meadow.
Although the marl ﬂat reference is useful in considering plant characteristics broadly, we
do not expect that F. alnus removal areas will move toward conditions in that reference in

the long term, unless they occur in a speciﬁc area of mineral-rich groundwater upwelling.

Long-term Restoration Success

Invasive species management is a continued concern in many restorations.
Invasives have become increasingly common in wetlands since the 1970’s (Bowles and
Jones 2006). We do not know the exact history of invasion at the study site, but photos
from the 1930’s show it as a sedge-dominated community and F. alnus may have been
present at this site for upwards of 60 years. Propagules for F. alnus remain in the seed
bank, and other invasives, including P. arundinaceae, T ypha sp., and L. salicaria, are all
present at the site. In addition to occurring locally, these species are now part of the
regional species pool and can readily be brought to the site by wildlife and humans. To
prevent re—invasion of non-native species, both the local and regional species pool will
need to be managed (Matthews et al. 2009b) in the long term or expectations for
successful restoration may need to shift to include patches of non-native invasives. North
American prairie fens are historically ﬁre-adapted systems (Curtis 1971) and re-
introduction of periodic, patchy ﬁre (Williams 2000) has been shown to help remove
litter buildup and maintain species diversity in fen communities (Middleton et al. 2006)

84

and other wetlands (Bowles and Jones 2006). Once sedge and grass material have begun
to grow in prairie fens after F. alnus removal, ﬁre is part of the long-term management
scheme for fen restoration, and it helps prevent re-invasion of F. alnus, which is not ﬁre
tolerant.

Despite restoration of the historic disturbance regime, some changes in previously
invaded prairie fen may lead them to alternative stable states (Suding et a1. 2004). We
found evidence that soil structure and pH are altered in prairie fens; each of these factors
is likely to change only over long periods of time, and may make restoration more
difﬁcult (Zedler 2000). Our study reinforces the need to measure multiple metrics across
trophic levels in assessment of restoration success (Ruiz-Jaen and Aide 2005). If we had
measured only plant structure, we would have classiﬁed this restoration as a success, and
would not have had evidence pointing toward longer-term microhabitat and plant
diversity changes in this ecosystem. We have performed top-down removal of invasive
species, but have not addressed possible bottom-up factors that could shift the plant
community in the long term (D'Antonio and Chambers 2006). One solution to nutrient-
based alternative stable states is to put greater inputs into a system for a greater amount of
time, in hopes of shifting community dynamics. However, this approach has had little
evidence of long-term success (Lindig-Cisneros et al. 2003, Corbin and D'Antonio 2004).

Finally, while we found evidence of conditions in restored plots moving in the
direction of those in references, there are a number of considerations in measuring the
long-term success of restoration. Novel approaches to restoration which involve changes
that are not part of the historic disturbance regime may be required to shift the restoration

toward the goal and away from an alternative stable state (F im et al. 2010). In this case,

85

continued site-wide F. alnus removal and management will be required to ensure the
persistence of prairie fen. The research site currently is an ideal area for performing
adaptive restoration research (Hobbs and Harris 2001 , Zedler and Kercher 2005), and
management practices into the future could include novel, lower input methods to support

plant diversity.

Implications for practice
0 Best practice methods used for invasive species removal and control of prairie fen
wetlands were successful in the short-term.
0 Mid— and long-term restoration success will likely be based on continued invasive
control.
0 Rare and special concern plant species may require seeding/planting if the local

species pool cannot provide these sources.

Acknowledgements

We would like to thank G. Kowalewski, J. Curtis, M. Trimmer and J. Vigneron
from MSU forestry, R. Zuniga-Villegas and D. Tison of The Nature Conservancy and the
Landis Lab, especially C. Sebolt, for help with plot establishment. We greatly appreciate
the ﬁeld assistance provided by C. Phillippo and M. Lettow. A. Reznicek and R.
O’Connor assisted with plant specimen identiﬁcation. P. Doran, R. Isaacs, and C.
Malmstrom reviewed and helped improve this manuscript. Research funding was
provided by the MSU MacCready Endowment and National Fish and Wildlife

Foundation. Funding support for A.K.F. was provided through a sustainable agriculture

86

predoctoral fellowship by the CS. Mott foundation and a MSU Barnett Rosenberg

predoctoral fellowship in biological sciences.

87

Literature cited

Amon, J. P., C. A. Thompson, Q. J. Carpenter, and J. Miner. 2002. Temperate zone fens
of the glaciated Midwestern USA. Wetlands 22(2):301-317.

Bakker, J. P., and F. Berendse. 1999. Constraints in the restoration of ecological diversity
in grassland and heathland communities. Trends in Ecology & Evolution
14(2):63-68.

Baron, J. S., N. L. Poff, P. L. Angermeier, C. N. Dahm, P. H. Gleick, N. G. Hairston, R.
B. Jackson, C. A. Johnston, B. D. Richter, and A. D. Steinman. 2002. Meeting
ecological and societal needs for freshwater. Ecological Applications 12(5): 1247-
1260.

Bedford, B. L., and K. S. Godwin. 2003. Fens of the United States: distribution,
characteristics, and scientiﬁc connection versus legal isolation. Wetlands
23(3):608-629.

Bedford, B. L., M. R. Walbridge, and A. Aldous. 1999. Patterns in nutrient availability
and plant diversity of temperate North American wetlands. Ecology 80(7):215 1-
2169.

Bowles, M., and M. Jones. 2006. Trends of change in composition and structure of
Chicago region wetland vegetation. Chicago Wilderness Journal 4(3):25-34.

Bowles, M. L., P. D. Kelsey, and J. L. McBride. 2005. Relationships among
environmental factors, vegetation zones, and species richness in a North
American calcareous prairie fen. Wetlands 25(3):685-696.

Brinson, M. M., and R. Rheinhardt. 1996. The role of reference wetlands in functional
assessment and mitigation. Ecological Applications 6(1):69-76.

Brown, J. R., editor. 1998. Recommended soil test procedures for the North Central
region. Missouri Agricultural Experiment Station

Clarke, K. R., and R. N. Gorley. 2006a. Primer v6. Primer-E, Plymouth, UK.

Clarke, K. R., and R. N. Gorley. 2006b. Primer v6: User Manual/Tutorial. Primer-E,
Plymouth, UK

Cobbaert, D., L. Rochefort, and J. S. Price. 2004. Experimental restoration of a fen plant
community after peat mining. Applied Vegetation Science 7(2):209-220.

Converse, C. K. 1984. Element stewardship abstract for Rhamnus cathartica, Rhamnus
frangula (syn. F rangula alnus). The Nature Conservancy, Arlington, VA.

Corbin, J. D., and C. M. D'Antonio. 2004. Can carbon addition increase competitiveness
of native grasses? A case study from California. Restoration Ecology 12(1):36-43.

88

Curtis, J. T. 1971. The vegetation of Wisconsin; an ordination of plant communities.
University of Wisconsin Press, Madison, WI. 657.

D'Antonio, C. M., and J. C. Chambers. 2006. Using ecological theory to manage or
restore ecosystems affected by invasive plant species. Pages xii, 364 p. in D. A.
Falk, M. A. Palmer, and J. B. Zedler, editors. Foundations of restoration ecology.
Island Press, Washington, DC.

Drexler, J. 2., and B. L. Bedford. 2002. Pathways of nutrient loading and impacts on
plant diversity in a New York peatland. Wetlands 22(2):263-281.

Ehrenfeld, J. G. 2003. Effects of exotic plant invasions on soil nutrient cycling processes.
Ecosystems 6(6):503-523.

Fim, J ., A. P. N. House, and Y. M. Buckley. 2010. Alternative states models provide an
effective framework for invasive species control and restoration of native
communities. Journal of Applied Ecology 47(1):96-105.

Galatowitsch, S. M., N. 0. Anderson, and P. D. Ascher. 1999. Invasiveness in wetland
plants in temperate North America. Wetlands 19(4):733-755.

Godwin, H. 1943. F rangula alnus Miller (Rhamnusfrangula L.). Journal of Ecology
31:77-92.

Grootjans, A. P., E. B. Adema, W. Bleuten, H. Joostn, M. Madaras, and M. J anakova.
2006. Hydrological landscape settings of base-rich fen mires and fen meadows: an
overview. Applied Vegetation Science 9(2):]75-184.

Heneghan, L., F. F atemi, L. Umek, K. Grady, K. F agen, and M. Workman. 2006. The
invasive shrub European buckthom (Rhamnus cathartica L.) alters soil properties
in Midwestern US woodlands. Applied Soil Ecology 32(1): 142-148.

Herman, K. D., L. A. Masters, M. R. Penskar, A. A. Reznicek, G. S. Wilhelm, W. W.
Brodovich, and K. P. Gardiner. 2001. Floristic quality assessment with wetland
categories and exapmles of computer applications for the state of Michigan.
Michigan DNR, Wildlife Division, Natural Heritage Program, Lansing, MI.

Hobbs, R. J ., and J. A. Harris. 2001. Restoration ecology: Repairing the Earth's
ecosystems in the new millennium. Restoration Ecology 9(2):239-246.

Houlahan, J. E., and C. S. Findlay. 2004. Effect of invasive plant species on temperate
wetland plant diversity. Conservation Biology 18(4):] 132-1138.

Huebner, C. D., R. S. Morin, A. Zurbriggen, R. L. White, A. Moore, and D. Twardus.
2009. Patterns of exotic plant invasions in Pennsylvania's Allegheny National
Forest using intensive Forest Inventory and Analysis plots. Pages 258-270 Forest
Ecology and Management.

89

Kercher, S. M., and J. B. Zedler. 2004a. Flood tolerance in wetland angiosperrns: a
comparison of invasive and noninvasive species. Aquatic Botany 80(2):89-102.

Kercher, S. M., and J. B. Zedler. 2004b. Multiple disturbances accelerate invasion of reed
canary grass (Phalaris arundinacea L.) in a mesocosm study. Oecologia
138(3):455-464.

Klimkowska, A., R. van Diggelen, S. den Held, R. Brienen, S. Verbeek, and K. Vegelin.
2009. Seed production in fens and fen meadows along a disturbance gradient.
Applied Vegetation Science 12(3):304-315.

Lindig-Cisneros, R., J. Desmond, K. E. Boyer, and J. B. Zedler. 2003. Wetland
restoration thresholds: Can a degradation transition be reversed with increased
effort? Ecological Applications 13(1):]93-205.

MacDougall, A. S., and R. Turkington. 2005. Are invasive species the drivers or
passengers of change in degraded ecosystems? Ecology 86(1):42-55.

Magurran, A. E. 2004. Measuring biological diversity. Blackwell Publishers, Malden,
MA. viii, 256 p.

Matthews, J. W., and A. G. Endress. 2008. Performance criteria, compliance success, and
vegetation development in compensatory mitigation wetlands. Environmental
Management 41(1):]30-141.

Matthews, J. W., A. L. Peralta, D. N. Flanagan, P. M. Baldwin, A. Soni, A. D. Kent, and
A. G. Endress. 2009a. Relative inﬂuence of landscape vs. local factors on plant
community assembly in restored wetlands. Ecological Applications 19(8):2108-
2123.

Matthews, J. W., A. L. Peralta, A. Soni, P. Baldwin, A. D. Kent, and A. G. Endress.
2009b. Local and landscape correlates of non-native species invasion in restored
wetlands. Eco graphy 32(6): 103 1-1039.

Matthews, J. W., G. Spyreas, and A. G. Endress. 2009c. Trajectories of vegetation-based
indicators used to assess wetland restoration progress. Ecological Applications
19(8):2093-2107.

Middleton, B. A., B. Holsten, and R. van Diggelen. 2006. Biodiversity management of
fens and fen meadows by grazing, cutting and burning. Applied Vegetation
Science 9(2):307-316.

Nekola, J. C. 1994. The environment and vascular ﬂora of northeastern Iowa fen
communities. Rhodora 96(886): 121-169.

Panno, S. V., V. A. Nuzzo, K. Cartwright, B. R. Hensel, and I. G. Krapac. 1999. Impact
of urban development on the chemical composition of ground water in a fen-
wetland complex. Wetlands 19(1):236-245.

90

Peach, M., and J. B. Zedler. 2006. How tussocks structure sedge meadow vegetation.
Wetlands 26(2):322—335.

Quinn, G. P., and M. J. Keough. 2002. Experimental design and data analysis for
biologists. Cambridge University Press, Cambridge, UK ; New York. xvii, 537 p.

Ruiz-Jaen, M. C., and T. M. Aide. 2005. Restoration success: How is it being measured?
Restoration Ecology 13(3):569-577.

SAS Institute. 2003. SAS/STAT User's guide for personal computers. SAS Institute,
Cary, NC.

Satterthwaite, F. E. 1946. An approximate distribution of estimates of variance
components. Biometrics Bulletin 2:1 10-114.

SER Working Group. 2004. The SER international primer on ecological restoration,
www.ser.org & Tucson: Society for Ecological Restoration International.

Sinclair, A., and P. M. Catling. 1999. The value of cutting in the management of glossy
buckthom (Rhamnusfrangula L.). Wetland Journal 11(2):25-27.

Spieles, J. B., P. J. Comer, D. A. Albert, and M. A. Kost. 1999. Natural community
abstract for prairie fen. Michigan Natural Features Inventory, Lansing, MI.

Suding, K. N., K. L. Gross, and G. R. Houseman. 2004. Alternative states and positive
feedbacks in restoration ecology. Trends in Ecology & Evolution 19(1):46-53.

Swink, F ., and G. Wilhelm. 1994. Plants of the Chicago region. 4th edition. Indiana
Academy of Science, Indianapolis. xiv, 921 p.

Tuchman, N. C., D. J. Larkin, P. Geddes, R. Wildova, K. Jankowski, and D. E. Goldberg.
2009. Patterns of environmental change associated with Typha x glauca invasion
in a great lakes coastal wetland. Wetlands 29(3):964-975.

USDA. 2010. Plants Database.

van Diggelen, R., B. Middleton, J. Bakker, A. Grootjans, and M. Wassen. 2006. Fens and
ﬂoodplains of the temperate zone: Present status, threats, conservation and
restoration. Applied Vegetation Science 9(2): 157-162.

Verhoeven, J. T. A., W. Koerselman, and A. F. M. Meuleman. 1996. Nitrogen- or
phosphorus-limited growth in herbaceous, wet vegetation: Relations with
atmospheric inputs and management regimes. Trends in Ecology & Evolution
1 1(12):494-497.

Voss, E. G. 1985. Michigan ﬂora; a guide to the identiﬁcation and occurrence of the
native and naturalized seed-plants of the State, Part II: Dicots (Saururaceae-

91

Comaceae). Cranbrook Institute of Science. University of Michigan Herbarium,
Bloomﬁeld Hills, MI. 3 v.

White, P. S., and J. L. Walker. 1997. Approximating nature's variation: Selecting and
using reference information in restoration ecology. Restoration Ecology 5(4):338-
349.

Wilcove, D. S., D. Rothstein, J. Dubow, A. Phillips, and E. Losos. 1998. Quantifying
threats to imperiled species in the United States. Bioscience 48(8):607-615.

Williams, G. W. 2000. Reintroducing Indian-type ﬁre: implications for land managers.
Fire Management Today 60(3):40-48.

Zedler, J. B. 2000. Progress in wetland restoration ecology. Trends in Ecology &
Evolution 1 5(10):402-407.

Zedler, J. B., and S. Kercher. 2005. Wetland resources: Status, trends, ecosystem
services, and restorability. Annual Review of Environment and Resources 30:39-
74.

92

Table 5. Differences in resource availability between invaded, removal, and uninvaded
treatments pre-restoration and year 1 to 2 post-restoration (September 2007, 2009, and
2009, respectively). Signiﬁcant p—values (a=0.05) are bolded. Treatrnent*time
interactions not shown if neither treatment nor time were signiﬁcant.

 

Factor Treatment Time Treatment * time
F P F P F P

Light availability % full sun atl m~ 47.6 <.001 11.9 <.001 8.9 <.001
3°"S 9” 15.7 0.001 0.5 0.585 0.3 0.905
NH4(ppm) 2.5 0.126 13.4 <.001 2.5 0.068

1403(ppm)~ 0.7 0.500 8.1 0.001 0.9 0.486

Phosphorous (ppm)n 2.2 0.153 0.2 0.808 - -

Potassium (ppm) 0.5 0.612 9.8 0.001 0.7 0.599

% organic matter 1.4 0.278 0.6 0.548 - -

Microtopography #hummocks/m2~ 13.3 0.001 5.8 0.009 3.2 0.031

"' Log(x+1) transformed data used for ANOVA analysis.
" Square root transformed data used for analysis.

93

Table 6. Differences in plant community metrics between invaded, removal, and
uninvaded treatments pre-restoration and year 1 to 2 post-restoration (September 2007,
2009, and 2009, respectively). Signiﬁcant p-values (or=0.05) are bolded.

 

Factor l;l'reatmerlr’t F Time P Triatment l;Ime
% herbaceous cover 22.9 <.001 40.0 <.001 19.8 <.001
Relative graminoid cove; 25.7 <.001 1.3 0.301 3.5 0.023.
Relative exotic cover* 6.11 0.003 0.8 0.452 1.8 0.125
Plant species richness 32.6 <.001 70.0 <.001 8.8 <.001
Mean c°° 16.6 0.001 1.3 0.28 0.5 0.772
Floristic quality index 39.7 <.001 20.8 <.001 1.6 0.190
Simpsons DA 2.9 0.098 9.5 0.001 1.8 0.132

Berger-Parker dominance 1.7 0.219 5.4 0.005 1.3 0.263

* Square root arcsine transformed data used for analysis.
" Square root transformed data used for analysis.
00 Coefﬁcient of conservativeness on a scale of 1 to 10.

94

Table 7. ANOSIM based on similarity matrix by time. All resource factors and plant
factors presented in tables 1 and 2 included, plus % bare ground, soil Ca and Mg.
Signiﬁcant p-values are bolded. R ranges up to 1, a value closer to 1 indicates greater

between treatment differences.

 

Overall Test Invaded: Invaded: Removal:

Time Removal Reference Reference

R P R P R P R P
Pre restoration 0.31 0.017 -0.07 0.693 0.92 0.036 0.92 0.036
Year 1 0.50 0.001 0.27 0.043 0.88 0.012 0.67 0.012
Year 2 0.66 0.001 0.58 0.002 0.96 0.012 0.79 0.012

 

95

.8508 000 00003308 .00 800.00— .802.08 000000 m4
mam: 00000.50 4200080308 0003 0804000000000 00 000000000 00055 .modud .Emm+ 000 8000 000m .8003 00002000:
000 4050000 6000.00. .3200 N 0003000 0005000000000 000 Em: 5:50:05 00000800 :08 .00 800800090000 .8 00am...—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N .00> ...00> 00.00.0000. 05 N .00> .. .00> 00.00.0000. En.
0 . 0 o 0 p 0 o
m
m r —. D
8 m 0 n 8
m r N M. w
0 m r.
m . w. 94 m
a a m m
w w
. e m 8 m
0 m. .i
< w 0
Z < . on
D. 8 1 TO
0. 00..
r t 0.0
4 or
8:80.50 + m
.0580”. Jrl -4 . ON T 11Wl\\um . m6
.. 889... 161 d In 0
e
. on m . 0.0
m
r O I . S
0... m N... “H
. 8 m. - 6.0 H
W
. 8 an
. E.
< . 8..

 

 

 

 

om

 

 

 

96

.8527. 2w 3§omﬁwmm mo £032 .maouaaaom 3on

m4 mam: “castaway banaowawmm 803 vautcogaob 8 “853.5 82;? .83 «8330 880% Ema H8632: mo Sam
05 ﬁgmoae B>oo £6836: o>u§u< .8505 0.8 88:3,. 98 808305 ﬁrs <>OZ< >~B-N a 88m maoobo “88888
.modud .Emm+ 0.5 mag “ohm .329 832:5. BB 332: 35c M 50253 8an Ema .«o 3859800 6 0.5me

 

 

 

 

 

 

N Eo> F 58, c8898. mi N 30> F 50> c8828.. 2n.
n h - mw.o _ p L m
mu
. and m
m l c
d S
1 mad m w
e m
H - m 0
1 oto W m
m m
m . o m
L mvd m w.
w «M
m. moceeom Ill
- and m Ease”. 1T - h
D m 88>... IOI 0

 

 

 

 

 

 

and

 

 

f

 

 

 

 

 

 

 

v
”-2 a cum

1 n

INwd Q M

m! H w o 0 .8m
0 m

-20 1

m m

< .8W 9 m

m .own

INN s

w m nod

m.& a. a (foam

 

 

mm

97

 

 

 

 

 

 

 

N .mo> F 56> 5:893. 2n.
m
< I
4 <
8533”. III -
.9281 I51
38>... + “—

 

 

cod

mod

0
F.
O

to
".
O

o
N.
o

no
N
o

JOAOO ogoxa uoguodmd

c3898. 9n.

 

 

 

 

0.0

"'.
O

0!
o

‘7.
o

to. :0
O 6
Jero pgougwms uoguodmd

‘Q
c

Figure 6 cont’d.

98

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A: Pre restoration 20 Stress: 0.08
A B i
l ‘ ° l l l
A A
o ‘0 ' I
O
O ‘ o
A Invaded
A
0 Removal
I Reference
B: Year 1 2D Stress: 0.09
O I
A I
N ‘ o ° °
.2 A
E O
A A O I
C: Year 2 20 Stress—636
A B C
I
l ‘ H I l
‘ O
A Q) o
A
O
I
O I
A
A
Axis 1

Figure 7. Two dimensional NMDS plots based on a mean of each resource and plant
metric by treatment and replicate using the Euclidian dissimilarity index. There were two
reference plots pre-restoration; a third was added before year 1. Pre-restoration and year 2
statistical differences by treatment are shown with letters.

99

Chapter 4

For Restoration Ecology

INVASIVE SPECIES REMOVAL LEADS TO RAPID RECOVERY OF
POLLINATOR COMMUNITIES IN PRAIRIE FEN

Anna Katherine F iedler

100

Abstract
Habitat restoration is increasingly used to reverse degradation of rare ecosystems and
maintain biological diversity. In the face of increasing concerns about overall biodiversity
decline, and speciﬁcally pollinator loss, we ask whether restoration of a prairie fen
wetland increases plant and pollinator diversity. Here, we compare herbaceous plant, bee,
and butterﬂy abundance, diversity, and species composition in prairie fen invaded by
glossy buckthom (F rangula alnus), in restored areas where glossy buckthom was
removed, and in uninvaded reference areas. We found striking differences between ‘
invaded and uninvaded prairie fen in pollinator abundance and species composition. Plant
species diversity and composition in restored plots were signiﬁcantly different than
invaded plots two years after restoration, but also remained signiﬁcantly lower than
reference plots over this time period. In contrast, bee and butterﬂy abundance, diversity,
and composition were similar in restored and reference plots, and distinct from invaded
plots, in the second growing season following restoration. Our ﬁndings indicate that a
diverse community of mobile generalist pollinators rapidly re-colonizes restored areas,
while the plant community may take longer to recover. This work implies that, in areas
with intact pollinator metapopulations, restoration will likely prevent ﬁrrther loss of
mobile generalist pollinators. On the other hand, targeted efforts will likely be required to

restore populations of rare insects.

Keywords: conservation, restoration, pollinator, prairie fen, plant community.

101

Introduction

In recent years, concerns about losses of pollinator diversity have mounted. Bees
have long been considered important crop pollinators (Klein et a1. 2007), but their
importance in natural areas is also increasingly recognized (Potts 2010, Winfree 2010),
and pollinator losses extend beyond the managed honeybee, Apis mellifera L., to key
native pollinator groups, e.g. Bombus sp. (Goulson et al. 2008). There are likely multiple
drivers of pollinator diversity loss, including land-use change, introduction of non-native
plants, pollinators, grazers, and pathogens, altered hydrology, and climate change (Kevan
1999, Potts 2010). The primary threat to bee diversity is habitat loss, with invasive
species, parasites, and disease as secondary factors (Brown and Paxton 2009). Because
pollinators and wild plant populations are tightly linked, pollinator declines and
extinctions have the potential to result in trophic cascades that affect plant diversity.
About 90% of angiosperms are animal pollinated (Nabhan and Buchmann 1997) and over
60% of plant species are dependent on pollination by insects for maximal fruit and seed
set (Burd 1994, Ashman et al. 2004). Concomitant, large-scale declines have been found
in pollinator abundance and insect-pollinated plant diversity pre- versus post-1980 in
England and the Netherlands (Biesmeijer et al. 2006). With multiple lines of evidence
for losses in both plants and pollinators, maintaining species diversity and pollination is
an increasing priority for the long-term persistence of rare communities and habitats.

The use of habitat restoration to reverse losses of rare communities and the
species within them is rising. One of the primary goals of ecosystem restoration is to
increase diversity within protected areas (Hobbs and Norton 1996, SER Working Group
2004). Ecosystems are considered restored when the structure, function, and composition

are returned to a goal state, often based on historic conditions (SER Working Group
1 02

2004). The effect of restoration on plant species diversity is relatively well-documented
(Ruiz-Jaen and Aide 2005) and plant conservation is well-represented in scientiﬁc studies
(Clark and May 2002). However, the effects of restoration on the pollinator community
are less well known.

Because pollinators are infrequently considered in restoration, we do not know
whether typical restoration activities prevent further pollinator species losses, and there is
concern that they do not (Dixon 2009). Two insect groups, bees and butterﬂies, play an
important role in maintenance of pollination and species diversity. Bees are a species-rich
group, and the dominant group of pollinators in many regions (Williams et al. 2001).
They are well known for their prevalence as effective pollinators of both crops and wild
plants (Kearns et al. 1998). Butterﬂies comprise a well-known, charismatic group of
invertebrates and some may be effective umbrella species for conservation whose
protection may lead to protection of other species (New 1997). Because butterﬂies are
herbivorous as larvae and nectivarous as adults, they provide a potential link between
ecosystem functions of herbivory and pollination (Waltz and Covington 2004). In
addition, butterﬂy species range from specialists to generalists, and are known to recover
rapidly following restoration (Waltz and Covington 2004). Finally, in the Midwestern
U.S., butterﬂies and skippers form the most prevalent group of insects that are strongly
dependent on remnant savanna and prairie ecosystems (Panzer et a1. 1995).

In the Midwestern U.S., there is a relatively high frequency of globally rare
prairie fen wetlands. These ecosystems contain extremely high species diversity in small,
isolated patches, and are threatened by the same factors affecting species diversity

globally, (Spieles et al. 1999, Amon et a1. 2002, Bedford and Godwin 2003, Nekola

103

2004) e. g. habitat loss, invasive species, and pollution (W ilcove et a1. 1998). Prairie fens
contain a number of rare and endangered organisms, including 19 plant species and 25
insect species, 4 of which are butterﬂies (MNF I 2007). Many of the fens in the Midwest
are degraded, primarily by invasive species and changes to hydrology. Restoration
activities in prairie fen frequently involve removal of invasive species, but the effect on
pollinators and the plants on which they depend has not been studied.

Measuring changes in plants and insects is important to gain an understanding of
whether restoration activities lead to restoration of pollinator communities. In addition,
patterns in prairie fen provide an indication of what is likely in other restored ecosystems.
We examined plant and pollinator response to removal of invasive F rangula alnus L.
(glossy buckthom) from a prairie fen wetland in Michigan, USA. Our goal was to
examine the effect of restoration on the diversity, abundance, and community of both
plants and pollinators in the ﬁrst two years of a long-term restoration project. We
hypothesized that under restoration treatments 1) forb abundance and plant diversity
would increase and 2) bee and butterﬂy abundance and diversity would increase, with the
community becoming distinct from that in invaded plots but remaining distinct from
uninvaded plots. We also predicted that plant and pollinator communities would not

reach reference community states within this short time period.

Methods

Experimental design

We conducted this study in a prairie fen at the Michigan State University
MacCready Reserve in Clarklake, Liberty Township, Jackson County, Michigan. In

104

25x25 rn plots, we cleared invasive F. alnus ﬁ'om 6 replicate plots, left F. alnus in 6
replicate plots, and delineated three 25x25 m uninvaded reference areas (Chapter 2). We
cleared F. alnus in February 2008, following best management practices (The Nature
Conservancy, R. Villegas pers. comm.) of cutting and treating F. alnus stumps with
herbicide (glyphosate 25% AI, Rodeo, Monsanto, St. Louis, MO) with Cide Kick 11
surfactant, adj uvant, and penetrator (Brewer International, Vero Beach, FL), and stacked
brush piles outside of sampling areas (Chapter 2). Brush was burned in early spring 2009.
During the 2008 and 2009 growing seasons, we applied foliar herbicide to F. alnus re-
sprouts, P. arundinacea L., Rosa mulzﬂora Thunb., and Cirsium arvense (L.) Scop., and
cut and treated cut stems of Typha sp. and Populus tremuloides Michx. within restored

plots.

Sampling

During the 2008 and 2009 growing seasons, we assessed plant diversity and cover
in 9 m2 quadrats laid out on a grid within each plot (Chapter 2) three times during the
growitng season: 2-6 June 2008, 1-5 June 2009, 25-30 July 2008, 29 July-6 August 2009,
5-8 Sept. 2008, and 9—11 Sept. 2009. We identiﬁed and estimated percent cover of each
monocot, forb, and shrub less than 1.5 m height that was rooted within each quadrat.

We sampled pollinators using two methods, observational sampling and pollinator
bowl traps, monthly from June-September 2008 and 2009 (12 June, 8 August, 6 Sept.
2008 and 4 June, 3 July, 5 August, and 31 August 2009). Both techniques were
performed on the same day, in sunny, calm weather, when insects are most likely to be
active. We performed pollinator observation between 10am and 3pm EST. On each date,
two people stood back to back in the center of each replicate plot. Upon arrival at the

105

location, observers waited 1 minute, noting plant species blooming within 2 m of the
sample point. Each then observed all bees and butterﬂies that entered a 2 m radius half
circle around them for 5 minutes (10 minutes in June 2009), identifying all pollinators in
the sampling area within that time to genus and, when possible, collecting unknowns for
identiﬁcation. Although small pollinators are less likely to be detected using this
technique we did observe small bees including Hylaeus and Lasioglossum sp. using this
method.

Pollinator traps consisted of 3.25 oz. white cups (Solo Cup Company, Lake
Forest, IL). One third of the cups were painted blue ﬂuorescent, and one third were
painted yellow ﬂuorescent (Guerra Paint and Pigment, New York, NY, Droege 2010).
We placed two traps on opposite sides of all 9 plant quadrats per replicate, with starting
cup color randomized on each date and color alternating in the order: blue, white, yellow.
Cups were placed within 40 cm of the quadrat on the ground in locations where they were
not obscured by low growing vegetation. Each cup was ﬁlled 1/3 ﬁlll with soapy water
(Droege 2010) and placed in the ﬁeld between 8-10am. Cups were removed in the same
order they were placed in the ﬁeld, from 4-7pm on a given sampling day, soap water was
strained out, and all insects were strained into ethanol into samples pooled by replicate.
All bees and butterﬂies were removed from samples, washed (Droege 2010), pinned, and

identiﬁed to species.

Statistical analysis

We compared the abundance, diversity, and community similarity of plants, bees,
and butterﬂies in invaded, restored, and uninvaded plots in 2008 and 2009. For plants, we
used maximum percent cover per species within a growing season. For bee abundance,

106

we used season-long means per 5-minute observation, and for bee diversity and
community similarity we used pollinator bowl data. For butterﬂy abundance, diversity,
and community similarity, we used season-long means per S-minute observation. In all
cases season-long means were calculated by sampling date and replicate of each
treatment. Pollinator abundance during the lO-minute sampling period in June 2009 was
divided by 2 to make this sampling effort comparable with those on other dates. Bees in
pollinator bowls were summed per replicate on a given date. We used Simpsons diversity
of each plants, bees, and butterﬂies as a metric representing both species richness and
evenness, using D=l/z(p,.)2 where p, is the proportion of cover of the ith plant species per
quadrat or the abundance of a pollinator (Magurran 2004).

To investigate differences in plant and pollinator abundance and Simpsons
diversity, we performed 2-way ANOVAs for each group with treatment: invaded,
removal, and uninvaded reference, and time since restoration: year 1 and year 2 (PROC
MIXED, SAS Institute 2010). Bee and butterﬂy abundance were log(x+1) transformed to
meet assumptions of normality and homogeneity of variances, and we used Satterthwaite
adjusted degrees of freedom to account for the unequal number of replicates between
treatments (Satterthwaite 1946).

To examine community-level shifts in plant, bee, and butterﬂy abundance with
restoration, we created a similarity matrix with square root transformed data using the
Bray-Curtis index, which is well suited to species abundance data (Quinn and Keough
2002). In the case of insect abundances, a “dummy variable” of l was added to all values,
so that treatments and replicates with zero insects stayed in the analysis. We used

nonmetric multidimensional scaling (N MDS) to visualize the differences by treatment

107

and year on community metrics. The NMDS ordinates data by ranking variables so that
the closer their location is in 2 or 3 dimensional space, the more similar they are. We
performed the NMDS with 2 and 3 dimensions, 25 random starting conﬁgurations and a
minimum stress of 0.01 (Clarke and Gorley 2006a). All stress values for 2-dimensional
ﬁgures were 0.14 or less. Values under 0.10 indicate that 2-dimensional representation is
an accurate representation of the relationship between points. Values from 0.10 to 0.20,
while still representative, will not likely accurately show the small-scale relationships
between points (Clarke and Gorley 2006b). To examine statistical differences by
treatment in community metrics, we performed analysis of similarity (ANOSIM) using
the Bray-Curtis resemblance matrix of plant, bee, and butterﬂy data for year 1 and 2
separately, with 999 random permutations (Clarke and Gorley 2006b). Overall tests were

signiﬁcant in all cases, so pairwise tests were appropriate and are reported herein.

Results

We observed rapid changes in plant and pollinator abundance, diversity, and
communities within the ﬁrst two years of restoration. We examined forb cover as an
indicator of potential ﬂowering resources for pollinators. There was signiﬁcantly greater
percent forb cover in removal and reference plots than in invaded plots in both year 1 and
year 2 (Figure 8A) (P=0.011, F 5.42.253), with no signiﬁcant difference by year (P=0.072,
F 3.61332) or treatment by year (P=0.275, F 1.42332). Plant diversity also shifted in the
ﬁrst two growing seasons following restoration, but not as completely. We found
signiﬁcant differences by treatment in season-long Simpsons diversity of all plants less
than 1.5m tall (P=0.013, F 31 92,1 1.6), with signiﬁcantly greater diversity in uninvaded

reference plots than removal plots, and signiﬁcantly greater diversity in removal plots

108

than those that remained invaded by buckthom. There were no signiﬁcant differences in
plant diversity by year (P=0.085, F 3.0”“) or treatment by year (P=0.32, F 112,141).

An NMDS ordination of the plant community indicated a shift in the overall
community following restoration (Figure 9). In year 1, there was no signiﬁcant difference
between the plant community of removal and buckthom invaded plots (R 0.17, P=O. 123)
while by year 2, removal and invaded plant communities were signiﬁcantly different (R
0.5, P=0.002). In both years, there were signiﬁcant differences between the invaded and
reference plant communities (Year 1: R 0.86 P=0.012, Year 2: R 0.82 P=0.012) and the
plant community in removal plots remained signiﬁcantly different than that in reference
plots through year 2 after restoration (Year 1: R 0.71 P=0.012, Year 2: R 0.83 P=0.012)
(Figure 9). A number of shade-tolerant species were more common in invaded areas;
forbs included Circaea lutetiana, Symplocarpusfoetidus, Rosa multiﬂora, and Viola
nephymphylla (Table 8). A second group of disturbance tolerant species were more
common in recently restored areas, including monocots Carex bebbii, C. hystericina, C.
stipata and Leersia oryzoides and forbs Bidensﬁondosa, Epilobium coloratum,
Erechtites hieraciifolia, Eupatorium perfoliatum, Senecio aureus, Solanum dulcamara,
and Pilea pumila (Table 8). A number of species remained far more abundant in
uninvaded areas, including monocots Calamagrostis canadensis, Carex stricta, C.
tetanica, Muhlenbergia glomerata, Schizachyrium scoparium, Scirpus acutus, S.
americanus, and Sorghastrum nutans, as well as forbs Aster lanceolatus, Eupatorium
macalatum, Gentianopsis crinita, Hypoxis hirsuta, Lycopus americanus, Lysimachia
quadriﬂora, Polygonum amphibium, Potentillaﬁuticosa, Rudbeckia hirta, and Zigadenus

elegans, and the fern T helypteris palustris (Table 8).

109

Bee abundance and diversity shifted rapidly and more completely with restoration
than the plant community. Bee abundance based on observational sampling was
signiﬁcantly different by treatment (P<0.001, F 32.82,”), year (P=0.002, F 1581,12), and
treatment*year (P=0.003, F 10.12,”), with signiﬁcantly lower abundances in invaded
treatments in both years than all other treatments. In year 1, there were no signiﬁcant
differences between removal and uninvaded reference plots (Figure 10A). In year 2, there
were signiﬁcantly more bees in removal than reference plots (Figure 10A). Apis mellifera
dominated the bee community according to observational sampling, followed by Bombus,
Hylaeus, and other Halictidae in year 1. In year 2, Apis mellifera, Hylaeus, and Bombus
remained dominant groups, with a number of other genera represented by small numbers
of insects (Table 9).

We used pollinator trap data to assess bee diversity at the species level. Diversity
of bees was signiﬁcantly lower in invaded than removal and reference plots in both study
years (Figure 10B), with no signiﬁcant differences between removal and reference plots.
There were no signiﬁcant differences in bee diversity by year (P=0.91, F 0.01 1,24) or
treatrnent*year (P=0.12, F 232,24), but there were signiﬁcant differences by treatment
(P=0.001, F 12.4234). A number of bee species were represented by only one individual
in both year 1 and year 2, with 31% (9 of 29) of species (range in abundance 1-24) and
46% (23 of 50) of bee species (range in abundance 1-84) represented by one specimen in
years 1 and 2, respectively. This means that there is a high likelihood of species turnover
based on our sampling technique between years. Lasioglossum Dialictus sp. 2 was more
abundant in reference plots in both years than in other treatments (Table 10). Several bee

species were more abundant in removal plots in year 2 than in year 1, including Ceratina

110

calcarata/dupla, Augochlora pura, Augochlorella aurata, Lasioglossum Dialictus sp. 2
and Lasioglossum rohweri.

Patterns of pollinator abundance between observational and pollinator trap data
were similar with the notable exception that Apis mellifera and Bombus species were two
of the most abundant groups observed, for a combined total of 51.2 and 49.1 of observed
pollinating bees in year 1 and 2 after restoration. In contrast, these two groups comprised
less than 4% of total pollinators collected with pollinator traps in both year 1 and 2.

An NMDS analysis of the bee community showed evidence of a shifting
pollinator community within the ﬁrst season following restoration (Figure 1 1). There
were no signiﬁcant differences in the bee community between the removal and reference
plots (Year 1: R 0.15 P=O.19, Year 2: R -0.24 P=O.88), while the bee community in
buckthom invaded plots was signiﬁcantly different than removal plots in both years 1 and
2 (Year 1: R 0.76 P=0.001, Year 2: R 0.63 P=0.004). Invaded and reference plots
remained signiﬁcantly different (Year 1: R 0.85 P=0.012, Year 2: R 0.39 P=0.036).
Several of the most abundant pollinators were abundant in both sampling years, including
Ceratz'na calcarata/dupla, Lasioglossum Dialictus sp. 2, and Augochlora aurata (Table
10). There were also several notable changes between years. Lasioglossum rohweri
increased in relative abundance in year two. There was a nearly complete shift in
Megachilidae species between years 1 and 2 so that species represented in year 1 were
absent or nearly so in year 2, with others collected in year 2 (Table 10). In addition, a
greater number of Lasioglossum Dialictus sp. and Andrena species were collected in year
2 than in year 1. The only species which decreased greatly in abundance between year 1

and year 2 was Augochlora pura (Table 10).

111

Butterﬂy abundance and diversity also shifted rapidly post-restoration. We found
signiﬁcant differences in butterﬂy abundance by treatment (P=0.005, F 8.42,”) but not by
year (P=0.35, F 0.1 Hz) or treatment*year (P=0.32, F 1.32,;2). There were signiﬁcantly
fewer butterﬂies in invaded plots than other treatments in both study years, and no
signiﬁcant difference between removal and on-site reference plots (Figure 10C). There
were signiﬁcant effects on butterﬂy diversity by treatment (P=0.003, F 10.3 2,! 2), year
(P=0.02l, F 7.1 Hz), and treatrnent*year (P=0.041, F 422,12). Butterﬂy diversity was
signiﬁcantly lower in invaded plots than in restored and uninvaded reference plots in both
years, with no signiﬁcant differences between restored and reference plots (Figure 10D).
A number of butterﬂy species were singletons in both study years, with 45.5% (5 of 1 l)
of species (range in abundance 1-22) and 40% (6 of 15) of species (range in abundance 1-
15) represented by one individual in years 1 and 2, respectively. At the species level,
Pieris rapae was more abundant than any other butterﬂy in year 1, while in year 2
Phyciodes tharos and Poanes Massasoit composed a greater proportion of butterﬂies
observed.

An NMDS analysis on the butterﬂy community also indicated a rapid response to
restoration (Figure 12). The butterﬂy community in invaded plots was signiﬁcantly
different than that in the removal in years 1 and 2 (Year 1: R 0.44 P=0.02, Year 2: R 0.40
P=0.002). In year 1, removal and reference plots had similar butterﬂy communities (R -
0.17 P=0.79), as did invaded and reference plots (R 0.50 P=0.083). No butterﬂies were
collected in invaded areas, with a mean of li0.58 butterﬂies/5 minute sampling collected
in uninvaded references. In year 2, the butterﬂy community was signiﬁcantly different in

invaded than both the removal (R 0.40 P=0.002) and reference (R 0.74 P=0.012) plots

112

and was not different in removal versus reference plots (R 0.21 P=0.17). The only
butterﬂy species seen in invaded plots was Megisto cymela (Table 11). Despite no
signiﬁcant community-wide differences in butterﬂies in year 2 between reference and
removal plots, Epargyreus clams and Speyeria cybele cybele were more abundant in
removal than reference plots, while Ancyloxypha numitor, Poanes Massasoit, and

Phyciodes tharos were more abundant in reference than removal plots (Table 11).

Discussion

There are increasing concerns about ecosystem degradation and loss of the plants
and pollinators within ecosystems. The assessment of ﬂowering plant resources as well as
pollinator diversity, abundance, and community recovery are vital to evaluating the long-
term stability of pollinator populations in degraded and restored ecosystems. In this
study, we examined plant and pollinator communities with restoration of prairie fen
wetland, and found rapid responses in both species groups. We found distinctly lower
abundance and diversity of plants, bees, and butterﬂies in areas of the fen invaded by F.
alnus, while the plant and pollinator communities in restored areas shifted rapidly toward
those in uninvaded reference plots. Our results indicate that, within two years of F. alnus
removal, generalist pollinator community and diversity were restored in this ecosystem.
In contrast, forb abundance, plant diversity, and the plant community in restored areas
were intermediate between invaded and reference plots.

We did ﬁnd support for our hypothesis that plant diversity would increase in
restored, F. alnus removal plots, although in the short-term, season-long plant diversity
remained lower in removal than uninvaded plots. This is similar to other ﬁndings in

wetland restorations, where goals related to percent cover are more frequently met than

113

those of species diversity (Matthews et al. 2009c). Remnant ecosystems frequently have
greater native plant species richness and diversity than restored or re-created systems
(Polley et al. 2005, Shepherd and Debinski 2005). Even if the local plant community is
manipulated, the plant species pool and quality are limited by landscape structure and
mesoscale dynamics (Matthews et al. 2009a). For plants, regional pools also determine
the likelihood of reinvasion by non-natives (Matthews et al. 2009b). This points to the
potential need for management of regional plant species pools in restoration.

There may be multiple factors playing a role in the rapid insect response to .
restoration in this prairie fen. The primary one is likely the fact that areas of intact,
uninvaded fen were within 300m of invaded areas, so that there were existing ﬂoral
resources nearby at a scale within foraging distance for even the smallest pollinators we
collected (Steffan-Dewenter et al. 2002, Greenleaf et al. 2007). The link between plants
and pollinators is well known; a number of studies have found positive relationships
between richness or abundance of ﬂoral resources and pollinator diversity or activity at
local (Erhardt 1985, Hegland and Boeke 2006, Tuell et al. 2008) and landscape scales
(Steffan-Dewenter and Tschamtke 1999, Steffan-Dewenter et al. 2002, Potts et al. 2003).
For butterﬂies, availability of both nectar resources and larval host plants affect their
distribution (waell et al. 2004).

While we did not measure ﬂoral resources directly in this study, we found that in
restored areas, forb cover rapidly increased to levels similar to those in uninvaded plots.
Inasmuch as percent cover of forbs may act as a proxy for ﬂoral resources this may
indicate that ﬂoral abundance increased in restored areas. Light availability has been

documented to affect butterﬂy abundance in other restorations with no changes in nectar

114

plant species richness (Waltz and Covington 2004). Our previous work shows that light
availability increased rapidly to reference conditions following removal treatments
(Chapter 3). Butterﬂies are known to use habitat openings (Pollard et al. 1975) and
butterﬂy richness is greatest in the open-structured phase of grassland clearing (Erhardt
1985). In Mediterranean systems, bee abundance of common species was greater in areas
with more recent disturbance (Potts et al. 2003). In time, fewer pollinators may be found
in restored fen plots if they are again encroached by shrubs and light availability
decreases.

Rapid shifts in the pollinator community contradicted our hypothesis that the
pollinator community in restored areas would remain distinct from that in uninvaded
areas. In fact, all pollinator community metrics became similar to those in uninvaded fen
within the ﬁrst year following restoration. In our study, bee abundance in restored plots
surpassed that in reference plots in the second year after restoration. We did not see the
same pattern in bee diversity, however, indicating a potential maximum bee diversity was
reached within a plot even as bee abundance continued to increase. Similarly, Ebeling et
al. (2008) found a leveling off in pollinator species richness with plant species richness
and ﬂoral area, but a continued increase in the ﬁ'equency of pollinator visits to ﬂowers.

In addition to resource density, habitat size and quality play a role in pollinator
abundance and diversity. Habitat area inﬂuences which pollinator species use a patch. In
calcareous grasslands in Europe, increased species richness of monophagous butterﬂies
in calcareous grasslands with habitat area (Steffan-Dewenter and Tschamtke 2000)
provide evidence that specialists may be more likely to ﬁnd the resources they require in

larger areas. Scale also affects wild, solitary bees at much smaller scales than bumble

115

bees and honeybees (Steffan-Dewenter et al. 2002). Our study site includes a relatively
small total area and many intact prairie fen wetlands are small, isolated communities; it
would be interesting to see whether there is a relationship between the size of calcareous
grasslands in North America and insect diversity, especially of rare species.

The regional species pool is also a key element of species diversity and
abundance. In general, animals are more mobile than plants. However, fauna also have a
range of responses to restoration due to differences in their mobility and habitat
requirements, the presence of predators, and the rate of revegetation (Nichols and Nichols
2003). For more mobile ﬂying insects, the regional species pool and quality of the
landscape surrounding restored areas may also be key factors (Sumrnerville et al. 2005).
More mobile, common insects are likely to recolonize restored areas more rapidly than
rare species. None of the six butterﬂy species of concern that inhabit prairie fens were
found at our study site, despite the presence of small intact fen patches throughout the
study area. For those species of concern, habitat suitability is clearly not the only factor
determining their presence at a site, and larger scale restoration and connection of
currently disparate metapopulations, or even assisted migration, may be necessary for
rare species to inhabit this prairie fen and other restored areas.

Despite evidence of increased ﬂowering plant and pollinator diversity and
abundance following habitat restoration and the prevalence of generalization in
pollination systems (W aser et al. 1996) we do not know whether the plant-pollinator
interactions that lead to persistence of insect and plant diversity have been restored.
Plant-pollinator communities may be successfully restored but species interactions may

remain more complex on remnant than restored habitat (F orup et al. 2008). Different

116

species may be carrying out pollination, with unknown long-term repercussions. For
example, plant abundance and insect richness and abundance are restored in old ﬁeld
meadows, although the species comprising each system are distinct (Forup and Memmott
2005). In addition, pollinator, especially bee, communities vary in species composition
and abundance by year (Williams et al. 2001, Petanidou et al. 2008). Patterns of bee
abundance in our study matched that pattern; a large proportion of species collected in
both years were singleton species. This means that longer-term studies of pollination will
more likely assess the ability of plants and pollinator communities to persist in restored
habitats. Insect pollinators may be required in the long term to maintain genetic
variability in plant populations, as mobile insects can ﬁrnction as mobile link organisms,
linking genetic material through pollination (Lundberg and Moberg 2003).

Our study indicates that, in a landscape with intact habitat patches, mobile insect
abundance, diversity, and community were actually more rapidly restored than plant
communities in the short term. At our study site, patches of degraded habitat were
interspersed with patches of intact prairie fen, as well as neighboring undeveloped
uplands. This landscape likely provided sufﬁcient resources for the persistence of
generalist pollinators (W infree et al. 2009), which readily re-colonized restored areas. In
an area without generalist pollinators, we would expect to see cascading effects on plant
community diversity over time (Memmott et al. 2004). Finally, whether current
restoration activities are sufﬁcient for conservation of both plants and pollinators that are

endemic to this habitat will need to be assessed over decades.

Implications for practice

117

0 Current restoration best practices for prairie fen restoration are sufﬁcient for
restoration of the insect pollinator community in the short-term.

0 Rare plants may require re-introduction and long-term monitoring to ensure their
persistence in this habitat.

0 Rare and endangered insects may require well-maintained corridors or even
relocation assistance to enter this site in the medium-term, and continued

monitoring will be required to assess long-term persistence.

Acknowledgements

C. Phillippo and M. Lettow provided invaluable ﬁeld assistance. We are grateful
to T. Reznicek for plant identiﬁcation and B. DeMarco, J. Tuell and M. Nielsen for insect
identiﬁcation. Thanks to J. McCarthy for creating the bee, butterﬂy, and plant symbols.
Research funding was provided by National Fish and Wildlife Foundation, the MSU
MacCready Endowment, and a Hanes Trust Fellowship of the Michigan Botanical
Society. Funding support for A.K.F. was provided through a sustainable agriculture
predoctoral fellowship by the CS. Mott foundation, a MSU Barnett Rosenberg
predoctoral fellowship in biological sciences, and a MSU College of Agriculture

dissertation completion fellowship.

118

Literature cited

Amon, J. P., C. A. Thompson, Q. J. Carpenter, and J. Miner. 2002. Temperate zone fens
of the glaciated Midwestern USA. Wetlands 22(2):301-317.

Ashman, T. L., T. M. Knight, J. A. Steets, P. Amarasekare, M. Burd, D. R. Campbell, M.
R. Dudash, M. O. Johnston, S. J. Mazer, R. J. Mitchell, M. T. Morgan, and W. G. Wilson.
2004. Pollen limitation of plant reproduction: Ecological and evolutionary causes and
consequences. Ecology 85(9):2408-2421.

Bedford, B. L., and K. S. Godwin. 2003. Fens of the United States: distribution,
characteristics, and scientiﬁc connection versus legal isolation. Wetlands 23(3):608-629.

Biesmeijer, J. C., S. P. M. Roberts, M. Reemer, R. Ohlemuller, M. Edwards, T. Peeters,
A. P. Schaffers, S. G. Potts, R. Kleukers, C. D. Thomas, J. Settele, and W. E. Kunin.
2006. Parallel declines in pollinators and insect-pollinated plants in Britain and the
Netherlands. Science 313(5785):351-354.

Brown, M. J. F., and R. J. Paxton. 2009. The conservation of bees: a global perspective.
Apidologie 40(3):410-416.

Burd, M. 1994. Bateman principle and plant reproduction - the role of pollen limitation in
ﬁuit and seed set. Botanical Review 60(1):83-139.

Clark, J . A., and R. M. May. 2002. Taxonomic bias in conservation research. Science
297(5579):191-192.

Clarke, K. R., and R. N. Gorley. 2006a. Primer v6. Primer-E, Plymouth, UK.

Clarke, K. R., and R. N. Gorley. 2006b. Primer v6: User Manual/Tutorial. Primer-E,
Plymouth, UK

Dixon, K. W. 2009. Pollination and restoration. Science 325(5940):571-573.

Droege, S., editor. 2010. The very handy manual: How to catch and identify bees and
manage a collection. ftp://ftpext.usgs.gov/pub/er/md/laurel/Droegel.

Ebeling, A., A. M. Klein, J. Schumacher, W. W. Weisser, and T. Tschamtke. 2008. How
does plant richness affect pollinator richness and temporal stability of ﬂower visits?
Oikos 117(12):1808-1815.

Erhardt, A. 1985. Diurnal lepidoptera: sensitive indicators of cultivated and abandoned
grassland. Journal of Applied Ecology 22(3):849-861.

Forup, M. L., K. S. E. Henson, P. G. Craze, and J. Memmott. 2008. The restoration of
ecological interactions: plant-pollinator networks on ancient and restored heatlrlands.

Journal of Applied Ecology 45(3):742-752.

119

Pomp, M. L., and J. Memmott. 2005. The restoration of plant-pollinator interactions in
hay meadows. Restoration Ecology 13(2):265-274.

Goulson, D., G. C. Lye, and B. Darvill. 2008. Decline and conservation of bumble bees.
Annual Review of Entomology 53:191-208.

Greenleaf, S. S., N. M. Williams, R. Winfree, and C. Kremen. 2007. Bee foraging ranges
and their relationship to body size. Oecologia 153(3):589-596.

Hegland, S. J ., and L. Boeke. 2006. Relationships between the density and diversity of
ﬂoral resources and ﬂower visitor activity in a temperate grassland community.
Ecological Entomology 31(5):532-538.

Hobbs, R. J ., and D. A. Norton. 1996. Towards a conceptual framework for restoration
ecology. Restoration Ecology 4(2):93-1 10.

Kearns, C. A., D. W. Inouye, and N. M. Waser. 1998. Endangered mutualisms: The
conservation of plant-pollinator interactions. Annual Review of Ecology and Systematics
29:83-1 12.

Kevan, P. G. 1999. Pollinators as bioindicators of the state of the environment: species,
activity and diversity. Agriculture Ecosystems & Environment 74(1-3):373-393.

Klein, A. M., B. E. Vaissiere, J. H. Cane, I. Steffan-Dewenter, S. A. Cunningham, C.
Krernen, and T. Tschamtke. 2007. Importance of pollinators in changing landscapes for
world crops. Proceedings of the Royal Society B-Biological Sciences 274(1608):303-
313.

Landis, D. A., A. K. Fiedler, C. A. Hamm, D. L. Cuthrell, E. H. Schools, D. R. Pearsall,
M. E. Herbert, and P. J. Doran. submitted 2010. Insect conservation in prairie fen:
addressing multiple challenges. Journal of Insect Conservation.

Lundberg, J ., and F. Moberg. 2003. Mobile link organisms and ecosystem functioning:
Implications for ecosystem resilience and management. Ecosystems 6(1):87-98.

Magurran, A. E. 2004. Measuring biological diversity. Blackwell Publishers, Malden,
MA. viii, 256 p.

Matthews, J. W., A. L. Peralta, D. N. Flanagan, P. M. Baldwin, A. Soni, A. D. Kent, and
A. G. Endress. 2009a. Relative inﬂuence of landscape vs. local factors on plant
community assembly in restored wetlands. Ecological Applications 19(8):2108-2123.

Matthews, J. W., A. L. Peralta, A. Soni, P. Baldwin, A. D. Kent, and A. G. Endress.
2009b. Local and landscape correlates of non-native species invasion in restored
wetlands. Ecography 32(6): 103 1-1039.

120

Matthews, J. W., G. Spyreas, and A. G. Endress. 2009c. Trajectories of vegetation-based
indicators used to assess wetland restoration progress. Ecological Applications
19(8):2093-2107.

Memmott, J ., N. M. Waser, and M. V. Price. 2004. Tolerance of pollination networks to
species extinctions. Proceedings of the Royal Society of London Series B-Biological
Sciences 271(1557):2605-2611.

MNFI. 2007. Rare Species Explorer.in M. N. F. Inventory, editor., Lansing, MI.

Nabhan, G. P., and S. L. Buchmann. 1997. Services provided by pollinators. Page 392 in
G. C. Daily, editor. Nature's Services : Societal Dependence on Natural Ecosystems.
Island Press, Washington, DC.

Nekola, J. C. 2004. Vascular plant compositional gradients within and between Iowa
fens. Journal of Vegetation Science 15(6):771-780.

New, T. R. 1997. Are lepidoptera an effective 'umbrella group' for biodiversity
conservation? Journal of Insect Conservation 1:5-12.

Nichols, O. G., and F. M. Nichols. 2003. Long-term trends in fauna] recolonization after
bauxite mining in the jarrah forest of southwestern Australia. Restoration Ecology
ll(3):26l-272.

Panzer, R., D. Stillwaugh, R. Gnaedinger, and G. Derkovitz. 1995. Prevalence of remnant
dependence among the prairie-inhabiting and savanna-inhabiting insects of the Chicago
region. Natural Areas Journal 15(2):101-l 16.

Petanidou, T., A. S. Kallimanis, J. Tzanopoulos, S. P. Sgardelis, and J. D. Pantis. 2008.
Long-term observation of a pollination network: ﬂuctuation in species and interactions,

relative invariance of network structure and implications for estimates of specialization.
Ecology Letters 11(6):564-575.

Pollard, E., D. O. Elias, M. J. Skelton, and J. A. Thomas. 1975. A method of assessing
the abundance of butterﬂies in Monks Wood National Nature Reserve in 1973.
Entomolists's Gazette 26:79-88.

Polley, H. W., J. D. Demer, and B. J. Wilsey. 2005. Patterns of plant species diversity in
remnant and restored tall grass prairies. Restoration Ecology l3(3):480-487.

Potts, S. G., J. C. Biesmeijer, C. Kremen, P. Neumann, O. Schweiger, and W. E. Kunin.
2010. Global pollinator declines: trends, impacts and drivers. Trends in Ecology &
Evolution 25(6):345-353.

Potts, S. G., B. Vulliamy, A. Dafni, G. Ne'eman, and P. Willmer. 2003. Linking bees and
ﬂowers: How do ﬂoral communities structure pollinator communities? Ecology
84(10):2628-2642.

121

waell, R. F., E. A. Warrnan, T. H. Sparks, J. N. Greatorex-Davies, K. J. Walker, W. R.
Meek, C. Carvell, S. Petit, and L. G. Firbank. 2004. Assessing habitat quality for
butterﬂies on intensively managed arable farmland. Biological Conservation 118(3):313-
325.

Quinn, G. P., and M. J. Keough. 2002. Experimental design and data analysis for
biologists. Cambridge University Press, Cambridge, UK ; New York. xvii, 537 p.

Ruiz-Jaen, M. C., and T. M. Aide. 2005. Restoration success: How is itbeing measured?
Restoration Ecology 13(3):569-577.

SAS Institute. 2010. SAS/STAT User's guide for personal computers. SAS Institute,
Cary, NC.

Satterthwaite, F. E. 1946. An approximate distribution of estimates of variance
components. Biometrics Bulletin 22110-1 14.

SER Working Group. 2004. The SER international primer on ecological restoration,
www.ser.org & Tucson: Society for Ecological Restoration International.

Shepherd, S., and D. M. Debinski. 2005. Evaluation of isolated and integrated prairie
reconstructions as habitat for prairie butterﬂies. Biological Conservation 126(1):5 1-61.

Spieles, J. B., P. J. Comer, D. A. Albert, and M. A. Kost. 1999. Natural community
abstract for prairie fen. Michigan Natural Features Inventory, Lansing, MI.

Steffan-Dewenter, I., U. Munzenberg, C. Burger, C. Thies, and T. Tschamtke. 2002.
Scale-dependent effects of landscape context on three pollinator guilds. Ecology
83(5):]421-1432.

Steffan-Dewenter, I., and T. Tschamtke. 1999. Effects of habitat isolation on pollinator
communities and seed set. Oecologia 121(3):432-440.

Steffan-Dewenter, 1., and T. T scharntke. 2000. Butterﬂy community structure in
fragmented habitats. Ecology Letters 3(5):449-456.

Summerville, K. S., R. M. Steichen, and M. N. Lewis. 2005. Restoring lepidopteran
communities to oak savannas: Contrasting inﬂuences of habitat quantity and quality.
Restoration Ecology 13(1):120-128.

Tuell, J. K., A. K. Fiedler, D. Landis, and R. Isaacs. 2008. Visitation by wild and
managed bees (Hymenoptera : Apoidea) to eastern US native plants for use in
conservation programs. Environmental Entomology 37(3):707-718.

Waltz, A. E. M., and W. W. Covington. 2004. Ecological restoration treatments increase
butterﬂy richness and abundance: Mechanisms of response. Restoration Ecology
12(1):85-96.

122

Waser, N. M., L. Chittka, M. V. Price, N. M. Williams, and J. Ollerton. 1996.
Generalization in pollination systems, and why it matters. Ecology 77 (4):]043-1060.

Wilcove, D. S., D. Rothstein, J. Dubow, A. Phillips, and E. Losos. 1998. Quantifying
threats to imperiled species in the United States. Bioscience 48(8):607-615.

Williams, N. M., R. L. Minckley, and F. A. Silveira. 2001. Variation in native bee faunas
and its implications for detecting community changes. Conservation Ecology 5(7).

Winfree, R. 2010. The conservation and restoration of wild bees. Annals NY Academy of
Sciences pre-print.

Winfree, R., R. Aguilar, D. P. Vazquez, G. LeBuhn, and M. A. Aizen. 2009. A meta-
analysis of bees' responses to anthropogenic disturbance. Ecology 90(8):2068-2076.

123

Table 8. Plant species identiﬁed at research site and the percent of quadrats each species
occupied in year 2 (2009), 177 total. Bolded values are species whose frequency varied
most by treatment, Inv=Invaded, Remo=Removal and Ref=Reference.

 

 

Family Genus species Common Name C~ Native Inv Remo Ref
Aceraceae Acer rubrum L. red maple 7 Y 27.78 37.04 29.63
Asteraceae Achillea millefolium L. common yarrow 0 N 1.85 1.85 0.00
Ranunculaceae Actaea rubra red baneberry 10 Y 1.85 0.00 0.00
Scrophulariaceae Agalinis purpurea (L.) Pennell purple false foxglove 6 Y 0.00 3.70 0.00
Rosaceae Agrimonia gryposepala Wallr. tall hairy agrimony 2 Y 0.00 0.00 3.70
Rosaceae Agrimonia parviﬂora Aiton harvestlice 7 Y 5.56 7.41 7.41
Poaceae Agrop yr on trachycaulum (L'nk) slender wheatgrass 8 Y 0.00 0.00 3.70
Malte ex H.F. Lewrs
Fabaceae :gp'gléarpaea bracteata a”) American hogpeanut 4 Y 12.96 12.96 14.81
Poaceae Andropogon gerardii \fitman big bluestem 5 Y 0.00 0.00 11.11
Fabaceae Apios americana Medik. groundnut 7 Y 3.70 11.11 0.00
Araceae Arisaema triphyllum (L.) Schott jack in the pulpit 4 Y 7.41 1.85 0.00
Asclepladaceae Asclepias incamata swamp milkweed 4 Y 1.85 12.96 25.93
Asclepiadaceae Asclepias syn’aca common milkweed 0 Y 1.85 0.00 0.00
Asteraceae Aster Ianceolatus Willd. white panicle aster 3 Y 0.00 5.56 55.56
Asteraceae Aster puniceus L. purplestem aster 8 Y 12.96 24.07 48.15
Asteraceae Aster umbellatus Mill. parasol whitetop 9 Y 1.85 0.00 7.41
Asteraceae Bidens coronata (L. ) Britton crowned beggarticks 9 Y 0.00 3.70 7.41
Asteraceae Bidens frondosa (L.) devil‘s beggartick 1 Y 25.93 27.78 0.00
Urticaceae Boehmen'a cylindrica (L.) Sw. false nettle 2 Y 35.19 50.00 29.63
Poaceae Bromus ciliatus L. fringed brome 10 Y 0.00 1.85 14.81
Poaceae Calamagrostis canadensis (Michx.) bluejoint 3 Y 0.00 1.85 51 .85
Ranunculaceae Caltha palustris L. yellow marsh marigold 5 Y 3.70 3.70 7.41
Convolvuiaceae Calyslegia sepium (L.) R. Br. hedge false bindweed 1 Y 0.00 1.85 11.11
Campanulaceae Campanula aparinoides Pursh marsh bellﬂower 8 Y 7.41 12.96 33.33
. Cardamine bulbosa Schreb. ex
Brassrcaceae Muhl.) Britton, Sterns( & Poggenb. bulbous bittercress 6 Y 0.00 11 .11 0.00
Cyperaceae Carex bebbii Olney ex Fernald Bebb's sedge 6 Y 1.85 16.67 0.00
Cyperaceae Carex cryptolepis Mack. northeastern sedge 9 Y 0.00 0.00 3.70
Cyperaceae Carex ﬂava L. yellow sedge 10 Y 0.00 0.00 7.41
Cyperaceae Carex granularis Muhl. ex Willd. limestone meadow sedge 4 Y 1.85 1.85 0.00
Cyperaceae Carex hystericina Muhl. ex Willd. bottlebrush sedge 5 Y 0.00 38.89 3.70
Cyperaceae Carex leplalea Wahlenb. bristlystalked sedge 10 Y 3.70 14.81 0.00
Cyperaceae Carex pellita Muhl. ex Willd. woolly sedge 4 Y 3.70 5.56 29.63
Cyperaceae _ 33$; ”“3”“ Dewey ‘3" Am” prairie sedge 10 Y 0.00 0.00 25.93
Cyperaceae - Carex sartwellii Dewey Sartwell's sedge 6 Y 0.00 0.00 14.81
Cyperaceae Carex sterilis Willd. dioecious sedge 10 Y 1.85 0.00 44.44
Cyperaceae Carex stipata Muhl. ex Willd. awlfruit sedge 3 Y 33.3 42.59 0.00
Cyperaceae Carex stricta Lam. upright sedge 5 Y 25.93 27.78 92.59
Cyperaceae Carex tetanica Schkuhr rigid sedge 9 Y 0.00 3.70 40.74
Cyperaceae Carex vulpinoidea Michx. fox sedge 2 Y 1.85 5.56 11.11
Juglandaceae Carya seedling plgnut or shagbark 5 Y 3.70 5.56 0.00
Celastraceae Celastrus orbiculatus Thumb oriental bittersweet O N 14.81 1.85 7.41
Apiaceae Cicuta maculata L. spotted water hemlock 6 Y 0.00 1.85 0.00
Poaceae Cinna arundinacea L. sweet woodreed 5 Y 9.26 7.41 3.70
Onagraceae Circaea Iutetiana L. :‘igzgﬁaasgmmme' s 1 Y seas 43.15 11.11
Asteraceae Cirsium arvense (L.) Scop. Canada thistle O N 0.00 9.26 0.00
Asteraceae Cirsium muticum Michx. swamp thistle 10 Y 7.41 14.81 51.85
Asteraceae Cirsium vulgare (Savi) Ten. bull thistle O N 0.00 1.85 0.00
Cyperaceae Cladium mariscoides (Muhl.) Torr. smooth sawgrass 10 Y 0.00 0.00 62.96
Ranunculaceae Clematis virginiana L. western blue vigginsbower 4 Y 0.00 3.70 0.00

 

124

Table 8 cont’d.

 

 

Family Genus species Common Name C~ Native inv Remo Ref
Asteraceae Conyza canadensis (L.) Cronquist Canadian horseweed 0 Y 0.00 3.70 0.00
Comaceae Cornus foemina Mill. stiff dogwood 1 Y 16.67 5.56 14.81
Comaceae Cornus stolonifera Michx. redosier dogwood 6 Y 1.85 7.41 37.04
Dryopteris carthusiana (\fill.) H.P.
Dryopteridaceae Fuchs spinulose woodfern 8 Y 7.41 3.70 0.00
Dryopteridaceae Dryopteris cristata (L) A Gray crested woodfem 10 Y 9.26 3.70 0.00
Cyperaceae Eleocharis elliptica Kunth elliptic spikerush 8 Y 0.00 1.85 40.74
Cyperaceae Eleochan's rostellata (Torr.) Torr. beaked spikerush 10 Y 0.00 9.26 3.70
Onagraceae Epilobium coloratum Biehler purpleleaf willowherb 3 Y 24.07 64.61 11.11
Equisetaceae Equisetum arvense (L.) ﬁeld horsetail 0 Y 27.78 37.04 11.11
Asteraceae g’cecmes "’e’ac’”°”a (L) Ra" 9" American burnweed 2 Y 18.52 29.63 0.00
Asteraceae Erigeron annuus ( L.) Pers. eastern daisy ﬂeabane 0 N 1.85 7.41 0.00
Asteraceae Erigeron philadelphicus L. Philadelphia ﬂeabane 4 Y 0.00 3.70 0.00
Asteraceae Eupatorium macalatum L. spotted joe pye weed 4 Y 5.56 18.52 55.56
Asteraceae Eupatorium perfo/iatum L. common boneset 4 Y 29.63 72.22 44.44
Asteraceae Euthamia graminifo/ia (L.) Nutt. ﬂat-top goldentop 4 Y 0.00 5.56 7.41
Rosaceae Fragan‘a virginiana Duchesne Virginia strawberry 1 Y 5.56 1.85 7.41
Rhamnaceae Frangula alnus Mill. glossy buckthom 0 N 79.63 79.63 48.15
Oleaceae Fraxinus pennsylvanica Marsh. green ash 5 Y 1.85 0.00 0.00
Rubiaceae Galium boreale L. northern bedstraw 7 Y 7.41 5.56 7.41
Rubiaceae Galium circaezans Michx. licorice bedstraw 10 Y 1.85 5.56 0.00
Rubiaceae Galium triﬂorum Michx. fragrant bedstraw 5 Y 18.52 16.67 7.41
Gentianaceae Gentianopsis crinita (Froel.) Ma greater fringed gentian 10 Y 0.00 0.00 16.52
Geraniaceae Geranium maculatum L. spotted geranium 4 Y 0.00 1.85 0.00
Rosaceae Geum canadense Jacq. white avens 1 Y 25.93 14.81 11.11
Poaceae Glyceria striata (Lam) Hitchc. fowl mannagrass 4 Y 31.48 57.41 48.15
Boraginaceae fﬂsetha wrgmrana (L') "M' beggarsiice 0 Y 1.85 1.85 0.00
Asteraceae Helenium autumnale L. common sneezeweed 5 Y 0.00 0.00 18.52
Asteraceae Helianthus giganteus L. giant sunﬂower 9 Y 0.00 1.85 0.00
Poaceae Hierochloe odorata (L.) P. Beauv. sweetgrass 9 Y 0.00 5.56 3.70
Liliaceae Hypoxis hirsute (L.) Coville common goldstar 9 Y 0.00 0.00 46.15
Aquifoliaceae llex verticillata (L.) A. Gray common winterberry 9 Y 3.70 1.85 0.00
Balsaminaceae Impatiens capensis Meerb. jewelweed 3 Y 22.2 24.07 22.22
lridaceae Iris versicolor L. harlequin blueﬂag 5 Y 0.00 0.00 3.70
Juncaceae Juncus b'aChycepha’us (Enge'm') smallhead rush 9 Y 0.00 0.00 11.11
Buchenau
Juncaceae Juncus dud/eyi Wiegand Dudley's rush 4 Y 1.85 18.52 7.41
Cupressaceae Juniperus virginiana L. eastern redcedar 2 Y 3.70 0.00 0.00
Fabaceae Lathyrus palustris L. marsh pea 8 Y 5.56 16.67 22.22
Poaceae Leersia oryzoides (L.) Sw. rice cut grass 4 Y 16.67 36.69 7.41
Asteraceae Liatris spicata (L.) Willd. dense blazing star 6 Y 0.00 0.00 11.11
Lauraceae Lindera benzoin (L) Blume Northern spicebush 7 Y 18.52 7.41 0.00
Campanulaceae Lobelia kalmii L. Ontario lobelia 10 Y 0.00 0.00 11.11
Campanulaceae Lobe/fa siphilitica L. great blue lobelia 6 Y 9.26 11.11 0.00
Caprifoliaceae Lonicera L. honeysuckle 0 N 0.00 1.85 0.00
Lamiaceae 32%;? americanus Muhl. ex W. American water horehound 5 Y 11.11 12.96 66.67
Lamiaceae Lycopus uniﬂorus Michx. northern bugleweed 7 Y 18.52 33.33 55.56
Primulaceae Lysimachia quadriﬂora Sims fourﬂower yellow ioosestrife 9 Y 0.00 3.70 46.15
Lythraceae Lythrum salicaria (L.) purple Ioosestrife 0 N 5.56 24.07 18.52
Lamiaceae Mentha arvensis L. wild mint 5 Y 3.70 5.56 22.22
Poaceae Muhlenbergia glomerata (Willd.) Tri: spiked muhly 10 Y 0.00 0.00 46.15
Poaceae Muhlenbergia mexicana Mexican muhly 5 Y 7.41 18.52 22.22
Dryopteridaceae Onoclea sensibilis L. Sensitive fern 8 Y 27.78 12.96 11.11

 

125

Table 8 cont’d.

 

 

Family Genus species _ Common Name C~ Native lnv Remo Ref
Apiaceae 8mm” damn” (M'Chx') 03' Clayton's sweetroot 3 Y 7.41 3.70 0.00
Apiaceae Oxypolis rigidor (L.) Raf. stiff cowbane 7 Y 5.56 3.70 14.81
Poaceae Panicum capillare L. witchgrass 1 Y 0.00 1.85 0.00
Saxifragaceae Parnassia glauca grass of pamassus 10 Y 0.00 0.00 37.04
Vitaceae Srgfm’ssus q“’"°“e'°”a (L') Virginia creeper 2 Y 27.78 29.63 22.22
Scrophulariaceae Pedicuian‘s lanceolata Michx. swamp iousewort 9 Y 0.00 0.00 7.41
Crassuiaceae Penthorum sedoides L. ditch stonecrop 5 Y 5.56 16.67 0.00
Poaceae Phalaris arundinacea L. reed canarygrass 0 N 16.67 25.93 18.52
Phytolaccaceae Phytolacca americana L. American pokeweed 1 Y 1.85 0.00 0.00
Urticaceae Pilea pumila (L.) A. Gray clearweed 5 Y 68.52 61.46 18.52
Plantaginaceae Plantago major (L.) common plantain O N 0.00 0.00 3.70
Poaceae Poa praiensis L. Kentucky bluegrass O N 1.85 1.85 3.70
Polygonaceae Polygonum amphibium L. water knotweed 4 Y 3.70 11.11 55.56
Polygonaceae Segomm by drop me roides swamp smartweed 7 Y 0.00 1.85 0.00
Polygonaceae Polygonum punctatum Elliot dotted smartweed 6 Y 11.11 12.96 3.70
Polygonaceae Polygonum sagittatum L. arrowleaf tearthumb 8 Y 14.81 27.78 7.41
Polygonaceae Polygonum virginianum L. jumpseed 2 Y 7.41 5.56 0.00
Salicaceae Populus tremuloides Michx. quaking aspen 4 Y 5.56 24.07 7.41
Rosaceae Potentilla fruticosa auct. non L. shrubby cinquefoil 10 Y 0.00 0.00 62.96
Lamiaceae Prune/Ia vulgaris L. selﬂteal O Y 0.00 3.70 0.00
Rosaceae Prunus serotina Ehrh. black cherry 1 Y 24.07 12.96 3.70
Lamiaceae gﬁféznéégfng’c’kg'gﬂ’gi‘f’;0‘:)J' ngnia mountainmint 5 Y 1.85 0.00 14.81
Fagaceae Quercus sp. oak 5 Y 3.70 3.70 0.00
Ranunculaceae Ranunculus recurvatus Poir. blisterwort 5 Y 12.96 0.00 0.00
Rhamnaceae Rhamnus alnifolia L'Her. alderleaf buckthom 10 Y 0.00 0.00 3.70
Rhamnaceae Rhamnus cathartica L. common buckthom O N 18.52 9.26 7.41
Cyperaceae Rhynchospora alba (L.) Vahl white beaksedge 10 Y 0.00 0.00 7.41
Grossulariaceae Ribes cynosbati L. eastern prickly gooseberry 5 Y 5.56 3.70 0.00
Brassicaceae Rorippa palustris (L.) Besser bog yellowcress 4 Y 5.56 3.70 0.00
Rosaceae Rosa multiﬂora Thunb. multiﬂora rose 0 N 27.76 7.41 3.70
Rosaceae Rosa palustris Marsh. swamp rose 7 Y 1.85 5.56 3.70
Rosaceae Rubus pubescens Raf. dwarf red blackberry 10 Y 20.37 3.70 3.70
Rosaceae Rubus strigosus Michx. grayleaf red raspberry 3 Y 16.67 18.52 14.81
Asteraceae Rudbeckia hirta L. blackeyed Susan 1 Y 0.00 1.85 40.74
Polygonaceae Rumex orbiculatus A. Gray greater water dock 8 Y 1.85 3.70 3.70
Saliwceae Salix discolor Muhl. pussy willow 2 Y 0.00 38.89 33.33
Caprifoliaceae Sambucus canadensis L. American black elderberry 1 Y 9.26 3.70 3.70
Apiaceae Sanicula gregarta black snakeroot 2 Y 0.00 12.96 0.00
Lauraceae Sassafras albidum (Nutt.) Nees sassafras 3 Y 0.00 1.85 0.00
Poaceae ﬁggﬁachym’" swpar’u’" (M'Cm little bluestem 5 Y 0.00 0.00 37.04
Cyperaceae Scirpus acutus Muhl. ex Bigelow hardstem bullrush 6 Y 0.00 0.00 33.33
Cyperaceae Scirpus americanus Pers. threesquare 5 Y 0.00 0.00 33.33
Cyperaceae Scirpus atrovirens Willd. green bulrush 4 Y 0.00 27.78 0.00
Lamiaceae Scutellaria galericulata L. marsh skullcap 5 Y 3.70 20.37 11.11
Lamiaceae Scutellaria Iateriﬂora L. blue skullcap 5 Y 3.70 1.85 3.70

 

126

Table 8 cont’d.

 

 

Family Genus species Common Name C~ Native lnv Remo Ref
Asteraceae Senecio aureus L. golden ragwort 7 Y 38.89 51.65 18.52
Asteraceae Senecio pauperculus Michx. balsam groundsel 6 Y 0.00 0.00 33.33
Liliaceae Smilaclna stellata (L.) Desf. starry false lily of the valley 5 Y 5.56 3.70 0.00
Smilacaceae Smilax tamnoides L. bristly greenbrier 5 Y 12.96 1.85 0.00
Solanaceae Solanum dulcamara L. climbing nightshade O Y 11.11 29.63 0.00
Asteraceae Solidago altissima L. Canada goldenrod 1 Y 1.85 9.26 11.11
Asteraceae Solidago gigantea Aiton giant goldenrod 4 Y 3.70 3.70 11.11
Asteraceae Egg?” ohioenSIs Frank ex Ohio goldenrod 9 Y 0.00 0.00 33.33
Asteraceae Solidago patuia Muhl. ex Willd. roundleaf goldenrod 9 Y 35.19 37.04 33.33
Asteraceae Solidago riddellii Frank ex Riddell Riddell's goldenrod 7 Y 1.85 1.85 37.04
Asteraceae Solidago rugosa Mill. rough-Ieaved goldenrod 6 Y 35.19 31.48 18.52
Poaceae Sorghastrum nutans (L.) Nash indiangrass 5 Y 0.00 0.00 33.33
Poaceae Sphenopholis intermedia (Rydb.) R) slender wedgescale 4 Y 0.00 0.00 7.41
Rosaceae Spiraea alba Du Roi white meadowsweet 7 Y 0.00 0.00 3.70
Araceae :{ﬁﬁ’t‘t’ca’pus ”ems (L) 38"5‘" American skunk cabbage 8 Y 24.07 14.81 3.70
Asteraceae Taraxacum oﬁicinale F.H. Wigg. common dandelion 0 N 0.00 5.56 0.00
Ranunculaceae 13:13:," dasy carp um F'SCh' 8‘ purple meadow-rue 5 Y 5.56 3.70 11.11
Thelypteridaceae Thelypteris palustris Schott eastern marsh fern 6 Y 14.81 9.26 70.37
Anacardiaceae ;Zx‘gdendron radicans (L') Eastern poison ivy 2 Y 3.70 3.70 3.70
Anacardiaceae Toxicodendron vemix (L) Kuntze poison sumac 10 Y 1.85 5.56 0.00
Fraser's marsh St.
Clusiaceae Tn'adenum fraseri (Spach) Gleason Johnswort 8 Y 0.00 0.00 3.70
Fabaceae Trifolium repens L. white clover O N 0.00 1.85 0.00
Typhaceae Typha latifolia L. broadleaf cattail 1 N 3.70 20.37 7.41
Ulmaceae Ulmus sp. elm 3 Y 24.07 27.78 3.70
Urticaceae Urtica dioica L. stinging nettle 0 N 5.56 12.96 0.00
Verbenaceae Verbena hastata L. swamp verbena 4 Y 0.00 5.56 0.00
Verbenaceae Verbena urticifolia L. white vervain 5 Y 0.00 3.70 0.00
Violaceae Viola nephrophylla Greene northern bog violet 9 Y 50.00 35.19 7.41
Wtaceae Vitis riparia Michx. riverbank grape 2 Y 5.56 7.41 0.00
Liliaceae Zigadenus elegans Pursh ssp. mountain deathcamas 10 Y 0.00 0.00 16.52

glaucus (Nutt.) Hultén

~Coefﬁcient of conservatism from Herman et al. (2001)

127

Table 9. Number of bees observed using observational sampling per replicate and

treatment in years 1 and 2 post restoration. The percent of bees each genus composed of
the total number observed is in the ﬁrst column for years 1 and 2. Bolded values are the
most common species whose abundance varied most by treatment. Dashes indicate no
insects of that species were seen in a given year. Values are by replicate and treatment,

averaged over sample dates.

 

 

 

 

Family Genus % obs Invaded Removal Reference % obs Invaded Removal Reference
Andrenidae
Andrena 0 1 .8 0 0.25
Apidae
Apis mellifera 30.5 36.8 3.90 1.17
Anthophora 0 0.3 0.04 0
Bombus 21.2 12.3 1.33 0.25
X ylocopa 0 0.9 0.1 3 0
Colletidae
Ceratina 0 7.1 0.52 0.46
Halictidae
Hylaeus 16.9 15.6 1.83 0.29
Agapostemon 0.6 0.08 0
Augochlora 0 5.9 0.65 0.1 7
Augochlorella 1 .7 7.0 0.90 0.04
Halictus 0 1 .7 0.06 0.1 7
Lasioglossum 2.5 7.1 0.69 0.29
other Haiictidae" 23.7 0 - -
Megachilidae
Heriades 0 1 .5 0 ' 0.04 0.1 7
Hoplitus 0 0.2 0 0.02 0
Megachile O 0.6 0 0 0.08
Osmia 0 0.2 0 0.02 0
Perdita 0 0 0 0 0
* Other 3.4 0.6 . 2 0.02 0.04
Genus richness '7 17 4 15 12

 

*Represents bees that were not identiﬁed to genus; may be composed of more than 1

genus and family.

 

Table 10. Number of bees collected using pollinator bowl sampling per replicate and
treatment, Inv=Invaded, Remo=Removal and Ref=Reference, in years 1 and 2 post
restoration. The percent of bees each species composed of the total number trapped is in
the ﬁrst column for years 1 and 2. Bolded values are the most common species whose
abundance varied most by treatment. Dashes indicate no insects of that species were seen
in a given year. Values are by replicate and treatment, averaged over sample dates.

 

 

 

Year 1 Year 2
Family Genus species % trapped lnv Remo Ref 96 trapped Inv Remo Ref
Andrenidae
Andrena allegheniensis Viereck 0 - - - 0.2 0 0.04 0
A. car/ini Cockereli 0 - - - 0.2 0 0.04 0
A. cressonii Robertson 0 - - - 0.6 0 0.04 0.08
A. nasonii Robertson 0 - - - 0.6 0 0.04 0.08
A. perplexa Smith ‘ 0.6 0 0.06 0 0 - - -
Apidae
Anthophora (ermine/is Cresson 0.6 0.06 0 0 0.6 0.04 0 0.08
A. ursina Cresson O - - - 0.2 0 0.04 0
Apis mellifera L. 3.2 0 0.06 0.22 2.0 0 0.17 0.25
Bombus impatiens Cresson 0 - - - 0.2 0 0.04 0
B. vagans Smith 0 - - — 0.2 0 0.04 0
Ceratina calcarata/dupla a 13.2 0.11 1.00 0.44 19.9 0.33 1.93 1.92
C. strenua Smith 1.9 0 0.06 0.11 1.6 0 0.17 0.17
Colletidae
Hylaeus afﬁnus Smith 0 — - - 0.2 0 0.04 0
Hylaeus sp. 1 5.8 0 0.17 0.33 3.9 0.04 0.33 0.42
Hylaeus sp. 3 0 - - - 0.6 0 0.04 0.08
Halictidae
Agapostemon sericeus (Forster) 0 - - - 0.4 0 0 0.08
Ag. virescens (F.) 1.3 0 0.11 0 0.0 - - -
Augochlora pura (Say) 8.4 0.06 0.56 0.11 0.4 0.04 0.04 0
Aug. aurata (Smith) 7.8 0.06 0.50 0.11 11.2 0.08 1.54 0.67
Augochloropsis metal/ica (F .) 0 - - - 0.4 0 0.08 0
Halictus confusus Smith 1.3 0 0.11 0 4.5 0 0.33 0.58
Halictus Iigatus Say 1.3 0 0.11 0 1.0 0 0.04 0.17
Lasioglossum
atlanticum (Mitchell) 3.9 O 0.11 0.22 0.2 0 0.04 0
L. bruneri (Crawford) 0 - - - 0.2 0.04 0 0
L. coriaceum (Smith) 5.2 0.28 0.06 0.11 1.4 0.13 0.08 0.08
L. cressonii (Robertson) 0.6 0 0.06 0 0.6 0.04 0.08 0
L. Dialictus sp. 1 4.5 0 0.17 0.22 0.8 0.13 0.04 0
L. Dialictus sp. 2 11.0 0 0.06 0.69 14.0 0.42 0.96 1.50
L. Dialictus sp. 3 0 - - - 3.3 0.13 0.21 0.33
L. Dialictus sp. 4 0 - - - 0.2 0 0.04 0
L. Dialictus sp. 5 0 - - - 0.4 0 0 0.08
L. Dialictus sp. 6 0 - - - 0.2 0.04 0 0
L. Dialictus sp. 7 0 - - - 0.2 0.04 0 0
L. Dialictus sp. 8 0 - - - 0.4 0 0 0.08
L. Dialictus spp. 1.9 0 0.17 O 1.2 0.04 0.04 0.17
L. divergens (Lovell) 0.6 0.06 0 0 0 - - -
L. illinoense (Robertson) 0 - - - 0.2 0.04 0 0
L. leucozonium (Schrank) 1.9 0 0 17 0 0.2 0 0.04 0
L. macoupinense (Robertson) 0 - - - 0.2 0.04 O 0
L. nelumbonis (Robertson) 0.6 0 0.06 0 0 - - -
L. nymphaearum (Robertson) 0 - - - 0.4 0 0.08 0

 

129

Table 10 cont’d.

 

 

 

 

Year 1 Year 2
Family Genus species % trapped Inv Remo Ref 16 trapped lnv Remo Ref
Halictidae
L. paradmirandum (Knerer 8.
AI l) 1 .3 0 0.1 1 0 0 - - -
L. pectorale (Smith) 2.6 0 0.11 0.11 1.0 0 0.13 0.08
L. pilosum (Smith) 1.3 0 0 0.11 0.8 0 0 0.17
L. rohwen’ (Ellis) 5.2 0 0.22 0.22 21 .1 0.04 2.63 1.67
L. versans (Lovell) 3.2 0.06 0.22 0 0 - - -
Megachilidae
Hop/itis producta (Cresson) 0 - - - 0.6 0 0.13 0
H. spoliata (Provancher) 0 - - - 0.4 0 0.08 0
Megachile campanulae
(Robertson) 0 - - - 0.4 0 O 0.08
M. inermis Provancher 0 - - - 0.2 0 0.04 0
M. montivaga Cresson O - - - 0.2 0 0.04 0
M. pugnata Say 0.6 0 0.06 0 0 - - -
M. relativa Cresson 0.6 0 0.06 0 0 - - -
Osmia georgica Cresson 0 - - - 0.6 0 0.13 0
O. michiganensis Mitchell 0 - - - 0.6 0 0.04 0.08
O. pumila Cresson 2.6 0 0.22 0 0.2 0 0.04 0
O. similima Smith 1.3 0 0 0.11 0 - - -
Species richness 29 7 25 14 48 17 37 23

 

afemales of these species, C. calcarata Robertson and C. dupla Say, are morphologically

indistinguishable; only 1 male of each species was collected.

130

 

Table 11. Number and identity of butterﬂies seen during 5 minute observational sampling

in years 1 and 2 post restoration, % obs=percent of total observed per year, Inv=Invaded,

Removal and Ref=Reference. Bolded values are the most abundant species whose

abundance varied by treatment. Dashes indicate no insects of that species were seen in a
given year. Values are by replicate and treatment, averaged over sample dates.

Remo=

 

 

 

 

: I
m m w 9 m m o o 98:50: 363$
96 0 mod No - - - 0 5mm coo; 25. A5829 Sonic 96.502
e858
- - - o «a... 36 o New 223 8893 $3035.: 0398 areal
canted
8.0 cod o to :e 8e o m. : .5535 538.8 3885: 85a cause
0 mod 0 On 0 mod 0 Qm .6323?» Leo: ﬁnance; «3:63 0:..ch
omn_co___amn_
mod 0 o mam - - - o 6m. 3:3 Awacmccﬁv .328 cmwocm>
mod «to o mom :5 :d o #2 agate 8.98... see $22.8“: sense 28.8 seesaw
«16 mod o mdm - - - o E888 bacon $.29 8.65 «33.931
vod o 0 NF - - - o BEEF. 3.623 62?. €3.89 3th 3:23 mtoBm
8235.52
o sod o 5 - - - o sane Lassen 32953 1.5 sesame: 8.538
omEccwoi
and o O Q? 0 mod o Wm 83% ac? >5$3.2: Couczowv «oneness: meccol
o mod 0 ad - - - o Lonnie ononc; 3:55 «0530: mocmol
o n to o Nm - - - o 53% 8:03 526 coEEov «Ego machLmqm
2.6 3.0 o Em 0 mod 0 ad .5336 ammo. 322.com: 82.25:: mcqxxoiosx
mod mod 0 2. Ed mod 0 m. S Lonnie. noﬁEcEc: .I
82:88:
- - - o o mod o Qm 55:08 $339.51: 33963 2655
cmamcmo
3m. oEom >5 ago 3 new. 063. >5 one .x. 2:6: 56:30 302.» 3:00 2.25.
N 30> F as;

 

 

131

 

40’A

- Invaded
[2:] Removal
- Reference

 

i—_r.>

 

 

 

30

 

1—1>

 

10

 

 

°/o forb cover/ m2, season-long maximum
N
O

 

 

 

 

 

 

 

 

 

30

25

HQ

 

 

15L

1

1O

 

 

 

 

 

 

 

 

 

 

 

 

Season-long plant Simpsons diversity/ m2
o

0 . .
Year 1 Year 2

Figure 8. Comparisons of A) percent cover of forbs and B) Simpsons diversity between
F. alnus invaded, removal, and uninvaded plots in the ﬁrst and second years following
restoration. Maximum values for percent cover recorded for each growing season are
used. Error bars are +SEM, ci=0.05. Treatment effects from a 2-way AN OVA with
treatment and time are shown; year effects were not signiﬁcant.

132

 

1 '6‘

 

 

 

 

 

 

 

 

Year 1
A O A A
I O O
I O A
I
A
O A
O
A invaded
0 Removal
I Reference
N
(D
E Year 2
A
A
A
I A ‘
I
I 0 o 0
o
A
0
o

 

 

 

Axus 1
Figure 9. Two dimensional NMDS ordinations of the plant community show no
difference between invaded and removal plots in year 1 but a shift in removal plots away
from invaded plots by year 2. The ordination is based a Bray-Curtis dissimilarity matrix
for the ﬁrst and second years following restoration. Maximum values for percent cover
recorded for each growing season are used, data were square root transformed. Stress
values were 0.08 and 0.09 in years 1 and 2.

133

 

 

 

A a - invaded
12 - Removal
I: Reference

 

 

 

Bee abundance / 5 minute observation

 

 

 

 

 

 

 

 

 

 

Season-long bee diversity (Simpsons Index)

Year 1 Year 2

Figure 10. Comparisons of A) bee abundance B) bee diversity C) butterﬂy abundance and
D) butterﬂy diversity in the ﬁrst and second seasons following restoration. Bee
abundance, butterﬂy abundance, and diversity are based on the season-long mean number
of individuals observed in S-minute observational sampling periods in invaded, removal,
on-site reference (reference) and off-site reference (off-site). Bee diversity is based on
bees collected with bowl sampling in invaded, removal, and on-site reference plots.
Treatment effects from 2-way AN OVAs with treatment and time are shown. Letters in
upper and lower case within the same graph indicate that treatments differed by year or
treatrnent*year.

134

 

 

 

 

1: Reference

 

 

 

C - Invaded
Removal

 

 

 

F

—

 

 

 

 

 

 

 

 

 

- p - — l? — —

 

5. o.
2 2

5

1

0.

1

5.
0

0.
0

commZomno 935:. m \ 8:853 3:325

0.50.50.50.50.
433221100

coszewno 355:. m \ meLezzn o mcomaém

Year 2

Year 1

Figure 10 cont’d.

135

 

 

 

 

 

 

 

 

 

Year 1 0 .
O A
A
A
O A
I A
0 I A
A Invaded 0
0 Removal
f 0
N I Re erence ‘
i
<15 Year 2 A
I
o ‘ ‘
O
I
A
l A
O
O O
A

 

 

 

Axus 1
Figure 11. Two dimensional NMDS ordinations of the bee community show distinct bee
communities in invaded plots in both years 1 and 2, with bee communities in removal and
reference plots overlapping. The ordination is based a Bray-Curtis dissimilarity matrix
(using n+1), using mean values per replicate and treatment; data were square root
transformed. Stress values were 0.14 in both years.

136

 

Year 1

 

 

A Invaded I
0 Removal
I Reference

 

 

 

 

 

 

 

 

Year 2 A

Axis 2

 

 

 

AXIS 1
Figure 12. Two dimensional NMDS ordinations of the butterﬂy community show that the
invaded and removal butterﬂy communities differ in both years 1 and 2. In year 2, all six
replicates of invaded contained 0 butterﬂies, and all points overlap. In year 2, 1 butterﬂy
was seen in invaded plots and 5 of 6 points overlap. The ordination is based a Bray-Curtis
dissimilarity matrix (using n+1), with mean values per replicate and treatment; data were
square root transformed. Stress values were 0.04 for year 1 and 0.11 for year 2.

137

Chapter 5
For Oecologia

RAPID RESTORATION OF PLANT-POLLINATOR NETWORKS AND FUNCTION
FOLLOWING HABITAT RESTORATION

Anna Katherine Fiedler

138

 

Abstract

Species diversity is often used as the primary indicator of successful habitat
restoration; however, it is also critical that species interactions for long—term ecosystem
persistence are restored. Here, we determine whether plant—pollinator networks and the
key ecosystem function of pollination are restored following invasive species removal
from a degraded prairie fen wetland. We examined 1) quantitative food webs describing
plant—pollinator interactions, 2) pollinator diversity and abundance, and 3) pollinator
function using sentinel plant species, Asclepias incarnata, in invaded, uninvaded
reference, and restored plots, following invasive species removal. Plant-pollinator
networks in invaded plots were highly depauperate, with signiﬁcantly lower plant and
pollinator species richness than restored or reference plots. In contrast, network
connectance, compartment diversity, generality, and vulnerability did not differ between
restored and reference plots, with marginally higher interaction evenness in reference
than restored plots. Moreover, pollinator function was restored in cleared areas in the
initial growing season following restoration. We found no signiﬁcant differences in
pollinator abundance or diversity between cleared and reference areas, while these
remained low in uncleared areas. In the second year following restoration, we found no
signiﬁcant differences in pollinator function or abundance, potentially due to spillover
effects of restoration or stochastic pollination by year. Pairing plant—pollinator networks
with measures of pollination shows that, although plant and insect community
composition are not yet fully restored, function has been. Consideration of ecosystem
function and species interactions in combination provide a more process—based analysis

that can inform both ecological theory and restoration.

139

 

Introduction

Habitat loss is the number one cause of species extinctions in North America
(Wilcove et al. 1998), as well as the leading cause of pollinator extinctions globally
(Brown and Paxton 2009). Drivers of global environmental change, including land use,
elevated C02, and biotic invasions, lead not only to species extinctions but also changes
in species interactions within ecosystems (Tylianakis et al. 2008). Plant—pollinator I;
interactions are of particular concern as evidence mounts that pollinators are under threat
(Potts et al. 2010, Winfree 2010). Over 90% of angiosperms are animal—pollinated

(N abhan and Buchmann 1997), and the fossil record indicates that major plant extinctions

 

IT

in the past led to a decline in insect diversity (Bascompte and Jordano 2007). Recent
documentation of concomitant declines in both pollinators and insect—pollinated plants
(Biesmeijer et al. 2006) has reinforced the need to address plant and pollinator
populations in tandem.

Construction of quantitative plant—pollinator networks is one way to assess the
structure, and stability, of plant—pollinator interactions. Quantitative interaction networks
are increasingly used to examine similarities and differences between species interactions
across taxa and trophic levels (Bascompte and J ordano 2007 ). This characterization of
networks has resulted in the identiﬁcation of recurring patterns which facilitate species
persistence and long-term maintenance of biodiversity (Bascompte et a1. 2006). Such
patterns include: 1) asymmetry, where specialist insects tend to visit generalist plants and
specialist plants tend to be visited by generalist insects, 2) nestedness, with specialist
species interacting with a subset of those in generalist interactions and 3) heterogeneity,
with the majority of species linked in few interactions, but a few species more connected

than expected by chance (Bascompte and J ordano 2007). Due to the prevalence of
140

generalist interactions between pollinators and plants (W aser et al. 1996, Olesen et al.
2008), plant—pollinator networks are likely to be relatively stable over time. However, if
generalist pollinator species are lost, additional plant extinctions are more likely than if
specialist pollinators become locally extinct (Memmott et a1. 2004).

An increasingly common solution to habitat degradation is to restore fragmented
habitats, removing invasive species and reintroducing historic disturbance regimes. The
plant community is commonly the focus of restoration (Young et al. 2005), with species
diversity the primary metric of success (Ruiz-Jaen and Aide 2005). Diversity metrics,
however, cannot assess whether recovery of species interactions and ecosystem function
have occurred. In contrast, a quantitative evaluation of species interactions could reveal
such patterns and help determine. Ecosystem function is rarely accounted for in
restoration (but see: Lomov et a1. 2009, Lomov 2010, in press), and structure and
function may not be on the same trajectory.

Whether species network properties in restored systems are comparable to those
in intact ecosystems has been examined in only a handful of cases (Forup and Memmott
2005, Forup et al. 2008). Species interactions were found to be more complex in ancient
than restored areas in heathlands with low plant diversity (Forup et al. 2008). In old ﬁeld
hay meadows, there were no differences in insect richness or abundance between old and
restored meadows, but there was a greater proportion of potential species links in old than
restored meadows (F orup and Memmott 2005). In both cases, not all differences in
species identity between restored and reference sites that resulted in differences in

network structure.

141

 

Over 60% of plant species are dependent on pollination by insects for maximal
fruit and seed set (Burd 1994, Ashman et a1. 2004). The removal of obligate pollinators
from an ecosystem may result in populations which persist but cannot reproduce
(Eriksson 2000). Plants may be pollen limited over long timeframes due to random
ﬂuctuations in populations and weather, plant adaptation to stochastic pollinator
communities, or habitat alteration (Ashman et a1. 2004). Invasive plants may also lead to
increased competition for pollinators (Bartomeus et al. 2008) and affect pollination rates
at related native species (Brown et al. 2002). Assessing pollen limitation at the population

level, however, is extremely difﬁcult due to plant reproductive tradeoffs. For example, r

 

many self—incompatible plants compensate for dependence on pollination by reproducing E;
vegetatively, rather than depending exclusively on seeds for reproduction (Bond 1994).
However, in systems without this compensation the risk for plant extinction is high (Bond
1994). Habitat fragmentation can negatively effect pollination and plant reproduction,
with self—incompatible species more negatively affected by ﬁ'agmentation (Aguilar et al.
2006). In degraded, fragmented ecosystems, the soil seed bank is often lost and seed
dispersal is limited, so that plant reproduction is entirely dependent on the species that
remain or are reintroduced (Bakker and Berendse 1999). Given this overall pattern and
the prevalence of habitat fragmentation worldwide, it is vital that we gain a better
understanding of whether pollinator function is re—established following habitat
restoration and how function relates to plant—pollinator network structure.

Pollen transport webs assessing the likelihood that ﬂower—visiting insects are
potential pollinators have been previously studied (Forup and Memmott 2005, Forup et

a1. 2008). However, the question of whether network properties and pollen transport webs

142

actually translate into the key ecosystem function of pollination has never been explored
in a restoration context. Our goal was to document the effect of a non—native invasive
plant on plant—pollinator networks in prairie fen wetlands (Chapters 2—4). Here we
examine the effect of restoration on plant-pollinator networks, the function of

pollination, and their interactions.

Methods

Study system and experimental design

 

'1‘.“ war. ‘aﬂJ'n

This study took place in prairie fen, a globally rare groundwater fed wetland
ecosystem with high species density (N ekola 1994, Amon et al. 2002). Parts of the
glaciated Midwestern US, where prairie fens have the highest density on Earth, fens
contain from 7— 18% of the rare taxa occurring in their respective states. Threats to
prairie fen include habitat fragmentation, invasive species encroachment, hydrological
changes, and pollution (Spieles et al. 1999, Bedford and Godwin 2003). The primary
invasive species in North American prairie fens is F rangula alnus L. (glossy buckthom).
This species can form a 7 m tall closed—canopy system that alters the typical sedge-
meadow nature of prairie fen (Sinclair and Catling 1999). The conservation and
restoration of fens is paramount to survival of the species within them, as well as the
water ﬁltration and ﬂood control services these wetlands provide.

The study sites were two prairie fens in Jackson County, Liberty Township,
Michigan: the Michigan State University MacCready Reserve in Clarklake, MI, and the
Grand River Fen, in Liberty MI. At the MacCready Reserve, we established six replicate
plots containing mature F. alnus plants (invaded), six replicate plots where mature F.

143

alnus plants were removed in February 2008 (restored), and three replicate uninvaded
plots (on—site reference) (Fiedler and Landis in review). Four replicate plots separated by
a minimum of 15m were also laid out at the Grand River Fen as a larger, contiguous

reference area (off—site reference). All plots measured approximately 25x25 m.

Sampling
Pollinator networks

We established four circular sampling areas of 2 m diameter on a grid within each

 

replicate. We sampled for plant-pollinator interactions every three weeks throughout the

1m ~ 1.;

2009 growing season, on 15 June, 6 July, 27 July, 19 August, and 1 September. Sampling
was performed between 1000 and 1600 EST on clear, calm, sunny days, with the initial
replicate for sampling rotated on each date. All ﬂowering forbs within each sampling area
were identiﬁed, and the number of ﬂoral units counted using the method of Dicks et al.
(2002). Upon arrival at a sampling plot, we waited one minute for insects to resume
normal activity, then counted and identiﬁed all ﬂower—visiting arthropods during a 5—
minute period, noting the plant species they visited. Flower—visiting species were
collected using a Bioquip Insect Vac (Rancho Dominguez, CA) or visually identiﬁed to
species in the case of Apis mellifera, Agapostemon sericeus, Ceratina calcarata/dupla,
Augochlorella aurata, Dolz'chovespula maculata, Hylaeus sp., T oxomerus sp.,
Gymnosoma occidua, Plagiognathus politis, Phymata pennsylvanica, and Lepidoptera.
All identiﬁed species were subsequently included in plant—pollinator networks.
Pollinator function

In April 2008, we tested three species for their potential use as pollination sentinel
plants: Asclepias incarnata L., LobeIia siphilitica, and Eupatorium perfoliatum. All three

144

species are native to Michigan prairie fens, are known to be attractive to a number of
pollinators (Tuell et al. 2008), and have short lifespans which allow successful
propagation and reliable ﬂowering. To test for pollinator dependency, we grew each
species obtained as plant plugs (W ildtype Native Plant Nursery, Mason MI) in a
greenhouse, watered with 100 ppm N (Peters Professional 20-20-20, Scotts, Maryville,
OH) and 24/22 degrees C day/night and 20/4 hours light/dark. When plants were
ﬂowering, we placed them into a 2x2 m mesh cage in the greenhouse, 3 without and 3
with mesh bags over inﬂorescences to exclude pollinators. We placed a “research mini”

hive of Bombus impatiens (Koppert, Romulus, MI) in the cage with ﬂowering plants of

 

A. incarnata, L. siphilitica and E. perfoliatum for 17, 22, and 35 days, respectively,
during April-May 2008. We then allowed plants to develop, and counted seeds and fruits
after ﬁ'uits matured. We found no signiﬁcant differences using a 2-tailed, paired t-test in
the number of seeds per ﬂower on L. siphilitica (P=0.556) or E. perfoliatum with versus
without pollination, but did ﬁnd signiﬁcant differences in A. incarnata (P=0.012). We
therefore used paired sentinel A. incarnata L. plants to assess the level of pollination in
the ﬁeld. There is some controversy regarding the degree to which A. incarnata is a self—
incompatible species (Lipow and Wyatt 2000), but at minimum pollinia must be
transferred within a ﬂower for pollination to occur, so insects are key to pollination of
this species (Judd 2002).

The ﬁeld study was carried out in 2008 and 2009 using A. incarnata plants of
genotypes originally collected from Jackson and Ingham counties in Michigan (W ildtype
Native Plant Nursery, Mason, MI). Plants were potted into 2.5 gallon pots and individuals

with similar numbers of buds were paired. Plants were placed in all 15 plots at the .,

145

MacCready site, near the center of each replicate, and in 4 replicates 15 m from each
other at the Grand River site from 8 — 22 July 2008 and 13 -— 30 July 2009. In both years
timing of the ﬁeld experiment corresponded with natural ﬂowering phenology of A.
incarnata at these sites. All open ﬂowers were removed before placement in the ﬁeld and
one plant was bagged with no—see—um netting (Skeeta, Bradenton, FL) to exclude
pollinators while the paired plant was left unbagged to allow full access by pollinators. :-
This design accounts for any incidental self—pollination within plants. In 2009, we also
sprayed Tangle—trap insect trap coating (Tanglefoot Company, Grand Rapids, M1) on all

plant stems of pollinator excluded plants to prevent small insects from climbing the stems

 

and accessing ﬂowers.

We counted the number of ﬂower clusters that opened on each plant during its
time in the ﬁeld. When most plants had ﬁnished blooming, all plants were removed from
the ﬁeld, and unopened buds were excised and plants were bagged to prevent insect
visitation so that subsequent fruit set was due to ﬁeld pollination only. Bagged plants
were held outdoors while fruits developed. Fruits were collected when mature but before
dehiscence, dried, and frozen for 24h to kill seed—feeding insects. The length and weight
of each pod, and number of viable seeds produced were measured for each fruit. In 2009,
a systemic insecticide was applied to all plants after removal ﬁ'om the ﬁeld site at a rate
of 0.37g/liter (imidacloprid, Marathon granular, OHP Inc, Mainland, PA) on 1 August to
prevent oleander aphid (Aphis neriz' Boyer de Fonscolombe) feeding on plants.

During A. incarnata deployment in the ﬁeld, we observed all ﬂower—visiting
insects for 10 minutes/ replicate three times during each ﬁeld season (10, 14—15, and 17

July 2008, 16, 20, and 24 July 2009) between 1000 and 1600 EST on calm, sunny days.

146

 

To prevent bias, the starting replicate for sampling was rotated on each date. Insects were
collected whenever possible and identiﬁed to species and to genus when only observed in

the ﬁeld.

Statistical analysis
Pollinator networks

We condensed all species interactions into one plant—pollinator network per
replicate of each treatment, and created one matrix for each, representing the number of
visits at each plant species by a particular insect species. Due to the low number of
sample replicates and uneven sample sizes, all 7 replicates were included together in the
plant—pollinator network analyses. We calculated a set of quantitative network
parameters for each matrix using R (Bipartite v. 1.11, R Project 2009). Metrics included:
1) connectance, a measure of the proportion of possible links that were actually observed,
2) compartment diversity, Shannon’s diversity index across compartment sizes, 3)
vulnerability, the mean number of links per plant, 4) generality, the mean number of
plants per pollinator and 5) interaction evenness, based on Shannon’s diversity with
matrix cells as species and interactions as abundance (Tylianakis et al. 2007). A number
of metrics examining the topology of plant—pollinator networks are correlated (T ylianakis
et a1. 2007 , Dormann et al. 2009). We performed a Pearson correlation between values
for these metrics, eliminating those that were signiﬁcantly correlated at a Bonferroni—
adjusted P of 0.0021. Although we used weighted, quantitative metrics in all cases
(Banasek-Richter et al. 2004), a number of metrics are known to be affected by the
richness of both lower and higher trophic levels in the web. We accounted for these
effects by performing an ANCOVA with treatment as the main effect and the number of

147

 

1‘?"

 

plant and pollinator species within the network as covariates (J MP v. 8.0.2, SAS Institute
2009). We report p-values for treatment effects using type III sums of squares,
accounting for the effect of the number of species in the web. We assessed residuals for
adherence to the assumption of homogeneity of variances, and performed reciprocal
transformations on connectance and vulnerability to meet this assumption.

Because plant—pollinator interactions were nearly absent in unrestored areas, we
could not calculate quantitative parameters on these networks. To examine differences in
fundamental properties of the webs, including: observed ﬂowering plant richness,

proportion of plant species not visited by pollinators, pollinator richness, sampling time,

 

and ﬂoral density per replicate, we performed ANOVAs with treatment as the main effect

II F

(SAS Institute 2010). To examine differences in the insect and plant community between
treatments, we created a Bray—Curtis similarity matrix using square root transformed
data, adding 1 to all values so that replicates with low insect and plant abundance
remained in the analysis. This matrix was used in non—metric multidimensional scaling
(N MDS) to examine the relationships between treatments, with points closer to each
other in 2—dimensional space being more similar. We performed the NMDS using 2 and
3 dimensions, with a minimum stress of 0.01 and 25 random starting conﬁgurations
(Clarke and Gorley 2006a). Stress values for 2—dimensional ﬁgures were 0.14 or less,
indicating that the 2—dimensional representation shown is an accurate representation of
the relationship between the points overall, although likely not necessarily between
clustered points (Clarke and Gorley 2006b). We compared treatments using analysis of
similarity (ANOSIM) with the resemblance matrix for each treatment, including 999

random permutations (Clarke and Gorley 2006b).

148

Pollinator function

We compared treatment and time effects on two characteristics of A. incarnata
plants open to pollinators: the number of seeds/ﬁ'uit and the number of fruits/ﬂower
cluster open during deployment (accounting for incidental pollination on paired,
pollinator exclusion plants). We also compared treatment and time effects on ﬂower—
visiting insect richness and abundance, as well as bee and butterﬂy abundance,
standardizing all values by the number of ﬂower clusters open during sampling. We used
2—way ANOVAs with treatment (invaded, restored, on—site reference, off—site reference)
and time (i.e. growing season since restoration; year 1 and year 2), as well as their

interaction (PROC Mixed, SAS Institute 2010).

Results

Plant-pollinator networks

We observed a total of 684 plant—insect interactions in 19 plant—insect webs,
including 100 insect and 52 ﬂower species (Table 12, Table 13). We found striking
differences in plant—pollinator web structure between invaded and both restored and
uninvaded plots. Networks at the herbaceous plant level were greatly simpliﬁed, and
nearly absent, in invaded plots (Figure 13A), with few plants in the web, as well as low
ﬂoral density (Table 14). In restored and reference plots, there were a greater number of
interactions, as well as more species of both plants and pollinators (Figure 13B,C).
Interestingly, the restored replicates contained greater plant species richness, ﬂoral
density, and a greater number of plant—pollinator interactions than the uninvaded
reference plots (Table 14).

149

 

 

Analyses of a number of network factors, accounting for the number of
pollinators and plants in the network, show no signiﬁcant differences between the
interactions within restored and uninvaded reference networks (Table 15). There was,
however, a marginally signiﬁcant difference in interaction evenness between restored
(0.843 :L 0.025) and reference networks (0.940 d: 0.019) with greater evenness in
reference networks, indicating increased uniformity of interactions between trophic levels
along different pathways in reference networks.

An NMDS ordination of the plant and insect communities shows that both the

insect and plant communities differ between the reference and restored plots (Figure 14).

 

The reference and restored areas cluster separately and are signiﬁcantly different in both i
plant (R 0.96, P=0.012) and insect (R 0.81, P=0.012) species identity and abundance. The

ﬂowering plant (R 0.74, P=0.002) and insect (R 0.96, P=0.002) communities in restored

plots are distinct from those in invaded plots, as well. The two reference sites (on— and

off—site), which were lumped for the prior analysis, are not signiﬁcantly different in their

plant communities (R 0.00, P=0.486). Notably, the ﬂowering shrub Potentillaﬁuticosa

auct. non L. was the most abundant and most visited plant species in all but one of the

reference replicates (Figure 13C), while it was entirely absent from the restored

replicates. The on—and off—site references differed in insect communities (R 0.56,

P=0.029), and the insect communities in them were more similar to each other than the

references were to the restored or invaded treatments (Figure 14B).

Pollinator function

We observed 202 total insects in 40 genera in 2008, and 96 insects in 26 genera in

2009, with a total of 329 A. incamata ﬂower clusters blooming on uncaged plants in the

150

ﬁeld in 2008 and 403 ﬂower clusters in 2009. We found signiﬁcant differences in the
number of fruits produced in 2008, but not in 2009 (Table 16, Figure 15A). In 2008, we
found signiﬁcantly greater number of ﬁ'uits per ﬂower cluster in the restored, on— and
off—site reference than the invaded plots, indicating that pollinator function was fully
restored in the ﬁrst year following invasive removal. We found no signiﬁcant differences
in the abundance of seeds per ﬁ'uit by treatment or year, with a mean of 50.1:l: 2.4 seeds
per fruit (Table 14), indicating that the number of fruits produced does, in fact, represent
plant reproductive output.

We found differences by treatment in the rate of insect visitation at ﬂowers (Table
14, Figure 15B), with signiﬁcantly greater insect density at the off—site reference than all
other treatments, no signiﬁcant difference between the on-site reference and restored
plots, and a signiﬁcantly lower ﬂower—visiting insect density in the invaded plots. This
difference indicates that ﬂower—visiting insect rates we observed in the off—site reference
may be greater than those required for pollen saturation at A. incarnata. In 2009, we
observed no signiﬁcant differences in rates of insect visitation by treatment (Figure 15B),
which ﬁts with the similar rates of fruit production in that year (Figure 15A).

We examined pollinator species richness to assess whether fruit set was affected
by the number of ﬂower—visiting insect species. We found signiﬁcant differences in
pollinator richness by treatment and by year, with no signiﬁcant treatrnent*year
interaction (Table 16, Figure 15C) with greatest richness at on—site reference plots,
followed by the off—site reference, restored, and invaded plots.

We also assessed whether restoration had differential effects on different key

pollinator groups; hymenoptera have been shown to be the most abundant and effective

151

 

 

pollinators of related Asclepias species, with Apis mellifera inefﬁcient at pollen transfer
(Theiss et al. 2007). Lepidoptera were less abundant and less likely to pick up Asclepias
pollinia, so were overall less effective pollinators (Theiss et a1. 2007). Therefore, we
examined the visitation rates of Lepidoptera and bees separately. We found signiﬁcant
differences in butterﬂy abundance by treatment and time (Table 16). Bee abundance
differed signiﬁcantly by treatment, but not by year (Table 16, Figure 15D), with
signiﬁcantly greater bee visitation rates at ﬂowers at the off—site reference than the
restored and invaded plots, and intermediate rates at the on—site reference. There were no
signiﬁcant differences in the evenness of plant—visiting insects by treatment or year,

based on the Berger—Parker index of dominance (Table 16).

Discussion

Incorporation of plant-pollinator networks and the resultant levels of pollination
ﬁmction could greatly strengthen our ability to assess the success of ecological
restorations. Our examination of shifts in a plant—pollinator networks under restoration
indicated a near complete absence of plant—pollinator interactions in invaded areas, with
rapid network recovery following restoration. We found that although both plant and
insect identities varied between restored and uninvaded reference plots, plant—pollinator
network topologies were similar. In addition, we linked these ﬁndings to pollination of
one species within that web, A. incamata, and found recovery of pollinator function in
restored areas within the ﬁrst growing season following restoration.

We found no signiﬁcant differences in network properties between restored and
reference webs, in properties including web connectance, compartment diversity, the

generalization of pollinators, and vulnerability of plants. We did ﬁnd marginally greater
152

 

interaction evenness in reference plots. Our ﬁndings match those of other plant-pollinator
networks in restored areas, including heath (F orup et al. 2008) and old ﬁeld hay meadows
(F orup and Memmott 2005), where no signiﬁcant differences in network topologies were
found between restored and reference areas, although there were trends in both studies
toward greater connectance and more complex species interactions in reference than
restored areas. Our study also indicated subtle differences between network structures in
restored and intact ecosystems. Despite ﬁnding no differences in network structure
between restored and reference plots, we did ﬁnd signiﬁcant differences in both the plant
and pollinator communities between restored and reference plots, indicating that species
interactions were restored independent of species identity. This ﬁnding is also consistent
with other restorations in which network structure is restored while species identities
remain distinct (Forup and Memmott 2005, Forup et al. 2008).

One unique component of our ﬁndings was the near complete lack of a plant—
pollinator web under F. alnus. The fact that F. alnus is not wind-pollinated and produces
seed readily suggests that pollinators are visiting its ﬂowers in the canopy, however this
network does not extend to the understory under F. alnus. One caveat is that there may be
a network of pollinators visiting the shade-tolerant herbaceous plants that bloom earlier in
the growing season than we measured. In other plant-pollinator networks, invasive plants
have similar growth habits to the natives, and their effects on plant-pollinator webs have
been measured directly (Aizen et al. 2008, Bartomeus et al. 2008) and vary, depending on
their ﬂowering phenology compared to the native community (Bartomeus et al. 2008). In
addition, non-native plant species have been found to be potential stabilizers of plant-

pollinator networks, even as they fundamentally modify their structure (Aizen et al. 2008,

153

 

 

Valdovinos et al. 2009). The drastic and rapid modiﬁcation of the plant—pollinator web at
ground level that we found with restoration indicates that in cases where invasive plants
alter plant community structure, plant—pollinator networks between multiple species may
be less likely to persist.

Our examination of pollination on the sentinel plant A. incarnata indicated that
pollinator function was rapidly restored in this system. In the ﬁrst year following
restoration we saw greater ﬁ'uit set in restored and reference than invaded plots. The
signiﬁcantly greater pollinator visitation rates at off—site referenCe plots indicated that A.
incarnata may have a fruit production threshold that was reached at a fraction of the
pollinator density. The insect species richness patterns that only loosely match patterns in
fruit production also ﬁt with prior ﬁndings that, for A. incarnata, pollinator effectiveness,
rather than the rate of insect visitation, is the best predictor for pollination (Sahli and
Conner 2006). Pollinator constancy at A. incarnata is known to be high, but pollinator
effectiveness is determined both by an insect’s ability to remove pollinia and the
proportion of pollinia removed that are inserted into the ﬂower (Theiss et al. 2007).
Therefore, low pollinator diversity can lead to high rates of pollination when pollinators
that are most effective are also most abundant (Perfectti et al. 2009).

We found no signiﬁcant differences in fruit production or pollinator visitation
rates between A. incarnata in restored or invaded plots in the second season following
restoration even though pollinator networks remained nearly absent in invaded plots.
There are several, related, possible explanations for these patterns. The ﬁrst is that both
plant and pollinator species in networks are known to vary from year to year, although

web properties can remain consistent between years (Alarcon et al. 2008, Olesen et al.

154

 

2008, Petanidou et al. 2008). Thus, our ﬁnding of varying pollinator function between
year 1 and year 2 may have been driven by stochastic factors alone. A second explanation
is that ﬂower density changed in restoration plots over time, inﬂuencing pollinator
behavior. In the ﬁrst growing season, vegetative cover and ﬂowering plant density
remained low in restored plots, while by the second growing season, ﬂowering plant
density in restored plots was greater than that in reference plots. This could have led to
greater competition between plants for pollinators in the restored plots (Sargent and
Ackerly 2008). Finally, there are known, positive spillover effects on plant diversity

beyond the edges of restored areas (Brudvig et al. 2009). In this case, we may have

 

measured a spillover effect of restoration on pollinator function as pollinators became
more likely to search for ﬂoral resources in invaded areas that neighbored restored areas
greater ﬂoral densities.

Generalist pollinators may be playing a key role in our system which in the initial
phases of restoration represents an early successional plant community (Chapter 3). Many
early successional species have adaptations, including self—compatibility and wind—
dispersed seeds, that may lead to lower likelihood of pollen limitation (Bond 1994,
Ashman et al. 2004). In contrast, the plant community of high quality fens may be
reproducing vegetatively (Klimkowska et al. 2009), so that fen species that did not persist
with F. alnus invasion may no easily return to restored areas. Determining potential
mechanisms for patterns of restored function even as differences in plant communities
persist would also allow for explicit restoration of pollinator function in communities
with plant species that are likely to be pollen—limited (Bond 1994) and could increase

relocation success of pollinator-dependent plant species. In our study, intact areas of

155

 

prairie fen were within 400m of restored areas, so that in a landscape context, there was a
potential source of both plants and pollinators nearby (Pauw 2007).

Our approach combining ﬁrnctional elements of restoration with plant—pollinator
network structure allows potential assessment of the mechanisms of pollinator limitation,
as well as providing a meaningful metric for the recovery of function with restoration.
Assessment of ﬁmction and species networks approached in tandem with species
diversity measurements could provide mechanisms for assessing why restorations aren’t
successful, as well as providing a more complete understanding of ecosystem processes

in degraded and restored habitats.

 

Acknowledgements

Research funding was provided by National Fish and Wildlife Foundation, the
MSU MacCready Endowment, and a Hanes Trust Fellowship of the Michigan Botanical
Society. We are grateful to The Nature Conservancy of Michigan for allowing our use of
the Grand River fen as a ﬁeld site. Funding support for A.K.F. was provided through a
sustainable agriculture predoctoral fellowship by the CS. Mott foundation a MSU
Barnett Rosenberg predoctoral fellowship in biological sciences, and a MSU College of
Agriculture dissertation completion fellowship. T. Rezniceck provided valuable plant
identiﬁcation, M. Arduser, J. Tuell, G. Parsons, B. DeMarco, F. C. Thompson, and M.

Neilsen provided valuable insect identiﬁcations.

156

Literature Cited

Aguilar, R., L. Ashworth, L. Galetto, and M. A. Aizen. 2006. Plant reproductive
susceptibility to habitat fragmentation: review and synthesis through a meta-
analysis. Ecology Letters 9(8):968-980.

Aizen, M. A., C. L. Morales, and J. M. Morales. 2008. Invasive mutualists erode native
pollination webs. Plos Biology 6(2):396-403.

Alarcon, R., N. M. Waser, and J. Ollerton. 2008. Year-to-year variation in the topology
of a plant-pollinator interaction network. Oikos 1 17(12): 1796-1807.

Amon, J. P., C. A. Thompson, Q. J. Carpenter, and J. Miner. 2002. Temperate zone fens
of the glaciated Midwestern USA. Wetlands 22(2):301-317.

Ashman, T. L., T. M. Knight, J. A. Steets, P. Amarasekare, M. Burd, D. R. Campbell, M.
R. Dudash, M. O. Johnston, S. J. Mazer, R. J. Mitchell, M. T. Morgan, and W. G.
Wilson. 2004. Pollen limitation of plant reproduction: Ecological and
evolutionary causes and consequences. Ecology 85(9):2408-2421.

Bakker, J. P., and F. Berendse. 1999. Constraints in the restoration of ecological diversity
in grassland and heathland communities. Trends in Ecology & Evolution
14(2):63-68.

Banasek-Richter, C., M. F. Cattin, and L. F. Bersier. 2004. Sampling effects and the
robustness of quantitative and qualitative food-web descriptors. Journal of
Theoretical Biology 226(1):23-32.

Bartomeus, 1., M. Vila, and L. Santarnaria. 2008. Contrasting effects of invasive plants in
plant-pollinator networks. Oecologia 155(4):761-770.

Bascompte, J ., and P. J ordano. 2007. Plant-animal mutualistic networks: The architecture
of biodiversity. Annual Review of Ecology Evolution and Systematics 38:567-
593.

Bascompte, J ., P. J ordano, and J. M. Olesen. 2006. Asymmetric coevolutionary networks
facilitate biodiversity maintenance. Science 312(5772):431-433.

Bedford, B. L., and K. S. Godwin. 2003. Fens of the United States: distribution,
characteristics, and scientiﬁc connection versus legal isolation. Wetlands
23(3):608-629.

Biesmeijer, J. C., S. P. M. Roberts, M. Reemer, R. Ohlemuller, M. Edwards, T. Peeters,
A. P. Schaffers, S. G. Potts, R. Kleukers, C. D. Thomas, J. Settele, and W. E.
Kunin. 2006. Parallel declines in pollinators and insect-pollinated plants in Britain
and the Netherlands. Science 313(5785):351-354.

157

“W"

 

Bond, W. J. 1994. Do Mutualisms Matter - Assessing the Impact of Pollinator and
Disperser Disruption on Plant Extinction. Philosophical Transactions of the Royal
Society B-Biological Sciences 344(1307):83-90.

Brown, B. J ., R. J. Mitchell, and S. A. Graham. 2002. Competition for pollination
between an invasive species (purple Ioosestrife) and a native congener. Ecology
83(8):2328-2336.

Brown, M. J. F., and R. J. Paxton. 2009. The conservation of bees: a global perspective.
Apidologie 40(3):410-416.

Brudvig, L. A., E. l. Damschen, J. J. Tewksbury, N. M. Haddad, and D. J. Levey. 2009.
Landscape connectivity promotes plant biodiversity spillover into non-target
habitats. Proceedings of the National Academy of Sciences of the United States of
America 106(23):9328-9332.

Burd, M. 1994. Bateman principle and plant reproduction - the role of pollen limitation in i
limit and seed set. Botanical Review 60(1):83-139.

 

Clarke, K. R., and R. N. Gorley. 2006a. Primer v6. Primer-E, Plymouth, UK.

Clarke, K. R., and R. N. Gorley. 2006b. Primer v6: User Manual/'1‘ utorial. Primer-E,
Plymouth, UK

Dicks, L. V., S. A. Corbet, and R. F. waell. 2002. Compartrnentalization in plant-insect
ﬂower visitor webs. Journal of Animal Ecology 71(1):32-43.

Dormann, C. F., J. Fund, N. Bluthgen, and B. Gruber. 2009. Indices, graphs and null
models: Analyzing bipartite ecological networks. The Open Ecology Journal 2:7-
24.

Eriksson, O. 2000. Functional roles of remnant plant populations in communities and
ecosystems. Global Ecology and Biogeography 9(6):443-449.

Fiedler, A. K., and D. Landis. in review. Biotic and abiotic conditions in Michigan prairie
fen invaded by glossy buckthom (F rangula alnus). Natural Areas Journal in
review.

Forup, M. L., K. S. E. Henson, P. G. Craze, and J. Memmott. 2008. The restoration of
ecological interactions: plant-pollinator networks on ancient and restored
heathlands. Journal of Applied Ecology 45(3):742-752.

F orup, M. L., and J. Memmott. 2005. The restoration of plant-pollinator interactions in
hay meadows. Restoration Ecology 13(2):265-274.

Judd, W. S. 2002. Plant systematics : a phylogenetic approach. 2nd edition. Sinauer
Associates, Sunderland, Mass. xvi, 576 p.

158

Klirnkowska, A., R. van Diggelen, S. den Held, R. Brienen, S. Verbeek, and K. Vegelin.
2009. Seed production in fens and fen meadows along a disturbance gradient.
Applied Vegetation Science l2(3):304-315.

Lipow, S. R., and R. Wyatt. 2000. Towards an understanding of the mixed breeding
system of swamp milkweed (Asclepias incarnata). Journal of the Torrey Botanical
Society 127(3):193-199.

Lomov, B. 2010, in press. Pollination and plant reproductive success in restored urban
landscapes dominated by a pervasive exotic pollinator. Landscape and Urban
Planning.

Lomov, B., D. A. Keith, and D. F. Hochuli. 2009. Linking ecological function to species
composition in ecological restoration: Seed removal by ants in recreated
woodland. Austral Ecology 34(7):751-760.

Memmott, J ., N. M. Waser, and M. V. Price. 2004. Tolerance of pollination networks to
species extinctions. Proceedings of the Royal Society of London Series B-
Biological Sciences 271(1557):2605-261 1.

 

Nabhan, G. P., and S. L. Buchmann. 1997. Services provided by pollinators. Page 392 in
G. C. Daily, editor. Nature's Services : Societal Dependence on Natural
Ecosystems. Island Press, Washington, DC.

Nekola, J. C. 1994. The environment and vascular ﬂora of northeastern Iowa fen
communities. Rhodora 96(886):]21-169.

Olesen, J. M., J. Bascompte, H. Elberling, and P. Jordano. 2008. Temporal dynamics in a
pollination network. Ecology 89(6): 1 573-1 582.

Pauw, A. 2007 . Collapse of a pollination web in small conservation areas. Ecology
88(7):]759-1769.

Perfectti, F ., J. M. Gomez, and J. Bosch. 2009. The functional consequences of diversity
in plant-pollinator interactions. Oikos 118(9): 1430-1440.

Petanidou, T., A. S. Kallimanis, J. Tzanopoulos, S. P. Sgardelis, and J. D. Pantis. 2008.
Long-term observation of a pollination network: ﬂuctuation in species and
interactions, relative invariance of network structure and implications for
estimates of specialization. Ecology Letters 11(6):564-575.

Potts, S. G., J. C. Biesmeijer, C. Kremen, P. Neumann, O. Schweiger, and W. E. Kunin.
2010. Global pollinator declines: trends, impacts and drivers. Trends in Ecology
& Evolution 25(6):345-353.

R Project, C. T. 2009. A language and environment for statistical computing R
Foundation for Statistical Computing, Vienna, Austria.

159

Ruiz-Jaen, M. C., and T. M. Aide. 2005. Restoration success: How is it being measured?
Restoration Ecology 13(3):569-577.

Sahli, H. F., and J. K. Conner. 2006. Characterizing ecological generalization in plant-
pollination systems. Oecologia l48(3):365-372.

Sargent, R. D., and D. D. Ackerly. 2008. Plant-pollinator interactions and the assembly of
plant communities. Trends in Ecology & Evolution 23(3): 123-130.

SAS Institute. 2009. JMP v. 8.0.2. SAS Institute, Cary, NC.

SAS Institute. 2010. SAS/STAT User's guide for personal computers. SAS Institute,
Cary, NC.

Seavey, S. R., and P. H. Raven. 1977. Chromosomal evolution in Epilobium sect
Epilobium (Onagraceae). Plant Systematics and Evolution 127(2-3):107-119.

Sinclair, A., and P. M. Catling. 1999. The value of cutting in the management of glossy
buckthom (Rhamnusﬁ'angula L.). Wetland Journal 11(2):25-27.

Spieles, J. B., P. J. Comer, D. A. Albert, and M. A. Kost. 1999. Natural community
abstract for prairie fen. Michigan Natural Features Inventory, Lansing, MI.

Theiss, K., S. chhart, and C. T. Ivey. 2007 . Pollinator effectiveness on co-occurring
milkweeds (Asclepias; Apocynaceae, Asclepiadoideae). Annals of the Missouri
Botanical Garden 94(2):505-516.

Tuell, J. K., A. K. F iedler, D. Landis, and R. Isaacs. 2008. Visitation by wild and
managed bees (Hymenoptera : Apoidea) to eastern US native plants for use in
conservation programs. Environmental Entomology 37(3):707-718.

Tylianakis, J. M., R. K. Didharn, J. Bascompte, and D. A. Wardle. 2008. Global change
and species interactions in terrestrial ecosystems. Ecology Letters 11(1 2):1351-
1363.

T ylianakis, J. M., T. Tschamtke, and O. T. Lewis. 2007. Habitat modiﬁcation alters the
structure of tropical host-parasitoid food webs. Nature 445(7124):202-205.

Valdovinos, F. S., R. Ramos-Jiliberto, J. D. Flores, C. Espinoza, and G. Lopez. 2009.
Structure and dynamics of pollination networks: the role of alien plants. Oikos
118(8):]190—1200.

Waser, N. M., L. Chittka, M. V. Price, N. M. Williams, and J. Ollerton. 1996.
Generalization in pollination systems, and Why it matters. Ecology 77(4): 1043-
1060.

Wilcove, D. S., D. Rothstein, J. Dubow, A. Phillips, and E. Losos. 1998. Quantifying
threats to imperiled species in the United States. Bioscience 48(8):607-615.

160

 

ltz.«..-.Jf'al" ‘ '-. '

Winfree, R. 2010. The conservation and restoration of wild bees. Annals NY Academy of
Sciences pre-print.

Young, T. P., D. A. Petersen, and J. J. Clary. 2005. The ecology of restoration: historical
links, emerging issues and unexplored realms. Ecology Letters 8(6):662-673.

161

 

 

 

 

Table 12. Plant species in quantitative ﬂower visitation webs included in Figure 13.
Code Family Genus gpecies Common Name
ASTLAN Asteraceae Aster lanceolatus Willd. white panicle aster
BOECYL Urticaceae Boehmeria cylindrica (L.) Sw. false nettle

CIRLUT Onagraceae Circaea lutetiana L. broadleaf enchanter's nightshade
ClRMUT Asteraceae Cirsium muticum Michx. swamp thistle

EPICOL Onagraceae Epilobium coloratum Biehler purpleleaf willowherb
EUPPER Asteraceae Eupatorium perfoliatum L. common boneset
FRAALN Rhamnaceae Frangula alnus Mill. glossy buckthom
GALBOR Rubiaceae Galium boreale L. northern bedstraw
GEUCAN Rosaceae Geum canadense Jacq. white avens

HYPHIR Liliaceae Hypoxis hirsute (L.) Coville common goldstar
lMPCAP Balsaminaceae Impatiens capensis Meerb. jewelweed

LOBKAH Campanulaceae Lobelia kalmii L. Ontario lobelia

LOBSIP Campanulaceae Lobelia siphilitica L. great blue lobelia
LYSQUA Primulaceae Lysimachia quadriﬂora Sims fourﬂower yellow Ioosestrife
LYCUNI Lamiaceae Lycopus uniﬂorus Michx. northern bugleweed
MENARV Lamiaceae Mentha arvensis L. wild mint

MIMRIN Scrophuiariaceae Mimulus ringens L. Allegheny monkeyﬂower
PARGLA Saxifragaceae Parnassia glauca L. grass of pamassus
PEDLAN Scrophuiariaceae Pedicularis lanceolata Michx. swamp lousewort
PENSED Crassuiaceae Penthorum sedoides L. ditch stonecrop

POTF RU Rosaceae Potenti/Ia fruticosa auct. non L. shrubby cinquefoil
POLPUN Polygonaceae Polygonum punctatum Elliot dotted smartweed
POLSAG Polygonaceae Polygonum sagittatum L. arrowleaf tearthumb
ROSMUL Rosaceae Rosa multiﬂora Thunb. multiﬂora rose

SCUGAL Lamiaceae Scutellaria galericulata L. marsh skullcap
SOLDUL Solanaceae Solanum du/camara L. climbing nightshade
SOLOHI Asteraceae Solidago ohioensis Frank ex Riddell Ohio goldenrod
SOLPAT Asteraceae Solidago patula Muhl. ex Willd. roundleaf goldenrod
SOLRUG Asteraceae Solidago rugosa Mill. rough-leaved goldenrod
VALULl Valerianaceae Valenana ulrgmosa (Torr. 8‘ A' Gray) mountain valen'an

Rydb.

162

 

_
i

 

Table 13. Insect species in quantitative ﬂower visitation webs included in Figure 13.

 

 

 

Code Order Family Genus species
Hymenoptera
1 Apidae Anthophora terminalis Cresson
2 Apidae Apis mellifera L.
3 Apidae Bombus fervidus (F.)
4 Apidae Bombus impatiens Cresson
5 Apidae Bombus vegans Smith
6 Apidae Ceratina calcarata/dupla a
7 Colletidae Hylaeus sp. 1!"-
8 Halictidae Augochlora pura (Say)
9 Halictidae Augochlorella aurata (Smith)
10 Halictidae Augochloropsis metallica fulgida (Smith)
11 Halictidae Halictus rubicundus (Christ)
12 Halictidae Lasioglossum lineatulum (Crawford)
13 ichneumonidae "ichneum2"
14 Tiphiidae Myzinum sp. ,_
15 Vespidae Dolichovespula arenaria (Fabricius) H
16 Vespidae Dolichovespula maculata (Linnaeus)
Diptera
17 Syrphidae Sphaerophoria sp.
18 Syrphidae Syrphus ribesii (Linnaeus)
19 Syrphidae Toxomerus marginatus Say
20 Tachinidae Jurinopsis adjusta Van der Wulp
21 Tephritidae Urophora sp.
Coleoptera
22 Cerambycidae Euderces picipes (F .)
23 Chrysomelldae Diabrotica undecimpunctata (Barber)
24 Coccinellidae Harmonia axyridis (Pallas)
25 Mordellidae Mordella atrata (Melsheimer)
26 Mordellidae Mordella marginata (Melsheimer)
Hemiptera
27 Miridae Neurocolpus nubilis (Say)
28 Miridae Plagiognathus sp.
29 Phymatidae Phymata pennsylvanica (Handlirsch)
30 Reduviidae Zelus Iuridus Stat

 

 

163

Table 14. Characterization of plant-pollinator networks in invaded, restored, and
uninvaded reference treatments. Mean iSEM.

 

Treatment Plant richness insect richness ,6 Of plants not

 

visited Floral density- ff of interactions
invaded 1.2 1 0.5 1.7 1 0.9 59.7 1 15.0 1.3 1 0.4 2.0 1 1.1
Restored 9.8 1 1.5 22.5 1 2.9 32.6 1 9.3 14.1 11.7 77.5 1 18.4
Reference 5.6 1 0.8 13.6 1 1.2 55.5 1 5.1 5.2 1 0.8 25.9 1 4.9

2

 

~Per m

 

 

164

Table 15. Differences in plant—pollinator network structure metrics between restored and
uninvaded reference plots, based on an analysis of covariance with the number of plant
and pollinator species in each network accounted for, with DF 3,9.

Factor F p R2

 

Connectance* 0.3 0.580 0.958
Compartment diversity 0.2 0.652 0.172
Generality 3.0 0.796 0.894
Vulnerability* 0.5 0.484 0.662
Interaction evenness 4.4 0.066 0.447

 

* Reciprocal transformation performed for data analysis

 

165

 

Table 16. Differences in pollinator function and insect visitation at A. incarnata plants

between invaded, restored, and uninvaded treatments. Insect abundances are scaled by
the number of A. incarnata ﬂower clusters open during sampling. Signiﬁcant p-values

(or=0.05) are bolded.

 

Treatment Year Treatment * ar

Fad" F P F P F y:
Seeds/fruit 0.5 0.664 2.4 0.134 2.0 0.147
Fruits/ﬂower cluster*~ 12.5 <.001 3.3 0.081 4.9 0.007
Total insect abudance“ 8.0 0.001 10.2 0.004 4.9 0.008
Bee abundance“ 3.1 0.042 0.1 0.911 1.6 0.214
Butterﬂy abundance* 3.9 0.019 4.8 0.037 1.1 0.371

Pollinator species richness 5.5 0.005 4.4 0.047 0.7 0.577
Berger-Parker dominance 0.3 0.852 2.1 0.156 1 .5 0.238

 

* log(x+l) transformed data used for analysis.
ASquare root transformed data used for analysis.
~Accounting for pods produced in the unpollinated pair (i.e. incidental pollination)

166

 

1‘i‘. -"_' I" I 1' A ._'

 

Figure 13. Quantitative ﬂower visitation webs for representative replicates of the A)
invaded, B) restored, and C) reference prairie fen plots. Plant species are represented by
rectangles on the bottom of the ﬁgure, insects are shown on the top level. The width of a
bar represents the relative abundance of pollinators or plants and the frequency of the
interaction is represented by the width of the line from pollinators to plants. Flowers that
were observed but no pollinators were observed at are present in these webs, but with no
line connecting them to pollinators. Hymenopteran ﬂower—visitors are shown in black,
insects from all other orders are in grey.

167

1e Al
ented by
width Oil
of the
M615 [1131
it “1111 110
[1 black

Scale:

10
10

30

   

— |1i11V/\

28
E
O
.1
O
U)

7 917 19

2

— 9VS10d
Nnd10d
— N1WEH

  
 
 

$2
2 I rmuio
,8 .uoaoa
. .2 *3
= re
< 8 E
Figure 13.

168

9118108
lVd1OS

100108

WOODS
111WSOH

131415161920 22 23 24 25

9VS10d

12

10 11

     
 

- Nl'ld10d
- ClElSNBd
NIHWIW

AHVNEW

I INnOA1

9

" .. dVOdWI
—Nvonae

HaddnEl

100|d3

_ .lf‘lWUIO
._ - 1111810

 

 

 

 

 

 

A. Flowering plants 0 2D Stress: 0.14
. o
O o 00 A invaded
O Restored
I On-site Ref
e O Off-site Ref
I
e- A
A
6‘ A

D
" " "mi

 

 

 

 

 

 

 

A
A
N
E 3- insegts I 20 Stress: 0.11 ' '
e
O
o
O
‘ A
I
o A
o
o ‘ ‘
o
o 0
Axis 1

 

Figure 14. Two dimensional NMDS ordinations of A) the ﬂowering plant and B) ﬂower—-
visiting insect communities in plant—pollinator interactions. Both A) and B) indicate that
there are distinct plant and pollinator communities at the uncleared and cleared plots. A)
shows no distinction between the plant communities in the on— and off—site references,
while B) reveals signiﬁcantly different plant—visiting insect communities between on—
and off—site references. The ordination is based on a Bray—Curtis dissimilarity matrix
(using n+1) with mean values per replicate and treatment; data were square root
transformed. Stress values were 0.14 and 0.11 for ﬂowering plants and insects,

respectively.

169

Figure 15. Patterns of A) pollinator function and at A. incarnata ﬂower—visiting B) insect
richness C) insect abundance and D) bee abundance in the ﬁrst and second seasons
following restoration. Insect richness and abundances are based on means by treatment
and replicate over three sampling dates. Error bars are +SEM, D=0.05. Treatment effects
from a 2—way ANOVA with treatment and time are shown where signiﬁcant. Letters in
upper and lower case within the same graph indicate that treatments differed by year or

treatrnent*year.

170

g B) libel“!
ions
eatmenl
:nl elicits
.eiters in
year or

 

 

 

 

 

I I Y T j
. m. o. s. N.
F O O O O

mu: 0; [raisnlo Jeri/log / sroasul

20
18-
.6
4
12a
0

I
o.
O

 

 

- 2008
[:1 2009

 

 

 

 

 

 

 

 

reisnp ramou/ sum; to requinN
Figure 15.

 

171

 

 

 

 

 

 

Invaded Restored On-site Ref Off-site Ref

 

 

.I. e s. N o
O O O O O
ugw or [Jersnio JBMOU / seas #

 

 

 

 

 

 

 

 

 

 

 

 

 

 

E
9 .9
‘6
8'
9—
0)
j___ (K
< .3
9
C
o
8
0 8
°° as
6)
a:
B
8
>
E
.3 l. .1. s.

sseuuou segoeds Joreumod

 

 

Chapter 6

 

CONCLUSION

IMPROVING THE PRACTICE OF RESTORATION THROUGH RESTORATION
ECOLOGY

 

172

Ecosystem restoration was ﬁrst practiced in the 1940’s with Aldo Leopold’s
pioneering prairie restoration efforts in Wisconsin (Jordan et al. 1987). The subsequent
emergence of ecosystem restoration as a science is, in fact, a measure of our increasing
understanding of complex ecological systems (Jordan et al. 1987) as well as a means to
uncover gaps in our scientiﬁc knowledge. The practice of ecosystem restoration provides

a unique opportunity to address those knowledge gaps through development of ecological

I

theory and paradigms. The applied aspects of restoration dominated the science until the
1990’s so that the branch of ecology that has stemmed from ecosystem restoration,

restoration ecology, is relatively young (Young et al. 2005). The goal of ecosystem

 

 

t! TIL”: .7} ' 9“". )er "shin“! “"-

restoration is the recovery of an ecosystem toward a healthy, functional system (SER
Working Group 2004). Therefore, ecosystem restoration involves manipulating species
and processes within ecosystems and by its very nature provides potential for large-scale
experimental and observational studies that inform basic questions about community and
ecosystem ecology. My PhD research spans several areas of ecological theory that could
be informed through restoration, including species competition, succession, examination
of mutualisms, recruitment limitation, and ecosystem function (Young et al. 2005).

This study is unique within a restoration ecology context in two ways. First, it
includes low-quality invaded areas and hi gh-quality uninvaded reference areas in addition
to restored areas. In many cases restoration projects are performed in areas where there is
no available reference community, so historic conditions are used instead (White and

Walker 1997). In other cases, the comparisons lie between areas that have already been

 

restored and high quality references (e. g. F orup and Memmott 2005, F orup et al. 2008)

which do not provide an indication of the change since factors degrading the system were

173

removed or altered. Second, it incorporates a multi-faceted assessment of restoration
including conditions in invaded and high quality areas before restoration began (Chapter
2) resource availability and plant community structure (Chapter 3), pollinator and plant
diversity (Chapter 4), and plant-pollinator network structure and pollination in invaded,
restored, and uninvaded plots (Chapter 5). This is the ﬁrst study, to my knowledge, that
incorporates examination of all of these facets within a single ecosystem.

This cross—cutting approach not only provides a more nuanced assessment of
restoration success, it combines two areas of ecology that were, until recently, seen as
distinct: community and ecosystem ecology. The community ecology paradigm includes
a smaller-scale focus on species diversity and constraints that lead to greater or lower
diversity, as well as species interactions (Loreau 2010). The ecosystem ecology paradigm
includes a larger-scale focus on energy ﬂow and nutrient cycling, with biogeochemical
processes as the primary focus (Loreau 2010). The new paradigm, incorporating the
strengths of both community and ecosystem ecology, is one focused on biodiversity and
ecosystem functioning (Schwartz et al. 2000, Naeem 2002, Loreau 2010).

My study incorporates plant and pollinator diversity, as well as examination of the
community-based ecosystem function of pollination. Although I do not incorporate
explicit examination of the relationship between biodiversity and ecosystem function in
my study (Loreau et al. 2001), I found no signiﬁcant differences in pollinator diversity
(Chapter 4) plant-pollinator network properties, or pollinator function (Chapter 5)
between restored and reference communities. Using this approach, I can assess whether
functional elements of a system are consistent over time even if there are shifts in species

diversity and identity, and incorporate a more current, Gleasonian understanding of plant

174

 

 

communities as dynamic, rather than moving toward equilibrium after restoration
(Gleason 1926). Because I have characterized abiotic factors in the system, I also have an
understanding of what the biotic limitations may be.

A number of the factors I measured yielded different trajectories toward
restoration. Interestingly, factors related to resource availability, including soil pH and
microhabitat, showed no change in the ﬁrst two growing seasons post-restoration
(Chapter 2). Plant diversity and quality had also not changed in restored areas. On the
other hand, a number of factors shifted rapidly post-restoration. Plant community
structure had recovered broadly after 2 growing seasons. All metrics based on the
pollinator community, including diversity, plant-pollinator networks, and pollination,
recovered nearly immediately following restoration.

While it is clear that the restored areas are broadly on a trajectory toward the state
in reference areas, I do not know whether the shift will continue over time. There may be
biotic or abiotic thresholds in this system, so that the changes I have observed plateau or
conditions shift in the direction of the invaded system (Suding et al. 2004, Suding and
Hobbs 2009). Although my study has been relatively comprehensive in the factors
measured post-restoration, the ﬁndings are limited by the short timescale represented,
with two growing seasons post-restoration. Continued monitoring would not only assess
whether the restoration is successful in the long-term, it could provide a link between
elements of ecosystem theory and restoration (Young et al. 2005). These include 1)
assessing dispersal or recruitment limitation of key fen species, especially sedges, 2)
determining whether the lack of microtopography limits plant diversity and if

microtopographic variability increases over scales of tens, rather than hundreds, of years,

175

 

 

3) tracking whether plant succession moves in the direction of species found in intact
prairie fens, and 4) examination of plant-pollinator mutualisms and pollinator function
over time, to determine whether structure and function of the pollinator system are stable
across growing seasons.

This work also informs timescales for restoration in different ecosystems. The
goal state for prairie fens is essentially one of arrested succession, with relatively ﬁ'equent
disturbance preventing extensive shrub growth. As a result, prairie fens are dominated by
herbaceous perennial sedges, forbs, and grasses, as well as shrubs, resulting in a rapidly
restorable community structure. This community structure is also part of the reason for
the rapid plant and insect response; the primary plants are herbaceous perennial species,
rather than much longer-lived trees. Timescales for restoration in open-structured
ecosystems are likely to be shorter than those in which longer-lived trees are an important
part of the structure.

Although the metrics of restoration success I measured broadly indicate success in
terms of restoration, there are a number of factors which remain to be addressed. The
research design of this site, with restoration of intentionally fragmented patches, would '
take a great deal of effort to maintain in the long-term. Even in more contiguous restored
tracts, long-term control of invasive species will be required, as invasives are part of the
regional species pool and can disperse into the restored area (Matthews et al. 2009).
Studies on habitat restoration typically focus on restoration of single habitat types,
despite the fact that habitats occur on gradients and were historically interconnected
There is mounting evidence that species will be more likely to persist in landscapes that

facilitate dispersal among habitat patches (Ricketts 2001, Franklin and Lindenmayer

176

 

 

2009). Explicit tests of the effect of restoration across habitats (e. g. of prairie fens and
neighboring upland prairie and oak savanna systems) would provide ideal conditions to
assess whether there are positive effects of restoration in linked habitats, including
increased native plant diversity within and outside of restored areas (Brudvig et al. 2009).
Because research related to restoration is performed with limited resources, on-
ground restoration that incorporates any type of monitoring is relatively rare. I call for
clear communication of restoration research ﬁndings as a vital component of improving
the practice of ecosystem restoration. This extends beyond scientiﬁc publications to
workshops, web-based tools, and contacts with conservation organizations. Both
restoration ecology and ecosystem restoration would be advanced if restoration were
regularly carried out in both the context of ecological and practical questions that need

addressing, in collaboration with land managers.

177

 

 

Literature Cited

Brudvig, L. A., E. I. Damschen, J. J. Tewksbury, N. M. Haddad, and D. J. Levey. 2009.
Landscape connectivity promotes plant biodiversity spillover into non-target habitats.
Proceedings of the National Academy of Sciences of the United States of America
106(23):9328-9332.

Forup, M. L., K. S. E. Henson, P. G. Craze, and J. Memmott. 2008. The restoration of
ecological interactions: plant-pollinator networks on ancient and restored heathlands.
Journal of Applied Ecology 45(3):742-752.

Forup, M. L., and J. Memmott. 2005. The restoration of plant-pollinator interactions in
hay meadows. Restoration Ecology 13(2):265-274.

Franklin, J. F., and D. B. Lindenmayer. 2009. Importance of matrix habitats in
maintaining biological diversity. Proceedings of the National Academy of SCiences of the
United States of America 106(2):349-350.

Gleason, H. A. 1926. The individualistic concept of the plant association. American
Midland Naturalist 21 :92- 100.

Jordan, W. R., M. E. Gilpin, and J. D. Aber. 1987. Restoration ecology: a synthetic
approach to ecological research. Cambridge University Press, Cambridge
[Cambridgeshire] ; New York. viii, 342 p.

Loreau, M. 2010. Linking biodiversity and ecosystems: towards a unifying ecological
theory. Philosophical Transactions of the Royal Society B-Biological Sciences
365(1537):49-60.

Loreau, M., S. Naeem, P. Inchausti, J. Bengtsson, J. P. Grime, A. Hector, D. U. Hooper,
M. A. Huston, D. Raffaelli, B. Schrnid, D. Tilman, and D. A. Wardle. 2001. Ecology -
Biodiversity and ecosystem functioning: Current knowledge and future challenges.
Science 294(5543):804-808.

Matthews, J. W., A. L. Peralta, A. Soni, P. Baldwin, A. D. Kent, and A. G. Endress.
2009. Local and landscape correlates of non-native species invasion in restored wetlands.
Ecography 32(6):1031-1039.

Naeem, S. 2002. Ecosystem consequences of biodiversity loss: The evolution of a
paradigm. Ecology 83(6):1537-1552.

Ricketts, T. H. 2001. The matrix matters: Effective isolation in fragmented landscapes.
American Naturalist 158(1):87-99.

Schwartz, M. W., C. A. Brigham, J. D. Hoeksema, K. G. Lyons, M. H. Mills, and P. J.
van Mantgem. 2000. Linking biodiversity to ecosystem function: implications for
conservation ecology. Oecologia 122(3):297-305.

178

SER Working Group. 2004. The SER international primer on ecological restoration,
www.ser.org & Tucson: Society for Ecological Restoration International.

Suding, K. N., K. L. Gross, and G. R. Houseman. 2004. Alternative states and positive
feedbacks in restoration ecology. Trends in Ecology & Evolution 19(1):46-53.

Suding, K. N., and R. J. Hobbs. 2009. Threshold models in restoration and conservation:
a developing framework. Trends in Ecology & Evolution 24(5):271-279.

White, P. S., and J. L. Walker. 1997. Approximating nature's variation: Selecting and
using reference information in restoration ecology. Restoration Ecology 5(4):338-349.

Young, T. P., D. A. Petersen, and J. J. Clary. 2005. The ecology of restoration: historical F.
links, emerging issues and unexplored realms. Ecology Letters 8(6):662-673. -:

 

179

Appendix A. Soil nutrient levels (Mean 1 SEM) ﬁom 1-10 cm soil depth cores.

 

 

Factor Time invaded Removal Reference
NH4(ppm) Pre-restoration 14.74 1 1.33 14.52 1 1.37 9.72 1 2.49
Year 1 9.16 1 0.81 9.79 1 1.08 6.13 1 1.00
Year 2 17.77 1 2.07 12.85 1 1.29 8.17 1 1.49
NO3(ppm) Pre-restoration 12.73 1 2.42 17.83 1 3.26 5.56 1 1.13
Year 1 6.70 1 1.02 10.40 1 1.72 6.92 1 2.00
Year 2 13.34 1 2.87 13.45 1 2.33 8.86 1 3.09
Phosphorous (ppm) Pre-restoration 7.78 1 0.77 9.67 1 1.30 6.17 1 1.19
Year 1 8.83 1 1.35 9.06 1 1.34 7.50 1 1.35
Year 2 10.17 1 0.78 11.39 1 1.08 4.78 1 1.31
Potassium (ppm) Pre-restoration 88.96 1 8.30 79.12 1 7.97 69.17 1 16.65
Year1 104.36 1 10.05 103.46 1 10.14 94.47 1 11.96
Year 2 101.98 1 8.32 85.89 1 6.76 80.59 1 10.07

 

180

 

Appendix B.

 

 

Plant Voucher Data
Page 1 of 5 Pages

Collecti Collection

on # Famiy Scientific Name Date location

151 Euphorbiaceae Acalypha rhomboidea Raf. 9/10/2008 Block 5

163 Aceraceae Acer rubrum L. 5/30/2009 near Ref 1

103 Asteraceae Achillea millefolium L. 6/6/2008 Block 6

157 Ranunculaceae Actaea rubra 5/27/2009 Block1 cleared

34 Scrophuiariaceae Agalinis purpurea (L.) Pennell 9/11/2007 interiake fen

133 Rosaceae Agrimonia parviﬂora Aiton 8/14/2008 N of Block 1

189 Poaceae fegcvquron trachycaulum (Link) Malte ex H.F. 8/3/2009 Ref 3

164 Brassicaceae Alliaria petiolata (M. Bieb.) Cavara & Grande 5/30/2009 Block 3 cleared

96 Fabaceae Amphicarpaea bracteata (L.) Fernald 6/5/2008 Block 3

225 Poaceae Andropogon gerardii Vitman 8/24/2009 Ref 3

156 Araceae Arisaema triphyllum (L.) Schott 5/27/2009 Block1 uncleared

248 Asclepladaceae Asclepias incarnata 7/16/2009 Ref 2

197 Asclepladaceae Asclepias syriaca 8/6/2009 interiake fen

16 Asteraceae Aster lanceolatus Willd. 9/9/2007 Ref1

12 Asteraceae Aster puniceus var. ﬁnnus 9/9/2007 Ref1

18 Asteraceae Aster umbellatus Mill. 9/9/2007 Ref1

257 Betulaceae Betula pumila (L.) 7/3/2009 Ref 3

14 Asteraceae Bidens coronata (L. ) Britton 9/9/2007 Ref1

13 Asteraceae Bldens frondosa (L.) 9/9/2007 R911

90 Urticaceae Boehmeria cylindrica (L.) Sw. 6/5/2008 Block 2

143 Poaceae Bromus ciliatus L. 9/5/2008 Ref 3

144 Poaceae Calamagrostis canadensis (Michx.) P. Beauv. 9/5/2008 Ref 3

137 Ranunculaceae Caltha palustris L. 5/15/2008 near Ref 1

57 Convolvuiaceae Calystegia sepium (L.) R. Br. 6/2/2008 Ref 3

246 Campanulaceae Campanula apan'noides Pursh 7/27/2009 interiake fen

14o Brassicaceae Cf’dam’m bum” (5°"'°b' 6" Mm") 5/1 512008 Block 2 cleared
Bntton, Stems 8 Poggenb.

123 Cyperaceae Carex bebbii Olney ex Fernald 7/30/2008 Block 5 cleared

101 Cyperaceae Carex blanda Dewey 6/6/2008 Block 6

190 Cyperaceae Carex cryptolepis Mack. 8/3/2009 Ref 3

52 Cyperaceae Carex ﬁava L. 6/2/2008 Ref 3

179 Cyperaceae Carex granulan's Muhl. ex Willd. 6/5/2009 Block 5 cleared

74 Cyperaceae Carex hystericina Muhl. ex Willd. 6/4/2008 Ref1

42 Cyperaceae Carex leptalea Wahlenb. 9/11/2007 Block1

73 Cyperaceae Carex pellita Muhl. ex Willd. 6/4/2008 Ref1

66 Cyperaceae Carex prairea Dewey ex Alph. Wood 6/3/2008 Ref 2

60 Cyperaceae Carex sartwellii Dewey 6/2/2008 Ref 3

51 Cyperaceae Carex sterilis Willd., 9 6/2/2008 Ref 3

54 Cyperaceae Carex sterilis Willd., 6‘ 6/2/2008 Ref 3

78 Cyperaceae Carex stipata Muhl. ex Willd. 6/4/2008 Block 1

72 Cyperaceae Carex stn'cta Lam. 6/4/2008 Ref1

53 Cyperaceae Carex tetanica Schkuhr 6/2/2008 Ref 3

 

181

 

Appendix B
Plant Voucher Data
Page 2 of 5 Pages

 

 

(Joﬂecﬂ Chukxnkur
on It Family Scientific Name Date location
198 Cyperaceae Carex vulpinoidea Michx. 8/6/2009 Block 5 cleared
154 Juglandaceae Carya sp. 5I27/2009 Block 3 uncleared
51 Celastraceae Celastrus orbiculatus Thumb 9/1312007 Block 5
175 Apiaceae Cicuta maculata L. 6/2/2009 Block 1 cleared
21 Poaceae Cinna arundinacea L. 9l9/2007 Block1
251 Onagraceae Circaea lutetiana L. 7/3/2009 Block 2 cleared
150 Asteraceae Cirsium arvense (L.) Scop. 9I9/2008 Block 6
213 Asteraceae Cirsium muticum Michx. 8/17/2009 near 5 cleared
244 Asteraceae Cirsium vulgare (Savi) Ten. 9/11/2009 Block 5 cleared
110 Cyperaceae Cladium man'scoides (Muhl.) Torr. 7/28/2008 Ref 3
224 Ranunculaceae Clematis virginiana L. 8/24/2009 Block 5 uncleared
62 Comaceae Camus foemina Mill. 6I2I2008 Ref 3
63 Comaceae Cornus stolonifera Michx. 6I2I2008 Ref 3
254 Fabaceae xix/‘13:: g’m’msu’" (Mum ex w'"d') 7/15/2009 1 cleared
Dichanthelium acuminatum (Sw.) Gould &
196 Poaceae C.A. Clark var. fascicu/atum (Torr.) 8/6/2009 Block 6 cleared
Freckm am
468 Dryopteridaceae Dryopteris carthusiana (\Iili.) H.P. Fuchs 9/12/2008 Block1
102 Dryopteridaceae Dryopteris cristata ( L.) A. Gray 6I6/2008 Block 5
55 Cyperaceae Eleocharis elliptica Kunth 6/2/2008 Ref 3
149 Cyperaceae Eliocharis elliptica Kunth 9I8/2008 Block 4
41 Equisetaceae Equisetum arvense (L.) 9111/2007 Block 1
119 Asteraceae Erechtites hieraciifolia (L.) Raf. ex DC. 7/29/2008 Block 6 cleared
162 Asteraceae Erigeron annuus (L.) Pers. 5/27/2009 Block 1 cleared
161 Asteraceae Erigeron philadelphicus L. 5/27/2009 Block 1 cleared
8 Asteraceae Eupatorium maculatum L. 9I9I2007 Ref1
260 Asteraceae Eupatorium perfoliatum L. 9/14I2009 Near Ref 1
24 Asteraceae Euthamia graminifolia (L.) Nutt. 9I11/2007
259 Rosaceae Fragaria virginiana Duchesne 9I14I2009 Upland near ref 1
153 Rhamnaceae Frangula alnus Mill. 5/27/2009 Block 3 uncleared
503 Oleaceae Fraxinus pennsylvanica Marsh. 9I17I2007 Block 6
80 Rubiaceae Galium boreale L. 6/4/2008 Block 1
100 Rubiaceae Galium clrcaezans Michx. 6/6/2008 Block 6
1 Rubiaceae Gallum triﬂorum Michx. 9I9/2007 Ref1
239 Gentianaceae Gentianopsis crinita (Froei.) Ma 9/10I2009 Ref 3
159 Geraniaceae Geranium macalatum L. 5/27/2009 Block 2 cleared
124 Rosaceae Geum canadense Jacq. 7130/2008 Block 5 cleared
61 Poaceae Glycerla striata (Lam.) Hitchc. 6/2/2008 Ref 3
118 Boraginaceae Hackelia virginiana ( L.) l.M. Johnst. 7/29/2008 Block 2cleared. out
145 Asteraceae Helenium autumnale L. 9/5I2008 Ref 3
236 Asteraceae Helianthus giganteus L. 9I9I2009 Block 3 cleared
141 Liliaceae Hypoxis hirsute (L.) Coville 5/22/2008 Ref 3
56 Aquifoliaceae Ilex vertlcillata (L.) A Gray 9I17I2007 Block 5
203 Balsaminaceae Impatiens capensis Meerb. 8/6/2009 Block 2 cleared

155

lridaceae

Iris versicolor L.

182

5/27/2009 Block 1 cl

 

Appendix B
Plant Voucher Data
Page 3 of 5 Pages

 

 

Collecti Collection
on # Family Scientific Name Date location
187 Juncaceae Juncus brachycephalus (Engelm.) Buchenau 8/3/2009 Ref 3
64 Juncaceae Juncus dud/eyi Wiegand 6I3/2008 Ref 2
218 Cupressaceae Juniperus virginiana L. 8/17/2009 Ref 2
183 Fabaceae Lathyrus palustris L. 6I15I2009 Ref 2
206 Poaceae Leersia oryzoides (L.) Sw. 8/6/2009 Block 2 cleared
237 Asteraceae Liatris spicata (L.) Willd. 9I10/2009 Ref 3
208 Lauraceae Lindera benzoin (L.) Blume 8/17/2009 Block 2 uncleared
219 Campanulaceae Lobelia kalmii L. 8/17/2009 interiake fen
209 Campanulaceae Lobelia siphilitica L. 8/17/2009 Block 3 uncleared
261 Caprifoliaceae Lonicera L. 9/14/2009 Near Ref1
3 Lamiaceae Lycopus americanus Muhl. ex W. Bartram 9/9I2007 Ref 1
5 Lamiaceae Lycopus uniﬂorus Michx. 9/9/2007 Ref1
113 Primulaceae Lysimachia quadriﬂora Sims 7/28/2008 Ref 3
165 Primulaceae Lysimachia thyrsiﬂora L. 5/30/2009 Block 6 cleared
52A Lythraceae Lythrum salicaria (L. ) 9/13I2007 Block 5
2 Lamiaceae Mentha arvensis L. 9/9I2007 Ref1
221 Scrophuiariaceae Mimulus ringens L. 8/24/2009 Block 1 cleared
28 Poaceae Muhlenbergia glomerata (Willd.) Trin. 9I11/2007 Ref 1
243 Poaceae Muhlenbergia mexicana (L.) Trin. 9I11I2009 Blocks 5-6
210 Dryopteridaceae Onoclea sensibilis L. 8/17/2009 Block 2 uncleared
95 Apiaceae Osmorhiza claytonii (Michx.) C.B. Clarke 6I5I2008 Block 2
217 Oxalidaceae Oxalis stricta L. 8/17/2009 Block 6 cleared
126 Apiaceae Oxypolis n‘gidor (L.) Raf. 7/25/2008 Block 2 uncleared
245 Poaceae Panicum capillare L. 9/11I2009 Block 6 cleared
30 Poaceae Panicum dichotomiﬁorum Michx. 9/11I2007 Ref 2
202 Vltaceae Parthenocissus quinquefolia (L.) Planch. 8/6/2009 Block 2 cleared
229 Scrophuiariaceae Pedicularis lanceolata Michx. 8/24/2009 Ref1
107 Crassuiaceae Penthorum sedoides L. 7/25/2008 Block 1 cleared
68 Poaceae Phalaris arundinacea L. 6I4I2008 Ref 1
200 Phytolaccaceae Phytolacca americana L. 8/6/2009 Block 2 cleared
204 Urticaceae Pilea pumila (L.) A. Gray 8/6/2009 Block 2 cleared
199 Plantaginaceae Plantago major (L.) 8/6/2009 Upland near Ref 1
71 Poaceae Poa pratensis L. 6I4I2008 Ref 1
249 Polygonaceae Polygonum amphibium L. 7/1612009 interiake fen
130 Polygonaceae Polygonum hydropiperoides Michx. 8/14/2008 Block 1 cleared
131 Polygonaceae Polygonum punctatum Elliot 8/14/2008 Block 1 cleared
6 Polygonaceae Polygonum sagittatum L. 9/9/2007 Ref1
207 Polygonaceae Polygonum virginianum L. 8/6/2009 Block 2 uncleared
129 Salicaceae Populus tremuloides Michx. 8/14/2008 Block 3 cleared
234 Rosaceae PotentiI/a fruticosa auct. non L. 8/24/2009 Ref 2
109 Lamiaceae Prune/Ia vulgaris L. 7/25/2008 Block 1 cleared
160 Rosaceae Prunus serotina Ehrh. 5/19/2009 Block 2 uncleared
223 Fagaceae Quercus sp. 8/24/2009 Block 5 uncleared
76 Ranunculaceae Ranunculus recurvatus Poir. 6I4I2008 Block 1
230 Rhamnaceae Rhamnus cathartica L. 8/24/2009 Block 3 uncl
33 Cyperaceae Rhynchospora alba (L) Vahl 9/11/2007 Ref 2

 

183

 

Appendix B
Plant Voucher Data
Page 4 of 5 Pages

 

 

Collecti Collection
on if Family Scientific Name Date location
89 Grossulariaceae Ribes cynosbati L. 6I5I2008 Block 2
106 Brassicaceae Rorippa palustris (L.) Besser 7/25/2008 Block 1 cleared
227 Rosaceae Rosa multiﬂora Thunb. 8/24/2009 near Ref 1
69 Rosaceae Rosa palustris Marsh. 6/4/2008 Ref1
253 Rosaceae Rubus pubescens Raf. 7I16I2009
67 Rosaceae Rubus strigosus Michx. 6I3I2008 Ref 2
250 Asteraceae Rudbeckia hirta L. 7/16/2009 Ref 2
92 Polygonaceae Rumex orbiculatus A. Gray 6I5I2008 Block 2
9 Salicaceae Salix disco/or Muhl. 919/2007 Ref 1
99 Caprifoliaceae Sambucus canadensis L. 6I5I2008 Block 3
205 Lauraceae Sassaﬁ'as albidum (Nutt.) Nees 8/6/2009 Block 1 uncleared
29 Poaceae Schizachyn'um scoparium (Michx.) Nash 9/11I2007 Ref 2
215 Cyperaceae Scirpus acutus Muhl. ex Bigelow 8/17/2009 interiake fen
111 Cyperaceae Scirpus americanus Pers. 7/28/2008 Ref 3
108 Cyperaceae Scirpus atrovirens Willd. 7/25/2008 Block 1 cleared
17 Lamiaceae Scutellaria galericulata L.(17) 9I9I2007 Ref 1
4 Lamiaceae Scutellaria Iateriﬂora L. 9I9I2007 Ref 1
138 Asteraceae Senecio aureus L. 5/15/2008 Ref 1
180 Asteraceae Senecio pauperculus Michx. 6I5I2009 Ref 2
79 Liliaceae Smilacina stellata (L.) Desf. 6I4I2008 Block 1
47B Smilacaceae Smilax tamnoides L. 9/12/2008 Block 3
158 Solanaceae Solanum dulcamara L. 5/27/2009 Block 1 cleared
220 Asteraceae Solidago altissima L. 8/24/2009 Ref1
59 Asteraceae Solidago gigantea Aiton 6I2/2008 Ref 3
228 Asteraceae Solidago ohioensis Frank ex Riddell 8/24/2009 Ref 2
241 Asteraceae Solidago patula Muhl. ex Willd. 9/10/2009 Ref 1
240 Asteraceae Solidago ride/Iii (Frank ex Riddell) Rydb. 9/10I2009 Ref 2
211 Asteraceae Solidago rugosa Mill. 8/17/2009 Block 2 cleared
226 Poaceae Sorghastrum nutans (L.) Nash 8124/2009 Ref 2
121 Poaceae Sphenopholis intermedia (Rydb.) Rydb. 7l30/2008 Block 3 cleared
238 Rosaceae Spiraea alba Du Roi 9l10/2009 Ref 3
181 Araceae Symp/ocarpus foetidus (L.) Salisb. ex Nutt. 6I5I2009 Block 1 uncleared
233 Asteraceae Taraxacum oﬂicinale F.H. Wigg. 8/24/2009 Upland near ref 1
182 Ranunculaceae Thalictrum dasycarpum Fisch. 8r Avé-Lali. 6/15/2009 Ref 2
232 Thelypteridaceae Thelypteris palustris Schott 8/24/2009 near Ref 1
114 Clusiaceae Tn'adenum fraseri (Spach) Gleason 7/28I2008 Ref 3
343 Liliaceae Triantha glutinosa (Michx.) Baker 9I11/2007 Ref 2
231 Typhaceae Typha sp. 8/24/2009 Block 6 cleared
222 Ulmaceae Ulmus sp. 3/24/2009 3:: 5:32:13"
212 Urticaceae Urtica dioica L. 8/17/2009 Block 2 cleared
184 Verbenaceae Verbena hastata L. 7/31/2009 Block 1 cleared
147 Verbenaceae Verbena urticifolia L. 9/7/2008 Block 2
139 Violaceae Viola nephrophylla Greene 5/9/2008 Block 5
201 Vltaceae Wis n’paria Michx. 8/6/2009 Block 2 cleared
192 Liliaceae Zigadenus eIegans Pursh ssp. glaucus 8 I6 [2009 Ref 3

(Nutt.) Hultén

 

184

 

Appendix B

 

 

Plant Voucher Data
Page 5 of 5 Pages
Collecti ‘ Collection
on it Family Scientific Name . Date location
88 Cyperaceae Eleocharis rostellata (T orr.) Torr. 6I4I2008 Grand River Fan
81 Cyperaceae Sgrq‘paqgmm vmdrcannatum (Engelm.) 6I4I2008 Grand River Fen
85 Poaceae Hierochloe odorata (L.) P. Beauv. 6/4/2008 Grand River Fen
87 Poaceae Muhlenbergia n‘chardsonis (T rin.) Rydb. 6I4I2008 Grand River Fen
136 Saxifragaceae Parnassia glauca ' 8/14/2008 Grand River Fen
84 Rosaceae gityssocatpus opullfollus (L') Maxim, orth. 6I4I2008 Grand River Fen
. . Pycnanthemum virginianum (L.) T. Dur. &
255 Lamiaceae B.D. Jacks. ex B.L. R ob. & Fe mal d 7/16/2009 Grand River Fen
256 Rhamnaceae Rhamnus alnifolia L'Hér. 7/16/2009 Grand River Fen

82 Valerianaceae Valeriana uliginosa (Torr. 8 A. Gray) Rydb. 6/4/2008 Grand River Fen

All specimens on pages 1-4 collected: USA Michigan, Jackson
County, Clarklake, MacCready Reserve. 42° 7’37N 84° 23’30W.
Elevation 300m, in prairie fen wetland. Collection date noted
above.

All specimens on page 5 collected: USA Michigan, Jackson
County, Liberty, Grand River Fen. 42° 5’02N 84° 27’38W.
Elevation 300m, in prairie fen wetland. Collection date noted
above.

Investigator’s Name:
Anna K. F iedler

Date: 24 August, 2010
Received the above listed specimens for
deposit in the Michigan State University

may: aw...

CuratorV Date

185

 

Appendix G] Record of Deposition of Insect Voucher Specimens*
The specimens listed on the following sheet(s) have been deposited in the named
museum(s) as samples of those species or other taxa, which were used in this research.

Voucher recognition labels bearing the Voucher No. have been attached or included in
ﬂuid-preserved specimens.

Voucher No.: 2010-04

Title of thesis or dissertation (or other research projects):

RESTORATION OF PRAIRIE F EN PLANTS, INSECTS, AND ABIOTIC
PROCESSES

Museum(s) where deposited and abbreviations for table on following sheets:

Entomology Museum, Michigan State University (MSU)

Other Museums:

Investigator’s Name(s) (typed)
Anna Katherine Fiedler

 

 

Date 24 August 2010

*Reference: Yoshimoto, C. M. 1978. Voucher Specimens for Entomology in North
America.
Bull. Entomol. Soc. Amer. 24: 141-42.

Deposit as follows:
Original: Include as Appendix 1 in ribbon copy of thesis or dissertation.

Copies: Include as Appendix 1 in copies of thesis or dissertation.
Museum(s) ﬁles.
Research project ﬁles.

This form is available ﬁ'om and the Voucher No. is assigned by the Curator, Michigan
. State University Entomology Museum.

186

 

Appendix C
Insect Voucher Specimen Data

Page 1 of 6 Pages

 

 

Number of.
.5
Species or other taxon Date collected 3
Hemiptera
Pentatomidae Acrostemum hilare 27 July 2009. food web#66 1

Reduviidae Zelus Iuridus (Stal)
Miridae Neurocolpus nubilus (Say)
Phymatidae Phymata pennsylvanica
(Handlirsch)

Phymatidae Phymata pennsylvanica
(Handlirsch)

Thyreocoridae Corimelaena pulicaria
Thyreocoridae Corimelaena pulicaria

Hymenoptera
Vespidae Dolichovespula arenaria (Fabricius)

Vespidae Vespula vidua

Vespidae Dolichovespula maculata
Vespidae Dolichovespula maculata
Vespidae Polistes fuscatus (Fabricius)

Braconidae

ichneumonidae

Crabrionldae Ectemnius tn'fasciatus (Say)
Crabrionidae Ectemnius continuus (Fabricius)
Crabrionidae Tachytes auru/entus (Fabricius)
Sieroiomorphidae

Tiphiidae Myznum sp.

Vespidae Ancistrocerus anti/ope anti/ope
(Panzer)

Vespidae Ancllstrocerus sp.

Vespidae Eumenes fratemus (Say)
Vespidae Ancistmcerus catskil
albopha/eralus (Saussure)

Vespidae Euodynerus foraminotum (Panzer)
Vespidae Stenodynerus sp.

19 August 2009, food web #235
5 July 2009,_food web #25, #34

27 July 2009. food web #54
19 August 2009. food web #137

19 August 2009. food web #213
27 July 2009. food web #80

 

1 September 2009. food web
#296, #268

1 September 2009. food web
#313, #299

19 August 2009. food web #1 89
1 Septemeber 2009, food web

1 September 2009. food web #275
19 August 2009, food web #143.
#168, #209

1 September 2009. food web
#320, #328

15 ,July 2008

27 July 2009. food web #77

20 July 2009

15 July 2008

19 August 2009. food web #190

19 August 2009, food web #141

19 August 2009, food web #222.
I#150

19 August 2009, food web #132

19 August 2009. food web #140
1 September 2009, food web #252
19 Amust 2009. food web #158

 

 

 

 

4.3....”

Add-LAN

 

 

*Aii specimens collected: USA Michigan, Jackson County, Clarklake, MacCready Reserve.

42° 7'37N 84° 23’30W. Elevation 300m, in prairie fen wetland. Date noted above.

Investigator: Anna K Fiedler

Ento

 

 

urat-

 

Date 24-Aug_-_10

depositlntheMichigan r: ; - ' ; ;
[Ha ,VYM "a
’ V

Received the above listed specimens for

lty

1'.— ’l ‘ .

Date

wro—

V- : ernumber 2010-04

187

I

 

 

Appendix C
Insect Voucher Specimen Data

Page 2 of 6 Pages

 

 

Number of:

5 s a
Species or other taxon Date collected g g g:

3"
(I)

Hymenoptera cont.
FormicidaeDoIichoderus pustulatus (Mayr) 15 June 2009. food web #22 1
FormicidaeDoIichoderus pustulatus (Mayr) 19 August 2009, food web #230 1
Formicidae Camponotus novaeboracensis 17 July 2008 1
FormicidaeCamponotus novaeboracensis 14 July 2008 1
FormicidaeCamponotus caryae (Fitch) 17 July 2008 1
FormicidaeCamponotus caryae (Fitch) 24 July 2009 1
FormicidaeFormica fusca (Linneaus) 24 July 2009 1
Formicidae Formica glacialis (Wheeler) 19 August 2009, food web #215 1
Formicidae Formica glacialis (Wheeler) 27 July 2009, food web #98 1
Apidae Xylocopa virginica (Linneaus) 31 August 2009 1
Anthophoridae Melissodes ur. illata 19 August 2009, food web #160 1
Halictidae Lasioglossum tegulare (Robt) 6 July 2009, food web #52 1
Halictidae Lasioglossum nelumbonis
(Robertson) 6 June 2008 1
Halictidae Lasioglossum nymphaearum (Robt. 5 August 2009 1
Halictidae Lasioglossum paradmirandum
(K. & At.) 8 August 2008 1
Halictidae Lasioglossum coriaceum (Smith) 31 August 2009 2
Halictidae Lasioglossum cressonii (Robertsonl3 July 2009 2
Halictidae Lasioglossum (Lasioglossum)
leucozonium (Schrank) 12 June 2008 1
Halictidae Lasioglossum (Lasioglossum)
leucozonium (Schrank) 6 September 2008 1
Halictidae Lasioglossum lineatulum (Crawford 19 August 2009, food web #229 1
Halictidae Lasioglossum lineatulum (Crawford 1 September 2009, food web #298 1
Halictidae Lasioglossum pectorale (Smith) 4 June 2009 1
Halictidae Lasioglossum pectorale (Smith) 3 July 2009 1
Halictidae Lasioglossum pilosum (Sm.) 31 August 2009 2
Halictidae Lasioglossum (Dialictus) rohweri 6 September 2008 3'
Halictidae Lasioglossum versans (Crawford) 6 September 2008 2
Halictidae Lasioglossum (Dialictus) sp. 2 4 June 2009 1
Halictidae Lasioglossum sp. 2 4 June 2009 2
Halictidae Lasioglossum (Dialictus) sp. 1 8 August 2008 1
Halictidae Lasioglossum (Dialictus) sp. 1 6 September 2008 1
Halictidae Lasioglossum (Dialictus) sp. 4 3 July 2009 1
Halictidae Lasioglossum (Dialictus) sp. 5 5 August 2009 1
Halictidae Lasioglossum (Dialictus) sp. 6 4 June 2009 1
Halictidae Lasioglossum (Dialictus) sp. 7 4 June 2009 1
Halictidae Lasioglossum (Dialictus) sp. 8 5 August 2009 1
Halictidae Lasioglossum (Dialictus) sp. 9 5 Agust 2009 1

 

 

 

 

 

 

 

*All specimens collected: USA Michigan. Jackson County. Clarklake, MacCready Reserve.

42° 7’37N 84° 23’30W. Elevation 300m. in prairle fen wetland. Date noted above.

Anna K. Fiedler. Voucher number 2010-04

188

 

 

Appendix C
Insect Voucher Specimen Data

Page 3 of 6 Pages

 

 

Number of:

z; 2 5
Species or other taxon Date colbcted g, a g

m

Hymenoptera cont.
Megachilidae Heriades carinatus (Cresson) 5 August 2009 2
Megachilidae Heriades leavitti (Crawford) 19 August 2009, food web #149 1
Megachilidae Heriades variolosus (Cresson) 5 August 2009 1
Megachilidae Hoplitis pruducta (Cresson) 4 June 2009 1
Megachilidae Hop/itis spoliata (Provancher) 4 June 2009 1
Megachilidae Megachile campanulae
(Robertson) 3 July 2009 1
Megachilidae Megachile inermis (Provancher) 31 August 2009 1
Megachilidae Megachile monﬂvaga (Cresson) 4 June 2009 1
Megachilidae Megachile pugnata (Say) 8 August 2008 1
Megachilidae Megachile relativa (Cresson) 3 July 2009 1
Megachilidae Osmia georgica (Cresson) 4 June 2009 2
Megachilidae Osmia michiganensis (Mitchell) 4 June 2009 1
Megachilidae Osmia pumila (Cresson) 12 June 2008 2
Megachilidae Osmia simillima (Smith) 12 June 2009 1
Megachilidae Coelioxys modesta (Smith) 27 July 2009, food web #58 1
Andrenidae Andrena alleghaniensis (Viereck) 4 June 2009 1
Andrenidae Andrena car/ini (Cockereli) 4 June 2009 1
Andrenidae Andrena cressonii (Robt.) 4 June 2009 1
Andrenidae Andrena nasonii (Robertson) 4 June 2009 1
Andrenidae Andrena (Tylandrena) perplexa
(Sm.) 12 June 2008 1
Andrenidae Andrena placate (Mitchell) 5 August 2009 1
Andrenidae Andrena (Trachandrena)
virginiana (Mitchell) 5 August 2009 1
Andrenidae Andrena (Cnemiandrena)
hirticincta (Prov.) 19 August 2009, food web #136 1
Apidae Anthophora terminalis (Cresson) 4 June 2009 1
Apidae Anthophora terminalis (Cr.) 6 July 2009, food web #36 1
Apidae Anthophora ursina (Cresson) 5 August 2009 1
Apidae Apis mellifera L. 31 August 2009 2
Apidae Bombus impatiens (Chandler) 31 August 2009 2
Apidae Bombus citn'nus (Smith) 5 August 2009 1
Apidae Bombus bimaculatus (Psithyrus) 3 July 2009 1
Apidae Bombus fevidus (Fabricius) 19 August 2009, food web #180 1
Apidae Ceratina calcarala/dupla (Robertson) 3 July 2009 2
Apidae Ceratina dupla 4 June 2009 1
Apidae Ceratina calcarata (Robertson) 31 August 2009 1
Apidae Ceratina strenua (Smith) 6 September 2008 2
Colletidae Hylaeus aﬂinis (Smith) 4 June 2009 1
Colletidae Hylaeus sp. 1 31 August 2009 2
Colletidae Hylaeus sp. 3 3 July 2009 _l ‘

 

 

 

;—
‘All specimens collected: USA Michigan, Jackson County, Clarklake, MacCready Reserve.
42° 7’37N 84° 23'30W. Elevation 300m, in prairie fen wetland. Date noted above.

Anna K. Fiedler. Voucher number 2010-04

189

 

 

 

 

 

Appendix C
Insect Voucher Specimen Data

Page 4 of 6 Pages

 

 

Halictidae Lassioglossum coeruleum (Robt.)
Halictidae Lassioglossum divergens (Lovell)
Halictidae Lassioglossum illinoense (Robt.)
Halictidae Lassioglossum macovpinense
Lepidoptera

Hesperiidae Euphyes conspicuus (W.H.
Hesperiidae Euphyes conspicuus (W.H.
Hesperiidae Polites mystic (W.H. Edwards)
Nymphalidae Phyciodes tharos (Drury)
Lycaenidae L ycaena helloides (Boisduval)
Hesperiidae Atrytone Iogan (W.H. Edwards)
Yponom eutidae Atteva punctellan (Cramer)
Nym phalidae Polygonia progne (Cramer)
Hesperiidae Euphyes origenes (Fabricius)
Hesperiidae Euphyes conspicuus (W.H.
Hesperiidae Euphyes vestis metacomet (Harri
Hesperiidae Euphyes vestis metacomet (Harri
Hesperiidae Poanes Massasoit (Scudder)
Pieridae Pieris rapae (Linnaeus)

Pieridae Pieris rapae (Linnaeus)

Satyridae Megisto cymela (Cramer)
Papilionidae Pepi/lo troilus (Linnaeus)
Hesperiidae Ancyloxypha numitor (Fabricius)
Hesperiidae Ancyloxypha numitor (Fabricius)
Hesperiidae Epargyreus Clarus (Cramer)
Nymphalidae Boloria bellona (Fabricius)
Nymphalidae Boloria selene (Fabricius)

Number of:
S E 2:
Species or other taxon Date collected é 13 gr
to
Hymenoptera cont.
Colletidae Hylaeus sp. 2 5 August 2009 1
Halictidae Agapostemon sericeus (Forster) 4 June 2009 1
Halictidae Agapostemon sericeus (Forster) 5 August 2009 1
Halictidae Agapostemon virescens (Fabricius) 6 September 2008 1
Halictidae Agochlora pura (Say) 31 August 2009 2
Halictidae Augochlorella aurata (Smith) 4 June 2009 2
Halictidae Augochloropsis meta/lice (Fabricius4 June 2009 2
Halictidae Halictus confusus (Smith) 5 August 2009 2
Halictidae Halictus Iigatus 4 June 2009 2
Halictidae Halictus parallelus (Say) 6 June 2008 1
Halictidae Halictus rubicundus 19 August 2009, food web #177 1
Halictidae Lassioglossum atlanticum (Mitchell) 6 September 2008 2
Halictidae Lassioglossum bruneri (Crawford) 31 August 2009 1
1
1
1
1

 

15 June 2009, food web #5
12 June 2008

31 August 2009

4 June 2009

10 July 2008

17 July 2008

19 August 2009, food web #165
4 June 2009

12 June 2008

3 July 2009

19 August 2009, food web #174
6 September 2008

4 June 2009

10 July 2009

5 August 2009

14 July 2008

3 July 2009

8 August 2008

6 September 2008

4 June 2009

31 August 2009

12 June 2008

15 June 2009. food web #7
31 August 2009

5 August 2009

6 September 2008

Nymphalidae Speyeria cybele cybele(Fabriciu12 July 2008

 

 

 

 

*All specimens collected: USA Michigan, Jackson County. Clarklake, MacCready Reserve.

42° 7'37N 84° 23'30W. Elevation 300m, In prairie fen wetland. Date noted above.

Anna K. Fiedler. Voucher number 2010-04

190

 

 

 

Appendix C
Insect Voucher Specimen Data

Page 5 of 6 Pages

 

 

Number of:

i? E Z
Species or other taxon Date collected 5. 13, g»

t a
Coleoptera
Mordellidae Mordella marginata (Melsheimer) 27 July 2009, food web #79, #91 2
Falsomordellistena pubescens (Fabricius) 24 July 2009 2
Mordellidae Mordella atrata (Melsheimer) 6 July 2009, food web #27, #28 2
Scarabaeidae Trichiotinus aﬂinus 15 July 2008 1
Cerambycidae Euderces picipes (Fabricius) 17 July 2008 1
Cerambycidae Euderces picipes (Fabricius) 15 June 2009, food web #1 1
Cerambycidae Typocerus velutinus (Olivier) 14 July 2008 1
Coccinellidae Harmonia axyridis (Pallas) 19 August 2009, food web #183 1
Chrysomelidae Ophraella notata 27 July 2009. food web #81 1
Chrysomelidae Diabrotica undecimpunctata
(Barber) 1 September 2009, food web #319 1
Lampyridae Pyropyga sp. 15 June 2009, food web #14 1
Lampyridae Pyropyga sp. 16 July 2009 1
Cantharidae Chauliguathus marginatus 17 July 2008 1

Diptera

Syrphidae Erlstalis transverse (Wiedemann)
Syrphidae Eristalis ﬂavipes (Walker)
Syrphidae Sericomyia chrysotoxoides
Syrphidae Helophilus fasciatus (Macquart)
Syrphidae Pipiza femorata (Loew)

Syrphidae Tropodia quadrata (Say)
Syrphidae Neoascia meta/lice (Williston)
Syrphidae Syrphus ribesii (Linnaeus)
Syrphidae Syrphus rectus (Osten Sacken)
Syrphidae Melanostoma mellinum (Linnaeus)
Syrphidae Syrphus vitripennis (Meigen)
Syrphidae Toxomarginatus

Syrphidae Toxomarginatus

Syrphidae Toxomarginatus genninatus
Syrphidae Sphaerophoria contigua (Macquart)
Stratiomyidae Odontomyia Virgo
Anthromyiidae Hylemya platura (Meigen)
Muscidae

Tachinidae Genoa texensis (Townsend)
Tachinidae Gymnosoma occidua (Walker)
Tachinidae Jurinopsis adjusta (Van der Wulp)
Bombyliidae Anthrax Pluto

Bombyliidae Villa sp.

Stratiomyidae Nemotelus sp.

 

Tachinidae Gymnosoma occidua (Walker)

19 August 2009, food web #129
19 August 2009, food web #139
19 August 2009, food web #172
19 August 2009, food web #228
20 July 2009

17 July 2008

27 July 2009, food web #93

19 August 2009, food web #128.
19 August 2009, food web #173,

1 September 2009, food web #323
1 September 2009, food web #266
14 July 2008

17 July 2008

19 August 2009, food web #126

1 September 2009, food web #330
15 June 2009, food web #13

1 September 2009, food web #322
24 July 2009

1 September 2009. food web #287
19 August 2009, food web #175
19 August 2009, food web #186
24 July 2009

19 August 2009, food web #223
15 June 2009, food web #15, #16
1 September 2009, food web #312

 

 

 

 

 

*All specimens collected: USA Michigan, Jackson County, Clarklake, MacCready Reserve.

42° 7'37N 84° 23'30W. Elevation 300m, in prairie fen wetland. Date noted above.

Anna K. Fiedler. Voucher number 2010-04

191

 

 

Appendix C
insect Voucher Specimen Data

Page 6 of 6 Pages

 

 

Number of:

5 z 2:
Species or other taxon Date collected é a g.-

tn
Hemiptera
Lygaeidae Lygaeus kalmii (Stal) 14 July 2008 1
Miridae Plagiognathus 27 July 2009, food web #116, #1 17 2
Miridae Plagiognathus 19 August 2009, food web #233 1
Hymenoptera
Vespidae Polistes fuscatus (Fabricius) 1 September 2009, food web #355 1
1 September 2009, food web

Formica ulkei (Emery) #348, #358 2
Formica fusca (Linneaus) 17 July 2008 1
Halictidae Duforea marginata (Cresson) 5 August 2009 1
Megachilidae Megachile relativa (Cresson) 31 August 2009 1
Andrenidae Andrena (Callandrena)
runcinatae Ckrll. 31 August 2009 1
Andrenidae Perdita octomaculata (Say) 31 August 2009 2
Andrenidae Pseudopanurgus nubruscensis
(Crawford) 1 September 2009. food web #338 1
Apidae Bombus griseocollis (DeGeer) 14 August 2008 1
Lepidoptera
Hesperiidae Polites mystic (W.H. Edwards) 5 August 2009 1
Hesperiidae Euphyes dion (W.H. Edwards) 14 July 2008 1
Coleoptera
Cantharidae Polemius Iaticornis (Say) 19 August 2009. food web #237 1
Cantharidae Polemius canadensis (Brown) 19 August 2009, food web #240 1
Rhagonycha luteicollis (Gerrnar) 16 July 2009 1
Rhagonycha luteicollis (Gennar) 24 July 2009 1
Diptera
Syrphidae Sphaerophon‘a sp. 15 June 2009. food web #23 1
Tachinidae 24 July 2009 1
Stratiomyidae Odontomyia sp. 10 July 2008 1
Calliphoridae 17 July 2008 1
Tachinidae 24 July 2009 1
Tachinidae 17 July 2008 1
Tephritidae 15 June 2009, food web #24 1

 

 

 

 

 

‘All specimens collected: USA Michigan, Jackson County, Liberty, Grand River Fen.
42° 5’02N 84° 27'38W. Elevation 300m, in prairle fen wetland. Date noted above.

Anna K. Fiedler. Voucher number 2010-04

192